/[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 608 by ph10, Sun Jun 12 16:25:55 2011 UTC revision 888 by ph10, Tue Jan 17 14:43:23 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, RM64 };         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    #define save_mark     data
630    
631  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
632  variables. */  variables. */
# Line 667  defined). However, RMATCH isn't like a f Line 652  defined). However, RMATCH isn't like a f
652  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,
653  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
654    
655  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
656  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
657  #else  #else
658  utf8 = FALSE;  utf = FALSE;
659  #endif  #endif
660    
661  /* 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 665  if (md->match_call_count++ >= md->match_
665  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);
666    
667  /* 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
668  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
669  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
670  up space on the stack. See also MATCH_CONDASSERT below.  up space on the stack. See also MATCH_CONDASSERT below.
671    
672  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 690  for (;;)
690    {    {
691    minimize = possessive = FALSE;    minimize = possessive = FALSE;
692    op = *ecode;    op = *ecode;
693    
694    switch(op)    switch(op)
695      {      {
696      case OP_MARK:      case OP_MARK:
697      markptr = ecode + 2;      md->nomatch_mark = ecode + 2;
698      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      md->mark = NULL;    /* In case previously set by assertion */
699        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
700        eptrb, RM55);        eptrb, RM55);
701        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
702             md->mark == NULL) md->mark = ecode + 2;
703    
704      /* 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
705      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 708  for (;;)
708      position and return MATCH_SKIP. Otherwise, pass back the return code      position and return MATCH_SKIP. Otherwise, pass back the return code
709      unaltered. */      unaltered. */
710    
711      if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
712          strcmp((char *)markptr, (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC(ecode + 2, md->start_match_ptr) == 0)
713        {        {
714        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
715        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
716        }        }
   
     if (md->mark == NULL) md->mark = markptr;  
717      RRETURN(rrc);      RRETURN(rrc);
718    
719      case OP_FAIL:      case OP_FAIL:
720      MRRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
721    
722      /* COMMIT overrides PRUNE, SKIP, and THEN */      /* COMMIT overrides PRUNE, SKIP, and THEN */
723    
724      case OP_COMMIT:      case OP_COMMIT:
725      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
726        eptrb, RM52);        eptrb, RM52);
727      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&
728          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&
729          rrc != MATCH_THEN)          rrc != MATCH_THEN)
730        RRETURN(rrc);        RRETURN(rrc);
731      MRRETURN(MATCH_COMMIT);      RRETURN(MATCH_COMMIT);
732    
733      /* PRUNE overrides THEN */      /* PRUNE overrides THEN */
734    
735      case OP_PRUNE:      case OP_PRUNE:
736      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
737        eptrb, RM51);        eptrb, RM51);
738      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
739      MRRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
740    
741      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
742      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      md->nomatch_mark = ecode + 2;
743        md->mark = NULL;    /* In case previously set by assertion */
744        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
745        eptrb, RM56);        eptrb, RM56);
746        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
747             md->mark == NULL) md->mark = ecode + 2;
748      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
     md->mark = ecode + 2;  
749      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
750    
751      /* SKIP overrides PRUNE and THEN */      /* SKIP overrides PRUNE and THEN */
752    
753      case OP_SKIP:      case OP_SKIP:
754      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
755        eptrb, RM53);        eptrb, RM53);
756      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
757        RRETURN(rrc);        RRETURN(rrc);
758      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
759      MRRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
760    
761        /* Note that, for Perl compatibility, SKIP with an argument does NOT set
762        nomatch_mark. There is a flag that disables this opcode when re-matching a
763        pattern that ended with a SKIP for which there was not a matching MARK. */
764    
765      case OP_SKIP_ARG:      case OP_SKIP_ARG:
766      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      if (md->ignore_skip_arg)
767          {
768          ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
769          break;
770          }
771        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
772        eptrb, RM57);        eptrb, RM57);
773      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
774        RRETURN(rrc);        RRETURN(rrc);
775    
776      /* 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
777      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
778      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
779      as PRUNE. */      with the md->ignore_skip_arg flag set. */
780    
781      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
782      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
783    
784      /* 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
785      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
786      to skip back past alternatives that precede the THEN within the current      match pointer to do this. */
     branch. */  
787    
788      case OP_THEN:      case OP_THEN:
789      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
790        eptrb, RM54);        eptrb, RM54);
791      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
792      md->start_match_ptr = ecode - GET(ecode, 1);      md->start_match_ptr = ecode;
793      MRRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
794    
795      case OP_THEN_ARG:      case OP_THEN_ARG:
796      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1+LINK_SIZE],      md->nomatch_mark = ecode + 2;
797        offset_top, md, eptrb, RM58);      md->mark = NULL;    /* In case previously set by assertion */
798        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
799          md, eptrb, RM58);
800        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
801             md->mark == NULL) md->mark = ecode + 2;
802      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
803      md->start_match_ptr = ecode - GET(ecode, 1);      md->start_match_ptr = ecode;
     md->mark = ecode + LINK_SIZE + 2;  
804      RRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
805    
806        /* Handle an atomic group that does not contain any capturing parentheses.
807        This can be handled like an assertion. Prior to 8.13, all atomic groups
808        were handled this way. In 8.13, the code was changed as below for ONCE, so
809        that backups pass through the group and thereby reset captured values.
810        However, this uses a lot more stack, so in 8.20, atomic groups that do not
811        contain any captures generate OP_ONCE_NC, which can be handled in the old,
812        less stack intensive way.
813    
814        Check the alternative branches in turn - the matching won't pass the KET
815        for this kind of subpattern. If any one branch matches, we carry on as at
816        the end of a normal bracket, leaving the subject pointer, but resetting
817        the start-of-match value in case it was changed by \K. */
818    
819        case OP_ONCE_NC:
820        prev = ecode;
821        saved_eptr = eptr;
822        save_mark = md->mark;
823        do
824          {
825          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
826          if (rrc == MATCH_MATCH)  /* Note: _not_ MATCH_ACCEPT */
827            {
828            mstart = md->start_match_ptr;
829            break;
830            }
831          if (rrc == MATCH_THEN)
832            {
833            next = ecode + GET(ecode,1);
834            if (md->start_match_ptr < next &&
835                (*ecode == OP_ALT || *next == OP_ALT))
836              rrc = MATCH_NOMATCH;
837            }
838    
839          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
840          ecode += GET(ecode,1);
841          md->mark = save_mark;
842          }
843        while (*ecode == OP_ALT);
844    
845        /* If hit the end of the group (which could be repeated), fail */
846    
847        if (*ecode != OP_ONCE_NC && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);
848    
849        /* Continue as from after the group, updating the offsets high water
850        mark, since extracts may have been taken. */
851    
852        do ecode += GET(ecode, 1); while (*ecode == OP_ALT);
853    
854        offset_top = md->end_offset_top;
855        eptr = md->end_match_ptr;
856    
857        /* For a non-repeating ket, just continue at this level. This also
858        happens for a repeating ket if no characters were matched in the group.
859        This is the forcible breaking of infinite loops as implemented in Perl
860        5.005. */
861    
862        if (*ecode == OP_KET || eptr == saved_eptr)
863          {
864          ecode += 1+LINK_SIZE;
865          break;
866          }
867    
868        /* The repeating kets try the rest of the pattern or restart from the
869        preceding bracket, in the appropriate order. The second "call" of match()
870        uses tail recursion, to avoid using another stack frame. */
871    
872        if (*ecode == OP_KETRMIN)
873          {
874          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM65);
875          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
876          ecode = prev;
877          goto TAIL_RECURSE;
878          }
879        else  /* OP_KETRMAX */
880          {
881          md->match_function_type = MATCH_CBEGROUP;
882          RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
883          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
884          ecode += 1 + LINK_SIZE;
885          goto TAIL_RECURSE;
886          }
887        /* Control never gets here */
888    
889      /* Handle a capturing bracket, other than those that are possessive with an      /* Handle a capturing bracket, other than those that are possessive with an
890      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
891      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
892      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
893      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
894      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,
895      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
896      of the same bracket.      the working value and also the values of the final offsets, in case they
897        were set by a previous iteration of the same bracket.
898    
899      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
900      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 904  for (;;)
904      case OP_SCBRA:      case OP_SCBRA:
905      number = GET2(ecode, 1+LINK_SIZE);      number = GET2(ecode, 1+LINK_SIZE);
906      offset = number << 1;      offset = number << 1;
907    
908  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
909      printf("start bracket %d\n", number);      printf("start bracket %d\n", number);
910      printf("subject=");      printf("subject=");
# Line 834  for (;;) Line 918  for (;;)
918        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
919        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
920        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
921          save_mark = md->mark;
922    
923        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
924        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 841  for (;;) Line 926  for (;;)
926    
927        for (;;)        for (;;)
928          {          {
929          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
930          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
931            eptrb, RM1);            eptrb, RM1);
932          if (rrc != MATCH_NOMATCH &&          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
933              (rrc != MATCH_THEN || md->start_match_ptr != ecode))  
934            RRETURN(rrc);          /* If we backed up to a THEN, check whether it is within the current
935            branch by comparing the address of the THEN that is passed back with
936            the end of the branch. If it is within the current branch, and the
937            branch is one of two or more alternatives (it either starts or ends
938            with OP_ALT), we have reached the limit of THEN's action, so convert
939            the return code to NOMATCH, which will cause normal backtracking to
940            happen from now on. Otherwise, THEN is passed back to an outer
941            alternative. This implements Perl's treatment of parenthesized groups,
942            where a group not containing | does not affect the current alternative,
943            that is, (X) is NOT the same as (X|(*F)). */
944    
945            if (rrc == MATCH_THEN)
946              {
947              next = ecode + GET(ecode,1);
948              if (md->start_match_ptr < next &&
949                  (*ecode == OP_ALT || *next == OP_ALT))
950                rrc = MATCH_NOMATCH;
951              }
952    
953            /* Anything other than NOMATCH is passed back. */
954    
955            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
956          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
957          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
958          if (*ecode != OP_ALT) break;          md->mark = save_mark;
959            if (*ecode != OP_ALT) break;
960          }          }
961    
962        DPRINTF(("bracket %d failed\n", number));        DPRINTF(("bracket %d failed\n", number));
   
963        md->offset_vector[offset] = save_offset1;        md->offset_vector[offset] = save_offset1;
964        md->offset_vector[offset+1] = save_offset2;        md->offset_vector[offset+1] = save_offset2;
965        md->offset_vector[md->offset_end - number] = save_offset3;        md->offset_vector[md->offset_end - number] = save_offset3;
966    
967        if (rrc != MATCH_THEN) md->mark = markptr;        /* At this point, rrc will be one of MATCH_ONCE or MATCH_NOMATCH. */
968        RRETURN(MATCH_NOMATCH);  
969          RRETURN(rrc);
970        }        }
971    
972      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
# Line 873  for (;;) Line 980  for (;;)
980      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
981      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
982    
983      /* Non-capturing bracket, except for possessive with unlimited repeat. Loop      /* Non-capturing or atomic group, except for possessive with unlimited
984      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.
985      brackets, we would return the result of a recursive call to match()  
986      whatever happened. We can reduce stack usage by turning this into a tail      When we get to the final alternative within the brackets, we used to return
987      recursion, except in the case of a possibly empty group.*/      the result of a recursive call to match() whatever happened so it was
988        possible to reduce stack usage by turning this into a tail recursion,
989        except in the case of a possibly empty group. However, now that there is
990        the possiblity of (*THEN) occurring in the final alternative, this
991        optimization is no longer always possible.
992    
993        We can optimize if we know there are no (*THEN)s in the pattern; at present
994        this is the best that can be done.
995    
996        MATCH_ONCE is returned when the end of an atomic group is successfully
997        reached, but subsequent matching fails. It passes back up the tree (causing
998        captured values to be reset) until the original atomic group level is
999        reached. This is tested by comparing md->once_target with the start of the
1000        group. At this point, the return is converted into MATCH_NOMATCH so that
1001        previous backup points can be taken. */
1002    
1003        case OP_ONCE:
1004      case OP_BRA:      case OP_BRA:
1005      case OP_SBRA:      case OP_SBRA:
1006      DPRINTF(("start non-capturing bracket\n"));      DPRINTF(("start non-capturing bracket\n"));
1007    
1008      for (;;)      for (;;)
1009        {        {
1010        if (ecode[GET(ecode, 1)] != OP_ALT)   /* Final alternative */        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;
1011    
1012          /* If this is not a possibly empty group, and there are no (*THEN)s in
1013          the pattern, and this is the final alternative, optimize as described
1014          above. */
1015    
1016          else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1017          {          {
1018          if (op >= OP_SBRA)   /* Possibly empty group */          ecode += PRIV(OP_lengths)[*ecode];
           {  
           md->match_function_type = MATCH_CBEGROUP;  
           RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,  
             RM48);  
           if (rrc == MATCH_NOMATCH) md->mark = markptr;  
           RRETURN(rrc);  
           }  
         /* Not a possibly empty group; use tail recursion */  
         ecode += _pcre_OP_lengths[*ecode];  
         DPRINTF(("bracket 0 tail recursion\n"));  
1019          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1020          }          }
1021    
1022        /* For non-final alternatives, continue the loop for a NOMATCH result;        /* In all other cases, we have to make another call to match(). */
       otherwise return. */  
1023    
1024        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        save_mark = md->mark;
1025        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1026          RM2);          RM2);
1027        if (rrc != MATCH_NOMATCH &&  
1028            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* See comment in the code for capturing groups above about handling
1029          THEN. */
1030    
1031          if (rrc == MATCH_THEN)
1032            {
1033            next = ecode + GET(ecode,1);
1034            if (md->start_match_ptr < next &&
1035                (*ecode == OP_ALT || *next == OP_ALT))
1036              rrc = MATCH_NOMATCH;
1037            }
1038    
1039          if (rrc != MATCH_NOMATCH)
1040            {
1041            if (rrc == MATCH_ONCE)
1042              {
1043              const pcre_uchar *scode = ecode;
1044              if (*scode != OP_ONCE)           /* If not at start, find it */
1045                {
1046                while (*scode == OP_ALT) scode += GET(scode, 1);
1047                scode -= GET(scode, 1);
1048                }
1049              if (md->once_target == scode) rrc = MATCH_NOMATCH;
1050              }
1051          RRETURN(rrc);          RRETURN(rrc);
1052            }
1053        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1054          md->mark = save_mark;
1055          if (*ecode != OP_ALT) break;
1056        }        }
     /* Control never reaches here. */  
1057    
1058      /* Handle possessive capturing brackets with an unlimited repeat. We come      RRETURN(MATCH_NOMATCH);
1059    
1060        /* Handle possessive capturing brackets with an unlimited repeat. We come
1061      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
1062      handled similarly to the normal case above. However, the matching is      handled similarly to the normal case above. However, the matching is
1063      different. The end of these brackets will always be OP_KETRPOS, which      different. The end of these brackets will always be OP_KETRPOS, which
1064      returns MATCH_KETRPOS without going further in the pattern. By this means      returns MATCH_KETRPOS without going further in the pattern. By this means
1065      we can handle the group by iteration rather than recursion, thereby      we can handle the group by iteration rather than recursion, thereby
1066      reducing the amount of stack needed. */      reducing the amount of stack needed. */
1067    
1068      case OP_CBRAPOS:      case OP_CBRAPOS:
1069      case OP_SCBRAPOS:      case OP_SCBRAPOS:
1070      allow_zero = FALSE;      allow_zero = FALSE;
1071    
1072      POSSESSIVE_CAPTURE:      POSSESSIVE_CAPTURE:
1073      number = GET2(ecode, 1+LINK_SIZE);      number = GET2(ecode, 1+LINK_SIZE);
1074      offset = number << 1;      offset = number << 1;
# Line 939  for (;;) Line 1083  for (;;)
1083      if (offset < md->offset_max)      if (offset < md->offset_max)
1084        {        {
1085        matched_once = FALSE;        matched_once = FALSE;
1086        code_offset = ecode - md->start_code;        code_offset = (int)(ecode - md->start_code);
1087    
1088        save_offset1 = md->offset_vector[offset];        save_offset1 = md->offset_vector[offset];
1089        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
# Line 947  for (;;) Line 1091  for (;;)
1091        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
1092    
1093        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
1094    
1095        /* Each time round the loop, save the current subject position for use        /* Each time round the loop, save the current subject position for use
1096        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
1097        restart it with a new subject starting position, remembering that we had        restart it with a new subject starting position, remembering that we had
1098        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
1099        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
1100        see if a previous iteration matched. If so, the group has matched;        see if a previous iteration matched. If so, the group has matched;
1101        continue from afterwards. Otherwise it has failed; restore the previous        continue from afterwards. Otherwise it has failed; restore the previous
1102        capture values before returning NOMATCH. */        capture values before returning NOMATCH. */
1103    
1104        for (;;)        for (;;)
1105          {          {
1106          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1107            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1108          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1109          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1110            eptrb, RM63);            eptrb, RM63);
1111          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1112            {            {
1113            offset_top = md->end_offset_top;            offset_top = md->end_offset_top;
1114            eptr = md->end_match_ptr;            eptr = md->end_match_ptr;
1115            ecode = md->start_code + code_offset;            ecode = md->start_code + code_offset;
1116            save_capture_last = md->capture_last;            save_capture_last = md->capture_last;
1117            matched_once = TRUE;            matched_once = TRUE;
1118            continue;            continue;
1119            }            }
1120          if (rrc != MATCH_NOMATCH &&  
1121              (rrc != MATCH_THEN || md->start_match_ptr != ecode))          /* See comment in the code for capturing groups above about handling
1122            RRETURN(rrc);          THEN. */
1123    
1124            if (rrc == MATCH_THEN)
1125              {
1126              next = ecode + GET(ecode,1);
1127              if (md->start_match_ptr < next &&
1128                  (*ecode == OP_ALT || *next == OP_ALT))
1129                rrc = MATCH_NOMATCH;
1130              }
1131    
1132            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1133          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
1134          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
1135          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
1136          }          }
1137    
1138        if (!matched_once)        if (!matched_once)
1139          {          {
1140          md->offset_vector[offset] = save_offset1;          md->offset_vector[offset] = save_offset1;
1141          md->offset_vector[offset+1] = save_offset2;          md->offset_vector[offset+1] = save_offset2;
1142          md->offset_vector[md->offset_end - number] = save_offset3;          md->offset_vector[md->offset_end - number] = save_offset3;
1143          }          }
1144    
       if (rrc != MATCH_THEN) md->mark = markptr;  
1145        if (allow_zero || matched_once)        if (allow_zero || matched_once)
1146          {          {
1147          ecode += 1 + LINK_SIZE;          ecode += 1 + LINK_SIZE;
1148          break;          break;
1149          }          }
1150    
1151        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1152        }        }
1153    
1154      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
1155      as a non-capturing bracket. */      as a non-capturing bracket. */
1156    
# Line 1009  for (;;) Line 1162  for (;;)
1162      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1163      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1164    
1165      /* Non-capturing possessive bracket with unlimited repeat. We come here      /* Non-capturing possessive bracket with unlimited repeat. We come here
1166      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,
1167      without the capturing complication. It is written out separately for speed      without the capturing complication. It is written out separately for speed
1168      and cleanliness. */      and cleanliness. */
1169    
1170      case OP_BRAPOS:      case OP_BRAPOS:
1171      case OP_SBRAPOS:      case OP_SBRAPOS:
1172      allow_zero = FALSE;      allow_zero = FALSE;
1173    
1174      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1175      matched_once = FALSE;      matched_once = FALSE;
1176      code_offset = ecode - md->start_code;      code_offset = (int)(ecode - md->start_code);
1177    
1178      for (;;)      for (;;)
1179        {        {
1180        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1181        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1182          eptrb, RM64);          eptrb, RM48);
1183        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1184          {          {
1185            offset_top = md->end_offset_top;
1186          eptr = md->end_match_ptr;          eptr = md->end_match_ptr;
1187          ecode = md->start_code + code_offset;          ecode = md->start_code + code_offset;
1188          matched_once = TRUE;          matched_once = TRUE;
1189          continue;          continue;
1190          }          }
1191        if (rrc != MATCH_NOMATCH &&  
1192            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* See comment in the code for capturing groups above about handling
1193          RRETURN(rrc);        THEN. */
1194    
1195          if (rrc == MATCH_THEN)
1196            {
1197            next = ecode + GET(ecode,1);
1198            if (md->start_match_ptr < next &&
1199                (*ecode == OP_ALT || *next == OP_ALT))
1200              rrc = MATCH_NOMATCH;
1201            }
1202    
1203          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1204        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1205        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1206        }        }
1207    
1208      if (matched_once || allow_zero)      if (matched_once || allow_zero)
1209        {        {
1210        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1211        break;        break;
1212        }        }
1213      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
1214    
1215      /* Control never reaches here. */      /* Control never reaches here. */
# Line 1053  for (;;) Line 1217  for (;;)
1217      /* Conditional group: compilation checked that there are no more than      /* Conditional group: compilation checked that there are no more than
1218      two branches. If the condition is false, skipping the first branch takes us      two branches. If the condition is false, skipping the first branch takes us
1219      past the end if there is only one branch, but that's OK because that is      past the end if there is only one branch, but that's OK because that is
1220      exactly what going to the ket would do. As there is only one branch to be      exactly what going to the ket would do. */
     obeyed, we can use tail recursion to avoid using another stack frame. */  
1221    
1222      case OP_COND:      case OP_COND:
1223      case OP_SCOND:      case OP_SCOND:
# Line 1065  for (;;) Line 1228  for (;;)
1228    
1229      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (ecode[LINK_SIZE+1] == OP_CALLOUT)
1230        {        {
1231        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
1232          {          {
1233          pcre_callout_block cb;          PUBL(callout_block) cb;
1234          cb.version          = 1;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1235          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1236          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1237    #ifdef COMPILE_PCRE8
1238          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1239    #else
1240            cb.subject          = (PCRE_SPTR16)md->start_subject;
1241    #endif
1242          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1243          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1244          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
# Line 1080  for (;;) Line 1247  for (;;)
1247          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1248          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last;
1249          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1250          if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);          cb.mark             = md->nomatch_mark;
1251            if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1252          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1253          }          }
1254        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1255        }        }
1256    
1257      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1100  for (;;) Line 1268  for (;;)
1268        else        else
1269          {          {
1270          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1271          condition =  (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1272    
1273          /* 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
1274          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
1275          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
1276          if any one is set. */          if any one is set. */
1277    
1278          if (!condition && condcode == OP_NRREF && recno != RREF_ANY)          if (!condition && condcode == OP_NRREF)
1279            {            {
1280            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1281            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1282              {              {
1283              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 1122  for (;;) Line 1290  for (;;)
1290    
1291            if (i < md->name_count)            if (i < md->name_count)
1292              {              {
1293              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1294              while (slotB > md->name_table)              while (slotB > md->name_table)
1295                {                {
1296                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1297                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1298                  {                  {
1299                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1300                  if (condition) break;                  if (condition) break;
# Line 1142  for (;;) Line 1310  for (;;)
1310                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1311                  {                  {
1312                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1313                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1314                    {                    {
1315                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1316                    if (condition) break;                    if (condition) break;
# Line 1155  for (;;) Line 1323  for (;;)
1323    
1324          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1325    
1326          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1327          }          }
1328        }        }
1329    
# Line 1172  for (;;) Line 1340  for (;;)
1340        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1341          {          {
1342          int refno = offset >> 1;          int refno = offset >> 1;
1343          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1344    
1345          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1346            {            {
# Line 1186  for (;;) Line 1354  for (;;)
1354    
1355          if (i < md->name_count)          if (i < md->name_count)
1356            {            {
1357            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1358            while (slotB > md->name_table)            while (slotB > md->name_table)
1359              {              {
1360              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1361              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1362                {                {
1363                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1364                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1208  for (;;) Line 1376  for (;;)
1376              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1377                {                {
1378                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1379                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1380                  {                  {
1381                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1382                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1223  for (;;) Line 1391  for (;;)
1391    
1392        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1393    
1394        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1395        }        }
1396    
1397      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1238  for (;;) Line 1406  for (;;)
1406    
1407      else      else
1408        {        {
1409        md->match_function_type = MATCH_CONDASSERT;        md->match_function_type = MATCH_CONDASSERT;
1410        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);
1411        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
1412          {          {
1413            if (md->end_offset_top > offset_top)
1414              offset_top = md->end_offset_top;  /* Captures may have happened */
1415          condition = TRUE;          condition = TRUE;
1416          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
1417          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
1418          }          }
1419        else if (rrc != MATCH_NOMATCH &&  
1420                (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an
1421          assertion; it is therefore treated as NOMATCH. */
1422    
1423          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1424          {          {
1425          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
1426          }          }
# Line 1258  for (;;) Line 1431  for (;;)
1431          }          }
1432        }        }
1433    
1434      /* 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
1435      we can use tail recursion to avoid using another stack frame, except when      use tail recursion to avoid using another stack frame, except when there is
1436      we have an unlimited repeat of a possibly empty group. If the second      unlimited repeat of a possibly empty group. In the latter case, a recursive
1437      alternative doesn't exist, we can just plough on. */      call to match() is always required, unless the second alternative doesn't
1438        exist, in which case we can just plough on. Note that, for compatibility
1439        with Perl, the | in a conditional group is NOT treated as creating two
1440        alternatives. If a THEN is encountered in the branch, it propagates out to
1441        the enclosing alternative (unless nested in a deeper set of alternatives,
1442        of course). */
1443    
1444      if (condition || *ecode == OP_ALT)      if (condition || *ecode == OP_ALT)
1445        {        {
1446        ecode += 1 + LINK_SIZE;        if (op != OP_SCOND)
       if (op == OP_SCOND)        /* Possibly empty group */  
1447          {          {
1448          md->match_function_type = MATCH_CBEGROUP;          ecode += 1 + LINK_SIZE;
1449          RMATCH(eptr, ecode, offset_top, md, eptrb, RM49);          goto TAIL_RECURSE;
         RRETURN(rrc);  
1450          }          }
1451        else goto TAIL_RECURSE;  
1452          md->match_function_type = MATCH_CBEGROUP;
1453          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);
1454          RRETURN(rrc);
1455        }        }
1456      else                         /* Condition false & no alternative */  
1457         /* Condition false & no alternative; continue after the group. */
1458    
1459        else
1460        {        {
1461        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1462        }        }
# Line 1301  for (;;) Line 1483  for (;;)
1483        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1484        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1485        }        }
1486      ecode += 3;      ecode += 1 + IMM2_SIZE;
1487      break;      break;
1488    
1489    
1490      /* 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. */  
1491    
     case OP_ACCEPT:  
1492      case OP_END:      case OP_END:
1493      if (md->recursive != NULL)      case OP_ACCEPT:
1494        {      case OP_ASSERT_ACCEPT:
1495        recursion_info *rec = md->recursive;  
1496        md->recursive = rec->prevrec;      /* If we have matched an empty string, fail if not in an assertion and not
1497        memmove(md->offset_vector, rec->offset_save,      in a recursion if either PCRE_NOTEMPTY is set, or if PCRE_NOTEMPTY_ATSTART
1498          rec->saved_max * sizeof(int));      is set and we have matched at the start of the subject. In both cases,
1499        offset_top = rec->save_offset_top;      backtracking will then try other alternatives, if any. */
1500        if (rec->group_num == 0)  
1501          {      if (eptr == mstart && op != OP_ASSERT_ACCEPT &&
1502          ecode = rec->after_call;           md->recursive == NULL &&
1503          break;           (md->notempty ||
1504          }             (md->notempty_atstart &&
1505        }               mstart == md->start_subject + md->start_offset)))
1506          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);  
1507    
1508      /* Otherwise, we have a match. */      /* Otherwise, we have a match. */
1509    
1510      md->end_match_ptr = eptr;           /* Record where we ended */      md->end_match_ptr = eptr;           /* Record where we ended */
1511      md->end_offset_top = offset_top;    /* and how many extracts were taken */      md->end_offset_top = offset_top;    /* and how many extracts were taken */
1512      md->start_match_ptr = mstart;       /* and the start (\K can modify) */      md->start_match_ptr = mstart;       /* and the start (\K can modify) */
1513    
1514      /* 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
1515      given as the argument to MRRETURN when the heap is in use. */      given as the argument to RRETURN when the heap is in use. */
1516    
1517      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;
1518      MRRETURN(rrc);      RRETURN(rrc);
1519    
1520      /* Assertion brackets. Check the alternative branches in turn - the      /* Assertion brackets. Check the alternative branches in turn - the
1521      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,
1522      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
1523      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
1524      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
1525      of a condition, we want to return immediately afterwards. The caller of      of a condition, we want to return immediately afterwards. The caller of
1526      this incarnation of the match() function will have set MATCH_CONDASSERT in      this incarnation of the match() function will have set MATCH_CONDASSERT in
1527      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
1528      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
1529      match() to remember this case. */      match() to remember this case. */
1530    
1531      case OP_ASSERT:      case OP_ASSERT:
1532      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1533        save_mark = md->mark;
1534      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1535        {        {
1536        condassert = TRUE;        condassert = TRUE;
1537        md->match_function_type = 0;        md->match_function_type = 0;
1538        }        }
1539      else condassert = FALSE;      else condassert = FALSE;
1540    
1541      do      do
1542        {        {
1543        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
# Line 1376  for (;;) Line 1546  for (;;)
1546          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1547          break;          break;
1548          }          }
1549        if (rrc != MATCH_NOMATCH &&  
1550            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE does not allow THEN to escape beyond an assertion; it is treated
1551          RRETURN(rrc);        as NOMATCH. */
1552    
1553          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1554        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1555          md->mark = save_mark;
1556        }        }
1557      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1558    
1559      if (*ecode == OP_KET) MRRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1560    
1561      /* If checking an assertion for a condition, return MATCH_MATCH. */      /* If checking an assertion for a condition, return MATCH_MATCH. */
1562    
# Line 1403  for (;;) Line 1576  for (;;)
1576    
1577      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1578      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1579        save_mark = md->mark;
1580      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1581        {        {
1582        condassert = TRUE;        condassert = TRUE;
1583        md->match_function_type = 0;        md->match_function_type = 0;
1584        }        }
1585      else condassert = FALSE;      else condassert = FALSE;
1586    
1587      do      do
1588        {        {
1589        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1590        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) MRRETURN(MATCH_NOMATCH);        md->mark = save_mark;
1591          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1592        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1593          {          {
1594          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1595          break;          break;
1596          }          }
1597        if (rrc != MATCH_NOMATCH &&  
1598            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE does not allow THEN to escape beyond an assertion; it is treated
1599          RRETURN(rrc);        as NOMATCH. */
1600    
1601          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1602        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1603        }        }
1604      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1605    
1606      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1607    
1608      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1609      continue;      continue;
1610    
# Line 1437  for (;;) Line 1614  for (;;)
1614      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1615    
1616      case OP_REVERSE:      case OP_REVERSE:
1617  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1618      if (utf8)      if (utf)
1619        {        {
1620        i = GET(ecode, 1);        i = GET(ecode, 1);
1621        while (i-- > 0)        while (i-- > 0)
1622          {          {
1623          eptr--;          eptr--;
1624          if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1625          BACKCHAR(eptr);          BACKCHAR(eptr);
1626          }          }
1627        }        }
# Line 1455  for (;;) Line 1632  for (;;)
1632    
1633        {        {
1634        eptr -= GET(ecode, 1);        eptr -= GET(ecode, 1);
1635        if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1636        }        }
1637    
1638      /* Save the earliest consulted character, then skip to next op code */      /* Save the earliest consulted character, then skip to next op code */
# Line 1469  for (;;) Line 1646  for (;;)
1646      function is able to force a failure. */      function is able to force a failure. */
1647    
1648      case OP_CALLOUT:      case OP_CALLOUT:
1649      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1650        {        {
1651        pcre_callout_block cb;        PUBL(callout_block) cb;
1652        cb.version          = 1;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1653        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1654        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1655    #ifdef COMPILE_PCRE8
1656        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1657    #else
1658          cb.subject          = (PCRE_SPTR16)md->start_subject;
1659    #endif
1660        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1661        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1662        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
# Line 1484  for (;;) Line 1665  for (;;)
1665        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1666        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last;
1667        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1668        if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);        cb.mark             = md->nomatch_mark;
1669          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1670        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1671        }        }
1672      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1494  for (;;) Line 1676  for (;;)
1676      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
1677      whole pattern. (This is so that it works from duplicated subpatterns.)      whole pattern. (This is so that it works from duplicated subpatterns.)
1678    
1679      If there are any capturing brackets started but not finished, we have to      The state of the capturing groups is preserved over recursion, and
1680      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
1681      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
1682      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
1683      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
1684      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
1685      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.  
1686    
1687      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
1688      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
1689      for the original version of this logic. */      for the original version of this logic. It has, however, been hacked around
1690        a lot, so he is not to blame for the current way it works. */
1691    
1692      case OP_RECURSE:      case OP_RECURSE:
1693        {        {
1694          recursion_info *ri;
1695          int recno;
1696    
1697        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1698        new_recursive.group_num = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
1699          GET2(callpat, 1 + LINK_SIZE);          GET2(callpat, 1 + LINK_SIZE);
1700    
1701          /* Check for repeating a recursion without advancing the subject pointer.
1702          This should catch convoluted mutual recursions. (Some simple cases are
1703          caught at compile time.) */
1704    
1705          for (ri = md->recursive; ri != NULL; ri = ri->prevrec)
1706            if (recno == ri->group_num && eptr == ri->subject_position)
1707              RRETURN(PCRE_ERROR_RECURSELOOP);
1708    
1709        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1710    
1711          new_recursive.group_num = recno;
1712          new_recursive.subject_position = eptr;
1713        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1714        md->recursive = &new_recursive;        md->recursive = &new_recursive;
1715    
1716        /* Find where to continue from afterwards */        /* Where to continue from afterwards */
1717    
1718        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
       new_recursive.after_call = ecode;  
1719    
1720        /* Now save the offset data. */        /* Now save the offset data */
1721    
1722        new_recursive.saved_max = md->offset_end;        new_recursive.saved_max = md->offset_end;
1723        if (new_recursive.saved_max <= REC_STACK_SAVE_MAX)        if (new_recursive.saved_max <= REC_STACK_SAVE_MAX)
# Line 1533  for (;;) Line 1725  for (;;)
1725        else        else
1726          {          {
1727          new_recursive.offset_save =          new_recursive.offset_save =
1728            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1729          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1730          }          }
   
1731        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1732              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1733        new_recursive.save_offset_top = offset_top;  
1734          /* OK, now we can do the recursion. After processing each alternative,
1735        /* OK, now we can do the recursion. For each top-level alternative we        restore the offset data. If there were nested recursions, md->recursive
1736        restore the offset and recursion data. */        might be changed, so reset it before looping. */
1737    
1738        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1739        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
1740        do        do
1741          {          {
1742          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1743          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1744            md, eptrb, RM6);            md, eptrb, RM6);
1745            memcpy(md->offset_vector, new_recursive.offset_save,
1746                new_recursive.saved_max * sizeof(int));
1747            md->recursive = new_recursive.prevrec;
1748          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1749            {            {
1750            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
           md->recursive = new_recursive.prevrec;  
1751            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1752              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1753            MRRETURN(MATCH_MATCH);  
1754              /* Set where we got to in the subject, and reset the start in case
1755              it was changed by \K. This *is* propagated back out of a recursion,
1756              for Perl compatibility. */
1757    
1758              eptr = md->end_match_ptr;
1759              mstart = md->start_match_ptr;
1760              goto RECURSION_MATCHED;        /* Exit loop; end processing */
1761            }            }
1762          else if (rrc != MATCH_NOMATCH &&  
1763                  (rrc != MATCH_THEN || md->start_match_ptr != ecode))          /* PCRE does not allow THEN to escape beyond a recursion; it is treated
1764            as NOMATCH. */
1765    
1766            else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1767            {            {
1768            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1769            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1770              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1771            RRETURN(rrc);            RRETURN(rrc);
1772            }            }
1773    
1774          md->recursive = &new_recursive;          md->recursive = &new_recursive;
         memcpy(md->offset_vector, new_recursive.offset_save,  
             new_recursive.saved_max * sizeof(int));  
1775          callpat += GET(callpat, 1);          callpat += GET(callpat, 1);
1776          }          }
1777        while (*callpat == OP_ALT);        while (*callpat == OP_ALT);
# Line 1578  for (;;) Line 1779  for (;;)
1779        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1780        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1781        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1782          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1783        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;  
1784        }        }
1785    
1786      /* The repeating kets try the rest of the pattern or restart from the      RECURSION_MATCHED:
1787      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 */  
1788    
1789      /* 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
1790      bracketed group and go to there. */      bracketed group and go to there. */
# Line 1667  for (;;) Line 1798  for (;;)
1798      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
1799      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
1800      optional ones preceded by BRAZERO or BRAMINZERO. */      optional ones preceded by BRAZERO or BRAMINZERO. */
1801    
1802      case OP_BRAZERO:      case OP_BRAZERO:
1803      next = ecode + 1;      next = ecode + 1;
1804      RMATCH(eptr, next, offset_top, md, eptrb, RM10);      RMATCH(eptr, next, offset_top, md, eptrb, RM10);
# Line 1675  for (;;) Line 1806  for (;;)
1806      do next += GET(next, 1); while (*next == OP_ALT);      do next += GET(next, 1); while (*next == OP_ALT);
1807      ecode = next + 1 + LINK_SIZE;      ecode = next + 1 + LINK_SIZE;
1808      break;      break;
1809    
1810      case OP_BRAMINZERO:      case OP_BRAMINZERO:
1811      next = ecode + 1;      next = ecode + 1;
1812      do next += GET(next, 1); while (*next == OP_ALT);      do next += GET(next, 1); while (*next == OP_ALT);
# Line 1689  for (;;) Line 1820  for (;;)
1820      do next += GET(next,1); while (*next == OP_ALT);      do next += GET(next,1); while (*next == OP_ALT);
1821      ecode = next + 1 + LINK_SIZE;      ecode = next + 1 + LINK_SIZE;
1822      break;      break;
1823    
1824      /* BRAPOSZERO occurs before a possessive bracket group. Don't do anything      /* BRAPOSZERO occurs before a possessive bracket group. Don't do anything
1825      here; just jump to the group, with allow_zero set TRUE. */      here; just jump to the group, with allow_zero set TRUE. */
1826    
1827      case OP_BRAPOSZERO:      case OP_BRAPOSZERO:
1828      op = *(++ecode);      op = *(++ecode);
1829      allow_zero = TRUE;      allow_zero = TRUE;
1830      if (op == OP_CBRAPOS || op == OP_SCBRAPOS) goto POSSESSIVE_CAPTURE;      if (op == OP_CBRAPOS || op == OP_SCBRAPOS) goto POSSESSIVE_CAPTURE;
1831        goto POSSESSIVE_NON_CAPTURE;        goto POSSESSIVE_NON_CAPTURE;
# Line 1704  for (;;) Line 1835  for (;;)
1835      case OP_KET:      case OP_KET:
1836      case OP_KETRMIN:      case OP_KETRMIN:
1837      case OP_KETRMAX:      case OP_KETRMAX:
1838      case OP_KETRPOS:      case OP_KETRPOS:
1839      prev = ecode - GET(ecode, 1);      prev = ecode - GET(ecode, 1);
1840    
1841      /* 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
1842      infinite repeats of empty string matches, retrieve the subject start from      infinite repeats of empty string matches, retrieve the subject start from
1843      the chain. Otherwise, set it NULL. */      the chain. Otherwise, set it NULL. */
1844    
1845      if (*prev >= OP_SBRA)      if (*prev >= OP_SBRA || *prev == OP_ONCE)
1846        {        {
1847        saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */        saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */
1848        eptrb = eptrb->epb_prev;              /* Backup to previous group */        eptrb = eptrb->epb_prev;              /* Backup to previous group */
1849        }        }
1850      else saved_eptr = NULL;      else saved_eptr = NULL;
1851    
1852      /* 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
1853      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
1854      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
1855      it was changed by \K. */      start in case it was changed by \K. */
1856    
1857      if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||      if ((*prev >= OP_ASSERT && *prev <= OP_ASSERTBACK_NOT) ||
1858          *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||           *prev == OP_ONCE_NC)
         *prev == OP_ONCE)  
1859        {        {
1860        md->end_match_ptr = eptr;      /* For ONCE */        md->end_match_ptr = eptr;      /* For ONCE_NC */
1861        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
1862        md->start_match_ptr = mstart;        md->start_match_ptr = mstart;
1863        MRRETURN(MATCH_MATCH);        RRETURN(MATCH_MATCH);         /* Sets md->mark */
1864        }        }
1865    
1866      /* 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
1867      and if necessary complete handling an extraction by setting the offsets and      and if necessary complete handling an extraction by setting the offsets and
1868      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
1869      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
1870      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
1871        the current subject position and start match pointer and give a MATCH
1872        return. */
1873    
1874      if (*prev == OP_CBRA || *prev == OP_SCBRA ||      if (*prev == OP_CBRA || *prev == OP_SCBRA ||
1875          *prev == OP_CBRAPOS || *prev == OP_SCBRAPOS)          *prev == OP_CBRAPOS || *prev == OP_SCBRAPOS)
# Line 1750  for (;;) Line 1882  for (;;)
1882        printf("\n");        printf("\n");
1883  #endif  #endif
1884    
1885          /* Handle a recursively called group. */
1886    
1887          if (md->recursive != NULL && md->recursive->group_num == number)
1888            {
1889            md->end_match_ptr = eptr;
1890            md->start_match_ptr = mstart;
1891            RRETURN(MATCH_MATCH);
1892            }
1893    
1894          /* Deal with capturing */
1895    
1896        md->capture_last = number;        md->capture_last = number;
1897        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->offset_overflow = TRUE; else
1898          {          {
1899            /* If offset is greater than offset_top, it means that we are
1900            "skipping" a capturing group, and that group's offsets must be marked
1901            unset. In earlier versions of PCRE, all the offsets were unset at the
1902            start of matching, but this doesn't work because atomic groups and
1903            assertions can cause a value to be set that should later be unset.
1904            Example: matching /(?>(a))b|(a)c/ against "ac". This sets group 1 as
1905            part of the atomic group, but this is not on the final matching path,
1906            so must be unset when 2 is set. (If there is no group 2, there is no
1907            problem, because offset_top will then be 2, indicating no capture.) */
1908    
1909            if (offset > offset_top)
1910              {
1911              register int *iptr = md->offset_vector + offset_top;
1912              register int *iend = md->offset_vector + offset;
1913              while (iptr < iend) *iptr++ = -1;
1914              }
1915    
1916            /* Now make the extraction */
1917    
1918          md->offset_vector[offset] =          md->offset_vector[offset] =
1919            md->offset_vector[md->offset_end - number];            md->offset_vector[md->offset_end - number];
1920          md->offset_vector[offset+1] = (int)(eptr - md->start_subject);          md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1921          if (offset_top <= offset) offset_top = offset + 2;          if (offset_top <= offset) offset_top = offset + 2;
1922          }          }
   
       /* 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;  
         }  
1923        }        }
1924    
1925      /* For a non-repeating ket, just continue at this level. This also      /* For an ordinary non-repeating ket, just continue at this level. This
1926      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
1927      This is the forcible breaking of infinite loops as implemented in Perl      group. This is the forcible breaking of infinite loops as implemented in
1928      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,
1929      course of events. */      establish a backup point by processing the rest of the pattern at a lower
1930        level. If this results in a NOMATCH return, pass MATCH_ONCE back to the
1931        original OP_ONCE level, thereby bypassing intermediate backup points, but
1932        resetting any captures that happened along the way. */
1933    
1934      if (*ecode == OP_KET || eptr == saved_eptr)      if (*ecode == OP_KET || eptr == saved_eptr)
1935        {        {
1936        ecode += 1 + LINK_SIZE;        if (*prev == OP_ONCE)
1937            {
1938            RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM12);
1939            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1940            md->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */
1941            RRETURN(MATCH_ONCE);
1942            }
1943          ecode += 1 + LINK_SIZE;    /* Carry on at this level */
1944        break;        break;
1945        }        }
1946    
1947      /* OP_KETRPOS is a possessive repeating ket. Remember the current position,      /* OP_KETRPOS is a possessive repeating ket. Remember the current position,
1948      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
1949      at a time from the outer level, thus saving stack. */      at a time from the outer level, thus saving stack. */
1950    
1951      if (*ecode == OP_KETRPOS)      if (*ecode == OP_KETRPOS)
1952        {        {
1953        md->end_match_ptr = eptr;        md->end_match_ptr = eptr;
1954        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
1955        RRETURN(MATCH_KETRPOS);        RRETURN(MATCH_KETRPOS);
1956        }        }
1957    
1958      /* 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
1959      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
1960      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
1961      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
1962        string. */
1963    
1964      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1965        {        {
1966        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM12);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM7);
1967        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1968          if (*prev == OP_ONCE)
1969            {
1970            RMATCH(eptr, prev, offset_top, md, eptrb, RM8);
1971            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1972            md->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */
1973            RRETURN(MATCH_ONCE);
1974            }
1975        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
1976          {          {
1977          md->match_function_type = MATCH_CBEGROUP;          md->match_function_type = MATCH_CBEGROUP;
1978          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
1979          RRETURN(rrc);          RRETURN(rrc);
1980          }          }
# Line 1818  for (;;) Line 1983  for (;;)
1983        }        }
1984      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
1985        {        {
1986        if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1987        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
1988          if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
1989        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1990          if (*prev == OP_ONCE)
1991            {
1992            RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM9);
1993            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1994            md->once_target = prev;
1995            RRETURN(MATCH_ONCE);
1996            }
1997        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1998        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1999        }        }
# Line 1829  for (;;) Line 2002  for (;;)
2002      /* Not multiline mode: start of subject assertion, unless notbol. */      /* Not multiline mode: start of subject assertion, unless notbol. */
2003    
2004      case OP_CIRC:      case OP_CIRC:
2005      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
2006    
2007      /* Start of subject assertion */      /* Start of subject assertion */
2008    
2009      case OP_SOD:      case OP_SOD:
2010      if (eptr != md->start_subject) MRRETURN(MATCH_NOMATCH);      if (eptr != md->start_subject) RRETURN(MATCH_NOMATCH);
2011      ecode++;      ecode++;
2012      break;      break;
2013    
2014      /* Multiline mode: start of subject unless notbol, or after any newline. */      /* Multiline mode: start of subject unless notbol, or after any newline. */
2015    
2016      case OP_CIRCM:      case OP_CIRCM:
2017      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
2018      if (eptr != md->start_subject &&      if (eptr != md->start_subject &&
2019          (eptr == md->end_subject || !WAS_NEWLINE(eptr)))          (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
2020        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2021      ecode++;      ecode++;
2022      break;      break;
2023    
2024      /* Start of match assertion */      /* Start of match assertion */
2025    
2026      case OP_SOM:      case OP_SOM:
2027      if (eptr != md->start_subject + md->start_offset) MRRETURN(MATCH_NOMATCH);      if (eptr != md->start_subject + md->start_offset) RRETURN(MATCH_NOMATCH);
2028      ecode++;      ecode++;
2029      break;      break;
2030    
# Line 1867  for (;;) Line 2040  for (;;)
2040    
2041      case OP_DOLLM:      case OP_DOLLM:
2042      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2043        { if (!IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH); }        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }
2044      else      else
2045        {        {
2046        if (md->noteol) MRRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
2047        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2048        }        }
2049      ecode++;      ecode++;
2050      break;      break;
2051    
2052      /* Not multiline mode: assert before a terminating newline or before end of      /* Not multiline mode: assert before a terminating newline or before end of
2053      subject unless noteol is set. */      subject unless noteol is set. */
2054    
2055      case OP_DOLL:      case OP_DOLL:
2056      if (md->noteol) MRRETURN(MATCH_NOMATCH);      if (md->noteol) RRETURN(MATCH_NOMATCH);
2057      if (!md->endonly) goto ASSERT_NL_OR_EOS;      if (!md->endonly) goto ASSERT_NL_OR_EOS;
2058    
2059      /* ... else fall through for endonly */      /* ... else fall through for endonly */
# Line 1888  for (;;) Line 2061  for (;;)
2061      /* End of subject assertion (\z) */      /* End of subject assertion (\z) */
2062    
2063      case OP_EOD:      case OP_EOD:
2064      if (eptr < md->end_subject) MRRETURN(MATCH_NOMATCH);      if (eptr < md->end_subject) RRETURN(MATCH_NOMATCH);
2065      SCHECK_PARTIAL();      SCHECK_PARTIAL();
2066      ecode++;      ecode++;
2067      break;      break;
# Line 1899  for (;;) Line 2072  for (;;)
2072      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2073      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2074          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2075        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2076    
2077      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2078    
# Line 1918  for (;;) Line 2091  for (;;)
2091        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2092        partial matching. */        partial matching. */
2093    
2094  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2095        if (utf8)        if (utf)
2096          {          {
2097          /* Get status of previous character */          /* Get status of previous character */
2098    
2099          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2100            {            {
2101            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2102            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2103            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2104            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2105  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 1991  for (;;) Line 2164  for (;;)
2164              }              }
2165            else            else
2166  #endif  #endif
2167            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2168                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2169            }            }
2170    
2171          /* Get status of next character */          /* Get status of next character */
# Line 2014  for (;;) Line 2188  for (;;)
2188            }            }
2189          else          else
2190  #endif  #endif
2191          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2192              && ((md->ctypes[*eptr] & ctype_word) != 0);
2193          }          }
2194    
2195        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
2196    
2197        if ((*ecode++ == OP_WORD_BOUNDARY)?        if ((*ecode++ == OP_WORD_BOUNDARY)?
2198             cur_is_word == prev_is_word : cur_is_word != prev_is_word)             cur_is_word == prev_is_word : cur_is_word != prev_is_word)
2199          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2200        }        }
2201      break;      break;
2202    
2203      /* Match a single character type; inline for speed */      /* Match a single character type; inline for speed */
2204    
2205      case OP_ANY:      case OP_ANY:
2206      if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2207      /* Fall through */      /* Fall through */
2208    
2209      case OP_ALLANY:      case OP_ALLANY:
2210      if (eptr++ >= md->end_subject)      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2211        {        {                            /* not be updated before SCHECK_PARTIAL. */
2212        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2213        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2214        }        }
2215      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;      eptr++;
2216    #ifdef SUPPORT_UTF
2217        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2218    #endif
2219      ecode++;      ecode++;
2220      break;      break;
2221    
# Line 2045  for (;;) Line 2223  for (;;)
2223      any byte, even newline, independent of the setting of PCRE_DOTALL. */      any byte, even newline, independent of the setting of PCRE_DOTALL. */
2224    
2225      case OP_ANYBYTE:      case OP_ANYBYTE:
2226      if (eptr++ >= md->end_subject)      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2227        {        {                            /* not be updated before SCHECK_PARTIAL. */
2228        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2229        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2230        }        }
2231        eptr++;
2232      ecode++;      ecode++;
2233      break;      break;
2234    
# Line 2057  for (;;) Line 2236  for (;;)
2236      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2237        {        {
2238        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2239        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2240        }        }
2241      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2242      if (      if (
2243  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2244         c < 256 &&         c < 256 &&
2245  #endif  #endif
2246         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
2247         )         )
2248        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2249      ecode++;      ecode++;
2250      break;      break;
2251    
# Line 2074  for (;;) Line 2253  for (;;)
2253      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2254        {        {
2255        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2256        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2257        }        }
2258      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2259      if (      if (
2260  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2261         c >= 256 ||         c > 255 ||
2262  #endif  #endif
2263         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2264         )         )
2265        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2266      ecode++;      ecode++;
2267      break;      break;
2268    
# Line 2091  for (;;) Line 2270  for (;;)
2270      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2271        {        {
2272        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2273        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2274        }        }
2275      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2276      if (      if (
2277  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2278         c < 256 &&         c < 256 &&
2279  #endif  #endif
2280         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
2281         )         )
2282        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2283      ecode++;      ecode++;
2284      break;      break;
2285    
# Line 2108  for (;;) Line 2287  for (;;)
2287      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2288        {        {
2289        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2290        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2291        }        }
2292      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2293      if (      if (
2294  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2295         c >= 256 ||         c > 255 ||
2296  #endif  #endif
2297         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2298         )         )
2299        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2300      ecode++;      ecode++;
2301      break;      break;
2302    
# Line 2125  for (;;) Line 2304  for (;;)
2304      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2305        {        {
2306        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2307        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2308        }        }
2309      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2310      if (      if (
2311  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2312         c < 256 &&         c < 256 &&
2313  #endif  #endif
2314         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
2315         )         )
2316        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2317      ecode++;      ecode++;
2318      break;      break;
2319    
# Line 2142  for (;;) Line 2321  for (;;)
2321      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2322        {        {
2323        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2324        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2325        }        }
2326      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2327      if (      if (
2328  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2329         c >= 256 ||         c > 255 ||
2330  #endif  #endif
2331         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2332         )         )
2333        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2334      ecode++;      ecode++;
2335      break;      break;
2336    
# Line 2159  for (;;) Line 2338  for (;;)
2338      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2339        {        {
2340        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2341        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2342        }        }
2343      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2344      switch(c)      switch(c)
2345        {        {
2346        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2347    
2348        case 0x000d:        case 0x000d:
2349        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
2350        break;        break;
# Line 2178  for (;;) Line 2357  for (;;)
2357        case 0x0085:        case 0x0085:
2358        case 0x2028:        case 0x2028:
2359        case 0x2029:        case 0x2029:
2360        if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2361        break;        break;
2362        }        }
2363      ecode++;      ecode++;
# Line 2188  for (;;) Line 2367  for (;;)
2367      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2368        {        {
2369        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2370        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2371        }        }
2372      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2373      switch(c)      switch(c)
# Line 2213  for (;;) Line 2392  for (;;)
2392        case 0x202f:    /* NARROW NO-BREAK SPACE */        case 0x202f:    /* NARROW NO-BREAK SPACE */
2393        case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */        case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
2394        case 0x3000:    /* IDEOGRAPHIC SPACE */        case 0x3000:    /* IDEOGRAPHIC SPACE */
2395        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2396        }        }
2397      ecode++;      ecode++;
2398      break;      break;
# Line 2222  for (;;) Line 2401  for (;;)
2401      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2402        {        {
2403        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2404        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2405        }        }
2406      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2407      switch(c)      switch(c)
2408        {        {
2409        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2410        case 0x09:      /* HT */        case 0x09:      /* HT */
2411        case 0x20:      /* SPACE */        case 0x20:      /* SPACE */
2412        case 0xa0:      /* NBSP */        case 0xa0:      /* NBSP */
# Line 2256  for (;;) Line 2435  for (;;)
2435      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2436        {        {
2437        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2438        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2439        }        }
2440      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2441      switch(c)      switch(c)
# Line 2269  for (;;) Line 2448  for (;;)
2448        case 0x85:      /* NEL */        case 0x85:      /* NEL */
2449        case 0x2028:    /* LINE SEPARATOR */        case 0x2028:    /* LINE SEPARATOR */
2450        case 0x2029:    /* PARAGRAPH SEPARATOR */        case 0x2029:    /* PARAGRAPH SEPARATOR */
2451        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2452        }        }
2453      ecode++;      ecode++;
2454      break;      break;
# Line 2278  for (;;) Line 2457  for (;;)
2457      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2458        {        {
2459        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2460        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2461        }        }
2462      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2463      switch(c)      switch(c)
2464        {        {
2465        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2466        case 0x0a:      /* LF */        case 0x0a:      /* LF */
2467        case 0x0b:      /* VT */        case 0x0b:      /* VT */
2468        case 0x0c:      /* FF */        case 0x0c:      /* FF */
# Line 2305  for (;;) Line 2484  for (;;)
2484      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2485        {        {
2486        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2487        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2488        }        }
2489      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2490        {        {
# Line 2314  for (;;) Line 2493  for (;;)
2493        switch(ecode[1])        switch(ecode[1])
2494          {          {
2495          case PT_ANY:          case PT_ANY:
2496          if (op == OP_NOTPROP) MRRETURN(MATCH_NOMATCH);          if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
2497          break;          break;
2498    
2499          case PT_LAMP:          case PT_LAMP:
2500          if ((prop->chartype == ucp_Lu ||          if ((prop->chartype == ucp_Lu ||
2501               prop->chartype == ucp_Ll ||               prop->chartype == ucp_Ll ||
2502               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))
2503            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2504          break;          break;
2505    
2506          case PT_GC:          case PT_GC:
2507          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2508            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2509          break;          break;
2510    
2511          case PT_PC:          case PT_PC:
2512          if ((ecode[2] != prop->chartype) == (op == OP_PROP))          if ((ecode[2] != prop->chartype) == (op == OP_PROP))
2513            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2514          break;          break;
2515    
2516          case PT_SC:          case PT_SC:
2517          if ((ecode[2] != prop->script) == (op == OP_PROP))          if ((ecode[2] != prop->script) == (op == OP_PROP))
2518            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2519          break;          break;
2520    
2521          /* These are specials */          /* These are specials */
2522    
2523          case PT_ALNUM:          case PT_ALNUM:
2524          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2525               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2526            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2527          break;          break;
2528    
2529          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2530          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2531               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2532                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2533            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2534          break;          break;
2535    
2536          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2537          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2538               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2539               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2540                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2541            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2542          break;          break;
2543    
2544          case PT_WORD:          case PT_WORD:
2545          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2546               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2547               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2548            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2549          break;          break;
2550    
2551          /* This should never occur */          /* This should never occur */
# Line 2386  for (;;) Line 2565  for (;;)
2565      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2566        {        {
2567        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2568        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2569        }        }
2570      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2571        if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);
2572        while (eptr < md->end_subject)
2573        {        {
2574        int category = UCD_CATEGORY(c);        int len = 1;
2575        if (category == ucp_M) MRRETURN(MATCH_NOMATCH);        if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2576        while (eptr < md->end_subject)        if (UCD_CATEGORY(c) != ucp_M) break;
2577          {        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;  
         }  
2578        }        }
2579      ecode++;      ecode++;
2580      break;      break;
# Line 2418  for (;;) Line 2590  for (;;)
2590      loops). */      loops). */
2591    
2592      case OP_REF:      case OP_REF:
2593      case OP_REFI:      case OP_REFI:
2594      caseless = op == OP_REFI;      caseless = op == OP_REFI;
2595      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2596      ecode += 3;      ecode += 1 + IMM2_SIZE;
2597    
2598      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2599    
# Line 2461  for (;;) Line 2633  for (;;)
2633        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2634        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2635        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2636        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2637        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2638        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2639        break;        break;
2640    
2641        default:               /* No repeat follows */        default:               /* No repeat follows */
2642        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2643          {          {
2644          CHECK_PARTIAL();          CHECK_PARTIAL();
2645          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2646          }          }
2647        eptr += length;        eptr += length;
2648        continue;              /* With the main loop */        continue;              /* With the main loop */
2649        }        }
2650    
2651      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2652      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2653        means the reference is unset in non-Java-compatible mode. If the minimum is
2654        zero, we can continue at the same level without recursion. For any other
2655        minimum, carrying on will result in NOMATCH. */
2656    
2657      if (length == 0) continue;      if (length == 0) continue;
2658        if (length < 0 && min == 0) continue;
2659    
2660      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2661      the length of the reference string explicitly rather than passing the      the length of the reference string explicitly rather than passing the
# Line 2487  for (;;) Line 2663  for (;;)
2663    
2664      for (i = 1; i <= min; i++)      for (i = 1; i <= min; i++)
2665        {        {
2666        int slength;        int slength;
2667        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2668          {          {
2669          CHECK_PARTIAL();          CHECK_PARTIAL();
2670          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2671          }          }
2672        eptr += slength;        eptr += slength;
2673        }        }
# Line 2507  for (;;) Line 2683  for (;;)
2683        {        {
2684        for (fi = min;; fi++)        for (fi = min;; fi++)
2685          {          {
2686          int slength;          int slength;
2687          RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);
2688          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2689          if (fi >= max) MRRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2690          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2691            {            {
2692            CHECK_PARTIAL();            CHECK_PARTIAL();
2693            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2694            }            }
2695          eptr += slength;          eptr += slength;
2696          }          }
# Line 2528  for (;;) Line 2704  for (;;)
2704        pp = eptr;        pp = eptr;
2705        for (i = min; i < max; i++)        for (i = min; i < max; i++)
2706          {          {
2707          int slength;          int slength;
2708          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2709            {            {
2710            CHECK_PARTIAL();            CHECK_PARTIAL();
# Line 2542  for (;;) Line 2718  for (;;)
2718          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2719          eptr -= length;          eptr -= length;
2720          }          }
2721        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2722        }        }
2723      /* Control never gets here */      /* Control never gets here */
2724    
# Line 2560  for (;;) Line 2736  for (;;)
2736      case OP_NCLASS:      case OP_NCLASS:
2737      case OP_CLASS:      case OP_CLASS:
2738        {        {
2739          /* The data variable is saved across frames, so the byte map needs to
2740          be stored there. */
2741    #define BYTE_MAP ((pcre_uint8 *)data)
2742        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2743        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2744    
2745        switch (*ecode)        switch (*ecode)
2746          {          {
# Line 2582  for (;;) Line 2761  for (;;)
2761          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2762          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2763          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2764          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2765          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2766          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2767          break;          break;
2768    
2769          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2594  for (;;) Line 2773  for (;;)
2773    
2774        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2775    
2776  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2777        /* UTF-8 mode */        if (utf)
       if (utf8)  
2778          {          {
2779          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2780            {            {
2781            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2782              {              {
2783              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2784              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2785              }              }
2786            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
2787            if (c > 255)            if (c > 255)
2788              {              {
2789              if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2790              }              }
2791            else            else
2792              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
             }  
2793            }            }
2794          }          }
2795        else        else
2796  #endif  #endif
2797        /* Not UTF-8 mode */        /* Not UTF mode */
2798          {          {
2799          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2800            {            {
2801            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2802              {              {
2803              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2804              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2805              }              }
2806            c = *eptr++;            c = *eptr++;
2807            if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2808              if (c > 255)
2809                {
2810                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2811                }
2812              else
2813    #endif
2814                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2815            }            }
2816          }          }
2817    
# Line 2642  for (;;) Line 2825  for (;;)
2825    
2826        if (minimize)        if (minimize)
2827          {          {
2828  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2829          /* UTF-8 mode */          if (utf)
         if (utf8)  
2830            {            {
2831            for (fi = min;; fi++)            for (fi = min;; fi++)
2832              {              {
2833              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);
2834              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2835              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2836              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
2837                {                {
2838                SCHECK_PARTIAL();                SCHECK_PARTIAL();
2839                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2840                }                }
2841              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
2842              if (c > 255)              if (c > 255)
2843                {                {
2844                if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2845                }                }
2846              else              else
2847                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
               }  
2848              }              }
2849            }            }
2850          else          else
2851  #endif  #endif
2852          /* Not UTF-8 mode */          /* Not UTF mode */
2853            {            {
2854            for (fi = min;; fi++)            for (fi = min;; fi++)
2855              {              {
2856              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);
2857              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2858              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2859              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
2860                {                {
2861                SCHECK_PARTIAL();                SCHECK_PARTIAL();
2862                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2863                }                }
2864              c = *eptr++;              c = *eptr++;
2865              if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2866                if (c > 255)
2867                  {
2868                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2869                  }
2870                else
2871    #endif
2872                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2873              }              }
2874            }            }
2875          /* Control never gets here */          /* Control never gets here */
# Line 2694  for (;;) Line 2881  for (;;)
2881          {          {
2882          pp = eptr;          pp = eptr;
2883    
2884  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2885          /* UTF-8 mode */          if (utf)
         if (utf8)  
2886            {            {
2887            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2888              {              {
# Line 2712  for (;;) Line 2898  for (;;)
2898                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
2899                }                }
2900              else              else
2901                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
2902              eptr += len;              eptr += len;
2903              }              }
2904            for (;;)            for (;;)
# Line 2727  for (;;) Line 2911  for (;;)
2911            }            }
2912          else          else
2913  #endif  #endif
2914            /* Not UTF-8 mode */            /* Not UTF mode */
2915            {            {
2916            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2917              {              {
# Line 2737  for (;;) Line 2921  for (;;)
2921                break;                break;
2922                }                }
2923              c = *eptr;              c = *eptr;
2924              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
2925                if (c > 255)
2926                  {
2927                  if (op == OP_CLASS) break;
2928                  }
2929                else
2930    #endif
2931                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
2932              eptr++;              eptr++;
2933              }              }
2934            while (eptr >= pp)            while (eptr >= pp)
# Line 2748  for (;;) Line 2939  for (;;)
2939              }              }
2940            }            }
2941    
2942          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2943          }          }
2944    #undef BYTE_MAP
2945        }        }
2946      /* Control never gets here */      /* Control never gets here */
2947    
# Line 2758  for (;;) Line 2950  for (;;)
2950      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
2951      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
2952    
2953  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
2954      case OP_XCLASS:      case OP_XCLASS:
2955        {        {
2956        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2783  for (;;) Line 2975  for (;;)
2975          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2976          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2977          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2978          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2979          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2980          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2981          break;          break;
2982    
2983          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2800  for (;;) Line 2992  for (;;)
2992          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
2993            {            {
2994            SCHECK_PARTIAL();            SCHECK_PARTIAL();
2995            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2996            }            }
2997          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
2998          if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
2999          }          }
3000    
3001        /* 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 2820  for (;;) Line 3012  for (;;)
3012            {            {
3013            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);
3014            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3015            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3016            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3017              {              {
3018              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3019              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3020              }              }
3021            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3022            if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3023            }            }
3024          /* Control never gets here */          /* Control never gets here */
3025          }          }
# Line 2845  for (;;) Line 3037  for (;;)
3037              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3038              break;              break;
3039              }              }
3040    #ifdef SUPPORT_UTF
3041            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3042            if (!_pcre_xclass(c, data)) break;  #else
3043              c = *eptr;
3044    #endif
3045              if (!PRIV(xclass)(c, data, utf)) break;
3046            eptr += len;            eptr += len;
3047            }            }
3048          for(;;)          for(;;)
# Line 2854  for (;;) Line 3050  for (;;)
3050            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3051            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3052            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3053            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3054              if (utf) BACKCHAR(eptr);
3055    #endif
3056            }            }
3057          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3058          }          }
3059    
3060        /* Control never gets here */        /* Control never gets here */
# Line 2866  for (;;) Line 3064  for (;;)
3064      /* Match a single character, casefully */      /* Match a single character, casefully */
3065    
3066      case OP_CHAR:      case OP_CHAR:
3067  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3068      if (utf8)      if (utf)
3069        {        {
3070        length = 1;        length = 1;
3071        ecode++;        ecode++;
# Line 2875  for (;;) Line 3073  for (;;)
3073        if (length > md->end_subject - eptr)        if (length > md->end_subject - eptr)
3074          {          {
3075          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3076          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3077          }          }
3078        while (length-- > 0) if (*ecode++ != *eptr++) MRRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);
3079        }        }
3080      else      else
3081  #endif  #endif
3082        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3083        {        {
3084        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3085          {          {
3086          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
3087          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3088          }          }
3089        if (ecode[1] != *eptr++) MRRETURN(MATCH_NOMATCH);        if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH);
3090        ecode += 2;        ecode += 2;
3091        }        }
3092      break;      break;
3093    
3094      /* Match a single character, caselessly */      /* Match a single character, caselessly. If we are at the end of the
3095        subject, give up immediately. */
3096    
3097        case OP_CHARI:
3098        if (eptr >= md->end_subject)
3099          {
3100          SCHECK_PARTIAL();
3101          RRETURN(MATCH_NOMATCH);
3102          }
3103    
3104      case OP_CHARI:  #ifdef SUPPORT_UTF
3105  #ifdef SUPPORT_UTF8      if (utf)
     if (utf8)  
3106        {        {
3107        length = 1;        length = 1;
3108        ecode++;        ecode++;
3109        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
3110    
       if (length > md->end_subject - eptr)  
         {  
         CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */  
         MRRETURN(MATCH_NOMATCH);  
         }  
   
3111        /* 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
3112        can use the fast lookup table. */        we know that its other case must also be one byte long, so we can use the
3113          fast lookup table. We know that there is at least one byte left in the
3114          subject. */
3115    
3116        if (fc < 128)        if (fc < 128)
3117          {          {
3118          if (md->lcc[*ecode++] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);          if (md->lcc[fc]
3119                != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3120            ecode++;
3121            eptr++;
3122          }          }
3123    
3124        /* Otherwise we must pick up the subject character */        /* Otherwise we must pick up the subject character. Note that we cannot
3125          use the value of "length" to check for sufficient bytes left, because the
3126          other case of the character may have more or fewer bytes.  */
3127    
3128        else        else
3129          {          {
# Line 2934  for (;;) Line 3139  for (;;)
3139  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3140            if (dc != UCD_OTHERCASE(fc))            if (dc != UCD_OTHERCASE(fc))
3141  #endif  #endif
3142              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3143            }            }
3144          }          }
3145        }        }
3146      else      else
3147  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3148    
3149      /* Non-UTF-8 mode */      /* Not UTF mode */
3150        {        {
3151        if (md->end_subject - eptr < 1)        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3152          {            != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3153          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */        eptr++;
         MRRETURN(MATCH_NOMATCH);  
         }  
       if (md->lcc[ecode[1]] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);  
3154        ecode += 2;        ecode += 2;
3155        }        }
3156      break;      break;
# Line 2958  for (;;) Line 3160  for (;;)
3160      case OP_EXACT:      case OP_EXACT:
3161      case OP_EXACTI:      case OP_EXACTI:
3162      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3163      ecode += 3;      ecode += 1 + IMM2_SIZE;
3164      goto REPEATCHAR;      goto REPEATCHAR;
3165    
3166      case OP_POSUPTO:      case OP_POSUPTO:
# Line 2973  for (;;) Line 3175  for (;;)
3175      min = 0;      min = 0;
3176      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3177      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3178      ecode += 3;      ecode += 1 + IMM2_SIZE;
3179      goto REPEATCHAR;      goto REPEATCHAR;
3180    
3181      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3021  for (;;) Line 3223  for (;;)
3223      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3224    
3225      REPEATCHAR:      REPEATCHAR:
3226  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3227      if (utf8)      if (utf)
3228        {        {
3229        length = 1;        length = 1;
3230        charptr = ecode;        charptr = ecode;
# Line 3038  for (;;) Line 3240  for (;;)
3240          unsigned int othercase;          unsigned int othercase;
3241          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3242              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3243            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3244          else oclength = 0;          else oclength = 0;
3245  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3246    
3247          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3248            {            {
3249            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3250              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3251  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3252            else if (oclength > 0 &&            else if (oclength > 0 &&
3253                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3254                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3255  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3256            else            else
3257              {              {
3258              CHECK_PARTIAL();              CHECK_PARTIAL();
3259              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3260              }              }
3261            }            }
3262    
# Line 3066  for (;;) Line 3268  for (;;)
3268              {              {
3269              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);
3270              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3271              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3272              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3273                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3274  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3275              else if (oclength > 0 &&              else if (oclength > 0 &&
3276                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3277                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3278  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3279              else              else
3280                {                {
3281                CHECK_PARTIAL();                CHECK_PARTIAL();
3282                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3283                }                }
3284              }              }
3285            /* Control never gets here */            /* Control never gets here */
# Line 3089  for (;;) Line 3291  for (;;)
3291            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3292              {              {
3293              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3294                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3295  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3296              else if (oclength > 0 &&              else if (oclength > 0 &&
3297                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3298                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3299  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3300              else              else
3301                {                {
# Line 3108  for (;;) Line 3310  for (;;)
3310              {              {
3311              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3312              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3313              if (eptr == pp) { MRRETURN(MATCH_NOMATCH); }              if (eptr == pp) { RRETURN(MATCH_NOMATCH); }
3314  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3315              eptr--;              eptr--;
3316              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3125  for (;;) Line 3327  for (;;)
3327        value of fc will always be < 128. */        value of fc will always be < 128. */
3328        }        }
3329      else      else
3330  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3331          /* When not in UTF-8 mode, load a single-byte character. */
3332      /* When not in UTF-8 mode, load a single-byte character. */        fc = *ecode++;
3333    
3334      fc = *ecode++;      /* The value of fc at this point is always one character, though we may
3335        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  
3336      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3337      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3338      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3145  for (;;) Line 3345  for (;;)
3345    
3346      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3347        {        {
3348        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3349          /* fc must be < 128 if UTF is enabled. */
3350          foc = md->fcc[fc];
3351    #else
3352    #ifdef SUPPORT_UTF
3353    #ifdef SUPPORT_UCP
3354          if (utf && fc > 127)
3355            foc = UCD_OTHERCASE(fc);
3356    #else
3357          if (utf && fc > 127)
3358            foc = fc;
3359    #endif /* SUPPORT_UCP */
3360          else
3361    #endif /* SUPPORT_UTF */
3362            foc = TABLE_GET(fc, md->fcc, fc);
3363    #endif /* COMPILE_PCRE8 */
3364    
3365        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3366          {          {
3367          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3368            {            {
3369            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3370            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3371            }            }
3372          if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3373            eptr++;
3374          }          }
3375        if (min == max) continue;        if (min == max) continue;
3376        if (minimize)        if (minimize)
# Line 3162  for (;;) Line 3379  for (;;)
3379            {            {
3380            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3381            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3382            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3383            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3384              {              {
3385              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3386              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3387              }              }
3388            if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3389              eptr++;
3390            }            }
3391          /* Control never gets here */          /* Control never gets here */
3392          }          }
# Line 3182  for (;;) Line 3400  for (;;)
3400              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3401              break;              break;
3402              }              }
3403            if (fc != md->lcc[*eptr]) break;            if (fc != *eptr && foc != *eptr) break;
3404            eptr++;            eptr++;
3405            }            }
3406    
# Line 3194  for (;;) Line 3412  for (;;)
3412            eptr--;            eptr--;
3413            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3414            }            }
3415          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3416          }          }
3417        /* Control never gets here */        /* Control never gets here */
3418        }        }
# Line 3208  for (;;) Line 3426  for (;;)
3426          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3427            {            {
3428            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3429            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3430            }            }
3431          if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
3432          }          }
3433    
3434        if (min == max) continue;        if (min == max) continue;
# Line 3221  for (;;) Line 3439  for (;;)
3439            {            {
3440            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);
3441            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3442            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3443            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3444              {              {
3445              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3446              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3447              }              }
3448            if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
3449            }            }
3450          /* Control never gets here */          /* Control never gets here */
3451          }          }
# Line 3252  for (;;) Line 3470  for (;;)
3470            eptr--;            eptr--;
3471            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3472            }            }
3473          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3474          }          }
3475        }        }
3476      /* Control never gets here */      /* Control never gets here */
# Line 3261  for (;;) Line 3479  for (;;)
3479      checking can be multibyte. */      checking can be multibyte. */
3480    
3481      case OP_NOT:      case OP_NOT:
3482      case OP_NOTI:      case OP_NOTI:
3483      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
3484        {        {
3485        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3486        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3487        }        }
3488      ecode++;      ecode++;
3489      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
3490      if (op == OP_NOTI)         /* The caseless case */      if (op == OP_NOTI)         /* The caseless case */
3491        {        {
3492  #ifdef SUPPORT_UTF8        register int ch, och;
3493        if (c < 256)        ch = *ecode++;
3494  #endif  #ifdef COMPILE_PCRE8
3495        c = md->lcc[c];        /* ch must be < 128 if UTF is enabled. */
3496        if (md->lcc[*ecode++] == c) MRRETURN(MATCH_NOMATCH);        och = md->fcc[ch];
3497    #else
3498    #ifdef SUPPORT_UTF
3499    #ifdef SUPPORT_UCP
3500          if (utf && ch > 127)
3501            och = UCD_OTHERCASE(ch);
3502    #else
3503          if (utf && ch > 127)
3504            och = ch;
3505    #endif /* SUPPORT_UCP */
3506          else
3507    #endif /* SUPPORT_UTF */
3508            och = TABLE_GET(ch, md->fcc, ch);
3509    #endif /* COMPILE_PCRE8 */
3510          if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3511        }        }
3512      else    /* Caseful */      else    /* Caseful */
3513        {        {
3514        if (*ecode++ == c) MRRETURN(MATCH_NOMATCH);        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);
3515        }        }
3516      break;      break;
3517    
# Line 3293  for (;;) Line 3525  for (;;)
3525      case OP_NOTEXACT:      case OP_NOTEXACT:
3526      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3527      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3528      ecode += 3;      ecode += 1 + IMM2_SIZE;
3529      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3530    
3531      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3303  for (;;) Line 3535  for (;;)
3535      min = 0;      min = 0;
3536      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3537      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3538      ecode += 3;      ecode += 1 + IMM2_SIZE;
3539      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3540    
3541      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3335  for (;;) Line 3567  for (;;)
3567      possessive = TRUE;      possessive = TRUE;
3568      min = 0;      min = 0;
3569      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3570      ecode += 3;      ecode += 1 + IMM2_SIZE;
3571      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3572    
3573      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3374  for (;;) Line 3606  for (;;)
3606    
3607      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3608        {        {
3609        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3610          /* fc must be < 128 if UTF is enabled. */
3611          foc = md->fcc[fc];
3612    #else
3613    #ifdef SUPPORT_UTF
3614    #ifdef SUPPORT_UCP
3615          if (utf && fc > 127)
3616            foc = UCD_OTHERCASE(fc);
3617    #else
3618          if (utf && fc > 127)
3619            foc = fc;
3620    #endif /* SUPPORT_UCP */
3621          else
3622    #endif /* SUPPORT_UTF */
3623            foc = TABLE_GET(fc, md->fcc, fc);
3624    #endif /* COMPILE_PCRE8 */
3625    
3626  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3627        /* UTF-8 mode */        if (utf)
       if (utf8)  
3628          {          {
3629          register unsigned int d;          register unsigned int d;
3630          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3386  for (;;) Line 3632  for (;;)
3632            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3633              {              {
3634              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3635              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3636              }              }
3637            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3638            if (d < 256) d = md->lcc[d];            if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) MRRETURN(MATCH_NOMATCH);  
3639            }            }
3640          }          }
3641        else        else
3642  #endif  #endif
3643          /* Not UTF mode */
       /* Not UTF-8 mode */  
3644          {          {
3645          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3646            {            {
3647            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3648              {              {
3649              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3650              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3651              }              }
3652            if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3653              eptr++;
3654            }            }
3655          }          }
3656    
# Line 3413  for (;;) Line 3658  for (;;)
3658    
3659        if (minimize)        if (minimize)
3660          {          {
3661  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3662          /* UTF-8 mode */          if (utf)
         if (utf8)  
3663            {            {
3664            register unsigned int d;            register unsigned int d;
3665            for (fi = min;; fi++)            for (fi = min;; fi++)
3666              {              {
3667              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
3668              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3669              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3670              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3671                {                {
3672                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3673                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3674                }                }
3675              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3676              if (d < 256) d = md->lcc[d];              if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) MRRETURN(MATCH_NOMATCH);  
3677              }              }
3678            }            }
3679          else          else
3680  #endif  #endif
3681          /* Not UTF-8 mode */          /* Not UTF mode */
3682            {            {
3683            for (fi = min;; fi++)            for (fi = min;; fi++)
3684              {              {
3685              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);
3686              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3687              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3688              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3689                {                {
3690                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3691                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3692                }                }
3693              if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3694                eptr++;
3695              }              }
3696            }            }
3697          /* Control never gets here */          /* Control never gets here */
# Line 3459  for (;;) Line 3703  for (;;)
3703          {          {
3704          pp = eptr;          pp = eptr;
3705    
3706  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3707          /* UTF-8 mode */          if (utf)
         if (utf8)  
3708            {            {
3709            register unsigned int d;            register unsigned int d;
3710            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3473  for (;;) Line 3716  for (;;)
3716                break;                break;
3717                }                }
3718              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3719              if (d < 256) d = md->lcc[d];              if (fc == d || foc == d) break;
             if (fc == d) break;  
3720              eptr += len;              eptr += len;
3721              }              }
3722          if (possessive) continue;          if (possessive) continue;
# Line 3488  for (;;) Line 3730  for (;;)
3730            }            }
3731          else          else
3732  #endif  #endif
3733          /* Not UTF-8 mode */          /* Not UTF mode */
3734            {            {
3735            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3736              {              {
# Line 3497  for (;;) Line 3739  for (;;)
3739                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3740                break;                break;
3741                }                }
3742              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3743              eptr++;              eptr++;
3744              }              }
3745            if (possessive) continue;            if (possessive) continue;
# Line 3509  for (;;) Line 3751  for (;;)
3751              }              }
3752            }            }
3753    
3754          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3755          }          }
3756        /* Control never gets here */        /* Control never gets here */
3757        }        }
# Line 3518  for (;;) Line 3760  for (;;)
3760    
3761      else      else
3762        {        {
3763  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3764        /* UTF-8 mode */        if (utf)
       if (utf8)  
3765          {          {
3766          register unsigned int d;          register unsigned int d;
3767          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3528  for (;;) Line 3769  for (;;)
3769            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3770              {              {
3771              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3772              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3773              }              }
3774            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3775            if (fc == d) MRRETURN(MATCH_NOMATCH);            if (fc == d) RRETURN(MATCH_NOMATCH);
3776            }            }
3777          }          }
3778        else        else
3779  #endif  #endif
3780        /* Not UTF-8 mode */        /* Not UTF mode */
3781          {          {
3782          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3783            {            {
3784            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3785              {              {
3786              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3787              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3788              }              }
3789            if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);            if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
3790            }            }
3791          }          }
3792    
# Line 3553  for (;;) Line 3794  for (;;)
3794    
3795        if (minimize)        if (minimize)
3796          {          {
3797  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3798          /* UTF-8 mode */          if (utf)
         if (utf8)  
3799            {            {
3800            register unsigned int d;            register unsigned int d;
3801            for (fi = min;; fi++)            for (fi = min;; fi++)
3802              {              {
3803              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
3804              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3805              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3806              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3807                {                {
3808                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3809                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3810                }                }
3811              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3812              if (fc == d) MRRETURN(MATCH_NOMATCH);              if (fc == d) RRETURN(MATCH_NOMATCH);
3813              }              }
3814            }            }
3815          else          else
3816  #endif  #endif
3817          /* Not UTF-8 mode */          /* Not UTF mode */
3818            {            {
3819            for (fi = min;; fi++)            for (fi = min;; fi++)
3820              {              {
3821              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);
3822              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3823              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3824              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3825                {                {
3826                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3827                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3828                }                }
3829              if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);              if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
3830              }              }
3831            }            }
3832          /* Control never gets here */          /* Control never gets here */
# Line 3598  for (;;) Line 3838  for (;;)
3838          {          {
3839          pp = eptr;          pp = eptr;
3840    
3841  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3842          /* UTF-8 mode */          if (utf)
         if (utf8)  
3843            {            {
3844            register unsigned int d;            register unsigned int d;
3845            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3626  for (;;) Line 3865  for (;;)
3865            }            }
3866          else          else
3867  #endif  #endif
3868          /* Not UTF-8 mode */          /* Not UTF mode */
3869            {            {
3870            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3871              {              {
# Line 3647  for (;;) Line 3886  for (;;)
3886              }              }
3887            }            }
3888    
3889          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3890          }          }
3891        }        }
3892      /* Control never gets here */      /* Control never gets here */
# Line 3659  for (;;) Line 3898  for (;;)
3898      case OP_TYPEEXACT:      case OP_TYPEEXACT:
3899      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3900      minimize = TRUE;      minimize = TRUE;
3901      ecode += 3;      ecode += 1 + IMM2_SIZE;
3902      goto REPEATTYPE;      goto REPEATTYPE;
3903    
3904      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3667  for (;;) Line 3906  for (;;)
3906      min = 0;      min = 0;
3907      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3908      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
3909      ecode += 3;      ecode += 1 + IMM2_SIZE;
3910      goto REPEATTYPE;      goto REPEATTYPE;
3911    
3912      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3695  for (;;) Line 3934  for (;;)
3934      possessive = TRUE;      possessive = TRUE;
3935      min = 0;      min = 0;
3936      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3937      ecode += 3;      ecode += 1 + IMM2_SIZE;
3938      goto REPEATTYPE;      goto REPEATTYPE;
3939    
3940      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 3741  for (;;) Line 3980  for (;;)
3980          switch(prop_type)          switch(prop_type)
3981            {            {
3982            case PT_ANY:            case PT_ANY:
3983            if (prop_fail_result) MRRETURN(MATCH_NOMATCH);            if (prop_fail_result) RRETURN(MATCH_NOMATCH);
3984            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3985              {              {
3986              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3987                {                {
3988                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3989                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3990                }                }
3991              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3992              }              }
# Line 3756  for (;;) Line 3995  for (;;)
3995            case PT_LAMP:            case PT_LAMP:
3996            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3997              {              {
3998                int chartype;
3999              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4000                {                {
4001                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4002                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4003                }                }
4004              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4005              prop_chartype = UCD_CHARTYPE(c);              chartype = UCD_CHARTYPE(c);
4006              if ((prop_chartype == ucp_Lu ||              if ((chartype == ucp_Lu ||
4007                   prop_chartype == ucp_Ll ||                   chartype == ucp_Ll ||
4008                   prop_chartype == ucp_Lt) == prop_fail_result)                   chartype == ucp_Lt) == prop_fail_result)
4009                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4010              }              }
4011            break;            break;
4012    
# Line 3776  for (;;) Line 4016  for (;;)
4016              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4017                {                {
4018                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4019                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4020                }                }
4021              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4022              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)
4023              if ((prop_category == prop_value) == prop_fail_result)                RRETURN(MATCH_NOMATCH);
               MRRETURN(MATCH_NOMATCH);  
4024              }              }
4025            break;            break;
4026    
# Line 3791  for (;;) Line 4030  for (;;)
4030              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4031                {                {
4032                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4033                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4034                }                }
4035              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4036              prop_chartype = UCD_CHARTYPE(c);              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)
4037              if ((prop_chartype == prop_value) == prop_fail_result)                RRETURN(MATCH_NOMATCH);
               MRRETURN(MATCH_NOMATCH);  
4038              }              }
4039            break;            break;
4040    
# Line 3806  for (;;) Line 4044  for (;;)
4044              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4045                {                {
4046                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4047                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4048                }                }
4049              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4050              prop_script = UCD_SCRIPT(c);              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)
4051              if ((prop_script == prop_value) == prop_fail_result)                RRETURN(MATCH_NOMATCH);
               MRRETURN(MATCH_NOMATCH);  
4052              }              }
4053            break;            break;
4054    
4055            case PT_ALNUM:            case PT_ALNUM:
4056            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4057              {              {
4058                int category;
4059              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4060                {                {
4061                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4062                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4063                }                }
4064              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4065              prop_category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4066              if ((prop_category == ucp_L || prop_category == ucp_N)              if ((category == ucp_L || category == ucp_N) == prop_fail_result)
4067                     == prop_fail_result)                RRETURN(MATCH_NOMATCH);
               MRRETURN(MATCH_NOMATCH);  
4068              }              }
4069            break;            break;
4070    
# Line 3837  for (;;) Line 4074  for (;;)
4074              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4075                {                {
4076                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4077                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4078                }                }
4079              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4080              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 ||  
4081                   c == CHAR_FF || c == CHAR_CR)                   c == CHAR_FF || c == CHAR_CR)
4082                     == prop_fail_result)                     == prop_fail_result)
4083                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4084              }              }
4085            break;            break;
4086    
# Line 3854  for (;;) Line 4090  for (;;)
4090              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4091                {                {
4092                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4093                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4094                }                }
4095              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4096              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 ||  
4097                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
4098                     == prop_fail_result)                     == prop_fail_result)
4099                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4100              }              }
4101            break;            break;
4102    
4103            case PT_WORD:            case PT_WORD:
4104            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4105              {              {
4106                int category;
4107              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4108                {                {
4109                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4110                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4111                }                }
4112              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4113              prop_category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4114              if ((prop_category == ucp_L || prop_category == ucp_N ||              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)
                  c == CHAR_UNDERSCORE)  
4115                     == prop_fail_result)                     == prop_fail_result)
4116                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4117              }              }
4118            break;            break;
4119    
# Line 3899  for (;;) Line 4134  for (;;)
4134            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4135              {              {
4136              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4137              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4138              }              }
4139            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
4140            prop_category = UCD_CATEGORY(c);            if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);
           if (prop_category == ucp_M) MRRETURN(MATCH_NOMATCH);  
4141            while (eptr < md->end_subject)            while (eptr < md->end_subject)
4142              {              {
4143              int len = 1;              int len = 1;
4144              if (!utf8) c = *eptr;              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4145                else { GETCHARLEN(c, eptr, len); }              if (UCD_CATEGORY(c) != ucp_M) break;
             prop_category = UCD_CATEGORY(c);  
             if (prop_category != ucp_M) break;  
4146              eptr += len;              eptr += len;
4147              }              }
4148            }            }
# Line 3921  for (;;) Line 4153  for (;;)
4153    
4154  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4155    
4156  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4157        if (utf8) switch(ctype)        if (utf) switch(ctype)
4158          {          {
4159          case OP_ANY:          case OP_ANY:
4160          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3930  for (;;) Line 4162  for (;;)
4162            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4163              {              {
4164              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4165              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4166              }              }
4167            if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4168            eptr++;            eptr++;
4169            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4170            }            }
4171          break;          break;
4172    
# Line 3944  for (;;) Line 4176  for (;;)
4176            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4177              {              {
4178              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4179              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);