/[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 915 by zherczeg, Tue Feb 14 13:05:39 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;  int matched_length = length;
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        PCRE_PUCHAR eptr_start = eptr;
190      while (p < endptr)      while (p < endptr)
191        {        {
192        int c, d;        int c, d;
193        if (eptr >= md->end_subject) return -1;        if (eptr >= md->end_subject) return -((int)(eptr - eptr_start) + 1);
194        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
195        GETCHARINC(d, p);        GETCHARINC(d, p);
196        if (c != d && c != UCD_OTHERCASE(d)) return -1;        if (c != d && c != UCD_OTHERCASE(d)) return -1;
197        }        }
198        matched_length = (int)(eptr - eptr_start);
199      }      }
200    else    else
201  #endif  #endif
# Line 209  if (caseless) Line 204  if (caseless)
204    /* 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
205    is no UCP support. */    is no UCP support. */
206      {      {
207      if (eptr + length > md->end_subject) return -1;      if (eptr + length > md->end_subject)
208          {
209          if (md->partial == 0)
210            return -1;
211          length = (int)(md->end_subject - eptr);
212          matched_length = -(length + 1);
213          }
214      while (length-- > 0)      while (length-- > 0)
215        { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }        {
216      }        if (TABLE_GET(*p, md->lcc, *p) != TABLE_GET(*eptr, md->lcc, *eptr)) return -1;
217          p++;
218          eptr++;
219          }
220        }
221    }    }
222    
223  /* 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
224  are in UTF-8 mode. */  are in UTF-8 mode. */
225    
226  else  else
227    {    {
228    if (eptr + length > md->end_subject) return -1;    if (eptr + length > md->end_subject)
229    while (length-- > 0) if (*p++ != *eptr++) return -1;      {
230        if (md->partial == 0)
231          return -1;
232        length = (int)(md->end_subject - eptr);
233        matched_length = -(length + 1);
234        }
235      while (length-- > 0) if (*p++ != *eptr++) return -1;
236    }    }
237    
238  return eptr - eptr_start;  return matched_length;
239  }  }
240    
241    
# Line 276  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 287  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
287         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
288         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
289         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
290         RM61,  RM62, RM63, RM64 };         RM61,  RM62, RM63, RM64, RM65, RM66 };
291    
292  /* 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
293  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 300  actually used in this definition. */
300  #define RMATCH(ra,rb,rc,rd,re,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
301    { \    { \
302    printf("match() called in line %d\n", __LINE__); \    printf("match() called in line %d\n", __LINE__); \
303    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1); \    rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1); \
304    printf("to line %d\n", __LINE__); \    printf("to line %d\n", __LINE__); \
305    }    }
306  #define RRETURN(ra) \  #define RRETURN(ra) \
# Line 299  actually used in this definition. */ Line 310  actually used in this definition. */
310    }    }
311  #else  #else
312  #define RMATCH(ra,rb,rc,rd,re,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
313    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1)    rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1)
314  #define RRETURN(ra) return ra  #define RRETURN(ra) return ra
315  #endif  #endif
316    
# Line 314  argument of match(), which never changes Line 325  argument of match(), which never changes
325    
326  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
327    {\    {\
328    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
329    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
330    frame->Xwhere = rw; \    frame->Xwhere = rw; \
331    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
332    newframe->Xecode = rb;\    newframe->Xecode = rb;\
333    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
   newframe->Xmarkptr = markptr;\  
334    newframe->Xoffset_top = rc;\    newframe->Xoffset_top = rc;\
335    newframe->Xeptrb = re;\    newframe->Xeptrb = re;\
336    newframe->Xrdepth = frame->Xrdepth + 1;\    newframe->Xrdepth = frame->Xrdepth + 1;\
# Line 336  argument of match(), which never changes Line 346  argument of match(), which never changes
346    {\    {\
347    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
348    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
349    (pcre_stack_free)(oldframe);\    if (oldframe != &frame_zero) (PUBL(stack_free))(oldframe);\
350    if (frame != NULL)\    if (frame != NULL)\
351      {\      {\
352      rrc = ra;\      rrc = ra;\
# Line 353  typedef struct heapframe { Line 363  typedef struct heapframe {
363    
364    /* Function arguments that may change */    /* Function arguments that may change */
365    
366    USPTR Xeptr;    PCRE_PUCHAR Xeptr;
367    const uschar *Xecode;    const pcre_uchar *Xecode;
368    USPTR Xmstart;    PCRE_PUCHAR Xmstart;
   USPTR Xmarkptr;  
369    int Xoffset_top;    int Xoffset_top;
370    eptrblock *Xeptrb;    eptrblock *Xeptrb;
371    unsigned int Xrdepth;    unsigned int Xrdepth;
372    
373    /* Function local variables */    /* Function local variables */
374    
375    USPTR Xcallpat;    PCRE_PUCHAR Xcallpat;
376  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
377    USPTR Xcharptr;    PCRE_PUCHAR Xcharptr;
378  #endif  #endif
379    USPTR Xdata;    PCRE_PUCHAR Xdata;
380    USPTR Xnext;    PCRE_PUCHAR Xnext;
381    USPTR Xpp;    PCRE_PUCHAR Xpp;
382    USPTR Xprev;    PCRE_PUCHAR Xprev;
383    USPTR Xsaved_eptr;    PCRE_PUCHAR Xsaved_eptr;
384    
385    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
386    
# Line 383  typedef struct heapframe { Line 392  typedef struct heapframe {
392    int Xprop_type;    int Xprop_type;
393    int Xprop_value;    int Xprop_value;
394    int Xprop_fail_result;    int Xprop_fail_result;
   int Xprop_category;  
   int Xprop_chartype;  
   int Xprop_script;  
395    int Xoclength;    int Xoclength;
396    uschar Xocchars[8];    pcre_uchar Xocchars[6];
397  #endif  #endif
398    
399    int Xcodelink;    int Xcodelink;
# Line 429  returns a negative (error) response, the Line 435  returns a negative (error) response, the
435  same response. */  same response. */
436    
437  /* 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
438  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
439  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.
440  something has been matched). For hard partial matching, we then return  something has been matched). For hard partial matching, we then return
441  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 446  the subject. */
446        eptr > md->start_used_ptr) \        eptr > md->start_used_ptr) \
447      { \      { \
448      md->hitend = TRUE; \      md->hitend = TRUE; \
449      if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \      if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \
450      }      }
451    
452  #define SCHECK_PARTIAL()\  #define SCHECK_PARTIAL()\
453    if (md->partial != 0 && eptr > md->start_used_ptr) \    if (md->partial != 0 && eptr > md->start_used_ptr) \
454      { \      { \
455      md->hitend = TRUE; \      md->hitend = TRUE; \
456      if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \      if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \
457      }      }
458    
459    
460  /* Performance note: It might be tempting to extract commonly used fields from  /* Performance note: It might be tempting to extract commonly used fields from
461  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
462  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
463  made performance worse.  made performance worse.
464    
# Line 461  Arguments: Line 467  Arguments:
467     ecode       pointer to current position in compiled code     ecode       pointer to current position in compiled code
468     mstart      pointer to the current match start position (can be modified     mstart      pointer to the current match start position (can be modified
469                   by encountering \K)                   by encountering \K)
    markptr     pointer to the most recent MARK name, or NULL  
470     offset_top  current top pointer     offset_top  current top pointer
471     md          pointer to "static" info for the match     md          pointer to "static" info for the match
472     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 481  Returns:       MATCH_MATCH if matched
481  */  */
482    
483  static int  static int
484  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode,
485    const uschar *markptr, int offset_top, match_data *md, eptrblock *eptrb,    PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb,
486    unsigned int rdepth)    unsigned int rdepth)
487  {  {
488  /* 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 492  so they can be ordinary variables in all
492  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
493  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
494  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register unsigned int c;   /* Character values not kept over RMATCH() calls */
495  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
496    
497  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
498  BOOL caseless;  BOOL caseless;
499  int condcode;  int condcode;
500    
501  /* When recursion is not being used, all "local" variables that have to be  /* When recursion is not being used, all "local" variables that have to be
502  preserved over calls to RMATCH() are part of a "frame" which is obtained from  preserved over calls to RMATCH() are part of a "frame". We set up the top-level
503  heap storage. Set up the top-level frame here; others are obtained from the  frame on the stack here; subsequent instantiations are obtained from the heap
504  heap whenever RMATCH() does a "recursion". See the macro definitions above. */  whenever RMATCH() does a "recursion". See the macro definitions above. Putting
505    the top-level on the stack rather than malloc-ing them all gives a performance
506    boost in many cases where there is not much "recursion". */
507    
508  #ifdef NO_RECURSE  #ifdef NO_RECURSE
509  heapframe *frame = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));  heapframe frame_zero;
510  if (frame == NULL) RRETURN(PCRE_ERROR_NOMEMORY);  heapframe *frame = &frame_zero;
511  frame->Xprevframe = NULL;            /* Marks the top level */  frame->Xprevframe = NULL;            /* Marks the top level */
512    
513  /* Copy in the original argument variables */  /* Copy in the original argument variables */
# Line 508  frame->Xprevframe = NULL;            /* Line 515  frame->Xprevframe = NULL;            /*
515  frame->Xeptr = eptr;  frame->Xeptr = eptr;
516  frame->Xecode = ecode;  frame->Xecode = ecode;
517  frame->Xmstart = mstart;  frame->Xmstart = mstart;
 frame->Xmarkptr = markptr;  
518  frame->Xoffset_top = offset_top;  frame->Xoffset_top = offset_top;
519  frame->Xeptrb = eptrb;  frame->Xeptrb = eptrb;
520  frame->Xrdepth = rdepth;  frame->Xrdepth = rdepth;
# Line 522  HEAP_RECURSE: Line 528  HEAP_RECURSE:
528  #define eptr               frame->Xeptr  #define eptr               frame->Xeptr
529  #define ecode              frame->Xecode  #define ecode              frame->Xecode
530  #define mstart             frame->Xmstart  #define mstart             frame->Xmstart
 #define markptr            frame->Xmarkptr  
531  #define offset_top         frame->Xoffset_top  #define offset_top         frame->Xoffset_top
532  #define eptrb              frame->Xeptrb  #define eptrb              frame->Xeptrb
533  #define rdepth             frame->Xrdepth  #define rdepth             frame->Xrdepth
534    
535  /* Ditto for the local variables */  /* Ditto for the local variables */
536    
537  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
538  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
539  #endif  #endif
540  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
# Line 550  HEAP_RECURSE: Line 555  HEAP_RECURSE:
555  #define prop_type          frame->Xprop_type  #define prop_type          frame->Xprop_type
556  #define prop_value         frame->Xprop_value  #define prop_value         frame->Xprop_value
557  #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  
558  #define oclength           frame->Xoclength  #define oclength           frame->Xoclength
559  #define occhars            frame->Xocchars  #define occhars            frame->Xocchars
560  #endif  #endif
# Line 590  declarations can be cut out in a block. Line 592  declarations can be cut out in a block.
592  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
593  to RMATCH(). */  to RMATCH(). */
594    
595  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
596  const uschar *charptr;  const pcre_uchar *charptr;
597  #endif  #endif
598  const uschar *callpat;  const pcre_uchar *callpat;
599  const uschar *data;  const pcre_uchar *data;
600  const uschar *next;  const pcre_uchar *next;
601  USPTR         pp;  PCRE_PUCHAR       pp;
602  const uschar *prev;  const pcre_uchar *prev;
603  USPTR         saved_eptr;  PCRE_PUCHAR       saved_eptr;
604    
605  recursion_info new_recursive;  recursion_info new_recursive;
606    
607  BOOL cur_is_word;  BOOL cur_is_word;
608  BOOL condition;  BOOL condition;
609  BOOL prev_is_word;  BOOL prev_is_word;
610    
# Line 610  BOOL prev_is_word; Line 612  BOOL prev_is_word;
612  int prop_type;  int prop_type;
613  int prop_value;  int prop_value;
614  int prop_fail_result;  int prop_fail_result;
 int prop_category;  
 int prop_chartype;  
 int prop_script;  
615  int oclength;  int oclength;
616  uschar occhars[8];  pcre_uchar occhars[6];
617  #endif  #endif
618    
619  int codelink;  int codelink;
# Line 630  int save_offset1, save_offset2, save_off Line 629  int save_offset1, save_offset2, save_off
629  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
630    
631  eptrblock newptrb;  eptrblock newptrb;
632    
633    /* There is a special fudge for calling match() in a way that causes it to
634    measure the size of its basic stack frame when the stack is being used for
635    recursion. The second argument (ecode) being NULL triggers this behaviour. It
636    cannot normally ever be NULL. The return is the negated value of the frame
637    size. */
638    
639    if (ecode == NULL)
640      {
641      if (rdepth == 0)
642        return match((PCRE_PUCHAR)&rdepth, NULL, NULL, 0, NULL, NULL, 1);
643      else
644        {
645        int len = (char *)&rdepth - (char *)eptr;
646        return (len > 0)? -len : len;
647        }
648      }
649  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
650    
651  /* 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
652  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
653  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
654  the alternative names that are used. */  the alternative names that are used. */
655    
656  #define allow_zero    cur_is_word  #define allow_zero    cur_is_word
# Line 642  the alternative names that are used. */ Line 658  the alternative names that are used. */
658  #define code_offset   codelink  #define code_offset   codelink
659  #define condassert    condition  #define condassert    condition
660  #define matched_once  prev_is_word  #define matched_once  prev_is_word
661    #define foc           number
662    #define save_mark     data
663    
664  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
665  variables. */  variables. */
# Line 667  defined). However, RMATCH isn't like a f Line 685  defined). However, RMATCH isn't like a f
685  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,
686  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
687    
688  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
689  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
690  #else  #else
691  utf8 = FALSE;  utf = FALSE;
692  #endif  #endif
693    
694  /* 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 698  if (md->match_call_count++ >= md->match_
698  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);
699    
700  /* 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
701  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
702  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
703  up space on the stack. See also MATCH_CONDASSERT below.  up space on the stack. See also MATCH_CONDASSERT below.
704    
705  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 723  for (;;)
723    {    {
724    minimize = possessive = FALSE;    minimize = possessive = FALSE;
725    op = *ecode;    op = *ecode;
726    
727    switch(op)    switch(op)
728      {      {
729      case OP_MARK:      case OP_MARK:
730      markptr = ecode + 2;      md->nomatch_mark = ecode + 2;
731      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      md->mark = NULL;    /* In case previously set by assertion */
732        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
733        eptrb, RM55);        eptrb, RM55);
734        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
735             md->mark == NULL) md->mark = ecode + 2;
736    
737      /* 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
738      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 741  for (;;)
741      position and return MATCH_SKIP. Otherwise, pass back the return code      position and return MATCH_SKIP. Otherwise, pass back the return code
742      unaltered. */      unaltered. */
743    
744      if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
745          strcmp((char *)markptr, (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC(ecode + 2, md->start_match_ptr) == 0)
746        {        {
747        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
748        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
749        }        }
   
     if (md->mark == NULL) md->mark = markptr;  
750      RRETURN(rrc);      RRETURN(rrc);
751    
752      case OP_FAIL:      case OP_FAIL:
753      MRRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
754    
755      /* COMMIT overrides PRUNE, SKIP, and THEN */      /* COMMIT overrides PRUNE, SKIP, and THEN */
756    
757      case OP_COMMIT:      case OP_COMMIT:
758      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
759        eptrb, RM52);        eptrb, RM52);
760      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&
761          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&
762          rrc != MATCH_THEN)          rrc != MATCH_THEN)
763        RRETURN(rrc);        RRETURN(rrc);
764      MRRETURN(MATCH_COMMIT);      RRETURN(MATCH_COMMIT);
765    
766      /* PRUNE overrides THEN */      /* PRUNE overrides THEN */
767    
768      case OP_PRUNE:      case OP_PRUNE:
769      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
770        eptrb, RM51);        eptrb, RM51);
771      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
772      MRRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
773    
774      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
775      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      md->nomatch_mark = ecode + 2;
776        md->mark = NULL;    /* In case previously set by assertion */
777        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
778        eptrb, RM56);        eptrb, RM56);
779        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
780             md->mark == NULL) md->mark = ecode + 2;
781      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
     md->mark = ecode + 2;  
782      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
783    
784      /* SKIP overrides PRUNE and THEN */      /* SKIP overrides PRUNE and THEN */
785    
786      case OP_SKIP:      case OP_SKIP:
787      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
788        eptrb, RM53);        eptrb, RM53);
789      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
790        RRETURN(rrc);        RRETURN(rrc);
791      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
792      MRRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
793    
794        /* Note that, for Perl compatibility, SKIP with an argument does NOT set
795        nomatch_mark. There is a flag that disables this opcode when re-matching a
796        pattern that ended with a SKIP for which there was not a matching MARK. */
797    
798      case OP_SKIP_ARG:      case OP_SKIP_ARG:
799      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      if (md->ignore_skip_arg)
800          {
801          ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
802          break;
803          }
804        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
805        eptrb, RM57);        eptrb, RM57);
806      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
807        RRETURN(rrc);        RRETURN(rrc);
808    
809      /* 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
810      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
811      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
812      as PRUNE. */      with the md->ignore_skip_arg flag set. */
813    
814      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
815      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
816    
817      /* 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
818      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
819      to skip back past alternatives that precede the THEN within the current      match pointer to do this. */
     branch. */  
820    
821      case OP_THEN:      case OP_THEN:
822      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
823        eptrb, RM54);        eptrb, RM54);
824      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
825      md->start_match_ptr = ecode - GET(ecode, 1);      md->start_match_ptr = ecode;
826      MRRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
827    
828      case OP_THEN_ARG:      case OP_THEN_ARG:
829      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1+LINK_SIZE],      md->nomatch_mark = ecode + 2;
830        offset_top, md, eptrb, RM58);      md->mark = NULL;    /* In case previously set by assertion */
831        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
832          md, eptrb, RM58);
833        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
834             md->mark == NULL) md->mark = ecode + 2;
835      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
836      md->start_match_ptr = ecode - GET(ecode, 1);      md->start_match_ptr = ecode;
     md->mark = ecode + LINK_SIZE + 2;  
837      RRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
838    
839        /* Handle an atomic group that does not contain any capturing parentheses.
840        This can be handled like an assertion. Prior to 8.13, all atomic groups
841        were handled this way. In 8.13, the code was changed as below for ONCE, so
842        that backups pass through the group and thereby reset captured values.
843        However, this uses a lot more stack, so in 8.20, atomic groups that do not
844        contain any captures generate OP_ONCE_NC, which can be handled in the old,
845        less stack intensive way.
846    
847        Check the alternative branches in turn - the matching won't pass the KET
848        for this kind of subpattern. If any one branch matches, we carry on as at
849        the end of a normal bracket, leaving the subject pointer, but resetting
850        the start-of-match value in case it was changed by \K. */
851    
852        case OP_ONCE_NC:
853        prev = ecode;
854        saved_eptr = eptr;
855        save_mark = md->mark;
856        do
857          {
858          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
859          if (rrc == MATCH_MATCH)  /* Note: _not_ MATCH_ACCEPT */
860            {
861            mstart = md->start_match_ptr;
862            break;
863            }
864          if (rrc == MATCH_THEN)
865            {
866            next = ecode + GET(ecode,1);
867            if (md->start_match_ptr < next &&
868                (*ecode == OP_ALT || *next == OP_ALT))
869              rrc = MATCH_NOMATCH;
870            }
871    
872          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
873          ecode += GET(ecode,1);
874          md->mark = save_mark;
875          }
876        while (*ecode == OP_ALT);
877    
878        /* If hit the end of the group (which could be repeated), fail */
879    
880        if (*ecode != OP_ONCE_NC && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);
881    
882        /* Continue as from after the group, updating the offsets high water
883        mark, since extracts may have been taken. */
884    
885        do ecode += GET(ecode, 1); while (*ecode == OP_ALT);
886    
887        offset_top = md->end_offset_top;
888        eptr = md->end_match_ptr;
889    
890        /* For a non-repeating ket, just continue at this level. This also
891        happens for a repeating ket if no characters were matched in the group.
892        This is the forcible breaking of infinite loops as implemented in Perl
893        5.005. */
894    
895        if (*ecode == OP_KET || eptr == saved_eptr)
896          {
897          ecode += 1+LINK_SIZE;
898          break;
899          }
900    
901        /* The repeating kets try the rest of the pattern or restart from the
902        preceding bracket, in the appropriate order. The second "call" of match()
903        uses tail recursion, to avoid using another stack frame. */
904    
905        if (*ecode == OP_KETRMIN)
906          {
907          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM65);
908          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
909          ecode = prev;
910          goto TAIL_RECURSE;
911          }
912        else  /* OP_KETRMAX */
913          {
914          md->match_function_type = MATCH_CBEGROUP;
915          RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
916          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
917          ecode += 1 + LINK_SIZE;
918          goto TAIL_RECURSE;
919          }
920        /* Control never gets here */
921    
922      /* Handle a capturing bracket, other than those that are possessive with an      /* Handle a capturing bracket, other than those that are possessive with an
923      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
924      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
925      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
926      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
927      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,
928      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
929      of the same bracket.      the working value and also the values of the final offsets, in case they
930        were set by a previous iteration of the same bracket.
931    
932      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
933      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 937  for (;;)
937      case OP_SCBRA:      case OP_SCBRA:
938      number = GET2(ecode, 1+LINK_SIZE);      number = GET2(ecode, 1+LINK_SIZE);
939      offset = number << 1;      offset = number << 1;
940    
941  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
942      printf("start bracket %d\n", number);      printf("start bracket %d\n", number);
943      printf("subject=");      printf("subject=");
# Line 834  for (;;) Line 951  for (;;)
951        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
952        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
953        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
954          save_mark = md->mark;
955    
956        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
957        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 841  for (;;) Line 959  for (;;)
959    
960        for (;;)        for (;;)
961          {          {
962          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
963          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
964            eptrb, RM1);            eptrb, RM1);
965          if (rrc != MATCH_NOMATCH &&          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
966              (rrc != MATCH_THEN || md->start_match_ptr != ecode))  
967            RRETURN(rrc);          /* If we backed up to a THEN, check whether it is within the current
968            branch by comparing the address of the THEN that is passed back with
969            the end of the branch. If it is within the current branch, and the
970            branch is one of two or more alternatives (it either starts or ends
971            with OP_ALT), we have reached the limit of THEN's action, so convert
972            the return code to NOMATCH, which will cause normal backtracking to
973            happen from now on. Otherwise, THEN is passed back to an outer
974            alternative. This implements Perl's treatment of parenthesized groups,
975            where a group not containing | does not affect the current alternative,
976            that is, (X) is NOT the same as (X|(*F)). */
977    
978            if (rrc == MATCH_THEN)
979              {
980              next = ecode + GET(ecode,1);
981              if (md->start_match_ptr < next &&
982                  (*ecode == OP_ALT || *next == OP_ALT))
983                rrc = MATCH_NOMATCH;
984              }
985    
986            /* Anything other than NOMATCH is passed back. */
987    
988            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
989          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
990          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
991          if (*ecode != OP_ALT) break;          md->mark = save_mark;
992            if (*ecode != OP_ALT) break;
993          }          }
994    
995        DPRINTF(("bracket %d failed\n", number));        DPRINTF(("bracket %d failed\n", number));
   
996        md->offset_vector[offset] = save_offset1;        md->offset_vector[offset] = save_offset1;
997        md->offset_vector[offset+1] = save_offset2;        md->offset_vector[offset+1] = save_offset2;
998        md->offset_vector[md->offset_end - number] = save_offset3;        md->offset_vector[md->offset_end - number] = save_offset3;
999    
1000        if (rrc != MATCH_THEN) md->mark = markptr;        /* At this point, rrc will be one of MATCH_ONCE or MATCH_NOMATCH. */
1001        RRETURN(MATCH_NOMATCH);  
1002          RRETURN(rrc);
1003        }        }
1004    
1005      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
# Line 873  for (;;) Line 1013  for (;;)
1013      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1014      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1015    
1016      /* Non-capturing bracket, except for possessive with unlimited repeat. Loop      /* Non-capturing or atomic group, except for possessive with unlimited
1017      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.
1018      brackets, we would return the result of a recursive call to match()  
1019      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
1020      recursion, except in the case of a possibly empty group.*/      the result of a recursive call to match() whatever happened so it was
1021        possible to reduce stack usage by turning this into a tail recursion,
1022        except in the case of a possibly empty group. However, now that there is
1023        the possiblity of (*THEN) occurring in the final alternative, this
1024        optimization is no longer always possible.
1025    
1026        We can optimize if we know there are no (*THEN)s in the pattern; at present
1027        this is the best that can be done.
1028    
1029        MATCH_ONCE is returned when the end of an atomic group is successfully
1030        reached, but subsequent matching fails. It passes back up the tree (causing
1031        captured values to be reset) until the original atomic group level is
1032        reached. This is tested by comparing md->once_target with the start of the
1033        group. At this point, the return is converted into MATCH_NOMATCH so that
1034        previous backup points can be taken. */
1035    
1036        case OP_ONCE:
1037      case OP_BRA:      case OP_BRA:
1038      case OP_SBRA:      case OP_SBRA:
1039      DPRINTF(("start non-capturing bracket\n"));      DPRINTF(("start non-capturing bracket\n"));
1040    
1041      for (;;)      for (;;)
1042        {        {
1043        if (ecode[GET(ecode, 1)] != OP_ALT)   /* Final alternative */        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;
1044    
1045          /* If this is not a possibly empty group, and there are no (*THEN)s in
1046          the pattern, and this is the final alternative, optimize as described
1047          above. */
1048    
1049          else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1050          {          {
1051          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"));  
1052          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1053          }          }
1054    
1055        /* 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. */  
1056    
1057        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        save_mark = md->mark;
1058        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1059          RM2);          RM2);
1060        if (rrc != MATCH_NOMATCH &&  
1061            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* See comment in the code for capturing groups above about handling
1062          THEN. */
1063    
1064          if (rrc == MATCH_THEN)
1065            {
1066            next = ecode + GET(ecode,1);
1067            if (md->start_match_ptr < next &&
1068                (*ecode == OP_ALT || *next == OP_ALT))
1069              rrc = MATCH_NOMATCH;
1070            }
1071    
1072          if (rrc != MATCH_NOMATCH)
1073            {
1074            if (rrc == MATCH_ONCE)
1075              {
1076              const pcre_uchar *scode = ecode;
1077              if (*scode != OP_ONCE)           /* If not at start, find it */
1078                {
1079                while (*scode == OP_ALT) scode += GET(scode, 1);
1080                scode -= GET(scode, 1);
1081                }
1082              if (md->once_target == scode) rrc = MATCH_NOMATCH;
1083              }
1084          RRETURN(rrc);          RRETURN(rrc);
1085            }
1086        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1087          md->mark = save_mark;
1088          if (*ecode != OP_ALT) break;
1089        }        }
     /* Control never reaches here. */  
1090    
1091      /* Handle possessive capturing brackets with an unlimited repeat. We come      RRETURN(MATCH_NOMATCH);
1092    
1093        /* Handle possessive capturing brackets with an unlimited repeat. We come
1094      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
1095      handled similarly to the normal case above. However, the matching is      handled similarly to the normal case above. However, the matching is
1096      different. The end of these brackets will always be OP_KETRPOS, which      different. The end of these brackets will always be OP_KETRPOS, which
1097      returns MATCH_KETRPOS without going further in the pattern. By this means      returns MATCH_KETRPOS without going further in the pattern. By this means
1098      we can handle the group by iteration rather than recursion, thereby      we can handle the group by iteration rather than recursion, thereby
1099      reducing the amount of stack needed. */      reducing the amount of stack needed. */
1100    
1101      case OP_CBRAPOS:      case OP_CBRAPOS:
1102      case OP_SCBRAPOS:      case OP_SCBRAPOS:
1103      allow_zero = FALSE;      allow_zero = FALSE;
1104    
1105      POSSESSIVE_CAPTURE:      POSSESSIVE_CAPTURE:
1106      number = GET2(ecode, 1+LINK_SIZE);      number = GET2(ecode, 1+LINK_SIZE);
1107      offset = number << 1;      offset = number << 1;
# Line 939  for (;;) Line 1116  for (;;)
1116      if (offset < md->offset_max)      if (offset < md->offset_max)
1117        {        {
1118        matched_once = FALSE;        matched_once = FALSE;
1119        code_offset = ecode - md->start_code;        code_offset = (int)(ecode - md->start_code);
1120    
1121        save_offset1 = md->offset_vector[offset];        save_offset1 = md->offset_vector[offset];
1122        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
# Line 947  for (;;) Line 1124  for (;;)
1124        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
1125    
1126        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
1127    
1128        /* Each time round the loop, save the current subject position for use        /* Each time round the loop, save the current subject position for use
1129        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
1130        restart it with a new subject starting position, remembering that we had        restart it with a new subject starting position, remembering that we had
1131        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
1132        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
1133        see if a previous iteration matched. If so, the group has matched;        see if a previous iteration matched. If so, the group has matched;
1134        continue from afterwards. Otherwise it has failed; restore the previous        continue from afterwards. Otherwise it has failed; restore the previous
1135        capture values before returning NOMATCH. */        capture values before returning NOMATCH. */
1136    
1137        for (;;)        for (;;)
1138          {          {
1139          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1140            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1141          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1142          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1143            eptrb, RM63);            eptrb, RM63);
1144          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1145            {            {
1146            offset_top = md->end_offset_top;            offset_top = md->end_offset_top;
1147            eptr = md->end_match_ptr;            eptr = md->end_match_ptr;
1148            ecode = md->start_code + code_offset;            ecode = md->start_code + code_offset;
1149            save_capture_last = md->capture_last;            save_capture_last = md->capture_last;
1150            matched_once = TRUE;            matched_once = TRUE;
1151            continue;            continue;
1152            }            }
1153          if (rrc != MATCH_NOMATCH &&  
1154              (rrc != MATCH_THEN || md->start_match_ptr != ecode))          /* See comment in the code for capturing groups above about handling
1155            RRETURN(rrc);          THEN. */
1156    
1157            if (rrc == MATCH_THEN)
1158              {
1159              next = ecode + GET(ecode,1);
1160              if (md->start_match_ptr < next &&
1161                  (*ecode == OP_ALT || *next == OP_ALT))
1162                rrc = MATCH_NOMATCH;
1163              }
1164    
1165            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1166          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
1167          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
1168          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
1169          }          }
1170    
1171        if (!matched_once)        if (!matched_once)
1172          {          {
1173          md->offset_vector[offset] = save_offset1;          md->offset_vector[offset] = save_offset1;
1174          md->offset_vector[offset+1] = save_offset2;          md->offset_vector[offset+1] = save_offset2;
1175          md->offset_vector[md->offset_end - number] = save_offset3;          md->offset_vector[md->offset_end - number] = save_offset3;
1176          }          }
1177    
       if (rrc != MATCH_THEN) md->mark = markptr;  
1178        if (allow_zero || matched_once)        if (allow_zero || matched_once)
1179          {          {
1180          ecode += 1 + LINK_SIZE;          ecode += 1 + LINK_SIZE;
1181          break;          break;
1182          }          }
1183    
1184        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1185        }        }
1186    
1187      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
1188      as a non-capturing bracket. */      as a non-capturing bracket. */
1189    
# Line 1009  for (;;) Line 1195  for (;;)
1195      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1196      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1197    
1198      /* Non-capturing possessive bracket with unlimited repeat. We come here      /* Non-capturing possessive bracket with unlimited repeat. We come here
1199      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,
1200      without the capturing complication. It is written out separately for speed      without the capturing complication. It is written out separately for speed
1201      and cleanliness. */      and cleanliness. */
1202    
1203      case OP_BRAPOS:      case OP_BRAPOS:
1204      case OP_SBRAPOS:      case OP_SBRAPOS:
1205      allow_zero = FALSE;      allow_zero = FALSE;
1206    
1207      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1208      matched_once = FALSE;      matched_once = FALSE;
1209      code_offset = ecode - md->start_code;      code_offset = (int)(ecode - md->start_code);
1210    
1211      for (;;)      for (;;)
1212        {        {
1213        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1214        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1215          eptrb, RM64);          eptrb, RM48);
1216        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1217          {          {
1218            offset_top = md->end_offset_top;
1219          eptr = md->end_match_ptr;          eptr = md->end_match_ptr;
1220          ecode = md->start_code + code_offset;          ecode = md->start_code + code_offset;
1221          matched_once = TRUE;          matched_once = TRUE;
1222          continue;          continue;
1223          }          }
1224        if (rrc != MATCH_NOMATCH &&  
1225            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* See comment in the code for capturing groups above about handling
1226          RRETURN(rrc);        THEN. */
1227    
1228          if (rrc == MATCH_THEN)
1229            {
1230            next = ecode + GET(ecode,1);
1231            if (md->start_match_ptr < next &&
1232                (*ecode == OP_ALT || *next == OP_ALT))
1233              rrc = MATCH_NOMATCH;
1234            }
1235    
1236          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1237        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1238        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1239        }        }
1240    
1241      if (matched_once || allow_zero)      if (matched_once || allow_zero)
1242        {        {
1243        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1244        break;        break;
1245        }        }
1246      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
1247    
1248      /* Control never reaches here. */      /* Control never reaches here. */
# Line 1053  for (;;) Line 1250  for (;;)
1250      /* Conditional group: compilation checked that there are no more than      /* Conditional group: compilation checked that there are no more than
1251      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
1252      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
1253      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. */  
1254    
1255      case OP_COND:      case OP_COND:
1256      case OP_SCOND:      case OP_SCOND:
# Line 1065  for (;;) Line 1261  for (;;)
1261    
1262      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (ecode[LINK_SIZE+1] == OP_CALLOUT)
1263        {        {
1264        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
1265          {          {
1266          pcre_callout_block cb;          PUBL(callout_block) cb;
1267          cb.version          = 1;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1268          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1269          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1270    #ifdef COMPILE_PCRE8
1271          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1272    #else
1273            cb.subject          = (PCRE_SPTR16)md->start_subject;
1274    #endif
1275          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1276          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1277          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
# Line 1080  for (;;) Line 1280  for (;;)
1280          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1281          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last;
1282          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1283          if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);          cb.mark             = md->nomatch_mark;
1284            if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1285          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1286          }          }
1287        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1288        }        }
1289    
1290      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1100  for (;;) Line 1301  for (;;)
1301        else        else
1302          {          {
1303          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1304          condition =  (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1305    
1306          /* 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
1307          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
1308          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
1309          if any one is set. */          if any one is set. */
1310    
1311          if (!condition && condcode == OP_NRREF && recno != RREF_ANY)          if (!condition && condcode == OP_NRREF)
1312            {            {
1313            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1314            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1315              {              {
1316              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 1122  for (;;) Line 1323  for (;;)
1323    
1324            if (i < md->name_count)            if (i < md->name_count)
1325              {              {
1326              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1327              while (slotB > md->name_table)              while (slotB > md->name_table)
1328                {                {
1329                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1330                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1331                  {                  {
1332                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1333                  if (condition) break;                  if (condition) break;
# Line 1142  for (;;) Line 1343  for (;;)
1343                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1344                  {                  {
1345                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1346                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1347                    {                    {
1348                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1349                    if (condition) break;                    if (condition) break;
# Line 1155  for (;;) Line 1356  for (;;)
1356    
1357          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1358    
1359          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1360          }          }
1361        }        }
1362    
# Line 1172  for (;;) Line 1373  for (;;)
1373        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1374          {          {
1375          int refno = offset >> 1;          int refno = offset >> 1;
1376          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1377    
1378          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1379            {            {
# Line 1186  for (;;) Line 1387  for (;;)
1387    
1388          if (i < md->name_count)          if (i < md->name_count)
1389            {            {
1390            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1391            while (slotB > md->name_table)            while (slotB > md->name_table)
1392              {              {
1393              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1394              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1395                {                {
1396                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1397                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1208  for (;;) Line 1409  for (;;)
1409              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1410                {                {
1411                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1412                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1413                  {                  {
1414                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1415                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1223  for (;;) Line 1424  for (;;)
1424    
1425        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1426    
1427        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1428        }        }
1429    
1430      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1238  for (;;) Line 1439  for (;;)
1439    
1440      else      else
1441        {        {
1442        md->match_function_type = MATCH_CONDASSERT;        md->match_function_type = MATCH_CONDASSERT;
1443        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);
1444        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
1445          {          {
1446            if (md->end_offset_top > offset_top)
1447              offset_top = md->end_offset_top;  /* Captures may have happened */
1448          condition = TRUE;          condition = TRUE;
1449          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
1450          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
1451          }          }
1452        else if (rrc != MATCH_NOMATCH &&  
1453                (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an
1454          assertion; it is therefore treated as NOMATCH. */
1455    
1456          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1457          {          {
1458          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
1459          }          }
# Line 1258  for (;;) Line 1464  for (;;)
1464          }          }
1465        }        }
1466    
1467      /* 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
1468      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
1469      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
1470      alternative doesn't exist, we can just plough on. */      call to match() is always required, unless the second alternative doesn't
1471        exist, in which case we can just plough on. Note that, for compatibility
1472        with Perl, the | in a conditional group is NOT treated as creating two
1473        alternatives. If a THEN is encountered in the branch, it propagates out to
1474        the enclosing alternative (unless nested in a deeper set of alternatives,
1475        of course). */
1476    
1477      if (condition || *ecode == OP_ALT)      if (condition || *ecode == OP_ALT)
1478        {        {
1479        ecode += 1 + LINK_SIZE;        if (op != OP_SCOND)
       if (op == OP_SCOND)        /* Possibly empty group */  
1480          {          {
1481          md->match_function_type = MATCH_CBEGROUP;          ecode += 1 + LINK_SIZE;
1482          RMATCH(eptr, ecode, offset_top, md, eptrb, RM49);          goto TAIL_RECURSE;
         RRETURN(rrc);  
1483          }          }
1484        else goto TAIL_RECURSE;  
1485          md->match_function_type = MATCH_CBEGROUP;
1486          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);
1487          RRETURN(rrc);
1488        }        }
1489      else                         /* Condition false & no alternative */  
1490         /* Condition false & no alternative; continue after the group. */
1491    
1492        else
1493        {        {
1494        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1495        }        }
# Line 1301  for (;;) Line 1516  for (;;)
1516        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1517        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1518        }        }
1519      ecode += 3;      ecode += 1 + IMM2_SIZE;
1520      break;      break;
1521    
1522    
1523      /* 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. */  
1524    
     case OP_ACCEPT:  
1525      case OP_END:      case OP_END:
1526      if (md->recursive != NULL)      case OP_ACCEPT:
1527        {      case OP_ASSERT_ACCEPT:
1528        recursion_info *rec = md->recursive;  
1529        md->recursive = rec->prevrec;      /* If we have matched an empty string, fail if not in an assertion and not
1530        memmove(md->offset_vector, rec->offset_save,      in a recursion if either PCRE_NOTEMPTY is set, or if PCRE_NOTEMPTY_ATSTART
1531          rec->saved_max * sizeof(int));      is set and we have matched at the start of the subject. In both cases,
1532        offset_top = rec->save_offset_top;      backtracking will then try other alternatives, if any. */
1533        if (rec->group_num == 0)  
1534          {      if (eptr == mstart && op != OP_ASSERT_ACCEPT &&
1535          ecode = rec->after_call;           md->recursive == NULL &&
1536          break;           (md->notempty ||
1537          }             (md->notempty_atstart &&
1538        }               mstart == md->start_subject + md->start_offset)))
1539          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);  
1540    
1541      /* Otherwise, we have a match. */      /* Otherwise, we have a match. */
1542    
1543      md->end_match_ptr = eptr;           /* Record where we ended */      md->end_match_ptr = eptr;           /* Record where we ended */
1544      md->end_offset_top = offset_top;    /* and how many extracts were taken */      md->end_offset_top = offset_top;    /* and how many extracts were taken */
1545      md->start_match_ptr = mstart;       /* and the start (\K can modify) */      md->start_match_ptr = mstart;       /* and the start (\K can modify) */
1546    
1547      /* 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
1548      given as the argument to MRRETURN when the heap is in use. */      given as the argument to RRETURN when the heap is in use. */
1549    
1550      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;
1551      MRRETURN(rrc);      RRETURN(rrc);
1552    
1553      /* Assertion brackets. Check the alternative branches in turn - the      /* Assertion brackets. Check the alternative branches in turn - the
1554      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,
1555      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
1556      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
1557      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
1558      of a condition, we want to return immediately afterwards. The caller of      of a condition, we want to return immediately afterwards. The caller of
1559      this incarnation of the match() function will have set MATCH_CONDASSERT in      this incarnation of the match() function will have set MATCH_CONDASSERT in
1560      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
1561      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
1562      match() to remember this case. */      match() to remember this case. */
1563    
1564      case OP_ASSERT:      case OP_ASSERT:
1565      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1566        save_mark = md->mark;
1567      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1568        {        {
1569        condassert = TRUE;        condassert = TRUE;
1570        md->match_function_type = 0;        md->match_function_type = 0;
1571        }        }
1572      else condassert = FALSE;      else condassert = FALSE;
1573    
1574      do      do
1575        {        {
1576        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 1579  for (;;)
1579          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1580          break;          break;
1581          }          }
1582        if (rrc != MATCH_NOMATCH &&  
1583            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE does not allow THEN to escape beyond an assertion; it is treated
1584          RRETURN(rrc);        as NOMATCH. */
1585    
1586          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1587        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1588          md->mark = save_mark;
1589        }        }
1590      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1591    
1592      if (*ecode == OP_KET) MRRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1593    
1594      /* If checking an assertion for a condition, return MATCH_MATCH. */      /* If checking an assertion for a condition, return MATCH_MATCH. */
1595    
# Line 1403  for (;;) Line 1609  for (;;)
1609    
1610      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1611      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1612        save_mark = md->mark;
1613      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1614        {        {
1615        condassert = TRUE;        condassert = TRUE;
1616        md->match_function_type = 0;        md->match_function_type = 0;
1617        }        }
1618      else condassert = FALSE;      else condassert = FALSE;
1619    
1620      do      do
1621        {        {
1622        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1623        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) MRRETURN(MATCH_NOMATCH);        md->mark = save_mark;
1624          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1625        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1626          {          {
1627          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1628          break;          break;
1629          }          }
1630        if (rrc != MATCH_NOMATCH &&  
1631            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE does not allow THEN to escape beyond an assertion; it is treated
1632          RRETURN(rrc);        as NOMATCH. */
1633    
1634          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1635        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1636        }        }
1637      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1638    
1639      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1640    
1641      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1642      continue;      continue;
1643    
# Line 1437  for (;;) Line 1647  for (;;)
1647      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1648    
1649      case OP_REVERSE:      case OP_REVERSE:
1650  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1651      if (utf8)      if (utf)
1652        {        {
1653        i = GET(ecode, 1);        i = GET(ecode, 1);
1654        while (i-- > 0)        while (i-- > 0)
1655          {          {
1656          eptr--;          eptr--;
1657          if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1658          BACKCHAR(eptr);          BACKCHAR(eptr);
1659          }          }
1660        }        }
# Line 1455  for (;;) Line 1665  for (;;)
1665    
1666        {        {
1667        eptr -= GET(ecode, 1);        eptr -= GET(ecode, 1);
1668        if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1669        }        }
1670    
1671      /* 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 1679  for (;;)
1679      function is able to force a failure. */      function is able to force a failure. */
1680    
1681      case OP_CALLOUT:      case OP_CALLOUT:
1682      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1683        {        {
1684        pcre_callout_block cb;        PUBL(callout_block) cb;
1685        cb.version          = 1;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1686        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1687        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1688    #ifdef COMPILE_PCRE8
1689        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1690    #else
1691          cb.subject          = (PCRE_SPTR16)md->start_subject;
1692    #endif
1693        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1694        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1695        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
# Line 1484  for (;;) Line 1698  for (;;)
1698        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1699        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last;
1700        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1701        if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);        cb.mark             = md->nomatch_mark;
1702          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1703        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1704        }        }
1705      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1494  for (;;) Line 1709  for (;;)
1709      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
1710      whole pattern. (This is so that it works from duplicated subpatterns.)      whole pattern. (This is so that it works from duplicated subpatterns.)
1711    
1712      If there are any capturing brackets started but not finished, we have to      The state of the capturing groups is preserved over recursion, and
1713      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
1714      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
1715      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
1716      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
1717      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
1718      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.  
1719    
1720      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
1721      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
1722      for the original version of this logic. */      for the original version of this logic. It has, however, been hacked around
1723        a lot, so he is not to blame for the current way it works. */
1724    
1725      case OP_RECURSE:      case OP_RECURSE:
1726        {        {
1727          recursion_info *ri;
1728          int recno;
1729    
1730        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1731        new_recursive.group_num = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
1732          GET2(callpat, 1 + LINK_SIZE);          GET2(callpat, 1 + LINK_SIZE);
1733    
1734          /* Check for repeating a recursion without advancing the subject pointer.
1735          This should catch convoluted mutual recursions. (Some simple cases are
1736          caught at compile time.) */
1737    
1738          for (ri = md->recursive; ri != NULL; ri = ri->prevrec)
1739            if (recno == ri->group_num && eptr == ri->subject_position)
1740              RRETURN(PCRE_ERROR_RECURSELOOP);
1741    
1742        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1743    
1744          new_recursive.group_num = recno;
1745          new_recursive.subject_position = eptr;
1746        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1747        md->recursive = &new_recursive;        md->recursive = &new_recursive;
1748    
1749        /* Find where to continue from afterwards */        /* Where to continue from afterwards */
1750    
1751        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
       new_recursive.after_call = ecode;  
1752    
1753        /* Now save the offset data. */        /* Now save the offset data */
1754    
1755        new_recursive.saved_max = md->offset_end;        new_recursive.saved_max = md->offset_end;
1756        if (new_recursive.saved_max <= REC_STACK_SAVE_MAX)        if (new_recursive.saved_max <= REC_STACK_SAVE_MAX)
# Line 1533  for (;;) Line 1758  for (;;)
1758        else        else
1759          {          {
1760          new_recursive.offset_save =          new_recursive.offset_save =
1761            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1762          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1763          }          }
   
1764        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1765              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1766        new_recursive.save_offset_top = offset_top;  
1767          /* OK, now we can do the recursion. After processing each alternative,
1768        /* OK, now we can do the recursion. For each top-level alternative we        restore the offset data. If there were nested recursions, md->recursive
1769        restore the offset and recursion data. */        might be changed, so reset it before looping. */
1770    
1771        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1772        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
1773        do        do
1774          {          {
1775          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1776          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1777            md, eptrb, RM6);            md, eptrb, RM6);
1778            memcpy(md->offset_vector, new_recursive.offset_save,
1779                new_recursive.saved_max * sizeof(int));
1780            md->recursive = new_recursive.prevrec;
1781          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1782            {            {
1783            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
           md->recursive = new_recursive.prevrec;  
1784            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1785              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1786            MRRETURN(MATCH_MATCH);  
1787              /* Set where we got to in the subject, and reset the start in case
1788              it was changed by \K. This *is* propagated back out of a recursion,
1789              for Perl compatibility. */
1790    
1791              eptr = md->end_match_ptr;
1792              mstart = md->start_match_ptr;
1793              goto RECURSION_MATCHED;        /* Exit loop; end processing */
1794            }            }
1795          else if (rrc != MATCH_NOMATCH &&  
1796                  (rrc != MATCH_THEN || md->start_match_ptr != ecode))          /* PCRE does not allow THEN to escape beyond a recursion; it is treated
1797            as NOMATCH. */
1798    
1799            else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1800            {            {
1801            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1802            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1803              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1804            RRETURN(rrc);            RRETURN(rrc);
1805            }            }
1806    
1807          md->recursive = &new_recursive;          md->recursive = &new_recursive;
         memcpy(md->offset_vector, new_recursive.offset_save,  
             new_recursive.saved_max * sizeof(int));  
1808          callpat += GET(callpat, 1);          callpat += GET(callpat, 1);
1809          }          }
1810        while (*callpat == OP_ALT);        while (*callpat == OP_ALT);
# Line 1578  for (;;) Line 1812  for (;;)
1812        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1813        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1814        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1815          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1816        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;  
1817        }        }
1818    
1819      /* The repeating kets try the rest of the pattern or restart from the      RECURSION_MATCHED:
1820      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 */  
1821    
1822      /* 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
1823      bracketed group and go to there. */      bracketed group and go to there. */
# Line 1667  for (;;) Line 1831  for (;;)
1831      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
1832      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
1833      optional ones preceded by BRAZERO or BRAMINZERO. */      optional ones preceded by BRAZERO or BRAMINZERO. */
1834    
1835      case OP_BRAZERO:      case OP_BRAZERO:
1836      next = ecode + 1;      next = ecode + 1;
1837      RMATCH(eptr, next, offset_top, md, eptrb, RM10);      RMATCH(eptr, next, offset_top, md, eptrb, RM10);
# Line 1675  for (;;) Line 1839  for (;;)
1839      do next += GET(next, 1); while (*next == OP_ALT);      do next += GET(next, 1); while (*next == OP_ALT);
1840      ecode = next + 1 + LINK_SIZE;      ecode = next + 1 + LINK_SIZE;
1841      break;      break;
1842    
1843      case OP_BRAMINZERO:      case OP_BRAMINZERO:
1844      next = ecode + 1;      next = ecode + 1;
1845      do next += GET(next, 1); while (*next == OP_ALT);      do next += GET(next, 1); while (*next == OP_ALT);
# Line 1689  for (;;) Line 1853  for (;;)
1853      do next += GET(next,1); while (*next == OP_ALT);      do next += GET(next,1); while (*next == OP_ALT);
1854      ecode = next + 1 + LINK_SIZE;      ecode = next + 1 + LINK_SIZE;
1855      break;      break;
1856    
1857      /* BRAPOSZERO occurs before a possessive bracket group. Don't do anything      /* BRAPOSZERO occurs before a possessive bracket group. Don't do anything
1858      here; just jump to the group, with allow_zero set TRUE. */      here; just jump to the group, with allow_zero set TRUE. */
1859    
1860      case OP_BRAPOSZERO:      case OP_BRAPOSZERO:
1861      op = *(++ecode);      op = *(++ecode);
1862      allow_zero = TRUE;      allow_zero = TRUE;
1863      if (op == OP_CBRAPOS || op == OP_SCBRAPOS) goto POSSESSIVE_CAPTURE;      if (op == OP_CBRAPOS || op == OP_SCBRAPOS) goto POSSESSIVE_CAPTURE;
1864        goto POSSESSIVE_NON_CAPTURE;        goto POSSESSIVE_NON_CAPTURE;
# Line 1704  for (;;) Line 1868  for (;;)
1868      case OP_KET:      case OP_KET:
1869      case OP_KETRMIN:      case OP_KETRMIN:
1870      case OP_KETRMAX:      case OP_KETRMAX:
1871      case OP_KETRPOS:      case OP_KETRPOS:
1872      prev = ecode - GET(ecode, 1);      prev = ecode - GET(ecode, 1);
1873    
1874      /* 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
1875      infinite repeats of empty string matches, retrieve the subject start from      infinite repeats of empty string matches, retrieve the subject start from
1876      the chain. Otherwise, set it NULL. */      the chain. Otherwise, set it NULL. */
1877    
1878      if (*prev >= OP_SBRA)      if (*prev >= OP_SBRA || *prev == OP_ONCE)
1879        {        {
1880        saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */        saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */
1881        eptrb = eptrb->epb_prev;              /* Backup to previous group */        eptrb = eptrb->epb_prev;              /* Backup to previous group */
1882        }        }
1883      else saved_eptr = NULL;      else saved_eptr = NULL;
1884    
1885      /* 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
1886      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
1887      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
1888      it was changed by \K. */      start in case it was changed by \K. */
1889    
1890      if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||      if ((*prev >= OP_ASSERT && *prev <= OP_ASSERTBACK_NOT) ||
1891          *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||           *prev == OP_ONCE_NC)
         *prev == OP_ONCE)  
1892        {        {
1893        md->end_match_ptr = eptr;      /* For ONCE */        md->end_match_ptr = eptr;      /* For ONCE_NC */
1894        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
1895        md->start_match_ptr = mstart;        md->start_match_ptr = mstart;
1896        MRRETURN(MATCH_MATCH);        RRETURN(MATCH_MATCH);         /* Sets md->mark */
1897        }        }
1898    
1899      /* 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
1900      and if necessary complete handling an extraction by setting the offsets and      and if necessary complete handling an extraction by setting the offsets and
1901      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
1902      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
1903      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
1904        the current subject position and start match pointer and give a MATCH
1905        return. */
1906    
1907      if (*prev == OP_CBRA || *prev == OP_SCBRA ||      if (*prev == OP_CBRA || *prev == OP_SCBRA ||
1908          *prev == OP_CBRAPOS || *prev == OP_SCBRAPOS)          *prev == OP_CBRAPOS || *prev == OP_SCBRAPOS)
# Line 1750  for (;;) Line 1915  for (;;)
1915        printf("\n");        printf("\n");
1916  #endif  #endif
1917    
1918          /* Handle a recursively called group. */
1919    
1920          if (md->recursive != NULL && md->recursive->group_num == number)
1921            {
1922            md->end_match_ptr = eptr;
1923            md->start_match_ptr = mstart;
1924            RRETURN(MATCH_MATCH);
1925            }
1926    
1927          /* Deal with capturing */
1928    
1929        md->capture_last = number;        md->capture_last = number;
1930        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->offset_overflow = TRUE; else
1931          {          {
1932            /* If offset is greater than offset_top, it means that we are
1933            "skipping" a capturing group, and that group's offsets must be marked
1934            unset. In earlier versions of PCRE, all the offsets were unset at the
1935            start of matching, but this doesn't work because atomic groups and
1936            assertions can cause a value to be set that should later be unset.
1937            Example: matching /(?>(a))b|(a)c/ against "ac". This sets group 1 as
1938            part of the atomic group, but this is not on the final matching path,
1939            so must be unset when 2 is set. (If there is no group 2, there is no
1940            problem, because offset_top will then be 2, indicating no capture.) */
1941    
1942            if (offset > offset_top)
1943              {
1944              register int *iptr = md->offset_vector + offset_top;
1945              register int *iend = md->offset_vector + offset;
1946              while (iptr < iend) *iptr++ = -1;
1947              }
1948    
1949            /* Now make the extraction */
1950    
1951          md->offset_vector[offset] =          md->offset_vector[offset] =
1952            md->offset_vector[md->offset_end - number];            md->offset_vector[md->offset_end - number];
1953          md->offset_vector[offset+1] = (int)(eptr - md->start_subject);          md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1954          if (offset_top <= offset) offset_top = offset + 2;          if (offset_top <= offset) offset_top = offset + 2;
1955          }          }
   
       /* 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;  
         }  
1956        }        }
1957    
1958      /* For a non-repeating ket, just continue at this level. This also      /* For an ordinary non-repeating ket, just continue at this level. This
1959      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
1960      This is the forcible breaking of infinite loops as implemented in Perl      group. This is the forcible breaking of infinite loops as implemented in
1961      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,
1962      course of events. */      establish a backup point by processing the rest of the pattern at a lower
1963        level. If this results in a NOMATCH return, pass MATCH_ONCE back to the
1964        original OP_ONCE level, thereby bypassing intermediate backup points, but
1965        resetting any captures that happened along the way. */
1966    
1967      if (*ecode == OP_KET || eptr == saved_eptr)      if (*ecode == OP_KET || eptr == saved_eptr)
1968        {        {
1969        ecode += 1 + LINK_SIZE;        if (*prev == OP_ONCE)
1970            {
1971            RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM12);
1972            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1973            md->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */
1974            RRETURN(MATCH_ONCE);
1975            }
1976          ecode += 1 + LINK_SIZE;    /* Carry on at this level */
1977        break;        break;
1978        }        }
1979    
1980      /* OP_KETRPOS is a possessive repeating ket. Remember the current position,      /* OP_KETRPOS is a possessive repeating ket. Remember the current position,
1981      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
1982      at a time from the outer level, thus saving stack. */      at a time from the outer level, thus saving stack. */
1983    
1984      if (*ecode == OP_KETRPOS)      if (*ecode == OP_KETRPOS)
1985        {        {
1986        md->end_match_ptr = eptr;        md->end_match_ptr = eptr;
1987        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
1988        RRETURN(MATCH_KETRPOS);        RRETURN(MATCH_KETRPOS);
1989        }        }
1990    
1991      /* 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
1992      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
1993      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
1994      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
1995        string. */
1996    
1997      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1998        {        {
1999        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM12);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM7);
2000        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2001          if (*prev == OP_ONCE)
2002            {
2003            RMATCH(eptr, prev, offset_top, md, eptrb, RM8);
2004            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2005            md->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */
2006            RRETURN(MATCH_ONCE);
2007            }
2008        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2009          {          {
2010          md->match_function_type = MATCH_CBEGROUP;          md->match_function_type = MATCH_CBEGROUP;
2011          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2012          RRETURN(rrc);          RRETURN(rrc);
2013          }          }
# Line 1818  for (;;) Line 2016  for (;;)
2016        }        }
2017      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2018        {        {
2019        if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
2020        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2021          if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2022        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2023          if (*prev == OP_ONCE)
2024            {
2025            RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM9);
2026            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2027            md->once_target = prev;
2028            RRETURN(MATCH_ONCE);
2029            }
2030        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
2031        goto TAIL_RECURSE;        goto TAIL_RECURSE;
2032        }        }
# Line 1829  for (;;) Line 2035  for (;;)
2035      /* Not multiline mode: start of subject assertion, unless notbol. */      /* Not multiline mode: start of subject assertion, unless notbol. */
2036    
2037      case OP_CIRC:      case OP_CIRC:
2038      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
2039    
2040      /* Start of subject assertion */      /* Start of subject assertion */
2041    
2042      case OP_SOD:      case OP_SOD:
2043      if (eptr != md->start_subject) MRRETURN(MATCH_NOMATCH);      if (eptr != md->start_subject) RRETURN(MATCH_NOMATCH);
2044      ecode++;      ecode++;
2045      break;      break;
2046    
2047      /* Multiline mode: start of subject unless notbol, or after any newline. */      /* Multiline mode: start of subject unless notbol, or after any newline. */
2048    
2049      case OP_CIRCM:      case OP_CIRCM:
2050      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
2051      if (eptr != md->start_subject &&      if (eptr != md->start_subject &&
2052          (eptr == md->end_subject || !WAS_NEWLINE(eptr)))          (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
2053        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2054      ecode++;      ecode++;
2055      break;      break;
2056    
2057      /* Start of match assertion */      /* Start of match assertion */
2058    
2059      case OP_SOM:      case OP_SOM:
2060      if (eptr != md->start_subject + md->start_offset) MRRETURN(MATCH_NOMATCH);      if (eptr != md->start_subject + md->start_offset) RRETURN(MATCH_NOMATCH);
2061      ecode++;      ecode++;
2062      break;      break;
2063    
# Line 1867  for (;;) Line 2073  for (;;)
2073    
2074      case OP_DOLLM:      case OP_DOLLM:
2075      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2076        { if (!IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH); }        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }
2077      else      else
2078        {        {
2079        if (md->noteol) MRRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
2080        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2081        }        }
2082      ecode++;      ecode++;
2083      break;      break;
2084    
2085      /* Not multiline mode: assert before a terminating newline or before end of      /* Not multiline mode: assert before a terminating newline or before end of
2086      subject unless noteol is set. */      subject unless noteol is set. */
2087    
2088      case OP_DOLL:      case OP_DOLL:
2089      if (md->noteol) MRRETURN(MATCH_NOMATCH);      if (md->noteol) RRETURN(MATCH_NOMATCH);
2090      if (!md->endonly) goto ASSERT_NL_OR_EOS;      if (!md->endonly) goto ASSERT_NL_OR_EOS;
2091    
2092      /* ... else fall through for endonly */      /* ... else fall through for endonly */
# Line 1888  for (;;) Line 2094  for (;;)
2094      /* End of subject assertion (\z) */      /* End of subject assertion (\z) */
2095    
2096      case OP_EOD:      case OP_EOD:
2097      if (eptr < md->end_subject) MRRETURN(MATCH_NOMATCH);      if (eptr < md->end_subject) RRETURN(MATCH_NOMATCH);
2098      SCHECK_PARTIAL();      SCHECK_PARTIAL();
2099      ecode++;      ecode++;
2100      break;      break;
# Line 1899  for (;;) Line 2105  for (;;)
2105      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2106      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2107          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2108        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2109    
2110      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2111    
# Line 1918  for (;;) Line 2124  for (;;)
2124        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2125        partial matching. */        partial matching. */
2126    
2127  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2128        if (utf8)        if (utf)
2129          {          {
2130          /* Get status of previous character */          /* Get status of previous character */
2131    
2132          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2133            {            {
2134            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2135            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2136            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2137            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2138  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 1991  for (;;) Line 2197  for (;;)
2197              }              }
2198            else            else
2199  #endif  #endif
2200            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2201                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2202            }            }
2203    
2204          /* Get status of next character */          /* Get status of next character */
# Line 2014  for (;;) Line 2221  for (;;)
2221            }            }
2222          else          else
2223  #endif  #endif
2224          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2225              && ((md->ctypes[*eptr] & ctype_word) != 0);
2226          }          }
2227    
2228        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
2229    
2230        if ((*ecode++ == OP_WORD_BOUNDARY)?        if ((*ecode++ == OP_WORD_BOUNDARY)?
2231             cur_is_word == prev_is_word : cur_is_word != prev_is_word)             cur_is_word == prev_is_word : cur_is_word != prev_is_word)
2232          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2233        }        }
2234      break;      break;
2235    
2236      /* Match a single character type; inline for speed */      /* Match a single character type; inline for speed */
2237    
2238      case OP_ANY:      case OP_ANY:
2239      if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2240      /* Fall through */      /* Fall through */
2241    
2242      case OP_ALLANY:      case OP_ALLANY:
2243      if (eptr++ >= md->end_subject)      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2244        {        {                            /* not be updated before SCHECK_PARTIAL. */
2245        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2246        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2247        }        }
2248      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;      eptr++;
2249    #ifdef SUPPORT_UTF
2250        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2251    #endif
2252      ecode++;      ecode++;
2253      break;      break;
2254    
# Line 2045  for (;;) Line 2256  for (;;)
2256      any byte, even newline, independent of the setting of PCRE_DOTALL. */      any byte, even newline, independent of the setting of PCRE_DOTALL. */
2257    
2258      case OP_ANYBYTE:      case OP_ANYBYTE:
2259      if (eptr++ >= md->end_subject)      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2260        {        {                            /* not be updated before SCHECK_PARTIAL. */
2261        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2262        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2263        }        }
2264        eptr++;
2265      ecode++;      ecode++;
2266      break;      break;
2267    
# Line 2057  for (;;) Line 2269  for (;;)
2269      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2270        {        {
2271        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2272        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2273        }        }
2274      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2275      if (      if (
2276  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2277         c < 256 &&         c < 256 &&
2278  #endif  #endif
2279         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
2280         )         )
2281        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2282      ecode++;      ecode++;
2283      break;      break;
2284    
# Line 2074  for (;;) Line 2286  for (;;)
2286      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2287        {        {
2288        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2289        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2290        }        }
2291      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2292      if (      if (
2293  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2294         c >= 256 ||         c > 255 ||
2295  #endif  #endif
2296         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2297         )         )
2298        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2299      ecode++;      ecode++;
2300      break;      break;
2301    
# Line 2091  for (;;) Line 2303  for (;;)
2303      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2304        {        {
2305        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2306        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2307        }        }
2308      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2309      if (      if (
2310  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2311         c < 256 &&         c < 256 &&
2312  #endif  #endif
2313         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
2314         )         )
2315        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2316      ecode++;      ecode++;
2317      break;      break;
2318    
# Line 2108  for (;;) Line 2320  for (;;)
2320      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2321        {        {
2322        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2323        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2324        }        }
2325      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2326      if (      if (
2327  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2328         c >= 256 ||         c > 255 ||
2329  #endif  #endif
2330         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2331         )         )
2332        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2333      ecode++;      ecode++;
2334      break;      break;
2335    
# Line 2125  for (;;) Line 2337  for (;;)
2337      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2338        {        {
2339        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2340        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2341        }        }
2342      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2343      if (      if (
2344  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2345         c < 256 &&         c < 256 &&
2346  #endif  #endif
2347         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
2348         )         )
2349        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2350      ecode++;      ecode++;
2351      break;      break;
2352    
# Line 2142  for (;;) Line 2354  for (;;)
2354      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2355        {        {
2356        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2357        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2358        }        }
2359      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2360      if (      if (
2361  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2362         c >= 256 ||         c > 255 ||
2363  #endif  #endif
2364         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2365         )         )
2366        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2367      ecode++;      ecode++;
2368      break;      break;
2369    
# Line 2159  for (;;) Line 2371  for (;;)
2371      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2372        {        {
2373        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2374        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2375        }        }
2376      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2377      switch(c)      switch(c)
2378        {        {
2379        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2380    
2381        case 0x000d:        case 0x000d:
2382        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
2383        break;        break;
# Line 2178  for (;;) Line 2390  for (;;)
2390        case 0x0085:        case 0x0085:
2391        case 0x2028:        case 0x2028:
2392        case 0x2029:        case 0x2029:
2393        if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2394        break;        break;
2395        }        }
2396      ecode++;      ecode++;
# Line 2188  for (;;) Line 2400  for (;;)
2400      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2401        {        {
2402        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2403        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2404        }        }
2405      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2406      switch(c)      switch(c)
# Line 2213  for (;;) Line 2425  for (;;)
2425        case 0x202f:    /* NARROW NO-BREAK SPACE */        case 0x202f:    /* NARROW NO-BREAK SPACE */
2426        case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */        case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
2427        case 0x3000:    /* IDEOGRAPHIC SPACE */        case 0x3000:    /* IDEOGRAPHIC SPACE */
2428        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2429        }        }
2430      ecode++;      ecode++;
2431      break;      break;
# Line 2222  for (;;) Line 2434  for (;;)
2434      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2435        {        {
2436        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2437        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2438        }        }
2439      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2440      switch(c)      switch(c)
2441        {        {
2442        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2443        case 0x09:      /* HT */        case 0x09:      /* HT */
2444        case 0x20:      /* SPACE */        case 0x20:      /* SPACE */
2445        case 0xa0:      /* NBSP */        case 0xa0:      /* NBSP */
# Line 2256  for (;;) Line 2468  for (;;)
2468      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2469        {        {
2470        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2471        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2472        }        }
2473      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2474      switch(c)      switch(c)
# Line 2269  for (;;) Line 2481  for (;;)
2481        case 0x85:      /* NEL */        case 0x85:      /* NEL */
2482        case 0x2028:    /* LINE SEPARATOR */        case 0x2028:    /* LINE SEPARATOR */
2483        case 0x2029:    /* PARAGRAPH SEPARATOR */        case 0x2029:    /* PARAGRAPH SEPARATOR */
2484        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2485        }        }
2486      ecode++;      ecode++;
2487      break;      break;
# Line 2278  for (;;) Line 2490  for (;;)
2490      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2491        {        {
2492        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2493        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2494        }        }
2495      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2496      switch(c)      switch(c)
2497        {        {
2498        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2499        case 0x0a:      /* LF */        case 0x0a:      /* LF */
2500        case 0x0b:      /* VT */        case 0x0b:      /* VT */
2501        case 0x0c:      /* FF */        case 0x0c:      /* FF */
# Line 2305  for (;;) Line 2517  for (;;)
2517      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2518        {        {
2519        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2520        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2521        }        }
2522      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2523        {        {
# Line 2314  for (;;) Line 2526  for (;;)
2526        switch(ecode[1])        switch(ecode[1])
2527          {          {
2528          case PT_ANY:          case PT_ANY:
2529          if (op == OP_NOTPROP) MRRETURN(MATCH_NOMATCH);          if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
2530          break;          break;
2531    
2532          case PT_LAMP:          case PT_LAMP:
2533          if ((prop->chartype == ucp_Lu ||          if ((prop->chartype == ucp_Lu ||
2534               prop->chartype == ucp_Ll ||               prop->chartype == ucp_Ll ||
2535               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))
2536            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2537          break;          break;
2538    
2539          case PT_GC:          case PT_GC:
2540          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2541            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2542          break;          break;
2543    
2544          case PT_PC:          case PT_PC:
2545          if ((ecode[2] != prop->chartype) == (op == OP_PROP))          if ((ecode[2] != prop->chartype) == (op == OP_PROP))
2546            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2547          break;          break;
2548    
2549          case PT_SC:          case PT_SC:
2550          if ((ecode[2] != prop->script) == (op == OP_PROP))          if ((ecode[2] != prop->script) == (op == OP_PROP))
2551            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2552          break;          break;
2553    
2554          /* These are specials */          /* These are specials */
2555    
2556          case PT_ALNUM:          case PT_ALNUM:
2557          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2558               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2559            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2560          break;          break;
2561    
2562          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2563          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2564               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2565                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2566            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2567          break;          break;
2568    
2569          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2570          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2571               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2572               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2573                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2574            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2575          break;          break;
2576    
2577          case PT_WORD:          case PT_WORD:
2578          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2579               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2580               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2581            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2582          break;          break;
2583    
2584          /* This should never occur */          /* This should never occur */
# Line 2386  for (;;) Line 2598  for (;;)
2598      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2599        {        {
2600        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2601        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2602        }        }
2603      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2604        if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);
2605        while (eptr < md->end_subject)
2606        {        {
2607        int category = UCD_CATEGORY(c);        int len = 1;
2608        if (category == ucp_M) MRRETURN(MATCH_NOMATCH);        if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2609        while (eptr < md->end_subject)        if (UCD_CATEGORY(c) != ucp_M) break;
2610          {        eptr += len;
2611          int len = 1;        }
2612          if (!utf8) c = *eptr; else      if (md->partial != 0 && eptr >= md->end_subject)
2613            {        {
2614            GETCHARLEN(c, eptr, len);        SCHECK_PARTIAL();
           }  
         category = UCD_CATEGORY(c);  
         if (category != ucp_M) break;  
         eptr += len;  
         }  
2615        }        }
2616      ecode++;      ecode++;
2617      break;      break;
# Line 2418  for (;;) Line 2627  for (;;)
2627      loops). */      loops). */
2628    
2629      case OP_REF:      case OP_REF:
2630      case OP_REFI:      case OP_REFI:
2631      caseless = op == OP_REFI;      caseless = op == OP_REFI;
2632      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2633      ecode += 3;      ecode += 1 + IMM2_SIZE;
2634    
2635      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2636    
# Line 2461  for (;;) Line 2670  for (;;)
2670        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2671        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2672        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2673        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2674        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2675        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2676        break;        break;
2677    
2678        default:               /* No repeat follows */        default:               /* No repeat follows */
2679        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2680          {          {
2681            eptr += -(length + 1);
2682          CHECK_PARTIAL();          CHECK_PARTIAL();
2683          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2684          }          }
2685        eptr += length;        eptr += length;
2686        continue;              /* With the main loop */        continue;              /* With the main loop */
2687        }        }
2688    
2689      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2690      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2691        means the reference is unset in non-Java-compatible mode. If the minimum is
2692        zero, we can continue at the same level without recursion. For any other
2693        minimum, carrying on will result in NOMATCH. */
2694    
2695      if (length == 0) continue;      if (length == 0) continue;
2696        if (length < 0 && min == 0) continue;
2697    
2698      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2699      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 2701  for (;;)
2701    
2702      for (i = 1; i <= min; i++)      for (i = 1; i <= min; i++)
2703        {        {
2704        int slength;        int slength;
2705        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2706          {          {
2707            eptr += -(slength + 1);
2708          CHECK_PARTIAL();          CHECK_PARTIAL();
2709          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2710          }          }
2711        eptr += slength;        eptr += slength;
2712        }        }
# Line 2507  for (;;) Line 2722  for (;;)
2722        {        {
2723        for (fi = min;; fi++)        for (fi = min;; fi++)
2724          {          {
2725          int slength;          int slength;
2726          RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);
2727          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2728          if (fi >= max) MRRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2729          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2730            {            {
2731              eptr += -(slength + 1);
2732            CHECK_PARTIAL();            CHECK_PARTIAL();
2733            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2734            }            }
2735          eptr += slength;          eptr += slength;
2736          }          }
# Line 2528  for (;;) Line 2744  for (;;)
2744        pp = eptr;        pp = eptr;
2745        for (i = min; i < max; i++)        for (i = min; i < max; i++)
2746          {          {
2747          int slength;          int slength;
2748          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2749            {            {
2750              /* Restore the eptr after the check. */
2751              eptr += -(slength + 1);
2752            CHECK_PARTIAL();            CHECK_PARTIAL();
2753              eptr -= -(slength + 1);
2754            break;            break;
2755            }            }
2756          eptr += slength;          eptr += slength;
# Line 2542  for (;;) Line 2761  for (;;)
2761          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2762          eptr -= length;          eptr -= length;
2763          }          }
2764        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2765        }        }
2766      /* Control never gets here */      /* Control never gets here */
2767    
# Line 2560  for (;;) Line 2779  for (;;)
2779      case OP_NCLASS:      case OP_NCLASS:
2780      case OP_CLASS:      case OP_CLASS:
2781        {        {
2782          /* The data variable is saved across frames, so the byte map needs to
2783          be stored there. */
2784    #define BYTE_MAP ((pcre_uint8 *)data)
2785        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2786        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2787    
2788        switch (*ecode)        switch (*ecode)
2789          {          {
# Line 2582  for (;;) Line 2804  for (;;)
2804          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2805          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2806          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2807          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2808          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2809          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2810          break;          break;
2811    
2812          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2594  for (;;) Line 2816  for (;;)
2816    
2817        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2818    
2819  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2820        /* UTF-8 mode */        if (utf)
       if (utf8)  
2821          {          {
2822          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2823            {            {
2824            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2825              {              {
2826              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2827              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2828              }              }
2829            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
2830            if (c > 255)            if (c > 255)
2831              {              {
2832              if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2833              }              }
2834            else            else
2835              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
             }  
2836            }            }
2837          }          }
2838        else        else
2839  #endif  #endif
2840        /* Not UTF-8 mode */        /* Not UTF mode */
2841          {          {
2842          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2843            {            {
2844            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2845              {              {
2846              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2847              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2848              }              }
2849            c = *eptr++;            c = *eptr++;
2850            if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2851              if (c > 255)
2852                {
2853                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2854                }
2855              else
2856    #endif
2857                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2858            }            }
2859          }          }
2860    
# Line 2642  for (;;) Line 2868  for (;;)
2868    
2869        if (minimize)        if (minimize)
2870          {          {
2871  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2872          /* UTF-8 mode */          if (utf)
         if (utf8)  
2873            {            {
2874            for (fi = min;; fi++)            for (fi = min;; fi++)
2875              {              {
2876              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);
2877              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2878              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2879              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
2880                {                {
2881                SCHECK_PARTIAL();                SCHECK_PARTIAL();
2882                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2883                }                }
2884              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
2885              if (c > 255)              if (c > 255)
2886                {                {
2887                if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2888                }                }
2889              else              else
2890                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
               }  
2891              }              }
2892            }            }
2893          else          else
2894  #endif  #endif
2895          /* Not UTF-8 mode */          /* Not UTF mode */
2896            {            {
2897            for (fi = min;; fi++)            for (fi = min;; fi++)
2898              {              {
2899              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);
2900              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2901              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2902              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
2903                {                {
2904                SCHECK_PARTIAL();                SCHECK_PARTIAL();
2905                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2906                }                }
2907              c = *eptr++;              c = *eptr++;
2908              if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2909                if (c > 255)
2910                  {
2911                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2912                  }
2913                else
2914    #endif
2915                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2916              }              }
2917            }            }
2918          /* Control never gets here */          /* Control never gets here */
# Line 2694  for (;;) Line 2924  for (;;)
2924          {          {
2925          pp = eptr;          pp = eptr;
2926    
2927  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2928          /* UTF-8 mode */          if (utf)
         if (utf8)  
2929            {            {
2930            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2931              {              {
# Line 2712  for (;;) Line 2941  for (;;)
2941                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
2942                }                }
2943              else              else
2944                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
2945              eptr += len;              eptr += len;
2946              }              }
2947            for (;;)            for (;;)
# Line 2727  for (;;) Line 2954  for (;;)
2954            }            }
2955          else          else
2956  #endif  #endif
2957            /* Not UTF-8 mode */            /* Not UTF mode */
2958            {            {
2959            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2960              {              {
# Line 2737  for (;;) Line 2964  for (;;)
2964                break;                break;
2965                }                }
2966              c = *eptr;              c = *eptr;
2967              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
2968                if (c > 255)
2969                  {
2970                  if (op == OP_CLASS) break;
2971                  }
2972                else
2973    #endif
2974                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
2975              eptr++;              eptr++;
2976              }              }
2977            while (eptr >= pp)            while (eptr >= pp)
# Line 2748  for (;;) Line 2982  for (;;)
2982              }              }
2983            }            }
2984    
2985          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2986          }          }
2987    #undef BYTE_MAP
2988        }        }
2989      /* Control never gets here */      /* Control never gets here */
2990    
# Line 2758  for (;;) Line 2993  for (;;)
2993      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
2994      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
2995    
2996  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
2997      case OP_XCLASS:      case OP_XCLASS:
2998        {        {
2999        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2783  for (;;) Line 3018  for (;;)
3018          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3019          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3020          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3021          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
3022          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
3023          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
3024          break;          break;
3025    
3026          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2800  for (;;) Line 3035  for (;;)
3035          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3036            {            {
3037            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3038            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3039            }            }
3040          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
3041          if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3042          }          }
3043    
3044        /* 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 3055  for (;;)
3055            {            {
3056            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);
3057            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3058            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3059            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3060              {              {
3061              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3062              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3063              }              }
3064            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3065            if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3066            }            }
3067          /* Control never gets here */          /* Control never gets here */
3068          }          }
# Line 2845  for (;;) Line 3080  for (;;)
3080              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3081              break;              break;
3082              }              }
3083    #ifdef SUPPORT_UTF
3084            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3085            if (!_pcre_xclass(c, data)) break;  #else
3086              c = *eptr;
3087    #endif
3088              if (!PRIV(xclass)(c, data, utf)) break;
3089            eptr += len;            eptr += len;
3090            }            }
3091          for(;;)          for(;;)
# Line 2854  for (;;) Line 3093  for (;;)
3093            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3094            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3095            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3096            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3097              if (utf) BACKCHAR(eptr);
3098    #endif
3099            }            }
3100          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3101          }          }
3102    
3103        /* Control never gets here */        /* Control never gets here */
# Line 2866  for (;;) Line 3107  for (;;)
3107      /* Match a single character, casefully */      /* Match a single character, casefully */
3108    
3109      case OP_CHAR:      case OP_CHAR:
3110  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3111      if (utf8)      if (utf)
3112        {        {
3113        length = 1;        length = 1;
3114        ecode++;        ecode++;
# Line 2875  for (;;) Line 3116  for (;;)
3116        if (length > md->end_subject - eptr)        if (length > md->end_subject - eptr)
3117          {          {
3118          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3119          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3120          }          }
3121        while (length-- > 0) if (*ecode++ != *eptr++) MRRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);
3122        }        }
3123      else      else
3124  #endif  #endif
3125        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3126        {        {
3127        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3128          {          {
3129          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
3130          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3131          }          }
3132        if (ecode[1] != *eptr++) MRRETURN(MATCH_NOMATCH);        if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH);
3133        ecode += 2;        ecode += 2;
3134        }        }
3135      break;      break;
3136    
3137      /* Match a single character, caselessly */      /* Match a single character, caselessly. If we are at the end of the
3138        subject, give up immediately. */
3139    
3140      case OP_CHARI:      case OP_CHARI:
3141  #ifdef SUPPORT_UTF8      if (eptr >= md->end_subject)
3142      if (utf8)        {
3143          SCHECK_PARTIAL();
3144          RRETURN(MATCH_NOMATCH);
3145          }
3146    
3147    #ifdef SUPPORT_UTF
3148        if (utf)
3149        {        {
3150        length = 1;        length = 1;
3151        ecode++;        ecode++;
3152        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
3153    
       if (length > md->end_subject - eptr)  
         {  
         CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */  
         MRRETURN(MATCH_NOMATCH);  
         }  
   
3154        /* 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
3155        can use the fast lookup table. */        we know that its other case must also be one byte long, so we can use the
3156          fast lookup table. We know that there is at least one byte left in the
3157          subject. */
3158    
3159        if (fc < 128)        if (fc < 128)
3160          {          {
3161          if (md->lcc[*ecode++] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);          if (md->lcc[fc]
3162                != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3163            ecode++;
3164            eptr++;
3165          }          }
3166    
3167        /* Otherwise we must pick up the subject character */        /* Otherwise we must pick up the subject character. Note that we cannot
3168          use the value of "length" to check for sufficient bytes left, because the
3169          other case of the character may have more or fewer bytes.  */
3170    
3171        else        else
3172          {          {
# Line 2934  for (;;) Line 3182  for (;;)
3182  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3183            if (dc != UCD_OTHERCASE(fc))            if (dc != UCD_OTHERCASE(fc))
3184  #endif  #endif
3185              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3186            }            }
3187          }          }
3188        }        }
3189      else      else
3190  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3191    
3192      /* Non-UTF-8 mode */      /* Not UTF mode */
3193        {        {
3194        if (md->end_subject - eptr < 1)        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3195          {            != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3196          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */        eptr++;
         MRRETURN(MATCH_NOMATCH);  
         }  
       if (md->lcc[ecode[1]] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);  
3197        ecode += 2;        ecode += 2;
3198        }        }
3199      break;      break;
# Line 2958  for (;;) Line 3203  for (;;)
3203      case OP_EXACT:      case OP_EXACT:
3204      case OP_EXACTI:      case OP_EXACTI:
3205      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3206      ecode += 3;      ecode += 1 + IMM2_SIZE;
3207      goto REPEATCHAR;      goto REPEATCHAR;
3208    
3209      case OP_POSUPTO:      case OP_POSUPTO:
# Line 2973  for (;;) Line 3218  for (;;)
3218      min = 0;      min = 0;
3219      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3220      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3221      ecode += 3;      ecode += 1 + IMM2_SIZE;
3222      goto REPEATCHAR;      goto REPEATCHAR;
3223    
3224      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3021  for (;;) Line 3266  for (;;)
3266      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3267    
3268      REPEATCHAR:      REPEATCHAR:
3269  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3270      if (utf8)      if (utf)
3271        {        {
3272        length = 1;        length = 1;
3273        charptr = ecode;        charptr = ecode;
# Line 3038  for (;;) Line 3283  for (;;)
3283          unsigned int othercase;          unsigned int othercase;
3284          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3285              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3286            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3287          else oclength = 0;          else oclength = 0;
3288  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3289    
3290          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3291            {            {
3292            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3293              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3294  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3295            else if (oclength > 0 &&            else if (oclength > 0 &&
3296                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3297                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3298  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3299            else            else
3300              {              {
3301              CHECK_PARTIAL();              CHECK_PARTIAL();
3302              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3303              }              }
3304            }            }
3305    
# Line 3066  for (;;) Line 3311  for (;;)
3311              {              {
3312              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);
3313              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3314              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3315              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3316                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3317  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3318              else if (oclength > 0 &&              else if (oclength > 0 &&
3319                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3320                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3321  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3322              else              else
3323                {                {
3324                CHECK_PARTIAL();                CHECK_PARTIAL();
3325                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3326                }                }
3327              }              }
3328            /* Control never gets here */            /* Control never gets here */
# Line 3089  for (;;) Line 3334  for (;;)
3334            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3335              {              {
3336              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3337                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3338  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3339              else if (oclength > 0 &&              else if (oclength > 0 &&
3340                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3341                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3342  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3343              else              else
3344                {                {
# Line 3108  for (;;) Line 3353  for (;;)
3353              {              {
3354              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3355              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3356              if (eptr == pp) { MRRETURN(MATCH_NOMATCH); }              if (eptr == pp) { RRETURN(MATCH_NOMATCH); }
3357  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3358              eptr--;              eptr--;
3359              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3125  for (;;) Line 3370  for (;;)
3370        value of fc will always be < 128. */        value of fc will always be < 128. */
3371        }        }
3372      else      else
3373  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3374          /* When not in UTF-8 mode, load a single-byte character. */
3375      /* When not in UTF-8 mode, load a single-byte character. */        fc = *ecode++;
   
     fc = *ecode++;  
3376    
3377      /* The value of fc at this point is always less than 256, though we may or      /* The value of fc at this point is always one character, though we may
3378      may not be in UTF-8 mode. The code is duplicated for the caseless and      or may not be in UTF mode. The code is duplicated for the caseless and
3379      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3380      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3381      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3145  for (;;) Line 3388  for (;;)
3388    
3389      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3390        {        {
3391        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3392          /* fc must be < 128 if UTF is enabled. */
3393          foc = md->fcc[fc];
3394    #else
3395    #ifdef SUPPORT_UTF
3396    #ifdef SUPPORT_UCP
3397          if (utf && fc > 127)
3398            foc = UCD_OTHERCASE(fc);
3399    #else
3400          if (utf && fc > 127)
3401            foc = fc;
3402    #endif /* SUPPORT_UCP */
3403          else
3404    #endif /* SUPPORT_UTF */
3405            foc = TABLE_GET(fc, md->fcc, fc);
3406    #endif /* COMPILE_PCRE8 */
3407    
3408        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3409          {          {
3410          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3411            {            {
3412            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3413            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3414            }            }
3415          if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3416            eptr++;
3417          }          }
3418        if (min == max) continue;        if (min == max) continue;
3419        if (minimize)        if (minimize)
# Line 3162  for (;;) Line 3422  for (;;)
3422            {            {
3423            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3424            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3425            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
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 != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3432              eptr++;
3433            }            }
3434          /* Control never gets here */          /* Control never gets here */
3435          }          }
# Line 3182  for (;;) Line 3443  for (;;)
3443              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3444              break;              break;
3445              }              }
3446            if (fc != md->lcc[*eptr]) break;            if (fc != *eptr && foc != *eptr) break;
3447            eptr++;            eptr++;
3448            }            }
3449    
# Line 3194  for (;;) Line 3455  for (;;)
3455            eptr--;            eptr--;
3456            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3457            }            }
3458          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3459          }          }
3460        /* Control never gets here */        /* Control never gets here */
3461        }        }
# Line 3208  for (;;) Line 3469  for (;;)
3469          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3470            {            {
3471            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3472            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3473            }            }
3474          if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
3475          }          }
3476    
3477        if (min == max) continue;        if (min == max) continue;
# Line 3221  for (;;) Line 3482  for (;;)
3482            {            {
3483            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);
3484            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3485            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3486            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3487              {              {
3488              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3489              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3490              }              }
3491            if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
3492            }            }
3493          /* Control never gets here */          /* Control never gets here */
3494          }          }
# Line 3252  for (;;) Line 3513  for (;;)
3513            eptr--;            eptr--;
3514            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3515            }            }
3516          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3517          }          }
3518        }        }
3519      /* Control never gets here */      /* Control never gets here */
# Line 3261  for (;;) Line 3522  for (;;)
3522      checking can be multibyte. */      checking can be multibyte. */
3523    
3524      case OP_NOT:      case OP_NOT:
3525      case OP_NOTI:      case OP_NOTI:
3526      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
3527        {        {
3528        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3529        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3530        }        }
3531      ecode++;      ecode++;
3532      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
3533      if (op == OP_NOTI)         /* The caseless case */      if (op == OP_NOTI)         /* The caseless case */
3534        {        {
3535  #ifdef SUPPORT_UTF8        register unsigned int ch, och;
3536        if (c < 256)        ch = *ecode++;
3537  #endif  #ifdef COMPILE_PCRE8
3538        c = md->lcc[c];        /* ch must be < 128 if UTF is enabled. */
3539        if (md->lcc[*ecode++] == c) MRRETURN(MATCH_NOMATCH);        och = md->fcc[ch];
3540    #else
3541    #ifdef SUPPORT_UTF
3542    #ifdef SUPPORT_UCP
3543          if (utf && ch > 127)
3544            och = UCD_OTHERCASE(ch);
3545    #else
3546          if (utf && ch > 127)
3547            och = ch;
3548    #endif /* SUPPORT_UCP */
3549          else
3550    #endif /* SUPPORT_UTF */
3551            och = TABLE_GET(ch, md->fcc, ch);
3552    #endif /* COMPILE_PCRE8 */
3553          if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3554        }        }
3555      else    /* Caseful */      else    /* Caseful */
3556        {        {
3557        if (*ecode++ == c) MRRETURN(MATCH_NOMATCH);        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);
3558        }        }
3559      break;      break;
3560    
# Line 3293  for (;;) Line 3568  for (;;)
3568      case OP_NOTEXACT:      case OP_NOTEXACT:
3569      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3570      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3571      ecode += 3;      ecode += 1 + IMM2_SIZE;
3572      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3573    
3574      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3303  for (;;) Line 3578  for (;;)
3578      min = 0;      min = 0;
3579      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3580      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3581      ecode += 3;      ecode += 1 + IMM2_SIZE;
3582      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3583    
3584      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3335  for (;;) Line 3610  for (;;)
3610      possessive = TRUE;      possessive = TRUE;
3611      min = 0;      min = 0;
3612      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3613      ecode += 3;      ecode += 1 + IMM2_SIZE;
3614      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3615    
3616      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3374  for (;;) Line 3649  for (;;)
3649    
3650      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3651        {        {
3652        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3653          /* fc must be < 128 if UTF is enabled. */
3654          foc = md->fcc[fc];
3655    #else
3656    #ifdef SUPPORT_UTF
3657    #ifdef SUPPORT_UCP
3658          if (utf && fc > 127)
3659            foc = UCD_OTHERCASE(fc);
3660    #else
3661          if (utf && fc > 127)
3662            foc = fc;
3663    #endif /* SUPPORT_UCP */
3664          else
3665    #endif /* SUPPORT_UTF */
3666            foc = TABLE_GET(fc, md->fcc, fc);
3667    #endif /* COMPILE_PCRE8 */
3668    
3669  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3670        /* UTF-8 mode */        if (utf)
       if (utf8)  
3671          {          {
3672          register unsigned int d;          register unsigned int d;
3673          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3386  for (;;) Line 3675  for (;;)
3675            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3676              {              {
3677              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3678              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3679              }              }
3680            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3681            if (d < 256) d = md->lcc[d];            if (fc == d || (unsigned int) foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) MRRETURN(MATCH_NOMATCH);  
3682            }            }
3683          }          }
3684        else        else
3685  #endif  #endif
3686          /* Not UTF mode */
       /* Not UTF-8 mode */  
3687          {          {
3688          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3689            {            {
3690            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3691              {              {
3692              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3693              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3694              }              }
3695            if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3696              eptr++;
3697            }            }
3698          }          }
3699    
# Line 3413  for (;;) Line 3701  for (;;)
3701    
3702        if (minimize)        if (minimize)
3703          {          {
3704  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3705          /* UTF-8 mode */          if (utf)
         if (utf8)  
3706            {            {
3707            register unsigned int d;            register unsigned int d;
3708            for (fi = min;; fi++)            for (fi = min;; fi++)
3709              {              {
3710              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
3711              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3712              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3713              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3714                {                {
3715                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3716                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3717                }                }
3718              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3719              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) MRRETURN(MATCH_NOMATCH);  
3720              }              }
3721            }            }
3722          else          else
3723  #endif  #endif
3724          /* Not UTF-8 mode */          /* Not UTF mode */
3725            {            {
3726            for (fi = min;; fi++)            for (fi = min;; fi++)
3727              {              {
3728              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);
3729              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3730              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3731              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3732                {                {
3733                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3734                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3735                }                }
3736              if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3737                eptr++;
3738              }              }
3739            }            }
3740          /* Control never gets here */          /* Control never gets here */
# Line 3459  for (;;) Line 3746  for (;;)
3746          {          {
3747          pp = eptr;          pp = eptr;
3748    
3749  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3750          /* UTF-8 mode */          if (utf)
         if (utf8)  
3751            {            {
3752            register unsigned int d;            register unsigned int d;
3753            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3473  for (;;) Line 3759  for (;;)
3759                break;                break;
3760                }                }
3761              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3762              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) break;
             if (fc == d) break;  
3763              eptr += len;              eptr += len;
3764              }              }
3765          if (possessive) continue;            if (possessive) continue;
3766          for(;;)            for(;;)
3767              {              {
3768              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3769              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 3488  for (;;) Line 3773  for (;;)
3773            }            }
3774          else          else
3775  #endif  #endif
3776          /* Not UTF-8 mode */          /* Not UTF mode */
3777            {            {
3778            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3779              {              {
# Line 3497  for (;;) Line 3782  for (;;)
3782                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3783                break;                break;
3784                }                }
3785              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3786              eptr++;              eptr++;
3787              }              }
3788            if (possessive) continue;            if (possessive) continue;
# Line 3509  for (;;) Line 3794  for (;;)
3794              }              }
3795            }            }
3796    
3797          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3798          }          }
3799        /* Control never gets here */        /* Control never gets here */
3800        }        }
# Line 3518  for (;;) Line 3803  for (;;)
3803    
3804      else      else
3805        {        {
3806  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3807        /* UTF-8 mode */        if (utf)
       if (utf8)  
3808          {          {
3809          register unsigned int d;          register unsigned int d;
3810          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3528  for (;;) Line 3812  for (;;)
3812            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3813              {              {
3814              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3815              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3816              }              }
3817            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3818            if (fc == d) MRRETURN(MATCH_NOMATCH);            if (fc == d) RRETURN(MATCH_NOMATCH);
3819            }            }
3820          }          }
3821        else        else
3822  #endif  #endif
3823        /* Not UTF-8 mode */        /* Not UTF mode */
3824          {          {
3825          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3826            {            {
3827            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3828              {              {
3829              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3830              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3831              }              }
3832            if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);            if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
3833            }            }
3834          }          }
3835    
# Line 3553  for (;;) Line 3837  for (;;)
3837    
3838        if (minimize)        if (minimize)
3839          {          {
3840  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3841          /* UTF-8 mode */          if (utf)
         if (utf8)  
3842            {            {
3843            register unsigned int d;            register unsigned int d;
3844            for (fi = min;; fi++)            for (fi = min;; fi++)
3845              {              {
3846              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
3847              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3848              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3849              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3850                {                {
3851                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3852                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3853                }                }
3854              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3855              if (fc == d) MRRETURN(MATCH_NOMATCH);              if (fc == d) RRETURN(MATCH_NOMATCH);
3856              }              }
3857            }            }
3858          else          else
3859  #endif  #endif
3860          /* Not UTF-8 mode */          /* Not UTF mode */
3861            {            {
3862            for (fi = min;; fi++)            for (fi = min;; fi++)
3863              {              {
3864              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);
3865              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3866              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3867              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3868                {                {
3869                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3870                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3871                }                }
3872              if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);              if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
3873              }              }
3874            }            }
3875          /* Control never gets here */          /* Control never gets here */
# Line 3598  for (;;) Line 3881  for (;;)
3881          {          {
3882          pp = eptr;          pp = eptr;
3883    
3884  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3885          /* UTF-8 mode */          if (utf)
         if (utf8)  
3886            {            {
3887            register unsigned int d;            register unsigned int d;
3888            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3626  for (;;) Line 3908  for (;;)
3908            }            }
3909          else          else
3910  #endif  #endif
3911          /* Not UTF-8 mode */          /* Not UTF mode */
3912            {            {
3913            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3914              {              {
# Line 3647  for (;;) Line 3929  for (;;)
3929              }              }
3930            }            }
3931    
3932          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3933          }          }
3934        }        }
3935      /* Control never gets here */      /* Control never gets here */
# Line 3659  for (;;) Line 3941  for (;;)
3941      case OP_TYPEEXACT:      case OP_TYPEEXACT:
3942      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3943      minimize = TRUE;      minimize = TRUE;
3944      ecode += 3;      ecode += 1 + IMM2_SIZE;
3945      goto REPEATTYPE;      goto REPEATTYPE;
3946    
3947      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3667  for (;;) Line 3949  for (;;)
3949      min = 0;      min = 0;
3950      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3951      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
3952      ecode += 3;      ecode += 1 + IMM2_SIZE;
3953      goto REPEATTYPE;      goto REPEATTYPE;
3954    
3955      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3695  for (;;) Line 3977  for (;;)
3977      possessive = TRUE;      possessive = TRUE;
3978      min = 0;      min = 0;
3979      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3980      ecode += 3;      ecode += 1 + IMM2_SIZE;
3981      goto REPEATTYPE;      goto REPEATTYPE;
3982    
3983      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 3741  for (;;) Line 4023  for (;;)
4023          switch(prop_type)          switch(prop_type)
4024            {            {
4025            case PT_ANY:            case PT_ANY:
4026            if (prop_fail_result) MRRETURN(MATCH_NOMATCH);            if (prop_fail_result) RRETURN(MATCH_NOMATCH);
4027            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4028              {              {
4029              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4030                {                {
4031                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4032                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4033                }                }
4034              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4035              }              }
# Line 3756  for (;;) Line 4038  for (;;)
4038            case PT_LAMP:            case PT_LAMP:
4039            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4040              {              {
4041                int chartype;
4042              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4043                {                {
4044                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4045                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4046                }                }
4047              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4048              prop_chartype = UCD_CHARTYPE(c);              chartype = UCD_CHARTYPE(c);
4049              if ((prop_chartype == ucp_Lu ||              if ((chartype == ucp_Lu ||
4050                   prop_chartype == ucp_Ll ||                   chartype == ucp_Ll ||
4051                   prop_chartype == ucp_Lt) == prop_fail_result)                   chartype == ucp_Lt) == prop_fail_result)
4052                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4053              }              }
4054            break;            break;
4055    
# Line 3776  for (;;) Line 4059  for (;;)
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);              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)
4066              if ((prop_category == prop_value) == prop_fail_result)                RRETURN(MATCH_NOMATCH);
               MRRETURN(MATCH_NOMATCH);  
4067              }              }
4068            break;            break;
4069    
# Line 3791  for (;;) Line 4073  for (;;)
4073              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4074                {                {
4075                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4076                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4077                }                }
4078              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4079              prop_chartype = UCD_CHARTYPE(c);              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)
4080              if ((prop_chartype == prop_value) == prop_fail_result)                RRETURN(MATCH_NOMATCH);
               MRRETURN(MATCH_NOMATCH);  
4081              }              }
4082            break;            break;
4083    
# Line 3806  for (;;) Line 4087  for (;;)
4087              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4088                {                {
4089                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4090                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4091                }                }
4092              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4093              prop_script = UCD_SCRIPT(c);              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)
4094              if ((prop_script == prop_value) == prop_fail_result)                RRETURN(MATCH_NOMATCH);
               MRRETURN(MATCH_NOMATCH);  
4095              }              }
4096            break;            break;
4097    
4098            case PT_ALNUM:            case PT_ALNUM:
4099            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4100              {              {
4101                int category;
4102              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4103                {                {
4104                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4105                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4106                }                }
4107              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4108              prop_category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4109              if ((prop_category == ucp_L || prop_category == ucp_N)              if ((category == ucp_L || category == ucp_N) == prop_fail_result)
4110                     == prop_fail_result)                RRETURN(MATCH_NOMATCH);
               MRRETURN(MATCH_NOMATCH);  
4111              }              }
4112            break;            break;
4113    
# Line 3837  for (;;) Line 4117  for (;;)
4117              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4118                {                {
4119                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4120                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4121                }                }
4122              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4123              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 ||  
4124                   c == CHAR_FF || c == CHAR_CR)                   c == CHAR_FF || c == CHAR_CR)
4125                     == prop_fail_result)                     == prop_fail_result)
4126                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4127              }              }
4128            break;            break;
4129    
# Line 3854  for (;;) Line 4133  for (;;)
4133              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4134                {                {
4135                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4136                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4137                }                }
4138              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4139              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 ||  
4140                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
4141                     == prop_fail_result)                     == prop_fail_result)
4142                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4143              }              }
4144            break;            break;