/[pcre]/code/branches/pcre16/pcre_exec.c
ViewVC logotype

Diff of /code/branches/pcre16/pcre_exec.c

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

revision 459 by ph10, Sun Oct 4 09:21:39 2009 UTC revision 630 by ph10, Fri Jul 22 10:00:10 2011 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) 1997-2009 University of Cambridge             Copyright (c) 1997-2011 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  /* Flag bits for the match() function */  /* 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,
62    as stack usage is to be discouraged. */
63    
64  #define match_condassert     0x01  /* Called to check a condition assertion */  #define MATCH_CONDASSERT     1  /* Called to check a condition assertion */
65  #define match_cbegroup       0x02  /* Could-be-empty unlimited repeat group */  #define MATCH_CBEGROUP       2  /* Could-be-empty unlimited repeat group */
66    
67  /* Non-error returns from the match() function. Error returns are externally  /* Non-error returns from the match() function. Error returns are externally
68  defined PCRE_ERROR_xxx codes, which are all negative. */  defined PCRE_ERROR_xxx codes, which are all negative. */
# Line 71  defined PCRE_ERROR_xxx codes, which are Line 73  defined PCRE_ERROR_xxx codes, which are
73  /* Special internal returns from the match() function. Make them sufficiently  /* Special internal returns from the match() function. Make them sufficiently
74  negative to avoid the external error codes. */  negative to avoid the external error codes. */
75    
76  #define MATCH_COMMIT       (-999)  #define MATCH_ACCEPT       (-999)
77  #define MATCH_PRUNE        (-998)  #define MATCH_COMMIT       (-998)
78  #define MATCH_SKIP         (-997)  #define MATCH_KETRPOS      (-997)
79  #define MATCH_THEN         (-996)  #define MATCH_ONCE         (-996)
80    #define MATCH_PRUNE        (-995)
81    #define MATCH_SKIP         (-994)
82    #define MATCH_SKIP_ARG     (-993)
83    #define MATCH_THEN         (-992)
84    
85    /* This is a convenience macro for code that occurs many times. */
86    
87    #define MRRETURN(ra) \
88      { \
89      md->mark = markptr; \
90      RRETURN(ra); \
91      }
92    
93  /* 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.
94  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 89  static const char rep_max[] = { 0, 0, 0, Line 103  static const char rep_max[] = { 0, 0, 0,
103    
104    
105    
106  #ifdef DEBUG  #ifdef PCRE_DEBUG
107  /*************************************************  /*************************************************
108  *        Debugging function to print chars       *  *        Debugging function to print chars       *
109  *************************************************/  *************************************************/
# Line 122  while (length-- > 0) Line 136  while (length-- > 0)
136  *          Match a back-reference                *  *          Match a back-reference                *
137  *************************************************/  *************************************************/
138    
139  /* If a back reference hasn't been set, the length that is passed is greater  /* Normally, if a back reference hasn't been set, the length that is passed is
140  than the number of characters left in the string, so the match fails.  negative, so the match always fails. However, in JavaScript compatibility mode,
141    the length passed is zero. Note that in caseless UTF-8 mode, the number of
142    subject bytes matched may be different to the number of reference bytes.
143    
144  Arguments:  Arguments:
145    offset      index into the offset vector    offset      index into the offset vector
146    eptr        points into the subject    eptr        pointer into the subject
147    length      length to be matched    length      length of reference to be matched (number of bytes)
148    md          points to match data block    md          points to match data block
149    ims         the ims flags    caseless    TRUE if caseless
150    
151  Returns:      TRUE if matched  Returns:      < 0 if not matched, otherwise the number of subject bytes matched
152  */  */
153    
154  static BOOL  static int
155  match_ref(int offset, register USPTR eptr, int length, match_data *md,  match_ref(int offset, register USPTR eptr, int length, match_data *md,
156    unsigned long int ims)    BOOL caseless)
157  {  {
158  USPTR p = md->start_subject + md->offset_vector[offset];  USPTR eptr_start = eptr;
159    register USPTR p = md->start_subject + md->offset_vector[offset];
160    
161  #ifdef DEBUG  #ifdef PCRE_DEBUG
162  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
163    printf("matching subject <null>");    printf("matching subject <null>");
164  else  else
# Line 154  pchars(p, length, FALSE, md); Line 171  pchars(p, length, FALSE, md);
171  printf("\n");  printf("\n");
172  #endif  #endif
173    
174  /* Always fail if not enough characters left */  /* Always fail if reference not set (and not JavaScript compatible). */
175    
176  if (length > md->end_subject - eptr) return FALSE;  if (length < 0) return -1;
177    
178  /* Separate the caseless case for speed. In UTF-8 mode we can only do this  /* Separate the caseless case for speed. In UTF-8 mode we can only do this
179  properly if Unicode properties are supported. Otherwise, we can check only  properly if Unicode properties are supported. Otherwise, we can check only
180  ASCII characters. */  ASCII characters. */
181    
182  if ((ims & PCRE_CASELESS) != 0)  if (caseless)
183    {    {
184  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
185  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
186    if (md->utf8)    if (md->utf8)
187      {      {
188      USPTR endptr = eptr + length;      /* Match characters up to the end of the reference. NOTE: the number of
189      while (eptr < endptr)      bytes matched may differ, because there are some characters whose upper and
190        lower case versions code as different numbers of bytes. For example, U+023A
191        (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);
192        a sequence of 3 of the former uses 6 bytes, as does a sequence of two of
193        the latter. It is important, therefore, to check the length along the
194        reference, not along the subject (earlier code did this wrong). */
195    
196        USPTR endptr = p + length;
197        while (p < endptr)
198        {        {
199        int c, d;        int c, d;
200          if (eptr >= md->end_subject) return -1;
201        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
202        GETCHARINC(d, p);        GETCHARINC(d, p);
203        if (c != d && c != UCD_OTHERCASE(d)) return FALSE;        if (c != d && c != UCD_OTHERCASE(d)) return -1;
204        }        }
205      }      }
206    else    else
# Line 183  if ((ims & PCRE_CASELESS) != 0) Line 209  if ((ims & PCRE_CASELESS) != 0)
209    
210    /* 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
211    is no UCP support. */    is no UCP support. */
212        {
213    while (length-- > 0)      if (eptr + length > md->end_subject) return -1;
214      { if (md->lcc[*p++] != md->lcc[*eptr++]) return FALSE; }      while (length-- > 0)
215          { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }
216        }
217    }    }
218    
219  /* 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
220  are in UTF-8 mode. */  are in UTF-8 mode. */
221    
222  else  else
223    { while (length-- > 0) if (*p++ != *eptr++) return FALSE; }    {
224      if (eptr + length > md->end_subject) return -1;
225      while (length-- > 0) if (*p++ != *eptr++) return -1;
226      }
227    
228  return TRUE;  return eptr - eptr_start;
229  }  }
230    
231    
# Line 245  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 276  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
276         RM21,  RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30,         RM21,  RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30,
277         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
278         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
279         RM51,  RM52, RM53, RM54 };         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
280           RM61,  RM62, RM63 };
281    
282  /* 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
283  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
284  actuall used in this definition. */  actually used in this definition. */
285    
286  #ifndef NO_RECURSE  #ifndef NO_RECURSE
287  #define REGISTER register  #define REGISTER register
288    
289  #ifdef DEBUG  #ifdef PCRE_DEBUG
290  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
291    { \    { \
292    printf("match() called in line %d\n", __LINE__); \    printf("match() called in line %d\n", __LINE__); \
293    rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1); \    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1); \
294    printf("to line %d\n", __LINE__); \    printf("to line %d\n", __LINE__); \
295    }    }
296  #define RRETURN(ra) \  #define RRETURN(ra) \
# Line 267  actuall used in this definition. */ Line 299  actuall used in this definition. */
299    return ra; \    return ra; \
300    }    }
301  #else  #else
302  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
303    rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1)    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1)
304  #define RRETURN(ra) return ra  #define RRETURN(ra) return ra
305  #endif  #endif
306    
# Line 281  argument of match(), which never changes Line 313  argument of match(), which never changes
313    
314  #define REGISTER  #define REGISTER
315    
316  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
317    {\    {\
318    heapframe *newframe = (pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\
319      if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
320    frame->Xwhere = rw; \    frame->Xwhere = rw; \
321    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
322    newframe->Xecode = rb;\    newframe->Xecode = rb;\
323    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
324      newframe->Xmarkptr = markptr;\
325    newframe->Xoffset_top = rc;\    newframe->Xoffset_top = rc;\
326    newframe->Xims = re;\    newframe->Xeptrb = re;\
   newframe->Xeptrb = rf;\  
   newframe->Xflags = rg;\  
327    newframe->Xrdepth = frame->Xrdepth + 1;\    newframe->Xrdepth = frame->Xrdepth + 1;\
328    newframe->Xprevframe = frame;\    newframe->Xprevframe = frame;\
329    frame = newframe;\    frame = newframe;\
# Line 303  argument of match(), which never changes Line 335  argument of match(), which never changes
335    
336  #define RRETURN(ra)\  #define RRETURN(ra)\
337    {\    {\
338    heapframe *newframe = frame;\    heapframe *oldframe = frame;\
339    frame = newframe->Xprevframe;\    frame = oldframe->Xprevframe;\
340    (pcre_stack_free)(newframe);\    (pcre_stack_free)(oldframe);\
341    if (frame != NULL)\    if (frame != NULL)\
342      {\      {\
343      rrc = ra;\      rrc = ra;\
# Line 325  typedef struct heapframe { Line 357  typedef struct heapframe {
357    USPTR Xeptr;    USPTR Xeptr;
358    const uschar *Xecode;    const uschar *Xecode;
359    USPTR Xmstart;    USPTR Xmstart;
360      USPTR Xmarkptr;
361    int Xoffset_top;    int Xoffset_top;
   long int Xims;  
362    eptrblock *Xeptrb;    eptrblock *Xeptrb;
   int Xflags;  
363    unsigned int Xrdepth;    unsigned int Xrdepth;
364    
365    /* Function local variables */    /* Function local variables */
# Line 349  typedef struct heapframe { Line 380  typedef struct heapframe {
380    BOOL Xcondition;    BOOL Xcondition;
381    BOOL Xprev_is_word;    BOOL Xprev_is_word;
382    
   unsigned long int Xoriginal_ims;  
   
383  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
384    int Xprop_type;    int Xprop_type;
385    int Xprop_value;    int Xprop_value;
386    int Xprop_fail_result;    int Xprop_fail_result;
   int Xprop_category;  
   int Xprop_chartype;  
   int Xprop_script;  
387    int Xoclength;    int Xoclength;
388    uschar Xocchars[8];    uschar Xocchars[8];
389  #endif  #endif
# Line 408  immediately. The second one is used when Line 434  immediately. The second one is used when
434  the subject. */  the subject. */
435    
436  #define CHECK_PARTIAL()\  #define CHECK_PARTIAL()\
437    if (md->partial != 0 && eptr >= md->end_subject && eptr > mstart)\    if (md->partial != 0 && eptr >= md->end_subject && \
438      {\        eptr > md->start_used_ptr) \
439      md->hitend = TRUE;\      { \
440      if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);\      md->hitend = TRUE; \
441        if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \
442      }      }
443    
444  #define SCHECK_PARTIAL()\  #define SCHECK_PARTIAL()\
445    if (md->partial && eptr > mstart)\    if (md->partial != 0 && eptr > md->start_used_ptr) \
446      {\      { \
447      md->hitend = TRUE;\      md->hitend = TRUE; \
448      if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);\      if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \
449      }      }
450    
451    
# Line 432  Arguments: Line 459  Arguments:
459     ecode       pointer to current position in compiled code     ecode       pointer to current position in compiled code
460     mstart      pointer to the current match start position (can be modified     mstart      pointer to the current match start position (can be modified
461                   by encountering \K)                   by encountering \K)
462       markptr     pointer to the most recent MARK name, or NULL
463     offset_top  current top pointer     offset_top  current top pointer
464     md          pointer to "static" info for the match     md          pointer to "static" info for the match
    ims         current /i, /m, and /s options  
465     eptrb       pointer to chain of blocks containing eptr at start of     eptrb       pointer to chain of blocks containing eptr at start of
466                   brackets - for testing for empty matches                   brackets - for testing for empty matches
    flags       can contain  
                  match_condassert - this is an assertion condition  
                  match_cbegroup - this is the start of an unlimited repeat  
                    group that can match an empty string  
467     rdepth      the recursion depth     rdepth      the recursion depth
468    
469  Returns:       MATCH_MATCH if matched            )  these values are >= 0  Returns:       MATCH_MATCH if matched            )  these values are >= 0
470                 MATCH_NOMATCH if failed to match  )                 MATCH_NOMATCH if failed to match  )
471                   a negative MATCH_xxx value for PRUNE, SKIP, etc
472                 a negative PCRE_ERROR_xxx value if aborted by an error condition                 a negative PCRE_ERROR_xxx value if aborted by an error condition
473                   (e.g. stopped by repeated call or recursion limit)                   (e.g. stopped by repeated call or recursion limit)
474  */  */
475    
476  static int  static int
477  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,
478    int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb,    const uschar *markptr, int offset_top, match_data *md, eptrblock *eptrb,
479    int flags, unsigned int rdepth)    unsigned int rdepth)
480  {  {
481  /* 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,
482  so they can be ordinary variables in all cases. Mark some of them with  so they can be ordinary variables in all cases. Mark some of them with
# Line 464  register unsigned int c;   /* Character Line 488  register unsigned int c;   /* Character
488  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */
489    
490  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
491    BOOL caseless;
492  int condcode;  int condcode;
493    
494  /* 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
# Line 472  heap storage. Set up the top-level frame Line 497  heap storage. Set up the top-level frame
497  heap whenever RMATCH() does a "recursion". See the macro definitions above. */  heap whenever RMATCH() does a "recursion". See the macro definitions above. */
498    
499  #ifdef NO_RECURSE  #ifdef NO_RECURSE
500  heapframe *frame = (pcre_stack_malloc)(sizeof(heapframe));  heapframe *frame = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));
501    if (frame == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
502  frame->Xprevframe = NULL;            /* Marks the top level */  frame->Xprevframe = NULL;            /* Marks the top level */
503    
504  /* Copy in the original argument variables */  /* Copy in the original argument variables */
# Line 480  frame->Xprevframe = NULL;            /* Line 506  frame->Xprevframe = NULL;            /*
506  frame->Xeptr = eptr;  frame->Xeptr = eptr;
507  frame->Xecode = ecode;  frame->Xecode = ecode;
508  frame->Xmstart = mstart;  frame->Xmstart = mstart;
509    frame->Xmarkptr = markptr;
510  frame->Xoffset_top = offset_top;  frame->Xoffset_top = offset_top;
 frame->Xims = ims;  
511  frame->Xeptrb = eptrb;  frame->Xeptrb = eptrb;
 frame->Xflags = flags;  
512  frame->Xrdepth = rdepth;  frame->Xrdepth = rdepth;
513    
514  /* This is where control jumps back to to effect "recursion" */  /* This is where control jumps back to to effect "recursion" */
# Line 495  HEAP_RECURSE: Line 520  HEAP_RECURSE:
520  #define eptr               frame->Xeptr  #define eptr               frame->Xeptr
521  #define ecode              frame->Xecode  #define ecode              frame->Xecode
522  #define mstart             frame->Xmstart  #define mstart             frame->Xmstart
523    #define markptr            frame->Xmarkptr
524  #define offset_top         frame->Xoffset_top  #define offset_top         frame->Xoffset_top
 #define ims                frame->Xims  
525  #define eptrb              frame->Xeptrb  #define eptrb              frame->Xeptrb
 #define flags              frame->Xflags  
526  #define rdepth             frame->Xrdepth  #define rdepth             frame->Xrdepth
527    
528  /* Ditto for the local variables */  /* Ditto for the local variables */
# Line 520  HEAP_RECURSE: Line 544  HEAP_RECURSE:
544  #define condition          frame->Xcondition  #define condition          frame->Xcondition
545  #define prev_is_word       frame->Xprev_is_word  #define prev_is_word       frame->Xprev_is_word
546    
 #define original_ims       frame->Xoriginal_ims  
   
547  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
548  #define prop_type          frame->Xprop_type  #define prop_type          frame->Xprop_type
549  #define prop_value         frame->Xprop_value  #define prop_value         frame->Xprop_value
550  #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  
551  #define oclength           frame->Xoclength  #define oclength           frame->Xoclength
552  #define occhars            frame->Xocchars  #define occhars            frame->Xocchars
553  #endif  #endif
# Line 558  i, and fc and c, can be the same variabl Line 577  i, and fc and c, can be the same variabl
577  #define fi i  #define fi i
578  #define fc c  #define fc c
579    
580    /* Many of the following variables are used only in small blocks of the code.
581    My normal style of coding would have declared them within each of those blocks.
582    However, in order to accommodate the version of this code that uses an external
583    "stack" implemented on the heap, it is easier to declare them all here, so the
584    declarations can be cut out in a block. The only declarations within blocks
585    below are for variables that do not have to be preserved over a recursive call
586    to RMATCH(). */
587    
588  #ifdef SUPPORT_UTF8                /* Many of these variables are used only  */  #ifdef SUPPORT_UTF8
589  const uschar *charptr;             /* in small blocks of the code. My normal */  const uschar *charptr;
590  #endif                             /* style of coding would have declared    */  #endif
591  const uschar *callpat;             /* them within each of those blocks.      */  const uschar *callpat;
592  const uschar *data;                /* However, in order to accommodate the   */  const uschar *data;
593  const uschar *next;                /* version of this code that uses an      */  const uschar *next;
594  USPTR         pp;                  /* external "stack" implemented on the    */  USPTR         pp;
595  const uschar *prev;                /* heap, it is easier to declare them all */  const uschar *prev;
596  USPTR         saved_eptr;          /* here, so the declarations can be cut   */  USPTR         saved_eptr;
597                                     /* out in a block. The only declarations  */  
598  recursion_info new_recursive;      /* within blocks below are for variables  */  recursion_info new_recursive;
599                                     /* that do not have to be preserved over  */  
600  BOOL cur_is_word;                  /* a recursive call to RMATCH().          */  BOOL cur_is_word;
601  BOOL condition;  BOOL condition;
602  BOOL prev_is_word;  BOOL prev_is_word;
603    
 unsigned long int original_ims;  
   
604  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
605  int prop_type;  int prop_type;
606  int prop_value;  int prop_value;
607  int prop_fail_result;  int prop_fail_result;
 int prop_category;  
 int prop_chartype;  
 int prop_script;  
608  int oclength;  int oclength;
609  uschar occhars[8];  uschar occhars[8];
610  #endif  #endif
# Line 603  int stacksave[REC_STACK_SAVE_MAX]; Line 624  int stacksave[REC_STACK_SAVE_MAX];
624  eptrblock newptrb;  eptrblock newptrb;
625  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
626    
627    /* To save space on the stack and in the heap frame, I have doubled up on some
628    of the local variables that are used only in localised parts of the code, but
629    still need to be preserved over recursive calls of match(). These macros define
630    the alternative names that are used. */
631    
632    #define allow_zero    cur_is_word
633    #define cbegroup      condition
634    #define code_offset   codelink
635    #define condassert    condition
636    #define matched_once  prev_is_word
637    
638  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
639  variables. */  variables. */
640    
# Line 622  TAIL_RECURSE: Line 654  TAIL_RECURSE:
654  /* OK, now we can get on with the real code of the function. Recursive calls  /* OK, now we can get on with the real code of the function. Recursive calls
655  are specified by the macro RMATCH and RRETURN is used to return. When  are specified by the macro RMATCH and RRETURN is used to return. When
656  NO_RECURSE is *not* defined, these just turn into a recursive call to match()  NO_RECURSE is *not* defined, these just turn into a recursive call to match()
657  and a "return", respectively (possibly with some debugging if DEBUG is  and a "return", respectively (possibly with some debugging if PCRE_DEBUG is
658  defined). However, RMATCH isn't like a function call because it's quite a  defined). However, RMATCH isn't like a function call because it's quite a
659  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,
660  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
# Line 639  haven't exceeded the recursive call limi Line 671  haven't exceeded the recursive call limi
671  if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT);  if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT);
672  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);
673    
 original_ims = ims;    /* Save for resetting on ')' */  
   
674  /* 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
675  string, the match_cbegroup flag is set. When this is the case, add the current  string, the variable md->match_function_type is set to MATCH_CBEGROUP. It is
676  subject pointer to the chain of such remembered pointers, to be checked when we  done this way to save having to use another function argument, which would take
677  hit the closing ket, in order to break infinite loops that match no characters.  up space on the stack. See also MATCH_CONDASSERT below.
678  When match() is called in other circumstances, don't add to the chain. The  
679  match_cbegroup flag must NOT be used with tail recursion, because the memory  When MATCH_CBEGROUP is set, add the current subject pointer to the chain of
680  block that is used is on the stack, so a new one may be required for each  such remembered pointers, to be checked when we hit the closing ket, in order
681  match(). */  to break infinite loops that match no characters. When match() is called in
682    other circumstances, don't add to the chain. The MATCH_CBEGROUP feature must
683    NOT be used with tail recursion, because the memory block that is used is on
684    the stack, so a new one may be required for each match(). */
685    
686  if ((flags & match_cbegroup) != 0)  if (md->match_function_type == MATCH_CBEGROUP)
687    {    {
688    newptrb.epb_saved_eptr = eptr;    newptrb.epb_saved_eptr = eptr;
689    newptrb.epb_prev = eptrb;    newptrb.epb_prev = eptrb;
690    eptrb = &newptrb;    eptrb = &newptrb;
691      md->match_function_type = 0;
692    }    }
693    
694  /* Now start processing the opcodes. */  /* Now start processing the opcodes. */
# Line 666  for (;;) Line 700  for (;;)
700    
701    switch(op)    switch(op)
702      {      {
703        case OP_MARK:
704        markptr = ecode + 2;
705        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,
706          eptrb, RM55);
707    
708        /* A return of MATCH_SKIP_ARG means that matching failed at SKIP with an
709        argument, and we must check whether that argument matches this MARK's
710        argument. It is passed back in md->start_match_ptr (an overloading of that
711        variable). If it does match, we reset that variable to the current subject
712        position and return MATCH_SKIP. Otherwise, pass back the return code
713        unaltered. */
714    
715        if (rrc == MATCH_SKIP_ARG &&
716            strcmp((char *)markptr, (char *)(md->start_match_ptr)) == 0)
717          {
718          md->start_match_ptr = eptr;
719          RRETURN(MATCH_SKIP);
720          }
721    
722        if (md->mark == NULL) md->mark = markptr;
723        RRETURN(rrc);
724    
725      case OP_FAIL:      case OP_FAIL:
726      RRETURN(MATCH_NOMATCH);      MRRETURN(MATCH_NOMATCH);
727    
728        /* COMMIT overrides PRUNE, SKIP, and THEN */
729    
730        case OP_COMMIT:
731        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
732          eptrb, RM52);
733        if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&
734            rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&
735            rrc != MATCH_THEN)
736          RRETURN(rrc);
737        MRRETURN(MATCH_COMMIT);
738    
739        /* PRUNE overrides THEN */
740    
741      case OP_PRUNE:      case OP_PRUNE:
742      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
743        ims, eptrb, flags, RM51);        eptrb, RM51);
744      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
745        MRRETURN(MATCH_PRUNE);
746    
747        case OP_PRUNE_ARG:
748        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,
749          eptrb, RM56);
750        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
751        md->mark = ecode + 2;
752      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
753    
754      case OP_COMMIT:      /* SKIP overrides PRUNE and THEN */
     RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,  
       ims, eptrb, flags, RM52);  
     if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
     RRETURN(MATCH_COMMIT);  
755    
756      case OP_SKIP:      case OP_SKIP:
757      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
758        ims, eptrb, flags, RM53);        eptrb, RM53);
759      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
760          RRETURN(rrc);
761      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
762      RRETURN(MATCH_SKIP);      MRRETURN(MATCH_SKIP);
763    
764        case OP_SKIP_ARG:
765        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,
766          eptrb, RM57);
767        if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
768          RRETURN(rrc);
769    
770        /* Pass back the current skip name by overloading md->start_match_ptr and
771        returning the special MATCH_SKIP_ARG return code. This will either be
772        caught by a matching MARK, or get to the top, where it is treated the same
773        as PRUNE. */
774    
775        md->start_match_ptr = ecode + 2;
776        RRETURN(MATCH_SKIP_ARG);
777    
778        /* For THEN (and THEN_ARG) we pass back the address of the bracket or
779        the alt that is at the start of the current branch. This makes it possible
780        to skip back past alternatives that precede the THEN within the current
781        branch. */
782    
783      case OP_THEN:      case OP_THEN:
784      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
785        ims, eptrb, flags, RM54);        eptrb, RM54);
786      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
787        md->start_match_ptr = ecode - GET(ecode, 1);
788        MRRETURN(MATCH_THEN);
789    
790        case OP_THEN_ARG:
791        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1+LINK_SIZE],
792          offset_top, md, eptrb, RM58);
793        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
794        md->start_match_ptr = ecode - GET(ecode, 1);
795        md->mark = ecode + LINK_SIZE + 2;
796      RRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
797    
798      /* Handle a capturing bracket. If there is space in the offset vector, save      /* Handle a capturing bracket, other than those that are possessive with an
799      the current subject position in the working slot at the top of the vector.      unlimited repeat. If there is space in the offset vector, save the current
800      We mustn't change the current values of the data slot, because they may be      subject position in the working slot at the top of the vector. We mustn't
801      set from a previous iteration of this group, and be referred to by a      change the current values of the data slot, because they may be set from a
802      reference inside the group.      previous iteration of this group, and be referred to by a reference inside
803        the group. A failure to match might occur after the group has succeeded,
804      If the bracket fails to match, we need to restore this value and also the      if something later on doesn't match. For this reason, we need to restore
805      values of the final offsets, in case they were set by a previous iteration      the working value and also the values of the final offsets, in case they
806      of the same bracket.      were set by a previous iteration of the same bracket.
807    
808      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
809      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 713  for (;;) Line 814  for (;;)
814      number = GET2(ecode, 1+LINK_SIZE);      number = GET2(ecode, 1+LINK_SIZE);
815      offset = number << 1;      offset = number << 1;
816    
817  #ifdef DEBUG  #ifdef PCRE_DEBUG
818      printf("start bracket %d\n", number);      printf("start bracket %d\n", number);
819      printf("subject=");      printf("subject=");
820      pchars(eptr, 16, TRUE, md);      pchars(eptr, 16, TRUE, md);
# Line 728  for (;;) Line 829  for (;;)
829        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
830    
831        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
832        md->offset_vector[md->offset_end - number] = eptr - md->start_subject;        md->offset_vector[md->offset_end - number] =
833            (int)(eptr - md->start_subject);
834    
835        flags = (op == OP_SCBRA)? match_cbegroup : 0;        for (;;)
       do  
836          {          {
837            if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
838          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
839            ims, eptrb, flags, RM1);            eptrb, RM1);
840          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
841            if (rrc != MATCH_NOMATCH &&
842                (rrc != MATCH_THEN || md->start_match_ptr != ecode))
843              RRETURN(rrc);
844          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
845          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
846            if (*ecode != OP_ALT) break;
847          }          }
       while (*ecode == OP_ALT);  
848    
849        DPRINTF(("bracket %d failed\n", number));        DPRINTF(("bracket %d failed\n", number));
   
850        md->offset_vector[offset] = save_offset1;        md->offset_vector[offset] = save_offset1;
851        md->offset_vector[offset+1] = save_offset2;        md->offset_vector[offset+1] = save_offset2;
852        md->offset_vector[md->offset_end - number] = save_offset3;        md->offset_vector[md->offset_end - number] = save_offset3;
853    
854        RRETURN(MATCH_NOMATCH);        /* At this point, rrc will be one of MATCH_ONCE, MATCH_NOMATCH, or
855          MATCH_THEN. */
856    
857          if (rrc != MATCH_THEN && md->mark == NULL) md->mark = markptr;
858          RRETURN(((rrc == MATCH_ONCE)? MATCH_ONCE:MATCH_NOMATCH));
859        }        }
860    
861      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
# Line 761  for (;;) Line 869  for (;;)
869      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
870      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
871    
872      /* Non-capturing bracket. Loop for all the alternatives. When we get to the      /* Non-capturing or atomic group, except for possessive with unlimited
873      final alternative within the brackets, we would return the result of a      repeat. Loop for all the alternatives. When we get to the final alternative
874      recursive call to match() whatever happened. We can reduce stack usage by      within the brackets, we used to return the result of a recursive call to
875      turning this into a tail recursion, except in the case when match_cbegroup      match() whatever happened so it was possible to reduce stack usage by
876      is set.*/      turning this into a tail recursion, except in the case of a possibly empty
877        group. However, now that there is the possiblity of (*THEN) occurring in
878        the final alternative, this optimization is no longer possible.
879    
880        MATCH_ONCE is returned when the end of an atomic group is successfully
881        reached, but subsequent matching fails. It passes back up the tree (causing
882        captured values to be reset) until the original atomic group level is
883        reached. This is tested by comparing md->once_target with the start of the
884        group. At this point, the return is converted into MATCH_NOMATCH so that
885        previous backup points can be taken. */
886    
887        case OP_ONCE:
888      case OP_BRA:      case OP_BRA:
889      case OP_SBRA:      case OP_SBRA:
890      DPRINTF(("start non-capturing bracket\n"));      DPRINTF(("start non-capturing bracket\n"));
891      flags = (op >= OP_SBRA)? match_cbegroup : 0;  
892      for (;;)      for (;;)
893        {        {
894        if (ecode[GET(ecode, 1)] != OP_ALT)   /* Final alternative */        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;
895          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,
896            RM2);
897          if (rrc != MATCH_NOMATCH &&
898              (rrc != MATCH_THEN || md->start_match_ptr != ecode))
899          {          {
900          if (flags == 0)    /* Not a possibly empty group */          if (rrc == MATCH_ONCE)
901            {            {
902            ecode += _pcre_OP_lengths[*ecode];            const uschar *scode = ecode;
903            DPRINTF(("bracket 0 tail recursion\n"));            if (*scode != OP_ONCE)           /* If not at start, find it */
904            goto TAIL_RECURSE;              {
905                while (*scode == OP_ALT) scode += GET(scode, 1);
906                scode -= GET(scode, 1);
907                }
908              if (md->once_target == scode) rrc = MATCH_NOMATCH;
909            }            }
910            RRETURN(rrc);
911            }
912          ecode += GET(ecode, 1);
913          if (*ecode != OP_ALT) break;
914          }
915        if (rrc != MATCH_THEN && md->mark == NULL) md->mark = markptr;
916        RRETURN(MATCH_NOMATCH);
917    
918          /* Possibly empty group; can't use tail recursion. */      /* Handle possessive capturing brackets with an unlimited repeat. We come
919        here from BRAZERO with allow_zero set TRUE. The offset_vector values are
920        handled similarly to the normal case above. However, the matching is
921        different. The end of these brackets will always be OP_KETRPOS, which
922        returns MATCH_KETRPOS without going further in the pattern. By this means
923        we can handle the group by iteration rather than recursion, thereby
924        reducing the amount of stack needed. */
925    
926        case OP_CBRAPOS:
927        case OP_SCBRAPOS:
928        allow_zero = FALSE;
929    
930          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,      POSSESSIVE_CAPTURE:
931            eptrb, flags, RM48);      number = GET2(ecode, 1+LINK_SIZE);
932          RRETURN(rrc);      offset = number << 1;
933    
934    #ifdef PCRE_DEBUG
935        printf("start possessive bracket %d\n", number);
936        printf("subject=");
937        pchars(eptr, 16, TRUE, md);
938        printf("\n");
939    #endif
940    
941        if (offset < md->offset_max)
942          {
943          matched_once = FALSE;
944          code_offset = ecode - md->start_code;
945    
946          save_offset1 = md->offset_vector[offset];
947          save_offset2 = md->offset_vector[offset+1];
948          save_offset3 = md->offset_vector[md->offset_end - number];
949          save_capture_last = md->capture_last;
950    
951          DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
952    
953          /* Each time round the loop, save the current subject position for use
954          when the group matches. For MATCH_MATCH, the group has matched, so we
955          restart it with a new subject starting position, remembering that we had
956          at least one match. For MATCH_NOMATCH, carry on with the alternatives, as
957          usual. If we haven't matched any alternatives in any iteration, check to
958          see if a previous iteration matched. If so, the group has matched;
959          continue from afterwards. Otherwise it has failed; restore the previous
960          capture values before returning NOMATCH. */
961    
962          for (;;)
963            {
964            md->offset_vector[md->offset_end - number] =
965              (int)(eptr - md->start_subject);
966            if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
967            RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
968              eptrb, RM63);
969            if (rrc == MATCH_KETRPOS)
970              {
971              offset_top = md->end_offset_top;
972              eptr = md->end_match_ptr;
973              ecode = md->start_code + code_offset;
974              save_capture_last = md->capture_last;
975              matched_once = TRUE;
976              continue;
977              }
978            if (rrc != MATCH_NOMATCH &&
979                (rrc != MATCH_THEN || md->start_match_ptr != ecode))
980              RRETURN(rrc);
981            md->capture_last = save_capture_last;
982            ecode += GET(ecode, 1);
983            if (*ecode != OP_ALT) break;
984          }          }
985    
986        /* For non-final alternatives, continue the loop for a NOMATCH result;        if (!matched_once)
987        otherwise return. */          {
988            md->offset_vector[offset] = save_offset1;
989            md->offset_vector[offset+1] = save_offset2;
990            md->offset_vector[md->offset_end - number] = save_offset3;
991            }
992    
993        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,        if (rrc != MATCH_THEN && md->mark == NULL) md->mark = markptr;
994          eptrb, flags, RM2);        if (allow_zero || matched_once)
995        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);          {
996            ecode += 1 + LINK_SIZE;
997            break;
998            }
999    
1000          RRETURN(MATCH_NOMATCH);
1001          }
1002    
1003        /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
1004        as a non-capturing bracket. */
1005    
1006        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1007        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1008    
1009        DPRINTF(("insufficient capture room: treat as non-capturing\n"));
1010    
1011        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1012        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1013    
1014        /* Non-capturing possessive bracket with unlimited repeat. We come here
1015        from BRAZERO with allow_zero = TRUE. The code is similar to the above,
1016        without the capturing complication. It is written out separately for speed
1017        and cleanliness. */
1018    
1019        case OP_BRAPOS:
1020        case OP_SBRAPOS:
1021        allow_zero = FALSE;
1022    
1023        POSSESSIVE_NON_CAPTURE:
1024        matched_once = FALSE;
1025        code_offset = ecode - md->start_code;
1026    
1027        for (;;)
1028          {
1029          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1030          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
1031            eptrb, RM48);
1032          if (rrc == MATCH_KETRPOS)
1033            {
1034            offset_top = md->end_offset_top;
1035            eptr = md->end_match_ptr;
1036            ecode = md->start_code + code_offset;
1037            matched_once = TRUE;
1038            continue;
1039            }
1040          if (rrc != MATCH_NOMATCH &&
1041              (rrc != MATCH_THEN || md->start_match_ptr != ecode))
1042            RRETURN(rrc);
1043        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1044          if (*ecode != OP_ALT) break;
1045        }        }
1046    
1047        if (matched_once || allow_zero)
1048          {
1049          ecode += 1 + LINK_SIZE;
1050          break;
1051          }
1052        RRETURN(MATCH_NOMATCH);
1053    
1054      /* Control never reaches here. */      /* Control never reaches here. */
1055    
1056      /* Conditional group: compilation checked that there are no more than      /* Conditional group: compilation checked that there are no more than
1057      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
1058      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
1059      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. */  
1060    
1061      case OP_COND:      case OP_COND:
1062      case OP_SCOND:      case OP_SCOND:
1063      codelink= GET(ecode, 1);      codelink = GET(ecode, 1);
1064    
1065      /* Because of the way auto-callout works during compile, a callout item is      /* Because of the way auto-callout works during compile, a callout item is
1066      inserted between OP_COND and an assertion condition. */      inserted between OP_COND and an assertion condition. */
# Line 821  for (;;) Line 1074  for (;;)
1074          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1075          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1076          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1077          cb.subject_length   = md->end_subject - md->start_subject;          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1078          cb.start_match      = mstart - md->start_subject;          cb.start_match      = (int)(mstart - md->start_subject);
1079          cb.current_position = eptr - md->start_subject;          cb.current_position = (int)(eptr - md->start_subject);
1080          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, LINK_SIZE + 3);
1081          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);
1082          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1083          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last;
1084          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1085          if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);
1086          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1087          }          }
1088        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += _pcre_OP_lengths[OP_CALLOUT];
# Line 843  for (;;) Line 1096  for (;;)
1096        {        {
1097        if (md->recursive == NULL)                /* Not recursing => FALSE */        if (md->recursive == NULL)                /* Not recursing => FALSE */
1098          {          {
1099          condition = FALSE;          condition = FALSE;
1100          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
1101          }          }
1102        else        else
1103          {          {
1104          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1105          condition =  (recno == RREF_ANY || recno == md->recursive->group_num);          condition =  (recno == RREF_ANY || recno == md->recursive->group_num);
1106    
1107          /* 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
1108          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
1109          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
1110          if any one is set. */          if any one is set. */
1111    
1112          if (!condition && condcode == OP_NRREF && recno != RREF_ANY)          if (!condition && condcode == OP_NRREF && recno != RREF_ANY)
1113            {            {
1114            uschar *slotA = md->name_table;            uschar *slotA = md->name_table;
1115            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1116              {              {
1117              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
1118              slotA += md->name_entry_size;              slotA += md->name_entry_size;
1119              }              }
1120    
1121            /* Found a name for the number - there can be only one; duplicate            /* Found a name for the number - there can be only one; duplicate
1122            names for different numbers are allowed, but not vice versa. First            names for different numbers are allowed, but not vice versa. First
1123            scan down for duplicates. */            scan down for duplicates. */
1124    
1125            if (i < md->name_count)            if (i < md->name_count)
1126              {              {
1127              uschar *slotB = slotA;              uschar *slotB = slotA;
1128              while (slotB > md->name_table)              while (slotB > md->name_table)
1129                {                {
# Line 878  for (;;) Line 1131  for (;;)
1131                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)
1132                  {                  {
1133                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1134                  if (condition) break;                  if (condition) break;
1135                  }                  }
1136                else break;                else break;
1137                }                }
1138    
1139              /* Scan up for duplicates */              /* Scan up for duplicates */
1140    
1141              if (!condition)              if (!condition)
1142                {                {
1143                slotB = slotA;                slotB = slotA;
1144                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1145                  {                  {
# Line 895  for (;;) Line 1148  for (;;)
1148                    {                    {
1149                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1150                    if (condition) break;                    if (condition) break;
1151                    }                    }
1152                  else break;                  else break;
1153                  }                  }
1154                }                }
1155              }              }
1156            }            }
1157    
1158          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1159    
1160          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 3 : GET(ecode, 1);
1161          }          }
1162        }        }
1163    
1164      else if (condcode == OP_CREF || condcode == OP_NCREF)  /* Group used test */      else if (condcode == OP_CREF || condcode == OP_NCREF)  /* Group used test */
1165        {        {
1166        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */
1167        condition = offset < offset_top && md->offset_vector[offset] >= 0;        condition = offset < offset_top && md->offset_vector[offset] >= 0;
1168    
1169        /* If the numbered capture is unset, but the reference was by name,        /* If the numbered capture is unset, but the reference was by name,
1170        scan the table to see if the name refers to any other numbers, and test        scan the table to see if the name refers to any other numbers, and test
1171        them. The condition is true if any one is set. This is tediously similar        them. The condition is true if any one is set. This is tediously similar
1172        to the code above, but not close enough to try to amalgamate. */        to the code above, but not close enough to try to amalgamate. */
1173    
1174        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1175          {          {
1176          int refno = offset >> 1;          int refno = offset >> 1;
1177          uschar *slotA = md->name_table;          uschar *slotA = md->name_table;
1178    
1179          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1180            {            {
1181            if (GET2(slotA, 0) == refno) break;            if (GET2(slotA, 0) == refno) break;
1182            slotA += md->name_entry_size;            slotA += md->name_entry_size;
1183            }            }
1184    
1185          /* Found a name for the number - there can be only one; duplicate names          /* Found a name for the number - there can be only one; duplicate names
1186          for different numbers are allowed, but not vice versa. First scan down          for different numbers are allowed, but not vice versa. First scan down
1187          for duplicates. */          for duplicates. */
1188    
1189          if (i < md->name_count)          if (i < md->name_count)
1190            {            {
1191            uschar *slotB = slotA;            uschar *slotB = slotA;
1192            while (slotB > md->name_table)            while (slotB > md->name_table)
1193              {              {
# Line 942  for (;;) Line 1195  for (;;)
1195              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)
1196                {                {
1197                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1198                condition = offset < offset_top &&                condition = offset < offset_top &&
1199                  md->offset_vector[offset] >= 0;                  md->offset_vector[offset] >= 0;
1200                if (condition) break;                if (condition) break;
1201                }                }
1202              else break;              else break;
1203              }              }
1204    
1205            /* Scan up for duplicates */            /* Scan up for duplicates */
1206    
1207            if (!condition)            if (!condition)
1208              {              {
1209              slotB = slotA;              slotB = slotA;
1210              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1211                {                {
# Line 960  for (;;) Line 1213  for (;;)
1213                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)
1214                  {                  {
1215                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1216                  condition = offset < offset_top &&                  condition = offset < offset_top &&
1217                    md->offset_vector[offset] >= 0;                    md->offset_vector[offset] >= 0;
1218                  if (condition) break;                  if (condition) break;
1219                  }                  }
1220                else break;                else break;
1221                }                }
1222              }              }
1223            }            }
1224          }          }
1225    
1226        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1227    
1228        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 3 : GET(ecode, 1);
# Line 982  for (;;) Line 1235  for (;;)
1235        }        }
1236    
1237      /* The condition is an assertion. Call match() to evaluate it - setting      /* The condition is an assertion. Call match() to evaluate it - setting
1238      the final argument match_condassert causes it to stop at the end of an      md->match_function_type to MATCH_CONDASSERT causes it to stop at the end of
1239      assertion. */      an assertion. */
1240    
1241      else      else
1242        {        {
1243        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        md->match_function_type = MATCH_CONDASSERT;
1244            match_condassert, RM3);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);
1245        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
1246          {          {
1247            if (md->end_offset_top > offset_top)
1248              offset_top = md->end_offset_top;  /* Captures may have happened */
1249          condition = TRUE;          condition = TRUE;
1250          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
1251          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
1252          }          }
1253        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)        else if (rrc != MATCH_NOMATCH &&
1254                  (rrc != MATCH_THEN || md->start_match_ptr != ecode))
1255          {          {
1256          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
1257          }          }
# Line 1007  for (;;) Line 1263  for (;;)
1263        }        }
1264    
1265      /* 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,
1266      we can use tail recursion to avoid using another stack frame, except when      we used to use tail recursion to avoid using another stack frame, except
1267      match_cbegroup is required for an unlimited repeat of a possibly empty      when there was unlimited repeat of a possibly empty group. However, that
1268      group. If the second alternative doesn't exist, we can just plough on. */      strategy no longer works because of the possibilty of (*THEN) being
1269        encountered in the branch. A recursive call to match() is always required,
1270        unless the second alternative doesn't exist, in which case we can just
1271        plough on. */
1272    
1273      if (condition || *ecode == OP_ALT)      if (condition || *ecode == OP_ALT)
1274        {        {
1275        ecode += 1 + LINK_SIZE;        if (op == OP_SCOND) md->match_function_type = MATCH_CBEGROUP;
1276        if (op == OP_SCOND)        /* Possibly empty group */        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);
1277          {        if (rrc == MATCH_THEN && md->start_match_ptr == ecode)
1278          RMATCH(eptr, ecode, offset_top, md, ims, eptrb, match_cbegroup, RM49);          rrc = MATCH_NOMATCH;
1279          RRETURN(rrc);        RRETURN(rrc);
         }  
       else                       /* Group must match something */  
         {  
         flags = 0;  
         goto TAIL_RECURSE;  
         }  
1280        }        }
1281      else                         /* Condition false & no alternative */      else                         /* Condition false & no alternative */
1282        {        {
1283        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1284        }        }
1285      break;      break;
1286    
1287    
1288      /* Before OP_ACCEPT there may be any number of OP_CLOSE opcodes,      /* Before OP_ACCEPT there may be any number of OP_CLOSE opcodes,
1289      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1290    
1291      case OP_CLOSE:      case OP_CLOSE:
1292      number = GET2(ecode, 1);      number = GET2(ecode, 1);
1293      offset = number << 1;      offset = number << 1;
1294    
1295  #ifdef DEBUG  #ifdef PCRE_DEBUG
1296        printf("end bracket %d at *ACCEPT", number);        printf("end bracket %d at *ACCEPT", number);
1297        printf("\n");        printf("\n");
1298  #endif  #endif
# Line 1049  for (;;) Line 1302  for (;;)
1302        {        {
1303        md->offset_vector[offset] =        md->offset_vector[offset] =
1304          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
1305        md->offset_vector[offset+1] = eptr - md->start_subject;        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1306        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1307        }        }
1308      ecode += 3;      ecode += 3;
1309      break;      break;
1310    
1311    
1312      /* End of the pattern, either real or forced. If we are in a top-level      /* End of the pattern, either real or forced. */
     recursion, we should restore the offsets appropriately and continue from  
     after the call. */  
1313    
     case OP_ACCEPT:  
1314      case OP_END:      case OP_END:
1315      if (md->recursive != NULL && md->recursive->group_num == 0)      case OP_ACCEPT:
1316        {      case OP_ASSERT_ACCEPT:
       recursion_info *rec = md->recursive;  
       DPRINTF(("End of pattern in a (?0) recursion\n"));  
       md->recursive = rec->prevrec;  
       memmove(md->offset_vector, rec->offset_save,  
         rec->saved_max * sizeof(int));  
       offset_top = rec->offset_top;  
       mstart = rec->save_start;  
       ims = original_ims;  
       ecode = rec->after_call;  
       break;  
       }  
1317    
1318      /* Otherwise, if we have matched an empty string, fail if PCRE_NOTEMPTY is      /* If we have matched an empty string, fail if not in an assertion and not
1319      set, or if PCRE_NOTEMPTY_ATSTART is set and we have matched at the start of      in a recursion if either PCRE_NOTEMPTY is set, or if PCRE_NOTEMPTY_ATSTART
1320      the subject. In both cases, backtracking will then try other alternatives,      is set and we have matched at the start of the subject. In both cases,
1321      if any. */      backtracking will then try other alternatives, if any. */
1322    
1323      if (eptr == mstart &&      if (eptr == mstart && op != OP_ASSERT_ACCEPT &&
1324          (md->notempty ||           md->recursive == NULL &&
1325            (md->notempty_atstart &&           (md->notempty ||
1326              mstart == md->start_subject + md->start_offset)))             (md->notempty_atstart &&
1327        RRETURN(MATCH_NOMATCH);               mstart == md->start_subject + md->start_offset)))
1328          MRRETURN(MATCH_NOMATCH);
1329    
1330      /* Otherwise, we have a match. */      /* Otherwise, we have a match. */
1331    
1332      md->end_match_ptr = eptr;           /* Record where we ended */      md->end_match_ptr = eptr;           /* Record where we ended */
1333      md->end_offset_top = offset_top;    /* and how many extracts were taken */      md->end_offset_top = offset_top;    /* and how many extracts were taken */
1334      md->start_match_ptr = mstart;       /* and the start (\K can modify) */      md->start_match_ptr = mstart;       /* and the start (\K can modify) */
     RRETURN(MATCH_MATCH);  
1335    
1336      /* Change option settings */      /* For some reason, the macros don't work properly if an expression is
1337        given as the argument to MRRETURN when the heap is in use. */
1338    
1339      case OP_OPT:      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;
1340      ims = ecode[1];      MRRETURN(rrc);
     ecode += 2;  
     DPRINTF(("ims set to %02lx\n", ims));  
     break;  
1341    
1342      /* Assertion brackets. Check the alternative branches in turn - the      /* Assertion brackets. Check the alternative branches in turn - the
1343      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,
1344      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
1345      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
1346      this level is identical to the lookahead case. */      this level is identical to the lookahead case. When the assertion is part
1347        of a condition, we want to return immediately afterwards. The caller of
1348        this incarnation of the match() function will have set MATCH_CONDASSERT in
1349        md->match_function type, and one of these opcodes will be the first opcode
1350        that is processed. We use a local variable that is preserved over calls to
1351        match() to remember this case. */
1352    
1353      case OP_ASSERT:      case OP_ASSERT:
1354      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1355        if (md->match_function_type == MATCH_CONDASSERT)
1356          {
1357          condassert = TRUE;
1358          md->match_function_type = 0;
1359          }
1360        else condassert = FALSE;
1361    
1362      do      do
1363        {        {
1364        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
1365          RM4);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1366        if (rrc == MATCH_MATCH) break;          {
1367        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);          mstart = md->start_match_ptr;   /* In case \K reset it */
1368            markptr = md->mark;
1369            break;
1370            }
1371          if (rrc != MATCH_NOMATCH &&
1372              (rrc != MATCH_THEN || md->start_match_ptr != ecode))
1373            RRETURN(rrc);
1374        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1375        }        }
1376      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1377      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);  
1378        if (*ecode == OP_KET) MRRETURN(MATCH_NOMATCH);
1379    
1380      /* If checking an assertion for a condition, return MATCH_MATCH. */      /* If checking an assertion for a condition, return MATCH_MATCH. */
1381    
1382      if ((flags & match_condassert) != 0) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);
1383    
1384      /* Continue from after the assertion, updating the offsets high water      /* Continue from after the assertion, updating the offsets high water
1385      mark, since extracts may have been taken during the assertion. */      mark, since extracts may have been taken during the assertion. */
# Line 1133  for (;;) Line 1389  for (;;)
1389      offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1390      continue;      continue;
1391    
1392      /* Negative assertion: all branches must fail to match */      /* Negative assertion: all branches must fail to match. Encountering SKIP,
1393        PRUNE, or COMMIT means we must assume failure without checking subsequent
1394        branches. */
1395    
1396      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1397      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1398        if (md->match_function_type == MATCH_CONDASSERT)
1399          {
1400          condassert = TRUE;
1401          md->match_function_type = 0;
1402          }
1403        else condassert = FALSE;
1404    
1405      do      do
1406        {        {
1407        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1408          RM5);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) MRRETURN(MATCH_NOMATCH);
1409        if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1410        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);          {
1411            do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1412            break;
1413            }
1414          if (rrc != MATCH_NOMATCH &&
1415              (rrc != MATCH_THEN || md->start_match_ptr != ecode))
1416            RRETURN(rrc);
1417        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1418        }        }
1419      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1420    
1421      if ((flags & match_condassert) != 0) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1422    
1423      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1424      continue;      continue;
# Line 1165  for (;;) Line 1436  for (;;)
1436        while (i-- > 0)        while (i-- > 0)
1437          {          {
1438          eptr--;          eptr--;
1439          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);          if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);
1440          BACKCHAR(eptr);          BACKCHAR(eptr);
1441          }          }
1442        }        }
# Line 1176  for (;;) Line 1447  for (;;)
1447    
1448        {        {
1449        eptr -= GET(ecode, 1);        eptr -= GET(ecode, 1);
1450        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);        if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);
1451        }        }
1452    
1453      /* Save the earliest consulted character, then skip to next op code */      /* Save the earliest consulted character, then skip to next op code */
# Line 1197  for (;;) Line 1468  for (;;)
1468        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1469        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1470        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1471        cb.subject_length   = md->end_subject - md->start_subject;        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1472        cb.start_match      = mstart - md->start_subject;        cb.start_match      = (int)(mstart - md->start_subject);
1473        cb.current_position = eptr - md->start_subject;        cb.current_position = (int)(eptr - md->start_subject);
1474        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1475        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1476        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1477        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last;
1478        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1479        if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);
1480        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1481        }        }
1482      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1215  for (;;) Line 1486  for (;;)
1486      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
1487      whole pattern. (This is so that it works from duplicated subpatterns.)      whole pattern. (This is so that it works from duplicated subpatterns.)
1488    
1489      If there are any capturing brackets started but not finished, we have to      The state of the capturing groups is preserved over recursion, and
1490      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
1491      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
1492      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
1493      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
1494      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
1495      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.  
1496    
1497      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
1498      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
1499      for the original version of this logic. */      for the original version of this logic. It has, however, been hacked around
1500        a lot, so he is not to blame for the current way it works. */
1501    
1502      case OP_RECURSE:      case OP_RECURSE:
1503        {        {
# Line 1241  for (;;) Line 1510  for (;;)
1510        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1511        md->recursive = &new_recursive;        md->recursive = &new_recursive;
1512    
1513        /* Find where to continue from afterwards */        /* Where to continue from afterwards */
1514    
1515        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
       new_recursive.after_call = ecode;  
1516    
1517        /* Now save the offset data. */        /* Now save the offset data */
1518    
1519        new_recursive.saved_max = md->offset_end;        new_recursive.saved_max = md->offset_end;
1520        if (new_recursive.saved_max <= REC_STACK_SAVE_MAX)        if (new_recursive.saved_max <= REC_STACK_SAVE_MAX)
# Line 1257  for (;;) Line 1525  for (;;)
1525            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));
1526          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1527          }          }
   
1528        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1529              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
       new_recursive.save_start = mstart;  
       new_recursive.offset_top = offset_top;  
       mstart = eptr;  
1530    
1531        /* OK, now we can do the recursion. For each top-level alternative we        /* OK, now we can do the recursion. After processing each alternative,
1532        restore the offset and recursion data. */        restore the offset data. If there were nested recursions, md->recursive
1533          might be changed, so reset it before looping. */
1534    
1535        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1536        flags = (*callpat >= OP_SBRA)? match_cbegroup : 0;        cbegroup = (*callpat >= OP_SBRA);
1537        do        do
1538          {          {
1539            if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1540          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,
1541            md, ims, eptrb, flags, RM6);            md, eptrb, RM6);
1542          if (rrc == MATCH_MATCH)          memcpy(md->offset_vector, new_recursive.offset_save,
1543                new_recursive.saved_max * sizeof(int));
1544            if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1545            {            {
1546            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
1547            md->recursive = new_recursive.prevrec;            md->recursive = new_recursive.prevrec;
1548            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1549              (pcre_free)(new_recursive.offset_save);              (pcre_free)(new_recursive.offset_save);
1550            RRETURN(MATCH_MATCH);  
1551              /* Set where we got to in the subject, and reset the start in case
1552              it was changed by \K. This *is* propagated back out of a recursion,
1553              for Perl compatibility. */
1554    
1555              eptr = md->end_match_ptr;
1556              mstart = md->start_match_ptr;
1557              goto RECURSION_MATCHED;        /* Exit loop; end processing */
1558            }            }
1559          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH &&
1560                    (rrc != MATCH_THEN || md->start_match_ptr != ecode))
1561            {            {
1562            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1563            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
# Line 1290  for (;;) Line 1566  for (;;)
1566            }            }
1567    
1568          md->recursive = &new_recursive;          md->recursive = &new_recursive;
         memcpy(md->offset_vector, new_recursive.offset_save,  
             new_recursive.saved_max * sizeof(int));  
1569          callpat += GET(callpat, 1);          callpat += GET(callpat, 1);
1570          }          }
1571        while (*callpat == OP_ALT);        while (*callpat == OP_ALT);
# Line 1300  for (;;) Line 1574  for (;;)
1574        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1575        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1576          (pcre_free)(new_recursive.offset_save);          (pcre_free)(new_recursive.offset_save);
1577        RRETURN(MATCH_NOMATCH);        MRRETURN(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. */  
   
     case OP_ONCE:  
     prev = ecode;  
     saved_eptr = eptr;  
   
     do  
       {  
       RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM7);  
       if (rrc == MATCH_MATCH) break;  
       if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) 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;  
1578        }        }
1579    
1580      /* The repeating kets try the rest of the pattern or restart from the      RECURSION_MATCHED:
1581      preceding bracket, in the appropriate order. The second "call" of match()      break;
     uses tail recursion, to avoid using another stack frame. We need to reset  
     any options that changed within the bracket before re-running it, so  
     check the next opcode. */  
   
     if (ecode[1+LINK_SIZE] == OP_OPT)  
       {  
       ims = (ims & ~PCRE_IMS) | ecode[4];  
       DPRINTF(("ims set to %02lx at group repeat\n", ims));  
       }  
   
     if (*ecode == OP_KETRMIN)  
       {  
       RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM8);  
       if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
       ecode = prev;  
       flags = 0;  
       goto TAIL_RECURSE;  
       }  
     else  /* OP_KETRMAX */  
       {  
       RMATCH(eptr, prev, offset_top, md, ims, eptrb, match_cbegroup, RM9);  
       if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
       ecode += 1 + LINK_SIZE;  
       flags = 0;  
       goto TAIL_RECURSE;  
       }  
     /* Control never gets here */  
1582    
1583      /* 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
1584      bracketed group and go to there. */      bracketed group and go to there. */
# Line 1392  for (;;) Line 1594  for (;;)
1594      optional ones preceded by BRAZERO or BRAMINZERO. */      optional ones preceded by BRAZERO or BRAMINZERO. */
1595    
1596      case OP_BRAZERO:      case OP_BRAZERO:
1597        {      next = ecode + 1;
1598        next = ecode+1;      RMATCH(eptr, next, offset_top, md, eptrb, RM10);
1599        RMATCH(eptr, next, offset_top, md, ims, eptrb, 0, RM10);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1600        if (rrc != MATCH_NOMATCH) RRETURN(rrc);      do next += GET(next, 1); while (*next == OP_ALT);
1601        do next += GET(next,1); while (*next == OP_ALT);      ecode = next + 1 + LINK_SIZE;
       ecode = next + 1 + LINK_SIZE;  
       }  
1602      break;      break;
1603    
1604      case OP_BRAMINZERO:      case OP_BRAMINZERO:
1605        {      next = ecode + 1;
1606        next = ecode+1;      do next += GET(next, 1); while (*next == OP_ALT);
1607        do next += GET(next, 1); while (*next == OP_ALT);      RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, eptrb, RM11);
1608        RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0, RM11);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1609        if (rrc != MATCH_NOMATCH) RRETURN(rrc);      ecode++;
       ecode++;  
       }  
1610      break;      break;
1611    
1612      case OP_SKIPZERO:      case OP_SKIPZERO:
1613        {      next = ecode+1;
1614        next = ecode+1;      do next += GET(next,1); while (*next == OP_ALT);
1615        do next += GET(next,1); while (*next == OP_ALT);      ecode = next + 1 + LINK_SIZE;
       ecode = next + 1 + LINK_SIZE;  
       }  
1616      break;      break;
1617    
1618        /* BRAPOSZERO occurs before a possessive bracket group. Don't do anything
1619        here; just jump to the group, with allow_zero set TRUE. */
1620    
1621        case OP_BRAPOSZERO:
1622        op = *(++ecode);
1623        allow_zero = TRUE;
1624        if (op == OP_CBRAPOS || op == OP_SCBRAPOS) goto POSSESSIVE_CAPTURE;
1625          goto POSSESSIVE_NON_CAPTURE;
1626    
1627      /* End of a group, repeated or non-repeating. */      /* End of a group, repeated or non-repeating. */
1628    
1629      case OP_KET:      case OP_KET:
1630      case OP_KETRMIN:      case OP_KETRMIN:
1631      case OP_KETRMAX:      case OP_KETRMAX:
1632        case OP_KETRPOS:
1633      prev = ecode - GET(ecode, 1);      prev = ecode - GET(ecode, 1);
1634    
1635      /* 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
1636      infinite repeats of empty string matches, retrieve the subject start from      infinite repeats of empty string matches, retrieve the subject start from
1637      the chain. Otherwise, set it NULL. */      the chain. Otherwise, set it NULL. */
1638    
1639      if (*prev >= OP_SBRA)      if (*prev >= OP_SBRA || *prev == OP_ONCE)
1640        {        {
1641        saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */        saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */
1642        eptrb = eptrb->epb_prev;              /* Backup to previous group */        eptrb = eptrb->epb_prev;              /* Backup to previous group */
# Line 1439  for (;;) Line 1645  for (;;)
1645    
1646      /* If we are at the end of an assertion group, stop matching and return      /* If we are at the end of an assertion group, stop matching and return
1647      MATCH_MATCH, but record the current high water mark for use by positive      MATCH_MATCH, but record the current high water mark for use by positive
1648      assertions. Do this also for the "once" (atomic) groups. */      assertions. We also need to record the match start in case it was changed
1649        by \K. */
1650    
1651      if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||      if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||
1652          *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||          *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT)
         *prev == OP_ONCE)  
1653        {        {
1654        md->end_match_ptr = eptr;      /* For ONCE */        md->end_match_ptr = eptr;      /* For ONCE */
1655        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
1656        RRETURN(MATCH_MATCH);        md->start_match_ptr = mstart;
1657          MRRETURN(MATCH_MATCH);         /* Sets md->mark */
1658        }        }
1659    
1660      /* 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
1661      and if necessary complete handling an extraction by setting the offsets and      and if necessary complete handling an extraction by setting the offsets and
1662      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
1663      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
1664      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
1665        the current subject position and start match pointer and give a MATCH
1666        return. */
1667    
1668      if (*prev == OP_CBRA || *prev == OP_SCBRA)      if (*prev == OP_CBRA || *prev == OP_SCBRA ||
1669            *prev == OP_CBRAPOS || *prev == OP_SCBRAPOS)
1670        {        {
1671        number = GET2(prev, 1+LINK_SIZE);        number = GET2(prev, 1+LINK_SIZE);
1672        offset = number << 1;        offset = number << 1;
1673    
1674  #ifdef DEBUG  #ifdef PCRE_DEBUG
1675        printf("end bracket %d", number);        printf("end bracket %d", number);
1676        printf("\n");        printf("\n");
1677  #endif  #endif
1678    
1679          /* Handle a recursively called group. */
1680    
1681          if (md->recursive != NULL && md->recursive->group_num == number)
1682            {
1683            md->end_match_ptr = eptr;
1684            md->start_match_ptr = mstart;
1685            RRETURN(MATCH_MATCH);
1686            }
1687    
1688          /* Deal with capturing */
1689    
1690        md->capture_last = number;        md->capture_last = number;
1691        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->offset_overflow = TRUE; else
1692          {          {
1693            /* If offset is greater than offset_top, it means that we are
1694            "skipping" a capturing group, and that group's offsets must be marked
1695            unset. In earlier versions of PCRE, all the offsets were unset at the
1696            start of matching, but this doesn't work because atomic groups and
1697            assertions can cause a value to be set that should later be unset.
1698            Example: matching /(?>(a))b|(a)c/ against "ac". This sets group 1 as
1699            part of the atomic group, but this is not on the final matching path,
1700            so must be unset when 2 is set. (If there is no group 2, there is no
1701            problem, because offset_top will then be 2, indicating no capture.) */
1702    
1703            if (offset > offset_top)
1704              {
1705              register int *iptr = md->offset_vector + offset_top;
1706              register int *iend = md->offset_vector + offset;
1707              while (iptr < iend) *iptr++ = -1;
1708              }
1709    
1710            /* Now make the extraction */
1711    
1712          md->offset_vector[offset] =          md->offset_vector[offset] =
1713            md->offset_vector[md->offset_end - number];            md->offset_vector[md->offset_end - number];
1714          md->offset_vector[offset+1] = eptr - md->start_subject;          md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1715          if (offset_top <= offset) offset_top = offset + 2;          if (offset_top <= offset) offset_top = offset + 2;
1716          }          }
1717          }
1718    
1719        /* Handle a recursively called group. Restore the offsets      /* For an ordinary non-repeating ket, just continue at this level. This
1720        appropriately and continue from after the call. */      also happens for a repeating ket if no characters were matched in the
1721        group. This is the forcible breaking of infinite loops as implemented in
1722        Perl 5.005. For a non-repeating atomic group, establish a backup point by
1723        processing the rest of the pattern at a lower level. If this results in a
1724        NOMATCH return, pass MATCH_ONCE back to the original OP_ONCE level, thereby
1725        bypassing intermediate backup points, but resetting any captures that
1726        happened along the way. */
1727    
1728        if (md->recursive != NULL && md->recursive->group_num == number)      if (*ecode == OP_KET || eptr == saved_eptr)
1729          {
1730          if (*prev == OP_ONCE)
1731          {          {
1732          recursion_info *rec = md->recursive;          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM12);
1733          DPRINTF(("Recursion (%d) succeeded - continuing\n", number));          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1734          md->recursive = rec->prevrec;          md->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */
1735          mstart = rec->save_start;          RRETURN(MATCH_ONCE);
         memcpy(md->offset_vector, rec->offset_save,  
           rec->saved_max * sizeof(int));  
         offset_top = rec->offset_top;  
         ecode = rec->after_call;  
         ims = original_ims;  
         break;  
1736          }          }
1737          ecode += 1 + LINK_SIZE;    /* Carry on at this level */
1738          break;
1739        }        }
1740    
1741      /* For both capturing and non-capturing groups, reset the value of the ims      /* OP_KETRPOS is a possessive repeating ket. Remember the current position,
1742      flags, in case they got changed during the group. */      and return the MATCH_KETRPOS. This makes it possible to do the repeats one
1743        at a time from the outer level, thus saving stack. */
     ims = original_ims;  
     DPRINTF(("ims reset to %02lx\n", ims));  
   
     /* 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. */  
1744    
1745      if (*ecode == OP_KET || eptr == saved_eptr)      if (*ecode == OP_KETRPOS)
1746        {        {
1747        ecode += 1 + LINK_SIZE;        md->end_match_ptr = eptr;
1748        break;        md->end_offset_top = offset_top;
1749          RRETURN(MATCH_KETRPOS);
1750        }        }
1751    
1752      /* The repeating kets try the rest of the pattern or restart from the      /* The normal repeating kets try the rest of the pattern or restart from
1753      preceding bracket, in the appropriate order. In the second case, we can use      the preceding bracket, in the appropriate order. In the second case, we can
1754      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
1755      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
1756        string. */
     flags = (*prev >= OP_SBRA)? match_cbegroup : 0;  
1757    
1758      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1759        {        {
1760        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM12);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM7);
1761        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1762        if (flags != 0)    /* Could match an empty string */        if (*prev == OP_ONCE)
1763          {          {
1764          RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM8);
1765            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1766            md->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */
1767            RRETURN(MATCH_ONCE);
1768            }
1769          if (*prev >= OP_SBRA)    /* Could match an empty string */
1770            {
1771            md->match_function_type = MATCH_CBEGROUP;
1772            RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
1773          RRETURN(rrc);          RRETURN(rrc);
1774          }          }
1775        ecode = prev;        ecode = prev;
# Line 1532  for (;;) Line 1777  for (;;)
1777        }        }
1778      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
1779        {        {
1780        RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13);        if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1781          RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
1782          if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
1783        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1784          if (*prev == OP_ONCE)
1785            {
1786            RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM9);
1787            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1788            md->once_target = prev;
1789            RRETURN(MATCH_ONCE);
1790            }
1791        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
       flags = 0;  
1792        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1793        }        }
1794      /* Control never gets here */      /* Control never gets here */
1795    
1796      /* Start of subject unless notbol, or after internal newline if multiline */      /* Not multiline mode: start of subject assertion, unless notbol. */
1797    
1798      case OP_CIRC:      case OP_CIRC:
1799      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);
     if ((ims & PCRE_MULTILINE) != 0)  
       {  
       if (eptr != md->start_subject &&  
           (eptr == md->end_subject || !WAS_NEWLINE(eptr)))  
         RRETURN(MATCH_NOMATCH);  
       ecode++;  
       break;  
       }  
     /* ... else fall through */  
1800    
1801      /* Start of subject assertion */      /* Start of subject assertion */
1802    
1803      case OP_SOD:      case OP_SOD:
1804      if (eptr != md->start_subject) RRETURN(MATCH_NOMATCH);      if (eptr != md->start_subject) MRRETURN(MATCH_NOMATCH);
1805        ecode++;
1806        break;
1807    
1808        /* Multiline mode: start of subject unless notbol, or after any newline. */
1809    
1810        case OP_CIRCM:
1811        if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);
1812        if (eptr != md->start_subject &&
1813            (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
1814          MRRETURN(MATCH_NOMATCH);
1815      ecode++;      ecode++;
1816      break;      break;
1817    
1818      /* Start of match assertion */      /* Start of match assertion */
1819    
1820      case OP_SOM:      case OP_SOM:
1821      if (eptr != md->start_subject + md->start_offset) RRETURN(MATCH_NOMATCH);      if (eptr != md->start_subject + md->start_offset) MRRETURN(MATCH_NOMATCH);
1822      ecode++;      ecode++;
1823      break;      break;
1824    
# Line 1575  for (;;) Line 1829  for (;;)
1829      ecode++;      ecode++;
1830      break;      break;
1831    
1832      /* Assert before internal newline if multiline, or before a terminating      /* Multiline mode: assert before any newline, or before end of subject
1833      newline unless endonly is set, else end of subject unless noteol is set. */      unless noteol is set. */
1834    
1835      case OP_DOLL:      case OP_DOLLM:
1836      if ((ims & PCRE_MULTILINE) != 0)      if (eptr < md->end_subject)
1837        {        { if (!IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH); }
       if (eptr < md->end_subject)  
         { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }  
       else  
         { if (md->noteol) RRETURN(MATCH_NOMATCH); }  
       ecode++;  
       break;  
       }  
1838      else      else
1839        {        {
1840        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) MRRETURN(MATCH_NOMATCH);
1841        if (!md->endonly)        SCHECK_PARTIAL();
         {  
         if (eptr != md->end_subject &&  
             (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))  
           RRETURN(MATCH_NOMATCH);  
         ecode++;  
         break;  
         }  
1842        }        }
1843        ecode++;
1844        break;
1845    
1846        /* Not multiline mode: assert before a terminating newline or before end of
1847        subject unless noteol is set. */
1848    
1849        case OP_DOLL:
1850        if (md->noteol) MRRETURN(MATCH_NOMATCH);
1851        if (!md->endonly) goto ASSERT_NL_OR_EOS;
1852    
1853      /* ... else fall through for endonly */      /* ... else fall through for endonly */
1854    
1855      /* End of subject assertion (\z) */      /* End of subject assertion (\z) */
1856    
1857      case OP_EOD:      case OP_EOD:
1858      if (eptr < md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr < md->end_subject) MRRETURN(MATCH_NOMATCH);
1859        SCHECK_PARTIAL();
1860      ecode++;      ecode++;
1861      break;      break;
1862    
1863      /* End of subject or ending \n assertion (\Z) */      /* End of subject or ending \n assertion (\Z) */
1864    
1865      case OP_EODN:      case OP_EODN:
1866      if (eptr != md->end_subject &&      ASSERT_NL_OR_EOS:
1867        if (eptr < md->end_subject &&
1868          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
1869        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
1870    
1871        /* Either at end of string or \n before end. */
1872    
1873        SCHECK_PARTIAL();
1874      ecode++;      ecode++;
1875      break;      break;
1876    
# Line 1632  for (;;) Line 1888  for (;;)
1888  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
1889        if (utf8)        if (utf8)
1890          {          {
1891            /* Get status of previous character */
1892    
1893          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
1894            {            {
1895            USPTR lastptr = eptr - 1;            USPTR lastptr = eptr - 1;
1896            while((*lastptr & 0xc0) == 0x80) lastptr--;            while((*lastptr & 0xc0) == 0x80) lastptr--;
1897            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
1898            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
1899    #ifdef SUPPORT_UCP
1900              if (md->use_ucp)
1901                {
1902                if (c == '_') prev_is_word = TRUE; else
1903                  {
1904                  int cat = UCD_CATEGORY(c);
1905                  prev_is_word = (cat == ucp_L || cat == ucp_N);
1906                  }
1907                }
1908              else
1909    #endif
1910            prev_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0;            prev_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0;
1911            }            }
1912    
1913            /* Get status of next character */
1914    
1915          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
1916            {            {
1917            SCHECK_PARTIAL();            SCHECK_PARTIAL();
# Line 1648  for (;;) Line 1920  for (;;)
1920          else          else
1921            {            {
1922            GETCHAR(c, eptr);            GETCHAR(c, eptr);
1923    #ifdef SUPPORT_UCP
1924              if (md->use_ucp)
1925                {
1926                if (c == '_') cur_is_word = TRUE; else
1927                  {
1928                  int cat = UCD_CATEGORY(c);
1929                  cur_is_word = (cat == ucp_L || cat == ucp_N);
1930                  }
1931                }
1932              else
1933    #endif
1934            cur_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0;            cur_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0;
1935            }            }
1936          }          }
1937        else        else
1938  #endif  #endif
1939    
1940        /* Not in UTF-8 mode */        /* Not in UTF-8 mode, but we may still have PCRE_UCP set, and for
1941          consistency with the behaviour of \w we do use it in this case. */
1942    
1943          {          {
1944            /* Get status of previous character */
1945    
1946          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
1947            {            {
1948            if (eptr <= md->start_used_ptr) md->start_used_ptr = eptr - 1;            if (eptr <= md->start_used_ptr) md->start_used_ptr = eptr - 1;
1949    #ifdef SUPPORT_UCP
1950              if (md->use_ucp)
1951                {
1952                c = eptr[-1];
1953                if (c == '_') prev_is_word = TRUE; else
1954                  {
1955                  int cat = UCD_CATEGORY(c);
1956                  prev_is_word = (cat == ucp_L || cat == ucp_N);
1957                  }
1958                }
1959              else
1960    #endif
1961            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);
1962            }            }
1963    
1964            /* Get status of next character */
1965    
1966          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
1967            {            {
1968            SCHECK_PARTIAL();            SCHECK_PARTIAL();
1969            cur_is_word = FALSE;            cur_is_word = FALSE;
1970            }            }
1971          else cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          else
1972    #ifdef SUPPORT_UCP
1973            if (md->use_ucp)
1974              {
1975              c = *eptr;
1976              if (c == '_') cur_is_word = TRUE; else
1977                {
1978                int cat = UCD_CATEGORY(c);
1979                cur_is_word = (cat == ucp_L || cat == ucp_N);
1980                }
1981              }
1982            else
1983    #endif
1984            cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);
1985          }          }
1986    
1987        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
1988    
1989        if ((*ecode++ == OP_WORD_BOUNDARY)?        if ((*ecode++ == OP_WORD_BOUNDARY)?
1990             cur_is_word == prev_is_word : cur_is_word != prev_is_word)             cur_is_word == prev_is_word : cur_is_word != prev_is_word)
1991          RRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
1992        }        }
1993      break;      break;
1994    
1995      /* Match a single character type; inline for speed */      /* Match a single character type; inline for speed */
1996    
1997      case OP_ANY:      case OP_ANY:
1998      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);
1999      /* Fall through */      /* Fall through */
2000    
2001      case OP_ALLANY:      case OP_ALLANY:
2002      if (eptr++ >= md->end_subject)      if (eptr++ >= md->end_subject)
2003        {        {
2004        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2005        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2006        }        }
2007      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
2008      ecode++;      ecode++;
# Line 1701  for (;;) Line 2015  for (;;)
2015      if (eptr++ >= md->end_subject)      if (eptr++ >= md->end_subject)
2016        {        {
2017        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2018        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2019        }        }
2020      ecode++;      ecode++;
2021      break;      break;
# Line 1710  for (;;) Line 2024  for (;;)
2024      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2025        {        {
2026        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2027        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2028        }        }
2029      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2030      if (      if (
# Line 1719  for (;;) Line 2033  for (;;)
2033  #endif  #endif
2034         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
2035         )         )
2036        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2037      ecode++;      ecode++;
2038      break;      break;
2039    
# Line 1727  for (;;) Line 2041  for (;;)
2041      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2042        {        {
2043        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2044        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2045        }        }
2046      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2047      if (      if (
# Line 1736  for (;;) Line 2050  for (;;)
2050  #endif  #endif
2051         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2052         )         )
2053        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2054      ecode++;      ecode++;
2055      break;      break;
2056    
# Line 1744  for (;;) Line 2058  for (;;)
2058      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2059        {        {
2060        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2061        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2062        }        }
2063      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2064      if (      if (
# Line 1753  for (;;) Line 2067  for (;;)
2067  #endif  #endif
2068         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
2069         )         )
2070        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2071      ecode++;      ecode++;
2072      break;      break;
2073    
# Line 1761  for (;;) Line 2075  for (;;)
2075      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2076        {        {
2077        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2078        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2079        }        }
2080      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2081      if (      if (
# Line 1770  for (;;) Line 2084  for (;;)
2084  #endif  #endif
2085         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2086         )         )
2087        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2088      ecode++;      ecode++;
2089      break;      break;
2090    
# Line 1778  for (;;) Line 2092  for (;;)
2092      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2093        {        {
2094        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2095        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2096        }        }
2097      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2098      if (      if (
# Line 1787  for (;;) Line 2101  for (;;)
2101  #endif  #endif
2102         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
2103         )         )
2104        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2105      ecode++;      ecode++;
2106      break;      break;
2107    
# Line 1795  for (;;) Line 2109  for (;;)
2109      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2110        {        {
2111        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2112        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2113        }        }
2114      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2115      if (      if (
# Line 1804  for (;;) Line 2118  for (;;)
2118  #endif  #endif
2119         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2120         )         )
2121        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2122      ecode++;      ecode++;
2123      break;      break;
2124    
# Line 1812  for (;;) Line 2126  for (;;)
2126      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2127        {        {
2128        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2129        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2130        }        }
2131      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2132      switch(c)      switch(c)
2133        {        {
2134        default: RRETURN(MATCH_NOMATCH);        default: MRRETURN(MATCH_NOMATCH);
2135    
2136        case 0x000d:        case 0x000d:
2137        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
2138        break;        break;
# Line 1830  for (;;) Line 2145  for (;;)
2145        case 0x0085:        case 0x0085:
2146        case 0x2028:        case 0x2028:
2147        case 0x2029:        case 0x2029:
2148        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);
2149        break;        break;
2150        }        }
2151      ecode++;      ecode++;
# Line 1840  for (;;) Line 2155  for (;;)
2155      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2156        {        {
2157        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2158        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2159        }        }
2160      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2161      switch(c)      switch(c)
# Line 1865  for (;;) Line 2180  for (;;)
2180        case 0x202f:    /* NARROW NO-BREAK SPACE */        case 0x202f:    /* NARROW NO-BREAK SPACE */
2181        case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */        case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
2182        case 0x3000:    /* IDEOGRAPHIC SPACE */        case 0x3000:    /* IDEOGRAPHIC SPACE */
2183        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2184        }        }
2185      ecode++;      ecode++;
2186      break;      break;
# Line 1874  for (;;) Line 2189  for (;;)
2189      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2190        {        {
2191        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2192        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2193        }        }
2194      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2195      switch(c)      switch(c)
2196        {        {
2197        default: RRETURN(MATCH_NOMATCH);        default: MRRETURN(MATCH_NOMATCH);
2198        case 0x09:      /* HT */        case 0x09:      /* HT */
2199        case 0x20:      /* SPACE */        case 0x20:      /* SPACE */
2200        case 0xa0:      /* NBSP */        case 0xa0:      /* NBSP */
# Line 1908  for (;;) Line 2223  for (;;)
2223      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2224        {        {
2225        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2226        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2227        }        }
2228      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2229      switch(c)      switch(c)
# Line 1921  for (;;) Line 2236  for (;;)
2236        case 0x85:      /* NEL */        case 0x85:      /* NEL */
2237        case 0x2028:    /* LINE SEPARATOR */        case 0x2028:    /* LINE SEPARATOR */
2238        case 0x2029:    /* PARAGRAPH SEPARATOR */        case 0x2029:    /* PARAGRAPH SEPARATOR */
2239        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2240        }        }
2241      ecode++;      ecode++;
2242      break;      break;
# Line 1930  for (;;) Line 2245  for (;;)
2245      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2246        {        {
2247        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2248        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2249        }        }
2250      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2251      switch(c)      switch(c)
2252        {        {
2253        default: RRETURN(MATCH_NOMATCH);        default: MRRETURN(MATCH_NOMATCH);
2254        case 0x0a:      /* LF */        case 0x0a:      /* LF */
2255        case 0x0b:      /* VT */        case 0x0b:      /* VT */
2256        case 0x0c:      /* FF */        case 0x0c:      /* FF */
# Line 1957  for (;;) Line 2272  for (;;)
2272      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2273        {        {
2274        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2275        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2276        }        }
2277      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2278        {        {
# Line 1966  for (;;) Line 2281  for (;;)
2281        switch(ecode[1])        switch(ecode[1])
2282          {          {
2283          case PT_ANY:          case PT_ANY:
2284          if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);          if (op == OP_NOTPROP) MRRETURN(MATCH_NOMATCH);
2285          break;          break;
2286    
2287          case PT_LAMP:          case PT_LAMP:
2288          if ((prop->chartype == ucp_Lu ||          if ((prop->chartype == ucp_Lu ||
2289               prop->chartype == ucp_Ll ||               prop->chartype == ucp_Ll ||
2290               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))
2291            RRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
2292           break;          break;
2293    
2294          case PT_GC:          case PT_GC:
2295          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))
2296            RRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
2297          break;          break;
2298    
2299          case PT_PC:          case PT_PC:
2300          if ((ecode[2] != prop->chartype) == (op == OP_PROP))          if ((ecode[2] != prop->chartype) == (op == OP_PROP))
2301            RRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
2302          break;          break;
2303    
2304          case PT_SC:          case PT_SC:
2305          if ((ecode[2] != prop->script) == (op == OP_PROP))          if ((ecode[2] != prop->script) == (op == OP_PROP))
2306            RRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
2307            break;
2308    
2309            /* These are specials */
2310    
2311            case PT_ALNUM:
2312            if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||
2313                 _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2314              MRRETURN(MATCH_NOMATCH);
2315          break;          break;
2316    
2317            case PT_SPACE:    /* Perl space */
2318            if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||
2319                 c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2320                   == (op == OP_NOTPROP))
2321              MRRETURN(MATCH_NOMATCH);
2322            break;
2323    
2324            case PT_PXSPACE:  /* POSIX space */
2325            if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||
2326                 c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2327                 c == CHAR_FF || c == CHAR_CR)
2328                   == (op == OP_NOTPROP))
2329              MRRETURN(MATCH_NOMATCH);
2330            break;
2331    
2332            case PT_WORD:
2333            if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||
2334                 _pcre_ucp_gentype[prop->chartype] == ucp_N ||
2335                 c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2336              MRRETURN(MATCH_NOMATCH);
2337            break;
2338    
2339            /* This should never occur */
2340    
2341          default:          default:
2342          RRETURN(PCRE_ERROR_INTERNAL);          RRETURN(PCRE_ERROR_INTERNAL);
2343          }          }
# Line 2006  for (;;) Line 2353  for (;;)
2353      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2354        {        {
2355        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2356        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2357        }        }
2358      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2359        if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH);
2360        while (eptr < md->end_subject)
2361        {        {
2362        int category = UCD_CATEGORY(c);        int len = 1;
2363        if (category == ucp_M) RRETURN(MATCH_NOMATCH);        if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2364        while (eptr < md->end_subject)        if (UCD_CATEGORY(c) != ucp_M) break;
2365          {        eptr += len;
         int len = 1;  
         if (!utf8) c = *eptr; else  
           {  
           GETCHARLEN(c, eptr, len);  
           }  
         category = UCD_CATEGORY(c);  
         if (category != ucp_M) break;  
         eptr += len;  
         }  
2366        }        }
2367      ecode++;      ecode++;
2368      break;      break;
# Line 2038  for (;;) Line 2378  for (;;)
2378      loops). */      loops). */
2379    
2380      case OP_REF:      case OP_REF:
2381        {      case OP_REFI:
2382        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      caseless = op == OP_REFI;
2383        ecode += 3;      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2384        ecode += 3;
2385    
2386        /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2387    
2388        (a) In the default, Perl-compatible state, set the length to be longer      (a) In the default, Perl-compatible state, set the length negative;
2389        than the amount of subject left; this ensures that every attempt at a      this ensures that every attempt at a match fails. We can't just fail
2390        match fails. We can't just fail here, because of the possibility of      here, because of the possibility of quantifiers with zero minima.
       quantifiers with zero minima.  
2391    
2392        (b) If the JavaScript compatibility flag is set, set the length to zero      (b) If the JavaScript compatibility flag is set, set the length to zero
2393        so that the back reference matches an empty string.      so that the back reference matches an empty string.
2394    
2395        Otherwise, set the length to the length of what was matched by the      Otherwise, set the length to the length of what was matched by the
2396        referenced subpattern. */      referenced subpattern. */
2397    
2398        if (offset >= offset_top || md->offset_vector[offset] < 0)      if (offset >= offset_top || md->offset_vector[offset] < 0)
2399          length = (md->jscript_compat)? 0 : md->end_subject - eptr + 1;        length = (md->jscript_compat)? 0 : -1;
2400        else      else
2401          length = md->offset_vector[offset+1] - md->offset_vector[offset];        length = md->offset_vector[offset+1] - md->offset_vector[offset];
2402    
2403        /* Set up for repetition, or handle the non-repeated case */      /* Set up for repetition, or handle the non-repeated case */
2404    
2405        switch (*ecode)      switch (*ecode)
2406          {        {
2407          case OP_CRSTAR:        case OP_CRSTAR:
2408          case OP_CRMINSTAR:        case OP_CRMINSTAR:
2409          case OP_CRPLUS:        case OP_CRPLUS:
2410          case OP_CRMINPLUS:        case OP_CRMINPLUS:
2411          case OP_CRQUERY:        case OP_CRQUERY:
2412          case OP_CRMINQUERY:        case OP_CRMINQUERY:
2413          c = *ecode++ - OP_CRSTAR;        c = *ecode++ - OP_CRSTAR;
2414          minimize = (c & 1) != 0;        minimize = (c & 1) != 0;
2415          min = rep_min[c];                 /* Pick up values from tables; */        min = rep_min[c];                 /* Pick up values from tables; */
2416          max = rep_max[c];                 /* zero for max => infinity */        max = rep_max[c];                 /* zero for max => infinity */
2417          if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2418          break;        break;
2419    
2420          case OP_CRRANGE:        case OP_CRRANGE:
2421          case OP_CRMINRANGE:        case OP_CRMINRANGE:
2422          minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2423          min = GET2(ecode, 1);        min = GET2(ecode, 1);
2424          max = GET2(ecode, 3);        max = GET2(ecode, 3);
2425          if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2426          ecode += 5;        ecode += 5;
2427          break;        break;
2428    
2429          default:               /* No repeat follows */        default:               /* No repeat follows */
2430          if (!match_ref(offset, eptr, length, md, ims))        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2431            {          {
2432            CHECK_PARTIAL();          CHECK_PARTIAL();
2433            RRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
           }  
         eptr += length;  
         continue;              /* With the main loop */  
2434          }          }
2435          eptr += length;
2436          continue;              /* With the main loop */
2437          }
2438    
2439        /* If the length of the reference is zero, just continue with the      /* Handle repeated back references. If the length of the reference is
2440        main loop. */      zero, just continue with the main loop. */
2441    
2442        if (length == 0) continue;      if (length == 0) continue;
2443    
2444        /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2445        the length of the reference string explicitly rather than passing the      the length of the reference string explicitly rather than passing the
2446        address of eptr, so that eptr can be a register variable. */      address of eptr, so that eptr can be a register variable. */
2447    
2448        for (i = 1; i <= min; i++)      for (i = 1; i <= min; i++)
2449          {
2450          int slength;
2451          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2452          {          {
2453          if (!match_ref(offset, eptr, length, md, ims))          CHECK_PARTIAL();
2454            {          MRRETURN(MATCH_NOMATCH);
           CHECK_PARTIAL();  
           RRETURN(MATCH_NOMATCH);  
           }  
         eptr += length;  
2455          }          }
2456          eptr += slength;
2457          }
2458    
2459        /* If min = max, continue at the same level without recursion.      /* If min = max, continue at the same level without recursion.
2460        They are not both allowed to be zero. */      They are not both allowed to be zero. */
2461    
2462        if (min == max) continue;      if (min == max) continue;
2463    
2464        /* If minimizing, keep trying and advancing the pointer */      /* If minimizing, keep trying and advancing the pointer */
2465    
2466        if (minimize)      if (minimize)
2467          {
2468          for (fi = min;; fi++)
2469          {          {
2470          for (fi = min;; fi++)          int slength;
2471            RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);
2472            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2473            if (fi >= max) MRRETURN(MATCH_NOMATCH);
2474            if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2475            {            {
2476            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM14);            CHECK_PARTIAL();
2477            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            MRRETURN(MATCH_NOMATCH);
           if (fi >= max) RRETURN(MATCH_NOMATCH);  
           if (!match_ref(offset, eptr, length, md, ims))  
             {  
             CHECK_PARTIAL();  
             RRETURN(MATCH_NOMATCH);  
             }  
           eptr += length;  
2478            }            }
2479          /* Control never gets here */          eptr += slength;
2480          }          }
2481          /* Control never gets here */
2482          }
2483    
2484        /* If maximizing, find the longest string and work backwards */      /* If maximizing, find the longest string and work backwards */
2485    
2486        else      else
2487          {
2488          pp = eptr;
2489          for (i = min; i < max; i++)
2490          {          {
2491          pp = eptr;          int slength;
2492          for (i = min; i < max; i++)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2493            {            {
2494            if (!match_ref(offset, eptr, length, md, ims)) break;            CHECK_PARTIAL();
2495            eptr += length;            break;
           }  
         while (eptr >= pp)  
           {  
           RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM15);  
           if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
           eptr -= length;  
2496            }            }
2497          RRETURN(MATCH_NOMATCH);          eptr += slength;
2498            }
2499          while (eptr >= pp)
2500            {
2501            RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
2502            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2503            eptr -= length;
2504          }          }
2505          MRRETURN(MATCH_NOMATCH);
2506        }        }
2507      /* Control never gets here */      /* Control never gets here */
2508    
# Line 2217  for (;;) Line 2563  for (;;)
2563            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2564              {              {
2565              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2566              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
2567              }              }
2568            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
2569            if (c > 255)            if (c > 255)
2570              {              {
2571              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);
2572              }              }
2573            else            else
2574              {              {
2575              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);              if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);
2576              }              }
2577            }            }
2578          }          }
# Line 2239  for (;;) Line 2585  for (;;)
2585            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2586              {              {
2587              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2588              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
2589              }              }
2590            c = *eptr++;            c = *eptr++;
2591            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);            if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);
2592            }            }
2593          }          }
2594    
# Line 2262  for (;;) Line 2608  for (;;)
2608            {            {
2609            for (fi = min;; fi++)            for (fi = min;; fi++)
2610              {              {
2611              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM16);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);
2612              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2613              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
2614              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
2615                {                {
2616                SCHECK_PARTIAL();                SCHECK_PARTIAL();
2617                RRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
2618                }                }
2619              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
2620              if (c > 255)              if (c > 255)
2621                {                {
2622                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);
2623                }                }
2624              else              else
2625                {                {
2626                if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);                if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);
2627                }                }
2628              }              }
2629            }            }
# Line 2287  for (;;) Line 2633  for (;;)
2633            {            {
2634            for (fi = min;; fi++)            for (fi = min;; fi++)
2635              {              {
2636              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM17);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);
2637              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2638              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
2639              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
2640                {                {
2641                SCHECK_PARTIAL();                SCHECK_PARTIAL();
2642                RRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
2643                }                }
2644              c = *eptr++;              c = *eptr++;
2645              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);              if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);
2646              }              }
2647            }            }
2648          /* Control never gets here */          /* Control never gets here */
# Line 2315  for (;;) Line 2661  for (;;)
2661            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2662              {              {
2663              int len = 1;              int len = 1;
2664              if (eptr >= md->end_subject) break;              if (eptr >= md->end_subject)
2665                  {
2666                  SCHECK_PARTIAL();
2667                  break;
2668                  }
2669              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
2670              if (c > 255)              if (c > 255)
2671                {                {
# Line 2329  for (;;) Line 2679  for (;;)
2679              }              }
2680            for (;;)            for (;;)
2681              {              {
2682              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM18);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM18);
2683              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2684              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2685              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2341  for (;;) Line 2691  for (;;)
2691            {            {
2692            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2693              {              {
2694              if (eptr >= md->end_subject) break;              if (eptr >= md->end_subject)
2695                  {
2696                  SCHECK_PARTIAL();
2697                  break;
2698                  }
2699              c = *eptr;              c = *eptr;
2700              if ((data[c/8] & (1 << (c&7))) == 0) break;              if ((data[c/8] & (1 << (c&7))) == 0) break;
2701              eptr++;              eptr++;
2702              }              }
2703            while (eptr >= pp)            while (eptr >= pp)
2704              {              {
2705              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM19);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM19);
2706              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2707              eptr--;              eptr--;
2708              }              }
2709            }            }
2710    
2711          RRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
2712          }          }
2713        }        }
2714      /* Control never gets here */      /* Control never gets here */
# Line 2406  for (;;) Line 2760  for (;;)
2760          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
2761            {            {
2762            SCHECK_PARTIAL();            SCHECK_PARTIAL();
2763            RRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
2764            }            }
2765          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
2766          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);          if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);
2767          }          }
2768    
2769        /* 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 2424  for (;;) Line 2778  for (;;)
2778          {          {
2779          for (fi = min;; fi++)          for (fi = min;; fi++)
2780            {            {
2781            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);
2782            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2783            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
2784            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2785              {              {
2786              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2787              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
2788              }              }
2789            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
2790            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);            if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);
2791            }            }
2792          /* Control never gets here */          /* Control never gets here */
2793          }          }
# Line 2446  for (;;) Line 2800  for (;;)
2800          for (i = min; i < max; i++)          for (i = min; i < max; i++)
2801            {            {
2802            int len = 1;            int len = 1;
2803            if (eptr >= md->end_subject) break;            if (eptr >= md->end_subject)
2804                {
2805                SCHECK_PARTIAL();
2806                break;
2807                }
2808            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
2809            if (!_pcre_xclass(c, data)) break;            if (!_pcre_xclass(c, data)) break;
2810            eptr += len;            eptr += len;
2811            }            }
2812          for(;;)          for(;;)
2813            {            {
2814            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
2815            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2816            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
2817            if (utf8) BACKCHAR(eptr);            if (utf8) BACKCHAR(eptr);
2818            }            }
2819          RRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
2820          }          }
2821    
2822        /* Control never gets here */        /* Control never gets here */
# Line 2477  for (;;) Line 2835  for (;;)
2835        if (length > md->end_subject - eptr)        if (length > md->end_subject - eptr)
2836          {          {
2837          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
2838          RRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
2839          }          }
2840        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != *eptr++) MRRETURN(MATCH_NOMATCH);
2841        }        }
2842      else      else
2843  #endif  #endif
# Line 2489  for (;;) Line 2847  for (;;)
2847        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
2848          {          {
2849          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
2850          RRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
2851          }          }
2852        if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH);        if (ecode[1] != *eptr++) MRRETURN(MATCH_NOMATCH);
2853        ecode += 2;        ecode += 2;
2854        }        }
2855      break;      break;
2856    
2857      /* Match a single character, caselessly */      /* Match a single character, caselessly */
2858    
2859      case OP_CHARNC:      case OP_CHARI:
2860  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2861      if (utf8)      if (utf8)
2862        {        {
# Line 2509  for (;;) Line 2867  for (;;)
2867        if (length > md->end_subject - eptr)        if (length > md->end_subject - eptr)
2868          {          {
2869          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
2870          RRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
2871          }          }
2872    
2873        /* 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
# Line 2517  for (;;) Line 2875  for (;;)
2875    
2876        if (fc < 128)        if (fc < 128)
2877          {          {
2878          if (md->lcc[*ecode++] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          if (md->lcc[*ecode++] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);
2879          }          }
2880    
2881        /* Otherwise we must pick up the subject character */        /* Otherwise we must pick up the subject character */
# Line 2536  for (;;) Line 2894  for (;;)
2894  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2895            if (dc != UCD_OTHERCASE(fc))            if (dc != UCD_OTHERCASE(fc))
2896  #endif  #endif
2897              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
2898            }            }
2899          }          }
2900        }        }
# Line 2548  for (;;) Line 2906  for (;;)
2906        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
2907          {          {
2908          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
2909          RRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
2910          }          }
2911        if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);        if (md->lcc[ecode[1]] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);
2912        ecode += 2;        ecode += 2;
2913        }        }
2914      break;      break;
# Line 2558  for (;;) Line 2916  for (;;)
2916      /* Match a single character repeatedly. */      /* Match a single character repeatedly. */
2917    
2918      case OP_EXACT:      case OP_EXACT:
2919        case OP_EXACTI:
2920      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
2921      ecode += 3;      ecode += 3;
2922      goto REPEATCHAR;      goto REPEATCHAR;
2923    
2924      case OP_POSUPTO:      case OP_POSUPTO:
2925        case OP_POSUPTOI:
2926      possessive = TRUE;      possessive = TRUE;
2927      /* Fall through */      /* Fall through */
2928    
2929      case OP_UPTO:      case OP_UPTO:
2930        case OP_UPTOI:
2931      case OP_MINUPTO:      case OP_MINUPTO:
2932        case OP_MINUPTOI:
2933      min = 0;      min = 0;
2934      max = GET2(ecode, 1);      max = GET2(ecode, 1);
2935      minimize = *ecode == OP_MINUPTO;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
2936      ecode += 3;      ecode += 3;
2937      goto REPEATCHAR;      goto REPEATCHAR;
2938    
2939      case OP_POSSTAR:      case OP_POSSTAR:
2940        case OP_POSSTARI:
2941      possessive = TRUE;      possessive = TRUE;
2942      min = 0;      min = 0;
2943      max = INT_MAX;      max = INT_MAX;
# Line 2582  for (;;) Line 2945  for (;;)
2945      goto REPEATCHAR;      goto REPEATCHAR;
2946    
2947      case OP_POSPLUS:      case OP_POSPLUS:
2948        case OP_POSPLUSI:
2949      possessive = TRUE;      possessive = TRUE;
2950      min = 1;      min = 1;
2951      max = INT_MAX;      max = INT_MAX;
# Line 2589  for (;;) Line 2953  for (;;)
2953      goto REPEATCHAR;      goto REPEATCHAR;
2954    
2955      case OP_POSQUERY:      case OP_POSQUERY:
2956        case OP_POSQUERYI:
2957      possessive = TRUE;      possessive = TRUE;
2958      min = 0;      min = 0;
2959      max = 1;      max = 1;
# Line 2596  for (;;) Line 2961  for (;;)
2961      goto REPEATCHAR;      goto REPEATCHAR;
2962    
2963      case OP_STAR:      case OP_STAR:
2964        case OP_STARI:
2965      case OP_MINSTAR:      case OP_MINSTAR:
2966        case OP_MINSTARI:
2967      case OP_PLUS:      case OP_PLUS:
2968        case OP_PLUSI:
2969      case OP_MINPLUS:      case OP_MINPLUS:
2970        case OP_MINPLUSI:
2971      case OP_QUERY:      case OP_QUERY:
2972        case OP_QUERYI:
2973      case OP_MINQUERY:      case OP_MINQUERY:
2974      c = *ecode++ - OP_STAR;      case OP_MINQUERYI:
2975        c = *ecode++ - ((op < OP_STARI)? OP_STAR : OP_STARI);
2976      minimize = (c & 1) != 0;      minimize = (c & 1) != 0;
   
2977      min = rep_min[c];                 /* Pick up values from tables; */      min = rep_min[c];                 /* Pick up values from tables; */
2978      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
2979      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
# Line 2626  for (;;) Line 2996  for (;;)
2996          {          {
2997  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2998          unsigned int othercase;          unsigned int othercase;
2999          if ((ims & PCRE_CASELESS) != 0 &&          if (op >= OP_STARI &&     /* Caseless */
3000              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3001            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = _pcre_ord2utf8(othercase, occhars);
3002          else oclength = 0;          else oclength = 0;
# Line 2644  for (;;) Line 3014  for (;;)
3014            else            else
3015              {              {
3016              CHECK_PARTIAL();              CHECK_PARTIAL();
3017              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
3018              }              }
3019            }            }
3020    
# Line 2654  for (;;) Line 3024  for (;;)
3024            {            {
3025            for (fi = min;; fi++)            for (fi = min;; fi++)
3026              {              {
3027              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM22);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);
3028              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3029              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3030              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3031                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, length) == 0) eptr += length;
3032  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2667  for (;;) Line 3037  for (;;)
3037              else              else
3038                {                {
3039                CHECK_PARTIAL();                CHECK_PARTIAL();
3040                RRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3041                }                }
3042              }              }
3043            /* Control never gets here */            /* Control never gets here */
# Line 2685  for (;;) Line 3055  for (;;)
3055                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3056                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;
3057  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3058              else break;              else
3059                  {
3060                  CHECK_PARTIAL();
3061                  break;
3062                  }
3063              }              }
3064    
3065            if (possessive) continue;            if (possessive) continue;
3066    
3067            for(;;)            for(;;)
3068              {              {
3069              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3070              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3071              if (eptr == pp) { RRETURN(MATCH_NOMATCH); }              if (eptr == pp) { MRRETURN(MATCH_NOMATCH); }
3072  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3073              eptr--;              eptr--;
3074              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2729  for (;;) Line 3103  for (;;)
3103      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3104        max, eptr));        max, eptr));
3105    
3106      if ((ims & PCRE_CASELESS) != 0)      if (op >= OP_STARI)  /* Caseless */
3107        {        {
3108        fc = md->lcc[fc];        fc = md->lcc[fc];
3109        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
# Line 2737  for (;;) Line 3111  for (;;)
3111          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3112            {            {
3113            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3114            RRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
3115            }            }
3116          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);
3117          }          }
3118        if (min == max) continue;        if (min == max) continue;
3119        if (minimize)        if (minimize)
3120          {          {
3121          for (fi = min;; fi++)          for (fi = min;; fi++)
3122            {            {
3123            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3124            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3125            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
3126            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3127              {              {
3128              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3129              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
3130              }              }
3131            if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);
3132            }            }
3133          /* Control never gets here */          /* Control never gets here */
3134          }          }
# Line 2763  for (;;) Line 3137  for (;;)
3137          pp = eptr;          pp = eptr;
3138          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3139            {            {
3140            if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break;            if (eptr >= md->end_subject)
3141                {
3142                SCHECK_PARTIAL();
3143                break;
3144                }
3145              if (fc != md->lcc[*eptr]) break;
3146            eptr++;            eptr++;
3147            }            }
3148    
# Line 2771  for (;;) Line 3150  for (;;)
3150    
3151          while (eptr >= pp)          while (eptr >= pp)
3152            {            {
3153            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM25);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);
3154            eptr--;            eptr--;
3155            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3156            }            }
3157          RRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
3158          }          }
3159        /* Control never gets here */        /* Control never gets here */
3160        }        }
# Line 2789  for (;;) Line 3168  for (;;)
3168          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3169            {            {
3170            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3171            RRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
3172            }            }
3173          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);
3174          }          }
3175    
3176        if (min == max) continue;        if (min == max) continue;
# Line 2800  for (;;) Line 3179  for (;;)
3179          {          {
3180          for (fi = min;; fi++)          for (fi = min;; fi++)
3181            {            {
3182            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM26);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);
3183            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3184            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
3185            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3186              {              {
3187              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3188              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
3189              }              }
3190            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);
3191            }            }
3192          /* Control never gets here */          /* Control never gets here */
3193          }          }
# Line 2817  for (;;) Line 3196  for (;;)
3196          pp = eptr;          pp = eptr;
3197          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3198            {            {
3199            if (eptr >= md->end_subject || fc != *eptr) break;            if (eptr >= md->end_subject)
3200                {
3201                SCHECK_PARTIAL();
3202                break;
3203                }
3204              if (fc != *eptr) break;
3205            eptr++;            eptr++;
3206            }            }
3207          if (possessive) continue;          if (possessive) continue;
3208    
3209          while (eptr >= pp)          while (eptr >= pp)
3210            {            {
3211            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);
3212            eptr--;            eptr--;
3213            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3214            }            }
3215          RRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
3216          }          }
3217        }        }
3218      /* Control never gets here */      /* Control never gets here */
# Line 2837  for (;;) Line 3221  for (;;)
3221      checking can be multibyte. */      checking can be multibyte. */
3222    
3223      case OP_NOT:      case OP_NOT:
3224        case OP_NOTI:
3225      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
3226        {        {
3227        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3228        RRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
3229        }        }
3230      ecode++;      ecode++;
3231      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
3232      if ((ims & PCRE_CASELESS) != 0)      if (op == OP_NOTI)         /* The caseless case */
3233        {        {
3234  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3235        if (c < 256)        if (c < 256)
3236  #endif  #endif
3237        c = md->lcc[c];        c = md->lcc[c];
3238        if (md->lcc[*ecode++] == c) RRETURN(MATCH_NOMATCH);        if (md->lcc[*ecode++] == c) MRRETURN(MATCH_NOMATCH);
3239        }        }
3240      else      else    /* Caseful */
3241        {        {
3242        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        if (*ecode++ == c) MRRETURN(MATCH_NOMATCH);
3243        }        }
3244      break;      break;
3245    
# Line 2866  for (;;) Line 3251  for (;;)
3251      about... */      about... */
3252    
3253      case OP_NOTEXACT:      case OP_NOTEXACT:
3254        case OP_NOTEXACTI:
3255      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3256      ecode += 3;      ecode += 3;
3257      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3258    
3259      case OP_NOTUPTO:      case OP_NOTUPTO:
3260        case OP_NOTUPTOI:
3261      case OP_NOTMINUPTO:      case OP_NOTMINUPTO:
3262        case OP_NOTMINUPTOI:
3263      min = 0;      min = 0;
3264      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3265      minimize = *ecode == OP_NOTMINUPTO;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3266      ecode += 3;      ecode += 3;
3267      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3268    
3269      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
3270        case OP_NOTPOSSTARI:
3271      possessive = TRUE;      possessive = TRUE;
3272      min = 0;      min = 0;
3273      max = INT_MAX;      max = INT_MAX;
# Line 2886  for (;;) Line 3275  for (;;)
3275      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3276    
3277      case OP_NOTPOSPLUS:      case OP_NOTPOSPLUS:
3278        case OP_NOTPOSPLUSI:
3279      possessive = TRUE;      possessive = TRUE;
3280      min = 1;      min = 1;
3281      max = INT_MAX;      max = INT_MAX;
# Line 2893  for (;;) Line 3283  for (;;)
3283      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3284    
3285      case OP_NOTPOSQUERY:      case OP_NOTPOSQUERY:
3286        case OP_NOTPOSQUERYI:
3287      possessive = TRUE;      possessive = TRUE;
3288      min = 0;      min = 0;
3289      max = 1;      max = 1;
# Line 2900  for (;;) Line 3291  for (;;)
3291      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3292    
3293      case OP_NOTPOSUPTO:      case OP_NOTPOSUPTO:
3294        case OP_NOTPOSUPTOI:
3295      possessive = TRUE;      possessive = TRUE;
3296      min = 0;      min = 0;
3297      max = GET2(ecode, 1);      max = GET2(ecode, 1);
# Line 2907  for (;;) Line 3299  for (;;)
3299      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3300    
3301      case OP_NOTSTAR:      case OP_NOTSTAR:
3302        case OP_NOTSTARI:
3303      case OP_NOTMINSTAR:      case OP_NOTMINSTAR:
3304        case OP_NOTMINSTARI:
3305      case OP_NOTPLUS:      case OP_NOTPLUS:
3306        case OP_NOTPLUSI:
3307      case OP_NOTMINPLUS:      case OP_NOTMINPLUS:
3308        case OP_NOTMINPLUSI:
3309      case OP_NOTQUERY:      case OP_NOTQUERY:
3310        case OP_NOTQUERYI:
3311      case OP_NOTMINQUERY:      case OP_NOTMINQUERY:
3312      c = *ecode++ - OP_NOTSTAR;      case OP_NOTMINQUERYI:
3313        c = *ecode++ - ((op >= OP_NOTSTARI)? OP_NOTSTARI: OP_NOTSTAR);
3314      minimize = (c & 1) != 0;      minimize = (c & 1) != 0;
3315      min = rep_min[c];                 /* Pick up values from tables; */      min = rep_min[c];                 /* Pick up values from tables; */
3316      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
# Line 2934  for (;;) Line 3332  for (;;)
3332      DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3333        max, eptr));        max, eptr));
3334    
3335      if ((ims & PCRE_CASELESS) != 0)      if (op >= OP_NOTSTARI)     /* Caseless */
3336        {        {
3337        fc = md->lcc[fc];        fc = md->lcc[fc];
3338    
# Line 2948  for (;;) Line 3346  for (;;)
3346            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3347              {              {
3348              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3349              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
3350              }              }
3351            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3352            if (d < 256) d = md->lcc[d];            if (d < 256) d = md->lcc[d];
3353            if (fc == d) RRETURN(MATCH_NOMATCH);            if (fc == d) MRRETURN(MATCH_NOMATCH);
3354            }            }
3355          }          }
3356        else        else
# Line 2965  for (;;) Line 3363  for (;;)
3363            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3364              {              {
3365              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3366              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
3367              }              }
3368            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);
3369            }            }
3370          }          }
3371    
# Line 2982  for (;;) Line 3380  for (;;)
3380            register unsigned int d;            register unsigned int d;
3381            for (fi = min;; fi++)            for (fi = min;; fi++)
3382              {              {
3383              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
3384              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3385              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3386              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3387                {                {
3388                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3389                RRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3390                }                }
3391              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3392              if (d < 256) d = md->lcc[d];              if (d < 256) d = md->lcc[d];
3393              if (fc == d) RRETURN(MATCH_NOMATCH);              if (fc == d) MRRETURN(MATCH_NOMATCH);
3394              }              }
3395            }            }
3396          else          else
# Line 3001  for (;;) Line 3399  for (;;)
3399            {            {
3400            for (fi = min;; fi++)            for (fi = min;; fi++)
3401              {              {
3402              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM29);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);
3403              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3404              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3405              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3406                {                {
3407                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3408                RRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3409                }                }
3410              if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);              if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);
3411              }              }
3412            }            }
3413          /* Control never gets here */          /* Control never gets here */
# Line 3029  for (;;) Line 3427  for (;;)
3427            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3428              {              {
3429              int len = 1;              int len = 1;
3430              if (eptr >= md->end_subject) break;              if (eptr >= md->end_subject)
3431                  {
3432                  SCHECK_PARTIAL();
3433                  break;
3434                  }
3435              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3436              if (d < 256) d = md->lcc[d];              if (d < 256) d = md->lcc[d];
3437              if (fc == d) break;              if (fc == d) break;
# Line 3038  for (;;) Line 3440  for (;;)
3440          if (possessive) continue;          if (possessive) continue;
3441          for(;;)          for(;;)
3442              {              {
3443              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3444              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3445              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
3446              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3050  for (;;) Line 3452  for (;;)
3452            {            {
3453            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3454              {              {
3455              if (eptr >= md->end_subject || fc == md->lcc[*eptr]) break;              if (eptr >= md->end_subject)
3456                  {
3457                  SCHECK_PARTIAL();
3458                  break;
3459                  }
3460                if (fc == md->lcc[*eptr]) break;
3461              eptr++;              eptr++;
3462              }              }
3463            if (possessive) continue;            if (possessive) continue;
3464            while (eptr >= pp)            while (eptr >= pp)
3465              {              {
3466              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM31);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);
3467              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3468              eptr--;              eptr--;
3469              }              }
3470            }            }
3471    
3472          RRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
3473          }          }
3474        /* Control never gets here */        /* Control never gets here */
3475        }        }
# Line 3081  for (;;) Line 3488  for (;;)
3488            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3489              {              {
3490              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3491              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
3492              }              }
3493            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3494            if (fc == d) RRETURN(MATCH_NOMATCH);            if (fc == d) MRRETURN(MATCH_NOMATCH);
3495            }            }
3496          }          }
3497        else        else
# Line 3096  for (;;) Line 3503  for (;;)
3503            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3504              {              {
3505              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3506              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
3507              }              }
3508            if (fc == *eptr++) RRETURN(MATCH_NOMATCH);            if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);
3509            }            }
3510          }          }
3511    
# Line 3113  for (;;) Line 3520  for (;;)
3520            register unsigned int d;            register unsigned int d;
3521            for (fi = min;; fi++)            for (fi = min;; fi++)
3522              {              {
3523              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
3524              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3525              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3526              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3527                {                {
3528                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3529                RRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3530                }                }
3531              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3532              if (fc == d) RRETURN(MATCH_NOMATCH);              if (fc == d) MRRETURN(MATCH_NOMATCH);
3533              }              }
3534            }            }
3535          else          else
# Line 3131  for (;;) Line 3538  for (;;)
3538            {            {
3539            for (fi = min;; fi++)            for (fi = min;; fi++)
3540              {              {
3541              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM33);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);
3542              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3543              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3544              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3545                {                {
3546                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3547                RRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3548                }                }
3549              if (fc == *eptr++) RRETURN(MATCH_NOMATCH);              if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);
3550              }              }
3551            }            }
3552          /* Control never gets here */          /* Control never gets here */
# Line 3159  for (;;) Line 3566  for (;;)
3566            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3567              {              {
3568              int len = 1;              int len = 1;
3569              if (eptr >= md->end_subject) break;              if (eptr >= md->end_subject)
3570                  {
3571                  SCHECK_PARTIAL();
3572                  break;
3573                  }
3574              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3575              if (fc == d) break;              if (fc == d) break;
3576              eptr += len;              eptr += len;
# Line 3167  for (;;) Line 3578  for (;;)
3578            if (possessive) continue;            if (possessive) continue;
3579            for(;;)            for(;;)
3580              {              {
3581              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM34);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);
3582              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3583              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
3584              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3179  for (;;) Line 3590  for (;;)
3590            {            {
3591            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3592              {              {
3593              if (eptr >= md->end_subject || fc == *eptr) break;              if (eptr >= md->end_subject)
3594                  {
3595                  SCHECK_PARTIAL();
3596                  break;
3597                  }
3598                if (fc == *eptr) break;
3599              eptr++;              eptr++;
3600              }              }
3601            if (possessive) continue;            if (possessive) continue;
3602            while (eptr >= pp)            while (eptr >= pp)
3603              {              {
3604              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM35);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);
3605              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3606              eptr--;              eptr--;
3607              }              }
3608            }            }
3609    
3610          RRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
3611          }          }
3612        }        }
3613      /* Control never gets here */      /* Control never gets here */
# Line 3285  for (;;) Line 3701  for (;;)
3701          switch(prop_type)          switch(prop_type)
3702            {            {
3703            case PT_ANY:            case PT_ANY:
3704            if (prop_fail_result) RRETURN(MATCH_NOMATCH);            if (prop_fail_result) MRRETURN(MATCH_NOMATCH);
3705            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3706              {              {
3707              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3708                {                {
3709                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3710                RRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3711                }                }
3712              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3713              }              }
# Line 3300  for (;;) Line 3716  for (;;)
3716            case PT_LAMP:            case PT_LAMP:
3717            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3718              {              {
3719                int chartype;
3720              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3721                {                {
3722                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3723                RRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3724                }                }
3725              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3726              prop_chartype = UCD_CHARTYPE(c);              chartype = UCD_CHARTYPE(c);
3727              if ((prop_chartype == ucp_Lu ||              if ((chartype == ucp_Lu ||
3728                   prop_chartype == ucp_Ll ||                   chartype == ucp_Ll ||
3729                   prop_chartype == ucp_Lt) == prop_fail_result)                   chartype == ucp_Lt) == prop_fail_result)
3730                RRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3731              }              }
3732            break;            break;
3733    
# Line 3320  for (;;) Line 3737  for (;;)
3737              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3738                {                {
3739                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3740                RRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3741                }                }
3742              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3743              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)
3744              if ((prop_category == prop_value) == prop_fail_result)                MRRETURN(MATCH_NOMATCH);
               RRETURN(MATCH_NOMATCH);  
3745              }              }
3746            break;            break;
3747    
# Line 3335  for (;;) Line 3751  for (;;)
3751              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3752                {                {
3753                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3754                RRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3755                }                }
3756              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3757              prop_chartype = UCD_CHARTYPE(c);              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)
3758              if ((prop_chartype == prop_value) == prop_fail_result)                MRRETURN(MATCH_NOMATCH);
               RRETURN(MATCH_NOMATCH);  
3759              }              }
3760            break;            break;
3761    
# Line 3350  for (;;) Line 3765  for (;;)
3765              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3766                {                {
3767                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3768                RRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3769                  }
3770                GETCHARINCTEST(c, eptr);
3771                if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)
3772                  MRRETURN(MATCH_NOMATCH);
3773                }
3774              break;
3775    
3776              case PT_ALNUM:
3777              for (i = 1; i <= min; i++)
3778                {
3779                int category;
3780                if (eptr >= md->end_subject)
3781                  {
3782                  SCHECK_PARTIAL();
3783                  MRRETURN(MATCH_NOMATCH);
3784                  }
3785                GETCHARINCTEST(c, eptr);
3786                category = UCD_CATEGORY(c);
3787                if ((category == ucp_L || category == ucp_N) == prop_fail_result)
3788                  MRRETURN(MATCH_NOMATCH);
3789                }
3790              break;
3791    
3792              case PT_SPACE:    /* Perl space */
3793              for (i = 1; i <= min; i++)
3794                {
3795                if (eptr >= md->end_subject)
3796                  {
3797                  SCHECK_PARTIAL();
3798                  MRRETURN(MATCH_NOMATCH);
3799                  }
3800                GETCHARINCTEST(c, eptr);
3801                if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
3802                     c == CHAR_FF || c == CHAR_CR)
3803                       == prop_fail_result)
3804                  MRRETURN(MATCH_NOMATCH);
3805                }
3806              break;
3807    
3808              case PT_PXSPACE:  /* POSIX space */
3809              for (i = 1; i <= min; i++)
3810                {
3811                if (eptr >= md->end_subject)
3812                  {
3813                  SCHECK_PARTIAL();
3814                  MRRETURN(MATCH_NOMATCH);
3815                }                }
3816              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3817              prop_script = UCD_SCRIPT(c);              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
3818              if ((prop_script == prop_value) == prop_fail_result)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
3819                RRETURN(MATCH_NOMATCH);                     == prop_fail_result)
3820                  MRRETURN(MATCH_NOMATCH);
3821              }              }
3822            break;            break;
3823    
3824              case PT_WORD:
3825              for (i = 1; i <= min; i++)
3826                {
3827                int category;
3828                if (eptr >= md->end_subject)
3829                  {
3830                  SCHECK_PARTIAL();
3831                  MRRETURN(MATCH_NOMATCH);
3832                  }
3833                GETCHARINCTEST(c, eptr);
3834                category = UCD_CATEGORY(c);
3835                if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)
3836                       == prop_fail_result)
3837                  MRRETURN(MATCH_NOMATCH);
3838                }
3839              break;
3840    
3841              /* This should not occur */
3842    
3843            default:            default:
3844            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
3845            }            }
# Line 3374  for (;;) Line 3855  for (;;)
3855            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3856              {              {
3857              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3858              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
3859              }              }
3860            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3861            prop_category = UCD_CATEGORY(c);            if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH);
           if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);  
3862            while (eptr < md->end_subject)            while (eptr < md->end_subject)
3863              {              {
3864              int len = 1;              int len = 1;
3865              if (!utf8) c = *eptr;              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }
3866                else { GETCHARLEN(c, eptr, len); }              if (UCD_CATEGORY(c) != ucp_M) break;
             prop_category = UCD_CATEGORY(c);  
             if (prop_category != ucp_M) break;  
3867              eptr += len;              eptr += len;
3868              }              }
3869            }            }
# Line 3405  for (;;) Line 3883  for (;;)
3883            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3884              {              {
3885              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3886              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
3887              }              }
3888            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);
3889            eptr++;            eptr++;
3890            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3891            }            }
# Line 3419  for (;;) Line 3897  for (;;)
3897            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3898              {              {
3899              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3900              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
3901              }              }
3902            eptr++;            eptr++;
3903            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
# Line 3427  for (;;) Line 3905  for (;;)
3905          break;          break;
3906    
3907          case OP_ANYBYTE:          case OP_ANYBYTE:
3908          if (eptr > md->end_subject - min) RRETURN(MATCH_NOMATCH);          if (eptr > md->end_subject - min) MRRETURN(MATCH_NOMATCH);
3909          eptr += min;          eptr += min;
3910          break;          break;
3911    
# Line 3437  for (;;) Line 3915  for (;;)
3915            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3916              {              {
3917              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3918              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
3919              }              }
3920            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
3921            switch(c)            switch(c)
3922              {              {
3923              default: RRETURN(MATCH_NOMATCH);              default: MRRETURN(MATCH_NOMATCH);
3924    
3925              case 0x000d:              case 0x000d:
3926              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3927              break;              break;
# Line 3455  for (;;) Line 3934  for (;;)
3934              case 0x0085:              case 0x0085:
3935              case 0x2028:              case 0x2028:
3936              case 0x2029:              case 0x2029:
3937              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);
3938              break;              break;
3939              }              }
3940            }            }
# Line 3467  for (;;) Line 3946  for (;;)
3946            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3947              {              {
3948              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3949              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
3950              }              }
3951            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
3952            switch(c)            switch(c)
# Line 3492  for (;;) Line 3971  for (;;)
3971              case 0x202f:    /* NARROW NO-BREAK SPACE */              case 0x202f:    /* NARROW NO-BREAK SPACE */
3972              case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */              case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3973              case 0x3000:    /* IDEOGRAPHIC SPACE */              case 0x3000:    /* IDEOGRAPHIC SPACE */
3974              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
3975              }              }
3976            }            }
3977          break;          break;
# Line 3503  for (;;) Line 3982  for (;;)
3982            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3983              {              {
3984              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3985              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
3986              }              }
3987            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
3988            switch(c)            switch(c)
3989              {              {
3990              default: RRETURN(MATCH_NOMATCH);              default: MRRETURN(MATCH_NOMATCH);
3991              case 0x09:      /* HT */              case 0x09:      /* HT */
3992              case 0x20:      /* SPACE */              case 0x20:      /* SPACE */
3993              case 0xa0:      /* NBSP */              case 0xa0:      /* NBSP */
# Line 3539  for (;;) Line 4018  for (;;)
4018            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4019              {              {
4020              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4021              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
4022              }              }
4023            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4024            switch(c)            switch(c)
# Line 3552  for (;;) Line 4031  for (;;)
4031              case 0x85:      /* NEL */              case 0x85:      /* NEL */
4032              case 0x2028:    /* LINE SEPARATOR */              case 0x2028:    /* LINE SEPARATOR */
4033              case 0x2029:    /* PARAGRAPH SEPARATOR */              case 0x2029:    /* PARAGRAPH SEPARATOR */
4034              RRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
4035              }              }
4036            }            }