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

Diff of /code/trunk/pcre_exec.c

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

revision 527 by ph10, Sat May 29 15:50:39 2010 UTC revision 609 by ph10, Wed Jun 15 18:09:23 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) 1997-2010 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 73  negative to avoid the external error cod Line 75  negative to avoid the external error cod
75    
76  #define MATCH_ACCEPT       (-999)  #define MATCH_ACCEPT       (-999)
77  #define MATCH_COMMIT       (-998)  #define MATCH_COMMIT       (-998)
78  #define MATCH_PRUNE        (-997)  #define MATCH_KETRPOS      (-997)
79  #define MATCH_SKIP         (-996)  #define MATCH_PRUNE        (-996)
80  #define MATCH_SKIP_ARG     (-995)  #define MATCH_SKIP         (-995)
81  #define MATCH_THEN         (-994)  #define MATCH_SKIP_ARG     (-994)
82    #define MATCH_THEN         (-993)
83    
84  /* This is a convenience macro for code that occurs many times. */  /* This is a convenience macro for code that occurs many times. */
85    
# Line 132  while (length-- > 0) Line 135  while (length-- > 0)
135  *          Match a back-reference                *  *          Match a back-reference                *
136  *************************************************/  *************************************************/
137    
138  /* 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
139  than the number of characters left in the string, so the match fails.  negative, so the match always fails. However, in JavaScript compatibility mode,
140    the length passed is zero. Note that in caseless UTF-8 mode, the number of
141    subject bytes matched may be different to the number of reference bytes.
142    
143  Arguments:  Arguments:
144    offset      index into the offset vector    offset      index into the offset vector
145    eptr        points into the subject    eptr        pointer into the subject
146    length      length to be matched    length      length of reference to be matched (number of bytes)
147    md          points to match data block    md          points to match data block
148    ims         the ims flags    caseless    TRUE if caseless
149    
150  Returns:      TRUE if matched  Returns:      < 0 if not matched, otherwise the number of subject bytes matched
151  */  */
152    
153  static BOOL  static int
154  match_ref(int offset, register USPTR eptr, int length, match_data *md,  match_ref(int offset, register USPTR eptr, int length, match_data *md,
155    unsigned long int ims)    BOOL caseless)
156  {  {
157  USPTR p = md->start_subject + md->offset_vector[offset];  USPTR eptr_start = eptr;
158    register USPTR p = md->start_subject + md->offset_vector[offset];
159    
160  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
161  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 164  pchars(p, length, FALSE, md); Line 170  pchars(p, length, FALSE, md);
170  printf("\n");  printf("\n");
171  #endif  #endif
172    
173  /* Always fail if not enough characters left */  /* Always fail if reference not set (and not JavaScript compatible). */
174    
175  if (length > md->end_subject - eptr) return FALSE;  if (length < 0) return -1;
176    
177  /* 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
178  properly if Unicode properties are supported. Otherwise, we can check only  properly if Unicode properties are supported. Otherwise, we can check only
179  ASCII characters. */  ASCII characters. */
180    
181  if ((ims & PCRE_CASELESS) != 0)  if (caseless)
182    {    {
183  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
184  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
185    if (md->utf8)    if (md->utf8)
186      {      {
187      USPTR endptr = eptr + length;      /* Match characters up to the end of the reference. NOTE: the number of
188      while (eptr < endptr)      bytes matched may differ, because there are some characters whose upper and
189        lower case versions code as different numbers of bytes. For example, U+023A
190        (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);
191        a sequence of 3 of the former uses 6 bytes, as does a sequence of two of
192        the latter. It is important, therefore, to check the length along the
193        reference, not along the subject (earlier code did this wrong). */
194    
195        USPTR endptr = p + length;
196        while (p < endptr)
197        {        {
198        int c, d;        int c, d;
199          if (eptr >= md->end_subject) return -1;
200        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
201        GETCHARINC(d, p);        GETCHARINC(d, p);
202        if (c != d && c != UCD_OTHERCASE(d)) return FALSE;        if (c != d && c != UCD_OTHERCASE(d)) return -1;
203        }        }
204      }      }
205    else    else
# Line 193  if ((ims & PCRE_CASELESS) != 0) Line 208  if ((ims & PCRE_CASELESS) != 0)
208    
209    /* 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
210    is no UCP support. */    is no UCP support. */
211        {
212    while (length-- > 0)      if (eptr + length > md->end_subject) return -1;
213      { if (md->lcc[*p++] != md->lcc[*eptr++]) return FALSE; }      while (length-- > 0)
214          { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }
215        }
216    }    }
217    
218  /* 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
219  are in UTF-8 mode. */  are in UTF-8 mode. */
220    
221  else  else
222    { while (length-- > 0) if (*p++ != *eptr++) return FALSE; }    {
223      if (eptr + length > md->end_subject) return -1;
224      while (length-- > 0) if (*p++ != *eptr++) return -1;
225      }
226    
227  return TRUE;  return eptr - eptr_start;
228  }  }
229    
230    
# Line 256  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 276  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
276         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
277         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
278         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
279         RM61,  RM62 };         RM61,  RM62, RM63};
280    
281  /* 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
282  versions and production versions. Note that the "rw" argument of RMATCH isn't  versions and production versions. Note that the "rw" argument of RMATCH isn't
# Line 266  actually used in this definition. */ Line 286  actually used in this definition. */
286  #define REGISTER register  #define REGISTER register
287    
288  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
289  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
290    { \    { \
291    printf("match() called in line %d\n", __LINE__); \    printf("match() called in line %d\n", __LINE__); \
292    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rf,rg,rdepth+1); \    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1); \
293    printf("to line %d\n", __LINE__); \    printf("to line %d\n", __LINE__); \
294    }    }
295  #define RRETURN(ra) \  #define RRETURN(ra) \
# Line 278  actually used in this definition. */ Line 298  actually used in this definition. */
298    return ra; \    return ra; \
299    }    }
300  #else  #else
301  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
302    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rf,rg,rdepth+1)    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1)
303  #define RRETURN(ra) return ra  #define RRETURN(ra) return ra
304  #endif  #endif
305    
# Line 292  argument of match(), which never changes Line 312  argument of match(), which never changes
312    
313  #define REGISTER  #define REGISTER
314    
315  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
316    {\    {\
317    heapframe *newframe = (pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\
318      if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
319    frame->Xwhere = rw; \    frame->Xwhere = rw; \
320    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
321    newframe->Xecode = rb;\    newframe->Xecode = rb;\
322    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
323    newframe->Xmarkptr = markptr;\    newframe->Xmarkptr = markptr;\
324    newframe->Xoffset_top = rc;\    newframe->Xoffset_top = rc;\
325    newframe->Xims = re;\    newframe->Xeptrb = re;\
   newframe->Xeptrb = rf;\  
   newframe->Xflags = rg;\  
326    newframe->Xrdepth = frame->Xrdepth + 1;\    newframe->Xrdepth = frame->Xrdepth + 1;\
327    newframe->Xprevframe = frame;\    newframe->Xprevframe = frame;\
328    frame = newframe;\    frame = newframe;\
# Line 339  typedef struct heapframe { Line 358  typedef struct heapframe {
358    USPTR Xmstart;    USPTR Xmstart;
359    USPTR Xmarkptr;    USPTR Xmarkptr;
360    int Xoffset_top;    int Xoffset_top;
   long int Xims;  
361    eptrblock *Xeptrb;    eptrblock *Xeptrb;
   int Xflags;  
362    unsigned int Xrdepth;    unsigned int Xrdepth;
363    
364    /* Function local variables */    /* Function local variables */
# Line 362  typedef struct heapframe { Line 379  typedef struct heapframe {
379    BOOL Xcondition;    BOOL Xcondition;
380    BOOL Xprev_is_word;    BOOL Xprev_is_word;
381    
   unsigned long int Xoriginal_ims;  
   
382  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
383    int Xprop_type;    int Xprop_type;
384    int Xprop_value;    int Xprop_value;
# Line 421  immediately. The second one is used when Line 436  immediately. The second one is used when
436  the subject. */  the subject. */
437    
438  #define CHECK_PARTIAL()\  #define CHECK_PARTIAL()\
439    if (md->partial != 0 && eptr >= md->end_subject && eptr > mstart)\    if (md->partial != 0 && eptr >= md->end_subject && \
440      {\        eptr > md->start_used_ptr) \
441      md->hitend = TRUE;\      { \
442      if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL);\      md->hitend = TRUE; \
443        if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \
444      }      }
445    
446  #define SCHECK_PARTIAL()\  #define SCHECK_PARTIAL()\
447    if (md->partial != 0 && eptr > mstart)\    if (md->partial != 0 && eptr > md->start_used_ptr) \
448      {\      { \
449      md->hitend = TRUE;\      md->hitend = TRUE; \
450      if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL);\      if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \
451      }      }
452    
453    
# Line 448  Arguments: Line 464  Arguments:
464     markptr     pointer to the most recent MARK name, or NULL     markptr     pointer to the most recent MARK name, or NULL
465     offset_top  current top pointer     offset_top  current top pointer
466     md          pointer to "static" info for the match     md          pointer to "static" info for the match
    ims         current /i, /m, and /s options  
467     eptrb       pointer to chain of blocks containing eptr at start of     eptrb       pointer to chain of blocks containing eptr at start of
468                   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  
469     rdepth      the recursion depth     rdepth      the recursion depth
470    
471  Returns:       MATCH_MATCH if matched            )  these values are >= 0  Returns:       MATCH_MATCH if matched            )  these values are >= 0
# Line 466  Returns:       MATCH_MATCH if matched Line 477  Returns:       MATCH_MATCH if matched
477    
478  static int  static int
479  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,
480    const uschar *markptr, int offset_top, match_data *md, unsigned long int ims,    const uschar *markptr, int offset_top, match_data *md, eptrblock *eptrb,
481    eptrblock *eptrb, int flags, unsigned int rdepth)    unsigned int rdepth)
482  {  {
483  /* 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,
484  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 479  register unsigned int c;   /* Character Line 490  register unsigned int c;   /* Character
490  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */
491    
492  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
493    BOOL caseless;
494  int condcode;  int condcode;
495    
496  /* 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 487  heap storage. Set up the top-level frame Line 499  heap storage. Set up the top-level frame
499  heap whenever RMATCH() does a "recursion". See the macro definitions above. */  heap whenever RMATCH() does a "recursion". See the macro definitions above. */
500    
501  #ifdef NO_RECURSE  #ifdef NO_RECURSE
502  heapframe *frame = (pcre_stack_malloc)(sizeof(heapframe));  heapframe *frame = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));
503    if (frame == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
504  frame->Xprevframe = NULL;            /* Marks the top level */  frame->Xprevframe = NULL;            /* Marks the top level */
505    
506  /* Copy in the original argument variables */  /* Copy in the original argument variables */
# Line 497  frame->Xecode = ecode; Line 510  frame->Xecode = ecode;
510  frame->Xmstart = mstart;  frame->Xmstart = mstart;
511  frame->Xmarkptr = markptr;  frame->Xmarkptr = markptr;
512  frame->Xoffset_top = offset_top;  frame->Xoffset_top = offset_top;
 frame->Xims = ims;  
513  frame->Xeptrb = eptrb;  frame->Xeptrb = eptrb;
 frame->Xflags = flags;  
514  frame->Xrdepth = rdepth;  frame->Xrdepth = rdepth;
515    
516  /* This is where control jumps back to to effect "recursion" */  /* This is where control jumps back to to effect "recursion" */
# Line 513  HEAP_RECURSE: Line 524  HEAP_RECURSE:
524  #define mstart             frame->Xmstart  #define mstart             frame->Xmstart
525  #define markptr            frame->Xmarkptr  #define markptr            frame->Xmarkptr
526  #define offset_top         frame->Xoffset_top  #define offset_top         frame->Xoffset_top
 #define ims                frame->Xims  
527  #define eptrb              frame->Xeptrb  #define eptrb              frame->Xeptrb
 #define flags              frame->Xflags  
528  #define rdepth             frame->Xrdepth  #define rdepth             frame->Xrdepth
529    
530  /* Ditto for the local variables */  /* Ditto for the local variables */
# Line 537  HEAP_RECURSE: Line 546  HEAP_RECURSE:
546  #define condition          frame->Xcondition  #define condition          frame->Xcondition
547  #define prev_is_word       frame->Xprev_is_word  #define prev_is_word       frame->Xprev_is_word
548    
 #define original_ims       frame->Xoriginal_ims  
   
549  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
550  #define prop_type          frame->Xprop_type  #define prop_type          frame->Xprop_type
551  #define prop_value         frame->Xprop_value  #define prop_value         frame->Xprop_value
# Line 575  i, and fc and c, can be the same variabl Line 582  i, and fc and c, can be the same variabl
582  #define fi i  #define fi i
583  #define fc c  #define fc c
584    
585    /* Many of the following variables are used only in small blocks of the code.
586  #ifdef SUPPORT_UTF8                /* Many of these variables are used only  */  My normal style of coding would have declared them within each of those blocks.
587  const uschar *charptr;             /* in small blocks of the code. My normal */  However, in order to accommodate the version of this code that uses an external
588  #endif                             /* style of coding would have declared    */  "stack" implemented on the heap, it is easier to declare them all here, so the
589  const uschar *callpat;             /* them within each of those blocks.      */  declarations can be cut out in a block. The only declarations within blocks
590  const uschar *data;                /* However, in order to accommodate the   */  below are for variables that do not have to be preserved over a recursive call
591  const uschar *next;                /* version of this code that uses an      */  to RMATCH(). */
592  USPTR         pp;                  /* external "stack" implemented on the    */  
593  const uschar *prev;                /* heap, it is easier to declare them all */  #ifdef SUPPORT_UTF8
594  USPTR         saved_eptr;          /* here, so the declarations can be cut   */  const uschar *charptr;
595                                     /* out in a block. The only declarations  */  #endif
596  recursion_info new_recursive;      /* within blocks below are for variables  */  const uschar *callpat;
597                                     /* that do not have to be preserved over  */  const uschar *data;
598  BOOL cur_is_word;                  /* a recursive call to RMATCH().          */  const uschar *next;
599    USPTR         pp;
600    const uschar *prev;
601    USPTR         saved_eptr;
602    
603    recursion_info new_recursive;
604    
605    BOOL cur_is_word;
606  BOOL condition;  BOOL condition;
607  BOOL prev_is_word;  BOOL prev_is_word;
608    
 unsigned long int original_ims;  
   
609  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
610  int prop_type;  int prop_type;
611  int prop_value;  int prop_value;
# Line 620  int stacksave[REC_STACK_SAVE_MAX]; Line 632  int stacksave[REC_STACK_SAVE_MAX];
632  eptrblock newptrb;  eptrblock newptrb;
633  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
634    
635    /* To save space on the stack and in the heap frame, I have doubled up on some
636    of the local variables that are used only in localised parts of the code, but
637    still need to be preserved over recursive calls of match(). These macros define
638    the alternative names that are used. */
639    
640    #define allow_zero    cur_is_word
641    #define cbegroup      condition
642    #define code_offset   codelink
643    #define condassert    condition
644    #define matched_once  prev_is_word
645    
646  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
647  variables. */  variables. */
648    
# Line 656  haven't exceeded the recursive call limi Line 679  haven't exceeded the recursive call limi
679  if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT);  if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT);
680  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);
681    
 original_ims = ims;    /* Save for resetting on ')' */  
   
682  /* 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
683  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
684  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
685  hit the closing ket, in order to break infinite loops that match no characters.  up space on the stack. See also MATCH_CONDASSERT below.
686  When match() is called in other circumstances, don't add to the chain. The  
687  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
688  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
689  match(). */  to break infinite loops that match no characters. When match() is called in
690    other circumstances, don't add to the chain. The MATCH_CBEGROUP feature must
691    NOT be used with tail recursion, because the memory block that is used is on
692    the stack, so a new one may be required for each match(). */
693    
694  if ((flags & match_cbegroup) != 0)  if (md->match_function_type == MATCH_CBEGROUP)
695    {    {
696    newptrb.epb_saved_eptr = eptr;    newptrb.epb_saved_eptr = eptr;
697    newptrb.epb_prev = eptrb;    newptrb.epb_prev = eptrb;
698    eptrb = &newptrb;    eptrb = &newptrb;
699      md->match_function_type = 0;
700    }    }
701    
702  /* Now start processing the opcodes. */  /* Now start processing the opcodes. */
# Line 680  for (;;) Line 705  for (;;)
705    {    {
706    minimize = possessive = FALSE;    minimize = possessive = FALSE;
707    op = *ecode;    op = *ecode;
708    
709    switch(op)    switch(op)
710      {      {
711      case OP_MARK:      case OP_MARK:
712      markptr = ecode + 2;      markptr = ecode + 2;
713      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,
714        ims, eptrb, flags, RM55);        eptrb, RM55);
715    
716      /* A return of MATCH_SKIP_ARG means that matching failed at SKIP with an      /* A return of MATCH_SKIP_ARG means that matching failed at SKIP with an
717      argument, and we must check whether that argument matches this MARK's      argument, and we must check whether that argument matches this MARK's
# Line 708  for (;;) Line 733  for (;;)
733      case OP_FAIL:      case OP_FAIL:
734      MRRETURN(MATCH_NOMATCH);      MRRETURN(MATCH_NOMATCH);
735    
736        /* COMMIT overrides PRUNE, SKIP, and THEN */
737    
738      case OP_COMMIT:      case OP_COMMIT:
739      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
740        ims, eptrb, flags, RM52);        eptrb, RM52);
741      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&
742            rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&
743            rrc != MATCH_THEN)
744          RRETURN(rrc);
745      MRRETURN(MATCH_COMMIT);      MRRETURN(MATCH_COMMIT);
746    
747        /* PRUNE overrides THEN */
748    
749      case OP_PRUNE:      case OP_PRUNE:
750      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
751        ims, eptrb, flags, RM51);        eptrb, RM51);
752      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
753      MRRETURN(MATCH_PRUNE);      MRRETURN(MATCH_PRUNE);
754    
755      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
756      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,
757        ims, eptrb, flags, RM56);        eptrb, RM56);
758      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
759      md->mark = ecode + 2;      md->mark = ecode + 2;
760      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
761    
762        /* SKIP overrides PRUNE and THEN */
763    
764      case OP_SKIP:      case OP_SKIP:
765      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
766        ims, eptrb, flags, RM53);        eptrb, RM53);
767      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
768          RRETURN(rrc);
769      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
770      MRRETURN(MATCH_SKIP);      MRRETURN(MATCH_SKIP);
771    
772      case OP_SKIP_ARG:      case OP_SKIP_ARG:
773      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,
774        ims, eptrb, flags, RM57);        eptrb, RM57);
775      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
776          RRETURN(rrc);
777    
778      /* Pass back the current skip name by overloading md->start_match_ptr and      /* Pass back the current skip name by overloading md->start_match_ptr and
779      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
# Line 747  for (;;) Line 783  for (;;)
783      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
784      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
785    
786        /* For THEN (and THEN_ARG) we pass back the address of the bracket or
787        the alt that is at the start of the current branch. This makes it possible
788        to skip back past alternatives that precede the THEN within the current
789        branch. */
790    
791      case OP_THEN:      case OP_THEN:
792      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
793        ims, eptrb, flags, RM54);        eptrb, RM54);
794      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
795        md->start_match_ptr = ecode - GET(ecode, 1);
796      MRRETURN(MATCH_THEN);      MRRETURN(MATCH_THEN);
797    
798      case OP_THEN_ARG:      case OP_THEN_ARG:
799      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1+LINK_SIZE],
800        ims, eptrb, flags, RM58);        offset_top, md, eptrb, RM58);
801      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
802      md->mark = ecode + 2;      md->start_match_ptr = ecode - GET(ecode, 1);
803        md->mark = ecode + LINK_SIZE + 2;
804      RRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
805    
806      /* 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
807      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
808      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
809      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
810      reference inside the group.      previous iteration of this group, and be referred to by a reference inside
811        the group. If we fail to match, we need to restore this value and also the
     If the bracket fails to match, we need to restore this value and also the  
812      values of the final offsets, in case they were set by a previous iteration      values of the final offsets, in case they were set by a previous iteration
813      of the same bracket.      of the same bracket.
814    
# Line 778  for (;;) Line 820  for (;;)
820      case OP_SCBRA:      case OP_SCBRA:
821      number = GET2(ecode, 1+LINK_SIZE);      number = GET2(ecode, 1+LINK_SIZE);
822      offset = number << 1;      offset = number << 1;
823    
824  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
825      printf("start bracket %d\n", number);      printf("start bracket %d\n", number);
826      printf("subject=");      printf("subject=");
# Line 794  for (;;) Line 836  for (;;)
836        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
837    
838        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
839        md->offset_vector[md->offset_end - number] = eptr - md->start_subject;        md->offset_vector[md->offset_end - number] =
840            (int)(eptr - md->start_subject);
841    
842        flags = (op == OP_SCBRA)? match_cbegroup : 0;        for (;;)
       do  
843          {          {
844          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
845            ims, eptrb, flags, RM1);          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
846          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);            eptrb, RM1);
847            if (rrc != MATCH_NOMATCH &&
848                (rrc != MATCH_THEN || md->start_match_ptr != ecode))
849              RRETURN(rrc);
850          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
851          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
852            if (*ecode != OP_ALT) break;
853          }          }
       while (*ecode == OP_ALT);  
854    
855        DPRINTF(("bracket %d failed\n", number));        DPRINTF(("bracket %d failed\n", number));
856    
# Line 813  for (;;) Line 858  for (;;)
858        md->offset_vector[offset+1] = save_offset2;        md->offset_vector[offset+1] = save_offset2;
859        md->offset_vector[md->offset_end - number] = save_offset3;        md->offset_vector[md->offset_end - number] = save_offset3;
860    
861        if (rrc != MATCH_THEN) md->mark = markptr;        if (rrc != MATCH_THEN && md->mark == NULL) md->mark = markptr;
862        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
863        }        }
864    
# Line 828  for (;;) Line 873  for (;;)
873      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
874      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
875    
876      /* Non-capturing bracket. Loop for all the alternatives. When we get to the      /* Non-capturing bracket, except for possessive with unlimited repeat. Loop
877      final alternative within the brackets, we would return the result of a      for all the alternatives. When we get to the final alternative within the
878      recursive call to match() whatever happened. We can reduce stack usage by      brackets, we used to return the result of a recursive call to match()
879      turning this into a tail recursion, except in the case when match_cbegroup      whatever happened so it was possible to reduce stack usage by turning this
880      is set.*/      into a tail recursion, except in the case of a possibly empty group.
881        However, now that there is the possiblity of (*THEN) occurring in the final
882        alternative, this optimization is no longer possible. */
883    
884      case OP_BRA:      case OP_BRA:
885      case OP_SBRA:      case OP_SBRA:
886      DPRINTF(("start non-capturing bracket\n"));      DPRINTF(("start non-capturing bracket\n"));
     flags = (op >= OP_SBRA)? match_cbegroup : 0;  
887      for (;;)      for (;;)
888        {        {
889        if (ecode[GET(ecode, 1)] != OP_ALT)   /* Final alternative */        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
890          {        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,
891          if (flags == 0)    /* Not a possibly empty group */          RM2);
892            {        if (rrc != MATCH_NOMATCH &&
893            ecode += _pcre_OP_lengths[*ecode];            (rrc != MATCH_THEN || md->start_match_ptr != ecode))
894            DPRINTF(("bracket 0 tail recursion\n"));          RRETURN(rrc);
895            goto TAIL_RECURSE;        ecode += GET(ecode, 1);
896            }        if (*ecode != OP_ALT) break;
897          }
898    
899          /* Possibly empty group; can't use tail recursion. */      if (rrc != MATCH_THEN && md->mark == NULL) md->mark = markptr;
900        RRETURN(MATCH_NOMATCH);
901    
902          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,      /* Handle possessive capturing brackets with an unlimited repeat. We come
903            eptrb, flags, RM48);      here from BRAZERO with allow_zero set TRUE. The offset_vector values are
904          if (rrc == MATCH_NOMATCH) md->mark = markptr;      handled similarly to the normal case above. However, the matching is
905          RRETURN(rrc);      different. The end of these brackets will always be OP_KETRPOS, which
906        returns MATCH_KETRPOS without going further in the pattern. By this means
907        we can handle the group by iteration rather than recursion, thereby
908        reducing the amount of stack needed. */
909    
910        case OP_CBRAPOS:
911        case OP_SCBRAPOS:
912        allow_zero = FALSE;
913    
914        POSSESSIVE_CAPTURE:
915        number = GET2(ecode, 1+LINK_SIZE);
916        offset = number << 1;
917    
918    #ifdef PCRE_DEBUG
919        printf("start possessive bracket %d\n", number);
920        printf("subject=");
921        pchars(eptr, 16, TRUE, md);
922        printf("\n");
923    #endif
924    
925        if (offset < md->offset_max)
926          {
927          matched_once = FALSE;
928          code_offset = ecode - md->start_code;
929    
930          save_offset1 = md->offset_vector[offset];
931          save_offset2 = md->offset_vector[offset+1];
932          save_offset3 = md->offset_vector[md->offset_end - number];
933          save_capture_last = md->capture_last;
934    
935          DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
936    
937          /* Each time round the loop, save the current subject position for use
938          when the group matches. For MATCH_MATCH, the group has matched, so we
939          restart it with a new subject starting position, remembering that we had
940          at least one match. For MATCH_NOMATCH, carry on with the alternatives, as
941          usual. If we haven't matched any alternatives in any iteration, check to
942          see if a previous iteration matched. If so, the group has matched;
943          continue from afterwards. Otherwise it has failed; restore the previous
944          capture values before returning NOMATCH. */
945    
946          for (;;)
947            {
948            md->offset_vector[md->offset_end - number] =
949              (int)(eptr - md->start_subject);
950            if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
951            RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
952              eptrb, RM63);
953            if (rrc == MATCH_KETRPOS)
954              {
955              offset_top = md->end_offset_top;
956              eptr = md->end_match_ptr;
957              ecode = md->start_code + code_offset;
958              save_capture_last = md->capture_last;
959              matched_once = TRUE;
960              continue;
961              }
962            if (rrc != MATCH_NOMATCH &&
963                (rrc != MATCH_THEN || md->start_match_ptr != ecode))
964              RRETURN(rrc);
965            md->capture_last = save_capture_last;
966            ecode += GET(ecode, 1);
967            if (*ecode != OP_ALT) break;
968          }          }
969    
970          if (!matched_once)
971            {
972            md->offset_vector[offset] = save_offset1;
973            md->offset_vector[offset+1] = save_offset2;
974            md->offset_vector[md->offset_end - number] = save_offset3;
975            }
976    
977          if (rrc != MATCH_THEN && md->mark == NULL) md->mark = markptr;
978          if (allow_zero || matched_once)
979            {
980            ecode += 1 + LINK_SIZE;
981            break;
982            }
983    
984          RRETURN(MATCH_NOMATCH);
985          }
986    
987        /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
988        as a non-capturing bracket. */
989    
990        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
991        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
992    
993        DPRINTF(("insufficient capture room: treat as non-capturing\n"));
994    
995        /* For non-final alternatives, continue the loop for a NOMATCH result;      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
996        otherwise return. */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
997    
998        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,      /* Non-capturing possessive bracket with unlimited repeat. We come here
999          eptrb, flags, RM2);      from BRAZERO with allow_zero = TRUE. The code is similar to the above,
1000        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      without the capturing complication. It is written out separately for speed
1001        and cleanliness. */
1002    
1003        case OP_BRAPOS:
1004        case OP_SBRAPOS:
1005        allow_zero = FALSE;
1006    
1007        POSSESSIVE_NON_CAPTURE:
1008        matched_once = FALSE;
1009        code_offset = ecode - md->start_code;
1010    
1011        for (;;)
1012          {
1013          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1014          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
1015            eptrb, RM48);
1016          if (rrc == MATCH_KETRPOS)
1017            {
1018            eptr = md->end_match_ptr;
1019            ecode = md->start_code + code_offset;
1020            matched_once = TRUE;
1021            continue;
1022            }
1023          if (rrc != MATCH_NOMATCH &&
1024              (rrc != MATCH_THEN || md->start_match_ptr != ecode))
1025            RRETURN(rrc);
1026        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1027          if (*ecode != OP_ALT) break;
1028        }        }
1029    
1030        if (matched_once || allow_zero)
1031          {
1032          ecode += 1 + LINK_SIZE;
1033          break;
1034          }
1035        RRETURN(MATCH_NOMATCH);
1036    
1037      /* Control never reaches here. */      /* Control never reaches here. */
1038    
1039      /* Conditional group: compilation checked that there are no more than      /* Conditional group: compilation checked that there are no more than
1040      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
1041      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
1042      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. */  
1043    
1044      case OP_COND:      case OP_COND:
1045      case OP_SCOND:      case OP_SCOND:
1046      codelink= GET(ecode, 1);      codelink = GET(ecode, 1);
1047    
1048      /* 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
1049      inserted between OP_COND and an assertion condition. */      inserted between OP_COND and an assertion condition. */
# Line 889  for (;;) Line 1057  for (;;)
1057          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1058          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1059          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1060          cb.subject_length   = md->end_subject - md->start_subject;          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1061          cb.start_match      = mstart - md->start_subject;          cb.start_match      = (int)(mstart - md->start_subject);
1062          cb.current_position = eptr - md->start_subject;          cb.current_position = (int)(eptr - md->start_subject);
1063          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, LINK_SIZE + 3);
1064          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);
1065          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
# Line 1050  for (;;) Line 1218  for (;;)
1218        }        }
1219    
1220      /* The condition is an assertion. Call match() to evaluate it - setting      /* The condition is an assertion. Call match() to evaluate it - setting
1221      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
1222      assertion. */      an assertion. */
1223    
1224      else      else
1225        {        {
1226        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        md->match_function_type = MATCH_CONDASSERT;
1227            match_condassert, RM3);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);
1228        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
1229          {          {
1230          condition = TRUE;          condition = TRUE;
1231          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
1232          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
1233          }          }
1234        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)        else if (rrc != MATCH_NOMATCH &&
1235                  (rrc != MATCH_THEN || md->start_match_ptr != ecode))
1236          {          {
1237          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
1238          }          }
# Line 1075  for (;;) Line 1244  for (;;)
1244        }        }
1245    
1246      /* 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,
1247      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
1248      match_cbegroup is required for an unlimited repeat of a possibly empty      when there was unlimited repeat of a possibly empty group. However, that
1249      group. If the second alternative doesn't exist, we can just plough on. */      strategy no longer works because of the possibilty of (*THEN) being
1250        encountered in the branch. A recursive call to match() is always required,
1251        unless the second alternative doesn't exist, in which case we can just
1252        plough on. */
1253    
1254      if (condition || *ecode == OP_ALT)      if (condition || *ecode == OP_ALT)
1255        {        {
1256        ecode += 1 + LINK_SIZE;        if (op == OP_SCOND) md->match_function_type = MATCH_CBEGROUP;
1257        if (op == OP_SCOND)        /* Possibly empty group */        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);
1258          {        if (rrc == MATCH_THEN && md->start_match_ptr == ecode)
1259          RMATCH(eptr, ecode, offset_top, md, ims, eptrb, match_cbegroup, RM49);          rrc = MATCH_NOMATCH;
1260          RRETURN(rrc);        RRETURN(rrc);
         }  
       else                       /* Group must match something */  
         {  
         flags = 0;  
         goto TAIL_RECURSE;  
         }  
1261        }        }
1262      else                         /* Condition false & no alternative */      else                         /* Condition false & no alternative */
1263        {        {
# Line 1117  for (;;) Line 1283  for (;;)
1283        {        {
1284        md->offset_vector[offset] =        md->offset_vector[offset] =
1285          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
1286        md->offset_vector[offset+1] = eptr - md->start_subject;        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1287        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1288        }        }
1289      ecode += 3;      ecode += 3;
1290      break;      break;
1291    
1292    
1293      /* End of the pattern, either real or forced. If we are in a top-level      /* End of the pattern, either real or forced. If we are in a recursion, we
1294      recursion, we should restore the offsets appropriately and continue from      should restore the offsets appropriately, and if it's a top-level
1295      after the call. */      recursion, continue from after the call. */
1296    
1297      case OP_ACCEPT:      case OP_ACCEPT:
1298      case OP_END:      case OP_END:
1299      if (md->recursive != NULL && md->recursive->group_num == 0)      if (md->recursive != NULL)
1300        {        {
1301        recursion_info *rec = md->recursive;        recursion_info *rec = md->recursive;
       DPRINTF(("End of pattern in a (?0) recursion\n"));  
1302        md->recursive = rec->prevrec;        md->recursive = rec->prevrec;
1303        memmove(md->offset_vector, rec->offset_save,        memmove(md->offset_vector, rec->offset_save,
1304          rec->saved_max * sizeof(int));          rec->saved_max * sizeof(int));
1305        offset_top = rec->save_offset_top;        offset_top = rec->save_offset_top;
1306        ims = original_ims;        if (rec->group_num == 0)
1307        ecode = rec->after_call;          {
1308        break;          ecode = rec->after_call;
1309            break;
1310            }
1311        }        }
1312    
1313      /* Otherwise, if we have matched an empty string, fail if PCRE_NOTEMPTY is      /* Otherwise, if we have matched an empty string, fail if PCRE_NOTEMPTY is
# Line 1148  for (;;) Line 1315  for (;;)
1315      the subject. In both cases, backtracking will then try other alternatives,      the subject. In both cases, backtracking will then try other alternatives,
1316      if any. */      if any. */
1317    
1318      if (eptr == mstart &&      else if (eptr == mstart &&
1319          (md->notempty ||          (md->notempty ||
1320            (md->notempty_atstart &&            (md->notempty_atstart &&
1321              mstart == md->start_subject + md->start_offset)))              mstart == md->start_subject + md->start_offset)))
1322        MRRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
1323    
1324      /* Otherwise, we have a match. */      /* Otherwise, we have a match. */
1325    
1326      md->end_match_ptr = eptr;           /* Record where we ended */      md->end_match_ptr = eptr;           /* Record where we ended */
1327      md->end_offset_top = offset_top;    /* and how many extracts were taken */      md->end_offset_top = offset_top;    /* and how many extracts were taken */
1328      md->start_match_ptr = mstart;       /* and the start (\K can modify) */      md->start_match_ptr = mstart;       /* and the start (\K can modify) */
# Line 1166  for (;;) Line 1333  for (;;)
1333      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;
1334      MRRETURN(rrc);      MRRETURN(rrc);
1335    
     /* Change option settings */  
   
     case OP_OPT:  
     ims = ecode[1];  
     ecode += 2;  
     DPRINTF(("ims set to %02lx\n", ims));  
     break;  
   
1336      /* Assertion brackets. Check the alternative branches in turn - the      /* Assertion brackets. Check the alternative branches in turn - the
1337      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,
1338      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
1339      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
1340      this level is identical to the lookahead case. */      this level is identical to the lookahead case. When the assertion is part
1341        of a condition, we want to return immediately afterwards. The caller of
1342        this incarnation of the match() function will have set MATCH_CONDASSERT in
1343        md->match_function type, and one of these opcodes will be the first opcode
1344        that is processed. We use a local variable that is preserved over calls to
1345        match() to remember this case. */
1346    
1347      case OP_ASSERT:      case OP_ASSERT:
1348      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1349        if (md->match_function_type == MATCH_CONDASSERT)
1350          {
1351          condassert = TRUE;
1352          md->match_function_type = 0;
1353          }
1354        else condassert = FALSE;
1355    
1356      do      do
1357        {        {
1358        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
         RM4);  
1359        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1360          {          {
1361          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1362          break;          break;
1363          }          }
1364        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH &&
1365              (rrc != MATCH_THEN || md->start_match_ptr != ecode))
1366            RRETURN(rrc);
1367        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1368        }        }
1369      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1370    
1371      if (*ecode == OP_KET) MRRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) MRRETURN(MATCH_NOMATCH);
1372    
1373      /* If checking an assertion for a condition, return MATCH_MATCH. */      /* If checking an assertion for a condition, return MATCH_MATCH. */
1374    
1375      if ((flags & match_condassert) != 0) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);
1376    
1377      /* Continue from after the assertion, updating the offsets high water      /* Continue from after the assertion, updating the offsets high water
1378      mark, since extracts may have been taken during the assertion. */      mark, since extracts may have been taken during the assertion. */
# Line 1215  for (;;) Line 1388  for (;;)
1388    
1389      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1390      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1391        if (md->match_function_type == MATCH_CONDASSERT)
1392          {
1393          condassert = TRUE;
1394          md->match_function_type = 0;
1395          }
1396        else condassert = FALSE;
1397    
1398      do      do
1399        {        {
1400        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
         RM5);  
1401        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) MRRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) MRRETURN(MATCH_NOMATCH);
1402        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1403          {          {
1404          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1405          break;          break;
1406          }          }
1407        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH &&
1408              (rrc != MATCH_THEN || md->start_match_ptr != ecode))
1409            RRETURN(rrc);
1410        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1411        }        }
1412      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1413    
1414      if ((flags & match_condassert) != 0) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1415    
1416      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1417      continue;      continue;
1418    
# Line 1280  for (;;) Line 1461  for (;;)
1461        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1462        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1463        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1464        cb.subject_length   = md->end_subject - md->start_subject;        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1465        cb.start_match      = mstart - md->start_subject;        cb.start_match      = (int)(mstart - md->start_subject);
1466        cb.current_position = eptr - md->start_subject;        cb.current_position = (int)(eptr - md->start_subject);
1467        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1468        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1469        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
# Line 1344  for (;;) Line 1525  for (;;)
1525        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1526              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1527        new_recursive.save_offset_top = offset_top;        new_recursive.save_offset_top = offset_top;
1528    
1529        /* OK, now we can do the recursion. For each top-level alternative we        /* OK, now we can do the recursion. For each top-level alternative we
1530        restore the offset and recursion data. */        restore the offset and recursion data. */
1531    
1532        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1533        flags = (*callpat >= OP_SBRA)? match_cbegroup : 0;        cbegroup = (*callpat >= OP_SBRA);
1534        do        do
1535          {          {
1536            if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1537          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,
1538            md, ims, eptrb, flags, RM6);            md, eptrb, RM6);
1539          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1540            {            {
1541            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
# Line 1362  for (;;) Line 1544  for (;;)
1544              (pcre_free)(new_recursive.offset_save);              (pcre_free)(new_recursive.offset_save);
1545            MRRETURN(MATCH_MATCH);            MRRETURN(MATCH_MATCH);
1546            }            }
1547          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH &&
1548                    (rrc != MATCH_THEN || md->start_match_ptr != ecode))
1549            {            {
1550            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1551            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
# Line 1399  for (;;) Line 1582  for (;;)
1582    
1583      do      do
1584        {        {
1585        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM7);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM7);
1586        if (rrc == MATCH_MATCH)  /* Note: _not_ MATCH_ACCEPT */        if (rrc == MATCH_MATCH)  /* Note: _not_ MATCH_ACCEPT */
1587          {          {
1588          mstart = md->start_match_ptr;          mstart = md->start_match_ptr;
1589          break;          break;
1590          }          }
1591        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH &&
1592              (rrc != MATCH_THEN || md->start_match_ptr != ecode))
1593            RRETURN(rrc);
1594        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1595        }        }
1596      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 1436  for (;;) Line 1621  for (;;)
1621    
1622      /* The repeating kets try the rest of the pattern or restart from the      /* The repeating kets try the rest of the pattern or restart from the
1623      preceding bracket, in the appropriate order. The second "call" of match()      preceding bracket, in the appropriate order. The second "call" of match()
1624      uses tail recursion, to avoid using another stack frame. We need to reset      uses tail recursion, to avoid using another stack frame. */
     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));  
       }  
1625    
1626      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1627        {        {
1628        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM8);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM8);
1629        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1630        ecode = prev;        ecode = prev;
       flags = 0;  
1631        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1632        }        }
1633      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
1634        {        {
1635        RMATCH(eptr, prev, offset_top, md, ims, eptrb, match_cbegroup, RM9);        md->match_function_type = MATCH_CBEGROUP;
1636          RMATCH(eptr, prev, offset_top, md, eptrb, RM9);
1637        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1638        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
       flags = 0;  
1639        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1640        }        }
1641      /* Control never gets here */      /* Control never gets here */
# Line 1476  for (;;) Line 1652  for (;;)
1652      at all - i.e. it could be ()* or ()? or even (){0} in the pattern. Brackets      at all - i.e. it could be ()* or ()? or even (){0} in the pattern. Brackets
1653      with fixed upper repeat limits are compiled as a number of copies, with the      with fixed upper repeat limits are compiled as a number of copies, with the
1654      optional ones preceded by BRAZERO or BRAMINZERO. */      optional ones preceded by BRAZERO or BRAMINZERO. */
1655    
1656      case OP_BRAZERO:      case OP_BRAZERO:
1657        {      next = ecode + 1;
1658        next = ecode+1;      RMATCH(eptr, next, offset_top, md, eptrb, RM10);
1659        RMATCH(eptr, next, offset_top, md, ims, eptrb, 0, RM10);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1660        if (rrc != MATCH_NOMATCH) RRETURN(rrc);      do next += GET(next, 1); while (*next == OP_ALT);
1661        do next += GET(next,1); while (*next == OP_ALT);      ecode = next + 1 + LINK_SIZE;
       ecode = next + 1 + LINK_SIZE;  
       }  
1662      break;      break;
1663    
1664      case OP_BRAMINZERO:      case OP_BRAMINZERO:
1665        {      next = ecode + 1;
1666        next = ecode+1;      do next += GET(next, 1); while (*next == OP_ALT);
1667        do next += GET(next, 1); while (*next == OP_ALT);      RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, eptrb, RM11);
1668        RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0, RM11);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1669        if (rrc != MATCH_NOMATCH) RRETURN(rrc);      ecode++;
       ecode++;  
       }  
1670      break;      break;
1671    
1672      case OP_SKIPZERO:      case OP_SKIPZERO:
1673        {      next = ecode+1;
1674        next = ecode+1;      do next += GET(next,1); while (*next == OP_ALT);
1675        do next += GET(next,1); while (*next == OP_ALT);      ecode = next + 1 + LINK_SIZE;
       ecode = next + 1 + LINK_SIZE;  
       }  
1676      break;      break;
1677    
1678        /* BRAPOSZERO occurs before a possessive bracket group. Don't do anything
1679        here; just jump to the group, with allow_zero set TRUE. */
1680    
1681        case OP_BRAPOSZERO:
1682        op = *(++ecode);
1683        allow_zero = TRUE;
1684        if (op == OP_CBRAPOS || op == OP_SCBRAPOS) goto POSSESSIVE_CAPTURE;
1685          goto POSSESSIVE_NON_CAPTURE;
1686    
1687      /* End of a group, repeated or non-repeating. */      /* End of a group, repeated or non-repeating. */
1688    
1689      case OP_KET:      case OP_KET:
1690      case OP_KETRMIN:      case OP_KETRMIN:
1691      case OP_KETRMAX:      case OP_KETRMAX:
1692        case OP_KETRPOS:
1693      prev = ecode - GET(ecode, 1);      prev = ecode - GET(ecode, 1);
1694    
1695      /* 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
# Line 1544  for (;;) Line 1724  for (;;)
1724      a recurse into group 0, so it won't be picked up here. Instead, we catch it      a recurse into group 0, so it won't be picked up here. Instead, we catch it
1725      when the OP_END is reached. Other recursion is handled here. */      when the OP_END is reached. Other recursion is handled here. */
1726    
1727      if (*prev == OP_CBRA || *prev == OP_SCBRA)      if (*prev == OP_CBRA || *prev == OP_SCBRA ||
1728            *prev == OP_CBRAPOS || *prev == OP_SCBRAPOS)
1729        {        {
1730        number = GET2(prev, 1+LINK_SIZE);        number = GET2(prev, 1+LINK_SIZE);
1731        offset = number << 1;        offset = number << 1;
# Line 1559  for (;;) Line 1740  for (;;)
1740          {          {
1741          md->offset_vector[offset] =          md->offset_vector[offset] =
1742            md->offset_vector[md->offset_end - number];            md->offset_vector[md->offset_end - number];
1743          md->offset_vector[offset+1] = eptr - md->start_subject;          md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1744          if (offset_top <= offset) offset_top = offset + 2;          if (offset_top <= offset) offset_top = offset + 2;
1745          }          }
1746    
# Line 1575  for (;;) Line 1756  for (;;)
1756            rec->saved_max * sizeof(int));            rec->saved_max * sizeof(int));
1757          offset_top = rec->save_offset_top;          offset_top = rec->save_offset_top;
1758          ecode = rec->after_call;          ecode = rec->after_call;
         ims = original_ims;  
1759          break;          break;
1760          }          }
1761        }        }
1762    
     /* For both capturing and non-capturing groups, reset the value of the ims  
     flags, in case they got changed during the group. */  
   
     ims = original_ims;  
     DPRINTF(("ims reset to %02lx\n", ims));  
   
1763      /* For a non-repeating ket, just continue at this level. This also      /* For a non-repeating ket, just continue at this level. This also
1764      happens for a repeating ket if no characters were matched in the group.      happens for a repeating ket if no characters were matched in the group.
1765      This is the forcible breaking of infinite loops as implemented in Perl      This is the forcible breaking of infinite loops as implemented in Perl
# Line 1597  for (;;) Line 1771  for (;;)
1771        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1772        break;        break;
1773        }        }
1774    
1775      /* The repeating kets try the rest of the pattern or restart from the      /* OP_KETRPOS is a possessive repeating ket. Remember the current position,
1776      preceding bracket, in the appropriate order. In the second case, we can use      and return the MATCH_KETRPOS. This makes it possible to do the repeats one
1777      tail recursion to avoid using another stack frame, unless we have an      at a time from the outer level, thus saving stack. */
1778    
1779        if (*ecode == OP_KETRPOS)
1780          {
1781          md->end_match_ptr = eptr;
1782          md->end_offset_top = offset_top;
1783          RRETURN(MATCH_KETRPOS);
1784          }
1785    
1786        /* The normal repeating kets try the rest of the pattern or restart from
1787        the preceding bracket, in the appropriate order. In the second case, we can
1788        use tail recursion to avoid using another stack frame, unless we have an
1789      unlimited repeat of a group that can match an empty string. */      unlimited repeat of a group that can match an empty string. */
1790    
     flags = (*prev >= OP_SBRA)? match_cbegroup : 0;  
   
1791      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1792        {        {
1793        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM12);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM12);
1794        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1795        if (flags != 0)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
1796          {          {
1797          RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM50);          md->match_function_type = MATCH_CBEGROUP;
1798            RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
1799          RRETURN(rrc);          RRETURN(rrc);
1800          }          }
1801        ecode = prev;        ecode = prev;
# Line 1619  for (;;) Line 1803  for (;;)
1803        }        }
1804      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
1805        {        {
1806        RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13);        if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1807          RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
1808        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1809        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
       flags = 0;  
1810        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1811        }        }
1812      /* Control never gets here */      /* Control never gets here */
1813    
1814      /* Start of subject unless notbol, or after internal newline if multiline */      /* Not multiline mode: start of subject assertion, unless notbol. */
1815    
1816      case OP_CIRC:      case OP_CIRC:
1817      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);
1818      if ((ims & PCRE_MULTILINE) != 0)  
       {  
       if (eptr != md->start_subject &&  
           (eptr == md->end_subject || !WAS_NEWLINE(eptr)))  
         MRRETURN(MATCH_NOMATCH);  
       ecode++;  
       break;  
       }  
     /* ... else fall through */  
   
1819      /* Start of subject assertion */      /* Start of subject assertion */
1820    
1821      case OP_SOD:      case OP_SOD:
1822      if (eptr != md->start_subject) MRRETURN(MATCH_NOMATCH);      if (eptr != md->start_subject) MRRETURN(MATCH_NOMATCH);
1823      ecode++;      ecode++;
1824      break;      break;
1825    
1826        /* Multiline mode: start of subject unless notbol, or after any newline. */
1827    
1828        case OP_CIRCM:
1829        if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);
1830        if (eptr != md->start_subject &&
1831            (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
1832          MRRETURN(MATCH_NOMATCH);
1833        ecode++;
1834        break;
1835    
1836      /* Start of match assertion */      /* Start of match assertion */
1837    
# Line 1662  for (;;) Line 1847  for (;;)
1847      ecode++;      ecode++;
1848      break;      break;
1849    
1850      /* Assert before internal newline if multiline, or before a terminating      /* Multiline mode: assert before any newline, or before end of subject
1851      newline unless endonly is set, else end of subject unless noteol is set. */      unless noteol is set. */
1852    
1853      case OP_DOLL:      case OP_DOLLM:
1854      if ((ims & PCRE_MULTILINE) != 0)      if (eptr < md->end_subject)
1855        {        { if (!IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH); }
       if (eptr < md->end_subject)  
         { if (!IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH); }  
       else  
         { if (md->noteol) MRRETURN(MATCH_NOMATCH); }  
       ecode++;  
       break;  
       }  
1856      else      else
1857        {        {
1858        if (md->noteol) MRRETURN(MATCH_NOMATCH);        if (md->noteol) MRRETURN(MATCH_NOMATCH);
1859        if (!md->endonly)        SCHECK_PARTIAL();
         {  
         if (eptr != md->end_subject &&  
             (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))  
           MRRETURN(MATCH_NOMATCH);  
         ecode++;  
         break;  
         }  
1860        }        }
1861        ecode++;
1862        break;
1863    
1864        /* Not multiline mode: assert before a terminating newline or before end of
1865        subject unless noteol is set. */
1866    
1867        case OP_DOLL:
1868        if (md->noteol) MRRETURN(MATCH_NOMATCH);
1869        if (!md->endonly) goto ASSERT_NL_OR_EOS;
1870    
1871      /* ... else fall through for endonly */      /* ... else fall through for endonly */
1872    
1873      /* End of subject assertion (\z) */      /* End of subject assertion (\z) */
1874    
1875      case OP_EOD:      case OP_EOD:
1876      if (eptr < md->end_subject) MRRETURN(MATCH_NOMATCH);      if (eptr < md->end_subject) MRRETURN(MATCH_NOMATCH);
1877        SCHECK_PARTIAL();
1878      ecode++;      ecode++;
1879      break;      break;
1880    
1881      /* End of subject or ending \n assertion (\Z) */      /* End of subject or ending \n assertion (\Z) */
1882    
1883      case OP_EODN:      case OP_EODN:
1884      if (eptr != md->end_subject &&      ASSERT_NL_OR_EOS:
1885        if (eptr < md->end_subject &&
1886          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
1887        MRRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
1888    
1889        /* Either at end of string or \n before end. */
1890    
1891        SCHECK_PARTIAL();
1892      ecode++;      ecode++;
1893      break;      break;
1894    
# Line 1963  for (;;) Line 2150  for (;;)
2150      switch(c)      switch(c)
2151        {        {
2152        default: MRRETURN(MATCH_NOMATCH);        default: MRRETURN(MATCH_NOMATCH);
2153    
2154        case 0x000d:        case 0x000d:
2155        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
2156        break;        break;
# Line 2215  for (;;) Line 2403  for (;;)
2403      loops). */      loops). */
2404    
2405      case OP_REF:      case OP_REF:
2406        {      case OP_REFI:
2407        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      caseless = op == OP_REFI;
2408        ecode += 3;      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2409        ecode += 3;
2410    
2411        /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2412    
2413        (a) In the default, Perl-compatible state, set the length to be longer      (a) In the default, Perl-compatible state, set the length negative;
2414        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
2415        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.  
2416    
2417        (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
2418        so that the back reference matches an empty string.      so that the back reference matches an empty string.
2419    
2420        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
2421        referenced subpattern. */      referenced subpattern. */
2422    
2423        if (offset >= offset_top || md->offset_vector[offset] < 0)      if (offset >= offset_top || md->offset_vector[offset] < 0)
2424          length = (md->jscript_compat)? 0 : md->end_subject - eptr + 1;        length = (md->jscript_compat)? 0 : -1;
2425        else      else
2426          length = md->offset_vector[offset+1] - md->offset_vector[offset];        length = md->offset_vector[offset+1] - md->offset_vector[offset];
2427    
2428        /* Set up for repetition, or handle the non-repeated case */      /* Set up for repetition, or handle the non-repeated case */
2429    
2430        switch (*ecode)      switch (*ecode)
2431          {        {
2432          case OP_CRSTAR:        case OP_CRSTAR:
2433          case OP_CRMINSTAR:        case OP_CRMINSTAR:
2434          case OP_CRPLUS:        case OP_CRPLUS:
2435          case OP_CRMINPLUS:        case OP_CRMINPLUS:
2436          case OP_CRQUERY:        case OP_CRQUERY:
2437          case OP_CRMINQUERY:        case OP_CRMINQUERY:
2438          c = *ecode++ - OP_CRSTAR;        c = *ecode++ - OP_CRSTAR;
2439          minimize = (c & 1) != 0;        minimize = (c & 1) != 0;
2440          min = rep_min[c];                 /* Pick up values from tables; */        min = rep_min[c];                 /* Pick up values from tables; */
2441          max = rep_max[c];                 /* zero for max => infinity */        max = rep_max[c];                 /* zero for max => infinity */
2442          if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2443          break;        break;
2444    
2445          case OP_CRRANGE:        case OP_CRRANGE:
2446          case OP_CRMINRANGE:        case OP_CRMINRANGE:
2447          minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2448          min = GET2(ecode, 1);        min = GET2(ecode, 1);
2449          max = GET2(ecode, 3);        max = GET2(ecode, 3);
2450          if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2451          ecode += 5;        ecode += 5;
2452          break;        break;
2453    
2454          default:               /* No repeat follows */        default:               /* No repeat follows */
2455          if (!match_ref(offset, eptr, length, md, ims))        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2456            {          {
2457            CHECK_PARTIAL();          CHECK_PARTIAL();
2458            MRRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
           }  
         eptr += length;  
         continue;              /* With the main loop */  
2459          }          }
2460          eptr += length;
2461          continue;              /* With the main loop */
2462          }
2463    
2464        /* If the length of the reference is zero, just continue with the      /* Handle repeated back references. If the length of the reference is
2465        main loop. */      zero, just continue with the main loop. */
2466    
2467        if (length == 0) continue;      if (length == 0) continue;
2468    
2469        /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2470        the length of the reference string explicitly rather than passing the      the length of the reference string explicitly rather than passing the
2471        address of eptr, so that eptr can be a register variable. */      address of eptr, so that eptr can be a register variable. */
2472    
2473        for (i = 1; i <= min; i++)      for (i = 1; i <= min; i++)
2474          {
2475          int slength;
2476          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2477          {          {
2478          if (!match_ref(offset, eptr, length, md, ims))          CHECK_PARTIAL();
2479            {          MRRETURN(MATCH_NOMATCH);
           CHECK_PARTIAL();  
           MRRETURN(MATCH_NOMATCH);  
           }  
         eptr += length;  
2480          }          }
2481          eptr += slength;
2482          }
2483    
2484        /* If min = max, continue at the same level without recursion.      /* If min = max, continue at the same level without recursion.
2485        They are not both allowed to be zero. */      They are not both allowed to be zero. */
2486    
2487        if (min == max) continue;      if (min == max) continue;
2488    
2489        /* If minimizing, keep trying and advancing the pointer */      /* If minimizing, keep trying and advancing the pointer */
2490    
2491        if (minimize)      if (minimize)
2492          {
2493          for (fi = min;; fi++)
2494          {          {
2495          for (fi = min;; fi++)          int slength;
2496            RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);
2497            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2498            if (fi >= max) MRRETURN(MATCH_NOMATCH);
2499            if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2500            {            {
2501            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM14);            CHECK_PARTIAL();
2502            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            MRRETURN(MATCH_NOMATCH);
           if (fi >= max) MRRETURN(MATCH_NOMATCH);  
           if (!match_ref(offset, eptr, length, md, ims))  
             {  
             CHECK_PARTIAL();  
             MRRETURN(MATCH_NOMATCH);  
             }  
           eptr += length;  
2503            }            }
2504          /* Control never gets here */          eptr += slength;
2505          }          }
2506          /* Control never gets here */
2507          }
2508    
2509        /* If maximizing, find the longest string and work backwards */      /* If maximizing, find the longest string and work backwards */
2510    
2511        else      else
2512          {
2513          pp = eptr;
2514          for (i = min; i < max; i++)
2515          {          {
2516          pp = eptr;          int slength;
2517          for (i = min; i < max; i++)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
           {  
           if (!match_ref(offset, eptr, length, md, ims))  
             {  
             CHECK_PARTIAL();  
             break;  
             }  
           eptr += length;  
           }  
         while (eptr >= pp)  
2518            {            {
2519            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM15);            CHECK_PARTIAL();
2520            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            break;
           eptr -= length;  
2521            }            }
2522          MRRETURN(MATCH_NOMATCH);          eptr += slength;
2523          }          }
2524          while (eptr >= pp)
2525            {
2526            RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
2527            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2528            eptr -= length;
2529            }
2530          MRRETURN(MATCH_NOMATCH);
2531        }        }
2532      /* Control never gets here */      /* Control never gets here */
2533    
# Line 2443  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, RM16);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);
2637              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2638              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
2639              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 2468  for (;;) Line 2658  for (;;)
2658            {            {
2659            for (fi = min;; fi++)            for (fi = min;; fi++)
2660              {              {
2661              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM17);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);
2662              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2663              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
2664              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 2514  for (;;) Line 2704  for (;;)
2704              }              }
2705            for (;;)            for (;;)
2706              {              {
2707              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM18);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM18);
2708              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2709              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2710              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2537  for (;;) Line 2727  for (;;)
2727              }              }
2728            while (eptr >= pp)            while (eptr >= pp)
2729              {              {
2730              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM19);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM19);
2731              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2732              eptr--;              eptr--;
2733              }              }
# Line 2613  for (;;) Line 2803  for (;;)
2803          {          {
2804          for (fi = min;; fi++)          for (fi = min;; fi++)
2805            {            {
2806            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);
2807            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2808            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
2809            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 2646  for (;;) Line 2836  for (;;)
2836            }            }
2837          for(;;)          for(;;)
2838            {            {
2839            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
2840            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2841            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
2842            if (utf8) BACKCHAR(eptr);            if (utf8) BACKCHAR(eptr);
# Line 2691  for (;;) Line 2881  for (;;)
2881    
2882      /* Match a single character, caselessly */      /* Match a single character, caselessly */
2883    
2884      case OP_CHARNC:      case OP_CHARI:
2885  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2886      if (utf8)      if (utf8)
2887        {        {
# Line 2751  for (;;) Line 2941  for (;;)
2941      /* Match a single character repeatedly. */      /* Match a single character repeatedly. */
2942    
2943      case OP_EXACT:      case OP_EXACT:
2944        case OP_EXACTI:
2945      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
2946      ecode += 3;      ecode += 3;
2947      goto REPEATCHAR;      goto REPEATCHAR;
2948    
2949      case OP_POSUPTO:      case OP_POSUPTO:
2950        case OP_POSUPTOI:
2951      possessive = TRUE;      possessive = TRUE;
2952      /* Fall through */      /* Fall through */
2953    
2954      case OP_UPTO:      case OP_UPTO:
2955        case OP_UPTOI:
2956      case OP_MINUPTO:      case OP_MINUPTO:
2957        case OP_MINUPTOI:
2958      min = 0;      min = 0;
2959      max = GET2(ecode, 1);      max = GET2(ecode, 1);
2960      minimize = *ecode == OP_MINUPTO;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
2961      ecode += 3;      ecode += 3;
2962      goto REPEATCHAR;      goto REPEATCHAR;
2963    
2964      case OP_POSSTAR:      case OP_POSSTAR:
2965        case OP_POSSTARI:
2966      possessive = TRUE;      possessive = TRUE;
2967      min = 0;      min = 0;
2968      max = INT_MAX;      max = INT_MAX;
# Line 2775  for (;;) Line 2970  for (;;)
2970      goto REPEATCHAR;      goto REPEATCHAR;
2971    
2972      case OP_POSPLUS:      case OP_POSPLUS:
2973        case OP_POSPLUSI:
2974      possessive = TRUE;      possessive = TRUE;
2975      min = 1;      min = 1;
2976      max = INT_MAX;      max = INT_MAX;
# Line 2782  for (;;) Line 2978  for (;;)
2978      goto REPEATCHAR;      goto REPEATCHAR;
2979    
2980      case OP_POSQUERY:      case OP_POSQUERY:
2981        case OP_POSQUERYI:
2982      possessive = TRUE;      possessive = TRUE;
2983      min = 0;      min = 0;
2984      max = 1;      max = 1;
# Line 2789  for (;;) Line 2986  for (;;)
2986      goto REPEATCHAR;      goto REPEATCHAR;
2987    
2988      case OP_STAR:      case OP_STAR:
2989        case OP_STARI:
2990      case OP_MINSTAR:      case OP_MINSTAR:
2991        case OP_MINSTARI:
2992      case OP_PLUS:      case OP_PLUS:
2993        case OP_PLUSI:
2994      case OP_MINPLUS:      case OP_MINPLUS:
2995        case OP_MINPLUSI:
2996      case OP_QUERY:      case OP_QUERY:
2997        case OP_QUERYI:
2998      case OP_MINQUERY:      case OP_MINQUERY:
2999      c = *ecode++ - OP_STAR;      case OP_MINQUERYI:
3000        c = *ecode++ - ((op < OP_STARI)? OP_STAR : OP_STARI);
3001      minimize = (c & 1) != 0;      minimize = (c & 1) != 0;
   
3002      min = rep_min[c];                 /* Pick up values from tables; */      min = rep_min[c];                 /* Pick up values from tables; */
3003      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
3004      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
# Line 2819  for (;;) Line 3021  for (;;)
3021          {          {
3022  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3023          unsigned int othercase;          unsigned int othercase;
3024          if ((ims & PCRE_CASELESS) != 0 &&          if (op >= OP_STARI &&     /* Caseless */
3025              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3026            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = _pcre_ord2utf8(othercase, occhars);
3027          else oclength = 0;          else oclength = 0;
# Line 2847  for (;;) Line 3049  for (;;)
3049            {            {
3050            for (fi = min;; fi++)            for (fi = min;; fi++)
3051              {              {
3052              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM22);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);
3053              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3054              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3055              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
# Line 2889  for (;;) Line 3091  for (;;)
3091    
3092            for(;;)            for(;;)
3093              {              {
3094              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3095              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3096              if (eptr == pp) { MRRETURN(MATCH_NOMATCH); }              if (eptr == pp) { MRRETURN(MATCH_NOMATCH); }
3097  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2926  for (;;) Line 3128  for (;;)
3128      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3129        max, eptr));        max, eptr));
3130    
3131      if ((ims & PCRE_CASELESS) != 0)      if (op >= OP_STARI)  /* Caseless */
3132        {        {
3133        fc = md->lcc[fc];        fc = md->lcc[fc];
3134        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
# Line 2943  for (;;) Line 3145  for (;;)
3145          {          {
3146          for (fi = min;; fi++)          for (fi = min;; fi++)
3147            {            {
3148            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3149            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3150            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
3151            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 2973  for (;;) Line 3175  for (;;)
3175    
3176          while (eptr >= pp)          while (eptr >= pp)
3177            {            {
3178            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM25);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);
3179            eptr--;            eptr--;
3180            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3181            }            }
# Line 3002  for (;;) Line 3204  for (;;)
3204          {          {
3205          for (fi = min;; fi++)          for (fi = min;; fi++)
3206            {            {
3207            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM26);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);
3208            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3209            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
3210            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3031  for (;;) Line 3233  for (;;)
3233    
3234          while (eptr >= pp)          while (eptr >= pp)
3235            {            {
3236            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);
3237            eptr--;            eptr--;
3238            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3239            }            }
# Line 3044  for (;;) Line 3246  for (;;)
3246      checking can be multibyte. */      checking can be multibyte. */
3247    
3248      case OP_NOT:      case OP_NOT:
3249        case OP_NOTI:
3250      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
3251        {        {
3252        SCHECK_PARTIAL();        SCHECK_PARTIAL();
# Line 3051  for (;;) Line 3254  for (;;)
3254        }        }
3255      ecode++;      ecode++;
3256      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
3257      if ((ims & PCRE_CASELESS) != 0)      if (op == OP_NOTI)         /* The caseless case */
3258        {        {
3259  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3260        if (c < 256)        if (c < 256)
# Line 3059  for (;;) Line 3262  for (;;)
3262        c = md->lcc[c];        c = md->lcc[c];
3263        if (md->lcc[*ecode++] == c) MRRETURN(MATCH_NOMATCH);        if (md->lcc[*ecode++] == c) MRRETURN(MATCH_NOMATCH);
3264        }        }
3265      else      else    /* Caseful */
3266        {        {
3267        if (*ecode++ == c) MRRETURN(MATCH_NOMATCH);        if (*ecode++ == c) MRRETURN(MATCH_NOMATCH);
3268        }        }
# Line 3073  for (;;) Line 3276  for (;;)
3276      about... */      about... */
3277    
3278      case OP_NOTEXACT:      case OP_NOTEXACT:
3279        case OP_NOTEXACTI:
3280      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3281      ecode += 3;      ecode += 3;
3282      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3283    
3284      case OP_NOTUPTO:      case OP_NOTUPTO:
3285        case OP_NOTUPTOI:
3286      case OP_NOTMINUPTO:      case OP_NOTMINUPTO:
3287        case OP_NOTMINUPTOI:
3288      min = 0;      min = 0;
3289      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3290      minimize = *ecode == OP_NOTMINUPTO;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3291      ecode += 3;      ecode += 3;
3292      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3293    
3294      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
3295        case OP_NOTPOSSTARI:
3296      possessive = TRUE;      possessive = TRUE;
3297      min = 0;      min = 0;
3298      max = INT_MAX;      max = INT_MAX;
# Line 3093  for (;;) Line 3300  for (;;)
3300      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3301    
3302      case OP_NOTPOSPLUS:      case OP_NOTPOSPLUS:
3303        case OP_NOTPOSPLUSI:
3304      possessive = TRUE;      possessive = TRUE;
3305      min = 1;      min = 1;
3306      max = INT_MAX;      max = INT_MAX;
# Line 3100  for (;;) Line 3308  for (;;)
3308      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3309    
3310      case OP_NOTPOSQUERY:      case OP_NOTPOSQUERY:
3311        case OP_NOTPOSQUERYI:
3312      possessive = TRUE;      possessive = TRUE;
3313      min = 0;      min = 0;
3314      max = 1;      max = 1;
# Line 3107  for (;;) Line 3316  for (;;)
3316      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3317    
3318      case OP_NOTPOSUPTO:      case OP_NOTPOSUPTO:
3319        case OP_NOTPOSUPTOI:
3320      possessive = TRUE;      possessive = TRUE;
3321      min = 0;      min = 0;
3322      max = GET2(ecode, 1);      max = GET2(ecode, 1);
# Line 3114  for (;;) Line 3324  for (;;)
3324      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3325    
3326      case OP_NOTSTAR:      case OP_NOTSTAR:
3327        case OP_NOTSTARI:
3328      case OP_NOTMINSTAR:      case OP_NOTMINSTAR:
3329        case OP_NOTMINSTARI:
3330      case OP_NOTPLUS:      case OP_NOTPLUS:
3331        case OP_NOTPLUSI:
3332      case OP_NOTMINPLUS:      case OP_NOTMINPLUS:
3333        case OP_NOTMINPLUSI:
3334      case OP_NOTQUERY:      case OP_NOTQUERY:
3335        case OP_NOTQUERYI:
3336      case OP_NOTMINQUERY:      case OP_NOTMINQUERY:
3337      c = *ecode++ - OP_NOTSTAR;      case OP_NOTMINQUERYI:
3338        c = *ecode++ - ((op >= OP_NOTSTARI)? OP_NOTSTARI: OP_NOTSTAR);
3339      minimize = (c & 1) != 0;      minimize = (c & 1) != 0;
3340      min = rep_min[c];                 /* Pick up values from tables; */      min = rep_min[c];                 /* Pick up values from tables; */
3341      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
# Line 3141  for (;;) Line 3357  for (;;)
3357      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,
3358        max, eptr));        max, eptr));
3359    
3360      if ((ims & PCRE_CASELESS) != 0)      if (op >= OP_NOTSTARI)     /* Caseless */
3361        {        {
3362        fc = md->lcc[fc];        fc = md->lcc[fc];
3363    
# Line 3189  for (;;) Line 3405  for (;;)
3405            register unsigned int d;            register unsigned int d;
3406            for (fi = min;; fi++)            for (fi = min;; fi++)
3407              {              {
3408              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
3409              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3410              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3411              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 3208  for (;;) Line 3424  for (;;)
3424            {            {
3425            for (fi = min;; fi++)            for (fi = min;; fi++)
3426              {              {
3427              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM29);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);
3428              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3429              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3430              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 3249  for (;;) Line 3465  for (;;)
3465          if (possessive) continue;          if (possessive) continue;
3466          for(;;)          for(;;)
3467              {              {
3468              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3469              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3470              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
3471              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3272  for (;;) Line 3488  for (;;)
3488            if (possessive) continue;            if (possessive) continue;
3489            while (eptr >= pp)            while (eptr >= pp)
3490              {              {
3491              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM31);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);
3492              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3493              eptr--;              eptr--;
3494              }              }
# Line 3329  for (;;) Line 3545  for (;;)
3545            register unsigned int d;            register unsigned int d;
3546            for (fi = min;; fi++)            for (fi = min;; fi++)
3547              {              {
3548              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
3549              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3550              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3551              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 3347  for (;;) Line 3563  for (;;)
3563            {            {
3564            for (fi = min;; fi++)            for (fi = min;; fi++)
3565              {              {
3566              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM33);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);
3567              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3568              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3569              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 3387  for (;;) Line 3603  for (;;)
3603            if (possessive) continue;            if (possessive) continue;
3604            for(;;)            for(;;)
3605              {              {
3606              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM34);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);
3607              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3608              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
3609              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3410  for (;;) Line 3626  for (;;)
3626            if (possessive) continue;            if (possessive) continue;
3627            while (eptr >= pp)            while (eptr >= pp)
3628              {              {
3629              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM35);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);
3630              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3631              eptr--;              eptr--;
3632              }              }
# Line 3737  for (;;) Line 3953  for (;;)
3953            switch(c)            switch(c)
3954              {              {
3955              default: MRRETURN(MATCH_NOMATCH);              default: MRRETURN(MATCH_NOMATCH);
3956    
3957              case 0x000d:              case 0x000d:
3958              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3959              break;              break;
# Line 4013  for (;;) Line 4230  for (;;)
4230            switch(*eptr++)            switch(*eptr++)
4231              {              {
4232              default: MRRETURN(MATCH_NOMATCH);              default: MRRETURN(MATCH_NOMATCH);
4233    
4234              case 0x000d:              case 0x000d:
4235              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
4236              break;              break;
4237    
4238              case 0x000a:              case 0x000a:
4239              break;              break;
4240    
# Line 4205  for (;;) Line 4424  for (;;)
4424            case PT_ANY:            case PT_ANY:
4425            for (fi = min;; fi++)            for (fi = min;; fi++)
4426              {              {
4427              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM36);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM36);
4428              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4429              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4430              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4213  for (;;) Line 4432  for (;;)
4432                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4433                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4434                }                }
4435              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
4436              if (prop_fail_result) MRRETURN(MATCH_NOMATCH);              if (prop_fail_result) MRRETURN(MATCH_NOMATCH);
4437              }              }
4438            /* Control never gets here */            /* Control never gets here */
# Line 4221  for (;;) Line 4440  for (;;)
4440            case PT_LAMP:            case PT_LAMP:
4441            for (fi = min;; fi++)            for (fi = min;; fi++)
4442              {              {
4443              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM37);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM37);
4444              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4445              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4446              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4229  for (;;) Line 4448  for (;;)
4448                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4449                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4450                }                }
4451              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
4452              prop_chartype = UCD_CHARTYPE(c);              prop_chartype = UCD_CHARTYPE(c);
4453              if ((prop_chartype == ucp_Lu ||              if ((prop_chartype == ucp_Lu ||
4454                   prop_chartype == ucp_Ll ||                   prop_chartype == ucp_Ll ||
# Line 4241  for (;;) Line 4460  for (;;)
4460            case PT_GC:            case PT_GC:
4461            for (fi = min;; fi++)            for (fi = min;; fi++)
4462              {              {
4463              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM38);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM38);
4464              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4465              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4466              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4249  for (;;) Line 4468  for (;;)
4468                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4469                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4470                }                }
4471              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
4472              prop_category = UCD_CATEGORY(c);              prop_category = UCD_CATEGORY(c);
4473              if ((prop_category == prop_value) == prop_fail_result)              if ((prop_category == prop_value) == prop_fail_result)
4474                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
# Line 4259  for (;;) Line 4478  for (;;)
4478            case PT_PC:            case PT_PC:
4479            for (fi = min;; fi++)            for (fi = min;; fi++)
4480              {              {
4481              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM39);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM39);
4482              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4483              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4484              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4267  for (;;) Line 4486  for (;;)
4486                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4487                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4488                }                }
4489              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
4490              prop_chartype = UCD_CHARTYPE(c);              prop_chartype = UCD_CHARTYPE(c);
4491              if ((prop_chartype == prop_value) == prop_fail_result)              if ((prop_chartype == prop_value) == prop_fail_result)
4492                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
# Line 4277  for (;;) Line 4496  for (;;)
4496            case PT_SC:            case PT_SC:
4497            for (fi = min;; fi++)            for (fi = min;; fi++)
4498              {              {
4499              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM40);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM40);
4500              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4501              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4502              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4285  for (;;) Line 4504  for (;;)
4504                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4505                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4506                }                }
4507              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
4508              prop_script = UCD_SCRIPT(c);              prop_script = UCD_SCRIPT(c);
4509              if ((prop_script == prop_value) == prop_fail_result)              if ((prop_script == prop_value) == prop_fail_result)
4510                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
# Line 4295  for (;;) Line 4514  for (;;)
4514            case PT_ALNUM:            case PT_ALNUM:
4515            for (fi = min;; fi++)            for (fi = min;; fi++)
4516              {              {
4517              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM59);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM59);
4518              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4519              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4520              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4303  for (;;) Line 4522  for (;;)
4522                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4523                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4524                }                }
4525              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
4526              prop_category = UCD_CATEGORY(c);              prop_category = UCD_CATEGORY(c);
4527              if ((prop_category == ucp_L || prop_category == ucp_N)              if ((prop_category == ucp_L || prop_category == ucp_N)
4528                     == prop_fail_result)                     == prop_fail_result)
# Line 4314  for (;;) Line 4533  for (;;)
4533            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
4534            for (fi = min;; fi++)            for (fi = min;; fi++)
4535              {              {
4536              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM60);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM60);
4537              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4538              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4539              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4322  for (;;) Line 4541  for (;;)
4541                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4542                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4543                }                }
4544              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
4545              prop_category = UCD_CATEGORY(c);              prop_category = UCD_CATEGORY(c);
4546              if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
4547                   c == CHAR_FF || c == CHAR_CR)                   c == CHAR_FF || c == CHAR_CR)
# Line 4334  for (;;) Line 4553  for (;;)
4553            case PT_PXSPACE:  /* POSIX space */            case PT_PXSPACE:  /* POSIX space */
4554            for (fi = min;; fi++)            for (fi = min;; fi++)
4555              {              {
4556              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM61);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM61);
4557              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4558              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4559              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4342  for (;;) Line 4561  for (;;)
4561                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4562                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4563                }                }
4564              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
4565              prop_category = UCD_CATEGORY(c);              prop_category = UCD_CATEGORY(c);
4566              if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
4567                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
# Line 4354  for (;;) Line 4573  for (;;)
4573            case PT_WORD:            case PT_WORD:
4574            for (fi = min;; fi++)            for (fi = min;; fi++)
4575              {              {
4576              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM62);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM62);
4577              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4578              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4579              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4362  for (;;) Line 4581  for (;;)
4581                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4582                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4583                }                }
4584              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
4585              prop_category = UCD_CATEGORY(c);              prop_category = UCD_CATEGORY(c);
4586              if ((prop_category == ucp_L ||              if ((prop_category == ucp_L ||
4587                   prop_category == ucp_N ||                   prop_category == ucp_N ||
# Line 4386  for (;;) Line 4605  for (;;)
4605          {          {
4606          for (fi = min;; fi++)          for (fi = min;; fi++)
4607            {            {
4608            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM41);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM41);
4609            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4610            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
4611            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 4418  for (;;) Line 4637  for (;;)
4637          {          {
4638          for (fi = min;; fi++)          for (fi = min;; fi++)
4639            {            {
4640            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM42);
4641            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4642            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
4643            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 4581  for (;;) Line 4800  for (;;)
4800          {          {
4801          for (fi = min;; fi++)          for (fi = min;; fi++)
4802            {            {
4803            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM43);
4804            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4805            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
4806            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 4720  for (;;) Line 4939  for (;;)
4939                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4940                break;                break;
4941                }                }
4942              GETCHARLEN(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
4943              if (prop_fail_result) break;              if (prop_fail_result) break;
4944              eptr+= len;              eptr+= len;
4945              }              }
# Line 4735  for (;;) Line 4954  for (;;)
4954                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4955                break;                break;
4956                }                }
4957              GETCHARLEN(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
4958              prop_chartype = UCD_CHARTYPE(c);              prop_chartype = UCD_CHARTYPE(c);
4959              if ((prop_chartype == ucp_Lu ||              if ((prop_chartype == ucp_Lu ||
4960                   prop_chartype == ucp_Ll ||                   prop_chartype == ucp_Ll ||
# Line 4754  for (;;) Line 4973  for (;;)
4973                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4974                break;                break;
4975                }                }
4976              GETCHARLEN(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
4977              prop_category = UCD_CATEGORY(c);              prop_category = UCD_CATEGORY(c);
4978              if ((prop_category == prop_value) == prop_fail_result)              if ((prop_category == prop_value) == prop_fail_result)
4979                break;                break;
# Line 4771  for (;;) Line 4990  for (;;)
4990                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4991                break;                break;
4992                }                }
4993              GETCHARLEN(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
4994              prop_chartype = UCD_CHARTYPE(c);              prop_chartype = UCD_CHARTYPE(c);
4995              if ((prop_chartype == prop_value) == prop_fail_result)              if ((prop_chartype == prop_value) == prop_fail_result)
4996                break;                break;
# Line 4788  for (;;) Line 5007  for (;;)
5007                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5008                break;                break;
5009                }                }
5010              GETCHARLEN(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5011              prop_script = UCD_SCRIPT(c);              prop_script = UCD_SCRIPT(c);
5012              if ((prop_script == prop_value) == prop_fail_result)              if ((prop_script == prop_value) == prop_fail_result)
5013                break;                break;
# Line 4805  for (;;) Line 5024  for (;;)
5024                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5025                break;                break;
5026                }                }
5027              GETCHARLEN(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5028              prop_category = UCD_CATEGORY(c);              prop_category = UCD_CATEGORY(c);
5029              if ((prop_category == ucp_L || prop_category == ucp_N)              if ((prop_category == ucp_L || prop_category == ucp_N)
5030                   == prop_fail_result)                   == prop_fail_result)
# Line 4823  for (;;) Line 5042  for (;;)
5042                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5043                break;                break;
5044                }                }
5045              GETCHARLEN(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5046              prop_category = UCD_CATEGORY(c);              prop_category = UCD_CATEGORY(c);
5047              if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
5048                   c == CHAR_FF || c == CHAR_CR)                   c == CHAR_FF || c == CHAR_CR)
# Line 4842  for (;;) Line 5061  for (;;)
5061                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5062                break;                break;
5063                }                }
5064              GETCHARLEN(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5065              prop_category = UCD_CATEGORY(c);              prop_category = UCD_CATEGORY(c);
5066              if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
5067                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
# Line 4861  for (;;) Line 5080  for (;;)
5080                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5081                break;                break;
5082                }                }
5083              GETCHARLEN(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5084              prop_category = UCD_CATEGORY(c);              prop_category = UCD_CATEGORY(c);
5085              if ((prop_category == ucp_L || prop_category == ucp_N ||              if ((prop_category == ucp_L || prop_category == ucp_N ||
5086                   c == CHAR_UNDERSCORE) == prop_fail_result)                   c == CHAR_UNDERSCORE) == prop_fail_result)
# Line 4879  for (;;) Line 5098  for (;;)
5098          if (possessive) continue;          if (possessive) continue;
5099          for(;;)          for(;;)
5100            {            {
5101            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5102            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5103            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5104            if (utf8) BACKCHAR(eptr);            if (utf8) BACKCHAR(eptr);
# Line 4920  for (;;) Line 5139  for (;;)
5139    
5140          for(;;)          for(;;)
5141            {            {
5142            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM45);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM45);
5143            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5144            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5145            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
# Line 5204  for (;;) Line 5423  for (;;)
5423            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5424            }            }
5425    
5426          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run. If possessive, we are
5427            done (no backing up). Otherwise, match at this position; anything other
5428            than no match is immediately returned. For nomatch, back up one
5429            character, unless we are matching \R and the last thing matched was
5430            \r\n, in which case, back up two bytes. */
5431    
5432          if (possessive) continue;          if (possessive) continue;
5433          for(;;)          for(;;)
5434            {            {
5435            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM46);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM46);
5436            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5437            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5438            BACKCHAR(eptr);            BACKCHAR(eptr);
5439              if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&
5440                  eptr[-1] == '\r') eptr--;
5441            }            }
5442          }          }
5443        else        else
# Line 5411  for (;;) Line 5636  for (;;)
5636            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5637            }            }
5638    
5639          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run. If possessive, we are
5640            done (no backing up). Otherwise, match at this position; anything other
5641            than no match is immediately returned. For nomatch, back up one
5642            character (byte), unless we are matching \R and the last thing matched
5643            was \r\n, in which case, back up two bytes. */
5644    
5645          if (possessive) continue;          if (possessive) continue;
5646          while (eptr >= pp)          while (eptr >= pp)
5647            {            {
5648            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM47);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);
           eptr--;  
5649            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5650              eptr--;
5651              if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&
5652                  eptr[-1] == '\r') eptr--;
5653            }            }
5654          }          }
5655    
# Line 5457  switch (frame->Xwhere) Line 5688  switch (frame->Xwhere)
5688    LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(17)    LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(17)
5689    LBL(19) LBL(24) LBL(25) LBL(26) LBL(27) LBL(29) LBL(31) LBL(33)    LBL(19) LBL(24) LBL(25) LBL(26) LBL(27) LBL(29) LBL(31) LBL(33)
5690    LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52)    LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52)
5691    LBL(53) LBL(54) LBL(55) LBL(56) LBL(57) LBL(58)    LBL(53) LBL(54) LBL(55) LBL(56) LBL(57) LBL(58) LBL(63)
5692  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
5693    LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30)    LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30)
5694    LBL(32) LBL(34) LBL(42) LBL(46)    LBL(32) LBL(34) LBL(42) LBL(46)
# Line 5486  Undefine all the macros that were define Line 5717  Undefine all the macros that were define
5717  #undef ecode  #undef ecode
5718  #undef mstart  #undef mstart
5719  #undef offset_top  #undef offset_top
 #undef ims  
5720  #undef eptrb  #undef eptrb
5721  #undef flags  #undef flags
5722    
# Line 5504  Undefine all the macros that were define Line 5734  Undefine all the macros that were define
5734  #undef condition  #undef condition
5735  #undef prev_is_word  #undef prev_is_word
5736    
 #undef original_ims  
   
5737  #undef ctype  #undef ctype
5738  #undef length  #undef length
5739  #undef max  #undef max
# Line 5567  int first_byte = -1; Line 5795  int first_byte = -1;
5795  int req_byte = -1;  int req_byte = -1;
5796  int req_byte2 = -1;  int req_byte2 = -1;
5797  int newline;  int newline;
 unsigned long int ims;  
5798  BOOL using_temporary_offsets = FALSE;  BOOL using_temporary_offsets = FALSE;
5799  BOOL anchored;  BOOL anchored;
5800  BOOL startline;  BOOL startline;
# Line 5597  if ((options & ~PUBLIC_EXEC_OPTIONS) != Line 5824  if ((options & ~PUBLIC_EXEC_OPTIONS) !=
5824  if (re == NULL || subject == NULL ||  if (re == NULL || subject == NULL ||
5825     (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL;     (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL;
5826  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
5827    if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
5828    
5829  /* This information is for finding all the numbers associated with a given  /* This information is for finding all the numbers associated with a given
5830  name, for condition testing. */  name, for condition testing. */
# Line 5757  defined (though never set). So there's n Line 5985  defined (though never set). So there's n
5985  if (md->partial && (re->flags & PCRE_NOPARTIAL) != 0)  if (md->partial && (re->flags & PCRE_NOPARTIAL) != 0)
5986    return PCRE_ERROR_BADPARTIAL;    return PCRE_ERROR_BADPARTIAL;
5987    
5988  /* Check a UTF-8 string if required. Unfortunately there's no way of passing  /* Check a UTF-8 string if required. Pass back the character offset and error
5989  back the character offset. */  code for an invalid string if a results vector is available. */
5990    
5991  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
5992  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)
5993    {    {
5994    if (_pcre_valid_utf8((USPTR)subject, length) >= 0)    int erroroffset;
5995      return PCRE_ERROR_BADUTF8;    int errorcode = _pcre_valid_utf8((USPTR)subject, length, &erroroffset);
5996    if (start_offset > 0 && start_offset < length)    if (errorcode != 0)
5997      {      {
5998      int tb = ((USPTR)subject)[start_offset];      if (offsetcount >= 2)
     if (tb > 127)  
5999        {        {
6000        tb &= 0xc0;        offsets[0] = erroroffset;
6001        if (tb != 0 && tb != 0xc0) return PCRE_ERROR_BADUTF8_OFFSET;        offsets[1] = errorcode;
6002        }        }
6003      }      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?
6004          PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
6005        }
6006    
6007      /* Check that a start_offset points to the start of a UTF-8 character. */
6008    
6009      if (start_offset > 0 && start_offset < length &&
6010          (((USPTR)subject)[start_offset] & 0xc0) == 0x80)
6011        return PCRE_ERROR_BADUTF8_OFFSET;
6012    }    }
6013  #endif  #endif
6014    
 /* The ims options can vary during the matching as a result of the presence  
 of (?ims) items in the pattern. They are kept in a local variable so that  
 restoring at the exit of a group is easy. */  
   
 ims = re->options & (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL);  
   
6015  /* If the expression has got more back references than the offsets supplied can  /* If the expression has got more back references than the offsets supplied can
6016  hold, we get a temporary chunk of working store to use during the matching.  hold, we get a temporary chunk of working store to use during the matching.
6017  Otherwise, we can use the vector supplied, rounding down its size to a multiple  Otherwise, we can use the vector supplied, rounding down its size to a multiple
# Line 5900  for(;;) Line 6129  for(;;)
6129    /* There are some optimizations that avoid running the match if a known    /* There are some optimizations that avoid running the match if a known
6130    starting point is not found, or if a known later character is not present.    starting point is not found, or if a known later character is not present.
6131    However, there is an option that disables these, for testing and for ensuring    However, there is an option that disables these, for testing and for ensuring
6132    that all callouts do actually occur. */    that all callouts do actually occur. The option can be set in the regex by
6133      (*NO_START_OPT) or passed in match-time options. */
6134    
6135    if ((options & PCRE_NO_START_OPTIMIZE) == 0)    if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0)
6136      {      {
6137      /* Advance to a unique first byte if there is one. */      /* Advance to a unique first byte if there is one. */
6138    
# Line 5956  for(;;) Line 6186  for(;;)
6186        while (start_match < end_subject)        while (start_match < end_subject)
6187          {          {
6188          register unsigned int c = *start_match;          register unsigned int c = *start_match;
6189          if ((start_bits[c/8] & (1 << (c&7))) == 0) start_match++;          if ((start_bits[c/8] & (1 << (c&7))) == 0)
6190            else break;            {
6191              start_match++;
6192    #ifdef SUPPORT_UTF8
6193              if (utf8)
6194                while(start_match < end_subject && (*start_match & 0xc0) == 0x80)
6195                  start_match++;
6196    #endif
6197              }
6198            else break;
6199          }          }
6200        }        }
6201      }   /* Starting optimizations */      }   /* Starting optimizations */
# Line 6052  for(;;) Line 6290  for(;;)
6290    md->start_match_ptr = start_match;    md->start_match_ptr = start_match;
6291    md->start_used_ptr = start_match;    md->start_used_ptr = start_match;
6292    md->match_call_count = 0;    md->match_call_count = 0;
6293    rc = match(start_match, md->start_code, start_match, NULL, 2, md, ims, NULL,    md->match_function_type = 0;
6294      0, 0);    rc = match(start_match, md->start_code, start_match, NULL, 2, md, NULL, 0);
6295    if (md->hitend && start_partial == NULL) start_partial = md->start_used_ptr;    if (md->hitend && start_partial == NULL) start_partial = md->start_used_ptr;
6296    
6297    switch(rc)    switch(rc)
6298      {      {
6299      /* NOMATCH and PRUNE advance by one character. If MATCH_SKIP_ARG reaches      /* SKIP passes back the next starting point explicitly, but if it is the
6300      this level it means that a MARK that matched the SKIP's arg was not found.      same as the match we have just done, treat it as NOMATCH. */
6301      We treat this as NOMATCH. THEN at this level acts exactly like PRUNE. */  
6302        case MATCH_SKIP:
6303        if (md->start_match_ptr != start_match)
6304          {
6305          new_start_match = md->start_match_ptr;
6306          break;
6307          }
6308        /* Fall through */
6309    
6310        /* If MATCH_SKIP_ARG reaches this level it means that a MARK that matched
6311        the SKIP's arg was not found. We also treat this as NOMATCH. */
6312    
6313        case MATCH_SKIP_ARG:
6314        /* Fall through */
6315    
6316        /* NOMATCH and PRUNE advance by one character. THEN at this level acts
6317        exactly like PRUNE. */
6318    
6319      case MATCH_NOMATCH:      case MATCH_NOMATCH:
6320      case MATCH_PRUNE:      case MATCH_PRUNE:
     case MATCH_SKIP_ARG:  
6321      case MATCH_THEN:      case MATCH_THEN:
6322      new_start_match = start_match + 1;      new_start_match = start_match + 1;
6323  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 6074  for(;;) Line 6327  for(;;)
6327  #endif  #endif
6328      break;      break;
6329    
     /* SKIP passes back the next starting point explicitly. */  
   
     case MATCH_SKIP:  
     new_start_match = md->start_match_ptr;  
     break;  
   
6330      /* COMMIT disables the bumpalong, but otherwise behaves as NOMATCH. */      /* COMMIT disables the bumpalong, but otherwise behaves as NOMATCH. */
6331    
6332      case MATCH_COMMIT:      case MATCH_COMMIT:
# Line 6167  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 6414  if (rc == MATCH_MATCH || rc == MATCH_ACC
6414    
6415    /* Set the return code to the number of captured strings, or 0 if there are    /* Set the return code to the number of captured strings, or 0 if there are
6416    too many to fit into the vector. */    too many to fit into the vector. */
6417    
6418    rc = md->offset_overflow? 0 : md->end_offset_top/2;    rc = md->offset_overflow? 0 : md->end_offset_top/2;
6419    
6420    /* If there is space, set up the whole thing as substring 0. The value of    /* If there is space, set up the whole thing as substring 0. The value of
# Line 6176  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 6423  if (rc == MATCH_MATCH || rc == MATCH_ACC
6423    
6424    if (offsetcount < 2) rc = 0; else    if (offsetcount < 2) rc = 0; else
6425      {      {
6426      offsets[0] = md->start_match_ptr - md->start_subject;      offsets[0] = (int)(md->start_match_ptr - md->start_subject);
6427      offsets[1] = md->end_match_ptr - md->start_subject;      offsets[1] = (int)(md->end_match_ptr - md->start_subject);
6428      }      }
6429    
6430    DPRINTF((">>>> returning %d\n", rc));    DPRINTF((">>>> returning %d\n", rc));
# Line 6209  if (start_partial != NULL) Line 6456  if (start_partial != NULL)
6456    md->mark = NULL;    md->mark = NULL;
6457    if (offsetcount > 1)    if (offsetcount > 1)
6458      {      {
6459      offsets[0] = start_partial - (USPTR)subject;      offsets[0] = (int)(start_partial - (USPTR)subject);
6460      offsets[1] = end_subject - (USPTR)subject;      offsets[1] = (int)(end_subject - (USPTR)subject);
6461      }      }
6462    rc = PCRE_ERROR_PARTIAL;    rc = PCRE_ERROR_PARTIAL;
6463    }    }

Legend:
Removed from v.527  
changed lines
  Added in v.609

  ViewVC Help
Powered by ViewVC 1.1.5