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

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

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

code/trunk/pcre_exec.c revision 597 by ph10, Mon May 2 17:08:52 2011 UTC code/branches/pcre16/pcre_exec.c revision 786 by zherczeg, Tue Dec 6 11:33:41 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_ONCE         (-996)
80  #define MATCH_SKIP_ARG     (-995)  #define MATCH_PRUNE        (-995)
81  #define MATCH_THEN         (-994)  #define MATCH_SKIP         (-994)
82    #define MATCH_SKIP_ARG     (-993)
83    #define MATCH_THEN         (-992)
84    
85  /* This is a convenience macro for code that occurs many times. */  /* This is a convenience macro for code that occurs many times. */
86    
# Line 117  Returns:     nothing Line 121  Returns:     nothing
121  */  */
122    
123  static void  static void
124  pchars(const uschar *p, int length, BOOL is_subject, match_data *md)  pchars(const pcre_uchar *p, int length, BOOL is_subject, match_data *md)
125  {  {
126  unsigned int c;  unsigned int c;
127  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;
# Line 134  while (length-- > 0) Line 138  while (length-- > 0)
138    
139  /* Normally, if a back reference hasn't been set, the length that is passed is  /* Normally, if a back reference hasn't been set, the length that is passed is
140  negative, so the match always fails. However, in JavaScript compatibility mode,  negative, so the match always fails. However, in JavaScript compatibility mode,
141  the length passed is zero. Note that in caseless UTF-8 mode, the number of  the length passed is zero. Note that in caseless UTF-8 mode, the number of
142  subject bytes matched may be different to the number of reference bytes.  subject bytes matched may be different to the number of reference bytes.
143    
144  Arguments:  Arguments:
# Line 142  Arguments: Line 146  Arguments:
146    eptr        pointer into the subject    eptr        pointer into the subject
147    length      length of reference to be matched (number of bytes)    length      length of reference to be matched (number of bytes)
148    md          points to match data block    md          points to match data block
149    ims         the ims flags    caseless    TRUE if caseless
150    
151  Returns:      < 0 if not matched, otherwise the number of subject bytes matched  Returns:      < 0 if not matched, otherwise the number of subject bytes matched
152  */  */
153    
154  static int  static int
155  match_ref(int offset, register USPTR eptr, int length, match_data *md,  match_ref(int offset, register PCRE_PUCHAR eptr, int length, match_data *md,
156    unsigned long int ims)    BOOL caseless)
157  {  {
158  USPTR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
159  register USPTR p = md->start_subject + md->offset_vector[offset];  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
160    
161  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
162  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 175  if (length < 0) return -1; Line 179  if (length < 0) return -1;
179  properly if Unicode properties are supported. Otherwise, we can check only  properly if Unicode properties are supported. Otherwise, we can check only
180  ASCII characters. */  ASCII characters. */
181    
182  if ((ims & PCRE_CASELESS) != 0)  if (caseless)
183    {    {
184  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
185  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
186    if (md->utf8)    if (md->utf)
187      {      {
188      /* Match characters up to the end of the reference. NOTE: the number of      /* Match characters up to the end of the reference. NOTE: the number of
189      bytes matched may differ, because there are some characters whose upper and      bytes matched may differ, because there are some characters whose upper and
190      lower case versions code as different numbers of bytes. For example, U+023A      lower case versions code as different numbers of bytes. For example, U+023A
191      (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);      (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);
192      a sequence of 3 of the former uses 6 bytes, as does a sequence of two of      a sequence of 3 of the former uses 6 bytes, as does a sequence of two of
193      the latter. It is important, therefore, to check the length along the      the latter. It is important, therefore, to check the length along the
194      reference, not along the subject (earlier code did this wrong). */      reference, not along the subject (earlier code did this wrong). */
195    
196      USPTR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
197      while (p < endptr)      while (p < endptr)
198        {        {
199        int c, d;        int c, d;
# Line 206  if ((ims & PCRE_CASELESS) != 0) Line 210  if ((ims & PCRE_CASELESS) != 0)
210    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there
211    is no UCP support. */    is no UCP support. */
212      {      {
213      if (eptr + length > md->end_subject) return -1;      if (eptr + length > md->end_subject) return -1;
214      while (length-- > 0)      while (length-- > 0)
215        { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }        { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }
216      }      }
217    }    }
218    
219  /* In the caseful case, we can just compare the bytes, whether or not we  /* In the caseful case, we can just compare the bytes, whether or not we
220  are in UTF-8 mode. */  are in UTF-8 mode. */
221    
222  else  else
223    {    {
224    if (eptr + length > md->end_subject) return -1;    if (eptr + length > md->end_subject) return -1;
225    while (length-- > 0) if (*p++ != *eptr++) return -1;    while (length-- > 0) if (*p++ != *eptr++) return -1;
226    }    }
227    
228  return eptr - eptr_start;  return eptr - eptr_start;
# Line 273  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 277  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
277         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
278         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
279         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
280         RM61,  RM62 };         RM61,  RM62, RM63, RM64, RM65, RM66 };
281    
282  /* These versions of the macros use the stack, as normal. There are debugging  /* These versions of the macros use the stack, as normal. There are debugging
283  versions and production versions. Note that the "rw" argument of RMATCH isn't  versions and production versions. Note that the "rw" argument of RMATCH isn't
# Line 283  actually used in this definition. */ Line 287  actually used in this definition. */
287  #define REGISTER register  #define REGISTER register
288    
289  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
290  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
291    { \    { \
292    printf("match() called in line %d\n", __LINE__); \    printf("match() called in line %d\n", __LINE__); \
293    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rf,rg,rdepth+1); \    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1); \
294    printf("to line %d\n", __LINE__); \    printf("to line %d\n", __LINE__); \
295    }    }
296  #define RRETURN(ra) \  #define RRETURN(ra) \
# Line 295  actually used in this definition. */ Line 299  actually used in this definition. */
299    return ra; \    return ra; \
300    }    }
301  #else  #else
302  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
303    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rf,rg,rdepth+1)    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1)
304  #define RRETURN(ra) return ra  #define RRETURN(ra) return ra
305  #endif  #endif
306    
# Line 309  argument of match(), which never changes Line 313  argument of match(), which never changes
313    
314  #define REGISTER  #define REGISTER
315    
316  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
317    {\    {\
318    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\
319    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
# Line 319  argument of match(), which never changes Line 323  argument of match(), which never changes
323    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
324    newframe->Xmarkptr = markptr;\    newframe->Xmarkptr = markptr;\
325    newframe->Xoffset_top = rc;\    newframe->Xoffset_top = rc;\
326    newframe->Xims = re;\    newframe->Xeptrb = re;\
   newframe->Xeptrb = rf;\  
   newframe->Xflags = rg;\  
327    newframe->Xrdepth = frame->Xrdepth + 1;\    newframe->Xrdepth = frame->Xrdepth + 1;\
328    newframe->Xprevframe = frame;\    newframe->Xprevframe = frame;\
329    frame = newframe;\    frame = newframe;\
# Line 352  typedef struct heapframe { Line 354  typedef struct heapframe {
354    
355    /* Function arguments that may change */    /* Function arguments that may change */
356    
357    USPTR Xeptr;    PCRE_PUCHAR Xeptr;
358    const uschar *Xecode;    const pcre_uchar *Xecode;
359    USPTR Xmstart;    PCRE_PUCHAR Xmstart;
360    USPTR Xmarkptr;    PCRE_PUCHAR Xmarkptr;
361    int Xoffset_top;    int Xoffset_top;
   long int Xims;  
362    eptrblock *Xeptrb;    eptrblock *Xeptrb;
   int Xflags;  
363    unsigned int Xrdepth;    unsigned int Xrdepth;
364    
365    /* Function local variables */    /* Function local variables */
366    
367    USPTR Xcallpat;    PCRE_PUCHAR Xcallpat;
368  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
369    USPTR Xcharptr;    PCRE_PUCHAR Xcharptr;
370  #endif  #endif
371    USPTR Xdata;    PCRE_PUCHAR Xdata;
372    USPTR Xnext;    PCRE_PUCHAR Xnext;
373    USPTR Xpp;    PCRE_PUCHAR Xpp;
374    USPTR Xprev;    PCRE_PUCHAR Xprev;
375    USPTR Xsaved_eptr;    PCRE_PUCHAR Xsaved_eptr;
376    
377    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
378    
# Line 380  typedef struct heapframe { Line 380  typedef struct heapframe {
380    BOOL Xcondition;    BOOL Xcondition;
381    BOOL Xprev_is_word;    BOOL Xprev_is_word;
382    
   unsigned long int Xoriginal_ims;  
   
383  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
384    int Xprop_type;    int Xprop_type;
385    int Xprop_value;    int Xprop_value;
386    int Xprop_fail_result;    int Xprop_fail_result;
   int Xprop_category;  
   int Xprop_chartype;  
   int Xprop_script;  
387    int Xoclength;    int Xoclength;
388    uschar Xocchars[8];    pcre_uchar Xocchars[6];
389  #endif  #endif
390    
391    int Xcodelink;    int Xcodelink;
# Line 455  the subject. */ Line 450  the subject. */
450    
451    
452  /* Performance note: It might be tempting to extract commonly used fields from  /* Performance note: It might be tempting to extract commonly used fields from
453  the md structure (e.g. utf8, end_subject) into individual variables to improve  the md structure (e.g. utf, end_subject) into individual variables to improve
454  performance. Tests using gcc on a SPARC disproved this; in the first case, it  performance. Tests using gcc on a SPARC disproved this; in the first case, it
455  made performance worse.  made performance worse.
456    
# Line 467  Arguments: Line 462  Arguments:
462     markptr     pointer to the most recent MARK name, or NULL     markptr     pointer to the most recent MARK name, or NULL
463     offset_top  current top pointer     offset_top  current top pointer
464     md          pointer to "static" info for the match     md          pointer to "static" info for the match
    ims         current /i, /m, and /s options  
465     eptrb       pointer to chain of blocks containing eptr at start of     eptrb       pointer to chain of blocks containing eptr at start of
466                   brackets - for testing for empty matches                   brackets - for testing for empty matches
    flags       can contain  
                  match_condassert - this is an assertion condition  
                  match_cbegroup - this is the start of an unlimited repeat  
                    group that can match an empty string  
467     rdepth      the recursion depth     rdepth      the recursion depth
468    
469  Returns:       MATCH_MATCH if matched            )  these values are >= 0  Returns:       MATCH_MATCH if matched            )  these values are >= 0
# Line 484  Returns:       MATCH_MATCH if matched Line 474  Returns:       MATCH_MATCH if matched
474  */  */
475    
476  static int  static int
477  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode,
478    const uschar *markptr, int offset_top, match_data *md, unsigned long int ims,    PCRE_PUCHAR mstart, const pcre_uchar *markptr, int offset_top,
479    eptrblock *eptrb, int flags, unsigned int rdepth)    match_data *md, eptrblock *eptrb, unsigned int rdepth)
480  {  {
481  /* These variables do not need to be preserved over recursion in this function,  /* These variables do not need to be preserved over recursion in this function,
482  so they can be ordinary variables in all cases. Mark some of them with  so they can be ordinary variables in all cases. Mark some of them with
# Line 495  so they can be ordinary variables in all Line 485  so they can be ordinary variables in all
485  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
486  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
487  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register unsigned int c;   /* Character values not kept over RMATCH() calls */
488  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
489    
490  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
491    BOOL caseless;
492  int condcode;  int condcode;
493    
494  /* When recursion is not being used, all "local" variables that have to be  /* When recursion is not being used, all "local" variables that have to be
# Line 517  frame->Xecode = ecode; Line 508  frame->Xecode = ecode;
508  frame->Xmstart = mstart;  frame->Xmstart = mstart;
509  frame->Xmarkptr = markptr;  frame->Xmarkptr = markptr;
510  frame->Xoffset_top = offset_top;  frame->Xoffset_top = offset_top;
 frame->Xims = ims;  
511  frame->Xeptrb = eptrb;  frame->Xeptrb = eptrb;
 frame->Xflags = flags;  
512  frame->Xrdepth = rdepth;  frame->Xrdepth = rdepth;
513    
514  /* This is where control jumps back to to effect "recursion" */  /* This is where control jumps back to to effect "recursion" */
# Line 533  HEAP_RECURSE: Line 522  HEAP_RECURSE:
522  #define mstart             frame->Xmstart  #define mstart             frame->Xmstart
523  #define markptr            frame->Xmarkptr  #define markptr            frame->Xmarkptr
524  #define offset_top         frame->Xoffset_top  #define offset_top         frame->Xoffset_top
 #define ims                frame->Xims  
525  #define eptrb              frame->Xeptrb  #define eptrb              frame->Xeptrb
 #define flags              frame->Xflags  
526  #define rdepth             frame->Xrdepth  #define rdepth             frame->Xrdepth
527    
528  /* Ditto for the local variables */  /* Ditto for the local variables */
# Line 557  HEAP_RECURSE: Line 544  HEAP_RECURSE:
544  #define condition          frame->Xcondition  #define condition          frame->Xcondition
545  #define prev_is_word       frame->Xprev_is_word  #define prev_is_word       frame->Xprev_is_word
546    
 #define original_ims       frame->Xoriginal_ims  
   
547  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
548  #define prop_type          frame->Xprop_type  #define prop_type          frame->Xprop_type
549  #define prop_value         frame->Xprop_value  #define prop_value         frame->Xprop_value
550  #define prop_fail_result   frame->Xprop_fail_result  #define prop_fail_result   frame->Xprop_fail_result
 #define prop_category      frame->Xprop_category  
 #define prop_chartype      frame->Xprop_chartype  
 #define prop_script        frame->Xprop_script  
551  #define oclength           frame->Xoclength  #define oclength           frame->Xoclength
552  #define occhars            frame->Xocchars  #define occhars            frame->Xocchars
553  #endif  #endif
# Line 595  i, and fc and c, can be the same variabl Line 577  i, and fc and c, can be the same variabl
577  #define fi i  #define fi i
578  #define fc c  #define fc c
579    
580    /* Many of the following variables are used only in small blocks of the code.
581    My normal style of coding would have declared them within each of those blocks.
582    However, in order to accommodate the version of this code that uses an external
583    "stack" implemented on the heap, it is easier to declare them all here, so the
584    declarations can be cut out in a block. The only declarations within blocks
585    below are for variables that do not have to be preserved over a recursive call
586    to RMATCH(). */
587    
588    #ifdef SUPPORT_UTF8
589    const pcre_uchar *charptr;
590    #endif
591    const pcre_uchar *callpat;
592    const pcre_uchar *data;
593    const pcre_uchar *next;
594    PCRE_PUCHAR       pp;
595    const pcre_uchar *prev;
596    PCRE_PUCHAR       saved_eptr;
597    
598  #ifdef SUPPORT_UTF8                /* Many of these variables are used only  */  recursion_info new_recursive;
599  const uschar *charptr;             /* in small blocks of the code. My normal */  
600  #endif                             /* style of coding would have declared    */  BOOL cur_is_word;
 const uschar *callpat;             /* them within each of those blocks.      */  
 const uschar *data;                /* However, in order to accommodate the   */  
 const uschar *next;                /* version of this code that uses an      */  
 USPTR         pp;                  /* external "stack" implemented on the    */  
 const uschar *prev;                /* heap, it is easier to declare them all */  
 USPTR         saved_eptr;          /* here, so the declarations can be cut   */  
                                    /* out in a block. The only declarations  */  
 recursion_info new_recursive;      /* within blocks below are for variables  */  
                                    /* that do not have to be preserved over  */  
 BOOL cur_is_word;                  /* a recursive call to RMATCH().          */  
601  BOOL condition;  BOOL condition;
602  BOOL prev_is_word;  BOOL prev_is_word;
603    
 unsigned long int original_ims;  
   
604  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
605  int prop_type;  int prop_type;
606  int prop_value;  int prop_value;
607  int prop_fail_result;  int prop_fail_result;
 int prop_category;  
 int prop_chartype;  
 int prop_script;  
608  int oclength;  int oclength;
609  uschar occhars[8];  pcre_uchar occhars[6];
610  #endif  #endif
611    
612  int codelink;  int codelink;
# Line 640  int stacksave[REC_STACK_SAVE_MAX]; Line 624  int stacksave[REC_STACK_SAVE_MAX];
624  eptrblock newptrb;  eptrblock newptrb;
625  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
626    
627    /* To save space on the stack and in the heap frame, I have doubled up on some
628    of the local variables that are used only in localised parts of the code, but
629    still need to be preserved over recursive calls of match(). These macros define
630    the alternative names that are used. */
631    
632    #define allow_zero    cur_is_word
633    #define cbegroup      condition
634    #define code_offset   codelink
635    #define condassert    condition
636    #define matched_once  prev_is_word
637    
638  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
639  variables. */  variables. */
640    
# Line 665  complicated macro. It has to be used in Line 660  complicated macro. It has to be used in
660  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
661    
662  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
663  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
664  #else  #else
665  utf8 = FALSE;  utf = FALSE;
666  #endif  #endif
667    
668  /* First check that we haven't called match() too many times, or that we  /* First check that we haven't called match() too many times, or that we
# Line 676  haven't exceeded the recursive call limi Line 671  haven't exceeded the recursive call limi
671  if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT);  if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT);
672  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);
673    
 original_ims = ims;    /* Save for resetting on ')' */  
   
674  /* At the start of a group with an unlimited repeat that may match an empty  /* At the start of a group with an unlimited repeat that may match an empty
675  string, the match_cbegroup flag is set. When this is the case, add the current  string, the variable md->match_function_type is set to MATCH_CBEGROUP. It is
676  subject pointer to the chain of such remembered pointers, to be checked when we  done this way to save having to use another function argument, which would take
677  hit the closing ket, in order to break infinite loops that match no characters.  up space on the stack. See also MATCH_CONDASSERT below.
678  When match() is called in other circumstances, don't add to the chain. The  
679  match_cbegroup flag must NOT be used with tail recursion, because the memory  When MATCH_CBEGROUP is set, add the current subject pointer to the chain of
680  block that is used is on the stack, so a new one may be required for each  such remembered pointers, to be checked when we hit the closing ket, in order
681  match(). */  to break infinite loops that match no characters. When match() is called in
682    other circumstances, don't add to the chain. The MATCH_CBEGROUP feature must
683    NOT be used with tail recursion, because the memory block that is used is on
684    the stack, so a new one may be required for each match(). */
685    
686  if ((flags & match_cbegroup) != 0)  if (md->match_function_type == MATCH_CBEGROUP)
687    {    {
688    newptrb.epb_saved_eptr = eptr;    newptrb.epb_saved_eptr = eptr;
689    newptrb.epb_prev = eptrb;    newptrb.epb_prev = eptrb;
690    eptrb = &newptrb;    eptrb = &newptrb;
691      md->match_function_type = 0;
692    }    }
693    
694  /* Now start processing the opcodes. */  /* Now start processing the opcodes. */
# Line 705  for (;;) Line 702  for (;;)
702      {      {
703      case OP_MARK:      case OP_MARK:
704      markptr = ecode + 2;      markptr = ecode + 2;
705      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
706        ims, eptrb, flags, RM55);        eptrb, RM55);
707    
708      /* 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
709      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 716  for (;;) Line 713  for (;;)
713      unaltered. */      unaltered. */
714    
715      if (rrc == MATCH_SKIP_ARG &&      if (rrc == MATCH_SKIP_ARG &&
716          strcmp((char *)markptr, (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC(markptr, md->start_match_ptr) == 0)
717        {        {
718        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
719        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
# Line 731  for (;;) Line 728  for (;;)
728      /* COMMIT overrides PRUNE, SKIP, and THEN */      /* COMMIT overrides PRUNE, SKIP, and THEN */
729    
730      case OP_COMMIT:      case OP_COMMIT:
731      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
732        ims, eptrb, flags, RM52);        eptrb, RM52);
733      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&
734          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&
735          rrc != MATCH_THEN)          rrc != MATCH_THEN)
# Line 742  for (;;) Line 739  for (;;)
739      /* PRUNE overrides THEN */      /* PRUNE overrides THEN */
740    
741      case OP_PRUNE:      case OP_PRUNE:
742      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
743        ims, eptrb, flags, RM51);        eptrb, RM51);
744      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
745      MRRETURN(MATCH_PRUNE);      MRRETURN(MATCH_PRUNE);
746    
747      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
748      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
749        ims, eptrb, flags, RM56);        eptrb, RM56);
750      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
751      md->mark = ecode + 2;      md->mark = ecode + 2;
752      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
# Line 757  for (;;) Line 754  for (;;)
754      /* SKIP overrides PRUNE and THEN */      /* SKIP overrides PRUNE and THEN */
755    
756      case OP_SKIP:      case OP_SKIP:
757      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
758        ims, eptrb, flags, RM53);        eptrb, RM53);
759      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
760        RRETURN(rrc);        RRETURN(rrc);
761      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
762      MRRETURN(MATCH_SKIP);      MRRETURN(MATCH_SKIP);
763    
764      case OP_SKIP_ARG:      case OP_SKIP_ARG:
765      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
766        ims, eptrb, flags, RM57);        eptrb, RM57);
767      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
768        RRETURN(rrc);        RRETURN(rrc);
769    
# Line 778  for (;;) Line 775  for (;;)
775      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
776      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
777    
778      /* For THEN (and THEN_ARG) we pass back the address of the bracket or      /* For THEN (and THEN_ARG) we pass back the address of the opcode, so that
779      the alt that is at the start of the current branch. This makes it possible      the branch in which it occurs can be determined. Overload the start of
780      to skip back past alternatives that precede the THEN within the current      match pointer to do this. */
     branch. */  
781    
782      case OP_THEN:      case OP_THEN:
783      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
784        ims, eptrb, flags, RM54);        eptrb, RM54);
785      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
786      md->start_match_ptr = ecode - GET(ecode, 1);      md->start_match_ptr = ecode;
787      MRRETURN(MATCH_THEN);      MRRETURN(MATCH_THEN);
788    
789      case OP_THEN_ARG:      case OP_THEN_ARG:
790      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1+LINK_SIZE],      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
791        offset_top, md, ims, eptrb, flags, RM58);        md, eptrb, RM58);
792      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
793      md->start_match_ptr = ecode - GET(ecode, 1);      md->start_match_ptr = ecode;
794      md->mark = ecode + LINK_SIZE + 2;      md->mark = ecode + 2;
795      RRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
796    
797      /* Handle a capturing bracket. If there is space in the offset vector, save      /* Handle an atomic group that does not contain any capturing parentheses.
798      the current subject position in the working slot at the top of the vector.      This can be handled like an assertion. Prior to 8.13, all atomic groups
799      We mustn't change the current values of the data slot, because they may be      were handled this way. In 8.13, the code was changed as below for ONCE, so
800      set from a previous iteration of this group, and be referred to by a      that backups pass through the group and thereby reset captured values.
801      reference inside the group.      However, this uses a lot more stack, so in 8.20, atomic groups that do not
802        contain any captures generate OP_ONCE_NC, which can be handled in the old,
803      If the bracket fails to match, we need to restore this value and also the      less stack intensive way.
804      values of the final offsets, in case they were set by a previous iteration  
805      of the same bracket.      Check the alternative branches in turn - the matching won't pass the KET
806        for this kind of subpattern. If any one branch matches, we carry on as at
807        the end of a normal bracket, leaving the subject pointer, but resetting
808        the start-of-match value in case it was changed by \K. */
809    
810        case OP_ONCE_NC:
811        prev = ecode;
812        saved_eptr = eptr;
813        do
814          {
815          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
816          if (rrc == MATCH_MATCH)  /* Note: _not_ MATCH_ACCEPT */
817            {
818            mstart = md->start_match_ptr;
819            markptr = md->mark;
820            break;
821            }
822          if (rrc == MATCH_THEN)
823            {
824            next = ecode + GET(ecode,1);
825            if (md->start_match_ptr < next &&
826                (*ecode == OP_ALT || *next == OP_ALT))
827              rrc = MATCH_NOMATCH;
828            }
829    
830          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
831          ecode += GET(ecode,1);
832          }
833        while (*ecode == OP_ALT);
834    
835        /* If hit the end of the group (which could be repeated), fail */
836    
837        if (*ecode != OP_ONCE_NC && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);
838    
839        /* Continue as from after the group, updating the offsets high water
840        mark, since extracts may have been taken. */
841    
842        do ecode += GET(ecode, 1); while (*ecode == OP_ALT);
843    
844        offset_top = md->end_offset_top;
845        eptr = md->end_match_ptr;
846    
847        /* For a non-repeating ket, just continue at this level. This also
848        happens for a repeating ket if no characters were matched in the group.
849        This is the forcible breaking of infinite loops as implemented in Perl
850        5.005. */
851    
852        if (*ecode == OP_KET || eptr == saved_eptr)
853          {
854          ecode += 1+LINK_SIZE;
855          break;
856          }
857    
858        /* The repeating kets try the rest of the pattern or restart from the
859        preceding bracket, in the appropriate order. The second "call" of match()
860        uses tail recursion, to avoid using another stack frame. */
861    
862        if (*ecode == OP_KETRMIN)
863          {
864          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM65);
865          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
866          ecode = prev;
867          goto TAIL_RECURSE;
868          }
869        else  /* OP_KETRMAX */
870          {
871          md->match_function_type = MATCH_CBEGROUP;
872          RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
873          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
874          ecode += 1 + LINK_SIZE;
875          goto TAIL_RECURSE;
876          }
877        /* Control never gets here */
878    
879        /* Handle a capturing bracket, other than those that are possessive with an
880        unlimited repeat. If there is space in the offset vector, save the current
881        subject position in the working slot at the top of the vector. We mustn't
882        change the current values of the data slot, because they may be set from a
883        previous iteration of this group, and be referred to by a reference inside
884        the group. A failure to match might occur after the group has succeeded,
885        if something later on doesn't match. For this reason, we need to restore
886        the working value and also the values of the final offsets, in case they
887        were set by a previous iteration of the same bracket.
888    
889      If there isn't enough space in the offset vector, treat this as if it were      If there isn't enough space in the offset vector, treat this as if it were
890      a non-capturing bracket. Don't worry about setting the flag for the error      a non-capturing bracket. Don't worry about setting the flag for the error
# Line 835  for (;;) Line 913  for (;;)
913        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
914          (int)(eptr - md->start_subject);          (int)(eptr - md->start_subject);
915    
916        flags = (op == OP_SCBRA)? match_cbegroup : 0;        for (;;)
       do  
917          {          {
918          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
919            ims, eptrb, flags, RM1);          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
920          if (rrc != MATCH_NOMATCH &&            eptrb, RM1);
921              (rrc != MATCH_THEN || md->start_match_ptr != ecode))          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
922            RRETURN(rrc);  
923            /* If we backed up to a THEN, check whether it is within the current
924            branch by comparing the address of the THEN that is passed back with
925            the end of the branch. If it is within the current branch, and the
926            branch is one of two or more alternatives (it either starts or ends
927            with OP_ALT), we have reached the limit of THEN's action, so convert
928            the return code to NOMATCH, which will cause normal backtracking to
929            happen from now on. Otherwise, THEN is passed back to an outer
930            alternative. This implements Perl's treatment of parenthesized groups,
931            where a group not containing | does not affect the current alternative,
932            that is, (X) is NOT the same as (X|(*F)). */
933    
934            if (rrc == MATCH_THEN)
935              {
936              next = ecode + GET(ecode,1);
937              if (md->start_match_ptr < next &&
938                  (*ecode == OP_ALT || *next == OP_ALT))
939                rrc = MATCH_NOMATCH;
940              }
941    
942            /* Anything other than NOMATCH is passed back. */
943    
944            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
945          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
946          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
947            if (*ecode != OP_ALT) break;
948          }          }
       while (*ecode == OP_ALT);  
949    
950        DPRINTF(("bracket %d failed\n", number));        DPRINTF(("bracket %d failed\n", number));
   
951        md->offset_vector[offset] = save_offset1;        md->offset_vector[offset] = save_offset1;
952        md->offset_vector[offset+1] = save_offset2;        md->offset_vector[offset+1] = save_offset2;
953        md->offset_vector[md->offset_end - number] = save_offset3;        md->offset_vector[md->offset_end - number] = save_offset3;
954    
955        if (rrc != MATCH_THEN) md->mark = markptr;        /* At this point, rrc will be one of MATCH_ONCE or MATCH_NOMATCH. */
956        RRETURN(MATCH_NOMATCH);  
957          if (md->mark == NULL) md->mark = markptr;
958          RRETURN(rrc);
959        }        }
960    
961      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
# Line 869  for (;;) Line 969  for (;;)
969      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
970      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
971    
972      /* Non-capturing bracket. Loop for all the alternatives. When we get to the      /* Non-capturing or atomic group, except for possessive with unlimited
973      final alternative within the brackets, we would return the result of a      repeat and ONCE group with no captures. Loop for all the alternatives.
     recursive call to match() whatever happened. We can reduce stack usage by  
     turning this into a tail recursion, except in the case when match_cbegroup  
     is set.*/  
974    
975        When we get to the final alternative within the brackets, we used to return
976        the result of a recursive call to match() whatever happened so it was
977        possible to reduce stack usage by turning this into a tail recursion,
978        except in the case of a possibly empty group. However, now that there is
979        the possiblity of (*THEN) occurring in the final alternative, this
980        optimization is no longer always possible.
981    
982        We can optimize if we know there are no (*THEN)s in the pattern; at present
983        this is the best that can be done.
984    
985        MATCH_ONCE is returned when the end of an atomic group is successfully
986        reached, but subsequent matching fails. It passes back up the tree (causing
987        captured values to be reset) until the original atomic group level is
988        reached. This is tested by comparing md->once_target with the start of the
989        group. At this point, the return is converted into MATCH_NOMATCH so that
990        previous backup points can be taken. */
991    
992        case OP_ONCE:
993      case OP_BRA:      case OP_BRA:
994      case OP_SBRA:      case OP_SBRA:
995      DPRINTF(("start non-capturing bracket\n"));      DPRINTF(("start non-capturing bracket\n"));
996      flags = (op >= OP_SBRA)? match_cbegroup : 0;  
997      for (;;)      for (;;)
998        {        {
999        if (ecode[GET(ecode, 1)] != OP_ALT)   /* Final alternative */        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;
1000    
1001          /* If this is not a possibly empty group, and there are no (*THEN)s in
1002          the pattern, and this is the final alternative, optimize as described
1003          above. */
1004    
1005          else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1006            {
1007            ecode += PRIV(OP_lengths)[*ecode];
1008            goto TAIL_RECURSE;
1009            }
1010    
1011          /* In all other cases, we have to make another call to match(). */
1012    
1013          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1014            RM2);
1015    
1016          /* See comment in the code for capturing groups above about handling
1017          THEN. */
1018    
1019          if (rrc == MATCH_THEN)
1020            {
1021            next = ecode + GET(ecode,1);
1022            if (md->start_match_ptr < next &&
1023                (*ecode == OP_ALT || *next == OP_ALT))
1024              rrc = MATCH_NOMATCH;
1025            }
1026    
1027          if (rrc != MATCH_NOMATCH)
1028          {          {
1029          if (flags == 0)    /* Not a possibly empty group */          if (rrc == MATCH_ONCE)
1030            {            {
1031            ecode += _pcre_OP_lengths[*ecode];            const pcre_uchar *scode = ecode;
1032            DPRINTF(("bracket 0 tail recursion\n"));            if (*scode != OP_ONCE)           /* If not at start, find it */
1033            goto TAIL_RECURSE;              {
1034                while (*scode == OP_ALT) scode += GET(scode, 1);
1035                scode -= GET(scode, 1);
1036                }
1037              if (md->once_target == scode) rrc = MATCH_NOMATCH;
1038            }            }
1039            RRETURN(rrc);
1040            }
1041          ecode += GET(ecode, 1);
1042          if (*ecode != OP_ALT) break;
1043          }
1044    
1045        if (md->mark == NULL) md->mark = markptr;
1046        RRETURN(MATCH_NOMATCH);
1047    
1048          /* Possibly empty group; can't use tail recursion. */      /* Handle possessive capturing brackets with an unlimited repeat. We come
1049        here from BRAZERO with allow_zero set TRUE. The offset_vector values are
1050        handled similarly to the normal case above. However, the matching is
1051        different. The end of these brackets will always be OP_KETRPOS, which
1052        returns MATCH_KETRPOS without going further in the pattern. By this means
1053        we can handle the group by iteration rather than recursion, thereby
1054        reducing the amount of stack needed. */
1055    
1056        case OP_CBRAPOS:
1057        case OP_SCBRAPOS:
1058        allow_zero = FALSE;
1059    
1060          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,      POSSESSIVE_CAPTURE:
1061            eptrb, flags, RM48);      number = GET2(ecode, 1+LINK_SIZE);
1062          if (rrc == MATCH_NOMATCH) md->mark = markptr;      offset = number << 1;
1063          RRETURN(rrc);  
1064    #ifdef PCRE_DEBUG
1065        printf("start possessive bracket %d\n", number);
1066        printf("subject=");
1067        pchars(eptr, 16, TRUE, md);
1068        printf("\n");
1069    #endif
1070    
1071        if (offset < md->offset_max)
1072          {
1073          matched_once = FALSE;
1074          code_offset = ecode - md->start_code;
1075    
1076          save_offset1 = md->offset_vector[offset];
1077          save_offset2 = md->offset_vector[offset+1];
1078          save_offset3 = md->offset_vector[md->offset_end - number];
1079          save_capture_last = md->capture_last;
1080    
1081          DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
1082    
1083          /* Each time round the loop, save the current subject position for use
1084          when the group matches. For MATCH_MATCH, the group has matched, so we
1085          restart it with a new subject starting position, remembering that we had
1086          at least one match. For MATCH_NOMATCH, carry on with the alternatives, as
1087          usual. If we haven't matched any alternatives in any iteration, check to
1088          see if a previous iteration matched. If so, the group has matched;
1089          continue from afterwards. Otherwise it has failed; restore the previous
1090          capture values before returning NOMATCH. */
1091    
1092          for (;;)
1093            {
1094            md->offset_vector[md->offset_end - number] =
1095              (int)(eptr - md->start_subject);
1096            if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1097            RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1098              eptrb, RM63);
1099            if (rrc == MATCH_KETRPOS)
1100              {
1101              offset_top = md->end_offset_top;
1102              eptr = md->end_match_ptr;
1103              ecode = md->start_code + code_offset;
1104              save_capture_last = md->capture_last;
1105              matched_once = TRUE;
1106              continue;
1107              }
1108    
1109            /* See comment in the code for capturing groups above about handling
1110            THEN. */
1111    
1112            if (rrc == MATCH_THEN)
1113              {
1114              next = ecode + GET(ecode,1);
1115              if (md->start_match_ptr < next &&
1116                  (*ecode == OP_ALT || *next == OP_ALT))
1117                rrc = MATCH_NOMATCH;
1118              }
1119    
1120            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1121            md->capture_last = save_capture_last;
1122            ecode += GET(ecode, 1);
1123            if (*ecode != OP_ALT) break;
1124          }          }
1125    
1126        /* For non-final alternatives, continue the loop for a NOMATCH result;        if (!matched_once)
1127        otherwise return. */          {
1128            md->offset_vector[offset] = save_offset1;
1129            md->offset_vector[offset+1] = save_offset2;
1130            md->offset_vector[md->offset_end - number] = save_offset3;
1131            }
1132    
1133        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,        if (md->mark == NULL) md->mark = markptr;
1134          eptrb, flags, RM2);        if (allow_zero || matched_once)
1135        if (rrc != MATCH_NOMATCH &&          {
1136            (rrc != MATCH_THEN || md->start_match_ptr != ecode))          ecode += 1 + LINK_SIZE;
1137          RRETURN(rrc);          break;
1138            }
1139    
1140          RRETURN(MATCH_NOMATCH);
1141          }
1142    
1143        /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
1144        as a non-capturing bracket. */
1145    
1146        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1147        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1148    
1149        DPRINTF(("insufficient capture room: treat as non-capturing\n"));
1150    
1151        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1152        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1153    
1154        /* Non-capturing possessive bracket with unlimited repeat. We come here
1155        from BRAZERO with allow_zero = TRUE. The code is similar to the above,
1156        without the capturing complication. It is written out separately for speed
1157        and cleanliness. */
1158    
1159        case OP_BRAPOS:
1160        case OP_SBRAPOS:
1161        allow_zero = FALSE;
1162    
1163        POSSESSIVE_NON_CAPTURE:
1164        matched_once = FALSE;
1165        code_offset = ecode - md->start_code;
1166    
1167        for (;;)
1168          {
1169          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1170          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1171            eptrb, RM48);
1172          if (rrc == MATCH_KETRPOS)
1173            {
1174            offset_top = md->end_offset_top;
1175            eptr = md->end_match_ptr;
1176            ecode = md->start_code + code_offset;
1177            matched_once = TRUE;
1178            continue;
1179            }
1180    
1181          /* See comment in the code for capturing groups above about handling
1182          THEN. */
1183    
1184          if (rrc == MATCH_THEN)
1185            {
1186            next = ecode + GET(ecode,1);
1187            if (md->start_match_ptr < next &&
1188                (*ecode == OP_ALT || *next == OP_ALT))
1189              rrc = MATCH_NOMATCH;
1190            }
1191    
1192          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1193        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1194          if (*ecode != OP_ALT) break;
1195        }        }
1196    
1197        if (matched_once || allow_zero)
1198          {
1199          ecode += 1 + LINK_SIZE;
1200          break;
1201          }
1202        RRETURN(MATCH_NOMATCH);
1203    
1204      /* Control never reaches here. */      /* Control never reaches here. */
1205    
1206      /* Conditional group: compilation checked that there are no more than      /* Conditional group: compilation checked that there are no more than
1207      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
1208      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
1209      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. */  
1210    
1211      case OP_COND:      case OP_COND:
1212      case OP_SCOND:      case OP_SCOND:
1213      codelink= GET(ecode, 1);      codelink = GET(ecode, 1);
1214    
1215      /* 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
1216      inserted between OP_COND and an assertion condition. */      inserted between OP_COND and an assertion condition. */
# Line 928  for (;;) Line 1220  for (;;)
1220        if (pcre_callout != NULL)        if (pcre_callout != NULL)
1221          {          {
1222          pcre_callout_block cb;          pcre_callout_block cb;
1223          cb.version          = 1;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1224          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1225          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1226          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
# Line 940  for (;;) Line 1232  for (;;)
1232          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1233          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last;
1234          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1235            cb.mark             = (unsigned char *)markptr;
1236          if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);          if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);
1237          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1238          }          }
1239        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1240        }        }
1241    
1242      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 960  for (;;) Line 1253  for (;;)
1253        else        else
1254          {          {
1255          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1256          condition =  (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1257    
1258          /* If the test is for recursion into a specific subpattern, and it is          /* If the test is for recursion into a specific subpattern, and it is
1259          false, but the test was set up by name, scan the table to see if the          false, but the test was set up by name, scan the table to see if the
1260          name refers to any other numbers, and test them. The condition is true          name refers to any other numbers, and test them. The condition is true
1261          if any one is set. */          if any one is set. */
1262    
1263          if (!condition && condcode == OP_NRREF && recno != RREF_ANY)          if (!condition && condcode == OP_NRREF)
1264            {            {
1265            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1266            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1267              {              {
1268              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 982  for (;;) Line 1275  for (;;)
1275    
1276            if (i < md->name_count)            if (i < md->name_count)
1277              {              {
1278              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1279              while (slotB > md->name_table)              while (slotB > md->name_table)
1280                {                {
1281                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1282                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1283                  {                  {
1284                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1285                  if (condition) break;                  if (condition) break;
# Line 1002  for (;;) Line 1295  for (;;)
1295                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1296                  {                  {
1297                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1298                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1299                    {                    {
1300                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1301                    if (condition) break;                    if (condition) break;
# Line 1015  for (;;) Line 1308  for (;;)
1308    
1309          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1310    
1311          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1312          }          }
1313        }        }
1314    
# Line 1032  for (;;) Line 1325  for (;;)
1325        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1326          {          {
1327          int refno = offset >> 1;          int refno = offset >> 1;
1328          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1329    
1330          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1331            {            {
# Line 1046  for (;;) Line 1339  for (;;)
1339    
1340          if (i < md->name_count)          if (i < md->name_count)
1341            {            {
1342            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1343            while (slotB > md->name_table)            while (slotB > md->name_table)
1344              {              {
1345              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1346              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1347                {                {
1348                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1349                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1068  for (;;) Line 1361  for (;;)
1361              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1362                {                {
1363                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1364                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1365                  {                  {
1366                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1367                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1083  for (;;) Line 1376  for (;;)
1376    
1377        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1378    
1379        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1380        }        }
1381    
1382      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1093  for (;;) Line 1386  for (;;)
1386        }        }
1387    
1388      /* The condition is an assertion. Call match() to evaluate it - setting      /* The condition is an assertion. Call match() to evaluate it - setting
1389      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
1390      assertion. */      an assertion. */
1391    
1392      else      else
1393        {        {
1394        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        md->match_function_type = MATCH_CONDASSERT;
1395            match_condassert, RM3);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);
1396        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
1397          {          {
1398            if (md->end_offset_top > offset_top)
1399              offset_top = md->end_offset_top;  /* Captures may have happened */
1400          condition = TRUE;          condition = TRUE;
1401          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
1402          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
1403          }          }
1404        else if (rrc != MATCH_NOMATCH &&  
1405                (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an
1406          assertion; it is therefore treated as NOMATCH. */
1407    
1408          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1409          {          {
1410          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
1411          }          }
# Line 1118  for (;;) Line 1416  for (;;)
1416          }          }
1417        }        }
1418    
1419      /* We are now at the branch that is to be obeyed. As there is only one,      /* We are now at the branch that is to be obeyed. As there is only one, can
1420      we can use tail recursion to avoid using another stack frame, except when      use tail recursion to avoid using another stack frame, except when there is
1421      match_cbegroup is required for an unlimited repeat of a possibly empty      unlimited repeat of a possibly empty group. In the latter case, a recursive
1422      group. If the second alternative doesn't exist, we can just plough on. */      call to match() is always required, unless the second alternative doesn't
1423        exist, in which case we can just plough on. Note that, for compatibility
1424        with Perl, the | in a conditional group is NOT treated as creating two
1425        alternatives. If a THEN is encountered in the branch, it propagates out to
1426        the enclosing alternative (unless nested in a deeper set of alternatives,
1427        of course). */
1428    
1429      if (condition || *ecode == OP_ALT)      if (condition || *ecode == OP_ALT)
1430        {        {
1431        ecode += 1 + LINK_SIZE;        if (op != OP_SCOND)
       if (op == OP_SCOND)        /* Possibly empty group */  
         {  
         RMATCH(eptr, ecode, offset_top, md, ims, eptrb, match_cbegroup, RM49);  
         RRETURN(rrc);  
         }  
       else                       /* Group must match something */  
1432          {          {
1433          flags = 0;          ecode += 1 + LINK_SIZE;
1434          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1435          }          }
1436    
1437          md->match_function_type = MATCH_CBEGROUP;
1438          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);
1439          RRETURN(rrc);
1440        }        }
1441      else                         /* Condition false & no alternative */  
1442         /* Condition false & no alternative; continue after the group. */
1443    
1444        else
1445        {        {
1446        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1447        }        }
# Line 1164  for (;;) Line 1468  for (;;)
1468        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1469        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1470        }        }
1471      ecode += 3;      ecode += 1 + IMM2_SIZE;
1472      break;      break;
1473    
1474    
1475      /* End of the pattern, either real or forced. If we are in a top-level      /* End of the pattern, either real or forced. */
     recursion, we should restore the offsets appropriately and continue from  
     after the call. */  
1476    
     case OP_ACCEPT:  
1477      case OP_END:      case OP_END:
1478      if (md->recursive != NULL && md->recursive->group_num == 0)      case OP_ACCEPT:
1479        {      case OP_ASSERT_ACCEPT:
       recursion_info *rec = md->recursive;  
       DPRINTF(("End of pattern in a (?0) recursion\n"));  
       md->recursive = rec->prevrec;  
       memmove(md->offset_vector, rec->offset_save,  
         rec->saved_max * sizeof(int));  
       offset_top = rec->save_offset_top;  
       ims = original_ims;  
       ecode = rec->after_call;  
       break;  
       }  
1480    
1481      /* Otherwise, if we have matched an empty string, fail if PCRE_NOTEMPTY is      /* If we have matched an empty string, fail if not in an assertion and not
1482      set, or if PCRE_NOTEMPTY_ATSTART is set and we have matched at the start of      in a recursion if either PCRE_NOTEMPTY is set, or if PCRE_NOTEMPTY_ATSTART
1483      the subject. In both cases, backtracking will then try other alternatives,      is set and we have matched at the start of the subject. In both cases,
1484      if any. */      backtracking will then try other alternatives, if any. */
1485    
1486      if (eptr == mstart &&      if (eptr == mstart && op != OP_ASSERT_ACCEPT &&
1487          (md->notempty ||           md->recursive == NULL &&
1488            (md->notempty_atstart &&           (md->notempty ||
1489              mstart == md->start_subject + md->start_offset)))             (md->notempty_atstart &&
1490                 mstart == md->start_subject + md->start_offset)))
1491        MRRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
1492    
1493      /* Otherwise, we have a match. */      /* Otherwise, we have a match. */
# Line 1210  for (;;) Line 1502  for (;;)
1502      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;
1503      MRRETURN(rrc);      MRRETURN(rrc);
1504    
     /* Change option settings */  
   
     case OP_OPT:  
     ims = ecode[1];  
     ecode += 2;  
     DPRINTF(("ims set to %02lx\n", ims));  
     break;  
   
1505      /* Assertion brackets. Check the alternative branches in turn - the      /* Assertion brackets. Check the alternative branches in turn - the
1506      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,
1507      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
1508      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
1509      this level is identical to the lookahead case. */      this level is identical to the lookahead case. When the assertion is part
1510        of a condition, we want to return immediately afterwards. The caller of
1511        this incarnation of the match() function will have set MATCH_CONDASSERT in
1512        md->match_function type, and one of these opcodes will be the first opcode
1513        that is processed. We use a local variable that is preserved over calls to
1514        match() to remember this case. */
1515    
1516      case OP_ASSERT:      case OP_ASSERT:
1517      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1518        if (md->match_function_type == MATCH_CONDASSERT)
1519          {
1520          condassert = TRUE;
1521          md->match_function_type = 0;
1522          }
1523        else condassert = FALSE;
1524    
1525      do      do
1526        {        {
1527        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
         RM4);  
1528        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1529          {          {
1530          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1531            markptr = md->mark;
1532          break;          break;
1533          }          }
1534        if (rrc != MATCH_NOMATCH &&  
1535            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE does not allow THEN to escape beyond an assertion; it is treated
1536          RRETURN(rrc);        as NOMATCH. */
1537    
1538          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1539        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1540        }        }
1541      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1542    
1543      if (*ecode == OP_KET) MRRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) MRRETURN(MATCH_NOMATCH);
1544    
1545      /* If checking an assertion for a condition, return MATCH_MATCH. */      /* If checking an assertion for a condition, return MATCH_MATCH. */
1546    
1547      if ((flags & match_condassert) != 0) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);
1548    
1549      /* Continue from after the assertion, updating the offsets high water      /* Continue from after the assertion, updating the offsets high water
1550      mark, since extracts may have been taken during the assertion. */      mark, since extracts may have been taken during the assertion. */
# Line 1261  for (;;) Line 1560  for (;;)
1560    
1561      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1562      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1563        if (md->match_function_type == MATCH_CONDASSERT)
1564          {
1565          condassert = TRUE;
1566          md->match_function_type = 0;
1567          }
1568        else condassert = FALSE;
1569    
1570      do      do
1571        {        {
1572        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
         RM5);  
1573        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) MRRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) MRRETURN(MATCH_NOMATCH);
1574        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1575          {          {
1576          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1577          break;          break;
1578          }          }
1579        if (rrc != MATCH_NOMATCH &&  
1580            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE does not allow THEN to escape beyond an assertion; it is treated
1581          RRETURN(rrc);        as NOMATCH. */
1582    
1583          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1584        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1585        }        }
1586      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1587    
1588      if ((flags & match_condassert) != 0) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1589    
1590      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1591      continue;      continue;
# Line 1290  for (;;) Line 1597  for (;;)
1597    
1598      case OP_REVERSE:      case OP_REVERSE:
1599  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
1600      if (utf8)      if (utf)
1601        {        {
1602        i = GET(ecode, 1);        i = GET(ecode, 1);
1603        while (i-- > 0)        while (i-- > 0)
# Line 1324  for (;;) Line 1631  for (;;)
1631      if (pcre_callout != NULL)      if (pcre_callout != NULL)
1632        {        {
1633        pcre_callout_block cb;        pcre_callout_block cb;
1634        cb.version          = 1;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1635        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1636        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1637        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
# Line 1336  for (;;) Line 1643  for (;;)
1643        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1644        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last;
1645        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1646          cb.mark             = (unsigned char *)markptr;
1647        if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);        if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);
1648        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1649        }        }
# Line 1346  for (;;) Line 1654  for (;;)
1654      offset data is the offset to the starting bracket from the start of the      offset data is the offset to the starting bracket from the start of the
1655      whole pattern. (This is so that it works from duplicated subpatterns.)      whole pattern. (This is so that it works from duplicated subpatterns.)
1656    
1657      If there are any capturing brackets started but not finished, we have to      The state of the capturing groups is preserved over recursion, and
1658      save their starting points and reinstate them after the recursion. However,      re-instated afterwards. We don't know how many are started and not yet
1659      we don't know how many such there are (offset_top records the completed      finished (offset_top records the completed total) so we just have to save
1660      total) so we just have to save all the potential data. There may be up to      all the potential data. There may be up to 65535 such values, which is too
1661      65535 such values, which is too large to put on the stack, but using malloc      large to put on the stack, but using malloc for small numbers seems
1662      for small numbers seems expensive. As a compromise, the stack is used when      expensive. As a compromise, the stack is used when there are no more than
1663      there are no more than REC_STACK_SAVE_MAX values to store; otherwise malloc      REC_STACK_SAVE_MAX values to store; otherwise malloc is used.
     is used. A problem is what to do if the malloc fails ... there is no way of  
     returning to the top level with an error. Save the top REC_STACK_SAVE_MAX  
     values on the stack, and accept that the rest may be wrong.  
1664    
1665      There are also other values that have to be saved. We use a chained      There are also other values that have to be saved. We use a chained
1666      sequence of blocks that actually live on the stack. Thanks to Robin Houston      sequence of blocks that actually live on the stack. Thanks to Robin Houston
1667      for the original version of this logic. */      for the original version of this logic. It has, however, been hacked around
1668        a lot, so he is not to blame for the current way it works. */
1669    
1670      case OP_RECURSE:      case OP_RECURSE:
1671        {        {
1672          recursion_info *ri;
1673          int recno;
1674    
1675        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1676        new_recursive.group_num = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
1677          GET2(callpat, 1 + LINK_SIZE);          GET2(callpat, 1 + LINK_SIZE);
1678    
1679          /* Check for repeating a recursion without advancing the subject pointer.
1680          This should catch convoluted mutual recursions. (Some simple cases are
1681          caught at compile time.) */
1682    
1683          for (ri = md->recursive; ri != NULL; ri = ri->prevrec)
1684            if (recno == ri->group_num && eptr == ri->subject_position)
1685              RRETURN(PCRE_ERROR_RECURSELOOP);
1686    
1687        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1688    
1689          new_recursive.group_num = recno;
1690          new_recursive.subject_position = eptr;
1691        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1692        md->recursive = &new_recursive;        md->recursive = &new_recursive;
1693    
1694        /* Find where to continue from afterwards */        /* Where to continue from afterwards */
1695    
1696        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
       new_recursive.after_call = ecode;  
1697    
1698        /* Now save the offset data. */        /* Now save the offset data */
1699    
1700        new_recursive.saved_max = md->offset_end;        new_recursive.saved_max = md->offset_end;
1701        if (new_recursive.saved_max <= REC_STACK_SAVE_MAX)        if (new_recursive.saved_max <= REC_STACK_SAVE_MAX)
# Line 1388  for (;;) Line 1706  for (;;)
1706            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));
1707          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1708          }          }
   
1709        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1710              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
       new_recursive.save_offset_top = offset_top;  
1711    
1712        /* OK, now we can do the recursion. For each top-level alternative we        /* OK, now we can do the recursion. After processing each alternative,
1713        restore the offset and recursion data. */        restore the offset data. If there were nested recursions, md->recursive
1714          might be changed, so reset it before looping. */
1715    
1716        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1717        flags = (*callpat >= OP_SBRA)? match_cbegroup : 0;        cbegroup = (*callpat >= OP_SBRA);
1718        do        do
1719          {          {
1720          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1721            md, ims, eptrb, flags, RM6);          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1722              md, eptrb, RM6);
1723            memcpy(md->offset_vector, new_recursive.offset_save,
1724                new_recursive.saved_max * sizeof(int));
1725            md->recursive = new_recursive.prevrec;
1726          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1727            {            {
1728            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
           md->recursive = new_recursive.prevrec;  
1729            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1730              (pcre_free)(new_recursive.offset_save);              (pcre_free)(new_recursive.offset_save);
1731            MRRETURN(MATCH_MATCH);  
1732              /* Set where we got to in the subject, and reset the start in case
1733              it was changed by \K. This *is* propagated back out of a recursion,
1734              for Perl compatibility. */
1735    
1736              eptr = md->end_match_ptr;
1737              mstart = md->start_match_ptr;
1738              goto RECURSION_MATCHED;        /* Exit loop; end processing */
1739            }            }
1740          else if (rrc != MATCH_NOMATCH &&  
1741                  (rrc != MATCH_THEN || md->start_match_ptr != ecode))          /* PCRE does not allow THEN to escape beyond a recursion; it is treated
1742            as NOMATCH. */
1743    
1744            else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1745            {            {
1746            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1747            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
# Line 1420  for (;;) Line 1750  for (;;)
1750            }            }
1751    
1752          md->recursive = &new_recursive;          md->recursive = &new_recursive;
         memcpy(md->offset_vector, new_recursive.offset_save,  
             new_recursive.saved_max * sizeof(int));  
1753          callpat += GET(callpat, 1);          callpat += GET(callpat, 1);
1754          }          }
1755        while (*callpat == OP_ALT);        while (*callpat == OP_ALT);
# Line 1432  for (;;) Line 1760  for (;;)
1760          (pcre_free)(new_recursive.offset_save);          (pcre_free)(new_recursive.offset_save);
1761        MRRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
1762        }        }
     /* Control never reaches here */  
   
     /* "Once" brackets are like assertion brackets except that after a match,  
     the point in the subject string is not moved back. Thus there can never be  
     a move back into the brackets. Friedl calls these "atomic" subpatterns.  
     Check the alternative branches in turn - the matching won't pass the KET  
     for this kind of subpattern. If any one branch matches, we carry on as at  
     the end of a normal bracket, leaving the subject pointer, but resetting  
     the start-of-match value in case it was changed by \K. */  
   
     case OP_ONCE:  
     prev = ecode;  
     saved_eptr = eptr;  
   
     do  
       {  
       RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM7);  
       if (rrc == MATCH_MATCH)  /* Note: _not_ MATCH_ACCEPT */  
         {  
         mstart = md->start_match_ptr;  
         break;  
         }  
       if (rrc != MATCH_NOMATCH &&  
           (rrc != MATCH_THEN || md->start_match_ptr != ecode))  
         RRETURN(rrc);  
       ecode += GET(ecode,1);  
       }  
     while (*ecode == OP_ALT);  
   
     /* If hit the end of the group (which could be repeated), fail */  
   
     if (*ecode != OP_ONCE && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);  
   
     /* Continue as from after the assertion, updating the offsets high water  
     mark, since extracts may have been taken. */  
   
     do ecode += GET(ecode, 1); while (*ecode == OP_ALT);  
   
     offset_top = md->end_offset_top;  
     eptr = md->end_match_ptr;  
   
     /* For a non-repeating ket, just continue at this level. This also  
     happens for a repeating ket if no characters were matched in the group.  
     This is the forcible breaking of infinite loops as implemented in Perl  
     5.005. If there is an options reset, it will get obeyed in the normal  
     course of events. */  
   
     if (*ecode == OP_KET || eptr == saved_eptr)  
       {  
       ecode += 1+LINK_SIZE;  
       break;  
       }  
   
     /* The repeating kets try the rest of the pattern or restart from the  
     preceding bracket, in the appropriate order. The second "call" of match()  
     uses tail recursion, to avoid using another stack frame. We need to reset  
     any options that changed within the bracket before re-running it, so  
     check the next opcode. */  
   
     if (ecode[1+LINK_SIZE] == OP_OPT)  
       {  
       ims = (ims & ~PCRE_IMS) | ecode[4];  
       DPRINTF(("ims set to %02lx at group repeat\n", ims));  
       }  
1763    
1764      if (*ecode == OP_KETRMIN)      RECURSION_MATCHED:
1765        {      break;
       RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM8);  
       if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
       ecode = prev;  
       flags = 0;  
       goto TAIL_RECURSE;  
       }  
     else  /* OP_KETRMAX */  
       {  
       RMATCH(eptr, prev, offset_top, md, ims, eptrb, match_cbegroup, RM9);  
       if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
       ecode += 1 + LINK_SIZE;  
       flags = 0;  
       goto TAIL_RECURSE;  
       }  
     /* Control never gets here */  
1766    
1767      /* An alternation is the end of a branch; scan along to find the end of the      /* An alternation is the end of a branch; scan along to find the end of the
1768      bracketed group and go to there. */      bracketed group and go to there. */
# Line 1529  for (;;) Line 1778  for (;;)
1778      optional ones preceded by BRAZERO or BRAMINZERO. */      optional ones preceded by BRAZERO or BRAMINZERO. */
1779    
1780      case OP_BRAZERO:      case OP_BRAZERO:
1781        {      next = ecode + 1;
1782        next = ecode+1;      RMATCH(eptr, next, offset_top, md, eptrb, RM10);
1783        RMATCH(eptr, next, offset_top, md, ims, eptrb, 0, RM10);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1784        if (rrc != MATCH_NOMATCH) RRETURN(rrc);      do next += GET(next, 1); while (*next == OP_ALT);
1785        do next += GET(next,1); while (*next == OP_ALT);      ecode = next + 1 + LINK_SIZE;
       ecode = next + 1 + LINK_SIZE;  
       }  
1786      break;      break;
1787    
1788      case OP_BRAMINZERO:      case OP_BRAMINZERO:
1789        {      next = ecode + 1;
1790        next = ecode+1;      do next += GET(next, 1); while (*next == OP_ALT);
1791        do next += GET(next, 1); while (*next == OP_ALT);      RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, eptrb, RM11);
1792        RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0, RM11);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1793        if (rrc != MATCH_NOMATCH) RRETURN(rrc);      ecode++;
       ecode++;  
       }  
1794      break;      break;
1795    
1796      case OP_SKIPZERO:      case OP_SKIPZERO:
1797        {      next = ecode+1;
1798        next = ecode+1;      do next += GET(next,1); while (*next == OP_ALT);
1799        do next += GET(next,1); while (*next == OP_ALT);      ecode = next + 1 + LINK_SIZE;
       ecode = next + 1 + LINK_SIZE;  
       }  
1800      break;      break;
1801    
1802        /* BRAPOSZERO occurs before a possessive bracket group. Don't do anything
1803        here; just jump to the group, with allow_zero set TRUE. */
1804    
1805        case OP_BRAPOSZERO:
1806        op = *(++ecode);
1807        allow_zero = TRUE;
1808        if (op == OP_CBRAPOS || op == OP_SCBRAPOS) goto POSSESSIVE_CAPTURE;
1809          goto POSSESSIVE_NON_CAPTURE;
1810    
1811      /* End of a group, repeated or non-repeating. */      /* End of a group, repeated or non-repeating. */
1812    
1813      case OP_KET:      case OP_KET:
1814      case OP_KETRMIN:      case OP_KETRMIN:
1815      case OP_KETRMAX:      case OP_KETRMAX:
1816        case OP_KETRPOS:
1817      prev = ecode - GET(ecode, 1);      prev = ecode - GET(ecode, 1);
1818    
1819      /* 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
1820      infinite repeats of empty string matches, retrieve the subject start from      infinite repeats of empty string matches, retrieve the subject start from
1821      the chain. Otherwise, set it NULL. */      the chain. Otherwise, set it NULL. */
1822    
1823      if (*prev >= OP_SBRA)      if (*prev >= OP_SBRA || *prev == OP_ONCE)
1824        {        {
1825        saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */        saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */
1826        eptrb = eptrb->epb_prev;              /* Backup to previous group */        eptrb = eptrb->epb_prev;              /* Backup to previous group */
1827        }        }
1828      else saved_eptr = NULL;      else saved_eptr = NULL;
1829    
1830      /* If we are at the end of an assertion group or an atomic group, stop      /* If we are at the end of an assertion group or a non-capturing atomic
1831      matching and return MATCH_MATCH, but record the current high water mark for      group, stop matching and return MATCH_MATCH, but record the current high
1832      use by positive assertions. We also need to record the match start in case      water mark for use by positive assertions. We also need to record the match
1833      it was changed by \K. */      start in case it was changed by \K. */
1834    
1835      if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||      if ((*prev >= OP_ASSERT && *prev <= OP_ASSERTBACK_NOT) ||
1836          *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||           *prev == OP_ONCE_NC)
         *prev == OP_ONCE)  
1837        {        {
1838        md->end_match_ptr = eptr;      /* For ONCE */        md->end_match_ptr = eptr;      /* For ONCE_NC */
1839        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
1840        md->start_match_ptr = mstart;        md->start_match_ptr = mstart;
1841        MRRETURN(MATCH_MATCH);        MRRETURN(MATCH_MATCH);         /* Sets md->mark */
1842        }        }
1843    
1844      /* For capturing groups we have to check the group number back at the start      /* For capturing groups we have to check the group number back at the start
1845      and if necessary complete handling an extraction by setting the offsets and      and if necessary complete handling an extraction by setting the offsets and
1846      bumping the high water mark. Note that whole-pattern recursion is coded as      bumping the high water mark. Whole-pattern recursion is coded as a recurse
1847      a recurse into group 0, so it won't be picked up here. Instead, we catch it      into group 0, so it won't be picked up here. Instead, we catch it when the
1848      when the OP_END is reached. Other recursion is handled here. */      OP_END is reached. Other recursion is handled here. We just have to record
1849        the current subject position and start match pointer and give a MATCH
1850        return. */
1851    
1852      if (*prev == OP_CBRA || *prev == OP_SCBRA)      if (*prev == OP_CBRA || *prev == OP_SCBRA ||
1853            *prev == OP_CBRAPOS || *prev == OP_SCBRAPOS)
1854        {        {
1855        number = GET2(prev, 1+LINK_SIZE);        number = GET2(prev, 1+LINK_SIZE);
1856        offset = number << 1;        offset = number << 1;
# Line 1605  for (;;) Line 1860  for (;;)
1860        printf("\n");        printf("\n");
1861  #endif  #endif
1862    
1863          /* Handle a recursively called group. */
1864    
1865          if (md->recursive != NULL && md->recursive->group_num == number)
1866            {
1867            md->end_match_ptr = eptr;
1868            md->start_match_ptr = mstart;
1869            RRETURN(MATCH_MATCH);
1870            }
1871    
1872          /* Deal with capturing */
1873    
1874        md->capture_last = number;        md->capture_last = number;
1875        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->offset_overflow = TRUE; else
1876          {          {
1877            /* If offset is greater than offset_top, it means that we are
1878            "skipping" a capturing group, and that group's offsets must be marked
1879            unset. In earlier versions of PCRE, all the offsets were unset at the
1880            start of matching, but this doesn't work because atomic groups and
1881            assertions can cause a value to be set that should later be unset.
1882            Example: matching /(?>(a))b|(a)c/ against "ac". This sets group 1 as
1883            part of the atomic group, but this is not on the final matching path,
1884            so must be unset when 2 is set. (If there is no group 2, there is no
1885            problem, because offset_top will then be 2, indicating no capture.) */
1886    
1887            if (offset > offset_top)
1888              {
1889              register int *iptr = md->offset_vector + offset_top;
1890              register int *iend = md->offset_vector + offset;
1891              while (iptr < iend) *iptr++ = -1;
1892              }
1893    
1894            /* Now make the extraction */
1895    
1896          md->offset_vector[offset] =          md->offset_vector[offset] =
1897            md->offset_vector[md->offset_end - number];            md->offset_vector[md->offset_end - number];
1898          md->offset_vector[offset+1] = (int)(eptr - md->start_subject);          md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1899          if (offset_top <= offset) offset_top = offset + 2;          if (offset_top <= offset) offset_top = offset + 2;
1900          }          }
1901          }
1902    
1903        /* Handle a recursively called group. Restore the offsets      /* For an ordinary non-repeating ket, just continue at this level. This
1904        appropriately and continue from after the call. */      also happens for a repeating ket if no characters were matched in the
1905        group. This is the forcible breaking of infinite loops as implemented in
1906        Perl 5.005. For a non-repeating atomic group that includes captures,
1907        establish a backup point by processing the rest of the pattern at a lower
1908        level. If this results in a NOMATCH return, pass MATCH_ONCE back to the
1909        original OP_ONCE level, thereby bypassing intermediate backup points, but
1910        resetting any captures that happened along the way. */
1911    
1912        if (md->recursive != NULL && md->recursive->group_num == number)      if (*ecode == OP_KET || eptr == saved_eptr)
1913          {
1914          if (*prev == OP_ONCE)
1915          {          {
1916          recursion_info *rec = md->recursive;          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM12);
1917          DPRINTF(("Recursion (%d) succeeded - continuing\n", number));          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1918          md->recursive = rec->prevrec;          md->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */
1919          memcpy(md->offset_vector, rec->offset_save,          RRETURN(MATCH_ONCE);
           rec->saved_max * sizeof(int));  
         offset_top = rec->save_offset_top;  
         ecode = rec->after_call;  
         ims = original_ims;  
         break;  
1920          }          }
1921          ecode += 1 + LINK_SIZE;    /* Carry on at this level */
1922          break;
1923        }        }
1924    
1925      /* For both capturing and non-capturing groups, reset the value of the ims      /* OP_KETRPOS is a possessive repeating ket. Remember the current position,
1926      flags, in case they got changed during the group. */      and return the MATCH_KETRPOS. This makes it possible to do the repeats one
1927        at a time from the outer level, thus saving stack. */
1928    
1929      ims = original_ims;      if (*ecode == OP_KETRPOS)
     DPRINTF(("ims reset to %02lx\n", ims));  
   
     /* For a non-repeating ket, just continue at this level. This also  
     happens for a repeating ket if no characters were matched in the group.  
     This is the forcible breaking of infinite loops as implemented in Perl  
     5.005. If there is an options reset, it will get obeyed in the normal  
     course of events. */  
   
     if (*ecode == OP_KET || eptr == saved_eptr)  
1930        {        {
1931        ecode += 1 + LINK_SIZE;        md->end_match_ptr = eptr;
1932        break;        md->end_offset_top = offset_top;
1933          RRETURN(MATCH_KETRPOS);
1934        }        }
1935    
1936      /* The repeating kets try the rest of the pattern or restart from the      /* The normal repeating kets try the rest of the pattern or restart from
1937      preceding bracket, in the appropriate order. In the second case, we can use      the preceding bracket, in the appropriate order. In the second case, we can
1938      tail recursion to avoid using another stack frame, unless we have an      use tail recursion to avoid using another stack frame, unless we have an
1939      unlimited repeat of a group that can match an empty string. */      an atomic group or an unlimited repeat of a group that can match an empty
1940        string. */
     flags = (*prev >= OP_SBRA)? match_cbegroup : 0;  
1941    
1942      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1943        {        {
1944        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM12);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM7);
1945        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1946        if (flags != 0)    /* Could match an empty string */        if (*prev == OP_ONCE)
1947          {          {
1948          RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM8);
1949            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1950            md->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */
1951            RRETURN(MATCH_ONCE);
1952            }
1953          if (*prev >= OP_SBRA)    /* Could match an empty string */
1954            {
1955            md->match_function_type = MATCH_CBEGROUP;
1956            RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
1957          RRETURN(rrc);          RRETURN(rrc);
1958          }          }
1959        ecode = prev;        ecode = prev;
# Line 1670  for (;;) Line 1961  for (;;)
1961        }        }
1962      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
1963        {        {
1964        RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13);        if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1965          RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
1966          if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
1967        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1968          if (*prev == OP_ONCE)
1969            {
1970            RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM9);
1971            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1972            md->once_target = prev;
1973            RRETURN(MATCH_ONCE);
1974            }
1975        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
       flags = 0;  
1976        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1977        }        }
1978      /* Control never gets here */      /* Control never gets here */
1979    
1980      /* Start of subject unless notbol, or after internal newline if multiline */      /* Not multiline mode: start of subject assertion, unless notbol. */
1981    
1982      case OP_CIRC:      case OP_CIRC:
1983      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);
     if ((ims & PCRE_MULTILINE) != 0)  
       {  
       if (eptr != md->start_subject &&  
           (eptr == md->end_subject || !WAS_NEWLINE(eptr)))  
         MRRETURN(MATCH_NOMATCH);  
       ecode++;  
       break;  
       }  
     /* ... else fall through */  
1984    
1985      /* Start of subject assertion */      /* Start of subject assertion */
1986    
# Line 1699  for (;;) Line 1989  for (;;)
1989      ecode++;      ecode++;
1990      break;      break;
1991    
1992        /* Multiline mode: start of subject unless notbol, or after any newline. */
1993    
1994        case OP_CIRCM:
1995        if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);
1996        if (eptr != md->start_subject &&
1997            (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
1998          MRRETURN(MATCH_NOMATCH);
1999        ecode++;
2000        break;
2001    
2002      /* Start of match assertion */      /* Start of match assertion */
2003    
2004      case OP_SOM:      case OP_SOM:
# Line 1713  for (;;) Line 2013  for (;;)
2013      ecode++;      ecode++;
2014      break;      break;
2015    
2016      /* Assert before internal newline if multiline, or before a terminating      /* Multiline mode: assert before any newline, or before end of subject
2017      newline unless endonly is set, else end of subject unless noteol is set. */      unless noteol is set. */
2018    
2019      case OP_DOLL:      case OP_DOLLM:
2020      if ((ims & PCRE_MULTILINE) != 0)      if (eptr < md->end_subject)
2021        {        { if (!IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH); }
2022        if (eptr < md->end_subject)      else
         { if (!IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH); }  
       else  
         {  
         if (md->noteol) MRRETURN(MATCH_NOMATCH);  
         SCHECK_PARTIAL();  
         }  
       ecode++;  
       break;  
       }  
     else  /* Not multiline */  
2023        {        {
2024        if (md->noteol) MRRETURN(MATCH_NOMATCH);        if (md->noteol) MRRETURN(MATCH_NOMATCH);
2025        if (!md->endonly) goto ASSERT_NL_OR_EOS;        SCHECK_PARTIAL();
2026        }        }
2027        ecode++;
2028        break;
2029    
2030        /* Not multiline mode: assert before a terminating newline or before end of
2031        subject unless noteol is set. */
2032    
2033        case OP_DOLL:
2034        if (md->noteol) MRRETURN(MATCH_NOMATCH);
2035        if (!md->endonly) goto ASSERT_NL_OR_EOS;
2036    
2037      /* ... else fall through for endonly */      /* ... else fall through for endonly */
2038    
# Line 1770  for (;;) Line 2069  for (;;)
2069        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2070        partial matching. */        partial matching. */
2071    
2072  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2073        if (utf8)        if (utf)
2074          {          {
2075          /* Get status of previous character */          /* Get status of previous character */
2076    
2077          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2078            {            {
2079            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2080            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2081            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2082            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2083  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 1884  for (;;) Line 2183  for (;;)
2183      /* Fall through */      /* Fall through */
2184    
2185      case OP_ALLANY:      case OP_ALLANY:
2186      if (eptr++ >= md->end_subject)      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2187        {        {                            /* not be updated before SCHECK_PARTIAL. */
2188        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2189        MRRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2190        }        }
2191      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;      eptr++;
2192    #ifdef SUPPORT_UTF
2193        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2194    #endif
2195      ecode++;      ecode++;
2196      break;      break;
2197    
# Line 1897  for (;;) Line 2199  for (;;)
2199      any byte, even newline, independent of the setting of PCRE_DOTALL. */      any byte, even newline, independent of the setting of PCRE_DOTALL. */
2200    
2201      case OP_ANYBYTE:      case OP_ANYBYTE:
2202      if (eptr++ >= md->end_subject)      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2203        {        {                            /* not be updated before SCHECK_PARTIAL. */
2204        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2205        MRRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2206        }        }
2207        eptr++;
2208      ecode++;      ecode++;
2209      break;      break;
2210    
# Line 2017  for (;;) Line 2320  for (;;)
2320      switch(c)      switch(c)
2321        {        {
2322        default: MRRETURN(MATCH_NOMATCH);        default: MRRETURN(MATCH_NOMATCH);
2323    
2324        case 0x000d:        case 0x000d:
2325        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
2326        break;        break;
# Line 2176  for (;;) Line 2480  for (;;)
2480          break;          break;
2481    
2482          case PT_GC:          case PT_GC:
2483          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2484            MRRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
2485          break;          break;
2486    
# Line 2193  for (;;) Line 2497  for (;;)
2497          /* These are specials */          /* These are specials */
2498    
2499          case PT_ALNUM:          case PT_ALNUM:
2500          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2501               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2502            MRRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
2503          break;          break;
2504    
2505          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2506          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2507               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2508                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2509            MRRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
2510          break;          break;
2511    
2512          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2513          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2514               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2515               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2516                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
# Line 2214  for (;;) Line 2518  for (;;)
2518          break;          break;
2519    
2520          case PT_WORD:          case PT_WORD:
2521          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2522               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2523               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2524            MRRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
2525          break;          break;
# Line 2240  for (;;) Line 2544  for (;;)
2544        MRRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2545        }        }
2546      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2547        if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH);
2548        while (eptr < md->end_subject)
2549        {        {
2550        int category = UCD_CATEGORY(c);        int len = 1;
2551        if (category == ucp_M) MRRETURN(MATCH_NOMATCH);        if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2552        while (eptr < md->end_subject)        if (UCD_CATEGORY(c) != ucp_M) break;
2553          {        eptr += len;
         int len = 1;  
         if (!utf8) c = *eptr; else  
           {  
           GETCHARLEN(c, eptr, len);  
           }  
         category = UCD_CATEGORY(c);  
         if (category != ucp_M) break;  
         eptr += len;  
         }  
2554        }        }
2555      ecode++;      ecode++;
2556      break;      break;
# Line 2269  for (;;) Line 2566  for (;;)
2566      loops). */      loops). */
2567    
2568      case OP_REF:      case OP_REF:
2569        case OP_REFI:
2570        caseless = op == OP_REFI;
2571      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2572      ecode += 3;      ecode += 1 + IMM2_SIZE;
2573    
2574      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2575    
# Line 2310  for (;;) Line 2609  for (;;)
2609        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2610        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2611        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2612        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2613        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2614        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2615        break;        break;
2616    
2617        default:               /* No repeat follows */        default:               /* No repeat follows */
2618        if ((length = match_ref(offset, eptr, length, md, ims)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2619          {          {
2620          CHECK_PARTIAL();          CHECK_PARTIAL();
2621          MRRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
# Line 2336  for (;;) Line 2635  for (;;)
2635    
2636      for (i = 1; i <= min; i++)      for (i = 1; i <= min; i++)
2637        {        {
2638        int slength;        int slength;
2639        if ((slength = match_ref(offset, eptr, length, md, ims)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2640          {          {
2641          CHECK_PARTIAL();          CHECK_PARTIAL();
2642          MRRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
# Line 2356  for (;;) Line 2655  for (;;)
2655        {        {
2656        for (fi = min;; fi++)        for (fi = min;; fi++)
2657          {          {
2658          int slength;          int slength;
2659          RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM14);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);
2660          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2661          if (fi >= max) MRRETURN(MATCH_NOMATCH);          if (fi >= max) MRRETURN(MATCH_NOMATCH);
2662          if ((slength = match_ref(offset, eptr, length, md, ims)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2663            {            {
2664            CHECK_PARTIAL();            CHECK_PARTIAL();
2665            MRRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
# Line 2377  for (;;) Line 2676  for (;;)
2676        pp = eptr;        pp = eptr;
2677        for (i = min; i < max; i++)        for (i = min; i < max; i++)
2678          {          {
2679          int slength;          int slength;
2680          if ((slength = match_ref(offset, eptr, length, md, ims)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2681            {            {
2682            CHECK_PARTIAL();            CHECK_PARTIAL();
2683            break;            break;
# Line 2387  for (;;) Line 2686  for (;;)
2686          }          }
2687        while (eptr >= pp)        while (eptr >= pp)
2688          {          {
2689          RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
2690          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2691          eptr -= length;          eptr -= length;
2692          }          }
# Line 2409  for (;;) Line 2708  for (;;)
2708      case OP_NCLASS:      case OP_NCLASS:
2709      case OP_CLASS:      case OP_CLASS:
2710        {        {
2711          /* The data variable is saved across frames, so the byte map needs to
2712          be stored there. */
2713    #define BYTE_MAP ((pcre_uint8 *)data)
2714        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2715        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2716    
2717        switch (*ecode)        switch (*ecode)
2718          {          {
# Line 2431  for (;;) Line 2733  for (;;)
2733          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2734          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2735          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2736          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2737          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2738          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2739          break;          break;
2740    
2741          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2443  for (;;) Line 2745  for (;;)
2745    
2746        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2747    
2748  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2749        /* UTF-8 mode */        if (utf)
       if (utf8)  
2750          {          {
2751          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2752            {            {
# Line 2460  for (;;) Line 2761  for (;;)
2761              if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);
2762              }              }
2763            else            else
2764              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
             }  
2765            }            }
2766          }          }
2767        else        else
2768  #endif  #endif
2769        /* Not UTF-8 mode */        /* Not UTF mode */
2770          {          {
2771          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2772            {            {
# Line 2477  for (;;) Line 2776  for (;;)
2776              MRRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
2777              }              }
2778            c = *eptr++;            c = *eptr++;
2779            if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2780              if (c > 255)
2781                {
2782                if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);
2783                }
2784              else
2785    #endif
2786                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);
2787            }            }
2788          }          }
2789    
# Line 2491  for (;;) Line 2797  for (;;)
2797    
2798        if (minimize)        if (minimize)
2799          {          {
2800  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2801          /* UTF-8 mode */          if (utf)
         if (utf8)  
2802            {            {
2803            for (fi = min;; fi++)            for (fi = min;; fi++)
2804              {              {
2805              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM16);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);
2806              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2807              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
2808              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 2511  for (;;) Line 2816  for (;;)
2816                if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);
2817                }                }
2818              else              else
2819                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
               }  
2820              }              }
2821            }            }
2822          else          else
2823  #endif  #endif
2824          /* Not UTF-8 mode */          /* Not UTF mode */
2825            {            {
2826            for (fi = min;; fi++)            for (fi = min;; fi++)
2827              {              {
2828              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM17);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);
2829              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2830              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
2831              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 2531  for (;;) Line 2834  for (;;)
2834                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
2835                }                }
2836              c = *eptr++;              c = *eptr++;
2837              if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2838                if (c > 255)
2839                  {
2840                  if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);
2841                  }
2842                else
2843    #endif
2844                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);
2845              }              }
2846            }            }
2847          /* Control never gets here */          /* Control never gets here */
# Line 2543  for (;;) Line 2853  for (;;)
2853          {          {
2854          pp = eptr;          pp = eptr;
2855    
2856  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2857          /* UTF-8 mode */          if (utf)
         if (utf8)  
2858            {            {
2859            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2860              {              {
# Line 2561  for (;;) Line 2870  for (;;)
2870                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
2871                }                }
2872              else              else
2873                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
2874              eptr += len;              eptr += len;
2875              }              }
2876            for (;;)            for (;;)
2877              {              {
2878              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM18);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM18);
2879              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2880              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2881              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2576  for (;;) Line 2883  for (;;)
2883            }            }
2884          else          else
2885  #endif  #endif
2886            /* Not UTF-8 mode */            /* Not UTF mode */
2887            {            {
2888            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2889              {              {
# Line 2586  for (;;) Line 2893  for (;;)
2893                break;                break;
2894                }                }
2895              c = *eptr;              c = *eptr;
2896              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
2897                if (c > 255)
2898                  {
2899                  if (op == OP_CLASS) break;
2900                  }
2901                else
2902    #endif
2903                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
2904              eptr++;              eptr++;
2905              }              }
2906            while (eptr >= pp)            while (eptr >= pp)
2907              {              {
2908              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM19);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM19);
2909              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2910              eptr--;              eptr--;
2911              }              }
# Line 2599  for (;;) Line 2913  for (;;)
2913    
2914          MRRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
2915          }          }
2916    #undef BYTE_MAP
2917        }        }
2918      /* Control never gets here */      /* Control never gets here */
2919    
# Line 2607  for (;;) Line 2922  for (;;)
2922      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8
2923      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
2924    
2925  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
2926      case OP_XCLASS:      case OP_XCLASS:
2927        {        {
2928        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2632  for (;;) Line 2947  for (;;)
2947          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2948          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2949          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2950          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2951          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2952          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2953          break;          break;
2954    
2955          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2652  for (;;) Line 2967  for (;;)
2967            MRRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
2968            }            }
2969          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
2970          if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data)) MRRETURN(MATCH_NOMATCH);
2971          }          }
2972    
2973        /* If max == min we can continue with the main loop without the        /* If max == min we can continue with the main loop without the
# Line 2667  for (;;) Line 2982  for (;;)
2982          {          {
2983          for (fi = min;; fi++)          for (fi = min;; fi++)
2984            {            {
2985            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);
2986            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2987            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
2988            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 2676  for (;;) Line 2991  for (;;)
2991              MRRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
2992              }              }
2993            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
2994            if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data)) MRRETURN(MATCH_NOMATCH);
2995            }            }
2996          /* Control never gets here */          /* Control never gets here */
2997          }          }
# Line 2694  for (;;) Line 3009  for (;;)
3009              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3010              break;              break;
3011              }              }
3012    #ifdef SUPPORT_UTF
3013            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3014            if (!_pcre_xclass(c, data)) break;  #else
3015              c = *eptr;
3016    #endif
3017              if (!PRIV(xclass)(c, data)) break;
3018            eptr += len;            eptr += len;
3019            }            }
3020          for(;;)          for(;;)
3021            {            {
3022            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3023            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3024            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3025            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3026              if (utf) BACKCHAR(eptr);
3027    #endif
3028            }            }
3029          MRRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
3030          }          }
# Line 2716  for (;;) Line 3037  for (;;)
3037    
3038      case OP_CHAR:      case OP_CHAR:
3039  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3040      if (utf8)      if (utf)
3041        {        {
3042        length = 1;        length = 1;
3043        ecode++;        ecode++;
# Line 2730  for (;;) Line 3051  for (;;)
3051        }        }
3052      else      else
3053  #endif  #endif
3054        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3055        {        {
3056        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3057          {          {
# Line 2745  for (;;) Line 3065  for (;;)
3065    
3066      /* Match a single character, caselessly */      /* Match a single character, caselessly */
3067    
3068      case OP_CHARNC:      case OP_CHARI:
3069  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3070      if (utf8)      if (utf)
3071        {        {
3072        length = 1;        length = 1;
3073        ecode++;        ecode++;
# Line 2790  for (;;) Line 3110  for (;;)
3110      else      else
3111  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF8 */
3112    
3113      /* Non-UTF-8 mode */      /* Not UTF mode */
3114        {        {
3115        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3116          {          {
# Line 2805  for (;;) Line 3125  for (;;)
3125      /* Match a single character repeatedly. */      /* Match a single character repeatedly. */
3126    
3127      case OP_EXACT:      case OP_EXACT:
3128        case OP_EXACTI:
3129      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3130      ecode += 3;      ecode += 1 + IMM2_SIZE;
3131      goto REPEATCHAR;      goto REPEATCHAR;
3132    
3133      case OP_POSUPTO:      case OP_POSUPTO:
3134        case OP_POSUPTOI:
3135      possessive = TRUE;      possessive = TRUE;
3136      /* Fall through */      /* Fall through */
3137    
3138      case OP_UPTO:      case OP_UPTO:
3139        case OP_UPTOI:
3140      case OP_MINUPTO:      case OP_MINUPTO:
3141        case OP_MINUPTOI:
3142      min = 0;      min = 0;
3143      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3144      minimize = *ecode == OP_MINUPTO;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3145      ecode += 3;      ecode += 1 + IMM2_SIZE;
3146      goto REPEATCHAR;      goto REPEATCHAR;
3147    
3148      case OP_POSSTAR:      case OP_POSSTAR:
3149        case OP_POSSTARI:
3150      possessive = TRUE;      possessive = TRUE;
3151      min = 0;      min = 0;
3152      max = INT_MAX;      max = INT_MAX;
# Line 2829  for (;;) Line 3154  for (;;)
3154      goto REPEATCHAR;      goto REPEATCHAR;
3155    
3156      case OP_POSPLUS:      case OP_POSPLUS:
3157        case OP_POSPLUSI:
3158      possessive = TRUE;      possessive = TRUE;
3159      min = 1;      min = 1;
3160      max = INT_MAX;      max = INT_MAX;
# Line 2836  for (;;) Line 3162  for (;;)
3162      goto REPEATCHAR;      goto REPEATCHAR;
3163    
3164      case OP_POSQUERY:      case OP_POSQUERY:
3165        case OP_POSQUERYI:
3166      possessive = TRUE;      possessive = TRUE;
3167      min = 0;      min = 0;
3168      max = 1;      max = 1;
# Line 2843  for (;;) Line 3170  for (;;)
3170      goto REPEATCHAR;      goto REPEATCHAR;
3171    
3172      case OP_STAR:      case OP_STAR:
3173        case OP_STARI:
3174      case OP_MINSTAR:      case OP_MINSTAR:
3175        case OP_MINSTARI:
3176      case OP_PLUS:      case OP_PLUS:
3177        case OP_PLUSI:
3178      case OP_MINPLUS:      case OP_MINPLUS:
3179        case OP_MINPLUSI:
3180      case OP_QUERY:      case OP_QUERY:
3181        case OP_QUERYI:
3182      case OP_MINQUERY:      case OP_MINQUERY:
3183      c = *ecode++ - OP_STAR;      case OP_MINQUERYI:
3184        c = *ecode++ - ((op < OP_STARI)? OP_STAR : OP_STARI);
3185      minimize = (c & 1) != 0;      minimize = (c & 1) != 0;
   
3186      min = rep_min[c];                 /* Pick up values from tables; */      min = rep_min[c];                 /* Pick up values from tables; */
3187      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
3188      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
# Line 2859  for (;;) Line 3191  for (;;)
3191    
3192      REPEATCHAR:      REPEATCHAR:
3193  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3194      if (utf8)      if (utf)
3195        {        {
3196        length = 1;        length = 1;
3197        charptr = ecode;        charptr = ecode;
# Line 2873  for (;;) Line 3205  for (;;)
3205          {          {
3206  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3207          unsigned int othercase;          unsigned int othercase;
3208          if ((ims & PCRE_CASELESS) != 0 &&          if (op >= OP_STARI &&     /* Caseless */
3209              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3210            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3211          else oclength = 0;          else oclength = 0;
3212  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3213    
# Line 2886  for (;;) Line 3218  for (;;)
3218  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3219            else if (oclength > 0 &&            else if (oclength > 0 &&
3220                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3221                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3222  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3223            else            else
3224              {              {
# Line 2901  for (;;) Line 3233  for (;;)
3233            {            {
3234            for (fi = min;; fi++)            for (fi = min;; fi++)
3235              {              {
3236              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM22);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);
3237              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3238              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3239              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
# Line 2909  for (;;) Line 3241  for (;;)
3241  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3242              else if (oclength > 0 &&              else if (oclength > 0 &&
3243                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3244                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3245  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3246              else              else
3247                {                {
# Line 2930  for (;;) Line 3262  for (;;)
3262  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3263              else if (oclength > 0 &&              else if (oclength > 0 &&
3264                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3265                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3266  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3267              else              else
3268                {                {
# Line 2943  for (;;) Line 3275  for (;;)
3275    
3276            for(;;)            for(;;)
3277              {              {
3278              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3279              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3280              if (eptr == pp) { MRRETURN(MATCH_NOMATCH); }              if (eptr == pp) { MRRETURN(MATCH_NOMATCH); }
3281  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2980  for (;;) Line 3312  for (;;)
3312      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3313        max, eptr));        max, eptr));
3314    
3315      if ((ims & PCRE_CASELESS) != 0)      if (op >= OP_STARI)  /* Caseless */
3316        {        {
3317        fc = md->lcc[fc];        fc = md->lcc[fc];
3318        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
# Line 2997  for (;;) Line 3329  for (;;)
3329          {          {
3330          for (fi = min;; fi++)          for (fi = min;; fi++)
3331            {            {
3332            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3333            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3334            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
3335            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3027  for (;;) Line 3359  for (;;)
3359    
3360          while (eptr >= pp)          while (eptr >= pp)
3361            {            {
3362            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM25);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);
3363            eptr--;            eptr--;
3364            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3365            }            }
# Line 3056  for (;;) Line 3388  for (;;)
3388          {          {
3389          for (fi = min;; fi++)          for (fi = min;; fi++)
3390            {            {
3391            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM26);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);
3392            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3393            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
3394            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3085  for (;;) Line 3417  for (;;)
3417    
3418          while (eptr >= pp)          while (eptr >= pp)
3419            {            {
3420            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);
3421            eptr--;            eptr--;
3422            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3423            }            }
# Line 3098  for (;;) Line 3430  for (;;)
3430      checking can be multibyte. */      checking can be multibyte. */
3431    
3432      case OP_NOT:      case OP_NOT:
3433        case OP_NOTI:
3434      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
3435        {        {
3436        SCHECK_PARTIAL();        SCHECK_PARTIAL();
# Line 3105  for (;;) Line 3438  for (;;)
3438        }        }
3439      ecode++;      ecode++;
3440      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
3441      if ((ims & PCRE_CASELESS) != 0)      if (op == OP_NOTI)         /* The caseless case */
3442        {        {
3443  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3444        if (c < 256)        if (c < 256)
# Line 3113  for (;;) Line 3446  for (;;)
3446        c = md->lcc[c];        c = md->lcc[c];
3447        if (md->lcc[*ecode++] == c) MRRETURN(MATCH_NOMATCH);        if (md->lcc[*ecode++] == c) MRRETURN(MATCH_NOMATCH);
3448        }        }
3449      else      else    /* Caseful */
3450        {        {
3451        if (*ecode++ == c) MRRETURN(MATCH_NOMATCH);        if (*ecode++ == c) MRRETURN(MATCH_NOMATCH);
3452        }        }
# Line 3127  for (;;) Line 3460  for (;;)
3460      about... */      about... */
3461    
3462      case OP_NOTEXACT:      case OP_NOTEXACT:
3463        case OP_NOTEXACTI:
3464      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3465      ecode += 3;      ecode += 1 + IMM2_SIZE;
3466      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3467    
3468      case OP_NOTUPTO:      case OP_NOTUPTO:
3469        case OP_NOTUPTOI:
3470      case OP_NOTMINUPTO:      case OP_NOTMINUPTO:
3471        case OP_NOTMINUPTOI:
3472      min = 0;      min = 0;
3473      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3474      minimize = *ecode == OP_NOTMINUPTO;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3475      ecode += 3;      ecode += 1 + IMM2_SIZE;
3476      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3477    
3478      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
3479        case OP_NOTPOSSTARI:
3480      possessive = TRUE;      possessive = TRUE;
3481      min = 0;      min = 0;
3482      max = INT_MAX;      max = INT_MAX;
# Line 3147  for (;;) Line 3484  for (;;)
3484      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3485    
3486      case OP_NOTPOSPLUS:      case OP_NOTPOSPLUS:
3487        case OP_NOTPOSPLUSI:
3488      possessive = TRUE;      possessive = TRUE;
3489      min = 1;      min = 1;
3490      max = INT_MAX;      max = INT_MAX;
# Line 3154  for (;;) Line 3492  for (;;)
3492      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3493    
3494      case OP_NOTPOSQUERY:      case OP_NOTPOSQUERY:
3495        case OP_NOTPOSQUERYI:
3496      possessive = TRUE;      possessive = TRUE;
3497      min = 0;      min = 0;
3498      max = 1;      max = 1;
# Line 3161  for (;;) Line 3500  for (;;)
3500      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3501    
3502      case OP_NOTPOSUPTO:      case OP_NOTPOSUPTO:
3503        case OP_NOTPOSUPTOI:
3504      possessive = TRUE;      possessive = TRUE;
3505      min = 0;      min = 0;
3506      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3507      ecode += 3;      ecode += 1 + IMM2_SIZE;
3508      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3509    
3510      case OP_NOTSTAR:      case OP_NOTSTAR:
3511        case OP_NOTSTARI:
3512      case OP_NOTMINSTAR:      case OP_NOTMINSTAR:
3513        case OP_NOTMINSTARI:
3514      case OP_NOTPLUS:      case OP_NOTPLUS:
3515        case OP_NOTPLUSI:
3516      case OP_NOTMINPLUS:      case OP_NOTMINPLUS:
3517        case OP_NOTMINPLUSI:
3518      case OP_NOTQUERY:      case OP_NOTQUERY:
3519        case OP_NOTQUERYI:
3520      case OP_NOTMINQUERY:      case OP_NOTMINQUERY:
3521      c = *ecode++ - OP_NOTSTAR;      case OP_NOTMINQUERYI:
3522        c = *ecode++ - ((op >= OP_NOTSTARI)? OP_NOTSTARI: OP_NOTSTAR);
3523      minimize = (c & 1) != 0;      minimize = (c & 1) != 0;
3524      min = rep_min[c];                 /* Pick up values from tables; */      min = rep_min[c];                 /* Pick up values from tables; */
3525      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
# Line 3195  for (;;) Line 3541  for (;;)
3541      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,
3542        max, eptr));        max, eptr));
3543    
3544      if ((ims & PCRE_CASELESS) != 0)      if (op >= OP_NOTSTARI)     /* Caseless */
3545        {        {
3546        fc = md->lcc[fc];        fc = md->lcc[fc];
3547    
3548  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3549        /* UTF-8 mode */        if (utf)
       if (utf8)  
3550          {          {
3551          register unsigned int d;          register unsigned int d;
3552          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3218  for (;;) Line 3563  for (;;)
3563          }          }
3564        else        else
3565  #endif  #endif
3566          /* Not UTF mode */
       /* Not UTF-8 mode */  
3567          {          {
3568          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3569            {            {
# Line 3237  for (;;) Line 3581  for (;;)
3581        if (minimize)        if (minimize)
3582          {          {
3583  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3584          /* UTF-8 mode */          if (utf)
         if (utf8)  
3585            {            {
3586            register unsigned int d;            register unsigned int d;
3587            for (fi = min;; fi++)            for (fi = min;; fi++)
3588              {              {
3589              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
3590              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3591              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3592              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 3258  for (;;) Line 3601  for (;;)
3601            }            }
3602          else          else
3603  #endif  #endif
3604          /* Not UTF-8 mode */          /* Not UTF mode */
3605            {            {
3606            for (fi = min;; fi++)            for (fi = min;; fi++)
3607              {              {
3608              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM29);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);
3609              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3610              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3611              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 3283  for (;;) Line 3626  for (;;)
3626          pp = eptr;          pp = eptr;
3627    
3628  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3629          /* UTF-8 mode */          if (utf)
         if (utf8)  
3630            {            {
3631            register unsigned int d;            register unsigned int d;
3632            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3303  for (;;) Line 3645  for (;;)
3645          if (possessive) continue;          if (possessive) continue;
3646          for(;;)          for(;;)
3647              {              {
3648              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3649              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3650              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
3651              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3311  for (;;) Line 3653  for (;;)
3653            }            }
3654          else          else
3655  #endif  #endif
3656          /* Not UTF-8 mode */          /* Not UTF mode */
3657            {            {
3658            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3659              {              {
# Line 3326  for (;;) Line 3668  for (;;)
3668            if (possessive) continue;            if (possessive) continue;
3669            while (eptr >= pp)            while (eptr >= pp)
3670              {              {
3671              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM31);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);
3672              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3673              eptr--;              eptr--;
3674              }              }
# Line 3342  for (;;) Line 3684  for (;;)
3684      else      else
3685        {        {
3686  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3687        /* UTF-8 mode */        if (utf)
       if (utf8)  
3688          {          {
3689          register unsigned int d;          register unsigned int d;
3690          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3359  for (;;) Line 3700  for (;;)
3700          }          }
3701        else        else
3702  #endif  #endif
3703        /* Not UTF-8 mode */        /* Not UTF mode */
3704          {          {
3705          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3706            {            {
# Line 3377  for (;;) Line 3718  for (;;)
3718        if (minimize)        if (minimize)
3719          {          {
3720  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3721          /* UTF-8 mode */          if (utf)
         if (utf8)  
3722            {            {
3723            register unsigned int d;            register unsigned int d;
3724            for (fi = min;; fi++)            for (fi = min;; fi++)
3725              {              {
3726              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
3727              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3728              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3729              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 3397  for (;;) Line 3737  for (;;)
3737            }            }
3738          else          else
3739  #endif  #endif
3740          /* Not UTF-8 mode */          /* Not UTF mode */
3741            {            {
3742            for (fi = min;; fi++)            for (fi = min;; fi++)
3743              {              {
3744              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM33);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);
3745              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3746              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3747              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 3422  for (;;) Line 3762  for (;;)
3762          pp = eptr;          pp = eptr;
3763    
3764  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3765          /* UTF-8 mode */          if (utf)
         if (utf8)  
3766            {            {
3767            register unsigned int d;            register unsigned int d;
3768            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3441  for (;;) Line 3780  for (;;)
3780            if (possessive) continue;            if (possessive) continue;
3781            for(;;)            for(;;)
3782              {              {
3783              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM34);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);
3784              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3785              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
3786              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3449  for (;;) Line 3788  for (;;)
3788            }            }
3789          else          else
3790  #endif  #endif
3791          /* Not UTF-8 mode */          /* Not UTF mode */
3792            {            {
3793            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3794              {              {
# Line 3464  for (;;) Line 3803  for (;;)
3803            if (possessive) continue;            if (possessive) continue;
3804            while (eptr >= pp)            while (eptr >= pp)
3805              {              {
3806              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM35);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);
3807              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3808              eptr--;              eptr--;
3809              }              }
# Line 3482  for (;;) Line 3821  for (;;)
3821      case OP_TYPEEXACT:      case OP_TYPEEXACT:
3822      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3823      minimize = TRUE;      minimize = TRUE;
3824      ecode += 3;      ecode += 1 + IMM2_SIZE;
3825      goto REPEATTYPE;      goto REPEATTYPE;
3826    
3827      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3490  for (;;) Line 3829  for (;;)
3829      min = 0;      min = 0;
3830      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3831      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
3832      ecode += 3;      ecode += 1 + IMM2_SIZE;
3833      goto REPEATTYPE;      goto REPEATTYPE;
3834    
3835      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3518  for (;;) Line 3857  for (;;)
3857      possessive = TRUE;      possessive = TRUE;
3858      min = 0;      min = 0;
3859      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3860      ecode += 3;      ecode += 1 + IMM2_SIZE;
3861      goto REPEATTYPE;      goto REPEATTYPE;
3862    
3863      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 3579  for (;;) Line 3918  for (;;)
3918            case PT_LAMP:            case PT_LAMP:
3919            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3920              {              {
3921                int chartype;
3922              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3923                {                {
3924                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3925                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3926                }                }
3927              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3928              prop_chartype = UCD_CHARTYPE(c);              chartype = UCD_CHARTYPE(c);
3929              if ((prop_chartype == ucp_Lu ||              if ((chartype == ucp_Lu ||
3930                   prop_chartype == ucp_Ll ||                   chartype == ucp_Ll ||
3931                   prop_chartype == ucp_Lt) == prop_fail_result)                   chartype == ucp_Lt) == prop_fail_result)
3932                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3933              }              }
3934            break;            break;
# Line 3602  for (;;) Line 3942  for (;;)
3942                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3943                }                }
3944              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3945              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)
             if ((prop_category == prop_value) == prop_fail_result)  
3946                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3947              }              }
3948            break;            break;
# Line 3617  for (;;) Line 3956  for (;;)
3956                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3957                }                }
3958              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3959              prop_chartype = UCD_CHARTYPE(c);              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)
             if ((prop_chartype == prop_value) == prop_fail_result)  
3960                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3961              }              }
3962            break;            break;
# Line 3632  for (;;) Line 3970  for (;;)
3970                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3971                }                }
3972              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3973              prop_script = UCD_SCRIPT(c);              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)
             if ((prop_script == prop_value) == prop_fail_result)  
3974                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3975              }              }
3976            break;            break;
# Line 3641  for (;;) Line 3978  for (;;)
3978            case PT_ALNUM:            case PT_ALNUM:
3979            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3980              {              {
3981                int category;
3982              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3983                {                {
3984                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3985                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3986                }                }
3987              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3988              prop_category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
3989              if ((prop_category == ucp_L || prop_category == ucp_N)              if ((category == ucp_L || category == ucp_N) == prop_fail_result)
                    == prop_fail_result)  
3990                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3991              }              }
3992            break;            break;
# Line 3663  for (;;) Line 4000  for (;;)
4000                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4001                }                }
4002              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4003              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
             if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||  
4004                   c == CHAR_FF || c == CHAR_CR)                   c == CHAR_FF || c == CHAR_CR)
4005                     == prop_fail_result)                     == prop_fail_result)
4006                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
# Line 3680  for (;;) Line 4016  for (;;)
4016                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4017                }                }
4018              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4019              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
             if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||  
4020                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
4021                     == prop_fail_result)                     == prop_fail_result)
4022                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
# Line 3691  for (;;) Line 4026  for (;;)
4026            case PT_WORD:            case PT_WORD:
4027            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4028              {              {
4029                int category;
4030              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4031                {                {
4032                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4033                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4034                }                }
4035              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4036              prop_category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4037              if ((prop_category == ucp_L || prop_category == ucp_N ||              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)
                  c == CHAR_UNDERSCORE)  
4038                     == prop_fail_result)                     == prop_fail_result)
4039                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4040              }              }
# Line 3725  for (;;) Line 4060  for (;;)
4060              MRRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
4061              }              }
4062            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
4063            prop_category = UCD_CATEGORY(c);            if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH);
           if (prop_category == ucp_M) MRRETURN(MATCH_NOMATCH);  
4064            while (eptr < md->end_subject)            while (eptr < md->end_subject)
4065              {              {
4066              int len = 1;              int len = 1;
4067              if (!utf8) c = *eptr;              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4068                else { GETCHARLEN(c, eptr, len); }              if (UCD_CATEGORY(c) != ucp_M) break;
             prop_category = UCD_CATEGORY(c);  
             if (prop_category != ucp_M) break;  
4069              eptr += len;              eptr += len;
4070              }              }
4071            }            }
# Line 3744  for (;;) Line 4076  for (;;)
4076    
4077  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4078    
4079  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4080        if (utf8) switch(ctype)        if (utf) switch(ctype)
4081          {          {
4082          case OP_ANY:          case OP_ANY:
4083          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3757  for (;;) Line 4089  for (;;)
4089              }              }
4090            if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);
4091            eptr++;            eptr++;
4092            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4093            }            }
4094          break;          break;
4095    
# Line 3770  for (;;) Line 4102  for (;;)
4102              MRRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
4103              }              }
4104            eptr++;            eptr++;
4105            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4106            }            }
4107          break;          break;
4108    
# Line 3791  for (;;) Line 4123  for (;;)
4123            switch(c)            switch(c)
4124              {              {
4125              default: MRRETURN(MATCH_NOMATCH);              default: MRRETURN(MATCH_NOMATCH);
4126    
4127              case 0x000d:              case 0x000d:
4128              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
4129              break;              break;
# Line 3967  for (;;) Line 4300  for (;;)
4300              }              }
4301            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)
4302              MRRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
4303            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4304              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4305            }            }
4306          break;          break;
4307    
# Line 3995  for (;;) Line 4329  for (;;)
4329              }              }
4330            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)
4331              MRRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
4332            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4333              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4334            }            }
4335          break;          break;
4336    
# Line 4067  for (;;) Line 4402  for (;;)
4402            switch(*eptr++)            switch(*eptr++)
4403              {              {
4404              default: MRRETURN(MATCH_NOMATCH);              default: MRRETURN(MATCH_NOMATCH);
4405    
4406              case 0x000d:              case 0x000d:
4407              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
4408              break;              break;
4409    
4410              case 0x000a:              case 0x000a:
4411              break;              break;
4412    
# Line 4259  for (;;) Line 4596  for (;;)
4596            case PT_ANY:            case PT_ANY:
4597            for (fi = min;; fi++)            for (fi = min;; fi++)
4598              {              {
4599              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM36);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM36);
4600              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4601              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4602              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4275  for (;;) Line 4612  for (;;)
4612            case PT_LAMP:            case PT_LAMP:
4613            for (fi = min;; fi++)            for (fi = min;; fi++)
4614              {              {
4615              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM37);              int chartype;
4616                RMATCH(eptr, ecode, offset_top, md, eptrb, RM37);
4617              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4618              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4619              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4284  for (;;) Line 4622  for (;;)
4622                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4623                }                }
4624              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4625              prop_chartype = UCD_CHARTYPE(c);              chartype = UCD_CHARTYPE(c);
4626              if ((prop_chartype == ucp_Lu ||              if ((chartype == ucp_Lu ||
4627                   prop_chartype == ucp_Ll ||                   chartype == ucp_Ll ||
4628                   prop_chartype == ucp_Lt) == prop_fail_result)                   chartype == ucp_Lt) == prop_fail_result)
4629                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4630              }              }
4631            /* Control never gets here */            /* Control never gets here */
# Line 4295  for (;;) Line 4633  for (;;)
4633            case PT_GC:            case PT_GC:
4634            for (fi = min;; fi++)            for (fi = min;; fi++)
4635              {              {
4636              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM38);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM38);
4637              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4638              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4639              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4304  for (;;) Line 4642  for (;;)
4642                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4643                }                }
4644              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4645              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)
             if ((prop_category == prop_value) == prop_fail_result)  
4646                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4647              }              }
4648            /* Control never gets here */            /* Control never gets here */
# Line 4313  for (;;) Line 4650  for (;;)
4650            case PT_PC:            case PT_PC:
4651            for (fi = min;; fi++)            for (fi = min;; fi++)
4652              {              {
4653              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM39);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM39);
4654              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4655              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4656              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4322  for (;;) Line 4659  for (;;)
4659                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4660                }                }
4661              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4662              prop_chartype = UCD_CHARTYPE(c);              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)
             if ((prop_chartype == prop_value) == prop_fail_result)  
4663                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4664              }              }
4665            /* Control never gets here */            /* Control never gets here */
# Line 4331  for (;;) Line 4667  for (;;)
4667            case PT_SC:            case PT_SC:
4668            for (fi = min;; fi++)            for (fi = min;; fi++)
4669              {              {
4670              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM40);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM40);
4671              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4672              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4673              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4340  for (;;) Line 4676  for (;;)
4676                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4677                }                }
4678              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4679              prop_script = UCD_SCRIPT(c);              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)
             if ((prop_script == prop_value) == prop_fail_result)  
4680                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4681              }              }
4682            /* Control never gets here */            /* Control never gets here */
# Line 4349  for (;;) Line 4684  for (;;)
4684            case PT_ALNUM:            case PT_ALNUM:
4685            for (fi = min;; fi++)            for (fi = min;; fi++)
4686              {              {
4687              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM59);              int category;
4688                RMATCH(eptr, ecode, offset_top, md, eptrb, RM59);
4689              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4690              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4691              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4358  for (;;) Line 4694  for (;;)
4694                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4695                }                }
4696              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4697              prop_category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4698              if ((prop_category == ucp_L || prop_category == ucp_N)              if ((category == ucp_L || category == ucp_N) == prop_fail_result)
                    == prop_fail_result)  
4699                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4700              }              }
4701            /* Control never gets here */            /* Control never gets here */
# Line 4368  for (;;) Line 4703  for (;;)
4703            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
4704            for (fi = min;; fi++)            for (fi = min;; fi++)
4705              {              {
4706              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM60);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM60);
4707              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4708              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4709              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4377  for (;;) Line 4712  for (;;)
4712                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4713                }                }
4714              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4715              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
             if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||  
4716                   c == CHAR_FF || c == CHAR_CR)                   c == CHAR_FF || c == CHAR_CR)
4717                     == prop_fail_result)                     == prop_fail_result)
4718                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
# Line 4388  for (;;) Line 4722  for (;;)
4722            case PT_PXSPACE:  /* POSIX space */            case PT_PXSPACE:  /* POSIX space */
4723            for (fi = min;; fi++)            for (fi = min;; fi++)
4724              {              {
4725              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM61);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM61);
4726              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4727              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4728              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4397  for (;;) Line 4731  for (;;)
4731                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4732                }                }
4733              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4734              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
             if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||  
4735                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
4736                     == prop_fail_result)                     == prop_fail_result)
4737                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
# Line 4408  for (;;) Line 4741  for (;;)
4741            case PT_WORD:            case PT_WORD:
4742            for (fi = min;; fi++)            for (fi = min;; fi++)
4743              {              {
4744              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM62);              int category;
4745                RMATCH(eptr, ecode, offset_top, md, eptrb, RM62);
4746              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4747              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4748              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4417  for (;;) Line 4751  for (;;)
4751                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4752                }                }
4753              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4754              prop_category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4755              if ((prop_category == ucp_L ||              if ((category == ucp_L ||
4756                   prop_category == ucp_N ||                   category == ucp_N ||
4757                   c == CHAR_UNDERSCORE)                   c == CHAR_UNDERSCORE)
4758                     == prop_fail_result)                     == prop_fail_result)
4759                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
# Line 4440  for (;;) Line 4774  for (;;)
4774          {          {
4775          for (fi = min;; fi++)          for (fi = min;; fi++)
4776            {            {
4777            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM41);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM41);
4778            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4779            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
4780            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 4449  for (;;) Line 4783  for (;;)
4783              MRRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
4784              }              }
4785            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
4786            prop_category = UCD_CATEGORY(c);            if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH);
           if (prop_category == ucp_M) MRRETURN(MATCH_NOMATCH);  
4787            while (eptr < md->end_subject)            while (eptr < md->end_subject)
4788              {              {
4789              int len = 1;              int len = 1;
4790              if (!utf8) c = *eptr;              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4791                else { GETCHARLEN(c, eptr, len); }              if (UCD_CATEGORY(c) != ucp_M) break;
             prop_category = UCD_CATEGORY(c);  
             if (prop_category != ucp_M) break;  
4792              eptr += len;              eptr += len;
4793              }              }
4794            }            }
4795          }          }
   
4796        else        else
4797  #endif     /* SUPPORT_UCP */  #endif     /* SUPPORT_UCP */
4798    
4799  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4800        /* UTF-8 mode */        if (utf)
       if (utf8)  
4801          {          {
4802          for (fi = min;; fi++)          for (fi = min;; fi++)
4803            {            {
4804            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM42);
4805            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4806            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
4807            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 4631  for (;;) Line 4960  for (;;)
4960          }          }
4961        else        else
4962  #endif  #endif
4963        /* Not UTF-8 mode */        /* Not UTF mode */
4964          {          {
4965          for (fi = min;; fi++)          for (fi = min;; fi++)
4966            {            {
4967            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM43);
4968            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4969            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
4970            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 4783  for (;;) Line 5112  for (;;)
5112            case PT_LAMP:            case PT_LAMP:
5113            for (i = min; i < max; i++)            for (i = min; i < max; i++)
5114              {              {
5115                int chartype;
5116              int len = 1;              int len = 1;
5117              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
5118                {                {
# Line 4790  for (;;) Line 5120  for (;;)
5120                break;                break;
5121                }                }
5122              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5123              prop_chartype = UCD_CHARTYPE(c);              chartype = UCD_CHARTYPE(c);
5124              if ((prop_chartype == ucp_Lu ||              if ((chartype == ucp_Lu ||
5125                   prop_chartype == ucp_Ll ||                   chartype == ucp_Ll ||
5126                   prop_chartype == ucp_Lt) == prop_fail_result)                   chartype == ucp_Lt) == prop_fail_result)
5127                break;                break;
5128              eptr+= len;              eptr+= len;
5129              }              }
# Line 4809  for (;;) Line 5139  for (;;)
5139                break;                break;
5140                }                }
5141              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5142              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result) break;
             if ((prop_category == prop_value) == prop_fail_result)  
               break;  
5143              eptr+= len;              eptr+= len;
5144              }              }
5145            break;            break;
# Line 4826  for (;;) Line 5154  for (;;)
5154                break;                break;
5155                }                }
5156              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5157              prop_chartype = UCD_CHARTYPE(c);              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result) break;
             if ((prop_chartype == prop_value) == prop_fail_result)  
               break;  
5158              eptr+= len;              eptr+= len;
5159              }              }
5160            break;            break;
# Line 4843  for (;;) Line 5169  for (;;)
5169                break;                break;
5170                }                }
5171              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5172              prop_script = UCD_SCRIPT(c);              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result) break;
             if ((prop_script == prop_value) == prop_fail_result)  
               break;  
5173              eptr+= len;              eptr+= len;
5174              }              }
5175            break;            break;
# Line 4853  for (;;) Line 5177  for (;;)
5177            case PT_ALNUM:            case PT_ALNUM:
5178            for (i = min; i < max; i++)            for (i = min; i < max; i++)
5179              {              {
5180                int category;
5181              int len = 1;              int len = 1;
5182              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
5183                {                {
# Line 4860  for (;;) Line 5185  for (;;)
5185                break;                break;
5186                }                }
5187              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5188              prop_category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
5189              if ((prop_category == ucp_L || prop_category == ucp_N)              if ((category == ucp_L || category == ucp_N) == prop_fail_result)
                  == prop_fail_result)  
5190                break;                break;
5191              eptr+= len;              eptr+= len;
5192              }              }
# Line 4878  for (;;) Line 5202  for (;;)
5202                break;                break;
5203                }                }
5204              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5205              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
             if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||  
5206                   c == CHAR_FF || c == CHAR_CR)                   c == CHAR_FF || c == CHAR_CR)
5207                   == prop_fail_result)                   == prop_fail_result)
5208                break;                break;
# Line 4897  for (;;) Line 5220  for (;;)
5220                break;                break;
5221                }                }
5222              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5223              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
             if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||  
5224                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
5225                   == prop_fail_result)                   == prop_fail_result)
5226                break;                break;
# Line 4909  for (;;) Line 5231  for (;;)
5231            case PT_WORD:            case PT_WORD:
5232            for (i = min; i < max; i++)            for (i = min; i < max; i++)
5233              {              {
5234                int category;
5235              int len = 1;              int len = 1;
5236              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
5237                {                {
# Line 4916  for (;;) Line 5239  for (;;)
5239                break;                break;
5240                }                }
5241              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5242              prop_category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
5243              if ((prop_category == ucp_L || prop_category == ucp_N ||              if ((category == ucp_L || category == ucp_N ||
5244                   c == CHAR_UNDERSCORE) == prop_fail_result)                   c == CHAR_UNDERSCORE) == prop_fail_result)
5245                break;                break;
5246              eptr+= len;              eptr+= len;
# Line 4933  for (;;) Line 5256  for (;;)
5256          if (possessive) continue;          if (possessive) continue;
5257          for(;;)          for(;;)
5258            {            {
5259            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5260            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5261            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5262            if (utf8) BACKCHAR(eptr);            if (utf) BACKCHAR(eptr);
5263            }            }
5264          }          }
5265    
# Line 4947  for (;;) Line 5270  for (;;)
5270          {          {
5271          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5272            {            {
5273              int len = 1;
5274            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5275              {              {
5276              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5277              break;              break;
5278              }              }
5279            GETCHARINCTEST(c, eptr);            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5280            prop_category = UCD_CATEGORY(c);            if (UCD_CATEGORY(c) == ucp_M) break;
5281            if (prop_category == ucp_M) break;            eptr += len;
5282            while (eptr < md->end_subject)            while (eptr < md->end_subject)
5283              {              {
5284              int len = 1;              len = 1;
5285              if (!utf8) c = *eptr; else              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5286                {              if (UCD_CATEGORY(c) != ucp_M) break;
               GETCHARLEN(c, eptr, len);  
               }  
             prop_category = UCD_CATEGORY(c);  
             if (prop_category != ucp_M) break;  
5287              eptr += len;              eptr += len;
5288              }              }
5289            }            }
# Line 4974  for (;;) Line 5294  for (;;)
5294    
5295          for(;;)          for(;;)
5296            {            {
5297            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM45);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM45);
5298            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5299            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5300            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
5301              {              {
5302              int len = 1;              if (!utf) c = *eptr; else
             if (!utf8) c = *eptr; else  
5303                {                {
5304                BACKCHAR(eptr);                BACKCHAR(eptr);
5305                GETCHARLEN(c, eptr, len);                GETCHAR(c, eptr);
5306                }                }
5307              prop_category = UCD_CATEGORY(c);              if (UCD_CATEGORY(c) != ucp_M) break;
             if (prop_category != ucp_M) break;  
5308              eptr--;              eptr--;
5309              }              }
5310            }            }
# Line 4995  for (;;) Line 5313  for (;;)
5313        else        else
5314  #endif   /* SUPPORT_UCP */  #endif   /* SUPPORT_UCP */
5315    
5316  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5317        /* UTF-8 mode */        if (utf)
   
       if (utf8)  
5318          {          {
5319          switch(ctype)          switch(ctype)
5320            {            {
# Line 5014  for (;;) Line 5330  for (;;)
5330                  }                  }
5331                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5332                eptr++;                eptr++;
5333                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5334                }                }
5335              }              }
5336    
# Line 5031  for (;;) Line 5347  for (;;)
5347                  }                  }
5348                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5349                eptr++;                eptr++;
5350                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5351                }                }
5352              }              }
5353            break;            break;
# Line 5047  for (;;) Line 5363  for (;;)
5363                  break;                  break;
5364                  }                  }
5365                eptr++;                eptr++;
5366                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5367                }                }
5368              }              }
5369            else eptr = md->end_subject;   /* Unlimited UTF-8 repeat */            else
5370                {
5371                eptr = md->end_subject;   /* Unlimited UTF-8 repeat */
5372                SCHECK_PARTIAL();
5373                }
5374            break;            break;
5375    
5376            /* The byte case is the same as non-UTF8 */            /* The byte case is the same as non-UTF8 */
# Line 5258  for (;;) Line 5578  for (;;)
5578            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5579            }            }