/[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 795 by zherczeg, Sat Dec 10 02:20:06 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_UTF
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_UTF
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 */
529    
530  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
531  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
532  #endif  #endif
533  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
# 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_UTF
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    recursion_info new_recursive;
599    
600  #ifdef SUPPORT_UTF8                /* Many of these variables are used only  */  BOOL cur_is_word;
 const uschar *charptr;             /* in small blocks of the code. My normal */  
 #endif                             /* style of coding would have declared    */  
 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    #define foc           number
638    
639  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
640  variables. */  variables. */
641    
# Line 664  defined). However, RMATCH isn't like a f Line 660  defined). However, RMATCH isn't like a f
660  complicated macro. It has to be used in one particular way. This shouldn't,  complicated macro. It has to be used in one particular way. This shouldn't,
661  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
662    
663  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
664  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
665  #else  #else
666  utf8 = FALSE;  utf = FALSE;
667  #endif  #endif
668    
669  /* 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 672  haven't exceeded the recursive call limi
672  if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT);  if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT);
673  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);
674    
 original_ims = ims;    /* Save for resetting on ')' */  
   
675  /* 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
676  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
677  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
678  hit the closing ket, in order to break infinite loops that match no characters.  up space on the stack. See also MATCH_CONDASSERT below.
679  When match() is called in other circumstances, don't add to the chain. The  
680  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
681  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
682  match(). */  to break infinite loops that match no characters. When match() is called in
683    other circumstances, don't add to the chain. The MATCH_CBEGROUP feature must
684    NOT be used with tail recursion, because the memory block that is used is on
685    the stack, so a new one may be required for each match(). */
686    
687  if ((flags & match_cbegroup) != 0)  if (md->match_function_type == MATCH_CBEGROUP)
688    {    {
689    newptrb.epb_saved_eptr = eptr;    newptrb.epb_saved_eptr = eptr;
690    newptrb.epb_prev = eptrb;    newptrb.epb_prev = eptrb;
691    eptrb = &newptrb;    eptrb = &newptrb;
692      md->match_function_type = 0;
693    }    }
694    
695  /* Now start processing the opcodes. */  /* Now start processing the opcodes. */
# Line 705  for (;;) Line 703  for (;;)
703      {      {
704      case OP_MARK:      case OP_MARK:
705      markptr = ecode + 2;      markptr = ecode + 2;
706      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
707        ims, eptrb, flags, RM55);        eptrb, RM55);
708    
709      /* 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
710      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 714  for (;;)
714      unaltered. */      unaltered. */
715    
716      if (rrc == MATCH_SKIP_ARG &&      if (rrc == MATCH_SKIP_ARG &&
717          strcmp((char *)markptr, (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC(markptr, md->start_match_ptr) == 0)
718        {        {
719        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
720        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
# Line 731  for (;;) Line 729  for (;;)
729      /* COMMIT overrides PRUNE, SKIP, and THEN */      /* COMMIT overrides PRUNE, SKIP, and THEN */
730    
731      case OP_COMMIT:      case OP_COMMIT:
732      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
733        ims, eptrb, flags, RM52);        eptrb, RM52);
734      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&
735          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&
736          rrc != MATCH_THEN)          rrc != MATCH_THEN)
# Line 742  for (;;) Line 740  for (;;)
740      /* PRUNE overrides THEN */      /* PRUNE overrides THEN */
741    
742      case OP_PRUNE:      case OP_PRUNE:
743      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
744        ims, eptrb, flags, RM51);        eptrb, RM51);
745      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
746      MRRETURN(MATCH_PRUNE);      MRRETURN(MATCH_PRUNE);
747    
748      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
749      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
750        ims, eptrb, flags, RM56);        eptrb, RM56);
751      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
752      md->mark = ecode + 2;      md->mark = ecode + 2;
753      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
# Line 757  for (;;) Line 755  for (;;)
755      /* SKIP overrides PRUNE and THEN */      /* SKIP overrides PRUNE and THEN */
756    
757      case OP_SKIP:      case OP_SKIP:
758      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
759        ims, eptrb, flags, RM53);        eptrb, RM53);
760      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
761        RRETURN(rrc);        RRETURN(rrc);
762      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
763      MRRETURN(MATCH_SKIP);      MRRETURN(MATCH_SKIP);
764    
765      case OP_SKIP_ARG:      case OP_SKIP_ARG:
766      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
767        ims, eptrb, flags, RM57);        eptrb, RM57);
768      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
769        RRETURN(rrc);        RRETURN(rrc);
770    
# Line 778  for (;;) Line 776  for (;;)
776      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
777      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
778    
779      /* 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
780      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
781      to skip back past alternatives that precede the THEN within the current      match pointer to do this. */
     branch. */  
782    
783      case OP_THEN:      case OP_THEN:
784      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
785        ims, eptrb, flags, RM54);        eptrb, RM54);
786      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
787      md->start_match_ptr = ecode - GET(ecode, 1);      md->start_match_ptr = ecode;
788      MRRETURN(MATCH_THEN);      MRRETURN(MATCH_THEN);
789    
790      case OP_THEN_ARG:      case OP_THEN_ARG:
791      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1+LINK_SIZE],      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
792        offset_top, md, ims, eptrb, flags, RM58);        md, eptrb, RM58);
793      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
794      md->start_match_ptr = ecode - GET(ecode, 1);      md->start_match_ptr = ecode;
795      md->mark = ecode + LINK_SIZE + 2;      md->mark = ecode + 2;
796      RRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
797    
798      /* Handle a capturing bracket. If there is space in the offset vector, save      /* Handle an atomic group that does not contain any capturing parentheses.
799      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
800      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
801      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.
802      reference inside the group.      However, this uses a lot more stack, so in 8.20, atomic groups that do not
803        contain any captures generate OP_ONCE_NC, which can be handled in the old,
804      If the bracket fails to match, we need to restore this value and also the      less stack intensive way.
805      values of the final offsets, in case they were set by a previous iteration  
806      of the same bracket.      Check the alternative branches in turn - the matching won't pass the KET
807        for this kind of subpattern. If any one branch matches, we carry on as at
808        the end of a normal bracket, leaving the subject pointer, but resetting
809        the start-of-match value in case it was changed by \K. */
810    
811        case OP_ONCE_NC:
812        prev = ecode;
813        saved_eptr = eptr;
814        do
815          {
816          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
817          if (rrc == MATCH_MATCH)  /* Note: _not_ MATCH_ACCEPT */
818            {
819            mstart = md->start_match_ptr;
820            markptr = md->mark;
821            break;
822            }
823          if (rrc == MATCH_THEN)
824            {
825            next = ecode + GET(ecode,1);
826            if (md->start_match_ptr < next &&
827                (*ecode == OP_ALT || *next == OP_ALT))
828              rrc = MATCH_NOMATCH;
829            }
830    
831          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
832          ecode += GET(ecode,1);
833          }
834        while (*ecode == OP_ALT);
835    
836        /* If hit the end of the group (which could be repeated), fail */
837    
838        if (*ecode != OP_ONCE_NC && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);
839    
840        /* Continue as from after the group, updating the offsets high water
841        mark, since extracts may have been taken. */
842    
843        do ecode += GET(ecode, 1); while (*ecode == OP_ALT);
844    
845        offset_top = md->end_offset_top;
846        eptr = md->end_match_ptr;
847    
848        /* For a non-repeating ket, just continue at this level. This also
849        happens for a repeating ket if no characters were matched in the group.
850        This is the forcible breaking of infinite loops as implemented in Perl
851        5.005. */
852    
853        if (*ecode == OP_KET || eptr == saved_eptr)
854          {
855          ecode += 1+LINK_SIZE;
856          break;
857          }
858    
859        /* The repeating kets try the rest of the pattern or restart from the
860        preceding bracket, in the appropriate order. The second "call" of match()
861        uses tail recursion, to avoid using another stack frame. */
862    
863        if (*ecode == OP_KETRMIN)
864          {
865          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM65);
866          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
867          ecode = prev;
868          goto TAIL_RECURSE;
869          }
870        else  /* OP_KETRMAX */
871          {
872          md->match_function_type = MATCH_CBEGROUP;
873          RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
874          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
875          ecode += 1 + LINK_SIZE;
876          goto TAIL_RECURSE;
877          }
878        /* Control never gets here */
879    
880        /* Handle a capturing bracket, other than those that are possessive with an
881        unlimited repeat. If there is space in the offset vector, save the current
882        subject position in the working slot at the top of the vector. We mustn't
883        change the current values of the data slot, because they may be set from a
884        previous iteration of this group, and be referred to by a reference inside
885        the group. A failure to match might occur after the group has succeeded,
886        if something later on doesn't match. For this reason, we need to restore
887        the working value and also the values of the final offsets, in case they
888        were set by a previous iteration of the same bracket.
889    
890      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
891      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 914  for (;;)
914        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
915          (int)(eptr - md->start_subject);          (int)(eptr - md->start_subject);
916    
917        flags = (op == OP_SCBRA)? match_cbegroup : 0;        for (;;)
       do  
918          {          {
919          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
920            ims, eptrb, flags, RM1);          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
921          if (rrc != MATCH_NOMATCH &&            eptrb, RM1);
922              (rrc != MATCH_THEN || md->start_match_ptr != ecode))          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
923            RRETURN(rrc);  
924            /* If we backed up to a THEN, check whether it is within the current
925            branch by comparing the address of the THEN that is passed back with
926            the end of the branch. If it is within the current branch, and the
927            branch is one of two or more alternatives (it either starts or ends
928            with OP_ALT), we have reached the limit of THEN's action, so convert
929            the return code to NOMATCH, which will cause normal backtracking to
930            happen from now on. Otherwise, THEN is passed back to an outer
931            alternative. This implements Perl's treatment of parenthesized groups,
932            where a group not containing | does not affect the current alternative,
933            that is, (X) is NOT the same as (X|(*F)). */
934    
935            if (rrc == MATCH_THEN)
936              {
937              next = ecode + GET(ecode,1);
938              if (md->start_match_ptr < next &&
939                  (*ecode == OP_ALT || *next == OP_ALT))
940                rrc = MATCH_NOMATCH;
941              }
942    
943            /* Anything other than NOMATCH is passed back. */
944    
945            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
946          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
947          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
948            if (*ecode != OP_ALT) break;
949          }          }
       while (*ecode == OP_ALT);  
950    
951        DPRINTF(("bracket %d failed\n", number));        DPRINTF(("bracket %d failed\n", number));
   
952        md->offset_vector[offset] = save_offset1;        md->offset_vector[offset] = save_offset1;
953        md->offset_vector[offset+1] = save_offset2;        md->offset_vector[offset+1] = save_offset2;
954        md->offset_vector[md->offset_end - number] = save_offset3;        md->offset_vector[md->offset_end - number] = save_offset3;
955    
956        if (rrc != MATCH_THEN) md->mark = markptr;        /* At this point, rrc will be one of MATCH_ONCE or MATCH_NOMATCH. */
957        RRETURN(MATCH_NOMATCH);  
958          if (md->mark == NULL) md->mark = markptr;
959          RRETURN(rrc);
960        }        }
961    
962      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
# Line 869  for (;;) Line 970  for (;;)
970      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
971      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
972    
973      /* Non-capturing bracket. Loop for all the alternatives. When we get to the      /* Non-capturing or atomic group, except for possessive with unlimited
974      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.*/  
975    
976        When we get to the final alternative within the brackets, we used to return
977        the result of a recursive call to match() whatever happened so it was
978        possible to reduce stack usage by turning this into a tail recursion,
979        except in the case of a possibly empty group. However, now that there is
980        the possiblity of (*THEN) occurring in the final alternative, this
981        optimization is no longer always possible.
982    
983        We can optimize if we know there are no (*THEN)s in the pattern; at present
984        this is the best that can be done.
985    
986        MATCH_ONCE is returned when the end of an atomic group is successfully
987        reached, but subsequent matching fails. It passes back up the tree (causing
988        captured values to be reset) until the original atomic group level is
989        reached. This is tested by comparing md->once_target with the start of the
990        group. At this point, the return is converted into MATCH_NOMATCH so that
991        previous backup points can be taken. */
992    
993        case OP_ONCE:
994      case OP_BRA:      case OP_BRA:
995      case OP_SBRA:      case OP_SBRA:
996      DPRINTF(("start non-capturing bracket\n"));      DPRINTF(("start non-capturing bracket\n"));
997      flags = (op >= OP_SBRA)? match_cbegroup : 0;  
998      for (;;)      for (;;)
999        {        {
1000        if (ecode[GET(ecode, 1)] != OP_ALT)   /* Final alternative */        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;
1001    
1002          /* If this is not a possibly empty group, and there are no (*THEN)s in
1003          the pattern, and this is the final alternative, optimize as described
1004          above. */
1005    
1006          else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1007          {          {
1008          if (flags == 0)    /* Not a possibly empty group */          ecode += PRIV(OP_lengths)[*ecode];
1009            goto TAIL_RECURSE;
1010            }
1011    
1012          /* In all other cases, we have to make another call to match(). */
1013    
1014          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1015            RM2);
1016    
1017          /* See comment in the code for capturing groups above about handling
1018          THEN. */
1019    
1020          if (rrc == MATCH_THEN)
1021            {
1022            next = ecode + GET(ecode,1);
1023            if (md->start_match_ptr < next &&
1024                (*ecode == OP_ALT || *next == OP_ALT))
1025              rrc = MATCH_NOMATCH;
1026            }
1027    
1028          if (rrc != MATCH_NOMATCH)
1029            {
1030            if (rrc == MATCH_ONCE)
1031            {            {
1032            ecode += _pcre_OP_lengths[*ecode];            const pcre_uchar *scode = ecode;
1033            DPRINTF(("bracket 0 tail recursion\n"));            if (*scode != OP_ONCE)           /* If not at start, find it */
1034            goto TAIL_RECURSE;              {
1035                while (*scode == OP_ALT) scode += GET(scode, 1);
1036                scode -= GET(scode, 1);
1037                }
1038              if (md->once_target == scode) rrc = MATCH_NOMATCH;
1039            }            }
1040            RRETURN(rrc);
1041            }
1042          ecode += GET(ecode, 1);
1043          if (*ecode != OP_ALT) break;
1044          }
1045    
1046          /* Possibly empty group; can't use tail recursion. */      if (md->mark == NULL) md->mark = markptr;
1047        RRETURN(MATCH_NOMATCH);
1048    
1049          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,      /* Handle possessive capturing brackets with an unlimited repeat. We come
1050            eptrb, flags, RM48);      here from BRAZERO with allow_zero set TRUE. The offset_vector values are
1051          if (rrc == MATCH_NOMATCH) md->mark = markptr;      handled similarly to the normal case above. However, the matching is
1052          RRETURN(rrc);      different. The end of these brackets will always be OP_KETRPOS, which
1053        returns MATCH_KETRPOS without going further in the pattern. By this means
1054        we can handle the group by iteration rather than recursion, thereby
1055        reducing the amount of stack needed. */
1056    
1057        case OP_CBRAPOS:
1058        case OP_SCBRAPOS:
1059        allow_zero = FALSE;
1060    
1061        POSSESSIVE_CAPTURE:
1062        number = GET2(ecode, 1+LINK_SIZE);
1063        offset = number << 1;
1064    
1065    #ifdef PCRE_DEBUG
1066        printf("start possessive bracket %d\n", number);
1067        printf("subject=");
1068        pchars(eptr, 16, TRUE, md);
1069        printf("\n");
1070    #endif
1071    
1072        if (offset < md->offset_max)
1073          {
1074          matched_once = FALSE;
1075          code_offset = ecode - md->start_code;
1076    
1077          save_offset1 = md->offset_vector[offset];
1078          save_offset2 = md->offset_vector[offset+1];
1079          save_offset3 = md->offset_vector[md->offset_end - number];
1080          save_capture_last = md->capture_last;
1081    
1082          DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
1083    
1084          /* Each time round the loop, save the current subject position for use
1085          when the group matches. For MATCH_MATCH, the group has matched, so we
1086          restart it with a new subject starting position, remembering that we had
1087          at least one match. For MATCH_NOMATCH, carry on with the alternatives, as
1088          usual. If we haven't matched any alternatives in any iteration, check to
1089          see if a previous iteration matched. If so, the group has matched;
1090          continue from afterwards. Otherwise it has failed; restore the previous
1091          capture values before returning NOMATCH. */
1092    
1093          for (;;)
1094            {
1095            md->offset_vector[md->offset_end - number] =
1096              (int)(eptr - md->start_subject);
1097            if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1098            RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1099              eptrb, RM63);
1100            if (rrc == MATCH_KETRPOS)
1101              {
1102              offset_top = md->end_offset_top;
1103              eptr = md->end_match_ptr;
1104              ecode = md->start_code + code_offset;
1105              save_capture_last = md->capture_last;
1106              matched_once = TRUE;
1107              continue;
1108              }
1109    
1110            /* See comment in the code for capturing groups above about handling
1111            THEN. */
1112    
1113            if (rrc == MATCH_THEN)
1114              {
1115              next = ecode + GET(ecode,1);
1116              if (md->start_match_ptr < next &&
1117                  (*ecode == OP_ALT || *next == OP_ALT))
1118                rrc = MATCH_NOMATCH;
1119              }
1120    
1121            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1122            md->capture_last = save_capture_last;
1123            ecode += GET(ecode, 1);
1124            if (*ecode != OP_ALT) break;
1125          }          }
1126    
1127        /* For non-final alternatives, continue the loop for a NOMATCH result;        if (!matched_once)
1128        otherwise return. */          {
1129            md->offset_vector[offset] = save_offset1;
1130            md->offset_vector[offset+1] = save_offset2;
1131            md->offset_vector[md->offset_end - number] = save_offset3;
1132            }
1133    
1134        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,        if (md->mark == NULL) md->mark = markptr;
1135          eptrb, flags, RM2);        if (allow_zero || matched_once)
1136        if (rrc != MATCH_NOMATCH &&          {
1137            (rrc != MATCH_THEN || md->start_match_ptr != ecode))          ecode += 1 + LINK_SIZE;
1138          RRETURN(rrc);          break;
1139            }
1140    
1141          RRETURN(MATCH_NOMATCH);
1142          }
1143    
1144        /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
1145        as a non-capturing bracket. */
1146    
1147        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1148        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1149    
1150        DPRINTF(("insufficient capture room: treat as non-capturing\n"));
1151    
1152        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1153        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1154    
1155        /* Non-capturing possessive bracket with unlimited repeat. We come here
1156        from BRAZERO with allow_zero = TRUE. The code is similar to the above,
1157        without the capturing complication. It is written out separately for speed
1158        and cleanliness. */
1159    
1160        case OP_BRAPOS:
1161        case OP_SBRAPOS:
1162        allow_zero = FALSE;
1163    
1164        POSSESSIVE_NON_CAPTURE:
1165        matched_once = FALSE;
1166        code_offset = ecode - md->start_code;
1167    
1168        for (;;)
1169          {
1170          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1171          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1172            eptrb, RM48);
1173          if (rrc == MATCH_KETRPOS)
1174            {
1175            offset_top = md->end_offset_top;
1176            eptr = md->end_match_ptr;
1177            ecode = md->start_code + code_offset;
1178            matched_once = TRUE;
1179            continue;
1180            }
1181    
1182          /* See comment in the code for capturing groups above about handling
1183          THEN. */
1184    
1185          if (rrc == MATCH_THEN)
1186            {
1187            next = ecode + GET(ecode,1);
1188            if (md->start_match_ptr < next &&
1189                (*ecode == OP_ALT || *next == OP_ALT))
1190              rrc = MATCH_NOMATCH;
1191            }
1192    
1193          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1194        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1195          if (*ecode != OP_ALT) break;
1196        }        }
1197    
1198        if (matched_once || allow_zero)
1199          {
1200          ecode += 1 + LINK_SIZE;
1201          break;
1202          }
1203        RRETURN(MATCH_NOMATCH);
1204    
1205      /* Control never reaches here. */      /* Control never reaches here. */
1206    
1207      /* Conditional group: compilation checked that there are no more than      /* Conditional group: compilation checked that there are no more than
1208      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
1209      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
1210      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. */  
1211    
1212      case OP_COND:      case OP_COND:
1213      case OP_SCOND:      case OP_SCOND:
1214      codelink= GET(ecode, 1);      codelink = GET(ecode, 1);
1215    
1216      /* 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
1217      inserted between OP_COND and an assertion condition. */      inserted between OP_COND and an assertion condition. */
# Line 928  for (;;) Line 1221  for (;;)
1221        if (pcre_callout != NULL)        if (pcre_callout != NULL)
1222          {          {
1223          pcre_callout_block cb;          pcre_callout_block cb;
1224          cb.version          = 1;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1225          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1226          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1227          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
# Line 940  for (;;) Line 1233  for (;;)
1233          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1234          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last;
1235          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1236            cb.mark             = (unsigned char *)markptr;
1237          if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);          if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);
1238          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1239          }          }
1240        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1241        }        }
1242    
1243      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 960  for (;;) Line 1254  for (;;)
1254        else        else
1255          {          {
1256          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1257          condition =  (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1258    
1259          /* 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
1260          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
1261          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
1262          if any one is set. */          if any one is set. */
1263    
1264          if (!condition && condcode == OP_NRREF && recno != RREF_ANY)          if (!condition && condcode == OP_NRREF)
1265            {            {
1266            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1267            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1268              {              {
1269              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 982  for (;;) Line 1276  for (;;)
1276    
1277            if (i < md->name_count)            if (i < md->name_count)
1278              {              {
1279              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1280              while (slotB > md->name_table)              while (slotB > md->name_table)
1281                {                {
1282                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1283                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1284                  {                  {
1285                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1286                  if (condition) break;                  if (condition) break;
# Line 1002  for (;;) Line 1296  for (;;)
1296                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1297                  {                  {
1298                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1299                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1300                    {                    {
1301                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1302                    if (condition) break;                    if (condition) break;
# Line 1015  for (;;) Line 1309  for (;;)
1309    
1310          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1311    
1312          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1313          }          }
1314        }        }
1315    
# Line 1032  for (;;) Line 1326  for (;;)
1326        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1327          {          {
1328          int refno = offset >> 1;          int refno = offset >> 1;
1329          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1330    
1331          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1332            {            {
# Line 1046  for (;;) Line 1340  for (;;)
1340    
1341          if (i < md->name_count)          if (i < md->name_count)
1342            {            {
1343            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1344            while (slotB > md->name_table)            while (slotB > md->name_table)
1345              {              {
1346              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1347              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1348                {                {
1349                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1350                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1068  for (;;) Line 1362  for (;;)
1362              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1363                {                {
1364                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1365                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1366                  {                  {
1367                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1368                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1083  for (;;) Line 1377  for (;;)
1377    
1378        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1379    
1380        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1381        }        }
1382    
1383      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1093  for (;;) Line 1387  for (;;)
1387        }        }
1388    
1389      /* The condition is an assertion. Call match() to evaluate it - setting      /* The condition is an assertion. Call match() to evaluate it - setting
1390      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
1391      assertion. */      an assertion. */
1392    
1393      else      else
1394        {        {
1395        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        md->match_function_type = MATCH_CONDASSERT;
1396            match_condassert, RM3);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);
1397        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
1398          {          {
1399            if (md->end_offset_top > offset_top)
1400              offset_top = md->end_offset_top;  /* Captures may have happened */
1401          condition = TRUE;          condition = TRUE;
1402          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
1403          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
1404          }          }
1405        else if (rrc != MATCH_NOMATCH &&  
1406                (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an
1407          assertion; it is therefore treated as NOMATCH. */
1408    
1409          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1410          {          {
1411          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
1412          }          }
# Line 1118  for (;;) Line 1417  for (;;)
1417          }          }
1418        }        }
1419    
1420      /* 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
1421      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
1422      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
1423      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
1424        exist, in which case we can just plough on. Note that, for compatibility
1425        with Perl, the | in a conditional group is NOT treated as creating two
1426        alternatives. If a THEN is encountered in the branch, it propagates out to
1427        the enclosing alternative (unless nested in a deeper set of alternatives,
1428        of course). */
1429    
1430      if (condition || *ecode == OP_ALT)      if (condition || *ecode == OP_ALT)
1431        {        {
1432        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 */  
1433          {          {
1434          flags = 0;          ecode += 1 + LINK_SIZE;
1435          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1436          }          }
1437    
1438          md->match_function_type = MATCH_CBEGROUP;
1439          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);
1440          RRETURN(rrc);
1441        }        }
1442      else                         /* Condition false & no alternative */  
1443         /* Condition false & no alternative; continue after the group. */
1444    
1445        else
1446        {        {
1447        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1448        }        }
# Line 1164  for (;;) Line 1469  for (;;)
1469        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1470        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1471        }        }
1472      ecode += 3;      ecode += 1 + IMM2_SIZE;
1473      break;      break;
1474    
1475    
1476      /* 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. */  
1477    
     case OP_ACCEPT:  
1478      case OP_END:      case OP_END:
1479      if (md->recursive != NULL && md->recursive->group_num == 0)      case OP_ACCEPT:
1480        {      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;  
       }  
1481    
1482      /* 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
1483      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
1484      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,
1485      if any. */      backtracking will then try other alternatives, if any. */
1486    
1487      if (eptr == mstart &&      if (eptr == mstart && op != OP_ASSERT_ACCEPT &&
1488          (md->notempty ||           md->recursive == NULL &&
1489            (md->notempty_atstart &&           (md->notempty ||
1490              mstart == md->start_subject + md->start_offset)))             (md->notempty_atstart &&
1491                 mstart == md->start_subject + md->start_offset)))
1492        MRRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
1493    
1494      /* Otherwise, we have a match. */      /* Otherwise, we have a match. */
# Line 1210  for (;;) Line 1503  for (;;)
1503      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;
1504      MRRETURN(rrc);      MRRETURN(rrc);
1505    
     /* Change option settings */  
   
     case OP_OPT:  
     ims = ecode[1];  
     ecode += 2;  
     DPRINTF(("ims set to %02lx\n", ims));  
     break;  
   
1506      /* Assertion brackets. Check the alternative branches in turn - the      /* Assertion brackets. Check the alternative branches in turn - the
1507      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,
1508      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
1509      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
1510      this level is identical to the lookahead case. */      this level is identical to the lookahead case. When the assertion is part
1511        of a condition, we want to return immediately afterwards. The caller of
1512        this incarnation of the match() function will have set MATCH_CONDASSERT in
1513        md->match_function type, and one of these opcodes will be the first opcode
1514        that is processed. We use a local variable that is preserved over calls to
1515        match() to remember this case. */
1516    
1517      case OP_ASSERT:      case OP_ASSERT:
1518      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1519        if (md->match_function_type == MATCH_CONDASSERT)
1520          {
1521          condassert = TRUE;
1522          md->match_function_type = 0;
1523          }
1524        else condassert = FALSE;
1525    
1526      do      do
1527        {        {
1528        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
         RM4);  
1529        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1530          {          {
1531          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1532            markptr = md->mark;
1533          break;          break;
1534          }          }
1535        if (rrc != MATCH_NOMATCH &&  
1536            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE does not allow THEN to escape beyond an assertion; it is treated
1537          RRETURN(rrc);        as NOMATCH. */
1538    
1539          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1540        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1541        }        }
1542      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1543    
1544      if (*ecode == OP_KET) MRRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) MRRETURN(MATCH_NOMATCH);
1545    
1546      /* If checking an assertion for a condition, return MATCH_MATCH. */      /* If checking an assertion for a condition, return MATCH_MATCH. */
1547    
1548      if ((flags & match_condassert) != 0) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);
1549    
1550      /* Continue from after the assertion, updating the offsets high water      /* Continue from after the assertion, updating the offsets high water
1551      mark, since extracts may have been taken during the assertion. */      mark, since extracts may have been taken during the assertion. */
# Line 1261  for (;;) Line 1561  for (;;)
1561    
1562      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1563      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1564        if (md->match_function_type == MATCH_CONDASSERT)
1565          {
1566          condassert = TRUE;
1567          md->match_function_type = 0;
1568          }
1569        else condassert = FALSE;
1570    
1571      do      do
1572        {        {
1573        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
         RM5);  
1574        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) MRRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) MRRETURN(MATCH_NOMATCH);
1575        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1576          {          {
1577          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1578          break;          break;
1579          }          }
1580        if (rrc != MATCH_NOMATCH &&  
1581            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE does not allow THEN to escape beyond an assertion; it is treated
1582          RRETURN(rrc);        as NOMATCH. */
1583    
1584          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1585        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1586        }        }
1587      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1588    
1589      if ((flags & match_condassert) != 0) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1590    
1591      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1592      continue;      continue;
# Line 1289  for (;;) Line 1597  for (;;)
1597      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1598    
1599      case OP_REVERSE:      case OP_REVERSE:
1600  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1601      if (utf8)      if (utf)
1602        {        {
1603        i = GET(ecode, 1);        i = GET(ecode, 1);
1604        while (i-- > 0)        while (i-- > 0)
# Line 1324  for (;;) Line 1632  for (;;)
1632      if (pcre_callout != NULL)      if (pcre_callout != NULL)
1633        {        {
1634        pcre_callout_block cb;        pcre_callout_block cb;
1635        cb.version          = 1;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1636        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1637        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1638        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
# Line 1336  for (;;) Line 1644  for (;;)
1644        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1645        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last;
1646        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1647          cb.mark             = (unsigned char *)markptr;
1648        if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);        if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);
1649        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1650        }        }
# Line 1346  for (;;) Line 1655  for (;;)
1655      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
1656      whole pattern. (This is so that it works from duplicated subpatterns.)      whole pattern. (This is so that it works from duplicated subpatterns.)
1657    
1658      If there are any capturing brackets started but not finished, we have to      The state of the capturing groups is preserved over recursion, and
1659      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
1660      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
1661      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
1662      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
1663      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
1664      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.  
1665    
1666      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
1667      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
1668      for the original version of this logic. */      for the original version of this logic. It has, however, been hacked around
1669        a lot, so he is not to blame for the current way it works. */
1670    
1671      case OP_RECURSE:      case OP_RECURSE:
1672        {        {
1673          recursion_info *ri;
1674          int recno;
1675    
1676        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1677        new_recursive.group_num = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
1678          GET2(callpat, 1 + LINK_SIZE);          GET2(callpat, 1 + LINK_SIZE);
1679    
1680          /* Check for repeating a recursion without advancing the subject pointer.
1681          This should catch convoluted mutual recursions. (Some simple cases are
1682          caught at compile time.) */
1683    
1684          for (ri = md->recursive; ri != NULL; ri = ri->prevrec)
1685            if (recno == ri->group_num && eptr == ri->subject_position)
1686              RRETURN(PCRE_ERROR_RECURSELOOP);
1687    
1688        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1689    
1690          new_recursive.group_num = recno;
1691          new_recursive.subject_position = eptr;
1692        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1693        md->recursive = &new_recursive;        md->recursive = &new_recursive;
1694    
1695        /* Find where to continue from afterwards */        /* Where to continue from afterwards */
1696    
1697        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
       new_recursive.after_call = ecode;  
1698    
1699        /* Now save the offset data. */        /* Now save the offset data */
1700    
1701        new_recursive.saved_max = md->offset_end;        new_recursive.saved_max = md->offset_end;
1702        if (new_recursive.saved_max <= REC_STACK_SAVE_MAX)        if (new_recursive.saved_max <= REC_STACK_SAVE_MAX)
# Line 1388  for (;;) Line 1707  for (;;)
1707            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));
1708          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1709          }          }
   
1710        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1711              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
       new_recursive.save_offset_top = offset_top;  
1712    
1713        /* OK, now we can do the recursion. For each top-level alternative we        /* OK, now we can do the recursion. After processing each alternative,
1714        restore the offset and recursion data. */        restore the offset data. If there were nested recursions, md->recursive
1715          might be changed, so reset it before looping. */
1716    
1717        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1718        flags = (*callpat >= OP_SBRA)? match_cbegroup : 0;        cbegroup = (*callpat >= OP_SBRA);
1719        do        do
1720          {          {
1721          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1722            md, ims, eptrb, flags, RM6);          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1723              md, eptrb, RM6);
1724            memcpy(md->offset_vector, new_recursive.offset_save,
1725                new_recursive.saved_max * sizeof(int));
1726            md->recursive = new_recursive.prevrec;
1727          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1728            {            {
1729            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
           md->recursive = new_recursive.prevrec;  
1730            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1731              (pcre_free)(new_recursive.offset_save);              (pcre_free)(new_recursive.offset_save);
1732            MRRETURN(MATCH_MATCH);  
1733              /* Set where we got to in the subject, and reset the start in case
1734              it was changed by \K. This *is* propagated back out of a recursion,
1735              for Perl compatibility. */
1736    
1737              eptr = md->end_match_ptr;
1738              mstart = md->start_match_ptr;
1739              goto RECURSION_MATCHED;        /* Exit loop; end processing */
1740            }            }
1741          else if (rrc != MATCH_NOMATCH &&  
1742                  (rrc != MATCH_THEN || md->start_match_ptr != ecode))          /* PCRE does not allow THEN to escape beyond a recursion; it is treated
1743            as NOMATCH. */
1744    
1745            else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1746            {            {
1747            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1748            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
# Line 1420  for (;;) Line 1751  for (;;)
1751            }            }
1752    
1753          md->recursive = &new_recursive;          md->recursive = &new_recursive;
         memcpy(md->offset_vector, new_recursive.offset_save,  
             new_recursive.saved_max * sizeof(int));  
1754          callpat += GET(callpat, 1);          callpat += GET(callpat, 1);
1755          }          }
1756        while (*callpat == OP_ALT);        while (*callpat == OP_ALT);
# Line 1432  for (;;) Line 1761  for (;;)
1761          (pcre_free)(new_recursive.offset_save);          (pcre_free)(new_recursive.offset_save);
1762        MRRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
1763        }        }
     /* 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;  
1764    
1765      /* For a non-repeating ket, just continue at this level. This also      RECURSION_MATCHED:
1766      happens for a repeating ket if no characters were matched in the group.      break;
     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));  
       }  
   
     if (*ecode == OP_KETRMIN)  
       {  
       RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM8);  
       if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
       ecode = prev;  
       flags = 0;  
       goto TAIL_RECURSE;  
       }  
     else  /* OP_KETRMAX */  
       {  
       RMATCH(eptr, prev, offset_top, md, ims, eptrb, match_cbegroup, RM9);  
       if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
       ecode += 1 + LINK_SIZE;  
       flags = 0;  
       goto TAIL_RECURSE;  
       }  
     /* Control never gets here */  
1767    
1768      /* 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
1769      bracketed group and go to there. */      bracketed group and go to there. */
# Line 1529  for (;;) Line 1779  for (;;)
1779      optional ones preceded by BRAZERO or BRAMINZERO. */      optional ones preceded by BRAZERO or BRAMINZERO. */
1780    
1781      case OP_BRAZERO:      case OP_BRAZERO:
1782        {      next = ecode + 1;
1783        next = ecode+1;      RMATCH(eptr, next, offset_top, md, eptrb, RM10);
1784        RMATCH(eptr, next, offset_top, md, ims, eptrb, 0, RM10);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1785        if (rrc != MATCH_NOMATCH) RRETURN(rrc);      do next += GET(next, 1); while (*next == OP_ALT);
1786        do next += GET(next,1); while (*next == OP_ALT);      ecode = next + 1 + LINK_SIZE;
       ecode = next + 1 + LINK_SIZE;  
       }  
1787      break;      break;
1788    
1789      case OP_BRAMINZERO:      case OP_BRAMINZERO:
1790        {      next = ecode + 1;
1791        next = ecode+1;      do next += GET(next, 1); while (*next == OP_ALT);
1792        do next += GET(next, 1); while (*next == OP_ALT);      RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, eptrb, RM11);
1793        RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0, RM11);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1794        if (rrc != MATCH_NOMATCH) RRETURN(rrc);      ecode++;
       ecode++;  
       }  
1795      break;      break;
1796    
1797      case OP_SKIPZERO:      case OP_SKIPZERO:
1798        {      next = ecode+1;
1799        next = ecode+1;      do next += GET(next,1); while (*next == OP_ALT);
1800        do next += GET(next,1); while (*next == OP_ALT);      ecode = next + 1 + LINK_SIZE;
       ecode = next + 1 + LINK_SIZE;  
       }  
1801      break;      break;
1802    
1803        /* BRAPOSZERO occurs before a possessive bracket group. Don't do anything
1804        here; just jump to the group, with allow_zero set TRUE. */
1805    
1806        case OP_BRAPOSZERO:
1807        op = *(++ecode);
1808        allow_zero = TRUE;
1809        if (op == OP_CBRAPOS || op == OP_SCBRAPOS) goto POSSESSIVE_CAPTURE;
1810          goto POSSESSIVE_NON_CAPTURE;
1811    
1812      /* End of a group, repeated or non-repeating. */      /* End of a group, repeated or non-repeating. */
1813    
1814      case OP_KET:      case OP_KET:
1815      case OP_KETRMIN:      case OP_KETRMIN:
1816      case OP_KETRMAX:      case OP_KETRMAX:
1817        case OP_KETRPOS:
1818      prev = ecode - GET(ecode, 1);      prev = ecode - GET(ecode, 1);
1819    
1820      /* 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
1821      infinite repeats of empty string matches, retrieve the subject start from      infinite repeats of empty string matches, retrieve the subject start from
1822      the chain. Otherwise, set it NULL. */      the chain. Otherwise, set it NULL. */
1823    
1824      if (*prev >= OP_SBRA)      if (*prev >= OP_SBRA || *prev == OP_ONCE)
1825        {        {
1826        saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */        saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */
1827        eptrb = eptrb->epb_prev;              /* Backup to previous group */        eptrb = eptrb->epb_prev;              /* Backup to previous group */
1828        }        }
1829      else saved_eptr = NULL;      else saved_eptr = NULL;
1830    
1831      /* 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
1832      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
1833      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
1834      it was changed by \K. */      start in case it was changed by \K. */
1835    
1836      if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||      if ((*prev >= OP_ASSERT && *prev <= OP_ASSERTBACK_NOT) ||
1837          *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||           *prev == OP_ONCE_NC)
         *prev == OP_ONCE)  
1838        {        {
1839        md->end_match_ptr = eptr;      /* For ONCE */        md->end_match_ptr = eptr;      /* For ONCE_NC */
1840        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
1841        md->start_match_ptr = mstart;        md->start_match_ptr = mstart;
1842        MRRETURN(MATCH_MATCH);        MRRETURN(MATCH_MATCH);         /* Sets md->mark */
1843        }        }
1844    
1845      /* 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
1846      and if necessary complete handling an extraction by setting the offsets and      and if necessary complete handling an extraction by setting the offsets and
1847      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
1848      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
1849      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
1850        the current subject position and start match pointer and give a MATCH
1851        return. */
1852    
1853      if (*prev == OP_CBRA || *prev == OP_SCBRA)      if (*prev == OP_CBRA || *prev == OP_SCBRA ||
1854            *prev == OP_CBRAPOS || *prev == OP_SCBRAPOS)
1855        {        {
1856        number = GET2(prev, 1+LINK_SIZE);        number = GET2(prev, 1+LINK_SIZE);
1857        offset = number << 1;        offset = number << 1;
# Line 1605  for (;;) Line 1861  for (;;)
1861        printf("\n");        printf("\n");
1862  #endif  #endif
1863    
1864          /* Handle a recursively called group. */
1865    
1866          if (md->recursive != NULL && md->recursive->group_num == number)
1867            {
1868            md->end_match_ptr = eptr;
1869            md->start_match_ptr = mstart;
1870            RRETURN(MATCH_MATCH);
1871            }
1872    
1873          /* Deal with capturing */
1874    
1875        md->capture_last = number;        md->capture_last = number;
1876        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->offset_overflow = TRUE; else
1877          {          {
1878            /* If offset is greater than offset_top, it means that we are
1879            "skipping" a capturing group, and that group's offsets must be marked
1880            unset. In earlier versions of PCRE, all the offsets were unset at the
1881            start of matching, but this doesn't work because atomic groups and
1882            assertions can cause a value to be set that should later be unset.
1883            Example: matching /(?>(a))b|(a)c/ against "ac". This sets group 1 as
1884            part of the atomic group, but this is not on the final matching path,
1885            so must be unset when 2 is set. (If there is no group 2, there is no
1886            problem, because offset_top will then be 2, indicating no capture.) */
1887    
1888            if (offset > offset_top)
1889              {
1890              register int *iptr = md->offset_vector + offset_top;
1891              register int *iend = md->offset_vector + offset;
1892              while (iptr < iend) *iptr++ = -1;
1893              }
1894    
1895            /* Now make the extraction */
1896    
1897          md->offset_vector[offset] =          md->offset_vector[offset] =
1898            md->offset_vector[md->offset_end - number];            md->offset_vector[md->offset_end - number];
1899          md->offset_vector[offset+1] = (int)(eptr - md->start_subject);          md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1900          if (offset_top <= offset) offset_top = offset + 2;          if (offset_top <= offset) offset_top = offset + 2;
1901          }          }
1902          }
1903    
1904        /* Handle a recursively called group. Restore the offsets      /* For an ordinary non-repeating ket, just continue at this level. This
1905        appropriately and continue from after the call. */      also happens for a repeating ket if no characters were matched in the
1906        group. This is the forcible breaking of infinite loops as implemented in
1907        Perl 5.005. For a non-repeating atomic group that includes captures,
1908        establish a backup point by processing the rest of the pattern at a lower
1909        level. If this results in a NOMATCH return, pass MATCH_ONCE back to the
1910        original OP_ONCE level, thereby bypassing intermediate backup points, but
1911        resetting any captures that happened along the way. */
1912    
1913        if (md->recursive != NULL && md->recursive->group_num == number)      if (*ecode == OP_KET || eptr == saved_eptr)
1914          {
1915          if (*prev == OP_ONCE)
1916          {          {
1917          recursion_info *rec = md->recursive;          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM12);
1918          DPRINTF(("Recursion (%d) succeeded - continuing\n", number));          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1919          md->recursive = rec->prevrec;          md->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */
1920          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;  
1921          }          }
1922          ecode += 1 + LINK_SIZE;    /* Carry on at this level */
1923          break;
1924        }        }
1925    
1926      /* For both capturing and non-capturing groups, reset the value of the ims      /* OP_KETRPOS is a possessive repeating ket. Remember the current position,
1927      flags, in case they got changed during the group. */      and return the MATCH_KETRPOS. This makes it possible to do the repeats one
1928        at a time from the outer level, thus saving stack. */
     ims = original_ims;  
     DPRINTF(("ims reset to %02lx\n", ims));  
   
     /* For a non-repeating ket, just continue at this level. This also  
     happens for a repeating ket if no characters were matched in the group.  
     This is the forcible breaking of infinite loops as implemented in Perl  
     5.005. If there is an options reset, it will get obeyed in the normal  
     course of events. */  
1929    
1930      if (*ecode == OP_KET || eptr == saved_eptr)      if (*ecode == OP_KETRPOS)
1931        {        {
1932        ecode += 1 + LINK_SIZE;        md->end_match_ptr = eptr;
1933        break;        md->end_offset_top = offset_top;
1934          RRETURN(MATCH_KETRPOS);
1935        }        }
1936    
1937      /* 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
1938      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
1939      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
1940      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
1941        string. */
     flags = (*prev >= OP_SBRA)? match_cbegroup : 0;  
1942    
1943      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1944        {        {
1945        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM12);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM7);
1946        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1947        if (flags != 0)    /* Could match an empty string */        if (*prev == OP_ONCE)
1948          {          {
1949          RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM8);
1950            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1951            md->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */
1952            RRETURN(MATCH_ONCE);
1953            }
1954          if (*prev >= OP_SBRA)    /* Could match an empty string */
1955            {
1956            md->match_function_type = MATCH_CBEGROUP;
1957            RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
1958          RRETURN(rrc);          RRETURN(rrc);
1959          }          }
1960        ecode = prev;        ecode = prev;
# Line 1670  for (;;) Line 1962  for (;;)
1962        }        }
1963      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
1964        {        {
1965        RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13);        if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1966          RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
1967          if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
1968        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1969          if (*prev == OP_ONCE)
1970            {
1971            RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM9);
1972            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1973            md->once_target = prev;
1974            RRETURN(MATCH_ONCE);
1975            }
1976        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
       flags = 0;  
1977        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1978        }        }
1979      /* Control never gets here */      /* Control never gets here */
1980    
1981      /* Start of subject unless notbol, or after internal newline if multiline */      /* Not multiline mode: start of subject assertion, unless notbol. */
1982    
1983      case OP_CIRC:      case OP_CIRC:
1984      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 */  
1985    
1986      /* Start of subject assertion */      /* Start of subject assertion */
1987    
# Line 1699  for (;;) Line 1990  for (;;)
1990      ecode++;      ecode++;
1991      break;      break;
1992    
1993        /* Multiline mode: start of subject unless notbol, or after any newline. */
1994    
1995        case OP_CIRCM:
1996        if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);
1997        if (eptr != md->start_subject &&
1998            (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
1999          MRRETURN(MATCH_NOMATCH);
2000        ecode++;
2001        break;
2002    
2003      /* Start of match assertion */      /* Start of match assertion */
2004    
2005      case OP_SOM:      case OP_SOM:
# Line 1713  for (;;) Line 2014  for (;;)
2014      ecode++;      ecode++;
2015      break;      break;
2016    
2017      /* Assert before internal newline if multiline, or before a terminating      /* Multiline mode: assert before any newline, or before end of subject
2018      newline unless endonly is set, else end of subject unless noteol is set. */      unless noteol is set. */
2019    
2020      case OP_DOLL:      case OP_DOLLM:
2021      if ((ims & PCRE_MULTILINE) != 0)      if (eptr < md->end_subject)
2022        {        { if (!IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH); }
2023        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 */  
2024        {        {
2025        if (md->noteol) MRRETURN(MATCH_NOMATCH);        if (md->noteol) MRRETURN(MATCH_NOMATCH);
2026        if (!md->endonly) goto ASSERT_NL_OR_EOS;        SCHECK_PARTIAL();
2027        }        }
2028        ecode++;
2029        break;
2030    
2031        /* Not multiline mode: assert before a terminating newline or before end of
2032        subject unless noteol is set. */
2033    
2034        case OP_DOLL:
2035        if (md->noteol) MRRETURN(MATCH_NOMATCH);
2036        if (!md->endonly) goto ASSERT_NL_OR_EOS;
2037    
2038      /* ... else fall through for endonly */      /* ... else fall through for endonly */
2039    
# Line 1770  for (;;) Line 2070  for (;;)
2070        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2071        partial matching. */        partial matching. */
2072    
2073  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2074        if (utf8)        if (utf)
2075          {          {
2076          /* Get status of previous character */          /* Get status of previous character */
2077    
2078          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2079            {            {
2080            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2081            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2082            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2083            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2084  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 1884  for (;;) Line 2184  for (;;)
2184      /* Fall through */      /* Fall through */
2185    
2186      case OP_ALLANY:      case OP_ALLANY:
2187      if (eptr++ >= md->end_subject)      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2188        {        {                            /* not be updated before SCHECK_PARTIAL. */
2189        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2190        MRRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2191        }        }
2192      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;      eptr++;
2193    #ifdef SUPPORT_UTF
2194        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2195    #endif
2196      ecode++;      ecode++;
2197      break;      break;
2198    
# Line 1897  for (;;) Line 2200  for (;;)
2200      any byte, even newline, independent of the setting of PCRE_DOTALL. */      any byte, even newline, independent of the setting of PCRE_DOTALL. */
2201    
2202      case OP_ANYBYTE:      case OP_ANYBYTE:
2203      if (eptr++ >= md->end_subject)      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2204        {        {                            /* not be updated before SCHECK_PARTIAL. */
2205        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2206        MRRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2207        }        }
2208        eptr++;
2209      ecode++;      ecode++;
2210      break;      break;
2211    
# Line 1913  for (;;) Line 2217  for (;;)
2217        }        }
2218      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2219      if (      if (
2220  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2221         c < 256 &&         c < 256 &&
2222  #endif  #endif
2223         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
# Line 1930  for (;;) Line 2234  for (;;)
2234        }        }
2235      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2236      if (      if (
2237  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2238         c >= 256 ||         c > 255 ||
2239  #endif  #endif
2240         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2241         )         )
# Line 1947  for (;;) Line 2251  for (;;)
2251        }        }
2252      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2253      if (      if (
2254  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2255         c < 256 &&         c < 256 &&
2256  #endif  #endif
2257         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
# Line 1964  for (;;) Line 2268  for (;;)
2268        }        }
2269      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2270      if (      if (
2271  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2272         c >= 256 ||         c > 255 ||
2273  #endif  #endif
2274         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2275         )         )
# Line 1981  for (;;) Line 2285  for (;;)
2285        }        }
2286      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2287      if (      if (
2288  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2289         c < 256 &&         c < 256 &&
2290  #endif  #endif
2291         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
# Line 1998  for (;;) Line 2302  for (;;)
2302        }        }
2303      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2304      if (      if (
2305  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2306         c >= 256 ||         c > 255 ||
2307  #endif  #endif
2308         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2309         )         )
# Line 2017  for (;;) Line 2321  for (;;)
2321      switch(c)      switch(c)
2322        {        {
2323        default: MRRETURN(MATCH_NOMATCH);        default: MRRETURN(MATCH_NOMATCH);
2324    
2325        case 0x000d:        case 0x000d:
2326        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
2327        break;        break;
# Line 2176  for (;;) Line 2481  for (;;)
2481          break;          break;
2482    
2483          case PT_GC:          case PT_GC:
2484          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2485            MRRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
2486          break;          break;
2487    
# Line 2193  for (;;) Line 2498  for (;;)
2498          /* These are specials */          /* These are specials */
2499    
2500          case PT_ALNUM:          case PT_ALNUM:
2501          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2502               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2503            MRRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
2504          break;          break;
2505    
2506          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2507          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2508               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2509                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2510            MRRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
2511          break;          break;
2512    
2513          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2514          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2515               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2516               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2517                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
# Line 2214  for (;;) Line 2519  for (;;)
2519          break;          break;
2520    
2521          case PT_WORD:          case PT_WORD:
2522          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2523               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2524               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2525            MRRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
2526          break;          break;
# Line 2240  for (;;) Line 2545  for (;;)
2545        MRRETURN(MATCH_NOMATCH);        MRRETURN(MATCH_NOMATCH);
2546        }        }
2547      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2548        if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH);
2549        while (eptr < md->end_subject)
2550        {        {
2551        int category = UCD_CATEGORY(c);        int len = 1;
2552        if (category == ucp_M) MRRETURN(MATCH_NOMATCH);        if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2553        while (eptr < md->end_subject)        if (UCD_CATEGORY(c) != ucp_M) break;
2554          {        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;  
         }  
2555        }        }
2556      ecode++;      ecode++;
2557      break;      break;
# Line 2269  for (;;) Line 2567  for (;;)
2567      loops). */      loops). */
2568    
2569      case OP_REF:      case OP_REF:
2570        case OP_REFI:
2571        caseless = op == OP_REFI;
2572      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2573      ecode += 3;      ecode += 1 + IMM2_SIZE;
2574    
2575      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2576    
# Line 2310  for (;;) Line 2610  for (;;)
2610        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2611        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2612        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2613        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2614        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2615        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2616        break;        break;
2617    
2618        default:               /* No repeat follows */        default:               /* No repeat follows */
2619        if ((length = match_ref(offset, eptr, length, md, ims)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2620          {          {
2621          CHECK_PARTIAL();          CHECK_PARTIAL();
2622          MRRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
# Line 2336  for (;;) Line 2636  for (;;)
2636    
2637      for (i = 1; i <= min; i++)      for (i = 1; i <= min; i++)
2638        {        {
2639        int slength;        int slength;
2640        if ((slength = match_ref(offset, eptr, length, md, ims)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2641          {          {
2642          CHECK_PARTIAL();          CHECK_PARTIAL();
2643          MRRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
# Line 2356  for (;;) Line 2656  for (;;)
2656        {        {
2657        for (fi = min;; fi++)        for (fi = min;; fi++)
2658          {          {
2659          int slength;          int slength;
2660          RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM14);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);
2661          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2662          if (fi >= max) MRRETURN(MATCH_NOMATCH);          if (fi >= max) MRRETURN(MATCH_NOMATCH);
2663          if ((slength = match_ref(offset, eptr, length, md, ims)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2664            {            {
2665            CHECK_PARTIAL();            CHECK_PARTIAL();
2666            MRRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
# Line 2377  for (;;) Line 2677  for (;;)
2677        pp = eptr;        pp = eptr;
2678        for (i = min; i < max; i++)        for (i = min; i < max; i++)
2679          {          {
2680          int slength;          int slength;
2681          if ((slength = match_ref(offset, eptr, length, md, ims)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2682            {            {
2683            CHECK_PARTIAL();            CHECK_PARTIAL();
2684            break;            break;
# Line 2387  for (;;) Line 2687  for (;;)
2687          }          }
2688        while (eptr >= pp)        while (eptr >= pp)
2689          {          {
2690          RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
2691          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2692          eptr -= length;          eptr -= length;
2693          }          }
# Line 2409  for (;;) Line 2709  for (;;)
2709      case OP_NCLASS:      case OP_NCLASS:
2710      case OP_CLASS:      case OP_CLASS:
2711        {        {
2712          /* The data variable is saved across frames, so the byte map needs to
2713          be stored there. */
2714    #define BYTE_MAP ((pcre_uint8 *)data)
2715        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2716        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2717    
2718        switch (*ecode)        switch (*ecode)
2719          {          {
# Line 2431  for (;;) Line 2734  for (;;)
2734          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2735          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2736          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2737          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2738          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2739          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2740          break;          break;
2741    
2742          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2443  for (;;) Line 2746  for (;;)
2746    
2747        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2748    
2749  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2750        /* UTF-8 mode */        if (utf)
       if (utf8)  
2751          {          {
2752          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2753            {            {
# Line 2460  for (;;) Line 2762  for (;;)
2762              if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);
2763              }              }
2764            else            else
2765              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
             }  
2766            }            }
2767          }          }
2768        else        else
2769  #endif  #endif
2770        /* Not UTF-8 mode */        /* Not UTF mode */
2771          {          {
2772          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2773            {            {
# Line 2477  for (;;) Line 2777  for (;;)
2777              MRRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
2778              }              }
2779            c = *eptr++;            c = *eptr++;
2780            if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2781              if (c > 255)
2782                {
2783                if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);
2784                }
2785              else
2786    #endif
2787                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);
2788            }            }
2789          }          }
2790    
# Line 2491  for (;;) Line 2798  for (;;)
2798    
2799        if (minimize)        if (minimize)
2800          {          {
2801  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2802          /* UTF-8 mode */          if (utf)
         if (utf8)  
2803            {            {
2804            for (fi = min;; fi++)            for (fi = min;; fi++)
2805              {              {
2806              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM16);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);
2807              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2808              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
2809              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 2511  for (;;) Line 2817  for (;;)
2817                if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);
2818                }                }
2819              else              else
2820                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
               }  
2821              }              }
2822            }            }
2823          else          else
2824  #endif  #endif
2825          /* Not UTF-8 mode */          /* Not UTF mode */
2826            {            {
2827            for (fi = min;; fi++)            for (fi = min;; fi++)
2828              {              {
2829              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM17);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);
2830              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2831              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
2832              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 2531  for (;;) Line 2835  for (;;)
2835                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
2836                }                }
2837              c = *eptr++;              c = *eptr++;
2838              if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2839                if (c > 255)
2840                  {
2841                  if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);
2842                  }
2843                else
2844    #endif
2845                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);
2846              }              }
2847            }            }
2848          /* Control never gets here */          /* Control never gets here */
# Line 2543  for (;;) Line 2854  for (;;)
2854          {          {
2855          pp = eptr;          pp = eptr;
2856    
2857  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2858          /* UTF-8 mode */          if (utf)
         if (utf8)  
2859            {            {
2860            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2861              {              {
# Line 2561  for (;;) Line 2871  for (;;)
2871                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
2872                }                }
2873              else              else
2874                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
2875              eptr += len;              eptr += len;
2876              }              }
2877            for (;;)            for (;;)
2878              {              {
2879              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM18);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM18);
2880              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2881              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2882              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2576  for (;;) Line 2884  for (;;)
2884            }            }
2885          else          else
2886  #endif  #endif
2887            /* Not UTF-8 mode */            /* Not UTF mode */
2888            {            {
2889            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2890              {              {
# Line 2586  for (;;) Line 2894  for (;;)
2894                break;                break;
2895                }                }
2896              c = *eptr;              c = *eptr;
2897              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
2898                if (c > 255)
2899                  {
2900                  if (op == OP_CLASS) break;
2901                  }
2902                else
2903    #endif
2904                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
2905              eptr++;              eptr++;
2906              }              }
2907            while (eptr >= pp)            while (eptr >= pp)
2908              {              {
2909              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM19);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM19);
2910              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2911              eptr--;              eptr--;
2912              }              }
# Line 2599  for (;;) Line 2914  for (;;)
2914    
2915          MRRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
2916          }          }
2917    #undef BYTE_MAP
2918        }        }
2919      /* Control never gets here */      /* Control never gets here */
2920    
# Line 2607  for (;;) Line 2923  for (;;)
2923      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
2924      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
2925    
2926  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
2927      case OP_XCLASS:      case OP_XCLASS:
2928        {        {
2929        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2632  for (;;) Line 2948  for (;;)
2948          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2949          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2950          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2951          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2952          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2953          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2954          break;          break;
2955    
2956          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2652  for (;;) Line 2968  for (;;)
2968            MRRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
2969            }            }
2970          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
2971          if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) MRRETURN(MATCH_NOMATCH);
2972          }          }
2973    
2974        /* 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 2983  for (;;)
2983          {          {
2984          for (fi = min;; fi++)          for (fi = min;; fi++)
2985            {            {
2986            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);
2987            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2988            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
2989            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 2676  for (;;) Line 2992  for (;;)
2992              MRRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
2993              }              }
2994            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
2995            if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) MRRETURN(MATCH_NOMATCH);
2996            }            }
2997          /* Control never gets here */          /* Control never gets here */
2998          }          }
# Line 2694  for (;;) Line 3010  for (;;)
3010              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3011              break;              break;
3012              }              }
3013    #ifdef SUPPORT_UTF
3014            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3015            if (!_pcre_xclass(c, data)) break;  #else
3016              c = *eptr;
3017    #endif
3018              if (!PRIV(xclass)(c, data, utf)) break;
3019            eptr += len;            eptr += len;
3020            }            }
3021          for(;;)          for(;;)
3022            {            {
3023            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3024            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3025            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3026            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3027              if (utf) BACKCHAR(eptr);
3028    #endif
3029            }            }
3030          MRRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
3031          }          }
# Line 2715  for (;;) Line 3037  for (;;)
3037      /* Match a single character, casefully */      /* Match a single character, casefully */
3038    
3039      case OP_CHAR:      case OP_CHAR:
3040  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3041      if (utf8)      if (utf)
3042        {        {
3043        length = 1;        length = 1;
3044        ecode++;        ecode++;
# Line 2730  for (;;) Line 3052  for (;;)
3052        }        }
3053      else      else
3054  #endif  #endif
3055        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3056        {        {
3057        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3058          {          {
# Line 2745  for (;;) Line 3066  for (;;)
3066    
3067      /* Match a single character, caselessly */      /* Match a single character, caselessly */
3068    
3069      case OP_CHARNC:      case OP_CHARI:
3070  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3071      if (utf8)      if (utf)
3072        {        {
3073        length = 1;        length = 1;
3074        ecode++;        ecode++;
# Line 2788  for (;;) Line 3109  for (;;)
3109          }          }
3110        }        }
3111      else      else
3112  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3113    
3114      /* Non-UTF-8 mode */      /* Not UTF mode */
3115        {        {
3116        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3117          {          {
3118          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
3119          MRRETURN(MATCH_NOMATCH);          MRRETURN(MATCH_NOMATCH);
3120          }          }
3121        if (md->lcc[ecode[1]] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3122              != TABLE_GET(*eptr, md->lcc, *eptr)) MRRETURN(MATCH_NOMATCH);
3123          eptr++;
3124        ecode += 2;        ecode += 2;
3125        }        }
3126      break;      break;
# Line 2805  for (;;) Line 3128  for (;;)
3128      /* Match a single character repeatedly. */      /* Match a single character repeatedly. */
3129    
3130      case OP_EXACT:      case OP_EXACT:
3131        case OP_EXACTI:
3132      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3133      ecode += 3;      ecode += 1 + IMM2_SIZE;
3134      goto REPEATCHAR;      goto REPEATCHAR;
3135    
3136      case OP_POSUPTO:      case OP_POSUPTO:
3137        case OP_POSUPTOI:
3138      possessive = TRUE;      possessive = TRUE;
3139      /* Fall through */      /* Fall through */
3140    
3141      case OP_UPTO:      case OP_UPTO:
3142        case OP_UPTOI:
3143      case OP_MINUPTO:      case OP_MINUPTO:
3144        case OP_MINUPTOI:
3145      min = 0;      min = 0;
3146      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3147      minimize = *ecode == OP_MINUPTO;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3148      ecode += 3;      ecode += 1 + IMM2_SIZE;
3149      goto REPEATCHAR;      goto REPEATCHAR;
3150    
3151      case OP_POSSTAR:      case OP_POSSTAR:
3152        case OP_POSSTARI:
3153      possessive = TRUE;      possessive = TRUE;
3154      min = 0;      min = 0;
3155      max = INT_MAX;      max = INT_MAX;
# Line 2829  for (;;) Line 3157  for (;;)
3157      goto REPEATCHAR;      goto REPEATCHAR;
3158    
3159      case OP_POSPLUS:      case OP_POSPLUS:
3160        case OP_POSPLUSI:
3161      possessive = TRUE;      possessive = TRUE;
3162      min = 1;      min = 1;
3163      max = INT_MAX;      max = INT_MAX;
# Line 2836  for (;;) Line 3165  for (;;)
3165      goto REPEATCHAR;      goto REPEATCHAR;
3166    
3167      case OP_POSQUERY:      case OP_POSQUERY:
3168        case OP_POSQUERYI:
3169      possessive = TRUE;      possessive = TRUE;
3170      min = 0;      min = 0;
3171      max = 1;      max = 1;
# Line 2843  for (;;) Line 3173  for (;;)
3173      goto REPEATCHAR;      goto REPEATCHAR;
3174    
3175      case OP_STAR:      case OP_STAR:
3176        case OP_STARI:
3177      case OP_MINSTAR:      case OP_MINSTAR:
3178        case OP_MINSTARI:
3179      case OP_PLUS:      case OP_PLUS:
3180        case OP_PLUSI:
3181      case OP_MINPLUS:      case OP_MINPLUS:
3182        case OP_MINPLUSI:
3183      case OP_QUERY:      case OP_QUERY:
3184        case OP_QUERYI:
3185      case OP_MINQUERY:      case OP_MINQUERY:
3186      c = *ecode++ - OP_STAR;      case OP_MINQUERYI:
3187        c = *ecode++ - ((op < OP_STARI)? OP_STAR : OP_STARI);
3188      minimize = (c & 1) != 0;      minimize = (c & 1) != 0;
   
3189      min = rep_min[c];                 /* Pick up values from tables; */      min = rep_min[c];                 /* Pick up values from tables; */
3190      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
3191      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
# Line 2858  for (;;) Line 3193  for (;;)
3193      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3194    
3195      REPEATCHAR:      REPEATCHAR:
3196  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3197      if (utf8)      if (utf)
3198        {        {
3199        length = 1;        length = 1;
3200        charptr = ecode;        charptr = ecode;
# Line 2873  for (;;) Line 3208  for (;;)
3208          {          {
3209  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3210          unsigned int othercase;          unsigned int othercase;
3211          if ((ims & PCRE_CASELESS) != 0 &&          if (op >= OP_STARI &&     /* Caseless */
3212              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3213            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3214          else oclength = 0;          else oclength = 0;
3215  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3216    
3217          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3218            {            {
3219            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3220              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3221  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3222            else if (oclength > 0 &&            else if (oclength > 0 &&
3223                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3224                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3225  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3226            else            else
3227              {              {
# Line 2901  for (;;) Line 3236  for (;;)
3236            {            {
3237            for (fi = min;; fi++)            for (fi = min;; fi++)
3238              {              {
3239              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM22);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);
3240              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3241              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3242              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3243                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3244  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3245              else if (oclength > 0 &&              else if (oclength > 0 &&
3246                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3247                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3248  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3249              else              else
3250                {                {
# Line 2926  for (;;) Line 3261  for (;;)
3261            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3262              {              {
3263              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3264                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3265  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3266              else if (oclength > 0 &&              else if (oclength > 0 &&
3267                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3268                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3269  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3270              else              else
3271                {                {
# Line 2943  for (;;) Line 3278  for (;;)
3278    
3279            for(;;)            for(;;)
3280              {              {
3281              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3282              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3283              if (eptr == pp) { MRRETURN(MATCH_NOMATCH); }              if (eptr == pp) { MRRETURN(MATCH_NOMATCH); }
3284  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2962  for (;;) Line 3297  for (;;)
3297        value of fc will always be < 128. */        value of fc will always be < 128. */
3298        }        }
3299      else      else
3300  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3301          /* When not in UTF-8 mode, load a single-byte character. */
3302      /* When not in UTF-8 mode, load a single-byte character. */        fc = *ecode++;
3303    
3304      fc = *ecode++;      /* The value of fc at this point is always one character, though we may
3305        or may not be in UTF mode. The code is duplicated for the caseless and
     /* The value of fc at this point is always less than 256, though we may or  
     may not be in UTF-8 mode. The code is duplicated for the caseless and  
3306      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3307      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3308      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 2980  for (;;) Line 3313  for (;;)
3313      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3314        max, eptr));        max, eptr));
3315    
3316      if ((ims & PCRE_CASELESS) != 0)      if (op >= OP_STARI)  /* Caseless */
3317        {        {
3318        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3319          /* fc must be < 128 */
3320          foc = md->fcc[fc];
3321    #else
3322    #ifdef SUPPORT_UTF
3323    #ifdef SUPPORT_UCP
3324          if (utf && fc > 127)
3325            foc = UCD_OTHERCASE(fc);
3326    #else
3327          if (utf && fc > 127)
3328            foc = fc;
3329    #endif /* SUPPORT_UCP */
3330          else
3331    #endif /* SUPPORT_UTF */
3332            foc = TABLE_GET(fc, md->fcc, fc);
3333    #endif /* COMPILE_PCRE8 */
3334    
3335        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3336          {          {
3337          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
# Line 2990  for (;;) Line 3339  for (;;)
3339            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3340            MRRETURN(MATCH_NOMATCH);            MRRETURN(MATCH_NOMATCH);
3341            }            }
3342          if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);          if (fc != *eptr && foc != *eptr) MRRETURN(MATCH_NOMATCH);
3343            eptr++;
3344          }          }
3345        if (min == max) continue;        if (min == max) continue;
3346        if (minimize)        if (minimize)
3347          {          {
3348          for (fi = min;; fi++)          for (fi = min;; fi++)
3349            {            {
3350            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3351            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3352            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
3353            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3005  for (;;) Line 3355  for (;;)
3355              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3356              MRRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
3357              }              }
3358            if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);            if (fc != *eptr && foc != *eptr) MRRETURN(MATCH_NOMATCH);
3359              eptr++;
3360            }            }
3361          /* Control never gets here */          /* Control never gets here */
3362          }          }
# Line 3019  for (;;) Line 3370  for (;;)
3370              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3371              break;              break;
3372              }              }
3373            if (fc != md->lcc[*eptr]) break;            if (fc != *eptr && foc != *eptr) break;
3374            eptr++;            eptr++;
3375            }            }
3376    
# Line 3027  for (;;) Line 3378  for (;;)
3378    
3379          while (eptr >= pp)          while (eptr >= pp)
3380            {            {
3381            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM25);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);
3382            eptr--;            eptr--;
3383            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3384            }            }
# Line 3056  for (;;) Line 3407  for (;;)
3407          {          {
3408          for (fi = min;; fi++)          for (fi = min;; fi++)
3409            {            {
3410            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM26);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);
3411            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3412            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
3413            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3085  for (;;) Line 3436  for (;;)
3436    
3437          while (eptr >= pp)          while (eptr >= pp)
3438            {            {
3439            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);
3440            eptr--;            eptr--;
3441            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3442            }            }
# Line 3098  for (;;) Line 3449  for (;;)
3449      checking can be multibyte. */      checking can be multibyte. */
3450    
3451      case OP_NOT:      case OP_NOT:
3452        case OP_NOTI:
3453      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
3454        {        {
3455        SCHECK_PARTIAL();        SCHECK_PARTIAL();
# Line 3105  for (;;) Line 3457  for (;;)
3457        }        }
3458      ecode++;      ecode++;
3459      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
3460      if ((ims & PCRE_CASELESS) != 0)      if (op == OP_NOTI)         /* The caseless case */
3461        {        {
3462  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
3463        if (c < 256)        if (c < 256)
3464  #endif  #endif
3465        c = md->lcc[c];          c = md->lcc[c];
3466        if (md->lcc[*ecode++] == c) MRRETURN(MATCH_NOMATCH);        if (md->lcc[*ecode++] == c) MRRETURN(MATCH_NOMATCH);
3467        }        }
3468      else      else    /* Caseful */
3469        {        {
3470        if (*ecode++ == c) MRRETURN(MATCH_NOMATCH);        if (*ecode++ == c) MRRETURN(MATCH_NOMATCH);
3471        }        }
# Line 3127  for (;;) Line 3479  for (;;)
3479      about... */      about... */
3480    
3481      case OP_NOTEXACT:      case OP_NOTEXACT:
3482        case OP_NOTEXACTI:
3483      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3484      ecode += 3;      ecode += 1 + IMM2_SIZE;
3485      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3486    
3487      case OP_NOTUPTO:      case OP_NOTUPTO:
3488        case OP_NOTUPTOI:
3489      case OP_NOTMINUPTO:      case OP_NOTMINUPTO:
3490        case OP_NOTMINUPTOI:
3491      min = 0;      min = 0;
3492      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3493      minimize = *ecode == OP_NOTMINUPTO;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3494      ecode += 3;      ecode += 1 + IMM2_SIZE;
3495      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3496    
3497      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
3498        case OP_NOTPOSSTARI:
3499      possessive = TRUE;      possessive = TRUE;
3500      min = 0;      min = 0;
3501      max = INT_MAX;      max = INT_MAX;
# Line 3147  for (;;) Line 3503  for (;;)
3503      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3504    
3505      case OP_NOTPOSPLUS:      case OP_NOTPOSPLUS:
3506        case OP_NOTPOSPLUSI:
3507      possessive = TRUE;      possessive = TRUE;
3508      min = 1;      min = 1;
3509      max = INT_MAX;      max = INT_MAX;
# Line 3154  for (;;) Line 3511  for (;;)
3511      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3512    
3513      case OP_NOTPOSQUERY:      case OP_NOTPOSQUERY:
3514        case OP_NOTPOSQUERYI:
3515      possessive = TRUE;      possessive = TRUE;
3516      min = 0;      min = 0;
3517      max = 1;      max = 1;
# Line 3161  for (;;) Line 3519  for (;;)
3519      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3520    
3521      case OP_NOTPOSUPTO:      case OP_NOTPOSUPTO:
3522        case OP_NOTPOSUPTOI:
3523      possessive = TRUE;      possessive = TRUE;
3524      min = 0;      min = 0;
3525      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3526      ecode += 3;      ecode += 1 + IMM2_SIZE;
3527      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3528    
3529      case OP_NOTSTAR:      case OP_NOTSTAR:
3530        case OP_NOTSTARI:
3531      case OP_NOTMINSTAR:      case OP_NOTMINSTAR:
3532        case OP_NOTMINSTARI:
3533      case OP_NOTPLUS:      case OP_NOTPLUS:
3534        case OP_NOTPLUSI:
3535      case OP_NOTMINPLUS:      case OP_NOTMINPLUS:
3536        case OP_NOTMINPLUSI:
3537      case OP_NOTQUERY:      case OP_NOTQUERY:
3538        case OP_NOTQUERYI:
3539      case OP_NOTMINQUERY:      case OP_NOTMINQUERY:
3540      c = *ecode++ - OP_NOTSTAR;      case OP_NOTMINQUERYI:
3541        c = *ecode++ - ((op >= OP_NOTSTARI)? OP_NOTSTARI: OP_NOTSTAR);
3542      minimize = (c & 1) != 0;      minimize = (c & 1) != 0;
3543      min = rep_min[c];                 /* Pick up values from tables; */      min = rep_min[c];                 /* Pick up values from tables; */
3544      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
# Line 3195  for (;;) Line 3560  for (;;)
3560      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,
3561        max, eptr));        max, eptr));
3562    
3563      if ((ims & PCRE_CASELESS) != 0)      if (op >= OP_NOTSTARI)     /* Caseless */
3564        {        {
3565        fc = md->lcc[fc];        fc = TABLE_GET(fc, md->lcc, fc);
3566    
3567  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3568        /* UTF-8 mode */        if (utf)
       if (utf8)  
3569          {          {
3570          register unsigned int d;          register unsigned int d;
3571          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3218  for (;;) Line 3582  for (;;)
3582          }          }
3583        else        else
3584  #endif  #endif
3585          /* Not UTF mode */
       /* Not UTF-8 mode */  
3586          {          {
3587          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3588            {            {
# Line 3236  for (;;) Line 3599  for (;;)
3599    
3600        if (minimize)        if (minimize)
3601          {          {
3602  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3603          /* UTF-8 mode */          if (utf)
         if (utf8)  
3604            {            {
3605            register unsigned int d;            register unsigned int d;
3606            for (fi = min;; fi++)            for (fi = min;; fi++)
3607              {              {
3608              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
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 3258  for (;;) Line 3620  for (;;)
3620            }            }
3621          else          else
3622  #endif  #endif
3623          /* Not UTF-8 mode */          /* Not UTF mode */
3624            {            {
3625            for (fi = min;; fi++)            for (fi = min;; fi++)
3626              {              {
3627              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM29);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);
3628              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3629              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3630              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 3282  for (;;) Line 3644  for (;;)
3644          {          {
3645          pp = eptr;          pp = eptr;
3646    
3647  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3648          /* UTF-8 mode */          if (utf)
         if (utf8)  
3649            {            {
3650            register unsigned int d;            register unsigned int d;
3651            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3303  for (;;) Line 3664  for (;;)
3664          if (possessive) continue;          if (possessive) continue;
3665          for(;;)          for(;;)
3666              {              {
3667              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3668              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3669              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
3670              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3311  for (;;) Line 3672  for (;;)
3672            }            }
3673          else          else
3674  #endif  #endif
3675          /* Not UTF-8 mode */          /* Not UTF mode */
3676            {            {
3677            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3678              {              {
# Line 3326  for (;;) Line 3687  for (;;)
3687            if (possessive) continue;            if (possessive) continue;
3688            while (eptr >= pp)            while (eptr >= pp)
3689              {              {
3690              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM31);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);
3691              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3692              eptr--;              eptr--;
3693              }              }
# Line 3341  for (;;) Line 3702  for (;;)
3702    
3703      else      else
3704        {        {
3705  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3706        /* UTF-8 mode */        if (utf)
       if (utf8)  
3707          {          {
3708          register unsigned int d;          register unsigned int d;
3709          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3359  for (;;) Line 3719  for (;;)
3719          }          }
3720        else        else
3721  #endif  #endif
3722        /* Not UTF-8 mode */        /* Not UTF mode */
3723          {          {
3724          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3725            {            {
# Line 3376  for (;;) Line 3736  for (;;)
3736    
3737        if (minimize)        if (minimize)
3738          {          {
3739  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3740          /* UTF-8 mode */          if (utf)
         if (utf8)  
3741            {            {
3742            register unsigned int d;            register unsigned int d;
3743            for (fi = min;; fi++)            for (fi = min;; fi++)
3744              {              {
3745              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
3746              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3747              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3748              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 3397  for (;;) Line 3756  for (;;)
3756            }            }
3757          else          else
3758  #endif  #endif
3759          /* Not UTF-8 mode */          /* Not UTF mode */
3760            {            {
3761            for (fi = min;; fi++)            for (fi = min;; fi++)
3762              {              {
3763              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM33);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);
3764              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3765              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
3766              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 3421  for (;;) Line 3780  for (;;)
3780          {          {
3781          pp = eptr;          pp = eptr;
3782    
3783  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3784          /* UTF-8 mode */          if (utf)
         if (utf8)  
3785            {            {
3786            register unsigned int d;            register unsigned int d;
3787            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3441  for (;;) Line 3799  for (;;)
3799            if (possessive) continue;            if (possessive) continue;
3800            for(;;)            for(;;)
3801              {              {
3802              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM34);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);
3803              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3804              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
3805              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3449  for (;;) Line 3807  for (;;)
3807            }            }
3808          else          else
3809  #endif  #endif
3810          /* Not UTF-8 mode */          /* Not UTF mode */
3811            {            {
3812            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3813              {              {
# Line 3464  for (;;) Line 3822  for (;;)
3822            if (possessive) continue;            if (possessive) continue;
3823            while (eptr >= pp)            while (eptr >= pp)
3824              {              {
3825              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM35);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);
3826              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3827              eptr--;              eptr--;
3828              }              }
# Line 3482  for (;;) Line 3840  for (;;)
3840      case OP_TYPEEXACT:      case OP_TYPEEXACT:
3841      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3842      minimize = TRUE;      minimize = TRUE;
3843      ecode += 3;      ecode += 1 + IMM2_SIZE;
3844      goto REPEATTYPE;      goto REPEATTYPE;
3845    
3846      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3490  for (;;) Line 3848  for (;;)
3848      min = 0;      min = 0;
3849      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3850      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
3851      ecode += 3;      ecode += 1 + IMM2_SIZE;
3852      goto REPEATTYPE;      goto REPEATTYPE;
3853    
3854      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3518  for (;;) Line 3876  for (;;)
3876      possessive = TRUE;      possessive = TRUE;
3877      min = 0;      min = 0;
3878      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3879      ecode += 3;      ecode += 1 + IMM2_SIZE;
3880      goto REPEATTYPE;      goto REPEATTYPE;
3881    
3882      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 3579  for (;;) Line 3937  for (;;)
3937            case PT_LAMP:            case PT_LAMP:
3938            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3939              {              {
3940                int chartype;
3941              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3942                {                {
3943                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3944                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3945                }                }
3946              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3947              prop_chartype = UCD_CHARTYPE(c);              chartype = UCD_CHARTYPE(c);
3948              if ((prop_chartype == ucp_Lu ||              if ((chartype == ucp_Lu ||
3949                   prop_chartype == ucp_Ll ||                   chartype == ucp_Ll ||
3950                   prop_chartype == ucp_Lt) == prop_fail_result)                   chartype == ucp_Lt) == prop_fail_result)
3951                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3952              }              }
3953            break;            break;
# Line 3602  for (;;) Line 3961  for (;;)
3961                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3962                }                }
3963              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3964              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)
             if ((prop_category == prop_value) == prop_fail_result)  
3965                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3966              }              }
3967            break;            break;
# Line 3617  for (;;) Line 3975  for (;;)
3975                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3976                }                }
3977              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3978              prop_chartype = UCD_CHARTYPE(c);              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)
             if ((prop_chartype == prop_value) == prop_fail_result)  
3979                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3980              }              }
3981            break;            break;
# Line 3632  for (;;) Line 3989  for (;;)
3989                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3990                }                }
3991              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3992              prop_script = UCD_SCRIPT(c);              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)
             if ((prop_script == prop_value) == prop_fail_result)  
3993                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
3994              }              }
3995            break;            break;
# Line 3641  for (;;) Line 3997  for (;;)
3997            case PT_ALNUM:            case PT_ALNUM:
3998            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3999              {              {
4000                int category;
4001              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4002                {                {
4003                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4004                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4005                }                }
4006              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4007              prop_category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4008              if ((prop_category == ucp_L || prop_category == ucp_N)              if ((category == ucp_L || category == ucp_N) == prop_fail_result)
                    == prop_fail_result)  
4009                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4010              }              }
4011            break;            break;
# Line 3663  for (;;) Line 4019  for (;;)
4019                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4020                }                }
4021              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4022              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 ||  
4023                   c == CHAR_FF || c == CHAR_CR)                   c == CHAR_FF || c == CHAR_CR)
4024                     == prop_fail_result)                     == prop_fail_result)
4025                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
# Line 3680  for (;;) Line 4035  for (;;)
4035                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4036                }                }
4037              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4038              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 ||  
4039                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
4040                     == prop_fail_result)                     == prop_fail_result)
4041                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
# Line 3691  for (;;) Line 4045  for (;;)
4045            case PT_WORD:            case PT_WORD:
4046            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4047              {              {
4048                int category;
4049              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4050                {                {
4051                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4052                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4053                }                }
4054              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4055              prop_category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4056              if ((prop_category == ucp_L || prop_category == ucp_N ||              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)
                  c == CHAR_UNDERSCORE)  
4057                     == prop_fail_result)                     == prop_fail_result)
4058                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4059              }              }
# Line 3725  for (;;) Line 4079  for (;;)
4079              MRRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
4080              }              }
4081            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
4082            prop_category = UCD_CATEGORY(c);            if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH);
           if (prop_category == ucp_M) MRRETURN(MATCH_NOMATCH);  
4083            while (eptr < md->end_subject)            while (eptr < md->end_subject)
4084              {              {
4085              int len = 1;              int len = 1;
4086              if (!utf8) c = *eptr;              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4087                else { GETCHARLEN(c, eptr, len); }              if (UCD_CATEGORY(c) != ucp_M) break;
             prop_category = UCD_CATEGORY(c);  
             if (prop_category != ucp_M) break;  
4088              eptr += len;              eptr += len;
4089              }              }
4090            }            }
# Line 3744  for (;;) Line 4095  for (;;)
4095    
4096  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4097    
4098  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4099        if (utf8) switch(ctype)        if (utf) switch(ctype)
4100          {          {
4101          case OP_ANY:          case OP_ANY:
4102          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3757  for (;;) Line 4108  for (;;)
4108              }              }
4109            if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);
4110            eptr++;            eptr++;
4111            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4112            }            }
4113          break;          break;
4114    
# Line 3770  for (;;) Line 4121  for (;;)
4121              MRRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
4122              }              }
4123            eptr++;            eptr++;
4124            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4125            }            }
4126          break;          break;
4127    
# Line 3791  for (;;) Line 4142  for (;;)
4142            switch(c)            switch(c)
4143              {              {
4144              default: MRRETURN(MATCH_NOMATCH);              default: MRRETURN(MATCH_NOMATCH);
4145    
4146              case 0x000d:              case 0x000d:
4147              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
4148              break;              break;
# Line 3967  for (;;) Line 4319  for (;;)
4319              }              }
4320            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)
4321              MRRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
4322            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4323              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4324            }            }
4325          break;          break;
4326    
# Line 3995  for (;;) Line 4348  for (;;)
4348              }              }
4349            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)
4350              MRRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
4351            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4352              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4353            }            }
4354          break;          break;
4355    
# Line 4018  for (;;) Line 4372  for (;;)
4372          }  /* End switch(ctype) */          }  /* End switch(ctype) */
4373    
4374        else        else
4375  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF */
4376    
4377        /* Code for the non-UTF-8 case for minimum matching of operators other        /* Code for the non-UTF-8 case for minimum matching of operators other
4378        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. */
# Line 4067  for (;;) Line 4421  for (;;)
4421            switch(*eptr++)            switch(*eptr++)
4422              {              {
4423              default: MRRETURN(MATCH_NOMATCH);              default: MRRETURN(MATCH_NOMATCH);
4424    
4425              case 0x000d:              case 0x000d:
4426              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
4427              break;              break;
4428    
4429              case 0x000a:              case 0x000a:
4430              break;              break;
4431    
# Line 4259  for (;;) Line 4615  for (;;)
4615            case PT_ANY:            case PT_ANY:
4616            for (fi = min;; fi++)            for (fi = min;; fi++)
4617              {              {
4618              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM36);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM36);
4619              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4620              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4621              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4275  for (;;) Line 4631  for (;;)
4631            case PT_LAMP:            case PT_LAMP:
4632            for (fi = min;; fi++)            for (fi = min;; fi++)
4633              {              {
4634              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM37);              int chartype;
4635                RMATCH(eptr, ecode, offset_top, md, eptrb, RM37);
4636              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4637              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4638              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4284  for (;;) Line 4641  for (;;)
4641                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4642                }                }
4643              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4644              prop_chartype = UCD_CHARTYPE(c);              chartype = UCD_CHARTYPE(c);
4645              if ((prop_chartype == ucp_Lu ||              if ((chartype == ucp_Lu ||
4646                   prop_chartype == ucp_Ll ||                   chartype == ucp_Ll ||
4647                   prop_chartype == ucp_Lt) == prop_fail_result)                   chartype == ucp_Lt) == prop_fail_result)
4648                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4649              }              }
4650            /* Control never gets here */            /* Control never gets here */
# Line 4295  for (;;) Line 4652  for (;;)
4652            case PT_GC:            case PT_GC:
4653            for (fi = min;; fi++)            for (fi = min;; fi++)
4654              {              {
4655              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM38);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM38);
4656              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4657              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4658              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4304  for (;;) Line 4661  for (;;)
4661                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4662                }                }
4663              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4664              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)
             if ((prop_category == prop_value) == prop_fail_result)  
4665                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4666              }              }
4667            /* Control never gets here */            /* Control never gets here */
# Line 4313  for (;;) Line 4669  for (;;)
4669            case PT_PC:            case PT_PC:
4670            for (fi = min;; fi++)            for (fi = min;; fi++)
4671              {              {
4672              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM39);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM39);
4673              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4674              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4675              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4322  for (;;) Line 4678  for (;;)
4678                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4679                }                }
4680              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4681              prop_chartype = UCD_CHARTYPE(c);              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)
             if ((prop_chartype == prop_value) == prop_fail_result)  
4682                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4683              }              }
4684            /* Control never gets here */            /* Control never gets here */
# Line 4331  for (;;) Line 4686  for (;;)
4686            case PT_SC:            case PT_SC:
4687            for (fi = min;; fi++)            for (fi = min;; fi++)
4688              {              {
4689              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM40);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM40);
4690              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4691              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4692              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4340  for (;;) Line 4695  for (;;)
4695                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4696                }                }
4697              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4698              prop_script = UCD_SCRIPT(c);              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)
             if ((prop_script == prop_value) == prop_fail_result)  
4699                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4700              }              }
4701            /* Control never gets here */            /* Control never gets here */
# Line 4349  for (;;) Line 4703  for (;;)
4703            case PT_ALNUM:            case PT_ALNUM:
4704            for (fi = min;; fi++)            for (fi = min;; fi++)
4705              {              {
4706              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM59);              int category;
4707                RMATCH(eptr, ecode, offset_top, md, eptrb, RM59);
4708              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4709              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4710              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4358  for (;;) Line 4713  for (;;)
4713                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4714                }                }
4715              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4716              prop_category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4717              if ((prop_category == ucp_L || prop_category == ucp_N)              if ((category == ucp_L || category == ucp_N) == prop_fail_result)
                    == prop_fail_result)  
4718                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4719              }              }
4720            /* Control never gets here */            /* Control never gets here */
# Line 4368  for (;;) Line 4722  for (;;)
4722            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
4723            for (fi = min;; fi++)            for (fi = min;; fi++)
4724              {              {
4725              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM60);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM60);
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 4377  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_FF || c == CHAR_CR)                   c == CHAR_FF || c == CHAR_CR)
4736                     == prop_fail_result)                     == prop_fail_result)
4737                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
# Line 4388  for (;;) Line 4741  for (;;)
4741            case PT_PXSPACE:  /* POSIX space */            case PT_PXSPACE:  /* POSIX space */
4742            for (fi = min;; fi++)            for (fi = min;; fi++)
4743              {              {
4744              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM61);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM61);
4745              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4746              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4747              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4397  for (;;) Line 4750  for (;;)
4750                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4751                }                }
4752              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4753              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 ||  
4754                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
4755                     == prop_fail_result)                     == prop_fail_result)
4756                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
# Line 4408  for (;;) Line 4760  for (;;)
4760            case PT_WORD:            case PT_WORD:
4761            for (fi = min;; fi++)            for (fi = min;; fi++)
4762              {              {
4763              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM62);              int category;
4764                RMATCH(eptr, ecode, offset_top, md, eptrb, RM62);
4765              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4766              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) MRRETURN(MATCH_NOMATCH);
4767              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4417  for (;;) Line 4770  for (;;)
4770                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
4771                }                }
4772              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4773              prop_category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4774              if ((prop_category == ucp_L ||              if ((category == ucp_L ||
4775                   prop_category == ucp_N ||                   category == ucp_N ||
4776                   c == CHAR_UNDERSCORE)                   c == CHAR_UNDERSCORE)
4777                     == prop_fail_result)                     == prop_fail_result)
4778                MRRETURN(MATCH_NOMATCH);                MRRETURN(MATCH_NOMATCH);
# Line 4440  for (;;) Line 4793  for (;;)
4793          {          {
4794          for (fi = min;; fi++)          for (fi = min;; fi++)
4795            {            {
4796            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM41);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM41);
4797            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4798            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
4799            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 4449  for (;;) Line 4802  for (;;)
4802              MRRETURN(MATCH_NOMATCH);              MRRETURN(MATCH_NOMATCH);
4803              }              }
4804            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
4805            prop_category = UCD_CATEGORY(c);            if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH);
           if (prop_category == ucp_M) MRRETURN(MATCH_NOMATCH);  
4806            while (eptr < md->end_subject)            while (eptr < md->end_subject)
4807              {              {
4808              int len = 1;              int len = 1;
4809              if (!utf8) c = *eptr;              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4810                else { GETCHARLEN(c, eptr, len); }              if (UCD_CATEGORY(c) != ucp_M) break;
             prop_category = UCD_CATEGORY(c);  
             if (prop_category != ucp_M) break;  
4811              eptr += len;              eptr += len;
4812              }              }
4813            }            }
4814          }          }
   
4815        else        else
4816  #endif     /* SUPPORT_UCP */  #endif     /* SUPPORT_UCP */
4817    
4818  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4819        /* UTF-8 mode */        if (utf)
       if (utf8)  
4820          {          {
4821          for (fi = min;; fi++)          for (fi = min;; fi++)
4822            {            {
4823            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM42);
4824            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4825            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
4826            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 4631  for (;;) Line 4979  for (;;)
4979          }          }
4980        else        else
4981  #endif  #endif
4982        /* Not UTF-8 mode */        /* Not UTF mode */
4983          {          {
4984          for (fi = min;; fi++)          for (fi = min;; fi++)
4985            {            {
4986            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM43);
4987            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4988            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) MRRETURN(MATCH_NOMATCH);
4989            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 4783  for (;;) Line 5131  for (;;)
5131            case PT_LAMP:            case PT_LAMP:
5132            for (i = min; i < max; i++)            for (i = min; i < max; i++)
5133              {              {
5134                int chartype;
5135              int len = 1;              int len = 1;
5136              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
5137                {                {
# Line 4790  for (;;) Line 5139  for (;;)
5139                break;                break;
5140                }                }
5141              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5142              prop_chartype = UCD_CHARTYPE(c);              chartype = UCD_CHARTYPE(c);
5143              if ((prop_chartype == ucp_Lu ||              if ((chartype == ucp_Lu ||
5144                   prop_chartype == ucp_Ll ||                   chartype == ucp_Ll ||
5145                   prop_chartype == ucp_Lt) == prop_fail_result)                   chartype == ucp_Lt) == prop_fail_result)
5146                break;                break;
5147              eptr+= len;              eptr+= len;
5148              }              }
# Line 4809  for (;;) Line 5158  for (;;)
5158                break;                break;
5159                }                }
5160              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5161              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result) break;
             if ((prop_category == prop_value) == prop_fail_result)  
               break;  
5162              eptr+= len;              eptr+= len;
5163              }              }
5164            break;            break;
# Line 4826  for (;;) Line 5173  for (;;)
5173                break;                break;
5174                }                }
5175              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5176              prop_chartype = UCD_CHARTYPE(c);              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result) break;
             if ((prop_chartype == prop_value) == prop_fail_result)  
               break;  
5177              eptr+= len;              eptr+= len;
5178              }              }
5179            break;            break;
# Line 4843  for (;;) Line 5188  for (;;)
5188                break;                break;
5189                }                }
5190              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5191              prop_script = UCD_SCRIPT(c);              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result) break;
             if ((prop_script == prop_value) == prop_fail_result)  
               break;  
5192              eptr+= len;              eptr+= len;
5193              }              }
5194            break;            break;
# Line 4853  for (;;) Line 5196  for (;;)
5196            case PT_ALNUM:            case PT_ALNUM:
5197            for (i = min; i < max; i++)            for (i = min; i < max; i++)
5198              {              {
5199                int category;
5200              int len = 1;              int len = 1;
5201              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
5202                {                {
# Line 4860  for (;;) Line 5204  for (;;)
5204                break;                break;
5205                }                }
5206             &nbs