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

Diff of /code/trunk/pcre_exec.c

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

revision 716 by ph10, Tue Oct 4 16:38:05 2011 UTC revision 1376 by ph10, Sat Oct 12 18:02:11 2013 UTC
# Line 6  Line 6 
6  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
7    
8                         Written by Philip Hazel                         Written by Philip Hazel
9             Copyright (c) 1997-2011 University of Cambridge             Copyright (c) 1997-2013 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 37  POSSIBILITY OF SUCH DAMAGE. Line 37  POSSIBILITY OF SUCH DAMAGE.
37  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
38  */  */
39    
   
40  /* This module contains pcre_exec(), the externally visible function that does  /* This module contains pcre_exec(), the externally visible function that does
41  pattern matching using an NFA algorithm, trying to mimic Perl as closely as  pattern matching using an NFA algorithm, trying to mimic Perl as closely as
42  possible. There are also some static supporting functions. */  possible. There are also some static supporting functions. */
# Line 57  possible. There are also some static sup Line 56  possible. There are also some static sup
56  #undef min  #undef min
57  #undef max  #undef max
58    
59    /* The md->capture_last field uses the lower 16 bits for the last captured
60    substring (which can never be greater than 65535) and a bit in the top half
61    to mean "capture vector overflowed". This odd way of doing things was
62    implemented when it was realized that preserving and restoring the overflow bit
63    whenever the last capture number was saved/restored made for a neater
64    interface, and doing it this way saved on (a) another variable, which would
65    have increased the stack frame size (a big NO-NO in PCRE) and (b) another
66    separate set of save/restore instructions. The following defines are used in
67    implementing this. */
68    
69    #define CAPLMASK    0x0000ffff    /* The bits used for last_capture */
70    #define OVFLMASK    0xffff0000    /* The bits used for the overflow flag */
71    #define OVFLBIT     0x00010000    /* The bit that is set for overflow */
72    
73  /* Values for setting in md->match_function_type to indicate two special types  /* Values for setting in md->match_function_type to indicate two special types
74  of call to match(). We do it this way to save on using another stack variable,  of call to match(). We do it this way to save on using another stack variable,
75  as stack usage is to be discouraged. */  as stack usage is to be discouraged. */
# Line 74  defined PCRE_ERROR_xxx codes, which are Line 87  defined PCRE_ERROR_xxx codes, which are
87  negative to avoid the external error codes. */  negative to avoid the external error codes. */
88    
89  #define MATCH_ACCEPT       (-999)  #define MATCH_ACCEPT       (-999)
90  #define MATCH_COMMIT       (-998)  #define MATCH_KETRPOS      (-998)
91  #define MATCH_KETRPOS      (-997)  #define MATCH_ONCE         (-997)
92  #define MATCH_ONCE         (-996)  /* The next 5 must be kept together and in sequence so that a test that checks
93    for any one of them can use a range. */
94    #define MATCH_COMMIT       (-996)
95  #define MATCH_PRUNE        (-995)  #define MATCH_PRUNE        (-995)
96  #define MATCH_SKIP         (-994)  #define MATCH_SKIP         (-994)
97  #define MATCH_SKIP_ARG     (-993)  #define MATCH_SKIP_ARG     (-993)
98  #define MATCH_THEN         (-992)  #define MATCH_THEN         (-992)
99    #define MATCH_BACKTRACK_MAX MATCH_THEN
100  /* This is a convenience macro for code that occurs many times. */  #define MATCH_BACKTRACK_MIN MATCH_COMMIT
   
 #define MRRETURN(ra) \  
   { \  
   md->mark = markptr; \  
   RRETURN(ra); \  
   }  
101    
102  /* Maximum number of ints of offset to save on the stack for recursive calls.  /* Maximum number of ints of offset to save on the stack for recursive calls.
103  If the offset vector is bigger, malloc is used. This should be a multiple of 3,  If the offset vector is bigger, malloc is used. This should be a multiple of 3,
# Line 101  because the offset vector is always a mu Line 110  because the offset vector is always a mu
110  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };
111  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };
112    
   
   
113  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
114  /*************************************************  /*************************************************
115  *        Debugging function to print chars       *  *        Debugging function to print chars       *
# Line 121  Returns:     nothing Line 128  Returns:     nothing
128  */  */
129    
130  static void  static void
131  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)
132  {  {
133  unsigned int c;  pcre_uint32 c;
134    BOOL utf = md->utf;
135  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;
136  while (length-- > 0)  while (length-- > 0)
137    if (isprint(c = *(p++))) printf("%c", c); else printf("\\x%02x", c);    if (isprint(c = RAWUCHARINCTEST(p))) printf("%c", (char)c); else printf("\\x{%02x}", c);
138  }  }
139  #endif  #endif
140    
# Line 148  Arguments: Line 156  Arguments:
156    md          points to match data block    md          points to match data block
157    caseless    TRUE if caseless    caseless    TRUE if caseless
158    
159  Returns:      < 0 if not matched, otherwise the number of subject bytes matched  Returns:      >= 0 the number of subject bytes matched
160                  -1 no match
161                  -2 partial match; always given if at end subject
162  */  */
163    
164  static int  static int
165  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,
166    BOOL caseless)    BOOL caseless)
167  {  {
168  USPTR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
169  register USPTR p = md->start_subject + md->offset_vector[offset];  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
170    #ifdef SUPPORT_UTF
171    BOOL utf = md->utf;
172    #endif
173    
174  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
175  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 171  pchars(p, length, FALSE, md); Line 184  pchars(p, length, FALSE, md);
184  printf("\n");  printf("\n");
185  #endif  #endif
186    
187  /* Always fail if reference not set (and not JavaScript compatible). */  /* Always fail if reference not set (and not JavaScript compatible - in that
188    case the length is passed as zero). */
189    
190  if (length < 0) return -1;  if (length < 0) return -1;
191    
# Line 181  ASCII characters. */ Line 195  ASCII characters. */
195    
196  if (caseless)  if (caseless)
197    {    {
198  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
199  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
200    if (md->utf8)    if (utf)
201      {      {
202      /* 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
203      bytes matched may differ, because there are some characters whose upper and      data units matched may differ, because in UTF-8 there are some characters
204      lower case versions code as different numbers of bytes. For example, U+023A      whose upper and lower case versions code have different numbers of bytes.
205      (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);      For example, U+023A (2 bytes in UTF-8) is the upper case version of U+2C65
206      a sequence of 3 of the former uses 6 bytes, as does a sequence of two of      (3 bytes in UTF-8); a sequence of 3 of the former uses 6 bytes, as does a
207      the latter. It is important, therefore, to check the length along the      sequence of two of the latter. It is important, therefore, to check the
208      reference, not along the subject (earlier code did this wrong). */      length along the reference, not along the subject (earlier code did this
209        wrong). */
210    
211      USPTR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
212      while (p < endptr)      while (p < endptr)
213        {        {
214        int c, d;        pcre_uint32 c, d;
215        if (eptr >= md->end_subject) return -1;        const ucd_record *ur;
216          if (eptr >= md->end_subject) return -2;   /* Partial match */
217        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
218        GETCHARINC(d, p);        GETCHARINC(d, p);
219        if (c != d && c != UCD_OTHERCASE(d)) return -1;        ur = GET_UCD(d);
220          if (c != d && c != d + ur->other_case)
221            {
222            const pcre_uint32 *pp = PRIV(ucd_caseless_sets) + ur->caseset;
223            for (;;)
224              {
225              if (c < *pp) return -1;
226              if (c == *pp++) break;
227              }
228            }
229        }        }
230      }      }
231    else    else
# Line 210  if (caseless) Line 235  if (caseless)
235    /* 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
236    is no UCP support. */    is no UCP support. */
237      {      {
     if (eptr + length > md->end_subject) return -1;  
238      while (length-- > 0)      while (length-- > 0)
239        { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }        {
240          pcre_uint32 cc, cp;
241          if (eptr >= md->end_subject) return -2;   /* Partial match */
242          cc = RAWUCHARTEST(eptr);
243          cp = RAWUCHARTEST(p);
244          if (TABLE_GET(cp, md->lcc, cp) != TABLE_GET(cc, md->lcc, cc)) return -1;
245          p++;
246          eptr++;
247          }
248      }      }
249    }    }
250    
# Line 221  are in UTF-8 mode. */ Line 253  are in UTF-8 mode. */
253    
254  else  else
255    {    {
256    if (eptr + length > md->end_subject) return -1;    while (length-- > 0)
257    while (length-- > 0) if (*p++ != *eptr++) return -1;      {
258        if (eptr >= md->end_subject) return -2;   /* Partial match */
259        if (RAWUCHARINCTEST(p) != RAWUCHARINCTEST(eptr)) return -1;
260        }
261    }    }
262    
263  return eptr - eptr_start;  return (int)(eptr - eptr_start);
264  }  }
265    
266    
# Line 277  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 312  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
312         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
313         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
314         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
315         RM61,  RM62, RM63 };         RM61,  RM62, RM63, RM64, RM65, RM66, RM67, RM68 };
316    
317  /* 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
318  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 290  actually used in this definition. */ Line 325  actually used in this definition. */
325  #define RMATCH(ra,rb,rc,rd,re,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
326    { \    { \
327    printf("match() called in line %d\n", __LINE__); \    printf("match() called in line %d\n", __LINE__); \
328    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1); \    rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1); \
329    printf("to line %d\n", __LINE__); \    printf("to line %d\n", __LINE__); \
330    }    }
331  #define RRETURN(ra) \  #define RRETURN(ra) \
332    { \    { \
333    printf("match() returned %d from line %d ", ra, __LINE__); \    printf("match() returned %d from line %d\n", ra, __LINE__); \
334    return ra; \    return ra; \
335    }    }
336  #else  #else
337  #define RMATCH(ra,rb,rc,rd,re,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
338    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1)    rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1)
339  #define RRETURN(ra) return ra  #define RRETURN(ra) return ra
340  #endif  #endif
341    
# Line 315  argument of match(), which never changes Line 350  argument of match(), which never changes
350    
351  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
352    {\    {\
353    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = frame->Xnextframe;\
354    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL)\
355    frame->Xwhere = rw; \      {\
356        newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
357        if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
358        newframe->Xnextframe = NULL;\
359        frame->Xnextframe = newframe;\
360        }\
361      frame->Xwhere = rw;\
362    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
363    newframe->Xecode = rb;\    newframe->Xecode = rb;\
364    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
   newframe->Xmarkptr = markptr;\  
365    newframe->Xoffset_top = rc;\    newframe->Xoffset_top = rc;\
366    newframe->Xeptrb = re;\    newframe->Xeptrb = re;\
367    newframe->Xrdepth = frame->Xrdepth + 1;\    newframe->Xrdepth = frame->Xrdepth + 1;\
# Line 337  argument of match(), which never changes Line 377  argument of match(), which never changes
377    {\    {\
378    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
379    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
   (pcre_stack_free)(oldframe);\  
380    if (frame != NULL)\    if (frame != NULL)\
381      {\      {\
382      rrc = ra;\      rrc = ra;\
# Line 351  argument of match(), which never changes Line 390  argument of match(), which never changes
390    
391  typedef struct heapframe {  typedef struct heapframe {
392    struct heapframe *Xprevframe;    struct heapframe *Xprevframe;
393      struct heapframe *Xnextframe;
394    
395    /* Function arguments that may change */    /* Function arguments that may change */
396    
397    USPTR Xeptr;    PCRE_PUCHAR Xeptr;
398    const uschar *Xecode;    const pcre_uchar *Xecode;
399    USPTR Xmstart;    PCRE_PUCHAR Xmstart;
   USPTR Xmarkptr;  
400    int Xoffset_top;    int Xoffset_top;
401    eptrblock *Xeptrb;    eptrblock *Xeptrb;
402    unsigned int Xrdepth;    unsigned int Xrdepth;
403    
404    /* Function local variables */    /* Function local variables */
405    
406    USPTR Xcallpat;    PCRE_PUCHAR Xcallpat;
407  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
408    USPTR Xcharptr;    PCRE_PUCHAR Xcharptr;
409  #endif  #endif
410    USPTR Xdata;    PCRE_PUCHAR Xdata;
411    USPTR Xnext;    PCRE_PUCHAR Xnext;
412    USPTR Xpp;    PCRE_PUCHAR Xpp;
413    USPTR Xprev;    PCRE_PUCHAR Xprev;
414    USPTR Xsaved_eptr;    PCRE_PUCHAR Xsaved_eptr;
415    
416    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
417    
# Line 382  typedef struct heapframe { Line 421  typedef struct heapframe {
421    
422  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
423    int Xprop_type;    int Xprop_type;
424    int Xprop_value;    unsigned int Xprop_value;
425    int Xprop_fail_result;    int Xprop_fail_result;
426    int Xoclength;    int Xoclength;
427    uschar Xocchars[8];    pcre_uchar Xocchars[6];
428  #endif  #endif
429    
430    int Xcodelink;    int Xcodelink;
# Line 395  typedef struct heapframe { Line 434  typedef struct heapframe {
434    int Xlength;    int Xlength;
435    int Xmax;    int Xmax;
436    int Xmin;    int Xmin;
437    int Xnumber;    unsigned int Xnumber;
438    int Xoffset;    int Xoffset;
439    int Xop;    unsigned int Xop;
440    int Xsave_capture_last;    pcre_int32 Xsave_capture_last;
441    int Xsave_offset1, Xsave_offset2, Xsave_offset3;    int Xsave_offset1, Xsave_offset2, Xsave_offset3;
442    int Xstacksave[REC_STACK_SAVE_MAX];    int Xstacksave[REC_STACK_SAVE_MAX];
443    
# Line 427  returns a negative (error) response, the Line 466  returns a negative (error) response, the
466  same response. */  same response. */
467    
468  /* These macros pack up tests that are used for partial matching, and which  /* These macros pack up tests that are used for partial matching, and which
469  appears several times in the code. We set the "hit end" flag if the pointer is  appear several times in the code. We set the "hit end" flag if the pointer is
470  at the end of the subject and also past the start of the subject (i.e.  at the end of the subject and also past the start of the subject (i.e.
471  something has been matched). For hard partial matching, we then return  something has been matched). For hard partial matching, we then return
472  immediately. The second one is used when we already know we are past the end of  immediately. The second one is used when we already know we are past the end of
# Line 438  the subject. */ Line 477  the subject. */
477        eptr > md->start_used_ptr) \        eptr > md->start_used_ptr) \
478      { \      { \
479      md->hitend = TRUE; \      md->hitend = TRUE; \
480      if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \      if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \
481      }      }
482    
483  #define SCHECK_PARTIAL()\  #define SCHECK_PARTIAL()\
484    if (md->partial != 0 && eptr > md->start_used_ptr) \    if (md->partial != 0 && eptr > md->start_used_ptr) \
485      { \      { \
486      md->hitend = TRUE; \      md->hitend = TRUE; \
487      if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \      if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \
488      }      }
489    
490    
491  /* Performance note: It might be tempting to extract commonly used fields from  /* Performance note: It might be tempting to extract commonly used fields from
492  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
493  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
494  made performance worse.  made performance worse.
495    
# Line 459  Arguments: Line 498  Arguments:
498     ecode       pointer to current position in compiled code     ecode       pointer to current position in compiled code
499     mstart      pointer to the current match start position (can be modified     mstart      pointer to the current match start position (can be modified
500                   by encountering \K)                   by encountering \K)
    markptr     pointer to the most recent MARK name, or NULL  
501     offset_top  current top pointer     offset_top  current top pointer
502     md          pointer to "static" info for the match     md          pointer to "static" info for the match
503     eptrb       pointer to chain of blocks containing eptr at start of     eptrb       pointer to chain of blocks containing eptr at start of
# Line 474  Returns:       MATCH_MATCH if matched Line 512  Returns:       MATCH_MATCH if matched
512  */  */
513    
514  static int  static int
515  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode,
516    const uschar *markptr, int offset_top, match_data *md, eptrblock *eptrb,    PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb,
517    unsigned int rdepth)    unsigned int rdepth)
518  {  {
519  /* These variables do not need to be preserved over recursion in this function,  /* These variables do not need to be preserved over recursion in this function,
# Line 484  so they can be ordinary variables in all Line 522  so they can be ordinary variables in all
522    
523  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
524  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
525  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register pcre_uint32 c;    /* Character values not kept over RMATCH() calls */
526  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
527    
528  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
529  BOOL caseless;  BOOL caseless;
530  int condcode;  int condcode;
531    
532  /* 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
533  preserved over calls to RMATCH() are part of a "frame" which is obtained from  preserved over calls to RMATCH() are part of a "frame". We set up the top-level
534  heap storage. Set up the top-level frame here; others are obtained from the  frame on the stack here; subsequent instantiations are obtained from the heap
535  heap whenever RMATCH() does a "recursion". See the macro definitions above. */  whenever RMATCH() does a "recursion". See the macro definitions above. Putting
536    the top-level on the stack rather than malloc-ing them all gives a performance
537    boost in many cases where there is not much "recursion". */
538    
539  #ifdef NO_RECURSE  #ifdef NO_RECURSE
540  heapframe *frame = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));  heapframe *frame = (heapframe *)md->match_frames_base;
 if (frame == NULL) RRETURN(PCRE_ERROR_NOMEMORY);  
 frame->Xprevframe = NULL;            /* Marks the top level */  
541    
542  /* Copy in the original argument variables */  /* Copy in the original argument variables */
543    
544  frame->Xeptr = eptr;  frame->Xeptr = eptr;
545  frame->Xecode = ecode;  frame->Xecode = ecode;
546  frame->Xmstart = mstart;  frame->Xmstart = mstart;
 frame->Xmarkptr = markptr;  
547  frame->Xoffset_top = offset_top;  frame->Xoffset_top = offset_top;
548  frame->Xeptrb = eptrb;  frame->Xeptrb = eptrb;
549  frame->Xrdepth = rdepth;  frame->Xrdepth = rdepth;
# Line 520  HEAP_RECURSE: Line 557  HEAP_RECURSE:
557  #define eptr               frame->Xeptr  #define eptr               frame->Xeptr
558  #define ecode              frame->Xecode  #define ecode              frame->Xecode
559  #define mstart             frame->Xmstart  #define mstart             frame->Xmstart
 #define markptr            frame->Xmarkptr  
560  #define offset_top         frame->Xoffset_top  #define offset_top         frame->Xoffset_top
561  #define eptrb              frame->Xeptrb  #define eptrb              frame->Xeptrb
562  #define rdepth             frame->Xrdepth  #define rdepth             frame->Xrdepth
563    
564  /* Ditto for the local variables */  /* Ditto for the local variables */
565    
566  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
567  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
568  #endif  #endif
569  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
# Line 585  declarations can be cut out in a block. Line 621  declarations can be cut out in a block.
621  below are for variables that do not have to be preserved over a recursive call  below are for variables that do not have to be preserved over a recursive call
622  to RMATCH(). */  to RMATCH(). */
623    
624  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
625  const uschar *charptr;  const pcre_uchar *charptr;
626  #endif  #endif
627  const uschar *callpat;  const pcre_uchar *callpat;
628  const uschar *data;  const pcre_uchar *data;
629  const uschar *next;  const pcre_uchar *next;
630  USPTR         pp;  PCRE_PUCHAR       pp;
631  const uschar *prev;  const pcre_uchar *prev;
632  USPTR         saved_eptr;  PCRE_PUCHAR       saved_eptr;
633    
634  recursion_info new_recursive;  recursion_info new_recursive;
635    
# Line 603  BOOL prev_is_word; Line 639  BOOL prev_is_word;
639    
640  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
641  int prop_type;  int prop_type;
642  int prop_value;  unsigned int prop_value;
643  int prop_fail_result;  int prop_fail_result;
644  int oclength;  int oclength;
645  uschar occhars[8];  pcre_uchar occhars[6];
646  #endif  #endif
647    
648  int codelink;  int codelink;
# Line 614  int ctype; Line 650  int ctype;
650  int length;  int length;
651  int max;  int max;
652  int min;  int min;
653  int number;  unsigned int number;
654  int offset;  int offset;
655  int op;  unsigned int op;
656  int save_capture_last;  pcre_int32 save_capture_last;
657  int save_offset1, save_offset2, save_offset3;  int save_offset1, save_offset2, save_offset3;
658  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
659    
660  eptrblock newptrb;  eptrblock newptrb;
661    
662    /* There is a special fudge for calling match() in a way that causes it to
663    measure the size of its basic stack frame when the stack is being used for
664    recursion. The second argument (ecode) being NULL triggers this behaviour. It
665    cannot normally ever be NULL. The return is the negated value of the frame
666    size. */
667    
668    if (ecode == NULL)
669      {
670      if (rdepth == 0)
671        return match((PCRE_PUCHAR)&rdepth, NULL, NULL, 0, NULL, NULL, 1);
672      else
673        {
674        int len = (char *)&rdepth - (char *)eptr;
675        return (len > 0)? -len : len;
676        }
677      }
678  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
679    
680  /* To save space on the stack and in the heap frame, I have doubled up on some  /* To save space on the stack and in the heap frame, I have doubled up on some
# Line 634  the alternative names that are used. */ Line 687  the alternative names that are used. */
687  #define code_offset   codelink  #define code_offset   codelink
688  #define condassert    condition  #define condassert    condition
689  #define matched_once  prev_is_word  #define matched_once  prev_is_word
690    #define foc           number
691    #define save_mark     data
692    
693  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
694  variables. */  variables. */
# Line 659  defined). However, RMATCH isn't like a f Line 714  defined). However, RMATCH isn't like a f
714  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,
715  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
716    
717  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
718  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
719  #else  #else
720  utf8 = FALSE;  utf = FALSE;
721  #endif  #endif
722    
723  /* 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 701  for (;;) Line 756  for (;;)
756    switch(op)    switch(op)
757      {      {
758      case OP_MARK:      case OP_MARK:
759      markptr = ecode + 2;      md->nomatch_mark = ecode + 2;
760      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      md->mark = NULL;    /* In case previously set by assertion */
761        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
762        eptrb, RM55);        eptrb, RM55);
763        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
764             md->mark == NULL) md->mark = ecode + 2;
765    
766      /* 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
767      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 712  for (;;) Line 770  for (;;)
770      position and return MATCH_SKIP. Otherwise, pass back the return code      position and return MATCH_SKIP. Otherwise, pass back the return code
771      unaltered. */      unaltered. */
772    
773      if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
774          strcmp((char *)markptr, (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC_TEST(ecode + 2, md->start_match_ptr) == 0)
775        {        {
776        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
777        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
778        }        }
   
     if (md->mark == NULL) md->mark = markptr;  
779      RRETURN(rrc);      RRETURN(rrc);
780    
781      case OP_FAIL:      case OP_FAIL:
782      MRRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
   
     /* COMMIT overrides PRUNE, SKIP, and THEN */  
783    
784      case OP_COMMIT:      case OP_COMMIT:
785      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
786        eptrb, RM52);        eptrb, RM52);
787      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
788          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&      RRETURN(MATCH_COMMIT);
         rrc != MATCH_THEN)  
       RRETURN(rrc);  
     MRRETURN(MATCH_COMMIT);  
   
     /* PRUNE overrides THEN */  
789    
790      case OP_PRUNE:      case OP_PRUNE:
791      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
792        eptrb, RM51);        eptrb, RM51);
793      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
794      MRRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
795    
796      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
797      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      md->nomatch_mark = ecode + 2;
798        md->mark = NULL;    /* In case previously set by assertion */
799        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
800        eptrb, RM56);        eptrb, RM56);
801      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
802      md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
803        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
804      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
805    
     /* SKIP overrides PRUNE and THEN */  
   
806      case OP_SKIP:      case OP_SKIP:
807      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
808        eptrb, RM53);        eptrb, RM53);
809      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
       RRETURN(rrc);  
810      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
811      MRRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
812    
813        /* Note that, for Perl compatibility, SKIP with an argument does NOT set
814        nomatch_mark. When a pattern match ends with a SKIP_ARG for which there was
815        not a matching mark, we have to re-run the match, ignoring the SKIP_ARG
816        that failed and any that precede it (either they also failed, or were not
817        triggered). To do this, we maintain a count of executed SKIP_ARGs. If a
818        SKIP_ARG gets to top level, the match is re-run with md->ignore_skip_arg
819        set to the count of the one that failed. */
820    
821      case OP_SKIP_ARG:      case OP_SKIP_ARG:
822      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      md->skip_arg_count++;
823        if (md->skip_arg_count <= md->ignore_skip_arg)
824          {
825          ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
826          break;
827          }
828        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
829        eptrb, RM57);        eptrb, RM57);
830      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
       RRETURN(rrc);  
831    
832      /* Pass back the current skip name by overloading md->start_match_ptr and      /* Pass back the current skip name by overloading md->start_match_ptr and
833      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
834      caught by a matching MARK, or get to the top, where it is treated the same      caught by a matching MARK, or get to the top, where it causes a rematch
835      as PRUNE. */      with md->ignore_skip_arg set to the value of md->skip_arg_count. */
836    
837      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
838      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
# Line 780  for (;;) Line 842  for (;;)
842      match pointer to do this. */      match pointer to do this. */
843    
844      case OP_THEN:      case OP_THEN:
845      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
846        eptrb, RM54);        eptrb, RM54);
847      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
848      md->start_match_ptr = ecode;      md->start_match_ptr = ecode;
849      MRRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
850    
851      case OP_THEN_ARG:      case OP_THEN_ARG:
852      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top,      md->nomatch_mark = ecode + 2;
853        md->mark = NULL;    /* In case previously set by assertion */
854        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
855        md, eptrb, RM58);        md, eptrb, RM58);
856        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
857             md->mark == NULL) md->mark = ecode + 2;
858      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
859      md->start_match_ptr = ecode;      md->start_match_ptr = ecode;
     md->mark = ecode + 2;  
860      RRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
861    
862        /* Handle an atomic group that does not contain any capturing parentheses.
863        This can be handled like an assertion. Prior to 8.13, all atomic groups
864        were handled this way. In 8.13, the code was changed as below for ONCE, so
865        that backups pass through the group and thereby reset captured values.
866        However, this uses a lot more stack, so in 8.20, atomic groups that do not
867        contain any captures generate OP_ONCE_NC, which can be handled in the old,
868        less stack intensive way.
869    
870        Check the alternative branches in turn - the matching won't pass the KET
871        for this kind of subpattern. If any one branch matches, we carry on as at
872        the end of a normal bracket, leaving the subject pointer, but resetting
873        the start-of-match value in case it was changed by \K. */
874    
875        case OP_ONCE_NC:
876        prev = ecode;
877        saved_eptr = eptr;
878        save_mark = md->mark;
879        do
880          {
881          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
882          if (rrc == MATCH_MATCH)  /* Note: _not_ MATCH_ACCEPT */
883            {
884            mstart = md->start_match_ptr;
885            break;
886            }
887          if (rrc == MATCH_THEN)
888            {
889            next = ecode + GET(ecode,1);
890            if (md->start_match_ptr < next &&
891                (*ecode == OP_ALT || *next == OP_ALT))
892              rrc = MATCH_NOMATCH;
893            }
894    
895          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
896          ecode += GET(ecode,1);
897          md->mark = save_mark;
898          }
899        while (*ecode == OP_ALT);
900    
901        /* If hit the end of the group (which could be repeated), fail */
902    
903        if (*ecode != OP_ONCE_NC && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);
904    
905        /* Continue as from after the group, updating the offsets high water
906        mark, since extracts may have been taken. */
907    
908        do ecode += GET(ecode, 1); while (*ecode == OP_ALT);
909    
910        offset_top = md->end_offset_top;
911        eptr = md->end_match_ptr;
912    
913        /* For a non-repeating ket, just continue at this level. This also
914        happens for a repeating ket if no characters were matched in the group.
915        This is the forcible breaking of infinite loops as implemented in Perl
916        5.005. */
917    
918        if (*ecode == OP_KET || eptr == saved_eptr)
919          {
920          ecode += 1+LINK_SIZE;
921          break;
922          }
923    
924        /* The repeating kets try the rest of the pattern or restart from the
925        preceding bracket, in the appropriate order. The second "call" of match()
926        uses tail recursion, to avoid using another stack frame. */
927    
928        if (*ecode == OP_KETRMIN)
929          {
930          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM65);
931          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
932          ecode = prev;
933          goto TAIL_RECURSE;
934          }
935        else  /* OP_KETRMAX */
936          {
937          RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
938          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
939          ecode += 1 + LINK_SIZE;
940          goto TAIL_RECURSE;
941          }
942        /* Control never gets here */
943    
944      /* Handle a capturing bracket, other than those that are possessive with an      /* Handle a capturing bracket, other than those that are possessive with an
945      unlimited repeat. If there is space in the offset vector, save the current      unlimited repeat. If there is space in the offset vector, save the current
946      subject position in the working slot at the top of the vector. We mustn't      subject position in the working slot at the top of the vector. We mustn't
# Line 826  for (;;) Line 973  for (;;)
973        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
974        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
975        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
976          save_mark = md->mark;
977    
978        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
979        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 834  for (;;) Line 982  for (;;)
982        for (;;)        for (;;)
983          {          {
984          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
985          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
986            eptrb, RM1);            eptrb, RM1);
987          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
988    
989          /* If we backed up to a THEN, check whether it is within the current          /* If we backed up to a THEN, check whether it is within the current
990          branch by comparing the address of the THEN that is passed back with          branch by comparing the address of the THEN that is passed back with
991          the end of the branch. If it is within the current branch, and the          the end of the branch. If it is within the current branch, and the
992          branch is one of two or more alternatives (it either starts or ends          branch is one of two or more alternatives (it either starts or ends
993          with OP_ALT), we have reached the limit of THEN's action, so convert          with OP_ALT), we have reached the limit of THEN's action, so convert
994          the return code to NOMATCH, which will cause normal backtracking to          the return code to NOMATCH, which will cause normal backtracking to
995          happen from now on. Otherwise, THEN is passed back to an outer          happen from now on. Otherwise, THEN is passed back to an outer
996          alternative. This implements Perl's treatment of parenthesized groups,          alternative. This implements Perl's treatment of parenthesized groups,
997          where a group not containing | does not affect the current alternative,          where a group not containing | does not affect the current alternative,
998          that is, (X) is NOT the same as (X|(*F)). */          that is, (X) is NOT the same as (X|(*F)). */
999    
1000          if (rrc == MATCH_THEN)          if (rrc == MATCH_THEN)
1001            {            {
1002            next = ecode + GET(ecode,1);            next = ecode + GET(ecode,1);
1003            if (md->start_match_ptr < next &&            if (md->start_match_ptr < next &&
1004                (*ecode == OP_ALT || *next == OP_ALT))                (*ecode == OP_ALT || *next == OP_ALT))
1005              rrc = MATCH_NOMATCH;              rrc = MATCH_NOMATCH;
1006            }            }
1007    
1008          /* Anything other than NOMATCH is passed back. */          /* Anything other than NOMATCH is passed back. */
1009    
1010          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1011          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
1012          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
1013            md->mark = save_mark;
1014          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
1015          }          }
1016    
# Line 872  for (;;) Line 1021  for (;;)
1021    
1022        /* At this point, rrc will be one of MATCH_ONCE or MATCH_NOMATCH. */        /* At this point, rrc will be one of MATCH_ONCE or MATCH_NOMATCH. */
1023    
       if (md->mark == NULL) md->mark = markptr;  
1024        RRETURN(rrc);        RRETURN(rrc);
1025        }        }
1026    
# Line 888  for (;;) Line 1036  for (;;)
1036      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1037    
1038      /* Non-capturing or atomic group, except for possessive with unlimited      /* Non-capturing or atomic group, except for possessive with unlimited
1039      repeat. Loop for all the alternatives.      repeat and ONCE group with no captures. Loop for all the alternatives.
1040    
1041      When we get to the final alternative within the brackets, we used to return      When we get to the final alternative within the brackets, we used to return
1042      the result of a recursive call to match() whatever happened so it was      the result of a recursive call to match() whatever happened so it was
# Line 914  for (;;) Line 1062  for (;;)
1062    
1063      for (;;)      for (;;)
1064        {        {
1065        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1066            md->match_function_type = MATCH_CBEGROUP;
1067    
1068        /* If this is not a possibly empty group, and there are no (*THEN)s in        /* If this is not a possibly empty group, and there are no (*THEN)s in
1069        the pattern, and this is the final alternative, optimize as described        the pattern, and this is the final alternative, optimize as described
# Line 922  for (;;) Line 1071  for (;;)
1071    
1072        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1073          {          {
1074          ecode += _pcre_OP_lengths[*ecode];          ecode += PRIV(OP_lengths)[*ecode];
1075          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1076          }          }
1077    
1078        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1079    
1080        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,        save_mark = md->mark;
1081          save_capture_last = md->capture_last;
1082          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1083          RM2);          RM2);
1084    
1085        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
1086        THEN. */        THEN. */
1087    
1088        if (rrc == MATCH_THEN)        if (rrc == MATCH_THEN)
1089          {          {
1090          next = ecode + GET(ecode,1);          next = ecode + GET(ecode,1);
1091          if (md->start_match_ptr < next &&          if (md->start_match_ptr < next &&
1092              (*ecode == OP_ALT || *next == OP_ALT))              (*ecode == OP_ALT || *next == OP_ALT))
1093            rrc = MATCH_NOMATCH;            rrc = MATCH_NOMATCH;
1094          }          }
1095    
1096        if (rrc != MATCH_NOMATCH)        if (rrc != MATCH_NOMATCH)
1097          {          {
1098          if (rrc == MATCH_ONCE)          if (rrc == MATCH_ONCE)
1099            {            {
1100            const uschar *scode = ecode;            const pcre_uchar *scode = ecode;
1101            if (*scode != OP_ONCE)           /* If not at start, find it */            if (*scode != OP_ONCE)           /* If not at start, find it */
1102              {              {
1103              while (*scode == OP_ALT) scode += GET(scode, 1);              while (*scode == OP_ALT) scode += GET(scode, 1);
# Line 957  for (;;) Line 1108  for (;;)
1108          RRETURN(rrc);          RRETURN(rrc);
1109          }          }
1110        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1111          md->mark = save_mark;
1112        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1113          md->capture_last = save_capture_last;
1114        }        }
1115    
     if (md->mark == NULL) md->mark = markptr;  
1116      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
1117    
1118      /* Handle possessive capturing brackets with an unlimited repeat. We come      /* Handle possessive capturing brackets with an unlimited repeat. We come
# Line 989  for (;;) Line 1141  for (;;)
1141      if (offset < md->offset_max)      if (offset < md->offset_max)
1142        {        {
1143        matched_once = FALSE;        matched_once = FALSE;
1144        code_offset = ecode - md->start_code;        code_offset = (int)(ecode - md->start_code);
1145    
1146        save_offset1 = md->offset_vector[offset];        save_offset1 = md->offset_vector[offset];
1147        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
# Line 1012  for (;;) Line 1164  for (;;)
1164          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1165            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1166          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1167          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1168            eptrb, RM63);            eptrb, RM63);
1169          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1170            {            {
# Line 1023  for (;;) Line 1175  for (;;)
1175            matched_once = TRUE;            matched_once = TRUE;
1176            continue;            continue;
1177            }            }
1178    
1179          /* See comment in the code for capturing groups above about handling          /* See comment in the code for capturing groups above about handling
1180          THEN. */          THEN. */
1181    
1182          if (rrc == MATCH_THEN)          if (rrc == MATCH_THEN)
1183            {            {
1184            next = ecode + GET(ecode,1);            next = ecode + GET(ecode,1);
1185            if (md->start_match_ptr < next &&            if (md->start_match_ptr < next &&
1186                (*ecode == OP_ALT || *next == OP_ALT))                (*ecode == OP_ALT || *next == OP_ALT))
1187              rrc = MATCH_NOMATCH;              rrc = MATCH_NOMATCH;
1188            }            }
1189    
1190          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1191          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
# Line 1048  for (;;) Line 1200  for (;;)
1200          md->offset_vector[md->offset_end - number] = save_offset3;          md->offset_vector[md->offset_end - number] = save_offset3;
1201          }          }
1202    
       if (md->mark == NULL) md->mark = markptr;  
1203        if (allow_zero || matched_once)        if (allow_zero || matched_once)
1204          {          {
1205          ecode += 1 + LINK_SIZE;          ecode += 1 + LINK_SIZE;
# Line 1080  for (;;) Line 1231  for (;;)
1231    
1232      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1233      matched_once = FALSE;      matched_once = FALSE;
1234      code_offset = ecode - md->start_code;      code_offset = (int)(ecode - md->start_code);
1235        save_capture_last = md->capture_last;
1236    
1237      for (;;)      for (;;)
1238        {        {
1239        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1240        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1241          eptrb, RM48);          eptrb, RM48);
1242        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1243          {          {
# Line 1095  for (;;) Line 1247  for (;;)
1247          matched_once = TRUE;          matched_once = TRUE;
1248          continue;          continue;
1249          }          }
1250    
1251        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
1252        THEN. */        THEN. */
1253    
1254        if (rrc == MATCH_THEN)        if (rrc == MATCH_THEN)
1255          {          {
1256          next = ecode + GET(ecode,1);          next = ecode + GET(ecode,1);
1257          if (md->start_match_ptr < next &&          if (md->start_match_ptr < next &&
1258              (*ecode == OP_ALT || *next == OP_ALT))              (*ecode == OP_ALT || *next == OP_ALT))
1259            rrc = MATCH_NOMATCH;            rrc = MATCH_NOMATCH;
1260          }          }
1261    
1262        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1263        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1264        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1265          md->capture_last = save_capture_last;
1266        }        }
1267    
1268      if (matched_once || allow_zero)      if (matched_once || allow_zero)
# Line 1121  for (;;) Line 1274  for (;;)
1274    
1275      /* Control never reaches here. */      /* Control never reaches here. */
1276    
1277      /* Conditional group: compilation checked that there are no more than      /* Conditional group: compilation checked that there are no more than two
1278      two branches. If the condition is false, skipping the first branch takes us      branches. If the condition is false, skipping the first branch takes us
1279      past the end if there is only one branch, but that's OK because that is      past the end of the item if there is only one branch, but that's exactly
1280      exactly what going to the ket would do. */      what we want. */
1281    
1282      case OP_COND:      case OP_COND:
1283      case OP_SCOND:      case OP_SCOND:
1284      codelink = GET(ecode, 1);  
1285        /* The variable codelink will be added to ecode when the condition is
1286        false, to get to the second branch. Setting it to the offset to the ALT
1287        or KET, then incrementing ecode achieves this effect. We now have ecode
1288        pointing to the condition or callout. */
1289    
1290        codelink = GET(ecode, 1);   /* Offset to the second branch */
1291        ecode += 1 + LINK_SIZE;     /* From this opcode */
1292    
1293      /* 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
1294      inserted between OP_COND and an assertion condition. */      inserted between OP_COND and an assertion condition. */
1295    
1296      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (*ecode == OP_CALLOUT)
1297        {        {
1298        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
1299          {          {
1300          pcre_callout_block cb;          PUBL(callout_block) cb;
1301          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1302          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[1];
1303          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1304    #if defined COMPILE_PCRE8
1305          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1306    #elif defined COMPILE_PCRE16
1307            cb.subject          = (PCRE_SPTR16)md->start_subject;
1308    #elif defined COMPILE_PCRE32
1309            cb.subject          = (PCRE_SPTR32)md->start_subject;
1310    #endif
1311          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1312          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1313          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
1314          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, 2);
1315          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1316          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1317          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last & CAPLMASK;
1318            /* Internal change requires this for API compatibility. */
1319            if (cb.capture_last == 0) cb.capture_last = -1;
1320          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1321          cb.mark             = markptr;          cb.mark             = md->nomatch_mark;
1322          if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1323          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1324          }          }
       ecode += _pcre_OP_lengths[OP_CALLOUT];  
       }  
1325    
1326      condcode = ecode[LINK_SIZE+1];        /* Advance ecode past the callout, so it now points to the condition. We
1327          must adjust codelink so that the value of ecode+codelink is unchanged. */
1328    
1329          ecode += PRIV(OP_lengths)[OP_CALLOUT];
1330          codelink -= PRIV(OP_lengths)[OP_CALLOUT];
1331          }
1332    
1333      /* Now see what the actual condition is */      /* Test the various possible conditions */
1334    
1335      if (condcode == OP_RREF || condcode == OP_NRREF)    /* Recursion test */      condition = FALSE;
1336        switch(condcode = *ecode)
1337        {        {
1338        if (md->recursive == NULL)                /* Not recursing => FALSE */        case OP_RREF:         /* Numbered group recursion test */
1339          if (md->recursive != NULL)     /* Not recursing => FALSE */
1340          {          {
1341          condition = FALSE;          unsigned int recno = GET2(ecode, 1);   /* Recursion group number*/
1342          ecode += GET(ecode, 1);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1343          }          }
1344        else        break;
         {  
         int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/  
         condition =  (recno == RREF_ANY || recno == md->recursive->group_num);  
   
         /* If the test is for recursion into a specific subpattern, and it is  
         false, but the test was set up by name, scan the table to see if the  
         name refers to any other numbers, and test them. The condition is true  
         if any one is set. */  
   
         if (!condition && condcode == OP_NRREF && recno != RREF_ANY)  
           {  
           uschar *slotA = md->name_table;  
           for (i = 0; i < md->name_count; i++)  
             {  
             if (GET2(slotA, 0) == recno) break;  
             slotA += md->name_entry_size;  
             }  
   
           /* Found a name for the number - there can be only one; duplicate  
           names for different numbers are allowed, but not vice versa. First  
           scan down for duplicates. */  
   
           if (i < md->name_count)  
             {  
             uschar *slotB = slotA;  
             while (slotB > md->name_table)  
               {  
               slotB -= md->name_entry_size;  
               if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)  
                 {  
                 condition = GET2(slotB, 0) == md->recursive->group_num;  
                 if (condition) break;  
                 }  
               else break;  
               }  
   
             /* Scan up for duplicates */  
1345    
1346              if (!condition)        case OP_DNRREF:       /* Duplicate named group recursion test */
1347                {        if (md->recursive != NULL)
1348                slotB = slotA;          {
1349                for (i++; i < md->name_count; i++)          int count = GET2(ecode, 1 + IMM2_SIZE);
1350                  {          pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
1351                  slotB += md->name_entry_size;          while (count-- > 0)
1352                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)            {
1353                    {            unsigned int recno = GET2(slot, 0);
1354                    condition = GET2(slotB, 0) == md->recursive->group_num;            condition = recno == md->recursive->group_num;
1355                    if (condition) break;            if (condition) break;
1356                    }            slot += md->name_entry_size;
                 else break;  
                 }  
               }  
             }  
1357            }            }
   
         /* Chose branch according to the condition */  
   
         ecode += condition? 3 : GET(ecode, 1);  
1358          }          }
1359        }        break;
1360    
1361      else if (condcode == OP_CREF || condcode == OP_NCREF)  /* Group used test */        case OP_CREF:         /* Numbered group used test */
1362        {        offset = GET2(ecode, 1) << 1;  /* Doubled ref number */
       offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */  
1363        condition = offset < offset_top && md->offset_vector[offset] >= 0;        condition = offset < offset_top && md->offset_vector[offset] >= 0;
1364          break;
1365    
1366        /* If the numbered capture is unset, but the reference was by name,        case OP_DNCREF:      /* Duplicate named group used test */
       scan the table to see if the name refers to any other numbers, and test  
       them. The condition is true if any one is set. This is tediously similar  
       to the code above, but not close enough to try to amalgamate. */  
   
       if (!condition && condcode == OP_NCREF)  
1367          {          {
1368          int refno = offset >> 1;          int count = GET2(ecode, 1 + IMM2_SIZE);
1369          uschar *slotA = md->name_table;          pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
1370            while (count-- > 0)
1371          for (i = 0; i < md->name_count; i++)            {
1372            {            offset = GET2(slot, 0) << 1;
1373            if (GET2(slotA, 0) == refno) break;            condition = offset < offset_top && md->offset_vector[offset] >= 0;
1374            slotA += md->name_entry_size;            if (condition) break;
1375            }            slot += md->name_entry_size;
   
         /* Found a name for the number - there can be only one; duplicate names  
         for different numbers are allowed, but not vice versa. First scan down  
         for duplicates. */  
   
         if (i < md->name_count)  
           {  
           uschar *slotB = slotA;  
           while (slotB > md->name_table)  
             {  
             slotB -= md->name_entry_size;  
             if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)  
               {  
               offset = GET2(slotB, 0) << 1;  
               condition = offset < offset_top &&  
                 md->offset_vector[offset] >= 0;  
               if (condition) break;  
               }  
             else break;  
             }  
   
           /* Scan up for duplicates */  
   
           if (!condition)  
             {  
             slotB = slotA;  
             for (i++; i < md->name_count; i++)  
               {  
               slotB += md->name_entry_size;  
               if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)  
                 {  
                 offset = GET2(slotB, 0) << 1;  
                 condition = offset < offset_top &&  
                   md->offset_vector[offset] >= 0;  
                 if (condition) break;  
                 }  
               else break;  
               }  
             }  
1376            }            }
1377          }          }
1378          break;
1379    
1380        /* Chose branch according to the condition */        case OP_DEF:     /* DEFINE - always false */
1381          break;
       ecode += condition? 3 : GET(ecode, 1);  
       }  
   
     else if (condcode == OP_DEF)     /* DEFINE - always false */  
       {  
       condition = FALSE;  
       ecode += GET(ecode, 1);  
       }  
1382    
1383      /* The condition is an assertion. Call match() to evaluate it - setting        /* The condition is an assertion. Call match() to evaluate it - setting
1384      md->match_function_type to MATCH_CONDASSERT causes it to stop at the end of        md->match_function_type to MATCH_CONDASSERT causes it to stop at the end
1385      an assertion. */        of an assertion. */
1386    
1387      else        default:
       {  
1388        md->match_function_type = MATCH_CONDASSERT;        md->match_function_type = MATCH_CONDASSERT;
1389        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);        RMATCH(eptr, ecode, offset_top, md, NULL, RM3);
1390        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
1391          {          {
1392          if (md->end_offset_top > offset_top)          if (md->end_offset_top > offset_top)
1393            offset_top = md->end_offset_top;  /* Captures may have happened */            offset_top = md->end_offset_top;  /* Captures may have happened */
1394          condition = TRUE;          condition = TRUE;
1395          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);  
1396            /* Advance ecode past the assertion to the start of the first branch,
1397            but adjust it so that the general choosing code below works. */
1398    
1399            ecode += GET(ecode, 1);
1400          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
1401            ecode += 1 + LINK_SIZE - PRIV(OP_lengths)[condcode];
1402          }          }
1403    
1404        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an
1405        assertion; it is therefore treated as NOMATCH. */        assertion; it is therefore treated as NOMATCH. Any other return is an
1406          error. */
1407    
1408        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1409          {          {
1410          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
1411          }          }
1412        else        break;
         {  
         condition = FALSE;  
         ecode += codelink;  
         }  
1413        }        }
1414    
1415      /* We are now at the branch that is to be obeyed. As there is only one, can      /* Choose branch according to the condition */
     use tail recursion to avoid using another stack frame, except when there is  
     unlimited repeat of a possibly empty group. In the latter case, a recursive  
     call to match() is always required, unless the second alternative doesn't  
     exist, in which case we can just plough on. Note that, for compatibility  
     with Perl, the | in a conditional group is NOT treated as creating two  
     alternatives. If a THEN is encountered in the branch, it propagates out to  
     the enclosing alternative (unless nested in a deeper set of alternatives,  
     of course). */  
1416    
1417      if (condition || *ecode == OP_ALT)      ecode += condition? PRIV(OP_lengths)[condcode] : codelink;
1418    
1419        /* We are now at the branch that is to be obeyed. As there is only one, we
1420        can use tail recursion to avoid using another stack frame, except when
1421        there is unlimited repeat of a possibly empty group. In the latter case, a
1422        recursive call to match() is always required, unless the second alternative
1423        doesn't exist, in which case we can just plough on. Note that, for
1424        compatibility with Perl, the | in a conditional group is NOT treated as
1425        creating two alternatives. If a THEN is encountered in the branch, it
1426        propagates out to the enclosing alternative (unless nested in a deeper set
1427        of alternatives, of course). */
1428    
1429        if (condition || ecode[-(1+LINK_SIZE)] == OP_ALT)
1430        {        {
1431        if (op != OP_SCOND)        if (op != OP_SCOND)
1432          {          {
         ecode += 1 + LINK_SIZE;  
1433          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1434          }          }
1435    
1436        md->match_function_type = MATCH_CBEGROUP;        md->match_function_type = MATCH_CBEGROUP;
1437        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);        RMATCH(eptr, ecode, offset_top, md, eptrb, RM49);
1438        RRETURN(rrc);        RRETURN(rrc);
1439        }        }
1440    
# Line 1361  for (;;) Line 1442  for (;;)
1442    
1443      else      else
1444        {        {
       ecode += 1 + LINK_SIZE;  
1445        }        }
1446      break;      break;
1447    
# Line 1370  for (;;) Line 1450  for (;;)
1450      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1451    
1452      case OP_CLOSE:      case OP_CLOSE:
1453      number = GET2(ecode, 1);      number = GET2(ecode, 1);   /* Must be less than 65536 */
1454      offset = number << 1;      offset = number << 1;
1455    
1456  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 1378  for (;;) Line 1458  for (;;)
1458        printf("\n");        printf("\n");
1459  #endif  #endif
1460    
1461      md->capture_last = number;      md->capture_last = (md->capture_last & OVFLMASK) | number;
1462      if (offset >= md->offset_max) md->offset_overflow = TRUE; else      if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1463        {        {
1464        md->offset_vector[offset] =        md->offset_vector[offset] =
1465          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
1466        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1467        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1468        }        }
1469      ecode += 3;      ecode += 1 + IMM2_SIZE;
1470      break;      break;
1471    
1472    
# Line 1406  for (;;) Line 1486  for (;;)
1486           (md->notempty ||           (md->notempty ||
1487             (md->notempty_atstart &&             (md->notempty_atstart &&
1488               mstart == md->start_subject + md->start_offset)))               mstart == md->start_subject + md->start_offset)))
1489        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1490    
1491      /* Otherwise, we have a match. */      /* Otherwise, we have a match. */
1492    
# Line 1415  for (;;) Line 1495  for (;;)
1495      md->start_match_ptr = mstart;       /* and the start (\K can modify) */      md->start_match_ptr = mstart;       /* and the start (\K can modify) */
1496    
1497      /* For some reason, the macros don't work properly if an expression is      /* For some reason, the macros don't work properly if an expression is
1498      given as the argument to MRRETURN when the heap is in use. */      given as the argument to RRETURN when the heap is in use. */
1499    
1500      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;
1501      MRRETURN(rrc);      RRETURN(rrc);
1502    
1503      /* Assertion brackets. Check the alternative branches in turn - the      /* Assertion brackets. Check the alternative branches in turn - the
1504      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,
# Line 1433  for (;;) Line 1513  for (;;)
1513    
1514      case OP_ASSERT:      case OP_ASSERT:
1515      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1516        save_mark = md->mark;
1517      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1518        {        {
1519        condassert = TRUE;        condassert = TRUE;
# Line 1440  for (;;) Line 1521  for (;;)
1521        }        }
1522      else condassert = FALSE;      else condassert = FALSE;
1523    
1524        /* Loop for each branch */
1525    
1526      do      do
1527        {        {
1528        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
1529    
1530          /* A match means that the assertion is true; break out of the loop
1531          that matches its alternatives. */
1532    
1533        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1534          {          {
1535          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
         markptr = md->mark;  
1536          break;          break;
1537          }          }
1538    
1539        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* If not matched, restore the previous mark setting. */
1540        as NOMATCH. */  
1541          md->mark = save_mark;
1542        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);  
1543          /* See comment in the code for capturing groups above about handling
1544          THEN. */
1545    
1546          if (rrc == MATCH_THEN)
1547            {
1548            next = ecode + GET(ecode,1);
1549            if (md->start_match_ptr < next &&
1550                (*ecode == OP_ALT || *next == OP_ALT))
1551              rrc = MATCH_NOMATCH;
1552            }
1553    
1554          /* Anything other than NOMATCH causes the entire assertion to fail,
1555          passing back the return code. This includes COMMIT, SKIP, PRUNE and an
1556          uncaptured THEN, which means they take their normal effect. This
1557          consistent approach does not always have exactly the same effect as in
1558          Perl. */
1559    
1560          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1561        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1562        }        }
1563      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);   /* Continue for next alternative */
1564    
1565      if (*ecode == OP_KET) MRRETURN(MATCH_NOMATCH);      /* If we have tried all the alternative branches, the assertion has
1566        failed. If not, we broke out after a match. */
1567    
1568        if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1569    
1570      /* If checking an assertion for a condition, return MATCH_MATCH. */      /* If checking an assertion for a condition, return MATCH_MATCH. */
1571    
1572      if (condassert) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);
1573    
1574      /* Continue from after the assertion, updating the offsets high water      /* Continue from after a successful assertion, updating the offsets high
1575      mark, since extracts may have been taken during the assertion. */      water mark, since extracts may have been taken during the assertion. */
1576    
1577      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1578      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1579      offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1580      continue;      continue;
1581    
1582      /* Negative assertion: all branches must fail to match. Encountering SKIP,      /* Negative assertion: all branches must fail to match for the assertion to
1583      PRUNE, or COMMIT means we must assume failure without checking subsequent      succeed. */
     branches. */  
1584    
1585      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1586      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1587        save_mark = md->mark;
1588      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1589        {        {
1590        condassert = TRUE;        condassert = TRUE;
# Line 1485  for (;;) Line 1592  for (;;)
1592        }        }
1593      else condassert = FALSE;      else condassert = FALSE;
1594    
1595        /* Loop for each alternative branch. */
1596    
1597      do      do
1598        {        {
1599        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1600        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) MRRETURN(MATCH_NOMATCH);        md->mark = save_mark;   /* Always restore the mark setting */
1601        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)  
1602          switch(rrc)
1603          {          {
1604          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          case MATCH_MATCH:            /* A successful match means */
1605            case MATCH_ACCEPT:           /* the assertion has failed. */
1606            RRETURN(MATCH_NOMATCH);
1607    
1608            case MATCH_NOMATCH:          /* Carry on with next branch */
1609          break;          break;
1610    
1611            /* See comment in the code for capturing groups above about handling
1612            THEN. */
1613    
1614            case MATCH_THEN:
1615            next = ecode + GET(ecode,1);
1616            if (md->start_match_ptr < next &&
1617                (*ecode == OP_ALT || *next == OP_ALT))
1618              {
1619              rrc = MATCH_NOMATCH;
1620              break;
1621              }
1622            /* Otherwise fall through. */
1623    
1624            /* COMMIT, SKIP, PRUNE, and an uncaptured THEN cause the whole
1625            assertion to fail to match, without considering any more alternatives.
1626            Failing to match means the assertion is true. This is a consistent
1627            approach, but does not always have the same effect as in Perl. */
1628    
1629            case MATCH_COMMIT:
1630            case MATCH_SKIP:
1631            case MATCH_SKIP_ARG:
1632            case MATCH_PRUNE:
1633            do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1634            goto NEG_ASSERT_TRUE;   /* Break out of alternation loop */
1635    
1636            /* Anything else is an error */
1637    
1638            default:
1639            RRETURN(rrc);
1640          }          }
1641    
1642        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* Continue with next branch */
       as NOMATCH. */  
1643    
       if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);  
1644        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1645        }        }
1646      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1647    
1648      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */      /* All branches in the assertion failed to match. */
1649    
1650      ecode += 1 + LINK_SIZE;      NEG_ASSERT_TRUE:
1651        if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1652        ecode += 1 + LINK_SIZE;                /* Continue with current branch */
1653      continue;      continue;
1654    
1655      /* Move the subject pointer back. This occurs only at the start of      /* Move the subject pointer back. This occurs only at the start of
# Line 1514  for (;;) Line 1658  for (;;)
1658      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1659    
1660      case OP_REVERSE:      case OP_REVERSE:
1661  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1662      if (utf8)      if (utf)
1663        {        {
1664        i = GET(ecode, 1);        i = GET(ecode, 1);
1665        while (i-- > 0)        while (i-- > 0)
1666          {          {
1667          eptr--;          eptr--;
1668          if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1669          BACKCHAR(eptr);          BACKCHAR(eptr);
1670          }          }
1671        }        }
# Line 1532  for (;;) Line 1676  for (;;)
1676    
1677        {        {
1678        eptr -= GET(ecode, 1);        eptr -= GET(ecode, 1);
1679        if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1680        }        }
1681    
1682      /* Save the earliest consulted character, then skip to next op code */      /* Save the earliest consulted character, then skip to next op code */
# Line 1546  for (;;) Line 1690  for (;;)
1690      function is able to force a failure. */      function is able to force a failure. */
1691    
1692      case OP_CALLOUT:      case OP_CALLOUT:
1693      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1694        {        {
1695        pcre_callout_block cb;        PUBL(callout_block) cb;
1696        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1697        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1698        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1699    #if defined COMPILE_PCRE8
1700        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1701    #elif defined COMPILE_PCRE16
1702          cb.subject          = (PCRE_SPTR16)md->start_subject;
1703    #elif defined COMPILE_PCRE32
1704          cb.subject          = (PCRE_SPTR32)md->start_subject;
1705    #endif
1706        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1707        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1708        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
1709        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1710        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1711        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1712        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1713          /* Internal change requires this for API compatibility. */
1714          if (cb.capture_last == 0) cb.capture_last = -1;
1715        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1716        cb.mark             = markptr;        cb.mark             = md->nomatch_mark;
1717        if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1718        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1719        }        }
1720      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1588  for (;;) Line 1740  for (;;)
1740      case OP_RECURSE:      case OP_RECURSE:
1741        {        {
1742        recursion_info *ri;        recursion_info *ri;
1743        int recno;        unsigned int recno;
1744    
1745        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1746        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1605  for (;;) Line 1757  for (;;)
1757        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1758    
1759        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1760          new_recursive.saved_capture_last = md->capture_last;
1761        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1762        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1763        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1621  for (;;) Line 1774  for (;;)
1774        else        else
1775          {          {
1776          new_recursive.offset_save =          new_recursive.offset_save =
1777            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1778          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1779          }          }
1780        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1781              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1782    
1783        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1784        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1785        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1786          */
1787    
1788        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1789        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
1790        do        do
1791          {          {
1792          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1793          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1794            md, eptrb, RM6);            md, eptrb, RM6);
1795          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1796              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1797            md->capture_last = new_recursive.saved_capture_last;
1798          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1799          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1800            {            {
1801            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
1802            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1803              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1804    
1805            /* Set where we got to in the subject, and reset the start in case            /* Set where we got to in the subject, and reset the start in case
1806            it was changed by \K. This *is* propagated back out of a recursion,            it was changed by \K. This *is* propagated back out of a recursion,
# Line 1656  for (;;) Line 1811  for (;;)
1811            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1812            }            }
1813    
1814          /* PCRE does not allow THEN to escape beyond a recursion; it is treated          /* PCRE does not allow THEN, SKIP, PRUNE or COMMIT to escape beyond a
1815          as NOMATCH. */          recursion; they cause a NOMATCH for the entire recursion. These codes
1816            are defined in a range that can be tested for. */
1817    
1818            if (rrc >= MATCH_BACKTRACK_MIN && rrc <= MATCH_BACKTRACK_MAX)
1819              RRETURN(MATCH_NOMATCH);
1820    
1821          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          /* Any return code other than NOMATCH is an error. */
1822    
1823            if (rrc != MATCH_NOMATCH)
1824            {            {
1825            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1826            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1827              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1828            RRETURN(rrc);            RRETURN(rrc);
1829            }            }
1830    
# Line 1675  for (;;) Line 1836  for (;;)
1836        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1837        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1838        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1839          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1840        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1841        }        }
1842    
1843      RECURSION_MATCHED:      RECURSION_MATCHED:
# Line 1745  for (;;) Line 1906  for (;;)
1906        }        }
1907      else saved_eptr = NULL;      else saved_eptr = NULL;
1908    
1909      /* If we are at the end of an assertion group, stop matching and return      /* If we are at the end of an assertion group or a non-capturing atomic
1910      MATCH_MATCH, but record the current high water mark for use by positive      group, stop matching and return MATCH_MATCH, but record the current high
1911      assertions. We also need to record the match start in case it was changed      water mark for use by positive assertions. We also need to record the match
1912      by \K. */      start in case it was changed by \K. */
1913    
1914      if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||      if ((*prev >= OP_ASSERT && *prev <= OP_ASSERTBACK_NOT) ||
1915          *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT)           *prev == OP_ONCE_NC)
1916        {        {
1917        md->end_match_ptr = eptr;      /* For ONCE */        md->end_match_ptr = eptr;      /* For ONCE_NC */
1918        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
1919        md->start_match_ptr = mstart;        md->start_match_ptr = mstart;
1920        MRRETURN(MATCH_MATCH);         /* Sets md->mark */        RRETURN(MATCH_MATCH);         /* Sets md->mark */
1921        }        }
1922    
1923      /* 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
# Line 1789  for (;;) Line 1950  for (;;)
1950    
1951        /* Deal with capturing */        /* Deal with capturing */
1952    
1953        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
1954        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1955          {          {
1956          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
1957          "skipping" a capturing group, and that group's offsets must be marked          "skipping" a capturing group, and that group's offsets must be marked
# Line 1821  for (;;) Line 1982  for (;;)
1982      /* For an ordinary non-repeating ket, just continue at this level. This      /* For an ordinary non-repeating ket, just continue at this level. This
1983      also happens for a repeating ket if no characters were matched in the      also happens for a repeating ket if no characters were matched in the
1984      group. This is the forcible breaking of infinite loops as implemented in      group. This is the forcible breaking of infinite loops as implemented in
1985      Perl 5.005. For a non-repeating atomic group, establish a backup point by      Perl 5.005. For a non-repeating atomic group that includes captures,
1986      processing the rest of the pattern at a lower level. If this results in a      establish a backup point by processing the rest of the pattern at a lower
1987      NOMATCH return, pass MATCH_ONCE back to the original OP_ONCE level, thereby      level. If this results in a NOMATCH return, pass MATCH_ONCE back to the
1988      bypassing intermediate backup points, but resetting any captures that      original OP_ONCE level, thereby bypassing intermediate backup points, but
1989      happened along the way. */      resetting any captures that happened along the way. */
1990    
1991      if (*ecode == OP_KET || eptr == saved_eptr)      if (*ecode == OP_KET || eptr == saved_eptr)
1992        {        {
# Line 1870  for (;;) Line 2031  for (;;)
2031          }          }
2032        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2033          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2034          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2035          RRETURN(rrc);          RRETURN(rrc);
2036          }          }
# Line 1879  for (;;) Line 2039  for (;;)
2039        }        }
2040      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2041        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2042        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2043        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2044        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 1898  for (;;) Line 2057  for (;;)
2057      /* Not multiline mode: start of subject assertion, unless notbol. */      /* Not multiline mode: start of subject assertion, unless notbol. */
2058    
2059      case OP_CIRC:      case OP_CIRC:
2060      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
2061    
2062      /* Start of subject assertion */      /* Start of subject assertion */
2063    
2064      case OP_SOD:      case OP_SOD:
2065      if (eptr != md->start_subject) MRRETURN(MATCH_NOMATCH);      if (eptr != md->start_subject) RRETURN(MATCH_NOMATCH);
2066      ecode++;      ecode++;
2067      break;      break;
2068    
2069      /* Multiline mode: start of subject unless notbol, or after any newline. */      /* Multiline mode: start of subject unless notbol, or after any newline. */
2070    
2071      case OP_CIRCM:      case OP_CIRCM:
2072      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
2073      if (eptr != md->start_subject &&      if (eptr != md->start_subject &&
2074          (eptr == md->end_subject || !WAS_NEWLINE(eptr)))          (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
2075        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2076      ecode++;      ecode++;
2077      break;      break;
2078    
2079      /* Start of match assertion */      /* Start of match assertion */
2080    
2081      case OP_SOM:      case OP_SOM:
2082      if (eptr != md->start_subject + md->start_offset) MRRETURN(MATCH_NOMATCH);      if (eptr != md->start_subject + md->start_offset) RRETURN(MATCH_NOMATCH);
2083      ecode++;      ecode++;
2084      break;      break;
2085    
# Line 1936  for (;;) Line 2095  for (;;)
2095    
2096      case OP_DOLLM:      case OP_DOLLM:
2097      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2098        { if (!IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH); }        {
2099          if (!IS_NEWLINE(eptr))
2100            {
2101            if (md->partial != 0 &&
2102                eptr + 1 >= md->end_subject &&
2103                NLBLOCK->nltype == NLTYPE_FIXED &&
2104                NLBLOCK->nllen == 2 &&
2105                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2106              {
2107              md->hitend = TRUE;
2108              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2109              }
2110            RRETURN(MATCH_NOMATCH);
2111            }
2112          }
2113      else      else
2114        {        {
2115        if (md->noteol) MRRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
2116        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2117        }        }
2118      ecode++;      ecode++;
# Line 1949  for (;;) Line 2122  for (;;)
2122      subject unless noteol is set. */      subject unless noteol is set. */
2123    
2124      case OP_DOLL:      case OP_DOLL:
2125      if (md->noteol) MRRETURN(MATCH_NOMATCH);      if (md->noteol) RRETURN(MATCH_NOMATCH);
2126      if (!md->endonly) goto ASSERT_NL_OR_EOS;      if (!md->endonly) goto ASSERT_NL_OR_EOS;
2127    
2128      /* ... else fall through for endonly */      /* ... else fall through for endonly */
# Line 1957  for (;;) Line 2130  for (;;)
2130      /* End of subject assertion (\z) */      /* End of subject assertion (\z) */
2131    
2132      case OP_EOD:      case OP_EOD:
2133      if (eptr < md->end_subject) MRRETURN(MATCH_NOMATCH);      if (eptr < md->end_subject) RRETURN(MATCH_NOMATCH);
2134      SCHECK_PARTIAL();      SCHECK_PARTIAL();
2135      ecode++;      ecode++;
2136      break;      break;
# Line 1968  for (;;) Line 2141  for (;;)
2141      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2142      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2143          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2144        MRRETURN(MATCH_NOMATCH);        {
2145          if (md->partial != 0 &&
2146              eptr + 1 >= md->end_subject &&
2147              NLBLOCK->nltype == NLTYPE_FIXED &&
2148              NLBLOCK->nllen == 2 &&
2149              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2150            {
2151            md->hitend = TRUE;
2152            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2153            }
2154          RRETURN(MATCH_NOMATCH);
2155          }
2156    
2157      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2158    
# Line 1987  for (;;) Line 2171  for (;;)
2171        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2172        partial matching. */        partial matching. */
2173    
2174  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2175        if (utf8)        if (utf)
2176          {          {
2177          /* Get status of previous character */          /* Get status of previous character */
2178    
2179          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2180            {            {
2181            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2182            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2183            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2184            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2185  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2060  for (;;) Line 2244  for (;;)
2244              }              }
2245            else            else
2246  #endif  #endif
2247            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2248                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2249            }            }
2250    
2251          /* Get status of next character */          /* Get status of next character */
# Line 2083  for (;;) Line 2268  for (;;)
2268            }            }
2269          else          else
2270  #endif  #endif
2271          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2272              && ((md->ctypes[*eptr] & ctype_word) != 0);
2273          }          }
2274    
2275        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
2276    
2277        if ((*ecode++ == OP_WORD_BOUNDARY)?        if ((*ecode++ == OP_WORD_BOUNDARY)?
2278             cur_is_word == prev_is_word : cur_is_word != prev_is_word)             cur_is_word == prev_is_word : cur_is_word != prev_is_word)
2279          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2280        }        }
2281      break;      break;
2282    
2283      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2284        CRLF newlines and partial matching. */
2285    
2286      case OP_ANY:      case OP_ANY:
2287      if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2288        if (md->partial != 0 &&
2289            eptr + 1 >= md->end_subject &&
2290            NLBLOCK->nltype == NLTYPE_FIXED &&
2291            NLBLOCK->nllen == 2 &&
2292            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2293          {
2294          md->hitend = TRUE;
2295          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2296          }
2297    
2298      /* Fall through */      /* Fall through */
2299    
2300        /* Match any single character whatsoever. */
2301    
2302      case OP_ALLANY:      case OP_ALLANY:
2303      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2304        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
2305        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2306        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2307        }        }
2308      eptr++;      eptr++;
2309      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  #ifdef SUPPORT_UTF
2310        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2311    #endif
2312      ecode++;      ecode++;
2313      break;      break;
2314    
# Line 2118  for (;;) Line 2319  for (;;)
2319      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2320        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
2321        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2322        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2323        }        }
2324      eptr++;      eptr++;
2325      ecode++;      ecode++;
# Line 2128  for (;;) Line 2329  for (;;)
2329      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2330        {        {
2331        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2332        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2333        }        }
2334      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2335      if (      if (
2336  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2337         c < 256 &&         c < 256 &&
2338  #endif  #endif
2339         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
2340         )         )
2341        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2342      ecode++;      ecode++;
2343      break;      break;
2344    
# Line 2145  for (;;) Line 2346  for (;;)
2346      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2347        {        {
2348        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2349        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2350        }        }
2351      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2352      if (      if (
2353  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2354         c >= 256 ||         c > 255 ||
2355  #endif  #endif
2356         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2357         )         )
2358        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2359      ecode++;      ecode++;
2360      break;      break;
2361    
# Line 2162  for (;;) Line 2363  for (;;)
2363      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2364        {        {
2365        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2366        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2367        }        }
2368      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2369      if (      if (
2370  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2371         c < 256 &&         c < 256 &&
2372  #endif  #endif
2373         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
2374         )         )
2375        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2376      ecode++;      ecode++;
2377      break;      break;
2378    
# Line 2179  for (;;) Line 2380  for (;;)
2380      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2381        {        {
2382        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2383        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2384        }        }
2385      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2386      if (      if (
2387  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2388         c >= 256 ||         c > 255 ||
2389  #endif  #endif
2390         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2391         )         )
2392        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2393      ecode++;      ecode++;
2394      break;      break;
2395    
# Line 2196  for (;;) Line 2397  for (;;)
2397      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2398        {        {
2399        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2400        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2401        }        }
2402      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2403      if (      if (
2404  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2405         c < 256 &&         c < 256 &&
2406  #endif  #endif
2407         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
2408         )         )
2409        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2410      ecode++;      ecode++;
2411      break;      break;
2412    
# Line 2213  for (;;) Line 2414  for (;;)
2414      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2415        {        {
2416        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2417        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2418        }        }
2419      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2420      if (      if (
2421  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2422         c >= 256 ||         c > 255 ||
2423  #endif  #endif
2424         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2425         )         )
2426        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2427      ecode++;      ecode++;
2428      break;      break;
2429    
# Line 2230  for (;;) Line 2431  for (;;)
2431      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2432        {        {
2433        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2434        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2435        }        }
2436      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2437      switch(c)      switch(c)
2438        {        {
2439        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2440    
2441        case 0x000d:        case CHAR_CR:
2442        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2443            {
2444            SCHECK_PARTIAL();
2445            }
2446          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2447        break;        break;
2448    
2449        case 0x000a:        case CHAR_LF:
2450        break;        break;
2451    
2452        case 0x000b:        case CHAR_VT:
2453        case 0x000c:        case CHAR_FF:
2454        case 0x0085:        case CHAR_NEL:
2455    #ifndef EBCDIC
2456        case 0x2028:        case 0x2028:
2457        case 0x2029:        case 0x2029:
2458        if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);  #endif  /* Not EBCDIC */
2459          if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2460        break;        break;
2461        }        }
2462      ecode++;      ecode++;
# Line 2259  for (;;) Line 2466  for (;;)
2466      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2467        {        {
2468        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2469        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2470        }        }
2471      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2472      switch(c)      switch(c)
2473        {        {
2474          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2475        default: break;        default: break;
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       MRRETURN(MATCH_NOMATCH);  
2476        }        }
2477      ecode++;      ecode++;
2478      break;      break;
# Line 2293  for (;;) Line 2481  for (;;)
2481      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2482        {        {
2483        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2484        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2485        }        }
2486      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2487      switch(c)      switch(c)
2488        {        {
2489        default: MRRETURN(MATCH_NOMATCH);        HSPACE_CASES: break;  /* Byte and multibyte cases */
2490        case 0x09:      /* HT */        default: RRETURN(MATCH_NOMATCH);
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       break;  
2491        }        }
2492      ecode++;      ecode++;
2493      break;      break;
# Line 2327  for (;;) Line 2496  for (;;)
2496      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2497        {        {
2498        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2499        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2500        }        }
2501      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2502      switch(c)      switch(c)
2503        {        {
2504          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2505        default: break;        default: break;
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       MRRETURN(MATCH_NOMATCH);  
2506        }        }
2507      ecode++;      ecode++;
2508      break;      break;
# Line 2349  for (;;) Line 2511  for (;;)
2511      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2512        {        {
2513        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2514        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2515        }        }
2516      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2517      switch(c)      switch(c)
2518        {        {
2519        default: MRRETURN(MATCH_NOMATCH);        VSPACE_CASES: break;
2520        case 0x0a:      /* LF */        default: RRETURN(MATCH_NOMATCH);
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       break;  
2521        }        }
2522      ecode++;      ecode++;
2523      break;      break;
# Line 2376  for (;;) Line 2531  for (;;)
2531      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2532        {        {
2533        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2534        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2535        }        }
2536      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2537        {        {
2538          const pcre_uint32 *cp;
2539        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2540    
2541        switch(ecode[1])        switch(ecode[1])
2542          {          {
2543          case PT_ANY:          case PT_ANY:
2544          if (op == OP_NOTPROP) MRRETURN(MATCH_NOMATCH);          if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
2545          break;          break;
2546    
2547          case PT_LAMP:          case PT_LAMP:
2548          if ((prop->chartype == ucp_Lu ||          if ((prop->chartype == ucp_Lu ||
2549               prop->chartype == ucp_Ll ||               prop->chartype == ucp_Ll ||
2550               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))
2551            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2552          break;          break;
2553    
2554          case PT_GC:          case PT_GC:
2555          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2556            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2557          break;          break;
2558    
2559          case PT_PC:          case PT_PC:
2560          if ((ecode[2] != prop->chartype) == (op == OP_PROP))          if ((ecode[2] != prop->chartype) == (op == OP_PROP))
2561            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2562          break;          break;
2563    
2564          case PT_SC:          case PT_SC:
2565          if ((ecode[2] != prop->script) == (op == OP_PROP))          if ((ecode[2] != prop->script) == (op == OP_PROP))
2566            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2567          break;          break;
2568    
2569          /* These are specials */          /* These are specials */
2570    
2571          case PT_ALNUM:          case PT_ALNUM:
2572          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2573               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2574            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2575          break;          break;
2576    
2577          case PT_SPACE:    /* Perl space */          /* Perl space used to exclude VT, but from Perl 5.18 it is included,
2578          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          which means that Perl space and POSIX space are now identical. PCRE
2579               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)          was changed at release 8.34. */
                == (op == OP_NOTPROP))  
           MRRETURN(MATCH_NOMATCH);  
         break;  
2580    
2581            case PT_SPACE:    /* Perl space */
2582          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2583          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          switch(c)
2584               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||            {
2585               c == CHAR_FF || c == CHAR_CR)            HSPACE_CASES:
2586                 == (op == OP_NOTPROP))            VSPACE_CASES:
2587            MRRETURN(MATCH_NOMATCH);            if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
2588              break;
2589    
2590              default:
2591              if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z) ==
2592                (op == OP_NOTPROP)) RRETURN(MATCH_NOMATCH);
2593              break;
2594              }
2595          break;          break;
2596    
2597          case PT_WORD:          case PT_WORD:
2598          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2599               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2600               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2601            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2602            break;
2603    
2604            case PT_CLIST:
2605            cp = PRIV(ucd_caseless_sets) + ecode[2];
2606            for (;;)
2607              {
2608              if (c < *cp)
2609                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2610              if (c == *cp++)
2611                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2612              }
2613            break;
2614    
2615            case PT_UCNC:
2616            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
2617                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
2618                 c >= 0xe000) == (op == OP_NOTPROP))
2619              RRETURN(MATCH_NOMATCH);
2620          break;          break;
2621    
2622          /* This should never occur */          /* This should never occur */
# Line 2457  for (;;) Line 2636  for (;;)
2636      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2637        {        {
2638        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2639        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2640        }        }
2641      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2642        {        {
2643        int len = 1;        int lgb, rgb;
2644        if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2645        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2646        eptr += len;        while (eptr < md->end_subject)
2647            {
2648            int len = 1;
2649            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2650            rgb = UCD_GRAPHBREAK(c);
2651            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2652            lgb = rgb;
2653            eptr += len;
2654            }
2655        }        }
2656        CHECK_PARTIAL();
2657      ecode++;      ecode++;
2658      break;      break;
2659  #endif  #endif  /* SUPPORT_UCP */
2660    
2661    
2662      /* Match a back reference, possibly repeatedly. Look past the end of the      /* Match a back reference, possibly repeatedly. Look past the end of the
# Line 2479  for (;;) Line 2665  for (;;)
2665      similar code to character type repeats - written out again for speed.      similar code to character type repeats - written out again for speed.
2666      However, if the referenced string is the empty string, always treat      However, if the referenced string is the empty string, always treat
2667      it as matched, any number of times (otherwise there could be infinite      it as matched, any number of times (otherwise there could be infinite
2668      loops). */      loops). If the reference is unset, there are two possibilities:
   
     case OP_REF:  
     case OP_REFI:  
     caseless = op == OP_REFI;  
     offset = GET2(ecode, 1) << 1;               /* Doubled ref number */  
     ecode += 3;  
   
     /* If the reference is unset, there are two possibilities:  
2669    
2670      (a) In the default, Perl-compatible state, set the length negative;      (a) In the default, Perl-compatible state, set the length negative;
2671      this ensures that every attempt at a match fails. We can't just fail      this ensures that every attempt at a match fails. We can't just fail
# Line 2497  for (;;) Line 2675  for (;;)
2675      so that the back reference matches an empty string.      so that the back reference matches an empty string.
2676    
2677      Otherwise, set the length to the length of what was matched by the      Otherwise, set the length to the length of what was matched by the
2678      referenced subpattern. */      referenced subpattern.
2679    
2680        The OP_REF and OP_REFI opcodes are used for a reference to a numbered group
2681        or to a non-duplicated named group. For a duplicated named group, OP_DNREF
2682        and OP_DNREFI are used. In this case we must scan the list of groups to
2683        which the name refers, and use the first one that is set. */
2684    
2685        case OP_DNREF:
2686        case OP_DNREFI:
2687        caseless = op == OP_DNREFI;
2688          {
2689          int count = GET2(ecode, 1+IMM2_SIZE);
2690          pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
2691          ecode += 1 + 2*IMM2_SIZE;
2692    
2693          while (count-- > 0)
2694            {
2695            offset = GET2(slot, 0) << 1;
2696            if (offset < offset_top && md->offset_vector[offset] >= 0) break;
2697            slot += md->name_entry_size;
2698            }
2699          if (count < 0)
2700            length = (md->jscript_compat)? 0 : -1;
2701          else
2702            length = md->offset_vector[offset+1] - md->offset_vector[offset];
2703          }
2704        goto REF_REPEAT;
2705    
2706        case OP_REF:
2707        case OP_REFI:
2708        caseless = op == OP_REFI;
2709        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2710        ecode += 1 + IMM2_SIZE;
2711      if (offset >= offset_top || md->offset_vector[offset] < 0)      if (offset >= offset_top || md->offset_vector[offset] < 0)
2712        length = (md->jscript_compat)? 0 : -1;        length = (md->jscript_compat)? 0 : -1;
2713      else      else
# Line 2506  for (;;) Line 2715  for (;;)
2715    
2716      /* Set up for repetition, or handle the non-repeated case */      /* Set up for repetition, or handle the non-repeated case */
2717    
2718        REF_REPEAT:
2719      switch (*ecode)      switch (*ecode)
2720        {        {
2721        case OP_CRSTAR:        case OP_CRSTAR:
# Line 2525  for (;;) Line 2735  for (;;)
2735        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2736        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2737        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2738        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2739        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2740        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2741        break;        break;
2742    
2743        default:               /* No repeat follows */        default:               /* No repeat follows */
2744        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2745          {          {
2746            if (length == -2) eptr = md->end_subject;   /* Partial match */
2747          CHECK_PARTIAL();          CHECK_PARTIAL();
2748          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2749          }          }
2750        eptr += length;        eptr += length;
2751        continue;              /* With the main loop */        continue;              /* With the main loop */
2752        }        }
2753    
2754      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2755      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2756        means the reference is unset in non-Java-compatible mode. If the minimum is
2757        zero, we can continue at the same level without recursion. For any other
2758        minimum, carrying on will result in NOMATCH. */
2759    
2760      if (length == 0) continue;      if (length == 0) continue;
2761        if (length < 0 && min == 0) continue;
2762    
2763      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2764      the length of the reference string explicitly rather than passing the      the length of the reference string explicitly rather than passing the
# Line 2554  for (;;) Line 2769  for (;;)
2769        int slength;        int slength;
2770        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2771          {          {
2772            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2773          CHECK_PARTIAL();          CHECK_PARTIAL();
2774          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2775          }          }
2776        eptr += slength;        eptr += slength;
2777        }        }
# Line 2574  for (;;) Line 2790  for (;;)
2790          int slength;          int slength;
2791          RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);
2792          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2793          if (fi >= max) MRRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2794          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2795            {            {
2796              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2797            CHECK_PARTIAL();            CHECK_PARTIAL();
2798            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2799            }            }
2800          eptr += slength;          eptr += slength;
2801          }          }
# Line 2595  for (;;) Line 2812  for (;;)
2812          int slength;          int slength;
2813          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2814            {            {
2815            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2816              the soft partial matching case. */
2817    
2818              if (slength == -2 && md->partial != 0 &&
2819                  md->end_subject > md->start_used_ptr)
2820                {
2821                md->hitend = TRUE;
2822                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2823                }
2824            break;            break;
2825            }            }
2826          eptr += slength;          eptr += slength;
2827          }          }
2828    
2829        while (eptr >= pp)        while (eptr >= pp)
2830          {          {
2831          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
2832          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2833          eptr -= length;          eptr -= length;
2834          }          }
2835        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2836        }        }
2837      /* Control never gets here */      /* Control never gets here */
2838    
# Line 2624  for (;;) Line 2850  for (;;)
2850      case OP_NCLASS:      case OP_NCLASS:
2851      case OP_CLASS:      case OP_CLASS:
2852        {        {
2853          /* The data variable is saved across frames, so the byte map needs to
2854          be stored there. */
2855    #define BYTE_MAP ((pcre_uint8 *)data)
2856        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2857        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2858    
2859        switch (*ecode)        switch (*ecode)
2860          {          {
# Line 2646  for (;;) Line 2875  for (;;)
2875          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2876          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2877          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2878          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2879          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2880          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2881          break;          break;
2882    
2883          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2658  for (;;) Line 2887  for (;;)
2887    
2888        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2889    
2890  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2891        /* UTF-8 mode */        if (utf)
       if (utf8)  
2892          {          {
2893          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2894            {            {
2895            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2896              {              {
2897              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2898              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2899              }              }
2900            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
2901            if (c > 255)            if (c > 255)
2902              {              {
2903              if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2904              }              }
2905            else            else
2906              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
             }  
2907            }            }
2908          }          }
2909        else        else
2910  #endif  #endif
2911        /* Not UTF-8 mode */        /* Not UTF mode */
2912          {          {
2913          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2914            {            {
2915            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2916              {              {
2917              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2918              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2919              }              }
2920            c = *eptr++;            c = *eptr++;
2921            if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2922              if (c > 255)
2923                {
2924                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2925                }
2926              else
2927    #endif
2928                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2929            }            }
2930          }          }
2931    
# Line 2706  for (;;) Line 2939  for (;;)
2939    
2940        if (minimize)        if (minimize)
2941          {          {
2942  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2943          /* UTF-8 mode */          if (utf)
         if (utf8)  
2944            {            {
2945            for (fi = min;; fi++)            for (fi = min;; fi++)
2946              {              {
2947              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);
2948              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2949              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2950              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
2951                {                {
2952                SCHECK_PARTIAL();                SCHECK_PARTIAL();
2953                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2954                }                }
2955              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
2956              if (c > 255)              if (c > 255)
2957                {                {
2958                if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2959                }                }
2960              else              else
2961                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
               }  
2962              }              }
2963            }            }
2964          else          else
2965  #endif  #endif
2966          /* Not UTF-8 mode */          /* Not UTF mode */
2967            {            {
2968            for (fi = min;; fi++)            for (fi = min;; fi++)
2969              {              {
2970              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);
2971              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2972              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2973              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
2974                {                {
2975                SCHECK_PARTIAL();                SCHECK_PARTIAL();
2976                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2977                }                }
2978              c = *eptr++;              c = *eptr++;
2979              if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2980                if (c > 255)
2981                  {
2982                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2983                  }
2984                else
2985    #endif
2986                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2987              }              }
2988            }            }
2989          /* Control never gets here */          /* Control never gets here */
# Line 2758  for (;;) Line 2995  for (;;)
2995          {          {
2996          pp = eptr;          pp = eptr;
2997    
2998  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2999          /* UTF-8 mode */          if (utf)
         if (utf8)  
3000            {            {
3001            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3002              {              {
# Line 2776  for (;;) Line 3012  for (;;)
3012                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
3013                }                }
3014              else              else
3015                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
3016              eptr += len;              eptr += len;
3017              }              }
3018            for (;;)            for (;;)
# Line 2791  for (;;) Line 3025  for (;;)
3025            }            }
3026          else          else
3027  #endif  #endif
3028            /* Not UTF-8 mode */            /* Not UTF mode */
3029            {            {
3030            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3031              {              {
# Line 2801  for (;;) Line 3035  for (;;)
3035                break;                break;
3036                }                }
3037              c = *eptr;              c = *eptr;
3038              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
3039                if (c > 255)
3040                  {
3041                  if (op == OP_CLASS) break;
3042                  }
3043                else
3044    #endif
3045                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3046              eptr++;              eptr++;
3047              }              }
3048            while (eptr >= pp)            while (eptr >= pp)
# Line 2812  for (;;) Line 3053  for (;;)
3053              }              }
3054            }            }
3055    
3056          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3057          }          }
3058    #undef BYTE_MAP
3059        }        }
3060      /* Control never gets here */      /* Control never gets here */
3061    
# Line 2822  for (;;) Line 3064  for (;;)
3064      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
3065      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
3066    
3067  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3068      case OP_XCLASS:      case OP_XCLASS:
3069        {        {
3070        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2847  for (;;) Line 3089  for (;;)
3089          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3090          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3091          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3092          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
3093          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
3094          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
3095          break;          break;
3096    
3097          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2864  for (;;) Line 3106  for (;;)
3106          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3107            {            {
3108            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3109            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3110            }            }
3111          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
3112          if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3113          }          }
3114    
3115        /* 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 2884  for (;;) Line 3126  for (;;)
3126            {            {
3127            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);
3128            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3129            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3130            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3131              {              {
3132              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3133              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3134              }              }
3135            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3136            if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3137            }            }
3138          /* Control never gets here */          /* Control never gets here */
3139          }          }
# Line 2909  for (;;) Line 3151  for (;;)
3151              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3152              break;              break;
3153              }              }
3154    #ifdef SUPPORT_UTF
3155            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3156            if (!_pcre_xclass(c, data)) break;  #else
3157              c = *eptr;
3158    #endif
3159              if (!PRIV(xclass)(c, data, utf)) break;
3160            eptr += len;            eptr += len;
3161            }            }
3162          for(;;)          for(;;)
# Line 2918  for (;;) Line 3164  for (;;)
3164            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3165            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3166            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3167            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3168              if (utf) BACKCHAR(eptr);
3169    #endif
3170            }            }
3171          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3172          }          }
3173    
3174        /* Control never gets here */        /* Control never gets here */
# Line 2930  for (;;) Line 3178  for (;;)
3178      /* Match a single character, casefully */      /* Match a single character, casefully */
3179    
3180      case OP_CHAR:      case OP_CHAR:
3181  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3182      if (utf8)      if (utf)
3183        {        {
3184        length = 1;        length = 1;
3185        ecode++;        ecode++;
# Line 2939  for (;;) Line 3187  for (;;)
3187        if (length > md->end_subject - eptr)        if (length > md->end_subject - eptr)
3188          {          {
3189          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3190          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3191          }          }
3192        while (length-- > 0) if (*ecode++ != *eptr++) MRRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3193        }        }
3194      else      else
3195  #endif  #endif
3196        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3197        {        {
3198        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3199          {          {
3200          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
3201          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3202          }          }
3203        if (ecode[1] != *eptr++) MRRETURN(MATCH_NOMATCH);        if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH);
3204        ecode += 2;        ecode += 2;
3205        }        }
3206      break;      break;
3207    
3208      /* Match a single character, caselessly */      /* Match a single character, caselessly. If we are at the end of the
3209        subject, give up immediately. */
3210    
3211      case OP_CHARI:      case OP_CHARI:
3212  #ifdef SUPPORT_UTF8      if (eptr >= md->end_subject)
3213      if (utf8)        {
3214          SCHECK_PARTIAL();
3215          RRETURN(MATCH_NOMATCH);
3216          }
3217    
3218    #ifdef SUPPORT_UTF
3219        if (utf)
3220        {        {
3221        length = 1;        length = 1;
3222        ecode++;        ecode++;
3223        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
3224    
       if (length > md->end_subject - eptr)  
         {  
         CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */  
         MRRETURN(MATCH_NOMATCH);  
         }  
   
3225        /* If the pattern character's value is < 128, we have only one byte, and        /* If the pattern character's value is < 128, we have only one byte, and
3226        can use the fast lookup table. */        we know that its other case must also be one byte long, so we can use the
3227          fast lookup table. We know that there is at least one byte left in the
3228          subject. */
3229    
3230        if (fc < 128)        if (fc < 128)
3231          {          {
3232          if (md->lcc[*ecode++] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);          pcre_uint32 cc = RAWUCHAR(eptr);
3233            if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3234            ecode++;
3235            eptr++;
3236          }          }
3237    
3238        /* Otherwise we must pick up the subject character */        /* Otherwise we must pick up the subject character. Note that we cannot
3239          use the value of "length" to check for sufficient bytes left, because the
3240          other case of the character may have more or fewer bytes.  */
3241    
3242        else        else
3243          {          {
3244          unsigned int dc;          pcre_uint32 dc;
3245          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3246          ecode += length;          ecode += length;
3247    
# Line 2998  for (;;) Line 3253  for (;;)
3253  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3254            if (dc != UCD_OTHERCASE(fc))            if (dc != UCD_OTHERCASE(fc))
3255  #endif  #endif
3256              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3257            }            }
3258          }          }
3259        }        }
3260      else      else
3261  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3262    
3263      /* Non-UTF-8 mode */      /* Not UTF mode */
3264        {        {
3265        if (md->end_subject - eptr < 1)        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3266          {            != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3267          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */        eptr++;
         MRRETURN(MATCH_NOMATCH);  
         }  
       if (md->lcc[ecode[1]] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);  
3268        ecode += 2;        ecode += 2;
3269        }        }
3270      break;      break;
# Line 3022  for (;;) Line 3274  for (;;)
3274      case OP_EXACT:      case OP_EXACT:
3275      case OP_EXACTI:      case OP_EXACTI:
3276      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3277      ecode += 3;      ecode += 1 + IMM2_SIZE;
3278      goto REPEATCHAR;      goto REPEATCHAR;
3279    
3280      case OP_POSUPTO:      case OP_POSUPTO:
# Line 3037  for (;;) Line 3289  for (;;)
3289      min = 0;      min = 0;
3290      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3291      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3292      ecode += 3;      ecode += 1 + IMM2_SIZE;
3293      goto REPEATCHAR;      goto REPEATCHAR;
3294    
3295      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3082  for (;;) Line 3334  for (;;)
3334      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
3335      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
3336    
3337      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. We first check
3338        for the minimum number of characters. If the minimum equals the maximum, we
3339        are done. Otherwise, if minimizing, check the rest of the pattern for a
3340        match; if there isn't one, advance up to the maximum, one character at a
3341        time.
3342    
3343        If maximizing, advance up to the maximum number of matching characters,
3344        until eptr is past the end of the maximum run. If possessive, we are
3345        then done (no backing up). Otherwise, match at this position; anything
3346        other than no match is immediately returned. For nomatch, back up one
3347        character, unless we are matching \R and the last thing matched was
3348        \r\n, in which case, back up two bytes. When we reach the first optional
3349        character position, we can save stack by doing a tail recurse.
3350    
3351        The various UTF/non-UTF and caseful/caseless cases are handled separately,
3352        for speed. */
3353    
3354      REPEATCHAR:      REPEATCHAR:
3355  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3356      if (utf8)      if (utf)
3357        {        {
3358        length = 1;        length = 1;
3359        charptr = ecode;        charptr = ecode;
# Line 3099  for (;;) Line 3366  for (;;)
3366        if (length > 1)        if (length > 1)
3367          {          {
3368  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3369          unsigned int othercase;          pcre_uint32 othercase;
3370          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3371              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3372            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3373          else oclength = 0;          else oclength = 0;
3374  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3375    
3376          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3377            {            {
3378            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3379              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3380  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3381            else if (oclength > 0 &&            else if (oclength > 0 &&
3382                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3383                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3384  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3385            else            else
3386              {              {
3387              CHECK_PARTIAL();              CHECK_PARTIAL();
3388              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3389              }              }
3390            }            }
3391    
# Line 3130  for (;;) Line 3397  for (;;)
3397              {              {
3398              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);
3399              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3400              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3401              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3402                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3403  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3404              else if (oclength > 0 &&              else if (oclength > 0 &&
3405                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3406                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3407  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3408              else              else
3409                {                {
3410                CHECK_PARTIAL();                CHECK_PARTIAL();
3411                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3412                }                }
3413              }              }
3414            /* Control never gets here */            /* Control never gets here */
# Line 3153  for (;;) Line 3420  for (;;)
3420            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3421              {              {
3422              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3423                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3424  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3425              else if (oclength > 0 &&              else if (oclength > 0 &&
3426                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3427                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3428  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3429              else              else
3430                {                {
# Line 3166  for (;;) Line 3433  for (;;)
3433                }                }
3434              }              }
3435    
3436            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
   
3437            for(;;)            for(;;)
3438              {              {
3439                if (eptr == pp) goto TAIL_RECURSE;
3440              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3441              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
             if (eptr == pp) { MRRETURN(MATCH_NOMATCH); }  
3442  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3443              eptr--;              eptr--;
3444              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3189  for (;;) Line 3455  for (;;)
3455        value of fc will always be < 128. */        value of fc will always be < 128. */
3456        }        }
3457      else      else
3458  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3459          /* When not in UTF-8 mode, load a single-byte character. */
3460          fc = *ecode++;
3461    
3462      /* When not in UTF-8 mode, load a single-byte character. */      /* The value of fc at this point is always one character, though we may
3463        or may not be in UTF mode. The code is duplicated for the caseless and
     fc = *ecode++;  
   
     /* 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  
3464      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3465      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3466      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3205  for (;;) Line 3469  for (;;)
3469      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3470    
3471      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3472        max, eptr));        max, (char *)eptr));
3473    
3474      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3475        {        {
3476        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3477          /* fc must be < 128 if UTF is enabled. */
3478          foc = md->fcc[fc];
3479    #else
3480    #ifdef SUPPORT_UTF
3481    #ifdef SUPPORT_UCP
3482          if (utf && fc > 127)
3483            foc = UCD_OTHERCASE(fc);
3484    #else
3485          if (utf && fc > 127)
3486            foc = fc;
3487    #endif /* SUPPORT_UCP */
3488          else
3489    #endif /* SUPPORT_UTF */
3490            foc = TABLE_GET(fc, md->fcc, fc);
3491    #endif /* COMPILE_PCRE8 */
3492    
3493        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3494          {          {
3495            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3496          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3497            {            {
3498            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3499            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3500            }            }
3501          if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3502            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3503            eptr++;
3504          }          }
3505        if (min == max) continue;        if (min == max) continue;
3506        if (minimize)        if (minimize)
3507          {          {
3508          for (fi = min;; fi++)          for (fi = min;; fi++)
3509            {            {
3510              pcre_uint32 cc;               /* Faster than pcre_uchar */
3511            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3512            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3513            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3514            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3515              {              {
3516              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3517              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3518              }              }
3519            if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3520              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3521              eptr++;
3522            }            }
3523          /* Control never gets here */          /* Control never gets here */
3524          }          }
# Line 3241  for (;;) Line 3527  for (;;)
3527          pp = eptr;          pp = eptr;
3528          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3529            {            {
3530              pcre_uint32 cc;               /* Faster than pcre_uchar */
3531            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3532              {              {
3533              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3534              break;              break;
3535              }              }
3536            if (fc != md->lcc[*eptr]) break;            cc = RAWUCHARTEST(eptr);
3537              if (fc != cc && foc != cc) break;
3538            eptr++;            eptr++;
3539            }            }
3540            if (possessive) continue;       /* No backtracking */
3541          if (possessive) continue;          for (;;)
   
         while (eptr >= pp)  
3542            {            {
3543              if (eptr == pp) goto TAIL_RECURSE;
3544            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);
3545            eptr--;            eptr--;
3546            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3547            }            }
3548          MRRETURN(MATCH_NOMATCH);          /* Control never gets here */
3549          }          }
       /* Control never gets here */  
3550        }        }
3551    
3552      /* Caseful comparisons (includes all multi-byte characters) */      /* Caseful comparisons (includes all multi-byte characters) */
# Line 3272  for (;;) Line 3558  for (;;)
3558          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3559            {            {
3560            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3561            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3562            }            }
3563          if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3564          }          }
3565    
3566        if (min == max) continue;        if (min == max) continue;
# Line 3285  for (;;) Line 3571  for (;;)
3571            {            {
3572            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);
3573            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3574            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3575            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3576              {              {
3577              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3578              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3579              }              }
3580            if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3581            }            }
3582          /* Control never gets here */          /* Control never gets here */
3583          }          }
# Line 3305  for (;;) Line 3591  for (;;)
3591              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3592              break;              break;
3593              }              }
3594            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3595            eptr++;            eptr++;
3596            }            }
3597          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
3598            for (;;)
         while (eptr >= pp)  
3599            {            {
3600              if (eptr == pp) goto TAIL_RECURSE;
3601            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);
3602            eptr--;            eptr--;
3603            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3604            }            }
3605          MRRETURN(MATCH_NOMATCH);          /* Control never gets here */
3606          }          }
3607        }        }
3608      /* Control never gets here */      /* Control never gets here */
# Line 3329  for (;;) Line 3615  for (;;)
3615      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
3616        {        {
3617        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3618        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3619        }        }
3620      ecode++;  #ifdef SUPPORT_UTF
3621      GETCHARINCTEST(c, eptr);      if (utf)
     if (op == OP_NOTI)         /* The caseless case */  
3622        {        {
3623  #ifdef SUPPORT_UTF8        register pcre_uint32 ch, och;
3624        if (c < 256)  
3625  #endif        ecode++;
3626        c = md->lcc[c];        GETCHARINC(ch, ecode);
3627        if (md->lcc[*ecode++] == c) MRRETURN(MATCH_NOMATCH);        GETCHARINC(c, eptr);
3628    
3629          if (op == OP_NOT)
3630            {
3631            if (ch == c) RRETURN(MATCH_NOMATCH);
3632            }
3633          else
3634            {
3635    #ifdef SUPPORT_UCP
3636            if (ch > 127)
3637              och = UCD_OTHERCASE(ch);
3638    #else
3639            if (ch > 127)
3640              och = ch;
3641    #endif /* SUPPORT_UCP */
3642            else
3643              och = TABLE_GET(ch, md->fcc, ch);
3644            if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3645            }
3646        }        }
3647      else    /* Caseful */      else
3648    #endif
3649        {        {
3650        if (*ecode++ == c) MRRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3651          c = *eptr++;
3652          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3653            RRETURN(MATCH_NOMATCH);
3654          ecode += 2;
3655        }        }
3656      break;      break;
3657    
# Line 3357  for (;;) Line 3665  for (;;)
3665      case OP_NOTEXACT:      case OP_NOTEXACT:
3666      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3667      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3668      ecode += 3;      ecode += 1 + IMM2_SIZE;
3669      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3670    
3671      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3367  for (;;) Line 3675  for (;;)
3675      min = 0;      min = 0;
3676      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3677      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3678      ecode += 3;      ecode += 1 + IMM2_SIZE;
3679      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3680    
3681      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3399  for (;;) Line 3707  for (;;)
3707      possessive = TRUE;      possessive = TRUE;
3708      min = 0;      min = 0;
3709      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3710      ecode += 3;      ecode += 1 + IMM2_SIZE;
3711      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3712    
3713      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3423  for (;;) Line 3731  for (;;)
3731      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3732    
3733      REPEATNOTCHAR:      REPEATNOTCHAR:
3734      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3735    
3736      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3737      since matching characters is likely to be quite common. First, ensure the      since matching characters is likely to be quite common. First, ensure the
# Line 3434  for (;;) Line 3742  for (;;)
3742      characters and work backwards. */      characters and work backwards. */
3743    
3744      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,
3745        max, eptr));        max, (char *)eptr));
3746    
3747      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3748        {        {
3749        fc = md->lcc[fc];  #ifdef SUPPORT_UTF
3750    #ifdef SUPPORT_UCP
3751          if (utf && fc > 127)
3752            foc = UCD_OTHERCASE(fc);
3753    #else
3754          if (utf && fc > 127)
3755            foc = fc;
3756    #endif /* SUPPORT_UCP */
3757          else
3758    #endif /* SUPPORT_UTF */
3759            foc = TABLE_GET(fc, md->fcc, fc);
3760    
3761  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3762        /* UTF-8 mode */        if (utf)
       if (utf8)  
3763          {          {
3764          register unsigned int d;          register pcre_uint32 d;
3765          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3766            {            {
3767            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3768              {              {
3769              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3770              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3771              }              }
3772            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3773            if (d < 256) d = md->lcc[d];            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) MRRETURN(MATCH_NOMATCH);  
3774            }            }
3775          }          }
3776        else        else
3777  #endif  #endif  /* SUPPORT_UTF */
3778          /* Not UTF mode */
       /* Not UTF-8 mode */  
3779          {          {
3780          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3781            {            {
3782            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3783              {              {
3784              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3785              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3786              }              }
3787            if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3788              eptr++;
3789            }            }
3790          }          }
3791    
# Line 3477  for (;;) Line 3793  for (;;)
3793    
3794        if (minimize)        if (minimize)
3795          {          {
3796  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3797          /* UTF-8 mode */          if (utf)
         if (utf8)  
3798            {            {
3799            register unsigned int d;            register pcre_uint32 d;
3800            for (fi = min;; fi++)            for (fi = min;; fi++)
3801              {              {
3802              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
3803              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3804              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3805              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3806                {                {
3807                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3808                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3809                }                }
3810              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3811              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) MRRETURN(MATCH_NOMATCH);  
3812              }              }
3813            }            }
3814          else          else
3815  #endif  #endif  /*SUPPORT_UTF */
3816          /* Not UTF-8 mode */          /* Not UTF mode */
3817            {            {
3818            for (fi = min;; fi++)            for (fi = min;; fi++)
3819              {              {
3820              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);
3821              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3822              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3823              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3824                {                {
3825                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3826                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3827                }                }
3828              if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3829                eptr++;
3830              }              }
3831            }            }
3832          /* Control never gets here */          /* Control never gets here */
# Line 3523  for (;;) Line 3838  for (;;)
3838          {          {
3839          pp = eptr;          pp = eptr;
3840    
3841  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3842          /* UTF-8 mode */          if (utf)
         if (utf8)  
3843            {            {
3844            register unsigned int d;            register pcre_uint32 d;
3845            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3846              {              {
3847              int len = 1;              int len = 1;
# Line 3537  for (;;) Line 3851  for (;;)
3851                break;                break;
3852                }                }
3853              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3854              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) break;
             if (fc == d) break;  
3855              eptr += len;              eptr += len;
3856              }              }
3857          if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3858          for(;;)            for(;;)
3859              {              {
3860                if (eptr == pp) goto TAIL_RECURSE;
3861              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3862              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3863              if (eptr-- == pp) break;        /* Stop if tried at original pos */              eptr--;
3864              BACKCHAR(eptr);              BACKCHAR(eptr);
3865              }              }
3866            }            }
3867          else          else
3868  #endif  #endif  /* SUPPORT_UTF */
3869          /* Not UTF-8 mode */          /* Not UTF mode */
3870            {            {
3871            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3872              {              {
# Line 3561  for (;;) Line 3875  for (;;)
3875                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3876                break;                break;
3877                }                }
3878              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3879              eptr++;              eptr++;
3880              }              }
3881            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3882            while (eptr >= pp)            for (;;)
3883              {              {
3884                if (eptr == pp) goto TAIL_RECURSE;
3885              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);
3886              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3887              eptr--;              eptr--;
3888              }              }
3889            }            }
3890            /* Control never gets here */
         MRRETURN(MATCH_NOMATCH);  
3891          }          }
       /* Control never gets here */  
3892        }        }
3893    
3894      /* Caseful comparisons */      /* Caseful comparisons */
3895    
3896      else      else
3897        {        {
3898  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3899        /* UTF-8 mode */        if (utf)
       if (utf8)  
3900          {          {
3901          register unsigned int d;          register pcre_uint32 d;
3902          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3903            {            {
3904            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3905              {              {
3906              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3907              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3908              }              }
3909            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3910            if (fc == d) MRRETURN(MATCH_NOMATCH);            if (fc == d) RRETURN(MATCH_NOMATCH);
3911            }            }
3912          }          }
3913        else        else
3914  #endif  #endif
3915        /* Not UTF-8 mode */        /* Not UTF mode */
3916          {          {
3917          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3918            {            {
3919            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3920              {              {
3921              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3922              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3923              }              }
3924            if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);            if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
3925            }            }
3926          }          }
3927    
# Line 3617  for (;;) Line 3929  for (;;)
3929    
3930        if (minimize)        if (minimize)
3931          {          {
3932  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3933          /* UTF-8 mode */          if (utf)
         if (utf8)  
3934            {            {
3935            register unsigned int d;            register pcre_uint32 d;
3936            for (fi = min;; fi++)            for (fi = min;; fi++)
3937              {              {
3938              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
3939              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3940              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3941              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3942                {                {
3943                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3944                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3945                }                }
3946              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3947              if (fc == d) MRRETURN(MATCH_NOMATCH);              if (fc == d) RRETURN(MATCH_NOMATCH);
3948              }              }
3949            }            }
3950          else          else
3951  #endif  #endif
3952          /* Not UTF-8 mode */          /* Not UTF mode */
3953            {            {
3954            for (fi = min;; fi++)            for (fi = min;; fi++)
3955              {              {
3956              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);
3957              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3958              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3959              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3960                {                {
3961                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3962                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3963                }                }
3964              if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);              if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
3965              }              }
3966            }            }
3967          /* Control never gets here */          /* Control never gets here */
# Line 3662  for (;;) Line 3973  for (;;)
3973          {          {
3974          pp = eptr;          pp = eptr;
3975    
3976  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3977          /* UTF-8 mode */          if (utf)
         if (utf8)  
3978            {            {
3979            register unsigned int d;            register pcre_uint32 d;
3980            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3981              {              {
3982              int len = 1;              int len = 1;
# Line 3679  for (;;) Line 3989  for (;;)
3989              if (fc == d) break;              if (fc == d) break;
3990              eptr += len;              eptr += len;
3991              }              }
3992            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3993            for(;;)            for(;;)
3994              {              {
3995                if (eptr == pp) goto TAIL_RECURSE;
3996              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);
3997              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3998              if (eptr-- == pp) break;        /* Stop if tried at original pos */              eptr--;
3999              BACKCHAR(eptr);              BACKCHAR(eptr);
4000              }              }
4001            }            }
4002          else          else
4003  #endif  #endif
4004          /* Not UTF-8 mode */          /* Not UTF mode */
4005            {            {
4006            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4007              {              {
# Line 3702  for (;;) Line 4013  for (;;)
4013              if (fc == *eptr) break;              if (fc == *eptr) break;
4014              eptr++;              eptr++;
4015              }              }
4016            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
4017            while (eptr >= pp)            for (;;)
4018              {              {
4019                if (eptr == pp) goto TAIL_RECURSE;
4020              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);
4021              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4022              eptr--;              eptr--;
4023              }              }
4024            }            }
4025            /* Control never gets here */
         MRRETURN(MATCH_NOMATCH);  
4026          }          }
4027        }        }
4028      /* Control never gets here */      /* Control never gets here */
# Line 3723  for (;;) Line 4034  for (;;)
4034      case OP_TYPEEXACT:      case OP_TYPEEXACT:
4035      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
4036      minimize = TRUE;      minimize = TRUE;
4037      ecode += 3;      ecode += 1 + IMM2_SIZE;
4038      goto REPEATTYPE;      goto REPEATTYPE;
4039    
4040      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3731  for (;;) Line 4042  for (;;)
4042      min = 0;      min = 0;
4043      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4044      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
4045      ecode += 3;      ecode += 1 + IMM2_SIZE;
4046      goto REPEATTYPE;      goto REPEATTYPE;
4047    
4048      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3759  for (;;) Line 4070  for (;;)
4070      possessive = TRUE;      possessive = TRUE;
4071      min = 0;      min = 0;
4072      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4073      ecode += 3;      ecode += 1 + IMM2_SIZE;
4074      goto REPEATTYPE;      goto REPEATTYPE;
4075    
4076      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 3805  for (;;) Line 4116  for (;;)
4116          switch(prop_type)          switch(prop_type)
4117            {            {
4118            case PT_ANY:            case PT_ANY:
4119            if (prop_fail_result) MRRETURN(MATCH_NOMATCH);            if (prop_fail_result) RRETURN(MATCH_NOMATCH);
4120            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4121              {              {
4122              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4123                {                {
4124                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4125                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4126                }                }
4127              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4128              }              }
# Line 3824  for (;;) Line 4135  for (;;)
4135              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4136                {                {
4137                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4138                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4139                }                }
4140              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4141              chartype = UCD_CHARTYPE(c);              chartype = UCD_CHARTYPE(c);
4142              if ((chartype == ucp_Lu ||              if ((chartype == ucp_Lu ||
4143                   chartype == ucp_Ll ||                   chartype == ucp_Ll ||
4144                   chartype == ucp_Lt) == prop_fail_result)                   chartype == ucp_Lt) == prop_fail_result)
4145                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4146              }              }
4147            break;            break;
4148    
# Line 3841  for (;;) Line 4152  for (;;)
4152              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4153                {                {
4154                SCHECK_PARTIAL();       &nb