/[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 87 by nigel, Sat Feb 24 21:41:21 2007 UTC revision 366 by ph10, Mon Jul 14 15:45:32 2008 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) 1997-2006 University of Cambridge             Copyright (c) 1997-2008 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 42  POSSIBILITY OF SUCH DAMAGE. Line 42  POSSIBILITY OF SUCH DAMAGE.
42  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
43  possible. There are also some static supporting functions. */  possible. There are also some static supporting functions. */
44    
45    #ifdef HAVE_CONFIG_H
46    #include "config.h"
47    #endif
48    
49    #define NLBLOCK md             /* Block containing newline information */
50    #define PSSTART start_subject  /* Field containing processed string start */
51    #define PSEND   end_subject    /* Field containing processed string end */
52    
53  #include "pcre_internal.h"  #include "pcre_internal.h"
54    
55    /* Undefine some potentially clashing cpp symbols */
56    
57  /* Structure for building a chain of data that actually lives on the  #undef min
58  stack, for holding the values of the subject pointer at the start of each  #undef max
 subpattern, so as to detect when an empty string has been matched by a  
 subpattern - to break infinite loops. When NO_RECURSE is set, these blocks  
 are on the heap, not on the stack. */  
   
 typedef struct eptrblock {  
   struct eptrblock *epb_prev;  
   USPTR epb_saved_eptr;  
 } eptrblock;  
59    
60  /* Flag bits for the match() function */  /* Flag bits for the match() function */
61    
62  #define match_condassert   0x01    /* Called to check a condition assertion */  #define match_condassert     0x01  /* Called to check a condition assertion */
63  #define match_isgroup      0x02    /* Set if start of bracketed group */  #define match_cbegroup       0x02  /* Could-be-empty unlimited repeat group */
64    
65  /* Non-error returns from the match() function. Error returns are externally  /* Non-error returns from the match() function. Error returns are externally
66  defined PCRE_ERROR_xxx codes, which are all negative. */  defined PCRE_ERROR_xxx codes, which are all negative. */
# Line 68  defined PCRE_ERROR_xxx codes, which are Line 68  defined PCRE_ERROR_xxx codes, which are
68  #define MATCH_MATCH        1  #define MATCH_MATCH        1
69  #define MATCH_NOMATCH      0  #define MATCH_NOMATCH      0
70    
71    /* Special internal returns from the match() function. Make them sufficiently
72    negative to avoid the external error codes. */
73    
74    #define MATCH_COMMIT       (-999)
75    #define MATCH_PRUNE        (-998)
76    #define MATCH_SKIP         (-997)
77    #define MATCH_THEN         (-996)
78    
79  /* 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.
80  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,
81  because the offset vector is always a multiple of 3 long. */  because the offset vector is always a multiple of 3 long. */
# Line 101  Returns:     nothing Line 109  Returns:     nothing
109  static void  static void
110  pchars(const uschar *p, int length, BOOL is_subject, match_data *md)  pchars(const uschar *p, int length, BOOL is_subject, match_data *md)
111  {  {
112  int c;  unsigned int c;
113  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;
114  while (length-- > 0)  while (length-- > 0)
115    if (isprint(c = *(p++))) printf("%c", c); else printf("\\x%02x", c);    if (isprint(c = *(p++))) printf("%c", c); else printf("\\x%02x", c);
# Line 150  printf("\n"); Line 158  printf("\n");
158    
159  if (length > md->end_subject - eptr) return FALSE;  if (length > md->end_subject - eptr) return FALSE;
160    
161  /* Separate the caselesss case for speed */  /* Separate the caseless case for speed. In UTF-8 mode we can only do this
162    properly if Unicode properties are supported. Otherwise, we can check only
163    ASCII characters. */
164    
165  if ((ims & PCRE_CASELESS) != 0)  if ((ims & PCRE_CASELESS) != 0)
166    {    {
167    #ifdef SUPPORT_UTF8
168    #ifdef SUPPORT_UCP
169      if (md->utf8)
170        {
171        USPTR endptr = eptr + length;
172        while (eptr < endptr)
173          {
174          int c, d;
175          GETCHARINC(c, eptr);
176          GETCHARINC(d, p);
177          if (c != d && c != UCD_OTHERCASE(d)) return FALSE;
178          }
179        }
180      else
181    #endif
182    #endif
183    
184      /* The same code works when not in UTF-8 mode and in UTF-8 mode when there
185      is no UCP support. */
186    
187    while (length-- > 0)    while (length-- > 0)
188      if (md->lcc[*p++] != md->lcc[*eptr++]) return FALSE;      { if (md->lcc[*p++] != md->lcc[*eptr++]) return FALSE; }
189    }    }
190    
191    /* In the caseful case, we can just compare the bytes, whether or not we
192    are in UTF-8 mode. */
193    
194  else  else
195    { while (length-- > 0) if (*p++ != *eptr++) return FALSE; }    { while (length-- > 0) if (*p++ != *eptr++) return FALSE; }
196    
# Line 186  calls by keeping local variables that ne Line 220  calls by keeping local variables that ne
220  obtained from malloc() instead instead of on the stack. Macros are used to  obtained from malloc() instead instead of on the stack. Macros are used to
221  achieve this so that the actual code doesn't look very different to what it  achieve this so that the actual code doesn't look very different to what it
222  always used to.  always used to.
223    
224    The original heap-recursive code used longjmp(). However, it seems that this
225    can be very slow on some operating systems. Following a suggestion from Stan
226    Switzer, the use of longjmp() has been abolished, at the cost of having to
227    provide a unique number for each call to RMATCH. There is no way of generating
228    a sequence of numbers at compile time in C. I have given them names, to make
229    them stand out more clearly.
230    
231    Crude tests on x86 Linux show a small speedup of around 5-8%. However, on
232    FreeBSD, avoiding longjmp() more than halves the time taken to run the standard
233    tests. Furthermore, not using longjmp() means that local dynamic variables
234    don't have indeterminate values; this has meant that the frame size can be
235    reduced because the result can be "passed back" by straight setting of the
236    variable instead of being passed in the frame.
237  ****************************************************************************  ****************************************************************************
238  ***************************************************************************/  ***************************************************************************/
239    
240    /* Numbers for RMATCH calls. When this list is changed, the code at HEAP_RETURN
241    below must be updated in sync.  */
242    
243    enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM6,  RM7,  RM8,  RM9,  RM10,
244           RM11,  RM12, RM13, RM14, RM15, RM16, RM17, RM18, RM19, RM20,
245           RM21,  RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30,
246           RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
247           RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
248           RM51,  RM52, RM53, RM54 };
249    
250  /* 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
251  versions and production versions. */  versions and production versions. Note that the "rw" argument of RMATCH isn't
252    actuall used in this definition. */
253    
254  #ifndef NO_RECURSE  #ifndef NO_RECURSE
255  #define REGISTER register  #define REGISTER register
256    
257  #ifdef DEBUG  #ifdef DEBUG
258  #define RMATCH(rx,ra,rb,rc,rd,re,rf,rg) \  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \
259    { \    { \
260    printf("match() called in line %d\n", __LINE__); \    printf("match() called in line %d\n", __LINE__); \
261    rx = match(ra,rb,rc,rd,re,rf,rg,rdepth+1); \    rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1); \
262    printf("to line %d\n", __LINE__); \    printf("to line %d\n", __LINE__); \
263    }    }
264  #define RRETURN(ra) \  #define RRETURN(ra) \
# Line 208  versions and production versions. */ Line 267  versions and production versions. */
267    return ra; \    return ra; \
268    }    }
269  #else  #else
270  #define RMATCH(rx,ra,rb,rc,rd,re,rf,rg) \  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \
271    rx = match(ra,rb,rc,rd,re,rf,rg,rdepth+1)    rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1)
272  #define RRETURN(ra) return ra  #define RRETURN(ra) return ra
273  #endif  #endif
274    
275  #else  #else
276    
277    
278  /* These versions of the macros manage a private stack on the heap. Note  /* These versions of the macros manage a private stack on the heap. Note that
279  that the rd argument of RMATCH isn't actually used. It's the md argument of  the "rd" argument of RMATCH isn't actually used in this definition. It's the md
280  match(), which never changes. */  argument of match(), which never changes. */
281    
282  #define REGISTER  #define REGISTER
283    
284  #define RMATCH(rx,ra,rb,rc,rd,re,rf,rg)\  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw)\
285    {\    {\
286    heapframe *newframe = (pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = (pcre_stack_malloc)(sizeof(heapframe));\
287    if (setjmp(frame->Xwhere) == 0)\    frame->Xwhere = rw; \
288      {\    newframe->Xeptr = ra;\
289      newframe->Xeptr = ra;\    newframe->Xecode = rb;\
290      newframe->Xecode = rb;\    newframe->Xmstart = mstart;\
291      newframe->Xoffset_top = rc;\    newframe->Xoffset_top = rc;\
292      newframe->Xims = re;\    newframe->Xims = re;\
293      newframe->Xeptrb = rf;\    newframe->Xeptrb = rf;\
294      newframe->Xflags = rg;\    newframe->Xflags = rg;\
295      newframe->Xrdepth = frame->Xrdepth + 1;\    newframe->Xrdepth = frame->Xrdepth + 1;\
296      newframe->Xprevframe = frame;\    newframe->Xprevframe = frame;\
297      frame = newframe;\    frame = newframe;\
298      DPRINTF(("restarting from line %d\n", __LINE__));\    DPRINTF(("restarting from line %d\n", __LINE__));\
299      goto HEAP_RECURSE;\    goto HEAP_RECURSE;\
300      }\    L_##rw:\
301    else\    DPRINTF(("jumped back to line %d\n", __LINE__));\
     {\  
     DPRINTF(("longjumped back to line %d\n", __LINE__));\  
     frame = md->thisframe;\  
     rx = frame->Xresult;\  
     }\  
302    }    }
303    
304  #define RRETURN(ra)\  #define RRETURN(ra)\
# Line 254  match(), which never changes. */ Line 308  match(), which never changes. */
308    (pcre_stack_free)(newframe);\    (pcre_stack_free)(newframe);\
309    if (frame != NULL)\    if (frame != NULL)\
310      {\      {\
311      frame->Xresult = ra;\      rrc = ra;\
312      md->thisframe = frame;\      goto HEAP_RETURN;\
     longjmp(frame->Xwhere, 1);\  
313      }\      }\
314    return ra;\    return ra;\
315    }    }
# Line 271  typedef struct heapframe { Line 324  typedef struct heapframe {
324    
325    const uschar *Xeptr;    const uschar *Xeptr;
326    const uschar *Xecode;    const uschar *Xecode;
327      const uschar *Xmstart;
328    int Xoffset_top;    int Xoffset_top;
329    long int Xims;    long int Xims;
330    eptrblock *Xeptrb;    eptrblock *Xeptrb;
331    int Xflags;    int Xflags;
332    int Xrdepth;    unsigned int Xrdepth;
333    
334    /* Function local variables */    /* Function local variables */
335    
# Line 291  typedef struct heapframe { Line 345  typedef struct heapframe {
345    
346    BOOL Xcur_is_word;    BOOL Xcur_is_word;
347    BOOL Xcondition;    BOOL Xcondition;
   BOOL Xminimize;  
348    BOOL Xprev_is_word;    BOOL Xprev_is_word;
349    
350    unsigned long int Xoriginal_ims;    unsigned long int Xoriginal_ims;
# Line 303  typedef struct heapframe { Line 356  typedef struct heapframe {
356    int Xprop_category;    int Xprop_category;
357    int Xprop_chartype;    int Xprop_chartype;
358    int Xprop_script;    int Xprop_script;
359    int *Xprop_test_variable;    int Xoclength;
360      uschar Xocchars[8];
361  #endif  #endif
362    
363    int Xctype;    int Xctype;
364    int Xfc;    unsigned int Xfc;
365    int Xfi;    int Xfi;
366    int Xlength;    int Xlength;
367    int Xmax;    int Xmax;
# Line 321  typedef struct heapframe { Line 375  typedef struct heapframe {
375    
376    eptrblock Xnewptrb;    eptrblock Xnewptrb;
377    
378    /* Place to pass back result, and where to jump back to */    /* Where to jump back to */
379    
380    int  Xresult;    int Xwhere;
   jmp_buf Xwhere;  
381    
382  } heapframe;  } heapframe;
383    
# Line 340  typedef struct heapframe { Line 393  typedef struct heapframe {
393  *         Match from current position            *  *         Match from current position            *
394  *************************************************/  *************************************************/
395    
396  /* On entry ecode points to the first opcode, and eptr to the first character  /* This function is called recursively in many circumstances. Whenever it
 in the subject string, while eptrb holds the value of eptr at the start of the  
 last bracketed group - used for breaking infinite loops matching zero-length  
 strings. This function is called recursively in many circumstances. Whenever it  
397  returns a negative (error) response, the outer incarnation must also return the  returns a negative (error) response, the outer incarnation must also return the
398  same response.  same response.
399    
# Line 353  performance. Tests using gcc on a SPARC Line 403  performance. Tests using gcc on a SPARC
403  made performance worse.  made performance worse.
404    
405  Arguments:  Arguments:
406     eptr        pointer in subject     eptr        pointer to current character in subject
407     ecode       position in code     ecode       pointer to current position in compiled code
408       mstart      pointer to the current match start position (can be modified
409                     by encountering \K)
410     offset_top  current top pointer     offset_top  current top pointer
411     md          pointer to "static" info for the match     md          pointer to "static" info for the match
412     ims         current /i, /m, and /s options     ims         current /i, /m, and /s options
# Line 362  Arguments: Line 414  Arguments:
414                   brackets - for testing for empty matches                   brackets - for testing for empty matches
415     flags       can contain     flags       can contain
416                   match_condassert - this is an assertion condition                   match_condassert - this is an assertion condition
417                   match_isgroup - this is the start of a bracketed group                   match_cbegroup - this is the start of an unlimited repeat
418                       group that can match an empty string
419     rdepth      the recursion depth     rdepth      the recursion depth
420    
421  Returns:       MATCH_MATCH if matched            )  these values are >= 0  Returns:       MATCH_MATCH if matched            )  these values are >= 0
# Line 372  Returns:       MATCH_MATCH if matched Line 425  Returns:       MATCH_MATCH if matched
425  */  */
426    
427  static int  static int
428  match(REGISTER USPTR eptr, REGISTER const uschar *ecode,  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, const uschar *mstart,
429    int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb,    int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb,
430    int flags, int rdepth)    int flags, unsigned int rdepth)
431  {  {
432  /* 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,
433  so they can be ordinary variables in all cases. Mark them with "register"  so they can be ordinary variables in all cases. Mark some of them with
434  because they are used a lot in loops. */  "register" because they are used a lot in loops. */
435    
436  register int  rrc;    /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
437  register int  i;      /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
438  register int  c;      /* Character values not kept over RMATCH() calls */  register unsigned int c;   /* Character values not kept over RMATCH() calls */
439  register BOOL utf8;   /* Local copy of UTF-8 flag for speed */  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */
440    
441    BOOL minimize, possessive; /* Quantifier options */
442    
443  /* 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
444  preserved over calls to RMATCH() are part of a "frame" which is obtained from  preserved over calls to RMATCH() are part of a "frame" which is obtained from
# Line 398  frame->Xprevframe = NULL;            /* Line 453  frame->Xprevframe = NULL;            /*
453    
454  frame->Xeptr = eptr;  frame->Xeptr = eptr;
455  frame->Xecode = ecode;  frame->Xecode = ecode;
456    frame->Xmstart = mstart;
457  frame->Xoffset_top = offset_top;  frame->Xoffset_top = offset_top;
458  frame->Xims = ims;  frame->Xims = ims;
459  frame->Xeptrb = eptrb;  frame->Xeptrb = eptrb;
# Line 412  HEAP_RECURSE: Line 468  HEAP_RECURSE:
468    
469  #define eptr               frame->Xeptr  #define eptr               frame->Xeptr
470  #define ecode              frame->Xecode  #define ecode              frame->Xecode
471    #define mstart             frame->Xmstart
472  #define offset_top         frame->Xoffset_top  #define offset_top         frame->Xoffset_top
473  #define ims                frame->Xims  #define ims                frame->Xims
474  #define eptrb              frame->Xeptrb  #define eptrb              frame->Xeptrb
# Line 434  HEAP_RECURSE: Line 491  HEAP_RECURSE:
491    
492  #define cur_is_word        frame->Xcur_is_word  #define cur_is_word        frame->Xcur_is_word
493  #define condition          frame->Xcondition  #define condition          frame->Xcondition
 #define minimize           frame->Xminimize  
494  #define prev_is_word       frame->Xprev_is_word  #define prev_is_word       frame->Xprev_is_word
495    
496  #define original_ims       frame->Xoriginal_ims  #define original_ims       frame->Xoriginal_ims
# Line 446  HEAP_RECURSE: Line 502  HEAP_RECURSE:
502  #define prop_category      frame->Xprop_category  #define prop_category      frame->Xprop_category
503  #define prop_chartype      frame->Xprop_chartype  #define prop_chartype      frame->Xprop_chartype
504  #define prop_script        frame->Xprop_script  #define prop_script        frame->Xprop_script
505  #define prop_test_variable frame->Xprop_test_variable  #define oclength           frame->Xoclength
506    #define occhars            frame->Xocchars
507  #endif  #endif
508    
509  #define ctype              frame->Xctype  #define ctype              frame->Xctype
# Line 470  HEAP_RECURSE: Line 527  HEAP_RECURSE:
527  get preserved during recursion in the normal way. In this environment, fi and  get preserved during recursion in the normal way. In this environment, fi and
528  i, and fc and c, can be the same variables. */  i, and fc and c, can be the same variables. */
529    
530  #else  #else         /* NO_RECURSE not defined */
531  #define fi i  #define fi i
532  #define fc c  #define fc c
533    
# Line 489  recursion_info new_recursive;      /* wi Line 546  recursion_info new_recursive;      /* wi
546                                     /* that do not have to be preserved over  */                                     /* that do not have to be preserved over  */
547  BOOL cur_is_word;                  /* a recursive call to RMATCH().          */  BOOL cur_is_word;                  /* a recursive call to RMATCH().          */
548  BOOL condition;  BOOL condition;
 BOOL minimize;  
549  BOOL prev_is_word;  BOOL prev_is_word;
550    
551  unsigned long int original_ims;  unsigned long int original_ims;
# Line 501  int prop_fail_result; Line 557  int prop_fail_result;
557  int prop_category;  int prop_category;
558  int prop_chartype;  int prop_chartype;
559  int prop_script;  int prop_script;
560  int *prop_test_variable;  int oclength;
561    uschar occhars[8];
562  #endif  #endif
563    
564  int ctype;  int ctype;
# Line 516  int save_offset1, save_offset2, save_off Line 573  int save_offset1, save_offset2, save_off
573  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
574    
575  eptrblock newptrb;  eptrblock newptrb;
576  #endif  #endif     /* NO_RECURSE */
577    
578  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
579  variables. */  variables. */
# Line 524  variables. */ Line 581  variables. */
581  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
582  prop_value = 0;  prop_value = 0;
583  prop_fail_result = 0;  prop_fail_result = 0;
 prop_test_variable = NULL;  
584  #endif  #endif
585    
586    
587    /* This label is used for tail recursion, which is used in a few cases even
588    when NO_RECURSE is not defined, in order to reduce the amount of stack that is
589    used. Thanks to Ian Taylor for noticing this possibility and sending the
590    original patch. */
591    
592    TAIL_RECURSE:
593    
594  /* OK, now we can get on with the real code of the function. Recursive calls  /* OK, now we can get on with the real code of the function. Recursive calls
595  are specified by the macro RMATCH and RRETURN is used to return. When  are specified by the macro RMATCH and RRETURN is used to return. When
596  NO_RECURSE is *not* defined, these just turn into a recursive call to match()  NO_RECURSE is *not* defined, these just turn into a recursive call to match()
# Line 535  defined). However, RMATCH isn't like a f Line 599  defined). However, RMATCH isn't like a f
599  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,
600  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
601    
602    #ifdef SUPPORT_UTF8
603    utf8 = md->utf8;       /* Local copy of the flag */
604    #else
605    utf8 = FALSE;
606    #endif
607    
608  /* 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
609  haven't exceeded the recursive call limit. */  haven't exceeded the recursive call limit. */
610    
# Line 542  if (md->match_call_count++ >= md->match_ Line 612  if (md->match_call_count++ >= md->match_
612  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);
613    
614  original_ims = ims;    /* Save for resetting on ')' */  original_ims = ims;    /* Save for resetting on ')' */
 utf8 = md->utf8;       /* Local copy of the flag */  
615    
616  /* At the start of a bracketed group, add the current subject pointer to the  /* At the start of a group with an unlimited repeat that may match an empty
617  stack of such pointers, to be re-instated at the end of the group when we hit  string, the match_cbegroup flag is set. When this is the case, add the current
618  the closing ket. When match() is called in other circumstances, we don't add to  subject pointer to the chain of such remembered pointers, to be checked when we
619  this stack. */  hit the closing ket, in order to break infinite loops that match no characters.
620    When match() is called in other circumstances, don't add to the chain. The
621    match_cbegroup flag must NOT be used with tail recursion, because the memory
622    block that is used is on the stack, so a new one may be required for each
623    match(). */
624    
625  if ((flags & match_isgroup) != 0)  if ((flags & match_cbegroup) != 0)
626    {    {
   newptrb.epb_prev = eptrb;  
627    newptrb.epb_saved_eptr = eptr;    newptrb.epb_saved_eptr = eptr;
628      newptrb.epb_prev = eptrb;
629    eptrb = &newptrb;    eptrb = &newptrb;
630    }    }
631    
632  /* Now start processing the operations. */  /* Now start processing the opcodes. */
633    
634  for (;;)  for (;;)
635    {    {
636      minimize = possessive = FALSE;
637    op = *ecode;    op = *ecode;
   minimize = FALSE;  
638    
639    /* For partial matching, remember if we ever hit the end of the subject after    /* For partial matching, remember if we ever hit the end of the subject after
640    matching at least one subject character. */    matching at least one subject character. */
641    
642    if (md->partial &&    if (md->partial &&
643        eptr >= md->end_subject &&        eptr >= md->end_subject &&
644        eptr > md->start_match)        eptr > mstart)
645      md->hitend = TRUE;      md->hitend = TRUE;
646    
647    /* Opening capturing bracket. If there is space in the offset vector, save    switch(op)
   the current subject position in the working slot at the top of the vector. We  
   mustn't change the current values of the data slot, because they may be set  
   from a previous iteration of this group, and be referred to by a reference  
   inside the group.  
   
   If the bracket fails to match, we need to restore this value and also the  
   values of the final offsets, in case they were set by a previous iteration of  
   the same bracket.  
   
   If there isn't enough space in the offset vector, treat this as if it were a  
   non-capturing bracket. Don't worry about setting the flag for the error case  
   here; that is handled in the code for KET. */  
   
   if (op > OP_BRA)  
648      {      {
649      number = op - OP_BRA;      case OP_FAIL:
650        RRETURN(MATCH_NOMATCH);
     /* For extended extraction brackets (large number), we have to fish out the  
     number from a dummy opcode at the start. */  
651    
652      if (number > EXTRACT_BASIC_MAX)      case OP_PRUNE:
653        number = GET2(ecode, 2+LINK_SIZE);      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
654          ims, eptrb, flags, RM51);
655        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
656        RRETURN(MATCH_PRUNE);
657    
658        case OP_COMMIT:
659        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
660          ims, eptrb, flags, RM52);
661        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
662        RRETURN(MATCH_COMMIT);
663    
664        case OP_SKIP:
665        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
666          ims, eptrb, flags, RM53);
667        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
668        md->start_match_ptr = eptr;   /* Pass back current position */
669        RRETURN(MATCH_SKIP);
670    
671        case OP_THEN:
672        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
673          ims, eptrb, flags, RM54);
674        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
675        RRETURN(MATCH_THEN);
676    
677        /* Handle a capturing bracket. If there is space in the offset vector, save
678        the current subject position in the working slot at the top of the vector.
679        We mustn't change the current values of the data slot, because they may be
680        set from a previous iteration of this group, and be referred to by a
681        reference inside the group.
682    
683        If the bracket fails to match, we need to restore this value and also the
684        values of the final offsets, in case they were set by a previous iteration
685        of the same bracket.
686    
687        If there isn't enough space in the offset vector, treat this as if it were
688        a non-capturing bracket. Don't worry about setting the flag for the error
689        case here; that is handled in the code for KET. */
690    
691        case OP_CBRA:
692        case OP_SCBRA:
693        number = GET2(ecode, 1+LINK_SIZE);
694      offset = number << 1;      offset = number << 1;
695    
696  #ifdef DEBUG  #ifdef DEBUG
697      printf("start bracket %d subject=", number);      printf("start bracket %d\n", number);
698        printf("subject=");
699      pchars(eptr, 16, TRUE, md);      pchars(eptr, 16, TRUE, md);
700      printf("\n");      printf("\n");
701  #endif  #endif
# Line 612  for (;;) Line 710  for (;;)
710        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
711        md->offset_vector[md->offset_end - number] = eptr - md->start_subject;        md->offset_vector[md->offset_end - number] = eptr - md->start_subject;
712    
713          flags = (op == OP_SCBRA)? match_cbegroup : 0;
714        do        do
715          {          {
716          RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb,          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
717            match_isgroup);            ims, eptrb, flags, RM1);
718          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
719          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
720          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
721          }          }
# Line 631  for (;;) Line 730  for (;;)
730        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
731        }        }
732    
733      /* Insufficient room for saving captured contents */      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
734        as a non-capturing bracket. */
735    
736      else op = OP_BRA;      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
737      }      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
738    
739    /* Other types of node can be handled by a switch */      DPRINTF(("insufficient capture room: treat as non-capturing\n"));
740    
741    switch(op)      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
742      {      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
743      case OP_BRA:     /* Non-capturing bracket: optimized */  
744      DPRINTF(("start bracket 0\n"));      /* Non-capturing bracket. Loop for all the alternatives. When we get to the
745      do      final alternative within the brackets, we would return the result of a
746        recursive call to match() whatever happened. We can reduce stack usage by
747        turning this into a tail recursion, except in the case when match_cbegroup
748        is set.*/
749    
750        case OP_BRA:
751        case OP_SBRA:
752        DPRINTF(("start non-capturing bracket\n"));
753        flags = (op >= OP_SBRA)? match_cbegroup : 0;
754        for (;;)
755        {        {
756        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb,        if (ecode[GET(ecode, 1)] != OP_ALT)   /* Final alternative */
757          match_isgroup);          {
758        if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (flags == 0)    /* Not a possibly empty group */
759              {
760              ecode += _pcre_OP_lengths[*ecode];
761              DPRINTF(("bracket 0 tail recursion\n"));
762              goto TAIL_RECURSE;
763              }
764    
765            /* Possibly empty group; can't use tail recursion. */
766    
767            RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
768              eptrb, flags, RM48);
769            RRETURN(rrc);
770            }
771    
772          /* For non-final alternatives, continue the loop for a NOMATCH result;
773          otherwise return. */
774    
775          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
776            eptrb, flags, RM2);
777          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
778        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
779        }        }
780      while (*ecode == OP_ALT);      /* Control never reaches here. */
     DPRINTF(("bracket 0 failed\n"));  
     RRETURN(MATCH_NOMATCH);  
781    
782      /* Conditional group: compilation checked that there are no more than      /* Conditional group: compilation checked that there are no more than
783      two branches. If the condition is false, skipping the first branch takes us      two branches. If the condition is false, skipping the first branch takes us
784      past the end if there is only one branch, but that's OK because that is      past the end if there is only one branch, but that's OK because that is
785      exactly what going to the ket would do. */      exactly what going to the ket would do. As there is only one branch to be
786        obeyed, we can use tail recursion to avoid using another stack frame. */
787    
788      case OP_COND:      case OP_COND:
789      if (ecode[LINK_SIZE+1] == OP_CREF) /* Condition extract or recurse test */      case OP_SCOND:
790        if (ecode[LINK_SIZE+1] == OP_RREF)         /* Recursion test */
791          {
792          offset = GET2(ecode, LINK_SIZE + 2);     /* Recursion group number*/
793          condition = md->recursive != NULL &&
794            (offset == RREF_ANY || offset == md->recursive->group_num);
795          ecode += condition? 3 : GET(ecode, 1);
796          }
797    
798        else if (ecode[LINK_SIZE+1] == OP_CREF)    /* Group used test */
799        {        {
800        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */
801        condition = (offset == CREF_RECURSE * 2)?        condition = offset < offset_top && md->offset_vector[offset] >= 0;
802          (md->recursive != NULL) :        ecode += condition? 3 : GET(ecode, 1);
803          (offset < offset_top && md->offset_vector[offset] >= 0);        }
804        RMATCH(rrc, eptr, ecode + (condition?  
805          (LINK_SIZE + 4) : (LINK_SIZE + 1 + GET(ecode, 1))),      else if (ecode[LINK_SIZE+1] == OP_DEF)     /* DEFINE - always false */
806          offset_top, md, ims, eptrb, match_isgroup);        {
807        RRETURN(rrc);        condition = FALSE;
808          ecode += GET(ecode, 1);
809        }        }
810    
811      /* The condition is an assertion. Call match() to evaluate it - setting      /* The condition is an assertion. Call match() to evaluate it - setting
812      the final argument TRUE causes it to stop at the end of an assertion. */      the final argument match_condassert causes it to stop at the end of an
813        assertion. */
814    
815      else      else
816        {        {
817        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,
818            match_condassert | match_isgroup);            match_condassert, RM3);
819        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
820          {          {
821          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE+2);          condition = TRUE;
822            ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
823          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
824          }          }
825        else if (rrc != MATCH_NOMATCH)        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
826          {          {
827          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
828          }          }
829        else ecode += GET(ecode, 1);        else
830        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb,          {
831          match_isgroup);          condition = FALSE;
832        RRETURN(rrc);          ecode += GET(ecode, 1);
833            }
834        }        }
     /* Control never reaches here */  
835    
836      /* Skip over conditional reference or large extraction number data if      /* We are now at the branch that is to be obeyed. As there is only one,
837      encountered. */      we can use tail recursion to avoid using another stack frame, except when
838        match_cbegroup is required for an unlimited repeat of a possibly empty
839        group. If the second alternative doesn't exist, we can just plough on. */
840    
841      case OP_CREF:      if (condition || *ecode == OP_ALT)
842      case OP_BRANUMBER:        {
843      ecode += 3;        ecode += 1 + LINK_SIZE;
844          if (op == OP_SCOND)        /* Possibly empty group */
845            {
846            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, match_cbegroup, RM49);
847            RRETURN(rrc);
848            }
849          else                       /* Group must match something */
850            {
851            flags = 0;
852            goto TAIL_RECURSE;
853            }
854          }
855        else                         /* Condition false & no 2nd alternative */
856          {
857          ecode += 1 + LINK_SIZE;
858          }
859      break;      break;
860    
     /* End of the pattern. If we are in a recursion, we should restore the  
     offsets appropriately and continue from after the call. */  
861    
862        /* End of the pattern, either real or forced. If we are in a top-level
863        recursion, we should restore the offsets appropriately and continue from
864        after the call. */
865    
866        case OP_ACCEPT:
867      case OP_END:      case OP_END:
868      if (md->recursive != NULL && md->recursive->group_num == 0)      if (md->recursive != NULL && md->recursive->group_num == 0)
869        {        {
# Line 713  for (;;) Line 872  for (;;)
872        md->recursive = rec->prevrec;        md->recursive = rec->prevrec;
873        memmove(md->offset_vector, rec->offset_save,        memmove(md->offset_vector, rec->offset_save,
874          rec->saved_max * sizeof(int));          rec->saved_max * sizeof(int));
875        md->start_match = rec->save_start;        mstart = rec->save_start;
876        ims = original_ims;        ims = original_ims;
877        ecode = rec->after_call;        ecode = rec->after_call;
878        break;        break;
# Line 722  for (;;) Line 881  for (;;)
881      /* Otherwise, if PCRE_NOTEMPTY is set, fail if we have matched an empty      /* Otherwise, if PCRE_NOTEMPTY is set, fail if we have matched an empty
882      string - backtracking will then try other alternatives, if any. */      string - backtracking will then try other alternatives, if any. */
883    
884      if (md->notempty && eptr == md->start_match) RRETURN(MATCH_NOMATCH);      if (md->notempty && eptr == mstart) RRETURN(MATCH_NOMATCH);
885      md->end_match_ptr = eptr;          /* Record where we ended */      md->end_match_ptr = eptr;           /* Record where we ended */
886      md->end_offset_top = offset_top;   /* and how many extracts were taken */      md->end_offset_top = offset_top;    /* and how many extracts were taken */
887        md->start_match_ptr = mstart;       /* and the start (\K can modify) */
888      RRETURN(MATCH_MATCH);      RRETURN(MATCH_MATCH);
889    
890      /* Change option settings */      /* Change option settings */
# Line 745  for (;;) Line 905  for (;;)
905      case OP_ASSERTBACK:      case OP_ASSERTBACK:
906      do      do
907        {        {
908        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
909          match_isgroup);          RM4);
910        if (rrc == MATCH_MATCH) break;        if (rrc == MATCH_MATCH) break;
911        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
912        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
913        }        }
914      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 772  for (;;) Line 932  for (;;)
932      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
933      do      do
934        {        {
935        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
936          match_isgroup);          RM5);
937        if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH);
938        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
939        ecode += GET(ecode,1);        ecode += GET(ecode,1);
940        }        }
941      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 794  for (;;) Line 954  for (;;)
954  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
955      if (utf8)      if (utf8)
956        {        {
957        c = GET(ecode,1);        i = GET(ecode, 1);
958        for (i = 0; i < c; i++)        while (i-- > 0)
959          {          {
960          eptr--;          eptr--;
961          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
962          BACKCHAR(eptr)          BACKCHAR(eptr);
963          }          }
964        }        }
965      else      else
# Line 808  for (;;) Line 968  for (;;)
968      /* No UTF-8 support, or not in UTF-8 mode: count is byte count */      /* No UTF-8 support, or not in UTF-8 mode: count is byte count */
969    
970        {        {
971        eptr -= GET(ecode,1);        eptr -= GET(ecode, 1);
972        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
973        }        }
974    
# Line 830  for (;;) Line 990  for (;;)
990        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
991        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
992        cb.subject_length   = md->end_subject - md->start_subject;        cb.subject_length   = md->end_subject - md->start_subject;
993        cb.start_match      = md->start_match - md->start_subject;        cb.start_match      = mstart - md->start_subject;
994        cb.current_position = eptr - md->start_subject;        cb.current_position = eptr - md->start_subject;
995        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
996        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
# Line 865  for (;;) Line 1025  for (;;)
1025      case OP_RECURSE:      case OP_RECURSE:
1026        {        {
1027        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1028        new_recursive.group_num = *callpat - OP_BRA;        new_recursive.group_num = (callpat == md->start_code)? 0 :
1029            GET2(callpat, 1 + LINK_SIZE);
       /* For extended extraction brackets (large number), we have to fish out  
       the number from a dummy opcode at the start. */  
   
       if (new_recursive.group_num > EXTRACT_BASIC_MAX)  
         new_recursive.group_num = GET2(callpat, 2+LINK_SIZE);  
1030    
1031        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1032    
# Line 897  for (;;) Line 1052  for (;;)
1052    
1053        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1054              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1055        new_recursive.save_start = md->start_match;        new_recursive.save_start = mstart;
1056        md->start_match = eptr;        mstart = eptr;
1057    
1058        /* OK, now we can do the recursion. For each top-level alternative we        /* OK, now we can do the recursion. For each top-level alternative we
1059        restore the offset and recursion data. */        restore the offset and recursion data. */
1060    
1061        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1062          flags = (*callpat >= OP_SBRA)? match_cbegroup : 0;
1063        do        do
1064          {          {
1065          RMATCH(rrc, eptr, callpat + 1 + LINK_SIZE, offset_top, md, ims,          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,
1066              eptrb, match_isgroup);            md, ims, eptrb, flags, RM6);
1067          if (rrc == MATCH_MATCH)          if (rrc == MATCH_MATCH)
1068            {            {
1069            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
# Line 916  for (;;) Line 1072  for (;;)
1072              (pcre_free)(new_recursive.offset_save);              (pcre_free)(new_recursive.offset_save);
1073            RRETURN(MATCH_MATCH);            RRETURN(MATCH_MATCH);
1074            }            }
1075          else if (rrc != MATCH_NOMATCH)          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1076            {            {
1077            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1078            RRETURN(rrc);            RRETURN(rrc);
# Line 945  for (;;) Line 1101  for (;;)
1101      the end of a normal bracket, leaving the subject pointer. */      the end of a normal bracket, leaving the subject pointer. */
1102    
1103      case OP_ONCE:      case OP_ONCE:
1104        {      prev = ecode;
1105        prev = ecode;      saved_eptr = eptr;
       saved_eptr = eptr;  
1106    
1107        do      do
1108          {        {
1109          RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM7);
1110            eptrb, match_isgroup);        if (rrc == MATCH_MATCH) break;
1111          if (rrc == MATCH_MATCH) break;        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1112          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        ecode += GET(ecode,1);
1113          ecode += GET(ecode,1);        }
1114          }      while (*ecode == OP_ALT);
       while (*ecode == OP_ALT);  
1115    
1116        /* If hit the end of the group (which could be repeated), fail */      /* If hit the end of the group (which could be repeated), fail */
1117    
1118        if (*ecode != OP_ONCE && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);      if (*ecode != OP_ONCE && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);
1119    
1120        /* Continue as from after the assertion, updating the offsets high water      /* Continue as from after the assertion, updating the offsets high water
1121        mark, since extracts may have been taken. */      mark, since extracts may have been taken. */
1122    
1123        do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode, 1); while (*ecode == OP_ALT);
1124    
1125        offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1126        eptr = md->end_match_ptr;      eptr = md->end_match_ptr;
1127    
1128        /* For a non-repeating ket, just continue at this level. This also      /* For a non-repeating ket, just continue at this level. This also
1129        happens for a repeating ket if no characters were matched in the group.      happens for a repeating ket if no characters were matched in the group.
1130        This is the forcible breaking of infinite loops as implemented in Perl      This is the forcible breaking of infinite loops as implemented in Perl
1131        5.005. If there is an options reset, it will get obeyed in the normal      5.005. If there is an options reset, it will get obeyed in the normal
1132        course of events. */      course of events. */
1133    
1134        if (*ecode == OP_KET || eptr == saved_eptr)      if (*ecode == OP_KET || eptr == saved_eptr)
1135          {        {
1136          ecode += 1+LINK_SIZE;        ecode += 1+LINK_SIZE;
1137          break;        break;
1138          }        }
1139    
1140        /* The repeating kets try the rest of the pattern or restart from the      /* The repeating kets try the rest of the pattern or restart from the
1141        preceding bracket, in the appropriate order. We need to reset any options      preceding bracket, in the appropriate order. The second "call" of match()
1142        that changed within the bracket before re-running it, so check the next      uses tail recursion, to avoid using another stack frame. We need to reset
1143        opcode. */      any options that changed within the bracket before re-running it, so
1144        check the next opcode. */
1145    
1146        if (ecode[1+LINK_SIZE] == OP_OPT)      if (ecode[1+LINK_SIZE] == OP_OPT)
1147          {        {
1148          ims = (ims & ~PCRE_IMS) | ecode[4];        ims = (ims & ~PCRE_IMS) | ecode[4];
1149          DPRINTF(("ims set to %02lx at group repeat\n", ims));        DPRINTF(("ims set to %02lx at group repeat\n", ims));
1150          }        }
1151    
1152        if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1153          {        {
1154          RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM8);
1155          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1156          RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);        ecode = prev;
1157          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        flags = 0;
1158          }        goto TAIL_RECURSE;
       else  /* OP_KETRMAX */  
         {  
         RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);  
         if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
         RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0);  
         if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
         }  
1159        }        }
1160      RRETURN(MATCH_NOMATCH);      else  /* OP_KETRMAX */
1161          {
1162          RMATCH(eptr, prev, offset_top, md, ims, eptrb, match_cbegroup, RM9);
1163          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1164          ecode += 1 + LINK_SIZE;
1165          flags = 0;
1166          goto TAIL_RECURSE;
1167          }
1168        /* Control never gets here */
1169    
1170      /* An alternation is the end of a branch; scan along to find the end of the      /* An alternation is the end of a branch; scan along to find the end of the
1171      bracketed group and go to there. */      bracketed group and go to there. */
# Line 1018  for (;;) Line 1174  for (;;)
1174      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1175      break;      break;
1176    
1177      /* BRAZERO and BRAMINZERO occur just before a bracket group, indicating      /* BRAZERO, BRAMINZERO and SKIPZERO occur just before a bracket group,
1178      that it may occur zero times. It may repeat infinitely, or not at all -      indicating that it may occur zero times. It may repeat infinitely, or not
1179      i.e. it could be ()* or ()? in the pattern. Brackets with fixed upper      at all - i.e. it could be ()* or ()? or even (){0} in the pattern. Brackets
1180      repeat limits are compiled as a number of copies, with the optional ones      with fixed upper repeat limits are compiled as a number of copies, with the
1181      preceded by BRAZERO or BRAMINZERO. */      optional ones preceded by BRAZERO or BRAMINZERO. */
1182    
1183      case OP_BRAZERO:      case OP_BRAZERO:
1184        {        {
1185        next = ecode+1;        next = ecode+1;
1186        RMATCH(rrc, eptr, next, offset_top, md, ims, eptrb, match_isgroup);        RMATCH(eptr, next, offset_top, md, ims, eptrb, 0, RM10);
1187        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1188        do next += GET(next,1); while (*next == OP_ALT);        do next += GET(next,1); while (*next == OP_ALT);
1189        ecode = next + 1+LINK_SIZE;        ecode = next + 1 + LINK_SIZE;
1190        }        }
1191      break;      break;
1192    
1193      case OP_BRAMINZERO:      case OP_BRAMINZERO:
1194        {        {
1195        next = ecode+1;        next = ecode+1;
1196        do next += GET(next,1); while (*next == OP_ALT);        do next += GET(next, 1); while (*next == OP_ALT);
1197        RMATCH(rrc, eptr, next + 1+LINK_SIZE, offset_top, md, ims, eptrb,        RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0, RM11);
         match_isgroup);  
1198        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1199        ecode++;        ecode++;
1200        }        }
1201      break;      break;
1202    
1203      /* End of a group, repeated or non-repeating. If we are at the end of      case OP_SKIPZERO:
1204      an assertion "group", stop matching and return MATCH_MATCH, but record the        {
1205      current high water mark for use by positive assertions. Do this also        next = ecode+1;
1206      for the "once" (not-backup up) groups. */        do next += GET(next,1); while (*next == OP_ALT);
1207          ecode = next + 1 + LINK_SIZE;
1208          }
1209        break;
1210    
1211        /* End of a group, repeated or non-repeating. */
1212    
1213      case OP_KET:      case OP_KET:
1214      case OP_KETRMIN:      case OP_KETRMIN:
1215      case OP_KETRMAX:      case OP_KETRMAX:
1216        {      prev = ecode - GET(ecode, 1);
       prev = ecode - GET(ecode, 1);  
       saved_eptr = eptrb->epb_saved_eptr;  
   
       /* Back up the stack of bracket start pointers. */  
1217    
1218        eptrb = eptrb->epb_prev;      /* If this was a group that remembered the subject start, in order to break
1219        infinite repeats of empty string matches, retrieve the subject start from
1220        the chain. Otherwise, set it NULL. */
1221    
1222        if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||      if (*prev >= OP_SBRA)
1223            *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||        {
1224            *prev == OP_ONCE)        saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */
1225          {        eptrb = eptrb->epb_prev;              /* Backup to previous group */
1226          md->end_match_ptr = eptr;      /* For ONCE */        }
1227          md->end_offset_top = offset_top;      else saved_eptr = NULL;
         RRETURN(MATCH_MATCH);  
         }  
1228    
1229        /* In all other cases except a conditional group we have to check the      /* If we are at the end of an assertion group, stop matching and return
1230        group number back at the start and if necessary complete handling an      MATCH_MATCH, but record the current high water mark for use by positive
1231        extraction by setting the offsets and bumping the high water mark. */      assertions. Do this also for the "once" (atomic) groups. */
1232    
1233        if (*prev != OP_COND)      if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||
1234          {          *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||
1235          number = *prev - OP_BRA;          *prev == OP_ONCE)
1236          {
1237          md->end_match_ptr = eptr;      /* For ONCE */
1238          md->end_offset_top = offset_top;
1239          RRETURN(MATCH_MATCH);
1240          }
1241    
1242          /* For extended extraction brackets (large number), we have to fish out      /* For capturing groups we have to check the group number back at the start
1243          the number from a dummy opcode at the start. */      and if necessary complete handling an extraction by setting the offsets and
1244        bumping the high water mark. Note that whole-pattern recursion is coded as
1245        a recurse into group 0, so it won't be picked up here. Instead, we catch it
1246        when the OP_END is reached. Other recursion is handled here. */
1247    
1248          if (number > EXTRACT_BASIC_MAX) number = GET2(prev, 2+LINK_SIZE);      if (*prev == OP_CBRA || *prev == OP_SCBRA)
1249          offset = number << 1;        {
1250          number = GET2(prev, 1+LINK_SIZE);
1251          offset = number << 1;
1252    
1253  #ifdef DEBUG  #ifdef DEBUG
1254          printf("end bracket %d", number);        printf("end bracket %d", number);
1255          printf("\n");        printf("\n");
1256  #endif  #endif
1257    
1258          /* Test for a numbered group. This includes groups called as a result        md->capture_last = number;
1259          of recursion. Note that whole-pattern recursion is coded as a recurse        if (offset >= md->offset_max) md->offset_overflow = TRUE; else
1260          into group 0, so it won't be picked up here. Instead, we catch it when          {
1261          the OP_END is reached. */          md->offset_vector[offset] =
1262              md->offset_vector[md->offset_end - number];
1263          if (number > 0)          md->offset_vector[offset+1] = eptr - md->start_subject;
1264            {          if (offset_top <= offset) offset_top = offset + 2;
1265            md->capture_last = number;          }
           if (offset >= md->offset_max) md->offset_overflow = TRUE; else  
             {  
             md->offset_vector[offset] =  
               md->offset_vector[md->offset_end - number];  
             md->offset_vector[offset+1] = eptr - md->start_subject;  
             if (offset_top <= offset) offset_top = offset + 2;  
             }  
1266    
1267            /* Handle a recursively called group. Restore the offsets        /* Handle a recursively called group. Restore the offsets
1268            appropriately and continue from after the call. */        appropriately and continue from after the call. */
1269    
1270            if (md->recursive != NULL && md->recursive->group_num == number)        if (md->recursive != NULL && md->recursive->group_num == number)
1271              {          {
1272              recursion_info *rec = md->recursive;          recursion_info *rec = md->recursive;
1273              DPRINTF(("Recursion (%d) succeeded - continuing\n", number));          DPRINTF(("Recursion (%d) succeeded - continuing\n", number));
1274              md->recursive = rec->prevrec;          md->recursive = rec->prevrec;
1275              md->start_match = rec->save_start;          mstart = rec->save_start;
1276              memcpy(md->offset_vector, rec->offset_save,          memcpy(md->offset_vector, rec->offset_save,
1277                rec->saved_max * sizeof(int));            rec->saved_max * sizeof(int));
1278              ecode = rec->after_call;          ecode = rec->after_call;
1279              ims = original_ims;          ims = original_ims;
1280              break;          break;
             }  
           }  
1281          }          }
1282          }
1283    
1284        /* Reset the value of the ims flags, in case they got changed during      /* For both capturing and non-capturing groups, reset the value of the ims
1285        the group. */      flags, in case they got changed during the group. */
1286    
1287        ims = original_ims;      ims = original_ims;
1288        DPRINTF(("ims reset to %02lx\n", ims));      DPRINTF(("ims reset to %02lx\n", ims));
1289    
1290        /* For a non-repeating ket, just continue at this level. This also      /* For a non-repeating ket, just continue at this level. This also
1291        happens for a repeating ket if no characters were matched in the group.      happens for a repeating ket if no characters were matched in the group.
1292        This is the forcible breaking of infinite loops as implemented in Perl      This is the forcible breaking of infinite loops as implemented in Perl
1293        5.005. If there is an options reset, it will get obeyed in the normal      5.005. If there is an options reset, it will get obeyed in the normal
1294        course of events. */      course of events. */
1295    
1296        if (*ecode == OP_KET || eptr == saved_eptr)      if (*ecode == OP_KET || eptr == saved_eptr)
1297          {        {
1298          ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1299          break;        break;
1300          }        }
1301    
1302        /* The repeating kets try the rest of the pattern or restart from the      /* The repeating kets try the rest of the pattern or restart from the
1303        preceding bracket, in the appropriate order. */      preceding bracket, in the appropriate order. In the second case, we can use
1304        tail recursion to avoid using another stack frame, unless we have an
1305        unlimited repeat of a group that can match an empty string. */
1306    
1307        if (*ecode == OP_KETRMIN)      flags = (*prev >= OP_SBRA)? match_cbegroup : 0;
1308          {  
1309          RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0);      if (*ecode == OP_KETRMIN)
1310          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        {
1311          RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM12);
1312          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1313          }        if (flags != 0)    /* Could match an empty string */
       else  /* OP_KETRMAX */  
1314          {          {
1315          RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);          RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM50);
1316          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          RRETURN(rrc);
         RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0);  
         if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
1317          }          }
1318          ecode = prev;
1319          goto TAIL_RECURSE;
1320        }        }
1321        else  /* OP_KETRMAX */
1322      RRETURN(MATCH_NOMATCH);        {
1323          RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13);
1324          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1325          ecode += 1 + LINK_SIZE;
1326          flags = 0;
1327          goto TAIL_RECURSE;
1328          }
1329        /* Control never gets here */
1330    
1331      /* Start of subject unless notbol, or after internal newline if multiline */      /* Start of subject unless notbol, or after internal newline if multiline */
1332    
# Line 1168  for (;;) Line 1334  for (;;)
1334      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
1335      if ((ims & PCRE_MULTILINE) != 0)      if ((ims & PCRE_MULTILINE) != 0)
1336        {        {
1337        if (eptr != md->start_subject && eptr[-1] != NEWLINE)        if (eptr != md->start_subject &&
1338              (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
1339          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
1340        ecode++;        ecode++;
1341        break;        break;
# Line 1189  for (;;) Line 1356  for (;;)
1356      ecode++;      ecode++;
1357      break;      break;
1358    
1359        /* Reset the start of match point */
1360    
1361        case OP_SET_SOM:
1362        mstart = eptr;
1363        ecode++;
1364        break;
1365    
1366      /* Assert before internal newline if multiline, or before a terminating      /* Assert before internal newline if multiline, or before a terminating
1367      newline unless endonly is set, else end of subject unless noteol is set. */      newline unless endonly is set, else end of subject unless noteol is set. */
1368    
# Line 1196  for (;;) Line 1370  for (;;)
1370      if ((ims & PCRE_MULTILINE) != 0)      if ((ims & PCRE_MULTILINE) != 0)
1371        {        {
1372        if (eptr < md->end_subject)        if (eptr < md->end_subject)
1373          { if (*eptr != NEWLINE) RRETURN(MATCH_NOMATCH); }          { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }
1374        else        else
1375          { if (md->noteol) RRETURN(MATCH_NOMATCH); }          { if (md->noteol) RRETURN(MATCH_NOMATCH); }
1376        ecode++;        ecode++;
# Line 1207  for (;;) Line 1381  for (;;)
1381        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
1382        if (!md->endonly)        if (!md->endonly)
1383          {          {
1384          if (eptr < md->end_subject - 1 ||          if (eptr != md->end_subject &&
1385             (eptr == md->end_subject - 1 && *eptr != NEWLINE))              (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
1386            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1387          ecode++;          ecode++;
1388          break;          break;
1389          }          }
1390        }        }
1391      /* ... else fall through */      /* ... else fall through for endonly */
1392    
1393      /* End of subject assertion (\z) */      /* End of subject assertion (\z) */
1394    
# Line 1226  for (;;) Line 1400  for (;;)
1400      /* End of subject or ending \n assertion (\Z) */      /* End of subject or ending \n assertion (\Z) */
1401    
1402      case OP_EODN:      case OP_EODN:
1403      if (eptr < md->end_subject - 1 ||      if (eptr != md->end_subject &&
1404         (eptr == md->end_subject - 1 && *eptr != NEWLINE)) RRETURN(MATCH_NOMATCH);          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
1405          RRETURN(MATCH_NOMATCH);
1406      ecode++;      ecode++;
1407      break;      break;
1408    
# Line 1280  for (;;) Line 1455  for (;;)
1455      /* Match a single character type; inline for speed */      /* Match a single character type; inline for speed */
1456    
1457      case OP_ANY:      case OP_ANY:
1458      if ((ims & PCRE_DOTALL) == 0 && eptr < md->end_subject && *eptr == NEWLINE)      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
1459        RRETURN(MATCH_NOMATCH);      /* Fall through */
1460    
1461        case OP_ALLANY:
1462      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);
1463  #ifdef SUPPORT_UTF8      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
     if (utf8)  
       while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  
 #endif  
1464      ecode++;      ecode++;
1465      break;      break;
1466    
# Line 1376  for (;;) Line 1550  for (;;)
1550      ecode++;      ecode++;
1551      break;      break;
1552    
1553        case OP_ANYNL:
1554        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1555        GETCHARINCTEST(c, eptr);
1556        switch(c)
1557          {
1558          default: RRETURN(MATCH_NOMATCH);
1559          case 0x000d:
1560          if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
1561          break;
1562    
1563          case 0x000a:
1564          break;
1565    
1566          case 0x000b:
1567          case 0x000c:
1568          case 0x0085:
1569          case 0x2028:
1570          case 0x2029:
1571          if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
1572          break;
1573          }
1574        ecode++;
1575        break;
1576    
1577        case OP_NOT_HSPACE:
1578        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1579        GETCHARINCTEST(c, eptr);
1580        switch(c)
1581          {
1582          default: break;
1583          case 0x09:      /* HT */
1584          case 0x20:      /* SPACE */
1585          case 0xa0:      /* NBSP */
1586          case 0x1680:    /* OGHAM SPACE MARK */
1587          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1588          case 0x2000:    /* EN QUAD */
1589          case 0x2001:    /* EM QUAD */
1590          case 0x2002:    /* EN SPACE */
1591          case 0x2003:    /* EM SPACE */
1592          case 0x2004:    /* THREE-PER-EM SPACE */
1593          case 0x2005:    /* FOUR-PER-EM SPACE */
1594          case 0x2006:    /* SIX-PER-EM SPACE */
1595          case 0x2007:    /* FIGURE SPACE */
1596          case 0x2008:    /* PUNCTUATION SPACE */
1597          case 0x2009:    /* THIN SPACE */
1598          case 0x200A:    /* HAIR SPACE */
1599          case 0x202f:    /* NARROW NO-BREAK SPACE */
1600          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1601          case 0x3000:    /* IDEOGRAPHIC SPACE */
1602          RRETURN(MATCH_NOMATCH);
1603          }
1604        ecode++;
1605        break;
1606    
1607        case OP_HSPACE:
1608        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1609        GETCHARINCTEST(c, eptr);
1610        switch(c)
1611          {
1612          default: RRETURN(MATCH_NOMATCH);
1613          case 0x09:      /* HT */
1614          case 0x20:      /* SPACE */
1615          case 0xa0:      /* NBSP */
1616          case 0x1680:    /* OGHAM SPACE MARK */
1617          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1618          case 0x2000:    /* EN QUAD */
1619          case 0x2001:    /* EM QUAD */
1620          case 0x2002:    /* EN SPACE */
1621          case 0x2003:    /* EM SPACE */
1622          case 0x2004:    /* THREE-PER-EM SPACE */
1623          case 0x2005:    /* FOUR-PER-EM SPACE */
1624          case 0x2006:    /* SIX-PER-EM SPACE */
1625          case 0x2007:    /* FIGURE SPACE */
1626          case 0x2008:    /* PUNCTUATION SPACE */
1627          case 0x2009:    /* THIN SPACE */
1628          case 0x200A:    /* HAIR SPACE */
1629          case 0x202f:    /* NARROW NO-BREAK SPACE */
1630          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1631          case 0x3000:    /* IDEOGRAPHIC SPACE */
1632          break;
1633          }
1634        ecode++;
1635        break;
1636    
1637        case OP_NOT_VSPACE:
1638        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1639        GETCHARINCTEST(c, eptr);
1640        switch(c)
1641          {
1642          default: break;
1643          case 0x0a:      /* LF */
1644          case 0x0b:      /* VT */
1645          case 0x0c:      /* FF */
1646          case 0x0d:      /* CR */
1647          case 0x85:      /* NEL */
1648          case 0x2028:    /* LINE SEPARATOR */
1649          case 0x2029:    /* PARAGRAPH SEPARATOR */
1650          RRETURN(MATCH_NOMATCH);
1651          }
1652        ecode++;
1653        break;
1654    
1655        case OP_VSPACE:
1656        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1657        GETCHARINCTEST(c, eptr);
1658        switch(c)
1659          {
1660          default: RRETURN(MATCH_NOMATCH);
1661          case 0x0a:      /* LF */
1662          case 0x0b:      /* VT */
1663          case 0x0c:      /* FF */
1664          case 0x0d:      /* CR */
1665          case 0x85:      /* NEL */
1666          case 0x2028:    /* LINE SEPARATOR */
1667          case 0x2029:    /* PARAGRAPH SEPARATOR */
1668          break;
1669          }
1670        ecode++;
1671        break;
1672    
1673  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
1674      /* Check the next character by Unicode property. We will get here only      /* Check the next character by Unicode property. We will get here only
1675      if the support is in the binary; otherwise a compile-time error occurs. */      if the support is in the binary; otherwise a compile-time error occurs. */
# Line 1385  for (;;) Line 1679  for (;;)
1679      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1680      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1681        {        {
1682        int chartype, script;        const ucd_record * prop = GET_UCD(c);
       int category = _pcre_ucp_findprop(c, &chartype, &script);  
1683    
1684        switch(ecode[1])        switch(ecode[1])
1685          {          {
# Line 1395  for (;;) Line 1688  for (;;)
1688          break;          break;
1689    
1690          case PT_LAMP:          case PT_LAMP:
1691          if ((chartype == ucp_Lu ||          if ((prop->chartype == ucp_Lu ||
1692               chartype == ucp_Ll ||               prop->chartype == ucp_Ll ||
1693               chartype == ucp_Lt) == (op == OP_NOTPROP))               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))
1694            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1695           break;           break;
1696    
1697          case PT_GC:          case PT_GC:
1698          if ((ecode[2] != category) == (op == OP_PROP))          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))
1699            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1700          break;          break;
1701    
1702          case PT_PC:          case PT_PC:
1703          if ((ecode[2] != chartype) == (op == OP_PROP))          if ((ecode[2] != prop->chartype) == (op == OP_PROP))
1704            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1705          break;          break;
1706    
1707          case PT_SC:          case PT_SC:
1708          if ((ecode[2] != script) == (op == OP_PROP))          if ((ecode[2] != prop->script) == (op == OP_PROP))
1709            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1710          break;          break;
1711    
1712          default:          default:
1713          RRETURN(PCRE_ERROR_INTERNAL);          RRETURN(PCRE_ERROR_INTERNAL);
         break;  
1714          }          }
1715    
1716        ecode += 3;        ecode += 3;
# Line 1432  for (;;) Line 1724  for (;;)
1724      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1725      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1726        {        {
1727        int chartype, script;        int category = UCD_CATEGORY(c);
       int category = _pcre_ucp_findprop(c, &chartype, &script);  
1728        if (category == ucp_M) RRETURN(MATCH_NOMATCH);        if (category == ucp_M) RRETURN(MATCH_NOMATCH);
1729        while (eptr < md->end_subject)        while (eptr < md->end_subject)
1730          {          {
# Line 1442  for (;;) Line 1733  for (;;)
1733            {            {
1734            GETCHARLEN(c, eptr, len);            GETCHARLEN(c, eptr, len);
1735            }            }
1736          category = _pcre_ucp_findprop(c, &chartype, &script);          category = UCD_CATEGORY(c);
1737          if (category != ucp_M) break;          if (category != ucp_M) break;
1738          eptr += len;          eptr += len;
1739          }          }
# Line 1463  for (;;) Line 1754  for (;;)
1754      case OP_REF:      case OP_REF:
1755        {        {
1756        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
1757        ecode += 3;                                 /* Advance past item */        ecode += 3;
1758    
1759          /* If the reference is unset, there are two possibilities:
1760    
1761          (a) In the default, Perl-compatible state, set the length to be longer
1762          than the amount of subject left; this ensures that every attempt at a
1763          match fails. We can't just fail here, because of the possibility of
1764          quantifiers with zero minima.
1765    
1766        /* If the reference is unset, set the length to be longer than the amount        (b) If the JavaScript compatibility flag is set, set the length to zero
1767        of subject left; this ensures that every attempt at a match fails. We        so that the back reference matches an empty string.
1768        can't just fail here, because of the possibility of quantifiers with zero  
1769        minima. */        Otherwise, set the length to the length of what was matched by the
1770          referenced subpattern. */
1771        length = (offset >= offset_top || md->offset_vector[offset] < 0)?  
1772          md->end_subject - eptr + 1 :        if (offset >= offset_top || md->offset_vector[offset] < 0)
1773          md->offset_vector[offset+1] - md->offset_vector[offset];          length = (md->jscript_compat)? 0 : md->end_subject - eptr + 1;
1774          else
1775            length = md->offset_vector[offset+1] - md->offset_vector[offset];
1776    
1777        /* Set up for repetition, or handle the non-repeated case */        /* Set up for repetition, or handle the non-repeated case */
1778    
# Line 1532  for (;;) Line 1832  for (;;)
1832          {          {
1833          for (fi = min;; fi++)          for (fi = min;; fi++)
1834            {            {
1835            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM14);
1836            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1837            if (fi >= max || !match_ref(offset, eptr, length, md, ims))            if (fi >= max || !match_ref(offset, eptr, length, md, ims))
1838              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
# Line 1553  for (;;) Line 1853  for (;;)
1853            }            }
1854          while (eptr >= pp)          while (eptr >= pp)
1855            {            {
1856            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM15);
1857            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1858            eptr -= length;            eptr -= length;
1859            }            }
# Line 1658  for (;;) Line 1958  for (;;)
1958            {            {
1959            for (fi = min;; fi++)            for (fi = min;; fi++)
1960              {              {
1961              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM16);
1962              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1963              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1964              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 1678  for (;;) Line 1978  for (;;)
1978            {            {
1979            for (fi = min;; fi++)            for (fi = min;; fi++)
1980              {              {
1981              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM17);
1982              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1983              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1984              c = *eptr++;              c = *eptr++;
# Line 1715  for (;;) Line 2015  for (;;)
2015              }              }
2016            for (;;)            for (;;)
2017              {              {
2018              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM18);
2019              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2020              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2021              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 1734  for (;;) Line 2034  for (;;)
2034              }              }
2035            while (eptr >= pp)            while (eptr >= pp)
2036              {              {
2037              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM19);
2038              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2039              eptr--;              eptr--;
2040              }              }
# Line 1805  for (;;) Line 2105  for (;;)
2105          {          {
2106          for (fi = min;; fi++)          for (fi = min;; fi++)
2107            {            {
2108            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20);
2109            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2110            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2111            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
# Line 1829  for (;;) Line 2129  for (;;)
2129            }            }
2130          for(;;)          for(;;)
2131            {            {
2132            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);
2133            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2134            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
2135            BACKCHAR(eptr)            if (utf8) BACKCHAR(eptr);
2136            }            }
2137          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2138          }          }
# Line 1888  for (;;) Line 2188  for (;;)
2188    
2189        else        else
2190          {          {
2191          int dc;          unsigned int dc;
2192          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
2193          ecode += length;          ecode += length;
2194    
# Line 1898  for (;;) Line 2198  for (;;)
2198          if (fc != dc)          if (fc != dc)
2199            {            {
2200  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2201            if (dc != _pcre_ucp_othercase(fc))            if (dc != UCD_OTHERCASE(fc))
2202  #endif  #endif
2203              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2204            }            }
# Line 1915  for (;;) Line 2215  for (;;)
2215        }        }
2216      break;      break;
2217    
2218      /* Match a single character repeatedly; different opcodes share code. */      /* Match a single character repeatedly. */
2219    
2220      case OP_EXACT:      case OP_EXACT:
2221      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
2222      ecode += 3;      ecode += 3;
2223      goto REPEATCHAR;      goto REPEATCHAR;
2224    
2225        case OP_POSUPTO:
2226        possessive = TRUE;
2227        /* Fall through */
2228    
2229      case OP_UPTO:      case OP_UPTO:
2230      case OP_MINUPTO:      case OP_MINUPTO:
2231      min = 0;      min = 0;
# Line 1930  for (;;) Line 2234  for (;;)
2234      ecode += 3;      ecode += 3;
2235      goto REPEATCHAR;      goto REPEATCHAR;
2236    
2237        case OP_POSSTAR:
2238        possessive = TRUE;
2239        min = 0;
2240        max = INT_MAX;
2241        ecode++;
2242        goto REPEATCHAR;
2243    
2244        case OP_POSPLUS:
2245        possessive = TRUE;
2246        min = 1;
2247        max = INT_MAX;
2248        ecode++;
2249        goto REPEATCHAR;
2250    
2251        case OP_POSQUERY:
2252        possessive = TRUE;
2253        min = 0;
2254        max = 1;
2255        ecode++;
2256        goto REPEATCHAR;
2257    
2258      case OP_STAR:      case OP_STAR:
2259      case OP_MINSTAR:      case OP_MINSTAR:
2260      case OP_PLUS:      case OP_PLUS:
# Line 1961  for (;;) Line 2286  for (;;)
2286    
2287        if (length > 1)        if (length > 1)
2288          {          {
         int oclength = 0;  
         uschar occhars[8];  
   
2289  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2290          int othercase;          unsigned int othercase;
2291          if ((ims & PCRE_CASELESS) != 0 &&          if ((ims & PCRE_CASELESS) != 0 &&
2292              (othercase = _pcre_ucp_othercase(fc)) >= 0 &&              (othercase = UCD_OTHERCASE(fc)) != fc)
              othercase >= 0)  
2293            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = _pcre_ord2utf8(othercase, occhars);
2294            else oclength = 0;
2295  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2296    
2297          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2298            {            {
2299            if (memcmp(eptr, charptr, length) == 0) eptr += length;            if (memcmp(eptr, charptr, length) == 0) eptr += length;
2300    #ifdef SUPPORT_UCP
2301            /* Need braces because of following else */            /* Need braces because of following else */
2302            else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }            else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }
2303            else            else
# Line 1982  for (;;) Line 2305  for (;;)
2305              if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);              if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);
2306              eptr += oclength;              eptr += oclength;
2307              }              }
2308    #else   /* without SUPPORT_UCP */
2309              else { RRETURN(MATCH_NOMATCH); }
2310    #endif  /* SUPPORT_UCP */
2311            }            }
2312    
2313          if (min == max) continue;          if (min == max) continue;
# Line 1990  for (;;) Line 2316  for (;;)
2316            {            {
2317            for (fi = min;; fi++)            for (fi = min;; fi++)
2318              {              {
2319              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM22);
2320              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2321              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2322              if (memcmp(eptr, charptr, length) == 0) eptr += length;              if (memcmp(eptr, charptr, length) == 0) eptr += length;
2323    #ifdef SUPPORT_UCP
2324              /* Need braces because of following else */              /* Need braces because of following else */
2325              else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }              else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }
2326              else              else
# Line 2001  for (;;) Line 2328  for (;;)
2328                if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);                if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);
2329                eptr += oclength;                eptr += oclength;
2330                }                }
2331    #else   /* without SUPPORT_UCP */
2332                else { RRETURN (MATCH_NOMATCH); }
2333    #endif  /* SUPPORT_UCP */
2334              }              }
2335            /* Control never gets here */            /* Control never gets here */
2336            }            }
2337          else  
2338            else  /* Maximize */
2339            {            {
2340            pp = eptr;            pp = eptr;
2341            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2342              {              {
2343              if (eptr > md->end_subject - length) break;              if (eptr > md->end_subject - length) break;
2344              if (memcmp(eptr, charptr, length) == 0) eptr += length;              if (memcmp(eptr, charptr, length) == 0) eptr += length;
2345    #ifdef SUPPORT_UCP
2346              else if (oclength == 0) break;              else if (oclength == 0) break;
2347              else              else
2348                {                {
2349                if (memcmp(eptr, occhars, oclength) != 0) break;                if (memcmp(eptr, occhars, oclength) != 0) break;
2350                eptr += oclength;                eptr += oclength;
2351                }                }
2352    #else   /* without SUPPORT_UCP */
2353                else break;
2354    #endif  /* SUPPORT_UCP */
2355              }              }
2356            while (eptr >= pp)  
2357              if (possessive) continue;
2358              for(;;)
2359             {             {
2360             RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);             RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM23);
2361             if (rrc != MATCH_NOMATCH) RRETURN(rrc);             if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2362               if (eptr == pp) RRETURN(MATCH_NOMATCH);
2363    #ifdef SUPPORT_UCP
2364               eptr--;
2365               BACKCHAR(eptr);
2366    #else   /* without SUPPORT_UCP */
2367             eptr -= length;             eptr -= length;
2368    #endif  /* SUPPORT_UCP */
2369             }             }
           RRETURN(MATCH_NOMATCH);  
2370            }            }
2371          /* Control never gets here */          /* Control never gets here */
2372          }          }
# Line 2064  for (;;) Line 2406  for (;;)
2406          {          {
2407          for (fi = min;; fi++)          for (fi = min;; fi++)
2408            {            {
2409            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM24);
2410            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2411            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
2412                fc != md->lcc[*eptr++])                fc != md->lcc[*eptr++])
# Line 2072  for (;;) Line 2414  for (;;)
2414            }            }
2415          /* Control never gets here */          /* Control never gets here */
2416          }          }
2417        else        else  /* Maximize */
2418          {          {
2419          pp = eptr;          pp = eptr;
2420          for (i = min; i < max; i++)          for (i = min; i < max; i++)
# Line 2080  for (;;) Line 2422  for (;;)
2422            if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break;            if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break;
2423            eptr++;            eptr++;
2424            }            }
2425            if (possessive) continue;
2426          while (eptr >= pp)          while (eptr >= pp)
2427            {            {
2428            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM25);
2429            eptr--;            eptr--;
2430            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2431            }            }
# Line 2101  for (;;) Line 2444  for (;;)
2444          {          {
2445          for (fi = min;; fi++)          for (fi = min;; fi++)
2446            {            {
2447            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM26);
2448            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2449            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)
2450              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2451            }            }
2452          /* Control never gets here */          /* Control never gets here */
2453          }          }
2454        else        else  /* Maximize */
2455          {          {
2456          pp = eptr;          pp = eptr;
2457          for (i = min; i < max; i++)          for (i = min; i < max; i++)
# Line 2116  for (;;) Line 2459  for (;;)
2459            if (eptr >= md->end_subject || fc != *eptr) break;            if (eptr >= md->end_subject || fc != *eptr) break;
2460            eptr++;            eptr++;
2461            }            }
2462            if (possessive) continue;
2463          while (eptr >= pp)          while (eptr >= pp)
2464            {            {
2465            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27);
2466            eptr--;            eptr--;
2467            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2468            }            }
# Line 2168  for (;;) Line 2512  for (;;)
2512      ecode += 3;      ecode += 3;
2513      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
2514    
2515        case OP_NOTPOSSTAR:
2516        possessive = TRUE;
2517        min = 0;
2518        max = INT_MAX;
2519        ecode++;
2520        goto REPEATNOTCHAR;
2521    
2522        case OP_NOTPOSPLUS:
2523        possessive = TRUE;
2524        min = 1;
2525        max = INT_MAX;
2526        ecode++;
2527        goto REPEATNOTCHAR;
2528    
2529        case OP_NOTPOSQUERY:
2530        possessive = TRUE;
2531        min = 0;
2532        max = 1;
2533        ecode++;
2534        goto REPEATNOTCHAR;
2535    
2536        case OP_NOTPOSUPTO:
2537        possessive = TRUE;
2538        min = 0;
2539        max = GET2(ecode, 1);
2540        ecode += 3;
2541        goto REPEATNOTCHAR;
2542    
2543      case OP_NOTSTAR:      case OP_NOTSTAR:
2544      case OP_NOTMINSTAR:      case OP_NOTMINSTAR:
2545      case OP_NOTPLUS:      case OP_NOTPLUS:
# Line 2207  for (;;) Line 2579  for (;;)
2579        /* UTF-8 mode */        /* UTF-8 mode */
2580        if (utf8)        if (utf8)
2581          {          {
2582          register int d;          register unsigned int d;
2583          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2584            {            {
2585            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
# Line 2232  for (;;) Line 2604  for (;;)
2604          /* UTF-8 mode */          /* UTF-8 mode */
2605          if (utf8)          if (utf8)
2606            {            {
2607            register int d;            register unsigned int d;
2608            for (fi = min;; fi++)            for (fi = min;; fi++)
2609              {              {
2610              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28);
2611              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2612                if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2613              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2614              if (d < 256) d = md->lcc[d];              if (d < 256) d = md->lcc[d];
2615              if (fi >= max || eptr >= md->end_subject || fc == d)              if (fc == d) RRETURN(MATCH_NOMATCH);
2616                RRETURN(MATCH_NOMATCH);  
2617              }              }
2618            }            }
2619          else          else
# Line 2249  for (;;) Line 2622  for (;;)
2622            {            {
2623            for (fi = min;; fi++)            for (fi = min;; fi++)
2624              {              {
2625              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM29);
2626              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2627              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])
2628                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2268  for (;;) Line 2641  for (;;)
2641          /* UTF-8 mode */          /* UTF-8 mode */
2642          if (utf8)          if (utf8)
2643            {            {
2644            register int d;            register unsigned int d;
2645            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2646              {              {
2647              int len = 1;              int len = 1;
# Line 2278  for (;;) Line 2651  for (;;)
2651              if (fc == d) break;              if (fc == d) break;
2652              eptr += len;              eptr += len;
2653              }              }
2654            for(;;)          if (possessive) continue;
2655            for(;;)
2656              {              {
2657              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM30);
2658              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2659              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2660              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2295  for (;;) Line 2669  for (;;)
2669              if (eptr >= md->end_subject || fc == md->lcc[*eptr]) break;              if (eptr >= md->end_subject || fc == md->lcc[*eptr]) break;
2670              eptr++;              eptr++;
2671              }              }
2672              if (possessive) continue;
2673            while (eptr >= pp)            while (eptr >= pp)
2674              {              {
2675              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM31);
2676              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2677              eptr--;              eptr--;
2678              }              }
# Line 2316  for (;;) Line 2691  for (;;)
2691        /* UTF-8 mode */        /* UTF-8 mode */
2692        if (utf8)        if (utf8)
2693          {          {
2694          register int d;          register unsigned int d;
2695          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2696            {            {
2697            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
# Line 2339  for (;;) Line 2714  for (;;)
2714          /* UTF-8 mode */          /* UTF-8 mode */
2715          if (utf8)          if (utf8)
2716            {            {
2717            register int d;            register unsigned int d;
2718            for (fi = min;; fi++)            for (fi = min;; fi++)
2719              {              {
2720              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32);
2721              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2722                if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2723              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2724              if (fi >= max || eptr >= md->end_subject || fc == d)              if (fc == d) RRETURN(MATCH_NOMATCH);
               RRETURN(MATCH_NOMATCH);  
2725              }              }
2726            }            }
2727          else          else
# Line 2355  for (;;) Line 2730  for (;;)
2730            {            {
2731            for (fi = min;; fi++)            for (fi = min;; fi++)
2732              {              {
2733              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM33);
2734              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2735              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)
2736                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2374  for (;;) Line 2749  for (;;)
2749          /* UTF-8 mode */          /* UTF-8 mode */
2750          if (utf8)          if (utf8)
2751            {            {
2752            register int d;            register unsigned int d;
2753            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2754              {              {
2755              int len = 1;              int len = 1;
# Line 2383  for (;;) Line 2758  for (;;)
2758              if (fc == d) break;              if (fc == d) break;
2759              eptr += len;              eptr += len;
2760              }              }
2761              if (possessive) continue;
2762            for(;;)            for(;;)
2763              {              {
2764              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM34);
2765              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2766              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2767              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2400  for (;;) Line 2776  for (;;)
2776              if (eptr >= md->end_subject || fc == *eptr) break;              if (eptr >= md->end_subject || fc == *eptr) break;
2777              eptr++;              eptr++;
2778              }              }
2779              if (possessive) continue;
2780            while (eptr >= pp)            while (eptr >= pp)
2781              {              {
2782              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM35);
2783              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2784              eptr--;              eptr--;
2785              }              }
# Line 2431  for (;;) Line 2808  for (;;)
2808      ecode += 3;      ecode += 3;
2809      goto REPEATTYPE;      goto REPEATTYPE;
2810    
2811        case OP_TYPEPOSSTAR:
2812        possessive = TRUE;
2813        min = 0;
2814        max = INT_MAX;
2815        ecode++;
2816        goto REPEATTYPE;
2817    
2818        case OP_TYPEPOSPLUS:
2819        possessive = TRUE;
2820        min = 1;
2821        max = INT_MAX;
2822        ecode++;
2823        goto REPEATTYPE;
2824    
2825        case OP_TYPEPOSQUERY:
2826        possessive = TRUE;
2827        min = 0;
2828        max = 1;
2829        ecode++;
2830        goto REPEATTYPE;
2831    
2832        case OP_TYPEPOSUPTO:
2833        possessive = TRUE;
2834        min = 0;
2835        max = GET2(ecode, 1);
2836        ecode += 3;
2837        goto REPEATTYPE;
2838    
2839      case OP_TYPESTAR:      case OP_TYPESTAR:
2840      case OP_TYPEMINSTAR:      case OP_TYPEMINSTAR:
2841      case OP_TYPEPLUS:      case OP_TYPEPLUS:
# Line 2481  for (;;) Line 2886  for (;;)
2886            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2887              {              {
2888              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2889              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2890              }              }
2891            break;            break;
2892    
# Line 2489  for (;;) Line 2894  for (;;)
2894            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2895              {              {
2896              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2897              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2898              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_chartype = UCD_CHARTYPE(c);
2899              if ((prop_chartype == ucp_Lu ||              if ((prop_chartype == ucp_Lu ||
2900                   prop_chartype == ucp_Ll ||                   prop_chartype == ucp_Ll ||
2901                   prop_chartype == ucp_Lt) == prop_fail_result)                   prop_chartype == ucp_Lt) == prop_fail_result)
# Line 2502  for (;;) Line 2907  for (;;)
2907            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2908              {              {
2909              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2910              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2911              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
2912              if ((prop_category == prop_value) == prop_fail_result)              if ((prop_category == prop_value) == prop_fail_result)
2913                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2914              }              }
# Line 2513  for (;;) Line 2918  for (;;)
2918            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2919              {              {
2920              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2921              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2922              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_chartype = UCD_CHARTYPE(c);
2923              if ((prop_chartype == prop_value) == prop_fail_result)              if ((prop_chartype == prop_value) == prop_fail_result)
2924                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2925              }              }
# Line 2524  for (;;) Line 2929  for (;;)
2929            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2930              {              {
2931              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2932              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2933              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_script = UCD_SCRIPT(c);
2934              if ((prop_script == prop_value) == prop_fail_result)              if ((prop_script == prop_value) == prop_fail_result)
2935                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2936              }              }
# Line 2533  for (;;) Line 2938  for (;;)
2938    
2939            default:            default:
2940            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
           break;  
2941            }            }
2942          }          }
2943    
# Line 2545  for (;;) Line 2949  for (;;)
2949          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2950            {            {
2951            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
2952            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);            prop_category = UCD_CATEGORY(c);
2953            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);
2954            while (eptr < md->end_subject)            while (eptr < md->end_subject)
2955              {              {
# Line 2554  for (;;) Line 2958  for (;;)
2958                {                {
2959                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
2960                }                }
2961              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
2962              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
2963              eptr += len;              eptr += len;
2964              }              }
# Line 2572  for (;;) Line 2976  for (;;)
2976          case OP_ANY:          case OP_ANY:
2977          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2978            {            {
2979            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject || IS_NEWLINE(eptr))
              (*eptr++ == NEWLINE && (ims & PCRE_DOTALL) == 0))  
2980              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2981              eptr++;
2982              while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
2983              }
2984            break;
2985    
2986            case OP_ALLANY:
2987            for (i = 1; i <= min; i++)
2988              {
2989              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2990              eptr++;
2991            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
2992            }            }
2993          break;          break;
# Line 2583  for (;;) Line 2996  for (;;)
2996          eptr += min;          eptr += min;
2997          break;          break;
2998    
2999            case OP_ANYNL:
3000            for (i = 1; i <= min; i++)
3001              {
3002              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3003              GETCHARINC(c, eptr);
3004              switch(c)
3005                {
3006                default: RRETURN(MATCH_NOMATCH);
3007                case 0x000d:
3008                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3009                break;
3010    
3011                case 0x000a:
3012                break;
3013    
3014                case 0x000b:
3015                case 0x000c:
3016                case 0x0085:
3017                case 0x2028:
3018                case 0x2029:
3019                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3020                break;
3021                }
3022              }
3023            break;
3024    
3025            case OP_NOT_HSPACE:
3026            for (i = 1; i <= min; i++)
3027              {
3028              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3029              GETCHARINC(c, eptr);
3030              switch(c)
3031                {
3032                default: break;
3033                case 0x09:      /* HT */
3034                case 0x20:      /* SPACE */
3035                case 0xa0:      /* NBSP */
3036                case 0x1680:    /* OGHAM SPACE MARK */
3037                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3038                case 0x2000:    /* EN QUAD */
3039                case 0x2001:    /* EM QUAD */
3040                case 0x2002:    /* EN SPACE */
3041                case 0x2003:    /* EM SPACE */
3042                case 0x2004:    /* THREE-PER-EM SPACE */
3043                case 0x2005:    /* FOUR-PER-EM SPACE */
3044                case 0x2006:    /* SIX-PER-EM SPACE */
3045                case 0x2007:    /* FIGURE SPACE */
3046                case 0x2008:    /* PUNCTUATION SPACE */
3047                case 0x2009:    /* THIN SPACE */
3048                case 0x200A:    /* HAIR SPACE */
3049                case 0x202f:    /* NARROW NO-BREAK SPACE */
3050                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3051                case 0x3000:    /* IDEOGRAPHIC SPACE */
3052                RRETURN(MATCH_NOMATCH);
3053                }
3054              }
3055            break;
3056    
3057            case OP_HSPACE:
3058            for (i = 1; i <= min; i++)
3059              {
3060              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3061              GETCHARINC(c, eptr);
3062              switch(c)
3063                {
3064                default: RRETURN(MATCH_NOMATCH);
3065                case 0x09:      /* HT */
3066                case 0x20:      /* SPACE */
3067                case 0xa0:      /* NBSP */
3068                case 0x1680:    /* OGHAM SPACE MARK */
3069                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3070                case 0x2000:    /* EN QUAD */
3071                case 0x2001:    /* EM QUAD */
3072                case 0x2002:    /* EN SPACE */
3073                case 0x2003:    /* EM SPACE */
3074                case 0x2004:    /* THREE-PER-EM SPACE */
3075                case 0x2005:    /* FOUR-PER-EM SPACE */
3076                case 0x2006:    /* SIX-PER-EM SPACE */
3077                case 0x2007:    /* FIGURE SPACE */
3078                case 0x2008:    /* PUNCTUATION SPACE */
3079                case 0x2009:    /* THIN SPACE */
3080                case 0x200A:    /* HAIR SPACE */
3081                case 0x202f:    /* NARROW NO-BREAK SPACE */
3082                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3083                case 0x3000:    /* IDEOGRAPHIC SPACE */
3084                break;
3085                }
3086              }
3087            break;
3088    
3089            case OP_NOT_VSPACE:
3090            for (i = 1; i <= min; i++)
3091              {
3092              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3093              GETCHARINC(c, eptr);
3094              switch(c)
3095                {
3096                default: break;
3097                case 0x0a:      /* LF */
3098                case 0x0b:      /* VT */
3099                case 0x0c:      /* FF */
3100                case 0x0d:      /* CR */
3101                case 0x85:      /* NEL */
3102                case 0x2028:    /* LINE SEPARATOR */
3103                case 0x2029:    /* PARAGRAPH SEPARATOR */
3104                RRETURN(MATCH_NOMATCH);
3105                }
3106              }
3107            break;
3108    
3109            case OP_VSPACE:
3110            for (i = 1; i <= min; i++)
3111              {
3112              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3113              GETCHARINC(c, eptr);
3114              switch(c)
3115                {
3116                default: RRETURN(MATCH_NOMATCH);
3117                case 0x0a:      /* LF */
3118                case 0x0b:      /* VT */
3119                case 0x0c:      /* FF */
3120                case 0x0d:      /* CR */
3121                case 0x85:      /* NEL */
3122                case 0x2028:    /* LINE SEPARATOR */
3123                case 0x2029:    /* PARAGRAPH SEPARATOR */
3124                break;
3125                }
3126              }
3127            break;
3128    
3129          case OP_NOT_DIGIT:          case OP_NOT_DIGIT:
3130          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3131            {            {
# Line 2607  for (;;) Line 3150  for (;;)
3150          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3151            {            {
3152            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3153               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_space) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0))
3154              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3155            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3156            }            }
3157          break;          break;
3158    
# Line 2627  for (;;) Line 3170  for (;;)
3170          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3171            {            {
3172            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3173               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_word) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0))
3174              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3175            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3176            }            }
3177          break;          break;
3178    
# Line 2651  for (;;) Line 3194  for (;;)
3194  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF8 */
3195    
3196        /* Code for the non-UTF-8 case for minimum matching of operators other        /* Code for the non-UTF-8 case for minimum matching of operators other
3197        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. We can assume that there are the minimum
3198          number of bytes present, as this was tested above. */
3199    
3200        switch(ctype)        switch(ctype)
3201          {          {
3202          case OP_ANY:          case OP_ANY:
3203          if ((ims & PCRE_DOTALL) == 0)          for (i = 1; i <= min; i++)
3204            {            {
3205            for (i = 1; i <= min; i++)            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
3206              if (*eptr++ == NEWLINE) RRETURN(MATCH_NOMATCH);            eptr++;
3207            }            }
3208          else eptr += min;          break;
3209    
3210            case OP_ALLANY:
3211            eptr += min;
3212          break;          break;
3213    
3214          case OP_ANYBYTE:          case OP_ANYBYTE:
3215          eptr += min;          eptr += min;
3216          break;          break;
3217    
3218            /* Because of the CRLF case, we can't assume the minimum number of
3219            bytes are present in this case. */
3220    
3221            case OP_ANYNL:
3222            for (i = 1; i <= min; i++)
3223              {
3224              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3225              switch(*eptr++)
3226                {
3227                default: RRETURN(MATCH_NOMATCH);
3228                case 0x000d:
3229                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3230                break;
3231                case 0x000a:
3232                break;
3233    
3234                case 0x000b:
3235                case 0x000c:
3236                case 0x0085:
3237                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3238                break;
3239                }
3240              }
3241            break;
3242    
3243            case OP_NOT_HSPACE:
3244            for (i = 1; i <= min; i++)
3245              {
3246              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3247              switch(*eptr++)
3248                {
3249                default: break;
3250                case 0x09:      /* HT */
3251                case 0x20:      /* SPACE */
3252                case 0xa0:      /* NBSP */
3253                RRETURN(MATCH_NOMATCH);
3254                }
3255              }
3256            break;
3257    
3258            case OP_HSPACE:
3259            for (i = 1; i <= min; i++)
3260              {
3261              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3262              switch(*eptr++)
3263                {
3264                default: RRETURN(MATCH_NOMATCH);
3265                case 0x09:      /* HT */
3266                case 0x20:      /* SPACE */
3267                case 0xa0:      /* NBSP */
3268                break;
3269                }
3270              }
3271            break;
3272    
3273            case OP_NOT_VSPACE:
3274            for (i = 1; i <= min; i++)
3275              {
3276              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3277              switch(*eptr++)
3278                {
3279                default: break;
3280                case 0x0a:      /* LF */
3281                case 0x0b:      /* VT */
3282                case 0x0c:      /* FF */
3283                case 0x0d:      /* CR */
3284                case 0x85:      /* NEL */
3285                RRETURN(MATCH_NOMATCH);
3286                }
3287              }
3288            break;
3289    
3290            case OP_VSPACE:
3291            for (i = 1; i <= min; i++)
3292              {
3293              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3294              switch(*eptr++)
3295                {
3296                default: RRETURN(MATCH_NOMATCH);
3297                case 0x0a:      /* LF */
3298                case 0x0b:      /* VT */
3299                case 0x0c:      /* FF */
3300                case 0x0d:      /* CR */
3301                case 0x85:      /* NEL */
3302                break;
3303                }
3304              }
3305            break;
3306    
3307          case OP_NOT_DIGIT:          case OP_NOT_DIGIT:
3308          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3309            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
# Line 2723  for (;;) Line 3359  for (;;)
3359            case PT_ANY:            case PT_ANY:
3360            for (fi = min;; fi++)            for (fi = min;; fi++)
3361              {              {
3362              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM36);
3363              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3364              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3365              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
3366              if (prop_fail_result) RRETURN(MATCH_NOMATCH);              if (prop_fail_result) RRETURN(MATCH_NOMATCH);
3367              }              }
3368            break;            /* Control never gets here */
3369    
3370            case PT_LAMP:            case PT_LAMP:
3371            for (fi = min;; fi++)            for (fi = min;; fi++)
3372              {              {
3373              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM37);
3374              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3375              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3376              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
3377              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_chartype = UCD_CHARTYPE(c);
3378              if ((prop_chartype == ucp_Lu ||              if ((prop_chartype == ucp_Lu ||
3379                   prop_chartype == ucp_Ll ||                   prop_chartype == ucp_Ll ||
3380                   prop_chartype == ucp_Lt) == prop_fail_result)                   prop_chartype == ucp_Lt) == prop_fail_result)
3381                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3382              }              }
3383            break;            /* Control never gets here */
3384    
3385            case PT_GC:            case PT_GC:
3386            for (fi = min;; fi++)            for (fi = min;; fi++)
3387              {              {
3388              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM38);
3389              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3390              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3391              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
3392              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
3393              if ((prop_category == prop_value) == prop_fail_result)              if ((prop_category == prop_value) == prop_fail_result)
3394                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3395              }              }
3396            break;            /* Control never gets here */
3397    
3398            case PT_PC:            case PT_PC:
3399            for (fi = min;; fi++)            for (fi = min;; fi++)
3400              {              {
3401              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM39);
3402              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3403              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3404              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
3405              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_chartype = UCD_CHARTYPE(c);
3406              if ((prop_chartype == prop_value) == prop_fail_result)              if ((prop_chartype == prop_value) == prop_fail_result)
3407                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3408              }              }
3409            break;            /* Control never gets here */
3410    
3411            case PT_SC:            case PT_SC:
3412            for (fi = min;; fi++)            for (fi = min;; fi++)
3413              {              {
3414              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM40);
3415              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3416              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3417              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
3418              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_script = UCD_SCRIPT(c);
3419              if ((prop_script == prop_value) == prop_fail_result)              if ((prop_script == prop_value) == prop_fail_result)
3420                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3421              }              }
3422            break;            /* Control never gets here */
3423    
3424            default:            default:
3425            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
           break;  
3426            }            }
3427          }          }
3428    
# Line 2798  for (;;) Line 3433  for (;;)
3433          {          {
3434          for (fi = min;; fi++)          for (fi = min;; fi++)
3435            {            {
3436            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM41);
3437            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3438            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3439            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3440            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);            prop_category = UCD_CATEGORY(c);
3441            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);
3442            while (eptr < md->end_subject)            while (eptr < md->end_subject)
3443              {              {
# Line 2811  for (;;) Line 3446  for (;;)
3446                {                {
3447                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3448                }                }
3449              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
3450              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3451              eptr += len;              eptr += len;
3452              }              }
# Line 2827  for (;;) Line 3462  for (;;)
3462          {          {
3463          for (fi = min;; fi++)          for (fi = min;; fi++)
3464            {            {
3465            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42);
3466            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3467            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject ||
3468                   (ctype == OP_ANY && IS_NEWLINE(eptr)))
3469                RRETURN(MATCH_NOMATCH);
3470    
3471            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
3472            switch(ctype)            switch(ctype)
3473              {              {
3474              case OP_ANY:              case OP_ANY:        /* This is the non-NL case */
3475              if ((ims & PCRE_DOTALL) == 0 && c == NEWLINE) RRETURN(MATCH_NOMATCH);              case OP_ALLANY:
3476                case OP_ANYBYTE:
3477              break;              break;
3478    
3479              case OP_ANYBYTE:              case OP_ANYNL:
3480                switch(c)
3481                  {
3482                  default: RRETURN(MATCH_NOMATCH);
3483                  case 0x000d:
3484                  if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3485                  break;
3486                  case 0x000a:
3487                  break;
3488    
3489                  case 0x000b:
3490                  case 0x000c:
3491                  case 0x0085:
3492                  case 0x2028:
3493                  case 0x2029:
3494                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3495                  break;
3496                  }
3497                break;
3498    
3499                case OP_NOT_HSPACE:
3500                switch(c)
3501                  {
3502                  default: break;
3503                  case 0x09:      /* HT */
3504                  case 0x20:      /* SPACE */
3505                  case 0xa0:      /* NBSP */
3506                  case 0x1680:    /* OGHAM SPACE MARK */
3507                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3508                  case 0x2000:    /* EN QUAD */
3509                  case 0x2001:    /* EM QUAD */
3510                  case 0x2002:    /* EN SPACE */
3511                  case 0x2003:    /* EM SPACE */
3512                  case 0x2004:    /* THREE-PER-EM SPACE */
3513                  case 0x2005:    /* FOUR-PER-EM SPACE */
3514                  case 0x2006:    /* SIX-PER-EM SPACE */
3515                  case 0x2007:    /* FIGURE SPACE */
3516                  case 0x2008:    /* PUNCTUATION SPACE */
3517                  case 0x2009:    /* THIN SPACE */
3518                  case 0x200A:    /* HAIR SPACE */
3519                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3520                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3521                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3522                  RRETURN(MATCH_NOMATCH);
3523                  }
3524                break;
3525    
3526                case OP_HSPACE:
3527                switch(c)
3528                  {
3529                  default: RRETURN(MATCH_NOMATCH);
3530                  case 0x09:      /* HT */
3531                  case 0x20:      /* SPACE */
3532                  case 0xa0:      /* NBSP */
3533                  case 0x1680:    /* OGHAM SPACE MARK */
3534                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3535                  case 0x2000:    /* EN QUAD */
3536                  case 0x2001:    /* EM QUAD */
3537                  case 0x2002:    /* EN SPACE */
3538                  case 0x2003:    /* EM SPACE */
3539                  case 0x2004:    /* THREE-PER-EM SPACE */
3540                  case 0x2005:    /* FOUR-PER-EM SPACE */
3541                  case 0x2006:    /* SIX-PER-EM SPACE */
3542                  case 0x2007:    /* FIGURE SPACE */
3543                  case 0x2008:    /* PUNCTUATION SPACE */
3544                  case 0x2009:    /* THIN SPACE */
3545                  case 0x200A:    /* HAIR SPACE */
3546                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3547                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3548                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3549                  break;
3550                  }
3551                break;
3552    
3553                case OP_NOT_VSPACE:
3554                switch(c)
3555                  {
3556                  default: break;
3557                  case 0x0a:      /* LF */
3558                  case 0x0b:      /* VT */
3559                  case 0x0c:      /* FF */
3560                  case 0x0d:      /* CR */
3561                  case 0x85:      /* NEL */
3562                  case 0x2028:    /* LINE SEPARATOR */
3563                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3564                  RRETURN(MATCH_NOMATCH);
3565                  }
3566                break;
3567    
3568                case OP_VSPACE:
3569                switch(c)
3570                  {
3571                  default: RRETURN(MATCH_NOMATCH);
3572                  case 0x0a:      /* LF */
3573                  case 0x0b:      /* VT */
3574                  case 0x0c:      /* FF */
3575                  case 0x0d:      /* CR */
3576                  case 0x85:      /* NEL */
3577                  case 0x2028:    /* LINE SEPARATOR */
3578                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3579                  break;
3580                  }
3581              break;              break;
3582    
3583              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
# Line 2882  for (;;) Line 3621  for (;;)
3621          {          {
3622          for (fi = min;; fi++)          for (fi = min;; fi++)
3623            {            {
3624            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43);
3625            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3626            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject ||
3627                   (ctype == OP_ANY && IS_NEWLINE(eptr)))
3628                RRETURN(MATCH_NOMATCH);
3629    
3630            c = *eptr++;            c = *eptr++;
3631            switch(ctype)            switch(ctype)
3632              {              {
3633              case OP_ANY:              case OP_ANY:     /* This is the non-NL case */
3634              if ((ims & PCRE_DOTALL) == 0 && c == NEWLINE) RRETURN(MATCH_NOMATCH);              case OP_ALLANY:
3635                case OP_ANYBYTE:
3636              break;              break;
3637    
3638              case OP_ANYBYTE:              case OP_ANYNL:
3639                switch(c)
3640                  {
3641                  default: RRETURN(MATCH_NOMATCH);
3642                  case 0x000d:
3643                  if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3644                  break;
3645    
3646                  case 0x000a:
3647                  break;
3648    
3649                  case 0x000b:
3650                  case 0x000c:
3651                  case 0x0085:
3652                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3653                  break;
3654                  }
3655                break;
3656    
3657                case OP_NOT_HSPACE:
3658                switch(c)
3659                  {
3660                  default: break;
3661                  case 0x09:      /* HT */
3662                  case 0x20:      /* SPACE */
3663                  case 0xa0:      /* NBSP */
3664                  RRETURN(MATCH_NOMATCH);
3665                  }
3666                break;
3667    
3668                case OP_HSPACE:
3669                switch(c)
3670                  {
3671                  default: RRETURN(MATCH_NOMATCH);
3672                  case 0x09:      /* HT */
3673                  case 0x20:      /* SPACE */
3674                  case 0xa0:      /* NBSP */
3675                  break;
3676                  }
3677                break;
3678    
3679                case OP_NOT_VSPACE:
3680                switch(c)
3681                  {
3682                  default: break;
3683                  case 0x0a:      /* LF */
3684                  case 0x0b:      /* VT */
3685                  case 0x0c:      /* FF */
3686                  case 0x0d:      /* CR */
3687                  case 0x85:      /* NEL */
3688                  RRETURN(MATCH_NOMATCH);
3689                  }
3690                break;
3691    
3692                case OP_VSPACE:
3693                switch(c)
3694                  {
3695                  default: RRETURN(MATCH_NOMATCH);
3696                  case 0x0a:      /* LF */
3697                  case 0x0b:      /* VT */
3698                  case 0x0c:      /* FF */
3699                  case 0x0d:      /* CR */
3700                  case 0x85:      /* NEL */
3701                  break;
3702                  }
3703              break;              break;
3704    
3705              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
# Line 2927  for (;;) Line 3734  for (;;)
3734        /* Control never gets here */        /* Control never gets here */
3735        }        }
3736    
3737      /* If maximizing it is worth using inline code for speed, doing the type      /* If maximizing, it is worth using inline code for speed, doing the type
3738      test once at the start (i.e. keep it out of the loop). Again, keep the      test once at the start (i.e. keep it out of the loop). Again, keep the
3739      UTF-8 and UCP stuff separate. */      UTF-8 and UCP stuff separate. */
3740    
# Line 2957  for (;;) Line 3764  for (;;)
3764              int len = 1;              int len = 1;
3765              if (eptr >= md->end_subject) break;              if (eptr >= md->end_subject) break;
3766              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
3767              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_chartype = UCD_CHARTYPE(c);
3768              if ((prop_chartype == ucp_Lu ||              if ((prop_chartype == ucp_Lu ||
3769                   prop_chartype == ucp_Ll ||                   prop_chartype == ucp_Ll ||
3770                   prop_chartype == ucp_Lt) == prop_fail_result)                   prop_chartype == ucp_Lt) == prop_fail_result)
# Line 2972  for (;;) Line 3779  for (;;)
3779              int len = 1;              int len = 1;
3780              if (eptr >= md->end_subject) break;              if (eptr >= md->end_subject) break;
3781              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
3782              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
3783              if ((prop_category == prop_value) == prop_fail_result)              if ((prop_category == prop_value) == prop_fail_result)
3784                break;                break;
3785              eptr+= len;              eptr+= len;
# Line 2985  for (;;) Line 3792  for (;;)
3792              int len = 1;              int len = 1;
3793              if (eptr >= md->end_subject) break;              if (eptr >= md->end_subject) break;
3794              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
3795              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_chartype = UCD_CHARTYPE(c);
3796              if ((prop_chartype == prop_value) == prop_fail_result)              if ((prop_chartype == prop_value) == prop_fail_result)
3797                break;                break;
3798              eptr+= len;              eptr+= len;
# Line 2998  for (;;) Line 3805  for (;;)
3805              int len = 1;              int len = 1;
3806              if (eptr >= md->end_subject) break;              if (eptr >= md->end_subject) break;
3807              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
3808              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_script = UCD_SCRIPT(c);
3809              if ((prop_script == prop_value) == prop_fail_result)              if ((prop_script == prop_value) == prop_fail_result)
3810                break;                break;
3811              eptr+= len;              eptr+= len;
# Line 3008  for (;;) Line 3815  for (;;)
3815    
3816          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
3817    
3818            if (possessive) continue;
3819          for(;;)          for(;;)
3820            {            {
3821            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44);
3822            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3823            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3824            BACKCHAR(eptr);            if (utf8) BACKCHAR(eptr);
3825            }            }
3826          }          }
3827    
# Line 3026  for (;;) Line 3834  for (;;)
3834            {            {
3835            if (eptr >= md->end_subject) break;            if (eptr >= md->end_subject) break;
3836            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3837            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);            prop_category = UCD_CATEGORY(c);
3838            if (prop_category == ucp_M) break;            if (prop_category == ucp_M) break;
3839            while (eptr < md->end_subject)            while (eptr < md->end_subject)
3840              {              {
# Line 3035  for (;;) Line 3843  for (;;)
3843                {                {
3844                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3845                }                }
3846              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
3847              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3848              eptr += len;              eptr += len;
3849              }              }
# Line 3043  for (;;) Line 3851  for (;;)
3851    
3852          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
3853    
3854            if (possessive) continue;
3855          for(;;)          for(;;)
3856            {            {
3857            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM45);
3858            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3859            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3860            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
3861              {              {
3862              int len = 1;              int len = 1;
             BACKCHAR(eptr);  
3863              if (!utf8) c = *eptr; else              if (!utf8) c = *eptr; else
3864                {                {
3865                  BACKCHAR(eptr);
3866                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3867                }                }
3868              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
3869              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3870              eptr--;              eptr--;
3871              }              }
# Line 3074  for (;;) Line 3883  for (;;)
3883          switch(ctype)          switch(ctype)
3884            {            {
3885            case OP_ANY:            case OP_ANY:
   
           /* Special code is required for UTF8, but when the maximum is unlimited  
           we don't need it, so we repeat the non-UTF8 code. This is probably  
           worth it, because .* is quite a common idiom. */  
   
3886            if (max < INT_MAX)            if (max < INT_MAX)
3887              {              {
3888              if ((ims & PCRE_DOTALL) == 0)              for (i = min; i < max; i++)
               {  
               for (i = min; i < max; i++)  
                 {  
                 if (eptr >= md->end_subject || *eptr == NEWLINE) break;  
                 eptr++;  
                 while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  
                 }  
               }  
             else  
3889                {                {
3890                for (i = min; i < max; i++)                if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
3891                  {                eptr++;
3892                  eptr++;                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
                 while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  
                 }  
3893                }                }
3894              }              }
3895    
# Line 3104  for (;;) Line 3897  for (;;)
3897    
3898            else            else
3899              {              {
3900              if ((ims & PCRE_DOTALL) == 0)              for (i = min; i < max; i++)
3901                {                {
3902                for (i = min; i < max; i++)                if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
3903                  {                eptr++;
3904                  if (eptr >= md->end_subject || *eptr == NEWLINE) break;                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
                 eptr++;  
                 }  
               break;  
3905                }                }
3906              else              }
3907              break;
3908    
3909              case OP_ALLANY:
3910              if (max < INT_MAX)
3911                {
3912                for (i = min; i < max; i++)
3913                {                {
3914                c = max - min;                if (eptr >= md->end_subject) break;
3915                if (c > md->end_subject - eptr) c = md->end_subject - eptr;                eptr++;
3916                eptr += c;                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3917                }                }
3918              }              }
3919              else eptr = md->end_subject;   /* Unlimited UTF-8 repeat */
3920            break;            break;
3921    
3922            /* The byte case is the same as non-UTF8 */            /* The byte case is the same as non-UTF8 */
3923    
3924            case OP_ANYBYTE:            case OP_ANYBYTE:
3925            c = max - min;            c = max - min;
3926            if (c > md->end_subject - eptr) c = md->end_subject - eptr;            if (c > (unsigned int)(md->end_subject - eptr))
3927                c = md->end_subject - eptr;
3928            eptr += c;            eptr += c;
3929            break;            break;
3930    
3931              case OP_ANYNL:
3932              for (i = min; i < max; i++)
3933                {
3934                int len = 1;
3935                if (eptr >= md->end_subject) break;
3936                GETCHARLEN(c, eptr, len);
3937                if (c == 0x000d)
3938                  {
3939                  if (++eptr >= md->end_subject) break;
3940                  if (*eptr == 0x000a) eptr++;
3941                  }
3942                else
3943                  {
3944                  if (c != 0x000a &&
3945                      (md->bsr_anycrlf ||
3946                       (c != 0x000b && c != 0x000c &&
3947                        c != 0x0085 && c != 0x2028 && c != 0x2029)))
3948                    break;
3949                  eptr += len;
3950                  }
3951                }
3952              break;
3953    
3954              case OP_NOT_HSPACE:
3955              case OP_HSPACE:
3956              for (i = min; i < max; i++)
3957                {
3958                BOOL gotspace;
3959                int len = 1;
3960                if (eptr >= md->end_subject) break;
3961                GETCHARLEN(c, eptr, len);
3962                switch(c)
3963                  {
3964                  default: gotspace = FALSE; break;
3965                  case 0x09:      /* HT */
3966                  case 0x20:      /* SPACE */
3967                  case 0xa0:      /* NBSP */
3968                  case 0x1680:    /* OGHAM SPACE MARK */
3969                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3970                  case 0x2000:    /* EN QUAD */
3971                  case 0x2001:    /* EM QUAD */
3972                  case 0x2002:    /* EN SPACE */
3973                  case 0x2003:    /* EM SPACE */
3974                  case 0x2004:    /* THREE-PER-EM SPACE */
3975                  case 0x2005:    /* FOUR-PER-EM SPACE */
3976                  case 0x2006:    /* SIX-PER-EM SPACE */
3977                  case 0x2007:    /* FIGURE SPACE */
3978                  case 0x2008:    /* PUNCTUATION SPACE */
3979                  case 0x2009:    /* THIN SPACE */
3980                  case 0x200A:    /* HAIR SPACE */
3981                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3982                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3983                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3984                  gotspace = TRUE;
3985                  break;
3986                  }
3987                if (gotspace == (ctype == OP_NOT_HSPACE)) break;
3988                eptr += len;
3989                }
3990              break;
3991    
3992              case OP_NOT_VSPACE:
3993              case OP_VSPACE:
3994              for (i = min; i < max; i++)
3995                {
3996                BOOL gotspace;
3997                int len = 1;
3998                if (eptr >= md->end_subject) break;
3999                GETCHARLEN(c, eptr, len);
4000                switch(c)
4001                  {
4002                  default: gotspace = FALSE; break;
4003                  case 0x0a:      /* LF */
4004                  case 0x0b:      /* VT */
4005                  case 0x0c:      /* FF */
4006                  case 0x0d:      /* CR */
4007                  case 0x85:      /* NEL */
4008                  case 0x2028:    /* LINE SEPARATOR */
4009                  case 0x2029:    /* PARAGRAPH SEPARATOR */
4010                  gotspace = TRUE;
4011                  break;
4012                  }
4013                if (gotspace == (ctype == OP_NOT_VSPACE)) break;
4014                eptr += len;
4015                }
4016              break;
4017    
4018            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
4019            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4020              {              {
# Line 3202  for (;;) Line 4087  for (;;)
4087    
4088          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
4089    
4090            if (possessive) continue;
4091          for(;;)          for(;;)
4092            {            {
4093            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM46);
4094            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4095            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
4096            BACKCHAR(eptr);            BACKCHAR(eptr);
4097            }            }
4098          }          }
4099        else        else
4100  #endif  #endif  /* SUPPORT_UTF8 */
4101    
4102        /* Not UTF-8 mode */        /* Not UTF-8 mode */
4103          {          {
4104          switch(ctype)          switch(ctype)
4105            {            {
4106            case OP_ANY:            case OP_ANY:
4107            if ((ims & PCRE_DOTALL) == 0)            for (i = min; i < max; i++)
4108              {              {
4109              for (i = min; i < max; i++)              if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
4110                {              eptr++;
               if (eptr >= md->end_subject || *eptr == NEWLINE) break;  
               eptr++;  
               }  
             break;  
4111              }              }
4112            /* For DOTALL case, fall through and treat as \C */            break;
4113    
4114              case OP_ALLANY:
4115            case OP_ANYBYTE:            case OP_ANYBYTE:
4116            c = max - min;            c = max - min;
4117            if (c > md->end_subject - eptr) c = md->end_subject - eptr;            if (c > (unsigned int)(md->end_subject - eptr))
4118                c = md->end_subject - eptr;
4119            eptr += c;            eptr += c;
4120            break;            break;
4121    
4122              case OP_ANYNL:
4123              for (i = min; i < max; i++)
4124                {
4125                if (eptr >= md->end_subject) break;
4126                c = *eptr;
4127                if (c == 0x000d)
4128                  {
4129                  if (++eptr >= md->end_subject) break;
4130                  if (*eptr == 0x000a) eptr++;
4131                  }
4132                else
4133                  {
4134                  if (c != 0x000a &&
4135                      (md->bsr_anycrlf ||
4136                        (c != 0x000b && c != 0x000c && c != 0x0085)))
4137                    break;
4138                  eptr++;
4139                  }
4140                }
4141              break;
4142    
4143              case OP_NOT_HSPACE:
4144              for (i = min; i < max; i++)
4145                {
4146                if (eptr >= md->end_subject) break;
4147                c = *eptr;
4148                if (c == 0x09 || c == 0x20 || c == 0xa0) break;
4149                eptr++;
4150                }
4151              break;
4152    
4153              case OP_HSPACE:
4154              for (i = min; i < max; i++)
4155                {
4156                if (eptr >= md->end_subject) break;
4157                c = *eptr;
4158                if (c != 0x09 && c != 0x20 && c != 0xa0) break;
4159                eptr++;
4160                }
4161              break;
4162    
4163              case OP_NOT_VSPACE:
4164              for (i = min; i < max; i++)
4165                {
4166                if (eptr >= md->end_subject) break;
4167                c = *eptr;
4168                if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85)
4169                  break;
4170                eptr++;
4171                }
4172              break;
4173    
4174              case OP_VSPACE:
4175              for (i = min; i < max; i++)
4176                {
4177                if (eptr >= md->end_subject) break;
4178                c = *eptr;
4179                if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85)
4180                  break;
4181                eptr++;
4182                }
4183              break;
4184    
4185            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
4186            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4187              {              {
# Line 3295  for (;;) Line 4242  for (;;)
4242    
4243          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
4244    
4245            if (possessive) continue;
4246          while (eptr >= pp)          while (eptr >= pp)
4247            {            {
4248            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM47);
4249            eptr--;            eptr--;
4250            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4251            }            }
# Line 3309  for (;;) Line 4257  for (;;)
4257        }        }
4258      /* Control never gets here */      /* Control never gets here */
4259    
4260      /* There's been some horrible disaster. Since all codes > OP_BRA are      /* There's been some horrible disaster. Arrival here can only mean there is
4261      for capturing brackets, and there shouldn't be any gaps between 0 and      something seriously wrong in the code above or the OP_xxx definitions. */
     OP_BRA, arrival here can only mean there is something seriously wrong  
     in the code above or the OP_xxx definitions. */  
4262    
4263      default:      default:
4264      DPRINTF(("Unknown opcode %d\n", *ecode));      DPRINTF(("Unknown opcode %d\n", *ecode));
4265      RRETURN(PCRE_ERROR_UNKNOWN_NODE);      RRETURN(PCRE_ERROR_UNKNOWN_OPCODE);
4266      }      }
4267    
4268    /* Do not stick any code in here without much thought; it is assumed    /* Do not stick any code in here without much thought; it is assumed
# Line 3325  for (;;) Line 4271  for (;;)
4271    
4272    }             /* End of main loop */    }             /* End of main loop */
4273  /* Control never reaches here */  /* Control never reaches here */
4274    
4275    
4276    /* When compiling to use the heap rather than the stack for recursive calls to
4277    match(), the RRETURN() macro jumps here. The number that is saved in
4278    frame->Xwhere indicates which label we actually want to return to. */
4279    
4280    #ifdef NO_RECURSE
4281    #define LBL(val) case val: goto L_RM##val;
4282    HEAP_RETURN:
4283    switch (frame->Xwhere)
4284      {
4285      LBL( 1) LBL( 2) LBL( 3) LBL( 4) LBL( 5) LBL( 6) LBL( 7) LBL( 8)
4286      LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(17)
4287      LBL(19) LBL(24) LBL(25) LBL(26) LBL(27) LBL(29) LBL(31) LBL(33)
4288      LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52)
4289      LBL(53) LBL(54)
4290    #ifdef SUPPORT_UTF8
4291      LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30)
4292      LBL(32) LBL(34) LBL(42) LBL(46)
4293    #ifdef SUPPORT_UCP
4294      LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45)
4295    #endif  /* SUPPORT_UCP */
4296    #endif  /* SUPPORT_UTF8 */
4297      default:
4298      DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));
4299      return PCRE_ERROR_INTERNAL;
4300      }
4301    #undef LBL
4302    #endif  /* NO_RECURSE */
4303  }  }
4304    
4305    
# Line 3337  Undefine all the macros that were define Line 4312  Undefine all the macros that were define
4312  #ifdef NO_RECURSE  #ifdef NO_RECURSE
4313  #undef eptr  #undef eptr
4314  #undef ecode  #undef ecode
4315    #undef mstart
4316  #undef offset_top  #undef offset_top
4317  #undef ims  #undef ims
4318  #undef eptrb  #undef eptrb
# Line 3354  Undefine all the macros that were define Line 4330  Undefine all the macros that were define
4330    
4331  #undef cur_is_word  #undef cur_is_word
4332  #undef condition  #undef condition
 #undef minimize  
4333  #undef prev_is_word  #undef prev_is_word
4334    
4335  #undef original_ims  #undef original_ims
# Line 3410  Returns:          > 0 => success; value Line 4385  Returns:          > 0 => success; value
4385                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
4386  */  */
4387    
4388  PCRE_DATA_SCOPE int  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
4389  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
4390    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
4391    int offsetcount)    int offsetcount)
# Line 3419  int rc, resetcount, ocount; Line 4394  int rc, resetcount, ocount;
4394  int first_byte = -1;  int first_byte = -1;
4395  int req_byte = -1;  int req_byte = -1;
4396  int req_byte2 = -1;  int req_byte2 = -1;
4397  unsigned long int ims = 0;  int newline;
4398    unsigned long int ims;
4399  BOOL using_temporary_offsets = FALSE;  BOOL using_temporary_offsets = FALSE;
4400  BOOL anchored;  BOOL anchored;
4401  BOOL startline;  BOOL startline;
4402  BOOL firstline;  BOOL firstline;
4403  BOOL first_byte_caseless = FALSE;  BOOL first_byte_caseless = FALSE;
4404  BOOL req_byte_caseless = FALSE;  BOOL req_byte_caseless = FALSE;
4405    BOOL utf8;
4406  match_data match_block;  match_data match_block;
4407    match_data *md = &match_block;
4408  const uschar *tables;  const uschar *tables;
4409  const uschar *start_bits = NULL;  const uschar *start_bits = NULL;
4410  USPTR start_match = (USPTR)subject + start_offset;  USPTR start_match = (USPTR)subject + start_offset;
# Line 3451  if (offsetcount < 0) return PCRE_ERROR_B Line 4429  if (offsetcount < 0) return PCRE_ERROR_B
4429  the default values. */  the default values. */
4430    
4431  study = NULL;  study = NULL;
4432  match_block.match_limit = MATCH_LIMIT;  md->match_limit = MATCH_LIMIT;
4433  match_block.match_limit_recursion = MATCH_LIMIT_RECURSION;  md->match_limit_recursion = MATCH_LIMIT_RECURSION;
4434  match_block.callout_data = NULL;  md->callout_data = NULL;
4435    
4436  /* The table pointer is always in native byte order. */  /* The table pointer is always in native byte order. */
4437    
# Line 3465  if (extra_data != NULL) Line 4443  if (extra_data != NULL)
4443    if ((flags & PCRE_EXTRA_STUDY_DATA) != 0)    if ((flags & PCRE_EXTRA_STUDY_DATA) != 0)
4444      study = (const pcre_study_data *)extra_data->study_data;      study = (const pcre_study_data *)extra_data->study_data;
4445    if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0)    if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0)
4446      match_block.match_limit = extra_data->match_limit;      md->match_limit = extra_data->match_limit;
4447    if ((flags & PCRE_EXTRA_MATCH_LIMIT_RECURSION) != 0)    if ((flags & PCRE_EXTRA_MATCH_LIMIT_RECURSION) != 0)
4448      match_block.match_limit_recursion = extra_data->match_limit_recursion;      md->match_limit_recursion = extra_data->match_limit_recursion;
4449    if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0)    if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0)
4450      match_block.callout_data = extra_data->callout_data;      md->callout_data = extra_data->callout_data;
4451    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;
4452    }    }
4453    
# Line 3494  if (re->magic_number != MAGIC_NUMBER) Line 4472  if (re->magic_number != MAGIC_NUMBER)
4472  /* Set up other data */  /* Set up other data */
4473    
4474  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;
4475  startline = (re->options & PCRE_STARTLINE) != 0;  startline = (re->flags & PCRE_STARTLINE) != 0;
4476  firstline = (re->options & PCRE_FIRSTLINE) != 0;  firstline = (re->options & PCRE_FIRSTLINE) != 0;
4477    
4478  /* The code starts after the real_pcre block and the capture name table. */  /* The code starts after the real_pcre block and the capture name table. */
4479    
4480  match_block.start_code = (const uschar *)external_re + re->name_table_offset +  md->start_code = (const uschar *)external_re + re->name_table_offset +
4481    re->name_count * re->name_entry_size;    re->name_count * re->name_entry_size;
4482    
4483  match_block.start_subject = (USPTR)subject;  md->start_subject = (USPTR)subject;
4484  match_block.start_offset = start_offset;  md->start_offset = start_offset;
4485  match_block.end_subject = match_block.start_subject + length;  md->end_subject = md->start_subject + length;
4486  end_subject = match_block.end_subject;  end_subject = md->end_subject;
4487    
4488  match_block.endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;
4489  match_block.utf8 = (re->options & PCRE_UTF8) != 0;  utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;
4490    md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;
4491  match_block.notbol = (options & PCRE_NOTBOL) != 0;  
4492  match_block.noteol = (options & PCRE_NOTEOL) != 0;  md->notbol = (options & PCRE_NOTBOL) != 0;
4493  match_block.notempty = (options & PCRE_NOTEMPTY) != 0;  md->noteol = (options & PCRE_NOTEOL) != 0;
4494  match_block.partial = (options & PCRE_PARTIAL) != 0;  md->notempty = (options & PCRE_NOTEMPTY) != 0;
4495  match_block.hitend = FALSE;  md->partial = (options & PCRE_PARTIAL) != 0;
4496    md->hitend = FALSE;
4497    
4498  match_block.recursive = NULL;                   /* No recursion at top level */  md->recursive = NULL;                   /* No recursion at top level */
4499    
4500  match_block.lcc = tables + lcc_offset;  md->lcc = tables + lcc_offset;
4501  match_block.ctypes = tables + ctypes_offset;  md->ctypes = tables + ctypes_offset;
4502    
4503    /* Handle different \R options. */
4504    
4505    switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE))
4506      {
4507      case 0:
4508      if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0)
4509        md->bsr_anycrlf = (re->options & PCRE_BSR_ANYCRLF) != 0;
4510      else
4511    #ifdef BSR_ANYCRLF
4512      md->bsr_anycrlf = TRUE;
4513    #else
4514      md->bsr_anycrlf = FALSE;
4515    #endif
4516      break;
4517    
4518      case PCRE_BSR_ANYCRLF:
4519      md->bsr_anycrlf = TRUE;
4520      break;
4521    
4522      case PCRE_BSR_UNICODE:
4523      md->bsr_anycrlf = FALSE;
4524      break;
4525    
4526      default: return PCRE_ERROR_BADNEWLINE;
4527      }
4528    
4529    /* Handle different types of newline. The three bits give eight cases. If
4530    nothing is set at run time, whatever was used at compile time applies. */
4531    
4532    switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options :
4533            (pcre_uint32)options) & PCRE_NEWLINE_BITS)
4534      {
4535      case 0: newline = NEWLINE; break;   /* Compile-time default */
4536      case PCRE_NEWLINE_CR: newline = '\r'; break;
4537      case PCRE_NEWLINE_LF: newline = '\n'; break;
4538      case PCRE_NEWLINE_CR+
4539           PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;
4540      case PCRE_NEWLINE_ANY: newline = -1; break;
4541      case PCRE_NEWLINE_ANYCRLF: newline = -2; break;
4542      default: return PCRE_ERROR_BADNEWLINE;
4543      }
4544    
4545    if (newline == -2)
4546      {
4547      md->nltype = NLTYPE_ANYCRLF;
4548      }
4549    else if (newline < 0)
4550      {
4551      md->nltype = NLTYPE_ANY;
4552      }
4553    else
4554      {
4555      md->nltype = NLTYPE_FIXED;
4556      if (newline > 255)
4557        {
4558        md->nllen = 2;
4559        md->nl[0] = (newline >> 8) & 255;
4560        md->nl[1] = newline & 255;
4561        }
4562      else
4563        {
4564        md->nllen = 1;
4565        md->nl[0] = newline;
4566        }
4567      }
4568    
4569  /* Partial matching is supported only for a restricted set of regexes at the  /* Partial matching is supported only for a restricted set of regexes at the
4570  moment. */  moment. */
4571    
4572  if (match_block.partial && (re->options & PCRE_NOPARTIAL) != 0)  if (md->partial && (re->flags & PCRE_NOPARTIAL) != 0)
4573    return PCRE_ERROR_BADPARTIAL;    return PCRE_ERROR_BADPARTIAL;
4574    
4575  /* Check a UTF-8 string if required. Unfortunately there's no way of passing  /* Check a UTF-8 string if required. Unfortunately there's no way of passing
4576  back the character offset. */  back the character offset. */
4577    
4578  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4579  if (match_block.utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)
4580    {    {
4581    if (_pcre_valid_utf8((uschar *)subject, length) >= 0)    if (_pcre_valid_utf8((uschar *)subject, length) >= 0)
4582      return PCRE_ERROR_BADUTF8;      return PCRE_ERROR_BADUTF8;
# Line 3563  ocount = offsetcount - (offsetcount % 3) Line 4608  ocount = offsetcount - (offsetcount % 3)
4608  if (re->top_backref > 0 && re->top_backref >= ocount/3)  if (re->top_backref > 0 && re->top_backref >= ocount/3)
4609    {    {
4610    ocount = re->top_backref * 3 + 3;    ocount = re->top_backref * 3 + 3;
4611    match_block.offset_vector = (int *)(pcre_malloc)(ocount * sizeof(int));    md->offset_vector = (int *)(pcre_malloc)(ocount * sizeof(int));
4612    if (match_block.offset_vector == NULL) return PCRE_ERROR_NOMEMORY;    if (md->offset_vector == NULL) return PCRE_ERROR_NOMEMORY;
4613    using_temporary_offsets = TRUE;    using_temporary_offsets = TRUE;
4614    DPRINTF(("Got memory to hold back references\n"));    DPRINTF(("Got memory to hold back references\n"));
4615    }    }
4616  else match_block.offset_vector = offsets;  else md->offset_vector = offsets;
4617    
4618  match_block.offset_end = ocount;  md->offset_end = ocount;
4619  match_block.offset_max = (2*ocount)/3;  md->offset_max = (2*ocount)/3;
4620  match_block.offset_overflow = FALSE;  md->offset_overflow = FALSE;
4621  match_block.capture_last = -1;  md->capture_last = -1;
4622    
4623  /* Compute the minimum number of offsets that we need to reset each time. Doing  /* Compute the minimum number of offsets that we need to reset each time. Doing
4624  this makes a huge difference to execution time when there aren't many brackets  this makes a huge difference to execution time when there aren't many brackets
# Line 3586  if (resetcount > offsetcount) resetcount Line 4631  if (resetcount > offsetcount) resetcount
4631  never be used unless previously set, but they get saved and restored, and so we  never be used unless previously set, but they get saved and restored, and so we
4632  initialize them to avoid reading uninitialized locations. */  initialize them to avoid reading uninitialized locations. */
4633    
4634  if (match_block.offset_vector != NULL)  if (md->offset_vector != NULL)
4635    {    {
4636    register int *iptr = match_block.offset_vector + ocount;    register int *iptr = md->offset_vector + ocount;
4637    register int *iend = iptr - resetcount/2 + 1;    register int *iend = iptr - resetcount/2 + 1;
4638    while (--iptr >= iend) *iptr = -1;    while (--iptr >= iend) *iptr = -1;
4639    }    }
# Line 3601  studied, there may be a bitmap of possib Line 4646  studied, there may be a bitmap of possib
4646    
4647  if (!anchored)  if (!anchored)
4648    {    {
4649    if ((re->options & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
4650      {      {
4651      first_byte = re->first_byte & 255;      first_byte = re->first_byte & 255;
4652      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)
4653        first_byte = match_block.lcc[first_byte];        first_byte = md->lcc[first_byte];
4654      }      }
4655    else    else
4656      if (!startline && study != NULL &&      if (!startline && study != NULL &&
# Line 3616  if (!anchored) Line 4661  if (!anchored)
4661  /* For anchored or unanchored matches, there may be a "last known required  /* For anchored or unanchored matches, there may be a "last known required
4662  character" set. */  character" set. */
4663    
4664  if ((re->options & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
4665    {    {
4666    req_byte = re->req_byte & 255;    req_byte = re->req_byte & 255;
4667    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;
4668    req_byte2 = (tables + fcc_offset)[req_byte];  /* case flipped */    req_byte2 = (tables + fcc_offset)[req_byte];  /* case flipped */
4669    }    }
4670    
4671    
4672    /* ==========================================================================*/
4673    
4674  /* Loop for handling unanchored repeated matching attempts; for anchored regexs  /* Loop for handling unanchored repeated matching attempts; for anchored regexs
4675  the loop runs just once. */  the loop runs just once. */
4676    
4677  do  for(;;)
4678    {    {
4679    USPTR save_end_subject = end_subject;    USPTR save_end_subject = end_subject;
4680      USPTR new_start_match;
4681    
4682    /* Reset the maximum number of extractions we might see. */    /* Reset the maximum number of extractions we might see. */
4683    
4684    if (match_block.offset_vector != NULL)    if (md->offset_vector != NULL)
4685      {      {
4686      register int *iptr = match_block.offset_vector;      register int *iptr = md->offset_vector;
4687      register int *iend = iptr + resetcount;      register int *iend = iptr + resetcount;
4688      while (iptr < iend) *iptr++ = -1;      while (iptr < iend) *iptr++ = -1;
4689      }      }
4690    
4691    /* Advance to a unique first char if possible. If firstline is TRUE, the    /* Advance to a unique first char if possible. If firstline is TRUE, the
4692    start of the match is constrained to the first line of a multiline string.    start of the match is constrained to the first line of a multiline string.
4693    Implement this by temporarily adjusting end_subject so that we stop scanning    That is, the match must be before or at the first newline. Implement this by
4694    at a newline. If the match fails at the newline, later code breaks this loop.    temporarily adjusting end_subject so that we stop scanning at a newline. If
4695    */    the match fails at the newline, later code breaks this loop. */
4696    
4697    if (firstline)    if (firstline)
4698      {      {
4699      USPTR t = start_match;      USPTR t = start_match;
4700      while (t < save_end_subject && *t != '\n') t++;  #ifdef SUPPORT_UTF8
4701        if (utf8)
4702          {
4703          while (t < md->end_subject && !IS_NEWLINE(t))
4704            {
4705            t++;
4706            while (t < end_subject && (*t & 0xc0) == 0x80) t++;
4707            }
4708          }
4709        else
4710    #endif
4711        while (t < md->end_subject && !IS_NEWLINE(t)) t++;
4712      end_subject = t;      end_subject = t;
4713      }      }
4714    
4715    /* Now test for a unique first byte */    /* Now advance to a unique first byte if there is one. */
4716    
4717    if (first_byte >= 0)    if (first_byte >= 0)
4718      {      {
4719      if (first_byte_caseless)      if (first_byte_caseless)
4720        while (start_match < end_subject &&        while (start_match < end_subject && md->lcc[*start_match] != first_byte)
              match_block.lcc[*start_match] != first_byte)  
4721          start_match++;          start_match++;
4722      else      else
4723        while (start_match < end_subject && *start_match != first_byte)        while (start_match < end_subject && *start_match != first_byte)
4724          start_match++;          start_match++;
4725      }      }
4726    
4727    /* Or to just after \n for a multiline match if possible */    /* Or to just after a linebreak for a multiline match */
4728    
4729    else if (startline)    else if (startline)
4730      {      {
4731      if (start_match > match_block.start_subject + start_offset)      if (start_match > md->start_subject + start_offset)
4732        {        {
4733        while (start_match < end_subject && start_match[-1] != NEWLINE)  #ifdef SUPPORT_UTF8
4734          if (utf8)
4735            {
4736            while (start_match < end_subject && !WAS_NEWLINE(start_match))
4737              {
4738              start_match++;
4739              while(start_match < end_subject && (*start_match & 0xc0) == 0x80)
4740                start_match++;
4741              }
4742            }
4743          else
4744    #endif
4745          while (start_match < end_subject && !WAS_NEWLINE(start_match))
4746            start_match++;
4747    
4748          /* If we have just passed a CR and the newline option is ANY or ANYCRLF,
4749          and we are now at a LF, advance the match position by one more character.
4750          */
4751    
4752          if (start_match[-1] == '\r' &&
4753               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
4754               start_match < end_subject &&
4755               *start_match == '\n')
4756          start_match++;          start_match++;
4757        }        }
4758      }      }
4759    
4760    /* Or to a non-unique first char after study */    /* Or to a non-unique first byte after study */
4761    
4762    else if (start_bits != NULL)    else if (start_bits != NULL)
4763      {      {
4764      while (start_match < end_subject)      while (start_match < end_subject)
4765        {        {
4766        register unsigned int c = *start_match;        register unsigned int c = *start_match;
4767        if ((start_bits[c/8] & (1 << (c&7))) == 0) start_match++; else break;        if ((start_bits[c/8] & (1 << (c&7))) == 0) start_match++;
4768            else break;
4769        }        }
4770      }      }
4771    
# Line 3693  do Line 4775  do
4775    
4776  #ifdef DEBUG  /* Sigh. Some compilers never learn. */  #ifdef DEBUG  /* Sigh. Some compilers never learn. */
4777    printf(">>>> Match against: ");    printf(">>>> Match against: ");
4778    pchars(start_match, end_subject - start_match, TRUE, &match_block);    pchars(start_match, end_subject - start_match, TRUE, md);
4779    printf("\n");    printf("\n");
4780  #endif  #endif
4781    
# Line 3707  do Line 4789  do
4789    
4790    HOWEVER: when the subject string is very, very long, searching to its end can    HOWEVER: when the subject string is very, very long, searching to its end can
4791    take a long time, and give bad performance on quite ordinary patterns. This    take a long time, and give bad performance on quite ordinary patterns. This
4792    showed up when somebody was matching /^C/ on a 32-megabyte string... so we    showed up when somebody was matching something like /^\d+C/ on a 32-megabyte
4793    don't do this when the string is sufficiently long.    string... so we don't do this when the string is sufficiently long.
4794    
4795    ALSO: this processing is disabled when partial matching is requested.    ALSO: this processing is disabled when partial matching is requested.
4796    */    */
4797    
4798    if (req_byte >= 0 &&    if (req_byte >= 0 &&
4799        end_subject - start_match < REQ_BYTE_MAX &&        end_subject - start_match < REQ_BYTE_MAX &&
4800        !match_block.partial)        !md->partial)
4801      {      {
4802      register USPTR p = start_match + ((first_byte >= 0)? 1 : 0);      register USPTR p = start_match + ((first_byte >= 0)? 1 : 0);
4803    
# Line 3740  do Line 4822  do
4822            }            }
4823          }          }
4824    
4825        /* If we can't find the required character, break the matching loop */        /* If we can't find the required character, break the matching loop,
4826          forcing a match failure. */
4827    
4828        if (p >= end_subject) break;        if (p >= end_subject)
4829            {
4830            rc = MATCH_NOMATCH;
4831            break;
4832            }
4833    
4834        /* If we have found the required character, save the point where we        /* If we have found the required character, save the point where we
4835        found it, so that we don't search again next time round the loop if        found it, so that we don't search again next time round the loop if
# Line 3752  do Line 4839  do
4839        }        }
4840      }      }
4841    
4842    /* When a match occurs, substrings will be set for all internal extractions;    /* OK, we can now run the match. */
4843    we just need to set up the whole thing as substring 0 before returning. If  
4844    there were too many extractions, set the return code to zero. In the case    md->start_match_ptr = start_match;
4845    where we had to get some local store to hold offsets for backreferences, copy    md->match_call_count = 0;
4846    those back references that we can. In this case there need not be overflow    rc = match(start_match, md->start_code, start_match, 2, md, ims, NULL, 0, 0);
   if certain parts of the pattern were not used. */  
   
   match_block.start_match = start_match;  
   match_block.match_call_count = 0;  
   
   rc = match(start_match, match_block.start_code, 2, &match_block, ims, NULL,  
     match_isgroup, 0);  
   
   /* When the result is no match, if the subject's first character was a  
   newline and the PCRE_FIRSTLINE option is set, break (which will return  
   PCRE_ERROR_NOMATCH). The option requests that a match occur before the first  
   newline in the subject. Otherwise, advance the pointer to the next character  
   and continue - but the continuation will actually happen only when the  
   pattern is not anchored. */  
4847    
4848    if (rc == MATCH_NOMATCH)    switch(rc)
4849      {      {
4850      if (firstline && *start_match == NEWLINE) break;      /* NOMATCH and PRUNE advance by one character. THEN at this level acts
4851      start_match++;      exactly like PRUNE. */
4852    
4853        case MATCH_NOMATCH:
4854        case MATCH_PRUNE:
4855 &nb