/[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 91 by nigel, Sat Feb 24 21:41:34 2007 UTC revision 294 by ph10, Sat Dec 29 16:55:59 2007 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) 1997-2007 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  #define NLBLOCK md           /* The block containing newline information */  #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 186  calls by keeping local variables that ne Line 194  calls by keeping local variables that ne
194  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
195  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
196  always used to.  always used to.
197    
198    The original heap-recursive code used longjmp(). However, it seems that this
199    can be very slow on some operating systems. Following a suggestion from Stan
200    Switzer, the use of longjmp() has been abolished, at the cost of having to
201    provide a unique number for each call to RMATCH. There is no way of generating
202    a sequence of numbers at compile time in C. I have given them names, to make
203    them stand out more clearly.
204    
205    Crude tests on x86 Linux show a small speedup of around 5-8%. However, on
206    FreeBSD, avoiding longjmp() more than halves the time taken to run the standard
207    tests. Furthermore, not using longjmp() means that local dynamic variables
208    don't have indeterminate values; this has meant that the frame size can be
209    reduced because the result can be "passed back" by straight setting of the
210    variable instead of being passed in the frame.
211  ****************************************************************************  ****************************************************************************
212  ***************************************************************************/  ***************************************************************************/
213    
214    /* Numbers for RMATCH calls. When this list is changed, the code at HEAP_RETURN
215    below must be updated in sync.  */
216    
217    enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM6,  RM7,  RM8,  RM9,  RM10,
218           RM11,  RM12, RM13, RM14, RM15, RM16, RM17, RM18, RM19, RM20,
219           RM21,  RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30,
220           RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
221           RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
222           RM51,  RM52, RM53, RM54 };
223    
224  /* 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
225  versions and production versions. */  versions and production versions. Note that the "rw" argument of RMATCH isn't
226    actuall used in this definition. */
227    
228  #ifndef NO_RECURSE  #ifndef NO_RECURSE
229  #define REGISTER register  #define REGISTER register
230    
231  #ifdef DEBUG  #ifdef DEBUG
232  #define RMATCH(rx,ra,rb,rc,rd,re,rf,rg) \  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \
233    { \    { \
234    printf("match() called in line %d\n", __LINE__); \    printf("match() called in line %d\n", __LINE__); \
235    rx = match(ra,rb,rc,rd,re,rf,rg,rdepth+1); \    rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1); \
236    printf("to line %d\n", __LINE__); \    printf("to line %d\n", __LINE__); \
237    }    }
238  #define RRETURN(ra) \  #define RRETURN(ra) \
# Line 208  versions and production versions. */ Line 241  versions and production versions. */
241    return ra; \    return ra; \
242    }    }
243  #else  #else
244  #define RMATCH(rx,ra,rb,rc,rd,re,rf,rg) \  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \
245    rx = match(ra,rb,rc,rd,re,rf,rg,rdepth+1)    rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1)
246  #define RRETURN(ra) return ra  #define RRETURN(ra) return ra
247  #endif  #endif
248    
249  #else  #else
250    
251    
252  /* 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
253  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
254  match(), which never changes. */  argument of match(), which never changes. */
255    
256  #define REGISTER  #define REGISTER
257    
258  #define RMATCH(rx,ra,rb,rc,rd,re,rf,rg)\  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw)\
259    {\    {\
260    heapframe *newframe = (pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = (pcre_stack_malloc)(sizeof(heapframe));\
261    if (setjmp(frame->Xwhere) == 0)\    frame->Xwhere = rw; \
262      {\    newframe->Xeptr = ra;\
263      newframe->Xeptr = ra;\    newframe->Xecode = rb;\
264      newframe->Xecode = rb;\    newframe->Xmstart = mstart;\
265      newframe->Xoffset_top = rc;\    newframe->Xoffset_top = rc;\
266      newframe->Xims = re;\    newframe->Xims = re;\
267      newframe->Xeptrb = rf;\    newframe->Xeptrb = rf;\
268      newframe->Xflags = rg;\    newframe->Xflags = rg;\
269      newframe->Xrdepth = frame->Xrdepth + 1;\    newframe->Xrdepth = frame->Xrdepth + 1;\
270      newframe->Xprevframe = frame;\    newframe->Xprevframe = frame;\
271      frame = newframe;\    frame = newframe;\
272      DPRINTF(("restarting from line %d\n", __LINE__));\    DPRINTF(("restarting from line %d\n", __LINE__));\
273      goto HEAP_RECURSE;\    goto HEAP_RECURSE;\
274      }\    L_##rw:\
275    else\    DPRINTF(("jumped back to line %d\n", __LINE__));\
     {\  
     DPRINTF(("longjumped back to line %d\n", __LINE__));\  
     frame = md->thisframe;\  
     rx = frame->Xresult;\  
     }\  
276    }    }
277    
278  #define RRETURN(ra)\  #define RRETURN(ra)\
# Line 254  match(), which never changes. */ Line 282  match(), which never changes. */
282    (pcre_stack_free)(newframe);\    (pcre_stack_free)(newframe);\
283    if (frame != NULL)\    if (frame != NULL)\
284      {\      {\
285      frame->Xresult = ra;\      rrc = ra;\
286      md->thisframe = frame;\      goto HEAP_RETURN;\
     longjmp(frame->Xwhere, 1);\  
287      }\      }\
288    return ra;\    return ra;\
289    }    }
# Line 271  typedef struct heapframe { Line 298  typedef struct heapframe {
298    
299    const uschar *Xeptr;    const uschar *Xeptr;
300    const uschar *Xecode;    const uschar *Xecode;
301      const uschar *Xmstart;
302    int Xoffset_top;    int Xoffset_top;
303    long int Xims;    long int Xims;
304    eptrblock *Xeptrb;    eptrblock *Xeptrb;
# Line 291  typedef struct heapframe { Line 319  typedef struct heapframe {
319    
320    BOOL Xcur_is_word;    BOOL Xcur_is_word;
321    BOOL Xcondition;    BOOL Xcondition;
   BOOL Xminimize;  
322    BOOL Xprev_is_word;    BOOL Xprev_is_word;
323    
324    unsigned long int Xoriginal_ims;    unsigned long int Xoriginal_ims;
# Line 303  typedef struct heapframe { Line 330  typedef struct heapframe {
330    int Xprop_category;    int Xprop_category;
331    int Xprop_chartype;    int Xprop_chartype;
332    int Xprop_script;    int Xprop_script;
333    int *Xprop_test_variable;    int Xoclength;
334      uschar Xocchars[8];
335  #endif  #endif
336    
337    int Xctype;    int Xctype;
338    int Xfc;    unsigned int Xfc;
339    int Xfi;    int Xfi;
340    int Xlength;    int Xlength;
341    int Xmax;    int Xmax;
# Line 321  typedef struct heapframe { Line 349  typedef struct heapframe {
349    
350    eptrblock Xnewptrb;    eptrblock Xnewptrb;
351    
352    /* Place to pass back result, and where to jump back to */    /* Where to jump back to */
353    
354    int  Xresult;    int Xwhere;
   jmp_buf Xwhere;  
355    
356  } heapframe;  } heapframe;
357    
# Line 340  typedef struct heapframe { Line 367  typedef struct heapframe {
367  *         Match from current position            *  *         Match from current position            *
368  *************************************************/  *************************************************/
369    
370  /* 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  
371  returns a negative (error) response, the outer incarnation must also return the  returns a negative (error) response, the outer incarnation must also return the
372  same response.  same response.
373    
# Line 353  performance. Tests using gcc on a SPARC Line 377  performance. Tests using gcc on a SPARC
377  made performance worse.  made performance worse.
378    
379  Arguments:  Arguments:
380     eptr        pointer in subject     eptr        pointer to current character in subject
381     ecode       position in code     ecode       pointer to current position in compiled code
382       mstart      pointer to the current match start position (can be modified
383                     by encountering \K)
384     offset_top  current top pointer     offset_top  current top pointer
385     md          pointer to "static" info for the match     md          pointer to "static" info for the match
386     ims         current /i, /m, and /s options     ims         current /i, /m, and /s options
# Line 362  Arguments: Line 388  Arguments:
388                   brackets - for testing for empty matches                   brackets - for testing for empty matches
389     flags       can contain     flags       can contain
390                   match_condassert - this is an assertion condition                   match_condassert - this is an assertion condition
391                   match_isgroup - this is the start of a bracketed group                   match_cbegroup - this is the start of an unlimited repeat
392                       group that can match an empty string
393     rdepth      the recursion depth     rdepth      the recursion depth
394    
395  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 399  Returns:       MATCH_MATCH if matched
399  */  */
400    
401  static int  static int
402  match(REGISTER USPTR eptr, REGISTER const uschar *ecode,  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, const uschar *mstart,
403    int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb,    int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb,
404    int flags, unsigned int rdepth)    int flags, unsigned int rdepth)
405  {  {
406  /* 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,
407  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
408  because they are used a lot in loops. */  "register" because they are used a lot in loops. */
409    
410  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
411  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
412  register unsigned int  c;  /* Character values not kept over RMATCH() calls */  register unsigned int c;   /* Character values not kept over RMATCH() calls */
413  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */
414    
415    BOOL minimize, possessive; /* Quantifier options */
416    
417  /* 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
418  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
419  heap storage. Set up the top-level frame here; others are obtained from the  heap storage. Set up the top-level frame here; others are obtained from the
# Line 398  frame->Xprevframe = NULL;            /* Line 427  frame->Xprevframe = NULL;            /*
427    
428  frame->Xeptr = eptr;  frame->Xeptr = eptr;
429  frame->Xecode = ecode;  frame->Xecode = ecode;
430    frame->Xmstart = mstart;
431  frame->Xoffset_top = offset_top;  frame->Xoffset_top = offset_top;
432  frame->Xims = ims;  frame->Xims = ims;
433  frame->Xeptrb = eptrb;  frame->Xeptrb = eptrb;
# Line 412  HEAP_RECURSE: Line 442  HEAP_RECURSE:
442    
443  #define eptr               frame->Xeptr  #define eptr               frame->Xeptr
444  #define ecode              frame->Xecode  #define ecode              frame->Xecode
445    #define mstart             frame->Xmstart
446  #define offset_top         frame->Xoffset_top  #define offset_top         frame->Xoffset_top
447  #define ims                frame->Xims  #define ims                frame->Xims
448  #define eptrb              frame->Xeptrb  #define eptrb              frame->Xeptrb
# Line 434  HEAP_RECURSE: Line 465  HEAP_RECURSE:
465    
466  #define cur_is_word        frame->Xcur_is_word  #define cur_is_word        frame->Xcur_is_word
467  #define condition          frame->Xcondition  #define condition          frame->Xcondition
 #define minimize           frame->Xminimize  
468  #define prev_is_word       frame->Xprev_is_word  #define prev_is_word       frame->Xprev_is_word
469    
470  #define original_ims       frame->Xoriginal_ims  #define original_ims       frame->Xoriginal_ims
# Line 446  HEAP_RECURSE: Line 476  HEAP_RECURSE:
476  #define prop_category      frame->Xprop_category  #define prop_category      frame->Xprop_category
477  #define prop_chartype      frame->Xprop_chartype  #define prop_chartype      frame->Xprop_chartype
478  #define prop_script        frame->Xprop_script  #define prop_script        frame->Xprop_script
479  #define prop_test_variable frame->Xprop_test_variable  #define oclength           frame->Xoclength
480    #define occhars            frame->Xocchars
481  #endif  #endif
482    
483  #define ctype              frame->Xctype  #define ctype              frame->Xctype
# Line 470  HEAP_RECURSE: Line 501  HEAP_RECURSE:
501  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
502  i, and fc and c, can be the same variables. */  i, and fc and c, can be the same variables. */
503    
504  #else  #else         /* NO_RECURSE not defined */
505  #define fi i  #define fi i
506  #define fc c  #define fc c
507    
# Line 489  recursion_info new_recursive;      /* wi Line 520  recursion_info new_recursive;      /* wi
520                                     /* that do not have to be preserved over  */                                     /* that do not have to be preserved over  */
521  BOOL cur_is_word;                  /* a recursive call to RMATCH().          */  BOOL cur_is_word;                  /* a recursive call to RMATCH().          */
522  BOOL condition;  BOOL condition;
 BOOL minimize;  
523  BOOL prev_is_word;  BOOL prev_is_word;
524    
525  unsigned long int original_ims;  unsigned long int original_ims;
# Line 501  int prop_fail_result; Line 531  int prop_fail_result;
531  int prop_category;  int prop_category;
532  int prop_chartype;  int prop_chartype;
533  int prop_script;  int prop_script;
534  int *prop_test_variable;  int oclength;
535    uschar occhars[8];
536  #endif  #endif
537    
538  int ctype;  int ctype;
# Line 516  int save_offset1, save_offset2, save_off Line 547  int save_offset1, save_offset2, save_off
547  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
548    
549  eptrblock newptrb;  eptrblock newptrb;
550  #endif  #endif     /* NO_RECURSE */
551    
552  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
553  variables. */  variables. */
# Line 524  variables. */ Line 555  variables. */
555  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
556  prop_value = 0;  prop_value = 0;
557  prop_fail_result = 0;  prop_fail_result = 0;
 prop_test_variable = NULL;  
558  #endif  #endif
559    
560    
561  /* This label is used for tail recursion, which is used in a few cases even  /* This label is used for tail recursion, which is used in a few cases even
562  when NO_RECURSE is not defined, in order to reduce the amount of stack that is  when NO_RECURSE is not defined, in order to reduce the amount of stack that is
563  used. Thanks to Ian Taylor for noticing this possibility and sending the  used. Thanks to Ian Taylor for noticing this possibility and sending the
# Line 542  defined). However, RMATCH isn't like a f Line 573  defined). However, RMATCH isn't like a f
573  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,
574  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
575    
576    #ifdef SUPPORT_UTF8
577    utf8 = md->utf8;       /* Local copy of the flag */
578    #else
579    utf8 = FALSE;
580    #endif
581    
582  /* 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
583  haven't exceeded the recursive call limit. */  haven't exceeded the recursive call limit. */
584    
# Line 550  if (rdepth >= md->match_limit_recursion) Line 587  if (rdepth >= md->match_limit_recursion)
587    
588  original_ims = ims;    /* Save for resetting on ')' */  original_ims = ims;    /* Save for resetting on ')' */
589    
590  #ifdef SUPPORT_UTF8  /* At the start of a group with an unlimited repeat that may match an empty
591  utf8 = md->utf8;       /* Local copy of the flag */  string, the match_cbegroup flag is set. When this is the case, add the current
592  #else  subject pointer to the chain of such remembered pointers, to be checked when we
593  utf8 = FALSE;  hit the closing ket, in order to break infinite loops that match no characters.
594  #endif  When match() is called in other circumstances, don't add to the chain. The
595    match_cbegroup flag must NOT be used with tail recursion, because the memory
596  /* At the start of a bracketed group, add the current subject pointer to the  block that is used is on the stack, so a new one may be required for each
597  stack of such pointers, to be re-instated at the end of the group when we hit  match(). */
 the closing ket. When match() is called in other circumstances, we don't add to  
 this stack. */  
598    
599  if ((flags & match_isgroup) != 0)  if ((flags & match_cbegroup) != 0)
600    {    {
   newptrb.epb_prev = eptrb;  
601    newptrb.epb_saved_eptr = eptr;    newptrb.epb_saved_eptr = eptr;
602      newptrb.epb_prev = eptrb;
603    eptrb = &newptrb;    eptrb = &newptrb;
604    }    }
605    
606  /* Now start processing the operations. */  /* Now start processing the opcodes. */
607    
608  for (;;)  for (;;)
609    {    {
610      minimize = possessive = FALSE;
611    op = *ecode;    op = *ecode;
   minimize = FALSE;  
612    
613    /* 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
614    matching at least one subject character. */    matching at least one subject character. */
615    
616    if (md->partial &&    if (md->partial &&
617        eptr >= md->end_subject &&        eptr >= md->end_subject &&
618        eptr > md->start_match)        eptr > mstart)
619      md->hitend = TRUE;      md->hitend = TRUE;
620    
621    /* 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)  
622      {      {
623      number = op - OP_BRA;      case OP_FAIL:
624        RRETURN(MATCH_NOMATCH);
625    
626      /* For extended extraction brackets (large number), we have to fish out the      case OP_PRUNE:
627      number from a dummy opcode at the start. */      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
628          ims, eptrb, flags, RM51);
629      if (number > EXTRACT_BASIC_MAX)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
630        number = GET2(ecode, 2+LINK_SIZE);      RRETURN(MATCH_PRUNE);
631    
632        case OP_COMMIT:
633        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
634          ims, eptrb, flags, RM52);
635        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
636        RRETURN(MATCH_COMMIT);
637    
638        case OP_SKIP:
639        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
640          ims, eptrb, flags, RM53);
641        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
642        md->start_match_ptr = eptr;   /* Pass back current position */
643        RRETURN(MATCH_SKIP);
644    
645        case OP_THEN:
646        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
647          ims, eptrb, flags, RM54);
648        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
649        RRETURN(MATCH_THEN);
650    
651        /* Handle a capturing bracket. If there is space in the offset vector, save
652        the current subject position in the working slot at the top of the vector.
653        We mustn't change the current values of the data slot, because they may be
654        set from a previous iteration of this group, and be referred to by a
655        reference inside the group.
656    
657        If the bracket fails to match, we need to restore this value and also the
658        values of the final offsets, in case they were set by a previous iteration
659        of the same bracket.
660    
661        If there isn't enough space in the offset vector, treat this as if it were
662        a non-capturing bracket. Don't worry about setting the flag for the error
663        case here; that is handled in the code for KET. */
664    
665        case OP_CBRA:
666        case OP_SCBRA:
667        number = GET2(ecode, 1+LINK_SIZE);
668      offset = number << 1;      offset = number << 1;
669    
670  #ifdef DEBUG  #ifdef DEBUG
671      printf("start bracket %d subject=", number);      printf("start bracket %d\n", number);
672        printf("subject=");
673      pchars(eptr, 16, TRUE, md);      pchars(eptr, 16, TRUE, md);
674      printf("\n");      printf("\n");
675  #endif  #endif
# Line 624  for (;;) Line 684  for (;;)
684        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
685        md->offset_vector[md->offset_end - number] = eptr - md->start_subject;        md->offset_vector[md->offset_end - number] = eptr - md->start_subject;
686    
687          flags = (op == OP_SCBRA)? match_cbegroup : 0;
688        do        do
689          {          {
690          RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb,          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
691            match_isgroup);            ims, eptrb, flags, RM1);
692          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
693          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
694          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
695          }          }
# Line 643  for (;;) Line 704  for (;;)
704        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
705        }        }
706    
707      /* Insufficient room for saving captured contents */      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
708        as a non-capturing bracket. */
     else op = OP_BRA;  
     }  
   
   /* Other types of node can be handled by a switch */  
709    
710    switch(op)      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
711      {      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
     case OP_BRA:     /* Non-capturing bracket: optimized */  
     DPRINTF(("start bracket 0\n"));  
712    
713      /* Loop for all the alternatives */      DPRINTF(("insufficient capture room: treat as non-capturing\n"));
714    
715        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
716        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
717    
718        /* Non-capturing bracket. Loop for all the alternatives. When we get to the
719        final alternative within the brackets, we would return the result of a
720        recursive call to match() whatever happened. We can reduce stack usage by
721        turning this into a tail recursion, except in the case when match_cbegroup
722        is set.*/
723    
724        case OP_BRA:
725        case OP_SBRA:
726        DPRINTF(("start non-capturing bracket\n"));
727        flags = (op >= OP_SBRA)? match_cbegroup : 0;
728      for (;;)      for (;;)
729        {        {
730        /* When we get to the final alternative within the brackets, we would        if (ecode[GET(ecode, 1)] != OP_ALT)   /* Final alternative */
731        return the result of a recursive call to match() whatever happened. We          {
732        can reduce stack usage by turning this into a tail recursion. */          if (flags == 0)    /* Not a possibly empty group */
733              {
734        if (ecode[GET(ecode, 1)] != OP_ALT)            ecode += _pcre_OP_lengths[*ecode];
735         {            DPRINTF(("bracket 0 tail recursion\n"));
736         ecode += 1 + LINK_SIZE;            goto TAIL_RECURSE;
737         flags = match_isgroup;            }
738         DPRINTF(("bracket 0 tail recursion\n"));  
739         goto TAIL_RECURSE;          /* Possibly empty group; can't use tail recursion. */
740         }  
741            RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
742              eptrb, flags, RM48);
743            RRETURN(rrc);
744            }
745    
746        /* For non-final alternatives, continue the loop for a NOMATCH result;        /* For non-final alternatives, continue the loop for a NOMATCH result;
747        otherwise return. */        otherwise return. */
748    
749        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb,        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
750          match_isgroup);          eptrb, flags, RM2);
751        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
752        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
753        }        }
754      /* Control never reaches here. */      /* Control never reaches here. */
# Line 688  for (;;) Line 760  for (;;)
760      obeyed, we can use tail recursion to avoid using another stack frame. */      obeyed, we can use tail recursion to avoid using another stack frame. */
761    
762      case OP_COND:      case OP_COND:
763      if (ecode[LINK_SIZE+1] == OP_CREF) /* Condition extract or recurse test */      case OP_SCOND:
764        if (ecode[LINK_SIZE+1] == OP_RREF)         /* Recursion test */
765          {
766          offset = GET2(ecode, LINK_SIZE + 2);     /* Recursion group number*/
767          condition = md->recursive != NULL &&
768            (offset == RREF_ANY || offset == md->recursive->group_num);
769          ecode += condition? 3 : GET(ecode, 1);
770          }
771    
772        else if (ecode[LINK_SIZE+1] == OP_CREF)    /* Group used test */
773        {        {
774        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */
775        condition = (offset == CREF_RECURSE * 2)?        condition = offset < offset_top && md->offset_vector[offset] >= 0;
776          (md->recursive != NULL) :        ecode += condition? 3 : GET(ecode, 1);
777          (offset < offset_top && md->offset_vector[offset] >= 0);        }
778        ecode += condition? (LINK_SIZE + 4) : (LINK_SIZE + 1 + GET(ecode, 1));  
779        flags = match_isgroup;      else if (ecode[LINK_SIZE+1] == OP_DEF)     /* DEFINE - always false */
780        goto TAIL_RECURSE;        {
781          condition = FALSE;
782          ecode += GET(ecode, 1);
783        }        }
784    
785      /* The condition is an assertion. Call match() to evaluate it - setting      /* The condition is an assertion. Call match() to evaluate it - setting
786      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
787        assertion. */
788    
789      else      else
790        {        {
791        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,
792            match_condassert | match_isgroup);            match_condassert, RM3);
793        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
794          {          {
795          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE+2);          condition = TRUE;
796            ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
797          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
798          }          }
799        else if (rrc != MATCH_NOMATCH)        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
800          {          {
801          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
802          }          }
803        else ecode += GET(ecode, 1);        else
804            {
805            condition = FALSE;
806            ecode += GET(ecode, 1);
807            }
808          }
809    
810        /* We are now at the branch that is to be obeyed. As there is only one,      /* We are now at the branch that is to be obeyed. As there is only one,
811        we can use tail recursion to avoid using another stack frame. */      we can use tail recursion to avoid using another stack frame, except when
812        match_cbegroup is required for an unlimited repeat of a possibly empty
813        group. If the second alternative doesn't exist, we can just plough on. */
814    
815        if (condition || *ecode == OP_ALT)
816          {
817          ecode += 1 + LINK_SIZE;
818          if (op == OP_SCOND)        /* Possibly empty group */
819            {
820            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, match_cbegroup, RM49);
821            RRETURN(rrc);
822            }
823          else                       /* Group must match something */
824            {
825            flags = 0;
826            goto TAIL_RECURSE;
827            }
828          }
829        else                         /* Condition false & no 2nd alternative */
830          {
831        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
       flags = match_isgroup;  
       goto TAIL_RECURSE;  
832        }        }
     /* Control never reaches here */  
   
     /* Skip over conditional reference or large extraction number data if  
     encountered. */  
   
     case OP_CREF:  
     case OP_BRANUMBER:  
     ecode += 3;  
833      break;      break;
834    
     /* End of the pattern. If we are in a recursion, we should restore the  
     offsets appropriately and continue from after the call. */  
835    
836        /* End of the pattern, either real or forced. If we are in a top-level
837        recursion, we should restore the offsets appropriately and continue from
838        after the call. */
839    
840        case OP_ACCEPT:
841      case OP_END:      case OP_END:
842      if (md->recursive != NULL && md->recursive->group_num == 0)      if (md->recursive != NULL && md->recursive->group_num == 0)
843        {        {
# Line 745  for (;;) Line 846  for (;;)
846        md->recursive = rec->prevrec;        md->recursive = rec->prevrec;
847        memmove(md->offset_vector, rec->offset_save,        memmove(md->offset_vector, rec->offset_save,
848          rec->saved_max * sizeof(int));          rec->saved_max * sizeof(int));
849        md->start_match = rec->save_start;        mstart = rec->save_start;
850        ims = original_ims;        ims = original_ims;
851        ecode = rec->after_call;        ecode = rec->after_call;
852        break;        break;
# Line 754  for (;;) Line 855  for (;;)
855      /* 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
856      string - backtracking will then try other alternatives, if any. */      string - backtracking will then try other alternatives, if any. */
857    
858      if (md->notempty && eptr == md->start_match) RRETURN(MATCH_NOMATCH);      if (md->notempty && eptr == mstart) RRETURN(MATCH_NOMATCH);
859      md->end_match_ptr = eptr;          /* Record where we ended */      md->end_match_ptr = eptr;           /* Record where we ended */
860      md->end_offset_top = offset_top;   /* and how many extracts were taken */      md->end_offset_top = offset_top;    /* and how many extracts were taken */
861        md->start_match_ptr = mstart;       /* and the start (\K can modify) */
862      RRETURN(MATCH_MATCH);      RRETURN(MATCH_MATCH);
863    
864      /* Change option settings */      /* Change option settings */
# Line 777  for (;;) Line 879  for (;;)
879      case OP_ASSERTBACK:      case OP_ASSERTBACK:
880      do      do
881        {        {
882        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
883          match_isgroup);          RM4);
884        if (rrc == MATCH_MATCH) break;        if (rrc == MATCH_MATCH) break;
885        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
886        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
887        }        }
888      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 804  for (;;) Line 906  for (;;)
906      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
907      do      do
908        {        {
909        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
910          match_isgroup);          RM5);
911        if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH);
912        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
913        ecode += GET(ecode,1);        ecode += GET(ecode,1);
914        }        }
915      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 826  for (;;) Line 928  for (;;)
928  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
929      if (utf8)      if (utf8)
930        {        {
931        c = GET(ecode,1);        i = GET(ecode, 1);
932        for (i = 0; i < c; i++)        while (i-- > 0)
933          {          {
934          eptr--;          eptr--;
935          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
936          BACKCHAR(eptr)          BACKCHAR(eptr);
937          }          }
938        }        }
939      else      else
# Line 840  for (;;) Line 942  for (;;)
942      /* 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 */
943    
944        {        {
945        eptr -= GET(ecode,1);        eptr -= GET(ecode, 1);
946        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
947        }        }
948    
# Line 862  for (;;) Line 964  for (;;)
964        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
965        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
966        cb.subject_length   = md->end_subject - md->start_subject;        cb.subject_length   = md->end_subject - md->start_subject;
967        cb.start_match      = md->start_match - md->start_subject;        cb.start_match      = mstart - md->start_subject;
968        cb.current_position = eptr - md->start_subject;        cb.current_position = eptr - md->start_subject;
969        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
970        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
# Line 897  for (;;) Line 999  for (;;)
999      case OP_RECURSE:      case OP_RECURSE:
1000        {        {
1001        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1002        new_recursive.group_num = *callpat - OP_BRA;        new_recursive.group_num = (callpat == md->start_code)? 0 :
1003            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);  
1004    
1005        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1006    
# Line 929  for (;;) Line 1026  for (;;)
1026    
1027        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1028              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1029        new_recursive.save_start = md->start_match;        new_recursive.save_start = mstart;
1030        md->start_match = eptr;        mstart = eptr;
1031    
1032        /* 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
1033        restore the offset and recursion data. */        restore the offset and recursion data. */
1034    
1035        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1036          flags = (*callpat >= OP_SBRA)? match_cbegroup : 0;
1037        do        do
1038          {          {
1039          RMATCH(rrc, eptr, callpat + 1 + LINK_SIZE, offset_top, md, ims,          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,
1040              eptrb, match_isgroup);            md, ims, eptrb, flags, RM6);
1041          if (rrc == MATCH_MATCH)          if (rrc == MATCH_MATCH)
1042            {            {
1043            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
# Line 948  for (;;) Line 1046  for (;;)
1046              (pcre_free)(new_recursive.offset_save);              (pcre_free)(new_recursive.offset_save);
1047            RRETURN(MATCH_MATCH);            RRETURN(MATCH_MATCH);
1048            }            }
1049          else if (rrc != MATCH_NOMATCH)          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1050            {            {
1051            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1052            RRETURN(rrc);            RRETURN(rrc);
# Line 982  for (;;) Line 1080  for (;;)
1080    
1081      do      do
1082        {        {
1083        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM7);
         eptrb, match_isgroup);  
1084        if (rrc == MATCH_MATCH) break;        if (rrc == MATCH_MATCH) break;
1085        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1086        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1087        }        }
1088      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 997  for (;;) Line 1094  for (;;)
1094      /* Continue as from after the assertion, updating the offsets high water      /* Continue as from after the assertion, updating the offsets high water
1095      mark, since extracts may have been taken. */      mark, since extracts may have been taken. */
1096    
1097      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode, 1); while (*ecode == OP_ALT);
1098    
1099      offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1100      eptr = md->end_match_ptr;      eptr = md->end_match_ptr;
# Line 1028  for (;;) Line 1125  for (;;)
1125    
1126      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1127        {        {
1128        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);
1129        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1130        ecode = prev;        ecode = prev;
1131        flags = match_isgroup;        flags = 0;
1132        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1133        }        }
1134      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
1135        {        {
1136        RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);        RMATCH(eptr, prev, offset_top, md, ims, eptrb, match_cbegroup, RM9);
1137        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1138        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1139        flags = 0;        flags = 0;
# Line 1060  for (;;) Line 1157  for (;;)
1157      case OP_BRAZERO:      case OP_BRAZERO:
1158        {        {
1159        next = ecode+1;        next = ecode+1;
1160        RMATCH(rrc, eptr, next, offset_top, md, ims, eptrb, match_isgroup);        RMATCH(eptr, next, offset_top, md, ims, eptrb, 0, RM10);
1161        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1162        do next += GET(next,1); while (*next == OP_ALT);        do next += GET(next,1); while (*next == OP_ALT);
1163        ecode = next + 1+LINK_SIZE;        ecode = next + 1 + LINK_SIZE;
1164        }        }
1165      break;      break;
1166    
1167      case OP_BRAMINZERO:      case OP_BRAMINZERO:
1168        {        {
1169        next = ecode+1;        next = ecode+1;
1170        do next += GET(next,1); while (*next == OP_ALT);        do next += GET(next, 1); while (*next == OP_ALT);
1171        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);  
1172        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1173        ecode++;        ecode++;
1174        }        }
1175      break;      break;
1176    
1177      /* End of a group, repeated or non-repeating. If we are at the end of      /* End of a group, repeated or non-repeating. */
     an assertion "group", stop matching and return MATCH_MATCH, but record the  
     current high water mark for use by positive assertions. Do this also  
     for the "once" (not-backup up) groups. */  
1178    
1179      case OP_KET:      case OP_KET:
1180      case OP_KETRMIN:      case OP_KETRMIN:
1181      case OP_KETRMAX:      case OP_KETRMAX:
1182      prev = ecode - GET(ecode, 1);      prev = ecode - GET(ecode, 1);
     saved_eptr = eptrb->epb_saved_eptr;  
1183    
1184      /* Back up the stack of bracket start pointers. */      /* If this was a group that remembered the subject start, in order to break
1185        infinite repeats of empty string matches, retrieve the subject start from
1186        the chain. Otherwise, set it NULL. */
1187    
1188        if (*prev >= OP_SBRA)
1189          {
1190          saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */
1191          eptrb = eptrb->epb_prev;              /* Backup to previous group */
1192          }
1193        else saved_eptr = NULL;
1194    
1195      eptrb = eptrb->epb_prev;      /* If we are at the end of an assertion group, stop matching and return
1196        MATCH_MATCH, but record the current high water mark for use by positive
1197        assertions. Do this also for the "once" (atomic) groups. */
1198    
1199      if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||      if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||
1200          *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||          *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||
# Line 1102  for (;;) Line 1205  for (;;)
1205        RRETURN(MATCH_MATCH);        RRETURN(MATCH_MATCH);
1206        }        }
1207    
1208      /* In all other cases except a conditional group we have to check the      /* For capturing groups we have to check the group number back at the start
1209      group number back at the start and if necessary complete handling an      and if necessary complete handling an extraction by setting the offsets and
1210      extraction by setting the offsets and bumping the high water mark. */      bumping the high water mark. Note that whole-pattern recursion is coded as
1211        a recurse into group 0, so it won't be picked up here. Instead, we catch it
1212        when the OP_END is reached. Other recursion is handled here. */
1213    
1214      if (*prev != OP_COND)      if (*prev == OP_CBRA || *prev == OP_SCBRA)
1215        {        {
1216        number = *prev - OP_BRA;        number = GET2(prev, 1+LINK_SIZE);
   
       /* For extended extraction brackets (large number), we have to fish out  
       the number from a dummy opcode at the start. */  
   
       if (number > EXTRACT_BASIC_MAX) number = GET2(prev, 2+LINK_SIZE);  
1217        offset = number << 1;        offset = number << 1;
1218    
1219  #ifdef DEBUG  #ifdef DEBUG
# Line 1121  for (;;) Line 1221  for (;;)
1221        printf("\n");        printf("\n");
1222  #endif  #endif
1223    
1224        /* Test for a numbered group. This includes groups called as a result        md->capture_last = number;
1225        of recursion. Note that whole-pattern recursion is coded as a recurse        if (offset >= md->offset_max) md->offset_overflow = TRUE; else
       into group 0, so it won't be picked up here. Instead, we catch it when  
       the OP_END is reached. */  
   
       if (number > 0)  
1226          {          {
1227          md->capture_last = number;          md->offset_vector[offset] =
1228          if (offset >= md->offset_max) md->offset_overflow = TRUE; else            md->offset_vector[md->offset_end - number];
1229            {          md->offset_vector[offset+1] = eptr - md->start_subject;
1230            md->offset_vector[offset] =          if (offset_top <= offset) offset_top = offset + 2;
1231              md->offset_vector[md->offset_end - number];          }
1232            md->offset_vector[offset+1] = eptr - md->start_subject;  
1233            if (offset_top <= offset) offset_top = offset + 2;        /* Handle a recursively called group. Restore the offsets
1234            }        appropriately and continue from after the call. */
1235    
1236          /* Handle a recursively called group. Restore the offsets        if (md->recursive != NULL && md->recursive->group_num == number)
1237          appropriately and continue from after the call. */          {
1238            recursion_info *rec = md->recursive;
1239          if (md->recursive != NULL && md->recursive->group_num == number)          DPRINTF(("Recursion (%d) succeeded - continuing\n", number));
1240            {          md->recursive = rec->prevrec;
1241            recursion_info *rec = md->recursive;          mstart = rec->save_start;
1242            DPRINTF(("Recursion (%d) succeeded - continuing\n", number));          memcpy(md->offset_vector, rec->offset_save,
1243            md->recursive = rec->prevrec;            rec->saved_max * sizeof(int));
1244            md->start_match = rec->save_start;          ecode = rec->after_call;
1245            memcpy(md->offset_vector, rec->offset_save,          ims = original_ims;
1246              rec->saved_max * sizeof(int));          break;
           ecode = rec->after_call;  
           ims = original_ims;  
           break;  
           }  
1247          }          }
1248        }        }
1249    
1250      /* 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
1251      the group. */      flags, in case they got changed during the group. */
1252    
1253      ims = original_ims;      ims = original_ims;
1254      DPRINTF(("ims reset to %02lx\n", ims));      DPRINTF(("ims reset to %02lx\n", ims));
# Line 1175  for (;;) Line 1267  for (;;)
1267    
1268      /* 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
1269      preceding bracket, in the appropriate order. In the second case, we can use      preceding bracket, in the appropriate order. In the second case, we can use
1270      tail recursion to avoid using another stack frame. */      tail recursion to avoid using another stack frame, unless we have an
1271        unlimited repeat of a group that can match an empty string. */
1272    
1273        flags = (*prev >= OP_SBRA)? match_cbegroup : 0;
1274    
1275      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1276        {        {
1277        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, RM12);
1278        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1279          if (flags != 0)    /* Could match an empty string */
1280            {
1281            RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM50);
1282            RRETURN(rrc);
1283            }
1284        ecode = prev;        ecode = prev;
       flags = match_isgroup;  
1285        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1286        }        }
1287      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
1288        {        {
1289        RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);        RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13);
1290        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1291        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1292        flags = 0;        flags = 0;
# Line 1202  for (;;) Line 1301  for (;;)
1301      if ((ims & PCRE_MULTILINE) != 0)      if ((ims & PCRE_MULTILINE) != 0)
1302        {        {
1303        if (eptr != md->start_subject &&        if (eptr != md->start_subject &&
1304            (eptr == md->end_subject ||            (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
            eptr < md->start_subject + md->nllen ||  
            !IS_NEWLINE(eptr - md->nllen)))  
1305          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
1306        ecode++;        ecode++;
1307        break;        break;
# Line 1225  for (;;) Line 1322  for (;;)
1322      ecode++;      ecode++;
1323      break;      break;
1324    
1325        /* Reset the start of match point */
1326    
1327        case OP_SET_SOM:
1328        mstart = eptr;
1329        ecode++;
1330        break;
1331    
1332      /* Assert before internal newline if multiline, or before a terminating      /* Assert before internal newline if multiline, or before a terminating
1333      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. */
1334    
# Line 1244  for (;;) Line 1348  for (;;)
1348        if (!md->endonly)        if (!md->endonly)
1349          {          {
1350          if (eptr != md->end_subject &&          if (eptr != md->end_subject &&
1351              (eptr != md->end_subject - md->nllen || !IS_NEWLINE(eptr)))              (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
1352            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1353          ecode++;          ecode++;
1354          break;          break;
# Line 1263  for (;;) Line 1367  for (;;)
1367    
1368      case OP_EODN:      case OP_EODN:
1369      if (eptr != md->end_subject &&      if (eptr != md->end_subject &&
1370          (eptr != md->end_subject - md->nllen || !IS_NEWLINE(eptr)))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
1371        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1372      ecode++;      ecode++;
1373      break;      break;
# Line 1319  for (;;) Line 1423  for (;;)
1423      case OP_ANY:      case OP_ANY:
1424      if ((ims & PCRE_DOTALL) == 0)      if ((ims & PCRE_DOTALL) == 0)
1425        {        {
1426        if (eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr))        if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
         RRETURN(MATCH_NOMATCH);  
1427        }        }
1428      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);
1429      if (utf8)      if (utf8)
# Line 1414  for (;;) Line 1517  for (;;)
1517      ecode++;      ecode++;
1518      break;      break;
1519    
1520        case OP_ANYNL:
1521        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1522        GETCHARINCTEST(c, eptr);
1523        switch(c)
1524          {
1525          default: RRETURN(MATCH_NOMATCH);
1526          case 0x000d:
1527          if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
1528          break;
1529    
1530          case 0x000a:
1531          break;
1532    
1533          case 0x000b:
1534          case 0x000c:
1535          case 0x0085:
1536          case 0x2028:
1537          case 0x2029:
1538          if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
1539          break;
1540          }
1541        ecode++;
1542        break;
1543    
1544        case OP_NOT_HSPACE:
1545        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1546        GETCHARINCTEST(c, eptr);
1547        switch(c)
1548          {
1549          default: break;
1550          case 0x09:      /* HT */
1551          case 0x20:      /* SPACE */
1552          case 0xa0:      /* NBSP */
1553          case 0x1680:    /* OGHAM SPACE MARK */
1554          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1555          case 0x2000:    /* EN QUAD */
1556          case 0x2001:    /* EM QUAD */
1557          case 0x2002:    /* EN SPACE */
1558          case 0x2003:    /* EM SPACE */
1559          case 0x2004:    /* THREE-PER-EM SPACE */
1560          case 0x2005:    /* FOUR-PER-EM SPACE */
1561          case 0x2006:    /* SIX-PER-EM SPACE */
1562          case 0x2007:    /* FIGURE SPACE */
1563          case 0x2008:    /* PUNCTUATION SPACE */
1564          case 0x2009:    /* THIN SPACE */
1565          case 0x200A:    /* HAIR SPACE */
1566          case 0x202f:    /* NARROW NO-BREAK SPACE */
1567          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1568          case 0x3000:    /* IDEOGRAPHIC SPACE */
1569          RRETURN(MATCH_NOMATCH);
1570          }
1571        ecode++;
1572        break;
1573    
1574        case OP_HSPACE:
1575        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1576        GETCHARINCTEST(c, eptr);
1577        switch(c)
1578          {
1579          default: RRETURN(MATCH_NOMATCH);
1580          case 0x09:      /* HT */
1581          case 0x20:      /* SPACE */
1582          case 0xa0:      /* NBSP */
1583          case 0x1680:    /* OGHAM SPACE MARK */
1584          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1585          case 0x2000:    /* EN QUAD */
1586          case 0x2001:    /* EM QUAD */
1587          case 0x2002:    /* EN SPACE */
1588          case 0x2003:    /* EM SPACE */
1589          case 0x2004:    /* THREE-PER-EM SPACE */
1590          case 0x2005:    /* FOUR-PER-EM SPACE */
1591          case 0x2006:    /* SIX-PER-EM SPACE */
1592          case 0x2007:    /* FIGURE SPACE */
1593          case 0x2008:    /* PUNCTUATION SPACE */
1594          case 0x2009:    /* THIN SPACE */
1595          case 0x200A:    /* HAIR SPACE */
1596          case 0x202f:    /* NARROW NO-BREAK SPACE */
1597          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1598          case 0x3000:    /* IDEOGRAPHIC SPACE */
1599          break;
1600          }
1601        ecode++;
1602        break;
1603    
1604        case OP_NOT_VSPACE:
1605        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1606        GETCHARINCTEST(c, eptr);
1607        switch(c)
1608          {
1609          default: break;
1610          case 0x0a:      /* LF */
1611          case 0x0b:      /* VT */
1612          case 0x0c:      /* FF */
1613          case 0x0d:      /* CR */
1614          case 0x85:      /* NEL */
1615          case 0x2028:    /* LINE SEPARATOR */
1616          case 0x2029:    /* PARAGRAPH SEPARATOR */
1617          RRETURN(MATCH_NOMATCH);
1618          }
1619        ecode++;
1620        break;
1621    
1622        case OP_VSPACE:
1623        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1624        GETCHARINCTEST(c, eptr);
1625        switch(c)
1626          {
1627          default: RRETURN(MATCH_NOMATCH);
1628          case 0x0a:      /* LF */
1629          case 0x0b:      /* VT */
1630          case 0x0c:      /* FF */
1631          case 0x0d:      /* CR */
1632          case 0x85:      /* NEL */
1633          case 0x2028:    /* LINE SEPARATOR */
1634          case 0x2029:    /* PARAGRAPH SEPARATOR */
1635          break;
1636          }
1637        ecode++;
1638        break;
1639    
1640  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
1641      /* Check the next character by Unicode property. We will get here only      /* Check the next character by Unicode property. We will get here only
1642      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 1456  for (;;) Line 1679  for (;;)
1679    
1680          default:          default:
1681          RRETURN(PCRE_ERROR_INTERNAL);          RRETURN(PCRE_ERROR_INTERNAL);
         break;  
1682          }          }
1683    
1684        ecode += 3;        ecode += 3;
# Line 1570  for (;;) Line 1792  for (;;)
1792          {          {
1793          for (fi = min;; fi++)          for (fi = min;; fi++)
1794            {            {
1795            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM14);
1796            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1797            if (fi >= max || !match_ref(offset, eptr, length, md, ims))            if (fi >= max || !match_ref(offset, eptr, length, md, ims))
1798              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
# Line 1591  for (;;) Line 1813  for (;;)
1813            }            }
1814          while (eptr >= pp)          while (eptr >= pp)
1815            {            {
1816            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM15);
1817            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1818            eptr -= length;            eptr -= length;
1819            }            }
# Line 1696  for (;;) Line 1918  for (;;)
1918            {            {
1919            for (fi = min;; fi++)            for (fi = min;; fi++)
1920              {              {
1921              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM16);
1922              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1923              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1924              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 1716  for (;;) Line 1938  for (;;)
1938            {            {
1939            for (fi = min;; fi++)            for (fi = min;; fi++)
1940              {              {
1941              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM17);
1942              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1943              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1944              c = *eptr++;              c = *eptr++;
# Line 1753  for (;;) Line 1975  for (;;)
1975              }              }
1976            for (;;)            for (;;)
1977              {              {
1978              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM18);
1979              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1980              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
1981              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 1772  for (;;) Line 1994  for (;;)
1994              }              }
1995            while (eptr >= pp)            while (eptr >= pp)
1996              {              {
1997              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM19);
1998              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1999              eptr--;              eptr--;
2000              }              }
# Line 1843  for (;;) Line 2065  for (;;)
2065          {          {
2066          for (fi = min;; fi++)          for (fi = min;; fi++)
2067            {            {
2068            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20);
2069            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2070            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2071            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
# Line 1867  for (;;) Line 2089  for (;;)
2089            }            }
2090          for(;;)          for(;;)
2091            {            {
2092            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);
2093            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2094            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
2095            BACKCHAR(eptr)            if (utf8) BACKCHAR(eptr);
2096            }            }
2097          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2098          }          }
# Line 1926  for (;;) Line 2148  for (;;)
2148    
2149        else        else
2150          {          {
2151          int dc;          unsigned int dc;
2152          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
2153          ecode += length;          ecode += length;
2154    
# Line 1953  for (;;) Line 2175  for (;;)
2175        }        }
2176      break;      break;
2177    
2178      /* Match a single character repeatedly; different opcodes share code. */      /* Match a single character repeatedly. */
2179    
2180      case OP_EXACT:      case OP_EXACT:
2181      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
2182      ecode += 3;      ecode += 3;
2183      goto REPEATCHAR;      goto REPEATCHAR;
2184    
2185        case OP_POSUPTO:
2186        possessive = TRUE;
2187        /* Fall through */
2188    
2189      case OP_UPTO:      case OP_UPTO:
2190      case OP_MINUPTO:      case OP_MINUPTO:
2191      min = 0;      min = 0;
# Line 1968  for (;;) Line 2194  for (;;)
2194      ecode += 3;      ecode += 3;
2195      goto REPEATCHAR;      goto REPEATCHAR;
2196    
2197        case OP_POSSTAR:
2198        possessive = TRUE;
2199        min = 0;
2200        max = INT_MAX;
2201        ecode++;
2202        goto REPEATCHAR;
2203    
2204        case OP_POSPLUS:
2205        possessive = TRUE;
2206        min = 1;
2207        max = INT_MAX;
2208        ecode++;
2209        goto REPEATCHAR;
2210    
2211        case OP_POSQUERY:
2212        possessive = TRUE;
2213        min = 0;
2214        max = 1;
2215        ecode++;
2216        goto REPEATCHAR;
2217    
2218      case OP_STAR:      case OP_STAR:
2219      case OP_MINSTAR:      case OP_MINSTAR:
2220      case OP_PLUS:      case OP_PLUS:
# Line 1999  for (;;) Line 2246  for (;;)
2246    
2247        if (length > 1)        if (length > 1)
2248          {          {
         int oclength = 0;  
         uschar occhars[8];  
   
2249  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2250          int othercase;          unsigned int othercase;
2251          if ((ims & PCRE_CASELESS) != 0 &&          if ((ims & PCRE_CASELESS) != 0 &&
2252              (othercase = _pcre_ucp_othercase(fc)) >= 0 &&              (othercase = _pcre_ucp_othercase(fc)) != NOTACHAR)
              othercase >= 0)  
2253            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = _pcre_ord2utf8(othercase, occhars);
2254            else oclength = 0;
2255  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2256    
2257          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2258            {            {
2259            if (memcmp(eptr, charptr, length) == 0) eptr += length;            if (memcmp(eptr, charptr, length) == 0) eptr += length;
2260    #ifdef SUPPORT_UCP
2261            /* Need braces because of following else */            /* Need braces because of following else */
2262            else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }            else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }
2263            else            else
# Line 2020  for (;;) Line 2265  for (;;)
2265              if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);              if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);
2266              eptr += oclength;              eptr += oclength;
2267              }              }
2268    #else   /* without SUPPORT_UCP */
2269              else { RRETURN(MATCH_NOMATCH); }
2270    #endif  /* SUPPORT_UCP */
2271            }            }
2272    
2273          if (min == max) continue;          if (min == max) continue;
# Line 2028  for (;;) Line 2276  for (;;)
2276            {            {
2277            for (fi = min;; fi++)            for (fi = min;; fi++)
2278              {              {
2279              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM22);
2280              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2281              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2282              if (memcmp(eptr, charptr, length) == 0) eptr += length;              if (memcmp(eptr, charptr, length) == 0) eptr += length;
2283    #ifdef SUPPORT_UCP
2284              /* Need braces because of following else */              /* Need braces because of following else */
2285              else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }              else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }
2286              else              else
# Line 2039  for (;;) Line 2288  for (;;)
2288                if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);                if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);
2289                eptr += oclength;                eptr += oclength;
2290                }                }
2291    #else   /* without SUPPORT_UCP */
2292                else { RRETURN (MATCH_NOMATCH); }
2293    #endif  /* SUPPORT_UCP */
2294              }              }
2295            /* Control never gets here */            /* Control never gets here */
2296            }            }
2297          else  
2298            else  /* Maximize */
2299            {            {
2300            pp = eptr;            pp = eptr;
2301            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2302              {              {
2303              if (eptr > md->end_subject - length) break;              if (eptr > md->end_subject - length) break;
2304              if (memcmp(eptr, charptr, length) == 0) eptr += length;              if (memcmp(eptr, charptr, length) == 0) eptr += length;
2305    #ifdef SUPPORT_UCP
2306              else if (oclength == 0) break;              else if (oclength == 0) break;
2307              else              else
2308                {                {
2309                if (memcmp(eptr, occhars, oclength) != 0) break;                if (memcmp(eptr, occhars, oclength) != 0) break;
2310                eptr += oclength;                eptr += oclength;
2311                }                }
2312    #else   /* without SUPPORT_UCP */
2313                else break;
2314    #endif  /* SUPPORT_UCP */
2315              }              }
2316            while (eptr >= pp)  
2317              if (possessive) continue;
2318              for(;;)
2319             {             {
2320             RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);             RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM23);
2321             if (rrc != MATCH_NOMATCH) RRETURN(rrc);             if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2322               if (eptr == pp) RRETURN(MATCH_NOMATCH);
2323    #ifdef SUPPORT_UCP
2324               eptr--;
2325               BACKCHAR(eptr);
2326    #else   /* without SUPPORT_UCP */
2327             eptr -= length;             eptr -= length;
2328    #endif  /* SUPPORT_UCP */
2329             }             }
           RRETURN(MATCH_NOMATCH);  
2330            }            }
2331          /* Control never gets here */          /* Control never gets here */
2332          }          }
# Line 2102  for (;;) Line 2366  for (;;)
2366          {          {
2367          for (fi = min;; fi++)          for (fi = min;; fi++)
2368            {            {
2369            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM24);
2370            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2371            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
2372                fc != md->lcc[*eptr++])                fc != md->lcc[*eptr++])
# Line 2110  for (;;) Line 2374  for (;;)
2374            }            }
2375          /* Control never gets here */          /* Control never gets here */
2376          }          }
2377        else        else  /* Maximize */
2378          {          {
2379          pp = eptr;          pp = eptr;
2380          for (i = min; i < max; i++)          for (i = min; i < max; i++)
# Line 2118  for (;;) Line 2382  for (;;)
2382            if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break;            if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break;
2383            eptr++;            eptr++;
2384            }            }
2385            if (possessive) continue;
2386          while (eptr >= pp)          while (eptr >= pp)
2387            {            {
2388            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM25);
2389            eptr--;            eptr--;
2390            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2391            }            }
# Line 2139  for (;;) Line 2404  for (;;)
2404          {          {
2405          for (fi = min;; fi++)          for (fi = min;; fi++)
2406            {            {
2407            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM26);
2408            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2409            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)
2410              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2411            }            }
2412          /* Control never gets here */          /* Control never gets here */
2413          }          }
2414        else        else  /* Maximize */
2415          {          {
2416          pp = eptr;          pp = eptr;
2417          for (i = min; i < max; i++)          for (i = min; i < max; i++)
# Line 2154  for (;;) Line 2419  for (;;)
2419            if (eptr >= md->end_subject || fc != *eptr) break;            if (eptr >= md->end_subject || fc != *eptr) break;
2420            eptr++;            eptr++;
2421            }            }
2422            if (possessive) continue;
2423          while (eptr >= pp)          while (eptr >= pp)
2424            {            {
2425            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27);
2426            eptr--;            eptr--;
2427            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2428            }            }
# Line 2206  for (;;) Line 2472  for (;;)
2472      ecode += 3;      ecode += 3;
2473      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
2474    
2475        case OP_NOTPOSSTAR:
2476        possessive = TRUE;
2477        min = 0;
2478        max = INT_MAX;
2479        ecode++;
2480        goto REPEATNOTCHAR;
2481    
2482        case OP_NOTPOSPLUS:
2483        possessive = TRUE;
2484        min = 1;
2485        max = INT_MAX;
2486        ecode++;
2487        goto REPEATNOTCHAR;
2488    
2489        case OP_NOTPOSQUERY:
2490        possessive = TRUE;
2491        min = 0;
2492        max = 1;
2493        ecode++;
2494        goto REPEATNOTCHAR;
2495    
2496        case OP_NOTPOSUPTO:
2497        possessive = TRUE;
2498        min = 0;
2499        max = GET2(ecode, 1);
2500        ecode += 3;
2501        goto REPEATNOTCHAR;
2502    
2503      case OP_NOTSTAR:      case OP_NOTSTAR:
2504      case OP_NOTMINSTAR:      case OP_NOTMINSTAR:
2505      case OP_NOTPLUS:      case OP_NOTPLUS:
# Line 2245  for (;;) Line 2539  for (;;)
2539        /* UTF-8 mode */        /* UTF-8 mode */
2540        if (utf8)        if (utf8)
2541          {          {
2542          register int d;          register unsigned int d;
2543          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2544            {            {
2545            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
# Line 2270  for (;;) Line 2564  for (;;)
2564          /* UTF-8 mode */          /* UTF-8 mode */
2565          if (utf8)          if (utf8)
2566            {            {
2567            register int d;            register unsigned int d;
2568            for (fi = min;; fi++)            for (fi = min;; fi++)
2569              {              {
2570              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28);
2571              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2572              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2573              if (d < 256) d = md->lcc[d];              if (d < 256) d = md->lcc[d];
# Line 2287  for (;;) Line 2581  for (;;)
2581            {            {
2582            for (fi = min;; fi++)            for (fi = min;; fi++)
2583              {              {
2584              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM29);
2585              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2586              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])
2587                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2306  for (;;) Line 2600  for (;;)
2600          /* UTF-8 mode */          /* UTF-8 mode */
2601          if (utf8)          if (utf8)
2602            {            {
2603            register int d;            register unsigned int d;
2604            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2605              {              {
2606              int len = 1;              int len = 1;
# Line 2316  for (;;) Line 2610  for (;;)
2610              if (fc == d) break;              if (fc == d) break;
2611              eptr += len;              eptr += len;
2612              }              }
2613            for(;;)          if (possessive) continue;
2614            for(;;)
2615              {              {
2616              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM30);
2617              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2618              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2619              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2333  for (;;) Line 2628  for (;;)
2628              if (eptr >= md->end_subject || fc == md->lcc[*eptr]) break;              if (eptr >= md->end_subject || fc == md->lcc[*eptr]) break;
2629              eptr++;              eptr++;
2630              }              }
2631              if (possessive) continue;
2632            while (eptr >= pp)            while (eptr >= pp)
2633              {              {
2634              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM31);
2635              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2636              eptr--;              eptr--;
2637              }              }
# Line 2354  for (;;) Line 2650  for (;;)
2650        /* UTF-8 mode */        /* UTF-8 mode */
2651        if (utf8)        if (utf8)
2652          {          {
2653          register int d;          register unsigned int d;
2654          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2655            {            {
2656            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
# Line 2377  for (;;) Line 2673  for (;;)
2673          /* UTF-8 mode */          /* UTF-8 mode */
2674          if (utf8)          if (utf8)
2675            {            {
2676            register int d;            register unsigned int d;
2677            for (fi = min;; fi++)            for (fi = min;; fi++)
2678              {              {
2679              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32);
2680              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2681              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2682              if (fi >= max || eptr >= md->end_subject || fc == d)              if (fi >= max || eptr >= md->end_subject || fc == d)
# Line 2393  for (;;) Line 2689  for (;;)
2689            {            {
2690            for (fi = min;; fi++)            for (fi = min;; fi++)
2691              {              {
2692              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM33);
2693              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2694              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)
2695                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2412  for (;;) Line 2708  for (;;)
2708          /* UTF-8 mode */          /* UTF-8 mode */
2709          if (utf8)          if (utf8)
2710            {            {
2711            register int d;            register unsigned int d;
2712            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2713              {              {
2714              int len = 1;              int len = 1;
# Line 2421  for (;;) Line 2717  for (;;)
2717              if (fc == d) break;              if (fc == d) break;
2718              eptr += len;              eptr += len;
2719              }              }
2720              if (possessive) continue;
2721            for(;;)            for(;;)
2722              {              {
2723              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM34);
2724              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2725              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2726              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2438  for (;;) Line 2735  for (;;)
2735              if (eptr >= md->end_subject || fc == *eptr) break;              if (eptr >= md->end_subject || fc == *eptr) break;
2736              eptr++;              eptr++;
2737              }              }
2738              if (possessive) continue;
2739            while (eptr >= pp)            while (eptr >= pp)
2740              {              {
2741              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM35);
2742              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2743              eptr--;              eptr--;
2744              }              }
# Line 2469  for (;;) Line 2767  for (;;)
2767      ecode += 3;      ecode += 3;
2768      goto REPEATTYPE;      goto REPEATTYPE;
2769    
2770        case OP_TYPEPOSSTAR:
2771        possessive = TRUE;
2772        min = 0;
2773        max = INT_MAX;
2774        ecode++;
2775        goto REPEATTYPE;
2776    
2777        case OP_TYPEPOSPLUS:
2778        possessive = TRUE;
2779        min = 1;
2780        max = INT_MAX;
2781        ecode++;
2782        goto REPEATTYPE;
2783    
2784        case OP_TYPEPOSQUERY:
2785        possessive = TRUE;
2786        min = 0;
2787        max = 1;
2788        ecode++;
2789        goto REPEATTYPE;
2790    
2791        case OP_TYPEPOSUPTO:
2792        possessive = TRUE;
2793        min = 0;
2794        max = GET2(ecode, 1);
2795        ecode += 3;
2796        goto REPEATTYPE;
2797    
2798      case OP_TYPESTAR:      case OP_TYPESTAR:
2799      case OP_TYPEMINSTAR:      case OP_TYPEMINSTAR:
2800      case OP_TYPEPLUS:      case OP_TYPEPLUS:
# Line 2519  for (;;) Line 2845  for (;;)
2845            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2846              {              {
2847              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2848              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2849              }              }
2850            break;            break;
2851    
# Line 2527  for (;;) Line 2853  for (;;)
2853            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2854              {              {
2855              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2856              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2857              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2858              if ((prop_chartype == ucp_Lu ||              if ((prop_chartype == ucp_Lu ||
2859                   prop_chartype == ucp_Ll ||                   prop_chartype == ucp_Ll ||
# Line 2540  for (;;) Line 2866  for (;;)
2866            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2867              {              {
2868              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2869              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2870              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2871              if ((prop_category == prop_value) == prop_fail_result)              if ((prop_category == prop_value) == prop_fail_result)
2872                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2551  for (;;) Line 2877  for (;;)
2877            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2878              {              {
2879              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2880              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2881              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2882              if ((prop_chartype == prop_value) == prop_fail_result)              if ((prop_chartype == prop_value) == prop_fail_result)
2883                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2562  for (;;) Line 2888  for (;;)
2888            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2889              {              {
2890              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2891              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2892              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2893              if ((prop_script == prop_value) == prop_fail_result)              if ((prop_script == prop_value) == prop_fail_result)
2894                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2571  for (;;) Line 2897  for (;;)
2897    
2898            default:            default:
2899            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
           break;  
2900            }            }
2901          }          }
2902    
# Line 2611  for (;;) Line 2936  for (;;)
2936          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2937            {            {
2938            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
2939                 ((ims & PCRE_DOTALL) == 0 &&                 ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))
                  eptr <= md->end_subject - md->nllen &&  
                  IS_NEWLINE(eptr)))  
2940              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2941            eptr++;            eptr++;
2942            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
# Line 2624  for (;;) Line 2947  for (;;)
2947          eptr += min;          eptr += min;
2948          break;          break;
2949    
2950          case OP_NOT_DIGIT:          case OP_ANYNL:
2951          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2952            {            {
2953            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2954            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
2955            if (c < 128 && (md->ctypes[c] & ctype_digit) != 0)            switch(c)
2956              RRETURN(MATCH_NOMATCH);              {
2957                default: RRETURN(MATCH_NOMATCH);
2958                case 0x000d:
2959                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
2960                break;
2961    
2962                case 0x000a:
2963                break;
2964    
2965                case 0x000b:
2966                case 0x000c:
2967                case 0x0085:
2968                case 0x2028:
2969                case 0x2029:
2970                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2971                break;
2972                }
2973            }            }
2974          break;          break;
2975    
2976          case OP_DIGIT:          case OP_NOT_HSPACE:
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) RRETURN(MATCH_NOMATCH);
2980               *eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            GETCHARINC(c, eptr);
2981              switch(c)
2982                {
2983                default: break;
2984                case 0x09:      /* HT */
2985                case 0x20:      /* SPACE */
2986                case 0xa0:      /* NBSP */
2987                case 0x1680:    /* OGHAM SPACE MARK */
2988                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
2989                case 0x2000:    /* EN QUAD */
2990                case 0x2001:    /* EM QUAD */
2991                case 0x2002:    /* EN SPACE */
2992                case 0x2003:    /* EM SPACE */
2993                case 0x2004:    /* THREE-PER-EM SPACE */
2994                case 0x2005:    /* FOUR-PER-EM SPACE */
2995                case 0x2006:    /* SIX-PER-EM SPACE */
2996                case 0x2007:    /* FIGURE SPACE */
2997                case 0x2008:    /* PUNCTUATION SPACE */
2998                case 0x2009:    /* THIN SPACE */
2999                case 0x200A:    /* HAIR SPACE */
3000                case 0x202f:    /* NARROW NO-BREAK SPACE */
3001                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3002                case 0x3000:    /* IDEOGRAPHIC SPACE */
3003              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3004            /* No need to skip more bytes - we know it's a 1-byte character */              }
3005            }            }
3006          break;          break;
3007    
3008          case OP_NOT_WHITESPACE:          case OP_HSPACE:
3009          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3010            {            {
3011            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3012               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_space) != 0))            GETCHARINC(c, eptr);
3013              RRETURN(MATCH_NOMATCH);            switch(c)
3014            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;              {
3015                default: RRETURN(MATCH_NOMATCH);
3016                case 0x09:      /* HT */
3017                case 0x20:      /* SPACE */
3018                case 0xa0:      /* NBSP */
3019                case 0x1680:    /* OGHAM SPACE MARK */
3020                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3021                case 0x2000:    /* EN QUAD */
3022                case 0x2001:    /* EM QUAD */
3023                case 0x2002:    /* EN SPACE */
3024                case 0x2003:    /* EM SPACE */
3025                case 0x2004:    /* THREE-PER-EM SPACE */
3026                case 0x2005:    /* FOUR-PER-EM SPACE */
3027                case 0x2006:    /* SIX-PER-EM SPACE */
3028                case 0x2007:    /* FIGURE SPACE */
3029                case 0x2008:    /* PUNCTUATION SPACE */
3030                case 0x2009:    /* THIN SPACE */
3031                case 0x200A:    /* HAIR SPACE */
3032                case 0x202f:    /* NARROW NO-BREAK SPACE */
3033                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3034                case 0x3000:    /* IDEOGRAPHIC SPACE */
3035                break;
3036                }
3037            }            }
3038          break;          break;
3039    
3040          case OP_WHITESPACE:          case OP_NOT_VSPACE:
3041          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3042            {            {
3043            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3044               *eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            GETCHARINC(c, eptr);
3045              switch(c)
3046                {
3047                default: break;
3048                case 0x0a:      /* LF */
3049                case 0x0b:      /* VT */
3050                case 0x0c:      /* FF */
3051                case 0x0d:      /* CR */
3052                case 0x85:      /* NEL */
3053                case 0x2028:    /* LINE SEPARATOR */
3054                case 0x2029:    /* PARAGRAPH SEPARATOR */
3055              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3056            /* No need to skip more bytes - we know it's a 1-byte character */              }
3057            }            }
3058          break;          break;
3059    
3060          case OP_NOT_WORDCHAR:          case OP_VSPACE:
3061          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3062            {            {
3063            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3064               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_word) != 0))            GETCHARINC(c, eptr);
3065              RRETURN(MATCH_NOMATCH);            switch(c)
3066            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;              {
3067                default: RRETURN(MATCH_NOMATCH);
3068                case 0x0a:      /* LF */
3069                case 0x0b:      /* VT */
3070                case 0x0c:      /* FF */
3071                case 0x0d:      /* CR */
3072                case 0x85:      /* NEL */
3073                case 0x2028:    /* LINE SEPARATOR */
3074                case 0x2029:    /* PARAGRAPH SEPARATOR */
3075                break;
3076                }
3077              }
3078            break;
3079    
3080            case OP_NOT_DIGIT:
3081            for (i = 1; i <= min; i++)
3082              {
3083              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3084              GETCHARINC(c, eptr);
3085              if (c < 128 && (md->ctypes[c] & ctype_digit) != 0)
3086                RRETURN(MATCH_NOMATCH);
3087              }
3088            break;
3089    
3090            case OP_DIGIT:
3091            for (i = 1; i <= min; i++)
3092              {
3093              if (eptr >= md->end_subject ||
3094                 *eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)
3095                RRETURN(MATCH_NOMATCH);
3096              /* No need to skip more bytes - we know it's a 1-byte character */
3097              }
3098            break;
3099    
3100            case OP_NOT_WHITESPACE:
3101            for (i = 1; i <= min; i++)
3102              {
3103              if (eptr >= md->end_subject ||
3104                 (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0))
3105                RRETURN(MATCH_NOMATCH);
3106              while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3107              }
3108            break;
3109    
3110            case OP_WHITESPACE:
3111            for (i = 1; i <= min; i++)
3112              {
3113              if (eptr >= md->end_subject ||
3114                 *eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)
3115                RRETURN(MATCH_NOMATCH);
3116              /* No need to skip more bytes - we know it's a 1-byte character */
3117              }
3118            break;
3119    
3120            case OP_NOT_WORDCHAR:
3121            for (i = 1; i <= min; i++)
3122              {
3123              if (eptr >= md->end_subject ||
3124                 (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0))
3125                RRETURN(MATCH_NOMATCH);
3126              while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3127            }            }
3128          break;          break;
3129    
# Line 2692  for (;;) Line 3145  for (;;)
3145  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF8 */
3146    
3147        /* 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
3148        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. We can assume that there are the minimum
3149          number of bytes present, as this was tested above. */
3150    
3151        switch(ctype)        switch(ctype)
3152          {          {
# Line 2701  for (;;) Line 3155  for (;;)
3155            {            {
3156            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3157              {              {
3158              if (eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr))              if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
               RRETURN(MATCH_NOMATCH);  
3159              eptr++;              eptr++;
3160              }              }
3161            }            }
# Line 2713  for (;;) Line 3166  for (;;)
3166          eptr += min;          eptr += min;
3167          break;          break;
3168    
3169            /* Because of the CRLF case, we can't assume the minimum number of
3170            bytes are present in this case. */
3171    
3172            case OP_ANYNL:
3173            for (i = 1; i <= min; i++)
3174              {
3175              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3176              switch(*eptr++)
3177                {
3178                default: RRETURN(MATCH_NOMATCH);
3179                case 0x000d:
3180                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3181                break;
3182                case 0x000a:
3183                break;
3184    
3185                case 0x000b:
3186                case 0x000c:
3187                case 0x0085:
3188                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3189                break;
3190                }
3191              }
3192            break;
3193    
3194            case OP_NOT_HSPACE:
3195            for (i = 1; i <= min; i++)
3196              {
3197              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3198              switch(*eptr++)
3199                {
3200                default: break;
3201                case 0x09:      /* HT */
3202                case 0x20:      /* SPACE */
3203                case 0xa0:      /* NBSP */
3204                RRETURN(MATCH_NOMATCH);
3205                }
3206              }
3207            break;
3208    
3209            case OP_HSPACE:
3210            for (i = 1; i <= min; i++)
3211              {
3212              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3213              switch(*eptr++)
3214                {
3215                default: RRETURN(MATCH_NOMATCH);
3216                case 0x09:      /* HT */
3217                case 0x20:      /* SPACE */
3218                case 0xa0:      /* NBSP */
3219                break;
3220                }
3221              }
3222            break;
3223    
3224            case OP_NOT_VSPACE:
3225            for (i = 1; i <= min; i++)
3226              {
3227              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3228              switch(*eptr++)
3229                {
3230                default: break;
3231                case 0x0a:      /* LF */
3232                case 0x0b:      /* VT */
3233                case 0x0c:      /* FF */
3234                case 0x0d:      /* CR */
3235                case 0x85:      /* NEL */
3236                RRETURN(MATCH_NOMATCH);
3237                }
3238              }
3239            break;
3240    
3241            case OP_VSPACE:
3242            for (i = 1; i <= min; i++)
3243              {
3244              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3245              switch(*eptr++)
3246                {
3247                default: RRETURN(MATCH_NOMATCH);
3248                case 0x0a:      /* LF */
3249                case 0x0b:      /* VT */
3250                case 0x0c:      /* FF */
3251                case 0x0d:      /* CR */
3252                case 0x85:      /* NEL */
3253                break;
3254                }
3255              }
3256            break;
3257    
3258          case OP_NOT_DIGIT:          case OP_NOT_DIGIT:
3259          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3260            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
# Line 2768  for (;;) Line 3310  for (;;)
3310            case PT_ANY:            case PT_ANY:
3311            for (fi = min;; fi++)            for (fi = min;; fi++)
3312              {              {
3313              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM36);
3314              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3315              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3316              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
3317              if (prop_fail_result) RRETURN(MATCH_NOMATCH);              if (prop_fail_result) RRETURN(MATCH_NOMATCH);
3318              }              }
3319            break;            /* Control never gets here */
3320    
3321            case PT_LAMP:            case PT_LAMP:
3322            for (fi = min;; fi++)            for (fi = min;; fi++)
3323              {              {
3324              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM37);
3325              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3326              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3327              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 2789  for (;;) Line 3331  for (;;)
3331                   prop_chartype == ucp_Lt) == prop_fail_result)                   prop_chartype == ucp_Lt) == prop_fail_result)
3332                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3333              }              }
3334            break;            /* Control never gets here */
3335    
3336            case PT_GC:            case PT_GC:
3337            for (fi = min;; fi++)            for (fi = min;; fi++)
3338              {              {
3339              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM38);
3340              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3341              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3342              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 2802  for (;;) Line 3344  for (;;)
3344              if ((prop_category == prop_value) == prop_fail_result)              if ((prop_category == prop_value) == prop_fail_result)
3345                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3346              }              }
3347            break;            /* Control never gets here */
3348    
3349            case PT_PC:            case PT_PC:
3350            for (fi = min;; fi++)            for (fi = min;; fi++)
3351              {              {
3352              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM39);
3353              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3354              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3355              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 2815  for (;;) Line 3357  for (;;)
3357              if ((prop_chartype == prop_value) == prop_fail_result)              if ((prop_chartype == prop_value) == prop_fail_result)
3358                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3359              }              }
3360            break;            /* Control never gets here */
3361    
3362            case PT_SC:            case PT_SC:
3363            for (fi = min;; fi++)            for (fi = min;; fi++)
3364              {              {
3365              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM40);
3366              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3367              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3368              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 2828  for (;;) Line 3370  for (;;)
3370              if ((prop_script == prop_value) == prop_fail_result)              if ((prop_script == prop_value) == prop_fail_result)
3371                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3372              }              }
3373            break;            /* Control never gets here */
3374    
3375            default:            default:
3376            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
           break;  
3377            }            }
3378          }          }
3379    
# Line 2843  for (;;) Line 3384  for (;;)
3384          {          {
3385          for (fi = min;; fi++)          for (fi = min;; fi++)
3386            {            {
3387            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM41);
3388            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3389            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3390            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
# Line 2872  for (;;) Line 3413  for (;;)
3413          {          {
3414          for (fi = min;; fi++)          for (fi = min;; fi++)
3415            {            {
3416            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42);
3417            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3418            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
3419                 (ctype == OP_ANY && (ims & PCRE_DOTALL) == 0 &&                 (ctype == OP_ANY && (ims & PCRE_DOTALL) == 0 &&
3420                  eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr)))                  IS_NEWLINE(eptr)))
3421              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3422    
3423            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
# Line 2888  for (;;) Line 3429  for (;;)
3429              case OP_ANYBYTE:              case OP_ANYBYTE:
3430              break;              break;
3431    
3432                case OP_ANYNL:
3433                switch(c)
3434                  {
3435                  default: RRETURN(MATCH_NOMATCH);
3436                  case 0x000d:
3437                  if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3438                  break;
3439                  case 0x000a:
3440                  break;
3441    
3442                  case 0x000b:
3443                  case 0x000c:
3444                  case 0x0085:
3445                  case 0x2028:
3446                  case 0x2029:
3447                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3448                  break;
3449                  }
3450                break;
3451    
3452                case OP_NOT_HSPACE:
3453                switch(c)
3454                  {
3455                  default: break;
3456                  case 0x09:      /* HT */
3457                  case 0x20:      /* SPACE */
3458                  case 0xa0:      /* NBSP */
3459                  case 0x1680:    /* OGHAM SPACE MARK */
3460                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3461                  case 0x2000:    /* EN QUAD */
3462                  case 0x2001:    /* EM QUAD */
3463                  case 0x2002:    /* EN SPACE */
3464                  case 0x2003:    /* EM SPACE */
3465                  case 0x2004:    /* THREE-PER-EM SPACE */
3466                  case 0x2005:    /* FOUR-PER-EM SPACE */
3467                  case 0x2006:    /* SIX-PER-EM SPACE */
3468                  case 0x2007:    /* FIGURE SPACE */
3469                  case 0x2008:    /* PUNCTUATION SPACE */
3470                  case 0x2009:    /* THIN SPACE */
3471                  case 0x200A:    /* HAIR SPACE */
3472                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3473                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3474                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3475                  RRETURN(MATCH_NOMATCH);
3476                  }
3477                break;
3478    
3479                case OP_HSPACE:
3480                switch(c)
3481                  {
3482                  default: RRETURN(MATCH_NOMATCH);
3483                  case 0x09:      /* HT */
3484                  case 0x20:      /* SPACE */
3485                  case 0xa0:      /* NBSP */
3486                  case 0x1680:    /* OGHAM SPACE MARK */
3487                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3488                  case 0x2000:    /* EN QUAD */
3489                  case 0x2001:    /* EM QUAD */
3490                  case 0x2002:    /* EN SPACE */
3491                  case 0x2003:    /* EM SPACE */
3492                  case 0x2004:    /* THREE-PER-EM SPACE */
3493                  case 0x2005:    /* FOUR-PER-EM SPACE */
3494                  case 0x2006:    /* SIX-PER-EM SPACE */
3495                  case 0x2007:    /* FIGURE SPACE */
3496                  case 0x2008:    /* PUNCTUATION SPACE */
3497                  case 0x2009:    /* THIN SPACE */
3498                  case 0x200A:    /* HAIR SPACE */
3499                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3500                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3501                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3502                  break;
3503                  }
3504                break;
3505    
3506                case OP_NOT_VSPACE:
3507                switch(c)
3508                  {
3509                  default: break;
3510                  case 0x0a:      /* LF */
3511                  case 0x0b:      /* VT */
3512                  case 0x0c:      /* FF */
3513                  case 0x0d:      /* CR */
3514                  case 0x85:      /* NEL */
3515                  case 0x2028:    /* LINE SEPARATOR */
3516                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3517                  RRETURN(MATCH_NOMATCH);
3518                  }
3519                break;
3520    
3521                case OP_VSPACE:
3522                switch(c)
3523                  {
3524                  default: RRETURN(MATCH_NOMATCH);
3525                  case 0x0a:      /* LF */
3526                  case 0x0b:      /* VT */
3527                  case 0x0c:      /* FF */
3528                  case 0x0d:      /* CR */
3529                  case 0x85:      /* NEL */
3530                  case 0x2028:    /* LINE SEPARATOR */
3531                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3532                  break;
3533                  }
3534                break;
3535    
3536              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
3537              if (c < 256 && (md->ctypes[c] & ctype_digit) != 0)              if (c < 256 && (md->ctypes[c] & ctype_digit) != 0)
3538                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2929  for (;;) Line 3574  for (;;)
3574          {          {
3575          for (fi = min;; fi++)          for (fi = min;; fi++)
3576            {            {
3577            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43);
3578            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3579            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
3580                 ((ims & PCRE_DOTALL) == 0 &&                 ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))
                  eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr)))  
3581              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3582    
3583            c = *eptr++;            c = *eptr++;
# Line 2945  for (;;) Line 3589  for (;;)
3589              case OP_ANYBYTE:              case OP_ANYBYTE:
3590              break;              break;
3591    
3592                case OP_ANYNL:
3593                switch(c)
3594                  {
3595                  default: RRETURN(MATCH_NOMATCH);
3596                  case 0x000d:
3597                  if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3598                  break;
3599    
3600                  case 0x000a:
3601                  break;
3602    
3603                  case 0x000b:
3604                  case 0x000c:
3605                  case 0x0085:
3606                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3607                  break;
3608                  }
3609                break;
3610    
3611                case OP_NOT_HSPACE:
3612                switch(c)
3613                  {
3614                  default: break;
3615                  case 0x09:      /* HT */
3616                  case 0x20:      /* SPACE */
3617                  case 0xa0:      /* NBSP */
3618                  RRETURN(MATCH_NOMATCH);
3619                  }
3620                break;
3621    
3622                case OP_HSPACE:
3623                switch(c)
3624                  {
3625                  default: RRETURN(MATCH_NOMATCH);
3626                  case 0x09:      /* HT */
3627                  case 0x20:      /* SPACE */
3628                  case 0xa0:      /* NBSP */
3629                  break;
3630                  }
3631                break;
3632    
3633                case OP_NOT_VSPACE:
3634                switch(c)
3635                  {
3636                  default: break;
3637                  case 0x0a:      /* LF */
3638                  case 0x0b:      /* VT */
3639                  case 0x0c:      /* FF */
3640                  case 0x0d:      /* CR */
3641                  case 0x85:      /* NEL */
3642                  RRETURN(MATCH_NOMATCH);
3643                  }
3644                break;
3645    
3646                case OP_VSPACE:
3647                switch(c)
3648                  {
3649                  default: RRETURN(MATCH_NOMATCH);
3650                  case 0x0a:      /* LF */
3651                  case 0x0b:      /* VT */
3652                  case 0x0c:      /* FF */
3653                  case 0x0d:      /* CR */
3654                  case 0x85:      /* NEL */
3655                  break;
3656                  }
3657                break;
3658    
3659              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
3660              if ((md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);              if ((md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
3661              break;              break;
# Line 2977  for (;;) Line 3688  for (;;)
3688        /* Control never gets here */        /* Control never gets here */
3689        }        }
3690    
3691      /* 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
3692      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
3693      UTF-8 and UCP stuff separate. */      UTF-8 and UCP stuff separate. */
3694    
# Line 3058  for (;;) Line 3769  for (;;)
3769    
3770          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
3771    
3772            if (possessive) continue;
3773          for(;;)          for(;;)
3774            {            {
3775            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44);
3776            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3777            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3778            BACKCHAR(eptr);            if (utf8) BACKCHAR(eptr);
3779            }            }
3780          }          }
3781    
# Line 3093  for (;;) Line 3805  for (;;)
3805    
3806          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
3807    
3808            if (possessive) continue;
3809          for(;;)          for(;;)
3810            {            {
3811            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM45);
3812            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3813            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3814            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
3815              {              {
3816              int len = 1;              int len = 1;
             BACKCHAR(eptr);  
3817              if (!utf8) c = *eptr; else              if (!utf8) c = *eptr; else
3818                {                {
3819                  BACKCHAR(eptr);
3820                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3821                }                }
3822              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
# Line 3124  for (;;) Line 3837  for (;;)
3837          switch(ctype)          switch(ctype)
3838            {            {
3839            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. */  
   
3840            if (max < INT_MAX)            if (max < INT_MAX)
3841              {              {
3842              if ((ims & PCRE_DOTALL) == 0)              if ((ims & PCRE_DOTALL) == 0)
3843                {                {
3844                for (i = min; i < max; i++)                for (i = min; i < max; i++)
3845                  {                  {
3846                  if (eptr >= md->end_subject ||                  if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
                     (eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr)))  
                   break;  
3847                  eptr++;                  eptr++;
3848                  while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                  while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3849                  }                  }
# Line 3161  for (;;) Line 3867  for (;;)
3867                {                {
3868                for (i = min; i < max; i++)                for (i = min; i < max; i++)
3869                  {                  {
3870                  if (eptr >= md->end_subject ||                  if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
                     (eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr)))  
                   break;  
3871                  eptr++;                  eptr++;
3872                    while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3873                  }                  }
               break;  
3874                }                }
3875              else              else
3876                {                {
3877                c = max - min;                eptr = md->end_subject;
               if (c > md->end_subject - eptr) c = md->end_subject - eptr;  
               eptr += c;  
3878                }                }
3879              }              }
3880            break;            break;
# Line 3181  for (;;) Line 3883  for (;;)
3883    
3884            case OP_ANYBYTE:            case OP_ANYBYTE:
3885            c = max - min;            c = max - min;
3886            if (c > md->end_subject - eptr) c = md->end_subject - eptr;            if (c > (unsigned int)(md->end_subject - eptr))
3887                c = md->end_subject - eptr;
3888            eptr += c;            eptr += c;
3889            break;            break;
3890    
3891              case OP_ANYNL:
3892              for (i = min; i < max; i++)
3893                {
3894                int len = 1;
3895                if (eptr >= md->end_subject) break;
3896                GETCHARLEN(c, eptr, len);
3897                if (c == 0x000d)
3898                  {
3899                  if (++eptr >= md->end_subject) break;
3900                  if (*eptr == 0x000a) eptr++;
3901                  }
3902                else
3903                  {
3904                  if (c != 0x000a &&
3905                      (md->bsr_anycrlf ||
3906                       (c != 0x000b && c != 0x000c &&
3907                        c != 0x0085 && c != 0x2028 && c != 0x2029)))
3908                    break;
3909                  eptr += len;
3910                  }
3911                }
3912              break;
3913    
3914              case OP_NOT_HSPACE:
3915              case OP_HSPACE:
3916              for (i = min; i < max; i++)
3917                {
3918                BOOL gotspace;
3919                int len = 1;
3920                if (eptr >= md->end_subject) break;
3921                GETCHARLEN(c, eptr, len);
3922                switch(c)
3923                  {
3924                  default: gotspace = FALSE; break;
3925                  case 0x09:      /* HT */
3926                  case 0x20:      /* SPACE */
3927                  case 0xa0:      /* NBSP */
3928                  case 0x1680:    /* OGHAM SPACE MARK */
3929                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3930                  case 0x2000:    /* EN QUAD */
3931                  case 0x2001:    /* EM QUAD */
3932                  case 0x2002:    /* EN SPACE */
3933                  case 0x2003:    /* EM SPACE */
3934                  case 0x2004:    /* THREE-PER-EM SPACE */
3935                  case 0x2005:    /* FOUR-PER-EM SPACE */
3936                  case 0x2006:    /* SIX-PER-EM SPACE */
3937                  case 0x2007:    /* FIGURE SPACE */
3938                  case 0x2008:    /* PUNCTUATION SPACE */
3939                  case 0x2009:    /* THIN SPACE */
3940                  case 0x200A:    /* HAIR SPACE */
3941                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3942                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3943                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3944                  gotspace = TRUE;
3945                  break;
3946                  }
3947                if (gotspace == (ctype == OP_NOT_HSPACE)) break;
3948                eptr += len;
3949                }
3950              break;
3951    
3952              case OP_NOT_VSPACE:
3953              case OP_VSPACE:
3954              for (i = min; i < max; i++)
3955                {
3956                BOOL gotspace;
3957                int len = 1;
3958                if (eptr >= md->end_subject) break;
3959                GETCHARLEN(c, eptr, len);
3960                switch(c)
3961                  {
3962                  default: gotspace = FALSE; break;
3963                  case 0x0a:      /* LF */
3964                  case 0x0b:      /* VT */
3965                  case 0x0c:      /* FF */
3966                  case 0x0d:      /* CR */
3967                  case 0x85:      /* NEL */
3968                  case 0x2028:    /* LINE SEPARATOR */
3969                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3970                  gotspace = TRUE;
3971                  break;
3972                  }
3973                if (gotspace == (ctype == OP_NOT_VSPACE)) break;
3974                eptr += len;
3975                }
3976              break;
3977    
3978            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
3979            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3980              {              {
# Line 3257  for (;;) Line 4047  for (;;)
4047    
4048          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
4049    
4050            if (possessive) continue;
4051          for(;;)          for(;;)
4052            {            {
4053            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM46);
4054            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4055            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
4056            BACKCHAR(eptr);            BACKCHAR(eptr);
4057            }            }
4058          }          }
4059        else        else
4060  #endif  #endif  /* SUPPORT_UTF8 */
4061    
4062        /* Not UTF-8 mode */        /* Not UTF-8 mode */
4063          {          {
# Line 3277  for (;;) Line 4068  for (;;)
4068              {              {
4069              for (i = min; i < max; i++)              for (i = min; i < max; i++)
4070                {                {
4071                if (eptr >= md->end_subject ||                if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
                   (eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr)))  
                 break;  
4072                eptr++;                eptr++;
4073                }                }
4074              break;              break;
# Line 3288  for (;;) Line 4077  for (;;)
4077    
4078            case OP_ANYBYTE:            case OP_ANYBYTE:
4079            c = max - min;            c = max - min;
4080            if (c > md->end_subject - eptr) c = md->end_subject - eptr;            if (c > (unsigned int)(md->end_subject - eptr))
4081                c = md->end_subject - eptr;
4082            eptr += c;            eptr += c;
4083            break;            break;
4084    
4085              case OP_ANYNL:
4086              for (i = min; i < max; i++)
4087                {
4088                if (eptr >= md->end_subject) break;
4089                c = *eptr;
4090                if (c == 0x000d)
4091                  {
4092                  if (++eptr >= md->end_subject) break;
4093                  if (*eptr == 0x000a) eptr++;
4094                  }
4095                else
4096                  {
4097                  if (c != 0x000a &&
4098                      (md->bsr_anycrlf ||
4099                        (c != 0x000b && c != 0x000c && c != 0x0085)))
4100                    break;
4101                  eptr++;
4102                  }
4103                }
4104              break;
4105    
4106              case OP_NOT_HSPACE:
4107              for (i = min; i < max; i++)
4108                {
4109                if (eptr >= md->end_subject) break;
4110                c = *eptr;
4111                if (c == 0x09 || c == 0x20 || c == 0xa0) break;
4112                eptr++;
4113                }
4114              break;
4115    
4116              case OP_HSPACE:
4117              for (i = min; i < max; i++)
4118                {
4119                if (eptr >= md->end_subject) break;
4120                c = *eptr;
4121                if (c != 0x09 && c != 0x20 && c != 0xa0) break;
4122                eptr++;
4123                }
4124              break;
4125    
4126              case OP_NOT_VSPACE:
4127              for (i = min; i < max; i++)
4128                {
4129                if (eptr >= md->end_subject) break;
4130                c = *eptr;
4131                if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85)
4132                  break;
4133                eptr++;
4134                }
4135              break;
4136    
4137              case OP_VSPACE:
4138              for (i = min; i < max; i++)
4139                {
4140                if (eptr >= md->end_subject) break;
4141                c = *eptr;
4142                if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85)
4143                  break;
4144                eptr++;
4145                }
4146              break;
4147    
4148            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
4149            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4150              {              {
# Line 3352  for (;;) Line 4205  for (;;)
4205    
4206          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
4207    
4208            if (possessive) continue;
4209          while (eptr >= pp)          while (eptr >= pp)
4210            {            {
4211            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM47);
4212            eptr--;            eptr--;
4213            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4214            }            }
# Line 3366  for (;;) Line 4220  for (;;)
4220        }        }
4221      /* Control never gets here */      /* Control never gets here */
4222    
4223      /* 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
4224      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. */  
4225    
4226      default:      default:
4227      DPRINTF(("Unknown opcode %d\n", *ecode));      DPRINTF(("Unknown opcode %d\n", *ecode));
4228      RRETURN(PCRE_ERROR_UNKNOWN_NODE);      RRETURN(PCRE_ERROR_UNKNOWN_OPCODE);
4229      }      }
4230    
4231    /* 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 3382  for (;;) Line 4234  for (;;)
4234    
4235    }             /* End of main loop */    }             /* End of main loop */
4236  /* Control never reaches here */  /* Control never reaches here */
4237    
4238    
4239    /* When compiling to use the heap rather than the stack for recursive calls to
4240    match(), the RRETURN() macro jumps here. The number that is saved in
4241    frame->Xwhere indicates which label we actually want to return to. */
4242    
4243    #ifdef NO_RECURSE
4244    #define LBL(val) case val: goto L_RM##val;
4245    HEAP_RETURN:
4246    switch (frame->Xwhere)
4247      {
4248      LBL( 1) LBL( 2) LBL( 3) LBL( 4) LBL( 5) LBL( 6) LBL( 7) LBL( 8)
4249      LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(17)
4250      LBL(19) LBL(24) LBL(25) LBL(26) LBL(27) LBL(29) LBL(31) LBL(33)
4251      LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52)
4252      LBL(53) LBL(54)
4253    #ifdef SUPPORT_UTF8
4254      LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30)
4255      LBL(32) LBL(34) LBL(42) LBL(46)
4256    #ifdef SUPPORT_UCP
4257      LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45)
4258    #endif  /* SUPPORT_UCP */
4259    #endif  /* SUPPORT_UTF8 */
4260      default:
4261      DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));
4262      return PCRE_ERROR_INTERNAL;
4263      }
4264    #undef LBL
4265    #endif  /* NO_RECURSE */
4266  }  }
4267    
4268    
# Line 3394  Undefine all the macros that were define Line 4275  Undefine all the macros that were define
4275  #ifdef NO_RECURSE  #ifdef NO_RECURSE
4276  #undef eptr  #undef eptr
4277  #undef ecode  #undef ecode
4278    #undef mstart
4279  #undef offset_top  #undef offset_top
4280  #undef ims  #undef ims
4281  #undef eptrb  #undef eptrb
# Line 3411  Undefine all the macros that were define Line 4293  Undefine all the macros that were define
4293    
4294  #undef cur_is_word  #undef cur_is_word
4295  #undef condition  #undef condition
 #undef minimize  
4296  #undef prev_is_word  #undef prev_is_word
4297    
4298  #undef original_ims  #undef original_ims
# Line 3467  Returns:          > 0 => success; value Line 4348  Returns:          > 0 => success; value
4348                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
4349  */  */
4350    
4351  PCRE_DATA_SCOPE int  PCRE_EXP_DEFN int
4352  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
4353    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
4354    int offsetcount)    int offsetcount)
# Line 3484  BOOL startline; Line 4365  BOOL startline;
4365  BOOL firstline;  BOOL firstline;
4366  BOOL first_byte_caseless = FALSE;  BOOL first_byte_caseless = FALSE;
4367  BOOL req_byte_caseless = FALSE;  BOOL req_byte_caseless = FALSE;
4368    BOOL utf8;
4369  match_data match_block;  match_data match_block;
4370  match_data *md = &match_block;  match_data *md = &match_block;
4371  const uschar *tables;  const uschar *tables;
# Line 3553  if (re->magic_number != MAGIC_NUMBER) Line 4435  if (re->magic_number != MAGIC_NUMBER)
4435  /* Set up other data */  /* Set up other data */
4436    
4437  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;
4438  startline = (re->options & PCRE_STARTLINE) != 0;  startline = (re->flags & PCRE_STARTLINE) != 0;
4439  firstline = (re->options & PCRE_FIRSTLINE) != 0;  firstline = (re->options & PCRE_FIRSTLINE) != 0;
4440    
4441  /* 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. */
# Line 3567  md->end_subject = md->start_subject + le Line 4449  md->end_subject = md->start_subject + le
4449  end_subject = md->end_subject;  end_subject = md->end_subject;
4450    
4451  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;
4452  md->utf8 = (re->options & PCRE_UTF8) != 0;  utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;
4453    
4454  md->notbol = (options & PCRE_NOTBOL) != 0;  md->notbol = (options & PCRE_NOTBOL) != 0;
4455  md->noteol = (options & PCRE_NOTEOL) != 0;  md->noteol = (options & PCRE_NOTEOL) != 0;
# Line 3580  md->recursive = NULL; Line 4462  md->recursive = NULL;
4462  md->lcc = tables + lcc_offset;  md->lcc = tables + lcc_offset;
4463  md->ctypes = tables + ctypes_offset;  md->ctypes = tables + ctypes_offset;
4464    
4465  /* Handle different types of newline. The two bits give four cases. If nothing  /* Handle different \R options. */
 is set at run time, whatever was used at compile time applies. */  
4466    
4467  switch ((((options & PCRE_NEWLINE_CRLF) == 0)? re->options : options) &  switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE))
          PCRE_NEWLINE_CRLF)  
4468    {    {
4469    default:              newline = NEWLINE; break;   /* Compile-time default */    case 0:
4470      if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0)
4471        md->bsr_anycrlf = (re->options & PCRE_BSR_ANYCRLF) != 0;
4472      else
4473    #ifdef BSR_ANYCRLF
4474      md->bsr_anycrlf = TRUE;
4475    #else
4476      md->bsr_anycrlf = FALSE;
4477    #endif
4478      break;
4479    
4480      case PCRE_BSR_ANYCRLF:
4481      md->bsr_anycrlf = TRUE;
4482      break;
4483    
4484      case PCRE_BSR_UNICODE:
4485      md->bsr_anycrlf = FALSE;
4486      break;
4487    
4488      default: return PCRE_ERROR_BADNEWLINE;
4489      }
4490    
4491    /* Handle different types of newline. The three bits give eight cases. If
4492    nothing is set at run time, whatever was used at compile time applies. */
4493    
4494    switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options :
4495            (pcre_uint32)options) & PCRE_NEWLINE_BITS)
4496      {
4497      case 0: newline = NEWLINE; break;   /* Compile-time default */
4498    case PCRE_NEWLINE_CR: newline = '\r'; break;    case PCRE_NEWLINE_CR: newline = '\r'; break;
4499    case PCRE_NEWLINE_LF: newline = '\n'; break;    case PCRE_NEWLINE_LF: newline = '\n'; break;
4500    case PCRE_NEWLINE_CR+    case PCRE_NEWLINE_CR+
4501         PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;         PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;
4502      case PCRE_NEWLINE_ANY: newline = -1; break;
4503      case PCRE_NEWLINE_ANYCRLF: newline = -2; break;
4504      default: return PCRE_ERROR_BADNEWLINE;
4505    }    }
4506    
4507  if (newline > 255)  if (newline == -2)
4508      {
4509      md->nltype = NLTYPE_ANYCRLF;
4510      }
4511    else if (newline < 0)
4512    {    {
4513    md->nllen = 2;    md->nltype = NLTYPE_ANY;
   md->nl[0] = (newline >> 8) & 255;  
   md->nl[1] = newline & 255;  
4514    }    }
4515  else  else
4516    {    {
4517    md->nllen = 1;    md->nltype = NLTYPE_FIXED;
4518    md->nl[0] = newline;    if (newline > 255)
4519        {
4520        md->nllen = 2;
4521        md->nl[0] = (newline >> 8) & 255;
4522        md->nl[1] = newline & 255;
4523        }
4524      else
4525        {
4526        md->nllen = 1;
4527        md->nl[0] = newline;
4528        }
4529    }    }
4530    
4531  /* 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
4532  moment. */  moment. */
4533    
4534  if (md->partial && (re->options & PCRE_NOPARTIAL) != 0)  if (md->partial && (re->flags & PCRE_NOPARTIAL) != 0)
4535    return PCRE_ERROR_BADPARTIAL;    return PCRE_ERROR_BADPARTIAL;
4536    
4537  /* 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
4538  back the character offset. */  back the character offset. */
4539    
4540  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4541  if (md->utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)
4542    {    {
4543    if (_pcre_valid_utf8((uschar *)subject, length) >= 0)    if (_pcre_valid_utf8((uschar *)subject, length) >= 0)
4544      return PCRE_ERROR_BADUTF8;      return PCRE_ERROR_BADUTF8;
# Line 3685  studied, there may be a bitmap of possib Line 4608  studied, there may be a bitmap of possib
4608    
4609  if (!anchored)  if (!anchored)
4610    {    {
4611    if ((re->options & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
4612      {      {
4613      first_byte = re->first_byte & 255;      first_byte = re->first_byte & 255;
4614      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)
# Line 3700  if (!anchored) Line 4623  if (!anchored)
4623  /* For anchored or unanchored matches, there may be a "last known required  /* For anchored or unanchored matches, there may be a "last known required
4624  character" set. */  character" set. */
4625    
4626  if ((re->options & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
4627    {    {
4628    req_byte = re->req_byte & 255;    req_byte = re->req_byte & 255;
4629    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;
4630    req_byte2 = (tables + fcc_offset)[req_byte];  /* case flipped */    req_byte2 = (tables + fcc_offset)[req_byte];  /* case flipped */
4631    }    }
4632    
4633    
4634    /* ==========================================================================*/
4635    
4636  /* Loop for handling unanchored repeated matching attempts; for anchored regexs  /* Loop for handling unanchored repeated matching attempts; for anchored regexs
4637  the loop runs just once. */  the loop runs just once. */
4638    
4639  do  for(;;)
4640    {    {
4641    USPTR save_end_subject = end_subject;    USPTR save_end_subject = end_subject;
4642      USPTR new_start_match;
4643    
4644    /* Reset the maximum number of extractions we might see. */    /* Reset the maximum number of extractions we might see. */
4645    
# Line 3725  do Line 4652  do
4652    
4653    /* 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
4654    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.
4655    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
4656    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
4657    */    the match fails at the newline, later code breaks this loop. */
4658    
4659    if (firstline)    if (firstline)
4660      {      {
4661      USPTR t = start_match;      USPTR t = start_match;
4662      while (t <= save_end_subject - md->nllen && !IS_NEWLINE(t)) t++;      while (t < md->end_subject && !IS_NEWLINE(t)) t++;
4663      end_subject = t;      end_subject = t;
4664      }      }
4665    
# Line 3743  do Line 4670  do
4670      if (first_byte_caseless)      if (first_byte_caseless)
4671        while (start_match < end_subject &&        while (start_match < end_subject &&
4672               md->lcc[*start_match] != first_byte)               md->lcc[*start_match] != first_byte)
4673          start_match++;          { NEXTCHAR(start_match); }
4674      else      else
4675        while (start_match < end_subject && *start_match != first_byte)        while (start_match < end_subject && *start_match != first_byte)
4676          start_match++;          { NEXTCHAR(start_match); }
4677      }      }
4678    
4679    /* Or to just after a linebreak for a multiline match if possible */    /* Or to just after a linebreak for a multiline match if possible */
4680    
4681    else if (startline)    else if (startline)
4682      {      {
4683      if (start_match >= md->start_subject + md->nllen +      if (start_match > md->start_subject + start_offset)
           start_offset)  
4684        {        {
4685        while (start_match <= end_subject &&        while (start_match <= end_subject && !WAS_NEWLINE(start_match))
4686               !IS_NEWLINE(start_match - md->nllen))          { NEXTCHAR(start_match); }
4687    
4688          /* If we have just passed a CR and the newline option is ANY or ANYCRLF,
4689          and we are now at a LF, advance the match position by one more character.
4690          */
4691    
4692          if (start_match[-1] == '\r' &&
4693               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
4694               start_match < end_subject &&
4695               *start_match == '\n')
4696          start_match++;          start_match++;
4697        }        }
4698      }      }
# Line 3769  do Line 4704  do
4704      while (start_match < end_subject)      while (start_match < end_subject)
4705        {        {
4706        register unsigned int c = *start_match;        register unsigned int c = *start_match;
4707        if ((start_bits[c/8] & (1 << (c&7))) == 0) start_match++; else break;        if ((start_bits[c/8] & (1 << (c&7))) == 0)
4708            { NEXTCHAR(start_match); }
4709          else break;
4710        }        }
4711      }      }
4712    
# Line 3793  do Line 4730  do
4730    
4731    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
4732    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
4733    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
4734    don't do this when the string is sufficiently long.    string... so we don't do this when the string is sufficiently long.
4735    
4736    ALSO: this processing is disabled when partial matching is requested.    ALSO: this processing is disabled when partial matching is requested.
4737    */    */
# Line 3826  do Line 4763  do
4763            }            }
4764          }          }
4765    
4766        /* If we can't find the required character, break the matching loop */        /* If we can't find the required character, break the matching loop,
4767          forcing a match failure. */
4768    
4769        if (p >= end_subject) break;        if (p >= end_subject)
4770            {
4771            rc = MATCH_NOMATCH;
4772            break;
4773            }
4774    
4775        /* If we have found the required character, save the point where we        /* If we have found the required character, save the point where we
4776        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 3838  do Line 4780  do
4780        }        }
4781      }      }
4782    
4783    /* When a match occurs, substrings will be set for all internal extractions;    /* OK, we can now run the match. */
   we just need to set up the whole thing as substring 0 before returning. If  
   there were too many extractions, set the return code to zero. In the case  
   where we had to get some local store to hold offsets for backreferences, copy  
   those back references that we can. In this case there need not be overflow  
   if certain parts of the pattern were not used. */  
4784    
4785    md->start_match = start_match;    md->start_match_ptr = start_match;
4786    md->match_call_count = 0;    md->match_call_count = 0;
4787      rc = match(start_match, md->start_code, start_match, 2, md, ims, NULL, 0, 0);
4788    
4789    rc = match(start_match, md->start_code, 2, md, ims, NULL, match_isgroup, 0);    switch(rc)
   
   /* 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. */  
   
   if (rc == MATCH_NOMATCH)  
4790      {      {
4791      if (firstline &&      /* NOMATCH and PRUNE advance by one character. THEN at this level acts
4792          start_match <= md->end_subject - md->nllen &&      exactly like PRUNE. */
4793          IS_NEWLINE(start_match))  
4794        break;      case MATCH_NOMATCH:
4795      start_match++;      case MATCH_PRUNE:
4796        case MATCH_THEN:
4797        new_start_match = start_match + 1;
4798  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4799      if (md->utf8)      if (utf8)
4800        while(start_match < end_subject && (*start_match & 0xc0) == 0x80)        while(new_start_match < end_subject && (*new_start_match & 0xc0) == 0x80)
4801          start_match++;          new_start_match++;
4802  #endif  #endif
4803      continue;      break;
     }  
4804    
4805    if (rc != MATCH_MATCH)      /* SKIP passes back the next starting point explicitly. */
4806      {  
4807      DPRINTF((">>>> error: returning %d\n", rc));      case MATCH_SKIP:
4808      return rc;      new_start_match = md->start_match_ptr;
4809        break;
4810    
4811        /* COMMIT disables the bumpalong, but otherwise behaves as NOMATCH. */
4812    
4813        case MATCH_COMMIT:
4814        rc = MATCH_NOMATCH;
4815        goto ENDLOOP;
4816    
4817        /* Any other return is some kind of error. */
4818    
4819        default:
4820        goto ENDLOOP;
4821      }      }
4822    
4823    /* We have a match! Copy the offset information from temporary store if    /* Control reaches here for the various types of "no match at this point"
4824    necessary */    result. Reset the code to MATCH_NOMATCH for subsequent checking. */
4825    
4826      rc = MATCH_NOMATCH;
4827    
4828      /* If PCRE_FIRSTLINE is set, the match must happen before or at the first
4829      newline in the subject (though it may continue over the newline). Therefore,
4830      if we have just failed to match, starting at a newline, do not continue. */
4831    
4832      if (firstline && IS_NEWLINE(start_match)) break;
4833    
4834      /* Advance to new matching position */
4835    
4836      start_match = new_start_match;
4837    
4838      /* Break the loop if the pattern is anchored or if we have passed the end of
4839      the subject. */
4840    
4841      if (anchored || start_match > end_subject) break;
4842    
4843      /* If we have just passed a CR and we are now at a LF, and the pattern does
4844      not contain any explicit matches for \r or \n, and the newline option is CRLF
4845      or ANY or ANYCRLF, advance the match position by one more character. */
4846    
4847      if (start_match[-1] == '\r' &&
4848          start_match < end_subject &&
4849          *start_match == '\n' &&
4850          (re->flags & PCRE_HASCRORLF) == 0 &&
4851            (md->nltype == NLTYPE_ANY ||
4852             md->nltype == NLTYPE_ANYCRLF ||
4853             md->nllen == 2))
4854        start_match++;
4855    
4856      }   /* End of for(;;) "bumpalong" loop */
4857    
4858    /* ==========================================================================*/
4859    
4860    /* We reach here when rc is not MATCH_NOMATCH, or if one of the stopping
4861    conditions is true:
4862    
4863    (1) The pattern is anchored or the match was failed by (*COMMIT);
4864    
4865    (2) We are past the end of the subject;
4866    
4867    (3) PCRE_FIRSTLINE is set and we have failed to match at a newline, because
4868        this option requests that a match occur at or before the first newline in
4869        the subject.
4870    
4871    When we have a match and the offset vector is big enough to deal with any
4872    backreferences, captured substring offsets will already be set up. In the case
4873    where we had to get some local store to hold offsets for backreference
4874    processing, copy those that we can. In this case there need not be overflow if
4875    certain parts of the pattern were not used, even though there are more
4876    capturing parentheses than vector slots. */
4877    
4878    ENDLOOP:
4879    
4880    if (rc == MATCH_MATCH)
4881      {
4882    if (using_temporary_offsets)    if (using_temporary_offsets)
4883      {      {
4884      if (offsetcount >= 4)      if (offsetcount >= 4)
# Line 3889  do Line 4887  do
4887          (offsetcount - 2) * sizeof(int));          (offsetcount - 2) * sizeof(int));
4888        DPRINTF(("Copied offsets from temporary memory\n"));        DPRINTF(("Copied offsets from temporary memory\n"));
4889        }        }
4890      if (md->end_offset_top > offsetcount)      if (md->end_offset_top > offsetcount) md->offset_overflow = TRUE;
       md->offset_overflow = TRUE;  
   
4891      DPRINTF(("Freeing temporary memory\n"));      DPRINTF(("Freeing temporary memory\n"));
4892      (pcre_free)(md->offset_vector);      (pcre_free)(md->offset_vector);
4893      }      }
4894    
4895      /* Set the return code to the number of captured strings, or 0 if there are
4896      too many to fit into the vector. */
4897    
4898    rc = md->offset_overflow? 0 : md->end_offset_top/2;    rc = md->offset_overflow? 0 : md->end_offset_top/2;
4899    
4900      /* If there is space, set up the whole thing as substring 0. The value of
4901      md->start_match_ptr might be modified if \K was encountered on the success
4902      matching path. */
4903    
4904    if (offsetcount < 2) rc = 0; else    if (offsetcount < 2) rc = 0; else
4905      {      {
4906      offsets[0] = start_match - md->start_subject;      offsets[0] = md->start_match_ptr - md->start_subject;
4907      offsets[1] = md->end_match_ptr - md->start_subject;      offsets[1] = md->end_match_ptr - md->start_subject;
4908      }      }
4909    
# Line 3908  do Line 4911  do
4911    return rc;    return rc;
4912    }    }
4913    
4914  /* This "while" is the end of the "do" above */  /* Control gets here if there has been an error, or if the overall match
4915    attempt has failed at all permitted starting positions. */
 while (!anchored && start_match <= end_subject);  
4916    
4917  if (using_temporary_offsets)  if (using_temporary_offsets)
4918    {    {
# Line 3918  if (using_temporary_offsets) Line 4920  if (using_temporary_offsets)
4920    (pcre_free)(md->offset_vector);    (pcre_free)(md->offset_vector);
4921    }    }
4922    
4923  if (md->partial && md->hitend)  if (rc != MATCH_NOMATCH)
4924      {
4925      DPRINTF((">>>> error: returning %d\n", rc));
4926      return rc;
4927      }
4928    else if (md->partial && md->hitend)
4929    {    {
4930    DPRINTF((">>>> returning PCRE_ERROR_PARTIAL\n"));    DPRINTF((">>>> returning PCRE_ERROR_PARTIAL\n"));
4931    return PCRE_ERROR_PARTIAL;    return PCRE_ERROR_PARTIAL;

Legend:
Removed from v.91  
changed lines
  Added in v.294

  ViewVC Help
Powered by ViewVC 1.1.5