/[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 77 by nigel, Sat Feb 24 21:40:45 2007 UTC revision 444 by ph10, Sun Sep 13 16:26:39 2009 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) 1997-2005 University of Cambridge             Copyright (c) 1997-2009 University of Cambridge
10    
11  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
12  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 42  POSSIBILITY OF SUCH DAMAGE. Line 42  POSSIBILITY OF SUCH DAMAGE.
42  pattern matching using an NFA algorithm, trying to mimic Perl as closely as  pattern matching using an NFA algorithm, trying to mimic Perl as closely as
43  possible. There are also some static supporting functions. */  possible. There are also some static supporting functions. */
44    
45    #ifdef HAVE_CONFIG_H
46    #include "config.h"
47    #endif
48    
49    #define NLBLOCK md             /* Block containing newline information */
50    #define PSSTART start_subject  /* Field containing processed string start */
51    #define PSEND   end_subject    /* Field containing processed string end */
52    
53  #include "pcre_internal.h"  #include "pcre_internal.h"
54    
55    /* Undefine some potentially clashing cpp symbols */
56    
57  /* Structure for building a chain of data that actually lives on the  #undef min
58  stack, for holding the values of the subject pointer at the start of each  #undef max
 subpattern, so as to detect when an empty string has been matched by a  
 subpattern - to break infinite loops. When NO_RECURSE is set, these blocks  
 are on the heap, not on the stack. */  
   
 typedef struct eptrblock {  
   struct eptrblock *epb_prev;  
   const uschar *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 128  Returns: TRUE if matched Line 136  Returns: TRUE if matched
136  */  */
137    
138  static BOOL  static BOOL
139  match_ref(int offset, register const uschar *eptr, int length, match_data *md,  match_ref(int offset, register USPTR eptr, int length, match_data *md,
140    unsigned long int ims)    unsigned long int ims)
141  {  {
142  const uschar *p = md->start_subject + md->offset_vector[offset];  USPTR p = md->start_subject + md->offset_vector[offset];
143    
144  #ifdef DEBUG  #ifdef DEBUG
145  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 150  printf("\n"); Line 158  printf("\n");
158    
159  if (length > md->end_subject - eptr) return FALSE;  if (length > md->end_subject - eptr) return FALSE;
160    
161  /* Separate the caselesss case for speed */  /* Separate the caseless case for speed. In UTF-8 mode we can only do this
162    properly if Unicode properties are supported. Otherwise, we can check only
163    ASCII characters. */
164    
165  if ((ims & PCRE_CASELESS) != 0)  if ((ims & PCRE_CASELESS) != 0)
166    {    {
167    #ifdef SUPPORT_UTF8
168    #ifdef SUPPORT_UCP
169      if (md->utf8)
170        {
171        USPTR endptr = eptr + length;
172        while (eptr < endptr)
173          {
174          int c, d;
175          GETCHARINC(c, eptr);
176          GETCHARINC(d, p);
177          if (c != d && c != UCD_OTHERCASE(d)) return FALSE;
178          }
179        }
180      else
181    #endif
182    #endif
183    
184      /* The same code works when not in UTF-8 mode and in UTF-8 mode when there
185      is no UCP support. */
186    
187    while (length-- > 0)    while (length-- > 0)
188      if (md->lcc[*p++] != md->lcc[*eptr++]) return FALSE;      { if (md->lcc[*p++] != md->lcc[*eptr++]) return FALSE; }
189    }    }
190    
191    /* In the caseful case, we can just compare the bytes, whether or not we
192    are in UTF-8 mode. */
193    
194  else  else
195    { while (length-- > 0) if (*p++ != *eptr++) return FALSE; }    { while (length-- > 0) if (*p++ != *eptr++) return FALSE; }
196    
# Line 169  return TRUE; Line 203  return TRUE;
203  ****************************************************************************  ****************************************************************************
204                     RECURSION IN THE match() FUNCTION                     RECURSION IN THE match() FUNCTION
205    
206  The match() function is highly recursive. Some regular expressions can cause  The match() function is highly recursive, though not every recursive call
207  it to recurse thousands of times. I was writing for Unix, so I just let it  increases the recursive depth. Nevertheless, some regular expressions can cause
208  call itself recursively. This uses the stack for saving everything that has  it to recurse to a great depth. I was writing for Unix, so I just let it call
209  to be saved for a recursive call. On Unix, the stack can be large, and this  itself recursively. This uses the stack for saving everything that has to be
210  works fine.  saved for a recursive call. On Unix, the stack can be large, and this works
211    fine.
212  It turns out that on non-Unix systems there are problems with programs that  
213  use a lot of stack. (This despite the fact that every last chip has oodles  It turns out that on some non-Unix-like systems there are problems with
214  of memory these days, and techniques for extending the stack have been known  programs that use a lot of stack. (This despite the fact that every last chip
215  for decades.) So....  has oodles of memory these days, and techniques for extending the stack have
216    been known for decades.) So....
217    
218  There is a fudge, triggered by defining NO_RECURSE, which avoids recursive  There is a fudge, triggered by defining NO_RECURSE, which avoids recursive
219  calls by keeping local variables that need to be preserved in blocks of memory  calls by keeping local variables that need to be preserved in blocks of memory
220  obtained from malloc instead instead of on the stack. Macros are used to  obtained from malloc() instead instead of on the stack. Macros are used to
221  achieve this so that the actual code doesn't look very different to what it  achieve this so that the actual code doesn't look very different to what it
222  always used to.  always used to.
223    
224    The original heap-recursive code used longjmp(). However, it seems that this
225    can be very slow on some operating systems. Following a suggestion from Stan
226    Switzer, the use of longjmp() has been abolished, at the cost of having to
227    provide a unique number for each call to RMATCH. There is no way of generating
228    a sequence of numbers at compile time in C. I have given them names, to make
229    them stand out more clearly.
230    
231    Crude tests on x86 Linux show a small speedup of around 5-8%. However, on
232    FreeBSD, avoiding longjmp() more than halves the time taken to run the standard
233    tests. Furthermore, not using longjmp() means that local dynamic variables
234    don't have indeterminate values; this has meant that the frame size can be
235    reduced because the result can be "passed back" by straight setting of the
236    variable instead of being passed in the frame.
237  ****************************************************************************  ****************************************************************************
238  ***************************************************************************/  ***************************************************************************/
239    
240    /* Numbers for RMATCH calls. When this list is changed, the code at HEAP_RETURN
241    below must be updated in sync.  */
242    
243  /* These versions of the macros use the stack, as normal */  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM6,  RM7,  RM8,  RM9,  RM10,
244           RM11,  RM12, RM13, RM14, RM15, RM16, RM17, RM18, RM19, RM20,
245           RM21,  RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30,
246           RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
247           RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
248           RM51,  RM52, RM53, RM54 };
249    
250    /* These versions of the macros use the stack, as normal. There are debugging
251    versions and production versions. Note that the "rw" argument of RMATCH isn't
252    actuall used in this definition. */
253    
254  #ifndef NO_RECURSE  #ifndef NO_RECURSE
255  #define REGISTER register  #define REGISTER register
256  #define RMATCH(rx,ra,rb,rc,rd,re,rf,rg) rx = match(ra,rb,rc,rd,re,rf,rg)  
257    #ifdef DEBUG
258    #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \
259      { \
260      printf("match() called in line %d\n", __LINE__); \
261      rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1); \
262      printf("to line %d\n", __LINE__); \
263      }
264    #define RRETURN(ra) \
265      { \
266      printf("match() returned %d from line %d ", ra, __LINE__); \
267      return ra; \
268      }
269    #else
270    #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \
271      rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1)
272  #define RRETURN(ra) return ra  #define RRETURN(ra) return ra
273    #endif
274    
275  #else  #else
276    
277    
278  /* These versions of the macros manage a private stack on the heap. Note  /* These versions of the macros manage a private stack on the heap. Note that
279  that the rd argument of RMATCH isn't actually used. It's the md argument of  the "rd" argument of RMATCH isn't actually used in this definition. It's the md
280  match(), which never changes. */  argument of match(), which never changes. */
281    
282  #define REGISTER  #define REGISTER
283    
284  #define RMATCH(rx,ra,rb,rc,rd,re,rf,rg)\  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw)\
285    {\    {\
286    heapframe *newframe = (pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = (pcre_stack_malloc)(sizeof(heapframe));\
287    if (setjmp(frame->Xwhere) == 0)\    frame->Xwhere = rw; \
288      {\    newframe->Xeptr = ra;\
289      newframe->Xeptr = ra;\    newframe->Xecode = rb;\
290      newframe->Xecode = rb;\    newframe->Xmstart = mstart;\
291      newframe->Xoffset_top = rc;\    newframe->Xoffset_top = rc;\
292      newframe->Xims = re;\    newframe->Xims = re;\
293      newframe->Xeptrb = rf;\    newframe->Xeptrb = rf;\
294      newframe->Xflags = rg;\    newframe->Xflags = rg;\
295      newframe->Xprevframe = frame;\    newframe->Xrdepth = frame->Xrdepth + 1;\
296      frame = newframe;\    newframe->Xprevframe = frame;\
297      DPRINTF(("restarting from line %d\n", __LINE__));\    frame = newframe;\
298      goto HEAP_RECURSE;\    DPRINTF(("restarting from line %d\n", __LINE__));\
299      }\    goto HEAP_RECURSE;\
300    else\    L_##rw:\
301      {\    DPRINTF(("jumped back to line %d\n", __LINE__));\
     DPRINTF(("longjumped back to line %d\n", __LINE__));\  
     frame = md->thisframe;\  
     rx = frame->Xresult;\  
     }\  
302    }    }
303    
304  #define RRETURN(ra)\  #define RRETURN(ra)\
# Line 235  match(), which never changes. */ Line 308  match(), which never changes. */
308    (pcre_stack_free)(newframe);\    (pcre_stack_free)(newframe);\
309    if (frame != NULL)\    if (frame != NULL)\
310      {\      {\
311      frame->Xresult = ra;\      rrc = ra;\
312      md->thisframe = frame;\      goto HEAP_RETURN;\
     longjmp(frame->Xwhere, 1);\  
313      }\      }\
314    return ra;\    return ra;\
315    }    }
# Line 250  typedef struct heapframe { Line 322  typedef struct heapframe {
322    
323    /* Function arguments that may change */    /* Function arguments that may change */
324    
325    const uschar *Xeptr;    USPTR Xeptr;
326    const uschar *Xecode;    const uschar *Xecode;
327      USPTR Xmstart;
328    int Xoffset_top;    int Xoffset_top;
329    long int Xims;    long int Xims;
330    eptrblock *Xeptrb;    eptrblock *Xeptrb;
331    int Xflags;    int Xflags;
332      unsigned int Xrdepth;
333    
334    /* Function local variables */    /* Function local variables */
335    
336    const uschar *Xcallpat;    USPTR Xcallpat;
337    const uschar *Xcharptr;  #ifdef SUPPORT_UTF8
338    const uschar *Xdata;    USPTR Xcharptr;
339    const uschar *Xnext;  #endif
340    const uschar *Xpp;    USPTR Xdata;
341    const uschar *Xprev;    USPTR Xnext;
342    const uschar *Xsaved_eptr;    USPTR Xpp;
343      USPTR Xprev;
344      USPTR Xsaved_eptr;
345    
346    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
347    
348    BOOL Xcur_is_word;    BOOL Xcur_is_word;
349    BOOL Xcondition;    BOOL Xcondition;
   BOOL Xminimize;  
350    BOOL Xprev_is_word;    BOOL Xprev_is_word;
351    
352    unsigned long int Xoriginal_ims;    unsigned long int Xoriginal_ims;
353    
354  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
355    int Xprop_type;    int Xprop_type;
356      int Xprop_value;
357    int Xprop_fail_result;    int Xprop_fail_result;
358    int Xprop_category;    int Xprop_category;
359    int Xprop_chartype;    int Xprop_chartype;
360    int Xprop_othercase;    int Xprop_script;
361    int Xprop_test_against;    int Xoclength;
362    int *Xprop_test_variable;    uschar Xocchars[8];
363  #endif  #endif
364    
365      int Xcodelink;
366    int Xctype;    int Xctype;
367    int Xfc;    unsigned int Xfc;
368    int Xfi;    int Xfi;
369    int Xlength;    int Xlength;
370    int Xmax;    int Xmax;
# Line 301  typedef struct heapframe { Line 378  typedef struct heapframe {
378    
379    eptrblock Xnewptrb;    eptrblock Xnewptrb;
380    
381    /* Place to pass back result, and where to jump back to */    /* Where to jump back to */
382    
383    int  Xresult;    int Xwhere;
   jmp_buf Xwhere;  
384    
385  } heapframe;  } heapframe;
386    
# Line 320  typedef struct heapframe { Line 396  typedef struct heapframe {
396  *         Match from current position            *  *         Match from current position            *
397  *************************************************/  *************************************************/
398    
399  /* 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  
400  returns a negative (error) response, the outer incarnation must also return the  returns a negative (error) response, the outer incarnation must also return the
401  same response.  same response. */
402    
403    /* These macros pack up tests that are used for partial matching, and which
404    appears several times in the code. We set the "hit end" flag if the pointer is
405    at the end of the subject and also past the start of the subject (i.e.
406    something has been matched). For hard partial matching, we then return
407    immediately. The second one is used when we already know we are past the end of
408    the subject. */
409    
410    #define CHECK_PARTIAL()\
411      if (md->partial != 0 && eptr >= md->end_subject && eptr > mstart)\
412        {\
413        md->hitend = TRUE;\
414        if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);\
415        }
416    
417    #define SCHECK_PARTIAL()\
418      if (md->partial && eptr > mstart)\
419        {\
420        md->hitend = TRUE;\
421        if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);\
422        }
423    
424  Performance note: It might be tempting to extract commonly used fields from the  
425  md structure (e.g. utf8, end_subject) into individual variables to improve  /* Performance note: It might be tempting to extract commonly used fields from
426    the md structure (e.g. utf8, end_subject) into individual variables to improve
427  performance. Tests using gcc on a SPARC disproved this; in the first case, it  performance. Tests using gcc on a SPARC disproved this; in the first case, it
428  made performance worse.  made performance worse.
429    
430  Arguments:  Arguments:
431     eptr        pointer in subject     eptr        pointer to current character in subject
432     ecode       position in code     ecode       pointer to current position in compiled code
433       mstart      pointer to the current match start position (can be modified
434                     by encountering \K)
435     offset_top  current top pointer     offset_top  current top pointer
436     md          pointer to "static" info for the match     md          pointer to "static" info for the match
437     ims         current /i, /m, and /s options     ims         current /i, /m, and /s options
# Line 342  Arguments: Line 439  Arguments:
439                   brackets - for testing for empty matches                   brackets - for testing for empty matches
440     flags       can contain     flags       can contain
441                   match_condassert - this is an assertion condition                   match_condassert - this is an assertion condition
442                   match_isgroup - this is the start of a bracketed group                   match_cbegroup - this is the start of an unlimited repeat
443                       group that can match an empty string
444       rdepth      the recursion depth
445    
446  Returns:       MATCH_MATCH if matched            )  these values are >= 0  Returns:       MATCH_MATCH if matched            )  these values are >= 0
447                 MATCH_NOMATCH if failed to match  )                 MATCH_NOMATCH if failed to match  )
448                 a negative PCRE_ERROR_xxx value if aborted by an error condition                 a negative PCRE_ERROR_xxx value if aborted by an error condition
449                   (e.g. stopped by recursion limit)                   (e.g. stopped by repeated call or recursion limit)
450  */  */
451    
452  static int  static int
453  match(REGISTER const uschar *eptr, REGISTER const uschar *ecode,  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,
454    int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb,    int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb,
455    int flags)    int flags, unsigned int rdepth)
456  {  {
457  /* 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,
458  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
459  because they are used a lot in loops. */  "register" because they are used a lot in loops. */
460    
461    register int  rrc;         /* Returns from recursive calls */
462    register int  i;           /* Used for loops not involving calls to RMATCH() */
463    register unsigned int c;   /* Character values not kept over RMATCH() calls */
464    register BOOL utf8;        /* Local copy of UTF-8 flag for speed */
465    
466  register int  rrc;    /* Returns from recursive calls */  BOOL minimize, possessive; /* Quantifier options */
467  register int  i;      /* Used for loops not involving calls to RMATCH() */  int condcode;
 register int  c;      /* Character values not kept over RMATCH() calls */  
 register BOOL utf8;   /* Local copy of UTF-8 flag for speed */  
468    
469  /* 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
470  preserved over calls to RMATCH() are part of a "frame" which is obtained from  preserved over calls to RMATCH() are part of a "frame" which is obtained from
# Line 377  frame->Xprevframe = NULL; /* Line 479  frame->Xprevframe = NULL; /*
479    
480  frame->Xeptr = eptr;  frame->Xeptr = eptr;
481  frame->Xecode = ecode;  frame->Xecode = ecode;
482    frame->Xmstart = mstart;
483  frame->Xoffset_top = offset_top;  frame->Xoffset_top = offset_top;
484  frame->Xims = ims;  frame->Xims = ims;
485  frame->Xeptrb = eptrb;  frame->Xeptrb = eptrb;
486  frame->Xflags = flags;  frame->Xflags = flags;
487    frame->Xrdepth = rdepth;
488    
489  /* This is where control jumps back to to effect "recursion" */  /* This is where control jumps back to to effect "recursion" */
490    
# Line 390  HEAP_RECURSE: Line 494  HEAP_RECURSE:
494    
495  #define eptr               frame->Xeptr  #define eptr               frame->Xeptr
496  #define ecode              frame->Xecode  #define ecode              frame->Xecode
497    #define mstart             frame->Xmstart
498  #define offset_top         frame->Xoffset_top  #define offset_top         frame->Xoffset_top
499  #define ims                frame->Xims  #define ims                frame->Xims
500  #define eptrb              frame->Xeptrb  #define eptrb              frame->Xeptrb
501  #define flags              frame->Xflags  #define flags              frame->Xflags
502    #define rdepth             frame->Xrdepth
503    
504  /* Ditto for the local variables */  /* Ditto for the local variables */
505    
# Line 401  HEAP_RECURSE: Line 507  HEAP_RECURSE:
507  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
508  #endif  #endif
509  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
510    #define codelink           frame->Xcodelink
511  #define data               frame->Xdata  #define data               frame->Xdata
512  #define next               frame->Xnext  #define next               frame->Xnext
513  #define pp                 frame->Xpp  #define pp                 frame->Xpp
# Line 411  HEAP_RECURSE: Line 518  HEAP_RECURSE:
518    
519  #define cur_is_word        frame->Xcur_is_word  #define cur_is_word        frame->Xcur_is_word
520  #define condition          frame->Xcondition  #define condition          frame->Xcondition
 #define minimize           frame->Xminimize  
521  #define prev_is_word       frame->Xprev_is_word  #define prev_is_word       frame->Xprev_is_word
522    
523  #define original_ims       frame->Xoriginal_ims  #define original_ims       frame->Xoriginal_ims
524    
525  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
526  #define prop_type          frame->Xprop_type  #define prop_type          frame->Xprop_type
527    #define prop_value         frame->Xprop_value
528  #define prop_fail_result   frame->Xprop_fail_result  #define prop_fail_result   frame->Xprop_fail_result
529  #define prop_category      frame->Xprop_category  #define prop_category      frame->Xprop_category
530  #define prop_chartype      frame->Xprop_chartype  #define prop_chartype      frame->Xprop_chartype
531  #define prop_othercase     frame->Xprop_othercase  #define prop_script        frame->Xprop_script
532  #define prop_test_against  frame->Xprop_test_against  #define oclength           frame->Xoclength
533  #define prop_test_variable frame->Xprop_test_variable  #define occhars            frame->Xocchars
534  #endif  #endif
535    
536  #define ctype              frame->Xctype  #define ctype              frame->Xctype
# Line 447  HEAP_RECURSE: Line 554  HEAP_RECURSE:
554  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
555  i, and fc and c, can be the same variables. */  i, and fc and c, can be the same variables. */
556    
557  #else  #else         /* NO_RECURSE not defined */
558  #define fi i  #define fi i
559  #define fc c  #define fc c
560    
561    
562  #ifdef SUPPORT_UTF8                /* Many of these variables are used ony */  #ifdef SUPPORT_UTF8                /* Many of these variables are used only  */
563  const uschar *charptr;             /* small blocks of the code. My normal  */  const uschar *charptr;             /* in small blocks of the code. My normal */
564  #endif                             /* style of coding would have declared  */  #endif                             /* style of coding would have declared    */
565  const uschar *callpat;             /* them within each of those blocks.    */  const uschar *callpat;             /* them within each of those blocks.      */
566  const uschar *data;                /* However, in order to accommodate the */  const uschar *data;                /* However, in order to accommodate the   */
567  const uschar *next;                /* version of this code that uses an    */  const uschar *next;                /* version of this code that uses an      */
568  const uschar *pp;                  /* external "stack" implemented on the  */  USPTR         pp;                  /* external "stack" implemented on the    */
569  const uschar *prev;                /* heap, it is easier to declare them   */  const uschar *prev;                /* heap, it is easier to declare them all */
570  const uschar *saved_eptr;          /* all here, so the declarations can    */  USPTR         saved_eptr;          /* here, so the declarations can be cut   */
571                                     /* be cut out in a block. The only      */                                     /* out in a block. The only declarations  */
572  recursion_info new_recursive;      /* declarations within blocks below are */  recursion_info new_recursive;      /* within blocks below are for variables  */
573                                     /* for variables that do not have to    */                                     /* that do not have to be preserved over  */
574  BOOL cur_is_word;                  /* be preserved over a recursive call   */  BOOL cur_is_word;                  /* a recursive call to RMATCH().          */
575  BOOL condition;                    /* to RMATCH().                         */  BOOL condition;
 BOOL minimize;  
576  BOOL prev_is_word;  BOOL prev_is_word;
577    
578  unsigned long int original_ims;  unsigned long int original_ims;
579    
580  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
581  int prop_type;  int prop_type;
582    int prop_value;
583  int prop_fail_result;  int prop_fail_result;
584  int prop_category;  int prop_category;
585  int prop_chartype;  int prop_chartype;
586  int prop_othercase;  int prop_script;
587  int prop_test_against;  int oclength;
588  int *prop_test_variable;  uschar occhars[8];
589  #endif  #endif
590    
591    int codelink;
592  int ctype;  int ctype;
593  int length;  int length;
594  int max;  int max;
# Line 493  int save_offset1, save_offset2, save_off Line 601  int save_offset1, save_offset2, save_off
601  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
602    
603  eptrblock newptrb;  eptrblock newptrb;
604  #endif  #endif     /* NO_RECURSE */
605    
606  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
607  variables. */  variables. */
608    
609  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
610    prop_value = 0;
611  prop_fail_result = 0;  prop_fail_result = 0;
 prop_test_against = 0;  
 prop_test_variable = NULL;  
612  #endif  #endif
613    
614  /* OK, now we can get on with the real code of the function. Recursion is  
615  specified by the macros RMATCH and RRETURN. When NO_RECURSE is *not* defined,  /* This label is used for tail recursion, which is used in a few cases even
616  these just turn into a recursive call to match() and a "return", respectively.  when NO_RECURSE is not defined, in order to reduce the amount of stack that is
617  However, RMATCH isn't like a function call because it's quite a complicated  used. Thanks to Ian Taylor for noticing this possibility and sending the
618  macro. It has to be used in one particular way. This shouldn't, however, impact  original patch. */
619  performance when true recursion is being used. */  
620    TAIL_RECURSE:
621    
622    /* OK, now we can get on with the real code of the function. Recursive calls
623    are specified by the macro RMATCH and RRETURN is used to return. When
624    NO_RECURSE is *not* defined, these just turn into a recursive call to match()
625    and a "return", respectively (possibly with some debugging if DEBUG is
626    defined). However, RMATCH isn't like a function call because it's quite a
627    complicated macro. It has to be used in one particular way. This shouldn't,
628    however, impact performance when true recursion is being used. */
629    
630    #ifdef SUPPORT_UTF8
631    utf8 = md->utf8;       /* Local copy of the flag */
632    #else
633    utf8 = FALSE;
634    #endif
635    
636    /* First check that we haven't called match() too many times, or that we
637    haven't exceeded the recursive call limit. */
638    
639  if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT);  if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT);
640    if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);
641    
642  original_ims = ims;    /* Save for resetting on ')' */  original_ims = ims;    /* Save for resetting on ')' */
 utf8 = md->utf8;       /* Local copy of the flag */  
643    
644  /* At the start of a bracketed group, add the current subject pointer to the  /* At the start of a group with an unlimited repeat that may match an empty
645  stack of such pointers, to be re-instated at the end of the group when we hit  string, the match_cbegroup flag is set. When this is the case, add the current
646  the closing ket. When match() is called in other circumstances, we don't add to  subject pointer to the chain of such remembered pointers, to be checked when we
647  this stack. */  hit the closing ket, in order to break infinite loops that match no characters.
648    When match() is called in other circumstances, don't add to the chain. The
649    match_cbegroup flag must NOT be used with tail recursion, because the memory
650    block that is used is on the stack, so a new one may be required for each
651    match(). */
652    
653  if ((flags & match_isgroup) != 0)  if ((flags & match_cbegroup) != 0)
654    {    {
   newptrb.epb_prev = eptrb;  
655    newptrb.epb_saved_eptr = eptr;    newptrb.epb_saved_eptr = eptr;
656      newptrb.epb_prev = eptrb;
657    eptrb = &newptrb;    eptrb = &newptrb;
658    }    }
659    
660  /* Now start processing the operations. */  /* Now start processing the opcodes. */
661    
662  for (;;)  for (;;)
663    {    {
664      minimize = possessive = FALSE;
665    op = *ecode;    op = *ecode;
   minimize = FALSE;  
   
   /* For partial matching, remember if we ever hit the end of the subject after  
   matching at least one subject character. */  
   
   if (md->partial &&  
       eptr >= md->end_subject &&  
       eptr > md->start_match)  
     md->hitend = TRUE;  
   
   /* Opening capturing bracket. If there is space in the offset vector, save  
   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. */  
666    
667    if (op > OP_BRA)    switch(op)
668      {      {
669      number = op - OP_BRA;      case OP_FAIL:
670        RRETURN(MATCH_NOMATCH);
     /* For extended extraction brackets (large number), we have to fish out the  
     number from a dummy opcode at the start. */  
671    
672      if (number > EXTRACT_BASIC_MAX)      case OP_PRUNE:
673        number = GET2(ecode, 2+LINK_SIZE);      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
674          ims, eptrb, flags, RM51);
675        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
676        RRETURN(MATCH_PRUNE);
677    
678        case OP_COMMIT:
679        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
680          ims, eptrb, flags, RM52);
681        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
682        RRETURN(MATCH_COMMIT);
683    
684        case OP_SKIP:
685        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
686          ims, eptrb, flags, RM53);
687        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
688        md->start_match_ptr = eptr;   /* Pass back current position */
689        RRETURN(MATCH_SKIP);
690    
691        case OP_THEN:
692        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
693          ims, eptrb, flags, RM54);
694        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
695        RRETURN(MATCH_THEN);
696    
697        /* Handle a capturing bracket. If there is space in the offset vector, save
698        the current subject position in the working slot at the top of the vector.
699        We mustn't change the current values of the data slot, because they may be
700        set from a previous iteration of this group, and be referred to by a
701        reference inside the group.
702    
703        If the bracket fails to match, we need to restore this value and also the
704        values of the final offsets, in case they were set by a previous iteration
705        of the same bracket.
706    
707        If there isn't enough space in the offset vector, treat this as if it were
708        a non-capturing bracket. Don't worry about setting the flag for the error
709        case here; that is handled in the code for KET. */
710    
711        case OP_CBRA:
712        case OP_SCBRA:
713        number = GET2(ecode, 1+LINK_SIZE);
714      offset = number << 1;      offset = number << 1;
715    
716  #ifdef DEBUG  #ifdef DEBUG
717      printf("start bracket %d subject=", number);      printf("start bracket %d\n", number);
718        printf("subject=");
719      pchars(eptr, 16, TRUE, md);      pchars(eptr, 16, TRUE, md);
720      printf("\n");      printf("\n");
721  #endif  #endif
# Line 584  for (;;) Line 730  for (;;)
730        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
731        md->offset_vector[md->offset_end - number] = eptr - md->start_subject;        md->offset_vector[md->offset_end - number] = eptr - md->start_subject;
732    
733          flags = (op == OP_SCBRA)? match_cbegroup : 0;
734        do        do
735          {          {
736          RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb,          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
737            match_isgroup);            ims, eptrb, flags, RM1);
738          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
739          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
740          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
741          }          }
# Line 603  for (;;) Line 750  for (;;)
750        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
751        }        }
752    
753      /* Insufficient room for saving captured contents */      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
754        as a non-capturing bracket. */
755    
756      else op = OP_BRA;      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
757      }      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
758    
759    /* Other types of node can be handled by a switch */      DPRINTF(("insufficient capture room: treat as non-capturing\n"));
760    
761    switch(op)      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
762      {      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
763      case OP_BRA:     /* Non-capturing bracket: optimized */  
764      DPRINTF(("start bracket 0\n"));      /* Non-capturing bracket. Loop for all the alternatives. When we get to the
765      do      final alternative within the brackets, we would return the result of a
766        recursive call to match() whatever happened. We can reduce stack usage by
767        turning this into a tail recursion, except in the case when match_cbegroup
768        is set.*/
769    
770        case OP_BRA:
771        case OP_SBRA:
772        DPRINTF(("start non-capturing bracket\n"));
773        flags = (op >= OP_SBRA)? match_cbegroup : 0;
774        for (;;)
775        {        {
776        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb,        if (ecode[GET(ecode, 1)] != OP_ALT)   /* Final alternative */
777          match_isgroup);          {
778        if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (flags == 0)    /* Not a possibly empty group */
779              {
780              ecode += _pcre_OP_lengths[*ecode];
781              DPRINTF(("bracket 0 tail recursion\n"));
782              goto TAIL_RECURSE;
783              }
784    
785            /* Possibly empty group; can't use tail recursion. */
786    
787            RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
788              eptrb, flags, RM48);
789            RRETURN(rrc);
790            }
791    
792          /* For non-final alternatives, continue the loop for a NOMATCH result;
793          otherwise return. */
794    
795          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
796            eptrb, flags, RM2);
797          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
798        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
799        }        }
800      while (*ecode == OP_ALT);      /* Control never reaches here. */
     DPRINTF(("bracket 0 failed\n"));  
     RRETURN(MATCH_NOMATCH);  
801    
802      /* Conditional group: compilation checked that there are no more than      /* Conditional group: compilation checked that there are no more than
803      two branches. If the condition is false, skipping the first branch takes us      two branches. If the condition is false, skipping the first branch takes us
804      past the end if there is only one branch, but that's OK because that is      past the end if there is only one branch, but that's OK because that is
805      exactly what going to the ket would do. */      exactly what going to the ket would do. As there is only one branch to be
806        obeyed, we can use tail recursion to avoid using another stack frame. */
807    
808      case OP_COND:      case OP_COND:
809      if (ecode[LINK_SIZE+1] == OP_CREF) /* Condition extract or recurse test */      case OP_SCOND:
810        codelink= GET(ecode, 1);
811    
812        /* Because of the way auto-callout works during compile, a callout item is
813        inserted between OP_COND and an assertion condition. */
814    
815        if (ecode[LINK_SIZE+1] == OP_CALLOUT)
816          {
817          if (pcre_callout != NULL)
818            {
819            pcre_callout_block cb;
820            cb.version          = 1;   /* Version 1 of the callout block */
821            cb.callout_number   = ecode[LINK_SIZE+2];
822            cb.offset_vector    = md->offset_vector;
823            cb.subject          = (PCRE_SPTR)md->start_subject;
824            cb.subject_length   = md->end_subject - md->start_subject;
825            cb.start_match      = mstart - md->start_subject;
826            cb.current_position = eptr - md->start_subject;
827            cb.pattern_position = GET(ecode, LINK_SIZE + 3);
828            cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);
829            cb.capture_top      = offset_top/2;
830            cb.capture_last     = md->capture_last;
831            cb.callout_data     = md->callout_data;
832            if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);
833            if (rrc < 0) RRETURN(rrc);
834            }
835          ecode += _pcre_OP_lengths[OP_CALLOUT];
836          }
837    
838        condcode = ecode[LINK_SIZE+1];
839    
840        /* Now see what the actual condition is */
841    
842        if (condcode == OP_RREF)         /* Recursion test */
843          {
844          offset = GET2(ecode, LINK_SIZE + 2);     /* Recursion group number*/
845          condition = md->recursive != NULL &&
846            (offset == RREF_ANY || offset == md->recursive->group_num);
847          ecode += condition? 3 : GET(ecode, 1);
848          }
849    
850        else if (condcode == OP_CREF)    /* Group used test */
851        {        {
852        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */
853        condition = (offset == CREF_RECURSE * 2)?        condition = offset < offset_top && md->offset_vector[offset] >= 0;
854          (md->recursive != NULL) :        ecode += condition? 3 : GET(ecode, 1);
855          (offset < offset_top && md->offset_vector[offset] >= 0);        }
856        RMATCH(rrc, eptr, ecode + (condition?  
857          (LINK_SIZE + 4) : (LINK_SIZE + 1 + GET(ecode, 1))),      else if (condcode == OP_DEF)     /* DEFINE - always false */
858          offset_top, md, ims, eptrb, match_isgroup);        {
859        RRETURN(rrc);        condition = FALSE;
860          ecode += GET(ecode, 1);
861        }        }
862    
863      /* The condition is an assertion. Call match() to evaluate it - setting      /* The condition is an assertion. Call match() to evaluate it - setting
864      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
865        assertion. */
866    
867      else      else
868        {        {
869        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,
870            match_condassert | match_isgroup);            match_condassert, RM3);
871        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
872          {          {
873          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE+2);          condition = TRUE;
874            ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
875          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
876          }          }
877        else if (rrc != MATCH_NOMATCH)        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
878          {          {
879          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
880          }          }
881        else ecode += GET(ecode, 1);        else
882        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb,          {
883          match_isgroup);          condition = FALSE;
884        RRETURN(rrc);          ecode += codelink;
885            }
886        }        }
     /* Control never reaches here */  
887    
888      /* Skip over conditional reference or large extraction number data if      /* We are now at the branch that is to be obeyed. As there is only one,
889      encountered. */      we can use tail recursion to avoid using another stack frame, except when
890        match_cbegroup is required for an unlimited repeat of a possibly empty
891        group. If the second alternative doesn't exist, we can just plough on. */
892    
893      case OP_CREF:      if (condition || *ecode == OP_ALT)
894      case OP_BRANUMBER:        {
895      ecode += 3;        ecode += 1 + LINK_SIZE;
896          if (op == OP_SCOND)        /* Possibly empty group */
897            {
898            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, match_cbegroup, RM49);
899            RRETURN(rrc);
900            }
901          else                       /* Group must match something */
902            {
903            flags = 0;
904            goto TAIL_RECURSE;
905            }
906          }
907        else                         /* Condition false & no alternative */
908          {
909          ecode += 1 + LINK_SIZE;
910          }
911      break;      break;
912    
     /* End of the pattern. If we are in a recursion, we should restore the  
     offsets appropriately and continue from after the call. */  
913    
914        /* End of the pattern, either real or forced. If we are in a top-level
915        recursion, we should restore the offsets appropriately and continue from
916        after the call. */
917    
918        case OP_ACCEPT:
919      case OP_END:      case OP_END:
920      if (md->recursive != NULL && md->recursive->group_num == 0)      if (md->recursive != NULL && md->recursive->group_num == 0)
921        {        {
922        recursion_info *rec = md->recursive;        recursion_info *rec = md->recursive;
923        DPRINTF(("Hit the end in a (?0) recursion\n"));        DPRINTF(("End of pattern in a (?0) recursion\n"));
924        md->recursive = rec->prevrec;        md->recursive = rec->prevrec;
925        memmove(md->offset_vector, rec->offset_save,        memmove(md->offset_vector, rec->offset_save,
926          rec->saved_max * sizeof(int));          rec->saved_max * sizeof(int));
927        md->start_match = rec->save_start;        mstart = rec->save_start;
928        ims = original_ims;        ims = original_ims;
929        ecode = rec->after_call;        ecode = rec->after_call;
930        break;        break;
931        }        }
932    
933      /* Otherwise, if PCRE_NOTEMPTY is set, fail if we have matched an empty      /* Otherwise, if we have matched an empty string, fail if PCRE_NOTEMPTY is
934      string - backtracking will then try other alternatives, if any. */      set, or if PCRE_NOTEMPTY_ATSTART is set and we have matched at the start of
935        the subject. In both cases, backtracking will then try other alternatives,
936        if any. */
937    
938        if (eptr == mstart &&
939            (md->notempty ||
940              (md->notempty_atstart &&
941                mstart == md->start_subject + md->start_offset)))
942          RRETURN(MATCH_NOMATCH);
943    
944        /* Otherwise, we have a match. */
945    
946      if (md->notempty && eptr == md->start_match) RRETURN(MATCH_NOMATCH);      md->end_match_ptr = eptr;           /* Record where we ended */
947      md->end_match_ptr = eptr;          /* Record where we ended */      md->end_offset_top = offset_top;    /* and how many extracts were taken */
948      md->end_offset_top = offset_top;   /* and how many extracts were taken */      md->start_match_ptr = mstart;       /* and the start (\K can modify) */
949      RRETURN(MATCH_MATCH);      RRETURN(MATCH_MATCH);
950    
951      /* Change option settings */      /* Change option settings */
# Line 717  for (;;) Line 966  for (;;)
966      case OP_ASSERTBACK:      case OP_ASSERTBACK:
967      do      do
968        {        {
969        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
970          match_isgroup);          RM4);
971        if (rrc == MATCH_MATCH) break;        if (rrc == MATCH_MATCH) break;
972        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
973        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
974        }        }
975      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 744  for (;;) Line 993  for (;;)
993      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
994      do      do
995        {        {
996        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
997          match_isgroup);          RM5);
998        if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH);
999        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1000        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1001        }        }
1002      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 766  for (;;) Line 1015  for (;;)
1015  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
1016      if (utf8)      if (utf8)
1017        {        {
1018        c = GET(ecode,1);        i = GET(ecode, 1);
1019        for (i = 0; i < c; i++)        while (i-- > 0)
1020          {          {
1021          eptr--;          eptr--;
1022          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1023          BACKCHAR(eptr)          BACKCHAR(eptr);
1024          }          }
1025        }        }
1026      else      else
# Line 780  for (;;) Line 1029  for (;;)
1029      /* 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 */
1030    
1031        {        {
1032        eptr -= GET(ecode,1);        eptr -= GET(ecode, 1);
1033        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1034        }        }
1035    
1036      /* Skip to next op code */      /* Save the earliest consulted character, then skip to next op code */
1037    
1038        if (eptr < md->start_used_ptr) md->start_used_ptr = eptr;
1039      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1040      break;      break;
1041    
# Line 800  for (;;) Line 1050  for (;;)
1050        cb.version          = 1;   /* Version 1 of the callout block */        cb.version          = 1;   /* Version 1 of the callout block */
1051        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1052        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1053        cb.subject          = (const char *)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1054        cb.subject_length   = md->end_subject - md->start_subject;        cb.subject_length   = md->end_subject - md->start_subject;
1055        cb.start_match      = md->start_match - md->start_subject;        cb.start_match      = mstart - md->start_subject;
1056        cb.current_position = eptr - md->start_subject;        cb.current_position = eptr - md->start_subject;
1057        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1058        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
# Line 837  for (;;) Line 1087  for (;;)
1087      case OP_RECURSE:      case OP_RECURSE:
1088        {        {
1089        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1090        new_recursive.group_num = *callpat - OP_BRA;        new_recursive.group_num = (callpat == md->start_code)? 0 :
1091            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);  
1092    
1093        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1094    
# Line 869  for (;;) Line 1114  for (;;)
1114    
1115        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1116              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1117        new_recursive.save_start = md->start_match;        new_recursive.save_start = mstart;
1118        md->start_match = eptr;        mstart = eptr;
1119    
1120        /* 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
1121        restore the offset and recursion data. */        restore the offset and recursion data. */
1122    
1123        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1124          flags = (*callpat >= OP_SBRA)? match_cbegroup : 0;
1125        do        do
1126          {          {
1127          RMATCH(rrc, eptr, callpat + 1 + LINK_SIZE, offset_top, md, ims,          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,
1128              eptrb, match_isgroup);            md, ims, eptrb, flags, RM6);
1129          if (rrc == MATCH_MATCH)          if (rrc == MATCH_MATCH)
1130            {            {
1131              DPRINTF(("Recursion matched\n"));
1132            md->recursive = new_recursive.prevrec;            md->recursive = new_recursive.prevrec;
1133            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1134              (pcre_free)(new_recursive.offset_save);              (pcre_free)(new_recursive.offset_save);
1135            RRETURN(MATCH_MATCH);            RRETURN(MATCH_MATCH);
1136            }            }
1137          else if (rrc != MATCH_NOMATCH) RRETURN(rrc);          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1138              {
1139              DPRINTF(("Recursion gave error %d\n", rrc));
1140              if (new_recursive.offset_save != stacksave)
1141                (pcre_free)(new_recursive.offset_save);
1142              RRETURN(rrc);
1143              }
1144    
1145          md->recursive = &new_recursive;          md->recursive = &new_recursive;
1146          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
# Line 912  for (;;) Line 1165  for (;;)
1165      the end of a normal bracket, leaving the subject pointer. */      the end of a normal bracket, leaving the subject pointer. */
1166    
1167      case OP_ONCE:      case OP_ONCE:
1168        {      prev = ecode;
1169        prev = ecode;      saved_eptr = eptr;
       saved_eptr = eptr;  
1170    
1171        do      do
1172          {        {
1173          RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM7);
1174            eptrb, match_isgroup);        if (rrc == MATCH_MATCH) break;
1175          if (rrc == MATCH_MATCH) break;        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1176          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        ecode += GET(ecode,1);
1177          ecode += GET(ecode,1);        }
1178          }      while (*ecode == OP_ALT);
       while (*ecode == OP_ALT);  
1179    
1180        /* If hit the end of the group (which could be repeated), fail */      /* If hit the end of the group (which could be repeated), fail */
1181    
1182        if (*ecode != OP_ONCE && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);      if (*ecode != OP_ONCE && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);
1183    
1184        /* Continue as from after the assertion, updating the offsets high water      /* Continue as from after the assertion, updating the offsets high water
1185        mark, since extracts may have been taken. */      mark, since extracts may have been taken. */
1186    
1187        do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode, 1); while (*ecode == OP_ALT);
1188    
1189        offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1190        eptr = md->end_match_ptr;      eptr = md->end_match_ptr;
1191    
1192        /* For a non-repeating ket, just continue at this level. This also      /* For a non-repeating ket, just continue at this level. This also
1193        happens for a repeating ket if no characters were matched in the group.      happens for a repeating ket if no characters were matched in the group.
1194        This is the forcible breaking of infinite loops as implemented in Perl      This is the forcible breaking of infinite loops as implemented in Perl
1195        5.005. If there is an options reset, it will get obeyed in the normal      5.005. If there is an options reset, it will get obeyed in the normal
1196        course of events. */      course of events. */
1197    
1198        if (*ecode == OP_KET || eptr == saved_eptr)      if (*ecode == OP_KET || eptr == saved_eptr)
1199          {        {
1200          ecode += 1+LINK_SIZE;        ecode += 1+LINK_SIZE;
1201          break;        break;
1202          }        }
1203    
1204        /* 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
1205        preceding bracket, in the appropriate order. We need to reset any options      preceding bracket, in the appropriate order. The second "call" of match()
1206        that changed within the bracket before re-running it, so check the next      uses tail recursion, to avoid using another stack frame. We need to reset
1207        opcode. */      any options that changed within the bracket before re-running it, so
1208        check the next opcode. */
1209    
1210        if (ecode[1+LINK_SIZE] == OP_OPT)      if (ecode[1+LINK_SIZE] == OP_OPT)
1211          {        {
1212          ims = (ims & ~PCRE_IMS) | ecode[4];        ims = (ims & ~PCRE_IMS) | ecode[4];
1213          DPRINTF(("ims set to %02lx at group repeat\n", ims));        DPRINTF(("ims set to %02lx at group repeat\n", ims));
1214          }        }
1215    
1216        if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1217          {        {
1218          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);
1219          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1220          RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);        ecode = prev;
1221          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        flags = 0;
1222          }        goto TAIL_RECURSE;
       else  /* OP_KETRMAX */  
         {  
         RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);  
         if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
         RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0);  
         if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
         }  
1223        }        }
1224      RRETURN(MATCH_NOMATCH);      else  /* OP_KETRMAX */
1225          {
1226          RMATCH(eptr, prev, offset_top, md, ims, eptrb, match_cbegroup, RM9);
1227          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1228          ecode += 1 + LINK_SIZE;
1229          flags = 0;
1230          goto TAIL_RECURSE;
1231          }
1232        /* Control never gets here */
1233    
1234      /* An alternation is the end of a branch; scan along to find the end of the      /* An alternation is the end of a branch; scan along to find the end of the
1235      bracketed group and go to there. */      bracketed group and go to there. */
# Line 985  for (;;) Line 1238  for (;;)
1238      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1239      break;      break;
1240    
1241      /* BRAZERO and BRAMINZERO occur just before a bracket group, indicating      /* BRAZERO, BRAMINZERO and SKIPZERO occur just before a bracket group,
1242      that it may occur zero times. It may repeat infinitely, or not at all -      indicating that it may occur zero times. It may repeat infinitely, or not
1243      i.e. it could be ()* or ()? in the pattern. Brackets with fixed upper      at all - i.e. it could be ()* or ()? or even (){0} in the pattern. Brackets
1244      repeat limits are compiled as a number of copies, with the optional ones      with fixed upper repeat limits are compiled as a number of copies, with the
1245      preceded by BRAZERO or BRAMINZERO. */      optional ones preceded by BRAZERO or BRAMINZERO. */
1246    
1247      case OP_BRAZERO:      case OP_BRAZERO:
1248        {        {
1249        next = ecode+1;        next = ecode+1;
1250        RMATCH(rrc, eptr, next, offset_top, md, ims, eptrb, match_isgroup);        RMATCH(eptr, next, offset_top, md, ims, eptrb, 0, RM10);
1251        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1252        do next += GET(next,1); while (*next == OP_ALT);        do next += GET(next,1); while (*next == OP_ALT);
1253        ecode = next + 1+LINK_SIZE;        ecode = next + 1 + LINK_SIZE;
1254        }        }
1255      break;      break;
1256    
1257      case OP_BRAMINZERO:      case OP_BRAMINZERO:
1258        {        {
1259        next = ecode+1;        next = ecode+1;
1260        do next += GET(next,1); while (*next == OP_ALT);        do next += GET(next, 1); while (*next == OP_ALT);
1261        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);  
1262        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1263        ecode++;        ecode++;
1264        }        }
1265      break;      break;
1266    
1267      /* End of a group, repeated or non-repeating. If we are at the end of      case OP_SKIPZERO:
1268      an assertion "group", stop matching and return MATCH_MATCH, but record the        {
1269      current high water mark for use by positive assertions. Do this also        next = ecode+1;
1270      for the "once" (not-backup up) groups. */        do next += GET(next,1); while (*next == OP_ALT);
1271          ecode = next + 1 + LINK_SIZE;
1272          }
1273        break;
1274    
1275        /* End of a group, repeated or non-repeating. */
1276    
1277      case OP_KET:      case OP_KET:
1278      case OP_KETRMIN:      case OP_KETRMIN:
1279      case OP_KETRMAX:      case OP_KETRMAX:
1280        {      prev = ecode - GET(ecode, 1);
       prev = ecode - GET(ecode, 1);  
       saved_eptr = eptrb->epb_saved_eptr;  
1281    
1282        /* Back up the stack of bracket start pointers. */      /* If this was a group that remembered the subject start, in order to break
1283        infinite repeats of empty string matches, retrieve the subject start from
1284        the chain. Otherwise, set it NULL. */
1285    
1286        eptrb = eptrb->epb_prev;      if (*prev >= OP_SBRA)
1287          {
1288        if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||        saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */
1289            *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||        eptrb = eptrb->epb_prev;              /* Backup to previous group */
1290            *prev == OP_ONCE)        }
1291          {      else saved_eptr = NULL;
         md->end_match_ptr = eptr;      /* For ONCE */  
         md->end_offset_top = offset_top;  
         RRETURN(MATCH_MATCH);  
         }  
1292    
1293        /* In all other cases except a conditional group we have to check the      /* If we are at the end of an assertion group, stop matching and return
1294        group number back at the start and if necessary complete handling an      MATCH_MATCH, but record the current high water mark for use by positive
1295        extraction by setting the offsets and bumping the high water mark. */      assertions. Do this also for the "once" (atomic) groups. */
1296    
1297        if (*prev != OP_COND)      if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||
1298          {          *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||
1299          number = *prev - OP_BRA;          *prev == OP_ONCE)
1300          {
1301          md->end_match_ptr = eptr;      /* For ONCE */
1302          md->end_offset_top = offset_top;
1303          RRETURN(MATCH_MATCH);
1304          }
1305    
1306          /* For extended extraction brackets (large number), we have to fish out      /* For capturing groups we have to check the group number back at the start
1307          the number from a dummy opcode at the start. */      and if necessary complete handling an extraction by setting the offsets and
1308        bumping the high water mark. Note that whole-pattern recursion is coded as
1309        a recurse into group 0, so it won't be picked up here. Instead, we catch it
1310        when the OP_END is reached. Other recursion is handled here. */
1311    
1312          if (number > EXTRACT_BASIC_MAX) number = GET2(prev, 2+LINK_SIZE);      if (*prev == OP_CBRA || *prev == OP_SCBRA)
1313          offset = number << 1;        {
1314          number = GET2(prev, 1+LINK_SIZE);
1315          offset = number << 1;
1316    
1317  #ifdef DEBUG  #ifdef DEBUG
1318          printf("end bracket %d", number);        printf("end bracket %d", number);
1319          printf("\n");        printf("\n");
1320  #endif  #endif
1321    
1322          /* Test for a numbered group. This includes groups called as a result        md->capture_last = number;
1323          of recursion. Note that whole-pattern recursion is coded as a recurse        if (offset >= md->offset_max) md->offset_overflow = TRUE; else
1324          into group 0, so it won't be picked up here. Instead, we catch it when          {
1325          the OP_END is reached. */          md->offset_vector[offset] =
1326              md->offset_vector[md->offset_end - number];
1327          if (number > 0)          md->offset_vector[offset+1] = eptr - md->start_subject;
1328            {          if (offset_top <= offset) offset_top = offset + 2;
1329            md->capture_last = number;          }
1330            if (offset >= md->offset_max) md->offset_overflow = TRUE; else  
1331              {        /* Handle a recursively called group. Restore the offsets
1332              md->offset_vector[offset] =        appropriately and continue from after the call. */
1333                md->offset_vector[md->offset_end - number];  
1334              md->offset_vector[offset+1] = eptr - md->start_subject;        if (md->recursive != NULL && md->recursive->group_num == number)
1335              if (offset_top <= offset) offset_top = offset + 2;          {
1336              }          recursion_info *rec = md->recursive;
1337            DPRINTF(("Recursion (%d) succeeded - continuing\n", number));
1338            /* Handle a recursively called group. Restore the offsets          md->recursive = rec->prevrec;
1339            appropriately and continue from after the call. */          mstart = rec->save_start;
1340            memcpy(md->offset_vector, rec->offset_save,
1341            if (md->recursive != NULL && md->recursive->group_num == number)            rec->saved_max * sizeof(int));
1342              {          ecode = rec->after_call;
1343              recursion_info *rec = md->recursive;          ims = original_ims;
1344              DPRINTF(("Recursion (%d) succeeded - continuing\n", number));          break;
             md->recursive = rec->prevrec;  
             md->start_match = rec->save_start;  
             memcpy(md->offset_vector, rec->offset_save,  
               rec->saved_max * sizeof(int));  
             ecode = rec->after_call;  
             ims = original_ims;  
             break;  
             }  
           }  
1345          }          }
1346          }
1347    
1348        /* 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
1349        the group. */      flags, in case they got changed during the group. */
1350    
1351        ims = original_ims;      ims = original_ims;
1352        DPRINTF(("ims reset to %02lx\n", ims));      DPRINTF(("ims reset to %02lx\n", ims));
1353    
1354        /* For a non-repeating ket, just continue at this level. This also      /* For a non-repeating ket, just continue at this level. This also
1355        happens for a repeating ket if no characters were matched in the group.      happens for a repeating ket if no characters were matched in the group.
1356        This is the forcible breaking of infinite loops as implemented in Perl      This is the forcible breaking of infinite loops as implemented in Perl
1357        5.005. If there is an options reset, it will get obeyed in the normal      5.005. If there is an options reset, it will get obeyed in the normal
1358        course of events. */      course of events. */
1359    
1360        if (*ecode == OP_KET || eptr == saved_eptr)      if (*ecode == OP_KET || eptr == saved_eptr)
1361          {        {
1362          ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1363          break;        break;
1364          }        }
1365    
1366        /* 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
1367        preceding bracket, in the appropriate order. */      preceding bracket, in the appropriate order. In the second case, we can use
1368        tail recursion to avoid using another stack frame, unless we have an
1369        unlimited repeat of a group that can match an empty string. */
1370    
1371        if (*ecode == OP_KETRMIN)      flags = (*prev >= OP_SBRA)? match_cbegroup : 0;
1372          {  
1373          RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0);      if (*ecode == OP_KETRMIN)
1374          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        {
1375          RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM12);
1376          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1377          }        if (flags != 0)    /* Could match an empty string */
       else  /* OP_KETRMAX */  
1378          {          {
1379          RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);          RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM50);
1380          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          RRETURN(rrc);
         RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0);  
         if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
1381          }          }
1382          ecode = prev;
1383          goto TAIL_RECURSE;
1384        }        }
1385        else  /* OP_KETRMAX */
1386      RRETURN(MATCH_NOMATCH);        {
1387          RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13);
1388          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1389          ecode += 1 + LINK_SIZE;
1390          flags = 0;
1391          goto TAIL_RECURSE;
1392          }
1393        /* Control never gets here */
1394    
1395      /* Start of subject unless notbol, or after internal newline if multiline */      /* Start of subject unless notbol, or after internal newline if multiline */
1396    
# Line 1135  for (;;) Line 1398  for (;;)
1398      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
1399      if ((ims & PCRE_MULTILINE) != 0)      if ((ims & PCRE_MULTILINE) != 0)
1400        {        {
1401        if (eptr != md->start_subject && eptr[-1] != NEWLINE)        if (eptr != md->start_subject &&
1402              (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
1403          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
1404        ecode++;        ecode++;
1405        break;        break;
# Line 1156  for (;;) Line 1420  for (;;)
1420      ecode++;      ecode++;
1421      break;      break;
1422    
1423        /* Reset the start of match point */
1424    
1425        case OP_SET_SOM:
1426        mstart = eptr;
1427        ecode++;
1428        break;
1429    
1430      /* Assert before internal newline if multiline, or before a terminating      /* Assert before internal newline if multiline, or before a terminating
1431      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. */
1432    
# Line 1163  for (;;) Line 1434  for (;;)
1434      if ((ims & PCRE_MULTILINE) != 0)      if ((ims & PCRE_MULTILINE) != 0)
1435        {        {
1436        if (eptr < md->end_subject)        if (eptr < md->end_subject)
1437          { if (*eptr != NEWLINE) RRETURN(MATCH_NOMATCH); }          { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }
1438        else        else
1439          { if (md->noteol) RRETURN(MATCH_NOMATCH); }          { if (md->noteol) RRETURN(MATCH_NOMATCH); }
1440        ecode++;        ecode++;
# Line 1174  for (;;) Line 1445  for (;;)
1445        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
1446        if (!md->endonly)        if (!md->endonly)
1447          {          {
1448          if (eptr < md->end_subject - 1 ||          if (eptr != md->end_subject &&
1449             (eptr == md->end_subject - 1 && *eptr != NEWLINE))              (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
1450            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1451          ecode++;          ecode++;
1452          break;          break;
1453          }          }
1454        }        }
1455      /* ... else fall through */      /* ... else fall through for endonly */
1456    
1457      /* End of subject assertion (\z) */      /* End of subject assertion (\z) */
1458    
# Line 1193  for (;;) Line 1464  for (;;)
1464      /* End of subject or ending \n assertion (\Z) */      /* End of subject or ending \n assertion (\Z) */
1465    
1466      case OP_EODN:      case OP_EODN:
1467      if (eptr < md->end_subject - 1 ||      if (eptr != md->end_subject &&
1468         (eptr == md->end_subject - 1 && *eptr != NEWLINE)) RRETURN(MATCH_NOMATCH);          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
1469          RRETURN(MATCH_NOMATCH);
1470      ecode++;      ecode++;
1471      break;      break;
1472    
# Line 1206  for (;;) Line 1478  for (;;)
1478    
1479        /* Find out if the previous and current characters are "word" characters.        /* Find out if the previous and current characters are "word" characters.
1480        It takes a bit more work in UTF-8 mode. Characters > 255 are assumed to        It takes a bit more work in UTF-8 mode. Characters > 255 are assumed to
1481        be "non-word" characters. */        be "non-word" characters. Remember the earliest consulted character for
1482          partial matching. */
1483    
1484  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
1485        if (utf8)        if (utf8)
1486          {          {
1487          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
1488            {            {
1489            const uschar *lastptr = eptr - 1;            USPTR lastptr = eptr - 1;
1490            while((*lastptr & 0xc0) == 0x80) lastptr--;            while((*lastptr & 0xc0) == 0x80) lastptr--;
1491              if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
1492            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
1493            prev_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0;            prev_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0;
1494            }            }
1495          if (eptr >= md->end_subject) cur_is_word = FALSE; else          if (eptr >= md->end_subject)
1496              {
1497              SCHECK_PARTIAL();
1498              cur_is_word = FALSE;
1499              }
1500            else
1501            {            {
1502            GETCHAR(c, eptr);            GETCHAR(c, eptr);
1503            cur_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0;            cur_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0;
# Line 1227  for (;;) Line 1506  for (;;)
1506        else        else
1507  #endif  #endif
1508    
1509        /* More streamlined when not in UTF-8 mode */        /* Not in UTF-8 mode */
1510    
1511          {          {
1512          prev_is_word = (eptr != md->start_subject) &&          if (eptr == md->start_subject) prev_is_word = FALSE; else
1513            ((md->ctypes[eptr[-1]] & ctype_word) != 0);            {
1514          cur_is_word = (eptr < md->end_subject) &&            if (eptr <= md->start_used_ptr) md->start_used_ptr = eptr - 1;
1515            ((md->ctypes[*eptr] & ctype_word) != 0);            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);
1516              }
1517            if (eptr >= md->end_subject)
1518              {
1519              SCHECK_PARTIAL();
1520              cur_is_word = FALSE;
1521              }
1522            else cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);
1523          }          }
1524    
1525        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
# Line 1247  for (;;) Line 1533  for (;;)
1533      /* Match a single character type; inline for speed */      /* Match a single character type; inline for speed */
1534    
1535      case OP_ANY:      case OP_ANY:
1536      if ((ims & PCRE_DOTALL) == 0 && eptr < md->end_subject && *eptr == NEWLINE)      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
1537        /* Fall through */
1538    
1539        case OP_ALLANY:
1540        if (eptr++ >= md->end_subject)
1541          {
1542          SCHECK_PARTIAL();
1543        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1544      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);        }
1545  #ifdef SUPPORT_UTF8      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
     if (utf8)  
       while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  
 #endif  
1546      ecode++;      ecode++;
1547      break;      break;
1548    
# Line 1261  for (;;) Line 1550  for (;;)
1550      any byte, even newline, independent of the setting of PCRE_DOTALL. */      any byte, even newline, independent of the setting of PCRE_DOTALL. */
1551    
1552      case OP_ANYBYTE:      case OP_ANYBYTE:
1553      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr++ >= md->end_subject)
1554          {
1555          SCHECK_PARTIAL();
1556          RRETURN(MATCH_NOMATCH);
1557          }
1558      ecode++;      ecode++;
1559      break;      break;
1560    
1561      case OP_NOT_DIGIT:      case OP_NOT_DIGIT:
1562      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
1563          {
1564          SCHECK_PARTIAL();
1565          RRETURN(MATCH_NOMATCH);
1566          }
1567      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1568      if (      if (
1569  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1279  for (;;) Line 1576  for (;;)
1576      break;      break;
1577    
1578      case OP_DIGIT:      case OP_DIGIT:
1579      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
1580          {
1581          SCHECK_PARTIAL();
1582          RRETURN(MATCH_NOMATCH);
1583          }
1584      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1585      if (      if (
1586  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1292  for (;;) Line 1593  for (;;)
1593      break;      break;
1594    
1595      case OP_NOT_WHITESPACE:      case OP_NOT_WHITESPACE:
1596      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
1597          {
1598          SCHECK_PARTIAL();
1599          RRETURN(MATCH_NOMATCH);
1600          }
1601      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1602      if (      if (
1603  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1305  for (;;) Line 1610  for (;;)
1610      break;      break;
1611    
1612      case OP_WHITESPACE:      case OP_WHITESPACE:
1613      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
1614          {
1615          SCHECK_PARTIAL();
1616          RRETURN(MATCH_NOMATCH);
1617          }
1618      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1619      if (      if (
1620  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1318  for (;;) Line 1627  for (;;)
1627      break;      break;
1628    
1629      case OP_NOT_WORDCHAR:      case OP_NOT_WORDCHAR:
1630      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
1631          {
1632          SCHECK_PARTIAL();
1633          RRETURN(MATCH_NOMATCH);
1634          }
1635      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1636      if (      if (
1637  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1331  for (;;) Line 1644  for (;;)
1644      break;      break;
1645    
1646      case OP_WORDCHAR:      case OP_WORDCHAR:
1647      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
1648          {
1649          SCHECK_PARTIAL();
1650          RRETURN(MATCH_NOMATCH);
1651          }
1652      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1653      if (      if (
1654  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1343  for (;;) Line 1660  for (;;)
1660      ecode++;      ecode++;
1661      break;      break;
1662    
1663  #ifdef SUPPORT_UCP      case OP_ANYNL:
1664      /* Check the next character by Unicode property. We will get here only      if (eptr >= md->end_subject)
1665      if the support is in the binary; otherwise a compile-time error occurs. */        {
1666          SCHECK_PARTIAL();
1667      case OP_PROP:        RRETURN(MATCH_NOMATCH);
1668      case OP_NOTPROP:        }
     if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);  
1669      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1670        switch(c)
1671        {        {
1672        int chartype, rqdtype;        default: RRETURN(MATCH_NOMATCH);
1673        int othercase;        case 0x000d:
1674        int category = ucp_findchar(c, &chartype, &othercase);        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
1675          break;
1676    
1677        rqdtype = *(++ecode);        case 0x000a:
1678        ecode++;        break;
1679    
1680        if (rqdtype >= 128)        case 0x000b:
1681          {        case 0x000c:
1682          if ((rqdtype - 128 != category) == (op == OP_PROP))        case 0x0085:
1683            RRETURN(MATCH_NOMATCH);        case 0x2028:
1684          }        case 0x2029:
1685        else        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
1686          break;
1687          }
1688        ecode++;
1689        break;
1690    
1691        case OP_NOT_HSPACE:
1692        if (eptr >= md->end_subject)
1693          {
1694          SCHECK_PARTIAL();
1695          RRETURN(MATCH_NOMATCH);
1696          }
1697        GETCHARINCTEST(c, eptr);
1698        switch(c)
1699          {
1700          default: break;
1701          case 0x09:      /* HT */
1702          case 0x20:      /* SPACE */
1703          case 0xa0:      /* NBSP */
1704          case 0x1680:    /* OGHAM SPACE MARK */
1705          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1706          case 0x2000:    /* EN QUAD */
1707          case 0x2001:    /* EM QUAD */
1708          case 0x2002:    /* EN SPACE */
1709          case 0x2003:    /* EM SPACE */
1710          case 0x2004:    /* THREE-PER-EM SPACE */
1711          case 0x2005:    /* FOUR-PER-EM SPACE */
1712          case 0x2006:    /* SIX-PER-EM SPACE */
1713          case 0x2007:    /* FIGURE SPACE */
1714          case 0x2008:    /* PUNCTUATION SPACE */
1715          case 0x2009:    /* THIN SPACE */
1716          case 0x200A:    /* HAIR SPACE */
1717          case 0x202f:    /* NARROW NO-BREAK SPACE */
1718          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1719          case 0x3000:    /* IDEOGRAPHIC SPACE */
1720          RRETURN(MATCH_NOMATCH);
1721          }
1722        ecode++;
1723        break;
1724    
1725        case OP_HSPACE:
1726        if (eptr >= md->end_subject)
1727          {
1728          SCHECK_PARTIAL();
1729          RRETURN(MATCH_NOMATCH);
1730          }
1731        GETCHARINCTEST(c, eptr);
1732        switch(c)
1733          {
1734          default: RRETURN(MATCH_NOMATCH);
1735          case 0x09:      /* HT */
1736          case 0x20:      /* SPACE */
1737          case 0xa0:      /* NBSP */
1738          case 0x1680:    /* OGHAM SPACE MARK */
1739          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1740          case 0x2000:    /* EN QUAD */
1741          case 0x2001:    /* EM QUAD */
1742          case 0x2002:    /* EN SPACE */
1743          case 0x2003:    /* EM SPACE */
1744          case 0x2004:    /* THREE-PER-EM SPACE */
1745          case 0x2005:    /* FOUR-PER-EM SPACE */
1746          case 0x2006:    /* SIX-PER-EM SPACE */
1747          case 0x2007:    /* FIGURE SPACE */
1748          case 0x2008:    /* PUNCTUATION SPACE */
1749          case 0x2009:    /* THIN SPACE */
1750          case 0x200A:    /* HAIR SPACE */
1751          case 0x202f:    /* NARROW NO-BREAK SPACE */
1752          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1753          case 0x3000:    /* IDEOGRAPHIC SPACE */
1754          break;
1755          }
1756        ecode++;
1757        break;
1758    
1759        case OP_NOT_VSPACE:
1760        if (eptr >= md->end_subject)
1761          {
1762          SCHECK_PARTIAL();
1763          RRETURN(MATCH_NOMATCH);
1764          }
1765        GETCHARINCTEST(c, eptr);
1766        switch(c)
1767          {
1768          default: break;
1769          case 0x0a:      /* LF */
1770          case 0x0b:      /* VT */
1771          case 0x0c:      /* FF */
1772          case 0x0d:      /* CR */
1773          case 0x85:      /* NEL */
1774          case 0x2028:    /* LINE SEPARATOR */
1775          case 0x2029:    /* PARAGRAPH SEPARATOR */
1776          RRETURN(MATCH_NOMATCH);
1777          }
1778        ecode++;
1779        break;
1780    
1781        case OP_VSPACE:
1782        if (eptr >= md->end_subject)
1783          {
1784          SCHECK_PARTIAL();
1785          RRETURN(MATCH_NOMATCH);
1786          }
1787        GETCHARINCTEST(c, eptr);
1788        switch(c)
1789          {
1790          default: RRETURN(MATCH_NOMATCH);
1791          case 0x0a:      /* LF */
1792          case 0x0b:      /* VT */
1793          case 0x0c:      /* FF */
1794          case 0x0d:      /* CR */
1795          case 0x85:      /* NEL */
1796          case 0x2028:    /* LINE SEPARATOR */
1797          case 0x2029:    /* PARAGRAPH SEPARATOR */
1798          break;
1799          }
1800        ecode++;
1801        break;
1802    
1803    #ifdef SUPPORT_UCP
1804        /* Check the next character by Unicode property. We will get here only
1805        if the support is in the binary; otherwise a compile-time error occurs. */
1806    
1807        case OP_PROP:
1808        case OP_NOTPROP:
1809        if (eptr >= md->end_subject)
1810          {
1811          SCHECK_PARTIAL();
1812          RRETURN(MATCH_NOMATCH);
1813          }
1814        GETCHARINCTEST(c, eptr);
1815          {
1816          const ucd_record *prop = GET_UCD(c);
1817    
1818          switch(ecode[1])
1819          {          {
1820          if ((rqdtype != chartype) == (op == OP_PROP))          case PT_ANY:
1821            if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
1822            break;
1823    
1824            case PT_LAMP:
1825            if ((prop->chartype == ucp_Lu ||
1826                 prop->chartype == ucp_Ll ||
1827                 prop->chartype == ucp_Lt) == (op == OP_NOTPROP))
1828              RRETURN(MATCH_NOMATCH);
1829             break;
1830    
1831            case PT_GC:
1832            if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))
1833            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1834            break;
1835    
1836            case PT_PC:
1837            if ((ecode[2] != prop->chartype) == (op == OP_PROP))
1838              RRETURN(MATCH_NOMATCH);
1839            break;
1840    
1841            case PT_SC:
1842            if ((ecode[2] != prop->script) == (op == OP_PROP))
1843              RRETURN(MATCH_NOMATCH);
1844            break;
1845    
1846            default:
1847            RRETURN(PCRE_ERROR_INTERNAL);
1848          }          }
1849    
1850          ecode += 3;
1851        }        }
1852      break;      break;
1853    
# Line 1376  for (;;) Line 1855  for (;;)
1855      is in the binary; otherwise a compile-time error occurs. */      is in the binary; otherwise a compile-time error occurs. */
1856    
1857      case OP_EXTUNI:      case OP_EXTUNI:
1858      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
1859          {
1860          SCHECK_PARTIAL();
1861          RRETURN(MATCH_NOMATCH);
1862          }
1863      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1864        {        {
1865        int chartype;        int category = UCD_CATEGORY(c);
       int othercase;  
       int category = ucp_findchar(c, &chartype, &othercase);  
1866        if (category == ucp_M) RRETURN(MATCH_NOMATCH);        if (category == ucp_M) RRETURN(MATCH_NOMATCH);
1867        while (eptr < md->end_subject)        while (eptr < md->end_subject)
1868          {          {
# Line 1390  for (;;) Line 1871  for (;;)
1871            {            {
1872            GETCHARLEN(c, eptr, len);            GETCHARLEN(c, eptr, len);
1873            }            }
1874          category = ucp_findchar(c, &chartype, &othercase);          category = UCD_CATEGORY(c);
1875          if (category != ucp_M) break;          if (category != ucp_M) break;
1876          eptr += len;          eptr += len;
1877          }          }
# Line 1411  for (;;) Line 1892  for (;;)
1892      case OP_REF:      case OP_REF:
1893        {        {
1894        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
1895        ecode += 3;                                 /* Advance past item */        ecode += 3;
1896    
1897          /* If the reference is unset, there are two possibilities:
1898    
1899          (a) In the default, Perl-compatible state, set the length to be longer
1900          than the amount of subject left; this ensures that every attempt at a
1901          match fails. We can't just fail here, because of the possibility of
1902          quantifiers with zero minima.
1903    
1904        /* If the reference is unset, set the length to be longer than the amount        (b) If the JavaScript compatibility flag is set, set the length to zero
1905        of subject left; this ensures that every attempt at a match fails. We        so that the back reference matches an empty string.
1906        can't just fail here, because of the possibility of quantifiers with zero  
1907        minima. */        Otherwise, set the length to the length of what was matched by the
1908          referenced subpattern. */
1909        length = (offset >= offset_top || md->offset_vector[offset] < 0)?  
1910          md->end_subject - eptr + 1 :        if (offset >= offset_top || md->offset_vector[offset] < 0)
1911          md->offset_vector[offset+1] - md->offset_vector[offset];          length = (md->jscript_compat)? 0 : md->end_subject - eptr + 1;
1912          else
1913            length = md->offset_vector[offset+1] - md->offset_vector[offset];
1914    
1915        /* Set up for repetition, or handle the non-repeated case */        /* Set up for repetition, or handle the non-repeated case */
1916    
# Line 1449  for (;;) Line 1939  for (;;)
1939          break;          break;
1940    
1941          default:               /* No repeat follows */          default:               /* No repeat follows */
1942          if (!match_ref(offset, eptr, length, md, ims)) RRETURN(MATCH_NOMATCH);          if (!match_ref(offset, eptr, length, md, ims))
1943              {
1944              CHECK_PARTIAL();
1945              RRETURN(MATCH_NOMATCH);
1946              }
1947          eptr += length;          eptr += length;
1948          continue;              /* With the main loop */          continue;              /* With the main loop */
1949          }          }
# Line 1465  for (;;) Line 1959  for (;;)
1959    
1960        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
1961          {          {
1962          if (!match_ref(offset, eptr, length, md, ims)) RRETURN(MATCH_NOMATCH);          if (!match_ref(offset, eptr, length, md, ims))
1963              {
1964              CHECK_PARTIAL();
1965              RRETURN(MATCH_NOMATCH);
1966              }
1967          eptr += length;          eptr += length;
1968          }          }
1969    
# Line 1480  for (;;) Line 1978  for (;;)
1978          {          {
1979          for (fi = min;; fi++)          for (fi = min;; fi++)
1980            {            {
1981            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM14);
1982            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1983            if (fi >= max || !match_ref(offset, eptr, length, md, ims))            if (fi >= max) RRETURN(MATCH_NOMATCH);
1984              if (!match_ref(offset, eptr, length, md, ims))
1985                {
1986                CHECK_PARTIAL();
1987              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
1988                }
1989            eptr += length;            eptr += length;
1990            }            }
1991          /* Control never gets here */          /* Control never gets here */
# Line 1501  for (;;) Line 2003  for (;;)
2003            }            }
2004          while (eptr >= pp)          while (eptr >= pp)
2005            {            {
2006            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM15);
2007            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2008            eptr -= length;            eptr -= length;
2009            }            }
# Line 1510  for (;;) Line 2012  for (;;)
2012        }        }
2013      /* Control never gets here */      /* Control never gets here */
2014    
   
   
2015      /* Match a bit-mapped character class, possibly repeatedly. This op code is      /* Match a bit-mapped character class, possibly repeatedly. This op code is
2016      used when all the characters in the class have values in the range 0-255,      used when all the characters in the class have values in the range 0-255,
2017      and either the matching is caseful, or the characters are in the range      and either the matching is caseful, or the characters are in the range
# Line 1566  for (;;) Line 2066  for (;;)
2066          {          {
2067          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2068            {            {
2069            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (eptr >= md->end_subject)
2070                {
2071                SCHECK_PARTIAL();
2072                RRETURN(MATCH_NOMATCH);
2073                }
2074            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
2075            if (c > 255)            if (c > 255)
2076              {              {
# Line 1584  for (;;) Line 2088  for (;;)
2088          {          {
2089          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2090            {            {
2091            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (eptr >= md->end_subject)
2092                {
2093                SCHECK_PARTIAL();
2094                RRETURN(MATCH_NOMATCH);
2095                }
2096            c = *eptr++;            c = *eptr++;
2097            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2098            }            }
# Line 1606  for (;;) Line 2114  for (;;)
2114            {            {
2115            for (fi = min;; fi++)            for (fi = min;; fi++)
2116              {              {
2117              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM16);
2118              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2119              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2120                if (eptr >= md->end_subject)
2121                  {
2122                  SCHECK_PARTIAL();
2123                  RRETURN(MATCH_NOMATCH);
2124                  }
2125              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
2126              if (c > 255)              if (c > 255)
2127                {                {
# Line 1626  for (;;) Line 2139  for (;;)
2139            {            {
2140            for (fi = min;; fi++)            for (fi = min;; fi++)
2141              {              {
2142              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM17);
2143              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2144              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2145                if (eptr >= md->end_subject)
2146                  {
2147                  SCHECK_PARTIAL();
2148                  RRETURN(MATCH_NOMATCH);
2149                  }
2150              c = *eptr++;              c = *eptr++;
2151              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2152              }              }
# Line 1663  for (;;) Line 2181  for (;;)
2181              }              }
2182            for (;;)            for (;;)
2183              {              {
2184              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM18);
2185              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2186              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2187              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 1682  for (;;) Line 2200  for (;;)
2200              }              }
2201            while (eptr >= pp)            while (eptr >= pp)
2202              {              {
2203              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM19);
             eptr--;  
2204              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2205                eptr--;
2206              }              }
2207            }            }
2208    
# Line 1695  for (;;) Line 2213  for (;;)
2213    
2214    
2215      /* Match an extended character class. This opcode is encountered only      /* Match an extended character class. This opcode is encountered only
2216      in UTF-8 mode, because that's the only time it is compiled. */      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8
2217        mode, because Unicode properties are supported in non-UTF-8 mode. */
2218    
2219  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2220      case OP_XCLASS:      case OP_XCLASS:
# Line 1736  for (;;) Line 2255  for (;;)
2255    
2256        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
2257          {          {
2258          if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);          if (eptr >= md->end_subject)
2259          GETCHARINC(c, eptr);            {
2260              SCHECK_PARTIAL();
2261              RRETURN(MATCH_NOMATCH);
2262              }
2263            GETCHARINCTEST(c, eptr);
2264          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);
2265          }          }
2266    
# Line 1753  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, RM20);
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) RRETURN(MATCH_NOMATCH);
2282            GETCHARINC(c, eptr);            if (eptr >= md->end_subject)
2283                {
2284                SCHECK_PARTIAL();
2285                RRETURN(MATCH_NOMATCH);
2286                }
2287              GETCHARINCTEST(c, eptr);
2288            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);
2289            }            }
2290          /* Control never gets here */          /* Control never gets here */
# Line 1771  for (;;) Line 2299  for (;;)
2299            {            {
2300            int len = 1;            int len = 1;
2301            if (eptr >= md->end_subject) break;            if (eptr >= md->end_subject) break;
2302            GETCHARLEN(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
2303            if (!_pcre_xclass(c, data)) break;            if (!_pcre_xclass(c, data)) break;
2304            eptr += len;            eptr += len;
2305            }            }
2306          for(;;)          for(;;)
2307            {            {
2308            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);
2309            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2310            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
2311            BACKCHAR(eptr)            if (utf8) BACKCHAR(eptr);
2312            }            }
2313          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2314          }          }
# Line 1798  for (;;) Line 2326  for (;;)
2326        length = 1;        length = 1;
2327        ecode++;        ecode++;
2328        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
2329        if (length > md->end_subject - eptr) RRETURN(MATCH_NOMATCH);        if (length > md->end_subject - eptr)
2330            {
2331            CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
2332            RRETURN(MATCH_NOMATCH);
2333            }
2334        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);
2335        }        }
2336      else      else
# Line 1806  for (;;) Line 2338  for (;;)
2338    
2339      /* Non-UTF-8 mode */      /* Non-UTF-8 mode */
2340        {        {
2341        if (md->end_subject - eptr < 1) RRETURN(MATCH_NOMATCH);        if (md->end_subject - eptr < 1)
2342            {
2343            SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
2344            RRETURN(MATCH_NOMATCH);
2345            }
2346        if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH);        if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH);
2347        ecode += 2;        ecode += 2;
2348        }        }
# Line 1822  for (;;) Line 2358  for (;;)
2358        ecode++;        ecode++;
2359        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
2360    
2361        if (length > md->end_subject - eptr) RRETURN(MATCH_NOMATCH);        if (length > md->end_subject - eptr)
2362            {
2363            CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
2364            RRETURN(MATCH_NOMATCH);
2365            }
2366    
2367        /* If the pattern character's value is < 128, we have only one byte, and        /* If the pattern character's value is < 128, we have only one byte, and
2368        can use the fast lookup table. */        can use the fast lookup table. */
# Line 1836  for (;;) Line 2376  for (;;)
2376    
2377        else        else
2378          {          {
2379          int dc;          unsigned int dc;
2380          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
2381          ecode += length;          ecode += length;
2382    
2383          /* If we have Unicode property support, we can use it to test the other          /* If we have Unicode property support, we can use it to test the other
2384          case of the character, if there is one. The result of ucp_findchar() is          case of the character, if there is one. */
         < 0 if the char isn't found, and othercase is returned as zero if there  
         isn't one. */  
2385    
2386          if (fc != dc)          if (fc != dc)
2387            {            {
2388  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2389            int chartype;            if (dc != UCD_OTHERCASE(fc))
           int othercase;  
           if (ucp_findchar(fc, &chartype, &othercase) < 0 || dc != othercase)  
2390  #endif  #endif
2391              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2392            }            }
# Line 1861  for (;;) Line 2397  for (;;)
2397    
2398      /* Non-UTF-8 mode */      /* Non-UTF-8 mode */
2399        {        {
2400        if (md->end_subject - eptr < 1) RRETURN(MATCH_NOMATCH);        if (md->end_subject - eptr < 1)
2401            {
2402            SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
2403            RRETURN(MATCH_NOMATCH);
2404            }
2405        if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);        if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);
2406        ecode += 2;        ecode += 2;
2407        }        }
2408      break;      break;
2409    
2410      /* Match a single character repeatedly; different opcodes share code. */      /* Match a single character repeatedly. */
2411    
2412      case OP_EXACT:      case OP_EXACT:
2413      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
2414      ecode += 3;      ecode += 3;
2415      goto REPEATCHAR;      goto REPEATCHAR;
2416    
2417        case OP_POSUPTO:
2418        possessive = TRUE;
2419        /* Fall through */
2420    
2421      case OP_UPTO:      case OP_UPTO:
2422      case OP_MINUPTO:      case OP_MINUPTO:
2423      min = 0;      min = 0;
# Line 1882  for (;;) Line 2426  for (;;)
2426      ecode += 3;      ecode += 3;
2427      goto REPEATCHAR;      goto REPEATCHAR;
2428    
2429        case OP_POSSTAR:
2430        possessive = TRUE;
2431        min = 0;
2432        max = INT_MAX;
2433        ecode++;
2434        goto REPEATCHAR;
2435    
2436        case OP_POSPLUS:
2437        possessive = TRUE;
2438        min = 1;
2439        max = INT_MAX;
2440        ecode++;
2441        goto REPEATCHAR;
2442    
2443        case OP_POSQUERY:
2444        possessive = TRUE;
2445        min = 0;
2446        max = 1;
2447        ecode++;
2448        goto REPEATCHAR;
2449    
2450      case OP_STAR:      case OP_STAR:
2451      case OP_MINSTAR:      case OP_MINSTAR:
2452      case OP_PLUS:      case OP_PLUS:
# Line 1890  for (;;) Line 2455  for (;;)
2455      case OP_MINQUERY:      case OP_MINQUERY:
2456      c = *ecode++ - OP_STAR;      c = *ecode++ - OP_STAR;
2457      minimize = (c & 1) != 0;      minimize = (c & 1) != 0;
2458    
2459      min = rep_min[c];                 /* Pick up values from tables; */      min = rep_min[c];                 /* Pick up values from tables; */
2460      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
2461      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
2462    
2463      /* Common code for all repeated single-character matches. We can give      /* Common code for all repeated single-character matches. */
     up quickly if there are fewer than the minimum number of characters left in  
     the subject. */  
2464    
2465      REPEATCHAR:      REPEATCHAR:
2466  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1905  for (;;) Line 2469  for (;;)
2469        length = 1;        length = 1;
2470        charptr = ecode;        charptr = ecode;
2471        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
       if (min * length > md->end_subject - eptr) RRETURN(MATCH_NOMATCH);  
2472        ecode += length;        ecode += length;
2473    
2474        /* Handle multibyte character matching specially here. There is        /* Handle multibyte character matching specially here. There is
# Line 1913  for (;;) Line 2476  for (;;)
2476    
2477        if (length > 1)        if (length > 1)
2478          {          {
         int oclength = 0;  
         uschar occhars[8];  
   
2479  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2480          int othercase;          unsigned int othercase;
         int chartype;  
2481          if ((ims & PCRE_CASELESS) != 0 &&          if ((ims & PCRE_CASELESS) != 0 &&
2482               ucp_findchar(fc, &chartype, &othercase) >= 0 &&              (othercase = UCD_OTHERCASE(fc)) != fc)
              othercase > 0)  
2483            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = _pcre_ord2utf8(othercase, occhars);
2484            else oclength = 0;
2485  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2486    
2487          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2488            {            {
2489            if (memcmp(eptr, charptr, length) == 0) eptr += length;            if (eptr <= md->end_subject - length &&
2490            /* Need braces because of following else */              memcmp(eptr, charptr, length) == 0) eptr += length;
2491            else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }  #ifdef SUPPORT_UCP
2492              else if (oclength > 0 &&
2493                       eptr <= md->end_subject - oclength &&
2494                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;
2495    #endif  /* SUPPORT_UCP */
2496            else            else
2497              {              {
2498              if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);              CHECK_PARTIAL();
2499              eptr += oclength;              RRETURN(MATCH_NOMATCH);
2500              }              }
2501            }            }
2502    
# Line 1943  for (;;) Line 2506  for (;;)
2506            {            {
2507            for (fi = min;; fi++)            for (fi = min;; fi++)
2508              {              {
2509              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM22);
2510              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2511              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2512              if (memcmp(eptr, charptr, length) == 0) eptr += length;              if (eptr <= md->end_subject - length &&
2513              /* Need braces because of following else */                memcmp(eptr, charptr, length) == 0) eptr += length;
2514              else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }  #ifdef SUPPORT_UCP
2515                else if (oclength > 0 &&
2516                         eptr <= md->end_subject - oclength &&
2517                         memcmp(eptr, occhars, oclength) == 0) eptr += oclength;
2518    #endif  /* SUPPORT_UCP */
2519              else              else
2520                {                {
2521                if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);                CHECK_PARTIAL();
2522                eptr += oclength;                RRETURN(MATCH_NOMATCH);
2523                }                }
2524              }              }
2525            /* Control never gets here */            /* Control never gets here */
2526            }            }
2527          else  
2528            else  /* Maximize */
2529            {            {
2530            pp = eptr;            pp = eptr;
2531            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2532              {              {
2533              if (eptr > md->end_subject - length) break;              if (eptr <= md->end_subject - length &&
2534              if (memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, length) == 0) eptr += length;
2535              else if (oclength == 0) break;  #ifdef SUPPORT_UCP
2536              else              else if (oclength > 0 &&
2537                {                       eptr <= md->end_subject - oclength &&
2538                if (memcmp(eptr, occhars, oclength) != 0) break;                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;
2539                eptr += oclength;  #endif  /* SUPPORT_UCP */
2540                }              else break;
2541                }
2542    
2543              if (possessive) continue;
2544    
2545              for(;;)
2546                {
2547                RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM23);
2548                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2549                if (eptr == pp) { RRETURN(MATCH_NOMATCH); }
2550    #ifdef SUPPORT_UCP
2551                eptr--;
2552                BACKCHAR(eptr);
2553    #else   /* without SUPPORT_UCP */
2554                eptr -= length;
2555    #endif  /* SUPPORT_UCP */
2556              }              }
           while (eptr >= pp)  
            {  
            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);  
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
            eptr -= length;  
            }  
           RRETURN(MATCH_NOMATCH);  
2557            }            }
2558          /* Control never gets here */          /* Control never gets here */
2559          }          }
# Line 1990  for (;;) Line 2566  for (;;)
2566  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF8 */
2567    
2568      /* When not in UTF-8 mode, load a single-byte character. */      /* When not in UTF-8 mode, load a single-byte character. */
2569        {  
2570        if (min > md->end_subject - eptr) RRETURN(MATCH_NOMATCH);      fc = *ecode++;
       fc = *ecode++;  
       }  
2571    
2572      /* The value of fc at this point is always less than 256, though we may or      /* The value of fc at this point is always less than 256, though we may or
2573      may not be in UTF-8 mode. The code is duplicated for the caseless and      may not be in UTF-8 mode. The code is duplicated for the caseless and
# Line 2011  for (;;) Line 2585  for (;;)
2585        {        {
2586        fc = md->lcc[fc];        fc = md->lcc[fc];
2587        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
2588            {
2589            if (eptr >= md->end_subject)
2590              {
2591              SCHECK_PARTIAL();
2592              RRETURN(MATCH_NOMATCH);
2593              }
2594          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);
2595            }
2596        if (min == max) continue;        if (min == max) continue;
2597        if (minimize)        if (minimize)
2598          {          {
2599          for (fi = min;; fi++)          for (fi = min;; fi++)
2600            {            {
2601            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM24);
2602            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2603            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max) RRETURN(MATCH_NOMATCH);
2604                fc != md->lcc[*eptr++])            if (eptr >= md->end_subject)
2605                {
2606                SCHECK_PARTIAL();
2607              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2608                }
2609              if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);
2610            }            }
2611          /* Control never gets here */          /* Control never gets here */
2612          }          }
2613        else        else  /* Maximize */
2614          {          {
2615          pp = eptr;          pp = eptr;
2616          for (i = min; i < max; i++)          for (i = min; i < max; i++)
# Line 2033  for (;;) Line 2618  for (;;)
2618            if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break;            if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break;
2619            eptr++;            eptr++;
2620            }            }
2621    
2622            if (possessive) continue;
2623    
2624          while (eptr >= pp)          while (eptr >= pp)
2625            {            {
2626            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM25);
2627            eptr--;            eptr--;
2628            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2629            }            }
# Line 2048  for (;;) Line 2636  for (;;)
2636    
2637      else      else
2638        {        {
2639        for (i = 1; i <= min; i++) if (fc != *eptr++) RRETURN(MATCH_NOMATCH);        for (i = 1; i <= min; i++)
2640            {
2641            if (eptr >= md->end_subject)
2642              {
2643              SCHECK_PARTIAL();
2644              RRETURN(MATCH_NOMATCH);
2645              }
2646            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
2647            }
2648    
2649        if (min == max) continue;        if (min == max) continue;
2650    
2651        if (minimize)        if (minimize)
2652          {          {
2653          for (fi = min;; fi++)          for (fi = min;; fi++)
2654            {            {
2655            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM26);
2656            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2657            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)            if (fi >= max) RRETURN(MATCH_NOMATCH);
2658              if (eptr >= md->end_subject)
2659                {
2660                SCHECK_PARTIAL();
2661              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2662                }
2663              if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
2664            }            }
2665          /* Control never gets here */          /* Control never gets here */
2666          }          }
2667        else        else  /* Maximize */
2668          {          {
2669          pp = eptr;          pp = eptr;
2670          for (i = min; i < max; i++)          for (i = min; i < max; i++)
# Line 2069  for (;;) Line 2672  for (;;)
2672            if (eptr >= md->end_subject || fc != *eptr) break;            if (eptr >= md->end_subject || fc != *eptr) break;
2673            eptr++;            eptr++;
2674            }            }
2675            if (possessive) continue;
2676    
2677          while (eptr >= pp)          while (eptr >= pp)
2678            {            {
2679            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27);
2680            eptr--;            eptr--;
2681            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2682            }            }
# Line 2084  for (;;) Line 2689  for (;;)
2689      checking can be multibyte. */      checking can be multibyte. */
2690    
2691      case OP_NOT:      case OP_NOT:
2692      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
2693          {
2694          SCHECK_PARTIAL();
2695          RRETURN(MATCH_NOMATCH);
2696          }
2697      ecode++;      ecode++;
2698      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2699      if ((ims & PCRE_CASELESS) != 0)      if ((ims & PCRE_CASELESS) != 0)
# Line 2121  for (;;) Line 2730  for (;;)
2730      ecode += 3;      ecode += 3;
2731      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
2732    
2733        case OP_NOTPOSSTAR:
2734        possessive = TRUE;
2735        min = 0;
2736        max = INT_MAX;
2737        ecode++;
2738        goto REPEATNOTCHAR;
2739    
2740        case OP_NOTPOSPLUS:
2741        possessive = TRUE;
2742        min = 1;
2743        max = INT_MAX;
2744        ecode++;
2745        goto REPEATNOTCHAR;
2746    
2747        case OP_NOTPOSQUERY:
2748        possessive = TRUE;
2749        min = 0;
2750        max = 1;
2751        ecode++;
2752        goto REPEATNOTCHAR;
2753    
2754        case OP_NOTPOSUPTO:
2755        possessive = TRUE;
2756        min = 0;
2757        max = GET2(ecode, 1);
2758        ecode += 3;
2759        goto REPEATNOTCHAR;
2760    
2761      case OP_NOTSTAR:      case OP_NOTSTAR:
2762      case OP_NOTMINSTAR:      case OP_NOTMINSTAR:
2763      case OP_NOTPLUS:      case OP_NOTPLUS:
# Line 2133  for (;;) Line 2770  for (;;)
2770      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
2771      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
2772    
2773      /* Common code for all repeated single-byte matches. We can give up quickly      /* Common code for all repeated single-byte matches. */
     if there are fewer than the minimum number of bytes left in the  
     subject. */  
2774    
2775      REPEATNOTCHAR:      REPEATNOTCHAR:
     if (min > md->end_subject - eptr) RRETURN(MATCH_NOMATCH);  
2776      fc = *ecode++;      fc = *ecode++;
2777    
2778      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
# Line 2160  for (;;) Line 2794  for (;;)
2794        /* UTF-8 mode */        /* UTF-8 mode */
2795        if (utf8)        if (utf8)
2796          {          {
2797          register int d;          register unsigned int d;
2798          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2799            {            {
2800              if (eptr >= md->end_subject)
2801                {
2802                SCHECK_PARTIAL();
2803                RRETURN(MATCH_NOMATCH);
2804                }
2805            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
2806            if (d < 256) d = md->lcc[d];            if (d < 256) d = md->lcc[d];
2807            if (fc == d) RRETURN(MATCH_NOMATCH);            if (fc == d) RRETURN(MATCH_NOMATCH);
# Line 2174  for (;;) Line 2813  for (;;)
2813        /* Not UTF-8 mode */        /* Not UTF-8 mode */
2814          {          {
2815          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2816              {
2817              if (eptr >= md->end_subject)
2818                {
2819                SCHECK_PARTIAL();
2820                RRETURN(MATCH_NOMATCH);
2821                }
2822            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);
2823              }
2824          }          }
2825    
2826        if (min == max) continue;        if (min == max) continue;
# Line 2185  for (;;) Line 2831  for (;;)
2831          /* UTF-8 mode */          /* UTF-8 mode */
2832          if (utf8)          if (utf8)
2833            {            {
2834            register int d;            register unsigned int d;
2835            for (fi = min;; fi++)            for (fi = min;; fi++)
2836              {              {
2837              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28);
2838              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2839                if (fi >= max) RRETURN(MATCH_NOMATCH);
2840                if (eptr >= md->end_subject)
2841                  {
2842                  SCHECK_PARTIAL();
2843                  RRETURN(MATCH_NOMATCH);
2844                  }
2845              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2846              if (d < 256) d = md->lcc[d];              if (d < 256) d = md->lcc[d];
2847              if (fi >= max || eptr >= md->end_subject || fc == d)              if (fc == d) RRETURN(MATCH_NOMATCH);
               RRETURN(MATCH_NOMATCH);  
2848              }              }
2849            }            }
2850          else          else
# Line 2202  for (;;) Line 2853  for (;;)
2853            {            {
2854            for (fi = min;; fi++)            for (fi = min;; fi++)
2855              {              {
2856              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM29);
2857              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2858              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])              if (fi >= max) RRETURN(MATCH_NOMATCH);
2859                if (eptr >= md->end_subject)
2860                  {
2861                  SCHECK_PARTIAL();
2862                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2863                  }
2864                if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);
2865              }              }
2866            }            }
2867          /* Control never gets here */          /* Control never gets here */
# Line 2221  for (;;) Line 2877  for (;;)
2877          /* UTF-8 mode */          /* UTF-8 mode */
2878          if (utf8)          if (utf8)
2879            {            {
2880            register int d;            register unsigned int d;
2881            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2882              {              {
2883              int len = 1;              int len = 1;
# Line 2231  for (;;) Line 2887  for (;;)
2887              if (fc == d) break;              if (fc == d) break;
2888              eptr += len;              eptr += len;
2889              }              }
2890            for(;;)          if (possessive) continue;
2891            for(;;)
2892              {              {
2893              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM30);
2894              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2895              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2896              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2248  for (;;) Line 2905  for (;;)
2905              if (eptr >= md->end_subject || fc == md->lcc[*eptr]) break;              if (eptr >= md->end_subject || fc == md->lcc[*eptr]) break;
2906              eptr++;              eptr++;
2907              }              }
2908              if (possessive) continue;
2909            while (eptr >= pp)            while (eptr >= pp)
2910              {              {
2911              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM31);
2912              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2913              eptr--;              eptr--;
2914              }              }
# Line 2269  for (;;) Line 2927  for (;;)
2927        /* UTF-8 mode */        /* UTF-8 mode */
2928        if (utf8)        if (utf8)
2929          {          {
2930          register int d;          register unsigned int d;
2931          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2932            {            {
2933              if (eptr >= md->end_subject)
2934                {
2935                SCHECK_PARTIAL();
2936                RRETURN(MATCH_NOMATCH);
2937                }
2938            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
2939            if (fc == d) RRETURN(MATCH_NOMATCH);            if (fc == d) RRETURN(MATCH_NOMATCH);
2940            }            }
# Line 2281  for (;;) Line 2944  for (;;)
2944        /* Not UTF-8 mode */        /* Not UTF-8 mode */
2945          {          {
2946          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2947              {
2948              if (eptr >= md->end_subject)
2949                {
2950                SCHECK_PARTIAL();
2951                RRETURN(MATCH_NOMATCH);
2952                }
2953            if (fc == *eptr++) RRETURN(MATCH_NOMATCH);            if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
2954              }
2955          }          }
2956    
2957        if (min == max) continue;        if (min == max) continue;
# Line 2292  for (;;) Line 2962  for (;;)
2962          /* UTF-8 mode */          /* UTF-8 mode */
2963          if (utf8)          if (utf8)
2964            {            {
2965            register int d;            register unsigned int d;
2966            for (fi = min;; fi++)            for (fi = min;; fi++)
2967              {              {
2968              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32);
2969              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2970              GETCHARINC(d, eptr);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2971              if (fi >= max || eptr >= md->end_subject || fc == d)              if (eptr >= md->end_subject)
2972                  {
2973                  SCHECK_PARTIAL();
2974                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2975                  }
2976                GETCHARINC(d, eptr);
2977                if (fc == d) RRETURN(MATCH_NOMATCH);
2978              }              }
2979            }            }
2980          else          else
# Line 2308  for (;;) Line 2983  for (;;)
2983            {            {
2984            for (fi = min;; fi++)            for (fi = min;; fi++)
2985              {              {
2986              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM33);
2987              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2988              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)              if (fi >= max) RRETURN(MATCH_NOMATCH);
2989                if (eptr >= md->end_subject)
2990                  {
2991                  SCHECK_PARTIAL();
2992                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2993                  }
2994                if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
2995              }              }
2996            }            }
2997          /* Control never gets here */          /* Control never gets here */
# Line 2327  for (;;) Line 3007  for (;;)
3007          /* UTF-8 mode */          /* UTF-8 mode */
3008          if (utf8)          if (utf8)
3009            {            {
3010            register int d;            register unsigned int d;
3011            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3012              {              {
3013              int len = 1;              int len = 1;
# Line 2336  for (;;) Line 3016  for (;;)
3016              if (fc == d) break;              if (fc == d) break;
3017              eptr += len;              eptr += len;
3018              }              }
3019              if (possessive) continue;
3020            for(;;)            for(;;)
3021              {              {
3022              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM34);
3023              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3024              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
3025              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2353  for (;;) Line 3034  for (;;)
3034              if (eptr >= md->end_subject || fc == *eptr) break;              if (eptr >= md->end_subject || fc == *eptr) break;
3035              eptr++;              eptr++;
3036              }              }
3037              if (possessive) continue;
3038            while (eptr >= pp)            while (eptr >= pp)
3039              {              {
3040              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM35);
3041              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3042              eptr--;              eptr--;
3043              }              }
# Line 2384  for (;;) Line 3066  for (;;)
3066      ecode += 3;      ecode += 3;
3067      goto REPEATTYPE;      goto REPEATTYPE;
3068    
3069      case OP_TYPESTAR:      case OP_TYPEPOSSTAR:
3070      case OP_TYPEMINSTAR:      possessive = TRUE;
3071      case OP_TYPEPLUS:      min = 0;
3072      case OP_TYPEMINPLUS:      max = INT_MAX;
3073      case OP_TYPEQUERY:      ecode++;
3074      case OP_TYPEMINQUERY:      goto REPEATTYPE;
3075      c = *ecode++ - OP_TYPESTAR;  
3076      minimize = (c & 1) != 0;      case OP_TYPEPOSPLUS:
3077      min = rep_min[c];                 /* Pick up values from tables; */      possessive = TRUE;
3078      max = rep_max[c];                 /* zero for max => infinity */      min = 1;
3079      if (max == 0) max = INT_MAX;      max = INT_MAX;
3080        ecode++;
3081        goto REPEATTYPE;
3082    
3083        case OP_TYPEPOSQUERY:
3084        possessive = TRUE;
3085        min = 0;
3086        max = 1;
3087        ecode++;
3088        goto REPEATTYPE;
3089    
3090        case OP_TYPEPOSUPTO:
3091        possessive = TRUE;
3092        min = 0;
3093        max = GET2(ecode, 1);
3094        ecode += 3;
3095        goto REPEATTYPE;
3096    
3097        case OP_TYPESTAR:
3098        case OP_TYPEMINSTAR:
3099        case OP_TYPEPLUS:
3100        case OP_TYPEMINPLUS:
3101        case OP_TYPEQUERY:
3102        case OP_TYPEMINQUERY:
3103        c = *ecode++ - OP_TYPESTAR;
3104        minimize = (c & 1) != 0;
3105        min = rep_min[c];                 /* Pick up values from tables; */
3106        max = rep_max[c];                 /* zero for max => infinity */
3107        if (max == 0) max = INT_MAX;
3108    
3109      /* Common code for all repeated single character type matches. Note that      /* Common code for all repeated single character type matches. Note that
3110      in UTF-8 mode, '.' matches a character of any length, but for the other      in UTF-8 mode, '.' matches a character of any length, but for the other
# Line 2408  for (;;) Line 3118  for (;;)
3118        {        {
3119        prop_fail_result = ctype == OP_NOTPROP;        prop_fail_result = ctype == OP_NOTPROP;
3120        prop_type = *ecode++;        prop_type = *ecode++;
3121        if (prop_type >= 128)        prop_value = *ecode++;
         {  
         prop_test_against = prop_type - 128;  
         prop_test_variable = &prop_category;  
         }  
       else  
         {  
         prop_test_against = prop_type;  
         prop_test_variable = &prop_chartype;  
         }  
3122        }        }
3123      else prop_type = -1;      else prop_type = -1;
3124  #endif  #endif
3125    
3126      /* First, ensure the minimum number of matches are present. Use inline      /* First, ensure the minimum number of matches are present. Use inline
3127      code for maximizing the speed, and do the type test once at the start      code for maximizing the speed, and do the type test once at the start
3128      (i.e. keep it out of the loop). Also we can test that there are at least      (i.e. keep it out of the loop). Separate the UTF-8 code completely as that
     the minimum number of bytes before we start. This isn't as effective in  
     UTF-8 mode, but it does no harm. Separate the UTF-8 code completely as that  
3129      is tidier. Also separate the UCP code, which can be the same for both UTF-8      is tidier. Also separate the UCP code, which can be the same for both UTF-8
3130      and single-bytes. */      and single-bytes. */
3131    
     if (min > md->end_subject - eptr) RRETURN(MATCH_NOMATCH);  
3132      if (min > 0)      if (min > 0)
3133        {        {
3134  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3135        if (prop_type > 0)        if (prop_type >= 0)
3136          {          {
3137          for (i = 1; i <= min; i++)          switch(prop_type)
3138            {            {
3139            GETCHARINC(c, eptr);            case PT_ANY:
3140            prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase);            if (prop_fail_result) RRETURN(MATCH_NOMATCH);
3141            if ((*prop_test_variable == prop_test_against) == prop_fail_result)            for (i = 1; i <= min; i++)
3142              RRETURN(MATCH_NOMATCH);              {
3143                if (eptr >= md->end_subject)
3144                  {
3145                  SCHECK_PARTIAL();
3146                  RRETURN(MATCH_NOMATCH);
3147                  }
3148                GETCHARINCTEST(c, eptr);
3149                }
3150              break;
3151    
3152              case PT_LAMP:
3153              for (i = 1; i <= min; i++)
3154                {
3155                if (eptr >= md->end_subject)
3156                  {
3157                  SCHECK_PARTIAL();
3158                  RRETURN(MATCH_NOMATCH);
3159                  }
3160                GETCHARINCTEST(c, eptr);
3161                prop_chartype = UCD_CHARTYPE(c);
3162                if ((prop_chartype == ucp_Lu ||
3163                     prop_chartype == ucp_Ll ||
3164                     prop_chartype == ucp_Lt) == prop_fail_result)
3165                  RRETURN(MATCH_NOMATCH);
3166                }
3167              break;
3168    
3169              case PT_GC:
3170              for (i = 1; i <= min; i++)
3171                {
3172                if (eptr >= md->end_subject)
3173                  {
3174                  SCHECK_PARTIAL();
3175                  RRETURN(MATCH_NOMATCH);
3176                  }
3177                GETCHARINCTEST(c, eptr);
3178                prop_category = UCD_CATEGORY(c);
3179                if ((prop_category == prop_value) == prop_fail_result)
3180                  RRETURN(MATCH_NOMATCH);
3181                }
3182              break;
3183    
3184              case PT_PC:
3185              for (i = 1; i <= min; i++)
3186                {
3187                if (eptr >= md->end_subject)
3188                  {
3189                  SCHECK_PARTIAL();
3190                  RRETURN(MATCH_NOMATCH);
3191                  }
3192                GETCHARINCTEST(c, eptr);
3193                prop_chartype = UCD_CHARTYPE(c);
3194                if ((prop_chartype == prop_value) == prop_fail_result)
3195                  RRETURN(MATCH_NOMATCH);
3196                }
3197              break;
3198    
3199              case PT_SC:
3200              for (i = 1; i <= min; i++)
3201                {
3202                if (eptr >= md->end_subject)
3203                  {
3204                  SCHECK_PARTIAL();
3205                  RRETURN(MATCH_NOMATCH);
3206                  }
3207                GETCHARINCTEST(c, eptr);
3208                prop_script = UCD_SCRIPT(c);
3209                if ((prop_script == prop_value) == prop_fail_result)
3210                  RRETURN(MATCH_NOMATCH);
3211                }
3212              break;
3213    
3214              default:
3215              RRETURN(PCRE_ERROR_INTERNAL);
3216            }            }
3217          }          }
3218    
# Line 2452  for (;;) Line 3223  for (;;)
3223          {          {
3224          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3225            {            {
3226              if (eptr >= md->end_subject)
3227                {
3228                SCHECK_PARTIAL();
3229                RRETURN(MATCH_NOMATCH);
3230                }
3231            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3232            prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase);            prop_category = UCD_CATEGORY(c);
3233            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);
3234            while (eptr < md->end_subject)            while (eptr < md->end_subject)
3235              {              {
3236              int len = 1;              int len = 1;
3237              if (!utf8) c = *eptr; else              if (!utf8) c = *eptr;
3238                {                else { GETCHARLEN(c, eptr, len); }
3239                GETCHARLEN(c, eptr, len);              prop_category = UCD_CATEGORY(c);
               }  
             prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase);  
3240              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3241              eptr += len;              eptr += len;
3242              }              }
# Line 2480  for (;;) Line 3254  for (;;)
3254          case OP_ANY:          case OP_ANY:
3255          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3256            {            {
3257            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject)
3258               (*eptr++ == NEWLINE && (ims & PCRE_DOTALL) == 0))              {
3259                SCHECK_PARTIAL();
3260                RRETURN(MATCH_NOMATCH);
3261                }
3262              if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
3263              eptr++;
3264              while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3265              }
3266            break;
3267    
3268            case OP_ALLANY:
3269            for (i = 1; i <= min; i++)
3270              {
3271              if (eptr >= md->end_subject)
3272                {
3273                SCHECK_PARTIAL();
3274              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3275                }
3276              eptr++;
3277            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3278            }            }
3279          break;          break;
3280    
3281          case OP_ANYBYTE:          case OP_ANYBYTE:
3282            if (eptr > md->end_subject - min) RRETURN(MATCH_NOMATCH);
3283          eptr += min;          eptr += min;
3284          break;          break;
3285    
3286            case OP_ANYNL:
3287            for (i = 1; i <= min; i++)
3288              {
3289              if (eptr >= md->end_subject)
3290                {
3291                SCHECK_PARTIAL();
3292                RRETURN(MATCH_NOMATCH);
3293                }
3294              GETCHARINC(c, eptr);
3295              switch(c)
3296                {
3297                default: RRETURN(MATCH_NOMATCH);
3298                case 0x000d:
3299                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3300                break;
3301    
3302                case 0x000a:
3303                break;
3304    
3305                case 0x000b:
3306                case 0x000c:
3307                case 0x0085:
3308                case 0x2028:
3309                case 0x2029:
3310                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3311                break;
3312                }
3313              }
3314            break;
3315    
3316            case OP_NOT_HSPACE:
3317            for (i = 1; i <= min; i++)
3318              {
3319              if (eptr >= md->end_subject)
3320                {
3321                SCHECK_PARTIAL();
3322                RRETURN(MATCH_NOMATCH);
3323                }
3324              GETCHARINC(c, eptr);
3325              switch(c)
3326                {
3327                default: break;
3328                case 0x09:      /* HT */
3329                case 0x20:      /* SPACE */
3330                case 0xa0:      /* NBSP */
3331                case 0x1680:    /* OGHAM SPACE MARK */
3332                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3333                case 0x2000:    /* EN QUAD */
3334                case 0x2001:    /* EM QUAD */
3335                case 0x2002:    /* EN SPACE */
3336                case 0x2003:    /* EM SPACE */
3337                case 0x2004:    /* THREE-PER-EM SPACE */
3338                case 0x2005:    /* FOUR-PER-EM SPACE */
3339                case 0x2006:    /* SIX-PER-EM SPACE */
3340                case 0x2007:    /* FIGURE SPACE */
3341                case 0x2008:    /* PUNCTUATION SPACE */
3342                case 0x2009:    /* THIN SPACE */
3343                case 0x200A:    /* HAIR SPACE */
3344                case 0x202f:    /* NARROW NO-BREAK SPACE */
3345                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3346                case 0x3000:    /* IDEOGRAPHIC SPACE */
3347                RRETURN(MATCH_NOMATCH);
3348                }
3349              }
3350            break;
3351    
3352            case OP_HSPACE:
3353            for (i = 1; i <= min; i++)
3354              {
3355              if (eptr >= md->end_subject)
3356                {
3357                SCHECK_PARTIAL();
3358                RRETURN(MATCH_NOMATCH);
3359                }
3360              GETCHARINC(c, eptr);
3361              switch(c)
3362                {
3363                default: RRETURN(MATCH_NOMATCH);
3364                case 0x09:      /* HT */
3365                case 0x20:      /* SPACE */
3366                case 0xa0:      /* NBSP */
3367                case 0x1680:    /* OGHAM SPACE MARK */
3368                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3369                case 0x2000:    /* EN QUAD */
3370                case 0x2001:    /* EM QUAD */
3371                case 0x2002:    /* EN SPACE */
3372                case 0x2003:    /* EM SPACE */
3373                case 0x2004:    /* THREE-PER-EM SPACE */
3374                case 0x2005:    /* FOUR-PER-EM SPACE */
3375                case 0x2006:    /* SIX-PER-EM SPACE */
3376                case 0x2007:    /* FIGURE SPACE */
3377                case 0x2008:    /* PUNCTUATION SPACE */
3378                case 0x2009:    /* THIN SPACE */
3379                case 0x200A:    /* HAIR SPACE */
3380                case 0x202f:    /* NARROW NO-BREAK SPACE */
3381                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3382                case 0x3000:    /* IDEOGRAPHIC SPACE */
3383                break;
3384                }
3385              }
3386            break;
3387    
3388            case OP_NOT_VSPACE:
3389            for (i = 1; i <= min; i++)
3390              {
3391              if (eptr >= md->end_subject)
3392                {
3393                SCHECK_PARTIAL();
3394                RRETURN(MATCH_NOMATCH);
3395                }
3396              GETCHARINC(c, eptr);
3397              switch(c)
3398                {
3399                default: break;
3400                case 0x0a:      /* LF */
3401                case 0x0b:      /* VT */
3402                case 0x0c:      /* FF */
3403                case 0x0d:      /* CR */
3404                case 0x85:      /* NEL */
3405                case 0x2028:    /* LINE SEPARATOR */
3406                case 0x2029:    /* PARAGRAPH SEPARATOR */
3407                RRETURN(MATCH_NOMATCH);
3408                }
3409              }
3410            break;
3411    
3412            case OP_VSPACE:
3413            for (i = 1; i <= min; i++)
3414              {
3415              if (eptr >= md->end_subject)
3416                {
3417                SCHECK_PARTIAL();
3418                RRETURN(MATCH_NOMATCH);
3419                }
3420              GETCHARINC(c, eptr);
3421              switch(c)
3422                {
3423                default: RRETURN(MATCH_NOMATCH);
3424                case 0x0a:      /* LF */
3425                case 0x0b:      /* VT */
3426                case 0x0c:      /* FF */
3427                case 0x0d:      /* CR */
3428                case 0x85:      /* NEL */
3429                case 0x2028:    /* LINE SEPARATOR */
3430                case 0x2029:    /* PARAGRAPH SEPARATOR */
3431                break;
3432                }
3433              }
3434            break;
3435    
3436          case OP_NOT_DIGIT:          case OP_NOT_DIGIT:
3437          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3438            {            {
3439            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (eptr >= md->end_subject)
3440                {
3441                SCHECK_PARTIAL();
3442                RRETURN(MATCH_NOMATCH);
3443                }
3444            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
3445            if (c < 128 && (md->ctypes[c] & ctype_digit) != 0)            if (c < 128 && (md->ctypes[c] & ctype_digit) != 0)
3446              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
# Line 2504  for (;;) Line 3450  for (;;)
3450          case OP_DIGIT:          case OP_DIGIT:
3451          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3452            {            {
3453            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject)
3454               *eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)              {
3455                SCHECK_PARTIAL();
3456                RRETURN(MATCH_NOMATCH);
3457                }
3458              if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)
3459              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3460            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
3461            }            }
# Line 2514  for (;;) Line 3464  for (;;)
3464          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
3465          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3466            {            {
3467            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject)
3468               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_space) != 0))              {
3469                SCHECK_PARTIAL();
3470              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3471            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;              }
3472              if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)
3473                RRETURN(MATCH_NOMATCH);
3474              while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3475            }            }
3476          break;          break;
3477    
3478          case OP_WHITESPACE:          case OP_WHITESPACE:
3479          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3480            {            {
3481            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject)
3482               *eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)              {
3483                SCHECK_PARTIAL();
3484                RRETURN(MATCH_NOMATCH);
3485                }
3486              if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)
3487              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3488            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
3489            }            }
# Line 2535  for (;;) Line 3493  for (;;)
3493          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3494            {            {
3495            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3496               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_word) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0))
3497              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3498            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3499            }            }
3500          break;          break;
3501    
3502          case OP_WORDCHAR:          case OP_WORDCHAR:
3503          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3504            {            {
3505            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject)
3506               *eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)              {
3507                SCHECK_PARTIAL();
3508                RRETURN(MATCH_NOMATCH);
3509                }
3510              if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)
3511              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3512            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
3513            }            }
# Line 2564  for (;;) Line 3526  for (;;)
3526        switch(ctype)        switch(ctype)
3527          {          {
3528          case OP_ANY:          case OP_ANY:
3529          if ((ims & PCRE_DOTALL) == 0)          for (i = 1; i <= min; i++)
3530            {            {
3531            for (i = 1; i <= min; i++)            if (eptr >= md->end_subject)
3532              if (*eptr++ == NEWLINE) RRETURN(MATCH_NOMATCH);              {
3533                SCHECK_PARTIAL();
3534                RRETURN(MATCH_NOMATCH);
3535                }
3536              if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
3537              eptr++;
3538              }
3539            break;
3540    
3541            case OP_ALLANY:
3542            if (eptr > md->end_subject - min)
3543              {
3544              SCHECK_PARTIAL();
3545              RRETURN(MATCH_NOMATCH);
3546            }            }
3547          else eptr += min;          eptr += min;
3548          break;          break;
3549    
3550          case OP_ANYBYTE:          case OP_ANYBYTE:
3551            if (eptr > md->end_subject - min)
3552              {
3553              SCHECK_PARTIAL();
3554              RRETURN(MATCH_NOMATCH);
3555              }
3556          eptr += min;          eptr += min;
3557          break;          break;
3558    
3559            case OP_ANYNL:
3560            for (i = 1; i <= min; i++)
3561              {
3562              if (eptr >= md->end_subject)
3563                {
3564                SCHECK_PARTIAL();
3565                RRETURN(MATCH_NOMATCH);
3566                }
3567              switch(*eptr++)
3568                {
3569                default: RRETURN(MATCH_NOMATCH);
3570                case 0x000d:
3571                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3572                break;
3573                case 0x000a:
3574                break;
3575    
3576                case 0x000b:
3577                case 0x000c:
3578                case 0x0085:
3579                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3580                break;
3581                }
3582              }
3583            break;
3584    
3585            case OP_NOT_HSPACE:
3586            for (i = 1; i <= min; i++)
3587              {
3588              if (eptr >= md->end_subject)
3589                {
3590                SCHECK_PARTIAL();
3591                RRETURN(MATCH_NOMATCH);
3592                }
3593              switch(*eptr++)
3594                {
3595                default: break;
3596                case 0x09:      /* HT */
3597                case 0x20:      /* SPACE */
3598                case 0xa0:      /* NBSP */
3599                RRETURN(MATCH_NOMATCH);
3600                }
3601              }
3602            break;
3603    
3604            case OP_HSPACE:
3605            for (i = 1; i <= min; i++)
3606              {
3607              if (eptr >= md->end_subject)
3608                {
3609                SCHECK_PARTIAL();
3610                RRETURN(MATCH_NOMATCH);
3611                }
3612              switch(*eptr++)
3613                {
3614                default: RRETURN(MATCH_NOMATCH);
3615                case 0x09:      /* HT */
3616                case 0x20:      /* SPACE */
3617                case 0xa0:      /* NBSP */
3618                break;
3619                }
3620              }
3621            break;
3622    
3623            case OP_NOT_VSPACE:
3624            for (i = 1; i <= min; i++)
3625              {
3626              if (eptr >= md->end_subject)
3627                {
3628                SCHECK_PARTIAL();
3629                RRETURN(MATCH_NOMATCH);
3630                }
3631              switch(*eptr++)
3632                {
3633                default: break;
3634                case 0x0a:      /* LF */
3635                case 0x0b:      /* VT */
3636                case 0x0c:      /* FF */
3637                case 0x0d:      /* CR */
3638                case 0x85:      /* NEL */
3639                RRETURN(MATCH_NOMATCH);
3640                }
3641              }
3642            break;
3643    
3644            case OP_VSPACE:
3645            for (i = 1; i <= min; i++)
3646              {
3647              if (eptr >= md->end_subject)
3648                {
3649                SCHECK_PARTIAL();
3650                RRETURN(MATCH_NOMATCH);
3651                }
3652              switch(*eptr++)
3653                {
3654                default: RRETURN(MATCH_NOMATCH);
3655                case 0x0a:      /* LF */
3656                case 0x0b:      /* VT */
3657                case 0x0c:      /* FF */
3658                case 0x0d:      /* CR */
3659                case 0x85:      /* NEL */
3660                break;
3661                }
3662              }
3663            break;
3664    
3665          case OP_NOT_DIGIT:          case OP_NOT_DIGIT:
3666          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3667              {
3668              if (eptr >= md->end_subject)
3669                {
3670                SCHECK_PARTIAL();
3671                RRETURN(MATCH_NOMATCH);
3672                }
3673            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
3674              }
3675          break;          break;
3676    
3677          case OP_DIGIT:          case OP_DIGIT:
3678          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3679              {
3680              if (eptr >= md->end_subject)
3681                {
3682                SCHECK_PARTIAL();
3683                RRETURN(MATCH_NOMATCH);
3684                }
3685            if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);            if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);
3686              }
3687          break;          break;
3688    
3689          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
3690          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3691              {
3692              if (eptr >= md->end_subject)
3693                {
3694                SCHECK_PARTIAL();
3695                RRETURN(MATCH_NOMATCH);
3696                }
3697            if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);            if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);
3698              }
3699          break;          break;
3700    
3701          case OP_WHITESPACE:          case OP_WHITESPACE:
3702          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3703              {
3704              if (eptr >= md->end_subject)
3705                {
3706                SCHECK_PARTIAL();
3707                RRETURN(MATCH_NOMATCH);
3708                }
3709            if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);            if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);
3710              }
3711          break;          break;
3712    
3713          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
3714          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3715              {
3716              if (eptr >= md->end_subject)
3717                {
3718                SCHECK_PARTIAL();
3719                RRETURN(MATCH_NOMATCH);
3720                }
3721            if ((md->ctypes[*eptr++] & ctype_word) != 0)            if ((md->ctypes[*eptr++] & ctype_word) != 0)
3722              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3723              }
3724          break;          break;
3725    
3726          case OP_WORDCHAR:          case OP_WORDCHAR:
3727          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3728              {
3729              if (eptr >= md->end_subject)
3730                {
3731                SCHECK_PARTIAL();
3732                RRETURN(MATCH_NOMATCH);
3733                }
3734            if ((md->ctypes[*eptr++] & ctype_word) == 0)            if ((md->ctypes[*eptr++] & ctype_word) == 0)
3735              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3736              }
3737          break;          break;
3738    
3739          default:          default:
# Line 2624  for (;;) Line 3752  for (;;)
3752      if (minimize)      if (minimize)
3753        {        {
3754  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3755        if (prop_type > 0)        if (prop_type >= 0)
3756          {          {
3757          for (fi = min;; fi++)          switch(prop_type)
3758            {            {
3759            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            case PT_ANY:
3760            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            for (fi = min;; fi++)
3761            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              {
3762            GETCHARINC(c, eptr);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM36);
3763            prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3764            if ((*prop_test_variable == prop_test_against) == prop_fail_result)              if (fi >= max) RRETURN(MATCH_NOMATCH);
3765              RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject)
3766                  {
3767                  SCHECK_PARTIAL();
3768                  RRETURN(MATCH_NOMATCH);
3769                  }
3770                GETCHARINC(c, eptr);
3771                if (prop_fail_result) RRETURN(MATCH_NOMATCH);
3772                }
3773              /* Control never gets here */
3774    
3775              case PT_LAMP:
3776              for (fi = min;; fi++)
3777                {
3778                RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM37);
3779                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3780                if (fi >= max) RRETURN(MATCH_NOMATCH);
3781                if (eptr >= md->end_subject)
3782                  {
3783                  SCHECK_PARTIAL();
3784                  RRETURN(MATCH_NOMATCH);
3785                  }
3786                GETCHARINC(c, eptr);
3787                prop_chartype = UCD_CHARTYPE(c);
3788                if ((prop_chartype == ucp_Lu ||
3789                     prop_chartype == ucp_Ll ||
3790                     prop_chartype == ucp_Lt) == prop_fail_result)
3791                  RRETURN(MATCH_NOMATCH);
3792                }
3793              /* Control never gets here */
3794    
3795              case PT_GC:
3796              for (fi = min;; fi++)
3797                {
3798                RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM38);
3799                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3800                if (fi >= max) RRETURN(MATCH_NOMATCH);
3801                if (eptr >= md->end_subject)
3802                  {
3803                  SCHECK_PARTIAL();
3804                  RRETURN(MATCH_NOMATCH);
3805                  }
3806                GETCHARINC(c, eptr);
3807                prop_category = UCD_CATEGORY(c);
3808                if ((prop_category == prop_value) == prop_fail_result)
3809                  RRETURN(MATCH_NOMATCH);
3810                }
3811              /* Control never gets here */
3812    
3813              case PT_PC:
3814              for (fi = min;; fi++)
3815                {
3816                RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM39);
3817                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3818                if (fi >= max) RRETURN(MATCH_NOMATCH);
3819                if (eptr >= md->end_subject)
3820                  {
3821                  SCHECK_PARTIAL();
3822                  RRETURN(MATCH_NOMATCH);
3823                  }
3824                GETCHARINC(c, eptr);
3825                prop_chartype = UCD_CHARTYPE(c);
3826                if ((prop_chartype == prop_value) == prop_fail_result)
3827                  RRETURN(MATCH_NOMATCH);
3828                }
3829              /* Control never gets here */
3830    
3831              case PT_SC:
3832              for (fi = min;; fi++)
3833                {
3834                RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM40);
3835                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3836                if (fi >= max) RRETURN(MATCH_NOMATCH);
3837                if (eptr >= md->end_subject)
3838                  {
3839                  SCHECK_PARTIAL();
3840                  RRETURN(MATCH_NOMATCH);
3841                  }
3842                GETCHARINC(c, eptr);
3843                prop_script = UCD_SCRIPT(c);
3844                if ((prop_script == prop_value) == prop_fail_result)
3845                  RRETURN(MATCH_NOMATCH);
3846                }
3847              /* Control never gets here */
3848    
3849              default:
3850              RRETURN(PCRE_ERROR_INTERNAL);
3851            }            }
3852          }          }
3853    
# Line 2645  for (;;) Line 3858  for (;;)
3858          {          {
3859          for (fi = min;; fi++)          for (fi = min;; fi++)
3860            {            {
3861            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM41);
3862            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3863            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3864              if (eptr >= md->end_subject)
3865                {
3866                SCHECK_PARTIAL();
3867                RRETURN(MATCH_NOMATCH);
3868                }
3869            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3870            prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase);            prop_category = UCD_CATEGORY(c);
3871            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);
3872            while (eptr < md->end_subject)            while (eptr < md->end_subject)
3873              {              {
3874              int len = 1;              int len = 1;
3875              if (!utf8) c = *eptr; else              if (!utf8) c = *eptr;
3876                {                else { GETCHARLEN(c, eptr, len); }
3877                GETCHARLEN(c, eptr, len);              prop_category = UCD_CATEGORY(c);
               }  
             prop_category = ucp_findchar(c, &prop_chartype, &prop_othercase);  
3878              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3879              eptr += len;              eptr += len;
3880              }              }
# Line 2674  for (;;) Line 3890  for (;;)
3890          {          {
3891          for (fi = min;; fi++)          for (fi = min;; fi++)
3892            {            {
3893            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42);
3894            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3895            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3896              if (eptr >= md->end_subject)
3897                {
3898                SCHECK_PARTIAL();
3899                RRETURN(MATCH_NOMATCH);
3900                }
3901              if (ctype == OP_ANY && IS_NEWLINE(eptr))
3902                RRETURN(MATCH_NOMATCH);
3903            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
3904            switch(ctype)            switch(ctype)
3905              {              {
3906              case OP_ANY:              case OP_ANY:        /* This is the non-NL case */
3907              if ((ims & PCRE_DOTALL) == 0 && c == NEWLINE) RRETURN(MATCH_NOMATCH);              case OP_ALLANY:
3908                case OP_ANYBYTE:
3909              break;              break;
3910    
3911              case OP_ANYBYTE:              case OP_ANYNL:
3912                switch(c)
3913                  {
3914                  default: RRETURN(MATCH_NOMATCH);
3915                  case 0x000d:
3916                  if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3917                  break;
3918                  case 0x000a:
3919                  break;
3920    
3921                  case 0x000b:
3922                  case 0x000c:
3923                  case 0x0085:
3924                  case 0x2028:
3925                  case 0x2029:
3926                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3927                  break;
3928                  }
3929                break;
3930    
3931                case OP_NOT_HSPACE:
3932                switch(c)
3933                  {
3934                  default: break;
3935                  case 0x09:      /* HT */
3936                  case 0x20:      /* SPACE */
3937                  case 0xa0:      /* NBSP */
3938                  case 0x1680:    /* OGHAM SPACE MARK */
3939                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3940                  case 0x2000:    /* EN QUAD */
3941                  case 0x2001:    /* EM QUAD */
3942                  case 0x2002:    /* EN SPACE */
3943                  case 0x2003:    /* EM SPACE */
3944                  case 0x2004:    /* THREE-PER-EM SPACE */
3945                  case 0x2005:    /* FOUR-PER-EM SPACE */
3946                  case 0x2006:    /* SIX-PER-EM SPACE */
3947                  case 0x2007:    /* FIGURE SPACE */
3948                  case 0x2008:    /* PUNCTUATION SPACE */
3949                  case 0x2009:    /* THIN SPACE */
3950                  case 0x200A:    /* HAIR SPACE */
3951                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3952                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3953                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3954                  RRETURN(MATCH_NOMATCH);
3955                  }
3956                break;
3957    
3958                case OP_HSPACE:
3959                switch(c)
3960                  {
3961                  default: RRETURN(MATCH_NOMATCH);
3962                  case 0x09:      /* HT */
3963                  case 0x20:      /* SPACE */
3964                  case 0xa0:      /* NBSP */
3965                  case 0x1680:    /* OGHAM SPACE MARK */
3966                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3967                  case 0x2000:    /* EN QUAD */
3968                  case 0x2001:    /* EM QUAD */
3969                  case 0x2002:    /* EN SPACE */
3970                  case 0x2003:    /* EM SPACE */
3971                  case 0x2004:    /* THREE-PER-EM SPACE */
3972                  case 0x2005:    /* FOUR-PER-EM SPACE */
3973                  case 0x2006:    /* SIX-PER-EM SPACE */
3974                  case 0x2007:    /* FIGURE SPACE */
3975                  case 0x2008:    /* PUNCTUATION SPACE */
3976                  case 0x2009:    /* THIN SPACE */
3977                  case 0x200A:    /* HAIR SPACE */
3978                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3979                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3980                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3981                  break;
3982                  }
3983                break;
3984    
3985                case OP_NOT_VSPACE:
3986                switch(c)
3987                  {
3988                  default: break;
3989                  case 0x0a:      /* LF */
3990                  case 0x0b:      /* VT */
3991                  case 0x0c:      /* FF */
3992                  case 0x0d:      /* CR */
3993                  case 0x85:      /* NEL */
3994                  case 0x2028:    /* LINE SEPARATOR */
3995                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3996                  RRETURN(MATCH_NOMATCH);
3997                  }
3998                break;
3999    
4000                case OP_VSPACE:
4001                switch(c)
4002                  {
4003                  default: RRETURN(MATCH_NOMATCH);
4004                  case 0x0a:      /* LF */
4005                  case 0x0b:      /* VT */
4006                  case 0x0c:      /* FF */
4007                  case 0x0d:      /* CR */
4008                  case 0x85:      /* NEL */
4009                  case 0x2028:    /* LINE SEPARATOR */
4010                  case 0x2029:    /* PARAGRAPH SEPARATOR */
4011                  break;
4012                  }
4013              break;              break;
4014    
4015              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
# Line 2729  for (;;) Line 4053  for (;;)
4053          {          {
4054          for (fi = min;; fi++)          for (fi = min;; fi++)
4055            {            {
4056            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43);
4057            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4058            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
4059              if (eptr >= md->end_subject)
4060                {
4061                SCHECK_PARTIAL();
4062                RRETURN(MATCH_NOMATCH);
4063                }
4064              if (ctype == OP_ANY && IS_NEWLINE(eptr))
4065                RRETURN(MATCH_NOMATCH);
4066            c = *eptr++;            c = *eptr++;
4067            switch(ctype)            switch(ctype)
4068              {              {
4069              case OP_ANY:              case OP_ANY:     /* This is the non-NL case */
4070              if ((ims & PCRE_DOTALL) == 0 && c == NEWLINE) RRETURN(MATCH_NOMATCH);              case OP_ALLANY:
4071                case OP_ANYBYTE:
4072              break;              break;
4073    
4074              case OP_ANYBYTE:              case OP_ANYNL:
4075                switch(c)
4076                  {
4077                  default: RRETURN(MATCH_NOMATCH);
4078                  case 0x000d:
4079                  if (eptr < md->end_subject && *eptr == 0x0a) eptr++;