/[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 85 by nigel, Sat Feb 24 21:41:13 2007 UTC revision 336 by ph10, Sat Apr 12 15:59:03 2008 UTC
# Line 6  Line 6 
6  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
7    
8                         Written by Philip Hazel                         Written by Philip Hazel
9             Copyright (c) 1997-2005 University of Cambridge             Copyright (c) 1997-2008 University of Cambridge
10    
11  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
12  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 42  POSSIBILITY OF SUCH DAMAGE. Line 42  POSSIBILITY OF SUCH DAMAGE.
42  pattern matching using an NFA algorithm, trying to mimic Perl as closely as  pattern matching using an NFA algorithm, trying to mimic Perl as closely as
43  possible. There are also some static supporting functions. */  possible. There are also some static supporting functions. */
44    
45    #ifdef HAVE_CONFIG_H
46    #include "config.h"
47    #endif
48    
49    #define NLBLOCK md             /* Block containing newline information */
50    #define PSSTART start_subject  /* Field containing processed string start */
51    #define PSEND   end_subject    /* Field containing processed string end */
52    
53  #include "pcre_internal.h"  #include "pcre_internal.h"
54    
55    /* Undefine some potentially clashing cpp symbols */
56    
57  /* Structure for building a chain of data that actually lives on the  #undef min
58  stack, for holding the values of the subject pointer at the start of each  #undef max
 subpattern, so as to detect when an empty string has been matched by a  
 subpattern - to break infinite loops. When NO_RECURSE is set, these blocks  
 are on the heap, not on the stack. */  
   
 typedef struct eptrblock {  
   struct eptrblock *epb_prev;  
   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 169  return TRUE; Line 177  return TRUE;
177  ****************************************************************************  ****************************************************************************
178                     RECURSION IN THE match() FUNCTION                     RECURSION IN THE match() FUNCTION
179    
180  The match() function is highly recursive. Some regular expressions can cause  The match() function is highly recursive, though not every recursive call
181  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
182  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
183  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
184  works fine.  saved for a recursive call. On Unix, the stack can be large, and this works
185    fine.
186  It turns out that on non-Unix systems there are problems with programs that  
187  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
188  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
189  for decades.) So....  has oodles of memory these days, and techniques for extending the stack have
190    been known for decades.) So....
191    
192  There is a fudge, triggered by defining NO_RECURSE, which avoids recursive  There is a fudge, triggered by defining NO_RECURSE, which avoids recursive
193  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
194  obtained from malloc instead instead of on the stack. Macros are used to  obtained from malloc() instead instead of on the stack. Macros are used to
195  achieve this so that the actual code doesn't look very different to what it  achieve this so that the actual code doesn't look very different to what it
196  always used to.  always used to.
197    
198    The original heap-recursive code used longjmp(). However, it seems that this
199    can be very slow on some operating systems. Following a suggestion from Stan
200    Switzer, the use of longjmp() has been abolished, at the cost of having to
201    provide a unique number for each call to RMATCH. There is no way of generating
202    a sequence of numbers at compile time in C. I have given them names, to make
203    them stand out more clearly.
204    
205    Crude tests on x86 Linux show a small speedup of around 5-8%. However, on
206    FreeBSD, avoiding longjmp() more than halves the time taken to run the standard
207    tests. Furthermore, not using longjmp() means that local dynamic variables
208    don't have indeterminate values; this has meant that the frame size can be
209    reduced because the result can be "passed back" by straight setting of the
210    variable instead of being passed in the frame.
211  ****************************************************************************  ****************************************************************************
212  ***************************************************************************/  ***************************************************************************/
213    
214    /* Numbers for RMATCH calls. When this list is changed, the code at HEAP_RETURN
215    below must be updated in sync.  */
216    
217  /* These versions of the macros use the stack, as normal */  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM6,  RM7,  RM8,  RM9,  RM10,
218           RM11,  RM12, RM13, RM14, RM15, RM16, RM17, RM18, RM19, RM20,
219           RM21,  RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30,
220           RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
221           RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
222           RM51,  RM52, RM53, RM54 };
223    
224    /* These versions of the macros use the stack, as normal. There are debugging
225    versions and production versions. Note that the "rw" argument of RMATCH isn't
226    actuall used in this definition. */
227    
228  #ifndef NO_RECURSE  #ifndef NO_RECURSE
229  #define REGISTER register  #define REGISTER register
230  #define RMATCH(rx,ra,rb,rc,rd,re,rf,rg) rx = match(ra,rb,rc,rd,re,rf,rg)  
231    #ifdef DEBUG
232    #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \
233      { \
234      printf("match() called in line %d\n", __LINE__); \
235      rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1); \
236      printf("to line %d\n", __LINE__); \
237      }
238    #define RRETURN(ra) \
239      { \
240      printf("match() returned %d from line %d ", ra, __LINE__); \
241      return ra; \
242      }
243    #else
244    #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \
245      rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1)
246  #define RRETURN(ra) return ra  #define RRETURN(ra) return ra
247    #endif
248    
249  #else  #else
250    
251    
252  /* These versions of the macros manage a private stack on the heap. Note  /* These versions of the macros manage a private stack on the heap. Note that
253  that the rd argument of RMATCH isn't actually used. It's the md argument of  the "rd" argument of RMATCH isn't actually used in this definition. It's the md
254  match(), which never changes. */  argument of match(), which never changes. */
255    
256  #define REGISTER  #define REGISTER
257    
258  #define RMATCH(rx,ra,rb,rc,rd,re,rf,rg)\  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw)\
259    {\    {\
260    heapframe *newframe = (pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = (pcre_stack_malloc)(sizeof(heapframe));\
261    if (setjmp(frame->Xwhere) == 0)\    frame->Xwhere = rw; \
262      {\    newframe->Xeptr = ra;\
263      newframe->Xeptr = ra;\    newframe->Xecode = rb;\
264      newframe->Xecode = rb;\    newframe->Xmstart = mstart;\
265      newframe->Xoffset_top = rc;\    newframe->Xoffset_top = rc;\
266      newframe->Xims = re;\    newframe->Xims = re;\
267      newframe->Xeptrb = rf;\    newframe->Xeptrb = rf;\
268      newframe->Xflags = rg;\    newframe->Xflags = rg;\
269      newframe->Xprevframe = frame;\    newframe->Xrdepth = frame->Xrdepth + 1;\
270      frame = newframe;\    newframe->Xprevframe = frame;\
271      DPRINTF(("restarting from line %d\n", __LINE__));\    frame = newframe;\
272      goto HEAP_RECURSE;\    DPRINTF(("restarting from line %d\n", __LINE__));\
273      }\    goto HEAP_RECURSE;\
274    else\    L_##rw:\
275      {\    DPRINTF(("jumped back to line %d\n", __LINE__));\
     DPRINTF(("longjumped back to line %d\n", __LINE__));\  
     frame = md->thisframe;\  
     rx = frame->Xresult;\  
     }\  
276    }    }
277    
278  #define RRETURN(ra)\  #define RRETURN(ra)\
# Line 235  match(), which never changes. */ Line 282  match(), which never changes. */
282    (pcre_stack_free)(newframe);\    (pcre_stack_free)(newframe);\
283    if (frame != NULL)\    if (frame != NULL)\
284      {\      {\
285      frame->Xresult = ra;\      rrc = ra;\
286      md->thisframe = frame;\      goto HEAP_RETURN;\
     longjmp(frame->Xwhere, 1);\  
287      }\      }\
288    return ra;\    return ra;\
289    }    }
# Line 252  typedef struct heapframe { Line 298  typedef struct heapframe {
298    
299    const uschar *Xeptr;    const uschar *Xeptr;
300    const uschar *Xecode;    const uschar *Xecode;
301      const uschar *Xmstart;
302    int Xoffset_top;    int Xoffset_top;
303    long int Xims;    long int Xims;
304    eptrblock *Xeptrb;    eptrblock *Xeptrb;
305    int Xflags;    int Xflags;
306      unsigned int Xrdepth;
307    
308    /* Function local variables */    /* Function local variables */
309    
# Line 271  typedef struct heapframe { Line 319  typedef struct heapframe {
319    
320    BOOL Xcur_is_word;    BOOL Xcur_is_word;
321    BOOL Xcondition;    BOOL Xcondition;
   BOOL Xminimize;  
322    BOOL Xprev_is_word;    BOOL Xprev_is_word;
323    
324    unsigned long int Xoriginal_ims;    unsigned long int Xoriginal_ims;
325    
326  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
327    int Xprop_type;    int Xprop_type;
328      int Xprop_value;
329    int Xprop_fail_result;    int Xprop_fail_result;
330    int Xprop_category;    int Xprop_category;
331    int Xprop_chartype;    int Xprop_chartype;
332    int Xprop_othercase;    int Xprop_script;
333    int Xprop_test_against;    int Xoclength;
334    int *Xprop_test_variable;    uschar Xocchars[8];
335  #endif  #endif
336    
337    int Xctype;    int Xctype;
338    int Xfc;    unsigned int Xfc;
339    int Xfi;    int Xfi;
340    int Xlength;    int Xlength;
341    int Xmax;    int Xmax;
# Line 301  typedef struct heapframe { Line 349  typedef struct heapframe {
349    
350    eptrblock Xnewptrb;    eptrblock Xnewptrb;
351    
352    /* Place to pass back result, and where to jump back to */    /* Where to jump back to */
353    
354    int  Xresult;    int Xwhere;
   jmp_buf Xwhere;  
355    
356  } heapframe;  } heapframe;
357    
# Line 320  typedef struct heapframe { Line 367  typedef struct heapframe {
367  *         Match from current position            *  *         Match from current position            *
368  *************************************************/  *************************************************/
369    
370  /* On entry ecode points to the first opcode, and eptr to the first character  /* This function is called recursively in many circumstances. Whenever it
 in the subject string, while eptrb holds the value of eptr at the start of the  
 last bracketed group - used for breaking infinite loops matching zero-length  
 strings. This function is called recursively in many circumstances. Whenever it  
371  returns a negative (error) response, the outer incarnation must also return the  returns a negative (error) response, the outer incarnation must also return the
372  same response.  same response.
373    
# Line 333  performance. Tests using gcc on a SPARC Line 377  performance. Tests using gcc on a SPARC
377  made performance worse.  made performance worse.
378    
379  Arguments:  Arguments:
380     eptr        pointer in subject     eptr        pointer to current character in subject
381     ecode       position in code     ecode       pointer to current position in compiled code
382       mstart      pointer to the current match start position (can be modified
383                     by encountering \K)
384     offset_top  current top pointer     offset_top  current top pointer
385     md          pointer to "static" info for the match     md          pointer to "static" info for the match
386     ims         current /i, /m, and /s options     ims         current /i, /m, and /s options
# Line 342  Arguments: Line 388  Arguments:
388                   brackets - for testing for empty matches                   brackets - for testing for empty matches
389     flags       can contain     flags       can contain
390                   match_condassert - this is an assertion condition                   match_condassert - this is an assertion condition
391                   match_isgroup - this is the start of a bracketed group                   match_cbegroup - this is the start of an unlimited repeat
392                       group that can match an empty string
393       rdepth      the recursion depth
394    
395  Returns:       MATCH_MATCH if matched            )  these values are >= 0  Returns:       MATCH_MATCH if matched            )  these values are >= 0
396                 MATCH_NOMATCH if failed to match  )                 MATCH_NOMATCH if failed to match  )
397                 a negative PCRE_ERROR_xxx value if aborted by an error condition                 a negative PCRE_ERROR_xxx value if aborted by an error condition
398                   (e.g. stopped by recursion limit)                   (e.g. stopped by repeated call or recursion limit)
399  */  */
400    
401  static int  static int
402  match(REGISTER const uschar *eptr, REGISTER const uschar *ecode,  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, const uschar *mstart,
403    int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb,    int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb,
404    int flags)    int flags, unsigned int rdepth)
405  {  {
406  /* These variables do not need to be preserved over recursion in this function,  /* These variables do not need to be preserved over recursion in this function,
407  so they can be ordinary variables in all cases. Mark them with "register"  so they can be ordinary variables in all cases. Mark some of them with
408  because they are used a lot in loops. */  "register" because they are used a lot in loops. */
409    
410  register int  rrc;    /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
411  register int  i;      /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
412  register int  c;      /* Character values not kept over RMATCH() calls */  register unsigned int c;   /* Character values not kept over RMATCH() calls */
413  register BOOL utf8;   /* Local copy of UTF-8 flag for speed */  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */
414    
415    BOOL minimize, possessive; /* Quantifier options */
416    
417  /* When recursion is not being used, all "local" variables that have to be  /* When recursion is not being used, all "local" variables that have to be
418  preserved over calls to RMATCH() are part of a "frame" which is obtained from  preserved over calls to RMATCH() are part of a "frame" which is obtained from
# Line 377  frame->Xprevframe = NULL;            /* Line 427  frame->Xprevframe = NULL;            /*
427    
428  frame->Xeptr = eptr;  frame->Xeptr = eptr;
429  frame->Xecode = ecode;  frame->Xecode = ecode;
430    frame->Xmstart = mstart;
431  frame->Xoffset_top = offset_top;  frame->Xoffset_top = offset_top;
432  frame->Xims = ims;  frame->Xims = ims;
433  frame->Xeptrb = eptrb;  frame->Xeptrb = eptrb;
434  frame->Xflags = flags;  frame->Xflags = flags;
435    frame->Xrdepth = rdepth;
436    
437  /* This is where control jumps back to to effect "recursion" */  /* This is where control jumps back to to effect "recursion" */
438    
# Line 390  HEAP_RECURSE: Line 442  HEAP_RECURSE:
442    
443  #define eptr               frame->Xeptr  #define eptr               frame->Xeptr
444  #define ecode              frame->Xecode  #define ecode              frame->Xecode
445    #define mstart             frame->Xmstart
446  #define offset_top         frame->Xoffset_top  #define offset_top         frame->Xoffset_top
447  #define ims                frame->Xims  #define ims                frame->Xims
448  #define eptrb              frame->Xeptrb  #define eptrb              frame->Xeptrb
449  #define flags              frame->Xflags  #define flags              frame->Xflags
450    #define rdepth             frame->Xrdepth
451    
452  /* Ditto for the local variables */  /* Ditto for the local variables */
453    
# Line 411  HEAP_RECURSE: Line 465  HEAP_RECURSE:
465    
466  #define cur_is_word        frame->Xcur_is_word  #define cur_is_word        frame->Xcur_is_word
467  #define condition          frame->Xcondition  #define condition          frame->Xcondition
 #define minimize           frame->Xminimize  
468  #define prev_is_word       frame->Xprev_is_word  #define prev_is_word       frame->Xprev_is_word
469    
470  #define original_ims       frame->Xoriginal_ims  #define original_ims       frame->Xoriginal_ims
471    
472  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
473  #define prop_type          frame->Xprop_type  #define prop_type          frame->Xprop_type
474    #define prop_value         frame->Xprop_value
475  #define prop_fail_result   frame->Xprop_fail_result  #define prop_fail_result   frame->Xprop_fail_result
476  #define prop_category      frame->Xprop_category  #define prop_category      frame->Xprop_category
477  #define prop_chartype      frame->Xprop_chartype  #define prop_chartype      frame->Xprop_chartype
478  #define prop_othercase     frame->Xprop_othercase  #define prop_script        frame->Xprop_script
479  #define prop_test_against  frame->Xprop_test_against  #define oclength           frame->Xoclength
480  #define prop_test_variable frame->Xprop_test_variable  #define occhars            frame->Xocchars
481  #endif  #endif
482    
483  #define ctype              frame->Xctype  #define ctype              frame->Xctype
# Line 447  HEAP_RECURSE: Line 501  HEAP_RECURSE:
501  get preserved during recursion in the normal way. In this environment, fi and  get preserved during recursion in the normal way. In this environment, fi and
502  i, and fc and c, can be the same variables. */  i, and fc and c, can be the same variables. */
503    
504  #else  #else         /* NO_RECURSE not defined */
505  #define fi i  #define fi i
506  #define fc c  #define fc c
507    
508    
509  #ifdef SUPPORT_UTF8                /* Many of these variables are used ony */  #ifdef SUPPORT_UTF8                /* Many of these variables are used only  */
510  const uschar *charptr;             /* small blocks of the code. My normal  */  const uschar *charptr;             /* in small blocks of the code. My normal */
511  #endif                             /* style of coding would have declared  */  #endif                             /* style of coding would have declared    */
512  const uschar *callpat;             /* them within each of those blocks.    */  const uschar *callpat;             /* them within each of those blocks.      */
513  const uschar *data;                /* However, in order to accommodate the */  const uschar *data;                /* However, in order to accommodate the   */
514  const uschar *next;                /* version of this code that uses an    */  const uschar *next;                /* version of this code that uses an      */
515  const uschar *pp;                  /* external "stack" implemented on the  */  USPTR         pp;                  /* external "stack" implemented on the    */
516  const uschar *prev;                /* heap, it is easier to declare them   */  const uschar *prev;                /* heap, it is easier to declare them all */
517  const uschar *saved_eptr;          /* all here, so the declarations can    */  USPTR         saved_eptr;          /* here, so the declarations can be cut   */
518                                     /* be cut out in a block. The only      */                                     /* out in a block. The only declarations  */
519  recursion_info new_recursive;      /* declarations within blocks below are */  recursion_info new_recursive;      /* within blocks below are for variables  */
520                                     /* for variables that do not have to    */                                     /* that do not have to be preserved over  */
521  BOOL cur_is_word;                  /* be preserved over a recursive call   */  BOOL cur_is_word;                  /* a recursive call to RMATCH().          */
522  BOOL condition;                    /* to RMATCH().                         */  BOOL condition;
 BOOL minimize;  
523  BOOL prev_is_word;  BOOL prev_is_word;
524    
525  unsigned long int original_ims;  unsigned long int original_ims;
526    
527  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
528  int prop_type;  int prop_type;
529    int prop_value;
530  int prop_fail_result;  int prop_fail_result;
531  int prop_category;  int prop_category;
532  int prop_chartype;  int prop_chartype;
533  int prop_othercase;  int prop_script;
534  int prop_test_against;  int oclength;
535  int *prop_test_variable;  uschar occhars[8];
536  #endif  #endif
537    
538  int ctype;  int ctype;
# Line 493  int save_offset1, save_offset2, save_off Line 547  int save_offset1, save_offset2, save_off
547  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
548    
549  eptrblock newptrb;  eptrblock newptrb;
550  #endif  #endif     /* NO_RECURSE */
551    
552  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
553  variables. */  variables. */
554    
555  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
556    prop_value = 0;
557  prop_fail_result = 0;  prop_fail_result = 0;
 prop_test_against = 0;  
 prop_test_variable = NULL;  
558  #endif  #endif
559    
560  /* OK, now we can get on with the real code of the function. Recursion is  
561  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
562  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
563  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
564  macro. It has to be used in one particular way. This shouldn't, however, impact  original patch. */
565  performance when true recursion is being used. */  
566    TAIL_RECURSE:
567    
568    /* OK, now we can get on with the real code of the function. Recursive calls
569    are specified by the macro RMATCH and RRETURN is used to return. When
570    NO_RECURSE is *not* defined, these just turn into a recursive call to match()
571    and a "return", respectively (possibly with some debugging if DEBUG is
572    defined). However, RMATCH isn't like a function call because it's quite a
573    complicated macro. It has to be used in one particular way. This shouldn't,
574    however, impact performance when true recursion is being used. */
575    
576    #ifdef SUPPORT_UTF8
577    utf8 = md->utf8;       /* Local copy of the flag */
578    #else
579    utf8 = FALSE;
580    #endif
581    
582    /* First check that we haven't called match() too many times, or that we
583    haven't exceeded the recursive call limit. */
584    
585  if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT);  if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT);
586    if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);
587    
588  original_ims = ims;    /* Save for resetting on ')' */  original_ims = ims;    /* Save for resetting on ')' */
 utf8 = md->utf8;       /* Local copy of the flag */  
589    
590  /* 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
591  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
592  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
593  this stack. */  hit the closing ket, in order to break infinite loops that match no characters.
594    When match() is called in other circumstances, don't add to the chain. The
595    match_cbegroup flag must NOT be used with tail recursion, because the memory
596    block that is used is on the stack, so a new one may be required for each
597    match(). */
598    
599  if ((flags & match_isgroup) != 0)  if ((flags & match_cbegroup) != 0)
600    {    {
   newptrb.epb_prev = eptrb;  
601    newptrb.epb_saved_eptr = eptr;    newptrb.epb_saved_eptr = eptr;
602      newptrb.epb_prev = eptrb;
603    eptrb = &newptrb;    eptrb = &newptrb;
604    }    }
605    
606  /* Now start processing the operations. */  /* Now start processing the opcodes. */
607    
608  for (;;)  for (;;)
609    {    {
610      minimize = possessive = FALSE;
611    op = *ecode;    op = *ecode;
   minimize = FALSE;  
612    
613    /* For partial matching, remember if we ever hit the end of the subject after    /* For partial matching, remember if we ever hit the end of the subject after
614    matching at least one subject character. */    matching at least one subject character. */
615    
616    if (md->partial &&    if (md->partial &&
617        eptr >= md->end_subject &&        eptr >= md->end_subject &&
618        eptr > md->start_match)        eptr > mstart)
619      md->hitend = TRUE;      md->hitend = TRUE;
620    
621    /* Opening capturing bracket. If there is space in the offset vector, save    switch(op)
   the current subject position in the working slot at the top of the vector. We  
   mustn't change the current values of the data slot, because they may be set  
   from a previous iteration of this group, and be referred to by a reference  
   inside the group.  
   
   If the bracket fails to match, we need to restore this value and also the  
   values of the final offsets, in case they were set by a previous iteration of  
   the same bracket.  
   
   If there isn't enough space in the offset vector, treat this as if it were a  
   non-capturing bracket. Don't worry about setting the flag for the error case  
   here; that is handled in the code for KET. */  
   
   if (op > OP_BRA)  
622      {      {
623      number = op - OP_BRA;      case OP_FAIL:
624        RRETURN(MATCH_NOMATCH);
     /* For extended extraction brackets (large number), we have to fish out the  
     number from a dummy opcode at the start. */  
625    
626      if (number > EXTRACT_BASIC_MAX)      case OP_PRUNE:
627        number = GET2(ecode, 2+LINK_SIZE);      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
628          ims, eptrb, flags, RM51);
629        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
630        RRETURN(MATCH_PRUNE);
631    
632        case OP_COMMIT:
633        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
634          ims, eptrb, flags, RM52);
635        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
636        RRETURN(MATCH_COMMIT);
637    
638        case OP_SKIP:
639        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
640          ims, eptrb, flags, RM53);
641        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
642        md->start_match_ptr = eptr;   /* Pass back current position */
643        RRETURN(MATCH_SKIP);
644    
645        case OP_THEN:
646        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
647          ims, eptrb, flags, RM54);
648        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
649        RRETURN(MATCH_THEN);
650    
651        /* Handle a capturing bracket. If there is space in the offset vector, save
652        the current subject position in the working slot at the top of the vector.
653        We mustn't change the current values of the data slot, because they may be
654        set from a previous iteration of this group, and be referred to by a
655        reference inside the group.
656    
657        If the bracket fails to match, we need to restore this value and also the
658        values of the final offsets, in case they were set by a previous iteration
659        of the same bracket.
660    
661        If there isn't enough space in the offset vector, treat this as if it were
662        a non-capturing bracket. Don't worry about setting the flag for the error
663        case here; that is handled in the code for KET. */
664    
665        case OP_CBRA:
666        case OP_SCBRA:
667        number = GET2(ecode, 1+LINK_SIZE);
668      offset = number << 1;      offset = number << 1;
669    
670  #ifdef DEBUG  #ifdef DEBUG
671      printf("start bracket %d subject=", number);      printf("start bracket %d\n", number);
672        printf("subject=");
673      pchars(eptr, 16, TRUE, md);      pchars(eptr, 16, TRUE, md);
674      printf("\n");      printf("\n");
675  #endif  #endif
# Line 584  for (;;) Line 684  for (;;)
684        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
685        md->offset_vector[md->offset_end - number] = eptr - md->start_subject;        md->offset_vector[md->offset_end - number] = eptr - md->start_subject;
686    
687          flags = (op == OP_SCBRA)? match_cbegroup : 0;
688        do        do
689          {          {
690          RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb,          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
691            match_isgroup);            ims, eptrb, flags, RM1);
692          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
693          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
694          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
695          }          }
# Line 603  for (;;) Line 704  for (;;)
704        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
705        }        }
706    
707      /* Insufficient room for saving captured contents */      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
708        as a non-capturing bracket. */
709    
710      else op = OP_BRA;      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
711      }      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
712    
713    /* Other types of node can be handled by a switch */      DPRINTF(("insufficient capture room: treat as non-capturing\n"));
714    
715    switch(op)      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
716      {      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
717      case OP_BRA:     /* Non-capturing bracket: optimized */  
718      DPRINTF(("start bracket 0\n"));      /* Non-capturing bracket. Loop for all the alternatives. When we get to the
719      do      final alternative within the brackets, we would return the result of a
720        recursive call to match() whatever happened. We can reduce stack usage by
721        turning this into a tail recursion, except in the case when match_cbegroup
722        is set.*/
723    
724        case OP_BRA:
725        case OP_SBRA:
726        DPRINTF(("start non-capturing bracket\n"));
727        flags = (op >= OP_SBRA)? match_cbegroup : 0;
728        for (;;)
729        {        {
730        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb,        if (ecode[GET(ecode, 1)] != OP_ALT)   /* Final alternative */
731          match_isgroup);          {
732        if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (flags == 0)    /* Not a possibly empty group */
733              {
734              ecode += _pcre_OP_lengths[*ecode];
735              DPRINTF(("bracket 0 tail recursion\n"));
736              goto TAIL_RECURSE;
737              }
738    
739            /* Possibly empty group; can't use tail recursion. */
740    
741            RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
742              eptrb, flags, RM48);
743            RRETURN(rrc);
744            }
745    
746          /* For non-final alternatives, continue the loop for a NOMATCH result;
747          otherwise return. */
748    
749          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
750            eptrb, flags, RM2);
751          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
752        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
753        }        }
754      while (*ecode == OP_ALT);      /* Control never reaches here. */
     DPRINTF(("bracket 0 failed\n"));  
     RRETURN(MATCH_NOMATCH);  
755    
756      /* Conditional group: compilation checked that there are no more than      /* Conditional group: compilation checked that there are no more than
757      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
758      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
759      exactly what going to the ket would do. */      exactly what going to the ket would do. As there is only one branch to be
760        obeyed, we can use tail recursion to avoid using another stack frame. */
761    
762      case OP_COND:      case OP_COND:
763      if (ecode[LINK_SIZE+1] == OP_CREF) /* Condition extract or recurse test */      case OP_SCOND:
764        if (ecode[LINK_SIZE+1] == OP_RREF)         /* Recursion test */
765          {
766          offset = GET2(ecode, LINK_SIZE + 2);     /* Recursion group number*/
767          condition = md->recursive != NULL &&
768            (offset == RREF_ANY || offset == md->recursive->group_num);
769          ecode += condition? 3 : GET(ecode, 1);
770          }
771    
772        else if (ecode[LINK_SIZE+1] == OP_CREF)    /* Group used test */
773        {        {
774        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */
775        condition = (offset == CREF_RECURSE * 2)?        condition = offset < offset_top && md->offset_vector[offset] >= 0;
776          (md->recursive != NULL) :        ecode += condition? 3 : GET(ecode, 1);
777          (offset < offset_top && md->offset_vector[offset] >= 0);        }
778        RMATCH(rrc, eptr, ecode + (condition?  
779          (LINK_SIZE + 4) : (LINK_SIZE + 1 + GET(ecode, 1))),      else if (ecode[LINK_SIZE+1] == OP_DEF)     /* DEFINE - always false */
780          offset_top, md, ims, eptrb, match_isgroup);        {
781        RRETURN(rrc);        condition = FALSE;
782          ecode += GET(ecode, 1);
783        }        }
784    
785      /* The condition is an assertion. Call match() to evaluate it - setting      /* The condition is an assertion. Call match() to evaluate it - setting
786      the final argument TRUE causes it to stop at the end of an assertion. */      the final argument match_condassert causes it to stop at the end of an
787        assertion. */
788    
789      else      else
790        {        {
791        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,
792            match_condassert | match_isgroup);            match_condassert, RM3);
793        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
794          {          {
795          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE+2);          condition = TRUE;
796            ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
797          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
798          }          }
799        else if (rrc != MATCH_NOMATCH)        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
800          {          {
801          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
802          }          }
803        else ecode += GET(ecode, 1);        else
804        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb,          {
805          match_isgroup);          condition = FALSE;
806        RRETURN(rrc);          ecode += GET(ecode, 1);
807            }
808        }        }
     /* Control never reaches here */  
809    
810      /* 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,
811      encountered. */      we can use tail recursion to avoid using another stack frame, except when
812        match_cbegroup is required for an unlimited repeat of a possibly empty
813        group. If the second alternative doesn't exist, we can just plough on. */
814    
815      case OP_CREF:      if (condition || *ecode == OP_ALT)
816      case OP_BRANUMBER:        {
817      ecode += 3;        ecode += 1 + LINK_SIZE;
818          if (op == OP_SCOND)        /* Possibly empty group */
819            {
820            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, match_cbegroup, RM49);
821            RRETURN(rrc);
822            }
823          else                       /* Group must match something */
824            {
825            flags = 0;
826            goto TAIL_RECURSE;
827            }
828          }
829        else                         /* Condition false & no 2nd alternative */
830          {
831          ecode += 1 + LINK_SIZE;
832          }
833      break;      break;
834    
     /* End of the pattern. If we are in a recursion, we should restore the  
     offsets appropriately and continue from after the call. */  
835    
836        /* End of the pattern, either real or forced. If we are in a top-level
837        recursion, we should restore the offsets appropriately and continue from
838        after the call. */
839    
840        case OP_ACCEPT:
841      case OP_END:      case OP_END:
842      if (md->recursive != NULL && md->recursive->group_num == 0)      if (md->recursive != NULL && md->recursive->group_num == 0)
843        {        {
844        recursion_info *rec = md->recursive;        recursion_info *rec = md->recursive;
845        DPRINTF(("Hit the end in a (?0) recursion\n"));        DPRINTF(("End of pattern in a (?0) recursion\n"));
846        md->recursive = rec->prevrec;        md->recursive = rec->prevrec;
847        memmove(md->offset_vector, rec->offset_save,        memmove(md->offset_vector, rec->offset_save,
848          rec->saved_max * sizeof(int));          rec->saved_max * sizeof(int));
849        md->start_match = rec->save_start;        mstart = rec->save_start;
850        ims = original_ims;        ims = original_ims;
851        ecode = rec->after_call;        ecode = rec->after_call;
852        break;        break;
# Line 694  for (;;) Line 855  for (;;)
855      /* Otherwise, if PCRE_NOTEMPTY is set, fail if we have matched an empty      /* Otherwise, if PCRE_NOTEMPTY is set, fail if we have matched an empty
856      string - backtracking will then try other alternatives, if any. */      string - backtracking will then try other alternatives, if any. */
857    
858      if (md->notempty && eptr == md->start_match) RRETURN(MATCH_NOMATCH);      if (md->notempty && eptr == mstart) RRETURN(MATCH_NOMATCH);
859      md->end_match_ptr = eptr;          /* Record where we ended */      md->end_match_ptr = eptr;           /* Record where we ended */
860      md->end_offset_top = offset_top;   /* and how many extracts were taken */      md->end_offset_top = offset_top;    /* and how many extracts were taken */
861        md->start_match_ptr = mstart;       /* and the start (\K can modify) */
862      RRETURN(MATCH_MATCH);      RRETURN(MATCH_MATCH);
863    
864      /* Change option settings */      /* Change option settings */
# Line 717  for (;;) Line 879  for (;;)
879      case OP_ASSERTBACK:      case OP_ASSERTBACK:
880      do      do
881        {        {
882        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
883          match_isgroup);          RM4);
884        if (rrc == MATCH_MATCH) break;        if (rrc == MATCH_MATCH) break;
885        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
886        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
887        }        }
888      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 744  for (;;) Line 906  for (;;)
906      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
907      do      do
908        {        {
909        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
910          match_isgroup);          RM5);
911        if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH);
912        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
913        ecode += GET(ecode,1);        ecode += GET(ecode,1);
914        }        }
915      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 766  for (;;) Line 928  for (;;)
928  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
929      if (utf8)      if (utf8)
930        {        {
931        c = GET(ecode,1);        i = GET(ecode, 1);
932        for (i = 0; i < c; i++)        while (i-- > 0)
933          {          {
934          eptr--;          eptr--;
935          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
936          BACKCHAR(eptr)          BACKCHAR(eptr);
937          }          }
938        }        }
939      else      else
# Line 780  for (;;) Line 942  for (;;)
942      /* No UTF-8 support, or not in UTF-8 mode: count is byte count */      /* No UTF-8 support, or not in UTF-8 mode: count is byte count */
943    
944        {        {
945        eptr -= GET(ecode,1);        eptr -= GET(ecode, 1);
946        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
947        }        }
948    
# Line 800  for (;;) Line 962  for (;;)
962        cb.version          = 1;   /* Version 1 of the callout block */        cb.version          = 1;   /* Version 1 of the callout block */
963        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
964        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
965        cb.subject          = (const char *)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
966        cb.subject_length   = md->end_subject - md->start_subject;        cb.subject_length   = md->end_subject - md->start_subject;
967        cb.start_match      = md->start_match - md->start_subject;        cb.start_match      = mstart - md->start_subject;
968        cb.current_position = eptr - md->start_subject;        cb.current_position = eptr - md->start_subject;
969        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
970        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
# Line 837  for (;;) Line 999  for (;;)
999      case OP_RECURSE:      case OP_RECURSE:
1000        {        {
1001        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1002        new_recursive.group_num = *callpat - OP_BRA;        new_recursive.group_num = (callpat == md->start_code)? 0 :
1003            GET2(callpat, 1 + LINK_SIZE);
       /* For extended extraction brackets (large number), we have to fish out  
       the number from a dummy opcode at the start. */  
   
       if (new_recursive.group_num > EXTRACT_BASIC_MAX)  
         new_recursive.group_num = GET2(callpat, 2+LINK_SIZE);  
1004    
1005        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1006    
# Line 869  for (;;) Line 1026  for (;;)
1026    
1027        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1028              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1029        new_recursive.save_start = md->start_match;        new_recursive.save_start = mstart;
1030        md->start_match = eptr;        mstart = eptr;
1031    
1032        /* OK, now we can do the recursion. For each top-level alternative we        /* OK, now we can do the recursion. For each top-level alternative we
1033        restore the offset and recursion data. */        restore the offset and recursion data. */
1034    
1035        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1036          flags = (*callpat >= OP_SBRA)? match_cbegroup : 0;
1037        do        do
1038          {          {
1039          RMATCH(rrc, eptr, callpat + 1 + LINK_SIZE, offset_top, md, ims,          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,
1040              eptrb, match_isgroup);            md, ims, eptrb, flags, RM6);
1041          if (rrc == MATCH_MATCH)          if (rrc == MATCH_MATCH)
1042            {            {
1043              DPRINTF(("Recursion matched\n"));
1044            md->recursive = new_recursive.prevrec;            md->recursive = new_recursive.prevrec;
1045            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1046              (pcre_free)(new_recursive.offset_save);              (pcre_free)(new_recursive.offset_save);
1047            RRETURN(MATCH_MATCH);            RRETURN(MATCH_MATCH);
1048            }            }
1049          else if (rrc != MATCH_NOMATCH) RRETURN(rrc);          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1050              {
1051              DPRINTF(("Recursion gave error %d\n", rrc));
1052              RRETURN(rrc);
1053              }
1054    
1055          md->recursive = &new_recursive;          md->recursive = &new_recursive;
1056          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
# Line 912  for (;;) Line 1075  for (;;)
1075      the end of a normal bracket, leaving the subject pointer. */      the end of a normal bracket, leaving the subject pointer. */
1076    
1077      case OP_ONCE:      case OP_ONCE:
1078        {      prev = ecode;
1079        prev = ecode;      saved_eptr = eptr;
       saved_eptr = eptr;  
1080    
1081        do      do
1082          {        {
1083          RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM7);
1084            eptrb, match_isgroup);        if (rrc == MATCH_MATCH) break;
1085          if (rrc == MATCH_MATCH) break;        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1086          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        ecode += GET(ecode,1);
1087          ecode += GET(ecode,1);        }
1088          }      while (*ecode == OP_ALT);
       while (*ecode == OP_ALT);  
1089    
1090        /* If hit the end of the group (which could be repeated), fail */      /* If hit the end of the group (which could be repeated), fail */
1091    
1092        if (*ecode != OP_ONCE && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);      if (*ecode != OP_ONCE && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);
1093    
1094        /* Continue as from after the assertion, updating the offsets high water      /* Continue as from after the assertion, updating the offsets high water
1095        mark, since extracts may have been taken. */      mark, since extracts may have been taken. */
1096    
1097        do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode, 1); while (*ecode == OP_ALT);
1098    
1099        offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1100        eptr = md->end_match_ptr;      eptr = md->end_match_ptr;
1101    
1102        /* For a non-repeating ket, just continue at this level. This also      /* For a non-repeating ket, just continue at this level. This also
1103        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.
1104        This is the forcible breaking of infinite loops as implemented in Perl      This is the forcible breaking of infinite loops as implemented in Perl
1105        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
1106        course of events. */      course of events. */
1107    
1108        if (*ecode == OP_KET || eptr == saved_eptr)      if (*ecode == OP_KET || eptr == saved_eptr)
1109          {        {
1110          ecode += 1+LINK_SIZE;        ecode += 1+LINK_SIZE;
1111          break;        break;
1112          }        }
1113    
1114        /* 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
1115        preceding bracket, in the appropriate order. We need to reset any options      preceding bracket, in the appropriate order. The second "call" of match()
1116        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
1117        opcode. */      any options that changed within the bracket before re-running it, so
1118        check the next opcode. */
1119    
1120        if (ecode[1+LINK_SIZE] == OP_OPT)      if (ecode[1+LINK_SIZE] == OP_OPT)
1121          {        {
1122          ims = (ims & ~PCRE_IMS) | ecode[4];        ims = (ims & ~PCRE_IMS) | ecode[4];
1123          DPRINTF(("ims set to %02lx at group repeat\n", ims));        DPRINTF(("ims set to %02lx at group repeat\n", ims));
1124          }        }
1125    
1126        if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1127          {        {
1128          RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM8);
1129          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1130          RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);        ecode = prev;
1131          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        flags = 0;
1132          }        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);  
         }  
1133        }        }
1134      RRETURN(MATCH_NOMATCH);      else  /* OP_KETRMAX */
1135          {
1136          RMATCH(eptr, prev, offset_top, md, ims, eptrb, match_cbegroup, RM9);
1137          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1138          ecode += 1 + LINK_SIZE;
1139          flags = 0;
1140          goto TAIL_RECURSE;
1141          }
1142        /* Control never gets here */
1143    
1144      /* 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
1145      bracketed group and go to there. */      bracketed group and go to there. */
# Line 985  for (;;) Line 1148  for (;;)
1148      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1149      break;      break;
1150    
1151      /* BRAZERO and BRAMINZERO occur just before a bracket group, indicating      /* BRAZERO, BRAMINZERO and SKIPZERO occur just before a bracket group,
1152      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
1153      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
1154      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
1155      preceded by BRAZERO or BRAMINZERO. */      optional ones preceded by BRAZERO or BRAMINZERO. */
1156    
1157      case OP_BRAZERO:      case OP_BRAZERO:
1158        {        {
1159        next = ecode+1;        next = ecode+1;
1160        RMATCH(rrc, eptr, next, offset_top, md, ims, eptrb, match_isgroup);        RMATCH(eptr, next, offset_top, md, ims, eptrb, 0, RM10);
1161        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1162        do next += GET(next,1); while (*next == OP_ALT);        do next += GET(next,1); while (*next == OP_ALT);
1163        ecode = next + 1+LINK_SIZE;        ecode = next + 1 + LINK_SIZE;
1164        }        }
1165      break;      break;
1166    
1167      case OP_BRAMINZERO:      case OP_BRAMINZERO:
1168        {        {
1169        next = ecode+1;        next = ecode+1;
1170        do next += GET(next,1); while (*next == OP_ALT);        do next += GET(next, 1); while (*next == OP_ALT);
1171        RMATCH(rrc, eptr, next + 1+LINK_SIZE, offset_top, md, ims, eptrb,        RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0, RM11);
         match_isgroup);  
1172        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1173        ecode++;        ecode++;
1174        }        }
1175      break;      break;
1176    
1177      /* End of a group, repeated or non-repeating. If we are at the end of      case OP_SKIPZERO:
1178      an assertion "group", stop matching and return MATCH_MATCH, but record the        {
1179      current high water mark for use by positive assertions. Do this also        next = ecode+1;
1180      for the "once" (not-backup up) groups. */        do next += GET(next,1); while (*next == OP_ALT);
1181          ecode = next + 1 + LINK_SIZE;
1182          }
1183        break;
1184    
1185        /* End of a group, repeated or non-repeating. */
1186    
1187      case OP_KET:      case OP_KET:
1188      case OP_KETRMIN:      case OP_KETRMIN:
1189      case OP_KETRMAX:      case OP_KETRMAX:
1190        {      prev = ecode - GET(ecode, 1);
       prev = ecode - GET(ecode, 1);  
       saved_eptr = eptrb->epb_saved_eptr;  
1191    
1192        /* Back up the stack of bracket start pointers. */      /* If this was a group that remembered the subject start, in order to break
1193        infinite repeats of empty string matches, retrieve the subject start from
1194        the chain. Otherwise, set it NULL. */
1195    
1196        eptrb = eptrb->epb_prev;      if (*prev >= OP_SBRA)
1197          {
1198        if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||        saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */
1199            *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||        eptrb = eptrb->epb_prev;              /* Backup to previous group */
1200            *prev == OP_ONCE)        }
1201          {      else saved_eptr = NULL;
         md->end_match_ptr = eptr;      /* For ONCE */  
         md->end_offset_top = offset_top;  
         RRETURN(MATCH_MATCH);  
         }  
1202    
1203        /* 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
1204        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
1205        extraction by setting the offsets and bumping the high water mark. */      assertions. Do this also for the "once" (atomic) groups. */
1206    
1207        if (*prev != OP_COND)      if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||
1208          {          *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||
1209          number = *prev - OP_BRA;          *prev == OP_ONCE)
1210          {
1211          md->end_match_ptr = eptr;      /* For ONCE */
1212          md->end_offset_top = offset_top;
1213          RRETURN(MATCH_MATCH);
1214          }
1215    
1216          /* 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
1217          the number from a dummy opcode at the start. */      and if necessary complete handling an extraction by setting the offsets and
1218        bumping the high water mark. Note that whole-pattern recursion is coded as
1219        a recurse into group 0, so it won't be picked up here. Instead, we catch it
1220        when the OP_END is reached. Other recursion is handled here. */
1221    
1222          if (number > EXTRACT_BASIC_MAX) number = GET2(prev, 2+LINK_SIZE);      if (*prev == OP_CBRA || *prev == OP_SCBRA)
1223          offset = number << 1;        {
1224          number = GET2(prev, 1+LINK_SIZE);
1225          offset = number << 1;
1226    
1227  #ifdef DEBUG  #ifdef DEBUG
1228          printf("end bracket %d", number);        printf("end bracket %d", number);
1229          printf("\n");        printf("\n");
1230  #endif  #endif
1231    
1232          /* Test for a numbered group. This includes groups called as a result        md->capture_last = number;
1233          of recursion. Note that whole-pattern recursion is coded as a recurse        if (offset >= md->offset_max) md->offset_overflow = TRUE; else
1234          into group 0, so it won't be picked up here. Instead, we catch it when          {
1235          the OP_END is reached. */          md->offset_vector[offset] =
1236              md->offset_vector[md->offset_end - number];
1237          if (number > 0)          md->offset_vector[offset+1] = eptr - md->start_subject;
1238            {          if (offset_top <= offset) offset_top = offset + 2;
1239            md->capture_last = number;          }
1240            if (offset >= md->offset_max) md->offset_overflow = TRUE; else  
1241              {        /* Handle a recursively called group. Restore the offsets
1242              md->offset_vector[offset] =        appropriately and continue from after the call. */
1243                md->offset_vector[md->offset_end - number];  
1244              md->offset_vector[offset+1] = eptr - md->start_subject;        if (md->recursive != NULL && md->recursive->group_num == number)
1245              if (offset_top <= offset) offset_top = offset + 2;          {
1246              }          recursion_info *rec = md->recursive;
1247            DPRINTF(("Recursion (%d) succeeded - continuing\n", number));
1248            /* Handle a recursively called group. Restore the offsets          md->recursive = rec->prevrec;
1249            appropriately and continue from after the call. */          mstart = rec->save_start;
1250            memcpy(md->offset_vector, rec->offset_save,
1251            if (md->recursive != NULL && md->recursive->group_num == number)            rec->saved_max * sizeof(int));
1252              {          ecode = rec->after_call;
1253              recursion_info *rec = md->recursive;          ims = original_ims;
1254              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;  
             }  
           }  
1255          }          }
1256          }
1257    
1258        /* 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
1259        the group. */      flags, in case they got changed during the group. */
1260    
1261        ims = original_ims;      ims = original_ims;
1262        DPRINTF(("ims reset to %02lx\n", ims));      DPRINTF(("ims reset to %02lx\n", ims));
1263    
1264        /* For a non-repeating ket, just continue at this level. This also      /* For a non-repeating ket, just continue at this level. This also
1265        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.
1266        This is the forcible breaking of infinite loops as implemented in Perl      This is the forcible breaking of infinite loops as implemented in Perl
1267        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
1268        course of events. */      course of events. */
1269    
1270        if (*ecode == OP_KET || eptr == saved_eptr)      if (*ecode == OP_KET || eptr == saved_eptr)
1271          {        {
1272          ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1273          break;        break;
1274          }        }
1275    
1276        /* 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
1277        preceding bracket, in the appropriate order. */      preceding bracket, in the appropriate order. In the second case, we can use
1278        tail recursion to avoid using another stack frame, unless we have an
1279        unlimited repeat of a group that can match an empty string. */
1280    
1281        if (*ecode == OP_KETRMIN)      flags = (*prev >= OP_SBRA)? match_cbegroup : 0;
1282          {  
1283          RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0);      if (*ecode == OP_KETRMIN)
1284          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        {
1285          RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM12);
1286          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1287          }        if (flags != 0)    /* Could match an empty string */
       else  /* OP_KETRMAX */  
1288          {          {
1289          RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);          RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM50);
1290          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);  
1291          }          }
1292          ecode = prev;
1293          goto TAIL_RECURSE;
1294        }        }
1295        else  /* OP_KETRMAX */
1296      RRETURN(MATCH_NOMATCH);        {
1297          RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13);
1298          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1299          ecode += 1 + LINK_SIZE;
1300          flags = 0;
1301          goto TAIL_RECURSE;
1302          }
1303        /* Control never gets here */
1304    
1305      /* Start of subject unless notbol, or after internal newline if multiline */      /* Start of subject unless notbol, or after internal newline if multiline */
1306    
# Line 1135  for (;;) Line 1308  for (;;)
1308      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
1309      if ((ims & PCRE_MULTILINE) != 0)      if ((ims & PCRE_MULTILINE) != 0)
1310        {        {
1311        if (eptr != md->start_subject && eptr[-1] != NEWLINE)        if (eptr != md->start_subject &&
1312              (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
1313          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
1314        ecode++;        ecode++;
1315        break;        break;
# Line 1156  for (;;) Line 1330  for (;;)
1330      ecode++;      ecode++;
1331      break;      break;
1332    
1333        /* Reset the start of match point */
1334    
1335        case OP_SET_SOM:
1336        mstart = eptr;
1337        ecode++;
1338        break;
1339    
1340      /* Assert before internal newline if multiline, or before a terminating      /* Assert before internal newline if multiline, or before a terminating
1341      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. */
1342    
# Line 1163  for (;;) Line 1344  for (;;)
1344      if ((ims & PCRE_MULTILINE) != 0)      if ((ims & PCRE_MULTILINE) != 0)
1345        {        {
1346        if (eptr < md->end_subject)        if (eptr < md->end_subject)
1347          { if (*eptr != NEWLINE) RRETURN(MATCH_NOMATCH); }          { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }
1348        else        else
1349          { if (md->noteol) RRETURN(MATCH_NOMATCH); }          { if (md->noteol) RRETURN(MATCH_NOMATCH); }
1350        ecode++;        ecode++;
# Line 1174  for (;;) Line 1355  for (;;)
1355        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
1356        if (!md->endonly)        if (!md->endonly)
1357          {          {
1358          if (eptr < md->end_subject - 1 ||          if (eptr != md->end_subject &&
1359             (eptr == md->end_subject - 1 && *eptr != NEWLINE))              (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
1360            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1361          ecode++;          ecode++;
1362          break;          break;
1363          }          }
1364        }        }
1365      /* ... else fall through */      /* ... else fall through for endonly */
1366    
1367      /* End of subject assertion (\z) */      /* End of subject assertion (\z) */
1368    
# Line 1193  for (;;) Line 1374  for (;;)
1374      /* End of subject or ending \n assertion (\Z) */      /* End of subject or ending \n assertion (\Z) */
1375    
1376      case OP_EODN:      case OP_EODN:
1377      if (eptr < md->end_subject - 1 ||      if (eptr != md->end_subject &&
1378         (eptr == md->end_subject - 1 && *eptr != NEWLINE)) RRETURN(MATCH_NOMATCH);          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
1379          RRETURN(MATCH_NOMATCH);
1380      ecode++;      ecode++;
1381      break;      break;
1382    
# Line 1247  for (;;) Line 1429  for (;;)
1429      /* Match a single character type; inline for speed */      /* Match a single character type; inline for speed */
1430    
1431      case OP_ANY:      case OP_ANY:
1432      if ((ims & PCRE_DOTALL) == 0 && eptr < md->end_subject && *eptr == NEWLINE)      if ((ims & PCRE_DOTALL) == 0)
1433        RRETURN(MATCH_NOMATCH);        {
1434          if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
1435          }
1436      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);
 #ifdef SUPPORT_UTF8  
1437      if (utf8)      if (utf8)
1438        while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;        while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
 #endif  
1439      ecode++;      ecode++;
1440      break;      break;
1441    
# Line 1343  for (;;) Line 1525  for (;;)
1525      ecode++;      ecode++;
1526      break;      break;
1527    
1528        case OP_ANYNL:
1529        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1530        GETCHARINCTEST(c, eptr);
1531        switch(c)
1532          {
1533          default: RRETURN(MATCH_NOMATCH);
1534          case 0x000d:
1535          if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
1536          break;
1537    
1538          case 0x000a:
1539          break;
1540    
1541          case 0x000b:
1542          case 0x000c:
1543          case 0x0085:
1544          case 0x2028:
1545          case 0x2029:
1546          if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
1547          break;
1548          }
1549        ecode++;
1550        break;
1551    
1552        case OP_NOT_HSPACE:
1553        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1554        GETCHARINCTEST(c, eptr);
1555        switch(c)
1556          {
1557          default: break;
1558          case 0x09:      /* HT */
1559          case 0x20:      /* SPACE */
1560          case 0xa0:      /* NBSP */
1561          case 0x1680:    /* OGHAM SPACE MARK */
1562          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1563          case 0x2000:    /* EN QUAD */
1564          case 0x2001:    /* EM QUAD */
1565          case 0x2002:    /* EN SPACE */
1566          case 0x2003:    /* EM SPACE */
1567          case 0x2004:    /* THREE-PER-EM SPACE */
1568          case 0x2005:    /* FOUR-PER-EM SPACE */
1569          case 0x2006:    /* SIX-PER-EM SPACE */
1570          case 0x2007:    /* FIGURE SPACE */
1571          case 0x2008:    /* PUNCTUATION SPACE */
1572          case 0x2009:    /* THIN SPACE */
1573          case 0x200A:    /* HAIR SPACE */
1574          case 0x202f:    /* NARROW NO-BREAK SPACE */
1575          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1576          case 0x3000:    /* IDEOGRAPHIC SPACE */
1577          RRETURN(MATCH_NOMATCH);
1578          }
1579        ecode++;
1580        break;
1581    
1582        case OP_HSPACE:
1583        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1584        GETCHARINCTEST(c, eptr);
1585        switch(c)
1586          {
1587          default: RRETURN(MATCH_NOMATCH);
1588          case 0x09:      /* HT */
1589          case 0x20:      /* SPACE */
1590          case 0xa0:      /* NBSP */
1591          case 0x1680:    /* OGHAM SPACE MARK */
1592          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1593          case 0x2000:    /* EN QUAD */
1594          case 0x2001:    /* EM QUAD */
1595          case 0x2002:    /* EN SPACE */
1596          case 0x2003:    /* EM SPACE */
1597          case 0x2004:    /* THREE-PER-EM SPACE */
1598          case 0x2005:    /* FOUR-PER-EM SPACE */
1599          case 0x2006:    /* SIX-PER-EM SPACE */
1600          case 0x2007:    /* FIGURE SPACE */
1601          case 0x2008:    /* PUNCTUATION SPACE */
1602          case 0x2009:    /* THIN SPACE */
1603          case 0x200A:    /* HAIR SPACE */
1604          case 0x202f:    /* NARROW NO-BREAK SPACE */
1605          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1606          case 0x3000:    /* IDEOGRAPHIC SPACE */
1607          break;
1608          }
1609        ecode++;
1610        break;
1611    
1612        case OP_NOT_VSPACE:
1613        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1614        GETCHARINCTEST(c, eptr);
1615        switch(c)
1616          {
1617          default: break;
1618          case 0x0a:      /* LF */
1619          case 0x0b:      /* VT */
1620          case 0x0c:      /* FF */
1621          case 0x0d:      /* CR */
1622          case 0x85:      /* NEL */
1623          case 0x2028:    /* LINE SEPARATOR */
1624          case 0x2029:    /* PARAGRAPH SEPARATOR */
1625          RRETURN(MATCH_NOMATCH);
1626          }
1627        ecode++;
1628        break;
1629    
1630        case OP_VSPACE:
1631        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1632        GETCHARINCTEST(c, eptr);
1633        switch(c)
1634          {
1635          default: RRETURN(MATCH_NOMATCH);
1636          case 0x0a:      /* LF */
1637          case 0x0b:      /* VT */
1638          case 0x0c:      /* FF */
1639          case 0x0d:      /* CR */
1640          case 0x85:      /* NEL */
1641          case 0x2028:    /* LINE SEPARATOR */
1642          case 0x2029:    /* PARAGRAPH SEPARATOR */
1643          break;
1644          }
1645        ecode++;
1646        break;
1647    
1648  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
1649      /* Check the next character by Unicode property. We will get here only      /* Check the next character by Unicode property. We will get here only
1650      if the support is in the binary; otherwise a compile-time error occurs. */      if the support is in the binary; otherwise a compile-time error occurs. */
# Line 1352  for (;;) Line 1654  for (;;)
1654      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1655      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1656        {        {
1657        int chartype, rqdtype;        int chartype, script;
1658        int othercase;        int category = _pcre_ucp_findprop(c, &chartype, &script);
       int category = _pcre_ucp_findchar(c, &chartype, &othercase);  
   
       rqdtype = *(++ecode);  
       ecode++;  
1659    
1660        if (rqdtype >= 128)        switch(ecode[1])
1661          {          {
1662          if ((rqdtype - 128 != category) == (op == OP_PROP))          case PT_ANY:
1663            if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
1664            break;
1665    
1666            case PT_LAMP:
1667            if ((chartype == ucp_Lu ||
1668                 chartype == ucp_Ll ||
1669                 chartype == ucp_Lt) == (op == OP_NOTPROP))
1670            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1671          }           break;
1672        else  
1673          {          case PT_GC:
1674          if ((rqdtype != chartype) == (op == OP_PROP))          if ((ecode[2] != category) == (op == OP_PROP))
1675              RRETURN(MATCH_NOMATCH);
1676            break;
1677    
1678            case PT_PC:
1679            if ((ecode[2] != chartype) == (op == OP_PROP))
1680            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1681            break;
1682    
1683            case PT_SC:
1684            if ((ecode[2] != script) == (op == OP_PROP))
1685              RRETURN(MATCH_NOMATCH);
1686            break;
1687    
1688            default:
1689            RRETURN(PCRE_ERROR_INTERNAL);
1690          }          }
1691    
1692          ecode += 3;
1693        }        }
1694      break;      break;
1695    
# Line 1379  for (;;) Line 1700  for (;;)
1700      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1701      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1702        {        {
1703        int chartype;        int chartype, script;
1704        int othercase;        int category = _pcre_ucp_findprop(c, &chartype, &script);
       int category = _pcre_ucp_findchar(c, &chartype, &othercase);  
1705        if (category == ucp_M) RRETURN(MATCH_NOMATCH);        if (category == ucp_M) RRETURN(MATCH_NOMATCH);
1706        while (eptr < md->end_subject)        while (eptr < md->end_subject)
1707          {          {
# Line 1390  for (;;) Line 1710  for (;;)
1710            {            {
1711            GETCHARLEN(c, eptr, len);            GETCHARLEN(c, eptr, len);
1712            }            }
1713          category = _pcre_ucp_findchar(c, &chartype, &othercase);          category = _pcre_ucp_findprop(c, &chartype, &script);
1714          if (category != ucp_M) break;          if (category != ucp_M) break;
1715          eptr += len;          eptr += len;
1716          }          }
# Line 1411  for (;;) Line 1731  for (;;)
1731      case OP_REF:      case OP_REF:
1732        {        {
1733        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
1734        ecode += 3;                                 /* Advance past item */        ecode += 3;
1735    
1736        /* If the reference is unset, set the length to be longer than the amount        /* If the reference is unset, there are two possibilities:
1737        of subject left; this ensures that every attempt at a match fails. We  
1738        can't just fail here, because of the possibility of quantifiers with zero        (a) In the default, Perl-compatible state, set the length to be longer
1739        minima. */        than the amount of subject left; this ensures that every attempt at a
1740          match fails. We can't just fail here, because of the possibility of
1741        length = (offset >= offset_top || md->offset_vector[offset] < 0)?        quantifiers with zero minima.
1742          md->end_subject - eptr + 1 :  
1743          md->offset_vector[offset+1] - md->offset_vector[offset];        (b) If the JavaScript compatibility flag is set, set the length to zero
1744          so that the back reference matches an empty string.
1745    
1746          Otherwise, set the length to the length of what was matched by the
1747          referenced subpattern. */
1748    
1749          if (offset >= offset_top || md->offset_vector[offset] < 0)
1750            length = (md->jscript_compat)? 0 : md->end_subject - eptr + 1;
1751          else
1752            length = md->offset_vector[offset+1] - md->offset_vector[offset];
1753    
1754        /* Set up for repetition, or handle the non-repeated case */        /* Set up for repetition, or handle the non-repeated case */
1755    
# Line 1480  for (;;) Line 1809  for (;;)
1809          {          {
1810          for (fi = min;; fi++)          for (fi = min;; fi++)
1811            {            {
1812            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM14);
1813            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1814            if (fi >= max || !match_ref(offset, eptr, length, md, ims))            if (fi >= max || !match_ref(offset, eptr, length, md, ims))
1815              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
# Line 1501  for (;;) Line 1830  for (;;)
1830            }            }
1831          while (eptr >= pp)          while (eptr >= pp)
1832            {            {
1833            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM15);
1834            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1835            eptr -= length;            eptr -= length;
1836            }            }
# Line 1606  for (;;) Line 1935  for (;;)
1935            {            {
1936            for (fi = min;; fi++)            for (fi = min;; fi++)
1937              {              {
1938              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM16);
1939              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1940              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1941              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 1626  for (;;) Line 1955  for (;;)
1955            {            {
1956            for (fi = min;; fi++)            for (fi = min;; fi++)
1957              {              {
1958              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM17);
1959              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1960              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1961              c = *eptr++;              c = *eptr++;
# Line 1663  for (;;) Line 1992  for (;;)
1992              }              }
1993            for (;;)            for (;;)
1994              {              {
1995              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM18);
1996              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1997              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
1998              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 1682  for (;;) Line 2011  for (;;)
2011              }              }
2012            while (eptr >= pp)            while (eptr >= pp)
2013              {              {
2014              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM19);
             eptr--;  
2015              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2016                eptr--;
2017              }              }
2018            }            }
2019    
# Line 1753  for (;;) Line 2082  for (;;)
2082          {          {
2083          for (fi = min;; fi++)          for (fi = min;; fi++)
2084            {            {
2085            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20);
2086            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2087            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2088            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
# Line 1777  for (;;) Line 2106  for (;;)
2106            }            }
2107          for(;;)          for(;;)
2108            {            {
2109            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);
2110            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2111            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
2112            BACKCHAR(eptr)            if (utf8) BACKCHAR(eptr);
2113            }            }
2114          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2115          }          }
# Line 1836  for (;;) Line 2165  for (;;)
2165    
2166        else        else
2167          {          {
2168          int dc;          unsigned int dc;
2169          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
2170          ecode += length;          ecode += length;
2171    
2172          /* 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
2173          case of the character, if there is one. The result of _pcre_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. */  
2174    
2175          if (fc != dc)          if (fc != dc)
2176            {            {
2177  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2178            int chartype;            if (dc != _pcre_ucp_othercase(fc))
           int othercase;  
           if (_pcre_ucp_findchar(fc, &chartype, &othercase) < 0 || dc != othercase)  
2179  #endif  #endif
2180              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2181            }            }
# Line 1867  for (;;) Line 2192  for (;;)
2192        }        }
2193      break;      break;
2194    
2195      /* Match a single character repeatedly; different opcodes share code. */      /* Match a single character repeatedly. */
2196    
2197      case OP_EXACT:      case OP_EXACT:
2198      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
2199      ecode += 3;      ecode += 3;
2200      goto REPEATCHAR;      goto REPEATCHAR;
2201    
2202        case OP_POSUPTO:
2203        possessive = TRUE;
2204        /* Fall through */
2205    
2206      case OP_UPTO:      case OP_UPTO:
2207      case OP_MINUPTO:      case OP_MINUPTO:
2208      min = 0;      min = 0;
# Line 1882  for (;;) Line 2211  for (;;)
2211      ecode += 3;      ecode += 3;
2212      goto REPEATCHAR;      goto REPEATCHAR;
2213    
2214        case OP_POSSTAR:
2215        possessive = TRUE;
2216        min = 0;
2217        max = INT_MAX;
2218        ecode++;
2219        goto REPEATCHAR;
2220    
2221        case OP_POSPLUS:
2222        possessive = TRUE;
2223        min = 1;
2224        max = INT_MAX;
2225        ecode++;
2226        goto REPEATCHAR;
2227    
2228        case OP_POSQUERY:
2229        possessive = TRUE;
2230        min = 0;
2231        max = 1;
2232        ecode++;
2233        goto REPEATCHAR;
2234    
2235      case OP_STAR:      case OP_STAR:
2236      case OP_MINSTAR:      case OP_MINSTAR:
2237      case OP_PLUS:      case OP_PLUS:
# Line 1913  for (;;) Line 2263  for (;;)
2263    
2264        if (length > 1)        if (length > 1)
2265          {          {
         int oclength = 0;  
         uschar occhars[8];  
   
2266  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2267          int othercase;          unsigned int othercase;
         int chartype;  
2268          if ((ims & PCRE_CASELESS) != 0 &&          if ((ims & PCRE_CASELESS) != 0 &&
2269               _pcre_ucp_findchar(fc, &chartype, &othercase) >= 0 &&              (othercase = _pcre_ucp_othercase(fc)) != NOTACHAR)
              othercase > 0)  
2270            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = _pcre_ord2utf8(othercase, occhars);
2271            else oclength = 0;
2272  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2273    
2274          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2275            {            {
2276            if (memcmp(eptr, charptr, length) == 0) eptr += length;            if (memcmp(eptr, charptr, length) == 0) eptr += length;
2277    #ifdef SUPPORT_UCP
2278            /* Need braces because of following else */            /* Need braces because of following else */
2279            else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }            else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }
2280            else            else
# Line 1935  for (;;) Line 2282  for (;;)
2282              if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);              if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);
2283              eptr += oclength;              eptr += oclength;
2284              }              }
2285    #else   /* without SUPPORT_UCP */
2286              else { RRETURN(MATCH_NOMATCH); }
2287    #endif  /* SUPPORT_UCP */
2288            }            }
2289    
2290          if (min == max) continue;          if (min == max) continue;
# Line 1943  for (;;) Line 2293  for (;;)
2293            {            {
2294            for (fi = min;; fi++)            for (fi = min;; fi++)
2295              {              {
2296              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM22);
2297              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2298              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2299              if (memcmp(eptr, charptr, length) == 0) eptr += length;              if (memcmp(eptr, charptr, length) == 0) eptr += length;
2300    #ifdef SUPPORT_UCP
2301              /* Need braces because of following else */              /* Need braces because of following else */
2302              else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }              else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }
2303              else              else
# Line 1954  for (;;) Line 2305  for (;;)
2305                if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);                if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);
2306                eptr += oclength;                eptr += oclength;
2307                }                }
2308    #else   /* without SUPPORT_UCP */
2309                else { RRETURN (MATCH_NOMATCH); }
2310    #endif  /* SUPPORT_UCP */
2311              }              }
2312            /* Control never gets here */            /* Control never gets here */
2313            }            }
2314          else  
2315            else  /* Maximize */
2316            {            {
2317            pp = eptr;            pp = eptr;
2318            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2319              {              {
2320              if (eptr > md->end_subject - length) break;              if (eptr > md->end_subject - length) break;
2321              if (memcmp(eptr, charptr, length) == 0) eptr += length;              if (memcmp(eptr, charptr, length) == 0) eptr += length;
2322    #ifdef SUPPORT_UCP
2323              else if (oclength == 0) break;              else if (oclength == 0) break;
2324              else              else
2325                {                {
2326                if (memcmp(eptr, occhars, oclength) != 0) break;                if (memcmp(eptr, occhars, oclength) != 0) break;
2327                eptr += oclength;                eptr += oclength;
2328                }                }
2329    #else   /* without SUPPORT_UCP */
2330                else break;
2331    #endif  /* SUPPORT_UCP */
2332              }              }
2333            while (eptr >= pp)  
2334              if (possessive) continue;
2335              for(;;)
2336             {             {
2337             RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);             RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM23);
2338             if (rrc != MATCH_NOMATCH) RRETURN(rrc);             if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2339               if (eptr == pp) RRETURN(MATCH_NOMATCH);
2340    #ifdef SUPPORT_UCP
2341               eptr--;
2342               BACKCHAR(eptr);
2343    #else   /* without SUPPORT_UCP */
2344             eptr -= length;             eptr -= length;
2345    #endif  /* SUPPORT_UCP */
2346             }             }
           RRETURN(MATCH_NOMATCH);  
2347            }            }
2348          /* Control never gets here */          /* Control never gets here */
2349          }          }
# Line 2017  for (;;) Line 2383  for (;;)
2383          {          {
2384          for (fi = min;; fi++)          for (fi = min;; fi++)
2385            {            {
2386            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM24);
2387            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2388            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
2389                fc != md->lcc[*eptr++])                fc != md->lcc[*eptr++])
# Line 2025  for (;;) Line 2391  for (;;)
2391            }            }
2392          /* Control never gets here */          /* Control never gets here */
2393          }          }
2394        else        else  /* Maximize */
2395          {          {
2396          pp = eptr;          pp = eptr;
2397          for (i = min; i < max; i++)          for (i = min; i < max; i++)
# Line 2033  for (;;) Line 2399  for (;;)
2399            if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break;            if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break;
2400            eptr++;            eptr++;
2401            }            }
2402            if (possessive) continue;
2403          while (eptr >= pp)          while (eptr >= pp)
2404            {            {
2405            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM25);
2406            eptr--;            eptr--;
2407            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2408            }            }
# Line 2054  for (;;) Line 2421  for (;;)
2421          {          {
2422          for (fi = min;; fi++)          for (fi = min;; fi++)
2423            {            {
2424            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM26);
2425            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2426            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)
2427              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2428            }            }
2429          /* Control never gets here */          /* Control never gets here */
2430          }          }
2431        else        else  /* Maximize */
2432          {          {
2433          pp = eptr;          pp = eptr;
2434          for (i = min; i < max; i++)          for (i = min; i < max; i++)
# Line 2069  for (;;) Line 2436  for (;;)
2436            if (eptr >= md->end_subject || fc != *eptr) break;            if (eptr >= md->end_subject || fc != *eptr) break;
2437            eptr++;            eptr++;
2438            }            }
2439            if (possessive) continue;
2440          while (eptr >= pp)          while (eptr >= pp)
2441            {            {
2442            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27);
2443            eptr--;            eptr--;
2444            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2445            }            }
# Line 2121  for (;;) Line 2489  for (;;)
2489      ecode += 3;      ecode += 3;
2490      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
2491    
2492        case OP_NOTPOSSTAR:
2493        possessive = TRUE;
2494        min = 0;
2495        max = INT_MAX;
2496        ecode++;
2497        goto REPEATNOTCHAR;
2498    
2499        case OP_NOTPOSPLUS:
2500        possessive = TRUE;
2501        min = 1;
2502        max = INT_MAX;
2503        ecode++;
2504        goto REPEATNOTCHAR;
2505    
2506        case OP_NOTPOSQUERY:
2507        possessive = TRUE;
2508        min = 0;
2509        max = 1;
2510        ecode++;
2511        goto REPEATNOTCHAR;
2512    
2513        case OP_NOTPOSUPTO:
2514        possessive = TRUE;
2515        min = 0;
2516        max = GET2(ecode, 1);
2517        ecode += 3;
2518        goto REPEATNOTCHAR;
2519    
2520      case OP_NOTSTAR:      case OP_NOTSTAR:
2521      case OP_NOTMINSTAR:      case OP_NOTMINSTAR:
2522      case OP_NOTPLUS:      case OP_NOTPLUS:
# Line 2160  for (;;) Line 2556  for (;;)
2556        /* UTF-8 mode */        /* UTF-8 mode */
2557        if (utf8)        if (utf8)
2558          {          {
2559          register int d;          register unsigned int d;
2560          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2561            {            {
2562            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
# Line 2185  for (;;) Line 2581  for (;;)
2581          /* UTF-8 mode */          /* UTF-8 mode */
2582          if (utf8)          if (utf8)
2583            {            {
2584            register int d;            register unsigned int d;
2585            for (fi = min;; fi++)            for (fi = min;; fi++)
2586              {              {
2587              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28);
2588              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2589              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2590              if (d < 256) d = md->lcc[d];              if (d < 256) d = md->lcc[d];
# Line 2202  for (;;) Line 2598  for (;;)
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, RM29);
2602              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2603              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])
2604                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2221  for (;;) Line 2617  for (;;)
2617          /* UTF-8 mode */          /* UTF-8 mode */
2618          if (utf8)          if (utf8)
2619            {            {
2620            register int d;            register unsigned int d;
2621            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2622              {              {
2623              int len = 1;              int len = 1;
# Line 2231  for (;;) Line 2627  for (;;)
2627              if (fc == d) break;              if (fc == d) break;
2628              eptr += len;              eptr += len;
2629              }              }
2630            for(;;)          if (possessive) continue;
2631            for(;;)
2632              {              {
2633              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM30);
2634              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2635              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2636              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2248  for (;;) Line 2645  for (;;)
2645              if (eptr >= md->end_subject || fc == md->lcc[*eptr]) break;              if (eptr >= md->end_subject || fc == md->lcc[*eptr]) break;
2646              eptr++;              eptr++;
2647              }              }
2648              if (possessive) continue;
2649            while (eptr >= pp)            while (eptr >= pp)
2650              {              {
2651              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM31);
2652              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2653              eptr--;              eptr--;
2654              }              }
# Line 2269  for (;;) Line 2667  for (;;)
2667        /* UTF-8 mode */        /* UTF-8 mode */
2668        if (utf8)        if (utf8)
2669          {          {
2670          register int d;          register unsigned int d;
2671          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2672            {            {
2673            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
# Line 2292  for (;;) Line 2690  for (;;)
2690          /* UTF-8 mode */          /* UTF-8 mode */
2691          if (utf8)          if (utf8)
2692            {            {
2693            register int d;            register unsigned int d;
2694            for (fi = min;; fi++)            for (fi = min;; fi++)
2695              {              {
2696              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32);
2697              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2698              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2699              if (fi >= max || eptr >= md->end_subject || fc == d)              if (fi >= max || eptr >= md->end_subject || fc == d)
# Line 2308  for (;;) Line 2706  for (;;)
2706            {            {
2707            for (fi = min;; fi++)            for (fi = min;; fi++)
2708              {              {
2709              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM33);
2710              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2711              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)
2712                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2327  for (;;) Line 2725  for (;;)
2725          /* UTF-8 mode */          /* UTF-8 mode */
2726          if (utf8)          if (utf8)
2727            {            {
2728            register int d;            register unsigned int d;
2729            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2730              {              {
2731              int len = 1;              int len = 1;
# Line 2336  for (;;) Line 2734  for (;;)
2734              if (fc == d) break;              if (fc == d) break;
2735              eptr += len;              eptr += len;
2736              }              }
2737              if (possessive) continue;
2738            for(;;)            for(;;)
2739              {              {
2740              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM34);
2741              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2742              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2743              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2353  for (;;) Line 2752  for (;;)
2752              if (eptr >= md->end_subject || fc == *eptr) break;              if (eptr >= md->end_subject || fc == *eptr) break;
2753              eptr++;              eptr++;
2754              }              }
2755              if (possessive) continue;
2756            while (eptr >= pp)            while (eptr >= pp)
2757              {              {
2758              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM35);
2759              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2760              eptr--;              eptr--;
2761              }              }
# Line 2384  for (;;) Line 2784  for (;;)
2784      ecode += 3;      ecode += 3;
2785      goto REPEATTYPE;      goto REPEATTYPE;
2786    
2787        case OP_TYPEPOSSTAR:
2788        possessive = TRUE;
2789        min = 0;
2790        max = INT_MAX;
2791        ecode++;
2792        goto REPEATTYPE;
2793    
2794        case OP_TYPEPOSPLUS:
2795        possessive = TRUE;
2796        min = 1;
2797        max = INT_MAX;
2798        ecode++;
2799        goto REPEATTYPE;
2800    
2801        case OP_TYPEPOSQUERY:
2802        possessive = TRUE;
2803        min = 0;
2804        max = 1;
2805        ecode++;
2806        goto REPEATTYPE;
2807    
2808        case OP_TYPEPOSUPTO:
2809        possessive = TRUE;
2810        min = 0;
2811        max = GET2(ecode, 1);
2812        ecode += 3;
2813        goto REPEATTYPE;
2814    
2815      case OP_TYPESTAR:      case OP_TYPESTAR:
2816      case OP_TYPEMINSTAR:      case OP_TYPEMINSTAR:
2817      case OP_TYPEPLUS:      case OP_TYPEPLUS:
# Line 2408  for (;;) Line 2836  for (;;)
2836        {        {
2837        prop_fail_result = ctype == OP_NOTPROP;        prop_fail_result = ctype == OP_NOTPROP;
2838        prop_type = *ecode++;        prop_type = *ecode++;
2839        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;  
         }  
2840        }        }
2841      else prop_type = -1;      else prop_type = -1;
2842  #endif  #endif
# Line 2434  for (;;) Line 2853  for (;;)
2853      if (min > 0)      if (min > 0)
2854        {        {
2855  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2856        if (prop_type > 0)        if (prop_type >= 0)
2857          {          {
2858          for (i = 1; i <= min; i++)          switch(prop_type)
2859            {            {
2860            GETCHARINC(c, eptr);            case PT_ANY:
2861            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);            if (prop_fail_result) RRETURN(MATCH_NOMATCH);
2862            if ((*prop_test_variable == prop_test_against) == prop_fail_result)            for (i = 1; i <= min; i++)
2863              RRETURN(MATCH_NOMATCH);              {
2864                if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2865                GETCHARINCTEST(c, eptr);
2866                }
2867              break;
2868    
2869              case PT_LAMP:
2870              for (i = 1; i <= min; i++)
2871                {
2872                if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2873                GETCHARINCTEST(c, eptr);
2874                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2875                if ((prop_chartype == ucp_Lu ||
2876                     prop_chartype == ucp_Ll ||
2877                     prop_chartype == ucp_Lt) == prop_fail_result)
2878                  RRETURN(MATCH_NOMATCH);
2879                }
2880              break;
2881    
2882              case PT_GC:
2883              for (i = 1; i <= min; i++)
2884                {
2885                if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2886                GETCHARINCTEST(c, eptr);
2887                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2888                if ((prop_category == prop_value) == prop_fail_result)
2889                  RRETURN(MATCH_NOMATCH);
2890                }
2891              break;
2892    
2893              case PT_PC:
2894              for (i = 1; i <= min; i++)
2895                {
2896                if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2897                GETCHARINCTEST(c, eptr);
2898                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2899                if ((prop_chartype == prop_value) == prop_fail_result)
2900                  RRETURN(MATCH_NOMATCH);
2901                }
2902              break;
2903    
2904              case PT_SC:
2905              for (i = 1; i <= min; i++)
2906                {
2907                if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2908                GETCHARINCTEST(c, eptr);
2909                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2910                if ((prop_script == prop_value) == prop_fail_result)
2911                  RRETURN(MATCH_NOMATCH);
2912                }
2913              break;
2914    
2915              default:
2916              RRETURN(PCRE_ERROR_INTERNAL);
2917            }            }
2918          }          }
2919    
# Line 2453  for (;;) Line 2925  for (;;)
2925          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2926            {            {
2927            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
2928            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2929            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);
2930            while (eptr < md->end_subject)            while (eptr < md->end_subject)
2931              {              {
# Line 2462  for (;;) Line 2934  for (;;)
2934                {                {
2935                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
2936                }                }
2937              prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2938              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
2939              eptr += len;              eptr += len;
2940              }              }
# Line 2481  for (;;) Line 2953  for (;;)
2953          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2954            {            {
2955            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
2956               (*eptr++ == NEWLINE && (ims & PCRE_DOTALL) == 0))                 ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))
2957              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2958              eptr++;
2959            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
2960            }            }
2961          break;          break;
# Line 2491  for (;;) Line 2964  for (;;)
2964          eptr += min;          eptr += min;
2965          break;          break;
2966    
2967            case OP_ANYNL:
2968            for (i = 1; i <= min; i++)
2969              {
2970              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2971              GETCHARINC(c, eptr);
2972              switch(c)
2973                {
2974                default: RRETURN(MATCH_NOMATCH);
2975                case 0x000d:
2976                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
2977                break;
2978    
2979                case 0x000a:
2980                break;
2981    
2982                case 0x000b:
2983                case 0x000c:
2984                case 0x0085:
2985                case 0x2028:
2986                case 0x2029:
2987                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2988                break;
2989                }
2990              }
2991            break;
2992    
2993            case OP_NOT_HSPACE:
2994            for (i = 1; i <= min; i++)
2995              {
2996              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2997              GETCHARINC(c, eptr);
2998              switch(c)
2999                {
3000                default: break;
3001                case 0x09:      /* HT */
3002                case 0x20:      /* SPACE */
3003                case 0xa0:      /* NBSP */
3004                case 0x1680:    /* OGHAM SPACE MARK */
3005                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3006                case 0x2000:    /* EN QUAD */
3007                case 0x2001:    /* EM QUAD */
3008                case 0x2002:    /* EN SPACE */
3009                case 0x2003:    /* EM SPACE */
3010                case 0x2004:    /* THREE-PER-EM SPACE */
3011                case 0x2005:    /* FOUR-PER-EM SPACE */
3012                case 0x2006:    /* SIX-PER-EM SPACE */
3013                case 0x2007:    /* FIGURE SPACE */
3014                case 0x2008:    /* PUNCTUATION SPACE */
3015                case 0x2009:    /* THIN SPACE */
3016                case 0x200A:    /* HAIR SPACE */
3017                case 0x202f:    /* NARROW NO-BREAK SPACE */
3018                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3019                case 0x3000:    /* IDEOGRAPHIC SPACE */
3020                RRETURN(MATCH_NOMATCH);
3021                }
3022              }
3023            break;
3024    
3025            case OP_HSPACE:
3026            for (i = 1; i <= min; i++)
3027              {
3028              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3029              GETCHARINC(c, eptr);
3030              switch(c)
3031                {
3032                default: RRETURN(MATCH_NOMATCH);
3033                case 0x09:      /* HT */
3034                case 0x20:      /* SPACE */
3035                case 0xa0:      /* NBSP */
3036                case 0x1680:    /* OGHAM SPACE MARK */
3037                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3038                case 0x2000:    /* EN QUAD */
3039                case 0x2001:    /* EM QUAD */
3040                case 0x2002:    /* EN SPACE */
3041                case 0x2003:    /* EM SPACE */
3042                case 0x2004:    /* THREE-PER-EM SPACE */
3043                case 0x2005:    /* FOUR-PER-EM SPACE */
3044                case 0x2006:    /* SIX-PER-EM SPACE */
3045                case 0x2007:    /* FIGURE SPACE */
3046                case 0x2008:    /* PUNCTUATION SPACE */
3047                case 0x2009:    /* THIN SPACE */
3048                case 0x200A:    /* HAIR SPACE */
3049                case 0x202f:    /* NARROW NO-BREAK SPACE */
3050                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3051                case 0x3000:    /* IDEOGRAPHIC SPACE */
3052                break;
3053                }
3054              }
3055            break;
3056    
3057            case OP_NOT_VSPACE:
3058            for (i = 1; i <= min; i++)
3059              {
3060              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3061              GETCHARINC(c, eptr);
3062              switch(c)
3063                {
3064                default: break;
3065                case 0x0a:      /* LF */
3066                case 0x0b:      /* VT */
3067                case 0x0c:      /* FF */
3068                case 0x0d:      /* CR */
3069                case 0x85:      /* NEL */
3070                case 0x2028:    /* LINE SEPARATOR */
3071                case 0x2029:    /* PARAGRAPH SEPARATOR */
3072                RRETURN(MATCH_NOMATCH);
3073                }
3074              }
3075            break;
3076    
3077            case OP_VSPACE:
3078            for (i = 1; i <= min; i++)
3079              {
3080              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3081              GETCHARINC(c, eptr);
3082              switch(c)
3083                {
3084                default: RRETURN(MATCH_NOMATCH);
3085                case 0x0a:      /* LF */
3086                case 0x0b:      /* VT */
3087                case 0x0c:      /* FF */
3088                case 0x0d:      /* CR */
3089                case 0x85:      /* NEL */
3090                case 0x2028:    /* LINE SEPARATOR */
3091                case 0x2029:    /* PARAGRAPH SEPARATOR */
3092                break;
3093                }
3094              }
3095            break;
3096    
3097          case OP_NOT_DIGIT:          case OP_NOT_DIGIT:
3098          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3099            {            {
# Line 2515  for (;;) Line 3118  for (;;)
3118          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3119            {            {
3120            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3121               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_space) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0))
3122              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3123            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3124            }            }
3125          break;          break;
3126    
# Line 2535  for (;;) Line 3138  for (;;)
3138          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3139            {            {
3140            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3141               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_word) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0))
3142              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3143            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3144            }            }
3145          break;          break;
3146    
# Line 2559  for (;;) Line 3162  for (;;)
3162  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF8 */
3163    
3164        /* Code for the non-UTF-8 case for minimum matching of operators other        /* Code for the non-UTF-8 case for minimum matching of operators other
3165        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. We can assume that there are the minimum
3166          number of bytes present, as this was tested above. */
3167    
3168        switch(ctype)        switch(ctype)
3169          {          {
# Line 2567  for (;;) Line 3171  for (;;)
3171          if ((ims & PCRE_DOTALL) == 0)          if ((ims & PCRE_DOTALL) == 0)
3172            {            {
3173            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3174              if (*eptr++ == NEWLINE) RRETURN(MATCH_NOMATCH);              {
3175                if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
3176                eptr++;
3177                }
3178            }            }
3179          else eptr += min;          else eptr += min;
3180          break;          break;
# Line 2576  for (;;) Line 3183  for (;;)
3183          eptr += min;          eptr += min;
3184          break;          break;
3185    
3186            /* Because of the CRLF case, we can't assume the minimum number of
3187            bytes are present in this case. */
3188    
3189            case OP_ANYNL:
3190            for (i = 1; i <= min; i++)
3191              {
3192              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3193              switch(*eptr++)
3194                {
3195                default: RRETURN(MATCH_NOMATCH);
3196                case 0x000d:
3197                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3198                break;
3199                case 0x000a:
3200                break;
3201    
3202                case 0x000b:
3203                case 0x000c:
3204                case 0x0085:
3205                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3206                break;
3207                }
3208              }
3209            break;
3210    
3211            case OP_NOT_HSPACE:
3212            for (i = 1; i <= min; i++)
3213              {
3214              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3215              switch(*eptr++)
3216                {
3217                default: break;
3218                case 0x09:      /* HT */
3219                case 0x20:      /* SPACE */
3220                case 0xa0:      /* NBSP */
3221                RRETURN(MATCH_NOMATCH);
3222                }
3223              }
3224            break;
3225    
3226            case OP_HSPACE:
3227            for (i = 1; i <= min; i++)
3228              {
3229              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3230              switch(*eptr++)
3231                {
3232                default: RRETURN(MATCH_NOMATCH);
3233                case 0x09:      /* HT */
3234                case 0x20:      /* SPACE */
3235                case 0xa0:      /* NBSP */
3236                break;
3237                }
3238              }
3239            break;
3240    
3241            case OP_NOT_VSPACE:
3242            for (i = 1; i <= min; i++)
3243              {
3244              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3245              switch(*eptr++)
3246                {
3247                default: break;
3248                case 0x0a:      /* LF */
3249                case 0x0b:      /* VT */
3250                case 0x0c:      /* FF */
3251                case 0x0d:      /* CR */
3252                case 0x85:      /* NEL */
3253                RRETURN(MATCH_NOMATCH);
3254                }
3255              }
3256            break;
3257    
3258            case OP_VSPACE:
3259            for (i = 1; i <= min; i++)
3260              {
3261              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3262              switch(*eptr++)
3263                {
3264                default: RRETURN(MATCH_NOMATCH);
3265                case 0x0a:      /* LF */
3266                case 0x0b:      /* VT */
3267                case 0x0c:      /* FF */
3268                case 0x0d:      /* CR */
3269                case 0x85:      /* NEL */
3270                break;
3271                }
3272              }
3273            break;
3274    
3275          case OP_NOT_DIGIT:          case OP_NOT_DIGIT:
3276          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3277            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
# Line 2624  for (;;) Line 3320  for (;;)
3320      if (minimize)      if (minimize)
3321        {        {
3322  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3323        if (prop_type > 0)        if (prop_type >= 0)
3324          {          {
3325          for (fi = min;; fi++)          switch(prop_type)
3326            {            {
3327            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            case PT_ANY:
3328            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            for (fi = min;; fi++)
3329            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              {
3330            GETCHARINC(c, eptr);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM36);
3331            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3332            if ((*prop_test_variable == prop_test_against) == prop_fail_result)              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3333              RRETURN(MATCH_NOMATCH);              GETCHARINC(c, eptr);
3334                if (prop_fail_result) RRETURN(MATCH_NOMATCH);
3335                }
3336              /* Control never gets here */
3337    
3338              case PT_LAMP:
3339              for (fi = min;; fi++)
3340                {
3341                RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM37);
3342                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3343                if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3344                GETCHARINC(c, eptr);
3345                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3346                if ((prop_chartype == ucp_Lu ||
3347                     prop_chartype == ucp_Ll ||
3348                     prop_chartype == ucp_Lt) == prop_fail_result)
3349                  RRETURN(MATCH_NOMATCH);
3350                }
3351              /* Control never gets here */
3352    
3353              case PT_GC:
3354              for (fi = min;; fi++)
3355                {
3356                RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM38);
3357                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3358                if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3359                GETCHARINC(c, eptr);
3360                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3361                if ((prop_category == prop_value) == prop_fail_result)
3362                  RRETURN(MATCH_NOMATCH);
3363                }
3364              /* Control never gets here */
3365    
3366              case PT_PC:
3367              for (fi = min;; fi++)
3368                {
3369                RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM39);
3370                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3371                if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3372                GETCHARINC(c, eptr);
3373                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3374                if ((prop_chartype == prop_value) == prop_fail_result)
3375                  RRETURN(MATCH_NOMATCH);
3376                }
3377              /* Control never gets here */
3378    
3379              case PT_SC:
3380              for (fi = min;; fi++)
3381                {
3382                RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM40);
3383                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3384                if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3385                GETCHARINC(c, eptr);
3386                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3387                if ((prop_script == prop_value) == prop_fail_result)
3388                  RRETURN(MATCH_NOMATCH);
3389                }
3390              /* Control never gets here */
3391    
3392              default:
3393              RRETURN(PCRE_ERROR_INTERNAL);
3394            }            }
3395          }          }
3396    
# Line 2645  for (;;) Line 3401  for (;;)
3401          {          {
3402          for (fi = min;; fi++)          for (fi = min;; fi++)
3403            {            {
3404            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM41);
3405            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3406            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3407            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3408            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3409            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);
3410            while (eptr < md->end_subject)            while (eptr < md->end_subject)
3411              {              {
# Line 2658  for (;;) Line 3414  for (;;)
3414                {                {
3415                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3416                }                }
3417              prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3418              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3419              eptr += len;              eptr += len;
3420              }              }
# Line 2674  for (;;) Line 3430  for (;;)
3430          {          {
3431          for (fi = min;; fi++)          for (fi = min;; fi++)
3432            {            {
3433            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42);
3434            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3435            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject ||
3436                   (ctype == OP_ANY && (ims & PCRE_DOTALL) == 0 &&
3437                    IS_NEWLINE(eptr)))
3438                RRETURN(MATCH_NOMATCH);
3439    
3440            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
3441            switch(ctype)            switch(ctype)
3442              {              {
3443              case OP_ANY:              case OP_ANY:        /* This is the DOTALL case */
             if ((ims & PCRE_DOTALL) == 0 && c == NEWLINE) RRETURN(MATCH_NOMATCH);  
3444              break;              break;
3445    
3446              case OP_ANYBYTE:              case OP_ANYBYTE:
3447              break;              break;
3448    
3449                case OP_ANYNL:
3450                switch(c)
3451                  {
3452                  default: RRETURN(MATCH_NOMATCH);
3453                  case 0x000d:
3454                  if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3455                  break;
3456                  case 0x000a:
3457                  break;
3458    
3459                  case 0x000b:
3460                  case 0x000c:
3461                  case 0x0085:
3462                  case 0x2028:
3463                  case 0x2029:
3464                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3465                  break;
3466                  }
3467                break;
3468    
3469                case OP_NOT_HSPACE:
3470                switch(c)
3471                  {
3472                  default: break;
3473                  case 0x09:      /* HT */
3474                  case 0x20:      /* SPACE */
3475                  case 0xa0:      /* NBSP */
3476                  case 0x1680:    /* OGHAM SPACE MARK */
3477                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3478                  case 0x2000:    /* EN QUAD */
3479                  case 0x2001:    /* EM QUAD */
3480                  case 0x2002:    /* EN SPACE */
3481                  case 0x2003:    /* EM SPACE */
3482                  case 0x2004:    /* THREE-PER-EM SPACE */
3483                  case 0x2005:    /* FOUR-PER-EM SPACE */
3484                  case 0x2006:    /* SIX-PER-EM SPACE */
3485                  case 0x2007:    /* FIGURE SPACE */
3486                  case 0x2008:    /* PUNCTUATION SPACE */
3487                  case 0x2009:    /* THIN SPACE */
3488                  case 0x200A:    /* HAIR SPACE */
3489                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3490                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3491                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3492                  RRETURN(MATCH_NOMATCH);
3493                  }
3494                break;
3495    
3496                case OP_HSPACE:
3497                switch(c)
3498                  {
3499                  default: RRETURN(MATCH_NOMATCH);
3500                  case 0x09:      /* HT */
3501                  case 0x20:      /* SPACE */
3502                  case 0xa0:      /* NBSP */
3503                  case 0x1680:    /* OGHAM SPACE MARK */
3504                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3505                  case 0x2000:    /* EN QUAD */
3506                  case 0x2001:    /* EM QUAD */
3507                  case 0x2002:    /* EN SPACE */
3508                  case 0x2003:    /* EM SPACE */
3509                  case 0x2004:    /* THREE-PER-EM SPACE */
3510                  case 0x2005:    /* FOUR-PER-EM SPACE */
3511                  case 0x2006:    /* SIX-PER-EM SPACE */
3512                  case 0x2007:    /* FIGURE SPACE */
3513                  case 0x2008:    /* PUNCTUATION SPACE */
3514                  case 0x2009:    /* THIN SPACE */
3515                  case 0x200A:    /* HAIR SPACE */
3516                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3517                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3518                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3519                  break;
3520                  }
3521                break;
3522    
3523                case OP_NOT_VSPACE:
3524                switch(c)
3525                  {
3526                  default: break;
3527                  case 0x0a:      /* LF */
3528                  case 0x0b:      /* VT */
3529                  case 0x0c:      /* FF */
3530                  case 0x0d:      /* CR */
3531                  case 0x85:      /* NEL */
3532                  case 0x2028:    /* LINE SEPARATOR */
3533                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3534                  RRETURN(MATCH_NOMATCH);
3535                  }
3536                break;
3537    
3538                case OP_VSPACE:
3539                switch(c)
3540                  {
3541                  default: RRETURN(MATCH_NOMATCH);
3542                  case 0x0a:      /* LF */
3543                  case 0x0b:      /* VT */
3544                  case 0x0c:      /* FF */
3545                  case 0x0d:      /* CR */
3546                  case 0x85:      /* NEL */
3547                  case 0x2028:    /* LINE SEPARATOR */
3548                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3549                  break;
3550                  }
3551                break;
3552    
3553              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
3554              if (c < 256 && (md->ctypes[c] & ctype_digit) != 0)              if (c < 256 && (md->ctypes[c] & ctype_digit) != 0)
3555                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2729  for (;;) Line 3591  for (;;)
3591          {          {
3592          for (fi = min;; fi++)          for (fi = min;; fi++)
3593            {            {
3594            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43);
3595            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3596            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject ||
3597                   ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))
3598                RRETURN(MATCH_NOMATCH);
3599    
3600            c = *eptr++;            c = *eptr++;
3601            switch(ctype)            switch(ctype)
3602              {              {
3603              case OP_ANY:              case OP_ANY:   /* This is the DOTALL case */
3604              if ((ims & PCRE_DOTALL) == 0 && c == NEWLINE) RRETURN(MATCH_NOMATCH);              break;
3605    
3606                case OP_ANYBYTE:
3607                break;
3608    
3609                case OP_ANYNL:
3610                switch(c)
3611                  {
3612                  default: RRETURN(MATCH_NOMATCH);
3613                  case 0x000d:
3614                  if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3615                  break;
3616    
3617                  case 0x000a:
3618                  break;
3619    
3620                  case 0x000b:
3621                  case 0x000c:
3622                  case 0x0085:
3623                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3624                  break;
3625                  }
3626              break;              break;
3627    
3628              case OP_ANYBYTE:              case OP_NOT_HSPACE:
3629                switch(c)
3630                  {
3631                  default: break;
3632                  case 0x09:      /* HT */
3633                  case 0x20:      /* SPACE */
3634                  case 0xa0:      /* NBSP */
3635                  RRETURN(MATCH_NOMATCH);
3636                  }
3637                break;
3638    
3639                case OP_HSPACE:
3640                switch(c)
3641                  {
3642                  default: RRETURN(MATCH_NOMATCH);
3643                  case 0x09:      /* HT */
3644                  case 0x20:      /* SPACE */
3645                  case 0xa0:      /* NBSP */
3646                  break;
3647                  }
3648                break;
3649    
3650                case OP_NOT_VSPACE:
3651                switch(c)
3652                  {
3653                  default: break;
3654                  case 0x0a:      /* LF */
3655                  case 0x0b:      /* VT */
3656                  case 0x0c:      /* FF */
3657                  case 0x0d:      /* CR */
3658                  case 0x85:      /* NEL */
3659                  RRETURN(MATCH_NOMATCH);
3660                  }
3661                break;
3662    
3663                case OP_VSPACE:
3664                switch(c)
3665                  {
3666                  default: RRETURN(MATCH_NOMATCH);
3667                  case 0x0a:      /* LF */
3668                  case 0x0b:      /* VT */
3669                  case 0x0c:      /* FF */
3670                  case 0x0d:      /* CR */
3671                  case 0x85:      /* NEL */
3672                  break;
3673                  }
3674              break;              break;
3675    
3676              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
# Line 2774  for (;;) Line 3705  for (;;)
3705        /* Control never gets here */        /* Control never gets here */
3706        }        }
3707    
3708      /* If maximizing it is worth using inline code for speed, doing the type      /* If maximizing, it is worth using inline code for speed, doing the type
3709      test once at the start (i.e. keep it out of the loop). Again, keep the      test once at the start (i.e. keep it out of the loop). Again, keep the
3710      UTF-8 and UCP stuff separate. */      UTF-8 and UCP stuff separate. */
3711    
# Line 2783  for (;;) Line 3714  for (;;)
3714        pp = eptr;  /* Remember where we started */        pp = eptr;  /* Remember where we started */
3715    
3716  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3717        if (prop_type > 0)        if (prop_type >= 0)
3718          {          {
3719          for (i = min; i < max; i++)          switch(prop_type)
3720            {            {
3721            int len = 1;            case PT_ANY:
3722            if (eptr >= md->end_subject) break;            for (i = min; i < max; i++)
3723            GETCHARLEN(c, eptr, len);              {
3724            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              int len = 1;
3725            if ((*prop_test_variable == prop_test_against) == prop_fail_result)              if (eptr >= md->end_subject) break;
3726              break;              GETCHARLEN(c, eptr, len);
3727            eptr+= len;              if (prop_fail_result) break;
3728                eptr+= len;
3729                }
3730              break;
3731    
3732              case PT_LAMP:
3733              for (i = min; i < max; i++)
3734                {
3735                int len = 1;
3736                if (eptr >= md->end_subject) break;
3737                GETCHARLEN(c, eptr, len);
3738                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3739                if ((prop_chartype == ucp_Lu ||
3740                     prop_chartype == ucp_Ll ||
3741                     prop_chartype == ucp_Lt) == prop_fail_result)
3742                  break;
3743                eptr+= len;
3744                }
3745              break;
3746    
3747              case PT_GC:
3748              for (i = min; i < max; i++)
3749                {
3750                int len = 1;
3751                if (eptr >= md->end_subject) break;
3752                GETCHARLEN(c, eptr, len);
3753                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3754                if ((prop_category == prop_value) == prop_fail_result)
3755                  break;
3756                eptr+= len;
3757                }
3758              break;
3759    
3760              case PT_PC:
3761              for (i = min; i < max; i++)
3762                {
3763                int len = 1;
3764                if (eptr >= md->end_subject) break;
3765                GETCHARLEN(c, eptr, len);
3766                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3767                if ((prop_chartype == prop_value) == prop_fail_result)
3768                  break;
3769                eptr+= len;
3770                }
3771              break;
3772    
3773              case PT_SC:
3774              for (i = min; i < max; i++)
3775                {
3776                int len = 1;
3777                if (eptr >= md->end_subject) break;
3778                GETCHARLEN(c, eptr, len);
3779                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3780                if ((prop_script == prop_value) == prop_fail_result)
3781                  break;
3782                eptr+= len;
3783                }
3784              break;
3785            }            }
3786    
3787          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
3788    
3789            if (possessive) continue;
3790          for(;;)          for(;;)
3791            {            {
3792            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44);
3793            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3794            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3795            BACKCHAR(eptr);            if (utf8) BACKCHAR(eptr);
3796            }            }
3797          }          }
3798    
# Line 2816  for (;;) Line 3805  for (;;)
3805            {            {
3806            if (eptr >= md->end_subject) break;            if (eptr >= md->end_subject) break;
3807            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3808            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3809            if (prop_category == ucp_M) break;            if (prop_category == ucp_M) break;
3810            while (eptr < md->end_subject)            while (eptr < md->end_subject)
3811              {              {
# Line 2825  for (;;) Line 3814  for (;;)
3814                {                {
3815                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3816                }                }
3817              prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3818              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3819              eptr += len;              eptr += len;
3820              }              }
# Line 2833  for (;;) Line 3822  for (;;)
3822    
3823          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
3824    
3825            if (possessive) continue;
3826          for(;;)          for(;;)
3827            {            {
3828            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM45);
3829            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3830            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3831            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
3832              {              {
3833              int len = 1;              int len = 1;
             BACKCHAR(eptr);  
3834              if (!utf8) c = *eptr; else              if (!utf8) c = *eptr; else
3835                {                {
3836                  BACKCHAR(eptr);
3837                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3838                }                }
3839              prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3840              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3841              eptr--;              eptr--;
3842              }              }
# Line 2864  for (;;) Line 3854  for (;;)
3854          switch(ctype)          switch(ctype)
3855            {            {
3856            case OP_ANY:            case OP_ANY:
   
           /* Special code is required for UTF8, but when the maximum is unlimited  
           we don't need it, so we repeat the non-UTF8 code. This is probably  
           worth it, because .* is quite a common idiom. */  
   
3857            if (max < INT_MAX)            if (max < INT_MAX)
3858              {              {
3859              if ((ims & PCRE_DOTALL) == 0)              if ((ims & PCRE_DOTALL) == 0)
3860                {                {
3861                for (i = min; i < max; i++)                for (i = min; i < max; i++)
3862                  {                  {
3863                  if (eptr >= md->end_subject || *eptr == NEWLINE) break;                  if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
3864                  eptr++;                  eptr++;
3865                  while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                  while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3866                  }                  }
# Line 2884  for (;;) Line 3869  for (;;)
3869                {                {
3870                for (i = min; i < max; i++)                for (i = min; i < max; i++)
3871                  {                  {
3872                    if (eptr >= md->end_subject) break;
3873                  eptr++;                  eptr++;
3874                  while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                  while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3875                  }                  }
# Line 2898  for (;;) Line 3884  for (;;)
3884                {                {
3885                for (i = min; i < max; i++)                for (i = min; i < max; i++)
3886                  {                  {
3887                  if (eptr >= md->end_subject || *eptr == NEWLINE) break;                  if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
3888                  eptr++;                  eptr++;
3889                    while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3890                  }                  }
               break;  
3891                }                }
3892              else              else
3893                {                {
3894                c = max - min;                eptr = md->end_subject;
               if (c > md->end_subject - eptr) c = md->end_subject - eptr;  
               eptr += c;  
3895                }                }
3896              }              }
3897            break;            break;
# Line 2916  for (;;) Line 3900  for (;;)
3900    
3901            case OP_ANYBYTE:            case OP_ANYBYTE:
3902            c = max - min;            c = max - min;
3903            if (c > md->end_subject - eptr) c = md->end_subject - eptr;            if (c > (unsigned int)(md->end_subject - eptr))
3904                c = md->end_subject - eptr;
3905            eptr += c;            eptr += c;
3906            break;            break;
3907    
3908              case OP_ANYNL:
3909              for (i = min; i < max; i++)
3910                {
3911                int len = 1;
3912                if (eptr >= md->end_subject) break;
3913                GETCHARLEN(c, eptr, len);
3914                if (c == 0x000d)
3915                  {
3916                  if (++eptr >= md->end_subject) break;
3917                  if (*eptr == 0x000a) eptr++;
3918                  }
3919                else
3920                  {
3921                  if (c != 0x000a &&
3922                      (md->bsr_anycrlf ||
3923                       (c != 0x000b && c != 0x000c &&
3924                        c != 0x0085 && c != 0x2028 && c != 0x2029)))
3925                    break;
3926                  eptr += len;
3927                  }
3928                }
3929              break;
3930    
3931              case OP_NOT_HSPACE:
3932              case OP_HSPACE:
3933              for (i = min; i < max; i++)
3934                {
3935                BOOL gotspace;
3936                int len = 1;
3937                if (eptr >= md->end_subject) break;
3938                GETCHARLEN(c, eptr, len);
3939                switch(c)
3940                  {
3941                  default: gotspace = FALSE; break;
3942                  case 0x09:      /* HT */
3943                  case 0x20:      /* SPACE */
3944                  case 0xa0:      /* NBSP */
3945                  case 0x1680:    /* OGHAM SPACE MARK */
3946                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3947                  case 0x2000:    /* EN QUAD */
3948                  case 0x2001:    /* EM QUAD */
3949                  case 0x2002:    /* EN SPACE */
3950                  case 0x2003:    /* EM SPACE */
3951                  case 0x2004:    /* THREE-PER-EM SPACE */
3952                  case 0x2005:    /* FOUR-PER-EM SPACE */
3953                  case 0x2006:    /* SIX-PER-EM SPACE */
3954                  case 0x2007:    /* FIGURE SPACE */
3955                  case 0x2008:    /* PUNCTUATION SPACE */
3956                  case 0x2009:    /* THIN SPACE */
3957                  case 0x200A:    /* HAIR SPACE */
3958                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3959                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3960                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3961                  gotspace = TRUE;
3962                  break;
3963                  }
3964                if (gotspace == (ctype == OP_NOT_HSPACE)) break;
3965                eptr += len;
3966                }
3967              break;
3968    
3969              case OP_NOT_VSPACE:
3970              case OP_VSPACE:
3971              for (i = min; i < max; i++)
3972                {
3973                BOOL gotspace;
3974                int len = 1;
3975                if (eptr >= md->end_subject) break;
3976                GETCHARLEN(c, eptr, len);
3977                switch(c)
3978                  {
3979                  default: gotspace = FALSE; break;
3980                  case 0x0a:      /* LF */
3981                  case 0x0b:      /* VT */
3982                  case 0x0c:      /* FF */
3983                  case 0x0d:      /* CR */
3984                  case 0x85:      /* NEL */
3985                  case 0x2028:    /* LINE SEPARATOR */
3986                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3987                  gotspace = TRUE;
3988                  break;
3989                  }
3990                if (gotspace == (ctype == OP_NOT_VSPACE)) break;
3991                eptr += len;
3992                }
3993              break;
3994    
3995            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
3996            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3997              {              {
# Line 2992  for (;;) Line 4064  for (;;)
4064    
4065          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
4066    
4067            if (possessive) continue;
4068          for(;;)          for(;;)
4069            {            {
4070            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM46);
4071            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4072            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
4073            BACKCHAR(eptr);            BACKCHAR(eptr);
4074            }            }
4075          }          }
4076        else        else
4077  #endif  #endif  /* SUPPORT_UTF8 */
4078    
4079        /* Not UTF-8 mode */        /* Not UTF-8 mode */
4080          {          {
# Line 3012  for (;;) Line 4085  for (;;)
4085              {              {
4086              for (i = min; i < max; i++)              for (i = min; i < max; i++)
4087                {                {
4088                if (eptr >= md->end_subject || *eptr == NEWLINE) break;                if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
4089                eptr++;                eptr++;
4090                }                }
4091              break;              break;
# Line 3021  for (;;) Line 4094  for (;;)
4094    
4095            case OP_ANYBYTE:            case OP_ANYBYTE:
4096            c = max - min;            c = max - min;
4097            if (c > md->end_subject - eptr) c = md->end_subject - eptr;            if (c > (unsigned int)(md->end_subject - eptr))
4098                c = md->end_subject - eptr;
4099            eptr += c;            eptr += c;
4100            break;            break;
4101    
4102              case OP_ANYNL:
4103              for (i = min; i < max; i++)
4104                {
4105                if (eptr >= md->end_subject) break;
4106                c = *eptr;
4107                if (c == 0x000d)
4108                  {
4109                  if (++eptr >= md->end_subject) break;
4110                  if (*eptr == 0x000a) eptr++;
4111                  }
4112                else
4113                  {
4114                  if (c != 0x000a &&
4115                      (md->bsr_anycrlf ||
4116                        (c != 0x000b && c != 0x000c && c != 0x0085)))
4117                    break;
4118                  eptr++;
4119                  }
4120                }
4121              break;
4122    
4123              case OP_NOT_HSPACE:
4124              for (i = min; i < max; i++)
4125                {
4126                if (eptr >= md->end_subject) break;
4127                c = *eptr;
4128                if (c == 0x09 || c == 0x20 || c == 0xa0) break;
4129                eptr++;
4130                }
4131              break;
4132    
4133              case OP_HSPACE:
4134              for (i = min; i < max; i++)
4135                {
4136                if (eptr >= md->end_subject) break;
4137                c = *eptr;
4138                if (c != 0x09 && c != 0x20 && c != 0xa0) break;
4139                eptr++;
4140                }
4141              break;
4142    
4143              case OP_NOT_VSPACE:
4144              for (i = min; i < max; i++)
4145                {
4146                if (eptr >= md->end_subject) break;
4147                c = *eptr;
4148                if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85)
4149                  break;
4150                eptr++;
4151                }
4152              break;
4153    
4154              case OP_VSPACE:
4155              for (i = min; i < max; i++)
4156                {
4157                if (eptr >= md->end_subject) break;
4158                c = *eptr;
4159                if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85)
4160                  break;
4161                eptr++;
4162                }
4163              break;
4164    
4165            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
4166            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4167              {              {
# Line 3085  for (;;) Line 4222  for (;;)
4222    
4223          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
4224    
4225            if (possessive) continue;
4226          while (eptr >= pp)          while (eptr >= pp)
4227            {            {
4228            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM47);
4229            eptr--;            eptr--;
4230            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4231            }            }
# Line 3099  for (;;) Line 4237  for (;;)
4237        }        }
4238      /* Control never gets here */      /* Control never gets here */
4239    
4240      /* There's been some horrible disaster. Since all codes > OP_BRA are      /* There's been some horrible disaster. Arrival here can only mean there is
4241      for capturing brackets, and there shouldn't be any gaps between 0 and      something seriously wrong in the code above or the OP_xxx definitions. */
     OP_BRA, arrival here can only mean there is something seriously wrong  
     in the code above or the OP_xxx definitions. */  
4242    
4243      default:      default:
4244      DPRINTF(("Unknown opcode %d\n", *ecode));      DPRINTF(("Unknown opcode %d\n", *ecode));
4245      RRETURN(PCRE_ERROR_UNKNOWN_NODE);      RRETURN(PCRE_ERROR_UNKNOWN_OPCODE);
4246      }      }
4247    
4248    /* Do not stick any code in here without much thought; it is assumed    /* Do not stick any code in here without much thought; it is assumed
# Line 3115  for (;;) Line 4251  for (;;)
4251    
4252    }             /* End of main loop */    }             /* End of main loop */
4253  /* Control never reaches here */  /* Control never reaches here */
4254    
4255    
4256    /* When compiling to use the heap rather than the stack for recursive calls to
4257    match(), the RRETURN() macro jumps here. The number that is saved in
4258    frame->Xwhere indicates which label we actually want to return to. */
4259    
4260    #ifdef NO_RECURSE
4261    #define LBL(val) case val: goto L_RM##val;
4262    HEAP_RETURN:
4263    switch (frame->Xwhere)
4264      {
4265      LBL( 1) LBL( 2) LBL( 3) LBL( 4) LBL( 5) LBL( 6) LBL( 7) LBL( 8)
4266      LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(17)
4267      LBL(19) LBL(24) LBL(25) LBL(26) LBL(27) LBL(29) LBL(31) LBL(33)
4268      LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52)
4269      LBL(53) LBL(54)
4270    #ifdef SUPPORT_UTF8
4271      LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30)
4272      LBL(32) LBL(34) LBL(42) LBL(46)
4273    #ifdef SUPPORT_UCP
4274      LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45)
4275    #endif  /* SUPPORT_UCP */
4276    #endif  /* SUPPORT_UTF8 */
4277      default:
4278      DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));
4279      return PCRE_ERROR_INTERNAL;
4280      }
4281    #undef LBL
4282    #endif  /* NO_RECURSE */
4283  }  }
4284    
4285    
# Line 3127  Undefine all the macros that were define Line 4292  Undefine all the macros that were define
4292  #ifdef NO_RECURSE  #ifdef NO_RECURSE
4293  #undef eptr  #undef eptr
4294  #undef ecode  #undef ecode
4295    #undef mstart
4296  #undef offset_top  #undef offset_top
4297  #undef ims  #undef ims
4298  #undef eptrb  #undef eptrb
# Line 3144  Undefine all the macros that were define Line 4310  Undefine all the macros that were define
4310    
4311  #undef cur_is_word  #undef cur_is_word
4312  #undef condition  #undef condition
 #undef minimize  
4313  #undef prev_is_word  #undef prev_is_word
4314    
4315  #undef original_ims  #undef original_ims
# Line 3200  Returns:          > 0 => success; value Line 4365  Returns:          > 0 => success; value
4365                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
4366  */  */
4367    
4368  PCRE_EXPORT int  PCRE_EXP_DEFN int
4369  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
4370    const char *subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
4371    int offsetcount)    int offsetcount)
4372  {  {
4373  int rc, resetcount, ocount;  int rc, resetcount, ocount;
4374  int first_byte = -1;  int first_byte = -1;
4375  int req_byte = -1;  int req_byte = -1;
4376  int req_byte2 = -1;  int req_byte2 = -1;
4377  unsigned long int ims = 0;  int newline;
4378    unsigned long int ims;
4379  BOOL using_temporary_offsets = FALSE;  BOOL using_temporary_offsets = FALSE;
4380  BOOL anchored;  BOOL anchored;
4381  BOOL startline;  BOOL startline;
4382  BOOL firstline;  BOOL firstline;
4383  BOOL first_byte_caseless = FALSE;  BOOL first_byte_caseless = FALSE;
4384  BOOL req_byte_caseless = FALSE;  BOOL req_byte_caseless = FALSE;
4385    BOOL utf8;
4386  match_data match_block;  match_data match_block;
4387    match_data *md = &match_block;
4388  const uschar *tables;  const uschar *tables;
4389  const uschar *start_bits = NULL;  const uschar *start_bits = NULL;
4390  const uschar *start_match = (const uschar *)subject + start_offset;  USPTR start_match = (USPTR)subject + start_offset;
4391  const uschar *end_subject;  USPTR end_subject;
4392  const uschar *req_byte_ptr = start_match - 1;  USPTR req_byte_ptr = start_match - 1;
4393    
4394  pcre_study_data internal_study;  pcre_study_data internal_study;
4395  const pcre_study_data *study;  const pcre_study_data *study;
# Line 3241  if (offsetcount < 0) return PCRE_ERROR_B Line 4409  if (offsetcount < 0) return PCRE_ERROR_B
4409  the default values. */  the default values. */
4410    
4411  study = NULL;  study = NULL;
4412  match_block.match_limit = MATCH_LIMIT;  md->match_limit = MATCH_LIMIT;
4413  match_block.callout_data = NULL;  md->match_limit_recursion = MATCH_LIMIT_RECURSION;
4414    md->callout_data = NULL;
4415    
4416  /* The table pointer is always in native byte order. */  /* The table pointer is always in native byte order. */
4417    
# Line 3254  if (extra_data != NULL) Line 4423  if (extra_data != NULL)
4423    if ((flags & PCRE_EXTRA_STUDY_DATA) != 0)    if ((flags & PCRE_EXTRA_STUDY_DATA) != 0)
4424      study = (const pcre_study_data *)extra_data->study_data;      study = (const pcre_study_data *)extra_data->study_data;
4425    if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0)    if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0)
4426      match_block.match_limit = extra_data->match_limit;      md->match_limit = extra_data->match_limit;
4427      if ((flags & PCRE_EXTRA_MATCH_LIMIT_RECURSION) != 0)
4428        md->match_limit_recursion = extra_data->match_limit_recursion;
4429    if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0)    if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0)
4430      match_block.callout_data = extra_data->callout_data;      md->callout_data = extra_data->callout_data;
4431    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;
4432    }    }
4433    
# Line 3281  if (re->magic_number != MAGIC_NUMBER) Line 4452  if (re->magic_number != MAGIC_NUMBER)
4452  /* Set up other data */  /* Set up other data */
4453    
4454  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;
4455  startline = (re->options & PCRE_STARTLINE) != 0;  startline = (re->flags & PCRE_STARTLINE) != 0;
4456  firstline = (re->options & PCRE_FIRSTLINE) != 0;  firstline = (re->options & PCRE_FIRSTLINE) != 0;
4457    
4458  /* The code starts after the real_pcre block and the capture name table. */  /* The code starts after the real_pcre block and the capture name table. */
4459    
4460  match_block.start_code = (const uschar *)external_re + re->name_table_offset +  md->start_code = (const uschar *)external_re + re->name_table_offset +
4461    re->name_count * re->name_entry_size;    re->name_count * re->name_entry_size;
4462    
4463  match_block.start_subject = (const uschar *)subject;  md->start_subject = (USPTR)subject;
4464  match_block.start_offset = start_offset;  md->start_offset = start_offset;
4465  match_block.end_subject = match_block.start_subject + length;  md->end_subject = md->start_subject + length;
4466  end_subject = match_block.end_subject;  end_subject = md->end_subject;
   
 match_block.endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;  
 match_block.utf8 = (re->options & PCRE_UTF8) != 0;  
   
 match_block.notbol = (options & PCRE_NOTBOL) != 0;  
 match_block.noteol = (options & PCRE_NOTEOL) != 0;  
 match_block.notempty = (options & PCRE_NOTEMPTY) != 0;  
 match_block.partial = (options & PCRE_PARTIAL) != 0;  
 match_block.hitend = FALSE;  
4467    
4468  match_block.recursive = NULL;                   /* No recursion at top level */  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;
4469    utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;
4470    md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;
4471    
4472    md->notbol = (options & PCRE_NOTBOL) != 0;
4473    md->noteol = (options & PCRE_NOTEOL) != 0;
4474    md->notempty = (options & PCRE_NOTEMPTY) != 0;
4475    md->partial = (options & PCRE_PARTIAL) != 0;
4476    md->hitend = FALSE;
4477    
4478    md->recursive = NULL;                   /* No recursion at top level */
4479    
4480    md->lcc = tables + lcc_offset;
4481    md->ctypes = tables + ctypes_offset;
4482    
4483    /* Handle different \R options. */
4484    
4485    switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE))
4486      {
4487      case 0:
4488      if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0)
4489        md->bsr_anycrlf = (re->options & PCRE_BSR_ANYCRLF) != 0;
4490      else
4491    #ifdef BSR_ANYCRLF
4492      md->bsr_anycrlf = TRUE;
4493    #else
4494      md->bsr_anycrlf = FALSE;
4495    #endif
4496      break;
4497    
4498      case PCRE_BSR_ANYCRLF:
4499      md->bsr_anycrlf = TRUE;
4500      break;
4501    
4502      case PCRE_BSR_UNICODE:
4503      md->bsr_anycrlf = FALSE;
4504      break;
4505    
4506      default: return PCRE_ERROR_BADNEWLINE;
4507      }
4508    
4509    /* Handle different types of newline. The three bits give eight cases. If
4510    nothing is set at run time, whatever was used at compile time applies. */
4511    
4512    switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options :
4513            (pcre_uint32)options) & PCRE_NEWLINE_BITS)
4514      {
4515      case 0: newline = NEWLINE; break;   /* Compile-time default */
4516      case PCRE_NEWLINE_CR: newline = '\r'; break;
4517      case PCRE_NEWLINE_LF: newline = '\n'; break;
4518      case PCRE_NEWLINE_CR+
4519           PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;
4520      case PCRE_NEWLINE_ANY: newline = -1; break;
4521      case PCRE_NEWLINE_ANYCRLF: newline = -2; break;
4522      default: return PCRE_ERROR_BADNEWLINE;
4523      }
4524    
4525  match_block.lcc = tables + lcc_offset;  if (newline == -2)
4526  match_block.ctypes = tables + ctypes_offset;    {
4527      md->nltype = NLTYPE_ANYCRLF;
4528      }
4529    else if (newline < 0)
4530      {
4531      md->nltype = NLTYPE_ANY;
4532      }
4533    else
4534      {
4535      md->nltype = NLTYPE_FIXED;
4536      if (newline > 255)
4537        {
4538        md->nllen = 2;
4539        md->nl[0] = (newline >> 8) & 255;
4540        md->nl[1] = newline & 255;
4541        }
4542      else
4543        {
4544        md->nllen = 1;
4545        md->nl[0] = newline;
4546        }
4547      }
4548    
4549  /* Partial matching is supported only for a restricted set of regexes at the  /* Partial matching is supported only for a restricted set of regexes at the
4550  moment. */  moment. */
4551    
4552  if (match_block.partial && (re->options & PCRE_NOPARTIAL) != 0)  if (md->partial && (re->flags & PCRE_NOPARTIAL) != 0)
4553    return PCRE_ERROR_BADPARTIAL;    return PCRE_ERROR_BADPARTIAL;
4554    
4555  /* Check a UTF-8 string if required. Unfortunately there's no way of passing  /* Check a UTF-8 string if required. Unfortunately there's no way of passing
4556  back the character offset. */  back the character offset. */
4557    
4558  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4559  if (match_block.utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)
4560    {    {
4561    if (_pcre_valid_utf8((uschar *)subject, length) >= 0)    if (_pcre_valid_utf8((uschar *)subject, length) >= 0)
4562      return PCRE_ERROR_BADUTF8;      return PCRE_ERROR_BADUTF8;
# Line 3350  ocount = offsetcount - (offsetcount % 3) Line 4588  ocount = offsetcount - (offsetcount % 3)
4588  if (re->top_backref > 0 && re->top_backref >= ocount/3)  if (re->top_backref > 0 && re->top_backref >= ocount/3)
4589    {    {
4590    ocount = re->top_backref * 3 + 3;    ocount = re->top_backref * 3 + 3;
4591    match_block.offset_vector = (int *)(pcre_malloc)(ocount * sizeof(int));    md->offset_vector = (int *)(pcre_malloc)(ocount * sizeof(int));
4592    if (match_block.offset_vector == NULL) return PCRE_ERROR_NOMEMORY;    if (md->offset_vector == NULL) return PCRE_ERROR_NOMEMORY;
4593    using_temporary_offsets = TRUE;    using_temporary_offsets = TRUE;
4594    DPRINTF(("Got memory to hold back references\n"));    DPRINTF(("Got memory to hold back references\n"));
4595    }    }
4596  else match_block.offset_vector = offsets;  else md->offset_vector = offsets;
4597    
4598  match_block.offset_end = ocount;  md->offset_end = ocount;
4599  match_block.offset_max = (2*ocount)/3;  md->offset_max = (2*ocount)/3;
4600  match_block.offset_overflow = FALSE;  md->offset_overflow = FALSE;
4601  match_block.capture_last = -1;  md->capture_last = -1;
4602    
4603  /* Compute the minimum number of offsets that we need to reset each time. Doing  /* Compute the minimum number of offsets that we need to reset each time. Doing
4604  this makes a huge difference to execution time when there aren't many brackets  this makes a huge difference to execution time when there aren't many brackets
# Line 3373  if (resetcount > offsetcount) resetcount Line 4611  if (resetcount > offsetcount) resetcount
4611  never be used unless previously set, but they get saved and restored, and so we  never be used unless previously set, but they get saved and restored, and so we
4612  initialize them to avoid reading uninitialized locations. */  initialize them to avoid reading uninitialized locations. */
4613    
4614  if (match_block.offset_vector != NULL)  if (md->offset_vector != NULL)
4615    {    {
4616    register int *iptr = match_block.offset_vector + ocount;    register int *iptr = md->offset_vector + ocount;
4617    register int *iend = iptr - resetcount/2 + 1;    register int *iend = iptr - resetcount/2 + 1;
4618    while (--iptr >= iend) *iptr = -1;    while (--iptr >= iend) *iptr = -1;
4619    }    }
# Line 3388  studied, there may be a bitmap of possib Line 4626  studied, there may be a bitmap of possib
4626    
4627  if (!anchored)  if (!anchored)
4628    {    {
4629    if ((re->options & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
4630      {      {
4631      first_byte = re->first_byte & 255;      first_byte = re->first_byte & 255;
4632      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)
4633        first_byte = match_block.lcc[first_byte];        first_byte = md->lcc[first_byte];
4634      }      }
4635    else    else
4636      if (!startline && study != NULL &&      if (!startline && study != NULL &&
# Line 3403  if (!anchored) Line 4641  if (!anchored)
4641  /* For anchored or unanchored matches, there may be a "last known required  /* For anchored or unanchored matches, there may be a "last known required
4642  character" set. */  character" set. */
4643    
4644  if ((re->options & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
4645    {    {
4646    req_byte = re->req_byte & 255;    req_byte = re->req_byte & 255;
4647    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;
4648    req_byte2 = (tables + fcc_offset)[req_byte];  /* case flipped */    req_byte2 = (tables + fcc_offset)[req_byte];  /* case flipped */
4649    }    }
4650    
4651    
4652    /* ==========================================================================*/
4653    
4654  /* Loop for handling unanchored repeated matching attempts; for anchored regexs  /* Loop for handling unanchored repeated matching attempts; for anchored regexs
4655  the loop runs just once. */  the loop runs just once. */
4656    
4657  do  for(;;)
4658    {    {
4659    const uschar *save_end_subject = end_subject;    USPTR save_end_subject = end_subject;
4660      USPTR new_start_match;
4661    
4662    /* Reset the maximum number of extractions we might see. */    /* Reset the maximum number of extractions we might see. */
4663    
4664    if (match_block.offset_vector != NULL)    if (md->offset_vector != NULL)
4665      {      {
4666      register int *iptr = match_block.offset_vector;      register int *iptr = md->offset_vector;
4667      register int *iend = iptr + resetcount;      register int *iend = iptr + resetcount;
4668      while (iptr < iend) *iptr++ = -1;      while (iptr < iend) *iptr++ = -1;
4669      }      }
4670    
4671    /* Advance to a unique first char if possible. If firstline is TRUE, the    /* Advance to a unique first char if possible. If firstline is TRUE, the
4672    start of the match is constrained to the first line of a multiline string.    start of the match is constrained to the first line of a multiline string.
4673    Implement this by temporarily adjusting end_subject so that we stop scanning    That is, the match must be before or at the first newline. Implement this by
4674    at a newline. If the match fails at the newline, later code breaks this loop.    temporarily adjusting end_subject so that we stop scanning at a newline. If
4675    */    the match fails at the newline, later code breaks this loop. */
4676    
4677    if (firstline)    if (firstline)
4678      {      {
4679      const uschar *t = start_match;      USPTR t = start_match;
4680      while (t < save_end_subject && *t != '\n') t++;      while (t < md->end_subject && !IS_NEWLINE(t)) t++;
4681      end_subject = t;      end_subject = t;
4682      }      }
4683    
# Line 3445  do Line 4687  do
4687      {      {
4688      if (first_byte_caseless)      if (first_byte_caseless)
4689        while (start_match < end_subject &&        while (start_match < end_subject &&
4690               match_block.lcc[*start_match] != first_byte)               md->lcc[*start_match] != first_byte)
4691          start_match++;          { NEXTCHAR(start_match); }
4692      else      else
4693        while (start_match < end_subject && *start_match != first_byte)        while (start_match < end_subject && *start_match != first_byte)
4694          start_match++;          { NEXTCHAR(start_match); }
4695      }      }
4696    
4697    /* Or to just after \n for a multiline match if possible */    /* Or to just after a linebreak for a multiline match if possible */
4698    
4699    else if (startline)    else if (startline)
4700      {      {
4701      if (start_match > match_block.start_subject + start_offset)      if (start_match > md->start_subject + start_offset)
4702        {        {
4703        while (start_match < end_subject && start_match[-1] != NEWLINE)        while (start_match <= end_subject && !WAS_NEWLINE(start_match))
4704            { NEXTCHAR(start_match); }
4705    
4706          /* If we have just passed a CR and the newline option is ANY or ANYCRLF,
4707          and we are now at a LF, advance the match position by one more character.
4708          */
4709    
4710          if (start_match[-1] == '\r' &&
4711               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
4712               start_match < end_subject &&
4713               *start_match == '\n')
4714          start_match++;          start_match++;
4715        }        }
4716      }      }
# Line 3470  do Line 4722  do
4722      while (start_match < end_subject)      while (start_match < end_subject)
4723        {        {
4724        register unsigned int c = *start_match;        register unsigned int c = *start_match;
4725        if ((start_bits[c/8] & (1 << (c&7))) == 0) start_match++; else break;        if ((start_bits[c/8] & (1 << (c&7))) == 0)
4726            { NEXTCHAR(start_match); }
4727          else break;
4728        }        }
4729      }      }
4730    
# Line 3480  do Line 4734  do
4734    
4735  #ifdef DEBUG  /* Sigh. Some compilers never learn. */  #ifdef DEBUG  /* Sigh. Some compilers never learn. */
4736    printf(">>>> Match against: ");    printf(">>>> Match against: ");
4737    pchars(start_match, end_subject - start_match, TRUE, &match_block);    pchars(start_match, end_subject - start_match, TRUE, md);