/[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 341 by ph10, Sat Apr 19 16:41:04 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        /* Fall through */
1437    
1438        case OP_ALLANY:
1439      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);
 #ifdef SUPPORT_UTF8  
1440      if (utf8)      if (utf8)
1441        while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;        while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
 #endif  
1442      ecode++;      ecode++;
1443      break;      break;
1444    
# Line 1343  for (;;) Line 1528  for (;;)
1528      ecode++;      ecode++;
1529      break;      break;
1530    
1531        case OP_ANYNL:
1532        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1533        GETCHARINCTEST(c, eptr);
1534        switch(c)
1535          {
1536          default: RRETURN(MATCH_NOMATCH);
1537          case 0x000d:
1538          if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
1539          break;
1540    
1541          case 0x000a:
1542          break;
1543    
1544          case 0x000b:
1545          case 0x000c:
1546          case 0x0085:
1547          case 0x2028:
1548          case 0x2029:
1549          if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
1550          break;
1551          }
1552        ecode++;
1553        break;
1554    
1555        case OP_NOT_HSPACE:
1556        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1557        GETCHARINCTEST(c, eptr);
1558        switch(c)
1559          {
1560          default: break;
1561          case 0x09:      /* HT */
1562          case 0x20:      /* SPACE */
1563          case 0xa0:      /* NBSP */
1564          case 0x1680:    /* OGHAM SPACE MARK */
1565          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1566          case 0x2000:    /* EN QUAD */
1567          case 0x2001:    /* EM QUAD */
1568          case 0x2002:    /* EN SPACE */
1569          case 0x2003:    /* EM SPACE */
1570          case 0x2004:    /* THREE-PER-EM SPACE */
1571          case 0x2005:    /* FOUR-PER-EM SPACE */
1572          case 0x2006:    /* SIX-PER-EM SPACE */
1573          case 0x2007:    /* FIGURE SPACE */
1574          case 0x2008:    /* PUNCTUATION SPACE */
1575          case 0x2009:    /* THIN SPACE */
1576          case 0x200A:    /* HAIR SPACE */
1577          case 0x202f:    /* NARROW NO-BREAK SPACE */
1578          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1579          case 0x3000:    /* IDEOGRAPHIC SPACE */
1580          RRETURN(MATCH_NOMATCH);
1581          }
1582        ecode++;
1583        break;
1584    
1585        case OP_HSPACE:
1586        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1587        GETCHARINCTEST(c, eptr);
1588        switch(c)
1589          {
1590          default: RRETURN(MATCH_NOMATCH);
1591          case 0x09:      /* HT */
1592          case 0x20:      /* SPACE */
1593          case 0xa0:      /* NBSP */
1594          case 0x1680:    /* OGHAM SPACE MARK */
1595          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1596          case 0x2000:    /* EN QUAD */
1597          case 0x2001:    /* EM QUAD */
1598          case 0x2002:    /* EN SPACE */
1599          case 0x2003:    /* EM SPACE */
1600          case 0x2004:    /* THREE-PER-EM SPACE */
1601          case 0x2005:    /* FOUR-PER-EM SPACE */
1602          case 0x2006:    /* SIX-PER-EM SPACE */
1603          case 0x2007:    /* FIGURE SPACE */
1604          case 0x2008:    /* PUNCTUATION SPACE */
1605          case 0x2009:    /* THIN SPACE */
1606          case 0x200A:    /* HAIR SPACE */
1607          case 0x202f:    /* NARROW NO-BREAK SPACE */
1608          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1609          case 0x3000:    /* IDEOGRAPHIC SPACE */
1610          break;
1611          }
1612        ecode++;
1613        break;
1614    
1615        case OP_NOT_VSPACE:
1616        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1617        GETCHARINCTEST(c, eptr);
1618        switch(c)
1619          {
1620          default: break;
1621          case 0x0a:      /* LF */
1622          case 0x0b:      /* VT */
1623          case 0x0c:      /* FF */
1624          case 0x0d:      /* CR */
1625          case 0x85:      /* NEL */
1626          case 0x2028:    /* LINE SEPARATOR */
1627          case 0x2029:    /* PARAGRAPH SEPARATOR */
1628          RRETURN(MATCH_NOMATCH);
1629          }
1630        ecode++;
1631        break;
1632    
1633        case OP_VSPACE:
1634        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1635        GETCHARINCTEST(c, eptr);
1636        switch(c)
1637          {
1638          default: RRETURN(MATCH_NOMATCH);
1639          case 0x0a:      /* LF */
1640          case 0x0b:      /* VT */
1641          case 0x0c:      /* FF */
1642          case 0x0d:      /* CR */
1643          case 0x85:      /* NEL */
1644          case 0x2028:    /* LINE SEPARATOR */
1645          case 0x2029:    /* PARAGRAPH SEPARATOR */
1646          break;
1647          }
1648        ecode++;
1649        break;
1650    
1651  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
1652      /* Check the next character by Unicode property. We will get here only      /* Check the next character by Unicode property. We will get here only
1653      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 1657  for (;;)
1657      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1658      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1659        {        {
1660        int chartype, rqdtype;        int chartype, script;
1661        int othercase;        int category = _pcre_ucp_findprop(c, &chartype, &script);
       int category = _pcre_ucp_findchar(c, &chartype, &othercase);  
   
       rqdtype = *(++ecode);  
       ecode++;  
1662    
1663        if (rqdtype >= 128)        switch(ecode[1])
1664          {          {
1665          if ((rqdtype - 128 != category) == (op == OP_PROP))          case PT_ANY:
1666            if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
1667            break;
1668    
1669            case PT_LAMP:
1670            if ((chartype == ucp_Lu ||
1671                 chartype == ucp_Ll ||
1672                 chartype == ucp_Lt) == (op == OP_NOTPROP))
1673            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1674          }           break;
1675        else  
1676          {          case PT_GC:
1677          if ((rqdtype != chartype) == (op == OP_PROP))          if ((ecode[2] != category) == (op == OP_PROP))
1678              RRETURN(MATCH_NOMATCH);
1679            break;
1680    
1681            case PT_PC:
1682            if ((ecode[2] != chartype) == (op == OP_PROP))
1683            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1684            break;
1685    
1686            case PT_SC:
1687            if ((ecode[2] != script) == (op == OP_PROP))
1688              RRETURN(MATCH_NOMATCH);
1689            break;
1690    
1691            default:
1692            RRETURN(PCRE_ERROR_INTERNAL);
1693          }          }
1694    
1695          ecode += 3;
1696        }        }
1697      break;      break;
1698    
# Line 1379  for (;;) Line 1703  for (;;)
1703      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1704      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1705        {        {
1706        int chartype;        int chartype, script;
1707        int othercase;        int category = _pcre_ucp_findprop(c, &chartype, &script);
       int category = _pcre_ucp_findchar(c, &chartype, &othercase);  
1708        if (category == ucp_M) RRETURN(MATCH_NOMATCH);        if (category == ucp_M) RRETURN(MATCH_NOMATCH);
1709        while (eptr < md->end_subject)        while (eptr < md->end_subject)
1710          {          {
# Line 1390  for (;;) Line 1713  for (;;)
1713            {            {
1714            GETCHARLEN(c, eptr, len);            GETCHARLEN(c, eptr, len);
1715            }            }
1716          category = _pcre_ucp_findchar(c, &chartype, &othercase);          category = _pcre_ucp_findprop(c, &chartype, &script);
1717          if (category != ucp_M) break;          if (category != ucp_M) break;
1718          eptr += len;          eptr += len;
1719          }          }
# Line 1411  for (;;) Line 1734  for (;;)
1734      case OP_REF:      case OP_REF:
1735        {        {
1736        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
1737        ecode += 3;                                 /* Advance past item */        ecode += 3;
1738    
1739        /* If the reference is unset, set the length to be longer than the amount        /* If the reference is unset, there are two possibilities:
1740        of subject left; this ensures that every attempt at a match fails. We  
1741        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
1742        minima. */        than the amount of subject left; this ensures that every attempt at a
1743          match fails. We can't just fail here, because of the possibility of
1744        length = (offset >= offset_top || md->offset_vector[offset] < 0)?        quantifiers with zero minima.
1745          md->end_subject - eptr + 1 :  
1746          md->offset_vector[offset+1] - md->offset_vector[offset];        (b) If the JavaScript compatibility flag is set, set the length to zero
1747          so that the back reference matches an empty string.
1748    
1749          Otherwise, set the length to the length of what was matched by the
1750          referenced subpattern. */
1751    
1752          if (offset >= offset_top || md->offset_vector[offset] < 0)
1753            length = (md->jscript_compat)? 0 : md->end_subject - eptr + 1;
1754          else
1755            length = md->offset_vector[offset+1] - md->offset_vector[offset];
1756    
1757        /* Set up for repetition, or handle the non-repeated case */        /* Set up for repetition, or handle the non-repeated case */
1758    
# Line 1480  for (;;) Line 1812  for (;;)
1812          {          {
1813          for (fi = min;; fi++)          for (fi = min;; fi++)
1814            {            {
1815            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM14);
1816            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1817            if (fi >= max || !match_ref(offset, eptr, length, md, ims))            if (fi >= max || !match_ref(offset, eptr, length, md, ims))
1818              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
# Line 1501  for (;;) Line 1833  for (;;)
1833            }            }
1834          while (eptr >= pp)          while (eptr >= pp)
1835            {            {
1836            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM15);
1837            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1838            eptr -= length;            eptr -= length;
1839            }            }
# Line 1606  for (;;) Line 1938  for (;;)
1938            {            {
1939            for (fi = min;; fi++)            for (fi = min;; fi++)
1940              {              {
1941              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM16);
1942              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1943              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1944              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 1626  for (;;) Line 1958  for (;;)
1958            {            {
1959            for (fi = min;; fi++)            for (fi = min;; fi++)
1960              {              {
1961              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM17);
1962              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1963              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1964              c = *eptr++;              c = *eptr++;
# Line 1663  for (;;) Line 1995  for (;;)
1995              }              }
1996            for (;;)            for (;;)
1997              {              {
1998              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM18);
1999              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2000              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2001              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 1682  for (;;) Line 2014  for (;;)
2014              }              }
2015            while (eptr >= pp)            while (eptr >= pp)
2016              {              {
2017              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM19);
             eptr--;  
2018              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2019                eptr--;
2020              }              }
2021            }            }
2022    
# Line 1753  for (;;) Line 2085  for (;;)
2085          {          {
2086          for (fi = min;; fi++)          for (fi = min;; fi++)
2087            {            {
2088            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20);
2089            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2090            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2091            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
# Line 1777  for (;;) Line 2109  for (;;)
2109            }            }
2110          for(;;)          for(;;)
2111            {            {
2112            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);
2113            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2114            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
2115            BACKCHAR(eptr)            if (utf8) BACKCHAR(eptr);
2116            }            }
2117          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2118          }          }
# Line 1836  for (;;) Line 2168  for (;;)
2168    
2169        else        else
2170          {          {
2171          int dc;          unsigned int dc;
2172          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
2173          ecode += length;          ecode += length;
2174    
2175          /* 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
2176          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. */  
2177    
2178          if (fc != dc)          if (fc != dc)
2179            {            {
2180  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2181            int chartype;            if (dc != _pcre_ucp_othercase(fc))
           int othercase;  
           if (_pcre_ucp_findchar(fc, &chartype, &othercase) < 0 || dc != othercase)  
2182  #endif  #endif
2183              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2184            }            }
# Line 1867  for (;;) Line 2195  for (;;)
2195        }        }
2196      break;      break;
2197    
2198      /* Match a single character repeatedly; different opcodes share code. */      /* Match a single character repeatedly. */
2199    
2200      case OP_EXACT:      case OP_EXACT:
2201      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
2202      ecode += 3;      ecode += 3;
2203      goto REPEATCHAR;      goto REPEATCHAR;
2204    
2205        case OP_POSUPTO:
2206        possessive = TRUE;
2207        /* Fall through */
2208    
2209      case OP_UPTO:      case OP_UPTO:
2210      case OP_MINUPTO:      case OP_MINUPTO:
2211      min = 0;      min = 0;
# Line 1882  for (;;) Line 2214  for (;;)
2214      ecode += 3;      ecode += 3;
2215      goto REPEATCHAR;      goto REPEATCHAR;
2216    
2217        case OP_POSSTAR:
2218        possessive = TRUE;
2219        min = 0;
2220        max = INT_MAX;
2221        ecode++;
2222        goto REPEATCHAR;
2223    
2224        case OP_POSPLUS:
2225        possessive = TRUE;
2226        min = 1;
2227        max = INT_MAX;
2228        ecode++;
2229        goto REPEATCHAR;
2230    
2231        case OP_POSQUERY:
2232        possessive = TRUE;
2233        min = 0;
2234        max = 1;
2235        ecode++;
2236        goto REPEATCHAR;
2237    
2238      case OP_STAR:      case OP_STAR:
2239      case OP_MINSTAR:      case OP_MINSTAR:
2240      case OP_PLUS:      case OP_PLUS:
# Line 1913  for (;;) Line 2266  for (;;)
2266    
2267        if (length > 1)        if (length > 1)
2268          {          {
         int oclength = 0;  
         uschar occhars[8];  
   
2269  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2270          int othercase;          unsigned int othercase;
         int chartype;  
2271          if ((ims & PCRE_CASELESS) != 0 &&          if ((ims & PCRE_CASELESS) != 0 &&
2272               _pcre_ucp_findchar(fc, &chartype, &othercase) >= 0 &&              (othercase = _pcre_ucp_othercase(fc)) != NOTACHAR)
              othercase > 0)  
2273            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = _pcre_ord2utf8(othercase, occhars);
2274            else oclength = 0;
2275  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2276    
2277          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2278            {            {
2279            if (memcmp(eptr, charptr, length) == 0) eptr += length;            if (memcmp(eptr, charptr, length) == 0) eptr += length;
2280    #ifdef SUPPORT_UCP
2281            /* Need braces because of following else */            /* Need braces because of following else */
2282            else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }            else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }
2283            else            else
# Line 1935  for (;;) Line 2285  for (;;)
2285              if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);              if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);
2286              eptr += oclength;              eptr += oclength;
2287              }              }
2288    #else   /* without SUPPORT_UCP */
2289              else { RRETURN(MATCH_NOMATCH); }
2290    #endif  /* SUPPORT_UCP */
2291            }            }
2292    
2293          if (min == max) continue;          if (min == max) continue;
# Line 1943  for (;;) Line 2296  for (;;)
2296            {            {
2297            for (fi = min;; fi++)            for (fi = min;; fi++)
2298              {              {
2299              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM22);
2300              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2301              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2302              if (memcmp(eptr, charptr, length) == 0) eptr += length;              if (memcmp(eptr, charptr, length) == 0) eptr += length;
2303    #ifdef SUPPORT_UCP
2304              /* Need braces because of following else */              /* Need braces because of following else */
2305              else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }              else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }
2306              else              else
# Line 1954  for (;;) Line 2308  for (;;)
2308                if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);                if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);
2309                eptr += oclength;                eptr += oclength;
2310                }                }
2311    #else   /* without SUPPORT_UCP */
2312                else { RRETURN (MATCH_NOMATCH); }
2313    #endif  /* SUPPORT_UCP */
2314              }              }
2315            /* Control never gets here */            /* Control never gets here */
2316            }            }
2317          else  
2318            else  /* Maximize */
2319            {            {
2320            pp = eptr;            pp = eptr;
2321            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2322              {              {
2323              if (eptr > md->end_subject - length) break;              if (eptr > md->end_subject - length) break;
2324              if (memcmp(eptr, charptr, length) == 0) eptr += length;              if (memcmp(eptr, charptr, length) == 0) eptr += length;
2325    #ifdef SUPPORT_UCP
2326              else if (oclength == 0) break;              else if (oclength == 0) break;
2327              else              else
2328                {                {
2329                if (memcmp(eptr, occhars, oclength) != 0) break;                if (memcmp(eptr, occhars, oclength) != 0) break;
2330                eptr += oclength;                eptr += oclength;
2331                }                }
2332    #else   /* without SUPPORT_UCP */
2333                else break;
2334    #endif  /* SUPPORT_UCP */
2335              }              }
2336            while (eptr >= pp)  
2337              if (possessive) continue;
2338              for(;;)
2339             {             {
2340             RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);             RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM23);
2341             if (rrc != MATCH_NOMATCH) RRETURN(rrc);             if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2342               if (eptr == pp) RRETURN(MATCH_NOMATCH);
2343    #ifdef SUPPORT_UCP
2344               eptr--;
2345               BACKCHAR(eptr);
2346    #else   /* without SUPPORT_UCP */
2347             eptr -= length;             eptr -= length;
2348    #endif  /* SUPPORT_UCP */
2349             }             }
           RRETURN(MATCH_NOMATCH);  
2350            }            }
2351          /* Control never gets here */          /* Control never gets here */
2352          }          }
# Line 2017  for (;;) Line 2386  for (;;)
2386          {          {
2387          for (fi = min;; fi++)          for (fi = min;; fi++)
2388            {            {
2389            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM24);
2390            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2391            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
2392                fc != md->lcc[*eptr++])                fc != md->lcc[*eptr++])
# Line 2025  for (;;) Line 2394  for (;;)
2394            }            }
2395          /* Control never gets here */          /* Control never gets here */
2396          }          }
2397        else        else  /* Maximize */
2398          {          {
2399          pp = eptr;          pp = eptr;
2400          for (i = min; i < max; i++)          for (i = min; i < max; i++)
# Line 2033  for (;;) Line 2402  for (;;)
2402            if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break;            if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break;
2403            eptr++;            eptr++;
2404            }            }
2405            if (possessive) continue;
2406          while (eptr >= pp)          while (eptr >= pp)
2407            {            {
2408            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM25);
2409            eptr--;            eptr--;
2410            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2411            }            }
# Line 2054  for (;;) Line 2424  for (;;)
2424          {          {
2425          for (fi = min;; fi++)          for (fi = min;; fi++)
2426            {            {
2427            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM26);
2428            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2429            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)
2430              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2431            }            }
2432          /* Control never gets here */          /* Control never gets here */
2433          }          }
2434        else        else  /* Maximize */
2435          {          {
2436          pp = eptr;          pp = eptr;
2437          for (i = min; i < max; i++)          for (i = min; i < max; i++)
# Line 2069  for (;;) Line 2439  for (;;)
2439            if (eptr >= md->end_subject || fc != *eptr) break;            if (eptr >= md->end_subject || fc != *eptr) break;
2440            eptr++;            eptr++;
2441            }            }
2442            if (possessive) continue;
2443          while (eptr >= pp)          while (eptr >= pp)
2444            {            {
2445            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27);
2446            eptr--;            eptr--;
2447            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2448            }            }
# Line 2121  for (;;) Line 2492  for (;;)
2492      ecode += 3;      ecode += 3;
2493      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
2494    
2495        case OP_NOTPOSSTAR:
2496        possessive = TRUE;
2497        min = 0;
2498        max = INT_MAX;
2499        ecode++;
2500        goto REPEATNOTCHAR;
2501    
2502        case OP_NOTPOSPLUS:
2503        possessive = TRUE;
2504        min = 1;
2505        max = INT_MAX;
2506        ecode++;
2507        goto REPEATNOTCHAR;
2508    
2509        case OP_NOTPOSQUERY:
2510        possessive = TRUE;
2511        min = 0;
2512        max = 1;
2513        ecode++;
2514        goto REPEATNOTCHAR;
2515    
2516        case OP_NOTPOSUPTO:
2517        possessive = TRUE;
2518        min = 0;
2519        max = GET2(ecode, 1);
2520        ecode += 3;
2521        goto REPEATNOTCHAR;
2522    
2523      case OP_NOTSTAR:      case OP_NOTSTAR:
2524      case OP_NOTMINSTAR:      case OP_NOTMINSTAR:
2525      case OP_NOTPLUS:      case OP_NOTPLUS:
# Line 2160  for (;;) Line 2559  for (;;)
2559        /* UTF-8 mode */        /* UTF-8 mode */
2560        if (utf8)        if (utf8)
2561          {          {
2562          register int d;          register unsigned int d;
2563          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2564            {            {
2565            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
# Line 2185  for (;;) Line 2584  for (;;)
2584          /* UTF-8 mode */          /* UTF-8 mode */
2585          if (utf8)          if (utf8)
2586            {            {
2587            register int d;            register unsigned int d;
2588            for (fi = min;; fi++)            for (fi = min;; fi++)
2589              {              {
2590              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28);
2591              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2592              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2593              if (d < 256) d = md->lcc[d];              if (d < 256) d = md->lcc[d];
# Line 2202  for (;;) Line 2601  for (;;)
2601            {            {
2602            for (fi = min;; fi++)            for (fi = min;; fi++)
2603              {              {
2604              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM29);
2605              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2606              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])
2607                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2221  for (;;) Line 2620  for (;;)
2620          /* UTF-8 mode */          /* UTF-8 mode */
2621          if (utf8)          if (utf8)
2622            {            {
2623            register int d;            register unsigned int d;
2624            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2625              {              {
2626              int len = 1;              int len = 1;
# Line 2231  for (;;) Line 2630  for (;;)
2630              if (fc == d) break;              if (fc == d) break;
2631              eptr += len;              eptr += len;
2632              }              }
2633            for(;;)          if (possessive) continue;
2634            for(;;)
2635              {              {
2636              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM30);
2637              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2638              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2639              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2248  for (;;) Line 2648  for (;;)
2648              if (eptr >= md->end_subject || fc == md->lcc[*eptr]) break;              if (eptr >= md->end_subject || fc == md->lcc[*eptr]) break;
2649              eptr++;              eptr++;
2650              }              }
2651              if (possessive) continue;
2652            while (eptr >= pp)            while (eptr >= pp)
2653              {              {
2654              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM31);
2655              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2656              eptr--;              eptr--;
2657              }              }
# Line 2269  for (;;) Line 2670  for (;;)
2670        /* UTF-8 mode */        /* UTF-8 mode */
2671        if (utf8)        if (utf8)
2672          {          {
2673          register int d;          register unsigned int d;
2674          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2675            {            {
2676            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
# Line 2292  for (;;) Line 2693  for (;;)
2693          /* UTF-8 mode */          /* UTF-8 mode */
2694          if (utf8)          if (utf8)
2695            {            {
2696            register int d;            register unsigned int d;
2697            for (fi = min;; fi++)            for (fi = min;; fi++)
2698              {              {
2699              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32);
2700              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2701              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2702              if (fi >= max || eptr >= md->end_subject || fc == d)              if (fi >= max || eptr >= md->end_subject || fc == d)
# Line 2308  for (;;) Line 2709  for (;;)
2709            {            {
2710            for (fi = min;; fi++)            for (fi = min;; fi++)
2711              {              {
2712              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM33);
2713              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2714              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)
2715                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2327  for (;;) Line 2728  for (;;)
2728          /* UTF-8 mode */          /* UTF-8 mode */
2729          if (utf8)          if (utf8)
2730            {            {
2731            register int d;            register unsigned int d;
2732            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2733              {              {
2734              int len = 1;              int len = 1;
# Line 2336  for (;;) Line 2737  for (;;)
2737              if (fc == d) break;              if (fc == d) break;
2738              eptr += len;              eptr += len;
2739              }              }
2740              if (possessive) continue;
2741            for(;;)            for(;;)
2742              {              {
2743              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM34);
2744              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2745              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2746              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2353  for (;;) Line 2755  for (;;)
2755              if (eptr >= md->end_subject || fc == *eptr) break;              if (eptr >= md->end_subject || fc == *eptr) break;
2756              eptr++;              eptr++;
2757              }              }
2758              if (possessive) continue;
2759            while (eptr >= pp)            while (eptr >= pp)
2760              {              {
2761              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM35);
2762              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2763              eptr--;              eptr--;
2764              }              }
# Line 2384  for (;;) Line 2787  for (;;)
2787      ecode += 3;      ecode += 3;
2788      goto REPEATTYPE;      goto REPEATTYPE;
2789    
2790        case OP_TYPEPOSSTAR:
2791        possessive = TRUE;
2792        min = 0;
2793        max = INT_MAX;
2794        ecode++;
2795        goto REPEATTYPE;
2796    
2797        case OP_TYPEPOSPLUS:
2798        possessive = TRUE;
2799        min = 1;
2800        max = INT_MAX;
2801        ecode++;
2802        goto REPEATTYPE;
2803    
2804        case OP_TYPEPOSQUERY:
2805        possessive = TRUE;
2806        min = 0;
2807        max = 1;
2808        ecode++;
2809        goto REPEATTYPE;
2810    
2811        case OP_TYPEPOSUPTO:
2812        possessive = TRUE;
2813        min = 0;
2814        max = GET2(ecode, 1);
2815        ecode += 3;
2816        goto REPEATTYPE;
2817    
2818      case OP_TYPESTAR:      case OP_TYPESTAR:
2819      case OP_TYPEMINSTAR:      case OP_TYPEMINSTAR:
2820      case OP_TYPEPLUS:      case OP_TYPEPLUS:
# Line 2408  for (;;) Line 2839  for (;;)
2839        {        {
2840        prop_fail_result = ctype == OP_NOTPROP;        prop_fail_result = ctype == OP_NOTPROP;
2841        prop_type = *ecode++;        prop_type = *ecode++;
2842        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;  
         }  
2843        }        }
2844      else prop_type = -1;      else prop_type = -1;
2845  #endif  #endif
# Line 2434  for (;;) Line 2856  for (;;)
2856      if (min > 0)      if (min > 0)
2857        {        {
2858  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2859        if (prop_type > 0)        if (prop_type >= 0)
2860          {          {
2861          for (i = 1; i <= min; i++)          switch(prop_type)
2862            {            {
2863            GETCHARINC(c, eptr);            case PT_ANY:
2864            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);            if (prop_fail_result) RRETURN(MATCH_NOMATCH);
2865            if ((*prop_test_variable == prop_test_against) == prop_fail_result)            for (i = 1; i <= min; i++)
2866              RRETURN(MATCH_NOMATCH);              {
2867                if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2868                GETCHARINCTEST(c, eptr);
2869                }
2870              break;
2871    
2872              case PT_LAMP:
2873              for (i = 1; i <= min; i++)
2874                {
2875                if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2876                GETCHARINCTEST(c, eptr);
2877                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2878                if ((prop_chartype == ucp_Lu ||
2879                     prop_chartype == ucp_Ll ||
2880                     prop_chartype == ucp_Lt) == prop_fail_result)
2881                  RRETURN(MATCH_NOMATCH);
2882                }
2883              break;
2884    
2885              case PT_GC:
2886              for (i = 1; i <= min; i++)
2887                {
2888                if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2889                GETCHARINCTEST(c, eptr);
2890                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2891                if ((prop_category == prop_value) == prop_fail_result)
2892                  RRETURN(MATCH_NOMATCH);
2893                }
2894              break;
2895    
2896              case PT_PC:
2897              for (i = 1; i <= min; i++)
2898                {
2899                if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2900                GETCHARINCTEST(c, eptr);
2901                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2902                if ((prop_chartype == prop_value) == prop_fail_result)
2903                  RRETURN(MATCH_NOMATCH);
2904                }
2905              break;
2906    
2907              case PT_SC:
2908              for (i = 1; i <= min; i++)
2909                {
2910                if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2911                GETCHARINCTEST(c, eptr);
2912                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2913                if ((prop_script == prop_value) == prop_fail_result)
2914                  RRETURN(MATCH_NOMATCH);
2915                }
2916              break;
2917    
2918              default:
2919              RRETURN(PCRE_ERROR_INTERNAL);
2920            }            }
2921          }          }
2922    
# Line 2453  for (;;) Line 2928  for (;;)
2928          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2929            {            {
2930            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
2931            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2932            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);
2933            while (eptr < md->end_subject)            while (eptr < md->end_subject)
2934              {              {
# Line 2462  for (;;) Line 2937  for (;;)
2937                {                {
2938                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
2939                }                }
2940              prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2941              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
2942              eptr += len;              eptr += len;
2943              }              }
# Line 2481  for (;;) Line 2956  for (;;)
2956          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2957            {            {
2958            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
2959               (*eptr++ == NEWLINE && (ims & PCRE_DOTALL) == 0))                 ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))
2960              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2961              eptr++;
2962              while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
2963              }
2964            break;
2965    
2966            case OP_ALLANY:
2967            for (i = 1; i <= min; i++)
2968              {
2969              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2970              eptr++;
2971            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
2972            }            }
2973          break;          break;
# Line 2491  for (;;) Line 2976  for (;;)
2976          eptr += min;          eptr += min;
2977          break;          break;
2978    
2979            case OP_ANYNL:
2980            for (i = 1; i <= min; i++)
2981              {
2982              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2983              GETCHARINC(c, eptr);
2984              switch(c)
2985                {
2986                default: RRETURN(MATCH_NOMATCH);
2987                case 0x000d:
2988                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
2989                break;
2990    
2991                case 0x000a:
2992                break;
2993    
2994                case 0x000b:
2995                case 0x000c:
2996                case 0x0085:
2997                case 0x2028:
2998                case 0x2029:
2999                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3000                break;
3001                }
3002              }
3003            break;
3004    
3005            case OP_NOT_HSPACE:
3006            for (i = 1; i <= min; i++)
3007              {
3008              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3009              GETCHARINC(c, eptr);
3010              switch(c)
3011                {
3012                default: break;
3013                case 0x09:      /* HT */
3014                case 0x20:      /* SPACE */
3015                case 0xa0:      /* NBSP */
3016                case 0x1680:    /* OGHAM SPACE MARK */
3017                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3018                case 0x2000:    /* EN QUAD */
3019                case 0x2001:    /* EM QUAD */
3020                case 0x2002:    /* EN SPACE */
3021                case 0x2003:    /* EM SPACE */
3022                case 0x2004:    /* THREE-PER-EM SPACE */
3023                case 0x2005:    /* FOUR-PER-EM SPACE */
3024                case 0x2006:    /* SIX-PER-EM SPACE */
3025                case 0x2007:    /* FIGURE SPACE */
3026                case 0x2008:    /* PUNCTUATION SPACE */
3027                case 0x2009:    /* THIN SPACE */
3028                case 0x200A:    /* HAIR SPACE */
3029                case 0x202f:    /* NARROW NO-BREAK SPACE */
3030                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3031                case 0x3000:    /* IDEOGRAPHIC SPACE */
3032                RRETURN(MATCH_NOMATCH);
3033                }
3034              }
3035            break;
3036    
3037            case OP_HSPACE:
3038            for (i = 1; i <= min; i++)
3039              {
3040              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3041              GETCHARINC(c, eptr);
3042              switch(c)
3043                {
3044                default: RRETURN(MATCH_NOMATCH);
3045                case 0x09:      /* HT */
3046                case 0x20:      /* SPACE */
3047                case 0xa0:      /* NBSP */
3048                case 0x1680:    /* OGHAM SPACE MARK */
3049                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3050                case 0x2000:    /* EN QUAD */
3051                case 0x2001:    /* EM QUAD */
3052                case 0x2002:    /* EN SPACE */
3053                case 0x2003:    /* EM SPACE */
3054                case 0x2004:    /* THREE-PER-EM SPACE */
3055                case 0x2005:    /* FOUR-PER-EM SPACE */
3056                case 0x2006:    /* SIX-PER-EM SPACE */
3057                case 0x2007:    /* FIGURE SPACE */
3058                case 0x2008:    /* PUNCTUATION SPACE */
3059                case 0x2009:    /* THIN SPACE */
3060                case 0x200A:    /* HAIR SPACE */
3061                case 0x202f:    /* NARROW NO-BREAK SPACE */
3062                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3063                case 0x3000:    /* IDEOGRAPHIC SPACE */
3064                break;
3065                }
3066              }
3067            break;
3068    
3069            case OP_NOT_VSPACE:
3070            for (i = 1; i <= min; i++)
3071              {
3072              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3073              GETCHARINC(c, eptr);
3074              switch(c)
3075                {
3076                default: break;
3077                case 0x0a:      /* LF */
3078                case 0x0b:      /* VT */
3079                case 0x0c:      /* FF */
3080                case 0x0d:      /* CR */
3081                case 0x85:      /* NEL */
3082                case 0x2028:    /* LINE SEPARATOR */
3083                case 0x2029:    /* PARAGRAPH SEPARATOR */
3084                RRETURN(MATCH_NOMATCH);
3085                }
3086              }
3087            break;
3088    
3089            case OP_VSPACE:
3090            for (i = 1; i <= min; i++)
3091              {
3092              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3093              GETCHARINC(c, eptr);
3094              switch(c)
3095                {
3096                default: RRETURN(MATCH_NOMATCH);
3097                case 0x0a:      /* LF */
3098                case 0x0b:      /* VT */
3099                case 0x0c:      /* FF */
3100                case 0x0d:      /* CR */
3101                case 0x85:      /* NEL */
3102                case 0x2028:    /* LINE SEPARATOR */
3103                case 0x2029:    /* PARAGRAPH SEPARATOR */
3104                break;
3105                }
3106              }
3107            break;
3108    
3109          case OP_NOT_DIGIT:          case OP_NOT_DIGIT:
3110          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3111            {            {
# Line 2515  for (;;) Line 3130  for (;;)
3130          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3131            {            {
3132            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3133               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_space) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0))
3134              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3135            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3136            }            }
3137          break;          break;
3138    
# Line 2535  for (;;) Line 3150  for (;;)
3150          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3151            {            {
3152            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3153               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_word) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0))
3154              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3155            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3156            }            }
3157          break;          break;
3158    
# Line 2559  for (;;) Line 3174  for (;;)
3174  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF8 */
3175    
3176        /* 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
3177        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. We can assume that there are the minimum
3178          number of bytes present, as this was tested above. */
3179    
3180        switch(ctype)        switch(ctype)
3181          {          {
# Line 2567  for (;;) Line 3183  for (;;)
3183          if ((ims & PCRE_DOTALL) == 0)          if ((ims & PCRE_DOTALL) == 0)
3184            {            {
3185            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3186              if (*eptr++ == NEWLINE) RRETURN(MATCH_NOMATCH);              {
3187                if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
3188                eptr++;
3189                }
3190            }            }
3191          else eptr += min;          else eptr += min;
3192          break;          break;
3193    
3194            case OP_ALLANY:
3195            eptr += min;
3196            break;
3197    
3198          case OP_ANYBYTE:          case OP_ANYBYTE:
3199          eptr += min;          eptr += min;
3200          break;          break;
3201    
3202            /* Because of the CRLF case, we can't assume the minimum number of
3203            bytes are present in this case. */
3204    
3205            case OP_ANYNL:
3206            for (i = 1; i <= min; i++)
3207              {
3208              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3209              switch(*eptr++)
3210                {
3211                default: RRETURN(MATCH_NOMATCH);
3212                case 0x000d:
3213                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3214                break;
3215                case 0x000a:
3216                break;
3217    
3218                case 0x000b:
3219                case 0x000c:
3220                case 0x0085:
3221                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3222                break;
3223                }
3224              }
3225            break;
3226    
3227            case OP_NOT_HSPACE:
3228            for (i = 1; i <= min; i++)
3229              {
3230              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3231              switch(*eptr++)
3232                {
3233                default: break;
3234                case 0x09:      /* HT */
3235                case 0x20:      /* SPACE */
3236                case 0xa0:      /* NBSP */
3237                RRETURN(MATCH_NOMATCH);
3238                }
3239              }
3240            break;
3241    
3242            case OP_HSPACE:
3243            for (i = 1; i <= min; i++)
3244              {
3245              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3246              switch(*eptr++)
3247                {
3248                default: RRETURN(MATCH_NOMATCH);
3249                case 0x09:      /* HT */
3250                case 0x20:      /* SPACE */
3251                case 0xa0:      /* NBSP */
3252                break;
3253                }
3254              }
3255            break;
3256    
3257            case OP_NOT_VSPACE:
3258            for (i = 1; i <= min; i++)
3259              {
3260              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3261              switch(*eptr++)
3262                {
3263                default: break;
3264                case 0x0a:      /* LF */
3265                case 0x0b:      /* VT */
3266                case 0x0c:      /* FF */
3267                case 0x0d:      /* CR */
3268                case 0x85:      /* NEL */
3269                RRETURN(MATCH_NOMATCH);
3270                }
3271              }
3272            break;
3273    
3274            case OP_VSPACE:
3275            for (i = 1; i <= min; i++)
3276              {
3277              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3278              switch(*eptr++)
3279                {
3280                default: RRETURN(MATCH_NOMATCH);
3281                case 0x0a:      /* LF */
3282                case 0x0b:      /* VT */
3283                case 0x0c:      /* FF */
3284                case 0x0d:      /* CR */
3285                case 0x85:      /* NEL */
3286                break;
3287                }
3288              }
3289            break;
3290    
3291          case OP_NOT_DIGIT:          case OP_NOT_DIGIT:
3292          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3293            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
# Line 2624  for (;;) Line 3336  for (;;)
3336      if (minimize)      if (minimize)
3337        {        {
3338  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3339        if (prop_type > 0)        if (prop_type >= 0)
3340          {          {
3341          for (fi = min;; fi++)          switch(prop_type)
3342            {            {
3343            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            case PT_ANY:
3344            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            for (fi = min;; fi++)
3345            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              {
3346            GETCHARINC(c, eptr);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM36);
3347            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3348            if ((*prop_test_variable == prop_test_against) == prop_fail_result)              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3349              RRETURN(MATCH_NOMATCH);              GETCHARINC(c, eptr);
3350                if (prop_fail_result) RRETURN(MATCH_NOMATCH);
3351                }
3352              /* Control never gets here */
3353    
3354              case PT_LAMP:
3355              for (fi = min;; fi++)
3356                {
3357                RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM37);
3358                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3359                if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3360                GETCHARINC(c, eptr);
3361                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3362                if ((prop_chartype == ucp_Lu ||
3363                     prop_chartype == ucp_Ll ||
3364                     prop_chartype == ucp_Lt) == prop_fail_result)
3365                  RRETURN(MATCH_NOMATCH);
3366                }
3367              /* Control never gets here */
3368    
3369              case PT_GC:
3370              for (fi = min;; fi++)
3371                {
3372                RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM38);
3373                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3374                if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3375                GETCHARINC(c, eptr);
3376                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3377                if ((prop_category == prop_value) == prop_fail_result)
3378                  RRETURN(MATCH_NOMATCH);
3379                }
3380              /* Control never gets here */
3381    
3382              case PT_PC:
3383              for (fi = min;; fi++)
3384                {
3385                RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM39);
3386                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3387                if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3388                GETCHARINC(c, eptr);
3389                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3390                if ((prop_chartype == prop_value) == prop_fail_result)
3391                  RRETURN(MATCH_NOMATCH);
3392                }
3393              /* Control never gets here */
3394    
3395              case PT_SC:
3396              for (fi = min;; fi++)
3397                {
3398                RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM40);
3399                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3400                if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3401                GETCHARINC(c, eptr);
3402                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3403                if ((prop_script == prop_value) == prop_fail_result)
3404                  RRETURN(MATCH_NOMATCH);
3405                }
3406              /* Control never gets here */
3407    
3408              default:
3409              RRETURN(PCRE_ERROR_INTERNAL);
3410            }            }
3411          }          }
3412    
# Line 2645  for (;;) Line 3417  for (;;)
3417          {          {
3418          for (fi = min;; fi++)          for (fi = min;; fi++)
3419            {            {
3420            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM41);
3421            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3422            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3423            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3424            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3425            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);
3426            while (eptr < md->end_subject)            while (eptr < md->end_subject)
3427              {              {
# Line 2658  for (;;) Line 3430  for (;;)
3430                {                {
3431                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3432                }                }
3433              prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3434              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3435              eptr += len;              eptr += len;
3436              }              }
# Line 2674  for (;;) Line 3446  for (;;)
3446          {          {
3447          for (fi = min;; fi++)          for (fi = min;; fi++)
3448            {            {
3449            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42);
3450            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3451            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject ||
3452                   (ctype == OP_ANY && (ims & PCRE_DOTALL) == 0 &&
3453                    IS_NEWLINE(eptr)))
3454                RRETURN(MATCH_NOMATCH);
3455    
3456            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
3457            switch(ctype)            switch(ctype)
3458              {              {
3459              case OP_ANY:              case OP_ANY:        /* This is the DOTALL case */
3460              if ((ims & PCRE_DOTALL) == 0 && c == NEWLINE) RRETURN(MATCH_NOMATCH);              case OP_ALLANY:
3461                case OP_ANYBYTE:
3462              break;              break;
3463    
3464              case OP_ANYBYTE:              case OP_ANYNL:
3465                switch(c)
3466                  {
3467                  default: RRETURN(MATCH_NOMATCH);
3468                  case 0x000d:
3469                  if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3470                  break;
3471                  case 0x000a:
3472                  break;
3473    
3474                  case 0x000b:
3475                  case 0x000c:
3476                  case 0x0085:
3477                  case 0x2028:
3478                  case 0x2029:
3479                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3480                  break;
3481                  }
3482                break;
3483    
3484                case OP_NOT_HSPACE:
3485                switch(c)
3486                  {
3487                  default: break;
3488                  case 0x09:      /* HT */
3489                  case 0x20:      /* SPACE */
3490                  case 0xa0:      /* NBSP */
3491                  case 0x1680:    /* OGHAM SPACE MARK */
3492                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3493                  case 0x2000:    /* EN QUAD */
3494                  case 0x2001:    /* EM QUAD */
3495                  case 0x2002:    /* EN SPACE */
3496                  case 0x2003:    /* EM SPACE */
3497                  case 0x2004:    /* THREE-PER-EM SPACE */
3498                  case 0x2005:    /* FOUR-PER-EM SPACE */
3499                  case 0x2006:    /* SIX-PER-EM SPACE */
3500                  case 0x2007:    /* FIGURE SPACE */
3501                  case 0x2008:    /* PUNCTUATION SPACE */
3502                  case 0x2009:    /* THIN SPACE */
3503                  case 0x200A:    /* HAIR SPACE */
3504                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3505                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3506                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3507                  RRETURN(MATCH_NOMATCH);
3508                  }
3509                break;
3510    
3511                case OP_HSPACE:
3512                switch(c)
3513                  {
3514                  default: RRETURN(MATCH_NOMATCH);
3515                  case 0x09:      /* HT */
3516                  case 0x20:      /* SPACE */
3517                  case 0xa0:      /* NBSP */
3518                  case 0x1680:    /* OGHAM SPACE MARK */
3519                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3520                  case 0x2000:    /* EN QUAD */
3521                  case 0x2001:    /* EM QUAD */
3522                  case 0x2002:    /* EN SPACE */
3523                  case 0x2003:    /* EM SPACE */
3524                  case 0x2004:    /* THREE-PER-EM SPACE */
3525                  case 0x2005:    /* FOUR-PER-EM SPACE */
3526                  case 0x2006:    /* SIX-PER-EM SPACE */
3527                  case 0x2007:    /* FIGURE SPACE */
3528                  case 0x2008:    /* PUNCTUATION SPACE */
3529                  case 0x2009:    /* THIN SPACE */
3530                  case 0x200A:    /* HAIR SPACE */
3531                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3532                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3533                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3534                  break;
3535                  }
3536                break;
3537    
3538                case OP_NOT_VSPACE:
3539                switch(c)
3540                  {
3541                  default: break;
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                  RRETURN(MATCH_NOMATCH);
3550                  }
3551                break;
3552    
3553                case OP_VSPACE:
3554                switch(c)
3555                  {
3556                  default: RRETURN(MATCH_NOMATCH);
3557                  case 0x0a:      /* LF */
3558                  case 0x0b:      /* VT */
3559                  case 0x0c:      /* FF */
3560                  case 0x0d:      /* CR */
3561                  case 0x85:      /* NEL */
3562                  case 0x2028:    /* LINE SEPARATOR */
3563                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3564                  break;
3565                  }
3566              break;              break;
3567    
3568              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
# Line 2729  for (;;) Line 3606  for (;;)
3606          {          {
3607          for (fi = min;; fi++)          for (fi = min;; fi++)
3608            {            {
3609            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43);
3610            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3611            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject ||
3612                   ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))
3613                RRETURN(MATCH_NOMATCH);
3614    
3615            c = *eptr++;            c = *eptr++;
3616            switch(ctype)            switch(ctype)
3617              {              {
3618              case OP_ANY:              case OP_ANY:     /* This is the DOTALL case */
3619              if ((ims & PCRE_DOTALL) == 0 && c == NEWLINE) RRETURN(MATCH_NOMATCH);              case OP_ALLANY:
3620                case OP_ANYBYTE:
3621                break;
3622    
3623                case OP_ANYNL:
3624                switch(c)
3625                  {
3626                  default: RRETURN(MATCH_NOMATCH);
3627                  case 0x000d:
3628                  if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3629                  break;
3630    
3631                  case 0x000a:
3632                  break;
3633    
3634                  case 0x000b:
3635                  case 0x000c:
3636                  case 0x0085:
3637                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3638                  break;
3639                  }
3640                break;
3641    
3642                case OP_NOT_HSPACE:
3643                switch(c)
3644                  {
3645                  default: break;
3646                  case 0x09:      /* HT */
3647                  case 0x20:      /* SPACE */
3648                  case 0xa0:      /* NBSP */
3649                  RRETURN(MATCH_NOMATCH);
3650                  }
3651                break;
3652    
3653                case OP_HSPACE:
3654                switch(c)
3655                  {
3656                  default: RRETURN(MATCH_NOMATCH);
3657                  case 0x09:      /* HT */
3658                  case 0x20:      /* SPACE */
3659                  case 0xa0:      /* NBSP */
3660                  break;
3661                  }
3662              break;              break;
3663    
3664              case OP_ANYBYTE:              case OP_NOT_VSPACE:
3665                switch(c)
3666                  {
3667                  default: break;
3668                  case 0x0a:      /* LF */
3669                  case 0x0b:      /* VT */
3670                  case 0x0c:      /* FF */
3671                  case 0x0d:      /* CR */
3672                  case 0x85:      /* NEL */
3673                  RRETURN(MATCH_NOMATCH);
3674                  }
3675                break;
3676    
3677                case OP_VSPACE:
3678                switch(c)
3679                  {
3680                  default: RRETURN(MATCH_NOMATCH);
3681                  case 0x0a:      /* LF */
3682                  case 0x0b:      /* VT */
3683                  case 0x0c:      /* FF */
3684                  case 0x0d:      /* CR */
3685                  case 0x85:      /* NEL */
3686                  break;
3687                  }
3688              break;              break;
3689    
3690              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
# Line 2774  for (;;) Line 3719  for (;;)
3719        /* Control never gets here */        /* Control never gets here */
3720        }        }
3721    
3722      /* 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
3723      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
3724      UTF-8 and UCP stuff separate. */      UTF-8 and UCP stuff separate. */
3725    
# Line 2783  for (;;) Line 3728  for (;;)
3728        pp = eptr;  /* Remember where we started */        pp = eptr;  /* Remember where we started */
3729    
3730  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3731        if (prop_type > 0)        if (prop_type >= 0)
3732          {          {
3733          for (i = min; i < max; i++)          switch(prop_type)
3734            {            {
3735            int len = 1;            case PT_ANY:
3736            if (eptr >= md->end_subject) break;            for (i = min; i < max; i++)
3737            GETCHARLEN(c, eptr, len);              {
3738            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              int len = 1;
3739            if ((*prop_test_variable == prop_test_against) == prop_fail_result)              if (eptr >= md->end_subject) break;
3740              break;              GETCHARLEN(c, eptr, len);
3741            eptr+= len;              if (prop_fail_result) break;
3742                eptr+= len;
3743                }
3744              break;
3745    
3746              case PT_LAMP:
3747              for (i = min; i < max; i++)
3748                {
3749                int len = 1;
3750                if (eptr >= md->end_subject) break;
3751                GETCHARLEN(c, eptr, len);
3752                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3753                if ((prop_chartype == ucp_Lu ||
3754                     prop_chartype == ucp_Ll ||
3755                     prop_chartype == ucp_Lt) == prop_fail_result)
3756                  break;
3757                eptr+= len;
3758                }
3759              break;
3760    
3761              case PT_GC:
3762              for (i = min; i < max; i++)
3763                {
3764                int len = 1;
3765                if (eptr >= md->end_subject) break;
3766                GETCHARLEN(c, eptr, len);
3767                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3768                if ((prop_category == prop_value) == prop_fail_result)
3769                  break;
3770                eptr+= len;
3771                }
3772              break;
3773    
3774              case PT_PC:
3775              for (i = min; i < max; i++)
3776                {
3777                int len = 1;
3778                if (eptr >= md->end_subject) break;
3779                GETCHARLEN(c, eptr, len);
3780                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3781                if ((prop_chartype == prop_value) == prop_fail_result)
3782                  break;
3783                eptr+= len;
3784                }
3785              break;
3786    
3787              case PT_SC:
3788              for (i = min; i < max; i++)
3789                {
3790                int len = 1;
3791                if (eptr >= md->end_subject) break;
3792                GETCHARLEN(c, eptr, len);
3793                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3794                if ((prop_script == prop_value) == prop_fail_result)
3795                  break;
3796                eptr+= len;
3797                }
3798              break;
3799            }            }
3800    
3801          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
3802    
3803            if (possessive) continue;
3804          for(;;)          for(;;)
3805            {            {
3806            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44);
3807            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3808            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3809            BACKCHAR(eptr);            if (utf8) BACKCHAR(eptr);
3810            }            }
3811          }          }
3812    
# Line 2816  for (;;) Line 3819  for (;;)
3819            {            {
3820            if (eptr >= md->end_subject) break;            if (eptr >= md->end_subject) break;
3821            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3822            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3823            if (prop_category == ucp_M) break;            if (prop_category == ucp_M) break;
3824            while (eptr < md->end_subject)            while (eptr < md->end_subject)
3825              {              {
# Line 2825  for (;;) Line 3828  for (;;)
3828                {                {
3829                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3830                }                }
3831              prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3832              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3833              eptr += len;              eptr += len;
3834              }              }
# Line 2833  for (;;) Line 3836  for (;;)
3836    
3837          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
3838    
3839            if (possessive) continue;
3840          for(;;)          for(;;)
3841            {            {
3842            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM45);
3843            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3844            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3845            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
3846              {              {
3847              int len = 1;              int len = 1;
             BACKCHAR(eptr);  
3848              if (!utf8) c = *eptr; else              if (!utf8) c = *eptr; else
3849                {                {
3850                  BACKCHAR(eptr);
3851                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3852                }                }
3853              prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3854              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3855              eptr--;              eptr--;
3856              }              }
# Line 2864  for (;;) Line 3868  for (;;)
3868          switch(ctype)          switch(ctype)
3869            {            {
3870            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. */  
   
3871            if (max < INT_MAX)            if (max < INT_MAX)
3872              {              {
3873              if ((ims & PCRE_DOTALL) == 0)              if ((ims & PCRE_DOTALL) == 0)
3874                {                {
3875                for (i = min; i < max; i++)                for (i = min; i < max; i++)
3876                  {                  {
3877                  if (eptr >= md->end_subject || *eptr == NEWLINE) break;                  if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
3878                  eptr++;                  eptr++;
3879                  while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                  while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3880                  }                  }
# Line 2884  for (;;) Line 3883  for (;;)
3883                {                {
3884                for (i = min; i < max; i++)                for (i = min; i < max; i++)
3885                  {                  {
3886                    if (eptr >= md->end_subject) break;
3887                  eptr++;                  eptr++;
3888                  while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                  while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3889                  }                  }
# Line 2898  for (;;) Line 3898  for (;;)
3898                {                {
3899                for (i = min; i < max; i++)                for (i = min; i < max; i++)
3900                  {                  {
3901                  if (eptr >= md->end_subject || *eptr == NEWLINE) break;                  if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
3902                  eptr++;                  eptr++;
3903                    while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3904                  }                  }
               break;  
3905                }                }
3906              else              else
3907                {                {
3908                c = max - min;                eptr = md->end_subject;
3909                if (c > md->end_subject - eptr) c = md->end_subject - eptr;                }
3910                eptr += c;              }
3911              break;
3912    
3913              case OP_ALLANY:
3914              if (max < INT_MAX)
3915                {
3916                for (i = min; i < max; i++)
3917                  {
3918                  if (eptr >= md->end_subject) break;
3919                  eptr++;
3920                  while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3921                }                }
3922              }              }
3923              else eptr = md->end_subject;   /* Unlimited UTF-8 repeat */
3924            break;            break;
3925    
3926            /* The byte case is the same as non-UTF8 */            /* The byte case is the same as non-UTF8 */
3927    
3928            case OP_ANYBYTE:            case OP_ANYBYTE:
3929            c = max - min;            c = max - min;
3930            if (c > md->end_subject - eptr) c = md->end_subject - eptr;            if (c > (unsigned int)(md->end_subject - eptr))
3931                c = md->end_subject - eptr;
3932            eptr += c;            eptr += c;
3933            break;            break;
3934    
3935              case OP_ANYNL:
3936              for (i = min; i < max; i++)
3937                {
3938                int len = 1;
3939                if (eptr >= md->end_subject) break;
3940                GETCHARLEN(c, eptr, len);
3941                if (c == 0x000d)
3942                  {
3943                  if (++eptr >= md->end_subject) break;
3944                  if (*eptr == 0x000a) eptr++;
3945                  }
3946                else
3947                  {
3948                  if (c != 0x000a &&
3949                      (md->bsr_anycrlf ||
3950                       (c != 0x000b && c != 0x000c &&
3951                        c != 0x0085 && c != 0x2028 && c != 0x2029)))
3952                    break;
3953                  eptr += len;
3954                  }
3955                }
3956              break;
3957    
3958              case OP_NOT_HSPACE:
3959              case OP_HSPACE:
3960              for (i = min; i < max; i++)
3961                {
3962                BOOL gotspace;
3963                int len = 1;
3964                if (eptr >= md->end_subject) break;
3965                GETCHARLEN(c, eptr, len);
3966                switch(c)
3967                  {
3968                  default: gotspace = FALSE; break;
3969                  case 0x09:      /* HT */
3970                  case 0x20:      /* SPACE */
3971                  case 0xa0:      /* NBSP */
3972                  case 0x1680:    /* OGHAM SPACE MARK */
3973                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3974                  case 0x2000:    /* EN QUAD */
3975                  case 0x2001:    /* EM QUAD */
3976                  case 0x2002:    /* EN SPACE */
3977                  case 0x2003:    /* EM SPACE */
3978                  case 0x2004:    /* THREE-PER-EM SPACE */
3979                  case 0x2005:    /* FOUR-PER-EM SPACE */
3980                  case 0x2006:    /* SIX-PER-EM SPACE */
3981                  case 0x2007:    /* FIGURE SPACE */
3982                  case 0x2008:    /* PUNCTUATION SPACE */
3983                  case 0x2009:    /* THIN SPACE */
3984                  case 0x200A:    /* HAIR SPACE */
3985                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3986                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3987                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3988                  gotspace = TRUE;
3989                  break;
3990                  }
3991                if (gotspace == (ctype == OP_NOT_HSPACE)) break;
3992                eptr += len;
3993                }
3994              break;
3995    
3996              case OP_NOT_VSPACE:
3997              case OP_VSPACE:
3998              for (i = min; i < max; i++)
3999                {
4000                BOOL gotspace;
4001                int len = 1;
4002                if (eptr >= md->end_subject) break;
4003                GETCHARLEN(c, eptr, len);
4004                switch(c)
4005                  {
4006                  default: gotspace = FALSE; break;
4007                  case 0x0a:      /* LF */
4008                  case 0x0b:      /* VT */
4009                  case 0x0c:      /* FF */
4010                  case 0x0d:      /* CR */
4011                  case 0x85:      /* NEL */
4012                  case 0x2028:    /* LINE SEPARATOR */
4013                  case 0x2029:    /* PARAGRAPH SEPARATOR */
4014                  gotspace = TRUE;
4015                  break;
4016                  }
4017                if (gotspace == (ctype == OP_NOT_VSPACE)) break;
4018                eptr += len;
4019                }
4020              break;
4021    
4022            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
4023            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4024              {              {
# Line 2992  for (;;) Line 4091  for (;;)
4091    
4092          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
4093    
4094            if (possessive) continue;
4095          for(;;)          for(;;)
4096            {            {
4097            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM46);
4098            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4099            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
4100            BACKCHAR(eptr);            BACKCHAR(eptr);
4101            }            }
4102          }          }
4103        else        else
4104  #endif  #endif  /* SUPPORT_UTF8 */
4105    
4106        /* Not UTF-8 mode */        /* Not UTF-8 mode */
4107          {          {
# Line 3012  for (;;) Line 4112  for (;;)
4112              {              {
4113              for (i = min; i < max; i++)              for (i = min; i < max; i++)
4114                {                {
4115                if (eptr >= md->end_subject || *eptr == NEWLINE) break;                if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
4116                eptr++;                eptr++;
4117                }                }
4118              break;              break;
4119              }              }
4120            /* For DOTALL case, fall through and treat as \C */            /* For DOTALL case, fall through */
4121    
4122              case OP_ALLANY:
4123            case OP_ANYBYTE:            case OP_ANYBYTE:
4124            c = max - min;            c = max - min;
4125            if (c > md->end_subject - eptr) c = md->end_subject - eptr;            if (c > (unsigned int)(md->end_subject - eptr))
4126                c = md->end_subject - eptr;
4127            eptr += c;            eptr += c;
4128            break;            break;
4129    
4130              case OP_ANYNL:
4131              for (i = min; i < max; i++)
4132                {
4133                if (eptr >= md->end_subject) break;
4134                c = *eptr;
4135                if (c == 0x000d)
4136                  {
4137                  if (++eptr >= md->end_subject) break;
4138                  if (*eptr == 0x000a) eptr++;
4139                  }
4140                else
4141                  {
4142                  if (c != 0x000a &&
4143                      (md->bsr_anycrlf ||
4144                        (c != 0x000b && c != 0x000c && c != 0x0085)))
4145                    break;
4146                  eptr++;
4147                  }
4148                }
4149              break;
4150    
4151              case OP_NOT_HSPACE:
4152              for (i = min; i < max; i++)
4153                {
4154                if (eptr >= md->end_subject) break;
4155                c = *eptr;
4156                if (c == 0x09 || c == 0x20 || c == 0xa0) break;
4157                eptr++;
4158                }
4159              break;
4160    
4161              case OP_HSPACE:
4162              for (i = min; i < max; i++)
4163                {
4164                if (eptr >= md->end_subject) break;
4165                c = *eptr;
4166                if (c != 0x09 && c != 0x20 && c != 0xa0) break;
4167                eptr++;
4168                }
4169              break;
4170    
4171              case OP_NOT_VSPACE:
4172              for (i = min; i < max; i++)
4173                {
4174                if (eptr >= md->end_subject) break;
4175                c = *eptr;
4176                if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85)
4177                  break;
4178                eptr++;
4179                }
4180              break;
4181    
4182              case OP_VSPACE:
4183              for (i = min; i < max; i++)
4184                {
4185                if (eptr >= md->end_subject) break;
4186                c = *eptr;
4187                if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85)
4188                  break;
4189                eptr++;
4190                }
4191              break;
4192    
4193            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
4194            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4195              {              {
# Line 3085  for (;;) Line 4250  for (;;)
4250    
4251          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
4252    
4253            if (possessive) continue;
4254          while (eptr >= pp)          while (eptr >= pp)
4255            {            {
4256            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM47);
4257            eptr--;            eptr--;
4258            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4259            }            }
# Line 3099  for (;;) Line 4265  for (;;)
4265        }        }
4266      /* Control never gets here */      /* Control never gets here */
4267    
4268      /* 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
4269      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. */  
4270    
4271      default:      default:
4272      DPRINTF(("Unknown opcode %d\n", *ecode));      DPRINTF(("Unknown opcode %d\n", *ecode));
4273      RRETURN(PCRE_ERROR_UNKNOWN_NODE);      RRETURN(PCRE_ERROR_UNKNOWN_OPCODE);
4274      }      }
4275    
4276    /* 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 4279  for (;;)
4279    
4280    }             /* End of main loop */    }             /* End of main loop */
4281  /* Control never reaches here */  /* Control never reaches here */
4282    
4283    
4284    /* When compiling to use the heap rather than the stack for recursive calls to
4285    match(), the RRETURN() macro jumps here. The number that is saved in
4286    frame->Xwhere indicates which label we actually want to return to. */
4287    
4288    #ifdef NO_RECURSE
4289    #define LBL(val) case val: goto L_RM##val;
4290    HEAP_RETURN:
4291    switch (frame->Xwhere)
4292      {
4293      LBL( 1) LBL( 2) LBL( 3) LBL( 4) LBL( 5) LBL( 6) LBL( 7) LBL( 8)
4294      LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(17)
4295      LBL(19) LBL(24) LBL(25) LBL(26) LBL(27) LBL(29) LBL(31) LBL(33)
4296      LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52)
4297      LBL(53) LBL(54)
4298    #ifdef SUPPORT_UTF8
4299      LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30)
4300      LBL(32) LBL(34) LBL(42) LBL(46)
4301    #ifdef SUPPORT_UCP
4302      LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45)
4303    #endif  /* SUPPORT_UCP */
4304    #endif  /* SUPPORT_UTF8 */
4305      default:
4306      DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));
4307      return PCRE_ERROR_INTERNAL;
4308      }
4309    #undef LBL
4310    #endif  /* NO_RECURSE */
4311  }  }
4312    
4313    
# Line 3127  Undefine all the macros that were define Line 4320  Undefine all the macros that were define
4320  #ifdef NO_RECURSE  #ifdef NO_RECURSE
4321  #undef eptr  #undef eptr
4322  #undef ecode  #undef ecode
4323    #undef mstart
4324  #undef offset_top  #undef offset_top
4325  #undef ims  #undef ims
4326  #undef eptrb  #undef eptrb
# Line 3144  Undefine all the macros that were define Line 4338  Undefine all the macros that were define
4338    
4339  #undef cur_is_word  #undef cur_is_word
4340  #undef condition  #undef condition
 #undef minimize  
4341  #undef prev_is_word  #undef prev_is_word
4342    
4343  #undef original_ims  #undef original_ims
# Line 3200  Returns:          > 0 => success; value Line 4393  Returns:          > 0 => success; value
4393                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
4394  */  */
4395    
4396  PCRE_EXPORT int  PCRE_EXP_DEFN int
4397  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
4398    const char *subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
4399    int offsetcount)    int offsetcount)
4400  {  {
4401  int rc, resetcount, ocount;  int rc, resetcount, ocount;
4402  int first_byte = -1;  int first_byte = -1;
4403  int req_byte = -1;  int req_byte = -1;
4404  int req_byte2 = -1;  int req_byte2 = -1;
4405  unsigned long int ims = 0;  int newline;
4406    unsigned long int ims;
4407  BOOL using_temporary_offsets = FALSE;  BOOL using_temporary_offsets = FALSE;
4408  BOOL anchored;  BOOL anchored;
4409  BOOL startline;  BOOL startline;
4410  BOOL firstline;  BOOL firstline;
4411  BOOL first_byte_caseless = FALSE;  BOOL first_byte_caseless = FALSE;
4412  BOOL req_byte_caseless = FALSE;  BOOL req_byte_caseless = FALSE;
4413    BOOL utf8;
4414  match_data match_block;  match_data match_block;
4415    match_data *md = &match_block;
4416  const uschar *tables;  const uschar *tables;
4417  const uschar *start_bits = NULL;  const uschar *start_bits = NULL;
4418  const uschar *start_match = (const uschar *)subject + start_offset;  USPTR start_match = (USPTR)subject + start_offset;
4419  const uschar *end_subject;  USPTR end_subject;
4420  const uschar *req_byte_ptr = start_match - 1;  USPTR req_byte_ptr = start_match - 1;
4421    
4422  pcre_study_data internal_study;  pcre_study_data internal_study;
4423  const pcre_study_data *study;  const pcre_study_data *study;
# Line 3241  if (offsetcount < 0) return PCRE_ERROR_B Line 4437  if (offsetcount < 0) return PCRE_ERROR_B
4437  the default values. */  the default values. */
4438    
4439  study = NULL;  study = NULL;
4440  match_block.match_limit = MATCH_LIMIT;  md->match_limit = MATCH_LIMIT;
4441  match_block.callout_data = NULL;  md->match_limit_recursion = MATCH_LIMIT_RECURSION;
4442    md->callout_data = NULL;
4443    
4444  /* The table pointer is always in native byte order. */  /* The table pointer is always in native byte order. */
4445    
# Line 3254  if (extra_data != NULL) Line 4451  if (extra_data != NULL)
4451    if ((flags & PCRE_EXTRA_STUDY_DATA) != 0)    if ((flags & PCRE_EXTRA_STUDY_DATA) != 0)
4452      study = (const pcre_study_data *)extra_data->study_data;      study = (const pcre_study_data *)extra_data->study_data;
4453    if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0)    if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0)
4454      match_block.match_limit = extra_data->match_limit;      md->match_limit = extra_data->match_limit;
4455      if ((flags & PCRE_EXTRA_MATCH_LIMIT_RECURSION) != 0)
4456        md->match_limit_recursion = extra_data->match_limit_recursion;
4457    if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0)    if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0)
4458      match_block.callout_data = extra_data->callout_data;      md->callout_data = extra_data->callout_data;
4459    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;
4460    }    }
4461    
# Line 3281  if (re->magic_number != MAGIC_NUMBER) Line 4480  if (re->magic_number != MAGIC_NUMBER)
4480  /* Set up other data */  /* Set up other data */
4481    
4482  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;
4483  startline = (re->options & PCRE_STARTLINE) != 0;  startline = (re->flags & PCRE_STARTLINE) != 0;
4484  firstline = (re->options & PCRE_FIRSTLINE) != 0;  firstline = (re->options & PCRE_FIRSTLINE) != 0;
4485    
4486  /* 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. */
4487    
4488  match_block.start_code = (const uschar *)external_re + re->name_table_offset +  md->start_code = (const uschar *)external_re + re->name_table_offset +
4489    re->name_count * re->name_entry_size;    re->name_count * re->name_entry_size;
4490    
4491  match_block.start_subject = (const uschar *)subject;  md->start_subject = (USPTR)subject;
4492  match_block.start_offset = start_offset;  md->start_offset = start_offset;
4493  match_block.end_subject = match_block.start_subject + length;  md->end_subject = md->start_subject + length;
4494  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;  
4495    
4496  match_block.recursive = NULL;                   /* No recursion at top level */  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;
4497    utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;
4498    md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;
4499    
4500    md->notbol = (options & PCRE_NOTBOL) != 0;
4501    md->noteol = (options & PCRE_NOTEOL) != 0;
4502    md->notempty = (options & PCRE_NOTEMPTY) != 0;
4503    md->partial = (options & PCRE_PARTIAL) != 0;
4504    md->hitend = FALSE;
4505    
4506    md->recursive = NULL;                   /* No recursion at top level */
4507    
4508    md->lcc = tables + lcc_offset;
4509    md->ctypes = tables + ctypes_offset;
4510    
4511    /* Handle different \R options. */
4512    
4513    switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE))
4514      {
4515      case 0:
4516      if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0)
4517        md->bsr_anycrlf = (re->options & PCRE_BSR_ANYCRLF) != 0;
4518      else
4519    #ifdef BSR_ANYCRLF
4520      md->bsr_anycrlf = TRUE;
4521    #else
4522      md->bsr_anycrlf = FALSE;
4523    #endif
4524      break;
4525    
4526      case PCRE_BSR_ANYCRLF:
4527      md->bsr_anycrlf = TRUE;
4528      break;
4529    
4530      case PCRE_BSR_UNICODE:
4531      md->bsr_anycrlf = FALSE;
4532      break;
4533    
4534      default: return PCRE_ERROR_BADNEWLINE;
4535      }
4536    
4537    /* Handle different types of newline. The three bits give eight cases. If
4538    nothing is set at run time, whatever was used at compile time applies. */
4539    
4540    switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options :
4541            (pcre_uint32)options) & PCRE_NEWLINE_BITS)
4542      {
4543      case 0: newline = NEWLINE; break;   /* Compile-time default */
4544      case PCRE_NEWLINE_CR: newline = '\r'; break;
4545      case PCRE_NEWLINE_LF: newline = '\n'; break;
4546      case PCRE_NEWLINE_CR+
4547           PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;
4548      case PCRE_NEWLINE_ANY: newline = -1; break;
4549      case PCRE_NEWLINE_ANYCRLF: newline = -2; break;
4550      default: return PCRE_ERROR_BADNEWLINE;
4551      }
4552    
4553  match_block.lcc = tables + lcc_offset;  if (newline == -2)
4554  match_block.ctypes = tables + ctypes_offset;    {
4555      md->nltype = NLTYPE_ANYCRLF;
4556      }
4557    else if (newline < 0)
4558      {
4559      md->nltype = NLTYPE_ANY;
4560      }
4561    else
4562      {
4563      md->nltype = NLTYPE_FIXED;
4564      if (newline > 255)
4565        {
4566        md->nllen = 2;
4567        md->nl[0] = (newline >> 8) & 255;
4568        md->nl[1] = newline & 255;
4569        }
4570      else
4571        {
4572        md->nllen = 1;
4573        md->nl[0] = newline;
4574        }
4575      }
4576    
4577  /* 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
4578  moment. */  moment. */
4579    
4580  if (match_block.partial && (re->options & PCRE_NOPARTIAL) != 0)  if (md->partial && (re->flags & PCRE_NOPARTIAL) != 0)
4581    return PCRE_ERROR_BADPARTIAL;    return PCRE_ERROR_BADPARTIAL;
4582    
4583  /* 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
4584  back the character offset. */  back the character offset. */
4585    
4586  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4587  if (match_block.utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)
4588    {    {
4589    if (_pcre_valid_utf8((uschar *)subject, length) >= 0)    if (_pcre_valid_utf8((uschar *)subject, length) >= 0)
4590      return PCRE_ERROR_BADUTF8;      return PCRE_ERROR_BADUTF8;
# Line 3350  ocount = offsetcount - (offsetcount % 3) Line 4616  ocount = offsetcount - (offsetcount % 3)
4616  if (re->top_backref > 0 && re->top_backref >= ocount/3)  if (re->top_backref > 0 && re->top_backref >= ocount/3)
4617    {    {
4618    ocount = re->top_backref * 3 + 3;    ocount = re->top_backref * 3 + 3;
4619    match_block.offset_vector = (int *)(pcre_malloc)(ocount * sizeof(int));    md->offset_vector = (int *)(pcre_malloc)(ocount * sizeof(int));
4620    if (match_block.offset_vector == NULL) return PCRE_ERROR_NOMEMORY;    if (md->offset_vector == NULL) return PCRE_ERROR_NOMEMORY;
4621    using_temporary_offsets = TRUE;    using_temporary_offsets = TRUE;
4622    DPRINTF(("Got memory to hold back references\n"));    DPRINTF(("Got memory to hold back references\n"));
4623    }    }
4624  else match_block.offset_vector = offsets;  else md->offset_vector = offsets;
4625    
4626  match_block.offset_end = ocount;  md->offset_end = ocount;
4627  match_block.offset_max = (2*ocount)/3;  md->offset_max = (2*ocount)/3;
4628  match_block.offset_overflow = FALSE;  md->offset_overflow = FALSE;
4629  match_block.capture_last = -1;  md->capture_last = -1;
4630    
4631  /* 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
4632  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 4639  if (resetcount > offsetcount) resetcount
4639  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
4640  initialize them to avoid reading uninitialized locations. */  initialize them to avoid reading uninitialized locations. */
4641    
4642  if (match_block.offset_vector != NULL)  if (md->offset_vector != NULL)
4643    {    {
4644    register int *iptr = match_block.offset_vector + ocount;    register int *iptr = md->offset_vector + ocount;
4645    register int *iend = iptr - resetcount/2 + 1;    register int *iend = iptr - resetcount/2 + 1;
4646    while (--iptr >= iend) *iptr = -1;    while (--iptr >= iend) *iptr = -1;
4647    }    }
# Line 3388  studied, there may be a bitmap of possib Line 4654  studied, there may be a bitmap of possib
4654    
4655  if (!anchored)  if (!anchored)
4656    {    {
4657    if ((re->options & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
4658      {      {
4659      first_byte = re->first_byte & 255;      first_byte = re->first_byte & 255;
4660      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)
4661        first_byte = match_block.lcc[first_byte];        first_byte = md->lcc[first_byte];
4662      }      }
4663    else    else
4664      if (!startline && study != NULL &&      if (!startline && study != NULL &&
# Line 3403  if (!anchored) Line 4669  if (!anchored)
4669  /* For anchored or unanchored matches, there may be a "last known required  /* For anchored or unanchored matches, there may be a "last known required
4670  character" set. */  character" set. */
4671    
4672  if ((re->options & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
4673    {    {
4674    req_byte = re->req_byte & 255;    req_byte = re->req_byte & 255;
4675    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;
4676    req_byte2 = (tables + fcc_offset)[req_byte];  /* case flipped */    req_byte2 = (tables + fcc_offset)[req_byte];  /* case flipped */
4677    }    }
4678    
4679    
4680    /* ==========================================================================*/
4681    
4682  /* Loop for handling unanchored repeated matching attempts; for anchored regexs  /* Loop for handling unanchored repeated matching attempts; for anchored regexs
4683  the loop runs just once. */  the loop runs just once. */
4684    
4685  do  for(;;)
4686    {    {
4687    const uschar *save_end_subject = end_subject;    USPTR save_end_subject = end_subject;
4688      USPTR new_start_match;
4689    
4690    /* Reset the maximum number of extractions we might see. */    /* Reset the maximum number of extractions we might see. */
4691    
4692    if (match_block.offset_vector != NULL)    if (md->offset_vector != NULL)
4693      {      {
4694      register int *iptr = match_block.offset_vector;      register int *iptr = md->offset_vector;
4695      register int *iend = iptr + resetcount;      register int *iend = iptr + resetcount;
4696      while (iptr < iend) *iptr++ = -1;      while (iptr < iend) *iptr++ = -1;
4697      }      }
4698    
4699    /* 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
4700    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.
4701    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
4702    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
4703    */    the match fails at the newline, later code breaks this loop. */
4704    
4705    if (firstline)    if (firstline)
4706      {      {
4707      const uschar *t = start_match;      USPTR t = start_match;
4708      while (t < save_end_subject && *t != '\n') t++;      while (t < md->end_subject && !IS_NEWLINE(t)) t++;
4709      end_subject = t;      end_subject = t;
4710      }      }
4711    
# Line 3445  do Line 4715  do
4715      {      {
4716      if (first_byte_caseless)      if (first_byte_caseless)
4717        while (start_match < end_subject &&        while (start_match < end_subject &&
4718               match_block.lcc[*start_match] != first_byte)               md->lcc[*start_match] != first_byte)
4719          start_match++;          { NEXTCHAR(start_match); }
4720      else      else
4721        while (start_match < end_subject && *start_match != first_byte)        while (start_match < end_subject && *start_match != first_byte)
4722          start_match++;          { NEXTCHAR(start_match); }
4723      }      }
4724    
4725    /* Or to just after \n for a multiline match if possible */    /* Or to just after a linebreak for a multiline match if possible */
4726    
4727    else if (startline)    else if (startline)
4728      {      {
4729      if (start_match > match_block.start_subject + start_offset)      if (start_match > md->start_subject + start_offset)
4730        {        {
4731        while (start_match < end_subject && start_match[-1] != NEWLINE)        while (start_match <= end_subject && !WAS_NEWLINE(start_match))
4732            { NEXTCHAR(start_match); }
4733    
4734          /* If we have just passed a CR and the newline option is ANY or ANYCRLF,
4735          and we are now at a LF, advance the match position by one more character.
4736          */
4737    
4738          if (start_match[-1] == '\r' &&
4739               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
4740               start_match < end_subject &&
4741               *start_match == '\n')
4742          start_match++;          start_match++;
4743        }        }
4744      }      }
# Line 3470  do Line 4750  do
4750      while (start_match < end_subject)      while (start_match < end_subject)
4751        {        {
4752        register unsigned int c = *start_match;        register unsigned int c = *start_match;
4753        if ((start_bits[c/8] & (1 << (c&7))) == 0) start_match++; else break;        if ((start_bits[c/8] & (1 << (c&7))) == 0)
4754            { NEXTCHAR(start_match); }
4755          else break;
4756        }        }
4757      }      }
4758    
# Line 3480  do Line 4762  do
4762    
4763  #ifdef DEBUG  /* Sigh. Some compilers never learn. */  #ifdef DEBUG  /* Sigh. Some compilers never learn. */
4764    printf(">>>> Match against: ");    printf(">>>> Match against: ");
4765    pchars(start_match, end_subject - start_match, TRUE, &match_block);    pchars(start_match, end_subject - start_match, TRUE, md);
4766    printf("\n");    printf("\n");
4767  #endif  #endif
4768    
# Line 3494  do Line 4776  do
4776    
4777    HOWEVER: when the subject string is very, very long, searching to its end can    HOWEVER: when the subject string is very, very long, searching to its end can
4778    take a long time, and give bad performance on quite ordinary patterns. This    take a long time, and give bad performance on quite ordinary patterns. This
4779    showed up when somebody was matching /^C/ on a 32-megabyte string... so we    showed up when somebody was matching something like /^\d+C/ on a 32-megabyte
4780    don't do this when the string is sufficiently long.    string... so we don't do this when the string is sufficiently long.
4781    
4782    ALSO: this processing is disabled when partial matching is requested.    ALSO: this processing is disabled when partial matching is requested.
4783    */    */
4784    
4785    if (req_byte >= 0 &&    if (req_byte >= 0 &&
4786        end_subject - start_match < REQ_BYTE_MAX &&        end_subject - start_match < REQ_BYTE_MAX &&
4787        !match_block.partial)        !md->partial)
4788      {      {
4789      register const uschar *p = start_match + ((first_byte >= 0)? 1 : 0);      register USPTR p = start_match + ((first_byte >= 0)? 1 : 0);
4790    
4791      /* We don't need to repeat the search if we haven't yet reached the      /* We don't need to repeat the search if we haven't yet reached the
4792      place we found it at last time. */      place we found it at last time. */
# Line 3527  do Line 4809  do
4809            }            }
4810          }          }
4811    
4812        /* If we can't find the required character, break the matching loop */        /* If we can't find the required character, break the matching loop,
4813          forcing a match failure. */
4814    
4815        if (p >= end_subject) break;        if (p >= end_subject)
4816            {
4817            rc = MATCH_NOMATCH;
4818            break;
4819            }
4820    
4821        /* If we have found the required character, save the point where we        /* If we have found the required character, save the point where we
4822        found it, so that we don't search again next time round the loop if        found it, so that we don't search again next time round the loop if
# Line 3539  do Line 4826  do
4826        }        }
4827      }      }
4828    
4829    /* When a match occurs, substrings will be set for all internal extractions;    /* OK, we can now run the match. */
   we just need to set up the whole thing as substring 0 before returning. If  
   there were too many extractions, set the return code to zero. In the case  
   where we had to get some local store to hold offsets for backreferences, copy  
   those back references that we can. In this case there need not be overflow  
   if certain parts of the pattern were not used. */  
   
   match_block.start_match = start_match;  
   match_block.match_call_count = 0;  
   
   rc = match(start_match, match_block.start_code, 2, &match_block, ims, NULL,  
     match_isgroup);  
   
   /* When the result is no match, if the subject's first character was a  
   newline and the PCRE_FIRSTLINE option is set, break (which will return  
   PCRE_ERROR_NOMATCH). The option requests that a match occur before the first  
   newline in the subject. Otherwise, advance the pointer to the next character  
   and continue - but the continuation will actually happen only when the  
   pattern is not anchored. */  
4830    
4831    if (rc == MATCH_NOMATCH)    md->start_match_ptr = start_match;
4832      md->match_call_count = 0;
4833      rc = match(start_match, md->start_code, start_match, 2, md, ims, NULL, 0, 0);
4834    
4835      switch(rc)
4836      {      {
4837      if (firstline && *start_match == NEWLINE) break;      /* NOMATCH and PRUNE advance by one character. THEN at this level acts
4838      start_match++;      exactly like PRUNE. */
4839    
4840        case MATCH_NOMATCH:
4841        case MATCH_PRUNE:
4842        case MATCH_THEN:
4843        new_start_match = start_match + 1;
4844  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4845      if (match_block.utf8)      if (utf8)
4846        while(start_match < end_subject && (*start_match & 0xc0) == 0x80)        while(new_start_match < end_subject && (*new_start_match & 0xc0) == 0x80)
4847          start_match++;          new_start_match++;
4848  #endif  #endif
4849      continue;      break;
     }  
4850    
4851    if (rc != MATCH_MATCH)      /* SKIP passes back the next starting point explicitly. */
4852      {  
4853      DPRINTF((">>>> error: returning %d\n", rc));      case MATCH_SKIP:
4854      return rc;      new_start_match = md->start_match_ptr;
4855        break;
4856    
4857        /* COMMIT disables the bumpalong, but otherwise behaves as NOMATCH. */
4858    
4859        case MATCH_COMMIT:
4860        rc = MATCH_NOMATCH;
4861        goto ENDLOOP;
4862    
4863        /* Any other return is some kind of error. */
4864    
4865        default:
4866        goto ENDLOOP;
4867      }      }
4868    
4869    /* We have a match! Copy the offset information from temporary store if    /* Control reaches here for the various types of "no match at this point"
4870    necessary */    result. Reset the code to MATCH_NOMATCH for subsequent checking. */
4871    
4872      rc = MATCH_NOMATCH;
4873    
4874      /* If PCRE_FIRSTLINE is set, the match must happen before or at the first
4875      newline in the subject (though it may continue over the newline). Therefore,
4876      if we have just failed to match, starting at a newline, do not continue. */
4877    
4878      if (firstline && IS_NEWLINE(start_match)) break;
4879    
4880      /* Advance to new matching position */
4881    
4882      start_match = new_start_match;
4883    
4884      /* Break the loop if the pattern is anchored or if we have passed the end of
4885      the subject. */
4886    
4887      if (anchored || start_match > end_subject) break;
4888    
4889      /* If we have just passed a CR and we are now at a LF, and the pattern does
4890      not contain any explicit matches for \r or \n, and the newline option is CRLF
4891      or ANY or ANYCRLF, advance the match position by one more character. */
4892    
4893      if (start_match[-1] == '\r' &&
4894          start_match < end_subject &&
4895          *start_match == '\n' &&
4896          (re->flags & PCRE_HASCRORLF) == 0 &&
4897            (md->nltype == NLTYPE_ANY ||
4898             md->nltype == NLTYPE_ANYCRLF ||
4899             md->nllen == 2))