/[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 335 by ph10, Sat Apr 12 14:36:14 2008 UTC
# Line 6  Line 6 
6  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
7    
8                         Written by Philip Hazel                         Written by Philip Hazel
9             Copyright (c) 1997-2005 University of Cambridge             Copyright (c) 1997-2008 University of Cambridge
10    
11  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
12  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 42  POSSIBILITY OF SUCH DAMAGE. Line 42  POSSIBILITY OF SUCH DAMAGE.
42  pattern matching using an NFA algorithm, trying to mimic Perl as closely as  pattern matching using an NFA algorithm, trying to mimic Perl as closely as
43  possible. There are also some static supporting functions. */  possible. There are also some static supporting functions. */
44    
45    #ifdef HAVE_CONFIG_H
46    #include "config.h"
47    #endif
48    
49    #define NLBLOCK md             /* Block containing newline information */
50    #define PSSTART start_subject  /* Field containing processed string start */
51    #define PSEND   end_subject    /* Field containing processed string end */
52    
53  #include "pcre_internal.h"  #include "pcre_internal.h"
54    
55    /* Undefine some potentially clashing cpp symbols */
56    
57  /* Structure for building a chain of data that actually lives on the  #undef min
58  stack, for holding the values of the subject pointer at the start of each  #undef max
 subpattern, so as to detect when an empty string has been matched by a  
 subpattern - to break infinite loops. When NO_RECURSE is set, these blocks  
 are on the heap, not on the stack. */  
   
 typedef struct eptrblock {  
   struct eptrblock *epb_prev;  
   const uschar *epb_saved_eptr;  
 } eptrblock;  
59    
60  /* Flag bits for the match() function */  /* Flag bits for the match() function */
61    
62  #define match_condassert   0x01    /* Called to check a condition assertion */  #define match_condassert     0x01  /* Called to check a condition assertion */
63  #define match_isgroup      0x02    /* Set if start of bracketed group */  #define match_cbegroup       0x02  /* Could-be-empty unlimited repeat group */
64    
65  /* Non-error returns from the match() function. Error returns are externally  /* Non-error returns from the match() function. Error returns are externally
66  defined PCRE_ERROR_xxx codes, which are all negative. */  defined PCRE_ERROR_xxx codes, which are all negative. */
# Line 68  defined PCRE_ERROR_xxx codes, which are Line 68  defined PCRE_ERROR_xxx codes, which are
68  #define MATCH_MATCH        1  #define MATCH_MATCH        1
69  #define MATCH_NOMATCH      0  #define MATCH_NOMATCH      0
70    
71    /* Special internal returns from the match() function. Make them sufficiently
72    negative to avoid the external error codes. */
73    
74    #define MATCH_COMMIT       (-999)
75    #define MATCH_PRUNE        (-998)
76    #define MATCH_SKIP         (-997)
77    #define MATCH_THEN         (-996)
78    
79  /* Maximum number of ints of offset to save on the stack for recursive calls.  /* Maximum number of ints of offset to save on the stack for recursive calls.
80  If the offset vector is bigger, malloc is used. This should be a multiple of 3,  If the offset vector is bigger, malloc is used. This should be a multiple of 3,
81  because the offset vector is always a multiple of 3 long. */  because the offset vector is always a multiple of 3 long. */
# Line 101  Returns:     nothing Line 109  Returns:     nothing
109  static void  static void
110  pchars(const uschar *p, int length, BOOL is_subject, match_data *md)  pchars(const uschar *p, int length, BOOL is_subject, match_data *md)
111  {  {
112  int c;  unsigned int c;
113  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;
114  while (length-- > 0)  while (length-- > 0)
115    if (isprint(c = *(p++))) printf("%c", c); else printf("\\x%02x", c);    if (isprint(c = *(p++))) printf("%c", c); else printf("\\x%02x", c);
# Line 128  Returns:      TRUE if matched Line 136  Returns:      TRUE if matched
136  */  */
137    
138  static BOOL  static BOOL
139  match_ref(int offset, register const uschar *eptr, int length, match_data *md,  match_ref(int offset, register USPTR eptr, int length, match_data *md,
140    unsigned long int ims)    unsigned long int ims)
141  {  {
142  const uschar *p = md->start_subject + md->offset_vector[offset];  USPTR p = md->start_subject + md->offset_vector[offset];
143    
144  #ifdef DEBUG  #ifdef DEBUG
145  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 169  return TRUE; Line 177  return TRUE;
177  ****************************************************************************  ****************************************************************************
178                     RECURSION IN THE match() FUNCTION                     RECURSION IN THE match() FUNCTION
179    
180  The match() function is highly recursive. Some regular expressions can cause  The match() function is highly recursive, though not every recursive call
181  it to recurse thousands of times. I was writing for Unix, so I just let it  increases the recursive depth. Nevertheless, some regular expressions can cause
182  call itself recursively. This uses the stack for saving everything that has  it to recurse to a great depth. I was writing for Unix, so I just let it call
183  to be saved for a recursive call. On Unix, the stack can be large, and this  itself recursively. This uses the stack for saving everything that has to be
184  works fine.  saved for a recursive call. On Unix, the stack can be large, and this works
185    fine.
186  It turns out that on non-Unix systems there are problems with programs that  
187  use a lot of stack. (This despite the fact that every last chip has oodles  It turns out that on some non-Unix-like systems there are problems with
188  of memory these days, and techniques for extending the stack have been known  programs that use a lot of stack. (This despite the fact that every last chip
189  for decades.) So....  has oodles of memory these days, and techniques for extending the stack have
190    been known for decades.) So....
191    
192  There is a fudge, triggered by defining NO_RECURSE, which avoids recursive  There is a fudge, triggered by defining NO_RECURSE, which avoids recursive
193  calls by keeping local variables that need to be preserved in blocks of memory  calls by keeping local variables that need to be preserved in blocks of memory
194  obtained from malloc instead instead of on the stack. Macros are used to  obtained from malloc() instead instead of on the stack. Macros are used to
195  achieve this so that the actual code doesn't look very different to what it  achieve this so that the actual code doesn't look very different to what it
196  always used to.  always used to.
197    
198    The original heap-recursive code used longjmp(). However, it seems that this
199    can be very slow on some operating systems. Following a suggestion from Stan
200    Switzer, the use of longjmp() has been abolished, at the cost of having to
201    provide a unique number for each call to RMATCH. There is no way of generating
202    a sequence of numbers at compile time in C. I have given them names, to make
203    them stand out more clearly.
204    
205    Crude tests on x86 Linux show a small speedup of around 5-8%. However, on
206    FreeBSD, avoiding longjmp() more than halves the time taken to run the standard
207    tests. Furthermore, not using longjmp() means that local dynamic variables
208    don't have indeterminate values; this has meant that the frame size can be
209    reduced because the result can be "passed back" by straight setting of the
210    variable instead of being passed in the frame.
211  ****************************************************************************  ****************************************************************************
212  ***************************************************************************/  ***************************************************************************/
213    
214    /* Numbers for RMATCH calls. When this list is changed, the code at HEAP_RETURN
215    below must be updated in sync.  */
216    
217  /* These versions of the macros use the stack, as normal */  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM6,  RM7,  RM8,  RM9,  RM10,
218           RM11,  RM12, RM13, RM14, RM15, RM16, RM17, RM18, RM19, RM20,
219           RM21,  RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30,
220           RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
221           RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
222           RM51,  RM52, RM53, RM54 };
223    
224    /* These versions of the macros use the stack, as normal. There are debugging
225    versions and production versions. Note that the "rw" argument of RMATCH isn't
226    actuall used in this definition. */
227    
228  #ifndef NO_RECURSE  #ifndef NO_RECURSE
229  #define REGISTER register  #define REGISTER register
230  #define RMATCH(rx,ra,rb,rc,rd,re,rf,rg) rx = match(ra,rb,rc,rd,re,rf,rg)  
231    #ifdef DEBUG
232    #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \
233      { \
234      printf("match() called in line %d\n", __LINE__); \
235      rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1); \
236      printf("to line %d\n", __LINE__); \
237      }
238    #define RRETURN(ra) \
239      { \
240      printf("match() returned %d from line %d ", ra, __LINE__); \
241      return ra; \
242      }
243    #else
244    #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \
245      rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1)
246  #define RRETURN(ra) return ra  #define RRETURN(ra) return ra
247    #endif
248    
249  #else  #else
250    
251    
252  /* These versions of the macros manage a private stack on the heap. Note  /* These versions of the macros manage a private stack on the heap. Note that
253  that the rd argument of RMATCH isn't actually used. It's the md argument of  the "rd" argument of RMATCH isn't actually used in this definition. It's the md
254  match(), which never changes. */  argument of match(), which never changes. */
255    
256  #define REGISTER  #define REGISTER
257    
258  #define RMATCH(rx,ra,rb,rc,rd,re,rf,rg)\  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw)\
259    {\    {\
260    heapframe *newframe = (pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = (pcre_stack_malloc)(sizeof(heapframe));\
261    if (setjmp(frame->Xwhere) == 0)\    frame->Xwhere = rw; \
262      {\    newframe->Xeptr = ra;\
263      newframe->Xeptr = ra;\    newframe->Xecode = rb;\
264      newframe->Xecode = rb;\    newframe->Xmstart = mstart;\
265      newframe->Xoffset_top = rc;\    newframe->Xoffset_top = rc;\
266      newframe->Xims = re;\    newframe->Xims = re;\
267      newframe->Xeptrb = rf;\    newframe->Xeptrb = rf;\
268      newframe->Xflags = rg;\    newframe->Xflags = rg;\
269      newframe->Xprevframe = frame;\    newframe->Xrdepth = frame->Xrdepth + 1;\
270      frame = newframe;\    newframe->Xprevframe = frame;\
271      DPRINTF(("restarting from line %d\n", __LINE__));\    frame = newframe;\
272      goto HEAP_RECURSE;\    DPRINTF(("restarting from line %d\n", __LINE__));\
273      }\    goto HEAP_RECURSE;\
274    else\    L_##rw:\
275      {\    DPRINTF(("jumped back to line %d\n", __LINE__));\
     DPRINTF(("longjumped back to line %d\n", __LINE__));\  
     frame = md->thisframe;\  
     rx = frame->Xresult;\  
     }\  
276    }    }
277    
278  #define RRETURN(ra)\  #define RRETURN(ra)\
# Line 235  match(), which never changes. */ Line 282  match(), which never changes. */
282    (pcre_stack_free)(newframe);\    (pcre_stack_free)(newframe);\
283    if (frame != NULL)\    if (frame != NULL)\
284      {\      {\
285      frame->Xresult = ra;\      rrc = ra;\
286      md->thisframe = frame;\      goto HEAP_RETURN;\
     longjmp(frame->Xwhere, 1);\  
287      }\      }\
288    return ra;\    return ra;\
289    }    }
# Line 252  typedef struct heapframe { Line 298  typedef struct heapframe {
298    
299    const uschar *Xeptr;    const uschar *Xeptr;
300    const uschar *Xecode;    const uschar *Xecode;
301      const uschar *Xmstart;
302    int Xoffset_top;    int Xoffset_top;
303    long int Xims;    long int Xims;
304    eptrblock *Xeptrb;    eptrblock *Xeptrb;
305    int Xflags;    int Xflags;
306      unsigned int Xrdepth;
307    
308    /* Function local variables */    /* Function local variables */
309    
# Line 271  typedef struct heapframe { Line 319  typedef struct heapframe {
319    
320    BOOL Xcur_is_word;    BOOL Xcur_is_word;
321    BOOL Xcondition;    BOOL Xcondition;
   BOOL Xminimize;  
322    BOOL Xprev_is_word;    BOOL Xprev_is_word;
323    
324    unsigned long int Xoriginal_ims;    unsigned long int Xoriginal_ims;
325    
326  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
327    int Xprop_type;    int Xprop_type;
328      int Xprop_value;
329    int Xprop_fail_result;    int Xprop_fail_result;
330    int Xprop_category;    int Xprop_category;
331    int Xprop_chartype;    int Xprop_chartype;
332    int Xprop_othercase;    int Xprop_script;
333    int Xprop_test_against;    int Xoclength;
334    int *Xprop_test_variable;    uschar Xocchars[8];
335  #endif  #endif
336    
337    int Xctype;    int Xctype;
338    int Xfc;    unsigned int Xfc;
339    int Xfi;    int Xfi;
340    int Xlength;    int Xlength;
341    int Xmax;    int Xmax;
# Line 301  typedef struct heapframe { Line 349  typedef struct heapframe {
349    
350    eptrblock Xnewptrb;    eptrblock Xnewptrb;
351    
352    /* Place to pass back result, and where to jump back to */    /* Where to jump back to */
353    
354    int  Xresult;    int Xwhere;
   jmp_buf Xwhere;  
355    
356  } heapframe;  } heapframe;
357    
# Line 320  typedef struct heapframe { Line 367  typedef struct heapframe {
367  *         Match from current position            *  *         Match from current position            *
368  *************************************************/  *************************************************/
369    
370  /* On entry ecode points to the first opcode, and eptr to the first character  /* This function is called recursively in many circumstances. Whenever it
 in the subject string, while eptrb holds the value of eptr at the start of the  
 last bracketed group - used for breaking infinite loops matching zero-length  
 strings. This function is called recursively in many circumstances. Whenever it  
371  returns a negative (error) response, the outer incarnation must also return the  returns a negative (error) response, the outer incarnation must also return the
372  same response.  same response.
373    
# Line 333  performance. Tests using gcc on a SPARC Line 377  performance. Tests using gcc on a SPARC
377  made performance worse.  made performance worse.
378    
379  Arguments:  Arguments:
380     eptr        pointer in subject     eptr        pointer to current character in subject
381     ecode       position in code     ecode       pointer to current position in compiled code
382       mstart      pointer to the current match start position (can be modified
383                     by encountering \K)
384     offset_top  current top pointer     offset_top  current top pointer
385     md          pointer to "static" info for the match     md          pointer to "static" info for the match
386     ims         current /i, /m, and /s options     ims         current /i, /m, and /s options
# Line 342  Arguments: Line 388  Arguments:
388                   brackets - for testing for empty matches                   brackets - for testing for empty matches
389     flags       can contain     flags       can contain
390                   match_condassert - this is an assertion condition                   match_condassert - this is an assertion condition
391                   match_isgroup - this is the start of a bracketed group                   match_cbegroup - this is the start of an unlimited repeat
392                       group that can match an empty string
393       rdepth      the recursion depth
394    
395  Returns:       MATCH_MATCH if matched            )  these values are >= 0  Returns:       MATCH_MATCH if matched            )  these values are >= 0
396                 MATCH_NOMATCH if failed to match  )                 MATCH_NOMATCH if failed to match  )
397                 a negative PCRE_ERROR_xxx value if aborted by an error condition                 a negative PCRE_ERROR_xxx value if aborted by an error condition
398                   (e.g. stopped by recursion limit)                   (e.g. stopped by repeated call or recursion limit)
399  */  */
400    
401  static int  static int
402  match(REGISTER const uschar *eptr, REGISTER const uschar *ecode,  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, const uschar *mstart,
403    int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb,    int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb,
404    int flags)    int flags, unsigned int rdepth)
405  {  {
406  /* These variables do not need to be preserved over recursion in this function,  /* These variables do not need to be preserved over recursion in this function,
407  so they can be ordinary variables in all cases. Mark them with "register"  so they can be ordinary variables in all cases. Mark some of them with
408  because they are used a lot in loops. */  "register" because they are used a lot in loops. */
409    
410  register int  rrc;    /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
411  register int  i;      /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
412  register int  c;      /* Character values not kept over RMATCH() calls */  register unsigned int c;   /* Character values not kept over RMATCH() calls */
413  register BOOL utf8;   /* Local copy of UTF-8 flag for speed */  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */
414    
415    BOOL minimize, possessive; /* Quantifier options */
416    
417  /* When recursion is not being used, all "local" variables that have to be  /* When recursion is not being used, all "local" variables that have to be
418  preserved over calls to RMATCH() are part of a "frame" which is obtained from  preserved over calls to RMATCH() are part of a "frame" which is obtained from
# Line 377  frame->Xprevframe = NULL;            /* Line 427  frame->Xprevframe = NULL;            /*
427    
428  frame->Xeptr = eptr;  frame->Xeptr = eptr;
429  frame->Xecode = ecode;  frame->Xecode = ecode;
430    frame->Xmstart = mstart;
431  frame->Xoffset_top = offset_top;  frame->Xoffset_top = offset_top;
432  frame->Xims = ims;  frame->Xims = ims;
433  frame->Xeptrb = eptrb;  frame->Xeptrb = eptrb;
434  frame->Xflags = flags;  frame->Xflags = flags;
435    frame->Xrdepth = rdepth;
436    
437  /* This is where control jumps back to to effect "recursion" */  /* This is where control jumps back to to effect "recursion" */
438    
# Line 390  HEAP_RECURSE: Line 442  HEAP_RECURSE:
442    
443  #define eptr               frame->Xeptr  #define eptr               frame->Xeptr
444  #define ecode              frame->Xecode  #define ecode              frame->Xecode
445    #define mstart             frame->Xmstart
446  #define offset_top         frame->Xoffset_top  #define offset_top         frame->Xoffset_top
447  #define ims                frame->Xims  #define ims                frame->Xims
448  #define eptrb              frame->Xeptrb  #define eptrb              frame->Xeptrb
449  #define flags              frame->Xflags  #define flags              frame->Xflags
450    #define rdepth             frame->Xrdepth
451    
452  /* Ditto for the local variables */  /* Ditto for the local variables */
453    
# Line 411  HEAP_RECURSE: Line 465  HEAP_RECURSE:
465    
466  #define cur_is_word        frame->Xcur_is_word  #define cur_is_word        frame->Xcur_is_word
467  #define condition          frame->Xcondition  #define condition          frame->Xcondition
 #define minimize           frame->Xminimize  
468  #define prev_is_word       frame->Xprev_is_word  #define prev_is_word       frame->Xprev_is_word
469    
470  #define original_ims       frame->Xoriginal_ims  #define original_ims       frame->Xoriginal_ims
471    
472  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
473  #define prop_type          frame->Xprop_type  #define prop_type          frame->Xprop_type
474    #define prop_value         frame->Xprop_value
475  #define prop_fail_result   frame->Xprop_fail_result  #define prop_fail_result   frame->Xprop_fail_result
476  #define prop_category      frame->Xprop_category  #define prop_category      frame->Xprop_category
477  #define prop_chartype      frame->Xprop_chartype  #define prop_chartype      frame->Xprop_chartype
478  #define prop_othercase     frame->Xprop_othercase  #define prop_script        frame->Xprop_script
479  #define prop_test_against  frame->Xprop_test_against  #define oclength           frame->Xoclength
480  #define prop_test_variable frame->Xprop_test_variable  #define occhars            frame->Xocchars
481  #endif  #endif
482    
483  #define ctype              frame->Xctype  #define ctype              frame->Xctype
# Line 447  HEAP_RECURSE: Line 501  HEAP_RECURSE:
501  get preserved during recursion in the normal way. In this environment, fi and  get preserved during recursion in the normal way. In this environment, fi and
502  i, and fc and c, can be the same variables. */  i, and fc and c, can be the same variables. */
503    
504  #else  #else         /* NO_RECURSE not defined */
505  #define fi i  #define fi i
506  #define fc c  #define fc c
507    
508    
509  #ifdef SUPPORT_UTF8                /* Many of these variables are used ony */  #ifdef SUPPORT_UTF8                /* Many of these variables are used only  */
510  const uschar *charptr;             /* small blocks of the code. My normal  */  const uschar *charptr;             /* in small blocks of the code. My normal */
511  #endif                             /* style of coding would have declared  */  #endif                             /* style of coding would have declared    */
512  const uschar *callpat;             /* them within each of those blocks.    */  const uschar *callpat;             /* them within each of those blocks.      */
513  const uschar *data;                /* However, in order to accommodate the */  const uschar *data;                /* However, in order to accommodate the   */
514  const uschar *next;                /* version of this code that uses an    */  const uschar *next;                /* version of this code that uses an      */
515  const uschar *pp;                  /* external "stack" implemented on the  */  USPTR         pp;                  /* external "stack" implemented on the    */
516  const uschar *prev;                /* heap, it is easier to declare them   */  const uschar *prev;                /* heap, it is easier to declare them all */
517  const uschar *saved_eptr;          /* all here, so the declarations can    */  USPTR         saved_eptr;          /* here, so the declarations can be cut   */
518                                     /* be cut out in a block. The only      */                                     /* out in a block. The only declarations  */
519  recursion_info new_recursive;      /* declarations within blocks below are */  recursion_info new_recursive;      /* within blocks below are for variables  */
520                                     /* for variables that do not have to    */                                     /* that do not have to be preserved over  */
521  BOOL cur_is_word;                  /* be preserved over a recursive call   */  BOOL cur_is_word;                  /* a recursive call to RMATCH().          */
522  BOOL condition;                    /* to RMATCH().                         */  BOOL condition;
 BOOL minimize;  
523  BOOL prev_is_word;  BOOL prev_is_word;
524    
525  unsigned long int original_ims;  unsigned long int original_ims;
526    
527  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
528  int prop_type;  int prop_type;
529    int prop_value;
530  int prop_fail_result;  int prop_fail_result;
531  int prop_category;  int prop_category;
532  int prop_chartype;  int prop_chartype;
533  int prop_othercase;  int prop_script;
534  int prop_test_against;  int oclength;
535  int *prop_test_variable;  uschar occhars[8];
536  #endif  #endif
537    
538  int ctype;  int ctype;
# Line 493  int save_offset1, save_offset2, save_off Line 547  int save_offset1, save_offset2, save_off
547  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
548    
549  eptrblock newptrb;  eptrblock newptrb;
550  #endif  #endif     /* NO_RECURSE */
551    
552  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
553  variables. */  variables. */
554    
555  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
556    prop_value = 0;
557  prop_fail_result = 0;  prop_fail_result = 0;
 prop_test_against = 0;  
 prop_test_variable = NULL;  
558  #endif  #endif
559    
560  /* OK, now we can get on with the real code of the function. Recursion is  
561  specified by the macros RMATCH and RRETURN. When NO_RECURSE is *not* defined,  /* This label is used for tail recursion, which is used in a few cases even
562  these just turn into a recursive call to match() and a "return", respectively.  when NO_RECURSE is not defined, in order to reduce the amount of stack that is
563  However, RMATCH isn't like a function call because it's quite a complicated  used. Thanks to Ian Taylor for noticing this possibility and sending the
564  macro. It has to be used in one particular way. This shouldn't, however, impact  original patch. */
565  performance when true recursion is being used. */  
566    TAIL_RECURSE:
567    
568    /* OK, now we can get on with the real code of the function. Recursive calls
569    are specified by the macro RMATCH and RRETURN is used to return. When
570    NO_RECURSE is *not* defined, these just turn into a recursive call to match()
571    and a "return", respectively (possibly with some debugging if DEBUG is
572    defined). However, RMATCH isn't like a function call because it's quite a
573    complicated macro. It has to be used in one particular way. This shouldn't,
574    however, impact performance when true recursion is being used. */
575    
576    #ifdef SUPPORT_UTF8
577    utf8 = md->utf8;       /* Local copy of the flag */
578    #else
579    utf8 = FALSE;
580    #endif
581    
582    /* First check that we haven't called match() too many times, or that we
583    haven't exceeded the recursive call limit. */
584    
585  if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT);  if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT);
586    if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);
587    
588  original_ims = ims;    /* Save for resetting on ')' */  original_ims = ims;    /* Save for resetting on ')' */
 utf8 = md->utf8;       /* Local copy of the flag */  
589    
590  /* At the start of a bracketed group, add the current subject pointer to the  /* At the start of a group with an unlimited repeat that may match an empty
591  stack of such pointers, to be re-instated at the end of the group when we hit  string, the match_cbegroup flag is set. When this is the case, add the current
592  the closing ket. When match() is called in other circumstances, we don't add to  subject pointer to the chain of such remembered pointers, to be checked when we
593  this stack. */  hit the closing ket, in order to break infinite loops that match no characters.
594    When match() is called in other circumstances, don't add to the chain. The
595    match_cbegroup flag must NOT be used with tail recursion, because the memory
596    block that is used is on the stack, so a new one may be required for each
597    match(). */
598    
599  if ((flags & match_isgroup) != 0)  if ((flags & match_cbegroup) != 0)
600    {    {
   newptrb.epb_prev = eptrb;  
601    newptrb.epb_saved_eptr = eptr;    newptrb.epb_saved_eptr = eptr;
602      newptrb.epb_prev = eptrb;
603    eptrb = &newptrb;    eptrb = &newptrb;
604    }    }
605    
606  /* Now start processing the operations. */  /* Now start processing the opcodes. */
607    
608  for (;;)  for (;;)
609    {    {
610      minimize = possessive = FALSE;
611    op = *ecode;    op = *ecode;
   minimize = FALSE;  
612    
613    /* For partial matching, remember if we ever hit the end of the subject after    /* For partial matching, remember if we ever hit the end of the subject after
614    matching at least one subject character. */    matching at least one subject character. */
615    
616    if (md->partial &&    if (md->partial &&
617        eptr >= md->end_subject &&        eptr >= md->end_subject &&
618        eptr > md->start_match)        eptr > mstart)
619      md->hitend = TRUE;      md->hitend = TRUE;
620    
621    /* Opening capturing bracket. If there is space in the offset vector, save    switch(op)
   the current subject position in the working slot at the top of the vector. We  
   mustn't change the current values of the data slot, because they may be set  
   from a previous iteration of this group, and be referred to by a reference  
   inside the group.  
   
   If the bracket fails to match, we need to restore this value and also the  
   values of the final offsets, in case they were set by a previous iteration of  
   the same bracket.  
   
   If there isn't enough space in the offset vector, treat this as if it were a  
   non-capturing bracket. Don't worry about setting the flag for the error case  
   here; that is handled in the code for KET. */  
   
   if (op > OP_BRA)  
622      {      {
623      number = op - OP_BRA;      case OP_FAIL:
624        RRETURN(MATCH_NOMATCH);
     /* For extended extraction brackets (large number), we have to fish out the  
     number from a dummy opcode at the start. */  
625    
626      if (number > EXTRACT_BASIC_MAX)      case OP_PRUNE:
627        number = GET2(ecode, 2+LINK_SIZE);      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
628          ims, eptrb, flags, RM51);
629        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
630        RRETURN(MATCH_PRUNE);
631    
632        case OP_COMMIT:
633        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
634          ims, eptrb, flags, RM52);
635        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
636        RRETURN(MATCH_COMMIT);
637    
638        case OP_SKIP:
639        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
640          ims, eptrb, flags, RM53);
641        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
642        md->start_match_ptr = eptr;   /* Pass back current position */
643        RRETURN(MATCH_SKIP);
644    
645        case OP_THEN:
646        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
647          ims, eptrb, flags, RM54);
648        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
649        RRETURN(MATCH_THEN);
650    
651        /* Handle a capturing bracket. If there is space in the offset vector, save
652        the current subject position in the working slot at the top of the vector.
653        We mustn't change the current values of the data slot, because they may be
654        set from a previous iteration of this group, and be referred to by a
655        reference inside the group.
656    
657        If the bracket fails to match, we need to restore this value and also the
658        values of the final offsets, in case they were set by a previous iteration
659        of the same bracket.
660    
661        If there isn't enough space in the offset vector, treat this as if it were
662        a non-capturing bracket. Don't worry about setting the flag for the error
663        case here; that is handled in the code for KET. */
664    
665        case OP_CBRA:
666        case OP_SCBRA:
667        number = GET2(ecode, 1+LINK_SIZE);
668      offset = number << 1;      offset = number << 1;
669    
670  #ifdef DEBUG  #ifdef DEBUG
671      printf("start bracket %d subject=", number);      printf("start bracket %d\n", number);
672        printf("subject=");
673      pchars(eptr, 16, TRUE, md);      pchars(eptr, 16, TRUE, md);
674      printf("\n");      printf("\n");
675  #endif  #endif
# Line 584  for (;;) Line 684  for (;;)
684        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
685        md->offset_vector[md->offset_end - number] = eptr - md->start_subject;        md->offset_vector[md->offset_end - number] = eptr - md->start_subject;
686    
687          flags = (op == OP_SCBRA)? match_cbegroup : 0;
688        do        do
689          {          {
690          RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb,          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
691            match_isgroup);            ims, eptrb, flags, RM1);
692          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
693          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
694          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
695          }          }
# Line 603  for (;;) Line 704  for (;;)
704        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
705        }        }
706    
707      /* Insufficient room for saving captured contents */      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
708        as a non-capturing bracket. */
709    
710      else op = OP_BRA;      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
711      }      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
712    
713    /* Other types of node can be handled by a switch */      DPRINTF(("insufficient capture room: treat as non-capturing\n"));
714    
715    switch(op)      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
716      {      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
717      case OP_BRA:     /* Non-capturing bracket: optimized */  
718      DPRINTF(("start bracket 0\n"));      /* Non-capturing bracket. Loop for all the alternatives. When we get to the
719      do      final alternative within the brackets, we would return the result of a
720        recursive call to match() whatever happened. We can reduce stack usage by
721        turning this into a tail recursion, except in the case when match_cbegroup
722        is set.*/
723    
724        case OP_BRA:
725        case OP_SBRA:
726        DPRINTF(("start non-capturing bracket\n"));
727        flags = (op >= OP_SBRA)? match_cbegroup : 0;
728        for (;;)
729        {        {
730        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb,        if (ecode[GET(ecode, 1)] != OP_ALT)   /* Final alternative */
731          match_isgroup);          {
732        if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (flags == 0)    /* Not a possibly empty group */
733              {
734              ecode += _pcre_OP_lengths[*ecode];
735              DPRINTF(("bracket 0 tail recursion\n"));
736              goto TAIL_RECURSE;
737              }
738    
739            /* Possibly empty group; can't use tail recursion. */
740    
741            RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
742              eptrb, flags, RM48);
743            RRETURN(rrc);
744            }
745    
746          /* For non-final alternatives, continue the loop for a NOMATCH result;
747          otherwise return. */
748    
749          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
750            eptrb, flags, RM2);
751          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
752        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
753        }        }
754      while (*ecode == OP_ALT);      /* Control never reaches here. */
     DPRINTF(("bracket 0 failed\n"));  
     RRETURN(MATCH_NOMATCH);  
755    
756      /* Conditional group: compilation checked that there are no more than      /* Conditional group: compilation checked that there are no more than
757      two branches. If the condition is false, skipping the first branch takes us      two branches. If the condition is false, skipping the first branch takes us
758      past the end if there is only one branch, but that's OK because that is      past the end if there is only one branch, but that's OK because that is
759      exactly what going to the ket would do. */      exactly what going to the ket would do. As there is only one branch to be
760        obeyed, we can use tail recursion to avoid using another stack frame. */
761    
762      case OP_COND:      case OP_COND:
763      if (ecode[LINK_SIZE+1] == OP_CREF) /* Condition extract or recurse test */      case OP_SCOND:
764        if (ecode[LINK_SIZE+1] == OP_RREF)         /* Recursion test */
765          {
766          offset = GET2(ecode, LINK_SIZE + 2);     /* Recursion group number*/
767          condition = md->recursive != NULL &&
768            (offset == RREF_ANY || offset == md->recursive->group_num);
769          ecode += condition? 3 : GET(ecode, 1);
770          }
771    
772        else if (ecode[LINK_SIZE+1] == OP_CREF)    /* Group used test */
773        {        {
774        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */
775        condition = (offset == CREF_RECURSE * 2)?        condition = offset < offset_top && md->offset_vector[offset] >= 0;
776          (md->recursive != NULL) :        ecode += condition? 3 : GET(ecode, 1);
777          (offset < offset_top && md->offset_vector[offset] >= 0);        }
778        RMATCH(rrc, eptr, ecode + (condition?  
779          (LINK_SIZE + 4) : (LINK_SIZE + 1 + GET(ecode, 1))),      else if (ecode[LINK_SIZE+1] == OP_DEF)     /* DEFINE - always false */
780          offset_top, md, ims, eptrb, match_isgroup);        {
781        RRETURN(rrc);        condition = FALSE;
782          ecode += GET(ecode, 1);
783        }        }
784    
785      /* The condition is an assertion. Call match() to evaluate it - setting      /* The condition is an assertion. Call match() to evaluate it - setting
786      the final argument TRUE causes it to stop at the end of an assertion. */      the final argument match_condassert causes it to stop at the end of an
787        assertion. */
788    
789      else      else
790        {        {
791        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,
792            match_condassert | match_isgroup);            match_condassert, RM3);
793        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
794          {          {
795          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE+2);          condition = TRUE;
796            ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
797          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
798          }          }
799        else if (rrc != MATCH_NOMATCH)        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
800          {          {
801          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
802          }          }
803        else ecode += GET(ecode, 1);        else
804        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb,          {
805          match_isgroup);          condition = FALSE;
806        RRETURN(rrc);          ecode += GET(ecode, 1);
807            }
808        }        }
     /* Control never reaches here */  
809    
810      /* Skip over conditional reference or large extraction number data if      /* We are now at the branch that is to be obeyed. As there is only one,
811      encountered. */      we can use tail recursion to avoid using another stack frame, except when
812        match_cbegroup is required for an unlimited repeat of a possibly empty
813        group. If the second alternative doesn't exist, we can just plough on. */
814    
815      case OP_CREF:      if (condition || *ecode == OP_ALT)
816      case OP_BRANUMBER:        {
817      ecode += 3;        ecode += 1 + LINK_SIZE;
818          if (op == OP_SCOND)        /* Possibly empty group */
819            {
820            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, match_cbegroup, RM49);
821            RRETURN(rrc);
822            }
823          else                       /* Group must match something */
824            {
825            flags = 0;
826            goto TAIL_RECURSE;
827            }
828          }
829        else                         /* Condition false & no 2nd alternative */
830          {
831          ecode += 1 + LINK_SIZE;
832          }
833      break;      break;
834    
     /* End of the pattern. If we are in a recursion, we should restore the  
     offsets appropriately and continue from after the call. */  
835    
836        /* End of the pattern, either real or forced. If we are in a top-level
837        recursion, we should restore the offsets appropriately and continue from
838        after the call. */
839    
840        case OP_ACCEPT:
841      case OP_END:      case OP_END:
842      if (md->recursive != NULL && md->recursive->group_num == 0)      if (md->recursive != NULL && md->recursive->group_num == 0)
843        {        {
844        recursion_info *rec = md->recursive;        recursion_info *rec = md->recursive;
845        DPRINTF(("Hit the end in a (?0) recursion\n"));        DPRINTF(("End of pattern in a (?0) recursion\n"));
846        md->recursive = rec->prevrec;        md->recursive = rec->prevrec;
847        memmove(md->offset_vector, rec->offset_save,        memmove(md->offset_vector, rec->offset_save,
848          rec->saved_max * sizeof(int));          rec->saved_max * sizeof(int));
849        md->start_match = rec->save_start;        mstart = rec->save_start;
850        ims = original_ims;        ims = original_ims;
851        ecode = rec->after_call;        ecode = rec->after_call;
852        break;        break;
# Line 694  for (;;) Line 855  for (;;)
855      /* Otherwise, if PCRE_NOTEMPTY is set, fail if we have matched an empty      /* Otherwise, if PCRE_NOTEMPTY is set, fail if we have matched an empty
856      string - backtracking will then try other alternatives, if any. */      string - backtracking will then try other alternatives, if any. */
857    
858      if (md->notempty && eptr == md->start_match) RRETURN(MATCH_NOMATCH);      if (md->notempty && eptr == mstart) RRETURN(MATCH_NOMATCH);
859      md->end_match_ptr = eptr;          /* Record where we ended */      md->end_match_ptr = eptr;           /* Record where we ended */
860      md->end_offset_top = offset_top;   /* and how many extracts were taken */      md->end_offset_top = offset_top;    /* and how many extracts were taken */
861        md->start_match_ptr = mstart;       /* and the start (\K can modify) */
862      RRETURN(MATCH_MATCH);      RRETURN(MATCH_MATCH);
863    
864      /* Change option settings */      /* Change option settings */
# Line 717  for (;;) Line 879  for (;;)
879      case OP_ASSERTBACK:      case OP_ASSERTBACK:
880      do      do
881        {        {
882        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
883          match_isgroup);          RM4);
884        if (rrc == MATCH_MATCH) break;        if (rrc == MATCH_MATCH) break;
885        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
886        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
887        }        }
888      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 744  for (;;) Line 906  for (;;)
906      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
907      do      do
908        {        {
909        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
910          match_isgroup);          RM5);
911        if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH);
912        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
913        ecode += GET(ecode,1);        ecode += GET(ecode,1);
914        }        }
915      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 766  for (;;) Line 928  for (;;)
928  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
929      if (utf8)      if (utf8)
930        {        {
931        c = GET(ecode,1);        i = GET(ecode, 1);
932        for (i = 0; i < c; i++)        while (i-- > 0)
933          {          {
934          eptr--;          eptr--;
935          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
936          BACKCHAR(eptr)          BACKCHAR(eptr);
937          }          }
938        }        }
939      else      else
# Line 780  for (;;) Line 942  for (;;)
942      /* No UTF-8 support, or not in UTF-8 mode: count is byte count */      /* No UTF-8 support, or not in UTF-8 mode: count is byte count */
943    
944        {        {
945        eptr -= GET(ecode,1);        eptr -= GET(ecode, 1);
946        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
947        }        }
948    
# Line 800  for (;;) Line 962  for (;;)
962        cb.version          = 1;   /* Version 1 of the callout block */        cb.version          = 1;   /* Version 1 of the callout block */
963        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
964        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
965        cb.subject          = (const char *)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
966        cb.subject_length   = md->end_subject - md->start_subject;        cb.subject_length   = md->end_subject - md->start_subject;
967        cb.start_match      = md->start_match - md->start_subject;        cb.start_match      = mstart - md->start_subject;
968        cb.current_position = eptr - md->start_subject;        cb.current_position = eptr - md->start_subject;
969        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
970        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
# Line 837  for (;;) Line 999  for (;;)
999      case OP_RECURSE:      case OP_RECURSE:
1000        {        {
1001        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1002        new_recursive.group_num = *callpat - OP_BRA;        new_recursive.group_num = (callpat == md->start_code)? 0 :
1003            GET2(callpat, 1 + LINK_SIZE);
       /* For extended extraction brackets (large number), we have to fish out  
       the number from a dummy opcode at the start. */  
   
       if (new_recursive.group_num > EXTRACT_BASIC_MAX)  
         new_recursive.group_num = GET2(callpat, 2+LINK_SIZE);  
1004    
1005        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1006    
# Line 869  for (;;) Line 1026  for (;;)
1026    
1027        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1028              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1029        new_recursive.save_start = md->start_match;        new_recursive.save_start = mstart;
1030        md->start_match = eptr;        mstart = eptr;
1031    
1032        /* OK, now we can do the recursion. For each top-level alternative we        /* OK, now we can do the recursion. For each top-level alternative we
1033        restore the offset and recursion data. */        restore the offset and recursion data. */
1034    
1035        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1036          flags = (*callpat >= OP_SBRA)? match_cbegroup : 0;
1037        do        do
1038          {          {
1039          RMATCH(rrc, eptr, callpat + 1 + LINK_SIZE, offset_top, md, ims,          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,
1040              eptrb, match_isgroup);            md, ims, eptrb, flags, RM6);
1041          if (rrc == MATCH_MATCH)          if (rrc == MATCH_MATCH)
1042            {            {
1043              DPRINTF(("Recursion matched\n"));
1044            md->recursive = new_recursive.prevrec;            md->recursive = new_recursive.prevrec;
1045            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1046              (pcre_free)(new_recursive.offset_save);              (pcre_free)(new_recursive.offset_save);
1047            RRETURN(MATCH_MATCH);            RRETURN(MATCH_MATCH);
1048            }            }
1049          else if (rrc != MATCH_NOMATCH) RRETURN(rrc);          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1050              {
1051              DPRINTF(("Recursion gave error %d\n", rrc));
1052              RRETURN(rrc);
1053              }
1054    
1055          md->recursive = &new_recursive;          md->recursive = &new_recursive;
1056          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
# Line 912  for (;;) Line 1075  for (;;)
1075      the end of a normal bracket, leaving the subject pointer. */      the end of a normal bracket, leaving the subject pointer. */
1076    
1077      case OP_ONCE:      case OP_ONCE:
1078        {      prev = ecode;
1079        prev = ecode;      saved_eptr = eptr;
       saved_eptr = eptr;  
1080    
1081        do      do
1082          {        {
1083          RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM7);
1084            eptrb, match_isgroup);        if (rrc == MATCH_MATCH) break;
1085          if (rrc == MATCH_MATCH) break;        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1086          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        ecode += GET(ecode,1);
1087          ecode += GET(ecode,1);        }
1088          }      while (*ecode == OP_ALT);
       while (*ecode == OP_ALT);  
1089    
1090        /* If hit the end of the group (which could be repeated), fail */      /* If hit the end of the group (which could be repeated), fail */
1091    
1092        if (*ecode != OP_ONCE && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);      if (*ecode != OP_ONCE && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);
1093    
1094        /* Continue as from after the assertion, updating the offsets high water      /* Continue as from after the assertion, updating the offsets high water
1095        mark, since extracts may have been taken. */      mark, since extracts may have been taken. */
1096    
1097        do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode, 1); while (*ecode == OP_ALT);
1098    
1099        offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1100        eptr = md->end_match_ptr;      eptr = md->end_match_ptr;
1101    
1102        /* For a non-repeating ket, just continue at this level. This also      /* For a non-repeating ket, just continue at this level. This also
1103        happens for a repeating ket if no characters were matched in the group.      happens for a repeating ket if no characters were matched in the group.
1104        This is the forcible breaking of infinite loops as implemented in Perl      This is the forcible breaking of infinite loops as implemented in Perl
1105        5.005. If there is an options reset, it will get obeyed in the normal      5.005. If there is an options reset, it will get obeyed in the normal
1106        course of events. */      course of events. */
1107    
1108        if (*ecode == OP_KET || eptr == saved_eptr)      if (*ecode == OP_KET || eptr == saved_eptr)
1109          {        {
1110          ecode += 1+LINK_SIZE;        ecode += 1+LINK_SIZE;
1111          break;        break;
1112          }        }
1113    
1114        /* The repeating kets try the rest of the pattern or restart from the      /* The repeating kets try the rest of the pattern or restart from the
1115        preceding bracket, in the appropriate order. We need to reset any options      preceding bracket, in the appropriate order. The second "call" of match()
1116        that changed within the bracket before re-running it, so check the next      uses tail recursion, to avoid using another stack frame. We need to reset
1117        opcode. */      any options that changed within the bracket before re-running it, so
1118        check the next opcode. */
1119    
1120        if (ecode[1+LINK_SIZE] == OP_OPT)      if (ecode[1+LINK_SIZE] == OP_OPT)
1121          {        {
1122          ims = (ims & ~PCRE_IMS) | ecode[4];        ims = (ims & ~PCRE_IMS) | ecode[4];
1123          DPRINTF(("ims set to %02lx at group repeat\n", ims));        DPRINTF(("ims set to %02lx at group repeat\n", ims));
1124          }        }
1125    
1126        if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1127          {        {
1128          RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM8);
1129          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1130          RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);        ecode = prev;
1131          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        flags = 0;
1132          }        goto TAIL_RECURSE;
       else  /* OP_KETRMAX */  
         {  
         RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);  
         if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
         RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0);  
         if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
         }  
1133        }        }
1134      RRETURN(MATCH_NOMATCH);      else  /* OP_KETRMAX */
1135          {
1136          RMATCH(eptr, prev, offset_top, md, ims, eptrb, match_cbegroup, RM9);
1137          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1138          ecode += 1 + LINK_SIZE;
1139          flags = 0;
1140          goto TAIL_RECURSE;
1141          }
1142        /* Control never gets here */
1143    
1144      /* An alternation is the end of a branch; scan along to find the end of the      /* An alternation is the end of a branch; scan along to find the end of the
1145      bracketed group and go to there. */      bracketed group and go to there. */
# Line 985  for (;;) Line 1148  for (;;)
1148      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1149      break;      break;
1150    
1151      /* BRAZERO and BRAMINZERO occur just before a bracket group, indicating      /* BRAZERO, BRAMINZERO and SKIPZERO occur just before a bracket group,
1152      that it may occur zero times. It may repeat infinitely, or not at all -      indicating that it may occur zero times. It may repeat infinitely, or not
1153      i.e. it could be ()* or ()? in the pattern. Brackets with fixed upper      at all - i.e. it could be ()* or ()? or even (){0} in the pattern. Brackets
1154      repeat limits are compiled as a number of copies, with the optional ones      with fixed upper repeat limits are compiled as a number of copies, with the
1155      preceded by BRAZERO or BRAMINZERO. */      optional ones preceded by BRAZERO or BRAMINZERO. */
1156    
1157      case OP_BRAZERO:      case OP_BRAZERO:
1158        {        {
1159        next = ecode+1;        next = ecode+1;
1160        RMATCH(rrc, eptr, next, offset_top, md, ims, eptrb, match_isgroup);        RMATCH(eptr, next, offset_top, md, ims, eptrb, 0, RM10);
1161        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1162        do next += GET(next,1); while (*next == OP_ALT);        do next += GET(next,1); while (*next == OP_ALT);
1163        ecode = next + 1+LINK_SIZE;        ecode = next + 1 + LINK_SIZE;
1164        }        }
1165      break;      break;
1166    
1167      case OP_BRAMINZERO:      case OP_BRAMINZERO:
1168        {        {
1169        next = ecode+1;        next = ecode+1;
1170        do next += GET(next,1); while (*next == OP_ALT);        do next += GET(next, 1); while (*next == OP_ALT);
1171        RMATCH(rrc, eptr, next + 1+LINK_SIZE, offset_top, md, ims, eptrb,        RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0, RM11);
         match_isgroup);  
1172        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1173        ecode++;        ecode++;
1174        }        }
1175      break;      break;
1176    
1177      /* End of a group, repeated or non-repeating. If we are at the end of      case OP_SKIPZERO:
1178      an assertion "group", stop matching and return MATCH_MATCH, but record the        {
1179      current high water mark for use by positive assertions. Do this also        next = ecode+1;
1180      for the "once" (not-backup up) groups. */        do next += GET(next,1); while (*next == OP_ALT);
1181          ecode = next + 1 + LINK_SIZE;
1182          }
1183        break;
1184    
1185        /* End of a group, repeated or non-repeating. */
1186    
1187      case OP_KET:      case OP_KET:
1188      case OP_KETRMIN:      case OP_KETRMIN:
1189      case OP_KETRMAX:      case OP_KETRMAX:
1190        {      prev = ecode - GET(ecode, 1);
       prev = ecode - GET(ecode, 1);  
       saved_eptr = eptrb->epb_saved_eptr;  
1191    
1192        /* Back up the stack of bracket start pointers. */      /* If this was a group that remembered the subject start, in order to break
1193        infinite repeats of empty string matches, retrieve the subject start from
1194        the chain. Otherwise, set it NULL. */
1195    
1196        eptrb = eptrb->epb_prev;      if (*prev >= OP_SBRA)
1197          {
1198        if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||        saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */
1199            *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||        eptrb = eptrb->epb_prev;              /* Backup to previous group */
1200            *prev == OP_ONCE)        }
1201          {      else saved_eptr = NULL;
         md->end_match_ptr = eptr;      /* For ONCE */  
         md->end_offset_top = offset_top;  
         RRETURN(MATCH_MATCH);  
         }  
1202    
1203        /* In all other cases except a conditional group we have to check the      /* If we are at the end of an assertion group, stop matching and return
1204        group number back at the start and if necessary complete handling an      MATCH_MATCH, but record the current high water mark for use by positive
1205        extraction by setting the offsets and bumping the high water mark. */      assertions. Do this also for the "once" (atomic) groups. */
1206    
1207        if (*prev != OP_COND)      if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||
1208          {          *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||
1209          number = *prev - OP_BRA;          *prev == OP_ONCE)
1210          {
1211          md->end_match_ptr = eptr;      /* For ONCE */
1212          md->end_offset_top = offset_top;
1213          RRETURN(MATCH_MATCH);
1214          }
1215    
1216          /* For extended extraction brackets (large number), we have to fish out      /* For capturing groups we have to check the group number back at the start
1217          the number from a dummy opcode at the start. */      and if necessary complete handling an extraction by setting the offsets and
1218        bumping the high water mark. Note that whole-pattern recursion is coded as
1219        a recurse into group 0, so it won't be picked up here. Instead, we catch it
1220        when the OP_END is reached. Other recursion is handled here. */
1221    
1222          if (number > EXTRACT_BASIC_MAX) number = GET2(prev, 2+LINK_SIZE);      if (*prev == OP_CBRA || *prev == OP_SCBRA)
1223          offset = number << 1;        {
1224          number = GET2(prev, 1+LINK_SIZE);
1225          offset = number << 1;
1226    
1227  #ifdef DEBUG  #ifdef DEBUG
1228          printf("end bracket %d", number);        printf("end bracket %d", number);
1229          printf("\n");        printf("\n");
1230  #endif  #endif
1231    
1232          /* Test for a numbered group. This includes groups called as a result        md->capture_last = number;
1233          of recursion. Note that whole-pattern recursion is coded as a recurse        if (offset >= md->offset_max) md->offset_overflow = TRUE; else
1234          into group 0, so it won't be picked up here. Instead, we catch it when          {
1235          the OP_END is reached. */          md->offset_vector[offset] =
1236              md->offset_vector[md->offset_end - number];
1237          if (number > 0)          md->offset_vector[offset+1] = eptr - md->start_subject;
1238            {          if (offset_top <= offset) offset_top = offset + 2;
1239            md->capture_last = number;          }
1240            if (offset >= md->offset_max) md->offset_overflow = TRUE; else  
1241              {        /* Handle a recursively called group. Restore the offsets
1242              md->offset_vector[offset] =        appropriately and continue from after the call. */
1243                md->offset_vector[md->offset_end - number];  
1244              md->offset_vector[offset+1] = eptr - md->start_subject;        if (md->recursive != NULL && md->recursive->group_num == number)
1245              if (offset_top <= offset) offset_top = offset + 2;          {
1246              }          recursion_info *rec = md->recursive;
1247            DPRINTF(("Recursion (%d) succeeded - continuing\n", number));
1248            /* Handle a recursively called group. Restore the offsets          md->recursive = rec->prevrec;
1249            appropriately and continue from after the call. */          mstart = rec->save_start;
1250            memcpy(md->offset_vector, rec->offset_save,
1251            if (md->recursive != NULL && md->recursive->group_num == number)            rec->saved_max * sizeof(int));
1252              {          ecode = rec->after_call;
1253              recursion_info *rec = md->recursive;          ims = original_ims;
1254              DPRINTF(("Recursion (%d) succeeded - continuing\n", number));          break;
             md->recursive = rec->prevrec;  
             md->start_match = rec->save_start;  
             memcpy(md->offset_vector, rec->offset_save,  
               rec->saved_max * sizeof(int));  
             ecode = rec->after_call;  
             ims = original_ims;  
             break;  
             }  
           }  
1255          }          }
1256          }
1257    
1258        /* Reset the value of the ims flags, in case they got changed during      /* For both capturing and non-capturing groups, reset the value of the ims
1259        the group. */      flags, in case they got changed during the group. */
1260    
1261        ims = original_ims;      ims = original_ims;
1262        DPRINTF(("ims reset to %02lx\n", ims));      DPRINTF(("ims reset to %02lx\n", ims));
1263    
1264        /* For a non-repeating ket, just continue at this level. This also      /* For a non-repeating ket, just continue at this level. This also
1265        happens for a repeating ket if no characters were matched in the group.      happens for a repeating ket if no characters were matched in the group.
1266        This is the forcible breaking of infinite loops as implemented in Perl      This is the forcible breaking of infinite loops as implemented in Perl
1267        5.005. If there is an options reset, it will get obeyed in the normal      5.005. If there is an options reset, it will get obeyed in the normal
1268        course of events. */      course of events. */
1269    
1270        if (*ecode == OP_KET || eptr == saved_eptr)      if (*ecode == OP_KET || eptr == saved_eptr)
1271          {        {
1272          ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1273          break;        break;
1274          }        }
1275    
1276        /* The repeating kets try the rest of the pattern or restart from the      /* The repeating kets try the rest of the pattern or restart from the
1277        preceding bracket, in the appropriate order. */      preceding bracket, in the appropriate order. In the second case, we can use
1278        tail recursion to avoid using another stack frame, unless we have an
1279        unlimited repeat of a group that can match an empty string. */
1280    
1281        if (*ecode == OP_KETRMIN)      flags = (*prev >= OP_SBRA)? match_cbegroup : 0;
1282          {  
1283          RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0);      if (*ecode == OP_KETRMIN)
1284          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        {
1285          RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM12);
1286          if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1287          }        if (flags != 0)    /* Could match an empty string */
       else  /* OP_KETRMAX */  
1288          {          {
1289          RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);          RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM50);
1290          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          RRETURN(rrc);
         RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0);  
         if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
1291          }          }
1292          ecode = prev;
1293          goto TAIL_RECURSE;
1294        }        }
1295        else  /* OP_KETRMAX */
1296      RRETURN(MATCH_NOMATCH);        {
1297          RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13);
1298          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1299          ecode += 1 + LINK_SIZE;
1300          flags = 0;
1301          goto TAIL_RECURSE;
1302          }
1303        /* Control never gets here */
1304    
1305      /* Start of subject unless notbol, or after internal newline if multiline */      /* Start of subject unless notbol, or after internal newline if multiline */
1306    
# Line 1135  for (;;) Line 1308  for (;;)
1308      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
1309      if ((ims & PCRE_MULTILINE) != 0)      if ((ims & PCRE_MULTILINE) != 0)
1310        {        {
1311        if (eptr != md->start_subject && eptr[-1] != NEWLINE)        if (eptr != md->start_subject &&
1312              (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
1313          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
1314        ecode++;        ecode++;
1315        break;        break;
# Line 1156  for (;;) Line 1330  for (;;)
1330      ecode++;      ecode++;
1331      break;      break;
1332    
1333        /* Reset the start of match point */
1334    
1335        case OP_SET_SOM:
1336        mstart = eptr;
1337        ecode++;
1338        break;
1339    
1340      /* Assert before internal newline if multiline, or before a terminating      /* Assert before internal newline if multiline, or before a terminating
1341      newline unless endonly is set, else end of subject unless noteol is set. */      newline unless endonly is set, else end of subject unless noteol is set. */
1342    
# Line 1163  for (;;) Line 1344  for (;;)
1344      if ((ims & PCRE_MULTILINE) != 0)      if ((ims & PCRE_MULTILINE) != 0)
1345        {        {
1346        if (eptr < md->end_subject)        if (eptr < md->end_subject)
1347          { if (*eptr != NEWLINE) RRETURN(MATCH_NOMATCH); }          { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }
1348        else        else
1349          { if (md->noteol) RRETURN(MATCH_NOMATCH); }          { if (md->noteol) RRETURN(MATCH_NOMATCH); }
1350        ecode++;        ecode++;
# Line 1174  for (;;) Line 1355  for (;;)
1355        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
1356        if (!md->endonly)        if (!md->endonly)
1357          {          {
1358          if (eptr < md->end_subject - 1 ||          if (eptr != md->end_subject &&
1359             (eptr == md->end_subject - 1 && *eptr != NEWLINE))              (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
1360            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1361          ecode++;          ecode++;
1362          break;          break;
1363          }          }
1364        }        }
1365      /* ... else fall through */      /* ... else fall through for endonly */
1366    
1367      /* End of subject assertion (\z) */      /* End of subject assertion (\z) */
1368    
# Line 1193  for (;;) Line 1374  for (;;)
1374      /* End of subject or ending \n assertion (\Z) */      /* End of subject or ending \n assertion (\Z) */
1375    
1376      case OP_EODN:      case OP_EODN:
1377      if (eptr < md->end_subject - 1 ||      if (eptr != md->end_subject &&
1378         (eptr == md->end_subject - 1 && *eptr != NEWLINE)) RRETURN(MATCH_NOMATCH);          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
1379          RRETURN(MATCH_NOMATCH);
1380      ecode++;      ecode++;
1381      break;      break;
1382    
# Line 1247  for (;;) Line 1429  for (;;)
1429      /* Match a single character type; inline for speed */      /* Match a single character type; inline for speed */
1430    
1431      case OP_ANY:      case OP_ANY:
1432      if ((ims & PCRE_DOTALL) == 0 && eptr < md->end_subject && *eptr == NEWLINE)      if ((ims & PCRE_DOTALL) == 0)
1433        RRETURN(MATCH_NOMATCH);        {
1434          if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
1435          }
1436      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);
 #ifdef SUPPORT_UTF8  
1437      if (utf8)      if (utf8)
1438        while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;        while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
 #endif  
1439      ecode++;      ecode++;
1440      break;      break;
1441    
# Line 1343  for (;;) Line 1525  for (;;)
1525      ecode++;      ecode++;
1526      break;      break;
1527    
1528        case OP_ANYNL:
1529        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1530        GETCHARINCTEST(c, eptr);
1531        switch(c)
1532          {
1533          default: RRETURN(MATCH_NOMATCH);
1534          case 0x000d:
1535          if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
1536          break;
1537    
1538          case 0x000a:
1539          break;
1540    
1541          case 0x000b:
1542          case 0x000c:
1543          case 0x0085:
1544          case 0x2028:
1545          case 0x2029:
1546          if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
1547          break;
1548          }
1549        ecode++;
1550        break;
1551    
1552        case OP_NOT_HSPACE:
1553        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1554        GETCHARINCTEST(c, eptr);
1555        switch(c)
1556          {
1557          default: break;
1558          case 0x09:      /* HT */
1559          case 0x20:      /* SPACE */
1560          case 0xa0:      /* NBSP */
1561          case 0x1680:    /* OGHAM SPACE MARK */
1562          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1563          case 0x2000:    /* EN QUAD */
1564          case 0x2001:    /* EM QUAD */
1565          case 0x2002:    /* EN SPACE */
1566          case 0x2003:    /* EM SPACE */
1567          case 0x2004:    /* THREE-PER-EM SPACE */
1568          case 0x2005:    /* FOUR-PER-EM SPACE */
1569          case 0x2006:    /* SIX-PER-EM SPACE */
1570          case 0x2007:    /* FIGURE SPACE */
1571          case 0x2008:    /* PUNCTUATION SPACE */
1572          case 0x2009:    /* THIN SPACE */
1573          case 0x200A:    /* HAIR SPACE */
1574          case 0x202f:    /* NARROW NO-BREAK SPACE */
1575          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1576          case 0x3000:    /* IDEOGRAPHIC SPACE */
1577          RRETURN(MATCH_NOMATCH);
1578          }
1579        ecode++;
1580        break;
1581    
1582        case OP_HSPACE:
1583        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1584        GETCHARINCTEST(c, eptr);
1585        switch(c)
1586          {
1587          default: RRETURN(MATCH_NOMATCH);
1588          case 0x09:      /* HT */
1589          case 0x20:      /* SPACE */
1590          case 0xa0:      /* NBSP */
1591          case 0x1680:    /* OGHAM SPACE MARK */
1592          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1593          case 0x2000:    /* EN QUAD */
1594          case 0x2001:    /* EM QUAD */
1595          case 0x2002:    /* EN SPACE */
1596          case 0x2003:    /* EM SPACE */
1597          case 0x2004:    /* THREE-PER-EM SPACE */
1598          case 0x2005:    /* FOUR-PER-EM SPACE */
1599          case 0x2006:    /* SIX-PER-EM SPACE */
1600          case 0x2007:    /* FIGURE SPACE */
1601          case 0x2008:    /* PUNCTUATION SPACE */
1602          case 0x2009:    /* THIN SPACE */
1603          case 0x200A:    /* HAIR SPACE */
1604          case 0x202f:    /* NARROW NO-BREAK SPACE */
1605          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1606          case 0x3000:    /* IDEOGRAPHIC SPACE */
1607          break;
1608          }
1609        ecode++;
1610        break;
1611    
1612        case OP_NOT_VSPACE:
1613        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1614        GETCHARINCTEST(c, eptr);
1615        switch(c)
1616          {
1617          default: break;
1618          case 0x0a:      /* LF */
1619          case 0x0b:      /* VT */
1620          case 0x0c:      /* FF */
1621          case 0x0d:      /* CR */
1622          case 0x85:      /* NEL */
1623          case 0x2028:    /* LINE SEPARATOR */
1624          case 0x2029:    /* PARAGRAPH SEPARATOR */
1625          RRETURN(MATCH_NOMATCH);
1626          }
1627        ecode++;
1628        break;
1629    
1630        case OP_VSPACE:
1631        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1632        GETCHARINCTEST(c, eptr);
1633        switch(c)
1634          {
1635          default: RRETURN(MATCH_NOMATCH);
1636          case 0x0a:      /* LF */
1637          case 0x0b:      /* VT */
1638          case 0x0c:      /* FF */
1639          case 0x0d:      /* CR */
1640          case 0x85:      /* NEL */
1641          case 0x2028:    /* LINE SEPARATOR */
1642          case 0x2029:    /* PARAGRAPH SEPARATOR */
1643          break;
1644          }
1645        ecode++;
1646        break;
1647    
1648  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
1649      /* Check the next character by Unicode property. We will get here only      /* Check the next character by Unicode property. We will get here only
1650      if the support is in the binary; otherwise a compile-time error occurs. */      if the support is in the binary; otherwise a compile-time error occurs. */
# Line 1352  for (;;) Line 1654  for (;;)
1654      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1655      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1656        {        {
1657        int chartype, rqdtype;        int chartype, script;
1658        int othercase;        int category = _pcre_ucp_findprop(c, &chartype, &script);
       int category = _pcre_ucp_findchar(c, &chartype, &othercase);  
   
       rqdtype = *(++ecode);  
       ecode++;  
1659    
1660        if (rqdtype >= 128)        switch(ecode[1])
1661          {          {
1662          if ((rqdtype - 128 != category) == (op == OP_PROP))          case PT_ANY:
1663            if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
1664            break;
1665    
1666            case PT_LAMP:
1667            if ((chartype == ucp_Lu ||
1668                 chartype == ucp_Ll ||
1669                 chartype == ucp_Lt) == (op == OP_NOTPROP))
1670            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1671          }           break;
1672        else  
1673          {          case PT_GC:
1674          if ((rqdtype != chartype) == (op == OP_PROP))          if ((ecode[2] != category) == (op == OP_PROP))
1675              RRETURN(MATCH_NOMATCH);
1676            break;
1677    
1678            case PT_PC:
1679            if ((ecode[2] != chartype) == (op == OP_PROP))
1680            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1681            break;
1682    
1683            case PT_SC:
1684            if ((ecode[2] != script) == (op == OP_PROP))
1685              RRETURN(MATCH_NOMATCH);
1686            break;
1687    
1688            default:
1689            RRETURN(PCRE_ERROR_INTERNAL);
1690          }          }
1691    
1692          ecode += 3;
1693        }        }
1694      break;      break;
1695    
# Line 1379  for (;;) Line 1700  for (;;)
1700      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1701      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1702        {        {
1703        int chartype;        int chartype, script;
1704        int othercase;        int category = _pcre_ucp_findprop(c, &chartype, &script);
       int category = _pcre_ucp_findchar(c, &chartype, &othercase);  
1705        if (category == ucp_M) RRETURN(MATCH_NOMATCH);        if (category == ucp_M) RRETURN(MATCH_NOMATCH);
1706        while (eptr < md->end_subject)        while (eptr < md->end_subject)
1707          {          {
# Line 1390  for (;;) Line 1710  for (;;)
1710            {            {
1711            GETCHARLEN(c, eptr, len);            GETCHARLEN(c, eptr, len);
1712            }            }
1713          category = _pcre_ucp_findchar(c, &chartype, &othercase);          category = _pcre_ucp_findprop(c, &chartype, &script);
1714          if (category != ucp_M) break;          if (category != ucp_M) break;
1715          eptr += len;          eptr += len;
1716          }          }
# Line 1480  for (;;) Line 1800  for (;;)
1800          {          {
1801          for (fi = min;; fi++)          for (fi = min;; fi++)
1802            {            {
1803            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM14);
1804            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1805            if (fi >= max || !match_ref(offset, eptr, length, md, ims))            if (fi >= max || !match_ref(offset, eptr, length, md, ims))
1806              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
# Line 1501  for (;;) Line 1821  for (;;)
1821            }            }
1822          while (eptr >= pp)          while (eptr >= pp)
1823            {            {
1824            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM15);
1825            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1826            eptr -= length;            eptr -= length;
1827            }            }
# Line 1606  for (;;) Line 1926  for (;;)
1926            {            {
1927            for (fi = min;; fi++)            for (fi = min;; fi++)
1928              {              {
1929              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM16);
1930              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1931              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1932              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 1626  for (;;) Line 1946  for (;;)
1946            {            {
1947            for (fi = min;; fi++)            for (fi = min;; fi++)
1948              {              {
1949              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM17);
1950              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1951              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1952              c = *eptr++;              c = *eptr++;
# Line 1663  for (;;) Line 1983  for (;;)
1983              }              }
1984            for (;;)            for (;;)
1985              {              {
1986              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM18);
1987              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1988              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
1989              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 1682  for (;;) Line 2002  for (;;)
2002              }              }
2003            while (eptr >= pp)            while (eptr >= pp)
2004              {              {
2005              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM19);
             eptr--;  
2006              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2007                eptr--;
2008              }              }
2009            }            }
2010    
# Line 1753  for (;;) Line 2073  for (;;)
2073          {          {
2074          for (fi = min;; fi++)          for (fi = min;; fi++)
2075            {            {
2076            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20);
2077            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2078            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2079            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
# Line 1777  for (;;) Line 2097  for (;;)
2097            }            }
2098          for(;;)          for(;;)
2099            {            {
2100            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);
2101            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2102            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
2103            BACKCHAR(eptr)            if (utf8) BACKCHAR(eptr);
2104            }            }
2105          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2106          }          }
# Line 1836  for (;;) Line 2156  for (;;)
2156    
2157        else        else
2158          {          {
2159          int dc;          unsigned int dc;
2160          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
2161          ecode += length;          ecode += length;
2162    
2163          /* 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
2164          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. */  
2165    
2166          if (fc != dc)          if (fc != dc)
2167            {            {
2168  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2169            int chartype;            if (dc != _pcre_ucp_othercase(fc))
           int othercase;  
           if (_pcre_ucp_findchar(fc, &chartype, &othercase) < 0 || dc != othercase)  
2170  #endif  #endif
2171              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2172            }            }
# Line 1867  for (;;) Line 2183  for (;;)
2183        }        }
2184      break;      break;
2185    
2186      /* Match a single character repeatedly; different opcodes share code. */      /* Match a single character repeatedly. */
2187    
2188      case OP_EXACT:      case OP_EXACT:
2189      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
2190      ecode += 3;      ecode += 3;
2191      goto REPEATCHAR;      goto REPEATCHAR;
2192    
2193        case OP_POSUPTO:
2194        possessive = TRUE;
2195        /* Fall through */
2196    
2197      case OP_UPTO:      case OP_UPTO:
2198      case OP_MINUPTO:      case OP_MINUPTO:
2199      min = 0;      min = 0;
# Line 1882  for (;;) Line 2202  for (;;)
2202      ecode += 3;      ecode += 3;
2203      goto REPEATCHAR;      goto REPEATCHAR;
2204    
2205        case OP_POSSTAR:
2206        possessive = TRUE;
2207        min = 0;
2208        max = INT_MAX;
2209        ecode++;
2210        goto REPEATCHAR;
2211    
2212        case OP_POSPLUS:
2213        possessive = TRUE;
2214        min = 1;
2215        max = INT_MAX;
2216        ecode++;
2217        goto REPEATCHAR;
2218    
2219        case OP_POSQUERY:
2220        possessive = TRUE;
2221        min = 0;
2222        max = 1;
2223        ecode++;
2224        goto REPEATCHAR;
2225    
2226      case OP_STAR:      case OP_STAR:
2227      case OP_MINSTAR:      case OP_MINSTAR:
2228      case OP_PLUS:      case OP_PLUS:
# Line 1913  for (;;) Line 2254  for (;;)
2254    
2255        if (length > 1)        if (length > 1)
2256          {          {
         int oclength = 0;  
         uschar occhars[8];  
   
2257  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2258          int othercase;          unsigned int othercase;
         int chartype;  
2259          if ((ims & PCRE_CASELESS) != 0 &&          if ((ims & PCRE_CASELESS) != 0 &&
2260               _pcre_ucp_findchar(fc, &chartype, &othercase) >= 0 &&              (othercase = _pcre_ucp_othercase(fc)) != NOTACHAR)
              othercase > 0)  
2261            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = _pcre_ord2utf8(othercase, occhars);
2262            else oclength = 0;
2263  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2264    
2265          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2266            {            {
2267            if (memcmp(eptr, charptr, length) == 0) eptr += length;            if (memcmp(eptr, charptr, length) == 0) eptr += length;
2268    #ifdef SUPPORT_UCP
2269            /* Need braces because of following else */            /* Need braces because of following else */
2270            else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }            else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }
2271            else            else
# Line 1935  for (;;) Line 2273  for (;;)
2273              if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);              if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);
2274              eptr += oclength;              eptr += oclength;
2275              }              }
2276    #else   /* without SUPPORT_UCP */
2277              else { RRETURN(MATCH_NOMATCH); }
2278    #endif  /* SUPPORT_UCP */
2279            }            }
2280    
2281          if (min == max) continue;          if (min == max) continue;
# Line 1943  for (;;) Line 2284  for (;;)
2284            {            {
2285            for (fi = min;; fi++)            for (fi = min;; fi++)
2286              {              {
2287              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM22);
2288              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2289              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2290              if (memcmp(eptr, charptr, length) == 0) eptr += length;              if (memcmp(eptr, charptr, length) == 0) eptr += length;
2291    #ifdef SUPPORT_UCP
2292              /* Need braces because of following else */              /* Need braces because of following else */
2293              else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }              else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }
2294              else              else
# Line 1954  for (;;) Line 2296  for (;;)
2296                if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);                if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);
2297                eptr += oclength;                eptr += oclength;
2298                }                }
2299    #else   /* without SUPPORT_UCP */
2300                else { RRETURN (MATCH_NOMATCH); }
2301    #endif  /* SUPPORT_UCP */
2302              }              }
2303            /* Control never gets here */            /* Control never gets here */
2304            }            }
2305          else  
2306            else  /* Maximize */
2307            {            {
2308            pp = eptr;            pp = eptr;
2309            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2310              {              {
2311              if (eptr > md->end_subject - length) break;              if (eptr > md->end_subject - length) break;
2312              if (memcmp(eptr, charptr, length) == 0) eptr += length;              if (memcmp(eptr, charptr, length) == 0) eptr += length;
2313    #ifdef SUPPORT_UCP
2314              else if (oclength == 0) break;              else if (oclength == 0) break;
2315              else              else
2316                {                {
2317                if (memcmp(eptr, occhars, oclength) != 0) break;                if (memcmp(eptr, occhars, oclength) != 0) break;
2318                eptr += oclength;                eptr += oclength;
2319                }                }
2320    #else   /* without SUPPORT_UCP */
2321                else break;
2322    #endif  /* SUPPORT_UCP */
2323              }              }
2324            while (eptr >= pp)  
2325              if (possessive) continue;
2326              for(;;)
2327             {             {
2328             RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);             RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM23);
2329             if (rrc != MATCH_NOMATCH) RRETURN(rrc);             if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2330               if (eptr == pp) RRETURN(MATCH_NOMATCH);
2331    #ifdef SUPPORT_UCP
2332               eptr--;
2333               BACKCHAR(eptr);
2334    #else   /* without SUPPORT_UCP */
2335             eptr -= length;             eptr -= length;
2336    #endif  /* SUPPORT_UCP */
2337             }             }
           RRETURN(MATCH_NOMATCH);  
2338            }            }
2339          /* Control never gets here */          /* Control never gets here */
2340          }          }
# Line 2017  for (;;) Line 2374  for (;;)
2374          {          {
2375          for (fi = min;; fi++)          for (fi = min;; fi++)
2376            {            {
2377            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM24);
2378            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2379            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
2380                fc != md->lcc[*eptr++])                fc != md->lcc[*eptr++])
# Line 2025  for (;;) Line 2382  for (;;)
2382            }            }
2383          /* Control never gets here */          /* Control never gets here */
2384          }          }
2385        else        else  /* Maximize */
2386          {          {
2387          pp = eptr;          pp = eptr;
2388          for (i = min; i < max; i++)          for (i = min; i < max; i++)
# Line 2033  for (;;) Line 2390  for (;;)
2390            if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break;            if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break;
2391            eptr++;            eptr++;
2392            }            }
2393            if (possessive) continue;
2394          while (eptr >= pp)          while (eptr >= pp)
2395            {            {
2396            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM25);
2397            eptr--;            eptr--;
2398            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2399            }            }
# Line 2054  for (;;) Line 2412  for (;;)
2412          {          {
2413          for (fi = min;; fi++)          for (fi = min;; fi++)
2414            {            {
2415            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM26);
2416            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2417            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)
2418              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2419            }            }
2420          /* Control never gets here */          /* Control never gets here */
2421          }          }
2422        else        else  /* Maximize */
2423          {          {
2424          pp = eptr;          pp = eptr;
2425          for (i = min; i < max; i++)          for (i = min; i < max; i++)
# Line 2069  for (;;) Line 2427  for (;;)
2427            if (eptr >= md->end_subject || fc != *eptr) break;            if (eptr >= md->end_subject || fc != *eptr) break;
2428            eptr++;            eptr++;
2429            }            }
2430            if (possessive) continue;
2431          while (eptr >= pp)          while (eptr >= pp)
2432            {            {
2433            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27);
2434            eptr--;            eptr--;
2435            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2436            }            }
# Line 2121  for (;;) Line 2480  for (;;)
2480      ecode += 3;      ecode += 3;
2481      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
2482    
2483        case OP_NOTPOSSTAR:
2484        possessive = TRUE;
2485        min = 0;
2486        max = INT_MAX;
2487        ecode++;
2488        goto REPEATNOTCHAR;
2489    
2490        case OP_NOTPOSPLUS:
2491        possessive = TRUE;
2492        min = 1;
2493        max = INT_MAX;
2494        ecode++;
2495        goto REPEATNOTCHAR;
2496    
2497        case OP_NOTPOSQUERY:
2498        possessive = TRUE;
2499        min = 0;
2500        max = 1;
2501        ecode++;
2502        goto REPEATNOTCHAR;
2503    
2504        case OP_NOTPOSUPTO:
2505        possessive = TRUE;
2506        min = 0;
2507        max = GET2(ecode, 1);
2508        ecode += 3;
2509        goto REPEATNOTCHAR;
2510    
2511      case OP_NOTSTAR:      case OP_NOTSTAR:
2512      case OP_NOTMINSTAR:      case OP_NOTMINSTAR:
2513      case OP_NOTPLUS:      case OP_NOTPLUS:
# Line 2160  for (;;) Line 2547  for (;;)
2547        /* UTF-8 mode */        /* UTF-8 mode */
2548        if (utf8)        if (utf8)
2549          {          {
2550          register int d;          register unsigned int d;
2551          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2552            {            {
2553            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
# Line 2185  for (;;) Line 2572  for (;;)
2572          /* UTF-8 mode */          /* UTF-8 mode */
2573          if (utf8)          if (utf8)
2574            {            {
2575            register int d;            register unsigned int d;
2576            for (fi = min;; fi++)            for (fi = min;; fi++)
2577              {              {
2578              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28);
2579              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2580              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2581              if (d < 256) d = md->lcc[d];              if (d < 256) d = md->lcc[d];
# Line 2202  for (;;) Line 2589  for (;;)
2589            {            {
2590            for (fi = min;; fi++)            for (fi = min;; fi++)
2591              {              {
2592              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM29);
2593              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2594              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])
2595                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2221  for (;;) Line 2608  for (;;)
2608          /* UTF-8 mode */          /* UTF-8 mode */
2609          if (utf8)          if (utf8)
2610            {            {
2611            register int d;            register unsigned int d;
2612            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2613              {              {
2614              int len = 1;              int len = 1;
# Line 2231  for (;;) Line 2618  for (;;)
2618              if (fc == d) break;              if (fc == d) break;
2619              eptr += len;              eptr += len;
2620              }              }
2621            for(;;)          if (possessive) continue;
2622            for(;;)
2623              {              {
2624              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM30);
2625              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2626              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2627              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2248  for (;;) Line 2636  for (;;)
2636              if (eptr >= md->end_subject || fc == md->lcc[*eptr]) break;              if (eptr >= md->end_subject || fc == md->lcc[*eptr]) break;
2637              eptr++;              eptr++;
2638              }              }
2639              if (possessive) continue;
2640            while (eptr >= pp)            while (eptr >= pp)
2641              {              {
2642              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM31);
2643              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2644              eptr--;              eptr--;
2645              }              }
# Line 2269  for (;;) Line 2658  for (;;)
2658        /* UTF-8 mode */        /* UTF-8 mode */
2659        if (utf8)        if (utf8)
2660          {          {
2661          register int d;          register unsigned int d;
2662          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2663            {            {
2664            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
# Line 2292  for (;;) Line 2681  for (;;)
2681          /* UTF-8 mode */          /* UTF-8 mode */
2682          if (utf8)          if (utf8)
2683            {            {
2684            register int d;            register unsigned int d;
2685            for (fi = min;; fi++)            for (fi = min;; fi++)
2686              {              {
2687              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32);
2688              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2689              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2690              if (fi >= max || eptr >= md->end_subject || fc == d)              if (fi >= max || eptr >= md->end_subject || fc == d)
# Line 2308  for (;;) Line 2697  for (;;)
2697            {            {
2698            for (fi = min;; fi++)            for (fi = min;; fi++)
2699              {              {
2700              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM33);
2701              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2702              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)
2703                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2327  for (;;) Line 2716  for (;;)
2716          /* UTF-8 mode */          /* UTF-8 mode */
2717          if (utf8)          if (utf8)
2718            {            {
2719            register int d;            register unsigned int d;
2720            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2721              {              {
2722              int len = 1;              int len = 1;
# Line 2336  for (;;) Line 2725  for (;;)
2725              if (fc == d) break;              if (fc == d) break;
2726              eptr += len;              eptr += len;
2727              }              }
2728              if (possessive) continue;
2729            for(;;)            for(;;)
2730              {              {
2731              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM34);
2732              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2733              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2734              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2353  for (;;) Line 2743  for (;;)
2743              if (eptr >= md->end_subject || fc == *eptr) break;              if (eptr >= md->end_subject || fc == *eptr) break;
2744              eptr++;              eptr++;
2745              }              }
2746              if (possessive) continue;
2747            while (eptr >= pp)            while (eptr >= pp)
2748              {              {
2749              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM35);
2750              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2751              eptr--;              eptr--;
2752              }              }
# Line 2384  for (;;) Line 2775  for (;;)
2775      ecode += 3;      ecode += 3;
2776      goto REPEATTYPE;      goto REPEATTYPE;
2777    
2778        case OP_TYPEPOSSTAR:
2779        possessive = TRUE;
2780        min = 0;
2781        max = INT_MAX;
2782        ecode++;
2783        goto REPEATTYPE;
2784    
2785        case OP_TYPEPOSPLUS:
2786        possessive = TRUE;
2787        min = 1;
2788        max = INT_MAX;
2789        ecode++;
2790        goto REPEATTYPE;
2791    
2792        case OP_TYPEPOSQUERY:
2793        possessive = TRUE;
2794        min = 0;
2795        max = 1;
2796        ecode++;
2797        goto REPEATTYPE;
2798    
2799        case OP_TYPEPOSUPTO:
2800        possessive = TRUE;
2801        min = 0;
2802        max = GET2(ecode, 1);
2803        ecode += 3;
2804        goto REPEATTYPE;
2805    
2806      case OP_TYPESTAR:      case OP_TYPESTAR:
2807      case OP_TYPEMINSTAR:      case OP_TYPEMINSTAR:
2808      case OP_TYPEPLUS:      case OP_TYPEPLUS:
# Line 2408  for (;;) Line 2827  for (;;)
2827        {        {
2828        prop_fail_result = ctype == OP_NOTPROP;        prop_fail_result = ctype == OP_NOTPROP;
2829        prop_type = *ecode++;        prop_type = *ecode++;
2830        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;  
         }  
2831        }        }
2832      else prop_type = -1;      else prop_type = -1;
2833  #endif  #endif
# Line 2434  for (;;) Line 2844  for (;;)
2844      if (min > 0)      if (min > 0)
2845        {        {
2846  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2847        if (prop_type > 0)        if (prop_type >= 0)
2848          {          {
2849          for (i = 1; i <= min; i++)          switch(prop_type)
2850            {            {
2851            GETCHARINC(c, eptr);            case PT_ANY:
2852            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);            if (prop_fail_result) RRETURN(MATCH_NOMATCH);
2853            if ((*prop_test_variable == prop_test_against) == prop_fail_result)            for (i = 1; i <= min; i++)
2854              RRETURN(MATCH_NOMATCH);              {
2855                if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2856                GETCHARINCTEST(c, eptr);
2857                }
2858              break;
2859    
2860              case PT_LAMP:
2861              for (i = 1; i <= min; i++)
2862                {
2863                if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2864                GETCHARINCTEST(c, eptr);
2865                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2866                if ((prop_chartype == ucp_Lu ||
2867                     prop_chartype == ucp_Ll ||
2868                     prop_chartype == ucp_Lt) == prop_fail_result)
2869                  RRETURN(MATCH_NOMATCH);
2870                }
2871              break;
2872    
2873              case PT_GC:
2874              for (i = 1; i <= min; i++)
2875                {
2876                if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2877                GETCHARINCTEST(c, eptr);
2878                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2879                if ((prop_category == prop_value) == prop_fail_result)
2880                  RRETURN(MATCH_NOMATCH);
2881                }
2882              break;
2883    
2884              case PT_PC:
2885              for (i = 1; i <= min; i++)
2886                {
2887                if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2888                GETCHARINCTEST(c, eptr);
2889                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2890                if ((prop_chartype == prop_value) == prop_fail_result)
2891                  RRETURN(MATCH_NOMATCH);
2892                }
2893              break;
2894    
2895              case PT_SC:
2896              for (i = 1; i <= min; i++)
2897                {
2898                if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2899                GETCHARINCTEST(c, eptr);
2900                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2901                if ((prop_script == prop_value) == prop_fail_result)
2902                  RRETURN(MATCH_NOMATCH);
2903                }
2904              break;
2905    
2906              default:
2907              RRETURN(PCRE_ERROR_INTERNAL);
2908            }            }
2909          }          }
2910    
# Line 2453  for (;;) Line 2916  for (;;)
2916          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2917            {            {
2918            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
2919            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2920            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);
2921            while (eptr < md->end_subject)            while (eptr < md->end_subject)
2922              {              {
# Line 2462  for (;;) Line 2925  for (;;)
2925                {                {
2926                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
2927                }                }
2928              prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2929              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
2930              eptr += len;              eptr += len;
2931              }              }
# Line 2481  for (;;) Line 2944  for (;;)
2944          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2945            {            {
2946            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
2947               (*eptr++ == NEWLINE && (ims & PCRE_DOTALL) == 0))                 ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))
2948              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2949              eptr++;
2950            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
2951            }            }
2952          break;          break;
# Line 2491  for (;;) Line 2955  for (;;)
2955          eptr += min;          eptr += min;
2956          break;          break;
2957    
2958            case OP_ANYNL:
2959            for (i = 1; i <= min; i++)
2960              {
2961              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2962              GETCHARINC(c, eptr);
2963              switch(c)
2964                {
2965                default: RRETURN(MATCH_NOMATCH);
2966                case 0x000d:
2967                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
2968                break;
2969    
2970                case 0x000a:
2971                break;
2972    
2973                case 0x000b:
2974                case 0x000c:
2975                case 0x0085:
2976                case 0x2028:
2977                case 0x2029:
2978                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2979                break;
2980                }
2981              }
2982            break;
2983    
2984            case OP_NOT_HSPACE:
2985            for (i = 1; i <= min; i++)
2986              {
2987              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2988              GETCHARINC(c, eptr);
2989              switch(c)
2990                {
2991                default: break;
2992                case 0x09:      /* HT */
2993                case 0x20:      /* SPACE */
2994                case 0xa0:      /* NBSP */
2995                case 0x1680:    /* OGHAM SPACE MARK */
2996                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
2997                case 0x2000:    /* EN QUAD */
2998                case 0x2001:    /* EM QUAD */
2999                case 0x2002:    /* EN SPACE */
3000                case 0x2003:    /* EM SPACE */
3001                case 0x2004:    /* THREE-PER-EM SPACE */
3002                case 0x2005:    /* FOUR-PER-EM SPACE */
3003                case 0x2006:    /* SIX-PER-EM SPACE */
3004                case 0x2007:    /* FIGURE SPACE */
3005                case 0x2008:    /* PUNCTUATION SPACE */
3006                case 0x2009:    /* THIN SPACE */
3007                case 0x200A:    /* HAIR SPACE */
3008                case 0x202f:    /* NARROW NO-BREAK SPACE */
3009                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3010                case 0x3000:    /* IDEOGRAPHIC SPACE */
3011                RRETURN(MATCH_NOMATCH);
3012                }
3013              }
3014            break;
3015    
3016            case OP_HSPACE:
3017            for (i = 1; i <= min; i++)
3018              {
3019              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3020              GETCHARINC(c, eptr);
3021              switch(c)
3022                {
3023                default: RRETURN(MATCH_NOMATCH);
3024                case 0x09:      /* HT */
3025                case 0x20:      /* SPACE */
3026                case 0xa0:      /* NBSP */
3027                case 0x1680:    /* OGHAM SPACE MARK */
3028                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3029                case 0x2000:    /* EN QUAD */
3030                case 0x2001:    /* EM QUAD */
3031                case 0x2002:    /* EN SPACE */
3032                case 0x2003:    /* EM SPACE */
3033                case 0x2004:    /* THREE-PER-EM SPACE */
3034                case 0x2005:    /* FOUR-PER-EM SPACE */
3035                case 0x2006:    /* SIX-PER-EM SPACE */
3036                case 0x2007:    /* FIGURE SPACE */
3037                case 0x2008:    /* PUNCTUATION SPACE */
3038                case 0x2009:    /* THIN SPACE */
3039                case 0x200A:    /* HAIR SPACE */
3040                case 0x202f:    /* NARROW NO-BREAK SPACE */
3041                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3042                case 0x3000:    /* IDEOGRAPHIC SPACE */
3043                break;
3044                }
3045              }
3046            break;
3047    
3048            case OP_NOT_VSPACE:
3049            for (i = 1; i <= min; i++)
3050              {
3051              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3052              GETCHARINC(c, eptr);
3053              switch(c)
3054                {
3055                default: break;
3056                case 0x0a:      /* LF */
3057                case 0x0b:      /* VT */
3058                case 0x0c:      /* FF */
3059                case 0x0d:      /* CR */
3060                case 0x85:      /* NEL */
3061                case 0x2028:    /* LINE SEPARATOR */
3062                case 0x2029:    /* PARAGRAPH SEPARATOR */
3063                RRETURN(MATCH_NOMATCH);
3064                }
3065              }
3066            break;
3067    
3068            case OP_VSPACE:
3069            for (i = 1; i <= min; i++)
3070              {
3071              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3072              GETCHARINC(c, eptr);
3073              switch(c)
3074                {
3075                default: RRETURN(MATCH_NOMATCH);
3076                case 0x0a:      /* LF */
3077                case 0x0b:      /* VT */
3078                case 0x0c:      /* FF */
3079                case 0x0d:      /* CR */
3080                case 0x85:      /* NEL */
3081                case 0x2028:    /* LINE SEPARATOR */
3082                case 0x2029:    /* PARAGRAPH SEPARATOR */
3083                break;
3084                }
3085              }
3086            break;
3087    
3088          case OP_NOT_DIGIT:          case OP_NOT_DIGIT:
3089          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3090            {            {
# Line 2515  for (;;) Line 3109  for (;;)
3109          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3110            {            {
3111            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3112               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_space) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0))
3113              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3114            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3115            }            }
3116          break;          break;
3117    
# Line 2535  for (;;) Line 3129  for (;;)
3129          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3130            {            {
3131            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3132               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_word) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0))
3133              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3134            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3135            }            }
3136          break;          break;
3137    
# Line 2559  for (;;) Line 3153  for (;;)
3153  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF8 */
3154    
3155        /* 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
3156        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. We can assume that there are the minimum
3157          number of bytes present, as this was tested above. */
3158    
3159        switch(ctype)        switch(ctype)
3160          {          {
# Line 2567  for (;;) Line 3162  for (;;)
3162          if ((ims & PCRE_DOTALL) == 0)          if ((ims & PCRE_DOTALL) == 0)
3163            {            {
3164            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3165              if (*eptr++ == NEWLINE) RRETURN(MATCH_NOMATCH);              {
3166                if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
3167                eptr++;
3168                }
3169            }            }
3170          else eptr += min;          else eptr += min;
3171          break;          break;
# Line 2576  for (;;) Line 3174  for (;;)
3174          eptr += min;          eptr += min;
3175          break;          break;
3176    
3177            /* Because of the CRLF case, we can't assume the minimum number of
3178            bytes are present in this case. */
3179    
3180            case OP_ANYNL:
3181            for (i = 1; i <= min; i++)
3182              {
3183              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3184              switch(*eptr++)
3185                {
3186                default: RRETURN(MATCH_NOMATCH);
3187                case 0x000d:
3188                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3189                break;
3190                case 0x000a:
3191                break;
3192    
3193                case 0x000b:
3194                case 0x000c:
3195                case 0x0085:
3196                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3197                break;
3198                }
3199              }
3200            break;
3201    
3202            case OP_NOT_HSPACE:
3203            for (i = 1; i <= min; i++)
3204              {
3205              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3206              switch(*eptr++)
3207                {
3208                default: break;
3209                case 0x09:      /* HT */
3210                case 0x20:      /* SPACE */
3211                case 0xa0:      /* NBSP */
3212                RRETURN(MATCH_NOMATCH);
3213                }
3214              }
3215            break;
3216    
3217            case OP_HSPACE:
3218            for (i = 1; i <= min; i++)
3219              {
3220              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3221              switch(*eptr++)
3222                {
3223                default: RRETURN(MATCH_NOMATCH);
3224                case 0x09:      /* HT */
3225                case 0x20:      /* SPACE */
3226                case 0xa0:      /* NBSP */
3227                break;
3228                }
3229              }
3230            break;
3231    
3232            case OP_NOT_VSPACE:
3233            for (i = 1; i <= min; i++)
3234              {
3235              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3236              switch(*eptr++)
3237                {
3238                default: break;
3239                case 0x0a:      /* LF */
3240                case 0x0b:      /* VT */
3241                case 0x0c:      /* FF */
3242                case 0x0d:      /* CR */
3243                case 0x85:      /* NEL */
3244                RRETURN(MATCH_NOMATCH);
3245                }
3246              }
3247            break;
3248    
3249            case OP_VSPACE:
3250            for (i = 1; i <= min; i++)
3251              {
3252              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3253              switch(*eptr++)
3254                {
3255                default: RRETURN(MATCH_NOMATCH);
3256                case 0x0a:      /* LF */
3257                case 0x0b:      /* VT */
3258                case 0x0c:      /* FF */
3259                case 0x0d:      /* CR */
3260                case 0x85:      /* NEL */
3261                break;
3262                }
3263              }
3264            break;
3265    
3266          case OP_NOT_DIGIT:          case OP_NOT_DIGIT:
3267          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3268            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
# Line 2624  for (;;) Line 3311  for (;;)
3311      if (minimize)      if (minimize)
3312        {        {
3313  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3314        if (prop_type > 0)        if (prop_type >= 0)
3315          {          {
3316          for (fi = min;; fi++)          switch(prop_type)
3317            {            {
3318            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            case PT_ANY:
3319            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            for (fi = min;; fi++)
3320            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              {
3321            GETCHARINC(c, eptr);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM36);
3322            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3323            if ((*prop_test_variable == prop_test_against) == prop_fail_result)              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3324              RRETURN(MATCH_NOMATCH);              GETCHARINC(c, eptr);
3325                if (prop_fail_result) RRETURN(MATCH_NOMATCH);
3326                }
3327              /* Control never gets here */
3328    
3329              case PT_LAMP:
3330              for (fi = min;; fi++)
3331                {
3332                RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM37);
3333                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3334                if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3335                GETCHARINC(c, eptr);
3336                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3337                if ((prop_chartype == ucp_Lu ||
3338                     prop_chartype == ucp_Ll ||
3339                     prop_chartype == ucp_Lt) == prop_fail_result)
3340                  RRETURN(MATCH_NOMATCH);
3341                }
3342              /* Control never gets here */
3343    
3344              case PT_GC:
3345              for (fi = min;; fi++)
3346                {
3347                RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM38);
3348                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3349                if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3350                GETCHARINC(c, eptr);
3351                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3352                if ((prop_category == prop_value) == prop_fail_result)
3353                  RRETURN(MATCH_NOMATCH);
3354                }
3355              /* Control never gets here */
3356    
3357              case PT_PC:
3358              for (fi = min;; fi++)
3359                {
3360                RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM39);
3361                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3362                if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3363                GETCHARINC(c, eptr);
3364                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3365                if ((prop_chartype == prop_value) == prop_fail_result)
3366                  RRETURN(MATCH_NOMATCH);
3367                }
3368              /* Control never gets here */
3369    
3370              case PT_SC:
3371              for (fi = min;; fi++)
3372                {
3373                RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM40);
3374                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3375                if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3376                GETCHARINC(c, eptr);
3377                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3378                if ((prop_script == prop_value) == prop_fail_result)
3379                  RRETURN(MATCH_NOMATCH);
3380                }
3381              /* Control never gets here */
3382    
3383              default:
3384              RRETURN(PCRE_ERROR_INTERNAL);
3385            }            }
3386          }          }
3387    
# Line 2645  for (;;) Line 3392  for (;;)
3392          {          {
3393          for (fi = min;; fi++)          for (fi = min;; fi++)
3394            {            {
3395            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM41);
3396            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3397            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3398            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3399            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3400            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);
3401            while (eptr < md->end_subject)            while (eptr < md->end_subject)
3402              {              {
# Line 2658  for (;;) Line 3405  for (;;)
3405                {                {
3406                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3407                }                }
3408              prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3409              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3410              eptr += len;              eptr += len;
3411              }              }
# Line 2674  for (;;) Line 3421  for (;;)
3421          {          {
3422          for (fi = min;; fi++)          for (fi = min;; fi++)
3423            {            {
3424            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42);
3425            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3426            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject ||
3427                   (ctype == OP_ANY && (ims & PCRE_DOTALL) == 0 &&
3428                    IS_NEWLINE(eptr)))
3429                RRETURN(MATCH_NOMATCH);
3430    
3431            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
3432            switch(ctype)            switch(ctype)
3433              {              {
3434              case OP_ANY:              case OP_ANY:        /* This is the DOTALL case */
             if ((ims & PCRE_DOTALL) == 0 && c == NEWLINE) RRETURN(MATCH_NOMATCH);  
3435              break;              break;
3436    
3437              case OP_ANYBYTE:              case OP_ANYBYTE:
3438              break;              break;
3439    
3440                case OP_ANYNL:
3441                switch(c)
3442                  {
3443                  default: RRETURN(MATCH_NOMATCH);
3444                  case 0x000d:
3445                  if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3446                  break;
3447                  case 0x000a:
3448                  break;
3449    
3450                  case 0x000b:
3451                  case 0x000c:
3452                  case 0x0085:
3453                  case 0x2028:
3454                  case 0x2029:
3455                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3456                  break;
3457                  }
3458                break;
3459    
3460                case OP_NOT_HSPACE:
3461                switch(c)
3462                  {
3463                  default: break;
3464                  case 0x09:      /* HT */
3465                  case 0x20:      /* SPACE */
3466                  case 0xa0:      /* NBSP */
3467                  case 0x1680:    /* OGHAM SPACE MARK */
3468                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3469                  case 0x2000:    /* EN QUAD */
3470                  case 0x2001:    /* EM QUAD */
3471                  case 0x2002:    /* EN SPACE */
3472                  case 0x2003:    /* EM SPACE */
3473                  case 0x2004:    /* THREE-PER-EM SPACE */
3474                  case 0x2005:    /* FOUR-PER-EM SPACE */
3475                  case 0x2006:    /* SIX-PER-EM SPACE */
3476                  case 0x2007:    /* FIGURE SPACE */
3477                  case 0x2008:    /* PUNCTUATION SPACE */
3478                  case 0x2009:    /* THIN SPACE */
3479                  case 0x200A:    /* HAIR SPACE */
3480                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3481                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3482                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3483                  RRETURN(MATCH_NOMATCH);
3484                  }
3485                break;
3486    
3487                case OP_HSPACE:
3488                switch(c)
3489                  {
3490                  default: RRETURN(MATCH_NOMATCH);
3491                  case 0x09:      /* HT */
3492                  case 0x20:      /* SPACE */
3493                  case 0xa0:      /* NBSP */
3494                  case 0x1680:    /* OGHAM SPACE MARK */
3495                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3496                  case 0x2000:    /* EN QUAD */
3497                  case 0x2001:    /* EM QUAD */
3498                  case 0x2002:    /* EN SPACE */
3499                  case 0x2003:    /* EM SPACE */
3500                  case 0x2004:    /* THREE-PER-EM SPACE */
3501                  case 0x2005:    /* FOUR-PER-EM SPACE */
3502                  case 0x2006:    /* SIX-PER-EM SPACE */
3503                  case 0x2007:    /* FIGURE SPACE */
3504                  case 0x2008:    /* PUNCTUATION SPACE */
3505                  case 0x2009:    /* THIN SPACE */
3506                  case 0x200A:    /* HAIR SPACE */
3507                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3508                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3509                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3510                  break;
3511                  }
3512                break;
3513    
3514                case OP_NOT_VSPACE:
3515                switch(c)
3516                  {
3517                  default: break;
3518                  case 0x0a:      /* LF */
3519                  case 0x0b:      /* VT */
3520                  case 0x0c:      /* FF */
3521                  case 0x0d:      /* CR */
3522                  case 0x85:      /* NEL */
3523                  case 0x2028:    /* LINE SEPARATOR */
3524                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3525                  RRETURN(MATCH_NOMATCH);
3526                  }
3527                break;
3528    
3529                case OP_VSPACE:
3530                switch(c)
3531                  {
3532                  default: RRETURN(MATCH_NOMATCH);
3533                  case 0x0a:      /* LF */
3534                  case 0x0b:      /* VT */
3535                  case 0x0c:      /* FF */
3536                  case 0x0d:      /* CR */
3537                  case 0x85:      /* NEL */
3538                  case 0x2028:    /* LINE SEPARATOR */
3539                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3540                  break;
3541                  }
3542                break;
3543    
3544              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
3545              if (c < 256 && (md->ctypes[c] & ctype_digit) != 0)              if (c < 256 && (md->ctypes[c] & ctype_digit) != 0)
3546                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2729  for (;;) Line 3582  for (;;)
3582          {          {
3583          for (fi = min;; fi++)          for (fi = min;; fi++)
3584            {            {
3585            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43);
3586            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3587            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject ||
3588                   ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))
3589                RRETURN(MATCH_NOMATCH);
3590    
3591            c = *eptr++;            c = *eptr++;
3592            switch(ctype)            switch(ctype)
3593              {              {
3594              case OP_ANY:              case OP_ANY:   /* This is the DOTALL case */
3595              if ((ims & PCRE_DOTALL) == 0 && c == NEWLINE) RRETURN(MATCH_NOMATCH);              break;
3596    
3597                case OP_ANYBYTE:
3598                break;
3599    
3600                case OP_ANYNL:
3601                switch(c)
3602                  {
3603                  default: RRETURN(MATCH_NOMATCH);
3604                  case 0x000d:
3605                  if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3606                  break;
3607    
3608                  case 0x000a:
3609                  break;
3610    
3611                  case 0x000b:
3612                  case 0x000c:
3613                  case 0x0085:
3614                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3615                  break;
3616                  }
3617              break;              break;
3618    
3619              case OP_ANYBYTE:              case OP_NOT_HSPACE:
3620                switch(c)
3621                  {
3622                  default: break;
3623                  case 0x09:      /* HT */
3624                  case 0x20:      /* SPACE */
3625                  case 0xa0:      /* NBSP */
3626                  RRETURN(MATCH_NOMATCH);
3627                  }
3628                break;
3629    
3630                case OP_HSPACE:
3631                switch(c)
3632                  {
3633                  default: RRETURN(MATCH_NOMATCH);
3634                  case 0x09:      /* HT */
3635                  case 0x20:      /* SPACE */
3636                  case 0xa0:      /* NBSP */
3637                  break;
3638                  }
3639                break;
3640    
3641                case OP_NOT_VSPACE:
3642                switch(c)
3643                  {
3644                  default: break;
3645                  case 0x0a:      /* LF */
3646                  case 0x0b:      /* VT */
3647                  case 0x0c:      /* FF */
3648                  case 0x0d:      /* CR */
3649                  case 0x85:      /* NEL */
3650                  RRETURN(MATCH_NOMATCH);
3651                  }
3652                break;
3653    
3654                case OP_VSPACE:
3655                switch(c)
3656                  {
3657                  default: RRETURN(MATCH_NOMATCH);
3658                  case 0x0a:      /* LF */
3659                  case 0x0b:      /* VT */
3660                  case 0x0c:      /* FF */
3661                  case 0x0d:      /* CR */
3662                  case 0x85:      /* NEL */
3663                  break;
3664                  }
3665              break;              break;
3666    
3667              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
# Line 2774  for (;;) Line 3696  for (;;)
3696        /* Control never gets here */        /* Control never gets here */
3697        }        }
3698    
3699      /* 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
3700      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
3701      UTF-8 and UCP stuff separate. */      UTF-8 and UCP stuff separate. */
3702    
# Line 2783  for (;;) Line 3705  for (;;)
3705        pp = eptr;  /* Remember where we started */        pp = eptr;  /* Remember where we started */
3706    
3707  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3708        if (prop_type > 0)        if (prop_type >= 0)
3709          {          {
3710          for (i = min; i < max; i++)          switch(prop_type)
3711            {            {
3712            int len = 1;            case PT_ANY:
3713            if (eptr >= md->end_subject) break;            for (i = min; i < max; i++)
3714            GETCHARLEN(c, eptr, len);              {
3715            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              int len = 1;
3716            if ((*prop_test_variable == prop_test_against) == prop_fail_result)              if (eptr >= md->end_subject) break;
3717              break;              GETCHARLEN(c, eptr, len);
3718            eptr+= len;              if (prop_fail_result) break;
3719                eptr+= len;
3720                }
3721              break;
3722    
3723              case PT_LAMP:
3724              for (i = min; i < max; i++)
3725                {
3726                int len = 1;
3727                if (eptr >= md->end_subject) break;
3728                GETCHARLEN(c, eptr, len);
3729                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3730                if ((prop_chartype == ucp_Lu ||
3731                     prop_chartype == ucp_Ll ||
3732                     prop_chartype == ucp_Lt) == prop_fail_result)
3733                  break;
3734                eptr+= len;
3735                }
3736              break;
3737    
3738              case PT_GC:
3739              for (i = min; i < max; i++)
3740                {
3741                int len = 1;
3742                if (eptr >= md->end_subject) break;
3743                GETCHARLEN(c, eptr, len);
3744                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3745                if ((prop_category == prop_value) == prop_fail_result)
3746                  break;
3747                eptr+= len;
3748                }
3749              break;
3750    
3751              case PT_PC:
3752              for (i = min; i < max; i++)
3753                {
3754                int len = 1;
3755                if (eptr >= md->end_subject) break;
3756                GETCHARLEN(c, eptr, len);
3757                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3758                if ((prop_chartype == prop_value) == prop_fail_result)
3759                  break;
3760                eptr+= len;
3761                }
3762              break;
3763    
3764              case PT_SC:
3765              for (i = min; i < max; i++)
3766                {
3767                int len = 1;
3768                if (eptr >= md->end_subject) break;
3769                GETCHARLEN(c, eptr, len);
3770                prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3771                if ((prop_script == prop_value) == prop_fail_result)
3772                  break;
3773                eptr+= len;
3774                }
3775              break;
3776            }            }
3777    
3778          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
3779    
3780            if (possessive) continue;
3781          for(;;)          for(;;)
3782            {            {
3783            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44);
3784            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3785            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3786            BACKCHAR(eptr);            if (utf8) BACKCHAR(eptr);
3787            }            }
3788          }          }
3789    
# Line 2816  for (;;) Line 3796  for (;;)
3796            {            {
3797            if (eptr >= md->end_subject) break;            if (eptr >= md->end_subject) break;
3798            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3799            prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3800            if (prop_category == ucp_M) break;            if (prop_category == ucp_M) break;
3801            while (eptr < md->end_subject)            while (eptr < md->end_subject)
3802              {              {
# Line 2825  for (;;) Line 3805  for (;;)
3805                {                {
3806                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3807                }                }
3808              prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3809              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3810              eptr += len;              eptr += len;
3811              }              }
# Line 2833  for (;;) Line 3813  for (;;)
3813    
3814          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
3815    
3816            if (possessive) continue;
3817          for(;;)          for(;;)
3818            {            {
3819            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM45);
3820            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3821            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3822            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
3823              {              {
3824              int len = 1;              int len = 1;
             BACKCHAR(eptr);  
3825              if (!utf8) c = *eptr; else              if (!utf8) c = *eptr; else
3826                {                {
3827                  BACKCHAR(eptr);
3828                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3829                }                }
3830              prop_category = _pcre_ucp_findchar(c, &prop_chartype, &prop_othercase);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
3831              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3832              eptr--;              eptr--;
3833              }              }
# Line 2864  for (;;) Line 3845  for (;;)
3845          switch(ctype)          switch(ctype)
3846            {            {
3847            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. */  
   
3848            if (max < INT_MAX)            if (max < INT_MAX)
3849              {              {
3850              if ((ims & PCRE_DOTALL) == 0)              if ((ims & PCRE_DOTALL) == 0)
3851                {                {
3852                for (i = min; i < max; i++)                for (i = min; i < max; i++)
3853                  {                  {
3854                  if (eptr >= md->end_subject || *eptr == NEWLINE) break;                  if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
3855                  eptr++;                  eptr++;
3856                  while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                  while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3857                  }                  }
# Line 2884  for (;;) Line 3860  for (;;)
3860                {                {
3861                for (i = min; i < max; i++)                for (i = min; i < max; i++)
3862                  {                  {
3863                    if (eptr >= md->end_subject) break;
3864                  eptr++;                  eptr++;
3865                  while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                  while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3866                  }                  }
# Line 2898  for (;;) Line 3875  for (;;)
3875                {                {
3876                for (i = min; i < max; i++)                for (i = min; i < max; i++)
3877                  {                  {
3878                  if (eptr >= md->end_subject || *eptr == NEWLINE) break;                  if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
3879                  eptr++;                  eptr++;
3880                    while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3881                  }                  }
               break;  
3882                }                }
3883              else              else
3884                {                {
3885                c = max - min;                eptr = md->end_subject;
               if (c > md->end_subject - eptr) c = md->end_subject - eptr;  
               eptr += c;  
3886                }                }
3887              }              }
3888            break;            break;
# Line 2916  for (;;) Line 3891  for (;;)
3891    
3892            case OP_ANYBYTE:            case OP_ANYBYTE:
3893            c = max - min;            c = max - min;
3894            if (c > md->end_subject - eptr) c = md->end_subject - eptr;            if (c > (unsigned int)(md->end_subject - eptr))
3895                c = md->end_subject - eptr;
3896            eptr += c;            eptr += c;
3897            break;            break;
3898    
3899              case OP_ANYNL:
3900              for (i = min; i < max; i++)
3901                {
3902                int len = 1;
3903                if (eptr >= md->end_subject) break;
3904                GETCHARLEN(c, eptr, len);
3905                if (c == 0x000d)
3906                  {
3907                  if (++eptr >= md->end_subject) break;
3908                  if (*eptr == 0x000a) eptr++;
3909                  }
3910                else
3911                  {
3912                  if (c != 0x000a &&
3913                      (md->bsr_anycrlf ||
3914                       (c != 0x000b && c != 0x000c &&
3915                        c != 0x0085 && c != 0x2028 && c != 0x2029)))
3916                    break;
3917                  eptr += len;
3918                  }
3919                }
3920              break;
3921    
3922              case OP_NOT_HSPACE:
3923              case OP_HSPACE:
3924              for (i = min; i < max; i++)
3925                {
3926                BOOL gotspace;
3927                int len = 1;
3928                if (eptr >= md->end_subject) break;
3929                GETCHARLEN(c, eptr, len);
3930                switch(c)
3931                  {
3932                  default: gotspace = FALSE; break;
3933                  case 0x09:      /* HT */
3934                  case 0x20:      /* SPACE */
3935                  case 0xa0:      /* NBSP */
3936                  case 0x1680:    /* OGHAM SPACE MARK */
3937                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3938                  case 0x2000:    /* EN QUAD */
3939                  case 0x2001:    /* EM QUAD */
3940                  case 0x2002:    /* EN SPACE */
3941                  case 0x2003:    /* EM SPACE */
3942                  case 0x2004:    /* THREE-PER-EM SPACE */
3943                  case 0x2005:    /* FOUR-PER-EM SPACE */
3944                  case 0x2006:    /* SIX-PER-EM SPACE */
3945                  case 0x2007:    /* FIGURE SPACE */
3946                  case 0x2008:    /* PUNCTUATION SPACE */
3947                  case 0x2009:    /* THIN SPACE */
3948                  case 0x200A:    /* HAIR SPACE */
3949                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3950                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3951                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3952                  gotspace = TRUE;
3953                  break;
3954                  }
3955                if (gotspace == (ctype == OP_NOT_HSPACE)) break;
3956                eptr += len;
3957                }
3958              break;
3959    
3960              case OP_NOT_VSPACE:
3961              case OP_VSPACE:
3962              for (i = min; i < max; i++)
3963                {
3964                BOOL gotspace;
3965                int len = 1;
3966                if (eptr >= md->end_subject) break;
3967                GETCHARLEN(c, eptr, len);
3968                switch(c)
3969                  {
3970                  default: gotspace = FALSE; break;
3971                  case 0x0a:      /* LF */
3972                  case 0x0b:      /* VT */
3973                  case 0x0c:      /* FF */
3974                  case 0x0d:      /* CR */
3975                  case 0x85:      /* NEL */
3976                  case 0x2028:    /* LINE SEPARATOR */
3977                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3978                  gotspace = TRUE;
3979                  break;
3980                  }
3981                if (gotspace == (ctype == OP_NOT_VSPACE)) break;
3982                eptr += len;
3983                }
3984              break;
3985    
3986            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
3987            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3988              {              {
# Line 2992  for (;;) Line 4055  for (;;)
4055    
4056          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
4057    
4058            if (possessive) continue;
4059          for(;;)          for(;;)
4060            {            {
4061            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM46);
4062            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4063            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
4064            BACKCHAR(eptr);            BACKCHAR(eptr);
4065            }            }
4066          }          }
4067        else        else
4068  #endif  #endif  /* SUPPORT_UTF8 */
4069    
4070        /* Not UTF-8 mode */        /* Not UTF-8 mode */
4071          {          {
# Line 3012  for (;;) Line 4076  for (;;)
4076              {              {
4077              for (i = min; i < max; i++)              for (i = min; i < max; i++)
4078                {                {
4079                if (eptr >= md->end_subject || *eptr == NEWLINE) break;                if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
4080                eptr++;                eptr++;
4081                }                }
4082              break;              break;
# Line 3021  for (;;) Line 4085  for (;;)
4085    
4086            case OP_ANYBYTE:            case OP_ANYBYTE:
4087            c = max - min;            c = max - min;
4088            if (c > md->end_subject - eptr) c = md->end_subject - eptr;            if (c > (unsigned int)(md->end_subject - eptr))
4089                c = md->end_subject - eptr;
4090            eptr += c;            eptr += c;
4091            break;            break;
4092    
4093              case OP_ANYNL:
4094              for (i = min; i < max; i++)
4095                {
4096                if (eptr >= md->end_subject) break;
4097                c = *eptr;
4098                if (c == 0x000d)
4099                  {
4100                  if (++eptr >= md->end_subject) break;
4101                  if (*eptr == 0x000a) eptr++;
4102                  }
4103                else
4104                  {
4105                  if (c != 0x000a &&
4106                      (md->bsr_anycrlf ||
4107                        (c != 0x000b && c != 0x000c && c != 0x0085)))
4108                    break;
4109                  eptr++;
4110                  }
4111                }
4112              break;
4113    
4114              case OP_NOT_HSPACE:
4115              for (i = min; i < max; i++)
4116                {
4117                if (eptr >= md->end_subject) break;
4118                c = *eptr;
4119                if (c == 0x09 || c == 0x20 || c == 0xa0) break;
4120                eptr++;
4121                }
4122              break;
4123    
4124              case OP_HSPACE:
4125              for (i = min; i < max; i++)
4126                {
4127                if (eptr >= md->end_subject) break;
4128                c = *eptr;
4129                if (c != 0x09 && c != 0x20 && c != 0xa0) break;
4130                eptr++;
4131                }
4132              break;
4133    
4134              case OP_NOT_VSPACE:
4135              for (i = min; i < max; i++)
4136                {
4137                if (eptr >= md->end_subject) break;
4138                c = *eptr;
4139                if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85)
4140                  break;
4141                eptr++;
4142                }
4143              break;
4144    
4145              case OP_VSPACE:
4146              for (i = min; i < max; i++)
4147                {
4148                if (eptr >= md->end_subject) break;
4149                c = *eptr;
4150                if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85)
4151                  break;
4152                eptr++;
4153                }
4154              break;
4155    
4156            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
4157            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4158              {              {
# Line 3085  for (;;) Line 4213  for (;;)
4213    
4214          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
4215    
4216            if (possessive) continue;
4217          while (eptr >= pp)          while (eptr >= pp)
4218            {            {
4219            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM47);
4220            eptr--;            eptr--;
4221            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4222            }            }
# Line 3099  for (;;) Line 4228  for (;;)
4228        }        }
4229      /* Control never gets here */      /* Control never gets here */
4230    
4231      /* 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
4232      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. */  
4233    
4234      default:      default:
4235      DPRINTF(("Unknown opcode %d\n", *ecode));      DPRINTF(("Unknown opcode %d\n", *ecode));
4236      RRETURN(PCRE_ERROR_UNKNOWN_NODE);      RRETURN(PCRE_ERROR_UNKNOWN_OPCODE);
4237      }      }
4238    
4239    /* 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 4242  for (;;)
4242    
4243    }             /* End of main loop */    }             /* End of main loop */
4244  /* Control never reaches here */  /* Control never reaches here */
4245    
4246    
4247    /* When compiling to use the heap rather than the stack for recursive calls to
4248    match(), the RRETURN() macro jumps here. The number that is saved in
4249    frame->Xwhere indicates which label we actually want to return to. */
4250    
4251    #ifdef NO_RECURSE
4252    #define LBL(val) case val: goto L_RM##val;
4253    HEAP_RETURN:
4254    switch (frame->Xwhere)
4255      {
4256      LBL( 1) LBL( 2) LBL( 3) LBL( 4) LBL( 5) LBL( 6) LBL( 7) LBL( 8)
4257      LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(17)
4258      LBL(19) LBL(24) LBL(25) LBL(26) LBL(27) LBL(29) LBL(31) LBL(33)
4259      LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52)
4260      LBL(53) LBL(54)
4261    #ifdef SUPPORT_UTF8
4262      LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30)
4263      LBL(32) LBL(34) LBL(42) LBL(46)
4264    #ifdef SUPPORT_UCP
4265      LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45)
4266    #endif  /* SUPPORT_UCP */
4267    #endif  /* SUPPORT_UTF8 */
4268      default:
4269      DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));
4270      return PCRE_ERROR_INTERNAL;
4271      }
4272    #undef LBL
4273    #endif  /* NO_RECURSE */
4274  }  }
4275    
4276    
# Line 3127  Undefine all the macros that were define Line 4283  Undefine all the macros that were define
4283  #ifdef NO_RECURSE  #ifdef NO_RECURSE
4284  #undef eptr  #undef eptr
4285  #undef ecode  #undef ecode
4286    #undef mstart
4287  #undef offset_top  #undef offset_top
4288  #undef ims  #undef ims
4289  #undef eptrb  #undef eptrb
# Line 3144  Undefine all the macros that were define Line 4301  Undefine all the macros that were define
4301    
4302  #undef cur_is_word  #undef cur_is_word
4303  #undef condition  #undef condition
 #undef minimize  
4304  #undef prev_is_word  #undef prev_is_word
4305    
4306  #undef original_ims  #undef original_ims
# Line 3200  Returns:          > 0 => success; value Line 4356  Returns:          > 0 => success; value
4356                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
4357  */  */
4358    
4359  PCRE_EXPORT int  PCRE_EXP_DEFN int
4360  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
4361    const char *subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
4362    int offsetcount)    int offsetcount)
4363  {  {
4364  int rc, resetcount, ocount;  int rc, resetcount, ocount;
4365  int first_byte = -1;  int first_byte = -1;
4366  int req_byte = -1;  int req_byte = -1;
4367  int req_byte2 = -1;  int req_byte2 = -1;
4368  unsigned long int ims = 0;  int newline;
4369    unsigned long int ims;
4370  BOOL using_temporary_offsets = FALSE;  BOOL using_temporary_offsets = FALSE;
4371  BOOL anchored;  BOOL anchored;
4372  BOOL startline;  BOOL startline;
4373  BOOL firstline;  BOOL firstline;
4374  BOOL first_byte_caseless = FALSE;  BOOL first_byte_caseless = FALSE;
4375  BOOL req_byte_caseless = FALSE;  BOOL req_byte_caseless = FALSE;
4376    BOOL utf8;
4377  match_data match_block;  match_data match_block;
4378    match_data *md = &match_block;
4379  const uschar *tables;  const uschar *tables;
4380  const uschar *start_bits = NULL;  const uschar *start_bits = NULL;
4381  const uschar *start_match = (const uschar *)subject + start_offset;  USPTR start_match = (USPTR)subject + start_offset;
4382  const uschar *end_subject;  USPTR end_subject;
4383  const uschar *req_byte_ptr = start_match - 1;  USPTR req_byte_ptr = start_match - 1;
4384    
4385  pcre_study_data internal_study;  pcre_study_data internal_study;
4386  const pcre_study_data *study;  const pcre_study_data *study;
# Line 3241  if (offsetcount < 0) return PCRE_ERROR_B Line 4400  if (offsetcount < 0) return PCRE_ERROR_B
4400  the default values. */  the default values. */
4401    
4402  study = NULL;  study = NULL;
4403  match_block.match_limit = MATCH_LIMIT;  md->match_limit = MATCH_LIMIT;
4404  match_block.callout_data = NULL;  md->match_limit_recursion = MATCH_LIMIT_RECURSION;
4405    md->callout_data = NULL;
4406    
4407  /* The table pointer is always in native byte order. */  /* The table pointer is always in native byte order. */
4408    
# Line 3254  if (extra_data != NULL) Line 4414  if (extra_data != NULL)
4414    if ((flags & PCRE_EXTRA_STUDY_DATA) != 0)    if ((flags & PCRE_EXTRA_STUDY_DATA) != 0)
4415      study = (const pcre_study_data *)extra_data->study_data;      study = (const pcre_study_data *)extra_data->study_data;
4416    if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0)    if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0)
4417      match_block.match_limit = extra_data->match_limit;      md->match_limit = extra_data->match_limit;
4418      if ((flags & PCRE_EXTRA_MATCH_LIMIT_RECURSION) != 0)
4419        md->match_limit_recursion = extra_data->match_limit_recursion;
4420    if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0)    if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0)
4421      match_block.callout_data = extra_data->callout_data;      md->callout_data = extra_data->callout_data;
4422    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;
4423    }    }
4424    
# Line 3281  if (re->magic_number != MAGIC_NUMBER) Line 4443  if (re->magic_number != MAGIC_NUMBER)
4443  /* Set up other data */  /* Set up other data */
4444    
4445  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;
4446  startline = (re->options & PCRE_STARTLINE) != 0;  startline = (re->flags & PCRE_STARTLINE) != 0;
4447  firstline = (re->options & PCRE_FIRSTLINE) != 0;  firstline = (re->options & PCRE_FIRSTLINE) != 0;
4448    
4449  /* 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. */
4450    
4451  match_block.start_code = (const uschar *)external_re + re->name_table_offset +  md->start_code = (const uschar *)external_re + re->name_table_offset +
4452    re->name_count * re->name_entry_size;    re->name_count * re->name_entry_size;
4453    
4454  match_block.start_subject = (const uschar *)subject;  md->start_subject = (USPTR)subject;
4455  match_block.start_offset = start_offset;  md->start_offset = start_offset;
4456  match_block.end_subject = match_block.start_subject + length;  md->end_subject = md->start_subject + length;
4457  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;  
4458    
4459  match_block.recursive = NULL;                   /* No recursion at top level */  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;
4460    utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;
4461    
4462    md->notbol = (options & PCRE_NOTBOL) != 0;
4463    md->noteol = (options & PCRE_NOTEOL) != 0;
4464    md->notempty = (options & PCRE_NOTEMPTY) != 0;
4465    md->partial = (options & PCRE_PARTIAL) != 0;
4466    md->hitend = FALSE;
4467    
4468    md->recursive = NULL;                   /* No recursion at top level */
4469    
4470    md->lcc = tables + lcc_offset;
4471    md->ctypes = tables + ctypes_offset;
4472    
4473    /* Handle different \R options. */
4474    
4475    switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE))
4476      {
4477      case 0:
4478      if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0)
4479        md->bsr_anycrlf = (re->options & PCRE_BSR_ANYCRLF) != 0;
4480      else
4481    #ifdef BSR_ANYCRLF
4482      md->bsr_anycrlf = TRUE;
4483    #else
4484      md->bsr_anycrlf = FALSE;
4485    #endif
4486      break;
4487    
4488      case PCRE_BSR_ANYCRLF:
4489      md->bsr_anycrlf = TRUE;
4490      break;
4491    
4492      case PCRE_BSR_UNICODE:
4493      md->bsr_anycrlf = FALSE;
4494      break;
4495    
4496      default: return PCRE_ERROR_BADNEWLINE;
4497      }
4498    
4499    /* Handle different types of newline. The three bits give eight cases. If
4500    nothing is set at run time, whatever was used at compile time applies. */
4501    
4502    switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options :
4503            (pcre_uint32)options) & PCRE_NEWLINE_BITS)
4504      {
4505      case 0: newline = NEWLINE; break;   /* Compile-time default */
4506      case PCRE_NEWLINE_CR: newline = '\r'; break;
4507      case PCRE_NEWLINE_LF: newline = '\n'; break;
4508      case PCRE_NEWLINE_CR+
4509           PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;
4510      case PCRE_NEWLINE_ANY: newline = -1; break;
4511      case PCRE_NEWLINE_ANYCRLF: newline = -2; break;
4512      default: return PCRE_ERROR_BADNEWLINE;
4513      }
4514    
4515  match_block.lcc = tables + lcc_offset;  if (newline == -2)
4516  match_block.ctypes = tables + ctypes_offset;    {
4517      md->nltype = NLTYPE_ANYCRLF;
4518      }
4519    else if (newline < 0)
4520      {
4521      md->nltype = NLTYPE_ANY;
4522      }
4523    else
4524      {
4525      md->nltype = NLTYPE_FIXED;
4526      if (newline > 255)
4527        {
4528        md->nllen = 2;
4529        md->nl[0] = (newline >> 8) & 255;
4530        md->nl[1] = newline & 255;
4531        }
4532      else
4533        {
4534        md->nllen = 1;
4535        md->nl[0] = newline;
4536        }
4537      }
4538    
4539  /* 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
4540  moment. */  moment. */
4541    
4542  if (match_block.partial && (re->options & PCRE_NOPARTIAL) != 0)  if (md->partial && (re->flags & PCRE_NOPARTIAL) != 0)
4543    return PCRE_ERROR_BADPARTIAL;    return PCRE_ERROR_BADPARTIAL;
4544    
4545  /* 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
4546  back the character offset. */  back the character offset. */
4547    
4548  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4549  if (match_block.utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)
4550    {    {
4551    if (_pcre_valid_utf8((uschar *)subject, length) >= 0)    if (_pcre_valid_utf8((uschar *)subject, length) >= 0)
4552      return PCRE_ERROR_BADUTF8;      return PCRE_ERROR_BADUTF8;
# Line 3350  ocount = offsetcount - (offsetcount % 3) Line 4578  ocount = offsetcount - (offsetcount % 3)
4578  if (re->top_backref > 0 && re->top_backref >= ocount/3)  if (re->top_backref > 0 && re->top_backref >= ocount/3)
4579    {    {
4580    ocount = re->top_backref * 3 + 3;    ocount = re->top_backref * 3 + 3;
4581    match_block.offset_vector = (int *)(pcre_malloc)(ocount * sizeof(int));    md->offset_vector = (int *)(pcre_malloc)(ocount * sizeof(int));
4582    if (match_block.offset_vector == NULL) return PCRE_ERROR_NOMEMORY;    if (md->offset_vector == NULL) return PCRE_ERROR_NOMEMORY;
4583    using_temporary_offsets = TRUE;    using_temporary_offsets = TRUE;
4584    DPRINTF(("Got memory to hold back references\n"));    DPRINTF(("Got memory to hold back references\n"));
4585    }    }
4586  else match_block.offset_vector = offsets;  else md->offset_vector = offsets;
4587    
4588  match_block.offset_end = ocount;  md->offset_end = ocount;
4589  match_block.offset_max = (2*ocount)/3;  md->offset_max = (2*ocount)/3;
4590  match_block.offset_overflow = FALSE;  md->offset_overflow = FALSE;
4591  match_block.capture_last = -1;  md->capture_last = -1;
4592    
4593  /* 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
4594  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 4601  if (resetcount > offsetcount) resetcount
4601  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
4602  initialize them to avoid reading uninitialized locations. */  initialize them to avoid reading uninitialized locations. */
4603    
4604  if (match_block.offset_vector != NULL)  if (md->offset_vector != NULL)
4605    {    {
4606    register int *iptr = match_block.offset_vector + ocount;    register int *iptr = md->offset_vector + ocount;
4607    register int *iend = iptr - resetcount/2 + 1;    register int *iend = iptr - resetcount/2 + 1;
4608    while (--iptr >= iend) *iptr = -1;    while (--iptr >= iend) *iptr = -1;
4609    }    }
# Line 3388  studied, there may be a bitmap of possib Line 4616  studied, there may be a bitmap of possib
4616    
4617  if (!anchored)  if (!anchored)
4618    {    {
4619    if ((re->options & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
4620      {      {
4621      first_byte = re->first_byte & 255;      first_byte = re->first_byte & 255;
4622      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)
4623        first_byte = match_block.lcc[first_byte];        first_byte = md->lcc[first_byte];
4624      }      }
4625    else    else
4626      if (!startline && study != NULL &&      if (!startline && study != NULL &&
# Line 3403  if (!anchored) Line 4631  if (!anchored)
4631  /* For anchored or unanchored matches, there may be a "last known required  /* For anchored or unanchored matches, there may be a "last known required
4632  character" set. */  character" set. */
4633    
4634  if ((re->options & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
4635    {    {
4636    req_byte = re->req_byte & 255;    req_byte = re->req_byte & 255;
4637    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;
4638    req_byte2 = (tables + fcc_offset)[req_byte];  /* case flipped */    req_byte2 = (tables + fcc_offset)[req_byte];  /* case flipped */
4639    }    }
4640    
4641    
4642    /* ==========================================================================*/
4643    
4644  /* Loop for handling unanchored repeated matching attempts; for anchored regexs  /* Loop for handling unanchored repeated matching attempts; for anchored regexs
4645  the loop runs just once. */  the loop runs just once. */
4646    
4647  do  for(;;)
4648    {    {
4649    const uschar *save_end_subject = end_subject;    USPTR save_end_subject = end_subject;
4650      USPTR new_start_match;
4651    
4652    /* Reset the maximum number of extractions we might see. */    /* Reset the maximum number of extractions we might see. */
4653    
4654    if (match_block.offset_vector != NULL)    if (md->offset_vector != NULL)
4655      {      {
4656      register int *iptr = match_block.offset_vector;      register int *iptr = md->offset_vector;
4657      register int *iend = iptr + resetcount;      register int *iend = iptr + resetcount;
4658      while (iptr < iend) *iptr++ = -1;      while (iptr < iend) *iptr++ = -1;
4659      }      }
4660    
4661    /* 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
4662    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.
4663    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
4664    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
4665    */    the match fails at the newline, later code breaks this loop. */
4666    
4667    if (firstline)    if (firstline)
4668      {      {
4669      const uschar *t = start_match;      USPTR t = start_match;
4670      while (t < save_end_subject && *t != '\n') t++;      while (t < md->end_subject && !IS_NEWLINE(t)) t++;
4671      end_subject = t;      end_subject = t;
4672      }      }
4673    
# Line 3445  do Line 4677  do
4677      {      {
4678      if (first_byte_caseless)      if (first_byte_caseless)
4679        while (start_match < end_subject &&        while (start_match < end_subject &&
4680               match_block.lcc[*start_match] != first_byte)               md->lcc[*start_match] != first_byte)
4681          start_match++;          { NEXTCHAR(start_match); }
4682      else      else
4683        while (start_match < end_subject && *start_match != first_byte)        while (start_match < end_subject && *start_match != first_byte)
4684          start_match++;          { NEXTCHAR(start_match); }
4685      }      }
4686    
4687    /* Or to just after \n for a multiline match if possible */    /* Or to just after a linebreak for a multiline match if possible */
4688    
4689    else if (startline)    else if (startline)
4690      {      {
4691      if (start_match > match_block.start_subject + start_offset)      if (start_match > md->start_subject + start_offset)
4692        {        {
4693        while (start_match < end_subject && start_match[-1] != NEWLINE)        while (start_match <= end_subject && !WAS_NEWLINE(start_match))
4694            { NEXTCHAR(start_match); }
4695    
4696          /* If we have just passed a CR and the newline option is ANY or ANYCRLF,
4697          and we are now at a LF, advance the match position by one more character.
4698          */
4699    
4700          if (start_match[-1] == '\r' &&
4701               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
4702               start_match < end_subject &&
4703               *start_match == '\n')
4704          start_match++;          start_match++;
4705        }        }
4706      }      }
# Line 3470  do Line 4712  do
4712      while (start_match < end_subject)      while (start_match < end_subject)
4713        {        {
4714        register unsigned int c = *start_match;        register unsigned int c = *start_match;
4715        if ((start_bits[c/8] & (1 << (c&7))) == 0) start_match++; else break;        if ((start_bits[c/8] & (1 << (c&7))) == 0)
4716            { NEXTCHAR(start_match); }
4717          else break;
4718        }        }
4719      }      }
4720    
# Line 3480  do Line 4724  do
4724    
4725  #ifdef DEBUG  /* Sigh. Some compilers never learn. */  #ifdef DEBUG  /* Sigh. Some compilers never learn. */
4726    printf(">>>> Match against: ");    printf(">>>> Match against: ");
4727    pchars(start_match, end_subject - start_match, TRUE, &match_block);    pchars(start_match, end_subject - start_match, TRUE, md);
4728    printf("\n");    printf("\n");
4729  #endif  #endif
4730    
# Line 3494  do Line 4738  do
4738    
4739    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
4740    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
4741    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
4742    don't do this when the string is sufficiently long.    string... so we don't do this when the string is sufficiently long.
4743    
4744    ALSO: this processing is disabled when partial matching is requested.    ALSO: this processing is disabled when partial matching is requested.
4745    */    */
4746    
4747    if (req_byte >= 0 &&    if (req_byte >= 0 &&
4748        end_subject - start_match < REQ_BYTE_MAX &&        end_subject - start_match < REQ_BYTE_MAX &&
4749        !match_block.partial)        !md->partial)
4750      {      {
4751      register const uschar *p = start_match + ((first_byte >= 0)? 1 : 0);      register USPTR p = start_match + ((first_byte >= 0)? 1 : 0);
4752    
4753      /* 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
4754      place we found it at last time. */      place we found it at last time. */
# Line 3527  do Line 4771  do
4771            }            }
4772          }          }
4773    
4774        /* If we can't find the required character, break the matching loop */        /* If we can't find the required character, break the matching loop,
4775          forcing a match failure. */
4776    
4777