/[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 771 by ph10, Tue Nov 29 15:34:12 2011 UTC revision 1248 by ph10, Wed Feb 13 17:36:38 2013 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) 1997-2011 University of Cambridge             Copyright (c) 1997-2012 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 37  POSSIBILITY OF SUCH DAMAGE. Line 37  POSSIBILITY OF SUCH DAMAGE.
37  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
38  */  */
39    
   
40  /* This module contains pcre_exec(), the externally visible function that does  /* This module contains pcre_exec(), the externally visible function that does
41  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
42  possible. There are also some static supporting functions. */  possible. There are also some static supporting functions. */
# Line 57  possible. There are also some static sup Line 56  possible. There are also some static sup
56  #undef min  #undef min
57  #undef max  #undef max
58    
59    /* The md->capture_last field uses the lower 16 bits for the last captured
60    substring (which can never be greater than 65535) and a bit in the top half
61    to mean "capture vector overflowed". This odd way of doing things was
62    implemented when it was realized that preserving and restoring the overflow bit
63    whenever the last capture number was saved/restored made for a neater
64    interface, and doing it this way saved on (a) another variable, which would
65    have increased the stack frame size (a big NO-NO in PCRE) and (b) another
66    separate set of save/restore instructions. The following defines are used in
67    implementing this. */
68    
69    #define CAPLMASK    0x0000ffff    /* The bits used for last_capture */
70    #define OVFLMASK    0xffff0000    /* The bits used for the overflow flag */
71    #define OVFLBIT     0x00010000    /* The bit that is set for overflow */
72    
73  /* Values for setting in md->match_function_type to indicate two special types  /* Values for setting in md->match_function_type to indicate two special types
74  of call to match(). We do it this way to save on using another stack variable,  of call to match(). We do it this way to save on using another stack variable,
75  as stack usage is to be discouraged. */  as stack usage is to be discouraged. */
# Line 93  because the offset vector is always a mu Line 106  because the offset vector is always a mu
106  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };
107  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };
108    
   
   
109  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
110  /*************************************************  /*************************************************
111  *        Debugging function to print chars       *  *        Debugging function to print chars       *
# Line 113  Returns:     nothing Line 124  Returns:     nothing
124  */  */
125    
126  static void  static void
127  pchars(const uschar *p, int length, BOOL is_subject, match_data *md)  pchars(const pcre_uchar *p, int length, BOOL is_subject, match_data *md)
128  {  {
129  unsigned int c;  pcre_uint32 c;
130    BOOL utf = md->utf;
131  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;
132  while (length-- > 0)  while (length-- > 0)
133    if (isprint(c = *(p++))) printf("%c", c); else printf("\\x%02x", c);    if (isprint(c = RAWUCHARINCTEST(p))) printf("%c", (char)c); else printf("\\x{%02x}", c);
134  }  }
135  #endif  #endif
136    
# Line 140  Arguments: Line 152  Arguments:
152    md          points to match data block    md          points to match data block
153    caseless    TRUE if caseless    caseless    TRUE if caseless
154    
155  Returns:      < 0 if not matched, otherwise the number of subject bytes matched  Returns:      >= 0 the number of subject bytes matched
156                  -1 no match
157                  -2 partial match; always given if at end subject
158  */  */
159    
160  static int  static int
161  match_ref(int offset, register USPTR eptr, int length, match_data *md,  match_ref(int offset, register PCRE_PUCHAR eptr, int length, match_data *md,
162    BOOL caseless)    BOOL caseless)
163  {  {
164  USPTR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
165  register USPTR p = md->start_subject + md->offset_vector[offset];  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
166    #ifdef SUPPORT_UTF
167    BOOL utf = md->utf;
168    #endif
169    
170  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
171  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 163  pchars(p, length, FALSE, md); Line 180  pchars(p, length, FALSE, md);
180  printf("\n");  printf("\n");
181  #endif  #endif
182    
183  /* Always fail if reference not set (and not JavaScript compatible). */  /* Always fail if reference not set (and not JavaScript compatible - in that
184    case the length is passed as zero). */
185    
186  if (length < 0) return -1;  if (length < 0) return -1;
187    
# Line 173  ASCII characters. */ Line 191  ASCII characters. */
191    
192  if (caseless)  if (caseless)
193    {    {
194  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
195  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
196    if (md->utf8)    if (utf)
197      {      {
198      /* Match characters up to the end of the reference. NOTE: the number of      /* Match characters up to the end of the reference. NOTE: the number of
199      bytes matched may differ, because there are some characters whose upper and      data units matched may differ, because in UTF-8 there are some characters
200      lower case versions code as different numbers of bytes. For example, U+023A      whose upper and lower case versions code have different numbers of bytes.
201      (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);      For example, U+023A (2 bytes in UTF-8) is the upper case version of U+2C65
202      a sequence of 3 of the former uses 6 bytes, as does a sequence of two of      (3 bytes in UTF-8); a sequence of 3 of the former uses 6 bytes, as does a
203      the latter. It is important, therefore, to check the length along the      sequence of two of the latter. It is important, therefore, to check the
204      reference, not along the subject (earlier code did this wrong). */      length along the reference, not along the subject (earlier code did this
205        wrong). */
206    
207      USPTR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
208      while (p < endptr)      while (p < endptr)
209        {        {
210        int c, d;        pcre_uint32 c, d;
211        if (eptr >= md->end_subject) return -1;        const ucd_record *ur;
212          if (eptr >= md->end_subject) return -2;   /* Partial match */
213        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
214        GETCHARINC(d, p);        GETCHARINC(d, p);
215        if (c != d && c != UCD_OTHERCASE(d)) return -1;        ur = GET_UCD(d);
216          if (c != d && c != d + ur->other_case)
217            {
218            const pcre_uint32 *pp = PRIV(ucd_caseless_sets) + ur->caseset;
219            for (;;)
220              {
221              if (c < *pp) return -1;
222              if (c == *pp++) break;
223              }
224            }
225        }        }
226      }      }
227    else    else
# Line 202  if (caseless) Line 231  if (caseless)
231    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there
232    is no UCP support. */    is no UCP support. */
233      {      {
     if (eptr + length > md->end_subject) return -1;  
234      while (length-- > 0)      while (length-- > 0)
235        { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }        {
236          pcre_uchar cc, cp;
237          if (eptr >= md->end_subject) return -2;   /* Partial match */
238          cc = RAWUCHARTEST(eptr);
239          cp = RAWUCHARTEST(p);
240          if (TABLE_GET(cp, md->lcc, cp) != TABLE_GET(cc, md->lcc, cc)) return -1;
241          p++;
242          eptr++;
243          }
244      }      }
245    }    }
246    
# Line 213  are in UTF-8 mode. */ Line 249  are in UTF-8 mode. */
249    
250  else  else
251    {    {
252    if (eptr + length > md->end_subject) return -1;    while (length-- > 0)
253    while (length-- > 0) if (*p++ != *eptr++) return -1;      {
254        if (eptr >= md->end_subject) return -2;   /* Partial match */
255        if (RAWUCHARINCTEST(p) != RAWUCHARINCTEST(eptr)) return -1;
256        }
257    }    }
258    
259  return eptr - eptr_start;  return (int)(eptr - eptr_start);
260  }  }
261    
262    
# Line 269  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 308  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
308         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
309         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
310         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
311         RM61,  RM62, RM63, RM64, RM65, RM66 };         RM61,  RM62, RM63, RM64, RM65, RM66, RM67 };
312    
313  /* These versions of the macros use the stack, as normal. There are debugging  /* These versions of the macros use the stack, as normal. There are debugging
314  versions and production versions. Note that the "rw" argument of RMATCH isn't  versions and production versions. Note that the "rw" argument of RMATCH isn't
# Line 287  actually used in this definition. */ Line 326  actually used in this definition. */
326    }    }
327  #define RRETURN(ra) \  #define RRETURN(ra) \
328    { \    { \
329    printf("match() returned %d from line %d ", ra, __LINE__); \    printf("match() returned %d from line %d\n", ra, __LINE__); \
330    return ra; \    return ra; \
331    }    }
332  #else  #else
# Line 307  argument of match(), which never changes Line 346  argument of match(), which never changes
346    
347  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
348    {\    {\
349    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = frame->Xnextframe;\
350    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL)\
351    frame->Xwhere = rw; \      {\
352        newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
353        if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
354        newframe->Xnextframe = NULL;\
355        frame->Xnextframe = newframe;\
356        }\
357      frame->Xwhere = rw;\
358    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
359    newframe->Xecode = rb;\    newframe->Xecode = rb;\
360    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
# Line 328  argument of match(), which never changes Line 373  argument of match(), which never changes
373    {\    {\
374    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
375    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
   (pcre_stack_free)(oldframe);\  
376    if (frame != NULL)\    if (frame != NULL)\
377      {\      {\
378      rrc = ra;\      rrc = ra;\
# Line 342  argument of match(), which never changes Line 386  argument of match(), which never changes
386    
387  typedef struct heapframe {  typedef struct heapframe {
388    struct heapframe *Xprevframe;    struct heapframe *Xprevframe;
389      struct heapframe *Xnextframe;
390    
391    /* Function arguments that may change */    /* Function arguments that may change */
392    
393    USPTR Xeptr;    PCRE_PUCHAR Xeptr;
394    const uschar *Xecode;    const pcre_uchar *Xecode;
395    USPTR Xmstart;    PCRE_PUCHAR Xmstart;
396    int Xoffset_top;    int Xoffset_top;
397    eptrblock *Xeptrb;    eptrblock *Xeptrb;
398    unsigned int Xrdepth;    unsigned int Xrdepth;
399    
400    /* Function local variables */    /* Function local variables */
401    
402    USPTR Xcallpat;    PCRE_PUCHAR Xcallpat;
403  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
404    USPTR Xcharptr;    PCRE_PUCHAR Xcharptr;
405  #endif  #endif
406    USPTR Xdata;    PCRE_PUCHAR Xdata;
407    USPTR Xnext;    PCRE_PUCHAR Xnext;
408    USPTR Xpp;    PCRE_PUCHAR Xpp;
409    USPTR Xprev;    PCRE_PUCHAR Xprev;
410    USPTR Xsaved_eptr;    PCRE_PUCHAR Xsaved_eptr;
411    
412    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
413    
# Line 372  typedef struct heapframe { Line 417  typedef struct heapframe {
417    
418  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
419    int Xprop_type;    int Xprop_type;
420    int Xprop_value;    unsigned int Xprop_value;
421    int Xprop_fail_result;    int Xprop_fail_result;
422    int Xoclength;    int Xoclength;
423    uschar Xocchars[8];    pcre_uchar Xocchars[6];
424  #endif  #endif
425    
426    int Xcodelink;    int Xcodelink;
# Line 385  typedef struct heapframe { Line 430  typedef struct heapframe {
430    int Xlength;    int Xlength;
431    int Xmax;    int Xmax;
432    int Xmin;    int Xmin;
433    int Xnumber;    unsigned int Xnumber;
434    int Xoffset;    int Xoffset;
435    int Xop;    unsigned int Xop;
436    int Xsave_capture_last;    pcre_int32 Xsave_capture_last;
437    int Xsave_offset1, Xsave_offset2, Xsave_offset3;    int Xsave_offset1, Xsave_offset2, Xsave_offset3;
438    int Xstacksave[REC_STACK_SAVE_MAX];    int Xstacksave[REC_STACK_SAVE_MAX];
439    
# Line 417  returns a negative (error) response, the Line 462  returns a negative (error) response, the
462  same response. */  same response. */
463    
464  /* These macros pack up tests that are used for partial matching, and which  /* These macros pack up tests that are used for partial matching, and which
465  appears several times in the code. We set the "hit end" flag if the pointer is  appear several times in the code. We set the "hit end" flag if the pointer is
466  at the end of the subject and also past the start of the subject (i.e.  at the end of the subject and also past the start of the subject (i.e.
467  something has been matched). For hard partial matching, we then return  something has been matched). For hard partial matching, we then return
468  immediately. The second one is used when we already know we are past the end of  immediately. The second one is used when we already know we are past the end of
# Line 440  the subject. */ Line 485  the subject. */
485    
486    
487  /* Performance note: It might be tempting to extract commonly used fields from  /* Performance note: It might be tempting to extract commonly used fields from
488  the md structure (e.g. utf8, end_subject) into individual variables to improve  the md structure (e.g. utf, end_subject) into individual variables to improve
489  performance. Tests using gcc on a SPARC disproved this; in the first case, it  performance. Tests using gcc on a SPARC disproved this; in the first case, it
490  made performance worse.  made performance worse.
491    
# Line 463  Returns:       MATCH_MATCH if matched Line 508  Returns:       MATCH_MATCH if matched
508  */  */
509    
510  static int  static int
511  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode,
512    int offset_top, match_data *md, eptrblock *eptrb, unsigned int rdepth)    PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb,
513      unsigned int rdepth)
514  {  {
515  /* 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,
516  so they can be ordinary variables in all cases. Mark some of them with  so they can be ordinary variables in all cases. Mark some of them with
# Line 472  so they can be ordinary variables in all Line 518  so they can be ordinary variables in all
518    
519  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
520  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
521  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register pcre_uint32 c;    /* Character values not kept over RMATCH() calls */
522  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
523    
524  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
525  BOOL caseless;  BOOL caseless;
526  int condcode;  int condcode;
527    
528  /* 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
529  preserved over calls to RMATCH() are part of a "frame" which is obtained from  preserved over calls to RMATCH() are part of a "frame". We set up the top-level
530  heap storage. Set up the top-level frame here; others are obtained from the  frame on the stack here; subsequent instantiations are obtained from the heap
531  heap whenever RMATCH() does a "recursion". See the macro definitions above. */  whenever RMATCH() does a "recursion". See the macro definitions above. Putting
532    the top-level on the stack rather than malloc-ing them all gives a performance
533    boost in many cases where there is not much "recursion". */
534    
535  #ifdef NO_RECURSE  #ifdef NO_RECURSE
536  heapframe *frame = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));  heapframe *frame = (heapframe *)md->match_frames_base;
 if (frame == NULL) RRETURN(PCRE_ERROR_NOMEMORY);  
 frame->Xprevframe = NULL;            /* Marks the top level */  
537    
538  /* Copy in the original argument variables */  /* Copy in the original argument variables */
539    
# Line 513  HEAP_RECURSE: Line 559  HEAP_RECURSE:
559    
560  /* Ditto for the local variables */  /* Ditto for the local variables */
561    
562  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
563  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
564  #endif  #endif
565  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
# Line 571  declarations can be cut out in a block. Line 617  declarations can be cut out in a block.
617  below are for variables that do not have to be preserved over a recursive call  below are for variables that do not have to be preserved over a recursive call
618  to RMATCH(). */  to RMATCH(). */
619    
620  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
621  const uschar *charptr;  const pcre_uchar *charptr;
622  #endif  #endif
623  const uschar *callpat;  const pcre_uchar *callpat;
624  const uschar *data;  const pcre_uchar *data;
625  const uschar *next;  const pcre_uchar *next;
626  USPTR         pp;  PCRE_PUCHAR       pp;
627  const uschar *prev;  const pcre_uchar *prev;
628  USPTR         saved_eptr;  PCRE_PUCHAR       saved_eptr;
629    
630  recursion_info new_recursive;  recursion_info new_recursive;
631    
# Line 589  BOOL prev_is_word; Line 635  BOOL prev_is_word;
635    
636  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
637  int prop_type;  int prop_type;
638  int prop_value;  unsigned int prop_value;
639  int prop_fail_result;  int prop_fail_result;
640  int oclength;  int oclength;
641  uschar occhars[8];  pcre_uchar occhars[6];
642  #endif  #endif
643    
644  int codelink;  int codelink;
# Line 600  int ctype; Line 646  int ctype;
646  int length;  int length;
647  int max;  int max;
648  int min;  int min;
649  int number;  unsigned int number;
650  int offset;  int offset;
651  int op;  unsigned int op;
652  int save_capture_last;  pcre_int32 save_capture_last;
653  int save_offset1, save_offset2, save_offset3;  int save_offset1, save_offset2, save_offset3;
654  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
655    
656  eptrblock newptrb;  eptrblock newptrb;
657    
658    /* There is a special fudge for calling match() in a way that causes it to
659    measure the size of its basic stack frame when the stack is being used for
660    recursion. The second argument (ecode) being NULL triggers this behaviour. It
661    cannot normally ever be NULL. The return is the negated value of the frame
662    size. */
663    
664    if (ecode == NULL)
665      {
666      if (rdepth == 0)
667        return match((PCRE_PUCHAR)&rdepth, NULL, NULL, 0, NULL, NULL, 1);
668      else
669        {
670        int len = (char *)&rdepth - (char *)eptr;
671        return (len > 0)? -len : len;
672        }
673      }
674  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
675    
676  /* To save space on the stack and in the heap frame, I have doubled up on some  /* To save space on the stack and in the heap frame, I have doubled up on some
# Line 620  the alternative names that are used. */ Line 683  the alternative names that are used. */
683  #define code_offset   codelink  #define code_offset   codelink
684  #define condassert    condition  #define condassert    condition
685  #define matched_once  prev_is_word  #define matched_once  prev_is_word
686    #define foc           number
687    #define save_mark     data
688    
689  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
690  variables. */  variables. */
# Line 645  defined). However, RMATCH isn't like a f Line 710  defined). However, RMATCH isn't like a f
710  complicated macro. It has to be used in one particular way. This shouldn't,  complicated macro. It has to be used in one particular way. This shouldn't,
711  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
712    
713  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
714  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
715  #else  #else
716  utf8 = FALSE;  utf = FALSE;
717  #endif  #endif
718    
719  /* First check that we haven't called match() too many times, or that we  /* First check that we haven't called match() too many times, or that we
# Line 689  for (;;) Line 754  for (;;)
754      case OP_MARK:      case OP_MARK:
755      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
756      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
757      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
758        eptrb, RM55);        eptrb, RM55);
759      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
760           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 702  for (;;) Line 767  for (;;)
767      unaltered. */      unaltered. */
768    
769      else if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
770          strcmp((char *)(ecode + 2), (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC_TEST(ecode + 2, md->start_match_ptr) == 0)
771        {        {
772        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
773        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
# Line 715  for (;;) Line 780  for (;;)
780      /* COMMIT overrides PRUNE, SKIP, and THEN */      /* COMMIT overrides PRUNE, SKIP, and THEN */
781    
782      case OP_COMMIT:      case OP_COMMIT:
783      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
784        eptrb, RM52);        eptrb, RM52);
785      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&
786          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&
# Line 726  for (;;) Line 791  for (;;)
791      /* PRUNE overrides THEN */      /* PRUNE overrides THEN */
792    
793      case OP_PRUNE:      case OP_PRUNE:
794      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
795        eptrb, RM51);        eptrb, RM51);
796      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
797      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
# Line 734  for (;;) Line 799  for (;;)
799      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
800      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
801      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
802      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
803        eptrb, RM56);        eptrb, RM56);
804      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
805           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 744  for (;;) Line 809  for (;;)
809      /* SKIP overrides PRUNE and THEN */      /* SKIP overrides PRUNE and THEN */
810    
811      case OP_SKIP:      case OP_SKIP:
812      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
813        eptrb, RM53);        eptrb, RM53);
814      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
815        RRETURN(rrc);        RRETURN(rrc);
# Line 752  for (;;) Line 817  for (;;)
817      RRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
818    
819      /* Note that, for Perl compatibility, SKIP with an argument does NOT set      /* Note that, for Perl compatibility, SKIP with an argument does NOT set
820      nomatch_mark. There is a flag that disables this opcode when re-matching a      nomatch_mark. There is a flag that disables this opcode when re-matching a
821      pattern that ended with a SKIP for which there was not a matching MARK. */      pattern that ended with a SKIP for which there was not a matching MARK. */
822    
823      case OP_SKIP_ARG:      case OP_SKIP_ARG:
824      if (md->ignore_skip_arg)      if (md->ignore_skip_arg)
825        {        {
826        ecode += _pcre_OP_lengths[*ecode] + ecode[1];        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
827        break;        break;
828        }        }
829      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
830        eptrb, RM57);        eptrb, RM57);
831      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
832        RRETURN(rrc);        RRETURN(rrc);
833    
834      /* Pass back the current skip name by overloading md->start_match_ptr and      /* Pass back the current skip name by overloading md->start_match_ptr and
835      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
836      caught by a matching MARK, or get to the top, where it causes a rematch      caught by a matching MARK, or get to the top, where it causes a rematch
837      with the md->ignore_skip_arg flag set. */      with the md->ignore_skip_arg flag set. */
838    
839      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
# Line 779  for (;;) Line 844  for (;;)
844      match pointer to do this. */      match pointer to do this. */
845    
846      case OP_THEN:      case OP_THEN:
847      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
848        eptrb, RM54);        eptrb, RM54);
849      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
850      md->start_match_ptr = ecode;      md->start_match_ptr = ecode;
# Line 788  for (;;) Line 853  for (;;)
853      case OP_THEN_ARG:      case OP_THEN_ARG:
854      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
855      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
856      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
857        md, eptrb, RM58);        md, eptrb, RM58);
858      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
859           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 812  for (;;) Line 877  for (;;)
877      case OP_ONCE_NC:      case OP_ONCE_NC:
878      prev = ecode;      prev = ecode;
879      saved_eptr = eptr;      saved_eptr = eptr;
880        save_mark = md->mark;
881      do      do
882        {        {
883        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
# Line 830  for (;;) Line 896  for (;;)
896    
897        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
898        ecode += GET(ecode,1);        ecode += GET(ecode,1);
899          md->mark = save_mark;
900        }        }
901      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
902    
# Line 869  for (;;) Line 936  for (;;)
936        }        }
937      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
938        {        {
       md->match_function_type = MATCH_CBEGROUP;  
939        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
940        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
941        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 909  for (;;) Line 975  for (;;)
975        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
976        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
977        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
978          save_mark = md->mark;
979    
980        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
981        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 917  for (;;) Line 984  for (;;)
984        for (;;)        for (;;)
985          {          {
986          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
987          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
988            eptrb, RM1);            eptrb, RM1);
989          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
990    
# Line 945  for (;;) Line 1012  for (;;)
1012          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1013          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
1014          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
1015            md->mark = save_mark;
1016          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
1017          }          }
1018    
# Line 996  for (;;) Line 1064  for (;;)
1064    
1065      for (;;)      for (;;)
1066        {        {
1067        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1068            md->match_function_type = MATCH_CBEGROUP;
1069    
1070        /* If this is not a possibly empty group, and there are no (*THEN)s in        /* If this is not a possibly empty group, and there are no (*THEN)s in
1071        the pattern, and this is the final alternative, optimize as described        the pattern, and this is the final alternative, optimize as described
# Line 1004  for (;;) Line 1073  for (;;)
1073    
1074        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1075          {          {
1076          ecode += _pcre_OP_lengths[*ecode];          ecode += PRIV(OP_lengths)[*ecode];
1077          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1078          }          }
1079    
1080        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1081    
1082        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,        save_mark = md->mark;
1083          save_capture_last = md->capture_last;
1084          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1085          RM2);          RM2);
1086    
1087        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
# Line 1028  for (;;) Line 1099  for (;;)
1099          {          {
1100          if (rrc == MATCH_ONCE)          if (rrc == MATCH_ONCE)
1101            {            {
1102            const uschar *scode = ecode;            const pcre_uchar *scode = ecode;
1103            if (*scode != OP_ONCE)           /* If not at start, find it */            if (*scode != OP_ONCE)           /* If not at start, find it */
1104              {              {
1105              while (*scode == OP_ALT) scode += GET(scode, 1);              while (*scode == OP_ALT) scode += GET(scode, 1);
# Line 1039  for (;;) Line 1110  for (;;)
1110          RRETURN(rrc);          RRETURN(rrc);
1111          }          }
1112        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1113          md->mark = save_mark;
1114        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1115          md->capture_last = save_capture_last;
1116        }        }
1117    
1118      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
# Line 1070  for (;;) Line 1143  for (;;)
1143      if (offset < md->offset_max)      if (offset < md->offset_max)
1144        {        {
1145        matched_once = FALSE;        matched_once = FALSE;
1146        code_offset = ecode - md->start_code;        code_offset = (int)(ecode - md->start_code);
1147    
1148        save_offset1 = md->offset_vector[offset];        save_offset1 = md->offset_vector[offset];
1149        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
# Line 1093  for (;;) Line 1166  for (;;)
1166          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1167            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1168          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1169          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1170            eptrb, RM63);            eptrb, RM63);
1171          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1172            {            {
# Line 1160  for (;;) Line 1233  for (;;)
1233    
1234      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1235      matched_once = FALSE;      matched_once = FALSE;
1236      code_offset = ecode - md->start_code;      code_offset = (int)(ecode - md->start_code);
1237        save_capture_last = md->capture_last;
1238    
1239      for (;;)      for (;;)
1240        {        {
1241        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1242        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1243          eptrb, RM48);          eptrb, RM48);
1244        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1245          {          {
# Line 1190  for (;;) Line 1264  for (;;)
1264        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1265        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1266        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1267          md->capture_last = save_capture_last;
1268        }        }
1269    
1270      if (matched_once || allow_zero)      if (matched_once || allow_zero)
# Line 1215  for (;;) Line 1290  for (;;)
1290    
1291      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (ecode[LINK_SIZE+1] == OP_CALLOUT)
1292        {        {
1293        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
1294          {          {
1295          pcre_callout_block cb;          PUBL(callout_block) cb;
1296          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1297          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1298          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1299    #if defined COMPILE_PCRE8
1300          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1301    #elif defined COMPILE_PCRE16
1302            cb.subject          = (PCRE_SPTR16)md->start_subject;
1303    #elif defined COMPILE_PCRE32
1304            cb.subject          = (PCRE_SPTR32)md->start_subject;
1305    #endif
1306          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1307          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1308          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
1309          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, LINK_SIZE + 3);
1310          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);
1311          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1312          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last & CAPLMASK;
1313            /* Internal change requires this for API compatibility. */
1314            if (cb.capture_last == 0) cb.capture_last = -1;
1315          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1316          cb.mark             = md->nomatch_mark;          cb.mark             = md->nomatch_mark;
1317          if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1318          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1319          }          }
1320        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1321        }        }
1322    
1323      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1250  for (;;) Line 1333  for (;;)
1333          }          }
1334        else        else
1335          {          {
1336          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          unsigned int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1337          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1338    
1339          /* If the test is for recursion into a specific subpattern, and it is          /* If the test is for recursion into a specific subpattern, and it is
# Line 1260  for (;;) Line 1343  for (;;)
1343    
1344          if (!condition && condcode == OP_NRREF)          if (!condition && condcode == OP_NRREF)
1345            {            {
1346            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1347            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1348              {              {
1349              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 1273  for (;;) Line 1356  for (;;)
1356    
1357            if (i < md->name_count)            if (i < md->name_count)
1358              {              {
1359              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1360              while (slotB > md->name_table)              while (slotB > md->name_table)
1361                {                {
1362                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1363                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1364                  {                  {
1365                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1366                  if (condition) break;                  if (condition) break;
# Line 1293  for (;;) Line 1376  for (;;)
1376                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1377                  {                  {
1378                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1379                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1380                    {                    {
1381                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1382                    if (condition) break;                    if (condition) break;
# Line 1306  for (;;) Line 1389  for (;;)
1389    
1390          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1391    
1392          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1393          }          }
1394        }        }
1395    
# Line 1322  for (;;) Line 1405  for (;;)
1405    
1406        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1407          {          {
1408          int refno = offset >> 1;          unsigned int refno = offset >> 1;
1409          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1410    
1411          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1412            {            {
# Line 1337  for (;;) Line 1420  for (;;)
1420    
1421          if (i < md->name_count)          if (i < md->name_count)
1422            {            {
1423            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1424            while (slotB > md->name_table)            while (slotB > md->name_table)
1425              {              {
1426              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1427              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1428                {                {
1429                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1430                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1359  for (;;) Line 1442  for (;;)
1442              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1443                {                {
1444                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1445                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1446                  {                  {
1447                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1448                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1374  for (;;) Line 1457  for (;;)
1457    
1458        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1459    
1460        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1461        }        }
1462    
1463      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1450  for (;;) Line 1533  for (;;)
1533      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1534    
1535      case OP_CLOSE:      case OP_CLOSE:
1536      number = GET2(ecode, 1);      number = GET2(ecode, 1);   /* Must be less than 65536 */
1537      offset = number << 1;      offset = number << 1;
1538    
1539  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 1458  for (;;) Line 1541  for (;;)
1541        printf("\n");        printf("\n");
1542  #endif  #endif
1543    
1544      md->capture_last = number;      md->capture_last = (md->capture_last & OVFLMASK) | number;
1545      if (offset >= md->offset_max) md->offset_overflow = TRUE; else      if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1546        {        {
1547        md->offset_vector[offset] =        md->offset_vector[offset] =
1548          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
1549        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1550        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1551        }        }
1552      ecode += 3;      ecode += 1 + IMM2_SIZE;
1553      break;      break;
1554    
1555    
# Line 1513  for (;;) Line 1596  for (;;)
1596    
1597      case OP_ASSERT:      case OP_ASSERT:
1598      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1599        save_mark = md->mark;
1600      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1601        {        {
1602        condassert = TRUE;        condassert = TRUE;
# Line 1528  for (;;) Line 1612  for (;;)
1612          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1613          break;          break;
1614          }          }
1615          md->mark = save_mark;
1616    
1617        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* A COMMIT failure must fail the entire assertion, without trying any
1618        as NOMATCH. */        subsequent branches. */
1619    
1620          if (rrc == MATCH_COMMIT) RRETURN(MATCH_NOMATCH);
1621    
1622          /* PCRE does not allow THEN to escape beyond an assertion; it
1623          is treated as NOMATCH. */
1624    
1625        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1626        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
# Line 1557  for (;;) Line 1647  for (;;)
1647    
1648      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1649      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1650        save_mark = md->mark;
1651      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1652        {        {
1653        condassert = TRUE;        condassert = TRUE;
# Line 1567  for (;;) Line 1658  for (;;)
1658      do      do
1659        {        {
1660        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1661          md->mark = save_mark;
1662        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1663        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1664          {          {
# Line 1593  for (;;) Line 1685  for (;;)
1685      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1686    
1687      case OP_REVERSE:      case OP_REVERSE:
1688  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1689      if (utf8)      if (utf)
1690        {        {
1691        i = GET(ecode, 1);        i = GET(ecode, 1);
1692        while (i-- > 0)        while (i-- > 0)
# Line 1625  for (;;) Line 1717  for (;;)
1717      function is able to force a failure. */      function is able to force a failure. */
1718    
1719      case OP_CALLOUT:      case OP_CALLOUT:
1720      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1721        {        {
1722        pcre_callout_block cb;        PUBL(callout_block) cb;
1723        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1724        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1725        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1726    #if defined COMPILE_PCRE8
1727        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1728    #elif defined COMPILE_PCRE16
1729          cb.subject          = (PCRE_SPTR16)md->start_subject;
1730    #elif defined COMPILE_PCRE32
1731          cb.subject          = (PCRE_SPTR32)md->start_subject;
1732    #endif
1733        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1734        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1735        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
1736        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1737        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1738        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1739        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1740          /* Internal change requires this for API compatibility. */
1741          if (cb.capture_last == 0) cb.capture_last = -1;
1742        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1743        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1744        if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1745        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1746        }        }
1747      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1667  for (;;) Line 1767  for (;;)
1767      case OP_RECURSE:      case OP_RECURSE:
1768        {        {
1769        recursion_info *ri;        recursion_info *ri;
1770        int recno;        unsigned int recno;
1771    
1772        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1773        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1684  for (;;) Line 1784  for (;;)
1784        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1785    
1786        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1787          new_recursive.saved_capture_last = md->capture_last;
1788        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1789        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1790        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1700  for (;;) Line 1801  for (;;)
1801        else        else
1802          {          {
1803          new_recursive.offset_save =          new_recursive.offset_save =
1804            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1805          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1806          }          }
1807        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1808              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1809    
1810        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1811        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1812        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1813          */
1814    
1815        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1816        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
1817        do        do
1818          {          {
1819          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1820          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1821            md, eptrb, RM6);            md, eptrb, RM6);
1822          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1823              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1824            md->capture_last = new_recursive.saved_capture_last;
1825          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1826          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1827            {            {
1828            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
1829            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1830              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1831    
1832            /* Set where we got to in the subject, and reset the start in case            /* Set where we got to in the subject, and reset the start in case
1833            it was changed by \K. This *is* propagated back out of a recursion,            it was changed by \K. This *is* propagated back out of a recursion,
# Line 1735  for (;;) Line 1838  for (;;)
1838            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1839            }            }
1840    
1841          /* PCRE does not allow THEN to escape beyond a recursion; it is treated          /* PCRE does not allow THEN or COMMIT to escape beyond a recursion; it
1842          as NOMATCH. */          is treated as NOMATCH. */
1843    
1844          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN &&
1845                     rrc != MATCH_COMMIT)
1846            {            {
1847            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1848            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1849              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1850            RRETURN(rrc);            RRETURN(rrc);
1851            }            }
1852    
# Line 1754  for (;;) Line 1858  for (;;)
1858        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1859        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1860        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1861          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1862        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1863        }        }
1864    
# Line 1868  for (;;) Line 1972  for (;;)
1972    
1973        /* Deal with capturing */        /* Deal with capturing */
1974    
1975        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
1976        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1977          {          {
1978          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
1979          "skipping" a capturing group, and that group's offsets must be marked          "skipping" a capturing group, and that group's offsets must be marked
# Line 1949  for (;;) Line 2053  for (;;)
2053          }          }
2054        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2055          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2056          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2057          RRETURN(rrc);          RRETURN(rrc);
2058          }          }
# Line 1958  for (;;) Line 2061  for (;;)
2061        }        }
2062      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2063        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2064        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2065        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2066        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2015  for (;;) Line 2117  for (;;)
2117    
2118      case OP_DOLLM:      case OP_DOLLM:
2119      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2120        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2121          if (!IS_NEWLINE(eptr))
2122            {
2123            if (md->partial != 0 &&
2124                eptr + 1 >= md->end_subject &&
2125                NLBLOCK->nltype == NLTYPE_FIXED &&
2126                NLBLOCK->nllen == 2 &&
2127                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2128              {
2129              md->hitend = TRUE;
2130              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2131              }
2132            RRETURN(MATCH_NOMATCH);
2133            }
2134          }
2135      else      else
2136        {        {
2137        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2047  for (;;) Line 2163  for (;;)
2163      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2164      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2165          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2166          {
2167          if (md->partial != 0 &&
2168              eptr + 1 >= md->end_subject &&
2169              NLBLOCK->nltype == NLTYPE_FIXED &&
2170              NLBLOCK->nllen == 2 &&
2171              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2172            {
2173            md->hitend = TRUE;
2174            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2175            }
2176        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2177          }
2178    
2179      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2180    
# Line 2066  for (;;) Line 2193  for (;;)
2193        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2194        partial matching. */        partial matching. */
2195    
2196  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2197        if (utf8)        if (utf)
2198          {          {
2199          /* Get status of previous character */          /* Get status of previous character */
2200    
2201          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2202            {            {
2203            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2204            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2205            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2206            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2207  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2139  for (;;) Line 2266  for (;;)
2266              }              }
2267            else            else
2268  #endif  #endif
2269            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2270                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2271            }            }
2272    
2273          /* Get status of next character */          /* Get status of next character */
# Line 2162  for (;;) Line 2290  for (;;)
2290            }            }
2291          else          else
2292  #endif  #endif
2293          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2294              && ((md->ctypes[*eptr] & ctype_word) != 0);
2295          }          }
2296    
2297        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
# Line 2173  for (;;) Line 2302  for (;;)
2302        }        }
2303      break;      break;
2304    
2305      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2306        CRLF newlines and partial matching. */
2307    
2308      case OP_ANY:      case OP_ANY:
2309      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2310        if (md->partial != 0 &&
2311            eptr + 1 >= md->end_subject &&
2312            NLBLOCK->nltype == NLTYPE_FIXED &&
2313            NLBLOCK->nllen == 2 &&
2314            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2315          {
2316          md->hitend = TRUE;
2317          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2318          }
2319    
2320      /* Fall through */      /* Fall through */
2321    
2322        /* Match any single character whatsoever. */
2323    
2324      case OP_ALLANY:      case OP_ALLANY:
2325      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2326        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2186  for (;;) Line 2328  for (;;)
2328        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2329        }        }
2330      eptr++;      eptr++;
2331      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  #ifdef SUPPORT_UTF
2332        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2333    #endif
2334      ecode++;      ecode++;
2335      break;      break;
2336    
# Line 2211  for (;;) Line 2355  for (;;)
2355        }        }
2356      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2357      if (      if (
2358  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2359         c < 256 &&         c < 256 &&
2360  #endif  #endif
2361         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
# Line 2228  for (;;) Line 2372  for (;;)
2372        }        }
2373      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2374      if (      if (
2375  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2376         c >= 256 ||         c > 255 ||
2377  #endif  #endif
2378         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2379         )         )
# Line 2245  for (;;) Line 2389  for (;;)
2389        }        }
2390      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2391      if (      if (
2392  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2393         c < 256 &&         c < 256 &&
2394  #endif  #endif
2395         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
# Line 2262  for (;;) Line 2406  for (;;)
2406        }        }
2407      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2408      if (      if (
2409  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2410         c >= 256 ||         c > 255 ||
2411  #endif  #endif
2412         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2413         )         )
# Line 2279  for (;;) Line 2423  for (;;)
2423        }        }
2424      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2425      if (      if (
2426  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2427         c < 256 &&         c < 256 &&
2428  #endif  #endif
2429         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
# Line 2296  for (;;) Line 2440  for (;;)
2440        }        }
2441      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2442      if (      if (
2443  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2444         c >= 256 ||         c > 255 ||
2445  #endif  #endif
2446         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2447         )         )
# Line 2316  for (;;) Line 2460  for (;;)
2460        {        {
2461        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2462    
2463        case 0x000d:        case CHAR_CR:
2464        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2465            {
2466            SCHECK_PARTIAL();
2467            }
2468          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2469        break;        break;
2470    
2471        case 0x000a:        case CHAR_LF:
2472        break;        break;
2473    
2474        case 0x000b:        case CHAR_VT:
2475        case 0x000c:        case CHAR_FF:
2476        case 0x0085:        case CHAR_NEL:
2477    #ifndef EBCDIC
2478        case 0x2028:        case 0x2028:
2479        case 0x2029:        case 0x2029:
2480    #endif  /* Not EBCDIC */
2481        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2482        break;        break;
2483        }        }
# Line 2343  for (;;) Line 2493  for (;;)
2493      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2494      switch(c)      switch(c)
2495        {        {
2496          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2497        default: break;        default: break;
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       RRETURN(MATCH_NOMATCH);  
2498        }        }
2499      ecode++;      ecode++;
2500      break;      break;
# Line 2377  for (;;) Line 2508  for (;;)
2508      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2509      switch(c)      switch(c)
2510        {        {
2511          HSPACE_CASES: break;  /* Byte and multibyte cases */
2512        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       break;  
2513        }        }
2514      ecode++;      ecode++;
2515      break;      break;
# Line 2411  for (;;) Line 2523  for (;;)
2523      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2524      switch(c)      switch(c)
2525        {        {
2526          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2527        default: break;        default: break;
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       RRETURN(MATCH_NOMATCH);  
2528        }        }
2529      ecode++;      ecode++;
2530      break;      break;
# Line 2433  for (;;) Line 2538  for (;;)
2538      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2539      switch(c)      switch(c)
2540        {        {
2541          VSPACE_CASES: break;
2542        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       break;  
2543        }        }
2544      ecode++;      ecode++;
2545      break;      break;
# Line 2459  for (;;) Line 2557  for (;;)
2557        }        }
2558      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2559        {        {
2560          const pcre_uint32 *cp;
2561        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2562    
2563        switch(ecode[1])        switch(ecode[1])
# Line 2475  for (;;) Line 2574  for (;;)
2574          break;          break;
2575    
2576          case PT_GC:          case PT_GC:
2577          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2578            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2579          break;          break;
2580    
# Line 2492  for (;;) Line 2591  for (;;)
2591          /* These are specials */          /* These are specials */
2592    
2593          case PT_ALNUM:          case PT_ALNUM:
2594          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2595               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2596            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2597          break;          break;
2598    
2599          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2600          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2601               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2602                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2603            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2604          break;          break;
2605    
2606          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2607          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2608               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2609               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2610                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
# Line 2513  for (;;) Line 2612  for (;;)
2612          break;          break;
2613    
2614          case PT_WORD:          case PT_WORD:
2615          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2616               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2617               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2618            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2619          break;          break;
2620    
2621            case PT_CLIST:
2622            cp = PRIV(ucd_caseless_sets) + ecode[2];
2623            for (;;)
2624              {
2625              if (c < *cp)
2626                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2627              if (c == *cp++)
2628                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2629              }
2630            break;
2631    
2632          /* This should never occur */          /* This should never occur */
2633    
2634          default:          default:
# Line 2538  for (;;) Line 2648  for (;;)
2648        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2649        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2650        }        }
2651      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2652        {        {
2653        int len = 1;        int lgb, rgb;
2654        if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2655        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2656        eptr += len;        while (eptr < md->end_subject)
2657            {
2658            int len = 1;
2659            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2660            rgb = UCD_GRAPHBREAK(c);
2661            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2662            lgb = rgb;
2663            eptr += len;
2664            }
2665        }        }
2666        CHECK_PARTIAL();
2667      ecode++;      ecode++;
2668      break;      break;
2669  #endif  #endif  /* SUPPORT_UCP */
2670    
2671    
2672      /* Match a back reference, possibly repeatedly. Look past the end of the      /* Match a back reference, possibly repeatedly. Look past the end of the
# Line 2564  for (;;) Line 2681  for (;;)
2681      case OP_REFI:      case OP_REFI:
2682      caseless = op == OP_REFI;      caseless = op == OP_REFI;
2683      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2684      ecode += 3;      ecode += 1 + IMM2_SIZE;
2685    
2686      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2687    
# Line 2604  for (;;) Line 2721  for (;;)
2721        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2722        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2723        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2724        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2725        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2726        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2727        break;        break;
2728    
2729        default:               /* No repeat follows */        default:               /* No repeat follows */
2730        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2731          {          {
2732            if (length == -2) eptr = md->end_subject;   /* Partial match */
2733          CHECK_PARTIAL();          CHECK_PARTIAL();
2734          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2735          }          }
# Line 2620  for (;;) Line 2738  for (;;)
2738        }        }
2739    
2740      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2741      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2742        means the reference is unset in non-Java-compatible mode. If the minimum is
2743        zero, we can continue at the same level without recursion. For any other
2744        minimum, carrying on will result in NOMATCH. */
2745    
2746      if (length == 0) continue;      if (length == 0) continue;
2747        if (length < 0 && min == 0) continue;
2748    
2749      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2750      the length of the reference string explicitly rather than passing the      the length of the reference string explicitly rather than passing the
# Line 2633  for (;;) Line 2755  for (;;)
2755        int slength;        int slength;
2756        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2757          {          {
2758            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2759          CHECK_PARTIAL();          CHECK_PARTIAL();
2760          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2761          }          }
# Line 2656  for (;;) Line 2779  for (;;)
2779          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2780          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2781            {            {
2782              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2783            CHECK_PARTIAL();            CHECK_PARTIAL();
2784            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2785            }            }
# Line 2674  for (;;) Line 2798  for (;;)
2798          int slength;          int slength;
2799          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2800            {            {
2801            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2802              the soft partial matching case. */
2803    
2804              if (slength == -2 && md->partial != 0 &&
2805                  md->end_subject > md->start_used_ptr)
2806                {
2807                md->hitend = TRUE;
2808                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2809                }
2810            break;            break;
2811            }            }
2812          eptr += slength;          eptr += slength;
2813          }          }
2814    
2815        while (eptr >= pp)        while (eptr >= pp)
2816          {          {
2817          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 2703  for (;;) Line 2836  for (;;)
2836      case OP_NCLASS:      case OP_NCLASS:
2837      case OP_CLASS:      case OP_CLASS:
2838        {        {
2839          /* The data variable is saved across frames, so the byte map needs to
2840          be stored there. */
2841    #define BYTE_MAP ((pcre_uint8 *)data)
2842        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2843        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2844    
2845        switch (*ecode)        switch (*ecode)
2846          {          {
# Line 2725  for (;;) Line 2861  for (;;)
2861          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2862          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2863          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2864          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2865          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2866          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2867          break;          break;
2868    
2869          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2737  for (;;) Line 2873  for (;;)
2873    
2874        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2875    
2876  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2877        /* UTF-8 mode */        if (utf)
       if (utf8)  
2878          {          {
2879          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2880            {            {
# Line 2754  for (;;) Line 2889  for (;;)
2889              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2890              }              }
2891            else            else
2892              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
             }  
2893            }            }
2894          }          }
2895        else        else
2896  #endif  #endif
2897        /* Not UTF-8 mode */        /* Not UTF mode */
2898          {          {
2899          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2900            {            {
# Line 2771  for (;;) Line 2904  for (;;)
2904              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2905              }              }
2906            c = *eptr++;            c = *eptr++;
2907            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2908              if (c > 255)
2909                {
2910                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2911                }
2912              else
2913    #endif
2914                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2915            }            }
2916          }          }
2917    
# Line 2785  for (;;) Line 2925  for (;;)
2925    
2926        if (minimize)        if (minimize)
2927          {          {
2928  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2929          /* UTF-8 mode */          if (utf)
         if (utf8)  
2930            {            {
2931            for (fi = min;; fi++)            for (fi = min;; fi++)
2932              {              {
# Line 2805  for (;;) Line 2944  for (;;)
2944                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2945                }                }
2946              else              else
2947                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
               }  
2948              }              }
2949            }            }
2950          else          else
2951  #endif  #endif
2952          /* Not UTF-8 mode */          /* Not UTF mode */
2953            {            {
2954            for (fi = min;; fi++)            for (fi = min;; fi++)
2955              {              {
# Line 2825  for (;;) Line 2962  for (;;)
2962                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2963                }                }
2964              c = *eptr++;              c = *eptr++;
2965              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2966                if (c > 255)
2967                  {
2968                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2969                  }
2970                else
2971    #endif
2972                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2973              }              }
2974            }            }
2975          /* Control never gets here */          /* Control never gets here */
# Line 2837  for (;;) Line 2981  for (;;)
2981          {          {
2982          pp = eptr;          pp = eptr;
2983    
2984  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2985          /* UTF-8 mode */          if (utf)
         if (utf8)  
2986            {            {
2987            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2988              {              {
# Line 2855  for (;;) Line 2998  for (;;)
2998                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
2999                }                }
3000              else              else
3001                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
3002              eptr += len;              eptr += len;
3003              }              }
3004            for (;;)            for (;;)
# Line 2870  for (;;) Line 3011  for (;;)
3011            }            }
3012          else          else
3013  #endif  #endif
3014            /* Not UTF-8 mode */            /* Not UTF mode */
3015            {            {
3016            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3017              {              {
# Line 2880  for (;;) Line 3021  for (;;)
3021                break;                break;
3022                }                }
3023              c = *eptr;              c = *eptr;
3024              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
3025                if (c > 255)
3026                  {
3027                  if (op == OP_CLASS) break;
3028                  }
3029                else
3030    #endif
3031                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3032              eptr++;              eptr++;
3033              }              }
3034            while (eptr >= pp)            while (eptr >= pp)
# Line 2893  for (;;) Line 3041  for (;;)
3041    
3042          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3043          }          }
3044    #undef BYTE_MAP
3045        }        }
3046      /* Control never gets here */      /* Control never gets here */
3047    
# Line 2901  for (;;) Line 3050  for (;;)
3050      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8
3051      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
3052    
3053  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3054      case OP_XCLASS:      case OP_XCLASS:
3055        {        {
3056        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2926  for (;;) Line 3075  for (;;)
3075          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3076          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3077          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3078          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
3079          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
3080          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
3081          break;          break;
3082    
3083          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2946  for (;;) Line 3095  for (;;)
3095            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3096            }            }
3097          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
3098          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3099          }          }
3100    
3101        /* If max == min we can continue with the main loop without the        /* If max == min we can continue with the main loop without the
# Line 2970  for (;;) Line 3119  for (;;)
3119              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3120              }              }
3121            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3122            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3123            }            }
3124          /* Control never gets here */          /* Control never gets here */
3125          }          }
# Line 2988  for (;;) Line 3137  for (;;)
3137              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3138              break;              break;
3139              }              }
3140    #ifdef SUPPORT_UTF
3141            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3142            if (!_pcre_xclass(c, data)) break;  #else
3143              c = *eptr;
3144    #endif
3145              if (!PRIV(xclass)(c, data, utf)) break;
3146            eptr += len;            eptr += len;
3147            }            }
3148          for(;;)          for(;;)
# Line 2997  for (;;) Line 3150  for (;;)
3150            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3151            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3152            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3153            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3154              if (utf) BACKCHAR(eptr);
3155    #endif
3156            }            }
3157          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3158          }          }
# Line 3009  for (;;) Line 3164  for (;;)
3164      /* Match a single character, casefully */      /* Match a single character, casefully */
3165    
3166      case OP_CHAR:      case OP_CHAR:
3167  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3168      if (utf8)      if (utf)
3169        {        {
3170        length = 1;        length = 1;
3171        ecode++;        ecode++;
# Line 3020  for (;;) Line 3175  for (;;)
3175          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3176          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3177          }          }
3178        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3179        }        }
3180      else      else
3181  #endif  #endif
3182        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3183        {        {
3184        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3185          {          {
# Line 3037  for (;;) Line 3191  for (;;)
3191        }        }
3192      break;      break;
3193    
3194      /* Match a single character, caselessly */      /* Match a single character, caselessly. If we are at the end of the
3195        subject, give up immediately. */
3196    
3197      case OP_CHARI:      case OP_CHARI:
3198  #ifdef SUPPORT_UTF8      if (eptr >= md->end_subject)
3199      if (utf8)        {
3200          SCHECK_PARTIAL();
3201          RRETURN(MATCH_NOMATCH);
3202          }
3203    
3204    #ifdef SUPPORT_UTF
3205        if (utf)
3206        {        {
3207        length = 1;        length = 1;
3208        ecode++;        ecode++;
3209        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
3210    
       if (length > md->end_subject - eptr)  
         {  
         CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */  
         RRETURN(MATCH_NOMATCH);  
         }  
   
3211        /* If the pattern character's value is < 128, we have only one byte, and        /* If the pattern character's value is < 128, we have only one byte, and
3212        can use the fast lookup table. */        we know that its other case must also be one byte long, so we can use the
3213          fast lookup table. We know that there is at least one byte left in the
3214          subject. */
3215    
3216        if (fc < 128)        if (fc < 128)
3217          {          {
3218          if (md->lcc[*ecode++] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          pcre_uchar cc = RAWUCHAR(eptr);
3219            if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3220            ecode++;
3221            eptr++;
3222          }          }
3223    
3224        /* Otherwise we must pick up the subject character */        /* Otherwise we must pick up the subject character. Note that we cannot
3225          use the value of "length" to check for sufficient bytes left, because the
3226          other case of the character may have more or fewer bytes.  */
3227    
3228        else        else
3229          {          {
3230          unsigned int dc;          pcre_uint32 dc;
3231          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3232          ecode += length;          ecode += length;
3233    
# Line 3082  for (;;) Line 3244  for (;;)
3244          }          }
3245        }        }
3246      else      else
3247  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3248    
3249      /* Non-UTF-8 mode */      /* Not UTF mode */
3250        {        {
3251        if (md->end_subject - eptr < 1)        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3252          {            != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3253          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */        eptr++;
         RRETURN(MATCH_NOMATCH);  
         }  
       if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);  
3254        ecode += 2;        ecode += 2;
3255        }        }
3256      break;      break;
# Line 3101  for (;;) Line 3260  for (;;)
3260      case OP_EXACT:      case OP_EXACT:
3261      case OP_EXACTI:      case OP_EXACTI:
3262      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3263      ecode += 3;      ecode += 1 + IMM2_SIZE;
3264      goto REPEATCHAR;      goto REPEATCHAR;
3265    
3266      case OP_POSUPTO:      case OP_POSUPTO:
# Line 3116  for (;;) Line 3275  for (;;)
3275      min = 0;      min = 0;
3276      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3277      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3278      ecode += 3;      ecode += 1 + IMM2_SIZE;
3279      goto REPEATCHAR;      goto REPEATCHAR;
3280    
3281      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3164  for (;;) Line 3323  for (;;)
3323      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3324    
3325      REPEATCHAR:      REPEATCHAR:
3326  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3327      if (utf8)      if (utf)
3328        {        {
3329        length = 1;        length = 1;
3330        charptr = ecode;        charptr = ecode;
# Line 3178  for (;;) Line 3337  for (;;)
3337        if (length > 1)        if (length > 1)
3338          {          {
3339  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3340          unsigned int othercase;          pcre_uint32 othercase;
3341          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3342              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3343            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3344          else oclength = 0;          else oclength = 0;
3345  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3346    
3347          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3348            {            {
3349            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3350              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3351  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3352            else if (oclength > 0 &&            else if (oclength > 0 &&
3353                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3354                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3355  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3356            else            else
3357              {              {
# Line 3211  for (;;) Line 3370  for (;;)
3370              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3371              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3372              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3373                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3374  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3375              else if (oclength > 0 &&              else if (oclength > 0 &&
3376                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3377                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3378  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3379              else              else
3380                {                {
# Line 3232  for (;;) Line 3391  for (;;)
3391            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3392              {              {
3393              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3394                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3395  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3396              else if (oclength > 0 &&              else if (oclength > 0 &&
3397                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3398                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3399  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3400              else              else
3401                {                {
# Line 3268  for (;;) Line 3427  for (;;)
3427        value of fc will always be < 128. */        value of fc will always be < 128. */
3428        }        }
3429      else      else
3430  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3431          /* When not in UTF-8 mode, load a single-byte character. */
3432      /* When not in UTF-8 mode, load a single-byte character. */        fc = *ecode++;
   
     fc = *ecode++;  
3433    
3434      /* The value of fc at this point is always less than 256, though we may or      /* The value of fc at this point is always one character, though we may
3435      may not be in UTF-8 mode. The code is duplicated for the caseless and      or may not be in UTF mode. The code is duplicated for the caseless and
3436      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3437      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3438      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3284  for (;;) Line 3441  for (;;)
3441      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3442    
3443      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3444        max, eptr));        max, (char *)eptr));
3445    
3446      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3447        {        {
3448        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3449          /* fc must be < 128 if UTF is enabled. */
3450          foc = md->fcc[fc];
3451    #else
3452    #ifdef SUPPORT_UTF
3453    #ifdef SUPPORT_UCP
3454          if (utf && fc > 127)
3455            foc = UCD_OTHERCASE(fc);
3456    #else
3457          if (utf && fc > 127)
3458            foc = fc;
3459    #endif /* SUPPORT_UCP */
3460          else
3461    #endif /* SUPPORT_UTF */
3462            foc = TABLE_GET(fc, md->fcc, fc);
3463    #endif /* COMPILE_PCRE8 */
3464    
3465        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3466          {          {
3467            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3468          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3469            {            {
3470            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3471            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3472            }            }
3473          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3474            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3475            eptr++;
3476          }          }
3477        if (min == max) continue;        if (min == max) continue;
3478        if (minimize)        if (minimize)
3479          {          {
3480          for (fi = min;; fi++)          for (fi = min;; fi++)
3481            {            {
3482              pcre_uint32 cc;               /* Faster than pcre_uchar */
3483            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3484            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3485            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3311  for (;;) Line 3488  for (;;)
3488              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3489              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3490              }              }
3491            if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3492              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3493              eptr++;
3494            }            }
3495          /* Control never gets here */          /* Control never gets here */
3496          }          }
# Line 3320  for (;;) Line 3499  for (;;)
3499          pp = eptr;          pp = eptr;
3500          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3501            {            {
3502              pcre_uint32 cc;               /* Faster than pcre_uchar */
3503            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3504              {              {
3505              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3506              break;              break;
3507              }              }
3508            if (fc != md->lcc[*eptr]) break;            cc = RAWUCHARTEST(eptr);
3509              if (fc != cc && foc != cc) break;
3510            eptr++;            eptr++;
3511            }            }
3512    
# Line 3353  for (;;) Line 3534  for (;;)
3534            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3535            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3536            }            }
3537          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3538          }          }
3539    
3540        if (min == max) continue;        if (min == max) continue;
# Line 3370  for (;;) Line 3551  for (;;)
3551              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3552              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3553              }              }
3554            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3555            }            }
3556          /* Control never gets here */          /* Control never gets here */
3557          }          }
# Line 3384  for (;;) Line 3565  for (;;)
3565              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3566              break;              break;
3567              }              }
3568            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3569            eptr++;            eptr++;
3570            }            }
3571          if (possessive) continue;          if (possessive) continue;
# Line 3410  for (;;) Line 3591  for (;;)
3591        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3592        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3593        }        }
3594      ecode++;  #ifdef SUPPORT_UTF
3595      GETCHARINCTEST(c, eptr);      if (utf)
     if (op == OP_NOTI)         /* The caseless case */  
3596        {        {
3597  #ifdef SUPPORT_UTF8        register pcre_uint32 ch, och;
3598        if (c < 256)  
3599  #endif        ecode++;
3600        c = md->lcc[c];        GETCHARINC(ch, ecode);
3601        if (md->lcc[*ecode++] == c) RRETURN(MATCH_NOMATCH);        GETCHARINC(c, eptr);
3602    
3603          if (op == OP_NOT)
3604            {
3605            if (ch == c) RRETURN(MATCH_NOMATCH);
3606            }
3607          else
3608            {
3609    #ifdef SUPPORT_UCP
3610            if (ch > 127)
3611              och = UCD_OTHERCASE(ch);
3612    #else
3613            if (ch > 127)
3614              och = ch;
3615    #endif /* SUPPORT_UCP */
3616            else
3617              och = TABLE_GET(ch, md->fcc, ch);
3618            if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3619            }
3620        }        }
3621      else    /* Caseful */      else
3622    #endif
3623        {        {
3624        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3625          c = *eptr++;
3626          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3627            RRETURN(MATCH_NOMATCH);
3628          ecode += 2;
3629        }        }
3630      break;      break;
3631    
# Line 3436  for (;;) Line 3639  for (;;)
3639      case OP_NOTEXACT:      case OP_NOTEXACT:
3640      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3641      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3642      ecode += 3;      ecode += 1 + IMM2_SIZE;
3643      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3644    
3645      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3446  for (;;) Line 3649  for (;;)
3649      min = 0;      min = 0;
3650      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3651      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3652      ecode += 3;      ecode += 1 + IMM2_SIZE;
3653      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3654    
3655      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3478  for (;;) Line 3681  for (;;)
3681      possessive = TRUE;      possessive = TRUE;
3682      min = 0;      min = 0;
3683      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3684      ecode += 3;      ecode += 1 + IMM2_SIZE;
3685      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3686    
3687      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3502  for (;;) Line 3705  for (;;)
3705      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3706    
3707      REPEATNOTCHAR:      REPEATNOTCHAR:
3708      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3709    
3710      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3711      since matching characters is likely to be quite common. First, ensure the      since matching characters is likely to be quite common. First, ensure the
# Line 3513  for (;;) Line 3716  for (;;)
3716      characters and work backwards. */      characters and work backwards. */
3717    
3718      DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3719        max, eptr));        max, (char *)eptr));
3720    
3721      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3722        {        {
3723        fc = md->lcc[fc];  #ifdef SUPPORT_UTF
3724    #ifdef SUPPORT_UCP
3725          if (utf && fc > 127)
3726            foc = UCD_OTHERCASE(fc);
3727    #else
3728          if (utf && fc > 127)
3729            foc = fc;
3730    #endif /* SUPPORT_UCP */
3731          else
3732    #endif /* SUPPORT_UTF */
3733            foc = TABLE_GET(fc, md->fcc, fc);
3734    
3735  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3736        /* UTF-8 mode */        if (utf)
       if (utf8)  
3737          {          {
3738          register unsigned int d;          register pcre_uint32 d;
3739          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3740            {            {
3741            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3532  for (;;) Line 3744  for (;;)
3744              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3745              }              }
3746            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3747            if (d < 256) d = md->lcc[d];            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) RRETURN(MATCH_NOMATCH);  
3748            }            }
3749          }          }
3750        else        else
3751  #endif  #endif
3752          /* Not UTF mode */
       /* Not UTF-8 mode */  
3753          {          {
3754          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3755            {            {
# Line 3548  for (;;) Line 3758  for (;;)
3758              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3759              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3760              }              }
3761            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3762              eptr++;
3763            }            }
3764          }          }
3765    
# Line 3556  for (;;) Line 3767  for (;;)
3767    
3768        if (minimize)        if (minimize)
3769          {          {
3770  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3771          /* UTF-8 mode */          if (utf)
         if (utf8)  
3772            {            {
3773            register unsigned int d;            register pcre_uint32 d;
3774            for (fi = min;; fi++)            for (fi = min;; fi++)
3775              {              {
3776              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3572  for (;;) Line 3782  for (;;)
3782                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3783                }                }
3784              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3785              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) RRETURN(MATCH_NOMATCH);  
3786              }              }
3787            }            }
3788          else          else
3789  #endif  #endif
3790          /* Not UTF-8 mode */          /* Not UTF mode */
3791            {            {
3792            for (fi = min;; fi++)            for (fi = min;; fi++)
3793              {              {
# Line 3590  for (;;) Line 3799  for (;;)
3799                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3800                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3801                }                }
3802              if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3803                eptr++;
3804              }              }
3805            }            }
3806          /* Control never gets here */          /* Control never gets here */
# Line 3602  for (;;) Line 3812  for (;;)
3812          {          {
3813          pp = eptr;          pp = eptr;
3814    
3815  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3816          /* UTF-8 mode */          if (utf)
         if (utf8)  
3817            {            {
3818            register unsigned int d;            register pcre_uint32 d;
3819            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3820              {              {
3821              int len = 1;              int len = 1;
# Line 3616  for (;;) Line 3825  for (;;)
3825                break;                break;
3826                }                }
3827              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3828              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) break;
             if (fc == d) break;  
3829              eptr += len;              eptr += len;
3830              }              }
3831          if (possessive) continue;            if (possessive) continue;
3832          for(;;)            for(;;)
3833              {              {
3834              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3835              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 3631  for (;;) Line 3839  for (;;)
3839            }            }
3840          else          else
3841  #endif  #endif
3842          /* Not UTF-8 mode */          /* Not UTF mode */
3843            {            {
3844            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3845              {              {
# Line 3640  for (;;) Line 3848  for (;;)
3848                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3849                break;                break;
3850                }                }
3851              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3852              eptr++;              eptr++;
3853              }              }
3854            if (possessive) continue;            if (possessive) continue;
# Line 3661  for (;;) Line 3869  for (;;)
3869    
3870      else      else
3871        {        {
3872  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3873        /* UTF-8 mode */        if (utf)
       if (utf8)  
3874          {          {
3875          register unsigned int d;          register pcre_uint32 d;
3876          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3877            {            {
3878            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3679  for (;;) Line 3886  for (;;)
3886          }          }
3887        else        else
3888  #endif  #endif
3889        /* Not UTF-8 mode */        /* Not UTF mode */
3890          {          {
3891          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3892            {            {
# Line 3696  for (;;) Line 3903  for (;;)
3903    
3904        if (minimize)        if (minimize)
3905          {          {
3906  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3907          /* UTF-8 mode */          if (utf)
         if (utf8)  
3908            {            {
3909            register unsigned int d;            register pcre_uint32 d;
3910            for (fi = min;; fi++)            for (fi = min;; fi++)
3911              {              {
3912              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3717  for (;;) Line 3923  for (;;)
3923            }            }
3924          else          else
3925  #endif  #endif
3926          /* Not UTF-8 mode */          /* Not UTF mode */
3927            {            {
3928            for (fi = min;; fi++)            for (fi = min;; fi++)
3929              {              {
# Line 3741  for (;;) Line 3947  for (;;)
3947          {          {
3948          pp = eptr;          pp = eptr;
3949    
3950  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3951          /* UTF-8 mode */          if (utf)
         if (utf8)  
3952            {            {
3953            register unsigned int d;            register pcre_uint32 d;
3954            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3955              {              {
3956              int len = 1;              int len = 1;
# Line 3769  for (;;) Line 3974  for (;;)
3974            }            }
3975          else          else
3976  #endif  #endif
3977          /* Not UTF-8 mode */          /* Not UTF mode */
3978            {            {
3979            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3980              {              {
# Line 3802  for (;;) Line 4007  for (;;)
4007      case OP_TYPEEXACT:      case OP_TYPEEXACT:
4008      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
4009      minimize = TRUE;      minimize = TRUE;
4010      ecode += 3;      ecode += 1 + IMM2_SIZE;
4011      goto REPEATTYPE;      goto REPEATTYPE;
4012    
4013      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3810  for (;;) Line 4015  for (;;)
4015      min = 0;      min = 0;
4016      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4017      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
4018      ecode += 3;      ecode += 1 + IMM2_SIZE;
4019      goto REPEATTYPE;      goto REPEATTYPE;
4020    
4021      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3838  for (;;) Line 4043  for (;;)
4043      possessive = TRUE;      possessive = TRUE;
4044      min = 0;      min = 0;
4045      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4046      ecode += 3;      ecode += 1 + IMM2_SIZE;
4047      goto REPEATTYPE;      goto REPEATTYPE;
4048    
4049      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 4021  for (;;) Line 4226  for (;;)
4226              }              }
4227            break;            break;
4228    
4229              case PT_CLIST:
4230              for (i = 1; i <= min; i++)
4231                {
4232                const pcre_uint32 *cp;
4233                if (eptr >= md->end_subject)
4234                  {
4235                  SCHECK_PARTIAL();
4236                  RRETURN(MATCH_NOMATCH);
4237                  }
4238                GETCHARINCTEST(c, eptr);
4239                cp = PRIV(ucd_caseless_sets) + prop_value;
4240                for (;;)
4241                  {
4242                  if (c < *cp)
4243                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4244                  if (c == *cp++)
4245                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4246                  }
4247                }
4248              break;
4249    
4250            /* This should not occur */            /* This should not occur */
4251    
4252            default:            default:
# Line 4040  for (;;) Line 4266  for (;;)
4266              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4267              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4268              }              }
4269            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4270              {              {
4271              int len = 1;              int lgb, rgb;
4272              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4273              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4274              eptr += len;             while (eptr < md->end_subject)
4275                  {
4276                  int len = 1;
4277                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4278                  rgb = UCD_GRAPHBREAK(c);
4279                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4280                  lgb = rgb;
4281                  eptr += len;
4282                  }
4283              }              }
4284              CHECK_PARTIAL();
4285            }            }
4286          }          }
4287    
# Line 4057  for (;;) Line 4290  for (;;)
4290    
4291  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4292    
4293  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4294        if (utf8) switch(ctype)        if (utf) switch(ctype)
4295          {          {
4296          case OP_ANY:          case OP_ANY:
4297          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 4069  for (;;) Line 4302  for (;;)
4302              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4303              }              }
4304            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4305              if (md->partial != 0 &&
4306                  eptr + 1 >= md->end_subject &&
4307                  NLBLOCK->nltype == NLTYPE_FIXED &&
4308                  NLBLOCK->nllen == 2 &&
4309                  RAWUCHAR(eptr) == NLBLOCK->nl[0])
4310                {
4311                md->hitend = TRUE;
4312                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4313                }
4314            eptr++;            eptr++;
4315            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4316            }            }
4317          break;          break;
4318    
# Line 4083  for (;;) Line 4325  for (;;)
4325              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4326              }              }
4327            eptr++;            eptr++;
4328            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4329            }            }
4330          break;          break;
4331    
# Line 4105  for (;;) Line 4347  for (;;)
4347              {              {
4348              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4349    
4350              case 0x000d:              case CHAR_CR:
4351              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
4352              break;              break;
4353    
4354              case 0x000a:              case CHAR_LF:
4355              break;              break;
4356    
4357              case 0x000b:              case CHAR_VT:
4358              case 0x000c:              case CHAR_FF:
4359              case 0x0085:              case CHAR_NEL:
4360    #ifndef EBCDIC
4361              case 0x2028:              case 0x2028:
4362              case 0x2029:              case 0x2029:
4363    #endif  /* Not EBCDIC */
4364              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4365              break;              break;
4366              }              }
# Line 4134  for (;;) Line 4378  for (;;)
4378            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4379            switch(c)            switch(c)
4380              {              {
4381                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4382              default: break;              default: break;
             case 0x09:      /* HT */  
             case 0x20:      /* SPACE */  
             case 0xa0:      /* NBSP */  
             case 0x1680:    /* OGHAM SPACE MARK */  
             case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
             case 0x2000:    /* EN QUAD */  
             case 0x2001:    /* EM QUAD */  
             case 0x2002:    /* EN SPACE */  
             case 0x2003:    /* EM SPACE */  
             case 0x2004:    /* THREE-PER-EM SPACE */  
             case 0x2005:    /* FOUR-PER-EM SPACE */  
             case 0x2006:    /* SIX-PER-EM SPACE */  
             case 0x2007:    /* FIGURE SPACE */  
             case 0x2008:    /* PUNCTUATION SPACE */  
             case 0x2009:    /* THIN SPACE */  
             case 0x200A:    /* HAIR SPACE */  
             case 0x202f:    /* NARROW NO-BREAK SPACE */  
             case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
             case 0x3000:    /* IDEOGRAPHIC SPACE */  
             RRETURN(MATCH_NOMATCH);  
4383              }              }
4384            }            }
4385          break;          break;
# Line 4170  for (;;) Line 4395  for (;;)
4395            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4396            switch(c)            switch(c)
4397              {              {
4398                HSPACE_CASES: break;  /* Byte and multibyte cases */
4399              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
             case 0x09:      /* HT */  
             case 0x20:      /* SPACE */  
             case 0xa0:      /* NBSP */  
             case 0x1680:    /* OGHAM SPACE MARK */  
             case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
             case 0x2000:    /* EN QUAD */  
             case 0x2001:    /* EM QUAD */  
             case 0x2002:    /* EN SPACE */  
             case 0x2003:    /* EM SPACE */  
             case 0x2004:    /* THREE-PER-EM SPACE */  
             case 0x2005:    /* FOUR-PER-EM SPACE */  
             case 0x2006:    /* SIX-PER-EM SPACE */  
             case 0x2007:    /* FIGURE SPACE */  
             case 0x2008:    /* PUNCTUATION SPACE */  
             case 0x2009:    /* THIN SPACE */  
             case 0x200A:    /* HAIR SPACE */  
             case 0x202f:    /* NARROW NO-BREAK SPACE */  
             case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
             case 0x3000:    /* IDEOGRAPHIC SPACE */  
             break;  
4400              }              }
4401            }            }
4402          break;          break;
# Line 4206  for (;;) Line 4412  for (;;)
4412            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4413            switch(c)            switch(c)
4414              {              {
4415                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4416              default: break;              default: break;
             case 0x0a:      /* LF */  
             case 0x0b:      /* VT */  
             case 0x0c:      /* FF */  
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
             RRETURN(MATCH_NOMATCH);  
4417              }              }
4418            }            }
4419          break;          break;
# Line 4230  for (;;) Line 4429  for (;;)
4429            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4430            switch(c)            switch(c)
4431              {              {
4432                VSPACE_CASES: break;
4433              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
             case 0x0a:      /* LF */  
             case 0x0b:      /* VT */  
             case 0x0c:      /* FF */  
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
             break;  
4434              }              }
4435            }            }
4436          break;          break;
# Line 4260  for (;;) Line 4452  for (;;)
4452          case OP_DIGIT:          case OP_DIGIT:
4453          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4454            {            {
4455              pcre_uchar cc;
4456    
4457            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4458              {              {
4459              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4460              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4461              }              }
4462            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            cc = RAWUCHAR(eptr);
4463              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4464              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4465              eptr++;
4466            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
4467            }            }
4468          break;          break;
# Line 4274  for (;;) Line 4470  for (;;)
4470          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4471          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4472            {            {
4473              pcre_uchar cc;
4474    
4475            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4476              {              {
4477              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4478              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4479              }              }
4480            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = RAWUCHAR(eptr);
4481              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4482              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4483            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4484              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4485            }            }
4486          break;          break;
4487    
4488          case OP_WHITESPACE:          case OP_WHITESPACE:
4489          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4490            {            {
4491              pcre_uchar cc;
4492    
4493            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4494              {              {
4495              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4496              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4497              }              }
4498            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            cc = RAWUCHAR(eptr);
4499              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4500              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4501              eptr++;
4502            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
4503            }            }
4504          break;          break;
# Line 4302  for (;;) Line 4506  for (;;)
4506          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4507          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4508            {            {
4509              pcre_uchar cc;
4510    
4511            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4512              {              {
4513              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4514              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4515              }              }
4516            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = RAWUCHAR(eptr);
4517              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4518              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4519            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4520              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4521            }            }
4522          break;          break;
4523    
4524          case OP_WORDCHAR:          case OP_WORDCHAR:
4525          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4526            {            {
4527              pcre_uchar cc;
4528    
4529            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4530              {              {
4531              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4532              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4533              }              }
4534            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)            cc = RAWUCHAR(eptr);
4535              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4536              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4537              eptr++;
4538            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
4539            }            }
4540          break;          break;
# Line 4332  for (;;) Line 4544  for (;;)
4544          }  /* End switch(ctype) */          }  /* End switch(ctype) */
4545    
4546        else        else
4547  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF */
4548    
4549        /* 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
4550        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. */
# Line 4348  for (;;) Line 4560  for (;;)
4560              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4561              }              }
4562            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4563              if (md->partial != 0 &&
4564                  eptr + 1 >= md->end_subject &&
4565                  NLBLOCK->nltype == NLTYPE_FIXED &&
4566                  NLBLOCK->nllen == 2 &&
4567                  *eptr == NLBLOCK->nl[0])
4568                {
4569                md->hitend = TRUE;
4570                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4571                }
4572            eptr++;            eptr++;
4573            }            }
4574          break;          break;
# Line 4382  for (;;) Line 4603  for (;;)
4603              {              {
4604              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4605    
4606              case 0x000d:              case CHAR_CR:
4607              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4608              break;              break;
4609    
4610              case 0x000a:              case CHAR_LF:
4611              break;              break;
4612    
4613              case 0x000b:              case CHAR_VT:
4614              case 0x000c:              case CHAR_FF:
4615              case 0x0085:              case CHAR_NEL:
4616    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4617                case 0x2028:
4618                case 0x2029:
4619    #endif
4620              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4621              break;              break;
4622              }              }
# Line 4409  for (;;) Line 4634  for (;;)
4634            switch(*eptr++)            switch(*eptr++)
4635              {              {
4636              default: break;              default: break;
4637              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4638              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4639              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4640    #endif
4641              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4642              }              }
4643            }            }
# Line 4428  for (;;) Line 4654  for (;;)
4654            switch(*eptr++)            switch(*eptr++)
4655              {              {
4656              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4657              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4658              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4659              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4660    #endif
4661              break;              break;
4662              }              }
4663            }            }
# Line 4446  for (;;) Line 4673  for (;;)
4673              }              }
4674            switch(*eptr++)            switch(*eptr++)
4675              {              {
4676              default: break;              VSPACE_BYTE_CASES:
4677              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4678              case 0x0b:      /* VT */              VSPACE_MULTIBYTE_CASES:
4679              case 0x0c:      /* FF */  #endif
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
4680              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4681                default: break;
4682              }              }
4683            }            }
4684          break;          break;
# Line 4468  for (;;) Line 4694  for (;;)
4694            switch(*eptr++)            switch(*eptr++)
4695              {              {
4696              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4697              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4698              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4699              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
4700              case 0x0d:      /* CR */  #endif
             case 0x85:      /* NEL */  
4701              break;              break;
4702              }              }
4703            }            }
# Line 4486  for (;;) Line 4711  for (;;)
4711              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4712              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4713              }              }
4714            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0)
4715                RRETURN(MATCH_NOMATCH);
4716              eptr++;
4717            }            }
4718          break;          break;
4719    
# Line 4498  for (;;) Line 4725  for (;;)
4725              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4726              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4727              }              }
4728            if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0)
4729                RRETURN(MATCH_NOMATCH);
4730              eptr++;
4731            }            }
4732          break;          break;
4733    
# Line 4510  for (;;) Line 4739  for (;;)
4739              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4740              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4741              }              }
4742            if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0)
4743                RRETURN(MATCH_NOMATCH);
4744              eptr++;
4745            }            }
4746          break;          break;
4747    
# Line 4522  for (;;) Line 4753  for (;;)
4753              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4754              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4755              }              }
4756            if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0)
4757                RRETURN(MATCH_NOMATCH);
4758              eptr++;
4759            }            }
4760          break;          break;
4761    
# Line 4534  for (;;) Line 4767  for (;;)
4767              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4768              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4769              }              }
4770            if ((md->ctypes[*eptr++] & ctype_word) != 0)            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0)
4771              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4772              eptr++;
4773            }            }
4774          break;          break;
4775    
# Line 4547  for (;;) Line 4781  for (;;)
4781              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4782              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4783              }              }
4784            if ((md->ctypes[*eptr++] & ctype_word) == 0)            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0)
4785              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4786              eptr++;
4787            }            }
4788          break;          break;
4789    
# Line 4739  for (;;) Line 4974  for (;;)
4974              }              }
4975            /* Control never gets here */            /* Control never gets here */
4976    
4977            /* This should never occur */            case PT_CLIST:
4978              for (fi = min;; fi++)
4979                {
4980                const pcre_uint32 *cp;
4981                RMATCH(eptr, ecode, offset_top, md, eptrb, RM67);
4982                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4983                if (fi >= max) RRETURN(MATCH_NOMATCH);
4984                if (eptr >= md->end_subject)
4985                  {
4986                  SCHECK_PARTIAL();
4987                  RRETURN(MATCH_NOMATCH);
4988                  }
4989                GETCHARINCTEST(c, eptr);
4990                cp = PRIV(ucd_caseless_sets) + prop_value;
4991                for (;;)
4992                  {
4993                  if (c < *cp)
4994                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4995                  if (c == *cp++)
4996                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4997                  }
4998                }
4999              /* Control never gets here */
5000    
5001              /* This should never occur */
5002            default:            default:
5003            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5004            }            }
# Line 4761  for (;;) Line 5019  for (;;)
5019              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5020              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5021              }              }
5022            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
5023              {              {
5024              int len = 1;              int lgb, rgb;
5025              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5026              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5027              eptr += len;              while (eptr < md->end_subject)
5028                  {
5029                  int len = 1;
5030                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5031                  rgb = UCD_GRAPHBREAK(c);
5032                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5033                  lgb = rgb;
5034                  eptr += len;
5035                  }
5036              }              }
5037              CHECK_PARTIAL();
5038            }            }
5039          }          }
5040        else        else
5041  #endif     /* SUPPORT_UCP */  #endif     /* SUPPORT_UCP */
5042    
5043  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5044        /* UTF-8 mode */        if (utf)
       if (utf8)  
5045          {          {
5046          for (fi = min;; fi++)          for (fi = min;; fi++)
5047            {            {
# Line 4794  for (;;) Line 5058  for (;;)
5058            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5059            switch(ctype)            switch(ctype)
5060              {              {
5061              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5062                if (md->partial != 0 &&    /* Take care with CRLF partial */
5063                    eptr >= md->end_subject &&
5064                    NLBLOCK->nltype == NLTYPE_FIXED &&
5065                    NLBLOCK->nllen == 2 &&
5066                    c == NLBLOCK->nl[0])
5067                  {
5068                  md->hitend = TRUE;
5069                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5070                  }
5071                break;
5072    
5073              case OP_ALLANY:              case OP_ALLANY:
5074              case OP_ANYBYTE:              case OP_ANYBYTE:
5075              break;              break;
# Line 4803  for (;;) Line 5078  for (;;)
5078              switch(c)              switch(c)
5079                {                {
5080                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5081                case 0x000d:                case CHAR_CR:
5082                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
5083                break;                break;
5084                case 0x000a:  
5085                  case CHAR_LF:
5086                break;                break;
5087    
5088                case 0x000b:                case CHAR_VT:
5089                case 0x000c:                case CHAR_FF:
5090                case 0x0085:                case CHAR_NEL:
5091    #ifndef EBCDIC
5092                case 0x2028:                case 0x2028:
5093                case 0x2029:                case 0x2029:
5094    #endif  /* Not EBCDIC */
5095                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5096                break;                break;
5097                }                }
# Line 4822  for (;;) Line 5100  for (;;)
5100              case OP_NOT_HSPACE:              case OP_NOT_HSPACE:
5101              switch(c)              switch(c)
5102                {                {
5103                  HSPACE_CASES: RRETURN(MATCH_NOMATCH);
5104                default: break;                default: break;
               case 0x09:      /* HT */  
               case 0x20:      /* SPACE */  
               case 0xa0:      /* NBSP */  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
               RRETURN(MATCH_NOMATCH);  
5105                }                }
5106              break;              break;
5107    
5108              case OP_HSPACE:              case OP_HSPACE:
5109              switch(c)              switch(c)
5110                {                {
5111                  HSPACE_CASES: break;
5112                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
               case 0x09:      /* HT */  
               case 0x20:      /* SPACE */  
               case 0xa0:      /* NBSP */  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
               break;  
5113                }                }
5114              break;              break;
5115    
5116              case OP_NOT_VSPACE:              case OP_NOT_VSPACE:
5117              switch(c)              switch(c)
5118                {                {
5119                  VSPACE_CASES: RRETURN(MATCH_NOMATCH);
5120                default: break;                default: break;
               case 0x0a:      /* LF */  
               case 0x0b:      /* VT */  
               case 0x0c:      /* FF */  
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
               RRETURN(MATCH_NOMATCH);  
5121                }                }
5122              break;              break;
5123    
5124              case OP_VSPACE:              case OP_VSPACE:
5125              switch(c)              switch(c)
5126                {                {
5127                  VSPACE_CASES: break;
5128                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
               case 0x0a:      /* LF */  
               case 0x0b:      /* VT */  
               case 0x0c:      /* FF */  
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
               break;  
5129                }                }
5130              break;              break;
5131    
# Line 4919  for (;;) Line 5145  for (;;)
5145              break;              break;
5146    
5147              case OP_WHITESPACE:              case OP_WHITESPACE:
5148              if  (c >= 256 || (md->ctypes[c] & ctype_space) == 0)              if (c >= 256 || (md->ctypes[c] & ctype_space) == 0)
5149                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5150              break;              break;
5151    
# Line 4940  for (;;) Line 5166  for (;;)
5166          }          }
5167        else        else
5168  #endif  #endif
5169        /* Not UTF-8 mode */        /* Not UTF mode */
5170          {          {
5171          for (fi = min;; fi++)          for (fi = min;; fi++)
5172            {            {
# Line 4957  for (;;) Line 5183  for (;;)
5183            c = *eptr++;            c = *eptr++;
5184            switch(ctype)            switch(ctype)
5185              {              {
5186              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5187                if (md->partial != 0 &&    /* Take care with CRLF partial */
5188                    eptr >= md->end_subject &&
5189                    NLBLOCK->nltype == NLTYPE_FIXED &&
5190                    NLBLOCK->nllen == 2 &&
5191                    c == NLBLOCK->nl[0])
5192                  {
5193                  md->hitend = TRUE;
5194                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5195                  }
5196                break;
5197    
5198              case OP_ALLANY:              case OP_ALLANY:
5199              case OP_ANYBYTE:              case OP_ANYBYTE:
5200              break;              break;
# Line 4966  for (;;) Line 5203  for (;;)
5203              switch(c)              switch(c)
5204                {                {
5205                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5206                case 0x000d:                case CHAR_CR:
5207                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5208                break;                break;
5209    
5210                case 0x000a:                case CHAR_LF:
5211                break;                break;
5212    
5213                case 0x000b:                case CHAR_VT:
5214                case 0x000c:                case CHAR_FF:
5215                case 0x0085:                case CHAR_NEL:
5216    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5217                  case 0x2028:
5218                  case 0x2029:
5219    #endif
5220                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5221                break;                break;
5222                }                }
# Line 4985  for (;;) Line 5226  for (;;)
5226              switch(c)              switch(c)
5227                {                {
5228                default: break;                default: break;
5229                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5230                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5231                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
5232    #endif
5233                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5234                }                }
5235              break;              break;
# Line 4996  for (;;) Line 5238  for (;;)
5238              switch(c)              switch(c)
5239                {                {
5240                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5241                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5242                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5243                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
5244    #endif
5245                break;                break;
5246                }                }
5247              break;              break;
# Line 5007  for (;;) Line 5250  for (;;)
5250              switch(c)              switch(c)
5251                {                {
5252                default: break;                default: break;
5253                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5254                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5255                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
5256                case 0x0d:      /* CR */  #endif
               case 0x85:      /* NEL */  
5257                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5258                }                }
5259              break;              break;
# Line 5020  for (;;) Line 5262  for (;;)
5262              switch(c)              switch(c)
5263                {                {
5264                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5265                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5266                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5267                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
5268                case 0x0d:      /* CR */  #endif
               case 0x85:      /* NEL */  
5269                break;                break;
5270                }                }
5271              break;              break;
5272    
5273              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
5274              if ((md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
5275              break;              break;
5276    
5277              case OP_DIGIT:              case OP_DIGIT:
5278              if ((md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);
5279              break;              break;
5280    
5281              case OP_NOT_WHITESPACE:              case OP_NOT_WHITESPACE:
5282              if ((md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);
5283              break;              break;
5284    
5285              case OP_WHITESPACE:              case OP_WHITESPACE:
5286              if  ((md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);
5287              break;              break;
5288    
5289              case OP_NOT_WORDCHAR:              case OP_NOT_WORDCHAR:
5290              if ((md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);
5291              break;              break;
5292    
5293              case OP_WORDCHAR:              case OP_WORDCHAR:
5294              if ((md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);
5295              break;              break;
5296    
5297              default:              default:
# Line 5227  for (;;) Line 5468  for (;;)
5468              }              }
5469            break;            break;
5470    
5471              case PT_CLIST:
5472              for (i = min; i < max; i++)
5473                {
5474                const pcre_uint32 *cp;
5475                int len = 1;
5476                if (eptr >= md->end_subject)
5477                  {
5478                  SCHECK_PARTIAL();
5479                  break;
5480                  }
5481                GETCHARLENTEST(c, eptr, len);
5482                cp = PRIV(ucd_caseless_sets) + prop_value;
5483                for (;;)
5484                  {
5485                  if (c < *cp)
5486                    { if (prop_fail_result) break; else goto GOT_MAX; }
5487                  if (c == *cp++)
5488                    { if (prop_fail_result) goto GOT_MAX; else break; }
5489                  }
5490                eptr += len;
5491                }
5492              GOT_MAX:
5493              break;
5494    
5495            default:            default:
5496            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5497            }            }
# Line 5239  for (;;) Line 5504  for (;;)
5504            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5505            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5506            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5507            if (utf8) BACKCHAR(eptr);            if (utf) BACKCHAR(eptr);
5508            }            }
5509          }          }
5510    
# Line 5250  for (;;) Line 5515  for (;;)
5515          {          {
5516          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5517            {            {
           int len = 1;  
5518            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5519              {              {
5520              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5521              break;              break;
5522              }              }
5523            if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
           if (UCD_CATEGORY(c) == ucp_M) break;  
           eptr += len;  
           while (eptr < md->end_subject)  
5524              {              {
5525              len = 1;              int lgb, rgb;
5526              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5527              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5528              eptr += len;              while (eptr < md->end_subject)
5529                  {
5530                  int len = 1;
5531                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5532                  rgb = UCD_GRAPHBREAK(c);
5533                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5534                  lgb = rgb;
5535                  eptr += len;
5536                  }
5537              }              }
5538              CHECK_PARTIAL();
5539            }            }
5540    
5541          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
# Line 5279  for (;;) Line 5549  for (;;)
5549            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5550            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
5551              {              {
5552              if (!utf8) c = *eptr; else              if (!utf) c = *eptr; else
5553                {                {
5554                BACKCHAR(eptr);                BACKCHAR(eptr);
5555                GETCHAR(c, eptr);                GETCHAR(c, eptr);
# Line 5293  for (;;) Line 5563  for (;;)
5563        else        else
5564  #endif   /* SUPPORT_UCP */  #endif   /* SUPPORT_UCP */
5565    
5566  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5567        /* UTF-8 mode */        if (utf)
   
       if (utf8)  
5568          {          {
5569          switch(ctype)          switch(ctype)
5570            {            {
# Line 5311  for (;;) Line 5579  for (;;)
5579                  break;                  break;
5580                  }                  }
5581                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5582                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5583                      eptr + 1 >= md->end_subject &&
5584                      NLBLOCK->nltype == NLTYPE_FIXED &&
5585                      NLBLOCK->nllen == 2 &&
5586                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5587                    {
5588                    md->hitend = TRUE;
5589                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5590                    }
5591                eptr++;                eptr++;
5592                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5593                }                }
5594              }              }
5595    
# Line 5328  for (;;) Line 5605  for (;;)
5605                  break;                  break;
5606                  }                  }
5607                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5608                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5609                      eptr + 1 >= md->end_subject &&
5610                      NLBLOCK->nltype == NLTYPE_FIXED &&
5611                      NLBLOCK->nllen == 2 &&
5612                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5613                    {
5614                    md->hitend = TRUE;
5615                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5616                    }
5617                eptr++;                eptr++;
5618                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5619                }                }
5620              }              }
5621            break;            break;
# Line 5345  for (;;) Line 5631  for (;;)
5631                  break;                  break;
5632                  }                  }
5633                eptr++;                eptr++;
5634                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5635                }                }
5636              }              }
5637            else            else
# Line 5377  for (;;) Line 5663  for (;;)
5663                break;                break;
5664                }                }
5665              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5666              if (c == 0x000d)              if (c == CHAR_CR)
5667                {                {
5668                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5669                if (*eptr == 0x000a) eptr++;                if (RAWUCHAR(eptr) == CHAR_LF) eptr++;
5670                }                }
5671              else              else
5672                {                {
5673                if (c != 0x000a &&                if (c != CHAR_LF &&
5674                    (md->bsr_anycrlf ||                    (md->bsr_anycrlf ||
5675                     (c != 0x000b && c != 0x000c &&                     (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5676                      c != 0x0085 && c != 0x2028 && c != 0x2029)))  #ifndef EBCDIC
5677                        && c != 0x2028 && c != 0x2029
5678    #endif  /* Not EBCDIC */
5679                        )))
5680                  break;                  break;
5681                eptr += len;                eptr += len;
5682                }                }
# Line 5408  for (;;) Line 5697  for (;;)
5697              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5698              switch(c)              switch(c)
5699                {                {
5700                  HSPACE_CASES: gotspace = TRUE; break;
5701                default: gotspace = FALSE; break;                default: gotspace = FALSE; break;
               case 0x09:      /* HT */  
               case 0x20:      /* SPACE */  
               case 0xa0:      /* NBSP */  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
               gotspace = TRUE;  
               break;  
5702                }                }
5703              if (gotspace == (ctype == OP_NOT_HSPACE)) break;              if (gotspace == (ctype == OP_NOT_HSPACE)) break;
5704              eptr += len;              eptr += len;
# Line 5450  for (;;) Line 5719  for (;;)
5719              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5720              switch(c)              switch(c)
5721                {                {
5722                  VSPACE_CASES: gotspace = TRUE; break;
5723                default: gotspace = FALSE; break;                default: gotspace = FALSE; break;
               case 0x0a:      /* LF */  
               case 0x0b:      /* VT */  
               case 0x0c:      /* FF */  
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
               gotspace = TRUE;  
               break;  
5724                }                }
5725              if (gotspace == (ctype == OP_NOT_VSPACE)) break;              if (gotspace == (ctype == OP_NOT_VSPACE)) break;
5726              eptr += len;              eptr += len;
# Line 5573  for (;;) Line 5834  for (;;)
5834            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5835            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5836            BACKCHAR(eptr);            BACKCHAR(eptr);
5837            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && RAWUCHAR(eptr) == CHAR_NL &&
5838                eptr[-1] == '\r') eptr--;                RAWUCHAR(eptr - 1) == CHAR_CR) eptr--;
5839            }            }
5840          }          }
5841        else        else
5842  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
5843          /* Not UTF mode */
       /* Not UTF-8 mode */  
5844          {          {
5845          switch(ctype)          switch(ctype)
5846            {            {
# Line 5593  for (;;) Line 5853  for (;;)
5853                break;                break;
5854                }                }
5855              if (IS_NEWLINE(eptr)) break;              if (IS_NEWLINE(eptr)) break;
5856                if (md->partial != 0 &&    /* Take care with CRLF partial */
5857                    eptr + 1 >= md->end_subject &&
5858                    NLBLOCK->nltype == NLTYPE_FIXED &&
5859                    NLBLOCK->nllen == 2 &&
5860                    *eptr == NLBLOCK->nl[0])
5861                  {
5862                  md->hitend = TRUE;
5863                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5864                  }
5865              eptr++;              eptr++;
5866              }              }
5867            break;            break;
# Line 5617  for (;;) Line 5886  for (;;)
5886                break;                break;
5887                }                }
5888              c = *eptr;              c = *eptr;
5889              if (c == 0x000d)              if (c == CHAR_CR)
5890                {                {
5891                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5892                if (*eptr == 0x000a) eptr++;                if (*eptr == CHAR_LF) eptr++;
5893                }                }
5894              else              else
5895                {                {
5896                if (c != 0x000a &&                if (c != CHAR_LF && (md->bsr_anycrlf ||
5897                    (md->bsr_anycrlf ||                   (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5898                      (c != 0x000b && c != 0x000c && c != 0x0085)))  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5899                  break;                   && c != 0x2028 && c != 0x2029
5900    #endif
5901                     ))) break;
5902                eptr++;                eptr++;
5903                }                }
5904              }              }
# Line 5641  for (;;) Line 5912  for (;;)
5912                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5913                break;                break;
5914                }                }
5915              c = *eptr;              switch(*eptr)
5916              if (c == 0x09 || c == 0x20 || c == 0xa0) break;                {
5917              eptr++;                default: eptr++; break;
5918                  HSPACE_BYTE_CASES:
5919    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5920                  HSPACE_MULTIBYTE_CASES:
5921    #endif
5922                  goto ENDLOOP00;
5923                  }
5924              }              }
5925              ENDLOOP00:
5926            break;            break;
5927    
5928            case OP_HSPACE:            case OP_HSPACE:
# Line 5655  for (;;) Line 5933  for (;;)
5933                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5934                break;                break;
5935                }                }
5936              c = *eptr;              switch(*eptr)
5937              if (c != 0x09 && c != 0x20 && c != 0xa0) break;                {
5938              eptr++;                default: goto ENDLOOP01;
5939                  HSPACE_BYTE_CASES:
5940    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5941                  HSPACE_MULTIBYTE_CASES:
5942    #endif
5943                  eptr++; break;
5944                  }
5945              }              }
5946              ENDLOOP01:
5947            break;            break;
5948    
5949            case OP_NOT_VSPACE:            case OP_NOT_VSPACE:
# Line 5669  for (;;) Line 5954  for (;;)
5954                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5955                break;                break;
5956                }                }
5957              c = *eptr;              switch(*eptr)
5958              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85)                {