/[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 788 by ph10, Tue Dec 6 15:38:01 2011 UTC revision 1265 by ph10, Sun Mar 3 10:42:46 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) 1997-2013 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_uint32 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 (int)(eptr - eptr_start);  return (int)(eptr - eptr_start);
# 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, RM68 };
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 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 758  for (;;) Line 823  for (;;)
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);
# 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 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 1161  for (;;) Line 1234  for (;;)
1234      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1235      matched_once = FALSE;      matched_once = FALSE;
1236      code_offset = (int)(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          codelink -= PRIV(OP_lengths)[OP_CALLOUT];
1322        }        }
1323    
1324      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1250  for (;;) Line 1334  for (;;)
1334          }          }
1335        else        else
1336          {          {
1337          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          unsigned int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1338          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1339    
1340          /* 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 1344  for (;;)
1344    
1345          if (!condition && condcode == OP_NRREF)          if (!condition && condcode == OP_NRREF)
1346            {            {
1347            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1348            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1349              {              {
1350              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 1273  for (;;) Line 1357  for (;;)
1357    
1358            if (i < md->name_count)            if (i < md->name_count)
1359              {              {
1360              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1361              while (slotB > md->name_table)              while (slotB > md->name_table)
1362                {                {
1363                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1364                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1365                  {                  {
1366                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1367                  if (condition) break;                  if (condition) break;
# Line 1293  for (;;) Line 1377  for (;;)
1377                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1378                  {                  {
1379                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1380                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1381                    {                    {
1382                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1383                    if (condition) break;                    if (condition) break;
# Line 1306  for (;;) Line 1390  for (;;)
1390    
1391          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1392    
1393          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1394          }          }
1395        }        }
1396    
# Line 1322  for (;;) Line 1406  for (;;)
1406    
1407        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1408          {          {
1409          int refno = offset >> 1;          unsigned int refno = offset >> 1;
1410          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1411    
1412          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1413            {            {
# Line 1337  for (;;) Line 1421  for (;;)
1421    
1422          if (i < md->name_count)          if (i < md->name_count)
1423            {            {
1424            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1425            while (slotB > md->name_table)            while (slotB > md->name_table)
1426              {              {
1427              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1428              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1429                {                {
1430                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1431                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1359  for (;;) Line 1443  for (;;)
1443              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1444                {                {
1445                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1446                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1447                  {                  {
1448                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1449                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1374  for (;;) Line 1458  for (;;)
1458    
1459        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1460    
1461        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1462        }        }
1463    
1464      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1450  for (;;) Line 1534  for (;;)
1534      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1535    
1536      case OP_CLOSE:      case OP_CLOSE:
1537      number = GET2(ecode, 1);      number = GET2(ecode, 1);   /* Must be less than 65536 */
1538      offset = number << 1;      offset = number << 1;
1539    
1540  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 1458  for (;;) Line 1542  for (;;)
1542        printf("\n");        printf("\n");
1543  #endif  #endif
1544    
1545      md->capture_last = number;      md->capture_last = (md->capture_last & OVFLMASK) | number;
1546      if (offset >= md->offset_max) md->offset_overflow = TRUE; else      if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1547        {        {
1548        md->offset_vector[offset] =        md->offset_vector[offset] =
1549          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
1550        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1551        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1552        }        }
1553      ecode += 3;      ecode += 1 + IMM2_SIZE;
1554      break;      break;
1555    
1556    
# Line 1513  for (;;) Line 1597  for (;;)
1597    
1598      case OP_ASSERT:      case OP_ASSERT:
1599      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1600        save_mark = md->mark;
1601      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1602        {        {
1603        condassert = TRUE;        condassert = TRUE;
# Line 1528  for (;;) Line 1613  for (;;)
1613          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1614          break;          break;
1615          }          }
1616          md->mark = save_mark;
1617    
1618        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* A COMMIT failure must fail the entire assertion, without trying any
1619        as NOMATCH. */        subsequent branches. */
1620    
1621          if (rrc == MATCH_COMMIT) RRETURN(MATCH_NOMATCH);
1622    
1623          /* PCRE does not allow THEN to escape beyond an assertion; it
1624          is treated as NOMATCH. */
1625    
1626        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1627        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
# Line 1557  for (;;) Line 1648  for (;;)
1648    
1649      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1650      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1651        save_mark = md->mark;
1652      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1653        {        {
1654        condassert = TRUE;        condassert = TRUE;
# Line 1567  for (;;) Line 1659  for (;;)
1659      do      do
1660        {        {
1661        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1662          md->mark = save_mark;
1663        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1664        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1665          {          {
# Line 1593  for (;;) Line 1686  for (;;)
1686      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1687    
1688      case OP_REVERSE:      case OP_REVERSE:
1689  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1690      if (utf8)      if (utf)
1691        {        {
1692        i = GET(ecode, 1);        i = GET(ecode, 1);
1693        while (i-- > 0)        while (i-- > 0)
# Line 1625  for (;;) Line 1718  for (;;)
1718      function is able to force a failure. */      function is able to force a failure. */
1719    
1720      case OP_CALLOUT:      case OP_CALLOUT:
1721      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1722        {        {
1723        pcre_callout_block cb;        PUBL(callout_block) cb;
1724        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1725        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1726        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1727    #if defined COMPILE_PCRE8
1728        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1729    #elif defined COMPILE_PCRE16
1730          cb.subject          = (PCRE_SPTR16)md->start_subject;
1731    #elif defined COMPILE_PCRE32
1732          cb.subject          = (PCRE_SPTR32)md->start_subject;
1733    #endif
1734        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1735        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1736        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
1737        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1738        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1739        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1740        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1741          /* Internal change requires this for API compatibility. */
1742          if (cb.capture_last == 0) cb.capture_last = -1;
1743        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1744        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1745        if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1746        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1747        }        }
1748      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1667  for (;;) Line 1768  for (;;)
1768      case OP_RECURSE:      case OP_RECURSE:
1769        {        {
1770        recursion_info *ri;        recursion_info *ri;
1771        int recno;        unsigned int recno;
1772    
1773        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1774        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1684  for (;;) Line 1785  for (;;)
1785        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1786    
1787        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1788          new_recursive.saved_capture_last = md->capture_last;
1789        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1790        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1791        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1700  for (;;) Line 1802  for (;;)
1802        else        else
1803          {          {
1804          new_recursive.offset_save =          new_recursive.offset_save =
1805            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1806          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1807          }          }
1808        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1809              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1810    
1811        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1812        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1813        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1814          */
1815    
1816        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1817        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
1818        do        do
1819          {          {
1820          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1821          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1822            md, eptrb, RM6);            md, eptrb, RM6);
1823          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1824              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1825            md->capture_last = new_recursive.saved_capture_last;
1826          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1827          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1828            {            {
1829            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
1830            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1831              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1832    
1833            /* 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
1834            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 1839  for (;;)
1839            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1840            }            }
1841    
1842          /* 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
1843          as NOMATCH. */          is treated as NOMATCH. */
1844    
1845          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN &&
1846                     rrc != MATCH_COMMIT)
1847            {            {
1848            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1849            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1850              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1851            RRETURN(rrc);            RRETURN(rrc);
1852            }            }
1853    
# Line 1754  for (;;) Line 1859  for (;;)
1859        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1860        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1861        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1862          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1863        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1864        }        }
1865    
# Line 1868  for (;;) Line 1973  for (;;)
1973    
1974        /* Deal with capturing */        /* Deal with capturing */
1975    
1976        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
1977        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1978          {          {
1979          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
1980          "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 2054  for (;;)
2054          }          }
2055        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2056          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2057          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2058          RRETURN(rrc);          RRETURN(rrc);
2059          }          }
# Line 1958  for (;;) Line 2062  for (;;)
2062        }        }
2063      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2064        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2065        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2066        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2067        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2015  for (;;) Line 2118  for (;;)
2118    
2119      case OP_DOLLM:      case OP_DOLLM:
2120      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2121        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2122          if (!IS_NEWLINE(eptr))
2123            {
2124            if (md->partial != 0 &&
2125                eptr + 1 >= md->end_subject &&
2126                NLBLOCK->nltype == NLTYPE_FIXED &&
2127                NLBLOCK->nllen == 2 &&
2128                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2129              {
2130              md->hitend = TRUE;
2131              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2132              }
2133            RRETURN(MATCH_NOMATCH);
2134            }
2135          }
2136      else      else
2137        {        {
2138        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2047  for (;;) Line 2164  for (;;)
2164      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2165      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2166          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2167          {
2168          if (md->partial != 0 &&
2169              eptr + 1 >= md->end_subject &&
2170              NLBLOCK->nltype == NLTYPE_FIXED &&
2171              NLBLOCK->nllen == 2 &&
2172              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2173            {
2174            md->hitend = TRUE;
2175            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2176            }
2177        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2178          }
2179    
2180      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2181    
# Line 2066  for (;;) Line 2194  for (;;)
2194        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2195        partial matching. */        partial matching. */
2196    
2197  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2198        if (utf8)        if (utf)
2199          {          {
2200          /* Get status of previous character */          /* Get status of previous character */
2201    
2202          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2203            {            {
2204            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2205            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2206            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2207            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2208  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2139  for (;;) Line 2267  for (;;)
2267              }              }
2268            else            else
2269  #endif  #endif
2270            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2271                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2272            }            }
2273    
2274          /* Get status of next character */          /* Get status of next character */
# Line 2162  for (;;) Line 2291  for (;;)
2291            }            }
2292          else          else
2293  #endif  #endif
2294          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2295              && ((md->ctypes[*eptr] & ctype_word) != 0);
2296          }          }
2297    
2298        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
# Line 2173  for (;;) Line 2303  for (;;)
2303        }        }
2304      break;      break;
2305    
2306      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2307        CRLF newlines and partial matching. */
2308    
2309      case OP_ANY:      case OP_ANY:
2310      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2311        if (md->partial != 0 &&
2312            eptr + 1 >= md->end_subject &&
2313            NLBLOCK->nltype == NLTYPE_FIXED &&
2314            NLBLOCK->nllen == 2 &&
2315            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2316          {
2317          md->hitend = TRUE;
2318          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2319          }
2320    
2321      /* Fall through */      /* Fall through */
2322    
2323        /* Match any single character whatsoever. */
2324    
2325      case OP_ALLANY:      case OP_ALLANY:
2326      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 */
2327        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2186  for (;;) Line 2329  for (;;)
2329        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2330        }        }
2331      eptr++;      eptr++;
2332      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  #ifdef SUPPORT_UTF
2333        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2334    #endif
2335      ecode++;      ecode++;
2336      break;      break;
2337    
# Line 2211  for (;;) Line 2356  for (;;)
2356        }        }
2357      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2358      if (      if (
2359  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2360         c < 256 &&         c < 256 &&
2361  #endif  #endif
2362         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
# Line 2228  for (;;) Line 2373  for (;;)
2373        }        }
2374      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2375      if (      if (
2376  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2377         c >= 256 ||         c > 255 ||
2378  #endif  #endif
2379         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2380         )         )
# Line 2245  for (;;) Line 2390  for (;;)
2390        }        }
2391      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2392      if (      if (
2393  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2394         c < 256 &&         c < 256 &&
2395  #endif  #endif
2396         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
# Line 2262  for (;;) Line 2407  for (;;)
2407        }        }
2408      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2409      if (      if (
2410  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2411         c >= 256 ||         c > 255 ||
2412  #endif  #endif
2413         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2414         )         )
# Line 2279  for (;;) Line 2424  for (;;)
2424        }        }
2425      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2426      if (      if (
2427  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2428         c < 256 &&         c < 256 &&
2429  #endif  #endif
2430         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
# Line 2296  for (;;) Line 2441  for (;;)
2441        }        }
2442      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2443      if (      if (
2444  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2445         c >= 256 ||         c > 255 ||
2446  #endif  #endif
2447         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2448         )         )
# Line 2316  for (;;) Line 2461  for (;;)
2461        {        {
2462        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2463    
2464        case 0x000d:        case CHAR_CR:
2465        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2466            {
2467            SCHECK_PARTIAL();
2468            }
2469          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2470        break;        break;
2471    
2472        case 0x000a:        case CHAR_LF:
2473        break;        break;
2474    
2475        case 0x000b:        case CHAR_VT:
2476        case 0x000c:        case CHAR_FF:
2477        case 0x0085:        case CHAR_NEL:
2478    #ifndef EBCDIC
2479        case 0x2028:        case 0x2028:
2480        case 0x2029:        case 0x2029:
2481    #endif  /* Not EBCDIC */
2482        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2483        break;        break;
2484        }        }
# Line 2343  for (;;) Line 2494  for (;;)
2494      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2495      switch(c)      switch(c)
2496        {        {
2497          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2498        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);  
2499        }        }
2500      ecode++;      ecode++;
2501      break;      break;
# Line 2377  for (;;) Line 2509  for (;;)
2509      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2510      switch(c)      switch(c)
2511        {        {
2512          HSPACE_CASES: break;  /* Byte and multibyte cases */
2513        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;  
2514        }        }
2515      ecode++;      ecode++;
2516      break;      break;
# Line 2411  for (;;) Line 2524  for (;;)
2524      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2525      switch(c)      switch(c)
2526        {        {
2527          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2528        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);  
2529        }        }
2530      ecode++;      ecode++;
2531      break;      break;
# Line 2433  for (;;) Line 2539  for (;;)
2539      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2540      switch(c)      switch(c)
2541        {        {
2542          VSPACE_CASES: break;
2543        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;  
2544        }        }
2545      ecode++;      ecode++;
2546      break;      break;
# Line 2459  for (;;) Line 2558  for (;;)
2558        }        }
2559      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2560        {        {
2561          const pcre_uint32 *cp;
2562        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2563    
2564        switch(ecode[1])        switch(ecode[1])
# Line 2475  for (;;) Line 2575  for (;;)
2575          break;          break;
2576    
2577          case PT_GC:          case PT_GC:
2578          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2579            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2580          break;          break;
2581    
# Line 2492  for (;;) Line 2592  for (;;)
2592          /* These are specials */          /* These are specials */
2593    
2594          case PT_ALNUM:          case PT_ALNUM:
2595          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2596               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2597            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2598          break;          break;
2599    
2600          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2601          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2602               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2603                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2604            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2605          break;          break;
2606    
2607          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2608          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2609               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2610               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2611                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
# Line 2513  for (;;) Line 2613  for (;;)
2613          break;          break;
2614    
2615          case PT_WORD:          case PT_WORD:
2616          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2617               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2618               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2619            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2620          break;          break;
2621    
2622            case PT_CLIST:
2623            cp = PRIV(ucd_caseless_sets) + ecode[2];
2624            for (;;)
2625              {
2626              if (c < *cp)
2627                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2628              if (c == *cp++)
2629                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2630              }
2631            break;
2632    
2633            case PT_UCNC:
2634            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
2635                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
2636                 c >= 0xe000) == (op == OP_NOTPROP))
2637              RRETURN(MATCH_NOMATCH);
2638            break;
2639    
2640          /* This should never occur */          /* This should never occur */
2641    
2642          default:          default:
# Line 2538  for (;;) Line 2656  for (;;)
2656        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2657        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2658        }        }
2659      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2660        {        {
2661        int len = 1;        int lgb, rgb;
2662        if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2663        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2664        eptr += len;        while (eptr < md->end_subject)
2665            {
2666            int len = 1;
2667            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2668            rgb = UCD_GRAPHBREAK(c);
2669            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2670            lgb = rgb;
2671            eptr += len;
2672            }
2673        }        }
2674        CHECK_PARTIAL();
2675      ecode++;      ecode++;
2676      break;      break;
2677  #endif  #endif  /* SUPPORT_UCP */
2678    
2679    
2680      /* 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 2689  for (;;)
2689      case OP_REFI:      case OP_REFI:
2690      caseless = op == OP_REFI;      caseless = op == OP_REFI;
2691      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2692      ecode += 3;      ecode += 1 + IMM2_SIZE;
2693    
2694      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2695    
# Line 2604  for (;;) Line 2729  for (;;)
2729        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2730        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2731        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2732        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2733        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2734        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2735        break;        break;
2736    
2737        default:               /* No repeat follows */        default:               /* No repeat follows */
2738        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2739          {          {
2740            if (length == -2) eptr = md->end_subject;   /* Partial match */
2741          CHECK_PARTIAL();          CHECK_PARTIAL();
2742          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2743          }          }
# Line 2620  for (;;) Line 2746  for (;;)
2746        }        }
2747    
2748      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2749      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2750        means the reference is unset in non-Java-compatible mode. If the minimum is
2751        zero, we can continue at the same level without recursion. For any other
2752        minimum, carrying on will result in NOMATCH. */
2753    
2754      if (length == 0) continue;      if (length == 0) continue;
2755        if (length < 0 && min == 0) continue;
2756    
2757      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2758      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 2763  for (;;)
2763        int slength;        int slength;
2764        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2765          {          {
2766            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2767          CHECK_PARTIAL();          CHECK_PARTIAL();
2768          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2769          }          }
# Line 2656  for (;;) Line 2787  for (;;)
2787          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2788          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2789            {            {
2790              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2791            CHECK_PARTIAL();            CHECK_PARTIAL();
2792            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2793            }            }
# Line 2674  for (;;) Line 2806  for (;;)
2806          int slength;          int slength;
2807          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2808            {            {
2809            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2810              the soft partial matching case. */
2811    
2812              if (slength == -2 && md->partial != 0 &&
2813                  md->end_subject > md->start_used_ptr)
2814                {
2815                md->hitend = TRUE;
2816                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2817                }
2818            break;            break;
2819            }            }
2820          eptr += slength;          eptr += slength;
2821          }          }
2822    
2823        while (eptr >= pp)        while (eptr >= pp)
2824          {          {
2825          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 2703  for (;;) Line 2844  for (;;)
2844      case OP_NCLASS:      case OP_NCLASS:
2845      case OP_CLASS:      case OP_CLASS:
2846        {        {
2847          /* The data variable is saved across frames, so the byte map needs to
2848          be stored there. */
2849    #define BYTE_MAP ((pcre_uint8 *)data)
2850        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2851        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2852    
2853        switch (*ecode)        switch (*ecode)
2854          {          {
# Line 2725  for (;;) Line 2869  for (;;)
2869          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2870          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2871          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2872          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2873          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2874          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2875          break;          break;
2876    
2877          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2737  for (;;) Line 2881  for (;;)
2881    
2882        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2883    
2884  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2885        /* UTF-8 mode */        if (utf)
       if (utf8)  
2886          {          {
2887          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2888            {            {
# Line 2754  for (;;) Line 2897  for (;;)
2897              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2898              }              }
2899            else            else
2900              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
             }  
2901            }            }
2902          }          }
2903        else        else
2904  #endif  #endif
2905        /* Not UTF-8 mode */        /* Not UTF mode */
2906          {          {
2907          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2908            {            {
# Line 2771  for (;;) Line 2912  for (;;)
2912              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2913              }              }
2914            c = *eptr++;            c = *eptr++;
2915            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2916              if (c > 255)
2917                {
2918                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2919                }
2920              else
2921    #endif
2922                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2923            }            }
2924          }          }
2925    
# Line 2785  for (;;) Line 2933  for (;;)
2933    
2934        if (minimize)        if (minimize)
2935          {          {
2936  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2937          /* UTF-8 mode */          if (utf)
         if (utf8)  
2938            {            {
2939            for (fi = min;; fi++)            for (fi = min;; fi++)
2940              {              {
# Line 2805  for (;;) Line 2952  for (;;)
2952                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2953                }                }
2954              else              else
2955                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
               }  
2956              }              }
2957            }            }
2958          else          else
2959  #endif  #endif
2960          /* Not UTF-8 mode */          /* Not UTF mode */
2961            {            {
2962            for (fi = min;; fi++)            for (fi = min;; fi++)
2963              {              {
# Line 2825  for (;;) Line 2970  for (;;)
2970                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2971                }                }
2972              c = *eptr++;              c = *eptr++;
2973              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2974                if (c > 255)
2975                  {
2976                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2977                  }
2978                else
2979    #endif
2980                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2981              }              }
2982            }            }
2983          /* Control never gets here */          /* Control never gets here */
# Line 2837  for (;;) Line 2989  for (;;)
2989          {          {
2990          pp = eptr;          pp = eptr;
2991    
2992  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2993          /* UTF-8 mode */          if (utf)
         if (utf8)  
2994            {            {
2995            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2996              {              {
# Line 2855  for (;;) Line 3006  for (;;)
3006                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
3007                }                }
3008              else              else
3009                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
3010              eptr += len;              eptr += len;
3011              }              }
3012            for (;;)            for (;;)
# Line 2870  for (;;) Line 3019  for (;;)
3019            }            }
3020          else          else
3021  #endif  #endif
3022            /* Not UTF-8 mode */            /* Not UTF mode */
3023            {            {
3024            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3025              {              {
# Line 2880  for (;;) Line 3029  for (;;)
3029                break;                break;
3030                }                }
3031              c = *eptr;              c = *eptr;
3032              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
3033                if (c > 255)
3034                  {
3035                  if (op == OP_CLASS) break;
3036                  }
3037                else
3038    #endif
3039                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3040              eptr++;              eptr++;
3041              }              }
3042            while (eptr >= pp)            while (eptr >= pp)
# Line 2893  for (;;) Line 3049  for (;;)
3049    
3050          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3051          }          }
3052    #undef BYTE_MAP
3053        }        }
3054      /* Control never gets here */      /* Control never gets here */
3055    
# Line 2901  for (;;) Line 3058  for (;;)
3058      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
3059      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
3060    
3061  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3062      case OP_XCLASS:      case OP_XCLASS:
3063        {        {
3064        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2926  for (;;) Line 3083  for (;;)
3083          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3084          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3085          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3086          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
3087          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
3088          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
3089          break;          break;
3090    
3091          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2946  for (;;) Line 3103  for (;;)
3103            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3104            }            }
3105          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
3106          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3107          }          }
3108    
3109        /* 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 3127  for (;;)
3127              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3128              }              }
3129            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3130            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3131            }            }
3132          /* Control never gets here */          /* Control never gets here */
3133          }          }
# Line 2988  for (;;) Line 3145  for (;;)
3145              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3146              break;              break;
3147              }              }
3148    #ifdef SUPPORT_UTF
3149            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3150            if (!_pcre_xclass(c, data)) break;  #else
3151              c = *eptr;
3152    #endif
3153              if (!PRIV(xclass)(c, data, utf)) break;
3154            eptr += len;            eptr += len;
3155            }            }
3156          for(;;)          for(;;)
# Line 2997  for (;;) Line 3158  for (;;)
3158            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3159            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3160            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3161            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3162              if (utf) BACKCHAR(eptr);
3163    #endif
3164            }            }
3165          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3166          }          }
# Line 3009  for (;;) Line 3172  for (;;)
3172      /* Match a single character, casefully */      /* Match a single character, casefully */
3173    
3174      case OP_CHAR:      case OP_CHAR:
3175  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3176      if (utf8)      if (utf)
3177        {        {
3178        length = 1;        length = 1;
3179        ecode++;        ecode++;
# Line 3020  for (;;) Line 3183  for (;;)
3183          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3184          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3185          }          }
3186        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3187        }        }
3188      else      else
3189  #endif  #endif
3190        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3191        {        {
3192        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3193          {          {
# Line 3047  for (;;) Line 3209  for (;;)
3209        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3210        }        }
3211    
3212  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3213      if (utf8)      if (utf)
3214        {        {
3215        length = 1;        length = 1;
3216        ecode++;        ecode++;
# Line 3061  for (;;) Line 3223  for (;;)
3223    
3224        if (fc < 128)        if (fc < 128)
3225          {          {
3226          if (md->lcc[*ecode++] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          pcre_uint32 cc = RAWUCHAR(eptr);
3227            if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3228            ecode++;
3229            eptr++;
3230          }          }
3231    
3232        /* Otherwise we must pick up the subject character. Note that we cannot        /* Otherwise we must pick up the subject character. Note that we cannot
# Line 3070  for (;;) Line 3235  for (;;)
3235    
3236        else        else
3237          {          {
3238          unsigned int dc;          pcre_uint32 dc;
3239          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3240          ecode += length;          ecode += length;
3241    
# Line 3087  for (;;) Line 3252  for (;;)
3252          }          }
3253        }        }
3254      else      else
3255  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3256    
3257      /* Non-UTF-8 mode */      /* Not UTF mode */
3258        {        {
3259        if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3260              != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3261          eptr++;
3262        ecode += 2;        ecode += 2;
3263        }        }
3264      break;      break;
# Line 3101  for (;;) Line 3268  for (;;)
3268      case OP_EXACT:      case OP_EXACT:
3269      case OP_EXACTI:      case OP_EXACTI:
3270      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3271      ecode += 3;      ecode += 1 + IMM2_SIZE;
3272      goto REPEATCHAR;      goto REPEATCHAR;
3273    
3274      case OP_POSUPTO:      case OP_POSUPTO:
# Line 3116  for (;;) Line 3283  for (;;)
3283      min = 0;      min = 0;
3284      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3285      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3286      ecode += 3;      ecode += 1 + IMM2_SIZE;
3287      goto REPEATCHAR;      goto REPEATCHAR;
3288    
3289      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3164  for (;;) Line 3331  for (;;)
3331      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3332    
3333      REPEATCHAR:      REPEATCHAR:
3334  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3335      if (utf8)      if (utf)
3336        {        {
3337        length = 1;        length = 1;
3338        charptr = ecode;        charptr = ecode;
# Line 3178  for (;;) Line 3345  for (;;)
3345        if (length > 1)        if (length > 1)
3346          {          {
3347  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3348          unsigned int othercase;          pcre_uint32 othercase;
3349          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3350              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3351            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3352          else oclength = 0;          else oclength = 0;
3353  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3354    
3355          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3356            {            {
3357            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3358              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3359  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3360            else if (oclength > 0 &&            else if (oclength > 0 &&
3361                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3362                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3363  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3364            else            else
3365              {              {
# Line 3211  for (;;) Line 3378  for (;;)
3378              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3379              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3380              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3381                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3382  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3383              else if (oclength > 0 &&              else if (oclength > 0 &&
3384                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3385                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3386  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3387              else              else
3388                {                {
# Line 3232  for (;;) Line 3399  for (;;)
3399            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3400              {              {
3401              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3402                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3403  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3404              else if (oclength > 0 &&              else if (oclength > 0 &&
3405                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3406                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3407  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3408              else              else
3409                {                {
# Line 3268  for (;;) Line 3435  for (;;)
3435        value of fc will always be < 128. */        value of fc will always be < 128. */
3436        }        }
3437      else      else
3438  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3439          /* When not in UTF-8 mode, load a single-byte character. */
3440      /* When not in UTF-8 mode, load a single-byte character. */        fc = *ecode++;
3441    
3442      fc = *ecode++;      /* The value of fc at this point is always one character, though we may
3443        or may not be in UTF mode. The code is duplicated for the caseless and
     /* The value of fc at this point is always less than 256, though we may or  
     may not be in UTF-8 mode. The code is duplicated for the caseless and  
3444      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3445      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3446      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3284  for (;;) Line 3449  for (;;)
3449      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3450    
3451      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3452        max, eptr));        max, (char *)eptr));
3453    
3454      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3455        {        {
3456        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3457          /* fc must be < 128 if UTF is enabled. */
3458          foc = md->fcc[fc];
3459    #else
3460    #ifdef SUPPORT_UTF
3461    #ifdef SUPPORT_UCP
3462          if (utf && fc > 127)
3463            foc = UCD_OTHERCASE(fc);
3464    #else
3465          if (utf && fc > 127)
3466            foc = fc;
3467    #endif /* SUPPORT_UCP */
3468          else
3469    #endif /* SUPPORT_UTF */
3470            foc = TABLE_GET(fc, md->fcc, fc);
3471    #endif /* COMPILE_PCRE8 */
3472    
3473        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3474          {          {
3475            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3476          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3477            {            {
3478            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3479            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3480            }            }
3481          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3482            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3483            eptr++;
3484          }          }
3485        if (min == max) continue;        if (min == max) continue;
3486        if (minimize)        if (minimize)
3487          {          {
3488          for (fi = min;; fi++)          for (fi = min;; fi++)
3489            {            {
3490              pcre_uint32 cc;               /* Faster than pcre_uchar */
3491            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3492            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3493            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3311  for (;;) Line 3496  for (;;)
3496              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3497              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3498              }              }
3499            if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3500              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3501              eptr++;
3502            }            }
3503          /* Control never gets here */          /* Control never gets here */
3504          }          }
# Line 3320  for (;;) Line 3507  for (;;)
3507          pp = eptr;          pp = eptr;
3508          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3509            {            {
3510              pcre_uint32 cc;               /* Faster than pcre_uchar */
3511            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3512              {              {
3513              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3514              break;              break;
3515              }              }
3516            if (fc != md->lcc[*eptr]) break;            cc = RAWUCHARTEST(eptr);
3517              if (fc != cc && foc != cc) break;
3518            eptr++;            eptr++;
3519            }            }
3520    
# Line 3353  for (;;) Line 3542  for (;;)
3542            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3543            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3544            }            }
3545          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3546          }          }
3547    
3548        if (min == max) continue;        if (min == max) continue;
# Line 3370  for (;;) Line 3559  for (;;)
3559              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3560              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3561              }              }
3562            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3563            }            }
3564          /* Control never gets here */          /* Control never gets here */
3565          }          }
# Line 3384  for (;;) Line 3573  for (;;)
3573              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3574              break;              break;
3575              }              }
3576            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3577            eptr++;            eptr++;
3578            }            }
3579          if (possessive) continue;          if (possessive) continue;
# Line 3410  for (;;) Line 3599  for (;;)
3599        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3600        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3601        }        }
3602      ecode++;  #ifdef SUPPORT_UTF
3603      GETCHARINCTEST(c, eptr);      if (utf)
     if (op == OP_NOTI)         /* The caseless case */  
3604        {        {
3605  #ifdef SUPPORT_UTF8        register pcre_uint32 ch, och;
3606        if (c < 256)  
3607  #endif        ecode++;
3608        c = md->lcc[c];        GETCHARINC(ch, ecode);
3609        if (md->lcc[*ecode++] == c) RRETURN(MATCH_NOMATCH);        GETCHARINC(c, eptr);
3610    
3611          if (op == OP_NOT)
3612            {
3613            if (ch == c) RRETURN(MATCH_NOMATCH);
3614            }
3615          else
3616            {
3617    #ifdef SUPPORT_UCP
3618            if (ch > 127)
3619              och = UCD_OTHERCASE(ch);
3620    #else
3621            if (ch > 127)
3622              och = ch;
3623    #endif /* SUPPORT_UCP */
3624            else
3625              och = TABLE_GET(ch, md->fcc, ch);
3626            if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3627            }
3628        }        }
3629      else    /* Caseful */      else
3630    #endif
3631        {        {
3632        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3633          c = *eptr++;
3634          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3635            RRETURN(MATCH_NOMATCH);
3636          ecode += 2;
3637        }        }
3638      break;      break;
3639    
# Line 3436  for (;;) Line 3647  for (;;)
3647      case OP_NOTEXACT:      case OP_NOTEXACT:
3648      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3649      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3650      ecode += 3;      ecode += 1 + IMM2_SIZE;
3651      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3652    
3653      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3446  for (;;) Line 3657  for (;;)
3657      min = 0;      min = 0;
3658      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3659      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3660      ecode += 3;      ecode += 1 + IMM2_SIZE;
3661      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3662    
3663      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3478  for (;;) Line 3689  for (;;)
3689      possessive = TRUE;      possessive = TRUE;
3690      min = 0;      min = 0;
3691      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3692      ecode += 3;      ecode += 1 + IMM2_SIZE;
3693      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3694    
3695      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3502  for (;;) Line 3713  for (;;)
3713      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3714    
3715      REPEATNOTCHAR:      REPEATNOTCHAR:
3716      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3717    
3718      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3719      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 3724  for (;;)
3724      characters and work backwards. */      characters and work backwards. */
3725    
3726      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,
3727        max, eptr));        max, (char *)eptr));
3728    
3729      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3730        {        {
3731        fc = md->lcc[fc];  #ifdef SUPPORT_UTF
3732    #ifdef SUPPORT_UCP
3733          if (utf && fc > 127)
3734            foc = UCD_OTHERCASE(fc);
3735    #else
3736          if (utf && fc > 127)
3737            foc = fc;
3738    #endif /* SUPPORT_UCP */
3739          else
3740    #endif /* SUPPORT_UTF */
3741            foc = TABLE_GET(fc, md->fcc, fc);
3742    
3743  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3744        /* UTF-8 mode */        if (utf)
       if (utf8)  
3745          {          {
3746          register unsigned int d;          register pcre_uint32 d;
3747          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3748            {            {
3749            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3532  for (;;) Line 3752  for (;;)
3752              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3753              }              }
3754            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3755            if (d < 256) d = md->lcc[d];            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) RRETURN(MATCH_NOMATCH);  
3756            }            }
3757          }          }
3758        else        else
3759  #endif  #endif
3760          /* Not UTF mode */
       /* Not UTF-8 mode */  
3761          {          {
3762          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3763            {            {
# Line 3548  for (;;) Line 3766  for (;;)
3766              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3767              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3768              }              }
3769            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3770              eptr++;
3771            }            }
3772          }          }
3773    
# Line 3556  for (;;) Line 3775  for (;;)
3775    
3776        if (minimize)        if (minimize)
3777          {          {
3778  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3779          /* UTF-8 mode */          if (utf)
         if (utf8)  
3780            {            {
3781            register unsigned int d;            register pcre_uint32 d;
3782            for (fi = min;; fi++)            for (fi = min;; fi++)
3783              {              {
3784              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3572  for (;;) Line 3790  for (;;)
3790                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3791                }                }
3792              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3793              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) RRETURN(MATCH_NOMATCH);  
3794              }              }
3795            }            }
3796          else          else
3797  #endif  #endif
3798          /* Not UTF-8 mode */          /* Not UTF mode */
3799            {            {
3800            for (fi = min;; fi++)            for (fi = min;; fi++)
3801              {              {
# Line 3590  for (;;) Line 3807  for (;;)
3807                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3808                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3809                }                }
3810              if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3811                eptr++;
3812              }              }
3813            }            }
3814          /* Control never gets here */          /* Control never gets here */
# Line 3602  for (;;) Line 3820  for (;;)
3820          {          {
3821          pp = eptr;          pp = eptr;
3822    
3823  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3824          /* UTF-8 mode */          if (utf)
         if (utf8)  
3825            {            {
3826            register unsigned int d;            register pcre_uint32 d;
3827            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3828              {              {
3829              int len = 1;              int len = 1;
# Line 3616  for (;;) Line 3833  for (;;)
3833                break;                break;
3834                }                }
3835              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3836              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) break;
             if (fc == d) break;  
3837              eptr += len;              eptr += len;
3838              }              }
3839          if (possessive) continue;            if (possessive) continue;
3840          for(;;)            for(;;)
3841              {              {
3842              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3843              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 3631  for (;;) Line 3847  for (;;)
3847            }            }
3848          else          else
3849  #endif  #endif
3850          /* Not UTF-8 mode */          /* Not UTF mode */
3851            {            {
3852            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3853              {              {
# Line 3640  for (;;) Line 3856  for (;;)
3856                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3857                break;                break;
3858                }                }
3859              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3860              eptr++;              eptr++;
3861              }              }
3862            if (possessive) continue;            if (possessive) continue;
# Line 3661  for (;;) Line 3877  for (;;)
3877    
3878      else      else
3879        {        {
3880  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3881        /* UTF-8 mode */        if (utf)
       if (utf8)  
3882          {          {
3883          register unsigned int d;          register pcre_uint32 d;
3884          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3885            {            {
3886            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3679  for (;;) Line 3894  for (;;)
3894          }          }
3895        else        else
3896  #endif  #endif
3897        /* Not UTF-8 mode */        /* Not UTF mode */
3898          {          {
3899          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3900            {            {
# Line 3696  for (;;) Line 3911  for (;;)
3911    
3912        if (minimize)        if (minimize)
3913          {          {
3914  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3915          /* UTF-8 mode */          if (utf)
         if (utf8)  
3916            {            {
3917            register unsigned int d;            register pcre_uint32 d;
3918            for (fi = min;; fi++)            for (fi = min;; fi++)
3919              {              {
3920              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3717  for (;;) Line 3931  for (;;)
3931            }            }
3932          else          else
3933  #endif  #endif
3934          /* Not UTF-8 mode */          /* Not UTF mode */
3935            {            {
3936            for (fi = min;; fi++)            for (fi = min;; fi++)
3937              {              {
# Line 3741  for (;;) Line 3955  for (;;)
3955          {          {
3956          pp = eptr;          pp = eptr;
3957    
3958  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3959          /* UTF-8 mode */          if (utf)
         if (utf8)  
3960            {            {
3961            register unsigned int d;            register pcre_uint32 d;
3962            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3963              {              {
3964              int len = 1;              int len = 1;
# Line 3769  for (;;) Line 3982  for (;;)
3982            }            }
3983          else          else
3984  #endif  #endif
3985          /* Not UTF-8 mode */          /* Not UTF mode */
3986            {            {
3987            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3988              {              {
# Line 3802  for (;;) Line 4015  for (;;)
4015      case OP_TYPEEXACT:      case OP_TYPEEXACT:
4016      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
4017      minimize = TRUE;      minimize = TRUE;
4018      ecode += 3;      ecode += 1 + IMM2_SIZE;
4019      goto REPEATTYPE;      goto REPEATTYPE;
4020    
4021      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3810  for (;;) Line 4023  for (;;)
4023      min = 0;      min = 0;
4024      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4025      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
4026      ecode += 3;      ecode += 1 + IMM2_SIZE;
4027      goto REPEATTYPE;      goto REPEATTYPE;
4028    
4029      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3838  for (;;) Line 4051  for (;;)
4051      possessive = TRUE;      possessive = TRUE;
4052      min = 0;      min = 0;
4053      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4054      ecode += 3;      ecode += 1 + IMM2_SIZE;
4055      goto REPEATTYPE;      goto REPEATTYPE;
4056    
4057      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 4021  for (;;) Line 4234  for (;;)
4234              }              }
4235            break;            break;
4236    
4237              case PT_CLIST:
4238              for (i = 1; i <= min; i++)
4239                {
4240                const pcre_uint32 *cp;
4241                if (eptr >= md->end_subject)
4242                  {
4243                  SCHECK_PARTIAL();
4244                  RRETURN(MATCH_NOMATCH);
4245                  }
4246                GETCHARINCTEST(c, eptr);
4247                cp = PRIV(ucd_caseless_sets) + prop_value;
4248                for (;;)
4249                  {
4250                  if (c < *cp)
4251                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4252                  if (c == *cp++)
4253                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4254                  }
4255                }
4256              break;
4257    
4258              case PT_UCNC:
4259              for (i = 1; i <= min; i++)
4260                {
4261                if (eptr >= md->end_subject)
4262                  {
4263                  SCHECK_PARTIAL();
4264                  RRETURN(MATCH_NOMATCH);
4265                  }
4266                GETCHARINCTEST(c, eptr);
4267                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
4268                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
4269                     c >= 0xe000) == prop_fail_result)
4270                  RRETURN(MATCH_NOMATCH);
4271                }
4272              break;
4273    
4274            /* This should not occur */            /* This should not occur */
4275    
4276            default:            default:
# Line 4040  for (;;) Line 4290  for (;;)
4290              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4291              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4292              }              }
4293            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4294              {              {
4295              int len = 1;              int lgb, rgb;
4296              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4297              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4298              eptr += len;             while (eptr < md->end_subject)
4299                  {
4300                  int len = 1;
4301                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4302                  rgb = UCD_GRAPHBREAK(c);
4303                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4304                  lgb = rgb;
4305                  eptr += len;
4306                  }
4307              }              }
4308              CHECK_PARTIAL();
4309            }            }
4310          }          }
4311    
# Line 4057  for (;;) Line 4314  for (;;)
4314    
4315  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4316    
4317  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4318        if (utf8) switch(ctype)        if (utf) switch(ctype)
4319          {          {
4320          case OP_ANY:          case OP_ANY:
4321          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 4069  for (;;) Line 4326  for (;;)
4326              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4327              }              }
4328            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4329              if (md->partial != 0 &&
4330                  eptr + 1 >= md->end_subject &&
4331                  NLBLOCK->nltype == NLTYPE_FIXED &&
4332                  NLBLOCK->nllen == 2 &&
4333                  RAWUCHAR(eptr) == NLBLOCK->nl[0])
4334                {
4335                md->hitend = TRUE;
4336                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4337                }
4338            eptr++;            eptr++;
4339            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4340            }            }
4341          break;          break;
4342    
# Line 4083  for (;;) Line 4349  for (;;)
4349              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4350              }              }
4351            eptr++;            eptr++;
4352            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4353            }            }
4354          break;          break;
4355    
# Line 4105  for (;;) Line 4371  for (;;)
4371              {              {
4372              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4373    
4374              case 0x000d:              case CHAR_CR:
4375              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
4376              break;              break;
4377    
4378              case 0x000a:              case CHAR_LF:
4379              break;              break;
4380    
4381              case 0x000b:              case CHAR_VT:
4382              case 0x000c:              case CHAR_FF:
4383              case 0x0085:              case CHAR_NEL:
4384    #ifndef EBCDIC
4385              case 0x2028:              case 0x2028:
4386              case 0x2029:              case 0x2029:
4387    #endif  /* Not EBCDIC */
4388              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4389              break;              break;
4390              }              }
# Line 4134  for (;;) Line 4402  for (;;)
4402            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4403            switch(c)            switch(c)
4404              {              {
4405                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4406              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);  
4407              }              }
4408            }            }
4409          break;          break;
# Line 4170  for (;;) Line 4419  for (;;)
4419            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4420            switch(c)            switch(c)
4421              {              {
4422                HSPACE_CASES: break;  /* Byte and multibyte cases */
4423              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;  
4424              }              }
4425            }            }
4426          break;          break;
# Line 4206  for (;;) Line 4436  for (;;)
4436            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4437            switch(c)            switch(c)
4438              {              {
4439                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4440              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);  
4441              }              }
4442            }            }
4443          break;          break;
# Line 4230  for (;;) Line 4453  for (;;)
4453            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4454            switch(c)            switch(c)
4455              {              {
4456                VSPACE_CASES: break;
4457              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;  
4458              }              }
4459            }            }
4460          break;          break;
# Line 4260  for (;;) Line 4476  for (;;)
4476          case OP_DIGIT:          case OP_DIGIT:
4477          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4478            {            {
4479              pcre_uint32 cc;
4480            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4481              {              {
4482              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4483              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4484              }              }
4485            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            cc = RAWUCHAR(eptr);
4486              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4487              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4488              eptr++;
4489            /* 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 */
4490            }            }
4491          break;          break;
# Line 4274  for (;;) Line 4493  for (;;)
4493          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4494          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4495            {            {
4496              pcre_uint32 cc;
4497            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4498              {              {
4499              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4500              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4501              }              }
4502            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = RAWUCHAR(eptr);
4503              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4504              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4505            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4506              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4507            }            }
4508          break;          break;
4509    
4510          case OP_WHITESPACE:          case OP_WHITESPACE:
4511          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4512            {            {
4513              pcre_uint32 cc;
4514            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4515              {              {
4516              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4517              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4518              }              }
4519            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            cc = RAWUCHAR(eptr);
4520              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4521              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4522              eptr++;
4523            /* 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 */
4524            }            }
4525          break;          break;
# Line 4302  for (;;) Line 4527  for (;;)
4527          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4528          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4529            {            {
4530              pcre_uint32 cc;
4531            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4532              {              {
4533              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4534              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4535              }              }
4536            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = RAWUCHAR(eptr);
4537              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4538              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4539            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4540              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4541            }            }
4542          break;          break;
4543    
4544          case OP_WORDCHAR:          case OP_WORDCHAR:
4545          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4546            {            {
4547              pcre_uint32 cc;
4548            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4549              {              {
4550              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4551              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4552              }              }
4553            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)            cc = RAWUCHAR(eptr);
4554              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4555              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4556              eptr++;
4557            /* 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 */
4558            }            }
4559          break;          break;
# Line 4332  for (;;) Line 4563  for (;;)
4563          }  /* End switch(ctype) */          }  /* End switch(ctype) */
4564    
4565        else        else
4566  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF */
4567    
4568        /* 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
4569        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. */
# Line 4348  for (;;) Line 4579  for (;;)
4579              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4580              }              }
4581            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4582              if (md->partial != 0 &&
4583                  eptr + 1 >= md->end_subject &&
4584                  NLBLOCK->nltype == NLTYPE_FIXED &&
4585                  NLBLOCK->nllen == 2 &&
4586                  *eptr == NLBLOCK->nl[0])
4587                {
4588                md->hitend = TRUE;
4589                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4590                }
4591            eptr++;            eptr++;
4592            }            }
4593          break;          break;
# Line 4382  for (;;) Line 4622  for (;;)
4622              {              {
4623              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4624    
4625              case 0x000d:              case CHAR_CR:
4626              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4627              break;              break;
4628    
4629              case 0x000a:              case CHAR_LF:
4630              break;              break;
4631    
4632              case 0x000b:              case CHAR_VT:
4633              case 0x000c:              case CHAR_FF:
4634              case 0x0085:              case CHAR_NEL:
4635    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4636                case 0x2028:
4637                case 0x2029:
4638    #endif
4639              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4640              break;              break;
4641              }              }
# Line 4409  for (;;) Line 4653  for (;;)
4653            switch(*eptr++)            switch(*eptr++)
4654              {              {
4655              default: break;              default: break;
4656              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4657              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4658              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4659    #endif
4660              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4661              }              }
4662            }            }
# Line 4428  for (;;) Line 4673  for (;;)
4673            switch(*eptr++)            switch(*eptr++)
4674              {              {
4675              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4676              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4677              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4678              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4679    #endif
4680              break;              break;
4681              }              }
4682            }            }
# Line 4446  for (;;) Line 4692  for (;;)
4692              }              }
4693            switch(*eptr++)            switch(*eptr++)
4694              {              {
4695              default: break;              VSPACE_BYTE_CASES:
4696              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4697              case 0x0b:      /* VT */              VSPACE_MULTIBYTE_CASES:
4698              case 0x0c:      /* FF */  #endif
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
4699              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4700                default: break;
4701              }              }
4702            }            }
4703          break;          break;
# Line 4468  for (;;) Line 4713  for (;;)
4713            switch(*eptr++)            switch(*eptr++)
4714              {              {
4715              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4716              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4717              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4718              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
4719              case 0x0d:      /* CR */  #endif
             case 0x85:      /* NEL */  
4720              break;              break;
4721              }              }
4722            }            }
# Line 4486  for (;;) Line 4730  for (;;)
4730              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4731              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4732              }              }
4733            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0)
4734                RRETURN(MATCH_NOMATCH);
4735              eptr++;
4736            }            }
4737          break;          break;
4738    
# Line 4498  for (;;) Line 4744  for (;;)
4744              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4745              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4746              }              }
4747            if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0)
4748                RRETURN(MATCH_NOMATCH);
4749              eptr++;
4750            }            }
4751          break;          break;
4752    
# Line 4510  for (;;) Line 4758  for (;;)
4758              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4759              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4760              }              }
4761            if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0)
4762                RRETURN(MATCH_NOMATCH);
4763              eptr++;
4764            }            }
4765          break;          break;
4766    
# Line 4522  for (;;) Line 4772  for (;;)
4772              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4773              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4774              }              }
4775            if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0)
4776                RRETURN(MATCH_NOMATCH);
4777              eptr++;
4778            }            }
4779          break;          break;
4780    
# Line 4534  for (;;) Line 4786  for (;;)
4786              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4787              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4788              }              }
4789            if ((md->ctypes[*eptr++] & ctype_word) != 0)            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0)
4790              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4791              eptr++;
4792            }            }
4793          break;          break;
4794    
# Line 4547  for (;;) Line 4800  for (;;)
4800              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4801              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4802              }              }
4803            if ((md->ctypes[*eptr++] & ctype_word) == 0)            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0)
4804              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4805              eptr++;
4806            }            }
4807          break;          break;
4808    
# Line 4739  for (;;) Line 4993  for (;;)
4993              }              }
4994            /* Control never gets here */            /* Control never gets here */
4995    
4996            /* This should never occur */            case PT_CLIST:
4997              for (fi = min;; fi++)
4998                {
4999                const pcre_uint32 *cp;
5000                RMATCH(eptr, ecode, offset_top, md, eptrb, RM67);
5001                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5002                if (fi >= max) RRETURN(MATCH_NOMATCH);
5003                if (eptr >= md->end_subject)
5004                  {
5005                  SCHECK_PARTIAL();
5006                  RRETURN(MATCH_NOMATCH);
5007                  }
5008                GETCHARINCTEST(c, eptr);
5009                cp = PRIV(ucd_caseless_sets) + prop_value;
5010                for (;;)
5011                  {
5012                  if (c < *cp)
5013                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
5014                  if (c == *cp++)
5015                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
5016                  }
5017                }
5018              /* Control never gets here */
5019    
5020              case PT_UCNC:
5021              for (fi = min;; fi++)
5022                {
5023                RMATCH(eptr, ecode, offset_top, md, eptrb, RM68);
5024                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5025                if (fi >= max) RRETURN(MATCH_NOMATCH);
5026                if (eptr >= md->end_subject)
5027                  {
5028                  SCHECK_PARTIAL();
5029                  RRETURN(MATCH_NOMATCH);
5030                  }
5031                GETCHARINCTEST(c, eptr);
5032                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5033                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5034                     c >= 0xe000) == prop_fail_result)
5035                  RRETURN(MATCH_NOMATCH);
5036                }
5037              /* Control never gets here */
5038    
5039              /* This should never occur */
5040            default:            default:
5041            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5042            }            }
# Line 4761  for (;;) Line 5057  for (;;)
5057              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5058              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5059              }              }
5060            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
5061              {              {
5062              int len = 1;              int lgb, rgb;
5063              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5064              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5065              eptr += len;              while (eptr < md->end_subject)
5066                  {
5067                  int len = 1;
5068                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5069                  rgb = UCD_GRAPHBREAK(c);
5070                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5071                  lgb = rgb;
5072                  eptr += len;
5073                  }
5074              }              }
5075              CHECK_PARTIAL();
5076            }            }
5077          }          }
5078        else        else
5079  #endif     /* SUPPORT_UCP */  #endif     /* SUPPORT_UCP */
5080    
5081  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5082        /* UTF-8 mode */        if (utf)
       if (utf8)  
5083          {          {
5084          for (fi = min;; fi++)          for (fi = min;; fi++)
5085            {            {
# Line 4794  for (;;) Line 5096  for (;;)
5096            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5097            switch(ctype)            switch(ctype)
5098              {              {
5099              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5100                if (md->partial != 0 &&    /* Take care with CRLF partial */
5101                    eptr >= md->end_subject &&
5102                    NLBLOCK->nltype == NLTYPE_FIXED &&
5103                    NLBLOCK->nllen == 2 &&
5104                    c == NLBLOCK->nl[0])
5105                  {
5106                  md->hitend = TRUE;
5107                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5108                  }
5109                break;
5110    
5111              case OP_ALLANY:              case OP_ALLANY:
5112              case OP_ANYBYTE:              case OP_ANYBYTE:
5113              break;              break;
# Line 4803  for (;;) Line 5116  for (;;)
5116              switch(c)              switch(c)
5117                {                {
5118                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5119                case 0x000d:                case CHAR_CR:
5120                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
5121                break;                break;
5122                case 0x000a:  
5123                  case CHAR_LF:
5124                break;                break;
5125    
5126                case 0x000b:                case CHAR_VT:
5127                case 0x000c:                case CHAR_FF:
5128                case 0x0085:                case CHAR_NEL:
5129    #ifndef EBCDIC
5130                case 0x2028:                case 0x2028:
5131                case 0x2029:                case 0x2029:
5132    #endif  /* Not EBCDIC */
5133                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5134                break;                break;
5135                }                }
# Line 4822  for (;;) Line 5138  for (;;)
5138              case OP_NOT_HSPACE:              case OP_NOT_HSPACE:
5139              switch(c)              switch(c)
5140                {                {
5141                  HSPACE_CASES: RRETURN(MATCH_NOMATCH);
5142                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);  
5143                }                }
5144              break;              break;
5145    
5146              case OP_HSPACE:              case OP_HSPACE:
5147              switch(c)              switch(c)
5148                {                {
5149                  HSPACE_CASES: break;
5150                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;  
5151                }                }
5152              break;              break;
5153    
5154              case OP_NOT_VSPACE:              case OP_NOT_VSPACE:
5155              switch(c)              switch(c)
5156                {                {
5157                  VSPACE_CASES: RRETURN(MATCH_NOMATCH);
5158                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);  
5159                }                }
5160              break;              break;
5161    
5162              case OP_VSPACE:              case OP_VSPACE:
5163              switch(c)              switch(c)
5164                {                {
5165                  VSPACE_CASES: break;
5166                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;  
5167                }                }
5168              break;              break;
5169    
# Line 4919  for (;;) Line 5183  for (;;)
5183              break;              break;
5184    
5185              case OP_WHITESPACE:              case OP_WHITESPACE:
5186              if  (c >= 256 || (md->ctypes[c] & ctype_space) == 0)              if (c >= 256 || (md->ctypes[c] & ctype_space) == 0)
5187                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5188              break;              break;
5189    
# Line 4940  for (;;) Line 5204  for (;;)
5204          }          }
5205        else        else
5206  #endif  #endif
5207        /* Not UTF-8 mode */        /* Not UTF mode */
5208          {          {
5209          for (fi = min;; fi++)          for (fi = min;; fi++)
5210            {            {
# Line 4957  for (;;) Line 5221  for (;;)
5221            c = *eptr++;            c = *eptr++;
5222            switch(ctype)            switch(ctype)
5223              {              {
5224              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5225                if (md->partial != 0 &&    /* Take care with CRLF partial */
5226                    eptr >= md->end_subject &&
5227                    NLBLOCK->nltype == NLTYPE_FIXED &&
5228                    NLBLOCK->nllen == 2 &&
5229                    c == NLBLOCK->nl[0])
5230                  {
5231                  md->hitend = TRUE;
5232                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5233                  }
5234                break;
5235    
5236              case OP_ALLANY:              case OP_ALLANY:
5237              case OP_ANYBYTE:              case OP_ANYBYTE:
5238              break;              break;
# Line 4966  for (;;) Line 5241  for (;;)
5241              switch(c)              switch(c)
5242                {                {
5243                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5244                case 0x000d:                case CHAR_CR:
5245                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5246                break;                break;
5247    
5248                case 0x000a:                case CHAR_LF:
5249                break;                break;
5250    
5251                case 0x000b:                case CHAR_VT:
5252                case 0x000c:                case CHAR_FF:
5253                case 0x0085:                case CHAR_NEL:
5254    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5255                  case 0x2028:
5256                  case 0x2029:
5257    #endif
5258                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5259                break;                break;
5260                }                }
# Line 4985  for (;;) Line 5264  for (;;)
5264              switch(c)              switch(c)
5265                {                {
5266                default: break;                default: break;
5267                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5268                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5269                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
5270    #endif
5271                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5272                }                }
5273              break;              break;
# Line 4996  for (;;) Line 5276  for (;;)
5276              switch(c)              switch(c)
5277                {                {
5278                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5279                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5280                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5281                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
5282    #endif
5283                break;                break;
5284                }                }
5285              break;              break;
# Line 5007  for (;;) Line 5288  for (;;)
5288              switch(c)              switch(c)
5289                {                {
5290                default: break;                default: break;
5291                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5292                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5293                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
5294                case 0x0d:      /* CR */  #endif
               case 0x85:      /* NEL */  
5295                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5296                }                }
5297              break;              break;
# Line 5020  for (;;) Line 5300  for (;;)
5300              switch(c)              switch(c)
5301                {                {
5302                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5303                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5304                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5305                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
5306                case 0x0d:      /* CR */  #endif
               case 0x85:      /* NEL */  
5307                break;                break;
5308                }                }
5309              break;              break;
5310    
5311              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
5312              if ((md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
5313              break;              break;
5314    
5315              case OP_DIGIT:              case OP_DIGIT:
5316              if ((md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);
5317              break;              break;
5318    
5319              case OP_NOT_WHITESPACE:              case OP_NOT_WHITESPACE:
5320              if ((md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);
5321              break;              break;
5322    
5323              case OP_WHITESPACE:              case OP_WHITESPACE:
5324              if  ((md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);
5325              break;              break;
5326    
5327              case OP_NOT_WORDCHAR:              case OP_NOT_WORDCHAR:
5328              if ((md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);
5329              break;              break;
5330    
5331              case OP_WORDCHAR:              case OP_WORDCHAR:
5332              if ((md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);
5333              break;              break;
5334    
5335              default:              default:
# Line 5227  for (;;) Line 5506  for (;;)
5506              }              }
5507            break;            break;
5508    
5509              case PT_CLIST:
5510              for (i = min; i < max; i++)
5511                {
5512                const pcre_uint32 *cp;
5513                int len = 1;
5514                if (eptr >= md->end_subject)
5515                  {
5516                  SCHECK_PARTIAL();
5517                  break;
5518                  }
5519                GETCHARLENTEST(c, eptr, len);
5520                cp = PRIV(ucd_caseless_sets) + prop_value;
5521                for (;;)
5522                  {
5523                  if (c < *cp)
5524                    { if (prop_fail_result) break; else goto GOT_MAX; }
5525                  if (c == *cp++)
5526                    { if (prop_fail_result) goto GOT_MAX; else break; }
5527                  }
5528                eptr += len;
5529                }
5530              GOT_MAX:
5531              break;
5532    
5533              case PT_UCNC:
5534              for (i = min; i < max; i++)
5535                {
5536                int len = 1;
5537                if (eptr >= md->end_subject)
5538                  {
5539                  SCHECK_PARTIAL();
5540                  break;
5541                  }
5542                GETCHARLENTEST(c, eptr, len);
5543                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5544                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5545                     c >= 0xe000) == prop_fail_result)
5546                  break;
5547                eptr += len;
5548                }
5549              break;
5550    
5551            default:            default:
5552            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5553            }            }
# Line 5239  for (;;) Line 5560  for (;;)
5560            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5561            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5562            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5563            if (utf8) BACKCHAR(eptr);            if (utf) BACKCHAR(eptr);
5564            }            }
5565          }          }
5566    
# Line 5250  for (;;) Line 5571  for (;;)
5571          {          {
5572          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5573            {            {
           int len = 1;  
5574            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5575              {              {
5576              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5577              break;              break;
5578              }              }
5579            if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
           if (UCD_CATEGORY(c) == ucp_M) break;  
           eptr += len;  
           while (eptr < md->end_subject)  
5580              {              {
5581              len = 1;              int lgb, rgb;
5582              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5583              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5584              eptr += len;              while (eptr < md->end_subject)
5585                  {
5586                  int len = 1;
5587                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5588                  rgb = UCD_GRAPHBREAK(c);
5589                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5590                  lgb = rgb;
5591                  eptr += len;
5592                  }
5593              }              }
5594              CHECK_PARTIAL();
5595            }            }
5596    
5597          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
# Line 5279  for (;;) Line 5605  for (;;)
5605            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5606            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
5607              {              {
5608              if (!utf8) c = *eptr; else              if (!utf) c = *eptr; else
5609                {                {
5610                BACKCHAR(eptr);                BACKCHAR(eptr);
5611                GETCHAR(c, eptr);                GETCHAR(c, eptr);
# Line 5293  for (;;) Line 5619  for (;;)
5619        else        else
5620  #endif   /* SUPPORT_UCP */  #endif   /* SUPPORT_UCP */
5621    
5622  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5623        /* UTF-8 mode */        if (utf)
   
       if (utf8)  
5624          {          {
5625          switch(ctype)          switch(ctype)
5626            {            {
# Line 5311  for (;;) Line 5635  for (;;)
5635                  break;                  break;
5636                  }                  }
5637                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5638                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5639                      eptr + 1 >= md->end_subject &&
5640                      NLBLOCK->nltype == NLTYPE_FIXED &&
5641                      NLBLOCK->nllen == 2 &&
5642                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5643                    {
5644                    md->hitend = TRUE;
5645                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5646                    }
5647                eptr++;                eptr++;
5648                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5649                }                }
5650              }              }
5651    
# Line 5328  for (;;) Line 5661  for (;;)
5661                  break;                  break;
5662                  }                  }
5663                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5664                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5665                      eptr + 1 >= md->end_subject &&
5666                      NLBLOCK->nltype == NLTYPE_FIXED &&
5667                      NLBLOCK->nllen == 2 &&
5668                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5669                    {
5670                    md->hitend = TRUE;
5671                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5672                    }
5673                eptr++;                eptr++;
5674                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5675                }                }
5676              }              }
5677            break;            break;
# Line 5345  for (;;) Line 5687  for (;;)
5687                  break;                  break;
5688                  }                  }
5689                eptr++;                eptr++;
5690                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5691                }                }
5692              }              }
5693            else            else
# Line 5377  for (;;) Line 5719  for (;;)
5719                break;                break;
5720                }                }
5721              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5722              if (c == 0x000d)              if (c == CHAR_CR)
5723                {                {
5724                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5725                if (*eptr == 0x000a) eptr++;                if (RAWUCHAR(eptr) == CHAR_LF) eptr++;
5726                }                }
5727              else              else
5728                {                {
5729                if (c != 0x000a &&                if (c != CHAR_LF &&
5730                    (md->bsr_anycrlf ||                    (md->bsr_anycrlf ||
5731                     (c != 0x000b && c != 0x000c &&                     (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5732                      c != 0x0085 && c != 0x2028 && c != 0x2029)))  #ifndef EBCDIC
5733                        && c != 0x2028 && c != 0x2029
5734    #endif  /* Not EBCDIC */
5735                        )))
5736                  break;                  break;
5737                eptr += len;                eptr += len;
5738                }                }
# Line 5408  for (;;) Line 5753  for (;;)
5753              GETCHARLEN(c, eptr, len);</