/[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 926 by ph10, Wed Feb 22 15:01:32 2012 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) 1997-2012 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 115  Returns:     nothing Line 126  Returns:     nothing
126  static void  static void
127  pchars(const pcre_uchar *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 142  Arguments: Line 154  Arguments:
154    
155  Returns:      >= 0 the number of subject bytes matched  Returns:      >= 0 the number of subject bytes matched
156                -1 no match                -1 no match
157                -2 partial match; always given if at end subject                -2 partial match; always given if at end subject
158  */  */
159    
160  static int  static int
# Line 151  match_ref(int offset, register PCRE_PUCH Line 163  match_ref(int offset, register PCRE_PUCH
163  {  {
164  PCRE_PUCHAR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
165  register PCRE_PUCHAR 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 165  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 - in that  /* Always fail if reference not set (and not JavaScript compatible - in that
184  case the length is passed as zero). */  case the length is passed as zero). */
185    
186  if (length < 0) return -1;  if (length < 0) return -1;
# Line 178  if (caseless) Line 193  if (caseless)
193    {    {
194  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
195  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
196    if (md->utf)    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      PCRE_PUCHAR 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          const ucd_record *ur;
212        if (eptr >= md->end_subject) return -2;   /* Partial match */        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 207  if (caseless) Line 233  if (caseless)
233      {      {
234      while (length-- > 0)      while (length-- > 0)
235        {        {
236          pcre_uint32 cc, cp;
237        if (eptr >= md->end_subject) return -2;   /* Partial match */        if (eptr >= md->end_subject) return -2;   /* Partial match */
238        if (TABLE_GET(*p, md->lcc, *p) != TABLE_GET(*eptr, md->lcc, *eptr)) return -1;        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++;        p++;
242        eptr++;        eptr++;
243        }        }
# Line 220  are in UTF-8 mode. */ Line 249  are in UTF-8 mode. */
249    
250  else  else
251    {    {
252    while (length-- > 0)    while (length-- > 0)
253      {      {
254      if (eptr >= md->end_subject) return -2;   /* Partial match */      if (eptr >= md->end_subject) return -2;   /* Partial match */
255      if (*p++ != *eptr++) return -1;      if (RAWUCHARINCTEST(p) != RAWUCHARINCTEST(eptr)) return -1;
256      }      }
257    }    }
258    
259  return (int)(eptr - eptr_start);  return (int)(eptr - eptr_start);
# Line 279  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 297  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 317  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 *)(PUBL(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 338  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;\
   if (oldframe != &frame_zero) (PUBL(stack_free))(oldframe);\  
376    if (frame != NULL)\    if (frame != NULL)\
377      {\      {\
378      rrc = ra;\      rrc = ra;\
# Line 352  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    
# Line 382  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    pcre_uchar Xocchars[6];    pcre_uchar Xocchars[6];
# Line 395  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 483  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 utf;         /* Local copy of UTF 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 */
# Line 498  the top-level on the stack rather than m Line 533  the top-level on the stack rather than m
533  boost in many cases where there is not much "recursion". */  boost in many cases where there is not much "recursion". */
534    
535  #ifdef NO_RECURSE  #ifdef NO_RECURSE
536  heapframe frame_zero;  heapframe *frame = (heapframe *)md->match_frames_base;
 heapframe *frame = &frame_zero;  
 frame->Xprevframe = NULL;            /* Marks the top level */  
537    
538  /* Copy in the original argument variables */  /* Copy in the original argument variables */
539    
# Line 602  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  pcre_uchar occhars[6];  pcre_uchar occhars[6];
# Line 613  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    
# Line 734  for (;;) Line 767  for (;;)
767      unaltered. */      unaltered. */
768    
769      else if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
770          STRCMP_UC_UC(ecode + 2, 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 903  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 1032  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 1047  for (;;) Line 1080  for (;;)
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        save_mark = md->mark;        save_mark = md->mark;
1083          save_capture_last = md->capture_last;
1084        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1085          RM2);          RM2);
1086    
# Line 1078  for (;;) Line 1112  for (;;)
1112        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1113        md->mark = save_mark;        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 1199  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        {        {
# Line 1228  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 1259  for (;;) Line 1296  for (;;)
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  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1300          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1301  #else  #elif defined COMPILE_PCRE16
1302          cb.subject          = (PCRE_SPTR16)md->start_subject;          cb.subject          = (PCRE_SPTR16)md->start_subject;
1303    #elif defined COMPILE_PCRE32
1304            cb.subject          = (PCRE_SPTR32)md->start_subject;
1305  #endif  #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);
# Line 1270  for (;;) Line 1309  for (;;)
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 = (*PUBL(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 += PRIV(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 1292  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 1364  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          pcre_uchar *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++)
# Line 1492  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 1500  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];
# Line 1571  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);
       md->mark = save_mark;  
1628        }        }
1629      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1630    
# Line 1677  for (;;) Line 1724  for (;;)
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  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1728        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1729  #else  #elif defined COMPILE_PCRE16
1730        cb.subject          = (PCRE_SPTR16)md->start_subject;        cb.subject          = (PCRE_SPTR16)md->start_subject;
1731    #elif defined COMPILE_PCRE32
1732          cb.subject          = (PCRE_SPTR32)md->start_subject;
1733  #endif  #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);
# Line 1688  for (;;) Line 1737  for (;;)
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 = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
# Line 1717  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 1734  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 1757  for (;;) Line 1809  for (;;)
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);
# Line 1769  for (;;) Line 1822  for (;;)
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            {            {
# Line 1785  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)
# Line 1918  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 1999  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 2008  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 2065  for (;;) Line 2118  for (;;)
2118    
2119      case OP_DOLLM:      case OP_DOLLM:
2120      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2121        {        {
2122        if (!IS_NEWLINE(eptr))        if (!IS_NEWLINE(eptr))
2123          {          {
2124          if (md->partial != 0 &&          if (md->partial != 0 &&
2125              eptr + 1 >= md->end_subject &&              eptr + 1 >= md->end_subject &&
2126              NLBLOCK->nltype == NLTYPE_FIXED &&              NLBLOCK->nltype == NLTYPE_FIXED &&
2127              NLBLOCK->nllen == 2 &&              NLBLOCK->nllen == 2 &&
2128              *eptr == NLBLOCK->nl[0])              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2129            {            {
2130            md->hitend = TRUE;            md->hitend = TRUE;
2131            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2132            }            }
2133          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2134          }          }
2135        }        }
2136      else      else
2137        {        {
# Line 2115  for (;;) Line 2168  for (;;)
2168        if (md->partial != 0 &&        if (md->partial != 0 &&
2169            eptr + 1 >= md->end_subject &&            eptr + 1 >= md->end_subject &&
2170            NLBLOCK->nltype == NLTYPE_FIXED &&            NLBLOCK->nltype == NLTYPE_FIXED &&
2171            NLBLOCK->nllen == 2 &&            NLBLOCK->nllen == 2 &&
2172            *eptr == NLBLOCK->nl[0])            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2173          {          {
2174          md->hitend = TRUE;          md->hitend = TRUE;
2175          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);          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 2258  for (;;) Line 2311  for (;;)
2311      if (md->partial != 0 &&      if (md->partial != 0 &&
2312          eptr + 1 >= md->end_subject &&          eptr + 1 >= md->end_subject &&
2313          NLBLOCK->nltype == NLTYPE_FIXED &&          NLBLOCK->nltype == NLTYPE_FIXED &&
2314          NLBLOCK->nllen == 2 &&          NLBLOCK->nllen == 2 &&
2315          *eptr == NLBLOCK->nl[0])          RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2316        {        {
2317        md->hitend = TRUE;        md->hitend = TRUE;
2318        if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);        if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2319        }        }
2320    
2321      /* Fall through */      /* Fall through */
2322    
2323      /* Match any single character whatsoever. */      /* 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 2408  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)        if (eptr >= md->end_subject)
2466          {          {
2467          SCHECK_PARTIAL();          SCHECK_PARTIAL();
2468          }          }
2469        else if (*eptr == 0x0a) eptr++;        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 2439  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 2473  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 2507  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 2529  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 2555  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 2615  for (;;) Line 2619  for (;;)
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 2634  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 (!utf) 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();      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 2709  for (;;) Line 2737  for (;;)
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 */          if (length == -2) eptr = md->end_subject;   /* Partial match */
2741          CHECK_PARTIAL();          CHECK_PARTIAL();
2742          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2743          }          }
# Line 2735  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 */          if (slength == -2) eptr = md->end_subject;   /* Partial match */
2767          CHECK_PARTIAL();          CHECK_PARTIAL();
2768          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2769          }          }
# Line 2759  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 */            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2791            CHECK_PARTIAL();            CHECK_PARTIAL();
2792            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2793            }            }
# Line 2778  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            /* Can't use CHECK_PARTIAL because we don't want to update eptr in            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2810            the soft partial matching case. */            the soft partial matching case. */
2811    
2812            if (slength == -2 && md->partial != 0 &&            if (slength == -2 && md->partial != 0 &&
2813                md->end_subject > md->start_used_ptr)                md->end_subject > md->start_used_ptr)
2814              {              {
2815              md->hitend = TRUE;              md->hitend = TRUE;
2816              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2817              }              }
# Line 2791  for (;;) Line 2819  for (;;)
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 3155  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
# Line 3195  for (;;) Line 3223  for (;;)
3223    
3224        if (fc < 128)        if (fc < 128)
3225          {          {
3226          if (md->lcc[fc]          pcre_uint32 cc = RAWUCHAR(eptr);
3227              != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);          if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3228          ecode++;          ecode++;
3229          eptr++;          eptr++;
3230          }          }
# Line 3207  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 3317  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 = PRIV(ord2utf)(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
# Line 3421  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        {        {
# Line 3444  for (;;) Line 3472  for (;;)
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 != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3482            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3483          eptr++;          eptr++;
3484          }          }
3485        if (min == max) continue;        if (min == max) continue;
# Line 3457  for (;;) Line 3487  for (;;)
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 3465  for (;;) Line 3496  for (;;)
3496              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3497              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3498              }              }
3499            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3500              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3501            eptr++;            eptr++;
3502            }            }
3503          /* Control never gets here */          /* Control never gets here */
# Line 3475  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 != *eptr && foc != *eptr) break;            cc = RAWUCHARTEST(eptr);
3517              if (fc != cc && foc != cc) break;
3518            eptr++;            eptr++;
3519            }            }
3520    
# Line 3508  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 3525  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 3539  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 3568  for (;;) Line 3602  for (;;)
3602  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3603      if (utf)      if (utf)
3604        {        {
3605        register unsigned int ch, och;        register pcre_uint32 ch, och;
3606    
3607        ecode++;        ecode++;
3608        GETCHARINC(ch, ecode);        GETCHARINC(ch, ecode);
# Line 3595  for (;;) Line 3629  for (;;)
3629      else      else
3630  #endif  #endif
3631        {        {
3632        register unsigned int ch = ecode[1];        register pcre_uint32 ch = ecode[1];
3633        c = *eptr++;        c = *eptr++;
3634        if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))        if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3635          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
# Line 3690  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        {        {
# Line 3709  for (;;) Line 3743  for (;;)
3743  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3744        if (utf)        if (utf)
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 3744  for (;;) Line 3778  for (;;)
3778  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3779          if (utf)          if (utf)
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 3789  for (;;) Line 3823  for (;;)
3823  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3824          if (utf)          if (utf)
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 3846  for (;;) Line 3880  for (;;)
3880  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3881        if (utf)        if (utf)
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 3880  for (;;) Line 3914  for (;;)
3914  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3915          if (utf)          if (utf)
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 3924  for (;;) Line 3958  for (;;)
3958  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3959          if (utf)          if (utf)
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 4200  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 4219  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 (!utf) 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();            CHECK_PARTIAL();
4309            }            }
4310          }          }
4311    
# Line 4252  for (;;) Line 4329  for (;;)
4329            if (md->partial != 0 &&            if (md->partial != 0 &&
4330                eptr + 1 >= md->end_subject &&                eptr + 1 >= md->end_subject &&
4331                NLBLOCK->nltype == NLTYPE_FIXED &&                NLBLOCK->nltype == NLTYPE_FIXED &&
4332                NLBLOCK->nllen == 2 &&                NLBLOCK->nllen == 2 &&
4333                *eptr == NLBLOCK->nl[0])                RAWUCHAR(eptr) == NLBLOCK->nl[0])
4334              {              {
4335              md->hitend = TRUE;              md->hitend = TRUE;
4336              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4337              }              }
# Line 4294  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 4323  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 4359  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 4395  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 4419  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 4449  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++;            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 */
# Line 4464  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            eptr++;            eptr++;
4506            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4479  for (;;) Line 4510  for (;;)
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++;            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 */
# Line 4494  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            eptr++;            eptr++;
4540            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4509  for (;;) Line 4544  for (;;)
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++;            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 */
# Line 4545  for (;;) Line 4582  for (;;)
4582            if (md->partial != 0 &&            if (md->partial != 0 &&
4583                eptr + 1 >= md->end_subject &&                eptr + 1 >= md->end_subject &&
4584                NLBLOCK->nltype == NLTYPE_FIXED &&                NLBLOCK->nltype == NLTYPE_FIXED &&
4585                NLBLOCK->nllen == 2 &&                NLBLOCK->nllen == 2 &&
4586                *eptr == NLBLOCK->nl[0])                *eptr == NLBLOCK->nl[0])
4587              {              {
4588              md->hitend = TRUE;              md->hitend = TRUE;
4589              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4590              }              }
# Line 4585  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  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4636              case 0x2028:              case 0x2028:
4637              case 0x2029:              case 0x2029:
4638  #endif  #endif
# Line 4616  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:
 #ifdef COMPILE_PCRE16  
             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 */  
4659  #endif  #endif
4660              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4661              }              }
# Line 4653  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:
 #ifdef COMPILE_PCRE16  
             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 */  
4679  #endif  #endif
4680              break;              break;
4681              }              }
# Line 4689  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:
             case 0x0c:      /* FF */  
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
4698  #endif  #endif
4699              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4700                default: break;
4701              }              }
4702            }            }
4703          break;          break;
# Line 4715  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:
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
4719  #endif  #endif
4720              break;              break;
4721              }              }
# Line 5000  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 5022  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 (!utf) 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();            CHECK_PARTIAL();
5076            }            }
5077          }          }
5078        else        else
# Line 5059  for (;;) Line 5100  for (;;)
5100              if (md->partial != 0 &&    /* Take care with CRLF partial */              if (md->partial != 0 &&    /* Take care with CRLF partial */
5101                  eptr >= md->end_subject &&                  eptr >= md->end_subject &&
5102                  NLBLOCK->nltype == NLTYPE_FIXED &&                  NLBLOCK->nltype == NLTYPE_FIXED &&
5103                  NLBLOCK->nllen == 2 &&                  NLBLOCK->nllen == 2 &&
5104                  c == NLBLOCK->nl[0])                  c == NLBLOCK->nl[0])
5105                {                {
5106                md->hitend = TRUE;                md->hitend = TRUE;
5107                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5108                }                }
5109              break;              break;
5110    
5111              case OP_ALLANY:              case OP_ALLANY:
5112              case OP_ANYBYTE:              case OP_ANYBYTE:
5113              break;              break;
# Line 5075  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 5094  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 5233  for (;;) Line 5225  for (;;)
5225              if (md->partial != 0 &&    /* Take care with CRLF partial */              if (md->partial != 0 &&    /* Take care with CRLF partial */
5226                  eptr >= md->end_subject &&                  eptr >= md->end_subject &&
5227                  NLBLOCK->nltype == NLTYPE_FIXED &&                  NLBLOCK->nltype == NLTYPE_FIXED &&
5228                  NLBLOCK->nllen == 2 &&                  NLBLOCK->nllen == 2 &&
5229                  c == NLBLOCK->nl[0])                  c == NLBLOCK->nl[0])
5230                {                {
5231                md->hitend = TRUE;                md->hitend = TRUE;
5232                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5233                }                }
5234              break;              break;
5235    
5236              case OP_ALLANY:              case OP_ALLANY:
5237              case OP_ANYBYTE:              case OP_ANYBYTE:
5238              break;              break;
# Line 5249  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  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5255                case 0x2028:                case 0x2028:
5256                case 0x2029:                case 0x2029:
5257  #endif  #endif
# Line 5272  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:
 #ifdef COMPILE_PCRE16  
               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 */  
5270  #endif  #endif
5271                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5272                }                }
# Line 5301  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:
 #ifdef COMPILE_PCRE16  
               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 */  
5282  #endif  #endif
5283                break;                break;
5284                }                }
# Line 5330  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:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5294  #endif  #endif
5295                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5296                }                }
# Line 5347  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:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5306  #endif  #endif
5307                break;                break;
5308                }                }
# Line 5558  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 5581  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 (!utf) 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 (!utf) 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();            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 5644  for (;;) Line 5638  for (;;)
5638                if (md->partial != 0 &&    /* Take care with CRLF partial */                if (md->partial != 0 &&    /* Take care with CRLF partial */
5639                    eptr + 1 >= md->end_subject &&                    eptr + 1 >= md->end_subject &&
5640                    NLBLOCK->nltype == NLTYPE_FIXED &&                    NLBLOCK->nltype == NLTYPE_FIXED &&
5641                    NLBLOCK->nllen == 2 &&                    NLBLOCK->nllen == 2 &&
5642                    *eptr == NLBLOCK->nl[0])                    RAWUCHAR(eptr) == NLBLOCK->nl[0])
5643                  {                  {
5644                  md->hitend = TRUE;                  md->hitend = TRUE;
5645                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5646                  }                  }
# Line 5670  for (;;) Line 5664  for (;;)
5664                if (md->partial != 0 &&    /* Take care with CRLF partial */                if (md->partial != 0 &&    /* Take care with CRLF partial */
5665                    eptr + 1 >= md->end_subject &&                    eptr + 1 >= md->end_subject &&
5666                    NLBLOCK->nltype == NLTYPE_FIXED &&                    NLBLOCK->nltype == NLTYPE_FIXED &&
5667                    NLBLOCK->nllen == 2 &&                    NLBLOCK->nllen == 2 &&
5668                    *eptr == NLBLOCK->nl[0])                    RAWUCHAR(eptr) == NLBLOCK->nl[0])
5669                  {                  {
5670                  md->hitend = TRUE;                  md->hitend = TRUE;
5671                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5672                  }                  }
# Line 5725  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 5756  for (;;) Line 5753  for (;;)
5753              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5754              switch(c)              switch(c)
5755                {                {
5756                  HSPACE_CASES: gotspace = TRUE; break;
5757                default: gotspace = FALSE; break;                default: gotspace = FALSE; break;
               case 0x09:      /* HT */  
               case 0x20:      /* SPACE */  
               case 0xa0:      /* NBSP */  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
               gotspace = TRUE;  
               break;  
5758                }                }
5759              if (gotspace == (ctype == OP_NOT_HSPACE)) break;              if (gotspace == (ctype == OP_NOT_HSPACE)) break;
5760              eptr += len;              eptr += len;
# Line 5798  for (;;) Line 5775  for (;;)
5775              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5776              switch(c)              switch(c)
5777                {                {
5778                  VSPACE_CASES: gotspace = TRUE; break;
5779                default: gotspace = FALSE; break;                default: gotspace = FALSE; break;
               case 0x0a:      /* LF */  
               case 0x0b:      /* VT */  
               case 0x0c:      /* FF */  
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
               gotspace = TRUE;  
               break;  
5780                }                }
5781              if (gotspace == (ctype == OP_NOT_VSPACE)) break;              if (gotspace == (ctype == OP_NOT_VSPACE)) break;
5782              eptr += len;              eptr += len;
# Line 5921  for (;;) Line 5890  for (;;)
5890            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5891            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5892            BACKCHAR(eptr);            BACKCHAR(eptr);
5893            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && RAWUCHAR(eptr) == CHAR_NL &&
5894                eptr[-1] == '\r') eptr--;                RAWUCHAR(eptr - 1) == CHAR_CR) eptr--;
5895            }            }
5896          }          }
5897        else        else
# Line 5943  for (;;) Line 5912  for (;;)
5912              if (md->partial != 0 &&    /* Take care with CRLF partial */              if (md->partial != 0 &&    /* Take care with CRLF partial */
5913                  eptr + 1 >= md->end_subject &&                  eptr + 1 >= md->end_subject &&
5914                  NLBLOCK->nltype == NLTYPE_FIXED &&                  NLBLOCK->nltype == NLTYPE_FIXED &&
5915                  NLBLOCK->nllen == 2 &&                  NLBLOCK->nllen == 2 &&
5916                  *eptr == NLBLOCK->nl[0])                  *eptr == NLBLOCK->nl[0])
5917                {                {
5918                md->hitend = TRUE;                md->hitend = TRUE;
5919                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5920                }                }
# Line 5973  for (;;) Line 5942  for (;;)
5942                break;                break;
5943                }                }
5944              c = *eptr;              c = *eptr;
5945              if (c == 0x000d)              if (c == CHAR_CR)
5946                {                {
5947                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5948                if (*eptr == 0x000a) eptr++;                if (*eptr == CHAR_LF) eptr++;
5949                }                }
5950              else              else
5951                {                {
5952                if (c != 0x000a && (md->bsr_anycrlf ||                if (c != CHAR_LF && (md->bsr_anycrlf ||
5953                  (c != 0x000b && c != 0x000c && c != 0x0085                   (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5954  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5955                  && c != 0x2028 && c != 0x2029                   && c != 0x2028 && c != 0x2029
5956  #endif  #endif
5957                  ))) break;                   ))) break;
5958                eptr++;                eptr++;
5959                }                }
5960              }              }
# Line 5999  for (;;) Line 5968  for (;;)
5968                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5969                break;                break;
5970                }                }
5971              c = *eptr;              switch(*eptr)
5972              if (c == 0x09 || c == 0x20 || c == 0xa0                {
5973  #ifdef COMPILE_PCRE16                default: eptr++; break;
5974                || c == 0x1680 || c == 0x180e || (c >= 0x2000 && c <= 0x200A)                HSPACE_BYTE_CASES:
5975                || c == 0x202f || c == 0x205f || c == 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5976                  HSPACE_MULTIBYTE_CASES:
5977  #endif  #endif
5978                ) break;                goto ENDLOOP00;
5979              eptr++;                }
5980              }              }
5981              ENDLOOP00:
5982            break;            break;
5983    
5984            case OP_HSPACE:            case OP_HSPACE:
# Line 6018  for (;;) Line 5989  for (;;)
5989                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5990                break;                break;
5991                }                }
5992              c = *eptr;              switch(*eptr)
5993              if (c != 0x09 && c != 0x20 && c != 0xa0                {
5994  #ifdef COMPILE_PCRE16                default: goto ENDLOOP01;
5995                && c != 0x1680 && c != 0x180e && (c < 0x2000 || c > 0x200A)                HSPACE_BYTE_CASES:
5996                && c != 0x202f && c != 0x205f && c != 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5997                  HSPACE_MULTIBYTE_CASES:
5998  #endif  #endif
5999                ) break;                eptr++; break;
6000              eptr++;                }
6001              }              }
6002              ENDLOOP01:
6003            break;            break;
6004    
6005            case OP_NOT_VSPACE:            case OP_NOT_VSPACE:
# Line 6037  for (;;) Line 6010  for (;;)
6010                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6011                break;                break;
6012                }                }
6013              c = *eptr;              switch(*eptr)
6014              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85                {
6015  #ifdef COMPILE_PCRE16                default: eptr++; break;
6016                || c == 0x2028 || c == 0x2029                VSPACE_BYTE_CASES:
6017    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6018                  VSPACE_MULTIBYTE_CASES:
6019  #endif  #endif
6020                ) break;                goto ENDLOOP02;
6021              eptr++;                }
6022              }              }
6023              ENDLOOP02:
6024            break;            break;
6025    
6026            case OP_VSPACE:            case OP_VSPACE:
# Line 6055  for (;;) Line 6031  for (;;)
6031                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6032                break;                break;
6033                }                }
6034              c = *eptr;              switch(*eptr)
6035              if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85                {
6036  #ifdef COMPILE_PCRE16                default: goto ENDLOOP03;
6037                && c != 0x2028 && c != 0x2029                VSPACE_BYTE_CASES:
6038    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6039                  VSPACE_MULTIBYTE_CASES:
6040  #endif  #endif
6041                ) break;                eptr++; break;
6042              eptr++;                }
6043              }              }
6044              ENDLOOP03:
6045            break;            break;
6046    
6047            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
# Line 6159  for (;;) Line 6138  for (;;)
6138            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);
6139            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
6140            eptr--;            eptr--;
6141            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && *eptr == CHAR_LF &&
6142                eptr[-1] == '\r') eptr--;                eptr[-1] == CHAR_CR) eptr--;
6143            }            }
6144          }          }
6145    
# Line 6210  switch (frame->Xwhere) Line 6189  switch (frame->Xwhere)
6189    LBL(32) LBL(34) LBL(42) LBL(46)    LBL(32) LBL(34) LBL(42) LBL(46)
6190  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6191    LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45)    LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45)
6192    LBL(59) LBL(60) LBL(61) LBL(62)    LBL(59) LBL(60) LBL(61) LBL(62) LBL(67) LBL(68)
6193  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
6194  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
6195    default:    default:
6196    DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));    DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));
   
 printf("+++jump error in pcre match: label %d non-existent\n", frame->Xwhere);  
   
6197    return PCRE_ERROR_INTERNAL;    return PCRE_ERROR_INTERNAL;
6198    }    }
6199  #undef LBL  #undef LBL
# Line 6279  Undefine all the macros that were define Line 6255  Undefine all the macros that were define
6255  ***************************************************************************/  ***************************************************************************/
6256    
6257    
6258    #ifdef NO_RECURSE
6259    /*************************************************
6260    *          Release allocated heap frames         *
6261    *************************************************/
6262    
6263    /* This function releases all the allocated frames. The base frame is on the
6264    machine stack, and so must not be freed.
6265    
6266    Argument: the address of the base frame
6267    Returns:  nothing
6268    */
6269    
6270    static void
6271    release_match_heapframes (heapframe *frame_base)
6272    {
6273    heapframe *nextframe = frame_base->Xnextframe;
6274    while (nextframe != NULL)
6275      {
6276      heapframe *oldframe = nextframe;
6277      nextframe = nextframe->Xnextframe;
6278      (PUBL(stack_free))(oldframe);
6279      }
6280    }
6281    #endif
6282    
6283    
6284  /*************************************************  /*************************************************
6285  *         Execute a Regular Expression           *  *         Execute a Regular Expression           *
# Line 6304  Returns:          > 0 => success; value Line 6305  Returns:          > 0 => success; value
6305                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
6306  */  */
6307    
6308  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
6309  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6310  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
6311    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
6312    int offsetcount)    int offsetcount)
6313  #else  #elif defined COMPILE_PCRE16
6314  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6315  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,
6316    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,
6317    int offsetcount)    int offsetcount)
6318    #elif defined COMPILE_PCRE32
6319    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6320    pcre32_exec(const pcre32 *argument_re, const pcre32_extra *extra_data,
6321      PCRE_SPTR32 subject, int length, int start_offset, int options, int *offsets,
6322      int offsetcount)
6323  #endif  #endif
6324  {  {
6325  int rc, ocount, arg_offset_max;  int rc, ocount, arg_offset_max;
# Line 6336  const pcre_uint8 *start_bits = NULL; Line 6342  const pcre_uint8 *start_bits = NULL;
6342  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;
6343  PCRE_PUCHAR end_subject;  PCRE_PUCHAR end_subject;
6344  PCRE_PUCHAR start_partial = NULL;  PCRE_PUCHAR start_partial = NULL;
6345    PCRE_PUCHAR match_partial;
6346  PCRE_PUCHAR req_char_ptr = start_match - 1;  PCRE_PUCHAR req_char_ptr = start_match - 1;
6347    
6348  const pcre_study_data *study;  const pcre_study_data *study;
6349  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;
6350    
6351    #ifdef NO_RECURSE
6352    heapframe frame_zero;
6353    frame_zero.Xprevframe = NULL;            /* Marks the top level */
6354    frame_zero.Xnextframe = NULL;            /* None are allocated yet */
6355    md->match_frames_base = &frame_zero;
6356    #endif
6357    
6358  /* Check for the special magic call that measures the size of the stack used  /* Check for the special magic call that measures the size of the stack used
6359  per recursive call of match(). Without the funny casting for sizeof, a Windows  per recursive call of match(). Without the funny casting for sizeof, a Windows
6360  compiler gave this error: "unary minus operator applied to unsigned type,  compiler gave this error: "unary minus operator applied to unsigned type,
6361  result still unsigned". Hopefully the cast fixes that. */  result still unsigned". Hopefully the cast fixes that. */
6362    
# Line 6360  if ((options & ~PUBLIC_EXEC_OPTIONS) != Line 6374  if ((options & ~PUBLIC_EXEC_OPTIONS) !=
6374  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))
6375    return PCRE_ERROR_NULL;    return PCRE_ERROR_NULL;
6376  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
6377    if (length < 0) return PCRE_ERROR_BADLENGTH;
6378  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
6379    
6380  /* Check that the first field in the block is the magic number. If it is not,  /* Check that the first field in the block is the magic number. If it is not,
# Line 6397  if (utf && (options & PCRE_NO_UTF8_CHECK Line 6412  if (utf && (options & PCRE_NO_UTF8_CHECK
6412        offsets[0] = erroroffset;        offsets[0] = erroroffset;
6413        offsets[1] = errorcode;        offsets[1] = errorcode;
6414        }        }
6415  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE8
     return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?  
       PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;  
 #else  
6416      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?
6417        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
6418    #elif defined COMPILE_PCRE16
6419        return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?
6420          PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;
6421    #elif defined COMPILE_PCRE32
6422        return PCRE_ERROR_BADUTF32;
6423  #endif  #endif
6424      }      }
6425    #if defined COMPILE_PCRE8 || defined COMPILE_PCRE16
6426    /* Check that a start_offset points to the start of a UTF character. */    /* Check that a start_offset points to the start of a UTF character. */
6427    if (start_offset > 0 && start_offset < length &&    if (start_offset > 0 && start_offset < length &&
6428        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))
6429      return PCRE_ERROR_BADUTF8_OFFSET;      return PCRE_ERROR_BADUTF8_OFFSET;
6430    #endif
6431    }    }
6432  #endif  #endif
6433    
# Line 6423  if (extra_data != NULL Line 6441  if (extra_data != NULL
6441      && (extra_data->flags & (PCRE_EXTRA_EXECUTABLE_JIT |      && (extra_data->flags & (PCRE_EXTRA_EXECUTABLE_JIT |
6442                               PCRE_EXTRA_TABLES)) == PCRE_EXTRA_EXECUTABLE_JIT                               PCRE_EXTRA_TABLES)) == PCRE_EXTRA_EXECUTABLE_JIT
6443      && extra_data->executable_jit != NULL      && extra_data->executable_jit != NULL
6444      && (options & ~(PCRE_NO_UTF8_CHECK | PCRE_NOTBOL | PCRE_NOTEOL |      && (options & ~PUBLIC_JIT_EXEC_OPTIONS) == 0)
                     PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART |  
                     PCRE_PARTIAL_SOFT | PCRE_PARTIAL_HARD)) == 0)  
6445    {    {
6446    rc = PRIV(jit_exec)(re, extra_data->executable_jit,    rc = PRIV(jit_exec)(extra_data, (const pcre_uchar *)subject, length,
6447      (const pcre_uchar *)subject, length, start_offset, options,         start_offset, options, offsets, offsetcount);
6448      ((extra_data->flags & PCRE_EXTRA_MATCH_LIMIT) == 0)  
     ? MATCH_LIMIT : extra_data->match_limit, offsets, offsetcount);  
   
6449    /* PCRE_ERROR_NULL means that the selected normal or partial matching    /* PCRE_ERROR_NULL means that the selected normal or partial matching
6450    mode is not compiled. In this case we simply fallback to interpreter. */    mode is not compiled. In this case we simply fallback to interpreter. */
6451    
6452    if (rc != PCRE_ERROR_NULL) return rc;    if (rc != PCRE_ERROR_JIT_BADOPTION) return rc;
6453    }    }
6454  #endif  #endif
6455    
# Line 6607  if (re->top_backref > 0 && re->top_backr Line 6621  if (re->top_backref > 0 && re->top_backr
6621    DPRINTF(("Got memory to hold back references\n"));    DPRINTF(("Got memory to hold back references\n"));
6622    }    }
6623  else md->offset_vector = offsets;  else md->offset_vector = offsets;
   
6624  md->offset_end = ocount;  md->offset_end = ocount;
6625  md->offset_max = (2*ocount)/3;  md->offset_max = (2*ocount)/3;
6626  md->offset_overflow = FALSE;  md->capture_last = 0;
 md->capture_last = -1;  
6627    
6628  /* Reset the working variable associated with each extraction. These should  /* Reset the working variable associated with each extraction. These should
6629  never be used unless previously set, but they get saved and restored, and so we  never be used unless previously set, but they get saved and restored, and so we
# Line 6719  for(;;) Line 6731  for(;;)
6731    
6732      if (has_first_char)      if (has_first_char)
6733        {        {
6734          pcre_uchar smc;
6735    
6736        if (first_char != first_char2)        if (first_char != first_char2)
6737          while (start_match < end_subject &&          while (start_match < end_subject &&
6738              *start_match != first_char && *start_match != first_char2)            (smc = RAWUCHARTEST(start_match)) != first_char && smc != first_char2)
6739            start_match++;            start_match++;
6740        else        else
6741          while (start_match < end_subject && *start_match != first_char)          while (start_match < end_subject && RAWUCHARTEST(start_match) != first_char)
6742            start_match++;            start_match++;
6743        }        }
6744    
# Line 6756  for(;;) Line 6770  for(;;)
6770          if (start_match[-1] == CHAR_CR &&          if (start_match[-1] == CHAR_CR &&
6771               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
6772               start_match < end_subject &&               start_match < end_subject &&
6773               *start_match == CHAR_NL)               RAWUCHARTEST(start_match) == CHAR_NL)
6774            start_match++;            start_match++;
6775          }          }
6776        }        }
# Line 6767  for(;;) Line 6781  for(;;)
6781        {        {
6782        while (start_match < end_subject)        while (start_match < end_subject)
6783          {          {
6784          register unsigned int c = *start_match;          register pcre_uint32 c = RAWUCHARTEST(start_match);
6785  #ifndef COMPILE_PCRE8  #ifndef COMPILE_PCRE8
6786          if (c > 255) c = 255;          if (c > 255) c = 255;
6787  #endif  #endif
# Line 6835  for(;;) Line 6849  for(;;)
6849            {            {
6850            while (p < end_subject)            while (p < end_subject)
6851              {              {
6852              register int pp = *p++;              register pcre_uint32 pp = RAWUCHARINCTEST(p);
6853              if (pp == req_char || pp == req_char2) { p--; break; }              if (pp == req_char || pp == req_char2) { p--; break; }
6854              }              }
6855            }            }
# Line 6843  for(;;) Line 6857  for(;;)
6857            {            {
6858            while (p < end_subject)            while (p < end_subject)
6859              {              {
6860              if (*p++ == req_char) { p--; break; }              if (RAWUCHARINCTEST(p) == req_char) { p--; break; }
6861              }              }
6862            }            }
6863    
# Line 6880  for(;;) Line 6894  for(;;)
6894    md->match_function_type = 0;    md->match_function_type = 0;
6895    md->end_offset_top = 0;    md->end_offset_top = 0;
6896    rc = match(start_match, md->start_code, start_match, 2, md, NULL, 0);    rc = match(start_match, md->start_code, start_match, 2, md, NULL, 0);
6897    if (md->hitend && start_partial == NULL) start_partial = md->start_used_ptr;    if (md->hitend && start_partial == NULL)
6898        {
6899        start_partial = md->start_used_ptr;
6900        match_partial = start_match;
6901        }
6902    
6903    switch(rc)    switch(rc)
6904      {      {
# Line 7006  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7024  if (rc == MATCH_MATCH || rc == MATCH_ACC
7024          (arg_offset_max - 2) * sizeof(int));          (arg_offset_max - 2) * sizeof(int));
7025        DPRINTF(("Copied offsets from temporary memory\n"));        DPRINTF(("Copied offsets from temporary memory\n"));
7026        }        }
7027      if (md->end_offset_top > arg_offset_max) md->offset_overflow = TRUE;      if (md->end_offset_top > arg_offset_max) md->capture_last |= OVFLBIT;
7028      DPRINTF(("Freeing temporary memory\n"));      DPRINTF(("Freeing temporary memory\n"));
7029      (PUBL(free))(md->offset_vector);      (PUBL(free))(md->offset_vector);
7030      }      }
# Line 7014  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7032  if (rc == MATCH_MATCH || rc == MATCH_ACC
7032    /* Set the return code to the number of captured strings, or 0 if there were    /* Set the return code to the number of captured strings, or 0 if there were
7033    too many to fit into the vector. */    too many to fit into the vector. */
7034    
7035    rc = (md->offset_overflow && md->end_offset_top >= arg_offset_max)?    rc = ((md->capture_last & OVFLBIT) != 0 &&
7036             md->end_offset_top >= arg_offset_max)?
7037      0 : md->end_offset_top/2;      0 : md->end_offset_top/2;
7038    
7039    /* If there is space in the offset vector, set any unused pairs at the end of    /* If there is space in the offset vector, set any unused pairs at the end of
# Line 7030  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7049  if (rc == MATCH_MATCH || rc == MATCH_ACC
7049      {      {
7050      register int *iptr, *iend;      register int *iptr, *iend;
7051      int resetcount = 2 + re->top_bracket * 2;      int resetcount = 2 + re->top_bracket * 2;
7052      if (resetcount > offsetcount) resetcount = ocount;      if (resetcount > offsetcount) resetcount = offsetcount;
7053      iptr = offsets + md->end_offset_top;      iptr = offsets + md->end_offset_top;
7054      iend = offsets + resetcount;      iend = offsets + resetcount;
7055      while (iptr < iend) *iptr++ = -1;      while (iptr < iend) *iptr++ = -1;
# Line 7051  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7070  if (rc == MATCH_MATCH || rc == MATCH_ACC
7070    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7071      *(extra_data->mark) = (pcre_uchar *)md->mark;      *(extra_data->mark) = (pcre_uchar *)md->mark;
7072    DPRINTF((">>>> returning %d\n", rc));    DPRINTF((">>>> returning %d\n", rc));
7073    #ifdef NO_RECURSE
7074      release_match_heapframes(&frame_zero);
7075    #endif
7076    return rc;    return rc;
7077    }    }
7078    
# Line 7068  if (using_temporary_offsets) Line 7090  if (using_temporary_offsets)
7090  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)
7091    {    {
7092    DPRINTF((">>>> error: returning %d\n", rc));    DPRINTF((">>>> error: returning %d\n", rc));
7093    #ifdef NO_RECURSE
7094      release_match_heapframes(&frame_zero);
7095    #endif
7096    return rc;    return rc;
7097    }    }
7098    
# Line 7081  if (start_partial != NULL) Line 7106  if (start_partial != NULL)
7106      {      {
7107      offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject);      offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject);
7108      offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject);      offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject);
7109        if (offsetcount > 2)
7110          offsets[2] = (int)(match_partial - (PCRE_PUCHAR)subject);
7111      }      }
7112    rc = PCRE_ERROR_PARTIAL;    rc = PCRE_ERROR_PARTIAL;
7113    }    }
# Line 7097  else Line 7124  else
7124    
7125  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7126    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;
7127    #ifdef NO_RECURSE
7128      release_match_heapframes(&frame_zero);
7129    #endif
7130  return rc;  return rc;
7131  }  }
7132    

Legend:
Removed from v.926  
changed lines
  Added in v.1265

  ViewVC Help
Powered by ViewVC 1.1.5