/[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 915 by zherczeg, Tue Feb 14 13:05:39 2012 UTC revision 1248 by ph10, Wed Feb 13 17:36:38 2013 UTC
# 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 140  Arguments: Line 152  Arguments:
152    md          points to match data block    md          points to match data block
153    caseless    TRUE if caseless    caseless    TRUE if caseless
154    
155  Returns:      < 0 if not matched, otherwise the number of subject bytes matched  Returns:      >= 0 the number of subject bytes matched
156                  -1 no match
157                  -2 partial match; always given if at end subject
158  */  */
159    
160  static int  static int
161  match_ref(int offset, register PCRE_PUCHAR eptr, int length, match_data *md,  match_ref(int offset, register PCRE_PUCHAR eptr, int length, match_data *md,
162    BOOL caseless)    BOOL caseless)
163  {  {
164  int matched_length = length;  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 163  pchars(p, length, FALSE, md); Line 180  pchars(p, length, FALSE, md);
180  printf("\n");  printf("\n");
181  #endif  #endif
182    
183  /* Always fail if reference not set (and not JavaScript compatible). */  /* Always fail if reference not set (and not JavaScript compatible - in that
184    case the length is passed as zero). */
185    
186  if (length < 0) return -1;  if (length < 0) return -1;
187    
# Line 175  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;
     PCRE_PUCHAR eptr_start = eptr;  
208      while (p < endptr)      while (p < endptr)
209        {        {
210        int c, d;        pcre_uint32 c, d;
211        if (eptr >= md->end_subject) return -((int)(eptr - eptr_start) + 1);        const ucd_record *ur;
212          if (eptr >= md->end_subject) return -2;   /* Partial match */
213        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
214        GETCHARINC(d, p);        GETCHARINC(d, p);
215        if (c != d && c != UCD_OTHERCASE(d)) return -1;        ur = GET_UCD(d);
216          if (c != d && c != d + ur->other_case)
217            {
218            const pcre_uint32 *pp = PRIV(ucd_caseless_sets) + ur->caseset;
219            for (;;)
220              {
221              if (c < *pp) return -1;
222              if (c == *pp++) break;
223              }
224            }
225        }        }
     matched_length = (int)(eptr - eptr_start);  
226      }      }
227    else    else
228  #endif  #endif
# Line 204  if (caseless) Line 231  if (caseless)
231    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there
232    is no UCP support. */    is no UCP support. */
233      {      {
     if (eptr + length > md->end_subject)  
       {  
       if (md->partial == 0)  
         return -1;  
       length = (int)(md->end_subject - eptr);  
       matched_length = -(length + 1);  
       }  
234      while (length-- > 0)      while (length-- > 0)
235        {        {
236        if (TABLE_GET(*p, md->lcc, *p) != TABLE_GET(*eptr, md->lcc, *eptr)) return -1;        pcre_uchar cc, cp;
237          if (eptr >= md->end_subject) return -2;   /* Partial match */
238          cc = RAWUCHARTEST(eptr);
239          cp = RAWUCHARTEST(p);
240          if (TABLE_GET(cp, md->lcc, cp) != TABLE_GET(cc, md->lcc, cc)) return -1;
241        p++;        p++;
242        eptr++;        eptr++;
243        }        }
# Line 225  are in UTF-8 mode. */ Line 249  are in UTF-8 mode. */
249    
250  else  else
251    {    {
252    if (eptr + length > md->end_subject)    while (length-- > 0)
253      {      {
254      if (md->partial == 0)      if (eptr >= md->end_subject) return -2;   /* Partial match */
255        return -1;      if (RAWUCHARINCTEST(p) != RAWUCHARINCTEST(eptr)) return -1;
     length = (int)(md->end_subject - eptr);  
     matched_length = -(length + 1);  
256      }      }
   while (length-- > 0) if (*p++ != *eptr++) return -1;  
257    }    }
258    
259  return matched_length;  return (int)(eptr - eptr_start);
260  }  }
261    
262    
# Line 287  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 308  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
308         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
309         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
310         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
311         RM61,  RM62, RM63, RM64, RM65, RM66 };         RM61,  RM62, RM63, RM64, RM65, RM66, RM67 };
312    
313  /* These versions of the macros use the stack, as normal. There are debugging  /* These versions of the macros use the stack, as normal. There are debugging
314  versions and production versions. Note that the "rw" argument of RMATCH isn't  versions and production versions. Note that the "rw" argument of RMATCH isn't
# Line 305  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 325  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 346  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 360  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 390  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 403  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 491  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 506  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 610  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 621  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 742  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 911  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 1040  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 1055  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 1086  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 1207  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 1236  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 1267  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 1278  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);
# Line 1300  for (;;) Line 1333  for (;;)
1333          }          }
1334        else        else
1335          {          {
1336          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          unsigned int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1337          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1338    
1339          /* If the test is for recursion into a specific subpattern, and it is          /* If the test is for recursion into a specific subpattern, and it is
# Line 1372  for (;;) Line 1405  for (;;)
1405    
1406        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1407          {          {
1408          int refno = offset >> 1;          unsigned int refno = offset >> 1;
1409          pcre_uchar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1410    
1411          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
# Line 1500  for (;;) Line 1533  for (;;)
1533      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1534    
1535      case OP_CLOSE:      case OP_CLOSE:
1536      number = GET2(ecode, 1);      number = GET2(ecode, 1);   /* Must be less than 65536 */
1537      offset = number << 1;      offset = number << 1;
1538    
1539  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 1508  for (;;) Line 1541  for (;;)
1541        printf("\n");        printf("\n");
1542  #endif  #endif
1543    
1544      md->capture_last = number;      md->capture_last = (md->capture_last & OVFLMASK) | number;
1545      if (offset >= md->offset_max) md->offset_overflow = TRUE; else      if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1546        {        {
1547        md->offset_vector[offset] =        md->offset_vector[offset] =
1548          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
# Line 1579  for (;;) Line 1612  for (;;)
1612          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1613          break;          break;
1614          }          }
1615          md->mark = save_mark;
1616    
1617        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* A COMMIT failure must fail the entire assertion, without trying any
1618        as NOMATCH. */        subsequent branches. */
1619    
1620          if (rrc == MATCH_COMMIT) RRETURN(MATCH_NOMATCH);
1621    
1622          /* PCRE does not allow THEN to escape beyond an assertion; it
1623          is treated as NOMATCH. */
1624    
1625        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1626        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
       md->mark = save_mark;  
1627        }        }
1628      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1629    
# Line 1685  for (;;) Line 1723  for (;;)
1723        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1724        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1725        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1726  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1727        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1728  #else  #elif defined COMPILE_PCRE16
1729        cb.subject          = (PCRE_SPTR16)md->start_subject;        cb.subject          = (PCRE_SPTR16)md->start_subject;
1730    #elif defined COMPILE_PCRE32
1731          cb.subject          = (PCRE_SPTR32)md->start_subject;
1732  #endif  #endif
1733        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1734        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
# Line 1696  for (;;) Line 1736  for (;;)
1736        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1737        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1738        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1739        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1740          /* Internal change requires this for API compatibility. */
1741          if (cb.capture_last == 0) cb.capture_last = -1;
1742        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1743        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1744        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
# Line 1725  for (;;) Line 1767  for (;;)
1767      case OP_RECURSE:      case OP_RECURSE:
1768        {        {
1769        recursion_info *ri;        recursion_info *ri;
1770        int recno;        unsigned int recno;
1771    
1772        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1773        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1742  for (;;) Line 1784  for (;;)
1784        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1785    
1786        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1787          new_recursive.saved_capture_last = md->capture_last;
1788        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1789        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1790        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1765  for (;;) Line 1808  for (;;)
1808              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1809    
1810        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1811        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1812        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1813          */
1814    
1815        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1816        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
# Line 1777  for (;;) Line 1821  for (;;)
1821            md, eptrb, RM6);            md, eptrb, RM6);
1822          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1823              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1824            md->capture_last = new_recursive.saved_capture_last;
1825          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1826          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1827            {            {
# Line 1793  for (;;) Line 1838  for (;;)
1838            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1839            }            }
1840    
1841          /* PCRE does not allow THEN to escape beyond a recursion; it is treated          /* PCRE does not allow THEN or COMMIT to escape beyond a recursion; it
1842          as NOMATCH. */          is treated as NOMATCH. */
1843    
1844          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN &&
1845                     rrc != MATCH_COMMIT)
1846            {            {
1847            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1848            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
# Line 1926  for (;;) Line 1972  for (;;)
1972    
1973        /* Deal with capturing */        /* Deal with capturing */
1974    
1975        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
1976        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1977          {          {
1978          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
1979          "skipping" a capturing group, and that group's offsets must be marked          "skipping" a capturing group, and that group's offsets must be marked
# Line 2007  for (;;) Line 2053  for (;;)
2053          }          }
2054        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2055          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2056          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2057          RRETURN(rrc);          RRETURN(rrc);
2058          }          }
# Line 2016  for (;;) Line 2061  for (;;)
2061        }        }
2062      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2063        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2064        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2065        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2066        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2073  for (;;) Line 2117  for (;;)
2117    
2118      case OP_DOLLM:      case OP_DOLLM:
2119      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2120        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2121          if (!IS_NEWLINE(eptr))
2122            {
2123            if (md->partial != 0 &&
2124                eptr + 1 >= md->end_subject &&
2125                NLBLOCK->nltype == NLTYPE_FIXED &&
2126                NLBLOCK->nllen == 2 &&
2127                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2128              {
2129              md->hitend = TRUE;
2130              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2131              }
2132            RRETURN(MATCH_NOMATCH);
2133            }
2134          }
2135      else      else
2136        {        {
2137        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2105  for (;;) Line 2163  for (;;)
2163      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2164      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2165          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2166          {
2167          if (md->partial != 0 &&
2168              eptr + 1 >= md->end_subject &&
2169              NLBLOCK->nltype == NLTYPE_FIXED &&
2170              NLBLOCK->nllen == 2 &&
2171              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2172            {
2173            md->hitend = TRUE;
2174            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2175            }
2176        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2177          }
2178    
2179      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2180    
# Line 2233  for (;;) Line 2302  for (;;)
2302        }        }
2303      break;      break;
2304    
2305      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2306        CRLF newlines and partial matching. */
2307    
2308      case OP_ANY:      case OP_ANY:
2309      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2310        if (md->partial != 0 &&
2311            eptr + 1 >= md->end_subject &&
2312            NLBLOCK->nltype == NLTYPE_FIXED &&
2313            NLBLOCK->nllen == 2 &&
2314            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2315          {
2316          md->hitend = TRUE;
2317          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2318          }
2319    
2320      /* Fall through */      /* Fall through */
2321    
2322        /* Match any single character whatsoever. */
2323    
2324      case OP_ALLANY:      case OP_ALLANY:
2325      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2326        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2378  for (;;) Line 2460  for (;;)
2460        {        {
2461        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2462    
2463        case 0x000d:        case CHAR_CR:
2464        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2465            {
2466            SCHECK_PARTIAL();
2467            }
2468          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2469        break;        break;
2470    
2471        case 0x000a:        case CHAR_LF:
2472        break;        break;
2473    
2474        case 0x000b:        case CHAR_VT:
2475        case 0x000c:        case CHAR_FF:
2476        case 0x0085:        case CHAR_NEL:
2477    #ifndef EBCDIC
2478        case 0x2028:        case 0x2028:
2479        case 0x2029:        case 0x2029:
2480    #endif  /* Not EBCDIC */
2481        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2482        break;        break;
2483        }        }
# Line 2405  for (;;) Line 2493  for (;;)
2493      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2494      switch(c)      switch(c)
2495        {        {
2496          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2497        default: break;        default: break;
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       RRETURN(MATCH_NOMATCH);  
2498        }        }
2499      ecode++;      ecode++;
2500      break;      break;
# Line 2439  for (;;) Line 2508  for (;;)
2508      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2509      switch(c)      switch(c)
2510        {        {
2511          HSPACE_CASES: break;  /* Byte and multibyte cases */
2512        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       break;  
2513        }        }
2514      ecode++;      ecode++;
2515      break;      break;
# Line 2473  for (;;) Line 2523  for (;;)
2523      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2524      switch(c)      switch(c)
2525        {        {
2526          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2527        default: break;        default: break;
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       RRETURN(MATCH_NOMATCH);  
2528        }        }
2529      ecode++;      ecode++;
2530      break;      break;
# Line 2495  for (;;) Line 2538  for (;;)
2538      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2539      switch(c)      switch(c)
2540        {        {
2541          VSPACE_CASES: break;
2542        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       break;  
2543        }        }
2544      ecode++;      ecode++;
2545      break;      break;
# Line 2521  for (;;) Line 2557  for (;;)
2557        }        }
2558      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2559        {        {
2560          const pcre_uint32 *cp;
2561        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2562    
2563        switch(ecode[1])        switch(ecode[1])
# Line 2581  for (;;) Line 2618  for (;;)
2618            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2619          break;          break;
2620    
2621            case PT_CLIST:
2622            cp = PRIV(ucd_caseless_sets) + ecode[2];
2623            for (;;)
2624              {
2625              if (c < *cp)
2626                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2627              if (c == *cp++)
2628                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2629              }
2630            break;
2631    
2632          /* This should never occur */          /* This should never occur */
2633    
2634          default:          default:
# Line 2600  for (;;) Line 2648  for (;;)
2648        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2649        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2650        }        }
2651      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
       {  
       int len = 1;  
       if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }  
       if (UCD_CATEGORY(c) != ucp_M) break;  
       eptr += len;  
       }  
     if (md->partial != 0 && eptr >= md->end_subject)  
2652        {        {
2653        SCHECK_PARTIAL();        int lgb, rgb;
2654          GETCHARINCTEST(c, eptr);
2655          lgb = UCD_GRAPHBREAK(c);
2656          while (eptr < md->end_subject)
2657            {
2658            int len = 1;
2659            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2660            rgb = UCD_GRAPHBREAK(c);
2661            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2662            lgb = rgb;
2663            eptr += len;
2664            }
2665        }        }
2666        CHECK_PARTIAL();
2667      ecode++;      ecode++;
2668      break;      break;
2669  #endif  #endif  /* SUPPORT_UCP */
2670    
2671    
2672      /* Match a back reference, possibly repeatedly. Look past the end of the      /* Match a back reference, possibly repeatedly. Look past the end of the
# Line 2678  for (;;) Line 2729  for (;;)
2729        default:               /* No repeat follows */        default:               /* No repeat follows */
2730        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2731          {          {
2732          eptr += -(length + 1);          if (length == -2) eptr = md->end_subject;   /* Partial match */
2733          CHECK_PARTIAL();          CHECK_PARTIAL();
2734          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2735          }          }
# Line 2704  for (;;) Line 2755  for (;;)
2755        int slength;        int slength;
2756        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2757          {          {
2758          eptr += -(slength + 1);          if (slength == -2) eptr = md->end_subject;   /* Partial match */
2759          CHECK_PARTIAL();          CHECK_PARTIAL();
2760          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2761          }          }
# Line 2728  for (;;) Line 2779  for (;;)
2779          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2780          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2781            {            {
2782            eptr += -(slength + 1);            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2783            CHECK_PARTIAL();            CHECK_PARTIAL();
2784            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2785            }            }
# Line 2747  for (;;) Line 2798  for (;;)
2798          int slength;          int slength;
2799          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2800            {            {
2801            /* Restore the eptr after the check. */            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2802            eptr += -(slength + 1);            the soft partial matching case. */
2803            CHECK_PARTIAL();  
2804            eptr -= -(slength + 1);            if (slength == -2 && md->partial != 0 &&
2805                  md->end_subject > md->start_used_ptr)
2806                {
2807                md->hitend = TRUE;
2808                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2809                }
2810            break;            break;
2811            }            }
2812          eptr += slength;          eptr += slength;
2813          }          }
2814    
2815        while (eptr >= pp)        while (eptr >= pp)
2816          {          {
2817          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 3118  for (;;) Line 3175  for (;;)
3175          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3176          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3177          }          }
3178        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3179        }        }
3180      else      else
3181  #endif  #endif
# Line 3158  for (;;) Line 3215  for (;;)
3215    
3216        if (fc < 128)        if (fc < 128)
3217          {          {
3218          if (md->lcc[fc]          pcre_uchar cc = RAWUCHAR(eptr);
3219              != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);          if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3220          ecode++;          ecode++;
3221          eptr++;          eptr++;
3222          }          }
# Line 3170  for (;;) Line 3227  for (;;)
3227    
3228        else        else
3229          {          {
3230          unsigned int dc;          pcre_uint32 dc;
3231          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3232          ecode += length;          ecode += length;
3233    
# Line 3280  for (;;) Line 3337  for (;;)
3337        if (length > 1)        if (length > 1)
3338          {          {
3339  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3340          unsigned int othercase;          pcre_uint32 othercase;
3341          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3342              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3343            oclength = PRIV(ord2utf)(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
# Line 3384  for (;;) Line 3441  for (;;)
3441      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3442    
3443      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3444        max, eptr));        max, (char *)eptr));
3445    
3446      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3447        {        {
# Line 3407  for (;;) Line 3464  for (;;)
3464    
3465        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3466          {          {
3467            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3468          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3469            {            {
3470            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3471            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3472            }            }
3473          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3474            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3475          eptr++;          eptr++;
3476          }          }
3477        if (min == max) continue;        if (min == max) continue;
# Line 3420  for (;;) Line 3479  for (;;)
3479          {          {
3480          for (fi = min;; fi++)          for (fi = min;; fi++)
3481            {            {
3482              pcre_uint32 cc;               /* Faster than pcre_uchar */
3483            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3484            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3485            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3428  for (;;) Line 3488  for (;;)
3488              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3489              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3490              }              }
3491            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3492              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3493            eptr++;            eptr++;
3494            }            }
3495          /* Control never gets here */          /* Control never gets here */
# Line 3438  for (;;) Line 3499  for (;;)
3499          pp = eptr;          pp = eptr;
3500          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3501            {            {
3502              pcre_uint32 cc;               /* Faster than pcre_uchar */
3503            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3504              {              {
3505              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3506              break;              break;
3507              }              }
3508            if (fc != *eptr && foc != *eptr) break;            cc = RAWUCHARTEST(eptr);
3509              if (fc != cc && foc != cc) break;
3510            eptr++;            eptr++;
3511            }            }
3512    
# Line 3471  for (;;) Line 3534  for (;;)
3534            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3535            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3536            }            }
3537          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3538          }          }
3539    
3540        if (min == max) continue;        if (min == max) continue;
# Line 3488  for (;;) Line 3551  for (;;)
3551              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3552              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3553              }              }
3554            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3555            }            }
3556          /* Control never gets here */          /* Control never gets here */
3557          }          }
# Line 3502  for (;;) Line 3565  for (;;)
3565              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3566              break;              break;
3567              }              }
3568            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3569            eptr++;            eptr++;
3570            }            }
3571          if (possessive) continue;          if (possessive) continue;
# Line 3528  for (;;) Line 3591  for (;;)
3591        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3592        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3593        }        }
     ecode++;  
     GETCHARINCTEST(c, eptr);  
     if (op == OP_NOTI)         /* The caseless case */  
       {  
       register unsigned int ch, och;  
       ch = *ecode++;  
 #ifdef COMPILE_PCRE8  
       /* ch must be < 128 if UTF is enabled. */  
       och = md->fcc[ch];  
 #else  
3594  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3595        if (utf)
3596          {
3597          register pcre_uint32 ch, och;
3598    
3599          ecode++;
3600          GETCHARINC(ch, ecode);
3601          GETCHARINC(c, eptr);
3602    
3603          if (op == OP_NOT)
3604            {
3605            if (ch == c) RRETURN(MATCH_NOMATCH);
3606            }
3607          else
3608            {
3609  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3610        if (utf && ch > 127)          if (ch > 127)
3611          och = UCD_OTHERCASE(ch);            och = UCD_OTHERCASE(ch);
3612  #else  #else
3613        if (utf && ch > 127)          if (ch > 127)
3614          och = ch;            och = ch;
3615  #endif /* SUPPORT_UCP */  #endif /* SUPPORT_UCP */
3616        else          else
3617  #endif /* SUPPORT_UTF */            och = TABLE_GET(ch, md->fcc, ch);
3618          och = TABLE_GET(ch, md->fcc, ch);          if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3619  #endif /* COMPILE_PCRE8 */          }
       if (ch == c || och == c) RRETURN(MATCH_NOMATCH);  
3620        }        }
3621      else    /* Caseful */      else
3622    #endif
3623        {        {
3624        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3625          c = *eptr++;
3626          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3627            RRETURN(MATCH_NOMATCH);
3628          ecode += 2;
3629        }        }
3630      break;      break;
3631    
# Line 3634  for (;;) Line 3705  for (;;)
3705      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3706    
3707      REPEATNOTCHAR:      REPEATNOTCHAR:
3708      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3709    
3710      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3711      since matching characters is likely to be quite common. First, ensure the      since matching characters is likely to be quite common. First, ensure the
# Line 3645  for (;;) Line 3716  for (;;)
3716      characters and work backwards. */      characters and work backwards. */
3717    
3718      DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3719        max, eptr));        max, (char *)eptr));
3720    
3721      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3722        {        {
 #ifdef COMPILE_PCRE8  
       /* fc must be < 128 if UTF is enabled. */  
       foc = md->fcc[fc];  
 #else  
3723  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3724  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3725        if (utf && fc > 127)        if (utf && fc > 127)
# Line 3664  for (;;) Line 3731  for (;;)
3731        else        else
3732  #endif /* SUPPORT_UTF */  #endif /* SUPPORT_UTF */
3733          foc = TABLE_GET(fc, md->fcc, fc);          foc = TABLE_GET(fc, md->fcc, fc);
 #endif /* COMPILE_PCRE8 */  
3734    
3735  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3736        if (utf)        if (utf)
3737          {          {
3738          register unsigned int d;          register pcre_uint32 d;
3739          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3740            {            {
3741            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3678  for (;;) Line 3744  for (;;)
3744              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3745              }              }
3746            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3747            if (fc == d || (unsigned int) foc == d) RRETURN(MATCH_NOMATCH);            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
3748            }            }
3749          }          }
3750        else        else
# Line 3704  for (;;) Line 3770  for (;;)
3770  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3771          if (utf)          if (utf)
3772            {            {
3773            register unsigned int d;            register pcre_uint32 d;
3774            for (fi = min;; fi++)            for (fi = min;; fi++)
3775              {              {
3776              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3749  for (;;) Line 3815  for (;;)
3815  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3816          if (utf)          if (utf)
3817            {            {
3818            register unsigned int d;            register pcre_uint32 d;
3819            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3820              {              {
3821              int len = 1;              int len = 1;
# Line 3806  for (;;) Line 3872  for (;;)
3872  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3873        if (utf)        if (utf)
3874          {          {
3875          register unsigned int d;          register pcre_uint32 d;
3876          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3877            {            {
3878            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3840  for (;;) Line 3906  for (;;)
3906  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3907          if (utf)          if (utf)
3908            {            {
3909            register unsigned int d;            register pcre_uint32 d;
3910            for (fi = min;; fi++)            for (fi = min;; fi++)
3911              {              {
3912              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3884  for (;;) Line 3950  for (;;)
3950  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3951          if (utf)          if (utf)
3952            {            {
3953            register unsigned int d;            register pcre_uint32 d;
3954            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3955              {              {
3956              int len = 1;              int len = 1;
# Line 4160  for (;;) Line 4226  for (;;)
4226              }              }
4227            break;            break;
4228    
4229              case PT_CLIST:
4230              for (i = 1; i <= min; i++)
4231                {
4232                const pcre_uint32 *cp;
4233                if (eptr >= md->end_subject)
4234                  {
4235                  SCHECK_PARTIAL();
4236                  RRETURN(MATCH_NOMATCH);
4237                  }
4238                GETCHARINCTEST(c, eptr);
4239                cp = PRIV(ucd_caseless_sets) + prop_value;
4240                for (;;)
4241                  {
4242                  if (c < *cp)
4243                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4244                  if (c == *cp++)
4245                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4246                  }
4247                }
4248              break;
4249    
4250            /* This should not occur */            /* This should not occur */
4251    
4252            default:            default:
# Line 4179  for (;;) Line 4266  for (;;)
4266              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4267              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4268              }              }
4269            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4270              {              {
4271              int len = 1;              int lgb, rgb;
4272              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4273              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4274              eptr += len;             while (eptr < md->end_subject)
4275                  {
4276                  int len = 1;
4277                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4278                  rgb = UCD_GRAPHBREAK(c);
4279                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4280                  lgb = rgb;
4281                  eptr += len;
4282                  }
4283              }              }
4284            }            CHECK_PARTIAL();
         if (md->partial != 0 && eptr >= md->end_subject)  
           {  
           SCHECK_PARTIAL();  
4285            }            }
4286          }          }
4287    
# Line 4212  for (;;) Line 4302  for (;;)
4302              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4303              }              }
4304            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4305              if (md->partial != 0 &&
4306                  eptr + 1 >= md->end_subject &&
4307                  NLBLOCK->nltype == NLTYPE_FIXED &&
4308                  NLBLOCK->nllen == 2 &&
4309                  RAWUCHAR(eptr) == NLBLOCK->nl[0])
4310                {
4311                md->hitend = TRUE;
4312                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4313                }
4314            eptr++;            eptr++;
4315            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4316            }            }
# Line 4248  for (;;) Line 4347  for (;;)
4347              {              {
4348              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4349    
4350              case 0x000d:              case CHAR_CR:
4351              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
4352              break;              break;
4353    
4354              case 0x000a:              case CHAR_LF:
4355              break;              break;
4356    
4357              case 0x000b:              case CHAR_VT:
4358              case 0x000c:              case CHAR_FF:
4359              case 0x0085:              case CHAR_NEL:
4360    #ifndef EBCDIC
4361              case 0x2028:              case 0x2028:
4362              case 0x2029:              case 0x2029:
4363    #endif  /* Not EBCDIC */
4364              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4365              break;              break;
4366              }              }
# Line 4277  for (;;) Line 4378  for (;;)
4378            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4379            switch(c)            switch(c)
4380              {              {
4381                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4382              default: break;              default: break;
             case 0x09:      /* HT */  
             case 0x20:      /* SPACE */  
             case 0xa0:      /* NBSP */  
             case 0x1680:    /* OGHAM SPACE MARK */  
             case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
             case 0x2000:    /* EN QUAD */  
             case 0x2001:    /* EM QUAD */  
             case 0x2002:    /* EN SPACE */  
             case 0x2003:    /* EM SPACE */  
             case 0x2004:    /* THREE-PER-EM SPACE */  
             case 0x2005:    /* FOUR-PER-EM SPACE */  
             case 0x2006:    /* SIX-PER-EM SPACE */  
             case 0x2007:    /* FIGURE SPACE */  
             case 0x2008:    /* PUNCTUATION SPACE */  
             case 0x2009:    /* THIN SPACE */  
             case 0x200A:    /* HAIR SPACE */  
             case 0x202f:    /* NARROW NO-BREAK SPACE */  
             case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
             case 0x3000:    /* IDEOGRAPHIC SPACE */  
             RRETURN(MATCH_NOMATCH);  
4383              }              }
4384            }            }
4385          break;          break;
# Line 4313  for (;;) Line 4395  for (;;)
4395            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4396            switch(c)            switch(c)
4397              {              {
4398                HSPACE_CASES: break;  /* Byte and multibyte cases */
4399              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
             case 0x09:      /* HT */  
             case 0x20:      /* SPACE */  
             case 0xa0:      /* NBSP */  
             case 0x1680:    /* OGHAM SPACE MARK */  
             case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
             case 0x2000:    /* EN QUAD */  
             case 0x2001:    /* EM QUAD */  
             case 0x2002:    /* EN SPACE */  
             case 0x2003:    /* EM SPACE */  
             case 0x2004:    /* THREE-PER-EM SPACE */  
             case 0x2005:    /* FOUR-PER-EM SPACE */  
             case 0x2006:    /* SIX-PER-EM SPACE */  
             case 0x2007:    /* FIGURE SPACE */  
             case 0x2008:    /* PUNCTUATION SPACE */  
             case 0x2009:    /* THIN SPACE */  
             case 0x200A:    /* HAIR SPACE */  
             case 0x202f:    /* NARROW NO-BREAK SPACE */  
             case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
             case 0x3000:    /* IDEOGRAPHIC SPACE */  
             break;  
4400              }              }
4401            }            }
4402          break;          break;
# Line 4349  for (;;) Line 4412  for (;;)
4412            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4413            switch(c)            switch(c)
4414              {              {
4415                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4416              default: break;              default: break;
             case 0x0a:      /* LF */  
             case 0x0b:      /* VT */  
             case 0x0c:      /* FF */  
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
             RRETURN(MATCH_NOMATCH);  
4417              }              }
4418            }            }
4419          break;          break;
# Line 4373  for (;;) Line 4429  for (;;)
4429            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4430            switch(c)            switch(c)
4431              {              {
4432                VSPACE_CASES: break;
4433              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
             case 0x0a:      /* LF */  
             case 0x0b:      /* VT */  
             case 0x0c:      /* FF */  
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
             break;  
4434              }              }
4435            }            }
4436          break;          break;
# Line 4403  for (;;) Line 4452  for (;;)
4452          case OP_DIGIT:          case OP_DIGIT:
4453          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4454            {            {
4455              pcre_uchar cc;
4456    
4457            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4458              {              {
4459              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4460              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4461              }              }
4462            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)            cc = RAWUCHAR(eptr);
4463              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4464              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4465            eptr++;            eptr++;
4466            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
# Line 4418  for (;;) Line 4470  for (;;)
4470          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4471          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4472            {            {
4473              pcre_uchar cc;
4474    
4475            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4476              {              {
4477              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4478              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4479              }              }
4480            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = RAWUCHAR(eptr);
4481              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4482              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4483            eptr++;            eptr++;
4484            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4433  for (;;) Line 4488  for (;;)
4488          case OP_WHITESPACE:          case OP_WHITESPACE:
4489          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4490            {            {
4491              pcre_uchar cc;
4492    
4493            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4494              {              {
4495              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4496              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4497              }              }
4498            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0)            cc = RAWUCHAR(eptr);
4499              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4500              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4501            eptr++;            eptr++;
4502            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
# Line 4448  for (;;) Line 4506  for (;;)
4506          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4507          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4508            {            {
4509              pcre_uchar cc;
4510    
4511            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4512              {              {
4513              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4514              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4515              }              }
4516            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = RAWUCHAR(eptr);
4517              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4518              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4519            eptr++;            eptr++;
4520            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4463  for (;;) Line 4524  for (;;)
4524          case OP_WORDCHAR:          case OP_WORDCHAR:
4525          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4526            {            {
4527              pcre_uchar cc;
4528    
4529            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4530              {              {
4531              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4532              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4533              }              }
4534            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0)            cc = RAWUCHAR(eptr);
4535              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4536              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4537            eptr++;            eptr++;
4538            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
# Line 4496  for (;;) Line 4560  for (;;)
4560              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4561              }              }
4562            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4563              if (md->partial != 0 &&
4564                  eptr + 1 >= md->end_subject &&
4565                  NLBLOCK->nltype == NLTYPE_FIXED &&
4566                  NLBLOCK->nllen == 2 &&
4567                  *eptr == NLBLOCK->nl[0])
4568                {
4569                md->hitend = TRUE;
4570                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4571                }
4572            eptr++;            eptr++;
4573            }            }
4574          break;          break;
# Line 4530  for (;;) Line 4603  for (;;)
4603              {              {
4604              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4605    
4606              case 0x000d:              case CHAR_CR:
4607              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4608              break;              break;
4609    
4610              case 0x000a:              case CHAR_LF:
4611              break;              break;
4612    
4613              case 0x000b:              case CHAR_VT:
4614              case 0x000c:              case CHAR_FF:
4615              case 0x0085:              case CHAR_NEL:
4616  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4617              case 0x2028:              case 0x2028:
4618              case 0x2029:              case 0x2029:
4619  #endif  #endif
# Line 4561  for (;;) Line 4634  for (;;)
4634            switch(*eptr++)            switch(*eptr++)
4635              {              {
4636              default: break;              default: break;
4637              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4638              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4639              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
 #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 */  
4640  #endif  #endif
4641              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4642              }              }
# Line 4598  for (;;) Line 4654  for (;;)
4654            switch(*eptr++)            switch(*eptr++)
4655              {              {
4656              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4657              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4658              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4659              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
 #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 */  
4660  #endif  #endif
4661              break;              break;
4662              }              }
# Line 4634  for (;;) Line 4673  for (;;)
4673              }              }
4674            switch(*eptr++)            switch(*eptr++)
4675              {              {
4676              default: break;              VSPACE_BYTE_CASES:
4677              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4678              case 0x0b:      /* VT */              VSPACE_MULTIBYTE_CASES:
             case 0x0c:      /* FF */  
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
4679  #endif  #endif
4680              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4681                default: break;
4682              }              }
4683            }            }
4684          break;          break;
# Line 4660  for (;;) Line 4694  for (;;)
4694            switch(*eptr++)            switch(*eptr++)
4695              {              {
4696              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4697              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4698              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4699              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
4700  #endif  #endif
4701              break;              break;
4702              }              }
# Line 4945  for (;;) Line 4974  for (;;)
4974              }              }
4975            /* Control never gets here */            /* Control never gets here */
4976    
4977            /* This should never occur */            case PT_CLIST:
4978              for (fi = min;; fi++)
4979                {
4980                const pcre_uint32 *cp;
4981                RMATCH(eptr, ecode, offset_top, md, eptrb, RM67);
4982                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4983                if (fi >= max) RRETURN(MATCH_NOMATCH);
4984                if (eptr >= md->end_subject)
4985                  {
4986                  SCHECK_PARTIAL();
4987                  RRETURN(MATCH_NOMATCH);
4988                  }
4989                GETCHARINCTEST(c, eptr);
4990                cp = PRIV(ucd_caseless_sets) + prop_value;
4991                for (;;)
4992                  {
4993                  if (c < *cp)
4994                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4995                  if (c == *cp++)
4996                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4997                  }
4998                }
4999              /* Control never gets here */
5000    
5001              /* This should never occur */
5002            default:            default:
5003            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5004            }            }
# Line 4967  for (;;) Line 5019  for (;;)
5019              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5020              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5021              }              }
5022            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
             {  
             int len = 1;  
             if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }  
             if (UCD_CATEGORY(c) != ucp_M) break;  
             eptr += len;  
             }  
           if (md->partial != 0 && eptr >= md->end_subject)  
5023              {              {
5024              SCHECK_PARTIAL();              int lgb, rgb;
5025                GETCHARINCTEST(c, eptr);
5026                lgb = UCD_GRAPHBREAK(c);
5027                while (eptr < md->end_subject)
5028                  {
5029                  int len = 1;
5030                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5031                  rgb = UCD_GRAPHBREAK(c);
5032                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5033                  lgb = rgb;
5034                  eptr += len;
5035                  }
5036              }              }
5037              CHECK_PARTIAL();
5038            }            }
5039          }          }
5040        else        else
# Line 5003  for (;;) Line 5058  for (;;)
5058            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5059            switch(ctype)            switch(ctype)
5060              {              {
5061              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5062                if (md->partial != 0 &&    /* Take care with CRLF partial */
5063                    eptr >= md->end_subject &&
5064                    NLBLOCK->nltype == NLTYPE_FIXED &&
5065                    NLBLOCK->nllen == 2 &&
5066                    c == NLBLOCK->nl[0])
5067                  {
5068                  md->hitend = TRUE;
5069                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5070                  }
5071                break;
5072    
5073              case OP_ALLANY:              case OP_ALLANY:
5074              case OP_ANYBYTE:              case OP_ANYBYTE:
5075              break;              break;
# Line 5012  for (;;) Line 5078  for (;;)
5078              switch(c)              switch(c)
5079                {                {
5080                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5081                case 0x000d:                case CHAR_CR:
5082                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
5083                break;                break;
5084                case 0x000a:  
5085                  case CHAR_LF:
5086                break;                break;
5087    
5088                case 0x000b:                case CHAR_VT:
5089                case 0x000c:                case CHAR_FF:
5090                case 0x0085:                case CHAR_NEL:
5091    #ifndef EBCDIC
5092                case 0x2028:                case 0x2028:
5093                case 0x2029:                case 0x2029:
5094    #endif  /* Not EBCDIC */
5095                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5096                break;                break;
5097                }                }
# Line 5031  for (;;) Line 5100  for (;;)
5100              case OP_NOT_HSPACE:              case OP_NOT_HSPACE:
5101              switch(c)              switch(c)
5102                {                {
5103                  HSPACE_CASES: RRETURN(MATCH_NOMATCH);
5104                default: break;                default: break;
               case 0x09:      /* HT */  
               case 0x20:      /* SPACE */  
               case 0xa0:      /* NBSP */  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
               RRETURN(MATCH_NOMATCH);  
5105                }                }
5106              break;              break;
5107    
5108              case OP_HSPACE:              case OP_HSPACE:
5109              switch(c)              switch(c)
5110                {                {
5111                  HSPACE_CASES: break;
5112                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
               case 0x09:      /* HT */  
               case 0x20:      /* SPACE */  
               case 0xa0:      /* NBSP */  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
               break;  
5113                }                }
5114              break;              break;
5115    
5116              case OP_NOT_VSPACE:              case OP_NOT_VSPACE:
5117              switch(c)              switch(c)
5118                {                {
5119                  VSPACE_CASES: RRETURN(MATCH_NOMATCH);
5120                default: break;                default: break;
               case 0x0a:      /* LF */  
               case 0x0b:      /* VT */  
               case 0x0c:      /* FF */  
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
               RRETURN(MATCH_NOMATCH);  
5121                }                }
5122              break;              break;
5123    
5124              case OP_VSPACE:              case OP_VSPACE:
5125              switch(c)              switch(c)
5126                {                {
5127                  VSPACE_CASES: break;
5128                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
               case 0x0a:      /* LF */  
               case 0x0b:      /* VT */  
               case 0x0c:      /* FF */  
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
               break;  
5129                }                }
5130              break;              break;
5131    
# Line 5166  for (;;) Line 5183  for (;;)
5183            c = *eptr++;            c = *eptr++;
5184            switch(ctype)            switch(ctype)
5185              {              {
5186              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5187                if (md->partial != 0 &&    /* Take care with CRLF partial */
5188                    eptr >= md->end_subject &&
5189                    NLBLOCK->nltype == NLTYPE_FIXED &&
5190                    NLBLOCK->nllen == 2 &&
5191                    c == NLBLOCK->nl[0])
5192                  {
5193                  md->hitend = TRUE;
5194                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5195                  }
5196                break;
5197    
5198              case OP_ALLANY:              case OP_ALLANY:
5199              case OP_ANYBYTE:              case OP_ANYBYTE:
5200              break;              break;
# Line 5175  for (;;) Line 5203  for (;;)
5203              switch(c)              switch(c)
5204                {                {
5205                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5206                case 0x000d:                case CHAR_CR:
5207                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5208                break;                break;
5209    
5210                case 0x000a:                case CHAR_LF:
5211                break;                break;
5212    
5213                case 0x000b:                case CHAR_VT:
5214                case 0x000c:                case CHAR_FF:
5215                case 0x0085:                case CHAR_NEL:
5216  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5217                case 0x2028:                case 0x2028:
5218                case 0x2029:                case 0x2029:
5219  #endif  #endif
# Line 5198  for (;;) Line 5226  for (;;)
5226              switch(c)              switch(c)
5227                {                {
5228                default: break;                default: break;
5229                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5230                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5231                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
 #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 */  
5232  #endif  #endif
5233                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5234                }                }
# Line 5227  for (;;) Line 5238  for (;;)
5238              switch(c)              switch(c)
5239                {                {
5240                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5241                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5242                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5243                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
 #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 */  
5244  #endif  #endif
5245                break;                break;
5246                }                }
# Line 5256  for (;;) Line 5250  for (;;)
5250              switch(c)              switch(c)
5251                {                {
5252                default: break;                default: break;
5253                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5254                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5255                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5256  #endif  #endif
5257                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5258                }                }
# Line 5273  for (;;) Line 5262  for (;;)
5262              switch(c)              switch(c)
5263                {                {
5264                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5265                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5266                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5267                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5268  #endif  #endif
5269                break;                break;
5270                }                }
# Line 5484  for (;;) Line 5468  for (;;)
5468              }              }
5469            break;            break;
5470    
5471              case PT_CLIST:
5472              for (i = min; i < max; i++)
5473                {
5474                const pcre_uint32 *cp;
5475                int len = 1;
5476                if (eptr >= md->end_subject)
5477                  {
5478                  SCHECK_PARTIAL();
5479                  break;
5480                  }
5481                GETCHARLENTEST(c, eptr, len);
5482                cp = PRIV(ucd_caseless_sets) + prop_value;
5483                for (;;)
5484                  {
5485                  if (c < *cp)
5486                    { if (prop_fail_result) break; else goto GOT_MAX; }
5487                  if (c == *cp++)
5488                    { if (prop_fail_result) goto GOT_MAX; else break; }
5489                  }
5490                eptr += len;
5491                }
5492              GOT_MAX:
5493              break;
5494    
5495            default:            default:
5496            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5497            }            }
# Line 5507  for (;;) Line 5515  for (;;)
5515          {          {
5516          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5517            {            {
           int len = 1;  
5518            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5519              {              {
5520              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5521              break;              break;
5522              }              }
5523            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
           if (UCD_CATEGORY(c) == ucp_M) break;  
           eptr += len;  
           while (eptr < md->end_subject)  
             {  
             len = 1;  
             if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }  
             if (UCD_CATEGORY(c) != ucp_M) break;  
             eptr += len;  
             }  
           if (eptr >= md->end_subject)  
5524              {              {
5525              SCHECK_PARTIAL();              int lgb, rgb;
5526                GETCHARINCTEST(c, eptr);
5527                lgb = UCD_GRAPHBREAK(c);
5528                while (eptr < md->end_subject)
5529                  {
5530                  int len = 1;
5531                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5532                  rgb = UCD_GRAPHBREAK(c);
5533                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5534                  lgb = rgb;
5535                  eptr += len;
5536                  }
5537              }              }
5538              CHECK_PARTIAL();
5539            }            }
5540    
5541          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
# Line 5570  for (;;) Line 5579  for (;;)
5579                  break;                  break;
5580                  }                  }
5581                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5582                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5583                      eptr + 1 >= md->end_subject &&
5584                      NLBLOCK->nltype == NLTYPE_FIXED &&
5585                      NLBLOCK->nllen == 2 &&
5586                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5587                    {
5588                    md->hitend = TRUE;
5589                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5590                    }
5591                eptr++;                eptr++;
5592                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5593                }                }
# Line 5587  for (;;) Line 5605  for (;;)
5605                  break;                  break;
5606                  }                  }
5607                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5608                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5609                      eptr + 1 >= md->end_subject &&
5610                      NLBLOCK->nltype == NLTYPE_FIXED &&
5611                      NLBLOCK->nllen == 2 &&
5612                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5613                    {
5614                    md->hitend = TRUE;
5615                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5616                    }
5617                eptr++;                eptr++;
5618                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5619                }                }
# Line 5636  for (;;) Line 5663  for (;;)
5663                break;                break;
5664                }                }
5665              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5666              if (c == 0x000d)              if (c == CHAR_CR)
5667                {                {
5668                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5669                if (*eptr == 0x000a) eptr++;                if (RAWUCHAR(eptr) == CHAR_LF) eptr++;
5670                }                }
5671              else              else
5672                {                {
5673                if (c != 0x000a &&                if (c != CHAR_LF &&
5674                    (md->bsr_anycrlf ||                    (md->bsr_anycrlf ||
5675                     (c != 0x000b && c != 0x000c &&                     (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5676                      c != 0x0085 && c != 0x2028 && c != 0x2029)))  #ifndef EBCDIC
5677                        && c != 0x2028 && c != 0x2029
5678    #endif  /* Not EBCDIC */
5679                        )))
5680                  break;                  break;
5681                eptr += len;                eptr += len;
5682                }                }
# Line 5667  for (;;) Line 5697  for (;;)
5697              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5698              switch(c)              switch(c)
5699                {                {
5700                  HSPACE_CASES: gotspace = TRUE; break;
5701                default: gotspace = FALSE; break;                default: gotspace = FALSE; break;
               case 0x09:      /* HT */  
               case 0x20:      /* SPACE */  
               case 0xa0:      /* NBSP */  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
               gotspace = TRUE;  
               break;  
5702                }                }
5703              if (gotspace == (ctype == OP_NOT_HSPACE)) break;              if (gotspace == (ctype == OP_NOT_HSPACE)) break;
5704              eptr += len;              eptr += len;
# Line 5709  for (;;) Line 5719  for (;;)
5719              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5720              switch(c)              switch(c)
5721                {                {
5722                  VSPACE_CASES: gotspace = TRUE; break;
5723                default: gotspace = FALSE; break;                default: gotspace = FALSE; break;
               case 0x0a:      /* LF */  
               case 0x0b:      /* VT */  
               case 0x0c:      /* FF */  
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
               gotspace = TRUE;  
               break;  
5724                }                }
5725              if (gotspace == (ctype == OP_NOT_VSPACE)) break;              if (gotspace == (ctype == OP_NOT_VSPACE)) break;
5726              eptr += len;              eptr += len;
# Line 5832  for (;;) Line 5834  for (;;)
5834            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5835            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5836            BACKCHAR(eptr);            BACKCHAR(eptr);
5837            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && RAWUCHAR(eptr) == CHAR_NL &&
5838                eptr[-1] == '\r') eptr--;                RAWUCHAR(eptr - 1) == CHAR_CR) eptr--;
5839            }            }
5840          }          }
5841        else        else
# Line 5851  for (;;) Line 5853  for (;;)
5853                break;                break;
5854                }                }
5855              if (IS_NEWLINE(eptr)) break;              if (IS_NEWLINE(eptr)) break;
5856                if (md->partial != 0 &&    /* Take care with CRLF partial */
5857                    eptr + 1 >= md->end_subject &&
5858                    NLBLOCK->nltype == NLTYPE_FIXED &&
5859                    NLBLOCK->nllen == 2 &&
5860                    *eptr == NLBLOCK->nl[0])
5861                  {
5862                  md->hitend = TRUE;
5863                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5864                  }
5865              eptr++;              eptr++;
5866              }              }
5867            break;            break;
# Line 5875  for (;;) Line 5886  for (;;)
5886                break;                break;
5887                }                }
5888              c = *eptr;              c = *eptr;
5889              if (c == 0x000d)              if (c == CHAR_CR)
5890                {                {
5891                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5892                if (*eptr == 0x000a) eptr++;                if (*eptr == CHAR_LF) eptr++;
5893                }                }
5894              else              else
5895                {                {
5896                if (c != 0x000a && (md->bsr_anycrlf ||                if (c != CHAR_LF && (md->bsr_anycrlf ||
5897                  (c != 0x000b && c != 0x000c && c != 0x0085                   (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5898  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5899                  && c != 0x2028 && c != 0x2029                   && c != 0x2028 && c != 0x2029
5900  #endif  #endif
5901                  ))) break;                   ))) break;
5902                eptr++;                eptr++;
5903                }                }
5904              }              }
# Line 5901  for (;;) Line 5912  for (;;)
5912                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5913                break;                break;
5914                }                }
5915              c = *eptr;              switch(*eptr)
5916              if (c == 0x09 || c == 0x20 || c == 0xa0                {
5917  #ifdef COMPILE_PCRE16                default: eptr++; break;
5918                || c == 0x1680 || c == 0x180e || (c >= 0x2000 && c <= 0x200A)                HSPACE_BYTE_CASES:
5919                || c == 0x202f || c == 0x205f || c == 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5920                  HSPACE_MULTIBYTE_CASES:
5921  #endif  #endif
5922                ) break;                goto ENDLOOP00;
5923              eptr++;                }
5924              }              }
5925              ENDLOOP00:
5926            break;            break;
5927    
5928            case OP_HSPACE:            case OP_HSPACE:
# Line 5920  for (;;) Line 5933  for (;;)
5933                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5934                break;                break;
5935                }                }
5936              c = *eptr;              switch(*eptr)
5937              if (c != 0x09 && c != 0x20 && c != 0xa0                {
5938  #ifdef COMPILE_PCRE16                default: goto ENDLOOP01;
5939                && c != 0x1680 && c != 0x180e && (c < 0x2000 || c > 0x200A)                HSPACE_BYTE_CASES:
5940                && c != 0x202f && c != 0x205f && c != 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5941                  HSPACE_MULTIBYTE_CASES:
5942  #endif  #endif
5943                ) break;                eptr++; break;
5944              eptr++;                }
5945              }              }
5946              ENDLOOP01:
5947            break;            break;
5948    
5949            case OP_NOT_VSPACE:            case OP_NOT_VSPACE:
# Line 5939  for (;;) Line 5954  for (;;)
5954                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5955                break;                break;
5956                }                }
5957              c = *eptr;              switch(*eptr)
5958              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85                {
5959  #ifdef COMPILE_PCRE16                default: eptr++; break;
5960                || c == 0x2028 || c == 0x2029                VSPACE_BYTE_CASES:
5961    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5962                  VSPACE_MULTIBYTE_CASES:
5963  #endif  #endif
5964                ) break;                goto ENDLOOP02;
5965              eptr++;                }
5966              }              }
5967              ENDLOOP02:
5968            break;            break;
5969    
5970            case OP_VSPACE:            case OP_VSPACE:
# Line 5957  for (;;) Line 5975  for (;;)
5975                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5976                break;                break;
5977                }                }
5978              c = *eptr;              switch(*eptr)
5979              if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85                {
5980  #ifdef COMPILE_PCRE16                default: goto ENDLOOP03;
5981                && c != 0x2028 && c != 0x2029                VSPACE_BYTE_CASES:
5982    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5983                  VSPACE_MULTIBYTE_CASES:
5984  #endif  #endif
5985                ) break;                eptr++; break;
5986              eptr++;                }
5987              }              }
5988              ENDLOOP03:
5989            break;            break;
5990    
5991            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
# Line 6061  for (;;) Line 6082  for (;;)
6082            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);
6083            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
6084            eptr--;            eptr--;
6085            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && *eptr == CHAR_LF &&
6086                eptr[-1] == '\r') eptr--;                eptr[-1] == CHAR_CR) eptr--;
6087            }            }
6088          }          }
6089    
# Line 6112  switch (frame->Xwhere) Line 6133  switch (frame->Xwhere)
6133    LBL(32) LBL(34) LBL(42) LBL(46)    LBL(32) LBL(34) LBL(42) LBL(46)
6134  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6135    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)
6136    LBL(59) LBL(60) LBL(61) LBL(62)    LBL(59) LBL(60) LBL(61) LBL(62) LBL(67)
6137  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
6138  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
6139    default:    default:
6140    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);  
   
6141    return PCRE_ERROR_INTERNAL;    return PCRE_ERROR_INTERNAL;
6142    }    }
6143  #undef LBL  #undef LBL
# Line 6181  Undefine all the macros that were define Line 6199  Undefine all the macros that were define
6199  ***************************************************************************/  ***************************************************************************/
6200    
6201    
6202    #ifdef NO_RECURSE
6203    /*************************************************
6204    *          Release allocated heap frames         *
6205    *************************************************/
6206    
6207    /* This function releases all the allocated frames. The base frame is on the
6208    machine stack, and so must not be freed.
6209    
6210    Argument: the address of the base frame
6211    Returns:  nothing
6212    */
6213    
6214    static void
6215    release_match_heapframes (heapframe *frame_base)
6216    {
6217    heapframe *nextframe = frame_base->Xnextframe;
6218    while (nextframe != NULL)
6219      {
6220      heapframe *oldframe = nextframe;
6221      nextframe = nextframe->Xnextframe;
6222      (PUBL(stack_free))(oldframe);
6223      }
6224    }
6225    #endif
6226    
6227    
6228  /*************************************************  /*************************************************
6229  *         Execute a Regular Expression           *  *         Execute a Regular Expression           *
# Line 6206  Returns:          > 0 => success; value Line 6249  Returns:          > 0 => success; value
6249                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
6250  */  */
6251    
6252  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
6253  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6254  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
6255    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
6256    int offsetcount)    int offsetcount)
6257  #else  #elif defined COMPILE_PCRE16
6258  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6259  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,
6260    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,
6261    int offsetcount)    int offsetcount)
6262    #elif defined COMPILE_PCRE32
6263    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6264    pcre32_exec(const pcre32 *argument_re, const pcre32_extra *extra_data,
6265      PCRE_SPTR32 subject, int length, int start_offset, int options, int *offsets,
6266      int offsetcount)
6267  #endif  #endif
6268  {  {
6269  int rc, ocount, arg_offset_max;  int rc, ocount, arg_offset_max;
# Line 6243  PCRE_PUCHAR req_char_ptr = start_match - Line 6291  PCRE_PUCHAR req_char_ptr = start_match -
6291  const pcre_study_data *study;  const pcre_study_data *study;
6292  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;
6293    
6294    #ifdef NO_RECURSE
6295    heapframe frame_zero;
6296    frame_zero.Xprevframe = NULL;            /* Marks the top level */
6297    frame_zero.Xnextframe = NULL;            /* None are allocated yet */
6298    md->match_frames_base = &frame_zero;
6299    #endif
6300    
6301  /* 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
6302  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
6303  compiler gave this error: "unary minus operator applied to unsigned type,  compiler gave this error: "unary minus operator applied to unsigned type,
6304  result still unsigned". Hopefully the cast fixes that. */  result still unsigned". Hopefully the cast fixes that. */
6305    
# Line 6262  if ((options & ~PUBLIC_EXEC_OPTIONS) != Line 6317  if ((options & ~PUBLIC_EXEC_OPTIONS) !=
6317  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))
6318    return PCRE_ERROR_NULL;    return PCRE_ERROR_NULL;
6319  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
6320    if (length < 0) return PCRE_ERROR_BADLENGTH;
6321  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
6322    
6323  /* 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 6299  if (utf && (options & PCRE_NO_UTF8_CHECK Line 6355  if (utf && (options & PCRE_NO_UTF8_CHECK
6355        offsets[0] = erroroffset;        offsets[0] = erroroffset;
6356        offsets[1] = errorcode;        offsets[1] = errorcode;
6357        }        }
6358  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE8
     return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?  
       PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;  
 #else  
6359      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?
6360        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
6361    #elif defined COMPILE_PCRE16
6362        return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?
6363          PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;
6364    #elif defined COMPILE_PCRE32
6365        return PCRE_ERROR_BADUTF32;
6366  #endif  #endif
6367      }      }
6368    #if defined COMPILE_PCRE8 || defined COMPILE_PCRE16
6369    /* 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. */
6370    if (start_offset > 0 && start_offset < length &&    if (start_offset > 0 && start_offset < length &&
6371        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))
6372      return PCRE_ERROR_BADUTF8_OFFSET;      return PCRE_ERROR_BADUTF8_OFFSET;
6373    #endif
6374    }    }
6375  #endif  #endif
6376    
6377  /* If the pattern was successfully studied with JIT support, run the JIT  /* If the pattern was successfully studied with JIT support, run the JIT
6378  executable instead of the rest of this function. Most options must be set at  executable instead of the rest of this function. Most options must be set at
6379  compile time for the JIT code to be usable. Fallback to the normal code path if  compile time for the JIT code to be usable. Fallback to the normal code path if
6380  an unsupported flag is set. In particular, JIT does not support partial  an unsupported flag is set. */
 matching. */  
6381    
6382  #ifdef SUPPORT_JIT  #ifdef SUPPORT_JIT
6383  if (extra_data != NULL  if (extra_data != NULL
6384      && (extra_data->flags & (PCRE_EXTRA_EXECUTABLE_JIT |      && (extra_data->flags & (PCRE_EXTRA_EXECUTABLE_JIT |
6385                               PCRE_EXTRA_TABLES)) == PCRE_EXTRA_EXECUTABLE_JIT                               PCRE_EXTRA_TABLES)) == PCRE_EXTRA_EXECUTABLE_JIT
6386      && extra_data->executable_jit != NULL      && extra_data->executable_jit != NULL
6387      && (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)  
6388    {    {
6389    rc = PRIV(jit_exec)(re, extra_data->executable_jit,    rc = PRIV(jit_exec)(extra_data, (const pcre_uchar *)subject, length,
6390      (const pcre_uchar *)subject, length, start_offset, options,         start_offset, options, offsets, offsetcount);
6391      ((extra_data->flags & PCRE_EXTRA_MATCH_LIMIT) == 0)  
     ? MATCH_LIMIT : extra_data->match_limit, offsets, offsetcount);  
6392    /* PCRE_ERROR_NULL means that the selected normal or partial matching    /* PCRE_ERROR_NULL means that the selected normal or partial matching
6393    mode is not compiled. In this case we simply fallback to interpreter. */    mode is not compiled. In this case we simply fallback to interpreter. */
6394    if (rc != PCRE_ERROR_NULL)  
6395      return rc;    if (rc != PCRE_ERROR_JIT_BADOPTION) return rc;
6396    }    }
6397  #endif  #endif
6398    
# Line 6509  if (re->top_backref > 0 && re->top_backr Line 6564  if (re->top_backref > 0 && re->top_backr
6564    DPRINTF(("Got memory to hold back references\n"));    DPRINTF(("Got memory to hold back references\n"));
6565    }    }
6566  else md->offset_vector = offsets;  else md->offset_vector = offsets;
   
6567  md->offset_end = ocount;  md->offset_end = ocount;
6568  md->offset_max = (2*ocount)/3;  md->offset_max = (2*ocount)/3;
6569  md->offset_overflow = FALSE;  md->capture_last = 0;
 md->capture_last = -1;  
6570    
6571  /* Reset the working variable associated with each extraction. These should  /* Reset the working variable associated with each extraction. These should
6572  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 6621  for(;;) Line 6674  for(;;)
6674    
6675      if (has_first_char)      if (has_first_char)
6676        {        {
6677          pcre_uchar smc;
6678    
6679        if (first_char != first_char2)        if (first_char != first_char2)
6680          while (start_match < end_subject &&          while (start_match < end_subject &&
6681              *start_match != first_char && *start_match != first_char2)            (smc = RAWUCHARTEST(start_match)) != first_char && smc != first_char2)
6682            start_match++;            start_match++;
6683        else        else
6684          while (start_match < end_subject && *start_match != first_char)          while (start_match < end_subject && RAWUCHARTEST(start_match) != first_char)
6685            start_match++;            start_match++;
6686        }        }
6687    
# Line 6658  for(;;) Line 6713  for(;;)
6713          if (start_match[-1] == CHAR_CR &&          if (start_match[-1] == CHAR_CR &&
6714               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
6715               start_match < end_subject &&               start_match < end_subject &&
6716               *start_match == CHAR_NL)               RAWUCHARTEST(start_match) == CHAR_NL)
6717            start_match++;            start_match++;
6718          }          }
6719        }        }
# Line 6669  for(;;) Line 6724  for(;;)
6724        {        {
6725        while (start_match < end_subject)        while (start_match < end_subject)
6726          {          {
6727          register unsigned int c = *start_match;          register pcre_uint32 c = RAWUCHARTEST(start_match);
6728  #ifndef COMPILE_PCRE8  #ifndef COMPILE_PCRE8
6729          if (c > 255) c = 255;          if (c > 255) c = 255;
6730  #endif  #endif
# Line 6737  for(;;) Line 6792  for(;;)
6792            {            {
6793            while (p < end_subject)            while (p < end_subject)
6794              {              {
6795              register int pp = *p++;              register pcre_uint32 pp = RAWUCHARINCTEST(p);
6796              if (pp == req_char || pp == req_char2) { p--; break; }              if (pp == req_char || pp == req_char2) { p--; break; }
6797              }              }
6798            }            }
# Line 6745  for(;;) Line 6800  for(;;)
6800            {            {
6801            while (p < end_subject)            while (p < end_subject)
6802              {              {
6803              if (*p++ == req_char) { p--; break; }              if (RAWUCHARINCTEST(p) == req_char) { p--; break; }
6804              }              }
6805            }            }
6806    
# Line 6908  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 6963  if (rc == MATCH_MATCH || rc == MATCH_ACC
6963          (arg_offset_max - 2) * sizeof(int));          (arg_offset_max - 2) * sizeof(int));
6964        DPRINTF(("Copied offsets from temporary memory\n"));        DPRINTF(("Copied offsets from temporary memory\n"));
6965        }        }
6966      if (md->end_offset_top > arg_offset_max) md->offset_overflow = TRUE;      if (md->end_offset_top > arg_offset_max) md->capture_last |= OVFLBIT;
6967      DPRINTF(("Freeing temporary memory\n"));      DPRINTF(("Freeing temporary memory\n"));
6968      (PUBL(free))(md->offset_vector);      (PUBL(free))(md->offset_vector);
6969      }      }
# Line 6916  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 6971  if (rc == MATCH_MATCH || rc == MATCH_ACC
6971    /* 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
6972    too many to fit into the vector. */    too many to fit into the vector. */
6973    
6974    rc = (md->offset_overflow && md->end_offset_top >= arg_offset_max)?    rc = ((md->capture_last & OVFLBIT) != 0 &&
6975             md->end_offset_top >= arg_offset_max)?
6976      0 : md->end_offset_top/2;      0 : md->end_offset_top/2;
6977    
6978    /* 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 6932  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 6988  if (rc == MATCH_MATCH || rc == MATCH_ACC
6988      {      {
6989      register int *iptr, *iend;      register int *iptr, *iend;
6990      int resetcount = 2 + re->top_bracket * 2;      int resetcount = 2 + re->top_bracket * 2;
6991      if (resetcount > offsetcount) resetcount = ocount;      if (resetcount > offsetcount) resetcount = offsetcount;
6992      iptr = offsets + md->end_offset_top;      iptr = offsets + md->end_offset_top;
6993      iend = offsets + resetcount;      iend = offsets + resetcount;
6994      while (iptr < iend) *iptr++ = -1;      while (iptr < iend) *iptr++ = -1;
# Line 6953  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7009  if (rc == MATCH_MATCH || rc == MATCH_ACC
7009    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7010      *(extra_data->mark) = (pcre_uchar *)md->mark;      *(extra_data->mark) = (pcre_uchar *)md->mark;
7011    DPRINTF((">>>> returning %d\n", rc));    DPRINTF((">>>> returning %d\n", rc));
7012    #ifdef NO_RECURSE
7013      release_match_heapframes(&frame_zero);
7014    #endif
7015    return rc;    return rc;
7016    }    }
7017    
# Line 6970  if (using_temporary_offsets) Line 7029  if (using_temporary_offsets)
7029  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)
7030    {    {
7031    DPRINTF((">>>> error: returning %d\n", rc));    DPRINTF((">>>> error: returning %d\n", rc));
7032    #ifdef NO_RECURSE
7033      release_match_heapframes(&frame_zero);
7034    #endif
7035    return rc;    return rc;
7036    }    }
7037    
# Line 6999  else Line 7061  else
7061    
7062  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7063    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;
7064    #ifdef NO_RECURSE
7065      release_match_heapframes(&frame_zero);
7066    #endif
7067  return rc;  return rc;
7068  }  }
7069    

Legend:
Removed from v.915  
changed lines
  Added in v.1248

  ViewVC Help
Powered by ViewVC 1.1.5