/[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 1313 by ph10, Wed Apr 24 12:07:09 2013 UTC revision 1431 by ph10, Thu Jan 2 17:41:28 2014 UTC
# Line 6  Line 6 
6  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
7    
8                         Written by Philip Hazel                         Written by Philip Hazel
9             Copyright (c) 1997-2013 University of Cambridge             Copyright (c) 1997-2014 University of Cambridge
10    
11  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
12  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 107  because the offset vector is always a mu Line 107  because the offset vector is always a mu
107    
108  /* Min and max values for the common repeats; for the maxima, 0 => infinity */  /* Min and max values for the common repeats; for the maxima, 0 => infinity */
109    
110  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };  static const char rep_min[] = { 0, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0, };
111  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };  static const char rep_max[] = { 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 1, };
112    
113  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
114  /*************************************************  /*************************************************
# Line 134  pcre_uint32 c; Line 134  pcre_uint32 c;
134  BOOL utf = md->utf;  BOOL utf = md->utf;
135  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;
136  while (length-- > 0)  while (length-- > 0)
137    if (isprint(c = RAWUCHARINCTEST(p))) printf("%c", (char)c); else printf("\\x{%02x}", c);    if (isprint(c = UCHAR21INCTEST(p))) printf("%c", (char)c); else printf("\\x{%02x}", c);
138  }  }
139  #endif  #endif
140    
# Line 167  match_ref(int offset, register PCRE_PUCH Line 167  match_ref(int offset, register PCRE_PUCH
167  {  {
168  PCRE_PUCHAR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
169  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
170  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && defined SUPPORT_UCP
171  BOOL utf = md->utf;  BOOL utf = md->utf;
172  #endif  #endif
173    
# Line 195  ASCII characters. */ Line 195  ASCII characters. */
195    
196  if (caseless)  if (caseless)
197    {    {
198  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && defined SUPPORT_UCP
 #ifdef SUPPORT_UCP  
199    if (utf)    if (utf)
200      {      {
201      /* 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
# Line 230  if (caseless) Line 229  if (caseless)
229      }      }
230    else    else
231  #endif  #endif
 #endif  
232    
233    /* 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
234    is no UCP support. */    is no UCP support. */
# Line 239  if (caseless) Line 237  if (caseless)
237        {        {
238        pcre_uint32 cc, cp;        pcre_uint32 cc, cp;
239        if (eptr >= md->end_subject) return -2;   /* Partial match */        if (eptr >= md->end_subject) return -2;   /* Partial match */
240        cc = RAWUCHARTEST(eptr);        cc = UCHAR21TEST(eptr);
241        cp = RAWUCHARTEST(p);        cp = UCHAR21TEST(p);
242        if (TABLE_GET(cp, md->lcc, cp) != TABLE_GET(cc, md->lcc, cc)) return -1;        if (TABLE_GET(cp, md->lcc, cp) != TABLE_GET(cc, md->lcc, cc)) return -1;
243        p++;        p++;
244        eptr++;        eptr++;
# Line 256  else Line 254  else
254    while (length-- > 0)    while (length-- > 0)
255      {      {
256      if (eptr >= md->end_subject) return -2;   /* Partial match */      if (eptr >= md->end_subject) return -2;   /* Partial match */
257      if (RAWUCHARINCTEST(p) != RAWUCHARINCTEST(eptr)) return -1;      if (UCHAR21INCTEST(p) != UCHAR21INCTEST(eptr)) return -1;
258      }      }
259    }    }
260    
# Line 312  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 310  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
310         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
311         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
312         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
313         RM61,  RM62, RM63, RM64, RM65, RM66, RM67, RM68 };         RM61,  RM62, RM63, RM64, RM65, RM66, RM67 };
314    
315  /* 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
316  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 811  for (;;) Line 809  for (;;)
809      RRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
810    
811      /* Note that, for Perl compatibility, SKIP with an argument does NOT set      /* Note that, for Perl compatibility, SKIP with an argument does NOT set
812      nomatch_mark. When a pattern match ends with a SKIP_ARG for which there was      nomatch_mark. When a pattern match ends with a SKIP_ARG for which there was
813      not a matching mark, we have to re-run the match, ignoring the SKIP_ARG      not a matching mark, we have to re-run the match, ignoring the SKIP_ARG
814      that failed and any that preceed it (either they also failed, or were not      that failed and any that precede it (either they also failed, or were not
815      triggered). To do this, we maintain a count of executed SKIP_ARGs. If a      triggered). To do this, we maintain a count of executed SKIP_ARGs. If a
816      SKIP_ARG gets to top level, the match is re-run with md->ignore_skip_arg      SKIP_ARG gets to top level, the match is re-run with md->ignore_skip_arg
817      set to the count of the one that failed. */      set to the count of the one that failed. */
818    
# Line 828  for (;;) Line 826  for (;;)
826      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
827        eptrb, RM57);        eptrb, RM57);
828      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
829    
830      /* Pass back the current skip name by overloading md->start_match_ptr and      /* Pass back the current skip name by overloading md->start_match_ptr and
831      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
832      caught by a matching MARK, or get to the top, where it causes a rematch      caught by a matching MARK, or get to the top, where it causes a rematch
# Line 1173  for (;;) Line 1171  for (;;)
1171            ecode = md->start_code + code_offset;            ecode = md->start_code + code_offset;
1172            save_capture_last = md->capture_last;            save_capture_last = md->capture_last;
1173            matched_once = TRUE;            matched_once = TRUE;
1174              mstart = md->start_match_ptr;    /* In case \K changed it */
1175            continue;            continue;
1176            }            }
1177    
# Line 1245  for (;;) Line 1244  for (;;)
1244          eptr = md->end_match_ptr;          eptr = md->end_match_ptr;
1245          ecode = md->start_code + code_offset;          ecode = md->start_code + code_offset;
1246          matched_once = TRUE;          matched_once = TRUE;
1247            mstart = md->start_match_ptr;   /* In case \K reset it */
1248          continue;          continue;
1249          }          }
1250    
# Line 1274  for (;;) Line 1274  for (;;)
1274    
1275      /* Control never reaches here. */      /* Control never reaches here. */
1276    
1277      /* Conditional group: compilation checked that there are no more than      /* Conditional group: compilation checked that there are no more than two
1278      two branches. If the condition is false, skipping the first branch takes us      branches. If the condition is false, skipping the first branch takes us
1279      past the end if there is only one branch, but that's OK because that is      past the end of the item if there is only one branch, but that's exactly
1280      exactly what going to the ket would do. */      what we want. */
1281    
1282      case OP_COND:      case OP_COND:
1283      case OP_SCOND:      case OP_SCOND:
1284      codelink = GET(ecode, 1);  
1285        /* The variable codelink will be added to ecode when the condition is
1286        false, to get to the second branch. Setting it to the offset to the ALT
1287        or KET, then incrementing ecode achieves this effect. We now have ecode
1288        pointing to the condition or callout. */
1289    
1290        codelink = GET(ecode, 1);   /* Offset to the second branch */
1291        ecode += 1 + LINK_SIZE;     /* From this opcode */
1292    
1293      /* Because of the way auto-callout works during compile, a callout item is      /* Because of the way auto-callout works during compile, a callout item is
1294      inserted between OP_COND and an assertion condition. */      inserted between OP_COND and an assertion condition. */
1295    
1296      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (*ecode == OP_CALLOUT)
1297        {        {
1298        if (PUBL(callout) != NULL)        if (PUBL(callout) != NULL)
1299          {          {
1300          PUBL(callout_block) cb;          PUBL(callout_block) cb;
1301          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1302          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[1];
1303          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1304  #if defined COMPILE_PCRE8  #if defined COMPILE_PCRE8
1305          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
# Line 1304  for (;;) Line 1311  for (;;)
1311          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1312          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1313          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
1314          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, 2);
1315          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1316          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1317          cb.capture_last     = md->capture_last & CAPLMASK;          cb.capture_last     = md->capture_last & CAPLMASK;
1318          /* Internal change requires this for API compatibility. */          /* Internal change requires this for API compatibility. */
# Line 1315  for (;;) Line 1322  for (;;)
1322          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1323          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1324          }          }
1325    
1326          /* Advance ecode past the callout, so it now points to the condition. We
1327          must adjust codelink so that the value of ecode+codelink is unchanged. */
1328    
1329        ecode += PRIV(OP_lengths)[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1330        codelink -= PRIV(OP_lengths)[OP_CALLOUT];        codelink -= PRIV(OP_lengths)[OP_CALLOUT];
1331        }        }
1332    
1333      condcode = ecode[LINK_SIZE+1];      /* Test the various possible conditions */
   
     /* Now see what the actual condition is */  
1334    
1335      if (condcode == OP_RREF || condcode == OP_NRREF)    /* Recursion test */      condition = FALSE;
1336        switch(condcode = *ecode)
1337        {        {
1338        if (md->recursive == NULL)                /* Not recursing => FALSE */        case OP_RREF:         /* Numbered group recursion test */
1339          {        if (md->recursive != NULL)     /* Not recursing => FALSE */
         condition = FALSE;  
         ecode += GET(ecode, 1);  
         }  
       else  
1340          {          {
1341          unsigned int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          unsigned int recno = GET2(ecode, 1);   /* Recursion group number*/
1342          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1343            }
1344          break;
1345    
1346          /* If the test is for recursion into a specific subpattern, and it is        case OP_DNRREF:       /* Duplicate named group recursion test */
1347          false, but the test was set up by name, scan the table to see if the        if (md->recursive != NULL)
1348          name refers to any other numbers, and test them. The condition is true          {
1349          if any one is set. */          int count = GET2(ecode, 1 + IMM2_SIZE);
1350            pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
1351          if (!condition && condcode == OP_NRREF)          while (count-- > 0)
1352            {            {
1353            pcre_uchar *slotA = md->name_table;            unsigned int recno = GET2(slot, 0);
1354            for (i = 0; i < md->name_count; i++)            condition = recno == md->recursive->group_num;
1355              {            if (condition) break;
1356              if (GET2(slotA, 0) == recno) break;            slot += md->name_entry_size;
             slotA += md->name_entry_size;  
             }  
   
           /* Found a name for the number - there can be only one; duplicate  
           names for different numbers are allowed, but not vice versa. First  
           scan down for duplicates. */  
   
           if (i < md->name_count)  
             {  
             pcre_uchar *slotB = slotA;  
             while (slotB > md->name_table)  
               {  
               slotB -= md->name_entry_size;  
               if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)  
                 {  
                 condition = GET2(slotB, 0) == md->recursive->group_num;  
                 if (condition) break;  
                 }  
               else break;  
               }  
   
             /* Scan up for duplicates */  
   
             if (!condition)  
               {  
               slotB = slotA;  
               for (i++; i < md->name_count; i++)  
                 {  
                 slotB += md->name_entry_size;  
                 if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)  
                   {  
                   condition = GET2(slotB, 0) == md->recursive->group_num;  
                   if (condition) break;  
                   }  
                 else break;  
                 }  
               }  
             }  
1357            }            }
   
         /* Chose branch according to the condition */  
   
         ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);  
1358          }          }
1359        }        break;
1360    
1361      else if (condcode == OP_CREF || condcode == OP_NCREF)  /* Group used test */        case OP_CREF:         /* Numbered group used test */
1362        {        offset = GET2(ecode, 1) << 1;  /* Doubled ref number */
       offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */  
1363        condition = offset < offset_top && md->offset_vector[offset] >= 0;        condition = offset < offset_top && md->offset_vector[offset] >= 0;
1364          break;
1365    
1366        /* If the numbered capture is unset, but the reference was by name,        case OP_DNCREF:      /* Duplicate named group used test */
       scan the table to see if the name refers to any other numbers, and test  
       them. The condition is true if any one is set. This is tediously similar  
       to the code above, but not close enough to try to amalgamate. */  
   
       if (!condition && condcode == OP_NCREF)  
1367          {          {
1368          unsigned int refno = offset >> 1;          int count = GET2(ecode, 1 + IMM2_SIZE);
1369          pcre_uchar *slotA = md->name_table;          pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
1370            while (count-- > 0)
1371          for (i = 0; i < md->name_count; i++)            {
1372            {            offset = GET2(slot, 0) << 1;
1373            if (GET2(slotA, 0) == refno) break;            condition = offset < offset_top && md->offset_vector[offset] >= 0;
1374            slotA += md->name_entry_size;            if (condition) break;
1375            }            slot += md->name_entry_size;
   
         /* Found a name for the number - there can be only one; duplicate names  
         for different numbers are allowed, but not vice versa. First scan down  
         for duplicates. */  
   
         if (i < md->name_count)  
           {  
           pcre_uchar *slotB = slotA;  
           while (slotB > md->name_table)  
             {  
             slotB -= md->name_entry_size;  
             if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)  
               {  
               offset = GET2(slotB, 0) << 1;  
               condition = offset < offset_top &&  
                 md->offset_vector[offset] >= 0;  
               if (condition) break;  
               }  
             else break;  
             }  
   
           /* Scan up for duplicates */  
   
           if (!condition)  
             {  
             slotB = slotA;  
             for (i++; i < md->name_count; i++)  
               {  
               slotB += md->name_entry_size;  
               if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)  
                 {  
                 offset = GET2(slotB, 0) << 1;  
                 condition = offset < offset_top &&  
                   md->offset_vector[offset] >= 0;  
                 if (condition) break;  
                 }  
               else break;  
               }  
             }  
1376            }            }
1377          }          }
1378          break;
1379    
1380        /* Chose branch according to the condition */        case OP_DEF:     /* DEFINE - always false */
1381          break;
       ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);  
       }  
   
     else if (condcode == OP_DEF)     /* DEFINE - always false */  
       {  
       condition = FALSE;  
       ecode += GET(ecode, 1);  
       }  
1382    
1383      /* The condition is an assertion. Call match() to evaluate it - setting        /* The condition is an assertion. Call match() to evaluate it - setting
1384      md->match_function_type to MATCH_CONDASSERT causes it to stop at the end of        md->match_function_type to MATCH_CONDASSERT causes it to stop at the end
1385      an assertion. */        of an assertion. */
1386    
1387      else        default:
       {  
1388        md->match_function_type = MATCH_CONDASSERT;        md->match_function_type = MATCH_CONDASSERT;
1389        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);        RMATCH(eptr, ecode, offset_top, md, NULL, RM3);
1390        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
1391          {          {
1392          if (md->end_offset_top > offset_top)          if (md->end_offset_top > offset_top)
1393            offset_top = md->end_offset_top;  /* Captures may have happened */            offset_top = md->end_offset_top;  /* Captures may have happened */
1394          condition = TRUE;          condition = TRUE;
1395          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);  
1396            /* Advance ecode past the assertion to the start of the first branch,
1397            but adjust it so that the general choosing code below works. */
1398    
1399            ecode += GET(ecode, 1);
1400          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
1401            ecode += 1 + LINK_SIZE - PRIV(OP_lengths)[condcode];
1402          }          }
1403    
1404        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an
1405        assertion; it is therefore treated as NOMATCH. */        assertion; it is therefore treated as NOMATCH. Any other return is an
1406          error. */
1407    
1408        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1409          {          {
1410          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
1411          }          }
1412        else        break;
         {  
         condition = FALSE;  
         ecode += codelink;  
         }  
1413        }        }
1414    
1415      /* We are now at the branch that is to be obeyed. As there is only one, can      /* Choose branch according to the condition */
     use tail recursion to avoid using another stack frame, except when there is  
     unlimited repeat of a possibly empty group. In the latter case, a recursive  
     call to match() is always required, unless the second alternative doesn't  
     exist, in which case we can just plough on. Note that, for compatibility  
     with Perl, the | in a conditional group is NOT treated as creating two  
     alternatives. If a THEN is encountered in the branch, it propagates out to  
     the enclosing alternative (unless nested in a deeper set of alternatives,  
     of course). */  
1416    
1417      if (condition || *ecode == OP_ALT)      ecode += condition? PRIV(OP_lengths)[condcode] : codelink;
1418    
1419        /* We are now at the branch that is to be obeyed. As there is only one, we
1420        can use tail recursion to avoid using another stack frame, except when
1421        there is unlimited repeat of a possibly empty group. In the latter case, a
1422        recursive call to match() is always required, unless the second alternative
1423        doesn't exist, in which case we can just plough on. Note that, for
1424        compatibility with Perl, the | in a conditional group is NOT treated as
1425        creating two alternatives. If a THEN is encountered in the branch, it
1426        propagates out to the enclosing alternative (unless nested in a deeper set
1427        of alternatives, of course). */
1428    
1429        if (condition || ecode[-(1+LINK_SIZE)] == OP_ALT)
1430        {        {
1431        if (op != OP_SCOND)        if (op != OP_SCOND)
1432          {          {
         ecode += 1 + LINK_SIZE;  
1433          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1434          }          }
1435    
1436        md->match_function_type = MATCH_CBEGROUP;        md->match_function_type = MATCH_CBEGROUP;
1437        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);        RMATCH(eptr, ecode, offset_top, md, eptrb, RM49);
1438        RRETURN(rrc);        RRETURN(rrc);
1439        }        }
1440    
# Line 1523  for (;;) Line 1442  for (;;)
1442    
1443      else      else
1444        {        {
       ecode += 1 + LINK_SIZE;  
1445        }        }
1446      break;      break;
1447    
# Line 1604  for (;;) Line 1522  for (;;)
1522      else condassert = FALSE;      else condassert = FALSE;
1523    
1524      /* Loop for each branch */      /* Loop for each branch */
1525    
1526      do      do
1527        {        {
1528        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
1529    
1530        /* A match means that the assertion is true; break out of the loop        /* A match means that the assertion is true; break out of the loop
1531        that matches its alternatives. */        that matches its alternatives. */
1532    
1533        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1534          {          {
1535          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1536          break;          break;
1537          }          }
1538    
1539        /* If not matched, restore the previous mark setting. */        /* If not matched, restore the previous mark setting. */
1540    
1541        md->mark = save_mark;        md->mark = save_mark;
1542    
1543        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
# Line 1632  for (;;) Line 1550  for (;;)
1550              (*ecode == OP_ALT || *next == OP_ALT))              (*ecode == OP_ALT || *next == OP_ALT))
1551            rrc = MATCH_NOMATCH;            rrc = MATCH_NOMATCH;
1552          }          }
1553    
1554        /* Anything other than NOMATCH causes the entire assertion to fail,        /* Anything other than NOMATCH causes the entire assertion to fail,
1555        passing back the return code. This includes COMMIT, SKIP, PRUNE and an        passing back the return code. This includes COMMIT, SKIP, PRUNE and an
1556        uncaptured THEN, which means they take their normal effect. This        uncaptured THEN, which means they take their normal effect. This
# Line 1643  for (;;) Line 1561  for (;;)
1561        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1562        }        }
1563      while (*ecode == OP_ALT);   /* Continue for next alternative */      while (*ecode == OP_ALT);   /* Continue for next alternative */
1564    
1565      /* If we have tried all the alternative branches, the assertion has      /* If we have tried all the alternative branches, the assertion has
1566      failed. If not, we broke out after a match. */      failed. If not, we broke out after a match. */
1567    
1568      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1569    
# Line 1661  for (;;) Line 1579  for (;;)
1579      offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1580      continue;      continue;
1581    
1582      /* Negative assertion: all branches must fail to match for the assertion to      /* Negative assertion: all branches must fail to match for the assertion to
1583      succeed. */      succeed. */
1584    
1585      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
# Line 1675  for (;;) Line 1593  for (;;)
1593      else condassert = FALSE;      else condassert = FALSE;
1594    
1595      /* Loop for each alternative branch. */      /* Loop for each alternative branch. */
1596    
1597      do      do
1598        {        {
1599        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1600        md->mark = save_mark;   /* Always restore the mark setting */        md->mark = save_mark;   /* Always restore the mark setting */
1601    
1602        switch(rrc)        switch(rrc)
1603          {          {
1604          case MATCH_MATCH:            /* A successful match means */          case MATCH_MATCH:            /* A successful match means */
1605          case MATCH_ACCEPT:           /* the assertion has failed. */          case MATCH_ACCEPT:           /* the assertion has failed. */
1606          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
1607    
1608          case MATCH_NOMATCH:          /* Carry on with next branch */          case MATCH_NOMATCH:          /* Carry on with next branch */
1609          break;          break;
1610    
1611          /* See comment in the code for capturing groups above about handling          /* See comment in the code for capturing groups above about handling
1612          THEN. */          THEN. */
# Line 1697  for (;;) Line 1615  for (;;)
1615          next = ecode + GET(ecode,1);          next = ecode + GET(ecode,1);
1616          if (md->start_match_ptr < next &&          if (md->start_match_ptr < next &&
1617              (*ecode == OP_ALT || *next == OP_ALT))              (*ecode == OP_ALT || *next == OP_ALT))
1618            {            {
1619            rrc = MATCH_NOMATCH;            rrc = MATCH_NOMATCH;
1620            break;            break;
1621            }            }
1622          /* Otherwise fall through. */          /* Otherwise fall through. */
1623    
1624          /* COMMIT, SKIP, PRUNE, and an uncaptured THEN cause the whole          /* COMMIT, SKIP, PRUNE, and an uncaptured THEN cause the whole
1625          assertion to fail to match, without considering any more alternatives.          assertion to fail to match, without considering any more alternatives.
1626          Failing to match means the assertion is true. This is a consistent          Failing to match means the assertion is true. This is a consistent
# Line 1710  for (;;) Line 1628  for (;;)
1628    
1629          case MATCH_COMMIT:          case MATCH_COMMIT:
1630          case MATCH_SKIP:          case MATCH_SKIP:
1631          case MATCH_SKIP_ARG:          case MATCH_SKIP_ARG:
1632          case MATCH_PRUNE:          case MATCH_PRUNE:
1633          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1634          goto NEG_ASSERT_TRUE;   /* Break out of alternation loop */          goto NEG_ASSERT_TRUE;   /* Break out of alternation loop */
1635    
1636          /* Anything else is an error */          /* Anything else is an error */
1637    
1638          default:          default:
1639          RRETURN(rrc);          RRETURN(rrc);
1640          }          }
1641    
1642        /* Continue with next branch */        /* Continue with next branch */
1643    
1644        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1645        }        }
1646      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1647    
1648      /* All branches in the assertion failed to match. */      /* All branches in the assertion failed to match. */
1649    
1650      NEG_ASSERT_TRUE:      NEG_ASSERT_TRUE:
1651      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1652      ecode += 1 + LINK_SIZE;                /* Continue with current branch */      ecode += 1 + LINK_SIZE;                /* Continue with current branch */
# Line 1896  for (;;) Line 1814  for (;;)
1814          /* PCRE does not allow THEN, SKIP, PRUNE or COMMIT to escape beyond a          /* PCRE does not allow THEN, SKIP, PRUNE or COMMIT to escape beyond a
1815          recursion; they cause a NOMATCH for the entire recursion. These codes          recursion; they cause a NOMATCH for the entire recursion. These codes
1816          are defined in a range that can be tested for. */          are defined in a range that can be tested for. */
1817    
1818          if (rrc >= MATCH_BACKTRACK_MIN && rrc <= MATCH_BACKTRACK_MAX)          if (rrc >= MATCH_BACKTRACK_MIN && rrc <= MATCH_BACKTRACK_MAX)
1819            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1820    
1821          /* Any return code other than NOMATCH is an error. */          /* Any return code other than NOMATCH is an error. */
1822    
1823          if (rrc != MATCH_NOMATCH)          if (rrc != MATCH_NOMATCH)
# Line 2089  for (;;) Line 2007  for (;;)
2007    
2008      if (*ecode == OP_KETRPOS)      if (*ecode == OP_KETRPOS)
2009        {        {
2010          md->start_match_ptr = mstart;    /* In case \K reset it */
2011        md->end_match_ptr = eptr;        md->end_match_ptr = eptr;
2012        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
2013        RRETURN(MATCH_KETRPOS);        RRETURN(MATCH_KETRPOS);
# Line 2184  for (;;) Line 2103  for (;;)
2103              eptr + 1 >= md->end_subject &&              eptr + 1 >= md->end_subject &&
2104              NLBLOCK->nltype == NLTYPE_FIXED &&              NLBLOCK->nltype == NLTYPE_FIXED &&
2105              NLBLOCK->nllen == 2 &&              NLBLOCK->nllen == 2 &&
2106              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])              UCHAR21TEST(eptr) == NLBLOCK->nl[0])
2107            {            {
2108            md->hitend = TRUE;            md->hitend = TRUE;
2109            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
# Line 2228  for (;;) Line 2147  for (;;)
2147            eptr + 1 >= md->end_subject &&            eptr + 1 >= md->end_subject &&
2148            NLBLOCK->nltype == NLTYPE_FIXED &&            NLBLOCK->nltype == NLTYPE_FIXED &&
2149            NLBLOCK->nllen == 2 &&            NLBLOCK->nllen == 2 &&
2150            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])            UCHAR21TEST(eptr) == NLBLOCK->nl[0])
2151          {          {
2152          md->hitend = TRUE;          md->hitend = TRUE;
2153          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
# Line 2371  for (;;) Line 2290  for (;;)
2290          eptr + 1 >= md->end_subject &&          eptr + 1 >= md->end_subject &&
2291          NLBLOCK->nltype == NLTYPE_FIXED &&          NLBLOCK->nltype == NLTYPE_FIXED &&
2292          NLBLOCK->nllen == 2 &&          NLBLOCK->nllen == 2 &&
2293          RAWUCHARTEST(eptr) == NLBLOCK->nl[0])          UCHAR21TEST(eptr) == NLBLOCK->nl[0])
2294        {        {
2295        md->hitend = TRUE;        md->hitend = TRUE;
2296        if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);        if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
# Line 2525  for (;;) Line 2444  for (;;)
2444          {          {
2445          SCHECK_PARTIAL();          SCHECK_PARTIAL();
2446          }          }
2447        else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;        else if (UCHAR21TEST(eptr) == CHAR_LF) eptr++;
2448        break;        break;
2449    
2450        case CHAR_LF:        case CHAR_LF:
# Line 2656  for (;;) Line 2575  for (;;)
2575            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2576          break;          break;
2577    
2578          case PT_SPACE:    /* Perl space */          /* Perl space used to exclude VT, but from Perl 5.18 it is included,
2579          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||          which means that Perl space and POSIX space are now identical. PCRE
2580               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)          was changed at release 8.34. */
                == (op == OP_NOTPROP))  
           RRETURN(MATCH_NOMATCH);  
         break;  
2581    
2582            case PT_SPACE:    /* Perl space */
2583          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2584          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||          switch(c)
2585               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||            {
2586               c == CHAR_FF || c == CHAR_CR)            HSPACE_CASES:
2587                 == (op == OP_NOTPROP))            VSPACE_CASES:
2588            RRETURN(MATCH_NOMATCH);            if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
2589              break;
2590    
2591              default:
2592              if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z) ==
2593                (op == OP_NOTPROP)) RRETURN(MATCH_NOMATCH);
2594              break;
2595              }
2596          break;          break;
2597    
2598          case PT_WORD:          case PT_WORD:
# Line 2742  for (;;) Line 2666  for (;;)
2666      similar code to character type repeats - written out again for speed.      similar code to character type repeats - written out again for speed.
2667      However, if the referenced string is the empty string, always treat      However, if the referenced string is the empty string, always treat
2668      it as matched, any number of times (otherwise there could be infinite      it as matched, any number of times (otherwise there could be infinite
2669      loops). */      loops). If the reference is unset, there are two possibilities:
   
     case OP_REF:  
     case OP_REFI:  
     caseless = op == OP_REFI;  
     offset = GET2(ecode, 1) << 1;               /* Doubled ref number */  
     ecode += 1 + IMM2_SIZE;  
   
     /* If the reference is unset, there are two possibilities:  
2670    
2671      (a) In the default, Perl-compatible state, set the length negative;      (a) In the default, Perl-compatible state, set the length negative;
2672      this ensures that every attempt at a match fails. We can't just fail      this ensures that every attempt at a match fails. We can't just fail
# Line 2760  for (;;) Line 2676  for (;;)
2676      so that the back reference matches an empty string.      so that the back reference matches an empty string.
2677    
2678      Otherwise, set the length to the length of what was matched by the      Otherwise, set the length to the length of what was matched by the
2679      referenced subpattern. */      referenced subpattern.
2680    
2681        The OP_REF and OP_REFI opcodes are used for a reference to a numbered group
2682        or to a non-duplicated named group. For a duplicated named group, OP_DNREF
2683        and OP_DNREFI are used. In this case we must scan the list of groups to
2684        which the name refers, and use the first one that is set. */
2685    
2686        case OP_DNREF:
2687        case OP_DNREFI:
2688        caseless = op == OP_DNREFI;
2689          {
2690          int count = GET2(ecode, 1+IMM2_SIZE);
2691          pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
2692          ecode += 1 + 2*IMM2_SIZE;
2693    
2694          /* Setting the default length first and initializing 'offset' avoids
2695          compiler warnings in the REF_REPEAT code. */
2696    
2697          length = (md->jscript_compat)? 0 : -1;
2698          offset = 0;
2699    
2700          while (count-- > 0)
2701            {
2702            offset = GET2(slot, 0) << 1;
2703            if (offset < offset_top && md->offset_vector[offset] >= 0)
2704              {
2705              length = md->offset_vector[offset+1] - md->offset_vector[offset];
2706              break;
2707              }
2708            slot += md->name_entry_size;
2709            }
2710          }
2711        goto REF_REPEAT;
2712    
2713        case OP_REF:
2714        case OP_REFI:
2715        caseless = op == OP_REFI;
2716        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2717        ecode += 1 + IMM2_SIZE;
2718      if (offset >= offset_top || md->offset_vector[offset] < 0)      if (offset >= offset_top || md->offset_vector[offset] < 0)
2719        length = (md->jscript_compat)? 0 : -1;        length = (md->jscript_compat)? 0 : -1;
2720      else      else
# Line 2769  for (;;) Line 2722  for (;;)
2722    
2723      /* Set up for repetition, or handle the non-repeated case */      /* Set up for repetition, or handle the non-repeated case */
2724    
2725        REF_REPEAT:
2726      switch (*ecode)      switch (*ecode)
2727        {        {
2728        case OP_CRSTAR:        case OP_CRSTAR:
# Line 2917  for (;;) Line 2871  for (;;)
2871          case OP_CRMINPLUS:          case OP_CRMINPLUS:
2872          case OP_CRQUERY:          case OP_CRQUERY:
2873          case OP_CRMINQUERY:          case OP_CRMINQUERY:
2874            case OP_CRPOSSTAR:
2875            case OP_CRPOSPLUS:
2876            case OP_CRPOSQUERY:
2877          c = *ecode++ - OP_CRSTAR;          c = *ecode++ - OP_CRSTAR;
2878          minimize = (c & 1) != 0;          if (c < OP_CRPOSSTAR - OP_CRSTAR) minimize = (c & 1) != 0;
2879            else possessive = TRUE;
2880          min = rep_min[c];                 /* Pick up values from tables; */          min = rep_min[c];                 /* Pick up values from tables; */
2881          max = rep_max[c];                 /* zero for max => infinity */          max = rep_max[c];                 /* zero for max => infinity */
2882          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
# Line 2926  for (;;) Line 2884  for (;;)
2884    
2885          case OP_CRRANGE:          case OP_CRRANGE:
2886          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2887            case OP_CRPOSRANGE:
2888          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2889            possessive = (*ecode == OP_CRPOSRANGE);
2890          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2891          max = GET2(ecode, 1 + IMM2_SIZE);          max = GET2(ecode, 1 + IMM2_SIZE);
2892          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
# Line 3068  for (;;) Line 3028  for (;;)
3028                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3029              eptr += len;              eptr += len;
3030              }              }
3031    
3032              if (possessive) continue;    /* No backtracking */
3033    
3034            for (;;)            for (;;)
3035              {              {
3036              RMATCH(eptr, ecode, offset_top, md, eptrb, RM18);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM18);
# Line 3098  for (;;) Line 3061  for (;;)
3061                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3062              eptr++;              eptr++;
3063              }              }
3064    
3065              if (possessive) continue;    /* No backtracking */
3066    
3067            while (eptr >= pp)            while (eptr >= pp)
3068              {              {
3069              RMATCH(eptr, ecode, offset_top, md, eptrb, RM19);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM19);
# Line 3113  for (;;) Line 3079  for (;;)
3079      /* Control never gets here */      /* Control never gets here */
3080    
3081    
3082      /* Match an extended character class. This opcode is encountered only      /* Match an extended character class. In the 8-bit library, this opcode is
3083      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8      encountered only when UTF-8 mode mode is supported. In the 16-bit and
3084      mode, because Unicode properties are supported in non-UTF-8 mode. */      32-bit libraries, codepoints greater than 255 may be encountered even when
3085        UTF is not supported. */
3086    
3087  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3088      case OP_XCLASS:      case OP_XCLASS:
# Line 3131  for (;;) Line 3098  for (;;)
3098          case OP_CRMINPLUS:          case OP_CRMINPLUS:
3099          case OP_CRQUERY:          case OP_CRQUERY:
3100          case OP_CRMINQUERY:          case OP_CRMINQUERY:
3101            case OP_CRPOSSTAR:
3102            case OP_CRPOSPLUS:
3103            case OP_CRPOSQUERY:
3104          c = *ecode++ - OP_CRSTAR;          c = *ecode++ - OP_CRSTAR;
3105          minimize = (c & 1) != 0;          if (c < OP_CRPOSSTAR - OP_CRSTAR) minimize = (c & 1) != 0;
3106            else possessive = TRUE;
3107          min = rep_min[c];                 /* Pick up values from tables; */          min = rep_min[c];                 /* Pick up values from tables; */
3108          max = rep_max[c];                 /* zero for max => infinity */          max = rep_max[c];                 /* zero for max => infinity */
3109          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
# Line 3140  for (;;) Line 3111  for (;;)
3111    
3112          case OP_CRRANGE:          case OP_CRRANGE:
3113          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3114            case OP_CRPOSRANGE:
3115          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3116            possessive = (*ecode == OP_CRPOSRANGE);
3117          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3118          max = GET2(ecode, 1 + IMM2_SIZE);          max = GET2(ecode, 1 + IMM2_SIZE);
3119          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
# Line 3212  for (;;) Line 3185  for (;;)
3185            if (!PRIV(xclass)(c, data, utf)) break;            if (!PRIV(xclass)(c, data, utf)) break;
3186            eptr += len;            eptr += len;
3187            }            }
3188    
3189            if (possessive) continue;    /* No backtracking */
3190    
3191          for(;;)          for(;;)
3192            {            {
3193            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
# Line 3242  for (;;) Line 3218  for (;;)
3218          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3219          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3220          }          }
3221        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != UCHAR21INC(eptr)) RRETURN(MATCH_NOMATCH);
3222        }        }
3223      else      else
3224  #endif  #endif
# Line 3282  for (;;) Line 3258  for (;;)
3258    
3259        if (fc < 128)        if (fc < 128)
3260          {          {
3261          pcre_uint32 cc = RAWUCHAR(eptr);          pcre_uint32 cc = UCHAR21(eptr);
3262          if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);          if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3263          ecode++;          ecode++;
3264          eptr++;          eptr++;
# Line 3387  for (;;) Line 3363  for (;;)
3363      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
3364      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
3365    
3366      /* Common code for all repeated single-character matches. We first check      /* Common code for all repeated single-character matches. We first check
3367      for the minimum number of characters. If the minimum equals the maximum, we      for the minimum number of characters. If the minimum equals the maximum, we
3368      are done. Otherwise, if minimizing, check the rest of the pattern for a      are done. Otherwise, if minimizing, check the rest of the pattern for a
3369      match; if there isn't one, advance up to the maximum, one character at a      match; if there isn't one, advance up to the maximum, one character at a
3370      time.      time.
3371    
3372      If maximizing, advance up to the maximum number of matching characters,      If maximizing, advance up to the maximum number of matching characters,
3373      until eptr is past the end of the maximum run. If possessive, we are      until eptr is past the end of the maximum run. If possessive, we are
3374      then done (no backing up). Otherwise, match at this position; anything      then done (no backing up). Otherwise, match at this position; anything
3375      other than no match is immediately returned. For nomatch, back up one      other than no match is immediately returned. For nomatch, back up one
3376      character, unless we are matching \R and the last thing matched was      character, unless we are matching \R and the last thing matched was
3377      \r\n, in which case, back up two bytes. When we reach the first optional      \r\n, in which case, back up two bytes. When we reach the first optional
3378      character position, we can save stack by doing a tail recurse.      character position, we can save stack by doing a tail recurse.
3379    
3380      The various UTF/non-UTF and caseful/caseless cases are handled separately,      The various UTF/non-UTF and caseful/caseless cases are handled separately,
3381      for speed. */      for speed. */
3382    
# Line 3551  for (;;) Line 3527  for (;;)
3527            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3528            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3529            }            }
3530          cc = RAWUCHARTEST(eptr);          cc = UCHAR21TEST(eptr);
3531          if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);          if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3532          eptr++;          eptr++;
3533          }          }
# Line 3569  for (;;) Line 3545  for (;;)
3545              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3546              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3547              }              }
3548            cc = RAWUCHARTEST(eptr);            cc = UCHAR21TEST(eptr);
3549            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3550            eptr++;            eptr++;
3551            }            }
# Line 3586  for (;;) Line 3562  for (;;)
3562              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3563              break;              break;
3564              }              }
3565            cc = RAWUCHARTEST(eptr);            cc = UCHAR21TEST(eptr);
3566            if (fc != cc && foc != cc) break;            if (fc != cc && foc != cc) break;
3567            eptr++;            eptr++;
3568            }            }
   
3569          if (possessive) continue;       /* No backtracking */          if (possessive) continue;       /* No backtracking */
3570          for (;;)          for (;;)
3571            {            {
3572            if (eptr == pp) goto TAIL_RECURSE;            if (eptr == pp) goto TAIL_RECURSE;
3573            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);
3574            eptr--;            eptr--;
3575            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3576            }            }
3577          RRETURN(MATCH_NOMATCH);          /* Control never gets here */
3578          }          }
       /* Control never gets here */  
3579        }        }
3580    
3581      /* Caseful comparisons (includes all multi-byte characters) */      /* Caseful comparisons (includes all multi-byte characters) */
# Line 3615  for (;;) Line 3589  for (;;)
3589            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3590            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3591            }            }
3592          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);          if (fc != UCHAR21INCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3593          }          }
3594    
3595        if (min == max) continue;        if (min == max) continue;
# Line 3632  for (;;) Line 3606  for (;;)
3606              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3607              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3608              }              }
3609            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);            if (fc != UCHAR21INCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3610            }            }
3611          /* Control never gets here */          /* Control never gets here */
3612          }          }
# Line 3646  for (;;) Line 3620  for (;;)
3620              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3621              break;              break;
3622              }              }
3623            if (fc != RAWUCHARTEST(eptr)) break;            if (fc != UCHAR21TEST(eptr)) break;
3624            eptr++;            eptr++;
3625            }            }
3626          if (possessive) continue;    /* No backtracking */          if (possessive) continue;    /* No backtracking */
# Line 3657  for (;;) Line 3631  for (;;)
3631            eptr--;            eptr--;
3632            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3633            }            }
3634          RRETURN(MATCH_NOMATCH);          /* Control never gets here */
3635          }          }
3636        }        }
3637      /* Control never gets here */      /* Control never gets here */
# Line 3942  for (;;) Line 3916  for (;;)
3916              eptr--;              eptr--;
3917              }              }
3918            }            }
3919            /* Control never gets here */
         RRETURN(MATCH_NOMATCH);  
3920          }          }
       /* Control never gets here */  
3921        }        }
3922    
3923      /* Caseful comparisons */      /* Caseful comparisons */
# Line 4073  for (;;) Line 4045  for (;;)
4045            if (possessive) continue;    /* No backtracking */            if (possessive) continue;    /* No backtracking */
4046            for (;;)            for (;;)
4047              {              {
4048              if (eptr == pp) goto TAIL_RECURSE;              if (eptr == pp) goto TAIL_RECURSE;
4049              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);
4050              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4051              eptr--;              eptr--;
4052              }              }
4053            }            }
4054            /* Control never gets here */
         RRETURN(MATCH_NOMATCH);  
4055          }          }
4056        }        }
4057      /* Control never gets here */      /* Control never gets here */
# Line 4262  for (;;) Line 4233  for (;;)
4233              }              }
4234            break;            break;
4235    
4236              /* Perl space used to exclude VT, but from Perl 5.18 it is included,
4237              which means that Perl space and POSIX space are now identical. PCRE
4238              was changed at release 8.34. */
4239    
4240            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
4241              case PT_PXSPACE:  /* POSIX space */
4242            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4243              {              {
4244              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4271  for (;;) Line 4247  for (;;)
4247                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4248                }                }
4249              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4250              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              switch(c)
                  c == CHAR_FF || c == CHAR_CR)  
                    == prop_fail_result)  
               RRETURN(MATCH_NOMATCH);  
             }  
           break;  
   
           case PT_PXSPACE:  /* POSIX space */  
           for (i = 1; i <= min; i++)  
             {  
             if (eptr >= md->end_subject)  
4251                {                {
4252                SCHECK_PARTIAL();                HSPACE_CASES:
4253                RRETURN(MATCH_NOMATCH);                VSPACE_CASES:
4254                  if (prop_fail_result) RRETURN(MATCH_NOMATCH);
4255                  break;
4256    
4257                  default:
4258                  if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result)
4259                    RRETURN(MATCH_NOMATCH);
4260                  break;
4261                }                }
             GETCHARINCTEST(c, eptr);  
             if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||  
                  c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)  
                    == prop_fail_result)  
               RRETURN(MATCH_NOMATCH);  
4262              }              }
4263            break;            break;
4264    
# Line 4407  for (;;) Line 4375  for (;;)
4375                eptr + 1 >= md->end_subject &&                eptr + 1 >= md->end_subject &&
4376                NLBLOCK->nltype == NLTYPE_FIXED &&                NLBLOCK->nltype == NLTYPE_FIXED &&
4377                NLBLOCK->nllen == 2 &&                NLBLOCK->nllen == 2 &&
4378                RAWUCHAR(eptr) == NLBLOCK->nl[0])                UCHAR21(eptr) == NLBLOCK->nl[0])
4379              {              {
4380              md->hitend = TRUE;              md->hitend = TRUE;
4381              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
# Line 4449  for (;;) Line 4417  for (;;)
4417              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4418    
4419              case CHAR_CR:              case CHAR_CR:
4420              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;              if (eptr < md->end_subject && UCHAR21(eptr) == CHAR_LF) eptr++;
4421              break;              break;
4422    
4423              case CHAR_LF:              case CHAR_LF:
# Line 4559  for (;;) Line 4527  for (;;)
4527              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4528              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4529              }              }
4530            cc = RAWUCHAR(eptr);            cc = UCHAR21(eptr);
4531            if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)            if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4532              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4533            eptr++;            eptr++;
# Line 4576  for (;;) Line 4544  for (;;)
4544              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4545              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4546              }              }
4547            cc = RAWUCHAR(eptr);            cc = UCHAR21(eptr);
4548            if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)            if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4549              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4550            eptr++;            eptr++;
# Line 4593  for (;;) Line 4561  for (;;)
4561              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4562              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4563              }              }
4564            cc = RAWUCHAR(eptr);            cc = UCHAR21(eptr);
4565            if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)            if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4566              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4567            eptr++;            eptr++;
# Line 4610  for (;;) Line 4578  for (;;)
4578              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4579              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4580              }              }
4581            cc = RAWUCHAR(eptr);            cc = UCHAR21(eptr);
4582            if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)            if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4583              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4584            eptr++;            eptr++;
# Line 4627  for (;;) Line 4595  for (;;)
4595              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4596              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4597              }              }
4598            cc = RAWUCHAR(eptr);            cc = UCHAR21(eptr);
4599            if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)            if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4600              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4601            eptr++;            eptr++;
# Line 5010  for (;;) Line 4978  for (;;)
4978              }              }
4979            /* Control never gets here */            /* Control never gets here */
4980    
4981            case PT_SPACE:    /* Perl space */            /* Perl space used to exclude VT, but from Perl 5.18 it is included,
4982            for (fi = min;; fi++)            which means that Perl space and POSIX space are now identical. PCRE
4983              {            was changed at release 8.34. */
             RMATCH(eptr, ecode, offset_top, md, eptrb, RM60);  
             if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
             if (fi >= max) RRETURN(MATCH_NOMATCH);  
             if (eptr >= md->end_subject)  
               {  
               SCHECK_PARTIAL();  
               RRETURN(MATCH_NOMATCH);  
               }  
             GETCHARINCTEST(c, eptr);  
             if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||  
                  c == CHAR_FF || c == CHAR_CR)  
                    == prop_fail_result)  
               RRETURN(MATCH_NOMATCH);  
             }  
           /* Control never gets here */  
4984    
4985              case PT_SPACE:    /* Perl space */
4986            case PT_PXSPACE:  /* POSIX space */            case PT_PXSPACE:  /* POSIX space */
4987            for (fi = min;; fi++)            for (fi = min;; fi++)
4988              {              {
# Line 5041  for (;;) Line 4995  for (;;)
4995                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4996                }                }
4997              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4998              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              switch(c)
4999                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)                {
5000                     == prop_fail_result)                HSPACE_CASES:
5001                RRETURN(MATCH_NOMATCH);                VSPACE_CASES:
5002                  if (prop_fail_result) RRETURN(MATCH_NOMATCH);
5003                  break;
5004    
5005                  default:
5006                  if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result)
5007                    RRETURN(MATCH_NOMATCH);
5008                  break;
5009                  }
5010              }              }
5011            /* Control never gets here */            /* Control never gets here */
5012    
# Line 5097  for (;;) Line 5059  for (;;)
5059            case PT_UCNC:            case PT_UCNC:
5060            for (fi = min;; fi++)            for (fi = min;; fi++)
5061              {              {
5062              RMATCH(eptr, ecode, offset_top, md, eptrb, RM68);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM60);
5063              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5064              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
5065              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 5194  for (;;) Line 5156  for (;;)
5156                {                {
5157                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5158                case CHAR_CR:                case CHAR_CR:
5159                if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;                if (eptr < md->end_subject && UCHAR21(eptr) == CHAR_LF) eptr++;
5160                break;                break;
5161    
5162                case CHAR_LF:                case CHAR_LF:
# Line 5528  for (;;) Line 5490  for (;;)
5490              }              }
5491            break;            break;
5492    
5493              /* Perl space used to exclude VT, but from Perl 5.18 it is included,
5494              which means that Perl space and POSIX space are now identical. PCRE
5495              was changed at release 8.34. */
5496    
5497            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
5498              case PT_PXSPACE:  /* POSIX space */
5499            for (i = min; i < max; i++)            for (i = min; i < max; i++)
5500              {              {
5501              int len = 1;              int len = 1;
# Line 5538  for (;;) Line 5505  for (;;)
5505                break;                break;
5506                }                }
5507              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5508              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              switch(c)
5509                   c == CHAR_FF || c == CHAR_CR)                {
5510                   == prop_fail_result)                HSPACE_CASES:
5511                  VSPACE_CASES:
5512                  if (prop_fail_result) goto ENDLOOP99;  /* Break the loop */
5513                break;                break;
             eptr+= len;  
             }  
           break;  
5514    
5515            case PT_PXSPACE:  /* POSIX space */                default:
5516            for (i = min; i < max; i++)                if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result)
5517              {                  goto ENDLOOP99;   /* Break the loop */
             int len = 1;  
             if (eptr >= md->end_subject)  
               {  
               SCHECK_PARTIAL();  
5518                break;                break;
5519                }                }
             GETCHARLENTEST(c, eptr, len);  
             if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||  
                  c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)  
                  == prop_fail_result)  
               break;  
5520              eptr+= len;              eptr+= len;
5521              }              }
5522              ENDLOOP99:
5523            break;            break;
5524    
5525            case PT_WORD:            case PT_WORD:
# Line 5642  for (;;) Line 5600  for (;;)
5600            }            }
5601          }          }
5602    
5603        /* Match extended Unicode sequences. We will get here only if the        /* Match extended Unicode grapheme clusters. We will get here only if the
5604        support is in the binary; otherwise a compile-time error occurs. */        support is in the binary; otherwise a compile-time error occurs. */
5605    
5606        else if (ctype == OP_EXTUNI)        else if (ctype == OP_EXTUNI)
# Line 5675  for (;;) Line 5633  for (;;)
5633          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
5634    
5635          if (possessive) continue;    /* No backtracking */          if (possessive) continue;    /* No backtracking */
5636    
5637          for(;;)          for(;;)
5638            {            {
5639            if (eptr == pp) goto TAIL_RECURSE;            int lgb, rgb;
5640              PCRE_PUCHAR fptr;
5641    
5642              if (eptr == pp) goto TAIL_RECURSE;   /* At start of char run */
5643            RMATCH(eptr, ecode, offset_top, md, eptrb, RM45);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM45);
5644            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5645            eptr--;  
5646            for (;;)                        /* Move back over one extended */            /* Backtracking over an extended grapheme cluster involves inspecting
5647              the previous two characters (if present) to see if a break is
5648              permitted between them. */
5649    
5650              eptr--;
5651              if (!utf) c = *eptr; else
5652                {
5653                BACKCHAR(eptr);
5654                GETCHAR(c, eptr);
5655                }
5656              rgb = UCD_GRAPHBREAK(c);
5657    
5658              for (;;)
5659              {              {
5660              if (!utf) c = *eptr; else              if (eptr == pp) goto TAIL_RECURSE;   /* At start of char run */
5661                fptr = eptr - 1;
5662                if (!utf) c = *fptr; else
5663                {                {
5664                BACKCHAR(eptr);                BACKCHAR(fptr);
5665                GETCHAR(c, eptr);                GETCHAR(c, fptr);
5666                }                }
5667              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5668              eptr--;              if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5669                eptr = fptr;
5670                rgb = lgb;
5671              }              }
5672            }            }
5673          }          }
# Line 5717  for (;;) Line 5695  for (;;)
5695                    eptr + 1 >= md->end_subject &&                    eptr + 1 >= md->end_subject &&
5696                    NLBLOCK->nltype == NLTYPE_FIXED &&                    NLBLOCK->nltype == NLTYPE_FIXED &&
5697                    NLBLOCK->nllen == 2 &&                    NLBLOCK->nllen == 2 &&
5698                    RAWUCHAR(eptr) == NLBLOCK->nl[0])                    UCHAR21(eptr) == NLBLOCK->nl[0])
5699                  {                  {
5700                  md->hitend = TRUE;                  md->hitend = TRUE;
5701                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
# Line 5743  for (;;) Line 5721  for (;;)
5721                    eptr + 1 >= md->end_subject &&                    eptr + 1 >= md->end_subject &&
5722                    NLBLOCK->nltype == NLTYPE_FIXED &&                    NLBLOCK->nltype == NLTYPE_FIXED &&
5723                    NLBLOCK->nllen == 2 &&                    NLBLOCK->nllen == 2 &&
5724                    RAWUCHAR(eptr) == NLBLOCK->nl[0])                    UCHAR21(eptr) == NLBLOCK->nl[0])
5725                  {                  {
5726                  md->hitend = TRUE;                  md->hitend = TRUE;
5727                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
# Line 5800  for (;;) Line 5778  for (;;)
5778              if (c == CHAR_CR)              if (c == CHAR_CR)
5779                {                {
5780                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5781                if (RAWUCHAR(eptr) == CHAR_LF) eptr++;                if (UCHAR21(eptr) == CHAR_LF) eptr++;
5782                }                }
5783              else              else
5784                {                {
# Line 5958  for (;;) Line 5936  for (;;)
5936          if (possessive) continue;    /* No backtracking */          if (possessive) continue;    /* No backtracking */
5937          for(;;)          for(;;)
5938            {            {
5939            if (eptr == pp) goto TAIL_RECURSE;            if (eptr == pp) goto TAIL_RECURSE;
5940            RMATCH(eptr, ecode, offset_top, md, eptrb, RM46);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM46);
5941            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5942            eptr--;            eptr--;
5943            BACKCHAR(eptr);            BACKCHAR(eptr);
5944            if (ctype == OP_ANYNL && eptr > pp  && RAWUCHAR(eptr) == CHAR_NL &&            if (ctype == OP_ANYNL && eptr > pp  && UCHAR21(eptr) == CHAR_NL &&
5945                RAWUCHAR(eptr - 1) == CHAR_CR) eptr--;                UCHAR21(eptr - 1) == CHAR_CR) eptr--;
5946            }            }
5947          }          }
5948        else        else
# Line 6211  for (;;) Line 6189  for (;;)
6189            }            }
6190          }          }
6191    
6192        /* Get here if we can't make it match with any permitted repetitions */        /* Control never gets here */
   
       RRETURN(MATCH_NOMATCH);  
6193        }        }
     /* Control never gets here */  
6194    
6195      /* There's been some horrible disaster. Arrival here can only mean there is      /* There's been some horrible disaster. Arrival here can only mean there is
6196      something seriously wrong in the code above or the OP_xxx definitions. */      something seriously wrong in the code above or the OP_xxx definitions. */
# Line 6249  switch (frame->Xwhere) Line 6224  switch (frame->Xwhere)
6224    LBL(53) LBL(54) LBL(55) LBL(56) LBL(57) LBL(58) LBL(63) LBL(64)    LBL(53) LBL(54) LBL(55) LBL(56) LBL(57) LBL(58) LBL(63) LBL(64)
6225    LBL(65) LBL(66)    LBL(65) LBL(66)
6226  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
6227    LBL(21)    LBL(20) LBL(21)
6228  #endif  #endif
6229  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
6230    LBL(16) LBL(18) LBL(20)    LBL(16) LBL(18)
6231    LBL(22) LBL(23) LBL(28) LBL(30)    LBL(22) LBL(23) LBL(28) LBL(30)
6232    LBL(32) LBL(34) LBL(42) LBL(46)    LBL(32) LBL(34) LBL(42) LBL(46)
6233  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6234    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)
6235    LBL(59) LBL(60) LBL(61) LBL(62) LBL(67) LBL(68)    LBL(59) LBL(60) LBL(61) LBL(62) LBL(67)
6236  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
6237  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
6238    default:    default:
# Line 6410  const pcre_uint8 *start_bits = NULL; Line 6385  const pcre_uint8 *start_bits = NULL;
6385  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;
6386  PCRE_PUCHAR end_subject;  PCRE_PUCHAR end_subject;
6387  PCRE_PUCHAR start_partial = NULL;  PCRE_PUCHAR start_partial = NULL;
6388  PCRE_PUCHAR match_partial;  PCRE_PUCHAR match_partial = NULL;
6389  PCRE_PUCHAR req_char_ptr = start_match - 1;  PCRE_PUCHAR req_char_ptr = start_match - 1;
6390    
6391  const pcre_study_data *study;  const pcre_study_data *study;
# Line 6511  if (extra_data != NULL Line 6486  if (extra_data != NULL
6486      && extra_data->executable_jit != NULL      && extra_data->executable_jit != NULL
6487      && (options & ~PUBLIC_JIT_EXEC_OPTIONS) == 0)      && (options & ~PUBLIC_JIT_EXEC_OPTIONS) == 0)
6488    {    {
   /* A facility for setting the match limit in the regex was added; this puts  
   a value in the compiled block. (Similarly for recursion limit, but the JIT  
   does not make use of that.) Because the regex is not passed to jit_exec, we  
   fudge up an alternative extra block, because we must not modify the extra  
   block that the user has passed. */  
   
 #if defined COMPILE_PCRE8  
   pcre_extra extra_data_copy;  
 #elif defined COMPILE_PCRE16  
   pcre16_extra extra_data_copy;  
 #elif defined COMPILE_PCRE32  
   pcre32_extra extra_data_copy;  
 #endif  
   
   if ((re->flags & PCRE_MLSET) != 0 &&  
       ((extra_data->flags & PCRE_EXTRA_MATCH_LIMIT) == 0 ||  
        re->limit_match < extra_data->match_limit))  
     {  
     extra_data_copy = *extra_data;  
     extra_data_copy.match_limit = re->limit_match;  
     extra_data_copy.flags |= PCRE_EXTRA_MATCH_LIMIT;  
     extra_data = &extra_data_copy;  
     }  
   
6489    rc = PRIV(jit_exec)(extra_data, (const pcre_uchar *)subject, length,    rc = PRIV(jit_exec)(extra_data, (const pcre_uchar *)subject, length,
6490         start_offset, options, offsets, offsetcount);         start_offset, options, offsets, offsetcount);
6491    
# Line 6581  if (extra_data != NULL) Line 6532  if (extra_data != NULL)
6532    }    }
6533    
6534  /* Limits in the regex override only if they are smaller. */  /* Limits in the regex override only if they are smaller. */
6535    
6536  if ((re->flags & PCRE_MLSET) != 0 && re->limit_match < md->match_limit)  if ((re->flags & PCRE_MLSET) != 0 && re->limit_match < md->match_limit)
6537    md->match_limit = re->limit_match;    md->match_limit = re->limit_match;
6538    
6539  if ((re->flags & PCRE_RLSET) != 0 &&  if ((re->flags & PCRE_RLSET) != 0 &&
6540      re->limit_recursion < md->match_limit_recursion)      re->limit_recursion < md->match_limit_recursion)
6541    md->match_limit_recursion = re->limit_recursion;    md->match_limit_recursion = re->limit_recursion;
6542    
# Line 6838  for(;;) Line 6789  for(;;)
6789    
6790        if (first_char != first_char2)        if (first_char != first_char2)
6791          while (start_match < end_subject &&          while (start_match < end_subject &&
6792            (smc = RAWUCHARTEST(start_match)) != first_char && smc != first_char2)            (smc = UCHAR21TEST(start_match)) != first_char && smc != first_char2)
6793            start_match++;            start_match++;
6794        else        else
6795          while (start_match < end_subject && RAWUCHARTEST(start_match) != first_char)          while (start_match < end_subject && UCHAR21TEST(start_match) != first_char)
6796            start_match++;            start_match++;
6797        }        }
6798    
# Line 6873  for(;;) Line 6824  for(;;)
6824          if (start_match[-1] == CHAR_CR &&          if (start_match[-1] == CHAR_CR &&
6825               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
6826               start_match < end_subject &&               start_match < end_subject &&
6827               RAWUCHARTEST(start_match) == CHAR_NL)               UCHAR21TEST(start_match) == CHAR_NL)
6828            start_match++;            start_match++;
6829          }          }
6830        }        }
# Line 6884  for(;;) Line 6835  for(;;)
6835        {        {
6836        while (start_match < end_subject)        while (start_match < end_subject)
6837          {          {
6838          register pcre_uint32 c = RAWUCHARTEST(start_match);          register pcre_uint32 c = UCHAR21TEST(start_match);
6839  #ifndef COMPILE_PCRE8  #ifndef COMPILE_PCRE8
6840          if (c > 255) c = 255;          if (c > 255) c = 255;
6841  #endif  #endif
6842          if ((start_bits[c/8] & (1 << (c&7))) == 0)          if ((start_bits[c/8] & (1 << (c&7))) != 0) break;
6843            {          start_match++;
           start_match++;  
 #if defined SUPPORT_UTF && defined COMPILE_PCRE8  
           /* In non 8-bit mode, the iteration will stop for  
           characters > 255 at the beginning or not stop at all. */  
           if (utf)  
             ACROSSCHAR(start_match < end_subject, *start_match,  
               start_match++);  
 #endif  
           }  
         else break;  
6844          }          }
6845        }        }
6846      }   /* Starting optimizations */      }   /* Starting optimizations */
# Line 6952  for(;;) Line 6893  for(;;)
6893            {            {
6894            while (p < end_subject)            while (p < end_subject)
6895              {              {
6896              register pcre_uint32 pp = RAWUCHARINCTEST(p);              register pcre_uint32 pp = UCHAR21INCTEST(p);
6897              if (pp == req_char || pp == req_char2) { p--; break; }              if (pp == req_char || pp == req_char2) { p--; break; }
6898              }              }
6899            }            }
# Line 6960  for(;;) Line 6901  for(;;)
6901            {            {
6902            while (p < end_subject)            while (p < end_subject)
6903              {              {
6904              if (RAWUCHARINCTEST(p) == req_char) { p--; break; }              if (UCHAR21INCTEST(p) == req_char) { p--; break; }
6905              }              }
6906            }            }
6907    
# Line 7202  if (rc != MATCH_NOMATCH && rc != PCRE_ER Line 7143  if (rc != MATCH_NOMATCH && rc != PCRE_ER
7143    
7144  /* Handle partial matches - disable any mark data */  /* Handle partial matches - disable any mark data */
7145    
7146  if (start_partial != NULL)  if (match_partial != NULL)
7147    {    {
7148    DPRINTF((">>>> returning PCRE_ERROR_PARTIAL\n"));    DPRINTF((">>>> returning PCRE_ERROR_PARTIAL\n"));
7149    md->mark = NULL;    md->mark = NULL;

Legend:
Removed from v.1313  
changed lines
  Added in v.1431

  ViewVC Help
Powered by ViewVC 1.1.5