/[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 1296 by ph10, Tue Mar 19 16:29:12 2013 UTC revision 1397 by ph10, Mon Nov 11 18:33:23 2013 UTC
# 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 811  for (;;) Line 811  for (;;)
811      RRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
812    
813      /* Note that, for Perl compatibility, SKIP with an argument does NOT set      /* Note that, for Perl compatibility, SKIP with an argument does NOT set
814      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
815      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
816      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
817      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
818      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
819      set to the count of the one that failed. */      set to the count of the one that failed. */
820    
# Line 828  for (;;) Line 828  for (;;)
828      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
829        eptrb, RM57);        eptrb, RM57);
830      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
831    
832      /* 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
833      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
834      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 1173  for (;;)
1173            ecode = md->start_code + code_offset;            ecode = md->start_code + code_offset;
1174            save_capture_last = md->capture_last;            save_capture_last = md->capture_last;
1175            matched_once = TRUE;            matched_once = TRUE;
1176              mstart = md->start_match_ptr;    /* In case \K changed it */
1177            continue;            continue;
1178            }            }
1179    
# Line 1245  for (;;) Line 1246  for (;;)
1246          eptr = md->end_match_ptr;          eptr = md->end_match_ptr;
1247          ecode = md->start_code + code_offset;          ecode = md->start_code + code_offset;
1248          matched_once = TRUE;          matched_once = TRUE;
1249            mstart = md->start_match_ptr;   /* In case \K reset it */
1250          continue;          continue;
1251          }          }
1252    
# Line 1274  for (;;) Line 1276  for (;;)
1276    
1277      /* Control never reaches here. */      /* Control never reaches here. */
1278    
1279      /* Conditional group: compilation checked that there are no more than      /* Conditional group: compilation checked that there are no more than two
1280      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
1281      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
1282      exactly what going to the ket would do. */      what we want. */
1283    
1284      case OP_COND:      case OP_COND:
1285      case OP_SCOND:      case OP_SCOND:
1286      codelink = GET(ecode, 1);  
1287        /* The variable codelink will be added to ecode when the condition is
1288        false, to get to the second branch. Setting it to the offset to the ALT
1289        or KET, then incrementing ecode achieves this effect. We now have ecode
1290        pointing to the condition or callout. */
1291    
1292        codelink = GET(ecode, 1);   /* Offset to the second branch */
1293        ecode += 1 + LINK_SIZE;     /* From this opcode */
1294    
1295      /* 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
1296      inserted between OP_COND and an assertion condition. */      inserted between OP_COND and an assertion condition. */
1297    
1298      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (*ecode == OP_CALLOUT)
1299        {        {
1300        if (PUBL(callout) != NULL)        if (PUBL(callout) != NULL)
1301          {          {
1302          PUBL(callout_block) cb;          PUBL(callout_block) cb;
1303          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1304          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[1];
1305          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1306  #if defined COMPILE_PCRE8  #if defined COMPILE_PCRE8
1307          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
# Line 1304  for (;;) Line 1313  for (;;)
1313          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1314          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1315          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
1316          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, 2);
1317          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1318          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1319          cb.capture_last     = md->capture_last & CAPLMASK;          cb.capture_last     = md->capture_last & CAPLMASK;
1320          /* Internal change requires this for API compatibility. */          /* Internal change requires this for API compatibility. */
# Line 1315  for (;;) Line 1324  for (;;)
1324          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1325          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1326          }          }
1327    
1328          /* Advance ecode past the callout, so it now points to the condition. We
1329          must adjust codelink so that the value of ecode+codelink is unchanged. */
1330    
1331        ecode += PRIV(OP_lengths)[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1332        codelink -= PRIV(OP_lengths)[OP_CALLOUT];        codelink -= PRIV(OP_lengths)[OP_CALLOUT];
1333        }        }
1334    
1335      condcode = ecode[LINK_SIZE+1];      /* Test the various possible conditions */
   
     /* Now see what the actual condition is */  
1336    
1337      if (condcode == OP_RREF || condcode == OP_NRREF)    /* Recursion test */      condition = FALSE;
1338        switch(condcode = *ecode)
1339        {        {
1340        if (md->recursive == NULL)                /* Not recursing => FALSE */        case OP_RREF:         /* Numbered group recursion test */
1341          if (md->recursive != NULL)     /* Not recursing => FALSE */
1342          {          {
1343          condition = FALSE;          unsigned int recno = GET2(ecode, 1);   /* Recursion group number*/
         ecode += GET(ecode, 1);  
         }  
       else  
         {  
         unsigned int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/  
1344          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1345            }
1346          break;
1347    
1348          /* If the test is for recursion into a specific subpattern, and it is        case OP_DNRREF:       /* Duplicate named group recursion test */
1349          false, but the test was set up by name, scan the table to see if the        if (md->recursive != NULL)
1350          name refers to any other numbers, and test them. The condition is true          {
1351          if any one is set. */          int count = GET2(ecode, 1 + IMM2_SIZE);
1352            pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
1353          if (!condition && condcode == OP_NRREF)          while (count-- > 0)
1354            {            {
1355            pcre_uchar *slotA = md->name_table;            unsigned int recno = GET2(slot, 0);
1356            for (i = 0; i < md->name_count; i++)            condition = recno == md->recursive->group_num;
1357              {            if (condition) break;
1358              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;  
                 }  
               }  
             }  
1359            }            }
   
         /* Chose branch according to the condition */  
   
         ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);  
1360          }          }
1361        }        break;
1362    
1363      else if (condcode == OP_CREF || condcode == OP_NCREF)  /* Group used test */        case OP_CREF:         /* Numbered group used test */
1364        {        offset = GET2(ecode, 1) << 1;  /* Doubled ref number */
       offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */  
1365        condition = offset < offset_top && md->offset_vector[offset] >= 0;        condition = offset < offset_top && md->offset_vector[offset] >= 0;
1366          break;
1367    
1368        /* 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)  
1369          {          {
1370          unsigned int refno = offset >> 1;          int count = GET2(ecode, 1 + IMM2_SIZE);
1371          pcre_uchar *slotA = md->name_table;          pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
1372            while (count-- > 0)
1373          for (i = 0; i < md->name_count; i++)            {
1374            {            offset = GET2(slot, 0) << 1;
1375            if (GET2(slotA, 0) == refno) break;            condition = offset < offset_top && md->offset_vector[offset] >= 0;
1376            slotA += md->name_entry_size;            if (condition) break;
1377            }            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;  
               }  
             }  
1378            }            }
1379          }          }
1380          break;
1381    
1382        /* Chose branch according to the condition */        case OP_DEF:     /* DEFINE - always false */
1383          break;
       ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);  
       }  
   
     else if (condcode == OP_DEF)     /* DEFINE - always false */  
       {  
       condition = FALSE;  
       ecode += GET(ecode, 1);  
       }  
1384    
1385      /* The condition is an assertion. Call match() to evaluate it - setting        /* The condition is an assertion. Call match() to evaluate it - setting
1386      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
1387      an assertion. */        of an assertion. */
1388    
1389      else        default:
       {  
1390        md->match_function_type = MATCH_CONDASSERT;        md->match_function_type = MATCH_CONDASSERT;
1391        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);        RMATCH(eptr, ecode, offset_top, md, NULL, RM3);
1392        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
1393          {          {
1394          if (md->end_offset_top > offset_top)          if (md->end_offset_top > offset_top)
1395            offset_top = md->end_offset_top;  /* Captures may have happened */            offset_top = md->end_offset_top;  /* Captures may have happened */
1396          condition = TRUE;          condition = TRUE;
1397          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);  
1398            /* Advance ecode past the assertion to the start of the first branch,
1399            but adjust it so that the general choosing code below works. */
1400    
1401            ecode += GET(ecode, 1);
1402          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
1403            ecode += 1 + LINK_SIZE - PRIV(OP_lengths)[condcode];
1404          }          }
1405    
1406        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an
1407        assertion; it is therefore treated as NOMATCH. */        assertion; it is therefore treated as NOMATCH. Any other return is an
1408          error. */
1409    
1410        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1411          {          {
1412          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
1413          }          }
1414        else        break;
         {  
         condition = FALSE;  
         ecode += codelink;  
         }  
1415        }        }
1416    
1417      /* We are now at the branch that is to be obeyed. As there is only one, can      /* Choose branch according to the condition */
1418      use tail recursion to avoid using another stack frame, except when there is  
1419      unlimited repeat of a possibly empty group. In the latter case, a recursive      ecode += condition? PRIV(OP_lengths)[condcode] : codelink;
     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). */  
1420    
1421      if (condition || *ecode == OP_ALT)      /* We are now at the branch that is to be obeyed. As there is only one, we
1422        can use tail recursion to avoid using another stack frame, except when
1423        there is unlimited repeat of a possibly empty group. In the latter case, a
1424        recursive call to match() is always required, unless the second alternative
1425        doesn't exist, in which case we can just plough on. Note that, for
1426        compatibility with Perl, the | in a conditional group is NOT treated as
1427        creating two alternatives. If a THEN is encountered in the branch, it
1428        propagates out to the enclosing alternative (unless nested in a deeper set
1429        of alternatives, of course). */
1430    
1431        if (condition || ecode[-(1+LINK_SIZE)] == OP_ALT)
1432        {        {
1433        if (op != OP_SCOND)        if (op != OP_SCOND)
1434          {          {
         ecode += 1 + LINK_SIZE;  
1435          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1436          }          }
1437    
1438        md->match_function_type = MATCH_CBEGROUP;        md->match_function_type = MATCH_CBEGROUP;
1439        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);        RMATCH(eptr, ecode, offset_top, md, eptrb, RM49);
1440        RRETURN(rrc);        RRETURN(rrc);
1441        }        }
1442    
# Line 1523  for (;;) Line 1444  for (;;)
1444    
1445      else      else
1446        {        {
       ecode += 1 + LINK_SIZE;  
1447        }        }
1448      break;      break;
1449    
# Line 1604  for (;;) Line 1524  for (;;)
1524      else condassert = FALSE;      else condassert = FALSE;
1525    
1526      /* Loop for each branch */      /* Loop for each branch */
1527    
1528      do      do
1529        {        {
1530        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
1531    
1532          /* A match means that the assertion is true; break out of the loop
1533          that matches its alternatives. */
1534    
1535        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1536          {          {
1537          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1538          break;          break;
1539          }          }
1540    
1541          /* If not matched, restore the previous mark setting. */
1542    
1543        md->mark = save_mark;        md->mark = save_mark;
1544    
1545        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
# Line 1625  for (;;) Line 1552  for (;;)
1552              (*ecode == OP_ALT || *next == OP_ALT))              (*ecode == OP_ALT || *next == OP_ALT))
1553            rrc = MATCH_NOMATCH;            rrc = MATCH_NOMATCH;
1554          }          }
1555    
1556        /* Anything other than NOMATCH causes the assertion to fail. This        /* Anything other than NOMATCH causes the entire assertion to fail,
1557        includes COMMIT, SKIP, and PRUNE. However, this consistent approach does        passing back the return code. This includes COMMIT, SKIP, PRUNE and an
1558        not always have exactly the same effect as in Perl. */        uncaptured THEN, which means they take their normal effect. This
1559          consistent approach does not always have exactly the same effect as in
1560          Perl. */
1561    
1562        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1563        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1564        }        }
1565      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);   /* Continue for next alternative */
1566    
1567      /* If we have tried all the alternative branches, the assertion has      /* If we have tried all the alternative branches, the assertion has
1568      failed. */      failed. If not, we broke out after a match. */
1569    
1570      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1571    
# Line 1652  for (;;) Line 1581  for (;;)
1581      offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1582      continue;      continue;
1583    
1584      /* Negative assertion: all branches must fail to match for the assertion to      /* Negative assertion: all branches must fail to match for the assertion to
1585      succeed. */      succeed. */
1586    
1587      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
# Line 1666  for (;;) Line 1595  for (;;)
1595      else condassert = FALSE;      else condassert = FALSE;
1596    
1597      /* Loop for each alternative branch. */      /* Loop for each alternative branch. */
1598    
1599      do      do
1600        {        {
1601        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1602        md->mark = save_mark;        md->mark = save_mark;   /* Always restore the mark setting */
   
       /* A successful match means the assertion has failed. */  
   
       if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);  
1603    
1604        /* See comment in the code for capturing groups above about handling        switch(rrc)
       THEN. */  
   
       if (rrc == MATCH_THEN)  
1605          {          {
1606            case MATCH_MATCH:            /* A successful match means */
1607            case MATCH_ACCEPT:           /* the assertion has failed. */
1608            RRETURN(MATCH_NOMATCH);
1609    
1610            case MATCH_NOMATCH:          /* Carry on with next branch */
1611            break;
1612    
1613            /* See comment in the code for capturing groups above about handling
1614            THEN. */
1615    
1616            case MATCH_THEN:
1617          next = ecode + GET(ecode,1);          next = ecode + GET(ecode,1);
1618          if (md->start_match_ptr < next &&          if (md->start_match_ptr < next &&
1619              (*ecode == OP_ALT || *next == OP_ALT))              (*ecode == OP_ALT || *next == OP_ALT))
1620              {
1621            rrc = MATCH_NOMATCH;            rrc = MATCH_NOMATCH;
1622              break;
1623              }
1624            /* Otherwise fall through. */
1625    
1626            /* COMMIT, SKIP, PRUNE, and an uncaptured THEN cause the whole
1627            assertion to fail to match, without considering any more alternatives.
1628            Failing to match means the assertion is true. This is a consistent
1629            approach, but does not always have the same effect as in Perl. */
1630    
1631            case MATCH_COMMIT:
1632            case MATCH_SKIP:
1633            case MATCH_SKIP_ARG:
1634            case MATCH_PRUNE:
1635            do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1636            goto NEG_ASSERT_TRUE;   /* Break out of alternation loop */
1637    
1638            /* Anything else is an error */
1639    
1640            default:
1641            RRETURN(rrc);
1642          }          }
   
       /* No match on a branch means we must carry on and try the next branch.  
       Anything else, in particular, SKIP, PRUNE, etc. causes a failure in the  
       enclosing branch. This is a consistent approach, but does not always have  
       the same effect as in Perl. */  
1643    
1644        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        /* Continue with next branch */
1645    
1646        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1647        }        }
1648      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1649    
1650      /* All branches in the assertion failed to match. */      /* All branches in the assertion failed to match. */
1651    
1652        NEG_ASSERT_TRUE:
1653      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1654      ecode += 1 + LINK_SIZE;                /* Continue with current branch */      ecode += 1 + LINK_SIZE;                /* Continue with current branch */
1655      continue;      continue;
# Line 1863  for (;;) Line 1814  for (;;)
1814            }            }
1815    
1816          /* 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
1817          recursion; they are treated as NOMATCH. These codes are defined in a          recursion; they cause a NOMATCH for the entire recursion. These codes
1818          range that can be tested for. Any other return code is an error. */          are defined in a range that can be tested for. */
1819    
1820          else if (rrc != MATCH_NOMATCH &&          if (rrc >= MATCH_BACKTRACK_MIN && rrc <= MATCH_BACKTRACK_MAX)
1821                   (rrc < MATCH_BACKTRACK_MIN || rrc > MATCH_BACKTRACK_MAX))            RRETURN(MATCH_NOMATCH);
1822    
1823            /* Any return code other than NOMATCH is an error. */
1824    
1825            if (rrc != MATCH_NOMATCH)
1826            {            {
1827            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1828            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
# Line 2054  for (;;) Line 2009  for (;;)
2009    
2010      if (*ecode == OP_KETRPOS)      if (*ecode == OP_KETRPOS)
2011        {        {
2012          md->start_match_ptr = mstart;    /* In case \K reset it */
2013        md->end_match_ptr = eptr;        md->end_match_ptr = eptr;
2014        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
2015        RRETURN(MATCH_KETRPOS);        RRETURN(MATCH_KETRPOS);
# Line 2621  for (;;) Line 2577  for (;;)
2577            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2578          break;          break;
2579    
2580          case PT_SPACE:    /* Perl space */          /* Perl space used to exclude VT, but from Perl 5.18 it is included,
2581          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||          which means that Perl space and POSIX space are now identical. PCRE
2582               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;  
2583    
2584            case PT_SPACE:    /* Perl space */
2585          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2586          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||          switch(c)
2587               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||            {
2588               c == CHAR_FF || c == CHAR_CR)            HSPACE_CASES:
2589                 == (op == OP_NOTPROP))            VSPACE_CASES:
2590            RRETURN(MATCH_NOMATCH);            if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
2591              break;
2592    
2593              default:
2594              if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z) ==
2595                (op == OP_NOTPROP)) RRETURN(MATCH_NOMATCH);
2596              break;
2597              }
2598          break;          break;
2599    
2600          case PT_WORD:          case PT_WORD:
# Line 2707  for (;;) Line 2668  for (;;)
2668      similar code to character type repeats - written out again for speed.      similar code to character type repeats - written out again for speed.
2669      However, if the referenced string is the empty string, always treat      However, if the referenced string is the empty string, always treat
2670      it as matched, any number of times (otherwise there could be infinite      it as matched, any number of times (otherwise there could be infinite
2671      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:  
2672    
2673      (a) In the default, Perl-compatible state, set the length negative;      (a) In the default, Perl-compatible state, set the length negative;
2674      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 2725  for (;;) Line 2678  for (;;)
2678      so that the back reference matches an empty string.      so that the back reference matches an empty string.
2679    
2680      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
2681      referenced subpattern. */      referenced subpattern.
2682    
2683        The OP_REF and OP_REFI opcodes are used for a reference to a numbered group
2684        or to a non-duplicated named group. For a duplicated named group, OP_DNREF
2685        and OP_DNREFI are used. In this case we must scan the list of groups to
2686        which the name refers, and use the first one that is set. */
2687    
2688        case OP_DNREF:
2689        case OP_DNREFI:
2690        caseless = op == OP_DNREFI;
2691          {
2692          int count = GET2(ecode, 1+IMM2_SIZE);
2693          pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
2694          ecode += 1 + 2*IMM2_SIZE;
2695    
2696          while (count-- > 0)
2697            {
2698            offset = GET2(slot, 0) << 1;
2699            if (offset < offset_top && md->offset_vector[offset] >= 0) break;
2700            slot += md->name_entry_size;
2701            }
2702          if (count < 0)
2703            length = (md->jscript_compat)? 0 : -1;
2704          else
2705            length = md->offset_vector[offset+1] - md->offset_vector[offset];
2706          }
2707        goto REF_REPEAT;
2708    
2709        case OP_REF:
2710        case OP_REFI:
2711        caseless = op == OP_REFI;
2712        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2713        ecode += 1 + IMM2_SIZE;
2714      if (offset >= offset_top || md->offset_vector[offset] < 0)      if (offset >= offset_top || md->offset_vector[offset] < 0)
2715        length = (md->jscript_compat)? 0 : -1;        length = (md->jscript_compat)? 0 : -1;
2716      else      else
# Line 2734  for (;;) Line 2718  for (;;)
2718    
2719      /* Set up for repetition, or handle the non-repeated case */      /* Set up for repetition, or handle the non-repeated case */
2720    
2721        REF_REPEAT:
2722      switch (*ecode)      switch (*ecode)
2723        {        {
2724        case OP_CRSTAR:        case OP_CRSTAR:
# Line 2882  for (;;) Line 2867  for (;;)
2867          case OP_CRMINPLUS:          case OP_CRMINPLUS:
2868          case OP_CRQUERY:          case OP_CRQUERY:
2869          case OP_CRMINQUERY:          case OP_CRMINQUERY:
2870            case OP_CRPOSSTAR:
2871            case OP_CRPOSPLUS:
2872            case OP_CRPOSQUERY:
2873          c = *ecode++ - OP_CRSTAR;          c = *ecode++ - OP_CRSTAR;
2874          minimize = (c & 1) != 0;          if (c < OP_CRPOSSTAR - OP_CRSTAR) minimize = (c & 1) != 0;
2875            else possessive = TRUE;
2876          min = rep_min[c];                 /* Pick up values from tables; */          min = rep_min[c];                 /* Pick up values from tables; */
2877          max = rep_max[c];                 /* zero for max => infinity */          max = rep_max[c];                 /* zero for max => infinity */
2878          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
# Line 2891  for (;;) Line 2880  for (;;)
2880    
2881          case OP_CRRANGE:          case OP_CRRANGE:
2882          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2883            case OP_CRPOSRANGE:
2884          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2885            possessive = (*ecode == OP_CRPOSRANGE);
2886          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2887          max = GET2(ecode, 1 + IMM2_SIZE);          max = GET2(ecode, 1 + IMM2_SIZE);
2888          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
# Line 3033  for (;;) Line 3024  for (;;)
3024                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3025              eptr += len;              eptr += len;
3026              }              }
3027    
3028              if (possessive) continue;    /* No backtracking */
3029    
3030            for (;;)            for (;;)
3031              {              {
3032              RMATCH(eptr, ecode, offset_top, md, eptrb, RM18);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM18);
# Line 3063  for (;;) Line 3057  for (;;)
3057                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3058              eptr++;              eptr++;
3059              }              }
3060    
3061              if (possessive) continue;    /* No backtracking */
3062    
3063            while (eptr >= pp)            while (eptr >= pp)
3064              {              {
3065              RMATCH(eptr, ecode, offset_top, md, eptrb, RM19);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM19);
# Line 3096  for (;;) Line 3093  for (;;)
3093          case OP_CRMINPLUS:          case OP_CRMINPLUS:
3094          case OP_CRQUERY:          case OP_CRQUERY:
3095          case OP_CRMINQUERY:          case OP_CRMINQUERY:
3096            case OP_CRPOSSTAR:
3097            case OP_CRPOSPLUS:
3098            case OP_CRPOSQUERY:
3099          c = *ecode++ - OP_CRSTAR;          c = *ecode++ - OP_CRSTAR;
3100          minimize = (c & 1) != 0;          if (c < OP_CRPOSSTAR - OP_CRSTAR) minimize = (c & 1) != 0;
3101            else possessive = TRUE;
3102          min = rep_min[c];                 /* Pick up values from tables; */          min = rep_min[c];                 /* Pick up values from tables; */
3103          max = rep_max[c];                 /* zero for max => infinity */          max = rep_max[c];                 /* zero for max => infinity */
3104          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
# Line 3105  for (;;) Line 3106  for (;;)
3106    
3107          case OP_CRRANGE:          case OP_CRRANGE:
3108          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3109            case OP_CRPOSRANGE:
3110          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3111            possessive = (*ecode == OP_CRPOSRANGE);
3112          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3113          max = GET2(ecode, 1 + IMM2_SIZE);          max = GET2(ecode, 1 + IMM2_SIZE);
3114          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
# Line 3177  for (;;) Line 3180  for (;;)
3180            if (!PRIV(xclass)(c, data, utf)) break;            if (!PRIV(xclass)(c, data, utf)) break;
3181            eptr += len;            eptr += len;
3182            }            }
3183    
3184            if (possessive) continue;    /* No backtracking */
3185    
3186          for(;;)          for(;;)
3187            {            {
3188            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
# Line 3352  for (;;) Line 3358  for (;;)
3358      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
3359      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
3360    
3361      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. We first check
3362        for the minimum number of characters. If the minimum equals the maximum, we
3363        are done. Otherwise, if minimizing, check the rest of the pattern for a
3364        match; if there isn't one, advance up to the maximum, one character at a
3365        time.
3366    
3367        If maximizing, advance up to the maximum number of matching characters,
3368        until eptr is past the end of the maximum run. If possessive, we are
3369        then done (no backing up). Otherwise, match at this position; anything
3370        other than no match is immediately returned. For nomatch, back up one
3371        character, unless we are matching \R and the last thing matched was
3372        \r\n, in which case, back up two bytes. When we reach the first optional
3373        character position, we can save stack by doing a tail recurse.
3374    
3375        The various UTF/non-UTF and caseful/caseless cases are handled separately,
3376        for speed. */
3377    
3378      REPEATCHAR:      REPEATCHAR:
3379  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
# Line 3436  for (;;) Line 3457  for (;;)
3457                }                }
3458              }              }
3459    
3460            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
   
3461            for(;;)            for(;;)
3462              {              {
3463                if (eptr == pp) goto TAIL_RECURSE;
3464              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3465              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
             if (eptr == pp) { RRETURN(MATCH_NOMATCH); }  
3466  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3467              eptr--;              eptr--;
3468              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3541  for (;;) Line 3561  for (;;)
3561            if (fc != cc && foc != cc) break;            if (fc != cc && foc != cc) break;
3562            eptr++;            eptr++;
3563            }            }
3564            if (possessive) continue;       /* No backtracking */
3565          if (possessive) continue;          for (;;)
   
         while (eptr >= pp)  
3566            {            {
3567              if (eptr == pp) goto TAIL_RECURSE;
3568            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);
3569            eptr--;            eptr--;
3570            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3571            }            }
3572          RRETURN(MATCH_NOMATCH);          /* Control never gets here */
3573          }          }
       /* Control never gets here */  
3574        }        }
3575    
3576      /* Caseful comparisons (includes all multi-byte characters) */      /* Caseful comparisons (includes all multi-byte characters) */
# Line 3600  for (;;) Line 3618  for (;;)
3618            if (fc != RAWUCHARTEST(eptr)) break;            if (fc != RAWUCHARTEST(eptr)) break;
3619            eptr++;            eptr++;
3620            }            }
3621          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
3622            for (;;)
         while (eptr >= pp)  
3623            {            {
3624              if (eptr == pp) goto TAIL_RECURSE;
3625            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);
3626            eptr--;            eptr--;
3627            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3628            }            }
3629          RRETURN(MATCH_NOMATCH);          /* Control never gets here */
3630          }          }
3631        }        }
3632      /* Control never gets here */      /* Control never gets here */
# Line 3780  for (;;) Line 3798  for (;;)
3798            }            }
3799          }          }
3800        else        else
3801  #endif  #endif  /* SUPPORT_UTF */
3802        /* Not UTF mode */        /* Not UTF mode */
3803          {          {
3804          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3818  for (;;) Line 3836  for (;;)
3836              }              }
3837            }            }
3838          else          else
3839  #endif  #endif  /*SUPPORT_UTF */
3840          /* Not UTF mode */          /* Not UTF mode */
3841            {            {
3842            for (fi = min;; fi++)            for (fi = min;; fi++)
# Line 3860  for (;;) Line 3878  for (;;)
3878              if (fc == d || (unsigned int)foc == d) break;              if (fc == d || (unsigned int)foc == d) break;
3879              eptr += len;              eptr += len;
3880              }              }
3881            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3882            for(;;)            for(;;)
3883              {              {
3884                if (eptr == pp) goto TAIL_RECURSE;
3885              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3886              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3887              if (eptr-- == pp) break;        /* Stop if tried at original pos */              eptr--;
3888              BACKCHAR(eptr);              BACKCHAR(eptr);
3889              }              }
3890            }            }
3891          else          else
3892  #endif  #endif  /* SUPPORT_UTF */
3893          /* Not UTF mode */          /* Not UTF mode */
3894            {            {
3895            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3883  for (;;) Line 3902  for (;;)
3902              if (fc == *eptr || foc == *eptr) break;              if (fc == *eptr || foc == *eptr) break;
3903              eptr++;              eptr++;
3904              }              }
3905            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3906            while (eptr >= pp)            for (;;)
3907              {              {
3908                if (eptr == pp) goto TAIL_RECURSE;
3909              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);
3910              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3911              eptr--;              eptr--;
3912              }              }
3913            }            }
3914            /* Control never gets here */
         RRETURN(MATCH_NOMATCH);  
3915          }          }
       /* Control never gets here */  
3916        }        }
3917    
3918      /* Caseful comparisons */      /* Caseful comparisons */
# Line 3995  for (;;) Line 4013  for (;;)
4013              if (fc == d) break;              if (fc == d) break;
4014              eptr += len;              eptr += len;
4015              }              }
4016            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
4017            for(;;)            for(;;)
4018              {              {
4019                if (eptr == pp) goto TAIL_RECURSE;
4020              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);
4021              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4022              if (eptr-- == pp) break;        /* Stop if tried at original pos */              eptr--;
4023              BACKCHAR(eptr);              BACKCHAR(eptr);
4024              }              }
4025            }            }
# Line 4018  for (;;) Line 4037  for (;;)
4037              if (fc == *eptr) break;              if (fc == *eptr) break;
4038              eptr++;              eptr++;
4039              }              }
4040            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
4041            while (eptr >= pp)            for (;;)
4042              {              {
4043                if (eptr == pp) goto TAIL_RECURSE;
4044              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);
4045              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4046              eptr--;              eptr--;
4047              }              }
4048            }            }
4049            /* Control never gets here */
         RRETURN(MATCH_NOMATCH);  
4050          }          }
4051        }        }
4052      /* Control never gets here */      /* Control never gets here */
# Line 4209  for (;;) Line 4228  for (;;)
4228              }              }
4229            break;            break;
4230    
4231              /* Perl space used to exclude VT, but from Perl 5.18 it is included,
4232              which means that Perl space and POSIX space are now identical. PCRE
4233              was changed at release 8.34. */
4234    
4235            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
4236              case PT_PXSPACE:  /* POSIX space */
4237            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4238              {              {
4239              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4218  for (;;) Line 4242  for (;;)
4242                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4243                }                }
4244              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4245              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)  
4246                {                {
4247                SCHECK_PARTIAL();                HSPACE_CASES:
4248                RRETURN(MATCH_NOMATCH);                VSPACE_CASES:
4249                  if (prop_fail_result) RRETURN(MATCH_NOMATCH);
4250                  break;
4251    
4252                  default:
4253                  if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result)
4254                    RRETURN(MATCH_NOMATCH);
4255                  break;
4256                }                }
             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);  
4257              }              }
4258            break;            break;
4259    
# Line 4957  for (;;) Line 4973  for (;;)
4973              }              }
4974            /* Control never gets here */            /* Control never gets here */
4975    
4976            case PT_SPACE:    /* Perl space */            /* Perl space used to exclude VT, but from Perl 5.18 it is included,
4977            for (fi = min;; fi++)            which means that Perl space and POSIX space are now identical. PCRE
4978              {            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 */  
4979    
4980              case PT_SPACE:    /* Perl space */
4981            case PT_PXSPACE:  /* POSIX space */            case PT_PXSPACE:  /* POSIX space */
4982            for (fi = min;; fi++)            for (fi = min;; fi++)
4983              {              {
# Line 4988  for (;;) Line 4990  for (;;)
4990                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4991                }                }
4992              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4993              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              switch(c)
4994                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)                {
4995                     == prop_fail_result)                HSPACE_CASES:
4996                RRETURN(MATCH_NOMATCH);                VSPACE_CASES:
4997                  if (prop_fail_result) RRETURN(MATCH_NOMATCH);
4998                  break;
4999    
5000                  default:
5001                  if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result)
5002                    RRETURN(MATCH_NOMATCH);
5003                  break;
5004                  }
5005              }              }
5006            /* Control never gets here */            /* Control never gets here */
5007    
# Line 5475  for (;;) Line 5485  for (;;)
5485              }              }
5486            break;            break;
5487    
5488              /* Perl space used to exclude VT, but from Perl 5.18 it is included,
5489              which means that Perl space and POSIX space are now identical. PCRE
5490              was changed at release 8.34. */
5491    
5492            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
5493              case PT_PXSPACE:  /* POSIX space */
5494            for (i = min; i < max; i++)            for (i = min; i < max; i++)
5495              {              {
5496              int len = 1;              int len = 1;
# Line 5485  for (;;) Line 5500  for (;;)
5500                break;                break;
5501                }                }
5502              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5503              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              switch(c)
5504                   c == CHAR_FF || c == CHAR_CR)                {
5505                   == prop_fail_result)                HSPACE_CASES:
5506                  VSPACE_CASES:
5507                  if (prop_fail_result) goto ENDLOOP99;  /* Break the loop */
5508                break;                break;
             eptr+= len;  
             }  
           break;  
5509    
5510            case PT_PXSPACE:  /* POSIX space */                default:
5511            for (i = min; i < max; i++)                if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result)
5512              {                  goto ENDLOOP99;   /* Break the loop */
             int len = 1;  
             if (eptr >= md->end_subject)  
               {  
               SCHECK_PARTIAL();  
5513                break;                break;
5514                }                }
             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;  
5515              eptr+= len;              eptr+= len;
5516              }              }
5517              ENDLOOP99:
5518            break;            break;
5519    
5520            case PT_WORD:            case PT_WORD:
# Line 5578  for (;;) Line 5584  for (;;)
5584    
5585          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
5586    
5587          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
5588          for(;;)          for(;;)
5589            {            {
5590              if (eptr == pp) goto TAIL_RECURSE;
5591            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5592            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5593            if (eptr-- == pp) break;        /* Stop if tried at original pos */            eptr--;
5594            if (utf) BACKCHAR(eptr);            if (utf) BACKCHAR(eptr);
5595            }            }
5596          }          }
5597    
5598        /* Match extended Unicode sequences. We will get here only if the        /* Match extended Unicode grapheme clusters. We will get here only if the
5599        support is in the binary; otherwise a compile-time error occurs. */        support is in the binary; otherwise a compile-time error occurs. */
5600    
5601        else if (ctype == OP_EXTUNI)        else if (ctype == OP_EXTUNI)
# Line 5620  for (;;) Line 5627  for (;;)
5627    
5628          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
5629    
5630          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
5631    
5632          for(;;)          for(;;)
5633            {            {
5634              int lgb, rgb;
5635              PCRE_PUCHAR fptr;
5636    
5637              if (eptr == pp) goto TAIL_RECURSE;   /* At start of char run */
5638            RMATCH(eptr, ecode, offset_top, md, eptrb, RM45);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM45);
5639            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5640            if (eptr-- == pp) break;        /* Stop if tried at original pos */  
5641            for (;;)                        /* Move back over one extended */            /* Backtracking over an extended grapheme cluster involves inspecting
5642              the previous two characters (if present) to see if a break is
5643              permitted between them. */
5644    
5645              eptr--;
5646              if (!utf) c = *eptr; else
5647              {              {
5648              if (!utf) c = *eptr; else              BACKCHAR(eptr);
5649                GETCHAR(c, eptr);
5650                }
5651              rgb = UCD_GRAPHBREAK(c);
5652    
5653              for (;;)
5654                {
5655                if (eptr == pp) goto TAIL_RECURSE;   /* At start of char run */
5656                fptr = eptr - 1;
5657                if (!utf) c = *fptr; else
5658                {                {
5659                BACKCHAR(eptr);                BACKCHAR(fptr);
5660                GETCHAR(c, eptr);                GETCHAR(c, fptr);
5661                }                }
5662              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5663              eptr--;              if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5664                eptr = fptr;
5665                rgb = lgb;
5666              }              }
5667            }            }
5668          }          }
# Line 5901  for (;;) Line 5928  for (;;)
5928            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5929            }            }
5930    
5931          /* eptr is now past the end of the maximum run. If possessive, we are          if (possessive) continue;    /* No backtracking */
         done (no backing up). Otherwise, match at this position; anything other  
         than no match is immediately returned. For nomatch, back up one  
         character, unless we are matching \R and the last thing matched was  
         \r\n, in which case, back up two bytes. */  
   
         if (possessive) continue;  
5932          for(;;)          for(;;)
5933            {            {
5934              if (eptr == pp) goto TAIL_RECURSE;
5935            RMATCH(eptr, ecode, offset_top, md, eptrb, RM46);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM46);
5936            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5937            if (eptr-- == pp) break;        /* Stop if tried at original pos */            eptr--;
5938            BACKCHAR(eptr);            BACKCHAR(eptr);
5939            if (ctype == OP_ANYNL && eptr > pp  && RAWUCHAR(eptr) == CHAR_NL &&            if (ctype == OP_ANYNL && eptr > pp  && RAWUCHAR(eptr) == CHAR_NL &&
5940                RAWUCHAR(eptr - 1) == CHAR_CR) eptr--;                RAWUCHAR(eptr - 1) == CHAR_CR) eptr--;
# Line 6150  for (;;) Line 6172  for (;;)
6172            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
6173            }            }
6174    
6175          /* eptr is now past the end of the maximum run. If possessive, we are          if (possessive) continue;    /* No backtracking */
6176          done (no backing up). Otherwise, match at this position; anything other          for (;;)
         than no match is immediately returned. For nomatch, back up one  
         character (byte), unless we are matching \R and the last thing matched  
         was \r\n, in which case, back up two bytes. */  
   
         if (possessive) continue;  
         while (eptr >= pp)  
6177            {            {
6178              if (eptr == pp) goto TAIL_RECURSE;
6179            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);
6180            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
6181            eptr--;            eptr--;
# Line 6167  for (;;) Line 6184  for (;;)
6184            }            }
6185          }          }
6186    
6187        /* Get here if we can't make it match with any permitted repetitions */        /* Control never gets here */
   
       RRETURN(MATCH_NOMATCH);  
6188        }        }
     /* Control never gets here */  
6189    
6190      /* 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
6191      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 6366  const pcre_uint8 *start_bits = NULL; Line 6380  const pcre_uint8 *start_bits = NULL;
6380  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;
6381  PCRE_PUCHAR end_subject;  PCRE_PUCHAR end_subject;
6382  PCRE_PUCHAR start_partial = NULL;  PCRE_PUCHAR start_partial = NULL;
6383  PCRE_PUCHAR match_partial;  PCRE_PUCHAR match_partial = NULL;
6384  PCRE_PUCHAR req_char_ptr = start_match - 1;  PCRE_PUCHAR req_char_ptr = start_match - 1;
6385    
6386  const pcre_study_data *study;  const pcre_study_data *study;
# Line 6496  md->callout_data = NULL; Line 6510  md->callout_data = NULL;
6510    
6511  tables = re->tables;  tables = re->tables;
6512    
6513    /* The two limit values override the defaults, whatever their value. */
6514    
6515  if (extra_data != NULL)  if (extra_data != NULL)
6516    {    {
6517    register unsigned int flags = extra_data->flags;    register unsigned int flags = extra_data->flags;
# Line 6510  if (extra_data != NULL) Line 6526  if (extra_data != NULL)
6526    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;
6527    }    }
6528    
6529    /* Limits in the regex override only if they are smaller. */
6530    
6531    if ((re->flags & PCRE_MLSET) != 0 && re->limit_match < md->match_limit)
6532      md->match_limit = re->limit_match;
6533    
6534    if ((re->flags & PCRE_RLSET) != 0 &&
6535        re->limit_recursion < md->match_limit_recursion)
6536      md->match_limit_recursion = re->limit_recursion;
6537    
6538  /* If the exec call supplied NULL for tables, use the inbuilt ones. This  /* If the exec call supplied NULL for tables, use the inbuilt ones. This
6539  is a feature that makes it possible to save compiled regex and re-use them  is a feature that makes it possible to save compiled regex and re-use them
6540  in other programs later. */  in other programs later. */
# Line 7123  if (rc != MATCH_NOMATCH && rc != PCRE_ER Line 7148  if (rc != MATCH_NOMATCH && rc != PCRE_ER
7148    
7149  /* Handle partial matches - disable any mark data */  /* Handle partial matches - disable any mark data */
7150    
7151  if (start_partial != NULL)  if (match_partial != NULL)
7152    {    {
7153    DPRINTF((">>>> returning PCRE_ERROR_PARTIAL\n"));    DPRINTF((">>>> returning PCRE_ERROR_PARTIAL\n"));
7154    md->mark = NULL;    md->mark = NULL;

Legend:
Removed from v.1296  
changed lines
  Added in v.1397

  ViewVC Help
Powered by ViewVC 1.1.5