/[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 1325 by ph10, Fri May 10 14:03:18 2013 UTC revision 1365 by ph10, Sun Oct 6 18:33:56 2013 UTC
# 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 */
1334    
1335      /* Now see what the actual condition is */      condition = FALSE;
1336        switch(condcode = *ecode)
1337      if (condcode == OP_RREF || condcode == OP_NRREF)    /* Recursion test */        {
1338        {        case OP_RREF:         /* Numbered group recursion test */
1339        if (md->recursive == NULL)                /* Not recursing => FALSE */        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            }            }
1358            }
1359          break;
1360    
1361          /* Chose branch according to the condition */        case OP_CREF:         /* Numbered group used test */
1362          offset = GET2(ecode, 1) << 1;  /* Doubled ref number */
         ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);  
         }  
       }  
   
     else if (condcode == OP_CREF || condcode == OP_NCREF)  /* Group used test */  
       {  
       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)
         for (i = 0; i < md->name_count; i++)  
           {  
           if (GET2(slotA, 0) == refno) break;  
           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)  
1371            {            {
1372            pcre_uchar *slotB = slotA;            offset = GET2(slot, 0) << 1;
1373            while (slotB > md->name_table)            condition = offset < offset_top && md->offset_vector[offset] >= 0;
1374              {            if (condition) break;
1375              slotB -= md->name_entry_size;            slot += 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);  
       }  
   
     /* The condition is an assertion. Call match() to evaluate it - setting  
     md->match_function_type to MATCH_CONDASSERT causes it to stop at the end of  
     an assertion. */  
1382    
1383      else        /* 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
1385          of an assertion. */
1386    
1387          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 */
1416      use tail recursion to avoid using another stack frame, except when there is  
1417      unlimited repeat of a possibly empty group. In the latter case, a recursive      ecode += condition? PRIV(OP_lengths)[condcode] : codelink;
1418      call to match() is always required, unless the second alternative doesn't  
1419      exist, in which case we can just plough on. Note that, for compatibility      /* We are now at the branch that is to be obeyed. As there is only one, we
1420      with Perl, the | in a conditional group is NOT treated as creating two      can use tail recursion to avoid using another stack frame, except when
1421      alternatives. If a THEN is encountered in the branch, it propagates out to      there is unlimited repeat of a possibly empty group. In the latter case, a
1422      the enclosing alternative (unless nested in a deeper set of alternatives,      recursive call to match() is always required, unless the second alternative
1423      of course). */      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      if (condition || *ecode == OP_ALT)      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 2656  for (;;) Line 2574  for (;;)
2574            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2575          break;          break;
2576    
2577            /* Perl space used to exclude VT, but from Perl 5.18 it is included,
2578            which means that Perl space and POSIX space are now identical. PCRE
2579            was changed at release 8.34. */
2580    
2581          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
         if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||  
              c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)  
                == (op == OP_NOTPROP))  
           RRETURN(MATCH_NOMATCH);  
         break;  
   
2582          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2583          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2584               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
# Line 2742  for (;;) Line 2658  for (;;)
2658      similar code to character type repeats - written out again for speed.      similar code to character type repeats - written out again for speed.
2659      However, if the referenced string is the empty string, always treat      However, if the referenced string is the empty string, always treat
2660      it as matched, any number of times (otherwise there could be infinite      it as matched, any number of times (otherwise there could be infinite
2661      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:  
2662    
2663      (a) In the default, Perl-compatible state, set the length negative;      (a) In the default, Perl-compatible state, set the length negative;
2664      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 2668  for (;;)
2668      so that the back reference matches an empty string.      so that the back reference matches an empty string.
2669    
2670      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
2671      referenced subpattern. */      referenced subpattern.
2672    
2673        The OP_REF and OP_REFI opcodes are used for a reference to a numbered group
2674        or to a non-duplicated named group. For a duplicated named group, OP_DNREF
2675        and OP_DNREFI are used. In this case we must scan the list of groups to
2676        which the name refers, and use the first one that is set. */
2677    
2678        case OP_DNREF:
2679        case OP_DNREFI:
2680        caseless = op == OP_DNREFI;
2681          {
2682          int count = GET2(ecode, 1+IMM2_SIZE);
2683          pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
2684          ecode += 1 + 2*IMM2_SIZE;
2685    
2686          while (count-- > 0)
2687            {
2688            offset = GET2(slot, 0) << 1;
2689            if (offset < offset_top && md->offset_vector[offset] >= 0) break;
2690            slot += md->name_entry_size;
2691            }
2692          if (count < 0)
2693            length = (md->jscript_compat)? 0 : -1;
2694          else
2695            length = md->offset_vector[offset+1] - md->offset_vector[offset];
2696          }
2697        goto REF_REPEAT;
2698    
2699        case OP_REF:
2700        case OP_REFI:
2701        caseless = op == OP_REFI;
2702        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2703        ecode += 1 + IMM2_SIZE;
2704      if (offset >= offset_top || md->offset_vector[offset] < 0)      if (offset >= offset_top || md->offset_vector[offset] < 0)
2705        length = (md->jscript_compat)? 0 : -1;        length = (md->jscript_compat)? 0 : -1;
2706      else      else
# Line 2769  for (;;) Line 2708  for (;;)
2708    
2709      /* Set up for repetition, or handle the non-repeated case */      /* Set up for repetition, or handle the non-repeated case */
2710    
2711        REF_REPEAT:
2712      switch (*ecode)      switch (*ecode)
2713        {        {
2714        case OP_CRSTAR:        case OP_CRSTAR:
# Line 3590  for (;;) Line 3530  for (;;)
3530            if (fc != cc && foc != cc) break;            if (fc != cc && foc != cc) break;
3531            eptr++;            eptr++;
3532            }            }
   
3533          if (possessive) continue;       /* No backtracking */          if (possessive) continue;       /* No backtracking */
3534          for (;;)          for (;;)
3535            {            {
# Line 3599  for (;;) Line 3538  for (;;)
3538            eptr--;            eptr--;
3539            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3540            }            }
3541          RRETURN(MATCH_NOMATCH);          /* Control never gets here */
3542          }          }
       /* Control never gets here */  
3543        }        }
3544    
3545      /* Caseful comparisons (includes all multi-byte characters) */      /* Caseful comparisons (includes all multi-byte characters) */
# Line 3657  for (;;) Line 3595  for (;;)
3595            eptr--;            eptr--;
3596            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3597            }            }
3598          RRETURN(MATCH_NOMATCH);          /* Control never gets here */
3599          }          }
3600        }        }
3601      /* Control never gets here */      /* Control never gets here */
# Line 3942  for (;;) Line 3880  for (;;)
3880              eptr--;              eptr--;
3881              }              }
3882            }            }
3883            /* Control never gets here */
         RRETURN(MATCH_NOMATCH);  
3884          }          }
       /* Control never gets here */  
3885        }        }
3886    
3887      /* Caseful comparisons */      /* Caseful comparisons */
# Line 4079  for (;;) Line 4015  for (;;)
4015              eptr--;              eptr--;
4016              }              }
4017            }            }
4018            /* Control never gets here */
         RRETURN(MATCH_NOMATCH);  
4019          }          }
4020        }        }
4021      /* Control never gets here */      /* Control never gets here */
# Line 4262  for (;;) Line 4197  for (;;)
4197              }              }
4198            break;            break;
4199    
4200              /* Perl space used to exclude VT, but from Perl 5.18 it is included,
4201              which means that Perl space and POSIX space are now identical. PCRE
4202              was changed at release 8.34. */
4203    
4204            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
           for (i = 1; i <= min; i++)  
             {  
             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);  
             }  
           break;  
   
4205            case PT_PXSPACE:  /* POSIX space */            case PT_PXSPACE:  /* POSIX space */
4206            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4207              {              {
# Line 5010  for (;;) Line 4934  for (;;)
4934              }              }
4935            /* Control never gets here */            /* Control never gets here */
4936    
4937              /* Perl space used to exclude VT, but from Perl 5.18 it is included,
4938              which means that Perl space and POSIX space are now identical. PCRE
4939              was changed at release 8.34. */
4940    
4941            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
           for (fi = min;; fi++)  
             {  
             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 */  
   
4942            case PT_PXSPACE:  /* POSIX space */            case PT_PXSPACE:  /* POSIX space */
4943            for (fi = min;; fi++)            for (fi = min;; fi++)
4944              {              {
# Line 5528  for (;;) Line 5438  for (;;)
5438              }              }
5439            break;            break;
5440    
5441              /* Perl space used to exclude VT, but from Perl 5.18 it is included,
5442              which means that Perl space and POSIX space are now identical. PCRE
5443              was changed at release 8.34. */
5444    
5445            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
           for (i = min; i < max; i++)  
             {  
             int len = 1;  
             if (eptr >= md->end_subject)  
               {  
               SCHECK_PARTIAL();  
               break;  
               }  
             GETCHARLENTEST(c, eptr, len);  
             if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||  
                  c == CHAR_FF || c == CHAR_CR)  
                  == prop_fail_result)  
               break;  
             eptr+= len;  
             }  
           break;  
   
5446            case PT_PXSPACE:  /* POSIX space */            case PT_PXSPACE:  /* POSIX space */
5447            for (i = min; i < max; i++)            for (i = min; i < max; i++)
5448              {              {
# Line 5642  for (;;) Line 5539  for (;;)
5539            }            }
5540          }          }
5541    
5542        /* Match extended Unicode sequences. We will get here only if the        /* Match extended Unicode grapheme clusters. We will get here only if the
5543        support is in the binary; otherwise a compile-time error occurs. */        support is in the binary; otherwise a compile-time error occurs. */
5544    
5545        else if (ctype == OP_EXTUNI)        else if (ctype == OP_EXTUNI)
# Line 5675  for (;;) Line 5572  for (;;)
5572          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
5573    
5574          if (possessive) continue;    /* No backtracking */          if (possessive) continue;    /* No backtracking */
5575    
5576          for(;;)          for(;;)
5577            {            {
5578            if (eptr == pp) goto TAIL_RECURSE;            int lgb, rgb;
5579              PCRE_PUCHAR fptr;
5580    
5581              if (eptr == pp) goto TAIL_RECURSE;   /* At start of char run */
5582            RMATCH(eptr, ecode, offset_top, md, eptrb, RM45);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM45);
5583            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5584    
5585              /* Backtracking over an extended grapheme cluster involves inspecting
5586              the previous two characters (if present) to see if a break is
5587              permitted between them. */
5588    
5589            eptr--;            eptr--;
5590            for (;;)                        /* Move back over one extended */            if (!utf) c = *eptr; else
5591              {              {
5592              if (!utf) c = *eptr; else              BACKCHAR(eptr);
5593                {              GETCHAR(c, eptr);
5594                BACKCHAR(eptr);              }
5595                GETCHAR(c, eptr);            rgb = UCD_GRAPHBREAK(c);
5596                }  
5597              if (UCD_CATEGORY(c) != ucp_M) break;            for (;;)
5598              eptr--;              {
5599                if (eptr == pp) goto TAIL_RECURSE;   /* At start of char run */
5600                fptr = eptr - 1;
5601                if (!utf) c = *fptr; else
5602                  {
5603                  BACKCHAR(fptr);
5604                  GETCHAR(c, fptr);
5605                  }
5606                lgb = UCD_GRAPHBREAK(c);
5607                if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5608                eptr = fptr;
5609                rgb = lgb;
5610              }              }
5611            }            }
5612          }          }
# Line 6210  for (;;) Line 6127  for (;;)
6127                eptr[-1] == CHAR_CR) eptr--;                eptr[-1] == CHAR_CR) eptr--;
6128            }            }
6129          }          }
6130    
6131        /* Get here if we can't make it match with any permitted repetitions */        /* Control never gets here */
   
       RRETURN(MATCH_NOMATCH);  
6132        }        }
     /* Control never gets here */  
6133    
6134      /* 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
6135      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 6410  const pcre_uint8 *start_bits = NULL; Line 6324  const pcre_uint8 *start_bits = NULL;
6324  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;
6325  PCRE_PUCHAR end_subject;  PCRE_PUCHAR end_subject;
6326  PCRE_PUCHAR start_partial = NULL;  PCRE_PUCHAR start_partial = NULL;
6327  PCRE_PUCHAR match_partial;  PCRE_PUCHAR match_partial = NULL;
6328  PCRE_PUCHAR req_char_ptr = start_match - 1;  PCRE_PUCHAR req_char_ptr = start_match - 1;
6329    
6330  const pcre_study_data *study;  const pcre_study_data *study;
# Line 7178  if (rc != MATCH_NOMATCH && rc != PCRE_ER Line 7092  if (rc != MATCH_NOMATCH && rc != PCRE_ER
7092    
7093  /* Handle partial matches - disable any mark data */  /* Handle partial matches - disable any mark data */
7094    
7095  if (start_partial != NULL)  if (match_partial != NULL)
7096    {    {
7097    DPRINTF((">>>> returning PCRE_ERROR_PARTIAL\n"));    DPRINTF((">>>> returning PCRE_ERROR_PARTIAL\n"));
7098    md->mark = NULL;    md->mark = NULL;

Legend:
Removed from v.1325  
changed lines
  Added in v.1365

  ViewVC Help
Powered by ViewVC 1.1.5