/[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 428 by ph10, Mon Aug 31 17:10:26 2009 UTC revision 455 by ph10, Sat Sep 26 19:12:32 2009 UTC
# Line 408  immediately. The second one is used when Line 408  immediately. The second one is used when
408  the subject. */  the subject. */
409    
410  #define CHECK_PARTIAL()\  #define CHECK_PARTIAL()\
411    if (md->partial && eptr >= md->end_subject && eptr > mstart)\    if (md->partial != 0 && eptr >= md->end_subject && eptr > mstart)\
412      {\      {\
413      md->hitend = TRUE;\      md->hitend = TRUE;\
414      if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);\      if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);\
# Line 663  for (;;) Line 663  for (;;)
663    {    {
664    minimize = possessive = FALSE;    minimize = possessive = FALSE;
665    op = *ecode;    op = *ecode;
666    
667    switch(op)    switch(op)
668      {      {
669      case OP_FAIL:      case OP_FAIL:
# Line 909  for (;;) Line 909  for (;;)
909        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
910        }        }
911      break;      break;
912    
913    
914        /* Before OP_ACCEPT there may be any number of OP_CLOSE opcodes,
915        to close any currently open capturing brackets. */
916    
917        case OP_CLOSE:
918        number = GET2(ecode, 1);
919        offset = number << 1;
920    
921    #ifdef DEBUG
922          printf("end bracket %d at *ACCEPT", number);
923          printf("\n");
924    #endif
925    
926        md->capture_last = number;
927        if (offset >= md->offset_max) md->offset_overflow = TRUE; else
928          {
929          md->offset_vector[offset] =
930            md->offset_vector[md->offset_end - number];
931          md->offset_vector[offset+1] = eptr - md->start_subject;
932          if (offset_top <= offset) offset_top = offset + 2;
933          }
934        ecode += 3;
935        break;
936    
937    
938      /* End of the pattern, either real or forced. If we are in a top-level      /* End of the pattern, either real or forced. If we are in a top-level
# Line 924  for (;;) Line 948  for (;;)
948        md->recursive = rec->prevrec;        md->recursive = rec->prevrec;
949        memmove(md->offset_vector, rec->offset_save,        memmove(md->offset_vector, rec->offset_save,
950          rec->saved_max * sizeof(int));          rec->saved_max * sizeof(int));
951          offset_top = rec->offset_top;
952        mstart = rec->save_start;        mstart = rec->save_start;
953        ims = original_ims;        ims = original_ims;
954        ecode = rec->after_call;        ecode = rec->after_call;
955        break;        break;
956        }        }
957    
958      /* Otherwise, if PCRE_NOTEMPTY is set, fail if we have matched an empty      /* Otherwise, if we have matched an empty string, fail if PCRE_NOTEMPTY is
959      string - backtracking will then try other alternatives, if any. */      set, or if PCRE_NOTEMPTY_ATSTART is set and we have matched at the start of
960        the subject. In both cases, backtracking will then try other alternatives,
961        if any. */
962    
963        if (eptr == mstart &&
964            (md->notempty ||
965              (md->notempty_atstart &&
966                mstart == md->start_subject + md->start_offset)))
967          RRETURN(MATCH_NOMATCH);
968    
969        /* Otherwise, we have a match. */
970    
     if (md->notempty && eptr == mstart) RRETURN(MATCH_NOMATCH);  
971      md->end_match_ptr = eptr;           /* Record where we ended */      md->end_match_ptr = eptr;           /* Record where we ended */
972      md->end_offset_top = offset_top;    /* and how many extracts were taken */      md->end_offset_top = offset_top;    /* and how many extracts were taken */
973      md->start_match_ptr = mstart;       /* and the start (\K can modify) */      md->start_match_ptr = mstart;       /* and the start (\K can modify) */
# Line 1024  for (;;) Line 1058  for (;;)
1058        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1059        }        }
1060    
1061      /* Skip to next op code */      /* Save the earliest consulted character, then skip to next op code */
1062    
1063        if (eptr < md->start_used_ptr) md->start_used_ptr = eptr;
1064      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1065      break;      break;
1066    
# Line 1105  for (;;) Line 1140  for (;;)
1140        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1141              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1142        new_recursive.save_start = mstart;        new_recursive.save_start = mstart;
1143          new_recursive.offset_top = offset_top;
1144        mstart = eptr;        mstart = eptr;
1145    
1146        /* OK, now we can do the recursion. For each top-level alternative we        /* OK, now we can do the recursion. For each top-level alternative we
# Line 1303  for (;;) Line 1339  for (;;)
1339        {        {
1340        number = GET2(prev, 1+LINK_SIZE);        number = GET2(prev, 1+LINK_SIZE);
1341        offset = number << 1;        offset = number << 1;
1342    
1343  #ifdef DEBUG  #ifdef DEBUG
1344        printf("end bracket %d", number);        printf("end bracket %d", number);
1345        printf("\n");        printf("\n");
# Line 1329  for (;;) Line 1365  for (;;)
1365          mstart = rec->save_start;          mstart = rec->save_start;
1366          memcpy(md->offset_vector, rec->offset_save,          memcpy(md->offset_vector, rec->offset_save,
1367            rec->saved_max * sizeof(int));            rec->saved_max * sizeof(int));
1368            offset_top = rec->offset_top;
1369          ecode = rec->after_call;          ecode = rec->after_call;
1370          ims = original_ims;          ims = original_ims;
1371          break;          break;
# Line 1468  for (;;) Line 1505  for (;;)
1505    
1506        /* Find out if the previous and current characters are "word" characters.        /* Find out if the previous and current characters are "word" characters.
1507        It takes a bit more work in UTF-8 mode. Characters > 255 are assumed to        It takes a bit more work in UTF-8 mode. Characters > 255 are assumed to
1508        be "non-word" characters. */        be "non-word" characters. Remember the earliest consulted character for
1509          partial matching. */
1510    
1511  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
1512        if (utf8)        if (utf8)
# Line 1477  for (;;) Line 1515  for (;;)
1515            {            {
1516            USPTR lastptr = eptr - 1;            USPTR lastptr = eptr - 1;
1517            while((*lastptr & 0xc0) == 0x80) lastptr--;            while((*lastptr & 0xc0) == 0x80) lastptr--;
1518              if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
1519            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
1520            prev_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0;            prev_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0;
1521            }            }
1522          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
1523            {            {
1524            SCHECK_PARTIAL();            SCHECK_PARTIAL();
1525            cur_is_word = FALSE;            cur_is_word = FALSE;
1526            }            }
1527          else          else
1528            {            {
# Line 1497  for (;;) Line 1536  for (;;)
1536        /* Not in UTF-8 mode */        /* Not in UTF-8 mode */
1537    
1538          {          {
1539          prev_is_word = (eptr != md->start_subject) &&          if (eptr == md->start_subject) prev_is_word = FALSE; else
1540            ((md->ctypes[eptr[-1]] & ctype_word) != 0);            {
1541          if (eptr >= md->end_subject)            if (eptr <= md->start_used_ptr) md->start_used_ptr = eptr - 1;
1542              prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);
1543              }
1544            if (eptr >= md->end_subject)
1545            {            {
1546            SCHECK_PARTIAL();            SCHECK_PARTIAL();
1547            cur_is_word = FALSE;            cur_is_word = FALSE;
1548            }            }
1549          else cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          else cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);
1550          }          }
# Line 1522  for (;;) Line 1564  for (;;)
1564      /* Fall through */      /* Fall through */
1565    
1566      case OP_ALLANY:      case OP_ALLANY:
1567      if (eptr++ >= md->end_subject)      if (eptr++ >= md->end_subject)
1568        {        {
1569        SCHECK_PARTIAL();        SCHECK_PARTIAL();
1570        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1571        }        }
1572      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
1573      ecode++;      ecode++;
1574      break;      break;
# Line 1535  for (;;) Line 1577  for (;;)
1577      any byte, even newline, independent of the setting of PCRE_DOTALL. */      any byte, even newline, independent of the setting of PCRE_DOTALL. */
1578    
1579      case OP_ANYBYTE:      case OP_ANYBYTE:
1580      if (eptr++ >= md->end_subject)      if (eptr++ >= md->end_subject)
1581        {        {
1582        SCHECK_PARTIAL();        SCHECK_PARTIAL();
1583        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1584        }        }
1585      ecode++;      ecode++;
1586      break;      break;
1587    
1588      case OP_NOT_DIGIT:      case OP_NOT_DIGIT:
1589      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
1590        {        {
1591        SCHECK_PARTIAL();        SCHECK_PARTIAL();
1592        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1593        }        }
1594      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1595      if (      if (
1596  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1561  for (;;) Line 1603  for (;;)
1603      break;      break;
1604    
1605      case OP_DIGIT:      case OP_DIGIT:
1606      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
1607        {        {
1608        SCHECK_PARTIAL();        SCHECK_PARTIAL();
1609        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1610        }        }
1611      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1612      if (      if (
1613  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1578  for (;;) Line 1620  for (;;)
1620      break;      break;
1621    
1622      case OP_NOT_WHITESPACE:      case OP_NOT_WHITESPACE:
1623      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
1624        {        {
1625        SCHECK_PARTIAL();        SCHECK_PARTIAL();
1626        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1627        }        }
1628      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1629      if (      if (
1630  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1595  for (;;) Line 1637  for (;;)
1637      break;      break;
1638    
1639      case OP_WHITESPACE:      case OP_WHITESPACE:
1640      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
1641        {        {
1642        SCHECK_PARTIAL();        SCHECK_PARTIAL();
1643        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1644        }        }
1645      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1646      if (      if (
1647  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1612  for (;;) Line 1654  for (;;)
1654      break;      break;
1655    
1656      case OP_NOT_WORDCHAR:      case OP_NOT_WORDCHAR:
1657      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
1658        {        {
1659        SCHECK_PARTIAL();        SCHECK_PARTIAL();
1660        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1661        }        }
1662      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1663      if (      if (
1664  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1629  for (;;) Line 1671  for (;;)
1671      break;      break;
1672    
1673      case OP_WORDCHAR:      case OP_WORDCHAR:
1674      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
1675        {        {
1676        SCHECK_PARTIAL();        SCHECK_PARTIAL();
1677        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1678        }        }
1679      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1680      if (      if (
1681  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1646  for (;;) Line 1688  for (;;)
1688      break;      break;
1689    
1690      case OP_ANYNL:      case OP_ANYNL:
1691      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
1692        {        {
1693        SCHECK_PARTIAL();        SCHECK_PARTIAL();
1694        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1695        }        }
1696      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1697      switch(c)      switch(c)
1698        {        {
# Line 1674  for (;;) Line 1716  for (;;)
1716      break;      break;
1717    
1718      case OP_NOT_HSPACE:      case OP_NOT_HSPACE:
1719      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
1720        {        {
1721        SCHECK_PARTIAL();        SCHECK_PARTIAL();
1722        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1723        }        }
1724      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1725      switch(c)      switch(c)
1726        {        {
# Line 1708  for (;;) Line 1750  for (;;)
1750      break;      break;
1751    
1752      case OP_HSPACE:      case OP_HSPACE:
1753      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
1754        {        {
1755        SCHECK_PARTIAL();        SCHECK_PARTIAL();
1756        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1757        }        }
1758      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1759      switch(c)      switch(c)
1760        {        {
# Line 1742  for (;;) Line 1784  for (;;)
1784      break;      break;
1785    
1786      case OP_NOT_VSPACE:      case OP_NOT_VSPACE:
1787      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
1788        {        {
1789        SCHECK_PARTIAL();        SCHECK_PARTIAL();
1790        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1791        }        }
1792      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1793      switch(c)      switch(c)
1794        {        {
# Line 1764  for (;;) Line 1806  for (;;)
1806      break;      break;
1807    
1808      case OP_VSPACE:      case OP_VSPACE:
1809      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
1810        {        {
1811        SCHECK_PARTIAL();        SCHECK_PARTIAL();
1812        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1813        }        }
1814      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1815      switch(c)      switch(c)
1816        {        {
# Line 1791  for (;;) Line 1833  for (;;)
1833    
1834      case OP_PROP:      case OP_PROP:
1835      case OP_NOTPROP:      case OP_NOTPROP:
1836      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
1837        {        {
1838        SCHECK_PARTIAL();        SCHECK_PARTIAL();
1839        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1840        }        }
1841      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1842        {        {
1843        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
# Line 1840  for (;;) Line 1882  for (;;)
1882      is in the binary; otherwise a compile-time error occurs. */      is in the binary; otherwise a compile-time error occurs. */
1883    
1884      case OP_EXTUNI:      case OP_EXTUNI:
1885      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
1886        {        {
1887        SCHECK_PARTIAL();        SCHECK_PARTIAL();
1888        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1889        }        }
1890      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1891        {        {
1892        int category = UCD_CATEGORY(c);        int category = UCD_CATEGORY(c);
# Line 1924  for (;;) Line 1966  for (;;)
1966          break;          break;
1967    
1968          default:               /* No repeat follows */          default:               /* No repeat follows */
1969          if (!match_ref(offset, eptr, length, md, ims))          if (!match_ref(offset, eptr, length, md, ims))
1970            {            {
1971            CHECK_PARTIAL();            CHECK_PARTIAL();
1972            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1973            }            }
1974          eptr += length;          eptr += length;
1975          continue;              /* With the main loop */          continue;              /* With the main loop */
1976          }          }
1977    
1978        /* If the length of the reference is zero, just continue with the        /* If the length of the reference is zero, just continue with the
1979        main loop. */        main loop. */
1980    
1981        if (length == 0) continue;        if (length == 0) continue;
1982    
1983        /* First, ensure the minimum number of matches are present. We get back        /* First, ensure the minimum number of matches are present. We get back
# Line 2311  for (;;) Line 2353  for (;;)
2353        length = 1;        length = 1;
2354        ecode++;        ecode++;
2355        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
2356        if (length > md->end_subject - eptr)        if (length > md->end_subject - eptr)
2357          {          {
2358          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
2359          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2360          }          }
2361        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);
2362        }        }
2363      else      else
# Line 2323  for (;;) Line 2365  for (;;)
2365    
2366      /* Non-UTF-8 mode */      /* Non-UTF-8 mode */
2367        {        {
2368        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
2369          {          {
2370          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
2371          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2372          }          }
2373        if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH);        if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH);
2374        ecode += 2;        ecode += 2;
2375        }        }
# Line 2343  for (;;) Line 2385  for (;;)
2385        ecode++;        ecode++;
2386        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
2387    
2388        if (length > md->end_subject - eptr)        if (length > md->end_subject - eptr)
2389          {          {
2390          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
2391          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2392          }          }
2393    
2394        /* If the pattern character's value is < 128, we have only one byte, and        /* If the pattern character's value is < 128, we have only one byte, and
2395        can use the fast lookup table. */        can use the fast lookup table. */
# Line 2382  for (;;) Line 2424  for (;;)
2424    
2425      /* Non-UTF-8 mode */      /* Non-UTF-8 mode */
2426        {        {
2427        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
2428          {          {
2429          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
2430          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2431          }          }
2432        if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);        if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);
2433        ecode += 2;        ecode += 2;
2434        }        }
# Line 2440  for (;;) Line 2482  for (;;)
2482      case OP_MINQUERY:      case OP_MINQUERY:
2483      c = *ecode++ - OP_STAR;      c = *ecode++ - OP_STAR;
2484      minimize = (c & 1) != 0;      minimize = (c & 1) != 0;
2485    
2486      min = rep_min[c];                 /* Pick up values from tables; */      min = rep_min[c];                 /* Pick up values from tables; */
2487      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
2488      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
# Line 2553  for (;;) Line 2595  for (;;)
2595      /* When not in UTF-8 mode, load a single-byte character. */      /* When not in UTF-8 mode, load a single-byte character. */
2596    
2597      fc = *ecode++;      fc = *ecode++;
2598    
2599      /* The value of fc at this point is always less than 256, though we may or      /* The value of fc at this point is always less than 256, though we may or
2600      may not be in UTF-8 mode. The code is duplicated for the caseless and      may not be in UTF-8 mode. The code is duplicated for the caseless and
2601      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
# Line 2630  for (;;) Line 2672  for (;;)
2672            }            }
2673          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
2674          }          }
2675    
2676        if (min == max) continue;        if (min == max) continue;
2677    
2678        if (minimize)        if (minimize)
2679          {          {
2680          for (fi = min;; fi++)          for (fi = min;; fi++)
# Line 2658  for (;;) Line 2700  for (;;)
2700            eptr++;            eptr++;
2701            }            }
2702          if (possessive) continue;          if (possessive) continue;
2703    
2704          while (eptr >= pp)          while (eptr >= pp)
2705            {            {
2706            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27);
# Line 2674  for (;;) Line 2716  for (;;)
2716      checking can be multibyte. */      checking can be multibyte. */
2717    
2718      case OP_NOT:      case OP_NOT:
2719      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2720        {        {
2721        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2722        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2723        }        }
2724      ecode++;      ecode++;
2725      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2726      if ((ims & PCRE_CASELESS) != 0)      if ((ims & PCRE_CASELESS) != 0)
# Line 3524  for (;;) Line 3566  for (;;)
3566          break;          break;
3567    
3568          case OP_ALLANY:          case OP_ALLANY:
3569          if (eptr > md->end_subject - min)          if (eptr > md->end_subject - min)
3570            {            {
3571            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3572            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3573            }            }
3574          eptr += min;          eptr += min;
3575          break;          break;
3576    
3577          case OP_ANYBYTE:          case OP_ANYBYTE:
3578          if (eptr > md->end_subject - min)          if (eptr > md->end_subject - min)
3579            {            {
3580            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3581            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3582            }            }
3583          eptr += min;          eptr += min;
3584          break;          break;
3585    
# Line 4914  md->jscript_compat = (re->options & PCRE Line 4956  md->jscript_compat = (re->options & PCRE
4956  md->notbol = (options & PCRE_NOTBOL) != 0;  md->notbol = (options & PCRE_NOTBOL) != 0;
4957  md->noteol = (options & PCRE_NOTEOL) != 0;  md->noteol = (options & PCRE_NOTEOL) != 0;
4958  md->notempty = (options & PCRE_NOTEMPTY) != 0;  md->notempty = (options & PCRE_NOTEMPTY) != 0;
4959    md->notempty_atstart = (options & PCRE_NOTEMPTY_ATSTART) != 0;
4960  md->partial = ((options & PCRE_PARTIAL_HARD) != 0)? 2 :  md->partial = ((options & PCRE_PARTIAL_HARD) != 0)? 2 :
4961                ((options & PCRE_PARTIAL_SOFT) != 0)? 1 : 0;                ((options & PCRE_PARTIAL_SOFT) != 0)? 1 : 0;
4962  md->hitend = FALSE;  md->hitend = FALSE;
# Line 5078  if (!anchored) Line 5121  if (!anchored)
5121      }      }
5122    else    else
5123      if (!startline && study != NULL &&      if (!startline && study != NULL &&
5124        (study->options & PCRE_STUDY_MAPPED) != 0)        (study->flags & PCRE_STUDY_MAPPED) != 0)
5125          start_bits = study->start_bits;          start_bits = study->start_bits;
5126    }    }
5127    
# Line 5204  for(;;) Line 5247  for(;;)
5247    /* Restore fudged end_subject */    /* Restore fudged end_subject */
5248    
5249    end_subject = save_end_subject;    end_subject = save_end_subject;
5250    
5251  #ifdef DEBUG  /* Sigh. Some compilers never learn. */    /* The following two optimizations are disabled for partial matching or if
   printf(">>>> Match against: ");  
   pchars(start_match, end_subject - start_match, TRUE, md);  
   printf("\n");  
 #endif  
   
   /* If req_byte is set, we know that that character must appear in the  
   subject for the match to succeed. If the first character is set, req_byte  
   must be later in the subject; otherwise the test starts at the match point.  
   This optimization can save a huge amount of backtracking in patterns with  
   nested unlimited repeats that aren't going to match. Writing separate code  
   for cased/caseless versions makes it go faster, as does using an  
   autoincrement and backing off on a match.  
   
   HOWEVER: when the subject string is very, very long, searching to its end  
   can take a long time, and give bad performance on quite ordinary patterns.  
   This showed up when somebody was matching something like /^\d+C/ on a  
   32-megabyte string... so we don't do this when the string is sufficiently  
   long.  
   
   ALSO: this processing is disabled when partial matching is requested, or if  
5252    disabling is explicitly requested. */    disabling is explicitly requested. */
5253    
5254    if ((options & PCRE_NO_START_OPTIMIZE) == 0 &&    if ((options & PCRE_NO_START_OPTIMIZE) == 0 && !md->partial)
5255        req_byte >= 0 &&      {
5256        end_subject - start_match < REQ_BYTE_MAX &&      /* If the pattern was studied, a minimum subject length may be set. This is
5257        !md->partial)      a lower bound; no actual string of that length may actually match the
5258      {      pattern. Although the value is, strictly, in characters, we treat it as
5259      register USPTR p = start_match + ((first_byte >= 0)? 1 : 0);      bytes to avoid spending too much time in this optimization. */
5260    
5261      /* We don't need to repeat the search if we haven't yet reached the      if (study != NULL && (study->flags & PCRE_STUDY_MINLEN) != 0 &&
5262      place we found it at last time. */          end_subject - start_match < study->minlength)
5263          {
5264      if (p > req_byte_ptr)        rc = MATCH_NOMATCH;
5265          break;
5266          }
5267    
5268        /* If req_byte is set, we know that that character must appear in the
5269        subject for the match to succeed. If the first character is set, req_byte
5270        must be later in the subject; otherwise the test starts at the match point.
5271        This optimization can save a huge amount of backtracking in patterns with
5272        nested unlimited repeats that aren't going to match. Writing separate code
5273        for cased/caseless versions makes it go faster, as does using an
5274        autoincrement and backing off on a match.
5275    
5276        HOWEVER: when the subject string is very, very long, searching to its end
5277        can take a long time, and give bad performance on quite ordinary patterns.
5278        This showed up when somebody was matching something like /^\d+C/ on a
5279        32-megabyte string... so we don't do this when the string is sufficiently
5280        long. */
5281    
5282        if (req_byte >= 0 && end_subject - start_match < REQ_BYTE_MAX)
5283        {        {
5284        if (req_byte_caseless)        register USPTR p = start_match + ((first_byte >= 0)? 1 : 0);
5285    
5286          /* We don't need to repeat the search if we haven't yet reached the
5287          place we found it at last time. */
5288    
5289          if (p > req_byte_ptr)
5290          {          {
5291          while (p < end_subject)          if (req_byte_caseless)
5292            {            {
5293            register int pp = *p++;            while (p < end_subject)
5294            if (pp == req_byte || pp == req_byte2) { p--; break; }              {
5295                register int pp = *p++;
5296                if (pp == req_byte || pp == req_byte2) { p--; break; }
5297                }
5298            }            }
5299          }          else
       else  
         {  
         while (p < end_subject)  
5300            {            {
5301            if (*p++ == req_byte) { p--; break; }            while (p < end_subject)
5302                {
5303                if (*p++ == req_byte) { p--; break; }
5304                }
5305            }            }
5306    
5307            /* If we can't find the required character, break the matching loop,
5308            forcing a match failure. */
5309    
5310            if (p >= end_subject)
5311              {
5312              rc = MATCH_NOMATCH;
5313              break;
5314              }
5315    
5316            /* If we have found the required character, save the point where we
5317            found it, so that we don't search again next time round the loop if
5318            the start hasn't passed this character yet. */
5319    
5320            req_byte_ptr = p;
5321          }          }
   
       /* If we can't find the required character, break the matching loop,  
       forcing a match failure. */  
   
       if (p >= end_subject)  
         {  
         rc = MATCH_NOMATCH;  
         break;  
         }  
   
       /* If we have found the required character, save the point where we  
       found it, so that we don't search again next time round the loop if  
       the start hasn't passed this character yet. */  
   
       req_byte_ptr = p;  
5322        }        }
5323      }      }
5324    
5325    #ifdef DEBUG  /* Sigh. Some compilers never learn. */
5326      printf(">>>> Match against: ");
5327      pchars(start_match, end_subject - start_match, TRUE, md);
5328      printf("\n");
5329    #endif
5330    
5331    /* OK, we can now run the match. If "hitend" is set afterwards, remember the    /* OK, we can now run the match. If "hitend" is set afterwards, remember the
5332    first starting point for which a partial match was found. */    first starting point for which a partial match was found. */
5333    
5334    md->start_match_ptr = start_match;    md->start_match_ptr = start_match;
5335      md->start_used_ptr = start_match;
5336    md->match_call_count = 0;    md->match_call_count = 0;
5337    rc = match(start_match, md->start_code, start_match, 2, md, ims, NULL, 0, 0);    rc = match(start_match, md->start_code, start_match, 2, md, ims, NULL, 0, 0);
5338    if (md->hitend && start_partial == NULL) start_partial = start_match;    if (md->hitend && start_partial == NULL) start_partial = md->start_used_ptr;
5339    
5340    switch(rc)    switch(rc)
5341      {      {
# Line 5309  for(;;) Line 5365  for(;;)
5365      rc = MATCH_NOMATCH;      rc = MATCH_NOMATCH;
5366      goto ENDLOOP;      goto ENDLOOP;
5367    
5368      /* Any other return is some kind of error. */      /* Any other return is either a match, or some kind of error. */
5369    
5370      default:      default:
5371      goto ENDLOOP;      goto ENDLOOP;
# Line 5391  if (rc == MATCH_MATCH) Line 5447  if (rc == MATCH_MATCH)
5447    too many to fit into the vector. */    too many to fit into the vector. */
5448    
5449    rc = md->offset_overflow? 0 : md->end_offset_top/2;    rc = md->offset_overflow? 0 : md->end_offset_top/2;
5450    
5451    /* If there is space, set up the whole thing as substring 0. The value of    /* If there is space, set up the whole thing as substring 0. The value of
5452    md->start_match_ptr might be modified if \K was encountered on the success    md->start_match_ptr might be modified if \K was encountered on the success
5453    matching path. */    matching path. */

Legend:
Removed from v.428  
changed lines
  Added in v.455

  ViewVC Help
Powered by ViewVC 1.1.5