/[pcre]/code/trunk/pcre_compile.c
ViewVC logotype

Diff of /code/trunk/pcre_compile.c

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1073 by chpe, Tue Oct 16 15:54:43 2012 UTC revision 1266 by ph10, Sun Mar 3 11:14:26 2013 UTC
# Line 80  to check them every time. */ Line 80  to check them every time. */
80  /* Definitions to allow mutual recursion */  /* Definitions to allow mutual recursion */
81    
82  static int  static int
83    add_list_to_class(pcre_uint8 *, pcre_uchar **, int, compile_data *,    add_list_to_class(pcre_uint8 *, pcre_uchar **, int, compile_data *,
84      const pcre_uint32 *, unsigned int);      const pcre_uint32 *, unsigned int);
85    
86  static BOOL  static BOOL
87    compile_regex(int, pcre_uchar **, const pcre_uchar **, int *, BOOL, BOOL,    compile_regex(int, pcre_uchar **, const pcre_uchar **, int *, BOOL, BOOL, int, int,
88      int, int, int *, int *, branch_chain *, compile_data *, int *);      pcre_uint32 *, pcre_int32 *, pcre_uint32 *, pcre_int32 *, branch_chain *,
89        compile_data *, int *);
90    
91    
92    
# Line 121  overrun before it actually does run off Line 122  overrun before it actually does run off
122    
123  /* Private flags added to firstchar and reqchar. */  /* Private flags added to firstchar and reqchar. */
124    
125  #define REQ_CASELESS   0x10000000l      /* Indicates caselessness */  #define REQ_CASELESS    (1 << 0)        /* Indicates caselessness */
126  #define REQ_VARY       0x20000000l      /* Reqchar followed non-literal item */  #define REQ_VARY        (1 << 1)        /* Reqchar followed non-literal item */
127  #define REQ_MASK       (REQ_CASELESS | REQ_VARY)  /* Negative values for the firstchar and reqchar flags */
128    #define REQ_UNSET       (-2)
129    #define REQ_NONE        (-1)
130    
131  /* Repeated character flags. */  /* Repeated character flags. */
132    
# Line 645  static const pcre_uint8 ebcdic_chartab[] Line 648  static const pcre_uint8 ebcdic_chartab[]
648    
649    
650    
   
651  /*************************************************  /*************************************************
652  *            Find an error text                  *  *            Find an error text                  *
653  *************************************************/  *************************************************/
# Line 665  find_error_text(int n) Line 667  find_error_text(int n)
667  const char *s = error_texts;  const char *s = error_texts;
668  for (; n > 0; n--)  for (; n > 0; n--)
669    {    {
670    while (*s++ != 0) {};    while (*s++ != CHAR_NULL) {};
671    if (*s == 0) return "Error text not found (please report)";    if (*s == CHAR_NULL) return "Error text not found (please report)";
672    }    }
673  return s;  return s;
674  }  }
# Line 771  Returns:         zero => a data characte Line 773  Returns:         zero => a data characte
773  */  */
774    
775  static int  static int
776  check_escape(const pcre_uchar **ptrptr, pcre_uint32 *chptr, int *errorcodeptr,  check_escape(const pcre_uchar **ptrptr, pcre_uint32 *chptr, int *errorcodeptr,
777    int bracount, int options, BOOL isclass)    int bracount, int options, BOOL isclass)
778  {  {
779  /* PCRE_UTF16 has the same value as PCRE_UTF8. */  /* PCRE_UTF16 has the same value as PCRE_UTF8. */
# Line 786  ptr--;                            /* Set Line 788  ptr--;                            /* Set
788    
789  /* If backslash is at the end of the pattern, it's an error. */  /* If backslash is at the end of the pattern, it's an error. */
790    
791  if (c == 0) *errorcodeptr = ERR1;  if (c == CHAR_NULL) *errorcodeptr = ERR1;
792    
793  /* Non-alphanumerics are literals. For digits or letters, do an initial lookup  /* Non-alphanumerics are literals. For digits or letters, do an initial lookup
794  in a table. A non-zero result is something that can be returned immediately.  in a table. A non-zero result is something that can be returned immediately.
# Line 795  Otherwise further processing may be requ Line 797  Otherwise further processing may be requ
797  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
798  /* Not alphanumeric */  /* Not alphanumeric */
799  else if (c < CHAR_0 || c > CHAR_z) {}  else if (c < CHAR_0 || c > CHAR_z) {}
800  else if ((i = escapes[c - CHAR_0]) != 0) { if (i > 0) c = (pcre_uint32)i; else escape = -i; }  else if ((i = escapes[c - CHAR_0]) != 0)
801      { if (i > 0) c = (pcre_uint32)i; else escape = -i; }
802    
803  #else           /* EBCDIC coding */  #else           /* EBCDIC coding */
804  /* Not alphanumeric */  /* Not alphanumeric */
# Line 845  else Line 848  else
848            }            }
849    
850  #if defined COMPILE_PCRE8  #if defined COMPILE_PCRE8
851          if (c > (utf ? 0x10ffff : 0xff))          if (c > (utf ? 0x10ffffU : 0xffU))
852  #elif defined COMPILE_PCRE16  #elif defined COMPILE_PCRE16
853          if (c > (utf ? 0x10ffff : 0xffff))          if (c > (utf ? 0x10ffffU : 0xffffU))
854  #elif defined COMPILE_PCRE32  #elif defined COMPILE_PCRE32
855          if (utf && c > 0x10ffff)          if (utf && c > 0x10ffffU)
856  #endif  #endif
857            {            {
858            *errorcodeptr = ERR76;            *errorcodeptr = ERR76;
# Line 896  else Line 899  else
899      if (ptr[1] == CHAR_LEFT_CURLY_BRACKET)      if (ptr[1] == CHAR_LEFT_CURLY_BRACKET)
900        {        {
901        const pcre_uchar *p;        const pcre_uchar *p;
902        for (p = ptr+2; *p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET; p++)        for (p = ptr+2; *p != CHAR_NULL && *p != CHAR_RIGHT_CURLY_BRACKET; p++)
903          if (*p != CHAR_MINUS && !IS_DIGIT(*p)) break;          if (*p != CHAR_MINUS && !IS_DIGIT(*p)) break;
904        if (*p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET)        if (*p != CHAR_NULL && *p != CHAR_RIGHT_CURLY_BRACKET)
905          {          {
906          escape = ESC_k;          escape = ESC_k;
907          break;          break;
# Line 1083  else Line 1086  else
1086  #endif  #endif
1087    
1088  #if defined COMPILE_PCRE8  #if defined COMPILE_PCRE8
1089          if (c > (utf ? 0x10ffff : 0xff)) { overflow = TRUE; break; }          if (c > (utf ? 0x10ffffU : 0xffU)) { overflow = TRUE; break; }
1090  #elif defined COMPILE_PCRE16  #elif defined COMPILE_PCRE16
1091          if (c > (utf ? 0x10ffff : 0xffff)) { overflow = TRUE; break; }          if (c > (utf ? 0x10ffffU : 0xffffU)) { overflow = TRUE; break; }
1092  #elif defined COMPILE_PCRE32  #elif defined COMPILE_PCRE32
1093          if (utf && c > 0x10ffff) { overflow = TRUE; break; }          if (utf && c > 0x10ffffU) { overflow = TRUE; break; }
1094  #endif  #endif
1095          }          }
1096    
# Line 1132  else Line 1135  else
1135    
1136      case CHAR_c:      case CHAR_c:
1137      c = *(++ptr);      c = *(++ptr);
1138      if (c == 0)      if (c == CHAR_NULL)
1139        {        {
1140        *errorcodeptr = ERR2;        *errorcodeptr = ERR2;
1141        break;        break;
# Line 1201  escape sequence. Line 1204  escape sequence.
1204  Argument:  Argument:
1205    ptrptr         points to the pattern position pointer    ptrptr         points to the pattern position pointer
1206    negptr         points to a boolean that is set TRUE for negation else FALSE    negptr         points to a boolean that is set TRUE for negation else FALSE
1207    dptr           points to an int that is set to the detailed property value    ptypeptr       points to an unsigned int that is set to the type value
1208      pdataptr       points to an unsigned int that is set to the detailed property value
1209    errorcodeptr   points to the error code variable    errorcodeptr   points to the error code variable
1210    
1211  Returns:         type value from ucp_type_table, or -1 for an invalid type  Returns:         TRUE if the type value was found, or FALSE for an invalid type
1212  */  */
1213    
1214  static int  static BOOL
1215  get_ucp(const pcre_uchar **ptrptr, BOOL *negptr, int *dptr, int *errorcodeptr)  get_ucp(const pcre_uchar **ptrptr, BOOL *negptr, unsigned int *ptypeptr,
1216      unsigned int *pdataptr, int *errorcodeptr)
1217  {  {
1218  pcre_uchar c;  pcre_uchar c;
1219  int i, bot, top;  int i, bot, top;
# Line 1216  const pcre_uchar *ptr = *ptrptr; Line 1221  const pcre_uchar *ptr = *ptrptr;
1221  pcre_uchar name[32];  pcre_uchar name[32];
1222    
1223  c = *(++ptr);  c = *(++ptr);
1224  if (c == 0) goto ERROR_RETURN;  if (c == CHAR_NULL) goto ERROR_RETURN;
1225    
1226  *negptr = FALSE;  *negptr = FALSE;
1227    
# Line 1233  if (c == CHAR_LEFT_CURLY_BRACKET) Line 1238  if (c == CHAR_LEFT_CURLY_BRACKET)
1238    for (i = 0; i < (int)(sizeof(name) / sizeof(pcre_uchar)) - 1; i++)    for (i = 0; i < (int)(sizeof(name) / sizeof(pcre_uchar)) - 1; i++)
1239      {      {
1240      c = *(++ptr);      c = *(++ptr);
1241      if (c == 0) goto ERROR_RETURN;      if (c == CHAR_NULL) goto ERROR_RETURN;
1242      if (c == CHAR_RIGHT_CURLY_BRACKET) break;      if (c == CHAR_RIGHT_CURLY_BRACKET) break;
1243      name[i] = c;      name[i] = c;
1244      }      }
# Line 1263  while (bot < top) Line 1268  while (bot < top)
1268    r = STRCMP_UC_C8(name, PRIV(utt_names) + PRIV(utt)[i].name_offset);    r = STRCMP_UC_C8(name, PRIV(utt_names) + PRIV(utt)[i].name_offset);
1269    if (r == 0)    if (r == 0)
1270      {      {
1271      *dptr = PRIV(utt)[i].value;      *ptypeptr = PRIV(utt)[i].type;
1272      return PRIV(utt)[i].type;      *pdataptr = PRIV(utt)[i].value;
1273        return TRUE;
1274      }      }
1275    if (r > 0) bot = i + 1; else top = i;    if (r > 0) bot = i + 1; else top = i;
1276    }    }
1277    
1278  *errorcodeptr = ERR47;  *errorcodeptr = ERR47;
1279  *ptrptr = ptr;  *ptrptr = ptr;
1280  return -1;  return FALSE;
1281    
1282  ERROR_RETURN:  ERROR_RETURN:
1283  *errorcodeptr = ERR46;  *errorcodeptr = ERR46;
1284  *ptrptr = ptr;  *ptrptr = ptr;
1285  return -1;  return FALSE;
1286  }  }
1287  #endif  #endif
1288    
# Line 1427  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1433  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1433    
1434    else if (ptr[2] == CHAR_NUMBER_SIGN)    else if (ptr[2] == CHAR_NUMBER_SIGN)
1435      {      {
1436      for (ptr += 3; *ptr != 0; ptr++) if (*ptr == CHAR_RIGHT_PARENTHESIS) break;      for (ptr += 3; *ptr != CHAR_NULL; ptr++)
1437          if (*ptr == CHAR_RIGHT_PARENTHESIS) break;
1438      goto FAIL_EXIT;      goto FAIL_EXIT;
1439      }      }
1440    
# Line 1440  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1447  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1447      ptr += 2;      ptr += 2;
1448      if (ptr[1] != CHAR_QUESTION_MARK)      if (ptr[1] != CHAR_QUESTION_MARK)
1449        {        {
1450        while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;        while (*ptr != CHAR_NULL && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
1451        if (*ptr != 0) ptr++;        if (*ptr != CHAR_NULL) ptr++;
1452        }        }
1453      }      }
1454    
# Line 1484  for (; ptr < cd->end_pattern; ptr++) Line 1491  for (; ptr < cd->end_pattern; ptr++)
1491    
1492    if (*ptr == CHAR_BACKSLASH)    if (*ptr == CHAR_BACKSLASH)
1493      {      {
1494      if (*(++ptr) == 0) goto FAIL_EXIT;      if (*(++ptr) == CHAR_NULL) goto FAIL_EXIT;
1495      if (*ptr == CHAR_Q) for (;;)      if (*ptr == CHAR_Q) for (;;)
1496        {        {
1497        while (*(++ptr) != 0 && *ptr != CHAR_BACKSLASH) {};        while (*(++ptr) != CHAR_NULL && *ptr != CHAR_BACKSLASH) {};
1498        if (*ptr == 0) goto FAIL_EXIT;        if (*ptr == CHAR_NULL) goto FAIL_EXIT;
1499        if (*(++ptr) == CHAR_E) break;        if (*(++ptr) == CHAR_E) break;
1500        }        }
1501      continue;      continue;
# Line 1532  for (; ptr < cd->end_pattern; ptr++) Line 1539  for (; ptr < cd->end_pattern; ptr++)
1539    
1540      while (*(++ptr) != CHAR_RIGHT_SQUARE_BRACKET)      while (*(++ptr) != CHAR_RIGHT_SQUARE_BRACKET)
1541        {        {
1542        if (*ptr == 0) return -1;        if (*ptr == CHAR_NULL) return -1;
1543        if (*ptr == CHAR_BACKSLASH)        if (*ptr == CHAR_BACKSLASH)
1544          {          {
1545          if (*(++ptr) == 0) goto FAIL_EXIT;          if (*(++ptr) == CHAR_NULL) goto FAIL_EXIT;
1546          if (*ptr == CHAR_Q) for (;;)          if (*ptr == CHAR_Q) for (;;)
1547            {            {
1548            while (*(++ptr) != 0 && *ptr != CHAR_BACKSLASH) {};            while (*(++ptr) != CHAR_NULL && *ptr != CHAR_BACKSLASH) {};
1549            if (*ptr == 0) goto FAIL_EXIT;            if (*ptr == CHAR_NULL) goto FAIL_EXIT;
1550            if (*(++ptr) == CHAR_E) break;            if (*(++ptr) == CHAR_E) break;
1551            }            }
1552          continue;          continue;
# Line 1553  for (; ptr < cd->end_pattern; ptr++) Line 1560  for (; ptr < cd->end_pattern; ptr++)
1560    if (xmode && *ptr == CHAR_NUMBER_SIGN)    if (xmode && *ptr == CHAR_NUMBER_SIGN)
1561      {      {
1562      ptr++;      ptr++;
1563      while (*ptr != 0)      while (*ptr != CHAR_NULL)
1564        {        {
1565        if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }        if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
1566        ptr++;        ptr++;
# Line 1561  for (; ptr < cd->end_pattern; ptr++) Line 1568  for (; ptr < cd->end_pattern; ptr++)
1568        if (utf) FORWARDCHAR(ptr);        if (utf) FORWARDCHAR(ptr);
1569  #endif  #endif
1570        }        }
1571      if (*ptr == 0) goto FAIL_EXIT;      if (*ptr == CHAR_NULL) goto FAIL_EXIT;
1572      continue;      continue;
1573      }      }
1574    
# Line 1571  for (; ptr < cd->end_pattern; ptr++) Line 1578  for (; ptr < cd->end_pattern; ptr++)
1578      {      {
1579      int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, count);      int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, count);
1580      if (rc > 0) return rc;      if (rc > 0) return rc;
1581      if (*ptr == 0) goto FAIL_EXIT;      if (*ptr == CHAR_NULL) goto FAIL_EXIT;
1582      }      }
1583    
1584    else if (*ptr == CHAR_RIGHT_PARENTHESIS)    else if (*ptr == CHAR_RIGHT_PARENTHESIS)
# Line 1636  matching closing parens. That is why we Line 1643  matching closing parens. That is why we
1643  for (;;)  for (;;)
1644    {    {
1645    rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, &count);    rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, &count);
1646    if (rc > 0 || *ptr++ == 0) break;    if (rc > 0 || *ptr++ == CHAR_NULL) break;
1647    }    }
1648    
1649  return rc;  return rc;
# Line 1852  for (;;) Line 1859  for (;;)
1859      case OP_NOTI:      case OP_NOTI:
1860      branchlength++;      branchlength++;
1861      cc += 2;      cc += 2;
1862  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
1863      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
1864  #endif  #endif
1865      break;      break;
# Line 1866  for (;;) Line 1873  for (;;)
1873      case OP_NOTEXACTI:      case OP_NOTEXACTI:
1874      branchlength += (int)GET2(cc,1);      branchlength += (int)GET2(cc,1);
1875      cc += 2 + IMM2_SIZE;      cc += 2 + IMM2_SIZE;
1876  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
1877      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
1878  #endif  #endif
1879      break;      break;
1880    
1881      case OP_TYPEEXACT:      case OP_TYPEEXACT:
1882      branchlength += GET2(cc,1);      branchlength += GET2(cc,1);
1883      if (cc[1 + IMM2_SIZE] == OP_PROP || cc[1 + IMM2_SIZE] == OP_NOTPROP)      if (cc[1 + IMM2_SIZE] == OP_PROP || cc[1 + IMM2_SIZE] == OP_NOTPROP)
1884        cc += 2;        cc += 2;
1885      cc += 1 + IMM2_SIZE + 1;      cc += 1 + IMM2_SIZE + 1;
1886      break;      break;
# Line 1909  for (;;) Line 1916  for (;;)
1916    
1917      /* Check a class for variable quantification */      /* Check a class for variable quantification */
1918    
 #if defined SUPPORT_UTF || defined COMPILE_PCRE16 || defined COMPILE_PCRE32  
     case OP_XCLASS:  
     cc += GET(cc, 1) - PRIV(OP_lengths)[OP_CLASS];  
     /* Fall through */  
 #endif  
   
1919      case OP_CLASS:      case OP_CLASS:
1920      case OP_NCLASS:      case OP_NCLASS:
1921    #if defined SUPPORT_UTF || defined COMPILE_PCRE16 || defined COMPILE_PCRE32
1922        case OP_XCLASS:
1923        /* The original code caused an unsigned overflow in 64 bit systems,
1924        so now we use a conditional statement. */
1925        if (op == OP_XCLASS)
1926          cc += GET(cc, 1);
1927        else
1928          cc += PRIV(OP_lengths)[OP_CLASS];
1929    #else
1930      cc += PRIV(OP_lengths)[OP_CLASS];      cc += PRIV(OP_lengths)[OP_CLASS];
1931    #endif
1932    
1933      switch (*cc)      switch (*cc)
1934        {        {
# Line 2112  for (;;) Line 2123  for (;;)
2123        case OP_TYPEMINUPTO:        case OP_TYPEMINUPTO:
2124        case OP_TYPEEXACT:        case OP_TYPEEXACT:
2125        case OP_TYPEPOSUPTO:        case OP_TYPEPOSUPTO:
2126        if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)        if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)
2127          code += 2;          code += 2;
2128        break;        break;
2129    
# Line 2232  for (;;) Line 2243  for (;;)
2243        case OP_TYPEUPTO:        case OP_TYPEUPTO:
2244        case OP_TYPEMINUPTO:        case OP_TYPEMINUPTO:
2245        case OP_TYPEEXACT:        case OP_TYPEEXACT:
2246        if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)        if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)
2247          code += 2;          code += 2;
2248        break;        break;
2249    
# Line 2558  for (code = first_significant_code(code Line 2569  for (code = first_significant_code(code
2569      case OP_TYPEUPTO:      case OP_TYPEUPTO:
2570      case OP_TYPEMINUPTO:      case OP_TYPEMINUPTO:
2571      case OP_TYPEPOSUPTO:      case OP_TYPEPOSUPTO:
2572      if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)      if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)
2573        code += 2;        code += 2;
2574      break;      break;
2575    
# Line 2708  check_posix_syntax(const pcre_uchar *ptr Line 2719  check_posix_syntax(const pcre_uchar *ptr
2719  {  {
2720  pcre_uchar terminator;          /* Don't combine these lines; the Solaris cc */  pcre_uchar terminator;          /* Don't combine these lines; the Solaris cc */
2721  terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */  terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */
2722  for (++ptr; *ptr != 0; ptr++)  for (++ptr; *ptr != CHAR_NULL; ptr++)
2723    {    {
2724    if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)    if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)
2725      ptr++;      ptr++;
# Line 2755  register int yield = 0; Line 2766  register int yield = 0;
2766  while (posix_name_lengths[yield] != 0)  while (posix_name_lengths[yield] != 0)
2767    {    {
2768    if (len == posix_name_lengths[yield] &&    if (len == posix_name_lengths[yield] &&
2769      STRNCMP_UC_C8(ptr, pn, len) == 0) return yield;      STRNCMP_UC_C8(ptr, pn, (unsigned int)len) == 0) return yield;
2770    pn += posix_name_lengths[yield] + 1;    pn += posix_name_lengths[yield] + 1;
2771    yield++;    yield++;
2772    }    }
# Line 2893  PUT(previous_callout, 2 + LINK_SIZE, len Line 2904  PUT(previous_callout, 2 + LINK_SIZE, len
2904  /* This function is passed the start and end of a class range, in UTF-8 mode  /* This function is passed the start and end of a class range, in UTF-8 mode
2905  with UCP support. It searches up the characters, looking for ranges of  with UCP support. It searches up the characters, looking for ranges of
2906  characters in the "other" case. Each call returns the next one, updating the  characters in the "other" case. Each call returns the next one, updating the
2907  start address. A character with multiple other cases is returned on its own  start address. A character with multiple other cases is returned on its own
2908  with a special return value.  with a special return value.
2909    
2910  Arguments:  Arguments:
# Line 2905  Arguments: Line 2916  Arguments:
2916  Yield:        -1 when no more  Yield:        -1 when no more
2917                 0 when a range is returned                 0 when a range is returned
2918                >0 the CASESET offset for char with multiple other cases                >0 the CASESET offset for char with multiple other cases
2919                  in this case, ocptr contains the original                  in this case, ocptr contains the original
2920  */  */
2921    
2922  static int  static int
# Line 2913  get_othercase_range(pcre_uint32 *cptr, p Line 2924  get_othercase_range(pcre_uint32 *cptr, p
2924    pcre_uint32 *odptr)    pcre_uint32 *odptr)
2925  {  {
2926  pcre_uint32 c, othercase, next;  pcre_uint32 c, othercase, next;
2927  int co;  unsigned int co;
2928    
2929  /* Find the first character that has an other case. If it has multiple other  /* Find the first character that has an other case. If it has multiple other
2930  cases, return its case offset value. */  cases, return its case offset value. */
2931    
2932  for (c = *cptr; c <= d; c++)  for (c = *cptr; c <= d; c++)
2933    {    {
2934    if ((co = UCD_CASESET(c)) != 0)    if ((co = UCD_CASESET(c)) != 0)
2935      {      {
2936      *ocptr = c++;   /* Character that has the set */      *ocptr = c++;   /* Character that has the set */
2937      *cptr = c;      /* Rest of input range */      *cptr = c;      /* Rest of input range */
2938      return co;      return (int)co;
2939      }      }
2940    if ((othercase = UCD_OTHERCASE(c)) != c) break;    if ((othercase = UCD_OTHERCASE(c)) != c) break;
2941    }    }
2942    
2943  if (c > d) return -1;  /* Reached end of range */  if (c > d) return -1;  /* Reached end of range */
# Line 2964  Returns:       TRUE if auto-possessifyin Line 2975  Returns:       TRUE if auto-possessifyin
2975  */  */
2976    
2977  static BOOL  static BOOL
2978  check_char_prop(pcre_uint32 c, int ptype, int pdata, BOOL negated)  check_char_prop(pcre_uint32 c, unsigned int ptype, unsigned int pdata, BOOL negated)
2979  {  {
2980  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2981  const pcre_uint32 *p;  const pcre_uint32 *p;
# Line 3009  switch(ptype) Line 3020  switch(ptype)
3020    return (PRIV(ucp_gentype)[prop->chartype] == ucp_L ||    return (PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
3021            PRIV(ucp_gentype)[prop->chartype] == ucp_N ||            PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
3022            c == CHAR_UNDERSCORE) == negated;            c == CHAR_UNDERSCORE) == negated;
3023    
3024  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3025    case PT_CLIST:    case PT_CLIST:
3026    p = PRIV(ucd_caseless_sets) + prop->caseset;    p = PRIV(ucd_caseless_sets) + prop->caseset;
3027    for (;;)    for (;;)
3028      {      {
3029      if ((unsigned int)c < *p) return !negated;      if (c < *p) return !negated;
3030      if ((unsigned int)c == *p++) return negated;      if (c == *p++) return negated;
3031      }      }
3032    break;  /* Control never reaches here */    break;  /* Control never reaches here */
3033  #endif  #endif
# Line 3053  check_auto_possessive(const pcre_uchar * Line 3064  check_auto_possessive(const pcre_uchar *
3064  pcre_uint32 c = NOTACHAR;  pcre_uint32 c = NOTACHAR;
3065  pcre_uint32 next;  pcre_uint32 next;
3066  int escape;  int escape;
3067  int op_code = *previous++;  pcre_uchar op_code = *previous++;
3068    
3069  /* Skip whitespace and comments in extended mode */  /* Skip whitespace and comments in extended mode */
3070    
# Line 3065  if ((options & PCRE_EXTENDED) != 0) Line 3076  if ((options & PCRE_EXTENDED) != 0)
3076      if (*ptr == CHAR_NUMBER_SIGN)      if (*ptr == CHAR_NUMBER_SIGN)
3077        {        {
3078        ptr++;        ptr++;
3079        while (*ptr != 0)        while (*ptr != CHAR_NULL)
3080          {          {
3081          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
3082          ptr++;          ptr++;
# Line 3084  value is a character, a negative value i Line 3095  value is a character, a negative value i
3095  if (*ptr == CHAR_BACKSLASH)  if (*ptr == CHAR_BACKSLASH)
3096    {    {
3097    int temperrorcode = 0;    int temperrorcode = 0;
3098    escape = check_escape(&ptr, &next, &temperrorcode, cd->bracount, options, FALSE);    escape = check_escape(&ptr, &next, &temperrorcode, cd->bracount, options,
3099        FALSE);
3100    if (temperrorcode != 0) return FALSE;    if (temperrorcode != 0) return FALSE;
3101    ptr++;    /* Point after the escape sequence */    ptr++;    /* Point after the escape sequence */
3102    }    }
# Line 3108  if ((options & PCRE_EXTENDED) != 0) Line 3120  if ((options & PCRE_EXTENDED) != 0)
3120      if (*ptr == CHAR_NUMBER_SIGN)      if (*ptr == CHAR_NUMBER_SIGN)
3121        {        {
3122        ptr++;        ptr++;
3123        while (*ptr != 0)        while (*ptr != CHAR_NULL)
3124          {          {
3125          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
3126          ptr++;          ptr++;
# Line 3129  if (*ptr == CHAR_ASTERISK || *ptr == CHA Line 3141  if (*ptr == CHAR_ASTERISK || *ptr == CHA
3141    
3142  /* If the previous item is a character, get its value. */  /* If the previous item is a character, get its value. */
3143    
3144  if (op_code == OP_CHAR || op_code == OP_CHARI ||  if (op_code == OP_CHAR || op_code == OP_CHARI ||
3145      op_code == OP_NOT || op_code == OP_NOTI)      op_code == OP_NOT || op_code == OP_NOTI)
   //if (escape == 0) switch(op_code)  
3146    {    {
3147  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3148    GETCHARTEST(c, previous);    GETCHARTEST(c, previous);
# Line 3147  if (escape == 0) Line 3158  if (escape == 0)
3158    {    {
3159    /* For a caseless UTF match, the next character may have more than one other    /* For a caseless UTF match, the next character may have more than one other
3160    case, which maps to the special PT_CLIST property. Check this first. */    case, which maps to the special PT_CLIST property. Check this first. */
3161    
3162  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3163    if (utf && c != NOTACHAR && (options & PCRE_CASELESS) != 0)    if (utf && c != NOTACHAR && (options & PCRE_CASELESS) != 0)
3164      {      {
3165      int ocs = UCD_CASESET(next);      unsigned int ocs = UCD_CASESET(next);
3166      if (ocs > 0) return check_char_prop(c, PT_CLIST, ocs, op_code >= OP_NOT);      if (ocs > 0) return check_char_prop(c, PT_CLIST, ocs, op_code >= OP_NOT);
3167      }      }
3168  #endif  #endif
# Line 3163  if (escape == 0) Line 3174  if (escape == 0)
3174    
3175      /* For CHARI (caseless character) we must check the other case. If we have      /* For CHARI (caseless character) we must check the other case. If we have
3176      Unicode property support, we can use it to test the other case of      Unicode property support, we can use it to test the other case of
3177      high-valued characters. We know that next can have only one other case,      high-valued characters. We know that next can have only one other case,
3178      because multi-other-case characters are dealt with above. */      because multi-other-case characters are dealt with above. */
3179    
3180      case OP_CHARI:      case OP_CHARI:
# Line 3183  if (escape == 0) Line 3194  if (escape == 0)
3194      else      else
3195  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
3196      return (c != TABLE_GET(next, cd->fcc, next));  /* Not UTF */      return (c != TABLE_GET(next, cd->fcc, next));  /* Not UTF */
3197    
3198      case OP_NOT:      case OP_NOT:
3199      return c == next;      return c == next;
3200    
3201      case OP_NOTI:      case OP_NOTI:
3202      if (c == next) return TRUE;      if (c == next) return TRUE;
3203  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
# Line 3230  if (escape == 0) Line 3241  if (escape == 0)
3241      case OP_NOT_HSPACE:      case OP_NOT_HSPACE:
3242      switch(next)      switch(next)
3243        {        {
3244        HSPACE_CASES:        HSPACE_CASES:
3245        return op_code == OP_NOT_HSPACE;        return op_code == OP_NOT_HSPACE;
3246    
3247        default:        default:
# Line 3242  if (escape == 0) Line 3253  if (escape == 0)
3253      case OP_NOT_VSPACE:      case OP_NOT_VSPACE:
3254      switch(next)      switch(next)
3255        {        {
3256        VSPACE_CASES:        VSPACE_CASES:
3257        return op_code == OP_NOT_VSPACE;        return op_code == OP_NOT_VSPACE;
3258    
3259        default:        default:
# Line 3296  switch(op_code) Line 3307  switch(op_code)
3307      case ESC_H:      case ESC_H:
3308      switch(c)      switch(c)
3309        {        {
3310        HSPACE_CASES:        HSPACE_CASES:
3311        return escape != ESC_h;        return escape != ESC_h;
3312    
3313        default:        default:
3314        return escape == ESC_h;        return escape == ESC_h;
3315        }        }
# Line 3307  switch(op_code) Line 3318  switch(op_code)
3318      case ESC_V:      case ESC_V:
3319      switch(c)      switch(c)
3320        {        {
3321        VSPACE_CASES:        VSPACE_CASES:
3322        return escape != ESC_v;        return escape != ESC_v;
3323    
3324        default:        default:
# Line 3337  switch(op_code) Line 3348  switch(op_code)
3348      case ESC_p:      case ESC_p:
3349      case ESC_P:      case ESC_P:
3350        {        {
3351        int ptype, pdata, errorcodeptr;        unsigned int ptype = 0, pdata = 0;
3352          int errorcodeptr;
3353        BOOL negated;        BOOL negated;
3354    
3355        ptr--;      /* Make ptr point at the p or P */        ptr--;      /* Make ptr point at the p or P */
3356        ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr);        if (!get_ucp(&ptr, &negated, &ptype, &pdata, &errorcodeptr))
3357        if (ptype < 0) return FALSE;          return FALSE;
3358        ptr++;      /* Point past the final curly ket */        ptr++;      /* Point past the final curly ket */
3359    
3360        /* If the property item is optional, we have to give up. (When generated        /* If the property item is optional, we have to give up. (When generated
# Line 3418  switch(op_code) Line 3430  switch(op_code)
3430  *************************************************/  *************************************************/
3431    
3432  /* This function packages up the logic of adding a character or range of  /* This function packages up the logic of adding a character or range of
3433  characters to a class. The character values in the arguments will be within the  characters to a class. The character values in the arguments will be within the
3434  valid values for the current mode (8-bit, 16-bit, UTF, etc). This function is  valid values for the current mode (8-bit, 16-bit, UTF, etc). This function is
3435  mutually recursive with the function immediately below.  mutually recursive with the function immediately below.
3436    
3437  Arguments:  Arguments:
3438    classbits     the bit map for characters < 256    classbits     the bit map for characters < 256
3439    uchardptr     points to the pointer for extra data    uchardptr     points to the pointer for extra data
3440    options       the options word    options       the options word
3441    cd            contains pointers to tables etc.    cd            contains pointers to tables etc.
3442    start         start of range character    start         start of range character
3443    end           end of range character    end           end of range character
3444    
3445  Returns:        the number of < 256 characters added  Returns:        the number of < 256 characters added
3446                  the pointer to extra data is updated                  the pointer to extra data is updated
3447  */  */
# Line 3441  add_to_class(pcre_uint8 *classbits, pcre Line 3453  add_to_class(pcre_uint8 *classbits, pcre
3453  pcre_uint32 c;  pcre_uint32 c;
3454  int n8 = 0;  int n8 = 0;
3455    
3456  /* If caseless matching is required, scan the range and process alternate  /* If caseless matching is required, scan the range and process alternate
3457  cases. In Unicode, there are 8-bit characters that have alternate cases that  cases. In Unicode, there are 8-bit characters that have alternate cases that
3458  are greater than 255 and vice-versa. Sometimes we can just extend the original  are greater than 255 and vice-versa. Sometimes we can just extend the original
3459  range. */  range. */
3460    
3461  if ((options & PCRE_CASELESS) != 0)  if ((options & PCRE_CASELESS) != 0)
3462    {    {
3463  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3464    if ((options & PCRE_UTF8) != 0)    if ((options & PCRE_UTF8) != 0)
3465      {      {
3466      int rc;      int rc;
3467      pcre_uint32 oc, od;      pcre_uint32 oc, od;
3468    
3469      options &= ~PCRE_CASELESS;   /* Remove for recursive calls */      options &= ~PCRE_CASELESS;   /* Remove for recursive calls */
3470      c = start;      c = start;
3471    
3472      while ((rc = get_othercase_range(&c, end, &oc, &od)) >= 0)      while ((rc = get_othercase_range(&c, end, &oc, &od)) >= 0)
3473        {        {
3474        /* Handle a single character that has more than one other case. */        /* Handle a single character that has more than one other case. */
3475    
3476        if (rc > 0) n8 += add_list_to_class(classbits, uchardptr, options, cd,        if (rc > 0) n8 += add_list_to_class(classbits, uchardptr, options, cd,
3477          PRIV(ucd_caseless_sets) + rc, oc);          PRIV(ucd_caseless_sets) + rc, oc);
3478    
3479        /* Do nothing if the other case range is within the original range. */        /* Do nothing if the other case range is within the original range. */
3480    
3481        else if (oc >= start && od <= end) continue;        else if (oc >= start && od <= end) continue;
3482    
3483        /* Extend the original range if there is overlap, noting that if oc < c, we        /* Extend the original range if there is overlap, noting that if oc < c, we
3484        can't have od > end because a subrange is always shorter than the basic        can't have od > end because a subrange is always shorter than the basic
3485        range. Otherwise, use a recursive call to add the additional range. */        range. Otherwise, use a recursive call to add the additional range. */
3486    
3487        else if (oc < start && od >= start - 1) start = oc; /* Extend downwards */        else if (oc < start && od >= start - 1) start = oc; /* Extend downwards */
3488        else if (od > end && oc <= end + 1) end = od;       /* Extend upwards */        else if (od > end && oc <= end + 1) end = od;       /* Extend upwards */
3489        else n8 += add_to_class(classbits, uchardptr, options, cd, oc, od);        else n8 += add_to_class(classbits, uchardptr, options, cd, oc, od);
# Line 3481  if ((options & PCRE_CASELESS) != 0) Line 3493  if ((options & PCRE_CASELESS) != 0)
3493  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3494    
3495    /* Not UTF-mode, or no UCP */    /* Not UTF-mode, or no UCP */
3496    
3497    for (c = start; c <= end && c < 256; c++)    for (c = start; c <= end && c < 256; c++)
3498      {      {
3499      SETBIT(classbits, cd->fcc[c]);      SETBIT(classbits, cd->fcc[c]);
3500      n8++;      n8++;
3501      }      }
3502    }    }
3503    
3504  /* Now handle the original range. Adjust the final value according to the bit  /* Now handle the original range. Adjust the final value according to the bit
3505  length - this means that the same lists of (e.g.) horizontal spaces can be used  length - this means that the same lists of (e.g.) horizontal spaces can be used
3506  in all cases. */  in all cases. */
# Line 3514  if (end < 0x100) Line 3526  if (end < 0x100)
3526    {    {
3527    for (c = start; c <= end; c++)    for (c = start; c <= end; c++)
3528      {      {
3529      n8++;      n8++;
3530      SETBIT(classbits, c);      SETBIT(classbits, c);
3531      }      }
3532    }    }
3533    
3534  else  else
3535    {    {
3536    pcre_uchar *uchardata = *uchardptr;    pcre_uchar *uchardata = *uchardptr;
3537    
3538  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3539    if ((options & PCRE_UTF8) != 0)  /* All UTFs use the same flag bit */    if ((options & PCRE_UTF8) != 0)  /* All UTFs use the same flag bit */
3540      {      {
3541      if (start < end)      if (start < end)
3542        {        {
3543        *uchardata++ = XCL_RANGE;        *uchardata++ = XCL_RANGE;
3544        uchardata += PRIV(ord2utf)(start, uchardata);        uchardata += PRIV(ord2utf)(start, uchardata);
3545        uchardata += PRIV(ord2utf)(end, uchardata);        uchardata += PRIV(ord2utf)(end, uchardata);
3546        }        }
3547      else if (start == end)      else if (start == end)
3548        {        {
3549        *uchardata++ = XCL_SINGLE;        *uchardata++ = XCL_SINGLE;
3550        uchardata += PRIV(ord2utf)(start, uchardata);        uchardata += PRIV(ord2utf)(start, uchardata);
3551        }        }
3552      }      }
3553    else    else
3554  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
3555    
3556    /* Without UTF support, character values are constrained by the bit length,    /* Without UTF support, character values are constrained by the bit length,
3557    and can only be > 256 for 16-bit and 32-bit libraries. */    and can only be > 256 for 16-bit and 32-bit libraries. */
3558    
3559  #ifdef COMPILE_PCRE8  #ifdef COMPILE_PCRE8
3560      {}      {}
3561  #else  #else
3562    if (start < end)    if (start < end)
3563      {      {
3564      *uchardata++ = XCL_RANGE;      *uchardata++ = XCL_RANGE;
# Line 3557  else Line 3569  else
3569      {      {
3570      *uchardata++ = XCL_SINGLE;      *uchardata++ = XCL_SINGLE;
3571      *uchardata++ = start;      *uchardata++ = start;
3572      }      }
3573  #endif  #endif
3574    
3575    *uchardptr = uchardata;   /* Updata extra data pointer */    *uchardptr = uchardata;   /* Updata extra data pointer */
3576    }    }
3577    
3578  return n8;    /* Number of 8-bit characters */  return n8;    /* Number of 8-bit characters */
3579  }  }
3580    
3581    
3582    
3583    
3584  /*************************************************  /*************************************************
3585  *        Add a list of characters to a class     *  *        Add a list of characters to a class     *
3586  *************************************************/  *************************************************/
3587    
3588  /* This function is used for adding a list of case-equivalent characters to a  /* This function is used for adding a list of case-equivalent characters to a
3589  class, and also for adding a list of horizontal or vertical whitespace. If the  class, and also for adding a list of horizontal or vertical whitespace. If the
3590  list is in order (which it should be), ranges of characters are detected and  list is in order (which it should be), ranges of characters are detected and
3591  handled appropriately. This function is mutually recursive with the function  handled appropriately. This function is mutually recursive with the function
# Line 3583  Arguments: Line 3595  Arguments:
3595    classbits     the bit map for characters < 256    classbits     the bit map for characters < 256
3596    uchardptr     points to the pointer for extra data    uchardptr     points to the pointer for extra data
3597    options       the options word    options       the options word
3598    cd            contains pointers to tables etc.    cd            contains pointers to tables etc.
3599    p             points to row of 32-bit values, terminated by NOTACHAR    p             points to row of 32-bit values, terminated by NOTACHAR
3600    except        character to omit; this is used when adding lists of    except        character to omit; this is used when adding lists of
3601                    case-equivalent characters to avoid including the one we                    case-equivalent characters to avoid including the one we
3602                    already know about                    already know about
3603    
3604  Returns:        the number of < 256 characters added  Returns:        the number of < 256 characters added
3605                  the pointer to extra data is updated                  the pointer to extra data is updated
3606  */  */
# Line 3602  while (p[0] < NOTACHAR) Line 3614  while (p[0] < NOTACHAR)
3614    {    {
3615    int n = 0;    int n = 0;
3616    if (p[0] != except)    if (p[0] != except)
3617      {      {
3618      while(p[n+1] == p[0] + n + 1) n++;      while(p[n+1] == p[0] + n + 1) n++;
3619      n8 += add_to_class(classbits, uchardptr, options, cd, p[0], p[n]);      n8 += add_to_class(classbits, uchardptr, options, cd, p[0], p[n]);
3620      }      }
3621    p += n + 1;    p += n + 1;
3622    }    }
3623  return n8;  return n8;
3624  }  }
3625    
3626    
3627    
# Line 3624  Arguments: Line 3636  Arguments:
3636    classbits     the bit map for characters < 256    classbits     the bit map for characters < 256
3637    uchardptr     points to the pointer for extra data    uchardptr     points to the pointer for extra data
3638    options       the options word    options       the options word
3639    cd            contains pointers to tables etc.    cd            contains pointers to tables etc.
3640    p             points to row of 32-bit values, terminated by NOTACHAR    p             points to row of 32-bit values, terminated by NOTACHAR
3641    
3642  Returns:        the number of < 256 characters added  Returns:        the number of < 256 characters added
3643                  the pointer to extra data is updated                  the pointer to extra data is updated
3644  */  */
3645    
3646  static int  static int
3647  add_not_list_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr,  add_not_list_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr,
3648    int options, compile_data *cd, const pcre_uint32 *p)    int options, compile_data *cd, const pcre_uint32 *p)
3649  {  {
3650  BOOL utf = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
# Line 3644  while (p[0] < NOTACHAR) Line 3656  while (p[0] < NOTACHAR)
3656    while (p[1] == p[0] + 1) p++;    while (p[1] == p[0] + 1) p++;
3657    n8 += add_to_class(classbits, uchardptr, options, cd, p[0] + 1,    n8 += add_to_class(classbits, uchardptr, options, cd, p[0] + 1,
3658      (p[1] == NOTACHAR) ? (utf ? 0x10ffffu : 0xffffffffu) : p[1] - 1);      (p[1] == NOTACHAR) ? (utf ? 0x10ffffu : 0xffffffffu) : p[1] - 1);
3659    p++;    p++;
3660    }    }
3661  return n8;  return n8;
3662  }  }
3663    
3664    
3665    
# Line 3666  Arguments: Line 3678  Arguments:
3678    codeptr        points to the pointer to the current code point    codeptr        points to the pointer to the current code point
3679    ptrptr         points to the current pattern pointer    ptrptr         points to the current pattern pointer
3680    errorcodeptr   points to error code variable    errorcodeptr   points to error code variable
3681    firstcharptr   set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE)    firstcharptr    place to put the first required character
3682    reqcharptr     set to the last literal character required, else < 0    firstcharflagsptr place to put the first character flags, or a negative number
3683      reqcharptr     place to put the last required character
3684      reqcharflagsptr place to put the last required character flags, or a negative number
3685    bcptr          points to current branch chain    bcptr          points to current branch chain
3686    cond_depth     conditional nesting depth    cond_depth     conditional nesting depth
3687    cd             contains pointers to tables etc.    cd             contains pointers to tables etc.
# Line 3680  Returns:         TRUE on success Line 3694  Returns:         TRUE on success
3694    
3695  static BOOL  static BOOL
3696  compile_branch(int *optionsptr, pcre_uchar **codeptr,  compile_branch(int *optionsptr, pcre_uchar **codeptr,
3697    const pcre_uchar **ptrptr, int *errorcodeptr, pcre_int32 *firstcharptr,    const pcre_uchar **ptrptr, int *errorcodeptr,
3698    pcre_int32 *reqcharptr, branch_chain *bcptr, int cond_depth,    pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr,
3699      pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr,
3700      branch_chain *bcptr, int cond_depth,
3701    compile_data *cd, int *lengthptr)    compile_data *cd, int *lengthptr)
3702  {  {
3703  int repeat_type, op_type;  int repeat_type, op_type;
3704  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */
3705  int bravalue = 0;  int bravalue = 0;
3706  int greedy_default, greedy_non_default;  int greedy_default, greedy_non_default;
3707  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
3708  pcre_int32 zeroreqchar, zerofirstchar;  pcre_int32 firstcharflags, reqcharflags;
3709    pcre_uint32 zeroreqchar, zerofirstchar;
3710    pcre_int32 zeroreqcharflags, zerofirstcharflags;
3711  pcre_int32 req_caseopt, reqvary, tempreqvary;  pcre_int32 req_caseopt, reqvary, tempreqvary;
3712  int options = *optionsptr;               /* May change dynamically */  int options = *optionsptr;               /* May change dynamically */
3713  int after_manual_callout = 0;  int after_manual_callout = 0;
# Line 3717  dynamically as we process the pattern. * Line 3735  dynamically as we process the pattern. *
3735  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3736  /* PCRE_UTF[16|32] have the same value as PCRE_UTF8. */  /* PCRE_UTF[16|32] have the same value as PCRE_UTF8. */
3737  BOOL utf = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
3738    #ifndef COMPILE_PCRE32
3739  pcre_uchar utf_chars[6];  pcre_uchar utf_chars[6];
3740    #endif
3741  #else  #else
3742  BOOL utf = FALSE;  BOOL utf = FALSE;
3743  #endif  #endif
3744    
3745  /* Helper variables for OP_XCLASS opcode (for characters > 255). We define  /* Helper variables for OP_XCLASS opcode (for characters > 255). We define
3746  class_uchardata always so that it can be passed to add_to_class() always,  class_uchardata always so that it can be passed to add_to_class() always,
3747  though it will not be used in non-UTF 8-bit cases. This avoids having to supply  though it will not be used in non-UTF 8-bit cases. This avoids having to supply
3748  alternative calls for the different cases. */  alternative calls for the different cases. */
3749    
3750  pcre_uchar *class_uchardata;  pcre_uchar *class_uchardata;
# Line 3752  to take the zero repeat into account. Th Line 3772  to take the zero repeat into account. Th
3772  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual
3773  item types that can be repeated set these backoff variables appropriately. */  item types that can be repeated set these backoff variables appropriately. */
3774    
3775  firstchar = reqchar = zerofirstchar = zeroreqchar = REQ_UNSET;  firstchar = reqchar = zerofirstchar = zeroreqchar = 0;
3776    firstcharflags = reqcharflags = zerofirstcharflags = zeroreqcharflags = REQ_UNSET;
3777    
3778  /* The variable req_caseopt contains either the REQ_CASELESS value  /* The variable req_caseopt contains either the REQ_CASELESS value
3779  or zero, according to the current setting of the caseless flag. The  or zero, according to the current setting of the caseless flag. The
# Line 3778  for (;; ptr++) Line 3799  for (;; ptr++)
3799    int recno;    int recno;
3800    int refsign;    int refsign;
3801    int skipbytes;    int skipbytes;
3802    int subreqchar;    pcre_uint32 subreqchar, subfirstchar;
3803    int subfirstchar;    pcre_int32 subreqcharflags, subfirstcharflags;
3804    int terminator;    int terminator;
3805    int mclength;    unsigned int mclength;
3806    int tempbracount;    unsigned int tempbracount;
3807    int ec; // FIXMEchpe pcre_uint32    pcre_uint32 ec;
3808    pcre_uchar mcbuffer[8];    pcre_uchar mcbuffer[8];
3809    
3810    /* Get next character in the pattern */    /* Get next character in the pattern */
# Line 3793  for (;; ptr++) Line 3814  for (;; ptr++)
3814    /* If we are at the end of a nested substitution, revert to the outer level    /* If we are at the end of a nested substitution, revert to the outer level
3815    string. Nesting only happens one level deep. */    string. Nesting only happens one level deep. */
3816    
3817    if (c == 0 && nestptr != NULL)    if (c == CHAR_NULL && nestptr != NULL)
3818      {      {
3819      ptr = nestptr;      ptr = nestptr;
3820      nestptr = NULL;      nestptr = NULL;
# Line 3868  for (;; ptr++) Line 3889  for (;; ptr++)
3889    
3890    /* If in \Q...\E, check for the end; if not, we have a literal */    /* If in \Q...\E, check for the end; if not, we have a literal */
3891    
3892    if (inescq && c != 0)    if (inescq && c != CHAR_NULL)
3893      {      {
3894      if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E)      if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E)
3895        {        {
# Line 3916  for (;; ptr++) Line 3937  for (;; ptr++)
3937      if (c == CHAR_NUMBER_SIGN)      if (c == CHAR_NUMBER_SIGN)
3938        {        {
3939        ptr++;        ptr++;
3940        while (*ptr != 0)        while (*ptr != CHAR_NULL)
3941          {          {
3942          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
3943          ptr++;          ptr++;
# Line 3924  for (;; ptr++) Line 3945  for (;; ptr++)
3945          if (utf) FORWARDCHAR(ptr);          if (utf) FORWARDCHAR(ptr);
3946  #endif  #endif
3947          }          }
3948        if (*ptr != 0) continue;        if (*ptr != CHAR_NULL) continue;
3949    
3950        /* Else fall through to handle end of string */        /* Else fall through to handle end of string */
3951        c = 0;        c = 0;
# Line 3946  for (;; ptr++) Line 3967  for (;; ptr++)
3967      case CHAR_VERTICAL_LINE:       /* or | or ) */      case CHAR_VERTICAL_LINE:       /* or | or ) */
3968      case CHAR_RIGHT_PARENTHESIS:      case CHAR_RIGHT_PARENTHESIS:
3969      *firstcharptr = firstchar;      *firstcharptr = firstchar;
3970        *firstcharflagsptr = firstcharflags;
3971      *reqcharptr = reqchar;      *reqcharptr = reqchar;
3972        *reqcharflagsptr = reqcharflags;
3973      *codeptr = code;      *codeptr = code;
3974      *ptrptr = ptr;      *ptrptr = ptr;
3975      if (lengthptr != NULL)      if (lengthptr != NULL)
# Line 3970  for (;; ptr++) Line 3993  for (;; ptr++)
3993      previous = NULL;      previous = NULL;
3994      if ((options & PCRE_MULTILINE) != 0)      if ((options & PCRE_MULTILINE) != 0)
3995        {        {
3996        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;        if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
3997        *code++ = OP_CIRCM;        *code++ = OP_CIRCM;
3998        }        }
3999      else *code++ = OP_CIRC;      else *code++ = OP_CIRC;
# Line 3985  for (;; ptr++) Line 4008  for (;; ptr++)
4008      repeats. The value of reqchar doesn't change either. */      repeats. The value of reqchar doesn't change either. */
4009    
4010      case CHAR_DOT:      case CHAR_DOT:
4011      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;      if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4012      zerofirstchar = firstchar;      zerofirstchar = firstchar;
4013        zerofirstcharflags = firstcharflags;
4014      zeroreqchar = reqchar;      zeroreqchar = reqchar;
4015        zeroreqcharflags = reqcharflags;
4016      previous = code;      previous = code;
4017      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;
4018      break;      break;
# Line 4061  for (;; ptr++) Line 4086  for (;; ptr++)
4086          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)
4087        {        {
4088        *code++ = negate_class? OP_ALLANY : OP_FAIL;        *code++ = negate_class? OP_ALLANY : OP_FAIL;
4089        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;        if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4090        zerofirstchar = firstchar;        zerofirstchar = firstchar;
4091          zerofirstcharflags = firstcharflags;
4092        break;        break;
4093        }        }
4094    
# Line 4097  for (;; ptr++) Line 4123  for (;; ptr++)
4123      means that an initial ] is taken as a data character. At the start of the      means that an initial ] is taken as a data character. At the start of the
4124      loop, c contains the first byte of the character. */      loop, c contains the first byte of the character. */
4125    
4126      if (c != 0) do      if (c != CHAR_NULL) do
4127        {        {
4128        const pcre_uchar *oldptr;        const pcre_uchar *oldptr;
4129    
4130  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
4131        if (utf && HAS_EXTRALEN(c))        if (utf && HAS_EXTRALEN(c))
4132          {                           /* Braces are required because the */          {                           /* Braces are required because the */
4133          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */
# Line 4112  for (;; ptr++) Line 4138  for (;; ptr++)
4138        /* In the pre-compile phase, accumulate the length of any extra        /* In the pre-compile phase, accumulate the length of any extra
4139        data and reset the pointer. This is so that very large classes that        data and reset the pointer. This is so that very large classes that
4140        contain a zillion > 255 characters no longer overwrite the work space        contain a zillion > 255 characters no longer overwrite the work space
4141        (which is on the stack). We have to remember that there was XCLASS data,        (which is on the stack). We have to remember that there was XCLASS data,
4142        however. */        however. */
4143    
4144        if (lengthptr != NULL && class_uchardata > class_uchardata_base)        if (lengthptr != NULL && class_uchardata > class_uchardata_base)
# Line 4176  for (;; ptr++) Line 4202  for (;; ptr++)
4202          alpha. This relies on the fact that the class table starts with          alpha. This relies on the fact that the class table starts with
4203          alpha, lower, upper as the first 3 entries. */          alpha, lower, upper as the first 3 entries. */
4204    
4205          if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)          if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)
4206            posix_class = 0;            posix_class = 0;
4207    
4208          /* When PCRE_UCP is set, some of the POSIX classes are converted to          /* When PCRE_UCP is set, some of the POSIX classes are converted to
# Line 4253  for (;; ptr++) Line 4279  for (;; ptr++)
4279    
4280        if (c == CHAR_BACKSLASH)        if (c == CHAR_BACKSLASH)
4281          {          {
4282          escape = check_escape(&ptr, &ec, errorcodeptr, cd->bracount, options, TRUE);          escape = check_escape(&ptr, &ec, errorcodeptr, cd->bracount, options,
4283              TRUE);
4284          if (*errorcodeptr != 0) goto FAILED;          if (*errorcodeptr != 0) goto FAILED;
4285            if (escape == 0) c = ec;
         if (escape == 0)  
           c = ec;  
4286          else if (escape == ESC_b) c = CHAR_BS; /* \b is backspace in a class */          else if (escape == ESC_b) c = CHAR_BS; /* \b is backspace in a class */
4287          else if (escape == ESC_N)            /* \N is not supported in a class */          else if (escape == ESC_N)          /* \N is not supported in a class */
4288            {            {
4289            *errorcodeptr = ERR71;            *errorcodeptr = ERR71;
4290            goto FAILED;            goto FAILED;
# Line 4332  for (;; ptr++) Line 4356  for (;; ptr++)
4356              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space];              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space];
4357              classbits[1] |= 0x08;    /* Perl 5.004 onwards omits VT from \s */              classbits[1] |= 0x08;    /* Perl 5.004 onwards omits VT from \s */
4358              continue;              continue;
4359    
4360              /* The rest apply in both UCP and non-UCP cases. */              /* The rest apply in both UCP and non-UCP cases. */
4361    
4362              case ESC_h:              case ESC_h:
4363              (void)add_list_to_class(classbits, &class_uchardata, options, cd,              (void)add_list_to_class(classbits, &class_uchardata, options, cd,
4364                PRIV(hspace_list), NOTACHAR);                PRIV(hspace_list), NOTACHAR);
4365              continue;              continue;
4366    
4367              case ESC_H:              case ESC_H:
4368              (void)add_not_list_to_class(classbits, &class_uchardata, options,              (void)add_not_list_to_class(classbits, &class_uchardata, options,
4369                cd, PRIV(hspace_list));                cd, PRIV(hspace_list));
4370              continue;              continue;
4371    
4372              case ESC_v:              case ESC_v:
4373              (void)add_list_to_class(classbits, &class_uchardata, options, cd,              (void)add_list_to_class(classbits, &class_uchardata, options, cd,
4374                PRIV(vspace_list), NOTACHAR);                PRIV(vspace_list), NOTACHAR);
4375              continue;              continue;
4376    
4377              case ESC_V:              case ESC_V:
4378              (void)add_not_list_to_class(classbits, &class_uchardata, options,              (void)add_not_list_to_class(classbits, &class_uchardata, options,
4379                cd, PRIV(vspace_list));                cd, PRIV(vspace_list));
4380              continue;              continue;
4381    
4382  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 4360  for (;; ptr++) Line 4384  for (;; ptr++)
4384              case ESC_P:              case ESC_P:
4385                {                {
4386                BOOL negated;                BOOL negated;
4387                int pdata;                unsigned int ptype = 0, pdata = 0;
4388                int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);                if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr))
4389                if (ptype < 0) goto FAILED;                  goto FAILED;
4390                *class_uchardata++ = ((escape == ESC_p) != negated)?                *class_uchardata++ = ((escape == ESC_p) != negated)?
4391                  XCL_PROP : XCL_NOTPROP;                  XCL_PROP : XCL_NOTPROP;
4392                *class_uchardata++ = ptype;                *class_uchardata++ = ptype;
# Line 4390  for (;; ptr++) Line 4414  for (;; ptr++)
4414    
4415          /* Fall through if the escape just defined a single character (c >= 0).          /* Fall through if the escape just defined a single character (c >= 0).
4416          This may be greater than 256. */          This may be greater than 256. */
4417    
4418          escape = 0;          escape = 0;
4419    
4420          }   /* End of backslash handling */          }   /* End of backslash handling */
# Line 4416  for (;; ptr++) Line 4440  for (;; ptr++)
4440    
4441        if (!inescq && ptr[1] == CHAR_MINUS)        if (!inescq && ptr[1] == CHAR_MINUS)
4442          {          {
4443          int d;          pcre_uint32 d;
4444          ptr += 2;          ptr += 2;
4445          while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2;          while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2;
4446    
# Line 4431  for (;; ptr++) Line 4455  for (;; ptr++)
4455            inescq = TRUE;            inescq = TRUE;
4456            break;            break;
4457            }            }
4458    
4459          /* Minus (hyphen) at the end of a class is treated as a literal, so put          /* Minus (hyphen) at the end of a class is treated as a literal, so put
4460          back the pointer and jump to handle the character that preceded it. */          back the pointer and jump to handle the character that preceded it. */
4461    
4462          if (*ptr == 0 || (!inescq && *ptr == CHAR_RIGHT_SQUARE_BRACKET))          if (*ptr == CHAR_NULL || (!inescq && *ptr == CHAR_RIGHT_SQUARE_BRACKET))
4463            {            {
4464            ptr = oldptr;            ptr = oldptr;
4465            goto CLASS_SINGLE_CHARACTER;            goto CLASS_SINGLE_CHARACTER;
4466            }            }
4467    
4468          /* Otherwise, we have a potential range; pick up the next character */          /* Otherwise, we have a potential range; pick up the next character */
4469    
4470  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
# Line 4487  for (;; ptr++) Line 4511  for (;; ptr++)
4511          /* We have found a character range, so single character optimizations          /* We have found a character range, so single character optimizations
4512          cannot be done anymore. Any value greater than 1 indicates that there          cannot be done anymore. Any value greater than 1 indicates that there
4513          is more than one character. */          is more than one character. */
4514    
4515          class_one_char = 2;          class_one_char = 2;
4516    
4517          /* Remember an explicit \r or \n, and add the range to the class. */          /* Remember an explicit \r or \n, and add the range to the class. */
4518    
4519          if (d == CHAR_CR || d == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF;          if (d == CHAR_CR || d == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF;
4520    
4521          class_has_8bitchar +=          class_has_8bitchar +=
4522            add_to_class(classbits, &class_uchardata, options, cd, c, d);            add_to_class(classbits, &class_uchardata, options, cd, c, d);
4523    
4524          continue;   /* Go get the next char in the class */          continue;   /* Go get the next char in the class */
4525          }          }
4526    
4527        /* Handle a single character - we can get here for a normal non-escape        /* Handle a single character - we can get here for a normal non-escape
4528        char, or after \ that introduces a single character or for an apparent        char, or after \ that introduces a single character or for an apparent
4529        range that isn't. Only the value 1 matters for class_one_char, so don't        range that isn't. Only the value 1 matters for class_one_char, so don't
4530        increase it if it is already 2 or more ... just in case there's a class        increase it if it is already 2 or more ... just in case there's a class
4531        with a zillion characters in it. */        with a zillion characters in it. */
4532    
4533        CLASS_SINGLE_CHARACTER:        CLASS_SINGLE_CHARACTER:
# Line 4522  for (;; ptr++) Line 4546  for (;; ptr++)
4546          {          {
4547          ptr++;          ptr++;
4548          zeroreqchar = reqchar;          zeroreqchar = reqchar;
4549            zeroreqcharflags = reqcharflags;
4550    
4551          if (negate_class)          if (negate_class)
4552            {            {
4553  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
             // FIXMEchpe pcreuint32?  
4554            int d;            int d;
4555  #endif  #endif
4556            if (firstchar == REQ_UNSET) firstchar = REQ_NONE;            if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4557            zerofirstchar = firstchar;            zerofirstchar = firstchar;
4558              zerofirstcharflags = firstcharflags;
4559    
4560            /* For caseless UTF-8 mode when UCP support is available, check            /* For caseless UTF-8 mode when UCP support is available, check
4561            whether this character has more than one other case. If so, generate            whether this character has more than one other case. If so, generate
4562            a special OP_NOTPROP item instead of OP_NOTI. */            a special OP_NOTPROP item instead of OP_NOTI. */
4563    
4564  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
4565            if (utf && (options & PCRE_CASELESS) != 0 &&            if (utf && (options & PCRE_CASELESS) != 0 &&
4566                (d = UCD_CASESET(c)) != 0)                (d = UCD_CASESET(c)) != 0)
4567              {              {
4568              *code++ = OP_NOTPROP;              *code++ = OP_NOTPROP;
4569              *code++ = PT_CLIST;              *code++ = PT_CLIST;
4570              *code++ = d;              *code++ = d;
4571              }              }
4572            else            else
4573  #endif  #endif
4574            /* Char has only one other case, or UCP not available */            /* Char has only one other case, or UCP not available */
4575    
# Line 4557  for (;; ptr++) Line 4582  for (;; ptr++)
4582  #endif  #endif
4583                *code++ = c;                *code++ = c;
4584              }              }
4585    
4586            /* We are finished with this character class */            /* We are finished with this character class */
4587    
4588            goto END_CLASS;            goto END_CLASS;
4589            }            }
4590    
# Line 4577  for (;; ptr++) Line 4602  for (;; ptr++)
4602            }            }
4603          goto ONE_CHAR;          goto ONE_CHAR;
4604          }       /* End of 1-char optimization */          }       /* End of 1-char optimization */
4605    
4606        /* There is more than one character in the class, or an XCLASS item        /* There is more than one character in the class, or an XCLASS item
4607        has been generated. Add this character to the class. */        has been generated. Add this character to the class. */
4608    
4609        class_has_8bitchar +=        class_has_8bitchar +=
4610          add_to_class(classbits, &class_uchardata, options, cd, c, c);          add_to_class(classbits, &class_uchardata, options, cd, c, c);
4611        }        }
4612    
# Line 4589  for (;; ptr++) Line 4614  for (;; ptr++)
4614      If we are at the end of an internal nested string, revert to the outer      If we are at the end of an internal nested string, revert to the outer
4615      string. */      string. */
4616    
4617      while (((c = *(++ptr)) != 0 ||      while (((c = *(++ptr)) != CHAR_NULL ||
4618             (nestptr != NULL &&             (nestptr != NULL &&
4619               (ptr = nestptr, nestptr = NULL, c = *(++ptr)) != 0)) &&               (ptr = nestptr, nestptr = NULL, c = *(++ptr)) != CHAR_NULL)) &&
4620             (c != CHAR_RIGHT_SQUARE_BRACKET || inescq));             (c != CHAR_RIGHT_SQUARE_BRACKET || inescq));
4621    
4622      /* Check for missing terminating ']' */      /* Check for missing terminating ']' */
4623    
4624      if (c == 0)      if (c == CHAR_NULL)
4625        {        {
4626        *errorcodeptr = ERR6;        *errorcodeptr = ERR6;
4627        goto FAILED;        goto FAILED;
4628        }        }
4629    
4630      /* We will need an XCLASS if data has been placed in class_uchardata. In      /* We will need an XCLASS if data has been placed in class_uchardata. In
4631      the second phase this is a sufficient test. However, in the pre-compile      the second phase this is a sufficient test. However, in the pre-compile
4632      phase, class_uchardata gets emptied to prevent workspace overflow, so it      phase, class_uchardata gets emptied to prevent workspace overflow, so it
4633      only if the very last character in the class needs XCLASS will it contain      only if the very last character in the class needs XCLASS will it contain
4634      anything at this point. For this reason, xclass gets set TRUE above when      anything at this point. For this reason, xclass gets set TRUE above when
4635      uchar_classdata is emptied, and that's why this code is the way it is here      uchar_classdata is emptied, and that's why this code is the way it is here
4636      instead of just doing a test on class_uchardata below. */      instead of just doing a test on class_uchardata below. */
4637    
4638  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
4639      if (class_uchardata > class_uchardata_base) xclass = TRUE;      if (class_uchardata > class_uchardata_base) xclass = TRUE;
4640  #endif  #endif
# Line 4618  for (;; ptr++) Line 4643  for (;; ptr++)
4643      setting, whatever the repeat count. Any reqchar setting must remain      setting, whatever the repeat count. Any reqchar setting must remain
4644      unchanged after any kind of repeat. */      unchanged after any kind of repeat. */
4645    
4646      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;      if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4647      zerofirstchar = firstchar;      zerofirstchar = firstchar;
4648        zerofirstcharflags = firstcharflags;
4649      zeroreqchar = reqchar;      zeroreqchar = reqchar;
4650        zeroreqcharflags = reqcharflags;
4651    
4652      /* If there are characters with values > 255, we have to compile an      /* If there are characters with values > 255, we have to compile an
4653      extended class, with its own opcode, unless there was a negated special      extended class, with its own opcode, unless there was a negated special
# Line 4676  for (;; ptr++) Line 4703  for (;; ptr++)
4703        memcpy(code, classbits, 32);        memcpy(code, classbits, 32);
4704        }        }
4705      code += 32 / sizeof(pcre_uchar);      code += 32 / sizeof(pcre_uchar);
4706    
4707      END_CLASS:      END_CLASS:
4708      break;      break;
4709    
# Line 4715  for (;; ptr++) Line 4742  for (;; ptr++)
4742      if (repeat_min == 0)      if (repeat_min == 0)
4743        {        {
4744        firstchar = zerofirstchar;    /* Adjust for zero repeat */        firstchar = zerofirstchar;    /* Adjust for zero repeat */
4745          firstcharflags = zerofirstcharflags;
4746        reqchar = zeroreqchar;        /* Ditto */        reqchar = zeroreqchar;        /* Ditto */
4747          reqcharflags = zeroreqcharflags;
4748        }        }
4749    
4750      /* Remember whether this is a variable length repeat */      /* Remember whether this is a variable length repeat */
# Line 4818  for (;; ptr++) Line 4847  for (;; ptr++)
4847          {          {
4848          c = code[-1];          c = code[-1];
4849          if (*previous <= OP_CHARI && repeat_min > 1)          if (*previous <= OP_CHARI && repeat_min > 1)
4850            reqchar = c | req_caseopt | cd->req_varyopt;            {
4851              reqchar = c;
4852              reqcharflags = req_caseopt | cd->req_varyopt;
4853              }
4854          }          }
4855    
4856        /* If the repetition is unlimited, it pays to see if the next thing on        /* If the repetition is unlimited, it pays to see if the next thing on
# Line 4875  for (;; ptr++) Line 4907  for (;; ptr++)
4907    
4908        if (repeat_max == 0) goto END_REPEAT;        if (repeat_max == 0) goto END_REPEAT;
4909    
       /*--------------------------------------------------------------------*/  
       /* This code is obsolete from release 8.00; the restriction was finally  
       removed: */  
   
       /* All real repeats make it impossible to handle partial matching (maybe  
       one day we will be able to remove this restriction). */  
   
       /* if (repeat_max != 1) cd->external_flags |= PCRE_NOPARTIAL; */  
       /*--------------------------------------------------------------------*/  
   
4910        /* Combine the op_type with the repeat_type */        /* Combine the op_type with the repeat_type */
4911    
4912        repeat_type += op_type;        repeat_type += op_type;
# Line 5031  for (;; ptr++) Line 5053  for (;; ptr++)
5053          goto END_REPEAT;          goto END_REPEAT;
5054          }          }
5055    
       /*--------------------------------------------------------------------*/  
       /* This code is obsolete from release 8.00; the restriction was finally  
       removed: */  
   
       /* All real repeats make it impossible to handle partial matching (maybe  
       one day we will be able to remove this restriction). */  
   
       /* if (repeat_max != 1) cd->external_flags |= PCRE_NOPARTIAL; */  
       /*--------------------------------------------------------------------*/  
   
5056        if (repeat_min == 0 && repeat_max == -1)        if (repeat_min == 0 && repeat_max == -1)
5057          *code++ = OP_CRSTAR + repeat_type;          *code++ = OP_CRSTAR + repeat_type;
5058        else if (repeat_min == 1 && repeat_max == -1)        else if (repeat_min == 1 && repeat_max == -1)
# Line 5200  for (;; ptr++) Line 5212  for (;; ptr++)
5212    
5213            else            else
5214              {              {
5215              if (groupsetfirstchar && reqchar < 0) reqchar = firstchar;              if (groupsetfirstchar && reqcharflags < 0)
5216                  {
5217                  reqchar = firstchar;
5218                  reqcharflags = firstcharflags;
5219                  }
5220    
5221              for (i = 1; i < repeat_min; i++)              for (i = 1; i < repeat_min; i++)
5222                {                {
# Line 5479  for (;; ptr++) Line 5495  for (;; ptr++)
5495        else if (*tempcode == OP_EXACT || *tempcode == OP_NOTEXACT)        else if (*tempcode == OP_EXACT || *tempcode == OP_NOTEXACT)
5496          {          {
5497          tempcode += PRIV(OP_lengths)[*tempcode];          tempcode += PRIV(OP_lengths)[*tempcode];
5498  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
5499          if (utf && HAS_EXTRALEN(tempcode[-1]))          if (utf && HAS_EXTRALEN(tempcode[-1]))
5500            tempcode += GET_EXTRALEN(tempcode[-1]);            tempcode += GET_EXTRALEN(tempcode[-1]);
5501  #endif  #endif
# Line 5575  for (;; ptr++) Line 5591  for (;; ptr++)
5591        if (*ptr == CHAR_COLON)        if (*ptr == CHAR_COLON)
5592          {          {
5593          arg = ++ptr;          arg = ++ptr;
5594          while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;          while (*ptr != CHAR_NULL && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
5595          arglen = (int)(ptr - arg);          arglen = (int)(ptr - arg);
5596          if ((unsigned int)arglen > MAX_MARK)          if ((unsigned int)arglen > MAX_MARK)
5597            {            {
# Line 5620  for (;; ptr++) Line 5636  for (;; ptr++)
5636                (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;                (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;
5637    
5638              /* Do not set firstchar after *ACCEPT */              /* Do not set firstchar after *ACCEPT */
5639              if (firstchar == REQ_UNSET) firstchar = REQ_NONE;              if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
5640              }              }
5641    
5642            /* Handle other cases with/without an argument */            /* Handle other cases with/without an argument */
# Line 5689  for (;; ptr++) Line 5705  for (;; ptr++)
5705          {          {
5706          case CHAR_NUMBER_SIGN:                 /* Comment; skip to ket */          case CHAR_NUMBER_SIGN:                 /* Comment; skip to ket */
5707          ptr++;          ptr++;
5708          while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;          while (*ptr != CHAR_NULL && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
5709          if (*ptr == 0)          if (*ptr == CHAR_NULL)
5710            {            {
5711            *errorcodeptr = ERR18;            *errorcodeptr = ERR18;
5712            goto FAILED;            goto FAILED;
# Line 5713  for (;; ptr++) Line 5729  for (;; ptr++)
5729          /* ------------------------------------------------------------ */          /* ------------------------------------------------------------ */
5730          case CHAR_LEFT_PARENTHESIS:          case CHAR_LEFT_PARENTHESIS:
5731          bravalue = OP_COND;       /* Conditional group */          bravalue = OP_COND;       /* Conditional group */
5732            tempptr = ptr;
5733    
5734          /* A condition can be an assertion, a number (referring to a numbered          /* A condition can be an assertion, a number (referring to a numbered
5735          group), a name (referring to a named group), or 'R', referring to          group), a name (referring to a named group), or 'R', referring to
# Line 5725  for (;; ptr++) Line 5742  for (;; ptr++)
5742          be the recursive thing or the name 'R' (and similarly for 'R' followed          be the recursive thing or the name 'R' (and similarly for 'R' followed
5743          by digits), and (b) a number could be a name that consists of digits.          by digits), and (b) a number could be a name that consists of digits.
5744          In both cases, we look for a name first; if not found, we try the other          In both cases, we look for a name first; if not found, we try the other
5745          cases. */          cases.
5746    
5747            For compatibility with auto-callouts, we allow a callout to be
5748            specified before a condition that is an assertion. First, check for the
5749            syntax of a callout; if found, adjust the temporary pointer that is
5750            used to check for an assertion condition. That's all that is needed! */
5751    
5752            if (ptr[1] == CHAR_QUESTION_MARK && ptr[2] == CHAR_C)
5753              {
5754              for (i = 3;; i++) if (!IS_DIGIT(ptr[i])) break;
5755              if (ptr[i] == CHAR_RIGHT_PARENTHESIS)
5756                tempptr += i + 1;
5757              }
5758    
5759          /* For conditions that are assertions, check the syntax, and then exit          /* For conditions that are assertions, check the syntax, and then exit
5760          the switch. This will take control down to where bracketed groups,          the switch. This will take control down to where bracketed groups,
5761          including assertions, are processed. */          including assertions, are processed. */
5762    
5763          if (ptr[1] == CHAR_QUESTION_MARK && (ptr[2] == CHAR_EQUALS_SIGN ||          if (tempptr[1] == CHAR_QUESTION_MARK &&
5764              ptr[2] == CHAR_EXCLAMATION_MARK || ptr[2] == CHAR_LESS_THAN_SIGN))                (tempptr[2] == CHAR_EQUALS_SIGN ||
5765                   tempptr[2] == CHAR_EXCLAMATION_MARK ||
5766                   tempptr[2] == CHAR_LESS_THAN_SIGN))
5767            break;            break;
5768    
5769          /* Most other conditions use OP_CREF (a couple change to OP_RREF          /* Most other conditions use OP_CREF (a couple change to OP_RREF
# Line 5766  for (;; ptr++) Line 5797  for (;; ptr++)
5797            }            }
5798          else          else
5799            {            {
5800            terminator = 0;            terminator = CHAR_NULL;
5801            if (ptr[1] == CHAR_MINUS || ptr[1] == CHAR_PLUS) refsign = *(++ptr);            if (ptr[1] == CHAR_MINUS || ptr[1] == CHAR_PLUS) refsign = *(++ptr);
5802            }            }
5803    
# Line 5786  for (;; ptr++) Line 5817  for (;; ptr++)
5817          while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0)          while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0)
5818            {            {
5819            if (recno >= 0)            if (recno >= 0)
5820              recno = (IS_DIGIT(*ptr))? recno * 10 + *ptr - CHAR_0 : -1;              recno = (IS_DIGIT(*ptr))? recno * 10 + (int)(*ptr - CHAR_0) : -1;
5821            ptr++;            ptr++;
5822            }            }
5823          namelen = (int)(ptr - name);          namelen = (int)(ptr - name);
5824    
5825          if ((terminator > 0 && *ptr++ != terminator) ||          if ((terminator > 0 && *ptr++ != (pcre_uchar)terminator) ||
5826              *ptr++ != CHAR_RIGHT_PARENTHESIS)              *ptr++ != CHAR_RIGHT_PARENTHESIS)
5827            {            {
5828            ptr--;      /* Error offset */            ptr--;      /* Error offset */
# Line 5856  for (;; ptr++) Line 5887  for (;; ptr++)
5887            code[1+LINK_SIZE]++;            code[1+LINK_SIZE]++;
5888            }            }
5889    
5890          /* If terminator == 0 it means that the name followed directly after          /* If terminator == CHAR_NULL it means that the name followed directly
5891          the opening parenthesis [e.g. (?(abc)...] and in this case there are          after the opening parenthesis [e.g. (?(abc)...] and in this case there
5892          some further alternatives to try. For the cases where terminator != 0          are some further alternatives to try. For the cases where terminator !=
5893          [things like (?(<name>... or (?('name')... or (?(R&name)... ] we have          0 [things like (?(<name>... or (?('name')... or (?(R&name)... ] we have
5894          now checked all the possibilities, so give an error. */          now checked all the possibilities, so give an error. */
5895    
5896          else if (terminator != 0)          else if (terminator != CHAR_NULL)
5897            {            {
5898            *errorcodeptr = ERR15;            *errorcodeptr = ERR15;
5899            goto FAILED;            goto FAILED;
# Line 6031  for (;; ptr++) Line 6062  for (;; ptr++)
6062    
6063            if (lengthptr != NULL)            if (lengthptr != NULL)
6064              {              {
6065              if (*ptr != terminator)              if (*ptr != (pcre_uchar)terminator)
6066                {                {
6067                *errorcodeptr = ERR42;                *errorcodeptr = ERR42;
6068                goto FAILED;                goto FAILED;
# Line 6173  for (;; ptr++) Line 6204  for (;; ptr++)
6204              *errorcodeptr = ERR62;              *errorcodeptr = ERR62;
6205              goto FAILED;              goto FAILED;
6206              }              }
6207            if (*ptr != terminator)            if (*ptr != (pcre_uchar)terminator)
6208              {              {
6209              *errorcodeptr = ERR42;              *errorcodeptr = ERR42;
6210              goto FAILED;              goto FAILED;
# Line 6279  for (;; ptr++) Line 6310  for (;; ptr++)
6310            while(IS_DIGIT(*ptr))            while(IS_DIGIT(*ptr))
6311              recno = recno * 10 + *ptr++ - CHAR_0;              recno = recno * 10 + *ptr++ - CHAR_0;
6312    
6313            if (*ptr != terminator)            if (*ptr != (pcre_uchar)terminator)
6314              {              {
6315              *errorcodeptr = ERR29;              *errorcodeptr = ERR29;
6316              goto FAILED;              goto FAILED;
# Line 6383  for (;; ptr++) Line 6414  for (;; ptr++)
6414    
6415          /* Can't determine a first byte now */          /* Can't determine a first byte now */
6416    
6417          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
6418          continue;          continue;
6419    
6420    
# Line 6517  for (;; ptr++) Line 6548  for (;; ptr++)
6548           cond_depth +           cond_depth +
6549             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */
6550           &subfirstchar,                   /* For possible first char */           &subfirstchar,                   /* For possible first char */
6551             &subfirstcharflags,
6552           &subreqchar,                     /* For possible last char */           &subreqchar,                     /* For possible last char */
6553             &subreqcharflags,
6554           bcptr,                           /* Current branch chain */           bcptr,                           /* Current branch chain */
6555           cd,                              /* Tables block */           cd,                              /* Tables block */
6556           (lengthptr == NULL)? NULL :      /* Actual compile phase */           (lengthptr == NULL)? NULL :      /* Actual compile phase */
# Line 6578  for (;; ptr++) Line 6611  for (;; ptr++)
6611            *errorcodeptr = ERR27;            *errorcodeptr = ERR27;
6612            goto FAILED;            goto FAILED;
6613            }            }
6614          if (condcount == 1) subfirstchar = subreqchar = REQ_NONE;          if (condcount == 1) subfirstcharflags = subreqcharflags = REQ_NONE;
6615          }          }
6616        }        }
6617    
# Line 6627  for (;; ptr++) Line 6660  for (;; ptr++)
6660      back off. */      back off. */
6661    
6662      zeroreqchar = reqchar;      zeroreqchar = reqchar;
6663        zeroreqcharflags = reqcharflags;
6664      zerofirstchar = firstchar;      zerofirstchar = firstchar;
6665        zerofirstcharflags = firstcharflags;
6666      groupsetfirstchar = FALSE;      groupsetfirstchar = FALSE;
6667    
6668      if (bravalue >= OP_ONCE)      if (bravalue >= OP_ONCE)
# Line 6638  for (;; ptr++) Line 6673  for (;; ptr++)
6673        no firstchar, set "none" for the whole branch. In both cases, a zero        no firstchar, set "none" for the whole branch. In both cases, a zero
6674        repeat forces firstchar to "none". */        repeat forces firstchar to "none". */
6675    
6676        if (firstchar == REQ_UNSET)        if (firstcharflags == REQ_UNSET)
6677          {          {
6678          if (subfirstchar >= 0)          if (subfirstcharflags >= 0)
6679            {            {
6680            firstchar = subfirstchar;            firstchar = subfirstchar;
6681              firstcharflags = subfirstcharflags;
6682            groupsetfirstchar = TRUE;            groupsetfirstchar = TRUE;
6683            }            }
6684          else firstchar = REQ_NONE;          else firstcharflags = REQ_NONE;
6685          zerofirstchar = REQ_NONE;          zerofirstcharflags = REQ_NONE;
6686          }          }
6687    
6688        /* If firstchar was previously set, convert the subpattern's firstchar        /* If firstchar was previously set, convert the subpattern's firstchar
6689        into reqchar if there wasn't one, using the vary flag that was in        into reqchar if there wasn't one, using the vary flag that was in
6690        existence beforehand. */        existence beforehand. */
6691    
6692        else if (subfirstchar >= 0 && subreqchar < 0)        else if (subfirstcharflags >= 0 && subreqcharflags < 0)
6693          subreqchar = subfirstchar | tempreqvary;          {
6694            subreqchar = subfirstchar;
6695            subreqcharflags = subfirstcharflags | tempreqvary;
6696            }
6697    
6698        /* If the subpattern set a required byte (or set a first byte that isn't        /* If the subpattern set a required byte (or set a first byte that isn't
6699        really the first byte - see above), set it. */        really the first byte - see above), set it. */
6700    
6701        if (subreqchar >= 0) reqchar = subreqchar;        if (subreqcharflags >= 0)
6702            {
6703            reqchar = subreqchar;
6704            reqcharflags = subreqcharflags;
6705            }
6706        }        }
6707    
6708      /* For a forward assertion, we take the reqchar, if set. This can be      /* For a forward assertion, we take the reqchar, if set. This can be
# Line 6670  for (;; ptr++) Line 6713  for (;; ptr++)
6713      of a firstchar. This is overcome by a scan at the end if there's no      of a firstchar. This is overcome by a scan at the end if there's no
6714      firstchar, looking for an asserted first char. */      firstchar, looking for an asserted first char. */
6715    
6716      else if (bravalue == OP_ASSERT && subreqchar >= 0) reqchar = subreqchar;      else if (bravalue == OP_ASSERT && subreqcharflags >= 0)
6717          {
6718          reqchar = subreqchar;
6719          reqcharflags = subreqcharflags;
6720          }
6721      break;     /* End of processing '(' */      break;     /* End of processing '(' */
6722    
6723    
# Line 6686  for (;; ptr++) Line 6733  for (;; ptr++)
6733      case CHAR_BACKSLASH:      case CHAR_BACKSLASH:
6734      tempptr = ptr;      tempptr = ptr;
6735      escape = check_escape(&ptr, &ec, errorcodeptr, cd->bracount, options, FALSE);      escape = check_escape(&ptr, &ec, errorcodeptr, cd->bracount, options, FALSE);
   
6736      if (*errorcodeptr != 0) goto FAILED;      if (*errorcodeptr != 0) goto FAILED;
6737    
6738      if (escape == 0)      if (escape == 0)                  /* The escape coded a single character */
6739        c = ec;        c = ec;
6740      else      else
6741        {        {
# Line 6706  for (;; ptr++) Line 6752  for (;; ptr++)
6752        /* For metasequences that actually match a character, we disable the        /* For metasequences that actually match a character, we disable the
6753        setting of a first character if it hasn't already been set. */        setting of a first character if it hasn't already been set. */
6754    
6755        if (firstchar == REQ_UNSET && escape > ESC_b && escape < ESC_Z)        if (firstcharflags == REQ_UNSET && escape > ESC_b && escape < ESC_Z)
6756          firstchar = REQ_NONE;          firstcharflags = REQ_NONE;
6757    
6758        /* Set values to reset to if this is followed by a zero repeat. */        /* Set values to reset to if this is followed by a zero repeat. */
6759    
6760        zerofirstchar = firstchar;        zerofirstchar = firstchar;
6761          zerofirstcharflags = firstcharflags;
6762        zeroreqchar = reqchar;        zeroreqchar = reqchar;
6763          zeroreqcharflags = reqcharflags;
6764    
6765        /* \g<name> or \g'name' is a subroutine call by name and \g<n> or \g'n'        /* \g<name> or \g'name' is a subroutine call by name and \g<n> or \g'n'
6766        is a subroutine call by number (Oniguruma syntax). In fact, the value        is a subroutine call by number (Oniguruma syntax). In fact, the value
# Line 6741  for (;; ptr++) Line 6789  for (;; ptr++)
6789          if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)          if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)
6790            {            {
6791            BOOL is_a_number = TRUE;            BOOL is_a_number = TRUE;
6792            for (p = ptr + 1; *p != 0 && *p != terminator; p++)            for (p = ptr + 1; *p != CHAR_NULL && *p != (pcre_uchar)terminator; p++)
6793              {              {
6794              if (!MAX_255(*p)) { is_a_number = FALSE; break; }              if (!MAX_255(*p)) { is_a_number = FALSE; break; }
6795              if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;              if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;
6796              if ((cd->ctypes[*p] & ctype_word) == 0) break;              if ((cd->ctypes[*p] & ctype_word) == 0) break;
6797              }              }
6798            if (*p != terminator)            if (*p != (pcre_uchar)terminator)
6799              {              {
6800              *errorcodeptr = ERR57;              *errorcodeptr = ERR57;
6801              break;              break;
# Line 6765  for (;; ptr++) Line 6813  for (;; ptr++)
6813    
6814          p = ptr + 2;          p = ptr + 2;
6815          while (IS_DIGIT(*p)) p++;          while (IS_DIGIT(*p)) p++;
6816          if (*p != terminator)          if (*p != (pcre_uchar)terminator)
6817            {            {
6818            *errorcodeptr = ERR57;            *errorcodeptr = ERR57;
6819            break;            break;
# Line 6802  for (;; ptr++) Line 6850  for (;; ptr++)
6850          recno = -escape;          recno = -escape;
6851    
6852          HANDLE_REFERENCE:    /* Come here from named backref handling */          HANDLE_REFERENCE:    /* Come here from named backref handling */
6853          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
6854          previous = code;          previous = code;
6855          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;
6856          PUT2INC(code, 0, recno);          PUT2INC(code, 0, recno);
# Line 6829  for (;; ptr++) Line 6877  for (;; ptr++)
6877        else if (escape == ESC_P || escape == ESC_p)        else if (escape == ESC_P || escape == ESC_p)
6878          {          {
6879          BOOL negated;          BOOL negated;
6880          int pdata;          unsigned int ptype = 0, pdata = 0;
6881          int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);          if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr))
6882          if (ptype < 0) goto FAILED;            goto FAILED;
6883          previous = code;          previous = code;
6884          *code++ = ((escape == ESC_p) != negated)? OP_PROP : OP_NOTPROP;          *code++ = ((escape == ESC_p) != negated)? OP_PROP : OP_NOTPROP;
6885          *code++ = ptype;          *code++ = ptype;
# Line 6853  for (;; ptr++) Line 6901  for (;; ptr++)
6901        can obtain the OP value by negating the escape value in the default        can obtain the OP value by negating the escape value in the default
6902        situation when PCRE_UCP is not set. When it *is* set, we substitute        situation when PCRE_UCP is not set. When it *is* set, we substitute
6903        Unicode property tests. Note that \b and \B do a one-character        Unicode property tests. Note that \b and \B do a one-character
6904        lookbehind. */        lookbehind, and \A also behaves as if it does. */
6905    
6906        else        else
6907          {          {
6908          if ((escape == ESC_b || escape == ESC_B) && cd->max_lookbehind == 0)          if ((escape == ESC_b || escape == ESC_B || escape == ESC_A) &&
6909                 cd->max_lookbehind == 0)
6910            cd->max_lookbehind = 1;            cd->max_lookbehind = 1;
6911  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6912          if (escape >= ESC_DU && escape <= ESC_wu)          if (escape >= ESC_DU && escape <= ESC_wu)
# Line 6905  for (;; ptr++) Line 6954  for (;; ptr++)
6954      mclength = 1;      mclength = 1;
6955      mcbuffer[0] = c;      mcbuffer[0] = c;
6956    
6957  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
6958      if (utf && HAS_EXTRALEN(c))      if (utf && HAS_EXTRALEN(c))
6959        ACROSSCHAR(TRUE, ptr[1], mcbuffer[mclength++] = *(++ptr));        ACROSSCHAR(TRUE, ptr[1], mcbuffer[mclength++] = *(++ptr));
6960  #endif  #endif
# Line 6915  for (;; ptr++) Line 6964  for (;; ptr++)
6964    
6965      ONE_CHAR:      ONE_CHAR:
6966      previous = code;      previous = code;
6967    
6968      /* For caseless UTF-8 mode when UCP support is available, check whether      /* For caseless UTF-8 mode when UCP support is available, check whether
6969      this character has more than one other case. If so, generate a special      this character has more than one other case. If so, generate a special
6970      OP_PROP item instead of OP_CHARI. */      OP_PROP item instead of OP_CHARI. */
6971    
6972  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6973      if (utf && (options & PCRE_CASELESS) != 0)      if (utf && (options & PCRE_CASELESS) != 0)
6974        {        {
# Line 6929  for (;; ptr++) Line 6978  for (;; ptr++)
6978          *code++ = OP_PROP;          *code++ = OP_PROP;
6979          *code++ = PT_CLIST;          *code++ = PT_CLIST;
6980          *code++ = c;          *code++ = c;
6981          if (firstchar == REQ_UNSET) firstchar = zerofirstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = zerofirstcharflags = REQ_NONE;
6982          break;          break;
6983          }          }
6984        }        }
6985  #endif  #endif
6986    
6987      /* Caseful matches, or not one of the multicase characters. */      /* Caseful matches, or not one of the multicase characters. */
6988    
6989      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR;      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR;
6990      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];
6991    
# Line 6950  for (;; ptr++) Line 6999  for (;; ptr++)
6999      Otherwise, leave the firstchar value alone, and don't change it on a zero      Otherwise, leave the firstchar value alone, and don't change it on a zero
7000      repeat. */      repeat. */
7001    
7002      if (firstchar == REQ_UNSET)      if (firstcharflags == REQ_UNSET)
7003        {        {
7004        zerofirstchar = REQ_NONE;        zerofirstcharflags = REQ_NONE;
7005        zeroreqchar = reqchar;        zeroreqchar = reqchar;
7006          zeroreqcharflags = reqcharflags;
7007    
7008        /* If the character is more than one byte long, we can set firstchar        /* If the character is more than one byte long, we can set firstchar
7009        only if it is not to be matched caselessly. */        only if it is not to be matched caselessly. */
# Line 6961  for (;; ptr++) Line 7011  for (;; ptr++)
7011        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
7012          {          {
7013          firstchar = mcbuffer[0] | req_caseopt;          firstchar = mcbuffer[0] | req_caseopt;
7014          if (mclength != 1) reqchar = code[-1] | cd->req_varyopt;          firstchar = mcbuffer[0];
7015            firstcharflags = req_caseopt;
7016    
7017            if (mclength != 1)
7018              {
7019              reqchar = code[-1];
7020              reqcharflags = cd->req_varyopt;
7021              }
7022          }          }
7023        else firstchar = reqchar = REQ_NONE;        else firstcharflags = reqcharflags = REQ_NONE;
7024        }        }
7025    
7026      /* firstchar was previously set; we can set reqchar only if the length is      /* firstchar was previously set; we can set reqchar only if the length is
# Line 6972  for (;; ptr++) Line 7029  for (;; ptr++)
7029      else      else
7030        {        {
7031        zerofirstchar = firstchar;        zerofirstchar = firstchar;
7032          zerofirstcharflags = firstcharflags;
7033        zeroreqchar = reqchar;        zeroreqchar = reqchar;
7034          zeroreqcharflags = reqcharflags;
7035        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
7036          reqchar = code[-1] | req_caseopt | cd->req_varyopt;          {
7037            reqchar = code[-1];
7038            reqcharflags = req_caseopt | cd->req_varyopt;
7039            }
7040        }        }
7041    
7042      break;            /* End of literal character handling */      break;            /* End of literal character handling */
# Line 6993  return FALSE; Line 7055  return FALSE;
7055    
7056    
7057    
   
7058  /*************************************************  /*************************************************
7059  *     Compile sequence of alternatives           *  *     Compile sequence of alternatives           *
7060  *************************************************/  *************************************************/
# Line 7014  Arguments: Line 7075  Arguments:
7075    reset_bracount TRUE to reset the count for each branch    reset_bracount TRUE to reset the count for each branch
7076    skipbytes      skip this many bytes at start (for brackets and OP_COND)    skipbytes      skip this many bytes at start (for brackets and OP_COND)
7077    cond_depth     depth of nesting for conditional subpatterns    cond_depth     depth of nesting for conditional subpatterns
7078    firstcharptr   place to put the first required character, or a negative number    firstcharptr    place to put the first required character
7079    reqcharptr     place to put the last required character, or a negative number    firstcharflagsptr place to put the first character flags, or a negative number
7080      reqcharptr     place to put the last required character
7081      reqcharflagsptr place to put the last required character flags, or a negative number
7082    bcptr          pointer to the chain of currently open branches    bcptr          pointer to the chain of currently open branches
7083    cd             points to the data block with tables pointers etc.    cd             points to the data block with tables pointers etc.
7084    lengthptr      NULL during the real compile phase    lengthptr      NULL during the real compile phase
# Line 7027  Returns:         TRUE on success Line 7090  Returns:         TRUE on success
7090  static BOOL  static BOOL
7091  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,
7092    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
7093    int cond_depth, pcre_int32 *firstcharptr, pcre_int32 *reqcharptr,    int cond_depth,
7094      pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr,
7095      pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr,
7096    branch_chain *bcptr, compile_data *cd, int *lengthptr)    branch_chain *bcptr, compile_data *cd, int *lengthptr)
7097  {  {
7098  const pcre_uchar *ptr = *ptrptr;  const pcre_uchar *ptr = *ptrptr;
# Line 7037  pcre_uchar *start_bracket = code; Line 7102  pcre_uchar *start_bracket = code;
7102  pcre_uchar *reverse_count = NULL;  pcre_uchar *reverse_count = NULL;
7103  open_capitem capitem;  open_capitem capitem;
7104  int capnumber = 0;  int capnumber = 0;
7105  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
7106  pcre_int32 branchfirstchar, branchreqchar;  pcre_int32 firstcharflags, reqcharflags;
7107    pcre_uint32 branchfirstchar, branchreqchar;
7108    pcre_int32 branchfirstcharflags, branchreqcharflags;
7109  int length;  int length;
7110  int orig_bracount;  unsigned int orig_bracount;
7111  int max_bracount;  unsigned int max_bracount;
7112  branch_chain bc;  branch_chain bc;
7113    
7114  bc.outer = bcptr;  bc.outer = bcptr;
7115  bc.current_branch = code;  bc.current_branch = code;
7116    
7117  firstchar = reqchar = REQ_UNSET;  firstchar = reqchar = 0;
7118    firstcharflags = reqcharflags = REQ_UNSET;
7119    
7120  /* Accumulate the length for use in the pre-compile phase. Start with the  /* Accumulate the length for use in the pre-compile phase. Start with the
7121  length of the BRA and KET and any extra bytes that are required at the  length of the BRA and KET and any extra bytes that are required at the
# Line 7107  for (;;) Line 7175  for (;;)
7175    into the length. */    into the length. */
7176    
7177    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,
7178          &branchreqchar, &bc, cond_depth, cd,          &branchfirstcharflags, &branchreqchar, &branchreqcharflags, &bc,
7179          (lengthptr == NULL)? NULL : &length))          cond_depth, cd, (lengthptr == NULL)? NULL : &length))
7180      {      {
7181      *ptrptr = ptr;      *ptrptr = ptr;
7182      return FALSE;      return FALSE;
# Line 7129  for (;;) Line 7197  for (;;)
7197      if (*last_branch != OP_ALT)      if (*last_branch != OP_ALT)
7198        {        {
7199        firstchar = branchfirstchar;        firstchar = branchfirstchar;
7200          firstcharflags = branchfirstcharflags;
7201        reqchar = branchreqchar;        reqchar = branchreqchar;
7202          reqcharflags = branchreqcharflags;
7203        }        }
7204    
7205      /* If this is not the first branch, the first char and reqchar have to      /* If this is not the first branch, the first char and reqchar have to
# Line 7143  for (;;) Line 7213  for (;;)
7213        we have to abandon the firstchar for the regex, but if there was        we have to abandon the firstchar for the regex, but if there was
7214        previously no reqchar, it takes on the value of the old firstchar. */        previously no reqchar, it takes on the value of the old firstchar. */
7215    
7216        if (firstchar >= 0 && firstchar != branchfirstchar)        if (firstcharflags >= 0 &&
7217              (firstcharflags != branchfirstcharflags || firstchar != branchfirstchar))
7218          {          {
7219          if (reqchar < 0) reqchar = firstchar;          if (reqcharflags < 0)
7220          firstchar = REQ_NONE;            {
7221              reqchar = firstchar;
7222              reqcharflags = firstcharflags;
7223              }
7224            firstcharflags = REQ_NONE;
7225          }          }
7226    
7227        /* If we (now or from before) have no firstchar, a firstchar from the        /* If we (now or from before) have no firstchar, a firstchar from the
7228        branch becomes a reqchar if there isn't a branch reqchar. */        branch becomes a reqchar if there isn't a branch reqchar. */
7229    
7230        if (firstchar < 0 && branchfirstchar >= 0 && branchreqchar < 0)        if (firstcharflags < 0 && branchfirstcharflags >= 0 && branchreqcharflags < 0)
7231            branchreqchar = branchfirstchar;          {
7232            branchreqchar = branchfirstchar;
7233            branchreqcharflags = branchfirstcharflags;
7234            }
7235    
7236        /* Now ensure that the reqchars match */        /* Now ensure that the reqchars match */
7237    
7238        if ((reqchar & ~REQ_VARY) != (branchreqchar & ~REQ_VARY))        if (((reqcharflags & ~REQ_VARY) != (branchreqcharflags & ~REQ_VARY)) ||
7239          reqchar = REQ_NONE;            reqchar != branchreqchar)
7240        else reqchar |= branchreqchar;   /* To "or" REQ_VARY */          reqcharflags = REQ_NONE;
7241          else
7242            {
7243            reqchar = branchreqchar;
7244            reqcharflags |= branchreqcharflags; /* To "or" REQ_VARY */
7245            }
7246        }        }
7247    
7248      /* If lookbehind, check that this branch matches a fixed-length string, and      /* If lookbehind, check that this branch matches a fixed-length string, and
# Line 7255  for (;;) Line 7338  for (;;)
7338      *codeptr = code;      *codeptr = code;
7339      *ptrptr = ptr;      *ptrptr = ptr;
7340      *firstcharptr = firstchar;      *firstcharptr = firstchar;
7341        *firstcharflagsptr = firstcharflags;
7342      *reqcharptr = reqchar;      *reqcharptr = reqchar;
7343        *reqcharflagsptr = reqcharflags;
7344      if (lengthptr != NULL)      if (lengthptr != NULL)
7345        {        {
7346        if (OFLOW_MAX - *lengthptr < length)        if (OFLOW_MAX - *lengthptr < length)
# Line 7490  do { Line 7575  do {
7575       {       {
7576       if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE;       if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE;
7577       }       }
7578    
7579     /* Atomic brackets */     /* Atomic brackets */
7580    
7581     else if (op == OP_ONCE || op == OP_ONCE_NC)     else if (op == OP_ONCE || op == OP_ONCE_NC)
# Line 7542  we return that char, otherwise -1. Line 7627  we return that char, otherwise -1.
7627    
7628  Arguments:  Arguments:
7629    code       points to start of expression (the bracket)    code       points to start of expression (the bracket)
7630      flags       points to the first char flags, or to REQ_NONE
7631    inassert   TRUE if in an assertion    inassert   TRUE if in an assertion
7632    
7633  Returns:     -1 or the fixed first char  Returns:     the fixed first char, or 0 with REQ_NONE in flags
7634  */  */
7635    
7636  static int  static pcre_uint32
7637  find_firstassertedchar(const pcre_uchar *code, BOOL inassert)  find_firstassertedchar(const pcre_uchar *code, pcre_int32 *flags,
7638      BOOL inassert)
7639  {  {
7640  register int c = -1;  register pcre_uint32 c = 0;
7641    int cflags = REQ_NONE;
7642    
7643    *flags = REQ_NONE;
7644  do {  do {
7645     int d;     pcre_uint32 d;
7646       int dflags;
7647     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||
7648               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0;               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0;
7649     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,
7650       TRUE);       TRUE);
7651     register int op = *scode;     register pcre_uchar op = *scode;
7652    
7653     switch(op)     switch(op)
7654       {       {
7655       default:       default:
7656       return -1;       return 0;
7657    
7658       case OP_BRA:       case OP_BRA:
7659       case OP_BRAPOS:       case OP_BRAPOS:
# Line 7574  do { Line 7665  do {
7665       case OP_ONCE:       case OP_ONCE:
7666       case OP_ONCE_NC:       case OP_ONCE_NC:
7667       case OP_COND:       case OP_COND:
7668       if ((d = find_firstassertedchar(scode, op == OP_ASSERT)) < 0)       d = find_firstassertedchar(scode, &dflags, op == OP_ASSERT);
7669         return -1;       if (dflags < 0)
7670       if (c < 0) c = d; else if (c != d) return -1;         return 0;
7671         if (cflags < 0) { c = d; cflags = dflags; } else if (c != d || cflags != dflags) return 0;
7672       break;       break;
7673    
7674       case OP_EXACT:       case OP_EXACT:
# Line 7587  do { Line 7679  do {
7679       case OP_PLUS:       case OP_PLUS:
7680       case OP_MINPLUS:       case OP_MINPLUS:
7681       case OP_POSPLUS:       case OP_POSPLUS:
7682       if (!inassert) return -1;       if (!inassert) return 0;
7683       if (c < 0) c = scode[1];       if (cflags < 0) { c = scode[1]; cflags = 0; }
7684         else if (c != scode[1]) return -1;         else if (c != scode[1]) return 0;
7685       break;       break;
7686    
7687       case OP_EXACTI:       case OP_EXACTI:
# Line 7600  do { Line 7692  do {
7692       case OP_PLUSI:       case OP_PLUSI:
7693       case OP_MINPLUSI:       case OP_MINPLUSI:
7694       case OP_POSPLUSI:       case OP_POSPLUSI:
7695       if (!inassert) return -1;       if (!inassert) return 0;
7696       if (c < 0) c = scode[1] | REQ_CASELESS;       if (cflags < 0) { c = scode[1]; cflags = REQ_CASELESS; }
7697         else if (c != scode[1]) return -1;         else if (c != scode[1]) return 0;
7698       break;       break;
7699       }       }
7700    
7701     code += GET(code, 1);     code += GET(code, 1);
7702     }     }
7703  while (*code == OP_ALT);  while (*code == OP_ALT);
7704    
7705    *flags = cflags;
7706  return c;  return c;
7707  }  }
7708    
# Line 7676  pcre32_compile2(PCRE_SPTR32 pattern, int Line 7770  pcre32_compile2(PCRE_SPTR32 pattern, int
7770  {  {
7771  REAL_PCRE *re;  REAL_PCRE *re;
7772  int length = 1;  /* For final END opcode */  int length = 1;  /* For final END opcode */
7773  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
7774    pcre_int32 firstcharflags, reqcharflags;
7775  int newline;  int newline;
7776  int errorcode = 0;  int errorcode = 0;
7777  int skipatstart = 0;  int skipatstart = 0;
# Line 7749  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 7844  while (ptr[skipatstart] == CHAR_LEFT_PAR
7844    int newnl = 0;    int newnl = 0;
7845    int newbsr = 0;    int newbsr = 0;
7846    
7847    /* For completeness and backward compatibility, (*UTFn) is supported in the
7848    relevant libraries, but (*UTF) is generic and always supported. Note that
7849    PCRE_UTF8 == PCRE_UTF16 == PCRE_UTF32. */
7850    
7851  #ifdef COMPILE_PCRE8  #ifdef COMPILE_PCRE8
7852    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 5) == 0)    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF8_RIGHTPAR, 5) == 0)
7853      { skipatstart += 7; options |= PCRE_UTF8; continue; }      { skipatstart += 7; options |= PCRE_UTF8; continue; }
7854  #endif  #endif
7855  #ifdef COMPILE_PCRE16  #ifdef COMPILE_PCRE16
7856    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF16_RIGHTPAR, 6) == 0)
7857      { skipatstart += 8; options |= PCRE_UTF16; continue; }      { skipatstart += 8; options |= PCRE_UTF16; continue; }
7858  #endif  #endif
7859  #ifdef COMPILE_PCRE32  #ifdef COMPILE_PCRE32
7860    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF32_RIGHTPAR, 6) == 0)
7861      { skipatstart += 8; options |= PCRE_UTF32; continue; }      { skipatstart += 8; options |= PCRE_UTF32; continue; }
7862  #endif  #endif
7863    
7864      else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 4) == 0)
7865        { skipatstart += 6; options |= PCRE_UTF8; continue; }
7866    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UCP_RIGHTPAR, 4) == 0)    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UCP_RIGHTPAR, 4) == 0)
7867      { skipatstart += 6; options |= PCRE_UCP; continue; }      { skipatstart += 6; options |= PCRE_UCP; continue; }
7868    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_NO_START_OPT_RIGHTPAR, 13) == 0)    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_NO_START_OPT_RIGHTPAR, 13) == 0)
# Line 7926  ptr += skipatstart; Line 8028  ptr += skipatstart;
8028  code = cworkspace;  code = cworkspace;
8029  *code = OP_BRA;  *code = OP_BRA;
8030  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,
8031    FALSE, 0, 0, &firstchar, &reqchar, NULL, cd, &length);    FALSE, 0, 0, &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL,
8032      cd, &length);
8033  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;
8034    
8035  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,
# Line 8004  ptr = (const pcre_uchar *)pattern + skip Line 8107  ptr = (const pcre_uchar *)pattern + skip
8107  code = (pcre_uchar *)codestart;  code = (pcre_uchar *)codestart;
8108  *code = OP_BRA;  *code = OP_BRA;
8109  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,
8110    &firstchar, &reqchar, NULL, cd, NULL);    &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL, cd, NULL);
8111  re->top_bracket = cd->bracount;  re->top_bracket = cd->bracount;
8112  re->top_backref = cd->top_backref;  re->top_backref = cd->top_backref;
8113  re->max_lookbehind = cd->max_lookbehind;  re->max_lookbehind = cd->max_lookbehind;
8114  re->flags = cd->external_flags | PCRE_MODE;  re->flags = cd->external_flags | PCRE_MODE;
8115    
8116  if (cd->had_accept) reqchar = REQ_NONE;   /* Must disable after (*ACCEPT) */  if (cd->had_accept)
8117      {
8118      reqchar = 0;              /* Must disable after (*ACCEPT) */
8119      reqcharflags = REQ_NONE;
8120      }
8121    
8122  /* If not reached end of pattern on success, there's an excess bracket. */  /* If not reached end of pattern on success, there's an excess bracket. */
8123    
8124  if (errorcode == 0 && *ptr != 0) errorcode = ERR22;  if (errorcode == 0 && *ptr != CHAR_NULL) errorcode = ERR22;
8125    
8126  /* Fill in the terminating state and check for disastrous overflow, but  /* Fill in the terminating state and check for disastrous overflow, but
8127  if debugging, leave the test till after things are printed out. */  if debugging, leave the test till after things are printed out. */
# Line 8025  if debugging, leave the test till after Line 8132  if debugging, leave the test till after
8132  if (code - codestart > length) errorcode = ERR23;  if (code - codestart > length) errorcode = ERR23;
8133  #endif  #endif
8134    
8135    #ifdef SUPPORT_VALGRIND
8136    /* If the estimated length exceeds the really used length, mark the extra
8137    allocated memory as unadressable, so that any out-of-bound reads can be
8138    detected. */
8139    VALGRIND_MAKE_MEM_NOACCESS(code, (length - (code - codestart)) * sizeof(pcre_uchar));
8140    #endif
8141    
8142  /* Fill in any forward references that are required. There may be repeated  /* Fill in any forward references that are required. There may be repeated
8143  references; optimize for them, as searching a large regex takes time. */  references; optimize for them, as searching a large regex takes time. */
8144    
# Line 8060  if (errorcode == 0 && re->top_backref > Line 8174  if (errorcode == 0 && re->top_backref >
8174    
8175  /* If there were any lookbehind assertions that contained OP_RECURSE  /* If there were any lookbehind assertions that contained OP_RECURSE
8176  (recursions or subroutine calls), a flag is set for them to be checked here,  (recursions or subroutine calls), a flag is set for them to be checked here,
8177  because they may contain forward references. Actual recursions can't be fixed  because they may contain forward references. Actual recursions cannot be fixed
8178  length, but subroutine calls can. It is done like this so that those without  length, but subroutine calls can. It is done like this so that those without
8179  OP_RECURSE that are not fixed length get a diagnosic with a useful offset. The  OP_RECURSE that are not fixed length get a diagnosic with a useful offset. The
8180  exceptional ones forgo this. We scan the pattern to check that they are fixed  exceptional ones forgo this. We scan the pattern to check that they are fixed
# Line 8131  if ((re->options & PCRE_ANCHORED) == 0) Line 8245  if ((re->options & PCRE_ANCHORED) == 0)
8245    if (is_anchored(codestart, 0, cd, 0)) re->options |= PCRE_ANCHORED;    if (is_anchored(codestart, 0, cd, 0)) re->options |= PCRE_ANCHORED;
8246    else    else
8247      {      {
8248      if (firstchar < 0)      if (firstcharflags < 0)
8249        firstchar = find_firstassertedchar(codestart, FALSE);        firstchar = find_firstassertedchar(codestart, &firstcharflags, FALSE);
8250      if (firstchar >= 0)   /* Remove caseless flag for non-caseable chars */      if (firstcharflags >= 0)   /* Remove caseless flag for non-caseable chars */
8251        {        {
8252  #if defined COMPILE_PCRE8  #if defined COMPILE_PCRE8
8253        re->first_char = firstchar & 0xff;        re->first_char = firstchar & 0xff;
8254  #elif defined COMPILE_PCRE16  #elif defined COMPILE_PCRE16
8255        re->first_char = firstchar & 0xffff;        re->first_char = firstchar & 0xffff;
8256  #elif defined COMPILE_PCRE32  #elif defined COMPILE_PCRE32
8257        re->first_char = firstchar & ~REQ_MASK;        re->first_char = firstchar;
8258  #endif  #endif
8259        if ((firstchar & REQ_CASELESS) != 0)        if ((firstcharflags & REQ_CASELESS) != 0)
8260          {          {
8261  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
8262          /* We ignore non-ASCII first chars in 8 bit mode. */          /* We ignore non-ASCII first chars in 8 bit mode. */
# Line 8174  if ((re->options & PCRE_ANCHORED) == 0) Line 8288  if ((re->options & PCRE_ANCHORED) == 0)
8288  variable length item in the regex. Remove the caseless flag for non-caseable  variable length item in the regex. Remove the caseless flag for non-caseable
8289  bytes. */  bytes. */
8290    
8291  if (reqchar >= 0 &&  if (reqcharflags >= 0 &&
8292       ((re->options & PCRE_ANCHORED) == 0 || (reqchar & REQ_VARY) != 0))       ((re->options & PCRE_ANCHORED) == 0 || (reqcharflags & REQ_VARY) != 0))
8293    {    {
8294  #if defined COMPILE_PCRE8  #if defined COMPILE_PCRE8
8295    re->req_char = reqchar & 0xff;    re->req_char = reqchar & 0xff;
8296  #elif defined COMPILE_PCRE16  #elif defined COMPILE_PCRE16
8297    re->req_char = reqchar & 0xffff;    re->req_char = reqchar & 0xffff;
8298  #elif defined COMPILE_PCRE32  #elif defined COMPILE_PCRE32
8299    re->req_char = reqchar & ~REQ_MASK;    re->req_char = reqchar;
8300  #endif  #endif
8301    if ((reqchar & REQ_CASELESS) != 0)    if ((reqcharflags & REQ_CASELESS) != 0)
8302      {      {
8303  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
8304      /* We ignore non-ASCII first chars in 8 bit mode. */      /* We ignore non-ASCII first chars in 8 bit mode. */

Legend:
Removed from v.1073  
changed lines
  Added in v.1266

  ViewVC Help
Powered by ViewVC 1.1.5