/[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 1059 by chpe, Tue Oct 16 15:53:53 2012 UTC revision 1231 by ph10, Fri Dec 7 12:05:53 2012 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 750  return (*p == CHAR_RIGHT_CURLY_BRACKET); Line 752  return (*p == CHAR_RIGHT_CURLY_BRACKET);
752    
753  /* This function is called when a \ has been encountered. It either returns a  /* This function is called when a \ has been encountered. It either returns a
754  positive value for a simple escape such as \n, or 0 for a data character  positive value for a simple escape such as \n, or 0 for a data character
755  which will be placed in chptr. A backreference to group  which will be placed in chptr. A backreference to group n is returned as
756  n is returned as ESC_REF + n; ESC_REF is the highest ESC_xxx macro. When  negative n. When UTF-8 is enabled, a positive value greater than 255 may
757  UTF-8 is enabled, a positive value greater than 255 may be returned in chptr.  be returned in chptr.
758  On entry,ptr is pointing at the \. On exit, it is on the final character of the  On entry,ptr is pointing at the \. On exit, it is on the final character of the
759  escape sequence.  escape sequence.
760    
# Line 766  Arguments: Line 768  Arguments:
768    
769  Returns:         zero => a data character  Returns:         zero => a data character
770                   positive => a special escape sequence                   positive => a special escape sequence
771                     negative => a back reference
772                   on error, errorcodeptr is set                   on error, errorcodeptr is set
773  */  */
774    
775  static int  static int
776  check_escape(const pcre_uchar **ptrptr, int *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. */
780  BOOL utf = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
781  const pcre_uchar *ptr = *ptrptr + 1;  const pcre_uchar *ptr = *ptrptr + 1;
782  pcre_int32 c;  pcre_uint32 c;
783  int escape = 0;  int escape = 0;
784  int i;  int i;
785    
# Line 785  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 794  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 = i; else escape = -i; }  else if ((i = escapes[c - CHAR_0]) != 0) { if (i > 0) c = (pcre_uint32)i; else escape = -i; }
801    
802  #else           /* EBCDIC coding */  #else           /* EBCDIC coding */
803  /* Not alphanumeric */  /* Not alphanumeric */
804  else if (c < CHAR_a || (!MAX_255(c) || (ebcdic_chartab[c] & 0x0E) == 0)) {}  else if (c < CHAR_a || (!MAX_255(c) || (ebcdic_chartab[c] & 0x0E) == 0)) {}
805  else if ((i = escapes[c - 0x48]) != 0)  { if (i > 0) c = i; else escape = -i; }  else if ((i = escapes[c - 0x48]) != 0)  { if (i > 0) c = (pcre_uint32)i; else escape = -i; }
806  #endif  #endif
807    
808  /* Escapes that need further processing, or are illegal. */  /* Escapes that need further processing, or are illegal. */
# Line 807  else if ((i = escapes[c - 0x48]) != 0) Line 810  else if ((i = escapes[c - 0x48]) != 0)
810  else  else
811    {    {
812    const pcre_uchar *oldptr;    const pcre_uchar *oldptr;
813    BOOL braced, negated;    BOOL braced, negated, overflow;
814      int s;
815    
816    switch (c)    switch (c)
817      {      {
# Line 832  else Line 836  else
836          c = 0;          c = 0;
837          for (i = 0; i < 4; ++i)          for (i = 0; i < 4; ++i)
838            {            {
839            register int cc = *(++ptr);            register pcre_uint32 cc = *(++ptr);
840  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
841            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
842            c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));            c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
# Line 894  else Line 898  else
898      if (ptr[1] == CHAR_LEFT_CURLY_BRACKET)      if (ptr[1] == CHAR_LEFT_CURLY_BRACKET)
899        {        {
900        const pcre_uchar *p;        const pcre_uchar *p;
901        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++)
902          if (*p != CHAR_MINUS && !IS_DIGIT(*p)) break;          if (*p != CHAR_MINUS && !IS_DIGIT(*p)) break;
903        if (*p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET)        if (*p != CHAR_NULL && *p != CHAR_RIGHT_CURLY_BRACKET)
904          {          {
905          escape = ESC_k;          escape = ESC_k;
906          break;          break;
# Line 914  else Line 918  else
918      else negated = FALSE;      else negated = FALSE;
919    
920      /* The integer range is limited by the machine's int representation. */      /* The integer range is limited by the machine's int representation. */
921      c = 0;      s = 0;
922        overflow = FALSE;
923      while (IS_DIGIT(ptr[1]))      while (IS_DIGIT(ptr[1]))
924        {        {
925        if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */        if (s > INT_MAX / 10 - 1) /* Integer overflow */
926          {          {
927          c = -1;          overflow = TRUE;
928          break;          break;
929          }          }
930        c = c * 10 + *(++ptr) - CHAR_0;        s = s * 10 + (int)(*(++ptr) - CHAR_0);
931        }        }
932      if (((unsigned int)c) > INT_MAX) /* Integer overflow */      if (overflow) /* Integer overflow */
933        {        {
934        while (IS_DIGIT(ptr[1]))        while (IS_DIGIT(ptr[1]))
935          ptr++;          ptr++;
# Line 938  else Line 943  else
943        break;        break;
944        }        }
945    
946      if (c == 0)      if (s == 0)
947        {        {
948        *errorcodeptr = ERR58;        *errorcodeptr = ERR58;
949        break;        break;
# Line 946  else Line 951  else
951    
952      if (negated)      if (negated)
953        {        {
954        if (c > bracount)        if (s > bracount)
955          {          {
956          *errorcodeptr = ERR15;          *errorcodeptr = ERR15;
957          break;          break;
958          }          }
959        c = bracount - (c - 1);        s = bracount - (s - 1);
960        }        }
961    
962      escape = ESC_REF + c;      escape = -s;
963      break;      break;
964    
965      /* The handling of escape sequences consisting of a string of digits      /* The handling of escape sequences consisting of a string of digits
# Line 976  else Line 981  else
981        {        {
982        oldptr = ptr;        oldptr = ptr;
983        /* The integer range is limited by the machine's int representation. */        /* The integer range is limited by the machine's int representation. */
984        c -= CHAR_0;        s = (int)(c -CHAR_0);
985          overflow = FALSE;
986        while (IS_DIGIT(ptr[1]))        while (IS_DIGIT(ptr[1]))
987          {          {
988          if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */          if (s > INT_MAX / 10 - 1) /* Integer overflow */
989            {            {
990            c = -1;            overflow = TRUE;
991            break;            break;
992            }            }
993          c = c * 10 + *(++ptr) - CHAR_0;          s = s * 10 + (int)(*(++ptr) - CHAR_0);
994          }          }
995        if (((unsigned int)c) > INT_MAX) /* Integer overflow */        if (overflow) /* Integer overflow */
996          {          {
997          while (IS_DIGIT(ptr[1]))          while (IS_DIGIT(ptr[1]))
998            ptr++;            ptr++;
999          *errorcodeptr = ERR61;          *errorcodeptr = ERR61;
1000          break;          break;
1001          }          }
1002        if (c < 10 || c <= bracount)        if (s < 10 || s <= bracount)
1003          {          {
1004          escape = ESC_REF + c;          escape = -s;
1005          break;          break;
1006          }          }
1007        ptr = oldptr;      /* Put the pointer back and fall through */        ptr = oldptr;      /* Put the pointer back and fall through */
# Line 1042  else Line 1048  else
1048          c = 0;          c = 0;
1049          for (i = 0; i < 2; ++i)          for (i = 0; i < 2; ++i)
1050            {            {
1051            register int cc = *(++ptr);            register pcre_uint32 cc = *(++ptr);
1052  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
1053            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
1054            c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));            c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
# Line 1060  else Line 1066  else
1066        const pcre_uchar *pt = ptr + 2;        const pcre_uchar *pt = ptr + 2;
1067    
1068        c = 0;        c = 0;
1069          overflow = FALSE;
1070        while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0)        while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0)
1071          {          {
1072          register int cc = *pt++;          register pcre_uint32 cc = *pt++;
1073          if (c == 0 && cc == CHAR_0) continue;     /* Leading zeroes */          if (c == 0 && cc == CHAR_0) continue;     /* Leading zeroes */
1074    
1075    #ifdef COMPILE_PCRE32
1076            if (c >= 0x10000000l) { overflow = TRUE; break; }
1077    #endif
1078    
1079  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
1080          if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */          if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
1081          c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));          c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
# Line 1074  else Line 1085  else
1085  #endif  #endif
1086    
1087  #if defined COMPILE_PCRE8  #if defined COMPILE_PCRE8
1088          if (c > (utf ? 0x10ffff : 0xff)) { c = -1; break; }          if (c > (utf ? 0x10ffff : 0xff)) { overflow = TRUE; break; }
1089  #elif defined COMPILE_PCRE16  #elif defined COMPILE_PCRE16
1090          if (c > (utf ? 0x10ffff : 0xffff)) { c = -1; break; }          if (c > (utf ? 0x10ffff : 0xffff)) { overflow = TRUE; break; }
1091  #elif defined COMPILE_PCRE32  #elif defined COMPILE_PCRE32
1092          if (utf && c > 0x10ffff) { c = -1; break; }          if (utf && c > 0x10ffff) { overflow = TRUE; break; }
1093  #endif  #endif
1094          }          }
1095    
1096        if (c < 0)        if (overflow)
1097          {          {
1098          while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0) pt++;          while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0) pt++;
1099          *errorcodeptr = ERR34;          *errorcodeptr = ERR34;
# Line 1104  else Line 1115  else
1115      c = 0;      c = 0;
1116      while (i++ < 2 && MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0)      while (i++ < 2 && MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0)
1117        {        {
1118        int cc;                                  /* Some compilers don't like */        pcre_uint32 cc;                          /* Some compilers don't like */
1119        cc = *(++ptr);                           /* ++ in initializers */        cc = *(++ptr);                           /* ++ in initializers */
1120  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
1121        if (cc >= CHAR_a) cc -= 32;              /* Convert to upper case */        if (cc >= CHAR_a) cc -= 32;              /* Convert to upper case */
# Line 1123  else Line 1134  else
1134    
1135      case CHAR_c:      case CHAR_c:
1136      c = *(++ptr);      c = *(++ptr);
1137      if (c == 0)      if (c == CHAR_NULL)
1138        {        {
1139        *errorcodeptr = ERR2;        *errorcodeptr = ERR2;
1140        break;        break;
# Line 1192  escape sequence. Line 1203  escape sequence.
1203  Argument:  Argument:
1204    ptrptr         points to the pattern position pointer    ptrptr         points to the pattern position pointer
1205    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
1206    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
1207      pdataptr       points to an unsigned int that is set to the detailed property value
1208    errorcodeptr   points to the error code variable    errorcodeptr   points to the error code variable
1209    
1210  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
1211  */  */
1212    
1213  static int  static BOOL
1214  get_ucp(const pcre_uchar **ptrptr, BOOL *negptr, int *dptr, int *errorcodeptr)  get_ucp(const pcre_uchar **ptrptr, BOOL *negptr, unsigned int *ptypeptr,
1215      unsigned int *pdataptr, int *errorcodeptr)
1216  {  {
1217  int c, i, bot, top;  pcre_uchar c;
1218    int i, bot, top;
1219  const pcre_uchar *ptr = *ptrptr;  const pcre_uchar *ptr = *ptrptr;
1220  pcre_uchar name[32];  pcre_uchar name[32];
1221    
1222  c = *(++ptr);  c = *(++ptr);
1223  if (c == 0) goto ERROR_RETURN;  if (c == CHAR_NULL) goto ERROR_RETURN;
1224    
1225  *negptr = FALSE;  *negptr = FALSE;
1226    
# Line 1223  if (c == CHAR_LEFT_CURLY_BRACKET) Line 1237  if (c == CHAR_LEFT_CURLY_BRACKET)
1237    for (i = 0; i < (int)(sizeof(name) / sizeof(pcre_uchar)) - 1; i++)    for (i = 0; i < (int)(sizeof(name) / sizeof(pcre_uchar)) - 1; i++)
1238      {      {
1239      c = *(++ptr);      c = *(++ptr);
1240      if (c == 0) goto ERROR_RETURN;      if (c == CHAR_NULL) goto ERROR_RETURN;
1241      if (c == CHAR_RIGHT_CURLY_BRACKET) break;      if (c == CHAR_RIGHT_CURLY_BRACKET) break;
1242      name[i] = c;      name[i] = c;
1243      }      }
# Line 1248  top = PRIV(utt_size); Line 1262  top = PRIV(utt_size);
1262    
1263  while (bot < top)  while (bot < top)
1264    {    {
1265      int r;
1266    i = (bot + top) >> 1;    i = (bot + top) >> 1;
1267    c = 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);
1268    if (c == 0)    if (r == 0)
1269      {      {
1270      *dptr = PRIV(utt)[i].value;      *ptypeptr = PRIV(utt)[i].type;
1271      return PRIV(utt)[i].type;      *pdataptr = PRIV(utt)[i].value;
1272        return TRUE;
1273      }      }
1274    if (c > 0) bot = i + 1; else top = i;    if (r > 0) bot = i + 1; else top = i;
1275    }    }
1276    
1277  *errorcodeptr = ERR47;  *errorcodeptr = ERR47;
1278  *ptrptr = ptr;  *ptrptr = ptr;
1279  return -1;  return FALSE;
1280    
1281  ERROR_RETURN:  ERROR_RETURN:
1282  *errorcodeptr = ERR46;  *errorcodeptr = ERR46;
1283  *ptrptr = ptr;  *ptrptr = ptr;
1284  return -1;  return FALSE;
1285  }  }
1286  #endif  #endif
1287    
# Line 1300  int max = -1; Line 1316  int max = -1;
1316  /* Read the minimum value and do a paranoid check: a negative value indicates  /* Read the minimum value and do a paranoid check: a negative value indicates
1317  an integer overflow. */  an integer overflow. */
1318    
1319  while (IS_DIGIT(*p)) min = min * 10 + *p++ - CHAR_0;  while (IS_DIGIT(*p)) min = min * 10 + (int)(*p++ - CHAR_0);
1320  if (min < 0 || min > 65535)  if (min < 0 || min > 65535)
1321    {    {
1322    *errorcodeptr = ERR5;    *errorcodeptr = ERR5;
# Line 1315  if (*p == CHAR_RIGHT_CURLY_BRACKET) max Line 1331  if (*p == CHAR_RIGHT_CURLY_BRACKET) max
1331    if (*(++p) != CHAR_RIGHT_CURLY_BRACKET)    if (*(++p) != CHAR_RIGHT_CURLY_BRACKET)
1332      {      {
1333      max = 0;      max = 0;
1334      while(IS_DIGIT(*p)) max = max * 10 + *p++ - CHAR_0;      while(IS_DIGIT(*p)) max = max * 10 + (int)(*p++ - CHAR_0);
1335      if (max < 0 || max > 65535)      if (max < 0 || max > 65535)
1336        {        {
1337        *errorcodeptr = ERR5;        *errorcodeptr = ERR5;
# Line 1416  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1432  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1432    
1433    else if (ptr[2] == CHAR_NUMBER_SIGN)    else if (ptr[2] == CHAR_NUMBER_SIGN)
1434      {      {
1435      for (ptr += 3; *ptr != 0; ptr++) if (*ptr == CHAR_RIGHT_PARENTHESIS) break;      for (ptr += 3; *ptr != CHAR_NULL; ptr++)
1436          if (*ptr == CHAR_RIGHT_PARENTHESIS) break;
1437      goto FAIL_EXIT;      goto FAIL_EXIT;
1438      }      }
1439    
# Line 1429  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1446  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1446      ptr += 2;      ptr += 2;
1447      if (ptr[1] != CHAR_QUESTION_MARK)      if (ptr[1] != CHAR_QUESTION_MARK)
1448        {        {
1449        while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;        while (*ptr != CHAR_NULL && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
1450        if (*ptr != 0) ptr++;        if (*ptr != CHAR_NULL) ptr++;
1451        }        }
1452      }      }
1453    
# Line 1446  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1463  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1463      if ((*ptr == CHAR_LESS_THAN_SIGN && ptr[1] != CHAR_EXCLAMATION_MARK &&      if ((*ptr == CHAR_LESS_THAN_SIGN && ptr[1] != CHAR_EXCLAMATION_MARK &&
1464          ptr[1] != CHAR_EQUALS_SIGN) || *ptr == CHAR_APOSTROPHE)          ptr[1] != CHAR_EQUALS_SIGN) || *ptr == CHAR_APOSTROPHE)
1465        {        {
1466        int term;        pcre_uchar term;
1467        const pcre_uchar *thisname;        const pcre_uchar *thisname;
1468        *count += 1;        *count += 1;
1469        if (name == NULL && *count == lorn) return *count;        if (name == NULL && *count == lorn) return *count;
# Line 1454  if (ptr[0] == CHAR_LEFT_PARENTHESIS) Line 1471  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
1471        if (term == CHAR_LESS_THAN_SIGN) term = CHAR_GREATER_THAN_SIGN;        if (term == CHAR_LESS_THAN_SIGN) term = CHAR_GREATER_THAN_SIGN;
1472        thisname = ptr;        thisname = ptr;
1473        while (*ptr != term) ptr++;        while (*ptr != term) ptr++;
1474        if (name != NULL && lorn == ptr - thisname &&        if (name != NULL && lorn == (int)(ptr - thisname) &&
1475            STRNCMP_UC_UC(name, thisname, lorn) == 0)            STRNCMP_UC_UC(name, thisname, (unsigned int)lorn) == 0)
1476          return *count;          return *count;
1477        term++;        term++;
1478        }        }
# Line 1473  for (; ptr < cd->end_pattern; ptr++) Line 1490  for (; ptr < cd->end_pattern; ptr++)
1490    
1491    if (*ptr == CHAR_BACKSLASH)    if (*ptr == CHAR_BACKSLASH)
1492      {      {
1493      if (*(++ptr) == 0) goto FAIL_EXIT;      if (*(++ptr) == CHAR_NULL) goto FAIL_EXIT;
1494      if (*ptr == CHAR_Q) for (;;)      if (*ptr == CHAR_Q) for (;;)
1495        {        {
1496        while (*(++ptr) != 0 && *ptr != CHAR_BACKSLASH) {};        while (*(++ptr) != CHAR_NULL && *ptr != CHAR_BACKSLASH) {};
1497        if (*ptr == 0) goto FAIL_EXIT;        if (*ptr == CHAR_NULL) goto FAIL_EXIT;
1498        if (*(++ptr) == CHAR_E) break;        if (*(++ptr) == CHAR_E) break;
1499        }        }
1500      continue;      continue;
# Line 1521  for (; ptr < cd->end_pattern; ptr++) Line 1538  for (; ptr < cd->end_pattern; ptr++)
1538    
1539      while (*(++ptr) != CHAR_RIGHT_SQUARE_BRACKET)      while (*(++ptr) != CHAR_RIGHT_SQUARE_BRACKET)
1540        {        {
1541        if (*ptr == 0) return -1;        if (*ptr == CHAR_NULL) return -1;
1542        if (*ptr == CHAR_BACKSLASH)        if (*ptr == CHAR_BACKSLASH)
1543          {          {
1544          if (*(++ptr) == 0) goto FAIL_EXIT;          if (*(++ptr) == CHAR_NULL) goto FAIL_EXIT;
1545          if (*ptr == CHAR_Q) for (;;)          if (*ptr == CHAR_Q) for (;;)
1546            {            {
1547            while (*(++ptr) != 0 && *ptr != CHAR_BACKSLASH) {};            while (*(++ptr) != CHAR_NULL && *ptr != CHAR_BACKSLASH) {};
1548            if (*ptr == 0) goto FAIL_EXIT;            if (*ptr == CHAR_NULL) goto FAIL_EXIT;
1549            if (*(++ptr) == CHAR_E) break;            if (*(++ptr) == CHAR_E) break;
1550            }            }
1551          continue;          continue;
# Line 1542  for (; ptr < cd->end_pattern; ptr++) Line 1559  for (; ptr < cd->end_pattern; ptr++)
1559    if (xmode && *ptr == CHAR_NUMBER_SIGN)    if (xmode && *ptr == CHAR_NUMBER_SIGN)
1560      {      {
1561      ptr++;      ptr++;
1562      while (*ptr != 0)      while (*ptr != CHAR_NULL)
1563        {        {
1564        if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }        if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
1565        ptr++;        ptr++;
# Line 1550  for (; ptr < cd->end_pattern; ptr++) Line 1567  for (; ptr < cd->end_pattern; ptr++)
1567        if (utf) FORWARDCHAR(ptr);        if (utf) FORWARDCHAR(ptr);
1568  #endif  #endif
1569        }        }
1570      if (*ptr == 0) goto FAIL_EXIT;      if (*ptr == CHAR_NULL) goto FAIL_EXIT;
1571      continue;      continue;
1572      }      }
1573    
# Line 1560  for (; ptr < cd->end_pattern; ptr++) Line 1577  for (; ptr < cd->end_pattern; ptr++)
1577      {      {
1578      int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, count);      int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, count);
1579      if (rc > 0) return rc;      if (rc > 0) return rc;
1580      if (*ptr == 0) goto FAIL_EXIT;      if (*ptr == CHAR_NULL) goto FAIL_EXIT;
1581      }      }
1582    
1583    else if (*ptr == CHAR_RIGHT_PARENTHESIS)    else if (*ptr == CHAR_RIGHT_PARENTHESIS)
# Line 1625  matching closing parens. That is why we Line 1642  matching closing parens. That is why we
1642  for (;;)  for (;;)
1643    {    {
1644    rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, &count);    rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, &count);
1645    if (rc > 0 || *ptr++ == 0) break;    if (rc > 0 || *ptr++ == CHAR_NULL) break;
1646    }    }
1647    
1648  return rc;  return rc;
# Line 1733  for (;;) Line 1750  for (;;)
1750    {    {
1751    int d;    int d;
1752    pcre_uchar *ce, *cs;    pcre_uchar *ce, *cs;
1753    register int op = *cc;    register pcre_uchar op = *cc;
1754    
1755    switch (op)    switch (op)
1756      {      {
# Line 1841  for (;;) Line 1858  for (;;)
1858      case OP_NOTI:      case OP_NOTI:
1859      branchlength++;      branchlength++;
1860      cc += 2;      cc += 2;
1861  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
1862      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
1863  #endif  #endif
1864      break;      break;
# Line 1853  for (;;) Line 1870  for (;;)
1870      case OP_EXACTI:      case OP_EXACTI:
1871      case OP_NOTEXACT:      case OP_NOTEXACT:
1872      case OP_NOTEXACTI:      case OP_NOTEXACTI:
1873      branchlength += GET2(cc,1);      branchlength += (int)GET2(cc,1);
1874      cc += 2 + IMM2_SIZE;      cc += 2 + IMM2_SIZE;
1875  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
1876      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
1877  #endif  #endif
1878      break;      break;
1879    
1880      case OP_TYPEEXACT:      case OP_TYPEEXACT:
1881      branchlength += GET2(cc,1);      branchlength += GET2(cc,1);
1882      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)
1883        cc += 2;        cc += 2;
1884      cc += 1 + IMM2_SIZE + 1;      cc += 1 + IMM2_SIZE + 1;
1885      break;      break;
# Line 1898  for (;;) Line 1915  for (;;)
1915    
1916      /* Check a class for variable quantification */      /* Check a class for variable quantification */
1917    
 #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  
   
1918      case OP_CLASS:      case OP_CLASS:
1919      case OP_NCLASS:      case OP_NCLASS:
1920    #if defined SUPPORT_UTF || defined COMPILE_PCRE16 || defined COMPILE_PCRE32
1921        case OP_XCLASS:
1922        /* The original code caused an unsigned overflow in 64 bit systems,
1923        so now we use a conditional statement. */
1924        if (op == OP_XCLASS)
1925          cc += GET(cc, 1);
1926        else
1927          cc += PRIV(OP_lengths)[OP_CLASS];
1928    #else
1929      cc += PRIV(OP_lengths)[OP_CLASS];      cc += PRIV(OP_lengths)[OP_CLASS];
1930    #endif
1931    
1932      switch (*cc)      switch (*cc)
1933        {        {
# Line 1921  for (;;) Line 1942  for (;;)
1942        case OP_CRRANGE:        case OP_CRRANGE:
1943        case OP_CRMINRANGE:        case OP_CRMINRANGE:
1944        if (GET2(cc,1) != GET2(cc,1+IMM2_SIZE)) return -1;        if (GET2(cc,1) != GET2(cc,1+IMM2_SIZE)) return -1;
1945        branchlength += GET2(cc,1);        branchlength += (int)GET2(cc,1);
1946        cc += 1 + 2 * IMM2_SIZE;        cc += 1 + 2 * IMM2_SIZE;
1947        break;        break;
1948    
# Line 2048  PRIV(find_bracket)(const pcre_uchar *cod Line 2069  PRIV(find_bracket)(const pcre_uchar *cod
2069  {  {
2070  for (;;)  for (;;)
2071    {    {
2072    register int c = *code;    register pcre_uchar c = *code;
2073    
2074    if (c == OP_END) return NULL;    if (c == OP_END) return NULL;
2075    
# Line 2071  for (;;) Line 2092  for (;;)
2092    else if (c == OP_CBRA || c == OP_SCBRA ||    else if (c == OP_CBRA || c == OP_SCBRA ||
2093             c == OP_CBRAPOS || c == OP_SCBRAPOS)             c == OP_CBRAPOS || c == OP_SCBRAPOS)
2094      {      {
2095      int n = GET2(code, 1+LINK_SIZE);      int n = (int)GET2(code, 1+LINK_SIZE);
2096      if (n == number) return (pcre_uchar *)code;      if (n == number) return (pcre_uchar *)code;
2097      code += PRIV(OP_lengths)[c];      code += PRIV(OP_lengths)[c];
2098      }      }
# Line 2101  for (;;) Line 2122  for (;;)
2122        case OP_TYPEMINUPTO:        case OP_TYPEMINUPTO:
2123        case OP_TYPEEXACT:        case OP_TYPEEXACT:
2124        case OP_TYPEPOSUPTO:        case OP_TYPEPOSUPTO:
2125        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)
2126          code += 2;          code += 2;
2127        break;        break;
2128    
# Line 2186  find_recurse(const pcre_uchar *code, BOO Line 2207  find_recurse(const pcre_uchar *code, BOO
2207  {  {
2208  for (;;)  for (;;)
2209    {    {
2210    register int c = *code;    register pcre_uchar c = *code;
2211    if (c == OP_END) return NULL;    if (c == OP_END) return NULL;
2212    if (c == OP_RECURSE) return code;    if (c == OP_RECURSE) return code;
2213    
# Line 2221  for (;;) Line 2242  for (;;)
2242        case OP_TYPEUPTO:        case OP_TYPEUPTO:
2243        case OP_TYPEMINUPTO:        case OP_TYPEMINUPTO:
2244        case OP_TYPEEXACT:        case OP_TYPEEXACT:
2245        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)
2246          code += 2;          code += 2;
2247        break;        break;
2248    
# Line 2340  static BOOL Line 2361  static BOOL
2361  could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode,  could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode,
2362    BOOL utf, compile_data *cd)    BOOL utf, compile_data *cd)
2363  {  {
2364  register int c;  register pcre_uchar c;
2365  for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE);  for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE);
2366       code < endcode;       code < endcode;
2367       code = first_significant_code(code + PRIV(OP_lengths)[c], TRUE))       code = first_significant_code(code + PRIV(OP_lengths)[c], TRUE))
# Line 2374  for (code = first_significant_code(code Line 2395  for (code = first_significant_code(code
2395      /* Test for forward reference */      /* Test for forward reference */
2396    
2397      for (scode = cd->start_workspace; scode < cd->hwm; scode += LINK_SIZE)      for (scode = cd->start_workspace; scode < cd->hwm; scode += LINK_SIZE)
2398        if (GET(scode, 0) == code + 1 - cd->start_code) return TRUE;        if ((int)GET(scode, 0) == (int)(code + 1 - cd->start_code)) return TRUE;
2399    
2400      /* Not a forward reference, test for completed backward reference */      /* Not a forward reference, test for completed backward reference */
2401    
# Line 2547  for (code = first_significant_code(code Line 2568  for (code = first_significant_code(code
2568      case OP_TYPEUPTO:      case OP_TYPEUPTO:
2569      case OP_TYPEMINUPTO:      case OP_TYPEMINUPTO:
2570      case OP_TYPEPOSUPTO:      case OP_TYPEPOSUPTO:
2571      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)
2572        code += 2;        code += 2;
2573      break;      break;
2574    
# Line 2695  Returns:   TRUE or FALSE Line 2716  Returns:   TRUE or FALSE
2716  static BOOL  static BOOL
2717  check_posix_syntax(const pcre_uchar *ptr, const pcre_uchar **endptr)  check_posix_syntax(const pcre_uchar *ptr, const pcre_uchar **endptr)
2718  {  {
2719  int terminator;          /* Don't combine these lines; the Solaris cc */  pcre_uchar terminator;          /* Don't combine these lines; the Solaris cc */
2720  terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */  terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */
2721  for (++ptr; *ptr != 0; ptr++)  for (++ptr; *ptr != CHAR_NULL; ptr++)
2722    {    {
2723    if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)    if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)
2724      ptr++;      ptr++;
# Line 2744  register int yield = 0; Line 2765  register int yield = 0;
2765  while (posix_name_lengths[yield] != 0)  while (posix_name_lengths[yield] != 0)
2766    {    {
2767    if (len == posix_name_lengths[yield] &&    if (len == posix_name_lengths[yield] &&
2768      STRNCMP_UC_C8(ptr, pn, len) == 0) return yield;      STRNCMP_UC_C8(ptr, pn, (unsigned int)len) == 0) return yield;
2769    pn += posix_name_lengths[yield] + 1;    pn += posix_name_lengths[yield] + 1;
2770    yield++;    yield++;
2771    }    }
# Line 2799  while ((ptr = (pcre_uchar *)find_recurse Line 2820  while ((ptr = (pcre_uchar *)find_recurse
2820    
2821    for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)    for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE)
2822      {      {
2823      offset = GET(hc, 0);      offset = (int)GET(hc, 0);
2824      if (cd->start_code + offset == ptr + 1)      if (cd->start_code + offset == ptr + 1)
2825        {        {
2826        PUT(hc, 0, offset + adjust);        PUT(hc, 0, offset + adjust);
# Line 2812  while ((ptr = (pcre_uchar *)find_recurse Line 2833  while ((ptr = (pcre_uchar *)find_recurse
2833    
2834    if (hc >= cd->hwm)    if (hc >= cd->hwm)
2835      {      {
2836      offset = GET(ptr, 1);      offset = (int)GET(ptr, 1);
2837      if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust);      if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust);
2838      }      }
2839    
# Line 2882  PUT(previous_callout, 2 + LINK_SIZE, len Line 2903  PUT(previous_callout, 2 + LINK_SIZE, len
2903  /* 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
2904  with UCP support. It searches up the characters, looking for ranges of  with UCP support. It searches up the characters, looking for ranges of
2905  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
2906  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
2907  with a special return value.  with a special return value.
2908    
2909  Arguments:  Arguments:
# Line 2894  Arguments: Line 2915  Arguments:
2915  Yield:        -1 when no more  Yield:        -1 when no more
2916                 0 when a range is returned                 0 when a range is returned
2917                >0 the CASESET offset for char with multiple other cases                >0 the CASESET offset for char with multiple other cases
2918                  in this case, ocptr contains the original                  in this case, ocptr contains the original
2919  */  */
2920    
2921  static int  static int
2922  get_othercase_range(unsigned int *cptr, unsigned int d, unsigned int *ocptr,  get_othercase_range(pcre_uint32 *cptr, pcre_uint32 d, pcre_uint32 *ocptr,
2923    unsigned int *odptr)    pcre_uint32 *odptr)
2924  {  {
2925  unsigned int c, othercase, next;  pcre_uint32 c, othercase, next;
2926  int co;  unsigned int co;
2927    
2928  /* 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
2929  cases, return its case offset value. */  cases, return its case offset value. */
2930    
2931  for (c = *cptr; c <= d; c++)  for (c = *cptr; c <= d; c++)
2932    {    {
2933    if ((co = UCD_CASESET(c)) != 0)    if ((co = UCD_CASESET(c)) != 0)
2934      {      {
2935      *ocptr = c++;   /* Character that has the set */      *ocptr = c++;   /* Character that has the set */
2936      *cptr = c;      /* Rest of input range */      *cptr = c;      /* Rest of input range */
2937      return co;      return (int)co;
2938      }      }
2939    if ((othercase = UCD_OTHERCASE(c)) != c) break;    if ((othercase = UCD_OTHERCASE(c)) != c) break;
2940    }    }
2941    
2942  if (c > d) return -1;  /* Reached end of range */  if (c > d) return -1;  /* Reached end of range */
# Line 2953  Returns:       TRUE if auto-possessifyin Line 2974  Returns:       TRUE if auto-possessifyin
2974  */  */
2975    
2976  static BOOL  static BOOL
2977  check_char_prop(int c, int ptype, int pdata, BOOL negated)  check_char_prop(pcre_uint32 c, unsigned int ptype, unsigned int pdata, BOOL negated)
2978  {  {
2979  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2980  const pcre_uint32 *p;  const pcre_uint32 *p;
# Line 2998  switch(ptype) Line 3019  switch(ptype)
3019    return (PRIV(ucp_gentype)[prop->chartype] == ucp_L ||    return (PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
3020            PRIV(ucp_gentype)[prop->chartype] == ucp_N ||            PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
3021            c == CHAR_UNDERSCORE) == negated;            c == CHAR_UNDERSCORE) == negated;
3022    
3023  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3024    case PT_CLIST:    case PT_CLIST:
3025    p = PRIV(ucd_caseless_sets) + prop->caseset;    p = PRIV(ucd_caseless_sets) + prop->caseset;
3026    for (;;)    for (;;)
3027      {      {
3028      if ((unsigned int)c < *p) return !negated;      if (c < *p) return !negated;
3029      if ((unsigned int)c == *p++) return negated;      if (c == *p++) return negated;
3030      }      }
3031    break;  /* Control never reaches here */    break;  /* Control never reaches here */
3032  #endif  #endif
# Line 3039  static BOOL Line 3060  static BOOL
3060  check_auto_possessive(const pcre_uchar *previous, BOOL utf,  check_auto_possessive(const pcre_uchar *previous, BOOL utf,
3061    const pcre_uchar *ptr, int options, compile_data *cd)    const pcre_uchar *ptr, int options, compile_data *cd)
3062  {  {
3063  pcre_int32 c = NOTACHAR; // FIXMEchpe pcre_uint32  pcre_uint32 c = NOTACHAR;
3064  pcre_int32 next;  pcre_uint32 next;
3065  int escape;  int escape;
3066  int op_code = *previous++;  pcre_uchar op_code = *previous++;
3067    
3068  /* Skip whitespace and comments in extended mode */  /* Skip whitespace and comments in extended mode */
3069    
# Line 3054  if ((options & PCRE_EXTENDED) != 0) Line 3075  if ((options & PCRE_EXTENDED) != 0)
3075      if (*ptr == CHAR_NUMBER_SIGN)      if (*ptr == CHAR_NUMBER_SIGN)
3076        {        {
3077        ptr++;        ptr++;
3078        while (*ptr != 0)        while (*ptr != CHAR_NULL)
3079          {          {
3080          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
3081          ptr++;          ptr++;
# Line 3097  if ((options & PCRE_EXTENDED) != 0) Line 3118  if ((options & PCRE_EXTENDED) != 0)
3118      if (*ptr == CHAR_NUMBER_SIGN)      if (*ptr == CHAR_NUMBER_SIGN)
3119        {        {
3120        ptr++;        ptr++;
3121        while (*ptr != 0)        while (*ptr != CHAR_NULL)
3122          {          {
3123          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
3124          ptr++;          ptr++;
# Line 3118  if (*ptr == CHAR_ASTERISK || *ptr == CHA Line 3139  if (*ptr == CHAR_ASTERISK || *ptr == CHA
3139    
3140  /* If the previous item is a character, get its value. */  /* If the previous item is a character, get its value. */
3141    
3142  if (op_code == OP_CHAR || op_code == OP_CHARI ||  if (op_code == OP_CHAR || op_code == OP_CHARI ||
3143      op_code == OP_NOT || op_code == OP_NOTI)      op_code == OP_NOT || op_code == OP_NOTI)
   //if (escape == 0) switch(op_code)  
3144    {    {
3145  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3146    GETCHARTEST(c, previous);    GETCHARTEST(c, previous);
# Line 3136  if (escape == 0) Line 3156  if (escape == 0)
3156    {    {
3157    /* 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
3158    case, which maps to the special PT_CLIST property. Check this first. */    case, which maps to the special PT_CLIST property. Check this first. */
3159    
3160  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3161    if (utf && (unsigned int)c != NOTACHAR && (options & PCRE_CASELESS) != 0)    if (utf && c != NOTACHAR && (options & PCRE_CASELESS) != 0)
3162      {      {
3163      int ocs = UCD_CASESET(next);      unsigned int ocs = UCD_CASESET(next);
3164      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);
3165      }      }
3166  #endif  #endif
# Line 3152  if (escape == 0) Line 3172  if (escape == 0)
3172    
3173      /* 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
3174      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
3175      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,
3176      because multi-other-case characters are dealt with above. */      because multi-other-case characters are dealt with above. */
3177    
3178      case OP_CHARI:      case OP_CHARI:
# Line 3160  if (escape == 0) Line 3180  if (escape == 0)
3180  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3181      if (utf)      if (utf)
3182        {        {
3183        unsigned int othercase;        pcre_uint32 othercase;
3184        if (next < 128) othercase = cd->fcc[next]; else        if (next < 128) othercase = cd->fcc[next]; else
3185  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3186        othercase = UCD_OTHERCASE((unsigned int)next);        othercase = UCD_OTHERCASE(next);
3187  #else  #else
3188        othercase = NOTACHAR;        othercase = NOTACHAR;
3189  #endif  #endif
3190        return (unsigned int)c != othercase;        return c != othercase;
3191        }        }
3192      else      else
3193  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
3194      return (c != TABLE_GET((unsigned int)next, cd->fcc, next));  /* Not UTF */      return (c != TABLE_GET(next, cd->fcc, next));  /* Not UTF */
3195    
3196      case OP_NOT:      case OP_NOT:
3197      return c == next;      return c == next;
3198    
3199      case OP_NOTI:      case OP_NOTI:
3200      if (c == next) return TRUE;      if (c == next) return TRUE;
3201  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3202      if (utf)      if (utf)
3203        {        {
3204        unsigned int othercase;        pcre_uint32 othercase;
3205        if (next < 128) othercase = cd->fcc[next]; else        if (next < 128) othercase = cd->fcc[next]; else
3206  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3207        othercase = UCD_OTHERCASE((unsigned int)next);        othercase = UCD_OTHERCASE(next);
3208  #else  #else
3209        othercase = NOTACHAR;        othercase = NOTACHAR;
3210  #endif  #endif
3211        return (unsigned int)c == othercase;        return c == othercase;
3212        }        }
3213      else      else
3214  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
3215      return (c == TABLE_GET((unsigned int)next, cd->fcc, next));  /* Not UTF */      return (c == TABLE_GET(next, cd->fcc, next));  /* Not UTF */
3216    
3217      /* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* set.      /* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* set.
3218      When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */      When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */
# Line 3219  if (escape == 0) Line 3239  if (escape == 0)
3239      case OP_NOT_HSPACE:      case OP_NOT_HSPACE:
3240      switch(next)      switch(next)
3241        {        {
3242        HSPACE_CASES:        HSPACE_CASES:
3243        return op_code == OP_NOT_HSPACE;        return op_code == OP_NOT_HSPACE;
3244    
3245        default:        default:
# Line 3231  if (escape == 0) Line 3251  if (escape == 0)
3251      case OP_NOT_VSPACE:      case OP_NOT_VSPACE:
3252      switch(next)      switch(next)
3253        {        {
3254        VSPACE_CASES:        VSPACE_CASES:
3255        return op_code == OP_NOT_VSPACE;        return op_code == OP_NOT_VSPACE;
3256    
3257        default:        default:
# Line 3285  switch(op_code) Line 3305  switch(op_code)
3305      case ESC_H:      case ESC_H:
3306      switch(c)      switch(c)
3307        {        {
3308        HSPACE_CASES:        HSPACE_CASES:
3309        return escape != ESC_h;        return escape != ESC_h;
3310    
3311        default:        default:
3312        return escape == ESC_h;        return escape == ESC_h;
3313        }        }
# Line 3296  switch(op_code) Line 3316  switch(op_code)
3316      case ESC_V:      case ESC_V:
3317      switch(c)      switch(c)
3318        {        {
3319        VSPACE_CASES:        VSPACE_CASES:
3320        return escape != ESC_v;        return escape != ESC_v;
3321    
3322        default:        default:
# Line 3326  switch(op_code) Line 3346  switch(op_code)
3346      case ESC_p:      case ESC_p:
3347      case ESC_P:      case ESC_P:
3348        {        {
3349        int ptype, pdata, errorcodeptr;        unsigned int ptype = 0, pdata = 0;
3350          int errorcodeptr;
3351        BOOL negated;        BOOL negated;
3352    
3353        ptr--;      /* Make ptr point at the p or P */        ptr--;      /* Make ptr point at the p or P */
3354        ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr);        if (!get_ucp(&ptr, &negated, &ptype, &pdata, &errorcodeptr))
3355        if (ptype < 0) return FALSE;          return FALSE;
3356        ptr++;      /* Point past the final curly ket */        ptr++;      /* Point past the final curly ket */
3357    
3358        /* 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 3407  switch(op_code) Line 3428  switch(op_code)
3428  *************************************************/  *************************************************/
3429    
3430  /* 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
3431  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
3432  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
3433  mutually recursive with the function immediately below.  mutually recursive with the function immediately below.
3434    
3435  Arguments:  Arguments:
3436    classbits     the bit map for characters < 256    classbits     the bit map for characters < 256
3437    uchardptr     points to the pointer for extra data    uchardptr     points to the pointer for extra data
3438    options       the options word    options       the options word
3439    cd            contains pointers to tables etc.    cd            contains pointers to tables etc.
3440    start         start of range character    start         start of range character
3441    end           end of range character    end           end of range character
3442    
3443  Returns:        the number of < 256 characters added  Returns:        the number of < 256 characters added
3444                  the pointer to extra data is updated                  the pointer to extra data is updated
3445  */  */
# Line 3430  add_to_class(pcre_uint8 *classbits, pcre Line 3451  add_to_class(pcre_uint8 *classbits, pcre
3451  pcre_uint32 c;  pcre_uint32 c;
3452  int n8 = 0;  int n8 = 0;
3453    
3454  /* If caseless matching is required, scan the range and process alternate  /* If caseless matching is required, scan the range and process alternate
3455  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
3456  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
3457  range. */  range. */
3458    
3459  if ((options & PCRE_CASELESS) != 0)  if ((options & PCRE_CASELESS) != 0)
3460    {    {
3461  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3462    if ((options & PCRE_UTF8) != 0)    if ((options & PCRE_UTF8) != 0)
3463      {      {
3464      int rc;      int rc;
3465      pcre_uint32 oc, od;      pcre_uint32 oc, od;
3466    
3467      options &= ~PCRE_CASELESS;   /* Remove for recursive calls */      options &= ~PCRE_CASELESS;   /* Remove for recursive calls */
3468      c = start;      c = start;
3469    
3470      while ((rc = get_othercase_range(&c, end, &oc, &od)) >= 0)      while ((rc = get_othercase_range(&c, end, &oc, &od)) >= 0)
3471        {        {
3472        /* Handle a single character that has more than one other case. */        /* Handle a single character that has more than one other case. */
3473    
3474        if (rc > 0) n8 += add_list_to_class(classbits, uchardptr, options, cd,        if (rc > 0) n8 += add_list_to_class(classbits, uchardptr, options, cd,
3475          PRIV(ucd_caseless_sets) + rc, oc);          PRIV(ucd_caseless_sets) + rc, oc);
3476    
3477        /* Do nothing if the other case range is within the original range. */        /* Do nothing if the other case range is within the original range. */
3478    
3479        else if (oc >= start && od <= end) continue;        else if (oc >= start && od <= end) continue;
3480    
3481        /* 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
3482        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
3483        range. Otherwise, use a recursive call to add the additional range. */        range. Otherwise, use a recursive call to add the additional range. */
3484    
3485        else if (oc < start && od >= start - 1) start = oc; /* Extend downwards */        else if (oc < start && od >= start - 1) start = oc; /* Extend downwards */
3486        else if (od > end && oc <= end + 1) end = od;       /* Extend upwards */        else if (od > end && oc <= end + 1) end = od;       /* Extend upwards */
3487        else n8 += add_to_class(classbits, uchardptr, options, cd, oc, od);        else n8 += add_to_class(classbits, uchardptr, options, cd, oc, od);
# Line 3470  if ((options & PCRE_CASELESS) != 0) Line 3491  if ((options & PCRE_CASELESS) != 0)
3491  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3492    
3493    /* Not UTF-mode, or no UCP */    /* Not UTF-mode, or no UCP */
3494    
3495    for (c = start; c <= end && c < 256; c++)    for (c = start; c <= end && c < 256; c++)
3496      {      {
3497      SETBIT(classbits, cd->fcc[c]);      SETBIT(classbits, cd->fcc[c]);
3498      n8++;      n8++;
3499      }      }
3500    }    }
3501    
3502  /* 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
3503  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
3504  in all cases. */  in all cases. */
# Line 3503  if (end < 0x100) Line 3524  if (end < 0x100)
3524    {    {
3525    for (c = start; c <= end; c++)    for (c = start; c <= end; c++)
3526      {      {
3527      n8++;      n8++;
3528      SETBIT(classbits, c);      SETBIT(classbits, c);
3529      }      }
3530    }    }
3531    
3532  else  else
3533    {    {
3534    pcre_uchar *uchardata = *uchardptr;    pcre_uchar *uchardata = *uchardptr;
3535    
3536  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3537    if ((options & PCRE_UTF8) != 0)  /* All UTFs use the same flag bit */    if ((options & PCRE_UTF8) != 0)  /* All UTFs use the same flag bit */
3538      {      {
3539      if (start < end)      if (start < end)
3540        {        {
3541        *uchardata++ = XCL_RANGE;        *uchardata++ = XCL_RANGE;
3542        uchardata += PRIV(ord2utf)(start, uchardata);        uchardata += PRIV(ord2utf)(start, uchardata);
3543        uchardata += PRIV(ord2utf)(end, uchardata);        uchardata += PRIV(ord2utf)(end, uchardata);
3544        }        }
3545      else if (start == end)      else if (start == end)
3546        {        {
3547        *uchardata++ = XCL_SINGLE;        *uchardata++ = XCL_SINGLE;
3548        uchardata += PRIV(ord2utf)(start, uchardata);        uchardata += PRIV(ord2utf)(start, uchardata);
3549        }        }
3550      }      }
3551    else    else
3552  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
3553    
3554    /* Without UTF support, character values are constrained by the bit length,    /* Without UTF support, character values are constrained by the bit length,
3555    and can only be > 256 for 16-bit and 32-bit libraries. */    and can only be > 256 for 16-bit and 32-bit libraries. */
3556    
3557  #ifdef COMPILE_PCRE8  #ifdef COMPILE_PCRE8
3558      {}      {}
3559  #else  #else
3560    if (start < end)    if (start < end)
3561      {      {
3562      *uchardata++ = XCL_RANGE;      *uchardata++ = XCL_RANGE;
# Line 3546  else Line 3567  else
3567      {      {
3568      *uchardata++ = XCL_SINGLE;      *uchardata++ = XCL_SINGLE;
3569      *uchardata++ = start;      *uchardata++ = start;
3570      }      }
3571  #endif  #endif
3572    
3573    *uchardptr = uchardata;   /* Updata extra data pointer */    *uchardptr = uchardata;   /* Updata extra data pointer */
3574    }    }
3575    
3576  return n8;    /* Number of 8-bit characters */  return n8;    /* Number of 8-bit characters */
3577  }  }
3578    
3579    
3580    
3581    
3582  /*************************************************  /*************************************************
3583  *        Add a list of characters to a class     *  *        Add a list of characters to a class     *
3584  *************************************************/  *************************************************/
3585    
3586  /* 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
3587  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
3588  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
3589  handled appropriately. This function is mutually recursive with the function  handled appropriately. This function is mutually recursive with the function
# Line 3572  Arguments: Line 3593  Arguments:
3593    classbits     the bit map for characters < 256    classbits     the bit map for characters < 256
3594    uchardptr     points to the pointer for extra data    uchardptr     points to the pointer for extra data
3595    options       the options word    options       the options word
3596    cd            contains pointers to tables etc.    cd            contains pointers to tables etc.
3597    p             points to row of 32-bit values, terminated by NOTACHAR    p             points to row of 32-bit values, terminated by NOTACHAR
3598    except        character to omit; this is used when adding lists of    except        character to omit; this is used when adding lists of
3599                    case-equivalent characters to avoid including the one we                    case-equivalent characters to avoid including the one we
3600                    already know about                    already know about
3601    
3602  Returns:        the number of < 256 characters added  Returns:        the number of < 256 characters added
3603                  the pointer to extra data is updated                  the pointer to extra data is updated
3604  */  */
# Line 3591  while (p[0] < NOTACHAR) Line 3612  while (p[0] < NOTACHAR)
3612    {    {
3613    int n = 0;    int n = 0;
3614    if (p[0] != except)    if (p[0] != except)
3615      {      {
3616      while(p[n+1] == p[0] + n + 1) n++;      while(p[n+1] == p[0] + n + 1) n++;
3617      n8 += add_to_class(classbits, uchardptr, options, cd, p[0], p[n]);      n8 += add_to_class(classbits, uchardptr, options, cd, p[0], p[n]);
3618      }      }
3619    p += n + 1;    p += n + 1;
3620    }    }
3621  return n8;  return n8;
3622  }  }
3623    
3624    
3625    
# Line 3613  Arguments: Line 3634  Arguments:
3634    classbits     the bit map for characters < 256    classbits     the bit map for characters < 256
3635    uchardptr     points to the pointer for extra data    uchardptr     points to the pointer for extra data
3636    options       the options word    options       the options word
3637    cd            contains pointers to tables etc.    cd            contains pointers to tables etc.
3638    p             points to row of 32-bit values, terminated by NOTACHAR    p             points to row of 32-bit values, terminated by NOTACHAR
3639    
3640  Returns:        the number of < 256 characters added  Returns:        the number of < 256 characters added
3641                  the pointer to extra data is updated                  the pointer to extra data is updated
3642  */  */
3643    
3644  static int  static int
3645  add_not_list_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr,  add_not_list_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr,
3646    int options, compile_data *cd, const pcre_uint32 *p)    int options, compile_data *cd, const pcre_uint32 *p)
3647  {  {
3648  BOOL utf = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
# Line 3633  while (p[0] < NOTACHAR) Line 3654  while (p[0] < NOTACHAR)
3654    while (p[1] == p[0] + 1) p++;    while (p[1] == p[0] + 1) p++;
3655    n8 += add_to_class(classbits, uchardptr, options, cd, p[0] + 1,    n8 += add_to_class(classbits, uchardptr, options, cd, p[0] + 1,
3656      (p[1] == NOTACHAR) ? (utf ? 0x10ffffu : 0xffffffffu) : p[1] - 1);      (p[1] == NOTACHAR) ? (utf ? 0x10ffffu : 0xffffffffu) : p[1] - 1);
3657    p++;    p++;
3658    }    }
3659  return n8;  return n8;
3660  }  }
3661    
3662    
3663    
# Line 3655  Arguments: Line 3676  Arguments:
3676    codeptr        points to the pointer to the current code point    codeptr        points to the pointer to the current code point
3677    ptrptr         points to the current pattern pointer    ptrptr         points to the current pattern pointer
3678    errorcodeptr   points to error code variable    errorcodeptr   points to error code variable
3679    firstcharptr   set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE)    firstcharptr    place to put the first required character
3680    reqcharptr     set to the last literal character required, else < 0    firstcharflagsptr place to put the first character flags, or a negative number
3681      reqcharptr     place to put the last required character
3682      reqcharflagsptr place to put the last required character flags, or a negative number
3683    bcptr          points to current branch chain    bcptr          points to current branch chain
3684    cond_depth     conditional nesting depth    cond_depth     conditional nesting depth
3685    cd             contains pointers to tables etc.    cd             contains pointers to tables etc.
# Line 3669  Returns:         TRUE on success Line 3692  Returns:         TRUE on success
3692    
3693  static BOOL  static BOOL
3694  compile_branch(int *optionsptr, pcre_uchar **codeptr,  compile_branch(int *optionsptr, pcre_uchar **codeptr,
3695    const pcre_uchar **ptrptr, int *errorcodeptr, pcre_int32 *firstcharptr,    const pcre_uchar **ptrptr, int *errorcodeptr,
3696    pcre_int32 *reqcharptr, branch_chain *bcptr, int cond_depth,    pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr,
3697      pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr,
3698      branch_chain *bcptr, int cond_depth,
3699    compile_data *cd, int *lengthptr)    compile_data *cd, int *lengthptr)
3700  {  {
3701  int repeat_type, op_type;  int repeat_type, op_type;
3702  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */
3703  int bravalue = 0;  int bravalue = 0;
3704  int greedy_default, greedy_non_default;  int greedy_default, greedy_non_default;
3705  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
3706  pcre_int32 zeroreqchar, zerofirstchar;  pcre_int32 firstcharflags, reqcharflags;
3707    pcre_uint32 zeroreqchar, zerofirstchar;
3708    pcre_int32 zeroreqcharflags, zerofirstcharflags;
3709  pcre_int32 req_caseopt, reqvary, tempreqvary;  pcre_int32 req_caseopt, reqvary, tempreqvary;
3710  int options = *optionsptr;               /* May change dynamically */  int options = *optionsptr;               /* May change dynamically */
3711  int after_manual_callout = 0;  int after_manual_callout = 0;
3712  int length_prevgroup = 0;  int length_prevgroup = 0;
3713  register int c;  register pcre_uint32 c;
3714  int escape;  int escape;
3715  register pcre_uchar *code = *codeptr;  register pcre_uchar *code = *codeptr;
3716  pcre_uchar *last_code = code;  pcre_uchar *last_code = code;
# Line 3706  dynamically as we process the pattern. * Line 3733  dynamically as we process the pattern. *
3733  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3734  /* PCRE_UTF[16|32] have the same value as PCRE_UTF8. */  /* PCRE_UTF[16|32] have the same value as PCRE_UTF8. */
3735  BOOL utf = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
3736    #ifndef COMPILE_PCRE32
3737  pcre_uchar utf_chars[6];  pcre_uchar utf_chars[6];
3738    #endif
3739  #else  #else
3740  BOOL utf = FALSE;  BOOL utf = FALSE;
3741  #endif  #endif
3742    
3743  /* Helper variables for OP_XCLASS opcode (for characters > 255). We define  /* Helper variables for OP_XCLASS opcode (for characters > 255). We define
3744  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,
3745  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
3746  alternative calls for the different cases. */  alternative calls for the different cases. */
3747    
3748  pcre_uchar *class_uchardata;  pcre_uchar *class_uchardata;
# Line 3741  to take the zero repeat into account. Th Line 3770  to take the zero repeat into account. Th
3770  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual
3771  item types that can be repeated set these backoff variables appropriately. */  item types that can be repeated set these backoff variables appropriately. */
3772    
3773  firstchar = reqchar = zerofirstchar = zeroreqchar = REQ_UNSET;  firstchar = reqchar = zerofirstchar = zeroreqchar = 0;
3774    firstcharflags = reqcharflags = zerofirstcharflags = zeroreqcharflags = REQ_UNSET;
3775    
3776  /* The variable req_caseopt contains either the REQ_CASELESS value  /* The variable req_caseopt contains either the REQ_CASELESS value
3777  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 3767  for (;; ptr++) Line 3797  for (;; ptr++)
3797    int recno;    int recno;
3798    int refsign;    int refsign;
3799    int skipbytes;    int skipbytes;
3800    int subreqchar;    pcre_uint32 subreqchar, subfirstchar;
3801    int subfirstchar;    pcre_int32 subreqcharflags, subfirstcharflags;
3802    int terminator;    int terminator;
3803    int mclength;    unsigned int mclength;
3804    int tempbracount;    unsigned int tempbracount;
3805    int ec; // FIXMEchpe pcre_uint32    pcre_uint32 ec;
3806    pcre_uchar mcbuffer[8];    pcre_uchar mcbuffer[8];
3807    
3808    /* Get next character in the pattern */    /* Get next character in the pattern */
# Line 3782  for (;; ptr++) Line 3812  for (;; ptr++)
3812    /* 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
3813    string. Nesting only happens one level deep. */    string. Nesting only happens one level deep. */
3814    
3815    if (c == 0 && nestptr != NULL)    if (c == CHAR_NULL && nestptr != NULL)
3816      {      {
3817      ptr = nestptr;      ptr = nestptr;
3818      nestptr = NULL;      nestptr = NULL;
# Line 3857  for (;; ptr++) Line 3887  for (;; ptr++)
3887    
3888    /* 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 */
3889    
3890    if (inescq && c != 0)    if (inescq && c != CHAR_NULL)
3891      {      {
3892      if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E)      if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E)
3893        {        {
# Line 3905  for (;; ptr++) Line 3935  for (;; ptr++)
3935      if (c == CHAR_NUMBER_SIGN)      if (c == CHAR_NUMBER_SIGN)
3936        {        {
3937        ptr++;        ptr++;
3938        while (*ptr != 0)        while (*ptr != CHAR_NULL)
3939          {          {
3940          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
3941          ptr++;          ptr++;
# Line 3913  for (;; ptr++) Line 3943  for (;; ptr++)
3943          if (utf) FORWARDCHAR(ptr);          if (utf) FORWARDCHAR(ptr);
3944  #endif  #endif
3945          }          }
3946        if (*ptr != 0) continue;        if (*ptr != CHAR_NULL) continue;
3947    
3948        /* Else fall through to handle end of string */        /* Else fall through to handle end of string */
3949        c = 0;        c = 0;
# Line 3935  for (;; ptr++) Line 3965  for (;; ptr++)
3965      case CHAR_VERTICAL_LINE:       /* or | or ) */      case CHAR_VERTICAL_LINE:       /* or | or ) */
3966      case CHAR_RIGHT_PARENTHESIS:      case CHAR_RIGHT_PARENTHESIS:
3967      *firstcharptr = firstchar;      *firstcharptr = firstchar;
3968        *firstcharflagsptr = firstcharflags;
3969      *reqcharptr = reqchar;      *reqcharptr = reqchar;
3970        *reqcharflagsptr = reqcharflags;
3971      *codeptr = code;      *codeptr = code;
3972      *ptrptr = ptr;      *ptrptr = ptr;
3973      if (lengthptr != NULL)      if (lengthptr != NULL)
# Line 3959  for (;; ptr++) Line 3991  for (;; ptr++)
3991      previous = NULL;      previous = NULL;
3992      if ((options & PCRE_MULTILINE) != 0)      if ((options & PCRE_MULTILINE) != 0)
3993        {        {
3994        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;        if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
3995        *code++ = OP_CIRCM;        *code++ = OP_CIRCM;
3996        }        }
3997      else *code++ = OP_CIRC;      else *code++ = OP_CIRC;
# Line 3974  for (;; ptr++) Line 4006  for (;; ptr++)
4006      repeats. The value of reqchar doesn't change either. */      repeats. The value of reqchar doesn't change either. */
4007    
4008      case CHAR_DOT:      case CHAR_DOT:
4009      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;      if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4010      zerofirstchar = firstchar;      zerofirstchar = firstchar;
4011        zerofirstcharflags = firstcharflags;
4012      zeroreqchar = reqchar;      zeroreqchar = reqchar;
4013        zeroreqcharflags = reqcharflags;
4014      previous = code;      previous = code;
4015      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;
4016      break;      break;
# Line 4050  for (;; ptr++) Line 4084  for (;; ptr++)
4084          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)
4085        {        {
4086        *code++ = negate_class? OP_ALLANY : OP_FAIL;        *code++ = negate_class? OP_ALLANY : OP_FAIL;
4087        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;        if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
4088        zerofirstchar = firstchar;        zerofirstchar = firstchar;
4089          zerofirstcharflags = firstcharflags;
4090        break;        break;
4091        }        }
4092    
# Line 4086  for (;; ptr++) Line 4121  for (;; ptr++)
4121      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
4122      loop, c contains the first byte of the character. */      loop, c contains the first byte of the character. */
4123    
4124      if (c != 0) do      if (c != CHAR_NULL) do
4125        {        {
4126        const pcre_uchar *oldptr;        const pcre_uchar *oldptr;
4127    
4128  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
4129        if (utf && HAS_EXTRALEN(c))        if (utf && HAS_EXTRALEN(c))
4130          {                           /* Braces are required because the */          {                           /* Braces are required because the */
4131          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */
# Line 4101  for (;; ptr++) Line 4136  for (;; ptr++)
4136        /* In the pre-compile phase, accumulate the length of any extra        /* In the pre-compile phase, accumulate the length of any extra
4137        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
4138        contain a zillion > 255 characters no longer overwrite the work space        contain a zillion > 255 characters no longer overwrite the work space
4139        (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,
4140        however. */        however. */
4141    
4142        if (lengthptr != NULL && class_uchardata > class_uchardata_base)        if (lengthptr != NULL && class_uchardata > class_uchardata_base)
# Line 4165  for (;; ptr++) Line 4200  for (;; ptr++)
4200          alpha. This relies on the fact that the class table starts with          alpha. This relies on the fact that the class table starts with
4201          alpha, lower, upper as the first 3 entries. */          alpha, lower, upper as the first 3 entries. */
4202    
4203          if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)          if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)
4204            posix_class = 0;            posix_class = 0;
4205    
4206          /* 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 4321  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 4349  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 4379  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 4405  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 4420  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 4453  for (;; ptr++) Line 4488  for (;; ptr++)
4488    
4489            /* \b is backspace; any other special means the '-' was literal. */            /* \b is backspace; any other special means the '-' was literal. */
4490    
4491            if (descape > 0)            if (descape != 0)
4492              {              {
4493              if (descape == ESC_b) d = CHAR_BS; else              if (descape == ESC_b) d = CHAR_BS; else
4494                {                {
# Line 4476  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 4511  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 4546  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 4566  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 4578  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 4607  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 4665  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 4704  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 4807  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 4864  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 5020  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 5189  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 5468  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 5564  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 5609  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 5678  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 5755  for (;; ptr++) Line 5782  for (;; ptr++)
5782            }            }
5783          else          else
5784            {            {
5785            terminator = 0;            terminator = CHAR_NULL;
5786            if (ptr[1] == CHAR_MINUS || ptr[1] == CHAR_PLUS) refsign = *(++ptr);            if (ptr[1] == CHAR_MINUS || ptr[1] == CHAR_PLUS) refsign = *(++ptr);
5787            }            }
5788    
# Line 5775  for (;; ptr++) Line 5802  for (;; ptr++)
5802          while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0)          while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0)
5803            {            {
5804            if (recno >= 0)            if (recno >= 0)
5805              recno = (IS_DIGIT(*ptr))? recno * 10 + *ptr - CHAR_0 : -1;              recno = (IS_DIGIT(*ptr))? recno * 10 + (int)(*ptr - CHAR_0) : -1;
5806            ptr++;            ptr++;
5807            }            }
5808          namelen = (int)(ptr - name);          namelen = (int)(ptr - name);
5809    
5810          if ((terminator > 0 && *ptr++ != terminator) ||          if ((terminator > 0 && *ptr++ != (pcre_uchar)terminator) ||
5811              *ptr++ != CHAR_RIGHT_PARENTHESIS)              *ptr++ != CHAR_RIGHT_PARENTHESIS)
5812            {            {
5813            ptr--;      /* Error offset */            ptr--;      /* Error offset */
# Line 5845  for (;; ptr++) Line 5872  for (;; ptr++)
5872            code[1+LINK_SIZE]++;            code[1+LINK_SIZE]++;
5873            }            }
5874    
5875          /* If terminator == 0 it means that the name followed directly after          /* If terminator == CHAR_NULL it means that the name followed directly
5876          the opening parenthesis [e.g. (?(abc)...] and in this case there are          after the opening parenthesis [e.g. (?(abc)...] and in this case there
5877          some further alternatives to try. For the cases where terminator != 0          are some further alternatives to try. For the cases where terminator !=
5878          [things like (?(<name>... or (?('name')... or (?(R&name)... ] we have          0 [things like (?(<name>... or (?('name')... or (?(R&name)... ] we have
5879          now checked all the possibilities, so give an error. */          now checked all the possibilities, so give an error. */
5880    
5881          else if (terminator != 0)          else if (terminator != CHAR_NULL)
5882            {            {
5883            *errorcodeptr = ERR15;            *errorcodeptr = ERR15;
5884            goto FAILED;            goto FAILED;
# Line 6020  for (;; ptr++) Line 6047  for (;; ptr++)
6047    
6048            if (lengthptr != NULL)            if (lengthptr != NULL)
6049              {              {
6050              if (*ptr != terminator)              if (*ptr != (pcre_uchar)terminator)
6051                {                {
6052                *errorcodeptr = ERR42;                *errorcodeptr = ERR42;
6053                goto FAILED;                goto FAILED;
# Line 6162  for (;; ptr++) Line 6189  for (;; ptr++)
6189              *errorcodeptr = ERR62;              *errorcodeptr = ERR62;
6190              goto FAILED;              goto FAILED;
6191              }              }
6192            if (*ptr != terminator)            if (*ptr != (pcre_uchar)terminator)
6193              {              {
6194              *errorcodeptr = ERR42;              *errorcodeptr = ERR42;
6195              goto FAILED;              goto FAILED;
# Line 6268  for (;; ptr++) Line 6295  for (;; ptr++)
6295            while(IS_DIGIT(*ptr))            while(IS_DIGIT(*ptr))
6296              recno = recno * 10 + *ptr++ - CHAR_0;              recno = recno * 10 + *ptr++ - CHAR_0;
6297    
6298            if (*ptr != terminator)            if (*ptr != (pcre_uchar)terminator)
6299              {              {
6300              *errorcodeptr = ERR29;              *errorcodeptr = ERR29;
6301              goto FAILED;              goto FAILED;
# Line 6372  for (;; ptr++) Line 6399  for (;; ptr++)
6399    
6400          /* Can't determine a first byte now */          /* Can't determine a first byte now */
6401    
6402          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
6403          continue;          continue;
6404    
6405    
# Line 6506  for (;; ptr++) Line 6533  for (;; ptr++)
6533           cond_depth +           cond_depth +
6534             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */
6535           &subfirstchar,                   /* For possible first char */           &subfirstchar,                   /* For possible first char */
6536             &subfirstcharflags,
6537           &subreqchar,                     /* For possible last char */           &subreqchar,                     /* For possible last char */
6538             &subreqcharflags,
6539           bcptr,                           /* Current branch chain */           bcptr,                           /* Current branch chain */
6540           cd,                              /* Tables block */           cd,                              /* Tables block */
6541           (lengthptr == NULL)? NULL :      /* Actual compile phase */           (lengthptr == NULL)? NULL :      /* Actual compile phase */
# Line 6567  for (;; ptr++) Line 6596  for (;; ptr++)
6596            *errorcodeptr = ERR27;            *errorcodeptr = ERR27;
6597            goto FAILED;            goto FAILED;
6598            }            }
6599          if (condcount == 1) subfirstchar = subreqchar = REQ_NONE;          if (condcount == 1) subfirstcharflags = subreqcharflags = REQ_NONE;
6600          }          }
6601        }        }
6602    
# Line 6616  for (;; ptr++) Line 6645  for (;; ptr++)
6645      back off. */      back off. */
6646    
6647      zeroreqchar = reqchar;      zeroreqchar = reqchar;
6648        zeroreqcharflags = reqcharflags;
6649      zerofirstchar = firstchar;      zerofirstchar = firstchar;
6650        zerofirstcharflags = firstcharflags;
6651      groupsetfirstchar = FALSE;      groupsetfirstchar = FALSE;
6652    
6653      if (bravalue >= OP_ONCE)      if (bravalue >= OP_ONCE)
# Line 6627  for (;; ptr++) Line 6658  for (;; ptr++)
6658        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
6659        repeat forces firstchar to "none". */        repeat forces firstchar to "none". */
6660    
6661        if (firstchar == REQ_UNSET)        if (firstcharflags == REQ_UNSET)
6662          {          {
6663          if (subfirstchar >= 0)          if (subfirstcharflags >= 0)
6664            {            {
6665            firstchar = subfirstchar;            firstchar = subfirstchar;
6666              firstcharflags = subfirstcharflags;
6667            groupsetfirstchar = TRUE;            groupsetfirstchar = TRUE;
6668            }            }
6669          else firstchar = REQ_NONE;          else firstcharflags = REQ_NONE;
6670          zerofirstchar = REQ_NONE;          zerofirstcharflags = REQ_NONE;
6671          }          }
6672    
6673        /* If firstchar was previously set, convert the subpattern's firstchar        /* If firstchar was previously set, convert the subpattern's firstchar
6674        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
6675        existence beforehand. */        existence beforehand. */
6676    
6677        else if (subfirstchar >= 0 && subreqchar < 0)        else if (subfirstcharflags >= 0 && subreqcharflags < 0)
6678          subreqchar = subfirstchar | tempreqvary;          {
6679            subreqchar = subfirstchar;
6680            subreqcharflags = subfirstcharflags | tempreqvary;
6681            }
6682    
6683        /* 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
6684        really the first byte - see above), set it. */        really the first byte - see above), set it. */
6685    
6686        if (subreqchar >= 0) reqchar = subreqchar;        if (subreqcharflags >= 0)
6687            {
6688            reqchar = subreqchar;
6689            reqcharflags = subreqcharflags;
6690            }
6691        }        }
6692    
6693      /* 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 6659  for (;; ptr++) Line 6698  for (;; ptr++)
6698      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
6699      firstchar, looking for an asserted first char. */      firstchar, looking for an asserted first char. */
6700    
6701      else if (bravalue == OP_ASSERT && subreqchar >= 0) reqchar = subreqchar;      else if (bravalue == OP_ASSERT && subreqcharflags >= 0)
6702          {
6703          reqchar = subreqchar;
6704          reqcharflags = subreqcharflags;
6705          }
6706      break;     /* End of processing '(' */      break;     /* End of processing '(' */
6707    
6708    
# Line 6667  for (;; ptr++) Line 6710  for (;; ptr++)
6710      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values
6711      are arranged to be the negation of the corresponding OP_values in the      are arranged to be the negation of the corresponding OP_values in the
6712      default case when PCRE_UCP is not set. For the back references, the values      default case when PCRE_UCP is not set. For the back references, the values
6713      are ESC_REF plus the reference number. Only back references and those types      are negative the reference number. Only back references and those types
6714      that consume a character may be repeated. We can test for values between      that consume a character may be repeated. We can test for values between
6715      ESC_b and ESC_Z for the latter; this may have to change if any new ones are      ESC_b and ESC_Z for the latter; this may have to change if any new ones are
6716      ever created. */      ever created. */
# Line 6695  for (;; ptr++) Line 6738  for (;; ptr++)
6738        /* For metasequences that actually match a character, we disable the        /* For metasequences that actually match a character, we disable the
6739        setting of a first character if it hasn't already been set. */        setting of a first character if it hasn't already been set. */
6740    
6741        if (firstchar == REQ_UNSET && escape > ESC_b && escape < ESC_Z)        if (firstcharflags == REQ_UNSET && escape > ESC_b && escape < ESC_Z)
6742          firstchar = REQ_NONE;          firstcharflags = REQ_NONE;
6743    
6744        /* 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. */
6745    
6746        zerofirstchar = firstchar;        zerofirstchar = firstchar;
6747          zerofirstcharflags = firstcharflags;
6748        zeroreqchar = reqchar;        zeroreqchar = reqchar;
6749          zeroreqcharflags = reqcharflags;
6750    
6751        /* \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'
6752        is a subroutine call by number (Oniguruma syntax). In fact, the value        is a subroutine call by number (Oniguruma syntax). In fact, the value
6753        ESC_g is returned only for these cases. So we don't need to check for <        ESC_g is returned only for these cases. So we don't need to check for <
6754        or ' if the value is ESC_g. For the Perl syntax \g{n} the value is        or ' if the value is ESC_g. For the Perl syntax \g{n} the value is
6755        ESC_REF+n, and for the Perl syntax \g{name} the result is ESC_k (as        -n, and for the Perl syntax \g{name} the result is ESC_k (as
6756        that is a synonym for a named back reference). */        that is a synonym for a named back reference). */
6757    
6758        if (escape == ESC_g)        if (escape == ESC_g)
# Line 6730  for (;; ptr++) Line 6775  for (;; ptr++)
6775          if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)          if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS)
6776            {            {
6777            BOOL is_a_number = TRUE;            BOOL is_a_number = TRUE;
6778            for (p = ptr + 1; *p != 0 && *p != terminator; p++)            for (p = ptr + 1; *p != CHAR_NULL && *p != (pcre_uchar)terminator; p++)
6779              {              {
6780              if (!MAX_255(*p)) { is_a_number = FALSE; break; }              if (!MAX_255(*p)) { is_a_number = FALSE; break; }
6781              if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;              if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE;
6782              if ((cd->ctypes[*p] & ctype_word) == 0) break;              if ((cd->ctypes[*p] & ctype_word) == 0) break;
6783              }              }
6784            if (*p != terminator)            if (*p != (pcre_uchar)terminator)
6785              {              {
6786              *errorcodeptr = ERR57;              *errorcodeptr = ERR57;
6787              break;              break;
# Line 6754  for (;; ptr++) Line 6799  for (;; ptr++)
6799    
6800          p = ptr + 2;          p = ptr + 2;
6801          while (IS_DIGIT(*p)) p++;          while (IS_DIGIT(*p)) p++;
6802          if (*p != terminator)          if (*p != (pcre_uchar)terminator)
6803            {            {
6804            *errorcodeptr = ERR57;            *errorcodeptr = ERR57;
6805            break;            break;
# Line 6785  for (;; ptr++) Line 6830  for (;; ptr++)
6830        not set to cope with cases like (?=(\w+))\1: which would otherwise set        not set to cope with cases like (?=(\w+))\1: which would otherwise set
6831        ':' later. */        ':' later. */
6832    
6833        if (escape >= ESC_REF)        if (escape < 0)
6834          {          {
6835          open_capitem *oc;          open_capitem *oc;
6836          recno = escape - ESC_REF;          recno = -escape;
6837    
6838          HANDLE_REFERENCE:    /* Come here from named backref handling */          HANDLE_REFERENCE:    /* Come here from named backref handling */
6839          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE;
6840          previous = code;          previous = code;
6841          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;
6842          PUT2INC(code, 0, recno);          PUT2INC(code, 0, recno);
# Line 6818  for (;; ptr++) Line 6863  for (;; ptr++)
6863        else if (escape == ESC_P || escape == ESC_p)        else if (escape == ESC_P || escape == ESC_p)
6864          {          {
6865          BOOL negated;          BOOL negated;
6866          int pdata;          unsigned int ptype = 0, pdata = 0;
6867          int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);          if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr))
6868          if (ptype < 0) goto FAILED;            goto FAILED;
6869          previous = code;          previous = code;
6870          *code++ = ((escape == ESC_p) != negated)? OP_PROP : OP_NOTPROP;          *code++ = ((escape == ESC_p) != negated)? OP_PROP : OP_NOTPROP;
6871          *code++ = ptype;          *code++ = ptype;
# Line 6894  for (;; ptr++) Line 6939  for (;; ptr++)
6939      mclength = 1;      mclength = 1;
6940      mcbuffer[0] = c;      mcbuffer[0] = c;
6941    
6942  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32  #ifdef SUPPORT_UTF
6943      if (utf && HAS_EXTRALEN(c))      if (utf && HAS_EXTRALEN(c))
6944        ACROSSCHAR(TRUE, ptr[1], mcbuffer[mclength++] = *(++ptr));        ACROSSCHAR(TRUE, ptr[1], mcbuffer[mclength++] = *(++ptr));
6945  #endif  #endif
# Line 6904  for (;; ptr++) Line 6949  for (;; ptr++)
6949    
6950      ONE_CHAR:      ONE_CHAR:
6951      previous = code;      previous = code;
6952    
6953      /* For caseless UTF-8 mode when UCP support is available, check whether      /* For caseless UTF-8 mode when UCP support is available, check whether
6954      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
6955      OP_PROP item instead of OP_CHARI. */      OP_PROP item instead of OP_CHARI. */
6956    
6957  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6958      if (utf && (options & PCRE_CASELESS) != 0)      if (utf && (options & PCRE_CASELESS) != 0)
6959        {        {
# Line 6918  for (;; ptr++) Line 6963  for (;; ptr++)
6963          *code++ = OP_PROP;          *code++ = OP_PROP;
6964          *code++ = PT_CLIST;          *code++ = PT_CLIST;
6965          *code++ = c;          *code++ = c;
6966          if (firstchar == REQ_UNSET) firstchar = zerofirstchar = REQ_NONE;          if (firstcharflags == REQ_UNSET) firstcharflags = zerofirstcharflags = REQ_NONE;
6967          break;          break;
6968          }          }
6969        }        }
6970  #endif  #endif
6971    
6972      /* Caseful matches, or not one of the multicase characters. */      /* Caseful matches, or not one of the multicase characters. */
6973    
6974      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR;      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR;
6975      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];
6976    
# Line 6939  for (;; ptr++) Line 6984  for (;; ptr++)
6984      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
6985      repeat. */      repeat. */
6986    
6987      if (firstchar == REQ_UNSET)      if (firstcharflags == REQ_UNSET)
6988        {        {
6989        zerofirstchar = REQ_NONE;        zerofirstcharflags = REQ_NONE;
6990        zeroreqchar = reqchar;        zeroreqchar = reqchar;
6991          zeroreqcharflags = reqcharflags;
6992    
6993        /* 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
6994        only if it is not to be matched caselessly. */        only if it is not to be matched caselessly. */
# Line 6950  for (;; ptr++) Line 6996  for (;; ptr++)
6996        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
6997          {          {
6998          firstchar = mcbuffer[0] | req_caseopt;          firstchar = mcbuffer[0] | req_caseopt;
6999          if (mclength != 1) reqchar = code[-1] | cd->req_varyopt;          firstchar = mcbuffer[0];
7000            firstcharflags = req_caseopt;
7001    
7002            if (mclength != 1)
7003              {
7004              reqchar = code[-1];
7005              reqcharflags = cd->req_varyopt;
7006              }
7007          }          }
7008        else firstchar = reqchar = REQ_NONE;        else firstcharflags = reqcharflags = REQ_NONE;
7009        }        }
7010    
7011      /* 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 6961  for (;; ptr++) Line 7014  for (;; ptr++)
7014      else      else
7015        {        {
7016        zerofirstchar = firstchar;        zerofirstchar = firstchar;
7017          zerofirstcharflags = firstcharflags;
7018        zeroreqchar = reqchar;        zeroreqchar = reqchar;
7019          zeroreqcharflags = reqcharflags;
7020        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
7021          reqchar = code[-1] | req_caseopt | cd->req_varyopt;          {
7022            reqchar = code[-1];
7023            reqcharflags = req_caseopt | cd->req_varyopt;
7024            }
7025        }        }
7026    
7027      break;            /* End of literal character handling */      break;            /* End of literal character handling */
# Line 6982  return FALSE; Line 7040  return FALSE;
7040    
7041    
7042    
   
7043  /*************************************************  /*************************************************
7044  *     Compile sequence of alternatives           *  *     Compile sequence of alternatives           *
7045  *************************************************/  *************************************************/
# Line 7003  Arguments: Line 7060  Arguments:
7060    reset_bracount TRUE to reset the count for each branch    reset_bracount TRUE to reset the count for each branch
7061    skipbytes      skip this many bytes at start (for brackets and OP_COND)    skipbytes      skip this many bytes at start (for brackets and OP_COND)
7062    cond_depth     depth of nesting for conditional subpatterns    cond_depth     depth of nesting for conditional subpatterns
7063    firstcharptr   place to put the first required character, or a negative number    firstcharptr    place to put the first required character
7064    reqcharptr     place to put the last required character, or a negative number    firstcharflagsptr place to put the first character flags, or a negative number
7065      reqcharptr     place to put the last required character
7066      reqcharflagsptr place to put the last required character flags, or a negative number
7067    bcptr          pointer to the chain of currently open branches    bcptr          pointer to the chain of currently open branches
7068    cd             points to the data block with tables pointers etc.    cd             points to the data block with tables pointers etc.
7069    lengthptr      NULL during the real compile phase    lengthptr      NULL during the real compile phase
# Line 7016  Returns:         TRUE on success Line 7075  Returns:         TRUE on success
7075  static BOOL  static BOOL
7076  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,
7077    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
7078    int cond_depth, pcre_int32 *firstcharptr, pcre_int32 *reqcharptr,    int cond_depth,
7079      pcre_uint32 *firstcharptr, pcre_int32 *firstcharflagsptr,
7080      pcre_uint32 *reqcharptr, pcre_int32 *reqcharflagsptr,
7081    branch_chain *bcptr, compile_data *cd, int *lengthptr)    branch_chain *bcptr, compile_data *cd, int *lengthptr)
7082  {  {
7083  const pcre_uchar *ptr = *ptrptr;  const pcre_uchar *ptr = *ptrptr;
# Line 7026  pcre_uchar *start_bracket = code; Line 7087  pcre_uchar *start_bracket = code;
7087  pcre_uchar *reverse_count = NULL;  pcre_uchar *reverse_count = NULL;
7088  open_capitem capitem;  open_capitem capitem;
7089  int capnumber = 0;  int capnumber = 0;
7090  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
7091  pcre_int32 branchfirstchar, branchreqchar;  pcre_int32 firstcharflags, reqcharflags;
7092    pcre_uint32 branchfirstchar, branchreqchar;
7093    pcre_int32 branchfirstcharflags, branchreqcharflags;
7094  int length;  int length;
7095  int orig_bracount;  unsigned int orig_bracount;
7096  int max_bracount;  unsigned int max_bracount;
7097  branch_chain bc;  branch_chain bc;
7098    
7099  bc.outer = bcptr;  bc.outer = bcptr;
7100  bc.current_branch = code;  bc.current_branch = code;
7101    
7102  firstchar = reqchar = REQ_UNSET;  firstchar = reqchar = 0;
7103    firstcharflags = reqcharflags = REQ_UNSET;
7104    
7105  /* 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
7106  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 7096  for (;;) Line 7160  for (;;)
7160    into the length. */    into the length. */
7161    
7162    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,
7163          &branchreqchar, &bc, cond_depth, cd,          &branchfirstcharflags, &branchreqchar, &branchreqcharflags, &bc,
7164          (lengthptr == NULL)? NULL : &length))          cond_depth, cd, (lengthptr == NULL)? NULL : &length))
7165      {      {
7166      *ptrptr = ptr;      *ptrptr = ptr;
7167      return FALSE;      return FALSE;
# Line 7118  for (;;) Line 7182  for (;;)
7182      if (*last_branch != OP_ALT)      if (*last_branch != OP_ALT)
7183        {        {
7184        firstchar = branchfirstchar;        firstchar = branchfirstchar;
7185          firstcharflags = branchfirstcharflags;
7186        reqchar = branchreqchar;        reqchar = branchreqchar;
7187          reqcharflags = branchreqcharflags;
7188        }        }
7189    
7190      /* 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 7132  for (;;) Line 7198  for (;;)
7198        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
7199        previously no reqchar, it takes on the value of the old firstchar. */        previously no reqchar, it takes on the value of the old firstchar. */
7200    
7201        if (firstchar >= 0 && firstchar != branchfirstchar)        if (firstcharflags >= 0 &&
7202              (firstcharflags != branchfirstcharflags || firstchar != branchfirstchar))
7203          {          {
7204          if (reqchar < 0) reqchar = firstchar;          if (reqcharflags < 0)
7205          firstchar = REQ_NONE;            {
7206              reqchar = firstchar;
7207              reqcharflags = firstcharflags;
7208              }
7209            firstcharflags = REQ_NONE;
7210          }          }
7211    
7212        /* 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
7213        branch becomes a reqchar if there isn't a branch reqchar. */        branch becomes a reqchar if there isn't a branch reqchar. */
7214    
7215        if (firstchar < 0 && branchfirstchar >= 0 && branchreqchar < 0)        if (firstcharflags < 0 && branchfirstcharflags >= 0 && branchreqcharflags < 0)
7216            branchreqchar = branchfirstchar;          {
7217            branchreqchar = branchfirstchar;
7218            branchreqcharflags = branchfirstcharflags;
7219            }
7220    
7221        /* Now ensure that the reqchars match */        /* Now ensure that the reqchars match */
7222    
7223        if ((reqchar & ~REQ_VARY) != (branchreqchar & ~REQ_VARY))        if (((reqcharflags & ~REQ_VARY) != (branchreqcharflags & ~REQ_VARY)) ||
7224          reqchar = REQ_NONE;            reqchar != branchreqchar)
7225        else reqchar |= branchreqchar;   /* To "or" REQ_VARY */          reqcharflags = REQ_NONE;
7226          else
7227            {
7228            reqchar = branchreqchar;
7229            reqcharflags |= branchreqcharflags; /* To "or" REQ_VARY */
7230            }
7231        }        }
7232    
7233      /* 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 7244  for (;;) Line 7323  for (;;)
7323      *codeptr = code;      *codeptr = code;
7324      *ptrptr = ptr;      *ptrptr = ptr;
7325      *firstcharptr = firstchar;      *firstcharptr = firstchar;
7326        *firstcharflagsptr = firstcharflags;
7327      *reqcharptr = reqchar;      *reqcharptr = reqchar;
7328        *reqcharflagsptr = reqcharflags;
7329      if (lengthptr != NULL)      if (lengthptr != NULL)
7330        {        {
7331        if (OFLOW_MAX - *lengthptr < length)        if (OFLOW_MAX - *lengthptr < length)
# Line 7479  do { Line 7560  do {
7560       {       {
7561       if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE;       if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE;
7562       }       }
7563    
7564     /* Atomic brackets */     /* Atomic brackets */
7565    
7566     else if (op == OP_ONCE || op == OP_ONCE_NC)     else if (op == OP_ONCE || op == OP_ONCE_NC)
# Line 7531  we return that char, otherwise -1. Line 7612  we return that char, otherwise -1.
7612    
7613  Arguments:  Arguments:
7614    code       points to start of expression (the bracket)    code       points to start of expression (the bracket)
7615      flags       points to the first char flags, or to REQ_NONE
7616    inassert   TRUE if in an assertion    inassert   TRUE if in an assertion
7617    
7618  Returns:     -1 or the fixed first char  Returns:     the fixed first char, or 0 with REQ_NONE in flags
7619  */  */
7620    
7621  static int  static pcre_uint32
7622  find_firstassertedchar(const pcre_uchar *code, BOOL inassert)  find_firstassertedchar(const pcre_uchar *code, pcre_int32 *flags,
7623      BOOL inassert)
7624  {  {
7625  register int c = -1;  register pcre_uint32 c = 0;
7626    int cflags = REQ_NONE;
7627    
7628    *flags = REQ_NONE;
7629  do {  do {
7630     int d;     pcre_uint32 d;
7631       int dflags;
7632     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||     int xl = (*code == OP_CBRA || *code == OP_SCBRA ||
7633               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0;               *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0;
7634     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,     const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl,
7635       TRUE);       TRUE);
7636     register int op = *scode;     register pcre_uchar op = *scode;
7637    
7638     switch(op)     switch(op)
7639       {       {
7640       default:       default:
7641       return -1;       return 0;
7642    
7643       case OP_BRA:       case OP_BRA:
7644       case OP_BRAPOS:       case OP_BRAPOS:
# Line 7563  do { Line 7650  do {
7650       case OP_ONCE:       case OP_ONCE:
7651       case OP_ONCE_NC:       case OP_ONCE_NC:
7652       case OP_COND:       case OP_COND:
7653       if ((d = find_firstassertedchar(scode, op == OP_ASSERT)) < 0)       d = find_firstassertedchar(scode, &dflags, op == OP_ASSERT);
7654         return -1;       if (dflags < 0)
7655       if (c < 0) c = d; else if (c != d) return -1;         return 0;
7656         if (cflags < 0) { c = d; cflags = dflags; } else if (c != d || cflags != dflags) return 0;
7657       break;       break;
7658    
7659       case OP_EXACT:       case OP_EXACT:
# Line 7576  do { Line 7664  do {
7664       case OP_PLUS:       case OP_PLUS:
7665       case OP_MINPLUS:       case OP_MINPLUS:
7666       case OP_POSPLUS:       case OP_POSPLUS:
7667       if (!inassert) return -1;       if (!inassert) return 0;
7668       if (c < 0) c = scode[1];       if (cflags < 0) { c = scode[1]; cflags = 0; }
7669         else if (c != scode[1]) return -1;         else if (c != scode[1]) return 0;
7670       break;       break;
7671    
7672       case OP_EXACTI:       case OP_EXACTI:
# Line 7589  do { Line 7677  do {
7677       case OP_PLUSI:       case OP_PLUSI:
7678       case OP_MINPLUSI:       case OP_MINPLUSI:
7679       case OP_POSPLUSI:       case OP_POSPLUSI:
7680       if (!inassert) return -1;       if (!inassert) return 0;
7681       if (c < 0) c = scode[1] | REQ_CASELESS;       if (cflags < 0) { c = scode[1]; cflags = REQ_CASELESS; }
7682         else if (c != scode[1]) return -1;         else if (c != scode[1]) return 0;
7683       break;       break;
7684       }       }
7685    
7686     code += GET(code, 1);     code += GET(code, 1);
7687     }     }
7688  while (*code == OP_ALT);  while (*code == OP_ALT);
7689    
7690    *flags = cflags;
7691  return c;  return c;
7692  }  }
7693    
# Line 7665  pcre32_compile2(PCRE_SPTR32 pattern, int Line 7755  pcre32_compile2(PCRE_SPTR32 pattern, int
7755  {  {
7756  REAL_PCRE *re;  REAL_PCRE *re;
7757  int length = 1;  /* For final END opcode */  int length = 1;  /* For final END opcode */
7758  pcre_int32 firstchar, reqchar;  pcre_uint32 firstchar, reqchar;
7759    pcre_int32 firstcharflags, reqcharflags;
7760  int newline;  int newline;
7761  int errorcode = 0;  int errorcode = 0;
7762  int skipatstart = 0;  int skipatstart = 0;
# Line 7738  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 7829  while (ptr[skipatstart] == CHAR_LEFT_PAR
7829    int newnl = 0;    int newnl = 0;
7830    int newbsr = 0;    int newbsr = 0;
7831    
7832    /* For completeness and backward compatibility, (*UTFn) is supported in the
7833    relevant libraries, but (*UTF) is generic and always supported. Note that
7834    PCRE_UTF8 == PCRE_UTF16 == PCRE_UTF32. */
7835    
7836  #ifdef COMPILE_PCRE8  #ifdef COMPILE_PCRE8
7837    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 5) == 0)    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF8_RIGHTPAR, 5) == 0)
7838      { skipatstart += 7; options |= PCRE_UTF8; continue; }      { skipatstart += 7; options |= PCRE_UTF8; continue; }
7839  #endif  #endif
7840  #ifdef COMPILE_PCRE16  #ifdef COMPILE_PCRE16
7841    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF16_RIGHTPAR, 6) == 0)
7842      { skipatstart += 8; options |= PCRE_UTF16; continue; }      { skipatstart += 8; options |= PCRE_UTF16; continue; }
7843  #endif  #endif
7844  #ifdef COMPILE_PCRE32  #ifdef COMPILE_PCRE32
7845    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF32_RIGHTPAR, 6) == 0)
7846      { skipatstart += 8; options |= PCRE_UTF32; continue; }      { skipatstart += 8; options |= PCRE_UTF32; continue; }
7847  #endif  #endif
7848    
7849      else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 4) == 0)
7850        { skipatstart += 6; options |= PCRE_UTF8; continue; }
7851    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)
7852      { skipatstart += 6; options |= PCRE_UCP; continue; }      { skipatstart += 6; options |= PCRE_UCP; continue; }
7853    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 7915  ptr += skipatstart; Line 8013  ptr += skipatstart;
8013  code = cworkspace;  code = cworkspace;
8014  *code = OP_BRA;  *code = OP_BRA;
8015  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,
8016    FALSE, 0, 0, &firstchar, &reqchar, NULL, cd, &length);    FALSE, 0, 0, &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL,
8017      cd, &length);
8018  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;
8019    
8020  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,
# Line 7993  ptr = (const pcre_uchar *)pattern + skip Line 8092  ptr = (const pcre_uchar *)pattern + skip
8092  code = (pcre_uchar *)codestart;  code = (pcre_uchar *)codestart;
8093  *code = OP_BRA;  *code = OP_BRA;
8094  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,
8095    &firstchar, &reqchar, NULL, cd, NULL);    &firstchar, &firstcharflags, &reqchar, &reqcharflags, NULL, cd, NULL);
8096  re->top_bracket = cd->bracount;  re->top_bracket = cd->bracount;
8097  re->top_backref = cd->top_backref;  re->top_backref = cd->top_backref;
8098  re->max_lookbehind = cd->max_lookbehind;  re->max_lookbehind = cd->max_lookbehind;
8099  re->flags = cd->external_flags | PCRE_MODE;  re->flags = cd->external_flags | PCRE_MODE;
8100    
8101  if (cd->had_accept) reqchar = REQ_NONE;   /* Must disable after (*ACCEPT) */  if (cd->had_accept)
8102      {
8103      reqchar = 0;              /* Must disable after (*ACCEPT) */
8104      reqcharflags = REQ_NONE;
8105      }
8106    
8107  /* 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. */
8108    
8109  if (errorcode == 0 && *ptr != 0) errorcode = ERR22;  if (errorcode == 0 && *ptr != CHAR_NULL) errorcode = ERR22;
8110    
8111  /* Fill in the terminating state and check for disastrous overflow, but  /* Fill in the terminating state and check for disastrous overflow, but
8112  if debugging, leave the test till after things are printed out. */  if debugging, leave the test till after things are printed out. */
# Line 8014  if debugging, leave the test till after Line 8117  if debugging, leave the test till after
8117  if (code - codestart > length) errorcode = ERR23;  if (code - codestart > length) errorcode = ERR23;
8118  #endif  #endif
8119    
8120    #ifdef SUPPORT_VALGRIND
8121    /* If the estimated length exceeds the really used length, mark the extra
8122    allocated memory as unadressable, so that any out-of-bound reads can be
8123    detected. */
8124    VALGRIND_MAKE_MEM_NOACCESS(code, (length - (code - codestart)) * sizeof(pcre_uchar));
8125    #endif
8126    
8127  /* Fill in any forward references that are required. There may be repeated  /* Fill in any forward references that are required. There may be repeated
8128  references; optimize for them, as searching a large regex takes time. */  references; optimize for them, as searching a large regex takes time. */
8129    
# Line 8049  if (errorcode == 0 && re->top_backref > Line 8159  if (errorcode == 0 && re->top_backref >
8159    
8160  /* If there were any lookbehind assertions that contained OP_RECURSE  /* If there were any lookbehind assertions that contained OP_RECURSE
8161  (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,
8162  because they may contain forward references. Actual recursions can't be fixed  because they may contain forward references. Actual recursions cannot be fixed
8163  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
8164  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
8165  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 8120  if ((re->options & PCRE_ANCHORED) == 0) Line 8230  if ((re->options & PCRE_ANCHORED) == 0)
8230    if (is_anchored(codestart, 0, cd, 0)) re->options |= PCRE_ANCHORED;    if (is_anchored(codestart, 0, cd, 0)) re->options |= PCRE_ANCHORED;
8231    else    else
8232      {      {
8233      if (firstchar < 0)      if (firstcharflags < 0)
8234        firstchar = find_firstassertedchar(codestart, FALSE);        firstchar = find_firstassertedchar(codestart, &firstcharflags, FALSE);
8235      if (firstchar >= 0)   /* Remove caseless flag for non-caseable chars */      if (firstcharflags >= 0)   /* Remove caseless flag for non-caseable chars */
8236        {        {
8237  #if defined COMPILE_PCRE8  #if defined COMPILE_PCRE8
8238        re->first_char = firstchar & 0xff;        re->first_char = firstchar & 0xff;
8239  #elif defined COMPILE_PCRE16  #elif defined COMPILE_PCRE16
8240        re->first_char = firstchar & 0xffff;        re->first_char = firstchar & 0xffff;
8241  #elif defined COMPILE_PCRE32  #elif defined COMPILE_PCRE32
8242        re->first_char = firstchar & ~REQ_MASK;        re->first_char = firstchar;
8243  #endif  #endif
8244        if ((firstchar & REQ_CASELESS) != 0)        if ((firstcharflags & REQ_CASELESS) != 0)
8245          {          {
8246  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
8247          /* We ignore non-ASCII first chars in 8 bit mode. */          /* We ignore non-ASCII first chars in 8 bit mode. */
# Line 8163  if ((re->options & PCRE_ANCHORED) == 0) Line 8273  if ((re->options & PCRE_ANCHORED) == 0)
8273  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
8274  bytes. */  bytes. */
8275    
8276  if (reqchar >= 0 &&  if (reqcharflags >= 0 &&
8277       ((re->options & PCRE_ANCHORED) == 0 || (reqchar & REQ_VARY) != 0))       ((re->options & PCRE_ANCHORED) == 0 || (reqcharflags & REQ_VARY) != 0))
8278    {    {
8279  #if defined COMPILE_PCRE8  #if defined COMPILE_PCRE8
8280    re->req_char = reqchar & 0xff;    re->req_char = reqchar & 0xff;
8281  #elif defined COMPILE_PCRE16  #elif defined COMPILE_PCRE16
8282    re->req_char = reqchar & 0xffff;    re->req_char = reqchar & 0xffff;
8283  #elif defined COMPILE_PCRE32  #elif defined COMPILE_PCRE32
8284    re->req_char = reqchar & ~REQ_MASK;    re->req_char = reqchar;
8285  #endif  #endif
8286    if ((reqchar & REQ_CASELESS) != 0)    if ((reqcharflags & REQ_CASELESS) != 0)
8287      {      {
8288  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)  #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
8289      /* We ignore non-ASCII first chars in 8 bit mode. */      /* We ignore non-ASCII first chars in 8 bit mode. */

Legend:
Removed from v.1059  
changed lines
  Added in v.1231

  ViewVC Help
Powered by ViewVC 1.1.5