/[pcre]/code/branches/pcre16/pcre_compile.c
ViewVC logotype

Diff of /code/branches/pcre16/pcre_compile.c

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

revision 769 by zherczeg, Sun Nov 27 17:40:33 2011 UTC revision 782 by zherczeg, Sat Dec 3 23:58:37 2011 UTC
# Line 97  overrun before it actually does run off Line 97  overrun before it actually does run off
97    
98  #define WORK_SIZE_CHECK (COMPILE_WORK_SIZE - 100)  #define WORK_SIZE_CHECK (COMPILE_WORK_SIZE - 100)
99    
100    /* Private flags added to firstchar and reqchar. */
101    
102    #define REQ_CASELESS   0x10000000l      /* Indicates caselessness */
103    #define REQ_VARY       0x20000000l      /* Reqchar followed non-literal item */
104    
105  /* Table for handling escaped characters in the range '0'-'z'. Positive returns  /* Table for handling escaped characters in the range '0'-'z'. Positive returns
106  are simple data values; negative values are for special things like \d and so  are simple data values; negative values are for special things like \d and so
# Line 466  static const char error_texts[] = Line 470  static const char error_texts[] =
470    "\\k is not followed by a braced, angle-bracketed, or quoted name\0"    "\\k is not followed by a braced, angle-bracketed, or quoted name\0"
471    /* 70 */    /* 70 */
472    "internal error: unknown opcode in find_fixedlength()\0"    "internal error: unknown opcode in find_fixedlength()\0"
473      "Not allowed UTF-8 / UTF-16 code point (>= 0xd800 && <= 0xdfff)\0"
474    ;    ;
475    
476  /* Table to identify digits and hex digits. This is used when compiling  /* Table to identify digits and hex digits. This is used when compiling
# Line 484  For convenience, we use the same bit def Line 489  For convenience, we use the same bit def
489    
490  Then we can use ctype_digit and ctype_xdigit in the code. */  Then we can use ctype_digit and ctype_xdigit in the code. */
491    
492    /* Using a simple comparison for decimal numbers rather than a memory read
493    is much faster, and the resulting code is simpler (the compiler turns it
494    into a subtraction and unsigned comparison). */
495    
496    #define IS_DIGIT(x) ((x) >= CHAR_0 && (x) <= CHAR_9)
497    
498  #ifndef EBCDIC  #ifndef EBCDIC
499    
500  /* This is the "normal" case, for ASCII systems, and EBCDIC systems running in  /* This is the "normal" case, for ASCII systems, and EBCDIC systems running in
501  UTF-8 mode. */  UTF-8 mode. */
502    
503  static const unsigned char digitab[] =  static const pcre_uint8 digitab[] =
504    {    {
505    0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7 */    0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7 */
506    0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   8- 15 */    0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   8- 15 */
# Line 528  static const unsigned char digitab[] = Line 539  static const unsigned char digitab[] =
539    
540  /* This is the "abnormal" case, for EBCDIC systems not running in UTF-8 mode. */  /* This is the "abnormal" case, for EBCDIC systems not running in UTF-8 mode. */
541    
542  static const unsigned char digitab[] =  static const pcre_uint8 digitab[] =
543    {    {
544    0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7  0 */    0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   0-  7  0 */
545    0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   8- 15    */    0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /*   8- 15    */
# Line 563  static const unsigned char digitab[] = Line 574  static const unsigned char digitab[] =
574    0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c, /*  0 - 7  F0 */    0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c, /*  0 - 7  F0 */
575    0x0c,0x0c,0x00,0x00,0x00,0x00,0x00,0x00};/*  8 -255    */    0x0c,0x0c,0x00,0x00,0x00,0x00,0x00,0x00};/*  8 -255    */
576    
577  static const unsigned char ebcdic_chartab[] = { /* chartable partial dup */  static const pcre_uint8 ebcdic_chartab[] = { /* chartable partial dup */
578    0x80,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /*   0-  7 */    0x80,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /*   0-  7 */
579    0x00,0x00,0x00,0x00,0x01,0x01,0x00,0x00, /*   8- 15 */    0x00,0x00,0x00,0x00,0x01,0x01,0x00,0x00, /*   8- 15 */
580    0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /*  16- 23 */    0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /*  16- 23 */
# Line 651  Returns:    TRUE or FALSE Line 662  Returns:    TRUE or FALSE
662  static BOOL  static BOOL
663  is_counted_repeat(const pcre_uchar *p)  is_counted_repeat(const pcre_uchar *p)
664  {  {
665  if ((digitab[*p++] & ctype_digit) == 0) return FALSE;  if (!IS_DIGIT(*p)) return FALSE;
666  while ((digitab[*p] & ctype_digit) != 0) p++;  p++;
667    while (IS_DIGIT(*p)) p++;
668  if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE;  if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE;
669    
670  if (*p++ != CHAR_COMMA) return FALSE;  if (*p++ != CHAR_COMMA) return FALSE;
671  if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE;  if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE;
672    
673  if ((digitab[*p++] & ctype_digit) == 0) return FALSE;  if (!IS_DIGIT(*p)) return FALSE;
674  while ((digitab[*p] & ctype_digit) != 0) p++;  p++;
675    while (IS_DIGIT(*p)) p++;
676    
677  return (*p == CHAR_RIGHT_CURLY_BRACKET);  return (*p == CHAR_RIGHT_CURLY_BRACKET);
678  }  }
# Line 694  static int Line 707  static int
707  check_escape(const pcre_uchar **ptrptr, int *errorcodeptr, int bracount,  check_escape(const pcre_uchar **ptrptr, int *errorcodeptr, int bracount,
708    int options, BOOL isclass)    int options, BOOL isclass)
709  {  {
710  BOOL utf8 = (options & PCRE_UTF8) != 0;  /* PCRE_UTF16 has the same value as PCRE_UTF8. */
711    BOOL utf = (options & PCRE_UTF8) != 0;
712  const pcre_uchar *ptr = *ptrptr + 1;  const pcre_uchar *ptr = *ptrptr + 1;
713  int c, i;  pcre_int32 c;
714    int i;
715    
716  GETCHARINCTEST(c, ptr);           /* Get character value, increment pointer */  GETCHARINCTEST(c, ptr);           /* Get character value, increment pointer */
717  ptr--;                            /* Set pointer back to the last byte */  ptr--;                            /* Set pointer back to the last byte */
# Line 710  in a table. A non-zero result is somethi Line 725  in a table. A non-zero result is somethi
725  Otherwise further processing may be required. */  Otherwise further processing may be required. */
726    
727  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
728  else if (c < CHAR_0 || c > CHAR_z) {}                     /* Not alphanumeric */  /* Not alphanumeric */
729    else if (c < CHAR_0 || c > CHAR_z) {}
730  else if ((i = escapes[c - CHAR_0]) != 0) c = i;  else if ((i = escapes[c - CHAR_0]) != 0) c = i;
731    
732  #else           /* EBCDIC coding */  #else           /* EBCDIC coding */
733  else if (c < 'a' || (ebcdic_chartab[c] & 0x0E) == 0) {}   /* Not alphanumeric */  /* Not alphanumeric */
734    else if (c < 'a' || (!MAX_255(c) || (ebcdic_chartab[c] & 0x0E) == 0)) {}
735  else if ((i = escapes[c - 0x48]) != 0)  c = i;  else if ((i = escapes[c - 0x48]) != 0)  c = i;
736  #endif  #endif
737    
# Line 740  else Line 757  else
757        {        {
758        /* In JavaScript, \u must be followed by four hexadecimal numbers.        /* In JavaScript, \u must be followed by four hexadecimal numbers.
759        Otherwise it is a lowercase u letter. */        Otherwise it is a lowercase u letter. */
760        if ((digitab[ptr[1]] & ctype_xdigit) != 0 && (digitab[ptr[2]] & ctype_xdigit) != 0        if (MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0
761             && (digitab[ptr[3]] & ctype_xdigit) != 0 && (digitab[ptr[4]] & ctype_xdigit) != 0)          && MAX_255(ptr[2]) && (digitab[ptr[2]] & ctype_xdigit) != 0
762            && MAX_255(ptr[3]) && (digitab[ptr[3]] & ctype_xdigit) != 0
763            && MAX_255(ptr[4]) && (digitab[ptr[4]] & ctype_xdigit) != 0)
764          {          {
765          c = 0;          c = 0;
766          for (i = 0; i < 4; ++i)          for (i = 0; i < 4; ++i)
# Line 797  else Line 816  else
816        {        {
817        const pcre_uchar *p;        const pcre_uchar *p;
818        for (p = ptr+2; *p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET; p++)        for (p = ptr+2; *p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET; p++)
819          if (*p != CHAR_MINUS && (digitab[*p] & ctype_digit) == 0) break;          if (*p != CHAR_MINUS && !IS_DIGIT(*p)) break;
820        if (*p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET)        if (*p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET)
821          {          {
822          c = -ESC_k;          c = -ESC_k;
# Line 815  else Line 834  else
834        }        }
835      else negated = FALSE;      else negated = FALSE;
836    
837        /* The integer range is limited by the machine's int representation. */
838      c = 0;      c = 0;
839      while ((digitab[ptr[1]] & ctype_digit) != 0)      while (IS_DIGIT(ptr[1]))
840          {
841          if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */
842            {
843            c = -1;
844            break;
845            }
846        c = c * 10 + *(++ptr) - CHAR_0;        c = c * 10 + *(++ptr) - CHAR_0;
847          }
848      if (c < 0)   /* Integer overflow */      if (((unsigned int)c) > INT_MAX) /* Integer overflow */
849        {        {
850          while (IS_DIGIT(ptr[1]))
851            ptr++;
852        *errorcodeptr = ERR61;        *errorcodeptr = ERR61;
853        break;        break;
854        }        }
# Line 868  else Line 896  else
896      if (!isclass)      if (!isclass)
897        {        {
898        oldptr = ptr;        oldptr = ptr;
899          /* The integer range is limited by the machine's int representation. */
900        c -= CHAR_0;        c -= CHAR_0;
901        while ((digitab[ptr[1]] & ctype_digit) != 0)        while (IS_DIGIT(ptr[1]))
902            {
903            if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */
904              {
905              c = -1;
906              break;
907              }
908          c = c * 10 + *(++ptr) - CHAR_0;          c = c * 10 + *(++ptr) - CHAR_0;
909        if (c < 0)    /* Integer overflow */          }
910          if (((unsigned int)c) > INT_MAX) /* Integer overflow */
911          {          {
912            while (IS_DIGIT(ptr[1]))
913              ptr++;
914          *errorcodeptr = ERR61;          *errorcodeptr = ERR61;
915          break;          break;
916          }          }
# Line 905  else Line 943  else
943      c -= CHAR_0;      c -= CHAR_0;
944      while(i++ < 2 && ptr[1] >= CHAR_0 && ptr[1] <= CHAR_7)      while(i++ < 2 && ptr[1] >= CHAR_0 && ptr[1] <= CHAR_7)
945          c = c * 8 + *(++ptr) - CHAR_0;          c = c * 8 + *(++ptr) - CHAR_0;
946      if (!utf8 && c > 255) *errorcodeptr = ERR51;      if (!utf && c > 0xff) *errorcodeptr = ERR51;
947      break;      break;
948    
949      /* \x is complicated. \x{ddd} is a character number which can be greater      /* \x is complicated. \x{ddd} is a character number which can be greater
950      than 0xff in utf8 mode, but only if the ddd are hex digits. If not, { is      than 0xff in utf or non-8bit mode, but only if the ddd are hex digits.
951      treated as a data character. */      If not, { is treated as a data character. */
952    
953      case CHAR_x:      case CHAR_x:
954      if ((options & PCRE_JAVASCRIPT_COMPAT) != 0)      if ((options & PCRE_JAVASCRIPT_COMPAT) != 0)
955        {        {
956        /* In JavaScript, \x must be followed by two hexadecimal numbers.        /* In JavaScript, \x must be followed by two hexadecimal numbers.
957        Otherwise it is a lowercase x letter. */        Otherwise it is a lowercase x letter. */
958        if ((digitab[ptr[1]] & ctype_xdigit) != 0 && (digitab[ptr[2]] & ctype_xdigit) != 0)        if (MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0
959            && MAX_255(ptr[2]) && (digitab[ptr[2]] & ctype_xdigit) != 0)
960          {          {
961          c = 0;          c = 0;
962          for (i = 0; i < 2; ++i)          for (i = 0; i < 2; ++i)
# Line 938  else Line 977  else
977      if (ptr[1] == CHAR_LEFT_CURLY_BRACKET)      if (ptr[1] == CHAR_LEFT_CURLY_BRACKET)
978        {        {
979        const pcre_uchar *pt = ptr + 2;        const pcre_uchar *pt = ptr + 2;
       int count = 0;  
980    
981        c = 0;        c = 0;
982        while ((digitab[*pt] & ctype_xdigit) != 0)        while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0)
983          {          {
984          register int cc = *pt++;          register int cc = *pt++;
985          if (c == 0 && cc == CHAR_0) continue;     /* Leading zeroes */          if (c == 0 && cc == CHAR_0) continue;     /* Leading zeroes */
         count++;  
986    
987  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
988          if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */          if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
# Line 954  else Line 991  else
991          if (cc >= CHAR_a && cc <= CHAR_z) cc += 64;  /* Convert to upper case */          if (cc >= CHAR_a && cc <= CHAR_z) cc += 64;  /* Convert to upper case */
992          c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));          c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));
993  #endif  #endif
994    
995    #ifdef COMPILE_PCRE8
996            if (c > (utf ? 0x10ffff : 0xff)) { c = -1; break; }
997    #else
998    #ifdef COMPILE_PCRE16
999            if (c > (utf ? 0x10ffff : 0xffff)) { c = -1; break; }
1000    #endif
1001    #endif
1002            }
1003    
1004          if (c < 0)
1005            {
1006            while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0) pt++;
1007            *errorcodeptr = ERR34;
1008          }          }
1009    
1010        if (*pt == CHAR_RIGHT_CURLY_BRACKET)        if (*pt == CHAR_RIGHT_CURLY_BRACKET)
1011          {          {
1012          if (c < 0 || count > (utf8? 8 : 2)) *errorcodeptr = ERR34;          if (utf && c >= 0xd800 && c <= 0xdfff) *errorcodeptr = ERR71;
1013          ptr = pt;          ptr = pt;
1014          break;          break;
1015          }          }
# Line 970  else Line 1021  else
1021      /* Read just a single-byte hex-defined char */      /* Read just a single-byte hex-defined char */
1022    
1023      c = 0;      c = 0;
1024      while (i++ < 2 && (digitab[ptr[1]] & ctype_xdigit) != 0)      while (i++ < 2 && MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0)
1025        {        {
1026        int cc;                                  /* Some compilers don't like */        int cc;                                  /* Some compilers don't like */
1027        cc = *(++ptr);                           /* ++ in initializers */        cc = *(++ptr);                           /* ++ in initializers */
# Line 1169  int max = -1; Line 1220  int max = -1;
1220  /* 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
1221  an integer overflow. */  an integer overflow. */
1222    
1223  while ((digitab[*p] & ctype_digit) != 0) min = min * 10 + *p++ - CHAR_0;  while (IS_DIGIT(*p)) min = min * 10 + *p++ - CHAR_0;
1224  if (min < 0 || min > 65535)  if (min < 0 || min > 65535)
1225    {    {
1226    *errorcodeptr = ERR5;    *errorcodeptr = ERR5;
# Line 1184  if (*p == CHAR_RIGHT_CURLY_BRACKET) max Line 1235  if (*p == CHAR_RIGHT_CURLY_BRACKET) max
1235    if (*(++p) != CHAR_RIGHT_CURLY_BRACKET)    if (*(++p) != CHAR_RIGHT_CURLY_BRACKET)
1236      {      {
1237      max = 0;      max = 0;
1238      while((digitab[*p] & ctype_digit) != 0) max = max * 10 + *p++ - CHAR_0;      while(IS_DIGIT(*p)) max = max * 10 + *p++ - CHAR_0;
1239      if (max < 0 || max > 65535)      if (max < 0 || max > 65535)
1240        {        {
1241        *errorcodeptr = ERR5;        *errorcodeptr = ERR5;
# Line 1239  Arguments: Line 1290  Arguments:
1290    name         name to seek, or NULL if seeking a numbered subpattern    name         name to seek, or NULL if seeking a numbered subpattern
1291    lorn         name length, or subpattern number if name is NULL    lorn         name length, or subpattern number if name is NULL
1292    xmode        TRUE if we are in /x mode    xmode        TRUE if we are in /x mode
1293    utf8         TRUE if we are in UTF-8 mode    utf          TRUE if we are in UTF-8 / UTF-16 mode
1294    count        pointer to the current capturing subpattern number (updated)    count        pointer to the current capturing subpattern number (updated)
1295    
1296  Returns:       the number of the named subpattern, or -1 if not found  Returns:       the number of the named subpattern, or -1 if not found
# Line 1247  Returns:       the number of the named s Line 1298  Returns:       the number of the named s
1298    
1299  static int  static int
1300  find_parens_sub(pcre_uchar **ptrptr, compile_data *cd, const pcre_uchar *name, int lorn,  find_parens_sub(pcre_uchar **ptrptr, compile_data *cd, const pcre_uchar *name, int lorn,
1301    BOOL xmode, BOOL utf8, int *count)    BOOL xmode, BOOL utf, int *count)
1302  {  {
1303  pcre_uchar *ptr = *ptrptr;  pcre_uchar *ptr = *ptrptr;
1304  int start_count = *count;  int start_count = *count;
# Line 1415  for (; ptr < cd->end_pattern; ptr++) Line 1466  for (; ptr < cd->end_pattern; ptr++)
1466        {        {
1467        if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }        if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
1468        ptr++;        ptr++;
1469  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1470        if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;        if (utf) FORWARDCHAR(ptr);
1471  #endif  #endif
1472        }        }
1473      if (*ptr == 0) goto FAIL_EXIT;      if (*ptr == 0) goto FAIL_EXIT;
# Line 1427  for (; ptr < cd->end_pattern; ptr++) Line 1478  for (; ptr < cd->end_pattern; ptr++)
1478    
1479    if (*ptr == CHAR_LEFT_PARENTHESIS)    if (*ptr == CHAR_LEFT_PARENTHESIS)
1480      {      {
1481      int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf8, count);      int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, count);
1482      if (rc > 0) return rc;      if (rc > 0) return rc;
1483      if (*ptr == 0) goto FAIL_EXIT;      if (*ptr == 0) goto FAIL_EXIT;
1484      }      }
# Line 1473  Arguments: Line 1524  Arguments:
1524    name         name to seek, or NULL if seeking a numbered subpattern    name         name to seek, or NULL if seeking a numbered subpattern
1525    lorn         name length, or subpattern number if name is NULL    lorn         name length, or subpattern number if name is NULL
1526    xmode        TRUE if we are in /x mode    xmode        TRUE if we are in /x mode
1527    utf8         TRUE if we are in UTF-8 mode    utf          TRUE if we are in UTF-8 / UTF-16 mode
1528    
1529  Returns:       the number of the found subpattern, or -1 if not found  Returns:       the number of the found subpattern, or -1 if not found
1530  */  */
1531    
1532  static int  static int
1533  find_parens(compile_data *cd, const pcre_uchar *name, int lorn, BOOL xmode,  find_parens(compile_data *cd, const pcre_uchar *name, int lorn, BOOL xmode,
1534    BOOL utf8)    BOOL utf)
1535  {  {
1536  pcre_uchar *ptr = (pcre_uchar *)cd->start_pattern;  pcre_uchar *ptr = (pcre_uchar *)cd->start_pattern;
1537  int count = 0;  int count = 0;
# Line 1493  matching closing parens. That is why we Line 1544  matching closing parens. That is why we
1544    
1545  for (;;)  for (;;)
1546    {    {
1547    rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf8, &count);    rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, &count);
1548    if (rc > 0 || *ptr++ == 0) break;    if (rc > 0 || *ptr++ == 0) break;
1549    }    }
1550    
# Line 1576  and doing the check at the end; a flag s Line 1627  and doing the check at the end; a flag s
1627    
1628  Arguments:  Arguments:
1629    code     points to the start of the pattern (the bracket)    code     points to the start of the pattern (the bracket)
1630    utf8     TRUE in UTF-8 mode    utf      TRUE in UTF-8 / UTF-16 mode
1631    atend    TRUE if called when the pattern is complete    atend    TRUE if called when the pattern is complete
1632    cd       the "compile data" structure    cd       the "compile data" structure
1633    
# Line 1588  Returns:   the fixed length, Line 1639  Returns:   the fixed length,
1639  */  */
1640    
1641  static int  static int
1642  find_fixedlength(pcre_uchar *code, BOOL utf8, BOOL atend, compile_data *cd)  find_fixedlength(pcre_uchar *code, BOOL utf, BOOL atend, compile_data *cd)
1643  {  {
1644  int length = -1;  int length = -1;
1645    
# Line 1615  for (;;) Line 1666  for (;;)
1666      case OP_ONCE:      case OP_ONCE:
1667      case OP_ONCE_NC:      case OP_ONCE_NC:
1668      case OP_COND:      case OP_COND:
1669      d = find_fixedlength(cc + ((op == OP_CBRA)? IMM2_SIZE : 0), utf8, atend, cd);      d = find_fixedlength(cc + ((op == OP_CBRA)? IMM2_SIZE : 0), utf, atend, cd);
1670      if (d < 0) return d;      if (d < 0) return d;
1671      branchlength += d;      branchlength += d;
1672      do cc += GET(cc, 1); while (*cc == OP_ALT);      do cc += GET(cc, 1); while (*cc == OP_ALT);
# Line 1649  for (;;) Line 1700  for (;;)
1700      cs = ce = (pcre_uchar *)cd->start_code + GET(cc, 1);  /* Start subpattern */      cs = ce = (pcre_uchar *)cd->start_code + GET(cc, 1);  /* Start subpattern */
1701      do ce += GET(ce, 1); while (*ce == OP_ALT);           /* End subpattern */      do ce += GET(ce, 1); while (*ce == OP_ALT);           /* End subpattern */
1702      if (cc > cs && cc < ce) return -1;                    /* Recursion */      if (cc > cs && cc < ce) return -1;                    /* Recursion */
1703      d = find_fixedlength(cs + 2, utf8, atend, cd);      d = find_fixedlength(cs + 2, utf, atend, cd);
1704      if (d < 0) return d;      if (d < 0) return d;
1705      branchlength += d;      branchlength += d;
1706      cc += 1 + LINK_SIZE;      cc += 1 + LINK_SIZE;
# Line 1708  for (;;) Line 1759  for (;;)
1759      case OP_NOTI:      case OP_NOTI:
1760      branchlength++;      branchlength++;
1761      cc += 2;      cc += 2;
1762  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1763      if (utf8 && cc[-1] >= 0xc0) cc += PRIV(utf8_table4)[cc[-1] & 0x3f];      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
1764  #endif  #endif
1765      break;      break;
1766    
# Line 1722  for (;;) Line 1773  for (;;)
1773      case OP_NOTEXACTI:      case OP_NOTEXACTI:
1774      branchlength += GET2(cc,1);      branchlength += GET2(cc,1);
1775      cc += 2 + IMM2_SIZE;      cc += 2 + IMM2_SIZE;
1776  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1777      if (utf8 && cc[-1] >= 0xc0) cc += PRIV(utf8_table4)[cc[-1] & 0x3f];      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
1778  #endif  #endif
1779      break;      break;
1780    
# Line 1764  for (;;) Line 1815  for (;;)
1815    
1816      /* Check a class for variable quantification */      /* Check a class for variable quantification */
1817    
1818  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || defined COMPILE_PCRE16
1819      case OP_XCLASS:      case OP_XCLASS:
1820      cc += GET(cc, 1) - 33;      cc += GET(cc, 1) - PRIV(OP_lengths)[OP_CLASS];
1821      /* Fall through */      /* Fall through */
1822  #endif  #endif
1823    
1824      case OP_CLASS:      case OP_CLASS:
1825      case OP_NCLASS:      case OP_NCLASS:
1826      cc += 33;      cc += PRIV(OP_lengths)[OP_CLASS];
1827    
1828      switch (*cc)      switch (*cc)
1829        {        {
# Line 1903  length. Line 1954  length.
1954    
1955  Arguments:  Arguments:
1956    code        points to start of expression    code        points to start of expression
1957    utf8        TRUE in UTF-8 mode    utf         TRUE in UTF-8 / UTF-16 mode
1958    number      the required bracket number or negative to find a lookbehind    number      the required bracket number or negative to find a lookbehind
1959    
1960  Returns:      pointer to the opcode for the bracket, or NULL if not found  Returns:      pointer to the opcode for the bracket, or NULL if not found
1961  */  */
1962    
1963  const pcre_uchar *  const pcre_uchar *
1964  PRIV(find_bracket)(const pcre_uchar *code, BOOL utf8, int number)  PRIV(find_bracket)(const pcre_uchar *code, BOOL utf, int number)
1965  {  {
1966  for (;;)  for (;;)
1967    {    {
# Line 1990  for (;;) Line 2041  for (;;)
2041    a multi-byte character. The length in the table is a minimum, so we have to    a multi-byte character. The length in the table is a minimum, so we have to
2042    arrange to skip the extra bytes. */    arrange to skip the extra bytes. */
2043    
2044  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2045      if (utf8) switch(c)      if (utf) switch(c)
2046        {        {
2047        case OP_CHAR:        case OP_CHAR:
2048        case OP_CHARI:        case OP_CHARI:
# Line 2021  for (;;) Line 2072  for (;;)
2072        case OP_MINQUERYI:        case OP_MINQUERYI:
2073        case OP_POSQUERY:        case OP_POSQUERY:
2074        case OP_POSQUERYI:        case OP_POSQUERYI:
2075        if (code[-1] >= 0xc0) code += PRIV(utf8_table4)[code[-1] & 0x3f];        if (HAS_EXTRALEN(code[-1])) code += GET_EXTRALEN(code[-1]);
2076        break;        break;
2077        }        }
2078  #else  #else
2079      (void)(utf8);  /* Keep compiler happy by referencing function argument */      (void)(utf);  /* Keep compiler happy by referencing function argument */
2080  #endif  #endif
2081      }      }
2082    }    }
# Line 2042  instance of OP_RECURSE. Line 2093  instance of OP_RECURSE.
2093    
2094  Arguments:  Arguments:
2095    code        points to start of expression    code        points to start of expression
2096    utf8        TRUE in UTF-8 mode    utf         TRUE in UTF-8 / UTF-16 mode
2097    
2098  Returns:      pointer to the opcode for OP_RECURSE, or NULL if not found  Returns:      pointer to the opcode for OP_RECURSE, or NULL if not found
2099  */  */
2100    
2101  static const pcre_uchar *  static const pcre_uchar *
2102  find_recurse(const pcre_uchar *code, BOOL utf8)  find_recurse(const pcre_uchar *code, BOOL utf)
2103  {  {
2104  for (;;)  for (;;)
2105    {    {
# Line 2110  for (;;) Line 2161  for (;;)
2161      by a multi-byte character. The length in the table is a minimum, so we have      by a multi-byte character. The length in the table is a minimum, so we have
2162      to arrange to skip the extra bytes. */      to arrange to skip the extra bytes. */
2163    
2164  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2165      if (utf8) switch(c)      if (utf) switch(c)
2166        {        {
2167        case OP_CHAR:        case OP_CHAR:
2168        case OP_CHARI:        case OP_CHARI:
# Line 2141  for (;;) Line 2192  for (;;)
2192        case OP_MINQUERYI:        case OP_MINQUERYI:
2193        case OP_POSQUERY:        case OP_POSQUERY:
2194        case OP_POSQUERYI:        case OP_POSQUERYI:
2195        if (code[-1] >= 0xc0) code += PRIV(utf8_table4)[code[-1] & 0x3f];        if (HAS_EXTRALEN(code[-1])) code += GET_EXTRALEN(code[-1]);
2196        break;        break;
2197        }        }
2198  #else  #else
2199      (void)(utf8);  /* Keep compiler happy by referencing function argument */      (void)(utf);  /* Keep compiler happy by referencing function argument */
2200  #endif  #endif
2201      }      }
2202    }    }
# Line 2168  bracket whose current branch will alread Line 2219  bracket whose current branch will alread
2219  Arguments:  Arguments:
2220    code        points to start of search    code        points to start of search
2221    endcode     points to where to stop    endcode     points to where to stop
2222    utf8        TRUE if in UTF8 mode    utf         TRUE if in UTF-8 / UTF-16 mode
2223    cd          contains pointers to tables etc.    cd          contains pointers to tables etc.
2224    
2225  Returns:      TRUE if what is matched could be empty  Returns:      TRUE if what is matched could be empty
# Line 2176  Returns:      TRUE if what is matched co Line 2227  Returns:      TRUE if what is matched co
2227    
2228  static BOOL  static BOOL
2229  could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode,  could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode,
2230    BOOL utf8, compile_data *cd)    BOOL utf, compile_data *cd)
2231  {  {
2232  register int c;  register int c;
2233  for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE);  for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE);
# Line 2224  for (code = first_significant_code(code Line 2275  for (code = first_significant_code(code
2275    
2276      do      do
2277        {        {
2278        if (could_be_empty_branch(scode, endcode, utf8, cd))        if (could_be_empty_branch(scode, endcode, utf, cd))
2279          {          {
2280          empty_branch = TRUE;          empty_branch = TRUE;
2281          break;          break;
# Line 2280  for (code = first_significant_code(code Line 2331  for (code = first_significant_code(code
2331        empty_branch = FALSE;        empty_branch = FALSE;
2332        do        do
2333          {          {
2334          if (!empty_branch && could_be_empty_branch(code, endcode, utf8, cd))          if (!empty_branch && could_be_empty_branch(code, endcode, utf, cd))
2335            empty_branch = TRUE;            empty_branch = TRUE;
2336          code += GET(code, 1);          code += GET(code, 1);
2337          }          }
# Line 2310  for (code = first_significant_code(code Line 2361  for (code = first_significant_code(code
2361    
2362      case OP_CLASS:      case OP_CLASS:
2363      case OP_NCLASS:      case OP_NCLASS:
2364      ccode = code + 33;      ccode = code + PRIV(OP_lengths)[OP_CLASS];
2365    
2366  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2367      CHECK_CLASS_REPEAT:      CHECK_CLASS_REPEAT:
# Line 2401  for (code = first_significant_code(code Line 2452  for (code = first_significant_code(code
2452      /* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO,      /* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO,
2453      MINUPTO, and POSUPTO may be followed by a multibyte character */      MINUPTO, and POSUPTO may be followed by a multibyte character */
2454    
2455  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2456      case OP_STAR:      case OP_STAR:
2457      case OP_STARI:      case OP_STARI:
2458      case OP_MINSTAR:      case OP_MINSTAR:
# Line 2414  for (code = first_significant_code(code Line 2465  for (code = first_significant_code(code
2465      case OP_MINQUERYI:      case OP_MINQUERYI:
2466      case OP_POSQUERY:      case OP_POSQUERY:
2467      case OP_POSQUERYI:      case OP_POSQUERYI:
2468      if (utf8 && code[1] >= 0xc0) code += PRIV(utf8_table4)[code[1] & 0x3f];      if (utf && HAS_EXTRALEN(code[1])) code += GET_EXTRALEN(code[1]);
2469      break;      break;
2470    
2471      case OP_UPTO:      case OP_UPTO:
# Line 2423  for (code = first_significant_code(code Line 2474  for (code = first_significant_code(code
2474      case OP_MINUPTOI:      case OP_MINUPTOI:
2475      case OP_POSUPTO:      case OP_POSUPTO:
2476      case OP_POSUPTOI:      case OP_POSUPTOI:
2477      if (utf8 && code[1 + IMM2_SIZE] >= 0xc0) code += PRIV(utf8_table4)[code[1 + IMM2_SIZE] & 0x3f];      if (utf && HAS_EXTRALEN(code[1 + IMM2_SIZE])) code += GET_EXTRALEN(code[1 + IMM2_SIZE]);
2478      break;      break;
2479  #endif  #endif
2480    
# Line 2467  Arguments: Line 2518  Arguments:
2518    code        points to start of the recursion    code        points to start of the recursion
2519    endcode     points to where to stop (current RECURSE item)    endcode     points to where to stop (current RECURSE item)
2520    bcptr       points to the chain of current (unclosed) branch starts    bcptr       points to the chain of current (unclosed) branch starts
2521    utf8        TRUE if in UTF-8 mode    utf         TRUE if in UTF-8 / UTF-16 mode
2522    cd          pointers to tables etc    cd          pointers to tables etc
2523    
2524  Returns:      TRUE if what is matched could be empty  Returns:      TRUE if what is matched could be empty
# Line 2475  Returns:      TRUE if what is matched co Line 2526  Returns:      TRUE if what is matched co
2526    
2527  static BOOL  static BOOL
2528  could_be_empty(const pcre_uchar *code, const pcre_uchar *endcode,  could_be_empty(const pcre_uchar *code, const pcre_uchar *endcode,
2529    branch_chain *bcptr, BOOL utf8, compile_data *cd)    branch_chain *bcptr, BOOL utf, compile_data *cd)
2530  {  {
2531  while (bcptr != NULL && bcptr->current_branch >= code)  while (bcptr != NULL && bcptr->current_branch >= code)
2532    {    {
2533    if (!could_be_empty_branch(bcptr->current_branch, endcode, utf8, cd))    if (!could_be_empty_branch(bcptr->current_branch, endcode, utf, cd))
2534      return FALSE;      return FALSE;
2535    bcptr = bcptr->outer;    bcptr = bcptr->outer;
2536    }    }
# Line 2614  value in the reference (which is a group Line 2665  value in the reference (which is a group
2665  Arguments:  Arguments:
2666    group      points to the start of the group    group      points to the start of the group
2667    adjust     the amount by which the group is to be moved    adjust     the amount by which the group is to be moved
2668    utf8       TRUE in UTF-8 mode    utf        TRUE in UTF-8 / UTF-16 mode
2669    cd         contains pointers to tables etc.    cd         contains pointers to tables etc.
2670    save_hwm   the hwm forward reference pointer at the start of the group    save_hwm   the hwm forward reference pointer at the start of the group
2671    
# Line 2622  Returns:     nothing Line 2673  Returns:     nothing
2673  */  */
2674    
2675  static void  static void
2676  adjust_recurse(pcre_uchar *group, int adjust, BOOL utf8, compile_data *cd,  adjust_recurse(pcre_uchar *group, int adjust, BOOL utf, compile_data *cd,
2677    pcre_uchar *save_hwm)    pcre_uchar *save_hwm)
2678  {  {
2679  pcre_uchar *ptr = group;  pcre_uchar *ptr = group;
2680    
2681  while ((ptr = (pcre_uchar *)find_recurse(ptr, utf8)) != NULL)  while ((ptr = (pcre_uchar *)find_recurse(ptr, utf)) != NULL)
2682    {    {
2683    int offset;    int offset;
2684    pcre_uchar *hc;    pcre_uchar *hc;
# Line 2833  sense to automatically possessify the re Line 2884  sense to automatically possessify the re
2884    
2885  Arguments:  Arguments:
2886    previous      pointer to the repeated opcode    previous      pointer to the repeated opcode
2887    utf8          TRUE in UTF-8 mode    utf           TRUE in UTF-8 / UTF-16 mode
2888    ptr           next character in pattern    ptr           next character in pattern
2889    options       options bits    options       options bits
2890    cd            contains pointers to tables etc.    cd            contains pointers to tables etc.
# Line 2842  Returns:        TRUE if possessifying is Line 2893  Returns:        TRUE if possessifying is
2893  */  */
2894    
2895  static BOOL  static BOOL
2896  check_auto_possessive(const pcre_uchar *previous, BOOL utf8,  check_auto_possessive(const pcre_uchar *previous, BOOL utf,
2897    const pcre_uchar *ptr, int options, compile_data *cd)    const pcre_uchar *ptr, int options, compile_data *cd)
2898  {  {
2899  int c, next;  int c, next;
# Line 2862  if ((options & PCRE_EXTENDED) != 0) Line 2913  if ((options & PCRE_EXTENDED) != 0)
2913          {          {
2914          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
2915          ptr++;          ptr++;
2916  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2917          if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;          if (utf) FORWARDCHAR(ptr);
2918  #endif  #endif
2919          }          }
2920        }        }
# Line 2885  if (*ptr == CHAR_BACKSLASH) Line 2936  if (*ptr == CHAR_BACKSLASH)
2936  else if ((cd->ctypes[*ptr] & ctype_meta) == 0)  else if ((cd->ctypes[*ptr] & ctype_meta) == 0)
2937    {    {
2938  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2939    if (utf8) { GETCHARINC(next, ptr); } else    if (utf) { GETCHARINC(next, ptr); } else
2940  #endif  #endif
2941    next = *ptr++;    next = *ptr++;
2942    }    }
# Line 2906  if ((options & PCRE_EXTENDED) != 0) Line 2957  if ((options & PCRE_EXTENDED) != 0)
2957          {          {
2958          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
2959          ptr++;          ptr++;
2960  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2961          if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;          if (utf) FORWARDCHAR(ptr);
2962  #endif  #endif
2963          }          }
2964        }        }
# Line 2946  if (next >= 0) switch(op_code) Line 2997  if (next >= 0) switch(op_code)
2997  #endif  #endif
2998    if (c == next) return FALSE;    if (c == next) return FALSE;
2999  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3000    if (utf8)    if (utf)
3001      {      {
3002      unsigned int othercase;      unsigned int othercase;
3003      if (next < 128) othercase = cd->fcc[next]; else      if (next < 128) othercase = cd->fcc[next]; else
# Line 2971  if (next >= 0) switch(op_code) Line 3022  if (next >= 0) switch(op_code)
3022    case OP_NOTI:    case OP_NOTI:
3023    if ((c = *previous) == next) return TRUE;    if ((c = *previous) == next) return TRUE;
3024  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3025    if (utf8)    if (utf)
3026      {      {
3027      unsigned int othercase;      unsigned int othercase;
3028      if (next < 128) othercase = cd->fcc[next]; else      if (next < 128) othercase = cd->fcc[next]; else
# Line 3258  Arguments: Line 3309  Arguments:
3309    codeptr        points to the pointer to the current code point    codeptr        points to the pointer to the current code point
3310    ptrptr         points to the current pattern pointer    ptrptr         points to the current pattern pointer
3311    errorcodeptr   points to error code variable    errorcodeptr   points to error code variable
3312    firstbyteptr   set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE)    firstcharptr   set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE)
3313    reqbyteptr     set to the last literal character required, else < 0    reqcharptr     set to the last literal character required, else < 0
3314    bcptr          points to current branch chain    bcptr          points to current branch chain
3315    cond_depth     conditional nesting depth    cond_depth     conditional nesting depth
3316    cd             contains pointers to tables etc.    cd             contains pointers to tables etc.
# Line 3272  Returns:         TRUE on success Line 3323  Returns:         TRUE on success
3323    
3324  static BOOL  static BOOL
3325  compile_branch(int *optionsptr, pcre_uchar **codeptr,  compile_branch(int *optionsptr, pcre_uchar **codeptr,
3326    const pcre_uchar **ptrptr, int *errorcodeptr, int *firstbyteptr,    const pcre_uchar **ptrptr, int *errorcodeptr, pcre_int32 *firstcharptr,
3327    int *reqbyteptr, branch_chain *bcptr, int cond_depth, compile_data *cd,    pcre_int32 *reqcharptr, branch_chain *bcptr, int cond_depth,
3328    int *lengthptr)    compile_data *cd, int *lengthptr)
3329  {  {
3330  int repeat_type, op_type;  int repeat_type, op_type;
3331  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */
3332  int bravalue = 0;  int bravalue = 0;
3333  int greedy_default, greedy_non_default;  int greedy_default, greedy_non_default;
3334  int firstbyte, reqbyte;  pcre_int32 firstchar, reqchar;
3335  int zeroreqbyte, zerofirstbyte;  pcre_int32 zeroreqchar, zerofirstchar;
3336  int req_caseopt, reqvary, tempreqvary;  pcre_int32 req_caseopt, reqvary, tempreqvary;
3337  int options = *optionsptr;               /* May change dynamically */  int options = *optionsptr;               /* May change dynamically */
3338  int after_manual_callout = 0;  int after_manual_callout = 0;
3339  int length_prevgroup = 0;  int length_prevgroup = 0;
# Line 3292  pcre_uchar *last_code = code; Line 3343  pcre_uchar *last_code = code;
3343  pcre_uchar *orig_code = code;  pcre_uchar *orig_code = code;
3344  pcre_uchar *tempcode;  pcre_uchar *tempcode;
3345  BOOL inescq = FALSE;  BOOL inescq = FALSE;
3346  BOOL groupsetfirstbyte = FALSE;  BOOL groupsetfirstchar = FALSE;
3347  const pcre_uchar *ptr = *ptrptr;  const pcre_uchar *ptr = *ptrptr;
3348  const pcre_uchar *tempptr;  const pcre_uchar *tempptr;
3349  const pcre_uchar *nestptr = NULL;  const pcre_uchar *nestptr = NULL;
3350  pcre_uchar *previous = NULL;  pcre_uchar *previous = NULL;
3351  pcre_uchar *previous_callout = NULL;  pcre_uchar *previous_callout = NULL;
3352  pcre_uchar *save_hwm = NULL;  pcre_uchar *save_hwm = NULL;
3353  pcre_uchar classbits[32];  pcre_uint8 classbits[32];
3354    
3355  /* We can fish out the UTF-8 setting once and for all into a BOOL, but we  /* We can fish out the UTF-8 setting once and for all into a BOOL, but we
3356  must not do this for other options (e.g. PCRE_EXTENDED) because they may change  must not do this for other options (e.g. PCRE_EXTENDED) because they may change
3357  dynamically as we process the pattern. */  dynamically as we process the pattern. */
3358    
3359  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3360  BOOL class_utf8;  /* PCRE_UTF16 has the same value as PCRE_UTF8. */
3361  BOOL utf8 = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
3362  pcre_uint8 *class_utf8data;  pcre_uchar utf_chars[6];
 pcre_uint8 *class_utf8data_base;  
 pcre_uint8 utf8_char[6];  
3363  #else  #else
3364  BOOL utf8 = FALSE;  BOOL utf = FALSE;
3365    #endif
3366    
3367    /* Helper variables for OP_XCLASS opcode (for characters > 255). */
3368    
3369    #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3370    BOOL xclass;
3371    pcre_uchar *class_uchardata;
3372    pcre_uchar *class_uchardata_base;
3373  #endif  #endif
3374    
3375  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 3326  greedy_non_default = greedy_default ^ 1; Line 3383  greedy_non_default = greedy_default ^ 1;
3383    
3384  /* Initialize no first byte, no required byte. REQ_UNSET means "no char  /* Initialize no first byte, no required byte. REQ_UNSET means "no char
3385  matching encountered yet". It gets changed to REQ_NONE if we hit something that  matching encountered yet". It gets changed to REQ_NONE if we hit something that
3386  matches a non-fixed char first char; reqbyte just remains unset if we never  matches a non-fixed char first char; reqchar just remains unset if we never
3387  find one.  find one.
3388    
3389  When we hit a repeat whose minimum is zero, we may have to adjust these values  When we hit a repeat whose minimum is zero, we may have to adjust these values
3390  to take the zero repeat into account. This is implemented by setting them to  to take the zero repeat into account. This is implemented by setting them to
3391  zerofirstbyte and zeroreqbyte when such a repeat is encountered. The individual  zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual
3392  item types that can be repeated set these backoff variables appropriately. */  item types that can be repeated set these backoff variables appropriately. */
3393    
3394  firstbyte = reqbyte = zerofirstbyte = zeroreqbyte = REQ_UNSET;  firstchar = reqchar = zerofirstchar = zeroreqchar = REQ_UNSET;
3395    
3396  /* The variable req_caseopt contains either the REQ_CASELESS value or zero,  /* The variable req_caseopt contains either the REQ_CASELESS value
3397  according to the current setting of the caseless flag. REQ_CASELESS is a bit  or zero, according to the current setting of the caseless flag. The
3398  value > 255. It is added into the firstbyte or reqbyte variables to record the  REQ_CASELESS leaves the lower 28 bit empty. It is added into the
3399  case status of the value. This is used only for ASCII characters. */  firstchar or reqchar variables to record the case status of the
3400    value. This is used only for ASCII characters. */
3401    
3402  req_caseopt = ((options & PCRE_CASELESS) != 0)? REQ_CASELESS : 0;  req_caseopt = ((options & PCRE_CASELESS) != 0)? REQ_CASELESS:0;
3403    
3404  /* Switch on next character until the end of the branch */  /* Switch on next character until the end of the branch */
3405    
# Line 3359  for (;; ptr++) Line 3417  for (;; ptr++)
3417    int recno;    int recno;
3418    int refsign;    int refsign;
3419    int skipbytes;    int skipbytes;
3420    int subreqbyte;    int subreqchar;
3421    int subfirstbyte;    int subfirstchar;
3422    int terminator;    int terminator;
3423    int mclength;    int mclength;
3424    int tempbracount;    int tempbracount;
3425    pcre_uchar mcbuffer[8];    pcre_uchar mcbuffer[8];
3426    
3427    /* Get next byte in the pattern */    /* Get next character in the pattern */
3428    
3429    c = *ptr;    c = *ptr;
3430    
# Line 3411  for (;; ptr++) Line 3469  for (;; ptr++)
3469        }        }
3470    
3471      *lengthptr += (int)(code - last_code);      *lengthptr += (int)(code - last_code);
3472      DPRINTF(("length=%d added %d c=%c\n", *lengthptr, (int)(code - last_code),      DPRINTF(("length=%d added %d c=%c (0x%x)\n", *lengthptr,
3473        c));        (int)(code - last_code), c, c));
3474    
3475      /* If "previous" is set and it is not at the start of the work space, move      /* If "previous" is set and it is not at the start of the work space, move
3476      it back to there, in order to avoid filling up the work space. Otherwise,      it back to there, in order to avoid filling up the work space. Otherwise,
# Line 3498  for (;; ptr++) Line 3556  for (;; ptr++)
3556          {          {
3557          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
3558          ptr++;          ptr++;
3559  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3560          if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;          if (utf) FORWARDCHAR(ptr);
3561  #endif  #endif
3562          }          }
3563        if (*ptr != 0) continue;        if (*ptr != 0) continue;
# Line 3523  for (;; ptr++) Line 3581  for (;; ptr++)
3581      case 0:                        /* The branch terminates at string end */      case 0:                        /* The branch terminates at string end */
3582      case CHAR_VERTICAL_LINE:       /* or | or ) */      case CHAR_VERTICAL_LINE:       /* or | or ) */
3583      case CHAR_RIGHT_PARENTHESIS:      case CHAR_RIGHT_PARENTHESIS:
3584      *firstbyteptr = firstbyte;      *firstcharptr = firstchar;
3585      *reqbyteptr = reqbyte;      *reqcharptr = reqchar;
3586      *codeptr = code;      *codeptr = code;
3587      *ptrptr = ptr;      *ptrptr = ptr;
3588      if (lengthptr != NULL)      if (lengthptr != NULL)
# Line 3548  for (;; ptr++) Line 3606  for (;; ptr++)
3606      previous = NULL;      previous = NULL;
3607      if ((options & PCRE_MULTILINE) != 0)      if ((options & PCRE_MULTILINE) != 0)
3608        {        {
3609        if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
3610        *code++ = OP_CIRCM;        *code++ = OP_CIRCM;
3611        }        }
3612      else *code++ = OP_CIRC;      else *code++ = OP_CIRC;
# Line 3560  for (;; ptr++) Line 3618  for (;; ptr++)
3618      break;      break;
3619    
3620      /* There can never be a first char if '.' is first, whatever happens about      /* There can never be a first char if '.' is first, whatever happens about
3621      repeats. The value of reqbyte doesn't change either. */      repeats. The value of reqchar doesn't change either. */
3622    
3623      case CHAR_DOT:      case CHAR_DOT:
3624      if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
3625      zerofirstbyte = firstbyte;      zerofirstchar = firstchar;
3626      zeroreqbyte = reqbyte;      zeroreqchar = reqchar;
3627      previous = code;      previous = code;
3628      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;      *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY;
3629      break;      break;
# Line 3620  for (;; ptr++) Line 3678  for (;; ptr++)
3678          {          {
3679          if (ptr[1] == CHAR_E)          if (ptr[1] == CHAR_E)
3680            ptr++;            ptr++;
3681          else if (STRNCMP_UC_C8(ptr + 1,          else if (STRNCMP_UC_C8(ptr + 1, STR_Q STR_BACKSLASH STR_E, 3) == 0)
                           STR_Q STR_BACKSLASH STR_E, 3) == 0)  
3682            ptr += 3;            ptr += 3;
3683          else          else
3684            break;            break;
# Line 3640  for (;; ptr++) Line 3697  for (;; ptr++)
3697          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)          (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0)
3698        {        {
3699        *code++ = negate_class? OP_ALLANY : OP_FAIL;        *code++ = negate_class? OP_ALLANY : OP_FAIL;
3700        if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;        if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
3701        zerofirstbyte = firstbyte;        zerofirstchar = firstchar;
3702        break;        break;
3703        }        }
3704    
# Line 3665  for (;; ptr++) Line 3722  for (;; ptr++)
3722    
3723      memset(classbits, 0, 32 * sizeof(pcre_uint8));      memset(classbits, 0, 32 * sizeof(pcre_uint8));
3724    
3725  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3726      class_utf8 = FALSE;                       /* No chars >= 256 */      xclass = FALSE;                           /* No chars >= 256 */
3727      class_utf8data = code + LINK_SIZE + 2;    /* For UTF-8 items */      class_uchardata = code + LINK_SIZE + 2;   /* For UTF-8 items */
3728      class_utf8data_base = class_utf8data;     /* For resetting in pass 1 */      class_uchardata_base = class_uchardata;   /* For resetting in pass 1 */
3729  #endif  #endif
3730    
3731      /* Process characters until ] is reached. By writing this as a "do" it      /* Process characters until ] is reached. By writing this as a "do" it
# Line 3680  for (;; ptr++) Line 3737  for (;; ptr++)
3737        const pcre_uchar *oldptr;        const pcre_uchar *oldptr;
3738    
3739  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
3740        if (utf8 && c > 127)        if (utf && c > 127)
3741          {                           /* Braces are required because the */          {                           /* Braces are required because the */
3742          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */
3743          }          }
3744    #endif
3745    
3746        /* In the pre-compile phase, accumulate the length of any UTF-8 extra  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3747          /* In the pre-compile phase, accumulate the length of any extra
3748        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
3749        contain a zillion UTF-8 characters no longer overwrite the work space        contain a zillion > 255 characters no longer overwrite the work space
3750        (which is on the stack). */        (which is on the stack). */
3751    
3752        if (lengthptr != NULL)        if (lengthptr != NULL)
3753          {          {
3754          *lengthptr += class_utf8data - class_utf8data_base;          *lengthptr += class_uchardata - class_uchardata_base;
3755          class_utf8data = class_utf8data_base;          class_uchardata = class_uchardata_base;
3756          }          }
   
3757  #endif  #endif
3758    
3759        /* Inside \Q...\E everything is literal except \E */        /* Inside \Q...\E everything is literal except \E */
# Line 3896  for (;; ptr++) Line 3954  for (;; ptr++)
3954              SETBIT(classbits, 0x09); /* VT */              SETBIT(classbits, 0x09); /* VT */
3955              SETBIT(classbits, 0x20); /* SPACE */              SETBIT(classbits, 0x20); /* SPACE */
3956              SETBIT(classbits, 0xa0); /* NSBP */              SETBIT(classbits, 0xa0); /* NSBP */
3957  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3958              if (utf8)              if (utf)
3959                {                {
3960                class_utf8 = TRUE;                xclass = TRUE;
3961                *class_utf8data++ = XCL_SINGLE;                *class_uchardata++ = XCL_SINGLE;
3962                class_utf8data += PRIV(ord2utf8)(0x1680, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x1680, class_uchardata);
3963                *class_utf8data++ = XCL_SINGLE;                *class_uchardata++ = XCL_SINGLE;
3964                class_utf8data += PRIV(ord2utf8)(0x180e, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x180e, class_uchardata);
3965                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
3966                class_utf8data += PRIV(ord2utf8)(0x2000, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x2000, class_uchardata);
3967                class_utf8data += PRIV(ord2utf8)(0x200A, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x200A, class_uchardata);
3968                *class_utf8data++ = XCL_SINGLE;                *class_uchardata++ = XCL_SINGLE;
3969                class_utf8data += PRIV(ord2utf8)(0x202f, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x202f, class_uchardata);
3970                *class_utf8data++ = XCL_SINGLE;                *class_uchardata++ = XCL_SINGLE;
3971                class_utf8data += PRIV(ord2utf8)(0x205f, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x205f, class_uchardata);
3972                *class_utf8data++ = XCL_SINGLE;                *class_uchardata++ = XCL_SINGLE;
3973                class_utf8data += PRIV(ord2utf8)(0x3000, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x3000, class_uchardata);
3974                }                }
3975  #endif  #endif
3976              continue;              continue;
# Line 3931  for (;; ptr++) Line 3989  for (;; ptr++)
3989                classbits[c] |= x;                classbits[c] |= x;
3990                }                }
3991    
3992  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3993              if (utf8)              if (utf)
3994                {                {
3995                class_utf8 = TRUE;                xclass = TRUE;
3996                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
3997                class_utf8data += PRIV(ord2utf8)(0x0100, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x0100, class_uchardata);
3998                class_utf8data += PRIV(ord2utf8)(0x167f, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x167f, class_uchardata);
3999                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4000                class_utf8data += PRIV(ord2utf8)(0x1681, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x1681, class_uchardata);
4001                class_utf8data += PRIV(ord2utf8)(0x180d, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x180d, class_uchardata);
4002                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4003                class_utf8data += PRIV(ord2utf8)(0x180f, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x180f, class_uchardata);
4004                class_utf8data += PRIV(ord2utf8)(0x1fff, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x1fff, class_uchardata);
4005                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4006                class_utf8data += PRIV(ord2utf8)(0x200B, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x200B, class_uchardata);
4007                class_utf8data += PRIV(ord2utf8)(0x202e, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x202e, class_uchardata);
4008                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4009                class_utf8data += PRIV(ord2utf8)(0x2030, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x2030, class_uchardata);
4010                class_utf8data += PRIV(ord2utf8)(0x205e, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x205e, class_uchardata);
4011                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4012                class_utf8data += PRIV(ord2utf8)(0x2060, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x2060, class_uchardata);
4013                class_utf8data += PRIV(ord2utf8)(0x2fff, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x2fff, class_uchardata);
4014                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4015                class_utf8data += PRIV(ord2utf8)(0x3001, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x3001, class_uchardata);
4016                class_utf8data += PRIV(ord2utf8)(0x7fffffff, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata);
4017                }                }
4018  #endif  #endif
4019              continue;              continue;
# Line 3966  for (;; ptr++) Line 4024  for (;; ptr++)
4024              SETBIT(classbits, 0x0c); /* FF */              SETBIT(classbits, 0x0c); /* FF */
4025              SETBIT(classbits, 0x0d); /* CR */              SETBIT(classbits, 0x0d); /* CR */
4026              SETBIT(classbits, 0x85); /* NEL */              SETBIT(classbits, 0x85); /* NEL */
4027  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4028              if (utf8)              if (utf)
4029                {                {
4030                class_utf8 = TRUE;                xclass = TRUE;
4031                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4032                class_utf8data += PRIV(ord2utf8)(0x2028, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x2028, class_uchardata);
4033                class_utf8data += PRIV(ord2utf8)(0x2029, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x2029, class_uchardata);
4034                }                }
4035  #endif  #endif
4036              continue;              continue;
# Line 3994  for (;; ptr++) Line 4052  for (;; ptr++)
4052                classbits[c] |= x;                classbits[c] |= x;
4053                }                }
4054    
4055  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4056              if (utf8)              if (utf)
4057                {                {
4058                class_utf8 = TRUE;                xclass = TRUE;
4059                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4060                class_utf8data += PRIV(ord2utf8)(0x0100, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x0100, class_uchardata);
4061                class_utf8data += PRIV(ord2utf8)(0x2027, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x2027, class_uchardata);
4062                *class_utf8data++ = XCL_RANGE;                *class_uchardata++ = XCL_RANGE;
4063                class_utf8data += PRIV(ord2utf8)(0x2029, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x2029, class_uchardata);
4064                class_utf8data += PRIV(ord2utf8)(0x7fffffff, class_utf8data);                class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata);
4065                }                }
4066  #endif  #endif
4067              continue;              continue;
# Line 4016  for (;; ptr++) Line 4074  for (;; ptr++)
4074                int pdata;                int pdata;
4075                int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);                int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);
4076                if (ptype < 0) goto FAILED;                if (ptype < 0) goto FAILED;
4077                class_utf8 = TRUE;                xclass = TRUE;
4078                *class_utf8data++ = ((-c == ESC_p) != negated)?                *class_uchardata++ = ((-c == ESC_p) != negated)?
4079                  XCL_PROP : XCL_NOTPROP;                  XCL_PROP : XCL_NOTPROP;
4080                *class_utf8data++ = ptype;                *class_uchardata++ = ptype;
4081                *class_utf8data++ = pdata;                *class_uchardata++ = pdata;
4082                class_charcount -= 2;   /* Not a < 256 character */                class_charcount -= 2;   /* Not a < 256 character */
4083                continue;                continue;
4084                }                }
# Line 4042  for (;; ptr++) Line 4100  for (;; ptr++)
4100            }            }
4101    
4102          /* Fall through if we have a single character (c >= 0). This may be          /* Fall through if we have a single character (c >= 0). This may be
4103          greater than 256 in UTF-8 mode. */          greater than 256 mode. */
4104    
4105          }   /* End of backslash handling */          }   /* End of backslash handling */
4106    
# Line 4091  for (;; ptr++) Line 4149  for (;; ptr++)
4149            }            }
4150    
4151  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4152          if (utf8)          if (utf)
4153            {                           /* Braces are required because the */            {                           /* Braces are required because the */
4154            GETCHARLEN(d, ptr, ptr);    /* macro generates multiple statements */            GETCHARLEN(d, ptr, ptr);    /* macro generates multiple statements */
4155            }            }
# Line 4140  for (;; ptr++) Line 4198  for (;; ptr++)
4198          matching for characters > 127 is available only if UCP support is          matching for characters > 127 is available only if UCP support is
4199          available. */          available. */
4200    
4201  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4202          if (utf8 && (d > 255 || ((options & PCRE_CASELESS) != 0 && d > 127)))          if (utf && (d > 255 || ((options & PCRE_CASELESS) != 0 && d > 127)))
4203    #endif
4204    #ifndef COMPILE_PCRE8
4205            if (d > 255)
4206    #endif
4207    #if defined SUPPORT_UTF || defined COMPILE_PCRE16
4208            {            {
4209            class_utf8 = TRUE;            xclass = TRUE;
4210    
4211            /* With UCP support, we can find the other case equivalents of            /* With UCP support, we can find the other case equivalents of
4212            the relevant characters. There may be several ranges. Optimize how            the relevant characters. There may be several ranges. Optimize how
# Line 4176  for (;; ptr++) Line 4239  for (;; ptr++)
4239    
4240                if (occ == ocd)                if (occ == ocd)
4241                  {                  {
4242                  *class_utf8data++ = XCL_SINGLE;                  *class_uchardata++ = XCL_SINGLE;
4243                  }                  }
4244                else                else
4245                  {                  {
4246                  *class_utf8data++ = XCL_RANGE;                  *class_uchardata++ = XCL_RANGE;
4247                  class_utf8data += PRIV(ord2utf8)(occ, class_utf8data);                  class_uchardata += PRIV(ord2utf)(occ, class_uchardata);
4248                  }                  }
4249                class_utf8data += PRIV(ord2utf8)(ocd, class_utf8data);                class_uchardata += PRIV(ord2utf)(ocd, class_uchardata);
4250                }                }
4251              }              }
4252  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
# Line 4191  for (;; ptr++) Line 4254  for (;; ptr++)
4254            /* Now record the original range, possibly modified for UCP caseless            /* Now record the original range, possibly modified for UCP caseless
4255            overlapping ranges. */            overlapping ranges. */
4256    
4257            *class_utf8data++ = XCL_RANGE;            *class_uchardata++ = XCL_RANGE;
4258            class_utf8data += PRIV(ord2utf8)(c, class_utf8data);  #ifdef SUPPORT_UTF
4259            class_utf8data += PRIV(ord2utf8)(d, class_utf8data);            class_uchardata += PRIV(ord2utf)(c, class_uchardata);
4260              class_uchardata += PRIV(ord2utf)(d, class_uchardata);
4261    #else
4262              *class_uchardata++ = c;
4263              *class_uchardata++ = d;
4264    #endif
4265    
4266            /* With UCP support, we are done. Without UCP support, there is no            /* With UCP support, we are done. Without UCP support, there is no
4267            caseless matching for UTF-8 characters > 127; we can use the bit map            caseless matching for UTF characters > 127; we can use the bit map
4268            for the smaller ones. */            for the smaller ones. As for 16 bit characters without UTF, we
4269              can still use  */
4270    
4271  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
4272            continue;    /* With next character in the class */            continue;    /* With next character in the class */
4273  #else  #else
4274    #ifdef SUPPORT_UTF
4275            if ((options & PCRE_CASELESS) == 0 || c > 127) continue;            if ((options & PCRE_CASELESS) == 0 || c > 127) continue;
   
4276            /* Adjust upper limit and fall through to set up the map */            /* Adjust upper limit and fall through to set up the map */
   
4277            d = 127;            d = 127;
4278    #else
4279              if (c > 255) continue;
4280              /* Adjust upper limit and fall through to set up the map */
4281              d = 255;
4282    #endif  /* SUPPORT_UTF */
4283  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
4284            }            }
4285  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF8 || COMPILE_PCRE16 */
4286    
4287          /* We use the bit map for all cases when not in UTF-8 mode; else          /* We use the bit map for 8 bit mode, or when the characters fall
4288          ranges that lie entirely within 0-127 when there is UCP support; else          partially or entirely to [0-255] ([0-127] for UCP) ranges. */
         for partial ranges without UCP support. */  
4289    
4290          class_charcount += d - c + 1;          class_charcount += d - c + 1;
4291          class_lastchar = d;          class_lastchar = d;
# Line 4242  for (;; ptr++) Line 4313  for (;; ptr++)
4313    
4314        /* Handle a character that cannot go in the bit map */        /* Handle a character that cannot go in the bit map */
4315    
4316  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4317        if (utf8 && (c > 255 || ((options & PCRE_CASELESS) != 0 && c > 127)))        if (utf && (c > 255 || ((options & PCRE_CASELESS) != 0 && c > 127)))
4318    #endif
4319    #ifndef COMPILE_PCRE8
4320          if (c > 255)
4321    #endif
4322    #if defined SUPPORT_UTF || defined COMPILE_PCRE16
4323          {          {
4324          class_utf8 = TRUE;          xclass = TRUE;
4325          *class_utf8data++ = XCL_SINGLE;          *class_uchardata++ = XCL_SINGLE;
4326          class_utf8data += PRIV(ord2utf8)(c, class_utf8data);  #ifdef SUPPORT_UTF
4327            class_uchardata += PRIV(ord2utf)(c, class_uchardata);
4328    #else
4329            *class_uchardata++ = c;
4330    #endif
4331    
4332  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
4333          if ((options & PCRE_CASELESS) != 0)          if ((options & PCRE_CASELESS) != 0)
# Line 4255  for (;; ptr++) Line 4335  for (;; ptr++)
4335            unsigned int othercase;            unsigned int othercase;
4336            if ((othercase = UCD_OTHERCASE(c)) != c)            if ((othercase = UCD_OTHERCASE(c)) != c)
4337              {              {
4338              *class_utf8data++ = XCL_SINGLE;              *class_uchardata++ = XCL_SINGLE;
4339              class_utf8data += PRIV(ord2utf8)(othercase, class_utf8data);              class_uchardata += PRIV(ord2utf)(othercase, class_uchardata);
4340              }              }
4341            }            }
4342  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
# Line 4308  for (;; ptr++) Line 4388  for (;; ptr++)
4388      The optimization throws away the bit map. We turn the item into a      The optimization throws away the bit map. We turn the item into a
4389      1-character OP_CHAR[I] if it's positive, or OP_NOT[I] if it's negative.      1-character OP_CHAR[I] if it's positive, or OP_NOT[I] if it's negative.
4390      Note that OP_NOT[I] does not support multibyte characters. In the positive      Note that OP_NOT[I] does not support multibyte characters. In the positive
4391      case, it can cause firstbyte to be set. Otherwise, there can be no first      case, it can cause firstchar to be set. Otherwise, there can be no first
4392      char if this item is first, whatever repeat count may follow. In the case      char if this item is first, whatever repeat count may follow. In the case
4393      of reqbyte, save the previous value for reinstating. */      of reqchar, save the previous value for reinstating. */
4394    
4395  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4396      if (class_charcount == 1 && !class_utf8 &&      if (class_charcount == 1 && !xclass &&
4397        (!utf8 || !negate_class || class_lastchar < 128))        (!utf || !negate_class || class_lastchar < 128))
4398  #else  #else
4399      if (class_charcount == 1)      if (class_charcount == 1)
4400  #endif  #endif
4401        {        {
4402        zeroreqbyte = reqbyte;        zeroreqchar = reqchar;
4403    
4404        /* The OP_NOT[I] opcodes work on one-byte characters only. */        /* The OP_NOT[I] opcodes work on one-byte characters only. */
4405    
4406        if (negate_class)        if (negate_class)
4407          {          {
4408          if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
4409          zerofirstbyte = firstbyte;          zerofirstchar = firstchar;
4410          *code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT;          *code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT;
4411          *code++ = class_lastchar;          *code++ = class_lastchar;
4412          break;          break;
# Line 4336  for (;; ptr++) Line 4416  for (;; ptr++)
4416        then we can handle this with the normal one-character code. */        then we can handle this with the normal one-character code. */
4417    
4418  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4419        if (utf8 && class_lastchar > 127)        if (utf && class_lastchar > 127)
4420          mclength = PRIV(ord2utf8)(class_lastchar, mcbuffer);          mclength = PRIV(ord2utf)(class_lastchar, mcbuffer);
4421        else        else
4422  #endif  #endif
4423          {          {
# Line 4349  for (;; ptr++) Line 4429  for (;; ptr++)
4429    
4430      /* The general case - not the one-char optimization. If this is the first      /* The general case - not the one-char optimization. If this is the first
4431      thing in the branch, there can be no first char setting, whatever the      thing in the branch, there can be no first char setting, whatever the
4432      repeat count. Any reqbyte setting must remain unchanged after any kind of      repeat count. Any reqchar setting must remain unchanged after any kind of
4433      repeat. */      repeat. */
4434    
4435      if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;      if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
4436      zerofirstbyte = firstbyte;      zerofirstchar = firstchar;
4437      zeroreqbyte = reqbyte;      zeroreqchar = reqchar;
4438    
4439      /* If there are characters with values > 255, we have to compile an      /* If there are characters with values > 255, we have to compile an
4440      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 4364  for (;; ptr++) Line 4444  for (;; ptr++)
4444      be listed) there are no characters < 256, we can omit the bitmap in the      be listed) there are no characters < 256, we can omit the bitmap in the
4445      actual compiled code. */      actual compiled code. */
4446    
4447  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4448      if (class_utf8 && (!should_flip_negation || (options & PCRE_UCP) != 0))      if (xclass && (!should_flip_negation || (options & PCRE_UCP) != 0))
4449    #endif
4450    #ifndef COMPILE_PCRE8
4451        if (xclass && !should_flip_negation)
4452    #endif
4453    #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
4454        {        {
4455        *class_utf8data++ = XCL_END;    /* Marks the end of extra data */        *class_uchardata++ = XCL_END;    /* Marks the end of extra data */
4456        *code++ = OP_XCLASS;        *code++ = OP_XCLASS;
4457        code += LINK_SIZE;        code += LINK_SIZE;
4458        *code = negate_class? XCL_NOT : 0;        *code = negate_class? XCL_NOT:0;
4459    
4460        /* If the map is required, move up the extra data to make room for it;        /* If the map is required, move up the extra data to make room for it;
4461        otherwise just move the code pointer to the end of the extra data. */        otherwise just move the code pointer to the end of the extra data. */
# Line 4378  for (;; ptr++) Line 4463  for (;; ptr++)
4463        if (class_charcount > 0)        if (class_charcount > 0)
4464          {          {
4465          *code++ |= XCL_MAP;          *code++ |= XCL_MAP;
4466          memmove(code + 32, code, class_utf8data - code);          memmove(code + (32 / sizeof(pcre_uchar)), code,
4467              IN_UCHARS(class_uchardata - code));
4468          memcpy(code, classbits, 32);          memcpy(code, classbits, 32);
4469          code = class_utf8data + 32;          code = class_uchardata + (32 / sizeof(pcre_uchar));
4470          }          }
4471        else code = class_utf8data;        else code = class_uchardata;
4472    
4473        /* Now fill in the complete length of the item */        /* Now fill in the complete length of the item */
4474    
# Line 4398  for (;; ptr++) Line 4484  for (;; ptr++)
4484      negating it if necessary. */      negating it if necessary. */
4485    
4486      *code++ = (negate_class == should_flip_negation) ? OP_CLASS : OP_NCLASS;      *code++ = (negate_class == should_flip_negation) ? OP_CLASS : OP_NCLASS;
4487      if (negate_class)      if (lengthptr == NULL)    /* Save time in the pre-compile phase */
       {  
       if (lengthptr == NULL)    /* Save time in the pre-compile phase */  
         for (c = 0; c < 32; c++) code[c] = ~classbits[c];  
       }  
     else  
4488        {        {
4489          if (negate_class)
4490            for (c = 0; c < 32; c++) classbits[c] = ~classbits[c];
4491        memcpy(code, classbits, 32);        memcpy(code, classbits, 32);
4492        }        }
4493      code += 32;      code += 32 / sizeof(pcre_uchar);
4494      break;      break;
4495    
4496    
# Line 4444  for (;; ptr++) Line 4527  for (;; ptr++)
4527    
4528      if (repeat_min == 0)      if (repeat_min == 0)
4529        {        {
4530        firstbyte = zerofirstbyte;    /* Adjust for zero repeat */        firstchar = zerofirstchar;    /* Adjust for zero repeat */
4531        reqbyte = zeroreqbyte;        /* Ditto */        reqchar = zeroreqchar;        /* Ditto */
4532        }        }
4533    
4534      /* Remember whether this is a variable length repeat */      /* Remember whether this is a variable length repeat */
# Line 4510  for (;; ptr++) Line 4593  for (;; ptr++)
4593    
4594      /* If previous was a character match, abolish the item and generate a      /* If previous was a character match, abolish the item and generate a
4595      repeat item instead. If a char item has a minumum of more than one, ensure      repeat item instead. If a char item has a minumum of more than one, ensure
4596      that it is set in reqbyte - it might not be if a sequence such as x{3} is      that it is set in reqchar - it might not be if a sequence such as x{3} is
4597      the first thing in a branch because the x will have gone into firstbyte      the first thing in a branch because the x will have gone into firstchar
4598      instead.  */      instead.  */
4599    
4600      if (*previous == OP_CHAR || *previous == OP_CHARI)      if (*previous == OP_CHAR || *previous == OP_CHARI)
4601        {        {
4602        op_type = (*previous == OP_CHAR)? 0 : OP_STARI - OP_STAR;        op_type = (*previous == OP_CHAR)? 0 : OP_STARI - OP_STAR;
4603    
4604        /* Deal with UTF-8 characters that take up more than one byte. It's        /* Deal with UTF characters that take up more than one character. It's
4605        easier to write this out separately than try to macrify it. Use c to        easier to write this out separately than try to macrify it. Use c to
4606        hold the length of the character in bytes, plus 0x80 to flag that it's a        hold the length of the character in bytes, plus 0x80 to flag that it's a
4607        length rather than a small character. */        length rather than a small character. */
4608    
4609  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4610        if (utf8 && (code[-1] & 0x80) != 0)        if (utf && (code[-1] & 0x80) != 0)
4611          {          {
4612          pcre_uchar *lastchar = code - 1;          pcre_uchar *lastchar = code - 1;
4613          while((*lastchar & 0xc0) == 0x80) lastchar--;          BACKCHAR(lastchar);
4614          c = code - lastchar;            /* Length of UTF-8 character */          c = code - lastchar;            /* Length of UTF-8 character */
4615          memcpy(utf8_char, lastchar, c); /* Save the char */          memcpy(utf_chars, lastchar, IN_UCHARS(c)); /* Save the char */
4616          c |= 0x80;                      /* Flag c as a length */          c |= 0x80;                      /* Flag c as a length */
4617          }          }
4618        else        else
4619  #endif  #endif
4620    
4621        /* Handle the case of a single byte - either with no UTF8 support, or        /* Handle the case of a single charater - either with no UTF support, or
4622        with UTF-8 disabled, or for a UTF-8 character < 128. */        with UTF disabled, or for a single character UTF character. */
4623    
4624          {          {
4625          c = code[-1];          c = code[-1];
4626          if (repeat_min > 1) reqbyte = c | req_caseopt | cd->req_varyopt;          if (repeat_min > 1) reqchar = c | req_caseopt | cd->req_varyopt;
4627          }          }
4628    
4629        /* 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 4550  for (;; ptr++) Line 4633  for (;; ptr++)
4633    
4634        if (!possessive_quantifier &&        if (!possessive_quantifier &&
4635            repeat_max < 0 &&            repeat_max < 0 &&
4636            check_auto_possessive(previous, utf8, ptr + 1, options, cd))            check_auto_possessive(previous, utf, ptr + 1, options, cd))
4637          {          {
4638          repeat_type = 0;    /* Force greedy */          repeat_type = 0;    /* Force greedy */
4639          possessive_quantifier = TRUE;          possessive_quantifier = TRUE;
# Line 4571  for (;; ptr++) Line 4654  for (;; ptr++)
4654        c = previous[1];        c = previous[1];
4655        if (!possessive_quantifier &&        if (!possessive_quantifier &&
4656            repeat_max < 0 &&            repeat_max < 0 &&
4657            check_auto_possessive(previous, utf8, ptr + 1, options, cd))            check_auto_possessive(previous, utf, ptr + 1, options, cd))
4658          {          {
4659          repeat_type = 0;    /* Force greedy */          repeat_type = 0;    /* Force greedy */
4660          possessive_quantifier = TRUE;          possessive_quantifier = TRUE;
# Line 4595  for (;; ptr++) Line 4678  for (;; ptr++)
4678    
4679        if (!possessive_quantifier &&        if (!possessive_quantifier &&
4680            repeat_max < 0 &&            repeat_max < 0 &&
4681            check_auto_possessive(previous, utf8, ptr + 1, options, cd))            check_auto_possessive(previous, utf, ptr + 1, options, cd))
4682          {          {
4683          repeat_type = 0;    /* Force greedy */          repeat_type = 0;    /* Force greedy */
4684          possessive_quantifier = TRUE;          possessive_quantifier = TRUE;
# Line 4680  for (;; ptr++) Line 4763  for (;; ptr++)
4763          if (repeat_max < 0)          if (repeat_max < 0)
4764            {            {
4765  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4766            if (utf8 && c >= 128)            if (utf && c >= 128)
4767              {              {
4768              memcpy(code, utf8_char, c & 7);              memcpy(code, utf_chars, c & 7);
4769              code += c & 7;              code += c & 7;
4770              }              }
4771            else            else
# Line 4705  for (;; ptr++) Line 4788  for (;; ptr++)
4788          else if (repeat_max != repeat_min)          else if (repeat_max != repeat_min)
4789            {            {
4790  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4791            if (utf8 && c >= 128)            if (utf && c >= 128)
4792              {              {
4793              memcpy(code, utf8_char, c & 7);              memcpy(code, utf_chars, c & 7);
4794              code += c & 7;              code += c & 7;
4795              }              }
4796            else            else
# Line 4735  for (;; ptr++) Line 4818  for (;; ptr++)
4818        /* The character or character type itself comes last in all cases. */        /* The character or character type itself comes last in all cases. */
4819    
4820  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4821        if (utf8 && c >= 128)        if (utf && c >= 128)
4822          {          {
4823          memcpy(code, utf8_char, c & 7);          memcpy(code, utf_chars, c & 7);
4824          code += c & 7;          code += c & 7;
4825          }          }
4826        else        else
# Line 4761  for (;; ptr++) Line 4844  for (;; ptr++)
4844    
4845      else if (*previous == OP_CLASS ||      else if (*previous == OP_CLASS ||
4846               *previous == OP_NCLASS ||               *previous == OP_NCLASS ||
4847  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF8 || !defined COMPILE_PCRE8
4848               *previous == OP_XCLASS ||               *previous == OP_XCLASS ||
4849  #endif  #endif
4850               *previous == OP_REF ||               *previous == OP_REF ||
# Line 4864  for (;; ptr++) Line 4947  for (;; ptr++)
4947          if (repeat_max <= 1)    /* Covers 0, 1, and unlimited */          if (repeat_max <= 1)    /* Covers 0, 1, and unlimited */
4948            {            {
4949            *code = OP_END;            *code = OP_END;
4950            adjust_recurse(previous, 1, utf8, cd, save_hwm);            adjust_recurse(previous, 1, utf, cd, save_hwm);
4951            memmove(previous + 1, previous, IN_UCHARS(len));            memmove(previous + 1, previous, IN_UCHARS(len));
4952            code++;            code++;
4953            if (repeat_max == 0)            if (repeat_max == 0)
# Line 4888  for (;; ptr++) Line 4971  for (;; ptr++)
4971            {            {
4972            int offset;            int offset;
4973            *code = OP_END;            *code = OP_END;
4974            adjust_recurse(previous, 2 + LINK_SIZE, utf8, cd, save_hwm);            adjust_recurse(previous, 2 + LINK_SIZE, utf, cd, save_hwm);
4975            memmove(previous + 2 + LINK_SIZE, previous, IN_UCHARS(len));            memmove(previous + 2 + LINK_SIZE, previous, IN_UCHARS(len));
4976            code += 2 + LINK_SIZE;            code += 2 + LINK_SIZE;
4977            *previous++ = OP_BRAZERO + repeat_type;            *previous++ = OP_BRAZERO + repeat_type;
# Line 4939  for (;; ptr++) Line 5022  for (;; ptr++)
5022    
5023            else            else
5024              {              {
5025              if (groupsetfirstbyte && reqbyte < 0) reqbyte = firstbyte;              if (groupsetfirstchar && reqchar < 0) reqchar = firstchar;
5026              for (i = 1; i < repeat_min; i++)              for (i = 1; i < repeat_min; i++)
5027                {                {
5028                pcre_uchar *hc;                pcre_uchar *hc;
# Line 5090  for (;; ptr++) Line 5173  for (;; ptr++)
5173              pcre_uchar *scode = bracode;              pcre_uchar *scode = bracode;
5174              do              do
5175                {                {
5176                if (could_be_empty_branch(scode, ketcode, utf8, cd))                if (could_be_empty_branch(scode, ketcode, utf, cd))
5177                  {                  {
5178                  *bracode += OP_SBRA - OP_BRA;                  *bracode += OP_SBRA - OP_BRA;
5179                  break;                  break;
# Line 5113  for (;; ptr++) Line 5196  for (;; ptr++)
5196                {                {
5197                int nlen = (int)(code - bracode);                int nlen = (int)(code - bracode);
5198                *code = OP_END;                *code = OP_END;
5199                adjust_recurse(bracode, 1 + LINK_SIZE, utf8, cd, save_hwm);                adjust_recurse(bracode, 1 + LINK_SIZE, utf, cd, save_hwm);
5200                memmove(bracode + 1 + LINK_SIZE, bracode, IN_UCHARS(nlen));                memmove(bracode + 1 + LINK_SIZE, bracode, IN_UCHARS(nlen));
5201                code += 1 + LINK_SIZE;                code += 1 + LINK_SIZE;
5202                nlen += 1 + LINK_SIZE;                nlen += 1 + LINK_SIZE;
# Line 5190  for (;; ptr++) Line 5273  for (;; ptr++)
5273        else if (*tempcode == OP_EXACT || *tempcode == OP_NOTEXACT)        else if (*tempcode == OP_EXACT || *tempcode == OP_NOTEXACT)
5274          {          {
5275          tempcode += PRIV(OP_lengths)[*tempcode];          tempcode += PRIV(OP_lengths)[*tempcode];
5276  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5277          if (utf8 && tempcode[-1] >= 0xc0)          if (utf && HAS_EXTRALEN(tempcode[-1]))
5278            tempcode += PRIV(utf8_table4)[tempcode[-1] & 0x3f];            tempcode += GET_EXTRALEN(tempcode[-1]);
5279  #endif  #endif
5280          }          }
5281    
# Line 5229  for (;; ptr++) Line 5312  for (;; ptr++)
5312    
5313          default:          default:
5314          *code = OP_END;          *code = OP_END;
5315          adjust_recurse(tempcode, 1 + LINK_SIZE, utf8, cd, save_hwm);          adjust_recurse(tempcode, 1 + LINK_SIZE, utf, cd, save_hwm);
5316          memmove(tempcode + 1 + LINK_SIZE, tempcode, IN_UCHARS(len));          memmove(tempcode + 1 + LINK_SIZE, tempcode, IN_UCHARS(len));
5317          code += 1 + LINK_SIZE;          code += 1 + LINK_SIZE;
5318          len += 1 + LINK_SIZE;          len += 1 + LINK_SIZE;
# Line 5242  for (;; ptr++) Line 5325  for (;; ptr++)
5325        }        }
5326    
5327      /* In all case we no longer have a previous item. We also set the      /* In all case we no longer have a previous item. We also set the
5328      "follows varying string" flag for subsequently encountered reqbytes if      "follows varying string" flag for subsequently encountered reqchars if
5329      it isn't already set and we have just passed a varying length item. */      it isn't already set and we have just passed a varying length item. */
5330    
5331      END_REPEAT:      END_REPEAT:
# Line 5320  for (;; ptr++) Line 5403  for (;; ptr++)
5403                }                }
5404              *code++ = (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;              *code++ = (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;
5405    
5406              /* Do not set firstbyte after *ACCEPT */              /* Do not set firstchar after *ACCEPT */
5407              if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;              if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
5408              }              }
5409    
5410            /* Handle other cases with/without an argument */            /* Handle other cases with/without an argument */
# Line 5474  for (;; ptr++) Line 5557  for (;; ptr++)
5557          while ((cd->ctypes[*ptr] & ctype_word) != 0)          while ((cd->ctypes[*ptr] & ctype_word) != 0)
5558            {            {
5559            if (recno >= 0)            if (recno >= 0)
5560              recno = ((digitab[*ptr] & ctype_digit) != 0)?              recno = (IS_DIGIT(*ptr))? recno * 10 + *ptr - CHAR_0 : -1;
               recno * 10 + *ptr - CHAR_0 : -1;  
5561            ptr++;            ptr++;
5562            }            }
5563          namelen = (int)(ptr - name);          namelen = (int)(ptr - name);
# Line 5539  for (;; ptr++) Line 5621  for (;; ptr++)
5621          /* Search the pattern for a forward reference */          /* Search the pattern for a forward reference */
5622    
5623          else if ((i = find_parens(cd, name, namelen,          else if ((i = find_parens(cd, name, namelen,
5624                          (options & PCRE_EXTENDED) != 0, utf8)) > 0)                          (options & PCRE_EXTENDED) != 0, utf)) > 0)
5625            {            {
5626            PUT2(code, 2+LINK_SIZE, i);            PUT2(code, 2+LINK_SIZE, i);
5627            code[1+LINK_SIZE]++;            code[1+LINK_SIZE]++;
# Line 5565  for (;; ptr++) Line 5647  for (;; ptr++)
5647            recno = 0;            recno = 0;
5648            for (i = 1; i < namelen; i++)            for (i = 1; i < namelen; i++)
5649              {              {
5650              if ((digitab[name[i]] & ctype_digit) == 0)              if (!IS_DIGIT(name[i]))
5651                {                {
5652                *errorcodeptr = ERR15;                *errorcodeptr = ERR15;
5653                goto FAILED;                goto FAILED;
# Line 5665  for (;; ptr++) Line 5747  for (;; ptr++)
5747          *code++ = OP_CALLOUT;          *code++ = OP_CALLOUT;
5748            {            {
5749            int n = 0;            int n = 0;
5750            while ((digitab[*(++ptr)] & ctype_digit) != 0)            ptr++;
5751              n = n * 10 + *ptr - CHAR_0;            while(IS_DIGIT(*ptr))
5752                n = n * 10 + *ptr++ - CHAR_0;
5753            if (*ptr != CHAR_RIGHT_PARENTHESIS)            if (*ptr != CHAR_RIGHT_PARENTHESIS)
5754              {              {
5755              *errorcodeptr = ERR39;              *errorcodeptr = ERR39;
# Line 5883  for (;; ptr++) Line 5966  for (;; ptr++)
5966            temp = cd->end_pattern;            temp = cd->end_pattern;
5967            cd->end_pattern = ptr;            cd->end_pattern = ptr;
5968            recno = find_parens(cd, name, namelen,            recno = find_parens(cd, name, namelen,
5969              (options & PCRE_EXTENDED) != 0, utf8);              (options & PCRE_EXTENDED) != 0, utf);
5970            cd->end_pattern = temp;            cd->end_pattern = temp;
5971            if (recno < 0) recno = 0;    /* Forward ref; set dummy number */            if (recno < 0) recno = 0;    /* Forward ref; set dummy number */
5972            }            }
# Line 5910  for (;; ptr++) Line 5993  for (;; ptr++)
5993              }              }
5994            else if ((recno =                /* Forward back reference */            else if ((recno =                /* Forward back reference */
5995                      find_parens(cd, name, namelen,                      find_parens(cd, name, namelen,
5996                        (options & PCRE_EXTENDED) != 0, utf8)) <= 0)                        (options & PCRE_EXTENDED) != 0, utf)) <= 0)
5997              {              {
5998              *errorcodeptr = ERR15;              *errorcodeptr = ERR15;
5999              goto FAILED;              goto FAILED;
# Line 5949  for (;; ptr++) Line 6032  for (;; ptr++)
6032            if ((refsign = *ptr) == CHAR_PLUS)            if ((refsign = *ptr) == CHAR_PLUS)
6033              {              {
6034              ptr++;              ptr++;
6035              if ((digitab[*ptr] & ctype_digit) == 0)              if (!IS_DIGIT(*ptr))
6036                {                {
6037                *errorcodeptr = ERR63;                *errorcodeptr = ERR63;
6038                goto FAILED;                goto FAILED;
# Line 5957  for (;; ptr++) Line 6040  for (;; ptr++)
6040              }              }
6041            else if (refsign == CHAR_MINUS)            else if (refsign == CHAR_MINUS)
6042              {              {
6043              if ((digitab[ptr[1]] & ctype_digit) == 0)              if (!IS_DIGIT(ptr[1]))
6044                goto OTHER_CHAR_AFTER_QUERY;                goto OTHER_CHAR_AFTER_QUERY;
6045              ptr++;              ptr++;
6046              }              }
6047    
6048            recno = 0;            recno = 0;
6049            while((digitab[*ptr] & ctype_digit) != 0)            while(IS_DIGIT(*ptr))
6050              recno = recno * 10 + *ptr++ - CHAR_0;              recno = recno * 10 + *ptr++ - CHAR_0;
6051    
6052            if (*ptr != terminator)            if (*ptr != terminator)
# Line 6014  for (;; ptr++) Line 6097  for (;; ptr++)
6097              {              {
6098              *code = OP_END;              *code = OP_END;
6099              if (recno != 0)              if (recno != 0)
6100                called = PRIV(find_bracket)(cd->start_code, utf8, recno);                called = PRIV(find_bracket)(cd->start_code, utf, recno);
6101    
6102              /* Forward reference */              /* Forward reference */
6103    
6104              if (called == NULL)              if (called == NULL)
6105                {                {
6106                if (find_parens(cd, NULL, recno,                if (find_parens(cd, NULL, recno,
6107                      (options & PCRE_EXTENDED) != 0, utf8) < 0)                      (options & PCRE_EXTENDED) != 0, utf) < 0)
6108                  {                  {
6109                  *errorcodeptr = ERR15;                  *errorcodeptr = ERR15;
6110                  goto FAILED;                  goto FAILED;
# Line 6045  for (;; ptr++) Line 6128  for (;; ptr++)
6128              conditional subpatterns will be picked up then. */              conditional subpatterns will be picked up then. */
6129    
6130              else if (GET(called, 1) == 0 && cond_depth <= 0 &&              else if (GET(called, 1) == 0 && cond_depth <= 0 &&
6131                       could_be_empty(called, code, bcptr, utf8, cd))                       could_be_empty(called, code, bcptr, utf, cd))
6132                {                {
6133                *errorcodeptr = ERR40;                *errorcodeptr = ERR40;
6134                goto FAILED;                goto FAILED;
# Line 6061  for (;; ptr++) Line 6144  for (;; ptr++)
6144    
6145          /* Can't determine a first byte now */          /* Can't determine a first byte now */
6146    
6147          if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
6148          continue;          continue;
6149    
6150    
# Line 6118  for (;; ptr++) Line 6201  for (;; ptr++)
6201          both phases.          both phases.
6202    
6203          If we are not at the pattern start, reset the greedy defaults and the          If we are not at the pattern start, reset the greedy defaults and the
6204          case value for firstbyte and reqbyte. */          case value for firstchar and reqchar. */
6205    
6206          if (*ptr == CHAR_RIGHT_PARENTHESIS)          if (*ptr == CHAR_RIGHT_PARENTHESIS)
6207            {            {
# Line 6131  for (;; ptr++) Line 6214  for (;; ptr++)
6214              {              {
6215              greedy_default = ((newoptions & PCRE_UNGREEDY) != 0);              greedy_default = ((newoptions & PCRE_UNGREEDY) != 0);
6216              greedy_non_default = greedy_default ^ 1;              greedy_non_default = greedy_default ^ 1;
6217              req_caseopt = ((newoptions & PCRE_CASELESS) != 0)? REQ_CASELESS : 0;              req_caseopt = ((newoptions & PCRE_CASELESS) != 0)? REQ_CASELESS:0;
6218              }              }
6219    
6220            /* Change options at this level, and pass them back for use            /* Change options at this level, and pass them back for use
# Line 6194  for (;; ptr++) Line 6277  for (;; ptr++)
6277           skipbytes,                       /* Skip over bracket number */           skipbytes,                       /* Skip over bracket number */
6278           cond_depth +           cond_depth +
6279             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */             ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */
6280           &subfirstbyte,                   /* For possible first char */           &subfirstchar,                   /* For possible first char */
6281           &subreqbyte,                     /* For possible last char */           &subreqchar,                     /* For possible last char */
6282           bcptr,                           /* Current branch chain */           bcptr,                           /* Current branch chain */
6283           cd,                              /* Tables block */           cd,                              /* Tables block */
6284           (lengthptr == NULL)? NULL :      /* Actual compile phase */           (lengthptr == NULL)? NULL :      /* Actual compile phase */
# Line 6246  for (;; ptr++) Line 6329  for (;; ptr++)
6329          }          }
6330    
6331        /* A "normal" conditional group. If there is just one branch, we must not        /* A "normal" conditional group. If there is just one branch, we must not
6332        make use of its firstbyte or reqbyte, because this is equivalent to an        make use of its firstchar or reqchar, because this is equivalent to an
6333        empty second branch. */        empty second branch. */
6334    
6335        else        else
# Line 6256  for (;; ptr++) Line 6339  for (;; ptr++)
6339            *errorcodeptr = ERR27;            *errorcodeptr = ERR27;
6340            goto FAILED;            goto FAILED;
6341            }            }
6342          if (condcount == 1) subfirstbyte = subreqbyte = REQ_NONE;          if (condcount == 1) subfirstchar = subreqchar = REQ_NONE;
6343          }          }
6344        }        }
6345    
# Line 6300  for (;; ptr++) Line 6383  for (;; ptr++)
6383      /* Handle updating of the required and first characters for other types of      /* Handle updating of the required and first characters for other types of
6384      group. Update for normal brackets of all kinds, and conditions with two      group. Update for normal brackets of all kinds, and conditions with two
6385      branches (see code above). If the bracket is followed by a quantifier with      branches (see code above). If the bracket is followed by a quantifier with
6386      zero repeat, we have to back off. Hence the definition of zeroreqbyte and      zero repeat, we have to back off. Hence the definition of zeroreqchar and
6387      zerofirstbyte outside the main loop so that they can be accessed for the      zerofirstchar outside the main loop so that they can be accessed for the
6388      back off. */      back off. */
6389    
6390      zeroreqbyte = reqbyte;      zeroreqchar = reqchar;
6391      zerofirstbyte = firstbyte;      zerofirstchar = firstchar;
6392      groupsetfirstbyte = FALSE;      groupsetfirstchar = FALSE;
6393    
6394      if (bravalue >= OP_ONCE)      if (bravalue >= OP_ONCE)
6395        {        {
6396        /* If we have not yet set a firstbyte in this branch, take it from the        /* If we have not yet set a firstchar in this branch, take it from the
6397        subpattern, remembering that it was set here so that a repeat of more        subpattern, remembering that it was set here so that a repeat of more
6398        than one can replicate it as reqbyte if necessary. If the subpattern has        than one can replicate it as reqchar if necessary. If the subpattern has
6399        no firstbyte, set "none" for the whole branch. In both cases, a zero        no firstchar, set "none" for the whole branch. In both cases, a zero
6400        repeat forces firstbyte to "none". */        repeat forces firstchar to "none". */
6401    
6402        if (firstbyte == REQ_UNSET)        if (firstchar == REQ_UNSET)
6403          {          {
6404          if (subfirstbyte >= 0)          if (subfirstchar >= 0)
6405            {            {
6406            firstbyte = subfirstbyte;            firstchar = subfirstchar;
6407            groupsetfirstbyte = TRUE;            groupsetfirstchar = TRUE;
6408            }            }
6409          else firstbyte = REQ_NONE;          else firstchar = REQ_NONE;
6410          zerofirstbyte = REQ_NONE;          zerofirstchar = REQ_NONE;
6411          }          }
6412    
6413        /* If firstbyte was previously set, convert the subpattern's firstbyte        /* If firstchar was previously set, convert the subpattern's firstchar
6414        into reqbyte 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
6415        existence beforehand. */        existence beforehand. */
6416    
6417        else if (subfirstbyte >= 0 && subreqbyte < 0)        else if (subfirstchar >= 0 && subreqchar < 0)
6418          subreqbyte = subfirstbyte | tempreqvary;          subreqchar = subfirstchar | tempreqvary;
6419    
6420        /* 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
6421        really the first byte - see above), set it. */        really the first byte - see above), set it. */
6422    
6423        if (subreqbyte >= 0) reqbyte = subreqbyte;        if (subreqchar >= 0) reqchar = subreqchar;
6424        }        }
6425    
6426      /* For a forward assertion, we take the reqbyte, if set. This can be      /* For a forward assertion, we take the reqchar, if set. This can be
6427      helpful if the pattern that follows the assertion doesn't set a different      helpful if the pattern that follows the assertion doesn't set a different
6428      char. For example, it's useful for /(?=abcde).+/. We can't set firstbyte      char. For example, it's useful for /(?=abcde).+/. We can't set firstchar
6429      for an assertion, however because it leads to incorrect effect for patterns      for an assertion, however because it leads to incorrect effect for patterns
6430      such as /(?=a)a.+/ when the "real" "a" would then become a reqbyte instead      such as /(?=a)a.+/ when the "real" "a" would then become a reqchar instead
6431      of a firstbyte. 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
6432      firstbyte, looking for an asserted first char. */      firstchar, looking for an asserted first char. */
6433    
6434      else if (bravalue == OP_ASSERT && subreqbyte >= 0) reqbyte = subreqbyte;      else if (bravalue == OP_ASSERT && subreqchar >= 0) reqchar = subreqchar;
6435      break;     /* End of processing '(' */      break;     /* End of processing '(' */
6436    
6437    
# Line 6381  for (;; ptr++) Line 6464  for (;; ptr++)
6464        /* For metasequences that actually match a character, we disable the        /* For metasequences that actually match a character, we disable the
6465        setting of a first character if it hasn't already been set. */        setting of a first character if it hasn't already been set. */
6466    
6467        if (firstbyte == REQ_UNSET && -c > ESC_b && -c < ESC_Z)        if (firstchar == REQ_UNSET && -c > ESC_b && -c < ESC_Z)
6468          firstbyte = REQ_NONE;          firstchar = REQ_NONE;
6469    
6470        /* 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. */
6471    
6472        zerofirstbyte = firstbyte;        zerofirstchar = firstchar;
6473        zeroreqbyte = reqbyte;        zeroreqchar = reqchar;
6474    
6475        /* \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'
6476        is a subroutine call by number (Oniguruma syntax). In fact, the value        is a subroutine call by number (Oniguruma syntax). In fact, the value
# Line 6438  for (;; ptr++) Line 6521  for (;; ptr++)
6521          /* Test a signed number in angle brackets or quotes. */          /* Test a signed number in angle brackets or quotes. */
6522    
6523          p = ptr + 2;          p = ptr + 2;
6524          while ((digitab[*p] & ctype_digit) != 0) p++;          while (IS_DIGIT(*p)) p++;
6525          if (*p != terminator)          if (*p != terminator)
6526            {            {
6527            *errorcodeptr = ERR57;            *errorcodeptr = ERR57;
# Line 6466  for (;; ptr++) Line 6549  for (;; ptr++)
6549          goto NAMED_REF_OR_RECURSE;          goto NAMED_REF_OR_RECURSE;
6550          }          }
6551    
6552        /* Back references are handled specially; must disable firstbyte if        /* Back references are handled specially; must disable firstchar if
6553        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
6554        ':' later. */        ':' later. */
6555    
# Line 6476  for (;; ptr++) Line 6559  for (;; ptr++)
6559          recno = -c - ESC_REF;          recno = -c - ESC_REF;
6560    
6561          HANDLE_REFERENCE:    /* Come here from named backref handling */          HANDLE_REFERENCE:    /* Come here from named backref handling */
6562          if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
6563          previous = code;          previous = code;
6564          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;          *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;
6565          PUT2INC(code, 0, recno);          PUT2INC(code, 0, recno);
# Line 6543  for (;; ptr++) Line 6626  for (;; ptr++)
6626    
6627            {            {
6628            previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;            previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;
6629            *code++ = (!utf8 && c == -ESC_C)? OP_ALLANY : -c;            *code++ = (!utf && c == -ESC_C)? OP_ALLANY : -c;
6630            }            }
6631          }          }
6632        continue;        continue;
# Line 6554  for (;; ptr++) Line 6637  for (;; ptr++)
6637      handle it as a data character. */      handle it as a data character. */
6638    
6639  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
6640      if (utf8 && c > 127)      if (utf && c > 127)
6641        mclength = PRIV(ord2utf8)(c, mcbuffer);        mclength = PRIV(ord2utf)(c, mcbuffer);
6642      else      else
6643  #endif  #endif
6644    
# Line 6576  for (;; ptr++) Line 6659  for (;; ptr++)
6659      mclength = 1;      mclength = 1;
6660      mcbuffer[0] = c;      mcbuffer[0] = c;
6661    
6662  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
6663      if (utf8 && c >= 0xc0)      if (utf && HAS_EXTRALEN(c))
6664        {        {
6665        while ((ptr[1] & 0xc0) == 0x80)        INTERNALCHAR(TRUE, ptr[1], mcbuffer[mclength++] = *(++ptr));
         mcbuffer[mclength++] = *(++ptr);  
6666        }        }
6667  #endif  #endif
6668    
# Line 6599  for (;; ptr++) Line 6681  for (;; ptr++)
6681    
6682      /* Set the first and required bytes appropriately. If no previous first      /* Set the first and required bytes appropriately. If no previous first
6683      byte, set it from this character, but revert to none on a zero repeat.      byte, set it from this character, but revert to none on a zero repeat.
6684      Otherwise, leave the firstbyte value alone, and don't change it on a zero      Otherwise, leave the firstchar value alone, and don't change it on a zero
6685      repeat. */      repeat. */
6686    
6687      if (firstbyte == REQ_UNSET)      if (firstchar == REQ_UNSET)
6688        {        {
6689        zerofirstbyte = REQ_NONE;        zerofirstchar = REQ_NONE;
6690        zeroreqbyte = reqbyte;        zeroreqchar = reqchar;
6691    
6692        /* If the character is more than one byte long, we can set firstbyte        /* If the character is more than one byte long, we can set firstchar
6693        only if it is not to be matched caselessly. */        only if it is not to be matched caselessly. */
6694    
6695        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
6696          {          {
6697          firstbyte = mcbuffer[0] | req_caseopt;          firstchar = mcbuffer[0] | req_caseopt;
6698          if (mclength != 1) reqbyte = code[-1] | cd->req_varyopt;          if (mclength != 1) reqchar = code[-1] | cd->req_varyopt;
6699          }          }
6700        else firstbyte = reqbyte = REQ_NONE;        else firstchar = reqchar = REQ_NONE;
6701        }        }
6702    
6703      /* firstbyte was previously set; we can set reqbyte only if the length is      /* firstchar was previously set; we can set reqchar only if the length is
6704      1 or the matching is caseful. */      1 or the matching is caseful. */
6705    
6706      else      else
6707        {        {
6708        zerofirstbyte = firstbyte;        zerofirstchar = firstchar;
6709        zeroreqbyte = reqbyte;        zeroreqchar = reqchar;
6710        if (mclength == 1 || req_caseopt == 0)        if (mclength == 1 || req_caseopt == 0)
6711          reqbyte = code[-1] | req_caseopt | cd->req_varyopt;          reqchar = code[-1] | req_caseopt | cd->req_varyopt;
6712        }        }
6713    
6714      break;            /* End of literal character handling */      break;            /* End of literal character handling */
# Line 6666  Arguments: Line 6748  Arguments:
6748    reset_bracount TRUE to reset the count for each branch    reset_bracount TRUE to reset the count for each branch
6749    skipbytes      skip this many bytes at start (for brackets and OP_COND)    skipbytes      skip this many bytes at start (for brackets and OP_COND)
6750    cond_depth     depth of nesting for conditional subpatterns    cond_depth     depth of nesting for conditional subpatterns
6751    firstbyteptr   place to put the first required character, or a negative number    firstcharptr   place to put the first required character, or a negative number
6752    reqbyteptr     place to put the last required character, or a negative number    reqcharptr     place to put the last required character, or a negative number
6753    bcptr          pointer to the chain of currently open branches    bcptr          pointer to the chain of currently open branches
6754    cd             points to the data block with tables pointers etc.    cd             points to the data block with tables pointers etc.
6755    lengthptr      NULL during the real compile phase    lengthptr      NULL during the real compile phase
# Line 6679  Returns:         TRUE on success Line 6761  Returns:         TRUE on success
6761  static BOOL  static BOOL
6762  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,  compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr,
6763    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
6764    int cond_depth, int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr,    int cond_depth, pcre_int32 *firstcharptr, pcre_int32 *reqcharptr,
6765    compile_data *cd, int *lengthptr)    branch_chain *bcptr, compile_data *cd, int *lengthptr)
6766  {  {
6767  const pcre_uchar *ptr = *ptrptr;  const pcre_uchar *ptr = *ptrptr;
6768  pcre_uchar *code = *codeptr;  pcre_uchar *code = *codeptr;
# Line 6689  pcre_uchar *start_bracket = code; Line 6771  pcre_uchar *start_bracket = code;
6771  pcre_uchar *reverse_count = NULL;  pcre_uchar *reverse_count = NULL;
6772  open_capitem capitem;  open_capitem capitem;
6773  int capnumber = 0;  int capnumber = 0;
6774  int firstbyte, reqbyte;  pcre_int32 firstchar, reqchar;
6775  int branchfirstbyte, branchreqbyte;  pcre_int32 branchfirstchar, branchreqchar;
6776  int length;  int length;
6777  int orig_bracount;  int orig_bracount;
6778  int max_bracount;  int max_bracount;
# Line 6699  branch_chain bc; Line 6781  branch_chain bc;
6781  bc.outer = bcptr;  bc.outer = bcptr;
6782  bc.current_branch = code;  bc.current_branch = code;
6783    
6784  firstbyte = reqbyte = REQ_UNSET;  firstchar = reqchar = REQ_UNSET;
6785    
6786  /* 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
6787  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 6758  for (;;) Line 6840  for (;;)
6840    /* Now compile the branch; in the pre-compile phase its length gets added    /* Now compile the branch; in the pre-compile phase its length gets added
6841    into the length. */    into the length. */
6842    
6843    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstbyte,    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar,
6844          &branchreqbyte, &bc, cond_depth, cd,          &branchreqchar, &bc, cond_depth, cd,
6845          (lengthptr == NULL)? NULL : &length))          (lengthptr == NULL)? NULL : &length))
6846      {      {
6847      *ptrptr = ptr;      *ptrptr = ptr;
# Line 6775  for (;;) Line 6857  for (;;)
6857    
6858    if (lengthptr == NULL)    if (lengthptr == NULL)
6859      {      {
6860      /* If this is the first branch, the firstbyte and reqbyte values for the      /* If this is the first branch, the firstchar and reqchar values for the
6861      branch become the values for the regex. */      branch become the values for the regex. */
6862    
6863      if (*last_branch != OP_ALT)      if (*last_branch != OP_ALT)
6864        {        {
6865        firstbyte = branchfirstbyte;        firstchar = branchfirstchar;
6866        reqbyte = branchreqbyte;        reqchar = branchreqchar;
6867        }        }
6868    
6869      /* If this is not the first branch, the first char and reqbyte have to      /* If this is not the first branch, the first char and reqchar have to
6870      match the values from all the previous branches, except that if the      match the values from all the previous branches, except that if the
6871      previous value for reqbyte didn't have REQ_VARY set, it can still match,      previous value for reqchar didn't have REQ_VARY set, it can still match,
6872      and we set REQ_VARY for the regex. */      and we set REQ_VARY for the regex. */
6873    
6874      else      else
6875        {        {
6876        /* If we previously had a firstbyte, but it doesn't match the new branch,        /* If we previously had a firstchar, but it doesn't match the new branch,
6877        we have to abandon the firstbyte for the regex, but if there was        we have to abandon the firstchar for the regex, but if there was
6878        previously no reqbyte, it takes on the value of the old firstbyte. */        previously no reqchar, it takes on the value of the old firstchar. */
6879    
6880        if (firstbyte >= 0 && firstbyte != branchfirstbyte)        if (firstchar >= 0 && firstchar != branchfirstchar)
6881          {          {
6882          if (reqbyte < 0) reqbyte = firstbyte;          if (reqchar < 0) reqchar = firstchar;
6883          firstbyte = REQ_NONE;          firstchar = REQ_NONE;
6884          }          }
6885    
6886        /* If we (now or from before) have no firstbyte, a firstbyte from the        /* If we (now or from before) have no firstchar, a firstchar from the
6887        branch becomes a reqbyte if there isn't a branch reqbyte. */        branch becomes a reqchar if there isn't a branch reqchar. */
6888    
6889        if (firstbyte < 0 && branchfirstbyte >= 0 && branchreqbyte < 0)        if (firstchar < 0 && branchfirstchar >= 0 && branchreqchar < 0)
6890            branchreqbyte = branchfirstbyte;            branchreqchar = branchfirstchar;
6891    
6892        /* Now ensure that the reqbytes match */        /* Now ensure that the reqchars match */
6893    
6894        if ((reqbyte & ~REQ_VARY) != (branchreqbyte & ~REQ_VARY))        if ((reqchar & ~REQ_VARY) != (branchreqchar & ~REQ_VARY))
6895          reqbyte = REQ_NONE;          reqchar = REQ_NONE;
6896        else reqbyte |= branchreqbyte;   /* To "or" REQ_VARY */        else reqchar |= branchreqchar;   /* To "or" REQ_VARY */
6897        }        }
6898    
6899      /* 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 6901  for (;;) Line 6983  for (;;)
6983    
6984      *codeptr = code;      *codeptr = code;
6985      *ptrptr = ptr;      *ptrptr = ptr;
6986      *firstbyteptr = firstbyte;      *firstcharptr = firstchar;
6987      *reqbyteptr = reqbyte;      *reqcharptr = reqchar;
6988      if (lengthptr != NULL)      if (lengthptr != NULL)
6989        {        {
6990        if (OFLOW_MAX - *lengthptr < length)        if (OFLOW_MAX - *lengthptr < length)
# Line 7281  pcre16_compile2(PCRE_SPTR16 pattern, int Line 7363  pcre16_compile2(PCRE_SPTR16 pattern, int
7363  {  {
7364  real_pcre *re;  real_pcre *re;
7365  int length = 1;  /* For final END opcode */  int length = 1;  /* For final END opcode */
7366  int firstbyte, reqbyte, newline;  pcre_int32 firstchar, reqchar;
7367    int newline;
7368  int errorcode = 0;  int errorcode = 0;
7369  int skipatstart = 0;  int skipatstart = 0;
7370  BOOL utf8;  BOOL utf;
7371  size_t size;  size_t size;
7372  pcre_uchar *code;  pcre_uchar *code;
7373  const pcre_uchar *codestart;  const pcre_uchar *codestart;
# Line 7382  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 7465  while (ptr[skipatstart] == CHAR_LEFT_PAR
7465    else break;    else break;
7466    }    }
7467    
7468  utf8 = (options & PCRE_UTF8) != 0;  /* PCRE_UTF16 has the same value as PCRE_UTF8. */
7469    utf = (options & PCRE_UTF8) != 0;
7470    
7471  /* Can't support UTF8 unless PCRE has been compiled to include the code. The  /* Can't support UTF8 unless PCRE has been compiled to include the code. The
7472  return of an error code from PRIV(valid_utf8)() is a new feature, introduced in  return of an error code from PRIV(valid_utf)() is a new feature, introduced in
7473  release 8.13. It is passed back from pcre_[dfa_]exec(), but at the moment is  release 8.13. It is passed back from pcre_[dfa_]exec(), but at the moment is
7474  not used here. */  not used here. */
7475    
7476  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
7477  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0 &&  if (utf && (options & PCRE_NO_UTF8_CHECK) == 0 &&
7478       (errorcode = PRIV(valid_utf8)((PCRE_PUCHAR)pattern, -1, erroroffset)) != 0)       (errorcode = PRIV(valid_utf)((PCRE_PUCHAR)pattern, -1, erroroffset)) != 0)
7479    {    {
7480    errorcode = ERR44;    errorcode = ERR44;
7481    goto PCRE_EARLY_ERROR_RETURN2;    goto PCRE_EARLY_ERROR_RETURN2;
7482    }    }
7483  #else  #else
7484  if (utf8)  if (utf)
7485    {    {
7486    errorcode = ERR32;    errorcode = ERR32;
7487    goto PCRE_EARLY_ERROR_RETURN;    goto PCRE_EARLY_ERROR_RETURN;
# Line 7509  ptr += skipatstart; Line 7593  ptr += skipatstart;
7593  code = cworkspace;  code = cworkspace;
7594  *code = OP_BRA;  *code = OP_BRA;
7595  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,  (void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE,
7596    FALSE, 0, 0, &firstbyte, &reqbyte, NULL, cd, &length);    FALSE, 0, 0, &firstchar, &reqchar, NULL, cd, &length);
7597  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;  if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN;
7598    
7599  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,  DPRINTF(("end pre-compile: length=%d workspace=%d\n", length,
# Line 7546  re->size = (int)size; Line 7630  re->size = (int)size;
7630  re->options = cd->external_options;  re->options = cd->external_options;
7631  re->flags = cd->external_flags;  re->flags = cd->external_flags;
7632  re->dummy1 = 0;  re->dummy1 = 0;
7633  re->first_byte = 0;  re->first_char = 0;
7634  re->req_byte = 0;  re->req_char = 0;
7635  re->name_table_offset = sizeof(real_pcre) / sizeof(pcre_uchar);  re->name_table_offset = sizeof(real_pcre) / sizeof(pcre_uchar);
7636  re->name_entry_size = cd->name_entry_size;  re->name_entry_size = cd->name_entry_size;
7637  re->name_count = cd->names_found;  re->name_count = cd->names_found;
# Line 7583  ptr = (const pcre_uchar *)pattern + skip Line 7667  ptr = (const pcre_uchar *)pattern + skip
7667  code = (pcre_uchar *)codestart;  code = (pcre_uchar *)codestart;
7668  *code = OP_BRA;  *code = OP_BRA;
7669  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,  (void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0,
7670    &firstbyte, &reqbyte, NULL, cd, NULL);    &firstchar, &reqchar, NULL, cd, NULL);
7671  re->top_bracket = cd->bracount;  re->top_bracket = cd->bracount;
7672  re->top_backref = cd->top_backref;  re->top_backref = cd->top_backref;
7673  re->flags = cd->external_flags;  re->flags = cd->external_flags;
7674    
7675  if (cd->had_accept) reqbyte = REQ_NONE;   /* Must disable after (*ACCEPT) */  if (cd->had_accept) reqchar = REQ_NONE;   /* Must disable after (*ACCEPT) */
7676    
7677  /* 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. */
7678    
# Line 7612  while (errorcode == 0 && cd->hwm > cwork Line 7696  while (errorcode == 0 && cd->hwm > cwork
7696    cd->hwm -= LINK_SIZE;    cd->hwm -= LINK_SIZE;
7697    offset = GET(cd->hwm, 0);    offset = GET(cd->hwm, 0);
7698    recno = GET(codestart, offset);    recno = GET(codestart, offset);
7699    groupptr = PRIV(find_bracket)(codestart, utf8, recno);    groupptr = PRIV(find_bracket)(codestart, utf, recno);
7700    if (groupptr == NULL) errorcode = ERR53;    if (groupptr == NULL) errorcode = ERR53;
7701      else PUT(((pcre_uchar *)codestart), offset, (int)(groupptr - codestart));      else PUT(((pcre_uchar *)codestart), offset, (int)(groupptr - codestart));
7702    }    }
# Line 7639  if (cd->check_lookbehind) Line 7723  if (cd->check_lookbehind)
7723    of zero, but that is a pathological case, and it does no harm.) When we find    of zero, but that is a pathological case, and it does no harm.) When we find
7724    one, we temporarily terminate the branch it is in while we scan it. */    one, we temporarily terminate the branch it is in while we scan it. */
7725    
7726    for (cc = (pcre_uchar *)PRIV(find_bracket)(codestart, utf8, -1);    for (cc = (pcre_uchar *)PRIV(find_bracket)(codestart, utf, -1);
7727         cc != NULL;         cc != NULL;
7728         cc = (pcre_uchar *)PRIV(find_bracket)(cc, utf8, -1))         cc = (pcre_uchar *)PRIV(find_bracket)(cc, utf, -1))
7729      {      {
7730      if (GET(cc, 1) == 0)      if (GET(cc, 1) == 0)
7731        {        {
# Line 7694  if ((re->options & PCRE_ANCHORED) == 0) Line 7778  if ((re->options & PCRE_ANCHORED) == 0)
7778      re->options |= PCRE_ANCHORED;      re->options |= PCRE_ANCHORED;
7779    else    else
7780      {      {
7781      if (firstbyte < 0)      if (firstchar < 0)
7782        firstbyte = find_firstassertedchar(codestart, FALSE);        firstchar = find_firstassertedchar(codestart, FALSE);
7783      if (firstbyte >= 0)   /* Remove caseless flag for non-caseable chars */      if (firstchar >= 0)   /* Remove caseless flag for non-caseable chars */
7784        {        {
7785        int ch = firstbyte & 255;  #ifdef COMPILE_PCRE8
7786        re->first_byte = ((firstbyte & REQ_CASELESS) != 0 &&        re->first_char = firstchar & 0xff;
7787           cd->fcc[ch] == ch)? ch : firstbyte;  #else
7788    #ifdef COMPILE_PCRE16
7789          re->first_char = firstchar & 0xffff;
7790    #endif
7791    #endif
7792          if ((firstchar & REQ_CASELESS) != 0 && MAX_255(re->first_char)
7793            && cd->fcc[re->first_char] != re->first_char)
7794            re->flags |= PCRE_FCH_CASELESS;
7795    
7796        re->flags |= PCRE_FIRSTSET;        re->flags |= PCRE_FIRSTSET;
7797        }        }
7798      else if (is_startline(codestart, 0, cd->backref_map))      else if (is_startline(codestart, 0, cd->backref_map))
# Line 7712  if ((re->options & PCRE_ANCHORED) == 0) Line 7804  if ((re->options & PCRE_ANCHORED) == 0)
7804  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
7805  bytes. */  bytes. */
7806    
7807  if (reqbyte >= 0 &&  if (reqchar >= 0 &&
7808       ((re->options & PCRE_ANCHORED) == 0 || (reqbyte & REQ_VARY) != 0))       ((re->options & PCRE_ANCHORED) == 0 || (reqchar & REQ_VARY) != 0))
7809    {    {
7810    int ch = reqbyte & 255;  #ifdef COMPILE_PCRE8
7811    re->req_byte = ((reqbyte & REQ_CASELESS) != 0 &&    re->req_char = reqchar & 0xff;
7812      cd->fcc[ch] == ch)? (reqbyte & ~REQ_CASELESS) : reqbyte;  #else
7813    #ifdef COMPILE_PCRE16
7814      re->req_char = reqchar & 0xffff;
7815    #endif
7816    #endif
7817      if ((reqchar & REQ_CASELESS) != 0 && MAX_255(re->req_char)
7818        && cd->fcc[re->req_char] != re->req_char)
7819        re->flags |= PCRE_RCH_CASELESS;
7820    
7821    re->flags |= PCRE_REQCHSET;    re->flags |= PCRE_REQCHSET;
7822    }    }
7823    
# Line 7732  printf("Options=%08x\n", re->options); Line 7832  printf("Options=%08x\n", re->options);
7832    
7833  if ((re->flags & PCRE_FIRSTSET) != 0)  if ((re->flags & PCRE_FIRSTSET) != 0)
7834    {    {
7835    int ch = re->first_byte & 255;    pcre_uchar ch = re->first_char;
7836    const char *caseless = ((re->first_byte & REQ_CASELESS) == 0)?    const char *caseless =
7837      "" : " (caseless)";      ((re->flags & PCRE_FCH_CASELESS) == 0)? "" : " (caseless)";
7838    if (isprint(ch)) printf("First char = %c%s\n", ch, caseless);    if (PRINTABLE(ch)) printf("First char = %c%s\n", ch, caseless);
7839      else printf("First char = \\x%02x%s\n", ch, caseless);      else printf("First char = \\x%02x%s\n", ch, caseless);
7840    }    }
7841    
7842  if ((re->flags & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
7843    {    {
7844    int ch = re->req_byte & 255;    pcre_uchar ch = re->req_char;
7845    const char *caseless = ((re->req_byte & REQ_CASELESS) == 0)?    const char *caseless =
7846      "" : " (caseless)";      ((re->flags & PCRE_RCH_CASELESS) == 0)? "" : " (caseless)";
7847    if (isprint(ch)) printf("Req char = %c%s\n", ch, caseless);    if (PRINTABLE(ch)) printf("Req char = %c%s\n", ch, caseless);
7848      else printf("Req char = \\x%02x%s\n", ch, caseless);      else printf("Req char = \\x%02x%s\n", ch, caseless);
7849    }    }
7850    

Legend:
Removed from v.769  
changed lines
  Added in v.782

  ViewVC Help
Powered by ViewVC 1.1.5