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

Diff of /code/trunk/pcre_compile.c

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

revision 1045 by ph10, Sun Sep 23 16:50:00 2012 UTC revision 1062 by chpe, Tue Oct 16 15:54:05 2012 UTC
# Line 53  supporting internal functions that are n Line 53  supporting internal functions that are n
53  #include "pcre_internal.h"  #include "pcre_internal.h"
54    
55    
56  /* When PCRE_DEBUG is defined, we need the pcre(16)_printint() function, which  /* When PCRE_DEBUG is defined, we need the pcre(16|32)_printint() function, which
57  is also used by pcretest. PCRE_DEBUG is not defined when building a production  is also used by pcretest. PCRE_DEBUG is not defined when building a production
58  library. We do not need to select pcre16_printint.c specially, because the  library. We do not need to select pcre16_printint.c specially, because the
59  COMPILE_PCREx macro will already be appropriately set. */  COMPILE_PCREx macro will already be appropriately set. */
# Line 123  overrun before it actually does run off Line 123  overrun before it actually does run off
123    
124  #define REQ_CASELESS   0x10000000l      /* Indicates caselessness */  #define REQ_CASELESS   0x10000000l      /* Indicates caselessness */
125  #define REQ_VARY       0x20000000l      /* Reqchar followed non-literal item */  #define REQ_VARY       0x20000000l      /* Reqchar followed non-literal item */
126    #define REQ_MASK       (REQ_CASELESS | REQ_VARY)
127    
128  /* Repeated character flags. */  /* Repeated character flags. */
129    
# Line 503  static const char error_texts[] = Line 504  static const char error_texts[] =
504    /* 75 */    /* 75 */
505    "name is too long in (*MARK), (*PRUNE), (*SKIP), or (*THEN)\0"    "name is too long in (*MARK), (*PRUNE), (*SKIP), or (*THEN)\0"
506    "character value in \\u.... sequence is too large\0"    "character value in \\u.... sequence is too large\0"
507      "invalid UTF-32 string\0"
508    ;    ;
509    
510  /* 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 747  return (*p == CHAR_RIGHT_CURLY_BRACKET); Line 749  return (*p == CHAR_RIGHT_CURLY_BRACKET);
749  *************************************************/  *************************************************/
750    
751  /* This function is called when a \ has been encountered. It either returns a  /* This function is called when a \ has been encountered. It either returns a
752  positive value for a simple escape such as \n, or a negative value which  positive value for a simple escape such as \n, or 0 for a data character
753  encodes one of the more complicated things such as \d. A backreference to group  which will be placed in chptr. A backreference to group n is returned as
754  n is returned as -(ESC_REF + n); ESC_REF is the highest ESC_xxx macro. When  negative n. When UTF-8 is enabled, a positive value greater than 255 may
755  UTF-8 is enabled, a positive value greater than 255 may be returned. On entry,  be returned in chptr.
756  ptr is pointing at the \. On exit, it is on the final character of the escape  On entry,ptr is pointing at the \. On exit, it is on the final character of the
757  sequence.  escape sequence.
758    
759  Arguments:  Arguments:
760    ptrptr         points to the pattern position pointer    ptrptr         points to the pattern position pointer
761      chptr          points to the data character
762    errorcodeptr   points to the errorcode variable    errorcodeptr   points to the errorcode variable
763    bracount       number of previous extracting brackets    bracount       number of previous extracting brackets
764    options        the options bits    options        the options bits
765    isclass        TRUE if inside a character class    isclass        TRUE if inside a character class
766    
767  Returns:         zero or positive => a data character  Returns:         zero => a data character
768                   negative => a special escape sequence                   positive => a special escape sequence
769                     negative => a back reference
770                   on error, errorcodeptr is set                   on error, errorcodeptr is set
771  */  */
772    
773  static int  static int
774  check_escape(const pcre_uchar **ptrptr, int *errorcodeptr, int bracount,  check_escape(const pcre_uchar **ptrptr, pcre_uint32 *chptr, int *errorcodeptr,
775    int options, BOOL isclass)    int bracount, int options, BOOL isclass)
776  {  {
777  /* PCRE_UTF16 has the same value as PCRE_UTF8. */  /* PCRE_UTF16 has the same value as PCRE_UTF8. */
778  BOOL utf = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
779  const pcre_uchar *ptr = *ptrptr + 1;  const pcre_uchar *ptr = *ptrptr + 1;
780  pcre_int32 c;  pcre_uint32 c;
781    int escape = 0;
782  int i;  int i;
783    
784  GETCHARINCTEST(c, ptr);           /* Get character value, increment pointer */  GETCHARINCTEST(c, ptr);           /* Get character value, increment pointer */
# Line 790  Otherwise further processing may be requ Line 795  Otherwise further processing may be requ
795  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
796  /* Not alphanumeric */  /* Not alphanumeric */
797  else if (c < CHAR_0 || c > CHAR_z) {}  else if (c < CHAR_0 || c > CHAR_z) {}
798  else if ((i = escapes[c - CHAR_0]) != 0) c = i;  else if ((i = escapes[c - CHAR_0]) != 0) { if (i > 0) c = (pcre_uint32)i; else escape = -i; }
799    
800  #else           /* EBCDIC coding */  #else           /* EBCDIC coding */
801  /* Not alphanumeric */  /* Not alphanumeric */
802  else if (c < CHAR_a || (!MAX_255(c) || (ebcdic_chartab[c] & 0x0E) == 0)) {}  else if (c < CHAR_a || (!MAX_255(c) || (ebcdic_chartab[c] & 0x0E) == 0)) {}
803  else if ((i = escapes[c - 0x48]) != 0)  c = i;  else if ((i = escapes[c - 0x48]) != 0)  { if (i > 0) c = (pcre_uint32)i; else escape = -i; }
804  #endif  #endif
805    
806  /* Escapes that need further processing, or are illegal. */  /* Escapes that need further processing, or are illegal. */
# Line 803  else if ((i = escapes[c - 0x48]) != 0) Line 808  else if ((i = escapes[c - 0x48]) != 0)
808  else  else
809    {    {
810    const pcre_uchar *oldptr;    const pcre_uchar *oldptr;
811    BOOL braced, negated;    BOOL braced, negated, overflow;
812      int s;
813    
814    switch (c)    switch (c)
815      {      {
# Line 828  else Line 834  else
834          c = 0;          c = 0;
835          for (i = 0; i < 4; ++i)          for (i = 0; i < 4; ++i)
836            {            {
837            register int cc = *(++ptr);            register pcre_uint32 cc = *(++ptr);
838  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
839            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
840            c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));            c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
# Line 838  else Line 844  else
844  #endif  #endif
845            }            }
846    
847  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
848          if (c > (utf ? 0x10ffff : 0xff))          if (c > (utf ? 0x10ffff : 0xff))
849  #else  #elif defined COMPILE_PCRE16
 #ifdef COMPILE_PCRE16  
850          if (c > (utf ? 0x10ffff : 0xffff))          if (c > (utf ? 0x10ffff : 0xffff))
851  #endif  #elif defined COMPILE_PCRE32
852            if (utf && c > 0x10ffff)
853  #endif  #endif
854            {            {
855            *errorcodeptr = ERR76;            *errorcodeptr = ERR76;
# Line 875  else Line 881  else
881      (3) For Oniguruma compatibility we also support \g followed by a name or a      (3) For Oniguruma compatibility we also support \g followed by a name or a
882      number either in angle brackets or in single quotes. However, these are      number either in angle brackets or in single quotes. However, these are
883      (possibly recursive) subroutine calls, _not_ backreferences. Just return      (possibly recursive) subroutine calls, _not_ backreferences. Just return
884      the -ESC_g code (cf \k). */      the ESC_g code (cf \k). */
885    
886      case CHAR_g:      case CHAR_g:
887      if (isclass) break;      if (isclass) break;
888      if (ptr[1] == CHAR_LESS_THAN_SIGN || ptr[1] == CHAR_APOSTROPHE)      if (ptr[1] == CHAR_LESS_THAN_SIGN || ptr[1] == CHAR_APOSTROPHE)
889        {        {
890        c = -ESC_g;        escape = ESC_g;
891        break;        break;
892        }        }
893    
# Line 894  else Line 900  else
900          if (*p != CHAR_MINUS && !IS_DIGIT(*p)) break;          if (*p != CHAR_MINUS && !IS_DIGIT(*p)) break;
901        if (*p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET)        if (*p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET)
902          {          {
903          c = -ESC_k;          escape = ESC_k;
904          break;          break;
905          }          }
906        braced = TRUE;        braced = TRUE;
# Line 910  else Line 916  else
916      else negated = FALSE;      else negated = FALSE;
917    
918      /* The integer range is limited by the machine's int representation. */      /* The integer range is limited by the machine's int representation. */
919      c = 0;      s = 0;
920        overflow = FALSE;
921      while (IS_DIGIT(ptr[1]))      while (IS_DIGIT(ptr[1]))
922        {        {
923        if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */        if (s > INT_MAX / 10 - 1) /* Integer overflow */
924          {          {
925          c = -1;          overflow = TRUE;
926          break;          break;
927          }          }
928        c = c * 10 + *(++ptr) - CHAR_0;        s = s * 10 + (int)(*(++ptr) - CHAR_0);
929        }        }
930      if (((unsigned int)c) > INT_MAX) /* Integer overflow */      if (overflow) /* Integer overflow */
931        {        {
932        while (IS_DIGIT(ptr[1]))        while (IS_DIGIT(ptr[1]))
933          ptr++;          ptr++;
# Line 934  else Line 941  else
941        break;        break;
942        }        }
943    
944      if (c == 0)      if (s == 0)
945        {        {
946        *errorcodeptr = ERR58;        *errorcodeptr = ERR58;
947        break;        break;
# Line 942  else Line 949  else
949    
950      if (negated)      if (negated)
951        {        {
952        if (c > bracount)        if (s > bracount)
953          {          {
954          *errorcodeptr = ERR15;          *errorcodeptr = ERR15;
955          break;          break;
956          }          }
957        c = bracount - (c - 1);        s = bracount - (s - 1);
958        }        }
959    
960      c = -(ESC_REF + c);      escape = -s;
961      break;      break;
962    
963      /* The handling of escape sequences consisting of a string of digits      /* The handling of escape sequences consisting of a string of digits
# Line 972  else Line 979  else
979        {        {
980        oldptr = ptr;        oldptr = ptr;
981        /* The integer range is limited by the machine's int representation. */        /* The integer range is limited by the machine's int representation. */
982        c -= CHAR_0;        s = (int)(c -CHAR_0);
983          overflow = FALSE;
984        while (IS_DIGIT(ptr[1]))        while (IS_DIGIT(ptr[1]))
985          {          {
986          if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */          if (s > INT_MAX / 10 - 1) /* Integer overflow */
987            {            {
988            c = -1;            overflow = TRUE;
989            break;            break;
990            }            }
991          c = c * 10 + *(++ptr) - CHAR_0;          s = s * 10 + (int)(*(++ptr) - CHAR_0);
992          }          }
993        if (((unsigned int)c) > INT_MAX) /* Integer overflow */        if (overflow) /* Integer overflow */
994          {          {
995          while (IS_DIGIT(ptr[1]))          while (IS_DIGIT(ptr[1]))
996            ptr++;            ptr++;
997          *errorcodeptr = ERR61;          *errorcodeptr = ERR61;
998          break;          break;
999          }          }
1000        if (c < 10 || c <= bracount)        if (s < 10 || s <= bracount)
1001          {          {
1002          c = -(ESC_REF + c);          escape = -s;
1003          break;          break;
1004          }          }
1005        ptr = oldptr;      /* Put the pointer back and fall through */        ptr = oldptr;      /* Put the pointer back and fall through */
# Line 1038  else Line 1046  else
1046          c = 0;          c = 0;
1047          for (i = 0; i < 2; ++i)          for (i = 0; i < 2; ++i)
1048            {            {
1049            register int cc = *(++ptr);            register pcre_uint32 cc = *(++ptr);
1050  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
1051            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */            if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
1052            c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));            c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
# Line 1056  else Line 1064  else
1064        const pcre_uchar *pt = ptr + 2;        const pcre_uchar *pt = ptr + 2;
1065    
1066        c = 0;        c = 0;
1067          overflow = FALSE;
1068        while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0)        while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0)
1069          {          {
1070          register int cc = *pt++;          register pcre_uint32 cc = *pt++;
1071          if (c == 0 && cc == CHAR_0) continue;     /* Leading zeroes */          if (c == 0 && cc == CHAR_0) continue;     /* Leading zeroes */
1072    
1073    #ifdef COMPILE_PCRE32
1074            if (c >= 0x10000000l) { overflow = TRUE; break; }
1075    #endif
1076    
1077  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
1078          if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */          if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
1079          c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));          c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
# Line 1069  else Line 1082  else
1082          c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));          c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));
1083  #endif  #endif
1084    
1085  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1086          if (c > (utf ? 0x10ffff : 0xff)) { c = -1; break; }          if (c > (utf ? 0x10ffff : 0xff)) { overflow = TRUE; break; }
1087  #else  #elif defined COMPILE_PCRE16
1088  #ifdef COMPILE_PCRE16          if (c > (utf ? 0x10ffff : 0xffff)) { overflow = TRUE; break; }
1089          if (c > (utf ? 0x10ffff : 0xffff)) { c = -1; break; }  #elif defined COMPILE_PCRE32
1090  #endif          if (utf && c > 0x10ffff) { overflow = TRUE; break; }
1091  #endif  #endif
1092          }          }
1093    
1094        if (c < 0)        if (overflow)
1095          {          {
1096          while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0) pt++;          while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0) pt++;
1097          *errorcodeptr = ERR34;          *errorcodeptr = ERR34;
# Line 1100  else Line 1113  else
1113      c = 0;      c = 0;
1114      while (i++ < 2 && MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0)      while (i++ < 2 && MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0)
1115        {        {
1116        int cc;                                  /* Some compilers don't like */        pcre_uint32 cc;                          /* Some compilers don't like */
1117        cc = *(++ptr);                           /* ++ in initializers */        cc = *(++ptr);                           /* ++ in initializers */
1118  #ifndef EBCDIC  /* ASCII/UTF-8 coding */  #ifndef EBCDIC  /* ASCII/UTF-8 coding */
1119        if (cc >= CHAR_a) cc -= 32;              /* Convert to upper case */        if (cc >= CHAR_a) cc -= 32;              /* Convert to upper case */
# Line 1159  else Line 1172  else
1172  newline". PCRE does not support \N{name}. However, it does support  newline". PCRE does not support \N{name}. However, it does support
1173  quantification such as \N{2,3}. */  quantification such as \N{2,3}. */
1174    
1175  if (c == -ESC_N && ptr[1] == CHAR_LEFT_CURLY_BRACKET &&  if (escape == ESC_N && ptr[1] == CHAR_LEFT_CURLY_BRACKET &&
1176       !is_counted_repeat(ptr+2))       !is_counted_repeat(ptr+2))
1177    *errorcodeptr = ERR37;    *errorcodeptr = ERR37;
1178    
1179  /* If PCRE_UCP is set, we change the values for \d etc. */  /* If PCRE_UCP is set, we change the values for \d etc. */
1180    
1181  if ((options & PCRE_UCP) != 0 && c <= -ESC_D && c >= -ESC_w)  if ((options & PCRE_UCP) != 0 && escape >= ESC_D && escape <= ESC_w)
1182    c -= (ESC_DU - ESC_D);    escape += (ESC_DU - ESC_D);
1183    
1184  /* Set the pointer to the final character before returning. */  /* Set the pointer to the final character before returning. */
1185    
1186  *ptrptr = ptr;  *ptrptr = ptr;
1187  return c;  *chptr = c;
1188    return escape;
1189  }  }
1190    
   
   
1191  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
1192  /*************************************************  /*************************************************
1193  *               Handle \P and \p                 *  *               Handle \P and \p                 *
# Line 1367  Arguments: Line 1379  Arguments:
1379    name         name to seek, or NULL if seeking a numbered subpattern    name         name to seek, or NULL if seeking a numbered subpattern
1380    lorn         name length, or subpattern number if name is NULL    lorn         name length, or subpattern number if name is NULL
1381    xmode        TRUE if we are in /x mode    xmode        TRUE if we are in /x mode
1382    utf          TRUE if we are in UTF-8 / UTF-16 mode    utf          TRUE if we are in UTF-8 / UTF-16 / UTF-32 mode
1383    count        pointer to the current capturing subpattern number (updated)    count        pointer to the current capturing subpattern number (updated)
1384    
1385  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 1601  Arguments: Line 1613  Arguments:
1613    name         name to seek, or NULL if seeking a numbered subpattern    name         name to seek, or NULL if seeking a numbered subpattern
1614    lorn         name length, or subpattern number if name is NULL    lorn         name length, or subpattern number if name is NULL
1615    xmode        TRUE if we are in /x mode    xmode        TRUE if we are in /x mode
1616    utf          TRUE if we are in UTF-8 / UTF-16 mode    utf          TRUE if we are in UTF-8 / UTF-16 / UTF-32 mode
1617    
1618  Returns:       the number of the found subpattern, or -1 if not found  Returns:       the number of the found subpattern, or -1 if not found
1619  */  */
# Line 1704  and doing the check at the end; a flag s Line 1716  and doing the check at the end; a flag s
1716    
1717  Arguments:  Arguments:
1718    code     points to the start of the pattern (the bracket)    code     points to the start of the pattern (the bracket)
1719    utf      TRUE in UTF-8 / UTF-16 mode    utf      TRUE in UTF-8 / UTF-16 / UTF-32 mode
1720    atend    TRUE if called when the pattern is complete    atend    TRUE if called when the pattern is complete
1721    cd       the "compile data" structure    cd       the "compile data" structure
1722    
# Line 1838  for (;;) Line 1850  for (;;)
1850      case OP_NOTI:      case OP_NOTI:
1851      branchlength++;      branchlength++;
1852      cc += 2;      cc += 2;
1853  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
1854      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
1855  #endif  #endif
1856      break;      break;
# Line 1852  for (;;) Line 1864  for (;;)
1864      case OP_NOTEXACTI:      case OP_NOTEXACTI:
1865      branchlength += GET2(cc,1);      branchlength += GET2(cc,1);
1866      cc += 2 + IMM2_SIZE;      cc += 2 + IMM2_SIZE;
1867  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
1868      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);      if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]);
1869  #endif  #endif
1870      break;      break;
1871    
1872      case OP_TYPEEXACT:      case OP_TYPEEXACT:
1873      branchlength += GET2(cc,1);      branchlength += GET2(cc,1);
1874      if (cc[1 + IMM2_SIZE] == OP_PROP || cc[1 + IMM2_SIZE] == OP_NOTPROP) cc += 2;      if (cc[1 + IMM2_SIZE] == OP_PROP || cc[1 + IMM2_SIZE] == OP_NOTPROP)
1875          cc += 2;
1876      cc += 1 + IMM2_SIZE + 1;      cc += 1 + IMM2_SIZE + 1;
1877      break;      break;
1878    
# Line 1894  for (;;) Line 1907  for (;;)
1907    
1908      /* Check a class for variable quantification */      /* Check a class for variable quantification */
1909    
1910  #if defined SUPPORT_UTF || defined COMPILE_PCRE16  #if defined SUPPORT_UTF || defined COMPILE_PCRE16 || defined COMPILE_PCRE32
1911      case OP_XCLASS:      case OP_XCLASS:
1912      cc += GET(cc, 1) - PRIV(OP_lengths)[OP_CLASS];      cc += GET(cc, 1) - PRIV(OP_lengths)[OP_CLASS];
1913      /* Fall through */      /* Fall through */
# Line 2033  length. Line 2046  length.
2046    
2047  Arguments:  Arguments:
2048    code        points to start of expression    code        points to start of expression
2049    utf         TRUE in UTF-8 / UTF-16 mode    utf         TRUE in UTF-8 / UTF-16 / UTF-32 mode
2050    number      the required bracket number or negative to find a lookbehind    number      the required bracket number or negative to find a lookbehind
2051    
2052  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
# Line 2097  for (;;) Line 2110  for (;;)
2110        case OP_TYPEMINUPTO:        case OP_TYPEMINUPTO:
2111        case OP_TYPEEXACT:        case OP_TYPEEXACT:
2112        case OP_TYPEPOSUPTO:        case OP_TYPEPOSUPTO:
2113        if (code[1 + IMM2_SIZE] == OP_PROP        if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)
2114          || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;          code += 2;
2115        break;        break;
2116    
2117        case OP_MARK:        case OP_MARK:
# Line 2120  for (;;) Line 2133  for (;;)
2133    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
2134    arrange to skip the extra bytes. */    arrange to skip the extra bytes. */
2135    
2136  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
2137      if (utf) switch(c)      if (utf) switch(c)
2138        {        {
2139        case OP_CHAR:        case OP_CHAR:
# Line 2172  instance of OP_RECURSE. Line 2185  instance of OP_RECURSE.
2185    
2186  Arguments:  Arguments:
2187    code        points to start of expression    code        points to start of expression
2188    utf         TRUE in UTF-8 / UTF-16 mode    utf         TRUE in UTF-8 / UTF-16 / UTF-32 mode
2189    
2190  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
2191  */  */
# Line 2217  for (;;) Line 2230  for (;;)
2230        case OP_TYPEUPTO:        case OP_TYPEUPTO:
2231        case OP_TYPEMINUPTO:        case OP_TYPEMINUPTO:
2232        case OP_TYPEEXACT:        case OP_TYPEEXACT:
2233        if (code[1 + IMM2_SIZE] == OP_PROP        if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)
2234          || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;          code += 2;
2235        break;        break;
2236    
2237        case OP_MARK:        case OP_MARK:
# Line 2240  for (;;) Line 2253  for (;;)
2253      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
2254      to arrange to skip the extra bytes. */      to arrange to skip the extra bytes. */
2255    
2256  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
2257      if (utf) switch(c)      if (utf) switch(c)
2258        {        {
2259        case OP_CHAR:        case OP_CHAR:
# Line 2326  bracket whose current branch will alread Line 2339  bracket whose current branch will alread
2339  Arguments:  Arguments:
2340    code        points to start of search    code        points to start of search
2341    endcode     points to where to stop    endcode     points to where to stop
2342    utf         TRUE if in UTF-8 / UTF-16 mode    utf         TRUE if in UTF-8 / UTF-16 / UTF-32 mode
2343    cd          contains pointers to tables etc.    cd          contains pointers to tables etc.
2344    
2345  Returns:      TRUE if what is matched could be empty  Returns:      TRUE if what is matched could be empty
# Line 2543  for (code = first_significant_code(code Line 2556  for (code = first_significant_code(code
2556      case OP_TYPEUPTO:      case OP_TYPEUPTO:
2557      case OP_TYPEMINUPTO:      case OP_TYPEMINUPTO:
2558      case OP_TYPEPOSUPTO:      case OP_TYPEPOSUPTO:
2559      if (code[1 + IMM2_SIZE] == OP_PROP      if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP)
2560        || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2;        code += 2;
2561      break;      break;
2562    
2563      /* End of branch */      /* End of branch */
# Line 2559  for (code = first_significant_code(code Line 2572  for (code = first_significant_code(code
2572      /* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO,      /* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO,
2573      MINUPTO, and POSUPTO may be followed by a multibyte character */      MINUPTO, and POSUPTO may be followed by a multibyte character */
2574    
2575  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
2576      case OP_STAR:      case OP_STAR:
2577      case OP_STARI:      case OP_STARI:
2578      case OP_MINSTAR:      case OP_MINSTAR:
# Line 2625  Arguments: Line 2638  Arguments:
2638    code        points to start of the recursion    code        points to start of the recursion
2639    endcode     points to where to stop (current RECURSE item)    endcode     points to where to stop (current RECURSE item)
2640    bcptr       points to the chain of current (unclosed) branch starts    bcptr       points to the chain of current (unclosed) branch starts
2641    utf         TRUE if in UTF-8 / UTF-16 mode    utf         TRUE if in UTF-8 / UTF-16 / UTF-32 mode
2642    cd          pointers to tables etc    cd          pointers to tables etc
2643    
2644  Returns:      TRUE if what is matched could be empty  Returns:      TRUE if what is matched could be empty
# Line 2772  value in the reference (which is a group Line 2785  value in the reference (which is a group
2785  Arguments:  Arguments:
2786    group      points to the start of the group    group      points to the start of the group
2787    adjust     the amount by which the group is to be moved    adjust     the amount by which the group is to be moved
2788    utf        TRUE in UTF-8 / UTF-16 mode    utf        TRUE in UTF-8 / UTF-16 / UTF-32 mode
2789    cd         contains pointers to tables etc.    cd         contains pointers to tables etc.
2790    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
2791    
# Line 2951  Returns:       TRUE if auto-possessifyin Line 2964  Returns:       TRUE if auto-possessifyin
2964  static BOOL  static BOOL
2965  check_char_prop(int c, int ptype, int pdata, BOOL negated)  check_char_prop(int c, int ptype, int pdata, BOOL negated)
2966  {  {
2967    #ifdef SUPPORT_UCP
2968    const pcre_uint32 *p;
2969    #endif
2970    
2971  const ucd_record *prop = GET_UCD(c);  const ucd_record *prop = GET_UCD(c);
2972    
2973  switch(ptype)  switch(ptype)
2974    {    {
2975    case PT_LAMP:    case PT_LAMP:
# Line 2989  switch(ptype) Line 3007  switch(ptype)
3007    return (PRIV(ucp_gentype)[prop->chartype] == ucp_L ||    return (PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
3008            PRIV(ucp_gentype)[prop->chartype] == ucp_N ||            PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
3009            c == CHAR_UNDERSCORE) == negated;            c == CHAR_UNDERSCORE) == negated;
3010    
3011    #ifdef SUPPORT_UCP
3012      case PT_CLIST:
3013      p = PRIV(ucd_caseless_sets) + prop->caseset;
3014      for (;;)
3015        {
3016        if ((unsigned int)c < *p) return !negated;
3017        if ((unsigned int)c == *p++) return negated;
3018        }
3019      break;  /* Control never reaches here */
3020    #endif
3021    }    }
3022    
3023  return FALSE;  return FALSE;
3024  }  }
3025  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
# Line 3006  sense to automatically possessify the re Line 3036  sense to automatically possessify the re
3036    
3037  Arguments:  Arguments:
3038    previous      pointer to the repeated opcode    previous      pointer to the repeated opcode
3039    utf           TRUE in UTF-8 / UTF-16 mode    utf           TRUE in UTF-8 / UTF-16 / UTF-32 mode
3040    ptr           next character in pattern    ptr           next character in pattern
3041    options       options bits    options       options bits
3042    cd            contains pointers to tables etc.    cd            contains pointers to tables etc.
# Line 3018  static BOOL Line 3048  static BOOL
3048  check_auto_possessive(const pcre_uchar *previous, BOOL utf,  check_auto_possessive(const pcre_uchar *previous, BOOL utf,
3049    const pcre_uchar *ptr, int options, compile_data *cd)    const pcre_uchar *ptr, int options, compile_data *cd)
3050  {  {
3051  pcre_int32 c, next;  pcre_int32 c = NOTACHAR; // FIXMEchpe pcre_uint32
3052    pcre_int32 next;
3053    int escape;
3054  int op_code = *previous++;  int op_code = *previous++;
3055    
3056  /* Skip whitespace and comments in extended mode */  /* Skip whitespace and comments in extended mode */
# Line 3050  value is a character, a negative value i Line 3082  value is a character, a negative value i
3082  if (*ptr == CHAR_BACKSLASH)  if (*ptr == CHAR_BACKSLASH)
3083    {    {
3084    int temperrorcode = 0;    int temperrorcode = 0;
3085    next = check_escape(&ptr, &temperrorcode, cd->bracount, options, FALSE);    escape = check_escape(&ptr, &next, &temperrorcode, cd->bracount, options, FALSE);
3086    if (temperrorcode != 0) return FALSE;    if (temperrorcode != 0) return FALSE;
3087    ptr++;    /* Point after the escape sequence */    ptr++;    /* Point after the escape sequence */
3088    }    }
3089  else if (!MAX_255(*ptr) || (cd->ctypes[*ptr] & ctype_meta) == 0)  else if (!MAX_255(*ptr) || (cd->ctypes[*ptr] & ctype_meta) == 0)
3090    {    {
3091      escape = 0;
3092  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3093    if (utf) { GETCHARINC(next, ptr); } else    if (utf) { GETCHARINC(next, ptr); } else
3094  #endif  #endif
# Line 3092  if (*ptr == CHAR_ASTERISK || *ptr == CHA Line 3125  if (*ptr == CHAR_ASTERISK || *ptr == CHA
3125    STRNCMP_UC_C8(ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0)    STRNCMP_UC_C8(ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0)
3126      return FALSE;      return FALSE;
3127    
3128  /* Now compare the next item with the previous opcode. First, handle cases when  /* If the previous item is a character, get its value. */
 the next item is a character. */  
3129    
3130  if (next >= 0) switch(op_code)  if (op_code == OP_CHAR || op_code == OP_CHARI ||
3131        op_code == OP_NOT || op_code == OP_NOTI)
3132      //if (escape == 0) switch(op_code)
3133    {    {
   case OP_CHAR:  
3134  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3135    GETCHARTEST(c, previous);    GETCHARTEST(c, previous);
3136  #else  #else
3137    c = *previous;    c = *previous;
3138  #endif  #endif
3139    return c != next;    }
3140    
3141    /* For CHARI (caseless character) we must check the other case. If we have  /* Now compare the next item with the previous opcode. First, handle cases when
3142    Unicode property support, we can use it to test the other case of  the next item is a character. */
   high-valued characters. */  
3143    
3144    case OP_CHARI:  if (escape == 0)
3145  #ifdef SUPPORT_UTF    {
3146    GETCHARTEST(c, previous);    /* For a caseless UTF match, the next character may have more than one other
3147  #else    case, which maps to the special PT_CLIST property. Check this first. */
3148    c = *previous;  
 #endif  
   if (c == next) return FALSE;  
 #ifdef SUPPORT_UTF  
   if (utf)  
     {  
     unsigned int othercase;  
     if (next < 128) othercase = cd->fcc[next]; else  
3149  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3150      othercase = UCD_OTHERCASE((unsigned int)next);    if (utf && (unsigned int)c != NOTACHAR && (options & PCRE_CASELESS) != 0)
3151  #else      {
3152      othercase = NOTACHAR;      int ocs = UCD_CASESET(next);
3153  #endif      if (ocs > 0) return check_char_prop(c, PT_CLIST, ocs, op_code >= OP_NOT);
     return (unsigned int)c != othercase;  
3154      }      }
   else  
 #endif  /* SUPPORT_UTF */  
   return (c != TABLE_GET((unsigned int)next, cd->fcc, next));  /* Non-UTF-8 mode */  
   
   case OP_NOT:  
 #ifdef SUPPORT_UTF  
   GETCHARTEST(c, previous);  
 #else  
   c = *previous;  
3155  #endif  #endif
   return c == next;  
3156    
3157    case OP_NOTI:    switch(op_code)
3158        {
3159        case OP_CHAR:
3160        return c != next;
3161    
3162        /* For CHARI (caseless character) we must check the other case. If we have
3163        Unicode property support, we can use it to test the other case of
3164        high-valued characters. We know that next can have only one other case,
3165        because multi-other-case characters are dealt with above. */
3166    
3167        case OP_CHARI:
3168        if (c == next) return FALSE;
3169  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3170    GETCHARTEST(c, previous);      if (utf)
3171          {
3172          unsigned int othercase;
3173          if (next < 128) othercase = cd->fcc[next]; else
3174    #ifdef SUPPORT_UCP
3175          othercase = UCD_OTHERCASE((unsigned int)next);
3176  #else  #else
3177    c = *previous;        othercase = NOTACHAR;
3178  #endif  #endif
3179    if (c == next) return TRUE;        return (unsigned int)c != othercase;
3180          }
3181        else
3182    #endif  /* SUPPORT_UTF */
3183        return (c != TABLE_GET((unsigned int)next, cd->fcc, next));  /* Not UTF */
3184    
3185        case OP_NOT:
3186        return c == next;
3187    
3188        case OP_NOTI:
3189        if (c == next) return TRUE;
3190  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3191    if (utf)      if (utf)
3192      {        {
3193      unsigned int othercase;        unsigned int othercase;
3194      if (next < 128) othercase = cd->fcc[next]; else        if (next < 128) othercase = cd->fcc[next]; else
3195  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3196      othercase = UCD_OTHERCASE((unsigned int)next);        othercase = UCD_OTHERCASE((unsigned int)next);
3197  #else  #else
3198      othercase = NOTACHAR;        othercase = NOTACHAR;
3199  #endif  #endif
3200      return (unsigned int)c == othercase;        return (unsigned int)c == othercase;
3201      }        }
3202    else      else
3203  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
3204    return (c == TABLE_GET((unsigned int)next, cd->fcc, next));  /* Non-UTF-8 mode */      return (c == TABLE_GET((unsigned int)next, cd->fcc, next));  /* Not UTF */
3205    
3206    /* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* set.      /* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* set.
3207    When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */      When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */
3208    
3209    case OP_DIGIT:      case OP_DIGIT:
3210    return next > 255 || (cd->ctypes[next] & ctype_digit) == 0;      return next > 255 || (cd->ctypes[next] & ctype_digit) == 0;
3211    
3212    case OP_NOT_DIGIT:      case OP_NOT_DIGIT:
3213    return next <= 255 && (cd->ctypes[next] & ctype_digit) != 0;      return next <= 255 && (cd->ctypes[next] & ctype_digit) != 0;
3214    
3215    case OP_WHITESPACE:      case OP_WHITESPACE:
3216    return next > 255 || (cd->ctypes[next] & ctype_space) == 0;      return next > 255 || (cd->ctypes[next] & ctype_space) == 0;
3217    
3218    case OP_NOT_WHITESPACE:      case OP_NOT_WHITESPACE:
3219    return next <= 255 && (cd->ctypes[next] & ctype_space) != 0;      return next <= 255 && (cd->ctypes[next] & ctype_space) != 0;
3220    
3221    case OP_WORDCHAR:      case OP_WORDCHAR:
3222    return next > 255 || (cd->ctypes[next] & ctype_word) == 0;      return next > 255 || (cd->ctypes[next] & ctype_word) == 0;
3223    
3224    case OP_NOT_WORDCHAR:      case OP_NOT_WORDCHAR:
3225    return next <= 255 && (cd->ctypes[next] & ctype_word) != 0;      return next <= 255 && (cd->ctypes[next] & ctype_word) != 0;
3226    
3227    case OP_HSPACE:      case OP_HSPACE:
3228    case OP_NOT_HSPACE:      case OP_NOT_HSPACE:
3229    switch(next)      switch(next)
3230      {        {
3231      HSPACE_CASES:        HSPACE_CASES:
3232      return op_code == OP_NOT_HSPACE;        return op_code == OP_NOT_HSPACE;
   
     default:  
     return op_code != OP_NOT_HSPACE;  
     }  
3233    
3234    case OP_ANYNL:        default:
3235    case OP_VSPACE:        return op_code != OP_NOT_HSPACE;
3236    case OP_NOT_VSPACE:        }
3237    switch(next)  
3238      {      case OP_ANYNL:
3239      VSPACE_CASES:      case OP_VSPACE:
3240      return op_code == OP_NOT_VSPACE;      case OP_NOT_VSPACE:
3241        switch(next)
3242      default:        {
3243      return op_code != OP_NOT_VSPACE;        VSPACE_CASES:
3244      }        return op_code == OP_NOT_VSPACE;
3245    
3246          default:
3247          return op_code != OP_NOT_VSPACE;
3248          }
3249    
3250  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3251    case OP_PROP:      case OP_PROP:
3252    return check_char_prop(next, previous[0], previous[1], FALSE);      return check_char_prop(next, previous[0], previous[1], FALSE);
3253    
3254    case OP_NOTPROP:      case OP_NOTPROP:
3255    return check_char_prop(next, previous[0], previous[1], TRUE);      return check_char_prop(next, previous[0], previous[1], TRUE);
3256  #endif  #endif
3257    
3258    default:      default:
3259    return FALSE;      return FALSE;
3260        }
3261    }    }
3262    
   
3263  /* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP  /* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP
3264  is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are  is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are
3265  generated only when PCRE_UCP is *not* set, that is, when only ASCII  generated only when PCRE_UCP is *not* set, that is, when only ASCII
# Line 3230  switch(op_code) Line 3270  switch(op_code)
3270    {    {
3271    case OP_CHAR:    case OP_CHAR:
3272    case OP_CHARI:    case OP_CHARI:
3273  #ifdef SUPPORT_UTF    switch(escape)
   GETCHARTEST(c, previous);  
 #else  
   c = *previous;  
 #endif  
   switch(-next)  
3274      {      {
3275      case ESC_d:      case ESC_d:
3276      return c > 255 || (cd->ctypes[c] & ctype_digit) == 0;      return c > 255 || (cd->ctypes[c] & ctype_digit) == 0;
# Line 3260  switch(op_code) Line 3295  switch(op_code)
3295      switch(c)      switch(c)
3296        {        {
3297        HSPACE_CASES:        HSPACE_CASES:
3298        return -next != ESC_h;        return escape != ESC_h;
3299    
3300        default:        default:
3301        return -next == ESC_h;        return escape == ESC_h;
3302        }        }
3303    
3304      case ESC_v:      case ESC_v:
# Line 3271  switch(op_code) Line 3306  switch(op_code)
3306      switch(c)      switch(c)
3307        {        {
3308        VSPACE_CASES:        VSPACE_CASES:
3309        return -next != ESC_v;        return escape != ESC_v;
3310    
3311        default:        default:
3312        return -next == ESC_v;        return escape == ESC_v;
3313        }        }
3314    
3315      /* When PCRE_UCP is set, these values get generated for \d etc. Find      /* When PCRE_UCP is set, these values get generated for \d etc. Find
3316      their substitutions and process them. The result will always be either      their substitutions and process them. The result will always be either
3317      -ESC_p or -ESC_P. Then fall through to process those values. */      ESC_p or ESC_P. Then fall through to process those values. */
3318    
3319  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3320      case ESC_du:      case ESC_du:
# Line 3290  switch(op_code) Line 3325  switch(op_code)
3325      case ESC_SU:      case ESC_SU:
3326        {        {
3327        int temperrorcode = 0;        int temperrorcode = 0;
3328        ptr = substitutes[-next - ESC_DU];        ptr = substitutes[escape - ESC_DU];
3329        next = check_escape(&ptr, &temperrorcode, 0, options, FALSE);        escape = check_escape(&ptr, &next, &temperrorcode, 0, options, FALSE);
3330        if (temperrorcode != 0) return FALSE;        if (temperrorcode != 0) return FALSE;
3331        ptr++;    /* For compatibility */        ptr++;    /* For compatibility */
3332        }        }
# Line 3318  switch(op_code) Line 3353  switch(op_code)
3353    
3354        /* Do the property check. */        /* Do the property check. */
3355    
3356        return check_char_prop(c, ptype, pdata, (next == -ESC_P) != negated);        return check_char_prop(c, ptype, pdata, (escape == ESC_P) != negated);
3357        }        }
3358  #endif  #endif
3359    
# Line 3333  switch(op_code) Line 3368  switch(op_code)
3368    these op-codes are never generated.) */    these op-codes are never generated.) */
3369    
3370    case OP_DIGIT:    case OP_DIGIT:
3371    return next == -ESC_D || next == -ESC_s || next == -ESC_W ||    return escape == ESC_D || escape == ESC_s || escape == ESC_W ||
3372           next == -ESC_h || next == -ESC_v || next == -ESC_R;           escape == ESC_h || escape == ESC_v || escape == ESC_R;
3373    
3374    case OP_NOT_DIGIT:    case OP_NOT_DIGIT:
3375    return next == -ESC_d;    return escape == ESC_d;
3376    
3377    case OP_WHITESPACE:    case OP_WHITESPACE:
3378    return next == -ESC_S || next == -ESC_d || next == -ESC_w;    return escape == ESC_S || escape == ESC_d || escape == ESC_w;
3379    
3380    case OP_NOT_WHITESPACE:    case OP_NOT_WHITESPACE:
3381    return next == -ESC_s || next == -ESC_h || next == -ESC_v || next == -ESC_R;    return escape == ESC_s || escape == ESC_h || escape == ESC_v || escape == ESC_R;
3382    
3383    case OP_HSPACE:    case OP_HSPACE:
3384    return next == -ESC_S || next == -ESC_H || next == -ESC_d ||    return escape == ESC_S || escape == ESC_H || escape == ESC_d ||
3385           next == -ESC_w || next == -ESC_v || next == -ESC_R;           escape == ESC_w || escape == ESC_v || escape == ESC_R;
3386    
3387    case OP_NOT_HSPACE:    case OP_NOT_HSPACE:
3388    return next == -ESC_h;    return escape == ESC_h;
3389    
3390    /* Can't have \S in here because VT matches \S (Perl anomaly) */    /* Can't have \S in here because VT matches \S (Perl anomaly) */
3391    case OP_ANYNL:    case OP_ANYNL:
3392    case OP_VSPACE:    case OP_VSPACE:
3393    return next == -ESC_V || next == -ESC_d || next == -ESC_w;    return escape == ESC_V || escape == ESC_d || escape == ESC_w;
3394    
3395    case OP_NOT_VSPACE:    case OP_NOT_VSPACE:
3396    return next == -ESC_v || next == -ESC_R;    return escape == ESC_v || escape == ESC_R;
3397    
3398    case OP_WORDCHAR:    case OP_WORDCHAR:
3399    return next == -ESC_W || next == -ESC_s || next == -ESC_h ||    return escape == ESC_W || escape == ESC_s || escape == ESC_h ||
3400           next == -ESC_v || next == -ESC_R;           escape == ESC_v || escape == ESC_R;
3401    
3402    case OP_NOT_WORDCHAR:    case OP_NOT_WORDCHAR:
3403    return next == -ESC_w || next == -ESC_d;    return escape == ESC_w || escape == ESC_d;
3404    
3405    default:    default:
3406    return FALSE;    return FALSE;
# Line 3399  Returns:        the number of < 256 char Line 3434  Returns:        the number of < 256 char
3434    
3435  static int  static int
3436  add_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr, int options,  add_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr, int options,
3437    compile_data *cd, unsigned int start, unsigned int end)    compile_data *cd, pcre_uint32 start, pcre_uint32 end)
3438  {  {
3439  unsigned int c;  pcre_uint32 c;
3440  int n8 = 0;  int n8 = 0;
3441    
3442  /* If caseless matching is required, scan the range and process alternate  /* If caseless matching is required, scan the range and process alternate
# Line 3415  if ((options & PCRE_CASELESS) != 0) Line 3450  if ((options & PCRE_CASELESS) != 0)
3450    if ((options & PCRE_UTF8) != 0)    if ((options & PCRE_UTF8) != 0)
3451      {      {
3452      int rc;      int rc;
3453      unsigned int oc, od;      pcre_uint32 oc, od;
3454    
3455      options &= ~PCRE_CASELESS;   /* Remove for recursive calls */      options &= ~PCRE_CASELESS;   /* Remove for recursive calls */
3456      c = start;      c = start;
# Line 3456  if ((options & PCRE_CASELESS) != 0) Line 3491  if ((options & PCRE_CASELESS) != 0)
3491  length - this means that the same lists of (e.g.) horizontal spaces can be used  length - this means that the same lists of (e.g.) horizontal spaces can be used
3492  in all cases. */  in all cases. */
3493    
3494  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
3495  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3496    if ((options & PCRE_UTF8) == 0)    if ((options & PCRE_UTF8) == 0)
3497  #endif  #endif
3498    if (end > 0xff) end = 0xff;    if (end > 0xff) end = 0xff;
 #endif  
3499    
3500  #ifdef COMPILE_PCRE16  #elif defined COMPILE_PCRE16
3501  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3502    if ((options & PCRE_UTF16) == 0)    if ((options & PCRE_UTF16) == 0)
3503  #endif  #endif
3504    if (end > 0xffff) end = 0xffff;    if (end > 0xffff) end = 0xffff;
3505  #endif  
3506    #endif /* COMPILE_PCRE[8|16] */
3507    
3508  /* If all characters are less than 256, use the bit map. Otherwise use extra  /* If all characters are less than 256, use the bit map. Otherwise use extra
3509  data. */  data. */
# Line 3598  static int Line 3633  static int
3633  add_not_list_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr,  add_not_list_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr,
3634    int options, compile_data *cd, const pcre_uint32 *p)    int options, compile_data *cd, const pcre_uint32 *p)
3635  {  {
3636    BOOL utf = (options & PCRE_UTF8) != 0;
3637  int n8 = 0;  int n8 = 0;
3638  if (p[0] > 0)  if (p[0] > 0)
3639    n8 += add_to_class(classbits, uchardptr, options, cd, 0, p[0] - 1);    n8 += add_to_class(classbits, uchardptr, options, cd, 0, p[0] - 1);
3640  while (p[0] < NOTACHAR)  while (p[0] < NOTACHAR)
3641    {    {
3642    while (p[1] == p[0] + 1) p++;    while (p[1] == p[0] + 1) p++;
3643    n8 += add_to_class(classbits, uchardptr, options, cd, p[0] + 1,    n8 += add_to_class(classbits, uchardptr, options, cd, p[0] + 1,
3644      (p[1] == NOTACHAR)? 0x10ffff : p[1] - 1);      (p[1] == NOTACHAR) ? (utf ? 0x10ffffu : 0xffffffffu) : p[1] - 1);
3645    p++;    p++;
3646    }    }
3647  return n8;  return n8;
# Line 3657  int options = *optionsptr; Line 3693  int options = *optionsptr;
3693  int after_manual_callout = 0;  int after_manual_callout = 0;
3694  int length_prevgroup = 0;  int length_prevgroup = 0;
3695  register int c;  register int c;
3696    int escape;
3697  register pcre_uchar *code = *codeptr;  register pcre_uchar *code = *codeptr;
3698  pcre_uchar *last_code = code;  pcre_uchar *last_code = code;
3699  pcre_uchar *orig_code = code;  pcre_uchar *orig_code = code;
# Line 3676  must not do this for other options (e.g. Line 3713  must not do this for other options (e.g.
3713  dynamically as we process the pattern. */  dynamically as we process the pattern. */
3714    
3715  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3716  /* PCRE_UTF16 has the same value as PCRE_UTF8. */  /* PCRE_UTF[16|32] have the same value as PCRE_UTF8. */
3717  BOOL utf = (options & PCRE_UTF8) != 0;  BOOL utf = (options & PCRE_UTF8) != 0;
3718  pcre_uchar utf_chars[6];  pcre_uchar utf_chars[6];
3719  #else  #else
3720  BOOL utf = FALSE;  BOOL utf = FALSE;
3721  #endif  #endif
3722    
3723  /* Helper variables for OP_XCLASS opcode (for characters > 255). */  /* Helper variables for OP_XCLASS opcode (for characters > 255). We define
3724    class_uchardata always so that it can be passed to add_to_class() always,
3725    though it will not be used in non-UTF 8-bit cases. This avoids having to supply
3726    alternative calls for the different cases. */
3727    
3728    pcre_uchar *class_uchardata;
3729  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3730  BOOL xclass;  BOOL xclass;
 pcre_uchar *class_uchardata;  
3731  pcre_uchar *class_uchardata_base;  pcre_uchar *class_uchardata_base;
3732  #endif  #endif
3733    
# Line 3741  for (;; ptr++) Line 3781  for (;; ptr++)
3781    int terminator;    int terminator;
3782    int mclength;    int mclength;
3783    int tempbracount;    int tempbracount;
3784      int ec; // FIXMEchpe pcre_uint32
3785    pcre_uchar mcbuffer[8];    pcre_uchar mcbuffer[8];
3786    
3787    /* Get next character in the pattern */    /* Get next character in the pattern */
# Line 4058  for (;; ptr++) Line 4099  for (;; ptr++)
4099        {        {
4100        const pcre_uchar *oldptr;        const pcre_uchar *oldptr;
4101    
4102  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4103        if (utf && HAS_EXTRALEN(c))        if (utf && HAS_EXTRALEN(c))
4104          {                           /* Braces are required because the */          {                           /* Braces are required because the */
4105          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */          GETCHARLEN(c, ptr, ptr);    /* macro generates multiple statements */
# Line 4133  for (;; ptr++) Line 4174  for (;; ptr++)
4174          alpha. This relies on the fact that the class table starts with          alpha. This relies on the fact that the class table starts with
4175          alpha, lower, upper as the first 3 entries. */          alpha, lower, upper as the first 3 entries. */
4176    
4177          if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)          if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)
4178            posix_class = 0;            posix_class = 0;
4179    
4180          /* When PCRE_UCP is set, some of the POSIX classes are converted to          /* When PCRE_UCP is set, some of the POSIX classes are converted to
# Line 4210  for (;; ptr++) Line 4251  for (;; ptr++)
4251    
4252        if (c == CHAR_BACKSLASH)        if (c == CHAR_BACKSLASH)
4253          {          {
4254          c = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE);          escape = check_escape(&ptr, &ec, errorcodeptr, cd->bracount, options, TRUE);
4255    
4256          if (*errorcodeptr != 0) goto FAILED;          if (*errorcodeptr != 0) goto FAILED;
4257    
4258          if (-c == ESC_b) c = CHAR_BS;    /* \b is backspace in a class */          if (escape == 0)
4259          else if (-c == ESC_N)            /* \N is not supported in a class */            c = ec;
4260            else if (escape == ESC_b) c = CHAR_BS; /* \b is backspace in a class */
4261            else if (escape == ESC_N)            /* \N is not supported in a class */
4262            {            {
4263            *errorcodeptr = ERR71;            *errorcodeptr = ERR71;
4264            goto FAILED;            goto FAILED;
4265            }            }
4266          else if (-c == ESC_Q)            /* Handle start of quoted string */          else if (escape == ESC_Q)            /* Handle start of quoted string */
4267            {            {
4268            if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)            if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)
4269              {              {
# Line 4228  for (;; ptr++) Line 4272  for (;; ptr++)
4272            else inescq = TRUE;            else inescq = TRUE;
4273            continue;            continue;
4274            }            }
4275          else if (-c == ESC_E) continue;  /* Ignore orphan \E */          else if (escape == ESC_E) continue;  /* Ignore orphan \E */
4276    
4277          if (c < 0)          else
4278            {            {
4279            register const pcre_uint8 *cbits = cd->cbits;            register const pcre_uint8 *cbits = cd->cbits;
4280            /* Every class contains at least two < 256 characters. */            /* Every class contains at least two < 256 characters. */
# Line 4238  for (;; ptr++) Line 4282  for (;; ptr++)
4282            /* Every class contains at least two characters. */            /* Every class contains at least two characters. */
4283            class_one_char += 2;            class_one_char += 2;
4284    
4285            switch (-c)            switch (escape)
4286              {              {
4287  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
4288              case ESC_du:     /* These are the values given for \d etc */              case ESC_du:     /* These are the values given for \d etc */
# Line 4248  for (;; ptr++) Line 4292  for (;; ptr++)
4292              case ESC_su:     /* of the default ASCII testing. */              case ESC_su:     /* of the default ASCII testing. */
4293              case ESC_SU:              case ESC_SU:
4294              nestptr = ptr;              nestptr = ptr;
4295              ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */              ptr = substitutes[escape - ESC_DU] - 1;  /* Just before substitute */
4296              class_has_8bitchar--;                /* Undo! */              class_has_8bitchar--;                /* Undo! */
4297              continue;              continue;
4298  #endif  #endif
# Line 4317  for (;; ptr++) Line 4361  for (;; ptr++)
4361                int pdata;                int pdata;
4362                int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);                int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);
4363                if (ptype < 0) goto FAILED;                if (ptype < 0) goto FAILED;
4364                *class_uchardata++ = ((-c == ESC_p) != negated)?                *class_uchardata++ = ((escape == ESC_p) != negated)?
4365                  XCL_PROP : XCL_NOTPROP;                  XCL_PROP : XCL_NOTPROP;
4366                *class_uchardata++ = ptype;                *class_uchardata++ = ptype;
4367                *class_uchardata++ = pdata;                *class_uchardata++ = pdata;
# Line 4345  for (;; ptr++) Line 4389  for (;; ptr++)
4389          /* Fall through if the escape just defined a single character (c >= 0).          /* Fall through if the escape just defined a single character (c >= 0).
4390          This may be greater than 256. */          This may be greater than 256. */
4391    
4392            escape = 0;
4393    
4394          }   /* End of backslash handling */          }   /* End of backslash handling */
4395    
4396        /* A character may be followed by '-' to form a range. However, Perl does        /* A character may be followed by '-' to form a range. However, Perl does
# Line 4410  for (;; ptr++) Line 4456  for (;; ptr++)
4456    
4457          if (!inescq && d == CHAR_BACKSLASH)          if (!inescq && d == CHAR_BACKSLASH)
4458            {            {
4459            d = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE);            int descape;
4460              descape = check_escape(&ptr, &d, errorcodeptr, cd->bracount, options, TRUE);
4461            if (*errorcodeptr != 0) goto FAILED;            if (*errorcodeptr != 0) goto FAILED;
4462    
4463            /* \b is backspace; any other special means the '-' was literal. */            /* \b is backspace; any other special means the '-' was literal. */
4464    
4465            if (d < 0)            if (descape != 0)
4466              {              {
4467              if (d == -ESC_b) d = CHAR_BS; else              if (descape == ESC_b) d = CHAR_BS; else
4468                {                {
4469                ptr = oldptr;                ptr = oldptr;
4470                goto CLASS_SINGLE_CHARACTER;  /* A few lines below */                goto CLASS_SINGLE_CHARACTER;  /* A few lines below */
# Line 4476  for (;; ptr++) Line 4523  for (;; ptr++)
4523    
4524          if (negate_class)          if (negate_class)
4525            {            {
4526    #ifdef SUPPORT_UCP
4527                // FIXMEchpe pcreuint32?
4528              int d;
4529    #endif
4530            if (firstchar == REQ_UNSET) firstchar = REQ_NONE;            if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
4531            zerofirstchar = firstchar;            zerofirstchar = firstchar;
4532            *code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT;  
4533  #ifdef SUPPORT_UTF            /* For caseless UTF-8 mode when UCP support is available, check
4534            if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)            whether this character has more than one other case. If so, generate
4535              code += PRIV(ord2utf)(c, code);            a special OP_NOTPROP item instead of OP_NOTI. */
4536            else  
4537    #ifdef SUPPORT_UCP
4538              if (utf && (options & PCRE_CASELESS) != 0 &&
4539                  (d = UCD_CASESET(c)) != 0)
4540                {
4541                *code++ = OP_NOTPROP;
4542                *code++ = PT_CLIST;
4543                *code++ = d;
4544                }
4545              else
4546  #endif  #endif
4547              *code++ = c;            /* Char has only one other case, or UCP not available */
4548            goto NOT_CHAR;  
4549                {
4550                *code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT;
4551    #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4552                if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)
4553                  code += PRIV(ord2utf)(c, code);
4554                else
4555    #endif
4556                  *code++ = c;
4557                }
4558    
4559              /* We are finished with this character class */
4560    
4561              goto END_CLASS;
4562            }            }
4563    
4564          /* For a single, positive character, get the value into mcbuffer, and          /* For a single, positive character, get the value into mcbuffer, and
4565          then we can handle this with the normal one-character code. */          then we can handle this with the normal one-character code. */
4566    
4567  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4568          if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)          if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)
4569            mclength = PRIV(ord2utf)(c, mcbuffer);            mclength = PRIV(ord2utf)(c, mcbuffer);
4570          else          else
# Line 4601  for (;; ptr++) Line 4674  for (;; ptr++)
4674        memcpy(code, classbits, 32);        memcpy(code, classbits, 32);
4675        }        }
4676      code += 32 / sizeof(pcre_uchar);      code += 32 / sizeof(pcre_uchar);
4677      NOT_CHAR:  
4678        END_CLASS:
4679      break;      break;
4680    
4681    
# Line 4725  for (;; ptr++) Line 4799  for (;; ptr++)
4799        hold the length of the character in bytes, plus UTF_LENGTH to flag that        hold the length of the character in bytes, plus UTF_LENGTH to flag that
4800        it's a length rather than a small character. */        it's a length rather than a small character. */
4801    
4802  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4803        if (utf && NOT_FIRSTCHAR(code[-1]))        if (utf && NOT_FIRSTCHAR(code[-1]))
4804          {          {
4805          pcre_uchar *lastchar = code - 1;          pcre_uchar *lastchar = code - 1;
# Line 4861  for (;; ptr++) Line 4935  for (;; ptr++)
4935    
4936          if (repeat_max < 0)          if (repeat_max < 0)
4937            {            {
4938  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4939            if (utf && (c & UTF_LENGTH) != 0)            if (utf && (c & UTF_LENGTH) != 0)
4940              {              {
4941              memcpy(code, utf_chars, IN_UCHARS(c & 7));              memcpy(code, utf_chars, IN_UCHARS(c & 7));
# Line 4886  for (;; ptr++) Line 4960  for (;; ptr++)
4960    
4961          else if (repeat_max != repeat_min)          else if (repeat_max != repeat_min)
4962            {            {
4963  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4964            if (utf && (c & UTF_LENGTH) != 0)            if (utf && (c & UTF_LENGTH) != 0)
4965              {              {
4966              memcpy(code, utf_chars, IN_UCHARS(c & 7));              memcpy(code, utf_chars, IN_UCHARS(c & 7));
# Line 4916  for (;; ptr++) Line 4990  for (;; ptr++)
4990    
4991        /* The character or character type itself comes last in all cases. */        /* The character or character type itself comes last in all cases. */
4992    
4993  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
4994        if (utf && (c & UTF_LENGTH) != 0)        if (utf && (c & UTF_LENGTH) != 0)
4995          {          {
4996          memcpy(code, utf_chars, IN_UCHARS(c & 7));          memcpy(code, utf_chars, IN_UCHARS(c & 7));
# Line 5403  for (;; ptr++) Line 5477  for (;; ptr++)
5477        else if (*tempcode == OP_EXACT || *tempcode == OP_NOTEXACT)        else if (*tempcode == OP_EXACT || *tempcode == OP_NOTEXACT)
5478          {          {
5479          tempcode += PRIV(OP_lengths)[*tempcode];          tempcode += PRIV(OP_lengths)[*tempcode];
5480  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
5481          if (utf && HAS_EXTRALEN(tempcode[-1]))          if (utf && HAS_EXTRALEN(tempcode[-1]))
5482            tempcode += GET_EXTRALEN(tempcode[-1]);            tempcode += GET_EXTRALEN(tempcode[-1]);
5483  #endif  #endif
# Line 5501  for (;; ptr++) Line 5575  for (;; ptr++)
5575          arg = ++ptr;          arg = ++ptr;
5576          while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;          while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
5577          arglen = (int)(ptr - arg);          arglen = (int)(ptr - arg);
5578          if (arglen > (int)MAX_MARK)          if ((unsigned int)arglen > MAX_MARK)
5579            {            {
5580            *errorcodeptr = ERR75;            *errorcodeptr = ERR75;
5581            goto FAILED;            goto FAILED;
# Line 6602  for (;; ptr++) Line 6676  for (;; ptr++)
6676      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values
6677      are arranged to be the negation of the corresponding OP_values in the      are arranged to be the negation of the corresponding OP_values in the
6678      default case when PCRE_UCP is not set. For the back references, the values      default case when PCRE_UCP is not set. For the back references, the values
6679      are ESC_REF plus the reference number. Only back references and those types      are negative the reference number. Only back references and those types
6680      that consume a character may be repeated. We can test for values between      that consume a character may be repeated. We can test for values between
6681      ESC_b and ESC_Z for the latter; this may have to change if any new ones are      ESC_b and ESC_Z for the latter; this may have to change if any new ones are
6682      ever created. */      ever created. */
6683    
6684      case CHAR_BACKSLASH:      case CHAR_BACKSLASH:
6685      tempptr = ptr;      tempptr = ptr;
6686      c = check_escape(&ptr, errorcodeptr, cd->bracount, options, FALSE);      escape = check_escape(&ptr, &ec, errorcodeptr, cd->bracount, options, FALSE);
6687    
6688      if (*errorcodeptr != 0) goto FAILED;      if (*errorcodeptr != 0) goto FAILED;
6689    
6690      if (c < 0)      if (escape == 0)
6691          c = ec;
6692        else
6693        {        {
6694        if (-c == ESC_Q)            /* Handle start of quoted string */        if (escape == ESC_Q)            /* Handle start of quoted string */
6695          {          {
6696          if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)          if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)
6697            ptr += 2;               /* avoid empty string */            ptr += 2;               /* avoid empty string */
# Line 6622  for (;; ptr++) Line 6699  for (;; ptr++)
6699          continue;          continue;
6700          }          }
6701    
6702        if (-c == ESC_E) continue;  /* Perl ignores an orphan \E */        if (escape == ESC_E) continue;  /* Perl ignores an orphan \E */
6703    
6704        /* For metasequences that actually match a character, we disable the        /* For metasequences that actually match a character, we disable the
6705        setting of a first character if it hasn't already been set. */        setting of a first character if it hasn't already been set. */
6706    
6707        if (firstchar == REQ_UNSET && -c > ESC_b && -c < ESC_Z)        if (firstchar == REQ_UNSET && escape > ESC_b && escape < ESC_Z)
6708          firstchar = REQ_NONE;          firstchar = REQ_NONE;
6709    
6710        /* 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. */
# Line 6637  for (;; ptr++) Line 6714  for (;; ptr++)
6714    
6715        /* \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'
6716        is a subroutine call by number (Oniguruma syntax). In fact, the value        is a subroutine call by number (Oniguruma syntax). In fact, the value
6717        -ESC_g is returned only for these cases. So we don't need to check for <        ESC_g is returned only for these cases. So we don't need to check for <
6718        or ' if the value is -ESC_g. For the Perl syntax \g{n} the value is        or ' if the value is ESC_g. For the Perl syntax \g{n} the value is
6719        -ESC_REF+n, and for the Perl syntax \g{name} the result is -ESC_k (as        -n, and for the Perl syntax \g{name} the result is ESC_k (as
6720        that is a synonym for a named back reference). */        that is a synonym for a named back reference). */
6721    
6722        if (-c == ESC_g)        if (escape == ESC_g)
6723          {          {
6724          const pcre_uchar *p;          const pcre_uchar *p;
6725          save_hwm = cd->hwm;   /* Normally this is set when '(' is read */          save_hwm = cd->hwm;   /* Normally this is set when '(' is read */
# Line 6698  for (;; ptr++) Line 6775  for (;; ptr++)
6775        /* \k<name> or \k'name' is a back reference by name (Perl syntax).        /* \k<name> or \k'name' is a back reference by name (Perl syntax).
6776        We also support \k{name} (.NET syntax).  */        We also support \k{name} (.NET syntax).  */
6777    
6778        if (-c == ESC_k)        if (escape == ESC_k)
6779          {          {
6780          if ((ptr[1] != CHAR_LESS_THAN_SIGN &&          if ((ptr[1] != CHAR_LESS_THAN_SIGN &&
6781            ptr[1] != CHAR_APOSTROPHE && ptr[1] != CHAR_LEFT_CURLY_BRACKET))            ptr[1] != CHAR_APOSTROPHE && ptr[1] != CHAR_LEFT_CURLY_BRACKET))
# Line 6717  for (;; ptr++) Line 6794  for (;; ptr++)
6794        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
6795        ':' later. */        ':' later. */
6796    
6797        if (-c >= ESC_REF)        if (escape < 0)
6798          {          {
6799          open_capitem *oc;          open_capitem *oc;
6800          recno = -c - ESC_REF;          recno = -escape;
6801    
6802          HANDLE_REFERENCE:    /* Come here from named backref handling */          HANDLE_REFERENCE:    /* Come here from named backref handling */
6803          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;          if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
# Line 6747  for (;; ptr++) Line 6824  for (;; ptr++)
6824        /* So are Unicode property matches, if supported. */        /* So are Unicode property matches, if supported. */
6825    
6826  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6827        else if (-c == ESC_P || -c == ESC_p)        else if (escape == ESC_P || escape == ESC_p)
6828          {          {
6829          BOOL negated;          BOOL negated;
6830          int pdata;          int pdata;
6831          int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);          int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);
6832          if (ptype < 0) goto FAILED;          if (ptype < 0) goto FAILED;
6833          previous = code;          previous = code;
6834          *code++ = ((-c == ESC_p) != negated)? OP_PROP : OP_NOTPROP;          *code++ = ((escape == ESC_p) != negated)? OP_PROP : OP_NOTPROP;
6835          *code++ = ptype;          *code++ = ptype;
6836          *code++ = pdata;          *code++ = pdata;
6837          }          }
# Line 6763  for (;; ptr++) Line 6840  for (;; ptr++)
6840        /* If Unicode properties are not supported, \X, \P, and \p are not        /* If Unicode properties are not supported, \X, \P, and \p are not
6841        allowed. */        allowed. */
6842    
6843        else if (-c == ESC_X || -c == ESC_P || -c == ESC_p)        else if (escape == ESC_X || escape == ESC_P || escape == ESC_p)
6844          {          {
6845          *errorcodeptr = ERR45;          *errorcodeptr = ERR45;
6846          goto FAILED;          goto FAILED;
# Line 6778  for (;; ptr++) Line 6855  for (;; ptr++)
6855    
6856        else        else
6857          {          {
6858          if ((-c == ESC_b || -c == ESC_B) && cd->max_lookbehind == 0)          if ((escape == ESC_b || escape == ESC_B) && cd->max_lookbehind == 0)
6859            cd->max_lookbehind = 1;            cd->max_lookbehind = 1;
6860  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6861          if (-c >= ESC_DU && -c <= ESC_wu)          if (escape >= ESC_DU && escape <= ESC_wu)
6862            {            {
6863            nestptr = ptr + 1;                   /* Where to resume */            nestptr = ptr + 1;                   /* Where to resume */
6864            ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */            ptr = substitutes[escape - ESC_DU] - 1;  /* Just before substitute */
6865            }            }
6866          else          else
6867  #endif  #endif
# Line 6792  for (;; ptr++) Line 6869  for (;; ptr++)
6869          so that it works in DFA mode and in lookbehinds. */          so that it works in DFA mode and in lookbehinds. */
6870    
6871            {            {
6872            previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;            previous = (escape > ESC_b && escape < ESC_Z)? code : NULL;
6873            *code++ = (!utf && c == -ESC_C)? OP_ALLANY : -c;            *code++ = (!utf && escape == ESC_C)? OP_ALLANY : escape;
6874            }            }
6875          }          }
6876        continue;        continue;
# Line 6803  for (;; ptr++) Line 6880  for (;; ptr++)
6880      a value > 127. We set its representation in the length/buffer, and then      a value > 127. We set its representation in the length/buffer, and then
6881      handle it as a data character. */      handle it as a data character. */
6882    
6883  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
6884      if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)      if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR)
6885        mclength = PRIV(ord2utf)(c, mcbuffer);        mclength = PRIV(ord2utf)(c, mcbuffer);
6886      else      else
# Line 6826  for (;; ptr++) Line 6903  for (;; ptr++)
6903      mclength = 1;      mclength = 1;
6904      mcbuffer[0] = c;      mcbuffer[0] = c;
6905    
6906  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && !defined COMPILE_PCRE32
6907      if (utf && HAS_EXTRALEN(c))      if (utf && HAS_EXTRALEN(c))
6908        ACROSSCHAR(TRUE, ptr[1], mcbuffer[mclength++] = *(++ptr));        ACROSSCHAR(TRUE, ptr[1], mcbuffer[mclength++] = *(++ptr));
6909  #endif  #endif
# Line 6836  for (;; ptr++) Line 6913  for (;; ptr++)
6913    
6914      ONE_CHAR:      ONE_CHAR:
6915      previous = code;      previous = code;
6916    
6917        /* For caseless UTF-8 mode when UCP support is available, check whether
6918        this character has more than one other case. If so, generate a special
6919        OP_PROP item instead of OP_CHARI. */
6920    
6921    #ifdef SUPPORT_UCP
6922        if (utf && (options & PCRE_CASELESS) != 0)
6923          {
6924          GETCHAR(c, mcbuffer);
6925          if ((c = UCD_CASESET(c)) != 0)
6926            {
6927            *code++ = OP_PROP;
6928            *code++ = PT_CLIST;
6929            *code++ = c;
6930            if (firstchar == REQ_UNSET) firstchar = zerofirstchar = REQ_NONE;
6931            break;
6932            }
6933          }
6934    #endif
6935    
6936        /* Caseful matches, or not one of the multicase characters. */
6937    
6938      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR;      *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR;
6939      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];
6940    
# Line 7535  Returns:        pointer to compiled data Line 7634  Returns:        pointer to compiled data
7634                  with errorptr and erroroffset set                  with errorptr and erroroffset set
7635  */  */
7636    
7637  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
7638  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION
7639  pcre_compile(const char *pattern, int options, const char **errorptr,  pcre_compile(const char *pattern, int options, const char **errorptr,
7640    int *erroroffset, const unsigned char *tables)    int *erroroffset, const unsigned char *tables)
7641  #else  #elif defined COMPILE_PCRE16
7642  PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION
7643  pcre16_compile(PCRE_SPTR16 pattern, int options, const char **errorptr,  pcre16_compile(PCRE_SPTR16 pattern, int options, const char **errorptr,
7644    int *erroroffset, const unsigned char *tables)    int *erroroffset, const unsigned char *tables)
7645    #elif defined COMPILE_PCRE32
7646    PCRE_EXP_DEFN pcre32 * PCRE_CALL_CONVENTION
7647    pcre32_compile(PCRE_SPTR32 pattern, int options, const char **errorptr,
7648      int *erroroffset, const unsigned char *tables)
7649  #endif  #endif
7650  {  {
7651  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
7652  return pcre_compile2(pattern, options, NULL, errorptr, erroroffset, tables);  return pcre_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
7653  #else  #elif defined COMPILE_PCRE16
7654  return pcre16_compile2(pattern, options, NULL, errorptr, erroroffset, tables);  return pcre16_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
7655    #elif defined COMPILE_PCRE32
7656    return pcre32_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
7657  #endif  #endif
7658  }  }
7659    
7660    
7661  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
7662  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION
7663  pcre_compile2(const char *pattern, int options, int *errorcodeptr,  pcre_compile2(const char *pattern, int options, int *errorcodeptr,
7664    const char **errorptr, int *erroroffset, const unsigned char *tables)    const char **errorptr, int *erroroffset, const unsigned char *tables)
7665  #else  #elif defined COMPILE_PCRE16
7666  PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION  PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION
7667  pcre16_compile2(PCRE_SPTR16 pattern, int options, int *errorcodeptr,  pcre16_compile2(PCRE_SPTR16 pattern, int options, int *errorcodeptr,
7668    const char **errorptr, int *erroroffset, const unsigned char *tables)    const char **errorptr, int *erroroffset, const unsigned char *tables)
7669    #elif defined COMPILE_PCRE32
7670    PCRE_EXP_DEFN pcre32 * PCRE_CALL_CONVENTION
7671    pcre32_compile2(PCRE_SPTR32 pattern, int options, int *errorcodeptr,
7672      const char **errorptr, int *erroroffset, const unsigned char *tables)
7673  #endif  #endif
7674  {  {
7675  REAL_PCRE *re;  REAL_PCRE *re;
# Line 7646  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 7755  while (ptr[skipatstart] == CHAR_LEFT_PAR
7755    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)    if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)
7756      { skipatstart += 8; options |= PCRE_UTF16; continue; }      { skipatstart += 8; options |= PCRE_UTF16; continue; }
7757  #endif  #endif
7758    #ifdef COMPILE_PCRE32
7759      if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0)
7760        { skipatstart += 8; options |= PCRE_UTF32; continue; }
7761    #endif
7762    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UCP_RIGHTPAR, 4) == 0)    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UCP_RIGHTPAR, 4) == 0)
7763      { skipatstart += 6; options |= PCRE_UCP; continue; }      { skipatstart += 6; options |= PCRE_UCP; continue; }
7764    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_NO_START_OPT_RIGHTPAR, 13) == 0)    else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_NO_START_OPT_RIGHTPAR, 13) == 0)
# Line 7674  while (ptr[skipatstart] == CHAR_LEFT_PAR Line 7787  while (ptr[skipatstart] == CHAR_LEFT_PAR
7787    else break;    else break;
7788    }    }
7789    
7790  /* PCRE_UTF16 has the same value as PCRE_UTF8. */  /* PCRE_UTF(16|32) have the same value as PCRE_UTF8. */
7791  utf = (options & PCRE_UTF8) != 0;  utf = (options & PCRE_UTF8) != 0;
7792    
7793  /* Can't support UTF unless PCRE has been compiled to include the code. The  /* Can't support UTF unless PCRE has been compiled to include the code. The
# Line 7686  not used here. */ Line 7799  not used here. */
7799  if (utf && (options & PCRE_NO_UTF8_CHECK) == 0 &&  if (utf && (options & PCRE_NO_UTF8_CHECK) == 0 &&
7800       (errorcode = PRIV(valid_utf)((PCRE_PUCHAR)pattern, -1, erroroffset)) != 0)       (errorcode = PRIV(valid_utf)((PCRE_PUCHAR)pattern, -1, erroroffset)) != 0)
7801    {    {
7802  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
7803    errorcode = ERR44;    errorcode = ERR44;
7804  #else  #elif defined COMPILE_PCRE16
7805    errorcode = ERR74;    errorcode = ERR74;
7806    #elif defined COMPILE_PCRE32
7807      errorcode = ERR77;
7808  #endif  #endif
7809    goto PCRE_EARLY_ERROR_RETURN2;    goto PCRE_EARLY_ERROR_RETURN2;
7810    }    }
# Line 7853  re->name_count = cd->names_found; Line 7968  re->name_count = cd->names_found;
7968  re->ref_count = 0;  re->ref_count = 0;
7969  re->tables = (tables == PRIV(default_tables))? NULL : tables;  re->tables = (tables == PRIV(default_tables))? NULL : tables;
7970  re->nullpad = NULL;  re->nullpad = NULL;
7971    #ifdef COMPILE_PCRE32
7972    re->dummy1 = re->dummy2 = 0;
7973    #endif
7974    
7975  /* The starting points of the name/number translation table and of the code are  /* The starting points of the name/number translation table and of the code are
7976  passed around in the compile data block. The start/end pattern and initial  passed around in the compile data block. The start/end pattern and initial
# Line 8015  if ((re->options & PCRE_ANCHORED) == 0) Line 8133  if ((re->options & PCRE_ANCHORED) == 0)
8133        firstchar = find_firstassertedchar(codestart, FALSE);        firstchar = find_firstassertedchar(codestart, FALSE);
8134      if (firstchar >= 0)   /* Remove caseless flag for non-caseable chars */      if (firstchar >= 0)   /* Remove caseless flag for non-caseable chars */
8135        {        {
8136  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
8137        re->first_char = firstchar & 0xff;        re->first_char = firstchar & 0xff;
8138  #else  #elif defined COMPILE_PCRE16
 #ifdef COMPILE_PCRE16  
8139        re->first_char = firstchar & 0xffff;        re->first_char = firstchar & 0xffff;
8140  #endif  #elif defined COMPILE_PCRE32
8141          re->first_char = firstchar & ~REQ_MASK;
8142  #endif  #endif
8143        if ((firstchar & REQ_CASELESS) != 0)        if ((firstchar & REQ_CASELESS) != 0)
8144          {          {
# Line 8057  bytes. */ Line 8175  bytes. */
8175  if (reqchar >= 0 &&  if (reqchar >= 0 &&
8176       ((re->options & PCRE_ANCHORED) == 0 || (reqchar & REQ_VARY) != 0))       ((re->options & PCRE_ANCHORED) == 0 || (reqchar & REQ_VARY) != 0))
8177    {    {
8178  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
8179    re->req_char = reqchar & 0xff;    re->req_char = reqchar & 0xff;
8180  #else  #elif defined COMPILE_PCRE16
 #ifdef COMPILE_PCRE16  
8181    re->req_char = reqchar & 0xffff;    re->req_char = reqchar & 0xffff;
8182  #endif  #elif defined COMPILE_PCRE32
8183      re->req_char = reqchar & ~REQ_MASK;
8184  #endif  #endif
8185    if ((reqchar & REQ_CASELESS) != 0)    if ((reqchar & REQ_CASELESS) != 0)
8186      {      {
# Line 8114  if ((re->flags & PCRE_REQCHSET) != 0) Line 8232  if ((re->flags & PCRE_REQCHSET) != 0)
8232      else printf("Req char = \\x%02x%s\n", ch, caseless);      else printf("Req char = \\x%02x%s\n", ch, caseless);
8233    }    }
8234    
8235  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
8236  pcre_printint((pcre *)re, stdout, TRUE);  pcre_printint((pcre *)re, stdout, TRUE);
8237  #else  #elif defined COMPILE_PCRE16
8238  pcre16_printint((pcre *)re, stdout, TRUE);  pcre16_printint((pcre *)re, stdout, TRUE);
8239    #elif defined COMPILE_PCRE32
8240    pcre32_printint((pcre *)re, stdout, TRUE);
8241  #endif  #endif
8242    
8243  /* This check is done here in the debugging case so that the code that  /* This check is done here in the debugging case so that the code that
# Line 8133  if (code - codestart > length) Line 8253  if (code - codestart > length)
8253    }    }
8254  #endif   /* PCRE_DEBUG */  #endif   /* PCRE_DEBUG */
8255    
8256  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
8257  return (pcre *)re;  return (pcre *)re;
8258  #else  #elif defined COMPILE_PCRE16
8259  return (pcre16 *)re;  return (pcre16 *)re;
8260    #elif defined COMPILE_PCRE32
8261    return (pcre32 *)re;
8262  #endif  #endif
8263  }  }
8264    

Legend:
Removed from v.1045  
changed lines
  Added in v.1062

  ViewVC Help
Powered by ViewVC 1.1.5