/[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 964 by ph10, Fri May 4 13:03:39 2012 UTC revision 1064 by chpe, Tue Oct 16 15:54:12 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 68  COMPILE_PCREx macro will already be appr Line 68  COMPILE_PCREx macro will already be appr
68    
69  /* Macro for setting individual bits in class bitmaps. */  /* Macro for setting individual bits in class bitmaps. */
70    
71  #define SETBIT(a,b) a[b/8] |= (1 << (b%8))  #define SETBIT(a,b) a[(b)/8] |= (1 << ((b)&7))
72    
73  /* Maximum length value to check against when making sure that the integer that  /* Maximum length value to check against when making sure that the integer that
74  holds the compiled pattern length does not overflow. We make it a bit less than  holds the compiled pattern length does not overflow. We make it a bit less than
# Line 77  to check them every time. */ Line 77  to check them every time. */
77    
78  #define OFLOW_MAX (INT_MAX - 20)  #define OFLOW_MAX (INT_MAX - 20)
79    
80    /* Definitions to allow mutual recursion */
81    
82    static int
83      add_list_to_class(pcre_uint8 *, pcre_uchar **, int, compile_data *,
84        const pcre_uint32 *, unsigned int);
85    
86    static BOOL
87      compile_regex(int, pcre_uchar **, const pcre_uchar **, int *, BOOL, BOOL,
88        int, int, int *, int *, branch_chain *, compile_data *, int *);
89    
90    
91    
92  /*************************************************  /*************************************************
93  *      Code parameters and static tables         *  *      Code parameters and static tables         *
# Line 112  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 490  static const char error_texts[] = Line 502  static const char error_texts[] =
502    "disallowed Unicode code point (>= 0xd800 && <= 0xdfff)\0"    "disallowed Unicode code point (>= 0xd800 && <= 0xdfff)\0"
503    "invalid UTF-16 string\0"    "invalid UTF-16 string\0"
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"
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 630  static const pcre_uint8 ebcdic_chartab[] Line 644  static const pcre_uint8 ebcdic_chartab[]
644  #endif  #endif
645    
646    
 /* Definition to allow mutual recursion */  
   
 static BOOL  
   compile_regex(int, pcre_uchar **, const pcre_uchar **, int *, BOOL, BOOL, int, int,  
     int *, int *, branch_chain *, compile_data *, int *);  
   
647    
648    
649  /*************************************************  /*************************************************
# Line 741  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 784  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 < '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 797  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 822  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 831  else Line 843  else
843            c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));            c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));
844  #endif  #endif
845            }            }
846    
847    #if defined COMPILE_PCRE8
848            if (c > (utf ? 0x10ffff : 0xff))
849    #elif defined COMPILE_PCRE16
850            if (c > (utf ? 0x10ffff : 0xffff))
851    #elif defined COMPILE_PCRE32
852            if (utf && c > 0x10ffff)
853    #endif
854              {
855              *errorcodeptr = ERR76;
856              }
857            else if (utf && c >= 0xd800 && c <= 0xdfff) *errorcodeptr = ERR73;
858          }          }
859        }        }
860      else      else
# Line 857  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 876  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 892  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 916  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 924  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 954  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 1020  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 1038  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 1051  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 1082  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 1141  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 1349  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 1583  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 1686  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 1820  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 1834  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 1876  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 2015  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 2079  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 2102  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 2154  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 2199  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 2222  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 2308  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 2525  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 2541  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 2607  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 2754  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 2858  PUT(previous_callout, 2 + LINK_SIZE, len Line 2889  PUT(previous_callout, 2 + LINK_SIZE, len
2889  *************************************************/  *************************************************/
2890    
2891  /* This function is passed the start and end of a class range, in UTF-8 mode  /* This function is passed the start and end of a class range, in UTF-8 mode
2892  with UCP support. It searches up the characters, looking for internal ranges of  with UCP support. It searches up the characters, looking for ranges of
2893  characters in the "other" case. Each call returns the next one, updating the  characters in the "other" case. Each call returns the next one, updating the
2894  start address.  start address. A character with multiple other cases is returned on its own
2895    with a special return value.
2896    
2897  Arguments:  Arguments:
2898    cptr        points to starting character value; updated    cptr        points to starting character value; updated
# Line 2868  Arguments: Line 2900  Arguments:
2900    ocptr       where to put start of othercase range    ocptr       where to put start of othercase range
2901    odptr       where to put end of othercase range    odptr       where to put end of othercase range
2902    
2903  Yield:        TRUE when range returned; FALSE when no more  Yield:        -1 when no more
2904                   0 when a range is returned
2905                  >0 the CASESET offset for char with multiple other cases
2906                    in this case, ocptr contains the original
2907  */  */
2908    
2909  static BOOL  static int
2910  get_othercase_range(unsigned int *cptr, unsigned int d, unsigned int *ocptr,  get_othercase_range(pcre_uint32 *cptr, pcre_uint32 d, pcre_uint32 *ocptr,
2911    unsigned int *odptr)    pcre_uint32 *odptr)
2912  {  {
2913  unsigned int c, othercase, next;  pcre_uint32 c, othercase, next;
2914    int co;
2915    
2916    /* Find the first character that has an other case. If it has multiple other
2917    cases, return its case offset value. */
2918    
2919  for (c = *cptr; c <= d; c++)  for (c = *cptr; c <= d; c++)
2920    { if ((othercase = UCD_OTHERCASE(c)) != c) break; }    {
2921      if ((co = UCD_CASESET(c)) != 0)
2922        {
2923        *ocptr = c++;   /* Character that has the set */
2924        *cptr = c;      /* Rest of input range */
2925        return co;
2926        }
2927      if ((othercase = UCD_OTHERCASE(c)) != c) break;
2928      }
2929    
2930  if (c > d) return FALSE;  if (c > d) return -1;  /* Reached end of range */
2931    
2932  *ocptr = othercase;  *ocptr = othercase;
2933  next = othercase + 1;  next = othercase + 1;
# Line 2891  for (++c; c <= d; c++) Line 2938  for (++c; c <= d; c++)
2938    next++;    next++;
2939    }    }
2940    
2941  *odptr = next - 1;  *odptr = next - 1;     /* End of othercase range */
2942  *cptr = c;  *cptr = c;             /* Rest of input range */
2943    return 0;
 return TRUE;  
2944  }  }
2945    
2946    
# Line 2916  Returns:       TRUE if auto-possessifyin Line 2962  Returns:       TRUE if auto-possessifyin
2962  */  */
2963    
2964  static BOOL  static BOOL
2965  check_char_prop(int c, int ptype, int pdata, BOOL negated)  check_char_prop(pcre_uint32 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 2956  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 2973  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 2985  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_uint32 c = NOTACHAR;
3052    pcre_uint32 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 3017  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 3059  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 && 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          pcre_uint32 othercase;
3173          if (next < 128) othercase = cd->fcc[next]; else
3174    #ifdef SUPPORT_UCP
3175          othercase = UCD_OTHERCASE(next);
3176  #else  #else
3177    c = *previous;        othercase = NOTACHAR;
3178  #endif  #endif
3179    if (c == next) return TRUE;        return c != othercase;
3180          }
3181        else
3182    #endif  /* SUPPORT_UTF */
3183        return (c != TABLE_GET(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;        pcre_uint32 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(next);
3197  #else  #else
3198      othercase = NOTACHAR;        othercase = NOTACHAR;
3199  #endif  #endif
3200      return (unsigned int)c == othercase;        return 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(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      case 0x09:        HSPACE_CASES:
3232      case 0x20:        return op_code == OP_NOT_HSPACE;
     case 0xa0:  
     case 0x1680:  
     case 0x180e:  
     case 0x2000:  
     case 0x2001:  
     case 0x2002:  
     case 0x2003:  
     case 0x2004:  
     case 0x2005:  
     case 0x2006:  
     case 0x2007:  
     case 0x2008:  
     case 0x2009:  
     case 0x200A:  
     case 0x202f:  
     case 0x205f:  
     case 0x3000:  
     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      case 0x0a:      case OP_VSPACE:
3240      case 0x0b:      case OP_NOT_VSPACE:
3241      case 0x0c:      switch(next)
3242      case 0x0d:        {
3243      case 0x85:        VSPACE_CASES:
3244      case 0x2028:        return op_code == OP_NOT_VSPACE;
3245      case 0x2029:  
3246      return op_code == OP_NOT_VSPACE;        default:
3247      default:        return op_code != OP_NOT_VSPACE;
3248      return op_code != OP_NOT_VSPACE;        }
     }  
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 3219  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 3248  switch(op_code) Line 3294  switch(op_code)
3294      case ESC_H:      case ESC_H:
3295      switch(c)      switch(c)
3296        {        {
3297        case 0x09:        HSPACE_CASES:
3298        case 0x20:        return escape != ESC_h;
3299        case 0xa0:  
       case 0x1680:  
       case 0x180e:  
       case 0x2000:  
       case 0x2001:  
       case 0x2002:  
       case 0x2003:  
       case 0x2004:  
       case 0x2005:  
       case 0x2006:  
       case 0x2007:  
       case 0x2008:  
       case 0x2009:  
       case 0x200A:  
       case 0x202f:  
       case 0x205f:  
       case 0x3000:  
       return -next != ESC_h;  
3300        default:        default:
3301        return -next == ESC_h;        return escape == ESC_h;
3302        }        }
3303    
3304      case ESC_v:      case ESC_v:
3305      case ESC_V:      case ESC_V:
3306      switch(c)      switch(c)
3307        {        {
3308        case 0x0a:        VSPACE_CASES:
3309        case 0x0b:        return escape != ESC_v;
3310        case 0x0c:  
       case 0x0d:  
       case 0x85:  
       case 0x2028:  
       case 0x2029:  
       return -next != ESC_v;  
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 3301  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 3329  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 3344  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 3388  switch(op_code) Line 3412  switch(op_code)
3412    
3413    
3414  /*************************************************  /*************************************************
3415    *        Add a character or range to a class     *
3416    *************************************************/
3417    
3418    /* This function packages up the logic of adding a character or range of
3419    characters to a class. The character values in the arguments will be within the
3420    valid values for the current mode (8-bit, 16-bit, UTF, etc). This function is
3421    mutually recursive with the function immediately below.
3422    
3423    Arguments:
3424      classbits     the bit map for characters < 256
3425      uchardptr     points to the pointer for extra data
3426      options       the options word
3427      cd            contains pointers to tables etc.
3428      start         start of range character
3429      end           end of range character
3430    
3431    Returns:        the number of < 256 characters added
3432                    the pointer to extra data is updated
3433    */
3434    
3435    static int
3436    add_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr, int options,
3437      compile_data *cd, pcre_uint32 start, pcre_uint32 end)
3438    {
3439    pcre_uint32 c;
3440    int n8 = 0;
3441    
3442    /* If caseless matching is required, scan the range and process alternate
3443    cases. In Unicode, there are 8-bit characters that have alternate cases that
3444    are greater than 255 and vice-versa. Sometimes we can just extend the original
3445    range. */
3446    
3447    if ((options & PCRE_CASELESS) != 0)
3448      {
3449    #ifdef SUPPORT_UCP
3450      if ((options & PCRE_UTF8) != 0)
3451        {
3452        int rc;
3453        pcre_uint32 oc, od;
3454    
3455        options &= ~PCRE_CASELESS;   /* Remove for recursive calls */
3456        c = start;
3457    
3458        while ((rc = get_othercase_range(&c, end, &oc, &od)) >= 0)
3459          {
3460          /* Handle a single character that has more than one other case. */
3461    
3462          if (rc > 0) n8 += add_list_to_class(classbits, uchardptr, options, cd,
3463            PRIV(ucd_caseless_sets) + rc, oc);
3464    
3465          /* Do nothing if the other case range is within the original range. */
3466    
3467          else if (oc >= start && od <= end) continue;
3468    
3469          /* Extend the original range if there is overlap, noting that if oc < c, we
3470          can't have od > end because a subrange is always shorter than the basic
3471          range. Otherwise, use a recursive call to add the additional range. */
3472    
3473          else if (oc < start && od >= start - 1) start = oc; /* Extend downwards */
3474          else if (od > end && oc <= end + 1) end = od;       /* Extend upwards */
3475          else n8 += add_to_class(classbits, uchardptr, options, cd, oc, od);
3476          }
3477        }
3478      else
3479    #endif  /* SUPPORT_UCP */
3480    
3481      /* Not UTF-mode, or no UCP */
3482    
3483      for (c = start; c <= end && c < 256; c++)
3484        {
3485        SETBIT(classbits, cd->fcc[c]);
3486        n8++;
3487        }
3488      }
3489    
3490    /* Now handle the original range. Adjust the final value according to the bit
3491    length - this means that the same lists of (e.g.) horizontal spaces can be used
3492    in all cases. */
3493    
3494    #if defined COMPILE_PCRE8
3495    #ifdef SUPPORT_UTF
3496      if ((options & PCRE_UTF8) == 0)
3497    #endif
3498      if (end > 0xff) end = 0xff;
3499    
3500    #elif defined COMPILE_PCRE16
3501    #ifdef SUPPORT_UTF
3502      if ((options & PCRE_UTF16) == 0)
3503    #endif
3504      if (end > 0xffff) end = 0xffff;
3505    
3506    #endif /* COMPILE_PCRE[8|16] */
3507    
3508    /* If all characters are less than 256, use the bit map. Otherwise use extra
3509    data. */
3510    
3511    if (end < 0x100)
3512      {
3513      for (c = start; c <= end; c++)
3514        {
3515        n8++;
3516        SETBIT(classbits, c);
3517        }
3518      }
3519    
3520    else
3521      {
3522      pcre_uchar *uchardata = *uchardptr;
3523    
3524    #ifdef SUPPORT_UTF
3525      if ((options & PCRE_UTF8) != 0)  /* All UTFs use the same flag bit */
3526        {
3527        if (start < end)
3528          {
3529          *uchardata++ = XCL_RANGE;
3530          uchardata += PRIV(ord2utf)(start, uchardata);
3531          uchardata += PRIV(ord2utf)(end, uchardata);
3532          }
3533        else if (start == end)
3534          {
3535          *uchardata++ = XCL_SINGLE;
3536          uchardata += PRIV(ord2utf)(start, uchardata);
3537          }
3538        }
3539      else
3540    #endif  /* SUPPORT_UTF */
3541    
3542      /* Without UTF support, character values are constrained by the bit length,
3543      and can only be > 256 for 16-bit and 32-bit libraries. */
3544    
3545    #ifdef COMPILE_PCRE8
3546        {}
3547    #else
3548      if (start < end)
3549        {
3550        *uchardata++ = XCL_RANGE;
3551        *uchardata++ = start;
3552        *uchardata++ = end;
3553        }
3554      else if (start == end)
3555        {
3556        *uchardata++ = XCL_SINGLE;
3557        *uchardata++ = start;
3558        }
3559    #endif
3560    
3561      *uchardptr = uchardata;   /* Updata extra data pointer */
3562      }
3563    
3564    return n8;    /* Number of 8-bit characters */
3565    }
3566    
3567    
3568    
3569    
3570    /*************************************************
3571    *        Add a list of characters to a class     *
3572    *************************************************/
3573    
3574    /* This function is used for adding a list of case-equivalent characters to a
3575    class, and also for adding a list of horizontal or vertical whitespace. If the
3576    list is in order (which it should be), ranges of characters are detected and
3577    handled appropriately. This function is mutually recursive with the function
3578    above.
3579    
3580    Arguments:
3581      classbits     the bit map for characters < 256
3582      uchardptr     points to the pointer for extra data
3583      options       the options word
3584      cd            contains pointers to tables etc.
3585      p             points to row of 32-bit values, terminated by NOTACHAR
3586      except        character to omit; this is used when adding lists of
3587                      case-equivalent characters to avoid including the one we
3588                      already know about
3589    
3590    Returns:        the number of < 256 characters added
3591                    the pointer to extra data is updated
3592    */
3593    
3594    static int
3595    add_list_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr, int options,
3596      compile_data *cd, const pcre_uint32 *p, unsigned int except)
3597    {
3598    int n8 = 0;
3599    while (p[0] < NOTACHAR)
3600      {
3601      int n = 0;
3602      if (p[0] != except)
3603        {
3604        while(p[n+1] == p[0] + n + 1) n++;
3605        n8 += add_to_class(classbits, uchardptr, options, cd, p[0], p[n]);
3606        }
3607      p += n + 1;
3608      }
3609    return n8;
3610    }
3611    
3612    
3613    
3614    /*************************************************
3615    *    Add characters not in a list to a class     *
3616    *************************************************/
3617    
3618    /* This function is used for adding the complement of a list of horizontal or
3619    vertical whitespace to a class. The list must be in order.
3620    
3621    Arguments:
3622      classbits     the bit map for characters < 256
3623      uchardptr     points to the pointer for extra data
3624      options       the options word
3625      cd            contains pointers to tables etc.
3626      p             points to row of 32-bit values, terminated by NOTACHAR
3627    
3628    Returns:        the number of < 256 characters added
3629                    the pointer to extra data is updated
3630    */
3631    
3632    static int
3633    add_not_list_to_class(pcre_uint8 *classbits, pcre_uchar **uchardptr,
3634      int options, compile_data *cd, const pcre_uint32 *p)
3635    {
3636    BOOL utf = (options & PCRE_UTF8) != 0;
3637    int n8 = 0;
3638    if (p[0] > 0)
3639      n8 += add_to_class(classbits, uchardptr, options, cd, 0, p[0] - 1);
3640    while (p[0] < NOTACHAR)
3641      {
3642      while (p[1] == p[0] + 1) p++;
3643      n8 += add_to_class(classbits, uchardptr, options, cd, p[0] + 1,
3644        (p[1] == NOTACHAR) ? (utf ? 0x10ffffu : 0xffffffffu) : p[1] - 1);
3645      p++;
3646      }
3647    return n8;
3648    }
3649    
3650    
3651    
3652    /*************************************************
3653  *           Compile one branch                   *  *           Compile one branch                   *
3654  *************************************************/  *************************************************/
3655    
# Line 3430  pcre_int32 req_caseopt, reqvary, tempreq Line 3692  pcre_int32 req_caseopt, reqvary, tempreq
3692  int options = *optionsptr;               /* May change dynamically */  int options = *optionsptr;               /* May change dynamically */
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 pcre_uint32 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 3450  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 3505  for (;; ptr++) Line 3771  for (;; ptr++)
3771    BOOL is_recurse;    BOOL is_recurse;
3772    BOOL reset_bracount;    BOOL reset_bracount;
3773    int class_has_8bitchar;    int class_has_8bitchar;
3774    int class_single_char;    int class_one_char;
3775    int newoptions;    int newoptions;
3776    int recno;    int recno;
3777    int refsign;    int refsign;
# Line 3515  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 3803  for (;; ptr++) Line 4070  for (;; ptr++)
4070    
4071      should_flip_negation = FALSE;      should_flip_negation = FALSE;
4072    
4073      /* For optimization purposes, we track some properties of the class.      /* For optimization purposes, we track some properties of the class:
4074      class_has_8bitchar will be non-zero, if the class contains at least one      class_has_8bitchar will be non-zero if the class contains at least one <
4075      < 256 character. class_single_char will be 1 if the class contains only      256 character; class_one_char will be 1 if the class contains just one
4076      a single character. */      character. */
4077    
4078      class_has_8bitchar = 0;      class_has_8bitchar = 0;
4079      class_single_char = 0;      class_one_char = 0;
4080    
4081      /* Initialize the 32-char bit map to all zeros. We build the map in a      /* Initialize the 32-char bit map to all zeros. We build the map in a
4082      temporary bit of memory, in case the class contains only 1 character (less      temporary bit of memory, in case the class contains fewer than two
4083      than 256), because in that case the compiled code doesn't use the bit map.      8-bit characters because in that case the compiled code doesn't use the bit
4084      */      map. */
4085    
4086      memset(classbits, 0, 32 * sizeof(pcre_uint8));      memset(classbits, 0, 32 * sizeof(pcre_uint8));
4087    
4088  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
4089      xclass = FALSE;                           /* No chars >= 256 */      xclass = FALSE;
4090      class_uchardata = code + LINK_SIZE + 2;   /* For UTF-8 items */      class_uchardata = code + LINK_SIZE + 2;   /* For XCLASS items */
4091      class_uchardata_base = class_uchardata;   /* For resetting in pass 1 */      class_uchardata_base = class_uchardata;   /* Save the start */
4092  #endif  #endif
4093    
4094      /* 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 3832  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 3843  for (;; ptr++) Line 4110  for (;; ptr++)
4110        /* In the pre-compile phase, accumulate the length of any extra        /* In the pre-compile phase, accumulate the length of any extra
4111        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
4112        contain a zillion > 255 characters no longer overwrite the work space        contain a zillion > 255 characters no longer overwrite the work space
4113        (which is on the stack). */        (which is on the stack). We have to remember that there was XCLASS data,
4114          however. */
4115    
4116        if (lengthptr != NULL)        if (lengthptr != NULL && class_uchardata > class_uchardata_base)
4117          {          {
4118            xclass = TRUE;
4119          *lengthptr += class_uchardata - class_uchardata_base;          *lengthptr += class_uchardata - class_uchardata_base;
4120          class_uchardata = class_uchardata_base;          class_uchardata = class_uchardata_base;
4121          }          }
# Line 3905  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 3948  for (;; ptr++) Line 4217  for (;; ptr++)
4217              for (c = 0; c < 32; c++) pbits[c] &= ~cbits[c + taboffset];              for (c = 0; c < 32; c++) pbits[c] &= ~cbits[c + taboffset];
4218            }            }
4219    
4220          /* Not see if we need to remove any special characters. An option          /* Now see if we need to remove any special characters. An option
4221          value of 1 removes vertical space and 2 removes underscore. */          value of 1 removes vertical space and 2 removes underscore. */
4222    
4223          if (tabopt < 0) tabopt = -tabopt;          if (tabopt < 0) tabopt = -tabopt;
# Line 3964  for (;; ptr++) Line 4233  for (;; ptr++)
4233            for (c = 0; c < 32; c++) classbits[c] |= pbits[c];            for (c = 0; c < 32; c++) classbits[c] |= pbits[c];
4234    
4235          ptr = tempptr + 1;          ptr = tempptr + 1;
4236          /* Every class contains at least one < 256 characters. */          /* Every class contains at least one < 256 character. */
4237          class_has_8bitchar = 1;          class_has_8bitchar = 1;
4238          /* Every class contains at least two characters. */          /* Every class contains at least two characters. */
4239          class_single_char = 2;          class_one_char = 2;
4240          continue;    /* End of POSIX syntax handling */          continue;    /* End of POSIX syntax handling */
4241          }          }
4242    
# Line 3975  for (;; ptr++) Line 4244  for (;; ptr++)
4244        of the specials, which just set a flag. The sequence \b is a special        of the specials, which just set a flag. The sequence \b is a special
4245        case. Inside a class (and only there) it is treated as backspace. We        case. Inside a class (and only there) it is treated as backspace. We
4246        assume that other escapes have more than one character in them, so        assume that other escapes have more than one character in them, so
4247        speculatively set both class_has_8bitchar and class_single_char bigger        speculatively set both class_has_8bitchar and class_one_char bigger
4248        than one. Unrecognized escapes fall through and are either treated        than one. Unrecognized escapes fall through and are either treated
4249        as literal characters (by default), or are faulted if        as literal characters (by default), or are faulted if
4250        PCRE_EXTRA is set. */        PCRE_EXTRA is set. */
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 4000  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. */
4281            class_has_8bitchar++;            class_has_8bitchar++;
4282            /* Every class contains at least two characters. */            /* Every class contains at least two characters. */
4283            class_single_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 4020  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 4044  for (;; ptr++) Line 4316  for (;; ptr++)
4316    
4317              /* Perl 5.004 onwards omits VT from \s, but we must preserve it              /* Perl 5.004 onwards omits VT from \s, but we must preserve it
4318              if it was previously set by something earlier in the character              if it was previously set by something earlier in the character
4319              class. */              class. Luckily, the value of CHAR_VT is 0x0b in both ASCII and
4320                EBCDIC, so we lazily just adjust the appropriate bit. */
4321    
4322              case ESC_s:              case ESC_s:
4323              classbits[0] |= cbits[cbit_space];              classbits[0] |= cbits[cbit_space];
# Line 4057  for (;; ptr++) Line 4330  for (;; ptr++)
4330              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space];              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space];
4331              classbits[1] |= 0x08;    /* Perl 5.004 onwards omits VT from \s */              classbits[1] |= 0x08;    /* Perl 5.004 onwards omits VT from \s */
4332              continue;              continue;
4333    
4334                /* The rest apply in both UCP and non-UCP cases. */
4335    
4336              case ESC_h:              case ESC_h:
4337              SETBIT(classbits, 0x09); /* VT */              (void)add_list_to_class(classbits, &class_uchardata, options, cd,
4338              SETBIT(classbits, 0x20); /* SPACE */                PRIV(hspace_list), NOTACHAR);
             SETBIT(classbits, 0xa0); /* NSBP */  
 #ifndef COMPILE_PCRE8  
             xclass = TRUE;  
             *class_uchardata++ = XCL_SINGLE;  
             *class_uchardata++ = 0x1680;  
             *class_uchardata++ = XCL_SINGLE;  
             *class_uchardata++ = 0x180e;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x2000;  
             *class_uchardata++ = 0x200a;  
             *class_uchardata++ = XCL_SINGLE;  
             *class_uchardata++ = 0x202f;  
             *class_uchardata++ = XCL_SINGLE;  
             *class_uchardata++ = 0x205f;  
             *class_uchardata++ = XCL_SINGLE;  
             *class_uchardata++ = 0x3000;  
 #elif defined SUPPORT_UTF  
             if (utf)  
               {  
               xclass = TRUE;  
               *class_uchardata++ = XCL_SINGLE;  
               class_uchardata += PRIV(ord2utf)(0x1680, class_uchardata);  
               *class_uchardata++ = XCL_SINGLE;  
               class_uchardata += PRIV(ord2utf)(0x180e, class_uchardata);  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x2000, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x200a, class_uchardata);  
               *class_uchardata++ = XCL_SINGLE;  
               class_uchardata += PRIV(ord2utf)(0x202f, class_uchardata);  
               *class_uchardata++ = XCL_SINGLE;  
               class_uchardata += PRIV(ord2utf)(0x205f, class_uchardata);  
               *class_uchardata++ = XCL_SINGLE;  
               class_uchardata += PRIV(ord2utf)(0x3000, class_uchardata);  
               }  
 #endif  
4339              continue;              continue;
4340    
4341              case ESC_H:              case ESC_H:
4342              for (c = 0; c < 32; c++)              (void)add_not_list_to_class(classbits, &class_uchardata, options,
4343                {                cd, PRIV(hspace_list));
               int x = 0xff;  
               switch (c)  
                 {  
                 case 0x09/8: x ^= 1 << (0x09%8); break;  
                 case 0x20/8: x ^= 1 << (0x20%8); break;  
                 case 0xa0/8: x ^= 1 << (0xa0%8); break;  
                 default: break;  
                 }  
               classbits[c] |= x;  
               }  
 #ifndef COMPILE_PCRE8  
             xclass = TRUE;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x0100;  
             *class_uchardata++ = 0x167f;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x1681;  
             *class_uchardata++ = 0x180d;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x180f;  
             *class_uchardata++ = 0x1fff;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x200b;  
             *class_uchardata++ = 0x202e;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x2030;  
             *class_uchardata++ = 0x205e;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x2060;  
             *class_uchardata++ = 0x2fff;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x3001;  
 #ifdef SUPPORT_UTF  
             if (utf)  
               class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata);  
             else  
 #endif  
               *class_uchardata++ = 0xffff;  
 #elif defined SUPPORT_UTF  
             if (utf)  
               {  
               xclass = TRUE;  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x0100, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x167f, class_uchardata);  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x1681, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x180d, class_uchardata);  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x180f, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x1fff, class_uchardata);  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x200b, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x202e, class_uchardata);  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x2030, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x205e, class_uchardata);  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x2060, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x2fff, class_uchardata);  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x3001, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata);  
               }  
 #endif  
4344              continue;              continue;
4345    
4346              case ESC_v:              case ESC_v:
4347              SETBIT(classbits, 0x0a); /* LF */              (void)add_list_to_class(classbits, &class_uchardata, options, cd,
4348              SETBIT(classbits, 0x0b); /* VT */                PRIV(vspace_list), NOTACHAR);
             SETBIT(classbits, 0x0c); /* FF */  
             SETBIT(classbits, 0x0d); /* CR */  
             SETBIT(classbits, 0x85); /* NEL */  
 #ifndef COMPILE_PCRE8  
             xclass = TRUE;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x2028;  
             *class_uchardata++ = 0x2029;  
 #elif defined SUPPORT_UTF  
             if (utf)  
               {  
               xclass = TRUE;  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x2028, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x2029, class_uchardata);  
               }  
 #endif  
4349              continue;              continue;
4350    
4351              case ESC_V:              case ESC_V:
4352              for (c = 0; c < 32; c++)              (void)add_not_list_to_class(classbits, &class_uchardata, options,
4353                {                cd, PRIV(vspace_list));
               int x = 0xff;  
               switch (c)  
                 {  
                 case 0x0a/8: x ^= 1 << (0x0a%8);  
                              x ^= 1 << (0x0b%8);  
                              x ^= 1 << (0x0c%8);  
                              x ^= 1 << (0x0d%8);  
                              break;  
                 case 0x85/8: x ^= 1 << (0x85%8); break;  
                 default: break;  
                 }  
               classbits[c] |= x;  
               }  
   
 #ifndef COMPILE_PCRE8  
             xclass = TRUE;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x0100;  
             *class_uchardata++ = 0x2027;  
             *class_uchardata++ = XCL_RANGE;  
             *class_uchardata++ = 0x202a;  
 #ifdef SUPPORT_UTF  
             if (utf)  
               class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata);  
             else  
 #endif  
               *class_uchardata++ = 0xffff;  
 #elif defined SUPPORT_UTF  
             if (utf)  
               {  
               xclass = TRUE;  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x0100, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x2027, class_uchardata);  
               *class_uchardata++ = XCL_RANGE;  
               class_uchardata += PRIV(ord2utf)(0x202a, class_uchardata);  
               class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata);  
               }  
 #endif  
4354              continue;              continue;
4355    
4356  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 4242  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                xclass = TRUE;                *class_uchardata++ = ((escape == ESC_p) != negated)?
               *class_uchardata++ = ((-c == 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 4262  for (;; ptr++) Line 4380  for (;; ptr++)
4380                goto FAILED;                goto FAILED;
4381                }                }
4382              class_has_8bitchar--;    /* Undo the speculative increase. */              class_has_8bitchar--;    /* Undo the speculative increase. */
4383              class_single_char -= 2;  /* Undo the speculative increase. */              class_one_char -= 2;     /* Undo the speculative increase. */
4384              c = *ptr;                /* Get the final character and fall through */              c = *ptr;                /* Get the final character and fall through */
4385              break;              break;
4386              }              }
4387            }            }
4388    
4389          /* Fall through if we have a single character (c >= 0). This may be          /* Fall through if the escape just defined a single character (c >= 0).
4390          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 single character may be followed by '-' to form a range. However,        /* A character may be followed by '-' to form a range. However, Perl does
4397        Perl does not permit ']' to be the end of the range. A '-' character        not permit ']' to be the end of the range. A '-' character at the end is
4398        at the end is treated as a literal. Perl ignores orphaned \E sequences        treated as a literal. Perl ignores orphaned \E sequences entirely. The
4399        entirely. The code for handling \Q and \E is messy. */        code for handling \Q and \E is messy. */
4400    
4401        CHECK_RANGE:        CHECK_RANGE:
4402        while (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)        while (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)
# Line 4284  for (;; ptr++) Line 4404  for (;; ptr++)
4404          inescq = FALSE;          inescq = FALSE;
4405          ptr += 2;          ptr += 2;
4406          }          }
   
4407        oldptr = ptr;        oldptr = ptr;
4408    
4409        /* Remember \r or \n */        /* Remember if \r or \n were explicitly used */
4410    
4411        if (c == CHAR_CR || c == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF;        if (c == CHAR_CR || c == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF;
4412    
# Line 4310  for (;; ptr++) Line 4429  for (;; ptr++)
4429            inescq = TRUE;            inescq = TRUE;
4430            break;            break;
4431            }            }
4432    
4433            /* Minus (hyphen) at the end of a class is treated as a literal, so put
4434            back the pointer and jump to handle the character that preceded it. */
4435    
4436          if (*ptr == 0 || (!inescq && *ptr == CHAR_RIGHT_SQUARE_BRACKET))          if (*ptr == 0 || (!inescq && *ptr == CHAR_RIGHT_SQUARE_BRACKET))
4437            {            {
4438            ptr = oldptr;            ptr = oldptr;
4439            goto LONE_SINGLE_CHARACTER;            goto CLASS_SINGLE_CHARACTER;
4440            }            }
4441    
4442            /* Otherwise, we have a potential range; pick up the next character */
4443    
4444  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
4445          if (utf)          if (utf)
# Line 4332  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 LONE_SINGLE_CHARACTER;  /* A few lines below */                goto CLASS_SINGLE_CHARACTER;  /* A few lines below */
4471                }                }
4472              }              }
4473            }            }
4474    
4475          /* Check that the two values are in the correct order. Optimize          /* Check that the two values are in the correct order. Optimize
4476          one-character ranges */          one-character ranges. */
4477    
4478          if (d < c)          if (d < c)
4479            {            {
4480            *errorcodeptr = ERR8;            *errorcodeptr = ERR8;
4481            goto FAILED;            goto FAILED;
4482            }            }
4483            if (d == c) goto CLASS_SINGLE_CHARACTER;  /* A few lines below */
4484    
4485          if (d == c) goto LONE_SINGLE_CHARACTER;  /* A few lines below */          /* We have found a character range, so single character optimizations
4486            cannot be done anymore. Any value greater than 1 indicates that there
4487            is more than one character. */
4488    
4489            class_one_char = 2;
4490    
4491          /* Remember \r or \n */          /* Remember an explicit \r or \n, and add the range to the class. */
4492    
4493          if (d == CHAR_CR || d == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF;          if (d == CHAR_CR || d == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF;
4494    
4495          /* Since we found a character range, single character optimizations          class_has_8bitchar +=
4496          cannot be done anymore. */            add_to_class(classbits, &class_uchardata, options, cd, c, d);
4497          class_single_char = 2;  
   
         /* In UTF-8 mode, if the upper limit is > 255, or > 127 for caseless  
         matching, we have to use an XCLASS with extra data items. Caseless  
         matching for characters > 127 is available only if UCP support is  
         available. */  
   
 #if defined SUPPORT_UTF && !(defined COMPILE_PCRE8)  
         if ((d > 255) || (utf && ((options & PCRE_CASELESS) != 0 && d > 127)))  
 #elif defined  SUPPORT_UTF  
         if (utf && (d > 255 || ((options & PCRE_CASELESS) != 0 && d > 127)))  
 #elif !(defined COMPILE_PCRE8)  
         if (d > 255)  
 #endif  
 #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)  
           {  
           xclass = TRUE;  
   
           /* With UCP support, we can find the other case equivalents of  
           the relevant characters. There may be several ranges. Optimize how  
           they fit with the basic range. */  
   
 #ifdef SUPPORT_UCP  
 #ifndef COMPILE_PCRE8  
           if (utf && (options & PCRE_CASELESS) != 0)  
 #else  
           if ((options & PCRE_CASELESS) != 0)  
 #endif  
             {  
             unsigned int occ, ocd;  
             unsigned int cc = c;  
             unsigned int origd = d;  
             while (get_othercase_range(&cc, origd, &occ, &ocd))  
               {  
               if (occ >= (unsigned int)c &&  
                   ocd <= (unsigned int)d)  
                 continue;                          /* Skip embedded ranges */  
   
               if (occ < (unsigned int)c  &&  
                   ocd >= (unsigned int)c - 1)      /* Extend the basic range */  
                 {                                  /* if there is overlap,   */  
                 c = occ;                           /* noting that if occ < c */  
                 continue;                          /* we can't have ocd > d  */  
                 }                                  /* because a subrange is  */  
               if (ocd > (unsigned int)d &&  
                   occ <= (unsigned int)d + 1)      /* always shorter than    */  
                 {                                  /* the basic range.       */  
                 d = ocd;  
                 continue;  
                 }  
   
               if (occ == ocd)  
                 {  
                 *class_uchardata++ = XCL_SINGLE;  
                 }  
               else  
                 {  
                 *class_uchardata++ = XCL_RANGE;  
                 class_uchardata += PRIV(ord2utf)(occ, class_uchardata);  
                 }  
               class_uchardata += PRIV(ord2utf)(ocd, class_uchardata);  
               }  
             }  
 #endif  /* SUPPORT_UCP */  
   
           /* Now record the original range, possibly modified for UCP caseless  
           overlapping ranges. */  
   
           *class_uchardata++ = XCL_RANGE;  
 #ifdef SUPPORT_UTF  
 #ifndef COMPILE_PCRE8  
           if (utf)  
             {  
             class_uchardata += PRIV(ord2utf)(c, class_uchardata);  
             class_uchardata += PRIV(ord2utf)(d, class_uchardata);  
             }  
           else  
             {  
             *class_uchardata++ = c;  
             *class_uchardata++ = d;  
             }  
 #else  
           class_uchardata += PRIV(ord2utf)(c, class_uchardata);  
           class_uchardata += PRIV(ord2utf)(d, class_uchardata);  
 #endif  
 #else /* SUPPORT_UTF */  
           *class_uchardata++ = c;  
           *class_uchardata++ = d;  
 #endif /* SUPPORT_UTF */  
   
           /* With UCP support, we are done. Without UCP support, there is no  
           caseless matching for UTF characters > 127; we can use the bit map  
           for the smaller ones. As for 16 bit characters without UTF, we  
           can still use  */  
   
 #ifdef SUPPORT_UCP  
 #ifndef COMPILE_PCRE8  
           if (utf)  
 #endif  
             continue;    /* With next character in the class */  
 #endif  /* SUPPORT_UCP */  
   
 #if defined SUPPORT_UTF && !defined(SUPPORT_UCP) && !(defined COMPILE_PCRE8)  
           if (utf)  
             {  
             if ((options & PCRE_CASELESS) == 0 || c > 127) continue;  
             /* Adjust upper limit and fall through to set up the map */  
             d = 127;  
             }  
           else  
             {  
             if (c > 255) continue;  
             /* Adjust upper limit and fall through to set up the map */  
             d = 255;  
             }  
 #elif defined SUPPORT_UTF && !defined(SUPPORT_UCP)  
           if ((options & PCRE_CASELESS) == 0 || c > 127) continue;  
           /* Adjust upper limit and fall through to set up the map */  
           d = 127;  
 #else  
           if (c > 255) continue;  
           /* Adjust upper limit and fall through to set up the map */  
           d = 255;  
 #endif  /* SUPPORT_UTF && !SUPPORT_UCP && !COMPILE_PCRE8 */  
           }  
 #endif  /* SUPPORT_UTF || !COMPILE_PCRE8 */  
   
         /* We use the bit map for 8 bit mode, or when the characters fall  
         partially or entirely to [0-255] ([0-127] for UCP) ranges. */  
   
         class_has_8bitchar = 1;  
   
         /* We can save a bit of time by skipping this in the pre-compile. */  
   
         if (lengthptr == NULL) for (; c <= d; c++)  
           {  
           classbits[c/8] |= (1 << (c&7));  
           if ((options & PCRE_CASELESS) != 0)  
             {  
             int uc = cd->fcc[c]; /* flip case */  
             classbits[uc/8] |= (1 << (uc&7));  
             }  
           }  
   
4498          continue;   /* Go get the next char in the class */          continue;   /* Go get the next char in the class */
4499          }          }
4500    
4501        /* Handle a lone single character - we can get here for a normal        /* Handle a single character - we can get here for a normal non-escape
4502        non-escape char, or after \ that introduces a single character or for an        char, or after \ that introduces a single character or for an apparent
4503        apparent range that isn't. */        range that isn't. Only the value 1 matters for class_one_char, so don't
4504          increase it if it is already 2 or more ... just in case there's a class
4505        LONE_SINGLE_CHARACTER:        with a zillion characters in it. */
4506    
4507          CLASS_SINGLE_CHARACTER:
4508          if (class_one_char < 2) class_one_char++;
4509    
4510          /* If class_one_char is 1, we have the first single character in the
4511          class, and there have been no prior ranges, or XCLASS items generated by
4512          escapes. If this is the final character in the class, we can optimize by
4513          turning the item into a 1-character OP_CHAR[I] if it's positive, or
4514          OP_NOT[I] if it's negative. In the positive case, it can cause firstchar
4515          to be set. Otherwise, there can be no first char if this item is first,
4516          whatever repeat count may follow. In the case of reqchar, save the
4517          previous value for reinstating. */
4518    
4519        /* Only the value of 1 matters for class_single_char. */        if (class_one_char == 1 && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)
   
       if (class_single_char < 2) class_single_char++;  
   
       /* If class_charcount is 1, we saw precisely one character. As long as  
       there was no use of \p or \P, in other words, no use of any XCLASS  
       features, we can optimize.  
   
       The optimization throws away the bit map. We turn the item into a  
       1-character OP_CHAR[I] if it's positive, or OP_NOT[I] if it's negative.  
       In the positive case, it can cause firstchar to be set. Otherwise, there  
       can be no first char if this item is first, whatever repeat count may  
       follow. In the case of reqchar, save the previous value for reinstating. */  
   
       if (class_single_char == 1 && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)  
4520          {          {
4521          ptr++;          ptr++;
4522          zeroreqchar = reqchar;          zeroreqchar = reqchar;
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 4564  for (;; ptr++) Line 4575  for (;; ptr++)
4575            }            }
4576          goto ONE_CHAR;          goto ONE_CHAR;
4577          }       /* End of 1-char optimization */          }       /* End of 1-char optimization */
4578    
4579        /* Handle a character that cannot go in the bit map. */        /* There is more than one character in the class, or an XCLASS item
4580          has been generated. Add this character to the class. */
4581  #if defined SUPPORT_UTF && !(defined COMPILE_PCRE8)  
4582        if ((c > 255) || (utf && ((options & PCRE_CASELESS) != 0 && c > 127)))        class_has_8bitchar +=
4583  #elif defined SUPPORT_UTF          add_to_class(classbits, &class_uchardata, options, cd, c, c);
       if (utf && (c > 255 || ((options & PCRE_CASELESS) != 0 && c > 127)))  
 #elif !(defined COMPILE_PCRE8)  
       if (c > 255)  
 #endif  
   
 #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)  
         {  
         xclass = TRUE;  
         *class_uchardata++ = XCL_SINGLE;  
 #ifdef SUPPORT_UTF  
 #ifndef COMPILE_PCRE8  
         /* In non 8 bit mode, we can get here even if we are not in UTF mode. */  
         if (!utf)  
           *class_uchardata++ = c;  
         else  
 #endif  
           class_uchardata += PRIV(ord2utf)(c, class_uchardata);  
 #else /* SUPPORT_UTF */  
         *class_uchardata++ = c;  
 #endif /* SUPPORT_UTF */  
   
 #ifdef SUPPORT_UCP  
 #ifdef COMPILE_PCRE8  
         if ((options & PCRE_CASELESS) != 0)  
 #else  
         /* In non 8 bit mode, we can get here even if we are not in UTF mode. */  
         if (utf && (options & PCRE_CASELESS) != 0)  
 #endif  
           {  
           unsigned int othercase;  
           if ((int)(othercase = UCD_OTHERCASE(c)) != c)  
             {  
             *class_uchardata++ = XCL_SINGLE;  
             class_uchardata += PRIV(ord2utf)(othercase, class_uchardata);  
             }  
           }  
 #endif  /* SUPPORT_UCP */  
   
         }  
       else  
 #endif  /* SUPPORT_UTF || COMPILE_PCRE16 */  
   
       /* Handle a single-byte character */  
         {  
         class_has_8bitchar = 1;  
         classbits[c/8] |= (1 << (c&7));  
         if ((options & PCRE_CASELESS) != 0)  
           {  
           c = cd->fcc[c]; /* flip case */  
           classbits[c/8] |= (1 << (c&7));  
           }  
         }  
4584        }        }
4585    
4586      /* Loop until ']' reached. This "while" is the end of the "do" far above.      /* Loop until ']' reached. This "while" is the end of the "do" far above.
# Line 4641  for (;; ptr++) Line 4600  for (;; ptr++)
4600        goto FAILED;        goto FAILED;
4601        }        }
4602    
4603        /* We will need an XCLASS if data has been placed in class_uchardata. In
4604        the second phase this is a sufficient test. However, in the pre-compile
4605        phase, class_uchardata gets emptied to prevent workspace overflow, so it
4606        only if the very last character in the class needs XCLASS will it contain
4607        anything at this point. For this reason, xclass gets set TRUE above when
4608        uchar_classdata is emptied, and that's why this code is the way it is here
4609        instead of just doing a test on class_uchardata below. */
4610    
4611    #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
4612        if (class_uchardata > class_uchardata_base) xclass = TRUE;
4613    #endif
4614    
4615      /* If this is the first thing in the branch, there can be no first char      /* If this is the first thing in the branch, there can be no first char
4616      setting, whatever the repeat count. Any reqchar setting must remain      setting, whatever the repeat count. Any reqchar setting must remain
4617      unchanged after any kind of repeat. */      unchanged after any kind of repeat. */
# Line 4703  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 4813  for (;; ptr++) Line 4785  for (;; ptr++)
4785      if (*previous == OP_CHAR || *previous == OP_CHARI      if (*previous == OP_CHAR || *previous == OP_CHARI
4786          || *previous == OP_NOT || *previous == OP_NOTI)          || *previous == OP_NOT || *previous == OP_NOTI)
4787        {        {
4788        switch (*previous)        switch (*previous)
4789          {          {
4790          default: /* Make compiler happy. */          default: /* Make compiler happy. */
4791          case OP_CHAR:  op_type = OP_STAR - OP_STAR; break;          case OP_CHAR:  op_type = OP_STAR - OP_STAR; break;
# Line 4827  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 4963  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 4988  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 5018  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 5505  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 5593  for (;; ptr++) Line 5565  for (;; ptr++)
5565        ptr++;        ptr++;
5566        while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_letter) != 0) ptr++;        while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_letter) != 0) ptr++;
5567        namelen = (int)(ptr - name);        namelen = (int)(ptr - name);
5568    
5569        /* It appears that Perl allows any characters whatsoever, other than        /* It appears that Perl allows any characters whatsoever, other than
5570        a closing parenthesis, to appear in arguments, so we no longer insist on        a closing parenthesis, to appear in arguments, so we no longer insist on
5571        letters, digits, and underscores. */        letters, digits, and underscores. */
# Line 5603  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;
5582            }            }
5583          }          }
5584    
5585        if (*ptr != CHAR_RIGHT_PARENTHESIS)        if (*ptr != CHAR_RIGHT_PARENTHESIS)
# Line 5623  for (;; ptr++) Line 5595  for (;; ptr++)
5595          if (namelen == verbs[i].len &&          if (namelen == verbs[i].len &&
5596              STRNCMP_UC_C8(name, vn, namelen) == 0)              STRNCMP_UC_C8(name, vn, namelen) == 0)
5597            {            {
5598              int setverb;
5599    
5600            /* Check for open captures before ACCEPT and convert it to            /* Check for open captures before ACCEPT and convert it to
5601            ASSERT_ACCEPT if in an assertion. */            ASSERT_ACCEPT if in an assertion. */
5602    
# Line 5640  for (;; ptr++) Line 5614  for (;; ptr++)
5614                *code++ = OP_CLOSE;                *code++ = OP_CLOSE;
5615                PUT2INC(code, 0, oc->number);                PUT2INC(code, 0, oc->number);
5616                }                }
5617              *code++ = (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;              setverb = *code++ =
5618                  (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;
5619    
5620              /* Do not set firstchar after *ACCEPT */              /* Do not set firstchar after *ACCEPT */
5621              if (firstchar == REQ_UNSET) firstchar = REQ_NONE;              if (firstchar == REQ_UNSET) firstchar = REQ_NONE;
# Line 5655  for (;; ptr++) Line 5630  for (;; ptr++)
5630                *errorcodeptr = ERR66;                *errorcodeptr = ERR66;
5631                goto FAILED;                goto FAILED;
5632                }                }
5633              *code = verbs[i].op;              setverb = *code++ = verbs[i].op;
             if (*code++ == OP_THEN) cd->external_flags |= PCRE_HASTHEN;  
5634              }              }
5635    
5636            else            else
# Line 5666  for (;; ptr++) Line 5640  for (;; ptr++)
5640                *errorcodeptr = ERR59;                *errorcodeptr = ERR59;
5641                goto FAILED;                goto FAILED;
5642                }                }
5643              *code = verbs[i].op_arg;              setverb = *code++ = verbs[i].op_arg;
             if (*code++ == OP_THEN_ARG) cd->external_flags |= PCRE_HASTHEN;  
5644              *code++ = arglen;              *code++ = arglen;
5645              memcpy(code, arg, IN_UCHARS(arglen));              memcpy(code, arg, IN_UCHARS(arglen));
5646              code += arglen;              code += arglen;
5647              *code++ = 0;              *code++ = 0;
5648              }              }
5649    
5650              switch (setverb)
5651                {
5652                case OP_THEN:
5653                case OP_THEN_ARG:
5654                cd->external_flags |= PCRE_HASTHEN;
5655                break;
5656    
5657                case OP_PRUNE:
5658                case OP_PRUNE_ARG:
5659                case OP_SKIP:
5660                case OP_SKIP_ARG:
5661                cd->had_pruneorskip = TRUE;
5662                break;
5663                }
5664    
5665            break;  /* Found verb, exit loop */            break;  /* Found verb, exit loop */
5666            }            }
5667    
# Line 6688  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 6708  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 6723  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 6784  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 6803  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 6833  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 6849  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 6859  for (;; ptr++) Line 6850  for (;; ptr++)
6850        /* For the rest (including \X when Unicode properties are supported), we        /* For the rest (including \X when Unicode properties are supported), we
6851        can obtain the OP value by negating the escape value in the default        can obtain the OP value by negating the escape value in the default
6852        situation when PCRE_UCP is not set. When it *is* set, we substitute        situation when PCRE_UCP is not set. When it *is* set, we substitute
6853        Unicode property tests. Note that \b and \B do a one-character        Unicode property tests. Note that \b and \B do a one-character
6854        lookbehind. */        lookbehind. */
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 6878  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 6889  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 6912  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 6922  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 7173  for (;;) Line 7186  for (;;)
7186          *ptrptr = ptr;          *ptrptr = ptr;
7187          return FALSE;          return FALSE;
7188          }          }
7189        else        else
7190          {          {
7191          if (fixed_length > cd->max_lookbehind)          if (fixed_length > cd->max_lookbehind)
7192            cd->max_lookbehind = fixed_length;            cd->max_lookbehind = fixed_length;
7193          PUT(reverse_count, 0, fixed_length);          PUT(reverse_count, 0, fixed_length);
7194          }          }
7195        }        }
7196      }      }
# Line 7310  and the highest back reference was great Line 7323  and the highest back reference was great
7323  However, by keeping a bitmap of the first 31 back references, we can catch some  However, by keeping a bitmap of the first 31 back references, we can catch some
7324  of the more common cases more precisely.  of the more common cases more precisely.
7325    
7326    ... A second exception is when the .* appears inside an atomic group, because
7327    this prevents the number of characters it matches from being adjusted.
7328    
7329  Arguments:  Arguments:
7330    code           points to start of expression (the bracket)    code           points to start of expression (the bracket)
7331    bracket_map    a bitmap of which brackets we are inside while testing; this    bracket_map    a bitmap of which brackets we are inside while testing; this
7332                    handles up to substring 31; after that we just have to take                    handles up to substring 31; after that we just have to take
7333                    the less precise approach                    the less precise approach
7334    backref_map    the back reference bitmap    cd             points to the compile data block
7335      atomcount      atomic group level
7336    
7337  Returns:     TRUE or FALSE  Returns:     TRUE or FALSE
7338  */  */
7339    
7340  static BOOL  static BOOL
7341  is_anchored(register const pcre_uchar *code, unsigned int bracket_map,  is_anchored(register const pcre_uchar *code, unsigned int bracket_map,
7342    unsigned int backref_map)    compile_data *cd, int atomcount)
7343  {  {
7344  do {  do {
7345     const pcre_uchar *scode = first_significant_code(     const pcre_uchar *scode = first_significant_code(
# Line 7334  do { Line 7351  do {
7351     if (op == OP_BRA  || op == OP_BRAPOS ||     if (op == OP_BRA  || op == OP_BRAPOS ||
7352         op == OP_SBRA || op == OP_SBRAPOS)         op == OP_SBRA || op == OP_SBRAPOS)
7353       {       {
7354       if (!is_anchored(scode, bracket_map, backref_map)) return FALSE;       if (!is_anchored(scode, bracket_map, cd, atomcount)) return FALSE;
7355       }       }
7356    
7357     /* Capturing brackets */     /* Capturing brackets */
# Line 7344  do { Line 7361  do {
7361       {       {
7362       int n = GET2(scode, 1+LINK_SIZE);       int n = GET2(scode, 1+LINK_SIZE);
7363       int new_map = bracket_map | ((n < 32)? (1 << n) : 1);       int new_map = bracket_map | ((n < 32)? (1 << n) : 1);
7364       if (!is_anchored(scode, new_map, backref_map)) return FALSE;       if (!is_anchored(scode, new_map, cd, atomcount)) return FALSE;
7365       }       }
7366    
7367     /* Other brackets */     /* Positive forward assertions and conditions */
7368    
7369     else if (op == OP_ASSERT || op == OP_ONCE || op == OP_ONCE_NC ||     else if (op == OP_ASSERT || op == OP_COND)
             op == OP_COND)  
7370       {       {
7371       if (!is_anchored(scode, bracket_map, backref_map)) return FALSE;       if (!is_anchored(scode, bracket_map, cd, atomcount)) return FALSE;
7372         }
7373    
7374       /* Atomic groups */
7375    
7376       else if (op == OP_ONCE || op == OP_ONCE_NC)
7377         {
7378         if (!is_anchored(scode, bracket_map, cd, atomcount + 1))
7379           return FALSE;
7380       }       }
7381    
7382     /* .* is not anchored unless DOTALL is set (which generates OP_ALLANY) and     /* .* is not anchored unless DOTALL is set (which generates OP_ALLANY) and
7383     it isn't in brackets that are or may be referenced. */     it isn't in brackets that are or may be referenced or inside an atomic
7384       group. */
7385    
7386     else if ((op == OP_TYPESTAR || op == OP_TYPEMINSTAR ||     else if ((op == OP_TYPESTAR || op == OP_TYPEMINSTAR ||
7387               op == OP_TYPEPOSSTAR))               op == OP_TYPEPOSSTAR))
7388       {       {
7389       if (scode[1] != OP_ALLANY || (bracket_map & backref_map) != 0)       if (scode[1] != OP_ALLANY || (bracket_map & cd->backref_map) != 0 ||
7390             atomcount > 0 || cd->had_pruneorskip)
7391         return FALSE;         return FALSE;
7392       }       }
7393    
7394     /* Check for explicit anchoring */     /* Check for explicit anchoring */
7395    
7396     else if (op != OP_SOD && op != OP_SOM && op != OP_CIRC) return FALSE;     else if (op != OP_SOD && op != OP_SOM && op != OP_CIRC) return FALSE;
7397    
7398     code += GET(code, 1);     code += GET(code, 1);
7399     }     }
7400  while (*code == OP_ALT);   /* Loop for each alternative */  while (*code == OP_ALT);   /* Loop for each alternative */
# Line 7385  return TRUE; Line 7412  return TRUE;
7412  matching and for non-DOTALL patterns that start with .* (which must start at  matching and for non-DOTALL patterns that start with .* (which must start at
7413  the beginning or after \n). As in the case of is_anchored() (see above), we  the beginning or after \n). As in the case of is_anchored() (see above), we
7414  have to take account of back references to capturing brackets that contain .*  have to take account of back references to capturing brackets that contain .*
7415  because in that case we can't make the assumption.  because in that case we can't make the assumption. Also, the appearance of .*
7416    inside atomic brackets or in a pattern that contains *PRUNE or *SKIP does not
7417    count, because once again the assumption no longer holds.
7418    
7419  Arguments:  Arguments:
7420    code           points to start of expression (the bracket)    code           points to start of expression (the bracket)
7421    bracket_map    a bitmap of which brackets we are inside while testing; this    bracket_map    a bitmap of which brackets we are inside while testing; this
7422                    handles up to substring 31; after that we just have to take                    handles up to substring 31; after that we just have to take
7423                    the less precise approach                    the less precise approach
7424    backref_map    the back reference bitmap    cd             points to the compile data
7425      atomcount      atomic group level
7426    
7427  Returns:         TRUE or FALSE  Returns:         TRUE or FALSE
7428  */  */
7429    
7430  static BOOL  static BOOL
7431  is_startline(const pcre_uchar *code, unsigned int bracket_map,  is_startline(const pcre_uchar *code, unsigned int bracket_map,
7432    unsigned int backref_map)    compile_data *cd, int atomcount)
7433  {  {
7434  do {  do {
7435     const pcre_uchar *scode = first_significant_code(     const pcre_uchar *scode = first_significant_code(
# Line 7425  do { Line 7455  do {
7455         return FALSE;         return FALSE;
7456    
7457         default:     /* Assertion */         default:     /* Assertion */
7458         if (!is_startline(scode, bracket_map, backref_map)) return FALSE;         if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE;
7459         do scode += GET(scode, 1); while (*scode == OP_ALT);         do scode += GET(scode, 1); while (*scode == OP_ALT);
7460         scode += 1 + LINK_SIZE;         scode += 1 + LINK_SIZE;
7461         break;         break;
# Line 7439  do { Line 7469  do {
7469     if (op == OP_BRA  || op == OP_BRAPOS ||     if (op == OP_BRA  || op == OP_BRAPOS ||
7470         op == OP_SBRA || op == OP_SBRAPOS)         op == OP_SBRA || op == OP_SBRAPOS)
7471       {       {
7472       if (!is_startline(scode, bracket_map, backref_map)) return FALSE;       if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE;
7473       }       }
7474    
7475     /* Capturing brackets */     /* Capturing brackets */
# Line 7449  do { Line 7479  do {
7479       {       {
7480       int n = GET2(scode, 1+LINK_SIZE);       int n = GET2(scode, 1+LINK_SIZE);
7481       int new_map = bracket_map | ((n < 32)? (1 << n) : 1);       int new_map = bracket_map | ((n < 32)? (1 << n) : 1);
7482       if (!is_startline(scode, new_map, backref_map)) return FALSE;       if (!is_startline(scode, new_map, cd, atomcount)) return FALSE;
7483       }       }
7484    
7485     /* Other brackets */     /* Positive forward assertions */
7486    
7487     else if (op == OP_ASSERT || op == OP_ONCE || op == OP_ONCE_NC)     else if (op == OP_ASSERT)
7488       {       {
7489       if (!is_startline(scode, bracket_map, backref_map)) return FALSE;       if (!is_startline(scode, bracket_map, cd, atomcount)) return FALSE;
7490       }       }
7491    
7492       /* Atomic brackets */
7493    
7494     /* .* means "start at start or after \n" if it isn't in brackets that     else if (op == OP_ONCE || op == OP_ONCE_NC)
7495     may be referenced. */       {
7496         if (!is_startline(scode, bracket_map, cd, atomcount + 1)) return FALSE;
7497         }
7498    
7499       /* .* means "start at start or after \n" if it isn't in atomic brackets or
7500       brackets that may be referenced, as long as the pattern does not contain
7501       *PRUNE or *SKIP, because these break the feature. Consider, for example,
7502       /.*?a(*PRUNE)b/ with the subject "aab", which matches "ab", i.e. not at the
7503       start of a line. */
7504    
7505     else if (op == OP_TYPESTAR || op == OP_TYPEMINSTAR || op == OP_TYPEPOSSTAR)     else if (op == OP_TYPESTAR || op == OP_TYPEMINSTAR || op == OP_TYPEPOSSTAR)
7506       {       {
7507       if (scode[1] != OP_ANY || (bracket_map & backref_map) != 0) return FALSE;       if (scode[1] != OP_ANY || (bracket_map & cd->backref_map) != 0 ||
7508             atomcount > 0 || cd->had_pruneorskip)
7509           return FALSE;
7510       }       }
7511    
7512     /* Check for explicit circumflex */     /* Check for explicit circumflex; anything else gives a FALSE result. Note
7513       in particular that this includes atomic brackets OP_ONCE and OP_ONCE_NC
7514       because the number of characters matched by .* cannot be adjusted inside
7515       them. */
7516    
7517     else if (op != OP_CIRC && op != OP_CIRCM) return FALSE;     else if (op != OP_CIRC && op != OP_CIRCM) return FALSE;
7518    
# Line 7589  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 7700  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 7728  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 7740  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 7907  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 7926  cd->start_code = codestart; Line 7990  cd->start_code = codestart;
7990  cd->hwm = (pcre_uchar *)(cd->start_workspace);  cd->hwm = (pcre_uchar *)(cd->start_workspace);
7991  cd->req_varyopt = 0;  cd->req_varyopt = 0;
7992  cd->had_accept = FALSE;  cd->had_accept = FALSE;
7993    cd->had_pruneorskip = FALSE;
7994  cd->check_lookbehind = FALSE;  cd->check_lookbehind = FALSE;
7995  cd->open_caps = NULL;  cd->open_caps = NULL;
7996    
# Line 8049  if (errorcode != 0) Line 8114  if (errorcode != 0)
8114    }    }
8115    
8116  /* If the anchored option was not passed, set the flag if we can determine that  /* If the anchored option was not passed, set the flag if we can determine that
8117  the pattern is anchored by virtue of ^ characters or \A or anything else (such  the pattern is anchored by virtue of ^ characters or \A or anything else, such
8118  as starting with .* when DOTALL is set).  as starting with non-atomic .* when DOTALL is set and there are no occurrences
8119    of *PRUNE or *SKIP.
8120    
8121  Otherwise, if we know what the first byte has to be, save it, because that  Otherwise, if we know what the first byte has to be, save it, because that
8122  speeds up unanchored matches no end. If not, see if we can set the  speeds up unanchored matches no end. If not, see if we can set the
8123  PCRE_STARTLINE flag. This is helpful for multiline matches when all branches  PCRE_STARTLINE flag. This is helpful for multiline matches when all branches
8124  start with ^. and also when all branches start with .* for non-DOTALL matches.  start with ^. and also when all branches start with non-atomic .* for
8125  */  non-DOTALL matches when *PRUNE and SKIP are not present. */
8126    
8127  if ((re->options & PCRE_ANCHORED) == 0)  if ((re->options & PCRE_ANCHORED) == 0)
8128    {    {
8129    if (is_anchored(codestart, 0, cd->backref_map))    if (is_anchored(codestart, 0, cd, 0)) re->options |= PCRE_ANCHORED;
     re->options |= PCRE_ANCHORED;  
8130    else    else
8131      {      {
8132      if (firstchar < 0)      if (firstchar < 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 8098  if ((re->options & PCRE_ANCHORED) == 0) Line 8163  if ((re->options & PCRE_ANCHORED) == 0)
8163    
8164        re->flags |= PCRE_FIRSTSET;        re->flags |= PCRE_FIRSTSET;
8165        }        }
8166      else if (is_startline(codestart, 0, cd->backref_map))  
8167        re->flags |= PCRE_STARTLINE;      else if (is_startline(codestart, 0, cd, 0)) re->flags |= PCRE_STARTLINE;
8168      }      }
8169    }    }
8170    
# Line 8110  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 8167  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 8186  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.964  
changed lines
  Added in v.1064

  ViewVC Help
Powered by ViewVC 1.1.5