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

Diff of /code/trunk/pcre_compile.c

Parent Directory | Revision Log | View Patch Patch

-revision 503 by ph10,
Sun Mar  7 17:35:52 2010 UTC
+revision 835 by ph10,
Wed Dec 28 16:10:09 2011 UTC
 Line 6
  and semantics are as close as possible to those of the Perl 5 language.
                         Written by Philip Hazel
-            Copyright (c) 1997-2010 University of Cambridge
+            Copyright (c) 1997-2011 University of Cambridge
  -----------------------------------------------------------------------------
  Redistribution and use in source and binary forms, with or without
 Line 92 
 is 4 there is plenty of room. */
  #define COMPILE_WORK_SIZE (4096)
+ /* The overrun tests check for a slightly smaller size so that they detect the
+ overrun before it actually does run off the end of the data block. */
+ #define WORK_SIZE_CHECK (COMPILE_WORK_SIZE - 100)
  /* Table for handling escaped characters in the range '0'-'z'. Positive returns
  are simple data values; negative values are for special things like \d and so
-Line 119 
 static const short int escapes[] = {
+Line 124 
 static const short int escapes[] = {
       -ESC_H,                  0,
 ,                       -ESC_K,
 ,                       0,
-,                       0,
+      -ESC_N,                  0,
       -ESC_P,                  -ESC_Q,
       -ESC_R,                  -ESC_S,
 ,                       0,
-Line 166 
 static const short int escapes[] = {
+Line 171 
 static const short int escapes[] = {
  /*  B8 */     0,     0,      0,       0,      0,   ']',    '=',    '-',
  /*  C0 */   '{',-ESC_A, -ESC_B,  -ESC_C, -ESC_D,-ESC_E,      0, -ESC_G,
  /*  C8 */-ESC_H,     0,      0,       0,      0,     0,      0,      0,
- /*  D0 */   '}',     0, -ESC_K,       0,      0,     0,      0, -ESC_P,
+ /*  D0 */   '}',     0, -ESC_K,       0,      0,-ESC_N,      0, -ESC_P,
  /*  D8 */-ESC_Q,-ESC_R,      0,       0,      0,     0,      0,      0,
  /*  E0 */  '\\',     0, -ESC_S,       0,      0,-ESC_V, -ESC_W, -ESC_X,
  /*  E8 */     0,-ESC_Z,      0,       0,      0,     0,      0,      0,
-Line 183 
 string is built from string macros so th
+Line 188 
 string is built from string macros so th
  platforms. */
  typedef struct verbitem {
-   int   len;
+   int   len;                 /* Length of verb name */
-   int   op;
+   int   op;                  /* Op when no arg, or -1 if arg mandatory */
+   int   op_arg;              /* Op when arg present, or -1 if not allowed */
  } verbitem;
  static const char verbnames[] =
+   "\0"                       /* Empty name is a shorthand for MARK */
+   STRING_MARK0
    STRING_ACCEPT0
    STRING_COMMIT0
    STRING_F0
-Line 197 
 static const char verbnames[] =
+Line 205 
 static const char verbnames[] =
    STRING_THEN;
  static const verbitem verbs[] = {
-   { 6, OP_ACCEPT },
+   { 0, -1,        OP_MARK },
-   { 6, OP_COMMIT },
+   { 4, -1,        OP_MARK },
-   { 1, OP_FAIL },
+   { 6, OP_ACCEPT, -1 },
-   { 4, OP_FAIL },
+   { 6, OP_COMMIT, -1 },
-   { 5, OP_PRUNE },
+   { 1, OP_FAIL,   -1 },
-   { 4, OP_SKIP  },
+   { 4, OP_FAIL,   -1 },
-   { 4, OP_THEN  }
+   { 5, OP_PRUNE,  OP_PRUNE_ARG },
+   { 4, OP_SKIP,   OP_SKIP_ARG  },
+   { 4, OP_THEN,   OP_THEN_ARG  }
  };
  static const int verbcount = sizeof(verbs)/sizeof(verbitem);
-Line 251 
 static const int posix_class_maps[] = {
+Line 261 
 static const int posix_class_maps[] = {
    cbit_xdigit,-1,          0              /* xdigit */
  };
+ /* Table of substitutes for \d etc when PCRE_UCP is set. The POSIX class
+ substitutes must be in the order of the names, defined above, and there are
+ both positive and negative cases. NULL means no substitute. */
+ #ifdef SUPPORT_UCP
+ static const uschar *substitutes[] = {
+   (uschar *)"\\P{Nd}",    /* \D */
+   (uschar *)"\\p{Nd}",    /* \d */
+   (uschar *)"\\P{Xsp}",   /* \S */       /* NOTE: Xsp is Perl space */
+   (uschar *)"\\p{Xsp}",   /* \s */
+   (uschar *)"\\P{Xwd}",   /* \W */
+   (uschar *)"\\p{Xwd}"    /* \w */
+ };
+ static const uschar *posix_substitutes[] = {
+   (uschar *)"\\p{L}",     /* alpha */
+   (uschar *)"\\p{Ll}",    /* lower */
+   (uschar *)"\\p{Lu}",    /* upper */
+   (uschar *)"\\p{Xan}",   /* alnum */
+   NULL,                   /* ascii */
+   (uschar *)"\\h",        /* blank */
+   NULL,                   /* cntrl */
+   (uschar *)"\\p{Nd}",    /* digit */
+   NULL,                   /* graph */
+   NULL,                   /* print */
+   NULL,                   /* punct */
+   (uschar *)"\\p{Xps}",   /* space */    /* NOTE: Xps is POSIX space */
+   (uschar *)"\\p{Xwd}",   /* word */
+   NULL,                   /* xdigit */
+   /* Negated cases */
+   (uschar *)"\\P{L}",     /* ^alpha */
+   (uschar *)"\\P{Ll}",    /* ^lower */
+   (uschar *)"\\P{Lu}",    /* ^upper */
+   (uschar *)"\\P{Xan}",   /* ^alnum */
+   NULL,                   /* ^ascii */
+   (uschar *)"\\H",        /* ^blank */
+   NULL,                   /* ^cntrl */
+   (uschar *)"\\P{Nd}",    /* ^digit */
+   NULL,                   /* ^graph */
+   NULL,                   /* ^print */
+   NULL,                   /* ^punct */
+   (uschar *)"\\P{Xps}",   /* ^space */   /* NOTE: Xps is POSIX space */
+   (uschar *)"\\P{Xwd}",   /* ^word */
+   NULL                    /* ^xdigit */
+ };
+ #define POSIX_SUBSIZE (sizeof(posix_substitutes)/sizeof(uschar *))
+ #endif
  #define STRING(a)  # a
  #define XSTRING(s) STRING(s)
-Line 263 
 the number of relocations needed when a
+Line 320 
 the number of relocations needed when a
  it is now one long string. We cannot use a table of offsets, because the
  lengths of inserts such as XSTRING(MAX_NAME_SIZE) are not known. Instead, we
  simply count through to the one we want - this isn't a performance issue
  because these strings are used only when there is a compilation error.
  Each substring ends with \0 to insert a null character. This includes the final
  substring, so that the whole string ends with \0\0, which can be detected when
  counting through. */
  static const char error_texts[] =
-Line 314 
 static const char error_texts[] =
+Line 371 
 static const char error_texts[] =
    /* 35 */
    "invalid condition (?(0)\0"
    "\\C not allowed in lookbehind assertion\0"
-   "PCRE does not support \\L, \\l, \\N, \\U, or \\u\0"
+   "PCRE does not support \\L, \\l, \\N{name}, \\U, or \\u\0"
    "number after (?C is > 255\0"
    "closing ) for (?C expected\0"
    /* 40 */
-Line 336 
 static const char error_texts[] =
+Line 393 
 static const char error_texts[] =
    "internal error: previously-checked referenced subpattern not found\0"
    "DEFINE group contains more than one branch\0"
    /* 55 */
-   "repeating a DEFINE group is not allowed\0"
+   "repeating a DEFINE group is not allowed\0"  /** DEAD **/
    "inconsistent NEWLINE options\0"
    "\\g is not followed by a braced, angle-bracketed, or quoted name/number or by a plain number\0"
    "a numbered reference must not be zero\0"
-   "(*VERB) with an argument is not supported\0"
+   "an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)\0"
    /* 60 */
    "(*VERB) not recognized\0"
    "number is too big\0"
-Line 348 
 static const char error_texts[] =
+Line 405 
 static const char error_texts[] =
    "digit expected after (?+\0"
    "] is an invalid data character in JavaScript compatibility mode\0"
    /* 65 */
-   "different names for subpatterns of the same number are not allowed\0";
+   "different names for subpatterns of the same number are not allowed\0"
+   "(*MARK) must have an argument\0"
+   "this version of PCRE is not compiled with PCRE_UCP support\0"
+   "\\c must be followed by an ASCII character\0"
+   "\\k is not followed by a braced, angle-bracketed, or quoted name\0"
+   /* 70 */
+   "internal error: unknown opcode in find_fixedlength()\0"
+   ;
  /* Table to identify digits and hex digits. This is used when compiling
  patterns. Note that the tables in chartables are dependent on the locale, and
-Line 484 
 static const unsigned char ebcdic_charta
+Line 548 
 static const unsigned char ebcdic_charta
  /* Definition to allow mutual recursion */
  static BOOL
-   compile_regex(int, int, uschar **, const uschar **, int *, BOOL, BOOL, int,
+   compile_regex(int, uschar **, const uschar **, int *, BOOL, BOOL, int, int,
      int *, int *, branch_chain *, compile_data *, int *);
-Line 506 
 static const char *
+Line 570 
 static const char *
  find_error_text(int n)
  {
  const char *s = error_texts;
  for (; n > 0; n--)
    {
    while (*s++ != 0) {};
    if (*s == 0) return "Error text not found (please report)";
    }
  return s;
  }
  /*************************************************
+ *            Check for counted repeat            *
+ *************************************************/
+ /* This function is called when a '{' is encountered in a place where it might
+ start a quantifier. It looks ahead to see if it really is a quantifier or not.
+ It is only a quantifier if it is one of the forms {ddd} {ddd,} or {ddd,ddd}
+ where the ddds are digits.
+ Arguments:
+   p         pointer to the first char after '{'
+ Returns:    TRUE or FALSE
+ */
+ static BOOL
+ is_counted_repeat(const uschar *p)
+ {
+ if ((digitab[*p++] & ctype_digit) == 0) return FALSE;
+ while ((digitab[*p] & ctype_digit) != 0) p++;
+ if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE;
+ if (*p++ != CHAR_COMMA) return FALSE;
+ if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE;
+ if ((digitab[*p++] & ctype_digit) == 0) return FALSE;
+ while ((digitab[*p] & ctype_digit) != 0) p++;
+ return (*p == CHAR_RIGHT_CURLY_BRACKET);
+ }
+ /*************************************************
  *            Handle escapes                      *
  *************************************************/
-Line 581 
 else
+Line 678 
 else
      case CHAR_l:
      case CHAR_L:
-     case CHAR_N:
+     *errorcodeptr = ERR37;
+     break;
      case CHAR_u:
+     if ((options & PCRE_JAVASCRIPT_COMPAT) != 0)
+       {
+       /* In JavaScript, \u must be followed by four hexadecimal numbers.
+       Otherwise it is a lowercase u letter. */
+       if ((digitab[ptr[1]] & ctype_xdigit) != 0 && (digitab[ptr[2]] & ctype_xdigit) != 0
+            && (digitab[ptr[3]] & ctype_xdigit) != 0 && (digitab[ptr[4]] & ctype_xdigit) != 0)
+         {
+         c = 0;
+         for (i = 0; i < 4; ++i)
+           {
+           register int cc = *(++ptr);
+ #ifndef EBCDIC  /* ASCII/UTF-8 coding */
+           if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
+           c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
+ #else           /* EBCDIC coding */
+           if (cc >= CHAR_a && cc <= CHAR_z) cc += 64;  /* Convert to upper case */
+           c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));
+ #endif
+           }
+         }
+       }
+     else
+       *errorcodeptr = ERR37;
+     break;
      case CHAR_U:
-     *errorcodeptr = ERR37;
+     /* In JavaScript, \U is an uppercase U letter. */
+     if ((options & PCRE_JAVASCRIPT_COMPAT) == 0) *errorcodeptr = ERR37;
      break;
-     /* \g must be followed by one of a number of specific things:
+     /* In a character class, \g is just a literal "g". Outside a character
+     class, \g must be followed by one of a number of specific things:
      (1) A number, either plain or braced. If positive, it is an absolute
      backreference. If negative, it is a relative backreference. This is a Perl
-Line 604 
 else
+Line 730 
 else
      the -ESC_g code (cf \k). */
      case CHAR_g:
+     if (isclass) break;
      if (ptr[1] == CHAR_LESS_THAN_SIGN || ptr[1] == CHAR_APOSTROPHE)
        {
        c = -ESC_g;
-Line 732 
 else
+Line 859 
 else
      treated as a data character. */
      case CHAR_x:
+     if ((options & PCRE_JAVASCRIPT_COMPAT) != 0)
+       {
+       /* In JavaScript, \x must be followed by two hexadecimal numbers.
+       Otherwise it is a lowercase x letter. */
+       if ((digitab[ptr[1]] & ctype_xdigit) != 0 && (digitab[ptr[2]] & ctype_xdigit) != 0)
+         {
+         c = 0;
+         for (i = 0; i < 2; ++i)
+           {
+           register int cc = *(++ptr);
+ #ifndef EBCDIC  /* ASCII/UTF-8 coding */
+           if (cc >= CHAR_a) cc -= 32;               /* Convert to upper case */
+           c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10));
+ #else           /* EBCDIC coding */
+           if (cc >= CHAR_a && cc <= CHAR_z) cc += 64;  /* Convert to upper case */
+           c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10));
+ #endif
+           }
+         }
+       break;
+       }
      if (ptr[1] == CHAR_LEFT_CURLY_BRACKET)
        {
        const uschar *pt = ptr + 2;
-Line 782 
 else
+Line 931 
 else
      break;
      /* For \c, a following letter is upper-cased; then the 0x40 bit is flipped.
-     This coding is ASCII-specific, but then the whole concept of \cx is
+     An error is given if the byte following \c is not an ASCII character. This
+     coding is ASCII-specific, but then the whole concept of \cx is
      ASCII-specific. (However, an EBCDIC equivalent has now been added.) */
      case CHAR_c:
-Line 792 
 else
+Line 942 
 else
        *errorcodeptr = ERR2;
        break;
        }
+ #ifndef EBCDIC    /* ASCII/UTF-8 coding */
- #ifndef EBCDIC  /* ASCII/UTF-8 coding */
+     if (c > 127)  /* Excludes all non-ASCII in either mode */
+       {
+       *errorcodeptr = ERR68;
+       break;
+       }
      if (c >= CHAR_a && c <= CHAR_z) c -= 32;
      c ^= 0x40;
- #else           /* EBCDIC coding */
+ #else             /* EBCDIC coding */
      if (c >= CHAR_a && c <= CHAR_z) c += 64;
      c ^= 0xC0;
  #endif
-Line 819 
 else
+Line 973 
 else
      }
    }
+ /* Perl supports \N{name} for character names, as well as plain \N for "not
+ newline". PCRE does not support \N{name}. However, it does support
+ quantification such as \N{2,3}. */
+ if (c == -ESC_N && ptr[1] == CHAR_LEFT_CURLY_BRACKET &&
+      !is_counted_repeat(ptr+2))
+   *errorcodeptr = ERR37;
+ /* If PCRE_UCP is set, we change the values for \d etc. */
+ if ((options & PCRE_UCP) != 0 && c <= -ESC_D && c >= -ESC_w)
+   c -= (ESC_DU - ESC_D);
+ /* Set the pointer to the final character before returning. */
  *ptrptr = ptr;
  return c;
  }
-Line 919 
 return -1;
+Line 1088 
 return -1;
  /*************************************************
- *            Check for counted repeat            *
- *************************************************/
- /* This function is called when a '{' is encountered in a place where it might
- start a quantifier. It looks ahead to see if it really is a quantifier or not.
- It is only a quantifier if it is one of the forms {ddd} {ddd,} or {ddd,ddd}
- where the ddds are digits.
- Arguments:
-   p         pointer to the first char after '{'
- Returns:    TRUE or FALSE
- */
- static BOOL
- is_counted_repeat(const uschar *p)
- {
- if ((digitab[*p++] & ctype_digit) == 0) return FALSE;
- while ((digitab[*p] & ctype_digit) != 0) p++;
- if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE;
- if (*p++ != CHAR_COMMA) return FALSE;
- if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE;
- if ((digitab[*p++] & ctype_digit) == 0) return FALSE;
- while ((digitab[*p] & ctype_digit) != 0) p++;
- return (*p == CHAR_RIGHT_CURLY_BRACKET);
- }
- /*************************************************
  *         Read repeat counts                     *
  *************************************************/
-Line 1027 
 top-level call starts at the beginning o
+Line 1163 
 top-level call starts at the beginning o
  start at a parenthesis. It scans along a pattern's text looking for capturing
  subpatterns, and counting them. If it finds a named pattern that matches the
  name it is given, it returns its number. Alternatively, if the name is NULL, it
- returns when it reaches a given numbered subpattern. We know that if (?P< is
+ returns when it reaches a given numbered subpattern. Recursion is used to keep
- encountered, the name will be terminated by '>' because that is checked in the
+ track of subpatterns that reset the capturing group numbers - the (?| feature.
- first pass. Recursion is used to keep track of subpatterns that reset the
- capturing group numbers - the (?| feature.
+ This function was originally called only from the second pass, in which we know
+ that if (?< or (?' or (?P< is encountered, the name will be correctly
+ terminated because that is checked in the first pass. There is now one call to
+ this function in the first pass, to check for a recursive back reference by
+ name (so that we can make the whole group atomic). In this case, we need check
+ only up to the current position in the pattern, and that is still OK because
+ and previous occurrences will have been checked. To make this work, the test
+ for "end of pattern" is a check against cd->end_pattern in the main loop,
+ instead of looking for a binary zero. This means that the special first-pass
+ call can adjust cd->end_pattern temporarily. (Checks for binary zero while
+ processing items within the loop are OK, because afterwards the main loop will
+ terminate.)
  Arguments:
    ptrptr       address of the current character pointer (updated)
-Line 1038 
 Arguments:
+Line 1185 
 Arguments:
    name         name to seek, or NULL if seeking a numbered subpattern
    lorn         name length, or subpattern number if name is NULL
    xmode        TRUE if we are in /x mode
+   utf8         TRUE if we are in UTF-8 mode
    count        pointer to the current capturing subpattern number (updated)
  Returns:       the number of the named subpattern, or -1 if not found
-Line 1045 
 Returns:       the number of the named s
+Line 1193 
 Returns:       the number of the named s
  static int
  find_parens_sub(uschar **ptrptr, compile_data *cd, const uschar *name, int lorn,
-   BOOL xmode, int *count)
+   BOOL xmode, BOOL utf8, int *count)
  {
  uschar *ptr = *ptrptr;
  int start_count = *count;
-Line 1057 
 dealing with. The very first call may no
+Line 1205 
 dealing with. The very first call may no
  if (ptr[0] == CHAR_LEFT_PARENTHESIS)
    {
-   if (ptr[1] == CHAR_QUESTION_MARK &&
+   /* Handle specials such as (*SKIP) or (*UTF8) etc. */
-       ptr[2] == CHAR_VERTICAL_LINE)
+   if (ptr[1] == CHAR_ASTERISK) ptr += 2;
+   /* Handle a normal, unnamed capturing parenthesis. */
+   else if (ptr[1] != CHAR_QUESTION_MARK)
+     {
+     *count += 1;
+     if (name == NULL && *count == lorn) return *count;
+     ptr++;
+     }
+   /* All cases now have (? at the start. Remember when we are in a group
+   where the parenthesis numbers are duplicated. */
+   else if (ptr[2] == CHAR_VERTICAL_LINE)
      {
      ptr += 3;
      dup_parens = TRUE;
      }
-   /* Handle a normal, unnamed capturing parenthesis */
+   /* Handle comments; all characters are allowed until a ket is reached. */
-   else if (ptr[1] != CHAR_QUESTION_MARK && ptr[1] != CHAR_ASTERISK)
+   else if (ptr[2] == CHAR_NUMBER_SIGN)
      {
-     *count += 1;
+     for (ptr += 3; *ptr != 0; ptr++) if (*ptr == CHAR_RIGHT_PARENTHESIS) break;
-     if (name == NULL && *count == lorn) return *count;
+     goto FAIL_EXIT;
-     ptr++;
      }
    /* Handle a condition. If it is an assertion, just carry on so that it
    is processed as normal. If not, skip to the closing parenthesis of the
-   condition (there can't be any nested parens. */
+   condition (there can't be any nested parens). */
    else if (ptr[2] == CHAR_LEFT_PARENTHESIS)
      {
-Line 1087 
 if (ptr[0] == CHAR_LEFT_PARENTHESIS)
+Line 1249 
 if (ptr[0] == CHAR_LEFT_PARENTHESIS)
        }
      }
-   /* We have either (? or (* and not a condition */
+   /* Start with (? but not a condition. */
    else
      {
-Line 1116 
 if (ptr[0] == CHAR_LEFT_PARENTHESIS)
+Line 1278 
 if (ptr[0] == CHAR_LEFT_PARENTHESIS)
    }
  /* Past any initial parenthesis handling, scan for parentheses or vertical
- bars. */
+ bars. Stop if we get to cd->end_pattern. Note that this is important for the
+ first-pass call when this value is temporarily adjusted to stop at the current
+ position. So DO NOT change this to a test for binary zero. */
- for (; *ptr != 0; ptr++)
+ for (; ptr < cd->end_pattern; ptr++)
    {
    /* Skip over backslashed characters and also entire \Q...\E */
-Line 1192 
 for (; *ptr != 0; ptr++)
+Line 1356 
 for (; *ptr != 0; ptr++)
    if (xmode && *ptr == CHAR_NUMBER_SIGN)
      {
-     while (*(++ptr) != 0 && *ptr != CHAR_NL) {};
+     ptr++;
+     while (*ptr != 0)
+       {
+       if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
+       ptr++;
+ #ifdef SUPPORT_UTF8
+       if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;
+ #endif
+       }
      if (*ptr == 0) goto FAIL_EXIT;
      continue;
      }
-Line 1201 
 for (; *ptr != 0; ptr++)
+Line 1373 
 for (; *ptr != 0; ptr++)
    if (*ptr == CHAR_LEFT_PARENTHESIS)
      {
-     int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, count);
+     int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf8, count);
      if (rc > 0) return rc;
      if (*ptr == 0) goto FAIL_EXIT;
      }
-Line 1209 
 for (; *ptr != 0; ptr++)
+Line 1381 
 for (; *ptr != 0; ptr++)
    else if (*ptr == CHAR_RIGHT_PARENTHESIS)
      {
      if (dup_parens && *count < hwm_count) *count = hwm_count;
-     *ptrptr = ptr;
+     goto FAIL_EXIT;
-     return -1;
      }
    else if (*ptr == CHAR_VERTICAL_LINE && dup_parens)
-Line 1248 
 Arguments:
+Line 1419 
 Arguments:
    name         name to seek, or NULL if seeking a numbered subpattern
    lorn         name length, or subpattern number if name is NULL
    xmode        TRUE if we are in /x mode
+   utf8         TRUE if we are in UTF-8 mode
  Returns:       the number of the found subpattern, or -1 if not found
  */
  static int
- find_parens(compile_data *cd, const uschar *name, int lorn, BOOL xmode)
+ find_parens(compile_data *cd, const uschar *name, int lorn, BOOL xmode,
+   BOOL utf8)
  {
  uschar *ptr = (uschar *)cd->start_pattern;
  int count = 0;
-Line 1266 
 matching closing parens. That is why we
+Line 1439 
 matching closing parens. That is why we
  for (;;)
    {
-   rc = find_parens_sub(&ptr, cd, name, lorn, xmode, &count);
+   rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf8, &count);
    if (rc > 0 || *ptr++ == 0) break;
    }
-Line 1282 
 return rc;
+Line 1455 
 return rc;
  /* This is called by several functions that scan a compiled expression looking
  for a fixed first character, or an anchoring op code etc. It skips over things
- that do not influence this. For some calls, a change of option is important.
+ that do not influence this. For some calls, it makes sense to skip negative
- For some calls, it makes sense to skip negative forward and all backward
+ forward and all backward assertions, and also the \b assertion; for others it
- assertions, and also the \b assertion; for others it does not.
+ does not.
  Arguments:
    code         pointer to the start of the group
-   options      pointer to external options
-   optbit       the option bit whose changing is significant, or
-                  zero if none are
    skipassert   TRUE if certain assertions are to be skipped
  Returns:       pointer to the first significant opcode
  */
  static const uschar*
- first_significant_code(const uschar *code, int *options, int optbit,
+ first_significant_code(const uschar *code, BOOL skipassert)
-   BOOL skipassert)
  {
  for (;;)
    {
    switch ((int)*code)
      {
-     case OP_OPT:
-     if (optbit > 0 && ((int)code[1] & optbit) != (*options & optbit))
-       *options = (int)code[1];
-     code += 2;
-     break;
      case OP_ASSERT_NOT:
      case OP_ASSERTBACK:
      case OP_ASSERTBACK_NOT:
-Line 1359 
 and doing the check at the end; a flag s
+Line 1522 
 and doing the check at the end; a flag s
  Arguments:
    code     points to the start of the pattern (the bracket)
-   options  the compiling options
+   utf8     TRUE in UTF-8 mode
    atend    TRUE if called when the pattern is complete
    cd       the "compile data" structure
  Returns:   the fixed length,
               or -1 if there is no fixed length,
-              or -2 if \C was encountered
+              or -2 if \C was encountered (in UTF-8 mode only)
               or -3 if an OP_RECURSE item was encountered and atend is FALSE
+              or -4 if an unknown opcode was encountered (internal error)
  */
  static int
- find_fixedlength(uschar *code, int options, BOOL atend, compile_data *cd)
+ find_fixedlength(uschar *code, BOOL utf8, BOOL atend, compile_data *cd)
  {
  int length = -1;
-Line 1387 
 for (;;)
+Line 1551 
 for (;;)
    register int op = *cc;
    switch (op)
      {
+     /* We only need to continue for OP_CBRA (normal capturing bracket) and
+     OP_BRA (normal non-capturing bracket) because the other variants of these
+     opcodes are all concerned with unlimited repeated groups, which of course
+     are not of fixed length. */
      case OP_CBRA:
      case OP_BRA:
      case OP_ONCE:
+     case OP_ONCE_NC:
      case OP_COND:
-     d = find_fixedlength(cc + ((op == OP_CBRA)? 2:0), options, atend, cd);
+     d = find_fixedlength(cc + ((op == OP_CBRA)? 2:0), utf8, atend, cd);
      if (d < 0) return d;
      branchlength += d;
      do cc += GET(cc, 1); while (*cc == OP_ALT);
      cc += 1 + LINK_SIZE;
      break;
-     /* Reached end of a branch; if it's a ket it is the end of a nested
+     /* Reached end of a branch; if it's a ket it is the end of a nested call.
-     call. If it's ALT it is an alternation in a nested call. If it is
+     If it's ALT it is an alternation in a nested call. An ACCEPT is effectively
-     END it's the end of the outer call. All can be handled by the same code. */
+     an ALT. If it is END it's the end of the outer call. All can be handled by
+     the same code. Note that we must not include the OP_KETRxxx opcodes here,
+     because they all imply an unlimited repeat. */
      case OP_ALT:
      case OP_KET:
-     case OP_KETRMAX:
-     case OP_KETRMIN:
      case OP_END:
+     case OP_ACCEPT:
+     case OP_ASSERT_ACCEPT:
      if (length < 0) length = branchlength;
        else if (length != branchlength) return -1;
      if (*cc != OP_ALT) return length;
-Line 1423 
 for (;;)
+Line 1595 
 for (;;)
      cs = ce = (uschar *)cd->start_code + GET(cc, 1);  /* Start subpattern */
      do ce += GET(ce, 1); while (*ce == OP_ALT);       /* End subpattern */
      if (cc > cs && cc < ce) return -1;                /* Recursion */
-     d = find_fixedlength(cs + 2, options, atend, cd);
+     d = find_fixedlength(cs + 2, utf8, atend, cd);
      if (d < 0) return d;
      branchlength += d;
      cc += 1 + LINK_SIZE;
-Line 1440 
 for (;;)
+Line 1612 
 for (;;)
      /* Skip over things that don't match chars */
-     case OP_REVERSE:
+     case OP_MARK:
+     case OP_PRUNE_ARG:
+     case OP_SKIP_ARG:
+     case OP_THEN_ARG:
+     cc += cc[1] + _pcre_OP_lengths[*cc];
+     break;
+     case OP_CALLOUT:
+     case OP_CIRC:
+     case OP_CIRCM:
+     case OP_CLOSE:
+     case OP_COMMIT:
      case OP_CREF:
-     case OP_NCREF:
-     case OP_RREF:
-     case OP_NRREF:
      case OP_DEF:
-     case OP_OPT:
+     case OP_DOLL:
-     case OP_CALLOUT:
+     case OP_DOLLM:
-     case OP_SOD:
-     case OP_SOM:
-     case OP_SET_SOM:
      case OP_EOD:
      case OP_EODN:
-     case OP_CIRC:
+     case OP_FAIL:
-     case OP_DOLL:
+     case OP_NCREF:
+     case OP_NRREF:
      case OP_NOT_WORD_BOUNDARY:
+     case OP_PRUNE:
+     case OP_REVERSE:
+     case OP_RREF:
+     case OP_SET_SOM:
+     case OP_SKIP:
+     case OP_SOD:
+     case OP_SOM:
+     case OP_THEN:
      case OP_WORD_BOUNDARY:
      cc += _pcre_OP_lengths[*cc];
      break;
-Line 1463 
 for (;;)
+Line 1649 
 for (;;)
      /* Handle literal characters */
      case OP_CHAR:
-     case OP_CHARNC:
+     case OP_CHARI:
      case OP_NOT:
+     case OP_NOTI:
      branchlength++;
      cc += 2;
  #ifdef SUPPORT_UTF8
-     if ((options & PCRE_UTF8) != 0 && cc[-1] >= 0xc0)
+     if (utf8 && cc[-1] >= 0xc0) cc += _pcre_utf8_table4[cc[-1] & 0x3f];
-       cc += _pcre_utf8_table4[cc[-1] & 0x3f];
  #endif
      break;
-Line 1477 
 for (;;)
+Line 1663 
 for (;;)
      need to skip over a multibyte character in UTF8 mode.  */
      case OP_EXACT:
+     case OP_EXACTI:
+     case OP_NOTEXACT:
+     case OP_NOTEXACTI:
      branchlength += GET2(cc,1);
      cc += 4;
  #ifdef SUPPORT_UTF8
-     if ((options & PCRE_UTF8) != 0 && cc[-1] >= 0xc0)
+     if (utf8 && cc[-1] >= 0xc0) cc += _pcre_utf8_table4[cc[-1] & 0x3f];
-       cc += _pcre_utf8_table4[cc[-1] & 0x3f];
  #endif
      break;
-Line 1498 
 for (;;)
+Line 1686 
 for (;;)
      cc += 2;
      /* Fall through */
+     case OP_HSPACE:
+     case OP_VSPACE:
+     case OP_NOT_HSPACE:
+     case OP_NOT_VSPACE:
      case OP_NOT_DIGIT:
      case OP_DIGIT:
      case OP_NOT_WHITESPACE:
-Line 1510 
 for (;;)
+Line 1702 
 for (;;)
      cc++;
      break;
-     /* The single-byte matcher isn't allowed */
+     /* The single-byte matcher isn't allowed. This only happens in UTF-8 mode;
+     otherwise \C is coded as OP_ALLANY. */
      case OP_ANYBYTE:
      return -2;
-Line 1529 
 for (;;)
+Line 1722 
 for (;;)
      switch (*cc)
        {
+       case OP_CRPLUS:
+       case OP_CRMINPLUS:
        case OP_CRSTAR:
        case OP_CRMINSTAR:
        case OP_CRQUERY:
-Line 1549 
 for (;;)
+Line 1744 
 for (;;)
      /* Anything else is variable length */
-     default:
+     case OP_ANYNL:
+     case OP_BRAMINZERO:
+     case OP_BRAPOS:
+     case OP_BRAPOSZERO:
+     case OP_BRAZERO:
+     case OP_CBRAPOS:
+     case OP_EXTUNI:
+     case OP_KETRMAX:
+     case OP_KETRMIN:
+     case OP_KETRPOS:
+     case OP_MINPLUS:
+     case OP_MINPLUSI:
+     case OP_MINQUERY:
+     case OP_MINQUERYI:
+     case OP_MINSTAR:
+     case OP_MINSTARI:
+     case OP_MINUPTO:
+     case OP_MINUPTOI:
+     case OP_NOTMINPLUS:
+     case OP_NOTMINPLUSI:
+     case OP_NOTMINQUERY:
+     case OP_NOTMINQUERYI:
+     case OP_NOTMINSTAR:
+     case OP_NOTMINSTARI:
+     case OP_NOTMINUPTO:
+     case OP_NOTMINUPTOI:
+     case OP_NOTPLUS:
+     case OP_NOTPLUSI:
+     case OP_NOTPOSPLUS:
+     case OP_NOTPOSPLUSI:
+     case OP_NOTPOSQUERY:
+     case OP_NOTPOSQUERYI:
+     case OP_NOTPOSSTAR:
+     case OP_NOTPOSSTARI:
+     case OP_NOTPOSUPTO:
+     case OP_NOTPOSUPTOI:
+     case OP_NOTQUERY:
+     case OP_NOTQUERYI:
+     case OP_NOTSTAR:
+     case OP_NOTSTARI:
+     case OP_NOTUPTO:
+     case OP_NOTUPTOI:
+     case OP_PLUS:
+     case OP_PLUSI:
+     case OP_POSPLUS:
+     case OP_POSPLUSI:
+     case OP_POSQUERY:
+     case OP_POSQUERYI:
+     case OP_POSSTAR:
+     case OP_POSSTARI:
+     case OP_POSUPTO:
+     case OP_POSUPTOI:
+     case OP_QUERY:
+     case OP_QUERYI:
+     case OP_REF:
+     case OP_REFI:
+     case OP_SBRA:
+     case OP_SBRAPOS:
+     case OP_SCBRA:
+     case OP_SCBRAPOS:
+     case OP_SCOND:
+     case OP_SKIPZERO:
+     case OP_STAR:
+     case OP_STARI:
+     case OP_TYPEMINPLUS:
+     case OP_TYPEMINQUERY:
+     case OP_TYPEMINSTAR:
+     case OP_TYPEMINUPTO:
+     case OP_TYPEPLUS:
+     case OP_TYPEPOSPLUS:
+     case OP_TYPEPOSQUERY:
+     case OP_TYPEPOSSTAR:
+     case OP_TYPEPOSUPTO:
+     case OP_TYPEQUERY:
+     case OP_TYPESTAR:
+     case OP_TYPEUPTO:
+     case OP_UPTO:
+     case OP_UPTOI:
      return -1;
+     /* Catch unrecognized opcodes so that when new ones are added they
+     are not forgotten, as has happened in the past. */
+     default:
+     return -4;
      }
    }
  /* Control never gets here */
-Line 1583 
 _pcre_find_bracket(const uschar *code, B
+Line 1861 
 _pcre_find_bracket(const uschar *code, B
  for (;;)
    {
    register int c = *code;
    if (c == OP_END) return NULL;
    /* XCLASS is used for classes that cannot be represented just by a bit
-Line 1601 
 for (;;)
+Line 1880 
 for (;;)
    /* Handle capturing bracket */
-   else if (c == OP_CBRA)
+   else if (c == OP_CBRA || c == OP_SCBRA ||
+            c == OP_CBRAPOS || c == OP_SCBRAPOS)
      {
      int n = GET2(code, 1+LINK_SIZE);
      if (n == number) return (uschar *)code;
-Line 1610 
 for (;;)
+Line 1890 
 for (;;)
    /* Otherwise, we can get the item's length from the table, except that for
    repeated character types, we have to test for \p and \P, which have an extra
-   two bytes of parameters. */
+   two bytes of parameters, and for MARK/PRUNE/SKIP/THEN with an argument, we
+   must add in its length. */
    else
      {
-Line 1634 
 for (;;)
+Line 1915 
 for (;;)
        case OP_TYPEPOSUPTO:
        if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2;
        break;
+       case OP_MARK:
+       case OP_PRUNE_ARG:
+       case OP_SKIP_ARG:
+       code += code[1];
+       break;
+       case OP_THEN_ARG:
+       code += code[1];
+       break;
        }
      /* Add in the fixed length from the table */
-Line 1648 
 for (;;)
+Line 1939 
 for (;;)
      if (utf8) switch(c)
        {
        case OP_CHAR:
-       case OP_CHARNC:
+       case OP_CHARI:
        case OP_EXACT:
+       case OP_EXACTI:
        case OP_UPTO:
+       case OP_UPTOI:
        case OP_MINUPTO:
+       case OP_MINUPTOI:
        case OP_POSUPTO:
+       case OP_POSUPTOI:
        case OP_STAR:
+       case OP_STARI:
        case OP_MINSTAR:
+       case OP_MINSTARI:
        case OP_POSSTAR:
+       case OP_POSSTARI:
        case OP_PLUS:
+       case OP_PLUSI:
        case OP_MINPLUS:
+       case OP_MINPLUSI:
        case OP_POSPLUS:
+       case OP_POSPLUSI:
        case OP_QUERY:
+       case OP_QUERYI:
        case OP_MINQUERY:
+       case OP_MINQUERYI:
        case OP_POSQUERY:
+       case OP_POSQUERYI:
        if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f];
        break;
        }
-Line 1705 
 for (;;)
+Line 2009 
 for (;;)
    /* Otherwise, we can get the item's length from the table, except that for
    repeated character types, we have to test for \p and \P, which have an extra
-   two bytes of parameters. */
+   two bytes of parameters, and for MARK/PRUNE/SKIP/THEN with an argument, we
+   must add in its length. */
    else
      {
-Line 1729 
 for (;;)
+Line 2034 
 for (;;)
        case OP_TYPEEXACT:
        if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2;
        break;
+       case OP_MARK:
+       case OP_PRUNE_ARG:
+       case OP_SKIP_ARG:
+       code += code[1];
+       break;
+       case OP_THEN_ARG:
+       code += code[1];
+       break;
        }
      /* Add in the fixed length from the table */
-Line 1743 
 for (;;)
+Line 2058 
 for (;;)
      if (utf8) switch(c)
        {
        case OP_CHAR:
-       case OP_CHARNC:
+       case OP_CHARI:
        case OP_EXACT:
+       case OP_EXACTI:
        case OP_UPTO:
+       case OP_UPTOI:
        case OP_MINUPTO:
+       case OP_MINUPTOI:
        case OP_POSUPTO:
+       case OP_POSUPTOI:
        case OP_STAR:
+       case OP_STARI:
        case OP_MINSTAR:
+       case OP_MINSTARI:
        case OP_POSSTAR:
+       case OP_POSSTARI:
        case OP_PLUS:
+       case OP_PLUSI:
        case OP_MINPLUS:
+       case OP_MINPLUSI:
        case OP_POSPLUS:
+       case OP_POSPLUSI:
        case OP_QUERY:
+       case OP_QUERYI:
        case OP_MINQUERY:
+       case OP_MINQUERYI:
        case OP_POSQUERY:
+       case OP_POSQUERYI:
        if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f];
        break;
        }
-Line 1795 
 could_be_empty_branch(const uschar *code
+Line 2123 
 could_be_empty_branch(const uschar *code
    compile_data *cd)
  {
  register int c;
- for (code = first_significant_code(code + _pcre_OP_lengths[*code], NULL, 0, TRUE);
+ for (code = first_significant_code(code + _pcre_OP_lengths[*code], TRUE);
       code < endcode;
-      code = first_significant_code(code + _pcre_OP_lengths[c], NULL, 0, TRUE))
+      code = first_significant_code(code + _pcre_OP_lengths[c], TRUE))
    {
    const uschar *ccode;
    c = *code;
    /* Skip over forward assertions; the other assertions are skipped by
    first_significant_code() with a TRUE final argument. */
-Line 1813 
 for (code = first_significant_code(code
+Line 2141 
 for (code = first_significant_code(code
      continue;
      }
-   /* Groups with zero repeats can of course be empty; skip them. */
-   if (c == OP_BRAZERO || c == OP_BRAMINZERO || c == OP_SKIPZERO)
-     {
-     code += _pcre_OP_lengths[c];
-     do code += GET(code, 1); while (*code == OP_ALT);
-     c = *code;
-     continue;
-     }
    /* For a recursion/subroutine call, if its end has been reached, which
-   implies a subroutine call, we can scan it. */
+   implies a backward reference subroutine call, we can scan it. If it's a
+   forward reference subroutine call, we can't. To detect forward reference
+   we have to scan up the list that is kept in the workspace. This function is
+   called only when doing the real compile, not during the pre-compile that
+   measures the size of the compiled pattern. */
    if (c == OP_RECURSE)
      {
-     const uschar *scode = cd->start_code + GET(code, 1);
+     const uschar *scode;
+     BOOL empty_branch;
+     /* Test for forward reference */
+     for (scode = cd->start_workspace; scode < cd->hwm; scode += LINK_SIZE)
+       if (GET(scode, 0) == code + 1 - cd->start_code) return TRUE;
+     /* Not a forward reference, test for completed backward reference */
+     empty_branch = FALSE;
+     scode = cd->start_code + GET(code, 1);
      if (GET(scode, 1) == 0) return TRUE;    /* Unclosed */
+     /* Completed backwards reference */
      do
        {
-       if (!could_be_empty_branch(scode, endcode, utf8, cd)) return FALSE;
+       if (could_be_empty_branch(scode, endcode, utf8, cd))
+         {
+         empty_branch = TRUE;
+         break;
+         }
        scode += GET(scode, 1);
        }
      while (*scode == OP_ALT);
+     if (!empty_branch) return FALSE;  /* All branches are non-empty */
      continue;
      }
+   /* Groups with zero repeats can of course be empty; skip them. */
+   if (c == OP_BRAZERO || c == OP_BRAMINZERO || c == OP_SKIPZERO ||
+       c == OP_BRAPOSZERO)
+     {
+     code += _pcre_OP_lengths[c];
+     do code += GET(code, 1); while (*code == OP_ALT);
+     c = *code;
+     continue;
+     }
+   /* A nested group that is already marked as "could be empty" can just be
+   skipped. */
+   if (c == OP_SBRA  || c == OP_SBRAPOS ||
+       c == OP_SCBRA || c == OP_SCBRAPOS)
+     {
+     do code += GET(code, 1); while (*code == OP_ALT);
+     c = *code;
+     continue;
+     }
    /* For other groups, scan the branches. */
-   if (c == OP_BRA || c == OP_CBRA || c == OP_ONCE || c == OP_COND)
+   if (c == OP_BRA  || c == OP_BRAPOS ||
+       c == OP_CBRA || c == OP_CBRAPOS ||
+       c == OP_ONCE || c == OP_ONCE_NC ||
+       c == OP_COND)
      {
      BOOL empty_branch;
      if (GET(code, 1) == 0) return TRUE;    /* Hit unclosed bracket */
-Line 1928 
 for (code = first_significant_code(code
+Line 2295 
 for (code = first_significant_code(code
      case OP_ALLANY:
      case OP_ANYBYTE:
      case OP_CHAR:
-     case OP_CHARNC:
+     case OP_CHARI:
      case OP_NOT:
+     case OP_NOTI:
      case OP_PLUS:
      case OP_MINPLUS:
      case OP_POSPLUS:
-Line 1969 
 for (code = first_significant_code(code
+Line 2337 
 for (code = first_significant_code(code
      case OP_KET:
      case OP_KETRMAX:
      case OP_KETRMIN:
+     case OP_KETRPOS:
      case OP_ALT:
      return TRUE;
-Line 1977 
 for (code = first_significant_code(code
+Line 2346 
 for (code = first_significant_code(code
  #ifdef SUPPORT_UTF8
      case OP_STAR:
+     case OP_STARI:
      case OP_MINSTAR:
+     case OP_MINSTARI:
      case OP_POSSTAR:
+     case OP_POSSTARI:
      case OP_QUERY:
+     case OP_QUERYI:
      case OP_MINQUERY:
+     case OP_MINQUERYI:
      case OP_POSQUERY:
+     case OP_POSQUERYI:
      if (utf8 && code[1] >= 0xc0) code += _pcre_utf8_table4[code[1] & 0x3f];
      break;
      case OP_UPTO:
+     case OP_UPTOI:
      case OP_MINUPTO:
+     case OP_MINUPTOI:
      case OP_POSUPTO:
+     case OP_POSUPTOI:
      if (utf8 && code[3] >= 0xc0) code += _pcre_utf8_table4[code[3] & 0x3f];
      break;
  #endif
+     /* MARK, and PRUNE/SKIP/THEN with an argument must skip over the argument
+     string. */
+     case OP_MARK:
+     case OP_PRUNE_ARG:
+     case OP_SKIP_ARG:
+     code += code[1];
+     break;
+     case OP_THEN_ARG:
+     code += code[1];
+     break;
      /* None of the remaining opcodes are required to match a character. */
      default:
      break;
      }
    }
-Line 2012 
 return TRUE;
+Line 2403 
 return TRUE;
  the current branch of the current pattern to see if it could match the empty
  string. If it could, we must look outwards for branches at other levels,
  stopping when we pass beyond the bracket which is the subject of the recursion.
+ This function is called only during the real compile, not during the
+ pre-compile.
  Arguments:
    code        points to start of the recursion
    endcode     points to where to stop (current RECURSE item)
    bcptr       points to the chain of current (unclosed) branch starts
    utf8        TRUE if in UTF-8 mode
    cd          pointers to tables etc
  Returns:      TRUE if what is matched could be empty
  */
-Line 2062 
 where Perl recognizes it as the POSIX cl
+Line 2455 
 where Perl recognizes it as the POSIX cl
  "l\ower". This is a lesser evil that not diagnosing bad classes when Perl does,
  I think.
+ A user pointed out that PCRE was rejecting [:a[:digit:]] whereas Perl was not.
+ It seems that the appearance of a nested POSIX class supersedes an apparent
+ external class. For example, [:a[:digit:]b:] matches "a", "b", ":", or
+ a digit.
+ In Perl, unescaped square brackets may also appear as part of class names. For
+ example, [:a[:abc]b:] gives unknown POSIX class "[:abc]b:]". However, for
+ [:a[:abc]b][b:] it gives unknown POSIX class "[:abc]b][b:]", which does not
+ seem right at all. PCRE does not allow closing square brackets in POSIX class
+ names.
  Arguments:
    ptr      pointer to the initial [
    endptr   where to return the end pointer
-Line 2076 
 int terminator;          /* Don't combin
+Line 2480 
 int terminator;          /* Don't combin
  terminator = *(++ptr);   /* compiler warns about "non-constant" initializer. */
  for (++ptr; *ptr != 0; ptr++)
    {
-   if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET) ptr++; else
+   if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)
+     ptr++;
+   else if (*ptr == CHAR_RIGHT_SQUARE_BRACKET) return FALSE;
+   else
      {
-     if (*ptr == CHAR_RIGHT_SQUARE_BRACKET) return FALSE;
      if (*ptr == terminator && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET)
        {
        *endptr = ptr;
        return TRUE;
        }
+     if (*ptr == CHAR_LEFT_SQUARE_BRACKET &&
+          (ptr[1] == CHAR_COLON || ptr[1] == CHAR_DOT ||
+           ptr[1] == CHAR_EQUALS_SIGN) &&
+         check_posix_syntax(ptr, endptr))
+       return FALSE;
      }
    }
  return FALSE;
-Line 2212 
 auto_callout(uschar *code, const uschar
+Line 2623 
 auto_callout(uschar *code, const uschar
  {
  *code++ = OP_CALLOUT;
  *code++ = 255;
- PUT(code, 0, ptr - cd->start_pattern);  /* Pattern offset */
+ PUT(code, 0, (int)(ptr - cd->start_pattern));  /* Pattern offset */
- PUT(code, LINK_SIZE, 0);                /* Default length */
+ PUT(code, LINK_SIZE, 0);                       /* Default length */
  return code + 2*LINK_SIZE;
  }
-Line 2238 
 Returns:             nothing
+Line 2649 
 Returns:             nothing
  static void
  complete_callout(uschar *previous_callout, const uschar *ptr, compile_data *cd)
  {
- int length = ptr - cd->start_pattern - GET(previous_callout, 2);
+ int length = (int)(ptr - cd->start_pattern - GET(previous_callout, 2));
  PUT(previous_callout, 2 + LINK_SIZE, length);
  }
-Line 2288 
 for (++c; c <= d; c++)
+Line 2699 
 for (++c; c <= d; c++)
  return TRUE;
  }
+ /*************************************************
+ *        Check a character and a property        *
+ *************************************************/
+ /* This function is called by check_auto_possessive() when a property item
+ is adjacent to a fixed character.
+ Arguments:
+   c            the character
+   ptype        the property type
+   pdata        the data for the type
+   negated      TRUE if it's a negated property (\P or \p{^)
+ Returns:       TRUE if auto-possessifying is OK
+ */
+ static BOOL
+ check_char_prop(int c, int ptype, int pdata, BOOL negated)
+ {
+ const ucd_record *prop = GET_UCD(c);
+ switch(ptype)
+   {
+   case PT_LAMP:
+   return (prop->chartype == ucp_Lu ||
+           prop->chartype == ucp_Ll ||
+           prop->chartype == ucp_Lt) == negated;
+   case PT_GC:
+   return (pdata == _pcre_ucp_gentype[prop->chartype]) == negated;
+   case PT_PC:
+   return (pdata == prop->chartype) == negated;
+   case PT_SC:
+   return (pdata == prop->script) == negated;
+   /* These are specials */
+   case PT_ALNUM:
+   return (_pcre_ucp_gentype[prop->chartype] == ucp_L ||
+           _pcre_ucp_gentype[prop->chartype] == ucp_N) == negated;
+   case PT_SPACE:    /* Perl space */
+   return (_pcre_ucp_gentype[prop->chartype] == ucp_Z ||
+           c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
+           == negated;
+   case PT_PXSPACE:  /* POSIX space */
+   return (_pcre_ucp_gentype[prop->chartype] == ucp_Z ||
+           c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
+           c == CHAR_FF || c == CHAR_CR)
+           == negated;
+   case PT_WORD:
+   return (_pcre_ucp_gentype[prop->chartype] == ucp_L ||
+           _pcre_ucp_gentype[prop->chartype] == ucp_N ||
+           c == CHAR_UNDERSCORE) == negated;
+   }
+ return FALSE;
+ }
  #endif  /* SUPPORT_UCP */
-Line 2301 
 whether the next thing could possibly ma
+Line 2775 
 whether the next thing could possibly ma
  sense to automatically possessify the repeated item.
  Arguments:
-   op_code       the repeated op code
+   previous      pointer to the repeated opcode
-   this          data for this item, depends on the opcode
    utf8          TRUE in UTF-8 mode
-   utf8_char     used for utf8 character bytes, NULL if not relevant
    ptr           next character in pattern
    options       options bits
    cd            contains pointers to tables etc.
-Line 2313 
 Returns:        TRUE if possessifying is
+Line 2785 
 Returns:        TRUE if possessifying is
  */
  static BOOL
- check_auto_possessive(int op_code, int item, BOOL utf8, uschar *utf8_char,
+ check_auto_possessive(const uschar *previous, BOOL utf8, const uschar *ptr,
-   const uschar *ptr, int options, compile_data *cd)
+   int options, compile_data *cd)
  {
- int next;
+ int c, next;
+ int op_code = *previous++;
  /* Skip whitespace and comments in extended mode */
-Line 2327 
 if ((options & PCRE_EXTENDED) != 0)
+Line 2800 
 if ((options & PCRE_EXTENDED) != 0)
      while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++;
      if (*ptr == CHAR_NUMBER_SIGN)
        {
-       while (*(++ptr) != 0)
+       ptr++;
+       while (*ptr != 0)
+         {
          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
+         ptr++;
+ #ifdef SUPPORT_UTF8
+         if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;
+ #endif
+         }
        }
      else break;
      }
-Line 2364 
 if ((options & PCRE_EXTENDED) != 0)
+Line 2844 
 if ((options & PCRE_EXTENDED) != 0)
      while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++;
      if (*ptr == CHAR_NUMBER_SIGN)
        {
-       while (*(++ptr) != 0)
+       ptr++;
+       while (*ptr != 0)
+         {
          if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; }
+         ptr++;
+ #ifdef SUPPORT_UTF8
+         if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;
+ #endif
+         }
        }
      else break;
      }
-Line 2377 
 if (*ptr == CHAR_ASTERISK || *ptr == CHA
+Line 2864 
 if (*ptr == CHAR_ASTERISK || *ptr == CHA
    strncmp((char *)ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0)
      return FALSE;
- /* Now compare the next item with the previous opcode. If the previous is a
+ /* Now compare the next item with the previous opcode. First, handle cases when
- positive single character match, "item" either contains the character or, if
+ the next item is a character. */
- "item" is greater than 127 in utf8 mode, the character's bytes are in
- utf8_char. */
- /* Handle cases when the next item is a character. */
  if (next >= 0) switch(op_code)
    {
    case OP_CHAR:
  #ifdef SUPPORT_UTF8
-   if (utf8 && item > 127) { GETCHAR(item, utf8_char); }
+   GETCHARTEST(c, previous);
  #else
-   (void)(utf8_char);  /* Keep compiler happy by referencing function argument */
+   c = *previous;
  #endif
-   return item != next;
+   return c != next;
-   /* For CHARNC (caseless character) we must check the other case. If we have
+   /* For CHARI (caseless character) we must check the other case. If we have
    Unicode property support, we can use it to test the other case of
    high-valued characters. */
-   case OP_CHARNC:
+   case OP_CHARI:
  #ifdef SUPPORT_UTF8
-   if (utf8 && item > 127) { GETCHAR(item, utf8_char); }
+   GETCHARTEST(c, previous);
+ #else
+   c = *previous;
  #endif
-   if (item == next) return FALSE;
+   if (c == next) return FALSE;
  #ifdef SUPPORT_UTF8
    if (utf8)
      {
-Line 2414 
 if (next >= 0) switch(op_code)
+Line 2898 
 if (next >= 0) switch(op_code)
  #else
      othercase = NOTACHAR;
  #endif
-     return (unsigned int)item != othercase;
+     return (unsigned int)c != othercase;
      }
    else
  #endif  /* SUPPORT_UTF8 */
-   return (item != cd->fcc[next]);  /* Non-UTF-8 mode */
+   return (c != cd->fcc[next]);  /* Non-UTF-8 mode */
-   /* For OP_NOT, "item" must be a single-byte character. */
+   /* For OP_NOT and OP_NOTI, the data is always a single-byte character. These
+   opcodes are not used for multi-byte characters, because they are coded using
+   an XCLASS instead. */
    case OP_NOT:
-   if (item == next) return TRUE;
+   return (c = *previous) == next;
-   if ((options & PCRE_CASELESS) == 0) return FALSE;
+   case OP_NOTI:
+   if ((c = *previous) == next) return TRUE;
  #ifdef SUPPORT_UTF8
    if (utf8)
      {
-Line 2435 
 if (next >= 0) switch(op_code)
+Line 2923 
 if (next >= 0) switch(op_code)
  #else
      othercase = NOTACHAR;
  #endif
-     return (unsigned int)item == othercase;
+     return (unsigned int)c == othercase;
      }
    else
  #endif  /* SUPPORT_UTF8 */
-   return (item == cd->fcc[next]);  /* Non-UTF-8 mode */
+   return (c == cd->fcc[next]);  /* Non-UTF-8 mode */
+   /* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* set.
+   When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */
    case OP_DIGIT:
    return next > 127 || (cd->ctypes[next] & ctype_digit) == 0;
-Line 2482 
 if (next >= 0) switch(op_code)
+Line 2973 
 if (next >= 0) switch(op_code)
      case 0x202f:
      case 0x205f:
      case 0x3000:
-     return op_code != OP_HSPACE;
+     return op_code == OP_NOT_HSPACE;
      default:
-     return op_code == OP_HSPACE;
+     return op_code != OP_NOT_HSPACE;
      }
+   case OP_ANYNL:
    case OP_VSPACE:
    case OP_NOT_VSPACE:
    switch(next)
-Line 2498 
 if (next >= 0) switch(op_code)
+Line 2990 
 if (next >= 0) switch(op_code)
      case 0x85:
      case 0x2028:
      case 0x2029:
-     return op_code != OP_VSPACE;
+     return op_code == OP_NOT_VSPACE;
      default:
-     return op_code == OP_VSPACE;
+     return op_code != OP_NOT_VSPACE;
      }
+ #ifdef SUPPORT_UCP
+   case OP_PROP:
+   return check_char_prop(next, previous[0], previous[1], FALSE);
+   case OP_NOTPROP:
+   return check_char_prop(next, previous[0], previous[1], TRUE);
+ #endif
    default:
    return FALSE;
    }
- /* Handle the case when the next item is \d, \s, etc. */
+ /* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP
+ is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are
+ generated only when PCRE_UCP is *not* set, that is, when only ASCII
+ characteristics are recognized. Similarly, the opcodes OP_DIGIT etc. are
+ replaced by OP_PROP codes when PCRE_UCP is set. */
  switch(op_code)
    {
    case OP_CHAR:
-   case OP_CHARNC:
+   case OP_CHARI:
  #ifdef SUPPORT_UTF8
-   if (utf8 && item > 127) { GETCHAR(item, utf8_char); }
+   GETCHARTEST(c, previous);
+ #else
+   c = *previous;
  #endif
    switch(-next)
      {
      case ESC_d:
-     return item > 127 || (cd->ctypes[item] & ctype_digit) == 0;
+     return c > 127 || (cd->ctypes[c] & ctype_digit) == 0;
      case ESC_D:
-     return item <= 127 && (cd->ctypes[item] & ctype_digit) != 0;
+     return c <= 127 && (cd->ctypes[c] & ctype_digit) != 0;
      case ESC_s:
-     return item > 127 || (cd->ctypes[item] & ctype_space) == 0;
+     return c > 127 || (cd->ctypes[c] & ctype_space) == 0;
      case ESC_S:
-     return item <= 127 && (cd->ctypes[item] & ctype_space) != 0;
+     return c <= 127 && (cd->ctypes[c] & ctype_space) != 0;
      case ESC_w:
-     return item > 127 || (cd->ctypes[item] & ctype_word) == 0;
+     return c > 127 || (cd->ctypes[c] & ctype_word) == 0;
      case ESC_W:
-     return item <= 127 && (cd->ctypes[item] & ctype_word) != 0;
+     return c <= 127 && (cd->ctypes[c] & ctype_word) != 0;
      case ESC_h:
      case ESC_H:
-     switch(item)
+     switch(c)
        {
        case 0x09:
        case 0x20:
-Line 2567 
 switch(op_code)
+Line 3073 
 switch(op_code)
      case ESC_v:
      case ESC_V:
-     switch(item)
+     switch(c)
        {
        case 0x0a:
        case 0x0b:
-Line 2581 
 switch(op_code)
+Line 3087 
 switch(op_code)
        return -next == ESC_v;
        }
+     /* When PCRE_UCP is set, these values get generated for \d etc. Find
+     their substitutions and process them. The result will always be either
+     -ESC_p or -ESC_P. Then fall through to process those values. */
+ #ifdef SUPPORT_UCP
+     case ESC_du:
+     case ESC_DU:
+     case ESC_wu:
+     case ESC_WU:
+     case ESC_su:
+     case ESC_SU:
+       {
+       int temperrorcode = 0;
+       ptr = substitutes[-next - ESC_DU];
+       next = check_escape(&ptr, &temperrorcode, 0, options, FALSE);
+       if (temperrorcode != 0) return FALSE;
+       ptr++;    /* For compatibility */
+       }
+     /* Fall through */
+     case ESC_p:
+     case ESC_P:
+       {
+       int ptype, pdata, errorcodeptr;
+       BOOL negated;
+       ptr--;      /* Make ptr point at the p or P */
+       ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr);
+       if (ptype < 0) return FALSE;
+       ptr++;      /* Point past the final curly ket */
+       /* If the property item is optional, we have to give up. (When generated
+       from \d etc by PCRE_UCP, this test will have been applied much earlier,
+       to the original \d etc. At this point, ptr will point to a zero byte. */
+       if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK ||
+         strncmp((char *)ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0)
+           return FALSE;
+       /* Do the property check. */
+       return check_char_prop(c, ptype, pdata, (next == -ESC_P) != negated);
+       }
+ #endif
      default:
      return FALSE;
      }
+   /* In principle, support for Unicode properties should be integrated here as
+   well. It means re-organizing the above code so as to get hold of the property
+   values before switching on the op-code. However, I wonder how many patterns
+   combine ASCII \d etc with Unicode properties? (Note that if PCRE_UCP is set,
+   these op-codes are never generated.) */
    case OP_DIGIT:
    return next == -ESC_D || next == -ESC_s || next == -ESC_W ||
-          next == -ESC_h || next == -ESC_v;
+          next == -ESC_h || next == -ESC_v || next == -ESC_R;
    case OP_NOT_DIGIT:
    return next == -ESC_d;
    case OP_WHITESPACE:
-   return next == -ESC_S || next == -ESC_d || next == -ESC_w;
+   return next == -ESC_S || next == -ESC_d || next == -ESC_w || next == -ESC_R;
    case OP_NOT_WHITESPACE:
    return next == -ESC_s || next == -ESC_h || next == -ESC_v;
    case OP_HSPACE:
-   return next == -ESC_S || next == -ESC_H || next == -ESC_d || next == -ESC_w;
+   return next == -ESC_S || next == -ESC_H || next == -ESC_d ||
+          next == -ESC_w || next == -ESC_v || next == -ESC_R;
    case OP_NOT_HSPACE:
    return next == -ESC_h;
    /* Can't have \S in here because VT matches \S (Perl anomaly) */
+   case OP_ANYNL:
    case OP_VSPACE:
    return next == -ESC_V || next == -ESC_d || next == -ESC_w;
    case OP_NOT_VSPACE:
-   return next == -ESC_v;
+   return next == -ESC_v || next == -ESC_R;
    case OP_WORDCHAR:
-   return next == -ESC_W || next == -ESC_s || next == -ESC_h || next == -ESC_v;
+   return next == -ESC_W || next == -ESC_s || next == -ESC_h ||
+          next == -ESC_v || next == -ESC_R;
    case OP_NOT_WORDCHAR:
    return next == -ESC_w || next == -ESC_d;
-Line 2644 
 Arguments:
+Line 3204 
 Arguments:
    firstbyteptr   set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE)
    reqbyteptr     set to the last literal character required, else < 0
    bcptr          points to current branch chain
+   cond_depth     conditional nesting depth
    cd             contains pointers to tables etc.
    lengthptr      NULL during the real compile phase
                   points to length accumulator during pre-compile phase
-Line 2655 
 Returns:         TRUE on success
+Line 3216 
 Returns:         TRUE on success
  static BOOL
  compile_branch(int *optionsptr, uschar **codeptr, const uschar **ptrptr,
    int *errorcodeptr, int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr,
-   compile_data *cd, int *lengthptr)
+   int cond_depth, compile_data *cd, int *lengthptr)
  {
  int repeat_type, op_type;
  int repeat_min = 0, repeat_max = 0;      /* To please picky compilers */
-Line 2664 
 int greedy_default, greedy_non_default;
+Line 3225 
 int greedy_default, greedy_non_default;
  int firstbyte, reqbyte;
  int zeroreqbyte, zerofirstbyte;
  int req_caseopt, reqvary, tempreqvary;
- int options = *optionsptr;
+ int options = *optionsptr;               /* May change dynamically */
  int after_manual_callout = 0;
  int length_prevgroup = 0;
  register int c;
-Line 2676 
 BOOL inescq = FALSE;
+Line 3237 
 BOOL inescq = FALSE;
  BOOL groupsetfirstbyte = FALSE;
  const uschar *ptr = *ptrptr;
  const uschar *tempptr;
+ const uschar *nestptr = NULL;
  uschar *previous = NULL;
  uschar *previous_callout = NULL;
  uschar *save_hwm = NULL;
  uschar classbits[32];
+ /* We can fish out the UTF-8 setting once and for all into a BOOL, but we
+ must not do this for other options (e.g. PCRE_EXTENDED) because they may change
+ dynamically as we process the pattern. */
  #ifdef SUPPORT_UTF8
  BOOL class_utf8;
  BOOL utf8 = (options & PCRE_UTF8) != 0;
-Line 2689 
 uschar *class_utf8data_base;
+Line 3255 
 uschar *class_utf8data_base;
  uschar utf8_char[6];
  #else
  BOOL utf8 = FALSE;
- uschar *utf8_char = NULL;
  #endif
  #ifdef PCRE_DEBUG
-Line 2740 
 for (;; ptr++)
+Line 3305 
 for (;; ptr++)
    int subfirstbyte;
    int terminator;
    int mclength;
+   int tempbracount;
    uschar mcbuffer[8];
    /* Get next byte in the pattern */
    c = *ptr;
+   /* If we are at the end of a nested substitution, revert to the outer level
+   string. Nesting only happens one level deep. */
+   if (c == 0 && nestptr != NULL)
+     {
+     ptr = nestptr;
+     nestptr = NULL;
+     c = *ptr;
+     }
    /* If we are in the pre-compile phase, accumulate the length used for the
    previous cycle of this loop. */
-Line 2754 
 for (;; ptr++)
+Line 3330 
 for (;; ptr++)
  #ifdef PCRE_DEBUG
      if (code > cd->hwm) cd->hwm = code;                 /* High water info */
  #endif
-     if (code > cd->start_workspace + COMPILE_WORK_SIZE) /* Check for overrun */
+     if (code > cd->start_workspace + WORK_SIZE_CHECK)   /* Check for overrun */
        {
        *errorcodeptr = ERR52;
        goto FAILED;
-Line 2776 
 for (;; ptr++)
+Line 3352 
 for (;; ptr++)
        goto FAILED;
        }
-     *lengthptr += code - last_code;
+     *lengthptr += (int)(code - last_code);
-     DPRINTF(("length=%d added %d c=%c\n", *lengthptr, code - last_code, c));
+     DPRINTF(("length=%d added %d c=%c\n", *lengthptr, (int)(code - last_code),
+       c));
      /* If "previous" is set and it is not at the start of the work space, move
      it back to there, in order to avoid filling up the work space. Otherwise,
-Line 2803 
 for (;; ptr++)
+Line 3380 
 for (;; ptr++)
    /* In the real compile phase, just check the workspace used by the forward
    reference list. */
-   else if (cd->hwm > cd->start_workspace + COMPILE_WORK_SIZE)
+   else if (cd->hwm > cd->start_workspace + WORK_SIZE_CHECK)
      {
      *errorcodeptr = ERR52;
      goto FAILED;
-Line 2851 
 for (;; ptr++)
+Line 3428 
 for (;; ptr++)
      previous_callout = NULL;
      }
-   /* In extended mode, skip white space and comments */
+   /* In extended mode, skip white space and comments. */
    if ((options & PCRE_EXTENDED) != 0)
      {
      if ((cd->ctypes[c] & ctype_space) != 0) continue;
      if (c == CHAR_NUMBER_SIGN)
        {
-       while (*(++ptr) != 0)
+       ptr++;
+       while (*ptr != 0)
          {
          if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; }
+         ptr++;
+ #ifdef SUPPORT_UTF8
+         if (utf8) while ((*ptr & 0xc0) == 0x80) ptr++;
+ #endif
          }
        if (*ptr != 0) continue;
-Line 2894 
 for (;; ptr++)
+Line 3476 
 for (;; ptr++)
          *errorcodeptr = ERR20;
          goto FAILED;
          }
-       *lengthptr += code - last_code;   /* To include callout length */
+       *lengthptr += (int)(code - last_code);   /* To include callout length */
        DPRINTF((">> end branch\n"));
        }
      return TRUE;
-Line 2905 
 for (;; ptr++)
+Line 3487 
 for (;; ptr++)
      the setting of any following char as a first character. */
      case CHAR_CIRCUMFLEX_ACCENT:
+     previous = NULL;
      if ((options & PCRE_MULTILINE) != 0)
        {
        if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;
+       *code++ = OP_CIRCM;
        }
-     previous = NULL;
+     else *code++ = OP_CIRC;
-     *code++ = OP_CIRC;
      break;
      case CHAR_DOLLAR_SIGN:
      previous = NULL;
-     *code++ = OP_DOLL;
+     *code++ = ((options & PCRE_MULTILINE) != 0)? OP_DOLLM : OP_DOLL;
      break;
      /* There can never be a first char if '.' is first, whatever happens about
-Line 3099 
 for (;; ptr++)
+Line 3682 
 for (;; ptr++)
            ptr++;
            }
-         posix_class = check_posix_name(ptr, tempptr - ptr);
+         posix_class = check_posix_name(ptr, (int)(tempptr - ptr));
          if (posix_class < 0)
            {
            *errorcodeptr = ERR30;
-Line 3113 
 for (;; ptr++)
+Line 3696 
 for (;; ptr++)
          if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)
            posix_class = 0;
-         /* We build the bit map for the POSIX class in a chunk of local store
+         /* When PCRE_UCP is set, some of the POSIX classes are converted to
-         because we may be adding and subtracting from it, and we don't want to
+         different escape sequences that use Unicode properties. */
-         subtract bits that may be in the main map already. At the end we or the
-         result into the bit map that is being built. */
+ #ifdef SUPPORT_UCP
+         if ((options & PCRE_UCP) != 0)
+           {
+           int pc = posix_class + ((local_negate)? POSIX_SUBSIZE/2 : 0);
+           if (posix_substitutes[pc] != NULL)
+             {
+             nestptr = tempptr + 1;
+             ptr = posix_substitutes[pc] - 1;
+             continue;
+             }
+           }
+ #endif
+         /* In the non-UCP case, we build the bit map for the POSIX class in a
+         chunk of local store because we may be adding and subtracting from it,
+         and we don't want to subtract bits that may be in the main map already.
+         At the end we or the result into the bit map that is being built. */
          posix_class *= 3;
-Line 3160 
 for (;; ptr++)
+Line 3758 
 for (;; ptr++)
        /* Backslash may introduce a single character, or it may introduce one
        of the specials, which just set a flag. The sequence \b is a special
-       case. Inside a class (and only there) it is treated as backspace.
+       case. Inside a class (and only there) it is treated as backspace. We
-       Elsewhere it marks a word boundary. Other escapes have preset maps ready
+       assume that other escapes have more than one character in them, so set
-       to 'or' into the one we are building. We assume they have more than one
+       class_charcount bigger than one. Unrecognized escapes fall through and
-       character in them, so set class_charcount bigger than one. */
+       are either treated as literal characters (by default), or are faulted if
+       PCRE_EXTRA is set. */
        if (c == CHAR_BACKSLASH)
          {
          c = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE);
          if (*errorcodeptr != 0) goto FAILED;
-         if (-c == ESC_b) c = CHAR_BS;       /* \b is backspace in a class */
+         if (-c == ESC_b) c = CHAR_BS;    /* \b is backspace in a class */
-         else if (-c == ESC_X) c = CHAR_X;   /* \X is literal X in a class */
-         else if (-c == ESC_R) c = CHAR_R;   /* \R is literal R in a class */
          else if (-c == ESC_Q)            /* Handle start of quoted string */
            {
            if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E)
-Line 3189 
 for (;; ptr++)
+Line 3786 
 for (;; ptr++)
            register const uschar *cbits = cd->cbits;
            class_charcount += 2;     /* Greater than 1 is what matters */
-           /* Save time by not doing this in the pre-compile phase. */
+           switch (-c)
-           if (lengthptr == NULL) switch (-c)
              {
+ #ifdef SUPPORT_UCP
+             case ESC_du:     /* These are the values given for \d etc */
+             case ESC_DU:     /* when PCRE_UCP is set. We replace the */
+             case ESC_wu:     /* escape sequence with an appropriate \p */
+             case ESC_WU:     /* or \P to test Unicode properties instead */
+             case ESC_su:     /* of the default ASCII testing. */
+             case ESC_SU:
+             nestptr = ptr;
+             ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */
+             class_charcount -= 2;                /* Undo! */
+             continue;
+ #endif
              case ESC_d:
              for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_digit];
              continue;
-Line 3211 
 for (;; ptr++)
+Line 3818 
 for (;; ptr++)
              for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word];
              continue;
+             /* Perl 5.004 onwards omits VT from \s, but we must preserve it
+             if it was previously set by something earlier in the character
+             class. */
              case ESC_s:
-             for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_space];
+             classbits[0] |= cbits[cbit_space];
-             classbits[1] &= ~0x08;   /* Perl 5.004 onwards omits VT from \s */
+             classbits[1] |= cbits[cbit_space+1] & ~0x08;
+             for (c = 2; c < 32; c++) classbits[c] |= cbits[c+cbit_space];
              continue;
              case ESC_S:
-Line 3222 
 for (;; ptr++)
+Line 3834 
 for (;; ptr++)
              classbits[1] |= 0x08;    /* Perl 5.004 onwards omits VT from \s */
              continue;
-             default:    /* Not recognized; fall through */
+             case ESC_h:
-             break;      /* Need "default" setting to stop compiler warning. */
-             }
-           /* In the pre-compile phase, just do the recognition. */
-           else if (c == -ESC_d || c == -ESC_D || c == -ESC_w ||
-                    c == -ESC_W || c == -ESC_s || c == -ESC_S) continue;
-           /* We need to deal with \H, \h, \V, and \v in both phases because
-           they use extra memory. */
-           if (-c == ESC_h)
-             {
              SETBIT(classbits, 0x09); /* VT */
              SETBIT(classbits, 0x20); /* SPACE */
              SETBIT(classbits, 0xa0); /* NSBP */
-Line 3259 
 for (;; ptr++)
+Line 3858 
 for (;; ptr++)
                }
  #endif
              continue;
-             }
-           if (-c == ESC_H)
+             case ESC_H:
-             {
              for (c = 0; c < 32; c++)
                {
                int x = 0xff;
-Line 3304 
 for (;; ptr++)
+Line 3901 
 for (;; ptr++)
                }
  #endif
              continue;
-             }
-           if (-c == ESC_v)
+             case ESC_v:
-             {
              SETBIT(classbits, 0x0a); /* LF */
              SETBIT(classbits, 0x0b); /* VT */
              SETBIT(classbits, 0x0c); /* FF */
-Line 3323 
 for (;; ptr++)
+Line 3918 
 for (;; ptr++)
                }
  #endif
              continue;
-             }
-           if (-c == ESC_V)
+             case ESC_V:
-             {
              for (c = 0; c < 32; c++)
                {
                int x = 0xff;
-Line 3356 
 for (;; ptr++)
+Line 3949 
 for (;; ptr++)
                }
  #endif
              continue;
-             }
-           /* We need to deal with \P and \p in both phases. */
  #ifdef SUPPORT_UCP
-           if (-c == ESC_p || -c == ESC_P)
+             case ESC_p:
-             {
+             case ESC_P:
-             BOOL negated;
+               {
-             int pdata;
+               BOOL negated;
-             int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);
+               int pdata;
-             if (ptype < 0) goto FAILED;
+               int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);
-             class_utf8 = TRUE;
+               if (ptype < 0) goto FAILED;
-             *class_utf8data++ = ((-c == ESC_p) != negated)?
+               class_utf8 = TRUE;
-               XCL_PROP : XCL_NOTPROP;
+               *class_utf8data++ = ((-c == ESC_p) != negated)?
-             *class_utf8data++ = ptype;
+                 XCL_PROP : XCL_NOTPROP;
-             *class_utf8data++ = pdata;
+               *class_utf8data++ = ptype;
-             class_charcount -= 2;   /* Not a < 256 character */
+               *class_utf8data++ = pdata;
-             continue;
+               class_charcount -= 2;   /* Not a < 256 character */
-             }
+               continue;
+               }
  #endif
-           /* Unrecognized escapes are faulted if PCRE is running in its
+             /* Unrecognized escapes are faulted if PCRE is running in its
-           strict mode. By default, for compatibility with Perl, they are
+             strict mode. By default, for compatibility with Perl, they are
-           treated as literals. */
+             treated as literals. */
-           if ((options & PCRE_EXTRA) != 0)
+             default:
-             {
+             if ((options & PCRE_EXTRA) != 0)
-             *errorcodeptr = ERR7;
+               {
-             goto FAILED;
+               *errorcodeptr = ERR7;
+               goto FAILED;
+               }
+             class_charcount -= 2;  /* Undo the default count from above */
+             c = *ptr;              /* Get the final character and fall through */
+             break;
              }
-           class_charcount -= 2;  /* Undo the default count from above */
-           c = *ptr;              /* Get the final character and fall through */
            }
          /* Fall through if we have a single character (c >= 0). This may be
-Line 3457 
 for (;; ptr++)
+Line 4050 
 for (;; ptr++)
            d = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE);
            if (*errorcodeptr != 0) goto FAILED;
-           /* \b is backspace; \X is literal X; \R is literal R; any other
+           /* \b is backspace; any other special means the '-' was literal */
-           special means the '-' was literal */
            if (d < 0)
              {
-             if (d == -ESC_b) d = CHAR_BS;
+             if (d == -ESC_b) d = CHAR_BS; else
-             else if (d == -ESC_X) d = CHAR_X;
-             else if (d == -ESC_R) d = CHAR_R; else
                {
                ptr = oldptr;
                goto LONE_SINGLE_CHARACTER;  /* A few lines below */
-Line 3630 
 for (;; ptr++)
+Line 4220 
 for (;; ptr++)
          }
        }
-     /* Loop until ']' reached. This "while" is the end of the "do" above. */
+     /* Loop until ']' reached. This "while" is the end of the "do" far above.
+     If we are at the end of an internal nested string, revert to the outer
+     string. */
+     while (((c = *(++ptr)) != 0 ||
+            (nestptr != NULL &&
+              (ptr = nestptr, nestptr = NULL, c = *(++ptr)) != 0)) &&
+            (c != CHAR_RIGHT_SQUARE_BRACKET || inescq));
-     while ((c = *(++ptr)) != 0 && (c != CHAR_RIGHT_SQUARE_BRACKET || inescq));
+     /* Check for missing terminating ']' */
-     if (c == 0)                          /* Missing terminating ']' */
+     if (c == 0)
        {
        *errorcodeptr = ERR6;
        goto FAILED;
        }
- /* This code has been disabled because it would mean that \s counts as
- an explicit \r or \n reference, and that's not really what is wanted. Now
- we set the flag only if there is a literal "\r" or "\n" in the class. */
- #if 0
-     /* Remember whether \r or \n are in this class */
-     if (negate_class)
-       {
-       if ((classbits[1] & 0x24) != 0x24) cd->external_flags |= PCRE_HASCRORLF;
-       }
-     else
-       {
-       if ((classbits[1] & 0x24) != 0) cd->external_flags |= PCRE_HASCRORLF;
-       }
- #endif
      /* If class_charcount is 1, we saw precisely one character whose value is
      less than 256. As long as there were no characters >= 128 and there was no
      use of \p or \P, in other words, no use of any XCLASS features, we can
-Line 3666 
 we set the flag only if there is a liter
+Line 4244 
 we set the flag only if there is a liter
      In UTF-8 mode, we can optimize the negative case only if there were no
      characters >= 128 because OP_NOT and the related opcodes like OP_NOTSTAR
-     operate on single-bytes only. This is an historical hangover. Maybe one day
+     operate on single-bytes characters only. This is an historical hangover.
-     we can tidy these opcodes to handle multi-byte characters.
+     Maybe one day we can tidy these opcodes to handle multi-byte characters.
      The optimization throws away the bit map. We turn the item into a
--character OP_CHAR[NC] if it's positive, or OP_NOT if it's negative. Note
+-character OP_CHAR[I] if it's positive, or OP_NOT[I] if it's negative.
-     that OP_NOT does not support multibyte characters. In the positive case, it
+     Note that OP_NOT[I] does not support multibyte characters. In the positive
-     can cause firstbyte to be set. Otherwise, there can be no first char if
+     case, it can cause firstbyte to be set. Otherwise, there can be no first
-     this item is first, whatever repeat count may follow. In the case of
+     char if this item is first, whatever repeat count may follow. In the case
-     reqbyte, save the previous value for reinstating. */
+     of reqbyte, save the previous value for reinstating. */
  #ifdef SUPPORT_UTF8
      if (class_charcount == 1 && !class_utf8 &&
-Line 3685 
 we set the flag only if there is a liter
+Line 4263 
 we set the flag only if there is a liter
        {
        zeroreqbyte = reqbyte;
-       /* The OP_NOT opcode works on one-byte characters only. */
+       /* The OP_NOT[I] opcodes work on one-byte characters only. */
        if (negate_class)
          {
          if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;
          zerofirstbyte = firstbyte;
-         *code++ = OP_NOT;
+         *code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT;
          *code++ = class_lastchar;
          break;
          }
-Line 3722 
 we set the flag only if there is a liter
+Line 4300 
 we set the flag only if there is a liter
      /* If there are characters with values > 255, we have to compile an
      extended class, with its own opcode, unless there was a negated special
-     such as \S in the class, because in that case all characters > 255 are in
+     such as \S in the class, and PCRE_UCP is not set, because in that case all
-     the class, so any that were explicitly given as well can be ignored. If
+     characters > 255 are in the class, so any that were explicitly given as
-     (when there are explicit characters > 255 that must be listed) there are no
+     well can be ignored. If (when there are explicit characters > 255 that must
-     characters < 256, we can omit the bitmap in the actual compiled code. */
+     be listed) there are no characters < 256, we can omit the bitmap in the
+     actual compiled code. */
  #ifdef SUPPORT_UTF8
-     if (class_utf8 && !should_flip_negation)
+     if (class_utf8 && (!should_flip_negation || (options & PCRE_UCP) != 0))
        {
        *class_utf8data++ = XCL_END;    /* Marks the end of extra data */
        *code++ = OP_XCLASS;
-Line 3754 
 we set the flag only if there is a liter
+Line 4333 
 we set the flag only if there is a liter
        }
  #endif
-     /* If there are no characters > 255, set the opcode to OP_CLASS or
+     /* If there are no characters > 255, or they are all to be included or
-     OP_NCLASS, depending on whether the whole class was negated and whether
+     excluded, set the opcode to OP_CLASS or OP_NCLASS, depending on whether the
-     there were negative specials such as \S in the class. Then copy the 32-byte
+     whole class was negated and whether there were negative specials such as \S
-     map into the code vector, negating it if necessary. */
+     (non-UCP) in the class. Then copy the 32-byte map into the code vector,
+     negating it if necessary. */
      *code++ = (negate_class == should_flip_negation) ? OP_CLASS : OP_NCLASS;
      if (negate_class)
-Line 3817 
 we set the flag only if there is a liter
+Line 4397 
 we set the flag only if there is a liter
      op_type = 0;                    /* Default single-char op codes */
      possessive_quantifier = FALSE;  /* Default not possessive quantifier */
-     /* Save start of previous item, in case we have to move it up to make space
+     /* Save start of previous item, in case we have to move it up in order to
-     for an inserted OP_ONCE for the additional '+' extension. */
+     insert something before it. */
      tempcode = previous;
-Line 3841 
 we set the flag only if there is a liter
+Line 4421 
 we set the flag only if there is a liter
        }
      else repeat_type = greedy_default;
+     /* If previous was a recursion call, wrap it in atomic brackets so that
+     previous becomes the atomic group. All recursions were so wrapped in the
+     past, but it no longer happens for non-repeated recursions. In fact, the
+     repeated ones could be re-implemented independently so as not to need this,
+     but for the moment we rely on the code for repeating groups. */
+     if (*previous == OP_RECURSE)
+       {
+       memmove(previous + 1 + LINK_SIZE, previous, 1 + LINK_SIZE);
+       *previous = OP_ONCE;
+       PUT(previous, 1, 2 + 2*LINK_SIZE);
+       previous[2 + 2*LINK_SIZE] = OP_KET;
+       PUT(previous, 3 + 2*LINK_SIZE, 2 + 2*LINK_SIZE);
+       code += 2 + 2 * LINK_SIZE;
+       length_prevgroup = 3 + 3*LINK_SIZE;
+       /* When actually compiling, we need to check whether this was a forward
+       reference, and if so, adjust the offset. */
+       if (lengthptr == NULL && cd->hwm >= cd->start_workspace + LINK_SIZE)
+         {
+         int offset = GET(cd->hwm, -LINK_SIZE);
+         if (offset == previous + 1 - cd->start_code)
+           PUT(cd->hwm, -LINK_SIZE, offset + 1 + LINK_SIZE);
+         }
+       }
+     /* Now handle repetition for the different types of item. */
      /* If previous was a character match, abolish the item and generate a
      repeat item instead. If a char item has a minumum of more than one, ensure
      that it is set in reqbyte - it might not be if a sequence such as x{3} is
      the first thing in a branch because the x will have gone into firstbyte
      instead.  */
-     if (*previous == OP_CHAR || *previous == OP_CHARNC)
+     if (*previous == OP_CHAR || *previous == OP_CHARI)
        {
+       op_type = (*previous == OP_CHAR)? 0 : OP_STARI - OP_STAR;
        /* Deal with UTF-8 characters that take up more than one byte. It's
        easier to write this out separately than try to macrify it. Use c to
        hold the length of the character in bytes, plus 0x80 to flag that it's a
-Line 3881 
 we set the flag only if there is a liter
+Line 4492 
 we set the flag only if there is a liter
        if (!possessive_quantifier &&
            repeat_max < 0 &&
-           check_auto_possessive(*previous, c, utf8, utf8_char, ptr + 1,
+           check_auto_possessive(previous, utf8, ptr + 1, options, cd))
-             options, cd))
          {
          repeat_type = 0;    /* Force greedy */
          possessive_quantifier = TRUE;
-Line 3894 
 we set the flag only if there is a liter
+Line 4504 
 we set the flag only if there is a liter
      /* If previous was a single negated character ([^a] or similar), we use
      one of the special opcodes, replacing it. The code is shared with single-
      character repeats by setting opt_type to add a suitable offset into
-     repeat_type. We can also test for auto-possessification. OP_NOT is
+     repeat_type. We can also test for auto-possessification. OP_NOT and OP_NOTI
-     currently used only for single-byte chars. */
+     are currently used only for single-byte chars. */
-     else if (*previous == OP_NOT)
+     else if (*previous == OP_NOT || *previous == OP_NOTI)
        {
-       op_type = OP_NOTSTAR - OP_STAR;  /* Use "not" opcodes */
+       op_type = ((*previous == OP_NOT)? OP_NOTSTAR : OP_NOTSTARI) - OP_STAR;
        c = previous[1];
        if (!possessive_quantifier &&
            repeat_max < 0 &&
-           check_auto_possessive(OP_NOT, c, utf8, NULL, ptr + 1, options, cd))
+           check_auto_possessive(previous, utf8, ptr + 1, options, cd))
          {
          repeat_type = 0;    /* Force greedy */
          possessive_quantifier = TRUE;
-Line 3927 
 we set the flag only if there is a liter
+Line 4537 
 we set the flag only if there is a liter
        if (!possessive_quantifier &&
            repeat_max < 0 &&
-           check_auto_possessive(c, 0, utf8, NULL, ptr + 1, options, cd))
+           check_auto_possessive(previous, utf8, ptr + 1, options, cd))
          {
          repeat_type = 0;    /* Force greedy */
          possessive_quantifier = TRUE;
-Line 4096 
 we set the flag only if there is a liter
+Line 4706 
 we set the flag only if there is a liter
  #ifdef SUPPORT_UTF8
               *previous == OP_XCLASS ||
  #endif
-              *previous == OP_REF)
+              *previous == OP_REF ||
+              *previous == OP_REFI)
        {
        if (repeat_max == 0)
          {
-Line 4130 
 we set the flag only if there is a liter
+Line 4741 
 we set the flag only if there is a liter
        }
      /* If previous was a bracket group, we may have to replicate it in certain
-     cases. */
+     cases. Note that at this point we can encounter only the "basic" bracket
+     opcodes such as BRA and CBRA, as this is the place where they get converted
+     into the more special varieties such as BRAPOS and SBRA. A test for >=
+     OP_ASSERT and <= OP_COND includes ASSERT, ASSERT_NOT, ASSERTBACK,
+     ASSERTBACK_NOT, ONCE, BRA, CBRA, and COND. Originally, PCRE did not allow
+     repetition of assertions, but now it does, for Perl compatibility. */
-     else if (*previous == OP_BRA  || *previous == OP_CBRA ||
+     else if (*previous >= OP_ASSERT && *previous <= OP_COND)
-              *previous == OP_ONCE || *previous == OP_COND)
        {
        register int i;
-       int ketoffset = 0;
+       int len = (int)(code - previous);
-       int len = code - previous;
        uschar *bralink = NULL;
+       uschar *brazeroptr = NULL;
-       /* Repeating a DEFINE group is pointless */
+       /* Repeating a DEFINE group is pointless, but Perl allows the syntax, so
+       we just ignore the repeat. */
        if (*previous == OP_COND && previous[LINK_SIZE+1] == OP_DEF)
-         {
+         goto END_REPEAT;
-         *errorcodeptr = ERR55;
-         goto FAILED;
-         }
-       /* If the maximum repeat count is unlimited, find the end of the bracket
+       /* There is no sense in actually repeating assertions. The only potential
-       by scanning through from the start, and compute the offset back to it
+       use of repetition is in cases when the assertion is optional. Therefore,
-       from the current code pointer. There may be an OP_OPT setting following
+       if the minimum is greater than zero, just ignore the repeat. If the
-       the final KET, so we can't find the end just by going back from the code
+       maximum is not not zero or one, set it to 1. */
-       pointer. */
+       if (*previous < OP_ONCE)    /* Assertion */
-       if (repeat_max == -1)
+         {
-         {
+         if (repeat_min > 0) goto END_REPEAT;
-         register uschar *ket = previous;
+         if (repeat_max < 0 || repeat_max > 1) repeat_max = 1;
-         do ket += GET(ket, 1); while (*ket != OP_KET);
-         ketoffset = code - ket;
          }
        /* The case of a zero minimum is special because of the need to stick
-Line 4179 
 we set the flag only if there is a liter
+Line 4790 
 we set the flag only if there is a liter
          **   goto END_REPEAT;
          **   }
-         However, that fails when a group is referenced as a subroutine from
+         However, that fails when a group or a subgroup within it is referenced
-         elsewhere in the pattern, so now we stick in OP_SKIPZERO in front of it
+         as a subroutine from elsewhere in the pattern, so now we stick in
-         so that it is skipped on execution. As we don't have a list of which
+         OP_SKIPZERO in front of it so that it is skipped on execution. As we
-         groups are referenced, we cannot do this selectively.
+         don't have a list of which groups are referenced, we cannot do this
+         selectively.
          If the maximum is 1 or unlimited, we just have to stick in the BRAZERO
          and do no more at this point. However, we do need to adjust any
-Line 4202 
 we set the flag only if there is a liter
+Line 4814 
 we set the flag only if there is a liter
              *previous++ = OP_SKIPZERO;
              goto END_REPEAT;
              }
+           brazeroptr = previous;    /* Save for possessive optimizing */
            *previous++ = OP_BRAZERO + repeat_type;
            }
-Line 4226 
 we set the flag only if there is a liter
+Line 4839 
 we set the flag only if there is a liter
            /* We chain together the bracket offset fields that have to be
            filled in later when the ends of the brackets are reached. */
-           offset = (bralink == NULL)? 0 : previous - bralink;
+           offset = (bralink == NULL)? 0 : (int)(previous - bralink);
            bralink = previous;
            PUTINC(previous, 0, offset);
            }
-Line 4335 
 we set the flag only if there is a liter
+Line 4948 
 we set the flag only if there is a liter
              {
              int offset;
              *code++ = OP_BRA;
-             offset = (bralink == NULL)? 0 : code - bralink;
+             offset = (bralink == NULL)? 0 : (int)(code - bralink);
              bralink = code;
              PUTINC(code, 0, offset);
              }
-Line 4356 
 we set the flag only if there is a liter
+Line 4969 
 we set the flag only if there is a liter
          while (bralink != NULL)
            {
            int oldlinkoffset;
-           int offset = code - bralink + 1;
+           int offset = (int)(code - bralink + 1);
            uschar *bra = code - offset;
            oldlinkoffset = GET(bra, 1);
            bralink = (oldlinkoffset == 0)? NULL : bralink - oldlinkoffset;
-Line 4366 
 we set the flag only if there is a liter
+Line 4979 
 we set the flag only if there is a liter
            }
          }
-       /* If the maximum is unlimited, set a repeater in the final copy. We
+       /* If the maximum is unlimited, set a repeater in the final copy. For
-       can't just offset backwards from the current code point, because we
+       ONCE brackets, that's all we need to do. However, possessively repeated
-       don't know if there's been an options resetting after the ket. The
+       ONCE brackets can be converted into non-capturing brackets, as the
-       correct offset was computed above.
+       behaviour of (?:xx)++ is the same as (?>xx)++ and this saves having to
+       deal with possessive ONCEs specially.
-       Then, when we are doing the actual compile phase, check to see whether
-       this group is a non-atomic one that could match an empty string. If so,
+       Otherwise, when we are doing the actual compile phase, check to see
+       whether this group is one that could match an empty string. If so,
        convert the initial operator to the S form (e.g. OP_BRA -> OP_SBRA) so
-       that runtime checking can be done. [This check is also applied to
+       that runtime checking can be done. [This check is also applied to ONCE
-       atomic groups at runtime, but in a different way.] */
+       groups at runtime, but in a different way.]
+       Then, if the quantifier was possessive and the bracket is not a
+       conditional, we convert the BRA code to the POS form, and the KET code to
+       KETRPOS. (It turns out to be convenient at runtime to detect this kind of
+       subpattern at both the start and at the end.) The use of special opcodes
+       makes it possible to reduce greatly the stack usage in pcre_exec(). If
+       the group is preceded by OP_BRAZERO, convert this to OP_BRAPOSZERO.
+       Then, if the minimum number of matches is 1 or 0, cancel the possessive
+       flag so that the default action below, of wrapping everything inside
+       atomic brackets, does not happen. When the minimum is greater than 1,
+       there will be earlier copies of the group, and so we still have to wrap
+       the whole thing. */
        else
          {
-         uschar *ketcode = code - ketoffset;
+         uschar *ketcode = code - 1 - LINK_SIZE;
          uschar *bracode = ketcode - GET(ketcode, 1);
-         *ketcode = OP_KETRMAX + repeat_type;
-         if (lengthptr == NULL && *bracode != OP_ONCE)
+         /* Convert possessive ONCE brackets to non-capturing */
+         if ((*bracode == OP_ONCE || *bracode == OP_ONCE_NC) &&
+             possessive_quantifier) *bracode = OP_BRA;
+         /* For non-possessive ONCE brackets, all we need to do is to
+         set the KET. */
+         if (*bracode == OP_ONCE || *bracode == OP_ONCE_NC)
+           *ketcode = OP_KETRMAX + repeat_type;
+         /* Handle non-ONCE brackets and possessive ONCEs (which have been
+         converted to non-capturing above). */
+         else
            {
-           uschar *scode = bracode;
+           /* In the compile phase, check for empty string matching. */
-           do
+           if (lengthptr == NULL)
              {
-             if (could_be_empty_branch(scode, ketcode, utf8, cd))
+             uschar *scode = bracode;
+             do
                {
-               *bracode += OP_SBRA - OP_BRA;
+               if (could_be_empty_branch(scode, ketcode, utf8, cd))
-               break;
+                 {
+                 *bracode += OP_SBRA - OP_BRA;
+                 break;
+                 }
+               scode += GET(scode, 1);
                }
-             scode += GET(scode, 1);
+             while (*scode == OP_ALT);
              }
-           while (*scode == OP_ALT);
+           /* Handle possessive quantifiers. */
+           if (possessive_quantifier)
+             {
+             /* For COND brackets, we wrap the whole thing in a possessively
+             repeated non-capturing bracket, because we have not invented POS
+             versions of the COND opcodes. Because we are moving code along, we
+             must ensure that any pending recursive references are updated. */
+             if (*bracode == OP_COND || *bracode == OP_SCOND)
+               {
+               int nlen = (int)(code - bracode);
+               *code = OP_END;
+               adjust_recurse(bracode, 1 + LINK_SIZE, utf8, cd, save_hwm);
+               memmove(bracode + 1+LINK_SIZE, bracode, nlen);
+               code += 1 + LINK_SIZE;
+               nlen += 1 + LINK_SIZE;
+               *bracode = OP_BRAPOS;
+               *code++ = OP_KETRPOS;
+               PUTINC(code, 0, nlen);
+               PUT(bracode, 1, nlen);
+               }
+             /* For non-COND brackets, we modify the BRA code and use KETRPOS. */
+             else
+               {
+               *bracode += 1;              /* Switch to xxxPOS opcodes */
+               *ketcode = OP_KETRPOS;
+               }
+             /* If the minimum is zero, mark it as possessive, then unset the
+             possessive flag when the minimum is 0 or 1. */
+             if (brazeroptr != NULL) *brazeroptr = OP_BRAPOSZERO;
+             if (repeat_min < 2) possessive_quantifier = FALSE;
+             }
+           /* Non-possessive quantifier */
+           else *ketcode = OP_KETRMAX + repeat_type;
            }
          }
        }
-Line 4415 
 we set the flag only if there is a liter
+Line 5103 
 we set the flag only if there is a liter
        }
      /* If the character following a repeat is '+', or if certain optimization
-     tests above succeeded, possessive_quantifier is TRUE. For some of the
+     tests above succeeded, possessive_quantifier is TRUE. For some opcodes,
-     simpler opcodes, there is an special alternative opcode for this. For
+     there are special alternative opcodes for this case. For anything else, we
-     anything else, we wrap the entire repeated item inside OP_ONCE brackets.
+     wrap the entire repeated item inside OP_ONCE brackets. Logically, the '+'
-     The '+' notation is just syntactic sugar, taken from Sun's Java package,
+     notation is just syntactic sugar, taken from Sun's Java package, but the
-     but the special opcodes can optimize it a bit. The repeated item starts at
+     special opcodes can optimize it.
-     tempcode, not at previous, which might be the first part of a string whose
-     (former) last char we repeated.
+     Some (but not all) possessively repeated subpatterns have already been
+     completely handled in the code just above. For them, possessive_quantifier
+     is always FALSE at this stage.
+     Note that the repeated item starts at tempcode, not at previous, which
+     might be the first part of a string whose (former) last char we repeated.
      Possessifying an 'exact' quantifier has no effect, so we can ignore it. But
      an 'upto' may follow. We skip over an 'exact' item, and then test the
-Line 4444 
 we set the flag only if there is a liter
+Line 5137 
 we set the flag only if there is a liter
  #endif
          }
-       len = code - tempcode;
+       len = (int)(code - tempcode);
        if (len > 0) switch (*tempcode)
          {
          case OP_STAR:  *tempcode = OP_POSSTAR; break;
-Line 4452 
 we set the flag only if there is a liter
+Line 5145 
 we set the flag only if there is a liter
          case OP_QUERY: *tempcode = OP_POSQUERY; break;
          case OP_UPTO:  *tempcode = OP_POSUPTO; break;
-         case OP_TYPESTAR:  *tempcode = OP_TYPEPOSSTAR; break;
+         case OP_STARI:  *tempcode = OP_POSSTARI; break;
-         case OP_TYPEPLUS:  *tempcode = OP_TYPEPOSPLUS; break;
+         case OP_PLUSI:  *tempcode = OP_POSPLUSI; break;
-         case OP_TYPEQUERY: *tempcode = OP_TYPEPOSQUERY; break;
+         case OP_QUERYI: *tempcode = OP_POSQUERYI; break;
-         case OP_TYPEUPTO:  *tempcode = OP_TYPEPOSUPTO; break;
+         case OP_UPTOI:  *tempcode = OP_POSUPTOI; break;
          case OP_NOTSTAR:  *tempcode = OP_NOTPOSSTAR; break;
          case OP_NOTPLUS:  *tempcode = OP_NOTPOSPLUS; break;
          case OP_NOTQUERY: *tempcode = OP_NOTPOSQUERY; break;
          case OP_NOTUPTO:  *tempcode = OP_NOTPOSUPTO; break;
+         case OP_NOTSTARI:  *tempcode = OP_NOTPOSSTARI; break;
+         case OP_NOTPLUSI:  *tempcode = OP_NOTPOSPLUSI; break;
+         case OP_NOTQUERYI: *tempcode = OP_NOTPOSQUERYI; break;
+         case OP_NOTUPTOI:  *tempcode = OP_NOTPOSUPTOI; break;
+         case OP_TYPESTAR:  *tempcode = OP_TYPEPOSSTAR; break;
+         case OP_TYPEPLUS:  *tempcode = OP_TYPEPOSPLUS; break;
+         case OP_TYPEQUERY: *tempcode = OP_TYPEPOSQUERY; break;
+         case OP_TYPEUPTO:  *tempcode = OP_TYPEPOSUPTO; break;
          /* Because we are moving code along, we must ensure that any
          pending recursive references are updated. */
          default:
          *code = OP_END;
          adjust_recurse(tempcode, 1 + LINK_SIZE, utf8, cd, save_hwm);
-Line 4503 
 we set the flag only if there is a liter
+Line 5206 
 we set the flag only if there is a liter
      /* First deal with various "verbs" that can be introduced by '*'. */
-     if (*(++ptr) == CHAR_ASTERISK && (cd->ctypes[ptr[1]] & ctype_letter) != 0)
+     if (*(++ptr) == CHAR_ASTERISK &&
+          ((cd->ctypes[ptr[1]] & ctype_letter) != 0 || ptr[1] == ':'))
        {
        int i, namelen;
+       int arglen = 0;
        const char *vn = verbnames;
-       const uschar *name = ++ptr;
+       const uschar *name = ptr + 1;
+       const uschar *arg = NULL;
        previous = NULL;
        while ((cd->ctypes[*++ptr] & ctype_letter) != 0) {};
+       namelen = (int)(ptr - name);
+       /* It appears that Perl allows any characters whatsoever, other than
+       a closing parenthesis, to appear in arguments, so we no longer insist on
+       letters, digits, and underscores. */
        if (*ptr == CHAR_COLON)
          {
-         *errorcodeptr = ERR59;   /* Not supported */
+         arg = ++ptr;
-         goto FAILED;
+         while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++;
+         arglen = (int)(ptr - arg);
          }
        if (*ptr != CHAR_RIGHT_PARENTHESIS)
          {
          *errorcodeptr = ERR60;
          goto FAILED;
          }
-       namelen = ptr - name;
+       /* Scan the table of verb names */
        for (i = 0; i < verbcount; i++)
          {
          if (namelen == verbs[i].len &&
              strncmp((char *)name, vn, namelen) == 0)
            {
-           /* Check for open captures before ACCEPT */
+           /* Check for open captures before ACCEPT and convert it to
+           ASSERT_ACCEPT if in an assertion. */
            if (verbs[i].op == OP_ACCEPT)
              {
              open_capitem *oc;
+             if (arglen != 0)
+               {
+               *errorcodeptr = ERR59;
+               goto FAILED;
+               }
              cd->had_accept = TRUE;
              for (oc = cd->open_caps; oc != NULL; oc = oc->next)
                {
                *code++ = OP_CLOSE;
                PUT2INC(code, 0, oc->number);
                }
+             *code++ = (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT;
+             /* Do not set firstbyte after *ACCEPT */
+             if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;
              }
-           *code++ = verbs[i].op;
-           break;
+           /* Handle other cases with/without an argument */
+           else if (arglen == 0)
+             {
+             if (verbs[i].op < 0)   /* Argument is mandatory */
+               {
+               *errorcodeptr = ERR66;
+               goto FAILED;
+               }
+             *code = verbs[i].op;
+             if (*code++ == OP_THEN) cd->external_flags |= PCRE_HASTHEN;
+             }
+           else
+             {
+             if (verbs[i].op_arg < 0)   /* Argument is forbidden */
+               {
+               *errorcodeptr = ERR59;
+               goto FAILED;
+               }
+             *code = verbs[i].op_arg;
+             if (*code++ == OP_THEN_ARG) cd->external_flags |= PCRE_HASTHEN;
+             *code++ = arglen;
+             memcpy(code, arg, arglen);
+             code += arglen;
+             *code++ = 0;
+             }
+           break;  /* Found verb, exit loop */
            }
          vn += verbs[i].len + 1;
          }
-       if (i < verbcount) continue;
-       *errorcodeptr = ERR60;
+       if (i < verbcount) continue;    /* Successfully handled a verb */
+       *errorcodeptr = ERR60;          /* Verb not recognized */
        goto FAILED;
        }
-Line 4663 
 we set the flag only if there is a liter
+Line 5419 
 we set the flag only if there is a liter
                recno * 10 + *ptr - CHAR_0 : -1;
            ptr++;
            }
-         namelen = ptr - name;
+         namelen = (int)(ptr - name);
          if ((terminator > 0 && *ptr++ != terminator) ||
              *ptr++ != CHAR_RIGHT_PARENTHESIS)
-Line 4724 
 we set the flag only if there is a liter
+Line 5480 
 we set the flag only if there is a liter
          /* Search the pattern for a forward reference */
          else if ((i = find_parens(cd, name, namelen,
-                         (options & PCRE_EXTENDED) != 0)) > 0)
+                         (options & PCRE_EXTENDED) != 0, utf8)) > 0)
            {
            PUT2(code, 2+LINK_SIZE, i);
            code[1+LINK_SIZE]++;
-Line 4792 
 we set the flag only if there is a liter
+Line 5548 
 we set the flag only if there is a liter
          /* ------------------------------------------------------------ */
          case CHAR_EQUALS_SIGN:                 /* Positive lookahead */
          bravalue = OP_ASSERT;
+         cd->assert_depth += 1;
          ptr++;
          break;
-Line 4806 
 we set the flag only if there is a liter
+Line 5563 
 we set the flag only if there is a liter
            continue;
            }
          bravalue = OP_ASSERT_NOT;
+         cd->assert_depth += 1;
          break;
-Line 4815 
 we set the flag only if there is a liter
+Line 5573 
 we set the flag only if there is a liter
            {
            case CHAR_EQUALS_SIGN:               /* Positive lookbehind */
            bravalue = OP_ASSERTBACK;
+           cd->assert_depth += 1;
            ptr += 2;
            break;
            case CHAR_EXCLAMATION_MARK:          /* Negative lookbehind */
            bravalue = OP_ASSERTBACK_NOT;
+           cd->assert_depth += 1;
            ptr += 2;
            break;
-Line 4841 
 we set the flag only if there is a liter
+Line 5601 
 we set the flag only if there is a liter
          /* ------------------------------------------------------------ */
          case CHAR_C:                 /* Callout - may be followed by digits; */
-         previous_callout = code;  /* Save for later completion */
+         previous_callout = code;     /* Save for later completion */
-         after_manual_callout = 1; /* Skip one item before completing */
+         after_manual_callout = 1;    /* Skip one item before completing */
          *code++ = OP_CALLOUT;
            {
            int n = 0;
-Line 4859 
 we set the flag only if there is a liter
+Line 5619 
 we set the flag only if there is a liter
              goto FAILED;
              }
            *code++ = n;
-           PUT(code, 0, ptr - cd->start_pattern + 1);  /* Pattern offset */
+           PUT(code, 0, (int)(ptr - cd->start_pattern + 1)); /* Pattern offset */
-           PUT(code, LINK_SIZE, 0);                    /* Default length */
+           PUT(code, LINK_SIZE, 0);                          /* Default length */
            code += 2 * LINK_SIZE;
            }
          previous = NULL;
-Line 4893 
 we set the flag only if there is a liter
+Line 5653 
 we set the flag only if there is a liter
            name = ++ptr;
            while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++;
-           namelen = ptr - name;
+           namelen = (int)(ptr - name);
            /* In the pre-compile phase, just do a syntax check. */
-Line 5023 
 we set the flag only if there is a liter
+Line 5783 
 we set the flag only if there is a liter
          NAMED_REF_OR_RECURSE:
          name = ++ptr;
          while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++;
-         namelen = ptr - name;
+         namelen = (int)(ptr - name);
-         /* In the pre-compile phase, do a syntax check and set a dummy
+         /* In the pre-compile phase, do a syntax check. We used to just set
-         reference number. */
+         a dummy reference number, because it was not used in the first pass.
+         However, with the change of recursive back references to be atomic,
+         we have to look for the number so that this state can be identified, as
+         otherwise the incorrect length is computed. If it's not a backwards
+         reference, the dummy number will do. */
          if (lengthptr != NULL)
            {
+           const uschar *temp;
            if (namelen == 0)
              {
              *errorcodeptr = ERR62;
-Line 5045 
 we set the flag only if there is a liter
+Line 5811 
 we set the flag only if there is a liter
              *errorcodeptr = ERR48;
              goto FAILED;
              }
-           recno = 0;
+           /* The name table does not exist in the first pass, so we cannot
+           do a simple search as in the code below. Instead, we have to scan the
+           pattern to find the number. It is important that we scan it only as
+           far as we have got because the syntax of named subpatterns has not
+           been checked for the rest of the pattern, and find_parens() assumes
+           correct syntax. In any case, it's a waste of resources to scan
+           further. We stop the scan at the current point by temporarily
+           adjusting the value of cd->endpattern. */
+           temp = cd->end_pattern;
+           cd->end_pattern = ptr;
+           recno = find_parens(cd, name, namelen,
+             (options & PCRE_EXTENDED) != 0, utf8);
+           cd->end_pattern = temp;
+           if (recno < 0) recno = 0;    /* Forward ref; set dummy number */
            }
          /* In the real compile, seek the name in the table. We check the name
-Line 5070 
 we set the flag only if there is a liter
+Line 5851 
 we set the flag only if there is a liter
              }
            else if ((recno =                /* Forward back reference */
                      find_parens(cd, name, namelen,
-                       (options & PCRE_EXTENDED) != 0)) <= 0)
+                       (options & PCRE_EXTENDED) != 0, utf8)) <= 0)
              {
              *errorcodeptr = ERR15;
              goto FAILED;
-Line 5181 
 we set the flag only if there is a liter
+Line 5962 
 we set the flag only if there is a liter
              if (called == NULL)
                {
                if (find_parens(cd, NULL, recno,
-                     (options & PCRE_EXTENDED) != 0) < 0)
+                     (options & PCRE_EXTENDED) != 0, utf8) < 0)
                  {
                  *errorcodeptr = ERR15;
                  goto FAILED;
                  }
                /* Fudge the value of "called" so that when it is inserted as an
                offset below, what it actually inserted is the reference number
-               of the group. */
+               of the group. Then remember the forward reference. */
                called = cd->start_code + recno;
-               PUTINC(cd->hwm, 0, code + 2 + LINK_SIZE - cd->start_code);
+               PUTINC(cd->hwm, 0, (int)(code + 1 - cd->start_code));
                }
              /* If not a forward reference, and the subpattern is still open,
              this is a recursive call. We check to see if this is a left
-             recursion that could loop for ever, and diagnose that case. */
+             recursion that could loop for ever, and diagnose that case. We
+             must not, however, do this check if we are in a conditional
+             subpattern because the condition might be testing for recursion in
+             a pattern such as /(?(R)a+|(?R)b)/, which is perfectly valid.
+             Forever loops are also detected at runtime, so those that occur in
+             conditional subpatterns will be picked up then. */
-             else if (GET(called, 1) == 0 &&
+             else if (GET(called, 1) == 0 && cond_depth <= 0 &&
                       could_be_empty(called, code, bcptr, utf8, cd))
                {
                *errorcodeptr = ERR40;
-Line 5207 
 we set the flag only if there is a liter
+Line 5993 
 we set the flag only if there is a liter
                }
              }
-           /* Insert the recursion/subroutine item, automatically wrapped inside
+           /* Insert the recursion/subroutine item. */
-           "once" brackets. Set up a "previous group" length so that a
-           subsequent quantifier will work. */
-           *code = OP_ONCE;
-           PUT(code, 1, 2 + 2*LINK_SIZE);
-           code += 1 + LINK_SIZE;
            *code = OP_RECURSE;
-           PUT(code, 1, called - cd->start_code);
+           PUT(code, 1, (int)(called - cd->start_code));
-           code += 1 + LINK_SIZE;
-           *code = OP_KET;
-           PUT(code, 1, 2 + 2*LINK_SIZE);
            code += 1 + LINK_SIZE;
-           length_prevgroup = 3 + 3*LINK_SIZE;
            }
          /* Can't determine a first byte now */
-Line 5284 
 we set the flag only if there is a liter
+Line 6058 
 we set the flag only if there is a liter
          is necessary to ensure we correctly detect the start of the pattern in
          both phases.
-         If we are not at the pattern start, compile code to change the ims
+         If we are not at the pattern start, reset the greedy defaults and the
-         options if this setting actually changes any of them, and reset the
+         case value for firstbyte and reqbyte. */
-         greedy defaults and the case value for firstbyte and reqbyte. */
          if (*ptr == CHAR_RIGHT_PARENTHESIS)
            {
-Line 5297 
 we set the flag only if there is a liter
+Line 6070 
 we set the flag only if there is a liter
              }
            else
              {
-             if ((options & PCRE_IMS) != (newoptions & PCRE_IMS))
-               {
-               *code++ = OP_OPT;
-               *code++ = newoptions & PCRE_IMS;
-               }
              greedy_default = ((newoptions & PCRE_UNGREEDY) != 0);
              greedy_non_default = greedy_default ^ 1;
              req_caseopt = ((newoptions & PCRE_CASELESS) != 0)? REQ_CASELESS : 0;
              }
            /* Change options at this level, and pass them back for use
-           in subsequent branches. When not at the start of the pattern, this
+           in subsequent branches. */
-           information is also necessary so that a resetting item can be
-           compiled at the end of a group (if we are in a group). */
            *optionsptr = options = newoptions;
            previous = NULL;       /* This item can't be repeated */
-Line 5327 
 we set the flag only if there is a liter
+Line 6093 
 we set the flag only if there is a liter
          }     /* End of switch for character following (? */
        }       /* End of (? handling */
-     /* Opening parenthesis not followed by '?'. If PCRE_NO_AUTO_CAPTURE is set,
+     /* Opening parenthesis not followed by '*' or '?'. If PCRE_NO_AUTO_CAPTURE
-     all unadorned brackets become non-capturing and behave like (?:...)
+     is set, all unadorned brackets become non-capturing and behave like (?:...)
      brackets. */
      else if ((options & PCRE_NO_AUTO_CAPTURE) != 0)
-Line 5346 
 we set the flag only if there is a liter
+Line 6112 
 we set the flag only if there is a liter
        skipbytes = 2;
        }
-     /* Process nested bracketed regex. Assertions may not be repeated, but
+     /* Process nested bracketed regex. Assertions used not to be repeatable,
-     other kinds can be. All their opcodes are >= OP_ONCE. We copy code into a
+     but this was changed for Perl compatibility, so all kinds can now be
-     non-register variable in order to be able to pass its address because some
+     repeated. We copy code into a non-register variable (tempcode) in order to
-     compilers complain otherwise. Pass in a new setting for the ims options if
+     be able to pass its address because some compilers complain otherwise. */
-     they have changed. */
-     previous = (bravalue >= OP_ONCE)? code : NULL;
+     previous = code;                      /* For handling repetition */
      *code = bravalue;
      tempcode = code;
-     tempreqvary = cd->req_varyopt;     /* Save value before bracket */
+     tempreqvary = cd->req_varyopt;        /* Save value before bracket */
-     length_prevgroup = 0;              /* Initialize for pre-compile phase */
+     tempbracount = cd->bracount;          /* Save value before bracket */
+     length_prevgroup = 0;                 /* Initialize for pre-compile phase */
      if (!compile_regex(
-          newoptions,                   /* The complete new option state */
+          newoptions,                      /* The complete new option state */
-          options & PCRE_IMS,           /* The previous ims option state */
+          &tempcode,                       /* Where to put code (updated) */
-          &tempcode,                    /* Where to put code (updated) */
+          &ptr,                            /* Input pointer (updated) */
-          &ptr,                         /* Input pointer (updated) */
+          errorcodeptr,                    /* Where to put an error message */
-          errorcodeptr,                 /* Where to put an error message */
           (bravalue == OP_ASSERTBACK ||
            bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */
-          reset_bracount,               /* True if (?| group */
+          reset_bracount,                  /* True if (?| group */
-          skipbytes,                    /* Skip over bracket number */
+          skipbytes,                       /* Skip over bracket number */
-          &subfirstbyte,                /* For possible first char */
+          cond_depth +
-          &subreqbyte,                  /* For possible last char */
+            ((bravalue == OP_COND)?1:0),   /* Depth of condition subpatterns */
-          bcptr,                        /* Current branch chain */
+          &subfirstbyte,                   /* For possible first char */
-          cd,                           /* Tables block */
+          &subreqbyte,                     /* For possible last char */
-          (lengthptr == NULL)? NULL :   /* Actual compile phase */
+          bcptr,                           /* Current branch chain */
-            &length_prevgroup           /* Pre-compile phase */
+          cd,                              /* Tables block */
+          (lengthptr == NULL)? NULL :      /* Actual compile phase */
+            &length_prevgroup              /* Pre-compile phase */
           ))
        goto FAILED;
+     /* If this was an atomic group and there are no capturing groups within it,
+     generate OP_ONCE_NC instead of OP_ONCE. */
+     if (bravalue == OP_ONCE && cd->bracount <= tempbracount)
+       *code = OP_ONCE_NC;
+     if (bravalue >= OP_ASSERT && bravalue <= OP_ASSERTBACK_NOT)
+       cd->assert_depth -= 1;
      /* At the end of compiling, code is still pointing to the start of the
-     group, while tempcode has been updated to point past the end of the group
+     group, while tempcode has been updated to point past the end of the group.
-     and any option resetting that may follow it. The pattern pointer (ptr)
+     The pattern pointer (ptr) is on the bracket.
-     is on the bracket. */
-     /* If this is a conditional bracket, check that there are no more than
+     If this is a conditional bracket, check that there are no more than
      two branches in the group, or just one if it's a DEFINE group. We do this
      in the real compile phase, not in the pre-pass, where the whole group may
      not be available. */
-Line 5447 
 we set the flag only if there is a liter
+Line 6222 
 we set the flag only if there is a liter
          goto FAILED;
          }
        *lengthptr += length_prevgroup - 2 - 2*LINK_SIZE;
-       *code++ = OP_BRA;
+       code++;   /* This already contains bravalue */
        PUTINC(code, 0, 1 + LINK_SIZE);
        *code++ = OP_KET;
        PUTINC(code, 0, 1 + LINK_SIZE);
-Line 5520 
 we set the flag only if there is a liter
+Line 6295 
 we set the flag only if there is a liter
      /* ===================================================================*/
      /* Handle metasequences introduced by \. For ones like \d, the ESC_ values
-     are arranged to be the negation of the corresponding OP_values. For the
+     are arranged to be the negation of the corresponding OP_values in the
-     back references, the values are ESC_REF plus the reference number. Only
+     default case when PCRE_UCP is not set. For the back references, the values
-     back references and those types that consume a character may be repeated.
+     are ESC_REF plus the reference number. Only back references and those types
-     We can test for values between ESC_b and ESC_Z for the latter; this may
+     that consume a character may be repeated. We can test for values between
-     have to change if any new ones are ever created. */
+     ESC_b and ESC_Z for the latter; this may have to change if any new ones are
+     ever created. */
      case CHAR_BACKSLASH:
      tempptr = ptr;
-Line 5614 
 we set the flag only if there is a liter
+Line 6390 
 we set the flag only if there is a liter
          }
        /* \k<name> or \k'name' is a back reference by name (Perl syntax).
-       We also support \k{name} (.NET syntax) */
+       We also support \k{name} (.NET syntax).  */
-       if (-c == ESC_k && (ptr[1] == CHAR_LESS_THAN_SIGN ||
+       if (-c == ESC_k)
-           ptr[1] == CHAR_APOSTROPHE || ptr[1] == CHAR_LEFT_CURLY_BRACKET))
          {
+         if ((ptr[1] != CHAR_LESS_THAN_SIGN &&
+           ptr[1] != CHAR_APOSTROPHE && ptr[1] != CHAR_LEFT_CURLY_BRACKET))
+           {
+           *errorcodeptr = ERR69;
+           break;
+           }
          is_recurse = FALSE;
          terminator = (*(++ptr) == CHAR_LESS_THAN_SIGN)?
            CHAR_GREATER_THAN_SIGN : (*ptr == CHAR_APOSTROPHE)?
-Line 5638 
 we set the flag only if there is a liter
+Line 6419 
 we set the flag only if there is a liter
          HANDLE_REFERENCE:    /* Come here from named backref handling */
          if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;
          previous = code;
-         *code++ = OP_REF;
+         *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF;
          PUT2INC(code, 0, recno);
          cd->backref_map |= (recno < 32)? (1 << recno) : 1;
          if (recno > cd->top_backref) cd->top_backref = recno;
-Line 5684 
 we set the flag only if there is a liter
+Line 6465 
 we set the flag only if there is a liter
  #endif
        /* For the rest (including \X when Unicode properties are supported), we
-       can obtain the OP value by negating the escape value. */
+       can obtain the OP value by negating the escape value in the default
+       situation when PCRE_UCP is not set. When it *is* set, we substitute
+       Unicode property tests. */
        else
          {
-         previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;
+ #ifdef SUPPORT_UCP
-         *code++ = -c;
+         if (-c >= ESC_DU && -c <= ESC_wu)
+           {
+           nestptr = ptr + 1;                   /* Where to resume */
+           ptr = substitutes[-c - ESC_DU] - 1;  /* Just before substitute */
+           }
+         else
+ #endif
+         /* In non-UTF-8 mode, we turn \C into OP_ALLANY instead of OP_ANYBYTE
+         so that it works in DFA mode and in lookbehinds. */
+           {
+           previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;
+           *code++ = (!utf8 && c == -ESC_C)? OP_ALLANY : -c;
+           }
          }
        continue;
        }
-Line 5734 
 we set the flag only if there is a liter
+Line 6530 
 we set the flag only if there is a liter
      ONE_CHAR:
      previous = code;
-     *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARNC : OP_CHAR;
+     *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR;
      for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];
      /* Remember if \r or \n were seen */
-Line 5763 
 we set the flag only if there is a liter
+Line 6559 
 we set the flag only if there is a liter
        else firstbyte = reqbyte = REQ_NONE;
        }
-     /* firstbyte was previously set; we can set reqbyte only the length is
+     /* firstbyte was previously set; we can set reqbyte only if the length is
 or the matching is caseful. */
      else
-Line 5798 
 return FALSE;
+Line 6594 
 return FALSE;
  /* On entry, ptr is pointing past the bracket character, but on return it
  points to the closing bracket, or vertical bar, or end of string. The code
  variable is pointing at the byte into which the BRA operator has been stored.
- If the ims options are changed at the start (for a (?ims: group) or during any
- branch, we need to insert an OP_OPT item at the start of every following branch
- to ensure they get set correctly at run time, and also pass the new options
- into every subsequent branch compile.
  This function is used during the pre-compile phase when we are trying to find
  out the amount of memory needed, as well as during the real compile phase. The
  value of lengthptr distinguishes the two phases.
  Arguments:
    options        option bits, including any changes for this subpattern
-   oldims         previous settings of ims option bits
    codeptr        -> the address of the current code pointer
    ptrptr         -> the address of the current pattern pointer
    errorcodeptr   -> pointer to error code variable
    lookbehind     TRUE if this is a lookbehind assertion
    reset_bracount TRUE to reset the count for each branch
    skipbytes      skip this many bytes at start (for brackets and OP_COND)
+   cond_depth     depth of nesting for conditional subpatterns
    firstbyteptr   place to put the first required character, or a negative number
    reqbyteptr     place to put the last required character, or a negative number
    bcptr          pointer to the chain of currently open branches
-Line 5827 
 Returns:         TRUE on success
+Line 6618 
 Returns:         TRUE on success
  */
  static BOOL
- compile_regex(int options, int oldims, uschar **codeptr, const uschar **ptrptr,
+ compile_regex(int options, uschar **codeptr, const uschar **ptrptr,
    int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes,
-   int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr, compile_data *cd,
+   int cond_depth, int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr,
-   int *lengthptr)
+   compile_data *cd, int *lengthptr)
  {
  const uschar *ptr = *ptrptr;
  uschar *code = *codeptr;
-Line 5844 
 int branchfirstbyte, branchreqbyte;
+Line 6635 
 int branchfirstbyte, branchreqbyte;
  int length;
  int orig_bracount;
  int max_bracount;
- int old_external_options = cd->external_options;
  branch_chain bc;
  bc.outer = bcptr;
-Line 5868 
 pre-compile phase to find out whether an
+Line 6658 
 pre-compile phase to find out whether an
  /* If this is a capturing subpattern, add to the chain of open capturing items
  so that we can detect them if (*ACCEPT) is encountered. This is also used to
- detect groups that contain recursive back references to themselves. */
+ detect groups that contain recursive back references to themselves. Note that
+ only OP_CBRA need be tested here; changing this opcode to one of its variants,
+ e.g. OP_SCBRAPOS, happens later, after the group has been compiled. */
  if (*code == OP_CBRA)
    {
-Line 5894 
 for (;;)
+Line 6686 
 for (;;)
    if (reset_bracount) cd->bracount = orig_bracount;
-   /* Handle a change of ims options at the start of the branch */
-   if ((options & PCRE_IMS) != oldims)
-     {
-     *code++ = OP_OPT;
-     *code++ = options & PCRE_IMS;
-     length += 2;
-     }
    /* Set up dummy OP_REVERSE if lookbehind assertion */
    if (lookbehind)
-Line 5917 
 for (;;)
+Line 6700 
 for (;;)
    into the length. */
    if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstbyte,
-         &branchreqbyte, &bc, cd, (lengthptr == NULL)? NULL : &length))
+         &branchreqbyte, &bc, cond_depth, cd,
+         (lengthptr == NULL)? NULL : &length))
      {
      *ptrptr = ptr;
      return FALSE;
      }
-   /* If the external options have changed during this branch, it means that we
-   are at the top level, and a leading option setting has been encountered. We
-   need to re-set the original option values to take account of this so that,
-   during the pre-compile phase, we know to allow for a re-set at the start of
-   subsequent branches. */
-   if (old_external_options != cd->external_options)
-     oldims = cd->external_options & PCRE_IMS;
    /* Keep the highest bracket count in case (?| was used and some branch
    has fewer than the rest. */
-Line 5992 
 for (;;)
+Line 6767 
 for (;;)
        {
        int fixed_length;
        *code = OP_END;
-       fixed_length = find_fixedlength(last_branch, options, FALSE, cd);
+       fixed_length = find_fixedlength(last_branch,  (options & PCRE_UTF8) != 0,
+         FALSE, cd);
        DPRINTF(("fixed length = %d\n", fixed_length));
        if (fixed_length == -3)
          {
-Line 6000 
 for (;;)
+Line 6776 
 for (;;)
          }
        else if (fixed_length < 0)
          {
-         *errorcodeptr = (fixed_length == -2)? ERR36 : ERR25;
+         *errorcodeptr = (fixed_length == -2)? ERR36 :
+                         (fixed_length == -4)? ERR70: ERR25;
          *ptrptr = ptr;
          return FALSE;
          }
-Line 6013 
 for (;;)
+Line 6790 
 for (;;)
    of offsets, with the field in the BRA item now becoming an offset to the
    first alternative. If there are no alternatives, it points to the end of the
    group. The length in the terminating ket is always the length of the whole
-   bracketed item. If any of the ims options were changed inside the group,
+   bracketed item. Return leaving the pointer at the terminating char. */
-   compile a resetting op-code following, except at the very end of the pattern.
-   Return leaving the pointer at the terminating char. */
    if (*ptr != CHAR_VERTICAL_LINE)
      {
      if (lengthptr == NULL)
        {
-       int branch_length = code - last_branch;
+       int branch_length = (int)(code - last_branch);
        do
          {
          int prev_length = GET(last_branch, 1);
-Line 6035 
 for (;;)
+Line 6810 
 for (;;)
      /* Fill in the ket */
      *code = OP_KET;
-     PUT(code, 1, code - start_bracket);
+     PUT(code, 1, (int)(code - start_bracket));
      code += 1 + LINK_SIZE;
      /* If it was a capturing subpattern, check to see if it contained any
-Line 6050 
 for (;;)
+Line 6825 
 for (;;)
            code - start_bracket);
          *start_bracket = OP_ONCE;
          code += 1 + LINK_SIZE;
-         PUT(start_bracket, 1, code - start_bracket);
+         PUT(start_bracket, 1, (int)(code - start_bracket));
          *code = OP_KET;
-         PUT(code, 1, code - start_bracket);
+         PUT(code, 1, (int)(code - start_bracket));
          code += 1 + LINK_SIZE;
          length += 2 + 2*LINK_SIZE;
          }
        cd->open_caps = cd->open_caps->next;
        }
-     /* Reset options if needed. */
-     if ((options & PCRE_IMS) != oldims && *ptr == CHAR_RIGHT_PARENTHESIS)
-       {
-       *code++ = OP_OPT;
-       *code++ = oldims;
-       length += 2;
-       }
      /* Retain the highest bracket number, in case resetting was used. */
      cd->bracount = max_bracount;
-Line 6107 
 for (;;)
+Line 6873 
 for (;;)
    else
      {
      *code = OP_ALT;
-     PUT(code, 1, code - last_branch);
+     PUT(code, 1, (int)(code - last_branch));
      bc.current_branch = last_branch = code;
      code += 1 + LINK_SIZE;
      }
-Line 6127 
 for (;;)
+Line 6893 
 for (;;)
  /* Try to find out if this is an anchored regular expression. Consider each
  alternative branch. If they all start with OP_SOD or OP_CIRC, or with a bracket
  all of whose alternatives start with OP_SOD or OP_CIRC (recurse ad lib), then
- it's anchored. However, if this is a multiline pattern, then only OP_SOD
+ it's anchored. However, if this is a multiline pattern, then only OP_SOD will
- counts, since OP_CIRC can match in the middle.
+ be found, because ^ generates OP_CIRCM in that mode.
  We can also consider a regex to be anchored if OP_SOM starts all its branches.
  This is the code for \G, which means "match at start of match position, taking
-Line 6149 
 of the more common cases more precisely.
+Line 6915 
 of the more common cases more precisely.
  Arguments:
    code           points to start of expression (the bracket)
-   options        points to the options setting
    bracket_map    a bitmap of which brackets we are inside while testing; this
                    handles up to substring 31; after that we just have to take
                    the less precise approach
-Line 6159 
 Returns:     TRUE or FALSE
+Line 6924 
 Returns:     TRUE or FALSE
  */
  static BOOL
- is_anchored(register const uschar *code, int *options, unsigned int bracket_map,
+ is_anchored(register const uschar *code, unsigned int bracket_map,
    unsigned int backref_map)
  {
  do {
     const uschar *scode = first_significant_code(code + _pcre_OP_lengths[*code],
-      options, PCRE_MULTILINE, FALSE);
+      FALSE);
     register int op = *scode;
     /* Non-capturing brackets */
-    if (op == OP_BRA)
+    if (op == OP_BRA  || op == OP_BRAPOS ||
+        op == OP_SBRA || op == OP_SBRAPOS)
       {
-      if (!is_anchored(scode, options, bracket_map, backref_map)) return FALSE;
+      if (!is_anchored(scode, bracket_map, backref_map)) return FALSE;
       }
     /* Capturing brackets */
-    else if (op == OP_CBRA)
+    else if (op == OP_CBRA  || op == OP_CBRAPOS ||
+             op == OP_SCBRA || op == OP_SCBRAPOS)
       {
       int n = GET2(scode, 1+LINK_SIZE);
       int new_map = bracket_map | ((n < 32)? (1 << n) : 1);
-      if (!is_anchored(scode, options, new_map, backref_map)) return FALSE;
+      if (!is_anchored(scode, new_map, backref_map)) return FALSE;
       }
     /* Other brackets */
-    else if (op == OP_ASSERT || op == OP_ONCE || op == OP_COND)
+    else if (op == OP_ASSERT || op == OP_ONCE || op == OP_ONCE_NC ||
+             op == OP_COND)
       {
-      if (!is_anchored(scode, options, bracket_map, backref_map)) return FALSE;
+      if (!is_anchored(scode, bracket_map, backref_map)) return FALSE;
       }
     /* .* is not anchored unless DOTALL is set (which generates OP_ALLANY) and
-Line 6202 
 do {
+Line 6970 
 do {
     /* Check for explicit anchoring */
-    else if (op != OP_SOD && op != OP_SOM &&
+    else if (op != OP_SOD && op != OP_SOM && op != OP_CIRC) return FALSE;
-            ((*options & PCRE_MULTILINE) != 0 || op != OP_CIRC))
-      return FALSE;
     code += GET(code, 1);
     }
  while (*code == OP_ALT);   /* Loop for each alternative */
-Line 6240 
 is_startline(const uschar *code, unsigne
+Line 7006 
 is_startline(const uschar *code, unsigne
  {
  do {
     const uschar *scode = first_significant_code(code + _pcre_OP_lengths[*code],
-      NULL, 0, FALSE);
+      FALSE);
     register int op = *scode;
     /* If we are at the start of a conditional assertion group, *both* the
-Line 6267 
 do {
+Line 7033 
 do {
         scode += 1 + LINK_SIZE;
         break;
         }
-      scode = first_significant_code(scode, NULL, 0, FALSE);
+      scode = first_significant_code(scode, FALSE);
       op = *scode;
       }
     /* Non-capturing brackets */
-    if (op == OP_BRA)
+    if (op == OP_BRA  || op == OP_BRAPOS ||
+        op == OP_SBRA || op == OP_SBRAPOS)
       {
       if (!is_startline(scode, bracket_map, backref_map)) return FALSE;
       }
     /* Capturing brackets */
-    else if (op == OP_CBRA)
+    else if (op == OP_CBRA  || op == OP_CBRAPOS ||
+             op == OP_SCBRA || op == OP_SCBRAPOS)
       {
       int n = GET2(scode, 1+LINK_SIZE);
       int new_map = bracket_map | ((n < 32)? (1 << n) : 1);
-Line 6289 
 do {
+Line 7057 
 do {
     /* Other brackets */
-    else if (op == OP_ASSERT || op == OP_ONCE)
+    else if (op == OP_ASSERT || op == OP_ONCE || op == OP_ONCE_NC)
       {
       if (!is_startline(scode, bracket_map, backref_map)) return FALSE;
       }
-Line 6304 
 do {
+Line 7072 
 do {
     /* Check for explicit circumflex */
-    else if (op != OP_CIRC) return FALSE;
+    else if (op != OP_CIRC && op != OP_CIRCM) return FALSE;
     /* Move on to the next alternative */
-Line 6330 
 we return that char, otherwise -1.
+Line 7098 
 we return that char, otherwise -1.
  Arguments:
    code       points to start of expression (the bracket)
-   options    pointer to the options (used to check casing changes)
    inassert   TRUE if in an assertion
  Returns:     -1 or the fixed first char
  */
  static int
- find_firstassertedchar(const uschar *code, int *options, BOOL inassert)
+ find_firstassertedchar(const uschar *code, BOOL inassert)
  {
  register int c = -1;
  do {
     int d;
-    const uschar *scode =
+    int xl = (*code == OP_CBRA || *code == OP_SCBRA ||
-      first_significant_code(code + 1+LINK_SIZE, options, PCRE_CASELESS, TRUE);
+              *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? 2:0;
+    const uschar *scode = first_significant_code(code + 1+LINK_SIZE + xl, TRUE);
     register int op = *scode;
     switch(op)
-Line 6352 
 do {
+Line 7120 
 do {
       return -1;
       case OP_BRA:
+      case OP_BRAPOS:
       case OP_CBRA:
+      case OP_SCBRA:
+      case OP_CBRAPOS:
+      case OP_SCBRAPOS:
       case OP_ASSERT:
       case OP_ONCE:
+      case OP_ONCE_NC:
       case OP_COND:
-      if ((d = find_firstassertedchar(scode, options, op == OP_ASSERT)) < 0)
+      if ((d = find_firstassertedchar(scode, op == OP_ASSERT)) < 0)
         return -1;
       if (c < 0) c = d; else if (c != d) return -1;
       break;
-      case OP_EXACT:       /* Fall through */
+      case OP_EXACT:
       scode += 2;
+      /* Fall through */
       case OP_CHAR:
-      case OP_CHARNC:
       case OP_PLUS:
       case OP_MINPLUS:
       case OP_POSPLUS:
       if (!inassert) return -1;
-      if (c < 0)
+      if (c < 0) c = scode[1];
-        {
+        else if (c != scode[1]) return -1;
-        c = scode[1];
+      break;
-        if ((*options & PCRE_CASELESS) != 0) c |= REQ_CASELESS;
-        }
+      case OP_EXACTI:
-      else if (c != scode[1]) return -1;
+      scode += 2;
+      /* Fall through */
+      case OP_CHARI:
+      case OP_PLUSI:
+      case OP_MINPLUSI:
+      case OP_POSPLUSI:
+      if (!inassert) return -1;
+      if (c < 0) c = scode[1] | REQ_CASELESS;
+        else if (c != scode[1]) return -1;
       break;
       }
-Line 6426 
 int length = 1;  /* For final END opcode
+Line 7208 
 int length = 1;  /* For final END opcode
  int firstbyte, reqbyte, newline;
  int errorcode = 0;
  int skipatstart = 0;
- BOOL utf8 = (options & PCRE_UTF8) != 0;
+ BOOL utf8;
  size_t size;
  uschar *code;
  const uschar *codestart;
-Line 6496 
 while (ptr[skipatstart] == CHAR_LEFT_PAR
+Line 7278 
 while (ptr[skipatstart] == CHAR_LEFT_PAR
    if (strncmp((char *)(ptr+skipatstart+2), STRING_UTF8_RIGHTPAR, 5) == 0)
      { skipatstart += 7; options |= PCRE_UTF8; continue; }
+   else if (strncmp((char *)(ptr+skipatstart+2), STRING_UCP_RIGHTPAR, 4) == 0)
+     { skipatstart += 6; options |= PCRE_UCP; continue; }
+   else if (strncmp((char *)(ptr+skipatstart+2), STRING_NO_START_OPT_RIGHTPAR, 13) == 0)
+     { skipatstart += 15; options |= PCRE_NO_START_OPTIMIZE; continue; }
    if (strncmp((char *)(ptr+skipatstart+2), STRING_CR_RIGHTPAR, 3) == 0)
      { skipatstart += 5; newnl = PCRE_NEWLINE_CR; }