p += Tcl_UniCharToUtf(-1, p);
    }
    Tcl_DStringSetLength(dsPtr, oldLength + (p - string));

    return string;
}

#if (TCL_UTF_MAX > 4) && (defined(__CYGWIN__) || defined(_WIN32))
char *
Tcl_Utf16ToUtfDString(
    const unsigned short *uniStr,	/* Utf-16 string to convert to UTF-8. */
    int uniLength,		/* Length of Utf-16 string (must be >= 0). */
    Tcl_DString *dsPtr)		/* UTF-8 representation of string is appended
				 * to this previously initialized DString. */
{

	 * Handles properly formed UTF-8 characters between 0x01 and 0x7F.
	 * Treats naked trail bytes 0x80 to 0x9F as valid characters from
	 * the cp1252 table. See: <https://en.wikipedia.org/wiki/UTF-8>
	 * Also treats \0 and other naked trail bytes 0xA0 to 0xBF as valid
	 * characters representing themselves.
	 */

#if TCL_UTF_MAX <= 4
	/* If *chPtr contains a high surrogate (produced by a previous
	 * Tcl_UtfToUniChar() call) and the next 3 bytes are UTF-8 continuation
	 * bytes, then we must produce a follow-up low surrogate. We only
	 * do that if the high surrogate matches the bits we encounter.
	 */
	if ((byte >= 0x80)
		&& (((((byte - 0x10) << 2) & 0xFC) | 0xD800) == (*chPtr & 0xFCFC))

	 */
    }
    else if (byte < 0xF8) {
	if (((src[1] & 0xC0) == 0x80) && ((src[2] & 0xC0) == 0x80) && ((src[3] & 0xC0) == 0x80)) {
	    /*
	     * Four-byte-character lead byte followed by three trail bytes.
	     */
#if TCL_UTF_MAX <= 4
	    Tcl_UniChar high = (((byte & 0x07) << 8) | ((src[1] & 0x3F) << 2)
		    | ((src[2] & 0x3F) >> 4)) - 0x40;
	    if (high >= 0x400) {
		/* out of range, < 0x10000 or > 0x10ffff */
	    } else {
		/* produce high surrogate, advance source pointer */
		*chPtr = 0xD800 + high;

	 */
    }

    *chPtr = byte;
    return 1;
}

#if (TCL_UTF_MAX > 4) && (defined(__CYGWIN__) || defined(_WIN32))
int
TclUtfToUtf16(
    const char *src,	/* The UTF-8 string. */
    unsigned short *chPtr)/* Filled with the Utf-16 representation of
				 * the UTF-8 string. */
{
    unsigned short byte;

    *w = '\0';
    Tcl_DStringSetLength(dsPtr,
	    oldLength + ((char *) w - (char *) wString));

    return wString;
}

#if (TCL_UTF_MAX > 4) && (defined(__CYGWIN__) || defined(_WIN32))
unsigned short *
Tcl_UtfToUtf16DString(
    const char *src,		/* UTF-8 string to convert to Unicode. */
    int length,			/* Length of UTF-8 string in bytes, or -1 for
				 * strlen(). */
    Tcl_DString *dsPtr)		/* Unicode representation of string is
				 * appended to this previously initialized

{
    int len, fullchar;
    Tcl_UniChar find = 0;

    while (1) {
	len = TclUtfToUniChar(src, &find);
	fullchar = find;
#if TCL_UTF_MAX <= 4
	if ((ch >= 0xD800) && (len < 3)) {
	    len += TclUtfToUniChar(src + len, &find);
	    fullchar = (((fullchar & 0x3ff) << 10) | (find & 0x3ff)) + 0x10000;
	}
#endif
	if (fullchar == ch) {
	    return src;

    Tcl_UniChar find = 0;
    const char *last;

    last = NULL;
    while (1) {
	len = TclUtfToUniChar(src, &find);
	fullchar = find;
#if TCL_UTF_MAX <= 4
	if ((ch >= 0xD800) && (len < 3)) {
	    len += TclUtfToUniChar(src + len, &find);
	    fullchar = (((fullchar & 0x3ff) << 10) | (find & 0x3ff)) + 0x10000;
	}
#endif
	if (fullchar == ch) {
	    last = src;

const char *
Tcl_UtfNext(
    const char *src)		/* The current location in the string. */
{
    Tcl_UniChar ch = 0;
    int len = TclUtfToUniChar(src, &ch);

#if TCL_UTF_MAX <= 4
    if ((ch >= 0xD800) && (len < 3)) {
	len += TclUtfToUniChar(src + len, &ch);
    }
#endif
    return src + len;
}


int
Tcl_UniCharAtIndex(
    register const char *src,	/* The UTF-8 string to dereference. */
    register int index)		/* The position of the desired character. */
{
    Tcl_UniChar ch = 0;
    int fullchar = 0;
#if TCL_UTF_MAX <= 4
	int len = 0;
#endif

    while (index-- >= 0) {
#if TCL_UTF_MAX <= 4
	src += (len = TclUtfToUniChar(src, &ch));
#else
	src += TclUtfToUniChar(src, &ch);
#endif
    }
    fullchar = ch;
#if TCL_UTF_MAX <= 4
    if ((ch >= 0xD800) && (len < 3)) {
	/* If last Tcl_UniChar was a high surrogate, combine with low surrogate */
	(void)TclUtfToUniChar(src, &ch);
	fullchar = (((fullchar & 0x3ff) << 10) | (ch & 0x3ff)) + 0x10000;
    }
#endif
    return fullchar;
}

/*
 *---------------------------------------------------------------------------
 *
 * Tcl_UtfAtIndex --
 *
 *	Returns a pointer to the specified character (not byte) position in
 *	the UTF-8 string. If TCL_UTF_MAX <= 4, characters > U+FFFF count as
 *	2 positions, but then the pointer should never be placed between
 *	the two positions.
 *
 * Results:
 *	As above.
 *
 * Side effects:

    Tcl_UniChar ch = 0;
    int len = 0;

    while (index-- > 0) {
	len = TclUtfToUniChar(src, &ch);
	src += len;
    }
#if TCL_UTF_MAX <= 4
    if ((ch >= 0xD800) && (len < 3)) {
	/* Index points at character following high Surrogate */
	src += TclUtfToUniChar(src, &ch);
    }
#endif
    return src;
}

     * Iterate over the string until we hit the terminating null.
     */

    src = dst = str;
    while (*src) {
	len = TclUtfToUniChar(src, &ch);
	upChar = ch;
#if TCL_UTF_MAX <= 4
	if ((ch >= 0xD800) && (len < 3)) {
	    len += TclUtfToUniChar(src + len, &ch);
	    /* Combine surrogates */
	    upChar = (((upChar & 0x3ff) << 10) | (ch & 0x3ff)) + 0x10000;
	}
#endif
	upChar = Tcl_UniCharToUpper(upChar);

     * Iterate over the string until we hit the terminating null.
     */

    src = dst = str;
    while (*src) {
	len = TclUtfToUniChar(src, &ch);
	lowChar = ch;
#if TCL_UTF_MAX <= 4
	if ((ch >= 0xD800) && (len < 3)) {
	    len += TclUtfToUniChar(src + len, &ch);
	    /* Combine surrogates */
	    lowChar = (((lowChar & 0x3ff) << 10) | (ch & 0x3ff)) + 0x10000;
	}
#endif
	lowChar = Tcl_UniCharToLower(lowChar);

     */

    src = dst = str;

    if (*src) {
	len = TclUtfToUniChar(src, &ch);
	titleChar = ch;
#if TCL_UTF_MAX <= 4
	if ((ch >= 0xD800) && (len < 3)) {
	    len += TclUtfToUniChar(src + len, &ch);
	    /* Combine surrogates */
	    titleChar = (((titleChar & 0x3ff) << 10) | (ch & 0x3ff)) + 0x10000;
	}
#endif
	titleChar = Tcl_UniCharToTitle(titleChar);

	if ((len < TclUtfCount(titleChar)) || ((titleChar & 0xF800) == 0xD800)) {
	    memcpy(dst, src, len);
	    dst += len;
	} else {
	    dst += Tcl_UniCharToUtf(titleChar, dst);
	}
	src += len;
    }
    while (*src) {
	len = TclUtfToUniChar(src, &ch);
	lowChar = ch;
#if TCL_UTF_MAX <= 4
	if ((ch >= 0xD800) && (len < 3)) {
	    len += TclUtfToUniChar(src + len, &ch);
	    /* Combine surrogates */
	    lowChar = (((lowChar & 0x3ff) << 10) | (ch & 0x3ff)) + 0x10000;
	}
#endif
	/* Special exception for Georgian Asomtavruli chars, no titlecase. */

	 * only when both strings are of at least n chars long (no need for \0
	 * check)
	 */

	cs += TclUtfToUniChar(cs, &ch1);
	ct += TclUtfToUniChar(ct, &ch2);
	if (ch1 != ch2) {
#if TCL_UTF_MAX <= 4
	    /* Surrogates always report higher than non-surrogates */
	    if (((ch1 & 0xFC00) == 0xD800)) {
	    if ((ch2 & 0xFC00) != 0xD800) {
		return ch1;
	    }
	    } else if ((ch2 & 0xFC00) == 0xD800) {
		return -ch2;

	 * n must be interpreted as chars, not bytes.
	 * This should be called only when both strings are of
	 * at least n chars long (no need for \0 check)
	 */
	cs += TclUtfToUniChar(cs, &ch1);
	ct += TclUtfToUniChar(ct, &ch2);
	if (ch1 != ch2) {
#if TCL_UTF_MAX <= 4
	    /* Surrogates always report higher than non-surrogates */
	    if (((ch1 & 0xFC00) == 0xD800)) {
	    if ((ch2 & 0xFC00) != 0xD800) {
		return ch1;
	    }
	    } else if ((ch2 & 0xFC00) == 0xD800) {
		return -ch2;

{
    Tcl_UniChar ch1 = 0, ch2 = 0;

    while (*cs && *ct) {
	cs += TclUtfToUniChar(cs, &ch1);
	ct += TclUtfToUniChar(ct, &ch2);
	if (ch1 != ch2) {
#if TCL_UTF_MAX <= 4
	    /* Surrogates always report higher than non-surrogates */
	    if (((ch1 & 0xFC00) == 0xD800)) {
	    if ((ch2 & 0xFC00) != 0xD800) {
		return ch1;
	    }
	    } else if ((ch2 & 0xFC00) == 0xD800) {
		return -ch2;

{
    Tcl_UniChar ch1 = 0, ch2 = 0;

    while (*cs && *ct) {
	cs += TclUtfToUniChar(cs, &ch1);
	ct += TclUtfToUniChar(ct, &ch2);
	if (ch1 != ch2) {
#if TCL_UTF_MAX <= 4
	    /* Surrogates always report higher than non-surrogates */
	    if (((ch1 & 0xFC00) == 0xD800)) {
	    if ((ch2 & 0xFC00) != 0xD800) {
		return ch1;
	    }
	    } else if ((ch2 & 0xFC00) == 0xD800) {
		return -ch2;