*
 *---------------------------------------------------------------------------
 */

char *
Tcl_UniCharToUtfDString(
    const Tcl_UniChar *uniStr,	/* Unicode string to convert to UTF-8. */
    size_t uniLength,		/* Length of Unicode string in Tcl_UniChars
				 * (must be >= 0). */
    Tcl_DString *dsPtr)		/* UTF-8 representation of string is appended
				 * to this previously initialized DString. */
{
    const Tcl_UniChar *w, *wEnd;
    char *p, *string;
    size_t oldLength;
    int len = 1;

    /*
     * UTF-8 string length in bytes will be <= Unicode string length * 4.
     */

    oldLength = Tcl_DStringLength(dsPtr);
    Tcl_DStringSetLength(dsPtr, oldLength + (uniLength + 1) * 4);

 *
 *---------------------------------------------------------------------------
 */

Tcl_UniChar *
Tcl_UtfToUniCharDString(
    const char *src,		/* UTF-8 string to convert to Unicode. */
    size_t 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
				 * DString. */
{
    Tcl_UniChar ch = 0, *w, *wString;
    const char *p, *end;
    size_t oldLength;

    if (length == TCL_AUTO_LENGTH) {
	length = strlen(src);
    }

    /*
     * Unicode string length in Tcl_UniChars will be <= UTF-8 string length in
     * bytes.
     */

 *---------------------------------------------------------------------------
 */

int
Tcl_UtfCharComplete(
    const char *src,		/* String to check if first few bytes contain
				 * a complete UTF-8 character. */
    size_t length)			/* Length of above string in bytes. */
{
   return length >= totalBytes[(unsigned char)*src];
}

/*
 *---------------------------------------------------------------------------
 *
 * Tcl_NumUtfChars --
 *
 *	Returns the number of characters (not bytes) in the UTF-8 string, not
 *	including the terminating NULL byte. This is equivalent to Plan 9
 *	utflen() and utfnlen().
 *
 * Results:
 *	As above.
 *
 * Side effects:
 *	None.
 *
 *---------------------------------------------------------------------------
 */

size_t
Tcl_NumUtfChars(
    register const char *src,	/* The UTF-8 string to measure. */
    size_t length)			/* The length of the string in bytes, or -1
				 * for strlen(string). */
{
    Tcl_UniChar ch = 0;
    register size_t i = 0;

    /*
     * The separate implementations are faster.
     *
     * Since this is a time-sensitive function, we also do the check for the
     * single-byte char case specially.
     */

    if (length == TCL_AUTO_LENGTH) {
	while (*src != '\0') {
	    src += TclUtfToUniChar(src, &ch);
	    i++;
	}

    } else {
	register const char *endPtr = src + length - 4;

	while (src < endPtr) {
	    src += TclUtfToUniChar(src, &ch);
	    i++;
	}

 *
 *---------------------------------------------------------------------------
 */

int
Tcl_UniCharAtIndex(
    register const char *src,	/* The UTF-8 string to dereference. */
    register size_t index)		/* The position of the desired character. */
{
    Tcl_UniChar ch = 0;
    int fullchar = 0;
#if TCL_UTF_MAX <= 4
	int len = 1;
#endif

    src += TclUtfToUniChar(src, &ch);
    while (index--) {
#if TCL_UTF_MAX <= 4
	src += (len = TclUtfToUniChar(src, &ch));
#else
	src += TclUtfToUniChar(src, &ch);
#endif
    }
    fullchar = ch;

 *
 *---------------------------------------------------------------------------
 */

const char *
Tcl_UtfAtIndex(
    register const char *src,	/* The UTF-8 string. */
    register size_t index)		/* The position of the desired character. */
{
    Tcl_UniChar ch = 0;
#if TCL_UTF_MAX <= 4
    int len = 1;
#endif

    if (index != TCL_AUTO_LENGTH) {
	while (index--) {
#if TCL_UTF_MAX <= 4
	    src += (len = TclUtfToUniChar(src, &ch));
#else
	    src += TclUtfToUniChar(src, &ch);
#endif
	}
#if TCL_UTF_MAX <= 4
	if (!len) {
	    /* Index points at character following High Surrogate */
	    src += TclUtfToUniChar(src, &ch);
	}
#endif
    }
    return src;
}

/*
 *---------------------------------------------------------------------------
 *
 * Tcl_UtfBackslash --

 *	that represent the Unicode character is at least as large as the
 *	source buffer from which the backslashed sequence was extracted, no
 *	buffer overruns should occur.
 *
 *---------------------------------------------------------------------------
 */

size_t
Tcl_UtfBackslash(
    const char *src,		/* Points to the backslash character of a
				 * backslash sequence. */
    int *readPtr,		/* Fill in with number of characters read from
				 * src, unless NULL. */
    char *dst)			/* Filled with the bytes represented by the
				 * backslash sequence. */
{
#define LINE_LENGTH 128
    size_t numRead, result;


    result = TclParseBackslash(src, LINE_LENGTH, &numRead, dst);
    if (numRead == LINE_LENGTH) {
	/*
	 * We ate a whole line. Pay the price of a strlen()
	 */

	result = TclParseBackslash(src, strlen(src), &numRead, dst);
    }
    if (readPtr != NULL) {
	*readPtr = numRead;
    }
    return result;
}


 *----------------------------------------------------------------------
 */

int
TclpUtfNcmp2(
    const char *cs,		/* UTF string to compare to ct. */
    const char *ct,		/* UTF string cs is compared to. */
    size_t numBytes)	/* Number of *bytes* to compare. */
{
    /*
     * We can't simply call 'memcmp(cs, ct, numBytes);' because we need to
     * check for Tcl's \xC0\x80 non-utf-8 null encoding. Otherwise utf-8 lexes
     * fine in the strcmp manner.
     */

 *----------------------------------------------------------------------
 */

int
Tcl_UtfNcmp(
    const char *cs,		/* UTF string to compare to ct. */
    const char *ct,		/* UTF string cs is compared to. */
    size_t numChars)	/* Number of UTF chars to compare. */
{
    Tcl_UniChar ch1 = 0, ch2 = 0;

    /*
     * Cannot use 'memcmp(cs, ct, n);' as byte representation of \u0000 (the
     * pair of bytes 0xC0,0x80) is larger than byte representation of \u0001
     * (the byte 0x01.)

 *----------------------------------------------------------------------
 */

int
Tcl_UtfNcasecmp(
    const char *cs,		/* UTF string to compare to ct. */
    const char *ct,		/* UTF string cs is compared to. */
    size_t numChars)	/* Number of UTF chars to compare. */
{
    Tcl_UniChar ch1 = 0, ch2 = 0;

    while (numChars-- > 0) {
	/*
	 * n must be interpreted as chars, not bytes.
	 * This should be called only when both strings are of

 *
 * Side effects:
 *	None.
 *
 *----------------------------------------------------------------------
 */

size_t
Tcl_UniCharLen(
    const Tcl_UniChar *uniStr)	/* Unicode string to find length of. */
{
    size_t len = 0;

    while (*uniStr != '\0') {
	len++;
	uniStr++;
    }
    return len;
}

 *----------------------------------------------------------------------
 */

int
Tcl_UniCharNcmp(
    const Tcl_UniChar *ucs,	/* Unicode string to compare to uct. */
    const Tcl_UniChar *uct,	/* Unicode string ucs is compared to. */
    size_t numChars)	/* Number of unichars to compare. */
{
#ifdef WORDS_BIGENDIAN
    /*
     * We are definitely on a big-endian machine; memcmp() is safe
     */

    return memcmp(ucs, uct, numChars*sizeof(Tcl_UniChar));

 *----------------------------------------------------------------------
 */

int
Tcl_UniCharNcasecmp(
    const Tcl_UniChar *ucs,	/* Unicode string to compare to uct. */
    const Tcl_UniChar *uct,	/* Unicode string ucs is compared to. */
    size_t numChars)	/* Number of unichars to compare. */
{
    for ( ; numChars != 0; numChars--, ucs++, uct++) {
	if (*ucs != *uct) {
	    Tcl_UniChar lcs = Tcl_UniCharToLower(*ucs);
	    Tcl_UniChar lct = Tcl_UniCharToLower(*uct);

	    if (lcs != lct) {

 *
 *----------------------------------------------------------------------
 */

int
TclUniCharMatch(
    const Tcl_UniChar *string,	/* Unicode String. */
    size_t strLen,			/* Length of String */
    const Tcl_UniChar *pattern,	/* Pattern, which may contain special
				 * characters. */
    size_t ptnLen,			/* Length of Pattern */
    int nocase)			/* 0 for case sensitive, 1 for insensitive */
{
    const Tcl_UniChar *stringEnd, *patternEnd;
    Tcl_UniChar p;

    stringEnd = string + strLen;
    patternEnd = pattern + ptnLen;