} *ranges;
} CharSet;

/*
 * Declarations for functions used only in this file.
 */

#undef UtfToUniChar
static int		UtfToUniChar(const char *string, int *chPtr);
static const char *	BuildCharSet(CharSet *cset, const char *format);
static int		CharInSet(CharSet *cset, int ch);
static void		ReleaseCharSet(CharSet *cset);
static int		ValidateFormat(Tcl_Interp *interp, const char *format,
			    int numVars, int *totalVars);

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


static int
UtfToUniChar(
    const char *src,
    int *chPtr)
{
    Tcl_UniChar ch;
    int uch, len;

Tcl_UniCharCaseMatch(
    const Tcl_UniChar *uniStr,	/* Unicode String. */
    const Tcl_UniChar *uniPattern,
				/* Pattern, which may contain special
				 * characters. */
    int nocase)			/* 0 for case sensitive, 1 for insensitive */
{
#if TCL_UTF_MAX == 3
    int strLen = 0, ptnLen = 0;

    while (uniStr[strLen] != 0) {
	strLen++;
    }
    while (uniPattern[ptnLen] != 0) {
	ptnLen++;
    }
    return TclUniCharMatch(uniStr, strLen, uniPattern, ptnLen, nocase);
#else
    Tcl_UniChar ch1 = 0, p;

    while (1) {
	p = *uniPattern;

	/*
	 * See if we're at the end of both the pattern and the string. If so,

	    }
	} else if (*uniStr != *uniPattern) {
	    return 0;
	}
	uniStr++;
	uniPattern++;
    }
#endif
}

/*
 *----------------------------------------------------------------------
 *
 * TclUniCharMatch --
 *

    int strLen,			/* Length of String */
    const Tcl_UniChar *pattern,	/* Pattern, which may contain special
				 * characters. */
    int ptnLen,			/* Length of Pattern */
    int nocase)			/* 0 for case sensitive, 1 for insensitive */
{
    const Tcl_UniChar *stringEnd, *patternEnd;
    int p;
#if TCL_UTF_MAX == 3
    int q;
#endif

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

    while (1) {
	/*
	 * See if we're at the end of both the pattern and the string. If so,
	 * we succeeded. If we're at the end of the pattern but not at the end
	 * of the string, we failed.
	 */

	if (pattern == patternEnd) {
	    return (string == stringEnd);
	}
	p = *pattern;
#if TCL_UTF_MAX == 3
	if ((p & 0xFC00) == 0xD800) {
	    if ((pattern + 1 < patternEnd) &&
		    ((pattern[1] & 0xFC00) == 0xDC00)) {
		p = (((p&0x3FF)<<10) | (pattern[1]&0x3FF)) + 0x10000;
		++pattern;
	    }
	}
#endif
	if ((string == stringEnd) && (p != '*')) {
	    return 0;
	}

	/*
	 * Check for a "*" as the next pattern character. It matches any
	 * substring. We handle this by skipping all the characters up to the

	    while (*(++pattern) == '*') {
		/* empty body */
	    }
	    if (pattern == patternEnd) {
		return 1;
	    }
	    p = *pattern;
#if TCL_UTF_MAX == 3
	    if ((p & 0xFC00) == 0xD800) {
		if ((pattern + 1 < patternEnd) &&
			((pattern[1] & 0xFC00) == 0xDC00)) {
		    p = (((p&0x3FF)<<10) | (pattern[1]&0x3FF)) + 0x10000;
		    ++pattern;
		}
	    }
#endif

	    if (nocase) {
		p = Tcl_UniCharToLower(p);
	    }
	    while (1) {

		/*
		 * Optimization for matching - cruise through the string
		 * quickly if the next char in the pattern isn't a special
		 * character.
		 */

		if ((p != '[') && (p != '?') && (p != '\\')) {
#if TCL_UTF_MAX == 3
		    while (string < stringEnd) {
			q = *string;
			if ((q & 0xFC00) == 0xD800) {
			    if ((string + 1 < stringEnd) &&
				    ((string[1] & 0xFC00) == 0xDC00)) {
				q = (((q&0x3FF)<<10) | (string[1]&0x3FF))
					+ 0x10000;
			    }
			}
			if ((p == q) || (nocase &&
				(p == Tcl_UniCharToLower(q)))) {
			    break;
			}
			if (q > 0xFFFF) {
			    string++;
			}
			string++;
		    }
#else
		    if (nocase) {
			while ((string < stringEnd) && (p != *string)
				&& (p != Tcl_UniCharToLower(*string))) {
			    string++;
			}
		    } else {
			while ((string < stringEnd) && (p != *string)) {
			    string++;
			}
		    }
#endif
		}
		if (TclUniCharMatch(string, stringEnd - string,
			pattern, patternEnd - pattern, nocase)) {
		    return 1;
		}
		if (string == stringEnd) {
		    return 0;
		}
#if TCL_UTF_MAX == 3
		if ((string[0] & 0xFC00) == 0xD800) {
		    if ((string + 1 < stringEnd) &&
			    ((string[1] & 0xFC00) == 0xDC00)) {
			string++;
		    }
		}
#endif
		string++;
	    }
	}

	/*
	 * Check for a "?" as the next pattern character. It matches any
	 * single character.
	 */

	if (p == '?') {
	    pattern++;
#if TCL_UTF_MAX == 3
	    if ((string[0] & 0xFC00) == 0xD800) {
		if ((string + 1 < stringEnd) &&
			((string[1] & 0xFC00) == 0xDC00)) {
		    string++;
		}
	    }
#endif
	    string++;
	    continue;
	}

	/*
	 * Check for a "[" as the next pattern character. It is followed by a
	 * list of characters that are acceptable, or by a range (two
	 * characters separated by "-").
	 */

	if (p == '[') {
	    int ch1, startChar, endChar;

	    pattern++;
#if TCL_UTF_MAX == 3
	    ch1 = *string;
	    if ((ch1 & 0xFC00) == 0xD800) {
		if ((string + 1 < stringEnd) &&
			((string[1] & 0xFC00) == 0xDC00)) {
		    ch1 = (((ch1&0x3FF)<<10) | (string[1]&0x3FF)) + 0x10000;
		    string++;
		}
	    }
	    if (nocase) {
		ch1 = Tcl_UniCharToLower(ch1);
	    }
#else
	    ch1 = (nocase ? Tcl_UniCharToLower(*string) : *string);
#endif
	    string++;
	    while (1) {
		if ((*pattern == ']') || (pattern == patternEnd)) {
		    return 0;
		}
#if TCL_UTF_MAX == 3
		startChar = *pattern;
		if ((startChar & 0xFC00) == 0xD800) {
		    if ((pattern + 1 < patternEnd) &&
			    ((pattern[1] & 0xFC00) == 0xDC00)) {
			startChar = (((startChar&0x3FF)<<10) |
					(pattern[1]&0x3FF)) + 0x10000;
			pattern++;
		    }
		}
	 	if (nocase) {
		    startChar = Tcl_UniCharToLower(startChar);
		}
#else
		startChar = (nocase ? Tcl_UniCharToLower(*pattern) : *pattern);
#endif
		pattern++;
		if (*pattern == '-') {
		    pattern++;
		    if (pattern == patternEnd) {
			return 0;
		    }
#if TCL_UTF_MAX == 3
		    endChar = *pattern;
		    if ((endChar & 0xFC00) == 0xD800) {
			if ((pattern + 1 < patternEnd) &&
				((pattern[1] & 0xFC00) == 0xDC00)) {
			    endChar = (((endChar&0x3FF)<<10) |
					    (pattern[1]&0x3FF)) + 0x10000;
			    pattern++;
			}
		    }
		    if (nocase) {
			endChar = Tcl_UniCharToLower(endChar);
		    }
#else
		    endChar = (nocase ? Tcl_UniCharToLower(*pattern)
			    : *pattern);
#endif
		    pattern++;
		    if (((startChar <= ch1) && (ch1 <= endChar))
			    || ((endChar <= ch1) && (ch1 <= startChar))) {
			/*
			 * Matches ranges of form [a-z] or [z-a].
			 */
			break;
		    }
		} else if (startChar == ch1) {
		    break;
		}
	    }
	    while (*pattern != ']') {
		if (pattern == patternEnd) {
		    pattern--;
		    break;
		}
		pattern++;
	    }
#if TCL_UTF_MAX == 3
	    if ((pattern[0] & 0xFC00) == 0xD800) {
		if ((pattern + 1 < patternEnd) &&
			((pattern[1] & 0xFC00) == 0xDC00)) {
		    pattern++;
		}
	    }
#endif
	    pattern++;
	    continue;
	}

	/*
	 * If the next pattern character is '\', just strip off the '\' so we
	 * do exact matching on the character that follows.
	 */

	if (p == '\\') {
	    if (++pattern == patternEnd) {
		return 0;
	    }
	}

	/*
	 * There's no special character. Just make sure that the next bytes of
	 * each string match.
	 */
#if TCL_UTF_MAX == 3
	p = *pattern;
	if ((p & 0xFC00) == 0xD800) {
	    if ((pattern + 1 < patternEnd) &&
		    ((pattern[1] & 0xFC00) == 0xDC00)) {
		p = (((p&0x3FF)<<10) | (pattern[1]&0x3FF)) + 0x10000;
		pattern++;
	    }
	}
	q = *string;
	if ((q & 0xFC00) == 0xD800) {
	    if ((string + 1 < stringEnd) &&
		    ((string[1] & 0xFC00) == 0xDC00)) {
		q = (((q&0x3FF)<<10) | (string[1]&0x3FF)) + 0x10000;
		string++;
	    }
	}
	if (nocase) {
	    if (Tcl_UniCharToLower(q) != Tcl_UniCharToLower(p)) {
		return 0;
	    }
	} else if (q != p) {
	    return 0;
	}
#else
	if (nocase) {
	    if (Tcl_UniCharToLower(*string) != Tcl_UniCharToLower(*pattern)) {
		return 0;
	    }
	} else if (*string != *pattern) {
	    return 0;
	}
#endif
	string++;
	pattern++;
    }
}

/*
 * Local Variables:
 * mode: c
 * c-basic-offset: 4
 * fill-column: 78
 * End:
 */

static Tcl_ThreadDataKey precisionKey;

/*
 * Prototypes for functions defined later in this file.
 */

#undef UtfToUniChar
static int		UtfToUniChar(const char *string, int *chPtr);
static void		ClearHash(Tcl_HashTable *tablePtr);
static void		FreeProcessGlobalValue(ClientData clientData);
static void		FreeThreadHash(ClientData clientData);
static int		GetEndOffsetFromObj(Tcl_Obj *objPtr, int endValue,
			    int *indexPtr);
static Tcl_HashTable *	GetThreadHash(Tcl_ThreadDataKey *keyPtr);
static int		SetEndOffsetFromAny(Tcl_Interp *interp,

 * balance for the list as a whole, while the current implementation achieves
 * this by establishing brace balance for every element.
 *
 * Finally, a reminder that the rules for parsing and formatting lists are
 * closely tied together with the rules for parsing and evaluating scripts,
 * and will need to evolve in sync.
 */

/*
 *----------------------------------------------------------------------
 *
 * UtfToUniChar --
 *
 *	Wrapper to Tcl_UtfToUniChar() capable of dealing with
 *	surrogate pairs when compiled with TCL_UTF_MAX == 3.
 *
 * Results:
 *	*chPtr is filled with the full unicode character, and the
 *	return value is the number of bytes from the UTF-8 string that
 *	were consumed.
 *
 * Side effects:
 *	None.
 *
 *----------------------------------------------------------------------
 */

static int
UtfToUniChar(
    const char *src,
    int *chPtr)
{
    Tcl_UniChar ch;
    int uch, len;

    len = TclUtfToUniChar(src, &ch);
    uch = ch;
#if TCL_UTF_MAX == 3
    if ((ch & 0xFC00) == 0xD800) {
	int len2 = TclUtfToUniChar(src + len, &ch);

	if (len2 && ((ch & 0xFC00) == 0xDC00)) {
	    uch = ((uch & 0x3FF) << 10) + 0x10000 + (ch & 0x3FF);
	    len += len2;
	}
    }
#endif
    *chPtr = uch;
    return len;
}

/*
 *----------------------------------------------------------------------
 *
 * TclMaxListLength --
 *
 *	Given 'bytes' pointing to 'numBytes' bytes, scan through them and

Tcl_Backslash(
    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 buf[TCL_UTF_MAX*2];
    int ch = 0;

    buf[0] = '\0';
    Tcl_UtfBackslash(src, readPtr, buf);
    UtfToUniChar(buf, &ch);
    return (char) ch;
}

/*
 *----------------------------------------------------------------------
 *
 * TclTrimRight --

     * Make trim string into unicode array.
     */

slowPath:
    Tcl_DStringInit(&ds);
    q = trim;
    do {

	int uch = 0, qInc = UtfToUniChar(q, &uch);

	q += qInc;












	Tcl_DStringAppend(&ds, (char *) &uch, sizeof(int));
    } while (q < trim + numTrim);
    numTrim = Tcl_DStringLength(&ds) / sizeof(int);

    /*
     * Outer loop: iterate over string to be trimmed.
     */

			 * rely on (bytes[numBytes] == '\0'). */
    const char *trim,	/* String of trim characters... */
    int numTrim)	/* ...and its length in bytes */
			/* Calls to TclUtfToUniChar() in this routine
			 * rely on (trim[numTrim] == '\0'). */
{
    const char *p = bytes, *q;

    int i;
    Tcl_DString ds;

    /* Empty strings -> nothing to do */
    if ((numBytes == 0) || (numTrim == 0)) {
	return 0;
    }

     * Make trim string into unicode array.
     */

slowPath:
    Tcl_DStringInit(&ds);
    q = trim;
    do {

	int uch = 0, qInc = UtfToUniChar(q, &uch);

	q += qInc;












	Tcl_DStringAppend(&ds, (char *) &uch, sizeof(int));
    } while (q < trim + numTrim);
    numTrim = Tcl_DStringLength(&ds) / sizeof(int);

    /*
     * Outer loop: iterate over string to be trimmed.
     */

    do {
	int uch = 0, pInc = UtfToUniChar(p, &uch);
















	/*
	 * Inner loop: scan trim string for match to current character.
	 */

	for (i = 0; i < numTrim; i++) {
	    if (uch == ((int *)Tcl_DStringValue(&ds))[i]) {

	/* When bytes is NUL-terminated, returns 0 <= trimLeft <= numBytes */
	trimLeft = TclTrimLeft(bytes, numBytes, trim, numTrim);
	numBytes -= trimLeft;

	/* If we did not trim the whole string, it starts with a character
	 * that we will not trim. Skip over it. */
	if (numBytes > 0) {

	    int len, uch = 0;
	    const char *first = bytes + trimLeft;

	    len = UtfToUniChar(first, &uch);









	    bytes += len;
	    numBytes -= (bytes - first);

	    if (numBytes > 0) {
		/* When bytes is NUL-terminated, returns
		 * 0 <= trimRight <= numBytes */
		trimRight = TclTrimRight(bytes, numBytes, trim, numTrim);

int
Tcl_StringCaseMatch(
    const char *str,		/* String. */
    const char *pattern,	/* Pattern, which may contain special
				 * characters. */
    int nocase)			/* 0 for case sensitive, 1 for insensitive */
{
    int p, charLen, ch1 = 0, ch2 = 0;


    while (1) {
	p = *pattern;

	/*
	 * See if we're at the end of both the pattern and the string. If so,
	 * we succeeded. If we're at the end of the pattern but not at the end

	    }

	    /*
	     * This is a special case optimization for single-byte utf.
	     */

	    if (UCHAR(*pattern) < 0x80) {
		ch2 = (int)
			(nocase ? tolower(UCHAR(*pattern)) : UCHAR(*pattern));
	    } else {
		UtfToUniChar(pattern, &ch2);
		if (nocase) {
		    ch2 = Tcl_UniCharToLower(ch2);
		}
	    }

	    while (1) {
		/*
		 * Optimization for matching - cruise through the string
		 * quickly if the next char in the pattern isn't a special
		 * character
		 */

		if ((p != '[') && (p != '?') && (p != '\\')) {
		    if (nocase) {
			while (*str) {
			    charLen = UtfToUniChar(str, &ch1);
			    if (ch2==ch1 || ch2==Tcl_UniCharToLower(ch1)) {
				break;
			    }
			    str += charLen;
			}
		    } else {
			/*
			 * There's no point in trying to make this code
			 * shorter, as the number of bytes you want to compare
			 * each time is non-constant.
			 */

			while (*str) {
			    charLen = UtfToUniChar(str, &ch1);
			    if (ch2 == ch1) {
				break;
			    }
			    str += charLen;
			}
		    }
		}
		if (Tcl_StringCaseMatch(str, pattern, nocase)) {
		    return 1;
		}
		if (*str == '\0') {
		    return 0;
		}
		str += UtfToUniChar(str, &ch1);
	    }
	}

	/*
	 * Check for a "?" as the next pattern character. It matches any
	 * single character.
	 */

	if (p == '?') {
	    pattern++;
	    str += UtfToUniChar(str, &ch1);
	    continue;
	}

	/*
	 * Check for a "[" as the next pattern character. It is followed by a
	 * list of characters that are acceptable, or by a range (two
	 * characters separated by "-").
	 */

	if (p == '[') {
	    int startChar = 0, endChar = 0;

	    pattern++;
	    if (UCHAR(*str) < 0x80) {
		ch1 = (int)
			(nocase ? tolower(UCHAR(*str)) : UCHAR(*str));
		str++;
	    } else {
		str += UtfToUniChar(str, &ch1);
		if (nocase) {
		    ch1 = Tcl_UniCharToLower(ch1);
		}
	    }
	    while (1) {
		if ((*pattern == ']') || (*pattern == '\0')) {
		    return 0;
		}
		if (UCHAR(*pattern) < 0x80) {
		    startChar = (int) (nocase
			    ? tolower(UCHAR(*pattern)) : UCHAR(*pattern));
		    pattern++;
		} else {
		    pattern += UtfToUniChar(pattern, &startChar);
		    if (nocase) {
			startChar = Tcl_UniCharToLower(startChar);
		    }
		}
		if (*pattern == '-') {
		    pattern++;
		    if (*pattern == '\0') {
			return 0;
		    }
		    if (UCHAR(*pattern) < 0x80) {
			endChar = (int) (nocase
				? tolower(UCHAR(*pattern)) : UCHAR(*pattern));
			pattern++;
		    } else {
			pattern += UtfToUniChar(pattern, &endChar);
			if (nocase) {
			    endChar = Tcl_UniCharToLower(endChar);
			}
		    }
		    if (((startChar <= ch1) && (ch1 <= endChar))
			    || ((endChar <= ch1) && (ch1 <= startChar))) {
			/*

	}

	/*
	 * There's no special character. Just make sure that the next bytes of
	 * each string match.
	 */

	str += UtfToUniChar(str, &ch1);
	pattern += UtfToUniChar(pattern, &ch2);
	if (nocase) {
	    if (Tcl_UniCharToLower(ch1) != Tcl_UniCharToLower(ch2)) {
		return 0;
	    }
	} else if (ch1 != ch2) {
	    return 0;
	}

test split-1.16 {basic split commands} {
    split "\uD83D\uDE02Hello\uD83D\uDE02World\uD83D\uDE02" \U0001F602
} {{} Hello World {}}
test split-1.17 {basic split commands} {
    split "\U0001F602Hello\U0001F602World\U0001F602" \uD83D\uDE02
} {{} Hello World {}}
test split-1.18 {basic split commands} {
    split "\U0001F602\U0001F602\U0001F602" \uD83D\uDE02
} {{} {} {} {}}
test split-1.19 {basic split commands} {
    proc foo args {
        tailcall split {*}$args
    }
    foo "\U0001F602Hello\U0001F602World\U0001F602" \U0001F602
} {{} Hello World {}}

test split-2.1 {split errors} {

		[string match *a*l*\u0000*123 $longString] \
		[string match *a*l*\u0000*123* $longString] \
		[string match *a*l*\u0000*cba* $longString] \
		[string match *===* $longString]
    }
    foo
} {0 1 1 1 0 0}
test stringComp-11.55 {string match, unicode} {
    string match *\U1F602* Hello\U1F602World
} 1
test stringComp-11.56 {string match, unicode} {
    string match *\[\U1F602\]* Hello\U1F602World
} 1
test stringComp-11.57 {string match, unicode} {
    string match *\[\U1F602-\U1F604\]* Hello\U1F603World
} 1
test stringComp-11.58 {string match, unicode} {
    proc foo {p s} {
        return [string match $p $s]
    }
    list \
        [foo *\[\U1F602-\U1F604\]* Hello\uD83D\uDE03World] \
        [foo *\[\U1F602-\U1F604\]* Hello\uD83D\uDE05World] \
        [foo *\[\U1F602-\U1F604\]* Hello\uD83DWorld] \
        [foo *\[\U1F602-\U1F604\]* Hello\uDE02World\uDE04] \
        [foo *\[\U1F602-\U1F604\]* Hello\uD83DW\uDE03]
} {1 0 0 0 0}

## string range
test stringComp-12.1 {Bug 3588366: end-offsets before start} {
    apply {s {
	string range $s 0 end-5
    }} 12345
} {}