static void		GrowUnicodeBuffer(Tcl_Obj *objPtr, size_t needed);
static int		SetStringFromAny(Tcl_Interp *interp, Tcl_Obj *objPtr);
static void		SetUnicodeObj(Tcl_Obj *objPtr,
			    const Tcl_UniChar *unicode, Tcl_Size numChars);
static Tcl_Size		UnicodeLength(const Tcl_UniChar *unicode);
static void		UpdateStringOfString(Tcl_Obj *objPtr);

#if TCL_UTF_MAX > 3
#define ISCONTINUATION(bytes) (\
	((bytes)[0] & 0xC0) == 0x80)
#else
#define ISCONTINUATION(bytes) (\
	((((bytes)[0] & 0xC0) == 0x80) || (((bytes)[0] == '\xED') \
	&& (((bytes)[1] & 0xF0) == 0xB0) && (((bytes)[2] & 0xC0) == 0x80))))
#endif


/*
 * The structure below defines the string Tcl object type by means of
 * functions that can be invoked by generic object code.
 */

const Tcl_ObjType tclStringType = {

    if (Tcl_IsShared(objPtr)) {
	Tcl_Panic("%s called with shared object", "Tcl_AppendLimitedToObj");
    }

    SetStringFromAny(NULL, objPtr);
    stringPtr = GET_STRING(objPtr);

    /* If appended string starts with a continuation byte or a lower surrogate,
     * force objPtr to unicode representation. See [7f1162a867] */
    if (bytes && ISCONTINUATION(bytes)) {
	Tcl_GetUnicode(objPtr);
	stringPtr = GET_STRING(objPtr);
    }
    if (stringPtr->hasUnicode && (stringPtr->numChars+1) > 1) {
	AppendUtfToUnicodeRep(objPtr, bytes, toCopy);
    } else {
	AppendUtfToUtfRep(objPtr, bytes, toCopy);
    }

    if (length <= limit) {

    /*
     * Must append as strings.
     */

    SetStringFromAny(NULL, objPtr);
    stringPtr = GET_STRING(objPtr);

    /* If appended string starts with a continuation byte or a lower surrogate,
     * force objPtr to unicode representation. See [7f1162a867]
     * This fixes append-3.4, append-3.7 and utf-1.18 testcases. */
    if (ISCONTINUATION(TclGetString(appendObjPtr))) {
	Tcl_GetUnicode(objPtr);
	stringPtr = GET_STRING(objPtr);
    }
    /*
     * If objPtr has a valid Unicode rep, then get a Unicode string from
     * appendObjPtr and append it.
     */

    if (stringPtr->hasUnicode) {
	/*

	    if (objPtr->length) {
		/*
		 * Non-empty string rep. Not a pure bytearray, so we won't
		 * create a pure bytearray.
		 */

	 	binary = 0;
	 	if (ov > objv+1 && ISCONTINUATION(TclGetString(objPtr))) {
	 	    forceUniChar = 1;
	 	} else if ((objPtr->typePtr) && (objPtr->typePtr != &tclStringType)) {
		    /* Prevent shimmer of non-string types. */
		    allowUniChar = 0;
		}
	    }
	} else {
	    /* assert (objPtr->typePtr != NULL) -- stork! */
	    binary = 0;

test append-3.6 {append surrogates} -body {
    set x \uDE02
    set x \uD83D$x
} -result \uD83D\uDE02
test append-3.7 {append \xC0 \x80} -constraints testbytestring -body {
    set x [testbytestring \xC0]
    string length [append x [testbytestring \x80]]
} -result 2
test append-3.8 {append \xC0 \x80} -constraints testbytestring -body {
    set x [testbytestring \xC0]
    string length $x[testbytestring \x80]
} -result 2
test append-3.9 {append \xC0 \x80} -constraints testbytestring -body {
    set x [testbytestring \x80]
    string length [testbytestring \xC0]$x
} -result 2
test append-3.10 {append surrogates} -body {
    set x \uD83D
    string range $x 0 end
    append x \uDE02
} -result [string range \uD83D\uDE02 0 end]

test append-4.1 {lappend command} {

test utf-1.15 {Tcl_UniCharToUtf: surrogate pairs from concat} {
    set hi \uD83D
    return $hi\uDE02
} \uD83D\uDE02
test utf-1.16 {Tcl_UniCharToUtf: \xC0 + \x80} testbytestring {
    set lo [testbytestring \x80]
    string length [testbytestring \xC0]$lo
} 2
test utf-1.17 {Tcl_UniCharToUtf: \xC0 + \x80} testbytestring {
    set hi [testbytestring \xC0]
    string length $hi[testbytestring \x80]
} 2
test utf-1.18 {Tcl_UniCharToUtf: surrogate pairs from concat} {
    string cat \uD83D \uDE02
} \uD83D\uDE02

test utf-2.1 {Tcl_UtfToUniChar: low ascii} {
    string length "abc"
} 3