ObjFW  Check-in [9817eab3be]

	ivars->cString = cString;
	ivars->cStringLength = cStringLength;

	switch (of_string_utf8_check(ivars->cString, ivars->cStringLength,
	    &ivars->length)) {
	case 1:
		ivars->UTF8 = YES;
		ivars->isUTF8 = YES;
		break;
	case -1:
		free(ivars);
		@throw [OFInvalidEncodingException
		    exceptionWithClass: [self class]];
	}

{
	of_unichar_t *unicodeString;
	size_t unicodeLen, newCStringLength;
	size_t i, j;
	char *newCString;
	BOOL isStart = YES;

	if (!s->UTF8) {
	if (!s->isUTF8) {
		uint8_t t;
		const of_unichar_t *const *table;

		assert(startTableSize >= 1 && middleTableSize >= 1);

		s->hashed = NO;

- (void)setCharacter: (of_unichar_t)character
	     atIndex: (size_t)index
{
	char buffer[4];
	of_unichar_t c;
	size_t lenNew, lenOld;

	if (s->UTF8)
	if (s->isUTF8)
		index = of_string_utf8_get_position(s->cString, index,
		    s->cStringLength);

	if (index > s->cStringLength)
		@throw [OFOutOfRangeException exceptionWithClass: [self class]];

	/* Shortcut if old and new character both are ASCII */

		memcpy(s->cString + index, buffer, lenNew);

		s->cStringLength -= lenOld;
		s->cStringLength += lenNew;
		s->cString[s->cStringLength] = '\0';

		if (character & 0x80)
			s->UTF8 = YES;
			s->isUTF8 = YES;
	} else if (lenNew < lenOld) {
		memmove(s->cString + index + lenNew,
		    s->cString + index + lenOld,
		    s->cStringLength - index - lenOld);
		memcpy(s->cString + index, buffer, lenNew);

		s->cStringLength -= lenOld;

	if (UTF8StringLength >= 3 && !memcmp(UTF8String, "\xEF\xBB\xBF", 3)) {
		UTF8String += 3;
		UTF8StringLength -= 3;
	}

	switch (of_string_utf8_check(UTF8String, UTF8StringLength, &length)) {
	case 1:
		s->UTF8 = YES;
		s->isUTF8 = YES;
		break;
	case -1:
		@throw [OFInvalidEncodingException
		    exceptionWithClass: [self class]];
	}

	s->hashed = NO;

	if (UTF8StringLength >= 3 && !memcmp(UTF8String, "\xEF\xBB\xBF", 3)) {
		UTF8String += 3;
		UTF8StringLength -= 3;
	}

	switch (of_string_utf8_check(UTF8String, UTF8StringLength, &length)) {
	case 1:
		s->UTF8 = YES;
		s->isUTF8 = YES;
		break;
	case -1:
		@throw [OFInvalidEncodingException
		    exceptionWithClass: [self class]];
	}

	s->hashed = NO;

	s->cStringLength += UTF8StringLength;
	s->length += [string length];

	s->cString[s->cStringLength] = 0;

	if ([string isKindOfClass: [OFString_UTF8 class]] ||
	    [string isKindOfClass: [OFMutableString_UTF8 class]]) {
		if (((OFString_UTF8*)string)->s->UTF8)
			s->UTF8 = YES;
		if (((OFString_UTF8*)string)->s->isUTF8)
			s->isUTF8 = YES;
	} else
		s->UTF8 = YES;
		s->isUTF8 = YES;
}

- (void)appendFormat: (OFConstantString*)format
       withArguments: (va_list)arguments
{
	char *UTF8String;
	int UTF8StringLength;

	for (i = 0, j = s->cStringLength - 1; i < s->cStringLength / 2;
	    i++, j--) {
		s->cString[i] ^= s->cString[j];
		s->cString[j] ^= s->cString[i];
		s->cString[i] ^= s->cString[j];
	}

	if (!s->UTF8)
	if (!s->isUTF8)
		return;

	for (i = 0; i < s->cStringLength; i++) {
		/* ASCII */
		if OF_LIKELY (!(s->cString[i] & 0x80))
			continue;

{
	size_t newCStringLength;

	if (index > s->length)
		@throw [OFOutOfRangeException
		    exceptionWithClass: [self class]];

	if (s->UTF8)
	if (s->isUTF8)
		index = of_string_utf8_get_position(s->cString, index,
		    s->cStringLength);

	newCStringLength = s->cStringLength + [string UTF8StringLength];
	s->hashed = NO;
	s->cString = [self resizeMemory: s->cString
				   size: newCStringLength + 1];

	memmove(s->cString + index + [string UTF8StringLength],
	    s->cString + index, s->cStringLength - index);
	memcpy(s->cString + index, [string UTF8String],
	    [string UTF8StringLength]);
	s->cString[newCStringLength] = '\0';

	s->cStringLength = newCStringLength;
	s->length += [string length];

	if ([string isKindOfClass: [OFString_UTF8 class]] ||
	    [string isKindOfClass: [OFMutableString_UTF8 class]]) {
		if (((OFString_UTF8*)string)->s->UTF8)
			s->UTF8 = YES;
		if (((OFString_UTF8*)string)->s->isUTF8)
			s->isUTF8 = YES;
	} else
		s->UTF8 = YES;
		s->isUTF8 = YES;
}

- (void)deleteCharactersInRange: (of_range_t)range
{
	size_t start = range.start;
	size_t end = range.start + range.length;

	if (end > s->length)
		@throw [OFOutOfRangeException exceptionWithClass: [self class]];

	s->hashed = NO;
	s->length -= end - start;

	if (s->UTF8) {
	if (s->isUTF8) {
		start = of_string_utf8_get_position(s->cString, start,
		    s->cStringLength);
		end = of_string_utf8_get_position(s->cString, end,
		    s->cStringLength);
	}

	memmove(s->cString + start, s->cString + end, s->cStringLength - end);

	size_t newCStringLength, newLength;

	if (end > s->length)
		@throw [OFOutOfRangeException exceptionWithClass: [self class]];

	newLength = s->length - (end - start) + [replacement length];

	if (s->UTF8) {
	if (s->isUTF8) {
		start = of_string_utf8_get_position(s->cString, start,
		    s->cStringLength);
		end = of_string_utf8_get_position(s->cString, end,
		    s->cStringLength);
	}

	newCStringLength = s->cStringLength - (end - start) +

	const char *searchString = [string UTF8String];
	const char *replacementString = [replacement UTF8String];
	size_t searchLength = [string UTF8StringLength];
	size_t replacementLength = [replacement UTF8StringLength];
	size_t i, last, newCStringLength, newLength;
	char *newCString;

	if (s->UTF8) {
	if (s->isUTF8) {
		range.start = of_string_utf8_get_position(s->cString,
		    range.start, s->cStringLength);
		range.length = of_string_utf8_get_position(s->cString,
		    range.start + range.length, s->cStringLength) - range.start;
	}

	if (range.start + range.length > [self UTF8StringLength])

 */
- (of_range_t)rangeOfString: (OFString*)string;

/**
 * \brief Returns the range of the string.
 *
 * \param string The string to search
 * \param options Options modifying search behaviour.
 * \param options Options modifying search behaviour
 * \return The range of the first occurrence of the string or a range with
 *	   OF_INVALID_INDEX as start position if it was not found
 */
- (of_range_t)rangeOfString: (OFString*)string
		    options: (of_string_search_options_t)options;

/**

	 * Since constant strings don't have s_store, they have to malloc it on
	 * the first access. Strings created at runtime just set the pointer to
	 * &s_store.
	 */
	struct of_string_utf8_ivars {
		char	 *cString;
		size_t	 cStringLength;
		BOOL	 UTF8;
		BOOL	 isUTF8;
		size_t	 length;
		BOOL	 hashed;
		uint32_t hash;
		char	 *freeWhenDone;
	} *restrict s;
	struct of_string_utf8_ivars s_store;
}

- OF_initWithUTF8String: (const char*)UTF8String
		 length: (size_t)UTF8StringLength
		storage: (char*)storage;
@end

		s->cString = storage;
		s->cStringLength = UTF8StringLength;

		switch (of_string_utf8_check(UTF8String, UTF8StringLength,
		    &s->length)) {
		case 1:
			s->UTF8 = YES;
			s->isUTF8 = YES;
			break;
		case -1:
			@throw [OFInvalidEncodingException
			    exceptionWithClass: [self class]];
		}

		memcpy(s->cString, UTF8String, UTF8StringLength);

			switch (of_string_utf8_check(cString, cStringLength,
			    &s->length)) {
			case 1:
				if (encoding == OF_STRING_ENCODING_ASCII)
					@throw [OFInvalidEncodingException
					    exceptionWithClass: [self class]];

				s->UTF8 = YES;
				s->isUTF8 = YES;
				break;
			case -1:
				@throw [OFInvalidEncodingException
				    exceptionWithClass: [self class]];
			}

			memcpy(s->cString, cString, cStringLength);

				size_t bytes;

				if (!(cString[i] & 0x80)) {
					s->cString[j++] = cString[i];
					continue;
				}

				s->UTF8 = YES;
				s->isUTF8 = YES;
				bytes = of_string_utf8_encode(
				    (uint8_t)cString[i], buffer);

				if (bytes == 0)
					@throw [OFInvalidEncodingException
					    exceptionWithClass: [self class]];

			character = table[(uint8_t)cString[i]];

			if (character == 0xFFFD)
				@throw [OFInvalidEncodingException
				    exceptionWithClass: [self class]];

			s->UTF8 = YES;
			s->isUTF8 = YES;
			characterBytes = of_string_utf8_encode(character,
			    buffer);

			if (characterBytes == 0)
				@throw [OFInvalidEncodingException
				    exceptionWithClass: [self class]];

		s->cString = (char*)UTF8String;
		s->cStringLength = UTF8StringLength;

		switch (of_string_utf8_check(UTF8String, UTF8StringLength,
		    &s->length)) {
		case 1:
			s->UTF8 = YES;
			s->isUTF8 = YES;
			break;
		case -1:
			@throw [OFInvalidEncodingException
			    exceptionWithClass: [self class]];
		}
	} @catch (id e) {
		[self release];

	@try {
		s = &s_store;

		s->cStringLength = [string UTF8StringLength];

		if ([string isKindOfClass: [OFString_UTF8 class]] ||
		    [string isKindOfClass: [OFMutableString_UTF8 class]])
			s->UTF8 = ((OFString_UTF8*)string)->s->UTF8;
			s->isUTF8 = ((OFString_UTF8*)string)->s->isUTF8;
		else
			s->UTF8 = YES;
			s->isUTF8 = YES;

		s->length = [string length];

		s->cString = [self allocMemoryWithSize: s->cStringLength + 1];
		memcpy(s->cString, [string UTF8String], s->cStringLength + 1);
	} @catch (id e) {
		[self release];

			    buffer);

			switch (characterLen) {
			case 1:
				s->cString[j++] = buffer[0];
				break;
			case 2:
				s->UTF8 = YES;
				s->isUTF8 = YES;
				s->cStringLength++;

				memcpy(s->cString + j, buffer, 2);
				j += 2;

				break;
			case 3:
				s->UTF8 = YES;
				s->isUTF8 = YES;
				s->cStringLength += 2;

				memcpy(s->cString + j, buffer, 3);
				j += 3;

				break;
			case 4:
				s->UTF8 = YES;
				s->isUTF8 = YES;
				s->cStringLength += 3;

				memcpy(s->cString + j, buffer, 4);
				j += 4;

				break;
			default:

			characterLen = of_string_utf8_encode(character, buffer);

			switch (characterLen) {
			case 1:
				s->cString[j++] = buffer[0];
				break;
			case 2:
				s->UTF8 = YES;
				s->isUTF8 = YES;
				s->cStringLength++;

				memcpy(s->cString + j, buffer, 2);
				j += 2;

				break;
			case 3:
				s->UTF8 = YES;
				s->isUTF8 = YES;
				s->cStringLength += 2;

				memcpy(s->cString + j, buffer, 3);
				j += 3;

				break;
			case 4:
				s->UTF8 = YES;
				s->isUTF8 = YES;
				s->cStringLength += 3;

				memcpy(s->cString + j, buffer, 4);
				j += 4;

				break;
			default:

		s->cStringLength = cStringLength;

		@try {
			switch (of_string_utf8_check(tmp, cStringLength,
			    &s->length)) {
			case 1:
				s->UTF8 = YES;
				s->isUTF8 = YES;
				break;
			case -1:
				@throw [OFInvalidEncodingException
				    exceptionWithClass: [self class]];
			}

			s->cString = [self

		s = &s_store;

		s->cStringLength = [firstComponent UTF8StringLength];

		if ([firstComponent isKindOfClass: [OFString_UTF8 class]] ||
		    [firstComponent isKindOfClass:
		    [OFMutableString_UTF8 class]])
			s->UTF8 = ((OFString_UTF8*)firstComponent)->s->UTF8;
			s->isUTF8 = ((OFString_UTF8*)firstComponent)->s->isUTF8;
		else
			s->UTF8 = YES;
			s->isUTF8 = YES;

		s->length = [firstComponent length];

		/* Calculate length and see if we need UTF-8 */
		va_copy(argumentsCopy, arguments);
		while ((component = va_arg(argumentsCopy, OFString*)) != nil) {
			s->cStringLength += 1 + [component UTF8StringLength];
			s->length += 1 + [component length];

			if ([component isKindOfClass: [OFString_UTF8 class]] ||
			    [component isKindOfClass:
			    [OFMutableString_UTF8 class]])
				s->isUTF8 =
				s->UTF8 = ((OFString_UTF8*)component)->s->UTF8;
				    ((OFString_UTF8*)component)->s->isUTF8;
			else
				s->UTF8 = YES;
				s->isUTF8 = YES;
		}

		s->cString = [self allocMemoryWithSize: s->cStringLength + 1];

		cStringLength = [firstComponent UTF8StringLength];
		memcpy(s->cString, [firstComponent UTF8String], cStringLength);
		i = cStringLength;

- (const char*)cStringWithEncoding: (of_string_encoding_t)encoding
{
	switch (encoding) {
	case OF_STRING_ENCODING_UTF_8:
		return s->cString;
	case OF_STRING_ENCODING_ASCII:
		if (s->UTF8)
		if (s->isUTF8)
			@throw [OFInvalidEncodingException
			    exceptionWithClass: [self class]];

		return s->cString;
	default:
		@throw [OFNotImplementedException
		    exceptionWithClass: [self class]

- (size_t)cStringLengthWithEncoding: (of_string_encoding_t)encoding
{
	switch (encoding) {
	case OF_STRING_ENCODING_UTF_8:
		return s->cStringLength;
	case OF_STRING_ENCODING_ASCII:
		if (s->UTF8)
		if (s->isUTF8)
			@throw [OFInvalidEncodingException
			    exceptionWithClass: [self class]];

		return s->cStringLength;
	default:
		@throw [OFNotImplementedException
		    exceptionWithClass: [self class]

		@throw [OFInvalidArgumentException
		    exceptionWithClass: [self class]
			      selector: _cmd];

	otherCString = [otherString UTF8String];
	otherCStringLength = [otherString UTF8StringLength];

	if (!s->UTF8) {
	if (!s->isUTF8) {
		minimumCStringLength = (s->cStringLength > otherCStringLength
		    ? otherCStringLength : s->cStringLength);

		if ((compare = memcasecmp(s->cString, otherCString,
		    minimumCStringLength)) == 0) {
			if (s->cStringLength > otherCStringLength)
				return OF_ORDERED_DESCENDING;

- (of_unichar_t)characterAtIndex: (size_t)index
{
	of_unichar_t character;

	if (index >= s->length)
		@throw [OFOutOfRangeException exceptionWithClass: [self class]];

	if (!s->UTF8)
	if (!s->isUTF8)
		return s->cString[index];

	index = of_string_utf8_get_position(s->cString, index,
	    s->cStringLength);

	if (!of_string_utf8_decode(s->cString + index, s->cStringLength - index,
	    &character))

{
	size_t start = range.start;
	size_t end = range.start + range.length;

	if (end > s->length)
		@throw [OFOutOfRangeException exceptionWithClass: [self class]];

	if (s->UTF8) {
	if (s->isUTF8) {
		start = of_string_utf8_get_position(s->cString, start,
		    s->cStringLength);
		end = of_string_utf8_get_position(s->cString, end,
		    s->cStringLength);
	}

	return [OFString stringWithUTF8String: s->cString + start