ObjFW  Diff

			return OF_ORDERED_ASCENDING;
	}

	return OF_ORDERED_SAME;
}

int
of_string_check_utf8(const char *cString, size_t cStringLength, size_t *length)
{
	size_t i, tmpLength = cStringLength;
	int isUTF8 = 0;

	madvise((void*)cString, cStringLength, MADV_SEQUENTIAL);

	for (i = 0; i < cStringLength; i++) {
		/* No sign of UTF-8 here */
		if (OF_LIKELY(!(cString[i] & 0x80)))
			continue;

		isUTF8 = 1;

		/* We're missing a start byte here */
		if (OF_UNLIKELY(!(cString[i] & 0x40))) {
			madvise((void*)cString, cStringLength, MADV_NORMAL);
			return -1;
		}

		/* 2 byte sequences for code points 0 - 127 are forbidden */
		if (OF_UNLIKELY((cString[i] & 0x7E) == 0x40)) {
			madvise((void*)cString, cStringLength, MADV_NORMAL);
			return -1;
		}

		/* We have at minimum a 2 byte character -> check next byte */
		if (OF_UNLIKELY(cStringLength <= i + 1 ||
		    (cString[i + 1] & 0xC0) != 0x80)) {
			madvise((void*)cString, cStringLength, MADV_NORMAL);
			return -1;
		}

		/* Check if we have at minimum a 3 byte character */
		if (OF_LIKELY(!(cString[i] & 0x20))) {
			i++;
			tmpLength--;
			continue;
		}

		/* We have at minimum a 3 byte char -> check second next byte */
		if (OF_UNLIKELY(cStringLength <= i + 2 ||
		    (cString[i + 2] & 0xC0) != 0x80)) {
			madvise((void*)cString, cStringLength, MADV_NORMAL);
			return -1;
		}

		/* Check if we have a 4 byte character */
		if (OF_LIKELY(!(cString[i] & 0x10))) {
			i += 2;
			tmpLength -= 2;
			continue;
		}

		/* We have a 4 byte character -> check third next byte */
		if (OF_UNLIKELY(cStringLength <= i + 3 ||
		    (cString[i + 3] & 0xC0) != 0x80)) {
			madvise((void*)cString, cStringLength, MADV_NORMAL);
			return -1;
		}

		/*
		 * Just in case, check if there's a 5th character, which is
		 * forbidden by UTF-8
		 */
		if (OF_UNLIKELY(cString[i] & 0x08)) {
			madvise((void*)cString, cStringLength, MADV_NORMAL);
			return -1;
		}

		i += 3;
		tmpLength -= 3;
	}

	madvise((void*)cString, cStringLength, MADV_NORMAL);

	if (length != NULL)
		*length = tmpLength;

	return isUTF8;
}

size_t
of_string_unicode_to_utf8(of_unichar_t character, char *buffer)
{

		s = [self allocMemoryWithSize: sizeof(*s)];
		memset(s, 0, sizeof(*s));

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

		if (encoding == OF_STRING_ENCODING_UTF_8) {
			switch (of_string_check_utf8(cString, cStringLength,
			    &s->length)) {
			case 1:
				s->isUTF8 = YES;
				break;
			case -1:
				@throw [OFInvalidEncodingException
				    newWithClass: isa];
			}

			memcpy(s->cString, cString, cStringLength);
			s->cString[cStringLength] = 0;

			return self;
		}

		/* All other encodings we support are single byte encodings */
		s->length = cStringLength;

		if (encoding == OF_STRING_ENCODING_ISO_8859_1) {
			for (i = j = 0; i < cStringLength; i++) {
				char buffer[4];
				size_t bytes;

				if (!(cString[i] & 0x80)) {

{
	self = [super init];

	@try {
		s = [self allocMemoryWithSize: sizeof(*s)];
		memset(s, 0, sizeof(*s));

		/*
		 * We need one call to make sure it's initialized (in case it's
		 * a constant string).
		 */
		s->cStringLength = [string cStringLength];
		s->isUTF8 = string->s->isUTF8;
		s->length = string->s->length;

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

	return self;
}

			swap = YES;

		s = [self allocMemoryWithSize: sizeof(*s)];
		memset(s, 0, sizeof(*s));

		s->cStringLength = length;
		s->cString = [self allocMemoryWithSize: (length * 4) + 1];
		s->length = length;

		for (i = 0; i < length; i++) {
			char buffer[4];
			size_t characterLen = of_string_unicode_to_utf8(
			    (swap ? of_bswap32(string[i]) : string[i]),
			    buffer);

			swap = YES;

		s = [self allocMemoryWithSize: sizeof(*s)];
		memset(s, 0, sizeof(*s));

		s->cStringLength = length;
		s->cString = [self allocMemoryWithSize: (length * 4) + 1];
		s->length = length;

		for (i = 0; i < length; i++) {
			char buffer[4];
			of_unichar_t character =
			    (swap ? of_bswap16(string[i]) : string[i]);
			size_t characterLen;

				    ? of_bswap16(string[i + 1])
				    : string[i + 1]);
				character = (((character & 0x3FF) << 10) |
				    (nextCharacter & 0x3FF)) + 0x10000;

				i++;
				s->cStringLength--;
				s->length--;
			}

			characterLen = of_string_unicode_to_utf8(
			    character, buffer);

			switch (characterLen) {
			case 1:

		    arguments)) == -1)
			@throw [OFInvalidFormatException newWithClass: isa];

		s->cStringLength = cStringLength;

		@try {
			switch (of_string_check_utf8(s->cString,
			    cStringLength, &s->length)) {
			case 1:
				s->isUTF8 = YES;
				break;
			case -1:
				@throw [OFInvalidEncodingException
				    newWithClass: isa];
			}

		/*
		 * First needs to be a call to be sure it is initialized, in
		 * case it's a constant string.
		 */
		s->cStringLength = [firstComponent cStringLength];
		s->isUTF8 = firstComponent->s->isUTF8;
		s->length = firstComponent->s->length;

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

			if (component->s->isUTF8)
				s->isUTF8 = YES;
		}

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

- (const char*)cString
{
	return s->cString;
}

- (size_t)length
{



	return s->length;


}

- (size_t)cStringLength
{
	return s->cStringLength;
}

- (BOOL)isEqual: (id)object
{
	OFString *otherString;

	if (![object isKindOfClass: [OFString class]])
		return NO;

	otherString = object;

	if ([otherString cStringLength] != s->cStringLength ||
	    otherString->s->length != s->length)
		return NO;

	if (strcmp(s->cString, otherString->s->cString))
		return NO;

	return YES;
}

- copy
{

	return element;
}

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


	if (index >= s->length)
		@throw [OFOutOfRangeException newWithClass: isa];

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


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




	if (!of_string_utf8_to_unicode(s->cString + index,
	    s->cStringLength - index, &character))
		@throw [OFInvalidEncodingException newWithClass: isa];

	return character;
}

			return OF_INVALID_INDEX;
	}
}

- (BOOL)containsString: (OFString*)string
{
	const char *cString = [string cString];
	size_t i, cStringLength = string->s->cStringLength;

	if (cStringLength == 0)
		return YES;

	if (cStringLength > s->cStringLength)
		return NO;

	for (i = 0; i <= s->cStringLength - cStringLength; i++)
		if (!memcmp(s->cString + i, cString, cStringLength))
			return YES;

	return NO;
}

- (OFString*)substringFromIndex: (size_t)start
			toIndex: (size_t)end
{
	if (start > end)
		@throw [OFInvalidArgumentException newWithClass: isa
						       selector: _cmd];

	if (end > s->length)
		@throw [OFOutOfRangeException newWithClass: isa];

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








	return [OFString stringWithCString: s->cString + start
				    length: end - start];
}

- (OFString*)substringWithRange: (of_range_t)range
{
	return [self substringFromIndex: range.start

- (of_unichar_t*)unicodeString
{
	OFObject *object = [[[OFObject alloc] init] autorelease];
	of_unichar_t *ret;
	size_t i, j;

	ret = [object allocMemoryForNItems: s->length + 2
				  withSize: sizeof(of_unichar_t)];

	i = 0;
	j = 0;

	ret[j++] = 0xFEFF;