ObjFW  Check-in [545b6f740e]

- initWithCharacters: (const of_unichar_t*)characters
	      length: (size_t)length
{
	return (id)[[OFMutableString_UTF8 alloc] initWithCharacters: characters
							     length: length];
}

- initWithUTF16String: (const of_char16_t*)string
{
	return (id)[[OFMutableString_UTF8 alloc] initWithUTF16String: string];
}

- initWithUTF16String: (const of_char16_t*)string
	       length: (size_t)length
{
	return (id)[[OFMutableString_UTF8 alloc] initWithUTF16String: string
							      length: length];
}

- initWithUTF16String: (const of_char16_t*)string
	    byteOrder: (of_byte_order_t)byteOrder
{
	return (id)[[OFMutableString_UTF8 alloc]
	    initWithUTF16String: string
		      byteOrder: byteOrder];
}

- initWithUTF16String: (const of_char16_t*)string
	       length: (size_t)length
	    byteOrder: (of_byte_order_t)byteOrder
{
	return (id)[[OFMutableString_UTF8 alloc]
	    initWithUTF16String: string
			 length: length
		      byteOrder: byteOrder];
}

- initWithUTF32String: (const of_char32_t*)string
{
	return (id)[[OFMutableString_UTF8 alloc] initWithUTF32String: string];
}

- initWithUTF32String: (const of_char32_t*)string
	       length: (size_t)length
{
	return (id)[[OFMutableString_UTF8 alloc] initWithUTF32String: string
							      length: length];
}

- initWithUTF32String: (const of_char32_t*)string
	    byteOrder: (of_byte_order_t)byteOrder
{
	return (id)[[OFMutableString_UTF8 alloc]
	    initWithUTF32String: string
		      byteOrder: byteOrder];
}

- initWithUTF32String: (const of_char32_t*)string
	       length: (size_t)length
	    byteOrder: (of_byte_order_t)byteOrder
{
	return (id)[[OFMutableString_UTF8 alloc]
	    initWithUTF32String: string
			 length: length
		      byteOrder: byteOrder];

#import "OFObject.h"
#import "OFSerialization.h"
#import "OFJSONRepresentation.h"

@class OFConstantString;

#if defined(__cplusplus) && __cplusplus >= 201103L
typedef char16_t of_char16_t;
typedef char32_t of_char32_t;
#else
typedef uint_least16_t of_char16_t;
typedef uint_least32_t of_char32_t;
#endif
typedef of_char32_t of_unichar_t;

/*!
 * @brief The encoding of a string.
 */
typedef enum of_string_encoding_t {
	OF_STRING_ENCODING_UTF_8,
	OF_STRING_ENCODING_ASCII,

/*!
 * @brief Creates a new OFString from a UTF-16 encoded string.
 *
 * @param string The UTF-16 string
 * @return A new autoreleased OFString
 */
+ (instancetype)stringWithUTF16String: (const of_char16_t*)string;

/*!
 * @brief Creates a new OFString from a UTF-16 encoded string with the
 *	  specified length.
 *
 * @param string The UTF-16 string
 * @param length The length of the UTF-16 string
 * @return A new autoreleased OFString
 */
+ (instancetype)stringWithUTF16String: (const of_char16_t*)string
			       length: (size_t)length;

/*!
 * @brief Creates a new OFString from a UTF-16 encoded string, assuming the
 *	  specified byte order if no BOM is found.
 *
 * @param string The UTF-16 string
 * @param byteOrder The byte order to assume if there is no BOM
 * @return A new autoreleased OFString
 */
+ (instancetype)stringWithUTF16String: (const of_char16_t*)string
			    byteOrder: (of_byte_order_t)byteOrder;

/*!
 * @brief Creates a new OFString from a UTF-16 encoded string with the
 *	  specified length, assuming the specified byte order if no BOM is
 *	  found.
 *
 * @param string The UTF-16 string
 * @param length The length of the UTF-16 string
 * @param byteOrder The byte order to assume if there is no BOM
 * @return A new autoreleased OFString
 */
+ (instancetype)stringWithUTF16String: (const of_char16_t*)string
			       length: (size_t)length
			    byteOrder: (of_byte_order_t)byteOrder;

/*!
 * @brief Creates a new OFString from a UTF-32 encoded string.
 *
 * @param string The UTF-32 string
 * @return A new autoreleased OFString
 */
+ (instancetype)stringWithUTF32String: (const of_char32_t*)string;

/*!
 * @brief Creates a new OFString from a UTF-32 encoded string with the
 *	  specified length.
 *
 * @param string The UTF-32 string
 * @param length The length of the UTF-32 string
 * @return A new autoreleased OFString
 */
+ (instancetype)stringWithUTF32String: (const of_char32_t*)string
			       length: (size_t)length;

/*!
 * @brief Creates a new OFString from a UTF-32 encoded string, assuming the
 *	  specified byte order if no BOM is found.
 *
 * @param string The UTF-32 string
 * @param byteOrder The byte order to assume if there is no BOM
 * @return A new autoreleased OFString
 */
+ (instancetype)stringWithUTF32String: (const of_char32_t*)string
			    byteOrder: (of_byte_order_t)byteOrder;

/*!
 * @brief Creates a new OFString from a UTF-32 encoded string with the
 *	  specified length, assuming the specified byte order if no BOM is
 *	  found.
 *
 * @param string The UTF-32 string
 * @param length The length of the UTF-32 string
 * @param byteOrder The byte order to assume if there is no BOM
 * @return A new autoreleased OFString
 */
+ (instancetype)stringWithUTF32String: (const of_char32_t*)string
			       length: (size_t)length
			    byteOrder: (of_byte_order_t)byteOrder;

/*!
 * @brief Creates a new OFString from a format string.
 *
 * See printf for the format syntax. As an addition, %@ is available as format

/*!
 * @brief Initializes an already allocated OFString with a UTF-16 string.
 *
 * @param string The UTF-16 string
 * @return An initialized OFString
 */
- initWithUTF16String: (const of_char16_t*)string;

/*!
 * @brief Initializes an already allocated OFString with a UTF-16 string with
 *	  the specified length.
 *
 * @param string The UTF-16 string
 * @param length The length of the UTF-16 string
 * @return An initialized OFString
 */
- initWithUTF16String: (const of_char16_t*)string
	       length: (size_t)length;

/*!
 * @brief Initializes an already allocated OFString with a UTF-16 string,
 *	  assuming the specified byte order if no BOM is found.
 *
 * @param string The UTF-16 string
 * @param byteOrder The byte order to assume if there is no BOM
 * @return An initialized OFString
 */
- initWithUTF16String: (const of_char16_t*)string
	    byteOrder: (of_byte_order_t)byteOrder;

/*!
 * @brief Initializes an already allocated OFString with a UTF-16 string with
 *	  the specified length, assuming the specified byte order if no BOM is
 *	  found.
 *
 * @param string The UTF-16 string
 * @param length The length of the UTF-16 string
 * @param byteOrder The byte order to assume if there is no BOM
 * @return An initialized OFString
 */
- initWithUTF16String: (const of_char16_t*)string
	       length: (size_t)length
	    byteOrder: (of_byte_order_t)byteOrder;

/*!
 * @brief Initializes an already allocated OFString with a UTF-32 string.
 *
 * @param string The UTF-32 string
 * @return An initialized OFString
 */
- initWithUTF32String: (const of_char32_t*)string;

/*!
 * @brief Initializes an already allocated OFString with a UTF-32 string with
 *	  the specified length
 *
 * @param string The UTF-32 string
 * @param length The length of the UTF-32 string
 * @return An initialized OFString
 */
- initWithUTF32String: (const of_char32_t*)string
	       length: (size_t)length;

/*!
 * @brief Initializes an already allocated OFString with a UTF-32 string,
 *	  assuming the specified byte order if no BOM is found.
 *
 * @param string The UTF-32 string
 * @param byteOrder The byte order to assume if there is no BOM
 * @return An initialized OFString
 */
- initWithUTF32String: (const of_char32_t*)string
	    byteOrder: (of_byte_order_t)byteOrder;

/*!
 * @brief Initializes an already allocated OFString with a UTF-32 string with
 *	  the specified length, assuming the specified byte order if no BOM is
 *	  found.
 *
 * @param string The UTF-32 string
 * @param length The length of the UTF-32 string
 * @param byteOrder The byte order to assume if there is no BOM
 * @return An initialized OFString
 */
- initWithUTF32String: (const of_char32_t*)string
	       length: (size_t)length
	    byteOrder: (of_byte_order_t)byteOrder;

/*!
 * @brief Initializes an already allocated OFString with a format string.
 *
 * See printf for the format syntax. As an addition, %@ is available as format

 *
 * The result is valid until the autorelease pool is released. If you want to
 * use the result outside the scope of the current autorelease pool, you have to
 * copy it.
 *
 * @return The string in UTF-16 encoding with native byte order
 */
- (const of_char16_t*)UTF16String OF_RETURNS_INNER_POINTER;

/*!
 * @brief Returns the string in UTF-16 encoding with the specified byte order.
 *
 * The result is valid until the autorelease pool is released. If you want to
 * use the result outside the scope of the current autorelease pool, you have to
 * copy it.
 *
 * @param byteOrder The byte order for the UTF-16 encoding
 * @return The string in UTF-16 encoding with the specified byte order
 */
- (const of_char16_t*)UTF16StringWithByteOrder: (of_byte_order_t)byteOrder
    OF_RETURNS_INNER_POINTER;

/*!
 * @brief Returns the length of the string in UTF-16 characters.
 *
 * @return The length of string in UTF-16 characters
 */
- (size_t)UTF16StringLength;

/*!
 * @brief Returns the string in UTF-32 encoding with native byte order.
 *
 * The result is valid until the autorelease pool is released. If you want to
 * use the result outside the scope of the current autorelease pool, you have to
 * copy it.
 *
 * @return The string in UTF-32 encoding with native byte order
 */
- (const of_char32_t*)UTF32String OF_RETURNS_INNER_POINTER;

/*!
 * @brief Returns the string in UTF-32 encoding with the specified byte order.
 *
 * The result is valid until the autorelease pool is released. If you want to
 * use the result outside the scope of the current autorelease pool, you have to
 * copy it.
 *
 * @param byteOrder The byte order for the UTF-32 encoding
 * @return The string in UTF-32 encoding with the specified byte order
 */
- (const of_char32_t*)UTF32StringWithByteOrder: (of_byte_order_t)byteOrder
    OF_RETURNS_INNER_POINTER;

/*!
 * @brief Writes the string into the specified file using UTF-8 encoding.
 *
 * @param path The path of the file to write to
 */

#endif

#ifdef __cplusplus
extern "C" {
#endif
extern size_t of_string_utf8_encode(of_unichar_t, char*);
extern size_t of_string_utf8_decode(const char*, size_t, of_unichar_t*);
extern size_t of_string_utf16_length(const of_char16_t*);
extern size_t of_string_utf32_length(const of_char32_t*);
#ifdef __cplusplus
}
#endif

		return 4;
	}

	return 0;
}

size_t
of_string_utf16_length(const of_char16_t *string)
{
	size_t length = 0;

	while (*string++ != 0)
		length++;

	return length;
}

size_t
of_string_utf32_length(const of_char32_t *string)
{
	size_t length = 0;

	while (*string++ != 0)
		length++;

	return length;

- initWithCharacters: (const of_unichar_t*)string
	      length: (size_t)length
{
	return (id)[[OFString_UTF8 alloc] initWithCharacters: string
						      length: length];
}

- initWithUTF16String: (const of_char16_t*)string
{
	return (id)[[OFString_UTF8 alloc] initWithUTF16String: string];
}

- initWithUTF16String: (const of_char16_t*)string
	       length: (size_t)length
{
	return (id)[[OFString_UTF8 alloc] initWithUTF16String: string
						       length: length];
}

- initWithUTF16String: (const of_char16_t*)string
	    byteOrder: (of_byte_order_t)byteOrder
{
	return (id)[[OFString_UTF8 alloc] initWithUTF16String: string
						    byteOrder: byteOrder];
}

- initWithUTF16String: (const of_char16_t*)string
	       length: (size_t)length
	    byteOrder: (of_byte_order_t)byteOrder
{
	return (id)[[OFString_UTF8 alloc] initWithUTF16String: string
						       length: length
						    byteOrder: byteOrder];
}

- initWithUTF32String: (const of_char32_t*)string
{
	return (id)[[OFString_UTF8 alloc] initWithUTF32String: string];
}

- initWithUTF32String: (const of_char32_t*)string
	       length: (size_t)length
{
	return (id)[[OFString_UTF8 alloc] initWithUTF32String: string
						       length: length];
}

- initWithUTF32String: (const of_char32_t*)string
	    byteOrder: (of_byte_order_t)byteOrder
{
	return (id)[[OFString_UTF8 alloc] initWithUTF32String: string
						    byteOrder: byteOrder];
}

- initWithUTF32String: (const of_char32_t*)string
	       length: (size_t)length
	    byteOrder: (of_byte_order_t)byteOrder
{
	return (id)[[OFString_UTF8 alloc] initWithUTF32String: string
						       length: length
						    byteOrder: byteOrder];
}

+ (instancetype)stringWithCharacters: (const of_unichar_t*)string
			      length: (size_t)length
{
	return [[[self alloc] initWithCharacters: string
					  length: length] autorelease];
}

+ (instancetype)stringWithUTF16String: (const of_char16_t*)string
{
	return [[[self alloc] initWithUTF16String: string] autorelease];
}

+ (instancetype)stringWithUTF16String: (const of_char16_t*)string
			       length: (size_t)length
{
	return [[[self alloc] initWithUTF16String: string
					   length: length] autorelease];
}

+ (instancetype)stringWithUTF16String: (const of_char16_t*)string
			    byteOrder: (of_byte_order_t)byteOrder
{
	return [[[self alloc] initWithUTF16String: string
					byteOrder: byteOrder] autorelease];
}

+ (instancetype)stringWithUTF16String: (const of_char16_t*)string
			       length: (size_t)length
			    byteOrder: (of_byte_order_t)byteOrder
{
	return [[[self alloc] initWithUTF16String: string
					   length: length
					byteOrder: byteOrder] autorelease];
}

+ (instancetype)stringWithUTF32String: (const of_char32_t*)string
{
	return [[[self alloc] initWithUTF32String: string] autorelease];
}

+ (instancetype)stringWithUTF32String: (const of_char32_t*)string
			       length: (size_t)length
{
	return [[[self alloc] initWithUTF32String: string
					   length: length] autorelease];
}

+ (instancetype)stringWithUTF32String: (const of_char32_t*)string
			    byteOrder: (of_byte_order_t)byteOrder
{
	return [[[self alloc] initWithUTF32String: string
					byteOrder: byteOrder] autorelease];
}

+ (instancetype)stringWithUTF32String: (const of_char32_t*)string
			       length: (size_t)length
			    byteOrder: (of_byte_order_t)byteOrder
{
	return [[[self alloc] initWithUTF32String: string
					   length: length
					byteOrder: byteOrder] autorelease];
}

		abort();
	} @catch (id e) {
		[self release];
		@throw e;
	}
}

- initWithUTF16String: (const of_char16_t*)string
{
	return [self initWithUTF16String: string
				  length: of_string_utf16_length(string)
			       byteOrder: OF_BYTE_ORDER_NATIVE];
}

- initWithUTF16String: (const of_char16_t*)string
	       length: (size_t)length
{
	return [self initWithUTF16String: string
				  length: length
			       byteOrder: OF_BYTE_ORDER_NATIVE];
}

- initWithUTF16String: (const of_char16_t*)string
	    byteOrder: (of_byte_order_t)byteOrder
{
	return [self initWithUTF16String: string
				  length: of_string_utf16_length(string)
			       byteOrder: byteOrder];
}

- initWithUTF16String: (const of_char16_t*)string
	       length: (size_t)length
	    byteOrder: (of_byte_order_t)byteOrder
{
	@try {
		[self doesNotRecognizeSelector: _cmd];
		abort();
	} @catch (id e) {
		[self release];
		@throw e;
	}
}

- initWithUTF32String: (const of_char32_t*)string
{
	return [self initWithUTF32String: string
				  length: of_string_utf32_length(string)
			       byteOrder: OF_BYTE_ORDER_NATIVE];
}

- initWithUTF32String: (const of_char32_t*)string
	       length: (size_t)length
{
	return [self initWithUTF32String: string
				  length: length
			       byteOrder: OF_BYTE_ORDER_NATIVE];
}

- initWithUTF32String: (const of_char32_t*)string
	    byteOrder: (of_byte_order_t)byteOrder
{
	return [self initWithUTF32String: string
				  length: of_string_utf32_length(string)
			       byteOrder: byteOrder];
}

- initWithUTF32String: (const of_char32_t*)string
	       length: (size_t)length
	    byteOrder: (of_byte_order_t)byteOrder
{
	@try {
		[self doesNotRecognizeSelector: _cmd];
		abort();
	} @catch (id e) {

				    count: length];
	[self getCharacters: ret
		    inRange: of_range(0, length)];

	return ret;
}

- (const of_char16_t*)UTF16String
{
	return [self UTF16StringWithByteOrder: OF_BYTE_ORDER_NATIVE];
}

- (const of_char16_t*)UTF16StringWithByteOrder: (of_byte_order_t)byteOrder
{
	OFObject *object = [[[OFObject alloc] init] autorelease];
	void *pool = objc_autoreleasePoolPush();
	const of_unichar_t *characters = [self characters];
	size_t length = [self length];
	of_char16_t *ret;
	size_t i, j;
	BOOL swap = (byteOrder != OF_BYTE_ORDER_NATIVE);

	/* Allocate memory for the worst case */
	ret = [object allocMemoryWithSize: sizeof(of_char16_t)
				    count: (length + 1) * 2];

	j = 0;

	for (i = 0; i < length; i++) {
		of_unichar_t c = characters[i];

				ret[j++] = c;
		}
	}
	ret[j] = 0;

	@try {
		ret = [object resizeMemory: ret
				      size: sizeof(of_char16_t)
				     count: j + 1];
	} @catch (OFOutOfMemoryException *e) {
		/* We don't care, as we only tried to make it smaller */
	}

	objc_autoreleasePoolPop(pool);

	for (i = 0; i < length; i++)
		if (characters[i] > 0xFFFF)
			UTF16StringLength++;

	return UTF16StringLength;
}

- (const of_char32_t*)UTF32String
{
	return [self UTF32StringWithByteOrder: OF_BYTE_ORDER_NATIVE];
}

- (const of_char32_t*)UTF32StringWithByteOrder: (of_byte_order_t)byteOrder
{
	OFObject *object = [[[OFObject alloc] init] autorelease];
	size_t length = [self length];
	of_char32_t *ret;

	ret = [object allocMemoryWithSize: sizeof(of_char32_t)
				    count: length + 1];
	[self getCharacters: ret
		    inRange: of_range(0, length)];
	ret[length] = 0;

	if (byteOrder != OF_BYTE_ORDER_NATIVE) {
		size_t i;

#import "OFOutOfRangeException.h"

#import "autorelease.h"
#import "macros.h"
#import "of_asprintf.h"
#import "unicode.h"

extern const of_char16_t of_iso_8859_15[256];
extern const of_char16_t of_windows_1252[256];

static inline int
memcasecmp(const char *first, const char *second, size_t length)
{
	size_t i;

	for (i = 0; i < length; i++) {

	 encoding: (of_string_encoding_t)encoding
	   length: (size_t)cStringLength
{
	self = [super init];

	@try {
		size_t i, j;
		const of_char16_t *table;

		if (encoding == OF_STRING_ENCODING_UTF_8 &&
		    cStringLength >= 3 && !memcmp(cString, "\xEF\xBB\xBF", 3)) {
			cString += 3;
			cStringLength -= 3;
		}

		[self release];
		@throw e;
	}

	return self;
}

- initWithUTF16String: (const of_char16_t*)string
	       length: (size_t)length
	    byteOrder: (of_byte_order_t)byteOrder
{
	self = [super init];

	@try {
		size_t i, j = 0;

			/* Missing high surrogate */
			if ((character & 0xFC00) == 0xDC00)
				@throw [OFInvalidEncodingException
				    exceptionWithClass: [self class]];

			if ((character & 0xFC00) == 0xD800) {
				of_char16_t nextCharacter;

				if (length <= i + 1)
					@throw [OFInvalidEncodingException
					    exceptionWithClass: [self class]];

				nextCharacter = (swap
				    ? OF_BSWAP16(string[i + 1])

		[self release];
		@throw e;
	}

	return self;
}

- initWithUTF32String: (const of_char32_t*)characters
	       length: (size_t)length
	    byteOrder: (of_byte_order_t)byteOrder
{
	self = [super init];

	@try {
		size_t i, j = 0;

		ret[j++] = c;
		i += cLen;
	}

	return ret;
}

- (const of_char32_t*)UTF32StringWithByteOrder: (of_byte_order_t)byteOrder
{
	OFObject *object = [[[OFObject alloc] init] autorelease];
	of_char32_t *ret;
	size_t i, j;

	ret = [object allocMemoryWithSize: sizeof(of_unichar_t)
				    count: s->length + 1];

	i = j = 0;

 *
 * Alternatively, it may be distributed under the terms of the GNU General
 * Public License, either version 2 or 3, which can be found in the file
 * LICENSE.GPLv2 or LICENSE.GPLv3 respectively included in the packaging of this
 * file.
 */

#import "OFString.h"

const of_char16_t of_iso_8859_15[256] = {
	0x0000, 0x0001, 0x0002, 0x0003, 0x0004, 0x0005, 0x0006, 0x0007,
	0x0008, 0x0009, 0x000A, 0x000B, 0x000C, 0x000D, 0x000E, 0x000F,
	0x0010, 0x0011, 0x0012, 0x0013, 0x0014, 0x0015, 0x0016, 0x0017,
	0x0018, 0x0019, 0x001A, 0x001B, 0x001C, 0x001D, 0x001E, 0x001F,
	0x0020, 0x0021, 0x0022, 0x0023, 0x0024, 0x0025, 0x0026, 0x0027,
	0x0028, 0x0029, 0x002A, 0x002B, 0x002C, 0x002D, 0x002E, 0x002F,
	0x0030, 0x0031, 0x0032, 0x0033, 0x0034, 0x0035, 0x0036, 0x0037,

 *
 * Alternatively, it may be distributed under the terms of the GNU General
 * Public License, either version 2 or 3, which can be found in the file
 * LICENSE.GPLv2 or LICENSE.GPLv3 respectively included in the packaging of this
 * file.
 */

#import "OFString.h"

const of_char16_t of_windows_1252[256] = {
	0x0000, 0x0001, 0x0002, 0x0003, 0x0004, 0x0005, 0x0006, 0x0007,
	0x0008, 0x0009, 0x000A, 0x000B, 0x000C, 0x000D, 0x000E, 0x000F,
	0x0010, 0x0011, 0x0012, 0x0013, 0x0014, 0x0015, 0x0016, 0x0017,
	0x0018, 0x0019, 0x001A, 0x001B, 0x001C, 0x001D, 0x001E, 0x001F,
	0x0020, 0x0021, 0x0022, 0x0023, 0x0024, 0x0025, 0x0026, 0x0027,
	0x0028, 0x0029, 0x002A, 0x002B, 0x002C, 0x002D, 0x002E, 0x002F,
	0x0030, 0x0031, 0x0032, 0x0033, 0x0034, 0x0035, 0x0036, 0x0037,