Merge branch 'master' into mm

Conflicts:
	engines/access/access.cpp
	engines/access/asurface.h
	engines/access/bubble_box.cpp
	engines/access/bubble_box.h
	engines/access/martian/martian_game.cpp
	engines/access/player.cpp
	engines/access/player.h
	engines/access/resources.cpp
	engines/access/screen.cpp
	engines/access/screen.h
	engines/access/sound.cpp
	engines/access/sound.h

6 files changed, 485 insertions, 78 deletions

diff --git a/common/dcl.cpp b/common/dcl.cpp
index 2f4cdeda6b..5993c218cb 100644
--- a/common/dcl.cpp
+++ b/common/dcl.cpp
@@ -30,17 +30,15 @@ namespace Common {
 
 class DecompressorDCL {
 public:
-	bool unpack(ReadStream *src, byte *dest, uint32 nPacked, uint32 nUnpacked);
+	bool unpack(SeekableReadStream *sourceStream, WriteStream *targetStream, uint32 targetSize, bool targetFixedSize);
 
 protected:
 	/**
 	 * Initialize decompressor.
-	 * @param src		source stream to read from
-	 * @param dest		destination stream to write to
-	 * @param nPacked	size of packed data
-	 * @param nUnpacked	size of unpacked data
+	 * @param sourceStream	source stream to read from
+	 * @param targetStream	target memory stream to write to
 	 */
-	void init(ReadStream *src, byte *dest, uint32 nPacked, uint32 nUnpacked);
+	void init(SeekableReadStream *sourceStream, WriteStream *targetStream, uint32 targetSize, bool targetFixedSize);
 
 	/**
 	 * Get a number of bits from _src stream, starting with the least
@@ -66,36 +64,38 @@ protected:
 
 	int huffman_lookup(const int *tree);
 
-	uint32 _dwBits;		///< bits buffer
-	byte _nBits;		///< number of unread bits in _dwBits
-	uint32 _szPacked;	///< size of the compressed data
-	uint32 _szUnpacked;	///< size of the decompressed data
-	uint32 _dwRead;		///< number of bytes read from _src
-	uint32 _dwWrote;	///< number of bytes written to _dest
-	ReadStream *_src;
-	byte *_dest;
+	uint32 _dwBits;			///< bits buffer
+	byte _nBits;			///< number of unread bits in _dwBits
+	uint32 _sourceSize;		///< size of the source stream
+	uint32 _targetSize;		///< size of the target stream (if fixed)
+	bool _targetFixedSize;  ///< if target stream is fixed size or dynamic size
+	uint32 _bytesRead;		///< number of bytes read from _sourceStream
+	uint32 _bytesWritten;	///< number of bytes written to _targetStream
+	SeekableReadStream *_sourceStream;
+	WriteStream *_targetStream;
 };
 
-void DecompressorDCL::init(ReadStream *src, byte *dest, uint32 nPacked, uint32 nUnpacked) {
-	_src = src;
-	_dest = dest;
-	_szPacked = nPacked;
-	_szUnpacked = nUnpacked;
+void DecompressorDCL::init(SeekableReadStream *sourceStream, WriteStream *targetStream, uint32 targetSize, bool targetFixedSize) {
+	_sourceStream = sourceStream;
+	_targetStream = targetStream;
+	_sourceSize = sourceStream->size();
+	_targetSize = targetSize;
+	_targetFixedSize = targetFixedSize;
 	_nBits = 0;
-	_dwRead = _dwWrote = 0;
+	_bytesRead = _bytesWritten = 0;
 	_dwBits = 0;
 }
 
 void DecompressorDCL::fetchBitsLSB() {
 	while (_nBits <= 24) {
-		_dwBits |= ((uint32)_src->readByte()) << _nBits;
+		_dwBits |= ((uint32)_sourceStream->readByte()) << _nBits;
 		_nBits += 8;
-		_dwRead++;
+		_bytesRead++;
 	}
 }
 
 uint32 DecompressorDCL::getBitsLSB(int n) {
-	// fetching more data to buffer if needed
+	// Fetching more data to buffer if needed
 	if (_nBits < n)
 		fetchBitsLSB();
 	uint32 ret = (_dwBits & ~((~0) << n));
@@ -109,7 +109,8 @@ byte DecompressorDCL::getByteLSB() {
 }
 
 void DecompressorDCL::putByte(byte b) {
-	_dest[_dwWrote++] = b;
+	_targetStream->writeByte(b);
+	_bytesWritten++;
 }
 
 #define HUFFMAN_LEAF 0x40000000
@@ -331,97 +332,189 @@ int DecompressorDCL::huffman_lookup(const int *tree) {
 #define DCL_BINARY_MODE 0
 #define DCL_ASCII_MODE 1
 
-bool DecompressorDCL::unpack(ReadStream *src, byte *dest, uint32 nPacked, uint32 nUnpacked) {
-	init(src, dest, nPacked, nUnpacked);
+#define MIDI_SETUP_BUNDLE_FILE_MAXIMUM_DICTIONARY_SIZE 4096
 
+bool DecompressorDCL::unpack(SeekableReadStream *sourceStream, WriteStream *targetStream, uint32 targetSize, bool targetFixedSize) {
+	byte   dictionary[MIDI_SETUP_BUNDLE_FILE_MAXIMUM_DICTIONARY_SIZE];
+	uint16 dictionaryPos = 0;
+	uint16 dictionarySize = 0;
+	uint16 dictionaryMask = 0;
 	int value;
-	uint32 val_distance, val_length;
+	uint16 tokenOffset = 0;
+	uint16 tokenLength = 0;
 
-	int mode = getByteLSB();
-	int length_param = getByteLSB();
+	init(sourceStream, targetStream, targetSize, targetFixedSize);
+
+	byte mode = getByteLSB();
+	byte dictionaryType = getByteLSB();
 
 	if (mode != DCL_BINARY_MODE && mode != DCL_ASCII_MODE) {
 		warning("DCL-INFLATE: Error: Encountered mode %02x, expected 00 or 01", mode);
 		return false;
 	}
 
-	if (length_param < 3 || length_param > 6)
-		warning("Unexpected length_param value %d (expected in [3,6])", length_param);
+	// TODO: original code supported 3 as well???
+	// Was this an accident or on purpose? And the original code did just give out a warning
+	// and didn't error out at all
+	switch (dictionaryType) {
+	case 4:
+		dictionarySize = 1024;
+		break;
+	case 5:
+		dictionarySize = 2048;
+		break;
+	case 6:
+		dictionarySize = 4096;
+		break;
+	default:
+		warning("DCL-INFLATE: Error: unsupported dictionary type %02x", dictionaryType);
+		return false;
+	}
+	dictionaryMask = dictionarySize - 1;
 
-	while (_dwWrote < _szUnpacked) {
+	while ((!targetFixedSize) || (_bytesWritten < _targetSize)) {
 		if (getBitsLSB(1)) { // (length,distance) pair
 			value = huffman_lookup(length_tree);
 
 			if (value < 8)
-				val_length = value + 2;
+				tokenLength = value + 2;
 			else
-				val_length = 8 + (1 << (value - 7)) + getBitsLSB(value - 7);
+				tokenLength = 8 + (1 << (value - 7)) + getBitsLSB(value - 7);
+
+			if (tokenLength == 519)
+				break; // End of stream signal
 
 			debug(8, " | ");
 
 			value = huffman_lookup(distance_tree);
 
-			if (val_length == 2)
-				val_distance = (value << 2) | getBitsLSB(2);
+			if (tokenLength == 2)
+				tokenOffset = (value << 2) | getBitsLSB(2);
 			else
-				val_distance = (value << length_param) | getBitsLSB(length_param);
-			val_distance ++;
+				tokenOffset = (value << dictionaryType) | getBitsLSB(dictionaryType);
+			tokenOffset++;
 
-			debug(8, "\nCOPY(%d from %d)\n", val_length, val_distance);
+			debug(8, "\nCOPY(%d from %d)\n", tokenLength, tokenOffset);
 
-			if (val_length + _dwWrote > _szUnpacked) {
-				warning("DCL-INFLATE Error: Write out of bounds while copying %d bytes (declared unpacked size is %d bytes, current is %d + %d bytes)",
-						val_length, _szUnpacked, _dwWrote, val_length);
-				return false;
+			if (_targetFixedSize) {
+				if (tokenLength + _bytesWritten > _targetSize) {
+					warning("DCL-INFLATE Error: Write out of bounds while copying %d bytes (declared unpacked size is %d bytes, current is %d + %d bytes)",
+							tokenLength, _targetSize, _bytesWritten, tokenLength);
+					return false;
+				}
 			}
 
-			if (_dwWrote < val_distance) {
+			if (_bytesWritten < tokenOffset) {
 				warning("DCL-INFLATE Error: Attempt to copy from before beginning of input stream (declared unpacked size is %d bytes, current is %d bytes)",
-						_szUnpacked, _dwWrote);
+						_targetSize, _bytesWritten);
 				return false;
 			}
 
-			while (val_length) {
-				uint32 copy_length = (val_length > val_distance) ? val_distance : val_length;
-				assert(val_distance >= copy_length);
-				uint32 pos = _dwWrote - val_distance;
-				for (uint32 i = 0; i < copy_length; i++)
-					putByte(dest[pos + i]);
+			uint16 dictionaryBaseIndex = (dictionaryPos - tokenOffset) & dictionaryMask;
+			uint16 dictionaryIndex = dictionaryBaseIndex;
+			uint16 dictionaryNextIndex = dictionaryPos;
+
+			while (tokenLength) {
+				// Write byte from dictionary
+				putByte(dictionary[dictionaryIndex]);
+				debug(9, "\33[32;31m%02x\33[37;37m ", dictionary[dictionaryIndex]);
+
+				dictionary[dictionaryNextIndex] = dictionary[dictionaryIndex];
 
-				for (uint32 i = 0; i < copy_length; i++)
-					debug(9, "\33[32;31m%02x\33[37;37m ", dest[pos + i]);
-				debug(9, "\n");
+				dictionaryNextIndex = (dictionaryNextIndex + 1) & dictionaryMask;
+				dictionaryIndex = (dictionaryIndex + 1) & dictionaryMask;
 
-				val_length -= copy_length;
-				val_distance += copy_length;
+				if (dictionaryIndex == dictionaryPos)
+					dictionaryIndex = dictionaryBaseIndex;
+				if (dictionaryNextIndex == dictionarySize)
+					dictionaryNextIndex = 0;
+
+				tokenLength--;
 			}
+			dictionaryPos = dictionaryNextIndex;
+			debug(9, "\n");
 
 		} else { // Copy byte verbatim
 			value = (mode == DCL_ASCII_MODE) ? huffman_lookup(ascii_tree) : getByteLSB();
 			putByte(value);
+
+			// Also remember it inside dictionary
+			dictionary[dictionaryPos] = value;
+			dictionaryPos++;
+			if (dictionaryPos >= dictionarySize)
+				dictionaryPos = 0;
+
 			debug(9, "\33[32;31m%02x \33[37;37m", value);
 		}
 	}
 
-	return _dwWrote == _szUnpacked;
+	if (_targetFixedSize) {
+		return _bytesWritten == _targetSize;
+	}
+	return true; // For targets featuring dynamic size we always succeed
 }
 
 bool decompressDCL(ReadStream *src, byte *dest, uint32 packedSize, uint32 unpackedSize) {
+	bool success = false;
+	DecompressorDCL dcl;
+
 	if (!src || !dest)
 		return false;
 
+	byte *sourceBufferPtr = (byte *)malloc(packedSize);
+	if (!sourceBufferPtr)
+		return false;
+
+	// Read source into memory
+	src->read(sourceBufferPtr, packedSize);
+
+	Common::MemoryReadStream  *sourceStream = new MemoryReadStream(sourceBufferPtr, packedSize, DisposeAfterUse::NO);
+	Common::MemoryWriteStream *targetStream = new MemoryWriteStream(dest, unpackedSize);
+
+	success = dcl.unpack(sourceStream, targetStream, unpackedSize, true);
+	delete sourceStream;
+	delete targetStream;
+	return success;
+}
+
+SeekableReadStream *decompressDCL(SeekableReadStream *sourceStream, uint32 packedSize, uint32 unpackedSize) {
+	bool success = false;
+	byte *targetPtr = nullptr;
+	Common::MemoryWriteStream *targetStream;
 	DecompressorDCL dcl;
-	return dcl.unpack(src, dest, packedSize, unpackedSize);
+
+	targetPtr = (byte *)malloc(unpackedSize);
+	if (!targetPtr)
+		return nullptr;
+
+	targetStream = new MemoryWriteStream(targetPtr, unpackedSize);
+
+	success = dcl.unpack(sourceStream, targetStream, unpackedSize, true);
+	delete targetStream;
+
+	if (!success) {
+		free(targetPtr);
+		return nullptr;
+	}
+	return new MemoryReadStream(targetPtr, unpackedSize, DisposeAfterUse::YES);
 }
 
-SeekableReadStream *decompressDCL(ReadStream *src, uint32 packedSize, uint32 unpackedSize) {
-	byte *data = (byte *)malloc(unpackedSize);
+// This one figures out the unpacked size by itself
+// Needed for at least Simon 2, because the unpacked size is not stored anywhere
+SeekableReadStream *decompressDCL(SeekableReadStream *sourceStream) {
+	Common::MemoryWriteStreamDynamic *targetStream;
+	DecompressorDCL dcl;
 
-	if (decompressDCL(src, data, packedSize, unpackedSize))
-		return new MemoryReadStream(data, unpackedSize, DisposeAfterUse::YES);
+	targetStream = new MemoryWriteStreamDynamic(DisposeAfterUse::NO);
 
-	free(data);
-	return 0;
+	if (dcl.unpack(sourceStream, targetStream, 0, false)) {
+		byte *targetPtr = targetStream->getData();
+		uint32 unpackedSize = targetStream->size();
+		delete targetStream;
+		return new MemoryReadStream(targetPtr, unpackedSize, DisposeAfterUse::YES);
+	}
+	delete targetStream;
+	return nullptr;
 }
 
 } // End of namespace Common
diff --git a/common/dcl.h b/common/dcl.h
index 0e96f74c07..f90bc23c8d 100644
--- a/common/dcl.h
+++ b/common/dcl.h
@@ -22,7 +22,8 @@
 
 /**
  * @file
- * PKWARE DCL ("explode") decompressor used in engines:
+ * PKWARE DCL ("explode") ("PKWARE data compression library") decompressor used in engines:
+ * - agos (exclusively for Simon 2 setup.shr file)
  * - mohawk
  * - sci
  */
@@ -38,16 +39,22 @@ class ReadStream;
 class SeekableReadStream;
 
 /**
- * Try to decompress a PKWARE DCL compressed stream. Returns true if
+ * Try to decompress a PKWARE DCL (PKWARE data compression library) compressed stream. Returns true if
  * successful.
  */
-bool decompressDCL(ReadStream *src, byte *dest, uint32 packedSize, uint32 unpackedSize);
+bool decompressDCL(ReadStream *sourceStream, byte *dest, uint32 packedSize, uint32 unpackedSize);
 
 /**
- * Try to decompress a PKWARE DCL compressed stream. Returns a valid pointer
+ * Try to decompress a PKWARE DCL (PKWARE data compression library) compressed stream. Returns a valid pointer
  * if successful and 0 otherwise.
  */
-SeekableReadStream *decompressDCL(ReadStream *src, uint32 packedSize, uint32 unpackedSize);
+SeekableReadStream *decompressDCL(SeekableReadStream *sourceStream, uint32 packedSize, uint32 unpackedSize);
+
+/**
+ * Try to decompress a PKWARE DCL (PKWARE data compression library) compressed stream. Returns a valid pointer
+ * if successful and 0 otherwise. This method is meant for cases, where the unpacked size is not known.
+ */
+SeekableReadStream *decompressDCL(SeekableReadStream *sourceStream);
 
 } // End of namespace Common
 
diff --git a/common/endian.h b/common/endian.h
index 6d6563f802..7278265961 100644
--- a/common/endian.h
+++ b/common/endian.h
@@ -49,6 +49,18 @@
 #	error No endianness defined
 #endif
 
+#ifdef HAVE_INT64
+#define SWAP_CONSTANT_64(a) \
+	((uint64)((((a) >> 56) & 0x000000FF) | \
+	          (((a) >> 40) & 0x0000FF00) | \
+	          (((a) >> 24) & 0x00FF0000) | \
+	          (((a) >>  8) & 0xFF000000) | \
+	          (((a) & 0xFF000000) <<  8) | \
+	          (((a) & 0x00FF0000) << 24) | \
+	          (((a) & 0x0000FF00) << 40) | \
+	          (((a) & 0x000000FF) << 56) ))
+#endif
+
 #define SWAP_CONSTANT_32(a) \
 	((uint32)((((a) >> 24) & 0x00FF) | \
 	          (((a) >>  8) & 0xFF00) | \
@@ -59,6 +71,36 @@
 	((uint16)((((a) >>  8) & 0x00FF) | \
 	          (((a) <<  8) & 0xFF00) ))
 
+
+
+/**
+ * Swap the bytes in a 16 bit word in order to convert LE encoded data to BE
+ * and vice versa.
+ */
+
+// compilerspecific variants come first, fallback last
+
+// Test for GCC and if the target has the MIPS rel.2 instructions (we know the psp does)
+#if defined(__GNUC__) && (defined(__psp__) || defined(_MIPS_ARCH_MIPS32R2) || defined(_MIPS_ARCH_MIPS64R2))
+
+	FORCEINLINE uint16 SWAP_BYTES_16(const uint16 a) {
+		if (__builtin_constant_p(a)) {
+			return SWAP_CONSTANT_16(a);
+		} else {
+			uint16 result;
+			__asm__ ("wsbh %0,%1" : "=r" (result) : "r" (a));
+			return result;
+		}
+	}
+#else
+
+	inline uint16 SWAP_BYTES_16(const uint16 a) {
+		return (a >> 8) | (a << 8);
+	}
+#endif
+
+
+
 /**
  * Swap the bytes in a 32 bit word in order to convert LE encoded data to BE
  * and vice versa.
@@ -108,32 +150,60 @@
 	}
 #endif
 
+#ifdef HAVE_INT64
 /**
- * Swap the bytes in a 16 bit word in order to convert LE encoded data to BE
+ * Swap the bytes in a 64 bit word in order to convert LE encoded data to BE
  * and vice versa.
  */
 
-// compilerspecific variants come first, fallback last
+// machine/compiler-specific variants come first, fallback last
 
 // Test for GCC and if the target has the MIPS rel.2 instructions (we know the psp does)
+//
 #if defined(__GNUC__) && (defined(__psp__) || defined(_MIPS_ARCH_MIPS32R2) || defined(_MIPS_ARCH_MIPS64R2))
 
-	FORCEINLINE uint16 SWAP_BYTES_16(const uint16 a) {
+	FORCEINLINE uint64 SWAP_BYTES_64(const uint64 a) {
 		if (__builtin_constant_p(a)) {
-			return SWAP_CONSTANT_16(a);
+			return SWAP_CONSTANT_64(a);
 		} else {
-			uint16 result;
-			__asm__ ("wsbh %0,%1" : "=r" (result) : "r" (a));
-			return result;
+			uint32 low = (uint32)a, high = (uint32)(a >> 32);
+			low = SWAP_BYTES_32(low);
+			high = SWAP_BYTES_32(high);
+
+			return (((uint64)low) << 32) | high;
 		}
 	}
+
+// Test for GCC >= 4.3.0 as this version added the bswap builtin
+#elif GCC_ATLEAST(4, 3)
+
+	FORCEINLINE uint64 SWAP_BYTES_64(uint64 a) {
+		return __builtin_bswap64(a);
+	}
+
+#elif defined(_MSC_VER)
+
+	FORCEINLINE uint64 SWAP_BYTES_64(uint64 a) {
+		return _byteswap_uint64(a);
+	}
+
+// generic fallback
 #else
 
-	inline uint16 SWAP_BYTES_16(const uint16 a) {
-		return (a >> 8) | (a << 8);
+	inline uint64 SWAP_BYTES_64(uint64 a) {
+		uint32 low = (uint32)a, high = (uint32)(a >> 32);
+		uint16 lowLow = (uint16)low, lowHigh = (uint16)(low >> 16),
+		       highLow = (uint16)high, highHigh = (uint16)(high >> 16);
+
+		return ((uint64)(((uint32)(uint16)((lowLow   >> 8) | (lowLow   << 8)) << 16) |
+		                          (uint16)((lowHigh  >> 8) | (lowHigh  << 8))) << 32) |
+		                (((uint32)(uint16)((highLow  >> 8) | (highLow  << 8)) << 16) |
+		                          (uint16)((highHigh >> 8) | (highHigh << 8)));
 	}
 #endif
 
+#endif // HAVE_INT64
+
 
 /**
  * A wrapper macro used around four character constants, like 'DATA', to
@@ -183,6 +253,18 @@
 		((Unaligned32 *)ptr)->val = value;
 	}
 
+#ifdef HAVE_INT64
+	FORCEINLINE uint64 READ_UINT64(const void *ptr) {
+		struct Unaligned64 { uint64 val; } __attribute__ ((__packed__, __may_alias__));
+		return ((const Unaligned64 *)ptr)->val;
+	}
+
+	FORCEINLINE void WRITE_UINT64(void *ptr, uint64 value) {
+		struct Unaligned64 { uint64 val; } __attribute__((__packed__, __may_alias__));
+		((Unaligned64 *)ptr)->val = value;
+	}
+#endif
+
 #elif !defined(SCUMM_NEED_ALIGNMENT)
 
 	FORCEINLINE uint16 READ_UINT16(const void *ptr) {
@@ -201,6 +283,16 @@
 		*(uint32 *)(ptr) = value;
 	}
 
+#ifdef HAVE_INT64
+	FORCEINLINE uint64 READ_UINT64(const void *ptr) {
+		return *(const uint64 *)(ptr);
+	}
+
+	FORCEINLINE void WRITE_UINT64(void *ptr, uint64 value) {
+		*(uint64 *)(ptr) = value;
+	}
+#endif
+
 
 // use software fallback by loading each byte explicitely
 #else
@@ -227,6 +319,23 @@
 			b[2] = (uint8)(value >> 16);
 			b[3] = (uint8)(value >> 24);
 		}
+#ifdef HAVE_INT64
+		inline uint64 READ_UINT64(const void *ptr) {
+			const uint8 *b = (const uint8 *)ptr;
+			return ((uint64)b[7] << 56) | ((uint64)b[6] << 48) | ((uint64)b[5] << 40) | ((uint64)b[4] << 32) | ((uint64)b[3] << 24) | ((uint64)b[2] << 16) | ((uint64)b[1] << 8) | ((uint64)b[0]);
+		}
+		inline void WRITE_UINT64(void *ptr, uint64 value) {
+			uint8 *b = (uint8 *)ptr;
+			b[0] = (uint8)(value >>  0);
+			b[1] = (uint8)(value >>  8);
+			b[2] = (uint8)(value >> 16);
+			b[3] = (uint8)(value >> 24);
+			b[4] = (uint8)(value >> 32);
+			b[5] = (uint8)(value >> 40);
+			b[6] = (uint8)(value >> 48);
+			b[7] = (uint8)(value >> 56);
+		}
+#endif
 
 #	elif defined(SCUMM_BIG_ENDIAN)
 
@@ -250,6 +359,23 @@
 			b[2] = (uint8)(value >>  8);
 			b[3] = (uint8)(value >>  0);
 		}
+#ifdef HAVE_INT64
+		inline uint64 READ_UINT64(const void *ptr) {
+			const uint8 *b = (const uint8 *)ptr;
+			return ((uint64)b[0] << 56) | ((uint64)b[1] << 48) | ((uint64)b[2] << 40) | ((uint64)b[3] << 32) | ((uint64)b[4] << 24) | ((uint64)b[5] << 16) | ((uint64)b[6] << 8) | ((uint64)b[7]);
+		}
+		inline void WRITE_UINT64(void *ptr, uint64 value) {
+			uint8 *b = (uint8 *)ptr;
+			b[0] = (uint8)(value >> 56);
+			b[1] = (uint8)(value >> 48);
+			b[2] = (uint8)(value >> 40);
+			b[3] = (uint8)(value >> 32);
+			b[4] = (uint8)(value >> 24);
+			b[5] = (uint8)(value >> 16);
+			b[6] = (uint8)(value >>  8);
+			b[7] = (uint8)(value >>  0);
+		}
+#endif
 
 #	endif
 
@@ -283,6 +409,17 @@
 	#define CONSTANT_BE_32(a) SWAP_CONSTANT_32(a)
 	#define CONSTANT_BE_16(a) SWAP_CONSTANT_16(a)
 
+#ifdef HAVE_INT64
+	#define READ_LE_UINT64(a) READ_UINT64(a)
+	#define WRITE_LE_UINT64(a, v) WRITE_UINT64(a, v)
+	#define FROM_LE_64(a) ((uint64)(a))
+	#define FROM_BE_64(a) SWAP_BYTES_64(a)
+	#define TO_LE_64(a) ((uint64)(a))
+	#define TO_BE_64(a) SWAP_BYTES_64(a)
+	#define CONSTANT_LE_64(a) ((uint64)(a))
+	#define CONSTANT_BE_64(a) SWAP_CONSTANT_64(a)
+#endif
+
 // if the unaligned load and the byteswap take alot instructions its better to directly read and invert
 #	if defined(SCUMM_NEED_ALIGNMENT) && !defined(__mips__)
 
@@ -306,6 +443,24 @@
 			b[2] = (uint8)(value >>  8);
 			b[3] = (uint8)(value >>  0);
 		}
+#ifdef HAVE_INT64
+		inline uint64 READ_BE_UINT64(const void *ptr) {
+			const uint8 *b = (const uint8 *)ptr;
+			return ((uint64)b[0] << 56) | ((uint64)b[1] << 48) | ((uint64)b[2] << 40) | ((uint64)b[3] << 32) | ((uint64)b[4] << 24) | ((uint64)b[5] << 16) | ((uint64)b[6] << 8) | ((uint64)b[7]);
+		}
+		inline void WRITE_BE_UINT64(void *ptr, uint64 value) {
+			uint8 *b = (uint8 *)ptr;
+			b[0] = (uint8)(value >> 56);
+			b[1] = (uint8)(value >> 48);
+			b[2] = (uint8)(value >> 40);
+			b[3] = (uint8)(value >> 32);
+			b[4] = (uint8)(value >> 24);
+			b[5] = (uint8)(value >> 16);
+			b[6] = (uint8)(value >> 8);
+			b[7] = (uint8)(value >> 0);
+		}
+#endif
+
 #	else
 
 		inline uint16 READ_BE_UINT16(const void *ptr) {
@@ -320,6 +475,14 @@
 		inline void WRITE_BE_UINT32(void *ptr, uint32 value) {
 			WRITE_UINT32(ptr, SWAP_BYTES_32(value));
 		}
+#ifdef HAVE_INT64
+		inline uint64 READ_BE_UINT64(const void *ptr) {
+			return SWAP_BYTES_64(READ_UINT64(ptr));
+		}
+		inline void WRITE_BE_UINT64(void *ptr, uint64 value) {
+			WRITE_UINT64(ptr, SWAP_BYTES_64(value));
+		}
+#endif
 
 #	endif	// if defined(SCUMM_NEED_ALIGNMENT)
 
@@ -349,6 +512,17 @@
 	#define CONSTANT_BE_32(a) ((uint32)(a))
 	#define CONSTANT_BE_16(a) ((uint16)(a))
 
+#ifdef HAVE_INT64
+	#define READ_BE_UINT64(a) READ_UINT64(a)
+	#define WRITE_BE_UINT64(a, v) WRITE_UINT64(a, v)
+	#define FROM_LE_64(a) SWAP_BYTES_64(a)
+	#define FROM_BE_64(a) ((uint64)(a))
+	#define TO_LE_64(a) SWAP_BYTES_64(a)
+	#define TO_BE_64(a) ((uint64)(a))
+	#define CONSTANT_LE_64(a) SWAP_CONSTANT_64(a)
+	#define CONSTANT_BE_64(a) ((uint64)(a))
+#endif
+
 // if the unaligned load and the byteswap take alot instructions its better to directly read and invert
 #	if defined(SCUMM_NEED_ALIGNMENT) && !defined(__mips__)
 
@@ -372,6 +546,25 @@
 		b[2] = (uint8)(value >> 16);
 		b[3] = (uint8)(value >> 24);
 	}
+
+#ifdef HAVE_INT64
+	inline uint64 READ_LE_UINT64(const void *ptr) {
+		const uint8 *b = (const uint8 *)ptr;
+		return ((uint64)b[7] << 56) | ((uint64)b[6] << 48) | ((uint64)b[5] << 40) | ((uint64)b[4] << 32) | ((uint64)b[3] << 24) | ((uint64)b[2] << 16) | ((uint64)b[1] << 8) | ((uint64)b[0]);
+	}
+	inline void WRITE_LE_UINT64(void *ptr, uint64 value) {
+		uint8 *b = (uint8 *)ptr;
+		b[0] = (uint8)(value >>  0);
+		b[1] = (uint8)(value >>  8);
+		b[2] = (uint8)(value >> 16);
+		b[3] = (uint8)(value >> 24);
+		b[4] = (uint8)(value >> 32);
+		b[5] = (uint8)(value >> 40);
+		b[6] = (uint8)(value >> 48);
+		b[7] = (uint8)(value >> 56);
+	}
+#endif
+
 #	else
 
 	inline uint16 READ_LE_UINT16(const void *ptr) {
@@ -386,6 +579,14 @@
 	inline void WRITE_LE_UINT32(void *ptr, uint32 value) {
 		WRITE_UINT32(ptr, SWAP_BYTES_32(value));
 	}
+#ifdef HAVE_INT64
+	inline uint64 READ_LE_UINT64(const void *ptr) {
+		return SWAP_BYTES_64(READ_UINT64(ptr));
+	}
+	inline void WRITE_LE_UINT64(void *ptr, uint64 value) {
+		WRITE_UINT64(ptr, SWAP_BYTES_64(value));
+	}
+#endif
 
 #	endif	// if defined(SCUMM_NEED_ALIGNMENT)
 
diff --git a/common/fft.cpp b/common/fft.cpp
index ac7386083f..27a04abb6a 100644
--- a/common/fft.cpp
+++ b/common/fft.cpp
@@ -56,11 +56,19 @@ FFT::FFT(int bits, int inverse) : _bits(bits), _inverse(inverse) {
 }
 
 FFT::~FFT() {
+	for (int i = 0; i < ARRAYSIZE(_cosTables); i++) {
+		delete _cosTables[i];
+	}
+
 	delete[] _revTab;
 	delete[] _expTab;
 	delete[] _tmpBuf;
 }
 
+const uint16 *FFT::getRevTab() const {
+	return _revTab;
+}
+
 void FFT::permute(Complex *z) {
 	int np = 1 << _bits;
 
diff --git a/common/fft.h b/common/fft.h
index 6eb72c3f84..ed66d32b71 100644
--- a/common/fft.h
+++ b/common/fft.h
@@ -47,6 +47,8 @@ public:
 	FFT(int bits, int inverse);
 	~FFT();
 
+	const uint16 *getRevTab() const;
+
 	/** Do the permutation needed BEFORE calling calc(). */
 	void permute(Complex *z);
 
diff --git a/common/stream.h b/common/stream.h
index 2702068cf3..abe5192b70 100644
--- a/common/stream.h
+++ b/common/stream.h
@@ -125,6 +125,13 @@ public:
 		write(&value, 4);
 	}
 
+#ifdef HAVE_INT64
+	void writeUint64LE(uint64 value) {
+		value = TO_LE_64(value);
+		write(&value, 8);
+	}
+#endif
+
 	void writeUint16BE(uint16 value) {
 		value = TO_BE_16(value);
 		write(&value, 2);
@@ -135,6 +142,13 @@ public:
 		write(&value, 4);
 	}
 
+#ifdef HAVE_INT64
+	void writeUint64BE(uint64 value) {
+		value = TO_BE_64(value);
+		write(&value, 8);
+	}
+#endif
+
 	FORCEINLINE void writeSint16LE(int16 value) {
 		writeUint16LE((uint16)value);
 	}
@@ -143,6 +157,12 @@ public:
 		writeUint32LE((uint32)value);
 	}
 
+#ifdef HAVE_INT64
+	FORCEINLINE void writeSint64LE(int64 value) {
+		writeUint64LE((uint64)value);
+	}
+#endif
+
 	FORCEINLINE void writeSint16BE(int16 value) {
 		writeUint16BE((uint16)value);
 	}
@@ -151,6 +171,12 @@ public:
 		writeUint32BE((uint32)value);
 	}
 
+#ifdef HAVE_INT64
+	FORCEINLINE void writeSint64BE(int64 value) {
+		writeUint64BE((uint64)value);
+	}
+#endif
+
 	/**
 	 * Write the given string to the stream.
 	 * This writes str.size() characters, but no terminating zero byte.
@@ -241,6 +267,21 @@ public:
 		return FROM_LE_32(val);
 	}
 
+#ifdef HAVE_INT64
+	/**
+	 * Read an unsigned 64-bit word stored in little endian (LSB first) order
+	 * from the stream and return it.
+	 * Performs no error checking. The return value is undefined
+	 * if a read error occurred (for which client code can check by
+	 * calling err() and eos() ).
+	 */
+	uint64 readUint64LE() {
+		uint64 val;
+		read(&val, 8);
+		return FROM_LE_64(val);
+	}
+#endif
+
 	/**
 	 * Read an unsigned 16-bit word stored in big endian (MSB first) order
 	 * from the stream and return it.
@@ -267,6 +308,21 @@ public:
 		return FROM_BE_32(val);
 	}
 
+#ifdef HAVE_INT64
+	/**
+	 * Read an unsigned 64-bit word stored in big endian (MSB first) order
+	 * from the stream and return it.
+	 * Performs no error checking. The return value is undefined
+	 * if a read error occurred (for which client code can check by
+	 * calling err() and eos() ).
+	 */
+	uint64 readUint64BE() {
+		uint64 val;
+		read(&val, 8);
+		return FROM_BE_64(val);
+	}
+#endif
+
 	/**
 	 * Read a signed 16-bit word stored in little endian (LSB first) order
 	 * from the stream and return it.
@@ -289,6 +345,19 @@ public:
 		return (int32)readUint32LE();
 	}
 
+#ifdef HAVE_INT64
+	/**
+	 * Read a signed 64-bit word stored in little endian (LSB first) order
+	 * from the stream and return it.
+	 * Performs no error checking. The return value is undefined
+	 * if a read error occurred (for which client code can check by
+	 * calling err() and eos() ).
+	 */
+	FORCEINLINE int64 readSint64LE() {
+		return (int64)readUint64LE();
+	}
+#endif
+
 	/**
 	 * Read a signed 16-bit word stored in big endian (MSB first) order
 	 * from the stream and return it.
@@ -311,6 +380,19 @@ public:
 		return (int32)readUint32BE();
 	}
 
+#ifdef HAVE_INT64
+	/**
+	 * Read a signed 64-bit word stored in big endian (MSB first) order
+	 * from the stream and return it.
+	 * Performs no error checking. The return value is undefined
+	 * if a read error occurred (for which client code can check by
+	 * calling err() and eos() ).
+	 */
+	FORCEINLINE int64 readSint64BE() {
+		return (int64)readUint64BE();
+	}
+#endif
+
 	/**
 	 * Read the specified amount of data into a malloc'ed buffer
 	 * which then is wrapped into a MemoryReadStream.
@@ -435,6 +517,14 @@ public:
 		return (_bigEndian) ? TO_BE_32(val) : TO_LE_32(val);
 	}
 
+#ifdef HAVE_INT64
+	uint64 readUint64() {
+		uint64 val;
+		read(&val, 8);
+		return (_bigEndian) ? TO_BE_64(val) : TO_LE_64(val);
+	}
+#endif
+
 	FORCEINLINE int16 readSint16() {
 		return (int16)readUint16();
 	}
@@ -442,6 +532,12 @@ public:
 	FORCEINLINE int32 readSint32() {
 		return (int32)readUint32();
 	}
+
+#ifdef HAVE_INT64
+	FORCEINLINE int64 readSint64() {
+		return (int64)readUint64();
+	}
+#endif
 };
 
 /**