diff --git a/lz4.cabal b/lz4.cabal
--- a/lz4.cabal
+++ b/lz4.cabal
@@ -1,5 +1,5 @@
 name:                lz4
-version:             0.2.1
+version:             0.2.2
 synopsis:            LZ4 compression for ByteStrings
 description:
   High level bindings to the LZ4 compression library.
diff --git a/src/cbits/lz4.c b/src/cbits/lz4.c
--- a/src/cbits/lz4.c
+++ b/src/cbits/lz4.c
@@ -1,860 +1,874 @@
-/*
-   LZ4 - Fast LZ compression algorithm
-   Copyright (C) 2011-2012, Yann Collet.
-   BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
-
-   Redistribution and use in source and binary forms, with or without
-   modification, are permitted provided that the following conditions are
-   met:
-
-       * Redistributions of source code must retain the above copyright
-   notice, this list of conditions and the following disclaimer.
-       * Redistributions in binary form must reproduce the above
-   copyright notice, this list of conditions and the following disclaimer
-   in the documentation and/or other materials provided with the
-   distribution.
-
-   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
-   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
-   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
-   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
-   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
-   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
-   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
-   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
-   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
-   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
-   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-
-   You can contact the author at :
-   - LZ4 homepage : http://fastcompression.blogspot.com/p/lz4.html
-   - LZ4 source repository : http://code.google.com/p/lz4/
-*/
-
-//**************************************
-// Tuning parameters
-//**************************************
-// MEMORY_USAGE :
-// Memory usage formula : N->2^N Bytes (examples : 10 -> 1KB; 12 -> 4KB ; 16 -> 64KB; 20 -> 1MB; etc.)
-// Increasing memory usage improves compression ratio
-// Reduced memory usage can improve speed, due to cache effect
-// Default value is 14, for 16KB, which nicely fits into Intel x86 L1 cache
-#define MEMORY_USAGE 14
-
-// NOTCOMPRESSIBLE_DETECTIONLEVEL :
-// Decreasing this value will make the algorithm skip faster data segments considered "incompressible"
-// This may decrease compression ratio dramatically, but will be faster on incompressible data
-// Increasing this value will make the algorithm search more before declaring a segment "incompressible"
-// This could improve compression a bit, but will be slower on incompressible data
-// The default value (6) is recommended
-#define NOTCOMPRESSIBLE_DETECTIONLEVEL 6
-
-// BIG_ENDIAN_NATIVE_BUT_INCOMPATIBLE :
-// This will provide a small boost to performance for big endian cpu, but the resulting compressed stream will be incompatible with little-endian CPU.
-// You can set this option to 1 in situations where data will remain within closed environment
-// This option is useless on Little_Endian CPU (such as x86)
-//#define BIG_ENDIAN_NATIVE_BUT_INCOMPATIBLE 1
-
-
-
-//**************************************
-// CPU Feature Detection
-//**************************************
-// 32 or 64 bits ?
-#if (defined(__x86_64__) || defined(__x86_64) || defined(__amd64__) || defined(__amd64) || defined(__ppc64__) || defined(_WIN64) || defined(__LP64__) || defined(_LP64) )   // Detects 64 bits mode
-#  define LZ4_ARCH64 1
-#else
-#  define LZ4_ARCH64 0
-#endif
-
-// Little Endian or Big Endian ?
-// Note : overwrite the below #define if you know your architecture endianess
-#if (defined(__BIG_ENDIAN__) || defined(__BIG_ENDIAN) || defined(_BIG_ENDIAN) || defined(_ARCH_PPC) || defined(__PPC__) || defined(__PPC) || defined(PPC) || defined(__powerpc__) || defined(__powerpc) || defined(powerpc) || ((defined(__BYTE_ORDER__)&&(__BYTE_ORDER__ == __ORDER_BIG_ENDIAN__))) )
-#  define LZ4_BIG_ENDIAN 1
-#else
-// Little Endian assumed. PDP Endian and other very rare endian format are unsupported.
-#endif
-
-// Unaligned memory access is automatically enabled for "common" CPU, such as x86.
-// For others CPU, the compiler will be more cautious, and insert extra code to ensure aligned access is respected
-// If you know your target CPU supports unaligned memory access, you may want to force this option manually to improve performance
-#if defined(__ARM_FEATURE_UNALIGNED)
-#  define LZ4_FORCE_UNALIGNED_ACCESS 1
-#endif
-
-// Define this parameter if your target system or compiler does not support hardware bit count
-#if defined(_MSC_VER) && defined(_WIN32_WCE)            // Visual Studio for Windows CE does not support Hardware bit count
-#  define LZ4_FORCE_SW_BITCOUNT
-#endif
-
-
-//**************************************
-// Compiler Options
-//**************************************
-#if __STDC_VERSION__ >= 199901L // C99
-/* "restrict" is a known keyword */
-#else
-#  define restrict // Disable restrict
-#endif
-
-#define GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__)
-
-#ifdef _MSC_VER  // Visual Studio
-#  define inline __forceinline           // Visual is not C99, but supports some kind of inline
-#  include <intrin.h>   // For Visual 2005
-#  if LZ4_ARCH64	// 64-bit
-#    pragma intrinsic(_BitScanForward64) // For Visual 2005
-#    pragma intrinsic(_BitScanReverse64) // For Visual 2005
-#  else
-#    pragma intrinsic(_BitScanForward)   // For Visual 2005
-#    pragma intrinsic(_BitScanReverse)   // For Visual 2005
-#  endif
-#endif
-
-#ifdef _MSC_VER
-#  define lz4_bswap16(x) _byteswap_ushort(x)
-#else
-#  define lz4_bswap16(x) ((unsigned short int) ((((x) >> 8) & 0xffu) | (((x) & 0xffu) << 8)))
-#endif
-
-#if (GCC_VERSION >= 302) || (__INTEL_COMPILER >= 800) || defined(__clang__)
-#  define expect(expr,value)    (__builtin_expect ((expr),(value)) )
-#else
-#  define expect(expr,value)    (expr)
-#endif
-
-#define likely(expr)     expect((expr) != 0, 1)
-#define unlikely(expr)   expect((expr) != 0, 0)
-
-
-//**************************************
-// Includes
-//**************************************
-#include <stdlib.h>   // for malloc
-#include <string.h>   // for memset
-#include "lz4.h"
-
-
-//**************************************
-// Basic Types
-//**************************************
-#if defined(_MSC_VER)    // Visual Studio does not support 'stdint' natively
-#  define BYTE	unsigned __int8
-#  define U16		unsigned __int16
-#  define U32		unsigned __int32
-#  define S32		__int32
-#  define U64		unsigned __int64
-#else
-#  include <stdint.h>
-#  define BYTE	uint8_t
-#  define U16		uint16_t
-#  define U32		uint32_t
-#  define S32		int32_t
-#  define U64		uint64_t
-#endif
-
-#ifndef LZ4_FORCE_UNALIGNED_ACCESS
-#  pragma pack(push, 1)
-#endif
-
-typedef struct _U16_S { U16 v; } U16_S;
-typedef struct _U32_S { U32 v; } U32_S;
-typedef struct _U64_S { U64 v; } U64_S;
-
-#ifndef LZ4_FORCE_UNALIGNED_ACCESS
-#  pragma pack(pop)
-#endif
-
-#define A64(x) (((U64_S *)(x))->v)
-#define A32(x) (((U32_S *)(x))->v)
-#define A16(x) (((U16_S *)(x))->v)
-
-
-//**************************************
-// Constants
-//**************************************
-#define MINMATCH 4
-
-#define HASH_LOG (MEMORY_USAGE-2)
-#define HASHTABLESIZE (1 << HASH_LOG)
-#define HASH_MASK (HASHTABLESIZE - 1)
-
-#define SKIPSTRENGTH (NOTCOMPRESSIBLE_DETECTIONLEVEL>2?NOTCOMPRESSIBLE_DETECTIONLEVEL:2)
-#define STACKLIMIT 13
-#define HEAPMODE (HASH_LOG>STACKLIMIT)  // Defines if memory is allocated into the stack (local variable), or into the heap (malloc()).
-#define COPYLENGTH 8
-#define LASTLITERALS 5
-#define MFLIMIT (COPYLENGTH+MINMATCH)
-#define MINLENGTH (MFLIMIT+1)
-
-#define MAXD_LOG 16
-#define MAX_DISTANCE ((1 << MAXD_LOG) - 1)
-
-#define ML_BITS 4
-#define ML_MASK ((1U<<ML_BITS)-1)
-#define RUN_BITS (8-ML_BITS)
-#define RUN_MASK ((1U<<RUN_BITS)-1)
-
-
-//**************************************
-// Architecture-specific macros
-//**************************************
-#if LZ4_ARCH64	// 64-bit
-#  define STEPSIZE 8
-#  define UARCH U64
-#  define AARCH A64
-#  define LZ4_COPYSTEP(s,d)		A64(d) = A64(s); d+=8; s+=8;
-#  define LZ4_COPYPACKET(s,d)		LZ4_COPYSTEP(s,d)
-#  define LZ4_SECURECOPY(s,d,e)	if (d<e) LZ4_WILDCOPY(s,d,e)
-#  define HTYPE U32
-#  define INITBASE(base)			const BYTE* const base = ip
-#else		// 32-bit
-#  define STEPSIZE 4
-#  define UARCH U32
-#  define AARCH A32
-#  define LZ4_COPYSTEP(s,d)		A32(d) = A32(s); d+=4; s+=4;
-#  define LZ4_COPYPACKET(s,d)		LZ4_COPYSTEP(s,d); LZ4_COPYSTEP(s,d);
-#  define LZ4_SECURECOPY			LZ4_WILDCOPY
-#  define HTYPE const BYTE*
-#  define INITBASE(base)			const int base = 0
-#endif
-
-#if (defined(LZ4_BIG_ENDIAN) && !defined(BIG_ENDIAN_NATIVE_BUT_INCOMPATIBLE))
-#  define LZ4_READ_LITTLEENDIAN_16(d,s,p) { U16 v = A16(p); v = lz4_bswap16(v); d = (s) - v; }
-#  define LZ4_WRITE_LITTLEENDIAN_16(p,i)  { U16 v = (U16)(i); v = lz4_bswap16(v); A16(p) = v; p+=2; }
-#else		// Little Endian
-#  define LZ4_READ_LITTLEENDIAN_16(d,s,p) { d = (s) - A16(p); }
-#  define LZ4_WRITE_LITTLEENDIAN_16(p,v)  { A16(p) = v; p+=2; }
-#endif
-
-
-//**************************************
-// Local structures
-//**************************************
-struct refTables
-{
-	HTYPE hashTable[HASHTABLESIZE];
-};
-
-
-//**************************************
-// Macros
-//**************************************
-#define LZ4_HASH_FUNCTION(i)	(((i) * 2654435761U) >> ((MINMATCH*8)-HASH_LOG))
-#define LZ4_HASH_VALUE(p)		LZ4_HASH_FUNCTION(A32(p))
-#define LZ4_WILDCOPY(s,d,e)		do { LZ4_COPYPACKET(s,d) } while (d<e);
-#define LZ4_BLINDCOPY(s,d,l)	{ BYTE* e=(d)+l; LZ4_WILDCOPY(s,d,e); d=e; }
-
-
-//****************************
-// Private functions
-//****************************
-#if LZ4_ARCH64
-
-static inline int LZ4_NbCommonBytes (register U64 val)
-{
-#if defined(LZ4_BIG_ENDIAN)
-    #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
-    unsigned long r = 0;
-    _BitScanReverse64( &r, val );
-    return (int)(r>>3);
-    #elif defined(__GNUC__) && (GCC_VERSION >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
-    return (__builtin_clzll(val) >> 3);
-    #else
-	int r;
-	if (!(val>>32)) { r=4; } else { r=0; val>>=32; }
-	if (!(val>>16)) { r+=2; val>>=8; } else { val>>=24; }
-	r += (!val);
-	return r;
-    #endif
-#else
-    #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
-    unsigned long r = 0;
-    _BitScanForward64( &r, val );
-    return (int)(r>>3);
-    #elif defined(__GNUC__) && (GCC_VERSION >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
-    return (__builtin_ctzll(val) >> 3);
-    #else
-	static const int DeBruijnBytePos[64] = { 0, 0, 0, 0, 0, 1, 1, 2, 0, 3, 1, 3, 1, 4, 2, 7, 0, 2, 3, 6, 1, 5, 3, 5, 1, 3, 4, 4, 2, 5, 6, 7, 7, 0, 1, 2, 3, 3, 4, 6, 2, 6, 5, 5, 3, 4, 5, 6, 7, 1, 2, 4, 6, 4, 4, 5, 7, 2, 6, 5, 7, 6, 7, 7 };
-	return DeBruijnBytePos[((U64)((val & -val) * 0x0218A392CDABBD3F)) >> 58];
-    #endif
-#endif
-}
-
-#else
-
-static inline int LZ4_NbCommonBytes (register U32 val)
-{
-#if defined(LZ4_BIG_ENDIAN)
-    #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
-    unsigned long r = 0;
-    _BitScanReverse( &r, val );
-    return (int)(r>>3);
-    #elif defined(__GNUC__) && (GCC_VERSION >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
-    return (__builtin_clz(val) >> 3);
-    #else
-	int r;
-	if (!(val>>16)) { r=2; val>>=8; } else { r=0; val>>=24; }
-	r += (!val);
-	return r;
-    #endif
-#else
-    #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
-    unsigned long r = 0;
-    _BitScanForward( &r, val );
-    return (int)(r>>3);
-    #elif defined(__GNUC__) && (GCC_VERSION >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
-    return (__builtin_ctz(val) >> 3);
-    #else
-	static const int DeBruijnBytePos[32] = { 0, 0, 3, 0, 3, 1, 3, 0, 3, 2, 2, 1, 3, 2, 0, 1, 3, 3, 1, 2, 2, 2, 2, 0, 3, 1, 2, 0, 1, 0, 1, 1 };
-	return DeBruijnBytePos[((U32)((val & -(S32)val) * 0x077CB531U)) >> 27];
-    #endif
-#endif
-}
-
-#endif
-
-
-
-//******************************
-// Compression functions
-//******************************
-
-// LZ4_compressCtx :
-// -----------------
-// Compress 'isize' bytes from 'source' into an output buffer 'dest' of maximum size 'maxOutputSize'.
-// If it cannot achieve it, compression will stop, and result of the function will be zero.
-// return : the number of bytes written in buffer 'dest', or 0 if the compression fails
-
-static inline int LZ4_compressCtx(void** ctx,
-				 const char* source,
-				 char* dest,
-				 int isize,
-				 int maxOutputSize)
-{
-#if HEAPMODE
-	struct refTables *srt = (struct refTables *) (*ctx);
-	HTYPE* HashTable;
-#else
-	HTYPE HashTable[HASHTABLESIZE] = {0};
-#endif
-
-	const BYTE* ip = (BYTE*) source;
-	INITBASE(base);
-	const BYTE* anchor = ip;
-	const BYTE* const iend = ip + isize;
-	const BYTE* const mflimit = iend - MFLIMIT;
-#define matchlimit (iend - LASTLITERALS)
-
-	BYTE* op = (BYTE*) dest;
-	BYTE* const oend = op + maxOutputSize;
-
-	int len, length;
-	const int skipStrength = SKIPSTRENGTH;
-	U32 forwardH;
-
-
-	// Init
-	if (isize<MINLENGTH) goto _last_literals;
-#if HEAPMODE
-	if (*ctx == NULL)
-	{
-		srt = (struct refTables *) malloc ( sizeof(struct refTables) );
-		*ctx = (void*) srt;
-	}
-	HashTable = (HTYPE*)(srt->hashTable);
-	memset((void*)HashTable, 0, sizeof(srt->hashTable));
-#else
-	(void) ctx;
-#endif
-
-
-	// First Byte
-	HashTable[LZ4_HASH_VALUE(ip)] = ip - base;
-	ip++; forwardH = LZ4_HASH_VALUE(ip);
-
-	// Main Loop
-    for ( ; ; )
-	{
-		int findMatchAttempts = (1U << skipStrength) + 3;
-		const BYTE* forwardIp = ip;
-		const BYTE* ref;
-		BYTE* token;
-
-		// Find a match
-		do {
-			U32 h = forwardH;
-			int step = findMatchAttempts++ >> skipStrength;
-			ip = forwardIp;
-			forwardIp = ip + step;
-
-			if unlikely(forwardIp > mflimit) { goto _last_literals; }
-
-			forwardH = LZ4_HASH_VALUE(forwardIp);
-			ref = base + HashTable[h];
-			HashTable[h] = ip - base;
-
-		} while ((ref < ip - MAX_DISTANCE) || (A32(ref) != A32(ip)));
-
-		// Catch up
-		while ((ip>anchor) && (ref>(BYTE*)source) && unlikely(ip[-1]==ref[-1])) { ip--; ref--; }
-
-		// Encode Literal length
-		length = (int)(ip - anchor);
-		token = op++;
-		if unlikely(op + length + (2 + 1 + LASTLITERALS) + (length>>8) >= oend) return 0; 		// Check output limit
-#ifdef _MSC_VER
-		if (length>=(int)RUN_MASK) 
-		{ 
-			int len = length-RUN_MASK; 
-			*token=(RUN_MASK<<ML_BITS); 
-			if (len>254)
-			{
-				do { *op++ = 255; len -= 255; } while (len>254);
-				*op++ = (BYTE)len; 
-				memcpy(op, anchor, length);
-				op += length;
-				goto _next_match;
-			}
-			else
-			*op++ = (BYTE)len; 
-		}
-		else *token = (length<<ML_BITS);
-#else
-		if (length>=(int)RUN_MASK) { *token=(RUN_MASK<<ML_BITS); len = length-RUN_MASK; for(; len > 254 ; len-=255) *op++ = 255; *op++ = (BYTE)len; }
-		else *token = (length<<ML_BITS);
-#endif
-
-		// Copy Literals
-		LZ4_BLINDCOPY(anchor, op, length);
-
-_next_match:
-		// Encode Offset
-		LZ4_WRITE_LITTLEENDIAN_16(op,(U16)(ip-ref));
-
-		// Start Counting
-		ip+=MINMATCH; ref+=MINMATCH;   // MinMatch verified
-		anchor = ip;
-		while likely(ip<matchlimit-(STEPSIZE-1))
-		{
-			UARCH diff = AARCH(ref) ^ AARCH(ip);
-			if (!diff) { ip+=STEPSIZE; ref+=STEPSIZE; continue; }
-			ip += LZ4_NbCommonBytes(diff);
-			goto _endCount;
-		}
-		if (LZ4_ARCH64) if ((ip<(matchlimit-3)) && (A32(ref) == A32(ip))) { ip+=4; ref+=4; }
-		if ((ip<(matchlimit-1)) && (A16(ref) == A16(ip))) { ip+=2; ref+=2; }
-		if ((ip<matchlimit) && (*ref == *ip)) ip++;
-_endCount:
-
-		// Encode MatchLength
-		len = (int)(ip - anchor);
-		if (len>=(int)ML_MASK) { *token+=ML_MASK; len-=ML_MASK; for(; len > 509 ; len-=510) { *op++ = 255; *op++ = 255; } if (len > 254) { len-=255; *op++ = 255; } *op++ = (BYTE)len; }
-		else *token += len;
-
-		// Test end of chunk
-		if (ip > mflimit) { anchor = ip;  break; }
-
-		// Fill table
-		HashTable[LZ4_HASH_VALUE(ip-2)] = ip - 2 - base;
-
-		// Test next position
-		ref = base + HashTable[LZ4_HASH_VALUE(ip)];
-		HashTable[LZ4_HASH_VALUE(ip)] = ip - base;
-		if ((ref > ip - (MAX_DISTANCE + 1)) && (A32(ref) == A32(ip))) { token = op++; *token=0; goto _next_match; }
-
-		// Prepare next loop
-		anchor = ip++;
-		forwardH = LZ4_HASH_VALUE(ip);
-	}
-
-_last_literals:
-	// Encode Last Literals
-	{
-		int lastRun = (int)(iend - anchor);
-		if (((char*)op - dest) + lastRun + 1 + ((lastRun-15)/255) >= maxOutputSize) return 0;
-		if (lastRun>=(int)RUN_MASK) { *op++=(RUN_MASK<<ML_BITS); lastRun-=RUN_MASK; for(; lastRun > 254 ; lastRun-=255) *op++ = 255; *op++ = (BYTE) lastRun; }
-		else *op++ = (lastRun<<ML_BITS);
-		memcpy(op, anchor, iend - anchor);
-		op += iend-anchor;
-	}
-
-	// End
-	return (int) (((char*)op)-dest);
-}
-
-
-
-// Note : this function is valid only if isize < LZ4_64KLIMIT
-#define LZ4_64KLIMIT ((1<<16) + (MFLIMIT-1))
-#define HASHLOG64K (HASH_LOG+1)
-#define HASH64KTABLESIZE (1U<<HASHLOG64K)
-#define LZ4_HASH64K_FUNCTION(i)	(((i) * 2654435761U) >> ((MINMATCH*8)-HASHLOG64K))
-#define LZ4_HASH64K_VALUE(p)	LZ4_HASH64K_FUNCTION(A32(p))
-static inline int LZ4_compress64kCtx(void** ctx,
-				 const char* source,
-				 char* dest,
-				 int isize,
-				 int maxOutputSize)
-{
-#if HEAPMODE
-	struct refTables *srt = (struct refTables *) (*ctx);
-	U16* HashTable;
-#else
-	U16 HashTable[HASH64KTABLESIZE] = {0};
-#endif
-
-	const BYTE* ip = (BYTE*) source;
-	const BYTE* anchor = ip;
-	const BYTE* const base = ip;
-	const BYTE* const iend = ip + isize;
-	const BYTE* const mflimit = iend - MFLIMIT;
-#define matchlimit (iend - LASTLITERALS)
-
-	BYTE* op = (BYTE*) dest;
-	BYTE* const oend = op + maxOutputSize;
-
-	int len, length;
-	const int skipStrength = SKIPSTRENGTH;
-	U32 forwardH;
-
-
-	// Init
-	if (isize<MINLENGTH) goto _last_literals;
-#if HEAPMODE
-	if (*ctx == NULL)
-	{
-		srt = (struct refTables *) malloc ( sizeof(struct refTables) );
-		*ctx = (void*) srt;
-	}
-	HashTable = (U16*)(srt->hashTable);
-	memset((void*)HashTable, 0, sizeof(srt->hashTable));
-#else
-	(void) ctx;
-#endif
-
-
-	// First Byte
-	ip++; forwardH = LZ4_HASH64K_VALUE(ip);
-
-	// Main Loop
-    for ( ; ; )
-	{
-		int findMatchAttempts = (1U << skipStrength) + 3;
-		const BYTE* forwardIp = ip;
-		const BYTE* ref;
-		BYTE* token;
-
-		// Find a match
-		do {
-			U32 h = forwardH;
-			int step = findMatchAttempts++ >> skipStrength;
-			ip = forwardIp;
-			forwardIp = ip + step;
-
-			if (forwardIp > mflimit) { goto _last_literals; }
-
-			forwardH = LZ4_HASH64K_VALUE(forwardIp);
-			ref = base + HashTable[h];
-			HashTable[h] = (U16)(ip - base);
-
-		} while (A32(ref) != A32(ip));
-
-		// Catch up
-		while ((ip>anchor) && (ref>(BYTE*)source) && (ip[-1]==ref[-1])) { ip--; ref--; }
-
-		// Encode Literal length
-		length = (int)(ip - anchor);
-		token = op++;
-		if unlikely(op + length + (2 + 1 + LASTLITERALS) + (length>>8) >= oend) return 0; 		// Check output limit
-#ifdef _MSC_VER
-		if (length>=(int)RUN_MASK) 
-		{ 
-			int len = length-RUN_MASK; 
-			*token=(RUN_MASK<<ML_BITS); 
-			if (len>254)
-			{
-				do { *op++ = 255; len -= 255; } while (len>254);
-				*op++ = (BYTE)len; 
-				memcpy(op, anchor, length);
-				op += length;
-				goto _next_match;
-			}
-			else
-			*op++ = (BYTE)len; 
-		}
-		else *token = (length<<ML_BITS);
-#else
-		if (length>=(int)RUN_MASK) { *token=(RUN_MASK<<ML_BITS); len = length-RUN_MASK; for(; len > 254 ; len-=255) *op++ = 255; *op++ = (BYTE)len; }
-		else *token = (length<<ML_BITS);
-#endif
-
-		// Copy Literals
-		LZ4_BLINDCOPY(anchor, op, length);
-
-_next_match:
-		// Encode Offset
-		LZ4_WRITE_LITTLEENDIAN_16(op,(U16)(ip-ref));
-
-		// Start Counting
-		ip+=MINMATCH; ref+=MINMATCH;   // MinMatch verified
-		anchor = ip;
-		while (ip<matchlimit-(STEPSIZE-1))
-		{
-			UARCH diff = AARCH(ref) ^ AARCH(ip);
-			if (!diff) { ip+=STEPSIZE; ref+=STEPSIZE; continue; }
-			ip += LZ4_NbCommonBytes(diff);
-			goto _endCount;
-		}
-		if (LZ4_ARCH64) if ((ip<(matchlimit-3)) && (A32(ref) == A32(ip))) { ip+=4; ref+=4; }
-		if ((ip<(matchlimit-1)) && (A16(ref) == A16(ip))) { ip+=2; ref+=2; }
-		if ((ip<matchlimit) && (*ref == *ip)) ip++;
-_endCount:
-
-		// Encode MatchLength
-		len = (int)(ip - anchor);
-		if (len>=(int)ML_MASK) { *token+=ML_MASK; len-=ML_MASK; for(; len > 509 ; len-=510) { *op++ = 255; *op++ = 255; } if (len > 254) { len-=255; *op++ = 255; } *op++ = (BYTE)len; }
-		else *token += len;
-
-		// Test end of chunk
-		if (ip > mflimit) { anchor = ip;  break; }
-
-		// Fill table
-		HashTable[LZ4_HASH64K_VALUE(ip-2)] = (U16)(ip - 2 - base);
-
-		// Test next position
-		ref = base + HashTable[LZ4_HASH64K_VALUE(ip)];
-		HashTable[LZ4_HASH64K_VALUE(ip)] = (U16)(ip - base);
-		if (A32(ref) == A32(ip)) { token = op++; *token=0; goto _next_match; }
-
-		// Prepare next loop
-		anchor = ip++;
-		forwardH = LZ4_HASH64K_VALUE(ip);
-	}
-
-_last_literals:
-	// Encode Last Literals
-	{
-		int lastRun = (int)(iend - anchor);
-		if (((char*)op - dest) + lastRun + 1 + ((lastRun)>>8) >= maxOutputSize) return 0;
-		if (lastRun>=(int)RUN_MASK) { *op++=(RUN_MASK<<ML_BITS); lastRun-=RUN_MASK; for(; lastRun > 254 ; lastRun-=255) *op++ = 255; *op++ = (BYTE) lastRun; }
-		else *op++ = (lastRun<<ML_BITS);
-		memcpy(op, anchor, iend - anchor);
-		op += iend-anchor;
-	}
-
-	// End
-	return (int) (((char*)op)-dest);
-}
-
-
-int LZ4_compress_limitedOutput(const char* source, 
-							   char* dest, 
-							   int isize, 
-							   int maxOutputSize)
-{
-#if HEAPMODE
-	void* ctx = malloc(sizeof(struct refTables));
-	int result;
-	if (isize < LZ4_64KLIMIT)
-		result = LZ4_compress64kCtx(&ctx, source, dest, isize, maxOutputSize);
-	else result = LZ4_compressCtx(&ctx, source, dest, isize, maxOutputSize);
-	free(ctx);
-	return result;
-#else
-	if (isize < (int)LZ4_64KLIMIT) return LZ4_compress64kCtx(NULL, source, dest, isize, maxOutputSize);
-	return LZ4_compressCtx(NULL, source, dest, isize, maxOutputSize);
-#endif
-}
-
-
-int LZ4_compress(const char* source,
-				 char* dest,
-				 int isize)
-{
-	return LZ4_compress_limitedOutput(source, dest, isize, LZ4_compressBound(isize));
-}
-
-
-
-
-//****************************
-// Decompression functions
-//****************************
-
-// Note : The decoding functions LZ4_uncompress() and LZ4_uncompress_unknownOutputSize()
-//		are safe against "buffer overflow" attack type.
-//		They will never write nor read outside of the provided output buffers.
-//      LZ4_uncompress_unknownOutputSize() also insures that it will never read outside of the input buffer.
-//		A corrupted input will produce an error result, a negative int, indicating the position of the error within input stream.
-
-int LZ4_uncompress(const char* source,
-				 char* dest,
-				 int osize)
-{
-	// Local Variables
-	const BYTE* restrict ip = (const BYTE*) source;
-	const BYTE* restrict ref;
-
-	BYTE* restrict op = (BYTE*) dest;
-	BYTE* const oend = op + osize;
-	BYTE* cpy;
-
-	BYTE token;
-
-	int	len, length;
-	size_t dec[] ={0, 3, 2, 3, 0, 0, 0, 0};
-
-
-	// Main Loop
-	while (1)
-	{
-		// get runlength
-		token = *ip++;
-		if ((length=(token>>ML_BITS)) == RUN_MASK)  { for (;(len=*ip++)==255;length+=255){} length += len; }
-
-		// copy literals
-		cpy = op+length;
-		if unlikely(cpy>oend-COPYLENGTH)
-		{
-			if (cpy > oend) goto _output_error;          // Error : request to write beyond destination buffer
-			memcpy(op, ip, length);
-			ip += length;
-			break;    // Necessarily EOF
-		}
-		LZ4_WILDCOPY(ip, op, cpy); ip -= (op-cpy); op = cpy;
-
-		// get offset
-		LZ4_READ_LITTLEENDIAN_16(ref,cpy,ip); ip+=2;
-		if (ref < (BYTE* const)dest) goto _output_error;   // Error : offset create reference outside destination buffer
-
-		// get matchlength
-		if ((length=(token&ML_MASK)) == ML_MASK) { for (;*ip==255;length+=255) {ip++;} length += *ip++; }
-
-		// copy repeated sequence
-		if unlikely(op-ref<STEPSIZE)
-		{
-#if LZ4_ARCH64
-			size_t dec2table[]={0, 0, 0, -1, 0, 1, 2, 3};
-			size_t dec2 = dec2table[op-ref];
-#else
-			const int dec2 = 0;
-#endif
-			*op++ = *ref++;
-			*op++ = *ref++;
-			*op++ = *ref++;
-			*op++ = *ref++;
-			ref -= dec[op-ref];
-			A32(op)=A32(ref); op += STEPSIZE-4;
-			ref -= dec2;
-		} else { LZ4_COPYSTEP(ref,op); }
-		cpy = op + length - (STEPSIZE-4);
-		if (cpy>oend-COPYLENGTH)
-		{
-			if (cpy > oend) goto _output_error;             // Error : request to write beyond destination buffer
-			LZ4_SECURECOPY(ref, op, (oend-COPYLENGTH));
-			while(op<cpy) *op++=*ref++;
-			op=cpy;
-			if (op == oend) break;    // Check EOF (should never happen, since last 5 bytes are supposed to be literals)
-			continue;
-		}
-		LZ4_SECURECOPY(ref, op, cpy);
-		op=cpy;		// correction
-	}
-
-	// end of decoding
-	return (int) (((char*)ip)-source);
-
-	// write overflow error detected
-_output_error:
-	return (int) (-(((char*)ip)-source));
-}
-
-
-int LZ4_uncompress_unknownOutputSize(
-				const char* source,
-				char* dest,
-				int isize,
-				int maxOutputSize)
-{
-	// Local Variables
-	const BYTE* restrict ip = (const BYTE*) source;
-	const BYTE* const iend = ip + isize;
-	const BYTE* restrict ref;
-
-	BYTE* restrict op = (BYTE*) dest;
-	BYTE* const oend = op + maxOutputSize;
-	BYTE* cpy;
-
-	size_t dec[] ={0, 3, 2, 3, 0, 0, 0, 0};
-
-
-	// Main Loop
-	while (ip<iend)
-	{
-		BYTE token;
-		int length;
-
-		// get runlength
-		token = *ip++;
-		if ((length=(token>>ML_BITS)) == RUN_MASK) { int s=255; while ((ip<iend) && (s==255)) { s=*ip++; length += s; } }
-
-		// copy literals
-		cpy = op+length;
-		if ((cpy>oend-COPYLENGTH) || (ip+length>iend-COPYLENGTH))
-		{
-			if (cpy > oend) goto _output_error;          // Error : request to write beyond destination buffer
-			if (ip+length > iend) goto _output_error;    // Error : request to read beyond source buffer
-			memcpy(op, ip, length);
-			op += length;
-			ip += length;
-			if (ip<iend) goto _output_error;             // Error : LZ4 format violation
-			break;    // Necessarily EOF, due to parsing restrictions
-		}
-		LZ4_WILDCOPY(ip, op, cpy); ip -= (op-cpy); op = cpy;
-
-		// get offset
-		LZ4_READ_LITTLEENDIAN_16(ref,cpy,ip); ip+=2;
-		if (ref < (BYTE* const)dest) goto _output_error;   // Error : offset creates reference outside of destination buffer
-
-		// get matchlength
-		if ((length=(token&ML_MASK)) == ML_MASK) { while (ip<iend) { int s = *ip++; length +=s; if (s==255) continue; break; } }
-
-		// copy repeated sequence
-		if unlikely(op-ref<STEPSIZE)
-		{
-#if LZ4_ARCH64
-			size_t dec2table[]={0, 0, 0, -1, 0, 1, 2, 3};
-			size_t dec2 = dec2table[op-ref];
-#else
-			const int dec2 = 0;
-#endif
-			*op++ = *ref++;
-			*op++ = *ref++;
-			*op++ = *ref++;
-			*op++ = *ref++;
-			ref -= dec[op-ref];
-			A32(op)=A32(ref); op += STEPSIZE-4;
-			ref -= dec2;
-		} else { LZ4_COPYSTEP(ref,op); }
-		cpy = op + length - (STEPSIZE-4);
-		if (cpy>oend-COPYLENGTH)
-		{
-			if (cpy > oend) goto _output_error;           // Error : request to write outside of destination buffer
-			LZ4_SECURECOPY(ref, op, (oend-COPYLENGTH));
-			while(op<cpy) *op++=*ref++;
-			op=cpy;
-			if (op == oend) break;    // Check EOF (should never happen, since last 5 bytes are supposed to be literals)
-			continue;
-		}
-		LZ4_SECURECOPY(ref, op, cpy);
-		op=cpy;		// correction
-	}
-
-	// end of decoding
-	return (int) (((char*)op)-dest);
-
-	// write overflow error detected
-_output_error:
-	return (int) (-(((char*)ip)-source));
-}
+/*
+   LZ4 - Fast LZ compression algorithm
+   Copyright (C) 2011-2012, Yann Collet.
+   BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
+
+   Redistribution and use in source and binary forms, with or without
+   modification, are permitted provided that the following conditions are
+   met:
+
+       * Redistributions of source code must retain the above copyright
+   notice, this list of conditions and the following disclaimer.
+       * Redistributions in binary form must reproduce the above
+   copyright notice, this list of conditions and the following disclaimer
+   in the documentation and/or other materials provided with the
+   distribution.
+
+   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+   You can contact the author at :
+   - LZ4 homepage : http://fastcompression.blogspot.com/p/lz4.html
+   - LZ4 source repository : http://code.google.com/p/lz4/
+*/
+
+//**************************************
+// Tuning parameters
+//**************************************
+// MEMORY_USAGE :
+// Memory usage formula : N->2^N Bytes (examples : 10 -> 1KB; 12 -> 4KB ; 16 -> 64KB; 20 -> 1MB; etc.)
+// Increasing memory usage improves compression ratio
+// Reduced memory usage can improve speed, due to cache effect
+// Default value is 14, for 16KB, which nicely fits into Intel x86 L1 cache
+#define MEMORY_USAGE 14
+
+// NOTCOMPRESSIBLE_DETECTIONLEVEL :
+// Decreasing this value will make the algorithm skip faster data segments considered "incompressible"
+// This may decrease compression ratio dramatically, but will be faster on incompressible data
+// Increasing this value will make the algorithm search more before declaring a segment "incompressible"
+// This could improve compression a bit, but will be slower on incompressible data
+// The default value (6) is recommended
+#define NOTCOMPRESSIBLE_DETECTIONLEVEL 6
+
+// BIG_ENDIAN_NATIVE_BUT_INCOMPATIBLE :
+// This will provide a small boost to performance for big endian cpu, but the resulting compressed stream will be incompatible with little-endian CPU.
+// You can set this option to 1 in situations where data will remain within closed environment
+// This option is useless on Little_Endian CPU (such as x86)
+//#define BIG_ENDIAN_NATIVE_BUT_INCOMPATIBLE 1
+
+
+
+//**************************************
+// CPU Feature Detection
+//**************************************
+// 32 or 64 bits ?
+#if (defined(__x86_64__) || defined(__x86_64) || defined(__amd64__) || defined(__amd64) || defined(__ppc64__) || defined(_WIN64) || defined(__LP64__) || defined(_LP64) )   // Detects 64 bits mode
+#  define LZ4_ARCH64 1
+#else
+#  define LZ4_ARCH64 0
+#endif
+
+// Little Endian or Big Endian ?
+// Overwrite the #define below if you know your architecture endianess
+#if defined (__GLIBC__)
+#  include <endian.h>
+#  if (__BYTE_ORDER == __BIG_ENDIAN)
+#     define LZ4_BIG_ENDIAN 1
+#  endif
+#elif (defined(__BIG_ENDIAN__) || defined(__BIG_ENDIAN) || defined(_BIG_ENDIAN)) && !(defined(__LITTLE_ENDIAN__) || defined(__LITTLE_ENDIAN) || defined(_LITTLE_ENDIAN))
+#  define LZ4_BIG_ENDIAN 1
+#elif defined(__sparc) || defined(__sparc__) \
+   || defined(__ppc__) || defined(_POWER) || defined(__powerpc__) || defined(_ARCH_PPC) || defined(__PPC__) || defined(__PPC) || defined(PPC) || defined(__powerpc__) || defined(__powerpc) || defined(powerpc) \
+   || defined(__hpux)  || defined(__hppa) \
+   || defined(_MIPSEB) || defined(__s390__)
+#  define LZ4_BIG_ENDIAN 1
+#else
+// Little Endian assumed. PDP Endian and other very rare endian format are unsupported.
+#endif
+
+// Unaligned memory access is automatically enabled for "common" CPU, such as x86.
+// For others CPU, the compiler will be more cautious, and insert extra code to ensure aligned access is respected
+// If you know your target CPU supports unaligned memory access, you want to force this option manually to improve performance
+#if defined(__ARM_FEATURE_UNALIGNED)
+#  define LZ4_FORCE_UNALIGNED_ACCESS 1
+#endif
+
+// Define this parameter if your target system or compiler does not support hardware bit count
+#if defined(_MSC_VER) && defined(_WIN32_WCE)            // Visual Studio for Windows CE does not support Hardware bit count
+#  define LZ4_FORCE_SW_BITCOUNT
+#endif
+
+
+//**************************************
+// Compiler Options
+//**************************************
+#if __STDC_VERSION__ >= 199901L   // C99
+/* "restrict" is a known keyword */
+#else
+#  define restrict // Disable restrict
+#endif
+
+#define GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__)
+
+#ifdef _MSC_VER  // Visual Studio
+#  include <intrin.h>   // For Visual 2005
+#  if LZ4_ARCH64	// 64-bit
+#    pragma intrinsic(_BitScanForward64) // For Visual 2005
+#    pragma intrinsic(_BitScanReverse64) // For Visual 2005
+#  else
+#    pragma intrinsic(_BitScanForward)   // For Visual 2005
+#    pragma intrinsic(_BitScanReverse)   // For Visual 2005
+#  endif
+#endif
+
+#ifdef _MSC_VER
+#  define lz4_bswap16(x) _byteswap_ushort(x)
+#else
+#  define lz4_bswap16(x) ((unsigned short int) ((((x) >> 8) & 0xffu) | (((x) & 0xffu) << 8)))
+#endif
+
+#if (GCC_VERSION >= 302) || (__INTEL_COMPILER >= 800) || defined(__clang__)
+#  define expect(expr,value)    (__builtin_expect ((expr),(value)) )
+#else
+#  define expect(expr,value)    (expr)
+#endif
+
+#define likely(expr)     expect((expr) != 0, 1)
+#define unlikely(expr)   expect((expr) != 0, 0)
+
+
+//**************************************
+// Includes
+//**************************************
+#include <stdlib.h>   // for malloc
+#include <string.h>   // for memset
+#include "lz4.h"
+
+
+//**************************************
+// Basic Types
+//**************************************
+#if defined(_MSC_VER)    // Visual Studio does not support 'stdint' natively
+#  define BYTE	unsigned __int8
+#  define U16		unsigned __int16
+#  define U32		unsigned __int32
+#  define S32		__int32
+#  define U64		unsigned __int64
+#else
+#  include <stdint.h>
+#  define BYTE	uint8_t
+#  define U16		uint16_t
+#  define U32		uint32_t
+#  define S32		int32_t
+#  define U64		uint64_t
+#endif
+
+#ifndef LZ4_FORCE_UNALIGNED_ACCESS
+#  pragma pack(push, 1)
+#endif
+
+typedef struct _U16_S { U16 v; } U16_S;
+typedef struct _U32_S { U32 v; } U32_S;
+typedef struct _U64_S { U64 v; } U64_S;
+
+#ifndef LZ4_FORCE_UNALIGNED_ACCESS
+#  pragma pack(pop)
+#endif
+
+#define A64(x) (((U64_S *)(x))->v)
+#define A32(x) (((U32_S *)(x))->v)
+#define A16(x) (((U16_S *)(x))->v)
+
+
+//**************************************
+// Constants
+//**************************************
+#define MINMATCH 4
+
+#define HASH_LOG (MEMORY_USAGE-2)
+#define HASHTABLESIZE (1 << HASH_LOG)
+#define HASH_MASK (HASHTABLESIZE - 1)
+
+#define SKIPSTRENGTH (NOTCOMPRESSIBLE_DETECTIONLEVEL>2?NOTCOMPRESSIBLE_DETECTIONLEVEL:2)
+#define STACKLIMIT 13
+#define HEAPMODE (HASH_LOG>STACKLIMIT)  // Defines if memory is allocated into the stack (local variable), or into the heap (malloc()).
+#define COPYLENGTH 8
+#define LASTLITERALS 5
+#define MFLIMIT (COPYLENGTH+MINMATCH)
+#define MINLENGTH (MFLIMIT+1)
+
+#define MAXD_LOG 16
+#define MAX_DISTANCE ((1 << MAXD_LOG) - 1)
+
+#define ML_BITS  4
+#define ML_MASK  ((1U<<ML_BITS)-1)
+#define RUN_BITS (8-ML_BITS)
+#define RUN_MASK ((1U<<RUN_BITS)-1)
+
+
+//**************************************
+// Architecture-specific macros
+//**************************************
+#if LZ4_ARCH64	// 64-bit
+#  define STEPSIZE 8
+#  define UARCH U64
+#  define AARCH A64
+#  define LZ4_COPYSTEP(s,d)       A64(d) = A64(s); d+=8; s+=8;
+#  define LZ4_COPYPACKET(s,d)     LZ4_COPYSTEP(s,d)
+#  define LZ4_SECURECOPY(s,d,e)   if (d<e) LZ4_WILDCOPY(s,d,e)
+#  define HTYPE                   U32
+#  define INITBASE(base)          const BYTE* const base = ip
+#else		// 32-bit
+#  define STEPSIZE 4
+#  define UARCH U32
+#  define AARCH A32
+#  define LZ4_COPYSTEP(s,d)       A32(d) = A32(s); d+=4; s+=4;
+#  define LZ4_COPYPACKET(s,d)     LZ4_COPYSTEP(s,d); LZ4_COPYSTEP(s,d);
+#  define LZ4_SECURECOPY          LZ4_WILDCOPY
+#  define HTYPE                   const BYTE*
+#  define INITBASE(base)          const int base = 0
+#endif
+
+#if (defined(LZ4_BIG_ENDIAN) && !defined(BIG_ENDIAN_NATIVE_BUT_INCOMPATIBLE))
+#  define LZ4_READ_LITTLEENDIAN_16(d,s,p) { U16 v = A16(p); v = lz4_bswap16(v); d = (s) - v; }
+#  define LZ4_WRITE_LITTLEENDIAN_16(p,i)  { U16 v = (U16)(i); v = lz4_bswap16(v); A16(p) = v; p+=2; }
+#else		// Little Endian
+#  define LZ4_READ_LITTLEENDIAN_16(d,s,p) { d = (s) - A16(p); }
+#  define LZ4_WRITE_LITTLEENDIAN_16(p,v)  { A16(p) = v; p+=2; }
+#endif
+
+
+//**************************************
+// Local structures
+//**************************************
+struct refTables
+{
+    HTYPE hashTable[HASHTABLESIZE];
+};
+
+
+//**************************************
+// Macros
+//**************************************
+#define LZ4_HASH_FUNCTION(i)	(((i) * 2654435761U) >> ((MINMATCH*8)-HASH_LOG))
+#define LZ4_HASH_VALUE(p)		LZ4_HASH_FUNCTION(A32(p))
+#define LZ4_WILDCOPY(s,d,e)		do { LZ4_COPYPACKET(s,d) } while (d<e);
+#define LZ4_BLINDCOPY(s,d,l)	{ BYTE* e=(d)+l; LZ4_WILDCOPY(s,d,e); d=e; }
+
+
+//****************************
+// Private functions
+//****************************
+#if LZ4_ARCH64
+
+static inline int LZ4_NbCommonBytes (register U64 val)
+{
+#if defined(LZ4_BIG_ENDIAN)
+    #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
+    unsigned long r = 0;
+    _BitScanReverse64( &r, val );
+    return (int)(r>>3);
+    #elif defined(__GNUC__) && (GCC_VERSION >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
+    return (__builtin_clzll(val) >> 3);
+    #else
+    int r;
+    if (!(val>>32)) { r=4; } else { r=0; val>>=32; }
+    if (!(val>>16)) { r+=2; val>>=8; } else { val>>=24; }
+    r += (!val);
+    return r;
+    #endif
+#else
+    #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
+    unsigned long r = 0;
+    _BitScanForward64( &r, val );
+    return (int)(r>>3);
+    #elif defined(__GNUC__) && (GCC_VERSION >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
+    return (__builtin_ctzll(val) >> 3);
+    #else
+    static const int DeBruijnBytePos[64] = { 0, 0, 0, 0, 0, 1, 1, 2, 0, 3, 1, 3, 1, 4, 2, 7, 0, 2, 3, 6, 1, 5, 3, 5, 1, 3, 4, 4, 2, 5, 6, 7, 7, 0, 1, 2, 3, 3, 4, 6, 2, 6, 5, 5, 3, 4, 5, 6, 7, 1, 2, 4, 6, 4, 4, 5, 7, 2, 6, 5, 7, 6, 7, 7 };
+    return DeBruijnBytePos[((U64)((val & -val) * 0x0218A392CDABBD3F)) >> 58];
+    #endif
+#endif
+}
+
+#else
+
+static inline int LZ4_NbCommonBytes (register U32 val)
+{
+#if defined(LZ4_BIG_ENDIAN)
+    #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
+    unsigned long r = 0;
+    _BitScanReverse( &r, val );
+    return (int)(r>>3);
+    #elif defined(__GNUC__) && (GCC_VERSION >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
+    return (__builtin_clz(val) >> 3);
+    #else
+    int r;
+    if (!(val>>16)) { r=2; val>>=8; } else { r=0; val>>=24; }
+    r += (!val);
+    return r;
+    #endif
+#else
+    #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
+    unsigned long r = 0;
+    _BitScanForward( &r, val );
+    return (int)(r>>3);
+    #elif defined(__GNUC__) && (GCC_VERSION >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
+    return (__builtin_ctz(val) >> 3);
+    #else
+    static const int DeBruijnBytePos[32] = { 0, 0, 3, 0, 3, 1, 3, 0, 3, 2, 2, 1, 3, 2, 0, 1, 3, 3, 1, 2, 2, 2, 2, 0, 3, 1, 2, 0, 1, 0, 1, 1 };
+    return DeBruijnBytePos[((U32)((val & -(S32)val) * 0x077CB531U)) >> 27];
+    #endif
+#endif
+}
+
+#endif
+
+
+
+//******************************
+// Compression functions
+//******************************
+
+// LZ4_compressCtx :
+// -----------------
+// Compress 'isize' bytes from 'source' into an output buffer 'dest' of maximum size 'maxOutputSize'.
+// If it cannot achieve it, compression will stop, and result of the function will be zero.
+// return : the number of bytes written in buffer 'dest', or 0 if the compression fails
+
+static inline int LZ4_compressCtx(void** ctx,
+                 const char* source,
+                 char* dest,
+                 int isize,
+                 int maxOutputSize)
+{
+#if HEAPMODE
+    struct refTables *srt = (struct refTables *) (*ctx);
+    HTYPE* HashTable;
+#else
+    HTYPE HashTable[HASHTABLESIZE] = {0};
+#endif
+
+    const BYTE* ip = (BYTE*) source;
+    INITBASE(base);
+    const BYTE* anchor = ip;
+    const BYTE* const iend = ip + isize;
+    const BYTE* const mflimit = iend - MFLIMIT;
+#define matchlimit (iend - LASTLITERALS)
+
+    BYTE* op = (BYTE*) dest;
+    BYTE* const oend = op + maxOutputSize;
+
+    int len, length;
+    const int skipStrength = SKIPSTRENGTH;
+    U32 forwardH;
+
+
+    // Init
+    if (isize<MINLENGTH) goto _last_literals;
+#if HEAPMODE
+    if (*ctx == NULL)
+    {
+        srt = (struct refTables *) malloc ( sizeof(struct refTables) );
+        *ctx = (void*) srt;
+    }
+    HashTable = (HTYPE*)(srt->hashTable);
+    memset((void*)HashTable, 0, sizeof(srt->hashTable));
+#else
+    (void) ctx;
+#endif
+
+
+    // First Byte
+    HashTable[LZ4_HASH_VALUE(ip)] = ip - base;
+    ip++; forwardH = LZ4_HASH_VALUE(ip);
+
+    // Main Loop
+    for ( ; ; )
+    {
+        int findMatchAttempts = (1U << skipStrength) + 3;
+        const BYTE* forwardIp = ip;
+        const BYTE* ref;
+        BYTE* token;
+
+        // Find a match
+        do {
+            U32 h = forwardH;
+            int step = findMatchAttempts++ >> skipStrength;
+            ip = forwardIp;
+            forwardIp = ip + step;
+
+            if unlikely(forwardIp > mflimit) { goto _last_literals; }
+
+            forwardH = LZ4_HASH_VALUE(forwardIp);
+            ref = base + HashTable[h];
+            HashTable[h] = ip - base;
+
+        } while ((ref < ip - MAX_DISTANCE) || (A32(ref) != A32(ip)));
+
+        // Catch up
+        while ((ip>anchor) && (ref>(BYTE*)source) && unlikely(ip[-1]==ref[-1])) { ip--; ref--; }
+
+        // Encode Literal length
+        length = (int)(ip - anchor);
+        token = op++;
+        if unlikely(op + length + (2 + 1 + LASTLITERALS) + (length>>8) > oend) return 0; 		// Check output limit
+#ifdef _MSC_VER
+        if (length>=(int)RUN_MASK) 
+        { 
+            int len = length-RUN_MASK; 
+            *token=(RUN_MASK<<ML_BITS); 
+            if (len>254)
+            {
+                do { *op++ = 255; len -= 255; } while (len>254);
+                *op++ = (BYTE)len; 
+                memcpy(op, anchor, length);
+                op += length;
+                goto _next_match;
+            }
+            else
+            *op++ = (BYTE)len; 
+        }
+        else *token = (length<<ML_BITS);
+#else
+        if (length>=(int)RUN_MASK) { *token=(RUN_MASK<<ML_BITS); len = length-RUN_MASK; for(; len > 254 ; len-=255) *op++ = 255; *op++ = (BYTE)len; }
+        else *token = (length<<ML_BITS);
+#endif
+
+        // Copy Literals
+        LZ4_BLINDCOPY(anchor, op, length);
+
+_next_match:
+        // Encode Offset
+        LZ4_WRITE_LITTLEENDIAN_16(op,(U16)(ip-ref));
+
+        // Start Counting
+        ip+=MINMATCH; ref+=MINMATCH;   // MinMatch verified
+        anchor = ip;
+        while likely(ip<matchlimit-(STEPSIZE-1))
+        {
+            UARCH diff = AARCH(ref) ^ AARCH(ip);
+            if (!diff) { ip+=STEPSIZE; ref+=STEPSIZE; continue; }
+            ip += LZ4_NbCommonBytes(diff);
+            goto _endCount;
+        }
+        if (LZ4_ARCH64) if ((ip<(matchlimit-3)) && (A32(ref) == A32(ip))) { ip+=4; ref+=4; }
+        if ((ip<(matchlimit-1)) && (A16(ref) == A16(ip))) { ip+=2; ref+=2; }
+        if ((ip<matchlimit) && (*ref == *ip)) ip++;
+_endCount:
+
+        // Encode MatchLength
+        len = (int)(ip - anchor);
+        if unlikely(op + (1 + LASTLITERALS) + (len>>8) > oend) return 0; 		// Check output limit
+        if (len>=(int)ML_MASK) { *token+=ML_MASK; len-=ML_MASK; for(; len > 509 ; len-=510) { *op++ = 255; *op++ = 255; } if (len > 254) { len-=255; *op++ = 255; } *op++ = (BYTE)len; }
+        else *token += len;
+
+        // Test end of chunk
+        if (ip > mflimit) { anchor = ip;  break; }
+
+        // Fill table
+        HashTable[LZ4_HASH_VALUE(ip-2)] = ip - 2 - base;
+
+        // Test next position
+        ref = base + HashTable[LZ4_HASH_VALUE(ip)];
+        HashTable[LZ4_HASH_VALUE(ip)] = ip - base;
+        if ((ref > ip - (MAX_DISTANCE + 1)) && (A32(ref) == A32(ip))) { token = op++; *token=0; goto _next_match; }
+
+        // Prepare next loop
+        anchor = ip++;
+        forwardH = LZ4_HASH_VALUE(ip);
+    }
+
+_last_literals:
+    // Encode Last Literals
+    {
+        int lastRun = (int)(iend - anchor);
+        if (((char*)op - dest) + lastRun + 1 + ((lastRun+255-RUN_MASK)/255) > (U32)maxOutputSize) return 0;
+        if (lastRun>=(int)RUN_MASK) { *op++=(RUN_MASK<<ML_BITS); lastRun-=RUN_MASK; for(; lastRun > 254 ; lastRun-=255) *op++ = 255; *op++ = (BYTE) lastRun; }
+        else *op++ = (lastRun<<ML_BITS);
+        memcpy(op, anchor, iend - anchor);
+        op += iend-anchor;
+    }
+
+    // End
+    return (int) (((char*)op)-dest);
+}
+
+
+
+// Note : this function is valid only if isize < LZ4_64KLIMIT
+#define LZ4_64KLIMIT ((1<<16) + (MFLIMIT-1))
+#define HASHLOG64K (HASH_LOG+1)
+#define HASH64KTABLESIZE (1U<<HASHLOG64K)
+#define LZ4_HASH64K_FUNCTION(i)	(((i) * 2654435761U) >> ((MINMATCH*8)-HASHLOG64K))
+#define LZ4_HASH64K_VALUE(p)	LZ4_HASH64K_FUNCTION(A32(p))
+static inline int LZ4_compress64kCtx(void** ctx,
+                 const char* source,
+                 char* dest,
+                 int isize,
+                 int maxOutputSize)
+{
+#if HEAPMODE
+    struct refTables *srt = (struct refTables *) (*ctx);
+    U16* HashTable;
+#else
+    U16 HashTable[HASH64KTABLESIZE] = {0};
+#endif
+
+    const BYTE* ip = (BYTE*) source;
+    const BYTE* anchor = ip;
+    const BYTE* const base = ip;
+    const BYTE* const iend = ip + isize;
+    const BYTE* const mflimit = iend - MFLIMIT;
+#define matchlimit (iend - LASTLITERALS)
+
+    BYTE* op = (BYTE*) dest;
+    BYTE* const oend = op + maxOutputSize;
+
+    int len, length;
+    const int skipStrength = SKIPSTRENGTH;
+    U32 forwardH;
+
+
+    // Init
+    if (isize<MINLENGTH) goto _last_literals;
+#if HEAPMODE
+    if (*ctx == NULL)
+    {
+        srt = (struct refTables *) malloc ( sizeof(struct refTables) );
+        *ctx = (void*) srt;
+    }
+    HashTable = (U16*)(srt->hashTable);
+    memset((void*)HashTable, 0, sizeof(srt->hashTable));
+#else
+    (void) ctx;
+#endif
+
+
+    // First Byte
+    ip++; forwardH = LZ4_HASH64K_VALUE(ip);
+
+    // Main Loop
+    for ( ; ; )
+    {
+        int findMatchAttempts = (1U << skipStrength) + 3;
+        const BYTE* forwardIp = ip;
+        const BYTE* ref;
+        BYTE* token;
+
+        // Find a match
+        do {
+            U32 h = forwardH;
+            int step = findMatchAttempts++ >> skipStrength;
+            ip = forwardIp;
+            forwardIp = ip + step;
+
+            if (forwardIp > mflimit) { goto _last_literals; }
+
+            forwardH = LZ4_HASH64K_VALUE(forwardIp);
+            ref = base + HashTable[h];
+            HashTable[h] = (U16)(ip - base);
+
+        } while (A32(ref) != A32(ip));
+
+        // Catch up
+        while ((ip>anchor) && (ref>(BYTE*)source) && (ip[-1]==ref[-1])) { ip--; ref--; }
+
+        // Encode Literal length
+        length = (int)(ip - anchor);
+        token = op++;
+        if unlikely(op + length + (2 + 1 + LASTLITERALS) + (length>>8) > oend) return 0; 		// Check output limit
+#ifdef _MSC_VER
+        if (length>=(int)RUN_MASK) 
+        { 
+            int len = length-RUN_MASK; 
+            *token=(RUN_MASK<<ML_BITS); 
+            if (len>254)
+            {
+                do { *op++ = 255; len -= 255; } while (len>254);
+                *op++ = (BYTE)len; 
+                memcpy(op, anchor, length);
+                op += length;
+                goto _next_match;
+            }
+            else
+            *op++ = (BYTE)len; 
+        }
+        else *token = (length<<ML_BITS);
+#else
+        if (length>=(int)RUN_MASK) { *token=(RUN_MASK<<ML_BITS); len = length-RUN_MASK; for(; len > 254 ; len-=255) *op++ = 255; *op++ = (BYTE)len; }
+        else *token = (length<<ML_BITS);
+#endif
+
+        // Copy Literals
+        LZ4_BLINDCOPY(anchor, op, length);
+
+_next_match:
+        // Encode Offset
+        LZ4_WRITE_LITTLEENDIAN_16(op,(U16)(ip-ref));
+
+        // Start Counting
+        ip+=MINMATCH; ref+=MINMATCH;   // MinMatch verified
+        anchor = ip;
+        while (ip<matchlimit-(STEPSIZE-1))
+        {
+            UARCH diff = AARCH(ref) ^ AARCH(ip);
+            if (!diff) { ip+=STEPSIZE; ref+=STEPSIZE; continue; }
+            ip += LZ4_NbCommonBytes(diff);
+            goto _endCount;
+        }
+        if (LZ4_ARCH64) if ((ip<(matchlimit-3)) && (A32(ref) == A32(ip))) { ip+=4; ref+=4; }
+        if ((ip<(matchlimit-1)) && (A16(ref) == A16(ip))) { ip+=2; ref+=2; }
+        if ((ip<matchlimit) && (*ref == *ip)) ip++;
+_endCount:
+
+        // Encode MatchLength
+        len = (int)(ip - anchor);
+        if unlikely(op + (1 + LASTLITERALS) + (len>>8) > oend) return 0; 		// Check output limit
+        if (len>=(int)ML_MASK) { *token+=ML_MASK; len-=ML_MASK; for(; len > 509 ; len-=510) { *op++ = 255; *op++ = 255; } if (len > 254) { len-=255; *op++ = 255; } *op++ = (BYTE)len; }
+        else *token += len;
+
+        // Test end of chunk
+        if (ip > mflimit) { anchor = ip;  break; }
+
+        // Fill table
+        HashTable[LZ4_HASH64K_VALUE(ip-2)] = (U16)(ip - 2 - base);
+
+        // Test next position
+        ref = base + HashTable[LZ4_HASH64K_VALUE(ip)];
+        HashTable[LZ4_HASH64K_VALUE(ip)] = (U16)(ip - base);
+        if (A32(ref) == A32(ip)) { token = op++; *token=0; goto _next_match; }
+
+        // Prepare next loop
+        anchor = ip++;
+        forwardH = LZ4_HASH64K_VALUE(ip);
+    }
+
+_last_literals:
+    // Encode Last Literals
+    {
+        int lastRun = (int)(iend - anchor);
+        if (op + lastRun + 1 + (lastRun-RUN_MASK+255)/255 > oend) return 0;
+        if (lastRun>=(int)RUN_MASK) { *op++=(RUN_MASK<<ML_BITS); lastRun-=RUN_MASK; for(; lastRun > 254 ; lastRun-=255) *op++ = 255; *op++ = (BYTE) lastRun; }
+        else *op++ = (lastRun<<ML_BITS);
+        memcpy(op, anchor, iend - anchor);
+        op += iend-anchor;
+    }
+
+    // End
+    return (int) (((char*)op)-dest);
+}
+
+
+int LZ4_compress_limitedOutput(const char* source, 
+                               char* dest, 
+                               int isize, 
+                               int maxOutputSize)
+{
+#if HEAPMODE
+    void* ctx = malloc(sizeof(struct refTables));
+    int result;
+    if (isize < LZ4_64KLIMIT)
+        result = LZ4_compress64kCtx(&ctx, source, dest, isize, maxOutputSize);
+    else result = LZ4_compressCtx(&ctx, source, dest, isize, maxOutputSize);
+    free(ctx);
+    return result;
+#else
+    if (isize < (int)LZ4_64KLIMIT) return LZ4_compress64kCtx(NULL, source, dest, isize, maxOutputSize);
+    return LZ4_compressCtx(NULL, source, dest, isize, maxOutputSize);
+#endif
+}
+
+
+int LZ4_compress(const char* source,
+                 char* dest,
+                 int isize)
+{
+    return LZ4_compress_limitedOutput(source, dest, isize, LZ4_compressBound(isize));
+}
+
+
+
+
+//****************************
+// Decompression functions
+//****************************
+
+// Note : The decoding functions LZ4_uncompress() and LZ4_uncompress_unknownOutputSize()
+//		are safe against "buffer overflow" attack type.
+//		They will never write nor read outside of the provided output buffers.
+//      LZ4_uncompress_unknownOutputSize() also insures that it will never read outside of the input buffer.
+//		A corrupted input will produce an error result, a negative int, indicating the position of the error within input stream.
+
+int LZ4_uncompress(const char* source,
+                 char* dest,
+                 int osize)
+{
+    // Local Variables
+    const BYTE* restrict ip = (const BYTE*) source;
+    const BYTE* ref;
+
+    BYTE* op = (BYTE*) dest;
+    BYTE* const oend = op + osize;
+    BYTE* cpy;
+
+    unsigned token;
+
+    size_t length;
+    size_t dec32table[] = {0, 3, 2, 3, 0, 0, 0, 0};
+#if LZ4_ARCH64
+    size_t dec64table[] = {0, 0, 0, -1, 0, 1, 2, 3};
+#endif
+
+
+    // Main Loop
+    while (1)
+    {
+        // get runlength
+        token = *ip++;
+        if ((length=(token>>ML_BITS)) == RUN_MASK)  { size_t len; for (;(len=*ip++)==255;length+=255){} length += len; }
+
+        // copy literals
+        cpy = op+length;
+        if unlikely(cpy>oend-COPYLENGTH)
+        {
+            if (cpy != oend) goto _output_error;         // Error : not enough place for another match (min 4) + 5 literals
+            memcpy(op, ip, length);
+            ip += length;
+            break;                                       // EOF
+        }
+        LZ4_WILDCOPY(ip, op, cpy); ip -= (op-cpy); op = cpy;
+
+        // get offset
+        LZ4_READ_LITTLEENDIAN_16(ref,cpy,ip); ip+=2;
+        if unlikely(ref < (BYTE* const)dest) goto _output_error;   // Error : offset create reference outside destination buffer
+
+        // get matchlength
+        if ((length=(token&ML_MASK)) == ML_MASK) { for (;*ip==255;length+=255) {ip++;} length += *ip++; }
+
+        // copy repeated sequence
+        if unlikely((op-ref)<STEPSIZE)
+        {
+#if LZ4_ARCH64
+            size_t dec64 = dec64table[op-ref];
+#else
+            const int dec64 = 0;
+#endif
+            op[0] = ref[0];
+            op[1] = ref[1];
+            op[2] = ref[2];
+            op[3] = ref[3];
+            op += 4, ref += 4; ref -= dec32table[op-ref];
+            A32(op) = A32(ref); 
+            op += STEPSIZE-4; ref -= dec64;
+        } else { LZ4_COPYSTEP(ref,op); }
+        cpy = op + length - (STEPSIZE-4);
+        if (cpy>oend-COPYLENGTH)
+        {
+            if (cpy > oend) goto _output_error;             // Error : request to write beyond destination buffer
+            LZ4_SECURECOPY(ref, op, (oend-COPYLENGTH));
+            while(op<cpy) *op++=*ref++;
+            op=cpy;
+            if (op == oend) goto _output_error;    // Check EOF (should never happen, since last 5 bytes are supposed to be literals)
+            continue;
+        }
+        LZ4_SECURECOPY(ref, op, cpy);
+        op=cpy;		// correction
+    }
+
+    // end of decoding
+    return (int) (((char*)ip)-source);
+
+    // write overflow error detected
+_output_error:
+    return (int) (-(((char*)ip)-source));
+}
+
+
+int LZ4_uncompress_unknownOutputSize(
+                const char* source,
+                char* dest,
+                int isize,
+                int maxOutputSize)
+{
+    // Local Variables
+    const BYTE* restrict ip = (const BYTE*) source;
+    const BYTE* const iend = ip + isize;
+    const BYTE* ref;
+
+    BYTE* op = (BYTE*) dest;
+    BYTE* const oend = op + maxOutputSize;
+    BYTE* cpy;
+
+    size_t dec32table[] = {0, 3, 2, 3, 0, 0, 0, 0};
+#if LZ4_ARCH64
+    size_t dec64table[] = {0, 0, 0, -1, 0, 1, 2, 3};
+#endif
+
+
+    // Main Loop
+    while (ip<iend)
+    {
+        unsigned token;
+        size_t length;
+
+        // get runlength
+        token = *ip++;
+        if ((length=(token>>ML_BITS)) == RUN_MASK) { int s=255; while ((ip<iend) && (s==255)) { s=*ip++; length += s; } }
+
+        // copy literals
+        cpy = op+length;
+        if ((cpy>oend-COPYLENGTH) || (ip+length>iend-COPYLENGTH))
+        {
+            if (cpy > oend) goto _output_error;          // Error : writes beyond output buffer
+            if (ip+length != iend) goto _output_error;   // Error : LZ4 format requires to consume all input at this stage
+            memcpy(op, ip, length);
+            op += length;
+            break;                                       // Necessarily EOF, due to parsing restrictions
+        }
+        LZ4_WILDCOPY(ip, op, cpy); ip -= (op-cpy); op = cpy;
+
+        // get offset
+        LZ4_READ_LITTLEENDIAN_16(ref,cpy,ip); ip+=2;
+        if (ref < (BYTE* const)dest) goto _output_error;   // Error : offset creates reference outside of destination buffer
+
+        // get matchlength
+        if ((length=(token&ML_MASK)) == ML_MASK) { while (ip<iend) { int s = *ip++; length +=s; if (s==255) continue; break; } }
+
+        // copy repeated sequence
+        if unlikely(op-ref<STEPSIZE)
+        {
+#if LZ4_ARCH64
+            size_t dec64 = dec64table[op-ref];
+#else
+            const int dec64 = 0;
+#endif
+            op[0] = ref[0];
+            op[1] = ref[1];
+            op[2] = ref[2];
+            op[3] = ref[3];
+            op += 4, ref += 4; ref -= dec32table[op-ref];
+            A32(op) = A32(ref); 
+            op += STEPSIZE-4; ref -= dec64;
+        } else { LZ4_COPYSTEP(ref,op); }
+        cpy = op + length - (STEPSIZE-4);
+        if (cpy>oend-COPYLENGTH)
+        {
+            if (cpy > oend) goto _output_error;    // Error : request to write outside of destination buffer
+            LZ4_SECURECOPY(ref, op, (oend-COPYLENGTH));
+            while(op<cpy) *op++=*ref++;
+            op=cpy;
+            if (op == oend) goto _output_error;    // Check EOF (should never happen, since last 5 bytes are supposed to be literals)
+            continue;
+        }
+        LZ4_SECURECOPY(ref, op, cpy);
+        op=cpy;		// correction
+    }
+
+    // end of decoding
+    return (int) (((char*)op)-dest);
+
+    // write overflow error detected
+_output_error:
+    return (int) (-(((char*)ip)-source));
+}
+
diff --git a/src/cbits/lz4.h b/src/cbits/lz4.h
--- a/src/cbits/lz4.h
+++ b/src/cbits/lz4.h
@@ -1,117 +1,128 @@
-/*
-   LZ4 - Fast LZ compression algorithm
-   Header File
-   Copyright (C) 2011-2012, Yann Collet.
-   BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
-
-   Redistribution and use in source and binary forms, with or without
-   modification, are permitted provided that the following conditions are
-   met:
-
-       * Redistributions of source code must retain the above copyright
-   notice, this list of conditions and the following disclaimer.
-       * Redistributions in binary form must reproduce the above
-   copyright notice, this list of conditions and the following disclaimer
-   in the documentation and/or other materials provided with the
-   distribution.
-
-   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
-   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
-   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
-   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
-   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
-   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
-   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
-   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
-   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
-   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
-   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-
-   You can contact the author at :
-   - LZ4 homepage : http://fastcompression.blogspot.com/p/lz4.html
-   - LZ4 source repository : http://code.google.com/p/lz4/
-*/
-#pragma once
-
-#if defined (__cplusplus)
-extern "C" {
-#endif
-
-
-//****************************
-// Simple Functions
-//****************************
-
-int LZ4_compress   (const char* source, char* dest, int isize);
-int LZ4_uncompress (const char* source, char* dest, int osize);
-
-/*
-LZ4_compress() :
-	Compresses 'isize' bytes from 'source' into 'dest'.
-	Destination buffer must be already allocated,
-	and must be sized to handle worst cases situations (input data not compressible)
-	Worst case size evaluation is provided by macro LZ4_compressBound()
-
-	isize  : is the input size. Max supported value is ~1.9GB
-	return : the number of bytes written in buffer dest
-
-
-LZ4_uncompress() :
-	osize  : is the output size, therefore the original size
-	return : the number of bytes read in the source buffer
-			 If the source stream is malformed, the function will stop decoding and return a negative result, indicating the byte position of the faulty instruction
-			 This function never writes beyond dest + osize, and is therefore protected against malicious data packets
-	note : destination buffer must be already allocated.
-		   its size must be a minimum of 'osize' bytes.
-*/
-
-
-//****************************
-// Advanced Functions
-//****************************
-
-#define LZ4_compressBound(isize)   (isize + (isize/255) + 16)
-
-/*
-LZ4_compressBound() :
-	Provides the maximum size that LZ4 may output in a "worst case" scenario (input data not compressible)
-	primarily useful for memory allocation of output buffer.
-
-	isize  : is the input size. Max supported value is ~1.9GB
-	return : maximum output size in a "worst case" scenario
-	note : this function is limited by "int" range (2^31-1)
-*/
-
-
-int LZ4_compress_limitedOutput   (const char* source, char* dest, int isize, int maxOutputSize);
-
-/*
-LZ4_compress_limitedOutput() :
-    Compress 'isize' bytes from 'source' into an output buffer 'dest' of maximum size 'maxOutputSize'.
-	If it cannot achieve it, compression will stop, and result of the function will be zero.
-	This function never writes outside of provided output buffer.
-
-	isize  : is the input size. Max supported value is ~1.9GB
-	maxOutputSize : is the size of the destination buffer (which must be already allocated)
-	return : the number of bytes written in buffer 'dest'
-			 or 0 if the compression fails
-*/
-
-
-int LZ4_uncompress_unknownOutputSize (const char* source, char* dest, int isize, int maxOutputSize);
-
-/*
-LZ4_uncompress_unknownOutputSize() :
-	isize  : is the input size, therefore the compressed size
-	maxOutputSize : is the size of the destination buffer (which must be already allocated)
-	return : the number of bytes decoded in the destination buffer (necessarily <= maxOutputSize)
-			 If the source stream is malformed, the function will stop decoding and return a negative result, indicating the byte position of the faulty instruction
-			 This function never writes beyond dest + maxOutputSize, and is therefore protected against malicious data packets
-	note   : Destination buffer must be already allocated.
-	         This version is slightly slower than LZ4_uncompress()
-*/
-
-
-#if defined (__cplusplus)
-}
-#endif
+/*
+   LZ4 - Fast LZ compression algorithm
+   Header File
+   Copyright (C) 2011-2012, Yann Collet.
+   BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
+
+   Redistribution and use in source and binary forms, with or without
+   modification, are permitted provided that the following conditions are
+   met:
+
+       * Redistributions of source code must retain the above copyright
+   notice, this list of conditions and the following disclaimer.
+       * Redistributions in binary form must reproduce the above
+   copyright notice, this list of conditions and the following disclaimer
+   in the documentation and/or other materials provided with the
+   distribution.
+
+   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+   You can contact the author at :
+   - LZ4 homepage : http://fastcompression.blogspot.com/p/lz4.html
+   - LZ4 source repository : http://code.google.com/p/lz4/
+*/
+#pragma once
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+
+//**************************************
+// Compiler Options
+//**************************************
+#if defined(_MSC_VER) && !defined(__cplusplus)   // Visual Studio
+#  define inline __inline           // Visual is not C99, but supports some kind of inline
+#endif
+
+
+//****************************
+// Simple Functions
+//****************************
+
+int LZ4_compress   (const char* source, char* dest, int isize);
+int LZ4_uncompress (const char* source, char* dest, int osize);
+
+/*
+LZ4_compress() :
+    Compresses 'isize' bytes from 'source' into 'dest'.
+    Destination buffer must be already allocated,
+    and must be sized to handle worst cases situations (input data not compressible)
+    Worst case size evaluation is provided by function LZ4_compressBound()
+
+    isize  : is the input size. Max supported value is ~1.9GB
+    return : the number of bytes written in buffer dest
+
+
+LZ4_uncompress() :
+    osize  : is the output size, therefore the original size
+    return : the number of bytes read in the source buffer
+             If the source stream is malformed, the function will stop decoding and return a negative result, indicating the byte position of the faulty instruction
+             This function never writes outside of provided buffers, and never modifies input buffer.
+    note : destination buffer must be already allocated.
+           its size must be a minimum of 'osize' bytes.
+*/
+
+
+//****************************
+// Advanced Functions
+//****************************
+
+static inline int LZ4_compressBound(int isize)   { return ((isize) + ((isize)/255) + 16); }
+#define           LZ4_COMPRESSBOUND(    isize)            ((isize) + ((isize)/255) + 16)
+
+/*
+LZ4_compressBound() :
+    Provides the maximum size that LZ4 may output in a "worst case" scenario (input data not compressible)
+    primarily useful for memory allocation of output buffer.
+	inline function is recommended for the general case,
+	but macro is also provided when results need to be evaluated at compile time (such as table size allocation).
+
+    isize  : is the input size. Max supported value is ~1.9GB
+    return : maximum output size in a "worst case" scenario
+    note : this function is limited by "int" range (2^31-1)
+*/
+
+
+int LZ4_compress_limitedOutput   (const char* source, char* dest, int isize, int maxOutputSize);
+
+/*
+LZ4_compress_limitedOutput() :
+    Compress 'isize' bytes from 'source' into an output buffer 'dest' of maximum size 'maxOutputSize'.
+    If it cannot achieve it, compression will stop, and result of the function will be zero.
+    This function never writes outside of provided output buffer.
+
+    isize  : is the input size. Max supported value is ~1.9GB
+    maxOutputSize : is the size of the destination buffer (which must be already allocated)
+    return : the number of bytes written in buffer 'dest'
+             or 0 if the compression fails
+*/
+
+
+int LZ4_uncompress_unknownOutputSize (const char* source, char* dest, int isize, int maxOutputSize);
+
+/*
+LZ4_uncompress_unknownOutputSize() :
+    isize  : is the input size, therefore the compressed size
+    maxOutputSize : is the size of the destination buffer (which must be already allocated)
+    return : the number of bytes decoded in the destination buffer (necessarily <= maxOutputSize)
+             If the source stream is malformed, the function will stop decoding and return a negative result, indicating the byte position of the faulty instruction
+             This function never writes beyond dest + maxOutputSize, and is therefore protected against malicious data packets
+    note   : Destination buffer must be already allocated.
+             This version is slightly slower than LZ4_uncompress()
+*/
+
+
+#if defined (__cplusplus)
+}
+#endif
diff --git a/src/cbits/lz4hc.c b/src/cbits/lz4hc.c
--- a/src/cbits/lz4hc.c
+++ b/src/cbits/lz4hc.c
@@ -1,671 +1,719 @@
-/*
-   LZ4 HC - High Compression Mode of LZ4
-   Copyright (C) 2011-2012, Yann Collet.
-   BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
-
-   Redistribution and use in source and binary forms, with or without
-   modification, are permitted provided that the following conditions are
-   met:
-
-       * Redistributions of source code must retain the above copyright
-   notice, this list of conditions and the following disclaimer.
-       * Redistributions in binary form must reproduce the above
-   copyright notice, this list of conditions and the following disclaimer
-   in the documentation and/or other materials provided with the
-   distribution.
-
-   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
-   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
-   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
-   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
-   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
-   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
-   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
-   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
-   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
-   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
-   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-
-   You can contact the author at :
-   - LZ4 homepage : http://fastcompression.blogspot.com/p/lz4.html
-   - LZ4 source repository : http://code.google.com/p/lz4/
-*/
-
-
-//**************************************
-// CPU Feature Detection
-//**************************************
-// 32 or 64 bits ?
-#if (defined(__x86_64__) || defined(__x86_64) || defined(__amd64__) || defined(__amd64) || defined(__ppc64__) || defined(_WIN64) || defined(__LP64__) || defined(_LP64) )   // Detects 64 bits mode
-#define LZ4_ARCH64 1
-#else
-#define LZ4_ARCH64 0
-#endif
-
-// Little Endian or Big Endian ? 
-#if (defined(__BIG_ENDIAN__) || defined(__BIG_ENDIAN) || defined(_BIG_ENDIAN) || defined(_ARCH_PPC) || defined(__PPC__) || defined(__PPC) || defined(PPC) || defined(__powerpc__) || defined(__powerpc) || defined(powerpc) || ((defined(__BYTE_ORDER__)&&(__BYTE_ORDER__ == __ORDER_BIG_ENDIAN__))) )
-#define LZ4_BIG_ENDIAN 1
-#else
-// Little Endian assumed. PDP Endian and other very rare endian format are unsupported.
-#endif
-
-// Unaligned memory access is automatically enabled for "common" CPU, such as x86.
-// For others CPU, the compiler will be more cautious, and insert extra code to ensure aligned access is respected
-// If you know your target CPU supports unaligned memory access, you may want to force this option manually to improve performance
-#if defined(__ARM_FEATURE_UNALIGNED)
-#define LZ4_FORCE_UNALIGNED_ACCESS 1
-#endif
-
-
-//**************************************
-// Compiler Options
-//**************************************
-#if __STDC_VERSION__ >= 199901L    // C99
-  /* "restrict" is a known keyword */
-#else
-#define restrict  // Disable restrict
-#endif
-
-#ifdef _MSC_VER
-#define inline __forceinline    // Visual is not C99, but supports some kind of inline
-#include <intrin.h>             // For Visual 2005
-#  if LZ4_ARCH64	// 64-bit
-#    pragma intrinsic(_BitScanForward64) // For Visual 2005
-#    pragma intrinsic(_BitScanReverse64) // For Visual 2005
-#  else
-#    pragma intrinsic(_BitScanForward)   // For Visual 2005
-#    pragma intrinsic(_BitScanReverse)   // For Visual 2005
-#  endif
-#endif
-
-#ifdef _MSC_VER  // Visual Studio
-#define bswap16(x) _byteswap_ushort(x)
-#else
-#define bswap16(x)  ((unsigned short int) ((((x) >> 8) & 0xffu) | (((x) & 0xffu) << 8)))
-#endif
-
-
-//**************************************
-// Includes
-//**************************************
-#include <stdlib.h>   // calloc, free
-#include <string.h>   // memset, memcpy
-#include "lz4hc.h"
-
-#define ALLOCATOR(s) calloc(1,s)
-#define FREEMEM free
-#define MEM_INIT memset
-
-
-//**************************************
-// Basic Types
-//**************************************
-#if defined(_MSC_VER)    // Visual Studio does not support 'stdint' natively
-#define BYTE	unsigned __int8
-#define U16		unsigned __int16
-#define U32		unsigned __int32
-#define S32		__int32
-#define U64		unsigned __int64
-#else
-#include <stdint.h>
-#define BYTE	uint8_t
-#define U16		uint16_t
-#define U32		uint32_t
-#define S32		int32_t
-#define U64		uint64_t
-#endif
-
-#ifndef LZ4_FORCE_UNALIGNED_ACCESS
-#pragma pack(push, 1) 
-#endif
-
-typedef struct _U16_S { U16 v; } U16_S;
-typedef struct _U32_S { U32 v; } U32_S;
-typedef struct _U64_S { U64 v; } U64_S;
-
-#ifndef LZ4_FORCE_UNALIGNED_ACCESS
-#pragma pack(pop) 
-#endif
-
-#define A64(x) (((U64_S *)(x))->v)
-#define A32(x) (((U32_S *)(x))->v)
-#define A16(x) (((U16_S *)(x))->v)
-
-
-//**************************************
-// Constants
-//**************************************
-#define MINMATCH 4
-
-#define DICTIONARY_LOGSIZE 16
-#define MAXD (1<<DICTIONARY_LOGSIZE)
-#define MAXD_MASK ((U32)(MAXD - 1))
-#define MAX_DISTANCE (MAXD - 1)
-
-#define HASH_LOG (DICTIONARY_LOGSIZE-1)
-#define HASHTABLESIZE (1 << HASH_LOG)
-#define HASH_MASK (HASHTABLESIZE - 1)
-
-#define MAX_NB_ATTEMPTS 256
-
-#define ML_BITS  4
-#define ML_MASK  (size_t)((1U<<ML_BITS)-1)
-#define RUN_BITS (8-ML_BITS)
-#define RUN_MASK ((1U<<RUN_BITS)-1)
-
-#define COPYLENGTH 8
-#define LASTLITERALS 5
-#define MFLIMIT (COPYLENGTH+MINMATCH)
-#define MINLENGTH (MFLIMIT+1)
-#define OPTIMAL_ML (int)((ML_MASK-1)+MINMATCH)
-
-
-//**************************************
-// Architecture-specific macros
-//**************************************
-#if LZ4_ARCH64	// 64-bit
-#define STEPSIZE 8
-#define LZ4_COPYSTEP(s,d)		A64(d) = A64(s); d+=8; s+=8;
-#define LZ4_COPYPACKET(s,d)		LZ4_COPYSTEP(s,d)
-#define UARCH U64
-#define AARCH A64
-#define HTYPE					U32
-#define INITBASE(b,s)			const BYTE* const b = s
-#else		// 32-bit
-#define STEPSIZE 4
-#define LZ4_COPYSTEP(s,d)		A32(d) = A32(s); d+=4; s+=4;
-#define LZ4_COPYPACKET(s,d)		LZ4_COPYSTEP(s,d); LZ4_COPYSTEP(s,d);
-#define UARCH U32
-#define AARCH A32
-#define HTYPE					const BYTE*
-#define INITBASE(b,s)		    const int b = 0
-#endif
-
-#if defined(LZ4_BIG_ENDIAN)
-#define LZ4_READ_LITTLEENDIAN_16(d,s,p) { U16 v = A16(p); v = bswap16(v); d = (s) - v; }
-#define LZ4_WRITE_LITTLEENDIAN_16(p,i)  { U16 v = (U16)(i); v = bswap16(v); A16(p) = v; p+=2; }
-#else		// Little Endian
-#define LZ4_READ_LITTLEENDIAN_16(d,s,p) { d = (s) - A16(p); }
-#define LZ4_WRITE_LITTLEENDIAN_16(p,v)  { A16(p) = v; p+=2; }
-#endif
-
-
-//************************************************************
-// Local Types
-//************************************************************
-typedef struct 
-{
-	const BYTE* base;
-	HTYPE hashTable[HASHTABLESIZE];
-	U16 chainTable[MAXD];
-	const BYTE* nextToUpdate;
-} LZ4HC_Data_Structure;
-
-
-//**************************************
-// Macros
-//**************************************
-#define LZ4_WILDCOPY(s,d,e)		do { LZ4_COPYPACKET(s,d) } while (d<e);
-#define LZ4_BLINDCOPY(s,d,l)	{ BYTE* e=d+l; LZ4_WILDCOPY(s,d,e); d=e; }
-#define HASH_FUNCTION(i)	(((i) * 2654435761U) >> ((MINMATCH*8)-HASH_LOG))
-#define HASH_VALUE(p)		HASH_FUNCTION(*(U32*)(p))
-#define HASH_POINTER(p)		(HashTable[HASH_VALUE(p)] + base)
-#define DELTANEXT(p)		chainTable[(size_t)(p) & MAXD_MASK] 
-#define GETNEXT(p)			((p) - (size_t)DELTANEXT(p))
-#define ADD_HASH(p)			{ size_t delta = (p) - HASH_POINTER(p); if (delta>MAX_DISTANCE) delta = MAX_DISTANCE; DELTANEXT(p) = (U16)delta; HashTable[HASH_VALUE(p)] = (p) - base; }
-
-
-//**************************************
-// Private functions
-//**************************************
-#if LZ4_ARCH64
-
-inline static int LZ4_NbCommonBytes (register U64 val)
-{
-#if defined(LZ4_BIG_ENDIAN)
-    #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
-    unsigned long r = 0;
-    _BitScanReverse64( &r, val );
-    return (int)(r>>3);
-    #elif defined(__GNUC__) && ((__GNUC__ * 100 + __GNUC_MINOR__) >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
-    return (__builtin_clzll(val) >> 3); 
-    #else
-	int r;
-	if (!(val>>32)) { r=4; } else { r=0; val>>=32; }
-	if (!(val>>16)) { r+=2; val>>=8; } else { val>>=24; }
-	r += (!val);
-	return r;
-    #endif
-#else
-    #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
-    unsigned long r = 0;
-    _BitScanForward64( &r, val );
-    return (int)(r>>3);
-    #elif defined(__GNUC__) && ((__GNUC__ * 100 + __GNUC_MINOR__) >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
-    return (__builtin_ctzll(val) >> 3); 
-    #else
-	static const int DeBruijnBytePos[64] = { 0, 0, 0, 0, 0, 1, 1, 2, 0, 3, 1, 3, 1, 4, 2, 7, 0, 2, 3, 6, 1, 5, 3, 5, 1, 3, 4, 4, 2, 5, 6, 7, 7, 0, 1, 2, 3, 3, 4, 6, 2, 6, 5, 5, 3, 4, 5, 6, 7, 1, 2, 4, 6, 4, 4, 5, 7, 2, 6, 5, 7, 6, 7, 7 };
-	return DeBruijnBytePos[((U64)((val & -val) * 0x0218A392CDABBD3F)) >> 58];
-    #endif
-#endif
-}
-
-#else
-
-inline static int LZ4_NbCommonBytes (register U32 val)
-{
-#if defined(LZ4_BIG_ENDIAN)
-    #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
-    unsigned long r = 0;
-    _BitScanReverse( &r, val );
-    return (int)(r>>3);
-    #elif defined(__GNUC__) && ((__GNUC__ * 100 + __GNUC_MINOR__) >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
-    return (__builtin_clz(val) >> 3); 
-    #else
-	int r;
-	if (!(val>>16)) { r=2; val>>=8; } else { r=0; val>>=24; }
-	r += (!val);
-	return r;
-    #endif
-#else
-    #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
-    unsigned long r = 0;
-    _BitScanForward( &r, val );
-    return (int)(r>>3);
-    #elif defined(__GNUC__) && ((__GNUC__ * 100 + __GNUC_MINOR__) >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
-    return (__builtin_ctz(val) >> 3); 
-    #else
-	static const int DeBruijnBytePos[32] = { 0, 0, 3, 0, 3, 1, 3, 0, 3, 2, 2, 1, 3, 2, 0, 1, 3, 3, 1, 2, 2, 2, 2, 0, 3, 1, 2, 0, 1, 0, 1, 1 };
-	return DeBruijnBytePos[((U32)((val & -(S32)val) * 0x077CB531U)) >> 27];
-    #endif
-#endif
-}
-
-#endif
-
-
-inline static int LZ4HC_Init (LZ4HC_Data_Structure* hc4, const BYTE* base)
-{
-	MEM_INIT((void*)hc4->hashTable, 0, sizeof(hc4->hashTable));
-	MEM_INIT(hc4->chainTable, 0xFF, sizeof(hc4->chainTable));
-	hc4->nextToUpdate = base + LZ4_ARCH64;
-	hc4->base = base;
-	return 1;
-}
-
-
-inline static void* LZ4HC_Create (const BYTE* base)
-{
-	void* hc4 = ALLOCATOR(sizeof(LZ4HC_Data_Structure));
-
-	LZ4HC_Init (hc4, base);
-	return hc4;
-}
-
-
-inline static int LZ4HC_Free (void** LZ4HC_Data)
-{
-	FREEMEM(*LZ4HC_Data);
-	*LZ4HC_Data = NULL;
-	return (1);
-}
-
-
-inline static void LZ4HC_Insert (LZ4HC_Data_Structure* hc4, const BYTE* ip)
-{
-	U16*   chainTable = hc4->chainTable;
-	HTYPE* HashTable  = hc4->hashTable;
-	INITBASE(base,hc4->base);
-
-	while(hc4->nextToUpdate < ip)
-	{
-		ADD_HASH(hc4->nextToUpdate);
-		hc4->nextToUpdate++;
-	}
-}
-
-
-inline static int LZ4HC_InsertAndFindBestMatch (LZ4HC_Data_Structure* hc4, const BYTE* ip, const BYTE* const matchlimit, const BYTE** matchpos)
-{
-	U16* const chainTable = hc4->chainTable;
-	HTYPE* const HashTable = hc4->hashTable;
-	const BYTE* ref;
-	INITBASE(base,hc4->base);
-	int nbAttempts=MAX_NB_ATTEMPTS;
-	int ml=0;
-
-	// HC4 match finder
-	LZ4HC_Insert(hc4, ip);
-	ref = HASH_POINTER(ip);
-	while ((ref > (ip-MAX_DISTANCE)) && (nbAttempts))
-	{
-		nbAttempts--;
-		if (*(ref+ml) == *(ip+ml))
-		if (*(U32*)ref == *(U32*)ip)
-		{
-			const BYTE* reft = ref+MINMATCH;
-			const BYTE* ipt = ip+MINMATCH;
-
-			while (ipt<matchlimit-(STEPSIZE-1))
-			{
-				UARCH diff = AARCH(reft) ^ AARCH(ipt);
-				if (!diff) { ipt+=STEPSIZE; reft+=STEPSIZE; continue; }
-				ipt += LZ4_NbCommonBytes(diff);
-				goto _endCount;
-			}
-			if (LZ4_ARCH64) if ((ipt<(matchlimit-3)) && (A32(reft) == A32(ipt))) { ipt+=4; reft+=4; }
-			if ((ipt<(matchlimit-1)) && (A16(reft) == A16(ipt))) { ipt+=2; reft+=2; }
-			if ((ipt<matchlimit) && (*reft == *ipt)) ipt++;
-_endCount:
-
-			if (ipt-ip > ml) { ml = (int)(ipt-ip); *matchpos = ref; }
-		}
-		ref = GETNEXT(ref);
-	}
-
-	return ml;
-}
-
-
-inline static int LZ4HC_InsertAndGetWiderMatch (LZ4HC_Data_Structure* hc4, const BYTE* ip, const BYTE* startLimit, const BYTE* matchlimit, int longest, const BYTE** matchpos, const BYTE** startpos)
-{
-	U16* const  chainTable = hc4->chainTable;
-	HTYPE* const HashTable = hc4->hashTable;
-	INITBASE(base,hc4->base);
-	const BYTE*  ref;
-	int nbAttempts = MAX_NB_ATTEMPTS;
-	int delta = (int)(ip-startLimit);
-
-	// First Match
-	LZ4HC_Insert(hc4, ip);
-	ref = HASH_POINTER(ip);
-
-	while ((ref > ip-MAX_DISTANCE) && (ref >= hc4->base) && (nbAttempts))
-	{
-		nbAttempts--;
-		if (*(startLimit + longest) == *(ref - delta + longest))
-		if (*(U32*)ref == *(U32*)ip)
-		{
-			const BYTE* reft = ref+MINMATCH;
-			const BYTE* ipt = ip+MINMATCH;
-			const BYTE* startt = ip;
-
-			while (ipt<matchlimit-(STEPSIZE-1))
-			{
-				UARCH diff = AARCH(reft) ^ AARCH(ipt);
-				if (!diff) { ipt+=STEPSIZE; reft+=STEPSIZE; continue; }
-				ipt += LZ4_NbCommonBytes(diff);
-				goto _endCount;
-			}
-			if (LZ4_ARCH64) if ((ipt<(matchlimit-3)) && (A32(reft) == A32(ipt))) { ipt+=4; reft+=4; }
-			if ((ipt<(matchlimit-1)) && (A16(reft) == A16(ipt))) { ipt+=2; reft+=2; }
-			if ((ipt<matchlimit) && (*reft == *ipt)) ipt++;
-_endCount:
-
-			reft = ref;
-			while ((startt>startLimit) && (reft > hc4->base) && (startt[-1] == reft[-1])) {startt--; reft--;}
-
-			if ((ipt-startt) > longest)
-			{
-				longest = (int)(ipt-startt);
-				*matchpos = reft;
-				*startpos = startt;
-			}
-		}
-		ref = GETNEXT(ref);
-	}
-
-	return longest;
-}
-
-
-inline static int LZ4_encodeSequence(const BYTE** ip, BYTE** op, const BYTE** anchor, int ml, const BYTE* ref)
-{
-	int length, len; 
-	BYTE* token;
-
-	// Encode Literal length
-	length = (int)(*ip - *anchor);
-	token = (*op)++;
-	if (length>=(int)RUN_MASK) { *token=(RUN_MASK<<ML_BITS); len = length-RUN_MASK; for(; len > 254 ; len-=255) *(*op)++ = 255;  *(*op)++ = (BYTE)len; } 
-	else *token = (length<<ML_BITS);
-
-	// Copy Literals
-	LZ4_BLINDCOPY(*anchor, *op, length);
-
-	// Encode Offset
-	LZ4_WRITE_LITTLEENDIAN_16(*op,(U16)(*ip-ref));
-
-	// Encode MatchLength
-	len = (int)(ml-MINMATCH);
-	if (len>=(int)ML_MASK) { *token+=ML_MASK; len-=ML_MASK; for(; len > 509 ; len-=510) { *(*op)++ = 255; *(*op)++ = 255; } if (len > 254) { len-=255; *(*op)++ = 255; } *(*op)++ = (BYTE)len; } 
-	else *token += len;	
-
-	// Prepare next loop
-	*ip += ml;
-	*anchor = *ip; 
-
-	return 0;
-}
-
-
-//****************************
-// Compression CODE
-//****************************
-
-int LZ4_compressHCCtx(LZ4HC_Data_Structure* ctx,
-				 const char* source, 
-				 char* dest,
-				 int isize)
-{	
-	const BYTE* ip = (const BYTE*) source;
-	const BYTE* anchor = ip;
-	const BYTE* const iend = ip + isize;
-	const BYTE* const mflimit = iend - MFLIMIT;
-	const BYTE* const matchlimit = (iend - LASTLITERALS);
-
-	BYTE* op = (BYTE*) dest;
-
-	int	ml, ml2, ml3, ml0;
-	const BYTE* ref=NULL;
-	const BYTE* start2=NULL;
-	const BYTE* ref2=NULL;
-	const BYTE* start3=NULL;
-	const BYTE* ref3=NULL;
-	const BYTE* start0;
-	const BYTE* ref0;
-
-	ip++;
-
-	// Main Loop
-	while (ip < mflimit)
-	{
-		ml = LZ4HC_InsertAndFindBestMatch (ctx, ip, matchlimit, (&ref));
-		if (!ml) { ip++; continue; }
-
-		// saved, in case we would skip too much
-		start0 = ip;
-		ref0 = ref;
-		ml0 = ml;
-
-_Search2:
-		if (ip+ml < mflimit)
-			ml2 = LZ4HC_InsertAndGetWiderMatch(ctx, ip + ml - 2, ip + 1, matchlimit, ml, &ref2, &start2);
-		else ml2=ml;
-
-		if (ml2 == ml)  // No better match
-		{
-			LZ4_encodeSequence(&ip, &op, &anchor, ml, ref);
-			continue;
-		}
-
-		if (start0 < ip)
-		{
-			if (start2 < ip + ml0)   // empirical
-			{
-				ip = start0;
-				ref = ref0;
-				ml = ml0;
-			}
-		}
-
-		// Here, start0==ip
-		if ((start2 - ip) < 3)   // First Match too small : removed
-		{
-			ml = ml2;
-			ip = start2;
-			ref =ref2;
-			goto _Search2;
-		}
-
-_Search3:
-		// Currently we have :
-		// ml2 > ml1, and
-		// ip1+3 <= ip2 (usually < ip1+ml1)
-		if ((start2 - ip) < OPTIMAL_ML)
-		{
-			int correction;
-			int new_ml = ml;
-			if (new_ml > OPTIMAL_ML) new_ml = OPTIMAL_ML;
-			if (ip+new_ml > start2 + ml2 - MINMATCH) new_ml = (int)(start2 - ip) + ml2 - MINMATCH;
-			correction = new_ml - (int)(start2 - ip);
-			if (correction > 0)
-			{
-				start2 += correction;
-				ref2 += correction;
-				ml2 -= correction;
-			}
-		}
-		// Now, we have start2 = ip+new_ml, with new_ml=min(ml, OPTIMAL_ML=18)
-
-		if (start2 + ml2 < mflimit)
-			ml3 = LZ4HC_InsertAndGetWiderMatch(ctx, start2 + ml2 - 3, start2, matchlimit, ml2, &ref3, &start3);
-		else ml3=ml2;
-
-		if (ml3 == ml2) // No better match : 2 sequences to encode
-		{
-			// ip & ref are known; Now for ml
-			if (start2 < ip+ml)
-			{
-				if ((start2 - ip) < OPTIMAL_ML)
-				{
-					int correction;
-					if (ml > OPTIMAL_ML) ml = OPTIMAL_ML;
-					if (ip+ml > start2 + ml2 - MINMATCH) ml = (int)(start2 - ip) + ml2 - MINMATCH;
-					correction = ml - (int)(start2 - ip);
-					if (correction > 0)
-					{
-						start2 += correction;
-						ref2 += correction;
-						ml2 -= correction;
-					}
-				}
-				else
-				{
-					ml = (int)(start2 - ip);
-				}
-			}
-			// Now, encode 2 sequences
-			LZ4_encodeSequence(&ip, &op, &anchor, ml, ref);
-			ip = start2;
-			LZ4_encodeSequence(&ip, &op, &anchor, ml2, ref2);
-			continue;
-		}
-
-		if (start3 < ip+ml+3) // Not enough space for match 2 : remove it
-		{
-			if (start3 >= (ip+ml)) // can write Seq1 immediately ==> Seq2 is removed, so Seq3 becomes Seq1
-			{
-				if (start2 < ip+ml)
-				{
-					int correction = (int)(ip+ml - start2);
-					start2 += correction;
-					ref2 += correction;
-					ml2 -= correction;
-					if (ml2 < MINMATCH)
-					{
-						start2 = start3;
-						ref2 = ref3;
-						ml2 = ml3;
-					}
-				}
-
-				LZ4_encodeSequence(&ip, &op, &anchor, ml, ref);
-				ip  = start3;
-				ref = ref3;
-				ml  = ml3;
-
-				start0 = start2;
-				ref0 = ref2;
-				ml0 = ml2;
-				goto _Search2;
-			}
-
-			start2 = start3;
-			ref2 = ref3;
-			ml2 = ml3;
-			goto _Search3;
-		}
-
-		// OK, now we have 3 ascending matches; let's write at least the first one
-		// ip & ref are known; Now for ml
-		if (start2 < ip+ml)
-		{
-			if ((start2 - ip) < (int)ML_MASK)
-			{
-				int correction;
-				if (ml > OPTIMAL_ML) ml = OPTIMAL_ML;
-				if (ip + ml > start2 + ml2 - MINMATCH) ml = (int)(start2 - ip) + ml2 - MINMATCH;
-				correction = ml - (int)(start2 - ip);
-				if (correction > 0)
-				{
-					start2 += correction;
-					ref2 += correction;
-					ml2 -= correction;
-				}
-			}
-			else
-			{
-				ml = (int)(start2 - ip);
-			}
-		}
-		LZ4_encodeSequence(&ip, &op, &anchor, ml, ref);
-
-		ip = start2;
-		ref = ref2;
-		ml = ml2;
-
-		start2 = start3;
-		ref2 = ref3;
-		ml2 = ml3;
-
-		goto _Search3;
-
-	}
-
-	// Encode Last Literals
-	{
-		int lastRun = (int)(iend - anchor);
-		if (lastRun>=(int)RUN_MASK) { *op++=(RUN_MASK<<ML_BITS); lastRun-=RUN_MASK; for(; lastRun > 254 ; lastRun-=255) *op++ = 255; *op++ = (BYTE) lastRun; } 
-		else *op++ = (lastRun<<ML_BITS);
-		memcpy(op, anchor, iend - anchor);
-		op += iend-anchor;
-	} 
-
-	// End
-	return (int) (((char*)op)-dest);
-}
-
-
-int LZ4_compressHC(const char* source, 
-				 char* dest,
-				 int isize)
-{
-	void* ctx = LZ4HC_Create((const BYTE*)source);
-	int result = LZ4_compressHCCtx(ctx, source, dest, isize);
-	LZ4HC_Free (&ctx);
-
-	return result;
-}
-
-
+/*
+   LZ4 HC - High Compression Mode of LZ4
+   Copyright (C) 2011-2012, Yann Collet.
+   BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
+
+   Redistribution and use in source and binary forms, with or without
+   modification, are permitted provided that the following conditions are
+   met:
+
+       * Redistributions of source code must retain the above copyright
+   notice, this list of conditions and the following disclaimer.
+       * Redistributions in binary form must reproduce the above
+   copyright notice, this list of conditions and the following disclaimer
+   in the documentation and/or other materials provided with the
+   distribution.
+
+   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+   You can contact the author at :
+   - LZ4 homepage : http://fastcompression.blogspot.com/p/lz4.html
+   - LZ4 source repository : http://code.google.com/p/lz4/
+*/
+
+
+//**************************************
+// CPU Feature Detection
+//**************************************
+// 32 or 64 bits ?
+#if (defined(__x86_64__) || defined(__x86_64) || defined(__amd64__) || defined(__amd64) || defined(__ppc64__) || defined(_WIN64) || defined(__LP64__) || defined(_LP64) )   // Detects 64 bits mode
+#  define LZ4_ARCH64 1
+#else
+#  define LZ4_ARCH64 0
+#endif
+
+// Little Endian or Big Endian ?
+// Overwrite the #define below if you know your architecture endianess
+#if defined (__GLIBC__)
+#  include <endian.h>
+#  if (__BYTE_ORDER == __BIG_ENDIAN)
+#     define LZ4_BIG_ENDIAN 1
+#  endif
+#elif (defined(__BIG_ENDIAN__) || defined(__BIG_ENDIAN) || defined(_BIG_ENDIAN)) && !(defined(__LITTLE_ENDIAN__) || defined(__LITTLE_ENDIAN) || defined(_LITTLE_ENDIAN))
+#  define LZ4_BIG_ENDIAN 1
+#elif defined(__sparc) || defined(__sparc__) \
+   || defined(__ppc__) || defined(_POWER) || defined(__powerpc__) || defined(_ARCH_PPC) || defined(__PPC__) || defined(__PPC) || defined(PPC) || defined(__powerpc__) || defined(__powerpc) || defined(powerpc) \
+   || defined(__hpux)  || defined(__hppa) \
+   || defined(_MIPSEB) || defined(__s390__)
+#  define LZ4_BIG_ENDIAN 1
+#else
+// Little Endian assumed. PDP Endian and other very rare endian format are unsupported.
+#endif
+
+// Unaligned memory access is automatically enabled for "common" CPU, such as x86.
+// For others CPU, the compiler will be more cautious, and insert extra code to ensure aligned access is respected
+// If you know your target CPU supports unaligned memory access, you want to force this option manually to improve performance
+#if defined(__ARM_FEATURE_UNALIGNED)
+#  define LZ4_FORCE_UNALIGNED_ACCESS 1
+#endif
+
+// Define this parameter if your target system or compiler does not support hardware bit count
+#if defined(_MSC_VER) && defined(_WIN32_WCE)            // Visual Studio for Windows CE does not support Hardware bit count
+#  define LZ4_FORCE_SW_BITCOUNT
+#endif
+
+
+//**************************************
+// Compiler Options
+//**************************************
+#if __STDC_VERSION__ >= 199901L    // C99
+  /* "restrict" is a known keyword */
+#else
+#  define restrict  // Disable restrict
+#endif
+
+#ifdef _MSC_VER
+#  define inline __inline             // Visual is not C99, but supports some kind of inline
+#  define forceinline __forceinline   
+#  include <intrin.h>                 // For Visual 2005
+#  if LZ4_ARCH64	// 64-bit
+#    pragma intrinsic(_BitScanForward64) // For Visual 2005
+#    pragma intrinsic(_BitScanReverse64) // For Visual 2005
+#  else
+#    pragma intrinsic(_BitScanForward)   // For Visual 2005
+#    pragma intrinsic(_BitScanReverse)   // For Visual 2005
+#  endif
+#else 
+#  ifdef __GNUC__
+#    define forceinline inline __attribute__((always_inline))
+#  else
+#    define forceinline inline
+#  endif
+#endif
+
+#ifdef _MSC_VER  // Visual Studio
+#define lz4_bswap16(x) _byteswap_ushort(x)
+#else
+#define lz4_bswap16(x)  ((unsigned short int) ((((x) >> 8) & 0xffu) | (((x) & 0xffu) << 8)))
+#endif
+
+
+//**************************************
+// Includes
+//**************************************
+#include <stdlib.h>   // calloc, free
+#include <string.h>   // memset, memcpy
+#include "lz4hc.h"
+
+#define ALLOCATOR(s) calloc(1,s)
+#define FREEMEM free
+#define MEM_INIT memset
+
+
+//**************************************
+// Basic Types
+//**************************************
+#if defined(_MSC_VER)    // Visual Studio does not support 'stdint' natively
+#define BYTE	unsigned __int8
+#define U16		unsigned __int16
+#define U32		unsigned __int32
+#define S32		__int32
+#define U64		unsigned __int64
+#else
+#include <stdint.h>
+#define BYTE	uint8_t
+#define U16		uint16_t
+#define U32		uint32_t
+#define S32		int32_t
+#define U64		uint64_t
+#endif
+
+#ifndef LZ4_FORCE_UNALIGNED_ACCESS
+#pragma pack(push, 1) 
+#endif
+
+typedef struct _U16_S { U16 v; } U16_S;
+typedef struct _U32_S { U32 v; } U32_S;
+typedef struct _U64_S { U64 v; } U64_S;
+
+#ifndef LZ4_FORCE_UNALIGNED_ACCESS
+#pragma pack(pop) 
+#endif
+
+#define A64(x) (((U64_S *)(x))->v)
+#define A32(x) (((U32_S *)(x))->v)
+#define A16(x) (((U16_S *)(x))->v)
+
+
+//**************************************
+// Constants
+//**************************************
+#define MINMATCH 4
+
+#define DICTIONARY_LOGSIZE 16
+#define MAXD (1<<DICTIONARY_LOGSIZE)
+#define MAXD_MASK ((U32)(MAXD - 1))
+#define MAX_DISTANCE (MAXD - 1)
+
+#define HASH_LOG (DICTIONARY_LOGSIZE-1)
+#define HASHTABLESIZE (1 << HASH_LOG)
+#define HASH_MASK (HASHTABLESIZE - 1)
+
+#define MAX_NB_ATTEMPTS 256
+
+#define ML_BITS  4
+#define ML_MASK  (size_t)((1U<<ML_BITS)-1)
+#define RUN_BITS (8-ML_BITS)
+#define RUN_MASK ((1U<<RUN_BITS)-1)
+
+#define COPYLENGTH 8
+#define LASTLITERALS 5
+#define MFLIMIT (COPYLENGTH+MINMATCH)
+#define MINLENGTH (MFLIMIT+1)
+#define OPTIMAL_ML (int)((ML_MASK-1)+MINMATCH)
+
+
+//**************************************
+// Architecture-specific macros
+//**************************************
+#if LZ4_ARCH64	// 64-bit
+#define STEPSIZE 8
+#define LZ4_COPYSTEP(s,d)		A64(d) = A64(s); d+=8; s+=8;
+#define LZ4_COPYPACKET(s,d)		LZ4_COPYSTEP(s,d)
+#define UARCH U64
+#define AARCH A64
+#define HTYPE					U32
+#define INITBASE(b,s)			const BYTE* const b = s
+#else		// 32-bit
+#define STEPSIZE 4
+#define LZ4_COPYSTEP(s,d)		A32(d) = A32(s); d+=4; s+=4;
+#define LZ4_COPYPACKET(s,d)		LZ4_COPYSTEP(s,d); LZ4_COPYSTEP(s,d);
+#define UARCH U32
+#define AARCH A32
+#define HTYPE					const BYTE*
+#define INITBASE(b,s)		    const int b = 0
+#endif
+
+#if defined(LZ4_BIG_ENDIAN)
+#define LZ4_READ_LITTLEENDIAN_16(d,s,p) { U16 v = A16(p); v = lz4_bswap16(v); d = (s) - v; }
+#define LZ4_WRITE_LITTLEENDIAN_16(p,i)  { U16 v = (U16)(i); v = lz4_bswap16(v); A16(p) = v; p+=2; }
+#else		// Little Endian
+#define LZ4_READ_LITTLEENDIAN_16(d,s,p) { d = (s) - A16(p); }
+#define LZ4_WRITE_LITTLEENDIAN_16(p,v)  { A16(p) = v; p+=2; }
+#endif
+
+
+//************************************************************
+// Local Types
+//************************************************************
+typedef struct 
+{
+    const BYTE* base;
+    HTYPE hashTable[HASHTABLESIZE];
+    U16 chainTable[MAXD];
+    const BYTE* nextToUpdate;
+} LZ4HC_Data_Structure;
+
+
+//**************************************
+// Macros
+//**************************************
+#define LZ4_WILDCOPY(s,d,e)    do { LZ4_COPYPACKET(s,d) } while (d<e);
+#define LZ4_BLINDCOPY(s,d,l)   { BYTE* e=d+l; LZ4_WILDCOPY(s,d,e); d=e; }
+#define HASH_FUNCTION(i)	   (((i) * 2654435761U) >> ((MINMATCH*8)-HASH_LOG))
+#define HASH_VALUE(p)		   HASH_FUNCTION(A32(p))
+#define HASH_POINTER(p)		   (HashTable[HASH_VALUE(p)] + base)
+#define DELTANEXT(p)		   chainTable[(size_t)(p) & MAXD_MASK] 
+#define GETNEXT(p)			   ((p) - (size_t)DELTANEXT(p))
+
+
+//**************************************
+// Private functions
+//**************************************
+#if LZ4_ARCH64
+
+inline static int LZ4_NbCommonBytes (register U64 val)
+{
+#if defined(LZ4_BIG_ENDIAN)
+    #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
+    unsigned long r = 0;
+    _BitScanReverse64( &r, val );
+    return (int)(r>>3);
+    #elif defined(__GNUC__) && ((__GNUC__ * 100 + __GNUC_MINOR__) >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
+    return (__builtin_clzll(val) >> 3); 
+    #else
+    int r;
+    if (!(val>>32)) { r=4; } else { r=0; val>>=32; }
+    if (!(val>>16)) { r+=2; val>>=8; } else { val>>=24; }
+    r += (!val);
+    return r;
+    #endif
+#else
+    #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
+    unsigned long r = 0;
+    _BitScanForward64( &r, val );
+    return (int)(r>>3);
+    #elif defined(__GNUC__) && ((__GNUC__ * 100 + __GNUC_MINOR__) >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
+    return (__builtin_ctzll(val) >> 3); 
+    #else
+    static const int DeBruijnBytePos[64] = { 0, 0, 0, 0, 0, 1, 1, 2, 0, 3, 1, 3, 1, 4, 2, 7, 0, 2, 3, 6, 1, 5, 3, 5, 1, 3, 4, 4, 2, 5, 6, 7, 7, 0, 1, 2, 3, 3, 4, 6, 2, 6, 5, 5, 3, 4, 5, 6, 7, 1, 2, 4, 6, 4, 4, 5, 7, 2, 6, 5, 7, 6, 7, 7 };
+    return DeBruijnBytePos[((U64)((val & -val) * 0x0218A392CDABBD3F)) >> 58];
+    #endif
+#endif
+}
+
+#else
+
+inline static int LZ4_NbCommonBytes (register U32 val)
+{
+#if defined(LZ4_BIG_ENDIAN)
+    #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
+    unsigned long r;
+    _BitScanReverse( &r, val );
+    return (int)(r>>3);
+    #elif defined(__GNUC__) && ((__GNUC__ * 100 + __GNUC_MINOR__) >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
+    return (__builtin_clz(val) >> 3); 
+    #else
+    int r;
+    if (!(val>>16)) { r=2; val>>=8; } else { r=0; val>>=24; }
+    r += (!val);
+    return r;
+    #endif
+#else
+    #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
+    unsigned long r;
+    _BitScanForward( &r, val );
+    return (int)(r>>3);
+    #elif defined(__GNUC__) && ((__GNUC__ * 100 + __GNUC_MINOR__) >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
+    return (__builtin_ctz(val) >> 3); 
+    #else
+    static const int DeBruijnBytePos[32] = { 0, 0, 3, 0, 3, 1, 3, 0, 3, 2, 2, 1, 3, 2, 0, 1, 3, 3, 1, 2, 2, 2, 2, 0, 3, 1, 2, 0, 1, 0, 1, 1 };
+    return DeBruijnBytePos[((U32)((val & -(S32)val) * 0x077CB531U)) >> 27];
+    #endif
+#endif
+}
+
+#endif
+
+
+inline static int LZ4HC_Init (LZ4HC_Data_Structure* hc4, const BYTE* base)
+{
+    MEM_INIT((void*)hc4->hashTable, 0, sizeof(hc4->hashTable));
+    MEM_INIT(hc4->chainTable, 0xFF, sizeof(hc4->chainTable));
+    hc4->nextToUpdate = base + LZ4_ARCH64;
+    hc4->base = base;
+    return 1;
+}
+
+
+inline static void* LZ4HC_Create (const BYTE* base)
+{
+    void* hc4 = ALLOCATOR(sizeof(LZ4HC_Data_Structure));
+
+    LZ4HC_Init ((LZ4HC_Data_Structure*)hc4, base);
+    return hc4;
+}
+
+
+inline static int LZ4HC_Free (void** LZ4HC_Data)
+{
+    FREEMEM(*LZ4HC_Data);
+    *LZ4HC_Data = NULL;
+    return (1);
+}
+
+
+// Update chains up to ip (excluded)
+forceinline static void LZ4HC_Insert (LZ4HC_Data_Structure* hc4, const BYTE* ip)
+{
+    U16*   chainTable = hc4->chainTable;
+    HTYPE* HashTable  = hc4->hashTable;
+    INITBASE(base,hc4->base);
+
+    while(hc4->nextToUpdate < ip)
+    {
+        const BYTE* p = hc4->nextToUpdate;
+        size_t delta = (p) - HASH_POINTER(p); 
+        if (delta>MAX_DISTANCE) delta = MAX_DISTANCE; 
+        DELTANEXT(p) = (U16)delta; 
+        HashTable[HASH_VALUE(p)] = (p) - base;
+        hc4->nextToUpdate++;
+    }
+}
+
+
+forceinline static size_t LZ4HC_CommonLength (const BYTE* p1, const BYTE* p2, const BYTE* const matchlimit)
+{
+    const BYTE* p1t = p1;
+
+    while (p1t<matchlimit-(STEPSIZE-1))
+    {
+        UARCH diff = AARCH(p2) ^ AARCH(p1t);
+        if (!diff) { p1t+=STEPSIZE; p2+=STEPSIZE; continue; }
+        p1t += LZ4_NbCommonBytes(diff);
+        return (p1t - p1);
+    }
+    if (LZ4_ARCH64) if ((p1t<(matchlimit-3)) && (A32(p2) == A32(p1t))) { p1t+=4; p2+=4; }
+    if ((p1t<(matchlimit-1)) && (A16(p2) == A16(p1t))) { p1t+=2; p2+=2; }
+    if ((p1t<matchlimit) && (*p2 == *p1t)) p1t++;
+    return (p1t - p1);
+}
+
+
+forceinline static int LZ4HC_InsertAndFindBestMatch (LZ4HC_Data_Structure* hc4, const BYTE* ip, const BYTE* const matchlimit, const BYTE** matchpos)
+{
+    U16* const chainTable = hc4->chainTable;
+    HTYPE* const HashTable = hc4->hashTable;
+    const BYTE* ref;
+    INITBASE(base,hc4->base);
+    int nbAttempts=MAX_NB_ATTEMPTS;
+    size_t ml=0;
+
+    // HC4 match finder
+    LZ4HC_Insert(hc4, ip);
+    ref = HASH_POINTER(ip);
+
+#if 1
+    if (ref >= ip-4)               // potential repetition
+    {
+        if (A32(ref) == A32(ip))   // confirmed
+        {
+            const U16 delta = (U16)(ip-ref);
+            const BYTE* ptr = ip;
+            const BYTE* end;
+            ml  = LZ4HC_CommonLength(ip+MINMATCH, ref+MINMATCH, matchlimit) + MINMATCH;
+            end = ip + ml - (MINMATCH-1);
+            while(ptr < end-delta)
+            {
+                DELTANEXT(ptr) = delta;    // Pre-Load
+                ptr++;
+            }
+            do
+            {
+                DELTANEXT(ptr) = delta;    
+                HashTable[HASH_VALUE(ptr)] = (ptr) - base;     // Head of chain
+                ptr++;
+            } while(ptr < end);
+            hc4->nextToUpdate = end;
+            *matchpos = ref;
+        }
+        ref = GETNEXT(ref);
+    }
+#endif
+
+    while ((ref >= ip-MAX_DISTANCE) && (ref >= hc4->base) && (nbAttempts))
+    {
+        nbAttempts--;
+        if (*(ref+ml) == *(ip+ml))
+        if (A32(ref) == A32(ip))
+        {
+            size_t mlt = LZ4HC_CommonLength(ip+MINMATCH, ref+MINMATCH, matchlimit) + MINMATCH;
+            if (mlt > ml) { ml = mlt; *matchpos = ref; }
+        }
+        ref = GETNEXT(ref);
+    }
+
+    return (int)ml;
+}
+
+
+forceinline static int LZ4HC_InsertAndGetWiderMatch (LZ4HC_Data_Structure* hc4, const BYTE* ip, const BYTE* startLimit, const BYTE* matchlimit, int longest, const BYTE** matchpos, const BYTE** startpos)
+{
+    U16* const  chainTable = hc4->chainTable;
+    HTYPE* const HashTable = hc4->hashTable;
+    INITBASE(base,hc4->base);
+    const BYTE*  ref;
+    int nbAttempts = MAX_NB_ATTEMPTS;
+    int delta = (int)(ip-startLimit);
+
+    // First Match
+    LZ4HC_Insert(hc4, ip);
+    ref = HASH_POINTER(ip);
+
+    while ((ref >= ip-MAX_DISTANCE) && (ref >= hc4->base) && (nbAttempts))
+    {
+        nbAttempts--;
+        if (*(startLimit + longest) == *(ref - delta + longest))
+        if (A32(ref) == A32(ip))
+        {
+#if 1
+            const BYTE* reft = ref+MINMATCH;
+            const BYTE* ipt = ip+MINMATCH;
+            const BYTE* startt = ip;
+
+            while (ipt<matchlimit-(STEPSIZE-1))
+            {
+                UARCH diff = AARCH(reft) ^ AARCH(ipt);
+                if (!diff) { ipt+=STEPSIZE; reft+=STEPSIZE; continue; }
+                ipt += LZ4_NbCommonBytes(diff);
+                goto _endCount;
+            }
+            if (LZ4_ARCH64) if ((ipt<(matchlimit-3)) && (A32(reft) == A32(ipt))) { ipt+=4; reft+=4; }
+            if ((ipt<(matchlimit-1)) && (A16(reft) == A16(ipt))) { ipt+=2; reft+=2; }
+            if ((ipt<matchlimit) && (*reft == *ipt)) ipt++;
+_endCount:
+            reft = ref;
+#else
+            // Easier for code maintenance, but unfortunately slower too
+            const BYTE* startt = ip;
+            const BYTE* reft = ref;
+            const BYTE* ipt = ip + MINMATCH + LZ4HC_CommonLength(ip+MINMATCH, ref+MINMATCH, matchlimit);
+#endif
+
+            while ((startt>startLimit) && (reft > hc4->base) && (startt[-1] == reft[-1])) {startt--; reft--;}
+
+            if ((ipt-startt) > longest)
+            {
+                longest = (int)(ipt-startt);
+                *matchpos = reft;
+                *startpos = startt;
+            }
+        }
+        ref = GETNEXT(ref);
+    }
+
+    return longest;
+}
+
+
+forceinline static int LZ4_encodeSequence(const BYTE** ip, BYTE** op, const BYTE** anchor, int ml, const BYTE* ref)
+{
+    int length, len; 
+    BYTE* token;
+
+    // Encode Literal length
+    length = (int)(*ip - *anchor);
+    token = (*op)++;
+    if (length>=(int)RUN_MASK) { *token=(RUN_MASK<<ML_BITS); len = length-RUN_MASK; for(; len > 254 ; len-=255) *(*op)++ = 255;  *(*op)++ = (BYTE)len; } 
+    else *token = (length<<ML_BITS);
+
+    // Copy Literals
+    LZ4_BLINDCOPY(*anchor, *op, length);
+
+    // Encode Offset
+    LZ4_WRITE_LITTLEENDIAN_16(*op,(U16)(*ip-ref));
+
+    // Encode MatchLength
+    len = (int)(ml-MINMATCH);
+    if (len>=(int)ML_MASK) { *token+=ML_MASK; len-=ML_MASK; for(; len > 509 ; len-=510) { *(*op)++ = 255; *(*op)++ = 255; } if (len > 254) { len-=255; *(*op)++ = 255; } *(*op)++ = (BYTE)len; } 
+    else *token += len;	
+
+    // Prepare next loop
+    *ip += ml;
+    *anchor = *ip; 
+
+    return 0;
+}
+
+
+//****************************
+// Compression CODE
+//****************************
+
+int LZ4_compressHCCtx(LZ4HC_Data_Structure* ctx,
+                 const char* source, 
+                 char* dest,
+                 int isize)
+{	
+    const BYTE* ip = (const BYTE*) source;
+    const BYTE* anchor = ip;
+    const BYTE* const iend = ip + isize;
+    const BYTE* const mflimit = iend - MFLIMIT;
+    const BYTE* const matchlimit = (iend - LASTLITERALS);
+
+    BYTE* op = (BYTE*) dest;
+
+    int	ml, ml2, ml3, ml0;
+    const BYTE* ref=NULL;
+    const BYTE* start2=NULL;
+    const BYTE* ref2=NULL;
+    const BYTE* start3=NULL;
+    const BYTE* ref3=NULL;
+    const BYTE* start0;
+    const BYTE* ref0;
+
+    ip++;
+
+    // Main Loop
+    while (ip < mflimit)
+    {
+        ml = LZ4HC_InsertAndFindBestMatch (ctx, ip, matchlimit, (&ref));
+        if (!ml) { ip++; continue; }
+
+        // saved, in case we would skip too much
+        start0 = ip;
+        ref0 = ref;
+        ml0 = ml;
+
+_Search2:
+        if (ip+ml < mflimit)
+            ml2 = LZ4HC_InsertAndGetWiderMatch(ctx, ip + ml - 2, ip + 1, matchlimit, ml, &ref2, &start2);
+        else ml2 = ml;
+
+        if (ml2 == ml)  // No better match
+        {
+            LZ4_encodeSequence(&ip, &op, &anchor, ml, ref);
+            continue;
+        }
+
+        if (start0 < ip)
+        {
+            if (start2 < ip + ml0)   // empirical
+            {
+                ip = start0;
+                ref = ref0;
+                ml = ml0;
+            }
+        }
+
+        // Here, start0==ip
+        if ((start2 - ip) < 3)   // First Match too small : removed
+        {
+            ml = ml2;
+            ip = start2;
+            ref =ref2;
+            goto _Search2;
+        }
+
+_Search3:
+        // Currently we have :
+        // ml2 > ml1, and
+        // ip1+3 <= ip2 (usually < ip1+ml1)
+        if ((start2 - ip) < OPTIMAL_ML)
+        {
+            int correction;
+            int new_ml = ml;
+            if (new_ml > OPTIMAL_ML) new_ml = OPTIMAL_ML;
+            if (ip+new_ml > start2 + ml2 - MINMATCH) new_ml = (int)(start2 - ip) + ml2 - MINMATCH;
+            correction = new_ml - (int)(start2 - ip);
+            if (correction > 0)
+            {
+                start2 += correction;
+                ref2 += correction;
+                ml2 -= correction;
+            }
+        }
+        // Now, we have start2 = ip+new_ml, with new_ml = min(ml, OPTIMAL_ML=18)
+
+        if (start2 + ml2 < mflimit)
+            ml3 = LZ4HC_InsertAndGetWiderMatch(ctx, start2 + ml2 - 3, start2, matchlimit, ml2, &ref3, &start3);
+        else ml3 = ml2;
+
+        if (ml3 == ml2) // No better match : 2 sequences to encode
+        {
+            // ip & ref are known; Now for ml
+            if (start2 < ip+ml)  ml = (int)(start2 - ip);
+            // Now, encode 2 sequences
+            LZ4_encodeSequence(&ip, &op, &anchor, ml, ref);
+            ip = start2;
+            LZ4_encodeSequence(&ip, &op, &anchor, ml2, ref2);
+            continue;
+        }
+
+        if (start3 < ip+ml+3) // Not enough space for match 2 : remove it
+        {
+            if (start3 >= (ip+ml)) // can write Seq1 immediately ==> Seq2 is removed, so Seq3 becomes Seq1
+            {
+                if (start2 < ip+ml)
+                {
+                    int correction = (int)(ip+ml - start2);
+                    start2 += correction;
+                    ref2 += correction;
+                    ml2 -= correction;
+                    if (ml2 < MINMATCH)
+                    {
+                        start2 = start3;
+                        ref2 = ref3;
+                        ml2 = ml3;
+                    }
+                }
+
+                LZ4_encodeSequence(&ip, &op, &anchor, ml, ref);
+                ip  = start3;
+                ref = ref3;
+                ml  = ml3;
+
+                start0 = start2;
+                ref0 = ref2;
+                ml0 = ml2;
+                goto _Search2;
+            }
+
+            start2 = start3;
+            ref2 = ref3;
+            ml2 = ml3;
+            goto _Search3;
+        }
+
+        // OK, now we have 3 ascending matches; let's write at least the first one
+        // ip & ref are known; Now for ml
+        if (start2 < ip+ml)
+        {
+            if ((start2 - ip) < (int)ML_MASK)
+            {
+                int correction;
+                if (ml > OPTIMAL_ML) ml = OPTIMAL_ML;
+                if (ip + ml > start2 + ml2 - MINMATCH) ml = (int)(start2 - ip) + ml2 - MINMATCH;
+                correction = ml - (int)(start2 - ip);
+                if (correction > 0)
+                {
+                    start2 += correction;
+                    ref2 += correction;
+                    ml2 -= correction;
+                }
+            }
+            else
+            {
+                ml = (int)(start2 - ip);
+            }
+        }
+        LZ4_encodeSequence(&ip, &op, &anchor, ml, ref);
+
+        ip = start2;
+        ref = ref2;
+        ml = ml2;
+
+        start2 = start3;
+        ref2 = ref3;
+        ml2 = ml3;
+
+        goto _Search3;
+
+    }
+
+    // Encode Last Literals
+    {
+        int lastRun = (int)(iend - anchor);
+        if (lastRun>=(int)RUN_MASK) { *op++=(RUN_MASK<<ML_BITS); lastRun-=RUN_MASK; for(; lastRun > 254 ; lastRun-=255) *op++ = 255; *op++ = (BYTE) lastRun; } 
+        else *op++ = (lastRun<<ML_BITS);
+        memcpy(op, anchor, iend - anchor);
+        op += iend-anchor;
+    } 
+
+    // End
+    return (int) (((char*)op)-dest);
+}
+
+
+int LZ4_compressHC(const char* source, 
+                 char* dest,
+                 int isize)
+{
+    void* ctx = LZ4HC_Create((const BYTE*)source);
+    int result = LZ4_compressHCCtx(ctx, source, dest, isize);
+    LZ4HC_Free (&ctx);
+
+    return result;
+}
+
+
diff --git a/src/cbits/lz4hc.h b/src/cbits/lz4hc.h
--- a/src/cbits/lz4hc.h
+++ b/src/cbits/lz4hc.h
@@ -1,60 +1,60 @@
-/*
-   LZ4 HC - High Compression Mode of LZ4
-   Header File
-   Copyright (C) 2011-2012, Yann Collet.
-   BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
-
-   Redistribution and use in source and binary forms, with or without
-   modification, are permitted provided that the following conditions are
-   met:
-
-       * Redistributions of source code must retain the above copyright
-   notice, this list of conditions and the following disclaimer.
-       * Redistributions in binary form must reproduce the above
-   copyright notice, this list of conditions and the following disclaimer
-   in the documentation and/or other materials provided with the
-   distribution.
-
-   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
-   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
-   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
-   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
-   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
-   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
-   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
-   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
-   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
-   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
-   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-
-   You can contact the author at :
-   - LZ4 homepage : http://fastcompression.blogspot.com/p/lz4.html
-   - LZ4 source repository : http://code.google.com/p/lz4/
-*/
-#pragma once
-
-
-#if defined (__cplusplus)
-extern "C" {
-#endif
-
-
-int LZ4_compressHC (const char* source, char* dest, int isize);
-
-/*
-LZ4_compressHC :
-	return : the number of bytes in compressed buffer dest
-	note : destination buffer must be already allocated. 
-		To avoid any problem, size it to handle worst cases situations (input data not compressible)
-		Worst case size evaluation is provided by function LZ4_compressBound() (see "lz4.h")
-*/
-
-
-/* Note :
-Decompression functions are provided within regular LZ4 source code (see "lz4.h") (BSD license)
-*/
-
-
-#if defined (__cplusplus)
-}
-#endif
+/*
+   LZ4 HC - High Compression Mode of LZ4
+   Header File
+   Copyright (C) 2011-2012, Yann Collet.
+   BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
+
+   Redistribution and use in source and binary forms, with or without
+   modification, are permitted provided that the following conditions are
+   met:
+
+       * Redistributions of source code must retain the above copyright
+   notice, this list of conditions and the following disclaimer.
+       * Redistributions in binary form must reproduce the above
+   copyright notice, this list of conditions and the following disclaimer
+   in the documentation and/or other materials provided with the
+   distribution.
+
+   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+   You can contact the author at :
+   - LZ4 homepage : http://fastcompression.blogspot.com/p/lz4.html
+   - LZ4 source repository : http://code.google.com/p/lz4/
+*/
+#pragma once
+
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+
+int LZ4_compressHC (const char* source, char* dest, int isize);
+
+/*
+LZ4_compressHC :
+	return : the number of bytes in compressed buffer dest
+	note : destination buffer must be already allocated. 
+		To avoid any problem, size it to handle worst cases situations (input data not compressible)
+		Worst case size evaluation is provided by function LZ4_compressBound() (see "lz4.h")
+*/
+
+
+/* Note :
+Decompression functions are provided within regular LZ4 source code (see "lz4.h") (BSD license)
+*/
+
+
+#if defined (__cplusplus)
+}
+#endif