xxhash 0.2.0 → 0.3.0

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: eb65711481aa5bbc9e1b7dc6c8338e0bda70f377
+  data.tar.gz: 5c4c2b5a262bd5831f9658aeccf35457a0556ce8
+SHA512:
+  metadata.gz: a1160fe558a79aaa74a9fabe74e902890bcb7a7512ad08ff4634839fa40dfb2abe74eef367b689c5d42f178380a77deba6d7d6a8cd15ebf94d52a6600c4a7e85
+  data.tar.gz: 5615f46479f9c86b2af11dc36cea59abd1205ba2b06f0364c572bb5b54aa5a816d5f643619b5c0f43965fc9928763ea3c8b112a8134a391beee09a7d527f4d0a

data/.gitignore

@@ -1,6 +1,7 @@
 *.gem
 *.rbc
 *.bundle
+**/*.iml
 .config
 .yardoc
 Gemfile.lock
@@ -16,3 +17,4 @@ test/tmp
 test/version_tmp
 tmp
 lib/xxhash/xxhash.so
+.idea/

data/CHANGELOG.md

@@ -1,3 +1,7 @@
+### master
+  * make seed param optional (by [@weakish](https://github.com/weakish))
+  * add 64-bit xxhash function (by [@justinwsmith](https://github.com/justinwsmith)) 
+
 ### 0.2.0 (September 4, 2013)
   * xxHash updated to [r32](https://code.google.com/p/xxhash/source/detail?r=32)
   * add `XXhash.xxh32_stream` method (by [@maltoe](https://github.com/maltoe))

data/Rakefile

@@ -14,6 +14,7 @@ task :default => :test
 
 require 'rake/extensiontask'
 Rake::ExtensionTask.new('xxhash') do |ext|
+  # Customizations can be set here
   ext.lib_dir = 'lib/xxhash'
 end
 

data/ext/xxhash/extconf.rb

@@ -1,8 +1,7 @@
 require 'mkmf'
 
-%w{g O3 Wall}.each do |flag|
-  flag = "-#{flag}"
-  $CPPFLAGS += " #{flag}" unless $CPPFLAGS.split.include? flag
+%w{ g O3  Wno-padded }.each do |flag|
+  $CFLAGS += " -#{flag}"
 end
 
 create_makefile('xxhash/xxhash')

data/ext/xxhash/libxxhash.c

@@ -1,6 +1,6 @@
 /*
 xxHash - Fast Hash algorithm
-Copyright (C) 2012-2013, Yann Collet.
+Copyright (C) 2012-2014, 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
@@ -28,6 +28,7 @@ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 You can contact the author at :
 - xxHash source repository : http://code.google.com/p/xxhash/
+- public discussion board : https://groups.google.com/forum/#!forum/lz4c
 */
 
 
@@ -47,7 +48,7 @@ You can contact the author at :
 // When this option is enabled, xxHash output for null input pointers will be the same as a null-length input.
 // This option has a very small performance cost (only measurable on small inputs).
 // By default, this option is disabled. To enable it, uncomment below define :
-//#define XXH_ACCEPT_NULL_INPUT_POINTER 1
+// #define XXH_ACCEPT_NULL_INPUT_POINTER 1
 
 // XXH_FORCE_NATIVE_FORMAT :
 // By default, xxHash library provides endian-independant Hash values, based on little-endian convention.
@@ -58,7 +59,6 @@ You can contact the author at :
 // This option has no impact on Little_Endian CPU.
 #define XXH_FORCE_NATIVE_FORMAT 0
 
-
 //**************************************
 // Compiler Specific Options
 //**************************************
@@ -68,28 +68,30 @@ You can contact the author at :
 #endif
 
 #ifdef _MSC_VER    // Visual Studio
-#  define forceinline static __forceinline
+#  define FORCE_INLINE static __forceinline
 #else
 #  ifdef __GNUC__
-#    define forceinline static inline __attribute__((always_inline))
+#    define FORCE_INLINE static inline __attribute__((always_inline))
 #  else
-#    define forceinline static inline
+#    define FORCE_INLINE static inline
 #  endif
 #endif
 
-
 //**************************************
 // Includes & Memory related functions
 //**************************************
 #include "libxxhash.h"
-// Modify the local functions below should you wish to use some other memory related routines
+// Modify the local functions below should you wish to use some other memory routines
 // for malloc(), free()
 #include <stdlib.h>
-forceinline void* XXH_malloc(size_t s) { return malloc(s); }
-forceinline void  XXH_free  (void* p)  { free(p); }
+static void* XXH_malloc(size_t s) { return malloc(s); }
+static void  XXH_free  (void* p)  { free(p); }
 // for memcpy()
 #include <string.h>
-forceinline void* XXH_memcpy(void* dest, const void* src, size_t size) { return memcpy(dest,src,size); }
+static void* XXH_memcpy(void* dest, const void* src, size_t size)
+{
+    return memcpy(dest,src,size);
+}
 
 
 //**************************************
@@ -97,17 +99,17 @@ forceinline void* XXH_memcpy(void* dest, const void* src, size_t size) { return
 //**************************************
 #if defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L   // C99
 # include <stdint.h>
-  typedef uint8_t  BYTE;
-  typedef uint16_t U16;
-  typedef uint32_t U32;
-  typedef  int32_t S32;
-  typedef uint64_t U64;
+typedef uint8_t  BYTE;
+typedef uint16_t U16;
+typedef uint32_t U32;
+typedef  int32_t S32;
+typedef uint64_t U64;
 #else
-  typedef unsigned char      BYTE;
-  typedef unsigned short     U16;
-  typedef unsigned int       U32;
-  typedef   signed int       S32;
-  typedef unsigned long long U64;
+typedef unsigned char      BYTE;
+typedef unsigned short     U16;
+typedef unsigned int       U32;
+typedef   signed int       S32;
+typedef unsigned long long U64;
 #endif
 
 #if defined(__GNUC__)  && !defined(XXH_USE_UNALIGNED_ACCESS)
@@ -124,13 +126,21 @@ forceinline void* XXH_memcpy(void* dest, const void* src, size_t size) { return
 #  endif
 #endif
 
-typedef struct _U32_S { U32 v; } _PACKED U32_S;
+typedef struct _U32_S
+{
+    U32 v;
+} _PACKED U32_S;
+typedef struct _U64_S
+{
+    U64 v;
+} _PACKED U64_S;
 
 #if !defined(XXH_USE_UNALIGNED_ACCESS) && !defined(__GNUC__)
 #  pragma pack(pop)
 #endif
 
 #define A32(x) (((U32_S *)(x))->v)
+#define A64(x) (((U64_S *)(x))->v)
 
 
 //***************************************
@@ -141,20 +151,37 @@ typedef struct _U32_S { U32 v; } _PACKED U32_S;
 // Note : although _rotl exists for minGW (GCC under windows), performance seems poor
 #if defined(_MSC_VER)
 #  define XXH_rotl32(x,r) _rotl(x,r)
+#  define XXH_rotl64(x,r) _rotl64(x,r)
 #else
 #  define XXH_rotl32(x,r) ((x << r) | (x >> (32 - r)))
+#  define XXH_rotl64(x,r) ((x << r) | (x >> (64 - r)))
 #endif
 
 #if defined(_MSC_VER)     // Visual Studio
 #  define XXH_swap32 _byteswap_ulong
+#  define XXH_swap64 _byteswap_uint64
 #elif GCC_VERSION >= 403
 #  define XXH_swap32 __builtin_bswap32
+#  define XXH_swap64 __builtin_bswap64
 #else
-static inline U32 XXH_swap32 (U32 x) {
+static inline U32 XXH_swap32 (U32 x)
+{
     return  ((x << 24) & 0xff000000 ) |
-        ((x <<  8) & 0x00ff0000 ) |
-        ((x >>  8) & 0x0000ff00 ) |
-        ((x >> 24) & 0x000000ff );}
+            ((x <<  8) & 0x00ff0000 ) |
+            ((x >>  8) & 0x0000ff00 ) |
+            ((x >> 24) & 0x000000ff );
+}
+static inline U64 XXH_swap64 (U64 x)
+{
+    return  ((x << 56) & 0xff00000000000000ULL) |
+            ((x << 40) & 0x00ff000000000000ULL) |
+            ((x << 24) & 0x0000ff0000000000ULL) |
+            ((x << 8)  & 0x000000ff00000000ULL) |
+            ((x >> 8)  & 0x00000000ff000000ULL) |
+            ((x >> 24) & 0x0000000000ff0000ULL) |
+            ((x >> 40) & 0x000000000000ff00ULL) |
+            ((x >> 56) & 0x00000000000000ffULL);
+}
 #endif
 
 
@@ -167,13 +194,18 @@ static inline U32 XXH_swap32 (U32 x) {
 #define PRIME32_4    668265263U
 #define PRIME32_5    374761393U
 
+#define PRIME64_1 11400714785074694791ULL
+#define PRIME64_2 14029467366897019727ULL
+#define PRIME64_3  1609587929392839161ULL
+#define PRIME64_4  9650029242287828579ULL
+#define PRIME64_5  2870177450012600261ULL
 
 //**************************************
 // Architecture Macros
 //**************************************
 typedef enum { XXH_bigEndian=0, XXH_littleEndian=1 } XXH_endianess;
 #ifndef XXH_CPU_LITTLE_ENDIAN   // It is possible to define XXH_CPU_LITTLE_ENDIAN externally, for example using a compiler switch
-    static const int one = 1;
+static const int one = 1;
 #   define XXH_CPU_LITTLE_ENDIAN   (*(char*)(&one))
 #endif
 
@@ -189,28 +221,49 @@ typedef enum { XXH_bigEndian=0, XXH_littleEndian=1 } XXH_endianess;
 //****************************
 typedef enum { XXH_aligned, XXH_unaligned } XXH_alignment;
 
-forceinline U32 XXH_readLE32_align(const U32* ptr, XXH_endianess endian, XXH_alignment align)
+FORCE_INLINE U32 XXH_readLE32_align(const void* ptr, XXH_endianess endian, XXH_alignment align)
 {
     if (align==XXH_unaligned)
         return endian==XXH_littleEndian ? A32(ptr) : XXH_swap32(A32(ptr));
     else
-        return endian==XXH_littleEndian ? *ptr : XXH_swap32(*ptr);
+        return endian==XXH_littleEndian ? *(U32*)ptr : XXH_swap32(*(U32*)ptr);
+}
+
+FORCE_INLINE U32 XXH_readLE32(const void* ptr, XXH_endianess endian)
+{
+    return XXH_readLE32_align(ptr, endian, XXH_unaligned);
+}
+
+FORCE_INLINE U64 XXH_readLE64_align(const void* ptr, XXH_endianess endian, XXH_alignment align)
+{
+    if (align==XXH_unaligned)
+        return endian==XXH_littleEndian ? A64(ptr) : XXH_swap64(A64(ptr));
+    else
+        return endian==XXH_littleEndian ? *(U64*)ptr : XXH_swap64(*(U64*)ptr);
 }
 
-forceinline U32 XXH_readLE32(const U32* ptr, XXH_endianess endian) { return XXH_readLE32_align(ptr, endian, XXH_unaligned); }
+FORCE_INLINE U64 XXH_readLE64(const void* ptr, XXH_endianess endian)
+{
+    return XXH_readLE64_align(ptr, endian, XXH_unaligned);
+}
 
 
 //****************************
 // Simple Hash Functions
 //****************************
-forceinline U32 XXH32_endian_align(const void* input, int len, U32 seed, XXH_endianess endian, XXH_alignment align)
+FORCE_INLINE U32 XXH32_endian_align(const void* input, size_t len, U32 seed, XXH_endianess endian, XXH_alignment align)
 {
     const BYTE* p = (const BYTE*)input;
-    const BYTE* const bEnd = p + len;
+    const BYTE* bEnd = p + len;
     U32 h32;
+#define XXH_get32bits(p) XXH_readLE32_align(p, endian, align)
 
 #ifdef XXH_ACCEPT_NULL_INPUT_POINTER
-    if (p==NULL) { len=0; p=(const BYTE*)(size_t)16; }
+    if (p==NULL)
+    {
+        len=0;
+        bEnd=p=(const BYTE*)(size_t)16;
+    }
 #endif
 
     if (len>=16)
@@ -223,11 +276,24 @@ forceinline U32 XXH32_endian_align(const void* input, int len, U32 seed, XXH_end
 
         do
         {
-            v1 += XXH_readLE32_align((const U32*)p, endian, align) * PRIME32_2; v1 = XXH_rotl32(v1, 13); v1 *= PRIME32_1; p+=4;
-            v2 += XXH_readLE32_align((const U32*)p, endian, align) * PRIME32_2; v2 = XXH_rotl32(v2, 13); v2 *= PRIME32_1; p+=4;
-            v3 += XXH_readLE32_align((const U32*)p, endian, align) * PRIME32_2; v3 = XXH_rotl32(v3, 13); v3 *= PRIME32_1; p+=4;
-            v4 += XXH_readLE32_align((const U32*)p, endian, align) * PRIME32_2; v4 = XXH_rotl32(v4, 13); v4 *= PRIME32_1; p+=4;
-        } while (p<=limit);
+            v1 += XXH_get32bits(p) * PRIME32_2;
+            v1 = XXH_rotl32(v1, 13);
+            v1 *= PRIME32_1;
+            p+=4;
+            v2 += XXH_get32bits(p) * PRIME32_2;
+            v2 = XXH_rotl32(v2, 13);
+            v2 *= PRIME32_1;
+            p+=4;
+            v3 += XXH_get32bits(p) * PRIME32_2;
+            v3 = XXH_rotl32(v3, 13);
+            v3 *= PRIME32_1;
+            p+=4;
+            v4 += XXH_get32bits(p) * PRIME32_2;
+            v4 = XXH_rotl32(v4, 13);
+            v4 *= PRIME32_1;
+            p+=4;
+        }
+        while (p<=limit);
 
         h32 = XXH_rotl32(v1, 1) + XXH_rotl32(v2, 7) + XXH_rotl32(v3, 12) + XXH_rotl32(v4, 18);
     }
@@ -238,9 +304,9 @@ forceinline U32 XXH32_endian_align(const void* input, int len, U32 seed, XXH_end
 
     h32 += (U32) len;
 
-    while (p<=bEnd-4)
+    while (p+4<=bEnd)
     {
-        h32 += XXH_readLE32_align((const U32*)p, endian, align) * PRIME32_3;
+        h32 += XXH_get32bits(p) * PRIME32_3;
         h32  = XXH_rotl32(h32, 17) * PRIME32_4 ;
         p+=4;
     }
@@ -262,18 +328,19 @@ forceinline U32 XXH32_endian_align(const void* input, int len, U32 seed, XXH_end
 }
 
 
-U32 XXH32(const void* input, int len, U32 seed)
+unsigned int XXH32 (const void* input, size_t len, unsigned seed)
 {
 #if 0
     // Simple version, good for code maintenance, but unfortunately slow for small inputs
-    void* state = XXH32_init(seed);
-    XXH32_update(state, input, len);
-    return XXH32_digest(state);
+    XXH32_state_t state;
+    XXH32_reset(&state, seed);
+    XXH32_update(&state, input, len);
+    return XXH32_digest(&state);
 #else
     XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
 
 #  if !defined(XXH_USE_UNALIGNED_ACCESS)
-    if ((((size_t)input) & 3))   // Input is aligned, let's leverage the speed advantage
+    if ((((size_t)input) & 3) == 0)   // Input is aligned, let's leverage the speed advantage
     {
         if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
             return XXH32_endian_align(input, len, seed, XXH_littleEndian, XXH_aligned);
@@ -289,12 +356,152 @@ U32 XXH32(const void* input, int len, U32 seed)
 #endif
 }
 
+FORCE_INLINE U64 XXH64_endian_align(const void* input, size_t len, U64 seed, XXH_endianess endian, XXH_alignment align)
+{
+    const BYTE* p = (const BYTE*)input;
+    const BYTE* bEnd = p + len;
+    U64 h64;
+#define XXH_get64bits(p) XXH_readLE64_align(p, endian, align)
 
-//****************************
-// Advanced Hash Functions
-//****************************
+#ifdef XXH_ACCEPT_NULL_INPUT_POINTER
+    if (p==NULL)
+    {
+        len=0;
+        bEnd=p=(const BYTE*)(size_t)32;
+    }
+#endif
+
+    if (len>=32)
+    {
+        const BYTE* const limit = bEnd - 32;
+        U64 v1 = seed + PRIME64_1 + PRIME64_2;
+        U64 v2 = seed + PRIME64_2;
+        U64 v3 = seed + 0;
+        U64 v4 = seed - PRIME64_1;
+
+        do
+        {
+            v1 += XXH_get64bits(p) * PRIME64_2;
+            p+=8;
+            v1 = XXH_rotl64(v1, 31);
+            v1 *= PRIME64_1;
+            v2 += XXH_get64bits(p) * PRIME64_2;
+            p+=8;
+            v2 = XXH_rotl64(v2, 31);
+            v2 *= PRIME64_1;
+            v3 += XXH_get64bits(p) * PRIME64_2;
+            p+=8;
+            v3 = XXH_rotl64(v3, 31);
+            v3 *= PRIME64_1;
+            v4 += XXH_get64bits(p) * PRIME64_2;
+            p+=8;
+            v4 = XXH_rotl64(v4, 31);
+            v4 *= PRIME64_1;
+        }
+        while (p<=limit);
+
+        h64 = XXH_rotl64(v1, 1) + XXH_rotl64(v2, 7) + XXH_rotl64(v3, 12) + XXH_rotl64(v4, 18);
+
+        v1 *= PRIME64_2;
+        v1 = XXH_rotl64(v1, 31);
+        v1 *= PRIME64_1;
+        h64 ^= v1;
+        h64 = h64 * PRIME64_1 + PRIME64_4;
+
+        v2 *= PRIME64_2;
+        v2 = XXH_rotl64(v2, 31);
+        v2 *= PRIME64_1;
+        h64 ^= v2;
+        h64 = h64 * PRIME64_1 + PRIME64_4;
+
+        v3 *= PRIME64_2;
+        v3 = XXH_rotl64(v3, 31);
+        v3 *= PRIME64_1;
+        h64 ^= v3;
+        h64 = h64 * PRIME64_1 + PRIME64_4;
+
+        v4 *= PRIME64_2;
+        v4 = XXH_rotl64(v4, 31);
+        v4 *= PRIME64_1;
+        h64 ^= v4;
+        h64 = h64 * PRIME64_1 + PRIME64_4;
+    }
+    else
+    {
+        h64  = seed + PRIME64_5;
+    }
+
+    h64 += (U64) len;
+
+    while (p+8<=bEnd)
+    {
+        U64 k1 = XXH_get64bits(p);
+        k1 *= PRIME64_2;
+        k1 = XXH_rotl64(k1,31);
+        k1 *= PRIME64_1;
+        h64 ^= k1;
+        h64 = XXH_rotl64(h64,27) * PRIME64_1 + PRIME64_4;
+        p+=8;
+    }
+
+    if (p+4<=bEnd)
+    {
+        h64 ^= (U64)(XXH_get32bits(p)) * PRIME64_1;
+        h64 = XXH_rotl64(h64, 23) * PRIME64_2 + PRIME64_3;
+        p+=4;
+    }
+
+    while (p<bEnd)
+    {
+        h64 ^= (*p) * PRIME64_5;
+        h64 = XXH_rotl64(h64, 11) * PRIME64_1;
+        p++;
+    }
+
+    h64 ^= h64 >> 33;
+    h64 *= PRIME64_2;
+    h64 ^= h64 >> 29;
+    h64 *= PRIME64_3;
+    h64 ^= h64 >> 32;
+
+    return h64;
+}
+
+
+unsigned long long XXH64 (const void* input, size_t len, unsigned long long seed)
+{
+#if 0
+    // Simple version, good for code maintenance, but unfortunately slow for small inputs
+    XXH64_state_t state;
+    XXH64_reset(&state, seed);
+    XXH64_update(&state, input, len);
+    return XXH64_digest(&state);
+#else
+    XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
+
+#  if !defined(XXH_USE_UNALIGNED_ACCESS)
+    if ((((size_t)input) & 7)==0)   // Input is aligned, let's leverage the speed advantage
+    {
+        if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
+            return XXH64_endian_align(input, len, seed, XXH_littleEndian, XXH_aligned);
+        else
+            return XXH64_endian_align(input, len, seed, XXH_bigEndian, XXH_aligned);
+    }
+#  endif
+
+    if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
+        return XXH64_endian_align(input, len, seed, XXH_littleEndian, XXH_unaligned);
+    else
+        return XXH64_endian_align(input, len, seed, XXH_bigEndian, XXH_unaligned);
+#endif
+}
 
-struct XXH_state32_t
+/****************************************************
+ *  Advanced Hash Functions
+****************************************************/
+
+/*** Allocation ***/
+typedef struct
 {
     U64 total_len;
     U32 seed;
@@ -302,21 +509,51 @@ struct XXH_state32_t
     U32 v2;
     U32 v3;
     U32 v4;
-    int memsize;
-    char memory[16];
-};
+    U32 mem32[4];   /* defined as U32 for alignment */
+    U32 memsize;
+} XXH_istate32_t;
+
+typedef struct
+{
+    U64 total_len;
+    U64 seed;
+    U64 v1;
+    U64 v2;
+    U64 v3;
+    U64 v4;
+    U64 mem64[4];   /* defined as U64 for alignment */
+    U32 memsize;
+} XXH_istate64_t;
 
 
-int XXH32_sizeofState()
+XXH32_state_t* XXH32_createState(void)
+{
+    XXH_STATIC_ASSERT(sizeof(XXH32_state_t) >= sizeof(XXH_istate32_t));   // A compilation error here means XXH32_state_t is not large enough
+    return (XXH32_state_t*)XXH_malloc(sizeof(XXH32_state_t));
+}
+XXH_errorcode XXH32_freeState(XXH32_state_t* statePtr)
+{
+    XXH_free(statePtr);
+    return XXH_OK;
+}
+
+XXH64_state_t* XXH64_createState(void)
+{
+    XXH_STATIC_ASSERT(sizeof(XXH64_state_t) >= sizeof(XXH_istate64_t));   // A compilation error here means XXH64_state_t is not large enough
+    return (XXH64_state_t*)XXH_malloc(sizeof(XXH64_state_t));
+}
+XXH_errorcode XXH64_freeState(XXH64_state_t* statePtr)
 {
-    XXH_STATIC_ASSERT(XXH32_SIZEOFSTATE >= sizeof(struct XXH_state32_t));   // A compilation error here means XXH32_SIZEOFSTATE is not large enough
-    return sizeof(struct XXH_state32_t);
+    XXH_free(statePtr);
+    return XXH_OK;
 }
 
 
-XXH_errorcode XXH32_resetState(void* state_in, U32 seed)
+/*** Hash feed ***/
+
+XXH_errorcode XXH32_reset(XXH32_state_t* state_in, U32 seed)
 {
-    struct XXH_state32_t * state = (struct XXH_state32_t *) state_in;
+    XXH_istate32_t* state = (XXH_istate32_t*) state_in;
     state->seed = seed;
     state->v1 = seed + PRIME32_1 + PRIME32_2;
     state->v2 = seed + PRIME32_2;
@@ -327,18 +564,23 @@ XXH_errorcode XXH32_resetState(void* state_in, U32 seed)
     return XXH_OK;
 }
 
-
-void* XXH32_init (U32 seed)
+XXH_errorcode XXH64_reset(XXH64_state_t* state_in, unsigned long long seed)
 {
-    void* state = XXH_malloc (sizeof(struct XXH_state32_t));
-    XXH32_resetState(state, seed);
-    return state;
+    XXH_istate64_t* state = (XXH_istate64_t*) state_in;
+    state->seed = seed;
+    state->v1 = seed + PRIME64_1 + PRIME64_2;
+    state->v2 = seed + PRIME64_2;
+    state->v3 = seed + 0;
+    state->v4 = seed - PRIME64_1;
+    state->total_len = 0;
+    state->memsize = 0;
+    return XXH_OK;
 }
 
 
-forceinline XXH_errorcode XXH32_update_endian (void* state_in, const void* input, int len, XXH_endianess endian)
+FORCE_INLINE XXH_errorcode XXH32_update_endian (XXH32_state_t* state_in, const void* input, size_t len, XXH_endianess endian)
 {
-    struct XXH_state32_t * state = (struct XXH_state32_t *) state_in;
+    XXH_istate32_t* state = (XXH_istate32_t *) state_in;
     const BYTE* p = (const BYTE*)input;
     const BYTE* const bEnd = p + len;
 
@@ -350,20 +592,32 @@ forceinline XXH_errorcode XXH32_update_endian (void* state_in, const void* input
 
     if (state->memsize + len < 16)   // fill in tmp buffer
     {
-        XXH_memcpy(state->memory + state->memsize, input, len);
-        state->memsize +=  len;
+        XXH_memcpy((BYTE*)(state->mem32) + state->memsize, input, len);
+        state->memsize += (U32)len;
         return XXH_OK;
     }
 
     if (state->memsize)   // some data left from previous update
     {
-        XXH_memcpy(state->memory + state->memsize, input, 16-state->memsize);
+        XXH_memcpy((BYTE*)(state->mem32) + state->memsize, input, 16-state->memsize);
         {
-            const U32* p32 = (const U32*)state->memory;
-            state->v1 += XXH_readLE32(p32, endian) * PRIME32_2; state->v1 = XXH_rotl32(state->v1, 13); state->v1 *= PRIME32_1; p32++;
-            state->v2 += XXH_readLE32(p32, endian) * PRIME32_2; state->v2 = XXH_rotl32(state->v2, 13); state->v2 *= PRIME32_1; p32++;
-            state->v3 += XXH_readLE32(p32, endian) * PRIME32_2; state->v3 = XXH_rotl32(state->v3, 13); state->v3 *= PRIME32_1; p32++;
-            state->v4 += XXH_readLE32(p32, endian) * PRIME32_2; state->v4 = XXH_rotl32(state->v4, 13); state->v4 *= PRIME32_1; p32++;
+            const U32* p32 = state->mem32;
+            state->v1 += XXH_readLE32(p32, endian) * PRIME32_2;
+            state->v1 = XXH_rotl32(state->v1, 13);
+            state->v1 *= PRIME32_1;
+            p32++;
+            state->v2 += XXH_readLE32(p32, endian) * PRIME32_2;
+            state->v2 = XXH_rotl32(state->v2, 13);
+            state->v2 *= PRIME32_1;
+            p32++;
+            state->v3 += XXH_readLE32(p32, endian) * PRIME32_2;
+            state->v3 = XXH_rotl32(state->v3, 13);
+            state->v3 *= PRIME32_1;
+            p32++;
+            state->v4 += XXH_readLE32(p32, endian) * PRIME32_2;
+            state->v4 = XXH_rotl32(state->v4, 13);
+            state->v4 *= PRIME32_1;
+            p32++;
         }
         p += 16-state->memsize;
         state->memsize = 0;
@@ -379,11 +633,24 @@ forceinline XXH_errorcode XXH32_update_endian (void* state_in, const void* input
 
         do
         {
-            v1 += XXH_readLE32((const U32*)p, endian) * PRIME32_2; v1 = XXH_rotl32(v1, 13); v1 *= PRIME32_1; p+=4;
-            v2 += XXH_readLE32((const U32*)p, endian) * PRIME32_2; v2 = XXH_rotl32(v2, 13); v2 *= PRIME32_1; p+=4;
-            v3 += XXH_readLE32((const U32*)p, endian) * PRIME32_2; v3 = XXH_rotl32(v3, 13); v3 *= PRIME32_1; p+=4;
-            v4 += XXH_readLE32((const U32*)p, endian) * PRIME32_2; v4 = XXH_rotl32(v4, 13); v4 *= PRIME32_1; p+=4;
-        } while (p<=limit);
+            v1 += XXH_readLE32(p, endian) * PRIME32_2;
+            v1 = XXH_rotl32(v1, 13);
+            v1 *= PRIME32_1;
+            p+=4;
+            v2 += XXH_readLE32(p, endian) * PRIME32_2;
+            v2 = XXH_rotl32(v2, 13);
+            v2 *= PRIME32_1;
+            p+=4;
+            v3 += XXH_readLE32(p, endian) * PRIME32_2;
+            v3 = XXH_rotl32(v3, 13);
+            v3 *= PRIME32_1;
+            p+=4;
+            v4 += XXH_readLE32(p, endian) * PRIME32_2;
+            v4 = XXH_rotl32(v4, 13);
+            v4 *= PRIME32_1;
+            p+=4;
+        }
+        while (p<=limit);
 
         state->v1 = v1;
         state->v2 = v2;
@@ -393,14 +660,14 @@ forceinline XXH_errorcode XXH32_update_endian (void* state_in, const void* input
 
     if (p < bEnd)
     {
-        XXH_memcpy(state->memory, p, bEnd-p);
-        state->memsize = (int)(bEnd-p);
+        XXH_memcpy(state->mem32, p, (size_t)(bEnd-p));
+        state->memsize = (U32)(bEnd-p);
     }
 
     return XXH_OK;
 }
 
-XXH_errorcode XXH32_update (void* state_in, const void* input, int len)
+XXH_errorcode XXH32_update (XXH32_state_t* state_in, const void* input, size_t len)
 {
     XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
 
@@ -412,11 +679,11 @@ XXH_errorcode XXH32_update (void* state_in, const void* input, int len)
 
 
 
-forceinline U32 XXH32_intermediateDigest_endian (void* state_in, XXH_endianess endian)
+FORCE_INLINE U32 XXH32_digest_endian (const XXH32_state_t* state_in, XXH_endianess endian)
 {
-    struct XXH_state32_t * state = (struct XXH_state32_t *) state_in;
-    const BYTE * p = (const BYTE*)state->memory;
-    BYTE* bEnd = (BYTE*)state->memory + state->memsize;
+    XXH_istate32_t* state = (XXH_istate32_t*) state_in;
+    const BYTE * p = (const BYTE*)state->mem32;
+    BYTE* bEnd = (BYTE*)(state->mem32) + state->memsize;
     U32 h32;
 
     if (state->total_len >= 16)
@@ -430,9 +697,9 @@ forceinline U32 XXH32_intermediateDigest_endian (void* state_in, XXH_endianess e
 
     h32 += (U32) state->total_len;
 
-    while (p<=bEnd-4)
+    while (p+4<=bEnd)
     {
-        h32 += XXH_readLE32((const U32*)p, endian) * PRIME32_3;
+        h32 += XXH_readLE32(p, endian) * PRIME32_3;
         h32  = XXH_rotl32(h32, 17) * PRIME32_4;
         p+=4;
     }
@@ -454,22 +721,207 @@ forceinline U32 XXH32_intermediateDigest_endian (void* state_in, XXH_endianess e
 }
 
 
-U32 XXH32_intermediateDigest (void* state_in)
+U32 XXH32_digest (const XXH32_state_t* state_in)
 {
     XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
 
     if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
-        return XXH32_intermediateDigest_endian(state_in, XXH_littleEndian);
+        return XXH32_digest_endian(state_in, XXH_littleEndian);
     else
-        return XXH32_intermediateDigest_endian(state_in, XXH_bigEndian);
+        return XXH32_digest_endian(state_in, XXH_bigEndian);
 }
 
 
-U32 XXH32_digest (void* state_in)
+FORCE_INLINE XXH_errorcode XXH64_update_endian (XXH64_state_t* state_in, const void* input, size_t len, XXH_endianess endian)
 {
-    U32 h32 = XXH32_intermediateDigest(state_in);
+    XXH_istate64_t * state = (XXH_istate64_t *) state_in;
+    const BYTE* p = (const BYTE*)input;
+    const BYTE* const bEnd = p + len;
 
-    XXH_free(state_in);
+#ifdef XXH_ACCEPT_NULL_INPUT_POINTER
+    if (input==NULL) return XXH_ERROR;
+#endif
 
-    return h32;
+    state->total_len += len;
+
+    if (state->memsize + len < 32)   // fill in tmp buffer
+    {
+        XXH_memcpy(((BYTE*)state->mem64) + state->memsize, input, len);
+        state->memsize += (U32)len;
+        return XXH_OK;
+    }
+
+    if (state->memsize)   // some data left from previous update
+    {
+        XXH_memcpy(((BYTE*)state->mem64) + state->memsize, input, 32-state->memsize);
+        {
+            const U64* p64 = state->mem64;
+            state->v1 += XXH_readLE64(p64, endian) * PRIME64_2;
+            state->v1 = XXH_rotl64(state->v1, 31);
+            state->v1 *= PRIME64_1;
+            p64++;
+            state->v2 += XXH_readLE64(p64, endian) * PRIME64_2;
+            state->v2 = XXH_rotl64(state->v2, 31);
+            state->v2 *= PRIME64_1;
+            p64++;
+            state->v3 += XXH_readLE64(p64, endian) * PRIME64_2;
+            state->v3 = XXH_rotl64(state->v3, 31);
+            state->v3 *= PRIME64_1;
+            p64++;
+            state->v4 += XXH_readLE64(p64, endian) * PRIME64_2;
+            state->v4 = XXH_rotl64(state->v4, 31);
+            state->v4 *= PRIME64_1;
+            p64++;
+        }
+        p += 32-state->memsize;
+        state->memsize = 0;
+    }
+
+    if (p+32 <= bEnd)
+    {
+        const BYTE* const limit = bEnd - 32;
+        U64 v1 = state->v1;
+        U64 v2 = state->v2;
+        U64 v3 = state->v3;
+        U64 v4 = state->v4;
+
+        do
+        {
+            v1 += XXH_readLE64(p, endian) * PRIME64_2;
+            v1 = XXH_rotl64(v1, 31);
+            v1 *= PRIME64_1;
+            p+=8;
+            v2 += XXH_readLE64(p, endian) * PRIME64_2;
+            v2 = XXH_rotl64(v2, 31);
+            v2 *= PRIME64_1;
+            p+=8;
+            v3 += XXH_readLE64(p, endian) * PRIME64_2;
+            v3 = XXH_rotl64(v3, 31);
+            v3 *= PRIME64_1;
+            p+=8;
+            v4 += XXH_readLE64(p, endian) * PRIME64_2;
+            v4 = XXH_rotl64(v4, 31);
+            v4 *= PRIME64_1;
+            p+=8;
+        }
+        while (p<=limit);
+
+        state->v1 = v1;
+        state->v2 = v2;
+        state->v3 = v3;
+        state->v4 = v4;
+    }
+
+    if (p < bEnd)
+    {
+        XXH_memcpy(state->mem64, p, (size_t)(bEnd-p));
+        state->memsize = (U32)(bEnd-p);
+    }
+
+    return XXH_OK;
+}
+
+XXH_errorcode XXH64_update (XXH64_state_t* state_in, const void* input, size_t len)
+{
+    XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
+
+    if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
+        return XXH64_update_endian(state_in, input, len, XXH_littleEndian);
+    else
+        return XXH64_update_endian(state_in, input, len, XXH_bigEndian);
+}
+
+
+
+FORCE_INLINE U64 XXH64_digest_endian (const XXH64_state_t* state_in, XXH_endianess endian)
+{
+    XXH_istate64_t * state = (XXH_istate64_t *) state_in;
+    const BYTE * p = (const BYTE*)state->mem64;
+    BYTE* bEnd = (BYTE*)state->mem64 + state->memsize;
+    U64 h64;
+
+    if (state->total_len >= 32)
+    {
+        U64 v1 = state->v1;
+        U64 v2 = state->v2;
+        U64 v3 = state->v3;
+        U64 v4 = state->v4;
+
+        h64 = XXH_rotl64(v1, 1) + XXH_rotl64(v2, 7) + XXH_rotl64(v3, 12) + XXH_rotl64(v4, 18);
+
+        v1 *= PRIME64_2;
+        v1 = XXH_rotl64(v1, 31);
+        v1 *= PRIME64_1;
+        h64 ^= v1;
+        h64 = h64*PRIME64_1 + PRIME64_4;
+
+        v2 *= PRIME64_2;
+        v2 = XXH_rotl64(v2, 31);
+        v2 *= PRIME64_1;
+        h64 ^= v2;
+        h64 = h64*PRIME64_1 + PRIME64_4;
+
+        v3 *= PRIME64_2;
+        v3 = XXH_rotl64(v3, 31);
+        v3 *= PRIME64_1;
+        h64 ^= v3;
+        h64 = h64*PRIME64_1 + PRIME64_4;
+
+        v4 *= PRIME64_2;
+        v4 = XXH_rotl64(v4, 31);
+        v4 *= PRIME64_1;
+        h64 ^= v4;
+        h64 = h64*PRIME64_1 + PRIME64_4;
+    }
+    else
+    {
+        h64  = state->seed + PRIME64_5;
+    }
+
+    h64 += (U64) state->total_len;
+
+    while (p+8<=bEnd)
+    {
+        U64 k1 = XXH_readLE64(p, endian);
+        k1 *= PRIME64_2;
+        k1 = XXH_rotl64(k1,31);
+        k1 *= PRIME64_1;
+        h64 ^= k1;
+        h64 = XXH_rotl64(h64,27) * PRIME64_1 + PRIME64_4;
+        p+=8;
+    }
+
+    if (p+4<=bEnd)
+    {
+        h64 ^= (U64)(XXH_readLE32(p, endian)) * PRIME64_1;
+        h64 = XXH_rotl64(h64, 23) * PRIME64_2 + PRIME64_3;
+        p+=4;
+    }
+
+    while (p<bEnd)
+    {
+        h64 ^= (*p) * PRIME64_5;
+        h64 = XXH_rotl64(h64, 11) * PRIME64_1;
+        p++;
+    }
+
+    h64 ^= h64 >> 33;
+    h64 *= PRIME64_2;
+    h64 ^= h64 >> 29;
+    h64 *= PRIME64_3;
+    h64 ^= h64 >> 32;
+
+    return h64;
 }
+
+
+unsigned long long XXH64_digest (const XXH64_state_t* state_in)
+{
+    XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
+
+    if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
+        return XXH64_digest_endian(state_in, XXH_littleEndian);
+    else
+        return XXH64_digest_endian(state_in, XXH_bigEndian);
+}
+