diff --git a/CompressSave/LZ4WrapC/CMakeLists.txt b/CompressSave/LZ4WrapC/CMakeLists.txt index d726611..53cd444 100644 --- a/CompressSave/LZ4WrapC/CMakeLists.txt +++ b/CompressSave/LZ4WrapC/CMakeLists.txt @@ -1,4 +1,4 @@ -cmake_minimum_required(VERSION 3.2) +cmake_minimum_required(VERSION 3.12) project(lz4wrap) diff --git a/CompressSave/LZ4WrapC/LZ4Wrap.c b/CompressSave/LZ4WrapC/LZ4Wrap.c index 542b318..12922f5 100644 --- a/CompressSave/LZ4WrapC/LZ4Wrap.c +++ b/CompressSave/LZ4WrapC/LZ4Wrap.c @@ -5,16 +5,19 @@ #include #include #include -#define Check(assert,errorcode) if(!(assert)) {printf(LZ4F_getErrorName(errorcode)); return errorcode;} +#define Check(assert, errorcode) \ + if (!(assert)) \ + { \ + printf("%s\n", LZ4F_getErrorName(errorcode)); \ + return errorcode; \ + } - - -static CContext* CreateCompressContext() +static CContext *CreateCompressContext() { - return (CContext*)malloc(sizeof(CContext)); + return (CContext *)malloc(sizeof(CContext)); } -void __stdcall CompressContextFree(CContext* ctx) +void __stdcall CompressContextFree(CContext *ctx) { if (ctx != NULL) { @@ -25,12 +28,12 @@ void __stdcall CompressContextFree(CContext* ctx) } static LZ4F_preferences_t kPrefs = { - { LZ4F_max4MB, LZ4F_blockLinked, LZ4F_contentChecksumEnabled, LZ4F_frame, - 0 /* unknown content size */, 0/* no dictID */ , LZ4F_blockChecksumEnabled }, - 0, /* compression level; 0 == default */ - 0, /* autoflush */ - 0, /* favor decompression speed */ - { 0, 0, 0 }, /* reserved, must be set to 0 */ + {LZ4F_max4MB, LZ4F_blockLinked, LZ4F_contentChecksumEnabled, LZ4F_frame, + 0 /* unknown content size */, 0 /* no dictID */, LZ4F_blockChecksumEnabled}, + 0, /* compression level; 0 == default */ + 0, /* autoflush */ + 0, /* favor decompression speed */ + {0, 0, 0}, /* reserved, must be set to 0 */ }; size_t __stdcall CompressBufferBound(size_t inBufferSize) @@ -38,8 +41,9 @@ size_t __stdcall CompressBufferBound(size_t inBufferSize) return LZ4F_compressBound(inBufferSize, &kPrefs) + LZ4F_HEADER_SIZE_MAX; } -CContext* CreateCompressContextFromBuffer(void* dict, size_t dictSize) { - CContext* ctx = CreateCompressContext(); +CContext *CreateCompressContextFromBuffer(void *dict, size_t dictSize) +{ + CContext *ctx = CreateCompressContext(); if (dict) ctx->dict = LZ4F_createCDict(dict, dictSize); else @@ -59,18 +63,20 @@ CContext* CreateCompressContextFromBuffer(void* dict, size_t dictSize) { return ctx; } -size_t __stdcall CompressBegin(CContext** pctx, int compressionLevel, void* outBuff , size_t outCapacity, void* dict, size_t dictSize) +size_t __stdcall CompressBegin(CContext **pctx, int compressionLevel, void *outBuff, size_t outCapacity, void *dict, size_t dictSize) { - CContext* ctx = CreateCompressContextFromBuffer(dict, dictSize); - if (ctx == NULL) return -1; + CContext *ctx = CreateCompressContextFromBuffer(dict, dictSize); + if (ctx == NULL) + return -1; - if (outCapacity < LZ4F_HEADER_SIZE_MAX || outCapacity < LZ4F_compressBound(0, &kPrefs)) return LZ4F_ERROR_dstMaxSize_tooSmall; + if (outCapacity < LZ4F_HEADER_SIZE_MAX || outCapacity < LZ4F_compressBound(0, &kPrefs)) + return LZ4F_ERROR_dstMaxSize_tooSmall; kPrefs.compressionLevel = compressionLevel; /* write frame header */ - size_t const headerSize = ctx->dict == NULL - ? LZ4F_compressBegin(ctx->cctx, outBuff, outCapacity, &kPrefs) - : LZ4F_compressBegin_usingCDict(ctx->cctx, outBuff, outCapacity, ctx->dict, &kPrefs); + size_t const headerSize = ctx->dict == NULL + ? LZ4F_compressBegin(ctx->cctx, outBuff, outCapacity, &kPrefs) + : LZ4F_compressBegin_usingCDict(ctx->cctx, outBuff, outCapacity, ctx->dict, &kPrefs); if (LZ4F_isError(headerSize)) { @@ -80,12 +86,11 @@ size_t __stdcall CompressBegin(CContext** pctx, int compressionLevel, void* outB return headerSize; } - -size_t __stdcall CompressUpdate(CContext* ctx,void* dstBuffer, size_t dstCapacity,const void* srcBuffer, size_t srcSize) +size_t __stdcall CompressUpdate(CContext *ctx, void *dstBuffer, size_t dstCapacity, const void *srcBuffer, size_t srcSize) { size_t result = ctx->dict == NULL - ? LZ4F_compressUpdate(ctx->cctx, dstBuffer, dstCapacity, srcBuffer, srcSize, NULL) - : LZ4F_compressFrame_usingCDict(ctx->cctx, dstBuffer, dstCapacity, srcBuffer, srcSize, ctx->dict, NULL); + ? LZ4F_compressUpdate(ctx->cctx, dstBuffer, dstCapacity, srcBuffer, srcSize, NULL) + : LZ4F_compressFrame_usingCDict(ctx->cctx, dstBuffer, dstCapacity, srcBuffer, srcSize, ctx->dict, NULL); if (LZ4F_isError(result)) { const char *str = LZ4F_getErrorName(result); @@ -94,32 +99,38 @@ size_t __stdcall CompressUpdate(CContext* ctx,void* dstBuffer, size_t dstCapacit return result; } -size_t __stdcall CompressEnd(CContext* ctx, void* dstBuffer, size_t dstCapacity) +size_t __stdcall CompressEnd(CContext *ctx, void *dstBuffer, size_t dstCapacity) { size_t writeSize = LZ4F_compressEnd(ctx->cctx, dstBuffer, dstCapacity, NULL); return writeSize; } -static size_t get_block_size(const LZ4F_frameInfo_t* info) { - switch (info->blockSizeID) { - case LZ4F_default: - case LZ4F_max64KB: return 1 << 16; - case LZ4F_max256KB: return 1 << 18; - case LZ4F_max1MB: return 1 << 20; - case LZ4F_max4MB: return 1 << 22; - default: - return -1; +static size_t get_block_size(const LZ4F_frameInfo_t *info) +{ + switch (info->blockSizeID) + { + case LZ4F_default: + case LZ4F_max64KB: + return 1 << 16; + case LZ4F_max256KB: + return 1 << 18; + case LZ4F_max1MB: + return 1 << 20; + case LZ4F_max4MB: + return 1 << 22; + default: + return -1; } } -//return: input bytes expects for next call -size_t __stdcall DecompressBegin(DContext **pdctx,void *inBuffer,size_t *inBufferSize, size_t *blockSize, void* dict, size_t dictSize) +// return: input bytes expects for next call +size_t __stdcall DecompressBegin(DContext **pdctx, void *inBuffer, size_t *inBufferSize, size_t *blockSize, void *dict, size_t dictSize) { - DContext* dctx = (DContext*)malloc(sizeof(DContext)); + DContext *dctx = (DContext *)malloc(sizeof(DContext)); size_t const dctxStatus = LZ4F_createDecompressionContext(&dctx->dctx, LZ4F_VERSION); Check(!LZ4F_isError(dctxStatus), dctxStatus); Check(*inBufferSize >= LZ4F_HEADER_SIZE_MAX, LZ4F_ERROR_dstMaxSize_tooSmall); - + LZ4F_frameInfo_t info; size_t const fires = LZ4F_getFrameInfo(dctx->dctx, &info, inBuffer, inBufferSize); Check(!LZ4F_isError(fires), fires); @@ -131,21 +142,21 @@ size_t __stdcall DecompressBegin(DContext **pdctx,void *inBuffer,size_t *inBuffe return fires; } -void __stdcall DecompressContextReset(DContext* dctx) +void __stdcall DecompressContextReset(DContext *dctx) { LZ4F_resetDecompressionContext(dctx->dctx); } -size_t __stdcall DecompressUpdate(DContext* dctx, void* outBuffer, size_t * outBufferSize, void* inBuffer, size_t * inBufferSize) +size_t __stdcall DecompressUpdate(DContext *dctx, void *outBuffer, size_t *outBufferSize, void *inBuffer, size_t *inBufferSize) { size_t ret = dctx->dict == NULL - ? LZ4F_decompress(dctx->dctx, outBuffer, outBufferSize, inBuffer, inBufferSize, NULL) - : LZ4F_decompress_usingDict(dctx->dctx, outBuffer, outBufferSize, inBuffer, inBufferSize, dctx->dict, dctx->dictSize, NULL); + ? LZ4F_decompress(dctx->dctx, outBuffer, outBufferSize, inBuffer, inBufferSize, NULL) + : LZ4F_decompress_usingDict(dctx->dctx, outBuffer, outBufferSize, inBuffer, inBufferSize, dctx->dict, dctx->dictSize, NULL); Check(!LZ4F_isError(ret), ret); return ret; } -size_t __stdcall DecompressEnd(DContext* dctx) +size_t __stdcall DecompressEnd(DContext *dctx) { if (!dctx) return 0; size_t r = LZ4F_freeDecompressionContext(dctx->dctx); diff --git a/CompressSave/LZ4WrapC/lz4/README.md b/CompressSave/LZ4WrapC/lz4/README.md index a86cd0d..a2d35bb 100644 --- a/CompressSave/LZ4WrapC/lz4/README.md +++ b/CompressSave/LZ4WrapC/lz4/README.md @@ -17,7 +17,7 @@ They generate and decode data using the [LZ4 block format]. #### Level 2 : High Compression variant -For more compression ratio at the cost of compression speed, +For better compression ratio at the cost of compression speed, the High Compression variant called **lz4hc** is available. Add files **`lz4hc.c`** and **`lz4hc.h`**. This variant also compresses data using the [LZ4 block format], @@ -108,6 +108,12 @@ The following build macro can be selected to adjust source code behavior at comp Remove support of dynamic memory allocation. For more details, see description of this macro in `lib/lz4.c`. +- `LZ4_STATIC_LINKING_ONLY_ENDIANNESS_INDEPENDENT_OUTPUT` : experimental feature aimed at producing the same + compressed output on platforms of different endianness (i.e. little-endian and big-endian). + Output on little-endian platforms shall remain unchanged, while big-endian platforms will start producing + the same output as little-endian ones. This isn't expected to impact backward- and forward-compatibility + in any way. + - `LZ4_FREESTANDING` : by setting this build macro to 1, LZ4/HC removes dependencies on the C standard library, including allocation functions and `memmove()`, `memcpy()`, and `memset()`. @@ -131,8 +137,8 @@ The following build macro can be selected to adjust source code behavior at comp #### Makefile variables The following `Makefile` variables can be selected to alter the profile of produced binaries : -- `BUILD_SHARED` : generate `libzstd` dynamic library (enabled by default) -- `BUILD_STATIC` : generate `libzstd` static library (enabled by default) +- `BUILD_SHARED` : generate `liblz4` dynamic library (enabled by default) +- `BUILD_STATIC` : generate `liblz4` static library (enabled by default) #### Amalgamation diff --git a/CompressSave/LZ4WrapC/lz4/lz4.c b/CompressSave/LZ4WrapC/lz4/lz4.c index 0982f95..1233a8e 100644 --- a/CompressSave/LZ4WrapC/lz4/lz4.c +++ b/CompressSave/LZ4WrapC/lz4/lz4.c @@ -1,6 +1,6 @@ /* LZ4 - Fast LZ compression algorithm - Copyright (C) 2011-2020, Yann Collet. + Copyright (c) Yann Collet. All rights reserved. BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) @@ -79,7 +79,7 @@ ( defined(__ARM_ARCH_6__) || defined(__ARM_ARCH_6J__) || defined(__ARM_ARCH_6K__) \ || defined(__ARM_ARCH_6Z__) || defined(__ARM_ARCH_6ZK__) || defined(__ARM_ARCH_6T2__) ) # define LZ4_FORCE_MEMORY_ACCESS 2 -# elif (defined(__INTEL_COMPILER) && !defined(_WIN32)) || defined(__GNUC__) +# elif (defined(__INTEL_COMPILER) && !defined(_WIN32)) || defined(__GNUC__) || defined(_MSC_VER) # define LZ4_FORCE_MEMORY_ACCESS 1 # endif #endif @@ -106,15 +106,13 @@ # define LZ4_SRC_INCLUDED 1 #endif -#ifndef LZ4_STATIC_LINKING_ONLY -#define LZ4_STATIC_LINKING_ONLY -#endif - #ifndef LZ4_DISABLE_DEPRECATE_WARNINGS -#define LZ4_DISABLE_DEPRECATE_WARNINGS /* due to LZ4_decompress_safe_withPrefix64k */ +# define LZ4_DISABLE_DEPRECATE_WARNINGS /* due to LZ4_decompress_safe_withPrefix64k */ #endif -#define LZ4_STATIC_LINKING_ONLY /* LZ4_DISTANCE_MAX */ +#ifndef LZ4_STATIC_LINKING_ONLY +# define LZ4_STATIC_LINKING_ONLY +#endif #include "lz4.h" /* see also "memory routines" below */ @@ -126,14 +124,17 @@ # include /* only present in VS2005+ */ # pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */ # pragma warning(disable : 6237) /* disable: C6237: conditional expression is always 0 */ +# pragma warning(disable : 6239) /* disable: C6239: ( && ) always evaluates to the result of */ +# pragma warning(disable : 6240) /* disable: C6240: ( && ) always evaluates to the result of */ +# pragma warning(disable : 6326) /* disable: C6326: Potential comparison of a constant with another constant */ #endif /* _MSC_VER */ #ifndef LZ4_FORCE_INLINE -# ifdef _MSC_VER /* Visual Studio */ +# if defined (_MSC_VER) && !defined (__clang__) /* MSVC */ # define LZ4_FORCE_INLINE static __forceinline # else # if defined (__cplusplus) || defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 */ -# ifdef __GNUC__ +# if defined (__GNUC__) || defined (__clang__) # define LZ4_FORCE_INLINE static inline __attribute__((always_inline)) # else # define LZ4_FORCE_INLINE static inline @@ -300,12 +301,12 @@ static int LZ4_isAligned(const void* ptr, size_t alignment) #include #if defined(__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) # include - typedef uint8_t BYTE; - typedef uint16_t U16; - typedef uint32_t U32; - typedef int32_t S32; - typedef uint64_t U64; - typedef uintptr_t uptrval; + typedef unsigned char BYTE; /*uint8_t not necessarily blessed to alias arbitrary type*/ + typedef uint16_t U16; + typedef uint32_t U32; + typedef int32_t S32; + typedef uint64_t U64; + typedef uintptr_t uptrval; #else # if UINT_MAX != 4294967295UL # error "LZ4 code (when not C++ or C99) assumes that sizeof(int) == 4" @@ -365,6 +366,11 @@ static unsigned LZ4_isLittleEndian(void) return one.c[0]; } +#if defined(__GNUC__) || defined(__INTEL_COMPILER) +#define LZ4_PACK( __Declaration__ ) __Declaration__ __attribute__((__packed__)) +#elif defined(_MSC_VER) +#define LZ4_PACK( __Declaration__ ) __pragma( pack(push, 1) ) __Declaration__ __pragma( pack(pop)) +#endif #if defined(LZ4_FORCE_MEMORY_ACCESS) && (LZ4_FORCE_MEMORY_ACCESS==2) /* lie to the compiler about data alignment; use with caution */ @@ -380,9 +386,9 @@ static void LZ4_write32(void* memPtr, U32 value) { *(U32*)memPtr = value; } /* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */ /* currently only defined for gcc and icc */ -typedef struct { U16 u16; } __attribute__((packed)) LZ4_unalign16; -typedef struct { U32 u32; } __attribute__((packed)) LZ4_unalign32; -typedef struct { reg_t uArch; } __attribute__((packed)) LZ4_unalignST; +LZ4_PACK(typedef struct { U16 u16; }) LZ4_unalign16; +LZ4_PACK(typedef struct { U32 u32; }) LZ4_unalign32; +LZ4_PACK(typedef struct { reg_t uArch; }) LZ4_unalignST; static U16 LZ4_read16(const void* ptr) { return ((const LZ4_unalign16*)ptr)->u16; } static U32 LZ4_read32(const void* ptr) { return ((const LZ4_unalign32*)ptr)->u32; } @@ -427,10 +433,22 @@ static U16 LZ4_readLE16(const void* memPtr) return LZ4_read16(memPtr); } else { const BYTE* p = (const BYTE*)memPtr; - return (U16)((U16)p[0] + (p[1]<<8)); + return (U16)((U16)p[0] | (p[1]<<8)); } } +#ifdef LZ4_STATIC_LINKING_ONLY_ENDIANNESS_INDEPENDENT_OUTPUT +static U32 LZ4_readLE32(const void* memPtr) +{ + if (LZ4_isLittleEndian()) { + return LZ4_read32(memPtr); + } else { + const BYTE* p = (const BYTE*)memPtr; + return (U32)p[0] | (p[1]<<8) | (p[2]<<16) | (p[3]<<24); + } +} +#endif + static void LZ4_writeLE16(void* memPtr, U16 value) { if (LZ4_isLittleEndian()) { @@ -460,13 +478,15 @@ static const int dec64table[8] = {0, 0, 0, -1, -4, 1, 2, 3}; #ifndef LZ4_FAST_DEC_LOOP # if defined __i386__ || defined _M_IX86 || defined __x86_64__ || defined _M_X64 # define LZ4_FAST_DEC_LOOP 1 -# elif defined(__aarch64__) && defined(__APPLE__) -# define LZ4_FAST_DEC_LOOP 1 -# elif defined(__aarch64__) && !defined(__clang__) - /* On non-Apple aarch64, we disable this optimization for clang because +# elif defined(__aarch64__) +# if defined(__clang__) && defined(__ANDROID__) + /* On Android aarch64, we disable this optimization for clang because * on certain mobile chipsets, performance is reduced with clang. For * more information refer to https://github.com/lz4/lz4/pull/707 */ -# define LZ4_FAST_DEC_LOOP 1 +# define LZ4_FAST_DEC_LOOP 0 +# else +# define LZ4_FAST_DEC_LOOP 1 +# endif # else # define LZ4_FAST_DEC_LOOP 0 # endif @@ -512,7 +532,7 @@ LZ4_wildCopy32(void* dstPtr, const void* srcPtr, void* dstEnd) /* LZ4_memcpy_using_offset() presumes : * - dstEnd >= dstPtr + MINMATCH - * - there is at least 8 bytes available to write after dstEnd */ + * - there is at least 12 bytes available to write after dstEnd */ LZ4_FORCE_INLINE void LZ4_memcpy_using_offset(BYTE* dstPtr, const BYTE* srcPtr, BYTE* dstEnd, const size_t offset) { @@ -527,12 +547,12 @@ LZ4_memcpy_using_offset(BYTE* dstPtr, const BYTE* srcPtr, BYTE* dstEnd, const si case 2: LZ4_memcpy(v, srcPtr, 2); LZ4_memcpy(&v[2], srcPtr, 2); -#if defined(_MSC_VER) && (_MSC_VER <= 1936) /* MSVC 2022 ver 17.6 or earlier */ +#if defined(_MSC_VER) && (_MSC_VER <= 1937) /* MSVC 2022 ver 17.7 or earlier */ # pragma warning(push) # pragma warning(disable : 6385) /* warning C6385: Reading invalid data from 'v'. */ #endif LZ4_memcpy(&v[4], v, 4); -#if defined(_MSC_VER) && (_MSC_VER <= 1936) /* MSVC 2022 ver 17.6 or earlier */ +#if defined(_MSC_VER) && (_MSC_VER <= 1937) /* MSVC 2022 ver 17.7 or earlier */ # pragma warning(pop) #endif break; @@ -779,7 +799,12 @@ LZ4_FORCE_INLINE U32 LZ4_hash5(U64 sequence, tableType_t const tableType) LZ4_FORCE_INLINE U32 LZ4_hashPosition(const void* const p, tableType_t const tableType) { if ((sizeof(reg_t)==8) && (tableType != byU16)) return LZ4_hash5(LZ4_read_ARCH(p), tableType); + +#ifdef LZ4_STATIC_LINKING_ONLY_ENDIANNESS_INDEPENDENT_OUTPUT + return LZ4_hash4(LZ4_readLE32(p), tableType); +#else return LZ4_hash4(LZ4_read32(p), tableType); +#endif } LZ4_FORCE_INLINE void LZ4_clearHash(U32 h, void* tableBase, tableType_t const tableType) @@ -873,7 +898,7 @@ LZ4_prepareTable(LZ4_stream_t_internal* const cctx, || tableType == byPtr || inputSize >= 4 KB) { - DEBUGLOG(4, "LZ4_prepareTable: Resetting table in %p", cctx); + DEBUGLOG(4, "LZ4_prepareTable: Resetting table in %p", (void*)cctx); MEM_INIT(cctx->hashTable, 0, LZ4_HASHTABLESIZE); cctx->currentOffset = 0; cctx->tableType = (U32)clearedTable; @@ -898,7 +923,7 @@ LZ4_prepareTable(LZ4_stream_t_internal* const cctx, cctx->dictSize = 0; } -/** LZ4_compress_generic() : +/** LZ4_compress_generic_validated() : * inlined, to ensure branches are decided at compilation time. * The following conditions are presumed already validated: * - source != NULL @@ -1080,7 +1105,10 @@ LZ4_FORCE_INLINE int LZ4_compress_generic_validated( /* Catch up */ filledIp = ip; - while (((ip>anchor) & (match > lowLimit)) && (unlikely(ip[-1]==match[-1]))) { ip--; match--; } + assert(ip > anchor); /* this is always true as ip has been advanced before entering the main loop */ + if ((match > lowLimit) && unlikely(ip[-1] == match[-1])) { + do { ip--; match--; } while (((ip > anchor) & (match > lowLimit)) && (unlikely(ip[-1] == match[-1]))); + } /* Encode Literals */ { unsigned const litLength = (unsigned)(ip - anchor); @@ -1095,7 +1123,7 @@ LZ4_FORCE_INLINE int LZ4_compress_generic_validated( goto _last_literals; } if (litLength >= RUN_MASK) { - int len = (int)(litLength - RUN_MASK); + unsigned len = litLength - RUN_MASK; *token = (RUN_MASK<= 255 ; len-=255) *op++ = 255; *op++ = (BYTE)len; @@ -1452,22 +1480,30 @@ int LZ4_compress_default(const char* src, char* dst, int srcSize, int dstCapacit /* Note!: This function leaves the stream in an unclean/broken state! * It is not safe to subsequently use the same state with a _fastReset() or * _continue() call without resetting it. */ -static int LZ4_compress_destSize_extState (LZ4_stream_t* state, const char* src, char* dst, int* srcSizePtr, int targetDstSize) +static int LZ4_compress_destSize_extState_internal(LZ4_stream_t* state, const char* src, char* dst, int* srcSizePtr, int targetDstSize, int acceleration) { void* const s = LZ4_initStream(state, sizeof (*state)); assert(s != NULL); (void)s; if (targetDstSize >= LZ4_compressBound(*srcSizePtr)) { /* compression success is guaranteed */ - return LZ4_compress_fast_extState(state, src, dst, *srcSizePtr, targetDstSize, 1); + return LZ4_compress_fast_extState(state, src, dst, *srcSizePtr, targetDstSize, acceleration); } else { if (*srcSizePtr < LZ4_64Klimit) { - return LZ4_compress_generic(&state->internal_donotuse, src, dst, *srcSizePtr, srcSizePtr, targetDstSize, fillOutput, byU16, noDict, noDictIssue, 1); + return LZ4_compress_generic(&state->internal_donotuse, src, dst, *srcSizePtr, srcSizePtr, targetDstSize, fillOutput, byU16, noDict, noDictIssue, acceleration); } else { tableType_t const addrMode = ((sizeof(void*)==4) && ((uptrval)src > LZ4_DISTANCE_MAX)) ? byPtr : byU32; - return LZ4_compress_generic(&state->internal_donotuse, src, dst, *srcSizePtr, srcSizePtr, targetDstSize, fillOutput, addrMode, noDict, noDictIssue, 1); + return LZ4_compress_generic(&state->internal_donotuse, src, dst, *srcSizePtr, srcSizePtr, targetDstSize, fillOutput, addrMode, noDict, noDictIssue, acceleration); } } } +int LZ4_compress_destSize_extState(void* state, const char* src, char* dst, int* srcSizePtr, int targetDstSize, int acceleration) +{ + int const r = LZ4_compress_destSize_extState_internal((LZ4_stream_t*)state, src, dst, srcSizePtr, targetDstSize, acceleration); + /* clean the state on exit */ + LZ4_initStream(state, sizeof (LZ4_stream_t)); + return r; +} + int LZ4_compress_destSize(const char* src, char* dst, int* srcSizePtr, int targetDstSize) { @@ -1479,7 +1515,7 @@ int LZ4_compress_destSize(const char* src, char* dst, int* srcSizePtr, int targe LZ4_stream_t* const ctx = &ctxBody; #endif - int result = LZ4_compress_destSize_extState(ctx, src, dst, srcSizePtr, targetDstSize); + int result = LZ4_compress_destSize_extState_internal(ctx, src, dst, srcSizePtr, targetDstSize, 1); #if (LZ4_HEAPMODE) FREEMEM(ctx); @@ -1498,7 +1534,7 @@ LZ4_stream_t* LZ4_createStream(void) { LZ4_stream_t* const lz4s = (LZ4_stream_t*)ALLOC(sizeof(LZ4_stream_t)); LZ4_STATIC_ASSERT(sizeof(LZ4_stream_t) >= sizeof(LZ4_stream_t_internal)); - DEBUGLOG(4, "LZ4_createStream %p", lz4s); + DEBUGLOG(4, "LZ4_createStream %p", (void*)lz4s); if (lz4s == NULL) return NULL; LZ4_initStream(lz4s, sizeof(*lz4s)); return lz4s; @@ -1529,7 +1565,7 @@ LZ4_stream_t* LZ4_initStream (void* buffer, size_t size) * prefer initStream() which is more general */ void LZ4_resetStream (LZ4_stream_t* LZ4_stream) { - DEBUGLOG(5, "LZ4_resetStream (ctx:%p)", LZ4_stream); + DEBUGLOG(5, "LZ4_resetStream (ctx:%p)", (void*)LZ4_stream); MEM_INIT(LZ4_stream, 0, sizeof(LZ4_stream_t_internal)); } @@ -1541,15 +1577,18 @@ void LZ4_resetStream_fast(LZ4_stream_t* ctx) { int LZ4_freeStream (LZ4_stream_t* LZ4_stream) { if (!LZ4_stream) return 0; /* support free on NULL */ - DEBUGLOG(5, "LZ4_freeStream %p", LZ4_stream); + DEBUGLOG(5, "LZ4_freeStream %p", (void*)LZ4_stream); FREEMEM(LZ4_stream); return (0); } #endif +typedef enum { _ld_fast, _ld_slow } LoadDict_mode_e; #define HASH_UNIT sizeof(reg_t) -int LZ4_loadDict (LZ4_stream_t* LZ4_dict, const char* dictionary, int dictSize) +int LZ4_loadDict_internal(LZ4_stream_t* LZ4_dict, + const char* dictionary, int dictSize, + LoadDict_mode_e _ld) { LZ4_stream_t_internal* const dict = &LZ4_dict->internal_donotuse; const tableType_t tableType = byU32; @@ -1557,7 +1596,7 @@ int LZ4_loadDict (LZ4_stream_t* LZ4_dict, const char* dictionary, int dictSize) const BYTE* const dictEnd = p + dictSize; U32 idx32; - DEBUGLOG(4, "LZ4_loadDict (%i bytes from %p into %p)", dictSize, dictionary, LZ4_dict); + DEBUGLOG(4, "LZ4_loadDict (%i bytes from %p into %p)", dictSize, (void*)dictionary, (void*)LZ4_dict); /* It's necessary to reset the context, * and not just continue it with prepareTable() @@ -1585,20 +1624,46 @@ int LZ4_loadDict (LZ4_stream_t* LZ4_dict, const char* dictionary, int dictSize) while (p <= dictEnd-HASH_UNIT) { U32 const h = LZ4_hashPosition(p, tableType); + /* Note: overwriting => favors positions end of dictionary */ LZ4_putIndexOnHash(idx32, h, dict->hashTable, tableType); p+=3; idx32+=3; } + if (_ld == _ld_slow) { + /* Fill hash table with additional references, to improve compression capability */ + p = dict->dictionary; + idx32 = dict->currentOffset - dict->dictSize; + while (p <= dictEnd-HASH_UNIT) { + U32 const h = LZ4_hashPosition(p, tableType); + U32 const limit = dict->currentOffset - 64 KB; + if (LZ4_getIndexOnHash(h, dict->hashTable, tableType) <= limit) { + /* Note: not overwriting => favors positions beginning of dictionary */ + LZ4_putIndexOnHash(idx32, h, dict->hashTable, tableType); + } + p++; idx32++; + } + } + return (int)dict->dictSize; } +int LZ4_loadDict(LZ4_stream_t* LZ4_dict, const char* dictionary, int dictSize) +{ + return LZ4_loadDict_internal(LZ4_dict, dictionary, dictSize, _ld_fast); +} + +int LZ4_loadDictSlow(LZ4_stream_t* LZ4_dict, const char* dictionary, int dictSize) +{ + return LZ4_loadDict_internal(LZ4_dict, dictionary, dictSize, _ld_slow); +} + void LZ4_attach_dictionary(LZ4_stream_t* workingStream, const LZ4_stream_t* dictionaryStream) { const LZ4_stream_t_internal* dictCtx = (dictionaryStream == NULL) ? NULL : &(dictionaryStream->internal_donotuse); DEBUGLOG(4, "LZ4_attach_dictionary (%p, %p, size %u)", - workingStream, dictionaryStream, + (void*)workingStream, (void*)dictionaryStream, dictCtx != NULL ? dictCtx->dictSize : 0); if (dictCtx != NULL) { @@ -1662,7 +1727,7 @@ int LZ4_compress_fast_continue (LZ4_stream_t* LZ4_stream, && (inputSize > 0) /* tolerance : don't lose history, in case next invocation would use prefix mode */ && (streamPtr->dictCtx == NULL) /* usingDictCtx */ ) { - DEBUGLOG(5, "LZ4_compress_fast_continue: dictSize(%u) at addr:%p is too small", streamPtr->dictSize, streamPtr->dictionary); + DEBUGLOG(5, "LZ4_compress_fast_continue: dictSize(%u) at addr:%p is too small", streamPtr->dictSize, (void*)streamPtr->dictionary); /* remove dictionary existence from history, to employ faster prefix mode */ streamPtr->dictSize = 0; streamPtr->dictionary = (const BYTE*)source; @@ -1752,7 +1817,7 @@ int LZ4_saveDict (LZ4_stream_t* LZ4_dict, char* safeBuffer, int dictSize) { LZ4_stream_t_internal* const dict = &LZ4_dict->internal_donotuse; - DEBUGLOG(5, "LZ4_saveDict : dictSize=%i, safeBuffer=%p", dictSize, safeBuffer); + DEBUGLOG(5, "LZ4_saveDict : dictSize=%i, safeBuffer=%p", dictSize, (void*)safeBuffer); if ((U32)dictSize > 64 KB) { dictSize = 64 KB; } /* useless to define a dictionary > 64 KB */ if ((U32)dictSize > dict->dictSize) { dictSize = (int)dict->dictSize; } @@ -1923,6 +1988,17 @@ read_variable_length(const BYTE** ip, const BYTE* ilimit, if (initial_check && unlikely((*ip) >= ilimit)) { /* read limit reached */ return rvl_error; } + s = **ip; + (*ip)++; + length += s; + if (unlikely((*ip) > ilimit)) { /* read limit reached */ + return rvl_error; + } + /* accumulator overflow detection (32-bit mode only) */ + if ((sizeof(length) < 8) && unlikely(length > ((Rvl_t)(-1)/2)) ) { + return rvl_error; + } + if (likely(s != 255)) return length; do { s = **ip; (*ip)++; @@ -1931,10 +2007,10 @@ read_variable_length(const BYTE** ip, const BYTE* ilimit, return rvl_error; } /* accumulator overflow detection (32-bit mode only) */ - if ((sizeof(length)<8) && unlikely(length > ((Rvl_t)(-1)/2)) ) { + if ((sizeof(length) < 8) && unlikely(length > ((Rvl_t)(-1)/2)) ) { return rvl_error; } - } while (s==255); + } while (s == 255); return length; } @@ -2000,7 +2076,7 @@ LZ4_decompress_generic( * note : fast loop may show a regression for some client arm chips. */ #if LZ4_FAST_DEC_LOOP if ((oend - op) < FASTLOOP_SAFE_DISTANCE) { - DEBUGLOG(6, "skip fast decode loop"); + DEBUGLOG(6, "move to safe decode loop"); goto safe_decode; } @@ -2012,6 +2088,7 @@ LZ4_decompress_generic( assert(ip < iend); token = *ip++; length = token >> ML_BITS; /* literal length */ + DEBUGLOG(7, "blockPos%6u: litLength token = %u", (unsigned)(op-(BYTE*)dst), (unsigned)length); /* decode literal length */ if (length == RUN_MASK) { @@ -2025,49 +2102,47 @@ LZ4_decompress_generic( if (unlikely((uptrval)(ip)+length<(uptrval)(ip))) { goto _output_error; } /* overflow detection */ /* copy literals */ - cpy = op+length; LZ4_STATIC_ASSERT(MFLIMIT >= WILDCOPYLENGTH); - if ((cpy>oend-32) || (ip+length>iend-32)) { goto safe_literal_copy; } - LZ4_wildCopy32(op, ip, cpy); - ip += length; op = cpy; - } else { - cpy = op+length; - DEBUGLOG(7, "copy %u bytes in a 16-bytes stripe", (unsigned)length); + if ((op+length>oend-32) || (ip+length>iend-32)) { goto safe_literal_copy; } + LZ4_wildCopy32(op, ip, op+length); + ip += length; op += length; + } else if (ip <= iend-(16 + 1/*max lit + offset + nextToken*/)) { /* We don't need to check oend, since we check it once for each loop below */ - if (ip > iend-(16 + 1/*max lit + offset + nextToken*/)) { goto safe_literal_copy; } + DEBUGLOG(7, "copy %u bytes in a 16-bytes stripe", (unsigned)length); /* Literals can only be <= 14, but hope compilers optimize better when copy by a register size */ LZ4_memcpy(op, ip, 16); - ip += length; op = cpy; + ip += length; op += length; + } else { + goto safe_literal_copy; } /* get offset */ offset = LZ4_readLE16(ip); ip+=2; - DEBUGLOG(6, " offset = %zu", offset); + DEBUGLOG(6, "blockPos%6u: offset = %u", (unsigned)(op-(BYTE*)dst), (unsigned)offset); match = op - offset; assert(match <= op); /* overflow check */ /* get matchlength */ length = token & ML_MASK; + DEBUGLOG(7, " match length token = %u (len==%u)", (unsigned)length, (unsigned)length+MINMATCH); if (length == ML_MASK) { size_t const addl = read_variable_length(&ip, iend - LASTLITERALS + 1, 0); if (addl == rvl_error) { - DEBUGLOG(6, "error reading long match length"); + DEBUGLOG(5, "error reading long match length"); goto _output_error; } length += addl; length += MINMATCH; + DEBUGLOG(7, " long match length == %u", (unsigned)length); if (unlikely((uptrval)(op)+length<(uptrval)op)) { goto _output_error; } /* overflow detection */ - if ((checkOffset) && (unlikely(match + dictSize < lowPrefix))) { - DEBUGLOG(6, "Error : offset outside buffers"); - goto _output_error; - } if (op + length >= oend - FASTLOOP_SAFE_DISTANCE) { goto safe_match_copy; } } else { length += MINMATCH; if (op + length >= oend - FASTLOOP_SAFE_DISTANCE) { + DEBUGLOG(7, "moving to safe_match_copy (ml==%u)", (unsigned)length); goto safe_match_copy; } @@ -2086,7 +2161,7 @@ LZ4_decompress_generic( } } } if ( checkOffset && (unlikely(match + dictSize < lowPrefix)) ) { - DEBUGLOG(6, "Error : pos=%zi, offset=%zi => outside buffers", op-lowPrefix, op-match); + DEBUGLOG(5, "Error : pos=%zi, offset=%zi => outside buffers", op-lowPrefix, op-match); goto _output_error; } /* match starting within external dictionary */ @@ -2143,6 +2218,7 @@ LZ4_decompress_generic( assert(ip < iend); token = *ip++; length = token >> ML_BITS; /* literal length */ + DEBUGLOG(7, "blockPos%6u: litLength token = %u", (unsigned)(op-(BYTE*)dst), (unsigned)length); /* A two-stage shortcut for the most common case: * 1) If the literal length is 0..14, and there is enough space, @@ -2163,6 +2239,7 @@ LZ4_decompress_generic( /* The second stage: prepare for match copying, decode full info. * If it doesn't work out, the info won't be wasted. */ length = token & ML_MASK; /* match length */ + DEBUGLOG(7, "blockPos%6u: matchLength token = %u (len=%u)", (unsigned)(op-(BYTE*)dst), (unsigned)length, (unsigned)length + 4); offset = LZ4_readLE16(ip); ip += 2; match = op - offset; assert(match <= op); /* check overflow */ @@ -2194,11 +2271,12 @@ LZ4_decompress_generic( if (unlikely((uptrval)(ip)+length<(uptrval)(ip))) { goto _output_error; } /* overflow detection */ } - /* copy literals */ - cpy = op+length; #if LZ4_FAST_DEC_LOOP safe_literal_copy: #endif + /* copy literals */ + cpy = op+length; + LZ4_STATIC_ASSERT(MFLIMIT >= WILDCOPYLENGTH); if ((cpy>oend-MFLIMIT) || (ip+length>iend-(2+1+LASTLITERALS))) { /* We've either hit the input parsing restriction or the output parsing restriction. @@ -2234,9 +2312,10 @@ LZ4_decompress_generic( * so check that we exactly consume the input and don't overrun the output buffer. */ if ((ip+length != iend) || (cpy > oend)) { - DEBUGLOG(6, "should have been last run of literals") - DEBUGLOG(6, "ip(%p) + length(%i) = %p != iend (%p)", ip, (int)length, ip+length, iend); - DEBUGLOG(6, "or cpy(%p) > oend(%p)", cpy, oend); + DEBUGLOG(5, "should have been last run of literals") + DEBUGLOG(5, "ip(%p) + length(%i) = %p != iend (%p)", (void*)ip, (int)length, (void*)(ip+length), (void*)iend); + DEBUGLOG(5, "or cpy(%p) > (oend-MFLIMIT)(%p)", (void*)cpy, (void*)(oend-MFLIMIT)); + DEBUGLOG(5, "after writing %u bytes / %i bytes available", (unsigned)(op-(BYTE*)dst), outputSize); goto _output_error; } } @@ -2262,6 +2341,7 @@ LZ4_decompress_generic( /* get matchlength */ length = token & ML_MASK; + DEBUGLOG(7, "blockPos%6u: matchLength token = %u", (unsigned)(op-(BYTE*)dst), (unsigned)length); _copy_match: if (length == ML_MASK) { @@ -2351,7 +2431,7 @@ LZ4_decompress_generic( while (op < cpy) { *op++ = *match++; } } else { LZ4_memcpy(op, match, 8); - if (length > 16) { LZ4_wildCopy8(op+8, match+8, cpy); } + if (length > 16) { LZ4_wildCopy8(op+8, match+8, cpy); } } op = cpy; /* wildcopy correction */ } diff --git a/CompressSave/LZ4WrapC/lz4/lz4.h b/CompressSave/LZ4WrapC/lz4/lz4.h index f85b038..5b147b2 100644 --- a/CompressSave/LZ4WrapC/lz4/lz4.h +++ b/CompressSave/LZ4WrapC/lz4/lz4.h @@ -1,7 +1,7 @@ /* * LZ4 - Fast LZ compression algorithm * Header File - * Copyright (C) 2011-2020, Yann Collet. + * Copyright (c) Yann Collet. All rights reserved. BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) @@ -129,8 +129,8 @@ extern "C" { /*------ Version ------*/ #define LZ4_VERSION_MAJOR 1 /* for breaking interface changes */ -#define LZ4_VERSION_MINOR 9 /* for new (non-breaking) interface capabilities */ -#define LZ4_VERSION_RELEASE 4 /* for tweaks, bug-fixes, or development */ +#define LZ4_VERSION_MINOR 10 /* for new (non-breaking) interface capabilities */ +#define LZ4_VERSION_RELEASE 0 /* for tweaks, bug-fixes, or development */ #define LZ4_VERSION_NUMBER (LZ4_VERSION_MAJOR *100*100 + LZ4_VERSION_MINOR *100 + LZ4_VERSION_RELEASE) @@ -144,23 +144,25 @@ LZ4LIB_API const char* LZ4_versionString (void); /**< library version string; /*-************************************ -* Tuning parameter +* Tuning memory usage **************************************/ -#define LZ4_MEMORY_USAGE_MIN 10 -#define LZ4_MEMORY_USAGE_DEFAULT 14 -#define LZ4_MEMORY_USAGE_MAX 20 - /*! * LZ4_MEMORY_USAGE : - * Memory usage formula : N->2^N Bytes (examples : 10 -> 1KB; 12 -> 4KB ; 16 -> 64KB; 20 -> 1MB; ) - * Increasing memory usage improves compression ratio, at the cost of speed. + * Can be selected at compile time, by setting LZ4_MEMORY_USAGE. + * Memory usage formula : N->2^N Bytes (examples : 10 -> 1KB; 12 -> 4KB ; 16 -> 64KB; 20 -> 1MB) + * Increasing memory usage improves compression ratio, generally at the cost of speed. * Reduced memory usage may improve speed at the cost of ratio, thanks to better cache locality. - * Default value is 14, for 16KB, which nicely fits into Intel x86 L1 cache + * Default value is 14, for 16KB, which nicely fits into most L1 caches. */ #ifndef LZ4_MEMORY_USAGE # define LZ4_MEMORY_USAGE LZ4_MEMORY_USAGE_DEFAULT #endif +/* These are absolute limits, they should not be changed by users */ +#define LZ4_MEMORY_USAGE_MIN 10 +#define LZ4_MEMORY_USAGE_DEFAULT 14 +#define LZ4_MEMORY_USAGE_MAX 20 + #if (LZ4_MEMORY_USAGE < LZ4_MEMORY_USAGE_MIN) # error "LZ4_MEMORY_USAGE is too small !" #endif @@ -191,7 +193,7 @@ LZ4LIB_API int LZ4_compress_default(const char* src, char* dst, int srcSize, int /*! LZ4_decompress_safe() : * @compressedSize : is the exact complete size of the compressed block. * @dstCapacity : is the size of destination buffer (which must be already allocated), - * is an upper bound of decompressed size. + * presumed an upper bound of decompressed size. * @return : the number of bytes decompressed into destination buffer (necessarily <= dstCapacity) * If destination buffer is not large enough, decoding will stop and output an error code (negative value). * If the source stream is detected malformed, the function will stop decoding and return a negative result. @@ -243,23 +245,25 @@ LZ4LIB_API int LZ4_compress_fast (const char* src, char* dst, int srcSize, int d LZ4LIB_API int LZ4_sizeofState(void); LZ4LIB_API int LZ4_compress_fast_extState (void* state, const char* src, char* dst, int srcSize, int dstCapacity, int acceleration); - /*! LZ4_compress_destSize() : * Reverse the logic : compresses as much data as possible from 'src' buffer - * into already allocated buffer 'dst', of size >= 'targetDestSize'. + * into already allocated buffer 'dst', of size >= 'dstCapacity'. * This function either compresses the entire 'src' content into 'dst' if it's large enough, * or fill 'dst' buffer completely with as much data as possible from 'src'. * note: acceleration parameter is fixed to "default". * - * *srcSizePtr : will be modified to indicate how many bytes where read from 'src' to fill 'dst'. + * *srcSizePtr : in+out parameter. Initially contains size of input. + * Will be modified to indicate how many bytes where read from 'src' to fill 'dst'. * New value is necessarily <= input value. - * @return : Nb bytes written into 'dst' (necessarily <= targetDestSize) + * @return : Nb bytes written into 'dst' (necessarily <= dstCapacity) * or 0 if compression fails. * - * Note : from v1.8.2 to v1.9.1, this function had a bug (fixed in v1.9.2+): - * the produced compressed content could, in specific circumstances, - * require to be decompressed into a destination buffer larger - * by at least 1 byte than the content to decompress. + * Note : 'targetDstSize' must be >= 1, because it's the smallest valid lz4 payload. + * + * Note 2:from v1.8.2 to v1.9.1, this function had a bug (fixed in v1.9.2+): + * the produced compressed content could, in rare circumstances, + * require to be decompressed into a destination buffer + * larger by at least 1 byte than decompressesSize. * If an application uses `LZ4_compress_destSize()`, * it's highly recommended to update liblz4 to v1.9.2 or better. * If this can't be done or ensured, @@ -267,8 +271,7 @@ LZ4LIB_API int LZ4_compress_fast_extState (void* state, const char* src, char* d * a dstCapacity which is > decompressedSize, by at least 1 byte. * See https://github.com/lz4/lz4/issues/859 for details */ -LZ4LIB_API int LZ4_compress_destSize (const char* src, char* dst, int* srcSizePtr, int targetDstSize); - +LZ4LIB_API int LZ4_compress_destSize(const char* src, char* dst, int* srcSizePtr, int targetDstSize); /*! LZ4_decompress_safe_partial() : * Decompress an LZ4 compressed block, of size 'srcSize' at position 'src', @@ -312,7 +315,7 @@ LZ4LIB_API int LZ4_decompress_safe_partial (const char* src, char* dst, int srcS ***********************************************/ typedef union LZ4_stream_u LZ4_stream_t; /* incomplete type (defined later) */ -/** +/*! Note about RC_INVOKED - RC_INVOKED is predefined symbol of rc.exe (the resource compiler which is part of MSVC/Visual Studio). @@ -362,13 +365,58 @@ LZ4LIB_API void LZ4_resetStream_fast (LZ4_stream_t* streamPtr); * LZ4_loadDict() triggers a reset, so any previous data will be forgotten. * The same dictionary will have to be loaded on decompression side for successful decoding. * Dictionary are useful for better compression of small data (KB range). - * While LZ4 accept any input as dictionary, - * results are generally better when using Zstandard's Dictionary Builder. + * While LZ4 itself accepts any input as dictionary, dictionary efficiency is also a topic. + * When in doubt, employ the Zstandard's Dictionary Builder. * Loading a size of 0 is allowed, and is the same as reset. - * @return : loaded dictionary size, in bytes (necessarily <= 64 KB) + * @return : loaded dictionary size, in bytes (note: only the last 64 KB are loaded) */ LZ4LIB_API int LZ4_loadDict (LZ4_stream_t* streamPtr, const char* dictionary, int dictSize); +/*! LZ4_loadDictSlow() : v1.10.0+ + * Same as LZ4_loadDict(), + * but uses a bit more cpu to reference the dictionary content more thoroughly. + * This is expected to slightly improve compression ratio. + * The extra-cpu cost is likely worth it if the dictionary is re-used across multiple sessions. + * @return : loaded dictionary size, in bytes (note: only the last 64 KB are loaded) + */ +LZ4LIB_API int LZ4_loadDictSlow(LZ4_stream_t* streamPtr, const char* dictionary, int dictSize); + +/*! LZ4_attach_dictionary() : stable since v1.10.0 + * + * This allows efficient re-use of a static dictionary multiple times. + * + * Rather than re-loading the dictionary buffer into a working context before + * each compression, or copying a pre-loaded dictionary's LZ4_stream_t into a + * working LZ4_stream_t, this function introduces a no-copy setup mechanism, + * in which the working stream references @dictionaryStream in-place. + * + * Several assumptions are made about the state of @dictionaryStream. + * Currently, only states which have been prepared by LZ4_loadDict() or + * LZ4_loadDictSlow() should be expected to work. + * + * Alternatively, the provided @dictionaryStream may be NULL, + * in which case any existing dictionary stream is unset. + * + * If a dictionary is provided, it replaces any pre-existing stream history. + * The dictionary contents are the only history that can be referenced and + * logically immediately precede the data compressed in the first subsequent + * compression call. + * + * The dictionary will only remain attached to the working stream through the + * first compression call, at the end of which it is cleared. + * @dictionaryStream stream (and source buffer) must remain in-place / accessible / unchanged + * through the completion of the compression session. + * + * Note: there is no equivalent LZ4_attach_*() method on the decompression side + * because there is no initialization cost, hence no need to share the cost across multiple sessions. + * To decompress LZ4 blocks using dictionary, attached or not, + * just employ the regular LZ4_setStreamDecode() for streaming, + * or the stateless LZ4_decompress_safe_usingDict() for one-shot decompression. + */ +LZ4LIB_API void +LZ4_attach_dictionary(LZ4_stream_t* workingStream, + const LZ4_stream_t* dictionaryStream); + /*! LZ4_compress_fast_continue() : * Compress 'src' content using data from previously compressed blocks, for better compression ratio. * 'dst' buffer must be already allocated. @@ -546,9 +594,9 @@ LZ4_decompress_safe_partial_usingDict(const char* src, char* dst, #define LZ4_STATIC_3504398509 #ifdef LZ4_PUBLISH_STATIC_FUNCTIONS -#define LZ4LIB_STATIC_API LZ4LIB_API +# define LZ4LIB_STATIC_API LZ4LIB_API #else -#define LZ4LIB_STATIC_API +# define LZ4LIB_STATIC_API #endif @@ -564,36 +612,11 @@ LZ4_decompress_safe_partial_usingDict(const char* src, char* dst, */ LZ4LIB_STATIC_API int LZ4_compress_fast_extState_fastReset (void* state, const char* src, char* dst, int srcSize, int dstCapacity, int acceleration); -/*! LZ4_attach_dictionary() : - * This is an experimental API that allows - * efficient use of a static dictionary many times. - * - * Rather than re-loading the dictionary buffer into a working context before - * each compression, or copying a pre-loaded dictionary's LZ4_stream_t into a - * working LZ4_stream_t, this function introduces a no-copy setup mechanism, - * in which the working stream references the dictionary stream in-place. - * - * Several assumptions are made about the state of the dictionary stream. - * Currently, only streams which have been prepared by LZ4_loadDict() should - * be expected to work. - * - * Alternatively, the provided dictionaryStream may be NULL, - * in which case any existing dictionary stream is unset. - * - * If a dictionary is provided, it replaces any pre-existing stream history. - * The dictionary contents are the only history that can be referenced and - * logically immediately precede the data compressed in the first subsequent - * compression call. - * - * The dictionary will only remain attached to the working stream through the - * first compression call, at the end of which it is cleared. The dictionary - * stream (and source buffer) must remain in-place / accessible / unchanged - * through the completion of the first compression call on the stream. +/*! LZ4_compress_destSize_extState() : introduced in v1.10.0 + * Same as LZ4_compress_destSize(), but using an externally allocated state. + * Also: exposes @acceleration */ -LZ4LIB_STATIC_API void -LZ4_attach_dictionary(LZ4_stream_t* workingStream, - const LZ4_stream_t* dictionaryStream); - +int LZ4_compress_destSize_extState(void* state, const char* src, char* dst, int* srcSizePtr, int targetDstSize, int acceleration); /*! In-place compression and decompression * @@ -677,10 +700,10 @@ LZ4_attach_dictionary(LZ4_stream_t* workingStream, #if defined(__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) # include - typedef int8_t LZ4_i8; - typedef uint8_t LZ4_byte; - typedef uint16_t LZ4_u16; - typedef uint32_t LZ4_u32; + typedef int8_t LZ4_i8; + typedef unsigned char LZ4_byte; + typedef uint16_t LZ4_u16; + typedef uint32_t LZ4_u32; #else typedef signed char LZ4_i8; typedef unsigned char LZ4_byte; @@ -705,7 +728,7 @@ struct LZ4_stream_t_internal { /* Implicit padding to ensure structure is aligned */ }; -#define LZ4_STREAM_MINSIZE ((1UL << LZ4_MEMORY_USAGE) + 32) /* static size, for inter-version compatibility */ +#define LZ4_STREAM_MINSIZE ((1UL << (LZ4_MEMORY_USAGE)) + 32) /* static size, for inter-version compatibility */ union LZ4_stream_u { char minStateSize[LZ4_STREAM_MINSIZE]; LZ4_stream_t_internal internal_donotuse; @@ -726,7 +749,7 @@ union LZ4_stream_u { * Note2: An LZ4_stream_t structure guarantees correct alignment and size. * Note3: Before v1.9.0, use LZ4_resetStream() instead **/ -LZ4LIB_API LZ4_stream_t* LZ4_initStream (void* buffer, size_t size); +LZ4LIB_API LZ4_stream_t* LZ4_initStream (void* stateBuffer, size_t size); /*! LZ4_streamDecode_t : @@ -838,11 +861,12 @@ LZ4_DEPRECATED("use LZ4_decompress_fast_usingDict() instead") LZ4LIB_API int LZ4 * But they may happen if input data is invalid (error or intentional tampering). * As a consequence, use these functions in trusted environments with trusted data **only**. */ -LZ4_DEPRECATED("This function is deprecated and unsafe. Consider using LZ4_decompress_safe() instead") +LZ4_DEPRECATED("This function is deprecated and unsafe. Consider using LZ4_decompress_safe_partial() instead") LZ4LIB_API int LZ4_decompress_fast (const char* src, char* dst, int originalSize); -LZ4_DEPRECATED("This function is deprecated and unsafe. Consider using LZ4_decompress_safe_continue() instead") +LZ4_DEPRECATED("This function is deprecated and unsafe. Consider migrating towards LZ4_decompress_safe_continue() instead. " + "Note that the contract will change (requires block's compressed size, instead of decompressed size)") LZ4LIB_API int LZ4_decompress_fast_continue (LZ4_streamDecode_t* LZ4_streamDecode, const char* src, char* dst, int originalSize); -LZ4_DEPRECATED("This function is deprecated and unsafe. Consider using LZ4_decompress_safe_usingDict() instead") +LZ4_DEPRECATED("This function is deprecated and unsafe. Consider using LZ4_decompress_safe_partial_usingDict() instead") LZ4LIB_API int LZ4_decompress_fast_usingDict (const char* src, char* dst, int originalSize, const char* dictStart, int dictSize); /*! LZ4_resetStream() : diff --git a/CompressSave/LZ4WrapC/lz4/lz4frame.c b/CompressSave/LZ4WrapC/lz4/lz4frame.c index 77857a5..6dab767 100644 --- a/CompressSave/LZ4WrapC/lz4/lz4frame.c +++ b/CompressSave/LZ4WrapC/lz4/lz4frame.c @@ -1,6 +1,6 @@ /* * LZ4 auto-framing library - * Copyright (C) 2011-2016, Yann Collet. + * Copyright (c) Yann Collet. All rights reserved. * * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) * @@ -44,6 +44,7 @@ /*-************************************ * Compiler Options **************************************/ +#include #ifdef _MSC_VER /* Visual Studio */ # pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */ #endif @@ -125,8 +126,9 @@ static void* LZ4F_malloc(size_t s, LZ4F_CustomMem cmem) static void LZ4F_free(void* p, LZ4F_CustomMem cmem) { - /* custom malloc defined : use it */ + if (p == NULL) return; if (cmem.customFree != NULL) { + /* custom allocation defined : use it */ cmem.customFree(cmem.opaqueState, p); return; } @@ -153,7 +155,7 @@ static void LZ4F_free(void* p, LZ4F_CustomMem cmem) static int g_debuglog_enable = 1; # define DEBUGLOG(l, ...) { \ if ((g_debuglog_enable) && (l<=LZ4_DEBUG)) { \ - fprintf(stderr, __FILE__ ": "); \ + fprintf(stderr, __FILE__ " %i: ", __LINE__ ); \ fprintf(stderr, __VA_ARGS__); \ fprintf(stderr, " \n"); \ } } @@ -186,9 +188,9 @@ static U32 LZ4F_readLE32 (const void* src) { const BYTE* const srcPtr = (const BYTE*)src; U32 value32 = srcPtr[0]; - value32 += ((U32)srcPtr[1])<< 8; - value32 += ((U32)srcPtr[2])<<16; - value32 += ((U32)srcPtr[3])<<24; + value32 |= ((U32)srcPtr[1])<< 8; + value32 |= ((U32)srcPtr[2])<<16; + value32 |= ((U32)srcPtr[3])<<24; return value32; } @@ -205,13 +207,13 @@ static U64 LZ4F_readLE64 (const void* src) { const BYTE* const srcPtr = (const BYTE*)src; U64 value64 = srcPtr[0]; - value64 += ((U64)srcPtr[1]<<8); - value64 += ((U64)srcPtr[2]<<16); - value64 += ((U64)srcPtr[3]<<24); - value64 += ((U64)srcPtr[4]<<32); - value64 += ((U64)srcPtr[5]<<40); - value64 += ((U64)srcPtr[6]<<48); - value64 += ((U64)srcPtr[7]<<56); + value64 |= ((U64)srcPtr[1]<<8); + value64 |= ((U64)srcPtr[2]<<16); + value64 |= ((U64)srcPtr[3]<<24); + value64 |= ((U64)srcPtr[4]<<32); + value64 |= ((U64)srcPtr[5]<<40); + value64 |= ((U64)srcPtr[6]<<48); + value64 |= ((U64)srcPtr[7]<<56); return value64; } @@ -257,7 +259,8 @@ static const size_t BFSize = LZ4F_BLOCK_CHECKSUM_SIZE; /* block footer : checks * Structures and local types **************************************/ -typedef enum { LZ4B_COMPRESSED, LZ4B_UNCOMPRESSED} LZ4F_blockCompression_t; +typedef enum { LZ4B_COMPRESSED, LZ4B_UNCOMPRESSED} LZ4F_BlockCompressMode_e; +typedef enum { ctxNone, ctxFast, ctxHC } LZ4F_CtxType_e; typedef struct LZ4F_cctx_s { @@ -275,8 +278,8 @@ typedef struct LZ4F_cctx_s XXH32_state_t xxh; void* lz4CtxPtr; U16 lz4CtxAlloc; /* sized for: 0 = none, 1 = lz4 ctx, 2 = lz4hc ctx */ - U16 lz4CtxState; /* in use as: 0 = none, 1 = lz4 ctx, 2 = lz4hc ctx */ - LZ4F_blockCompression_t blockCompression; + U16 lz4CtxType; /* in use as: 0 = none, 1 = lz4 ctx, 2 = lz4hc ctx */ + LZ4F_BlockCompressMode_e blockCompressMode; } LZ4F_cctx_t; @@ -314,7 +317,12 @@ static LZ4F_errorCode_t LZ4F_returnErrorCode(LZ4F_errorCodes code) #define RETURN_ERROR(e) return LZ4F_returnErrorCode(LZ4F_ERROR_ ## e) -#define RETURN_ERROR_IF(c,e) do { if (c) RETURN_ERROR(e); } while (0) +#define RETURN_ERROR_IF(c,e) do { \ + if (c) { \ + DEBUGLOG(3, "Error: " #c); \ + RETURN_ERROR(e); \ + } \ + } while (0) #define FORWARD_IF_ERROR(r) do { if (LZ4F_isError(r)) return (r); } while (0) @@ -429,6 +437,7 @@ size_t LZ4F_compressFrame_usingCDict(LZ4F_cctx* cctx, BYTE* dstPtr = dstStart; BYTE* const dstEnd = dstStart + dstCapacity; + DEBUGLOG(4, "LZ4F_compressFrame_usingCDict (srcSize=%u)", (unsigned)srcSize); if (preferencesPtr!=NULL) prefs = *preferencesPtr; else @@ -494,7 +503,7 @@ size_t LZ4F_compressFrame(void* dstBuffer, size_t dstCapacity, LZ4_initStream(&lz4ctx, sizeof(lz4ctx)); cctxPtr->lz4CtxPtr = &lz4ctx; cctxPtr->lz4CtxAlloc = 1; - cctxPtr->lz4CtxState = 1; + cctxPtr->lz4CtxType = ctxFast; } #endif DEBUGLOG(4, "LZ4F_compressFrame"); @@ -530,27 +539,35 @@ LZ4F_CDict* LZ4F_createCDict_advanced(LZ4F_CustomMem cmem, const void* dictBuffer, size_t dictSize) { const char* dictStart = (const char*)dictBuffer; - LZ4F_CDict* const cdict = (LZ4F_CDict*)LZ4F_malloc(sizeof(*cdict), cmem); + LZ4F_CDict* cdict = NULL; + DEBUGLOG(4, "LZ4F_createCDict_advanced"); - if (!cdict) return NULL; + + if (!dictStart) + return NULL; + cdict = (LZ4F_CDict*)LZ4F_malloc(sizeof(*cdict), cmem); + if (!cdict) + return NULL; + cdict->cmem = cmem; if (dictSize > 64 KB) { dictStart += dictSize - 64 KB; dictSize = 64 KB; } cdict->dictContent = LZ4F_malloc(dictSize, cmem); + /* note: using @cmem to allocate => can't use default create */ cdict->fastCtx = (LZ4_stream_t*)LZ4F_malloc(sizeof(LZ4_stream_t), cmem); - if (cdict->fastCtx) - LZ4_initStream(cdict->fastCtx, sizeof(LZ4_stream_t)); cdict->HCCtx = (LZ4_streamHC_t*)LZ4F_malloc(sizeof(LZ4_streamHC_t), cmem); - if (cdict->HCCtx) - LZ4_initStream(cdict->HCCtx, sizeof(LZ4_streamHC_t)); if (!cdict->dictContent || !cdict->fastCtx || !cdict->HCCtx) { LZ4F_freeCDict(cdict); return NULL; } memcpy(cdict->dictContent, dictStart, dictSize); - LZ4_loadDict (cdict->fastCtx, (const char*)cdict->dictContent, (int)dictSize); + LZ4_initStream(cdict->fastCtx, sizeof(LZ4_stream_t)); + LZ4_loadDictSlow(cdict->fastCtx, (const char*)cdict->dictContent, (int)dictSize); + LZ4_initStreamHC(cdict->HCCtx, sizeof(LZ4_streamHC_t)); + /* note: we don't know at this point which compression level is going to be used + * as a consequence, HCCtx is created for the more common HC mode */ LZ4_setCompressionLevel(cdict->HCCtx, LZ4HC_CLEVEL_DEFAULT); LZ4_loadDictHC(cdict->HCCtx, (const char*)cdict->dictContent, (int)dictSize); return cdict; @@ -616,7 +633,6 @@ LZ4F_createCompressionContext(LZ4F_cctx** LZ4F_compressionContextPtr, unsigned v return LZ4F_OK_NoError; } - LZ4F_errorCode_t LZ4F_freeCompressionContext(LZ4F_cctx* cctxPtr) { if (cctxPtr != NULL) { /* support free on NULL */ @@ -641,7 +657,7 @@ static void LZ4F_initStream(void* ctx, int level, LZ4F_blockMode_t blockMode) { if (level < LZ4HC_CLEVEL_MIN) { - if (cdict != NULL || blockMode == LZ4F_blockLinked) { + if (cdict || blockMode == LZ4F_blockLinked) { /* In these cases, we will call LZ4_compress_fast_continue(), * which needs an already reset context. Otherwise, we'll call a * one-shot API. The non-continued APIs internally perform their own @@ -649,11 +665,18 @@ static void LZ4F_initStream(void* ctx, * tableType they need the context to be in. So in that case this * would be misguided / wasted work. */ LZ4_resetStream_fast((LZ4_stream_t*)ctx); + if (cdict) + LZ4_attach_dictionary((LZ4_stream_t*)ctx, cdict->fastCtx); } - LZ4_attach_dictionary((LZ4_stream_t *)ctx, cdict ? cdict->fastCtx : NULL); + /* In these cases, we'll call a one-shot API. + * The non-continued APIs internally perform their own resets + * at the beginning of their calls, where they know + * which tableType they need the context to be in. + * Therefore, a reset here would be wasted work. */ } else { LZ4_resetStreamHC_fast((LZ4_streamHC_t*)ctx, level); - LZ4_attach_HC_dictionary((LZ4_streamHC_t *)ctx, cdict ? cdict->HCCtx : NULL); + if (cdict) + LZ4_attach_HC_dictionary((LZ4_streamHC_t*)ctx, cdict->HCCtx); } } @@ -668,14 +691,19 @@ static int ctxTypeID_to_size(int ctxTypeID) { } } -/*! LZ4F_compressBegin_usingCDict() : - * init streaming compression AND writes frame header into @dstBuffer. - * @dstCapacity must be >= LZ4F_HEADER_SIZE_MAX bytes. - * @return : number of bytes written into @dstBuffer for the header - * or an error code (can be tested using LZ4F_isError()) +size_t LZ4F_cctx_size(const LZ4F_cctx* cctx) { + if (cctx == NULL) { + return 0; + } + return sizeof(*cctx) + cctx->maxBufferSize + ctxTypeID_to_size(cctx->lz4CtxAlloc); +} + +/* LZ4F_compressBegin_internal() + * Note: only accepts @cdict _or_ @dictBuffer as non NULL. */ -size_t LZ4F_compressBegin_usingCDict(LZ4F_cctx* cctxPtr, +size_t LZ4F_compressBegin_internal(LZ4F_cctx* cctx, void* dstBuffer, size_t dstCapacity, + const void* dictBuffer, size_t dictSize, const LZ4F_CDict* cdict, const LZ4F_preferences_t* preferencesPtr) { @@ -685,70 +713,85 @@ size_t LZ4F_compressBegin_usingCDict(LZ4F_cctx* cctxPtr, RETURN_ERROR_IF(dstCapacity < maxFHSize, dstMaxSize_tooSmall); if (preferencesPtr == NULL) preferencesPtr = &prefNull; - cctxPtr->prefs = *preferencesPtr; + cctx->prefs = *preferencesPtr; + DEBUGLOG(5, "LZ4F_compressBegin_internal: Independent_blocks=%u", cctx->prefs.frameInfo.blockMode); /* cctx Management */ - { U16 const ctxTypeID = (cctxPtr->prefs.compressionLevel < LZ4HC_CLEVEL_MIN) ? 1 : 2; + { U16 const ctxTypeID = (cctx->prefs.compressionLevel < LZ4HC_CLEVEL_MIN) ? 1 : 2; int requiredSize = ctxTypeID_to_size(ctxTypeID); - int allocatedSize = ctxTypeID_to_size(cctxPtr->lz4CtxAlloc); + int allocatedSize = ctxTypeID_to_size(cctx->lz4CtxAlloc); if (allocatedSize < requiredSize) { /* not enough space allocated */ - LZ4F_free(cctxPtr->lz4CtxPtr, cctxPtr->cmem); - if (cctxPtr->prefs.compressionLevel < LZ4HC_CLEVEL_MIN) { + LZ4F_free(cctx->lz4CtxPtr, cctx->cmem); + if (cctx->prefs.compressionLevel < LZ4HC_CLEVEL_MIN) { /* must take ownership of memory allocation, * in order to respect custom allocator contract */ - cctxPtr->lz4CtxPtr = LZ4F_malloc(sizeof(LZ4_stream_t), cctxPtr->cmem); - if (cctxPtr->lz4CtxPtr) - LZ4_initStream(cctxPtr->lz4CtxPtr, sizeof(LZ4_stream_t)); + cctx->lz4CtxPtr = LZ4F_malloc(sizeof(LZ4_stream_t), cctx->cmem); + if (cctx->lz4CtxPtr) + LZ4_initStream(cctx->lz4CtxPtr, sizeof(LZ4_stream_t)); } else { - cctxPtr->lz4CtxPtr = LZ4F_malloc(sizeof(LZ4_streamHC_t), cctxPtr->cmem); - if (cctxPtr->lz4CtxPtr) - LZ4_initStreamHC(cctxPtr->lz4CtxPtr, sizeof(LZ4_streamHC_t)); + cctx->lz4CtxPtr = LZ4F_malloc(sizeof(LZ4_streamHC_t), cctx->cmem); + if (cctx->lz4CtxPtr) + LZ4_initStreamHC(cctx->lz4CtxPtr, sizeof(LZ4_streamHC_t)); } - RETURN_ERROR_IF(cctxPtr->lz4CtxPtr == NULL, allocation_failed); - cctxPtr->lz4CtxAlloc = ctxTypeID; - cctxPtr->lz4CtxState = ctxTypeID; - } else if (cctxPtr->lz4CtxState != ctxTypeID) { + RETURN_ERROR_IF(cctx->lz4CtxPtr == NULL, allocation_failed); + cctx->lz4CtxAlloc = ctxTypeID; + cctx->lz4CtxType = ctxTypeID; + } else if (cctx->lz4CtxType != ctxTypeID) { /* otherwise, a sufficient buffer is already allocated, * but we need to reset it to the correct context type */ - if (cctxPtr->prefs.compressionLevel < LZ4HC_CLEVEL_MIN) { - LZ4_initStream((LZ4_stream_t*)cctxPtr->lz4CtxPtr, sizeof(LZ4_stream_t)); + if (cctx->prefs.compressionLevel < LZ4HC_CLEVEL_MIN) { + LZ4_initStream((LZ4_stream_t*)cctx->lz4CtxPtr, sizeof(LZ4_stream_t)); } else { - LZ4_initStreamHC((LZ4_streamHC_t*)cctxPtr->lz4CtxPtr, sizeof(LZ4_streamHC_t)); - LZ4_setCompressionLevel((LZ4_streamHC_t*)cctxPtr->lz4CtxPtr, cctxPtr->prefs.compressionLevel); + LZ4_initStreamHC((LZ4_streamHC_t*)cctx->lz4CtxPtr, sizeof(LZ4_streamHC_t)); + LZ4_setCompressionLevel((LZ4_streamHC_t*)cctx->lz4CtxPtr, cctx->prefs.compressionLevel); } - cctxPtr->lz4CtxState = ctxTypeID; + cctx->lz4CtxType = ctxTypeID; } } /* Buffer Management */ - if (cctxPtr->prefs.frameInfo.blockSizeID == 0) - cctxPtr->prefs.frameInfo.blockSizeID = LZ4F_BLOCKSIZEID_DEFAULT; - cctxPtr->maxBlockSize = LZ4F_getBlockSize(cctxPtr->prefs.frameInfo.blockSizeID); + if (cctx->prefs.frameInfo.blockSizeID == 0) + cctx->prefs.frameInfo.blockSizeID = LZ4F_BLOCKSIZEID_DEFAULT; + cctx->maxBlockSize = LZ4F_getBlockSize(cctx->prefs.frameInfo.blockSizeID); { size_t const requiredBuffSize = preferencesPtr->autoFlush ? - ((cctxPtr->prefs.frameInfo.blockMode == LZ4F_blockLinked) ? 64 KB : 0) : /* only needs past data up to window size */ - cctxPtr->maxBlockSize + ((cctxPtr->prefs.frameInfo.blockMode == LZ4F_blockLinked) ? 128 KB : 0); + ((cctx->prefs.frameInfo.blockMode == LZ4F_blockLinked) ? 64 KB : 0) : /* only needs past data up to window size */ + cctx->maxBlockSize + ((cctx->prefs.frameInfo.blockMode == LZ4F_blockLinked) ? 128 KB : 0); - if (cctxPtr->maxBufferSize < requiredBuffSize) { - cctxPtr->maxBufferSize = 0; - LZ4F_free(cctxPtr->tmpBuff, cctxPtr->cmem); - cctxPtr->tmpBuff = (BYTE*)LZ4F_malloc(requiredBuffSize, cctxPtr->cmem); - RETURN_ERROR_IF(cctxPtr->tmpBuff == NULL, allocation_failed); - cctxPtr->maxBufferSize = requiredBuffSize; + if (cctx->maxBufferSize < requiredBuffSize) { + cctx->maxBufferSize = 0; + LZ4F_free(cctx->tmpBuff, cctx->cmem); + cctx->tmpBuff = (BYTE*)LZ4F_malloc(requiredBuffSize, cctx->cmem); + RETURN_ERROR_IF(cctx->tmpBuff == NULL, allocation_failed); + cctx->maxBufferSize = requiredBuffSize; } } - cctxPtr->tmpIn = cctxPtr->tmpBuff; - cctxPtr->tmpInSize = 0; - (void)XXH32_reset(&(cctxPtr->xxh), 0); + cctx->tmpIn = cctx->tmpBuff; + cctx->tmpInSize = 0; + (void)XXH32_reset(&(cctx->xxh), 0); /* context init */ - cctxPtr->cdict = cdict; - if (cctxPtr->prefs.frameInfo.blockMode == LZ4F_blockLinked) { + cctx->cdict = cdict; + if (cctx->prefs.frameInfo.blockMode == LZ4F_blockLinked) { /* frame init only for blockLinked : blockIndependent will be init at each block */ - LZ4F_initStream(cctxPtr->lz4CtxPtr, cdict, cctxPtr->prefs.compressionLevel, LZ4F_blockLinked); + LZ4F_initStream(cctx->lz4CtxPtr, cdict, cctx->prefs.compressionLevel, LZ4F_blockLinked); } if (preferencesPtr->compressionLevel >= LZ4HC_CLEVEL_MIN) { - LZ4_favorDecompressionSpeed((LZ4_streamHC_t*)cctxPtr->lz4CtxPtr, (int)preferencesPtr->favorDecSpeed); + LZ4_favorDecompressionSpeed((LZ4_streamHC_t*)cctx->lz4CtxPtr, (int)preferencesPtr->favorDecSpeed); } + if (dictBuffer) { + assert(cdict == NULL); + RETURN_ERROR_IF(dictSize > INT_MAX, parameter_invalid); + if (cctx->lz4CtxType == ctxFast) { + /* lz4 fast*/ + LZ4_loadDict((LZ4_stream_t*)cctx->lz4CtxPtr, (const char*)dictBuffer, (int)dictSize); + } else { + /* lz4hc */ + assert(cctx->lz4CtxType == ctxHC); + LZ4_loadDictHC((LZ4_streamHC_t*)cctx->lz4CtxPtr, (const char*)dictBuffer, (int)dictSize); + } + } + + /* Stage 2 : Write Frame Header */ /* Magic Number */ LZ4F_writeLE32(dstPtr, LZ4F_MAGICNUMBER); @@ -757,22 +800,22 @@ size_t LZ4F_compressBegin_usingCDict(LZ4F_cctx* cctxPtr, /* FLG Byte */ *dstPtr++ = (BYTE)(((1 & _2BITS) << 6) /* Version('01') */ - + ((cctxPtr->prefs.frameInfo.blockMode & _1BIT ) << 5) - + ((cctxPtr->prefs.frameInfo.blockChecksumFlag & _1BIT ) << 4) - + ((unsigned)(cctxPtr->prefs.frameInfo.contentSize > 0) << 3) - + ((cctxPtr->prefs.frameInfo.contentChecksumFlag & _1BIT ) << 2) - + (cctxPtr->prefs.frameInfo.dictID > 0) ); + + ((cctx->prefs.frameInfo.blockMode & _1BIT ) << 5) + + ((cctx->prefs.frameInfo.blockChecksumFlag & _1BIT ) << 4) + + ((unsigned)(cctx->prefs.frameInfo.contentSize > 0) << 3) + + ((cctx->prefs.frameInfo.contentChecksumFlag & _1BIT ) << 2) + + (cctx->prefs.frameInfo.dictID > 0) ); /* BD Byte */ - *dstPtr++ = (BYTE)((cctxPtr->prefs.frameInfo.blockSizeID & _3BITS) << 4); + *dstPtr++ = (BYTE)((cctx->prefs.frameInfo.blockSizeID & _3BITS) << 4); /* Optional Frame content size field */ - if (cctxPtr->prefs.frameInfo.contentSize) { - LZ4F_writeLE64(dstPtr, cctxPtr->prefs.frameInfo.contentSize); + if (cctx->prefs.frameInfo.contentSize) { + LZ4F_writeLE64(dstPtr, cctx->prefs.frameInfo.contentSize); dstPtr += 8; - cctxPtr->totalInSize = 0; + cctx->totalInSize = 0; } /* Optional dictionary ID field */ - if (cctxPtr->prefs.frameInfo.dictID) { - LZ4F_writeLE32(dstPtr, cctxPtr->prefs.frameInfo.dictID); + if (cctx->prefs.frameInfo.dictID) { + LZ4F_writeLE32(dstPtr, cctx->prefs.frameInfo.dictID); dstPtr += 4; } /* Header CRC Byte */ @@ -780,24 +823,54 @@ size_t LZ4F_compressBegin_usingCDict(LZ4F_cctx* cctxPtr, dstPtr++; } - cctxPtr->cStage = 1; /* header written, now request input data block */ + cctx->cStage = 1; /* header written, now request input data block */ return (size_t)(dstPtr - dstStart); } - -/*! LZ4F_compressBegin() : - * init streaming compression AND writes frame header into @dstBuffer. - * @dstCapacity must be >= LZ4F_HEADER_SIZE_MAX bytes. - * @preferencesPtr can be NULL, in which case default parameters are selected. - * @return : number of bytes written into dstBuffer for the header - * or an error code (can be tested using LZ4F_isError()) - */ -size_t LZ4F_compressBegin(LZ4F_cctx* cctxPtr, +size_t LZ4F_compressBegin(LZ4F_cctx* cctx, void* dstBuffer, size_t dstCapacity, const LZ4F_preferences_t* preferencesPtr) { - return LZ4F_compressBegin_usingCDict(cctxPtr, dstBuffer, dstCapacity, - NULL, preferencesPtr); + return LZ4F_compressBegin_internal(cctx, dstBuffer, dstCapacity, + NULL, 0, + NULL, preferencesPtr); +} + +/* LZ4F_compressBegin_usingDictOnce: + * Hidden implementation, + * employed for multi-threaded compression + * when frame defines linked blocks */ +size_t LZ4F_compressBegin_usingDictOnce(LZ4F_cctx* cctx, + void* dstBuffer, size_t dstCapacity, + const void* dict, size_t dictSize, + const LZ4F_preferences_t* preferencesPtr) +{ + return LZ4F_compressBegin_internal(cctx, dstBuffer, dstCapacity, + dict, dictSize, + NULL, preferencesPtr); +} + +size_t LZ4F_compressBegin_usingDict(LZ4F_cctx* cctx, + void* dstBuffer, size_t dstCapacity, + const void* dict, size_t dictSize, + const LZ4F_preferences_t* preferencesPtr) +{ + /* note : incorrect implementation : + * this will only use the dictionary once, + * instead of once *per* block when frames defines independent blocks */ + return LZ4F_compressBegin_usingDictOnce(cctx, dstBuffer, dstCapacity, + dict, dictSize, + preferencesPtr); +} + +size_t LZ4F_compressBegin_usingCDict(LZ4F_cctx* cctx, + void* dstBuffer, size_t dstCapacity, + const LZ4F_CDict* cdict, + const LZ4F_preferences_t* preferencesPtr) +{ + return LZ4F_compressBegin_internal(cctx, dstBuffer, dstCapacity, + NULL, 0, + cdict, preferencesPtr); } @@ -829,10 +902,11 @@ static size_t LZ4F_makeBlock(void* dst, LZ4F_blockChecksum_t crcFlag) { BYTE* const cSizePtr = (BYTE*)dst; + int dstCapacity = (srcSize > 1) ? (int)srcSize - 1 : 1; U32 cSize; assert(compress != NULL); cSize = (U32)compress(lz4ctx, (const char*)src, (char*)(cSizePtr+BHSize), - (int)(srcSize), (int)(srcSize-1), + (int)srcSize, dstCapacity, level, cdict); if (cSize == 0 || cSize >= srcSize) { @@ -891,9 +965,10 @@ static int LZ4F_doNotCompressBlock(void* ctx, const char* src, char* dst, int sr return 0; } -static compressFunc_t LZ4F_selectCompression(LZ4F_blockMode_t blockMode, int level, LZ4F_blockCompression_t compressMode) +static compressFunc_t LZ4F_selectCompression(LZ4F_blockMode_t blockMode, int level, LZ4F_BlockCompressMode_e compressMode) { - if (compressMode == LZ4B_UNCOMPRESSED) return LZ4F_doNotCompressBlock; + if (compressMode == LZ4B_UNCOMPRESSED) + return LZ4F_doNotCompressBlock; if (level < LZ4HC_CLEVEL_MIN) { if (blockMode == LZ4F_blockIndependent) return LZ4F_compressBlock; return LZ4F_compressBlock_continue; @@ -902,12 +977,13 @@ static compressFunc_t LZ4F_selectCompression(LZ4F_blockMode_t blockMode, int lev return LZ4F_compressBlockHC_continue; } -/* Save history (up to 64KB) into @tmpBuff */ -static int LZ4F_localSaveDict(LZ4F_cctx_t* cctxPtr) +/* Save or shorten history (up to 64KB) into @tmpBuff */ +static void LZ4F_localSaveDict(LZ4F_cctx_t* cctxPtr) { - if (cctxPtr->prefs.compressionLevel < LZ4HC_CLEVEL_MIN) - return LZ4_saveDict ((LZ4_stream_t*)(cctxPtr->lz4CtxPtr), (char*)(cctxPtr->tmpBuff), 64 KB); - return LZ4_saveDictHC ((LZ4_streamHC_t*)(cctxPtr->lz4CtxPtr), (char*)(cctxPtr->tmpBuff), 64 KB); + int const dictSize = (cctxPtr->prefs.compressionLevel < LZ4HC_CLEVEL_MIN) ? + LZ4_saveDict ((LZ4_stream_t*)(cctxPtr->lz4CtxPtr), (char*)(cctxPtr->tmpBuff), 64 KB) : + LZ4_saveDictHC ((LZ4_streamHC_t*)(cctxPtr->lz4CtxPtr), (char*)(cctxPtr->tmpBuff), 64 KB); + cctxPtr->tmpIn = cctxPtr->tmpBuff + dictSize; } typedef enum { notDone, fromTmpBuffer, fromSrcBuffer } LZ4F_lastBlockStatus; @@ -915,7 +991,7 @@ typedef enum { notDone, fromTmpBuffer, fromSrcBuffer } LZ4F_lastBlockStatus; static const LZ4F_compressOptions_t k_cOptionsNull = { 0, { 0, 0, 0 } }; - /*! LZ4F_compressUpdateImpl() : +/*! LZ4F_compressUpdateImpl() : * LZ4F_compressUpdate() can be called repetitively to compress as much data as necessary. * When successful, the function always entirely consumes @srcBuffer. * src data is either buffered or compressed into @dstBuffer. @@ -931,11 +1007,11 @@ static size_t LZ4F_compressUpdateImpl(LZ4F_cctx* cctxPtr, void* dstBuffer, size_t dstCapacity, const void* srcBuffer, size_t srcSize, const LZ4F_compressOptions_t* compressOptionsPtr, - LZ4F_blockCompression_t blockCompression) + LZ4F_BlockCompressMode_e blockCompression) { size_t const blockSize = cctxPtr->maxBlockSize; const BYTE* srcPtr = (const BYTE*)srcBuffer; - const BYTE* const srcEnd = srcPtr + srcSize; + const BYTE* const srcEnd = srcSize ? (assert(srcPtr!=NULL), srcPtr + srcSize) : srcPtr; BYTE* const dstStart = (BYTE*)dstBuffer; BYTE* dstPtr = dstStart; LZ4F_lastBlockStatus lastBlockCompressed = notDone; @@ -951,10 +1027,10 @@ static size_t LZ4F_compressUpdateImpl(LZ4F_cctx* cctxPtr, RETURN_ERROR(dstMaxSize_tooSmall); /* flush currently written block, to continue with new block compression */ - if (cctxPtr->blockCompression != blockCompression) { + if (cctxPtr->blockCompressMode != blockCompression) { bytesWritten = LZ4F_flush(cctxPtr, dstBuffer, dstCapacity, compressOptionsPtr); dstPtr += bytesWritten; - cctxPtr->blockCompression = blockCompression; + cctxPtr->blockCompressMode = blockCompression; } if (compressOptionsPtr == NULL) compressOptionsPtr = &k_cOptionsNull; @@ -1013,9 +1089,7 @@ static size_t LZ4F_compressUpdateImpl(LZ4F_cctx* cctxPtr, if (compressOptionsPtr->stableSrc) { cctxPtr->tmpIn = cctxPtr->tmpBuff; /* src is stable : dictionary remains in src across invocations */ } else { - int const realDictSize = LZ4F_localSaveDict(cctxPtr); - assert(0 <= realDictSize && realDictSize <= 64 KB); - cctxPtr->tmpIn = cctxPtr->tmpBuff + realDictSize; + LZ4F_localSaveDict(cctxPtr); } } @@ -1025,8 +1099,7 @@ static size_t LZ4F_compressUpdateImpl(LZ4F_cctx* cctxPtr, { /* only preserve 64KB within internal buffer. Ensures there is enough room for next block. * note: this situation necessarily implies lastBlockCompressed==fromTmpBuffer */ - int const realDictSize = LZ4F_localSaveDict(cctxPtr); - cctxPtr->tmpIn = cctxPtr->tmpBuff + realDictSize; + LZ4F_localSaveDict(cctxPtr); assert((cctxPtr->tmpIn + blockSize) <= (cctxPtr->tmpBuff + cctxPtr->maxBufferSize)); } @@ -1068,13 +1141,9 @@ size_t LZ4F_compressUpdate(LZ4F_cctx* cctxPtr, compressOptionsPtr, LZ4B_COMPRESSED); } -/*! LZ4F_compressUpdate() : - * LZ4F_compressUpdate() can be called repetitively to compress as much data as necessary. - * When successful, the function always entirely consumes @srcBuffer. - * src data is either buffered or compressed into @dstBuffer. - * If previously an uncompressed block was written, buffered data is flushed - * before appending compressed data is continued. - * This is only supported when LZ4F_blockIndependent is used +/*! LZ4F_uncompressedUpdate() : + * Same as LZ4F_compressUpdate(), but requests blocks to be sent uncompressed. + * This symbol is only supported when LZ4F_blockIndependent is used * @dstCapacity MUST be >= LZ4F_compressBound(srcSize, preferencesPtr). * @compressOptionsPtr is optional : provide NULL to mean "default". * @return : the number of bytes written into dstBuffer. It can be zero, meaning input data was just buffered. @@ -1084,7 +1153,8 @@ size_t LZ4F_compressUpdate(LZ4F_cctx* cctxPtr, size_t LZ4F_uncompressedUpdate(LZ4F_cctx* cctxPtr, void* dstBuffer, size_t dstCapacity, const void* srcBuffer, size_t srcSize, - const LZ4F_compressOptions_t* compressOptionsPtr) { + const LZ4F_compressOptions_t* compressOptionsPtr) +{ return LZ4F_compressUpdateImpl(cctxPtr, dstBuffer, dstCapacity, srcBuffer, srcSize, @@ -1094,7 +1164,7 @@ size_t LZ4F_uncompressedUpdate(LZ4F_cctx* cctxPtr, /*! LZ4F_flush() : * When compressed data must be sent immediately, without waiting for a block to be filled, - * invoke LZ4_flush(), which will immediately compress any remaining data stored within LZ4F_cctx. + * invoke LZ4F_flush(), which will immediately compress any remaining data stored within LZ4F_cctx. * The result of the function is the number of bytes written into dstBuffer. * It can be zero, this means there was no data left within LZ4F_cctx. * The function outputs an error code if it fails (can be tested using LZ4F_isError()) @@ -1108,13 +1178,15 @@ size_t LZ4F_flush(LZ4F_cctx* cctxPtr, BYTE* dstPtr = dstStart; compressFunc_t compress; + DEBUGLOG(5, "LZ4F_flush: %zu buffered bytes (saved dict size = %i) (dstCapacity=%u)", + cctxPtr->tmpInSize, (int)(cctxPtr->tmpIn - cctxPtr->tmpBuff), (unsigned)dstCapacity); if (cctxPtr->tmpInSize == 0) return 0; /* nothing to flush */ RETURN_ERROR_IF(cctxPtr->cStage != 1, compressionState_uninitialized); RETURN_ERROR_IF(dstCapacity < (cctxPtr->tmpInSize + BHSize + BFSize), dstMaxSize_tooSmall); (void)compressOptionsPtr; /* not useful (yet) */ /* select compression function */ - compress = LZ4F_selectCompression(cctxPtr->prefs.frameInfo.blockMode, cctxPtr->prefs.compressionLevel, cctxPtr->blockCompression); + compress = LZ4F_selectCompression(cctxPtr->prefs.frameInfo.blockMode, cctxPtr->prefs.compressionLevel, cctxPtr->blockCompressMode); /* compress tmp buffer */ dstPtr += LZ4F_makeBlock(dstPtr, @@ -1129,9 +1201,9 @@ size_t LZ4F_flush(LZ4F_cctx* cctxPtr, cctxPtr->tmpInSize = 0; /* keep tmpIn within limits */ - if ((cctxPtr->tmpIn + cctxPtr->maxBlockSize) > (cctxPtr->tmpBuff + cctxPtr->maxBufferSize)) { /* necessarily LZ4F_blockLinked */ - int const realDictSize = LZ4F_localSaveDict(cctxPtr); - cctxPtr->tmpIn = cctxPtr->tmpBuff + realDictSize; + if ((cctxPtr->tmpIn + cctxPtr->maxBlockSize) > (cctxPtr->tmpBuff + cctxPtr->maxBufferSize)) { + assert(cctxPtr->prefs.frameInfo.blockMode == LZ4F_blockLinked); + LZ4F_localSaveDict(cctxPtr); } return (size_t)(dstPtr - dstStart); @@ -1169,7 +1241,7 @@ size_t LZ4F_compressEnd(LZ4F_cctx* cctxPtr, if (cctxPtr->prefs.frameInfo.contentChecksumFlag == LZ4F_contentChecksumEnabled) { U32 const xxh = XXH32_digest(&(cctxPtr->xxh)); RETURN_ERROR_IF(dstCapacity < 8, dstMaxSize_tooSmall); - DEBUGLOG(5,"Writing 32-bit content checksum"); + DEBUGLOG(5,"Writing 32-bit content checksum (0x%0X)", xxh); LZ4F_writeLE32(dstPtr, xxh); dstPtr+=4; /* content Checksum */ } @@ -1266,15 +1338,25 @@ LZ4F_errorCode_t LZ4F_freeDecompressionContext(LZ4F_dctx* dctx) return result; } +size_t LZ4F_dctx_size(const LZ4F_dctx* dctx) { + if (dctx == NULL) { + return 0; + } + return sizeof(*dctx) + + (dctx->tmpIn != NULL ? dctx->maxBlockSize + BFSize : 0) + + (dctx->tmpOutBuffer != NULL ? dctx->maxBufferSize : 0); +} + /*==--- Streaming Decompression operations ---==*/ - void LZ4F_resetDecompressionContext(LZ4F_dctx* dctx) { + DEBUGLOG(5, "LZ4F_resetDecompressionContext"); dctx->dStage = dstage_getFrameHeader; dctx->dict = NULL; dctx->dictSize = 0; dctx->skipChecksum = 0; + dctx->frameRemainingSize = 0; } @@ -1331,6 +1413,7 @@ static size_t LZ4F_decodeHeader(LZ4F_dctx* dctx, const void* src, size_t srcSize if (((FLG>>1)&_1BIT) != 0) RETURN_ERROR(reservedFlag_set); /* Reserved bit */ if (version != 1) RETURN_ERROR(headerVersion_wrong); /* Version Number, only supported value */ } + DEBUGLOG(6, "contentSizeFlag: %u", contentSizeFlag); /* Frame Header Size */ frameHeaderSize = minFHSize + (contentSizeFlag?8:0) + (dictIDFlag?4:0); @@ -1367,8 +1450,9 @@ static size_t LZ4F_decodeHeader(LZ4F_dctx* dctx, const void* src, size_t srcSize dctx->frameInfo.contentChecksumFlag = (LZ4F_contentChecksum_t)contentChecksumFlag; dctx->frameInfo.blockSizeID = (LZ4F_blockSizeID_t)blockSizeID; dctx->maxBlockSize = LZ4F_getBlockSize((LZ4F_blockSizeID_t)blockSizeID); - if (contentSizeFlag) + if (contentSizeFlag) { dctx->frameRemainingSize = dctx->frameInfo.contentSize = LZ4F_readLE64(srcPtr+6); + } if (dictIDFlag) dctx->frameInfo.dictID = LZ4F_readLE32(srcPtr + frameHeaderSize - 5); @@ -1427,6 +1511,10 @@ LZ4F_errorCode_t LZ4F_getFrameInfo(LZ4F_dctx* dctx, LZ4F_frameInfo_t* frameInfoPtr, const void* srcBuffer, size_t* srcSizePtr) { + assert(dctx != NULL); + RETURN_ERROR_IF(frameInfoPtr == NULL, parameter_null); + RETURN_ERROR_IF(srcSizePtr == NULL, parameter_null); + LZ4F_STATIC_ASSERT(dstage_getFrameHeader < dstage_storeFrameHeader); if (dctx->dStage > dstage_storeFrameHeader) { /* frameInfo already decoded */ @@ -1568,7 +1656,7 @@ size_t LZ4F_decompress(LZ4F_dctx* dctx, size_t nextSrcSizeHint = 1; - DEBUGLOG(5, "LZ4F_decompress : %p,%u => %p,%u", + DEBUGLOG(5, "LZ4F_decompress: src[%p](%u) => dst[%p](%u)", srcBuffer, (unsigned)*srcSizePtr, dstBuffer, (unsigned)*dstSizePtr); if (dstBuffer == NULL) assert(*dstSizePtr == 0); MEM_INIT(&optionsNull, 0, sizeof(optionsNull)); @@ -1957,6 +2045,7 @@ size_t LZ4F_decompress(LZ4F_dctx* dctx, if (!dctx->skipChecksum) { U32 const readCRC = LZ4F_readLE32(selectedIn); U32 const resultCRC = XXH32_digest(&(dctx->xxh)); + DEBUGLOG(4, "frame checksum: stored 0x%0X vs 0x%0X processed", readCRC, resultCRC); #ifndef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION RETURN_ERROR_IF(readCRC != resultCRC, contentChecksum_invalid); #else diff --git a/CompressSave/LZ4WrapC/lz4/lz4frame.h b/CompressSave/LZ4WrapC/lz4/lz4frame.h index e8ce66d..18e183d 100644 --- a/CompressSave/LZ4WrapC/lz4/lz4frame.h +++ b/CompressSave/LZ4WrapC/lz4/lz4frame.h @@ -1,7 +1,8 @@ /* LZ4F - LZ4-Frame library Header File - Copyright (C) 2011-2020, Yann Collet. + Copyright (c) Yann Collet. All rights reserved. + BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) Redistribution and use in source and binary forms, with or without @@ -173,16 +174,16 @@ typedef LZ4F_contentChecksum_t contentChecksum_t; * setting all parameters to default. * It's then possible to update selectively some parameters */ typedef struct { - LZ4F_blockSizeID_t blockSizeID; /* max64KB, max256KB, max1MB, max4MB; 0 == default */ - LZ4F_blockMode_t blockMode; /* LZ4F_blockLinked, LZ4F_blockIndependent; 0 == default */ - LZ4F_contentChecksum_t contentChecksumFlag; /* 1: frame terminated with 32-bit checksum of decompressed data; 0: disabled (default) */ + LZ4F_blockSizeID_t blockSizeID; /* max64KB, max256KB, max1MB, max4MB; 0 == default (LZ4F_max64KB) */ + LZ4F_blockMode_t blockMode; /* LZ4F_blockLinked, LZ4F_blockIndependent; 0 == default (LZ4F_blockLinked) */ + LZ4F_contentChecksum_t contentChecksumFlag; /* 1: add a 32-bit checksum of frame's decompressed data; 0 == default (disabled) */ LZ4F_frameType_t frameType; /* read-only field : LZ4F_frame or LZ4F_skippableFrame */ unsigned long long contentSize; /* Size of uncompressed content ; 0 == unknown */ unsigned dictID; /* Dictionary ID, sent by compressor to help decoder select correct dictionary; 0 == no dictID provided */ - LZ4F_blockChecksum_t blockChecksumFlag; /* 1: each block followed by a checksum of block's compressed data; 0: disabled (default) */ + LZ4F_blockChecksum_t blockChecksumFlag; /* 1: each block followed by a checksum of block's compressed data; 0 == default (disabled) */ } LZ4F_frameInfo_t; -#define LZ4F_INIT_FRAMEINFO { LZ4F_default, LZ4F_blockLinked, LZ4F_noContentChecksum, LZ4F_frame, 0ULL, 0U, LZ4F_noBlockChecksum } /* v1.8.3+ */ +#define LZ4F_INIT_FRAMEINFO { LZ4F_max64KB, LZ4F_blockLinked, LZ4F_noContentChecksum, LZ4F_frame, 0ULL, 0U, LZ4F_noBlockChecksum } /* v1.8.3+ */ /*! LZ4F_preferences_t : * makes it possible to supply advanced compression instructions to streaming interface. @@ -204,16 +205,6 @@ typedef struct { * Simple compression function ***********************************/ -LZ4FLIB_API int LZ4F_compressionLevel_max(void); /* v1.8.0+ */ - -/*! LZ4F_compressFrameBound() : - * Returns the maximum possible compressed size with LZ4F_compressFrame() given srcSize and preferences. - * `preferencesPtr` is optional. It can be replaced by NULL, in which case, the function will assume default preferences. - * Note : this result is only usable with LZ4F_compressFrame(). - * It may also be relevant to LZ4F_compressUpdate() _only if_ no flush() operation is ever performed. - */ -LZ4FLIB_API size_t LZ4F_compressFrameBound(size_t srcSize, const LZ4F_preferences_t* preferencesPtr); - /*! LZ4F_compressFrame() : * Compress srcBuffer content into an LZ4-compressed frame. * It's a one shot operation, all input content is consumed, and all output is generated. @@ -235,6 +226,20 @@ LZ4FLIB_API size_t LZ4F_compressFrame(void* dstBuffer, size_t dstCapacity, const void* srcBuffer, size_t srcSize, const LZ4F_preferences_t* preferencesPtr); +/*! LZ4F_compressFrameBound() : + * Returns the maximum possible compressed size with LZ4F_compressFrame() given srcSize and preferences. + * `preferencesPtr` is optional. It can be replaced by NULL, in which case, the function will assume default preferences. + * Note : this result is only usable with LZ4F_compressFrame(). + * It may also be relevant to LZ4F_compressUpdate() _only if_ no flush() operation is ever performed. + */ +LZ4FLIB_API size_t LZ4F_compressFrameBound(size_t srcSize, const LZ4F_preferences_t* preferencesPtr); + + +/*! LZ4F_compressionLevel_max() : + * @return maximum allowed compression level (currently: 12) + */ +LZ4FLIB_API int LZ4F_compressionLevel_max(void); /* v1.8.0+ */ + /*-*********************************** * Advanced compression functions @@ -285,6 +290,9 @@ LZ4FLIB_API LZ4F_errorCode_t LZ4F_freeCompressionContext(LZ4F_cctx* cctx); /* Size in bytes of the content checksum. */ #define LZ4F_CONTENT_CHECKSUM_SIZE 4 +/* Size in bytes of the endmark. */ +#define LZ4F_ENDMARK_SIZE 4 + /*! LZ4F_compressBegin() : * will write the frame header into dstBuffer. * dstCapacity must be >= LZ4F_HEADER_SIZE_MAX bytes. @@ -365,8 +373,9 @@ typedef struct LZ4F_dctx_s LZ4F_dctx; /* incomplete type */ typedef LZ4F_dctx* LZ4F_decompressionContext_t; /* compatibility with previous API versions */ typedef struct { - unsigned stableDst; /* pledges that last 64KB decompressed data will remain available unmodified between invocations. - * This optimization skips storage operations in tmp buffers. */ + unsigned stableDst; /* pledges that last 64KB decompressed data is present right before @dstBuffer pointer. + * This optimization skips internal storage operations. + * Once set, this pledge must remain valid up to the end of current frame. */ unsigned skipChecksums; /* disable checksum calculation and verification, even when one is present in frame, to save CPU time. * Setting this option to 1 once disables all checksums for the rest of the frame. */ unsigned reserved1; /* must be set to zero for forward compatibility */ @@ -508,6 +517,109 @@ LZ4F_decompress(LZ4F_dctx* dctx, LZ4FLIB_API void LZ4F_resetDecompressionContext(LZ4F_dctx* dctx); /* always successful */ +/********************************** + * Dictionary compression API + *********************************/ + +/* A Dictionary is useful for the compression of small messages (KB range). + * It dramatically improves compression efficiency. + * + * LZ4 can ingest any input as dictionary, though only the last 64 KB are useful. + * Better results are generally achieved by using Zstandard's Dictionary Builder + * to generate a high-quality dictionary from a set of samples. + * + * The same dictionary will have to be used on the decompression side + * for decoding to be successful. + * To help identify the correct dictionary at decoding stage, + * the frame header allows optional embedding of a dictID field. + */ + +/*! LZ4F_compressBegin_usingDict() : stable since v1.10 + * Inits dictionary compression streaming, and writes the frame header into dstBuffer. + * @dstCapacity must be >= LZ4F_HEADER_SIZE_MAX bytes. + * @prefsPtr is optional : one may provide NULL as argument, + * however, it's the only way to provide dictID in the frame header. + * @dictBuffer must outlive the compression session. + * @return : number of bytes written into dstBuffer for the header, + * or an error code (which can be tested using LZ4F_isError()) + * NOTE: The LZ4Frame spec allows each independent block to be compressed with the dictionary, + * but this entry supports a more limited scenario, where only the first block uses the dictionary. + * This is still useful for small data, which only need one block anyway. + * For larger inputs, one may be more interested in LZ4F_compressFrame_usingCDict() below. + */ +LZ4FLIB_API size_t +LZ4F_compressBegin_usingDict(LZ4F_cctx* cctx, + void* dstBuffer, size_t dstCapacity, + const void* dictBuffer, size_t dictSize, + const LZ4F_preferences_t* prefsPtr); + +/*! LZ4F_decompress_usingDict() : stable since v1.10 + * Same as LZ4F_decompress(), using a predefined dictionary. + * Dictionary is used "in place", without any preprocessing. +** It must remain accessible throughout the entire frame decoding. */ +LZ4FLIB_API size_t +LZ4F_decompress_usingDict(LZ4F_dctx* dctxPtr, + void* dstBuffer, size_t* dstSizePtr, + const void* srcBuffer, size_t* srcSizePtr, + const void* dict, size_t dictSize, + const LZ4F_decompressOptions_t* decompressOptionsPtr); + +/***************************************** + * Bulk processing dictionary compression + *****************************************/ + +/* Loading a dictionary has a cost, since it involves construction of tables. + * The Bulk processing dictionary API makes it possible to share this cost + * over an arbitrary number of compression jobs, even concurrently, + * markedly improving compression latency for these cases. + * + * Note that there is no corresponding bulk API for the decompression side, + * because dictionary does not carry any initialization cost for decompression. + * Use the regular LZ4F_decompress_usingDict() there. + */ +typedef struct LZ4F_CDict_s LZ4F_CDict; + +/*! LZ4_createCDict() : stable since v1.10 + * When compressing multiple messages / blocks using the same dictionary, it's recommended to initialize it just once. + * LZ4_createCDict() will create a digested dictionary, ready to start future compression operations without startup delay. + * LZ4_CDict can be created once and shared by multiple threads concurrently, since its usage is read-only. + * @dictBuffer can be released after LZ4_CDict creation, since its content is copied within CDict. */ +LZ4FLIB_API LZ4F_CDict* LZ4F_createCDict(const void* dictBuffer, size_t dictSize); +LZ4FLIB_API void LZ4F_freeCDict(LZ4F_CDict* CDict); + +/*! LZ4_compressFrame_usingCDict() : stable since v1.10 + * Compress an entire srcBuffer into a valid LZ4 frame using a digested Dictionary. + * @cctx must point to a context created by LZ4F_createCompressionContext(). + * If @cdict==NULL, compress without a dictionary. + * @dstBuffer MUST be >= LZ4F_compressFrameBound(srcSize, preferencesPtr). + * If this condition is not respected, function will fail (@return an errorCode). + * The LZ4F_preferences_t structure is optional : one may provide NULL as argument, + * but it's not recommended, as it's the only way to provide @dictID in the frame header. + * @return : number of bytes written into dstBuffer. + * or an error code if it fails (can be tested using LZ4F_isError()) + * Note: for larger inputs generating multiple independent blocks, + * this entry point uses the dictionary for each block. */ +LZ4FLIB_API size_t +LZ4F_compressFrame_usingCDict(LZ4F_cctx* cctx, + void* dst, size_t dstCapacity, + const void* src, size_t srcSize, + const LZ4F_CDict* cdict, + const LZ4F_preferences_t* preferencesPtr); + +/*! LZ4F_compressBegin_usingCDict() : stable since v1.10 + * Inits streaming dictionary compression, and writes the frame header into dstBuffer. + * @dstCapacity must be >= LZ4F_HEADER_SIZE_MAX bytes. + * @prefsPtr is optional : one may provide NULL as argument, + * note however that it's the only way to insert a @dictID in the frame header. + * @cdict must outlive the compression session. + * @return : number of bytes written into dstBuffer for the header, + * or an error code, which can be tested using LZ4F_isError(). */ +LZ4FLIB_API size_t +LZ4F_compressBegin_usingCDict(LZ4F_cctx* cctx, + void* dstBuffer, size_t dstCapacity, + const LZ4F_CDict* cdict, + const LZ4F_preferences_t* prefsPtr); + #if defined (__cplusplus) } @@ -518,11 +630,8 @@ LZ4FLIB_API void LZ4F_resetDecompressionContext(LZ4F_dctx* dctx); /* always su #if defined(LZ4F_STATIC_LINKING_ONLY) && !defined(LZ4F_H_STATIC_09782039843) #define LZ4F_H_STATIC_09782039843 -#if defined (__cplusplus) -extern "C" { -#endif - -/* These declarations are not stable and may change in the future. +/* Note : + * The below declarations are not stable and may change in the future. * They are therefore only safe to depend on * when the caller is statically linked against the library. * To access their declarations, define LZ4F_STATIC_LINKING_ONLY. @@ -532,6 +641,11 @@ extern "C" { * by defining LZ4F_PUBLISH_STATIC_FUNCTIONS. * Use at your own risk. */ + +#if defined (__cplusplus) +extern "C" { +#endif + #ifdef LZ4F_PUBLISH_STATIC_FUNCTIONS # define LZ4FLIB_STATIC_API LZ4FLIB_API #else @@ -545,7 +659,7 @@ extern "C" { ITEM(ERROR_GENERIC) \ ITEM(ERROR_maxBlockSize_invalid) \ ITEM(ERROR_blockMode_invalid) \ - ITEM(ERROR_contentChecksumFlag_invalid) \ + ITEM(ERROR_parameter_invalid) \ ITEM(ERROR_compressionLevel_invalid) \ ITEM(ERROR_headerVersion_wrong) \ ITEM(ERROR_blockChecksum_invalid) \ @@ -563,6 +677,8 @@ extern "C" { ITEM(ERROR_frameDecoding_alreadyStarted) \ ITEM(ERROR_compressionState_uninitialized) \ ITEM(ERROR_parameter_null) \ + ITEM(ERROR_io_write) \ + ITEM(ERROR_io_read) \ ITEM(ERROR_maxCode) #define LZ4F_GENERATE_ENUM(ENUM) LZ4F_##ENUM, @@ -573,22 +689,26 @@ typedef enum { LZ4F_LIST_ERRORS(LZ4F_GENERATE_ENUM) LZ4FLIB_STATIC_API LZ4F_errorCodes LZ4F_getErrorCode(size_t functionResult); +/********************************** + * Advanced compression operations + *********************************/ /*! LZ4F_getBlockSize() : - * Return, in scalar format (size_t), - * the maximum block size associated with blockSizeID. + * @return, in scalar format (size_t), + * the maximum block size associated with @blockSizeID, + * or an error code (can be tested using LZ4F_isError()) if @blockSizeID is invalid. **/ LZ4FLIB_STATIC_API size_t LZ4F_getBlockSize(LZ4F_blockSizeID_t blockSizeID); /*! LZ4F_uncompressedUpdate() : - * LZ4F_uncompressedUpdate() can be called repetitively to add as much data uncompressed data as necessary. + * LZ4F_uncompressedUpdate() can be called repetitively to add data stored as uncompressed blocks. * Important rule: dstCapacity MUST be large enough to store the entire source buffer as * no compression is done for this operation * If this condition is not respected, LZ4F_uncompressedUpdate() will fail (result is an errorCode). * After an error, the state is left in a UB state, and must be re-initialized or freed. - * If previously a compressed block was written, buffered data is flushed + * If previously a compressed block was written, buffered data is flushed first, * before appending uncompressed data is continued. - * This is only supported when LZ4F_blockIndependent is used + * This operation is only supported when LZ4F_blockIndependent is used. * `cOptPtr` is optional : NULL can be provided, in which case all options are set to default. * @return : number of bytes written into `dstBuffer` (it can be zero, meaning input data was just buffered). * or an error code if it fails (which can be tested using LZ4F_isError()) @@ -600,81 +720,9 @@ LZ4F_uncompressedUpdate(LZ4F_cctx* cctx, const LZ4F_compressOptions_t* cOptPtr); /********************************** - * Bulk processing dictionary API + * Custom memory allocation *********************************/ -/* A Dictionary is useful for the compression of small messages (KB range). - * It dramatically improves compression efficiency. - * - * LZ4 can ingest any input as dictionary, though only the last 64 KB are useful. - * Best results are generally achieved by using Zstandard's Dictionary Builder - * to generate a high-quality dictionary from a set of samples. - * - * Loading a dictionary has a cost, since it involves construction of tables. - * The Bulk processing dictionary API makes it possible to share this cost - * over an arbitrary number of compression jobs, even concurrently, - * markedly improving compression latency for these cases. - * - * The same dictionary will have to be used on the decompression side - * for decoding to be successful. - * To help identify the correct dictionary at decoding stage, - * the frame header allows optional embedding of a dictID field. - */ -typedef struct LZ4F_CDict_s LZ4F_CDict; - -/*! LZ4_createCDict() : - * When compressing multiple messages / blocks using the same dictionary, it's recommended to load it just once. - * LZ4_createCDict() will create a digested dictionary, ready to start future compression operations without startup delay. - * LZ4_CDict can be created once and shared by multiple threads concurrently, since its usage is read-only. - * `dictBuffer` can be released after LZ4_CDict creation, since its content is copied within CDict */ -LZ4FLIB_STATIC_API LZ4F_CDict* LZ4F_createCDict(const void* dictBuffer, size_t dictSize); -LZ4FLIB_STATIC_API void LZ4F_freeCDict(LZ4F_CDict* CDict); - - -/*! LZ4_compressFrame_usingCDict() : - * Compress an entire srcBuffer into a valid LZ4 frame using a digested Dictionary. - * cctx must point to a context created by LZ4F_createCompressionContext(). - * If cdict==NULL, compress without a dictionary. - * dstBuffer MUST be >= LZ4F_compressFrameBound(srcSize, preferencesPtr). - * If this condition is not respected, function will fail (@return an errorCode). - * The LZ4F_preferences_t structure is optional : you may provide NULL as argument, - * but it's not recommended, as it's the only way to provide dictID in the frame header. - * @return : number of bytes written into dstBuffer. - * or an error code if it fails (can be tested using LZ4F_isError()) */ -LZ4FLIB_STATIC_API size_t -LZ4F_compressFrame_usingCDict(LZ4F_cctx* cctx, - void* dst, size_t dstCapacity, - const void* src, size_t srcSize, - const LZ4F_CDict* cdict, - const LZ4F_preferences_t* preferencesPtr); - - -/*! LZ4F_compressBegin_usingCDict() : - * Inits streaming dictionary compression, and writes the frame header into dstBuffer. - * dstCapacity must be >= LZ4F_HEADER_SIZE_MAX bytes. - * `prefsPtr` is optional : you may provide NULL as argument, - * however, it's the only way to provide dictID in the frame header. - * @return : number of bytes written into dstBuffer for the header, - * or an error code (which can be tested using LZ4F_isError()) */ -LZ4FLIB_STATIC_API size_t -LZ4F_compressBegin_usingCDict(LZ4F_cctx* cctx, - void* dstBuffer, size_t dstCapacity, - const LZ4F_CDict* cdict, - const LZ4F_preferences_t* prefsPtr); - - -/*! LZ4F_decompress_usingDict() : - * Same as LZ4F_decompress(), using a predefined dictionary. - * Dictionary is used "in place", without any preprocessing. -** It must remain accessible throughout the entire frame decoding. */ -LZ4FLIB_STATIC_API size_t -LZ4F_decompress_usingDict(LZ4F_dctx* dctxPtr, - void* dstBuffer, size_t* dstSizePtr, - const void* srcBuffer, size_t* srcSizePtr, - const void* dict, size_t dictSize, - const LZ4F_decompressOptions_t* decompressOptionsPtr); - - /*! Custom memory allocation : v1.9.4+ * These prototypes make it possible to pass custom allocation/free functions. * LZ4F_customMem is provided at state creation time, using LZ4F_create*_advanced() listed below. @@ -699,6 +747,11 @@ LZ4FLIB_STATIC_API LZ4F_cctx* LZ4F_createCompressionContext_advanced(LZ4F_Custom LZ4FLIB_STATIC_API LZ4F_dctx* LZ4F_createDecompressionContext_advanced(LZ4F_CustomMem customMem, unsigned version); LZ4FLIB_STATIC_API LZ4F_CDict* LZ4F_createCDict_advanced(LZ4F_CustomMem customMem, const void* dictBuffer, size_t dictSize); +/*! Context size inspection : v1.10.1+ + * These functions return the total memory footprint of the provided context. + */ +LZ4FLIB_STATIC_API size_t LZ4F_cctx_size(const LZ4F_cctx* cctx); +LZ4FLIB_STATIC_API size_t LZ4F_dctx_size(const LZ4F_dctx* dctx); #if defined (__cplusplus) } diff --git a/CompressSave/LZ4WrapC/lz4/lz4frame_static.h b/CompressSave/LZ4WrapC/lz4/lz4frame_static.h index 2b44a63..dad4b8f 100644 --- a/CompressSave/LZ4WrapC/lz4/lz4frame_static.h +++ b/CompressSave/LZ4WrapC/lz4/lz4frame_static.h @@ -1,7 +1,7 @@ /* LZ4 auto-framing library Header File for static linking only - Copyright (C) 2011-2020, Yann Collet. + Copyright (c) Yann Collet. All rights reserved. BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) diff --git a/CompressSave/LZ4WrapC/lz4/lz4hc.c b/CompressSave/LZ4WrapC/lz4/lz4hc.c index 651f190..df0142b 100644 --- a/CompressSave/LZ4WrapC/lz4/lz4hc.c +++ b/CompressSave/LZ4WrapC/lz4/lz4hc.c @@ -1,6 +1,6 @@ /* LZ4 HC - High Compression Mode of LZ4 - Copyright (C) 2011-2020, Yann Collet. + Copyright (c) Yann Collet. All rights reserved. BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) @@ -52,19 +52,19 @@ /*=== Dependency ===*/ #define LZ4_HC_STATIC_LINKING_ONLY #include "lz4hc.h" +#include -/*=== Common definitions ===*/ -#if defined(__GNUC__) -# pragma GCC diagnostic ignored "-Wunused-function" -#endif -#if defined (__clang__) -# pragma clang diagnostic ignored "-Wunused-function" -#endif - -#define LZ4_COMMONDEFS_ONLY +/*=== Shared lz4.c code ===*/ #ifndef LZ4_SRC_INCLUDED -#include "lz4.c" /* LZ4_count, constants, mem */ +# if defined(__GNUC__) +# pragma GCC diagnostic ignored "-Wunused-function" +# endif +# if defined (__clang__) +# pragma clang diagnostic ignored "-Wunused-function" +# endif +# define LZ4_COMMONDEFS_ONLY +# include "lz4.c" /* LZ4_count, constants, mem */ #endif @@ -80,18 +80,158 @@ typedef enum { noDictCtx, usingDictCtxHc } dictCtx_directive; /*=== Macros ===*/ #define MIN(a,b) ( (a) < (b) ? (a) : (b) ) #define MAX(a,b) ( (a) > (b) ? (a) : (b) ) -#define HASH_FUNCTION(i) (((i) * 2654435761U) >> ((MINMATCH*8)-LZ4HC_HASH_LOG)) -#define DELTANEXTMAXD(p) chainTable[(p) & LZ4HC_MAXD_MASK] /* flexible, LZ4HC_MAXD dependent */ + + +/*=== Levels definition ===*/ +typedef enum { lz4mid, lz4hc, lz4opt } lz4hc_strat_e; +typedef struct { + lz4hc_strat_e strat; + int nbSearches; + U32 targetLength; +} cParams_t; +static const cParams_t k_clTable[LZ4HC_CLEVEL_MAX+1] = { + { lz4mid, 2, 16 }, /* 0, unused */ + { lz4mid, 2, 16 }, /* 1, unused */ + { lz4mid, 2, 16 }, /* 2 */ + { lz4hc, 4, 16 }, /* 3 */ + { lz4hc, 8, 16 }, /* 4 */ + { lz4hc, 16, 16 }, /* 5 */ + { lz4hc, 32, 16 }, /* 6 */ + { lz4hc, 64, 16 }, /* 7 */ + { lz4hc, 128, 16 }, /* 8 */ + { lz4hc, 256, 16 }, /* 9 */ + { lz4opt, 96, 64 }, /*10==LZ4HC_CLEVEL_OPT_MIN*/ + { lz4opt, 512,128 }, /*11 */ + { lz4opt,16384,LZ4_OPT_NUM }, /* 12==LZ4HC_CLEVEL_MAX */ +}; + +static cParams_t LZ4HC_getCLevelParams(int cLevel) +{ + /* note : clevel convention is a bit different from lz4frame, + * possibly something worth revisiting for consistency */ + if (cLevel < 1) + cLevel = LZ4HC_CLEVEL_DEFAULT; + cLevel = MIN(LZ4HC_CLEVEL_MAX, cLevel); + return k_clTable[cLevel]; +} + + +/*=== Hashing ===*/ +#define LZ4HC_HASHSIZE 4 +#define HASH_FUNCTION(i) (((i) * 2654435761U) >> ((MINMATCH*8)-LZ4HC_HASH_LOG)) +static U32 LZ4HC_hashPtr(const void* ptr) { return HASH_FUNCTION(LZ4_read32(ptr)); } + +#if defined(LZ4_FORCE_MEMORY_ACCESS) && (LZ4_FORCE_MEMORY_ACCESS==2) +/* lie to the compiler about data alignment; use with caution */ +static U64 LZ4_read64(const void* memPtr) { return *(const U64*) memPtr; } + +#elif defined(LZ4_FORCE_MEMORY_ACCESS) && (LZ4_FORCE_MEMORY_ACCESS==1) +/* __pack instructions are safer, but compiler specific */ +LZ4_PACK(typedef struct { U64 u64; }) LZ4_unalign64; +static U64 LZ4_read64(const void* ptr) { return ((const LZ4_unalign64*)ptr)->u64; } + +#else /* safe and portable access using memcpy() */ +static U64 LZ4_read64(const void* memPtr) +{ + U64 val; LZ4_memcpy(&val, memPtr, sizeof(val)); return val; +} + +#endif /* LZ4_FORCE_MEMORY_ACCESS */ + +#define LZ4MID_HASHSIZE 8 +#define LZ4MID_HASHLOG (LZ4HC_HASH_LOG-1) +#define LZ4MID_HASHTABLESIZE (1 << LZ4MID_HASHLOG) + +static U32 LZ4MID_hash4(U32 v) { return (v * 2654435761U) >> (32-LZ4MID_HASHLOG); } +static U32 LZ4MID_hash4Ptr(const void* ptr) { return LZ4MID_hash4(LZ4_read32(ptr)); } +/* note: hash7 hashes the lower 56-bits. + * It presumes input was read using little endian.*/ +static U32 LZ4MID_hash7(U64 v) { return (U32)(((v << (64-56)) * 58295818150454627ULL) >> (64-LZ4MID_HASHLOG)) ; } +static U64 LZ4_readLE64(const void* memPtr); +static U32 LZ4MID_hash8Ptr(const void* ptr) { return LZ4MID_hash7(LZ4_readLE64(ptr)); } + +static U64 LZ4_readLE64(const void* memPtr) +{ + if (LZ4_isLittleEndian()) { + return LZ4_read64(memPtr); + } else { + const BYTE* p = (const BYTE*)memPtr; + /* note: relies on the compiler to simplify this expression */ + return (U64)p[0] | ((U64)p[1]<<8) | ((U64)p[2]<<16) | ((U64)p[3]<<24) + | ((U64)p[4]<<32) | ((U64)p[5]<<40) | ((U64)p[6]<<48) | ((U64)p[7]<<56); + } +} + + +/*=== Count match length ===*/ +LZ4_FORCE_INLINE +unsigned LZ4HC_NbCommonBytes32(U32 val) +{ + assert(val != 0); + if (LZ4_isLittleEndian()) { +# if defined(_MSC_VER) && (_MSC_VER >= 1400) && !defined(LZ4_FORCE_SW_BITCOUNT) + unsigned long r; + _BitScanReverse(&r, val); + return (unsigned)((31 - r) >> 3); +# elif (defined(__clang__) || (defined(__GNUC__) && ((__GNUC__ > 3) || \ + ((__GNUC__ == 3) && (__GNUC_MINOR__ >= 4))))) && \ + !defined(LZ4_FORCE_SW_BITCOUNT) + return (unsigned)__builtin_clz(val) >> 3; +# else + val >>= 8; + val = ((((val + 0x00FFFF00) | 0x00FFFFFF) + val) | + (val + 0x00FF0000)) >> 24; + return (unsigned)val ^ 3; +# endif + } else { +# if defined(_MSC_VER) && (_MSC_VER >= 1400) && !defined(LZ4_FORCE_SW_BITCOUNT) + unsigned long r; + _BitScanForward(&r, val); + return (unsigned)(r >> 3); +# elif (defined(__clang__) || (defined(__GNUC__) && ((__GNUC__ > 3) || \ + ((__GNUC__ == 3) && (__GNUC_MINOR__ >= 4))))) && \ + !defined(LZ4_FORCE_SW_BITCOUNT) + return (unsigned)__builtin_ctz(val) >> 3; +# else + const U32 m = 0x01010101; + return (unsigned)((((val - 1) ^ val) & (m - 1)) * m) >> 24; +# endif + } +} + +/** LZ4HC_countBack() : + * @return : negative value, nb of common bytes before ip/match */ +LZ4_FORCE_INLINE +int LZ4HC_countBack(const BYTE* const ip, const BYTE* const match, + const BYTE* const iMin, const BYTE* const mMin) +{ + int back = 0; + int const min = (int)MAX(iMin - ip, mMin - match); + assert(min <= 0); + assert(ip >= iMin); assert((size_t)(ip-iMin) < (1U<<31)); + assert(match >= mMin); assert((size_t)(match - mMin) < (1U<<31)); + + while ((back - min) > 3) { + U32 const v = LZ4_read32(ip + back - 4) ^ LZ4_read32(match + back - 4); + if (v) { + return (back - (int)LZ4HC_NbCommonBytes32(v)); + } else back -= 4; /* 4-byte step */ + } + /* check remainder if any */ + while ( (back > min) + && (ip[back-1] == match[back-1]) ) + back--; + return back; +} + +/*=== Chain table updates ===*/ #define DELTANEXTU16(table, pos) table[(U16)(pos)] /* faster */ /* Make fields passed to, and updated by LZ4HC_encodeSequence explicit */ #define UPDATABLE(ip, op, anchor) &ip, &op, &anchor -#define LZ4HC_HASHSIZE 4 -static U32 LZ4HC_hashPtr(const void* ptr) { return HASH_FUNCTION(LZ4_read32(ptr)); } - /************************************** -* HC Compression +* Init **************************************/ static void LZ4HC_clearTables (LZ4HC_CCtx_internal* hc4) { @@ -119,6 +259,554 @@ static void LZ4HC_init_internal (LZ4HC_CCtx_internal* hc4, const BYTE* start) } +/************************************** +* Encode +**************************************/ +#if defined(LZ4_DEBUG) && (LZ4_DEBUG >= 2) +# define RAWLOG(...) fprintf(stderr, __VA_ARGS__) +void LZ4HC_hexOut(const void* src, size_t len) +{ + const BYTE* p = (const BYTE*)src; + size_t n; + for (n=0; n= _lev) { \ + RAWLOG("match bytes: "); \ + LZ4HC_hexOut(_ptr, _len); \ + RAWLOG("ref bytes: "); \ + LZ4HC_hexOut(_ref, _len); \ + } + +#else +# define HEX_CMP(l,p,r,_l) +#endif + +/* LZ4HC_encodeSequence() : + * @return : 0 if ok, + * 1 if buffer issue detected */ +LZ4_FORCE_INLINE int LZ4HC_encodeSequence ( + const BYTE** _ip, + BYTE** _op, + const BYTE** _anchor, + int matchLength, + int offset, + limitedOutput_directive limit, + BYTE* oend) +{ +#define ip (*_ip) +#define op (*_op) +#define anchor (*_anchor) + + BYTE* const token = op++; + +#if defined(LZ4_DEBUG) && (LZ4_DEBUG >= 6) + static const BYTE* start = NULL; + static U32 totalCost = 0; + U32 const pos = (start==NULL) ? 0 : (U32)(anchor - start); /* only works for single segment */ + U32 const ll = (U32)(ip - anchor); + U32 const llAdd = (ll>=15) ? ((ll-15) / 255) + 1 : 0; + U32 const mlAdd = (matchLength>=19) ? ((matchLength-19) / 255) + 1 : 0; + U32 const cost = 1 + llAdd + ll + 2 + mlAdd; + if (start==NULL) start = anchor; /* only works for single segment */ + DEBUGLOG(6, "pos:%7u -- literals:%4u, match:%4i, offset:%5i, cost:%4u + %5u", + pos, + (U32)(ip - anchor), matchLength, offset, + cost, totalCost); +# if 1 /* only works on single segment data */ + HEX_CMP(7, ip, ip-offset, matchLength); +# endif + totalCost += cost; +#endif + + /* Encode Literal length */ + { size_t litLen = (size_t)(ip - anchor); + LZ4_STATIC_ASSERT(notLimited == 0); + /* Check output limit */ + if (limit && ((op + (litLen / 255) + litLen + (2 + 1 + LASTLITERALS)) > oend)) { + DEBUGLOG(6, "Not enough room to write %i literals (%i bytes remaining)", + (int)litLen, (int)(oend - op)); + return 1; + } + if (litLen >= RUN_MASK) { + size_t len = litLen - RUN_MASK; + *token = (RUN_MASK << ML_BITS); + for(; len >= 255 ; len -= 255) *op++ = 255; + *op++ = (BYTE)len; + } else { + *token = (BYTE)(litLen << ML_BITS); + } + + /* Copy Literals */ + LZ4_wildCopy8(op, anchor, op + litLen); + op += litLen; + } + + /* Encode Offset */ + assert(offset <= LZ4_DISTANCE_MAX ); + assert(offset > 0); + LZ4_writeLE16(op, (U16)(offset)); op += 2; + + /* Encode MatchLength */ + assert(matchLength >= MINMATCH); + { size_t mlCode = (size_t)matchLength - MINMATCH; + if (limit && (op + (mlCode / 255) + (1 + LASTLITERALS) > oend)) { + DEBUGLOG(6, "Not enough room to write match length"); + return 1; /* Check output limit */ + } + if (mlCode >= ML_MASK) { + *token += ML_MASK; + mlCode -= ML_MASK; + for(; mlCode >= 510 ; mlCode -= 510) { *op++ = 255; *op++ = 255; } + if (mlCode >= 255) { mlCode -= 255; *op++ = 255; } + *op++ = (BYTE)mlCode; + } else { + *token += (BYTE)(mlCode); + } } + + /* Prepare next loop */ + ip += matchLength; + anchor = ip; + + return 0; + +#undef ip +#undef op +#undef anchor +} + + +typedef struct { + int off; + int len; + int back; /* negative value */ +} LZ4HC_match_t; + +LZ4HC_match_t LZ4HC_searchExtDict(const BYTE* ip, U32 ipIndex, + const BYTE* const iLowLimit, const BYTE* const iHighLimit, + const LZ4HC_CCtx_internal* dictCtx, U32 gDictEndIndex, + int currentBestML, int nbAttempts) +{ + size_t const lDictEndIndex = (size_t)(dictCtx->end - dictCtx->prefixStart) + dictCtx->dictLimit; + U32 lDictMatchIndex = dictCtx->hashTable[LZ4HC_hashPtr(ip)]; + U32 matchIndex = lDictMatchIndex + gDictEndIndex - (U32)lDictEndIndex; + int offset = 0, sBack = 0; + assert(lDictEndIndex <= 1 GB); + if (lDictMatchIndex>0) + DEBUGLOG(7, "lDictEndIndex = %zu, lDictMatchIndex = %u", lDictEndIndex, lDictMatchIndex); + while (ipIndex - matchIndex <= LZ4_DISTANCE_MAX && nbAttempts--) { + const BYTE* const matchPtr = dictCtx->prefixStart - dictCtx->dictLimit + lDictMatchIndex; + + if (LZ4_read32(matchPtr) == LZ4_read32(ip)) { + int mlt; + int back = 0; + const BYTE* vLimit = ip + (lDictEndIndex - lDictMatchIndex); + if (vLimit > iHighLimit) vLimit = iHighLimit; + mlt = (int)LZ4_count(ip+MINMATCH, matchPtr+MINMATCH, vLimit) + MINMATCH; + back = (ip > iLowLimit) ? LZ4HC_countBack(ip, matchPtr, iLowLimit, dictCtx->prefixStart) : 0; + mlt -= back; + if (mlt > currentBestML) { + currentBestML = mlt; + offset = (int)(ipIndex - matchIndex); + sBack = back; + DEBUGLOG(7, "found match of length %i within extDictCtx", currentBestML); + } } + + { U32 const nextOffset = DELTANEXTU16(dictCtx->chainTable, lDictMatchIndex); + lDictMatchIndex -= nextOffset; + matchIndex -= nextOffset; + } } + + { LZ4HC_match_t md; + md.len = currentBestML; + md.off = offset; + md.back = sBack; + return md; + } +} + +typedef LZ4HC_match_t (*LZ4MID_searchIntoDict_f)(const BYTE* ip, U32 ipIndex, + const BYTE* const iHighLimit, + const LZ4HC_CCtx_internal* dictCtx, U32 gDictEndIndex); + +static LZ4HC_match_t LZ4MID_searchHCDict(const BYTE* ip, U32 ipIndex, + const BYTE* const iHighLimit, + const LZ4HC_CCtx_internal* dictCtx, U32 gDictEndIndex) +{ + return LZ4HC_searchExtDict(ip,ipIndex, + ip, iHighLimit, + dictCtx, gDictEndIndex, + MINMATCH-1, 2); +} + +static LZ4HC_match_t LZ4MID_searchExtDict(const BYTE* ip, U32 ipIndex, + const BYTE* const iHighLimit, + const LZ4HC_CCtx_internal* dictCtx, U32 gDictEndIndex) +{ + size_t const lDictEndIndex = (size_t)(dictCtx->end - dictCtx->prefixStart) + dictCtx->dictLimit; + const U32* const hash4Table = dictCtx->hashTable; + const U32* const hash8Table = hash4Table + LZ4MID_HASHTABLESIZE; + DEBUGLOG(7, "LZ4MID_searchExtDict (ipIdx=%u)", ipIndex); + + /* search long match first */ + { U32 l8DictMatchIndex = hash8Table[LZ4MID_hash8Ptr(ip)]; + U32 m8Index = l8DictMatchIndex + gDictEndIndex - (U32)lDictEndIndex; + assert(lDictEndIndex <= 1 GB); + if (ipIndex - m8Index <= LZ4_DISTANCE_MAX) { + const BYTE* const matchPtr = dictCtx->prefixStart - dictCtx->dictLimit + l8DictMatchIndex; + const size_t safeLen = MIN(lDictEndIndex - l8DictMatchIndex, (size_t)(iHighLimit - ip)); + int mlt = (int)LZ4_count(ip, matchPtr, ip + safeLen); + if (mlt >= MINMATCH) { + LZ4HC_match_t md; + DEBUGLOG(7, "Found long ExtDict match of len=%u", mlt); + md.len = mlt; + md.off = (int)(ipIndex - m8Index); + md.back = 0; + return md; + } + } + } + + /* search for short match second */ + { U32 l4DictMatchIndex = hash4Table[LZ4MID_hash4Ptr(ip)]; + U32 m4Index = l4DictMatchIndex + gDictEndIndex - (U32)lDictEndIndex; + if (ipIndex - m4Index <= LZ4_DISTANCE_MAX) { + const BYTE* const matchPtr = dictCtx->prefixStart - dictCtx->dictLimit + l4DictMatchIndex; + const size_t safeLen = MIN(lDictEndIndex - l4DictMatchIndex, (size_t)(iHighLimit - ip)); + int mlt = (int)LZ4_count(ip, matchPtr, ip + safeLen); + if (mlt >= MINMATCH) { + LZ4HC_match_t md; + DEBUGLOG(7, "Found short ExtDict match of len=%u", mlt); + md.len = mlt; + md.off = (int)(ipIndex - m4Index); + md.back = 0; + return md; + } + } + } + + /* nothing found */ + { LZ4HC_match_t const md = {0, 0, 0 }; + return md; + } +} + +/************************************** +* Mid Compression (level 2) +**************************************/ + +LZ4_FORCE_INLINE void +LZ4MID_addPosition(U32* hTable, U32 hValue, U32 index) +{ + hTable[hValue] = index; +} + +#define ADDPOS8(_p, _idx) LZ4MID_addPosition(hash8Table, LZ4MID_hash8Ptr(_p), _idx) +#define ADDPOS4(_p, _idx) LZ4MID_addPosition(hash4Table, LZ4MID_hash4Ptr(_p), _idx) + +/* Fill hash tables with references into dictionary. + * The resulting table is only exploitable by LZ4MID (level 2) */ +static void +LZ4MID_fillHTable (LZ4HC_CCtx_internal* cctx, const void* dict, size_t size) +{ + U32* const hash4Table = cctx->hashTable; + U32* const hash8Table = hash4Table + LZ4MID_HASHTABLESIZE; + const BYTE* const prefixPtr = (const BYTE*)dict; + U32 const prefixIdx = cctx->dictLimit; + U32 const target = prefixIdx + (U32)size - LZ4MID_HASHSIZE; + U32 idx = cctx->nextToUpdate; + assert(dict == cctx->prefixStart); + DEBUGLOG(4, "LZ4MID_fillHTable (size:%zu)", size); + if (size <= LZ4MID_HASHSIZE) + return; + + for (; idx < target; idx += 3) { + ADDPOS4(prefixPtr+idx-prefixIdx, idx); + ADDPOS8(prefixPtr+idx+1-prefixIdx, idx+1); + } + + idx = (size > 32 KB + LZ4MID_HASHSIZE) ? target - 32 KB : cctx->nextToUpdate; + for (; idx < target; idx += 1) { + ADDPOS8(prefixPtr+idx-prefixIdx, idx); + } + + cctx->nextToUpdate = target; +} + +static LZ4MID_searchIntoDict_f select_searchDict_function(const LZ4HC_CCtx_internal* dictCtx) +{ + if (dictCtx == NULL) return NULL; + if (LZ4HC_getCLevelParams(dictCtx->compressionLevel).strat == lz4mid) + return LZ4MID_searchExtDict; + return LZ4MID_searchHCDict; +} + +/* preconditions: + * - *srcSizePtr within [1, LZ4_MAX_INPUT_SIZE] + * - src is valid + * - maxOutputSize >= 1 + * - dst is valid + */ +static int LZ4MID_compress ( + LZ4HC_CCtx_internal* const ctx, + const char* const src, + char* const dst, + int* srcSizePtr, + int const maxOutputSize, + const limitedOutput_directive limit, + const dictCtx_directive dict + ) +{ + U32* const hash4Table = ctx->hashTable; + U32* const hash8Table = hash4Table + LZ4MID_HASHTABLESIZE; + const BYTE* ip = (const BYTE*)src; + const BYTE* anchor = ip; + const BYTE* const iend = ip + *srcSizePtr; + const BYTE* const mflimit = iend - MFLIMIT; + const BYTE* const matchlimit = (iend - LASTLITERALS); + const BYTE* const ilimit = (iend - LZ4MID_HASHSIZE); + BYTE* op = (BYTE*)dst; + BYTE* oend = op + maxOutputSize; + + const BYTE* const prefixPtr = ctx->prefixStart; + const U32 prefixIdx = ctx->dictLimit; + const U32 ilimitIdx = (U32)(ilimit - prefixPtr) + prefixIdx; + const BYTE* const dictStart = ctx->dictStart; + const U32 dictIdx = ctx->lowLimit; + const U32 gDictEndIndex = ctx->lowLimit; + const LZ4MID_searchIntoDict_f searchIntoDict = (dict == usingDictCtxHc) ? select_searchDict_function(ctx->dictCtx) : NULL; + unsigned matchLength; + unsigned matchDistance; + + DEBUGLOG(5, "LZ4MID_compress (%i bytes)", *srcSizePtr); + + /* preconditions verifications */ + if (dict == usingDictCtxHc) DEBUGLOG(5, "usingDictCtxHc"); + assert(*srcSizePtr > 0); + assert(*srcSizePtr <= LZ4_MAX_INPUT_SIZE); + assert(src != NULL); + assert(maxOutputSize >= 1); + assert(dst != NULL); + + if (limit == fillOutput) oend -= LASTLITERALS; /* Hack for support LZ4 format restriction */ + if (*srcSizePtr < LZ4_minLength) + goto _lz4mid_last_literals; /* Input too small, no compression (all literals) */ + + /* main loop */ + while (ip <= mflimit) { + const U32 ipIndex = (U32)(ip - prefixPtr) + prefixIdx; + /* search long match */ + { U32 const h8 = LZ4MID_hash8Ptr(ip); + U32 const pos8 = hash8Table[h8]; + assert(h8 < LZ4MID_HASHTABLESIZE); + assert(pos8 < ipIndex); + LZ4MID_addPosition(hash8Table, h8, ipIndex); + if (ipIndex - pos8 <= LZ4_DISTANCE_MAX) { + /* match candidate found */ + if (pos8 >= prefixIdx) { + const BYTE* const matchPtr = prefixPtr + pos8 - prefixIdx; + assert(matchPtr < ip); + matchLength = LZ4_count(ip, matchPtr, matchlimit); + if (matchLength >= MINMATCH) { + DEBUGLOG(7, "found long match at pos %u (len=%u)", pos8, matchLength); + matchDistance = ipIndex - pos8; + goto _lz4mid_encode_sequence; + } + } else { + if (pos8 >= dictIdx) { + /* extDict match candidate */ + const BYTE* const matchPtr = dictStart + (pos8 - dictIdx); + const size_t safeLen = MIN(prefixIdx - pos8, (size_t)(matchlimit - ip)); + matchLength = LZ4_count(ip, matchPtr, ip + safeLen); + if (matchLength >= MINMATCH) { + DEBUGLOG(7, "found long match at ExtDict pos %u (len=%u)", pos8, matchLength); + matchDistance = ipIndex - pos8; + goto _lz4mid_encode_sequence; + } + } + } + } } + /* search short match */ + { U32 const h4 = LZ4MID_hash4Ptr(ip); + U32 const pos4 = hash4Table[h4]; + assert(h4 < LZ4MID_HASHTABLESIZE); + assert(pos4 < ipIndex); + LZ4MID_addPosition(hash4Table, h4, ipIndex); + if (ipIndex - pos4 <= LZ4_DISTANCE_MAX) { + /* match candidate found */ + if (pos4 >= prefixIdx) { + /* only search within prefix */ + const BYTE* const matchPtr = prefixPtr + (pos4 - prefixIdx); + assert(matchPtr < ip); + assert(matchPtr >= prefixPtr); + matchLength = LZ4_count(ip, matchPtr, matchlimit); + if (matchLength >= MINMATCH) { + /* short match found, let's just check ip+1 for longer */ + U32 const h8 = LZ4MID_hash8Ptr(ip+1); + U32 const pos8 = hash8Table[h8]; + U32 const m2Distance = ipIndex + 1 - pos8; + matchDistance = ipIndex - pos4; + if ( m2Distance <= LZ4_DISTANCE_MAX + && pos8 >= prefixIdx /* only search within prefix */ + && likely(ip < mflimit) + ) { + const BYTE* const m2Ptr = prefixPtr + (pos8 - prefixIdx); + unsigned ml2 = LZ4_count(ip+1, m2Ptr, matchlimit); + if (ml2 > matchLength) { + LZ4MID_addPosition(hash8Table, h8, ipIndex+1); + ip++; + matchLength = ml2; + matchDistance = m2Distance; + } } + goto _lz4mid_encode_sequence; + } + } else { + if (pos4 >= dictIdx) { + /* extDict match candidate */ + const BYTE* const matchPtr = dictStart + (pos4 - dictIdx); + const size_t safeLen = MIN(prefixIdx - pos4, (size_t)(matchlimit - ip)); + matchLength = LZ4_count(ip, matchPtr, ip + safeLen); + if (matchLength >= MINMATCH) { + DEBUGLOG(7, "found match at ExtDict pos %u (len=%u)", pos4, matchLength); + matchDistance = ipIndex - pos4; + goto _lz4mid_encode_sequence; + } + } + } + } } + /* no match found in prefix */ + if ( (dict == usingDictCtxHc) + && (ipIndex - gDictEndIndex < LZ4_DISTANCE_MAX - 8) ) { + /* search a match into external dictionary */ + LZ4HC_match_t dMatch = searchIntoDict(ip, ipIndex, + matchlimit, + ctx->dictCtx, gDictEndIndex); + if (dMatch.len >= MINMATCH) { + DEBUGLOG(7, "found Dictionary match (offset=%i)", dMatch.off); + assert(dMatch.back == 0); + matchLength = (unsigned)dMatch.len; + matchDistance = (unsigned)dMatch.off; + goto _lz4mid_encode_sequence; + } + } + /* no match found */ + ip += 1 + ((ip-anchor) >> 9); /* skip faster over incompressible data */ + continue; + +_lz4mid_encode_sequence: + /* catch back */ + while (((ip > anchor) & ((U32)(ip-prefixPtr) > matchDistance)) && (unlikely(ip[-1] == ip[-(int)matchDistance-1]))) { + ip--; matchLength++; + }; + + /* fill table with beginning of match */ + ADDPOS8(ip+1, ipIndex+1); + ADDPOS8(ip+2, ipIndex+2); + ADDPOS4(ip+1, ipIndex+1); + + /* encode */ + { BYTE* const saved_op = op; + /* LZ4HC_encodeSequence always updates @op; on success, it updates @ip and @anchor */ + if (LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), + (int)matchLength, (int)matchDistance, + limit, oend) ) { + op = saved_op; /* restore @op value before failed LZ4HC_encodeSequence */ + goto _lz4mid_dest_overflow; + } + } + + /* fill table with end of match */ + { U32 endMatchIdx = (U32)(ip-prefixPtr) + prefixIdx; + U32 pos_m2 = endMatchIdx - 2; + if (pos_m2 < ilimitIdx) { + if (likely(ip - prefixPtr > 5)) { + ADDPOS8(ip-5, endMatchIdx - 5); + } + ADDPOS8(ip-3, endMatchIdx - 3); + ADDPOS8(ip-2, endMatchIdx - 2); + ADDPOS4(ip-2, endMatchIdx - 2); + ADDPOS4(ip-1, endMatchIdx - 1); + } + } + } + +_lz4mid_last_literals: + /* Encode Last Literals */ + { size_t lastRunSize = (size_t)(iend - anchor); /* literals */ + size_t llAdd = (lastRunSize + 255 - RUN_MASK) / 255; + size_t const totalSize = 1 + llAdd + lastRunSize; + if (limit == fillOutput) oend += LASTLITERALS; /* restore correct value */ + if (limit && (op + totalSize > oend)) { + if (limit == limitedOutput) return 0; /* not enough space in @dst */ + /* adapt lastRunSize to fill 'dest' */ + lastRunSize = (size_t)(oend - op) - 1 /*token*/; + llAdd = (lastRunSize + 256 - RUN_MASK) / 256; + lastRunSize -= llAdd; + } + DEBUGLOG(6, "Final literal run : %i literals", (int)lastRunSize); + ip = anchor + lastRunSize; /* can be != iend if limit==fillOutput */ + + if (lastRunSize >= RUN_MASK) { + size_t accumulator = lastRunSize - RUN_MASK; + *op++ = (RUN_MASK << ML_BITS); + for(; accumulator >= 255 ; accumulator -= 255) + *op++ = 255; + *op++ = (BYTE) accumulator; + } else { + *op++ = (BYTE)(lastRunSize << ML_BITS); + } + assert(lastRunSize <= (size_t)(oend - op)); + LZ4_memcpy(op, anchor, lastRunSize); + op += lastRunSize; + } + + /* End */ + DEBUGLOG(5, "compressed %i bytes into %i bytes", *srcSizePtr, (int)((char*)op - dst)); + assert(ip >= (const BYTE*)src); + assert(ip <= iend); + *srcSizePtr = (int)(ip - (const BYTE*)src); + assert((char*)op >= dst); + assert(op <= oend); + assert((char*)op - dst < INT_MAX); + return (int)((char*)op - dst); + +_lz4mid_dest_overflow: + if (limit == fillOutput) { + /* Assumption : @ip, @anchor, @optr and @matchLength must be set correctly */ + size_t const ll = (size_t)(ip - anchor); + size_t const ll_addbytes = (ll + 240) / 255; + size_t const ll_totalCost = 1 + ll_addbytes + ll; + BYTE* const maxLitPos = oend - 3; /* 2 for offset, 1 for token */ + DEBUGLOG(6, "Last sequence is overflowing : %u literals, %u remaining space", + (unsigned)ll, (unsigned)(oend-op)); + if (op + ll_totalCost <= maxLitPos) { + /* ll validated; now adjust match length */ + size_t const bytesLeftForMl = (size_t)(maxLitPos - (op+ll_totalCost)); + size_t const maxMlSize = MINMATCH + (ML_MASK-1) + (bytesLeftForMl * 255); + assert(maxMlSize < INT_MAX); + if ((size_t)matchLength > maxMlSize) matchLength= (unsigned)maxMlSize; + if ((oend + LASTLITERALS) - (op + ll_totalCost + 2) - 1 + matchLength >= MFLIMIT) { + DEBUGLOG(6, "Let's encode a last sequence (ll=%u, ml=%u)", (unsigned)ll, matchLength); + LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), + (int)matchLength, (int)matchDistance, + notLimited, oend); + } } + DEBUGLOG(6, "Let's finish with a run of literals (%u bytes left)", (unsigned)(oend-op)); + goto _lz4mid_last_literals; + } + /* compression failed */ + return 0; +} + + +/************************************** +* HC Compression - Search +**************************************/ + /* Update chains up to ip (excluded) */ LZ4_FORCE_INLINE void LZ4HC_Insert (LZ4HC_CCtx_internal* hc4, const BYTE* ip) { @@ -143,23 +831,6 @@ LZ4_FORCE_INLINE void LZ4HC_Insert (LZ4HC_CCtx_internal* hc4, const BYTE* ip) hc4->nextToUpdate = target; } -/** LZ4HC_countBack() : - * @return : negative value, nb of common bytes before ip/match */ -LZ4_FORCE_INLINE -int LZ4HC_countBack(const BYTE* const ip, const BYTE* const match, - const BYTE* const iMin, const BYTE* const mMin) -{ - int back = 0; - int const min = (int)MAX(iMin - ip, mMin - match); - assert(min <= 0); - assert(ip >= iMin); assert((size_t)(ip-iMin) < (1U<<31)); - assert(match >= mMin); assert((size_t)(match - mMin) < (1U<<31)); - while ( (back > min) - && (ip[back-1] == match[back-1]) ) - back--; - return back; -} - #if defined(_MSC_VER) # define LZ4HC_rotl32(x,r) _rotl(x,r) #else @@ -239,10 +910,6 @@ static int LZ4HC_protectDictEnd(U32 const dictLimit, U32 const matchIndex) typedef enum { rep_untested, rep_not, rep_confirmed } repeat_state_e; typedef enum { favorCompressionRatio=0, favorDecompressionSpeed } HCfavor_e; -typedef struct { - int off; - int len; -} LZ4HC_match_t; LZ4_FORCE_INLINE LZ4HC_match_t LZ4HC_InsertAndGetWiderMatch ( @@ -250,7 +917,6 @@ LZ4HC_InsertAndGetWiderMatch ( const BYTE* const ip, const BYTE* const iLowLimit, const BYTE* const iHighLimit, int longest, - const BYTE** startpos, const int maxNbAttempts, const int patternAnalysis, const int chainSwap, const dictCtx_directive dict, @@ -258,7 +924,7 @@ LZ4HC_InsertAndGetWiderMatch ( { U16* const chainTable = hc4->chainTable; U32* const hashTable = hc4->hashTable; - const LZ4HC_CCtx_internal * const dictCtx = hc4->dictCtx; + const LZ4HC_CCtx_internal* const dictCtx = hc4->dictCtx; const BYTE* const prefixPtr = hc4->prefixStart; const U32 prefixIdx = hc4->dictLimit; const U32 ipIndex = (U32)(ip - prefixPtr) + prefixIdx; @@ -274,11 +940,9 @@ LZ4HC_InsertAndGetWiderMatch ( U32 matchIndex; repeat_state_e repeat = rep_untested; size_t srcPatternLength = 0; - int offset = 0; + int offset = 0, sBack = 0; DEBUGLOG(7, "LZ4HC_InsertAndGetWiderMatch"); - assert(startpos != NULL); - *startpos = ip; /* in case there is no solution */ /* First Match */ LZ4HC_Insert(hc4, ip); /* insert all prior positions up to ip (excluded) */ matchIndex = hashTable[LZ4HC_hashPtr(ip)]; @@ -304,8 +968,9 @@ LZ4HC_InsertAndGetWiderMatch ( if (matchLength > longest) { longest = matchLength; offset = (int)(ipIndex - matchIndex); - *startpos = ip + back; + sBack = back; DEBUGLOG(7, "Found match of len=%i within prefix, offset=%i, back=%i", longest, offset, -back); + HEX_CMP(7, ip + back, ip + back - offset, (size_t)matchLength); } } } } else { /* lowestMatchIndex <= matchIndex < dictLimit : within Ext Dict */ const BYTE* const matchPtr = dictStart + (matchIndex - dictIdx); @@ -323,8 +988,9 @@ LZ4HC_InsertAndGetWiderMatch ( if (matchLength > longest) { longest = matchLength; offset = (int)(ipIndex - matchIndex); - *startpos = ip + back; + sBack = back; DEBUGLOG(7, "Found match of len=%i within dict, offset=%i, back=%i", longest, offset, -back); + HEX_CMP(7, ip + back, matchPtr + back, (size_t)matchLength); } } } if (chainSwap && matchLength==longest) { /* better match => select a better chain */ @@ -413,7 +1079,7 @@ LZ4HC_InsertAndGetWiderMatch ( assert(maxML < 2 GB); longest = (int)maxML; offset = (int)(ipIndex - matchIndex); - *startpos = ip; + assert(sBack == 0); DEBUGLOG(7, "Found repeat pattern match of len=%i, offset=%i", longest, offset); } { U32 const distToNextPattern = DELTANEXTU16(chainTable, matchIndex); @@ -431,7 +1097,7 @@ LZ4HC_InsertAndGetWiderMatch ( if ( dict == usingDictCtxHc && nbAttempts > 0 - && ipIndex - lowestMatchIndex < LZ4_DISTANCE_MAX) { + && withinStartDistance) { size_t const dictEndOffset = (size_t)(dictCtx->end - dictCtx->prefixStart) + dictCtx->dictLimit; U32 dictMatchIndex = dictCtx->hashTable[LZ4HC_hashPtr(ip)]; assert(dictEndOffset <= 1 GB); @@ -451,7 +1117,7 @@ LZ4HC_InsertAndGetWiderMatch ( if (mlt > longest) { longest = mlt; offset = (int)(ipIndex - matchIndex); - *startpos = ip + back; + sBack = back; DEBUGLOG(7, "found match of length %i within extDictCtx", longest); } } @@ -464,6 +1130,7 @@ LZ4HC_InsertAndGetWiderMatch ( assert(longest >= 0); md.len = longest; md.off = offset; + md.back = sBack; return md; } } @@ -475,108 +1142,24 @@ LZ4HC_InsertAndFindBestMatch(LZ4HC_CCtx_internal* const hc4, /* Index table wi const int patternAnalysis, const dictCtx_directive dict) { - const BYTE* uselessPtr = ip; DEBUGLOG(7, "LZ4HC_InsertAndFindBestMatch"); /* note : LZ4HC_InsertAndGetWiderMatch() is able to modify the starting position of a match (*startpos), * but this won't be the case here, as we define iLowLimit==ip, * so LZ4HC_InsertAndGetWiderMatch() won't be allowed to search past ip */ - return LZ4HC_InsertAndGetWiderMatch(hc4, ip, ip, iLimit, MINMATCH-1, &uselessPtr, maxNbAttempts, patternAnalysis, 0 /*chainSwap*/, dict, favorCompressionRatio); + return LZ4HC_InsertAndGetWiderMatch(hc4, ip, ip, iLimit, MINMATCH-1, maxNbAttempts, patternAnalysis, 0 /*chainSwap*/, dict, favorCompressionRatio); } -/* LZ4HC_encodeSequence() : - * @return : 0 if ok, - * 1 if buffer issue detected */ -LZ4_FORCE_INLINE int LZ4HC_encodeSequence ( - const BYTE** _ip, - BYTE** _op, - const BYTE** _anchor, - int matchLength, - int offset, - limitedOutput_directive limit, - BYTE* oend) -{ -#define ip (*_ip) -#define op (*_op) -#define anchor (*_anchor) - - size_t length; - BYTE* const token = op++; - -#if defined(LZ4_DEBUG) && (LZ4_DEBUG >= 6) - static const BYTE* start = NULL; - static U32 totalCost = 0; - U32 const pos = (start==NULL) ? 0 : (U32)(anchor - start); - U32 const ll = (U32)(ip - anchor); - U32 const llAdd = (ll>=15) ? ((ll-15) / 255) + 1 : 0; - U32 const mlAdd = (matchLength>=19) ? ((matchLength-19) / 255) + 1 : 0; - U32 const cost = 1 + llAdd + ll + 2 + mlAdd; - if (start==NULL) start = anchor; /* only works for single segment */ - /* g_debuglog_enable = (pos >= 2228) & (pos <= 2262); */ - DEBUGLOG(6, "pos:%7u -- literals:%4u, match:%4i, offset:%5i, cost:%4u + %5u", - pos, - (U32)(ip - anchor), matchLength, offset, - cost, totalCost); - totalCost += cost; -#endif - - /* Encode Literal length */ - length = (size_t)(ip - anchor); - LZ4_STATIC_ASSERT(notLimited == 0); - /* Check output limit */ - if (limit && ((op + (length / 255) + length + (2 + 1 + LASTLITERALS)) > oend)) { - DEBUGLOG(6, "Not enough room to write %i literals (%i bytes remaining)", - (int)length, (int)(oend - op)); - return 1; - } - if (length >= RUN_MASK) { - size_t len = length - RUN_MASK; - *token = (RUN_MASK << ML_BITS); - for(; len >= 255 ; len -= 255) *op++ = 255; - *op++ = (BYTE)len; - } else { - *token = (BYTE)(length << ML_BITS); - } - - /* Copy Literals */ - LZ4_wildCopy8(op, anchor, op + length); - op += length; - - /* Encode Offset */ - assert(offset <= LZ4_DISTANCE_MAX ); - assert(offset > 0); - LZ4_writeLE16(op, (U16)(offset)); op += 2; - - /* Encode MatchLength */ - assert(matchLength >= MINMATCH); - length = (size_t)matchLength - MINMATCH; - if (limit && (op + (length / 255) + (1 + LASTLITERALS) > oend)) { - DEBUGLOG(6, "Not enough room to write match length"); - return 1; /* Check output limit */ - } - if (length >= ML_MASK) { - *token += ML_MASK; - length -= ML_MASK; - for(; length >= 510 ; length -= 510) { *op++ = 255; *op++ = 255; } - if (length >= 255) { length -= 255; *op++ = 255; } - *op++ = (BYTE)length; - } else { - *token += (BYTE)(length); - } - - /* Prepare next loop */ - ip += matchLength; - anchor = ip; - - return 0; -} -#undef ip -#undef op -#undef anchor +/* preconditions: + * - *srcSizePtr within [1, LZ4_MAX_INPUT_SIZE] + * - src is valid + * - maxOutputSize >= 1 + * - dst is valid + */ LZ4_FORCE_INLINE int LZ4HC_compress_hashChain ( LZ4HC_CCtx_internal* const ctx, - const char* const source, - char* const dest, + const char* const src, + char* const dst, int* srcSizePtr, int const maxOutputSize, int maxNbAttempts, @@ -587,24 +1170,31 @@ LZ4_FORCE_INLINE int LZ4HC_compress_hashChain ( const int inputSize = *srcSizePtr; const int patternAnalysis = (maxNbAttempts > 128); /* levels 9+ */ - const BYTE* ip = (const BYTE*) source; + const BYTE* ip = (const BYTE*)src; const BYTE* anchor = ip; const BYTE* const iend = ip + inputSize; const BYTE* const mflimit = iend - MFLIMIT; const BYTE* const matchlimit = (iend - LASTLITERALS); - BYTE* optr = (BYTE*) dest; - BYTE* op = (BYTE*) dest; + BYTE* optr = (BYTE*) dst; + BYTE* op = (BYTE*) dst; BYTE* oend = op + maxOutputSize; const BYTE* start0; const BYTE* start2 = NULL; const BYTE* start3 = NULL; LZ4HC_match_t m0, m1, m2, m3; - const LZ4HC_match_t nomatch = {0, 0}; + const LZ4HC_match_t nomatch = {0, 0, 0}; /* init */ DEBUGLOG(5, "LZ4HC_compress_hashChain (dict?=>%i)", dict); + + /* preconditions verifications */ + assert(*srcSizePtr >= 1); + assert(src != NULL); + assert(maxOutputSize >= 1); + assert(dst != NULL); + *srcSizePtr = 0; if (limit == fillOutput) oend -= LASTLITERALS; /* Hack for support LZ4 format restriction */ if (inputSize < LZ4_minLength) goto _last_literals; /* Input too small, no compression (all literals) */ @@ -620,16 +1210,21 @@ LZ4_FORCE_INLINE int LZ4HC_compress_hashChain ( _Search2: DEBUGLOG(7, "_Search2 (currently found match of size %i)", m1.len); if (ip+m1.len <= mflimit) { + start2 = ip + m1.len - 2; m2 = LZ4HC_InsertAndGetWiderMatch(ctx, - ip + m1.len - 2, ip + 0, matchlimit, m1.len, &start2, + start2, ip + 0, matchlimit, m1.len, maxNbAttempts, patternAnalysis, 0, dict, favorCompressionRatio); + start2 += m2.back; } else { m2 = nomatch; /* do not search further */ } if (m2.len <= m1.len) { /* No better match => encode ML1 immediately */ optr = op; - if (LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), m1.len, m1.off, limit, oend)) goto _dest_overflow; + if (LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), + m1.len, m1.off, + limit, oend) ) + goto _dest_overflow; continue; } @@ -660,9 +1255,11 @@ _Search3: } if (start2 + m2.len <= mflimit) { + start3 = start2 + m2.len - 3; m3 = LZ4HC_InsertAndGetWiderMatch(ctx, - start2 + m2.len - 3, start2, matchlimit, m2.len, &start3, + start3, start2, matchlimit, m2.len, maxNbAttempts, patternAnalysis, 0, dict, favorCompressionRatio); + start3 += m3.back; } else { m3 = nomatch; /* do not search further */ } @@ -672,11 +1269,15 @@ _Search3: if (start2 < ip+m1.len) m1.len = (int)(start2 - ip); /* Now, encode 2 sequences */ optr = op; - if (LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), m1.len, m1.off, limit, oend)) + if (LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), + m1.len, m1.off, + limit, oend) ) goto _dest_overflow; ip = start2; optr = op; - if (LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), m2.len, m2.off, limit, oend)) { + if (LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), + m2.len, m2.off, + limit, oend) ) { m1 = m2; goto _dest_overflow; } @@ -696,7 +1297,10 @@ _Search3: } optr = op; - if (LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), m1.len, m1.off, limit, oend)) goto _dest_overflow; + if (LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), + m1.len, m1.off, + limit, oend) ) + goto _dest_overflow; ip = start3; m1 = m3; @@ -731,7 +1335,10 @@ _Search3: } } optr = op; - if (LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), m1.len, m1.off, limit, oend)) goto _dest_overflow; + if ( LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), + m1.len, m1.off, + limit, oend) ) + goto _dest_overflow; /* ML2 becomes ML1 */ ip = start2; m1 = m2; @@ -772,12 +1379,12 @@ _last_literals: } /* End */ - *srcSizePtr = (int) (((const char*)ip) - source); - return (int) (((char*)op)-dest); + *srcSizePtr = (int) (((const char*)ip) - src); + return (int) (((char*)op)-dst); _dest_overflow: if (limit == fillOutput) { - /* Assumption : ip, anchor, ml and ref must be set correctly */ + /* Assumption : @ip, @anchor, @optr and @m1 must be set correctly */ size_t const ll = (size_t)(ip - anchor); size_t const ll_addbytes = (ll + 240) / 255; size_t const ll_totalCost = 1 + ll_addbytes + ll; @@ -808,8 +1415,7 @@ static int LZ4HC_compress_optimal( LZ4HC_CCtx_internal* ctx, const dictCtx_directive dict, const HCfavor_e favorDecSpeed); - -LZ4_FORCE_INLINE int +static int LZ4HC_compress_generic_internal ( LZ4HC_CCtx_internal* const ctx, const char* const src, @@ -821,42 +1427,26 @@ LZ4HC_compress_generic_internal ( const dictCtx_directive dict ) { - typedef enum { lz4hc, lz4opt } lz4hc_strat_e; - typedef struct { - lz4hc_strat_e strat; - int nbSearches; - U32 targetLength; - } cParams_t; - static const cParams_t clTable[LZ4HC_CLEVEL_MAX+1] = { - { lz4hc, 2, 16 }, /* 0, unused */ - { lz4hc, 2, 16 }, /* 1, unused */ - { lz4hc, 2, 16 }, /* 2, unused */ - { lz4hc, 4, 16 }, /* 3 */ - { lz4hc, 8, 16 }, /* 4 */ - { lz4hc, 16, 16 }, /* 5 */ - { lz4hc, 32, 16 }, /* 6 */ - { lz4hc, 64, 16 }, /* 7 */ - { lz4hc, 128, 16 }, /* 8 */ - { lz4hc, 256, 16 }, /* 9 */ - { lz4opt, 96, 64 }, /*10==LZ4HC_CLEVEL_OPT_MIN*/ - { lz4opt, 512,128 }, /*11 */ - { lz4opt,16384,LZ4_OPT_NUM }, /* 12==LZ4HC_CLEVEL_MAX */ - }; + DEBUGLOG(5, "LZ4HC_compress_generic_internal(src=%p, srcSize=%d, dstCapacity=%d)", + src, *srcSizePtr, dstCapacity); - DEBUGLOG(5, "LZ4HC_compress_generic_internal(src=%p, srcSize=%d)", - src, *srcSizePtr); - - if (limit == fillOutput && dstCapacity < 1) return 0; /* Impossible to store anything */ - if ((U32)*srcSizePtr > (U32)LZ4_MAX_INPUT_SIZE) return 0; /* Unsupported input size (too large or negative) */ + /* input sanitization */ + if ((U32)*srcSizePtr > (U32)LZ4_MAX_INPUT_SIZE) return 0; /* Unsupported input size (too large or negative) */ + if (dstCapacity < 1) return 0; /* Invalid: impossible to store anything */ + assert(dst); /* since dstCapacity >= 1, dst must be valid */ + if (*srcSizePtr == 0) { *dst = 0; return 1; } + assert(src != NULL); /* since *srcSizePtr >= 1, src must be valid */ ctx->end += *srcSizePtr; - if (cLevel < 1) cLevel = LZ4HC_CLEVEL_DEFAULT; /* note : convention is different from lz4frame, maybe something to review */ - cLevel = MIN(LZ4HC_CLEVEL_MAX, cLevel); - { cParams_t const cParam = clTable[cLevel]; + { cParams_t const cParam = LZ4HC_getCLevelParams(cLevel); HCfavor_e const favor = ctx->favorDecSpeed ? favorDecompressionSpeed : favorCompressionRatio; int result; - if (cParam.strat == lz4hc) { + if (cParam.strat == lz4mid) { + result = LZ4MID_compress(ctx, + src, dst, srcSizePtr, dstCapacity, + limit, dict); + } else if (cParam.strat == lz4hc) { result = LZ4HC_compress_hashChain(ctx, src, dst, srcSizePtr, dstCapacity, cParam.nbSearches, limit, dict); @@ -865,7 +1455,7 @@ LZ4HC_compress_generic_internal ( result = LZ4HC_compress_optimal(ctx, src, dst, srcSizePtr, dstCapacity, cParam.nbSearches, cParam.targetLength, limit, - cLevel == LZ4HC_CLEVEL_MAX, /* ultra mode */ + cLevel >= LZ4HC_CLEVEL_MAX, /* ultra mode */ dict, favor); } if (result <= 0) ctx->dirty = 1; @@ -890,6 +1480,13 @@ LZ4HC_compress_generic_noDictCtx ( return LZ4HC_compress_generic_internal(ctx, src, dst, srcSizePtr, dstCapacity, cLevel, limit, noDictCtx); } +static int isStateCompatible(const LZ4HC_CCtx_internal* ctx1, const LZ4HC_CCtx_internal* ctx2) +{ + int const isMid1 = LZ4HC_getCLevelParams(ctx1->compressionLevel).strat == lz4mid; + int const isMid2 = LZ4HC_getCLevelParams(ctx2->compressionLevel).strat == lz4mid; + return !(isMid1 ^ isMid2); +} + static int LZ4HC_compress_generic_dictCtx ( LZ4HC_CCtx_internal* const ctx, @@ -906,7 +1503,7 @@ LZ4HC_compress_generic_dictCtx ( if (position >= 64 KB) { ctx->dictCtx = NULL; return LZ4HC_compress_generic_noDictCtx(ctx, src, dst, srcSizePtr, dstCapacity, cLevel, limit); - } else if (position == 0 && *srcSizePtr > 4 KB) { + } else if (position == 0 && *srcSizePtr > 4 KB && isStateCompatible(ctx, ctx->dictCtx)) { LZ4_memcpy(ctx, ctx->dictCtx, sizeof(LZ4HC_CCtx_internal)); LZ4HC_setExternalDict(ctx, (const BYTE *)src); ctx->compressionLevel = (short)cLevel; @@ -1079,7 +1676,9 @@ int LZ4_loadDictHC (LZ4_streamHC_t* LZ4_streamHCPtr, const char* dictionary, int dictSize) { LZ4HC_CCtx_internal* const ctxPtr = &LZ4_streamHCPtr->internal_donotuse; - DEBUGLOG(4, "LZ4_loadDictHC(ctx:%p, dict:%p, dictSize:%d)", LZ4_streamHCPtr, dictionary, dictSize); + cParams_t cp; + DEBUGLOG(4, "LZ4_loadDictHC(ctx:%p, dict:%p, dictSize:%d, clevel=%d)", LZ4_streamHCPtr, dictionary, dictSize, ctxPtr->compressionLevel); + assert(dictSize >= 0); assert(LZ4_streamHCPtr != NULL); if (dictSize > 64 KB) { dictionary += (size_t)dictSize - 64 KB; @@ -1089,10 +1688,15 @@ int LZ4_loadDictHC (LZ4_streamHC_t* LZ4_streamHCPtr, { int const cLevel = ctxPtr->compressionLevel; LZ4_initStreamHC(LZ4_streamHCPtr, sizeof(*LZ4_streamHCPtr)); LZ4_setCompressionLevel(LZ4_streamHCPtr, cLevel); + cp = LZ4HC_getCLevelParams(cLevel); } LZ4HC_init_internal (ctxPtr, (const BYTE*)dictionary); ctxPtr->end = (const BYTE*)dictionary + dictSize; - if (dictSize >= LZ4HC_HASHSIZE) LZ4HC_Insert (ctxPtr, ctxPtr->end-3); + if (cp.strat == lz4mid) { + LZ4MID_fillHTable (ctxPtr, dictionary, (size_t)dictSize); + } else { + if (dictSize >= LZ4HC_HASHSIZE) LZ4HC_Insert (ctxPtr, ctxPtr->end-3); + } return dictSize; } @@ -1105,8 +1709,10 @@ void LZ4_attach_HC_dictionary(LZ4_streamHC_t *working_stream, const LZ4_streamHC static void LZ4HC_setExternalDict(LZ4HC_CCtx_internal* ctxPtr, const BYTE* newBlock) { DEBUGLOG(4, "LZ4HC_setExternalDict(%p, %p)", ctxPtr, newBlock); - if (ctxPtr->end >= ctxPtr->prefixStart + 4) - LZ4HC_Insert (ctxPtr, ctxPtr->end-3); /* Referencing remaining dictionary content */ + if ( (ctxPtr->end >= ctxPtr->prefixStart + 4) + && (LZ4HC_getCLevelParams(ctxPtr->compressionLevel).strat != lz4mid) ) { + LZ4HC_Insert (ctxPtr, ctxPtr->end-3); /* Referencing remaining dictionary content */ + } /* Only one memory segment for extDict, so any previous extDict is lost at this stage */ ctxPtr->lowLimit = ctxPtr->dictLimit; @@ -1131,7 +1737,8 @@ LZ4_compressHC_continue_generic (LZ4_streamHC_t* LZ4_streamHCPtr, LZ4_streamHCPtr, src, *srcSizePtr, limit); assert(ctxPtr != NULL); /* auto-init if forgotten */ - if (ctxPtr->prefixStart == NULL) LZ4HC_init_internal (ctxPtr, (const BYTE*) src); + if (ctxPtr->prefixStart == NULL) + LZ4HC_init_internal (ctxPtr, (const BYTE*) src); /* Check overflow */ if ((size_t)(ctxPtr->end - ctxPtr->prefixStart) + ctxPtr->dictLimit > 2 GB) { @@ -1176,7 +1783,6 @@ int LZ4_compress_HC_continue_destSize (LZ4_streamHC_t* LZ4_streamHCPtr, const ch } - /* LZ4_saveDictHC : * save history content * into a user-provided buffer @@ -1207,75 +1813,6 @@ int LZ4_saveDictHC (LZ4_streamHC_t* LZ4_streamHCPtr, char* safeBuffer, int dictS } -/*************************************************** -* Deprecated Functions -***************************************************/ - -/* These functions currently generate deprecation warnings */ - -/* Wrappers for deprecated compression functions */ -int LZ4_compressHC(const char* src, char* dst, int srcSize) { return LZ4_compress_HC (src, dst, srcSize, LZ4_compressBound(srcSize), 0); } -int LZ4_compressHC_limitedOutput(const char* src, char* dst, int srcSize, int maxDstSize) { return LZ4_compress_HC(src, dst, srcSize, maxDstSize, 0); } -int LZ4_compressHC2(const char* src, char* dst, int srcSize, int cLevel) { return LZ4_compress_HC (src, dst, srcSize, LZ4_compressBound(srcSize), cLevel); } -int LZ4_compressHC2_limitedOutput(const char* src, char* dst, int srcSize, int maxDstSize, int cLevel) { return LZ4_compress_HC(src, dst, srcSize, maxDstSize, cLevel); } -int LZ4_compressHC_withStateHC (void* state, const char* src, char* dst, int srcSize) { return LZ4_compress_HC_extStateHC (state, src, dst, srcSize, LZ4_compressBound(srcSize), 0); } -int LZ4_compressHC_limitedOutput_withStateHC (void* state, const char* src, char* dst, int srcSize, int maxDstSize) { return LZ4_compress_HC_extStateHC (state, src, dst, srcSize, maxDstSize, 0); } -int LZ4_compressHC2_withStateHC (void* state, const char* src, char* dst, int srcSize, int cLevel) { return LZ4_compress_HC_extStateHC(state, src, dst, srcSize, LZ4_compressBound(srcSize), cLevel); } -int LZ4_compressHC2_limitedOutput_withStateHC (void* state, const char* src, char* dst, int srcSize, int maxDstSize, int cLevel) { return LZ4_compress_HC_extStateHC(state, src, dst, srcSize, maxDstSize, cLevel); } -int LZ4_compressHC_continue (LZ4_streamHC_t* ctx, const char* src, char* dst, int srcSize) { return LZ4_compress_HC_continue (ctx, src, dst, srcSize, LZ4_compressBound(srcSize)); } -int LZ4_compressHC_limitedOutput_continue (LZ4_streamHC_t* ctx, const char* src, char* dst, int srcSize, int maxDstSize) { return LZ4_compress_HC_continue (ctx, src, dst, srcSize, maxDstSize); } - - -/* Deprecated streaming functions */ -int LZ4_sizeofStreamStateHC(void) { return sizeof(LZ4_streamHC_t); } - -/* state is presumed correctly sized, aka >= sizeof(LZ4_streamHC_t) - * @return : 0 on success, !=0 if error */ -int LZ4_resetStreamStateHC(void* state, char* inputBuffer) -{ - LZ4_streamHC_t* const hc4 = LZ4_initStreamHC(state, sizeof(*hc4)); - if (hc4 == NULL) return 1; /* init failed */ - LZ4HC_init_internal (&hc4->internal_donotuse, (const BYTE*)inputBuffer); - return 0; -} - -#if !defined(LZ4_STATIC_LINKING_ONLY_DISABLE_MEMORY_ALLOCATION) -void* LZ4_createHC (const char* inputBuffer) -{ - LZ4_streamHC_t* const hc4 = LZ4_createStreamHC(); - if (hc4 == NULL) return NULL; /* not enough memory */ - LZ4HC_init_internal (&hc4->internal_donotuse, (const BYTE*)inputBuffer); - return hc4; -} - -int LZ4_freeHC (void* LZ4HC_Data) -{ - if (!LZ4HC_Data) return 0; /* support free on NULL */ - FREEMEM(LZ4HC_Data); - return 0; -} -#endif - -int LZ4_compressHC2_continue (void* LZ4HC_Data, const char* src, char* dst, int srcSize, int cLevel) -{ - return LZ4HC_compress_generic (&((LZ4_streamHC_t*)LZ4HC_Data)->internal_donotuse, src, dst, &srcSize, 0, cLevel, notLimited); -} - -int LZ4_compressHC2_limitedOutput_continue (void* LZ4HC_Data, const char* src, char* dst, int srcSize, int dstCapacity, int cLevel) -{ - return LZ4HC_compress_generic (&((LZ4_streamHC_t*)LZ4HC_Data)->internal_donotuse, src, dst, &srcSize, dstCapacity, cLevel, limitedOutput); -} - -char* LZ4_slideInputBufferHC(void* LZ4HC_Data) -{ - LZ4HC_CCtx_internal* const s = &((LZ4_streamHC_t*)LZ4HC_Data)->internal_donotuse; - const BYTE* const bufferStart = s->prefixStart - s->dictLimit + s->lowLimit; - LZ4_resetStreamHC_fast((LZ4_streamHC_t*)LZ4HC_Data, s->compressionLevel); - /* ugly conversion trick, required to evade (const char*) -> (char*) cast-qual warning :( */ - return (char*)(uptrval)bufferStart; -} - - /* ================================================ * LZ4 Optimal parser (levels [LZ4HC_CLEVEL_OPT_MIN - LZ4HC_CLEVEL_MAX]) * ===============================================*/ @@ -1296,7 +1833,6 @@ LZ4_FORCE_INLINE int LZ4HC_literalsPrice(int const litlen) return price; } - /* requires mlen >= MINMATCH */ LZ4_FORCE_INLINE int LZ4HC_sequencePrice(int litlen, int mlen) { @@ -1312,8 +1848,6 @@ LZ4_FORCE_INLINE int LZ4HC_sequencePrice(int litlen, int mlen) return price; } - - LZ4_FORCE_INLINE LZ4HC_match_t LZ4HC_FindLongerMatch(LZ4HC_CCtx_internal* const ctx, const BYTE* ip, const BYTE* const iHighLimit, @@ -1321,19 +1855,27 @@ LZ4HC_FindLongerMatch(LZ4HC_CCtx_internal* const ctx, const dictCtx_directive dict, const HCfavor_e favorDecSpeed) { - LZ4HC_match_t const match0 = { 0 , 0 }; + LZ4HC_match_t const match0 = { 0 , 0, 0 }; /* note : LZ4HC_InsertAndGetWiderMatch() is able to modify the starting position of a match (*startpos), * but this won't be the case here, as we define iLowLimit==ip, ** so LZ4HC_InsertAndGetWiderMatch() won't be allowed to search past ip */ - LZ4HC_match_t md = LZ4HC_InsertAndGetWiderMatch(ctx, ip, ip, iHighLimit, minLen, &ip, nbSearches, 1 /*patternAnalysis*/, 1 /*chainSwap*/, dict, favorDecSpeed); + LZ4HC_match_t md = LZ4HC_InsertAndGetWiderMatch(ctx, ip, ip, iHighLimit, minLen, nbSearches, 1 /*patternAnalysis*/, 1 /*chainSwap*/, dict, favorDecSpeed); + assert(md.back == 0); if (md.len <= minLen) return match0; if (favorDecSpeed) { - if ((md.len>18) & (md.len<=36)) md.len=18; /* favor shortcut */ + if ((md.len>18) & (md.len<=36)) md.len=18; /* favor dec.speed (shortcut) */ } return md; } + +/* preconditions: + * - *srcSizePtr within [1, LZ4_MAX_INPUT_SIZE] + * - src is valid + * - maxOutputSize >= 1 + * - dst is valid + */ static int LZ4HC_compress_optimal ( LZ4HC_CCtx_internal* ctx, const char* const source, char* dst, @@ -1351,7 +1893,7 @@ static int LZ4HC_compress_optimal ( LZ4HC_CCtx_internal* ctx, #if defined(LZ4HC_HEAPMODE) && LZ4HC_HEAPMODE==1 LZ4HC_optimal_t* const opt = (LZ4HC_optimal_t*)ALLOC(sizeof(LZ4HC_optimal_t) * (LZ4_OPT_NUM + TRAILING_LITERALS)); #else - LZ4HC_optimal_t opt[LZ4_OPT_NUM + TRAILING_LITERALS]; /* ~64 KB, which is a bit large for stack... */ + LZ4HC_optimal_t opt[LZ4_OPT_NUM + TRAILING_LITERALS]; /* ~64 KB, which can be a bit large for some stacks... */ #endif const BYTE* ip = (const BYTE*) source; @@ -1366,10 +1908,17 @@ static int LZ4HC_compress_optimal ( LZ4HC_CCtx_internal* ctx, int ovoff = 0; /* init */ + DEBUGLOG(5, "LZ4HC_compress_optimal(dst=%p, dstCapa=%u)", dst, (unsigned)dstCapacity); #if defined(LZ4HC_HEAPMODE) && LZ4HC_HEAPMODE==1 if (opt == NULL) goto _return_label; #endif - DEBUGLOG(5, "LZ4HC_compress_optimal(dst=%p, dstCapa=%u)", dst, (unsigned)dstCapacity); + + /* preconditions verifications */ + assert(dstCapacity > 0); + assert(dst != NULL); + assert(*srcSizePtr > 0); + assert(source != NULL); + *srcSizePtr = 0; if (limit == fillOutput) oend -= LASTLITERALS; /* Hack for support LZ4 format restriction */ if (sufficient_len >= LZ4_OPT_NUM) sufficient_len = LZ4_OPT_NUM-1; @@ -1407,11 +1956,11 @@ static int LZ4HC_compress_optimal ( LZ4HC_CCtx_internal* ctx, rPos, cost, opt[rPos].litlen); } } /* set prices using initial match */ - { int mlen = MINMATCH; - int const matchML = firstMatch.len; /* necessarily < sufficient_len < LZ4_OPT_NUM */ + { int const matchML = firstMatch.len; /* necessarily < sufficient_len < LZ4_OPT_NUM */ int const offset = firstMatch.off; + int mlen; assert(matchML < LZ4_OPT_NUM); - for ( ; mlen <= matchML ; mlen++) { + for (mlen = MINMATCH ; mlen <= matchML ; mlen++) { int const cost = LZ4HC_sequencePrice(llen, mlen); opt[mlen].mlen = mlen; opt[mlen].off = offset; @@ -1631,7 +2180,76 @@ if (limit == fillOutput) { } _return_label: #if defined(LZ4HC_HEAPMODE) && LZ4HC_HEAPMODE==1 - FREEMEM(opt); + if (opt) FREEMEM(opt); #endif return retval; } + + +/*************************************************** +* Deprecated Functions +***************************************************/ + +/* These functions currently generate deprecation warnings */ + +/* Wrappers for deprecated compression functions */ +int LZ4_compressHC(const char* src, char* dst, int srcSize) { return LZ4_compress_HC (src, dst, srcSize, LZ4_compressBound(srcSize), 0); } +int LZ4_compressHC_limitedOutput(const char* src, char* dst, int srcSize, int maxDstSize) { return LZ4_compress_HC(src, dst, srcSize, maxDstSize, 0); } +int LZ4_compressHC2(const char* src, char* dst, int srcSize, int cLevel) { return LZ4_compress_HC (src, dst, srcSize, LZ4_compressBound(srcSize), cLevel); } +int LZ4_compressHC2_limitedOutput(const char* src, char* dst, int srcSize, int maxDstSize, int cLevel) { return LZ4_compress_HC(src, dst, srcSize, maxDstSize, cLevel); } +int LZ4_compressHC_withStateHC (void* state, const char* src, char* dst, int srcSize) { return LZ4_compress_HC_extStateHC (state, src, dst, srcSize, LZ4_compressBound(srcSize), 0); } +int LZ4_compressHC_limitedOutput_withStateHC (void* state, const char* src, char* dst, int srcSize, int maxDstSize) { return LZ4_compress_HC_extStateHC (state, src, dst, srcSize, maxDstSize, 0); } +int LZ4_compressHC2_withStateHC (void* state, const char* src, char* dst, int srcSize, int cLevel) { return LZ4_compress_HC_extStateHC(state, src, dst, srcSize, LZ4_compressBound(srcSize), cLevel); } +int LZ4_compressHC2_limitedOutput_withStateHC (void* state, const char* src, char* dst, int srcSize, int maxDstSize, int cLevel) { return LZ4_compress_HC_extStateHC(state, src, dst, srcSize, maxDstSize, cLevel); } +int LZ4_compressHC_continue (LZ4_streamHC_t* ctx, const char* src, char* dst, int srcSize) { return LZ4_compress_HC_continue (ctx, src, dst, srcSize, LZ4_compressBound(srcSize)); } +int LZ4_compressHC_limitedOutput_continue (LZ4_streamHC_t* ctx, const char* src, char* dst, int srcSize, int maxDstSize) { return LZ4_compress_HC_continue (ctx, src, dst, srcSize, maxDstSize); } + + +/* Deprecated streaming functions */ +int LZ4_sizeofStreamStateHC(void) { return sizeof(LZ4_streamHC_t); } + +/* state is presumed correctly sized, aka >= sizeof(LZ4_streamHC_t) + * @return : 0 on success, !=0 if error */ +int LZ4_resetStreamStateHC(void* state, char* inputBuffer) +{ + LZ4_streamHC_t* const hc4 = LZ4_initStreamHC(state, sizeof(*hc4)); + if (hc4 == NULL) return 1; /* init failed */ + LZ4HC_init_internal (&hc4->internal_donotuse, (const BYTE*)inputBuffer); + return 0; +} + +#if !defined(LZ4_STATIC_LINKING_ONLY_DISABLE_MEMORY_ALLOCATION) +void* LZ4_createHC (const char* inputBuffer) +{ + LZ4_streamHC_t* const hc4 = LZ4_createStreamHC(); + if (hc4 == NULL) return NULL; /* not enough memory */ + LZ4HC_init_internal (&hc4->internal_donotuse, (const BYTE*)inputBuffer); + return hc4; +} + +int LZ4_freeHC (void* LZ4HC_Data) +{ + if (!LZ4HC_Data) return 0; /* support free on NULL */ + FREEMEM(LZ4HC_Data); + return 0; +} +#endif + +int LZ4_compressHC2_continue (void* LZ4HC_Data, const char* src, char* dst, int srcSize, int cLevel) +{ + return LZ4HC_compress_generic (&((LZ4_streamHC_t*)LZ4HC_Data)->internal_donotuse, src, dst, &srcSize, 0, cLevel, notLimited); +} + +int LZ4_compressHC2_limitedOutput_continue (void* LZ4HC_Data, const char* src, char* dst, int srcSize, int dstCapacity, int cLevel) +{ + return LZ4HC_compress_generic (&((LZ4_streamHC_t*)LZ4HC_Data)->internal_donotuse, src, dst, &srcSize, dstCapacity, cLevel, limitedOutput); +} + +char* LZ4_slideInputBufferHC(void* LZ4HC_Data) +{ + LZ4HC_CCtx_internal* const s = &((LZ4_streamHC_t*)LZ4HC_Data)->internal_donotuse; + const BYTE* const bufferStart = s->prefixStart - s->dictLimit + s->lowLimit; + LZ4_resetStreamHC_fast((LZ4_streamHC_t*)LZ4HC_Data, s->compressionLevel); + /* ugly conversion trick, required to evade (const char*) -> (char*) cast-qual warning :( */ + return (char*)(uptrval)bufferStart; +} diff --git a/CompressSave/LZ4WrapC/lz4/lz4hc.h b/CompressSave/LZ4WrapC/lz4/lz4hc.h index e937acf..17beed9 100644 --- a/CompressSave/LZ4WrapC/lz4/lz4hc.h +++ b/CompressSave/LZ4WrapC/lz4/lz4hc.h @@ -1,7 +1,7 @@ /* LZ4 HC - High Compression Mode of LZ4 Header File - Copyright (C) 2011-2020, Yann Collet. + Copyright (c) Yann Collet. All rights reserved. BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) Redistribution and use in source and binary forms, with or without @@ -44,7 +44,7 @@ extern "C" { /* --- Useful constants --- */ -#define LZ4HC_CLEVEL_MIN 3 +#define LZ4HC_CLEVEL_MIN 2 #define LZ4HC_CLEVEL_DEFAULT 9 #define LZ4HC_CLEVEL_OPT_MIN 10 #define LZ4HC_CLEVEL_MAX 12 @@ -126,6 +126,8 @@ LZ4LIB_API int LZ4_freeStreamHC (LZ4_streamHC_t* streamHCPtr); After reset, a first "fictional block" can be designated as initial dictionary, using LZ4_loadDictHC() (Optional). + Note: In order for LZ4_loadDictHC() to create the correct data structure, + it is essential to set the compression level _before_ loading the dictionary. Invoke LZ4_compress_HC_continue() to compress each successive block. The number of blocks is unlimited. @@ -135,12 +137,12 @@ LZ4LIB_API int LZ4_freeStreamHC (LZ4_streamHC_t* streamHCPtr); It's allowed to update compression level anytime between blocks, using LZ4_setCompressionLevel() (experimental). - 'dst' buffer should be sized to handle worst case scenarios + @dst buffer should be sized to handle worst case scenarios (see LZ4_compressBound(), it ensures compression success). In case of failure, the API does not guarantee recovery, so the state _must_ be reset. To ensure compression success - whenever `dst` buffer size cannot be made >= LZ4_compressBound(), + whenever @dst buffer size cannot be made >= LZ4_compressBound(), consider using LZ4_compress_HC_continue_destSize(). Whenever previous input blocks can't be preserved unmodified in-place during compression of next blocks, @@ -176,6 +178,34 @@ LZ4LIB_API int LZ4_compress_HC_continue_destSize(LZ4_streamHC_t* LZ4_streamHCPtr LZ4LIB_API int LZ4_saveDictHC (LZ4_streamHC_t* streamHCPtr, char* safeBuffer, int maxDictSize); +/*! LZ4_attach_HC_dictionary() : stable since v1.10.0 + * This API allows for the efficient re-use of a static dictionary many times. + * + * Rather than re-loading the dictionary buffer into a working context before + * each compression, or copying a pre-loaded dictionary's LZ4_streamHC_t into a + * working LZ4_streamHC_t, this function introduces a no-copy setup mechanism, + * in which the working stream references the dictionary stream in-place. + * + * Several assumptions are made about the state of the dictionary stream. + * Currently, only streams which have been prepared by LZ4_loadDictHC() should + * be expected to work. + * + * Alternatively, the provided dictionary stream pointer may be NULL, in which + * case any existing dictionary stream is unset. + * + * A dictionary should only be attached to a stream without any history (i.e., + * a stream that has just been reset). + * + * The dictionary will remain attached to the working stream only for the + * current stream session. Calls to LZ4_resetStreamHC(_fast) will remove the + * dictionary context association from the working stream. The dictionary + * stream (and source buffer) must remain in-place / accessible / unchanged + * through the lifetime of the stream session. + */ +LZ4LIB_API void +LZ4_attach_HC_dictionary(LZ4_streamHC_t* working_stream, + const LZ4_streamHC_t* dictionary_stream); + /*^********************************************** * !!!!!! STATIC LINKING ONLY !!!!!! @@ -204,18 +234,18 @@ LZ4LIB_API int LZ4_saveDictHC (LZ4_streamHC_t* streamHCPtr, char* safeBuffer, in typedef struct LZ4HC_CCtx_internal LZ4HC_CCtx_internal; struct LZ4HC_CCtx_internal { - LZ4_u32 hashTable[LZ4HC_HASHTABLESIZE]; - LZ4_u16 chainTable[LZ4HC_MAXD]; - const LZ4_byte* end; /* next block here to continue on current prefix */ + LZ4_u32 hashTable[LZ4HC_HASHTABLESIZE]; + LZ4_u16 chainTable[LZ4HC_MAXD]; + const LZ4_byte* end; /* next block here to continue on current prefix */ const LZ4_byte* prefixStart; /* Indexes relative to this position */ const LZ4_byte* dictStart; /* alternate reference for extDict */ - LZ4_u32 dictLimit; /* below that point, need extDict */ - LZ4_u32 lowLimit; /* below that point, no more dict */ - LZ4_u32 nextToUpdate; /* index from which to continue dictionary update */ - short compressionLevel; - LZ4_i8 favorDecSpeed; /* favor decompression speed if this flag set, - otherwise, favor compression ratio */ - LZ4_i8 dirty; /* stream has to be fully reset if this flag is set */ + LZ4_u32 dictLimit; /* below that point, need extDict */ + LZ4_u32 lowLimit; /* below that point, no more history */ + LZ4_u32 nextToUpdate; /* index from which to continue dictionary update */ + short compressionLevel; + LZ4_i8 favorDecSpeed; /* favor decompression speed if this flag set, + otherwise, favor compression ratio */ + LZ4_i8 dirty; /* stream has to be fully reset if this flag is set */ const LZ4HC_CCtx_internal* dictCtx; }; @@ -376,35 +406,6 @@ LZ4LIB_STATIC_API int LZ4_compress_HC_extStateHC_fastReset ( int srcSize, int dstCapacity, int compressionLevel); -/*! LZ4_attach_HC_dictionary() : - * This is an experimental API that allows for the efficient use of a - * static dictionary many times. - * - * Rather than re-loading the dictionary buffer into a working context before - * each compression, or copying a pre-loaded dictionary's LZ4_streamHC_t into a - * working LZ4_streamHC_t, this function introduces a no-copy setup mechanism, - * in which the working stream references the dictionary stream in-place. - * - * Several assumptions are made about the state of the dictionary stream. - * Currently, only streams which have been prepared by LZ4_loadDictHC() should - * be expected to work. - * - * Alternatively, the provided dictionary stream pointer may be NULL, in which - * case any existing dictionary stream is unset. - * - * A dictionary should only be attached to a stream without any history (i.e., - * a stream that has just been reset). - * - * The dictionary will remain attached to the working stream only for the - * current stream session. Calls to LZ4_resetStreamHC(_fast) will remove the - * dictionary context association from the working stream. The dictionary - * stream (and source buffer) must remain in-place / accessible / unchanged - * through the lifetime of the stream session. - */ -LZ4LIB_STATIC_API void LZ4_attach_HC_dictionary( - LZ4_streamHC_t *working_stream, - const LZ4_streamHC_t *dictionary_stream); - #if defined (__cplusplus) } #endif diff --git a/CompressSave/LZ4WrapC/lz4/xxhash.c b/CompressSave/LZ4WrapC/lz4/xxhash.c index ff28749..e132619 100644 --- a/CompressSave/LZ4WrapC/lz4/xxhash.c +++ b/CompressSave/LZ4WrapC/lz4/xxhash.c @@ -120,12 +120,12 @@ static void* XXH_memcpy(void* dest, const void* src, size_t size) { return memcp /* ************************************* * Compiler Specific Options ***************************************/ -#ifdef _MSC_VER /* Visual Studio */ +#if defined (_MSC_VER) && !defined (__clang__) /* MSVC */ # pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */ # define FORCE_INLINE static __forceinline #else # if defined (__cplusplus) || defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 */ -# ifdef __GNUC__ +# if defined (__GNUC__) || defined (__clang__) # define FORCE_INLINE static inline __attribute__((always_inline)) # else # define FORCE_INLINE static inline @@ -213,7 +213,7 @@ static U32 XXH_swap32 (U32 x) /* ************************************* * Architecture Macros ***************************************/ -typedef enum { XXH_bigEndian=0, XXH_littleEndian=1 } XXH_endianess; +typedef enum { XXH_bigEndian=0, XXH_littleEndian=1 } XXH_endianness; /* XXH_CPU_LITTLE_ENDIAN can be defined externally, for example on the compiler command line */ #ifndef XXH_CPU_LITTLE_ENDIAN @@ -231,7 +231,7 @@ static int XXH_isLittleEndian(void) *****************************/ typedef enum { XXH_aligned, XXH_unaligned } XXH_alignment; -FORCE_INLINE U32 XXH_readLE32_align(const void* ptr, XXH_endianess endian, XXH_alignment align) +FORCE_INLINE U32 XXH_readLE32_align(const void* ptr, XXH_endianness endian, XXH_alignment align) { if (align==XXH_unaligned) return endian==XXH_littleEndian ? XXH_read32(ptr) : XXH_swap32(XXH_read32(ptr)); @@ -239,7 +239,7 @@ FORCE_INLINE U32 XXH_readLE32_align(const void* ptr, XXH_endianess endian, XXH_a return endian==XXH_littleEndian ? *(const U32*)ptr : XXH_swap32(*(const U32*)ptr); } -FORCE_INLINE U32 XXH_readLE32(const void* ptr, XXH_endianess endian) +FORCE_INLINE U32 XXH_readLE32(const void* ptr, XXH_endianness endian) { return XXH_readLE32_align(ptr, endian, XXH_unaligned); } @@ -289,7 +289,7 @@ static U32 XXH32_avalanche(U32 h32) static U32 XXH32_finalize(U32 h32, const void* ptr, size_t len, - XXH_endianess endian, XXH_alignment align) + XXH_endianness endian, XXH_alignment align) { const BYTE* p = (const BYTE*)ptr; @@ -350,7 +350,7 @@ XXH32_finalize(U32 h32, const void* ptr, size_t len, FORCE_INLINE U32 XXH32_endian_align(const void* input, size_t len, U32 seed, - XXH_endianess endian, XXH_alignment align) + XXH_endianness endian, XXH_alignment align) { const BYTE* p = (const BYTE*)input; const BYTE* bEnd = p + len; @@ -398,7 +398,7 @@ XXH_PUBLIC_API unsigned int XXH32 (const void* input, size_t len, unsigned int s XXH32_update(&state, input, len); return XXH32_digest(&state); #else - XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; + XXH_endianness endian_detected = (XXH_endianness)XXH_CPU_LITTLE_ENDIAN; if (XXH_FORCE_ALIGN_CHECK) { if ((((size_t)input) & 3) == 0) { /* Input is 4-bytes aligned, leverage the speed benefit */ @@ -449,7 +449,7 @@ XXH_PUBLIC_API XXH_errorcode XXH32_reset(XXH32_state_t* statePtr, unsigned int s FORCE_INLINE XXH_errorcode -XXH32_update_endian(XXH32_state_t* state, const void* input, size_t len, XXH_endianess endian) +XXH32_update_endian(XXH32_state_t* state, const void* input, size_t len, XXH_endianness endian) { if (input==NULL) #if defined(XXH_ACCEPT_NULL_INPUT_POINTER) && (XXH_ACCEPT_NULL_INPUT_POINTER>=1) @@ -514,7 +514,7 @@ XXH32_update_endian(XXH32_state_t* state, const void* input, size_t len, XXH_end XXH_PUBLIC_API 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; + XXH_endianness endian_detected = (XXH_endianness)XXH_CPU_LITTLE_ENDIAN; if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) return XXH32_update_endian(state_in, input, len, XXH_littleEndian); @@ -524,7 +524,7 @@ XXH_PUBLIC_API XXH_errorcode XXH32_update (XXH32_state_t* state_in, const void* FORCE_INLINE U32 -XXH32_digest_endian (const XXH32_state_t* state, XXH_endianess endian) +XXH32_digest_endian (const XXH32_state_t* state, XXH_endianness endian) { U32 h32; @@ -545,7 +545,7 @@ XXH32_digest_endian (const XXH32_state_t* state, XXH_endianess endian) XXH_PUBLIC_API unsigned int XXH32_digest (const XXH32_state_t* state_in) { - XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; + XXH_endianness endian_detected = (XXH_endianness)XXH_CPU_LITTLE_ENDIAN; if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) return XXH32_digest_endian(state_in, XXH_littleEndian); @@ -642,7 +642,7 @@ static U64 XXH_swap64 (U64 x) } #endif -FORCE_INLINE U64 XXH_readLE64_align(const void* ptr, XXH_endianess endian, XXH_alignment align) +FORCE_INLINE U64 XXH_readLE64_align(const void* ptr, XXH_endianness endian, XXH_alignment align) { if (align==XXH_unaligned) return endian==XXH_littleEndian ? XXH_read64(ptr) : XXH_swap64(XXH_read64(ptr)); @@ -650,7 +650,7 @@ FORCE_INLINE U64 XXH_readLE64_align(const void* ptr, XXH_endianess endian, XXH_a return endian==XXH_littleEndian ? *(const U64*)ptr : XXH_swap64(*(const U64*)ptr); } -FORCE_INLINE U64 XXH_readLE64(const void* ptr, XXH_endianess endian) +FORCE_INLINE U64 XXH_readLE64(const void* ptr, XXH_endianness endian) { return XXH_readLE64_align(ptr, endian, XXH_unaligned); } @@ -700,7 +700,7 @@ static U64 XXH64_avalanche(U64 h64) static U64 XXH64_finalize(U64 h64, const void* ptr, size_t len, - XXH_endianess endian, XXH_alignment align) + XXH_endianness endian, XXH_alignment align) { const BYTE* p = (const BYTE*)ptr; @@ -809,7 +809,7 @@ XXH64_finalize(U64 h64, const void* ptr, size_t len, FORCE_INLINE U64 XXH64_endian_align(const void* input, size_t len, U64 seed, - XXH_endianess endian, XXH_alignment align) + XXH_endianness endian, XXH_alignment align) { const BYTE* p = (const BYTE*)input; const BYTE* bEnd = p + len; @@ -861,7 +861,7 @@ XXH_PUBLIC_API unsigned long long XXH64 (const void* input, size_t len, unsigned XXH64_update(&state, input, len); return XXH64_digest(&state); #else - XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; + XXH_endianness endian_detected = (XXH_endianness)XXH_CPU_LITTLE_ENDIAN; if (XXH_FORCE_ALIGN_CHECK) { if ((((size_t)input) & 7)==0) { /* Input is aligned, let's leverage the speed advantage */ @@ -909,7 +909,7 @@ XXH_PUBLIC_API XXH_errorcode XXH64_reset(XXH64_state_t* statePtr, unsigned long } FORCE_INLINE XXH_errorcode -XXH64_update_endian (XXH64_state_t* state, const void* input, size_t len, XXH_endianess endian) +XXH64_update_endian (XXH64_state_t* state, const void* input, size_t len, XXH_endianness endian) { if (input==NULL) #if defined(XXH_ACCEPT_NULL_INPUT_POINTER) && (XXH_ACCEPT_NULL_INPUT_POINTER>=1) @@ -970,7 +970,7 @@ XXH64_update_endian (XXH64_state_t* state, const void* input, size_t len, XXH_en XXH_PUBLIC_API 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; + XXH_endianness endian_detected = (XXH_endianness)XXH_CPU_LITTLE_ENDIAN; if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) return XXH64_update_endian(state_in, input, len, XXH_littleEndian); @@ -978,7 +978,7 @@ XXH_PUBLIC_API XXH_errorcode XXH64_update (XXH64_state_t* state_in, const void* return XXH64_update_endian(state_in, input, len, XXH_bigEndian); } -FORCE_INLINE U64 XXH64_digest_endian (const XXH64_state_t* state, XXH_endianess endian) +FORCE_INLINE U64 XXH64_digest_endian (const XXH64_state_t* state, XXH_endianness endian) { U64 h64; @@ -1004,7 +1004,7 @@ FORCE_INLINE U64 XXH64_digest_endian (const XXH64_state_t* state, XXH_endianess XXH_PUBLIC_API unsigned long long XXH64_digest (const XXH64_state_t* state_in) { - XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; + XXH_endianness endian_detected = (XXH_endianness)XXH_CPU_LITTLE_ENDIAN; if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) return XXH64_digest_endian(state_in, XXH_littleEndian); diff --git a/CompressSave/NoneWrapC/CMakeLists.txt b/CompressSave/NoneWrapC/CMakeLists.txt index 5f5fce5..e3defad 100644 --- a/CompressSave/NoneWrapC/CMakeLists.txt +++ b/CompressSave/NoneWrapC/CMakeLists.txt @@ -1,4 +1,4 @@ -cmake_minimum_required(VERSION 3.2) +cmake_minimum_required(VERSION 3.12) project(nonewrap) diff --git a/CompressSave/NoneWrapC/nonewrap.c b/CompressSave/NoneWrapC/nonewrap.c index 85bcc34..4ce512a 100644 --- a/CompressSave/NoneWrapC/nonewrap.c +++ b/CompressSave/NoneWrapC/nonewrap.c @@ -10,7 +10,7 @@ size_t __stdcall CompressBufferBound(size_t inBufferSize) return inBufferSize * 4; } -size_t __stdcall CompressBegin(Context** pctx, int compressionLevel, void* outBuff, size_t outCapacity, void* dict, size_t dictSize) +size_t __stdcall CompressBegin(Context **pctx, int compressionLevel, void *outBuff, size_t outCapacity, void *dict, size_t dictSize) { Context *ctx = (Context *)malloc(sizeof(Context)); if (ctx == NULL) return -1; @@ -18,23 +18,23 @@ size_t __stdcall CompressBegin(Context** pctx, int compressionLevel, void* outBu return 0; } -size_t __stdcall CompressUpdate(Context* ctx,void* dstBuffer, size_t dstCapacity,const void* srcBuffer, size_t srcSize) +size_t __stdcall CompressUpdate(Context *ctx, void *dstBuffer, size_t dstCapacity, const void *srcBuffer, size_t srcSize) { memcpy(dstBuffer, srcBuffer, srcSize); return srcSize; } -size_t __stdcall CompressEnd(Context* ctx, void* dstBuffer, size_t dstCapacity) +size_t __stdcall CompressEnd(Context *ctx, void *dstBuffer, size_t dstCapacity) { return 0; } -void __stdcall CompressContextFree(Context* ctx) +void __stdcall CompressContextFree(Context *ctx) { free(ctx); } -size_t __stdcall DecompressBegin(Context **pdctx,void *inBuffer,size_t *inBufferSize, size_t *blockSize, void* dict, size_t dictSize) +size_t __stdcall DecompressBegin(Context **pdctx, void *inBuffer, size_t *inBufferSize, size_t *blockSize, void *dict, size_t dictSize) { Context *ctx = (Context *)malloc(sizeof(Context)); if (ctx == NULL) return -1; @@ -44,19 +44,19 @@ size_t __stdcall DecompressBegin(Context **pdctx,void *inBuffer,size_t *inBuffer return 0; } -void __stdcall DecompressContextReset(Context* dctx) +void __stdcall DecompressContextReset(Context *dctx) { } -size_t __stdcall DecompressUpdate(Context* dctx, void* outBuffer, size_t * outBufferSize, void* inBuffer, size_t * inBufferSize) +size_t __stdcall DecompressUpdate(Context *dctx, void *outBuffer, size_t *outBufferSize, void *inBuffer, size_t *inBufferSize) { memcpy(outBuffer, inBuffer, *inBufferSize); *outBufferSize = *inBufferSize; return 1; } -size_t __stdcall DecompressEnd(Context* ctx) +size_t __stdcall DecompressEnd(Context *ctx) { free(ctx); return 0; -} \ No newline at end of file +} diff --git a/CompressSave/ZstdWrapC/CMakeLists.txt b/CompressSave/ZstdWrapC/CMakeLists.txt index 98c364c..841ba31 100644 --- a/CompressSave/ZstdWrapC/CMakeLists.txt +++ b/CompressSave/ZstdWrapC/CMakeLists.txt @@ -1,4 +1,4 @@ -cmake_minimum_required(VERSION 3.2) +cmake_minimum_required(VERSION 3.12) project(zstdwrap) diff --git a/CompressSave/ZstdWrapC/zstdwrap.c b/CompressSave/ZstdWrapC/zstdwrap.c index c28c539..7001620 100644 --- a/CompressSave/ZstdWrapC/zstdwrap.c +++ b/CompressSave/ZstdWrapC/zstdwrap.c @@ -10,65 +10,61 @@ size_t __stdcall CompressBufferBound(size_t inBufferSize) return ZSTD_COMPRESSBOUND(inBufferSize); } -size_t __stdcall CompressBegin(ZSTD_CStream** pctx, int compressionLevel, void* outBuff, size_t outCapacity, void* dict, size_t dictSize) +size_t __stdcall CompressBegin(ZSTD_CStream **pctx, int compressionLevel, void *outBuff, size_t outCapacity, void *dict, size_t dictSize) { ZSTD_CStream *ctx = ZSTD_createCStream(); - if (ctx == NULL) return -1; + if (ctx == NULL) + return -1; ZSTD_CCtx_setParameter(ctx, ZSTD_c_compressionLevel, compressionLevel); if (dict) - { ZSTD_CCtx_loadDictionary(ctx, dict, dictSize); - } *pctx = ctx; return 0; } - -size_t __stdcall CompressUpdate(ZSTD_CStream* ctx,void* dstBuffer, size_t dstCapacity,const void* srcBuffer, size_t srcSize) +size_t __stdcall CompressUpdate(ZSTD_CStream *ctx, void *dstBuffer, size_t dstCapacity, const void *srcBuffer, size_t srcSize) { ZSTD_outBuffer obuf = {dstBuffer, dstCapacity, 0}; ZSTD_inBuffer ibuf = {srcBuffer, srcSize, 0}; do { ZSTD_compressStream2(ctx, &obuf, &ibuf, ZSTD_e_continue); - } - while (ibuf.pos < ibuf.size); + } while (ibuf.pos < ibuf.size); return obuf.pos; } -size_t __stdcall CompressEnd(ZSTD_CStream* ctx, void* dstBuffer, size_t dstCapacity) +size_t __stdcall CompressEnd(ZSTD_CStream *ctx, void *dstBuffer, size_t dstCapacity) { ZSTD_outBuffer obuf = {dstBuffer, dstCapacity, 0}; ZSTD_inBuffer ibuf = {NULL, 0, 0}; - while (ZSTD_compressStream2(ctx, &obuf, &ibuf, ZSTD_e_end) > 0) {} + while (ZSTD_compressStream2(ctx, &obuf, &ibuf, ZSTD_e_end) > 0) { } return obuf.pos; } -void __stdcall CompressContextFree(ZSTD_CStream* ctx) +void __stdcall CompressContextFree(ZSTD_CStream *ctx) { ZSTD_freeCStream(ctx); } -size_t __stdcall DecompressBegin(ZSTD_DStream **pdctx,void *inBuffer,size_t *inBufferSize, size_t *blockSize, void* dict, size_t dictSize) +size_t __stdcall DecompressBegin(ZSTD_DStream **pdctx, void *inBuffer, size_t *inBufferSize, size_t *blockSize, void *dict, size_t dictSize) { ZSTD_DStream *ctx = ZSTD_createDStream(); - if (ctx == NULL) return -1; + if (ctx == NULL) + return -1; if (dict) - { ZSTD_DCtx_loadDictionary(ctx, dict, dictSize); - } *pdctx = ctx; *inBufferSize = 0; *blockSize = ZSTD_DStreamOutSize() << 2; return 0; } -void __stdcall DecompressContextReset(ZSTD_DStream* dctx) +void __stdcall DecompressContextReset(ZSTD_DStream *dctx) { ZSTD_DCtx_reset(dctx, ZSTD_reset_session_only); } -size_t __stdcall DecompressUpdate(ZSTD_DStream* dctx, void* outBuffer, size_t * outBufferSize, void* inBuffer, size_t * inBufferSize) +size_t __stdcall DecompressUpdate(ZSTD_DStream *dctx, void *outBuffer, size_t *outBufferSize, void *inBuffer, size_t *inBufferSize) { ZSTD_outBuffer obuf = {outBuffer, *outBufferSize, 0}; ZSTD_inBuffer ibuf = {inBuffer, *inBufferSize, 0}; @@ -78,7 +74,7 @@ size_t __stdcall DecompressUpdate(ZSTD_DStream* dctx, void* outBuffer, size_t * return r; } -size_t __stdcall DecompressEnd(ZSTD_DStream* ctx) +size_t __stdcall DecompressEnd(ZSTD_DStream *ctx) { return ZSTD_freeDStream(ctx); -} \ No newline at end of file +} diff --git a/CompressSave/package/plugins/x64/lz4wrap.dll b/CompressSave/package/plugins/x64/lz4wrap.dll index 28b26b9..195095b 100644 Binary files a/CompressSave/package/plugins/x64/lz4wrap.dll and b/CompressSave/package/plugins/x64/lz4wrap.dll differ diff --git a/CompressSave/package/plugins/x64/nonewrap.dll b/CompressSave/package/plugins/x64/nonewrap.dll index 846e38f..cc62e3a 100644 Binary files a/CompressSave/package/plugins/x64/nonewrap.dll and b/CompressSave/package/plugins/x64/nonewrap.dll differ diff --git a/CompressSave/package/plugins/x64/zstdwrap.dll b/CompressSave/package/plugins/x64/zstdwrap.dll index 2f1e2fc..efa7340 100644 Binary files a/CompressSave/package/plugins/x64/zstdwrap.dll and b/CompressSave/package/plugins/x64/zstdwrap.dll differ