IH334Q5ACWE4TNQYYOOF6GWV6CRXOEM6542NVNPA6HRIZ3CBFKEAC
/*
* Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
* LICENSE file in the root directory of this source tree) and the GPLv2 (found
* in the COPYING file in the root directory of this source tree).
* You may select, at your option, one of the above-listed licenses.
*/
#if defined (__cplusplus)
extern "C" {
#endif
#ifndef ZSTD_H_235446
#define ZSTD_H_235446
/* ====== Dependency ======*/
#include <limits.h> /* INT_MAX */
#include <stddef.h> /* size_t */
/* ===== ZSTDLIB_API : control library symbols visibility ===== */
#ifndef ZSTDLIB_VISIBILITY
# if defined(__GNUC__) && (__GNUC__ >= 4)
# define ZSTDLIB_VISIBILITY __attribute__ ((visibility ("default")))
# else
# define ZSTDLIB_VISIBILITY
# endif
#endif
#if defined(ZSTD_DLL_EXPORT) && (ZSTD_DLL_EXPORT==1)
# define ZSTDLIB_API __declspec(dllexport) ZSTDLIB_VISIBILITY
#elif defined(ZSTD_DLL_IMPORT) && (ZSTD_DLL_IMPORT==1)
# define ZSTDLIB_API __declspec(dllimport) ZSTDLIB_VISIBILITY /* It isn't required but allows to generate better code, saving a function pointer load from the IAT and an indirect jump.*/
#else
# define ZSTDLIB_API ZSTDLIB_VISIBILITY
#endif
/*******************************************************************************
Introduction
zstd, short for Zstandard, is a fast lossless compression algorithm, targeting
real-time compression scenarios at zlib-level and better compression ratios.
The zstd compression library provides in-memory compression and decompression
functions.
The library supports regular compression levels from 1 up to ZSTD_maxCLevel(),
which is currently 22. Levels >= 20, labeled `--ultra`, should be used with
caution, as they require more memory. The library also offers negative
compression levels, which extend the range of speed vs. ratio preferences.
The lower the level, the faster the speed (at the cost of compression).
Compression can be done in:
- a single step (described as Simple API)
- a single step, reusing a context (described as Explicit context)
- unbounded multiple steps (described as Streaming compression)
The compression ratio achievable on small data can be highly improved using
a dictionary. Dictionary compression can be performed in:
- a single step (described as Simple dictionary API)
- a single step, reusing a dictionary (described as Bulk-processing
dictionary API)
Advanced experimental functions can be accessed using
`#define ZSTD_STATIC_LINKING_ONLY` before including zstd.h.
Advanced experimental APIs should never be used with a dynamically-linked
library. They are not "stable"; their definitions or signatures may change in
the future. Only static linking is allowed.
*******************************************************************************/
/*------ Version ------*/
#define ZSTD_VERSION_MAJOR 1
#define ZSTD_VERSION_MINOR 4
#define ZSTD_VERSION_RELEASE 4
#define ZSTD_VERSION_NUMBER (ZSTD_VERSION_MAJOR *100*100 + ZSTD_VERSION_MINOR *100 + ZSTD_VERSION_RELEASE)
ZSTDLIB_API unsigned ZSTD_versionNumber(void); /**< to check runtime library version */
#define ZSTD_LIB_VERSION ZSTD_VERSION_MAJOR.ZSTD_VERSION_MINOR.ZSTD_VERSION_RELEASE
#define ZSTD_QUOTE(str) #str
#define ZSTD_EXPAND_AND_QUOTE(str) ZSTD_QUOTE(str)
#define ZSTD_VERSION_STRING ZSTD_EXPAND_AND_QUOTE(ZSTD_LIB_VERSION)
ZSTDLIB_API const char* ZSTD_versionString(void); /* requires v1.3.0+ */
/* *************************************
* Default constant
***************************************/
#ifndef ZSTD_CLEVEL_DEFAULT
# define ZSTD_CLEVEL_DEFAULT 3
#endif
/* *************************************
* Constants
***************************************/
/* All magic numbers are supposed read/written to/from files/memory using little-endian convention */
#define ZSTD_MAGICNUMBER 0xFD2FB528 /* valid since v0.8.0 */
#define ZSTD_MAGIC_DICTIONARY 0xEC30A437 /* valid since v0.7.0 */
#define ZSTD_MAGIC_SKIPPABLE_START 0x184D2A50 /* all 16 values, from 0x184D2A50 to 0x184D2A5F, signal the beginning of a skippable frame */
#define ZSTD_MAGIC_SKIPPABLE_MASK 0xFFFFFFF0
#define ZSTD_BLOCKSIZELOG_MAX 17
#define ZSTD_BLOCKSIZE_MAX (1<<ZSTD_BLOCKSIZELOG_MAX)
/***************************************
* Simple API
***************************************/
/*! ZSTD_compress() :
* Compresses `src` content as a single zstd compressed frame into already allocated `dst`.
* Hint : compression runs faster if `dstCapacity` >= `ZSTD_compressBound(srcSize)`.
* @return : compressed size written into `dst` (<= `dstCapacity),
* or an error code if it fails (which can be tested using ZSTD_isError()). */
ZSTDLIB_API size_t ZSTD_compress( void* dst, size_t dstCapacity,
const void* src, size_t srcSize,
int compressionLevel);
/*! ZSTD_decompress() :
* `compressedSize` : must be the _exact_ size of some number of compressed and/or skippable frames.
* `dstCapacity` is an upper bound of originalSize to regenerate.
* If user cannot imply a maximum upper bound, it's better to use streaming mode to decompress data.
* @return : the number of bytes decompressed into `dst` (<= `dstCapacity`),
* or an errorCode if it fails (which can be tested using ZSTD_isError()). */
ZSTDLIB_API size_t ZSTD_decompress( void* dst, size_t dstCapacity,
const void* src, size_t compressedSize);
/*! ZSTD_getFrameContentSize() : requires v1.3.0+
* `src` should point to the start of a ZSTD encoded frame.
* `srcSize` must be at least as large as the frame header.
* hint : any size >= `ZSTD_frameHeaderSize_max` is large enough.
* @return : - decompressed size of `src` frame content, if known
* - ZSTD_CONTENTSIZE_UNKNOWN if the size cannot be determined
* - ZSTD_CONTENTSIZE_ERROR if an error occurred (e.g. invalid magic number, srcSize too small)
* note 1 : a 0 return value means the frame is valid but "empty".
* note 2 : decompressed size is an optional field, it may not be present, typically in streaming mode.
* When `return==ZSTD_CONTENTSIZE_UNKNOWN`, data to decompress could be any size.
* In which case, it's necessary to use streaming mode to decompress data.
* Optionally, application can rely on some implicit limit,
* as ZSTD_decompress() only needs an upper bound of decompressed size.
* (For example, data could be necessarily cut into blocks <= 16 KB).
* note 3 : decompressed size is always present when compression is completed using single-pass functions,
* such as ZSTD_compress(), ZSTD_compressCCtx() ZSTD_compress_usingDict() or ZSTD_compress_usingCDict().
* note 4 : decompressed size can be very large (64-bits value),
* potentially larger than what local system can handle as a single memory segment.
* In which case, it's necessary to use streaming mode to decompress data.
* note 5 : If source is untrusted, decompressed size could be wrong or intentionally modified.
* Always ensure return value fits within application's authorized limits.
* Each application can set its own limits.
* note 6 : This function replaces ZSTD_getDecompressedSize() */
#define ZSTD_CONTENTSIZE_UNKNOWN (0ULL - 1)
#define ZSTD_CONTENTSIZE_ERROR (0ULL - 2)
ZSTDLIB_API unsigned long long ZSTD_getFrameContentSize(const void *src, size_t srcSize);
/*! ZSTD_getDecompressedSize() :
* NOTE: This function is now obsolete, in favor of ZSTD_getFrameContentSize().
* Both functions work the same way, but ZSTD_getDecompressedSize() blends
* "empty", "unknown" and "error" results to the same return value (0),
* while ZSTD_getFrameContentSize() gives them separate return values.
* @return : decompressed size of `src` frame content _if known and not empty_, 0 otherwise. */
ZSTDLIB_API unsigned long long ZSTD_getDecompressedSize(const void* src, size_t srcSize);
/*! ZSTD_findFrameCompressedSize() :
* `src` should point to the start of a ZSTD frame or skippable frame.
* `srcSize` must be >= first frame size
* @return : the compressed size of the first frame starting at `src`,
* suitable to pass as `srcSize` to `ZSTD_decompress` or similar,
* or an error code if input is invalid */
ZSTDLIB_API size_t ZSTD_findFrameCompressedSize(const void* src, size_t srcSize);
/*====== Helper functions ======*/
#define ZSTD_COMPRESSBOUND(srcSize) ((srcSize) + ((srcSize)>>8) + (((srcSize) < (128<<10)) ? (((128<<10) - (srcSize)) >> 11) /* margin, from 64 to 0 */ : 0)) /* this formula ensures that bound(A) + bound(B) <= bound(A+B) as long as A and B >= 128 KB */
ZSTDLIB_API size_t ZSTD_compressBound(size_t srcSize); /*!< maximum compressed size in worst case single-pass scenario */
ZSTDLIB_API unsigned ZSTD_isError(size_t code); /*!< tells if a `size_t` function result is an error code */
ZSTDLIB_API const char* ZSTD_getErrorName(size_t code); /*!< provides readable string from an error code */
ZSTDLIB_API int ZSTD_minCLevel(void); /*!< minimum negative compression level allowed */
ZSTDLIB_API int ZSTD_maxCLevel(void); /*!< maximum compression level available */
/***************************************
* Explicit context
***************************************/
/*= Compression context
* When compressing many times,
* it is recommended to allocate a context just once,
* and re-use it for each successive compression operation.
* This will make workload friendlier for system's memory.
* Note : re-using context is just a speed / resource optimization.
* It doesn't change the compression ratio, which remains identical.
* Note 2 : In multi-threaded environments,
* use one different context per thread for parallel execution.
*/
typedef struct ZSTD_CCtx_s ZSTD_CCtx;
ZSTDLIB_API ZSTD_CCtx* ZSTD_createCCtx(void);
ZSTDLIB_API size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx);
/*! ZSTD_compressCCtx() :
* Same as ZSTD_compress(), using an explicit ZSTD_CCtx.
* Important : in order to behave similarly to `ZSTD_compress()`,
* this function compresses at requested compression level,
* __ignoring any other parameter__ .
* If any advanced parameter was set using the advanced API,
* they will all be reset. Only `compressionLevel` remains.
*/
ZSTDLIB_API size_t ZSTD_compressCCtx(ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity,
const void* src, size_t srcSize,
int compressionLevel);
/*= Decompression context
* When decompressing many times,
* it is recommended to allocate a context only once,
* and re-use it for each successive compression operation.
* This will make workload friendlier for system's memory.
* Use one context per thread for parallel execution. */
typedef struct ZSTD_DCtx_s ZSTD_DCtx;
ZSTDLIB_API ZSTD_DCtx* ZSTD_createDCtx(void);
ZSTDLIB_API size_t ZSTD_freeDCtx(ZSTD_DCtx* dctx);
/*! ZSTD_decompressDCtx() :
* Same as ZSTD_decompress(),
* requires an allocated ZSTD_DCtx.
* Compatible with sticky parameters.
*/
ZSTDLIB_API size_t ZSTD_decompressDCtx(ZSTD_DCtx* dctx,
void* dst, size_t dstCapacity,
const void* src, size_t srcSize);
/***************************************
* Advanced compression API
***************************************/
/* API design :
* Parameters are pushed one by one into an existing context,
* using ZSTD_CCtx_set*() functions.
* Pushed parameters are sticky : they are valid for next compressed frame, and any subsequent frame.
* "sticky" parameters are applicable to `ZSTD_compress2()` and `ZSTD_compressStream*()` !
* __They do not apply to "simple" one-shot variants such as ZSTD_compressCCtx()__ .
*
* It's possible to reset all parameters to "default" using ZSTD_CCtx_reset().
*
* This API supercedes all other "advanced" API entry points in the experimental section.
* In the future, we expect to remove from experimental API entry points which are redundant with this API.
*/
/* Compression strategies, listed from fastest to strongest */
typedef enum { ZSTD_fast=1,
ZSTD_dfast=2,
ZSTD_greedy=3,
ZSTD_lazy=4,
ZSTD_lazy2=5,
ZSTD_btlazy2=6,
ZSTD_btopt=7,
ZSTD_btultra=8,
ZSTD_btultra2=9
/* note : new strategies _might_ be added in the future.
Only the order (from fast to strong) is guaranteed */
} ZSTD_strategy;
typedef enum {
/* compression parameters
* Note: When compressing with a ZSTD_CDict these parameters are superseded
* by the parameters used to construct the ZSTD_CDict.
* See ZSTD_CCtx_refCDict() for more info (superseded-by-cdict). */
ZSTD_c_compressionLevel=100, /* Set compression parameters according to pre-defined cLevel table.
* Note that exact compression parameters are dynamically determined,
* depending on both compression level and srcSize (when known).
* Default level is ZSTD_CLEVEL_DEFAULT==3.
* Special: value 0 means default, which is controlled by ZSTD_CLEVEL_DEFAULT.
* Note 1 : it's possible to pass a negative compression level.
* Note 2 : setting a level resets all other compression parameters to default */
/* Advanced compression parameters :
* It's possible to pin down compression parameters to some specific values.
* In which case, these values are no longer dynamically selected by the compressor */
ZSTD_c_windowLog=101, /* Maximum allowed back-reference distance, expressed as power of 2.
* This will set a memory budget for streaming decompression,
* with larger values requiring more memory
* and typically compressing more.
* Must be clamped between ZSTD_WINDOWLOG_MIN and ZSTD_WINDOWLOG_MAX.
* Special: value 0 means "use default windowLog".
* Note: Using a windowLog greater than ZSTD_WINDOWLOG_LIMIT_DEFAULT
* requires explicitly allowing such size at streaming decompression stage. */
ZSTD_c_hashLog=102, /* Size of the initial probe table, as a power of 2.
* Resulting memory usage is (1 << (hashLog+2)).
* Must be clamped between ZSTD_HASHLOG_MIN and ZSTD_HASHLOG_MAX.
* Larger tables improve compression ratio of strategies <= dFast,
* and improve speed of strategies > dFast.
* Special: value 0 means "use default hashLog". */
ZSTD_c_chainLog=103, /* Size of the multi-probe search table, as a power of 2.
* Resulting memory usage is (1 << (chainLog+2)).
* Must be clamped between ZSTD_CHAINLOG_MIN and ZSTD_CHAINLOG_MAX.
* Larger tables result in better and slower compression.
* This parameter is useless for "fast" strategy.
* It's still useful when using "dfast" strategy,
* in which case it defines a secondary probe table.
* Special: value 0 means "use default chainLog". */
ZSTD_c_searchLog=104, /* Number of search attempts, as a power of 2.
* More attempts result in better and slower compression.
* This parameter is useless for "fast" and "dFast" strategies.
* Special: value 0 means "use default searchLog". */
ZSTD_c_minMatch=105, /* Minimum size of searched matches.
* Note that Zstandard can still find matches of smaller size,
* it just tweaks its search algorithm to look for this size and larger.
* Larger values increase compression and decompression speed, but decrease ratio.
* Must be clamped between ZSTD_MINMATCH_MIN and ZSTD_MINMATCH_MAX.
* Note that currently, for all strategies < btopt, effective minimum is 4.
* , for all strategies > fast, effective maximum is 6.
* Special: value 0 means "use default minMatchLength". */
ZSTD_c_targetLength=106, /* Impact of this field depends on strategy.
* For strategies btopt, btultra & btultra2:
* Length of Match considered "good enough" to stop search.
* Larger values make compression stronger, and slower.
* For strategy fast:
* Distance between match sampling.
* Larger values make compression faster, and weaker.
* Special: value 0 means "use default targetLength". */
ZSTD_c_strategy=107, /* See ZSTD_strategy enum definition.
* The higher the value of selected strategy, the more complex it is,
* resulting in stronger and slower compression.
* Special: value 0 means "use default strategy". */
/* LDM mode parameters */
ZSTD_c_enableLongDistanceMatching=160, /* Enable long distance matching.
* This parameter is designed to improve compression ratio
* for large inputs, by finding large matches at long distance.
* It increases memory usage and window size.
* Note: enabling this parameter increases default ZSTD_c_windowLog to 128 MB
* except when expressly set to a different value. */
ZSTD_c_ldmHashLog=161, /* Size of the table for long distance matching, as a power of 2.
* Larger values increase memory usage and compression ratio,
* but decrease compression speed.
* Must be clamped between ZSTD_HASHLOG_MIN and ZSTD_HASHLOG_MAX
* default: windowlog - 7.
* Special: value 0 means "automatically determine hashlog". */
ZSTD_c_ldmMinMatch=162, /* Minimum match size for long distance matcher.
* Larger/too small values usually decrease compression ratio.
* Must be clamped between ZSTD_LDM_MINMATCH_MIN and ZSTD_LDM_MINMATCH_MAX.
* Special: value 0 means "use default value" (default: 64). */
ZSTD_c_ldmBucketSizeLog=163, /* Log size of each bucket in the LDM hash table for collision resolution.
* Larger values improve collision resolution but decrease compression speed.
* The maximum value is ZSTD_LDM_BUCKETSIZELOG_MAX.
* Special: value 0 means "use default value" (default: 3). */
ZSTD_c_ldmHashRateLog=164, /* Frequency of inserting/looking up entries into the LDM hash table.
* Must be clamped between 0 and (ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN).
* Default is MAX(0, (windowLog - ldmHashLog)), optimizing hash table usage.
* Larger values improve compression speed.
* Deviating far from default value will likely result in a compression ratio decrease.
* Special: value 0 means "automatically determine hashRateLog". */
/* frame parameters */
ZSTD_c_contentSizeFlag=200, /* Content size will be written into frame header _whenever known_ (default:1)
* Content size must be known at the beginning of compression.
* This is automatically the case when using ZSTD_compress2(),
* For streaming scenarios, content size must be provided with ZSTD_CCtx_setPledgedSrcSize() */
ZSTD_c_checksumFlag=201, /* A 32-bits checksum of content is written at end of frame (default:0) */
ZSTD_c_dictIDFlag=202, /* When applicable, dictionary's ID is written into frame header (default:1) */
/* multi-threading parameters */
/* These parameters are only useful if multi-threading is enabled (compiled with build macro ZSTD_MULTITHREAD).
* They return an error otherwise. */
ZSTD_c_nbWorkers=400, /* Select how many threads will be spawned to compress in parallel.
* When nbWorkers >= 1, triggers asynchronous mode when used with ZSTD_compressStream*() :
* ZSTD_compressStream*() consumes input and flush output if possible, but immediately gives back control to caller,
* while compression work is performed in parallel, within worker threads.
* (note : a strong exception to this rule is when first invocation of ZSTD_compressStream2() sets ZSTD_e_end :
* in which case, ZSTD_compressStream2() delegates to ZSTD_compress2(), which is always a blocking call).
* More workers improve speed, but also increase memory usage.
* Default value is `0`, aka "single-threaded mode" : no worker is spawned, compression is performed inside Caller's thread, all invocations are blocking */
ZSTD_c_jobSize=401, /* Size of a compression job. This value is enforced only when nbWorkers >= 1.
* Each compression job is completed in parallel, so this value can indirectly impact the nb of active threads.
* 0 means default, which is dynamically determined based on compression parameters.
* Job size must be a minimum of overlap size, or 1 MB, whichever is largest.
* The minimum size is automatically and transparently enforced. */
ZSTD_c_overlapLog=402, /* Control the overlap size, as a fraction of window size.
* The overlap size is an amount of data reloaded from previous job at the beginning of a new job.
* It helps preserve compression ratio, while each job is compressed in parallel.
* This value is enforced only when nbWorkers >= 1.
* Larger values increase compression ratio, but decrease speed.
* Possible values range from 0 to 9 :
* - 0 means "default" : value will be determined by the library, depending on strategy
* - 1 means "no overlap"
* - 9 means "full overlap", using a full window size.
* Each intermediate rank increases/decreases load size by a factor 2 :
* 9: full window; 8: w/2; 7: w/4; 6: w/8; 5:w/16; 4: w/32; 3:w/64; 2:w/128; 1:no overlap; 0:default
* default value varies between 6 and 9, depending on strategy */
/* note : additional experimental parameters are also available
* within the experimental section of the API.
* At the time of this writing, they include :
* ZSTD_c_rsyncable
* ZSTD_c_format
* ZSTD_c_forceMaxWindow
* ZSTD_c_forceAttachDict
* ZSTD_c_literalCompressionMode
* ZSTD_c_targetCBlockSize
* ZSTD_c_srcSizeHint
* Because they are not stable, it's necessary to define ZSTD_STATIC_LINKING_ONLY to access them.
* note : never ever use experimentalParam? names directly;
* also, the enums values themselves are unstable and can still change.
*/
ZSTD_c_experimentalParam1=500,
ZSTD_c_experimentalParam2=10,
ZSTD_c_experimentalParam3=1000,
ZSTD_c_experimentalParam4=1001,
ZSTD_c_experimentalParam5=1002,
ZSTD_c_experimentalParam6=1003,
ZSTD_c_experimentalParam7=1004
} ZSTD_cParameter;
typedef struct {
size_t error;
int lowerBound;
int upperBound;
} ZSTD_bounds;
/*! ZSTD_cParam_getBounds() :
* All parameters must belong to an interval with lower and upper bounds,
* otherwise they will either trigger an error or be automatically clamped.
* @return : a structure, ZSTD_bounds, which contains
* - an error status field, which must be tested using ZSTD_isError()
* - lower and upper bounds, both inclusive
*/
ZSTDLIB_API ZSTD_bounds ZSTD_cParam_getBounds(ZSTD_cParameter cParam);
/*! ZSTD_CCtx_setParameter() :
* Set one compression parameter, selected by enum ZSTD_cParameter.
* All parameters have valid bounds. Bounds can be queried using ZSTD_cParam_getBounds().
* Providing a value beyond bound will either clamp it, or trigger an error (depending on parameter).
* Setting a parameter is generally only possible during frame initialization (before starting compression).
* Exception : when using multi-threading mode (nbWorkers >= 1),
* the following parameters can be updated _during_ compression (within same frame):
* => compressionLevel, hashLog, chainLog, searchLog, minMatch, targetLength and strategy.
* new parameters will be active for next job only (after a flush()).
* @return : an error code (which can be tested using ZSTD_isError()).
*/
ZSTDLIB_API size_t ZSTD_CCtx_setParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int value);
/*! ZSTD_CCtx_setPledgedSrcSize() :
* Total input data size to be compressed as a single frame.
* Value will be written in frame header, unless if explicitly forbidden using ZSTD_c_contentSizeFlag.
* This value will also be controlled at end of frame, and trigger an error if not respected.
* @result : 0, or an error code (which can be tested with ZSTD_isError()).
* Note 1 : pledgedSrcSize==0 actually means zero, aka an empty frame.
* In order to mean "unknown content size", pass constant ZSTD_CONTENTSIZE_UNKNOWN.
* ZSTD_CONTENTSIZE_UNKNOWN is default value for any new frame.
* Note 2 : pledgedSrcSize is only valid once, for the next frame.
* It's discarded at the end of the frame, and replaced by ZSTD_CONTENTSIZE_UNKNOWN.
* Note 3 : Whenever all input data is provided and consumed in a single round,
* for example with ZSTD_compress2(),
* or invoking immediately ZSTD_compressStream2(,,,ZSTD_e_end),
* this value is automatically overridden by srcSize instead.
*/
ZSTDLIB_API size_t ZSTD_CCtx_setPledgedSrcSize(ZSTD_CCtx* cctx, unsigned long long pledgedSrcSize);
typedef enum {
ZSTD_reset_session_only = 1,
ZSTD_reset_parameters = 2,
ZSTD_reset_session_and_parameters = 3
} ZSTD_ResetDirective;
/*! ZSTD_CCtx_reset() :
* There are 2 different things that can be reset, independently or jointly :
* - The session : will stop compressing current frame, and make CCtx ready to start a new one.
* Useful after an error, or to interrupt any ongoing compression.
* Any internal data not yet flushed is cancelled.
* Compression parameters and dictionary remain unchanged.
* They will be used to compress next frame.
* Resetting session never fails.
* - The parameters : changes all parameters back to "default".
* This removes any reference to any dictionary too.
* Parameters can only be changed between 2 sessions (i.e. no compression is currently ongoing)
* otherwise the reset fails, and function returns an error value (which can be tested using ZSTD_isError())
* - Both : similar to resetting the session, followed by resetting parameters.
*/
ZSTDLIB_API size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset);
/*! ZSTD_compress2() :
* Behave the same as ZSTD_compressCCtx(), but compression parameters are set using the advanced API.
* ZSTD_compress2() always starts a new frame.
* Should cctx hold data from a previously unfinished frame, everything about it is forgotten.
* - Compression parameters are pushed into CCtx before starting compression, using ZSTD_CCtx_set*()
* - The function is always blocking, returns when compression is completed.
* Hint : compression runs faster if `dstCapacity` >= `ZSTD_compressBound(srcSize)`.
* @return : compressed size written into `dst` (<= `dstCapacity),
* or an error code if it fails (which can be tested using ZSTD_isError()).
*/
ZSTDLIB_API size_t ZSTD_compress2( ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity,
const void* src, size_t srcSize);
/***************************************
* Advanced decompression API
***************************************/
/* The advanced API pushes parameters one by one into an existing DCtx context.
* Parameters are sticky, and remain valid for all following frames
* using the same DCtx context.
* It's possible to reset parameters to default values using ZSTD_DCtx_reset().
* Note : This API is compatible with existing ZSTD_decompressDCtx() and ZSTD_decompressStream().
* Therefore, no new decompression function is necessary.
*/
typedef enum {
ZSTD_d_windowLogMax=100, /* Select a size limit (in power of 2) beyond which
* the streaming API will refuse to allocate memory buffer
* in order to protect the host from unreasonable memory requirements.
* This parameter is only useful in streaming mode, since no internal buffer is allocated in single-pass mode.
* By default, a decompression context accepts window sizes <= (1 << ZSTD_WINDOWLOG_LIMIT_DEFAULT).
* Special: value 0 means "use default maximum windowLog". */
/* note : additional experimental parameters are also available
* within the experimental section of the API.
* At the time of this writing, they include :
* ZSTD_c_format
* Because they are not stable, it's necessary to define ZSTD_STATIC_LINKING_ONLY to access them.
* note : never ever use experimentalParam? names directly
*/
ZSTD_d_experimentalParam1=1000
} ZSTD_dParameter;
/*! ZSTD_dParam_getBounds() :
* All parameters must belong to an interval with lower and upper bounds,
* otherwise they will either trigger an error or be automatically clamped.
* @return : a structure, ZSTD_bounds, which contains
* - an error status field, which must be tested using ZSTD_isError()
* - both lower and upper bounds, inclusive
*/
ZSTDLIB_API ZSTD_bounds ZSTD_dParam_getBounds(ZSTD_dParameter dParam);
/*! ZSTD_DCtx_setParameter() :
* Set one compression parameter, selected by enum ZSTD_dParameter.
* All parameters have valid bounds. Bounds can be queried using ZSTD_dParam_getBounds().
* Providing a value beyond bound will either clamp it, or trigger an error (depending on parameter).
* Setting a parameter is only possible during frame initialization (before starting decompression).
* @return : 0, or an error code (which can be tested using ZSTD_isError()).
*/
ZSTDLIB_API size_t ZSTD_DCtx_setParameter(ZSTD_DCtx* dctx, ZSTD_dParameter param, int value);
/*! ZSTD_DCtx_reset() :
* Return a DCtx to clean state.
* Session and parameters can be reset jointly or separately.
* Parameters can only be reset when no active frame is being decompressed.
* @return : 0, or an error code, which can be tested with ZSTD_isError()
*/
ZSTDLIB_API size_t ZSTD_DCtx_reset(ZSTD_DCtx* dctx, ZSTD_ResetDirective reset);
/****************************
* Streaming
****************************/
typedef struct ZSTD_inBuffer_s {
const void* src; /**< start of input buffer */
size_t size; /**< size of input buffer */
size_t pos; /**< position where reading stopped. Will be updated. Necessarily 0 <= pos <= size */
} ZSTD_inBuffer;
typedef struct ZSTD_outBuffer_s {
void* dst; /**< start of output buffer */
size_t size; /**< size of output buffer */
size_t pos; /**< position where writing stopped. Will be updated. Necessarily 0 <= pos <= size */
} ZSTD_outBuffer;
/*-***********************************************************************
* Streaming compression - HowTo
*
* A ZSTD_CStream object is required to track streaming operation.
* Use ZSTD_createCStream() and ZSTD_freeCStream() to create/release resources.
* ZSTD_CStream objects can be reused multiple times on consecutive compression operations.
* It is recommended to re-use ZSTD_CStream since it will play nicer with system's memory, by re-using already allocated memory.
*
* For parallel execution, use one separate ZSTD_CStream per thread.
*
* note : since v1.3.0, ZSTD_CStream and ZSTD_CCtx are the same thing.
*
* Parameters are sticky : when starting a new compression on the same context,
* it will re-use the same sticky parameters as previous compression session.
* When in doubt, it's recommended to fully initialize the context before usage.
* Use ZSTD_CCtx_reset() to reset the context and ZSTD_CCtx_setParameter(),
* ZSTD_CCtx_setPledgedSrcSize(), or ZSTD_CCtx_loadDictionary() and friends to
* set more specific parameters, the pledged source size, or load a dictionary.
*
* Use ZSTD_compressStream2() with ZSTD_e_continue as many times as necessary to
* consume input stream. The function will automatically update both `pos`
* fields within `input` and `output`.
* Note that the function may not consume the entire input, for example, because
* the output buffer is already full, in which case `input.pos < input.size`.
* The caller must check if input has been entirely consumed.
* If not, the caller must make some room to receive more compressed data,
* and then present again remaining input data.
* note: ZSTD_e_continue is guaranteed to make some forward progress when called,
* but doesn't guarantee maximal forward progress. This is especially relevant
* when compressing with multiple threads. The call won't block if it can
* consume some input, but if it can't it will wait for some, but not all,
* output to be flushed.
* @return : provides a minimum amount of data remaining to be flushed from internal buffers
* or an error code, which can be tested using ZSTD_isError().
*
* At any moment, it's possible to flush whatever data might remain stuck within internal buffer,
* using ZSTD_compressStream2() with ZSTD_e_flush. `output->pos` will be updated.
* Note that, if `output->size` is too small, a single invocation with ZSTD_e_flush might not be enough (return code > 0).
* In which case, make some room to receive more compressed data, and call again ZSTD_compressStream2() with ZSTD_e_flush.
* You must continue calling ZSTD_compressStream2() with ZSTD_e_flush until it returns 0, at which point you can change the
* operation.
* note: ZSTD_e_flush will flush as much output as possible, meaning when compressing with multiple threads, it will
* block until the flush is complete or the output buffer is full.
* @return : 0 if internal buffers are entirely flushed,
* >0 if some data still present within internal buffer (the value is minimal estimation of remaining size),
* or an error code, which can be tested using ZSTD_isError().
*
* Calling ZSTD_compressStream2() with ZSTD_e_end instructs to finish a frame.
* It will perform a flush and write frame epilogue.
* The epilogue is required for decoders to consider a frame completed.
* flush operation is the same, and follows same rules as calling ZSTD_compressStream2() with ZSTD_e_flush.
* You must continue calling ZSTD_compressStream2() with ZSTD_e_end until it returns 0, at which point you are free to
* start a new frame.
* note: ZSTD_e_end will flush as much output as possible, meaning when compressing with multiple threads, it will
* block until the flush is complete or the output buffer is full.
* @return : 0 if frame fully completed and fully flushed,
* >0 if some data still present within internal buffer (the value is minimal estimation of remaining size),
* or an error code, which can be tested using ZSTD_isError().
*
* *******************************************************************/
typedef ZSTD_CCtx ZSTD_CStream; /**< CCtx and CStream are now effectively same object (>= v1.3.0) */
/* Continue to distinguish them for compatibility with older versions <= v1.2.0 */
/*===== ZSTD_CStream management functions =====*/
ZSTDLIB_API ZSTD_CStream* ZSTD_createCStream(void);
ZSTDLIB_API size_t ZSTD_freeCStream(ZSTD_CStream* zcs);
/*===== Streaming compression functions =====*/
typedef enum {
ZSTD_e_continue=0, /* collect more data, encoder decides when to output compressed result, for optimal compression ratio */
ZSTD_e_flush=1, /* flush any data provided so far,
* it creates (at least) one new block, that can be decoded immediately on reception;
* frame will continue: any future data can still reference previously compressed data, improving compression.
* note : multithreaded compression will block to flush as much output as possible. */
ZSTD_e_end=2 /* flush any remaining data _and_ close current frame.
* note that frame is only closed after compressed data is fully flushed (return value == 0).
* After that point, any additional data starts a new frame.
* note : each frame is independent (does not reference any content from previous frame).
: note : multithreaded compression will block to flush as much output as possible. */
} ZSTD_EndDirective;
/*! ZSTD_compressStream2() :
* Behaves about the same as ZSTD_compressStream, with additional control on end directive.
* - Compression parameters are pushed into CCtx before starting compression, using ZSTD_CCtx_set*()
* - Compression parameters cannot be changed once compression is started (save a list of exceptions in multi-threading mode)
* - output->pos must be <= dstCapacity, input->pos must be <= srcSize
* - output->pos and input->pos will be updated. They are guaranteed to remain below their respective limit.
* - When nbWorkers==0 (default), function is blocking : it completes its job before returning to caller.
* - When nbWorkers>=1, function is non-blocking : it just acquires a copy of input, and distributes jobs to internal worker threads, flush whatever is available,
* and then immediately returns, just indicating that there is some data remaining to be flushed.
* The function nonetheless guarantees forward progress : it will return only after it reads or write at least 1+ byte.
* - Exception : if the first call requests a ZSTD_e_end directive and provides enough dstCapacity, the function delegates to ZSTD_compress2() which is always blocking.
* - @return provides a minimum amount of data remaining to be flushed from internal buffers
* or an error code, which can be tested using ZSTD_isError().
* if @return != 0, flush is not fully completed, there is still some data left within internal buffers.
* This is useful for ZSTD_e_flush, since in this case more flushes are necessary to empty all buffers.
* For ZSTD_e_end, @return == 0 when internal buffers are fully flushed and frame is completed.
* - after a ZSTD_e_end directive, if internal buffer is not fully flushed (@return != 0),
* only ZSTD_e_end or ZSTD_e_flush operations are allowed.
* Before starting a new compression job, or changing compression parameters,
* it is required to fully flush internal buffers.
*/
ZSTDLIB_API size_t ZSTD_compressStream2( ZSTD_CCtx* cctx,
ZSTD_outBuffer* output,
ZSTD_inBuffer* input,
ZSTD_EndDirective endOp);
/* These buffer sizes are softly recommended.
* They are not required : ZSTD_compressStream*() happily accepts any buffer size, for both input and output.
* Respecting the recommended size just makes it a bit easier for ZSTD_compressStream*(),
* reducing the amount of memory shuffling and buffering, resulting in minor performance savings.
*
* However, note that these recommendations are from the perspective of a C caller program.
* If the streaming interface is invoked from some other language,
* especially managed ones such as Java or Go, through a foreign function interface such as jni or cgo,
* a major performance rule is to reduce crossing such interface to an absolute minimum.
* It's not rare that performance ends being spent more into the interface, rather than compression itself.
* In which cases, prefer using large buffers, as large as practical,
* for both input and output, to reduce the nb of roundtrips.
*/
ZSTDLIB_API size_t ZSTD_CStreamInSize(void); /**< recommended size for input buffer */
ZSTDLIB_API size_t ZSTD_CStreamOutSize(void); /**< recommended size for output buffer. Guarantee to successfully flush at least one complete compressed block. */
/* *****************************************************************************
* This following is a legacy streaming API.
* It can be replaced by ZSTD_CCtx_reset() and ZSTD_compressStream2().
* It is redundant, but remains fully supported.
* Advanced parameters and dictionary compression can only be used through the
* new API.
******************************************************************************/
/*!
* Equivalent to:
*
* ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
* ZSTD_CCtx_refCDict(zcs, NULL); // clear the dictionary (if any)
* ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel);
*/
ZSTDLIB_API size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel);
/*!
* Alternative for ZSTD_compressStream2(zcs, output, input, ZSTD_e_continue).
* NOTE: The return value is different. ZSTD_compressStream() returns a hint for
* the next read size (if non-zero and not an error). ZSTD_compressStream2()
* returns the minimum nb of bytes left to flush (if non-zero and not an error).
*/
ZSTDLIB_API size_t ZSTD_compressStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input);
/*! Equivalent to ZSTD_compressStream2(zcs, output, &emptyInput, ZSTD_e_flush). */
ZSTDLIB_API size_t ZSTD_flushStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output);
/*! Equivalent to ZSTD_compressStream2(zcs, output, &emptyInput, ZSTD_e_end). */
ZSTDLIB_API size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output);
/*-***************************************************************************
* Streaming decompression - HowTo
*
* A ZSTD_DStream object is required to track streaming operations.
* Use ZSTD_createDStream() and ZSTD_freeDStream() to create/release resources.
* ZSTD_DStream objects can be re-used multiple times.
*
* Use ZSTD_initDStream() to start a new decompression operation.
* @return : recommended first input size
* Alternatively, use advanced API to set specific properties.
*
* Use ZSTD_decompressStream() repetitively to consume your input.
* The function will update both `pos` fields.
* If `input.pos < input.size`, some input has not been consumed.
* It's up to the caller to present again remaining data.
* The function tries to flush all data decoded immediately, respecting output buffer size.
* If `output.pos < output.size`, decoder has flushed everything it could.
* But if `output.pos == output.size`, there might be some data left within internal buffers.,
* In which case, call ZSTD_decompressStream() again to flush whatever remains in the buffer.
* Note : with no additional input provided, amount of data flushed is necessarily <= ZSTD_BLOCKSIZE_MAX.
* @return : 0 when a frame is completely decoded and fully flushed,
* or an error code, which can be tested using ZSTD_isError(),
* or any other value > 0, which means there is still some decoding or flushing to do to complete current frame :
* the return value is a suggested next input size (just a hint for better latency)
* that will never request more than the remaining frame size.
* *******************************************************************************/
typedef ZSTD_DCtx ZSTD_DStream; /**< DCtx and DStream are now effectively same object (>= v1.3.0) */
/* For compatibility with versions <= v1.2.0, prefer differentiating them. */
/*===== ZSTD_DStream management functions =====*/
ZSTDLIB_API ZSTD_DStream* ZSTD_createDStream(void);
ZSTDLIB_API size_t ZSTD_freeDStream(ZSTD_DStream* zds);
/*===== Streaming decompression functions =====*/
/* This function is redundant with the advanced API and equivalent to:
*
* ZSTD_DCtx_reset(zds);
* ZSTD_DCtx_refDDict(zds, NULL);
*/
ZSTDLIB_API size_t ZSTD_initDStream(ZSTD_DStream* zds);
ZSTDLIB_API size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inBuffer* input);
ZSTDLIB_API size_t ZSTD_DStreamInSize(void); /*!< recommended size for input buffer */
ZSTDLIB_API size_t ZSTD_DStreamOutSize(void); /*!< recommended size for output buffer. Guarantee to successfully flush at least one complete block in all circumstances. */
/**************************
* Simple dictionary API
***************************/
/*! ZSTD_compress_usingDict() :
* Compression at an explicit compression level using a Dictionary.
* A dictionary can be any arbitrary data segment (also called a prefix),
* or a buffer with specified information (see dictBuilder/zdict.h).
* Note : This function loads the dictionary, resulting in significant startup delay.
* It's intended for a dictionary used only once.
* Note 2 : When `dict == NULL || dictSize < 8` no dictionary is used. */
ZSTDLIB_API size_t ZSTD_compress_usingDict(ZSTD_CCtx* ctx,
void* dst, size_t dstCapacity,
const void* src, size_t srcSize,
const void* dict,size_t dictSize,
int compressionLevel);
/*! ZSTD_decompress_usingDict() :
* Decompression using a known Dictionary.
* Dictionary must be identical to the one used during compression.
* Note : This function loads the dictionary, resulting in significant startup delay.
* It's intended for a dictionary used only once.
* Note : When `dict == NULL || dictSize < 8` no dictionary is used. */
ZSTDLIB_API size_t ZSTD_decompress_usingDict(ZSTD_DCtx* dctx,
void* dst, size_t dstCapacity,
const void* src, size_t srcSize,
const void* dict,size_t dictSize);
/***********************************
* Bulk processing dictionary API
**********************************/
typedef struct ZSTD_CDict_s ZSTD_CDict;
/*! ZSTD_createCDict() :
* When compressing multiple messages or blocks using the same dictionary,
* it's recommended to digest the dictionary only once, since it's a costly operation.
* ZSTD_createCDict() will create a state from digesting a dictionary.
* The resulting state can be used for future compression operations with very limited startup cost.
* ZSTD_CDict can be created once and shared by multiple threads concurrently, since its usage is read-only.
* @dictBuffer can be released after ZSTD_CDict creation, because its content is copied within CDict.
* Note 1 : Consider experimental function `ZSTD_createCDict_byReference()` if you prefer to not duplicate @dictBuffer content.
* Note 2 : A ZSTD_CDict can be created from an empty @dictBuffer,
* in which case the only thing that it transports is the @compressionLevel.
* This can be useful in a pipeline featuring ZSTD_compress_usingCDict() exclusively,
* expecting a ZSTD_CDict parameter with any data, including those without a known dictionary. */
ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict(const void* dictBuffer, size_t dictSize,
int compressionLevel);
/*! ZSTD_freeCDict() :
* Function frees memory allocated by ZSTD_createCDict(). */
ZSTDLIB_API size_t ZSTD_freeCDict(ZSTD_CDict* CDict);
/*! ZSTD_compress_usingCDict() :
* Compression using a digested Dictionary.
* Recommended when same dictionary is used multiple times.
* Note : compression level is _decided at dictionary creation time_,
* and frame parameters are hardcoded (dictID=yes, contentSize=yes, checksum=no) */
ZSTDLIB_API size_t ZSTD_compress_usingCDict(ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity,
const void* src, size_t srcSize,
const ZSTD_CDict* cdict);
typedef struct ZSTD_DDict_s ZSTD_DDict;
/*! ZSTD_createDDict() :
* Create a digested dictionary, ready to start decompression operation without startup delay.
* dictBuffer can be released after DDict creation, as its content is copied inside DDict. */
ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict(const void* dictBuffer, size_t dictSize);
/*! ZSTD_freeDDict() :
* Function frees memory allocated with ZSTD_createDDict() */
ZSTDLIB_API size_t ZSTD_freeDDict(ZSTD_DDict* ddict);
/*! ZSTD_decompress_usingDDict() :
* Decompression using a digested Dictionary.
* Recommended when same dictionary is used multiple times. */
ZSTDLIB_API size_t ZSTD_decompress_usingDDict(ZSTD_DCtx* dctx,
void* dst, size_t dstCapacity,
const void* src, size_t srcSize,
const ZSTD_DDict* ddict);
/********************************
* Dictionary helper functions
*******************************/
/*! ZSTD_getDictID_fromDict() :
* Provides the dictID stored within dictionary.
* if @return == 0, the dictionary is not conformant with Zstandard specification.
* It can still be loaded, but as a content-only dictionary. */
ZSTDLIB_API unsigned ZSTD_getDictID_fromDict(const void* dict, size_t dictSize);
/*! ZSTD_getDictID_fromDDict() :
* Provides the dictID of the dictionary loaded into `ddict`.
* If @return == 0, the dictionary is not conformant to Zstandard specification, or empty.
* Non-conformant dictionaries can still be loaded, but as content-only dictionaries. */
ZSTDLIB_API unsigned ZSTD_getDictID_fromDDict(const ZSTD_DDict* ddict);
/*! ZSTD_getDictID_fromFrame() :
* Provides the dictID required to decompressed the frame stored within `src`.
* If @return == 0, the dictID could not be decoded.
* This could for one of the following reasons :
* - The frame does not require a dictionary to be decoded (most common case).
* - The frame was built with dictID intentionally removed. Whatever dictionary is necessary is a hidden information.
* Note : this use case also happens when using a non-conformant dictionary.
* - `srcSize` is too small, and as a result, the frame header could not be decoded (only possible if `srcSize < ZSTD_FRAMEHEADERSIZE_MAX`).
* - This is not a Zstandard frame.
* When identifying the exact failure cause, it's possible to use ZSTD_getFrameHeader(), which will provide a more precise error code. */
ZSTDLIB_API unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize);
/*******************************************************************************
* Advanced dictionary and prefix API
*
* This API allows dictionaries to be used with ZSTD_compress2(),
* ZSTD_compressStream2(), and ZSTD_decompress(). Dictionaries are sticky, and
* only reset with the context is reset with ZSTD_reset_parameters or
* ZSTD_reset_session_and_parameters. Prefixes are single-use.
******************************************************************************/
/*! ZSTD_CCtx_loadDictionary() :
* Create an internal CDict from `dict` buffer.
* Decompression will have to use same dictionary.
* @result : 0, or an error code (which can be tested with ZSTD_isError()).
* Special: Loading a NULL (or 0-size) dictionary invalidates previous dictionary,
* meaning "return to no-dictionary mode".
* Note 1 : Dictionary is sticky, it will be used for all future compressed frames.
* To return to "no-dictionary" situation, load a NULL dictionary (or reset parameters).
* Note 2 : Loading a dictionary involves building tables.
* It's also a CPU consuming operation, with non-negligible impact on latency.
* Tables are dependent on compression parameters, and for this reason,
* compression parameters can no longer be changed after loading a dictionary.
* Note 3 :`dict` content will be copied internally.
* Use experimental ZSTD_CCtx_loadDictionary_byReference() to reference content instead.
* In such a case, dictionary buffer must outlive its users.
* Note 4 : Use ZSTD_CCtx_loadDictionary_advanced()
* to precisely select how dictionary content must be interpreted. */
ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary(ZSTD_CCtx* cctx, const void* dict, size_t dictSize);
/*! ZSTD_CCtx_refCDict() :
* Reference a prepared dictionary, to be used for all next compressed frames.
* Note that compression parameters are enforced from within CDict,
* and supersede any compression parameter previously set within CCtx.
* The parameters ignored are labled as "superseded-by-cdict" in the ZSTD_cParameter enum docs.
* The ignored parameters will be used again if the CCtx is returned to no-dictionary mode.
* The dictionary will remain valid for future compressed frames using same CCtx.
* @result : 0, or an error code (which can be tested with ZSTD_isError()).
* Special : Referencing a NULL CDict means "return to no-dictionary mode".
* Note 1 : Currently, only one dictionary can be managed.
* Referencing a new dictionary effectively "discards" any previous one.
* Note 2 : CDict is just referenced, its lifetime must outlive its usage within CCtx. */
ZSTDLIB_API size_t ZSTD_CCtx_refCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict);
/*! ZSTD_CCtx_refPrefix() :
* Reference a prefix (single-usage dictionary) for next compressed frame.
* A prefix is **only used once**. Tables are discarded at end of frame (ZSTD_e_end).
* Decompression will need same prefix to properly regenerate data.
* Compressing with a prefix is similar in outcome as performing a diff and compressing it,
* but performs much faster, especially during decompression (compression speed is tunable with compression level).
* @result : 0, or an error code (which can be tested with ZSTD_isError()).
* Special: Adding any prefix (including NULL) invalidates any previous prefix or dictionary
* Note 1 : Prefix buffer is referenced. It **must** outlive compression.
* Its content must remain unmodified during compression.
* Note 2 : If the intention is to diff some large src data blob with some prior version of itself,
* ensure that the window size is large enough to contain the entire source.
* See ZSTD_c_windowLog.
* Note 3 : Referencing a prefix involves building tables, which are dependent on compression parameters.
* It's a CPU consuming operation, with non-negligible impact on latency.
* If there is a need to use the same prefix multiple times, consider loadDictionary instead.
* Note 4 : By default, the prefix is interpreted as raw content (ZSTD_dct_rawContent).
* Use experimental ZSTD_CCtx_refPrefix_advanced() to alter dictionary interpretation. */
ZSTDLIB_API size_t ZSTD_CCtx_refPrefix(ZSTD_CCtx* cctx,
const void* prefix, size_t prefixSize);
/*! ZSTD_DCtx_loadDictionary() :
* Create an internal DDict from dict buffer,
* to be used to decompress next frames.
* The dictionary remains valid for all future frames, until explicitly invalidated.
* @result : 0, or an error code (which can be tested with ZSTD_isError()).
* Special : Adding a NULL (or 0-size) dictionary invalidates any previous dictionary,
* meaning "return to no-dictionary mode".
* Note 1 : Loading a dictionary involves building tables,
* which has a non-negligible impact on CPU usage and latency.
* It's recommended to "load once, use many times", to amortize the cost
* Note 2 :`dict` content will be copied internally, so `dict` can be released after loading.
* Use ZSTD_DCtx_loadDictionary_byReference() to reference dictionary content instead.
* Note 3 : Use ZSTD_DCtx_loadDictionary_advanced() to take control of
* how dictionary content is loaded and interpreted.
*/
ZSTDLIB_API size_t ZSTD_DCtx_loadDictionary(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);
/*! ZSTD_DCtx_refDDict() :
* Reference a prepared dictionary, to be used to decompress next frames.
* The dictionary remains active for decompression of future frames using same DCtx.
* @result : 0, or an error code (which can be tested with ZSTD_isError()).
* Note 1 : Currently, only one dictionary can be managed.
* Referencing a new dictionary effectively "discards" any previous one.
* Special: referencing a NULL DDict means "return to no-dictionary mode".
* Note 2 : DDict is just referenced, its lifetime must outlive its usage from DCtx.
*/
ZSTDLIB_API size_t ZSTD_DCtx_refDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict);
/*! ZSTD_DCtx_refPrefix() :
* Reference a prefix (single-usage dictionary) to decompress next frame.
* This is the reverse operation of ZSTD_CCtx_refPrefix(),
* and must use the same prefix as the one used during compression.
* Prefix is **only used once**. Reference is discarded at end of frame.
* End of frame is reached when ZSTD_decompressStream() returns 0.
* @result : 0, or an error code (which can be tested with ZSTD_isError()).
* Note 1 : Adding any prefix (including NULL) invalidates any previously set prefix or dictionary
* Note 2 : Prefix buffer is referenced. It **must** outlive decompression.
* Prefix buffer must remain unmodified up to the end of frame,
* reached when ZSTD_decompressStream() returns 0.
* Note 3 : By default, the prefix is treated as raw content (ZSTD_dct_rawContent).
* Use ZSTD_CCtx_refPrefix_advanced() to alter dictMode (Experimental section)
* Note 4 : Referencing a raw content prefix has almost no cpu nor memory cost.
* A full dictionary is more costly, as it requires building tables.
*/
ZSTDLIB_API size_t ZSTD_DCtx_refPrefix(ZSTD_DCtx* dctx,
const void* prefix, size_t prefixSize);
/* === Memory management === */
/*! ZSTD_sizeof_*() :
* These functions give the _current_ memory usage of selected object.
* Note that object memory usage can evolve (increase or decrease) over time. */
ZSTDLIB_API size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx);
ZSTDLIB_API size_t ZSTD_sizeof_DCtx(const ZSTD_DCtx* dctx);
ZSTDLIB_API size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs);
ZSTDLIB_API size_t ZSTD_sizeof_DStream(const ZSTD_DStream* zds);
ZSTDLIB_API size_t ZSTD_sizeof_CDict(const ZSTD_CDict* cdict);
ZSTDLIB_API size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict);
#endif /* ZSTD_H_235446 */
/* **************************************************************************************
* ADVANCED AND EXPERIMENTAL FUNCTIONS
****************************************************************************************
* The definitions in the following section are considered experimental.
* They are provided for advanced scenarios.
* They should never be used with a dynamic library, as prototypes may change in the future.
* Use them only in association with static linking.
* ***************************************************************************************/
#if defined(ZSTD_STATIC_LINKING_ONLY) && !defined(ZSTD_H_ZSTD_STATIC_LINKING_ONLY)
#define ZSTD_H_ZSTD_STATIC_LINKING_ONLY
/****************************************************************************************
* experimental API (static linking only)
****************************************************************************************
* The following symbols and constants
* are not planned to join "stable API" status in the near future.
* They can still change in future versions.
* Some of them are planned to remain in the static_only section indefinitely.
* Some of them might be removed in the future (especially when redundant with existing stable functions)
* ***************************************************************************************/
#define ZSTD_FRAMEHEADERSIZE_PREFIX(format) ((format) == ZSTD_f_zstd1 ? 5 : 1) /* minimum input size required to query frame header size */
#define ZSTD_FRAMEHEADERSIZE_MIN(format) ((format) == ZSTD_f_zstd1 ? 6 : 2)
#define ZSTD_FRAMEHEADERSIZE_MAX 18 /* can be useful for static allocation */
#define ZSTD_SKIPPABLEHEADERSIZE 8
/* compression parameter bounds */
#define ZSTD_WINDOWLOG_MAX_32 30
#define ZSTD_WINDOWLOG_MAX_64 31
#define ZSTD_WINDOWLOG_MAX ((int)(sizeof(size_t) == 4 ? ZSTD_WINDOWLOG_MAX_32 : ZSTD_WINDOWLOG_MAX_64))
#define ZSTD_WINDOWLOG_MIN 10
#define ZSTD_HASHLOG_MAX ((ZSTD_WINDOWLOG_MAX < 30) ? ZSTD_WINDOWLOG_MAX : 30)
#define ZSTD_HASHLOG_MIN 6
#define ZSTD_CHAINLOG_MAX_32 29
#define ZSTD_CHAINLOG_MAX_64 30
#define ZSTD_CHAINLOG_MAX ((int)(sizeof(size_t) == 4 ? ZSTD_CHAINLOG_MAX_32 : ZSTD_CHAINLOG_MAX_64))
#define ZSTD_CHAINLOG_MIN ZSTD_HASHLOG_MIN
#define ZSTD_SEARCHLOG_MAX (ZSTD_WINDOWLOG_MAX-1)
#define ZSTD_SEARCHLOG_MIN 1
#define ZSTD_MINMATCH_MAX 7 /* only for ZSTD_fast, other strategies are limited to 6 */
#define ZSTD_MINMATCH_MIN 3 /* only for ZSTD_btopt+, faster strategies are limited to 4 */
#define ZSTD_TARGETLENGTH_MAX ZSTD_BLOCKSIZE_MAX
#define ZSTD_TARGETLENGTH_MIN 0 /* note : comparing this constant to an unsigned results in a tautological test */
#define ZSTD_STRATEGY_MIN ZSTD_fast
#define ZSTD_STRATEGY_MAX ZSTD_btultra2
#define ZSTD_OVERLAPLOG_MIN 0
#define ZSTD_OVERLAPLOG_MAX 9
#define ZSTD_WINDOWLOG_LIMIT_DEFAULT 27 /* by default, the streaming decoder will refuse any frame
* requiring larger than (1<<ZSTD_WINDOWLOG_LIMIT_DEFAULT) window size,
* to preserve host's memory from unreasonable requirements.
* This limit can be overridden using ZSTD_DCtx_setParameter(,ZSTD_d_windowLogMax,).
* The limit does not apply for one-pass decoders (such as ZSTD_decompress()), since no additional memory is allocated */
/* LDM parameter bounds */
#define ZSTD_LDM_HASHLOG_MIN ZSTD_HASHLOG_MIN
#define ZSTD_LDM_HASHLOG_MAX ZSTD_HASHLOG_MAX
#define ZSTD_LDM_MINMATCH_MIN 4
#define ZSTD_LDM_MINMATCH_MAX 4096
#define ZSTD_LDM_BUCKETSIZELOG_MIN 1
#define ZSTD_LDM_BUCKETSIZELOG_MAX 8
#define ZSTD_LDM_HASHRATELOG_MIN 0
#define ZSTD_LDM_HASHRATELOG_MAX (ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN)
/* Advanced parameter bounds */
#define ZSTD_TARGETCBLOCKSIZE_MIN 64
#define ZSTD_TARGETCBLOCKSIZE_MAX ZSTD_BLOCKSIZE_MAX
#define ZSTD_SRCSIZEHINT_MIN 0
#define ZSTD_SRCSIZEHINT_MAX INT_MAX
/* internal */
#define ZSTD_HASHLOG3_MAX 17
/* --- Advanced types --- */
typedef struct ZSTD_CCtx_params_s ZSTD_CCtx_params;
typedef struct {
unsigned int matchPos; /* Match pos in dst */
/* If seqDef.offset > 3, then this is seqDef.offset - 3
* If seqDef.offset < 3, then this is the corresponding repeat offset
* But if seqDef.offset < 3 and litLength == 0, this is the
* repeat offset before the corresponding repeat offset
* And if seqDef.offset == 3 and litLength == 0, this is the
* most recent repeat offset - 1
*/
unsigned int offset;
unsigned int litLength; /* Literal length */
unsigned int matchLength; /* Match length */
/* 0 when seq not rep and seqDef.offset otherwise
* when litLength == 0 this will be <= 4, otherwise <= 3 like normal
*/
unsigned int rep;
} ZSTD_Sequence;
typedef struct {
unsigned windowLog; /**< largest match distance : larger == more compression, more memory needed during decompression */
unsigned chainLog; /**< fully searched segment : larger == more compression, slower, more memory (useless for fast) */
unsigned hashLog; /**< dispatch table : larger == faster, more memory */
unsigned searchLog; /**< nb of searches : larger == more compression, slower */
unsigned minMatch; /**< match length searched : larger == faster decompression, sometimes less compression */
unsigned targetLength; /**< acceptable match size for optimal parser (only) : larger == more compression, slower */
ZSTD_strategy strategy; /**< see ZSTD_strategy definition above */
} ZSTD_compressionParameters;
typedef struct {
int contentSizeFlag; /**< 1: content size will be in frame header (when known) */
int checksumFlag; /**< 1: generate a 32-bits checksum using XXH64 algorithm at end of frame, for error detection */
int noDictIDFlag; /**< 1: no dictID will be saved into frame header (dictID is only useful for dictionary compression) */
} ZSTD_frameParameters;
typedef struct {
ZSTD_compressionParameters cParams;
ZSTD_frameParameters fParams;
} ZSTD_parameters;
typedef enum {
ZSTD_dct_auto = 0, /* dictionary is "full" when starting with ZSTD_MAGIC_DICTIONARY, otherwise it is "rawContent" */
ZSTD_dct_rawContent = 1, /* ensures dictionary is always loaded as rawContent, even if it starts with ZSTD_MAGIC_DICTIONARY */
ZSTD_dct_fullDict = 2 /* refuses to load a dictionary if it does not respect Zstandard's specification, starting with ZSTD_MAGIC_DICTIONARY */
} ZSTD_dictContentType_e;
typedef enum {
ZSTD_dlm_byCopy = 0, /**< Copy dictionary content internally */
ZSTD_dlm_byRef = 1 /**< Reference dictionary content -- the dictionary buffer must outlive its users. */
} ZSTD_dictLoadMethod_e;
typedef enum {
ZSTD_f_zstd1 = 0, /* zstd frame format, specified in zstd_compression_format.md (default) */
ZSTD_f_zstd1_magicless = 1 /* Variant of zstd frame format, without initial 4-bytes magic number.
* Useful to save 4 bytes per generated frame.
* Decoder cannot recognise automatically this format, requiring this instruction. */
} ZSTD_format_e;
typedef enum {
/* Note: this enum and the behavior it controls are effectively internal
* implementation details of the compressor. They are expected to continue
* to evolve and should be considered only in the context of extremely
* advanced performance tuning.
*
* Zstd currently supports the use of a CDict in three ways:
*
* - The contents of the CDict can be copied into the working context. This
* means that the compression can search both the dictionary and input
* while operating on a single set of internal tables. This makes
* the compression faster per-byte of input. However, the initial copy of
* the CDict's tables incurs a fixed cost at the beginning of the
* compression. For small compressions (< 8 KB), that copy can dominate
* the cost of the compression.
*
* - The CDict's tables can be used in-place. In this model, compression is
* slower per input byte, because the compressor has to search two sets of
* tables. However, this model incurs no start-up cost (as long as the
* working context's tables can be reused). For small inputs, this can be
* faster than copying the CDict's tables.
*
* - The CDict's tables are not used at all, and instead we use the working
* context alone to reload the dictionary and use params based on the source
* size. See ZSTD_compress_insertDictionary() and ZSTD_compress_usingDict().
* This method is effective when the dictionary sizes are very small relative
* to the input size, and the input size is fairly large to begin with.
*
* Zstd has a simple internal heuristic that selects which strategy to use
* at the beginning of a compression. However, if experimentation shows that
* Zstd is making poor choices, it is possible to override that choice with
* this enum.
*/
ZSTD_dictDefaultAttach = 0, /* Use the default heuristic. */
ZSTD_dictForceAttach = 1, /* Never copy the dictionary. */
ZSTD_dictForceCopy = 2, /* Always copy the dictionary. */
ZSTD_dictForceLoad = 3 /* Always reload the dictionary */
} ZSTD_dictAttachPref_e;
typedef enum {
ZSTD_lcm_auto = 0, /**< Automatically determine the compression mode based on the compression level.
* Negative compression levels will be uncompressed, and positive compression
* levels will be compressed. */
ZSTD_lcm_huffman = 1, /**< Always attempt Huffman compression. Uncompressed literals will still be
* emitted if Huffman compression is not profitable. */
ZSTD_lcm_uncompressed = 2 /**< Always emit uncompressed literals. */
} ZSTD_literalCompressionMode_e;
/***************************************
* Frame size functions
***************************************/
/*! ZSTD_findDecompressedSize() :
* `src` should point to the start of a series of ZSTD encoded and/or skippable frames
* `srcSize` must be the _exact_ size of this series
* (i.e. there should be a frame boundary at `src + srcSize`)
* @return : - decompressed size of all data in all successive frames
* - if the decompressed size cannot be determined: ZSTD_CONTENTSIZE_UNKNOWN
* - if an error occurred: ZSTD_CONTENTSIZE_ERROR
*
* note 1 : decompressed size is an optional field, that may not be present, especially in streaming mode.
* When `return==ZSTD_CONTENTSIZE_UNKNOWN`, data to decompress could be any size.
* In which case, it's necessary to use streaming mode to decompress data.
* note 2 : decompressed size is always present when compression is done with ZSTD_compress()
* note 3 : decompressed size can be very large (64-bits value),
* potentially larger than what local system can handle as a single memory segment.
* In which case, it's necessary to use streaming mode to decompress data.
* note 4 : If source is untrusted, decompressed size could be wrong or intentionally modified.
* Always ensure result fits within application's authorized limits.
* Each application can set its own limits.
* note 5 : ZSTD_findDecompressedSize handles multiple frames, and so it must traverse the input to
* read each contained frame header. This is fast as most of the data is skipped,
* however it does mean that all frame data must be present and valid. */
ZSTDLIB_API unsigned long long ZSTD_findDecompressedSize(const void* src, size_t srcSize);
/*! ZSTD_decompressBound() :
* `src` should point to the start of a series of ZSTD encoded and/or skippable frames
* `srcSize` must be the _exact_ size of this series
* (i.e. there should be a frame boundary at `src + srcSize`)
* @return : - upper-bound for the decompressed size of all data in all successive frames
* - if an error occured: ZSTD_CONTENTSIZE_ERROR
*
* note 1 : an error can occur if `src` contains an invalid or incorrectly formatted frame.
* note 2 : the upper-bound is exact when the decompressed size field is available in every ZSTD encoded frame of `src`.
* in this case, `ZSTD_findDecompressedSize` and `ZSTD_decompressBound` return the same value.
* note 3 : when the decompressed size field isn't available, the upper-bound for that frame is calculated by:
* upper-bound = # blocks * min(128 KB, Window_Size)
*/
ZSTDLIB_API unsigned long long ZSTD_decompressBound(const void* src, size_t srcSize);
/*! ZSTD_frameHeaderSize() :
* srcSize must be >= ZSTD_FRAMEHEADERSIZE_PREFIX.
* @return : size of the Frame Header,
* or an error code (if srcSize is too small) */
ZSTDLIB_API size_t ZSTD_frameHeaderSize(const void* src, size_t srcSize);
/*! ZSTD_getSequences() :
* Extract sequences from the sequence store
* zc can be used to insert custom compression params.
* This function invokes ZSTD_compress2
* @return : number of sequences extracted
*/
ZSTDLIB_API size_t ZSTD_getSequences(ZSTD_CCtx* zc, ZSTD_Sequence* outSeqs,
size_t outSeqsSize, const void* src, size_t srcSize);
/***************************************
* Memory management
***************************************/
/*! ZSTD_estimate*() :
* These functions make it possible to estimate memory usage of a future
* {D,C}Ctx, before its creation.
*
* ZSTD_estimateCCtxSize() will provide a budget large enough for any
* compression level up to selected one. Unlike ZSTD_estimateCStreamSize*(),
* this estimate does not include space for a window buffer, so this estimate
* is guaranteed to be enough for single-shot compressions, but not streaming
* compressions. It will however assume the input may be arbitrarily large,
* which is the worst case. If srcSize is known to always be small,
* ZSTD_estimateCCtxSize_usingCParams() can provide a tighter estimation.
* ZSTD_estimateCCtxSize_usingCParams() can be used in tandem with
* ZSTD_getCParams() to create cParams from compressionLevel.
* ZSTD_estimateCCtxSize_usingCCtxParams() can be used in tandem with
* ZSTD_CCtxParams_setParameter().
*
* Note: only single-threaded compression is supported. This function will
* return an error code if ZSTD_c_nbWorkers is >= 1. */
ZSTDLIB_API size_t ZSTD_estimateCCtxSize(int compressionLevel);
ZSTDLIB_API size_t ZSTD_estimateCCtxSize_usingCParams(ZSTD_compressionParameters cParams);
ZSTDLIB_API size_t ZSTD_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params);
ZSTDLIB_API size_t ZSTD_estimateDCtxSize(void);
/*! ZSTD_estimateCStreamSize() :
* ZSTD_estimateCStreamSize() will provide a budget large enough for any compression level up to selected one.
* It will also consider src size to be arbitrarily "large", which is worst case.
* If srcSize is known to always be small, ZSTD_estimateCStreamSize_usingCParams() can provide a tighter estimation.
* ZSTD_estimateCStreamSize_usingCParams() can be used in tandem with ZSTD_getCParams() to create cParams from compressionLevel.
* ZSTD_estimateCStreamSize_usingCCtxParams() can be used in tandem with ZSTD_CCtxParams_setParameter(). Only single-threaded compression is supported. This function will return an error code if ZSTD_c_nbWorkers is >= 1.
* Note : CStream size estimation is only correct for single-threaded compression.
* ZSTD_DStream memory budget depends on window Size.
* This information can be passed manually, using ZSTD_estimateDStreamSize,
* or deducted from a valid frame Header, using ZSTD_estimateDStreamSize_fromFrame();
* Note : if streaming is init with function ZSTD_init?Stream_usingDict(),
* an internal ?Dict will be created, which additional size is not estimated here.
* In this case, get total size by adding ZSTD_estimate?DictSize */
ZSTDLIB_API size_t ZSTD_estimateCStreamSize(int compressionLevel);
ZSTDLIB_API size_t ZSTD_estimateCStreamSize_usingCParams(ZSTD_compressionParameters cParams);
ZSTDLIB_API size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params);
ZSTDLIB_API size_t ZSTD_estimateDStreamSize(size_t windowSize);
ZSTDLIB_API size_t ZSTD_estimateDStreamSize_fromFrame(const void* src, size_t srcSize);
/*! ZSTD_estimate?DictSize() :
* ZSTD_estimateCDictSize() will bet that src size is relatively "small", and content is copied, like ZSTD_createCDict().
* ZSTD_estimateCDictSize_advanced() makes it possible to control compression parameters precisely, like ZSTD_createCDict_advanced().
* Note : dictionaries created by reference (`ZSTD_dlm_byRef`) are logically smaller.
*/
ZSTDLIB_API size_t ZSTD_estimateCDictSize(size_t dictSize, int compressionLevel);
ZSTDLIB_API size_t ZSTD_estimateCDictSize_advanced(size_t dictSize, ZSTD_compressionParameters cParams, ZSTD_dictLoadMethod_e dictLoadMethod);
ZSTDLIB_API size_t ZSTD_estimateDDictSize(size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod);
/*! ZSTD_initStatic*() :
* Initialize an object using a pre-allocated fixed-size buffer.
* workspace: The memory area to emplace the object into.
* Provided pointer *must be 8-bytes aligned*.
* Buffer must outlive object.
* workspaceSize: Use ZSTD_estimate*Size() to determine
* how large workspace must be to support target scenario.
* @return : pointer to object (same address as workspace, just different type),
* or NULL if error (size too small, incorrect alignment, etc.)
* Note : zstd will never resize nor malloc() when using a static buffer.
* If the object requires more memory than available,
* zstd will just error out (typically ZSTD_error_memory_allocation).
* Note 2 : there is no corresponding "free" function.
* Since workspace is allocated externally, it must be freed externally too.
* Note 3 : cParams : use ZSTD_getCParams() to convert a compression level
* into its associated cParams.
* Limitation 1 : currently not compatible with internal dictionary creation, triggered by
* ZSTD_CCtx_loadDictionary(), ZSTD_initCStream_usingDict() or ZSTD_initDStream_usingDict().
* Limitation 2 : static cctx currently not compatible with multi-threading.
* Limitation 3 : static dctx is incompatible with legacy support.
*/
ZSTDLIB_API ZSTD_CCtx* ZSTD_initStaticCCtx(void* workspace, size_t workspaceSize);
ZSTDLIB_API ZSTD_CStream* ZSTD_initStaticCStream(void* workspace, size_t workspaceSize); /**< same as ZSTD_initStaticCCtx() */
ZSTDLIB_API ZSTD_DCtx* ZSTD_initStaticDCtx(void* workspace, size_t workspaceSize);
ZSTDLIB_API ZSTD_DStream* ZSTD_initStaticDStream(void* workspace, size_t workspaceSize); /**< same as ZSTD_initStaticDCtx() */
ZSTDLIB_API const ZSTD_CDict* ZSTD_initStaticCDict(
void* workspace, size_t workspaceSize,
const void* dict, size_t dictSize,
ZSTD_dictLoadMethod_e dictLoadMethod,
ZSTD_dictContentType_e dictContentType,
ZSTD_compressionParameters cParams);
ZSTDLIB_API const ZSTD_DDict* ZSTD_initStaticDDict(
void* workspace, size_t workspaceSize,
const void* dict, size_t dictSize,
ZSTD_dictLoadMethod_e dictLoadMethod,
ZSTD_dictContentType_e dictContentType);
/*! Custom memory allocation :
* These prototypes make it possible to pass your own allocation/free functions.
* ZSTD_customMem is provided at creation time, using ZSTD_create*_advanced() variants listed below.
* All allocation/free operations will be completed using these custom variants instead of regular <stdlib.h> ones.
*/
typedef void* (*ZSTD_allocFunction) (void* opaque, size_t size);
typedef void (*ZSTD_freeFunction) (void* opaque, void* address);
typedef struct { ZSTD_allocFunction customAlloc; ZSTD_freeFunction customFree; void* opaque; } ZSTD_customMem;
static ZSTD_customMem const ZSTD_defaultCMem = { NULL, NULL, NULL }; /**< this constant defers to stdlib's functions */
ZSTDLIB_API ZSTD_CCtx* ZSTD_createCCtx_advanced(ZSTD_customMem customMem);
ZSTDLIB_API ZSTD_CStream* ZSTD_createCStream_advanced(ZSTD_customMem customMem);
ZSTDLIB_API ZSTD_DCtx* ZSTD_createDCtx_advanced(ZSTD_customMem customMem);
ZSTDLIB_API ZSTD_DStream* ZSTD_createDStream_advanced(ZSTD_customMem customMem);
ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict_advanced(const void* dict, size_t dictSize,
ZSTD_dictLoadMethod_e dictLoadMethod,
ZSTD_dictContentType_e dictContentType,
ZSTD_compressionParameters cParams,
ZSTD_customMem customMem);
ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict_advanced(const void* dict, size_t dictSize,
ZSTD_dictLoadMethod_e dictLoadMethod,
ZSTD_dictContentType_e dictContentType,
ZSTD_customMem customMem);
/***************************************
* Advanced compression functions
***************************************/
/*! ZSTD_createCDict_byReference() :
* Create a digested dictionary for compression
* Dictionary content is just referenced, not duplicated.
* As a consequence, `dictBuffer` **must** outlive CDict,
* and its content must remain unmodified throughout the lifetime of CDict.
* note: equivalent to ZSTD_createCDict_advanced(), with dictLoadMethod==ZSTD_dlm_byRef */
ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict_byReference(const void* dictBuffer, size_t dictSize, int compressionLevel);
/*! ZSTD_getCParams() :
* @return ZSTD_compressionParameters structure for a selected compression level and estimated srcSize.
* `estimatedSrcSize` value is optional, select 0 if not known */
ZSTDLIB_API ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize);
/*! ZSTD_getParams() :
* same as ZSTD_getCParams(), but @return a full `ZSTD_parameters` object instead of sub-component `ZSTD_compressionParameters`.
* All fields of `ZSTD_frameParameters` are set to default : contentSize=1, checksum=0, noDictID=0 */
ZSTDLIB_API ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize);
/*! ZSTD_checkCParams() :
* Ensure param values remain within authorized range.
* @return 0 on success, or an error code (can be checked with ZSTD_isError()) */
ZSTDLIB_API size_t ZSTD_checkCParams(ZSTD_compressionParameters params);
/*! ZSTD_adjustCParams() :
* optimize params for a given `srcSize` and `dictSize`.
* `srcSize` can be unknown, in which case use ZSTD_CONTENTSIZE_UNKNOWN.
* `dictSize` must be `0` when there is no dictionary.
* cPar can be invalid : all parameters will be clamped within valid range in the @return struct.
* This function never fails (wide contract) */
ZSTDLIB_API ZSTD_compressionParameters ZSTD_adjustCParams(ZSTD_compressionParameters cPar, unsigned long long srcSize, size_t dictSize);
/*! ZSTD_compress_advanced() :
* Note : this function is now DEPRECATED.
* It can be replaced by ZSTD_compress2(), in combination with ZSTD_CCtx_setParameter() and other parameter setters.
* This prototype will be marked as deprecated and generate compilation warning on reaching v1.5.x */
ZSTDLIB_API size_t ZSTD_compress_advanced(ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity,
const void* src, size_t srcSize,
const void* dict,size_t dictSize,
ZSTD_parameters params);
/*! ZSTD_compress_usingCDict_advanced() :
* Note : this function is now REDUNDANT.
* It can be replaced by ZSTD_compress2(), in combination with ZSTD_CCtx_loadDictionary() and other parameter setters.
* This prototype will be marked as deprecated and generate compilation warning in some future version */
ZSTDLIB_API size_t ZSTD_compress_usingCDict_advanced(ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity,
const void* src, size_t srcSize,
const ZSTD_CDict* cdict,
ZSTD_frameParameters fParams);
/*! ZSTD_CCtx_loadDictionary_byReference() :
* Same as ZSTD_CCtx_loadDictionary(), but dictionary content is referenced, instead of being copied into CCtx.
* It saves some memory, but also requires that `dict` outlives its usage within `cctx` */
ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary_byReference(ZSTD_CCtx* cctx, const void* dict, size_t dictSize);
/*! ZSTD_CCtx_loadDictionary_advanced() :
* Same as ZSTD_CCtx_loadDictionary(), but gives finer control over
* how to load the dictionary (by copy ? by reference ?)
* and how to interpret it (automatic ? force raw mode ? full mode only ?) */
ZSTDLIB_API size_t ZSTD_CCtx_loadDictionary_advanced(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictContentType_e dictContentType);
/*! ZSTD_CCtx_refPrefix_advanced() :
* Same as ZSTD_CCtx_refPrefix(), but gives finer control over
* how to interpret prefix content (automatic ? force raw mode (default) ? full mode only ?) */
ZSTDLIB_API size_t ZSTD_CCtx_refPrefix_advanced(ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType);
/* === experimental parameters === */
/* these parameters can be used with ZSTD_setParameter()
* they are not guaranteed to remain supported in the future */
/* Enables rsyncable mode,
* which makes compressed files more rsync friendly
* by adding periodic synchronization points to the compressed data.
* The target average block size is ZSTD_c_jobSize / 2.
* It's possible to modify the job size to increase or decrease
* the granularity of the synchronization point.
* Once the jobSize is smaller than the window size,
* it will result in compression ratio degradation.
* NOTE 1: rsyncable mode only works when multithreading is enabled.
* NOTE 2: rsyncable performs poorly in combination with long range mode,
* since it will decrease the effectiveness of synchronization points,
* though mileage may vary.
* NOTE 3: Rsyncable mode limits maximum compression speed to ~400 MB/s.
* If the selected compression level is already running significantly slower,
* the overall speed won't be significantly impacted.
*/
#define ZSTD_c_rsyncable ZSTD_c_experimentalParam1
/* Select a compression format.
* The value must be of type ZSTD_format_e.
* See ZSTD_format_e enum definition for details */
#define ZSTD_c_format ZSTD_c_experimentalParam2
/* Force back-reference distances to remain < windowSize,
* even when referencing into Dictionary content (default:0) */
#define ZSTD_c_forceMaxWindow ZSTD_c_experimentalParam3
/* Controls whether the contents of a CDict
* are used in place, or copied into the working context.
* Accepts values from the ZSTD_dictAttachPref_e enum.
* See the comments on that enum for an explanation of the feature. */
#define ZSTD_c_forceAttachDict ZSTD_c_experimentalParam4
/* Controls how the literals are compressed (default is auto).
* The value must be of type ZSTD_literalCompressionMode_e.
* See ZSTD_literalCompressionMode_t enum definition for details.
*/
#define ZSTD_c_literalCompressionMode ZSTD_c_experimentalParam5
/* Tries to fit compressed block size to be around targetCBlockSize.
* No target when targetCBlockSize == 0.
* There is no guarantee on compressed block size (default:0) */
#define ZSTD_c_targetCBlockSize ZSTD_c_experimentalParam6
/* User's best guess of source size.
* Hint is not valid when srcSizeHint == 0.
* There is no guarantee that hint is close to actual source size,
* but compression ratio may regress significantly if guess considerably underestimates */
#define ZSTD_c_srcSizeHint ZSTD_c_experimentalParam7
/*! ZSTD_CCtx_getParameter() :
* Get the requested compression parameter value, selected by enum ZSTD_cParameter,
* and store it into int* value.
* @return : 0, or an error code (which can be tested with ZSTD_isError()).
*/
ZSTDLIB_API size_t ZSTD_CCtx_getParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int* value);
/*! ZSTD_CCtx_params :
* Quick howto :
* - ZSTD_createCCtxParams() : Create a ZSTD_CCtx_params structure
* - ZSTD_CCtxParams_setParameter() : Push parameters one by one into
* an existing ZSTD_CCtx_params structure.
* This is similar to
* ZSTD_CCtx_setParameter().
* - ZSTD_CCtx_setParametersUsingCCtxParams() : Apply parameters to
* an existing CCtx.
* These parameters will be applied to
* all subsequent frames.
* - ZSTD_compressStream2() : Do compression using the CCtx.
* - ZSTD_freeCCtxParams() : Free the memory.
*
* This can be used with ZSTD_estimateCCtxSize_advanced_usingCCtxParams()
* for static allocation of CCtx for single-threaded compression.
*/
ZSTDLIB_API ZSTD_CCtx_params* ZSTD_createCCtxParams(void);
ZSTDLIB_API size_t ZSTD_freeCCtxParams(ZSTD_CCtx_params* params);
/*! ZSTD_CCtxParams_reset() :
* Reset params to default values.
*/
ZSTDLIB_API size_t ZSTD_CCtxParams_reset(ZSTD_CCtx_params* params);
/*! ZSTD_CCtxParams_init() :
* Initializes the compression parameters of cctxParams according to
* compression level. All other parameters are reset to their default values.
*/
ZSTDLIB_API size_t ZSTD_CCtxParams_init(ZSTD_CCtx_params* cctxParams, int compressionLevel);
/*! ZSTD_CCtxParams_init_advanced() :
* Initializes the compression and frame parameters of cctxParams according to
* params. All other parameters are reset to their default values.
*/
ZSTDLIB_API size_t ZSTD_CCtxParams_init_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_parameters params);
/*! ZSTD_CCtxParams_setParameter() :
* Similar to ZSTD_CCtx_setParameter.
* Set one compression parameter, selected by enum ZSTD_cParameter.
* Parameters must be applied to a ZSTD_CCtx using ZSTD_CCtx_setParametersUsingCCtxParams().
* @result : 0, or an error code (which can be tested with ZSTD_isError()).
*/
ZSTDLIB_API size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int value);
/*! ZSTD_CCtxParams_getParameter() :
* Similar to ZSTD_CCtx_getParameter.
* Get the requested value of one compression parameter, selected by enum ZSTD_cParameter.
* @result : 0, or an error code (which can be tested with ZSTD_isError()).
*/
ZSTDLIB_API size_t ZSTD_CCtxParams_getParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int* value);
/*! ZSTD_CCtx_setParametersUsingCCtxParams() :
* Apply a set of ZSTD_CCtx_params to the compression context.
* This can be done even after compression is started,
* if nbWorkers==0, this will have no impact until a new compression is started.
* if nbWorkers>=1, new parameters will be picked up at next job,
* with a few restrictions (windowLog, pledgedSrcSize, nbWorkers, jobSize, and overlapLog are not updated).
*/
ZSTDLIB_API size_t ZSTD_CCtx_setParametersUsingCCtxParams(
ZSTD_CCtx* cctx, const ZSTD_CCtx_params* params);
/*! ZSTD_compressStream2_simpleArgs() :
* Same as ZSTD_compressStream2(),
* but using only integral types as arguments.
* This variant might be helpful for binders from dynamic languages
* which have troubles handling structures containing memory pointers.
*/
ZSTDLIB_API size_t ZSTD_compressStream2_simpleArgs (
ZSTD_CCtx* cctx,
void* dst, size_t dstCapacity, size_t* dstPos,
const void* src, size_t srcSize, size_t* srcPos,
ZSTD_EndDirective endOp);
/***************************************
* Advanced decompression functions
***************************************/
/*! ZSTD_isFrame() :
* Tells if the content of `buffer` starts with a valid Frame Identifier.
* Note : Frame Identifier is 4 bytes. If `size < 4`, @return will always be 0.
* Note 2 : Legacy Frame Identifiers are considered valid only if Legacy Support is enabled.
* Note 3 : Skippable Frame Identifiers are considered valid. */
ZSTDLIB_API unsigned ZSTD_isFrame(const void* buffer, size_t size);
/*! ZSTD_createDDict_byReference() :
* Create a digested dictionary, ready to start decompression operation without startup delay.
* Dictionary content is referenced, and therefore stays in dictBuffer.
* It is important that dictBuffer outlives DDict,
* it must remain read accessible throughout the lifetime of DDict */
ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict_byReference(const void* dictBuffer, size_t dictSize);
/*! ZSTD_DCtx_loadDictionary_byReference() :
* Same as ZSTD_DCtx_loadDictionary(),
* but references `dict` content instead of copying it into `dctx`.
* This saves memory if `dict` remains around.,
* However, it's imperative that `dict` remains accessible (and unmodified) while being used, so it must outlive decompression. */
ZSTDLIB_API size_t ZSTD_DCtx_loadDictionary_byReference(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);
/*! ZSTD_DCtx_loadDictionary_advanced() :
* Same as ZSTD_DCtx_loadDictionary(),
* but gives direct control over
* how to load the dictionary (by copy ? by reference ?)
* and how to interpret it (automatic ? force raw mode ? full mode only ?). */
ZSTDLIB_API size_t ZSTD_DCtx_loadDictionary_advanced(ZSTD_DCtx* dctx, const void* dict, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictContentType_e dictContentType);
/*! ZSTD_DCtx_refPrefix_advanced() :
* Same as ZSTD_DCtx_refPrefix(), but gives finer control over
* how to interpret prefix content (automatic ? force raw mode (default) ? full mode only ?) */
ZSTDLIB_API size_t ZSTD_DCtx_refPrefix_advanced(ZSTD_DCtx* dctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType);
/*! ZSTD_DCtx_setMaxWindowSize() :
* Refuses allocating internal buffers for frames requiring a window size larger than provided limit.
* This protects a decoder context from reserving too much memory for itself (potential attack scenario).
* This parameter is only useful in streaming mode, since no internal buffer is allocated in single-pass mode.
* By default, a decompression context accepts all window sizes <= (1 << ZSTD_WINDOWLOG_LIMIT_DEFAULT)
* @return : 0, or an error code (which can be tested using ZSTD_isError()).
*/
ZSTDLIB_API size_t ZSTD_DCtx_setMaxWindowSize(ZSTD_DCtx* dctx, size_t maxWindowSize);
/* ZSTD_d_format
* experimental parameter,
* allowing selection between ZSTD_format_e input compression formats
*/
#define ZSTD_d_format ZSTD_d_experimentalParam1
/*! ZSTD_DCtx_setFormat() :
* Instruct the decoder context about what kind of data to decode next.
* This instruction is mandatory to decode data without a fully-formed header,
* such ZSTD_f_zstd1_magicless for example.
* @return : 0, or an error code (which can be tested using ZSTD_isError()). */
ZSTDLIB_API size_t ZSTD_DCtx_setFormat(ZSTD_DCtx* dctx, ZSTD_format_e format);
/*! ZSTD_decompressStream_simpleArgs() :
* Same as ZSTD_decompressStream(),
* but using only integral types as arguments.
* This can be helpful for binders from dynamic languages
* which have troubles handling structures containing memory pointers.
*/
ZSTDLIB_API size_t ZSTD_decompressStream_simpleArgs (
ZSTD_DCtx* dctx,
void* dst, size_t dstCapacity, size_t* dstPos,
const void* src, size_t srcSize, size_t* srcPos);
/********************************************************************
* Advanced streaming functions
* Warning : most of these functions are now redundant with the Advanced API.
* Once Advanced API reaches "stable" status,
* redundant functions will be deprecated, and then at some point removed.
********************************************************************/
/*===== Advanced Streaming compression functions =====*/
/**! ZSTD_initCStream_srcSize() :
* This function is deprecated, and equivalent to:
* ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
* ZSTD_CCtx_refCDict(zcs, NULL); // clear the dictionary (if any)
* ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel);
* ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
*
* pledgedSrcSize must be correct. If it is not known at init time, use
* ZSTD_CONTENTSIZE_UNKNOWN. Note that, for compatibility with older programs,
* "0" also disables frame content size field. It may be enabled in the future.
* Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
*/
ZSTDLIB_API size_t
ZSTD_initCStream_srcSize(ZSTD_CStream* zcs,
int compressionLevel,
unsigned long long pledgedSrcSize);
/**! ZSTD_initCStream_usingDict() :
* This function is deprecated, and is equivalent to:
* ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
* ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel);
* ZSTD_CCtx_loadDictionary(zcs, dict, dictSize);
*
* Creates of an internal CDict (incompatible with static CCtx), except if
* dict == NULL or dictSize < 8, in which case no dict is used.
* Note: dict is loaded with ZSTD_dct_auto (treated as a full zstd dictionary if
* it begins with ZSTD_MAGIC_DICTIONARY, else as raw content) and ZSTD_dlm_byCopy.
* Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
*/
ZSTDLIB_API size_t
ZSTD_initCStream_usingDict(ZSTD_CStream* zcs,
const void* dict, size_t dictSize,
int compressionLevel);
/**! ZSTD_initCStream_advanced() :
* This function is deprecated, and is approximately equivalent to:
* ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
* // Pseudocode: Set each zstd parameter and leave the rest as-is.
* for ((param, value) : params) {
* ZSTD_CCtx_setParameter(zcs, param, value);
* }
* ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
* ZSTD_CCtx_loadDictionary(zcs, dict, dictSize);
*
* dict is loaded with ZSTD_dct_auto and ZSTD_dlm_byCopy.
* pledgedSrcSize must be correct.
* If srcSize is not known at init time, use value ZSTD_CONTENTSIZE_UNKNOWN.
* Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
*/
ZSTDLIB_API size_t
ZSTD_initCStream_advanced(ZSTD_CStream* zcs,
const void* dict, size_t dictSize,
ZSTD_parameters params,
unsigned long long pledgedSrcSize);
/**! ZSTD_initCStream_usingCDict() :
* This function is deprecated, and equivalent to:
* ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
* ZSTD_CCtx_refCDict(zcs, cdict);
*
* note : cdict will just be referenced, and must outlive compression session
* Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
*/
ZSTDLIB_API size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict);
/**! ZSTD_initCStream_usingCDict_advanced() :
* This function is DEPRECATED, and is approximately equivalent to:
* ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
* // Pseudocode: Set each zstd frame parameter and leave the rest as-is.
* for ((fParam, value) : fParams) {
* ZSTD_CCtx_setParameter(zcs, fParam, value);
* }
* ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
* ZSTD_CCtx_refCDict(zcs, cdict);
*
* same as ZSTD_initCStream_usingCDict(), with control over frame parameters.
* pledgedSrcSize must be correct. If srcSize is not known at init time, use
* value ZSTD_CONTENTSIZE_UNKNOWN.
* Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
*/
ZSTDLIB_API size_t
ZSTD_initCStream_usingCDict_advanced(ZSTD_CStream* zcs,
const ZSTD_CDict* cdict,
ZSTD_frameParameters fParams,
unsigned long long pledgedSrcSize);
/*! ZSTD_resetCStream() :
* This function is deprecated, and is equivalent to:
* ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
* ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
*
* start a new frame, using same parameters from previous frame.
* This is typically useful to skip dictionary loading stage, since it will re-use it in-place.
* Note that zcs must be init at least once before using ZSTD_resetCStream().
* If pledgedSrcSize is not known at reset time, use macro ZSTD_CONTENTSIZE_UNKNOWN.
* If pledgedSrcSize > 0, its value must be correct, as it will be written in header, and controlled at the end.
* For the time being, pledgedSrcSize==0 is interpreted as "srcSize unknown" for compatibility with older programs,
* but it will change to mean "empty" in future version, so use macro ZSTD_CONTENTSIZE_UNKNOWN instead.
* @return : 0, or an error code (which can be tested using ZSTD_isError())
* Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
*/
ZSTDLIB_API size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize);
typedef struct {
unsigned long long ingested; /* nb input bytes read and buffered */
unsigned long long consumed; /* nb input bytes actually compressed */
unsigned long long produced; /* nb of compressed bytes generated and buffered */
unsigned long long flushed; /* nb of compressed bytes flushed : not provided; can be tracked from caller side */
unsigned currentJobID; /* MT only : latest started job nb */
unsigned nbActiveWorkers; /* MT only : nb of workers actively compressing at probe time */
} ZSTD_frameProgression;
/* ZSTD_getFrameProgression() :
* tells how much data has been ingested (read from input)
* consumed (input actually compressed) and produced (output) for current frame.
* Note : (ingested - consumed) is amount of input data buffered internally, not yet compressed.
* Aggregates progression inside active worker threads.
*/
ZSTDLIB_API ZSTD_frameProgression ZSTD_getFrameProgression(const ZSTD_CCtx* cctx);
/*! ZSTD_toFlushNow() :
* Tell how many bytes are ready to be flushed immediately.
* Useful for multithreading scenarios (nbWorkers >= 1).
* Probe the oldest active job, defined as oldest job not yet entirely flushed,
* and check its output buffer.
* @return : amount of data stored in oldest job and ready to be flushed immediately.
* if @return == 0, it means either :
* + there is no active job (could be checked with ZSTD_frameProgression()), or
* + oldest job is still actively compressing data,
* but everything it has produced has also been flushed so far,
* therefore flush speed is limited by production speed of oldest job
* irrespective of the speed of concurrent (and newer) jobs.
*/
ZSTDLIB_API size_t ZSTD_toFlushNow(ZSTD_CCtx* cctx);
/*===== Advanced Streaming decompression functions =====*/
/**
* This function is deprecated, and is equivalent to:
*
* ZSTD_DCtx_reset(zds, ZSTD_reset_session_only);
* ZSTD_DCtx_loadDictionary(zds, dict, dictSize);
*
* note: no dictionary will be used if dict == NULL or dictSize < 8
* Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
*/
ZSTDLIB_API size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize);
/**
* This function is deprecated, and is equivalent to:
*
* ZSTD_DCtx_reset(zds, ZSTD_reset_session_only);
* ZSTD_DCtx_refDDict(zds, ddict);
*
* note : ddict is referenced, it must outlive decompression session
* Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
*/
ZSTDLIB_API size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* zds, const ZSTD_DDict* ddict);
/**
* This function is deprecated, and is equivalent to:
*
* ZSTD_DCtx_reset(zds, ZSTD_reset_session_only);
*
* re-use decompression parameters from previous init; saves dictionary loading
* Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
*/
ZSTDLIB_API size_t ZSTD_resetDStream(ZSTD_DStream* zds);
/*********************************************************************
* Buffer-less and synchronous inner streaming functions
*
* This is an advanced API, giving full control over buffer management, for users which need direct control over memory.
* But it's also a complex one, with several restrictions, documented below.
* Prefer normal streaming API for an easier experience.
********************************************************************* */
/**
Buffer-less streaming compression (synchronous mode)
A ZSTD_CCtx object is required to track streaming operations.
Use ZSTD_createCCtx() / ZSTD_freeCCtx() to manage resource.
ZSTD_CCtx object can be re-used multiple times within successive compression operations.
Start by initializing a context.
Use ZSTD_compressBegin(), or ZSTD_compressBegin_usingDict() for dictionary compression,
or ZSTD_compressBegin_advanced(), for finer parameter control.
It's also possible to duplicate a reference context which has already been initialized, using ZSTD_copyCCtx()
Then, consume your input using ZSTD_compressContinue().
There are some important considerations to keep in mind when using this advanced function :
- ZSTD_compressContinue() has no internal buffer. It uses externally provided buffers only.
- Interface is synchronous : input is consumed entirely and produces 1+ compressed blocks.
- Caller must ensure there is enough space in `dst` to store compressed data under worst case scenario.
Worst case evaluation is provided by ZSTD_compressBound().
ZSTD_compressContinue() doesn't guarantee recover after a failed compression.
- ZSTD_compressContinue() presumes prior input ***is still accessible and unmodified*** (up to maximum distance size, see WindowLog).
It remembers all previous contiguous blocks, plus one separated memory segment (which can itself consists of multiple contiguous blocks)
- ZSTD_compressContinue() detects that prior input has been overwritten when `src` buffer overlaps.
In which case, it will "discard" the relevant memory section from its history.
Finish a frame with ZSTD_compressEnd(), which will write the last block(s) and optional checksum.
It's possible to use srcSize==0, in which case, it will write a final empty block to end the frame.
Without last block mark, frames are considered unfinished (hence corrupted) by compliant decoders.
`ZSTD_CCtx` object can be re-used (ZSTD_compressBegin()) to compress again.
*/
/*===== Buffer-less streaming compression functions =====*/
ZSTDLIB_API size_t ZSTD_compressBegin(ZSTD_CCtx* cctx, int compressionLevel);
ZSTDLIB_API size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, int compressionLevel);
ZSTDLIB_API size_t ZSTD_compressBegin_advanced(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, ZSTD_parameters params, unsigned long long pledgedSrcSize); /**< pledgedSrcSize : If srcSize is not known at init time, use ZSTD_CONTENTSIZE_UNKNOWN */
ZSTDLIB_API size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict); /**< note: fails if cdict==NULL */
ZSTDLIB_API size_t ZSTD_compressBegin_usingCDict_advanced(ZSTD_CCtx* const cctx, const ZSTD_CDict* const cdict, ZSTD_frameParameters const fParams, unsigned long long const pledgedSrcSize); /* compression parameters are already set within cdict. pledgedSrcSize must be correct. If srcSize is not known, use macro ZSTD_CONTENTSIZE_UNKNOWN */
ZSTDLIB_API size_t ZSTD_copyCCtx(ZSTD_CCtx* cctx, const ZSTD_CCtx* preparedCCtx, unsigned long long pledgedSrcSize); /**< note: if pledgedSrcSize is not known, use ZSTD_CONTENTSIZE_UNKNOWN */
ZSTDLIB_API size_t ZSTD_compressContinue(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
ZSTDLIB_API size_t ZSTD_compressEnd(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
/*-
Buffer-less streaming decompression (synchronous mode)
A ZSTD_DCtx object is required to track streaming operations.
Use ZSTD_createDCtx() / ZSTD_freeDCtx() to manage it.
A ZSTD_DCtx object can be re-used multiple times.
First typical operation is to retrieve frame parameters, using ZSTD_getFrameHeader().
Frame header is extracted from the beginning of compressed frame, so providing only the frame's beginning is enough.
Data fragment must be large enough to ensure successful decoding.
`ZSTD_frameHeaderSize_max` bytes is guaranteed to always be large enough.
@result : 0 : successful decoding, the `ZSTD_frameHeader` structure is correctly filled.
>0 : `srcSize` is too small, please provide at least @result bytes on next attempt.
errorCode, which can be tested using ZSTD_isError().
It fills a ZSTD_frameHeader structure with important information to correctly decode the frame,
such as the dictionary ID, content size, or maximum back-reference distance (`windowSize`).
Note that these values could be wrong, either because of data corruption, or because a 3rd party deliberately spoofs false information.
As a consequence, check that values remain within valid application range.
For example, do not allocate memory blindly, check that `windowSize` is within expectation.
Each application can set its own limits, depending on local restrictions.
For extended interoperability, it is recommended to support `windowSize` of at least 8 MB.
ZSTD_decompressContinue() needs previous data blocks during decompression, up to `windowSize` bytes.
ZSTD_decompressContinue() is very sensitive to contiguity,
if 2 blocks don't follow each other, make sure that either the compressor breaks contiguity at the same place,
or that previous contiguous segment is large enough to properly handle maximum back-reference distance.
There are multiple ways to guarantee this condition.
The most memory efficient way is to use a round buffer of sufficient size.
Sufficient size is determined by invoking ZSTD_decodingBufferSize_min(),
which can @return an error code if required value is too large for current system (in 32-bits mode).
In a round buffer methodology, ZSTD_decompressContinue() decompresses each block next to previous one,
up to the moment there is not enough room left in the buffer to guarantee decoding another full block,
which maximum size is provided in `ZSTD_frameHeader` structure, field `blockSizeMax`.
At which point, decoding can resume from the beginning of the buffer.
Note that already decoded data stored in the buffer should be flushed before being overwritten.
There are alternatives possible, for example using two or more buffers of size `windowSize` each, though they consume more memory.
Finally, if you control the compression process, you can also ignore all buffer size rules,
as long as the encoder and decoder progress in "lock-step",
aka use exactly the same buffer sizes, break contiguity at the same place, etc.
Once buffers are setup, start decompression, with ZSTD_decompressBegin().
If decompression requires a dictionary, use ZSTD_decompressBegin_usingDict() or ZSTD_decompressBegin_usingDDict().
Then use ZSTD_nextSrcSizeToDecompress() and ZSTD_decompressContinue() alternatively.
ZSTD_nextSrcSizeToDecompress() tells how many bytes to provide as 'srcSize' to ZSTD_decompressContinue().
ZSTD_decompressContinue() requires this _exact_ amount of bytes, or it will fail.
@result of ZSTD_decompressContinue() is the number of bytes regenerated within 'dst' (necessarily <= dstCapacity).
It can be zero : it just means ZSTD_decompressContinue() has decoded some metadata item.
It can also be an error code, which can be tested with ZSTD_isError().
A frame is fully decoded when ZSTD_nextSrcSizeToDecompress() returns zero.
Context can then be reset to start a new decompression.
Note : it's possible to know if next input to present is a header or a block, using ZSTD_nextInputType().
This information is not required to properly decode a frame.
== Special case : skippable frames ==
Skippable frames allow integration of user-defined data into a flow of concatenated frames.
Skippable frames will be ignored (skipped) by decompressor.
The format of skippable frames is as follows :
a) Skippable frame ID - 4 Bytes, Little endian format, any value from 0x184D2A50 to 0x184D2A5F
b) Frame Size - 4 Bytes, Little endian format, unsigned 32-bits
c) Frame Content - any content (User Data) of length equal to Frame Size
For skippable frames ZSTD_getFrameHeader() returns zfhPtr->frameType==ZSTD_skippableFrame.
For skippable frames ZSTD_decompressContinue() always returns 0 : it only skips the content.
*/
/*===== Buffer-less streaming decompression functions =====*/
typedef enum { ZSTD_frame, ZSTD_skippableFrame } ZSTD_frameType_e;
typedef struct {
unsigned long long frameContentSize; /* if == ZSTD_CONTENTSIZE_UNKNOWN, it means this field is not available. 0 means "empty" */
unsigned long long windowSize; /* can be very large, up to <= frameContentSize */
unsigned blockSizeMax;
ZSTD_frameType_e frameType; /* if == ZSTD_skippableFrame, frameContentSize is the size of skippable content */
unsigned headerSize;
unsigned dictID;
unsigned checksumFlag;
} ZSTD_frameHeader;
/*! ZSTD_getFrameHeader() :
* decode Frame Header, or requires larger `srcSize`.
* @return : 0, `zfhPtr` is correctly filled,
* >0, `srcSize` is too small, value is wanted `srcSize` amount,
* or an error code, which can be tested using ZSTD_isError() */
ZSTDLIB_API size_t ZSTD_getFrameHeader(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize); /**< doesn't consume input */
/*! ZSTD_getFrameHeader_advanced() :
* same as ZSTD_getFrameHeader(),
* with added capability to select a format (like ZSTD_f_zstd1_magicless) */
ZSTDLIB_API size_t ZSTD_getFrameHeader_advanced(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize, ZSTD_format_e format);
ZSTDLIB_API size_t ZSTD_decodingBufferSize_min(unsigned long long windowSize, unsigned long long frameContentSize); /**< when frame content size is not known, pass in frameContentSize == ZSTD_CONTENTSIZE_UNKNOWN */
ZSTDLIB_API size_t ZSTD_decompressBegin(ZSTD_DCtx* dctx);
ZSTDLIB_API size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);
ZSTDLIB_API size_t ZSTD_decompressBegin_usingDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict);
ZSTDLIB_API size_t ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx* dctx);
ZSTDLIB_API size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
/* misc */
ZSTDLIB_API void ZSTD_copyDCtx(ZSTD_DCtx* dctx, const ZSTD_DCtx* preparedDCtx);
typedef enum { ZSTDnit_frameHeader, ZSTDnit_blockHeader, ZSTDnit_block, ZSTDnit_lastBlock, ZSTDnit_checksum, ZSTDnit_skippableFrame } ZSTD_nextInputType_e;
ZSTDLIB_API ZSTD_nextInputType_e ZSTD_nextInputType(ZSTD_DCtx* dctx);
/* ============================ */
/** Block level API */
/* ============================ */
/*!
Block functions produce and decode raw zstd blocks, without frame metadata.
Frame metadata cost is typically ~12 bytes, which can be non-negligible for very small blocks (< 100 bytes).
But users will have to take in charge needed metadata to regenerate data, such as compressed and content sizes.
A few rules to respect :
- Compressing and decompressing require a context structure
+ Use ZSTD_createCCtx() and ZSTD_createDCtx()
- It is necessary to init context before starting
+ compression : any ZSTD_compressBegin*() variant, including with dictionary
+ decompression : any ZSTD_decompressBegin*() variant, including with dictionary
+ copyCCtx() and copyDCtx() can be used too
- Block size is limited, it must be <= ZSTD_getBlockSize() <= ZSTD_BLOCKSIZE_MAX == 128 KB
+ If input is larger than a block size, it's necessary to split input data into multiple blocks
+ For inputs larger than a single block, consider using regular ZSTD_compress() instead.
Frame metadata is not that costly, and quickly becomes negligible as source size grows larger than a block.
- When a block is considered not compressible enough, ZSTD_compressBlock() result will be 0 (zero) !
===> In which case, nothing is produced into `dst` !
+ User __must__ test for such outcome and deal directly with uncompressed data
+ A block cannot be declared incompressible if ZSTD_compressBlock() return value was != 0.
Doing so would mess up with statistics history, leading to potential data corruption.
+ ZSTD_decompressBlock() _doesn't accept uncompressed data as input_ !!
+ In case of multiple successive blocks, should some of them be uncompressed,
decoder must be informed of their existence in order to follow proper history.
Use ZSTD_insertBlock() for such a case.
*/
/*===== Raw zstd block functions =====*/
ZSTDLIB_API size_t ZSTD_getBlockSize (const ZSTD_CCtx* cctx);
ZSTDLIB_API size_t ZSTD_compressBlock (ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
ZSTDLIB_API size_t ZSTD_decompressBlock(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
ZSTDLIB_API size_t ZSTD_insertBlock (ZSTD_DCtx* dctx, const void* blockStart, size_t blockSize); /**< insert uncompressed block into `dctx` history. Useful for multi-blocks decompression. */
#endif /* ZSTD_H_ZSTD_STATIC_LINKING_ONLY */
#if defined (__cplusplus)
}
#endif
/*
* Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
* LICENSE file in the root directory of this source tree) and the GPLv2 (found
* in the COPYING file in the root directory of this source tree).
* You may select, at your option, one of the above-listed licenses.
*/
#ifndef ZSTD_ERRORS_H_398273423
#define ZSTD_ERRORS_H_398273423
#if defined (__cplusplus)
extern "C" {
#endif
/*===== dependency =====*/
#include <stddef.h> /* size_t */
/* ===== ZSTDERRORLIB_API : control library symbols visibility ===== */
#ifndef ZSTDERRORLIB_VISIBILITY
# if defined(__GNUC__) && (__GNUC__ >= 4)
# define ZSTDERRORLIB_VISIBILITY __attribute__ ((visibility ("default")))
# else
# define ZSTDERRORLIB_VISIBILITY
# endif
#endif
#if defined(ZSTD_DLL_EXPORT) && (ZSTD_DLL_EXPORT==1)
# define ZSTDERRORLIB_API __declspec(dllexport) ZSTDERRORLIB_VISIBILITY
#elif defined(ZSTD_DLL_IMPORT) && (ZSTD_DLL_IMPORT==1)
# define ZSTDERRORLIB_API __declspec(dllimport) ZSTDERRORLIB_VISIBILITY /* It isn't required but allows to generate better code, saving a function pointer load from the IAT and an indirect jump.*/
#else
# define ZSTDERRORLIB_API ZSTDERRORLIB_VISIBILITY
#endif
/*-*********************************************
* Error codes list
*-*********************************************
* Error codes _values_ are pinned down since v1.3.1 only.
* Therefore, don't rely on values if you may link to any version < v1.3.1.
*
* Only values < 100 are considered stable.
*
* note 1 : this API shall be used with static linking only.
* dynamic linking is not yet officially supported.
* note 2 : Prefer relying on the enum than on its value whenever possible
* This is the only supported way to use the error list < v1.3.1
* note 3 : ZSTD_isError() is always correct, whatever the library version.
**********************************************/
typedef enum {
ZSTD_error_no_error = 0,
ZSTD_error_GENERIC = 1,
ZSTD_error_prefix_unknown = 10,
ZSTD_error_version_unsupported = 12,
ZSTD_error_frameParameter_unsupported = 14,
ZSTD_error_frameParameter_windowTooLarge = 16,
ZSTD_error_corruption_detected = 20,
ZSTD_error_checksum_wrong = 22,
ZSTD_error_dictionary_corrupted = 30,
ZSTD_error_dictionary_wrong = 32,
ZSTD_error_dictionaryCreation_failed = 34,
ZSTD_error_parameter_unsupported = 40,
ZSTD_error_parameter_outOfBound = 42,
ZSTD_error_tableLog_tooLarge = 44,
ZSTD_error_maxSymbolValue_tooLarge = 46,
ZSTD_error_maxSymbolValue_tooSmall = 48,
ZSTD_error_stage_wrong = 60,
ZSTD_error_init_missing = 62,
ZSTD_error_memory_allocation = 64,
ZSTD_error_workSpace_tooSmall= 66,
ZSTD_error_dstSize_tooSmall = 70,
ZSTD_error_srcSize_wrong = 72,
ZSTD_error_dstBuffer_null = 74,
/* following error codes are __NOT STABLE__, they can be removed or changed in future versions */
ZSTD_error_frameIndex_tooLarge = 100,
ZSTD_error_seekableIO = 102,
ZSTD_error_maxCode = 120 /* never EVER use this value directly, it can change in future versions! Use ZSTD_isError() instead */
} ZSTD_ErrorCode;
/*! ZSTD_getErrorCode() :
convert a `size_t` function result into a `ZSTD_ErrorCode` enum type,
which can be used to compare with enum list published above */
ZSTDERRORLIB_API ZSTD_ErrorCode ZSTD_getErrorCode(size_t functionResult);
ZSTDERRORLIB_API const char* ZSTD_getErrorString(ZSTD_ErrorCode code); /**< Same as ZSTD_getErrorName, but using a `ZSTD_ErrorCode` enum argument */
#if defined (__cplusplus)
}
#endif
#endif /* ZSTD_ERRORS_H_398273423 */
/*
* xxHash - Extremely Fast Hash algorithm
* Header File
* Copyright (C) 2012-present, Yann Collet.
*
* BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php)
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are
* met:
*
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following disclaimer
* in the documentation and/or other materials provided with the
* distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
* A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
* OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
* You can contact the author at:
* - xxHash homepage: https://www.xxhash.com
* - xxHash source repository: https://github.com/Cyan4973/xxHash
*/
/* TODO: update */
/* Notice extracted from xxHash homepage:
xxHash is an extremely fast hash algorithm, running at RAM speed limits.
It also successfully passes all tests from the SMHasher suite.
Comparison (single thread, Windows Seven 32 bits, using SMHasher on a Core 2 Duo @3GHz)
Name Speed Q.Score Author
xxHash 5.4 GB/s 10
CrapWow 3.2 GB/s 2 Andrew
MumurHash 3a 2.7 GB/s 10 Austin Appleby
SpookyHash 2.0 GB/s 10 Bob Jenkins
SBox 1.4 GB/s 9 Bret Mulvey
Lookup3 1.2 GB/s 9 Bob Jenkins
SuperFastHash 1.2 GB/s 1 Paul Hsieh
CityHash64 1.05 GB/s 10 Pike & Alakuijala
FNV 0.55 GB/s 5 Fowler, Noll, Vo
CRC32 0.43 GB/s 9
MD5-32 0.33 GB/s 10 Ronald L. Rivest
SHA1-32 0.28 GB/s 10
Q.Score is a measure of quality of the hash function.
It depends on successfully passing SMHasher test set.
10 is a perfect score.
Note: SMHasher's CRC32 implementation is not the fastest one.
Other speed-oriented implementations can be faster,
especially in combination with PCLMUL instruction:
https://fastcompression.blogspot.com/2019/03/presenting-xxh3.html?showComment=1552696407071#c3490092340461170735
A 64-bit version, named XXH64, is available since r35.
It offers much better speed, but for 64-bit applications only.
Name Speed on 64 bits Speed on 32 bits
XXH64 13.8 GB/s 1.9 GB/s
XXH32 6.8 GB/s 6.0 GB/s
*/
#if defined (__cplusplus)
extern "C" {
#endif
/* ****************************
* INLINE mode
******************************/
/*!
* XXH_INLINE_ALL (and XXH_PRIVATE_API)
* Use these build macros to inline xxhash into the target unit.
* Inlining improves performance on small inputs, especially when the length is
* expressed as a compile-time constant:
*
* https://fastcompression.blogspot.com/2018/03/xxhash-for-small-keys-impressive-power.html
*
* It also keeps xxHash symbols private to the unit, so they are not exported.
*
* Usage:
* #define XXH_INLINE_ALL
* #include "xxhash.h"
*
* Do not compile and link xxhash.o as a separate object, as it is not useful.
*/
#if (defined(XXH_INLINE_ALL) || defined(XXH_PRIVATE_API)) \
&& !defined(XXH_INLINE_ALL_31684351384)
/* this section should be traversed only once */
# define XXH_INLINE_ALL_31684351384
/* give access to the advanced API, required to compile implementations */
# undef XXH_STATIC_LINKING_ONLY /* avoid macro redef */
# define XXH_STATIC_LINKING_ONLY
/* make all functions private */
# undef XXH_PUBLIC_API
# if defined(__GNUC__)
# define XXH_PUBLIC_API static __inline __attribute__((unused))
# elif defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */)
# define XXH_PUBLIC_API static inline
# elif defined(_MSC_VER)
# define XXH_PUBLIC_API static __inline
# else
/* note: this version may generate warnings for unused static functions */
# define XXH_PUBLIC_API static
# endif
/*
* This part deals with the special case where a unit wants to inline xxHash,
* but "xxhash.h" has previously been included without XXH_INLINE_ALL, such
* as part of some previously included *.h header file.
* Without further action, the new include would just be ignored,
* and functions would effectively _not_ be inlined (silent failure).
* The following macros solve this situation by prefixing all inlined names,
* avoiding naming collision with previous inclusions.
*/
# ifdef XXH_NAMESPACE
# error "XXH_INLINE_ALL with XXH_NAMESPACE is not supported"
/*
* Note: Alternative: #undef all symbols (it's a pretty large list).
* Without #error: it compiles, but functions are actually not inlined.
*/
# endif
# define XXH_NAMESPACE XXH_INLINE_
/*
* Some identifiers (enums, type names) are not symbols, but they must
* still be renamed to avoid redeclaration.
* Alternative solution: do not redeclare them.
* However, this requires some #ifdefs, and is a more dispersed action.
* Meanwhile, renaming can be achieved in a single block
*/
# define XXH_IPREF(Id) XXH_INLINE_ ## Id
# define XXH_OK XXH_IPREF(XXH_OK)
# define XXH_ERROR XXH_IPREF(XXH_ERROR)
# define XXH_errorcode XXH_IPREF(XXH_errorcode)
# define XXH32_canonical_t XXH_IPREF(XXH32_canonical_t)
# define XXH64_canonical_t XXH_IPREF(XXH64_canonical_t)
# define XXH128_canonical_t XXH_IPREF(XXH128_canonical_t)
# define XXH32_state_s XXH_IPREF(XXH32_state_s)
# define XXH32_state_t XXH_IPREF(XXH32_state_t)
# define XXH64_state_s XXH_IPREF(XXH64_state_s)
# define XXH64_state_t XXH_IPREF(XXH64_state_t)
# define XXH3_state_s XXH_IPREF(XXH3_state_s)
# define XXH3_state_t XXH_IPREF(XXH3_state_t)
# define XXH128_hash_t XXH_IPREF(XXH128_hash_t)
/* Ensure the header is parsed again, even if it was previously included */
# undef XXHASH_H_5627135585666179
# undef XXHASH_H_STATIC_13879238742
#endif /* XXH_INLINE_ALL || XXH_PRIVATE_API */
/* ****************************************************************
* Stable API
*****************************************************************/
#ifndef XXHASH_H_5627135585666179
#define XXHASH_H_5627135585666179 1
/* specific declaration modes for Windows */
#if !defined(XXH_INLINE_ALL) && !defined(XXH_PRIVATE_API)
# if defined(WIN32) && defined(_MSC_VER) && (defined(XXH_IMPORT) || defined(XXH_EXPORT))
# ifdef XXH_EXPORT
# define XXH_PUBLIC_API __declspec(dllexport)
# elif XXH_IMPORT
# define XXH_PUBLIC_API __declspec(dllimport)
# endif
# else
# define XXH_PUBLIC_API /* do nothing */
# endif
#endif
/*!
* XXH_NAMESPACE, aka Namespace Emulation:
*
* If you want to include _and expose_ xxHash functions from within your own
* library, but also want to avoid symbol collisions with other libraries which
* may also include xxHash, you can use XXH_NAMESPACE to automatically prefix
* any public symbol from xxhash library with the value of XXH_NAMESPACE
* (therefore, avoid empty or numeric values).
*
* Note that no change is required within the calling program as long as it
* includes `xxhash.h`: Regular symbol names will be automatically translated
* by this header.
*/
#ifdef XXH_NAMESPACE
# define XXH_CAT(A,B) A##B
# define XXH_NAME2(A,B) XXH_CAT(A,B)
# define XXH_versionNumber XXH_NAME2(XXH_NAMESPACE, XXH_versionNumber)
# define XXH32 XXH_NAME2(XXH_NAMESPACE, XXH32)
# define XXH32_createState XXH_NAME2(XXH_NAMESPACE, XXH32_createState)
# define XXH32_freeState XXH_NAME2(XXH_NAMESPACE, XXH32_freeState)
# define XXH32_reset XXH_NAME2(XXH_NAMESPACE, XXH32_reset)
# define XXH32_update XXH_NAME2(XXH_NAMESPACE, XXH32_update)
# define XXH32_digest XXH_NAME2(XXH_NAMESPACE, XXH32_digest)
# define XXH32_copyState XXH_NAME2(XXH_NAMESPACE, XXH32_copyState)
# define XXH32_canonicalFromHash XXH_NAME2(XXH_NAMESPACE, XXH32_canonicalFromHash)
# define XXH32_hashFromCanonical XXH_NAME2(XXH_NAMESPACE, XXH32_hashFromCanonical)
# define XXH64 XXH_NAME2(XXH_NAMESPACE, XXH64)
# define XXH64_createState XXH_NAME2(XXH_NAMESPACE, XXH64_createState)
# define XXH64_freeState XXH_NAME2(XXH_NAMESPACE, XXH64_freeState)
# define XXH64_reset XXH_NAME2(XXH_NAMESPACE, XXH64_reset)
# define XXH64_update XXH_NAME2(XXH_NAMESPACE, XXH64_update)
# define XXH64_digest XXH_NAME2(XXH_NAMESPACE, XXH64_digest)
# define XXH64_copyState XXH_NAME2(XXH_NAMESPACE, XXH64_copyState)
# define XXH64_canonicalFromHash XXH_NAME2(XXH_NAMESPACE, XXH64_canonicalFromHash)
# define XXH64_hashFromCanonical XXH_NAME2(XXH_NAMESPACE, XXH64_hashFromCanonical)
#endif
/* *************************************
* Version
***************************************/
#define XXH_VERSION_MAJOR 0
#define XXH_VERSION_MINOR 7
#define XXH_VERSION_RELEASE 3
#define XXH_VERSION_NUMBER (XXH_VERSION_MAJOR *100*100 + XXH_VERSION_MINOR *100 + XXH_VERSION_RELEASE)
XXH_PUBLIC_API unsigned XXH_versionNumber (void);
/* ****************************
* Definitions
******************************/
#include <stddef.h> /* size_t */
typedef enum { XXH_OK=0, XXH_ERROR } XXH_errorcode;
/*-**********************************************************************
* 32-bit hash
************************************************************************/
#if !defined (__VMS) \
&& (defined (__cplusplus) \
|| (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) )
# include <stdint.h>
typedef uint32_t XXH32_hash_t;
#else
# include <limits.h>
# if UINT_MAX == 0xFFFFFFFFUL
typedef unsigned int XXH32_hash_t;
# else
# if ULONG_MAX == 0xFFFFFFFFUL
typedef unsigned long XXH32_hash_t;
# else
# error "unsupported platform: need a 32-bit type"
# endif
# endif
#endif
/*!
* XXH32():
* Calculate the 32-bit hash of sequence "length" bytes stored at memory address "input".
* The memory between input & input+length must be valid (allocated and read-accessible).
* "seed" can be used to alter the result predictably.
* Speed on Core 2 Duo @ 3 GHz (single thread, SMHasher benchmark): 5.4 GB/s
*/
XXH_PUBLIC_API XXH32_hash_t XXH32 (const void* input, size_t length, XXH32_hash_t seed);
/******* Streaming *******/
/*
* Streaming functions generate the xxHash value from an incrememtal input.
* This method is slower than single-call functions, due to state management.
* For small inputs, prefer `XXH32()` and `XXH64()`, which are better optimized.
*
* An XXH state must first be allocated using `XXH*_createState()`.
*
* Start a new hash by initializing the state with a seed using `XXH*_reset()`.
*
* Then, feed the hash state by calling `XXH*_update()` as many times as necessary.
*
* The function returns an error code, with 0 meaning OK, and any other value
* meaning there is an error.
*
* Finally, a hash value can be produced anytime, by using `XXH*_digest()`.
* This function returns the nn-bits hash as an int or long long.
*
* It's still possible to continue inserting input into the hash state after a
* digest, and generate new hash values later on by invoking `XXH*_digest()`.
*
* When done, release the state using `XXH*_freeState()`.
*/
typedef struct XXH32_state_s XXH32_state_t; /* incomplete type */
XXH_PUBLIC_API XXH32_state_t* XXH32_createState(void);
XXH_PUBLIC_API XXH_errorcode XXH32_freeState(XXH32_state_t* statePtr);
XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* dst_state, const XXH32_state_t* src_state);
XXH_PUBLIC_API XXH_errorcode XXH32_reset (XXH32_state_t* statePtr, XXH32_hash_t seed);
XXH_PUBLIC_API XXH_errorcode XXH32_update (XXH32_state_t* statePtr, const void* input, size_t length);
XXH_PUBLIC_API XXH32_hash_t XXH32_digest (const XXH32_state_t* statePtr);
/******* Canonical representation *******/
/*
* The default return values from XXH functions are unsigned 32 and 64 bit
* integers.
* This the simplest and fastest format for further post-processing.
*
* However, this leaves open the question of what is the order on the byte level,
* since little and big endian conventions will store the same number differently.
*
* The canonical representation settles this issue by mandating big-endian
* convention, the same convention as human-readable numbers (large digits first).
*
* When writing hash values to storage, sending them over a network, or printing
* them, it's highly recommended to use the canonical representation to ensure
* portability across a wider range of systems, present and future.
*
* The following functions allow transformation of hash values to and from
* canonical format.
*/
typedef struct { unsigned char digest[4]; } XXH32_canonical_t;
XXH_PUBLIC_API void XXH32_canonicalFromHash(XXH32_canonical_t* dst, XXH32_hash_t hash);
XXH_PUBLIC_API XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src);
#ifndef XXH_NO_LONG_LONG
/*-**********************************************************************
* 64-bit hash
************************************************************************/
#if !defined (__VMS) \
&& (defined (__cplusplus) \
|| (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) )
# include <stdint.h>
typedef uint64_t XXH64_hash_t;
#else
/* the following type must have a width of 64-bit */
typedef unsigned long long XXH64_hash_t;
#endif
/*!
* XXH64():
* Returns the 64-bit hash of sequence of length @length stored at memory
* address @input.
* @seed can be used to alter the result predictably.
* This function usually runs faster on 64-bit systems, but slower on 32-bit
* systems (see benchmark).
*/
XXH_PUBLIC_API XXH64_hash_t XXH64 (const void* input, size_t length, XXH64_hash_t seed);
/******* Streaming *******/
typedef struct XXH64_state_s XXH64_state_t; /* incomplete type */
XXH_PUBLIC_API XXH64_state_t* XXH64_createState(void);
XXH_PUBLIC_API XXH_errorcode XXH64_freeState(XXH64_state_t* statePtr);
XXH_PUBLIC_API void XXH64_copyState(XXH64_state_t* dst_state, const XXH64_state_t* src_state);
XXH_PUBLIC_API XXH_errorcode XXH64_reset (XXH64_state_t* statePtr, XXH64_hash_t seed);
XXH_PUBLIC_API XXH_errorcode XXH64_update (XXH64_state_t* statePtr, const void* input, size_t length);
XXH_PUBLIC_API XXH64_hash_t XXH64_digest (const XXH64_state_t* statePtr);
/******* Canonical representation *******/
typedef struct { unsigned char digest[8]; } XXH64_canonical_t;
XXH_PUBLIC_API void XXH64_canonicalFromHash(XXH64_canonical_t* dst, XXH64_hash_t hash);
XXH_PUBLIC_API XXH64_hash_t XXH64_hashFromCanonical(const XXH64_canonical_t* src);
#endif /* XXH_NO_LONG_LONG */
#endif /* XXHASH_H_5627135585666179 */
#if defined(XXH_STATIC_LINKING_ONLY) && !defined(XXHASH_H_STATIC_13879238742)
#define XXHASH_H_STATIC_13879238742
/* ****************************************************************************
* This section contains declarations which are not guaranteed to remain stable.
* They may change in future versions, becoming incompatible with a different
* version of the library.
* These declarations should only be used with static linking.
* Never use them in association with dynamic linking!
***************************************************************************** */
/*
* These definitions are only present to allow static allocation of an XXH
* state, for example, on the stack or in a struct.
* Never **ever** access members directly.
*/
struct XXH32_state_s {
XXH32_hash_t total_len_32;
XXH32_hash_t large_len;
XXH32_hash_t v1;
XXH32_hash_t v2;
XXH32_hash_t v3;
XXH32_hash_t v4;
XXH32_hash_t mem32[4];
XXH32_hash_t memsize;
XXH32_hash_t reserved; /* never read nor write, might be removed in a future version */
}; /* typedef'd to XXH32_state_t */
#ifndef XXH_NO_LONG_LONG /* defined when there is no 64-bit support */
struct XXH64_state_s {
XXH64_hash_t total_len;
XXH64_hash_t v1;
XXH64_hash_t v2;
XXH64_hash_t v3;
XXH64_hash_t v4;
XXH64_hash_t mem64[4];
XXH32_hash_t memsize;
XXH32_hash_t reserved32; /* required for padding anyway */
XXH64_hash_t reserved64; /* never read nor write, might be removed in a future version */
}; /* typedef'd to XXH64_state_t */
/*-**********************************************************************
* XXH3
* New experimental hash
************************************************************************/
/* ************************************************************************
* XXH3 is a new hash algorithm featuring:
* - Improved speed for both small and large inputs
* - True 64-bit and 128-bit outputs
* - SIMD acceleration
* - Improved 32-bit viability
*
* Speed analysis methodology is explained here:
*
* https://fastcompression.blogspot.com/2019/03/presenting-xxh3.html
*
* In general, expect XXH3 to run about ~2x faster on large inputs and >3x
* faster on small ones compared to XXH64, though exact differences depend on
* the platform.
*
* The algorithm is portable: Like XXH32 and XXH64, it generates the same hash
* on all platforms.
*
* It benefits greatly from SIMD and 64-bit arithmetic, but does not require it.
*
* Almost all 32-bit and 64-bit targets that can run XXH32 smoothly can run
* XXH3 at usable speeds, even if XXH64 runs slowly. Further details are
* explained in the implementation.
*
* Optimized implementations are provided for AVX2, SSE2, NEON, POWER8, ZVector,
* and scalar targets. This can be controlled with the XXH_VECTOR macro.
*
* XXH3 offers 2 variants, _64bits and _128bits.
* When only 64 bits are needed, prefer calling the _64bits variant, as it
* reduces the amount of mixing, resulting in faster speed on small inputs.
*
* It's also generally simpler to manipulate a scalar return type than a struct.
*
* The 128-bit version adds additional strength, but it is slightly slower.
*
* The XXH3 algorithm is still in development.
* The results it produces may still change in future versions.
*
* Results produced by v0.7.x are not comparable with results from v0.7.y.
* However, the API is completely stable, and it can safely be used for
* ephemeral data (local sessions).
*
* Avoid storing values in long-term storage until the algorithm is finalized.
*
* The API supports one-shot hashing, streaming mode, and custom secrets.
*/
#ifdef XXH_NAMESPACE
# define XXH3_64bits XXH_NAME2(XXH_NAMESPACE, XXH3_64bits)
# define XXH3_64bits_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_withSecret)
# define XXH3_64bits_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_withSeed)
# define XXH3_createState XXH_NAME2(XXH_NAMESPACE, XXH3_createState)
# define XXH3_freeState XXH_NAME2(XXH_NAMESPACE, XXH3_freeState)
# define XXH3_copyState XXH_NAME2(XXH_NAMESPACE, XXH3_copyState)
# define XXH3_64bits_reset XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset)
# define XXH3_64bits_reset_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset_withSeed)
# define XXH3_64bits_reset_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset_withSecret)
# define XXH3_64bits_update XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_update)
# define XXH3_64bits_digest XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_digest)
#endif
/* XXH3_64bits():
* default 64-bit variant, using default secret and default seed of 0.
* It's the fastest variant. */
XXH_PUBLIC_API XXH64_hash_t XXH3_64bits(const void* data, size_t len);
/*
* XXH3_64bits_withSecret():
* It's possible to provide any blob of bytes as a "secret" to generate the hash.
* This makes it more difficult for an external actor to prepare an intentional
* collision.
* The secret *must* be large enough (>= XXH3_SECRET_SIZE_MIN).
* It should consist of random bytes.
* Avoid trivial sequences, such as repeating sequences and especially '\0',
* as this can cancel out itself.
* Failure to respect these conditions will result in a poor quality hash.
*/
#define XXH3_SECRET_SIZE_MIN 136
XXH_PUBLIC_API XXH64_hash_t XXH3_64bits_withSecret(const void* data, size_t len, const void* secret, size_t secretSize);
/*
* XXH3_64bits_withSeed():
* This variant generates a custom secret on the fly based on the default
* secret, altered using the `seed` value.
* While this operation is decently fast, note that it's not completely free.
* Note: seed==0 produces the same results as XXH3_64bits().
*/
XXH_PUBLIC_API XXH64_hash_t XXH3_64bits_withSeed(const void* data, size_t len, XXH64_hash_t seed);
/* streaming 64-bit */
#if defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L) /* C11+ */
# include <stdalign.h>
# define XXH_ALIGN(n) alignas(n)
#elif defined(__GNUC__)
# define XXH_ALIGN(n) __attribute__ ((aligned(n)))
#elif defined(_MSC_VER)
# define XXH_ALIGN(n) __declspec(align(n))
#else
# define XXH_ALIGN(n) /* disabled */
#endif
/* Old GCC versions only accept the attribute after the type in structures. */
#if !(defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L)) /* C11+ */ \
&& defined(__GNUC__)
# define XXH_ALIGN_MEMBER(align, type) type XXH_ALIGN(align)
#else
# define XXH_ALIGN_MEMBER(align, type) XXH_ALIGN(align) type
#endif
typedef struct XXH3_state_s XXH3_state_t;
#define XXH3_SECRET_DEFAULT_SIZE 192 /* minimum XXH3_SECRET_SIZE_MIN */
#define XXH3_INTERNALBUFFER_SIZE 256
struct XXH3_state_s {
XXH_ALIGN_MEMBER(64, XXH64_hash_t acc[8]);
/* used to store a custom secret generated from the seed. Makes state larger.
* Design might change */
XXH_ALIGN_MEMBER(64, unsigned char customSecret[XXH3_SECRET_DEFAULT_SIZE]);
XXH_ALIGN_MEMBER(64, unsigned char buffer[XXH3_INTERNALBUFFER_SIZE]);
XXH32_hash_t bufferedSize;
XXH32_hash_t nbStripesPerBlock;
XXH32_hash_t nbStripesSoFar;
XXH32_hash_t secretLimit;
XXH32_hash_t reserved32;
XXH32_hash_t reserved32_2;
XXH64_hash_t totalLen;
XXH64_hash_t seed;
XXH64_hash_t reserved64;
/* note: there is some padding after due to alignment on 64 bytes */
const unsigned char* secret;
}; /* typedef'd to XXH3_state_t */
#undef XXH_ALIGN_MEMBER
/*
* Streaming requires state maintenance.
* This operation costs memory and CPU.
* As a consequence, streaming is slower than one-shot hashing.
* For better performance, prefer one-shot functions whenever possible.
*/
XXH_PUBLIC_API XXH3_state_t* XXH3_createState(void);
XXH_PUBLIC_API XXH_errorcode XXH3_freeState(XXH3_state_t* statePtr);
XXH_PUBLIC_API void XXH3_copyState(XXH3_state_t* dst_state, const XXH3_state_t* src_state);
/*
* XXH3_64bits_reset():
* Initialize with the default parameters.
* The result will be equivalent to `XXH3_64bits()`.
*/
XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset(XXH3_state_t* statePtr);
/*
* XXH3_64bits_reset_withSeed():
* Generate a custom secret from `seed`, and store it into `statePtr`.
* digest will be equivalent to `XXH3_64bits_withSeed()`.
*/
XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSeed(XXH3_state_t* statePtr, XXH64_hash_t seed);
/*
* XXH3_64bits_reset_withSecret():
* `secret` is referenced, and must outlive the hash streaming session, so
* be careful when using stack arrays.
* `secretSize` must be >= `XXH3_SECRET_SIZE_MIN`.
*/
XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSecret(XXH3_state_t* statePtr, const void* secret, size_t secretSize);
XXH_PUBLIC_API XXH_errorcode XXH3_64bits_update (XXH3_state_t* statePtr, const void* input, size_t length);
XXH_PUBLIC_API XXH64_hash_t XXH3_64bits_digest (const XXH3_state_t* statePtr);
/* 128-bit */
#ifdef XXH_NAMESPACE
# define XXH128 XXH_NAME2(XXH_NAMESPACE, XXH128)
# define XXH3_128bits XXH_NAME2(XXH_NAMESPACE, XXH3_128bits)
# define XXH3_128bits_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_withSeed)
# define XXH3_128bits_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_withSecret)
# define XXH3_128bits_reset XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset)
# define XXH3_128bits_reset_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset_withSeed)
# define XXH3_128bits_reset_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset_withSecret)
# define XXH3_128bits_update XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_update)
# define XXH3_128bits_digest XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_digest)
# define XXH128_isEqual XXH_NAME2(XXH_NAMESPACE, XXH128_isEqual)
# define XXH128_cmp XXH_NAME2(XXH_NAMESPACE, XXH128_cmp)
# define XXH128_canonicalFromHash XXH_NAME2(XXH_NAMESPACE, XXH128_canonicalFromHash)
# define XXH128_hashFromCanonical XXH_NAME2(XXH_NAMESPACE, XXH128_hashFromCanonical)
#endif
typedef struct {
XXH64_hash_t low64;
XXH64_hash_t high64;
} XXH128_hash_t;
XXH_PUBLIC_API XXH128_hash_t XXH128(const void* data, size_t len, XXH64_hash_t seed);
XXH_PUBLIC_API XXH128_hash_t XXH3_128bits(const void* data, size_t len);
XXH_PUBLIC_API XXH128_hash_t XXH3_128bits_withSeed(const void* data, size_t len, XXH64_hash_t seed); /* == XXH128() */
XXH_PUBLIC_API XXH128_hash_t XXH3_128bits_withSecret(const void* data, size_t len, const void* secret, size_t secretSize);
XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset(XXH3_state_t* statePtr);
XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSeed(XXH3_state_t* statePtr, XXH64_hash_t seed);
XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSecret(XXH3_state_t* statePtr, const void* secret, size_t secretSize);
XXH_PUBLIC_API XXH_errorcode XXH3_128bits_update (XXH3_state_t* statePtr, const void* input, size_t length);
XXH_PUBLIC_API XXH128_hash_t XXH3_128bits_digest (const XXH3_state_t* statePtr);
/* Note: For better performance, these functions can be inlined using XXH_INLINE_ALL */
/*!
* XXH128_isEqual():
* Return: 1 if `h1` and `h2` are equal, 0 if they are not.
*/
XXH_PUBLIC_API int XXH128_isEqual(XXH128_hash_t h1, XXH128_hash_t h2);
/*!
* XXH128_cmp():
*
* This comparator is compatible with stdlib's `qsort()`/`bsearch()`.
*
* return: >0 if *h128_1 > *h128_2
* <0 if *h128_1 < *h128_2
* =0 if *h128_1 == *h128_2
*/
XXH_PUBLIC_API int XXH128_cmp(const void* h128_1, const void* h128_2);
/******* Canonical representation *******/
typedef struct { unsigned char digest[16]; } XXH128_canonical_t;
XXH_PUBLIC_API void XXH128_canonicalFromHash(XXH128_canonical_t* dst, XXH128_hash_t hash);
XXH_PUBLIC_API XXH128_hash_t XXH128_hashFromCanonical(const XXH128_canonical_t* src);
#endif /* XXH_NO_LONG_LONG */
#if defined(XXH_INLINE_ALL) || defined(XXH_PRIVATE_API)
# define XXH_IMPLEMENTATION
#endif
#endif /* defined(XXH_STATIC_LINKING_ONLY) && !defined(XXHASH_H_STATIC_13879238742) */
/* ======================================================================== */
/* ======================================================================== */
/* ======================================================================== */
/*-**********************************************************************
* xxHash implementation
*-**********************************************************************
* xxHash's implementation used to be found in xxhash.c.
*
* However, code inlining requires the implementation to be visible to the
* compiler, usually within the header.
*
* As a workaround, xxhash.c used to be included within xxhash.h. This caused
* some issues with some build systems, especially ones which treat .c files
* as source files.
*
* Therefore, the implementation is now directly integrated within xxhash.h.
* Another small advantage is that xxhash.c is no longer needed in /include.
************************************************************************/
#if ( defined(XXH_INLINE_ALL) || defined(XXH_PRIVATE_API) \
|| defined(XXH_IMPLEMENTATION) ) && !defined(XXH_IMPLEM_13a8737387)
# define XXH_IMPLEM_13a8737387
/* *************************************
* Tuning parameters
***************************************/
/*!
* XXH_FORCE_MEMORY_ACCESS:
* By default, access to unaligned memory is controlled by `memcpy()`, which is
* safe and portable.
*
* Unfortunately, on some target/compiler combinations, the generated assembly
* is sub-optimal.
*
* The below switch allow to select a different access method for improved
* performance.
* Method 0 (default):
* Use `memcpy()`. Safe and portable.
* Method 1:
* `__attribute__((packed))` statement. It depends on compiler extensions
* and is therefore not portable.
* This method is safe if your compiler supports it, and *generally* as
* fast or faster than `memcpy`.
* Method 2:
* Direct access via cast. This method doesn't depend on the compiler but
* violates the C standard.
* It can generate buggy code on targets which do not support unaligned
* memory accesses.
* But in some circumstances, it's the only known way to get the most
* performance (ie GCC + ARMv6)
* Method 3:
* Byteshift. This can generate the best code on old compilers which don't
* inline small `memcpy()` calls, and it might also be faster on big-endian
* systems which lack a native byteswap instruction.
* See https://stackoverflow.com/a/32095106/646947 for details.
* Prefer these methods in priority order (0 > 1 > 2 > 3)
*/
#ifndef XXH_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */
# if !defined(__clang__) && defined(__GNUC__) && defined(__ARM_FEATURE_UNALIGNED) && defined(__ARM_ARCH) && (__ARM_ARCH == 6)
# define XXH_FORCE_MEMORY_ACCESS 2
# elif !defined(__clang__) && ((defined(__INTEL_COMPILER) && !defined(_WIN32)) || \
(defined(__GNUC__) && (defined(__ARM_ARCH) && __ARM_ARCH >= 7)))
# define XXH_FORCE_MEMORY_ACCESS 1
# endif
#endif
/*!
*XXH_ACCEPT_NULL_INPUT_POINTER:
* If the input pointer is NULL, xxHash's default behavior is to dereference it,
* triggering a segfault.
* When this macro is enabled, xxHash actively checks the input for a null pointer.
* If it is, the result for null input pointers is the same as a zero-length input.
*/
#ifndef XXH_ACCEPT_NULL_INPUT_POINTER /* can be defined externally */
# define XXH_ACCEPT_NULL_INPUT_POINTER 0
#endif
/*!
* XXH_FORCE_ALIGN_CHECK:
* This is a minor performance trick, only useful with lots of very small keys.
* It means: check for aligned/unaligned input.
* The check costs one initial branch per hash;
* Set it to 0 when the input is guaranteed to be aligned or when alignment
* doesn't matter for performance.
*
* This option does not affect XXH3.
*/
#ifndef XXH_FORCE_ALIGN_CHECK /* can be defined externally */
# if defined(__i386) || defined(_M_IX86) || defined(__x86_64__) || defined(_M_X64)
# define XXH_FORCE_ALIGN_CHECK 0
# else
# define XXH_FORCE_ALIGN_CHECK 1
# endif
#endif
/*!
* XXH_NO_INLINE_HINTS:
*
* By default, xxHash tries to force the compiler to inline almost all internal
* functions.
*
* This can usually improve performance due to reduced jumping and improved
* constant folding, but significantly increases the size of the binary which
* might not be favorable.
*
* Additionally, sometimes the forced inlining can be detrimental to performance,
* depending on the architecture.
*
* XXH_NO_INLINE_HINTS marks all internal functions as static, giving the
* compiler full control on whether to inline or not.
*
* When not optimizing (-O0), optimizing for size (-Os, -Oz), or using
* -fno-inline with GCC or Clang, this will automatically be defined.
*/
#ifndef XXH_NO_INLINE_HINTS
# if defined(__OPTIMIZE_SIZE__) /* -Os, -Oz */ \
|| defined(__NO_INLINE__) /* -O0, -fno-inline */
# define XXH_NO_INLINE_HINTS 1
# else
# define XXH_NO_INLINE_HINTS 0
# endif
#endif
/*!
* XXH_REROLL:
* Whether to reroll XXH32_finalize, and XXH64_finalize,
* instead of using an unrolled jump table/if statement loop.
*
* This is automatically defined on -Os/-Oz on GCC and Clang.
*/
#ifndef XXH_REROLL
# if defined(__OPTIMIZE_SIZE__)
# define XXH_REROLL 1
# else
# define XXH_REROLL 0
# endif
#endif
/* *************************************
* Includes & Memory related functions
***************************************/
/*!
* Modify the local functions below should you wish to use some other memory
* routines for malloc() and free()
*/
#include <stdlib.h>
static void* XXH_malloc(size_t s) { return malloc(s); }
static void XXH_free (void* p) { free(p); }
/*! and for memcpy() */
#include <string.h>
static void* XXH_memcpy(void* dest, const void* src, size_t size)
{
return memcpy(dest,src,size);
}
#include <limits.h> /* ULLONG_MAX */
/* *************************************
* Compiler Specific Options
***************************************/
#ifdef _MSC_VER /* Visual Studio warning fix */
# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */
#endif
#if XXH_NO_INLINE_HINTS /* disable inlining hints */
# define XXH_FORCE_INLINE static
# define XXH_NO_INLINE static
#elif defined(_MSC_VER) /* Visual Studio */
# define XXH_FORCE_INLINE static __forceinline
# define XXH_NO_INLINE static __declspec(noinline)
#else
# if defined (__cplusplus) \
|| defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 */
# ifdef __GNUC__
# define XXH_FORCE_INLINE static inline __attribute__((always_inline))
# define XXH_NO_INLINE static __attribute__((noinline))
# else
# define XXH_FORCE_INLINE static inline
# define XXH_NO_INLINE static
# endif
# else
# define XXH_FORCE_INLINE static
# define XXH_NO_INLINE static
# endif /* __STDC_VERSION__ */
#endif
/* *************************************
* Debug
***************************************/
/*
* DEBUGLEVEL is expected to be defined externally, typically via the compiler's
* command line options. The value must be a number.
*/
#ifndef DEBUGLEVEL
# define DEBUGLEVEL 0
#endif
#if (DEBUGLEVEL>=1)
# include <assert.h> /* note: can still be disabled with NDEBUG */
# define XXH_ASSERT(c) assert(c)
#else
# define XXH_ASSERT(c) ((void)0)
#endif
/* note: use after variable declarations */
#define XXH_STATIC_ASSERT(c) { enum { XXH_sa = 1/(int)(!!(c)) }; }
/* *************************************
* Basic Types
***************************************/
#if !defined (__VMS) \
&& (defined (__cplusplus) \
|| (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) )
# include <stdint.h>
typedef uint8_t xxh_u8;
#else
typedef unsigned char xxh_u8;
#endif
typedef XXH32_hash_t xxh_u32;
/* *** Memory access *** */
#if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3))
/*
* Manual byteshift. Best for old compilers which don't inline memcpy.
* We actually directly use XXH_readLE32 and XXH_readBE32.
*/
#elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==2))
/*
* Force direct memory access. Only works on CPU which support unaligned memory
* access in hardware.
*/
static xxh_u32 XXH_read32(const void* memPtr) { return *(const xxh_u32*) memPtr; }
#elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==1))
/*
* __pack instructions are safer but compiler specific, hence potentially
* problematic for some compilers.
*
* Currently only defined for GCC and ICC.
*/
typedef union { xxh_u32 u32; } __attribute__((packed)) unalign;
static xxh_u32 XXH_read32(const void* ptr) { return ((const unalign*)ptr)->u32; }
#else
/*
* Portable and safe solution. Generally efficient.
* see: https://stackoverflow.com/a/32095106/646947
*/
static xxh_u32 XXH_read32(const void* memPtr)
{
xxh_u32 val;
memcpy(&val, memPtr, sizeof(val));
return val;
}
#endif /* XXH_FORCE_DIRECT_MEMORY_ACCESS */
/* *** Endianess *** */
typedef enum { XXH_bigEndian=0, XXH_littleEndian=1 } XXH_endianess;
/*!
* XXH_CPU_LITTLE_ENDIAN:
* Defined to 1 if the target is little endian, or 0 if it is big endian.
* It can be defined externally, for example on the compiler command line.
*
* If it is not defined, a runtime check (which is usually constant folded)
* is used instead.
*/
#ifndef XXH_CPU_LITTLE_ENDIAN
/*
* Try to detect endianness automatically, to avoid the nonstandard behavior
* in `XXH_isLittleEndian()`
*/
# if defined(_WIN32) /* Windows is always little endian */ \
|| defined(__LITTLE_ENDIAN__) \
|| (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__)
# define XXH_CPU_LITTLE_ENDIAN 1
# elif defined(__BIG_ENDIAN__) \
|| (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__)
# define XXH_CPU_LITTLE_ENDIAN 0
# else
static int XXH_isLittleEndian(void)
{
/*
* Nonstandard, but well-defined behavior in practice.
* Don't use static: it is detrimental to performance.
*/
const union { xxh_u32 u; xxh_u8 c[4]; } one = { 1 };
return one.c[0];
}
# define XXH_CPU_LITTLE_ENDIAN XXH_isLittleEndian()
# endif
#endif
/* ****************************************
* Compiler-specific Functions and Macros
******************************************/
#define XXH_GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__)
#ifndef __has_builtin
# define __has_builtin(x) 0
#endif
#if !defined(NO_CLANG_BUILTIN) && __has_builtin(__builtin_rotateleft32) \
&& __has_builtin(__builtin_rotateleft64)
# define XXH_rotl32 __builtin_rotateleft32
# define XXH_rotl64 __builtin_rotateleft64
/* Note: although _rotl exists for minGW (GCC under windows), performance seems poor */
#elif defined(_MSC_VER)
# define XXH_rotl32(x,r) _rotl(x,r)
# define XXH_rotl64(x,r) _rotl64(x,r)
#else
# define XXH_rotl32(x,r) (((x) << (r)) | ((x) >> (32 - (r))))
# define XXH_rotl64(x,r) (((x) << (r)) | ((x) >> (64 - (r))))
#endif
#if defined(_MSC_VER) /* Visual Studio */
# define XXH_swap32 _byteswap_ulong
#elif XXH_GCC_VERSION >= 403
# define XXH_swap32 __builtin_bswap32
#else
static xxh_u32 XXH_swap32 (xxh_u32 x)
{
return ((x << 24) & 0xff000000 ) |
((x << 8) & 0x00ff0000 ) |
((x >> 8) & 0x0000ff00 ) |
((x >> 24) & 0x000000ff );
}
#endif
/* ***************************
* Memory reads
*****************************/
typedef enum { XXH_aligned, XXH_unaligned } XXH_alignment;
/*
* XXH_FORCE_MEMORY_ACCESS==3 is an endian-independent byteshift load.
*
* This is ideal for older compilers which don't inline memcpy.
*/
#if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3))
XXH_FORCE_INLINE xxh_u32 XXH_readLE32(const void* memPtr)
{
const xxh_u8* bytePtr = (const xxh_u8 *)memPtr;
return bytePtr[0]
| ((xxh_u32)bytePtr[1] << 8)
| ((xxh_u32)bytePtr[2] << 16)
| ((xxh_u32)bytePtr[3] << 24);
}
XXH_FORCE_INLINE xxh_u32 XXH_readBE32(const void* memPtr)
{
const xxh_u8* bytePtr = (const xxh_u8 *)memPtr;
return bytePtr[3]
| ((xxh_u32)bytePtr[2] << 8)
| ((xxh_u32)bytePtr[1] << 16)
| ((xxh_u32)bytePtr[0] << 24);
}
#else
XXH_FORCE_INLINE xxh_u32 XXH_readLE32(const void* ptr)
{
return XXH_CPU_LITTLE_ENDIAN ? XXH_read32(ptr) : XXH_swap32(XXH_read32(ptr));
}
static xxh_u32 XXH_readBE32(const void* ptr)
{
return XXH_CPU_LITTLE_ENDIAN ? XXH_swap32(XXH_read32(ptr)) : XXH_read32(ptr);
}
#endif
XXH_FORCE_INLINE xxh_u32
XXH_readLE32_align(const void* ptr, XXH_alignment align)
{
if (align==XXH_unaligned) {
return XXH_readLE32(ptr);
} else {
return XXH_CPU_LITTLE_ENDIAN ? *(const xxh_u32*)ptr : XXH_swap32(*(const xxh_u32*)ptr);
}
}
/* *************************************
* Misc
***************************************/
XXH_PUBLIC_API unsigned XXH_versionNumber (void) { return XXH_VERSION_NUMBER; }
/* *******************************************************************
* 32-bit hash functions
*********************************************************************/
static const xxh_u32 PRIME32_1 = 0x9E3779B1U; /* 0b10011110001101110111100110110001 */
static const xxh_u32 PRIME32_2 = 0x85EBCA77U; /* 0b10000101111010111100101001110111 */
static const xxh_u32 PRIME32_3 = 0xC2B2AE3DU; /* 0b11000010101100101010111000111101 */
static const xxh_u32 PRIME32_4 = 0x27D4EB2FU; /* 0b00100111110101001110101100101111 */
static const xxh_u32 PRIME32_5 = 0x165667B1U; /* 0b00010110010101100110011110110001 */
static xxh_u32 XXH32_round(xxh_u32 acc, xxh_u32 input)
{
acc += input * PRIME32_2;
acc = XXH_rotl32(acc, 13);
acc *= PRIME32_1;
#if defined(__GNUC__) && defined(__SSE4_1__) && !defined(XXH_ENABLE_AUTOVECTORIZE)
/*
* UGLY HACK:
* This inline assembly hack forces acc into a normal register. This is the
* only thing that prevents GCC and Clang from autovectorizing the XXH32
* loop (pragmas and attributes don't work for some resason) without globally
* disabling SSE4.1.
*
* The reason we want to avoid vectorization is because despite working on
* 4 integers at a time, there are multiple factors slowing XXH32 down on
* SSE4:
* - There's a ridiculous amount of lag from pmulld (10 cycles of latency on
* newer chips!) making it slightly slower to multiply four integers at
* once compared to four integers independently. Even when pmulld was
* fastest, Sandy/Ivy Bridge, it is still not worth it to go into SSE
* just to multiply unless doing a long operation.
*
* - Four instructions are required to rotate,
* movqda tmp, v // not required with VEX encoding
* pslld tmp, 13 // tmp <<= 13
* psrld v, 19 // x >>= 19
* por v, tmp // x |= tmp
* compared to one for scalar:
* roll v, 13 // reliably fast across the board
* shldl v, v, 13 // Sandy Bridge and later prefer this for some reason
*
* - Instruction level parallelism is actually more beneficial here because
* the SIMD actually serializes this operation: While v1 is rotating, v2
* can load data, while v3 can multiply. SSE forces them to operate
* together.
*
* How this hack works:
* __asm__("" // Declare an assembly block but don't declare any instructions
* : // However, as an Input/Output Operand,
* "+r" // constrain a read/write operand (+) as a general purpose register (r).
* (acc) // and set acc as the operand
* );
*
* Because of the 'r', the compiler has promised that seed will be in a
* general purpose register and the '+' says that it will be 'read/write',
* so it has to assume it has changed. It is like volatile without all the
* loads and stores.
*
* Since the argument has to be in a normal register (not an SSE register),
* each time XXH32_round is called, it is impossible to vectorize.
*/
__asm__("" : "+r" (acc));
#endif
return acc;
}
/* mix all bits */
static xxh_u32 XXH32_avalanche(xxh_u32 h32)
{
h32 ^= h32 >> 15;
h32 *= PRIME32_2;
h32 ^= h32 >> 13;
h32 *= PRIME32_3;
h32 ^= h32 >> 16;
return(h32);
}
#define XXH_get32bits(p) XXH_readLE32_align(p, align)
static xxh_u32
XXH32_finalize(xxh_u32 h32, const xxh_u8* ptr, size_t len, XXH_alignment align)
{
#define PROCESS1 \
h32 += (*ptr++) * PRIME32_5; \
h32 = XXH_rotl32(h32, 11) * PRIME32_1 ;
#define PROCESS4 \
h32 += XXH_get32bits(ptr) * PRIME32_3; \
ptr+=4; \
h32 = XXH_rotl32(h32, 17) * PRIME32_4 ;
/* Compact rerolled version */
if (XXH_REROLL) {
len &= 15;
while (len >= 4) {
PROCESS4;
len -= 4;
}
while (len > 0) {
PROCESS1;
--len;
}
return XXH32_avalanche(h32);
} else {
switch(len&15) /* or switch(bEnd - p) */ {
case 12: PROCESS4;
/* fallthrough */
case 8: PROCESS4;
/* fallthrough */
case 4: PROCESS4;
return XXH32_avalanche(h32);
case 13: PROCESS4;
/* fallthrough */
case 9: PROCESS4;
/* fallthrough */
case 5: PROCESS4;
PROCESS1;
return XXH32_avalanche(h32);
case 14: PROCESS4;
/* fallthrough */
case 10: PROCESS4;
/* fallthrough */
case 6: PROCESS4;
PROCESS1;
PROCESS1;
return XXH32_avalanche(h32);
case 15: PROCESS4;
/* fallthrough */
case 11: PROCESS4;
/* fallthrough */
case 7: PROCESS4;
/* fallthrough */
case 3: PROCESS1;
/* fallthrough */
case 2: PROCESS1;
/* fallthrough */
case 1: PROCESS1;
/* fallthrough */
case 0: return XXH32_avalanche(h32);
}
XXH_ASSERT(0);
return h32; /* reaching this point is deemed impossible */
}
}
XXH_FORCE_INLINE xxh_u32
XXH32_endian_align(const xxh_u8* input, size_t len, xxh_u32 seed, XXH_alignment align)
{
const xxh_u8* bEnd = input + len;
xxh_u32 h32;
#if defined(XXH_ACCEPT_NULL_INPUT_POINTER) && (XXH_ACCEPT_NULL_INPUT_POINTER>=1)
if (input==NULL) {
len=0;
bEnd=input=(const xxh_u8*)(size_t)16;
}
#endif
if (len>=16) {
const xxh_u8* const limit = bEnd - 15;
xxh_u32 v1 = seed + PRIME32_1 + PRIME32_2;
xxh_u32 v2 = seed + PRIME32_2;
xxh_u32 v3 = seed + 0;
xxh_u32 v4 = seed - PRIME32_1;
do {
v1 = XXH32_round(v1, XXH_get32bits(input)); input += 4;
v2 = XXH32_round(v2, XXH_get32bits(input)); input += 4;
v3 = XXH32_round(v3, XXH_get32bits(input)); input += 4;
v4 = XXH32_round(v4, XXH_get32bits(input)); input += 4;
} while (input < limit);
h32 = XXH_rotl32(v1, 1) + XXH_rotl32(v2, 7)
+ XXH_rotl32(v3, 12) + XXH_rotl32(v4, 18);
} else {
h32 = seed + PRIME32_5;
}
h32 += (xxh_u32)len;
return XXH32_finalize(h32, input, len&15, align);
}
XXH_PUBLIC_API XXH32_hash_t XXH32 (const void* input, size_t len, XXH32_hash_t seed)
{
#if 0
/* Simple version, good for code maintenance, but unfortunately slow for small inputs */
XXH32_state_t state;
XXH32_reset(&state, seed);
XXH32_update(&state, (const xxh_u8*)input, len);
return XXH32_digest(&state);
#else
if (XXH_FORCE_ALIGN_CHECK) {
if ((((size_t)input) & 3) == 0) { /* Input is 4-bytes aligned, leverage the speed benefit */
return XXH32_endian_align((const xxh_u8*)input, len, seed, XXH_aligned);
} }
return XXH32_endian_align((const xxh_u8*)input, len, seed, XXH_unaligned);
#endif
}
/******* Hash streaming *******/
XXH_PUBLIC_API XXH32_state_t* XXH32_createState(void)
{
return (XXH32_state_t*)XXH_malloc(sizeof(XXH32_state_t));
}
XXH_PUBLIC_API XXH_errorcode XXH32_freeState(XXH32_state_t* statePtr)
{
XXH_free(statePtr);
return XXH_OK;
}
XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* dstState, const XXH32_state_t* srcState)
{
memcpy(dstState, srcState, sizeof(*dstState));
}
XXH_PUBLIC_API XXH_errorcode XXH32_reset(XXH32_state_t* statePtr, XXH32_hash_t seed)
{
XXH32_state_t state; /* using a local state to memcpy() in order to avoid strict-aliasing warnings */
memset(&state, 0, sizeof(state));
state.v1 = seed + PRIME32_1 + PRIME32_2;
state.v2 = seed + PRIME32_2;
state.v3 = seed + 0;
state.v4 = seed - PRIME32_1;
/* do not write into reserved, planned to be removed in a future version */
memcpy(statePtr, &state, sizeof(state) - sizeof(state.reserved));
return XXH_OK;
}
XXH_PUBLIC_API XXH_errorcode
XXH32_update(XXH32_state_t* state, const void* input, size_t len)
{
if (input==NULL)
#if defined(XXH_ACCEPT_NULL_INPUT_POINTER) && (XXH_ACCEPT_NULL_INPUT_POINTER>=1)
return XXH_OK;
#else
return XXH_ERROR;
#endif
{ const xxh_u8* p = (const xxh_u8*)input;
const xxh_u8* const bEnd = p + len;
state->total_len_32 += (XXH32_hash_t)len;
state->large_len |= (XXH32_hash_t)((len>=16) | (state->total_len_32>=16));
if (state->memsize + len < 16) { /* fill in tmp buffer */
XXH_memcpy((xxh_u8*)(state->mem32) + state->memsize, input, len);
state->memsize += (XXH32_hash_t)len;
return XXH_OK;
}
if (state->memsize) { /* some data left from previous update */
XXH_memcpy((xxh_u8*)(state->mem32) + state->memsize, input, 16-state->memsize);
{ const xxh_u32* p32 = state->mem32;
state->v1 = XXH32_round(state->v1, XXH_readLE32(p32)); p32++;
state->v2 = XXH32_round(state->v2, XXH_readLE32(p32)); p32++;
state->v3 = XXH32_round(state->v3, XXH_readLE32(p32)); p32++;
state->v4 = XXH32_round(state->v4, XXH_readLE32(p32));
}
p += 16-state->memsize;
state->memsize = 0;
}
if (p <= bEnd-16) {
const xxh_u8* const limit = bEnd - 16;
xxh_u32 v1 = state->v1;
xxh_u32 v2 = state->v2;
xxh_u32 v3 = state->v3;
xxh_u32 v4 = state->v4;
do {
v1 = XXH32_round(v1, XXH_readLE32(p)); p+=4;
v2 = XXH32_round(v2, XXH_readLE32(p)); p+=4;
v3 = XXH32_round(v3, XXH_readLE32(p)); p+=4;
v4 = XXH32_round(v4, XXH_readLE32(p)); p+=4;
} while (p<=limit);
state->v1 = v1;
state->v2 = v2;
state->v3 = v3;
state->v4 = v4;
}
if (p < bEnd) {
XXH_memcpy(state->mem32, p, (size_t)(bEnd-p));
state->memsize = (unsigned)(bEnd-p);
}
}
return XXH_OK;
}
XXH_PUBLIC_API XXH32_hash_t XXH32_digest (const XXH32_state_t* state)
{
xxh_u32 h32;
if (state->large_len) {
h32 = XXH_rotl32(state->v1, 1)
+ XXH_rotl32(state->v2, 7)
+ XXH_rotl32(state->v3, 12)
+ XXH_rotl32(state->v4, 18);
} else {
h32 = state->v3 /* == seed */ + PRIME32_5;
}
h32 += state->total_len_32;
return XXH32_finalize(h32, (const xxh_u8*)state->mem32, state->memsize, XXH_aligned);
}
/******* Canonical representation *******/
/*
* The default return values from XXH functions are unsigned 32 and 64 bit
* integers.
*
* The canonical representation uses big endian convention, the same convention
* as human-readable numbers (large digits first).
*
* This way, hash values can be written into a file or buffer, remaining
* comparable across different systems.
*
* The following functions allow transformation of hash values to and from their
* canonical format.
*/
XXH_PUBLIC_API void XXH32_canonicalFromHash(XXH32_canonical_t* dst, XXH32_hash_t hash)
{
XXH_STATIC_ASSERT(sizeof(XXH32_canonical_t) == sizeof(XXH32_hash_t));
if (XXH_CPU_LITTLE_ENDIAN) hash = XXH_swap32(hash);
memcpy(dst, &hash, sizeof(*dst));
}
XXH_PUBLIC_API XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src)
{
return XXH_readBE32(src);
}
#ifndef XXH_NO_LONG_LONG
/* *******************************************************************
* 64-bit hash functions
*********************************************************************/
/******* Memory access *******/
typedef XXH64_hash_t xxh_u64;
/*!
* XXH_REROLL_XXH64:
* Whether to reroll the XXH64_finalize() loop.
*
* Just like XXH32, we can unroll the XXH64_finalize() loop. This can be a
* performance gain on 64-bit hosts, as only one jump is required.
*
* However, on 32-bit hosts, because arithmetic needs to be done with two 32-bit
* registers, and 64-bit arithmetic needs to be simulated, it isn't beneficial
* to unroll. The code becomes ridiculously large (the largest function in the
* binary on i386!), and rerolling it saves anywhere from 3kB to 20kB. It is
* also slightly faster because it fits into cache better and is more likely
* to be inlined by the compiler.
*
* If XXH_REROLL is defined, this is ignored and the loop is always rerolled.
*/
#ifndef XXH_REROLL_XXH64
# if (defined(__ILP32__) || defined(_ILP32)) /* ILP32 is often defined on 32-bit GCC family */ \
|| !(defined(__x86_64__) || defined(_M_X64) || defined(_M_AMD64) /* x86-64 */ \
|| defined(_M_ARM64) || defined(__aarch64__) || defined(__arm64__) /* aarch64 */ \
|| defined(__PPC64__) || defined(__PPC64LE__) || defined(__ppc64__) || defined(__powerpc64__) /* ppc64 */ \
|| defined(__mips64__) || defined(__mips64)) /* mips64 */ \
|| (!defined(SIZE_MAX) || SIZE_MAX < ULLONG_MAX) /* check limits */
# define XXH_REROLL_XXH64 1
# else
# define XXH_REROLL_XXH64 0
# endif
#endif /* !defined(XXH_REROLL_XXH64) */
#if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3))
/*
* Manual byteshift. Best for old compilers which don't inline memcpy.
* We actually directly use XXH_readLE64 and XXH_readBE64.
*/
#elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==2))
/* Force direct memory access. Only works on CPU which support unaligned memory access in hardware */
static xxh_u64 XXH_read64(const void* memPtr) { return *(const xxh_u64*) memPtr; }
#elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==1))
/*
* __pack instructions are safer, but compiler specific, hence potentially
* problematic for some compilers.
*
* Currently only defined for GCC and ICC.
*/
typedef union { xxh_u32 u32; xxh_u64 u64; } __attribute__((packed)) unalign64;
static xxh_u64 XXH_read64(const void* ptr) { return ((const unalign64*)ptr)->u64; }
#else
/*
* Portable and safe solution. Generally efficient.
* see: https://stackoverflow.com/a/32095106/646947
*/
static xxh_u64 XXH_read64(const void* memPtr)
{
xxh_u64 val;
memcpy(&val, memPtr, sizeof(val));
return val;
}
#endif /* XXH_FORCE_DIRECT_MEMORY_ACCESS */
#if defined(_MSC_VER) /* Visual Studio */
# define XXH_swap64 _byteswap_uint64
#elif XXH_GCC_VERSION >= 403
# define XXH_swap64 __builtin_bswap64
#else
static xxh_u64 XXH_swap64 (xxh_u64 x)
{
return ((x << 56) & 0xff00000000000000ULL) |
((x << 40) & 0x00ff000000000000ULL) |
((x << 24) & 0x0000ff0000000000ULL) |
((x << 8) & 0x000000ff00000000ULL) |
((x >> 8) & 0x00000000ff000000ULL) |
((x >> 24) & 0x0000000000ff0000ULL) |
((x >> 40) & 0x000000000000ff00ULL) |
((x >> 56) & 0x00000000000000ffULL);
}
#endif
/* XXH_FORCE_MEMORY_ACCESS==3 is an endian-independent byteshift load. */
#if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3))
XXH_FORCE_INLINE xxh_u64 XXH_readLE64(const void* memPtr)
{
const xxh_u8* bytePtr = (const xxh_u8 *)memPtr;
return bytePtr[0]
| ((xxh_u64)bytePtr[1] << 8)
| ((xxh_u64)bytePtr[2] << 16)
| ((xxh_u64)bytePtr[3] << 24)
| ((xxh_u64)bytePtr[4] << 32)
| ((xxh_u64)bytePtr[5] << 40)
| ((xxh_u64)bytePtr[6] << 48)
| ((xxh_u64)bytePtr[7] << 56);
}
XXH_FORCE_INLINE xxh_u64 XXH_readBE64(const void* memPtr)
{
const xxh_u8* bytePtr = (const xxh_u8 *)memPtr;
return bytePtr[7]
| ((xxh_u64)bytePtr[6] << 8)
| ((xxh_u64)bytePtr[5] << 16)
| ((xxh_u64)bytePtr[4] << 24)
| ((xxh_u64)bytePtr[3] << 32)
| ((xxh_u64)bytePtr[2] << 40)
| ((xxh_u64)bytePtr[1] << 48)
| ((xxh_u64)bytePtr[0] << 56);
}
#else
XXH_FORCE_INLINE xxh_u64 XXH_readLE64(const void* ptr)
{
return XXH_CPU_LITTLE_ENDIAN ? XXH_read64(ptr) : XXH_swap64(XXH_read64(ptr));
}
static xxh_u64 XXH_readBE64(const void* ptr)
{
return XXH_CPU_LITTLE_ENDIAN ? XXH_swap64(XXH_read64(ptr)) : XXH_read64(ptr);
}
#endif
XXH_FORCE_INLINE xxh_u64
XXH_readLE64_align(const void* ptr, XXH_alignment align)
{
if (align==XXH_unaligned)
return XXH_readLE64(ptr);
else
return XXH_CPU_LITTLE_ENDIAN ? *(const xxh_u64*)ptr : XXH_swap64(*(const xxh_u64*)ptr);
}
/******* xxh64 *******/
static const xxh_u64 PRIME64_1 = 0x9E3779B185EBCA87ULL; /* 0b1001111000110111011110011011000110000101111010111100101010000111 */
static const xxh_u64 PRIME64_2 = 0xC2B2AE3D27D4EB4FULL; /* 0b1100001010110010101011100011110100100111110101001110101101001111 */
static const xxh_u64 PRIME64_3 = 0x165667B19E3779F9ULL; /* 0b0001011001010110011001111011000110011110001101110111100111111001 */
static const xxh_u64 PRIME64_4 = 0x85EBCA77C2B2AE63ULL; /* 0b1000010111101011110010100111011111000010101100101010111001100011 */
static const xxh_u64 PRIME64_5 = 0x27D4EB2F165667C5ULL; /* 0b0010011111010100111010110010111100010110010101100110011111000101 */
static xxh_u64 XXH64_round(xxh_u64 acc, xxh_u64 input)
{
acc += input * PRIME64_2;
acc = XXH_rotl64(acc, 31);
acc *= PRIME64_1;
return acc;
}
static xxh_u64 XXH64_mergeRound(xxh_u64 acc, xxh_u64 val)
{
val = XXH64_round(0, val);
acc ^= val;
acc = acc * PRIME64_1 + PRIME64_4;
return acc;
}
static xxh_u64 XXH64_avalanche(xxh_u64 h64)
{
h64 ^= h64 >> 33;
h64 *= PRIME64_2;
h64 ^= h64 >> 29;
h64 *= PRIME64_3;
h64 ^= h64 >> 32;
return h64;
}
#define XXH_get64bits(p) XXH_readLE64_align(p, align)
static xxh_u64
XXH64_finalize(xxh_u64 h64, const xxh_u8* ptr, size_t len, XXH_alignment align)
{
#define PROCESS1_64 \
h64 ^= (*ptr++) * PRIME64_5; \
h64 = XXH_rotl64(h64, 11) * PRIME64_1;
#define PROCESS4_64 \
h64 ^= (xxh_u64)(XXH_get32bits(ptr)) * PRIME64_1; \
ptr+=4; \
h64 = XXH_rotl64(h64, 23) * PRIME64_2 + PRIME64_3;
#define PROCESS8_64 { \
xxh_u64 const k1 = XXH64_round(0, XXH_get64bits(ptr)); \
ptr+=8; \
h64 ^= k1; \
h64 = XXH_rotl64(h64,27) * PRIME64_1 + PRIME64_4; \
}
/* Rerolled version for 32-bit targets is faster and much smaller. */
if (XXH_REROLL || XXH_REROLL_XXH64) {
len &= 31;
while (len >= 8) {
PROCESS8_64;
len -= 8;
}
if (len >= 4) {
PROCESS4_64;
len -= 4;
}
while (len > 0) {
PROCESS1_64;
--len;
}
return XXH64_avalanche(h64);
} else {
switch(len & 31) {
case 24: PROCESS8_64;
/* fallthrough */
case 16: PROCESS8_64;
/* fallthrough */
case 8: PROCESS8_64;
return XXH64_avalanche(h64);
case 28: PROCESS8_64;
/* fallthrough */
case 20: PROCESS8_64;
/* fallthrough */
case 12: PROCESS8_64;
/* fallthrough */
case 4: PROCESS4_64;
return XXH64_avalanche(h64);
case 25: PROCESS8_64;
/* fallthrough */
case 17: PROCESS8_64;
/* fallthrough */
case 9: PROCESS8_64;
PROCESS1_64;
return XXH64_avalanche(h64);
case 29: PROCESS8_64;
/* fallthrough */
case 21: PROCESS8_64;
/* fallthrough */
case 13: PROCESS8_64;
/* fallthrough */
case 5: PROCESS4_64;
PROCESS1_64;
return XXH64_avalanche(h64);
case 26: PROCESS8_64;
/* fallthrough */
case 18: PROCESS8_64;
/* fallthrough */
case 10: PROCESS8_64;
PROCESS1_64;
PROCESS1_64;
return XXH64_avalanche(h64);
case 30: PROCESS8_64;
/* fallthrough */
case 22: PROCESS8_64;
/* fallthrough */
case 14: PROCESS8_64;
/* fallthrough */
case 6: PROCESS4_64;
PROCESS1_64;
PROCESS1_64;
return XXH64_avalanche(h64);
case 27: PROCESS8_64;
/* fallthrough */
case 19: PROCESS8_64;
/* fallthrough */
case 11: PROCESS8_64;
PROCESS1_64;
PROCESS1_64;
PROCESS1_64;
return XXH64_avalanche(h64);
case 31: PROCESS8_64;
/* fallthrough */
case 23: PROCESS8_64;
/* fallthrough */
case 15: PROCESS8_64;
/* fallthrough */
case 7: PROCESS4_64;
/* fallthrough */
case 3: PROCESS1_64;
/* fallthrough */
case 2: PROCESS1_64;
/* fallthrough */
case 1: PROCESS1_64;
/* fallthrough */
case 0: return XXH64_avalanche(h64);
}
}
/* impossible to reach */
XXH_ASSERT(0);
return 0; /* unreachable, but some compilers complain without it */
}
XXH_FORCE_INLINE xxh_u64
XXH64_endian_align(const xxh_u8* input, size_t len, xxh_u64 seed, XXH_alignment align)
{
const xxh_u8* bEnd = input + len;
xxh_u64 h64;
#if defined(XXH_ACCEPT_NULL_INPUT_POINTER) && (XXH_ACCEPT_NULL_INPUT_POINTER>=1)
if (input==NULL) {
len=0;
bEnd=input=(const xxh_u8*)(size_t)32;
}
#endif
if (len>=32) {
const xxh_u8* const limit = bEnd - 32;
xxh_u64 v1 = seed + PRIME64_1 + PRIME64_2;
xxh_u64 v2 = seed + PRIME64_2;
xxh_u64 v3 = seed + 0;
xxh_u64 v4 = seed - PRIME64_1;
do {
v1 = XXH64_round(v1, XXH_get64bits(input)); input+=8;
v2 = XXH64_round(v2, XXH_get64bits(input)); input+=8;
v3 = XXH64_round(v3, XXH_get64bits(input)); input+=8;
v4 = XXH64_round(v4, XXH_get64bits(input)); input+=8;
} while (input<=limit);
h64 = XXH_rotl64(v1, 1) + XXH_rotl64(v2, 7) + XXH_rotl64(v3, 12) + XXH_rotl64(v4, 18);
h64 = XXH64_mergeRound(h64, v1);
h64 = XXH64_mergeRound(h64, v2);
h64 = XXH64_mergeRound(h64, v3);
h64 = XXH64_mergeRound(h64, v4);
} else {
h64 = seed + PRIME64_5;
}
h64 += (xxh_u64) len;
return XXH64_finalize(h64, input, len, align);
}
XXH_PUBLIC_API XXH64_hash_t XXH64 (const void* input, size_t len, XXH64_hash_t seed)
{
#if 0
/* Simple version, good for code maintenance, but unfortunately slow for small inputs */
XXH64_state_t state;
XXH64_reset(&state, seed);
XXH64_update(&state, (const xxh_u8*)input, len);
return XXH64_digest(&state);
#else
if (XXH_FORCE_ALIGN_CHECK) {
if ((((size_t)input) & 7)==0) { /* Input is aligned, let's leverage the speed advantage */
return XXH64_endian_align((const xxh_u8*)input, len, seed, XXH_aligned);
} }
return XXH64_endian_align((const xxh_u8*)input, len, seed, XXH_unaligned);
#endif
}
/******* Hash Streaming *******/
XXH_PUBLIC_API XXH64_state_t* XXH64_createState(void)
{
return (XXH64_state_t*)XXH_malloc(sizeof(XXH64_state_t));
}
XXH_PUBLIC_API XXH_errorcode XXH64_freeState(XXH64_state_t* statePtr)
{
XXH_free(statePtr);
return XXH_OK;
}
XXH_PUBLIC_API void XXH64_copyState(XXH64_state_t* dstState, const XXH64_state_t* srcState)
{
memcpy(dstState, srcState, sizeof(*dstState));
}
XXH_PUBLIC_API XXH_errorcode XXH64_reset(XXH64_state_t* statePtr, XXH64_hash_t seed)
{
XXH64_state_t state; /* use a local state to memcpy() in order to avoid strict-aliasing warnings */
memset(&state, 0, sizeof(state));
state.v1 = seed + PRIME64_1 + PRIME64_2;
state.v2 = seed + PRIME64_2;
state.v3 = seed + 0;
state.v4 = seed - PRIME64_1;
/* do not write into reserved64, might be removed in a future version */
memcpy(statePtr, &state, sizeof(state) - sizeof(state.reserved64));
return XXH_OK;
}
XXH_PUBLIC_API XXH_errorcode
XXH64_update (XXH64_state_t* state, const void* input, size_t len)
{
if (input==NULL)
#if defined(XXH_ACCEPT_NULL_INPUT_POINTER) && (XXH_ACCEPT_NULL_INPUT_POINTER>=1)
return XXH_OK;
#else
return XXH_ERROR;
#endif
{ const xxh_u8* p = (const xxh_u8*)input;
const xxh_u8* const bEnd = p + len;
state->total_len += len;
if (state->memsize + len < 32) { /* fill in tmp buffer */
XXH_memcpy(((xxh_u8*)state->mem64) + state->memsize, input, len);
state->memsize += (xxh_u32)len;
return XXH_OK;
}
if (state->memsize) { /* tmp buffer is full */
XXH_memcpy(((xxh_u8*)state->mem64) + state->memsize, input, 32-state->memsize);
state->v1 = XXH64_round(state->v1, XXH_readLE64(state->mem64+0));
state->v2 = XXH64_round(state->v2, XXH_readLE64(state->mem64+1));
state->v3 = XXH64_round(state->v3, XXH_readLE64(state->mem64+2));
state->v4 = XXH64_round(state->v4, XXH_readLE64(state->mem64+3));
p += 32-state->memsize;
state->memsize = 0;
}
if (p+32 <= bEnd) {
const xxh_u8* const limit = bEnd - 32;
xxh_u64 v1 = state->v1;
xxh_u64 v2 = state->v2;
xxh_u64 v3 = state->v3;
xxh_u64 v4 = state->v4;
do {
v1 = XXH64_round(v1, XXH_readLE64(p)); p+=8;
v2 = XXH64_round(v2, XXH_readLE64(p)); p+=8;
v3 = XXH64_round(v3, XXH_readLE64(p)); p+=8;
v4 = XXH64_round(v4, XXH_readLE64(p)); p+=8;
} while (p<=limit);
state->v1 = v1;
state->v2 = v2;
state->v3 = v3;
state->v4 = v4;
}
if (p < bEnd) {
XXH_memcpy(state->mem64, p, (size_t)(bEnd-p));
state->memsize = (unsigned)(bEnd-p);
}
}
return XXH_OK;
}
XXH_PUBLIC_API XXH64_hash_t XXH64_digest (const XXH64_state_t* state)
{
xxh_u64 h64;
if (state->total_len >= 32) {
xxh_u64 const v1 = state->v1;
xxh_u64 const v2 = state->v2;
xxh_u64 const v3 = state->v3;
xxh_u64 const v4 = state->v4;
h64 = XXH_rotl64(v1, 1) + XXH_rotl64(v2, 7) + XXH_rotl64(v3, 12) + XXH_rotl64(v4, 18);
h64 = XXH64_mergeRound(h64, v1);
h64 = XXH64_mergeRound(h64, v2);
h64 = XXH64_mergeRound(h64, v3);
h64 = XXH64_mergeRound(h64, v4);
} else {
h64 = state->v3 /*seed*/ + PRIME64_5;
}
h64 += (xxh_u64) state->total_len;
return XXH64_finalize(h64, (const xxh_u8*)state->mem64, (size_t)state->total_len, XXH_aligned);
}
/******* Canonical representation *******/
XXH_PUBLIC_API void XXH64_canonicalFromHash(XXH64_canonical_t* dst, XXH64_hash_t hash)
{
XXH_STATIC_ASSERT(sizeof(XXH64_canonical_t) == sizeof(XXH64_hash_t));
if (XXH_CPU_LITTLE_ENDIAN) hash = XXH_swap64(hash);
memcpy(dst, &hash, sizeof(*dst));
}
XXH_PUBLIC_API XXH64_hash_t XXH64_hashFromCanonical(const XXH64_canonical_t* src)
{
return XXH_readBE64(src);
}
/* *********************************************************************
* XXH3
* New generation hash designed for speed on small keys and vectorization
************************************************************************ */
#include "xxh3.h"
#endif /* XXH_NO_LONG_LONG */
#endif /* XXH_IMPLEMENTATION */
#if defined (__cplusplus)
}
#endif
/*
* Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
* LICENSE file in the root directory of this source tree) and the GPLv2 (found
* in the COPYING file in the root directory of this source tree).
* You may select, at your option, one of the above-listed licenses.
*/
#ifndef MEM_H_MODULE
#define MEM_H_MODULE
#if defined (__cplusplus)
extern "C" {
#endif
/*-****************************************
* Dependencies
******************************************/
#include <stddef.h> /* size_t, ptrdiff_t */
#include <string.h> /* memcpy */
/*-****************************************
* Compiler specifics
******************************************/
#if defined(_MSC_VER) /* Visual Studio */
# include <stdlib.h> /* _byteswap_ulong */
# include <intrin.h> /* _byteswap_* */
#endif
#if defined(__GNUC__)
# define MEM_STATIC static __inline __attribute__((unused))
#elif defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */)
# define MEM_STATIC static inline
#elif defined(_MSC_VER)
# define MEM_STATIC static __inline
#else
# define MEM_STATIC static /* this version may generate warnings for unused static functions; disable the relevant warning */
#endif
#ifndef __has_builtin
# define __has_builtin(x) 0 /* compat. with non-clang compilers */
#endif
/* code only tested on 32 and 64 bits systems */
#define MEM_STATIC_ASSERT(c) { enum { MEM_static_assert = 1/(int)(!!(c)) }; }
MEM_STATIC void MEM_check(void) { MEM_STATIC_ASSERT((sizeof(size_t)==4) || (sizeof(size_t)==8)); }
/*-**************************************************************
* Basic Types
*****************************************************************/
#if !defined (__VMS) && (defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) )
# include <stdint.h>
typedef uint8_t BYTE;
typedef uint16_t U16;
typedef int16_t S16;
typedef uint32_t U32;
typedef int32_t S32;
typedef uint64_t U64;
typedef int64_t S64;
#else
# include <limits.h>
#if CHAR_BIT != 8
# error "this implementation requires char to be exactly 8-bit type"
#endif
typedef unsigned char BYTE;
#if USHRT_MAX != 65535
# error "this implementation requires short to be exactly 16-bit type"
#endif
typedef unsigned short U16;
typedef signed short S16;
#if UINT_MAX != 4294967295
# error "this implementation requires int to be exactly 32-bit type"
#endif
typedef unsigned int U32;
typedef signed int S32;
/* note : there are no limits defined for long long type in C90.
* limits exist in C99, however, in such case, <stdint.h> is preferred */
typedef unsigned long long U64;
typedef signed long long S64;
#endif
/*-**************************************************************
* Memory I/O
*****************************************************************/
/* MEM_FORCE_MEMORY_ACCESS :
* By default, access to unaligned memory is controlled by `memcpy()`, which is safe and portable.
* Unfortunately, on some target/compiler combinations, the generated assembly is sub-optimal.
* The below switch allow to select different access method for improved performance.
* Method 0 (default) : use `memcpy()`. Safe and portable.
* Method 1 : `__packed` statement. It depends on compiler extension (i.e., not portable).
* This method is safe if your compiler supports it, and *generally* as fast or faster than `memcpy`.
* Method 2 : direct access. This method is portable but violate C standard.
* It can generate buggy code on targets depending on alignment.
* In some circumstances, it's the only known way to get the most performance (i.e. GCC + ARMv6)
* See http://fastcompression.blogspot.fr/2015/08/accessing-unaligned-memory.html for details.
* Prefer these methods in priority order (0 > 1 > 2)
*/
#ifndef MEM_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */
# if defined(__GNUC__) && ( 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 MEM_FORCE_MEMORY_ACCESS 2
# elif defined(__INTEL_COMPILER) || defined(__GNUC__)
# define MEM_FORCE_MEMORY_ACCESS 1
# endif
#endif
MEM_STATIC unsigned MEM_32bits(void) { return sizeof(size_t)==4; }
MEM_STATIC unsigned MEM_64bits(void) { return sizeof(size_t)==8; }
MEM_STATIC unsigned MEM_isLittleEndian(void)
{
const union { U32 u; BYTE c[4]; } one = { 1 }; /* don't use static : performance detrimental */
return one.c[0];
}
#if defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==2)
/* violates C standard, by lying on structure alignment.
Only use if no other choice to achieve best performance on target platform */
MEM_STATIC U16 MEM_read16(const void* memPtr) { return *(const U16*) memPtr; }
MEM_STATIC U32 MEM_read32(const void* memPtr) { return *(const U32*) memPtr; }
MEM_STATIC U64 MEM_read64(const void* memPtr) { return *(const U64*) memPtr; }
MEM_STATIC size_t MEM_readST(const void* memPtr) { return *(const size_t*) memPtr; }
MEM_STATIC void MEM_write16(void* memPtr, U16 value) { *(U16*)memPtr = value; }
MEM_STATIC void MEM_write32(void* memPtr, U32 value) { *(U32*)memPtr = value; }
MEM_STATIC void MEM_write64(void* memPtr, U64 value) { *(U64*)memPtr = value; }
#elif defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==1)
/* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */
/* currently only defined for gcc and icc */
#if defined(_MSC_VER) || (defined(__INTEL_COMPILER) && defined(WIN32))
__pragma( pack(push, 1) )
typedef struct { U16 v; } unalign16;
typedef struct { U32 v; } unalign32;
typedef struct { U64 v; } unalign64;
typedef struct { size_t v; } unalignArch;
__pragma( pack(pop) )
#else
typedef struct { U16 v; } __attribute__((packed)) unalign16;
typedef struct { U32 v; } __attribute__((packed)) unalign32;
typedef struct { U64 v; } __attribute__((packed)) unalign64;
typedef struct { size_t v; } __attribute__((packed)) unalignArch;
#endif
MEM_STATIC U16 MEM_read16(const void* ptr) { return ((const unalign16*)ptr)->v; }
MEM_STATIC U32 MEM_read32(const void* ptr) { return ((const unalign32*)ptr)->v; }
MEM_STATIC U64 MEM_read64(const void* ptr) { return ((const unalign64*)ptr)->v; }
MEM_STATIC size_t MEM_readST(const void* ptr) { return ((const unalignArch*)ptr)->v; }
MEM_STATIC void MEM_write16(void* memPtr, U16 value) { ((unalign16*)memPtr)->v = value; }
MEM_STATIC void MEM_write32(void* memPtr, U32 value) { ((unalign32*)memPtr)->v = value; }
MEM_STATIC void MEM_write64(void* memPtr, U64 value) { ((unalign64*)memPtr)->v = value; }
#else
/* default method, safe and standard.
can sometimes prove slower */
MEM_STATIC U16 MEM_read16(const void* memPtr)
{
U16 val; memcpy(&val, memPtr, sizeof(val)); return val;
}
MEM_STATIC U32 MEM_read32(const void* memPtr)
{
U32 val; memcpy(&val, memPtr, sizeof(val)); return val;
}
MEM_STATIC U64 MEM_read64(const void* memPtr)
{
U64 val; memcpy(&val, memPtr, sizeof(val)); return val;
}
MEM_STATIC size_t MEM_readST(const void* memPtr)
{
size_t val; memcpy(&val, memPtr, sizeof(val)); return val;
}
MEM_STATIC void MEM_write16(void* memPtr, U16 value)
{
memcpy(memPtr, &value, sizeof(value));
}
MEM_STATIC void MEM_write32(void* memPtr, U32 value)
{
memcpy(memPtr, &value, sizeof(value));
}
MEM_STATIC void MEM_write64(void* memPtr, U64 value)
{
memcpy(memPtr, &value, sizeof(value));
}
#endif /* MEM_FORCE_MEMORY_ACCESS */
MEM_STATIC U32 MEM_swap32(U32 in)
{
#if defined(_MSC_VER) /* Visual Studio */
return _byteswap_ulong(in);
#elif (defined (__GNUC__) && (__GNUC__ * 100 + __GNUC_MINOR__ >= 403)) \
|| (defined(__clang__) && __has_builtin(__builtin_bswap32))
return __builtin_bswap32(in);
#else
return ((in << 24) & 0xff000000 ) |
((in << 8) & 0x00ff0000 ) |
((in >> 8) & 0x0000ff00 ) |
((in >> 24) & 0x000000ff );
#endif
}
MEM_STATIC U64 MEM_swap64(U64 in)
{
#if defined(_MSC_VER) /* Visual Studio */
return _byteswap_uint64(in);
#elif (defined (__GNUC__) && (__GNUC__ * 100 + __GNUC_MINOR__ >= 403)) \
|| (defined(__clang__) && __has_builtin(__builtin_bswap64))
return __builtin_bswap64(in);
#else
return ((in << 56) & 0xff00000000000000ULL) |
((in << 40) & 0x00ff000000000000ULL) |
((in << 24) & 0x0000ff0000000000ULL) |
((in << 8) & 0x000000ff00000000ULL) |
((in >> 8) & 0x00000000ff000000ULL) |
((in >> 24) & 0x0000000000ff0000ULL) |
((in >> 40) & 0x000000000000ff00ULL) |
((in >> 56) & 0x00000000000000ffULL);
#endif
}
MEM_STATIC size_t MEM_swapST(size_t in)
{
if (MEM_32bits())
return (size_t)MEM_swap32((U32)in);
else
return (size_t)MEM_swap64((U64)in);
}
/*=== Little endian r/w ===*/
MEM_STATIC U16 MEM_readLE16(const void* memPtr)
{
if (MEM_isLittleEndian())
return MEM_read16(memPtr);
else {
const BYTE* p = (const BYTE*)memPtr;
return (U16)(p[0] + (p[1]<<8));
}
}
MEM_STATIC void MEM_writeLE16(void* memPtr, U16 val)
{
if (MEM_isLittleEndian()) {
MEM_write16(memPtr, val);
} else {
BYTE* p = (BYTE*)memPtr;
p[0] = (BYTE)val;
p[1] = (BYTE)(val>>8);
}
}
MEM_STATIC U32 MEM_readLE24(const void* memPtr)
{
return MEM_readLE16(memPtr) + (((const BYTE*)memPtr)[2] << 16);
}
MEM_STATIC void MEM_writeLE24(void* memPtr, U32 val)
{
MEM_writeLE16(memPtr, (U16)val);
((BYTE*)memPtr)[2] = (BYTE)(val>>16);
}
MEM_STATIC U32 MEM_readLE32(const void* memPtr)
{
if (MEM_isLittleEndian())
return MEM_read32(memPtr);
else
return MEM_swap32(MEM_read32(memPtr));
}
MEM_STATIC void MEM_writeLE32(void* memPtr, U32 val32)
{
if (MEM_isLittleEndian())
MEM_write32(memPtr, val32);
else
MEM_write32(memPtr, MEM_swap32(val32));
}
MEM_STATIC U64 MEM_readLE64(const void* memPtr)
{
if (MEM_isLittleEndian())
return MEM_read64(memPtr);
else
return MEM_swap64(MEM_read64(memPtr));
}
MEM_STATIC void MEM_writeLE64(void* memPtr, U64 val64)
{
if (MEM_isLittleEndian())
MEM_write64(memPtr, val64);
else
MEM_write64(memPtr, MEM_swap64(val64));
}
MEM_STATIC size_t MEM_readLEST(const void* memPtr)
{
if (MEM_32bits())
return (size_t)MEM_readLE32(memPtr);
else
return (size_t)MEM_readLE64(memPtr);
}
MEM_STATIC void MEM_writeLEST(void* memPtr, size_t val)
{
if (MEM_32bits())
MEM_writeLE32(memPtr, (U32)val);
else
MEM_writeLE64(memPtr, (U64)val);
}
/*=== Big endian r/w ===*/
MEM_STATIC U32 MEM_readBE32(const void* memPtr)
{
if (MEM_isLittleEndian())
return MEM_swap32(MEM_read32(memPtr));
else
return MEM_read32(memPtr);
}
MEM_STATIC void MEM_writeBE32(void* memPtr, U32 val32)
{
if (MEM_isLittleEndian())
MEM_write32(memPtr, MEM_swap32(val32));
else
MEM_write32(memPtr, val32);
}
MEM_STATIC U64 MEM_readBE64(const void* memPtr)
{
if (MEM_isLittleEndian())
return MEM_swap64(MEM_read64(memPtr));
else
return MEM_read64(memPtr);
}
MEM_STATIC void MEM_writeBE64(void* memPtr, U64 val64)
{
if (MEM_isLittleEndian())
MEM_write64(memPtr, MEM_swap64(val64));
else
MEM_write64(memPtr, val64);
}
MEM_STATIC size_t MEM_readBEST(const void* memPtr)
{
if (MEM_32bits())
return (size_t)MEM_readBE32(memPtr);
else
return (size_t)MEM_readBE64(memPtr);
}
MEM_STATIC void MEM_writeBEST(void* memPtr, size_t val)
{
if (MEM_32bits())
MEM_writeBE32(memPtr, (U32)val);
else
MEM_writeBE64(memPtr, (U64)val);
}
#if defined (__cplusplus)
}
#endif
#endif /* MEM_H_MODULE */
/*
* Copyright (c) 2017-present, Facebook, Inc.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
* LICENSE file in the root directory of this source tree) and the GPLv2 (found
* in the COPYING file in the root directory of this source tree).
* You may select, at your option, one of the above-listed licenses.
*/
/* *********************************************************
* Turn on Large Files support (>4GB) for 32-bit Linux/Unix
***********************************************************/
#if !defined(__64BIT__) || defined(__MINGW32__) /* No point defining Large file for 64 bit but MinGW-w64 requires it */
# if !defined(_FILE_OFFSET_BITS)
# define _FILE_OFFSET_BITS 64 /* turn off_t into a 64-bit type for ftello, fseeko */
# endif
# if !defined(_LARGEFILE_SOURCE) /* obsolete macro, replaced with _FILE_OFFSET_BITS */
# define _LARGEFILE_SOURCE 1 /* Large File Support extension (LFS) - fseeko, ftello */
# endif
# if defined(_AIX) || defined(__hpux)
# define _LARGE_FILES /* Large file support on 32-bits AIX and HP-UX */
# endif
#endif
/* ************************************************************
* Avoid fseek()'s 2GiB barrier with MSVC, macOS, *BSD, MinGW
***************************************************************/
#if defined(_MSC_VER) && _MSC_VER >= 1400
# define LONG_SEEK _fseeki64
#elif !defined(__64BIT__) && (PLATFORM_POSIX_VERSION >= 200112L) /* No point defining Large file for 64 bit */
# define LONG_SEEK fseeko
#elif defined(__MINGW32__) && !defined(__STRICT_ANSI__) && !defined(__NO_MINGW_LFS) && defined(__MSVCRT__)
# define LONG_SEEK fseeko64
#elif defined(_WIN32) && !defined(__DJGPP__)
# include <windows.h>
static int LONG_SEEK(FILE* file, __int64 offset, int origin) {
LARGE_INTEGER off;
DWORD method;
off.QuadPart = offset;
if (origin == SEEK_END)
method = FILE_END;
else if (origin == SEEK_CUR)
method = FILE_CURRENT;
else
method = FILE_BEGIN;
if (SetFilePointerEx((HANDLE) _get_osfhandle(_fileno(file)), off, NULL, method))
return 0;
else
return -1;
}
#else
# define LONG_SEEK fseek
#endif
#include <stdlib.h> /* malloc, free */
#include <stdio.h> /* FILE* */
#include <limits.h> /* UNIT_MAX */
#include <assert.h>
#define XXH_STATIC_LINKING_ONLY
// #define XXH_NAMESPACE ZSTD_
#include "xxhash.h"
#define ZSTD_STATIC_LINKING_ONLY
#include "zstd.h"
#include "zstd_errors.h"
#include "mem.h"
#include "zstd_seekable.h"
#undef ERROR
#define ERROR(name) ((size_t)-ZSTD_error_##name)
#define CHECK_IO(f) { int const errcod = (f); if (errcod < 0) return ERROR(seekableIO); }
#undef MIN
#undef MAX
#define MIN(a, b) ((a) < (b) ? (a) : (b))
#define MAX(a, b) ((a) > (b) ? (a) : (b))
/* Special-case callbacks for FILE* and in-memory modes, so that we can treat
* them the same way as the advanced API */
static int ZSTD_seekable_read_FILE(void* opaque, void* buffer, size_t n)
{
size_t const result = fread(buffer, 1, n, (FILE*)opaque);
if (result != n) {
return -1;
}
return 0;
}
static int ZSTD_seekable_seek_FILE(void* opaque, long long offset, int origin)
{
int const ret = LONG_SEEK((FILE*)opaque, offset, origin);
if (ret) return ret;
return fflush((FILE*)opaque);
}
typedef struct {
const void *ptr;
size_t size;
size_t pos;
} buffWrapper_t;
static int ZSTD_seekable_read_buff(void* opaque, void* buffer, size_t n)
{
buffWrapper_t* buff = (buffWrapper_t*) opaque;
if (buff->pos + n > buff->size) return -1;
memcpy(buffer, (const BYTE*)buff->ptr + buff->pos, n);
buff->pos += n;
return 0;
}
static int ZSTD_seekable_seek_buff(void* opaque, long long offset, int origin)
{
buffWrapper_t* const buff = (buffWrapper_t*) opaque;
unsigned long long newOffset;
switch (origin) {
case SEEK_SET:
newOffset = offset;
break;
case SEEK_CUR:
newOffset = (unsigned long long)buff->pos + offset;
break;
case SEEK_END:
newOffset = (unsigned long long)buff->size + offset;
break;
default:
assert(0); /* not possible */
}
if (newOffset > buff->size) {
return -1;
}
buff->pos = newOffset;
return 0;
}
typedef struct {
U64 cOffset;
U64 dOffset;
U32 checksum;
} seekEntry_t;
typedef struct {
seekEntry_t* entries;
size_t tableLen;
int checksumFlag;
} seekTable_t;
#define SEEKABLE_BUFF_SIZE ZSTD_BLOCKSIZE_MAX
struct ZSTD_seekable_s {
ZSTD_DStream* dstream;
seekTable_t seekTable;
ZSTD_seekable_customFile src;
U64 decompressedOffset;
U32 curFrame;
BYTE inBuff[SEEKABLE_BUFF_SIZE]; /* need to do our own input buffering */
BYTE outBuff[SEEKABLE_BUFF_SIZE]; /* so we can efficiently decompress the
starts of chunks before we get to the
desired section */
ZSTD_inBuffer in; /* maintain continuity across ZSTD_seekable_decompress operations */
buffWrapper_t buffWrapper; /* for `src.opaque` in in-memory mode */
XXH64_state_t xxhState;
};
ZSTD_seekable* ZSTD_seekable_create(void)
{
ZSTD_seekable* zs = malloc(sizeof(ZSTD_seekable));
if (zs == NULL) return NULL;
/* also initializes stage to zsds_init */
memset(zs, 0, sizeof(*zs));
zs->dstream = ZSTD_createDStream();
if (zs->dstream == NULL) {
free(zs);
return NULL;
}
return zs;
}
size_t ZSTD_seekable_free(ZSTD_seekable* zs)
{
if (zs == NULL) return 0; /* support free on null */
ZSTD_freeDStream(zs->dstream);
free(zs->seekTable.entries);
free(zs);
return 0;
}
/** ZSTD_seekable_offsetToFrameIndex() :
* Performs a binary search to find the last frame with a decompressed offset
* <= pos
* @return : the frame's index */
unsigned ZSTD_seekable_offsetToFrameIndex(ZSTD_seekable* const zs, unsigned long long pos)
{
U32 lo = 0;
U32 hi = (U32)zs->seekTable.tableLen;
assert(zs->seekTable.tableLen <= UINT_MAX);
if (pos >= zs->seekTable.entries[zs->seekTable.tableLen].dOffset) {
return (U32)zs->seekTable.tableLen;
}
while (lo + 1 < hi) {
U32 const mid = lo + ((hi - lo) >> 1);
if (zs->seekTable.entries[mid].dOffset <= pos) {
lo = mid;
} else {
hi = mid;
}
}
return lo;
}
unsigned ZSTD_seekable_getNumFrames(ZSTD_seekable* const zs)
{
assert(zs->seekTable.tableLen <= UINT_MAX);
return (unsigned)zs->seekTable.tableLen;
}
unsigned long long ZSTD_seekable_getFrameCompressedOffset(ZSTD_seekable* const zs, unsigned frameIndex)
{
if (frameIndex >= zs->seekTable.tableLen) return ZSTD_SEEKABLE_FRAMEINDEX_TOOLARGE;
return zs->seekTable.entries[frameIndex].cOffset;
}
unsigned long long ZSTD_seekable_getFrameDecompressedOffset(ZSTD_seekable* const zs, unsigned frameIndex)
{
if (frameIndex >= zs->seekTable.tableLen) return ZSTD_SEEKABLE_FRAMEINDEX_TOOLARGE;
return zs->seekTable.entries[frameIndex].dOffset;
}
size_t ZSTD_seekable_getFrameCompressedSize(ZSTD_seekable* const zs, unsigned frameIndex)
{
if (frameIndex >= zs->seekTable.tableLen) return ERROR(frameIndex_tooLarge);
return zs->seekTable.entries[frameIndex + 1].cOffset -
zs->seekTable.entries[frameIndex].cOffset;
}
size_t ZSTD_seekable_getFrameDecompressedSize(ZSTD_seekable* const zs, unsigned frameIndex)
{
if (frameIndex > zs->seekTable.tableLen) return ERROR(frameIndex_tooLarge);
return zs->seekTable.entries[frameIndex + 1].dOffset -
zs->seekTable.entries[frameIndex].dOffset;
}
static size_t ZSTD_seekable_loadSeekTable(ZSTD_seekable* zs)
{
int checksumFlag;
ZSTD_seekable_customFile src = zs->src;
/* read the footer, fixed size */
CHECK_IO(src.seek(src.opaque, -(int)ZSTD_seekTableFooterSize, SEEK_END));
CHECK_IO(src.read(src.opaque, zs->inBuff, ZSTD_seekTableFooterSize));
if (MEM_readLE32(zs->inBuff + 5) != ZSTD_SEEKABLE_MAGICNUMBER) {
return ERROR(prefix_unknown);
}
{ BYTE const sfd = zs->inBuff[4];
checksumFlag = sfd >> 7;
/* check reserved bits */
if ((checksumFlag >> 2) & 0x1f) {
return ERROR(corruption_detected);
}
}
{ U32 const numFrames = MEM_readLE32(zs->inBuff);
U32 const sizePerEntry = 8 + (checksumFlag?4:0);
U32 const tableSize = sizePerEntry * numFrames;
U32 const frameSize = tableSize + ZSTD_seekTableFooterSize + ZSTD_SKIPPABLEHEADERSIZE;
U32 remaining = frameSize - ZSTD_seekTableFooterSize; /* don't need to re-read footer */
{
U32 const toRead = MIN(remaining, SEEKABLE_BUFF_SIZE);
CHECK_IO(src.seek(src.opaque, -(S64)frameSize, SEEK_END));
CHECK_IO(src.read(src.opaque, zs->inBuff, toRead));
remaining -= toRead;
}
if (MEM_readLE32(zs->inBuff) != (ZSTD_MAGIC_SKIPPABLE_START | 0xE)) {
return ERROR(prefix_unknown);
}
if (MEM_readLE32(zs->inBuff+4) + ZSTD_SKIPPABLEHEADERSIZE != frameSize) {
return ERROR(prefix_unknown);
}
{ /* Allocate an extra entry at the end so that we can do size
* computations on the last element without special case */
seekEntry_t* entries = (seekEntry_t*)malloc(sizeof(seekEntry_t) * (numFrames + 1));
U32 idx = 0;
U32 pos = 8;
U64 cOffset = 0;
U64 dOffset = 0;
if (!entries) {
free(entries);
return ERROR(memory_allocation);
}
/* compute cumulative positions */
for (; idx < numFrames; idx++) {
if (pos + sizePerEntry > SEEKABLE_BUFF_SIZE) {
U32 const offset = SEEKABLE_BUFF_SIZE - pos;
U32 const toRead = MIN(remaining, SEEKABLE_BUFF_SIZE - offset);
memmove(zs->inBuff, zs->inBuff + pos, offset); /* move any data we haven't read yet */
CHECK_IO(src.read(src.opaque, zs->inBuff+offset, toRead));
remaining -= toRead;
pos = 0;
}
entries[idx].cOffset = cOffset;
entries[idx].dOffset = dOffset;
cOffset += MEM_readLE32(zs->inBuff + pos);
pos += 4;
dOffset += MEM_readLE32(zs->inBuff + pos);
pos += 4;
if (checksumFlag) {
entries[idx].checksum = MEM_readLE32(zs->inBuff + pos);
pos += 4;
}
}
entries[numFrames].cOffset = cOffset;
entries[numFrames].dOffset = dOffset;
zs->seekTable.entries = entries;
zs->seekTable.tableLen = numFrames;
zs->seekTable.checksumFlag = checksumFlag;
return 0;
}
}
}
size_t ZSTD_seekable_initBuff(ZSTD_seekable* zs, const void* src, size_t srcSize)
{
zs->buffWrapper = (buffWrapper_t){src, srcSize, 0};
{ ZSTD_seekable_customFile srcFile = {&zs->buffWrapper,
&ZSTD_seekable_read_buff,
&ZSTD_seekable_seek_buff};
return ZSTD_seekable_initAdvanced(zs, srcFile); }
}
size_t ZSTD_seekable_initFile(ZSTD_seekable* zs, FILE* src)
{
ZSTD_seekable_customFile srcFile = {src, &ZSTD_seekable_read_FILE,
&ZSTD_seekable_seek_FILE};
return ZSTD_seekable_initAdvanced(zs, srcFile);
}
size_t ZSTD_seekable_initAdvanced(ZSTD_seekable* zs, ZSTD_seekable_customFile src)
{
zs->src = src;
{ const size_t seekTableInit = ZSTD_seekable_loadSeekTable(zs);
if (ZSTD_isError(seekTableInit)) return seekTableInit; }
zs->decompressedOffset = (U64)-1;
zs->curFrame = (U32)-1;
{ const size_t dstreamInit = ZSTD_initDStream(zs->dstream);
if (ZSTD_isError(dstreamInit)) return dstreamInit; }
return 0;
}
size_t ZSTD_seekable_decompress(ZSTD_seekable* zs, void* dst, size_t len, unsigned long long offset)
{
U32 targetFrame = ZSTD_seekable_offsetToFrameIndex(zs, offset);
do {
/* check if we can continue from a previous decompress job */
if (targetFrame != zs->curFrame || offset != zs->decompressedOffset) {
zs->decompressedOffset = zs->seekTable.entries[targetFrame].dOffset;
zs->curFrame = targetFrame;
CHECK_IO(zs->src.seek(zs->src.opaque,
zs->seekTable.entries[targetFrame].cOffset,
SEEK_SET));
zs->in = (ZSTD_inBuffer){zs->inBuff, 0, 0};
XXH64_reset(&zs->xxhState, 0);
ZSTD_resetDStream(zs->dstream);
}
while (zs->decompressedOffset < offset + len) {
size_t toRead;
ZSTD_outBuffer outTmp;
size_t prevOutPos;
if (zs->decompressedOffset < offset) {
/* dummy decompressions until we get to the target offset */
outTmp = (ZSTD_outBuffer){zs->outBuff, MIN(SEEKABLE_BUFF_SIZE, offset - zs->decompressedOffset), 0};
} else {
outTmp = (ZSTD_outBuffer){dst, len, zs->decompressedOffset - offset};
}
prevOutPos = outTmp.pos;
toRead = ZSTD_decompressStream(zs->dstream, &outTmp, &zs->in);
if (ZSTD_isError(toRead)) {
return toRead;
}
if (zs->seekTable.checksumFlag) {
XXH64_update(&zs->xxhState, (BYTE*)outTmp.dst + prevOutPos,
outTmp.pos - prevOutPos);
}
zs->decompressedOffset += outTmp.pos - prevOutPos;
if (toRead == 0) {
/* frame complete */
/* verify checksum */
if (zs->seekTable.checksumFlag &&
(XXH64_digest(&zs->xxhState) & 0xFFFFFFFFU) !=
zs->seekTable.entries[targetFrame].checksum) {
return ERROR(corruption_detected);
}
if (zs->decompressedOffset < offset + len) {
/* go back to the start and force a reset of the stream */
targetFrame = ZSTD_seekable_offsetToFrameIndex(zs, zs->decompressedOffset);
}
break;
}
/* read in more data if we're done with this buffer */
if (zs->in.pos == zs->in.size) {
toRead = MIN(toRead, SEEKABLE_BUFF_SIZE);
CHECK_IO(zs->src.read(zs->src.opaque, zs->inBuff, toRead));
zs->in.size = toRead;
zs->in.pos = 0;
}
}
} while (zs->decompressedOffset != offset + len);
return len;
}
size_t ZSTD_seekable_decompressFrame(ZSTD_seekable* zs, void* dst, size_t dstSize, unsigned frameIndex)
{
if (frameIndex >= zs->seekTable.tableLen) {
return ERROR(frameIndex_tooLarge);
}
{
size_t const decompressedSize =
zs->seekTable.entries[frameIndex + 1].dOffset -
zs->seekTable.entries[frameIndex].dOffset;
if (dstSize < decompressedSize) {
return ERROR(dstSize_tooSmall);
}
return ZSTD_seekable_decompress(
zs, dst, decompressedSize,
zs->seekTable.entries[frameIndex].dOffset);
}
}
/*
* Copyright (c) 2017-present, Facebook, Inc.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
* LICENSE file in the root directory of this source tree) and the GPLv2 (found
* in the COPYING file in the root directory of this source tree).
*/
#include <stdlib.h> /* malloc, free */
#include <limits.h> /* UINT_MAX */
#include <assert.h>
#define XXH_STATIC_LINKING_ONLY
// #define XXH_NAMESPACE ZSTD_
#include "xxhash.h"
#define ZSTD_STATIC_LINKING_ONLY
#include "zstd.h"
#include "zstd_errors.h"
#include "mem.h"
#include "zstd_seekable.h"
#define CHECK_Z(f) { size_t const ret = (f); if (ret != 0) return ret; }
#undef ERROR
#define ERROR(name) ((size_t)-ZSTD_error_##name)
#undef MIN
#undef MAX
#define MIN(a, b) ((a) < (b) ? (a) : (b))
#define MAX(a, b) ((a) > (b) ? (a) : (b))
typedef struct {
U32 cSize;
U32 dSize;
U32 checksum;
} framelogEntry_t;
struct ZSTD_frameLog_s {
framelogEntry_t* entries;
U32 size;
U32 capacity;
int checksumFlag;
/* for use when streaming out the seek table */
U32 seekTablePos;
U32 seekTableIndex;
} framelog_t;
struct ZSTD_seekable_CStream_s {
ZSTD_CStream* cstream;
ZSTD_frameLog framelog;
U32 frameCSize;
U32 frameDSize;
XXH64_state_t xxhState;
U32 maxFrameSize;
int writingSeekTable;
};
size_t ZSTD_seekable_frameLog_allocVec(ZSTD_frameLog* fl)
{
/* allocate some initial space */
size_t const FRAMELOG_STARTING_CAPACITY = 16;
fl->entries = (framelogEntry_t*)malloc(
sizeof(framelogEntry_t) * FRAMELOG_STARTING_CAPACITY);
if (fl->entries == NULL) return ERROR(memory_allocation);
fl->capacity = FRAMELOG_STARTING_CAPACITY;
return 0;
}
size_t ZSTD_seekable_frameLog_freeVec(ZSTD_frameLog* fl)
{
if (fl != NULL) free(fl->entries);
return 0;
}
ZSTD_frameLog* ZSTD_seekable_createFrameLog(int checksumFlag)
{
ZSTD_frameLog* fl = malloc(sizeof(ZSTD_frameLog));
if (fl == NULL) return NULL;
if (ZSTD_isError(ZSTD_seekable_frameLog_allocVec(fl))) {
free(fl);
return NULL;
}
fl->checksumFlag = checksumFlag;
fl->seekTablePos = 0;
fl->seekTableIndex = 0;
fl->size = 0;
return fl;
}
size_t ZSTD_seekable_freeFrameLog(ZSTD_frameLog* fl)
{
ZSTD_seekable_frameLog_freeVec(fl);
free(fl);
return 0;
}
ZSTD_seekable_CStream* ZSTD_seekable_createCStream()
{
ZSTD_seekable_CStream* zcs = malloc(sizeof(ZSTD_seekable_CStream));
if (zcs == NULL) return NULL;
memset(zcs, 0, sizeof(*zcs));
zcs->cstream = ZSTD_createCStream();
if (zcs->cstream == NULL) goto failed1;
if (ZSTD_isError(ZSTD_seekable_frameLog_allocVec(&zcs->framelog))) goto failed2;
return zcs;
failed2:
ZSTD_freeCStream(zcs->cstream);
failed1:
free(zcs);
return NULL;
}
size_t ZSTD_seekable_freeCStream(ZSTD_seekable_CStream* zcs)
{
if (zcs == NULL) return 0; /* support free on null */
ZSTD_freeCStream(zcs->cstream);
ZSTD_seekable_frameLog_freeVec(&zcs->framelog);
free(zcs);
return 0;
}
size_t ZSTD_seekable_initCStream(ZSTD_seekable_CStream* zcs,
int compressionLevel,
int checksumFlag,
unsigned maxFrameSize)
{
zcs->framelog.size = 0;
zcs->frameCSize = 0;
zcs->frameDSize = 0;
/* make sure maxFrameSize has a reasonable value */
if (maxFrameSize > ZSTD_SEEKABLE_MAX_FRAME_DECOMPRESSED_SIZE) {
return ERROR(frameParameter_unsupported);
}
zcs->maxFrameSize = maxFrameSize
? maxFrameSize
: ZSTD_SEEKABLE_MAX_FRAME_DECOMPRESSED_SIZE;
zcs->framelog.checksumFlag = checksumFlag;
if (zcs->framelog.checksumFlag) {
XXH64_reset(&zcs->xxhState, 0);
}
zcs->framelog.seekTablePos = 0;
zcs->framelog.seekTableIndex = 0;
zcs->writingSeekTable = 0;
return ZSTD_initCStream(zcs->cstream, compressionLevel);
}
size_t ZSTD_seekable_logFrame(ZSTD_frameLog* fl,
unsigned compressedSize,
unsigned decompressedSize,
unsigned checksum)
{
if (fl->size == ZSTD_SEEKABLE_MAXFRAMES)
return ERROR(frameIndex_tooLarge);
/* grow the buffer if required */
if (fl->size == fl->capacity) {
/* exponential size increase for constant amortized runtime */
size_t const newCapacity = fl->capacity * 2;
framelogEntry_t* const newEntries = realloc(fl->entries,
sizeof(framelogEntry_t) * newCapacity);
if (newEntries == NULL) return ERROR(memory_allocation);
fl->entries = newEntries;
assert(newCapacity <= UINT_MAX);
fl->capacity = (U32)newCapacity;
}
fl->entries[fl->size] = (framelogEntry_t){
compressedSize, decompressedSize, checksum
};
fl->size++;
return 0;
}
size_t ZSTD_seekable_endFrame(ZSTD_seekable_CStream* zcs, ZSTD_outBuffer* output)
{
size_t const prevOutPos = output->pos;
/* end the frame */
size_t ret = ZSTD_endStream(zcs->cstream, output);
zcs->frameCSize += output->pos - prevOutPos;
/* need to flush before doing the rest */
if (ret) return ret;
/* frame done */
/* store the frame data for later */
ret = ZSTD_seekable_logFrame(
&zcs->framelog, zcs->frameCSize, zcs->frameDSize,
zcs->framelog.checksumFlag
? XXH64_digest(&zcs->xxhState) & 0xFFFFFFFFU
: 0);
if (ret) return ret;
/* reset for the next frame */
zcs->frameCSize = 0;
zcs->frameDSize = 0;
ZSTD_resetCStream(zcs->cstream, 0);
if (zcs->framelog.checksumFlag)
XXH64_reset(&zcs->xxhState, 0);
return 0;
}
size_t ZSTD_seekable_compressStream(ZSTD_seekable_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input)
{
const BYTE* const inBase = (const BYTE*) input->src + input->pos;
size_t inLen = input->size - input->pos;
inLen = MIN(inLen, (size_t)(zcs->maxFrameSize - zcs->frameDSize));
/* if we haven't finished flushing the last frame, don't start writing a new one */
if (inLen > 0) {
ZSTD_inBuffer inTmp = { inBase, inLen, 0 };
size_t const prevOutPos = output->pos;
size_t const ret = ZSTD_compressStream(zcs->cstream, output, &inTmp);
if (zcs->framelog.checksumFlag) {
XXH64_update(&zcs->xxhState, inBase, inTmp.pos);
}
zcs->frameCSize += output->pos - prevOutPos;
zcs->frameDSize += inTmp.pos;
input->pos += inTmp.pos;
if (ZSTD_isError(ret)) return ret;
}
if (zcs->maxFrameSize == zcs->frameDSize) {
/* log the frame and start over */
size_t const ret = ZSTD_seekable_endFrame(zcs, output);
if (ZSTD_isError(ret)) return ret;
/* get the client ready for the next frame */
return (size_t)zcs->maxFrameSize;
}
return (size_t)(zcs->maxFrameSize - zcs->frameDSize);
}
static inline size_t ZSTD_seekable_seekTableSize(const ZSTD_frameLog* fl)
{
size_t const sizePerFrame = 8 + (fl->checksumFlag?4:0);
size_t const seekTableLen = ZSTD_SKIPPABLEHEADERSIZE +
sizePerFrame * fl->size +
ZSTD_seekTableFooterSize;
return seekTableLen;
}
static inline size_t ZSTD_stwrite32(ZSTD_frameLog* fl,
ZSTD_outBuffer* output, U32 const value,
U32 const offset)
{
if (fl->seekTablePos < offset + 4) {
BYTE tmp[4]; /* so that we can work with buffers too small to write a whole word to */
size_t const lenWrite =
MIN(output->size - output->pos, offset + 4 - fl->seekTablePos);
MEM_writeLE32(tmp, value);
memcpy((BYTE*)output->dst + output->pos,
tmp + (fl->seekTablePos - offset), lenWrite);
output->pos += lenWrite;
fl->seekTablePos += lenWrite;
if (lenWrite < 4) return ZSTD_seekable_seekTableSize(fl) - fl->seekTablePos;
}
return 0;
}
size_t ZSTD_seekable_writeSeekTable(ZSTD_frameLog* fl, ZSTD_outBuffer* output)
{
/* seekTableIndex: the current index in the table and
* seekTableSize: the amount of the table written so far
*
* This function is written this way so that if it has to return early
* because of a small buffer, it can keep going where it left off.
*/
size_t const sizePerFrame = 8 + (fl->checksumFlag?4:0);
size_t const seekTableLen = ZSTD_seekable_seekTableSize(fl);
CHECK_Z(ZSTD_stwrite32(fl, output, ZSTD_MAGIC_SKIPPABLE_START | 0xE, 0));
assert(seekTableLen <= (size_t)UINT_MAX);
CHECK_Z(ZSTD_stwrite32(fl, output, (U32)seekTableLen - ZSTD_SKIPPABLEHEADERSIZE, 4));
while (fl->seekTableIndex < fl->size) {
unsigned long long const start = ZSTD_SKIPPABLEHEADERSIZE + sizePerFrame * fl->seekTableIndex;
assert(start + 8 <= UINT_MAX);
CHECK_Z(ZSTD_stwrite32(fl, output,
fl->entries[fl->seekTableIndex].cSize,
(U32)start + 0));
CHECK_Z(ZSTD_stwrite32(fl, output,
fl->entries[fl->seekTableIndex].dSize,
(U32)start + 4));
if (fl->checksumFlag) {
CHECK_Z(ZSTD_stwrite32(
fl, output, fl->entries[fl->seekTableIndex].checksum,
(U32)start + 8));
}
fl->seekTableIndex++;
}
assert(seekTableLen <= UINT_MAX);
CHECK_Z(ZSTD_stwrite32(fl, output, fl->size,
(U32)seekTableLen - ZSTD_seekTableFooterSize));
if (output->size - output->pos < 1) return seekTableLen - fl->seekTablePos;
if (fl->seekTablePos < seekTableLen - 4) {
BYTE sfd = 0;
sfd |= (fl->checksumFlag) << 7;
((BYTE*)output->dst)[output->pos] = sfd;
output->pos++;
fl->seekTablePos++;
}
CHECK_Z(ZSTD_stwrite32(fl, output, ZSTD_SEEKABLE_MAGICNUMBER,
(U32)seekTableLen - 4));
if (fl->seekTablePos != seekTableLen) return ERROR(GENERIC);
return 0;
}
size_t ZSTD_seekable_endStream(ZSTD_seekable_CStream* zcs, ZSTD_outBuffer* output)
{
if (!zcs->writingSeekTable && zcs->frameDSize) {
const size_t endFrame = ZSTD_seekable_endFrame(zcs, output);
if (ZSTD_isError(endFrame)) return endFrame;
/* return an accurate size hint */
if (endFrame) return endFrame + ZSTD_seekable_seekTableSize(&zcs->framelog);
}
zcs->writingSeekTable = 1;
return ZSTD_seekable_writeSeekTable(&zcs->framelog, output);
}
#ifndef SEEKABLE_H
#define SEEKABLE_H
#if defined (__cplusplus)
extern "C" {
#endif
#include <stdio.h>
#include "zstd.h" /* ZSTDLIB_API */
#define ZSTD_seekTableFooterSize 9
#define ZSTD_SEEKABLE_MAGICNUMBER 0x8F92EAB1
#define ZSTD_SEEKABLE_MAXFRAMES 0x8000000U
/* Limit the maximum size to avoid any potential issues storing the compressed size */
#define ZSTD_SEEKABLE_MAX_FRAME_DECOMPRESSED_SIZE 0x80000000U
/*-****************************************************************************
* Seekable Format
*
* The seekable format splits the compressed data into a series of "frames",
* each compressed individually so that decompression of a section in the
* middle of an archive only requires zstd to decompress at most a frame's
* worth of extra data, instead of the entire archive.
******************************************************************************/
typedef struct ZSTD_seekable_CStream_s ZSTD_seekable_CStream;
typedef struct ZSTD_seekable_s ZSTD_seekable;
/*-****************************************************************************
* Seekable compression - HowTo
* A ZSTD_seekable_CStream object is required to tracking streaming operation.
* Use ZSTD_seekable_createCStream() and ZSTD_seekable_freeCStream() to create/
* release resources.
*
* Streaming objects are reusable to avoid allocation and deallocation,
* to start a new compression operation call ZSTD_seekable_initCStream() on the
* compressor.
*
* Data streamed to the seekable compressor will automatically be split into
* frames of size `maxFrameSize` (provided in ZSTD_seekable_initCStream()),
* or if none is provided, will be cut off whenever ZSTD_seekable_endFrame() is
* called or when the default maximum frame size (2GB) is reached.
*
* Use ZSTD_seekable_initCStream() to initialize a ZSTD_seekable_CStream object
* for a new compression operation.
* `maxFrameSize` indicates the size at which to automatically start a new
* seekable frame. `maxFrameSize == 0` implies the default maximum size.
* `checksumFlag` indicates whether or not the seek table should include frame
* checksums on the uncompressed data for verification.
* @return : a size hint for input to provide for compression, or an error code
* checkable with ZSTD_isError()
*
* Use ZSTD_seekable_compressStream() repetitively to consume input stream.
* The function will automatically update both `pos` fields.
* Note that it may not consume the entire input, in which case `pos < size`,
* and it's up to the caller to present again remaining data.
* @return : a size hint, preferred nb of bytes to use as input for next
* function call or an error code, which can be tested using
* ZSTD_isError().
* Note 1 : it's just a hint, to help latency a little, any other
* value will work fine.
*
* At any time, call ZSTD_seekable_endFrame() to end the current frame and
* start a new one.
*
* ZSTD_seekable_endStream() will end the current frame, and then write the seek
* table so that decompressors can efficiently find compressed frames.
* ZSTD_seekable_endStream() may return a number > 0 if it was unable to flush
* all the necessary data to `output`. In this case, it should be called again
* until all remaining data is flushed out and 0 is returned.
******************************************************************************/
/*===== Seekable compressor management =====*/
ZSTDLIB_API ZSTD_seekable_CStream* ZSTD_seekable_createCStream(void);
ZSTDLIB_API size_t ZSTD_seekable_freeCStream(ZSTD_seekable_CStream* zcs);
/*===== Seekable compression functions =====*/
ZSTDLIB_API size_t ZSTD_seekable_initCStream(ZSTD_seekable_CStream* zcs, int compressionLevel, int checksumFlag, unsigned maxFrameSize);
ZSTDLIB_API size_t ZSTD_seekable_compressStream(ZSTD_seekable_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input);
ZSTDLIB_API size_t ZSTD_seekable_endFrame(ZSTD_seekable_CStream* zcs, ZSTD_outBuffer* output);
ZSTDLIB_API size_t ZSTD_seekable_endStream(ZSTD_seekable_CStream* zcs, ZSTD_outBuffer* output);
/*= Raw seek table API
* These functions allow for the seek table to be constructed directly.
* This table can then be appended to a file of concatenated frames.
* This allows the frames to be compressed independently, even in parallel,
* and compiled together afterward into a seekable archive.
*
* Use ZSTD_seekable_createFrameLog() to allocate and initialize a tracking
* structure.
*
* Call ZSTD_seekable_logFrame() once for each frame in the archive.
* checksum is optional, and will not be used if checksumFlag was 0 when the
* frame log was created. If present, it should be the least significant 32
* bits of the XXH64 hash of the uncompressed data.
*
* Call ZSTD_seekable_writeSeekTable to serialize the data into a seek table.
* If the entire table was written, the return value will be 0. Otherwise,
* it will be equal to the number of bytes left to write. */
typedef struct ZSTD_frameLog_s ZSTD_frameLog;
ZSTDLIB_API ZSTD_frameLog* ZSTD_seekable_createFrameLog(int checksumFlag);
ZSTDLIB_API size_t ZSTD_seekable_freeFrameLog(ZSTD_frameLog* fl);
ZSTDLIB_API size_t ZSTD_seekable_logFrame(ZSTD_frameLog* fl, unsigned compressedSize, unsigned decompressedSize, unsigned checksum);
ZSTDLIB_API size_t ZSTD_seekable_writeSeekTable(ZSTD_frameLog* fl, ZSTD_outBuffer* output);
/*-****************************************************************************
* Seekable decompression - HowTo
* A ZSTD_seekable object is required to tracking the seekTable.
*
* Call ZSTD_seekable_init* to initialize a ZSTD_seekable object with the
* the seek table provided in the input.
* There are three modes for ZSTD_seekable_init:
* - ZSTD_seekable_initBuff() : An in-memory API. The data contained in
* `src` should be the entire seekable file, including the seek table.
* `src` should be kept alive and unmodified until the ZSTD_seekable object
* is freed or reset.
* - ZSTD_seekable_initFile() : A simplified file API using stdio. fread and
* fseek will be used to access the required data for building the seek
* table and doing decompression operations. `src` should not be closed
* or modified until the ZSTD_seekable object is freed or reset.
* - ZSTD_seekable_initAdvanced() : A general API allowing the client to
* provide its own read and seek callbacks.
* + ZSTD_seekable_read() : read exactly `n` bytes into `buffer`.
* Premature EOF should be treated as an error.
* + ZSTD_seekable_seek() : seek the read head to `offset` from `origin`,
* where origin is either SEEK_SET (beginning of
* file), or SEEK_END (end of file).
* Both functions should return a non-negative value in case of success, and a
* negative value in case of failure. If implementing using this API and
* stdio, be careful with files larger than 4GB and fseek. All of these
* functions return an error code checkable with ZSTD_isError().
*
* Call ZSTD_seekable_decompress to decompress `dstSize` bytes at decompressed
* offset `offset`. ZSTD_seekable_decompress may have to decompress the entire
* prefix of the frame before the desired data if it has not already processed
* this section. If ZSTD_seekable_decompress is called multiple times for a
* consecutive range of data, it will efficiently retain the decompressor object
* and avoid redecompressing frame prefixes. The return value is the number of
* bytes decompressed, or an error code checkable with ZSTD_isError().
*
* The seek table access functions can be used to obtain the data contained
* in the seek table. If frameIndex is larger than the value returned by
* ZSTD_seekable_getNumFrames(), they will return error codes checkable with
* ZSTD_isError(). Note that since the offset access functions return
* unsigned long long instead of size_t, in this case they will instead return
* the value ZSTD_SEEKABLE_FRAMEINDEX_TOOLARGE.
******************************************************************************/
/*===== Seekable decompressor management =====*/
ZSTDLIB_API ZSTD_seekable* ZSTD_seekable_create(void);
ZSTDLIB_API size_t ZSTD_seekable_free(ZSTD_seekable* zs);
/*===== Seekable decompression functions =====*/
ZSTDLIB_API size_t ZSTD_seekable_initBuff(ZSTD_seekable* zs, const void* src, size_t srcSize);
ZSTDLIB_API size_t ZSTD_seekable_initFile(ZSTD_seekable* zs, FILE* src);
ZSTDLIB_API size_t ZSTD_seekable_decompress(ZSTD_seekable* zs, void* dst, size_t dstSize, unsigned long long offset);
ZSTDLIB_API size_t ZSTD_seekable_decompressFrame(ZSTD_seekable* zs, void* dst, size_t dstSize, unsigned frameIndex);
#define ZSTD_SEEKABLE_FRAMEINDEX_TOOLARGE (0ULL-2)
/*===== Seek Table access functions =====*/
ZSTDLIB_API unsigned ZSTD_seekable_getNumFrames(ZSTD_seekable* const zs);
ZSTDLIB_API unsigned long long ZSTD_seekable_getFrameCompressedOffset(ZSTD_seekable* const zs, unsigned frameIndex);
ZSTDLIB_API unsigned long long ZSTD_seekable_getFrameDecompressedOffset(ZSTD_seekable* const zs, unsigned frameIndex);
ZSTDLIB_API size_t ZSTD_seekable_getFrameCompressedSize(ZSTD_seekable* const zs, unsigned frameIndex);
ZSTDLIB_API size_t ZSTD_seekable_getFrameDecompressedSize(ZSTD_seekable* const zs, unsigned frameIndex);
ZSTDLIB_API unsigned ZSTD_seekable_offsetToFrameIndex(ZSTD_seekable* const zs, unsigned long long offset);
/*===== Seekable advanced I/O API =====*/
typedef int(ZSTD_seekable_read)(void* opaque, void* buffer, size_t n);
typedef int(ZSTD_seekable_seek)(void* opaque, long long offset, int origin);
typedef struct {
void* opaque;
ZSTD_seekable_read* read;
ZSTD_seekable_seek* seek;
} ZSTD_seekable_customFile;
ZSTDLIB_API size_t ZSTD_seekable_initAdvanced(ZSTD_seekable* zs, ZSTD_seekable_customFile src);
#if defined (__cplusplus)
}
#endif
#endif
#![allow(non_upper_case_globals)]
#![allow(non_camel_case_types)]
#![allow(non_snake_case)]
#[macro_use]
extern crate thiserror;
include!(concat!(env!("OUT_DIR"), "/bindings.rs"));
use libc::*;
use std::ffi::{CStr, CString};
pub fn out_size() -> size_t {
unsafe { ZSTD_DStreamOutSize() }
}
/// The type of seekable compressors.
pub struct SeekableCStream {
p: *mut ZSTD_seekable_CStream_s,
}
/// The type of compressors.
pub struct CStream {
p: *mut ZSTD_CStream,
}
/// The type of compressors.
pub struct FrameLog {
p: *mut ZSTD_frameLog,
}
/// The type of decompressors.
pub struct Seekable<'a, R> {
p: *mut ZSTD_seekable,
b: *mut R,
f: *mut libc::FILE,
marker: std::marker::PhantomData<&'a R>,
}
unsafe impl<R> Send for Seekable<'static, R> {}
unsafe impl Send for SeekableCStream {}
unsafe impl Send for FrameLog {}
unsafe impl<D: Dst> Send for CompressedFrame<D> {}
#[derive(Error)]
pub enum Error {
Null,
CouldNotOpenFile(String),
ZSTD(size_t),
}
impl std::fmt::Debug for Error {
fn fmt(&self, fmt: &mut std::fmt::Formatter) -> std::fmt::Result {
match *self {
Error::Null => write!(fmt, "Null"),
Error::CouldNotOpenFile(ref f) => write!(fmt, "Could not open {}", f),
Error::ZSTD(r) => unsafe {
let error = CStr::from_ptr(ZSTD_getErrorName(r));
write!(fmt, "ZSTD({:?})", error)
},
}
}
}
impl std::fmt::Display for Error {
fn fmt(&self, fmt: &mut std::fmt::Formatter) -> std::fmt::Result {
write!(fmt, "{:?}", self)
}
}
impl Drop for SeekableCStream {
fn drop(&mut self) {
unsafe {
if !self.p.is_null() {
ZSTD_seekable_freeCStream(self.p);
self.p = std::ptr::null_mut();
}
}
}
}
impl Drop for CStream {
fn drop(&mut self) {
unsafe {
if !self.p.is_null() {
ZSTD_freeCStream(self.p);
self.p = std::ptr::null_mut();
}
}
}
}
impl<'a, R> Drop for Seekable<'a, R> {
fn drop(&mut self) {
unsafe {
if !self.p.is_null() {
ZSTD_seekable_free(self.p);
self.p = std::ptr::null_mut();
}
if !self.f.is_null() {
fclose(self.f);
self.f = std::ptr::null_mut();
}
if !self.b.is_null() {
let b: Box<R> = Box::from_raw(self.b);
std::mem::drop(b);
self.b = std::ptr::null_mut();
}
}
}
}
impl Drop for FrameLog {
fn drop(&mut self) {
unsafe {
if !self.p.is_null() {
ZSTD_seekable_freeFrameLog(self.p);
self.p = std::ptr::null_mut()
}
}
}
}
unsafe extern "C" fn zstd_seekable_read<R: std::io::Read + std::io::Seek>(
opaque: *mut c_void,
buffer: *mut c_void,
n: size_t,
) -> c_int {
let mut b: Box<R> = Box::from_raw(opaque as *mut R);
let s = std::slice::from_raw_parts_mut(buffer as *mut u8, n as usize);
let result = if b.read_exact(s).is_ok() { 0 } else { 1 };
std::mem::forget(b);
result
}
unsafe extern "C" fn zstd_seekable_seek<R: std::io::Read + std::io::Seek>(
opaque: *mut c_void,
offset: c_longlong,
origin: c_int,
) -> c_int {
let mut b: Box<R> = Box::from_raw(opaque as *mut R);
use std::io::SeekFrom;
let origin = if origin == libc::SEEK_SET {
SeekFrom::Start(offset as u64)
} else if origin == libc::SEEK_CUR {
SeekFrom::Current(offset as i64)
} else {
SeekFrom::End(offset as i64)
};
let result = if b.seek(origin).is_ok() { 0 } else { 1 };
std::mem::forget(b);
result
}
impl SeekableCStream {
/// Create a compressor with the given level and frame size. When seeking in the file, frames are decompressed one by one, so this should be chosen of a size similar to the chunks that will be decompressed.
pub fn new(level: usize, frame_size: usize) -> Result<Self, Error> {
unsafe {
let p = ZSTD_seekable_createCStream();
if p.is_null() {
return Err(Error::Null);
}
let result = ZSTD_seekable_initCStream(p, level as c_int, 1, frame_size as c_uint);
if ZSTD_isError(result) != 0 {
return Err(Error::ZSTD(result));
}
Ok(SeekableCStream { p })
}
}
/// Compress one chunk of input, and write it into the output. The `output` array must be large enough to hold the result. If successful, this function returns two integers `(out_pos, in_pos)`, where `out_pos` is the number of bytes written in `output`, and `in_pos` is the number of input bytes consumed.
pub fn compress(&mut self, output: &mut [u8], input: &[u8]) -> Result<(usize, usize), Error> {
unsafe {
let mut input = ZSTD_inBuffer {
src: input.as_ptr() as *const c_void,
size: input.len() as size_t,
pos: 0,
};
let mut output = ZSTD_outBuffer {
dst: output.as_mut_ptr() as *mut c_void,
size: output.len() as size_t,
pos: 0,
};
let result = ZSTD_seekable_compressStream(self.p, &mut output, &mut input);
if ZSTD_isError(result) != 0 {
return Err(Error::ZSTD(result));
}
Ok((output.pos as usize, input.pos as usize))
}
}
/// Finish writing the message, i.e. write the remaining pending block.
pub fn end_stream(&mut self, output: &mut [u8]) -> Result<usize, Error> {
unsafe {
let mut output = ZSTD_outBuffer {
dst: output.as_mut_ptr() as *mut c_void,
size: output.len() as size_t,
pos: 0,
};
let result = ZSTD_seekable_endStream(self.p, &mut output);
if ZSTD_isError(result) != 0 {
return Err(Error::ZSTD(result));
}
Ok(output.pos as usize)
}
}
}
impl CStream {
pub fn new(level: usize) -> Result<Self, Error> {
unsafe {
let p = ZSTD_createCStream();
if p.is_null() {
return Err(Error::Null);
}
let result = ZSTD_initCStream(p, level as c_int);
if ZSTD_isError(result) != 0 {
return Err(Error::ZSTD(result));
}
Ok(CStream { p })
}
}
pub fn in_size() -> usize {
unsafe { ZSTD_CStreamInSize() as usize }
}
pub fn out_size() -> usize {
unsafe { ZSTD_CStreamOutSize() as usize }
}
/// Compress one chunk of input, and write it into the output. The `output` array must be large enough to hold the result. If successful, this function returns three integers `(out_pos, in_pos, next_read_size)`, where `out_pos` is the number of bytes written in `output`, `in_pos` is the number of input bytes consumed, and `next_read_size` is a hint for the next read size.
pub fn compress(
&mut self,
output: &mut [u8],
input: &[u8],
) -> Result<(usize, usize, usize), Error> {
unsafe {
let mut input = ZSTD_inBuffer {
src: input.as_ptr() as *const c_void,
size: input.len() as size_t,
pos: 0,
};
let mut output = ZSTD_outBuffer {
dst: output.as_mut_ptr() as *mut c_void,
size: output.len() as size_t,
pos: 0,
};
let result = ZSTD_compressStream(self.p, &mut output, &mut input);
Ok((output.pos as usize, input.pos as usize, result as usize))
}
}
pub fn compress2(
&mut self,
output: &mut [u8],
input: &[u8],
op: EndDirective,
) -> Result<(usize, usize, usize), Error> {
unsafe {
let mut input = ZSTD_inBuffer {
src: input.as_ptr() as *const c_void,
size: input.len() as size_t,
pos: 0,
};
let mut output = ZSTD_outBuffer {
dst: output.as_mut_ptr() as *mut c_void,
size: output.len() as size_t,
pos: 0,
};
let result = ZSTD_compressStream2(self.p, &mut output, &mut input, op as c_uint);
Ok((output.pos as usize, input.pos as usize, result as usize))
}
}
pub fn flush(&mut self, output: &mut [u8]) -> Result<(usize, usize), Error> {
unsafe {
let mut output = ZSTD_outBuffer {
dst: output.as_mut_ptr() as *mut c_void,
size: output.len() as size_t,
pos: 0,
};
let result = ZSTD_flushStream(self.p, &mut output);
Ok((output.pos as usize, result as usize))
}
}
/// Finish writing the message, i.e. write the remaining pending block.
pub fn end(&mut self, output: &mut [u8]) -> Result<usize, Error> {
unsafe {
let mut output = ZSTD_outBuffer {
dst: output.as_mut_ptr() as *mut c_void,
size: output.len() as size_t,
pos: 0,
};
let result = ZSTD_endStream(self.p, &mut output);
if ZSTD_isError(result) != 0 {
return Err(Error::ZSTD(result));
}
Ok(output.pos as usize)
}
}
}
pub enum EndDirective {
Continue = 0,
Flush = 1,
End = 2,
}
pub struct DStream {
p: *mut ZSTD_DStream,
}
impl Drop for DStream {
fn drop(&mut self) {
unsafe {
ZSTD_freeDStream(self.p);
self.p = std::ptr::null_mut();
}
}
}
impl DStream {
pub fn new() -> Result<Self, Error> {
unsafe {
let p = ZSTD_createDStream();
if p.is_null() {
return Err(Error::Null);
}
Ok(DStream { p })
}
}
pub fn decompress(&mut self, output: &mut [u8], input: &[u8]) -> Result<(usize, usize), Error> {
unsafe {
let mut input = ZSTD_inBuffer {
src: input.as_ptr() as *const c_void,
size: input.len() as size_t,
pos: 0,
};
let mut output = ZSTD_outBuffer {
dst: output.as_mut_ptr() as *mut c_void,
size: output.len() as size_t,
pos: 0,
};
let _result = ZSTD_decompressStream(self.p, &mut output, &mut input);
Ok((output.pos as usize, input.pos as usize))
}
}
}
impl<'a> Seekable<'a, ()> {
/// Initialise a decompressor with an input buffer.
pub fn init_buf(input: &'a [u8]) -> Result<Self, Error> {
unsafe {
let p = ZSTD_seekable_create();
if p.is_null() {
return Err(Error::Null);
}
let result =
ZSTD_seekable_initBuff(p, input.as_ptr() as *const c_void, input.len() as size_t);
if ZSTD_isError(result) != 0 {
return Err(Error::ZSTD(result));
}
Ok(Seekable {
p,
f: std::ptr::null_mut(),
b: std::ptr::null_mut(),
marker: std::marker::PhantomData,
})
}
}
/// Initialise a decompressor with a file. This method opens the file, and dropping the resulting `Seekable` closes the file.
pub fn init_file(name_: &str) -> Result<Self, Error> {
unsafe {
let name = CString::new(name_).unwrap();
let f: *mut libc::FILE = fopen(name.as_ptr(), "rb\0".as_ptr() as *const c_char);
if f.is_null() {
return Err(Error::CouldNotOpenFile(name_.to_string()));
}
let p = ZSTD_seekable_create();
if p.is_null() {
return Err(Error::Null);
}
let result = ZSTD_seekable_initFile(p, f as *mut FILE);
if ZSTD_isError(result) != 0 {
return Err(Error::ZSTD(result));
}
Ok(Seekable {
p,
f,
b: std::ptr::null_mut(),
marker: std::marker::PhantomData,
})
}
}
}
impl<'a, R: std::io::Read + std::io::Seek> Seekable<'a, R> {
/// Initialise a decompressor with a file. This method opens the file, and dropping the resulting `Seekable` closes the file.
pub fn init(r: Box<R>) -> Result<Self, Error> {
unsafe {
let p = ZSTD_seekable_create();
if p.is_null() {
return Err(Error::Null);
}
let opaque = Box::into_raw(r) as *mut R;
let adv = ZSTD_seekable_customFile {
opaque: opaque as *mut c_void,
read: Some(zstd_seekable_read::<R>),
seek: Some(zstd_seekable_seek::<R>),
};
let result = ZSTD_seekable_initAdvanced(p, adv);
if ZSTD_isError(result) != 0 {
return Err(Error::ZSTD(result));
}
Ok(Seekable {
p,
f: std::ptr::null_mut(),
b: opaque,
marker: std::marker::PhantomData,
})
}
}
pub fn into_inner(self) -> Box<R> {
unsafe { Box::from_raw(self.b) }
}
}
impl<'a, R> Seekable<'a, R> {
/// Decompress starting from an offset. The length of `out` must
/// be at most the size of the decompressed output.
///
/// This function finds the correct frame to start with, and takes
/// care of decompressing multiple frames in a row.
pub fn decompress(&mut self, out: &mut [u8], offset: u64) -> Result<usize, Error> {
unsafe {
let result = ZSTD_seekable_decompress(
self.p,
out.as_mut_ptr() as *mut c_void,
out.len() as size_t,
offset,
);
if ZSTD_isError(result) != 0 {
return Err(Error::ZSTD(result));
}
Ok(result as usize)
}
}
/// Number of frames in the message.
pub fn get_num_frames(&self) -> usize {
unsafe { ZSTD_seekable_getNumFrames(self.p) as usize }
}
/// Offset of the frame in the compressed data.
pub fn get_frame_compressed_offset(&self, frame_index: usize) -> size_t {
unsafe { ZSTD_seekable_getFrameCompressedOffset(self.p, frame_index as c_uint) }
}
/// Size of the frame in the compressed data.
pub fn get_frame_compressed_size(&self, frame_index: usize) -> usize {
unsafe { ZSTD_seekable_getFrameCompressedSize(self.p, frame_index as c_uint) as usize }
}
/// Offset of the frame in the decompressed data.
pub fn get_frame_decompressed_offset(&self, frame_index: usize) -> u64 {
unsafe { ZSTD_seekable_getFrameDecompressedOffset(self.p, frame_index as c_uint) }
}
/// Size of the frame in the decompressed data.
pub fn get_frame_decompressed_size(&self, frame_index: usize) -> usize {
unsafe { ZSTD_seekable_getFrameDecompressedSize(self.p, frame_index as c_uint) as usize }
}
/// Decompress a single frame. This method internally calls `decompress`, and `dest` must be exactly the size of the uncompressed frame.
pub fn decompress_frame(&mut self, dest: &mut [u8], index: usize) -> usize {
unsafe {
ZSTD_seekable_decompressFrame(
self.p,
dest.as_mut_ptr() as *mut c_void,
dest.len() as size_t,
index as c_uint,
) as usize
}
}
/// Perform a binary search to find the frame containing the offset.
pub fn seekable_offset_to_frame_index(&mut self, offset: u64) -> usize {
unsafe { ZSTD_seekable_offsetToFrameIndex(self.p, offset) as usize }
}
}
pub trait Dst: Send {
fn as_mut_ptr(&mut self) -> *mut u8;
fn as_slice(&self) -> &[u8];
fn len(&self) -> usize;
fn new() -> Self;
}
impl Dst for [u8; 256] {
fn as_mut_ptr(&mut self) -> *mut u8 {
self.as_mut().as_mut_ptr()
}
fn as_slice(&self) -> &[u8] {
self.as_ref()
}
fn len(&self) -> usize {
256
}
fn new() -> Self {
unsafe { std::mem::MaybeUninit::uninit().assume_init() }
}
}
impl Dst for [u8; 512] {
fn as_mut_ptr(&mut self) -> *mut u8 {
self.as_mut().as_mut_ptr()
}
fn as_slice(&self) -> &[u8] {
self.as_ref()
}
fn len(&self) -> usize {
512
}
fn new() -> Self {
unsafe { std::mem::MaybeUninit::uninit().assume_init() }
}
}
impl Dst for [u8; 1024] {
fn as_mut_ptr(&mut self) -> *mut u8 {
self.as_mut().as_mut_ptr()
}
fn as_slice(&self) -> &[u8] {
self.as_ref()
}
fn len(&self) -> usize {
1024
}
fn new() -> Self {
unsafe { std::mem::MaybeUninit::uninit().assume_init() }
}
}
impl Dst for [u8; 2048] {
fn as_mut_ptr(&mut self) -> *mut u8 {
self.as_mut().as_mut_ptr()
}
fn as_slice(&self) -> &[u8] {
self.as_ref()
}
fn len(&self) -> usize {
2048
}
fn new() -> Self {
unsafe { std::mem::MaybeUninit::uninit().assume_init() }
}
}
impl Dst for [u8; 4096] {
fn as_mut_ptr(&mut self) -> *mut u8 {
self.as_mut().as_mut_ptr()
}
fn as_slice(&self) -> &[u8] {
self.as_ref()
}
fn len(&self) -> usize {
4096
}
fn new() -> Self {
unsafe { std::mem::MaybeUninit::uninit().assume_init() }
}
}
pub struct CompressedFrame<D: Dst> {
src_size: size_t,
dst_size: size_t,
checksum: c_uint,
dst: D,
}
impl<D: Dst> CompressedFrame<D> {
fn as_slice(&self) -> &[u8] {
&self.dst.as_slice()[..self.dst_size as usize]
}
}
extern "C" {
fn xxh64(src: *const u8, len: c_int) -> c_uint;
}
fn compress_frame<D: Dst>(
src_ptr: *const u8,
src_len: size_t,
level: usize,
) -> Result<CompressedFrame<D>, Error> {
unsafe {
let mut dst = D::new();
let checksum = xxh64(src_ptr, src_len as c_int);
let ret = ZSTD_compress(
dst.as_mut_ptr() as *mut c_void,
dst.len() as size_t,
src_ptr as *const c_void,
src_len,
level as c_int,
);
if ZSTD_isError(ret) != 0 {
return Err(Error::ZSTD(ret));
}
println!("{:?}", dst.as_slice());
Ok(CompressedFrame {
src_size: src_len,
dst_size: ret,
checksum,
dst,
})
}
}
impl FrameLog {
pub fn new() -> Self {
let p = unsafe { ZSTD_seekable_createFrameLog(1) };
assert!(!p.is_null());
FrameLog { p }
}
pub fn log_frame<D: Dst>(&mut self, frame: &CompressedFrame<D>) -> usize {
unsafe {
ZSTD_seekable_logFrame(
self.p,
frame.dst_size as c_uint,
frame.src_size as c_uint,
frame.checksum,
) as usize
}
}
pub fn write_all<W: std::io::Write>(&self, mut w: W) -> Result<(), std::io::Error> {
let mut output = [0; 1024];
let mut output_ = ZSTD_outBuffer {
dst: output.as_mut_ptr() as *mut c_void,
size: 1024,
pos: 0,
};
unsafe {
while ZSTD_seekable_writeSeekTable(self.p, &mut output_) != 0 {
w.write_all(&output[..output_.pos as usize])?;
output_.pos = 0;
}
w.write_all(&output[..output_.pos as usize])?;
}
Ok(())
}
}
struct Chunk {
p: *const u8,
len: size_t,
}
unsafe impl Send for Chunk {}
pub fn parallel_compress<W: std::io::Write, D: Dst + 'static>(
src: &[u8],
mut output: W,
level: usize,
jobs: usize,
) -> Result<(), anyhow::Error> {
use std::sync::mpsc::channel;
use threadpool::ThreadPool;
let chunk_size = 256;
let n = src.len() / chunk_size + if src.len() % 256 == 0 { 0 } else { 1 };
let pool = ThreadPool::new(jobs);
let (tx, rx) = channel();
for (i, chunk) in src.chunks(chunk_size).enumerate() {
let tx = tx.clone();
let chunk = Chunk {
p: chunk.as_ptr(),
len: chunk.len() as size_t,
};
pool.execute(move || {
let frame = compress_frame(chunk.p, chunk.len, level).unwrap();
tx.send((i, frame))
.expect("channel will be there waiting for the pool");
});
}
let frames: Vec<CompressedFrame<D>> = unsafe {
let mut frames = Vec::with_capacity(n);
frames.set_len(n);
for (i, frame) in rx.iter().take(n) {
frames[i] = frame
}
frames
};
let mut log = FrameLog::new();
for frame in frames.iter() {
output.write_all(frame.as_slice())?;
log.log_frame(frame);
}
log.write_all(&mut output)?;
Ok(())
}
#[test]
fn test_par() {
let input = b"Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut diam ante, sollicitudin a dolor et, volutpat elementum nulla. Etiam nec efficitur nibh, quis rutrum risus. Maecenas quis lorem malesuada, aliquet mi vel, viverra nunc. Donec et nulla sed velit sagittis varius. Suspendisse vestibulum, neque lobortis ornare vestibulum, orci turpis vulputate nisi, ut sodales neque purus eget magna. Nunc condimentum, diam eu consequat venenatis, est nisl semper lorem, et lobortis velit justo sed nulla. Nunc sit amet tempor nunc, vel posuere ipsum. Cras erat tortor, pulvinar ac pretium eu, auctor ac nibh. Duis iaculis porta magna, eu lobortis elit. Duis vitae massa eros. Nulla non magna accumsan, egestas quam sit amet, laoreet lectus.";
let mut output = Vec::new();
parallel_compress::<&mut Vec<u8>, [u8; 256]>(input, &mut output, 10, 4).unwrap();
let mut decomp = Vec::new();
let mut s = { Seekable::init_buf(&output).unwrap() };
for frame in 0..s.get_num_frames() {
let size = s.get_frame_decompressed_size(frame);
let n = decomp.len();
decomp.extend(std::iter::repeat(0).take(size));
s.decompress_frame(&mut decomp[n..], frame);
}
println!("{:?}", std::str::from_utf8(&decomp).unwrap());
assert_eq!(&input[..], &decomp[..])
}
#[test]
fn test() {
let mut cstream = SeekableCStream::new(10, 256).unwrap();
let input = b"Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut diam ante, sollicitudin a dolor et, volutpat elementum nulla. Etiam nec efficitur nibh, quis rutrum risus. Maecenas quis lorem malesuada, aliquet mi vel, viverra nunc. Donec et nulla sed velit sagittis varius. Suspendisse vestibulum, neque lobortis ornare vestibulum, orci turpis vulputate nisi, ut sodales neque purus eget magna. Nunc condimentum, diam eu consequat venenatis, est nisl semper lorem, et lobortis velit justo sed nulla. Nunc sit amet tempor nunc, vel posuere ipsum. Cras erat tortor, pulvinar ac pretium eu, auctor ac nibh. Duis iaculis porta magna, eu lobortis elit. Duis vitae massa eros. Nulla non magna accumsan, egestas quam sit amet, laoreet lectus.";
let mut input_pos = 0;
let mut output = vec![0; input.len()];
let mut output_pos = 0;
while input_pos < input.len() {
let (a, b) = cstream
.compress(&mut output[output_pos..], &input[input_pos..])
.unwrap();
output_pos += a;
input_pos += b;
}
while let Ok(n) = cstream.end_stream(&mut output[output_pos..]) {
if n == 0 {
break;
}
output_pos += n;
}
output.truncate(output_pos);
{
use std::io::Write;
let mut file = std::fs::File::create("test").unwrap();
file.write_all(&output).unwrap();
}
println!("input len = {:?}, pos = {:?}", input.len(), output_pos);
let mut decomp = Vec::new();
let mut s = { Seekable::init_buf(&output).unwrap() };
for frame in 0..s.get_num_frames() {
let size = s.get_frame_decompressed_size(frame);
println!("{:?} {:?}", size, s.get_frame_decompressed_offset(frame));
let n = decomp.len();
decomp.extend(std::iter::repeat(0).take(size));
s.decompress_frame(&mut decomp[n..], frame);
}
println!("{:?}", std::str::from_utf8(&decomp).unwrap());
assert_eq!(&input[..], &decomp[..])
}
with import <nixpkgs> {};
clangStdenv.mkDerivation {
name = "zstdseek";
buildInputs = [ pkgconfig llvmPackages.libclang.lib cmake ];
LIBCLANG_PATH="${llvmPackages.libclang.lib}/lib";
}
use std::env;
use std::path::PathBuf;
fn main() {
cc::Build::new()
.file("zstd/contrib/seekable_format/zstdseek_compress.c")
.file("zstd/contrib/seekable_format/zstdseek_decompress.c")
.file("xxh64.c")
.opt_level(3)
.include("zstd/lib/common")
.include("zstd/lib")
.compile("zstdseek");
println!("cargo:rustc-link-lib=zstd");
println!("cargo:rustc-link-lib=xxhash");
let bindings = bindgen::Builder::default()
.clang_arg("-Izstd/lib/common")
.clang_arg("-Izstd/lib")
.header("zstd/contrib/seekable_format/zstd_seekable.h")
.whitelist_function("ZSTD_CStreamInSize")
.whitelist_function("ZSTD_CStreamOutSize")
.whitelist_function("ZSTD_seekable_createCStream")
.whitelist_function("ZSTD_seekable_create")
.whitelist_function("ZSTD_seekable_initCStream")
.whitelist_function("ZSTD_seekable_initBuff")
.whitelist_function("ZSTD_seekable_initFile")
.whitelist_function("ZSTD_seekable_compressStream")
.whitelist_function("ZSTD_seekable_decompress")
.whitelist_function("ZSTD_seekable_endStream")
.whitelist_function("ZSTD_seekable_free")
.whitelist_function("ZSTD_seekable_freeCStream")
.whitelist_function("ZSTD_DStreamOutSize")
.whitelist_function("ZSTD_getErrorName")
.whitelist_function("ZSTD_seekable_getNumFrames")
.whitelist_function("ZSTD_seekable_getFrameCompressedOffset")
.whitelist_function("ZSTD_seekable_getFrameCompressedSize")
.whitelist_function("ZSTD_seekable_getFrameDecompressedOffset")
.whitelist_function("ZSTD_seekable_getFrameDecompressedSize")
.whitelist_function("ZSTD_seekable_decompressFrame")
.whitelist_function("ZSTD_seekable_offsetToFrameIndex")
.whitelist_function("ZSTD_isError")
.whitelist_function("ZSTD_seekable_freeFrameLog")
.whitelist_function("ZSTD_seekable_logFrame")
.whitelist_function("ZSTD_seekable_createFrameLog")
.whitelist_function("ZSTD_seekable_writeSeekTable")
.whitelist_function("ZSTD_compress")
.whitelist_function("xxh64")
//
.whitelist_function("ZSTD_createCStream")
.whitelist_function("ZSTD_initCStream")
.whitelist_function("ZSTD_freeCStream")
.whitelist_function("ZSTD_flushStream")
.whitelist_function("ZSTD_endStream")
.whitelist_function("ZSTD_compressStream")
.whitelist_function("ZSTD_compressStream2")
.whitelist_function("ZSTD_CStreamInSize")
.whitelist_function("ZSTD_CStreamOutSize")
//
.whitelist_function("ZSTD_createDStream")
.whitelist_function("ZSTD_initDStream")
.whitelist_function("ZSTD_freeDStream")
.whitelist_function("ZSTD_decompressStream")
.whitelist_type("ZSTD_inBuffer")
.whitelist_type("ZSTD_outBuffer")
.whitelist_type("ZSTD_frameLog")
.whitelist_function("ZSTD_seekable_initAdvanced")
.whitelist_type("ZSTD_seekable_customFile")
.generate()
.expect("Unable to generate bindings");
let out_path = PathBuf::from(env::var("OUT_DIR").unwrap());
bindings
.write_to_file(out_path.join("bindings.rs"))
.expect("Couldn't write bindings!");
}
# zstd-seekable
Seekable compression standard developped by Facebook.
This format is expected, but not guaranteed, to be stable. Consult https://facebook.github.io/zstd/ for more detail.
[package]
name = "zstd-seekable"
description = "Bindings to the seekable version of ZStandard."
version = "0.1.1"
authors = ["Pierre-Étienne Meunier <pe@pijul.org>"]
edition = "2018"
license = "BSD-3-Clause"
documentation = "https://docs.rs/zstd-seekable"
repository = "https://nest.pijul.com/pijul/zstd-seekable"
homepage = "https://nest.pijul.com/pijul/zstd-seekable"
include = [
"Cargo.toml",
"build.rs",
"src/lib.rs",
"xxh64.c",
"zstd/contrib/seekable_format/zstd_seekable.h",
"zstd/contrib/seekable_format/zstdseek_compress.c",
"zstd/contrib/seekable_format/zstdseek_decompress.c",
"zstd/lib/common/mem.h",
"zstd/lib/common/xxhash.h",
"zstd/lib/zstd.h",
"zstd/lib/common/zstd_errors.h",
]
[dependencies]
libc = "0.2"
anyhow = "1.0"
thiserror = "1.0"
threadpool = "1.8"
[build-dependencies]
cc = "1.0"
bindgen = "0.55"
cmake = "0.1"