githubmirror/zstd

doc/zstd_manual.html
1 200 строк75 KB
Yann Collet
fixed a small error in decodeCorpus
08 окт 2017, 01:19
08 окт 2017, 01:197f6a783
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=ISO-8859-1">
<title>zstd 1.3.2 Manual</title>
</head>
<body>
<h1>zstd 1.3.2 Manual</h1>
<hr>
<a name="Contents"></a><h2>Contents</h2>
<ol>
<li><a href="#Chapter1">Introduction</a></li>
<li><a href="#Chapter2">Version</a></li>
<li><a href="#Chapter3">Simple API</a></li>
<li><a href="#Chapter4">Explicit memory management</a></li>
<li><a href="#Chapter5">Simple dictionary API</a></li>
<li><a href="#Chapter6">Bulk processing dictionary API</a></li>
<li><a href="#Chapter7">Streaming</a></li>
<li><a href="#Chapter8">Streaming compression - HowTo</a></li>
<li><a href="#Chapter9">Streaming decompression - HowTo</a></li>
<li><a href="#Chapter10">START OF ADVANCED AND EXPERIMENTAL FUNCTIONS</a></li>
<li><a href="#Chapter11">Advanced types</a></li>
<li><a href="#Chapter12">Frame size functions</a></li>
<li><a href="#Chapter13">Context memory usage</a></li>
<li><a href="#Chapter14">Advanced compression functions</a></li>
<li><a href="#Chapter15">Advanced decompression functions</a></li>
<li><a href="#Chapter16">Advanced streaming functions</a></li>
<li><a href="#Chapter17">Buffer-less and synchronous inner streaming functions</a></li>
<li><a href="#Chapter18">Buffer-less streaming compression (synchronous mode)</a></li>
<li><a href="#Chapter19">Buffer-less streaming decompression (synchronous mode)</a></li>
<li><a href="#Chapter20">New advanced API (experimental)</a></li>
<li><a href="#Chapter21">Block level API</a></li>
</ol>
<hr>
<a name="Chapter1"></a><h2>Introduction</h2><pre>
  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 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.
  Compression can be done in:
    - a single step (described as Simple API)
    - a single step, reusing a context (described as Explicit memory management)
    - unbounded multiple steps (described as Streaming compression)
  The compression ratio achievable on small data can be highly improved using a dictionary in:
    - a single step (described as Simple dictionary API)
    - a single step, reusing a dictionary (described as Fast dictionary API)

  Advanced experimental functions can be accessed using #define ZSTD_STATIC_LINKING_ONLY before including zstd.h.
  Advanced experimental APIs shall never be used with a dynamic library.
  They are not "stable", their definition may change in the future. Only static linking is allowed.
<BR></pre>

<a name="Chapter2"></a><h2>Version</h2><pre></pre>

<pre><b>unsigned ZSTD_versionNumber(void);   </b>/**< useful to check dll version */<b>
</b></pre><BR>
<a name="Chapter3"></a><h2>Simple API</h2><pre></pre>

<pre><b>size_t ZSTD_compress( void* dst, size_t dstCapacity,
                const void* src, size_t srcSize,
                      int compressionLevel);
</b><p>  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()). 
</p></pre><BR>

<pre><b>size_t ZSTD_decompress( void* dst, size_t dstCapacity,
                  const void* src, size_t compressedSize);
</b><p>  `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()). 
</p></pre><BR>

<pre><b>#define ZSTD_CONTENTSIZE_UNKNOWN (0ULL - 1)
#define ZSTD_CONTENTSIZE_ERROR   (0ULL - 2)
unsigned long long ZSTD_getFrameContentSize(const void *src, size_t srcSize);
</b><p>  `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 the frame in `src`, 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 done with ZSTD_compress()
   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() 
</p></pre><BR>

<pre><b>unsigned long long ZSTD_getDecompressedSize(const void* src, size_t srcSize);
</b><p>  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 in the same return value (0),
  while ZSTD_getFrameContentSize() distinguishes them.

  'src' is the start of a zstd compressed frame.
  @return : content size to be decompressed, as a 64-bits value _if known and not empty_, 0 otherwise. 
</p></pre><BR>

<h3>Helper functions</h3><pre></pre><b><pre>#define ZSTD_COMPRESSBOUND(srcSize)   ((srcSize) + ((srcSize)>>8) + (((srcSize) < 128 KB) ? ((128 KB - (srcSize)) >> 11) </b>/* 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 */<b>
size_t      ZSTD_compressBound(size_t srcSize); </b>/*!< maximum compressed size in worst case scenario */<b>
unsigned    ZSTD_isError(size_t code);          </b>/*!< tells if a `size_t` function result is an error code */<b>
const char* ZSTD_getErrorName(size_t code);     </b>/*!< provides readable string from an error code */<b>
int         ZSTD_maxCLevel(void);               </b>/*!< maximum compression level available */<b>
</pre></b><BR>
<a name="Chapter4"></a><h2>Explicit memory management</h2><pre></pre>

<h3>Compression context</h3><pre>  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.
  Use one context per thread for parallel execution in multi-threaded environments. 
</pre><b><pre>typedef struct ZSTD_CCtx_s ZSTD_CCtx;
ZSTD_CCtx* ZSTD_createCCtx(void);
size_t     ZSTD_freeCCtx(ZSTD_CCtx* cctx);
</pre></b><BR>
<pre><b>size_t ZSTD_compressCCtx(ZSTD_CCtx* ctx,
                         void* dst, size_t dstCapacity,
                   const void* src, size_t srcSize,
                         int compressionLevel);
</b><p>  Same as ZSTD_compress(), requires an allocated ZSTD_CCtx (see ZSTD_createCCtx()). 
</p></pre><BR>

<h3>Decompression context</h3><pre>  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. 
</pre><b><pre>typedef struct ZSTD_DCtx_s ZSTD_DCtx;
ZSTD_DCtx* ZSTD_createDCtx(void);
size_t     ZSTD_freeDCtx(ZSTD_DCtx* dctx);
</pre></b><BR>
<pre><b>size_t ZSTD_decompressDCtx(ZSTD_DCtx* ctx,
                           void* dst, size_t dstCapacity,
                     const void* src, size_t srcSize);
</b><p>  Same as ZSTD_decompress(), requires an allocated ZSTD_DCtx (see ZSTD_createDCtx()) 
</p></pre><BR>

<a name="Chapter5"></a><h2>Simple dictionary API</h2><pre></pre>

<pre><b>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);
</b><p>  Compression using a predefined Dictionary (see dictBuilder/zdict.h).
  Note : This function loads the dictionary, resulting in significant startup delay.
  Note : When `dict == NULL || dictSize < 8` no dictionary is used. 
</p></pre><BR>

<pre><b>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);
</b><p>  Decompression using a predefined Dictionary (see dictBuilder/zdict.h).
  Dictionary must be identical to the one used during compression.
  Note : This function loads the dictionary, resulting in significant startup delay.
  Note : When `dict == NULL || dictSize < 8` no dictionary is used. 
</p></pre><BR>

<a name="Chapter6"></a><h2>Bulk processing dictionary API</h2><pre></pre>

<pre><b>ZSTD_CDict* ZSTD_createCDict(const void* dictBuffer, size_t dictSize,
                             int compressionLevel);
</b><p>  When compressing multiple messages / blocks with the same dictionary, it's recommended to load it just once.
  ZSTD_createCDict() will create a digested dictionary, ready to start future compression operations without startup delay.
  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, since its content is copied within CDict 
</p></pre><BR>

<pre><b>size_t      ZSTD_freeCDict(ZSTD_CDict* CDict);
</b><p>  Function frees memory allocated by ZSTD_createCDict(). 
</p></pre><BR>

<pre><b>size_t ZSTD_compress_usingCDict(ZSTD_CCtx* cctx,
                                void* dst, size_t dstCapacity,
                          const void* src, size_t srcSize,
                          const ZSTD_CDict* cdict);
</b><p>  Compression using a digested Dictionary.
  Faster startup than ZSTD_compress_usingDict(), recommended when same dictionary is used multiple times.
  Note that compression level is decided during dictionary creation.
  Frame parameters are hardcoded (dictID=yes, contentSize=yes, checksum=no) 
</p></pre><BR>

<pre><b>ZSTD_DDict* ZSTD_createDDict(const void* dictBuffer, size_t dictSize);
</b><p>  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 
</p></pre><BR>

<pre><b>size_t      ZSTD_freeDDict(ZSTD_DDict* ddict);
</b><p>  Function frees memory allocated with ZSTD_createDDict() 
</p></pre><BR>

<pre><b>size_t ZSTD_decompress_usingDDict(ZSTD_DCtx* dctx,
                                  void* dst, size_t dstCapacity,
                            const void* src, size_t srcSize,
                            const ZSTD_DDict* ddict);
</b><p>  Decompression using a digested Dictionary.
  Faster startup than ZSTD_decompress_usingDict(), recommended when same dictionary is used multiple times. 
</p></pre><BR>

<a name="Chapter7"></a><h2>Streaming</h2><pre></pre>

<pre><b>typedef struct ZSTD_inBuffer_s {
  const void* src;    </b>/**< start of input buffer */<b>
  size_t size;        </b>/**< size of input buffer */<b>
  size_t pos;         </b>/**< position where reading stopped. Will be updated. Necessarily 0 <= pos <= size */<b>
} ZSTD_inBuffer;
</b></pre><BR>
<pre><b>typedef struct ZSTD_outBuffer_s {
  void*  dst;         </b>/**< start of output buffer */<b>
  size_t size;        </b>/**< size of output buffer */<b>
  size_t pos;         </b>/**< position where writing stopped. Will be updated. Necessarily 0 <= pos <= size */<b>
} ZSTD_outBuffer;
</b></pre><BR>
<a name="Chapter8"></a><h2>Streaming compression - HowTo</h2><pre>
  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 in situations where many streaming operations will be achieved consecutively,
  since it will play nicer with system's memory, by re-using already allocated memory.
  Use one separate ZSTD_CStream per thread for parallel execution.

  Start a new compression by initializing ZSTD_CStream.
  Use ZSTD_initCStream() to start a new compression operation.
  Use ZSTD_initCStream_usingDict() or ZSTD_initCStream_usingCDict() for a compression which requires a dictionary (experimental section)

  Use ZSTD_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.
            Note 2 : size hint is guaranteed to be <= ZSTD_CStreamInSize()

  At any moment, it's possible to flush whatever data remains within internal buffer, using ZSTD_flushStream().
  `output->pos` will be updated.
  Note that some content might still be left within internal buffer if `output->size` is too small.
  @return : nb of bytes still present within internal buffer (0 if it's empty)
            or an error code, which can be tested using ZSTD_isError().

  ZSTD_endStream() 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.
  ZSTD_endStream() may not be able to flush full data if `output->size` is too small.
  In which case, call again ZSTD_endStream() to complete the flush.
  @return : 0 if frame fully completed and fully flushed,
             or >0 if some data is still present within internal buffer
                  (value is minimum size estimation for remaining data to flush, but it could be more)
            or an error code, which can be tested using ZSTD_isError().

 
<BR></pre>

<pre><b>typedef ZSTD_CCtx ZSTD_CStream;  </b>/**< CCtx and CStream are now effectively same object (>= v1.3.0) */<b>
</b></pre><BR>
<h3>ZSTD_CStream management functions</h3><pre></pre><b><pre>ZSTD_CStream* ZSTD_createCStream(void);
size_t ZSTD_freeCStream(ZSTD_CStream* zcs);
</pre></b><BR>
<h3>Streaming compression functions</h3><pre></pre><b><pre>size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel);
size_t ZSTD_compressStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input);
size_t ZSTD_flushStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output);
size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output);
</pre></b><BR>
<pre><b>size_t ZSTD_CStreamInSize(void);    </b>/**< recommended size for input buffer */<b>
</b></pre><BR>
<pre><b>size_t ZSTD_CStreamOutSize(void);   </b>/**< recommended size for output buffer. Guarantee to successfully flush at least one complete compressed block in all circumstances. */<b>
</b></pre><BR>
<a name="Chapter9"></a><h2>Streaming decompression - HowTo</h2><pre>
  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,
   or ZSTD_initDStream_usingDict() if decompression requires a dictionary.
   @return : recommended first input size

  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.
  If `output.pos < output.size`, decoder has flushed everything it could.
  @return : 0 when a frame is completely decoded and fully flushed,
            an error code, which can be tested using ZSTD_isError(),
            any other value > 0, which means there is still some decoding to do to complete current frame.
            The return value is a suggested next input size (a hint to improve latency) that will never load more than the current frame.
 
<BR></pre>

<pre><b>typedef ZSTD_DCtx ZSTD_DStream;  </b>/**< DCtx and DStream are now effectively same object (>= v1.3.0) */<b>
</b></pre><BR>
<h3>ZSTD_DStream management functions</h3><pre></pre><b><pre>ZSTD_DStream* ZSTD_createDStream(void);
size_t ZSTD_freeDStream(ZSTD_DStream* zds);
</pre></b><BR>
<h3>Streaming decompression functions</h3><pre></pre><b><pre>size_t ZSTD_initDStream(ZSTD_DStream* zds);
size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inBuffer* input);
</pre></b><BR>
<pre><b>size_t ZSTD_DStreamInSize(void);    </b>/*!< recommended size for input buffer */<b>
</b></pre><BR>
<pre><b>size_t ZSTD_DStreamOutSize(void);   </b>/*!< recommended size for output buffer. Guarantee to successfully flush at least one complete block in all circumstances. */<b>
</b></pre><BR>
<a name="Chapter10"></a><h2>START OF ADVANCED AND EXPERIMENTAL FUNCTIONS</h2><pre> The definitions in this section are considered experimental.
 They should never be used with a dynamic library, as prototypes may change in the future.
 They are provided for advanced scenarios.
 Use them only in association with static linking.
 
<BR></pre>

<a name="Chapter11"></a><h2>Advanced types</h2><pre></pre>

<pre><b>typedef enum { ZSTD_fast=1, ZSTD_dfast, ZSTD_greedy, ZSTD_lazy, ZSTD_lazy2,
               ZSTD_btlazy2, ZSTD_btopt, ZSTD_btultra } ZSTD_strategy;   </b>/* from faster to stronger */<b>
</b></pre><BR>
<pre><b>typedef struct {
    unsigned windowLog;      </b>/**< largest match distance : larger == more compression, more memory needed during decompression */<b>
    unsigned chainLog;       </b>/**< fully searched segment : larger == more compression, slower, more memory (useless for fast) */<b>
    unsigned hashLog;        </b>/**< dispatch table : larger == faster, more memory */<b>
    unsigned searchLog;      </b>/**< nb of searches : larger == more compression, slower */<b>
    unsigned searchLength;   </b>/**< match length searched : larger == faster decompression, sometimes less compression */<b>
    unsigned targetLength;   </b>/**< acceptable match size for optimal parser (only) : larger == more compression, slower */<b>
    ZSTD_strategy strategy;
} ZSTD_compressionParameters;
</b></pre><BR>
<pre><b>typedef struct {
    unsigned contentSizeFlag; </b>/**< 1: content size will be in frame header (when known) */<b>
    unsigned checksumFlag;    </b>/**< 1: generate a 32-bits checksum at end of frame, for error detection */<b>
    unsigned noDictIDFlag;    </b>/**< 1: no dictID will be saved into frame header (if dictionary compression) */<b>
} ZSTD_frameParameters;
</b></pre><BR>
<pre><b>typedef struct {
    ZSTD_compressionParameters cParams;
    ZSTD_frameParameters fParams;
} ZSTD_parameters;
</b></pre><BR>
<h3>Custom memory allocation functions</h3><pre></pre><b><pre>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;
</b>/* use this constant to defer to stdlib's functions */<b>
static const ZSTD_customMem ZSTD_defaultCMem = { NULL, NULL, NULL };
</pre></b><BR>
<a name="Chapter12"></a><h2>Frame size functions</h2><pre></pre>

<pre><b>size_t ZSTD_findFrameCompressedSize(const void* src, size_t srcSize);
</b><p>  `src` should point to the start of a ZSTD encoded frame or skippable frame
  `srcSize` must be at least as large as the frame
  @return : the compressed size of the first frame starting at `src`,
            suitable to pass to `ZSTD_decompress` or similar,
            or an error code if input is invalid 
</p></pre><BR>

<pre><b>unsigned long long ZSTD_findDecompressedSize(const void* src, size_t srcSize);
</b><p>  `src` should point 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 exactly at `srcSize` bytes after `src`)
  @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. 
</p></pre><BR>

<pre><b>size_t ZSTD_frameHeaderSize(const void* src, size_t srcSize);
</b><p>   `src` should point to the start of a ZSTD frame
   `srcSize` must be >= ZSTD_frameHeaderSize_prefix.
   @return : size of the Frame Header 
</p></pre><BR>

<a name="Chapter13"></a><h2>Context memory usage</h2><pre></pre>

<pre><b>size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx);
size_t ZSTD_sizeof_DCtx(const ZSTD_DCtx* dctx);
size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs);
size_t ZSTD_sizeof_DStream(const ZSTD_DStream* zds);
size_t ZSTD_sizeof_CDict(const ZSTD_CDict* cdict);
size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict);
</b><p>  These functions give the current memory usage of selected object.
  Object memory usage can evolve when re-used multiple times. 
</p></pre><BR>

<pre><b>size_t ZSTD_estimateCCtxSize(int compressionLevel);
size_t ZSTD_estimateCCtxSize_usingCParams(ZSTD_compressionParameters cParams);
size_t ZSTD_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params);
size_t ZSTD_estimateDCtxSize(void);
</b><p>  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.
  It will also consider src size to be arbitrarily "large", which is 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_CCtxParam_setParameter(). Only single-threaded compression is supported. This function will return an error code if ZSTD_p_nbThreads is > 1.
  Note : CCtx estimation is only correct for single-threaded compression 
</p></pre><BR>

<pre><b>size_t ZSTD_estimateCStreamSize(int compressionLevel);
size_t ZSTD_estimateCStreamSize_usingCParams(ZSTD_compressionParameters cParams);
size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params);
size_t ZSTD_estimateDStreamSize(size_t windowSize);
size_t ZSTD_estimateDStreamSize_fromFrame(const void* src, size_t srcSize);
</b><p>  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_CCtxParam_setParameter(). Only single-threaded compression is supported. This function will return an error code if ZSTD_p_nbThreads is set to a value > 1.
  Note : CStream 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 
</p></pre><BR>

<pre><b>typedef enum {
    ZSTD_dlm_byCopy = 0,     </b>/**< Copy dictionary content internally */<b>
    ZSTD_dlm_byRef,          </b>/**< Reference dictionary content -- the dictionary buffer must outlive its users. */<b>
} ZSTD_dictLoadMethod_e;
</b></pre><BR>
<pre><b>size_t ZSTD_estimateCDictSize(size_t dictSize, int compressionLevel);
size_t ZSTD_estimateCDictSize_advanced(size_t dictSize, ZSTD_compressionParameters cParams, ZSTD_dictLoadMethod_e dictLoadMethod);
size_t ZSTD_estimateDDictSize(size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod);
</b><p>  ZSTD_estimateCDictSize() will bet that src size is relatively "small", and content is copied, like ZSTD_createCDict().
  ZSTD_estimateCStreamSize_advanced_usingCParams() makes it possible to control precisely compression parameters, like ZSTD_createCDict_advanced().
  Note : dictionary created by reference using ZSTD_dlm_byRef are smaller
 
</p></pre><BR>

<a name="Chapter14"></a><h2>Advanced compression functions</h2><pre></pre>

<pre><b>ZSTD_CCtx* ZSTD_createCCtx_advanced(ZSTD_customMem customMem);
</b><p>  Create a ZSTD compression context using external alloc and free functions 
</p></pre><BR>

<pre><b>ZSTD_CCtx* ZSTD_initStaticCCtx(void* workspace, size_t workspaceSize);
</b><p>  workspace: The memory area to emplace the context into.
             Provided pointer must 8-bytes aligned.
             It must outlive context usage.
  workspaceSize: Use ZSTD_estimateCCtxSize() or ZSTD_estimateCStreamSize()
                 to determine how large workspace must be to support scenario.
 @return : pointer to ZSTD_CCtx*, or NULL if error (size too small)
  Note : zstd will never resize nor malloc() when using a static cctx.
         If it needs more memory than available, it will simply error out.
  Note 2 : there is no corresponding "free" function.
           Since workspace was allocated externally, it must be freed externally too.
  Limitation 1 : currently not compatible with internal CDict creation, such as
                 ZSTD_CCtx_loadDictionary() or ZSTD_initCStream_usingDict().
  Limitation 2 : currently not compatible with multi-threading
 
</p></pre><BR>

<pre><b>ZSTD_CDict* ZSTD_createCDict_byReference(const void* dictBuffer, size_t dictSize, int compressionLevel);
</b><p>  Create a digested dictionary for compression
  Dictionary content is simply referenced, and therefore stays in dictBuffer.
  It is important that dictBuffer outlives CDict, it must remain read accessible throughout the lifetime of CDict 
</p></pre><BR>

<pre><b>typedef enum { ZSTD_dm_auto=0,        </b>/* dictionary is "full" if it starts with ZSTD_MAGIC_DICTIONARY, otherwise it is "rawContent" */<b>
               ZSTD_dm_rawContent,    </b>/* ensures dictionary is always loaded as rawContent, even if it starts with ZSTD_MAGIC_DICTIONARY */<b>
               ZSTD_dm_fullDict       </b>/* refuses to load a dictionary if it does not respect Zstandard's specification */<b>
} ZSTD_dictMode_e;
</b></pre><BR>
<pre><b>ZSTD_CDict* ZSTD_createCDict_advanced(const void* dict, size_t dictSize,
                                      ZSTD_dictLoadMethod_e dictLoadMethod,
                                      ZSTD_dictMode_e dictMode,
                                      ZSTD_compressionParameters cParams,
                                      ZSTD_customMem customMem);
</b><p>  Create a ZSTD_CDict using external alloc and free, and customized compression parameters 
</p></pre><BR>

<pre><b>ZSTD_CDict* ZSTD_initStaticCDict(
                void* workspace, size_t workspaceSize,
          const void* dict, size_t dictSize,
                ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictMode_e dictMode,
                ZSTD_compressionParameters cParams);
</b><p>  Generate a digested dictionary in provided memory area.
  workspace: The memory area to emplace the dictionary into.
             Provided pointer must 8-bytes aligned.
             It must outlive dictionary usage.
  workspaceSize: Use ZSTD_estimateCDictSize()
                 to determine how large workspace must be.
  cParams : use ZSTD_getCParams() to transform a compression level
            into its relevants cParams.
 @return : pointer to ZSTD_CDict*, or NULL if error (size too small)
  Note : there is no corresponding "free" function.
         Since workspace was allocated externally, it must be freed externally.
 
</p></pre><BR>

<pre><b>ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize);
</b><p>   @return ZSTD_compressionParameters structure for a selected compression level and estimated srcSize.
   `estimatedSrcSize` value is optional, select 0 if not known 
</p></pre><BR>

<pre><b>ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize);
</b><p>   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 (0) 
</p></pre><BR>

<pre><b>size_t ZSTD_checkCParams(ZSTD_compressionParameters params);
</b><p>   Ensure param values remain within authorized range 
</p></pre><BR>

<pre><b>ZSTD_compressionParameters ZSTD_adjustCParams(ZSTD_compressionParameters cPar, unsigned long long srcSize, size_t dictSize);
</b><p>  optimize params for a given `srcSize` and `dictSize`.
  both values are optional, select `0` if unknown. 
</p></pre><BR>

<pre><b>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);
</b><p>   Same as ZSTD_compress_usingDict(), with fine-tune control over each compression parameter 
</p></pre><BR>

<pre><b>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);
</b><p>   Same as ZSTD_compress_usingCDict(), with fine-tune control over frame parameters 
</p></pre><BR>

<a name="Chapter15"></a><h2>Advanced decompression functions</h2><pre></pre>

<pre><b>unsigned ZSTD_isFrame(const void* buffer, size_t size);
</b><p>  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. 
</p></pre><BR>

<pre><b>ZSTD_DCtx* ZSTD_createDCtx_advanced(ZSTD_customMem customMem);
</b><p>  Create a ZSTD decompression context using external alloc and free functions 
</p></pre><BR>

<pre><b>ZSTD_DCtx* ZSTD_initStaticDCtx(void* workspace, size_t workspaceSize);
</b><p>  workspace: The memory area to emplace the context into.
             Provided pointer must 8-bytes aligned.
             It must outlive context usage.
  workspaceSize: Use ZSTD_estimateDCtxSize() or ZSTD_estimateDStreamSize()
                 to determine how large workspace must be to support scenario.
 @return : pointer to ZSTD_DCtx*, or NULL if error (size too small)
  Note : zstd will never resize nor malloc() when using a static dctx.
         If it needs more memory than available, it will simply error out.
  Note 2 : static dctx is incompatible with legacy support
  Note 3 : there is no corresponding "free" function.
           Since workspace was allocated externally, it must be freed externally.
  Limitation : currently not compatible with internal DDict creation,
               such as ZSTD_initDStream_usingDict().
 
</p></pre><BR>

<pre><b>ZSTD_DDict* ZSTD_createDDict_byReference(const void* dictBuffer, size_t dictSize);
</b><p>  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 
</p></pre><BR>

<pre><b>ZSTD_DDict* ZSTD_createDDict_advanced(const void* dict, size_t dictSize,
                                      ZSTD_dictLoadMethod_e dictLoadMethod,
                                      ZSTD_customMem customMem);
</b><p>  Create a ZSTD_DDict using external alloc and free, optionally by reference 
</p></pre><BR>

<pre><b>ZSTD_DDict* ZSTD_initStaticDDict(void* workspace, size_t workspaceSize,
                                 const void* dict, size_t dictSize,
                                 ZSTD_dictLoadMethod_e dictLoadMethod);
</b><p>  Generate a digested dictionary in provided memory area.
  workspace: The memory area to emplace the dictionary into.
             Provided pointer must 8-bytes aligned.
             It must outlive dictionary usage.
  workspaceSize: Use ZSTD_estimateDDictSize()
                 to determine how large workspace must be.
 @return : pointer to ZSTD_DDict*, or NULL if error (size too small)
  Note : there is no corresponding "free" function.
         Since workspace was allocated externally, it must be freed externally.
 
</p></pre><BR>

<pre><b>unsigned ZSTD_getDictID_fromDict(const void* dict, size_t dictSize);
</b><p>  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. 
</p></pre><BR>

<pre><b>unsigned ZSTD_getDictID_fromDDict(const ZSTD_DDict* ddict);
</b><p>  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. 
</p></pre><BR>

<pre><b>unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize);
</b><p>  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. 
</p></pre><BR>

<a name="Chapter16"></a><h2>Advanced streaming functions</h2><pre></pre>

<h3>Advanced Streaming compression functions</h3><pre></pre><b><pre>ZSTD_CStream* ZSTD_createCStream_advanced(ZSTD_customMem customMem);
ZSTD_CStream* ZSTD_initStaticCStream(void* workspace, size_t workspaceSize);    </b>/**< same as ZSTD_initStaticCCtx() */<b>
size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, int compressionLevel, unsigned long long pledgedSrcSize);   </b>/**< pledgedSrcSize must be correct, a size of 0 means unknown.  for a frame size of 0 use initCStream_advanced */<b>
size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, const void* dict, size_t dictSize, int compressionLevel); </b>/**< 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_dm_auto (treated as a full zstd dictionary if it begins with ZSTD_MAGIC_DICTIONARY, else as raw content) and ZSTD_dlm_byCopy.*/<b>
size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs, const void* dict, size_t dictSize,
                                             ZSTD_parameters params, unsigned long long pledgedSrcSize);  </b>/**< pledgedSrcSize is optional and can be 0 (meaning unknown). note: if the contentSizeFlag is set, pledgedSrcSize == 0 means the source size is actually 0. dict is loaded with ZSTD_dm_auto and ZSTD_dlm_byCopy. */<b>
size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict);  </b>/**< note : cdict will just be referenced, and must outlive compression session */<b>
size_t ZSTD_initCStream_usingCDict_advanced(ZSTD_CStream* zcs, const ZSTD_CDict* cdict, ZSTD_frameParameters fParams, unsigned long long pledgedSrcSize);  </b>/**< same as ZSTD_initCStream_usingCDict(), with control over frame parameters */<b>
</pre></b><BR>
<pre><b>size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize);
</b><p>  start a new compression job, using same parameters from previous job.
  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().
  pledgedSrcSize==0 means "srcSize unknown".
  If pledgedSrcSize > 0, its value must be correct, as it will be written in header, and controlled at the end.
  @return : 0, or an error code (which can be tested using ZSTD_isError()) 
</p></pre><BR>

<h3>Advanced Streaming decompression functions</h3><pre></pre><b><pre>ZSTD_DStream* ZSTD_createDStream_advanced(ZSTD_customMem customMem);
ZSTD_DStream* ZSTD_initStaticDStream(void* workspace, size_t workspaceSize);    </b>/**< same as ZSTD_initStaticDCtx() */<b>
typedef enum { DStream_p_maxWindowSize } ZSTD_DStreamParameter_e;
size_t ZSTD_setDStreamParameter(ZSTD_DStream* zds, ZSTD_DStreamParameter_e paramType, unsigned paramValue);   </b>/* obsolete : this API will be removed in a future version */<b>
size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize); </b>/**< note: no dictionary will be used if dict == NULL or dictSize < 8 */<b>
size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* zds, const ZSTD_DDict* ddict);  </b>/**< note : ddict is referenced, it must outlive decompression session */<b>
size_t ZSTD_resetDStream(ZSTD_DStream* zds);  </b>/**< re-use decompression parameters from previous init; saves dictionary loading */<b>
</pre></b><BR>
<a name="Chapter17"></a><h2>Buffer-less and synchronous inner streaming functions</h2><pre>
  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.
 
<BR></pre>

<a name="Chapter18"></a><h2>Buffer-less streaming compression (synchronous mode)</h2><pre>
  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.
<BR></pre>

<h3>Buffer-less streaming compression functions</h3><pre></pre><b><pre>size_t ZSTD_compressBegin(ZSTD_CCtx* cctx, int compressionLevel);
size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, int compressionLevel);
size_t ZSTD_compressBegin_advanced(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, ZSTD_parameters params, unsigned long long pledgedSrcSize); </b>/**< pledgedSrcSize is optional and can be 0 (meaning unknown). note: if the contentSizeFlag is set, pledgedSrcSize == 0 means the source size is actually 0 */<b>
size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict); </b>/**< note: fails if cdict==NULL */<b>
size_t ZSTD_compressBegin_usingCDict_advanced(ZSTD_CCtx* const cctx, const ZSTD_CDict* const cdict, ZSTD_frameParameters const fParams, unsigned long long const pledgedSrcSize);   </b>/* compression parameters are already set within cdict. pledgedSrcSize=0 means null-size */<b>
size_t ZSTD_copyCCtx(ZSTD_CCtx* cctx, const ZSTD_CCtx* preparedCCtx, unsigned long long pledgedSrcSize); </b>/**<  note: if pledgedSrcSize can be 0, indicating unknown size.  if it is non-zero, it must be accurate.  for 0 size frames, use compressBegin_advanced */<b>
</pre></b><BR>
<a name="Chapter19"></a><h2>Buffer-less streaming decompression (synchronous mode)</h2><pre>
  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.
<BR></pre>

<h3>Buffer-less streaming decompression functions</h3><pre></pre><b><pre>typedef enum { ZSTD_frame, ZSTD_skippableFrame } ZSTD_frameType_e;
typedef struct {
    unsigned long long frameContentSize; </b>/* if == ZSTD_CONTENTSIZE_UNKNOWN, it means this field is not available. 0 means "empty" */<b>
    unsigned long long windowSize;       </b>/* can be very large, up to <= frameContentSize */<b>
    unsigned blockSizeMax;
    ZSTD_frameType_e frameType;          </b>/* if == ZSTD_skippableFrame, frameContentSize is the size of skippable content */<b>
    unsigned headerSize;
    unsigned dictID;
    unsigned checksumFlag;
} ZSTD_frameHeader;
size_t ZSTD_getFrameHeader(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize);   </b>/**< doesn't consume input */<b>
size_t ZSTD_decodingBufferSize_min(unsigned long long windowSize, unsigned long long frameContentSize);  </b>/**< when frame content size is not known, pass in frameContentSize == ZSTD_CONTENTSIZE_UNKNOWN */<b>
</pre></b><BR>
<pre><b>typedef enum { ZSTDnit_frameHeader, ZSTDnit_blockHeader, ZSTDnit_block, ZSTDnit_lastBlock, ZSTDnit_checksum, ZSTDnit_skippableFrame } ZSTD_nextInputType_e;
</b></pre><BR>
<a name="Chapter20"></a><h2>New advanced API (experimental)</h2><pre></pre>

<pre><b>typedef enum {
    </b>/* Question : should we have a format ZSTD_f_auto ?<b>
     * For the time being, it would mean exactly the same as ZSTD_f_zstd1.
     * But, in the future, should several formats be supported,
     * on the compression side, it would mean "default format".
     * On the decompression side, it would mean "multi format",
     * and ZSTD_f_zstd1 could be reserved to mean "accept *only* zstd frames".
     * Since meaning is a little different, another option could be to define different enums for compression and decompression.
     * This question could be kept for later, when there are actually multiple formats to support,
     * but there is also the question of pinning enum values, and pinning value `0` is especially important */
    ZSTD_f_zstd1 = 0,        </b>/* zstd frame format, specified in zstd_compression_format.md (default) */<b>
    ZSTD_f_zstd1_magicless,  </b>/* Variant of zstd frame format, without initial 4-bytes magic number.<b>
                              * Useful to save 4 bytes per generated frame.
                              * Decoder cannot recognise automatically this format, requiring instructions. */
} ZSTD_format_e;
</b></pre><BR>
<pre><b>typedef enum {
    </b>/* compression format */<b>
    ZSTD_p_format = 10,      </b>/* See ZSTD_format_e enum definition.<b>
                              * Cast selected format as unsigned for ZSTD_CCtx_setParameter() compatibility. */

    </b>/* compression parameters */<b>
    ZSTD_p_compressionLevel=100, </b>/* Update all compression parameters according to pre-defined cLevel table<b>
                              * Default level is ZSTD_CLEVEL_DEFAULT==3.
                              * Special: value 0 means "do not change cLevel". */
    ZSTD_p_windowLog,        </b>/* Maximum allowed back-reference distance, expressed as power of 2.<b>
                              * Must be clamped between ZSTD_WINDOWLOG_MIN and ZSTD_WINDOWLOG_MAX.
                              * Special: value 0 means "do not change windowLog".
                              * Note: Using a window size greater than ZSTD_MAXWINDOWSIZE_DEFAULT (default: 2^27)
                              * requires setting the maximum window size at least as large during decompression. */
    ZSTD_p_hashLog,          </b>/* Size of the probe table, as a power of 2.<b>
                              * Resulting table size 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 "do not change hashLog". */
    ZSTD_p_chainLog,         </b>/* Size of the full-search table, as a power of 2.<b>
                              * Resulting table size is (1 << (chainLog+2)).
                              * Larger tables result in better and slower compression.
                              * This parameter is useless when using "fast" strategy.
                              * Special: value 0 means "do not change chainLog". */
    ZSTD_p_searchLog,        </b>/* Number of search attempts, as a power of 2.<b>
                              * More attempts result in better and slower compression.
                              * This parameter is useless when using "fast" and "dFast" strategies.
                              * Special: value 0 means "do not change searchLog". */
    ZSTD_p_minMatch,         </b>/* Minimum size of searched matches (note : repCode matches can be smaller).<b>
                              * Larger values make faster compression and decompression, but decrease ratio.
                              * Must be clamped between ZSTD_SEARCHLENGTH_MIN and ZSTD_SEARCHLENGTH_MAX.
                              * Note that currently, for all strategies < btopt, effective minimum is 4.
                              * Note that currently, for all strategies > fast, effective maximum is 6.
                              * Special: value 0 means "do not change minMatchLength". */
    ZSTD_p_targetLength,     </b>/* Only useful for strategies >= btopt.<b>
                              * Length of Match considered "good enough" to stop search.
                              * Larger values make compression stronger and slower.
                              * Special: value 0 means "do not change targetLength". */
    ZSTD_p_compressionStrategy, </b>/* See ZSTD_strategy enum definition.<b>
                              * Cast selected strategy as unsigned for ZSTD_CCtx_setParameter() compatibility.
                              * The higher the value of selected strategy, the more complex it is,
                              * resulting in stronger and slower compression.
                              * Special: value 0 means "do not change strategy". */

    </b>/* frame parameters */<b>
    ZSTD_p_contentSizeFlag=200, </b>/* Content size is written into frame header _whenever known_ (default:1)<b>
                              * note that content size must be known at the beginning,
                              * it is sent using ZSTD_CCtx_setPledgedSrcSize() */
    ZSTD_p_checksumFlag,     </b>/* A 32-bits checksum of content is written at end of frame (default:0) */<b>
    ZSTD_p_dictIDFlag,       </b>/* When applicable, dictID of dictionary is provided in frame header (default:1) */<b>

    </b>/* multi-threading parameters */<b>
    ZSTD_p_nbThreads=400,    </b>/* Select how many threads a compression job can spawn (default:1)<b>
                              * More threads improve speed, but also increase memory usage.
                              * Can only receive a value > 1 if ZSTD_MULTITHREAD is enabled.
                              * Special: value 0 means "do not change nbThreads" */
    ZSTD_p_jobSize,          </b>/* Size of a compression job. Each compression job is completed in parallel.<b>
                              * 0 means default, which is dynamically determined based on compression parameters.
                              * Job size must be a minimum of overlapSize, or 1 KB, whichever is largest
                              * The minimum size is automatically and transparently enforced */
    ZSTD_p_overlapSizeLog,   </b>/* Size of previous input reloaded at the beginning of each job.<b>
                              * 0 => no overlap, 6(default) => use 1/8th of windowSize, >=9 => use full windowSize */

    </b>/* advanced parameters - may not remain available after API update */<b>
    ZSTD_p_forceMaxWindow=1100, </b>/* Force back-reference distances to remain < windowSize,<b>
                              * even when referencing into Dictionary content (default:0) */
    ZSTD_p_enableLongDistanceMatching=1200,  </b>/* Enable long distance matching.<b>
                                         * This parameter is designed to improve the compression
                                         * ratio for large inputs with long distance matches.
                                         * This increases the memory usage as well as window size.
                                         * Note: setting this parameter sets all the LDM parameters
                                         * as well as ZSTD_p_windowLog. It should be set after
                                         * ZSTD_p_compressionLevel and before ZSTD_p_windowLog and
                                         * other LDM parameters. Setting the compression level
                                         * after this parameter overrides the window log, though LDM
                                         * will remain enabled until explicitly disabled. */
    ZSTD_p_ldmHashLog,   </b>/* Size of the table for long distance matching, as a power of 2.<b>
                          * 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). */
    ZSTD_p_ldmMinMatch,  </b>/* Minimum size of searched matches for long distance matcher.<b>
                          * Larger/too small values usually decrease compression ratio.
                          * Must be clamped between ZSTD_LDM_MINMATCH_MIN
                          * and ZSTD_LDM_MINMATCH_MAX (default: 64). */
    ZSTD_p_ldmBucketSizeLog,  </b>/* Log size of each bucket in the LDM hash table for collision resolution.<b>
                               * Larger values usually improve collision resolution but may decrease
                               * compression speed.
                               * The maximum value is ZSTD_LDM_BUCKETSIZELOG_MAX (default: 3). */
    ZSTD_p_ldmHashEveryLog,  </b>/* Frequency of inserting/looking up entries in the LDM hash table.<b>
                              * The default is MAX(0, (windowLog - ldmHashLog)) to
                              * optimize hash table usage.
                              * Larger values improve compression speed. Deviating far from the
                              * default value will likely result in a decrease in compression ratio.
                              * Must be clamped between 0 and ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN. */

} ZSTD_cParameter;
</b></pre><BR>
<pre><b>size_t ZSTD_CCtx_setParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, unsigned value);
</b><p>  Set one compression parameter, selected by enum ZSTD_cParameter.
  Note : when `value` is an enum, cast it to unsigned for proper type checking.
  @result : 0, or an error code (which can be tested with ZSTD_isError()). 
</p></pre><BR>

<pre><b>size_t ZSTD_CCtx_setPledgedSrcSize(ZSTD_CCtx* cctx, unsigned long long pledgedSrcSize);
</b><p>  Total input data size to be compressed as a single frame.
  This value will be controlled at the end, and result in error if not respected.
 @result : 0, or an error code (which can be tested with ZSTD_isError()).
  Note 1 : 0 means zero, empty.
           In order to mean "unknown content size", pass constant ZSTD_CONTENTSIZE_UNKNOWN.
           Note that ZSTD_CONTENTSIZE_UNKNOWN is default value for new compression jobs.
  Note 2 : If all data is provided and consumed in a single round,
           this value is overriden by srcSize instead. 
</p></pre><BR>

<pre><b>size_t ZSTD_CCtx_loadDictionary(ZSTD_CCtx* cctx, const void* dict, size_t dictSize);
size_t ZSTD_CCtx_loadDictionary_byReference(ZSTD_CCtx* cctx, const void* dict, size_t dictSize);
size_t ZSTD_CCtx_loadDictionary_advanced(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictMode_e dictMode);
</b><p>  Create an internal CDict from dict buffer.
  Decompression will have to use same buffer.
 @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 : `dict` content will be copied internally. Use
            ZSTD_CCtx_loadDictionary_byReference() to reference dictionary
            content instead. The dictionary buffer must then outlive its
            users.
  Note 2 : Loading a dictionary involves building tables, which are dependent on compression parameters.
           For this reason, compression parameters cannot be changed anymore after loading a dictionary.
           It's also a CPU-heavy operation, with non-negligible impact on latency.
  Note 3 : Dictionary will be used for all future compression jobs.
           To return to "no-dictionary" situation, load a NULL dictionary
  Note 5 : Use ZSTD_CCtx_loadDictionary_advanced() to select how dictionary
           content will be interpreted.
 
</p></pre><BR>

<pre><b>size_t ZSTD_CCtx_refCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict);
</b><p>  Reference a prepared dictionary, to be used for all next compression jobs.
  Note that compression parameters are enforced from within CDict,
  and supercede any compression parameter previously set within CCtx.
  The dictionary will remain valid for future compression jobs using same CCtx.
 @result : 0, or an error code (which can be tested with ZSTD_isError()).
  Special : adding a NULL CDict means "return to no-dictionary mode".
  Note 1 : Currently, only one dictionary can be managed.
           Adding a new dictionary effectively "discards" any previous one.
  Note 2 : CDict is just referenced, its lifetime must outlive CCtx.
 
</p></pre><BR>

<pre><b>size_t ZSTD_CCtx_refPrefix(ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize);
size_t ZSTD_CCtx_refPrefix_advanced(ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize, ZSTD_dictMode_e dictMode);
</b><p>  Reference a prefix (single-usage dictionary) for next compression job.
  Decompression need same prefix to properly regenerate data.
  Prefix is **only used once**. Tables are discarded at end of compression job.
  Subsequent compression jobs will be done without prefix (if none is explicitly referenced).
  If there is a need to use same prefix multiple times, consider embedding it into a ZSTD_CDict instead.
 @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 job.
  Note 2 : Referencing a prefix involves building tables, which are dependent on compression parameters.
           It's a CPU-heavy operation, with non-negligible impact on latency.
  Note 3 : By default, the prefix is treated as raw content
           (ZSTD_dm_rawContent). Use ZSTD_CCtx_refPrefix_advanced() to alter
           dictMode. 
</p></pre><BR>

<pre><b>typedef enum {
    ZSTD_e_continue=0, </b>/* collect more data, encoder transparently decides when to output result, for optimal conditions */<b>
    ZSTD_e_flush,      </b>/* flush any data provided so far - frame will continue, future data can still reference previous data for better compression */<b>
    ZSTD_e_end         </b>/* flush any remaining data and close current frame. Any additional data starts a new frame. */<b>
} ZSTD_EndDirective;
</b></pre><BR>
<pre><b>size_t ZSTD_compress_generic (ZSTD_CCtx* cctx,
                              ZSTD_outBuffer* output,
                              ZSTD_inBuffer* input,
                              ZSTD_EndDirective endOp);
</b><p>  Behave about the same as ZSTD_compressStream. To note :
  - Compression parameters are pushed into CCtx before starting compression, using ZSTD_CCtx_setParameter()
  - Compression parameters cannot be changed once compression is started.
  - outpot->pos must be <= dstCapacity, input->pos must be <= srcSize
  - outpot->pos and input->pos will be updated. They are guaranteed to remain below their respective limit.
  - @return provides the minimum amount of data still to flush from internal buffers
            or an error code, which can be tested using ZSTD_isError().
            if @return != 0, flush is not fully completed, there is some data left within internal buffers.
  - after a ZSTD_e_end directive, if internal buffer is not fully flushed,
            only ZSTD_e_end or ZSTD_e_flush operations are allowed.
            It is necessary to fully flush internal buffers
            before starting a new compression job, or changing compression parameters.
 
</p></pre><BR>

<pre><b>void ZSTD_CCtx_reset(ZSTD_CCtx* cctx);   </b>/* Not ready yet ! */<b>
</b><p>  Return a CCtx to clean state.
  Useful after an error, or to interrupt an ongoing compression job and start a new one.
  Any internal data not yet flushed is cancelled.
  Dictionary (if any) is dropped.
  All parameters are back to default values.
  It's possible to modify compression parameters after a reset.
 
</p></pre><BR>

<pre><b>size_t ZSTD_compress_generic_simpleArgs (
                ZSTD_CCtx* cctx,
                void* dst, size_t dstCapacity, size_t* dstPos,
          const void* src, size_t srcSize, size_t* srcPos,
                ZSTD_EndDirective endOp);
</b><p>  Same as ZSTD_compress_generic(),
  but using only integral types as arguments.
  Argument list is larger than ZSTD_{in,out}Buffer,
  but can be helpful for binders from dynamic languages
  which have troubles handling structures containing memory pointers.
 
</p></pre><BR>

<pre><b>ZSTD_CCtx_params* ZSTD_createCCtxParams(void);
</b><p>  Quick howto :
  - ZSTD_createCCtxParams() : Create a ZSTD_CCtx_params structure
  - ZSTD_CCtxParam_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 compression jobs.
  - ZSTD_compress_generic() : Do compression using the CCtx.
  - ZSTD_freeCCtxParams() : Free the memory.

  This can be used with ZSTD_estimateCCtxSize_advanced_usingCCtxParams()
  for static allocation for single-threaded compression.
 
</p></pre><BR>

<pre><b>size_t ZSTD_resetCCtxParams(ZSTD_CCtx_params* params);
</b><p>  Reset params to default, with the default compression level.
 
</p></pre><BR>

<pre><b>size_t ZSTD_initCCtxParams(ZSTD_CCtx_params* cctxParams, int compressionLevel);
</b><p>  Initializes the compression parameters of cctxParams according to
  compression level. All other parameters are reset to their default values.
 
</p></pre><BR>

<pre><b>size_t ZSTD_initCCtxParams_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_parameters params);
</b><p>  Initializes the compression and frame parameters of cctxParams according to
  params. All other parameters are reset to their default values.
 
</p></pre><BR>

<pre><b>size_t ZSTD_CCtxParam_setParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, unsigned value);
</b><p>  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().
  Note : when `value` is an enum, cast it to unsigned for proper type checking.
 @result : 0, or an error code (which can be tested with ZSTD_isError()).
 
</p></pre><BR>

<pre><b>size_t ZSTD_CCtx_setParametersUsingCCtxParams(
        ZSTD_CCtx* cctx, const ZSTD_CCtx_params* params);
</b><p>  Apply a set of ZSTD_CCtx_params to the compression context.
  This must be done before the dictionary is loaded.
  The pledgedSrcSize is treated as unknown.
  Multithreading parameters are applied only if nbThreads > 1.
 
</p></pre><BR>

<h3>Advanced parameters for decompression API</h3><pre></pre><b><pre></pre></b><BR>
<pre><b>size_t ZSTD_DCtx_loadDictionary(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);   </b>/* not implemented */<b>
size_t ZSTD_DCtx_loadDictionary_byReference(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);   </b>/* not implemented */<b>
size_t ZSTD_DCtx_loadDictionary_advanced(ZSTD_DCtx* dctx, const void* dict, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictMode_e dictMode);   </b>/* not implemented */<b>
</b><p>  Create an internal DDict from dict buffer,
  to be used to decompress next frames.
 @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 : `dict` content will be copied internally.
            Use ZSTD_DCtx_loadDictionary_byReference()
            to reference dictionary content instead.
            In which case, the dictionary buffer must outlive its users.
  Note 2 : Loading a dictionary involves building tables,
           which has a non-negligible impact on CPU usage and latency.
  Note 3 : Use ZSTD_DCtx_loadDictionary_advanced() to select
           how dictionary content will be interpreted and loaded.
 
</p></pre><BR>

<pre><b>size_t ZSTD_DCtx_refDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict);   </b>/* not implemented */<b>
</b><p>  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 : adding a NULL DDict means "return to no-dictionary mode".
  Note 2 : DDict is just referenced, its lifetime must outlive its usage from DCtx.
 
</p></pre><BR>

<pre><b>size_t ZSTD_DCtx_refPrefix(ZSTD_DCtx* dctx, const void* prefix, size_t prefixSize);   </b>/* not implemented */<b>
size_t ZSTD_DCtx_refPrefix_advanced(ZSTD_DCtx* dctx, const void* prefix, size_t prefixSize, ZSTD_dictMode_e dictMode);   </b>/* not implemented */<b>
</b><p>  Reference a prefix (single-usage dictionary) for next compression job.
  Prefix is **only used once**. It must be explicitly referenced before each frame.
  If there is a need to use same prefix multiple times, consider embedding it into a ZSTD_DDict instead.
 @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 compression job.
  Note 3 : By default, the prefix is treated as raw content (ZSTD_dm_rawContent).
           Use ZSTD_CCtx_refPrefix_advanced() to alter dictMode.
  Note 4 : Referencing a raw content prefix has almost no cpu nor memory cost.
 
</p></pre><BR>

<pre><b>size_t ZSTD_DCtx_setMaxWindowSize(ZSTD_DCtx* dctx, size_t maxWindowSize);
</b><p>  Refuses allocating internal buffers for frames requiring a window size larger than provided limit.
  This is useful to prevent 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 direct mode.
  By default, a decompression context accepts all window sizes <= (1 << ZSTD_WINDOWLOG_MAX)
 @return : 0, or an error code (which can be tested using ZSTD_isError()).
 
</p></pre><BR>

<pre><b>size_t ZSTD_DCtx_setFormat(ZSTD_DCtx* dctx, ZSTD_format_e format);
</b><p>  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()).
 
</p></pre><BR>

<pre><b>size_t ZSTD_decompress_generic(ZSTD_DCtx* dctx,
                               ZSTD_outBuffer* output,
                               ZSTD_inBuffer* input);
</b><p>  Behave the same as ZSTD_decompressStream.
  Decompression parameters cannot be changed once decompression is started.
 @return : an error code, which can be tested using ZSTD_isError()
           if >0, a hint, nb of expected input bytes for next invocation.
           `0` means : a frame has just been fully decoded and flushed.
 
</p></pre><BR>

<pre><b>size_t ZSTD_decompress_generic_simpleArgs (
                ZSTD_DCtx* dctx,
                void* dst, size_t dstCapacity, size_t* dstPos,
          const void* src, size_t srcSize, size_t* srcPos);
</b><p>  Same as ZSTD_decompress_generic(),
  but using only integral types as arguments.
  Argument list is larger than ZSTD_{in,out}Buffer,
  but can be helpful for binders from dynamic languages
  which have troubles handling structures containing memory pointers.
 
</p></pre><BR>

<pre><b>void ZSTD_DCtx_reset(ZSTD_DCtx* dctx);
</b><p>  Return a DCtx to clean state.
  If a decompression was ongoing, any internal data not yet flushed is cancelled.
  All parameters are back to default values, including sticky ones.
  Dictionary (if any) is dropped.
  Parameters can be modified again after a reset.
 
</p></pre><BR>

<a name="Chapter21"></a><h2>Block level API</h2><pre></pre>

<pre><b></b><p>    Frame metadata cost is typically ~18 bytes, which can be non-negligible for very small blocks (< 100 bytes).
    User will have to take in charge required information 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 size, consider using the regular ZSTD_compress() instead.
        Frame metadata is not that costly, and quickly becomes negligible as source size grows larger.
    - When a block is considered not compressible enough, ZSTD_compressBlock() result will be zero.
      In which case, nothing is produced into `dst`.
      + User must test for such outcome and deal directly with uncompressed data
      + 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.
</p></pre><BR>

<h3>Raw zstd block functions</h3><pre></pre><b><pre>size_t ZSTD_getBlockSize   (const ZSTD_CCtx* cctx);
size_t ZSTD_compressBlock  (ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
size_t ZSTD_decompressBlock(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
size_t ZSTD_insertBlock(ZSTD_DCtx* dctx, const void* blockStart, size_t blockSize);  </b>/**< insert uncompressed block into `dctx` history. Useful for multi-blocks decompression */<b>
</pre></b><BR>
</html>
</body>