3 <meta http-equiv="Content-Type" content="text/html; charset=ISO-8859-1">
4 <title>zstd 1.4.8 Manual</title>
7 <h1>zstd 1.4.8 Manual</h1>
9 <a name="Contents"></a><h2>Contents</h2>
11 <li><a href="#Chapter1">Introduction</a></li>
12 <li><a href="#Chapter2">Version</a></li>
13 <li><a href="#Chapter3">Simple API</a></li>
14 <li><a href="#Chapter4">Explicit context</a></li>
15 <li><a href="#Chapter5">Advanced compression API</a></li>
16 <li><a href="#Chapter6">Advanced decompression API</a></li>
17 <li><a href="#Chapter7">Streaming</a></li>
18 <li><a href="#Chapter8">Streaming compression - HowTo</a></li>
19 <li><a href="#Chapter9">Streaming decompression - HowTo</a></li>
20 <li><a href="#Chapter10">Simple dictionary API</a></li>
21 <li><a href="#Chapter11">Bulk processing dictionary API</a></li>
22 <li><a href="#Chapter12">Dictionary helper functions</a></li>
23 <li><a href="#Chapter13">Advanced dictionary and prefix API</a></li>
24 <li><a href="#Chapter14">experimental API (static linking only)</a></li>
25 <li><a href="#Chapter15">Frame size functions</a></li>
26 <li><a href="#Chapter16">Memory management</a></li>
27 <li><a href="#Chapter17">Advanced compression functions</a></li>
28 <li><a href="#Chapter18">Advanced decompression functions</a></li>
29 <li><a href="#Chapter19">Advanced streaming functions</a></li>
30 <li><a href="#Chapter20">Buffer-less and synchronous inner streaming functions</a></li>
31 <li><a href="#Chapter21">Buffer-less streaming compression (synchronous mode)</a></li>
32 <li><a href="#Chapter22">Buffer-less streaming decompression (synchronous mode)</a></li>
33 <li><a href="#Chapter23">Block level API</a></li>
36 <a name="Chapter1"></a><h2>Introduction</h2><pre>
37 zstd, short for Zstandard, is a fast lossless compression algorithm, targeting
38 real-time compression scenarios at zlib-level and better compression ratios.
39 The zstd compression library provides in-memory compression and decompression
42 The library supports regular compression levels from 1 up to ZSTD_maxCLevel(),
43 which is currently 22. Levels >= 20, labeled `--ultra`, should be used with
44 caution, as they require more memory. The library also offers negative
45 compression levels, which extend the range of speed vs. ratio preferences.
46 The lower the level, the faster the speed (at the cost of compression).
48 Compression can be done in:
49 - a single step (described as Simple API)
50 - a single step, reusing a context (described as Explicit context)
51 - unbounded multiple steps (described as Streaming compression)
53 The compression ratio achievable on small data can be highly improved using
54 a dictionary. Dictionary compression can be performed in:
55 - a single step (described as Simple dictionary API)
56 - a single step, reusing a dictionary (described as Bulk-processing
59 Advanced experimental functions can be accessed using
60 `#define ZSTD_STATIC_LINKING_ONLY` before including zstd.h.
62 Advanced experimental APIs should never be used with a dynamically-linked
63 library. They are not "stable"; their definitions or signatures may change in
64 the future. Only static linking is allowed.
67 <a name="Chapter2"></a><h2>Version</h2><pre></pre>
69 <pre><b>unsigned ZSTD_versionNumber(void);
70 </b><p> Return runtime library version, the value is (MAJOR*100*100 + MINOR*100 + RELEASE).
73 <pre><b>const char* ZSTD_versionString(void);
74 </b><p> Return runtime library version, like "1.4.5". Requires v1.3.0+.
77 <a name="Chapter3"></a><h2>Simple API</h2><pre></pre>
79 <pre><b>size_t ZSTD_compress( void* dst, size_t dstCapacity,
80 const void* src, size_t srcSize,
81 int compressionLevel);
82 </b><p> Compresses `src` content as a single zstd compressed frame into already allocated `dst`.
83 Hint : compression runs faster if `dstCapacity` >= `ZSTD_compressBound(srcSize)`.
84 @return : compressed size written into `dst` (<= `dstCapacity),
85 or an error code if it fails (which can be tested using ZSTD_isError()).
88 <pre><b>size_t ZSTD_decompress( void* dst, size_t dstCapacity,
89 const void* src, size_t compressedSize);
90 </b><p> `compressedSize` : must be the _exact_ size of some number of compressed and/or skippable frames.
91 `dstCapacity` is an upper bound of originalSize to regenerate.
92 If user cannot imply a maximum upper bound, it's better to use streaming mode to decompress data.
93 @return : the number of bytes decompressed into `dst` (<= `dstCapacity`),
94 or an errorCode if it fails (which can be tested using ZSTD_isError()).
97 <pre><b>#define ZSTD_CONTENTSIZE_UNKNOWN (0ULL - 1)
98 #define ZSTD_CONTENTSIZE_ERROR (0ULL - 2)
99 unsigned long long ZSTD_getFrameContentSize(const void *src, size_t srcSize);
100 </b><p> `src` should point to the start of a ZSTD encoded frame.
101 `srcSize` must be at least as large as the frame header.
102 hint : any size >= `ZSTD_frameHeaderSize_max` is large enough.
103 @return : - decompressed size of `src` frame content, if known
104 - ZSTD_CONTENTSIZE_UNKNOWN if the size cannot be determined
105 - ZSTD_CONTENTSIZE_ERROR if an error occurred (e.g. invalid magic number, srcSize too small)
106 note 1 : a 0 return value means the frame is valid but "empty".
107 note 2 : decompressed size is an optional field, it may not be present, typically in streaming mode.
108 When `return==ZSTD_CONTENTSIZE_UNKNOWN`, data to decompress could be any size.
109 In which case, it's necessary to use streaming mode to decompress data.
110 Optionally, application can rely on some implicit limit,
111 as ZSTD_decompress() only needs an upper bound of decompressed size.
112 (For example, data could be necessarily cut into blocks <= 16 KB).
113 note 3 : decompressed size is always present when compression is completed using single-pass functions,
114 such as ZSTD_compress(), ZSTD_compressCCtx() ZSTD_compress_usingDict() or ZSTD_compress_usingCDict().
115 note 4 : decompressed size can be very large (64-bits value),
116 potentially larger than what local system can handle as a single memory segment.
117 In which case, it's necessary to use streaming mode to decompress data.
118 note 5 : If source is untrusted, decompressed size could be wrong or intentionally modified.
119 Always ensure return value fits within application's authorized limits.
120 Each application can set its own limits.
121 note 6 : This function replaces ZSTD_getDecompressedSize()
124 <pre><b>unsigned long long ZSTD_getDecompressedSize(const void* src, size_t srcSize);
125 </b><p> NOTE: This function is now obsolete, in favor of ZSTD_getFrameContentSize().
126 Both functions work the same way, but ZSTD_getDecompressedSize() blends
127 "empty", "unknown" and "error" results to the same return value (0),
128 while ZSTD_getFrameContentSize() gives them separate return values.
129 @return : decompressed size of `src` frame content _if known and not empty_, 0 otherwise.
132 <pre><b>size_t ZSTD_findFrameCompressedSize(const void* src, size_t srcSize);
133 </b><p> `src` should point to the start of a ZSTD frame or skippable frame.
134 `srcSize` must be >= first frame size
135 @return : the compressed size of the first frame starting at `src`,
136 suitable to pass as `srcSize` to `ZSTD_decompress` or similar,
137 or an error code if input is invalid
140 <h3>Helper functions</h3><pre></pre><b><pre>#define ZSTD_COMPRESSBOUND(srcSize) ((srcSize) + ((srcSize)>>8) + (((srcSize) < (128<<10)) ? (((128<<10) - (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>
141 size_t ZSTD_compressBound(size_t srcSize); </b>/*!< maximum compressed size in worst case single-pass scenario */<b>
142 unsigned ZSTD_isError(size_t code); </b>/*!< tells if a `size_t` function result is an error code */<b>
143 const char* ZSTD_getErrorName(size_t code); </b>/*!< provides readable string from an error code */<b>
144 int ZSTD_minCLevel(void); </b>/*!< minimum negative compression level allowed */<b>
145 int ZSTD_maxCLevel(void); </b>/*!< maximum compression level available */<b>
147 <a name="Chapter4"></a><h2>Explicit context</h2><pre></pre>
149 <h3>Compression context</h3><pre> When compressing many times,
150 it is recommended to allocate a context just once,
151 and re-use it for each successive compression operation.
152 This will make workload friendlier for system's memory.
153 Note : re-using context is just a speed / resource optimization.
154 It doesn't change the compression ratio, which remains identical.
155 Note 2 : In multi-threaded environments,
156 use one different context per thread for parallel execution.
158 </pre><b><pre>typedef struct ZSTD_CCtx_s ZSTD_CCtx;
159 ZSTD_CCtx* ZSTD_createCCtx(void);
160 size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx);
162 <pre><b>size_t ZSTD_compressCCtx(ZSTD_CCtx* cctx,
163 void* dst, size_t dstCapacity,
164 const void* src, size_t srcSize,
165 int compressionLevel);
166 </b><p> Same as ZSTD_compress(), using an explicit ZSTD_CCtx.
167 Important : in order to behave similarly to `ZSTD_compress()`,
168 this function compresses at requested compression level,
169 __ignoring any other parameter__ .
170 If any advanced parameter was set using the advanced API,
171 they will all be reset. Only `compressionLevel` remains.
175 <h3>Decompression context</h3><pre> When decompressing many times,
176 it is recommended to allocate a context only once,
177 and re-use it for each successive compression operation.
178 This will make workload friendlier for system's memory.
179 Use one context per thread for parallel execution.
180 </pre><b><pre>typedef struct ZSTD_DCtx_s ZSTD_DCtx;
181 ZSTD_DCtx* ZSTD_createDCtx(void);
182 size_t ZSTD_freeDCtx(ZSTD_DCtx* dctx);
184 <pre><b>size_t ZSTD_decompressDCtx(ZSTD_DCtx* dctx,
185 void* dst, size_t dstCapacity,
186 const void* src, size_t srcSize);
187 </b><p> Same as ZSTD_decompress(),
188 requires an allocated ZSTD_DCtx.
189 Compatible with sticky parameters.
193 <a name="Chapter5"></a><h2>Advanced compression API</h2><pre></pre>
195 <pre><b>typedef enum { ZSTD_fast=1,
204 </b>/* note : new strategies _might_ be added in the future.<b>
205 Only the order (from fast to strong) is guaranteed */
208 <pre><b>typedef enum {
210 </b>/* compression parameters<b>
211 * Note: When compressing with a ZSTD_CDict these parameters are superseded
212 * by the parameters used to construct the ZSTD_CDict.
213 * See ZSTD_CCtx_refCDict() for more info (superseded-by-cdict). */
214 ZSTD_c_compressionLevel=100, </b>/* Set compression parameters according to pre-defined cLevel table.<b>
215 * Note that exact compression parameters are dynamically determined,
216 * depending on both compression level and srcSize (when known).
217 * Default level is ZSTD_CLEVEL_DEFAULT==3.
218 * Special: value 0 means default, which is controlled by ZSTD_CLEVEL_DEFAULT.
219 * Note 1 : it's possible to pass a negative compression level.
220 * Note 2 : setting a level does not automatically set all other compression parameters
221 * to default. Setting this will however eventually dynamically impact the compression
222 * parameters which have not been manually set. The manually set
223 * ones will 'stick'. */
224 </b>/* Advanced compression parameters :<b>
225 * It's possible to pin down compression parameters to some specific values.
226 * In which case, these values are no longer dynamically selected by the compressor */
227 ZSTD_c_windowLog=101, </b>/* Maximum allowed back-reference distance, expressed as power of 2.<b>
228 * This will set a memory budget for streaming decompression,
229 * with larger values requiring more memory
230 * and typically compressing more.
231 * Must be clamped between ZSTD_WINDOWLOG_MIN and ZSTD_WINDOWLOG_MAX.
232 * Special: value 0 means "use default windowLog".
233 * Note: Using a windowLog greater than ZSTD_WINDOWLOG_LIMIT_DEFAULT
234 * requires explicitly allowing such size at streaming decompression stage. */
235 ZSTD_c_hashLog=102, </b>/* Size of the initial probe table, as a power of 2.<b>
236 * Resulting memory usage is (1 << (hashLog+2)).
237 * Must be clamped between ZSTD_HASHLOG_MIN and ZSTD_HASHLOG_MAX.
238 * Larger tables improve compression ratio of strategies <= dFast,
239 * and improve speed of strategies > dFast.
240 * Special: value 0 means "use default hashLog". */
241 ZSTD_c_chainLog=103, </b>/* Size of the multi-probe search table, as a power of 2.<b>
242 * Resulting memory usage is (1 << (chainLog+2)).
243 * Must be clamped between ZSTD_CHAINLOG_MIN and ZSTD_CHAINLOG_MAX.
244 * Larger tables result in better and slower compression.
245 * This parameter is useless for "fast" strategy.
246 * It's still useful when using "dfast" strategy,
247 * in which case it defines a secondary probe table.
248 * Special: value 0 means "use default chainLog". */
249 ZSTD_c_searchLog=104, </b>/* Number of search attempts, as a power of 2.<b>
250 * More attempts result in better and slower compression.
251 * This parameter is useless for "fast" and "dFast" strategies.
252 * Special: value 0 means "use default searchLog". */
253 ZSTD_c_minMatch=105, </b>/* Minimum size of searched matches.<b>
254 * Note that Zstandard can still find matches of smaller size,
255 * it just tweaks its search algorithm to look for this size and larger.
256 * Larger values increase compression and decompression speed, but decrease ratio.
257 * Must be clamped between ZSTD_MINMATCH_MIN and ZSTD_MINMATCH_MAX.
258 * Note that currently, for all strategies < btopt, effective minimum is 4.
259 * , for all strategies > fast, effective maximum is 6.
260 * Special: value 0 means "use default minMatchLength". */
261 ZSTD_c_targetLength=106, </b>/* Impact of this field depends on strategy.<b>
262 * For strategies btopt, btultra & btultra2:
263 * Length of Match considered "good enough" to stop search.
264 * Larger values make compression stronger, and slower.
266 * Distance between match sampling.
267 * Larger values make compression faster, and weaker.
268 * Special: value 0 means "use default targetLength". */
269 ZSTD_c_strategy=107, </b>/* See ZSTD_strategy enum definition.<b>
270 * The higher the value of selected strategy, the more complex it is,
271 * resulting in stronger and slower compression.
272 * Special: value 0 means "use default strategy". */
274 </b>/* LDM mode parameters */<b>
275 ZSTD_c_enableLongDistanceMatching=160, </b>/* Enable long distance matching.<b>
276 * This parameter is designed to improve compression ratio
277 * for large inputs, by finding large matches at long distance.
278 * It increases memory usage and window size.
279 * Note: enabling this parameter increases default ZSTD_c_windowLog to 128 MB
280 * except when expressly set to a different value.
281 * Note: will be enabled by default if ZSTD_c_windowLog >= 128 MB and
282 * compression strategy >= ZSTD_btopt (== compression level 16+) */
283 ZSTD_c_ldmHashLog=161, </b>/* Size of the table for long distance matching, as a power of 2.<b>
284 * Larger values increase memory usage and compression ratio,
285 * but decrease compression speed.
286 * Must be clamped between ZSTD_HASHLOG_MIN and ZSTD_HASHLOG_MAX
287 * default: windowlog - 7.
288 * Special: value 0 means "automatically determine hashlog". */
289 ZSTD_c_ldmMinMatch=162, </b>/* Minimum match size for long distance matcher.<b>
290 * Larger/too small values usually decrease compression ratio.
291 * Must be clamped between ZSTD_LDM_MINMATCH_MIN and ZSTD_LDM_MINMATCH_MAX.
292 * Special: value 0 means "use default value" (default: 64). */
293 ZSTD_c_ldmBucketSizeLog=163, </b>/* Log size of each bucket in the LDM hash table for collision resolution.<b>
294 * Larger values improve collision resolution but decrease compression speed.
295 * The maximum value is ZSTD_LDM_BUCKETSIZELOG_MAX.
296 * Special: value 0 means "use default value" (default: 3). */
297 ZSTD_c_ldmHashRateLog=164, </b>/* Frequency of inserting/looking up entries into the LDM hash table.<b>
298 * Must be clamped between 0 and (ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN).
299 * Default is MAX(0, (windowLog - ldmHashLog)), optimizing hash table usage.
300 * Larger values improve compression speed.
301 * Deviating far from default value will likely result in a compression ratio decrease.
302 * Special: value 0 means "automatically determine hashRateLog". */
304 </b>/* frame parameters */<b>
305 ZSTD_c_contentSizeFlag=200, </b>/* Content size will be written into frame header _whenever known_ (default:1)<b>
306 * Content size must be known at the beginning of compression.
307 * This is automatically the case when using ZSTD_compress2(),
308 * For streaming scenarios, content size must be provided with ZSTD_CCtx_setPledgedSrcSize() */
309 ZSTD_c_checksumFlag=201, </b>/* A 32-bits checksum of content is written at end of frame (default:0) */<b>
310 ZSTD_c_dictIDFlag=202, </b>/* When applicable, dictionary's ID is written into frame header (default:1) */<b>
312 </b>/* multi-threading parameters */<b>
313 </b>/* These parameters are only active if multi-threading is enabled (compiled with build macro ZSTD_MULTITHREAD).<b>
314 * Otherwise, trying to set any other value than default (0) will be a no-op and return an error.
315 * In a situation where it's unknown if the linked library supports multi-threading or not,
316 * setting ZSTD_c_nbWorkers to any value >= 1 and consulting the return value provides a quick way to check this property.
318 ZSTD_c_nbWorkers=400, </b>/* Select how many threads will be spawned to compress in parallel.<b>
319 * When nbWorkers >= 1, triggers asynchronous mode when invoking ZSTD_compressStream*() :
320 * ZSTD_compressStream*() consumes input and flush output if possible, but immediately gives back control to caller,
321 * while compression is performed in parallel, within worker thread(s).
322 * (note : a strong exception to this rule is when first invocation of ZSTD_compressStream2() sets ZSTD_e_end :
323 * in which case, ZSTD_compressStream2() delegates to ZSTD_compress2(), which is always a blocking call).
324 * More workers improve speed, but also increase memory usage.
325 * Default value is `0`, aka "single-threaded mode" : no worker is spawned,
326 * compression is performed inside Caller's thread, and all invocations are blocking */
327 ZSTD_c_jobSize=401, </b>/* Size of a compression job. This value is enforced only when nbWorkers >= 1.<b>
328 * Each compression job is completed in parallel, so this value can indirectly impact the nb of active threads.
329 * 0 means default, which is dynamically determined based on compression parameters.
330 * Job size must be a minimum of overlap size, or 1 MB, whichever is largest.
331 * The minimum size is automatically and transparently enforced. */
332 ZSTD_c_overlapLog=402, </b>/* Control the overlap size, as a fraction of window size.<b>
333 * The overlap size is an amount of data reloaded from previous job at the beginning of a new job.
334 * It helps preserve compression ratio, while each job is compressed in parallel.
335 * This value is enforced only when nbWorkers >= 1.
336 * Larger values increase compression ratio, but decrease speed.
337 * Possible values range from 0 to 9 :
338 * - 0 means "default" : value will be determined by the library, depending on strategy
339 * - 1 means "no overlap"
340 * - 9 means "full overlap", using a full window size.
341 * Each intermediate rank increases/decreases load size by a factor 2 :
342 * 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
343 * default value varies between 6 and 9, depending on strategy */
345 </b>/* note : additional experimental parameters are also available<b>
346 * within the experimental section of the API.
347 * At the time of this writing, they include :
350 * ZSTD_c_forceMaxWindow
351 * ZSTD_c_forceAttachDict
352 * ZSTD_c_literalCompressionMode
353 * ZSTD_c_targetCBlockSize
355 * ZSTD_c_enableDedicatedDictSearch
356 * ZSTD_c_stableInBuffer
357 * ZSTD_c_stableOutBuffer
358 * ZSTD_c_blockDelimiters
359 * ZSTD_c_validateSequences
360 * Because they are not stable, it's necessary to define ZSTD_STATIC_LINKING_ONLY to access them.
361 * note : never ever use experimentalParam? names directly;
362 * also, the enums values themselves are unstable and can still change.
364 ZSTD_c_experimentalParam1=500,
365 ZSTD_c_experimentalParam2=10,
366 ZSTD_c_experimentalParam3=1000,
367 ZSTD_c_experimentalParam4=1001,
368 ZSTD_c_experimentalParam5=1002,
369 ZSTD_c_experimentalParam6=1003,
370 ZSTD_c_experimentalParam7=1004,
371 ZSTD_c_experimentalParam8=1005,
372 ZSTD_c_experimentalParam9=1006,
373 ZSTD_c_experimentalParam10=1007,
374 ZSTD_c_experimentalParam11=1008,
375 ZSTD_c_experimentalParam12=1009
378 <pre><b>typedef struct {
384 <pre><b>ZSTD_bounds ZSTD_cParam_getBounds(ZSTD_cParameter cParam);
385 </b><p> All parameters must belong to an interval with lower and upper bounds,
386 otherwise they will either trigger an error or be automatically clamped.
387 @return : a structure, ZSTD_bounds, which contains
388 - an error status field, which must be tested using ZSTD_isError()
389 - lower and upper bounds, both inclusive
393 <pre><b>size_t ZSTD_CCtx_setParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int value);
394 </b><p> Set one compression parameter, selected by enum ZSTD_cParameter.
395 All parameters have valid bounds. Bounds can be queried using ZSTD_cParam_getBounds().
396 Providing a value beyond bound will either clamp it, or trigger an error (depending on parameter).
397 Setting a parameter is generally only possible during frame initialization (before starting compression).
398 Exception : when using multi-threading mode (nbWorkers >= 1),
399 the following parameters can be updated _during_ compression (within same frame):
400 => compressionLevel, hashLog, chainLog, searchLog, minMatch, targetLength and strategy.
401 new parameters will be active for next job only (after a flush()).
402 @return : an error code (which can be tested using ZSTD_isError()).
406 <pre><b>size_t ZSTD_CCtx_setPledgedSrcSize(ZSTD_CCtx* cctx, unsigned long long pledgedSrcSize);
407 </b><p> Total input data size to be compressed as a single frame.
408 Value will be written in frame header, unless if explicitly forbidden using ZSTD_c_contentSizeFlag.
409 This value will also be controlled at end of frame, and trigger an error if not respected.
410 @result : 0, or an error code (which can be tested with ZSTD_isError()).
411 Note 1 : pledgedSrcSize==0 actually means zero, aka an empty frame.
412 In order to mean "unknown content size", pass constant ZSTD_CONTENTSIZE_UNKNOWN.
413 ZSTD_CONTENTSIZE_UNKNOWN is default value for any new frame.
414 Note 2 : pledgedSrcSize is only valid once, for the next frame.
415 It's discarded at the end of the frame, and replaced by ZSTD_CONTENTSIZE_UNKNOWN.
416 Note 3 : Whenever all input data is provided and consumed in a single round,
417 for example with ZSTD_compress2(),
418 or invoking immediately ZSTD_compressStream2(,,,ZSTD_e_end),
419 this value is automatically overridden by srcSize instead.
423 <pre><b>typedef enum {
424 ZSTD_reset_session_only = 1,
425 ZSTD_reset_parameters = 2,
426 ZSTD_reset_session_and_parameters = 3
427 } ZSTD_ResetDirective;
429 <pre><b>size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset);
430 </b><p> There are 2 different things that can be reset, independently or jointly :
431 - The session : will stop compressing current frame, and make CCtx ready to start a new one.
432 Useful after an error, or to interrupt any ongoing compression.
433 Any internal data not yet flushed is cancelled.
434 Compression parameters and dictionary remain unchanged.
435 They will be used to compress next frame.
436 Resetting session never fails.
437 - The parameters : changes all parameters back to "default".
438 This removes any reference to any dictionary too.
439 Parameters can only be changed between 2 sessions (i.e. no compression is currently ongoing)
440 otherwise the reset fails, and function returns an error value (which can be tested using ZSTD_isError())
441 - Both : similar to resetting the session, followed by resetting parameters.
445 <pre><b>size_t ZSTD_compress2( ZSTD_CCtx* cctx,
446 void* dst, size_t dstCapacity,
447 const void* src, size_t srcSize);
448 </b><p> Behave the same as ZSTD_compressCCtx(), but compression parameters are set using the advanced API.
449 ZSTD_compress2() always starts a new frame.
450 Should cctx hold data from a previously unfinished frame, everything about it is forgotten.
451 - Compression parameters are pushed into CCtx before starting compression, using ZSTD_CCtx_set*()
452 - The function is always blocking, returns when compression is completed.
453 Hint : compression runs faster if `dstCapacity` >= `ZSTD_compressBound(srcSize)`.
454 @return : compressed size written into `dst` (<= `dstCapacity),
455 or an error code if it fails (which can be tested using ZSTD_isError()).
459 <a name="Chapter6"></a><h2>Advanced decompression API</h2><pre></pre>
461 <pre><b>typedef enum {
463 ZSTD_d_windowLogMax=100, </b>/* Select a size limit (in power of 2) beyond which<b>
464 * the streaming API will refuse to allocate memory buffer
465 * in order to protect the host from unreasonable memory requirements.
466 * This parameter is only useful in streaming mode, since no internal buffer is allocated in single-pass mode.
467 * By default, a decompression context accepts window sizes <= (1 << ZSTD_WINDOWLOG_LIMIT_DEFAULT).
468 * Special: value 0 means "use default maximum windowLog". */
470 </b>/* note : additional experimental parameters are also available<b>
471 * within the experimental section of the API.
472 * At the time of this writing, they include :
474 * ZSTD_d_stableOutBuffer
475 * ZSTD_d_forceIgnoreChecksum
476 * Because they are not stable, it's necessary to define ZSTD_STATIC_LINKING_ONLY to access them.
477 * note : never ever use experimentalParam? names directly
479 ZSTD_d_experimentalParam1=1000,
480 ZSTD_d_experimentalParam2=1001,
481 ZSTD_d_experimentalParam3=1002
485 <pre><b>ZSTD_bounds ZSTD_dParam_getBounds(ZSTD_dParameter dParam);
486 </b><p> All parameters must belong to an interval with lower and upper bounds,
487 otherwise they will either trigger an error or be automatically clamped.
488 @return : a structure, ZSTD_bounds, which contains
489 - an error status field, which must be tested using ZSTD_isError()
490 - both lower and upper bounds, inclusive
494 <pre><b>size_t ZSTD_DCtx_setParameter(ZSTD_DCtx* dctx, ZSTD_dParameter param, int value);
495 </b><p> Set one compression parameter, selected by enum ZSTD_dParameter.
496 All parameters have valid bounds. Bounds can be queried using ZSTD_dParam_getBounds().
497 Providing a value beyond bound will either clamp it, or trigger an error (depending on parameter).
498 Setting a parameter is only possible during frame initialization (before starting decompression).
499 @return : 0, or an error code (which can be tested using ZSTD_isError()).
503 <pre><b>size_t ZSTD_DCtx_reset(ZSTD_DCtx* dctx, ZSTD_ResetDirective reset);
504 </b><p> Return a DCtx to clean state.
505 Session and parameters can be reset jointly or separately.
506 Parameters can only be reset when no active frame is being decompressed.
507 @return : 0, or an error code, which can be tested with ZSTD_isError()
511 <a name="Chapter7"></a><h2>Streaming</h2><pre></pre>
513 <pre><b>typedef struct ZSTD_inBuffer_s {
514 const void* src; </b>/**< start of input buffer */<b>
515 size_t size; </b>/**< size of input buffer */<b>
516 size_t pos; </b>/**< position where reading stopped. Will be updated. Necessarily 0 <= pos <= size */<b>
519 <pre><b>typedef struct ZSTD_outBuffer_s {
520 void* dst; </b>/**< start of output buffer */<b>
521 size_t size; </b>/**< size of output buffer */<b>
522 size_t pos; </b>/**< position where writing stopped. Will be updated. Necessarily 0 <= pos <= size */<b>
525 <a name="Chapter8"></a><h2>Streaming compression - HowTo</h2><pre>
526 A ZSTD_CStream object is required to track streaming operation.
527 Use ZSTD_createCStream() and ZSTD_freeCStream() to create/release resources.
528 ZSTD_CStream objects can be reused multiple times on consecutive compression operations.
529 It is recommended to re-use ZSTD_CStream since it will play nicer with system's memory, by re-using already allocated memory.
531 For parallel execution, use one separate ZSTD_CStream per thread.
533 note : since v1.3.0, ZSTD_CStream and ZSTD_CCtx are the same thing.
535 Parameters are sticky : when starting a new compression on the same context,
536 it will re-use the same sticky parameters as previous compression session.
537 When in doubt, it's recommended to fully initialize the context before usage.
538 Use ZSTD_CCtx_reset() to reset the context and ZSTD_CCtx_setParameter(),
539 ZSTD_CCtx_setPledgedSrcSize(), or ZSTD_CCtx_loadDictionary() and friends to
540 set more specific parameters, the pledged source size, or load a dictionary.
542 Use ZSTD_compressStream2() with ZSTD_e_continue as many times as necessary to
543 consume input stream. The function will automatically update both `pos`
544 fields within `input` and `output`.
545 Note that the function may not consume the entire input, for example, because
546 the output buffer is already full, in which case `input.pos < input.size`.
547 The caller must check if input has been entirely consumed.
548 If not, the caller must make some room to receive more compressed data,
549 and then present again remaining input data.
550 note: ZSTD_e_continue is guaranteed to make some forward progress when called,
551 but doesn't guarantee maximal forward progress. This is especially relevant
552 when compressing with multiple threads. The call won't block if it can
553 consume some input, but if it can't it will wait for some, but not all,
554 output to be flushed.
555 @return : provides a minimum amount of data remaining to be flushed from internal buffers
556 or an error code, which can be tested using ZSTD_isError().
558 At any moment, it's possible to flush whatever data might remain stuck within internal buffer,
559 using ZSTD_compressStream2() with ZSTD_e_flush. `output->pos` will be updated.
560 Note that, if `output->size` is too small, a single invocation with ZSTD_e_flush might not be enough (return code > 0).
561 In which case, make some room to receive more compressed data, and call again ZSTD_compressStream2() with ZSTD_e_flush.
562 You must continue calling ZSTD_compressStream2() with ZSTD_e_flush until it returns 0, at which point you can change the
564 note: ZSTD_e_flush will flush as much output as possible, meaning when compressing with multiple threads, it will
565 block until the flush is complete or the output buffer is full.
566 @return : 0 if internal buffers are entirely flushed,
567 >0 if some data still present within internal buffer (the value is minimal estimation of remaining size),
568 or an error code, which can be tested using ZSTD_isError().
570 Calling ZSTD_compressStream2() with ZSTD_e_end instructs to finish a frame.
571 It will perform a flush and write frame epilogue.
572 The epilogue is required for decoders to consider a frame completed.
573 flush operation is the same, and follows same rules as calling ZSTD_compressStream2() with ZSTD_e_flush.
574 You must continue calling ZSTD_compressStream2() with ZSTD_e_end until it returns 0, at which point you are free to
576 note: ZSTD_e_end will flush as much output as possible, meaning when compressing with multiple threads, it will
577 block until the flush is complete or the output buffer is full.
578 @return : 0 if frame fully completed and fully flushed,
579 >0 if some data still present within internal buffer (the value is minimal estimation of remaining size),
580 or an error code, which can be tested using ZSTD_isError().
585 <pre><b>typedef ZSTD_CCtx ZSTD_CStream; </b>/**< CCtx and CStream are now effectively same object (>= v1.3.0) */<b>
587 <h3>ZSTD_CStream management functions</h3><pre></pre><b><pre>ZSTD_CStream* ZSTD_createCStream(void);
588 size_t ZSTD_freeCStream(ZSTD_CStream* zcs);
590 <h3>Streaming compression functions</h3><pre></pre><b><pre>typedef enum {
591 ZSTD_e_continue=0, </b>/* collect more data, encoder decides when to output compressed result, for optimal compression ratio */<b>
592 ZSTD_e_flush=1, </b>/* flush any data provided so far,<b>
593 * it creates (at least) one new block, that can be decoded immediately on reception;
594 * frame will continue: any future data can still reference previously compressed data, improving compression.
595 * note : multithreaded compression will block to flush as much output as possible. */
596 ZSTD_e_end=2 </b>/* flush any remaining data _and_ close current frame.<b>
597 * note that frame is only closed after compressed data is fully flushed (return value == 0).
598 * After that point, any additional data starts a new frame.
599 * note : each frame is independent (does not reference any content from previous frame).
600 : note : multithreaded compression will block to flush as much output as possible. */
603 <pre><b>size_t ZSTD_compressStream2( ZSTD_CCtx* cctx,
604 ZSTD_outBuffer* output,
605 ZSTD_inBuffer* input,
606 ZSTD_EndDirective endOp);
607 </b><p> Behaves about the same as ZSTD_compressStream, with additional control on end directive.
608 - Compression parameters are pushed into CCtx before starting compression, using ZSTD_CCtx_set*()
609 - Compression parameters cannot be changed once compression is started (save a list of exceptions in multi-threading mode)
610 - output->pos must be <= dstCapacity, input->pos must be <= srcSize
611 - output->pos and input->pos will be updated. They are guaranteed to remain below their respective limit.
612 - endOp must be a valid directive
613 - When nbWorkers==0 (default), function is blocking : it completes its job before returning to caller.
614 - When nbWorkers>=1, function is non-blocking : it copies a portion of input, distributes jobs to internal worker threads, flush to output whatever is available,
615 and then immediately returns, just indicating that there is some data remaining to be flushed.
616 The function nonetheless guarantees forward progress : it will return only after it reads or write at least 1+ byte.
617 - 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.
618 - @return provides a minimum amount of data remaining to be flushed from internal buffers
619 or an error code, which can be tested using ZSTD_isError().
620 if @return != 0, flush is not fully completed, there is still some data left within internal buffers.
621 This is useful for ZSTD_e_flush, since in this case more flushes are necessary to empty all buffers.
622 For ZSTD_e_end, @return == 0 when internal buffers are fully flushed and frame is completed.
623 - after a ZSTD_e_end directive, if internal buffer is not fully flushed (@return != 0),
624 only ZSTD_e_end or ZSTD_e_flush operations are allowed.
625 Before starting a new compression job, or changing compression parameters,
626 it is required to fully flush internal buffers.
630 <pre><b>size_t ZSTD_CStreamInSize(void); </b>/**< recommended size for input buffer */<b>
632 <pre><b>size_t ZSTD_CStreamOutSize(void); </b>/**< recommended size for output buffer. Guarantee to successfully flush at least one complete compressed block. */<b>
634 <pre><b>size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel);
636 * Alternative for ZSTD_compressStream2(zcs, output, input, ZSTD_e_continue).
637 * NOTE: The return value is different. ZSTD_compressStream() returns a hint for
638 * the next read size (if non-zero and not an error). ZSTD_compressStream2()
639 * returns the minimum nb of bytes left to flush (if non-zero and not an error).
641 size_t ZSTD_compressStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input);
642 </b>/*! Equivalent to ZSTD_compressStream2(zcs, output, &emptyInput, ZSTD_e_flush). */<b>
643 size_t ZSTD_flushStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output);
644 </b>/*! Equivalent to ZSTD_compressStream2(zcs, output, &emptyInput, ZSTD_e_end). */<b>
645 size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output);
647 ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
648 ZSTD_CCtx_refCDict(zcs, NULL); // clear the dictionary (if any)
649 ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel);
653 <a name="Chapter9"></a><h2>Streaming decompression - HowTo</h2><pre>
654 A ZSTD_DStream object is required to track streaming operations.
655 Use ZSTD_createDStream() and ZSTD_freeDStream() to create/release resources.
656 ZSTD_DStream objects can be re-used multiple times.
658 Use ZSTD_initDStream() to start a new decompression operation.
659 @return : recommended first input size
660 Alternatively, use advanced API to set specific properties.
662 Use ZSTD_decompressStream() repetitively to consume your input.
663 The function will update both `pos` fields.
664 If `input.pos < input.size`, some input has not been consumed.
665 It's up to the caller to present again remaining data.
666 The function tries to flush all data decoded immediately, respecting output buffer size.
667 If `output.pos < output.size`, decoder has flushed everything it could.
668 But if `output.pos == output.size`, there might be some data left within internal buffers.,
669 In which case, call ZSTD_decompressStream() again to flush whatever remains in the buffer.
670 Note : with no additional input provided, amount of data flushed is necessarily <= ZSTD_BLOCKSIZE_MAX.
671 @return : 0 when a frame is completely decoded and fully flushed,
672 or an error code, which can be tested using ZSTD_isError(),
673 or any other value > 0, which means there is still some decoding or flushing to do to complete current frame :
674 the return value is a suggested next input size (just a hint for better latency)
675 that will never request more than the remaining frame size.
679 <pre><b>typedef ZSTD_DCtx ZSTD_DStream; </b>/**< DCtx and DStream are now effectively same object (>= v1.3.0) */<b>
681 <h3>ZSTD_DStream management functions</h3><pre></pre><b><pre>ZSTD_DStream* ZSTD_createDStream(void);
682 size_t ZSTD_freeDStream(ZSTD_DStream* zds);
684 <h3>Streaming decompression functions</h3><pre></pre><b><pre></pre></b><BR>
685 <pre><b>size_t ZSTD_DStreamInSize(void); </b>/*!< recommended size for input buffer */<b>
687 <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>
689 <a name="Chapter10"></a><h2>Simple dictionary API</h2><pre></pre>
691 <pre><b>size_t ZSTD_compress_usingDict(ZSTD_CCtx* ctx,
692 void* dst, size_t dstCapacity,
693 const void* src, size_t srcSize,
694 const void* dict,size_t dictSize,
695 int compressionLevel);
696 </b><p> Compression at an explicit compression level using a Dictionary.
697 A dictionary can be any arbitrary data segment (also called a prefix),
698 or a buffer with specified information (see dictBuilder/zdict.h).
699 Note : This function loads the dictionary, resulting in significant startup delay.
700 It's intended for a dictionary used only once.
701 Note 2 : When `dict == NULL || dictSize < 8` no dictionary is used.
704 <pre><b>size_t ZSTD_decompress_usingDict(ZSTD_DCtx* dctx,
705 void* dst, size_t dstCapacity,
706 const void* src, size_t srcSize,
707 const void* dict,size_t dictSize);
708 </b><p> Decompression using a known Dictionary.
709 Dictionary must be identical to the one used during compression.
710 Note : This function loads the dictionary, resulting in significant startup delay.
711 It's intended for a dictionary used only once.
712 Note : When `dict == NULL || dictSize < 8` no dictionary is used.
715 <a name="Chapter11"></a><h2>Bulk processing dictionary API</h2><pre></pre>
717 <pre><b>ZSTD_CDict* ZSTD_createCDict(const void* dictBuffer, size_t dictSize,
718 int compressionLevel);
719 </b><p> When compressing multiple messages or blocks using the same dictionary,
720 it's recommended to digest the dictionary only once, since it's a costly operation.
721 ZSTD_createCDict() will create a state from digesting a dictionary.
722 The resulting state can be used for future compression operations with very limited startup cost.
723 ZSTD_CDict can be created once and shared by multiple threads concurrently, since its usage is read-only.
724 @dictBuffer can be released after ZSTD_CDict creation, because its content is copied within CDict.
725 Note 1 : Consider experimental function `ZSTD_createCDict_byReference()` if you prefer to not duplicate @dictBuffer content.
726 Note 2 : A ZSTD_CDict can be created from an empty @dictBuffer,
727 in which case the only thing that it transports is the @compressionLevel.
728 This can be useful in a pipeline featuring ZSTD_compress_usingCDict() exclusively,
729 expecting a ZSTD_CDict parameter with any data, including those without a known dictionary.
732 <pre><b>size_t ZSTD_freeCDict(ZSTD_CDict* CDict);
733 </b><p> Function frees memory allocated by ZSTD_createCDict().
736 <pre><b>size_t ZSTD_compress_usingCDict(ZSTD_CCtx* cctx,
737 void* dst, size_t dstCapacity,
738 const void* src, size_t srcSize,
739 const ZSTD_CDict* cdict);
740 </b><p> Compression using a digested Dictionary.
741 Recommended when same dictionary is used multiple times.
742 Note : compression level is _decided at dictionary creation time_,
743 and frame parameters are hardcoded (dictID=yes, contentSize=yes, checksum=no)
746 <pre><b>ZSTD_DDict* ZSTD_createDDict(const void* dictBuffer, size_t dictSize);
747 </b><p> Create a digested dictionary, ready to start decompression operation without startup delay.
748 dictBuffer can be released after DDict creation, as its content is copied inside DDict.
751 <pre><b>size_t ZSTD_freeDDict(ZSTD_DDict* ddict);
752 </b><p> Function frees memory allocated with ZSTD_createDDict()
755 <pre><b>size_t ZSTD_decompress_usingDDict(ZSTD_DCtx* dctx,
756 void* dst, size_t dstCapacity,
757 const void* src, size_t srcSize,
758 const ZSTD_DDict* ddict);
759 </b><p> Decompression using a digested Dictionary.
760 Recommended when same dictionary is used multiple times.
763 <a name="Chapter12"></a><h2>Dictionary helper functions</h2><pre></pre>
765 <pre><b>unsigned ZSTD_getDictID_fromDict(const void* dict, size_t dictSize);
766 </b><p> Provides the dictID stored within dictionary.
767 if @return == 0, the dictionary is not conformant with Zstandard specification.
768 It can still be loaded, but as a content-only dictionary.
771 <pre><b>unsigned ZSTD_getDictID_fromDDict(const ZSTD_DDict* ddict);
772 </b><p> Provides the dictID of the dictionary loaded into `ddict`.
773 If @return == 0, the dictionary is not conformant to Zstandard specification, or empty.
774 Non-conformant dictionaries can still be loaded, but as content-only dictionaries.
777 <pre><b>unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize);
778 </b><p> Provides the dictID required to decompressed the frame stored within `src`.
779 If @return == 0, the dictID could not be decoded.
780 This could for one of the following reasons :
781 - The frame does not require a dictionary to be decoded (most common case).
782 - The frame was built with dictID intentionally removed. Whatever dictionary is necessary is a hidden information.
783 Note : this use case also happens when using a non-conformant dictionary.
784 - `srcSize` is too small, and as a result, the frame header could not be decoded (only possible if `srcSize < ZSTD_FRAMEHEADERSIZE_MAX`).
785 - This is not a Zstandard frame.
786 When identifying the exact failure cause, it's possible to use ZSTD_getFrameHeader(), which will provide a more precise error code.
789 <a name="Chapter13"></a><h2>Advanced dictionary and prefix API</h2><pre>
790 This API allows dictionaries to be used with ZSTD_compress2(),
791 ZSTD_compressStream2(), and ZSTD_decompress(). Dictionaries are sticky, and
792 only reset with the context is reset with ZSTD_reset_parameters or
793 ZSTD_reset_session_and_parameters. Prefixes are single-use.
796 <pre><b>size_t ZSTD_CCtx_loadDictionary(ZSTD_CCtx* cctx, const void* dict, size_t dictSize);
797 </b><p> Create an internal CDict from `dict` buffer.
798 Decompression will have to use same dictionary.
799 @result : 0, or an error code (which can be tested with ZSTD_isError()).
800 Special: Loading a NULL (or 0-size) dictionary invalidates previous dictionary,
801 meaning "return to no-dictionary mode".
802 Note 1 : Dictionary is sticky, it will be used for all future compressed frames.
803 To return to "no-dictionary" situation, load a NULL dictionary (or reset parameters).
804 Note 2 : Loading a dictionary involves building tables.
805 It's also a CPU consuming operation, with non-negligible impact on latency.
806 Tables are dependent on compression parameters, and for this reason,
807 compression parameters can no longer be changed after loading a dictionary.
808 Note 3 :`dict` content will be copied internally.
809 Use experimental ZSTD_CCtx_loadDictionary_byReference() to reference content instead.
810 In such a case, dictionary buffer must outlive its users.
811 Note 4 : Use ZSTD_CCtx_loadDictionary_advanced()
812 to precisely select how dictionary content must be interpreted.
815 <pre><b>size_t ZSTD_CCtx_refCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict);
816 </b><p> Reference a prepared dictionary, to be used for all next compressed frames.
817 Note that compression parameters are enforced from within CDict,
818 and supersede any compression parameter previously set within CCtx.
819 The parameters ignored are labled as "superseded-by-cdict" in the ZSTD_cParameter enum docs.
820 The ignored parameters will be used again if the CCtx is returned to no-dictionary mode.
821 The dictionary will remain valid for future compressed frames using same CCtx.
822 @result : 0, or an error code (which can be tested with ZSTD_isError()).
823 Special : Referencing a NULL CDict means "return to no-dictionary mode".
824 Note 1 : Currently, only one dictionary can be managed.
825 Referencing a new dictionary effectively "discards" any previous one.
826 Note 2 : CDict is just referenced, its lifetime must outlive its usage within CCtx.
829 <pre><b>size_t ZSTD_CCtx_refPrefix(ZSTD_CCtx* cctx,
830 const void* prefix, size_t prefixSize);
831 </b><p> Reference a prefix (single-usage dictionary) for next compressed frame.
832 A prefix is **only used once**. Tables are discarded at end of frame (ZSTD_e_end).
833 Decompression will need same prefix to properly regenerate data.
834 Compressing with a prefix is similar in outcome as performing a diff and compressing it,
835 but performs much faster, especially during decompression (compression speed is tunable with compression level).
836 @result : 0, or an error code (which can be tested with ZSTD_isError()).
837 Special: Adding any prefix (including NULL) invalidates any previous prefix or dictionary
838 Note 1 : Prefix buffer is referenced. It **must** outlive compression.
839 Its content must remain unmodified during compression.
840 Note 2 : If the intention is to diff some large src data blob with some prior version of itself,
841 ensure that the window size is large enough to contain the entire source.
842 See ZSTD_c_windowLog.
843 Note 3 : Referencing a prefix involves building tables, which are dependent on compression parameters.
844 It's a CPU consuming operation, with non-negligible impact on latency.
845 If there is a need to use the same prefix multiple times, consider loadDictionary instead.
846 Note 4 : By default, the prefix is interpreted as raw content (ZSTD_dct_rawContent).
847 Use experimental ZSTD_CCtx_refPrefix_advanced() to alter dictionary interpretation.
850 <pre><b>size_t ZSTD_DCtx_loadDictionary(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);
851 </b><p> Create an internal DDict from dict buffer,
852 to be used to decompress next frames.
853 The dictionary remains valid for all future frames, until explicitly invalidated.
854 @result : 0, or an error code (which can be tested with ZSTD_isError()).
855 Special : Adding a NULL (or 0-size) dictionary invalidates any previous dictionary,
856 meaning "return to no-dictionary mode".
857 Note 1 : Loading a dictionary involves building tables,
858 which has a non-negligible impact on CPU usage and latency.
859 It's recommended to "load once, use many times", to amortize the cost
860 Note 2 :`dict` content will be copied internally, so `dict` can be released after loading.
861 Use ZSTD_DCtx_loadDictionary_byReference() to reference dictionary content instead.
862 Note 3 : Use ZSTD_DCtx_loadDictionary_advanced() to take control of
863 how dictionary content is loaded and interpreted.
867 <pre><b>size_t ZSTD_DCtx_refDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict);
868 </b><p> Reference a prepared dictionary, to be used to decompress next frames.
869 The dictionary remains active for decompression of future frames using same DCtx.
870 @result : 0, or an error code (which can be tested with ZSTD_isError()).
871 Note 1 : Currently, only one dictionary can be managed.
872 Referencing a new dictionary effectively "discards" any previous one.
873 Special: referencing a NULL DDict means "return to no-dictionary mode".
874 Note 2 : DDict is just referenced, its lifetime must outlive its usage from DCtx.
878 <pre><b>size_t ZSTD_DCtx_refPrefix(ZSTD_DCtx* dctx,
879 const void* prefix, size_t prefixSize);
880 </b><p> Reference a prefix (single-usage dictionary) to decompress next frame.
881 This is the reverse operation of ZSTD_CCtx_refPrefix(),
882 and must use the same prefix as the one used during compression.
883 Prefix is **only used once**. Reference is discarded at end of frame.
884 End of frame is reached when ZSTD_decompressStream() returns 0.
885 @result : 0, or an error code (which can be tested with ZSTD_isError()).
886 Note 1 : Adding any prefix (including NULL) invalidates any previously set prefix or dictionary
887 Note 2 : Prefix buffer is referenced. It **must** outlive decompression.
888 Prefix buffer must remain unmodified up to the end of frame,
889 reached when ZSTD_decompressStream() returns 0.
890 Note 3 : By default, the prefix is treated as raw content (ZSTD_dct_rawContent).
891 Use ZSTD_CCtx_refPrefix_advanced() to alter dictMode (Experimental section)
892 Note 4 : Referencing a raw content prefix has almost no cpu nor memory cost.
893 A full dictionary is more costly, as it requires building tables.
897 <pre><b>size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx);
898 size_t ZSTD_sizeof_DCtx(const ZSTD_DCtx* dctx);
899 size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs);
900 size_t ZSTD_sizeof_DStream(const ZSTD_DStream* zds);
901 size_t ZSTD_sizeof_CDict(const ZSTD_CDict* cdict);
902 size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict);
903 </b><p> These functions give the _current_ memory usage of selected object.
904 Note that object memory usage can evolve (increase or decrease) over time.
907 <a name="Chapter14"></a><h2>experimental API (static linking only)</h2><pre>
908 The following symbols and constants
909 are not planned to join "stable API" status in the near future.
910 They can still change in future versions.
911 Some of them are planned to remain in the static_only section indefinitely.
912 Some of them might be removed in the future (especially when redundant with existing stable functions)
916 <pre><b>typedef struct {
917 unsigned int offset; </b>/* The offset of the match. (NOT the same as the offset code)<b>
918 * If offset == 0 and matchLength == 0, this sequence represents the last
919 * literals in the block of litLength size.
922 unsigned int litLength; </b>/* Literal length of the sequence. */<b>
923 unsigned int matchLength; </b>/* Match length of the sequence. */<b>
925 </b>/* Note: Users of this API may provide a sequence with matchLength == litLength == offset == 0.<b>
926 * In this case, we will treat the sequence as a marker for a block boundary.
929 unsigned int rep; </b>/* Represents which repeat offset is represented by the field 'offset'.<b>
930 * Ranges from [0, 3].
932 * Repeat offsets are essentially previous offsets from previous sequences sorted in
933 * recency order. For more detail, see doc/zstd_compression_format.md
935 * If rep == 0, then 'offset' does not contain a repeat offset.
938 * rep == 1 --> offset == repeat_offset_1
939 * rep == 2 --> offset == repeat_offset_2
940 * rep == 3 --> offset == repeat_offset_3
942 * rep == 1 --> offset == repeat_offset_2
943 * rep == 2 --> offset == repeat_offset_3
944 * rep == 3 --> offset == repeat_offset_1 - 1
946 * Note: This field is optional. ZSTD_generateSequences() will calculate the value of
947 * 'rep', but repeat offsets do not necessarily need to be calculated from an external
948 * sequence provider's perspective. For example, ZSTD_compressSequences() does not
949 * use this 'rep' field at all (as of now).
953 <pre><b>typedef struct {
954 unsigned windowLog; </b>/**< largest match distance : larger == more compression, more memory needed during decompression */<b>
955 unsigned chainLog; </b>/**< fully searched segment : larger == more compression, slower, more memory (useless for fast) */<b>
956 unsigned hashLog; </b>/**< dispatch table : larger == faster, more memory */<b>
957 unsigned searchLog; </b>/**< nb of searches : larger == more compression, slower */<b>
958 unsigned minMatch; </b>/**< match length searched : larger == faster decompression, sometimes less compression */<b>
959 unsigned targetLength; </b>/**< acceptable match size for optimal parser (only) : larger == more compression, slower */<b>
960 ZSTD_strategy strategy; </b>/**< see ZSTD_strategy definition above */<b>
961 } ZSTD_compressionParameters;
963 <pre><b>typedef struct {
964 int contentSizeFlag; </b>/**< 1: content size will be in frame header (when known) */<b>
965 int checksumFlag; </b>/**< 1: generate a 32-bits checksum using XXH64 algorithm at end of frame, for error detection */<b>
966 int noDictIDFlag; </b>/**< 1: no dictID will be saved into frame header (dictID is only useful for dictionary compression) */<b>
967 } ZSTD_frameParameters;
969 <pre><b>typedef struct {
970 ZSTD_compressionParameters cParams;
971 ZSTD_frameParameters fParams;
974 <pre><b>typedef enum {
975 ZSTD_dct_auto = 0, </b>/* dictionary is "full" when starting with ZSTD_MAGIC_DICTIONARY, otherwise it is "rawContent" */<b>
976 ZSTD_dct_rawContent = 1, </b>/* ensures dictionary is always loaded as rawContent, even if it starts with ZSTD_MAGIC_DICTIONARY */<b>
977 ZSTD_dct_fullDict = 2 </b>/* refuses to load a dictionary if it does not respect Zstandard's specification, starting with ZSTD_MAGIC_DICTIONARY */<b>
978 } ZSTD_dictContentType_e;
980 <pre><b>typedef enum {
981 ZSTD_dlm_byCopy = 0, </b>/**< Copy dictionary content internally */<b>
982 ZSTD_dlm_byRef = 1 </b>/**< Reference dictionary content -- the dictionary buffer must outlive its users. */<b>
983 } ZSTD_dictLoadMethod_e;
985 <pre><b>typedef enum {
986 ZSTD_f_zstd1 = 0, </b>/* zstd frame format, specified in zstd_compression_format.md (default) */<b>
987 ZSTD_f_zstd1_magicless = 1 </b>/* Variant of zstd frame format, without initial 4-bytes magic number.<b>
988 * Useful to save 4 bytes per generated frame.
989 * Decoder cannot recognise automatically this format, requiring this instruction. */
992 <pre><b>typedef enum {
993 </b>/* Note: this enum controls ZSTD_d_forceIgnoreChecksum */<b>
994 ZSTD_d_validateChecksum = 0,
995 ZSTD_d_ignoreChecksum = 1
996 } ZSTD_forceIgnoreChecksum_e;
998 <pre><b>typedef enum {
999 </b>/* Note: this enum and the behavior it controls are effectively internal<b>
1000 * implementation details of the compressor. They are expected to continue
1001 * to evolve and should be considered only in the context of extremely
1002 * advanced performance tuning.
1004 * Zstd currently supports the use of a CDict in three ways:
1006 * - The contents of the CDict can be copied into the working context. This
1007 * means that the compression can search both the dictionary and input
1008 * while operating on a single set of internal tables. This makes
1009 * the compression faster per-byte of input. However, the initial copy of
1010 * the CDict's tables incurs a fixed cost at the beginning of the
1011 * compression. For small compressions (< 8 KB), that copy can dominate
1012 * the cost of the compression.
1014 * - The CDict's tables can be used in-place. In this model, compression is
1015 * slower per input byte, because the compressor has to search two sets of
1016 * tables. However, this model incurs no start-up cost (as long as the
1017 * working context's tables can be reused). For small inputs, this can be
1018 * faster than copying the CDict's tables.
1020 * - The CDict's tables are not used at all, and instead we use the working
1021 * context alone to reload the dictionary and use params based on the source
1022 * size. See ZSTD_compress_insertDictionary() and ZSTD_compress_usingDict().
1023 * This method is effective when the dictionary sizes are very small relative
1024 * to the input size, and the input size is fairly large to begin with.
1026 * Zstd has a simple internal heuristic that selects which strategy to use
1027 * at the beginning of a compression. However, if experimentation shows that
1028 * Zstd is making poor choices, it is possible to override that choice with
1031 ZSTD_dictDefaultAttach = 0, </b>/* Use the default heuristic. */<b>
1032 ZSTD_dictForceAttach = 1, </b>/* Never copy the dictionary. */<b>
1033 ZSTD_dictForceCopy = 2, </b>/* Always copy the dictionary. */<b>
1034 ZSTD_dictForceLoad = 3 </b>/* Always reload the dictionary */<b>
1035 } ZSTD_dictAttachPref_e;
1037 <pre><b>typedef enum {
1038 ZSTD_lcm_auto = 0, </b>/**< Automatically determine the compression mode based on the compression level.<b>
1039 * Negative compression levels will be uncompressed, and positive compression
1040 * levels will be compressed. */
1041 ZSTD_lcm_huffman = 1, </b>/**< Always attempt Huffman compression. Uncompressed literals will still be<b>
1042 * emitted if Huffman compression is not profitable. */
1043 ZSTD_lcm_uncompressed = 2 </b>/**< Always emit uncompressed literals. */<b>
1044 } ZSTD_literalCompressionMode_e;
1046 <a name="Chapter15"></a><h2>Frame size functions</h2><pre></pre>
1048 <pre><b>unsigned long long ZSTD_findDecompressedSize(const void* src, size_t srcSize);
1049 </b><p> `src` should point to the start of a series of ZSTD encoded and/or skippable frames
1050 `srcSize` must be the _exact_ size of this series
1051 (i.e. there should be a frame boundary at `src + srcSize`)
1052 @return : - decompressed size of all data in all successive frames
1053 - if the decompressed size cannot be determined: ZSTD_CONTENTSIZE_UNKNOWN
1054 - if an error occurred: ZSTD_CONTENTSIZE_ERROR
1056 note 1 : decompressed size is an optional field, that may not be present, especially in streaming mode.
1057 When `return==ZSTD_CONTENTSIZE_UNKNOWN`, data to decompress could be any size.
1058 In which case, it's necessary to use streaming mode to decompress data.
1059 note 2 : decompressed size is always present when compression is done with ZSTD_compress()
1060 note 3 : decompressed size can be very large (64-bits value),
1061 potentially larger than what local system can handle as a single memory segment.
1062 In which case, it's necessary to use streaming mode to decompress data.
1063 note 4 : If source is untrusted, decompressed size could be wrong or intentionally modified.
1064 Always ensure result fits within application's authorized limits.
1065 Each application can set its own limits.
1066 note 5 : ZSTD_findDecompressedSize handles multiple frames, and so it must traverse the input to
1067 read each contained frame header. This is fast as most of the data is skipped,
1068 however it does mean that all frame data must be present and valid.
1071 <pre><b>unsigned long long ZSTD_decompressBound(const void* src, size_t srcSize);
1072 </b><p> `src` should point to the start of a series of ZSTD encoded and/or skippable frames
1073 `srcSize` must be the _exact_ size of this series
1074 (i.e. there should be a frame boundary at `src + srcSize`)
1075 @return : - upper-bound for the decompressed size of all data in all successive frames
1076 - if an error occured: ZSTD_CONTENTSIZE_ERROR
1078 note 1 : an error can occur if `src` contains an invalid or incorrectly formatted frame.
1079 note 2 : the upper-bound is exact when the decompressed size field is available in every ZSTD encoded frame of `src`.
1080 in this case, `ZSTD_findDecompressedSize` and `ZSTD_decompressBound` return the same value.
1081 note 3 : when the decompressed size field isn't available, the upper-bound for that frame is calculated by:
1082 upper-bound = # blocks * min(128 KB, Window_Size)
1086 <pre><b>size_t ZSTD_frameHeaderSize(const void* src, size_t srcSize);
1087 </b><p> srcSize must be >= ZSTD_FRAMEHEADERSIZE_PREFIX.
1088 @return : size of the Frame Header,
1089 or an error code (if srcSize is too small)
1092 <pre><b>typedef enum {
1093 ZSTD_sf_noBlockDelimiters = 0, </b>/* Representation of ZSTD_Sequence has no block delimiters, sequences only */<b>
1094 ZSTD_sf_explicitBlockDelimiters = 1 </b>/* Representation of ZSTD_Sequence contains explicit block delimiters */<b>
1095 } ZSTD_sequenceFormat_e;
1097 <pre><b></b><p> Generate sequences using ZSTD_compress2, given a source buffer.
1099 Each block will end with a dummy sequence
1100 with offset == 0, matchLength == 0, and litLength == length of last literals.
1101 litLength may be == 0, and if so, then the sequence of (of: 0 ml: 0 ll: 0)
1102 simply acts as a block delimiter.
1104 zc can be used to insert custom compression params.
1105 This function invokes ZSTD_compress2
1107 The output of this function can be fed into ZSTD_compressSequences() with CCtx
1108 setting of ZSTD_c_blockDelimiters as ZSTD_sf_explicitBlockDelimiters
1109 @return : number of sequences generated
1113 <pre><b>size_t ZSTD_mergeBlockDelimiters(ZSTD_Sequence* sequences, size_t seqsSize);
1114 </b><p> Given an array of ZSTD_Sequence, remove all sequences that represent block delimiters/last literals
1115 by merging them into into the literals of the next sequence.
1117 As such, the final generated result has no explicit representation of block boundaries,
1118 and the final last literals segment is not represented in the sequences.
1120 The output of this function can be fed into ZSTD_compressSequences() with CCtx
1121 setting of ZSTD_c_blockDelimiters as ZSTD_sf_noBlockDelimiters
1122 @return : number of sequences left after merging
1126 <pre><b>size_t ZSTD_compressSequences(ZSTD_CCtx* const cctx, void* dst, size_t dstSize,
1127 const ZSTD_Sequence* inSeqs, size_t inSeqsSize,
1128 const void* src, size_t srcSize);
1129 </b><p> Compress an array of ZSTD_Sequence, generated from the original source buffer, into dst.
1130 If a dictionary is included, then the cctx should reference the dict. (see: ZSTD_CCtx_refCDict(), ZSTD_CCtx_loadDictionary(), etc.)
1131 The entire source is compressed into a single frame.
1133 The compression behavior changes based on cctx params. In particular:
1134 If ZSTD_c_blockDelimiters == ZSTD_sf_noBlockDelimiters, the array of ZSTD_Sequence is expected to contain
1135 no block delimiters (defined in ZSTD_Sequence). Block boundaries are roughly determined based on
1136 the block size derived from the cctx, and sequences may be split. This is the default setting.
1138 If ZSTD_c_blockDelimiters == ZSTD_sf_explicitBlockDelimiters, the array of ZSTD_Sequence is expected to contain
1139 block delimiters (defined in ZSTD_Sequence). Behavior is undefined if no block delimiters are provided.
1141 If ZSTD_c_validateSequences == 0, this function will blindly accept the sequences provided. Invalid sequences cause undefined
1142 behavior. If ZSTD_c_validateSequences == 1, then if sequence is invalid (see doc/zstd_compression_format.md for
1143 specifics regarding offset/matchlength requirements) then the function will bail out and return an error.
1145 In addition to the two adjustable experimental params, there are other important cctx params.
1146 - ZSTD_c_minMatch MUST be set as less than or equal to the smallest match generated by the match finder. It has a minimum value of ZSTD_MINMATCH_MIN.
1147 - ZSTD_c_compressionLevel accordingly adjusts the strength of the entropy coder, as it would in typical compression.
1148 - ZSTD_c_windowLog affects offset validation: this function will return an error at higher debug levels if a provided offset
1149 is larger than what the spec allows for a given window log and dictionary (if present). See: doc/zstd_compression_format.md
1151 Note: Repcodes are, as of now, always re-calculated within this function, so ZSTD_Sequence::rep is unused.
1152 Note 2: Once we integrate ability to ingest repcodes, the explicit block delims mode must respect those repcodes exactly,
1153 and cannot emit an RLE block that disagrees with the repcode history
1154 @return : final compressed size or a ZSTD error.
1158 <a name="Chapter16"></a><h2>Memory management</h2><pre></pre>
1160 <pre><b>size_t ZSTD_estimateCCtxSize(int compressionLevel);
1161 size_t ZSTD_estimateCCtxSize_usingCParams(ZSTD_compressionParameters cParams);
1162 size_t ZSTD_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params);
1163 size_t ZSTD_estimateDCtxSize(void);
1164 </b><p> These functions make it possible to estimate memory usage
1165 of a future {D,C}Ctx, before its creation.
1167 ZSTD_estimateCCtxSize() will provide a memory budget large enough
1168 for any compression level up to selected one.
1169 Note : Unlike ZSTD_estimateCStreamSize*(), this estimate
1170 does not include space for a window buffer.
1171 Therefore, the estimation is only guaranteed for single-shot compressions, not streaming.
1172 The estimate will assume the input may be arbitrarily large,
1173 which is the worst case.
1175 When srcSize can be bound by a known and rather "small" value,
1176 this fact can be used to provide a tighter estimation
1177 because the CCtx compression context will need less memory.
1178 This tighter estimation can be provided by more advanced functions
1179 ZSTD_estimateCCtxSize_usingCParams(), which can be used in tandem with ZSTD_getCParams(),
1180 and ZSTD_estimateCCtxSize_usingCCtxParams(), which can be used in tandem with ZSTD_CCtxParams_setParameter().
1181 Both can be used to estimate memory using custom compression parameters and arbitrary srcSize limits.
1183 Note 2 : only single-threaded compression is supported.
1184 ZSTD_estimateCCtxSize_usingCCtxParams() will return an error code if ZSTD_c_nbWorkers is >= 1.
1188 <pre><b>size_t ZSTD_estimateCStreamSize(int compressionLevel);
1189 size_t ZSTD_estimateCStreamSize_usingCParams(ZSTD_compressionParameters cParams);
1190 size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params);
1191 size_t ZSTD_estimateDStreamSize(size_t windowSize);
1192 size_t ZSTD_estimateDStreamSize_fromFrame(const void* src, size_t srcSize);
1193 </b><p> ZSTD_estimateCStreamSize() will provide a budget large enough for any compression level up to selected one.
1194 It will also consider src size to be arbitrarily "large", which is worst case.
1195 If srcSize is known to always be small, ZSTD_estimateCStreamSize_usingCParams() can provide a tighter estimation.
1196 ZSTD_estimateCStreamSize_usingCParams() can be used in tandem with ZSTD_getCParams() to create cParams from compressionLevel.
1197 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.
1198 Note : CStream size estimation is only correct for single-threaded compression.
1199 ZSTD_DStream memory budget depends on window Size.
1200 This information can be passed manually, using ZSTD_estimateDStreamSize,
1201 or deducted from a valid frame Header, using ZSTD_estimateDStreamSize_fromFrame();
1202 Note : if streaming is init with function ZSTD_init?Stream_usingDict(),
1203 an internal ?Dict will be created, which additional size is not estimated here.
1204 In this case, get total size by adding ZSTD_estimate?DictSize
1207 <pre><b>size_t ZSTD_estimateCDictSize(size_t dictSize, int compressionLevel);
1208 size_t ZSTD_estimateCDictSize_advanced(size_t dictSize, ZSTD_compressionParameters cParams, ZSTD_dictLoadMethod_e dictLoadMethod);
1209 size_t ZSTD_estimateDDictSize(size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod);
1210 </b><p> ZSTD_estimateCDictSize() will bet that src size is relatively "small", and content is copied, like ZSTD_createCDict().
1211 ZSTD_estimateCDictSize_advanced() makes it possible to control compression parameters precisely, like ZSTD_createCDict_advanced().
1212 Note : dictionaries created by reference (`ZSTD_dlm_byRef`) are logically smaller.
1216 <pre><b>ZSTD_CCtx* ZSTD_initStaticCCtx(void* workspace, size_t workspaceSize);
1217 ZSTD_CStream* ZSTD_initStaticCStream(void* workspace, size_t workspaceSize); </b>/**< same as ZSTD_initStaticCCtx() */<b>
1218 </b><p> Initialize an object using a pre-allocated fixed-size buffer.
1219 workspace: The memory area to emplace the object into.
1220 Provided pointer *must be 8-bytes aligned*.
1221 Buffer must outlive object.
1222 workspaceSize: Use ZSTD_estimate*Size() to determine
1223 how large workspace must be to support target scenario.
1224 @return : pointer to object (same address as workspace, just different type),
1225 or NULL if error (size too small, incorrect alignment, etc.)
1226 Note : zstd will never resize nor malloc() when using a static buffer.
1227 If the object requires more memory than available,
1228 zstd will just error out (typically ZSTD_error_memory_allocation).
1229 Note 2 : there is no corresponding "free" function.
1230 Since workspace is allocated externally, it must be freed externally too.
1231 Note 3 : cParams : use ZSTD_getCParams() to convert a compression level
1232 into its associated cParams.
1233 Limitation 1 : currently not compatible with internal dictionary creation, triggered by
1234 ZSTD_CCtx_loadDictionary(), ZSTD_initCStream_usingDict() or ZSTD_initDStream_usingDict().
1235 Limitation 2 : static cctx currently not compatible with multi-threading.
1236 Limitation 3 : static dctx is incompatible with legacy support.
1240 <pre><b>ZSTD_DStream* ZSTD_initStaticDStream(void* workspace, size_t workspaceSize); </b>/**< same as ZSTD_initStaticDCtx() */<b>
1242 <pre><b>typedef void* (*ZSTD_allocFunction) (void* opaque, size_t size);
1243 typedef void (*ZSTD_freeFunction) (void* opaque, void* address);
1244 typedef struct { ZSTD_allocFunction customAlloc; ZSTD_freeFunction customFree; void* opaque; } ZSTD_customMem;
1247 __attribute__((__unused__))
1249 ZSTD_customMem const ZSTD_defaultCMem = { NULL, NULL, NULL }; </b>/**< this constant defers to stdlib's functions */<b>
1250 </b><p> These prototypes make it possible to pass your own allocation/free functions.
1251 ZSTD_customMem is provided at creation time, using ZSTD_create*_advanced() variants listed below.
1252 All allocation/free operations will be completed using these custom variants instead of regular <stdlib.h> ones.
1256 <a name="Chapter17"></a><h2>Advanced compression functions</h2><pre></pre>
1258 <pre><b>ZSTD_CDict* ZSTD_createCDict_byReference(const void* dictBuffer, size_t dictSize, int compressionLevel);
1259 </b><p> Create a digested dictionary for compression
1260 Dictionary content is just referenced, not duplicated.
1261 As a consequence, `dictBuffer` **must** outlive CDict,
1262 and its content must remain unmodified throughout the lifetime of CDict.
1263 note: equivalent to ZSTD_createCDict_advanced(), with dictLoadMethod==ZSTD_dlm_byRef
1266 <pre><b>unsigned ZSTD_getDictID_fromCDict(const ZSTD_CDict* cdict);
1267 </b><p> Provides the dictID of the dictionary loaded into `cdict`.
1268 If @return == 0, the dictionary is not conformant to Zstandard specification, or empty.
1269 Non-conformant dictionaries can still be loaded, but as content-only dictionaries.
1272 <pre><b>ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize);
1273 </b><p> @return ZSTD_compressionParameters structure for a selected compression level and estimated srcSize.
1274 `estimatedSrcSize` value is optional, select 0 if not known
1277 <pre><b>ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize);
1278 </b><p> same as ZSTD_getCParams(), but @return a full `ZSTD_parameters` object instead of sub-component `ZSTD_compressionParameters`.
1279 All fields of `ZSTD_frameParameters` are set to default : contentSize=1, checksum=0, noDictID=0
1282 <pre><b>size_t ZSTD_checkCParams(ZSTD_compressionParameters params);
1283 </b><p> Ensure param values remain within authorized range.
1284 @return 0 on success, or an error code (can be checked with ZSTD_isError())
1287 <pre><b>ZSTD_compressionParameters ZSTD_adjustCParams(ZSTD_compressionParameters cPar, unsigned long long srcSize, size_t dictSize);
1288 </b><p> optimize params for a given `srcSize` and `dictSize`.
1289 `srcSize` can be unknown, in which case use ZSTD_CONTENTSIZE_UNKNOWN.
1290 `dictSize` must be `0` when there is no dictionary.
1291 cPar can be invalid : all parameters will be clamped within valid range in the @return struct.
1292 This function never fails (wide contract)
1295 <pre><b>size_t ZSTD_compress_advanced(ZSTD_CCtx* cctx,
1296 void* dst, size_t dstCapacity,
1297 const void* src, size_t srcSize,
1298 const void* dict,size_t dictSize,
1299 ZSTD_parameters params);
1300 </b><p> Note : this function is now DEPRECATED.
1301 It can be replaced by ZSTD_compress2(), in combination with ZSTD_CCtx_setParameter() and other parameter setters.
1302 This prototype will be marked as deprecated and generate compilation warning on reaching v1.5.x
1305 <pre><b>size_t ZSTD_compress_usingCDict_advanced(ZSTD_CCtx* cctx,
1306 void* dst, size_t dstCapacity,
1307 const void* src, size_t srcSize,
1308 const ZSTD_CDict* cdict,
1309 ZSTD_frameParameters fParams);
1310 </b><p> Note : this function is now REDUNDANT.
1311 It can be replaced by ZSTD_compress2(), in combination with ZSTD_CCtx_loadDictionary() and other parameter setters.
1312 This prototype will be marked as deprecated and generate compilation warning in some future version
1315 <pre><b>size_t ZSTD_CCtx_loadDictionary_byReference(ZSTD_CCtx* cctx, const void* dict, size_t dictSize);
1316 </b><p> Same as ZSTD_CCtx_loadDictionary(), but dictionary content is referenced, instead of being copied into CCtx.
1317 It saves some memory, but also requires that `dict` outlives its usage within `cctx`
1320 <pre><b>size_t ZSTD_CCtx_loadDictionary_advanced(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictContentType_e dictContentType);
1321 </b><p> Same as ZSTD_CCtx_loadDictionary(), but gives finer control over
1322 how to load the dictionary (by copy ? by reference ?)
1323 and how to interpret it (automatic ? force raw mode ? full mode only ?)
1326 <pre><b>size_t ZSTD_CCtx_refPrefix_advanced(ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType);
1327 </b><p> Same as ZSTD_CCtx_refPrefix(), but gives finer control over
1328 how to interpret prefix content (automatic ? force raw mode (default) ? full mode only ?)
1331 <pre><b>size_t ZSTD_CCtx_getParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int* value);
1332 </b><p> Get the requested compression parameter value, selected by enum ZSTD_cParameter,
1333 and store it into int* value.
1334 @return : 0, or an error code (which can be tested with ZSTD_isError()).
1338 <pre><b>ZSTD_CCtx_params* ZSTD_createCCtxParams(void);
1339 size_t ZSTD_freeCCtxParams(ZSTD_CCtx_params* params);
1340 </b><p> Quick howto :
1341 - ZSTD_createCCtxParams() : Create a ZSTD_CCtx_params structure
1342 - ZSTD_CCtxParams_setParameter() : Push parameters one by one into
1343 an existing ZSTD_CCtx_params structure.
1345 ZSTD_CCtx_setParameter().
1346 - ZSTD_CCtx_setParametersUsingCCtxParams() : Apply parameters to
1348 These parameters will be applied to
1349 all subsequent frames.
1350 - ZSTD_compressStream2() : Do compression using the CCtx.
1351 - ZSTD_freeCCtxParams() : Free the memory.
1353 This can be used with ZSTD_estimateCCtxSize_advanced_usingCCtxParams()
1354 for static allocation of CCtx for single-threaded compression.
1358 <pre><b>size_t ZSTD_CCtxParams_reset(ZSTD_CCtx_params* params);
1359 </b><p> Reset params to default values.
1363 <pre><b>size_t ZSTD_CCtxParams_init(ZSTD_CCtx_params* cctxParams, int compressionLevel);
1364 </b><p> Initializes the compression parameters of cctxParams according to
1365 compression level. All other parameters are reset to their default values.
1369 <pre><b>size_t ZSTD_CCtxParams_init_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_parameters params);
1370 </b><p> Initializes the compression and frame parameters of cctxParams according to
1371 params. All other parameters are reset to their default values.
1375 <pre><b>size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int value);
1376 </b><p> Similar to ZSTD_CCtx_setParameter.
1377 Set one compression parameter, selected by enum ZSTD_cParameter.
1378 Parameters must be applied to a ZSTD_CCtx using
1379 ZSTD_CCtx_setParametersUsingCCtxParams().
1380 @result : a code representing success or failure (which can be tested with
1385 <pre><b>size_t ZSTD_CCtxParams_getParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int* value);
1386 </b><p> Similar to ZSTD_CCtx_getParameter.
1387 Get the requested value of one compression parameter, selected by enum ZSTD_cParameter.
1388 @result : 0, or an error code (which can be tested with ZSTD_isError()).
1392 <pre><b>size_t ZSTD_CCtx_setParametersUsingCCtxParams(
1393 ZSTD_CCtx* cctx, const ZSTD_CCtx_params* params);
1394 </b><p> Apply a set of ZSTD_CCtx_params to the compression context.
1395 This can be done even after compression is started,
1396 if nbWorkers==0, this will have no impact until a new compression is started.
1397 if nbWorkers>=1, new parameters will be picked up at next job,
1398 with a few restrictions (windowLog, pledgedSrcSize, nbWorkers, jobSize, and overlapLog are not updated).
1402 <pre><b>size_t ZSTD_compressStream2_simpleArgs (
1404 void* dst, size_t dstCapacity, size_t* dstPos,
1405 const void* src, size_t srcSize, size_t* srcPos,
1406 ZSTD_EndDirective endOp);
1407 </b><p> Same as ZSTD_compressStream2(),
1408 but using only integral types as arguments.
1409 This variant might be helpful for binders from dynamic languages
1410 which have troubles handling structures containing memory pointers.
1414 <a name="Chapter18"></a><h2>Advanced decompression functions</h2><pre></pre>
1416 <pre><b>unsigned ZSTD_isFrame(const void* buffer, size_t size);
1417 </b><p> Tells if the content of `buffer` starts with a valid Frame Identifier.
1418 Note : Frame Identifier is 4 bytes. If `size < 4`, @return will always be 0.
1419 Note 2 : Legacy Frame Identifiers are considered valid only if Legacy Support is enabled.
1420 Note 3 : Skippable Frame Identifiers are considered valid.
1423 <pre><b>ZSTD_DDict* ZSTD_createDDict_byReference(const void* dictBuffer, size_t dictSize);
1424 </b><p> Create a digested dictionary, ready to start decompression operation without startup delay.
1425 Dictionary content is referenced, and therefore stays in dictBuffer.
1426 It is important that dictBuffer outlives DDict,
1427 it must remain read accessible throughout the lifetime of DDict
1430 <pre><b>size_t ZSTD_DCtx_loadDictionary_byReference(ZSTD_DCtx* dctx, const void* dict, size_t dictSize);
1431 </b><p> Same as ZSTD_DCtx_loadDictionary(),
1432 but references `dict` content instead of copying it into `dctx`.
1433 This saves memory if `dict` remains around.,
1434 However, it's imperative that `dict` remains accessible (and unmodified) while being used, so it must outlive decompression.
1437 <pre><b>size_t ZSTD_DCtx_loadDictionary_advanced(ZSTD_DCtx* dctx, const void* dict, size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictContentType_e dictContentType);
1438 </b><p> Same as ZSTD_DCtx_loadDictionary(),
1439 but gives direct control over
1440 how to load the dictionary (by copy ? by reference ?)
1441 and how to interpret it (automatic ? force raw mode ? full mode only ?).
1444 <pre><b>size_t ZSTD_DCtx_refPrefix_advanced(ZSTD_DCtx* dctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType);
1445 </b><p> Same as ZSTD_DCtx_refPrefix(), but gives finer control over
1446 how to interpret prefix content (automatic ? force raw mode (default) ? full mode only ?)
1449 <pre><b>size_t ZSTD_DCtx_setMaxWindowSize(ZSTD_DCtx* dctx, size_t maxWindowSize);
1450 </b><p> Refuses allocating internal buffers for frames requiring a window size larger than provided limit.
1451 This protects a decoder context from reserving too much memory for itself (potential attack scenario).
1452 This parameter is only useful in streaming mode, since no internal buffer is allocated in single-pass mode.
1453 By default, a decompression context accepts all window sizes <= (1 << ZSTD_WINDOWLOG_LIMIT_DEFAULT)
1454 @return : 0, or an error code (which can be tested using ZSTD_isError()).
1458 <pre><b>size_t ZSTD_DCtx_getParameter(ZSTD_DCtx* dctx, ZSTD_dParameter param, int* value);
1459 </b><p> Get the requested decompression parameter value, selected by enum ZSTD_dParameter,
1460 and store it into int* value.
1461 @return : 0, or an error code (which can be tested with ZSTD_isError()).
1465 <pre><b>size_t ZSTD_DCtx_setFormat(ZSTD_DCtx* dctx, ZSTD_format_e format);
1466 </b><p> Instruct the decoder context about what kind of data to decode next.
1467 This instruction is mandatory to decode data without a fully-formed header,
1468 such ZSTD_f_zstd1_magicless for example.
1469 @return : 0, or an error code (which can be tested using ZSTD_isError()).
1472 <pre><b>size_t ZSTD_decompressStream_simpleArgs (
1474 void* dst, size_t dstCapacity, size_t* dstPos,
1475 const void* src, size_t srcSize, size_t* srcPos);
1476 </b><p> Same as ZSTD_decompressStream(),
1477 but using only integral types as arguments.
1478 This can be helpful for binders from dynamic languages
1479 which have troubles handling structures containing memory pointers.
1483 <a name="Chapter19"></a><h2>Advanced streaming functions</h2><pre> Warning : most of these functions are now redundant with the Advanced API.
1484 Once Advanced API reaches "stable" status,
1485 redundant functions will be deprecated, and then at some point removed.
1488 <h3>Advanced Streaming compression functions</h3><pre></pre><b><pre></pre></b><BR>
1490 ZSTD_initCStream_srcSize(ZSTD_CStream* zcs,
1491 int compressionLevel,
1492 unsigned long long pledgedSrcSize);
1493 </b><p> This function is deprecated, and equivalent to:
1494 ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
1495 ZSTD_CCtx_refCDict(zcs, NULL); // clear the dictionary (if any)
1496 ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel);
1497 ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
1499 pledgedSrcSize must be correct. If it is not known at init time, use
1500 ZSTD_CONTENTSIZE_UNKNOWN. Note that, for compatibility with older programs,
1501 "0" also disables frame content size field. It may be enabled in the future.
1502 Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
1507 ZSTD_initCStream_usingDict(ZSTD_CStream* zcs,
1508 const void* dict, size_t dictSize,
1509 int compressionLevel);
1510 </b><p> This function is deprecated, and is equivalent to:
1511 ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
1512 ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel);
1513 ZSTD_CCtx_loadDictionary(zcs, dict, dictSize);
1515 Creates of an internal CDict (incompatible with static CCtx), except if
1516 dict == NULL or dictSize < 8, in which case no dict is used.
1517 Note: dict is loaded with ZSTD_dct_auto (treated as a full zstd dictionary if
1518 it begins with ZSTD_MAGIC_DICTIONARY, else as raw content) and ZSTD_dlm_byCopy.
1519 Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
1524 ZSTD_initCStream_advanced(ZSTD_CStream* zcs,
1525 const void* dict, size_t dictSize,
1526 ZSTD_parameters params,
1527 unsigned long long pledgedSrcSize);
1528 </b><p> This function is deprecated, and is approximately equivalent to:
1529 ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
1530 // Pseudocode: Set each zstd parameter and leave the rest as-is.
1531 for ((param, value) : params) {
1532 ZSTD_CCtx_setParameter(zcs, param, value);
1534 ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
1535 ZSTD_CCtx_loadDictionary(zcs, dict, dictSize);
1537 dict is loaded with ZSTD_dct_auto and ZSTD_dlm_byCopy.
1538 pledgedSrcSize must be correct.
1539 If srcSize is not known at init time, use value ZSTD_CONTENTSIZE_UNKNOWN.
1540 Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
1544 <pre><b>size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict);
1545 </b><p> This function is deprecated, and equivalent to:
1546 ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
1547 ZSTD_CCtx_refCDict(zcs, cdict);
1549 note : cdict will just be referenced, and must outlive compression session
1550 Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
1555 ZSTD_initCStream_usingCDict_advanced(ZSTD_CStream* zcs,
1556 const ZSTD_CDict* cdict,
1557 ZSTD_frameParameters fParams,
1558 unsigned long long pledgedSrcSize);
1559 </b><p> This function is DEPRECATED, and is approximately equivalent to:
1560 ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
1561 // Pseudocode: Set each zstd frame parameter and leave the rest as-is.
1562 for ((fParam, value) : fParams) {
1563 ZSTD_CCtx_setParameter(zcs, fParam, value);
1565 ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
1566 ZSTD_CCtx_refCDict(zcs, cdict);
1568 same as ZSTD_initCStream_usingCDict(), with control over frame parameters.
1569 pledgedSrcSize must be correct. If srcSize is not known at init time, use
1570 value ZSTD_CONTENTSIZE_UNKNOWN.
1571 Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
1575 <pre><b>size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize);
1576 </b><p> This function is deprecated, and is equivalent to:
1577 ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
1578 ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
1580 start a new frame, using same parameters from previous frame.
1581 This is typically useful to skip dictionary loading stage, since it will re-use it in-place.
1582 Note that zcs must be init at least once before using ZSTD_resetCStream().
1583 If pledgedSrcSize is not known at reset time, use macro ZSTD_CONTENTSIZE_UNKNOWN.
1584 If pledgedSrcSize > 0, its value must be correct, as it will be written in header, and controlled at the end.
1585 For the time being, pledgedSrcSize==0 is interpreted as "srcSize unknown" for compatibility with older programs,
1586 but it will change to mean "empty" in future version, so use macro ZSTD_CONTENTSIZE_UNKNOWN instead.
1587 @return : 0, or an error code (which can be tested using ZSTD_isError())
1588 Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
1592 <pre><b>typedef struct {
1593 unsigned long long ingested; </b>/* nb input bytes read and buffered */<b>
1594 unsigned long long consumed; </b>/* nb input bytes actually compressed */<b>
1595 unsigned long long produced; </b>/* nb of compressed bytes generated and buffered */<b>
1596 unsigned long long flushed; </b>/* nb of compressed bytes flushed : not provided; can be tracked from caller side */<b>
1597 unsigned currentJobID; </b>/* MT only : latest started job nb */<b>
1598 unsigned nbActiveWorkers; </b>/* MT only : nb of workers actively compressing at probe time */<b>
1599 } ZSTD_frameProgression;
1601 <pre><b>size_t ZSTD_toFlushNow(ZSTD_CCtx* cctx);
1602 </b><p> Tell how many bytes are ready to be flushed immediately.
1603 Useful for multithreading scenarios (nbWorkers >= 1).
1604 Probe the oldest active job, defined as oldest job not yet entirely flushed,
1605 and check its output buffer.
1606 @return : amount of data stored in oldest job and ready to be flushed immediately.
1607 if @return == 0, it means either :
1608 + there is no active job (could be checked with ZSTD_frameProgression()), or
1609 + oldest job is still actively compressing data,
1610 but everything it has produced has also been flushed so far,
1611 therefore flush speed is limited by production speed of oldest job
1612 irrespective of the speed of concurrent (and newer) jobs.
1616 <h3>Advanced Streaming decompression functions</h3><pre></pre><b><pre></pre></b><BR>
1617 <pre><b>size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize);
1619 ZSTD_DCtx_reset(zds, ZSTD_reset_session_only);
1620 ZSTD_DCtx_loadDictionary(zds, dict, dictSize);
1622 note: no dictionary will be used if dict == NULL or dictSize < 8
1623 Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
1627 <pre><b>size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* zds, const ZSTD_DDict* ddict);
1629 ZSTD_DCtx_reset(zds, ZSTD_reset_session_only);
1630 ZSTD_DCtx_refDDict(zds, ddict);
1632 note : ddict is referenced, it must outlive decompression session
1633 Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
1637 <pre><b>size_t ZSTD_resetDStream(ZSTD_DStream* zds);
1639 ZSTD_DCtx_reset(zds, ZSTD_reset_session_only);
1641 re-use decompression parameters from previous init; saves dictionary loading
1642 Note : this prototype will be marked as deprecated and generate compilation warnings on reaching v1.5.x
1646 <a name="Chapter20"></a><h2>Buffer-less and synchronous inner streaming functions</h2><pre>
1647 This is an advanced API, giving full control over buffer management, for users which need direct control over memory.
1648 But it's also a complex one, with several restrictions, documented below.
1649 Prefer normal streaming API for an easier experience.
1653 <a name="Chapter21"></a><h2>Buffer-less streaming compression (synchronous mode)</h2><pre>
1654 A ZSTD_CCtx object is required to track streaming operations.
1655 Use ZSTD_createCCtx() / ZSTD_freeCCtx() to manage resource.
1656 ZSTD_CCtx object can be re-used multiple times within successive compression operations.
1658 Start by initializing a context.
1659 Use ZSTD_compressBegin(), or ZSTD_compressBegin_usingDict() for dictionary compression,
1660 or ZSTD_compressBegin_advanced(), for finer parameter control.
1661 It's also possible to duplicate a reference context which has already been initialized, using ZSTD_copyCCtx()
1663 Then, consume your input using ZSTD_compressContinue().
1664 There are some important considerations to keep in mind when using this advanced function :
1665 - ZSTD_compressContinue() has no internal buffer. It uses externally provided buffers only.
1666 - Interface is synchronous : input is consumed entirely and produces 1+ compressed blocks.
1667 - Caller must ensure there is enough space in `dst` to store compressed data under worst case scenario.
1668 Worst case evaluation is provided by ZSTD_compressBound().
1669 ZSTD_compressContinue() doesn't guarantee recover after a failed compression.
1670 - ZSTD_compressContinue() presumes prior input ***is still accessible and unmodified*** (up to maximum distance size, see WindowLog).
1671 It remembers all previous contiguous blocks, plus one separated memory segment (which can itself consists of multiple contiguous blocks)
1672 - ZSTD_compressContinue() detects that prior input has been overwritten when `src` buffer overlaps.
1673 In which case, it will "discard" the relevant memory section from its history.
1675 Finish a frame with ZSTD_compressEnd(), which will write the last block(s) and optional checksum.
1676 It's possible to use srcSize==0, in which case, it will write a final empty block to end the frame.
1677 Without last block mark, frames are considered unfinished (hence corrupted) by compliant decoders.
1679 `ZSTD_CCtx` object can be re-used (ZSTD_compressBegin()) to compress again.
1682 <h3>Buffer-less streaming compression functions</h3><pre></pre><b><pre>size_t ZSTD_compressBegin(ZSTD_CCtx* cctx, int compressionLevel);
1683 size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, int compressionLevel);
1684 size_t ZSTD_compressBegin_advanced(ZSTD_CCtx* cctx, const void* dict, size_t dictSize, ZSTD_parameters params, unsigned long long pledgedSrcSize); </b>/**< pledgedSrcSize : If srcSize is not known at init time, use ZSTD_CONTENTSIZE_UNKNOWN */<b>
1685 size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict); </b>/**< note: fails if cdict==NULL */<b>
1686 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 must be correct. If srcSize is not known, use macro ZSTD_CONTENTSIZE_UNKNOWN */<b>
1687 size_t ZSTD_copyCCtx(ZSTD_CCtx* cctx, const ZSTD_CCtx* preparedCCtx, unsigned long long pledgedSrcSize); </b>/**< note: if pledgedSrcSize is not known, use ZSTD_CONTENTSIZE_UNKNOWN */<b>
1689 <a name="Chapter22"></a><h2>Buffer-less streaming decompression (synchronous mode)</h2><pre>
1690 A ZSTD_DCtx object is required to track streaming operations.
1691 Use ZSTD_createDCtx() / ZSTD_freeDCtx() to manage it.
1692 A ZSTD_DCtx object can be re-used multiple times.
1694 First typical operation is to retrieve frame parameters, using ZSTD_getFrameHeader().
1695 Frame header is extracted from the beginning of compressed frame, so providing only the frame's beginning is enough.
1696 Data fragment must be large enough to ensure successful decoding.
1697 `ZSTD_frameHeaderSize_max` bytes is guaranteed to always be large enough.
1698 @result : 0 : successful decoding, the `ZSTD_frameHeader` structure is correctly filled.
1699 >0 : `srcSize` is too small, please provide at least @result bytes on next attempt.
1700 errorCode, which can be tested using ZSTD_isError().
1702 It fills a ZSTD_frameHeader structure with important information to correctly decode the frame,
1703 such as the dictionary ID, content size, or maximum back-reference distance (`windowSize`).
1704 Note that these values could be wrong, either because of data corruption, or because a 3rd party deliberately spoofs false information.
1705 As a consequence, check that values remain within valid application range.
1706 For example, do not allocate memory blindly, check that `windowSize` is within expectation.
1707 Each application can set its own limits, depending on local restrictions.
1708 For extended interoperability, it is recommended to support `windowSize` of at least 8 MB.
1710 ZSTD_decompressContinue() needs previous data blocks during decompression, up to `windowSize` bytes.
1711 ZSTD_decompressContinue() is very sensitive to contiguity,
1712 if 2 blocks don't follow each other, make sure that either the compressor breaks contiguity at the same place,
1713 or that previous contiguous segment is large enough to properly handle maximum back-reference distance.
1714 There are multiple ways to guarantee this condition.
1716 The most memory efficient way is to use a round buffer of sufficient size.
1717 Sufficient size is determined by invoking ZSTD_decodingBufferSize_min(),
1718 which can @return an error code if required value is too large for current system (in 32-bits mode).
1719 In a round buffer methodology, ZSTD_decompressContinue() decompresses each block next to previous one,
1720 up to the moment there is not enough room left in the buffer to guarantee decoding another full block,
1721 which maximum size is provided in `ZSTD_frameHeader` structure, field `blockSizeMax`.
1722 At which point, decoding can resume from the beginning of the buffer.
1723 Note that already decoded data stored in the buffer should be flushed before being overwritten.
1725 There are alternatives possible, for example using two or more buffers of size `windowSize` each, though they consume more memory.
1727 Finally, if you control the compression process, you can also ignore all buffer size rules,
1728 as long as the encoder and decoder progress in "lock-step",
1729 aka use exactly the same buffer sizes, break contiguity at the same place, etc.
1731 Once buffers are setup, start decompression, with ZSTD_decompressBegin().
1732 If decompression requires a dictionary, use ZSTD_decompressBegin_usingDict() or ZSTD_decompressBegin_usingDDict().
1734 Then use ZSTD_nextSrcSizeToDecompress() and ZSTD_decompressContinue() alternatively.
1735 ZSTD_nextSrcSizeToDecompress() tells how many bytes to provide as 'srcSize' to ZSTD_decompressContinue().
1736 ZSTD_decompressContinue() requires this _exact_ amount of bytes, or it will fail.
1738 @result of ZSTD_decompressContinue() is the number of bytes regenerated within 'dst' (necessarily <= dstCapacity).
1739 It can be zero : it just means ZSTD_decompressContinue() has decoded some metadata item.
1740 It can also be an error code, which can be tested with ZSTD_isError().
1742 A frame is fully decoded when ZSTD_nextSrcSizeToDecompress() returns zero.
1743 Context can then be reset to start a new decompression.
1745 Note : it's possible to know if next input to present is a header or a block, using ZSTD_nextInputType().
1746 This information is not required to properly decode a frame.
1748 == Special case : skippable frames
1750 Skippable frames allow integration of user-defined data into a flow of concatenated frames.
1751 Skippable frames will be ignored (skipped) by decompressor.
1752 The format of skippable frames is as follows :
1753 a) Skippable frame ID - 4 Bytes, Little endian format, any value from 0x184D2A50 to 0x184D2A5F
1754 b) Frame Size - 4 Bytes, Little endian format, unsigned 32-bits
1755 c) Frame Content - any content (User Data) of length equal to Frame Size
1756 For skippable frames ZSTD_getFrameHeader() returns zfhPtr->frameType==ZSTD_skippableFrame.
1757 For skippable frames ZSTD_decompressContinue() always returns 0 : it only skips the content.
1760 <h3>Buffer-less streaming decompression functions</h3><pre></pre><b><pre>typedef enum { ZSTD_frame, ZSTD_skippableFrame } ZSTD_frameType_e;
1762 unsigned long long frameContentSize; </b>/* if == ZSTD_CONTENTSIZE_UNKNOWN, it means this field is not available. 0 means "empty" */<b>
1763 unsigned long long windowSize; </b>/* can be very large, up to <= frameContentSize */<b>
1764 unsigned blockSizeMax;
1765 ZSTD_frameType_e frameType; </b>/* if == ZSTD_skippableFrame, frameContentSize is the size of skippable content */<b>
1766 unsigned headerSize;
1768 unsigned checksumFlag;
1771 <pre><b>size_t ZSTD_getFrameHeader(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize); </b>/**< doesn't consume input */<b>
1772 </b>/*! ZSTD_getFrameHeader_advanced() :<b>
1773 * same as ZSTD_getFrameHeader(),
1774 * with added capability to select a format (like ZSTD_f_zstd1_magicless) */
1775 size_t ZSTD_getFrameHeader_advanced(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize, ZSTD_format_e format);
1776 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>
1777 </b><p> decode Frame Header, or requires larger `srcSize`.
1778 @return : 0, `zfhPtr` is correctly filled,
1779 >0, `srcSize` is too small, value is wanted `srcSize` amount,
1780 or an error code, which can be tested using ZSTD_isError()
1783 <pre><b>typedef enum { ZSTDnit_frameHeader, ZSTDnit_blockHeader, ZSTDnit_block, ZSTDnit_lastBlock, ZSTDnit_checksum, ZSTDnit_skippableFrame } ZSTD_nextInputType_e;
1785 <a name="Chapter23"></a><h2>Block level API</h2><pre></pre>
1787 <pre><b></b><p> Frame metadata cost is typically ~12 bytes, which can be non-negligible for very small blocks (< 100 bytes).
1788 But users will have to take in charge needed metadata to regenerate data, such as compressed and content sizes.
1790 A few rules to respect :
1791 - Compressing and decompressing require a context structure
1792 + Use ZSTD_createCCtx() and ZSTD_createDCtx()
1793 - It is necessary to init context before starting
1794 + compression : any ZSTD_compressBegin*() variant, including with dictionary
1795 + decompression : any ZSTD_decompressBegin*() variant, including with dictionary
1796 + copyCCtx() and copyDCtx() can be used too
1797 - Block size is limited, it must be <= ZSTD_getBlockSize() <= ZSTD_BLOCKSIZE_MAX == 128 KB
1798 + If input is larger than a block size, it's necessary to split input data into multiple blocks
1799 + For inputs larger than a single block, consider using regular ZSTD_compress() instead.
1800 Frame metadata is not that costly, and quickly becomes negligible as source size grows larger than a block.
1801 - When a block is considered not compressible enough, ZSTD_compressBlock() result will be 0 (zero) !
1802 ===> In which case, nothing is produced into `dst` !
1803 + User __must__ test for such outcome and deal directly with uncompressed data
1804 + A block cannot be declared incompressible if ZSTD_compressBlock() return value was != 0.
1805 Doing so would mess up with statistics history, leading to potential data corruption.
1806 + ZSTD_decompressBlock() _doesn't accept uncompressed data as input_ !!
1807 + In case of multiple successive blocks, should some of them be uncompressed,
1808 decoder must be informed of their existence in order to follow proper history.
1809 Use ZSTD_insertBlock() for such a case.
1812 <h3>Raw zstd block functions</h3><pre></pre><b><pre>size_t ZSTD_getBlockSize (const ZSTD_CCtx* cctx);
1813 size_t ZSTD_compressBlock (ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
1814 size_t ZSTD_decompressBlock(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize);
1815 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>