1v1.3.8 2perf: better decompression speed on large files (+7%) and cold dictionaries (+15%) 3api : supports decompressing empty frames into NULL (used to be an error) (#1385) 4api : fix : ZSTD_dictForceCopy advanced command, by @felixhandte 5build: fixed zstd-pgo, by @bmwiedemann 6doc : clarified zstd_compression_format.md, by @ulikunitz 7 8v1.3.7 9perf: slightly better decompression speed on clang (depending on hardware target) 10fix : performance of dictionary compression for small input < 4 KB at levels 9 and 10 11build: no longer build backtrace by default in release mode; restrict further automatic mode 12build: control backtrace support through build macro BACKTRACE 13misc: added man pages for zstdless and zstdgrep, by @samrussell 14 15v1.3.6 16perf: much faster dictionary builder, by @jenniferliu 17perf: faster dictionary compression on small data when using multiple contexts, by @felixhandte 18perf: faster dictionary decompression when using a very large number of dictionaries simultaneously 19cli : fix : does no longer overwrite destination when source does not exist (#1082) 20cli : new command --adapt, for automatic compression level adaptation 21api : fix : block api can be streamed with > 4 GB, reported by @catid 22api : reduced ZSTD_DDict size by 2 KB 23api : minimum negative compression level is defined, and can be queried using ZSTD_minCLevel(). 24build: support Haiku target, by @korli 25build: Read Legacy format is limited to v0.5+ by default. Can be changed at compile time with macro ZSTD_LEGACY_SUPPORT. 26doc : zstd_compression_format.md updated to match wording in IETF RFC 8478 27misc: tests/paramgrill, a parameter optimizer, by @GeorgeLu97 28 29v1.3.5 30perf: much faster dictionary compression, by @felixhandte 31perf: small quality improvement for dictionary generation, by @terrelln 32perf: slightly improved high compression levels (notably level 19) 33mem : automatic memory release for long duration contexts 34cli : fix : overlapLog can be manually set 35cli : fix : decoding invalid lz4 frames 36api : fix : performance degradation for dictionary compression when using advanced API, by @terrelln 37api : change : clarify ZSTD_CCtx_reset() vs ZSTD_CCtx_resetParameters(), by @terrelln 38build: select custom libzstd scope through control macros, by @GeorgeLu97 39build: OpenBSD patch, by @bket 40build: make and make all are compatible with -j 41doc : clarify zstd_compression_format.md, updated for IETF RFC process 42misc: pzstd compatible with reproducible compilation, by @lamby 43 44v1.3.4 45perf: faster speed (especially decoding speed) on recent cpus (haswell+) 46perf: much better performance associating --long with multi-threading, by @terrelln 47perf: better compression at levels 13-15 48cli : asynchronous compression by default, for faster experience (use --single-thread for former behavior) 49cli : smoother status report in multi-threading mode 50cli : added command --fast=#, for faster compression modes 51cli : fix crash when not overwriting existing files, by Pádraig Brady (@pixelb) 52api : `nbThreads` becomes `nbWorkers` : 1 triggers asynchronous mode 53api : compression levels can be negative, for even more speed 54api : ZSTD_getFrameProgression() : get precise progress status of ZSTDMT anytime 55api : ZSTDMT can accept new compression parameters during compression 56api : implemented all advanced dictionary decompression prototypes 57build: improved meson recipe, by Shawn Landden (@shawnl) 58build: VS2017 scripts, by @HaydnTrigg 59misc: all /contrib projects fixed 60misc: added /contrib/docker script by @gyscos 61 62v1.3.3 63perf: faster zstd_opt strategy (levels 16-19) 64fix : bug #944 : multithreading with shared ditionary and large data, reported by @gsliepen 65cli : fix : content size written in header by default 66cli : fix : improved LZ4 format support, by @felixhandte 67cli : new : hidden command `-S`, to benchmark multiple files while generating one result per file 68api : fix : support large skippable frames, by @terrelln 69api : fix : streaming interface was adding a useless 3-bytes null block to small frames 70api : change : when setting `pledgedSrcSize`, use `ZSTD_CONTENTSIZE_UNKNOWN` macro value to mean "unknown" 71build: fix : compilation under rhel6 and centos6, reported by @pixelb 72build: added `check` target 73 74v1.3.2 75new : long range mode, using --long command, by Stella Lau (@stellamplau) 76new : ability to generate and decode magicless frames (#591) 77changed : maximum nb of threads reduced to 200, to avoid address space exhaustion in 32-bits mode 78fix : multi-threading compression works with custom allocators 79fix : ZSTD_sizeof_CStream() was over-evaluating memory usage 80fix : a rare compression bug when compression generates very large distances and bunch of other conditions (only possible at --ultra -22) 81fix : 32-bits build can now decode large offsets (levels 21+) 82cli : added LZ4 frame support by default, by Felix Handte (@felixhandte) 83cli : improved --list output 84cli : new : can split input file for dictionary training, using command -B# 85cli : new : clean operation artefact on Ctrl-C interruption 86cli : fix : do not change /dev/null permissions when using command -t with root access, reported by @mike155 (#851) 87cli : fix : write file size in header in multiple-files mode 88api : added macro ZSTD_COMPRESSBOUND() for static allocation 89api : experimental : new advanced decompression API 90api : fix : sizeof_CCtx() used to over-estimate 91build: fix : no-multithread variant compiles without pool.c dependency, reported by Mitchell Blank Jr (@mitchblank) (#819) 92build: better compatibility with reproducible builds, by Bernhard M. Wiedemann (@bmwiedemann) (#818) 93example : added streaming_memory_usage 94license : changed /examples license to BSD + GPLv2 95license : fix a few header files to reflect new license (#825) 96 97v1.3.1 98New license : BSD + GPLv2 99perf: substantially decreased memory usage in Multi-threading mode, thanks to reports by Tino Reichardt (@mcmilk) 100perf: Multi-threading supports up to 256 threads. Cap at 256 when more are requested (#760) 101cli : improved and fixed --list command, by @ib (#772) 102cli : command -vV to list supported formats, by @ib (#771) 103build : fixed binary variants, reported by @svenha (#788) 104build : fix Visual compilation for non x86/x64 targets, reported by Greg Slazinski (@GregSlazinski) (#718) 105API exp : breaking change : ZSTD_getframeHeader() provides more information 106API exp : breaking change : pinned down values of error codes 107doc : fixed huffman example, by Ulrich Kunitz (@ulikunitz) 108new : contrib/adaptive-compression, I/O driven compression strength, by Paul Cruz (@paulcruz74) 109new : contrib/long_distance_matching, statistics by Stella Lau (@stellamplau) 110updated : contrib/linux-kernel, by Nick Terrell (@terrelln) 111 112v1.3.0 113cli : new : `--list` command, by Paul Cruz 114cli : changed : xz/lzma support enabled by default 115cli : changed : `-t *` continue processing list after a decompression error 116API : added : ZSTD_versionString() 117API : promoted to stable status : ZSTD_getFrameContentSize(), by Sean Purcell 118API exp : new advanced API : ZSTD_compress_generic(), ZSTD_CCtx_setParameter() 119API exp : new : API for static or external allocation : ZSTD_initStatic?Ctx() 120API exp : added : ZSTD_decompressBegin_usingDDict(), requested by Guy Riddle (#700) 121API exp : clarified memory estimation / measurement functions. 122API exp : changed : strongest strategy renamed ZSTD_btultra, fastest strategy ZSTD_fast set to 1 123tools : decodecorpus can generate random dictionary-compressed samples, by Paul Cruz 124new : contrib/seekable_format, demo and API, by Sean Purcell 125changed : contrib/linux-kernel, updated version and license, by Nick Terrell 126 127v1.2.0 128cli : changed : Multithreading enabled by default (use target zstd-nomt or HAVE_THREAD=0 to disable) 129cli : new : command -T0 means "detect and use nb of cores", by Sean Purcell 130cli : new : zstdmt symlink hardwired to `zstd -T0` 131cli : new : command --threads=# (#671) 132cli : changed : cover dictionary builder by default, for improved quality, by Nick Terrell 133cli : new : commands --train-cover and --train-legacy, to select dictionary algorithm and parameters 134cli : experimental targets `zstd4` and `xzstd4`, with support for lz4 format, by Sean Purcell 135cli : fix : does not output compressed data on console 136cli : fix : ignore symbolic links unless --force specified, 137API : breaking change : ZSTD_createCDict_advanced(), only use compressionParameters as argument 138API : added : prototypes ZSTD_*_usingCDict_advanced(), for direct control over frameParameters. 139API : improved: ZSTDMT_compressCCtx() reduced memory usage 140API : fix : ZSTDMT_compressCCtx() now provides srcSize in header (#634) 141API : fix : src size stored in frame header is controlled at end of frame 142API : fix : enforced consistent rules for pledgedSrcSize==0 (#641) 143API : fix : error code "GENERIC" replaced by "dstSizeTooSmall" when appropriate 144build: improved cmake script, by @Majlen 145build: enabled Multi-threading support for *BSD, by Baptiste Daroussin 146tools: updated Paramgrill. Command -O# provides best parameters for sample and speed target. 147new : contrib/linux-kernel version, by Nick Terrell 148 149v1.1.4 150cli : new : can compress in *.gz format, using --format=gzip command, by Przemyslaw Skibinski 151cli : new : advanced benchmark command --priority=rt 152cli : fix : write on sparse-enabled file systems in 32-bits mode, by @ds77 153cli : fix : --rm remains silent when input is stdin 154cli : experimental : xzstd, with support for xz/lzma decoding, by Przemyslaw Skibinski 155speed : improved decompression speed in streaming mode for single shot scenarios (+5%) 156memory: DDict (decompression dictionary) memory usage down from 150 KB to 20 KB 157arch: 32-bits variant able to generate and decode very long matches (>32 MB), by Sean Purcell 158API : new : ZSTD_findFrameCompressedSize(), ZSTD_getFrameContentSize(), ZSTD_findDecompressedSize() 159API : changed : dropped support of legacy versions <= v0.3 (can be changed by modifying ZSTD_LEGACY_SUPPORT value) 160build : new: meson build system in contrib/meson, by Dima Krasner 161build : improved cmake script, by @Majlen 162build : added -Wformat-security flag, as recommended by Padraig Brady 163doc : new : educational decoder, by Sean Purcell 164 165v1.1.3 166cli : zstd can decompress .gz files (can be disabled with `make zstd-nogz` or `make HAVE_ZLIB=0`) 167cli : new : experimental target `make zstdmt`, with multi-threading support 168cli : new : improved dictionary builder "cover" (experimental), by Nick Terrell, based on prior work by Giuseppe Ottaviano. 169cli : new : advanced commands for detailed parameters, by Przemyslaw Skibinski 170cli : fix zstdless on Mac OS-X, by Andrew Janke 171cli : fix #232 "compress non-files" 172dictBuilder : improved dictionary generation quality, thanks to Nick Terrell 173API : new : lib/compress/ZSTDMT_compress.h multithreading API (experimental) 174API : new : ZSTD_create?Dict_byReference(), requested by Bartosz Taudul 175API : new : ZDICT_finalizeDictionary() 176API : fix : ZSTD_initCStream_usingCDict() properly writes dictID into frame header, by Gregory Szorc (#511) 177API : fix : all symbols properly exposed in libzstd, by Nick Terrell 178build : support for Solaris target, by Przemyslaw Skibinski 179doc : clarified specification, by Sean Purcell 180 181v1.1.2 182API : streaming : decompression : changed : automatic implicit reset when chain-decoding new frames without init 183API : experimental : added : dictID retrieval functions, and ZSTD_initCStream_srcSize() 184API : zbuff : changed : prototypes now generate deprecation warnings 185lib : improved : faster decompression speed at ultra compression settings and 32-bits mode 186lib : changed : only public ZSTD_ symbols are now exposed 187lib : changed : reduced usage of stack memory 188lib : fixed : several corner case bugs, by Nick Terrell 189cli : new : gzstd, experimental version able to decode .gz files, by Przemyslaw Skibinski 190cli : new : preserve file attributes 191cli : new : added zstdless and zstdgrep tools 192cli : fixed : status displays total amount decoded, even for file consisting of multiple frames (like pzstd) 193cli : fixed : zstdcat 194zlib_wrapper : added support for gz* functions, by Przemyslaw Skibinski 195install : better compatibility with FreeBSD, by Dimitry Andric 196source tree : changed : zbuff source files moved to lib/deprecated 197 198v1.1.1 199New : command -M#, --memory=, --memlimit=, --memlimit-decompress= to limit allowed memory consumption 200New : doc/zstd_manual.html, by Przemyslaw Skibinski 201Improved : slightly better compression ratio at --ultra levels (>= 20) 202Improved : better memory usage when using streaming compression API, thanks to @Rogier-5 report 203Added : API : ZSTD_initCStream_usingCDict(), ZSTD_initDStream_usingDDict() (experimental section) 204Added : example/multiple_streaming_compression.c 205Changed : zstd_errors.h is now installed within /include (and replaces errors_public.h) 206Updated man page 207Fixed : zstd-small, zstd-compress and zstd-decompress compilation targets 208 209v1.1.0 210New : contrib/pzstd, parallel version of zstd, by Nick Terrell 211added : NetBSD install target (#338) 212Improved : speed for batches of small files 213Improved : speed of zlib wrapper, by Przemyslaw Skibinski 214Changed : libzstd on Windows supports legacy formats, by Christophe Chevalier 215Fixed : CLI -d output to stdout by default when input is stdin (#322) 216Fixed : CLI correctly detects console on Mac OS-X 217Fixed : CLI supports recursive mode `-r` on Mac OS-X 218Fixed : Legacy decoders use unified error codes, reported by benrg (#341), fixed by Przemyslaw Skibinski 219Fixed : compatibility with OpenBSD, reported by Juan Francisco Cantero Hurtado (#319) 220Fixed : compatibility with Hurd, by Przemyslaw Skibinski (#365) 221Fixed : zstd-pgo, reported by octoploid (#329) 222 223v1.0.0 224Change Licensing, all project is now BSD, Copyright Facebook 225Small decompression speed improvement 226API : Streaming API supports legacy format 227API : ZDICT_getDictID(), ZSTD_sizeof_{CCtx, DCtx, CStream, DStream}(), ZSTD_setDStreamParamter() 228CLI supports legacy formats v0.4+ 229Fixed : compression fails on certain huge files, reported by Jesse McGrew 230Enhanced documentation, by Przemyslaw Skibinski 231 232v0.8.1 233New streaming API 234Changed : --ultra now enables levels beyond 19 235Changed : -i# now selects benchmark time in second 236Fixed : ZSTD_compress* can now compress > 4 GB in a single pass, reported by Nick Terrell 237Fixed : speed regression on specific patterns (#272) 238Fixed : support for Z_SYNC_FLUSH, by Dmitry Krot (#291) 239Fixed : ICC compilation, by Przemyslaw Skibinski 240 241v0.8.0 242Improved : better speed on clang and gcc -O2, thanks to Eric Biggers 243New : Build on FreeBSD and DragonFly, thanks to JrMarino 244Changed : modified API : ZSTD_compressEnd() 245Fixed : legacy mode with ZSTD_HEAPMODE=0, by Christopher Bergqvist 246Fixed : premature end of frame when zero-sized raw block, reported by Eric Biggers 247Fixed : large dictionaries (> 384 KB), reported by Ilona Papava 248Fixed : checksum correctly checked in single-pass mode 249Fixed : combined --test amd --rm, reported by Andreas M. Nilsson 250Modified : minor compression level adaptations 251Updated : compression format specification to v0.2.0 252changed : zstd.h moved to /lib directory 253 254v0.7.5 255Transition version, supporting decoding of v0.8.x 256 257v0.7.4 258Added : homebrew for Mac, by Daniel Cade 259Added : more examples 260Fixed : segfault when using small dictionaries, reported by Felix Handte 261Modified : default compression level for CLI is now 3 262Updated : specification, to v0.1.1 263 264v0.7.3 265New : compression format specification 266New : `--` separator, stating that all following arguments are file names. Suggested by Chip Turner. 267New : `ZSTD_getDecompressedSize()` 268New : OpenBSD target, by Juan Francisco Cantero Hurtado 269New : `examples` directory 270fixed : dictBuilder using HC levels, reported by Bartosz Taudul 271fixed : legacy support from ZSTD_decompress_usingDDict(), reported by Felix Handte 272fixed : multi-blocks decoding with intermediate uncompressed blocks, reported by Greg Slazinski 273modified : removed "mem.h" and "error_public.h" dependencies from "zstd.h" (experimental section) 274modified : legacy functions no longer need magic number 275 276v0.7.2 277fixed : ZSTD_decompressBlock() using multiple consecutive blocks. Reported by Greg Slazinski. 278fixed : potential segfault on very large files (many gigabytes). Reported by Chip Turner. 279fixed : CLI displays system error message when destination file cannot be created (#231). Reported by Chip Turner. 280 281v0.7.1 282fixed : ZBUFF_compressEnd() called multiple times with too small `dst` buffer, reported by Christophe Chevalier 283fixed : dictBuilder fails if first sample is too small, reported by Руслан Ковалёв 284fixed : corruption issue, reported by cj 285modified : checksum enabled by default in command line mode 286 287v0.7.0 288New : Support for directory compression, using `-r`, thanks to Przemyslaw Skibinski 289New : Command `--rm`, to remove source file after successful de/compression 290New : Visual build scripts, by Christophe Chevalier 291New : Support for Sparse File-systems (do not use space for zero-filled sectors) 292New : Frame checksum support 293New : Support pass-through mode (when using `-df`) 294API : more efficient Dictionary API : `ZSTD_compress_usingCDict()`, `ZSTD_decompress_usingDDict()` 295API : create dictionary files from custom content, by Giuseppe Ottaviano 296API : support for custom malloc/free functions 297New : controllable Dictionary ID 298New : Support for skippable frames 299 300v0.6.1 301New : zlib wrapper API, thanks to Przemyslaw Skibinski 302New : Ability to compile compressor / decompressor separately 303Changed : new lib directory structure 304Fixed : Legacy codec v0.5 compatible with dictionary decompression 305Fixed : Decoder corruption error (#173) 306Fixed : null-string roundtrip (#176) 307New : benchmark mode can select directory as input 308Experimental : midipix support, VMS support 309 310v0.6.0 311Stronger high compression modes, thanks to Przemyslaw Skibinski 312API : ZSTD_getFrameParams() provides size of decompressed content 313New : highest compression modes require `--ultra` command to fully unleash their capacity 314Fixed : zstd cli return error code > 0 and removes dst file artifact when decompression fails, thanks to Chip Turner 315 316v0.5.1 317New : Optimal parsing => Very high compression modes, thanks to Przemyslaw Skibinski 318Changed : Dictionary builder integrated into libzstd and zstd cli 319Changed (!) : zstd cli now uses "multiple input files" as default mode. See `zstd -h`. 320Fix : high compression modes for big-endian platforms 321New : zstd cli : `-t` | `--test` command 322 323v0.5.0 324New : dictionary builder utility 325Changed : streaming & dictionary API 326Improved : better compression of small data 327 328v0.4.7 329Improved : small compression speed improvement in HC mode 330Changed : `zstd_decompress.c` has ZSTD_LEGACY_SUPPORT to 0 by default 331fix : bt search bug 332 333v0.4.6 334fix : fast compression mode on Windows 335New : cmake configuration file, thanks to Artyom Dymchenko 336Improved : high compression mode on repetitive data 337New : block-level API 338New : ZSTD_duplicateCCtx() 339 340v0.4.5 341new : -m/--multiple : compress/decompress multiple files 342 343v0.4.4 344Fixed : high compression modes for Windows 32 bits 345new : external dictionary API extended to buffered mode and accessible through command line 346new : windows DLL project, thanks to Christophe Chevalier 347 348v0.4.3 : 349new : external dictionary API 350new : zstd-frugal 351 352v0.4.2 : 353Generic minor improvements for small blocks 354Fixed : big-endian compatibility, by Peter Harris (#85) 355 356v0.4.1 357Fixed : ZSTD_LEGACY_SUPPORT=0 build mode (reported by Luben) 358removed `zstd.c` 359 360v0.4.0 361Command line utility compatible with high compression levels 362Removed zstdhc => merged into zstd 363Added : ZBUFF API (see zstd_buffered.h) 364Rolling buffer support 365 366v0.3.6 367small blocks params 368 369v0.3.5 370minor generic compression improvements 371 372v0.3.4 373Faster fast cLevels 374 375v0.3.3 376Small compression ratio improvement 377 378v0.3.2 379Fixed Visual Studio 380 381v0.3.1 : 382Small compression ratio improvement 383 384v0.3 385HC mode : compression levels 2-26 386 387v0.2.2 388Fix : Visual Studio 2013 & 2015 release compilation, by Christophe Chevalier 389 390v0.2.1 391Fix : Read errors, advanced fuzzer tests, by Hanno Böck 392 393v0.2.0 394**Breaking format change** 395Faster decompression speed 396Can still decode v0.1 format 397 398v0.1.3 399fix uninitialization warning, reported by Evan Nemerson 400 401v0.1.2 402frame concatenation support 403 404v0.1.1 405fix compression bug 406detects write-flush errors 407 408v0.1.0 409first release 410