11 * liblzma: Fixed lzma_stream_encoder_mt_memusage() when a preset
14 * xzdiff: If mktemp isn't installed, mkdir will be used as
15 a fallback to create a temporary directory. Installing mktemp
18 * Updated French, German, Italian, Polish, and Vietnamese
21 Summary of fixes and new features added in the 5.1.x development
26 - Added support for multi-threaded compression. See the
27 lzma_mt structure, lzma_stream_encoder_mt(), and
28 lzma_stream_encoder_mt_memusage() in <lzma/container.h>,
29 lzma_get_progress() in <lzma/base.h>, and lzma_cputhreads()
30 in <lzma/hardware.h> for details.
32 - Made the uses of lzma_allocator const correct.
34 - Added lzma_block_uncomp_encode() to create uncompressed
35 .xz Blocks using LZMA2 uncompressed chunks.
37 - Added support for LZMA_IGNORE_CHECK.
39 - A few speed optimizations were made.
41 - Added support for symbol versioning. It is enabled by default
42 on GNU/Linux, other GNU-based systems, and FreeBSD.
44 - liblzma (not the whole XZ Utils) should now be buildable
45 with MSVC 2013 update 2 or later using windows/config.h.
49 - Fixed a race condition in the signal handling. It was
50 possible that e.g. the first SIGINT didn't make xz exit
51 if reading or writing blocked and one had bad luck. The fix
52 is non-trivial, so as of writing it is unknown if it will be
53 backported to the v5.0 branch.
55 - Multi-threaded decompression can be enabled with the
56 --threads (-T) option.
58 - New command line options in xz: --single-stream,
59 --block-size=SIZE, --block-list=SIZES,
60 --flush-timeout=TIMEOUT, and --ignore-check.
62 - xz -lvv now shows the minimum xz version that is required to
63 decompress the file. Currently it is 5.0.0 for all supported
64 .xz files except files with empty LZMA2 streams require 5.0.2.
66 * xzdiff and xzgrep now support .lzo files if lzop is installed.
67 The .tzo suffix is also recognized as a shorthand for .tar.lzo.
70 5.1.4beta (2014-09-14)
72 * All fixes from 5.0.6
74 * liblzma: Fixed the use of presets in threaded encoder
77 * xz --block-list and --block-size can now be used together
78 in single-threaded mode. Previously the combination only
79 worked in multi-threaded mode.
81 * Added support for LZMA_IGNORE_CHECK to liblzma and made it
82 available in xz as --ignore-check.
84 * liblzma speed optimizations:
86 - Initialization of a new LZMA1 or LZMA2 encoder has been
87 optimized. (The speed of reinitializing an already-allocated
88 encoder isn't affected.) This helps when compressing many
89 small buffers with lzma_stream_buffer_encode() and other
90 similar situations where an already-allocated encoder state
91 isn't reused. This speed-up is visible in xz too if one
92 compresses many small files one at a time instead running xz
93 once and giving all files as command-line arguments.
95 - Buffer comparisons are now much faster when unaligned access
96 is allowed (configured with --enable-unaligned-access). This
97 speeds up encoding significantly. There is arch-specific code
98 for 32-bit and 64-bit x86 (32-bit needs SSE2 for the best
99 results and there's no run-time CPU detection for now).
100 For other archs there is only generic code which probably
101 isn't as optimal as arch-specific solutions could be.
103 - A few speed optimizations were made to the SHA-256 code.
104 (Note that the builtin SHA-256 code isn't used on all
107 * liblzma can now be built with MSVC 2013 update 2 or later
108 using windows/config.h.
110 * Vietnamese translation was added.
113 5.1.3alpha (2013-10-26)
115 * All fixes from 5.0.5
119 - Fixed a deadlock in the threaded encoder.
121 - Made the uses of lzma_allocator const correct.
123 - Added lzma_block_uncomp_encode() to create uncompressed
124 .xz Blocks using LZMA2 uncompressed chunks.
126 - Added support for native threads on Windows and the ability
127 to detect the number of CPU cores.
131 - Fixed a race condition in the signal handling. It was
132 possible that e.g. the first SIGINT didn't make xz exit
133 if reading or writing blocked and one had bad luck. The fix
134 is non-trivial, so as of writing it is unknown if it will be
135 backported to the v5.0 branch.
137 - Made the progress indicator work correctly in threaded mode.
139 - Threaded encoder now works together with --block-list=SIZES.
141 - Added preliminary support for --flush-timeout=TIMEOUT.
142 It can be useful for (somewhat) real-time streaming. For
143 now the decompression side has to be done with something
144 else than the xz tool due to how xz does buffering, but this
148 5.1.2alpha (2012-07-04)
150 * All fixes from 5.0.3 and 5.0.4
154 - Fixed a deadlock and an invalid free() in the threaded encoder.
156 - Added support for symbol versioning. It is enabled by default
157 on GNU/Linux, other GNU-based systems, and FreeBSD.
159 - Use SHA-256 implementation from the operating system if one is
160 available in libc, libmd, or libutil. liblzma won't use e.g.
161 OpenSSL or libgcrypt to avoid introducing new dependencies.
163 - Fixed liblzma.pc for static linking.
165 - Fixed a few portability bugs.
167 * xz --decompress --single-stream now fixes the input position after
168 successful decompression. Now the following works:
170 echo foo | xz > foo.xz
171 echo bar | xz >> foo.xz
172 ( xz -dc --single-stream ; xz -dc --single-stream ) < foo.xz
174 Note that it doesn't work if the input is not seekable
175 or if there is Stream Padding between the concatenated
178 * xz -lvv now shows the minimum xz version that is required to
179 decompress the file. Currently it is 5.0.0 for all supported .xz
180 files except files with empty LZMA2 streams require 5.0.2.
182 * Added an *incomplete* implementation of --block-list=SIZES to xz.
183 It only works correctly in single-threaded mode and when
184 --block-size isn't used at the same time. --block-list allows
185 specifying the sizes of Blocks which can be useful e.g. when
186 creating files for random-access reading.
189 5.1.1alpha (2011-04-12)
191 * All fixes from 5.0.2
193 * liblzma fixes that will also be included in 5.0.3:
195 - A memory leak was fixed.
197 - lzma_stream_buffer_encode() no longer creates an empty .xz
198 Block if encoding an empty buffer. Such an empty Block with
199 LZMA2 data would trigger a bug in 5.0.1 and older (see the
200 first bullet point in 5.0.2 notes). When releasing 5.0.2,
201 I thought that no encoder creates this kind of files but
204 - Validate function arguments better in a few functions. Most
205 importantly, specifying an unsupported integrity check to
206 lzma_stream_buffer_encode() no longer creates a corrupt .xz
207 file. Probably no application tries to do that, so this
208 shouldn't be a big problem in practice.
210 - Document that lzma_block_buffer_encode(),
211 lzma_easy_buffer_encode(), lzma_stream_encoder(), and
212 lzma_stream_buffer_encode() may return LZMA_UNSUPPORTED_CHECK.
214 - The return values of the _memusage() functions are now
217 * Support for multithreaded compression was added using the simplest
218 method, which splits the input data into blocks and compresses
219 them independently. Other methods will be added in the future.
220 The current method has room for improvement, e.g. it is possible
221 to reduce the memory usage.
223 * Added the options --single-stream and --block-size=SIZE to xz.
225 * xzdiff and xzgrep now support .lzo files if lzop is installed.
226 The .tzo suffix is also recognized as a shorthand for .tar.lzo.
228 * Support for short 8.3 filenames under DOS was added to xz. It is
229 experimental and may change before it gets into a stable release.
234 * Fixed an old bug in xzgrep that affected OpenBSD and probably
235 a few other operating systems too.
237 * Updated French and German translations.
239 * Added support for detecting the amount of RAM on AmigaOS/AROS.
241 * Minor build system updates.
246 * Fix regressions introduced in 5.0.6:
248 - Fix building with non-GNU make.
250 - Fix invalid Libs.private value in liblzma.pc which broke
251 static linking against liblzma if the linker flags were
252 taken from pkg-config.
257 * xzgrep now exits with status 0 if at least one file matched.
259 * A few minor portability and build system fixes
264 * lzmadec and liblzma's lzma_alone_decoder(): Support decompressing
265 .lzma files that have less common settings in the headers
266 (dictionary size other than 2^n or 2^n + 2^(n-1), or uncompressed
267 size greater than 256 GiB). The limitations existed to avoid false
268 positives when detecting .lzma files. The lc + lp <= 4 limitation
269 still remains since liblzma's LZMA decoder has that limitation.
271 NOTE: xz's .lzma support or liblzma's lzma_auto_decoder() are NOT
272 affected by this change. They still consider uncommon .lzma headers
273 as not being in the .lzma format. Changing this would give way too
274 many false positives.
278 - Interaction of preset and custom filter chain options was
279 made less illogical. This affects only certain less typical
280 uses cases so few people are expected to notice this change.
282 Now when a custom filter chain option (e.g. --lzma2) is
283 specified, all preset options (-0 ... -9, -e) earlier are on
284 the command line are completely forgotten. Similarly, when
285 a preset option is specified, all custom filter chain options
286 earlier on the command line are completely forgotten.
288 Example 1: "xz -9 --lzma2=preset=5 -e" is equivalent to "xz -e"
289 which is equivalent to "xz -6e". Earlier -e didn't put xz back
290 into preset mode and thus the example command was equivalent
291 to "xz --lzma2=preset=5".
293 Example 2: "xz -9e --lzma2=preset=5 -7" is equivalent to
294 "xz -7". Earlier a custom filter chain option didn't make
295 xz forget the -e option so the example was equivalent to
298 - Fixes and improvements to error handling.
300 - Various fixes to the man page.
302 * xzless: Fixed to work with "less" versions 448 and later.
304 * xzgrep: Made -h an alias for --no-filename.
306 * Include the previously missing debug/translation.bash which can
307 be useful for translators.
309 * Include a build script for Mac OS X. This has been in the Git
310 repository since 2010 but due to a mistake in Makefile.am the
311 script hasn't been included in a release tarball before.
318 - Fix lzma_index_init(). It could crash if memory allocation
321 - Fix the possibility of an incorrect LZMA_BUF_ERROR when a BCJ
322 filter is used and the application only provides exactly as
323 much output space as is the uncompressed size of the file.
325 - Fix a bug in doc/examples_old/xz_pipe_decompress.c. It didn't
326 check if the last call to lzma_code() really returned
327 LZMA_STREAM_END, which made the program think that truncated
330 - New example programs in doc/examples (old programs are now in
331 doc/examples_old). These have more comments and more detailed
334 * Fix "xz -lvv foo.xz". It could crash on some corrupted files.
336 * Fix output of "xz --robot -lv" and "xz --robot -lvv" which
337 incorrectly printed the filename also in the "foo (x/x)" format.
339 * Fix exit status of "xzdiff foo.xz bar.xz".
341 * Fix exit status of "xzgrep foo binary_file".
343 * Fix portability to EBCDIC systems.
345 * Fix a configure issue on AIX with the XL C compiler. See INSTALL
348 * Update French, German, Italian, and Polish translations.
355 - A memory leak was fixed.
357 - lzma_stream_buffer_encode() no longer creates an empty .xz
358 Block if encoding an empty buffer. Such an empty Block with
359 LZMA2 data would trigger a bug in 5.0.1 and older (see the
360 first bullet point in 5.0.2 notes). When releasing 5.0.2,
361 I thought that no encoder creates this kind of files but
364 - Validate function arguments better in a few functions. Most
365 importantly, specifying an unsupported integrity check to
366 lzma_stream_buffer_encode() no longer creates a corrupt .xz
367 file. Probably no application tries to do that, so this
368 shouldn't be a big problem in practice.
370 - Document that lzma_block_buffer_encode(),
371 lzma_easy_buffer_encode(), lzma_stream_encoder(), and
372 lzma_stream_buffer_encode() may return LZMA_UNSUPPORTED_CHECK.
374 - The return values of the _memusage() functions are now
377 * Fix command name detection in xzgrep. xzegrep and xzfgrep now
378 correctly use egrep and fgrep instead of grep.
380 * French translation was added.
385 * LZMA2 decompressor now correctly accepts LZMA2 streams with no
386 uncompressed data. Previously it considered them corrupt. The
387 bug can affect applications that use raw LZMA2 streams. It is
388 very unlikely to affect .xz files because no compressor creates
389 .xz files with empty LZMA2 streams. (Empty .xz files are a
390 different thing than empty LZMA2 streams.)
392 * "xz --suffix=.foo filename.foo" now refuses to compress the
393 file due to it already having the suffix .foo. It was already
394 documented on the man page, but the code lacked the test.
396 * "xzgrep -l foo bar.xz" works now.
398 * Polish translation was added.
403 * xz --force now (de)compresses files that have setuid, setgid,
404 or sticky bit set and files that have multiple hard links.
405 The man page had it documented this way already, but the code
408 * gzip and bzip2 support in xzdiff was fixed.
412 * Minor fix to Czech translation
417 Only the most important changes compared to 4.999.9beta are listed
418 here. One change is especially important:
420 * The memory usage limit is now disabled by default. Some scripts
421 written before this change may have used --memory=max on xz command
422 line or in XZ_OPT. THESE USES OF --memory=max SHOULD BE REMOVED
423 NOW, because they interfere with user's ability to set the memory
424 usage limit himself. If user-specified limit causes problems to
425 your script, blame the user.
427 Other significant changes:
429 * Added support for XZ_DEFAULTS environment variable. This variable
430 allows users to set default options for xz, e.g. default memory
431 usage limit or default compression level. Scripts that use xz
432 must never set or unset XZ_DEFAULTS. Scripts should use XZ_OPT
433 instead if they need a way to pass options to xz via an
434 environment variable.
436 * The compression settings associated with the preset levels
437 -0 ... -9 have been changed. --extreme was changed a little too.
438 It is now less likely to make compression worse, but with some
439 files the new --extreme may compress slightly worse than the old
442 * If a preset level (-0 ... -9) is specified after a custom filter
443 chain options have been used (e.g. --lzma2), the custom filter
444 chain will be forgotten. Earlier the preset options were
445 completely ignored after custom filter chain options had been
448 * xz will create sparse files when decompressing if the uncompressed
449 data contains long sequences of binary zeros. This is done even
450 when writing to standard output that is connected to a regular
451 file and certain additional conditions are met to make it safe.
453 * Support for "xz --list" was added. Combine with --verbose or
454 --verbose --verbose (-vv) for detailed output.
456 * I had hoped that liblzma API would have been stable after
457 4.999.9beta, but there have been a couple of changes in the
458 advanced features, which don't affect most applications:
460 - Index handling code was revised. If you were using the old
461 API, you will get a compiler error (so it's easy to notice).
463 - A subtle but important change was made to the Block handling
464 API. lzma_block.version has to be initialized even for
465 lzma_block_header_decode(). Code that doesn't do it will work
466 for now, but might break in the future, which makes this API
469 * The major soname has been bumped to 5.0.0. liblzma API and ABI
470 are now stable, so the need to recompile programs linking against
471 liblzma shouldn't arise soon.