| blob | 376abea893e61ab9833fe412ac8119426b742a7e |
2 XZ Utils Release Notes
3 ======================
5 5.2.1 (2015-02-26)
7 * Fixed a compression-ratio regression in fast mode of LZMA1 and
8 LZMA2. The bug is present in 5.1.4beta and 5.2.0 releases.
10 * Fixed a portability problem in xz that affected at least OpenBSD.
12 * Fixed xzdiff to be compatible with FreeBSD's mktemp which differs
13 from most other mktemp implementations.
15 * Changed CPU core count detection to use cpuset_getaffinity() on
16 FreeBSD.
19 5.2.0 (2014-12-21)
21 Since 5.1.4beta:
23 * All fixes from 5.0.8
25 * liblzma: Fixed lzma_stream_encoder_mt_memusage() when a preset
26 was used.
28 * xzdiff: If mktemp isn't installed, mkdir will be used as
29 a fallback to create a temporary directory. Installing mktemp
30 is still recommended.
32 * Updated French, German, Italian, Polish, and Vietnamese
33 translations.
35 Summary of fixes and new features added in the 5.1.x development
36 releases:
38 * liblzma:
40 - Added support for multi-threaded compression. See the
41 lzma_mt structure, lzma_stream_encoder_mt(), and
42 lzma_stream_encoder_mt_memusage() in <lzma/container.h>,
43 lzma_get_progress() in <lzma/base.h>, and lzma_cputhreads()
44 in <lzma/hardware.h> for details.
46 - Made the uses of lzma_allocator const correct.
48 - Added lzma_block_uncomp_encode() to create uncompressed
49 .xz Blocks using LZMA2 uncompressed chunks.
51 - Added support for LZMA_IGNORE_CHECK.
53 - A few speed optimizations were made.
55 - Added support for symbol versioning. It is enabled by default
56 on GNU/Linux, other GNU-based systems, and FreeBSD.
58 - liblzma (not the whole XZ Utils) should now be buildable
59 with MSVC 2013 update 2 or later using windows/config.h.
61 * xz:
63 - Fixed a race condition in the signal handling. It was
64 possible that e.g. the first SIGINT didn't make xz exit
65 if reading or writing blocked and one had bad luck. The fix
66 is non-trivial, so as of writing it is unknown if it will be
67 backported to the v5.0 branch.
69 - Multi-threaded decompression can be enabled with the
70 --threads (-T) option.
72 - New command line options in xz: --single-stream,
73 --block-size=SIZE, --block-list=SIZES,
74 --flush-timeout=TIMEOUT, and --ignore-check.
76 - xz -lvv now shows the minimum xz version that is required to
77 decompress the file. Currently it is 5.0.0 for all supported
78 .xz files except files with empty LZMA2 streams require 5.0.2.
80 * xzdiff and xzgrep now support .lzo files if lzop is installed.
81 The .tzo suffix is also recognized as a shorthand for .tar.lzo.
84 5.1.4beta (2014-09-14)
86 * All fixes from 5.0.6
88 * liblzma: Fixed the use of presets in threaded encoder
89 initialization.
91 * xz --block-list and --block-size can now be used together
92 in single-threaded mode. Previously the combination only
93 worked in multi-threaded mode.
95 * Added support for LZMA_IGNORE_CHECK to liblzma and made it
96 available in xz as --ignore-check.
98 * liblzma speed optimizations:
100 - Initialization of a new LZMA1 or LZMA2 encoder has been
101 optimized. (The speed of reinitializing an already-allocated
102 encoder isn't affected.) This helps when compressing many
103 small buffers with lzma_stream_buffer_encode() and other
104 similar situations where an already-allocated encoder state
105 isn't reused. This speed-up is visible in xz too if one
106 compresses many small files one at a time instead running xz
107 once and giving all files as command-line arguments.
109 - Buffer comparisons are now much faster when unaligned access
110 is allowed (configured with --enable-unaligned-access). This
111 speeds up encoding significantly. There is arch-specific code
112 for 32-bit and 64-bit x86 (32-bit needs SSE2 for the best
113 results and there's no run-time CPU detection for now).
114 For other archs there is only generic code which probably
115 isn't as optimal as arch-specific solutions could be.
117 - A few speed optimizations were made to the SHA-256 code.
118 (Note that the builtin SHA-256 code isn't used on all
119 operating systems.)
121 * liblzma can now be built with MSVC 2013 update 2 or later
122 using windows/config.h.
124 * Vietnamese translation was added.
127 5.1.3alpha (2013-10-26)
129 * All fixes from 5.0.5
131 * liblzma:
133 - Fixed a deadlock in the threaded encoder.
135 - Made the uses of lzma_allocator const correct.
137 - Added lzma_block_uncomp_encode() to create uncompressed
138 .xz Blocks using LZMA2 uncompressed chunks.
140 - Added support for native threads on Windows and the ability
141 to detect the number of CPU cores.
143 * xz:
145 - Fixed a race condition in the signal handling. It was
146 possible that e.g. the first SIGINT didn't make xz exit
147 if reading or writing blocked and one had bad luck. The fix
148 is non-trivial, so as of writing it is unknown if it will be
149 backported to the v5.0 branch.
151 - Made the progress indicator work correctly in threaded mode.
153 - Threaded encoder now works together with --block-list=SIZES.
155 - Added preliminary support for --flush-timeout=TIMEOUT.
156 It can be useful for (somewhat) real-time streaming. For
157 now the decompression side has to be done with something
158 else than the xz tool due to how xz does buffering, but this
159 should be fixed.
162 5.1.2alpha (2012-07-04)
164 * All fixes from 5.0.3 and 5.0.4
166 * liblzma:
168 - Fixed a deadlock and an invalid free() in the threaded encoder.
170 - Added support for symbol versioning. It is enabled by default
171 on GNU/Linux, other GNU-based systems, and FreeBSD.
173 - Use SHA-256 implementation from the operating system if one is
174 available in libc, libmd, or libutil. liblzma won't use e.g.
175 OpenSSL or libgcrypt to avoid introducing new dependencies.
177 - Fixed liblzma.pc for static linking.
179 - Fixed a few portability bugs.
181 * xz --decompress --single-stream now fixes the input position after
182 successful decompression. Now the following works:
184 echo foo | xz > foo.xz
185 echo bar | xz >> foo.xz
186 ( xz -dc --single-stream ; xz -dc --single-stream ) < foo.xz
188 Note that it doesn't work if the input is not seekable
189 or if there is Stream Padding between the concatenated
190 .xz Streams.
192 * xz -lvv now shows the minimum xz version that is required to
193 decompress the file. Currently it is 5.0.0 for all supported .xz
194 files except files with empty LZMA2 streams require 5.0.2.
196 * Added an *incomplete* implementation of --block-list=SIZES to xz.
197 It only works correctly in single-threaded mode and when
198 --block-size isn't used at the same time. --block-list allows
199 specifying the sizes of Blocks which can be useful e.g. when
200 creating files for random-access reading.
203 5.1.1alpha (2011-04-12)
205 * All fixes from 5.0.2
207 * liblzma fixes that will also be included in 5.0.3:
209 - A memory leak was fixed.
211 - lzma_stream_buffer_encode() no longer creates an empty .xz
212 Block if encoding an empty buffer. Such an empty Block with
213 LZMA2 data would trigger a bug in 5.0.1 and older (see the
214 first bullet point in 5.0.2 notes). When releasing 5.0.2,
215 I thought that no encoder creates this kind of files but
216 I was wrong.
218 - Validate function arguments better in a few functions. Most
219 importantly, specifying an unsupported integrity check to
220 lzma_stream_buffer_encode() no longer creates a corrupt .xz
221 file. Probably no application tries to do that, so this
222 shouldn't be a big problem in practice.
224 - Document that lzma_block_buffer_encode(),
225 lzma_easy_buffer_encode(), lzma_stream_encoder(), and
226 lzma_stream_buffer_encode() may return LZMA_UNSUPPORTED_CHECK.
228 - The return values of the _memusage() functions are now
229 documented better.
231 * Support for multithreaded compression was added using the simplest
232 method, which splits the input data into blocks and compresses
233 them independently. Other methods will be added in the future.
234 The current method has room for improvement, e.g. it is possible
235 to reduce the memory usage.
237 * Added the options --single-stream and --block-size=SIZE to xz.
239 * xzdiff and xzgrep now support .lzo files if lzop is installed.
240 The .tzo suffix is also recognized as a shorthand for .tar.lzo.
242 * Support for short 8.3 filenames under DOS was added to xz. It is
243 experimental and may change before it gets into a stable release.
246 5.0.8 (2014-12-21)
248 * Fixed an old bug in xzgrep that affected OpenBSD and probably
249 a few other operating systems too.
251 * Updated French and German translations.
253 * Added support for detecting the amount of RAM on AmigaOS/AROS.
255 * Minor build system updates.
258 5.0.7 (2014-09-20)
260 * Fix regressions introduced in 5.0.6:
262 - Fix building with non-GNU make.
264 - Fix invalid Libs.private value in liblzma.pc which broke
265 static linking against liblzma if the linker flags were
266 taken from pkg-config.
269 5.0.6 (2014-09-14)
271 * xzgrep now exits with status 0 if at least one file matched.
273 * A few minor portability and build system fixes
276 5.0.5 (2013-06-30)
278 * lzmadec and liblzma's lzma_alone_decoder(): Support decompressing
279 .lzma files that have less common settings in the headers
280 (dictionary size other than 2^n or 2^n + 2^(n-1), or uncompressed
281 size greater than 256 GiB). The limitations existed to avoid false
282 positives when detecting .lzma files. The lc + lp <= 4 limitation
283 still remains since liblzma's LZMA decoder has that limitation.
285 NOTE: xz's .lzma support or liblzma's lzma_auto_decoder() are NOT
286 affected by this change. They still consider uncommon .lzma headers
287 as not being in the .lzma format. Changing this would give way too
288 many false positives.
290 * xz:
292 - Interaction of preset and custom filter chain options was
293 made less illogical. This affects only certain less typical
294 uses cases so few people are expected to notice this change.
296 Now when a custom filter chain option (e.g. --lzma2) is
297 specified, all preset options (-0 ... -9, -e) earlier are on
298 the command line are completely forgotten. Similarly, when
299 a preset option is specified, all custom filter chain options
300 earlier on the command line are completely forgotten.
302 Example 1: "xz -9 --lzma2=preset=5 -e" is equivalent to "xz -e"
303 which is equivalent to "xz -6e". Earlier -e didn't put xz back
304 into preset mode and thus the example command was equivalent
305 to "xz --lzma2=preset=5".
307 Example 2: "xz -9e --lzma2=preset=5 -7" is equivalent to
308 "xz -7". Earlier a custom filter chain option didn't make
309 xz forget the -e option so the example was equivalent to
310 "xz -7e".
312 - Fixes and improvements to error handling.
314 - Various fixes to the man page.
316 * xzless: Fixed to work with "less" versions 448 and later.
318 * xzgrep: Made -h an alias for --no-filename.
320 * Include the previously missing debug/translation.bash which can
321 be useful for translators.
323 * Include a build script for Mac OS X. This has been in the Git
324 repository since 2010 but due to a mistake in Makefile.am the
325 script hasn't been included in a release tarball before.
328 5.0.4 (2012-06-22)
330 * liblzma:
332 - Fix lzma_index_init(). It could crash if memory allocation
333 failed.
335 - Fix the possibility of an incorrect LZMA_BUF_ERROR when a BCJ
336 filter is used and the application only provides exactly as
337 much output space as is the uncompressed size of the file.
339 - Fix a bug in doc/examples_old/xz_pipe_decompress.c. It didn't
340 check if the last call to lzma_code() really returned
341 LZMA_STREAM_END, which made the program think that truncated
342 files are valid.
344 - New example programs in doc/examples (old programs are now in
345 doc/examples_old). These have more comments and more detailed
346 error handling.
348 * Fix "xz -lvv foo.xz". It could crash on some corrupted files.
350 * Fix output of "xz --robot -lv" and "xz --robot -lvv" which
351 incorrectly printed the filename also in the "foo (x/x)" format.
353 * Fix exit status of "xzdiff foo.xz bar.xz".
355 * Fix exit status of "xzgrep foo binary_file".
357 * Fix portability to EBCDIC systems.
359 * Fix a configure issue on AIX with the XL C compiler. See INSTALL
360 for details.
362 * Update French, German, Italian, and Polish translations.
365 5.0.3 (2011-05-21)
367 * liblzma fixes:
369 - A memory leak was fixed.
371 - lzma_stream_buffer_encode() no longer creates an empty .xz
372 Block if encoding an empty buffer. Such an empty Block with
373 LZMA2 data would trigger a bug in 5.0.1 and older (see the
374 first bullet point in 5.0.2 notes). When releasing 5.0.2,
375 I thought that no encoder creates this kind of files but
376 I was wrong.
378 - Validate function arguments better in a few functions. Most
379 importantly, specifying an unsupported integrity check to
380 lzma_stream_buffer_encode() no longer creates a corrupt .xz
381 file. Probably no application tries to do that, so this
382 shouldn't be a big problem in practice.
384 - Document that lzma_block_buffer_encode(),
385 lzma_easy_buffer_encode(), lzma_stream_encoder(), and
386 lzma_stream_buffer_encode() may return LZMA_UNSUPPORTED_CHECK.
388 - The return values of the _memusage() functions are now
389 documented better.
391 * Fix command name detection in xzgrep. xzegrep and xzfgrep now
392 correctly use egrep and fgrep instead of grep.
394 * French translation was added.
397 5.0.2 (2011-04-01)
399 * LZMA2 decompressor now correctly accepts LZMA2 streams with no
400 uncompressed data. Previously it considered them corrupt. The
401 bug can affect applications that use raw LZMA2 streams. It is
402 very unlikely to affect .xz files because no compressor creates
403 .xz files with empty LZMA2 streams. (Empty .xz files are a
404 different thing than empty LZMA2 streams.)
406 * "xz --suffix=.foo filename.foo" now refuses to compress the
407 file due to it already having the suffix .foo. It was already
408 documented on the man page, but the code lacked the test.
410 * "xzgrep -l foo bar.xz" works now.
412 * Polish translation was added.
415 5.0.1 (2011-01-29)
417 * xz --force now (de)compresses files that have setuid, setgid,
418 or sticky bit set and files that have multiple hard links.
419 The man page had it documented this way already, but the code
420 had a bug.
422 * gzip and bzip2 support in xzdiff was fixed.
424 * Portability fixes
426 * Minor fix to Czech translation
429 5.0.0 (2010-10-23)
431 Only the most important changes compared to 4.999.9beta are listed
432 here. One change is especially important:
434 * The memory usage limit is now disabled by default. Some scripts
435 written before this change may have used --memory=max on xz command
436 line or in XZ_OPT. THESE USES OF --memory=max SHOULD BE REMOVED
437 NOW, because they interfere with user's ability to set the memory
438 usage limit himself. If user-specified limit causes problems to
439 your script, blame the user.
441 Other significant changes:
443 * Added support for XZ_DEFAULTS environment variable. This variable
444 allows users to set default options for xz, e.g. default memory
445 usage limit or default compression level. Scripts that use xz
446 must never set or unset XZ_DEFAULTS. Scripts should use XZ_OPT
447 instead if they need a way to pass options to xz via an
448 environment variable.
450 * The compression settings associated with the preset levels
451 -0 ... -9 have been changed. --extreme was changed a little too.
452 It is now less likely to make compression worse, but with some
453 files the new --extreme may compress slightly worse than the old
454 --extreme.
456 * If a preset level (-0 ... -9) is specified after a custom filter
457 chain options have been used (e.g. --lzma2), the custom filter
458 chain will be forgotten. Earlier the preset options were
459 completely ignored after custom filter chain options had been
460 seen.
462 * xz will create sparse files when decompressing if the uncompressed
463 data contains long sequences of binary zeros. This is done even
464 when writing to standard output that is connected to a regular
465 file and certain additional conditions are met to make it safe.
467 * Support for "xz --list" was added. Combine with --verbose or
468 --verbose --verbose (-vv) for detailed output.
470 * I had hoped that liblzma API would have been stable after
471 4.999.9beta, but there have been a couple of changes in the
472 advanced features, which don't affect most applications:
474 - Index handling code was revised. If you were using the old
475 API, you will get a compiler error (so it's easy to notice).
477 - A subtle but important change was made to the Block handling
478 API. lzma_block.version has to be initialized even for
479 lzma_block_header_decode(). Code that doesn't do it will work
480 for now, but might break in the future, which makes this API
481 change easy to miss.
483 * The major soname has been bumped to 5.0.0. liblzma API and ABI
484 are now stable, so the need to recompile programs linking against
485 liblzma shouldn't arise soon.