9 - A memory leak was fixed.
11 - lzma_stream_buffer_encode() no longer creates an empty .xz
12 Block if encoding an empty buffer. Such an empty Block with
13 LZMA2 data would trigger a bug in 5.0.1 and older (see the
14 first bullet point in 5.0.2 notes). When releasing 5.0.2,
15 I thought that no encoder creates this kind of files but
18 - Validate function arguments better in a few functions. Most
19 importantly, specifying an unsupported integrity check to
20 lzma_stream_buffer_encode() no longer creates a corrupt .xz
21 file. Probably no application tries to do that, so this
22 shouldn't be a big problem in practice.
24 - Document that lzma_block_buffer_encode(),
25 lzma_easy_buffer_encode(), lzma_stream_encoder(), and
26 lzma_stream_buffer_encode() may return LZMA_UNSUPPORTED_CHECK.
28 - The return values of the _memusage() functions are now
31 * Fix command name detection in xzgrep. xzegrep and xzfgrep now
32 correctly use egrep and fgrep instead of grep.
34 * French translation was added.
39 * LZMA2 decompressor now correctly accepts LZMA2 streams with no
40 uncompressed data. Previously it considered them corrupt. The
41 bug can affect applications that use raw LZMA2 streams. It is
42 very unlikely to affect .xz files because no compressor creates
43 .xz files with empty LZMA2 streams. (Empty .xz files are a
44 different thing than empty LZMA2 streams.)
46 * "xz --suffix=.foo filename.foo" now refuses to compress the
47 file due to it already having the suffix .foo. It was already
48 documented on the man page, but the code lacked the test.
50 * "xzgrep -l foo bar.xz" works now.
52 * Polish translation was added.
57 * xz --force now (de)compresses files that have setuid, setgid,
58 or sticky bit set and files that have multiple hard links.
59 The man page had it documented this way already, but the code
62 * gzip and bzip2 support in xzdiff was fixed.
66 * Minor fix to Czech translation
71 Only the most important changes compared to 4.999.9beta are listed
72 here. One change is especially important:
74 * The memory usage limit is now disabled by default. Some scripts
75 written before this change may have used --memory=max on xz command
76 line or in XZ_OPT. THESE USES OF --memory=max SHOULD BE REMOVED
77 NOW, because they interfere with user's ability to set the memory
78 usage limit himself. If user-specified limit causes problems to
79 your script, blame the user.
81 Other significant changes:
83 * Added support for XZ_DEFAULTS environment variable. This variable
84 allows users to set default options for xz, e.g. default memory
85 usage limit or default compression level. Scripts that use xz
86 must never set or unset XZ_DEFAULTS. Scripts should use XZ_OPT
87 instead if they need a way to pass options to xz via an
90 * The compression settings associated with the preset levels
91 -0 ... -9 have been changed. --extreme was changed a little too.
92 It is now less likely to make compression worse, but with some
93 files the new --extreme may compress slightly worse than the old
96 * If a preset level (-0 ... -9) is specified after a custom filter
97 chain options have been used (e.g. --lzma2), the custom filter
98 chain will be forgotten. Earlier the preset options were
99 completely ignored after custom filter chain options had been
102 * xz will create sparse files when decompressing if the uncompressed
103 data contains long sequences of binary zeros. This is done even
104 when writing to standard output that is connected to a regular
105 file and certain additional conditions are met to make it safe.
107 * Support for "xz --list" was added. Combine with --verbose or
108 --verbose --verbose (-vv) for detailed output.
110 * I had hoped that liblzma API would have been stable after
111 4.999.9beta, but there have been a couple of changes in the
112 advanced features, which don't affect most applications:
114 - Index handling code was revised. If you were using the old
115 API, you will get a compiler error (so it's easy to notice).
117 - A subtle but important change was made to the Block handling
118 API. lzma_block.version has to be initialized even for
119 lzma_block_header_decode(). Code that doesn't do it will work
120 for now, but might break in the future, which makes this API
123 * The major soname has been bumped to 5.0.0. liblzma API and ABI
124 are now stable, so the need to recompile programs linking against
125 liblzma shouldn't arise soon.