forked from xz-mirror/xz
-
Notifications
You must be signed in to change notification settings - Fork 2
/
README
240 lines (181 loc) · 10.1 KB
/
README
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
FXZ Utils
=========
0. Overview
1. Documentation
1.1. Overall documentation
1.2. Documentation for command-line tools
1.3. Documentation for libflzma
2. Version numbering
3. Reporting bugs
4. Translating the fxz tool
5. Other implementations of the .xz format
6. Contact information
0. Overview
-----------
FXZ Utils is a fork of XZ Utils. It adds a multi-threaded radix
match finder and optimized encoder. The documentation, including this
file, is based upon the XZ Utils documentation.
The fxz command-line tool, derived from xz, provides faster
compression than xz on most inputs, but a slightly larger output.
These tests on the 723 MiB 'raspbian-jessie-lite-20151121.tar' from
the Squash Compression Corpus demonstrate performance relative to xz
using 4 threads:
Preset Size Time MiB/s RAM MiB
xz 4 222.6 1:06 10.9 333
fxz 4 222.8 0:42 16.9 50
xz 6 213.4 1:33 7.7 661
fxz 6 214.2 0:53 13.6 92
xz 8 195.9 1:47 6.7 2,629
fxz 8 196.7 1:02 11.5 336
FXZ Utils provide a general-purpose data-compression library plus
command-line tools. The native file format is the .xz format, but
also the legacy .lzma format is supported. The .xz format supports
multiple compression algorithms, which are called "filters" in the
context of FXZ Utils. The primary filter is currently LZMA2. With
typical files, FXZ Utils create about 30 % smaller files than gzip.
To ease adapting support for the .xz format into existing applications
and scripts, the API of libflzma is somewhat similar to the API of the
popular zlib library. For the same reason, the command-line tool fxz
has a command-line syntax similar to that of gzip.
When aiming for the highest compression ratio, the LZMA2 encoder uses
a lot of CPU time and may use, depending on the settings, even
hundreds of megabytes of RAM. However, in fast modes, the LZMA2 encoder
competes with bzip2 in compression speed, RAM usage, and compression
ratio.
LZMA2 is reasonably fast to decompress. It is a little slower than
gzip, but a lot faster than bzip2. Being fast to decompress means
that the .xz format is especially nice when the same file will be
decompressed very many times (usually on different computers), which
is the case e.g. when distributing software packages. In such
situations, it's not too bad if the compression takes some time,
since that needs to be done only once to benefit many people.
With some file types, combining (or "chaining") LZMA2 with an
additional filter can improve the compression ratio. A filter chain may
contain up to four filters, although usually only one or two are used.
For example, putting a BCJ (Branch/Call/Jump) filter before LZMA2
in the filter chain can improve compression ratio of executable files.
Since the .xz format allows adding new filter IDs, it is possible that
some day there will be a filter that is, for example, much faster to
compress than LZMA2 (but probably with worse compression ratio).
Similarly, it is possible that some day there is a filter that will
compress better than LZMA2.
FXZ Utils supports multithreaded compression. FXZ Utils doesn't support
multithreaded decompression yet. It has been planned though and taken
into account when designing the .xz file format. In the future, files
that were created in threaded mode can be decompressed in threaded
mode too.
1. Documentation
----------------
1.1. Overall documentation
README This file
INSTALL.generic Generic install instructions for those not familiar
with packages using GNU Autotools
INSTALL Installation instructions specific to FXZ Utils
PACKAGERS Information to packagers of FXZ Utils
COPYING FXZ Utils copyright and license information
COPYING.GPLv2 GNU General Public License version 2
COPYING.GPLv3 GNU General Public License version 3
COPYING.LGPLv2.1 GNU Lesser General Public License version 2.1
AUTHORS The main authors of FXZ Utils
THANKS Incomplete list of people who have helped making
this software
NEWS User-visible changes between FXZ Utils releases
ChangeLog Detailed list of changes (commit log)
TODO Known bugs and some sort of to-do list
Note that only some of the above files are included in binary
packages.
1.2. Documentation for command-line tools
The command-line tools are documented as man pages. In source code
releases (and possibly also in some binary packages), the man pages
are also provided in plain text (ASCII only) and PDF formats in the
directory "doc/man" to make the man pages more accessible to those
whose operating system doesn't provide an easy way to view man pages.
1.3. Documentation for libflzma
The libflzma API headers include short docs about each function
and data type as Doxygen tags. These docs should be quite OK as
a quick reference.
There are a few example/tutorial programs that should help in
getting started with libflzma. In the source package the examples
are in "doc/examples" and in binary packages they may be under
"examples" in the same directory as this README.
Since the libflzma API has similarities to the zlib API, some people
may find it useful to read the zlib docs and tutorial too:
http://zlib.net/manual.html
http://zlib.net/zlib_how.html
2. Version numbering
--------------------
The version number format of FXZ Utils is X.Y.ZS:
- X is the major version. When this is incremented, the library
API and ABI break.
- Y is the minor version. It is incremented when new features
are added without breaking the existing API or ABI. An even Y
indicates a stable release and an odd Y indicates unstable
(alpha or beta version).
- Z is the revision. This has a different meaning for stable and
unstable releases:
* Stable: Z is incremented when bugs get fixed without adding
any new features. This is intended to be convenient for
downstream distributors that want bug fixes but don't want
any new features to minimize the risk of introducing new bugs.
* Unstable: Z is just a counter. API or ABI of features added
in earlier unstable releases having the same X.Y may break.
- S indicates stability of the release. It is missing from the
stable releases, where Y is an even number. When Y is odd, S
is either "alpha" or "beta" to make it very clear that such
versions are not stable releases. The same X.Y.Z combination is
not used for more than one stability level, i.e. after X.Y.Zalpha,
the next version can be X.Y.(Z+1)beta but not X.Y.Zbeta.
3. Reporting bugs
-----------------
If the bug exists in XZ Utils too, it would be better to report it
to the maintainer of that project, as it is widely used and XZ
fixes can easily be incorporated into FXZ. For information see
https://tukaani.org/xz/
Naturally it is easiest for me if you already know what causes the
unexpected behavior. Even better if you have a patch to propose.
However, quite often the reason for unexpected behavior is unknown,
so here are a few things to do before sending a bug report:
1. Try to create a small example how to reproduce the issue.
2. Compile FXZ Utils with debugging code using configure switches
--enable-debug and, if possible, --disable-shared. If you are
using GCC, use CFLAGS='-O0 -ggdb3'. Don't strip the resulting
binaries.
3. Turn on core dumps. The exact command depends on your shell;
for example in GNU bash it is done with "ulimit -c unlimited",
and in tcsh with "limit coredumpsize unlimited".
4. Try to reproduce the suspected bug. If you get "assertion failed"
message, be sure to include the complete message in your bug
report. If the application leaves a coredump, get a backtrace
using gdb:
$ gdb /path/to/app-binary # Load the app to the debugger.
(gdb) core core # Open the coredump.
(gdb) bt # Print the backtrace. Copy & paste to bug report.
(gdb) quit # Quit gdb.
Report your bug via email (see Contact information below).
Don't send core dump files or any executables. If you have a small
example file(s) (total size less than 256 KiB), please include
it/them as an attachment. If you have bigger test files, put them
online somewhere and include a URL to the file(s) in the bug report.
Always include the exact version number of FXZ Utils in the bug report.
If you are using a snapshot from the git repository, use "git describe"
to get the exact snapshot version. If you are using FXZ Utils shipped
in an operating system distribution, mention the distribution name,
distribution version, and exact fxz package version; if you cannot
repeat the bug with the code compiled from unpatched source code,
you probably need to report a bug to your distribution's bug tracking
system.
4. Other implementations of the .xz format
------------------------------------------
7-Zip and the p7zip port of 7-Zip support the .xz format starting
from the version 9.00alpha.
http://7-zip.org/
http://p7zip.sourceforge.net/
XZ Embedded is a limited implementation written for use in the Linux
kernel, but it is also suitable for other embedded use.
https://tukaani.org/xz/embedded.html
5. Contact information
----------------------
If you have questions, bug reports, patches etc. related to FXZ
Utils, contact Conor McCarthy <conor.mccarthy.444@gmail.com>.
I'm sometimes slow at replying. If you haven't got a reply within two
weeks, assume that your email has got lost and resend it.