1
2This is the README for bzip2/libzip2.
3This version is fully compatible with the previous public releases.
4
5------------------------------------------------------------------
6This file is part of bzip2/libbzip2, a program and library for
7lossless, block-sorting data compression.
8
9bzip2/libbzip2 version 1.0.8 of 13 July 2019
10Copyright (C) 1996-2019 Julian Seward <jseward@acm.org>
11
12Please read the WARNING, DISCLAIMER and PATENTS sections in this file.
13
14This program is released under the terms of the license contained
15in the file LICENSE.
16------------------------------------------------------------------
17
18Complete documentation is available in Postscript form (manual.ps),
19PDF (manual.pdf) or html (manual.html). A plain-text version of the
20manual page is available as bzip2.txt.
21
22
23HOW TO BUILD -- UNIX
24
25Type 'make'. This builds the library libbz2.a and then the programs
26bzip2 and bzip2recover. Six self-tests are run. If the self-tests
27complete ok, carry on to installation:
28
29To install in /usr/local/bin, /usr/local/lib, /usr/local/man and
30/usr/local/include, type
31
32 make install
33
34To install somewhere else, eg, /xxx/yyy/{bin,lib,man,include}, type
35
36 make install PREFIX=/xxx/yyy
37
38If you are (justifiably) paranoid and want to see what 'make install'
39is going to do, you can first do
40
41 make -n install or
42 make -n install PREFIX=/xxx/yyy respectively.
43
44The -n instructs make to show the commands it would execute, but not
45actually execute them.
46
47
48HOW TO BUILD -- UNIX, shared library libbz2.so.
49
50Do 'make -f Makefile-libbz2_so'. This Makefile seems to work for
51Linux-ELF (RedHat 7.2 on an x86 box), with gcc. I make no claims
52that it works for any other platform, though I suspect it probably
53will work for most platforms employing both ELF and gcc.
54
55bzip2-shared, a client of the shared library, is also built, but not
56self-tested. So I suggest you also build using the normal Makefile,
57since that conducts a self-test. A second reason to prefer the
58version statically linked to the library is that, on x86 platforms,
59building shared objects makes a valuable register (%ebx) unavailable
60to gcc, resulting in a slowdown of 10%-20%, at least for bzip2.
61
62Important note for people upgrading .so's from 0.9.0/0.9.5 to version
631.0.X. All the functions in the library have been renamed, from (eg)
64bzCompress to BZ2_bzCompress, to avoid namespace pollution.
65Unfortunately this means that the libbz2.so created by
66Makefile-libbz2_so will not work with any program which used an older
67version of the library. I do encourage library clients to make the
68effort to upgrade to use version 1.0, since it is both faster and more
69robust than previous versions.
70
71
72HOW TO BUILD -- Windows 95, NT, DOS, Mac, etc.
73
74It's difficult for me to support compilation on all these platforms.
75My approach is to collect binaries for these platforms, and put them
76on the master web site (https://sourceware.org/bzip2/). Look there. However
77(FWIW), bzip2-1.0.X is very standard ANSI C and should compile
78unmodified with MS Visual C. If you have difficulties building, you
79might want to read README.COMPILATION.PROBLEMS.
80
81At least using MS Visual C++ 6, you can build from the unmodified
82sources by issuing, in a command shell:
83
84 nmake -f makefile.msc
85
86(you may need to first run the MSVC-provided script VCVARS32.BAT
87 so as to set up paths to the MSVC tools correctly).
88
89
90VALIDATION
91
92Correct operation, in the sense that a compressed file can always be
93decompressed to reproduce the original, is obviously of paramount
94importance. To validate bzip2, I used a modified version of Mark
95Nelson's churn program. Churn is an automated test driver which
96recursively traverses a directory structure, using bzip2 to compress
97and then decompress each file it encounters, and checking that the
98decompressed data is the same as the original.
99
100
101
102Please read and be aware of the following:
103
104WARNING:
105
106 This program and library (attempts to) compress data by
107 performing several non-trivial transformations on it.
108 Unless you are 100% familiar with *all* the algorithms
109 contained herein, and with the consequences of modifying them,
110 you should NOT meddle with the compression or decompression
111 machinery. Incorrect changes can and very likely *will*
112 lead to disastrous loss of data.
113
114
115DISCLAIMER:
116
117 I TAKE NO RESPONSIBILITY FOR ANY LOSS OF DATA ARISING FROM THE
118 USE OF THIS PROGRAM/LIBRARY, HOWSOEVER CAUSED.
119
120 Every compression of a file implies an assumption that the
121 compressed file can be decompressed to reproduce the original.
122 Great efforts in design, coding and testing have been made to
123 ensure that this program works correctly. However, the complexity
124 of the algorithms, and, in particular, the presence of various
125 special cases in the code which occur with very low but non-zero
126 probability make it impossible to rule out the possibility of bugs
127 remaining in the program. DO NOT COMPRESS ANY DATA WITH THIS
128 PROGRAM UNLESS YOU ARE PREPARED TO ACCEPT THE POSSIBILITY, HOWEVER
129 SMALL, THAT THE DATA WILL NOT BE RECOVERABLE.
130
131 That is not to say this program is inherently unreliable.
132 Indeed, I very much hope the opposite is true. bzip2/libbzip2
133 has been carefully constructed and extensively tested.
134
135
136PATENTS:
137
138 To the best of my knowledge, bzip2/libbzip2 does not use any
139 patented algorithms. However, I do not have the resources
140 to carry out a patent search. Therefore I cannot give any
141 guarantee of the above statement.
142
143
144
145WHAT'S NEW IN 0.9.0 (as compared to 0.1pl2) ?
146
147 * Approx 10% faster compression, 30% faster decompression
148 * -t (test mode) is a lot quicker
149 * Can decompress concatenated compressed files
150 * Programming interface, so programs can directly read/write .bz2 files
151 * Less restrictive (BSD-style) licensing
152 * Flag handling more compatible with GNU gzip
153 * Much more documentation, i.e., a proper user manual
154 * Hopefully, improved portability (at least of the library)
155
156WHAT'S NEW IN 0.9.5 ?
157
158 * Compression speed is much less sensitive to the input
159 data than in previous versions. Specifically, the very
160 slow performance caused by repetitive data is fixed.
161 * Many small improvements in file and flag handling.
162 * A Y2K statement.
163
164WHAT'S NEW IN 1.0.x ?
165
166 See the CHANGES file.
167
168I hope you find bzip2 useful. Feel free to contact the developers at
169 bzip2-devel@sourceware.org
170if you have any suggestions or queries. Many people mailed me with
171comments, suggestions and patches after the releases of bzip-0.15,
172bzip-0.21, and bzip2 versions 0.1pl2, 0.9.0, 0.9.5, 1.0.0, 1.0.1,
1731.0.2 and 1.0.3, and the changes in bzip2 are largely a result of this
174feedback. I thank you for your comments.
175
176bzip2's "home" is https://sourceware.org/bzip2/
177
178Julian Seward
179jseward@acm.org
180Cambridge, UK.
181
18218 July 1996 (version 0.15)
18325 August 1996 (version 0.21)
184 7 August 1997 (bzip2, version 0.1)
18529 August 1997 (bzip2, version 0.1pl2)
18623 August 1998 (bzip2, version 0.9.0)
187 8 June 1999 (bzip2, version 0.9.5)
188 4 Sept 1999 (bzip2, version 0.9.5d)
189 5 May 2000 (bzip2, version 1.0pre8)
19030 December 2001 (bzip2, version 1.0.2pre1)
19115 February 2005 (bzip2, version 1.0.3)
19220 December 2006 (bzip2, version 1.0.4)
19310 December 2007 (bzip2, version 1.0.5)
194 6 Sept 2010 (bzip2, version 1.0.6)
19527 June 2019 (bzip2, version 1.0.7)
19613 July 2019 (bzip2, version 1.0.8)
197
1------------------------------------------------------------------
2This file is part of bzip2/libbzip2, a program and library for
3lossless, block-sorting data compression.
4
5bzip2/libbzip2 version 1.0.8 of 13 July 2019
6Copyright (C) 1996-2019 Julian Seward <jseward@acm.org>
7
8Please read the WARNING, DISCLAIMER and PATENTS sections in the
9README file.
10
11This program is released under the terms of the license contained
12in the file LICENSE.
13------------------------------------------------------------------
14
15bzip2 should compile without problems on the vast majority of
16platforms. Using the supplied Makefile, I've built and tested it
17myself for x86-linux and amd64-linux. With makefile.msc, Visual C++
186.0 and nmake, you can build a native Win32 version too. Large file
19support seems to work correctly on at least on amd64-linux.
20
21When I say "large file" I mean a file of size 2,147,483,648 (2^31)
22bytes or above. Many older OSs can't handle files above this size,
23but many newer ones can. Large files are pretty huge -- most files
24you'll encounter are not Large Files.
25
26Early versions of bzip2 (0.1, 0.9.0, 0.9.5) compiled on a wide variety
27of platforms without difficulty, and I hope this version will continue
28in that tradition. However, in order to support large files, I've had
29to include the define -D_FILE_OFFSET_BITS=64 in the Makefile. This
30can cause problems.
31
32The technique of adding -D_FILE_OFFSET_BITS=64 to get large file
33support is, as far as I know, the Recommended Way to get correct large
34file support. For more details, see the Large File Support
35Specification, published by the Large File Summit, at
36
37 http://ftp.sas.com/standards/large.file
38
39As a general comment, if you get compilation errors which you think
40are related to large file support, try removing the above define from
41the Makefile, ie, delete the line
42
43 BIGFILES=-D_FILE_OFFSET_BITS=64
44
45from the Makefile, and do 'make clean ; make'. This will give you a
46version of bzip2 without large file support, which, for most
47applications, is probably not a problem.
48
49Alternatively, try some of the platform-specific hints listed below.
50
51You can use the spewG.c program to generate huge files to test bzip2's
52large file support, if you are feeling paranoid. Be aware though that
53any compilation problems which affect bzip2 will also affect spewG.c,
54alas.
55
56AIX: I have reports that for large file support, you need to specify
57-D_LARGE_FILES rather than -D_FILE_OFFSET_BITS=64. I have not tested
58this myself.
59