1
2The .lzma File Format
3=====================
4
5        0. Preface
6           0.1. Notices and Acknowledgements
7           0.2. Changes
8        1. File Format
9           1.1. Header
10                1.1.1. Properties
11                1.1.2. Dictionary Size
12                1.1.3. Uncompressed Size
13           1.2. LZMA Compressed Data
14        2. References
15
16
170. Preface
18
19        This document describes the .lzma file format, which is
20        sometimes also called LZMA_Alone format. It is a legacy file
21        format, which is being or has been replaced by the .xz format.
22        The MIME type of the .lzma format is `application/x-lzma'.
23
24        The most commonly used software to handle .lzma files are
25        LZMA SDK, LZMA Utils, 7-Zip, and XZ Utils. This document
26        describes some of the differences between these implementations
27        and gives hints what subset of the .lzma format is the most
28        portable.
29
30
310.1. Notices and Acknowledgements
32
33        This file format was designed by Igor Pavlov for use in
34        LZMA SDK. This document was written by Lasse Collin
35        <lasse.collin@tukaani.org> using the documentation found
36        from the LZMA SDK.
37
38        This document has been put into the public domain.
39
40
410.2. Changes
42
43        Last modified: 2011-04-12 11:55+0300
44
45
461. File Format
47
48        +-+-+-+-+-+-+-+-+-+-+-+-+-+==========================+
49        |         Header          |   LZMA Compressed Data   |
50        +-+-+-+-+-+-+-+-+-+-+-+-+-+==========================+
51
52        The .lzma format file consist of 13-byte Header followed by
53        the LZMA Compressed Data.
54
55        Unlike the .gz, .bz2, and .xz formats, it is not possible to
56        concatenate multiple .lzma files as is and expect the
57        decompression tool to decode the resulting file as if it were
58        a single .lzma file.
59
60        For example, the command line tools from LZMA Utils and
61        LZMA SDK silently ignore all the data after the first .lzma
62        stream. In contrast, the command line tool from XZ Utils
63        considers the .lzma file to be corrupt if there is data after
64        the first .lzma stream.
65
66
671.1. Header
68
69        +------------+----+----+----+----+--+--+--+--+--+--+--+--+
70        | Properties |  Dictionary Size  |   Uncompressed Size   |
71        +------------+----+----+----+----+--+--+--+--+--+--+--+--+
72
73
741.1.1. Properties
75
76        The Properties field contains three properties. An abbreviation
77        is given in parentheses, followed by the value range of the
78        property. The field consists of
79
80            1) the number of literal context bits (lc, [0, 8]);
81            2) the number of literal position bits (lp, [0, 4]); and
82            3) the number of position bits (pb, [0, 4]).
83
84        The properties are encoded using the following formula:
85
86            Properties = (pb * 5 + lp) * 9 + lc
87
88        The following C code illustrates a straightforward way to
89        decode the Properties field:
90
91            uint8_t lc, lp, pb;
92            uint8_t prop = get_lzma_properties();
93            if (prop > (4 * 5 + 4) * 9 + 8)
94                return LZMA_PROPERTIES_ERROR;
95
96            pb = prop / (9 * 5);
97            prop -= pb * 9 * 5;
98            lp = prop / 9;
99            lc = prop - lp * 9;
100
101        XZ Utils has an additional requirement: lc + lp <= 4. Files
102        which don't follow this requirement cannot be decompressed
103        with XZ Utils. Usually this isn't a problem since the most
104        common lc/lp/pb values are 3/0/2. It is the only lc/lp/pb
105        combination that the files created by LZMA Utils can have,
106        but LZMA Utils can decompress files with any lc/lp/pb.
107
108
1091.1.2. Dictionary Size
110
111        Dictionary Size is stored as an unsigned 32-bit little endian
112        integer. Any 32-bit value is possible, but for maximum
113        portability, only sizes of 2^n and 2^n + 2^(n-1) should be
114        used.
115
116        LZMA Utils creates only files with dictionary size 2^n,
117        16 <= n <= 25. LZMA Utils can decompress files with any
118        dictionary size.
119
120        XZ Utils creates and decompresses .lzma files only with
121        dictionary sizes 2^n and 2^n + 2^(n-1). If some other
122        dictionary size is specified when compressing, the value
123        stored in the Dictionary Size field is a rounded up, but the
124        specified value is still used in the actual compression code.
125
126
1271.1.3. Uncompressed Size
128
129        Uncompressed Size is stored as unsigned 64-bit little endian
130        integer. A special value of 0xFFFF_FFFF_FFFF_FFFF indicates
131        that Uncompressed Size is unknown. End of Payload Marker (*)
132        is used if and only if Uncompressed Size is unknown.
133
134        XZ Utils rejects files whose Uncompressed Size field specifies
135        a known size that is 256 GiB or more. This is to reject false
136        positives when trying to guess if the input file is in the
137        .lzma format. When Uncompressed Size is unknown, there is no
138        limit for the uncompressed size of the file.
139
140        (*) Some tools use the term End of Stream (EOS) marker
141            instead of End of Payload Marker.
142
143
1441.2. LZMA Compressed Data
145
146        Detailed description of the format of this field is out of
147        scope of this document.
148
149
1502. References
151
152        LZMA SDK - The original LZMA implementation
153        http://7-zip.org/sdk.html
154
155        7-Zip
156        http://7-zip.org/
157
158        LZMA Utils - LZMA adapted to POSIX-like systems
159        http://tukaani.org/lzma/
160
161        XZ Utils - The next generation of LZMA Utils
162        http://tukaani.org/xz/
163
164        The .xz file format - The successor of the .lzma format
165        http://tukaani.org/xz/xz-file-format.txt
166
167