1LIBARCHIVE-FORMATS(5) BSD File Formats Manual LIBARCHIVE-FORMATS(5)
2
4 libarchive-formats — archive formats supported by the libarchive library
5
7 The libarchive(3) library reads and writes a variety of streaming archive
8 formats. Generally speaking, all of these archive formats consist of a
9 series of “entries”. Each entry stores a single file system object, such
10 as a file, directory, or symbolic link.
11 The following provides a brief description of each format supported by
13 libarchive, with some information about recognized extensions or limita‐
14 tions of the current library support. Note that just because a format is
15 supported by libarchive does not imply that a program that uses
16 libarchive will support that format. Applications that use libarchive
17 specify which formats they wish to support, though many programs do use
18 libarchive convenience functions to enable all supported formats.
19 Tar Formats
21 The libarchive(3) library can read most tar archives. It can write
22 POSIX-standard “ustar” and “pax interchange” formats as well as v7 tar
23 format and a subset of the legacy GNU tar format.
24 All tar formats store each entry in one or more 512-byte records. The
26 first record is used for file metadata, including filename, timestamp,
27 and mode information, and the file data is stored in subsequent records.
28 Later variants have extended this by either appropriating undefined areas
29 of the header record, extending the header to multiple records, or by
30 storing special entries that modify the interpretation of subsequent en‐
31 tries.
32 gnutar The libarchive(3) library can read most GNU-format tar archives.
34 It currently supports the most popular GNU extensions, including
35 modern long filename and linkname support, as well as atime and
36 ctime data. The libarchive library does not support multi-volume
37 archives, nor the old GNU long filename format. It can read GNU
38 sparse file entries, including the new POSIX-based formats.
39 The libarchive(3) library can write GNU tar format, including
41 long filename and linkname support, as well as atime and ctime
42 data.
43 pax The libarchive(3) library can read and write POSIX-compliant pax
45 interchange format archives. Pax interchange format archives are
46 an extension of the older ustar format that adds a separate entry
47 with additional attributes stored as key/value pairs immediately
48 before each regular entry. The presence of these additional en‐
49 tries is the only difference between pax interchange format and
50 the older ustar format. The extended attributes are of unlimited
51 length and are stored as UTF-8 Unicode strings. Keywords defined
52 in the standard are in all lowercase; vendors are allowed to de‐
53 fine custom keys by preceding them with the vendor name in all
54 uppercase. When writing pax archives, libarchive uses many of
55 the SCHILY keys defined by Joerg Schilling's “star” archiver and
56 a few LIBARCHIVE keys. The libarchive library can read most of
57 the SCHILY keys and most of the GNU keys introduced by GNU tar.
58 It silently ignores any keywords that it does not understand.
59 The pax interchange format converts filenames to Unicode and
61 stores them using the UTF-8 encoding. Prior to libarchive 3.0,
62 libarchive erroneously assumed that the system wide-character
63 routines natively supported Unicode. This caused it to mis-han‐
64 dle non-ASCII filenames on systems that did not satisfy this as‐
65 sumption.
66 restricted pax
68 The libarchive library can also write pax archives in which it
69 attempts to suppress the extended attributes entry whenever pos‐
70 sible. The result will be identical to a ustar archive unless
71 the extended attributes entry is required to store a long file
72 name, long linkname, extended ACL, file flags, or if any of the
73 standard ustar data (user name, group name, UID, GID, etc) cannot
74 be fully represented in the ustar header. In all cases, the re‐
75 sult can be dearchived by any program that can read POSIX-compli‐
76 ant pax interchange format archives. Programs that correctly
77 read ustar format (see below) will also be able to read this for‐
78 mat; any extended attributes will be extracted as separate files
79 stored in PaxHeader directories.
80 ustar The libarchive library can both read and write this format. This
82 format has the following limitations:
83 • Device major and minor numbers are limited to 21 bits. Nodes
84 with larger numbers will not be added to the archive.
85 • Path names in the archive are limited to 255 bytes. (Shorter
86 if there is no / character in exactly the right place.)
87 • Symbolic links and hard links are stored in the archive with
88 the name of the referenced file. This name is limited to 100
89 bytes.
90 • Extended attributes, file flags, and other extended security
91 information cannot be stored.
92 • Archive entries are limited to 8 gigabytes in size.
93 Note that the pax interchange format has none of these restric‐
94 tions. The ustar format is old and widely supported. It is rec‐
95 ommended when compatibility is the primary concern.
96 v7 The libarchive library can read and write the legacy v7 tar for‐
98 mat. This format has the following limitations:
99 • Only regular files, directories, and symbolic links can be
100 archived. Block and character device nodes, FIFOs, and sock‐
101 ets cannot be archived.
102 • Path names in the archive are limited to 100 bytes.
103 • Symbolic links and hard links are stored in the archive with
104 the name of the referenced file. This name is limited to 100
105 bytes.
106 • User and group information are stored as numeric IDs; there
107 is no provision for storing user or group names.
108 • Extended attributes, file flags, and other extended security
109 information cannot be stored.
110 • Archive entries are limited to 8 gigabytes in size.
111 Generally, users should prefer the ustar format for portability
112 as the v7 tar format is both less useful and less portable.
113 The libarchive library also reads a variety of commonly-used extensions
115 to the basic tar format. These extensions are recognized automatically
116 whenever they appear.
117 Numeric extensions.
119 The POSIX standards require fixed-length numeric fields to be
120 written with some character position reserved for terminators.
121 Libarchive allows these fields to be written without terminator
122 characters. This extends the allowable range; in particular, us‐
123 tar archives with this extension can support entries up to 64 gi‐
124 gabytes in size. Libarchive also recognizes base-256 values in
125 most numeric fields. This essentially removes all limitations on
126 file size, modification time, and device numbers.
127 Solaris extensions
129 Libarchive recognizes ACL and extended attribute records written
130 by Solaris tar.
131 The first tar program appeared in Seventh Edition Unix in 1979. The
133 first official standard for the tar file format was the “ustar” (Unix
134 Standard Tar) format defined by POSIX in 1988. POSIX.1-2001 extended the
135 ustar format to create the “pax interchange” format.
136 Cpio Formats
138 The libarchive library can read and write a number of common cpio vari‐
139 ants. A cpio archive stores each entry as a fixed-size header followed
140 by a variable-length filename and variable-length data. Unlike the tar
141 format, the cpio format does only minimal padding of the header or file
142 data. There are several cpio variants, which differ primarily in how
143 they store the initial header: some store the values as octal or hexadec‐
144 imal numbers in ASCII, others as binary values of varying byte order and
145 length.
146 binary The libarchive library transparently reads both big-endian and
148 little-endian variants of the the two binary cpio formats; the
149 original one from PWB/UNIX, and the later, more widely used,
150 variant. This format used 32-bit binary values for file size and
151 mtime, and 16-bit binary values for the other fields. The for‐
152 mats support only the file types present in UNIX at the time of
153 their creation. File sizes are limited to 24 bits in the PWB
154 format, because of the limits of the file system, and to 31 bits
155 in the newer binary format, where signed 32 bit longs were used.
156 odc This is the POSIX standardized format, which is officially known
158 as the “cpio interchange format” or the “octet-oriented cpio
159 archive format” and sometimes unofficially referred to as the
160 “old character format”. This format stores the header contents
161 as octal values in ASCII. It is standard, portable, and immune
162 from byte-order confusion. File sizes and mtime are limited to
163 33 bits (8GB file size), other fields are limited to 18 bits.
164 SVR4/newc
166 The libarchive library can read both CRC and non-CRC variants of
167 this format. The SVR4 format uses eight-digit hexadecimal values
168 for all header fields. This limits file size to 4GB, and also
169 limits the mtime and other fields to 32 bits. The SVR4 format
170 can optionally include a CRC of the file contents, although
171 libarchive does not currently verify this CRC.
172 Cpio first appeared in PWB/UNIX 1.0, which was released within AT&T in
174 1977. PWB/UNIX 1.0 formed the basis of System III Unix, released outside
175 of AT&T in 1981. This makes cpio older than tar, although cpio was not
176 included in Version 7 AT&T Unix. As a result, the tar command became
177 much better known in universities and research groups that used Version
178 7. The combination of the find and cpio utilities provided very precise
179 control over file selection. Unfortunately, the format has many limita‐
180 tions that make it unsuitable for widespread use. Only the POSIX format
181 permits files over 4GB, and its 18-bit limit for most other fields makes
182 it unsuitable for modern systems. In addition, cpio formats only store
183 numeric UID/GID values (not usernames and group names), which can make it
184 very difficult to correctly transfer archives across systems with dissim‐
185 ilar user numbering.
186 Shar Formats
188 A “shell archive” is a shell script that, when executed on a POSIX-com‐
189 pliant system, will recreate a collection of file system objects. The
190 libarchive library can write two different kinds of shar archives:
191 shar The traditional shar format uses a limited set of POSIX commands,
193 including echo(1), mkdir(1), and sed(1). It is suitable for
194 portably archiving small collections of plain text files. How‐
195 ever, it is not generally well-suited for large archives (many
196 implementations of sh(1) have limits on the size of a script) nor
197 should it be used with non-text files.
198 shardump
200 This format is similar to shar but encodes files using
201 uuencode(1) so that the result will be a plain text file regard‐
202 less of the file contents. It also includes additional shell
203 commands that attempt to reproduce as many file attributes as
204 possible, including owner, mode, and flags. The additional com‐
205 mands used to restore file attributes make shardump archives less
206 portable than plain shar archives.
207 ISO9660 format
209 Libarchive can read and extract from files containing ISO9660-compliant
210 CDROM images. In many cases, this can remove the need to burn a physical
211 CDROM just in order to read the files contained in an ISO9660 image. It
212 also avoids security and complexity issues that come with virtual mounts
213 and loopback devices. Libarchive supports the most common Rockridge ex‐
214 tensions and has partial support for Joliet extensions. If both exten‐
215 sions are present, the Joliet extensions will be used and the Rockridge
216 extensions will be ignored. In particular, this can create problems with
217 hardlinks and symlinks, which are supported by Rockridge but not by
218 Joliet.
219 Libarchive reads ISO9660 images using a streaming strategy. This allows
221 it to read compressed images directly (decompressing on the fly) and al‐
222 lows it to read images directly from network sockets, pipes, and other
223 non-seekable data sources. This strategy works well for optimized
224 ISO9660 images created by many popular programs. Such programs collect
225 all directory information at the beginning of the ISO9660 image so it can
226 be read from a physical disk with a minimum of seeking. However, not all
227 ISO9660 images can be read in this fashion.
228 Libarchive can also write ISO9660 images. Such images are fully opti‐
230 mized with the directory information preceding all file data. This is
231 done by storing all file data to a temporary file while collecting direc‐
232 tory information in memory. When the image is finished, libarchive
233 writes out the directory structure followed by the file data. The loca‐
234 tion used for the temporary file can be changed by the usual environment
235 variables.
236 Zip format
238 Libarchive can read and write zip format archives that have uncompressed
239 entries and entries compressed with the “deflate” algorithm. Other zip
240 compression algorithms are not supported. It can extract jar archives,
241 archives that use Zip64 extensions and self-extracting zip archives.
242 Libarchive can use either of two different strategies for reading Zip ar‐
243 chives: a streaming strategy which is fast and can handle extremely large
244 archives, and a seeking strategy which can correctly process self-ex‐
245 tracting Zip archives and archives with deleted members or other in-place
246 modifications.
247 The streaming reader processes Zip archives as they are read. It can
249 read archives of arbitrary size from tape or network sockets, and can de‐
250 code Zip archives that have been separately compressed or encoded. How‐
251 ever, self-extracting Zip archives and archives with certain types of
252 modifications cannot be correctly handled. Such archives require that
253 the reader first process the Central Directory, which is ordinarily lo‐
254 cated at the end of a Zip archive and is thus inaccessible to the stream‐
255 ing reader. If the program using libarchive has enabled seek support,
256 then libarchive will use this to processes the central directory first.
257 In particular, the seeking reader must be used to correctly handle self-
259 extracting archives. Such archives consist of a program followed by a
260 regular Zip archive. The streaming reader cannot parse the initial pro‐
261 gram portion, but the seeking reader starts by reading the Central Direc‐
262 tory from the end of the archive. Similarly, Zip archives that have been
263 modified in-place can have deleted entries or other garbage data that can
264 only be accurately detected by first reading the Central Directory.
265 Archive (library) file format
267 The Unix archive format (commonly created by the ar(1) archiver) is a
268 general-purpose format which is used almost exclusively for object files
269 to be read by the link editor ld(1). The ar format has never been stan‐
270 dardised. There are two common variants: the GNU format derived from
271 SVR4, and the BSD format, which first appeared in 4.4BSD. The two differ
272 primarily in their handling of filenames longer than 15 characters: the
273 GNU/SVR4 variant writes a filename table at the beginning of the archive;
274 the BSD format stores each long filename in an extension area adjacent to
275 the entry. Libarchive can read both extensions, including archives that
276 may include both types of long filenames. Programs using libarchive can
277 write GNU/SVR4 format if they provide an entry called // containing a
278 filename table to be written into the archive before any of the entries.
279 Any entries whose names are not in the filename table will be written us‐
280 ing BSD-style long filenames. This can cause problems for programs such
281 as GNU ld that do not support the BSD-style long filenames.
282 mtree
284 Libarchive can read and write files in mtree(5) format. This format is
285 not a true archive format, but rather a textual description of a file hi‐
286 erarchy in which each line specifies the name of a file and provides spe‐
287 cific metadata about that file. Libarchive can read all of the keywords
288 supported by both the NetBSD and FreeBSD versions of mtree(8), although
289 many of the keywords cannot currently be stored in an archive_entry ob‐
290 ject. When writing, libarchive supports use of the
291 archive_write_set_options(3) interface to specify which keywords should
292 be included in the output. If libarchive was compiled with access to
293 suitable cryptographic libraries (such as the OpenSSL libraries), it can
294 compute hash entries such as sha512 or md5 from file data being written
295 to the mtree writer.
296 When reading an mtree file, libarchive will locate the corresponding
298 files on disk using the contents keyword if present or the regular file‐
299 name. If it can locate and open the file on disk, it will use that to
300 fill in any metadata that is missing from the mtree file and will read
301 the file contents and return those to the program using libarchive. If
302 it cannot locate and open the file on disk, libarchive will return an er‐
303 ror for any attempt to read the entry body.
304 7-Zip
306 Libarchive can read and write 7-Zip format archives. TODO: Need more in‐
307 formation
308 CAB
310 Libarchive can read Microsoft Cabinet ( “CAB”) format archives. TODO:
311 Need more information.
312 LHA
314 TODO: Information about libarchive's LHA support
315 RAR
317 Libarchive has limited support for reading RAR format archives. Cur‐
318 rently, libarchive can read RARv3 format archives which have been either
319 created uncompressed, or compressed using any of the compression methods
320 supported by the RARv3 format. Libarchive can also read self-extracting
321 RAR archives.
322 Warc
324 Libarchive can read and write “web archives”. TODO: Need more informa‐
325 tion
326 XAR
328 Libarchive can read and write the XAR format used by many Apple tools.
329 TODO: Need more information
330
332 ar(1), cpio(1), mkisofs(1), shar(1), tar(1), zip(1), zlib(3), cpio(5),
333 mtree(5), tar(5)
334BSD December 27, 2016 BSD