TAR(5) File Formats Manual TAR(5)
NAME
tar — format of tape archive files
DESCRIPTION
The tar archive format collects any number of files, directories, and other file system objects (symbolic links, device nodes, etc.) into a single stream of bytes. The format was originally designed to be used with tape drives that operate with fixed-size blocks, but is widely used as a general packaging mechanism.
General Format
A tar archive consists of a series of 512-byte records. Each file system object requires a header record which stores basic metadata (pathname, owner, permissions, etc.) and zero or more records containing any file data. The end of the archive is indicated by two records consisting entirely of zero bytes.
For compatibility with tape drives that use fixed block sizes, programs that read or write tar files always read or write a fixed number of records with each I/O operation. These “blocks” are always a multiple of the record size. The maximum block size supported by early implementations was 10240 bytes or 20 records. This is still the default for most implementations although block sizes of 1MiB (2048 records) or larger are commonly used with modern high-speed tape drives. (Note: the terms “block” and “record” here are not entirely standard; this document follows the convention established by John Gilmore in documenting pdtar.)
Old-Style Archive Format
The original tar archive format has been extended many times to include additional information that various implementors found necessary. This section describes the variant implemented by the tar command included in Version 7 AT&T UNIX, which seems to be the earliest widely-used version of the tar program.
The header record for an old-style tar archive consists of the following:
struct header_old_tar {
|
char name[100]; |
||
|
char mode[8]; |
||
|
char uid[8]; |
||
|
char gid[8]; |
||
|
char size[12]; |
||
|
char mtime[12]; |
||
|
char checksum[8]; |
||
|
char linkflag[1]; |
||
|
char linkname[100]; |
||
|
char pad[255]; |
};
All unused bytes in the header record are filled with nulls.
name
Pathname, stored as a null-terminated string. Early tar implementations only stored regular files (including hardlinks to those files). One common early convention used a trailing "/" character to indicate a directory name, allowing directory permissions and owner information to be archived and restored.
mode
File mode, stored as an octal number in ASCII.
uid, gid
User id and group id of owner, as octal numbers in ASCII.
size
Size of file, as octal number in ASCII. For regular files only, this indicates the amount of data that follows the header. In particular, this field was ignored by early tar implementations when extracting hardlinks. Modern writers should always store a zero length for hardlink entries.
mtime
Modification time of file, as an octal number in ASCII. This indicates the number of seconds since the start of the epoch, 00:00:00 UTC January 1, 1970. Note that negative values should be avoided here, as they are handled inconsistently.
checksum
Header checksum, stored as an octal number in ASCII. To compute the checksum, set the checksum field to all spaces, then sum all bytes in the header using unsigned arithmetic. This field should be stored as six octal digits followed by a null and a space character. Note that many early implementations of tar used signed arithmetic for the checksum field, which can cause interoperability problems when transferring archives between systems. Modern robust readers compute the checksum both ways and accept the header if either computation matches.
linkflag, linkname
In order to preserve hardlinks and conserve tape, a file with multiple links is only written to the archive the first time it is encountered. The next time it is encountered, the linkflag is set to an ASCII ‘1’ and the linkname field holds the first name under which this file appears. (Note that regular files have a null value in the linkflag field.)
Early tar implementations varied in how they terminated these fields. The tar command in Version 7 AT&T UNIX used the following conventions (this is also documented in early BSD manpages): the pathname must be null-terminated; the mode, uid, and gid fields must end in a space and a null byte; the size and mtime fields must end in a space; the checksum is terminated by a null and a space. Early implementations filled the numeric fields with leading spaces. This seems to have been common practice until the IEEE Std 1003.1-1988 (“POSIX.1”) standard was released. For best portability, modern implementations should fill the numeric fields with leading zeros.
Pre-POSIX Archives
An early draft of IEEE Std 1003.1-1988 (“POSIX.1”) served as the basis for John Gilmore’s pdtar program and many system implementations from the late 1980s and early 1990s. These archives generally follow the POSIX ustar format described below with the following variations:
•
The magic value consists of the five characters “ustar” followed by a space. The version field contains a space character followed by a null.
•
The numeric fields are generally filled with leading spaces (not leading zeros as recommended in the final standard).
•
The prefix field is often not used, limiting pathnames to the 100 characters of old-style archives.
POSIX ustar Archives
IEEE Std 1003.1-1988 (“POSIX.1”) defined a standard tar file format to be read and written by compliant implementations of tar(1). This format is often called the “ustar” format, after the magic value used in the header. (The name is an acronym for “Unix Standard TAR”.) It extends the historic format with new fields:
struct header_posix_ustar {
|
char name[100]; |
||
|
char mode[8]; |
||
|
char uid[8]; |
||
|
char gid[8]; |
||
|
char size[12]; |
||
|
char mtime[12]; |
||
|
char checksum[8]; |
||
|
char typeflag[1]; |
||
|
char linkname[100]; |
||
|
char magic[6]; |
||
|
char version[2]; |
||
|
char uname[32]; |
||
|
char gname[32]; |
||
|
char devmajor[8]; |
||
|
char devminor[8]; |
||
|
char prefix[155]; |
||
|
char pad[12]; |
};
typeflag
Type of entry. POSIX extended the earlier linkflag field with several new type values:
“0”
Regular file. NUL should be treated as a synonym, for compatibility purposes.
“1”
Hard link.
“2”
Symbolic link.
“3”
Character device node.
“4”
Block device node.
“5”
Directory.
“6”
FIFO node.
“7”
Reserved.
Other
A POSIX-compliant implementation must treat any unrecognized typeflag value as a regular file. In particular, writers should ensure that all entries have a valid filename so that they can be restored by readers that do not support the corresponding extension. Uppercase letters "A" through "Z" are reserved for custom extensions. Note that sockets and whiteout entries are not archivable.
It is worth noting that the size field, in particular, has different meanings depending on the type. For regular files, of course, it indicates the amount of data following the header. For directories, it may be used to indicate the total size of all files in the directory, for use by operating systems that pre-allocate directory space. For all other types, it should be set to zero by writers and ignored by readers.
magic
Contains the magic value “ustar” followed by a NUL byte to indicate that this is a POSIX standard archive. Full compliance requires the uname and gname fields be properly set.
version
Version. This should be “00” (two copies of the ASCII digit zero) for POSIX standard archives.
uname, gname
User and group names, as null-terminated ASCII strings. These should be used in preference to the uid/gid values when they are set and the corresponding names exist on the system.
devmajor, devminor
Major and minor numbers for character device or block device entry.
name, prefix
If the pathname is too long to fit in the 100 bytes provided by the standard format, it can be split at any / character with the first portion going into the prefix field. If the prefix field is not empty, the reader will prepend the prefix value and a / character to the regular name field to obtain the full pathname. The standard does not require a trailing / character on directory names, though most implementations still include this for compatibility reasons.
Note that all unused bytes must be set to NUL.
Field termination is specified slightly differently by POSIX than by previous implementations. The magic, uname, and gname fields must have a trailing NUL. The pathname, linkname, and prefix fields must have a trailing NUL unless they fill the entire field. (In particular, it is possible to store a 256-character pathname if it happens to have a / as the 156th character.) POSIX requires numeric fields to be zero-padded in the front, and requires them to be terminated with either space or NUL characters.
Currently, most tar implementations comply with the ustar format, occasionally extending it by adding new fields to the blank area at the end of the header record.
Numeric Extensions
There have been several attempts to extend the range of sizes or times supported by modifying how numbers are stored in the header.
One obvious extension to increase the size of files is to eliminate the terminating characters from the various numeric fields. For example, the standard only allows the size field to contain 11 octal digits, reserving the twelfth byte for a trailing NUL character. Allowing 12 octal digits allows file sizes up to 64 GB.
Another extension, utilized by GNU tar, star, and other newer tar implementations, permits binary numbers in the standard numeric fields. This is flagged by setting the high bit of the first byte. The remainder of the field is treated as a signed twos-complement value. This permits 95-bit values for the length and time fields and 63-bit values for the uid, gid, and device numbers. In particular, this provides a consistent way to handle negative time values. GNU tar supports this extension for the length, mtime, ctime, and atime fields. Joerg Schilling’s star program and the libarchive library support this extension for all numeric fields. Note that this extension is largely obsoleted by the extended attribute record provided by the pax interchange format.
Another early GNU extension allowed base-64 values rather than octal. This extension was short-lived and is no longer supported by any implementation.
Pax Interchange Format
There are many attributes that cannot be portably stored in a POSIX ustar archive. IEEE Std 1003.1-2001 (“POSIX.1”) defined a “pax interchange format” that uses two new types of entries to hold text-formatted metadata that applies to following entries. Note that a pax interchange format archive is a ustar archive in every respect. The new data is stored in ustar-compatible archive entries that use the “x” or “g” typeflag. In particular, older implementations that do not fully support these extensions will extract the metadata into regular files, where the metadata can be examined as necessary.
An entry in a pax interchange format archive consists of one or two standard ustar entries, each with its own header and data. The first optional entry stores the extended attributes for the following entry. This optional first entry has an "x" typeflag and a size field that indicates the total size of the extended attributes. The extended attributes themselves are stored as a series of text-format lines encoded in the portable UTF-8 encoding. Each line consists of a decimal number, a space, a key string, an equals sign, a value string, and a new line. The decimal number indicates the length of the entire line, including the initial length field and the trailing newline. An example of such a field is:
25 ctime=1084839148.1212\n
Keys in all lowercase are standard keys. Vendors can add their own keys by prefixing them with an all uppercase vendor name and a period. Note that, unlike the historic header, numeric values are stored using decimal, not octal. A description of some common keys follows:
atime, ctime, mtime
File access, inode change, and modification times. These fields can be negative or include a decimal point and a fractional value.
hdrcharset
The character set used by the pax extension values. By default, all textual values in the pax extended attributes are assumed to be in UTF-8, including pathnames, user names, and group names. In some cases, it is not possible to translate local conventions into UTF-8. If this key is present and the value is the six-character ASCII string “BINARY”, then all textual values are assumed to be in a platform-dependent multi-byte encoding. Note that there are only two valid values for this key: “BINARY” or “ISO-IR 10646 2000 UTF-8”. No other values are permitted by the standard, and the latter value should generally not be used as it is the default when this key is not specified. In particular, this flag should not be used as a general mechanism to allow filenames to be stored in arbitrary encodings.
uname, uid, gname, gid
User name, group name, and numeric UID and GID values. The user name and group name stored here are encoded in UTF8 and can thus include non-ASCII characters. The UID and GID fields can be of arbitrary length.
linkpath
The full path of the linked-to file. Note that this is encoded in UTF8 and can thus include non-ASCII characters.
path
The full pathname of the entry. Note that this is encoded in UTF8 and can thus include non-ASCII characters.
realtime.*, security.*
These keys are reserved and may be used for future standardization.
size
The size of the file. Note that there is no length limit on this field, allowing conforming archives to store files much larger than the historic 8GB limit.
SCHILY.*
Vendor-specific attributes used by Joerg Schilling’s star implementation.
SCHILY.acl.access, SCHILY.acl.default, SCHILY.acl.ace
Stores the access, default and NFSv4 ACLs as textual strings in a format that is an extension of the format specified by POSIX.1e draft 17. In particular, each user or group access specification can include an additional colon-separated field with the numeric UID or GID. This allows ACLs to be restored on systems that may not have complete user or group information available (such as when NIS/YP or LDAP services are temporarily unavailable).
SCHILY.devminor, SCHILY.devmajor
The full minor and major numbers for device nodes.
SCHILY.fflags
The file flags.
SCHILY.realsize
The full size of the file on disk. XXX explain? XXX
SCHILY.dev, SCHILY.ino, SCHILY.nlinks
The device number, inode number, and link count for the entry. In particular, note that a pax interchange format archive using Joerg Schilling’s SCHILY.* extensions can store all of the data from struct stat.
LIBARCHIVE.*
Vendor-specific attributes used by the libarchive library and programs that use it.
LIBARCHIVE.creationtime
The time when the file was created. (This should not be confused with the POSIX “ctime” attribute, which refers to the time when the file metadata was last changed.)
LIBARCHIVE.xattr.namespace.key
Libarchive stores POSIX.1e-style extended attributes using keys of this form. The key value is URL-encoded: All non-ASCII characters and the two special characters “=” and “%” are encoded as “%” followed by two uppercase hexadecimal digits. The value of this key is the extended attribute value encoded in base 64. XXX Detail the base-64 format here XXX
VENDOR.*
XXX document other vendor-specific extensions XXX
Any values stored in an extended attribute override the corresponding values in the regular tar header. Note that compliant readers should ignore the regular fields when they are overridden. This is important, as existing archivers are known to store non-compliant values in the standard header fields in this situation. There are no limits on length for any of these fields. In particular, numeric fields can be arbitrarily large. All text fields are encoded in UTF8. Compliant writers should store only portable 7-bit ASCII characters in the standard ustar header and use extended attributes whenever a text value contains non-ASCII characters.
In addition to the x entry described above, the pax interchange format also supports a g entry. The g entry is identical in format, but specifies attributes that serve as defaults for all subsequent archive entries. The g entry is not widely used.
Besides the new x and g entries, the pax interchange format has a few other minor variations from the earlier ustar format. The most troubling one is that hardlinks are permitted to have data following them. This allows readers to restore any hardlink to a file without having to rewind the archive to find an earlier entry. However, it creates complications for robust readers, as it is no longer clear whether or not they should ignore the size field for hardlink entries.
GNU Tar Archives
The GNU tar program started with a pre-POSIX format similar to that described earlier and has extended it using several different mechanisms: It added new fields to the empty space in the header (some of which was later used by POSIX for conflicting purposes); it allowed the header to be continued over multiple records; and it defined new entries that modify following entries (similar in principle to the x entry described above, but each GNU special entry is single-purpose, unlike the general-purpose x entry). As a result, GNU tar archives are not POSIX compatible, although more lenient POSIX-compliant readers can successfully extract most GNU tar archives.
struct header_gnu_tar {
|
char name[100]; |
|||
|
char mode[8]; |
|||
|
char uid[8]; |
|||
|
char gid[8]; |
|||
|
char size[12]; |
|||
|
char mtime[12]; |
|||
|
char checksum[8]; |
|||
|
char typeflag[1]; |
|||
|
char linkname[100]; |
|||
|
char magic[6]; |
|||
|
char version[2]; |
|||
|
char uname[32]; |
|||
|
char gname[32]; |
|||
|
char devmajor[8]; |
|||
|
char devminor[8]; |
|||
|
char atime[12]; |
|||
|
char ctime[12]; |
|||
|
char offset[12]; |
|||
|
char longnames[4]; |
|||
|
char unused[1]; |
|||
|
struct { |
|||
|
char offset[12]; |
|||
|
char numbytes[12]; |
|||
|
} sparse[4]; |
|||
|
char isextended[1]; |
|||
|
char realsize[12]; |
|||
|
char pad[17]; |
};
typeflag
GNU tar uses the following special entry types, in addition to those defined by POSIX:
7
GNU tar treats type "7" records identically to type "0" records, except on one obscure RTOS where they are used to indicate the pre-allocation of a contiguous file on disk.
D
This indicates a directory entry. Unlike the POSIX-standard "5" typeflag, the header is followed by data records listing the names of files in this directory. Each name is preceded by an ASCII "Y" if the file is stored in this archive or "N" if the file is not stored in this archive. Each name is terminated with a null, and an extra null marks the end of the name list. The purpose of this entry is to support incremental backups; a program restoring from such an archive may wish to delete files on disk that did not exist in the directory when the archive was made.
Note that the "D" typeflag specifically violates POSIX, which requires that unrecognized typeflags be restored as normal files. In this case, restoring the "D" entry as a file could interfere with subsequent creation of the like-named directory.
K
The data for this entry is a long linkname for the following regular entry.
L
The data for this entry is a long pathname for the following regular entry.
M
This is a continuation of the last file on the previous volume. GNU multi-volume archives guarantee that each volume begins with a valid entry header. To ensure this, a file may be split, with part stored at the end of one volume, and part stored at the beginning of the next volume. The "M" typeflag indicates that this entry continues an existing file. Such entries can only occur as the first or second entry in an archive (the latter only if the first entry is a volume label). The size field specifies the size of this entry. The offset field at bytes 369-380 specifies the offset where this file fragment begins. The realsize field specifies the total size of the file (which must equal size plus offset). When extracting, GNU tar checks that the header file name is the one it is expecting, that the header offset is in the correct sequence, and that the sum of offset and size is equal to realsize.
N
Type "N" records are no longer generated by GNU tar. They contained a list of files to be renamed or symlinked after extraction; this was originally used to support long names. The contents of this record are a text description of the operations to be done, in the form “Rename %s to %s\n” or “Symlink %s to %s\n”; in either case, both filenames are escaped using K&R C syntax. Due to security concerns, "N" records are now generally ignored when reading archives.
S
This is a “sparse” regular file. Sparse files are stored as a series of fragments. The header contains a list of fragment offset/length pairs. If more than four such entries are required, the header is extended as necessary with “extra” header extensions (an older format that is no longer used), or “sparse” extensions.
V
The name field should be interpreted as a tape/volume header name. This entry should generally be ignored on extraction.
magic
The magic field holds the five characters “ustar” followed by a space. Note that POSIX ustar archives have a trailing null.
version
The version field holds a space character followed by a null. Note that POSIX ustar archives use two copies of the ASCII digit “0”.
atime, ctime
The time the file was last accessed and the time of last change of file information, stored in octal as with mtime.
longnames
This field is apparently no longer used.
Sparse offset / numbytes
Each such structure specifies a single fragment of a sparse file. The two fields store values as octal numbers. The fragments are each padded to a multiple of 512 bytes in the archive. On extraction, the list of fragments is collected from the header (including any extension headers), and the data is then read and written to the file at appropriate offsets.
isextended
If this is set to non-zero, the header will be followed by additional “sparse header” records. Each such record contains information about as many as 21 additional sparse blocks as shown here:
struct gnu_sparse_header {
|
struct { |
|||
|
char offset[12]; |
|||
|
char numbytes[12]; |
|||
|
} sparse[21]; |
|||
|
char isextended[1]; |
|||
|
char padding[7]; |
};
realsize
A binary representation of the file’s complete size, with a much larger range than the POSIX file size. In particular, with M type files, the current entry is only a portion of the file. In that case, the POSIX size field will indicate the size of this entry; the realsize field will indicate the total size of the file.
GNU tar pax archives
GNU tar 1.14 (XXX check this XXX) and later will write pax interchange format archives when you specify the --posix flag. This format follows the pax interchange format closely, using some SCHILY tags and introducing new keywords to store sparse file information. There have been three iterations of the sparse file support, referred to as “0.0”, “0.1”, and “1.0”.
GNU.sparse.numblocks, GNU.sparse.offset, GNU.sparse.numbytes, GNU.sparse.size
The “0.0” format used an initial GNU.sparse.numblocks attribute to indicate the number of blocks in the file, a pair of GNU.sparse.offset and GNU.sparse.numbytes to indicate the offset and size of each block, and a single GNU.sparse.size to indicate the full size of the file. This is not the same as the size in the tar header because the latter value does not include the size of any holes. This format required that the order of attributes be preserved and relied on readers accepting multiple appearances of the same attribute names, which is not officially permitted by the standards.
GNU.sparse.map
The “0.1” format used a single attribute that stored a comma-separated list of decimal numbers. Each pair of numbers indicated the offset and size, respectively, of a block of data. This does not work well if the archive is extracted by an archiver that does not recognize this extension, since many pax implementations simply discard unrecognized attributes.
GNU.sparse.major, GNU.sparse.minor, GNU.sparse.name, GNU.sparse.realsize
The “1.0” format stores the sparse block map in one or more 512-byte blocks prepended to the file data in the entry body. The pax attributes indicate the existence of this map (via the GNU.sparse.major and GNU.sparse.minor fields) and the full size of the file. The GNU.sparse.name holds the true name of the file. To avoid confusion, the name stored in the regular tar header is a modified name so that extraction errors will be apparent to users.
Solaris Tar
XXX More Details Needed XXX
Solaris tar (beginning with SunOS XXX 5.7 ?? XXX) supports an “extended” format that is fundamentally similar to pax interchange format, with the following differences:
•
Extended attributes are stored in an entry whose type is X, not x, as used by pax interchange format. The detailed format of this entry appears to be the same as detailed above for the x entry.
•
An additional A header is used to store an ACL for the following regular entry. The body of this entry contains a seven-digit octal number followed by a zero byte, followed by the textual ACL description. The octal value is the number of ACL entries plus a constant that indicates the ACL type: 01000000 for POSIX.1e ACLs and 03000000 for NFSv4 ACLs.
AIX Tar
XXX More details needed XXX
AIX Tar uses a ustar-formatted header with the type A for storing coded ACL information. Unlike the Solaris format, AIX tar writes this header after the regular file body to which it applies. The pathname in this header is either NFS4 or AIXC to indicate the type of ACL stored. The actual ACL is stored in platform-specific binary format.
Mac OS X Tar
The tar distributed with Apple’s Mac OS X stores most regular files as two separate files in the tar archive. The two files have the same name except that the first one has “._” prepended to the last path element. This special file stores an AppleDouble-encoded binary blob with additional metadata about the second file, including ACL, extended attributes, and resources. To recreate the original file on disk, each separate file can be extracted and the Mac OS X copyfile() function can be used to unpack the separate metadata file and apply it to th regular file. Conversely, the same function provides a “pack” option to encode the extended metadata from a file into a separate file whose contents can then be put into a tar archive.
Note that the Apple extended attributes interact badly with long filenames. Since each file is stored with the full name, a separate set of extensions needs to be included in the archive for each one, doubling the overhead required for files with long names.
Summary of tar type codes
The following list is a condensed summary of the type codes used in tar header records generated by different tar implementations. More details about specific implementations can be found above:
NUL
Early tar programs stored a zero byte for regular files.
0
POSIX standard type code for a regular file.
1
POSIX standard type code for a hard link description.
2
POSIX standard type code for a symbolic link description.
3
POSIX standard type code for a character device node.
4
POSIX standard type code for a block device node.
5
POSIX standard type code for a directory.
6
POSIX standard type code for a FIFO.
7
POSIX reserved.
7
GNU tar used for pre-allocated files on some systems.
A
Solaris tar ACL description stored prior to a regular file header.
A
AIX tar ACL description stored after the file body.
D
GNU tar directory dump.
K
GNU tar long linkname for the following header.
L
GNU tar long pathname for the following header.
M
GNU tar multivolume marker, indicating the file is a continuation of a file from the previous volume.
N
GNU tar long filename support. Deprecated.
S
GNU tar sparse regular file.
V
GNU tar tape/volume header name.
X
Solaris tar general-purpose extension header.
g
POSIX pax interchange format global extensions.
x
POSIX pax interchange format per-file extensions.
SEE ALSO
ar(1), pax(1), tar(1)
STANDARDS
The tar utility is no longer a part of POSIX or the Single Unix Standard. It last appeared in Version 2 of the Single UNIX Specification (“SUSv2”). It has been supplanted in subsequent standards by pax(1). The ustar format is currently part of the specification for the pax(1) utility. The pax interchange file format is new with IEEE Std 1003.1-2001 (“POSIX.1”).
HISTORY
A tar command appeared in Seventh Edition Unix, which was released in January, 1979. It replaced the tp program from Fourth Edition Unix which in turn replaced the tap program from First Edition Unix. John Gilmore’s pdtar public-domain implementation (circa 1987) was highly influential and formed the basis of GNU tar (circa 1988). Joerg Shilling’s star archiver is another open-source (CDDL) archiver (originally developed circa 1985) which features complete support for pax interchange format.
This documentation was written as part of the libarchive and bsdtar project by Tim Kientzle <kientzle@@FreeBSD.org>. Debian December 27, 2016 TAR(5)
TAR(5) BSD File Formats Manual TAR(5)
d27 2 a28 2tar — format of tape archive files
d32 1 a32 1The tar archive format d40 2 a41 8
General
Format
A tar archive consists of a series of 512-byte
records. Each file system object requires a header record
which stores basic metadata (pathname, owner, permissions,
etc.) and zero or more records containing any file data. The
end of the archive is indicated by two records consisting
entirely of zero bytes.
For d65 9 a73 8
Old-Style
Archive Format
The original tar archive format has been extended many times
to include additional information that various implementors
found necessary. This section describes the variant
implemented by the tar command included in Version 7
AT&T UNIX, which seems to be the earliest widely-used
version of the tar program.
The header d79 1 a79 1
struct d85 2 a86 2
};
d168 1 a168 1All unused bytes in the header d173 1 a173 1
Pathname, d183 1 a183 1
File mode, d188 1 a188 1
User id and group id of owner, d193 1 a193 1
Size of file, d202 1 a202 1
Modification d210 1 a210 1
Header checksum, stored as an d225 1 a225 1
In order to preserve hardlinks d229 1 a229 1 an ASCII ’1’ and the linkname field holds d234 1 a234 1
Early tar d248 6 a253 5
Pre-POSIX
Archives
An early draft of IEEE Std 1003.1-1988
(“POSIX.1”) served as the basis for John
Gilmore’s pdtar program and many system
d260 1
a260 1
The magic value consists of the d267 1 a267 1
The numeric fields are d273 1 a273 1
The prefix field is often not d277 7 a283 5
POSIX ustar
Archives
IEEE Std 1003.1-1988 (“POSIX.1”) defined a
standard tar file format to be read and written by compliant
implementations of tar(1). This format is often called the
d289 1
a289 1
struct d295 2 a296 2
};
d436 1 a436 1Type of entry. POSIX extended d442 1 a442 1
Regular file. d448 1 a448 1
Hard link.
d452 1 a452 1Symbolic d457 1 a457 1
Character d462 1 a462 1
Block device d467 1 a467 1
Directory.
d471 1 a471 1FIFO node.
d475 1 a475 1Reserved.
d479 1 a479 1A d489 1 a489 1
It is worth noting that the d500 1 a500 1
Contains the d508 1 a508 1
Version. This should be d514 1 a514 1
User and group names, as d522 1 a522 1
Major and minor numbers for d527 1 a527 1
If the pathname is too long to d538 1 a538 1
Note that all d541 1 a541 1
Field d553 1 a553 1
Currently, most d558 6 a563 5
Numeric
Extensions
There have been several attempts to extend the range of
sizes or times supported by modifying how numbers are stored
in the header.
One obvious d573 1 a573 1
Another d589 1 a589 1
Another early d594 10 a603 9
Pax
Interchange Format
There are many attributes that cannot be portably stored in
a POSIX ustar archive. IEEE Std 1003.1-2001
(“POSIX.1”) defined a “pax interchange
format” that uses two new types of entries to hold
text-formatted metadata that applies to following entries.
Note that a pax interchange format archive is a ustar
archive in every respect. The new data is stored in
d610 1
a610 1
An entry in a d625 2 a626 1
25 ctime=1084839148.1212\n
d628 1 a628 1Keys in all lowercase are d638 1 a638 1
File access, inode change, and d644 1 a644 1
The character set used by the d664 1 a664 1
User name, group name, and d672 1 a672 1
The full path of the linked-to d678 1 a678 1
The full d685 1 a685 1
These keys are reserved and may d690 1 a690 1
The size of the d697 1 a697 1
Vendor-specific attributes used d703 1 a703 1
Stores the access, default and d716 1 a716 1
The full minor and major d721 1 a721 1
The file flags.
d725 1 a725 1The full size of the file on d731 1 a731 1
The device number, inode d739 1 a739 1
Vendor-specific attributes used d746 1 a746 1
The time when the file was d754 1 a754 1
Libarchive stores d765 1 a765 1
XXX document other d768 1 a768 1
Any values d782 1 a782 1
In addition to d789 1 a789 1
Besides the new d800 16 a815 14
GNU Tar
Archives
The GNU tar program started with a pre-POSIX format similar
to that described earlier and has extended it using several
different mechanisms: It added new fields to the empty space
in the header (some of which was later used by POSIX for
conflicting purposes); it allowed the header to be continued
over multiple records; and it defined new entries that
modify following entries (similar in principle to the
x entry described above, but each GNU special entry
is single-purpose, unlike the general-purpose x
entry). As a result, GNU tar archives are not POSIX
compatible, although more lenient POSIX-compliant readers
can successfully extract most GNU tar archives.
struct d823 2 a824 2
};
d1073 1 a1073 1GNU tar uses the following d1079 1 a1079 1
GNU tar treats d1087 1 a1087 1
This indicates d1099 1 a1099 1
Note that the d1108 1 a1108 1
The data for d1114 1 a1114 1
The data for d1120 1 a1120 1
This is a d1141 1 a1141 1
Type d1154 1 a1154 1
This is a d1164 1 a1164 1
The name d1170 1 a1170 1
The magic field d1177 1 a1177 1
The version field holds a space d1183 1 a1183 1
The time the file was last d1189 1 a1189 1
This field is apparently no d1195 1 a1195 1
Each such structure specifies a d1205 1 a1205 1
If this is set to non-zero, the d1211 1 a1211 1
struct d1217 1 a1217 1
};
d1278 1 a1278 1A binary representation of the d1286 11 a1296 10
GNU tar pax
archives
GNU tar 1.14 (XXX check this XXX) and later will write pax
interchange format archives when you specify the
--posix flag. This format follows the pax interchange
format closely, using some SCHILY tags and
introducing new keywords to store sparse file information.
There have been three iterations of the sparse file support,
referred to as “0.0”, “0.1”, and
“1.0”.
The “0.0” format d1317 1 a1317 1
The “0.1” format d1329 1 a1329 1
The “1.0” format d1339 4 a1342 3
Solaris
Tar
XXX More Details Needed XXX
Solaris tar d1352 1 a1352 1
Extended attributes are stored d1360 1 a1360 1
An additional A header d1368 2 a1369 3
AIX Tar
XXX More details needed XXX
AIX Tar uses a d1381 14 a1394 12
Mac OS X
Tar
The tar distributed with Apple’s Mac OS X stores most
regular files as two separate files in the tar archive. The
two files have the same name except that the first one has
“._” prepended to the last path element. This
special file stores an AppleDouble-encoded binary blob with
additional metadata about the second file, including ACL,
extended attributes, and resources. To recreate the original
file on disk, each separate file can be extracted and the
Mac OS X copyfile() function can be used to unpack
the separate metadata file and apply it to th regular file.
d1400 1
a1400 1
Note that the d1407 8 a1414 6
Summary of
tar type codes
The following list is a condensed summary of the type codes
used in tar header records generated by different tar
implementations. More details about specific implementations
can be found above:
Early tar d1423 1 a1423 1
POSIX standard d1428 1 a1428 1
POSIX standard d1433 1 a1433 1
POSIX standard d1438 1 a1438 1
POSIX standard d1443 1 a1443 1
POSIX standard d1448 1 a1448 1
POSIX standard d1453 1 a1453 1
POSIX standard d1458 1 a1458 1
POSIX d1463 1 a1463 1
GNU tar used d1468 1 a1468 1
Solaris tar ACL d1473 1 a1473 1
AIX tar ACL d1478 1 a1478 1
GNU tar d1483 1 a1483 1
GNU tar long d1488 1 a1488 1
GNU tar long d1493 1 a1493 1
GNU tar d1499 1 a1499 1
GNU tar long d1504 1 a1504 1
GNU tar sparse d1509 1 a1509 1
GNU tar d1514 1 a1514 1
Solaris tar d1519 1 a1519 1
POSIX pax d1524 1 a1524 1
POSIX pax d1529 2 a1530 1
ar(1), pax(1), tar(1)
d1534 1 a1534 1The tar utility is no d1538 3 a1540 3 standards by pax(1). The ustar format is currently part of the specification for the pax(1) utility. The pax interchange file format is new with IEEE Std 1003.1-2001 d1545 1 a1545 1
A tar command appeared in d1557 1 a1557 1
This d1560 2 a1561 4 <kientzle@@FreeBSD.org>.
BSD December 27, 2016 BSD
@ 1.11 log @libarchive: updated to 3.7.4 Libarchive 3.7.4 is a bugfix and security release Security fixes: rar: Fix OOB in rar e8 filter (CVE-2024-26256) zip: Fix out of boundary access Important bugfixes: 7zip: Limit amount of properties bsdtar: Fix error handling around strtol() usages passphrase: Improve newline handling on Windows passphrase: Never allow empty passwords rar: Fix "File CRC Error" when extracting specific rar4 archives xar: Avoid infinite link loop zip: Update AppleDouble support for directories zstd: Implement core detection @ text @d2 1 a2 1 @ 1.10 log @libarchive: updated to 3.7.3 Libarchive 3.7.3 is a feature, security and bugfix release. New features: PCRE2 support add trailing letter b to bsdtar(1) substitute pattern add support for long options "--group" and "--owner" to tar(1) Security fixes: Fix possible vulnerability in tar error reporting introduced in f27c173 Important bugfixes: ISO9660: preserve the natural order of links rar5: fix decoding unicode filenames on Windows rar5: fix infinite loop if during rar5 decompression the last block produced no data xz filter: fix incorrect eof at the end of an lzip member zip: fix end-of-data marker processing when decompressing zip archives multiple bsdunzip(1) fixes filetime truncation fix on Windows @ text @d2 1 a2 1 @ 1.9 log @libarchive: updated to 3.7.2 Libarchive 3.7.2 is a security, bugfix and feature release. Security fixes: Multiple vulnerabilities have been fixed in the PAX writer (1b4e0d0) Important bugfixes: bsdunzip(1) now correctly handles arguments following an -x after the zipfile New features: bsdunzip(1) now supports the "--version" flag 7-zip reader now translates Windows permissions into UNIX permissions uudecode filter in raw mode now supports file name and file mode zstd filter now supports the "long" write option Libarchive 3.7.1 is a security, feature and bugfix release. Security fixes: SEGV and stack buffer overflow in verbose mode of cpio Feature updates: bsdunzip updated to match latest upstream code Important bugfixes: miscellaneous functional bugfixes build fixes on multiple platforms Libarchive 3.7.0 is a feature and bugfix release. New features: bsdunzip: new tool ported from FreeBSD drop-in replacement for Info-ZIP unzip, not yet ported for Windows 7zip reader: support for Zstandard compression 7zip reader: support for ARM64 filter zstd filter: support for multi-frame zstd archives Other notable bugfixes and improvements: pax: fix year 2038 problem on platforms with 64-bit time_t Windows: Universal Windows Platform (UWP) fixes and improvements Windows: bcrypt usage fixes and improvements Windows: time function usage fixes and improvements @ text @d2 1 a2 1 @ 1.8 log @libarchive: Update to 3.4.3 Libarchive 3.4.3 is a feature and bugfix release. New features: support for pzstd compressed files (#1357) support for RHT.security.selinux tar extended attribute (#1348) Important bugfixes: various zstd fixes and improvements (#1342 #1352 #1359) child process handling fixes (#1372) Libarchive 3.4.2 is a feature and security release. New features: support for atomic file extraction (bsdtar -x --safe-writes) (#1289) support for mbed TLS (PolarSSL) (#1301) Important bugfixes: security fixes in RAR5 reader (#1280 #1326) compression buffer fix in XAR writer (#1317) fix uname and gname longer than 32 characters in PAX writer (#1319) fix segfault when archiving hard links in ISO9660 and XAR writers (#1325) fix support for extracting 7z archive entries with Delta filter (#987) Libarchive 3.4.1 is a feature and security release. New features: Unicode filename support for reading lha/lzh archives New pax write option "xattrhdr" Important bugfixes: security fixes in wide string processing (#1276 #1298) security fixes in RAR5 reader (#1212 #1217 #1296) security fixes and optimizations to write filter logic (#351) security fix related to use of readlink(2) (1dae5a5) sparse file handling fixes (#1218 #1260) Thanks to all contributors and bug reporters. Special thanks to Christos Zoulas (@@zoulasc) from NetBSD for the atomic file extraction feature. @ text @d1 2 a2 2 d53 2 a54 2 ’’blocks’’ are always a multiple of the record size. The maximum block size supported by early d58 4 a61 5 high-speed tape drives. (Note: the terms ’’block’’ and ’’record’’ here are not entirely standard; this document follows the convention established by John Gilmore in documenting pdtar.) d241 3 a243 3 (’’POSIX.1’’) standard was released. For best portability, modern implementations should fill the numeric fields with leading zeros. d248 5 a252 5 (’’POSIX.1’’) served as the basis for John Gilmore’s pdtar program and many system implementations from the late 1980s and early 1990s. These archives generally follow the POSIX ustar format described below with the following variations: d257 3 a259 3 five characters ’’ustar’’ followed by a space. The version field contains a space character followed by a null. d275 7 a281 7 IEEE Std 1003.1-1988 (’’POSIX.1’’) defined a standard tar file format to be read and written by compliant implementations of tar(1). This format is often called the ’’ustar’’ format, after the magic value used in the header. (The name is an acronym for ’’Unix Standard TAR’’.) It extends the historic format with new fields: d434 1 a434 1’’0’’
d440 1 a440 1’’1’’
d444 1 a444 1’’2’’
d449 1 a449 1’’3’’
d454 1 a454 1’’4’’
d459 1 a459 1’’5’’
d463 1 a463 1’’6’’
d467 1 a467 1’’7’’
d495 4 a498 4 magic value ’’ustar’’ followed by a NUL byte to indicate that this is a POSIX standard archive. Full compliance requires the uname and gname fields be properly set. d503 2 a504 2 ’’00’’ (two copies of the ASCII digit zero) for POSIX standard archives. d591 10 a600 11 (’’POSIX.1’’) defined a ’’pax interchange format’’ that uses two new types of entries to hold text-formatted metadata that applies to following entries. Note that a pax interchange format archive is a ustar archive in every respect. The new data is stored in ustar-compatible archive entries that use the ’’x’’ or ’’g’’ typeflag. In particular, older implementations that do not fully support these extensions will extract the metadata into regular files, where the metadata can be examined as necessary. d641 10 a650 11 six-character ASCII string ’’BINARY’’, then all textual values are assumed to be in a platform-dependent multi-byte encoding. Note that there are only two valid values for this key: ’’BINARY’’ or ’’ISO-IR 10646 2000 UTF-8’’. No other values are permitted by the standard, and the latter value should generally not be used as it is the default when this key is not specified. In particular, this flag should not be used as a general mechanism to allow filenames to be stored in arbitrary encodings. d739 2 a740 2 ’’ctime’’ attribute, which refers to the time when the file metadata was last changed.) d748 5 a752 7 characters and the two special characters ’’=’’ and ’’%’’ are encoded as ’’%’’ followed by two uppercase hexadecimal digits. The value of this key is the extended attribute value encoded in base 64. XXX Detail the base-64 format here XXX d1135 5 a1139 6 the operations to be done, in the form ’’Rename %s to %s\n’’ or ’’Symlink %s to %s\n’’; in either case, both filenames are escaped using K&R C syntax. Due to security concerns, "N" records are now generally ignored when reading archives. d1144 6 a1149 7 ’’sparse’’ regular file. Sparse files are stored as a series of fragments. The header contains a list of fragment offset/length pairs. If more than four such entries are required, the header is extended as necessary with ’’extra’’ header extensions (an older format that is no longer used), or ’’sparse’’ extensions. d1160 3 a1162 3 holds the five characters ’’ustar’’ followed by a space. Note that POSIX ustar archives have a trailing null. d1168 1 a1168 2 use two copies of the ASCII digit ’’0’’. d1195 4 a1198 4 header will be followed by additional ’’sparse header’’ records. Each such record contains information about as many as 21 additional sparse blocks as shown here: d1283 2 a1284 3 referred to as ’’0.0’’, ’’0.1’’, and ’’1.0’’. d1290 12 a1301 12The ’’0.0’’ format used an initial GNU.sparse.numblocks attribute to indicate the number of blocks in the file, a pair of GNU.sparse.offset and GNU.sparse.numbytes to indicate the offset and size of each block, and a single GNU.sparse.size to indicate the full size of the file. This is not the same as the size in the tar header because the latter value does not include the size of any holes. This format required that the order of attributes be preserved and relied on readers accepting multiple appearances of the same attribute names, which is not officially permitted by the standards.
d1305 7 a1311 8The ’’0.1’’ format used a single attribute that stored a comma-separated list of decimal numbers. Each pair of numbers indicated the offset and size, respectively, of a block of data. This does not work well if the archive is extracted by an archiver that does not recognize this extension, since many pax implementations simply discard unrecognized attributes.
d1317 9 a1325 10The ’’1.0’’ format stores the sparse block map in one or more 512-byte blocks prepended to the file data in the entry body. The pax attributes indicate the existence of this map (via the GNU.sparse.major and GNU.sparse.minor fields) and the full size of the file. The GNU.sparse.name holds the true name of the file. To avoid confusion, the name stored in the regular tar header is a modified name so that extraction errors will be apparent to users.
d1333 3 a1335 3 ’’extended’’ format that is fundamentally similar to pax interchange format, with the following differences: d1372 11 a1382 11 ’’._’’ prepended to the last path element. This special file stores an AppleDouble-encoded binary blob with additional metadata about the second file, including ACL, extended attributes, and resources. To recreate the original file on disk, each separate file can be extracted and the Mac OS X copyfile() function can be used to unpack the separate metadata file and apply it to th regular file. Conversely, the same function provides a ’’pack’’ option to encode the extended metadata from a file into a separate file whose contents can then be put into a tar archive. d1518 5 a1522 5 (’’SUSv2’’). It has been supplanted in subsequent standards by pax(1). The ustar format is currently part of the specification for the pax(1) utility. The pax interchange file format is new with IEEE Std 1003.1-2001 (’’POSIX.1’’). @ 1.7 log @Update for libarchive-3.4.0: - improvements for Android APK and JAR archives - better support for non-recursive list and extract - tar --exclude-vcs support - fixes for file attributes and flags handling - zipx support - rar 5.0 reader @ text @d1 2 a2 2 d53 2 a54 2 “blocks” are always a multiple of the record size. The maximum block size supported by early d58 5 a62 4 high-speed tape drives. (Note: the terms “block” and “record” here are not entirely standard; this document follows the convention established by John Gilmore in documenting pdtar.) d242 3 a244 3 (“POSIX.1”) standard was released. For best portability, modern implementations should fill the numeric fields with leading zeros. d249 5 a253 5 (“POSIX.1”) served as the basis for John Gilmore’s pdtar program and many system implementations from the late 1980s and early 1990s. These archives generally follow the POSIX ustar format described below with the following variations: d258 3 a260 3 five characters “ustar” followed by a space. The version field contains a space character followed by a null. d276 7 a282 7 IEEE Std 1003.1-1988 (“POSIX.1”) defined a standard tar file format to be read and written by compliant implementations of tar(1). This format is often called the “ustar” format, after the magic value used in the header. (The name is an acronym for “Unix Standard TAR”.) It extends the historic format with new fields: d435 1 a435 1“0”
d441 1 a441 1“1”
d445 1 a445 1“2”
d450 1 a450 1“3”
d455 1 a455 1“4”
d460 1 a460 1“5”
d464 1 a464 1“6”
d468 1 a468 1“7”
d496 4 a499 4 magic value “ustar” followed by a NUL byte to indicate that this is a POSIX standard archive. Full compliance requires the uname and gname fields be properly set. d504 2 a505 2 “00” (two copies of the ASCII digit zero) for POSIX standard archives. d592 11 a602 10 (“POSIX.1”) defined a “pax interchange format” that uses two new types of entries to hold text-formatted metadata that applies to following entries. Note that a pax interchange format archive is a ustar archive in every respect. The new data is stored in ustar-compatible archive entries that use the “x” or “g” typeflag. In particular, older implementations that do not fully support these extensions will extract the metadata into regular files, where the metadata can be examined as necessary. d643 11 a653 10 six-character ASCII string “BINARY”, then all textual values are assumed to be in a platform-dependent multi-byte encoding. Note that there are only two valid values for this key: “BINARY” or “ISO-IR 10646 2000 UTF-8”. No other values are permitted by the standard, and the latter value should generally not be used as it is the default when this key is not specified. In particular, this flag should not be used as a general mechanism to allow filenames to be stored in arbitrary encodings. d695 1 a695 1 SCHILY.acl.default, SCHILY.acl.ace d722 2 a723 2SCHILY.dev, SCHILY.ino, SCHILY.nlinks
d742 2 a743 2 “ctime” attribute, which refers to the time when the file metadata was last changed.) d746 1 a746 1LIBARCHIVE.xattr.namespace.key
d751 7 a757 5 characters and the two special characters “=” and “%” are encoded as “%” followed by two uppercase hexadecimal digits. The value of this key is the extended attribute value encoded in base 64. XXX Detail the base-64 format here XXX d1140 6 a1145 5 the operations to be done, in the form “Rename %s to %s\n” or “Symlink %s to %s\n”; in either case, both filenames are escaped using K&R C syntax. Due to security concerns, "N" records are now generally ignored when reading archives. d1150 7 a1156 6 “sparse” regular file. Sparse files are stored as a series of fragments. The header contains a list of fragment offset/length pairs. If more than four such entries are required, the header is extended as necessary with “extra” header extensions (an older format that is no longer used), or “sparse” extensions. d1167 3 a1169 3 holds the five characters “ustar” followed by a space. Note that POSIX ustar archives have a trailing null. d1175 2 a1176 1 use two copies of the ASCII digit “0”. d1203 4 a1206 4 header will be followed by additional “sparse header” records. Each such record contains information about as many as 21 additional sparse blocks as shown here: d1291 3 a1293 2 referred to as “0.0”, “0.1”, and “1.0”. d1299 12 a1310 12The “0.0” format used an initial GNU.sparse.numblocks attribute to indicate the number of blocks in the file, a pair of GNU.sparse.offset and GNU.sparse.numbytes to indicate the offset and size of each block, and a single GNU.sparse.size to indicate the full size of the file. This is not the same as the size in the tar header because the latter value does not include the size of any holes. This format required that the order of attributes be preserved and relied on readers accepting multiple appearances of the same attribute names, which is not officially permitted by the standards.
d1314 8 a1321 7The “0.1” format used a single attribute that stored a comma-separated list of decimal numbers. Each pair of numbers indicated the offset and size, respectively, of a block of data. This does not work well if the archive is extracted by an archiver that does not recognize this extension, since many pax implementations simply discard unrecognized attributes.
d1327 10 a1336 9The “1.0” format stores the sparse block map in one or more 512-byte blocks prepended to the file data in the entry body. The pax attributes indicate the existence of this map (via the GNU.sparse.major and GNU.sparse.minor fields) and the full size of the file. The GNU.sparse.name holds the true name of the file. To avoid confusion, the name stored in the regular tar header is a modified name so that extraction errors will be apparent to users.
d1344 3 a1346 3 “extended” format that is fundamentally similar to pax interchange format, with the following differences: d1383 11 a1393 11 “._” prepended to the last path element. This special file stores an AppleDouble-encoded binary blob with additional metadata about the second file, including ACL, extended attributes, and resources. To recreate the original file on disk, each separate file can be extracted and the Mac OS X copyfile() function can be used to unpack the separate metadata file and apply it to th regular file. Conversely, the same function provides a “pack” option to encode the extended metadata from a file into a separate file whose contents can then be put into a tar archive. d1529 5 a1533 5 (“SUSv2”). It has been supplanted in subsequent standards by pax(1). The ustar format is currently part of the specification for the pax(1) utility. The pax interchange file format is new with IEEE Std 1003.1-2001 (“POSIX.1”). @ 1.6 log @libarchive: updated to 3.3.3 libarchive 3.3.3: Avoid super-linear slowdown on malformed mtree files Many fixes for building with Visual Studio NO_OVERWRITE doesn't change existing directory attributes New support for Zstandard read and write filters @ text @d1 2 a2 2 d53 2 a54 2 ’’blocks’’ are always a multiple of the record size. The maximum block size supported by early d58 4 a61 5 high-speed tape drives. (Note: the terms ’’block’’ and ’’record’’ here are not entirely standard; this document follows the convention established by John Gilmore in documenting pdtar.) d241 3 a243 3 (’’POSIX.1’’) standard was released. For best portability, modern implementations should fill the numeric fields with leading zeros. d248 5 a252 5 (’’POSIX.1’’) served as the basis for John Gilmore’s pdtar program and many system implementations from the late 1980s and early 1990s. These archives generally follow the POSIX ustar format described below with the following variations: d257 3 a259 3 five characters ’’ustar’’ followed by a space. The version field contains a space character followed by a null. d275 7 a281 7 IEEE Std 1003.1-1988 (’’POSIX.1’’) defined a standard tar file format to be read and written by compliant implementations of tar(1). This format is often called the ’’ustar’’ format, after the magic value used in the header. (The name is an acronym for ’’Unix Standard TAR’’.) It extends the historic format with new fields: d434 1 a434 1’’0’’
d440 1 a440 1’’1’’
d444 1 a444 1’’2’’
d449 1 a449 1’’3’’
d454 1 a454 1’’4’’
d459 1 a459 1’’5’’
d463 1 a463 1’’6’’
d467 1 a467 1’’7’’
d495 4 a498 4 magic value ’’ustar’’ followed by a NUL byte to indicate that this is a POSIX standard archive. Full compliance requires the uname and gname fields be properly set. d503 2 a504 2 ’’00’’ (two copies of the ASCII digit zero) for POSIX standard archives. d591 10 a600 11 (’’POSIX.1’’) defined a ’’pax interchange format’’ that uses two new types of entries to hold text-formatted metadata that applies to following entries. Note that a pax interchange format archive is a ustar archive in every respect. The new data is stored in ustar-compatible archive entries that use the ’’x’’ or ’’g’’ typeflag. In particular, older implementations that do not fully support these extensions will extract the metadata into regular files, where the metadata can be examined as necessary. d641 10 a650 11 six-character ASCII string ’’BINARY’’, then all textual values are assumed to be in a platform-dependent multi-byte encoding. Note that there are only two valid values for this key: ’’BINARY’’ or ’’ISO-IR 10646 2000 UTF-8’’. No other values are permitted by the standard, and the latter value should generally not be used as it is the default when this key is not specified. In particular, this flag should not be used as a general mechanism to allow filenames to be stored in arbitrary encodings. d739 2 a740 2 ’’ctime’’ attribute, which refers to the time when the file metadata was last changed.) d748 5 a752 7 characters and the two special characters ’’=’’ and ’’%’’ are encoded as ’’%’’ followed by two uppercase hexadecimal digits. The value of this key is the extended attribute value encoded in base 64. XXX Detail the base-64 format here XXX d1135 5 a1139 6 the operations to be done, in the form ’’Rename %s to %s\n’’ or ’’Symlink %s to %s\n’’; in either case, both filenames are escaped using K&R C syntax. Due to security concerns, "N" records are now generally ignored when reading archives. d1144 6 a1149 7 ’’sparse’’ regular file. Sparse files are stored as a series of fragments. The header contains a list of fragment offset/length pairs. If more than four such entries are required, the header is extended as necessary with ’’extra’’ header extensions (an older format that is no longer used), or ’’sparse’’ extensions. d1160 3 a1162 3 holds the five characters ’’ustar’’ followed by a space. Note that POSIX ustar archives have a trailing null. d1168 1 a1168 2 use two copies of the ASCII digit ’’0’’. d1195 4 a1198 4 header will be followed by additional ’’sparse header’’ records. Each such record contains information about as many as 21 additional sparse blocks as shown here: d1283 2 a1284 3 referred to as ’’0.0’’, ’’0.1’’, and ’’1.0’’. d1290 12 a1301 12The ’’0.0’’ format used an initial GNU.sparse.numblocks attribute to indicate the number of blocks in the file, a pair of GNU.sparse.offset and GNU.sparse.numbytes to indicate the offset and size of each block, and a single GNU.sparse.size to indicate the full size of the file. This is not the same as the size in the tar header because the latter value does not include the size of any holes. This format required that the order of attributes be preserved and relied on readers accepting multiple appearances of the same attribute names, which is not officially permitted by the standards.
d1305 7 a1311 8The ’’0.1’’ format used a single attribute that stored a comma-separated list of decimal numbers. Each pair of numbers indicated the offset and size, respectively, of a block of data. This does not work well if the archive is extracted by an archiver that does not recognize this extension, since many pax implementations simply discard unrecognized attributes.
d1317 9 a1325 10The ’’1.0’’ format stores the sparse block map in one or more 512-byte blocks prepended to the file data in the entry body. The pax attributes indicate the existence of this map (via the GNU.sparse.major and GNU.sparse.minor fields) and the full size of the file. The GNU.sparse.name holds the true name of the file. To avoid confusion, the name stored in the regular tar header is a modified name so that extraction errors will be apparent to users.
d1333 3 a1335 3 ’’extended’’ format that is fundamentally similar to pax interchange format, with the following differences: d1372 11 a1382 11 ’’._’’ prepended to the last path element. This special file stores an AppleDouble-encoded binary blob with additional metadata about the second file, including ACL, extended attributes, and resources. To recreate the original file on disk, each separate file can be extracted and the Mac OS X copyfile() function can be used to unpack the separate metadata file and apply it to th regular file. Conversely, the same function provides a ’’pack’’ option to encode the extended metadata from a file into a separate file whose contents can then be put into a tar archive. d1518 5 a1522 5 (’’SUSv2’’). It has been supplanted in subsequent standards by pax(1). The ustar format is currently part of the specification for the pax(1) utility. The pax interchange file format is new with IEEE Std 1003.1-2001 (’’POSIX.1’’). @ 1.5 log @Merge for libarchive-3.3.2. @ text @d2 1 a2 1 @ 1.4 log @Merge libarchive-3.3.1. @ text @d2 1 a2 1 d53 1 a53 1 ‘‘blocks’’ are always a multiple of d59 2 a60 2 ‘‘block’’ and ‘‘record’’ here are not entirely d227 1 a227 1 an ASCII ‘1’ and the linkname field holds d242 1 a242 1 (‘‘POSIX.1’’) standard was released. d249 1 a249 1 (‘‘POSIX.1’’) served as the basis d258 1 a258 1 five characters ‘‘ustar’’ followed d276 1 a276 1 IEEE Std 1003.1-1988 (‘‘POSIX.1’’) d279 1 a279 1 called the ‘‘ustar’’ format, after d281 1 a281 1 for ‘‘Unix Standard TAR’’.) It d435 1 a435 1‘‘0’’
d441 1 a441 1‘‘1’’
d445 1 a445 1‘‘2’’
d450 1 a450 1‘‘3’’
d455 1 a455 1‘‘4’’
d460 1 a460 1‘‘5’’
d464 1 a464 1‘‘6’’
d468 1 a468 1‘‘7’’
d496 1 a496 1 magic value ‘‘ustar’’ followed by a d504 1 a504 1 ‘‘00’’ (two copies of the ASCII d592 2 a593 2 (‘‘POSIX.1’’) defined a ‘‘pax interchange format’’ that uses d598 2 a599 2 entries that use the ‘‘x’’ or ‘‘g’’ typeflag. In particular, older d644 1 a644 1 ‘‘BINARY’’, then all textual values d647 2 a648 2 key: ‘‘BINARY’’ or ‘‘ISO-IR 10646 2000 UTF-8’’. d742 1 a742 1 ‘‘ctime’’ attribute, which refers to d752 3 a754 3 ‘‘=’’ and ‘‘%’’ are encoded as ‘‘%’’ followed by two uppercase d1140 2 a1141 2 the operations to be done, in the form ‘‘Rename %s to %s\n’’ or ‘‘Symlink %s to d1150 1 a1150 1 ‘‘sparse’’ regular file. Sparse d1154 1 a1154 1 as necessary with ‘‘extra’’ header d1156 1 a1156 1 ‘‘sparse’’ extensions. d1167 1 a1167 1 holds the five characters ‘‘ustar’’ d1176 1 a1176 1 ‘‘0’’. d1203 1 a1203 1 header will be followed by additional ‘‘sparse d1287 7 a1293 8 −-posix flag. This format follows the pax interchange format closely, using some SCHILY tags and introducing new keywords to store sparse file information. There have been three iterations of the sparse file support, referred to as ‘‘0.0’’, ‘‘0.1’’, and ‘‘1.0’’. d1300 1 a1300 1 ‘‘0.0’’ format used an initial d1315 1 a1315 1 ‘‘0.1’’ format used a single d1328 1 a1328 1 ‘‘1.0’’ format stores the sparse d1344 1 a1344 1 ‘‘extended’’ format that is d1383 1 a1383 1 ‘‘._’’ prepended to the last path d1391 1 a1391 1 ‘‘pack’’ option to encode the d1529 1 a1529 1 (‘‘SUSv2’’). It has been supplanted d1533 1 a1533 1 1003.1-2001 (‘‘POSIX.1’’). @ 1.3 log @Update for libarchive 3.2.1. @ text @d2 1 a2 1 d53 1 a53 1 ’’blocks’’ are always a multiple of d59 2 a60 2 ’’block’’ and ’’record’’ here are not entirely d227 1 a227 1 an ASCII ’1’ and the linkname field holds d242 1 a242 1 (’’POSIX.1’’) standard was released. d249 1 a249 1 (’’POSIX.1’’) served as the basis d258 1 a258 1 five characters ’’ustar’’ followed d276 1 a276 1 IEEE Std 1003.1-1988 (’’POSIX.1’’) d279 1 a279 1 called the ’’ustar’’ format, after d281 1 a281 1 for ’’Unix Standard TAR’’.) It d435 1 a435 1’’0’’
d441 1 a441 1’’1’’
d445 1 a445 1’’2’’
d450 1 a450 1’’3’’
d455 1 a455 1’’4’’
d460 1 a460 1’’5’’
d464 1 a464 1’’6’’
d468 1 a468 1’’7’’
d496 1 a496 1 magic value ’’ustar’’ followed by a d504 1 a504 1 ’’00’’ (two copies of the ASCII d592 2 a593 2 (’’POSIX.1’’) defined a ’’pax interchange format’’ that uses d598 2 a599 2 entries that use the ’’x’’ or ’’g’’ typeflag. In particular, older d644 1 a644 1 ’’BINARY’’, then all textual values d647 2 a648 2 key: ’’BINARY’’ or ’’ISO-IR 10646 2000 UTF-8’’. d695 1 a695 1 SCHILY.acl.default d697 9 a705 8Stores the access and default ACLs as textual strings in a format that is an extension of the format specified by POSIX.1e draft 17. In particular, each user or group access specification can include a fourth colon-separated field with the numeric UID or GID. This allows ACLs to be restored on systems that may not have complete user or group information available (such as when NIS/YP or LDAP services are temporarily unavailable).
d742 1 a742 1 ’’ctime’’ attribute, which refers to d752 3 a754 3 ’’=’’ and ’’%’’ are encoded as ’’%’’ followed by two uppercase d1140 2 a1141 2 the operations to be done, in the form ’’Rename %s to %s\n’’ or ’’Symlink %s to d1150 1 a1150 1 ’’sparse’’ regular file. Sparse d1154 1 a1154 1 as necessary with ’’extra’’ header d1156 1 a1156 1 ’’sparse’’ extensions. d1167 1 a1167 1 holds the five characters ’’ustar’’ d1176 1 a1176 1 ’’0’’. d1203 1 a1203 1 header will be followed by additional ’’sparse d1292 3 a1294 3 ’’0.0’’, ’’0.1’’, and ’’1.0’’. d1301 1 a1301 1 ’’0.0’’ format used an initial d1316 1 a1316 1 ’’0.1’’ format used a single d1329 1 a1329 1 ’’1.0’’ format stores the sparse d1345 1 a1345 1 ’’extended’’ format that is d1384 1 a1384 1 ’’._’’ prepended to the last path d1392 1 a1392 1 ’’pack’’ option to encode the d1530 1 a1530 1 (’’SUSv2’’). It has been supplanted d1534 1 a1534 1 1003.1-2001 (’’POSIX.1’’). d1556 1 a1556 1 December 23, 2011 BSD @ 1.2 log @Changes 3.1.2: This is a maintenance update to fix issues with the new RAR seeking feature. This new release also contains fixes for build failures when building libarchive using Visual Studio 2012 and MinGW. @ text @d1 2 a2 2 d1545 1 a1545 1 another open-source (GPL) archiver (originally developed d1551 2 a1552 2 and bsdtar project by Tim Kientzle ⟨ kientzle@@FreeBSD.org⟩ . @ 1.1 log @Initial revision @ text @d1 2 a2 2 d11 4 a14 3 p { margin-top: 0; margin-bottom: 0; } pre { margin-top: 0; margin-bottom: 0; } table { margin-top: 0; margin-bottom: 0; } d23 1 a23 2tar(5) FreeBSD File Formats Manual tar(5)
d25 1 a25 1NAME
d27 1 a27 1tar — format of d30 1 d32 1 a32 3
DESCRIPTION
The tar archive format d40 1 a40 1
General d49 1 a49 1
For d53 1 a53 1 ‘‘blocks’’ are always a multiple of d59 2 a60 2 ‘‘block’’ and ‘‘record’’ here are not entirely d64 1 a64 1
Old-Style d73 1 a73 1
The header d77 1 a77 1
struct d80 1 a80 1
|
d87 3
a89 3
char name[100]; |
d91 2 a92 2 |
d95 3
a97 3
char mode[8]; |
d99 2 a100 2 |
d103 3
a105 3
char uid[8]; |
d107 2 a108 2 |
d111 3
a113 3
char gid[8]; |
d115 2 a116 2 |
d119 3
a121 3
char size[12]; |
d123 2 a124 2 |
d127 3
a129 3
char mtime[12]; |
d131 2 a132 2 |
d135 3
a137 3
char checksum[8]; |
d139 2 a140 2 |
d143 3
a145 3
char linkflag[1]; |
d147 2 a148 2 |
d151 3
a153 3
char linkname[100]; |
d155 2 a156 2 |
d159 3
a161 3
char pad[255]; |
d164 1 a164 1 |
|
d294 3
a296 3
char name[100]; |
d298 2 a299 2 |
d302 3
a304 3
char mode[8]; |
d306 2 a307 2 |
d310 3
a312 3
char uid[8]; |
d314 2 a315 2 |
d318 3
a320 3
char gid[8]; |
d322 2 a323 2 |
d326 3
a328 3
char size[12]; |
d330 2 a331 2 |
d334 3
a336 3
char mtime[12]; |
d338 2 a339 2 |
d342 3
a344 3
char checksum[8]; |
d346 2 a347 2 |
d350 3
a352 3
char typeflag[1]; |
d354 2 a355 2 |
d358 3
a360 3
char linkname[100]; |
d362 2 a363 2 |
d366 3
a368 3
char magic[6]; |
d370 2 a371 2 |
d374 3
a376 3
char version[2]; |
d378 2 a379 2 |
d382 3
a384 3
char uname[32]; |
d386 2 a387 2 |
d390 3
a392 3
char gname[32]; |
d394 2 a395 2 |
d398 3
a400 3
char devmajor[8]; |
d402 2 a403 2 |
d406 3
a408 3
char devminor[8]; |
d410 2 a411 2 |
d414 3
a416 3
char prefix[155]; |
d418 2 a419 2 |
d422 3
a424 3
char pad[12]; |
d427 1 a427 1 |
|
d820 4
a823 4
char name[100]; |
d825 2 a826 2 |
d829 4
a832 4
char mode[8]; |
d834 2 a835 2 |
d838 4
a841 4
char uid[8]; |
d843 2 a844 2 |
d847 4
a850 4
char gid[8]; |
d852 2 a853 2 |
d856 4
a859 4
char size[12]; |
d861 2 a862 2 |
d865 4
a868 4
char mtime[12]; |
d870 2 a871 2 |
d874 4
a877 4
char checksum[8]; |
d879 2 a880 2 |
d883 4
a886 4
char typeflag[1]; |
d888 2 a889 2 |
d892 4
a895 4
char linkname[100]; |
d897 2 a898 2 |
d901 4
a904 4
char magic[6]; |
d906 2 a907 2 |
d910 4
a913 4
char version[2]; |
d915 2 a916 2 |
d919 4
a922 4
char uname[32]; |
d924 2 a925 2 |
d928 4
a931 4
char gname[32]; |
d933 2 a934 2 |
d937 4
a940 4
char devmajor[8]; |
d942 2 a943 2 |
d946 4
a949 4
char devminor[8]; |
d951 2 a952 2 |
d955 4
a958 4
char atime[12]; |
d960 2 a961 2 |
d964 4
a967 4
char ctime[12]; |
d969 2 a970 2 |
d973 4
a976 4
char offset[12]; |
d978 2 a979 2 |
d982 4
a985 4
char longnames[4]; |
d987 2 a988 2 |
d991 4
a994 4
char unused[1]; |
d996 2 a997 2 |
d1000 4
a1003 4
struct { |
d1005 2 a1006 2 | d1008 1 a1008 1 |
d1011 3
a1013 3
char offset[12]; |
d1015 2 a1016 2 | d1018 1 a1018 1 |
d1021 3
a1023 3
char numbytes[12]; |
d1025 2 a1026 2 |
d1029 4
a1032 4
} sparse[4]; |
d1034 2 a1035 2 |
d1038 4
a1041 4
char isextended[1]; |
d1043 2 a1044 2 |
d1047 4
a1050 4
char realsize[12]; |
d1052 2 a1053 2 |
d1056 4
a1059 4
char pad[17]; |
d1062 1 a1062 1 |
|
d1217 4
a1220 4
struct { |
d1222 2 a1223 2 | d1225 1 a1225 1 |
d1228 3
a1230 3
char offset[12]; |
d1232 2 a1233 2 | d1235 1 a1235 1 |
d1238 3
a1240 3
char numbytes[12]; |
d1242 2 a1243 2 |
d1246 4
a1249 4
} sparse[21]; |
d1251 2 a1252 2 |
d1255 4
a1258 4
char isextended[1]; |
d1260 2 a1261 2 |
d1264 4
a1267 4
char padding[7]; |
d1270 1 a1270 1 |
|
char name[100]; |
||
|
char mode[8]; |
||
|
char uid[8]; |
||
|
char gid[8]; |
||
|
char size[12]; |
||
|
char mtime[12]; |
||
|
char checksum[8]; |
||
|
char linkflag[1]; |
||
|
char linkname[100]; |
||
|
char pad[255]; |
};
All unused bytes in the header record are filled with nulls.
name
Pathname, stored as a null-terminated string. Early tar implementations only stored regular files (including hardlinks to those files). One common early convention used a trailing "/" character to indicate a directory name, allowing directory permissions and owner information to be archived and restored.
mode
File mode, stored as an octal number in ASCII.
uid, gid
User id and group id of owner, as octal numbers in ASCII.
size
Size of file, as octal number in ASCII. For regular files only, this indicates the amount of data that follows the header. In particular, this field was ignored by early tar implementations when extracting hardlinks. Modern writers should always store a zero length for hardlink entries.
mtime
Modification time of file, as an octal number in ASCII. This indicates the number of seconds since the start of the epoch, 00:00:00 UTC January 1, 1970. Note that negative values should be avoided here, as they are handled inconsistently.
checksum
Header checksum, stored as an octal number in ASCII. To compute the checksum, set the checksum field to all spaces, then sum all bytes in the header using unsigned arithmetic. This field should be stored as six octal digits followed by a null and a space character. Note that many early implementations of tar used signed arithmetic for the checksum field, which can cause interoperability problems when transferring archives between systems. Modern robust readers compute the checksum both ways and accept the header if either computation matches.
linkflag, linkname
In order to preserve hardlinks and conserve tape, a file with multiple links is only written to the archive the first time it is encountered. The next time it is encountered, the linkflag is set to an ASCII ‘1’ and the linkname field holds the first name under which this file appears. (Note that regular files have a null value in the linkflag field.)
Early tar implementations varied in how they terminated these fields. The tar command in Version 7 AT&T UNIX used the following conventions (this is also documented in early BSD manpages): the pathname must be null-terminated; the mode, uid, and gid fields must end in a space and a null byte; the size and mtime fields must end in a space; the checksum is terminated by a null and a space. Early implementations filled the numeric fields with leading spaces. This seems to have been common practice until the IEEE Std 1003.1-1988 (‘‘POSIX.1’’) standard was released. For best portability, modern implementations should fill the numeric fields with leading zeros.
Pre-POSIX
Archives
An early draft of IEEE Std 1003.1-1988
(‘‘POSIX.1’’) served as the basis
for John Gilmore’s pdtar program and many
system implementations from the late 1980s and early 1990s.
These archives generally follow the POSIX ustar format
described below with the following variations:
•
The magic value is ‘‘ustar ’’ (note the following space). The version field contains a space character followed by a null.
•
The numeric fields are generally filled with leading spaces (not leading zeros as recommended in the final standard).
•
The prefix field is often not used, limiting pathnames to the 100 characters of old-style archives.
POSIX ustar
Archives
IEEE Std 1003.1-1988 (‘‘POSIX.1’’)
defined a standard tar file format to be read and written by
compliant implementations of tar(1). This format is often
called the ‘‘ustar’’ format, after
the magic value used in the header. (The name is an acronym
for ‘‘Unix Standard TAR’’.) It
extends the historic format with new fields:
struct header_posix_ustar {
|
char name[100]; |
||
|
char mode[8]; |
||
|
char uid[8]; |
||
|
char gid[8]; |
||
|
char size[12]; |
||
|
char mtime[12]; |
||
|
char checksum[8]; |
||
|
char typeflag[1]; |
||
|
char linkname[100]; |
||
|
char magic[6]; |
||
|
char version[2]; |
||
|
char uname[32]; |
||
|
char gname[32]; |
||
|
char devmajor[8]; |
||
|
char devminor[8]; |
||
|
char prefix[155]; |
||
|
char pad[12]; |
};
typeflag
Type of entry. POSIX extended the earlier linkflag field with several new type values:
‘‘0’’
Regular file. NUL should be treated as a synonym, for compatibility purposes.
‘‘1’’
Hard link.
‘‘2’’
Symbolic link.
‘‘3’’
Character device node.
‘‘4’’
Block device node.
‘‘5’’
Directory.
‘‘6’’
FIFO node.
‘‘7’’
Reserved.
Other
A POSIX-compliant implementation must treat any unrecognized typeflag value as a regular file. In particular, writers should ensure that all entries have a valid filename so that they can be restored by readers that do not support the corresponding extension. Uppercase letters "A" through "Z" are reserved for custom extensions. Note that sockets and whiteout entries are not archivable.
It is worth noting that the size field, in particular, has different meanings depending on the type. For regular files, of course, it indicates the amount of data following the header. For directories, it may be used to indicate the total size of all files in the directory, for use by operating systems that pre-allocate directory space. For all other types, it should be set to zero by writers and ignored by readers.
magic
Contains the magic value ‘‘ustar’’ followed by a NUL byte to indicate that this is a POSIX standard archive. Full compliance requires the uname and gname fields be properly set.
version
Version. This should be ‘‘00’’ (two copies of the ASCII digit zero) for POSIX standard archives.
uname, gname
User and group names, as null-terminated ASCII strings. These should be used in preference to the uid/gid values when they are set and the corresponding names exist on the system.
devmajor, devminor
Major and minor numbers for character device or block device entry.
name, prefix
If the pathname is too long to fit in the 100 bytes provided by the standard format, it can be split at any / character with the first portion going into the prefix field. If the prefix field is not empty, the reader will prepend the prefix value and a / character to the regular name field to obtain the full pathname. The standard does not require a trailing / character on directory names, though most implementations still include this for compatibility reasons.
Note that all unused bytes must be set to NUL.
Field termination is specified slightly differently by POSIX than by previous implementations. The magic, uname, and gname fields must have a trailing NUL. The pathname, linkname, and prefix fields must have a trailing NUL unless they fill the entire field. (In particular, it is possible to store a 256-character pathname if it happens to have a / as the 156th character.) POSIX requires numeric fields to be zero-padded in the front, and requires them to be terminated with either space or NUL characters.
Currently, most tar implementations comply with the ustar format, occasionally extending it by adding new fields to the blank area at the end of the header record.
Pax
Interchange Format
There are many attributes that cannot be portably stored in
a POSIX ustar archive. IEEE Std 1003.1-2001
(‘‘POSIX.1’’) defined a
‘‘pax interchange format’’ that uses
two new types of entries to hold text-formatted metadata
that applies to following entries. Note that a pax
interchange format archive is a ustar archive in every
respect. The new data is stored in ustar-compatible archive
entries that use the ‘‘x’’ or
‘‘g’’ typeflag. In particular, older
implementations that do not fully support these extensions
will extract the metadata into regular files, where the
metadata can be examined as necessary.
An entry in a pax interchange format archive consists of one or two standard ustar entries, each with its own header and data. The first optional entry stores the extended attributes for the following entry. This optional first entry has an "x" typeflag and a size field that indicates the total size of the extended attributes. The extended attributes themselves are stored as a series of text-format lines encoded in the portable UTF-8 encoding. Each line consists of a decimal number, a space, a key string, an equals sign, a value string, and a new line. The decimal number indicates the length of the entire line, including the initial length field and the trailing newline. An example of such a field is:
25 ctime=1084839148.1212\n
Keys in all lowercase are standard keys. Vendors can add their own keys by prefixing them with an all uppercase vendor name and a period. Note that, unlike the historic header, numeric values are stored using decimal, not octal. A description of some common keys follows:
atime, ctime, mtime
File access, inode change, and modification times. These fields can be negative or include a decimal point and a fractional value.
uname, uid, gname, gid
User name, group name, and numeric UID and GID values. The user name and group name stored here are encoded in UTF8 and can thus include non-ASCII characters. The UID and GID fields can be of arbitrary length.
linkpath
The full path of the linked-to file. Note that this is encoded in UTF8 and can thus include non-ASCII characters.
path
The full pathname of the entry. Note that this is encoded in UTF8 and can thus include non-ASCII characters.
realtime.*, security.*
These keys are reserved and may be used for future standardization.
size
The size of the file. Note that there is no length limit on this field, allowing conforming archives to store files much larger than the historic 8GB limit.
SCHILY.*
Vendor-specific attributes used by Joerg Schilling’s star implementation.
SCHILY.acl.access, SCHILY.acl.default
Stores the access and default ACLs as textual strings in a format that is an extension of the format specified by POSIX.1e draft 17. In particular, each user or group access specification can include a fourth colon-separated field with the numeric UID or GID. This allows ACLs to be restored on systems that may not have complete user or group information available (such as when NIS/YP or LDAP services are temporarily unavailable).
SCHILY.devminor, SCHILY.devmajor
The full minor and major numbers for device nodes.
SCHILY.fflags
The file flags.
SCHILY.realsize
The full size of the file on disk. XXX explain? XXX
SCHILY.dev, SCHILY.ino, SCHILY.nlinks
The device number, inode number, and link count for the entry. In particular, note that a pax interchange format archive using Joerg Schilling’s SCHILY.* extensions can store all of the data from struct stat.
LIBARCHIVE.xattr.namespace.key
Libarchive stores POSIX.1e-style extended attributes using keys of this form. The key value is URL-encoded: All non-ASCII characters and the two special characters ‘‘=’’ and ‘‘%’’ are encoded as ‘‘%’’ followed by two uppercase hexadecimal digits. The value of this key is the extended attribute value encoded in base 64. XXX Detail the base-64 format here XXX
VENDOR.*
XXX document other vendor-specific extensions XXX
Any values stored in an extended attribute override the corresponding values in the regular tar header. Note that compliant readers should ignore the regular fields when they are overridden. This is important, as existing archivers are known to store non-compliant values in the standard header fields in this situation. There are no limits on length for any of these fields. In particular, numeric fields can be arbitrarily large. All text fields are encoded in UTF8. Compliant writers should store only portable 7-bit ASCII characters in the standard ustar header and use extended attributes whenever a text value contains non-ASCII characters.
In addition to the x entry described above, the pax interchange format also supports a g entry. The g entry is identical in format, but specifies attributes that serve as defaults for all subsequent archive entries. The g entry is not widely used.
Besides the new x and g entries, the pax interchange format has a few other minor variations from the earlier ustar format. The most troubling one is that hardlinks are permitted to have data following them. This allows readers to restore any hardlink to a file without having to rewind the archive to find an earlier entry. However, it creates complications for robust readers, as it is no longer clear whether or not they should ignore the size field for hardlink entries.
GNU Tar
Archives
The GNU tar program started with a pre-POSIX format similar
to that described earlier and has extended it using several
different mechanisms: It added new fields to the empty space
in the header (some of which was later used by POSIX for
conflicting purposes); it allowed the header to be continued
over multiple records; and it defined new entries that
modify following entries (similar in principle to the
x entry described above, but each GNU special entry
is single-purpose, unlike the general-purpose x
entry). As a result, GNU tar archives are not POSIX
compatible, although more lenient POSIX-compliant readers
can successfully extract most GNU tar archives.
struct header_gnu_tar {
|
char name[100]; |
|||
|
char mode[8]; |
|||
|
char uid[8]; |
|||
|
char gid[8]; |
|||
|
char size[12]; |
|||
|
char mtime[12]; |
|||
|
char checksum[8]; |
|||
|
char typeflag[1]; |
|||
|
char linkname[100]; |
|||
|
char magic[6]; |
|||
|
char version[2]; |
|||
|
char uname[32]; |
|||
|
char gname[32]; |
|||
|
char devmajor[8]; |
|||
|
char devminor[8]; |
|||
|
char atime[12]; |
|||
|
char ctime[12]; |
|||
|
char offset[12]; |
|||
|
char longnames[4]; |
|||
|
char unused[1]; |
|||
|
struct { |
|||
|
char offset[12]; |
|||
|
char numbytes[12]; |
|||
|
} sparse[4]; |
|||
|
char isextended[1]; |
|||
|
char realsize[12]; |
|||
|
char pad[17]; |
};
typeflag
GNU tar uses the following special entry types, in addition to those defined by POSIX:
7
GNU tar treats type "7" records identically to type "0" records, except on one obscure RTOS where they are used to indicate the pre-allocation of a contiguous file on disk.
D
This indicates a directory entry. Unlike the POSIX-standard "5" typeflag, the header is followed by data records listing the names of files in this directory. Each name is preceded by an ASCII "Y" if the file is stored in this archive or "N" if the file is not stored in this archive. Each name is terminated with a null, and an extra null marks the end of the name list. The purpose of this entry is to support incremental backups; a program restoring from such an archive may wish to delete files on disk that did not exist in the directory when the archive was made.
Note that the "D" typeflag specifically violates POSIX, which requires that unrecognized typeflags be restored as normal files. In this case, restoring the "D" entry as a file could interfere with subsequent creation of the like-named directory.
K
The data for this entry is a long linkname for the following regular entry.
L
The data for this entry is a long pathname for the following regular entry.
M
This is a continuation of the last file on the previous volume. GNU multi-volume archives guarantee that each volume begins with a valid entry header. To ensure this, a file may be split, with part stored at the end of one volume, and part stored at the beginning of the next volume. The "M" typeflag indicates that this entry continues an existing file. Such entries can only occur as the first or second entry in an archive (the latter only if the first entry is a volume label). The size field specifies the size of this entry. The offset field at bytes 369-380 specifies the offset where this file fragment begins. The realsize field specifies the total size of the file (which must equal size plus offset). When extracting, GNU tar checks that the header file name is the one it is expecting, that the header offset is in the correct sequence, and that the sum of offset and size is equal to realsize.
N
Type "N" records are no longer generated by GNU tar. They contained a list of files to be renamed or symlinked after extraction; this was originally used to support long names. The contents of this record are a text description of the operations to be done, in the form ‘‘Rename %s to %s\n’’ or ‘‘Symlink %s to %s\n’’; in either case, both filenames are escaped using K&R C syntax. Due to security concerns, "N" records are now generally ignored when reading archives.
S
This is a ‘‘sparse’’ regular file. Sparse files are stored as a series of fragments. The header contains a list of fragment offset/length pairs. If more than four such entries are required, the header is extended as necessary with ‘‘extra’’ header extensions (an older format that is no longer used), or ‘‘sparse’’ extensions.
V
The name field should be interpreted as a tape/volume header name. This entry should generally be ignored on extraction.
magic
The magic field holds the five characters ‘‘ustar’’ followed by a space. Note that POSIX ustar archives have a trailing null.
version
The version field holds a space character followed by a null. Note that POSIX ustar archives use two copies of the ASCII digit ‘‘0’’.
atime, ctime
The time the file was last accessed and the time of last change of file information, stored in octal as with mtime.
longnames
This field is apparently no longer used.
Sparse offset / numbytes
Each such structure specifies a single fragment of a sparse file. The two fields store values as octal numbers. The fragments are each padded to a multiple of 512 bytes in the archive. On extraction, the list of fragments is collected from the header (including any extension headers), and the data is then read and written to the file at appropriate offsets.
isextended
If this is set to non-zero, the header will be followed by additional ‘‘sparse header’’ records. Each such record contains information about as many as 21 additional sparse blocks as shown here:
struct gnu_sparse_header {
|
struct { |
|||
|
char offset[12]; |
|||
|
char numbytes[12]; |
|||
|
} sparse[21]; |
|||
|
char isextended[1]; |
|||
|
char padding[7]; |
};
realsize
A binary representation of the file’s complete size, with a much larger range than the POSIX file size. In particular, with M type files, the current entry is only a portion of the file. In that case, the POSIX size field will indicate the size of this entry; the realsize field will indicate the total size of the file.
GNU tar pax
archives
GNU tar 1.14 (XXX check this XXX) and later will write pax
interchange format archives when you specify the
−-posix flag. This format uses custom keywords
to store sparse file information. There have been three
iterations of this support, referred to as
‘‘0.0’’,
‘‘0.1’’, and
‘‘1.0’’.
GNU.sparse.numblocks, GNU.sparse.offset, GNU.sparse.numbytes, GNU.sparse.size
The ‘‘0.0’’ format used an initial GNU.sparse.numblocks attribute to indicate the number of blocks in the file, a pair of GNU.sparse.offset and GNU.sparse.numbytes to indicate the offset and size of each block, and a single GNU.sparse.size to indicate the full size of the file. This is not the same as the size in the tar header because the latter value does not include the size of any holes. This format required that the order of attributes be preserved and relied on readers accepting multiple appearances of the same attribute names, which is not officially permitted by the standards.
GNU.sparse.map
The ‘‘0.1’’ format used a single attribute that stored a comma-separated list of decimal numbers. Each pair of numbers indicated the offset and size, respectively, of a block of data. This does not work well if the archive is extracted by an archiver that does not recognize this extension, since many pax implementations simply discard unrecognized attributes.
GNU.sparse.major, GNU.sparse.minor, GNU.sparse.name, GNU.sparse.realsize
The ‘‘1.0’’ format stores the sparse block map in one or more 512-byte blocks prepended to the file data in the entry body. The pax attributes indicate the existence of this map (via the GNU.sparse.major and GNU.sparse.minor fields) and the full size of the file. The GNU.sparse.name holds the true name of the file. To avoid confusion, the name stored in the regular tar header is a modified name so that extraction errors will be apparent to users.
Solaris
Tar
XXX More Details Needed XXX
Solaris tar (beginning with SunOS XXX 5.7 ?? XXX) supports an ‘‘extended’’ format that is fundamentally similar to pax interchange format, with the following differences:
•
Extended attributes are stored in an entry whose type is X, not x, as used by pax interchange format. The detailed format of this entry appears to be the same as detailed above for the x entry.
•
An additional A entry is used to store an ACL for the following regular entry. The body of this entry contains a seven-digit octal number followed by a zero byte, followed by the textual ACL description. The octal value is the number of ACL entries plus a constant that indicates the ACL type: 01000000 for POSIX.1e ACLs and 03000000 for NFSv4 ACLs.
AIX Tar
XXX More details needed XXX
Mac OS X
Tar
The tar distributed with Apple’s Mac OS X stores most
regular files as two separate entries in the tar archive.
The two entries have the same name except that the first one
has ‘‘._’’ added to the beginning of
the name. This first entry stores the ‘‘resource
fork’’ with additional attributes for the file.
The Mac OS X CopyFile() API is used to separate a
file on disk into separate resource and data streams and to
reassemble those separate streams when the file is restored
to disk.
Other
Extensions
One obvious extension to increase the size of files is to
eliminate the terminating characters from the various
numeric fields. For example, the standard only allows the
size field to contain 11 octal digits, reserving the twelfth
byte for a trailing NUL character. Allowing 12 octal digits
allows file sizes up to 64 GB.
Another extension, utilized by GNU tar, star, and other newer tar implementations, permits binary numbers in the standard numeric fields. This is flagged by setting the high bit of the first byte. This permits 95-bit values for the length and time fields and 63-bit values for the uid, gid, and device numbers. GNU tar supports this extension for the length, mtime, ctime, and atime fields. Joerg Schilling’s star program supports this extension for all numeric fields. Note that this extension is largely obsoleted by the extended attribute record provided by the pax interchange format.
Another early GNU extension allowed base-64 values rather than octal. This extension was short-lived and is no longer supported by any implementation.
SEE ALSO
ar(1), pax(1), tar(1)
STANDARDS
The tar utility is no longer a part of POSIX or the Single Unix Standard. It last appeared in Version 2 of the Single UNIX Specification (‘‘SUSv2’’). It has been supplanted in subsequent standards by pax(1). The ustar format is currently part of the specification for the pax(1) utility. The pax interchange file format is new with IEEE Std 1003.1-2001 (‘‘POSIX.1’’).
HISTORY
A tar command appeared in Seventh Edition Unix, which was released in January, 1979. It replaced the tp program from Fourth Edition Unix which in turn replaced the tap program from First Edition Unix. John Gilmore’s pdtar public-domain implementation (circa 1987) was highly influential and formed the basis of GNU tar (circa 1988). Joerg Shilling’s star archiver is another open-source (GPL) archiver (originally developed circa 1985) which features complete support for pax interchange format.
This documentation was written as part of the libarchive and bsdtar project by Tim Kientzle ⟨kientzle@@FreeBSD.org⟩.
FreeBSD 9.0 December 27, 2009 FreeBSD 9.0
)36 48 Q -.15(Fo)0 60 S 2.5(rc).15 G(ompatibility with tape dri)-2.5 E -.15(ve) -.25 G 2.5(st).15 G(hat use \214x)-2.5 E(ed block sizes,)-.15 E (programs that read or write tar \214les al)0 72 Q -.1(wa)-.1 G (ys read or write a \214x).1 E(ed)-.15 E 0 Cg EP %%Page: 9 9 %%BeginPageSetup BP %%EndPageSetup /F0 10/Times-Roman@@0 SF(number of records with each I/O operation.)0 12 Q(These)0 24 Q -.74(``)0 36 S(blocks').74 E(')-.74 E(are al)0 48 Q -.1 (wa)-.1 G(ys a multiple of the record size.).1 E (The maximum block size supported by early)0 60 Q(implementations w)0 72 Q(as 10240 bytes or 20 records.)-.1 E 0 Cg EP %%Page: 10 10 %%BeginPageSetup BP %%EndPageSetup /F0 10/Times-Roman@@0 SF(This is still the def)0 12 Q (ault for most implementations)-.1 E (although block sizes of 1MiB \(2048 records\) or lar)0 24 Q(ger are) -.18 E(commonly used with modern high-speed tape dri)0 36 Q -.15(ve)-.25 G(s.).15 E(\(Note: the terms)0 48 Q -.74(``)0 60 S(block').74 E(')-.74 E (and)0 72 Q 0 Cg EP %%Page: 11 11 %%BeginPageSetup BP %%EndPageSetup /F0 10/Times-Roman@@0 SF -.74(``)0 12 S(record').74 E(')-.74 E (here are not entirely standard; this document follo)0 24 Q(ws the)-.25 E(con)0 36 Q -.15(ve)-.4 G (ntion established by John Gilmore in documenting).15 E(pdtar.\)) 0 48 Q(
A)1.11 E
(T&T UNIX,)-1.11 E(which seems to be the earliest widely-used v)0
24 Q(ersion of the tar program.)-.15 E()36 36 Q (The header record for an old-style)0 48 Q(tar)0 60 Q(archi)0 72 Q .3 -.15(ve c)-.25 H(onsists of the follo).15 E(wing:)-.25 E 0 Cg EP %%Page: 14 14 %%BeginPageSetup BP %%EndPageSetup /F0 10/Times-Roman@@0 SF(
)0 24 Q()
0 36 Q(struct header_old_tar {)0 48 Q(char name[100];)36 60 Q
(char mode[8];)36 72 Q 0 Cg EP
%%Page: 15 15
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(char uid[8];)36 12 Q(char gid[8];)36 24 Q
(char size[12];)36 36 Q(char mtime[12];)36 48 Q(char checksum[8];)36 60
Q(char link\215ag[1];)36 72 Q 0 Cg EP
%%Page: 16 16
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(char linkname[100];)36 12 Q(char pad[255];)36 24
Q(};)0 36 Q()0 48 Q()0 60 Q()36
60 Q(Early tar implementations v)0 72 Q(aried in ho)-.25 E 2.5(wt)-.25 G
(he)-2.5 E 2.5(yt)-.15 G(erminated these \214elds.)-2.5 E 0 Cg EP
%%Page: 26 26
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(The tar command in)0 12 Q -1.11(Ve)0 24 S
(rsion 7 A)1.11 E(T&T UNIX)-1.11 E(used the follo)0 36 Q
(wing con)-.25 E -.15(ve)-.4 G
(ntions \(this is also documented in early BSD manpages\):).15 E
(the pathname must be null-terminated;)0 48 Q
(the mode, uid, and gid \214elds must end in a space and a null byte;)0
60 Q(the size and mtime \214elds must end in a space;)0 72 Q 0 Cg EP
%%Page: 27 27
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF
(the checksum is terminated by a null and a space.)0 12 Q(Early impleme\
ntations \214lled the numeric \214elds with leading spaces.)0 24 Q
(This seems to ha)0 36 Q .3 -.15(ve b)-.2 H
(een common practice until the).15 E(IEEE Std 1003.1-1988 \(`)2.5 48 Q
(`POSIX.1')-.74 E('\))-.74 E(standard w)0 60 Q
(as released.)-.1 E -.15(Fo)0 72 S 2.5(rb).15 G(est portability)-2.5 E
2.5(,m)-.65 G(odern implementations should \214ll the numeric)-2.5 E 0
Cg EP
%%Page: 28 28
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(\214elds with leading zeros.)0 12 Q
(
)0 60 Q()0 72 Q 0 Cg EP
%%Page: 35 35
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(struct header_posix_ustar {)0 12 Q
(char name[100];)36 24 Q(char mode[8];)36 36 Q(char uid[8];)36 48 Q
(char gid[8];)36 60 Q(char size[12];)36 72 Q 0 Cg EP
%%Page: 36 36
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(char mtime[12];)36 12 Q(char checksum[8];)36 24
Q(char type\215ag[1];)36 36 Q(char linkname[100];)36 48 Q
(char magic[6];)36 60 Q(char v)36 72 Q(ersion[2];)-.15 E 0 Cg EP
%%Page: 37 37
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(char uname[32];)36 12 Q(char gname[32];)36 24 Q
(char de)36 36 Q(vmajor[8];)-.25 E(char de)36 48 Q(vminor[8];)-.25 E
(char pre\214x[155];)36 60 Q(char pad[12];)36 72 Q 0 Cg EP
%%Page: 38 38
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(};)0 12 Q()0 24 Q()0 36 Q(/)0 24 Q
(character with the \214rst portion going into the pre\214x \214eld.)0
36 Q(If the pre\214x \214eld is not empty)0 48 Q 2.5(,t)-.65 G
(he reader will prepend)-2.5 E(the pre\214x v)0 60 Q(alue and a)-.25 E
(/)0 72 Q 0 Cg EP
%%Page: 50 50
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(character to the re)0 12 Q
(gular name \214eld to obtain the full pathname.)-.15 E
(The standard does not require a trailing)0 24 Q
(/)0 36 Q
(character on directory names, though most implementations still)0 48 Q
(include this for compatibility reasons.)0 60 Q()36 12 Q
(Note that all unused bytes must be set to)0 24 Q
(NUL.)0 36 Q(
)36 48 Q
(Field termination is speci\214ed slightly dif)0 60 Q(ferently by POSIX)
-.25 E(than by pre)0 72 Q(vious implementations.)-.25 E 0 Cg EP
%%Page: 52 52
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(The)0 12 Q(magic,)0 24 Q
(uname,)0 36 Q(and)0 48 Q(gname)0 60 Q
(\214elds must ha)0 72 Q .3 -.15(ve a t)-.2 H(railing).15 E 0 Cg EP
%%Page: 53 53
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(NUL.)0 12 Q(The)0 24 Q
(pathname,)0 36 Q(linkname,)0 48 Q(and)0 60
Q(pre\214x)0 72 Q 0 Cg EP
%%Page: 54 54
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(\214elds must ha)0 12 Q .3 -.15(ve a t)-.2 H
(railing).15 E(NUL)0 24 Q(unless the)0 36 Q
2.5<798c>-.15 G(ll the entire \214eld.)-2.5 E(\(In particular)0 48 Q 2.5
(,i)-.4 G 2.5(ti)-2.5 G 2.5(sp)-2.5 G
(ossible to store a 256-character pathname if it)-2.5 E(happens to ha)0
60 Q .3 -.15(ve a)-.2 H(/)0 72 Q 0 Cg EP
%%Page: 55 55
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(as the 156th character)0 12 Q(.\))-.55 E(POSIX \
requires numeric \214elds to be zero-padded in the front, and requires)0
24 Q(them to be terminated with either space or)0 36 Q
(NUL)0 48 Q(characters.)0 60 Q(
)36 72 Q 0 Cg EP %%Page: 56 56 %%BeginPageSetup BP %%EndPageSetup /F0 10/Times-Roman@@0 SF(Currently)0 12 Q 2.5(,m)-.65 G (ost tar implementations comply with the ustar)-2.5 E (format, occasionally e)0 24 Q(xtending it by adding ne)-.15 E 2.5<778c> -.25 G(elds to the)-2.5 E(blank area at the end of the header record.)0 36 Q(
)36 72 Q 0 Cg EP
%%Page: 60 60
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(An entry in a pax interchange format archi)0 12
Q .3 -.15(ve c)-.25 H(onsists of one or).15 E(tw)0 24 Q 2.5(os)-.1 G
(tandard ustar entries, each with its o)-2.5 E(wn header and data.)-.25
E(The \214rst optional entry stores the e)0 36 Q(xtended attrib)-.15 E
(utes)-.2 E(for the follo)0 48 Q(wing entry)-.25 E(.)-.65 E(This option\
al \214rst entry has an "x" type\215ag and a size \214eld that)0 60 Q
(indicates the total size of the e)0 72 Q(xtended attrib)-.15 E(utes.)
-.2 E 0 Cg EP
%%Page: 61 61
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(The e)0 12 Q(xtended attrib)-.15 E
(utes themselv)-.2 E(es are stored as a series of te)-.15 E(xt-format)
-.15 E(lines encoded in the portable UTF-8 encoding.)0 24 Q
(Each line consists of a decimal number)0 36 Q 2.5(,as)-.4 G(pace, a k)
-2.5 E .3 -.15(ey s)-.1 H(tring, an equals).15 E(sign, a v)0 48 Q
(alue string, and a ne)-.25 E 2.5(wl)-.25 G(ine.)-2.5 E(The decimal num\
ber indicates the length of the entire line, including the)0 60 Q
(initial length \214eld and the trailing ne)0 72 Q(wline.)-.25 E 0 Cg EP
%%Page: 62 62
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(An e)0 12 Q(xample of such a \214eld is:)-.15 E
(
)0 24 Q( )0 36 Q
(25 ctime=1084839148.1212\\n)0 48
Q(
)0 60 Q -2.15 -.25(Ke y)0 72 T 2.5(si).25 G 2.5(na)-2.5 G(ll lo)
-2.5 E(wercase are standard k)-.25 E -.15(ey)-.1 G(s.).15 E 0 Cg EP
%%Page: 63 63
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF -1.11(Ve)0 12 S(ndors can add their o)1.11 E
(wn k)-.25 E -.15(ey)-.1 G 2.5(sb).15 G 2.5(yp)-2.5 G
(re\214xing them with an all uppercase)-2.5 E -.15(ve)0 24 S
(ndor name and a period.).15 E(Note that, unlik)0 36 Q 2.5(et)-.1 G
(he historic header)-2.5 E 2.5(,n)-.4 G(umeric v)-2.5 E
(alues are stored using)-.25 E(decimal, not octal.)0 48 Q 2.5(Ad)0 60 S
(escription of some common k)-2.5 E -.15(ey)-.1 G 2.5(sf).15 G(ollo)-2.5
E(ws:)-.25 E(
) 36 72 Q 0 Cg EP %%Page: 75 75 %%BeginPageSetup BP %%EndPageSetup /F0 10/Times-Roman@@0 SF(An)0 12 Q 2.5(yv)-.15 G(alues stored in an e) -2.75 E(xtended attrib)-.15 E(ute o)-.2 E -.15(ve)-.15 G (rride the corresponding).15 E -.25(va)0 24 S(lues in the re).25 E (gular tar header)-.15 E(.)-.55 E (Note that compliant readers should ignore the re)0 36 Q (gular \214elds when the)-.15 E(y)-.15 E(are o)0 48 Q -.15(ve)-.15 G (rridden.).15 E(This is important, as e)0 60 Q(xisting archi)-.15 E -.15 (ve)-.25 G(rs are kno).15 E(wn to store non-compliant)-.25 E -.25(va)0 72 S(lues in the standard header \214elds in this situation.).25 E 0 Cg EP %%Page: 76 76 %%BeginPageSetup BP %%EndPageSetup /F0 10/Times-Roman@@0 SF(There are no limits on length for an)0 12 Q 2.5 (yo)-.15 G 2.5(ft)-2.5 G(hese \214elds.)-2.5 E(In particular)0 24 Q 2.5 (,n)-.4 G(umeric \214elds can be arbitrarily lar)-2.5 E(ge.)-.18 E (All te)0 36 Q(xt \214elds are encoded in UTF8.)-.15 E (Compliant writers should store only portable 7-bit ASCII characters in) 0 48 Q(the standard ustar header and use e)0 60 Q(xtended)-.15 E(attrib) 0 72 Q(utes whene)-.2 E -.15(ve)-.25 G 2.5(rat).15 G -.15(ex)-2.5 G 2.5 (tv).15 G(alue contains non-ASCII characters.)-2.75 E 0 Cg EP %%Page: 77 77 %%BeginPageSetup BP %%EndPageSetup /F0 10/Times-Roman@@0 SF(
)36 12 Q(In addition to the)0 24 Q (x)0 36 Q(entry described abo)0 48 Q -.15(ve)-.15 G 2.5 (,t).15 G(he pax interchange format)-2.5 E(also supports a)0 60 Q (g)0 72 Q 0 Cg EP %%Page: 78 78 %%BeginPageSetup BP %%EndPageSetup /F0 10/Times-Roman@@0 SF(entry)0 12 Q(.)-.65 E(The)0 24 Q (g)0 36 Q(entry is identical in format, b)0 48 Q (ut speci\214es attrib)-.2 E(utes that serv)-.2 E 2.5(ea)-.15 G(s)-2.5 E (def)0 60 Q(aults for all subsequent archi)-.1 E .3 -.15(ve e)-.25 H (ntries.).15 E(The)0 72 Q 0 Cg EP %%Page: 79 79 %%BeginPageSetup BP %%EndPageSetup /F0 10/Times-Roman@@0 SF(g)0 12 Q (entry is not widely used.)0 24 Q(
)36 36 Q(Besides the ne)0 48 Q(w) -.25 E(x)0 60 Q(and)0 72 Q 0 Cg EP %%Page: 80 80 %%BeginPageSetup BP %%EndPageSetup /F0 10/Times-Roman@@0 SF(g)0 12 Q (entries, the pax interchange format has a fe)0 24 Q 2.5(wo)-.25 G (ther minor v)-2.5 E(ariations)-.25 E(from the earlier ustar format.)0 36 Q(The most troubling one is that hardlinks are permitted to ha)0 48 Q -.15(ve)-.2 G(data follo)0 60 Q(wing them.)-.25 E(This allo)0 72 Q (ws readers to restore an)-.25 E 2.5(yh)-.15 G (ardlink to a \214le without)-2.5 E 0 Cg EP %%Page: 81 81 %%BeginPageSetup BP %%EndPageSetup /F0 10/Times-Roman@@0 SF(ha)0 12 Q(ving to re)-.2 E(wind the archi)-.25 E .3 -.15(ve t)-.25 H 2.5<6f8c>.15 G(nd an earlier entry)-2.5 E(.)-.65 E (Ho)0 24 Q(we)-.25 E -.15(ve)-.25 G .8 -.4(r, i).15 H 2.5(tc).4 G (reates complications for rob)-2.5 E(ust readers, as it is no longer)-.2 E(clear whether or not the)0 36 Q 2.5(ys)-.15 G (hould ignore the size \214eld for hardlink entries.)-2.5 E (
)0 60 Q()0 72 Q 0 Cg EP
%%Page: 85 85
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(struct header_gnu_tar {)0 12 Q(char name[100];)
36 24 Q(char mode[8];)36 36 Q(char uid[8];)36 48 Q(char gid[8];)36 60 Q
(char size[12];)36 72 Q 0 Cg EP
%%Page: 86 86
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(char mtime[12];)36 12 Q(char checksum[8];)36 24
Q(char type\215ag[1];)36 36 Q(char linkname[100];)36 48 Q
(char magic[6];)36 60 Q(char v)36 72 Q(ersion[2];)-.15 E 0 Cg EP
%%Page: 87 87
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(char uname[32];)36 12 Q(char gname[32];)36 24 Q
(char de)36 36 Q(vmajor[8];)-.25 E(char de)36 48 Q(vminor[8];)-.25 E
(char atime[12];)36 60 Q(char ctime[12];)36 72 Q 0 Cg EP
%%Page: 88 88
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(char of)36 12 Q(fset[12];)-.25 E
(char longnames[4];)36 24 Q(char unused[1];)36 36 Q(struct {)36 48 Q
(char of)72 60 Q(fset[12];)-.25 E(char numbytes[12];)72 72 Q 0 Cg EP
%%Page: 89 89
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF 2.5(}s)36 12 S(parse[4];)-2.5 E(char ise)36 24 Q
(xtended[1];)-.15 E(char realsize[12];)36 36 Q(char pad[17];)36 48 Q(};)
0 60 Q()0 72 Q 0 Cg EP
%%Page: 90 90
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF()0 12 Q()36 24 Q(Note that the "D" type\215ag sp\ eci\214cally violates POSIX, which requires)0 36 Q (that unrecognized type\215ags be restored as normal \214les.)0 48 Q (In this case, restoring the "D" entry as a \214le could interfere)0 60 Q(with subsequent creation of the lik)0 72 Q(e-named directory)-.1 E(.) -.65 E 0 Cg EP %%Page: 95 95 %%BeginPageSetup BP %%EndPageSetup /F0 10/Times-Roman@@0 SF(
)0 36 Q()0 48 Q
(struct gnu_sparse_header {)0 60 Q(struct {)36 72 Q 0 Cg EP
%%Page: 111 111
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF(char of)72 12 Q(fset[12];)-.25 E
(char numbytes[12];)72 24 Q 2.5(}s)36 36 S(parse[21];)-2.5 E 7.5
(char ise)36 48 R(xtended[1];)-.15 E 7.5(char padding[7];)36 60 R(};)0
72 Q 0 Cg EP
%%Page: 112 112
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF()0 12 Q()0 24 Q()36 36 Q(Solaris tar \(be)0 48 Q (ginning with SunOS XXX 5.7 ?? XXX\) supports an)-.15 E -.74(``)0 60 S -.15(ex).74 G(tended').15 E(')-.74 E (format that is fundamentally similar to pax interchange format,)0 72 Q 0 Cg EP %%Page: 126 126 %%BeginPageSetup BP %%EndPageSetup /F0 10/Times-Roman@@0 SF(with the follo)0 12 Q(wing dif)-.25 E(ferences:) -.25 E(
)36 12 Q(Another e)0 24 Q (xtension, utilized by GNU tar)-.15 E 2.5(,s)-.4 G(tar)-2.5 E 2.5(,a)-.4 G(nd other ne)-2.5 E(wer)-.25 E(tar)0 36 Q(implementations, perm\ its binary numbers in the standard numeric \214elds.)0 48 Q (This is \215agged by setting the high bit of the \214rst byte.)0 60 Q (This permits 95-bit v)0 72 Q(alues for the length and time \214elds) -.25 E 0 Cg EP %%Page: 136 136 %%BeginPageSetup BP %%EndPageSetup /F0 10/Times-Roman@@0 SF(and 63-bit v)0 12 Q (alues for the uid, gid, and de)-.25 E(vice numbers.)-.25 E (GNU tar supports this e)0 24 Q(xtension for the)-.15 E (length, mtime, ctime, and atime \214elds.)0 36 Q(Joer)0 48 Q 2.5(gS) -.18 G(chilling')-2.5 E 2.5(ss)-.55 G(tar program supports this e)-2.5 E (xtension for)-.15 E(all numeric \214elds.)0 60 Q(Note that this e)0 72 Q(xtension is lar)-.15 E(gely obsoleted by the e)-.18 E(xtended attrib) -.15 E(ute)-.2 E 0 Cg EP %%Page: 137 137 %%BeginPageSetup BP %%EndPageSetup /F0 10/Times-Roman@@0 SF(record pro)0 12 Q (vided by the pax interchange format.)-.15 E(
)36 24 Q (Another early GNU e)0 36 Q(xtension allo)-.15 E(wed base-64 v)-.25 E (alues rather than octal.)-.25 E(This e)0 48 Q(xtension w)-.15 E (as short-li)-.1 E -.15(ve)-.25 G 2.5(da).15 G (nd is no longer supported by an)-2.5 E(y)-.15 E(implementation.)0 60 Q (
)36 12 Q(This documentation w)0 24 Q
(as written as part of the)-.1 E(libarchi)0 36 Q -.15(ve)-.25 G()
.15 E(and)0 48 Q(bsdtar)0 60 Q(project by)0 72 Q 0 Cg EP
%%Page: 146 146
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Times-Roman@@0 SF -.35(Ti)0 12 S 2.5(mK).35 G
(ientzle <kientzle@@FreeBSD.or)-2.5 E(g>
.)-.18 E()0 24 Q
()0 36 Q 0 Cg EP
%%Trailer
end
%%EOF
@
1.1.1.5
log
@Import libarchive-3.2.1:
- security fixes and other bugfixes
- support for multhreading in xz 5.2+
@
text
@d1 1657
a1657 1558
TAR(5) BSD File Formats Manual TAR(5)
NAME
tar — format of tape archive files
DESCRIPTION
The tar archive format collects any number of files, directories, and other file system objects (symbolic links, device nodes, etc.) into a single stream of bytes. The format was originally designed to be used with tape drives that operate with fixed-size blocks, but is widely used as a general packaging mechanism.
General
Format
A tar archive consists of a series of 512-byte
records. Each file system object requires a header record
which stores basic metadata (pathname, owner, permissions,
etc.) and zero or more records containing any file data. The
end of the archive is indicated by two records consisting
entirely of zero bytes.
For compatibility with tape drives that use fixed block sizes, programs that read or write tar files always read or write a fixed number of records with each I/O operation. These ’’blocks’’ are always a multiple of the record size. The maximum block size supported by early implementations was 10240 bytes or 20 records. This is still the default for most implementations although block sizes of 1MiB (2048 records) or larger are commonly used with modern high-speed tape drives. (Note: the terms ’’block’’ and ’’record’’ here are not entirely standard; this document follows the convention established by John Gilmore in documenting pdtar.)
Old-Style
Archive Format
The original tar archive format has been extended many times
to include additional information that various implementors
found necessary. This section describes the variant
implemented by the tar command included in Version 7
AT&T UNIX, which seems to be the earliest widely-used
version of the tar program.
The header record for an old-style tar archive consists of the following:
struct header_old_tar {
|
char name[100]; |
||
|
char mode[8]; |
||
|
char uid[8]; |
||
|
char gid[8]; |
||
|
char size[12]; |
||
|
char mtime[12]; |
||
|
char checksum[8]; |
||
|
char linkflag[1]; |
||
|
char linkname[100]; |
||
|
char pad[255]; |
};
All unused bytes in the header record are filled with nulls.
name
Pathname, stored as a null-terminated string. Early tar implementations only stored regular files (including hardlinks to those files). One common early convention used a trailing "/" character to indicate a directory name, allowing directory permissions and owner information to be archived and restored.
mode
File mode, stored as an octal number in ASCII.
uid, gid
User id and group id of owner, as octal numbers in ASCII.
size
Size of file, as octal number in ASCII. For regular files only, this indicates the amount of data that follows the header. In particular, this field was ignored by early tar implementations when extracting hardlinks. Modern writers should always store a zero length for hardlink entries.
mtime
Modification time of file, as an octal number in ASCII. This indicates the number of seconds since the start of the epoch, 00:00:00 UTC January 1, 1970. Note that negative values should be avoided here, as they are handled inconsistently.
checksum
Header checksum, stored as an octal number in ASCII. To compute the checksum, set the checksum field to all spaces, then sum all bytes in the header using unsigned arithmetic. This field should be stored as six octal digits followed by a null and a space character. Note that many early implementations of tar used signed arithmetic for the checksum field, which can cause interoperability problems when transferring archives between systems. Modern robust readers compute the checksum both ways and accept the header if either computation matches.
linkflag, linkname
In order to preserve hardlinks and conserve tape, a file with multiple links is only written to the archive the first time it is encountered. The next time it is encountered, the linkflag is set to an ASCII ’1’ and the linkname field holds the first name under which this file appears. (Note that regular files have a null value in the linkflag field.)
Early tar implementations varied in how they terminated these fields. The tar command in Version 7 AT&T UNIX used the following conventions (this is also documented in early BSD manpages): the pathname must be null-terminated; the mode, uid, and gid fields must end in a space and a null byte; the size and mtime fields must end in a space; the checksum is terminated by a null and a space. Early implementations filled the numeric fields with leading spaces. This seems to have been common practice until the IEEE Std 1003.1-1988 (’’POSIX.1’’) standard was released. For best portability, modern implementations should fill the numeric fields with leading zeros.
Pre-POSIX
Archives
An early draft of IEEE Std 1003.1-1988
(’’POSIX.1’’) served as the basis
for John Gilmore’s pdtar program and many
system implementations from the late 1980s and early 1990s.
These archives generally follow the POSIX ustar format
described below with the following variations:
•
The magic value consists of the five characters ’’ustar’’ followed by a space. The version field contains a space character followed by a null.
•
The numeric fields are generally filled with leading spaces (not leading zeros as recommended in the final standard).
•
The prefix field is often not used, limiting pathnames to the 100 characters of old-style archives.
POSIX ustar
Archives
IEEE Std 1003.1-1988 (’’POSIX.1’’)
defined a standard tar file format to be read and written by
compliant implementations of tar(1). This format is often
called the ’’ustar’’ format, after
the magic value used in the header. (The name is an acronym
for ’’Unix Standard TAR’’.) It
extends the historic format with new fields:
struct header_posix_ustar {
|
char name[100]; |
||
|
char mode[8]; |
||
|
char uid[8]; |
||
|
char gid[8]; |
||
|
char size[12]; |
||
|
char mtime[12]; |
||
|
char checksum[8]; |
||
|
char typeflag[1]; |
||
|
char linkname[100]; |
||
|
char magic[6]; |
||
|
char version[2]; |
||
|
char uname[32]; |
||
|
char gname[32]; |
||
|
char devmajor[8]; |
||
|
char devminor[8]; |
||
|
char prefix[155]; |
||
|
char pad[12]; |
};
typeflag
Type of entry. POSIX extended the earlier linkflag field with several new type values:
’’0’’
Regular file. NUL should be treated as a synonym, for compatibility purposes.
’’1’’
Hard link.
’’2’’
Symbolic link.
’’3’’
Character device node.
’’4’’
Block device node.
’’5’’
Directory.
’’6’’
FIFO node.
’’7’’
Reserved.
Other
A POSIX-compliant implementation must treat any unrecognized typeflag value as a regular file. In particular, writers should ensure that all entries have a valid filename so that they can be restored by readers that do not support the corresponding extension. Uppercase letters "A" through "Z" are reserved for custom extensions. Note that sockets and whiteout entries are not archivable.
It is worth noting that the size field, in particular, has different meanings depending on the type. For regular files, of course, it indicates the amount of data following the header. For directories, it may be used to indicate the total size of all files in the directory, for use by operating systems that pre-allocate directory space. For all other types, it should be set to zero by writers and ignored by readers.
magic
Contains the magic value ’’ustar’’ followed by a NUL byte to indicate that this is a POSIX standard archive. Full compliance requires the uname and gname fields be properly set.
version
Version. This should be ’’00’’ (two copies of the ASCII digit zero) for POSIX standard archives.
uname, gname
User and group names, as null-terminated ASCII strings. These should be used in preference to the uid/gid values when they are set and the corresponding names exist on the system.
devmajor, devminor
Major and minor numbers for character device or block device entry.
name, prefix
If the pathname is too long to fit in the 100 bytes provided by the standard format, it can be split at any / character with the first portion going into the prefix field. If the prefix field is not empty, the reader will prepend the prefix value and a / character to the regular name field to obtain the full pathname. The standard does not require a trailing / character on directory names, though most implementations still include this for compatibility reasons.
Note that all unused bytes must be set to NUL.
Field termination is specified slightly differently by POSIX than by previous implementations. The magic, uname, and gname fields must have a trailing NUL. The pathname, linkname, and prefix fields must have a trailing NUL unless they fill the entire field. (In particular, it is possible to store a 256-character pathname if it happens to have a / as the 156th character.) POSIX requires numeric fields to be zero-padded in the front, and requires them to be terminated with either space or NUL characters.
Currently, most tar implementations comply with the ustar format, occasionally extending it by adding new fields to the blank area at the end of the header record.
Numeric
Extensions
There have been several attempts to extend the range of
sizes or times supported by modifying how numbers are stored
in the header.
One obvious extension to increase the size of files is to eliminate the terminating characters from the various numeric fields. For example, the standard only allows the size field to contain 11 octal digits, reserving the twelfth byte for a trailing NUL character. Allowing 12 octal digits allows file sizes up to 64 GB.
Another extension, utilized by GNU tar, star, and other newer tar implementations, permits binary numbers in the standard numeric fields. This is flagged by setting the high bit of the first byte. The remainder of the field is treated as a signed twos-complement value. This permits 95-bit values for the length and time fields and 63-bit values for the uid, gid, and device numbers. In particular, this provides a consistent way to handle negative time values. GNU tar supports this extension for the length, mtime, ctime, and atime fields. Joerg Schilling’s star program and the libarchive library support this extension for all numeric fields. Note that this extension is largely obsoleted by the extended attribute record provided by the pax interchange format.
Another early GNU extension allowed base-64 values rather than octal. This extension was short-lived and is no longer supported by any implementation.
Pax
Interchange Format
There are many attributes that cannot be portably stored in
a POSIX ustar archive. IEEE Std 1003.1-2001
(’’POSIX.1’’) defined a
’’pax interchange format’’ that uses
two new types of entries to hold text-formatted metadata
that applies to following entries. Note that a pax
interchange format archive is a ustar archive in every
respect. The new data is stored in ustar-compatible archive
entries that use the ’’x’’ or
’’g’’ typeflag. In particular, older
implementations that do not fully support these extensions
will extract the metadata into regular files, where the
metadata can be examined as necessary.
An entry in a pax interchange format archive consists of one or two standard ustar entries, each with its own header and data. The first optional entry stores the extended attributes for the following entry. This optional first entry has an "x" typeflag and a size field that indicates the total size of the extended attributes. The extended attributes themselves are stored as a series of text-format lines encoded in the portable UTF-8 encoding. Each line consists of a decimal number, a space, a key string, an equals sign, a value string, and a new line. The decimal number indicates the length of the entire line, including the initial length field and the trailing newline. An example of such a field is:
25 ctime=1084839148.1212\n
Keys in all lowercase are standard keys. Vendors can add their own keys by prefixing them with an all uppercase vendor name and a period. Note that, unlike the historic header, numeric values are stored using decimal, not octal. A description of some common keys follows:
atime, ctime, mtime
File access, inode change, and modification times. These fields can be negative or include a decimal point and a fractional value.
hdrcharset
The character set used by the pax extension values. By default, all textual values in the pax extended attributes are assumed to be in UTF-8, including pathnames, user names, and group names. In some cases, it is not possible to translate local conventions into UTF-8. If this key is present and the value is the six-character ASCII string ’’BINARY’’, then all textual values are assumed to be in a platform-dependent multi-byte encoding. Note that there are only two valid values for this key: ’’BINARY’’ or ’’ISO-IR 10646 2000 UTF-8’’. No other values are permitted by the standard, and the latter value should generally not be used as it is the default when this key is not specified. In particular, this flag should not be used as a general mechanism to allow filenames to be stored in arbitrary encodings.
uname, uid, gname, gid
User name, group name, and numeric UID and GID values. The user name and group name stored here are encoded in UTF8 and can thus include non-ASCII characters. The UID and GID fields can be of arbitrary length.
linkpath
The full path of the linked-to file. Note that this is encoded in UTF8 and can thus include non-ASCII characters.
path
The full pathname of the entry. Note that this is encoded in UTF8 and can thus include non-ASCII characters.
realtime.*, security.*
These keys are reserved and may be used for future standardization.
size
The size of the file. Note that there is no length limit on this field, allowing conforming archives to store files much larger than the historic 8GB limit.
SCHILY.*
Vendor-specific attributes used by Joerg Schilling’s star implementation.
SCHILY.acl.access, SCHILY.acl.default
Stores the access and default ACLs as textual strings in a format that is an extension of the format specified by POSIX.1e draft 17. In particular, each user or group access specification can include a fourth colon-separated field with the numeric UID or GID. This allows ACLs to be restored on systems that may not have complete user or group information available (such as when NIS/YP or LDAP services are temporarily unavailable).
SCHILY.devminor, SCHILY.devmajor
The full minor and major numbers for device nodes.
SCHILY.fflags
The file flags.
SCHILY.realsize
The full size of the file on disk. XXX explain? XXX
SCHILY.dev, SCHILY.ino, SCHILY.nlinks
The device number, inode number, and link count for the entry. In particular, note that a pax interchange format archive using Joerg Schilling’s SCHILY.* extensions can store all of the data from struct stat.
LIBARCHIVE.*
Vendor-specific attributes used by the libarchive library and programs that use it.
LIBARCHIVE.creationtime
The time when the file was created. (This should not be confused with the POSIX ’’ctime’’ attribute, which refers to the time when the file metadata was last changed.)
LIBARCHIVE.xattr.namespace.key
Libarchive stores POSIX.1e-style extended attributes using keys of this form. The key value is URL-encoded: All non-ASCII characters and the two special characters ’’=’’ and ’’%’’ are encoded as ’’%’’ followed by two uppercase hexadecimal digits. The value of this key is the extended attribute value encoded in base 64. XXX Detail the base-64 format here XXX
VENDOR.*
XXX document other vendor-specific extensions XXX
Any values stored in an extended attribute override the corresponding values in the regular tar header. Note that compliant readers should ignore the regular fields when they are overridden. This is important, as existing archivers are known to store non-compliant values in the standard header fields in this situation. There are no limits on length for any of these fields. In particular, numeric fields can be arbitrarily large. All text fields are encoded in UTF8. Compliant writers should store only portable 7-bit ASCII characters in the standard ustar header and use extended attributes whenever a text value contains non-ASCII characters.
In addition to the x entry described above, the pax interchange format also supports a g entry. The g entry is identical in format, but specifies attributes that serve as defaults for all subsequent archive entries. The g entry is not widely used.
Besides the new x and g entries, the pax interchange format has a few other minor variations from the earlier ustar format. The most troubling one is that hardlinks are permitted to have data following them. This allows readers to restore any hardlink to a file without having to rewind the archive to find an earlier entry. However, it creates complications for robust readers, as it is no longer clear whether or not they should ignore the size field for hardlink entries.
GNU Tar
Archives
The GNU tar program started with a pre-POSIX format similar
to that described earlier and has extended it using several
different mechanisms: It added new fields to the empty space
in the header (some of which was later used by POSIX for
conflicting purposes); it allowed the header to be continued
over multiple records; and it defined new entries that
modify following entries (similar in principle to the
x entry described above, but each GNU special entry
is single-purpose, unlike the general-purpose x
entry). As a result, GNU tar archives are not POSIX
compatible, although more lenient POSIX-compliant readers
can successfully extract most GNU tar archives.
struct header_gnu_tar {
|
char name[100]; |
|||
|
char mode[8]; |
|||
|
char uid[8]; |
|||
|
char gid[8]; |
|||
|
char size[12]; |
|||
|
char mtime[12]; |
|||
|
char checksum[8]; |
|||
|
char typeflag[1]; |
|||
|
char linkname[100]; |
|||
|
char magic[6]; |
|||
|
char version[2]; |
|||
|
char uname[32]; |
|||
|
char gname[32]; |
|||
|
char devmajor[8]; |
|||
|
char devminor[8]; |
|||
|
char atime[12]; |
|||
|
char ctime[12]; |
|||
|
char offset[12]; |
|||
|
char longnames[4]; |
|||
|
char unused[1]; |
|||
|
struct { |
|||
|
char offset[12]; |
|||
|
char numbytes[12]; |
|||
|
} sparse[4]; |
|||
|
char isextended[1]; |
|||
|
char realsize[12]; |
|||
|
char pad[17]; |
};
typeflag
GNU tar uses the following special entry types, in addition to those defined by POSIX:
7
GNU tar treats type "7" records identically to type "0" records, except on one obscure RTOS where they are used to indicate the pre-allocation of a contiguous file on disk.
D
This indicates a directory entry. Unlike the POSIX-standard "5" typeflag, the header is followed by data records listing the names of files in this directory. Each name is preceded by an ASCII "Y" if the file is stored in this archive or "N" if the file is not stored in this archive. Each name is terminated with a null, and an extra null marks the end of the name list. The purpose of this entry is to support incremental backups; a program restoring from such an archive may wish to delete files on disk that did not exist in the directory when the archive was made.
Note that the "D" typeflag specifically violates POSIX, which requires that unrecognized typeflags be restored as normal files. In this case, restoring the "D" entry as a file could interfere with subsequent creation of the like-named directory.
K
The data for this entry is a long linkname for the following regular entry.
L
The data for this entry is a long pathname for the following regular entry.
M
This is a continuation of the last file on the previous volume. GNU multi-volume archives guarantee that each volume begins with a valid entry header. To ensure this, a file may be split, with part stored at the end of one volume, and part stored at the beginning of the next volume. The "M" typeflag indicates that this entry continues an existing file. Such entries can only occur as the first or second entry in an archive (the latter only if the first entry is a volume label). The size field specifies the size of this entry. The offset field at bytes 369-380 specifies the offset where this file fragment begins. The realsize field specifies the total size of the file (which must equal size plus offset). When extracting, GNU tar checks that the header file name is the one it is expecting, that the header offset is in the correct sequence, and that the sum of offset and size is equal to realsize.
N
Type "N" records are no longer generated by GNU tar. They contained a list of files to be renamed or symlinked after extraction; this was originally used to support long names. The contents of this record are a text description of the operations to be done, in the form ’’Rename %s to %s\n’’ or ’’Symlink %s to %s\n’’; in either case, both filenames are escaped using K&R C syntax. Due to security concerns, "N" records are now generally ignored when reading archives.
S
This is a ’’sparse’’ regular file. Sparse files are stored as a series of fragments. The header contains a list of fragment offset/length pairs. If more than four such entries are required, the header is extended as necessary with ’’extra’’ header extensions (an older format that is no longer used), or ’’sparse’’ extensions.
V
The name field should be interpreted as a tape/volume header name. This entry should generally be ignored on extraction.
magic
The magic field holds the five characters ’’ustar’’ followed by a space. Note that POSIX ustar archives have a trailing null.
version
The version field holds a space character followed by a null. Note that POSIX ustar archives use two copies of the ASCII digit ’’0’’.
atime, ctime
The time the file was last accessed and the time of last change of file information, stored in octal as with mtime.
longnames
This field is apparently no longer used.
Sparse offset / numbytes
Each such structure specifies a single fragment of a sparse file. The two fields store values as octal numbers. The fragments are each padded to a multiple of 512 bytes in the archive. On extraction, the list of fragments is collected from the header (including any extension headers), and the data is then read and written to the file at appropriate offsets.
isextended
If this is set to non-zero, the header will be followed by additional ’’sparse header’’ records. Each such record contains information about as many as 21 additional sparse blocks as shown here:
struct gnu_sparse_header {
|
struct { |
|||
|
char offset[12]; |
|||
|
char numbytes[12]; |
|||
|
} sparse[21]; |
|||
|
char isextended[1]; |
|||
|
char padding[7]; |
};
realsize
A binary representation of the file’s complete size, with a much larger range than the POSIX file size. In particular, with M type files, the current entry is only a portion of the file. In that case, the POSIX size field will indicate the size of this entry; the realsize field will indicate the total size of the file.
GNU tar pax
archives
GNU tar 1.14 (XXX check this XXX) and later will write pax
interchange format archives when you specify the
−-posix flag. This format follows the pax
interchange format closely, using some SCHILY tags
and introducing new keywords to store sparse file
information. There have been three iterations of the sparse
file support, referred to as
’’0.0’’,
’’0.1’’, and
’’1.0’’.
GNU.sparse.numblocks, GNU.sparse.offset, GNU.sparse.numbytes, GNU.sparse.size
The ’’0.0’’ format used an initial GNU.sparse.numblocks attribute to indicate the number of blocks in the file, a pair of GNU.sparse.offset and GNU.sparse.numbytes to indicate the offset and size of each block, and a single GNU.sparse.size to indicate the full size of the file. This is not the same as the size in the tar header because the latter value does not include the size of any holes. This format required that the order of attributes be preserved and relied on readers accepting multiple appearances of the same attribute names, which is not officially permitted by the standards.
GNU.sparse.map
The ’’0.1’’ format used a single attribute that stored a comma-separated list of decimal numbers. Each pair of numbers indicated the offset and size, respectively, of a block of data. This does not work well if the archive is extracted by an archiver that does not recognize this extension, since many pax implementations simply discard unrecognized attributes.
GNU.sparse.major, GNU.sparse.minor, GNU.sparse.name, GNU.sparse.realsize
The ’’1.0’’ format stores the sparse block map in one or more 512-byte blocks prepended to the file data in the entry body. The pax attributes indicate the existence of this map (via the GNU.sparse.major and GNU.sparse.minor fields) and the full size of the file. The GNU.sparse.name holds the true name of the file. To avoid confusion, the name stored in the regular tar header is a modified name so that extraction errors will be apparent to users.
Solaris
Tar
XXX More Details Needed XXX
Solaris tar (beginning with SunOS XXX 5.7 ?? XXX) supports an ’’extended’’ format that is fundamentally similar to pax interchange format, with the following differences:
•
Extended attributes are stored in an entry whose type is X, not x, as used by pax interchange format. The detailed format of this entry appears to be the same as detailed above for the x entry.
•
An additional A header is used to store an ACL for the following regular entry. The body of this entry contains a seven-digit octal number followed by a zero byte, followed by the textual ACL description. The octal value is the number of ACL entries plus a constant that indicates the ACL type: 01000000 for POSIX.1e ACLs and 03000000 for NFSv4 ACLs.
AIX Tar
XXX More details needed XXX
AIX Tar uses a ustar-formatted header with the type A for storing coded ACL information. Unlike the Solaris format, AIX tar writes this header after the regular file body to which it applies. The pathname in this header is either NFS4 or AIXC to indicate the type of ACL stored. The actual ACL is stored in platform-specific binary format.
Mac OS X
Tar
The tar distributed with Apple’s Mac OS X stores most
regular files as two separate files in the tar archive. The
two files have the same name except that the first one has
’’._’’ prepended to the last path
element. This special file stores an AppleDouble-encoded
binary blob with additional metadata about the second file,
including ACL, extended attributes, and resources. To
recreate the original file on disk, each separate file can
be extracted and the Mac OS X copyfile() function can
be used to unpack the separate metadata file and apply it to
th regular file. Conversely, the same function provides a
’’pack’’ option to encode the
extended metadata from a file into a separate file whose
contents can then be put into a tar archive.
Note that the Apple extended attributes interact badly with long filenames. Since each file is stored with the full name, a separate set of extensions needs to be included in the archive for each one, doubling the overhead required for files with long names.
Summary of
tar type codes
The following list is a condensed summary of the type codes
used in tar header records generated by different tar
implementations. More details about specific implementations
can be found above:
NUL
Early tar programs stored a zero byte for regular files.
0
POSIX standard type code for a regular file.
1
POSIX standard type code for a hard link description.
2
POSIX standard type code for a symbolic link description.
3
POSIX standard type code for a character device node.
4
POSIX standard type code for a block device node.
5
POSIX standard type code for a directory.
6
POSIX standard type code for a FIFO.
7
POSIX reserved.
7
GNU tar used for pre-allocated files on some systems.
A
Solaris tar ACL description stored prior to a regular file header.
A
AIX tar ACL description stored after the file body.
D
GNU tar directory dump.
K
GNU tar long linkname for the following header.
L
GNU tar long pathname for the following header.
M
GNU tar multivolume marker, indicating the file is a continuation of a file from the previous volume.
N
GNU tar long filename support. Deprecated.
S
GNU tar sparse regular file.
V
GNU tar tape/volume header name.
X
Solaris tar general-purpose extension header.
g
POSIX pax interchange format global extensions.
x
POSIX pax interchange format per-file extensions.
SEE ALSO
ar(1), pax(1), tar(1)
STANDARDS
The tar utility is no longer a part of POSIX or the Single Unix Standard. It last appeared in Version 2 of the Single UNIX Specification (’’SUSv2’’). It has been supplanted in subsequent standards by pax(1). The ustar format is currently part of the specification for the pax(1) utility. The pax interchange file format is new with IEEE Std 1003.1-2001 (’’POSIX.1’’).
HISTORY
A tar command appeared in Seventh Edition Unix, which was released in January, 1979. It replaced the tp program from Fourth Edition Unix which in turn replaced the tap program from First Edition Unix. John Gilmore’s pdtar public-domain implementation (circa 1987) was highly influential and formed the basis of GNU tar (circa 1988). Joerg Shilling’s star archiver is another open-source (CDDL) archiver (originally developed circa 1985) which features complete support for pax interchange format.
This documentation was written as part of the libarchive and bsdtar project by Tim Kientzle <kientzle@@FreeBSD.org>.
BSD December 23, 2011 BSD
‘‘0’’
d441 1 a441 1‘‘1’’
d445 1 a445 1‘‘2’’
d450 1 a450 1‘‘3’’
d455 1 a455 1‘‘4’’
d460 1 a460 1‘‘5’’
d464 1 a464 1‘‘6’’
d468 1 a468 1‘‘7’’
d496 1 a496 1 magic value ‘‘ustar’’ followed by a d504 1 a504 1 ‘‘00’’ (two copies of the ASCII d592 2 a593 2 (‘‘POSIX.1’’) defined a ‘‘pax interchange format’’ that uses d598 2 a599 2 entries that use the ‘‘x’’ or ‘‘g’’ typeflag. In particular, older d644 1 a644 1 ‘‘BINARY’’, then all textual values d647 2 a648 2 key: ‘‘BINARY’’ or ‘‘ISO-IR 10646 2000 UTF-8’’. d695 1 a695 1 SCHILY.acl.default, SCHILY.acl.ace d697 8 a704 9Stores the access, default and NFSv4 ACLs as textual strings in a format that is an extension of the format specified by POSIX.1e draft 17. In particular, each user or group access specification can include an additional colon-separated field with the numeric UID or GID. This allows ACLs to be restored on systems that may not have complete user or group information available (such as when NIS/YP or LDAP services are temporarily unavailable).
d741 1 a741 1 ‘‘ctime’’ attribute, which refers to d751 3 a753 3 ‘‘=’’ and ‘‘%’’ are encoded as ‘‘%’’ followed by two uppercase d1139 2 a1140 2 the operations to be done, in the form ‘‘Rename %s to %s\n’’ or ‘‘Symlink %s to d1149 1 a1149 1 ‘‘sparse’’ regular file. Sparse d1153 1 a1153 1 as necessary with ‘‘extra’’ header d1155 1 a1155 1 ‘‘sparse’’ extensions. d1166 1 a1166 1 holds the five characters ‘‘ustar’’ d1175 1 a1175 1 ‘‘0’’. d1202 1 a1202 1 header will be followed by additional ‘‘sparse d1291 3 a1293 3 ‘‘0.0’’, ‘‘0.1’’, and ‘‘1.0’’. d1300 1 a1300 1 ‘‘0.0’’ format used an initial d1315 1 a1315 1 ‘‘0.1’’ format used a single d1328 1 a1328 1 ‘‘1.0’’ format stores the sparse d1344 1 a1344 1 ‘‘extended’’ format that is d1383 1 a1383 1 ‘‘._’’ prepended to the last path d1391 1 a1391 1 ‘‘pack’’ option to encode the d1529 1 a1529 1 (‘‘SUSv2’’). It has been supplanted d1533 1 a1533 1 1003.1-2001 (‘‘POSIX.1’’). d1555 1 a1555 1 December 27, 2016 BSD @ 1.1.1.7 log @Import libarchive-3.3.2 + 9de5f3 + f9dacbf: - Support NFS4 ACLs on Linux - Bugfixes @ text @d2 1 a2 1 d53 1 a53 1 ’’blocks’’ are always a multiple of d59 2 a60 2 ’’block’’ and ’’record’’ here are not entirely d227 1 a227 1 an ASCII ’1’ and the linkname field holds d242 1 a242 1 (’’POSIX.1’’) standard was released. d249 1 a249 1 (’’POSIX.1’’) served as the basis d258 1 a258 1 five characters ’’ustar’’ followed d276 1 a276 1 IEEE Std 1003.1-1988 (’’POSIX.1’’) d279 1 a279 1 called the ’’ustar’’ format, after d281 1 a281 1 for ’’Unix Standard TAR’’.) It d435 1 a435 1’’0’’
d441 1 a441 1’’1’’
d445 1 a445 1’’2’’
d450 1 a450 1’’3’’
d455 1 a455 1’’4’’
d460 1 a460 1’’5’’
d464 1 a464 1’’6’’
d468 1 a468 1’’7’’
d496 1 a496 1 magic value ’’ustar’’ followed by a d504 1 a504 1 ’’00’’ (two copies of the ASCII d592 2 a593 2 (’’POSIX.1’’) defined a ’’pax interchange format’’ that uses d598 2 a599 2 entries that use the ’’x’’ or ’’g’’ typeflag. In particular, older d644 1 a644 1 ’’BINARY’’, then all textual values d647 2 a648 2 key: ’’BINARY’’ or ’’ISO-IR 10646 2000 UTF-8’’. d742 1 a742 1 ’’ctime’’ attribute, which refers to d752 3 a754 3 ’’=’’ and ’’%’’ are encoded as ’’%’’ followed by two uppercase d1140 2 a1141 2 the operations to be done, in the form ’’Rename %s to %s\n’’ or ’’Symlink %s to d1150 1 a1150 1 ’’sparse’’ regular file. Sparse d1154 1 a1154 1 as necessary with ’’extra’’ header d1156 1 a1156 1 ’’sparse’’ extensions. d1167 1 a1167 1 holds the five characters ’’ustar’’ d1176 1 a1176 1 ’’0’’. d1203 1 a1203 1 header will be followed by additional ’’sparse d1287 8 a1294 7 --posix flag. This format follows the pax interchange format closely, using some SCHILY tags and introducing new keywords to store sparse file information. There have been three iterations of the sparse file support, referred to as ’’0.0’’, ’’0.1’’, and ’’1.0’’. d1301 1 a1301 1 ’’0.0’’ format used an initial d1316 1 a1316 1 ’’0.1’’ format used a single d1329 1 a1329 1 ’’1.0’’ format stores the sparse d1345 1 a1345 1 ’’extended’’ format that is d1384 1 a1384 1 ’’._’’ prepended to the last path d1392 1 a1392 1 ’’pack’’ option to encode the d1530 1 a1530 1 (’’SUSv2’’). It has been supplanted d1534 1 a1534 1 1003.1-2001 (’’POSIX.1’’). @ 1.1.1.8 log @Import libarchive-3.3.3 as should have done originally. @ text @d2 1 a2 1 @ 1.1.1.9 log @Import libarchive 3.4.0 @ text @d1 2 a2 2 d53 2 a54 2 “blocks” are always a multiple of the record size. The maximum block size supported by early d58 5 a62 4 high-speed tape drives. (Note: the terms “block” and “record” here are not entirely standard; this document follows the convention established by John Gilmore in documenting pdtar.) d242 3 a244 3 (“POSIX.1”) standard was released. For best portability, modern implementations should fill the numeric fields with leading zeros. d249 5 a253 5 (“POSIX.1”) served as the basis for John Gilmore’s pdtar program and many system implementations from the late 1980s and early 1990s. These archives generally follow the POSIX ustar format described below with the following variations: d258 3 a260 3 five characters “ustar” followed by a space. The version field contains a space character followed by a null. d276 7 a282 7 IEEE Std 1003.1-1988 (“POSIX.1”) defined a standard tar file format to be read and written by compliant implementations of tar(1). This format is often called the “ustar” format, after the magic value used in the header. (The name is an acronym for “Unix Standard TAR”.) It extends the historic format with new fields: d435 1 a435 1“0”
d441 1 a441 1“1”
d445 1 a445 1“2”
d450 1 a450 1“3”
d455 1 a455 1“4”
d460 1 a460 1“5”
d464 1 a464 1“6”
d468 1 a468 1“7”
d496 4 a499 4 magic value “ustar” followed by a NUL byte to indicate that this is a POSIX standard archive. Full compliance requires the uname and gname fields be properly set. d504 2 a505 2 “00” (two copies of the ASCII digit zero) for POSIX standard archives. d592 11 a602 10 (“POSIX.1”) defined a “pax interchange format” that uses two new types of entries to hold text-formatted metadata that applies to following entries. Note that a pax interchange format archive is a ustar archive in every respect. The new data is stored in ustar-compatible archive entries that use the “x” or “g” typeflag. In particular, older implementations that do not fully support these extensions will extract the metadata into regular files, where the metadata can be examined as necessary. d643 11 a653 10 six-character ASCII string “BINARY”, then all textual values are assumed to be in a platform-dependent multi-byte encoding. Note that there are only two valid values for this key: “BINARY” or “ISO-IR 10646 2000 UTF-8”. No other values are permitted by the standard, and the latter value should generally not be used as it is the default when this key is not specified. In particular, this flag should not be used as a general mechanism to allow filenames to be stored in arbitrary encodings. d742 2 a743 2 “ctime” attribute, which refers to the time when the file metadata was last changed.) d751 7 a757 5 characters and the two special characters “=” and “%” are encoded as “%” followed by two uppercase hexadecimal digits. The value of this key is the extended attribute value encoded in base 64. XXX Detail the base-64 format here XXX d1140 6 a1145 5 the operations to be done, in the form “Rename %s to %s\n” or “Symlink %s to %s\n”; in either case, both filenames are escaped using K&R C syntax. Due to security concerns, "N" records are now generally ignored when reading archives. d1150 7 a1156 6 “sparse” regular file. Sparse files are stored as a series of fragments. The header contains a list of fragment offset/length pairs. If more than four such entries are required, the header is extended as necessary with “extra” header extensions (an older format that is no longer used), or “sparse” extensions. d1167 3 a1169 3 holds the five characters “ustar” followed by a space. Note that POSIX ustar archives have a trailing null. d1175 2 a1176 1 use two copies of the ASCII digit “0”. d1203 4 a1206 4 header will be followed by additional “sparse header” records. Each such record contains information about as many as 21 additional sparse blocks as shown here: d1291 3 a1293 2 referred to as “0.0”, “0.1”, and “1.0”. d1299 12 a1310 12The “0.0” format used an initial GNU.sparse.numblocks attribute to indicate the number of blocks in the file, a pair of GNU.sparse.offset and GNU.sparse.numbytes to indicate the offset and size of each block, and a single GNU.sparse.size to indicate the full size of the file. This is not the same as the size in the tar header because the latter value does not include the size of any holes. This format required that the order of attributes be preserved and relied on readers accepting multiple appearances of the same attribute names, which is not officially permitted by the standards.
d1314 8 a1321 7The “0.1” format used a single attribute that stored a comma-separated list of decimal numbers. Each pair of numbers indicated the offset and size, respectively, of a block of data. This does not work well if the archive is extracted by an archiver that does not recognize this extension, since many pax implementations simply discard unrecognized attributes.
d1327 10 a1336 9The “1.0” format stores the sparse block map in one or more 512-byte blocks prepended to the file data in the entry body. The pax attributes indicate the existence of this map (via the GNU.sparse.major and GNU.sparse.minor fields) and the full size of the file. The GNU.sparse.name holds the true name of the file. To avoid confusion, the name stored in the regular tar header is a modified name so that extraction errors will be apparent to users.
d1344 3 a1346 3 “extended” format that is fundamentally similar to pax interchange format, with the following differences: d1383 11 a1393 11 “._” prepended to the last path element. This special file stores an AppleDouble-encoded binary blob with additional metadata about the second file, including ACL, extended attributes, and resources. To recreate the original file on disk, each separate file can be extracted and the Mac OS X copyfile() function can be used to unpack the separate metadata file and apply it to th regular file. Conversely, the same function provides a “pack” option to encode the extended metadata from a file into a separate file whose contents can then be put into a tar archive. d1529 5 a1533 5 (“SUSv2”). It has been supplanted in subsequent standards by pax(1). The ustar format is currently part of the specification for the pax(1) utility. The pax interchange file format is new with IEEE Std 1003.1-2001 (“POSIX.1”). @ 1.1.1.10 log @Import libarchive 3.7.2 @ text @d2 1 a2 1 d692 1 a692 1 SCHILY.acl.default, SCHILY.acl.ace d719 2 a720 2SCHILY.dev, SCHILY.ino, SCHILY.nlinks
d743 1 a743 1LIBARCHIVE.xattr.namespace.key
@ 1.1.1.11 log @libarchive: import version 3.7.3 @ text @d2 1 a2 1 @ 1.1.1.12 log @Libarchive 3.7.4 is a bugfix and security release Security fixes: rar: Fix OOB in rar e8 filter (CVE-2024-26256) zip: Fix out of boundary access Important bugfixes: 7zip: Limit amount of properties bsdtar: Fix error handling around strtol() usages passphrase: Improve newline handling on Windows passphrase: Never allow empty passwords rar: Fix "File CRC Error" when extracting specific rar4 archives xar: Avoid infinite link loop zip: Update AppleDouble support for directories zstd: Implement core detection @ text @d2 1 a2 1 @ 1.1.1.13 log @libarchive: imported version 3.7.5 Libarchive 3.7.5 Security fixes: fix multiple vulnerabilities identified by SAST cpio: ignore out-of-range gid/uid/size/ino and harden AFIO parsing lzop: prevent integer overflow rar4: protect copy_from_lzss_window_to_unp() rar4: fix CVE-2024-26256 rar4: fix OOB in delta and audio filter rar4: fix out of boundary access with large files rar4: add boundary checks to rgb filter rar4: fix OOB access with unicode filenames rar5: clear 'data ready' cache on window buffer reallocs rpm: calculate huge header sizes correctly unzip: unify EOF handling util: fix out of boundary access in mktemp functions uu: stop processing if lines are too long Important bugfixes: 7zip: fix issue when skipping first file in 7zip archive that is a multiple of 65536 bytes ar: fix archive entries having no type lha: do not allow negative file sizes lha: fix integer truncation on 32-bit systems shar: check strdup return value rar5: don't try to read rediculously long names xar: fix another infinite loop and expat error handling many Windows fixes, cleanups and improvements @ text @d2 1 a2 1 @ 1.1.1.14 log @libarchove: import version 3.7.7 @ text @d1 2 a2 2 d23 1 a23 1TAR(5) File Formats Manual TAR(5)
d27 2 a28 2tar — format of tape archive files
d32 1 a32 1The tar archive format d40 8 a47 2
General Format
d49 1 a49 9A tar archive consists of a series of 512-byte records. Each file system object requires a header record which stores basic metadata (pathname, owner, permissions, etc.) and zero or more records containing any file data. The end of the archive is indicated by two records consisting entirely of zero bytes.
For d63 8 a70 9
Old-Style Archive Format
The original tar archive format has been extended many times to include additional information that various implementors found necessary. This section describes the variant implemented by the tar command included in Version 7 AT&T UNIX, which seems to be the earliest widely-used version of the tar program.
d72 1 a72 1The header d76 1 a76 1
struct d82 2 a83 2
};
d165 1 a165 1All unused bytes in the header d170 1 a170 1
Pathname, d180 1 a180 1
File mode, d185 1 a185 1
User id and group id of owner, d190 1 a190 1
Size of file, d199 1 a199 1
Modification d207 1 a207 1
Header checksum, stored as an d222 1 a222 1
In order to preserve hardlinks d226 1 a226 1 an ASCII ‘1’ and the linkname field holds d231 1 a231 1
Early tar d245 5 a249 6
Pre-POSIX Archives
An early draft of IEEE Std 1003.1-1988 (“POSIX.1”) served as the basis for John Gilmore’s pdtar program and many system d256 1 a256 1
The magic value consists of the d263 1 a263 1
The numeric fields are d269 1 a269 1
The prefix field is often not d273 5 a277 7
POSIX ustar Archives
IEEE Std 1003.1-1988 (“POSIX.1”) defined a standard tar file format to be read and written by compliant implementations of tar(1). This format is often called the d283 1 a283 1
struct d289 2 a290 2
};
d430 1 a430 1Type of entry. POSIX extended d436 1 a436 1
Regular file. d442 1 a442 1
Hard link.
d446 1 a446 1Symbolic d451 1 a451 1
Character d456 1 a456 1
Block device d461 1 a461 1
Directory.
d465 1 a465 1FIFO node.
d469 1 a469 1Reserved.
d473 1 a473 1A d483 1 a483 1
It is worth noting that the d494 1 a494 1
Contains the d502 1 a502 1
Version. This should be d508 1 a508 1
User and group names, as d516 1 a516 1
Major and minor numbers for d521 1 a521 1
If the pathname is too long to d532 1 a532 1
Note that all d535 1 a535 1
Field d547 1 a547 1
Currently, most d552 5 a556 6
Numeric Extensions
There have been several attempts to extend the range of sizes or times supported by modifying how numbers are stored in the header.
d558 1 a558 1One obvious d566 1 a566 1
Another d582 1 a582 1
Another early d587 9 a595 10
Pax Interchange Format
There are many attributes that cannot be portably stored in a POSIX ustar archive. IEEE Std 1003.1-2001 (“POSIX.1”) defined a “pax interchange format” that uses two new types of entries to hold text-formatted metadata that applies to following entries. Note that a pax interchange format archive is a ustar archive in every respect. The new data is stored in d602 1 a602 1
An entry in a d617 1 a617 2
25 ctime=1084839148.1212\n
d619 1 a619 1Keys in all lowercase are d629 1 a629 1
File access, inode change, and d635 1 a635 1
The character set used by the d655 1 a655 1
User name, group name, and d663 1 a663 1
The full path of the linked-to d669 1 a669 1
The full d676 1 a676 1
These keys are reserved and may d681 1 a681 1
The size of the d688 1 a688 1
Vendor-specific attributes used d694 1 a694 1
Stores the access, default and d707 1 a707 1
The full minor and major d712 1 a712 1
The file flags.
d716 1 a716 1The full size of the file on d722 1 a722 1
The device number, inode d730 1 a730 1
Vendor-specific attributes used d737 1 a737 1
The time when the file was d745 1 a745 1
Libarchive stores d756 1 a756 1
XXX document other d759 1 a759 1
Any values d773 1 a773 1
In addition to d780 1 a780 1
Besides the new d791 14 a804 16
GNU Tar Archives
The GNU tar program started with a pre-POSIX format similar to that described earlier and has extended it using several different mechanisms: It added new fields to the empty space in the header (some of which was later used by POSIX for conflicting purposes); it allowed the header to be continued over multiple records; and it defined new entries that modify following entries (similar in principle to the x entry described above, but each GNU special entry is single-purpose, unlike the general-purpose x entry). As a result, GNU tar archives are not POSIX compatible, although more lenient POSIX-compliant readers can successfully extract most GNU tar archives.
d806 1 a806 1struct d812 2 a813 2
};
d1062 1 a1062 1GNU tar uses the following d1068 1 a1068 1
GNU tar treats d1076 1 a1076 1
This indicates d1088 1 a1088 1
Note that the d1097 1 a1097 1
The data for d1103 1 a1103 1
The data for d1109 1 a1109 1
This is a d1130 1 a1130 1
Type d1143 1 a1143 1
This is a d1153 1 a1153 1
The name d1159 1 a1159 1
The magic field d1166 1 a1166 1
The version field holds a space d1172 1 a1172 1
The time the file was last d1178 1 a1178 1
This field is apparently no d1184 1 a1184 1
Each such structure specifies a d1194 1 a1194 1
If this is set to non-zero, the d1200 1 a1200 1
struct d1206 1 a1206 1
};
d1267 1 a1267 1A binary representation of the d1275 10 a1284 11
GNU tar pax archives
GNU tar 1.14 (XXX check this XXX) and later will write pax interchange format archives when you specify the --posix flag. This format follows the pax interchange format closely, using some SCHILY tags and introducing new keywords to store sparse file information. There have been three iterations of the sparse file support, referred to as “0.0”, “0.1”, and “1.0”.
d1290 1 a1290 1The “0.0” format d1305 1 a1305 1
The “0.1” format d1317 1 a1317 1
The “1.0” format d1327 3 a1329 4
Solaris Tar
XXX More Details Needed XXX
d1331 1 a1331 1Solaris tar d1339 1 a1339 1
Extended attributes are stored d1347 1 a1347 1
An additional A header d1355 3 a1357 2
AIX Tar
d1359 1 a1359 3XXX More details needed XXX
AIX Tar uses a d1367 12 a1378 14
Mac OS X Tar
The tar distributed with Apple’s Mac OS X stores most regular files as two separate files in the tar archive. The two files have the same name except that the first one has “._” prepended to the last path element. This special file stores an AppleDouble-encoded binary blob with additional metadata about the second file, including ACL, extended attributes, and resources. To recreate the original file on disk, each separate file can be extracted and the Mac OS X copyfile() function can be used to unpack the separate metadata file and apply it to th regular file. d1384 1 a1384 1
Note that the d1391 6 a1396 8
Summary of tar type codes
The following list is a condensed summary of the type codes used in tar header records generated by different tar implementations. More details about specific implementations can be found above:
d1400 1 a1400 1Early tar d1405 1 a1405 1
POSIX standard d1410 1 a1410 1
POSIX standard d1415 1 a1415 1
POSIX standard d1420 1 a1420 1
POSIX standard d1425 1 a1425 1
POSIX standard d1430 1 a1430 1
POSIX standard d1435 1 a1435 1
POSIX standard d1440 1 a1440 1
POSIX d1445 1 a1445 1
GNU tar used d1450 1 a1450 1
Solaris tar ACL d1455 1 a1455 1
AIX tar ACL d1460 1 a1460 1
GNU tar d1465 1 a1465 1
GNU tar long d1470 1 a1470 1
GNU tar long d1475 1 a1475 1
GNU tar d1481 1 a1481 1
GNU tar long d1486 1 a1486 1
GNU tar sparse d1491 1 a1491 1
GNU tar d1496 1 a1496 1
Solaris tar d1501 1 a1501 1
POSIX pax d1506 1 a1506 1
POSIX pax d1511 1 a1511 2
ar(1), pax(1), tar(1)
d1515 1 a1515 1The tar utility is no d1519 3 a1521 3 standards by pax(1). The ustar format is currently part of the specification for the pax(1) utility. The pax interchange file format is new with IEEE Std 1003.1-2001 d1526 1 a1526 1
A tar command appeared in d1538 1 a1538 1
This d1541 4 a1544 2 <kientzle@@FreeBSD.org>. Debian December 27, 2016 TAR(5)
@ 1.1.1.15 log @libarchive: imported version 3.7.9 @ text @d2 1 a2 1 @ 1.1.1.16 log @libarchive: import version 3.8.0 Libarchive 3.8.0 is a feature and bugfix release. New features: bsdtar: support --mtime and --clamp-mtime lib: mbedtls 3.x compatibility 7-zip reader: improve self-extracting archive detection xar: xmllite support for the XAR reader and writer zip writer: added XZ, LZMA, ZSTD and BZIP2 support zip writer: added LZMA + RISCV BCJ filter Notable security fixes: rar: do not skip past EOF while reading rar: fix double free with over 4 billion nodes rar: fix heap-buffer-overflow warc: prevent signed integer overflow tar: fix overflow in build_ustar_entry Notable bugfixes: bsdtar: don't hardlink negative inode files together gz: allow setting the original filename for gzip compressed files lib: improve lseek handling lib: support @@-prefixed Unix epoch timestamps as date strings rar: support large headers on 32 bit systems tar reader: Improve LFS support on 32 bit systems @ text @d2 1 a2 1 @ 1.1.1.17 log @libarchive: import version 3.8.1 @ text @d2 1 a2 1 @ 1.1.1.18 log @libarchive: imported version 3.8.2 @ text @d2 1 a2 1 @ 1.1.1.19 log @libarchive: import version 3.8.3 @ text @d2 1 a2 1 @ 1.1.1.20 log @libarchive: import 3.8.4 @ text @d2 1 a2 1 @ 1.1.1.21 log @libarchive: import version 3.8.5 @ text @d2 1 a2 1 @ 1.1.1.22 log @libarchive: imported version 3.8.6 @ text @d2 1 a2 1 @ 1.1.1.23 log @libarchive: imported version 3.8.7 @ text @d2 1 a2 1 @