lzip/tarlz
1
0
Fork 0
Code Issues Activity

Adding upstream version 0.16.

Browse source 
Signed-off-by: Daniel Baumann <daniel@debian.org>
This commit is contained in:
Daniel Baumann 2025-02-17 21:14:09 +01:00
parent 5d67ab9e97
commit bb26c2917c
Signed by: daniel
GPG key ID: FBB4F0E80A80222F
20 changed files with 854 additions and 662 deletions
Download patch file Download diff file Expand all files Collapse all files

232
doc/tarlz.info
View file

@ -11,12 +11,13 @@ File: tarlz.info, Node: Top, Next: Introduction, Up: (dir)
Tarlz Manual
************
This manual is for Tarlz (version 0.15, 11 April 2019).
This manual is for Tarlz (version 0.16, 8 October 2019).
* Menu:
* Introduction:: Purpose and features of tarlz
* Invoking tarlz:: Command line interface
* Portable character set:: POSIX portable filename character set
* File format:: Detailed format of the compressed archive
* Amendments to pax format:: The reasons for the differences with pax
* Multi-threaded tar:: Limitations of parallel tar decoding
@ -39,13 +40,19 @@ File: tarlz.info, Node: Introduction, Next: Invoking tarlz, Prev: Top, Up: T
Tarlz is a massively parallel (multi-threaded) combined implementation
of the tar archiver and the lzip compressor. Tarlz creates, lists and
extracts archives in a simplified posix pax format compressed with
lzip, keeping the alignment between tar members and lzip members. This
method adds an indexed lzip layer on top of the tar archive, making it
possible to decode the archive safely in parallel. The resulting
multimember tar.lz archive is fully backward compatible with standard
tar tools like GNU tar, which treat it like any other tar.lz archive.
Tarlz can append files to the end of such compressed archives.
extracts archives in a simplified and safer variant of the POSIX pax
format compressed with lzip, keeping the alignment between tar members
and lzip members. The resulting multimember tar.lz archive is fully
backward compatible with standard tar tools like GNU tar, which treat
it like any other tar.lz archive. Tarlz can append files to the end of
such compressed archives.
Keeping the alignment between tar members and lzip members has two
advantages. It adds an indexed lzip layer on top of the tar archive,
making it possible to decode the archive safely in parallel. It also
minimizes the amount of data lost in case of corruption. Compressing a
tar archive with plzip may even double the amount of files lost for
each lzip member damaged because it does not keep the members aligned.
Tarlz can create tar archives with five levels of compression
granularity; per file (--no-solid), per block (--bsolid, default), per
@ -62,7 +69,7 @@ archive, but it has the following advantages:
member), and unwanted members can be deleted from the archive. Just
like an uncompressed tar archive.
* It is a safe posix-style backup format. In case of corruption,
* It is a safe POSIX-style backup format. In case of corruption,
tarlz can extract all the undamaged members from the tar.lz
archive, skipping over the damaged members, just like the standard
(uncompressed) tar. Moreover, the option '--keep-damaged' can be
@ -77,10 +84,11 @@ archive, but it has the following advantages:
with standard tar tools. *Note crc32::.
Tarlz does not understand other tar formats like 'gnu', 'oldgnu',
'star' or 'v7'.
'star' or 'v7'. 'tarlz -tf archive.tar.lz > /dev/null' can be used to
verify that the format of the archive is compatible with tarlz.

File: tarlz.info, Node: Invoking tarlz, Next: File format, Prev: Introduction, Up: Top
File: tarlz.info, Node: Invoking tarlz, Next: Portable character set, Prev: Introduction, Up: Top
2 Invoking tarlz
****************
@ -94,9 +102,9 @@ FILE is a directory.
On archive creation or appending tarlz archives the files specified,
but removes from member names any leading and trailing slashes and any
filename prefixes containing a '..' component. On extraction, leading
file name prefixes containing a '..' component. On extraction, leading
and trailing slashes are also removed from member names, and archive
members containing a '..' component in the filename are skipped. Tarlz
members containing a '..' component in the file name are skipped. Tarlz
detects when the archive being created or enlarged is among the files
to be dumped, appended or concatenated, and skips it.
@ -149,30 +157,31 @@ equivalent to '-1 --solid'
Change to directory DIR. When creating or appending, the position
of each '-C' option in the command line is significant; it will
change the current working directory for the following FILES until
a new '-C' option appears in the command line. When extracting, all
the '-C' options are executed in sequence before starting the
extraction. Listing ignores any '-C' options specified. DIR is
relative to the then current working directory, perhaps changed by
a previous '-C' option.
a new '-C' option appears in the command line. When extracting or
comparing, all the '-C' options are executed in sequence before
reading the archive. Listing ignores any '-C' options specified.
DIR is relative to the then current working directory, perhaps
changed by a previous '-C' option.
Note that a process can only have one current working directory
(CWD). Therefore multi-threading can't be used to create an
archive if a '-C' option appears after a relative filename in the
archive if a '-C' option appears after a relative file name in the
command line.
'-d'
'--diff'
Find differences between archive and file system. For each tar
member in the archive, verify that the corresponding file exists
and is of the same type (regular file, directory, etc). Report on
standard output the differences found in type, mode (permissions),
owner and group IDs, modification time, file size, file contents
(of regular files), target (of symlinks) and device number (of
block/character special files).
Compare and report differences between archive and file system.
For each tar member in the archive, verify that the corresponding
file in the file system exists and is of the same type (regular
file, directory, etc). Report on standard output the differences
found in type, mode (permissions), owner and group IDs,
modification time, file size, file contents (of regular files),
target (of symlinks) and device number (of block/character special
files).
As tarlz removes leading slashes from member names, the '-C'
option may be used in combination with '--diff' when absolute
filenames were used on archive creation: 'tarlz -C / -d'.
option may be used in combination with '--diff' when absolute file
names were used on archive creation: 'tarlz -C / -d'.
Alternatively, tarlz may be run from the root directory to perform
the comparison.
@ -184,15 +193,22 @@ equivalent to '-1 --solid'
Delete the specified files and directories from an archive in
place. It currently can delete only from uncompressed archives and
from archives with individually compressed files ('--no-solid'
archives). To delete a directory without deleting the files under
it, use 'tarlz --delete -f foo --exclude='dir/*' dir'. Deleting in
place may be dangerous. A corrupt archive, a power cut, or an I/O
error may cause data loss.
archives). Note that files of about '--data-size' or larger are
compressed individually even if '--bsolid' is used, and can
therefore be deleted. Tarlz takes care to not delete a tar member
unless it is possible to do so. For example it won't try to delete
a tar member that is not individually compressed. To delete a
directory without deleting the files under it, use
'tarlz --delete -f foo --exclude='dir/*' dir'. Deleting in place
may be dangerous. A corrupt archive, a power cut, or an I/O error
may cause data loss.
'--exclude=PATTERN'
Exclude files matching a shell pattern like '*.o'. A file is
considered to match if any component of the filename matches. For
example, '*.o' matches 'foo.o', 'foo.o/bar' and 'foo/bar.o'.
considered to match if any component of the file name matches. For
example, '*.o' matches 'foo.o', 'foo.o/bar' and 'foo/bar.o'. If
PATTERN contains a '/', it matches a corresponding '/' in the file
name. For example, 'foo/*.o' matches 'foo/bar.o'.
'-f ARCHIVE'
'--file=ARCHIVE'
@ -234,13 +250,15 @@ equivalent to '-1 --solid'
Compressed members can't be appended to an uncompressed archive,
nor vice versa. If the archive is compressed, it must be a
multimember lzip file with the two end-of-file blocks plus any
zero padding contained in the last lzip member of the archive.
Appending works as follows; first the end-of-file blocks are
removed, then the new members are appended, and finally two new
end-of-file blocks are appended to the archive. If the archive is
uncompressed, tarlz parses and skips tar headers until it finds
the end-of-file blocks. Exit with status 0 without modifying the
archive if no FILES have been specified.
zero padding contained in the last lzip member of the archive. It
is possible to append files to an archive with a different
compression granularity. Appending works as follows; first the
end-of-file blocks are removed, then the new members are appended,
and finally two new end-of-file blocks are appended to the
archive. If the archive is uncompressed, tarlz parses and skips
tar headers until it finds the end-of-file blocks. Exit with
status 0 without modifying the archive if no FILES have been
specified.
'-t'
'--list'
@ -351,7 +369,7 @@ equivalent to '-1 --solid'
that a corrupt 'GNU.crc32' keyword, for example 'GNU.crc33', is
reported as a missing CRC instead of as a corrupt record. This
misleading 'Missing CRC' message is the consequence of a flaw in
the posix pax format; i.e., the lack of a mandatory check sequence
the POSIX pax format; i.e., the lack of a mandatory check sequence
in the extended records. *Note crc32::.
'--out-slots=N'
@ -369,9 +387,24 @@ invalid input file, 3 for an internal consistency error (eg, bug) which
caused tarlz to panic.

File: tarlz.info, Node: File format, Next: Amendments to pax format, Prev: Invoking tarlz, Up: Top
File: tarlz.info, Node: Portable character set, Next: File format, Prev: Invoking tarlz, Up: Top
3 File format
3 POSIX portable filename character set
***************************************
The set of characters from which portable file names are constructed.
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
a b c d e f g h i j k l m n o p q r s t u v w x y z
0 1 2 3 4 5 6 7 8 9 . _ -
The last three characters are the period, underscore, and
hyphen-minus characters, respectively.

File: tarlz.info, Node: File format, Next: Amendments to pax format, Prev: Portable character set, Up: Top
4 File format
*************
In the diagram below, a box like this:
@ -393,7 +426,7 @@ sets). The members simply appear one after another in the file, with no
additional information before, between, or after them.
Each lzip member contains one or more tar members in a simplified
posix pax interchange format. The only pax typeflag value supported by
POSIX pax interchange format. The only pax typeflag value supported by
tarlz (in addition to the typeflag values defined by the ustar format)
is 'x'. The pax format is an extension on top of the ustar format that
removes the size limitations of the ustar format.
@ -438,7 +471,7 @@ tar.lz
+===============+=================================================+========+
3.1 Pax header block
4.1 Pax header block
====================
The pax header block is identical to the ustar header block described
@ -492,7 +525,7 @@ conversion to UTF-8 nor any other transformation.
swapping of two bytes.
3.2 Ustar header block
4.2 Ustar header block
======================
The ustar header block has a length of 512 bytes and is structured as
@ -519,11 +552,10 @@ prefix 345 155
All characters in the header block are coded using the ISO/IEC
646:1991 (ASCII) standard, except in fields storing names for files,
users, and groups. For maximum portability between implementations,
names should only contain characters from the portable filename
character set. But if an implementation supports the use of characters
outside of '/' and the portable filename character set in names for
files, users, and groups, tarlz will use the byte values in these names
unmodified.
names should only contain characters from the portable character set.
But if an implementation supports the use of characters outside of '/'
and the portable character set in names for files, users, and groups,
tarlz will use the byte values in these names unmodified.
The fields name, linkname, and prefix are null-terminated character
strings except when all characters in the array contain non-null
@ -618,38 +650,45 @@ character.

File: tarlz.info, Node: Amendments to pax format, Next: Multi-threaded tar, Prev: File format, Up: Top
4 The reasons for the differences with pax
5 The reasons for the differences with pax
******************************************
Tarlz is meant to reliably detect invalid or corrupt metadata during
decoding, and to create safe archives where corrupt metadata can be
reliably detected. In order to achieve these goals, tarlz makes some
changes to the variant of the pax format that it uses. This chapter
describes these changes and the concrete reasons to implement them.
Tarlz creates safe archives that allow the reliable detection of
invalid or corrupt metadata during decoding even when the integrity
checking of lzip can't be used because the lzip members are only
decompressed partially, as it happens in parallel '--list' and
'--extract'. In order to achieve this goal, tarlz makes some changes to
the variant of the pax format that it uses. This chapter describes
these changes and the concrete reasons to implement them.
4.1 Add a CRC of the extended records
5.1 Add a CRC of the extended records
=====================================
The posix pax format has a serious flaw. The metadata stored in pax
The POSIX pax format has a serious flaw. The metadata stored in pax
extended records are not protected by any kind of check sequence.
Corruption in a long filename may cause the extraction of the file in
Corruption in a long file name may cause the extraction of the file in
the wrong place without warning. Corruption in a large file size may
cause the truncation of the file or the appending of garbage to the
file, both followed by a spurious warning about a corrupt header far
from the place of the undetected corruption.
Metadata like filename and file size must be always protected in an
Metadata like file name and file size must be always protected in an
archive format because of the adverse effects of undetected corruption
in them, potentially much worse that undetected corruption in the data.
Even more so in the case of pax because the amount of metadata it
stores is potentially large, making undetected corruption more probable.
Headers and metadata must be protected separately from data because
the integrity checking of lzip may not be able to detect the corruption
before the metadata has been used, for example, to create a new file in
the wrong place.
Because of the above, tarlz protects the extended records with a CRC
in a way compatible with standard tar tools. *Note key_crc32::.
4.2 Remove flawed backward compatibility
5.2 Remove flawed backward compatibility
========================================
In order to allow the extraction of pax archives by a tar utility
@ -657,12 +696,12 @@ conforming to the POSIX-2:1993 standard, POSIX.1-2008 recommends
selecting extended header field values that allow such tar to create a
regular file containing the extended header records as data. This
approach is broken because if the extended header is needed because of
a long filename, the name and prefix fields will be unable to contain
a long file name, the name and prefix fields will be unable to contain
the full pathname of the file. Therefore the files corresponding to
both the extended header and the overridden ustar header will be
extracted using truncated filenames, perhaps overwriting existing files
or directories. It may be a security risk to extract a file with a
truncated filename.
extracted using truncated file names, perhaps overwriting existing
files or directories. It may be a security risk to extract a file with
a truncated file name.
To avoid this problem, tarlz writes extended headers with all fields
zeroed except size, chksum, typeflag, magic and version. This prevents
@ -672,28 +711,29 @@ overridden by extended records.
If an extended header is required for any reason (for example a file
size larger than 8 GiB or a link name longer than 100 bytes), tarlz
moves the filename also to the extended header to prevent an ustar tool
from trying to extract the file or link. This also makes easier during
parallel decoding the detection of a tar member split between two lzip
members at the boundary between the extended header and the ustar
header.
moves the file name also to the extended header to prevent an ustar
tool from trying to extract the file or link. This also makes easier
during parallel decoding the detection of a tar member split between
two lzip members at the boundary between the extended header and the
ustar header.
4.3 As simple as possible (but not simpler)
5.3 As simple as possible (but not simpler)
===========================================
The tarlz format is mainly ustar. Extended pax headers are used only
when needed because the length of a filename or link name, or the size
when needed because the length of a file name or link name, or the size
of a file exceed the limits of the ustar format. Adding extended
headers to each member just to record subsecond timestamps seems
wasteful for a backup format.
wasteful for a backup format. Moreover, minimizing the overhead may
help recovering the archive with lziprecover in case of corruption.
Global pax headers are tolerated, but not supported; they are parsed
and ignored. Some operations may not behave as expected if the archive
contains global headers.
4.4 Avoid misconversions to/from UTF-8
5.4 Avoid misconversions to/from UTF-8
======================================
There is no portable way to tell what charset a text string is coded
@ -705,7 +745,7 @@ this behavior will be adjusted with a command line option in the future.

File: tarlz.info, Node: Multi-threaded tar, Next: Minimum archive sizes, Prev: Amendments to pax format, Up: Top
5 Limitations of parallel tar decoding
6 Limitations of parallel tar decoding
**************************************
Safely decoding an arbitrary tar archive in parallel is impossible. For
@ -753,7 +793,7 @@ example listing the Silesia corpus on a dual core machine:

File: tarlz.info, Node: Minimum archive sizes, Next: Examples, Prev: Multi-threaded tar, Up: Top
6 Minimum archive sizes required for multi-threaded block compression
7 Minimum archive sizes required for multi-threaded block compression
*********************************************************************
When creating or appending to a compressed archive using multi-threaded
@ -791,7 +831,7 @@ Level

File: tarlz.info, Node: Examples, Next: Problems, Prev: Minimum archive sizes, Up: Top
7 A small tutorial with examples
8 A small tutorial with examples
********************************
Example 1: Create a multimember compressed archive 'archive.tar.lz'
@ -850,7 +890,7 @@ Example 8: Copy the contents of directory 'sourcedir' to the directory

File: tarlz.info, Node: Problems, Next: Concept index, Prev: Examples, Up: Top
8 Reporting bugs
9 Reporting bugs
****************
There are probably bugs in tarlz. There are certainly errors and
@ -881,6 +921,9 @@ Concept index
* invoking: Invoking tarlz. (line 6)
* minimum archive sizes: Minimum archive sizes. (line 6)
* options: Invoking tarlz. (line 6)
* parallel tar decoding: Multi-threaded tar. (line 6)
* portable character set: Portable character set.
(line 6)
* usage: Invoking tarlz. (line 6)
* version: Invoking tarlz. (line 6)
@ -888,20 +931,21 @@ Concept index

Tag Table:
Node: Top223
Node: Introduction1086
Node: Invoking tarlz3337
Ref: --data-size5489
Ref: --bsolid12172
Node: File format15802
Ref: key_crc3220622
Node: Amendments to pax format26039
Ref: crc3226580
Ref: flawed-compat27605
Node: Multi-threaded tar30128
Node: Minimum archive sizes32667
Node: Examples34800
Node: Problems36517
Node: Concept index37043
Node: Introduction1155
Node: Invoking tarlz3841
Ref: --data-size6006
Ref: --bsolid13287
Node: Portable character set16917
Node: File format17420
Ref: key_crc3222248
Node: Amendments to pax format27647
Ref: crc3228304
Ref: flawed-compat29564
Node: Multi-threaded tar32198
Node: Minimum archive sizes34737
Node: Examples36870
Node: Problems38587
Node: Concept index39113

End Tag Table