Thursday·15·November·2012
Tools to handle archives conveniently //at 01:42 //by abe
TL;DR: There’s a summary at the end of the article.
Today I wanted to see why a dependency in a .deb-package
from an external APT repository changed so that it became
uninstallable. While dpkg-deb --info foobar.deb easily
shows the control information, the changelog is in the filesystem part
of the package.
I could extract that one dpkg-deb, too,
but I’d have to extract either to some temporary directory or pipe it
into tar which then can extract a single file from the archive and
sent it to STDOUT:
dpkg-deb --fsys-tarfile foobar.deb | tar xOf - ./usr/share/doc/foobar/changelog.Debian.gz | zless
But that’s tedious to type. The following command is clearly less to type and way easier to remember:
acat foobar.deb ./usr/share/doc/foobar/changelog.Debian.gz | zless
acat stands for “archive cat” is part of the atool suite of commands:
- als
- lists files in an archive.
$ als foobar.tgz drwxr-xr-x abe/abe 0 2012-11-15 00:19 foobar/ -rw-r--r-- abe/abe 13 2012-11-15 00:20 foobar/bar -rw-r--r-- abe/abe 13 2012-11-15 00:20 foobar/foo
- acat
- extracts files in an archive to standard out.
$ acat foobar.tgz foobar/foo foobar/bar foobar/bar bar contents foobar/foo foo contents
- adiff
- generates a diff between two archives using diff(1).
$ als quux.zip Archive: quux.zip Length Date Time Name --------- ---------- ----- ---- 0 2012-11-15 00:23 quux/ 16 2012-11-15 00:22 quux/foo 13 2012-11-15 00:20 quux/bar --------- ------- 29 3 files $ adiff foobar.tgz quux.zip diff -ru Unpack-3594/foobar/foo Unpack-7862/quux/foo --- Unpack-3594/foobar/foo 2012-11-15 00:20:46.000000000 +0100 +++ Unpack-7862/quux/foo 2012-11-15 00:22:56.000000000 +0100 @@ -1 +1 @@ -foo contents +foobar contents- arepack
- repacks archives to a different format. It does this by first
extracting all files of the old archive into a temporary directory,
then packing all files extracted to that directory to the new archive.
Use the
--each(-e) option in combination with--format(-F) to repack multiple archives using a single invocation ofatool. Note thatarepackwill not remove the old archive. $ arepack foobar.tgz foobar.txz foobar.tgz: extracted to `Unpack-7121/foobar' foobar.txz: grew 36 bytes
- apack
- creates archives (or compresses files). If no file arguments are specified, filenames to add are read from standard in.
- aunpack
- extracts files from an archive. Often one wants to extract all files in an archive to a single subdirectory. However, some archives contain multiple files in their root directories. The aunpack program overcomes this problem by first extracting files to a unique (temporary) directory, and then moving its contents back if possible. This also prevents local files from being overwritten by mistake.
(atool subcommand descriptions from the atool man page which is licensed under GPLv3+. Examples by me.)
I though miss
the existence of an agrep subcommand. Guess why?
atool supports a wealth of archive types: tar (gzip-,
bzip-, bzip2-, compress-/Z-, lzip-, lzop-, xz-, and 7zip-compressed),
zip, jar/war, rar, lha/lzh, 7zip, alzip/alz, ace, ar, arj, arc, rpm,
deb, cab, gzip, bzip, bzip2, compress/Z, lzip, lzop, xz, rzip, lrzip
and cpio. (Not all subcommands support all archive types.)
Similar Utilities
There are some utilities which cover parts of what atool does, too:
Tools from the mtools package
Yes, they come from the “handle MS-DOS floppy disks tool” package, don’t ask me why. :-)
- uz
gunzips and extracts agzip‘dtar‘d archives- Advantage over
aunpack: Less to type. :-) - Disadvantage compared to
aunpack: Supports only one archive format. - lz
gunzips and shows a listing of agzip‘dtar‘d archive- Advantage over
als: One character less to type. :-) - Disadvantage compared to
als: Supports only one archive format.
unp
unp extracts one or more files given as arguments on the
command line.
$ unp -s Known archive formats and tools: 7z: p7zip or p7zip-full ace: unace ar,deb: binutils arj: arj bz2: bzip2 cab: cabextract chm: libchm-bin or archmage cpio,afio: cpio or afio dat: tnef dms: xdms exe: maybe orange or unzip or unrar or unarj or lha gz: gzip hqx: macutils lha,lzh: lha lz: lzip lzma: xz-utils or lzma lzo: lzop lzx: unlzx mbox: formail and mpack pmd: ppmd rar: rar or unrar or unrar-free rpm: rpm2cpio and cpio sea,sea.bin: macutils shar: sharutils tar: tar tar.bz2,tbz2: tar with bzip2 tar.lzip: tar with lzip tar.lzop,tzo: tar with lzop tar.xz,txz: tar with xz-utils tar.z: tar with compress tgz,tar.gz: tar with gzip uu: sharutils xz: xz-utils zip,cbz,cbr,jar,war,ear,xpi,adf: unzip zoo: zoo
So it’s very similar to aunpack, just a shorter command
and it supports some more exotic archive formats which
atool doesn’t support.
Also part of the unp package is ucat which does
more or less the same as acat, just with unp
as backend.
dtrx
From the man page of dtrx:
In addition to providing one command to extract many different archive types,
dtrxalso aids the user by extracting contents consistently. By default, everything will be written to a dedicated directory that’s named after the archive. dtrx will also change the permissions to ensure that the owner can read and write all those files.Supported archive formats: tar, zip (including self-extracting .exe files), cpio, rpm, deb, gem, 7z, cab, rar, and InstallShield. It can also decompress files compressed with gzip, bzip2, lzma, or compress.
dtrx -l lists the contents of an archive, i.e. works like
als or lz.
dtrx has two features not present in the other tools mentioned so far:
- It can extract metadata instead of the normal contents from .deb and .gem files.
- It can extract archives recursively, i.e. can extract archives inside of archives.
Unfortunately you can’t mix those two features. But you can use the following tool for that purpose:
deepfind
deepfind is a command from the package strigi-utils and recursively lists files in archives, including archives in archives. I’ve already written a detailed blog-posting about deepfind and its friend deepgrep.
tardiff
tardiff was written to check what changed in source code
tarballs from one release to another. By default it just lists the
differences in the file lists, not in the files’ contents and hence
works different than adiff.
Summary
atool and friends are probably the first choice when it comes to
DWIM archive handling, also
because they have an easy to remember subcommand scheme.
uz and lz and the shortest way to extract or
list the contents of a .tar.gz file. But nothing more. And you have to
install mtools even if you don’t have a floppy drive.
unp comes in handy for exotic archive formats atool
doesn’t support. And it’s way easier to remember and type than
aunpack.
dtrx is neat if you want to extract archives in archives
or if you want to extract metadata from some package files with just a
few keystrokes.
For listing all files in recursive archives, use
deepfind.
Tagged as: 7zip, acat, adiff, als, apack, archives, atool, aunpack, bzip, bzip2, deb, deepfind, dtrx, floppy, gem, grep, gzip, lha, lrzip, lz, lzip, lzop, MS-DOS, mtools, rar, rpm, rzip, strigi-utils, tar, tardiff, ucat, unp, UUUT, uz, xz, zip
4 comments // write a comment // comments off
Related stories
Comments
Re: Tools to handle archives conveniently
Posted by: Axel Website: http://noone.org/blog Time: Sat, 17 Nov 2012 01:48 Yes and no. acat can handle nested stuff if its format is know to be nested, e.g. .deb (.tar inside .ar), .tgz (.tar inside .gz), etc. The very same examples shows that:
$ acat foobar.deb ./usr/share/doc/foobar/changelog.Debian.gz | file -
/dev/stdin: gzip compressed data, from Unix, max compressionIf it would have understood nested format, it would have uncompressed changelog.Debian.gz, too.
For the same reason it doesn't need to resolve conflicts as with .deb files it always looks for the data tar archive, never for the control tar archive. It actually uses dpkg-deb's --fsys-tarfile internally:
$ acat xulrunner-18.0_18.0~a2+20121109042012-1_amd64.deb ./DEBIAN/control
tar: ./DEBIAN/control: Not found in archive
tar: Exiting with failure status due to previous errors
acat: dpkg-deb ... | tar ...: non-zero return-codeRe: Tools to handle archives conveniently
Posted by: Joachim Breitner Website: http://www.joachim-breitner.de/ Time: Thu, 15 Nov 2012 11:01 Thanks for the post; especially adiff is something that I have been wanting several times.
Re: Tools to handle archives conveniently
Posted by: Pino Toscano Website: Time: Thu, 15 Nov 2012 15:22 Recently, in the glibc development mailing list was mentioned pkgdiff [1], and applied for recent glibc versions [2]. While it is more specific of the tools you mentioned (it is taylored to diff tarballs of source code), could be useful to mention.
[1] http://pkgdiff.github.com/pkgdiff/
[2] http://sourceware.org/glibc/wiki/Testing/Diff_tool
Your Comment
Spam Protection: To post a comment, you'll have to answer the following question: What is 42 minus 19?
Re: Tools to handle archives conveniently
Looks like acat can handle nested archives, like in your Debian package example. Do you know how it would resolve conflicts between the two archives that are inside de Debian package?
Reply