College of Engineering & CS Wright State University Dayton, Ohio 45435-0001

CEG 333: Introduction to Unix Prabhaker Mateti tar

Tar stands for "tape archiver", although tar files are more commonly found on Internet sites than tapes nowadays. It is the most common Unix archiving utility. Unlike some compression utilities (like pkzip for .zip files), tar can only archive files (combine them into one large file) and not compress them. In keeping with the Unix philosophy of small programs that are combined as needed, separate compressors are used in conjunction with tar.

Tar can automatically call a compressor for seamless compression and decompression of archives. There are three different major Unix compressors: compress, gzip, and bzip2 Each of these works with single files. Alone, they can't combine multiple files into one package.

A tar command line gives:

• Exactly one operation:
  • -x for extract
  • -t for list
  • , and -c for create.
• One compressor, or none for an uncompressed archive:
  • -Z for compress (.Z files)
  • -z for gzip (.gz files)
  • -j for bzip2 (.bz2 files)
  (Compressors listed in order of decreasing age of algorithm, increasing size efficiency, and increasing processor usage. Not all versions of tar support all three.)
• Optional: a verbosity option. -v to list every file tar operates on, -vv to provide a complete listing (like ls -l)
• -f followed by the name of the archive file. (If absent, tar operates on stdin/out)
• All remaining tokens are assumed to be filenames. When adding, these are required. When extracting, if any filenames are present tar will extract only those files and directories. (Note that tar can operate recursively; if given a directory it adds every file in that directory and preserves paths when extracting.)

For example, the tar xvvfj ~/P1-2003-Fall.tbz command given in P3 means: "Extract this is a tar file compressed with bzip2, listing every file as it's extracted."

WARNING: The file option ("-f" or just "f") uses the string immediately after it as the output filename! So tar -czfv output.tar input files will fail, because tar will write its output to a file named "v" and give an error when it doesn't find an input file named "output.tar".