Using the Shell

The shell is an interactive command prompt with many different features:

• Input and output redirection

• Background processing

• Job control

• History editing

• Built-in help

• Command-line completion

• Command-line editing

The shell interprets keyboard commands and is generally used to launch other commands or programs using the shell's interpreter language known as shell scripts.

NOTE

Shell scripts are discussed in Chapter 14.

The shell you use is assigned by the last field in your entry in the system's /etc/passwd file. This example, for a user named andrew, shows that the login shell is bash:

 andrew:x:502:502::/home/andrew:/bin/bash 

The default shell for most Linux distributions, including Fedora, is the GNU bash or Bourne Again SHell, but other shells, such as tcsh, ksh, and zsh are available for use. You can use a different shell by typing its name at the command line. Alternatively, the root operator might assign a user to another shell when creating that user account (see "Working as Root," later in this chapter).

CAUTION

If you're interested in trying a different shell with Linux, you can change your login shell using the chsh command, but make sure that the shell is actually installed on your system. For example, to change your default shell to tcsh, first use the which command to verify that it is installed:

 $ which tcsh /bin/tcsh 

This example shows that the tcsh shell is installed under the /bin directory. The tcsh shell should also be listed in your system's list of approved shells, /etc/shells. Check to make sure that it is listed:

 $ grep tcsh /etc/shells /bin/tcsh 

You can also use the chsh command's -l option to list valid system shells in order to verify that using tcsh is allowed. Because tcsh is installed and listed in /etc/shells, you can then change your shell using the chsh command:

 $ chsh Changing shell for andrew. Password: New shell [/bin/bash]: /bin/ksh chsh: "/bin/ksh" does not exist. 

Note that the chsh command will report an error if you enter the name of a shell not installed on your system.

 $ chsh -s /bin/tcsh Changing shell for andrew. Password: Shell changed. 

If you now take a look at your /etc/passwd entry, you will see /bin/tcsh as your default shell. The next time you log in, you'll use tcsh.

Using Environment Variables

A number of in-memory variables are assigned and loaded by default when the user logs in. These variables are known as shell environment variables, which can be used by various commands to get information about your environment, such as the type of system you are running, your home directory, and the shell in use. Environment variables are used by Linux operating systems to help tailor the computing environment of your system, and include helpful specifications and setup, such as default locations of executable files and software libraries. If you begin writing shell scripts, you might use environment variables in your scripts. Until then, you only need to be aware of what environment variables are and do.

The following list includes a number of environment variables, along with descriptions of how the shell uses them:

• PWD To provide the name of the current working directory, used by the pwd command (such as /home/andrew/foo)

• USER To declare the user's name, such as andrew

• LANG To set language defaults, such as English

• SHELL To declare the name and location of the current shell, such as /bin/bash

• PATH To set the default location of executable files, such as /bin, /usr/bin, and so on

• LD_LIBRARY_PATH To declare the location of important software libraries (because most, but not all, Linux commands use shared resources)

• TERM To set the type of terminal in use, such as vt100, which can be important when using screen-oriented programs, such as text editors

• MACHINE To declare system type, system architecture, and so on

NOTE

Each shell can have its own feature set and language syntax, as well as a unique set of default environment variables. See Chapter 14 for more information about using the different shells included with Fedora.

At the command line, you can use the env or printenv commands to display these environment variables, like so:

 $ env PWD=/home/andrew HOSTNAME=laptop.andbhudson.co.uk USER=andrew MACHTYPE=i386-redhat-linux-gnu MAIL=/var/spool/mail/andrew BASH_ENV=/home/andrew/.bashrc LANG=en_GB DISPLAY=0:0 LOGNAME=andrew SHLVL=1 PATH=/usr/kerberos/bin:/usr/local/bin:/bin:/usr/bin: \ /usr/X11R6/bin:/home/bball/bin SHELL=/bin/bash HOSTTYPE=i386 OSTYPE=linux-gnu HISTSIZE=1000 TERM=xterm HOME=/home/andrew 

This abbreviated list shows a few common variables. These variables are set by configuration or resource files contained in the /etc, /etc/skel, or user /home directory. You can find default settings for bash, for example, in /etc/profile, /etc/bashrc, the .bashrc, or .bash_profile files installed in your home directory. Read the man page for bash for details about using these configuration files.

One of the most important environment variables is $PATH, which defines the location of executable files. For example, if as a regular user, you try to use a command that is not located in your $PATH, such as the ifconfig command, you'll see something like this:

 $ ifconfig -bash: ifconfig: command not found 

However, you might know that ifconfig is definitely installed on your system, and you can verify this using the whereis command like so:

 $  whereis ifconfig ifconfig: /sbin/ifconfig /usr/share/man/man8/ifconfig.8.gz 

You can also run the command by typing its full pathname, or complete directory specification like this:

 $ /sbin/ifconfig 

As you can see in this example, the ifconfig command is indeed installed. What happened is that by default, the /sbin directory is not in your $PATH. One of the reasons for this is that commands under the /sbin directory are normally intended to only be run by root. You can add /sbin to your $PATH by editing the file .bash_profile in your home directory (if you use the bash shell by default, like most Linux users). Look for the following line:

 PATH=$PATH:$HOME/bin 

You can then edit this file, perhaps using the vi editor (discussed in this chapter), to add the /sbin directory like so:

 PATH=$PATH:/sbin:$HOME/bin 

Save the file. The next time you log in, the /sbin directory will be in your $PATH. One way to use this change right away is to read in the new settings in .bash_profile by using the bash shell's source command like so:

 $ source .bash_profile 

You'll now be able to run ifconfig without the need to explicitly type its full pathname.

Some Linux commands also use environment variables, for example, to acquire configuration information (such as a communications program looking for a variable such as BAUD_RATE, which might denote a default modem speed).

To experiment with the environment variables, you can modify the PS1 variable to manipulate the appearance of your shell prompt. If you are working with bash, you can use its built-in export command to change the shell prompt. For example, if your default shell prompt looks like

 [andrew@laptop ~]$  

You can change its appearance by using the PS1 variable like this:

 $  PS1='$OSTYPE r00lz ->' 

After you press Enter, you'll see

 linux-gnu r00lz ->  

NOTE

See the bash manual page for other variables you can use for prompt settings.

Navigating and Searching with the Shell

Use the cd command (built into the shell) to navigate through the Fedora file system. This command is generally used with a specific directory location, or pathname like this:

 $ cd /usr/X11R6/lib/X11/doc 

Under Fedora, the cd command can also be used with several shortcuts. For example, to quickly move up to the parent (higher-level) directory, use the cd command like this:

 $ cd .. 

To return to one's home directory from anywhere in the Linux file system, use the cd command like this:

 $ cd 

You can also use the $HOME shell environment variable to accomplish the same thing. Type this command and press Enter to return to your home directory:

 $ cd $HOME 

You can accomplish the same thing by using the tilde (~) like this:

 $ cd ~ 

Linux also includes a number of GNU commands you can use to search the file system. These include

• whereis command Returns the location of the command and its man page.

• whatis command Returns a one-line synopsis from the command's man page.

• locate file Returns locations of all matching file(s); an extremely fast method of searching your system, as locate searches a database containing an index of all files on your system; however, this database (about 4MB in size and named slocate.db under the /var/lib/slocate directory) is built daily at 4:20 a.m. by default, and will not contain pathnames to files created during the workday or in the evening.

• apropos subject Returns a list of commands related to subject.

Managing Files with the Shell

Managing files in your home directory involves using one or more easily remembered commands. If you have any familiarity with the now ancient DOS, you'll recognize some of these commands (though their names are different from those you remember). Basic file management operations include paging (reading), moving, renaming, copying, searching, and deleting files and directories. These commands include

• cat filename Outputs contents of filename to display

• less filename Allows scrolling while reading contents of filename

• mv file1 file2 Renames file1 to file2

• mv file dir Moves file to specified directory

• cp file1 file2 Copies file1 and creates file2

• rm file Deletes file

• rmdir dir Deletes directory (if empty)

• grep string file(s) Searches through files(s) and displays lines containing matching string

Note that each of these commands can be used with pattern-matching strings known as wildcards or expressions. For example, to delete all files in the current directory beginning with the letters abc, you can use an expression beginning with the first three letters of the desired filenames. An asterisk (*) is then appended to match all these files. Use a command line with the rm command like this:

 $ rm abc* 

Linux shells recognize many types of filenaming wildcards, but this is different from the capabilities of Linux commands supporting the use of more complex expressions. You learn more about using wildcards in Chapter 14.

NOTE

Learn more about using expressions by reading the ex or grep manual pages.

Compressing and Decompressing Files Through the Shell

Another file management operation is compression and decompression of files, or the creation, listing, and expansion of file and directory archives. Linux distros usually include several compression utilities that you can use to create, compress, expand, or list the contents of compressed files and archives. These commands include

• bunzip2 Expands a compressed file

• bzip2 Compresses or expands files and directories

• gunzip Expands a compressed file

• gzip Compresses or expands files and directories

• shar file Creates a shell archive of files

• tar Creates, expands, or lists the contents of compressed or uncompressed file or directory archives known as tape archives or "tarballs"

• unshar Reassembles files from the shell archive

• uudecode file.uu Decodes an uuencoded text file to its binary form

• uuencode file Encodes a binary file to text file format for transmission via email

NOTE

The rpm command can also be used to manage archived data, but is generally used for software development and management. See Chapter 7 for more information on using RPM.

Most of these commands are easy to use. The tar command, however, has a somewhat complex (though capable) set of command-line options and syntax. Even so, you can quickly learn to use tar by remembering a few simple invocations on the command line. For example, to create a compressed archive of a directory, use tar's czf options like this:

 $ tar czf dirname.tgz dirname 

The result will be a compressed archive (a file ending in .tgz) of the specified directory (and all files and directories under it). Add the letter v to the preceding options to view the list of files being added during compression and archiving. To list the contents of the compressed archive, substitute the c option with the letter t, like this:

 $ tar tzf archive 

Of course, if many files are in the archive, a better invocation (in order to easily read or scroll through the output) is

 $ tar tzf archive | less 

To expand the contents of a compressed archive, use tar's xzf options, like so:

 $ tar xzf archive 

tar decompresses the specified archive and extracts the contents in the current directory.