The Internet Encyclopedia (Volume 3)

P1: JDW

UNIX WL040/Bidgoli-Vol III-Ch-41 August 13, 2003 17:26 Char Count= 0

504 UNIXOPERATINGSYSTEM

Table 1Command Line Shortcuts Available in Shells

Supporting Command History

!:0 The previous command name

!:n Repeat argumentnof the

previous command

!:ˆ[or !ˆ, leaving out Repeat the first argument of the

the colon] previous command

!:$ [or !$, leaving out Repeat the last argument of the

the colon] previous command

!:n1–n2 Repeat the arguments between

n1andn2from the previous

command, inclusive

!:-n Repeat arguments 0 throughn

from the previous command

Table 1 lists some of the command line shortcuts that are

available in shells supporting command history.

Command Line Editing

Most modern Unix shells (with the exception of the orig-

inal Bourne shell) support a feature known as command

line editing. Command line editing allows one to use the

history buffer to move back to previous commands and

modify them using a line editing syntax from Unix edi-

tors, namely, vi and emacs. The user can also edit current

command lines as he or she types. For example, in the

Korn shell, a user can tell the shell that he wants to edit

command lines as if he were in the emacs editor, by is-

suing the command “set -o emacs” (“set -o vi” would set

the editing for vi emulation). Once he has set emacs em-

ulation, the user can navigate the command history and

command line using familiar emacs cursor movement and

editing syntax.

For example, to go to the previous line in an emacs

buffer, one would press Ctrl-p. To go to the next (subse-

quent) line, one would press Ctrl-n. To move backward

on a line, character by character, one would press Ctrl-b

repeatedly. To move forward character by character, one

would press Ctrl-f. A user can walk backward character

by character on a command line, by pressing Ctrl-b. The

ability to search backward through the command history

is also available in many shells through the Ctrl-r key com-

bination. Table 2 shows some of the more common emacs

command line editing commands.

Command Line Completion

Coupled with command line editing is command line

completion. Imagine that there are three files in the cur-

rent directory: file1000a, file2000b, and file2010c. A shell

that offers command line completion will allow the user

to type just enough of a filename and then press a com-

mand line completion key, and the shell will examine the

file system andcompletethe rest of the filename automa-

tically. For example, suppose a user is using the bash shell,

enters the command “ls file1,” and then presses the tab

key. The tab key is the completion key for bash and tcsh

(it’s ESC ESC in the Korn shell). The shell automatically

completes the filename, leaving you with “ls file1000a” on

the command line.

Table 2Common Emacs Command Line Editing Commands

Meta-b Move one word back on the

command line (the Meta key

is often the “Alt” key on a PC

keyboard)

Meta-f Move one word forward on the

command line

Meta-backspace Delete the word directly before the

cursor’s current position on the

command line

Ctrl-a Move to the beginning of the

command line

Ctrl-e Move to the end of the command

line

Meta-d Delete the next word in the

command line

Aliases

Sometimes users want shorter names for lengthy common

command lines. Suppose, for instance, that every time a

user wanted to run the emacs editor to edit a file, she

wanted to set the colors and font size. So for every file she

edits, she would need to type something like the following

line:

/usr/bin/emacs -fn 6x13 -fg black -bg cyan

-cr blue -ms white filename.txt

The user would certainly get tired of typing that long

command every time she wanted to edit a file. Aliases can

be used to store commands for later use. For example, one

could store the core of the previous command (minus the

filename) in an alias, and then simply use the alias to refer

to the longer command:

alias myemacs = "/usr/bin/emacs -fn 6x13 -

fg black -bg cyan -cr blue -ms white"

Once that command is issued, the user could then sim-

ply type “myemacs filename.txt,” and the myemacs alias

would expand into the longer command line associated

with the alias, setting all her desired colors and the font

size to 6×13. Aliases, and other shell commands (such as

the set -o emacs command) are generally placed in startup

command files, which are default shell scripts loaded by

the shell each time a user logs in. In bash, the main file is

.bashrc. In the Korn shell, it is.kshrc. For the C Shell, it is

.cshrc (.tcshrc for the TC shell).

Shell Variables

Shells also allow the creation of user-defined variables that

enable the shell programmer to store temporary values. A

shell variable is usually an all capital letter name refer-

encing a particular value. Shell variables are accessed by

placing a dollar sign in front of the variable name. For ex-

ample, one of the standard shell variables is called $USER,

which contains the username of the user currently logged

into a given terminal. The echo command will print out

to the terminal the value of any variable. So, issuing the