scripts - All about shell scripts
This entry is an addendum to the C-Shell (csh) entry. Refer to csh first for introductory details on shell script files.
Shell scripts are text files that contain commands to be executed as a batch operation by the shell. Anything that can be performed at the shell prompt can be executed from a shell script. Here’s a sample:
# First, set up some command aliases
alias ls ls -F
alias c "clear; ctime"
alias send sx -yl
alias recv rx -yl
# Now print out a random, hopefully humorous adage
fortune
mail -n # read new mail (if any)
# Lastly, set up some variables
set bob=$/usr/bob
set prompt="\$time[4]> "
A file such as this can be created using a text editor and saved in your user directory. You can execute it from the shell just by typing “source file”, where “file” is the name of the script you just created.
You can avoid having to enter the shell’s “source” command to run a shell script by changing the filetype attribute of the script to “cmd”. To change a text file to an executable “cmd” (command) file, use the setfile command. If you’ve just created a script and named it “wonka”, for example, the command:
setfile -t cmd wonka
makes it executable. To run it, you simply enter its name. Note that the editor can edit “cmd” files as if they were ordinary text files, so using setfile is only necessary once after the file is first created (editing the file does not change its type). If you plan on creating many shell scripts, they’re typically stored in a directory called “bin” in your $home directory (e.g. $home/bin/wonka).
There is a special “cmd” file named “login” in your directory which is automatically executed after logging in. You can use an editor to tailor it to your heart’s desire, making your session on ProLine a unique experience.
Your scripts have the ability to use some special commands that allow for
greater flexibility in the way they execute:
halt a running script and return control back to the shell for regular keyboard-based command input. The exit command accepts a numeric argument as a result code to pass back to the shell or calling process. Any non-zero result is considered an error. If no argument is given, zero is assumed (no error).
test for a certain condition, such as the existence of a file. If the condition is true, script execution continues with the commands which follow until an else or endif command is encountered. If the condition is false, the first else or endif command on a new line is searched for, and execution will continue with the commands that follow it. (See if for complete details on if, else, and endif)
read a line of input from the user and assign it to a shell variable.
You can place comments into your scripts by putting a pound-sign (#) before each comment. The shell ignores anything that follows.
Scripts can also make use of any arguments that are entered at the command line when the script was launched. This is done by making reference to the argc and argv shell variables which hold the argument count and argument list respectively.
Consider following script called “view”:
if $argc < 2 then
echo -n "Name of file to view? "
read file
else
set file=$argv[1]
endif
if ! -f $file then
echo Can't seem to find $file!
exit -1
endif
echo ------------------
cat $file
echo ------------------
View begins by checking the argument count in argc. Since the command itself is part of the argument list, the argument count is always one plus the number of arguments following the command.
If $argc is less than two, meaning it must be one because only the name of the script was entered, then it prompts the user to enter a file name. Otherwise, if $argc is two or more, it assigns the file variable to the first argument, $argv[1].
Next, the view script tests for the existence of the file. The “!” operator is a logical NOT, which reverses the logic. Thus, the statement reads, “if the file does not exist”. Should the file be non-existent, an error message is printed, and the script exits with -1, which sets the shell’s status variable to -1.
If the file does exist, it is displayed, bordered by lines of dashes.
Normally, if an error occurs (due to a command or process that
returns a non-zero result), a running script is halted. To avoid this,
unset the exit shell variable. To restore normal error handling,
set exit (no assignment value is necessary — it just has to exist).
A script can call another script as a command, but when the second script stops, the first one WILL NOT resume execution. To have execution of the calling script resume, launch the second script through a new shell. Example:
# This is the first script $shell source script2 # invoke 2nd script via a new shell echo It worked # when the 2nd is done, it returns here # end of first script
The “shell” variable contains the pathname to the current shell program.