Scripts

Scripts

Steven J Zeil

Old Dominion University Dept. of Computer Science

Apr. 9, 2006

Contents:

1 Parameters

2 Control Flow

3 The test and expr Programs

4 Scripting Example: The Student Test Manager

4.1 Sample Scenario

4.2 The Script

5 Debugging Your Scripts

5.1 Isolate the Problem

5.2 Add Debugging Output

5.3 Trace the Execution

(A printable version of this document is available here.)

You can put any sequence of Unix commands into a file and turn that file into a command. Such a file is called a script. For example, suppose that you are working on a program myprog and have several files of test data that you run through it each time you make a change. Create a file dotest1 with the following lines:

./myprog < test1.dat > test1.dat.out ./myprog < test2.dat > test2.dat.out ./myprog < test3.dat > test3.dat.out ./myprog < test4.dat > test4.dat.out ./myprog < test5.dat > test5.dat.out

You can't execute dotest1 yet, because you don't have execute permission. (Do ls -l dotest1 to see this.) So use the chmod command to add execute permission:

chmod u+x dotest1

Now you can execute dotest1 by simply typing

./dotest1

Most shells provide special facilities for use in scripts. Since these differ from one shell to another, it's a good idea to tell Unix which shell to use when running the script. You do this by placing the command #!/bin/csh in the first line of the script.

In fact, you can list any program there, not just /bin/csh, and Unix will use that program to process the remainder of the lines in the script.

Any text file with execute permission can be invoked as a program.
The first line must identify the program to be run (after a starting #!).
The remaining lines are fed to the standard input of that program.

So another possibility for executing your test code would be to put the line #!./myprog at the front of each of the test data input files test1.dat, test2.dat, …, and then execute those data files!