Chapter 15
Shell Features
This chapter looks at three specific tasks: installing an init script onto any distribution of Linux or Unix, reporting on installed RPM packages, and a Kickstart postinstall script.
The first recipe points out some techniques for dealing with portability issues. Probably the most important thing to take from this recipe is that it is broken into four distinct steps as outlined in the Structure section. Although it might seem more efficient to copy the script to the relevant location as soon as the distribution has been determined, the code is made much more maintainable by realizing that these are separate tasks and doing only one step at a time.
The RPM Report recipe uses arrays from Chapter 9 and some of the more complicated variable structures from Chapter 7 to produce a fairly short and simple script to perform a common task. It also shows how some simple HTML formatting can be applied in a shell script in order to exploit the greater display capabilities of a graphical web browser for a more professional-looking presentation than you get with plain text output.
The Kickstart postinstall recipe sources libraries of functions and also uses here documents and redirection to make for a simple and maintainable Kickstart environment. It also demonstrates how different forms of conditional execution can be used to match the situation; sometimes testing the return code in $? makes for neater code, sometimes [ command ] && command is more readable. A lot ...
