Shell Scripting: Expert Recipes for Linux, Bash, and More

Summary

Although it is quite possible to administer and write shell scripts for a GNU/Linux system without knowing any of the history behind it, a lot of apparent quirks will not make sense without some appreciation of how things came to be the way they are. There is a difference between scripting for a typical Linux distribution, such as RedHat, SuSE, or Ubuntu, and scripting for an embedded device, which is more likely to be running busybox than a full GNU set of tools. Scripting for commercial Unix is slightly different again, and much as a web developer has to take care to ensure that a website works in multiple browsers on multiple platforms, a certain amount of testing is required to write solid cross-platform shell scripts.

Even when writing for a typical Linux distribution, it is useful to know what is where, and how it came to be there. Is there an /etc/sysconfig? Are init scripts in /etc/rc.d/init.d or /etc/init.d, or do they even exist in that way? What features can be identified to see what tradition is being followed by this particular distribution? Knowing the history of the system helps one to understand whether the syntax is tar xzf or tar -xzf; whether to use /etc/fstab or /etc/vfstab; whether running killall httpd will stop just your Apache processes (as it would under GNU/Linux) or halt the entire system (as it would on Solaris)!

The next chapter follows on from this checkered history to compare the variety of choices available when selecting a Unix or GNU/Linux ...