Recipe 16-3: Configuration Files
Configuration files are common to almost every operating system and application. Some systems tend to favor binary files, as this is more convenient as a way of dumping the state of the application and reading the state back in. The Unix and GNU/Linux tradition is to use text files. This can be far more convenient for a shell script to use, and it also means that the files can be manipulated by human intervention, by other scripts and other software. It also means that there need be no lock-in to a particular application just because the data or configuration can only be read by one piece of software.
- source (.)
- Variable assignment
A shell script can be encompassed within a single file, including all of its code, data, and configuration. However, the script can be more flexible if these items are stored in separate files. The user can edit the configuration as a simple text file; this way the user does not ever have to see the implementation details of the script that reads and interprets them. This happens at all levels, from the application user at the high level down to system administration tasks such as editing /etc/hosts or /etc/sysconfig/network-scripts/ifcfg-eth0. These are all essentially configuration files, too.
One of the simplest formats for a configuration file is to use the same syntax that the shell uses. This way, you can use the source (.) command to bring the configuration into the environment directly. ...