Unix in a Nutshell, 4th Edition

Functions

A shell function is a grouping of commands within a shell script. Shell functions let you modularize your program by dividing it up into separate tasks. This way the code for each task need not be repeated every time you need to perform the task. The POSIX shell syntax for defining a function follows the Bourne shell:

name () {
        function body's code come here
    }

Functions are invoked just as are regular shell built-in commands or external commands. The command line parameters $1, $2, and so on receive the function’s arguments, temporarily hiding the global values of $1, etc. For example:

    # fatal --- print an error message and die:

    fatal () {
        echo "$0: fatal error:" "$@" >&2     # messages to standard error
        exit 1
    }
    ...
    if [ $# = 0 ]    # not enough arguments
    then
        fatal not enough arguments
    fi

A function may use the return command to return an exit value to the calling shell program. Be careful not to use exit from within a function unless you really wish to terminate the entire program.

Bash and the Korn shell allow you to define functions using an additional keyword, function, as follows:

    function fatal {
        echo "$0: fatal error:" "$@" >&2     # messages to standard error
        exit 1
    }

When working with the different shells and defining functions, there are semantic differences that should be kept in mind:

• In Bash, all functions share traps with the “parent” shell (except the DEBUG trap, if function tracing has been turned on). With the errtrace option enabled (either set -E or set -o errtrace ...