.SH File Format

Introduction

A shell script is a text file of commands for a Unix shell, the program that reads what is typed at a terminal. The Bourne shell of 1979 set the pattern. The KornShell, Bash and Zsh extended it, and POSIX standardised a common core.

The extension is only a convention, and many scripts have none. .sh is the general name, usually for a script that any POSIX shell can run, while .bash, .zsh and .ksh mark scripts written for one shell.

Structure

A script is plain text with Unix line ends. The first line is often a shebang such as #!/bin/sh or #!/usr/bin/env bash, which tells the system which program should run the file when it has execute permission. A carriage return at the end of a line, from a file saved on Windows, becomes part of the last word on that line and breaks the script.

Each line is a command and its arguments, split on spaces. Commands connect with |, && and ||, and > and < redirect files. $name expands a variable, $1 onwards are the script’s arguments, and $(command) inserts a command’s output. Double quotes keep a value as one word. Comments start with #.

Control flow closes with reversed words: if and fi, case and esac, and do and done for loops. Arrays, [[ ]] tests and many other features exist in Bash, Zsh and KornShell but not in POSIX sh.

#!/bin/sh
# Report the size of each file named on the command line.
set -eu

for file in "$@"; do
    if [ -f "$file" ]; then
        size=$(wc -c < "$file")
        echo "$file: $size bytes"
    fi
done

Links