shell: What is the purpose of ${var:-} when var is unset or null?

Question

In my Linux Mint 17.2 /etc/bash.bashrc I see the following:

# set variable identifying the chroot you work in (used in the prompt below)
if [ -z "${debian_chroot:-}" ] && [ -r /etc/debian_chroot ]; then
    debian_chroot=$(cat /etc/debian_chroot)
fi

This is the first reference to the token debian_chroot.

Why does this code use ${debian_chroot:-} instead of just $debian_chroot?

Bash's Shell Parameter Expansion says:

${parameter:-word}

If parameter is unset or null, the expansion of word is substituted. Otherwise, the value of parameter is substituted.

Here, "word" is null, so why bother substituting null for null?

Answer 1

The syntax ${debian_chroot:-} prevents the shell from exiting if it is running with set -u (crash when using undefined variables) and debian_chroot is unset at that point.

You don't want a normal interactive shell to have set -u (it would crash too easily), but it can be very useful in scripts.

To see this:

bash -c 'set -u; [ -z $a ]; echo ok'          # error
bash -c 'set -u; a=; [ -z $a ]; echo ok'      # ok
bash -c 'set -u; [ -z ${a:-} ]; echo ok'      # ok
bash -c 'set -u; a=; [ -z ${a:-} ]; echo ok'  # ok

Answer 2

The use of the "${variable:-}" notation protects the script from an error if the shell is somehow invoked with -u or executes set -u — that causes a complaint when you use an undefined variable.

-u Treat unset variables and parameters other than the special parameters ‘@’ or ‘*’ as an error when performing parameter expansion. An error message will be written to the standard error, and a non-interactive shell will exit.