set -u (Undefined Variables) and set -x (Trace Mode)

Catching Unset Variables

With set -u active, referencing a variable that was never assigned causes an immediate error and non-zero exit, which helps catch typos in variable names early instead of them silently expanding to nothing.

Example: Catching Unset Variables

#!/bin/bash
(
    set -u
    echo "$defined_var" 2>/dev/null
) ; defined_var="hello"
(
    set -u
    echo "now it's set: $defined_var"
)

Safely Demonstrating set -u's Error

Running the failing case inside a subshell and checking its overall result with || lets a script show what set -u catches without letting that failure end the whole script.

Example: Safely Demonstrating set -u's Error

#!/bin/bash
(set -u; echo "$totally_undefined_variable") 2>/dev/null || echo "caught: set -u stopped an unset variable reference"

Providing Defaults to Avoid set -u Errors

${var:-default} expands to default whenever var is unset (or empty), which is the standard way to make a script compatible with set -u while still allowing optional variables, like positional parameters that may not be given.

Example: Providing Defaults to Avoid set -u Errors

#!/bin/bash
set -u
name="${1:-anonymous}"
echo "Hello, $name"

Tracing Execution with set -x

set -x makes bash print each command to stderr, prefixed with +, right before running it, with variables already expanded to their actual values. This is one of the fastest ways to see exactly what a script is doing.

Note: Redirect the trace output separately with exec 2>trace.log if you want to keep it out of the normal terminal output.

Warning: set -x can print sensitive values like passwords or tokens; avoid it around code that handles secrets.

Example: Tracing Execution with set -x

#!/bin/bash
set -x
num=5
double=$((num * 2))
set +x
echo "tracing disabled again; double is $double"