Learning Bash Properly

The problem

As a systems engineer I live in the terminal, but a lot of my Bash has been cargo-culted — grab a one-liner from Stack Overflow, tweak it until it works, move on. It runs, but I couldn’t always tell you why. I wanted to close that gap and treat the shell as a language I actually know rather than a slot machine I keep pulling until the output looks right.

The plan

Work through The Complete Bash Scripting Course end to end, then cement it the only way that sticks: rewrite my own homelab glue scripts properly. The TrueNAS backup jobs, the container healthchecks, the ad-hoc one-liners — all rebuilt with real error handling instead of hope and a trailing || true.

What I expect to learn

• The mechanics I’ve always fudged: quoting, word splitting, globbing, and why "$var" and $var are not the same thing
• Pipes, redirection, and process substitution as a composable toolkit rather than magic incantations
• Writing robust scripts — set -euo pipefail, trap for cleanup, meaningful exit codes, argument parsing with getopts
• The classic text tools done properly — grep, sed, awk, xargs, find — instead of reaching for Python every time
• Functions, loops, and conditionals, and the judgement to know when a task has outgrown Bash and wants a real language

What’ll probably trip me up

• The quoting and word-splitting rules that bite everyone — spaces in filenames, empty variables expanding to nothing
• set -e not behaving the way you’d expect inside functions, subshells, and pipelines
• Portability: bashisms vs POSIX sh, and the differences across the shells scattered around the homelab (OPNsense, TrueNAS, Alpine containers)

What’s next

• Rewrite the familyHub and TrueNAS backup + healthcheck scripts with proper error handling and logging
• Build a small personal library of reusable shell functions
• Run ShellCheck over everything as a standing habit