A Practical Shell Scripting Introduction for JavaScript Engineers

Karolis Šarapnickis — Sat, 06 Nov 2021 13:53:10 +0000

Compact rundown though shell basics focusing on practical use cases in JavaScript projects.

When developing scripts, I always had a tendency to go for Node.js instead of shell, even for relatively simple things. Shell scripting has always been challenging for me, until I dedicated some time in learning the basics. Now I understand that, when used correctly, shell scripts can really simplify the logic. So, here are some core concepts and practical use cases that you can learn and hopefully apply in your projects.

Essential Commands

Here is a list of some of the more common shell commands that you'd encounter in JS projects:

echo - prints text terminal window (e.g., echo Build is complete!)
Files and folders:
- mkdir creates directories (e.g. create directories recursively mkdir -p dist/code)
- rm - removes files or directories (e.g. forcibly and recursively remove dist directory - rm -rf dist)
- cp copy files and directories (e.g., cp -f dist/404/index.html dist/404.html)
- mv move or rename files or directories (e.g., mv -f gen/robots.txt gen/manifest.json dist)
tar - archiving utility (e.g., create a gzipped archive and write it to a file - tar -czf bundle.tar.gz dist/)

There are many more commands (ls, cd, cat, etc.), but when encountered you can look the up in the manual pages (e.g., man tar). There's also a very nifty community-driven tool called tldr, that aims to simplify the beloved man pages with practical examples (e.g., tldr tar).

Essential Operators

Command Chaining Operators

Usage of a single command is rarely enough, so here some chaining operators:

&& (the AND operator) - executes second command only if the first one succeeds (e.g., yarn && yarn build && yarn publish)
; (the semicolon operator) - runs several commands, despite if previous one succeeded or not (e.g., yarn build && mv -f gen/sitemap.xml dist; echo Build done!)

There are many more shell operators, but these and npm-run-all should cover the majority of your chaining needs.

Output `>` and Output Append `>>` Operators

Both, output > and output append >> operators redirects content to a destination, but only >> appends to the target. E.g., creating a .env file in your CI pipeline:

echo "PORT=${PRODUCTION_PORT}" > .env
echo "API_URL=${PRODUCTION_API_URL}" >> .env

Command Substitution

Command substitution is a mechanism by which the shell performs a given set of commands and then exchanges their output in the place of the commands. E.g., combine Node.js script evaluation and command substitution to have JavaScript output in shell environment:

echo Version - $(node -e "console.log(require('./package.json').version)")

Conditional Statements

Just like JavaScript, shell scripts can have if statements. They can be written as both multi and single line statements. E.g., performing an optimized yarn install for CI environment only:

#!/bin/bash

if [[ $CI ]]; then
    yarn --cache-folder $PWD/.yarn --prefer-offline --frozen-lockfile
else
    yarn
fi

Environment Variables

Environment variables are a common way of passing dynamically configurable values. Here are some use cases:

Configurable environment values with defaults - e.g., optional PORT environment variable with defaults from npm config:

{
  "config": {
    "port": 1234
  },
  "srcipts": {
    "start": "serve -l ${PORT:-$npm_package_config_port}"
  }
}

Reusing .env file in shell sessions - e.g., a localhost release script for Lerna project that loads .env, and performs some necessary checks:

#!/bin/bash

set -a; source .env; set +a

if [[ ! $GH_TOKEN ]]; then
    echo "🚨 Missing GH_TOKEN env variable" && exit 1
fi

if [[ ! "$(npm whoami)" ]]; then
    echo "🚨 Not logged in to npm" && exit 1
fi

if [[ ! "$(git status --porcelain)" ]]; then
    git checkout main && git pull && yarn release
else
    git status --porcelain
    echo "🧹 Working directory not clean" && exit 1
fi

Conclusion

Shell scripting is a powerful and elegant way to perform some common operations. Dedicate some time and give it a proper try - it's a versatile skill to have at your tool belt. This blog post scratches only the surface of what's possible, so here are some resources for further reading:

How to Manage Multiple SSH Keys for GitHub, GitLab, etc.

Karolis Šarapnickis — Sun, 12 Sep 2021 17:15:56 +0000

Short tutorial on how to structure projects and setup multi-user Git SSH workflow.

When using a single machine with multiple git accounts, you might encounter some obstacles with SSH setup. The easiest way to achieve a multi-user setup is to structure git repositories by respective directories, e.g.:

projects
├── work
│   ├── enterprise-fiz-buz
│   └── ...
└── personal
    ├── karolis.sh
    └── ...

This short tutorial will help you set up SHH for such workflow.

Generating a new SSH key

To generate a new SSH key, use the provided template script (or check the docs):

export EMAIL=karolis.sarapnickis@work.com
export SSH_FILE=$HOME/.ssh/id_rsa_work
ssh-keygen -t rsa -b 4096 -C $EMAIL -f $SSH_FILE
ssh-add -K $SSH_FILE
echo "🔽 PUBLIC KEY BELOW 🔽" && cat $SSH_FILE.pub
pbcopy < $SSH_FILE.pub

What's left is to add the SSH key to your GitHub/GitLab/etc. account.

Multi-user Git SSH setup

The general idea is that you use specific ssh command based on your working directory with the help of git includes.

Update the ~/.gitconfig file:

[user]
  name = Karolis Šarapnickis
  email = pastas.k@gmail.com
[core]
  sshCommand = ssh -i ~/.ssh/id_rsa
[push]
  default = current
[includeIf "gitdir:~/projects/work/"]
  path = ~/.gitconfig.work

And the ~/.gitconfig.work file:

[user]
  email = karolis.sarapnickis@work.com
[core]
  sshCommand = ssh -i ~/.ssh/id_rsa_work

And that's it! Appropriate SSH setup will be used based on directory location, no extra actions are needed.

Forem: Karolis Šarapnickis