I watched the latest and very helpful Railscast episode this morning - on Spork and wound up tweaking all the projects I’m working on to incorporate some of the things I learned in that screencast.

I also wound up tweaking a checklist I maintain in an Evernote entry of what I do to setup a Rails project. Of the projects I’m working on, I wind up putting them all into a similar environment and figured it would be a useful blog post. So here it is:

  • Create a git repository: Git is pretty awesome. If you don’t care who sees your app, you can stick it up on Github for free! This is a no brainer.
    • Github: Cheap and easy for a team of developers
    • Roll your own: I have a Linode slice, and put up a private git repository
  • Create a generic .gitignore file for the app. Here’s what I start with (on MacOS):
    • .bundle
    • db/*.sqlite3
    • log/*.log
    • tmp/restart.txt
    • tmp/*
    • .DS_Store
    • public/system/documents*
    • .sass-cache/*
    • .powenv # I use pow
  • Use RVM: RVM helps you manage multiple version of Rubies and related gem environments. I try to use one rvm for all my “rails 3.1 apps that could end up on heroku” and call it rails31
    • Add a .rvmrc file to the root directory: You should do this before you run rails new because you want to make sure you’re using the version of ruby rails you want. This means you should have a ruby and gemset created already via rvm.
  • Create the app: rails new <app_name> (I create so few apps, I always have to look this up - although now that I just wrote it in a blog entry, I’ll probably never have to do it again.)
  • Add the gems you want to work with. I’ve always add a few to support BDD/TDD.
    • pg: Postgres is the default database on heroku, where most of my apps end up. On top of that, I like the pg admin tool on mac os. Much more pleasant that mysql or using sqlite.
    • metric_fu: Based on advice from the Railscast Metrics, Metrics, Metrics. Simplecov is great.
    • formtastic: great for form processing
    • rspec-rails: I like RSpec over TestUnit
    • cucumber-rails: I also like BDD - even though it can be expensive in terms of time to run a test suite.
    • capybara
    • nokogiri: helpful with cucumber and capybara
    • factory_girl_rails: test factories
  • Add gems for continous testing: Rails 3.1 takes a few seconds to load - and I’d rather adopt tools that make it easy to re-run tests without delays:
    • database-cleaner: help with speeding up test runs
    • spork: a distributed ruby server environment
    • guard: supports continuous testing
    • guard-spork: helps to configure guard and spork
    • guard-rspec: helps to configure guard and rspec (don’t forget the cli options :all_on_start => false)
    • guard-cucumber: helps to configure guard and cucumber rb-fsevent - a listener for guard
  • Did you Check in yet? Check in early and often
  • Change your spec_helper.rb file and move functionality up into the Spork pre fork block.
  • Change your environment.rb to allow for factory girl and cucumber integration (of lots of borrowable step definitions)
require 'factory_girl/step_definitions'
  • Create a database:
  • Change your database.yml:
adapter: postgresql
database: <your_db_name>
username: postgres
password: <your_password>
pool: 5
timeout: 5000
host: localhost
  • Install RSpec
rails generate rspec:install
  • Install Cucumber
rails generate cucumber:install --capybara --rspec --spork
  • Add a model
  • Migrate your database(s)
rake:db:create
rake:db:migrate
rake:db:test:prepare
  • Make sure the site runs with one of your object controllers being the default home page configure everything for running tests quickly and automatically
# bootstrap spork
spork --bootstrap
# configure spec helper to use database cleaner
guard init spork
guard init rspec
guard init cucumber
# make sure spark block comes first
# make sure rspec and cucumber both startup (in guard) with the --drb option
  • Add a project in Pivotal Tracker (or some kanban tool) to manage your stories and velocity
    • I’m partial to pivotal tracker, it’s much faster than your typical kanban tools
  • BDD TDD in your features
  • Profit!

Now this post doesn’t really delve into configuration for most of these setup options, but there’s a ton of great resources on the web. Feel free to comment with your own indispensable configuration…