Articles

Getting Started with RSpec

Fri, 03 Dec 2010 00:29:46 +0000

I'll admit it, I don't test as much as I should. I understood the importance of Test Driven Development but never invested time into getting up to speed with RSpec, or any other testing tool for that matter.

I decided to commit to using RSpec for testing and development of one of my latest projects and I will say it has been an amazing experience. There is a real feeling of comfort knowing that the change you just made is not only already covered in tests but also has not broken anything else.

The docs are pretty good for RSpec so I'm going to just cover getting a Rails 3 app setup to use RSpec 2, FactoryGirl, Mocha and Autotest. After those gems are installed and setup, there is plenty of documentation for you to bury yourself in!

Setup

We are going to write a book store app and keep it really simple so we can focus on getting the testing portion setup correctly.

rails new bookstore cd bookstore rm public/index.html rm -rf test rake db:migrate

Open up your Gemfile and add the testing libraries we will be using to the development group.

Gemfile

source 'http://rubygems.org' gem 'rails', '3.0.2' gem 'sqlite3-ruby', :require => 'sqlite3' gem 'nifty-generators' group :development, :test do gem "factory_girl" gem 'factory_girl_rails' gem 'autotest' gem 'rspec' gem 'rspec-rails' gem "mocha" end

Run bundle install to get the new gems installed and run the rspec generator.

bundle install rails generate rspec:install

The generator will create a directory called /spec under the root app directory as well as a spec/spec_helper.rb file.

RSpec Helper

Let's customize the spec_helper.rb file since we are using Mocha for mocking and FactoryGirl instead of fixtures.

spec/spec_helper.rb

# Comment Out Line 18 # config.mock_with :rspec # And add this in it's place config.mock_with :mocha # Comment out line 21 since we are # using FactoryGirl and not fixtures # config.fixture_path = "#{::Rails.root}/spec/fixtures"

Now we are ready to generate our book resource. I'm going to use NiftyGenerators to generate the book model which is especially nice since it will generate the correct RSpec controller and model tests!

rails g nifty:layout rails g nifty:scaffold Book title:string author:string published:boolean rake db:migrate

Let's modify our routes to make the book resource our root controller.

config/routes.rb

root :to => "books#index"

If you open your spec directory you will notice that NiftyGenerator has created a controller, models and fixtures directory for us. Then on top of that there are specs in those directories that provide a great base!

If we run rake spec now we will see that we have 10 examples, 0 failures.

Autotest

So far we have a good base and can already use Autotest by running autotest in terminal.

Coming Up

In the next article we will replace our fixtures with FactoryGirl. There is also a Railscast about replacing your fixtures with factories which I highly recommend!

Less Rails More Migrate

Fri, 01 Oct 2010 18:50:19 +0000

I’ve recently needed a way to migrate multiple databases, all with different connections and migrations, with a single rake task. Our first solution was to just create a new rails project for each database, setup our database.yml file as normal and drop the migrations into db/migrate.

We could have called it there and been done with it. But I’m lazy. Having to go into each directory and run rake db:migrate was not going to let me reach the level of lazy I strive for. Not to mention all of the bloat of having a full rails projects hanging around just for the migrations.

Let’s make it lazy friendly ™.

The Setup

I first created the directory structure that would hold each databases config file and migrations.

migrator
 -> db
   -> configs
   -> migrations
 -> logs

For each database that needed to be migrated I created the following files:

migrator/configs/sample_database_1.yml

adapter: mysql database: sample_database_name username: supersecret password: plaintext host: localhost

Just the standard database.yml file without the development, production, etc top level nodes since this would only be run in one environment.

Folder in migrator/migrations named after the config yml

Using the example above then the folder would be /migrator/migrations/sample_database_1. In that folder we could drop all of our migration files (001_first_migration.rb, 002_second_migration, etc).

The meat of our lazy sandwich

A rake file to run all these migrations.

migrator/Rakefile

require 'active_record' require 'yaml' task :default => :migrate desc "Migrate the database(s) in db/migrations." task :migrate do migration_dir = Dir.new(File.join(File.dirname(__FILE__),"db","migrations")) config_dir = Dir.new(File.join(File.dirname(__FILE__),"db","configs")).path logs_dir = Dir.new(File.join(File.dirname(__FILE__), "logs")).path migration_dir.each do |d| next if d.first == '.' puts "Migrating '#{d}'..." begin ActiveRecord::Base.establish_connection(YAML::load(File.open(File.join(config_dir, "#{d}.yml")))) ActiveRecord::Base.logger = Logger.new(File.open(File.join(logs_dir, "#{d}.log"), 'a')) ActiveRecord::Migrator.migrate(File.join(migration_dir.path, d), ENV["VERSION"] ? ENV["VERSION"].to_i : nil) puts "Done" rescue Exception => e puts "\n\n#{e.to_s}\n" end end end

Run It

Now to run our new migrator just go into the migrator directory and run

> rake

You can also pass it a version if you want to run migrate to a certain version. Just keep in mind it will migrate all of the databases to the same version. Great for the solution I needed maybe not great for yours.

I've put together a tar file that has the rake file and directory structure already setup. Grab it from here

Using mod-rewrite to return default images

Fri, 01 Oct 2010 15:26:00 +0000

The Problem

We have a web app running that displays pages of thumbnails. Instead of having a rails action that would return the thumbnail or a default image if no thumbnail was available, I used Apache’s mod_rewrite to do it for me.

The behavior I was looking for was if the image requested existed then pass the request through. If the image requested did not exist on the file system then rewrite the request to /images/image_not_available.jpg

The Rewrite Rules

The rewrite rules I needed were simple and may not work for anyone elses need but all of our thumbnails are stored as /thumbnails/image_filename.jpg

Using RewriteCond -f then I can check if the requested file name exists on the file system.

RewriteCond %{DOCUMENT_ROOT}%{REQUEST_FILENAME} !-f RewriteRule ^(/thumbnails/.+)\.jpg$ /images/image_not_available.jpg [L]

Saved me the overhead of having a Rails action that does the same thing.