I am a big proponent of testing. I don’t subscribe to a particular discipline, I use what works and is suitable for the task. I think that sometimes doctorine can get in the way of shipping, however that is a blog post that warrants its own entry.
As stated in a GitHub Issue I want to have test suite for this blog. This site is a staticly generated.
A valid test or test suite would be to check the HTML and any internal links are valid. That would be easy enough to test. Read in any files that are publishable, check the markup via a few regexs. However these aren’t real world tests. Testing that the HTML is valid is no longer enough with the advent of javascript.
To create real world acceptance tests for my site, I use a few technologies:
- RSpec - Behaviour Driven Development for Ruby -
- Capybara - Test web applications by simulating how a real user would interact with my site
- Selenium - Selenium automates browsers
These let me set up a way to write tests (RSpec) and drive the browser (Selenium) using an acceptance test API (Capybara). This gets me most of the way there, but one problem - how do I serve the static site, so it can actually be tested?
Most popular Ruby applications (such as Ruby on Rails and Sinatra) use the Rack interface to connect web requests to Ruby code. Since Jekyll generates static pages, it doesn’t have a Rack interface by default. However rack-jekyll is a gem that bridges the gap, and allows us to serve my pages to Selenium so I can test it using the browser.
Gemfile
source 'https://rubygems.org'
gem 'jekyll'
group :development, :test do
gem 'capybara'
gem 'chromedriver-helper'
gem 'rack-jekyll'
gem 'rspec'
gem 'selenium-webdriver'
end
bundle install
brew install chromedriver
This installs what is needed.
Next generate the necessary configuration files for RSpec using rspec --init.
$ rspec --init
create .rspec
create spec/spec_helper.rb
This creates two files - .rspec is a configuration file for RSpec output, and spec/spec_helper.rb is the file that loads the tests and configures the environment. To setup RSpec to use Capybara and Rack::Jekyll the spec_helper.rb file needs to look like:
require 'capybara/rspec'
require 'rspec'
require 'rack/jekyll'
require 'rack/test'
RSpec.configure do |config|
config.expect_with :rspec do |expectations|
expectations.include_chain_clauses_in_custom_matcher_descriptions = true
end
config.mock_with :rspec do |mocks|
mocks.verify_partial_doubles = true
end
Capybara.register_driver :selenium do |app|
Capybara::Selenium::Driver.new(app, :browser => :chrome)
end
Capybara.app = Rack::Jekyll.new(force_build: true)
end
I created a directory and a file in the spec/ directory called
features/homepage__spec.rb. The *_spec.rb suffix is needed for all files to
let RSpec run it automatically.
describe 'the homepage', type: :feature, js: true do
before { visit '/' }
it 'has the page title' do
expect(find('.site-title').text).to eq('Stevie Codes')
end
end
Run bundle exec rspec. This will
- Build the Jekyll site
- Open a Chrome browser
- Run the test(s)
- Pause after the page loads.
I am going to use this to ensure my site works as expected. It obviously needs extended. The next step is to add it to some sort of CI workflow. From there then I will add more comprehensive tests. I think it is a good first step.
