How should I test a selenium application with capybara? - selenium

Question

Here is my problem:
I am writing an application that uses selenium-webdriver to connect to a site a click/fill a bunch of things.
Obviously, I want to test my code... And this is where it gets difficult! How do I do this?
Here is my test:
require 'spec_helper'
require 'capybara/rspec'
module MyModule
Capybara.run_server = false
describe "the method", :type => :request do
it "should open a browser and go to the site" do
MyClass.open_site("http://google.com")
page.has_content? "Google"
end
end
end
Here is the code:
require 'selenium-webdriver'
module MyModule
class MyClass
def self.open_site(url)
#driver = Selenium::WebDriver.for :firefox
#driver.navigate.to url
end
end
end
Here is the error I am getting:
Failures:
1) the method should open a browser and go to the site
Failure/Error: page.has_content? "Google"
ArgumentError:
rack-test requires a rack application, but none was given
# (eval):2:in `has_content?'
# ./spec/integration/myclass_spec.rb:10
I can understand the test is confused because usually Capybara runs Selenium to vist a site and check that everything looks good. But here Selenium is running on its own as part of the code...
How could I tell rack-test to use the running Selenium as its application?
Is Capybara even the right solution to test this code?
Thanks for your help!

Answer 1

One of the functions that you are using must be using a rack application.
The problem should not be with the line:
page.has_content? "google"
because that works fine for me with selenium driver. I suspect it is the way you have set up the driver.
I had a similar problem with running my capybara tests until I discovered these posts:google groups
They gave me a few pointers to get my test running.
In the end I ended up having these lines to configure my tests.
I am using chrome but Internet explorer and firefox would work this way too.
require 'selenium-webdriver'
Capybara.register_driver :selenium_ie do |app|
Capybara::Selenium::Driver.new(app, :browser => :chrome)
end
Capybara.default_driver = :selenium_chrome
Capybara.app_host = 'http://www.google.com'

Answer 2

Here's what I did with my Sinatra application running MongoDB. Comment out the selenium code to try pure selenium without capybara. Install the test-unit, capybara, capybara-webkit, selenium-webdriver gems. For capybara-webkit do brew install qt4 (mac) or install qt4 another way.
require './app'
require 'test-unit'
require 'capybara'
require 'capybara-webkit'
require 'selenium-webdriver'
class IntegrationTest < Test::Unit::TestCase
include Capybara::DSL
def setup
# Clear database first
MongoMapper.database.collections.select { |c| c.name != 'system.indexes' }.each(&:drop)
# For pure selenium: #b = Selenium::WebDriver.for :firefox
# For pure selenium: #w = Selenium::WebDriver::Wait.new(:timeout => 15)
Capybara.run_server = false
Capybara.default_selector = :css
Capybara.default_wait_time = 5
Capybara.ignore_hidden_elements = false
Capybara.javascript_driver = :webkit # Comment out to use :selenium
Capybara.default_driver = Capybara.javascript_driver
Capybara.app = Sinatra::Application.new
Capybara.app_host = "http://crowdfundhq.dev:3001"
Capybara.server_port = 3001
end
def teardown
# For pure selenium: #b.quit
Capybara.reset_sessions!
Capybara.use_default_driver
end
def test_root
# For pure selenium: #b.get "http://crowdfundhq.dev:3001"
# For pure selenium: assert #b.page_source =~ /#pricing/
# Change driver during test: Capybara.current_driver = Capybara.javascript_driver
visit("/")
assert(page.body =~ /highlight/)
end
end