I'm currently running a Rails 6 App with Rspec and Capybara. When running the system specs, rails automatically generates screenshots. This makes my tests slow. I would like to disable the screenshots. How do i disable screenshots?
spec_helper.rb
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
config.shared_context_metadata_behavior = :apply_to_host_groups
end
rails_helper.rb
require 'spec_helper'
ENV['RAILS_ENV'] ||= 'test'
require File.expand_path('../config/environment', __dir__)
abort("The Rails environment is running in production mode!") if Rails.env.production?
require 'rspec/rails'
begin
ActiveRecord::Migration.maintain_test_schema!
rescue ActiveRecord::PendingMigrationError => e
puts e.to_s.strip
exit 1
end
RSpec.configure do |config|
config.fixture_path = "#{::Rails.root}/spec/fixtures"
config.use_transactional_fixtures = true
config.infer_spec_type_from_file_location!
config.filter_rails_from_backtrace!
config.include FactoryBot::Syntax::Methods# config.filter_gems_from_backtrace("gem name")
end
Capybara.default_driver = :selenium_chrome_headless
Currently, the only way to disable screenshot is by including a before block like this:
require 'rails_helper'
RSpec.describe 'Hello world', type: :system do
before do
driven_by(:selenium_chrome_headless)
end
describe 'index page' do
it 'shows the right content' do
get hello_world_index_path
expect(page).to have('hello world')
end
end
end
I'm looking for a more sustainable way to disable the screenshots by default.
Rails calls the take_failed_screenshot by default in the teardown phase of the tests https://github.com/rails/rails/blob/c5bf2b4736f2ddafbc477af5b9478dd7143e5466/actionpack/lib/action_dispatch/system_testing/test_helpers/setup_and_teardown.rb#L8
I don't see any configuration to turn that off.
The method is defined here https://github.com/rails/rails/blob/c5bf2b4736f2ddafbc477af5b9478dd7143e5466/actionpack/lib/action_dispatch/system_testing/test_helpers/screenshot_helper.rb#L44-L46
Maybe you can try to override that method to not take the screenshot? something like this is you use minitest:
# test/application_system_test_case.rb
class ApplicationSystemTestCase < ActionDispatch::SystemTestCase
driven_by :selenium, using: :headless_chrome, screen_size: [1400, 1400]
def take_failed_screenshot
end
end
or if you use RSpec:
# spec/rails_helper.rb
module NoFailedScreenshots
def take_failed_screenshot
end
end
RSpec.configure do |config|
config.include(NoFailedScreenshots)
...
But one comment, if these are the screenshots you refer to, then you have a problem with too many failing specs if that slows things down, you should fix the specs instead.
If this is not what's happening to you then you probably have some custom setup taking screenshots that it's not standard Rails
Related
I added selenium webdriver to my test setup to start writing some tests based on javascript. I added
Capybara.register_driver :selenium do |app|
Capybara::Selenium::Driver.new(app, browser: :chrome)
end
Capybara.javascript_driver = :selenium_chrome
Capybara.configure do |config|
config.default_max_wait_time = 5
config.default_driver = :selenium
end
to my rails helper which now look like this
rails_helper.rb
require "spec_helper"
ENV["RAILS_ENV"] ||= "test"
require File.expand_path("../config/environment", __dir__)
# Prevent database truncation if the environment is production
abort("The Rails environment is running in production mode!") if Rails.env.production?
require "rspec/rails"
require "capybara/rspec"
Dir[Rails.root.join("spec", "support", "**", "*.rb")].each { |f| require f }
begin
ActiveRecord::Migration.maintain_test_schema!
rescue ActiveRecord::PendingMigrationError => e
puts e.to_s.strip
exit 1
end
RSpec.configure do |config|
config.infer_spec_type_from_file_location!
Capybara.register_driver :selenium do |app|
Capybara::Selenium::Driver.new(app, browser: :chrome)
end
Capybara.javascript_driver = :selenium_chrome
Capybara.configure do |config|
config.default_max_wait_time = 5
config.default_driver = :selenium
end
# Filter lines from Rails gems in backtraces.
config.filter_rails_from_backtrace!
# arbitrary gems may also be filtered via:
# config.filter_gems_from_backtrace("gem name")
# Include Factory Bot syntax methods
config.include FactoryBot::Syntax::Methods
# Configure Database Cleaner
config.before(:suite) do
DatabaseCleaner.strategy = :transaction
DatabaseCleaner.clean_with(:truncation)
end
config.around(:each) do |example|
DatabaseCleaner.cleaning do
example.run
end
end
end
After these changes, my previous feature tests broke because now they do not use the fixtures created within the tests.
i.e features/article_spec.rb
require "rails_helper"
RSpec.describe "articles", type: :feature do
before :each do
create(:article, published: false)
create(
:article,
name: "Published article",
body_text: "Lorem ipsum",
published_at: Time.now,
published: true
)
end
describe "article feature" do
it "shows only published articles" do
visit("/articles")
expect(page).to have_css(".article-container", count: 1)
end
end
end
The tests are not running headless so I can follow along and see that no articles are present despite the creation within article_spec.rb. I assume the driver executes the tests using another server and/or another database but I'm not sure. In which case don't know how to change that or how to seed that database. You guys have any clue?
You can't use an around hook for database cleaner with RSpec and Capybara because of the order the hooks get run in. If you're going to continue using database cleaner you need to be using closer to the recommended configuration - https://github.com/DatabaseCleaner/database_cleaner#rspec-with-capybara-example
That being said, since Rails 5.1 assuming you're using a normal configuration (testing with local app, normal DB config, up to date rspec-rails, etc) you don't need to use database cleaner at all, just set config.use_transactional_fixtures = true in your RSpec configuration, remove all references to database cleaner, and the DB connection should be safely shared between your tests and the app under test.
When I trying to run single test, like
bundle exec rspec spec/some_path/some_spec.rb
It runs all specs anyway, like it would be if I run bundle exec rspec spec/
Even when it print out failed specs and I only copy failed examples and put it again, the Rspec will run all existed specs. I haven't found any information about such behaviour in rspec documentation. Thank you very much if you know, what is wrong with my code! (or configuration, I suppose)
spec_helper.rb:
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
config.shared_context_metadata_behavior = :apply_to_host_groups
config.filter_run_when_matching :focus
config.disable_monkey_patching!
if config.files_to_run.one?
config.default_formatter = 'doc'
end
config.profile_examples = 10
config.order = :random
Kernel.srand config.seed
end
require 'webmock/rspec'
def body_as_json
json_str_to_hash response.body
end
def json_str_to_hash(str)
JSON.parse(str).with_indifferent_access
end
.rspec:
--require rails_helper
--format progress
--color
--order random
--profile 10
rails_helper.rb
# frozen_string_literal: true
unless ENV['COVERAGE'].to_s == ''
require 'simplecov'
SimpleCov.start 'rails'
end
ENV['RAILS_ENV'] ||= 'test'
require 'spec_helper'
require File.expand_path('../config/environment', __dir__)
require 'rspec/rails'
require 'capybara/rspec'
require 'capybara/rails'
require 'capybara/poltergeist'
require 'capybara-screenshot/rspec'
require 'timecop'
require 'sidekiq/testing'
require 'database_cleaner'
require 'webmock/rspec'
require 'audited-rspec'
Dir[Rails.root.join('spec/support/**/*.rb')].each { |f| require f }
ActiveRecord::Base.connection.reconnect!
ActiveRecord::Migration.maintain_test_schema!
RSpec.configure do |config|
config.infer_base_class_for_anonymous_controllers = false
config.include Capybara::DSL
config.include FactoryBot::Syntax::Methods
config.include CapybaraHelpers
config.include ProcessingHelpers
config.include RequestsHelper, type: :request
config.include ApiHelper
config.fixture_path = "#{::Rails.root}/spec/fixtures"
config.before(:suite) do
DatabaseCleaner.strategy = :transaction
DatabaseCleaner.clean_with :truncation, except: %w[spatial_ref_sys schema_migrations]
Faker::Config.locale = :en
# Rails.application.load_seed
Sidekiq::Testing.fake!
end
config.around do |example|
DatabaseCleaner.cleaning do
example.run
end
end
# Remove this line if you're not using ActiveRecord or ActiveRecord fixtures
# config.fixture_path = "#{::Rails.root}/spec/fixtures"
# If you're not using ActiveRecord, or you'd prefer not to run each of your
# examples within a transaction, remove the following line or assign false
# instead of true.
config.use_transactional_fixtures = false
# RSpec Rails can automatically mix in different behaviours to your tests
# based on their file location, for example enabling you to call `get` and
# `post` in specs under `spec/controllers`.
#
# You can disable this behaviour by removing the line below, and instead
# explicitly tag your specs with their type, e.g.:
#
# RSpec.describe UsersController, :type => :controller do
# # ...
# end
#
# The different available types are documented in the features, such as in
# https://relishapp.com/rspec/rspec-rails/docs
config.infer_spec_type_from_file_location!
# Filter lines from Rails gems in backtraces.
config.filter_rails_from_backtrace!
# arbitrary gems may also be filtered via:
# config.filter_gems_from_backtrace("gem name")
end
Shoulda::Matchers.configure do |config|
config.integrate do |with|
with.test_framework :rspec
with.library :rails
end
end
Capybara.register_driver :poltergeist do |app|
Capybara::Poltergeist::Driver.new(
app,
phantomjs_logger: Rails.root.join('log', 'poltergeist.log'),
inspector: true
)
end
Capybara.javascript_driver = :poltergeist
Capybara.server_port = 3001
Capybara.app_host = 'http://localhost:3001'
I am following the rails api book but building the code in an engine. The test is at spec/controllers/concerns/handicap/authenticable_spec.rb and looks like this
require 'spec_helper'
require_relative '../../../../app/controllers/concerns/handicap/authenticable.rb'
class Authentication
include Handicap::Authenticable
end
module Handicap
describe Authenticable, type: :controlller do
let(:authentication) { Authentication.new }
subject { authentication }
describe "#current_user" do
before do
#user = FactoryGirl.create :handicap_user
request.headers["Authorization"] = #user.auth_token
authentication.stub(:request).and_return(request)
end
it "returns the user from the authorization header" do
expect(authentication.current_user.auth_token).to eql #user.auth_token
end
end
end
end
When I run the test directly i.e. rspec ./spec/controllers/concerns/handicap/authenticable_spec.rb I get an error:
uninitialized constant Handicap::FactoryGirl
However, when I run all the tests i.e. rspec spec, it does find the FactoryGirl constant and the test fails with
undefined local variable or method `request' for #<RSpec::ExampleGroups::HandicapAuthenticable::CurrentUser:0x007ff276ad5988>.
According to this github issue, I need to add < ActionController::Base to the Authentication class i.e.
class Authentication < ActionController::Base
but if I add this in, I get
uninitialized constant ActionController
I have also tried adding < Handicap::ApplicationController but get
uninitialized constant Handicap::ApplicationController
There appears to be something wrong with my namespacing. There are three symptoms, the fact that FactoryGirl cannot be found if I run the test by itself, but is found when all the tests are run. The second is that it cannot find ActionController even when all the tests are run. The third is that I need to add the line:
require_relative '../../../../app/controllers/concerns/handicap/authenticable.rb'
to find the module that is being tested.
How do I fix my namespacing?
The rails_helper.rb file is
ENV['RAILS_ENV'] ||= 'test'
require File.expand_path('../dummy/config/environment.rb', __FILE__)
require 'rspec/rails'
require 'capybara'
require 'capybara/rails'
require 'capybara/rspec'
require 'capybara-screenshot'
require 'capybara-screenshot/rspec'
require 'capybara/poltergeist'
require 'capybara/email/rspec'
require 'pp'
require 'chris_api_helpers'
# Prevent database truncation if the environment is production
abort("The Rails environment is running in production mode!") if Rails.env.production?
require 'spec_helper'
require 'rspec/rails'
require 'factory_girl_rails'
ActiveRecord::Migration.maintain_test_schema!
Shoulda::Matchers.configure do |config|
config.integrate do |with|
# Choose a test framework:
with.test_framework :rspec
with.library :rails
end
end
RSpec.configure do |config|
config.fixture_path = "#{::Rails.root}/spec/fixtures"
config.use_transactional_fixtures = true
config.infer_spec_type_from_file_location!
# Filter lines from Rails gems in backtraces.
config.filter_rails_from_backtrace!
# force test migrations for db:migrate
ActiveRecord::Migration.maintain_test_schema!
Capybara::Screenshot.prune_strategy = { keep: 20 }
Capybara::Screenshot.append_timestamp = false
config.include FactoryGirl::Syntax::Methods
FactoryGirl.definition_file_paths << File.join(File.dirname(__FILE__), 'factories')
FactoryGirl.find_definitions
config.include Devise::Test::ControllerHelpers, type: :controller
end
and the spec_helper.rb is
require 'simplecov' if ENV["COVERAGE"]
SimpleCov.start do
add_filter '/spec/'
add_filter '/config/'
add_filter '/lib/'
add_filter '/vendor/'
add_group 'Controllers', 'app/controllers'
add_group 'Models', 'app/models'
add_group 'Helpers', 'app/helpers'
add_group 'Mailers', 'app/mailers'
add_group 'Views', 'app/views'
end if ENV["COVERAGE"]
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
config.shared_context_metadata_behavior = :apply_to_host_groups
#http://stackoverflow.com/questions/30859037/suppress-backtrace-for-rspec-3
config.backtrace_exclusion_patterns = [
/\/lib\d*\/ruby\//,
/bin\//,
/gems/,
/spec\/spec_helper\.rb/,
/lib\/rspec\/(core|expectations|matchers|mocks)/
]
end
You should not be putting specs into modules. This is the cause of problem here. If you need to reference a namespaced class, reference it like RSpec.describe Handicap::Authenticatable.
In general, when you are within a namespace and need to reference something explicitly from the 'root' scope, you can prepend it with double-colons. Such as:
module Handicap
class Something
def do_stuff
::FactoryGirl.create(:person)
end
end
end
It turns out that at the top of my file I should have had require 'rails_helper', not require 'spec_helper'``. All my other files hadrequire 'rails_helper'`` so when I ran the whole test suite, the rails_helper was being loaded anyway.
Embarrassing, but this Q&A might help someone else that has trouble spotting simple errors.
I have an RSpec feature spec that tests the login for my app. It passes when I run it in RSpec with Capybara, but when I try to run it flagged with js: true using Capybara-webkit, it fails. This is a problem because my entire app is behind the login, and if I can't get this bit to run I don't know how to do feature specs for the rest of the app.
Here's what I have tried:
Installing all the Capybara-webkit dependencies listed here. I'm running my app in a Docker container built on the ruby:2.3 image, which is built on Jessie.
Setting up DatabaseCleaner per this blog post. My database_cleaner.rb file is below.
Using the Headless gem (headless.rb below)
Running RSpec like so: xvfb-run -a bin/rspec spec/features/log_in_spec.rb (seems no different than running it normally with Headless)
How do I get my login specs to work under Capybara-webkit? Some of my specs will need to be flagged for JS and some won't, but they'll all need the user to be logged in. Thank you.
log_in_spec.rb
require 'rails_helper'
RSpec.feature "Log in", type: :feature do
scenario "as admin" do
user = create(:admin)
# Tried this instead of with Capybara, works with Capybara but not capybara-webkit
# login_as user, scope: :user, run_callbacks: false
visit root_path
fill_in 'Email', with: user.email
fill_in 'Password', with: user.password
find('.btn-primary').click
expect(page).to have_content('Admin')
end
end
spec_helper.rb
require 'capybara/rspec'
require 'paperclip/matchers'
RSpec.configure do |config|
Capybara.javascript_driver = :webkit
Capybara.app_host = 'https://192.168.99.101'
config.include Paperclip::Shoulda::Matchers
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
mocks.verify_doubled_constant_names = true
end
config.filter_run :focus
config.run_all_when_everything_filtered = true
config.disable_monkey_patching!
if config.files_to_run.one?
config.default_formatter = 'doc'
end
end
rails_helper.rb
ENV['RAILS_ENV'] ||= 'test'
require File.expand_path('../../config/environment', __FILE__)
# Prevent database truncation if the environment is production
abort("The Rails environment is running in production mode!") if Rails.env.production?
require 'spec_helper'
require 'rspec/rails'
require 'capybara/rails'
require 'devise'
require 'support/controller_macros'
Dir[Rails.root.join('spec/support/**/*.rb')].each { |f| require f }
# Checks for pending migrations before tests are run.
# If you are not using ActiveRecord, you can remove this line.
ActiveRecord::Migration.maintain_test_schema!
RSpec.configure do |config|
config.use_transactional_fixtures = false
config.filter_rails_from_backtrace!
config.include Warden::Test::Helpers
config.before :suite do
Warden.test_mode!
end
config.after :each do
Warden.test_reset!
end
end
# Added headless gem and this code thanks to this post: http://stackoverflow.com/a/28706535/3043668
if ENV['HEADLESS']
require 'headless'
headless = Headless.new
headless.start
at_exit { headless.stop }
end
spec/support/database_cleaner.rb
RSpec.configure do |config|
config.before(:suite) do
DatabaseCleaner.clean_with(:truncation)
end
config.before(:each) do
DatabaseCleaner.strategy = :transaction
end
config.before(:each, js: true) do
DatabaseCleaner.strategy = :truncation
end
config.before(:each) do
DatabaseCleaner.start
end
config.after(:each) do
DatabaseCleaner.clean
end
end
spec/support/headless.rb
RSpec.configure do |config|
config.around type: :feature do |example|
Headless.ly do
example.run
end
end
end
spec/support/capybara.rb
Capybara::Webkit.configure do |config|
config.debug = true
config.allow_unknown_urls
config.timeout = 5
config.ignore_ssl_errors
config.skip_image_loading
end
Here is a gist of the debug output from Capybara-webkit when I run the test. It looks like it's trying the same thing over and over.
UPDATE
I removed my Capybara.app_host setting, and the non-JS test still passes, but when I run it under capybara-webkit, I see this in the debig output:
Received 0 from "https://127.0.0.1:37193/login"
Page finished with false
Load finished
Page load from command finished
Wrote response false "{"class":"InvalidResponseError","message":"Unable to load URL: http://127.0.0.1:37193/login because of error loading https://127.0.0.1:37193/login: Unknown error"}"
Received "Reset()"
Started "Reset()"
undefined|1|SecurityError: DOM Exception 18: An attempt was made to break through the security policy of the user agent.
It's trying to visit("/login") and it is being redirected to the https version, and this is making it fail. How do I make it succeed?
The first reason is that the record after save procedure c=keep not the password in plain view, but seems that it is good in this case:
user = create(:admin)
# ...
user = User.first # wrong way
#...
fill_in 'Password', with: 'password'
The second reason in fail to login is that factory girl and capybara uses separate connections, so and data created in one session isn't available in another. To fix it use single connection patch (put it to spec/support) as described here.
This was a hard one. But the issue was that I had set force_ssl = true in my application.rb, stupidly, instead of putting it in production.rb and development.rb like a normal person.
I also had set Capybara-webkit's app_host, which, as it turned out, I did not need to do. After removing that, and running capybara-webkit with debug on, I saw that it was trying to redirect from http://localhost:45362/login (or whatever port) to https://localhost:45362/login (note the https!) and that this was causing a DOM 18 security error or whatever, and this was making it choke. I turned off force_ssl and now it works like a champ. Hope this helps you not tear your hair out.
For a while I was using Selenium / Spork / Rspec, with cache_classes on false, and everything seemed to be working.
In switching over to webkit, I've started to get errors related to cache_classes (e.g. Expected User, got User), so I've been fighting with it to try to get cache_classes set to true.
However no matter what I do, I end up with the following error:
Capybara::Driver::Webkit::WebkitInvalidResponseError:
Unable to load URL: http://127.0.0.1:56398/login
I have tried all kinds of things... including:
ActiveSupport::Dependencies.clear in both the prefork and each _run blocks
The code here: http://my.rails-royce.org/2012/01/14/reloading-models-in-rails-3-1-when-usign-spork-and-cache_classes-true/
Starting to wonder if I should just live with cache_classes = false, and figure out how to avoid the Factory girl errors. Any help would be appreciated. My spork file is as follows:
require 'spork'
Spork.prefork do
ENV["RAILS_ENV"] ||= 'test'
require 'rspec/rails'
require 'rspec/autorun'
require 'capybara/rails'
require 'capybara/rspec'
require 'database_cleaner'
require 'factory_girl'
require 'authlogic/test_case'
require 'email_spec'
include Authlogic::TestCase
require File.expand_path("../../config/environment", __FILE__)
RSpec.configure do |config|
# == Mock Framework
#
# If you prefer to use mocha, flexmock or RR, uncomment the appropriate line:
#
# config.mock_with :mocha
# config.mock_with :flexmock
# config.mock_with :rr
config.mock_with :rspec
ApplicationController.skip_before_filter :activate_authlogic
config.include(EmailSpec::Helpers)
config.include(EmailSpec::Matchers)
config.include Capybara::DSL
# Remove this line if you're not using ActiveRecord or ActiveRecord fixtures
#config.fixture_path = "#{::Rails.root}/spec/fixtures"
# If you're not using ActiveRecord, or you'd prefer not to run each of your
# examples within a transaction, remove the following line or assign false
# instead of true.
config.use_transactional_fixtures = false
# If true, the base class of anonymous controllers will be inferred
# automatically. This will be the default behavior in future versions of
# rspec-rails.
config.infer_base_class_for_anonymous_controllers = false
config.before(:suite) do
DatabaseCleaner.strategy = :truncation, {:except => %w[ages coefficients weights1 weights2 weights3 weights4 weights5 weights6 weights7 weights8 etfs]}
DatabaseCleaner.clean
end
config.before(:each) do
DatabaseCleaner.start
Capybara.current_driver = :webkit if example.metadata[:js]
#Capybara.current_driver = :selenium if example.metadata[:js]
activate_authlogic
ActiveRecord::Base.instantiate_observers
end
config.after(:each) do
DatabaseCleaner.clean
Capybara.use_default_driver if example.metadata[:js]
end
Capybara.default_selector = :css
end
# Requires supporting ruby files with custom matchers and macros, etc,
# in spec/support/ and its subdirectories.
Dir[Rails.root.join("spec/support/**/*.rb")].each {|f| require f}
## SUPPORT METHODS ##
## (erased for clarity) ##
## ##
ActiveSupport::Dependencies.clear
end
Spork.each_run do
#FactoryGirl.reload
# Required to fix a recurring error when testing Active_Admin stuff
# See here: http://railsgotchas.wordpress.com/2012/01/31/activeadmin-spork-and-the-infamous-undefined-local-variable-or-method-view_factory/
# Delete at some point if active admin or whoever fixes this
ActionView::Template.register_template_handler :arb, lambda { |template|
"self.class.send :include, Arbre::Builder; #_helpers = self; self.extend ActiveAdmin::ViewHelpers; #__current_dom_element__ = Arbre::Context.new(assigns, self); begin; #{template.source}; end; current_dom_context"
}
#ActiveSupport::Dependencies.clear
end
UPDATE : Adding an example spec just in case it helps....
describe "Items" do
before(:each) do
#user = Factory.create(:user)
activate_authlogic
b = # something not important
end
describe "usage paths" do
it "the form directly from the basic_simulation show page should have correctly functioning javascript validation", :js => true do
request_sign_in(#user) # This is a helper method which goes through the login form
visit '/basic_simulation'
fill_in "amount", :with => "-5000"
click_button "Calculate"
page.should have_selector("label.jquery-validator.amount-error", :text => "Please enter a value greater than or")
fill_in "amount", :with => "5000"
click_button "Calculate"
page.should have_selector("input#amount", :value => "5000")
end
end
You are having issues due to threading problems with Capybara-webkit and the test suite.
Jose Valim explains it much more clearly in a recent blog post.
If you follow his recommendations then you should be able to turn on transactional fixtures, remove database cleaner altogether and no longer have issues with your data during tests while using capybara-webkit. You'll get a nice boost in testing performance as well.
The trick though is to make sure that Jose's suggestion is in the Spork.each_run block or it will not work. For clarity here are the relevant parts of my spec_helper.rb.
require 'spork'
Spork.prefork do
ENV["RAILS_ENV"] ||= 'test'
require File.expand_path("../../config/environment", __FILE__)
require 'rspec/rails'
require 'capybara/rspec'
Dir[Rails.root.join("spec/support/**/*.rb")].each {|f| require f}
RSpec.configure do |config|
config.mock_with :rspec
config.fixture_path = "#{::Rails.root}/spec/fixtures"
config.use_transactional_fixtures = true
Capybara.default_driver = :rack_test
Capybara.default_selector = :css
Capybara.javascript_driver = :webkit
end
end
Spork.each_run do
if Spork.using_spork?
ActiveRecord::Base.instantiate_observers
end
require 'factory_girl_rails'
# Forces all threads to share the same connection, works on Capybara because it starts the web server in a thread.
class ActiveRecord::Base
mattr_accessor :shared_connection
##shared_connection = nil
def self.connection
##shared_connection || retrieve_connection
end
end
ActiveRecord::Base.shared_connection = ActiveRecord::Base.connection
end
A few other small suggestions:
If you are using the latest version of factory_girl_rails then you
should be using require factory_girl_rails in the Spork.each_run
block and require factory_girl should be removed from the prefork
The latest factory_girl_rails also no longer requires ActiveSupport::Dependencies.clear at all, although some people are still having issues without it so you should test removing it.
I'm still not sure about the need for ActiveRecord::Base.instantiate_observers but in any case you would only need it if you are using observers and I understand that it should be in the each_run block.
Try all that and see if it works for you.