rspec-rails is a testing framework for Rails 3.x, 4.x and 5.0.
Use rspec-rails 1.x for Rails 2.x.
Add rspec-rails
to both the :development
and :test
groups in the
Gemfile
:
group :development, :test do
gem 'rspec-rails', '~> 3.5'
end
Want to run against the master
branch? You'll need to include the dependent
RSpec repos as well. Add the following to your Gemfile
:
%w[rspec-core rspec-expectations rspec-mocks rspec-rails rspec-support].each do |lib|
gem lib, :git => "https://github.com/rspec/#{lib}.git", :branch => 'master'
end
Download and install by running:
bundle install
Initialize the spec/
directory (where specs will reside) with:
rails generate rspec:install
This adds the following files which are used for configuration:
.rspec
spec/spec_helper.rb
spec/rails_helper.rb
Check the comments in each file for more information.
Use the rspec
command to run your specs:
bundle exec rspec
By default the above will run all _spec.rb
files in the spec
directory. For
more details about this see the RSpec spec file
docs.
To run only a subset of these specs use the following command:
# Run only model specs
bundle exec rspec spec/models
# Run only specs for AccountsController
bundle exec rspec spec/controllers/accounts_controller_spec.rb
Specs can also be run via rake spec
, though this command may be slower to
start than the rspec
command.
In Rails 4, you may want to create a binstub for the rspec
command so it can
be run via bin/rspec
:
bundle binstubs rspec-core
For detailed information on the general RSpec 3.x upgrade process see the RSpec Upgrade docs.
There are three particular rspec-rails
specific changes to be aware of:
- The default helper files created in RSpec 3.x have changed
- File-type inference disabled by default
- Rails 4.x
ActiveRecord::Migration
pending migration checks - Extraction of
stub_model
andmock_model
torspec-activemodel-mocks
Please see the RSpec Rails Upgrade docs for full details.
NOTE: Generators run in RSpec 3.x will now require rails_helper
instead
of spec_helper
.
Once installed, RSpec will generate spec files instead of Test::Unit test files
when commands like rails generate model
and rails generate controller
are
used.
You may also invoke RSpec generators independently. For instance,
running rails generate rspec:model
will generate a model spec. For more
information, see list of all
generators.
Once you've set up the environment, you'll need to cd into the working directory of whichever repo you want to work in. From there you can run the specs and cucumber features, and make patches.
NOTE: You do not need to use rspec-dev to work on a specific RSpec repo. You can treat each RSpec repo as an independent project. Please see the following files:
For rspec-rails
-specific development information, see
Use model specs to describe behavior of models (usually ActiveRecord-based) in the application.
Model specs default to residing in the spec/models
folder. Tagging any
context with the metadata :type => :model
treats its examples as model
specs.
For example:
require "rails_helper"
RSpec.describe User, :type => :model do
it "orders by last name" do
lindeman = User.create!(first_name: "Andy", last_name: "Lindeman")
chelimsky = User.create!(first_name: "David", last_name: "Chelimsky")
expect(User.ordered_by_last_name).to eq([chelimsky, lindeman])
end
end
For more information, see cucumber scenarios for model specs.
Use controller specs to describe behavior of Rails controllers.
Controller specs default to residing in the spec/controllers
folder. Tagging
any context with the metadata :type => :controller
treats its examples as
controller specs.
For example:
require "rails_helper"
RSpec.describe PostsController, :type => :controller do
describe "GET #index" do
it "responds successfully with an HTTP 200 status code" do
get :index
expect(response).to be_success
expect(response).to have_http_status(200)
end
it "renders the index template" do
get :index
expect(response).to render_template("index")
end
it "loads all of the posts into @posts" do
post1, post2 = Post.create!, Post.create!
get :index
expect(assigns(:posts)).to match_array([post1, post2])
end
end
end
For more information, see cucumber scenarios for controller specs.
Note: To encourage more isolated testing, views are not rendered by default in controller specs. If you are verifying discrete view logic, use a view spec. If you are verifying the behaviour of a controller and view together, consider a request spec. You can use render_views if you must verify the rendered view contents within a controller spec, but this is not recommended.
Use request specs to specify one or more request/response cycles from end to end using a black box approach.
Request specs default to residing in the spec/requests
, spec/api
, and
spec/integration
directories. Tagging any context with the metadata :type => :request
treats its examples as request specs.
Request specs mix in behavior from ActionDispatch::Integration::Runner, which is the basis for Rails' integration tests.
require 'rails_helper'
RSpec.describe "home page", :type => :request do
it "displays the user's username after successful login" do
user = User.create!(:username => "jdoe", :password => "secret")
get "/login"
assert_select "form.login" do
assert_select "input[name=?]", "username"
assert_select "input[name=?]", "password"
assert_select "input[type=?]", "submit"
end
post "/login", :username => "jdoe", :password => "secret"
assert_select ".header .username", :text => "jdoe"
end
end
The above example uses only standard Rails and RSpec APIs, but many RSpec/Rails users like to use extension libraries like FactoryGirl and Capybara:
require 'rails_helper'
RSpec.describe "home page", :type => :request do
it "displays the user's username after successful login" do
user = FactoryGirl.create(:user, :username => "jdoe", :password => "secret")
visit "/login"
fill_in "Username", :with => "jdoe"
fill_in "Password", :with => "secret"
click_button "Log in"
expect(page).to have_selector(".header .username", :text => "jdoe")
end
end
FactoryGirl decouples this example from changes to validation requirements, which can be encoded into the underlying factory definition without requiring changes to this example.
Among other benefits, Capybara binds the form post to the generated HTML, which means we don't need to specify them separately. Note that Capybara's DSL as shown is, by default, only available in specs in the spec/features directory. For more information, see the Capybara integration docs.
There are several other Ruby libs that implement the factory pattern or provide a DSL for request specs (a.k.a. acceptance or integration specs), but FactoryGirl and Capybara seem to be the most widely used. Whether you choose these or other libs, we strongly recommend using something for each of these roles.
Feature specs test your application from the outside by simulating a browser.
capybara
is used to manage the
simulated browser.
Feature specs default to residing in the spec/features
folder. Tagging any
context with the metadata :type => :feature
treats its examples as feature
specs.
Feature specs mix in functionality from the capybara gem, thus they require
capybara
to use. To use feature specs, add capybara
to the Gemfile
:
gem "capybara"
For more information, see the cucumber scenarios for feature specs.
By default Mailer specs reside in the spec/mailers
folder. Adding the metadata
:type => :mailer
to any context makes its examples be treated as mailer specs.
ActionMailer::TestCase::Behavior
is mixed into your mailer specs.
require "rails_helper"
RSpec.describe Notifications, :type => :mailer do
describe "notify" do
let(:mail) { Notifications.signup }
it "renders the headers" do
expect(mail.subject).to eq("Signup")
expect(mail.to).to eq(["to@example.org"])
expect(mail.from).to eq(["from@example.com"])
end
it "renders the body" do
expect(mail.body.encoded).to match("Hi")
end
end
end
For more information, see the cucumber scenarios for mailer specs .
View specs default to residing in the spec/views
folder. Tagging any context
with the metadata :type => :view
treats its examples as view specs.
View specs mix in ActionView::TestCase::Behavior
.
require 'rails_helper'
RSpec.describe "events/index", :type => :view do
it "renders _event partial for each event" do
assign(:events, [double(Event), double(Event)])
render
expect(view).to render_template(:partial => "_event", :count => 2)
end
end
RSpec.describe "events/show", :type => :view do
it "displays the event location" do
assign(:event, Event.new(:location => "Chicago"))
render
expect(rendered).to include("Chicago")
end
end
View specs infer the controller name and path from the path to the view
template. e.g. if the template is events/index.html.erb
then:
controller.controller_path == "events"
controller.request.path_parameters[:controller] == "events"
This means that most of the time you don't need to set these values. When spec'ing a partial that is included across different controllers, you may need to override these values before rendering the view.
To provide a layout for the render, you'll need to specify both the template and the layout explicitly. For example:
render :template => "events/show", :layout => "layouts/application"
Use this to assign values to instance variables in the view:
assign(:widget, Widget.new)
render
The code above assigns Widget.new
to the @widget
variable in the view, and
then renders the view.
Note that because view specs mix in ActionView::TestCase
behavior, any
instance variables you set will be transparently propagated into your views
(similar to how instance variables you set in controller actions are made
available in views). For example:
@widget = Widget.new
render # @widget is available inside the view
RSpec doesn't officially support this pattern, which only works as a
side-effect of the inclusion of ActionView::TestCase
. Be aware that it may be
made unavailable in the future.
# rspec-rails-1.x
assigns[key] = value
# rspec-rails-2.x+
assign(key, value)
This represents the rendered view.
render
expect(rendered).to match /Some text expected to appear on the page/
# rspec-rails-1.x
render
response.should xxx
# rspec-rails-2.x+
render
rendered.should xxx
# rspec-rails-2.x+ with expect syntax
render
expect(rendered).to xxx
Routing specs default to residing in the spec/routing
folder. Tagging any
context with the metadata :type => :routing
treats its examples as routing
specs.
require 'rails_helper'
RSpec.describe "routing to profiles", :type => :routing do
it "routes /profile/:username to profile#show for username" do
expect(:get => "/profiles/jsmith").to route_to(
:controller => "profiles",
:action => "show",
:username => "jsmith"
)
end
it "does not expose a list of profiles" do
expect(:get => "/profiles").not_to be_routable
end
end
route_for
from rspec-rails-1.x is gone. Use route_to
and be_routable
instead.
Helper specs default to residing in the spec/helpers
folder. Tagging any
context with the metadata :type => :helper
treats its examples as helper
specs.
Helper specs mix in ActionView::TestCase::Behavior. A helper
object is
provided which mixes in the helper module being spec'd, along with
ApplicationHelper
(if present).
require 'rails_helper'
RSpec.describe EventsHelper, :type => :helper do
describe "#link_to_event" do
it "displays the title, and formatted date" do
event = Event.new("Ruby Kaigi", Date.new(2010, 8, 27))
# helper is an instance of ActionView::Base configured with the
# EventsHelper and all of Rails' built-in helpers
expect(helper.link_to_event).to match /Ruby Kaigi, 27 Aug, 2010/
end
end
end
Several domain-specific matchers are provided to each of the example group types. Most simply delegate to their equivalent Rails' assertions.
- Available in all specs
- Primarily intended for controller specs
expect(object).to be_a_new(Widget)
Passes if the object is a Widget
and returns true for new_record?
- Delegates to Rails'
assert_template
- Available in request, controller, and view specs
In request and controller specs, apply to the response
object:
expect(response).to render_template("new")
In view specs, apply to the view
object:
expect(view).to render_template(:partial => "_form", :locals => { :widget => widget } )
- Delegates to
assert_redirect
- Available in request and controller specs
expect(response).to redirect_to(widgets_path)
- Delegates to Rails'
assert_routing
- Available in routing and controller specs
expect(:get => "/widgets").to route_to(:controller => "widgets", :action => "index")
Passes if the path is recognized by Rails' routing. This is primarily intended
to be used with not_to
to specify standard CRUD routes which should not be
routable.
expect(:get => "/widgets/1/edit").not_to be_routable
- Passes if
response
has a matching HTTP status code - The following symbolic status codes are allowed:
Rack::Utils::SYMBOL_TO_STATUS_CODE
- One of the defined
ActionDispatch::TestResponse
aliases::error
:missing
:redirect
:success
- Available in controller, feature, and request specs.
In controller and request specs, apply to the response
object:
expect(response).to have_http_status(201)
expect(response).not_to have_http_status(:created)
In feature specs, apply to the page
object:
expect(page).to have_http_status(:success)
Several rake tasks are provided as a convenience for working with RSpec. To run
the entire spec suite use rake spec
. To run a subset of specs use the
associated type task, for example rake spec:models
.
A full list of the available rake tasks can be seen by running rake -T | grep spec
.
If you want to customize the behavior of rake spec
, you may define your own
task in the Rakefile
for your
project.
However, you must first clear the task that rspec-rails defined:
task("spec").clear