I'm writing some super simple integration specs, and while Capybara does simulate the browser/user interaction, I suspect that changing through several pages in one test causes issues. Mainly that the variable page doesn't get changed if you visit a few pages in a row. Here's what I mean:
require "rails_helper"
describe "The Initial Experience", :type => :feature do
it "User visits Narrowcontent" do
visit root_path
expect(page).to have_content "Sign In" && "Register"
click_on "Sign In"
expect(page).to have_content("Login Access")
within(".middle-login") do
fill_in "user_email", :with => ENV["admin_username"]
fill_in "user_password", :with => ENV["admin_pw"]
end
click_button 'Sign in'
save_and_open_page
#expect(page).to have_content("Dashboard")
end
end
I had to comment out the last expectation because it causes the test to fail. The keyword "Dashboard" DOES exist, but the problem is that after running click_button 'Sign In', the variable page still represents the old one, as if click_button 'Sign In'has no impact whatsoever.
Any tips?
Related
Tell me please,why does it happen?
I can't understand, if I write:
feature "Article Creation" do
#here i write (:all)
before(:all) do
sign_up_helper
end
I get the error:
Article Creation allows user to visit to creating article page
Failure/Error: fill_in :article_title, :with => 'test_title'
Capybara::ElementNotFound:
Unable to find field :article_title
or
1) Article Creation allows user to visit to article page
Failure/Error: expect(page).to have_content I18n.t('articles.articles_new')
expected to find text "New Article:" in "Toggle navigation Blog Rails New Contacts Sign in --- !ruby/hash:ActionController::Parameters controller: devise/sessions action: new {\"controller\"=>\"devise/sessions\", \"action\"=>\"new\"} nil You need to sign in or sign up before continuing. Sign in: Email Password Remember me Sign up Forgot your password?"
but, if I write:
feature "Article Creation" do
#here i write(:each)
before(:each) do
sign_up_helper
end
It's Ok. All tests works. My question -WHY?
This is my test:
*#before all test visitor signs up
#here I've changed :all and :each*
feature "Article Creation" do
before(:all) do
sign_up_helper
end
scenario "allows user to visit to article page" do
visit new_article_path
expect(page).to have_content I18n.t('articles.articles_new')
end
scenario "allows user to visit to created article page" do
visit new_article_path
fill_in :article_title, :with => 'test_title'
fill_in :article_text, :with => 'example_text'
click_button 'Save Article'
expect(page).to have_content 'example_text'
end
This is sign_up_helper method:
#spec/support/session_helper.rb
def sign_up_helper
visit new_user_registration_path
fill_in :user_email, :with => 'user#example.com'
fill_in :user_username, :with => 'mike'
fill_in :user_password, :with => 'secure123!##'
fill_in :user_password_confirmation, :with => 'secure123!##'
click_button 'Sign up'
end
This is html form:
<p>
<label for="article_title">Title</label><br/>
<input type="text" name="article[title]" id="article_title" />
</p>
<p>
<label for="article_text">Text</label><br/>
<textarea name="article[text]" id="article_text">
</textarea>
</p>
Environment for each test is set anew, I think. New session, cookies, etc. In many cases, even brand new users are generated. So one "global" login is not possible.
Even if it were possible, it would still be a problem, as it introduces spec order dependency which is bad. Imagine that one of your specs logs user out. Then each subsequent spec would fail, because user is not logged in anymore.
To prevent this, make sure that each spec sets its own environment as it needs it (user logins, method stubs, etc.), without relying on side-effects from previously executed specs (which may or may not persist).
I'm trying requests tests with Capybara and it doesnt seem to work. This is my test file:
describe "Sessions", :type => :request do
let(:company) { FactoryGirl.create(:company) }
let(:user) { FactoryGirl.create(:admin, company_id: company.id ) }
describe "login page" do
it "signs me in" do
visit '/users/sign_in'
within("#new_user") do
fill_in 'Email', :with => user.email
fill_in 'Password', :with => user.password
end
click_button 'Sign in'
expect(page).to have_content 'Agenda'
end
end
end
Throws the following Error:
Failure/Error: expect(page).to have_content 'Agenda'
expected #has_content?("Agenda") to return true, got false
Im not sure if the problem is at logging in or at redirecting. But if i change the last line in the test for this:
expect(page).to have_content 'Invalid email'
I get the same Error.
Thanks in advance.
UPDATE:
i'm using devise for login
Simplest thing is to use screenshot_and_save_page to take a screenshot of the page just before the failing test, maybe you can determine what the problem is there.
With rspec, it's quite clear how you should organise your unit specs. The directory structure inside spec is very similar to that found in the app directory, so model specs go in the model directory, controller specs go in the controller directory and so on.
But it's not so clear with integration testing. I have just one file pertaining to integration testing: spec/features/integration.rb
Is the idea to create one elaborate spec that tests every faculty of your application? Something like this:
require 'spec_helper'
describe "Everything", js: true do
before do
#user_0 = FactoryGirl.build(:user_0)
#user_1 = FactoryGirl.build(:user_1)
#user_2 = FactoryGirl.build(:user_2)
#user_3 = FactoryGirl.build(:user_3)
end
it "can create a user" do
visit root_path
click_link 'Sign In'
ap #user_0
fill_in('Email', with: #user_0.email)
fill_in('Password', with: #user_0.password)
click_button 'Sign in'
visit('/user_friendships')
end
it "can create a user" do
end
it "can create a user" do
end
it "can create a user" do
end
it "GET /root_path" do
visit root_path
page.should have_content("All of our statuses")
click_link "Post a New Status"
page.should have_content("New status")
fill_in "status_content", with: "Oh my god I am going insaaaaaaaaane!!!"
click_button "Create Status"
page.should have_content("Status was successfully created.")
click_link "Statuses"
page.should have_content("All of our statuses")
page.should have_content("Jimmy balooney")
page.should have_content("Oh my god I am going insaaaaaaaaane!!! ")
end
end
But a lot longer?
Should I use more than one file? How should I use the describe blocks? I'm only using one at the moment and that doesn't feel right.
The short answer is: No, it's not meant to go into one file.
Bigger projects split their acceptance test-suite into files based on feature sets. If you have a lot of tests, they are often split up into different directories. The way that you organize those tests is up to you. I have seen a lot of different approaches here. I tend to group spec with similar requirements on database setup, aka test-data.
If you want to have great guide for your rspec tests, go and have a look at this site: http://betterspecs.org/
It's usually desirable to split integration tests per page or flow in your application.
For example, you could try:
# spec/users/sign_in_spec.rb
RSpec.describe 'Users Sign In', js: true do
let!(:user) { FactoryGirl.build(:user_0) }
it "can sign in" do
visit root_path
click_link 'Sign In'
fill_in('Email', with: user.email)
fill_in('Password', with: user.password)
click_button 'Sign in'
page.should have_content('Signed In')
end
end
# spec/statuses/manage_status.rb
RSpec.describe 'Statuses', js: true do
it "should be able to post a status" do
visit root_path
page.should have_content("All of our statuses")
click_link "Post a New Status"
fill_in "status_content", with: "Everything is Alright"
click_button "Create Status"
page.should have_content("Status was successfully created.")
click_link "Statuses"
page.should have_content("All of our statuses")
page.should have_content("Everything is Alright ")
end
end
If you are looking for a way to organize your specs, you could use Capybara Test Helpers to encapsulate the code and make the tests more readable. For example:
# spec/users/sign_in_spec.rb
RSpec.describe 'Users Sign In', js: true, test_helpers: [:login] do
let!(:user) { FactoryGirl.build(:user_0) }
it "can sign in" do
visit root_path
login.enter_credentials(email: user.email, password: user.password)
current_page.should.have_content('Signed In')
end
end
# spec/statuses/manage_status.rb
RSpec.describe 'Statuses', js: true, test_helpers: [:statuses] do
it "should be able to post a status" do
visit root_path
current_page.should.have_content('All of our statuses')
statuses.post('Everything is Alright')
statuses.should_now.have_status('Everything is Alright')
end
end
I am writing some integration test cases for an existing application. My test works fine if there is only one 'it' block. However, If I add more than one 'it' block it throws an error. Below is my code that works:
require 'spec_helper'
describe 'Group' do
before do
visit 'http://groups.caremonkey.com/users/sign_in'
fill_in "Email", :with => "email#example.com"
fill_in "Password", :with => "password"
click_button "Login"
page.should have_link('Account')
end
it 'Should check all the links and functionality of groups' do
#add new subgroup with valid data should save a new group
find("#group-squares").click_link("Add")
fill_in "Group Name", :with => "Melbourne futsal"
click_on("Save")
page.should_not have_content("can't be blank")
page.execute_script("parent.$.fancybox.close();")
page.should have_link('Account')
#test edit group: should be able to update group info provided valid data are given
first(".actual img").click
page.should have_content("Group")
page.should have_link("Cancel")
fill_in "Group name", :with => "Futsal club"
page.execute_script("$('#sub-group-color-options').find('.color23').click()")
click_button "Save"
click_on("Cancel")
page.should have_link('Account')
end
end
It works perfectly fine when I put all the 'it' block together in a single 'it' block. But when I split them in different 'it' block, it stops working. For example if I split this ("test edit group: should be able to update group info provided valid data are given") test case into separate 'it' block as follows
require 'spec_helper'
describe 'Group' do
before do
visit 'http://groups.caremonkey.com/users/sign_in'
fill_in "Email", :with => "email#example.com"
fill_in "Password", :with => "password"
click_button "Login"
page.should have_link('Account')
end
it 'add new subgroup with valid data should save a new group' do
find("#group-squares").click_link("Add")
fill_in "Group Name", :with => "Melbourne futsal"
click_on("Save")
page.should_not have_content("can't be blank")
page.execute_script("parent.$.fancybox.close();")
page.should have_link('Account')
end
it 'should be able to update group info provided valid data are given' do
first(".actual img").click
page.should have_content("Group")
page.should have_link("Cancel")
fill_in "Group name", :with => "Futsal club"
page.execute_script("$('#sub-group-color-options').find('.color23').click()")
click_button "Save"
click_on("Cancel")
page.should have_link('Account')
end
end
then rspec fails, it passes the first test, however second test gets failed throwing following error.
Failure/Error: visit 'http://groups.caremonkey.com/users/sign_in'
ActionController::RoutingError:
No route matches [GET] "/users/sign_in"
One more thing, I have to test all the features in remote(url: http://groups.caremonkey.com/). Because, I am writing integration tests for an existing application. In addition, I need to login to the system before I test rest of the features of my application. Thanks in advance for your help.
Have you followed the Capybara documentation for calling remote servers? It says you should have the following:
Capybara.current_driver = :selenium # Or anything but rack_test, probably
Capybara.run_server = false # Don't run your app in-process
Capybara.app_host = 'http://groups.caremonkey.com/'
My guess is that when you have visited the site once, future visit calls are trying to use relative routes, which then is routed to the default server. I can't think why you would get a ActionController::RoutingError if you don't have some kind of Rack server running. Are you running these tests in some other Rails application?
I guess something like this:
require 'spec_helper'
describe 'Group' do
before do
visit 'http://groups.caremonkey.com/users/sign_in'
fill_in "Email", :with => "email#example.com"
fill_in "Password", :with => "password"
click_button "Login"
page.should have_link('Account')
find("#group-squares").click_link("Add") #apperently both specs are "scoped" to this page
end
it 'Should check all the links and functionality of groups' do
fill_in "Group Name", :with => "Melbourne futsal"
click_on("Save")
page.should_not have_content("can't be blank")
page.execute_script("parent.$.fancybox.close();")
page.should have_link('Account')
end
it "test edit group: should be able to update group info provided valid data are given"
first(".actual img").click
page.should have_content("Group")
page.should have_link("Cancel")
fill_in "Group name", :with => "Futsal club"
page.execute_script("$('#sub-group-color-options').find('.color23').click()")
click_button "Save"
click_on("Cancel")
page.should have_link('Account')
end
end
My gut feeling tells me both test need the follow this: find("#group-squares").click_link("Add") so I added it to the before block This test however is cryptic, what is first(".actual img")?
I have two capybara tests, the first of which signs in a user, and the second which is intended to test functions only available to a logged in user.
However, I am not able to get the second test working as the session is not being maintained across tests (as, apparently, it should be).
require 'integration_test_helper'
class SignupTest < ActionController::IntegrationTest
test 'sign up' do
visit '/'
click_link 'Sign Up!'
fill_in 'Email', :with => 'bob#wagonlabs.com'
click_button 'Sign up'
assert page.has_content?("Password can't be blank")
fill_in 'Email', :with => 'bob#wagonlabs.com'
fill_in 'Password', :with => 'password'
fill_in 'Password confirmation', :with => 'password'
click_button 'Sign up'
assert page.has_content?("You have signed up successfully.")
end
test 'create a product' do
visit '/admin'
save_and_open_page
end
end
The page generated by the save_and_open_page call is the global login screen, not the admin homepage as I would expect (the signup logs you in). What am I doing wrong here?
The reason this is happening is that tests are transactional, so you lose your state between tests. To get around this you need to replicate the login functionality in a function, and then call it again:
def login
visit '/'
fill_in 'Email', :with => 'bob#wagonlabs.com'
fill_in 'Password', :with => 'password'
fill_in 'Password confirmation', :with => 'password'
click_button 'Sign up'
end
test 'sign up' do
...
login
assert page.has_content?("You have signed up successfully.")
end
test 'create a product' do
login
visit '/admin'
save_and_open_page
end
Each test is run in a clean environment. If you wish to do common setup and teardown tasks, define setup and teardown methods as described in the Rails guides.