CSC/ECE 517 Spring 2022 - E2206: Testing for users controller
About Expertiza
Expertiza is the software benefits for both instructors and students by supporting various types of submissions and providing reusable objects for peer review. It is an open-source project based on Ruby on Rails framework. It allows the instructors not only to create and customize new or existing assignments but also to create a list of topics the students can sign up for. Students can form teams to work on various projects and assignments. Expertiza also lets students peer-review other students' submissions, enabling them to work together to improve others' learning experiences.
Current Project Description
This Expertiza OSS project "E2206: Testing for users_controller" is focused on adding unit tests for users_controller.rb.
Files Involved
- users_controller.rb
- users_controller_spec.rb
Running Tests
rspec ./spec/controllers/users_controller_spec.rb
Testing
users_controller had 5 methods for which unit tests were missing as seen in this coverall. The RSpec unit tests added for these methods in the spec file 'users_controller_spec.rb' are discussed below.
action_allowed?
This method checks if a particular action is allowed when it is called for the current user. The action is passed in the params. In this case we are testing the action list_pending_requested. We stub a student, an instructor and an admin, and we test if only the admin has the right to perform this action.
context '#set_anonymized_view' do it 'admin list_pending_requested' do params = { action: 'list_pending_requested' } allow(controller).to receive(:params).and_return(params) expect(controller.action_allowed?).to be false stub_current_user(student7, student7.role.name, student7.role) allow(controller).to receive(:params).and_return(params) expect(controller.action_allowed?).to be false stub_current_user(instructor, instructor.role.name, instructor.role) allow(controller).to receive(:params).and_return(params) expect(controller.action_allowed?).to be false stub_current_user(admin, admin.role.name, admin.role) allow(controller).to receive(:params).and_return(params) expect(controller.action_allowed?).to be true end end
auto_complete_for_user_name
For auto_complete_for_user_name method there is another method auto_complete_result being called in the controller. This method does not exist, and thus when auto_complete_for_user_name method is called, it raises ActionView::Template::Error. In the first test case, it is being checked if the correct error is being thrown. While, in the second test case the method is being called with a GET request and is expected to redirect us to the home page.
context '#auto_complete_for_user_name' do it 'checks if auto_complete returns actionview error' do stub_current_user(student1, student1.role.name, student1.role) session = { user: student1 } @params = {user: student1} allow(controller).to receive(:params).and_return(@params) expect{controller.auto_complete_for_user_name}.to raise_error(ActionView::Template::Error) end it 'checks if get auto_complete redirects to test host' do stub_current_user(student1, student1.role.name, student1.role) session = { user: student1 } @params = {user: student1} allow(controller).to receive(:params).and_return(@params) get :auto_complete_for_user_name, @params, session expect(response).to redirect_to("http://test.host/") end end
destroy
This action is called in the process of deleting a user. When this action is successfully executed or user is successfully destroyed (when User is passed a valid value 'student1') a note alert is flashed, whereas in case user is not destroyed (when User is passed an invalid value 'nil') an error is flashed. Further, it is being tested that at the end in both cases it redirects to list action.
context '#destroy' do it 'check if user was successfully destroyed' do allow(User).to receive(:find).with(any_args).and_return(student1) allow(AssignmentParticipant).to receive(:delete).with(any_args).and_return(true) allow(TeamsUser).to receive(:delete).with(any_args).and_return(true) allow(AssignmentQuestionnaire).to receive(:destroy).with(any_args).and_return(true) allow(User).to receive(:destroy).with(any_args).and_return(true) post :destroy expect(flash[:note]).to be_present end it 'check if user was not successfully destroyed' do allow(User).to receive(:find).with(any_args).and_return(nil) allow(AssignmentParticipant).to receive(:delete).with(any_args).and_return(true) allow(TeamsUser).to receive(:delete).with(any_args).and_return(true) allow(AssignmentQuestionnaire).to receive(:destroy).with(any_args).and_return(true) allow(User).to receive(:destroy).with(any_args).and_return(true) post :destroy expect(flash[:error]).to be_present end it 'check if successful destroy leads to redirect' do allow(User).to receive(:find).with(any_args).and_return(student1) allow(AssignmentParticipant).to receive(:delete).with(any_args).and_return(true) allow(TeamsUser).to receive(:delete).with(any_args).and_return(true) allow(AssignmentQuestionnaire).to receive(:destroy).with(any_args).and_return(true) allow(User).to receive(:destroy).with(any_args).and_return(true) post :destroy expect(response).to redirect_to :action=> :list end it 'check if error during destroy leads to redirect' do allow(User).to receive(:find).with(any_args).and_return(nil) allow(AssignmentParticipant).to receive(:delete).with(any_args).and_return(true) allow(TeamsUser).to receive(:delete).with(any_args).and_return(true) allow(AssignmentQuestionnaire).to receive(:destroy).with(any_args).and_return(true) allow(User).to receive(:destroy).with(any_args).and_return(true) post :destroy expect(response).to redirect_to :action=> :list end end
role
This method is private and returns the role of the user as per their role_id. In case role_id is nil a new role with name '(none)' is created and returned. We test if this function is successfully called through the edit action even when the user's role_id is nil, or, the template is successfully rendered for edit and no error is raised.
context '#edit' do it 'checks if role renders through edit' do new_student = User.new new_student = student1 new_student.role_id = nil allow(User).to receive(:find).with(any_args).and_return(new_student) get :edit expect(response).to render_template(:edit) end it 'checks if role fails through edit' do new_student = User.new new_student = student1 new_student.role_id = nil allow(User).to receive(:find).with(any_args).and_return(new_student) expect{controller.edit}.not_to raise_error end end
paginate_list
This method is also private and needs to be tested through list action. The parameter paginate_options which decides the execution of statement we need to test is predefined in the paginate_list, and thus the required statement cannot be accessed or tested. So, we tested the other statements in the method.
context '#list' do it 'checks that paginate_list does not fail with controller' do expect{controller.list}.not_to raise_error end it 'checks that paginate_list does not fail with post' do post :list expect(response.status).to eq(200) end end
Conclusion
The users_controller has been thoroughly tested in its present state. The auto_complete_result method in auto_complete_for_user_name method needs to be defined. Further, the paginate_list method needs to be refactored so that all statements in the method are accessible and may be tested.
Related Links
The main repository can be found here.
The forked git repository for this project can be found here.
A video of all tests running can be seen here.