Effective Unit & Integration Testing - Pitfalls and How to Avoid Them

A presentation at Qinshift Tech Talk by Petyo Dimitrov

About us Atanas Gardev Software Architect 20 years’ experience in IT, Java Backend focus 3 kids :) Petyo Dimitrov Senior Software Architect 17 years’ experience in IT no kids (yet)

Agenda 01 What is a good test? 02 Anti-patterns 03 How to make a better test? 04 AI benefits in testing 05 Q&A 3

Spock structure 12 def “test method name”() { given: “prepare” … when: “stimulus” … then: “assertions” … }

Spock syntax def user = Mock(User) user.name >> “Johnny Bravo” user.isDeleted == false 1 * converter.convert(user) 0 * converter._ 13

Anti-pattern #1 def user = Mock(User) user.name >> “Johnny Bravo” user.address >> “Sofia” user.type >> UserType.Corporate user.status >> UserStatus.Active user.isDeleted >> false user.lastLogin >> new Date() 14

Mocking overuse – solution (1) def johnny = new User(“Johnny Bravo”, “Sofia”, UserType.Active) johnny.isDeleted = false johnny.lastLogin = new Date() (2) def johnny = UserMother.johnny() 15

Object Mother pattern UserMother.johnny() UserMother.littleSuzy() AddressMother.sofia() UserMother.johnny() .withAddress( AddressMother.sofia().build()) .build() 16

Anti-pattern #2 def “user conversion works”() { given: def input = new User(“Johnny Bravo”, “Sofia”, Active) when: def result = underTest.convert(input) then: result.name == “Johnny Bravo” result.address == “Sofia” result.type == Active } 17

Repeating yourself – solution def “user conversion works”() { given: def input = UserMother.johnny() when: def result = underTest.convert(input) then: result.name == input.name result.address == input.address result.type == input.type } 18

Utility classes def “user conversion works”() { given: def input = UserMother.johnny() when: def result = underTest.convert(input) then: Utils.matchProperties(result, input, “name”, “address”, “type”, …) } 19

Anti-pattern #3 def “user service can create user”() { given: def user = UserMother.johnny() def userForm = new UserForm(user.name, user.address, user.type) when: def result = underTest.createUser(userForm) then: 1 * userRepository.persist(_ as User) >> user 0 * userRepository._ result == user 20 }

Testing implementation vs contract – solution def “user service can create user”() { given: def user = UserMother.johnny() def userForm = new UserForm(user.name, user.address, user.type) userRepository.persist(_ as User) >> user when: def result = underTest.createUser(userForm) then: 0 * userRepository._ result == user 21 }

Anti-pattern #4 – setup @PostMapping @ResponseBody public UserDto create(@RequestBody @Valid UserForm userForm) { User user = userService.create(userForm); return userConverter.convert(user); } 22

Anti-pattern #4 def “controller calls the service and converter”() { given: def user = UserMother.johnny() def userForm = new UserForm(user.name, user.address, user.type) when: def result = underTest.create(form) then: 1 * userService.create(form) >> user 1 * userConverter.convert(user) >> … result.name == user.name result.address == user.address } 23

Don’t write unit tests, write integration tests! 24

Why? • A few Integration/API Tests replace numerous Unit Tests • Easy to develop (e.g. in Spring - @DataJpaTest, @MockBean, MockMvc) • Work with real REST requests / async events • Allow access to the database (where test data may be set up in advance) • 3rd party system calls are handled by mocking the API rather than our client code 25

Integration test example (1) def “user is created, updated and deactivated after some time”() { given: def user = UserMother.johnny() when: def createdUserDto = /** trigger create REST API */ then: createdUserDto.name == user.name userRepository.findByName(user.name) != null 27

Integration test example (2) when: user.lastLogin = now().minusDays(30) def updatedUserDto = /** trigger update REST API */ then: updatedUserDto.name == user.name updatedUserDto.lastLogin == user.lastLogin when: jobsUtil.triggerUserPeriodicJob() def userDto = /** trigger read REST API */ then: userDto.name == user.name userDto.status == UserStatus.Deactivated 28

Don’t write integration tests, write unit tests! 29

Source code example def updateState(newState: UserState) { var validStates = [] when (this.state) { Inactive -> validStates = [Active, Deleted] Active -> validStates = [Inactive] } if (validStates.contains(newState)) { state = newState } else { throw IllegalArgumentException() } } 30

Unit test example @Unroll def “updateState should change state from #currentState to #newState”() { given: def user = new User(state: currentState) when: user.updateState(newState) then: user.state == newState where: 31 } currentState | newState UserState.Inactive | UserState.Active UserState.Inactive | UserState.Deleted UserState.Active | UserState.Inactive

AI considerations • • • • • • 39 Syntax & logical errors Limited edge case coverage Not as creative* Difficulty with complex tests Data Privacy Intellectual Property

Summary • Don’t Repeat Yourself • Avoid mock overuse • Test contracts, not implementation internals • Don’t test for the sake of testing • Distribute your tests across the test pyramid • Let AI help you 42

To me, legacy code is simply code… without tests Michael Feathers Working Effectively With Legacy Code

Thanks! Atanas Gardev Software Architect atanas.gardev@musala.com Petyo Dimitrov Senior Software Architect petyo.dimitrov@musala.com

Petyo Dimitrov
@petyodimitrov

1 / 45

In today’s agile software development processes, effective testing is crucial. It must be an integral part of writing code to ensure the resulting product functions as specified. In this session, we will define what constitutes good unit and integration tests and share common bad practices in unit testing that we have encountered over time.

We’ll explore real-world examples of poor testing patterns, their negative impacts, and how to rework them into more effective practices. Additionally, we will discuss how AI can be used to improve unit testing by assisting in writing effective test cases and enhancing test maintenance.

By the end, you’ll have a structured approach to writing better tests!

Key Takeaways:

Characteristics of Effective Tests
Common Pitfalls in Testing
Efficient Test Creation
Integrating AI in Testing

Effective Unit & Integration Testing - Pitfalls and How to Avoid Them

Link for this presentation:

HTML code for embedding:

Share on social media: