Week 6 - Testing and Build Management

Question 1

What is a unit?

Answer 1

Typically a single function or method

Question 2

What is a good unit testing plan?

Answer 2

To have one test module/class per production module/class, which tests the public interface of the production class

Question 3

What does the Arrange-Act-Assert Pattern divide tests into?

Answer 3

The arrange section

The act section

The assert section

Question 4

What happens in the arrange section of the Arrange-Act-Assert Pattern?

Answer 4

You bring the system under test and its dependencies to a desired state

Question 5

What happens in the act section of the Arrange-Act-Assert Pattern?

Answer 5

You can call functions/methods on the system under test, passing the prepared dependencies, and capturing the output value (if any)

Question 6

What happens in the assert section of the Arrange-Act-Assert Pattern?

Answer 6

You verify the outcome. The outcome may be represented by the return value, the final state of the SUT and its collaborators, or the methods the SUT called on those collaborators

Question 7

What does SUT stand for?

Answer 7

System Under Test

Question 8

What are the phases of Test-Driven Development (TDD)?

Answer 8

Red, Green and Refactor

Question 9

What does TDD stand for?

Answer 9

Test-Driven Development

Question 10

Describe the red phase of TDD.

Answer 10

Write an automated test for some behaviour

Question 11

Describe the green phase of TDD.

Answer 11

Write just enough code to pass the test

Question 12

Describe the refactor phase of TDD.

Answer 12

Change the code so it becomes better

Question 13

What is build management driven by?

Answer 13

Version control

Question 14

What are the three benefits version control provides?

Answer 14

Change histories

Branching and merging

Traceability

Question 15

What do change histories do?

Answer 15

Change histories make it possible to quickly and reliably get any version of a file

Question 16

What does branching and merging do?

Answer 16

Makes it possible for developers to work independently, and to combine their work together, warning of any conflicts

Question 17

How does version control make traceability possible?

Answer 17

By annotating each change with a short message describing the purpose of the change

Question 18

What are the seven essential practices of continuous integration?

Answer 18

Do not check in on a broken build

Always run commit tests locally

Wait for commit tests to pass

Never go home on a broken build

Always be prepared to revert

Do not comment out failing tests

Take responsibility for breakages

Question 19

Why should you not check in on a broken build?

Answer 19

There is no point checking in further changes, triggering new builds, and compounding the failure

Question 20

Why should you always run commit tests locally and wait for them to pass?

Answer 20

So failures can be fixed right away before causing a broken build - only commit once a local build is successful

Question 21

Why should you never go home on a broken build?

Answer 21

Because it means nobody else can work so you should check in at least an hour before the end of the working day, or first thing next morning

Question 22

What should you do if a problem with a commit cannot be fixed quickly?

Answer 22

Revert it

Question 23

What are the two common branching models?

Answer 23

Feature-based development and trunk-based development

Question 24

Describe feature-based development.

Answer 24

A branch is created for each feature being developed. Merging branches is delayed until the feature is finished, possibly weeks later

Question 25

Describe trunk-based development

Answer 25

A branch is created for each feature being developed. Merging branches is done daily, with feature flags used to hide incomplete work

Question 26

What are the two common repo organisations?

Answer 26

Monorepo and multirepo

Question 27

What is a monorepo?

Answer 27

One giant repo for all microservices - any check-in triggers the production of multiple microservices. Should be avoided

Question 28

What is a multirepo?

Answer 28

One repo per microservice - any check-in triggers the production of a single microservice