Session 14.06.2020

[hnformentin]

DSD2, group of 6 experienced on tests

The benefits:

• The advantage is evident for long term and large team projects, but difficult to teach in other contexts
• With TDD, the quality of the code is improved even before starting to code. The discussions to understand the problem while peer-programming demonstrated that even simple problems can deserve more attention before implementing the corresponding feature.

What make a code base testable?

• Unit tests to have quick feedback
• Modules that reduce dependence and side effects
• Avoid stating variables which requires a lot of mocking

How can developers turn the key of non-tested code base?

• Start testing from where the team can (unit or integration tests)
• Test properties (e.g. increasing sequence, larger or smaller than)
• Make an internal rule: any code review is initiated if test coverage is not good

Practical situations:

• What to do when data to be tested is too big? (one solution would be to test properties)

How to develop the skill and discipline to TDD?

• Internal rule to do it for three consecutive months

Advanced test techniques used:

• @pytest.marker.parametrize()
• pytest.param
• after the assertion, use ", error message" to specify what is wrong