Doorgaan naar hoofdcontent

Write good unit tests


It is not enough for today’s software developers to know their programming language well. There are further skills, that more and more companies are expecting from there employees. One of the most important is Test Driven Development (TDD). This is not an introduction to TDD. If you want to learn it, I recommend Uncle Bob’s awesome Clean Code Videos (Episode 6 - TDD) or simply ask Google for it. But many developers are writing bad code and applying TDD does not make them writing good code. Instead it makes them also writing bad tests. So this is about writing better test code.

Applying simple rules

To understand why test code can be bad, you should understand, what it should do. It should work as the parachute, that keeps you alive, when refactoring your code. Tests may help you to be sure nothing breaks, when adding new features to your code. But tests may also work as sample code, that documents your APIs better than any other documentaion except the code itself.
But how can you make your tests better? It might help to follow some simple rules, that could be easily applied to every language or test style like BDD or Junit-style.

Tests should be a state machine

Many people do not like BDD at all, but there is a pretty nice idea in it - the given-when-then style some frameworks promote. This style forces you into a way of thinking about tests, that you should adapt. Even if you do not use a BDD framework.
Writing a test this way means there is a start state, something happens and than an end state is reached. If your test is broken, the state machine in it is broken. In BDD frameworks the first part of your test is the given block, where all the setup stuff is done. The second part is the when block, where an action is applied on the test object, created in the given block. At least you have a thenblock, where you assert, that the correct end state is reached.
It is very helpful to have this in mind while writing a new test. Keep these three parts seperated and do not mix them in some way. Do not write code in your test, where an if appears, or even more complex logic. In a test you should only do the above three steps. Do some simple setup, call a method on your test object or invoke the test function and assert the result is correct.
This might also make your code better. If you write messy code, tests written this way are harder to create and maintain. If you have to much of inheritance, dependencies on other objects or resources like IO, you will have to set it up in every test you write and that is no fun. But you should write your tests first and hopefully it makes writing messy code harder, if you have written a well structured tests first.
More on this great article can be read here.

Reacties

Populaire posts van deze blog

OSB 10gR3 and SWA and MTOM

This blog is about using soap with attachments and the use of MTOM within the OSB (10gR3). A service is created that accepts a soap with attachment (DocumentService) and translates it to a service that accepts a binary element. MTOM is used for performance reasons for the second. Some notes: * For the use of attachments you need RPC-style document instead of the usual document-style. This due to the fact that the document-style limits a message to a single . * A service can not have both SWA and MTOM within OSB. First a WSDL is setup for the DocumentService: The $attachments variable holds the attachments and the body holds the attachment data. Also other data is stored within the attachment element (see h

Microservices mindmap

"The tree" - See also   my photo page When you are fairly new within the Microservices land, there are a lot of terms fired at you. So also for my own understanding i have made a mindmap. I think it has a good status now, so that i can share it with you. As always feedback is very welcome ! You can download the mindmap here .

Book review: Data Management at Scale (Piethein Strengholt)

 This blog is a review of the book "Data Management at Scale (See also at bol.com ) Data Management is a hot topic nowadays and this book does a fantastic job at adding value to this topic. It is a must read and one of the few technical books I finished reading in a weekend. The book gives a fantastic overview on how to implement a Data Mesh data architecture. The Data Mesh concept is explained by Martin Fowler here . The book is a good mix between conceptual and implementation architecture level. It gives a lot of examples of how this architecture at scale can work, for both small and big companies. It is practical and I used it to implement it at one of my customers. The book describes an architecture in which the focus is on the DIAL (Data- and Integration Access Layer).  On a high level the book covers the following topics: The key principles for data management at scale - Domain-Driven Design  - Domain Data Stores - Meta data management Ready Data Store The concept of servin