You’ve just pushed the final commit for this new feature to the source control system. Only a few more hours for the CI pipeline to run all the tests and the system is ready to deploy. After the build finished, you notice that a couple of tests have failed. That’s strange, they don’t seem to be related to this new feature. Let’s just start the CI build again. Urgh, that means more time waiting for it to finish. So you wait, and wait, … and then, after a long while, you get notified of a “green” build. Then you suddenly realise that you’ve forgot to add some UI tests for the new feature. D’oh, this is going to keep you busy for the rest of the day. It takes forever to write a new test like that.
These are a small selection of the joys software developer have with UI tests and other kinds of integrated end-to-end tests. Why do we keep doing this to ourselves? In this talk, we’re going to address the underlying issues caused by integrated end-to-end tests and discuss an alternative approach that results in faster feedback, a lower investment cost and above all, a feeling of trust and confidence in your developer tests.
I'm a professional software developer since Y2K. A blogger since Y2K+5. Author of the book "Writing Maintainable Unit Tests". Provider of training and coaching in XP practices. Curator of the Awesome Talks list. Thinking and learning about all kinds of technologies since forever.See all speakers ›