Integration testing - search for your knowledge, tips and links!

Question

Hey guys, I'm after some tips and pointers on integration testing for a web application. Our project has been working for several years, and it is quite complicated. We are pretty good at unit tests, but we lack a decent set of integration tests. We have no documented use cases or even a reasonable set of test cases outside of our unit tests. "Integration testing" today consists of the developer’s knowledge of the likely impact of changes and manual, ad-hoc testing of the application. This is really not ideal - now we want to design and automate a continuous set of tests so that we can conduct regression testing and increase confidence in the quality of the application.

Finally, we created a platform (based on Selenium) so that we can quickly authorize and automate test execution. The problem now: we do not have tests, the page is good and really blank. The system has about 30 classes that interact with each other and affect the user interface. For a new user, registering there 40 properties that can be set, each time influencing the experience. During the user's lifetime, they will generate even more states. Given so many variables and possible states, this is a frightening prospect to start with, and probably that's why it has been neglected so far.

The pain of not having a decent test suite is now becoming destructive. I take the time to fix this problem - I am after some practical advice on test development. How do you guys approach him? Do you have links that may be helpful? How can I stop my mind running away with an infinitely large number of states for user data? How can I clean up cases of edges that fail (and our use is like searching)?

Help!

Answer 1

If this is a huge number of combinations that keep you from trying to generate test files, you should clearly look at all-pair testing .

We used Microsoft's PICT as a tool to minimize the number of test boxes, while maintaining reasonable assurance that most cases were covered.

the argument for testing all pairs is this: the simplest errors in programs, usually one input parameter. Next The simplest category of errors consists of interactions dependent between pairs of parameters that can be caught checking all pairs. 1 Errors associated with the interaction between three or more parameters are gradually less common 2 , while at the same time, gradually more expensive search for exhaustive testing, which has as its limit exhaustive testing of all possible inputs.