Dynamic Test Scope Selection in Continuous Integration Loop

Abstract: Continuous integration (CI) is a critical step in big software projects, and the telecommunications industry relies on it in 5G network deployment. The testing infrastructure during the implementation is significantly challenging, and the developers need to execute fewer tests to get faster feedback about their changes in the code. There are many different test case selection methods where techniques based on test history are commonly used in test prioritization or decrease the test scope. This thesis is about researching and implementing two dynamic models in Ericsson's environment to minimize integration test scope after code changes. One is based on dependencies between classes, and the second is based on the historical record of test executions. The TCS (Test Case Selector) is evaluated concerning time and fault detection effectiveness compared with the currently implemented technique. The results show that the model based on dependencies did not improve, while the model based on historical test executions shows impressive improvements in increasing the fault detection rate. That increases testing efficiency with less or the same time as the current technique, where only the most effective and essential test suites are being executed.