As software systems grow, so does the number of their tests. As a result, we increasingly often encounter test suites that run hours or days, which is painfully slow. To get fast feedback about new bugs despite slow tests, we can run a subset of all tests more frequently. If this subset is chosen well, these tests identify most bugs very quickly.
This approach heavily depends on the selection of frequently-executed tests. We developed two different approaches for selection:
In the workshop, we first present the key ideas and research studies for both, and then demo them live in our analysis tool Teamscale.
Upcoming live workshop: