I have created several universal procedures after the training that cover elementary processes in our software and you then can refer to them in other test cases. Also the webinar about Dependencies helped me to understand this topic and optimize a test.
At the beginning we focused on the accounting field of our software. We are now testing different processes e.g. interfaces, document creation and accounting transactions with different master data or creating invoices. We then also check the created PDF files. That works super!
We still face challenges concerning the independency of the tests. Our tests star with an empty data base, the data is created by the tests during the test run and then imported. But, when there is an error in test #5, tests 1 to 4 have to run through first that we can find the error and debug. At this point this is working, but with an increasing number of tests this may become difficult due to lack of time. We should maybe save the data base after every test case, but this would have a deep impact in the performance.
We automated the tests via our Jenkins with pipelines (not the QF-Test plugin), since we have several setups and cleanups of our tests, this makes the tests better controllable. After the creation of a new branch of our software we just have to change the path to the new program version – We haven’t had other changings or bugs on QF-Test’s side when checking a new version! In the near future we are going to update eGECKO on Java 11, but I am optimistic that we won’t face any problems there.
eGECKO is a modular and very comfortable business software for very different economic processes.