Currently we have the QA script (qa-test-tools) running the tests before and after the patches we are testing. Then we compare the results and raise errors in case of new introduced errors. It is causing several problems: 1. Bad detection of deleted/moved files 2. False positives and some errors we don't catch 3. It is slow to run, especially because of the git manipulation (switch of branches) Ideally we need the tests to pass on the whole codebase, so we just need to run the tests on the file modified by the patchset. This is an omnibus bug in order to track the different steps we need to make the tests pass on the whole codebase, before being able to adjust the QA script. See also https://gitlab.com/koha-community/qa-test-tools/-/issues/93