Greetings: The bug that I most recently started working on was this one. The latest commits on the feature branch I have for this can be found on Salsa...
https://salsa.debian.org/deadjoe/adduser/-/commits/DBTS-559423-delgroup-status N.B. This branch will probably be deleted when it's merged, so the above link may break one day. This bug is my first attempt at writing a new functional test. Notice that the test CI step is failing during autopkgtest execution with the proper expected failure (at commit d4f37a2a). Next steps include adding a fix for this new failing test to close this bug and writing a second test while implementing the proper set-up/tear- down routines to be called in between. I have arrived at a stopping point, having just a few questions... 1. I have commented out the invocation of the legacy test suite in the test metadata file (debian/tests/control) to isolate my work. This is temporary to make sure nothing interferes with my new tests. Should we continue to support the legacy test suite? 2. In light of the above, does autopkgtest run the tests from each stanza in debian/tests/control in a new/clean testbed? How good is the isolation between stanzas at test runtime, and what can I rely on regarding this? I would like to make sure old-style tests don't interfere with my new-style tests. Hopefully, a method for adding new tests can be established that is immediately obvious to reviewers and contributors. That will clear a path for other contributors to add tests when they provide patches. That's all for now. Let me know what you think! -- Jason Franklin
