At least when doing release tests but probably also in other contexts there is a need to run f-r on a large set of packages. Doing this in the normal way would effectively be a mass rebuild and requires to much resources. So, f-r needs to run the tests available without using mock at all. While this limits the scope, it's still useful.

Although there are ways to do this using the -x option this is fragile since the tests needed by all other tests are disabled. To get a more consistent use case for this we need a defined way to run these kind of tests, be it a flag or an option.