skopeo Cirrus: use test-integration target

Oct 17 '25 13:10 lsm5

For the record: this is motivated by the discussion in https://github.com/containers/skopeo/issues/2701 .

Note https://github.com/containers/skopeo/issues/2701#issuecomment-3415747878 : right now it seems that we have to choose between test coverage and meaningful containerization.

Also, from a quick comparison, this seems to add 2-3 minutes (~20 %) to test runtime. That’s looking at one test run (vs. https://github.com/containers/skopeo/commit/169091e282a312e7e854727c13cd248c198ad954 ), it might well be a random flake — but if not, it’s a bit annoying.

Oct 17 '25 14:10 mtrmac

RE: https://github.com/containers/skopeo/issues/2701#issuecomment-3415747878 , I've added another commit to run the container with cap_sys_admin and now it works. Thanks.

Given there are set of container tests that are skipped in -local, I prefer we test test-integration in CI until we get to rewriting tests without container. But, that still leaves us with the problem of setting up openshift / oc in the environment which I guess we don't wanna bother the end tester with.

So, unless the openshift setup is made convenient, maybe let's just continue with skopeo_cidev. I'll check if the openshift binary used in openshift tests can be replaced with the newer oc binary. It's inconvenient that they don't ship it via an rpm AFAIK.

If this is good enough for merge, we should only merge whenever https://github.com/containers/automation_images/pull/419 (or whatever PR that change is added to) is ready to merge.

Oct 25 '25 20:10 lsm5

LGTM but I only play a Cirrus knowledgable person on TV

Oct 28 '25 20:10 TomSweeneyRedHat

A friendly reminder that this PR had no activity for 30 days.

Nov 30 '25 00:11 github-actions[bot]

@lsm5?

Jan 15 '26 23:01 TomSweeneyRedHat