Fedora CI SIG - 10-Feb-2021
by Jim Bair
Attendees: jimbair, msrb, tflink, bgoncalv
Meeting log:
https://meetbot.fedoraproject.org/fedora-ci/2021-02-10/fedora_ci_sig.2021...
Items discussed:
1. Branching occurred, so rawhide is now f35
- Miro uploaded an f34 image to testing farm, so f34 should work now as
well
- Worth noting but not impacting, but as announced last week, master
branches are now main branches
Pretty light meeting today; reminder that we have it at the same time every
two weeks, so if you have a question or problem with fedora CI, come ping
us in freenode in #fedora-ci or drop into our next meeting.
Thanks!
-Jim
2 years, 9 months
zuul rpm-install-test false positives
by Zbigniew Jędrzejewski-Szmek
[resend, since the first one seems to have gotten stuck permanently in the moderation queue]
Hi,
I would like to make the CI pass for systemd builds. I added an rpminspect.yaml file [1]
to disable a bunch rpminspect checks, and rpminspect passes now. But I'm stuck on
rpm-install-test, which dutifully informs me that
> package systemd-247.2-3.fc34.x86_64 conflicts with
> systemd-standalone-tmpfiles < 247.2-3.fc34^ provided by systemd-standalone-tmpfiles-247.2-3.fc34.x86_64
Duh, no surprise, I added the Conflicts myself ;)
I do NOT want to disable the test completely. I am very much interested in CI catching the
case where packages cannot be installed or upgraded. I'd like just to either tell it that
it should ignore systemd-standalone-tmpfiles and systemd-standalone-sysusers, or ideally,
it would be able to guess on its own that if there are Conflicts between subpackages from
the same built, that is intentional and should be ignored.
I looked at the docs [2, 3], but I don't see anything relevant.
[1] https://src.fedoraproject.org/rpms/systemd/c/7e740c5252c3291d67def67d33e5...
[2] https://docs.fedoraproject.org/en-US/ci/generic_tests/#_installability
[3] https://lists.fedoraproject.org/archives/list/ci@lists.fedoraproject.org/...
Zbyszek
2 years, 9 months