Closed olifre closed 1 year ago
This is a good case for an acceptance test as well I'd say.
Indeed it is. I've addressed all comments apart from the acceptance test, will look into this later this week.
I've now fixed the Rubocop offenses in the existing test, and added an acceptance test. I don't have a box at hand (yet) to run the acceptance test, but I think it should work — CI will show ;-).
Nice, the acceptance test found a real bug: disable
ing a COPR repository without adding it first will fail, and not be idempotent on dnf
-based systems (which leave the repository file in place, but switch it to disabled
in this case). yum
-based systems effectively don't differentiate between disabled
or removed
.
I have fixed that in the last force-push by ensuring a COPR repository is added before disabling it, if disabled
is requested and it is not disabled
yet on dnf
-based systems.
Another issue was that the removed
acceptance test did not add a repository first, so it was not a real test whether the removing worked. This is now also fixed.
So I hope this might be the last iteration :-).
One of the problems with this if you are running with
resources{'yumrepo':
purge => true,
}
The repo file will be deleted won't it?
I don't see a good way around that.
The repo file will be deleted won't it?
Indeed...
I don't see a good way around that.
Same here, at least not as long as we want to use the official way (via dnf
/ yum
) instead of managing the repo "manually", but then all the magic hidden by the underlying package manager plugins is lost and we'd need to follow all updates on that end.
Good ideas are very welcome.
Sorry for the multiple force-pushes, I found a small logic issue in the last commit and fixed it — if there is any good proposal on how to upstream this, fixing / documenting the issue @traylenator pointed out, please let me know.
@olifre force pushes are fine. Can you take a look at the failing CI jobs?
@olifre force pushes are fine. Can you take a look at the failing CI jobs?
Thanls, I think I fixed all CI issues, the remaining issue seems to be a fluke of the GitHub actions runner (read timeout reached
). I will trigger CI once more with an empty force-push.
@bastelfreak That did the trick, now all is green :smile: .
@bastelfreak It seems CI went red again after you re-ran it yesterday, but checking the logs, that seems like an issue with the runners, correct?
One of the problems with this if you are running with
resources{'yumrepo': purge => true, }
The repo file will be deleted won't it?
I don't see a good way around that.
Given there's no sensible way around this and it only affects particular configurations I would just document it.
If the resources declaration can be adjusted to ignore the copr
repos based on name or something then makes sense to give that example.
https://forge.puppet.com/modules/crayfishx/purge/readme gives an example in fact.
Given there's no sensible way around this and it only affects particular configurations I would just document it. If the resources declaration can be adjusted to ignore the
copr
repos based on name or something then makes sense to give that example.
Thanks for confirming!
The names of yumrepo
resources resulting from COPR commands end up being formed such as:
copr:copr.fedorainfracloud.org:jdoss:wireguard
copr:copr.fedorainfracloud.org:copart:restic
(tested with RHEL 7 and 8). Since resources
itself does not seem to allow limiting what is purged, the only way out seems to be to use something like the purge
module you linked.
What do you think about the documentation addendum I added in the preceding commit?
I have finally come around to also regenerate REFERENCE.md
and rebase onto current master
, so things should now be green and mergable.
COPR (Cool Other Package Repo) is a Fedora project for third-party package repositories, see: https://copr.fedorainfracloud.org/
Pull Request (PR) description
This PR adds support to manage COPR repositories via a new
yum::copr
resource. Documentation and test are included.This Pull Request (PR) fixes the following issues
Fixes #149