CI: On vX.Y-rhel branches, ensure that some downstream Jira issue is linked

[alexlarsson]

In the RHEL specific branches we want to ensure that all MRs link to at least one downstream Jira ticket. To do this we add a new test in validate-source similar to the existing pr-should-include-tests. This test only runs on actual pull requests.

The syntax for linking to a Jira is "Fixes " followed by one jira links, like so:

Fixes https://issues.redhat.com/browse/RHEL-50506

Note: This is the same syntax as for a regular github issue reference.

[alexlarsson]

@baude @mheon @TomSweeneyRedHat @edsantiago follow up from the email discussion.

The exact syntax, as well as the label name for the override is just an initial proposal for me, and is up for discussion.

[alexlarsson]

Btw, the DEST_BRANCH in .cirrus.yml on the "v4.9-rhel" branch is wrong (it is "v4.9"), and we have to update it if we want this to work there.

[mheon]

I'm thinking we're going to mandate all PRs on upstream that fix an issue reference that issue as "Fixes " where issue could be a Github issue ID, a URL, or a Jira reference. So I think matching "Fixes Jira -\n" would be preferable. But the overall approach LGTM.
This will need to be backported to each -rhel branch, we take CI from per-branch config not main.

[mheon]

@cevich Is there any way we can enforce Cirrus DEST_BRANCH being set correctly, because I think we're especially bad about remembering to do this for -rhel branches?

[alexlarsson]

Maybe we can have a github action (not in cirrus) that validates the cirrus DEST_BRANCH value? I think you can get that from $GITHUB_BASE_REF

[cevich]

Maybe we can have a github action (not in cirrus) that validates the cirrus DEST_BRANCH value? I think you can get that from $GITHUB_BASE_REF

That's not a bad suggestion actually since github actions on this repo can send e-mail to our monitoring list.

OTOH, the base-branch name is exposed in a Cirrus-CI env. var. (can't recall the exact name) to PR-level tasks but nowhere else (i.e. not in tag or branch-level jobs). So maybe a simple check could be done if $CIRRUS_BASE_BRANCH != main && $CIRRUS_BASE_BRANCH != $DEST_BRANCH or something to that effect.

[Luap99]

Not against this per se, just marking that I don't think we should merge this until we have a general accept format defined and documented before we "enforce" anything

[mheon]

@Luap99 I'm going to proposal the generic Fixes [issue] formatting on Monday at standup.

[TomSweeneyRedHat]

This needs to be documented somwehere. Perhaps https://github.com/containers/podman/blob/main/CONTRIBUTING.md

[alexlarsson]

@Luap99 I'm going to proposal the generic Fixes [issue] formatting on Monday at standup.

Any chance I can join that?

[mheon]

@alexlarsson Sure, I can invite you.

[alexlarsson]

Hmm, github actually also allows a colon, like "Fixes: ", and that is what CONTRIBUTING.md says. I'll make the PR handle that as well.

[alexlarsson]

@mheon Updated the PR

[mheon]

Someone who is better at regexes than I should review, but LGTM

[edsantiago]

Good start, but a few suggestions inline

[edsantiago]

LGTM. One gripe, not worth a repush. Thank you!

[edsantiago]

This doesn't belong here (it should be on line 11) but it's not worth yet another repush.

[alexlarsson]

I added it there so i could test it with:

CIRRUS_REPO_CLONE_TOKEN=... ./pr-should-link-jira.t

Otherwise it would not get propagated to the test.
That said, one might very well not want to support that...

[cevich]

If you do want to keep this here, I would be helpful to add a comment suggesting it's needed for pr-should-link-jira.t to function.

[alexlarsson]

I moved it.

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: alexlarsson, edsantiago

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [edsantiago]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[alexlarsson]

The mac build always seems to time out on the linter. I assume that is not my fault?

[cevich]

The mac build always seems to time out on the linter. I assume that is not my fault?

I think there was a fix for that recently and rebasing on main should fix it. I might be wrong though.

[alexlarsson]

The mac build always seems to time out on the linter. I assume that is not my fault?

I think there was a fix for that recently and rebasing on main should fix it. I might be wrong though.

I tried rebasing, lets see.

[packit-as-a-service]

Cockpit tests failed for commit ba46e5651476150655ba3f250b230110a5542f6b. @martinpitt, @jelly, @mvollmer please check.

[edsantiago]

Still LGTM

[alexlarsson]

Ok, all requested changes are done, and now we passed CI. Can I get this merged?

[alexlarsson]

(Note: Once this is merged we need to backport to active rhel branches.

[mheon]

/lgtm from me

[mheon]

That might have confused the bot
/lgtm

[cevich]

Thanks for taking this on and following it through @alexlarsson