Make sure building with relative paths work correctly.

[rhatdan]

Fixes: #12763

Signed-off-by: Daniel J Walsh [email protected]

[flouthoc]

nit: extra line

[flouthoc]

@Luap99 Some of the unrelated netavark tests are failing, could it be that we have regressed netavark upstream ?

[vrothberg]

@Luap99 Some of the unrelated netavark tests are failing, could it be that we have regressed netavark upstream ?

Are we merging netavark PRs without exercising them here?

[flouthoc]

@vrothberg Afaik I think yes but i am not sure if changes are immediately reflected on podman's CI. @cevich @Luap99 Could tell this better.

[vrothberg]

@vrothberg Afaik I think yes but i am not sure if changes are immediately reflected on podman's CI. @cevich @Luap99 Could tell this better.

If that's the case, Podman CI should probably pick specific commits from netavark (instead of branch which I presume is what we're doing). Just as we do with all other dependencies and building blocks.

[Luap99]

Yes we use the main branch currently, The plan is to switch to the packaged netavark once we have a release.
I do not understand why we get timeouts now, the only netavark PR that was merged is this one: https://github.com/containers/netavark/pull/218/files and I do not think this causes a hang.
We also made some aardvark changes but I do not see why this would cause a hang since aardvark runs in the background.

[Luap99]

My netavark PR did cause the issue, I accidentally swapped the logic there.
containers/netavark#220 will fix it, once merged you need to wait ~15min and then you can restart the tests. A force push is not required.

[jwhonce]

LGTM

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: flouthoc, rhatdan

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 [flouthoc,rhatdan]

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

[cevich]

Are we merging netavark PRs without exercising them here?

Afaik I think yes but i am not sure if changes are immediately reflected on podman's CI. @cevich @Luap99 Could tell this better.

FYI (since this is already merged), yes this is the case. It's likely a signal of a testing gap in netavark CI. We could (in theory) run some select podman e2e tests under netavark CI, but it's not a small amount of work to get that going.

[vrothberg]

FYI (since this is already merged), yes this is the case. It's likely a signal of a testing gap in netavark CI. We could (in theory) run some select podman e2e tests under netavark CI, but it's not a small amount of work to get that going.

We're doing something similar in Skopeo and c/image. c/image always uses a branch of Skopeo (because those are the actual tests) but Skopeo uses a commit of c/image to make vendoring explicit and from another repo breaking CI.

[cevich]

This is also possible to do. I believe all that's needed is to find the CI run for the specific commit you want, and simply update the download URLs in .cirrus.yml accordingly. Unf. I don't believe there's a simple/easy path to find the artifact URL by lookup on commit ID, but I could be wrong.

If someone wants to spend time on it, there is a path from commit ID -> artifact URL that goes through the Cirrus GraphQL API. So a small script could (in theory) do this...with the caveat that not all Commit IDs have associated CI runs, so the script would need to handle that appropriately.

root -> [searchBuilds]()(
    repositoryOwner: [String]!
    repositoryName: [String]!
    SHA: [String]
): [[Build]]