Make Rawhide CI Green

[Mab879]

Description:

Use registry.fedoraproject.org/fedora:rawhide for rawhide test

Rationale:

It should be updated a bit more often DockerHub, hopefully.

[github-actions]

Start a new ephemeral environment with changes proposed in this pull request:

Fedora Environment

Oracle Linux 8 Environment

[github-actions]

🤖 A k8s content image for this PR is available at:
ghcr.io/complianceascode/k8scontent:12065
This image was built from commit: 15e7703

Click here to see how to deploy it

If you alread have Compliance Operator deployed:
utils/build_ds_container.py -i ghcr.io/complianceascode/k8scontent:12065

Otherwise deploy the content and operator together by checking out ComplianceAsCode/compliance-operator and:
CONTENT_IMAGE=ghcr.io/complianceascode/k8scontent:12065 make deploy-local

[jan-cerny]

Looking at the failing

    platform = ssg.build_yaml.Platform.from_text(
        "systemd and !yum and (ntp or chrony)", product_cpes)
    assert platform.get_remediation_conditional("bash") == "( rpm --quiet -q systemd && ( rpm --quiet -q chrony || rpm --quiet -q ntp ) && ! ( rpm --quiet -q yum ) )"

it's suspicious that the conditional is expected to have different ordering of operands than the expression

[Mab879]

@evgenyz Got any ideas on the CI failure that Jan mentioned above?

[jan-cerny]

I think the produced Bash and Ansible code should have the same order of operands as the platform expression. Therefore, the test should be reworked. Then, I assume it will start failing everywhere except Rawhide. It's unclear to me how to fix that. There probably is used a set or a dictionary somewhere where we should use a tuple or a list or other structure that preserves order of the elements. But I haven't found where is the place. I think it needs to be somewhere in the bundled "expressions" library.

[evgenyz]

I think that with introduction of new Python ssg/ext/boolean.py started to order elements of the expression differently. It might be not convenient, but technically these expressions are equal from boolean logic point of view.

We can either try and fix boolean.py or stop comparing expressions as strings.

[evgenyz]

I think the produced Bash and Ansible code should have the same order of operands as the platform expression.

Since we use normalization of boolean expressions we can't really depend on order (and even particular set of elements) of the initial expression.

[evgenyz]

or stop comparing expressions as strings.

There is an example of how one could test features of the expression rather than its string representation contents.

[codeclimate]

Code Climate has analyzed commit 15e7703 and detected 0 issues on this pull request.

The test coverage on the diff in this pull request is 100.0% (50% is the threshold).

This pull request will bring the total coverage in the repository to 59.4% (0.0% change).

View more on Code Climate.