ci: run unit tests in CI

[mykaul]

Problem

Unit tests (tests/unit/) are never executed in any CI workflow. The integration-tests.yml only runs tests/integration/standard/ and tests/integration/cqlengine/ — which require a running Scylla cluster via ccm.

This means bugs in unit-testable code can go undetected for weeks. Example: Session.wait_for_schema_agreement was added in commit 0d215f45b with a docstring and test promising ValueError for invalid scope, but the validation was never implemented. It was only caught now because a human looked at the test output.

Solution

Add a uv run pytest tests/unit/test_*.py -x step after the build and before the Scylla download/start. Unit tests:

• Have zero external dependencies (no Scylla, no ccm, no JDK)
• Run in ~30 seconds (vs ~15 minutes for the Scylla download alone)
• Would have caught both pre-existing bugs immediately

Pre-existing test failures

Two tests fail on master due to bugs introduced before CI was running them:

Test	Bug	Fix PR
test_wait_for_schema_agreement_rejects_unknown_scope	Scope validation never implemented	#917
test_set_keyspace_for_all_pools_reports_all_errors	callback(host_errors) should pass accumulated errors	#915

Both are marked @pytest.mark.xfail(strict=False) referencing their fix PRs. Once those PRs merge, these xfail markers should be removed.

[coderabbitai]

Warning

Review limit reached

@mykaul, you've reached your PR review limit, so we couldn't start this review.

Next review available in: 34 minutes

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 838e9184-f1db-4a71-967f-840381d12d3b

📥 Commits

Reviewing files that changed from the base of the PR and between c1bfd54 and 701a6e8.

📒 Files selected for processing (2)

• .github/workflows/integration-tests.yml
• tests/unit/test_cluster.py

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[Lorak-mmk]

Unit tests (tests/unit/) are never executed in any CI workflow.

Not true. We are running unit tests. More than once actually, because we run them for each target that we build wheels for. See for example test output from your other PR: https://github.com/scylladb/python-driver/actions/runs/28400830092/job/84151324434?pr=917#step:12:1693

The issue is not that we are not running unit tests, but that we are silently skipping some tests, as I described in #904

The solution is to disallow skipping in CI, and instead explicitly list skipped tests, which Sylwia is working on in #911

[mykaul]

Unit tests (tests/unit/) are never executed in any CI workflow.

Not true. We are running unit tests. More than once actually, because we run them for each target that we build wheels for. See for example test output from your other PR: https://github.com/scylladb/python-driver/actions/runs/28400830092/job/84151324434?pr=917#step:12:1693

The issue is not that we are not running unit tests, but that we are silently skipping some tests, as I described in #904

The solution is to disallow skipping in CI, and instead explicitly list skipped tests, which Sylwia is working on in #911

But I did not fail on libev specific tests... ?

[Lorak-mmk]

But I did not fail on libev specific tests... ?

What didn't?

As an example, test_wait_for_schema_agreement_rejects_unknown_scope which should fail because Scope was not validated ( #917 ) was silently skipped, alongside others SessionTest tests precisely because lack of libev.

setup of those tests has this:

    def setUp(self):
        if connection_class is None:
            raise unittest.SkipTest('libev does not appear to be installed correctly')
        connection_class.initialize_reactor()

connection_class, unless specified with env, is this:

    log.debug("Using default event loop (libev)")
    try:
        from cassandra.io.libevreactor import LibevConnection
        connection_class = LibevConnection
    except DependencyException as e:
        log.debug('Could not import LibevConnection, '
                  'using connection_class=None; '
                  'failed with error:\n {}'.format(
                      repr(e)
                  ))
        log.debug("Will attempt to set connection class at cluster initialization")
        connection_class = None

So if libev is not installed, it will be None and all SessionTest tests skipped.