aglinxinyuan opened a new issue, #7510:
URL: https://github.com/apache/texera/issues/7510
### Task Summary
`KubernetesClientSpec` covers the pure fabric8-to-map transforms
(`phasesByPodName`,
`metricsByPodName`) and the namespace-wide wrappers, and stops there. The
single-pod half of the
class is untested: `getPodByName`, `podExists`, `getPodLimits`, `createPod`,
`deletePod` and
`generatePodURI` — **42 of the file's 62 lines**.
None of it needs a cluster, and that is the point worth recording. The seam
is already built and
documented in the class's own scaladoc: `class KubernetesClient(client:
io.fabric8.kubernetes.client.KubernetesClient)`
takes the fabric8 client as a constructor parameter, with the production
singleton bound to a real
one via `object KubernetesClient extends KubernetesClient(new
KubernetesClientBuilder().build())`.
The existing spec's `stubbedClient` fixture already proves the Mockito
approach; it only needs
extending to the rest of the fluent chain — `withName(...).get()` for the
lookups and
`resource(pod).inNamespace(...).create()` for creation.
Two traps:
- `RETURNS_DEEP_STUBS` cannot be used — fabric8's fluent API returns type
variables, so each step
has to be mocked explicitly. The existing fixture already says this.
- An `ArgumentCaptor` cannot be read from inside the `thenAnswer` that stubs
`create()`; it is only
populated by `verify(...).capture()` afterwards. Stub `create()` to return
`null` and inspect the
captured pod after the call.
### Task Type
- [ ] Refactor / Cleanup
- [ ] DevOps / Deployment / CI
- [x] Testing / QA
- [ ] Documentation
- [ ] Performance
- [ ] Other
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
