ClusterMetadata, update_metadata(...) always fails when following request_update() call

[jackgene]

Describe the bug
On ClusterMetadata, an update_metadata(...) call following a request_update() call always fails.

This is because update_metadata(...) is calling:

• concurrent.future.Future#success(...)
• concurrent.future.Future#failure(...)

Both of which do not exist.

Expected behaviour
The correct concurrent.futures.Future API is used, and no exception is thrown.

Environment (please complete the following information):

• aiokafka version: 0.11.0
• Kafka Broker version: N/A (not a Kafka issue)
• Other information: N/A

Reproducible example

    cluster = ClusterMetadata()
    updated_cluster = cluster.request_update()
    # This always fails:
    cluster.update_metadata(
        MetadataResponse[0]([(0, "foo", 12), (1, "bar", 34)], []),
    )

[ods]

Hi @jackgene, thank you for the contribution! You're right, something weird is here, but it doesn't look as simple as in your PR. How did you get this error? I don't see any place where request_update() is called, so self._future should be always None.

[jackgene]

Hi @jackgene, thank you for the contribution! You're right, something weird is here, but it doesn't look as simple as in your PR. How did you get this error? I don't see any place where request_update() is called, so self._future should be always None.

Yeah, I don't think it's called within the library. I think it's just something that could potentially happen if a client calls request_update(). I discovered these two weird future calls as they were flagged by Pyright during type checking.