response-future: release request ids after send failure

[dkropachev]

Fixes #871. Reclaims the request id when the send path fails after registration, and keeps the cleanup covered by unit tests.

[Copilot]

Pull request overview

This PR fixes a request/stream ID leak in ResponseFuture._query() when the send path fails after a request ID has already been allocated, ensuring the ID is reclaimed and the connection state is cleaned up. This aligns with issue #871 by making request-id cleanup explicit for the “allocated-but-not-successfully-sent” failure path and keeping it covered by a unit test.

Changes:

• Track request_id in _query() and ensure it is released back to connection.request_ids on send/encode failures after allocation.
• Remove any partially-registered request callback state (connection._requests) associated with the failed send before returning the connection to the pool.
• Add a unit test to assert request-id reclamation and request map cleanup when a send fails after registration.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
cassandra/cluster.py	Adds request-id aware cleanup on _query() exception paths to prevent stream/request ID leaks.
tests/unit/test_response_future.py	Adds regression coverage ensuring request IDs and _requests entries are cleaned up after post-allocation send failures.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.