Fix KeyError and exit stack leak on empty servers in ClientSessionGroup (#3384)

[xinghewumeng]

Description

Fixes #3384.

Connecting a ClientSessionGroup to a server that exposes no tools, resources, or prompts previously caused two distinct issues due to del self._session_exit_stacks[session] in _aggregate_components:

1. When connecting via connect_with_session(), the caller manages the session lifecycle, so session is not present in _session_exit_stacks. The unconditional deletion raised a raw KeyError.
2. When connecting via connect_to_server(), the deletion dropped the tracking entry without closing the underlying session_stack. The session was still registered in self._sessions, but calling disconnect_from_server(session) subsequently failed to close the session's exit stack because session in self._session_exit_stacks was False.

Changes

• Remove the premature deletion of _session_exit_stacks[session] in _aggregate_components.
• Allow servers that expose no components to be registered and managed cleanly in ClientSessionGroup.
• Add unit tests verifying both connect_with_session and connect_to_server lifecycles on empty servers.

Testing

• Added regression tests in tests/client/test_session_group.py:
  • test_client_session_group_connect_empty_server_with_session
  • test_client_session_group_connect_empty_server_via_connect_to_server
• Passed all tests: uv run pytest tests/client/test_session_group.py
• Passed linters: uv run ruff check and uv run ruff format --check

[github-actions]

This PR has been closed automatically. This repo only keeps pull requests open when they come from a maintainer, or from a contributor a maintainer has assigned to the linked issue, and you aren't currently assigned to #3384.

If a maintainer assigns you to #3384, this PR reopens on its own and there's nothing more you need to do here. Assignment is a maintainer call based on capacity; comments that only ask to be assigned don't factor in. What does help is engaging on the issue itself by confirming the repro, explaining why it matters for your use case, or describing the approach you'd take.

You're welcome to keep pushing commits here (just avoid force-pushing, since GitHub can't reopen a rewritten branch), but that on its own won't get the PR reviewed or the issue assigned, and realistically most auto-closed PRs stay closed. There's no need to open a new PR either way.

CONTRIBUTING.md has the full reasoning, but in short:

• We're a small team with very little capacity to review community PRs right now.
• Many recent PRs are AI-generated with little human review, and reviewing one carefully still costs a maintainer as much time as it ever did. A well-described issue is usually more useful to us than the code.

Maintainers: reopen, remove missing-issue-link, or add bypass-issue-check to override.