Avoid Python test server port-allocation races

[kpumuk]

RunClientServer.py selected an unused port by binding a temporary socket, closing it, and then launching TestServer.py on the same port. Another process could claim the port during that gap, causing intermittent macOS failures with OSError: [Errno 48] Address already in use.

Let each test server bind port 0 so the kernel allocates and owns the port, then report the bound port to the parent before the client starts. The same reporting path covers regular, SSL, and HTTP test servers; ephemeral HTTP tests bind to loopback and avoid hostname lookup during startup.

Representative failures include macOS 26 with Python 3.11, macOS 14 with Python 3.13, Intel macOS 15 with Python 3.10, macOS 26 with Python 3.10, and macOS 15 with Python 3.13. Each failed in the Python check while binding a test-server socket with OSError: [Errno 48] Address already in use.

• Did you create an Apache Jira ticket? (Request account here, not required for trivial changes) — Skipped at maintainer request for this focused test-harness fix.
• If a ticket exists: Does your pull request title follow the pattern "THRIFT-NNNN: describe my issue"? — No ticket exists.
• Did you squash your changes to a single commit? (not required, but preferred)
• Did you do your best to avoid breaking changes? If one was needed, did you label the Jira ticket with "Breaking-Change"?
• If your change does not involve any code, include [skip ci] anywhere in the commit message to free up build resources. — Not applicable; this changes Python test-harness code.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.