CASE race condition?

[cecille]

Problem

There are some races in the cirque tests that point to a possible case race condition. If an address gets updated while a case session is being established, it looks like it will affect the session establishment by tearing it down as it is being established.

Proposed Solution

Look more into this and find a way to reliably test.

[cecille]

An addendum to this - if we get multiple A and AAAA records triggering the callback, we can swap the address between the unauthenticated setup and the encrypted exchange.

[cecille]

Obsolete - this was opened before the great case refactor