Sporadic test failures in BPD tests

Travis seems to fail surprisingly often in the BPD tests with this error:

```
[...]
File "/home/travis/build/beetbox/beets/beets/util/bluelet.py", line 404, in __init__
self.sock.bind((host, port))
OSError: [Errno 98] Address already in use
```

Here's an [example job](https://travis-ci.org/beetbox/beets/jobs/543462082#L1121), for instance.

The problem is that these tests, since #3190, use a socket to set up communication between a server and a client. But if two of these run at the same time, then the server runs into an error when trying to listen on that socket. I tried to mitigate this in a hacky way in https://github.com/beetbox/beets/commit/99778d9ece151736ea2447a70e3fe7efd146b6c8 by randomizing the port over 10,000 options, but failures of this kind are still happening pretty often on Travis.

A potential solution would be to catch this error and re-randomize a few times before giving up.

Start with the BPD tests and beets/util/bluelet.py around line 404, where the server binds its socket. Reproduce the tests under parallel execution and verify that address-in-use failures are handled with repeated port selection before giving up.

Written by the indexing model from the issue text.