Allow for control over TCP socket

[bigs]

In order to leverage the daemon for purposes of integration testing and the testbed more generally, we need to have remote-dialable daemons. I'd vote for abstracting the listen path into a multiaddr (similarly, client listen paths should be multiaddrs) and adding support to dial unix sockets to go-multiaddr-net.

[bigs]

oh... i just kind of realized the fundamental incompatibility with unix sockets and multiaddr...

[vyzo]

Well, we don't need to dial, we need to listen.

But yeah, for testing purposes agreed that this could be useful.

[raulk]

Does the multiaddr spec define a strategy to deal with values containing slashes?

[raulk]

@vyzo for incoming streams the daemon needs to dial as well, no? To deliver that steam to the app. Also there was a case in pubsub/DHT with validators?

[bigs]

yeah @raulk that was my thinking

[anacrolix]

I think this may have been implemented now? At least the tests have been?

[vyzo]

Yes, this has been implemented; closing.