util: waitUntilConnected

[travisperson]

I've been going through the code and identifying patterns or ways to help clean up the code to make things easier to read and understand.

(Note: I've been working off of the in-progress PR #290)

There are a few tests that require multiple nodes to be connected (ping, pubsub, dht).

Running through this I found this setup code for ping, which uses a method waitUntilConnected, which doesn't actually connect any nodes, but instead just waits around till the nodes find each other (as far as I can tell).

https://github.com/ipfs/interface-ipfs-core/blob/f238e9f0ba8132dba51d3bac1eb491701f505ccd/js/src/ping.js#L38-L52

It appears that it heavily relies on peer discovery, which is dependent on how the node is configured, particularity for js-ipfs-core.

@alanshaw I believe you wrote this code. I believe it was added so that the tests itself did not have to attempt to find and connect to the peers, and instead could rely on the peers already being connected before conducting the pings.

Any reasons to use waitUntilConnected over simply calling ipfs.swarm.connect?

[alanshaw]

In hindsight I think this was a mistake and I'm going to send a PR to just connect the nodes that need to be connected, if the test is not explicitly testing discovery.

To phrase it another way, we don't want to have to skip tests that are testing features that aren't discovery related.

I've come to appreciate the state of discovery a whole lot more. In windows AFAIK there's no MDNS so waiting could take an indeterminate amount of time while the DHT bootstraps and our peers get registered on it and there's also this weirdness where in 0.4.14+ our spawned nodes never acquire any peers.