Consider re-holepunching if latency is "not good enough"

[flub]

We used to have logic that would re-holepunch at regular intervals if the latency was "not good enough". We have some commented out constants related to this still lying around to remind us:

• GOOD_ENOUGH_LATENCY
• UPGRADE_INTERVAL

Currently we re-holepunch whenever we know of new addressing changes.

We need to figure out what we want here. Do we still need to re-holepunch on intervals if the latency is not good enough? Or is the address-change logic sufficient?

I'm tempted to believe that the address-change logic is sufficient and that re-holepunching on interval is more likely to result in needless churn, e.g. if you have an active path a different inactive path that is possible might have a better RTT and you switch. But once you send data the RTT goes to the same number anyway. Re-holepunching if there's new information OTOH seems like it covers this already in a better way?

So this might just be a noop and deleting the commented out constants and todos.