HIP-46: facilitate routing by netid-to-oui

[dpezely]

This facilitates functionality that will be required if HIP-46 passes:

• Create new chain var, routers_by_netid_to_oui that contains sorted list of tuples
• Each tuple associates a NetID to OUI for a roaming partner
• New functionality is primarily contained within State Channel client code

[dpezely]

@evanmcc Here's some context, since you've been added as reviewer. Tests pass for others yet fail on my local system. Not sure if it's because they're running macos and I'm on Linux.

e.g., full_test() fails for me in init_per_testcase() when running via:

SC_DIR=/tmp/sc/ ./rebar3 ct --suite test/blockchain_state_channel_SUITE --group sc_libp2p --case test_full

Once we can get beyond that, there are a few blocks of commented-out code (marked with FIXME) in that same suite, netid_to_oui_test(), that need to be finished. It started life as a copy of full_test().

[macpie]

rebar3 ct --suite=test/blockchain_simple_SUITE.erl --case=routing_netid_to_oui_test --readable=true

[dpezely]

Within blockchain_simple_SUITE.erl, two existing tests are failing for reasons that seem unrelated to the new code: reload_test() and restart_test(). Since they're unrelated, I would have approved this PR if GitHub's UI would allow me. Because I created the initial PR, someone else will have to approve.

[dpezely]

For anything directly related to blockchain-core, my confidence in whether this is complete enough (or at least sufficient for this piece) is far less than with router repo.

With that caveat, I think this PR is ready for final review.

TODO: There still needs to be an official means for populating the new chain var, routers_by_netid_to_oui.

[jadeallenx]

please squash before merge

[dpezely]

Superseded by PR 1412