lightninglabs / neutrino / 73

72%

Pull #61

travis-ci

multi: fix (mostly) issues with disconnect/reconnect and stall disconnection

This commit updates two places, and those which depend on them:

* The `queryPeers()` function is modified to re-check the connected peers
  before attempting to query them. This prevents queries failing due to
  peers disconnecting, since it allows to wait some time for new or
  reconnected peers. This refactor also reduces the number of goroutines
  used per query, which improves query speed. The per-peer query function
  is also factored out, allowing it to be used in other schedulers.

* The `sendGetcfheaders()` function is modified to wait for a minimum
  number of peers to be connected prior to broadcasting a request. This
  works *most* of the time in an environment of rapidly-changing peers
  such as the sync reorg tests; however, an update currently in
  development will replace this code with query-based code that's
  optimized for handling this case correctly.

In addition, tests are updated to provide more error information and
to support the new code; some small concurrency fixes are also added.

72 of 72 new or added lines in 2 files covered. (100.0%)

2788 of 3962 relevant lines covered (70.37%)

5527.11 hits per line