Add state to the validator-network and fix some network issues

[jsdanielh]

• lib: Go back to use 3 as the default network quorum value
  • Change the client code to use 3 as the default network quorum value.
    Also make the devnet scripts use this value for configuring the
    nodes.
• validator-network: Do less validator PeerId cache clears
  • Change the function that clears the validator PeerId to receive the request error such that it doesn't clear the cache if InboundRequestError::NoReceiver error was received. This error is an indicative that the requested peer has no receiver for such request which could happen when the peer is no running the specified aggregation and this is not an indicative of the PeerId cache entry being wrong. Thus in case of receiving this error, the network won't invalidate the cache entry for this validator.
• network-libp2p: Fix mishandling of FinishedWithNoAdditionalRecord
  • Fix the mishandling of FinishedWithNoAdditionalRecord since we could get such event even if the query.finish isn't called (which the network is calling once Quorum records have been received).
    This mishandling was causing some of the queries to hang indefinitely since no error nor result was being sent to the application layer.
    Now the network always pushes the result to the application layer once the FinishedWithNoAdditionalRecord is received.
• validator-network: Add some state to the PeerId cache
  • Add some state to the validator PeerId cache to make less queries as possible to the DHT.
    This also makes the validator-network able to quickly return if the DHT query is still in progress allowing upper layers to continue their process without blocking them until the query finishes or errors out.

Pull request checklist

• All tests pass. The project builds and runs.
• I have resolved any merge conflicts.
• I have resolved all clippy and rustfmt warnings.

[hrxi]

I found one potential issue, but the whole thing looks very good. :)