devp2p/client: upgrade to eth/66

[ryanio]

This PR upgrades devp2p and client to communicate eth/66 since geth plans to deprecate eth/65 after london is activated on mainnet.

I updated devp2p's peer-communication.ts which was working for me (would appreciate someone else to run it to verify everything looks good), along with client's full sync with npm run client:start.

devp2p has no breaking changes as the payload logic is handled in the client, but the client will no longer support versions prior to eth/66 since structural changes needed to be made to support the encoding and decoding of messages to handle the new reqId field.

[codecov]

Codecov Report

Merging #1331 (63fce78) into master (103402b) will decrease coverage by 1.34%.
The diff coverage is 65.78%.

Flag	Coverage Δ
block	85.68% <ø> (+0.16%)	⬆️
blockchain	83.49% <ø> (+0.06%)	⬆️
client	83.94% <64.86%> (+0.05%)	⬆️
common	94.01% <ø> (ø)
devp2p	83.32% <100.00%> (-0.38%)	⬇️
ethash	82.83% <ø> (ø)
tx	88.36% <ø> (+0.11%)	⬆️
vm	79.23% <ø> (-3.41%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

[acolytec3]

I was able to successfully run peer-communication.ts so that looks good. As far as syncing the chain, here are my observations:

• mainnet, ropsten, and goerli all start syncing just fine
• rinkeby doesn't seem to find any suitable peers so not sure if the bootnodes there just haven't adopted eth/66 or not
• kovan won't start syncing though that doesn't appear to be an issue related to this PR as it throws the same error when trying to sync from the master branch

[acolytec3]

Ironically, it looks like my suggested change is causing the CI to fail though I can say with confidence that this is not always falsy. Is this a case of //@ts-ignore or whatever the equivalent linter ignore command is?

error: Unnecessary conditional, value is always falsy (@typescript-eslint/no-unnecessary-condition) at lib/sync/fetcher/blockfetcher.ts:33:10:
  31 |       })
  32 |     )[1]
> 33 |     if (!headers) {

[holgerd77]

Cool! 🎉