p2p/discover: split out discv4 code

This change makes getBalance, getCode, getStorageAt, getProof,
call, getTransactionCount return an error if the block number in
the request doesn't exist. getHeaderByNumber still returns null
for missing headers.

* consensus,core,miner: avoid overhead of creating a new block

* consensus: nitpick dot

* consensus: fix some comment formatting nits

This change restructures the internals of p2p/discover to make room for
the discv5 code which will soon be added to this package.

- packet type names now have a "V4" suffix.
- ListenUDP returns *UDPv4 instead of *Table. This technically breaks
  the API but the only caller in go-ethereum is package p2p, which uses
  a compatible interface and doesn't need changes.
- The internal transport interface is changed to make Table reusable for v5.
- The 'lookup' code moves from table to transport. This required
  updating the lookup unit test to use udpTest instead of a custom transport.

eth: enforce chain above CHT before accepting txs into the pool

* typo fix

* change to stores

* swarm/network: fix data races in TestInitialPeersMsg test

* swarm/network: add Kademlia.Saturation method with lock

* swarm/network: add Hive.Peer method to safely retrieve a bzz peer

* swarm/network: remove duplicate comment

* p2p/testing: prevent goroutine leak in ProtocolTester

* swarm/network: fix data race in newBzzBaseTesterWithAddrs

* swarm/network: fix goroutone leaks in testInitialPeersMsg

* swarm/network: raise number of peer check attempts in testInitialPeersMsg

* swarm/network: use Hive.Peer in Hive.PeerInfo function

* swarm/network: reduce the scope of mutex lock in newBzzBaseTesterWithAddrs

* swarm/storage: disable TestCleanIndex with race detector

cmd/utils, node: switch over default peer count to 50

* core: lookup txs by block number instead of block hash

Transaction hashes now store a reference to their corresponding
block number as opposed to their hash. In benchmarks this was
shown to reduce storage by over 12 GB.

The main limitation of this approach is that transactions on
non-canonical blocks could never be looked up, however that is
currently not supported.

The database version has been upgraded to version 5 and the
transaction lookup process is backwards-compatible with the
prior two transaction lookup formats prexisting in the
database instance. Tests have been added to ensure this.

* core/rawdb: tiny review nit fixes

 accounts: switch Ledger derivation path to canonical one

* eth/downloader: enable unsync-protection for light client

* eth/downloader: fix tests

* cmd, eth, miner: disable advance sealing if user require

* cmd, console, miner, les, eth: wrap the miner config

* eth: remove todo

* cmd, miner: revert noadvance flag

The reason for this is: if the transaction execution is even longer
than block time, then this kind of transactions is DoS attack.

trie: approximate the wasted cache metaspace closer

eth, les, light: enforce CHT checkpoints on fast-sync too

appveyor.yml: Upgraded to Go 1.12.4

core, eth, les, light: avoid storing computable receipt metadata 

swarm/storage/feed: add context handling/cancellation to Swarm Feeds lookup, fix bad hint lookup bug (#19353)

* swarm/storage/feed/lookup: Add context handling/forwarding

* swarm/storage/feed/lookup: Add test to catch bad hint

* swarm/storage/feed/lookup: Added context cancellation test

 This makes geth fails instead of falling back to local keystore, if the command line flag `--signer` is used

* cmd/clef, signer: make fourbytes its own package, break dep cycle

* signer/fourbyte: pull in a sanitized 4byte database

* swarm/api: fix file descriptor leak in NewTestSwarmServer

Swarm storage (localstore) was not closed. That resulted a
"too many open files" error if `TestClientUploadDownloadRawEncrypted`
was run with `-count 1000`.

* cmd/swarm: speed up StartNewNodes() by parallelization

Reduce cluster startup time from 13s to 7s.

* swarm/api: disable flaky TestClientUploadDownloadRawEncrypted with -race

* swarm/storage: disable flaky TestLDBStoreCollectGarbage (-race)

With race detection turned on the disabled cases often fail with:
"ldbstore_test.go:535: expected surplus chunk 150 to be missing, but got no error"

* cmd/swarm: fix process leak in TestACT and TestSwarmUp

Each test run we start 3 nodes, but we did not terminate them. So
those 3 nodes continued eating up 1.2GB (3.4GB with -race) after test
completion.

6b6c4d1c changed how we start clusters
to speed up tests. The changeset merged together test cases
and introduced a global cluster. But "forgot" about termination.

Let's get rid of "global cluster" so we have a clear owner of
termination (some time sacrifice), while leaving subtests to use the
same cluster.

* clef: bundle 4byte db into clef, fix #19048

* clef: add go-generate directive, remove internal abidb parser tool

* cmd/clef: extend go generate to format asset file

* metrics/prometheus: added prometheus http server and metrics collector

* metrics/prometheus: minor cleanups

* metrics/prometheus: named keys instead name in tag

* metrics/prometheus: minor typo cleanups, sorted report