* core/state, cmd/geth: streaming json output dump cmd + optional code+storage

* dump: add option to continue even if preimages are missing

* core, evm: lint nits

* cmd: use local flags for dump, omit empty code/storage

* core/state: fix state dump test

This function searches for an event+parameters in the ABI and returns it if found.

Co-authored-by: Victor Tran <vu.tran54@gmail.com>
Co-authored-by: Guillaume Ballet <gballet@gmail.com>

This PR updates a comment about the maximum client subscription buffer
to reflect changes made previously, and fixes a test that wouldn't fail
when wantError == true but execution did not return an error.

* graphql: check the integrity of the cdn files

* graphql: omit go-bindata

* core: move TxPool reorg and events to background goroutine

This change moves internal queue re-shuffling work in TxPool to a
background goroutine, TxPool.runReorg. Requests to execute runReorg are
accumulated by the new scheduleReorgLoop. The new loop also accumulates
transaction events.

The motivation for this change is making sends to txFeed synchronous
instead of sending them in one-off goroutines launched by 'add' and
'promoteExecutables'. If a downstream consumer of txFeed is blocked for
a while, reorg requests and events will queue up.

* core: remove homestead check in TxPool

This change removes tracking of the homestead block number from TxPool.
The homestead field was used to enforce minimum gas of 53000 for
contract creations after the homestead fork, but not before it. Since
nobody would want configure a non-homestead chain nowadays and contract
creations usually take more than 53000 gas, the extra correctness is
redundant and can be removed.

* core: fixes for review comments

* core: remove BenchmarkPoolInsert

This is useless now because there is no separate code path for
individual transactions anymore.

* core: fix pending counter metric

* core: fix pool tests

* core: dedup txpool announced events, discard stales

* core: reorg tx promotion/demotion to avoid weird pending gaps

params: add new bootnodes

cmd, graphql, node: graphql flag polishes, les integration

When cancelling the context for a call on a HTTP-based client while the
call is running, the select in requestOp.wait may hit the <-context.Done()
case instead of the <-op.resp case. This doesn't happen often -- our
cancel test hasn't caught this even though it ran thousands of times
on CI since the RPC client was added.

Fixes #19714

* miner: don't update pending state when no transactions are added

* miner: avoid transaction processing when pending block is already full

accounts/abi/bind/backends: use EIP155 on the simulated chain

accounts/keystore: fix #19707, avoid keyword as variable name

metrics: gather and export threads and goroutines

* appveyor: bump to Go 1.12.6

* vendor/vendor.json: govendor fetch github.com/karalabe/usb/^

* les: reject client connection if it makes too much invalid req

* les: address comments

* les: use uint32

* les: fix variable name

* les: add invalid counter for duplicate invalid req

renamed NewKeyStoreFromTransactor to NewKeystoreTransactor
fixed godoc

les/handler: avoid lookup missing state

eth: enforce stricter known limits on idle peers

EIP-695 was written in 2017. Parity and Infura have support for this
method and we should, too.

The dialer limits itself to one attempt every 30s. Apply the same limit
in Server and reject peers which try to connect too eagerly. The check
against the limit happens right after accepting the connection.

Further changes in this commit ensure we pass the Server logger
down to Peer instances, discovery and dialState. Unit test logging now
works in all Server tests.

* eth, les: reject stale request

* les: reuse local head number

core, ethdb, metrics, p2p: expose various counter metrics for grafana

* Fix file system access for Windows

* Encapsulate file accesses

* Style fixes

* vendor: remove unused dependencies

These were used by swarm code, which has now migrated to its own repository.

* travis.yml: remove sudo requirement for test builders

These needed sudo to run FUSE tests for swarm.

* p2p/enr: add entries for for IPv4/IPv6 separation

This adds entry types for "ip6", "udp6", "tcp6" keys. The IP type stays
around because removing it would break a lot of code and force everyone
to care about the distinction.

* p2p/enode: track IPv4 and IPv6 address separately

LocalNode predicts the local node's UDP endpoint and updates the record.
This change makes it predict IPv4 and IPv6 endpoints separately since
they can now be in the record at the same time.

* p2p/enode: implement base64 text format
* all: switch to enode.Parse(...)

This allows passing base64-encoded node records to all the places that
previously accepted enode:// URLs. The URL format is still supported.

* cmd/bootnode, p2p: log node URL instead of ENR

...and return the base64 record in NodeInfo.