When the op-node is closed, it also closes the P2P node. To prevent future usage, `n.p2pNode` is set to `nil`. However, the AsyncGossiper can still sometimes send gossip messages to the `OpNode` struct after close. Since `n.p2pNode` is `nil` at this point, the node would panic. This PR updates the `OpNode` implementation to check for this case. Access is protected via a mutex to prevent concurrency problems.

Restores the correct asterisc RISCV bytecode, moves the contract
to the vendor folder, and changes CI so that it performs the check
on the asterisc bytecode when attempting to merge into develop.

* op-deployer: Load CLI options into the bootstrap disputegame config

* op-deployer: Simplify bootstrap disputegame to only deploy the disputegame contract

* update superchain-registry dependency

brings in the unichain sepolia holocene_time

* update dependency on op-geth

Sometimes, disconnection events happen after the sync client is closed. This can (among other things) lead to panics in tests due to logs occurring after the test has exited. This PR makes RemovePeer a noop when `s.closingPeers == true`, which is the same functionality as `AddPeer`.

This was causing flakes.

Now that multiple test processes are running at once, multiple threads are compiling the op-program client in the tests. This PR builds the client upfront to avoid this condition.

Requires moving from ethclient.Client to sources.L1Client since ethclient.Client requires a *rpc.Client which doesn't have any support for request timeouts.

* feat: delete unused deploy functions

* feat: Delete some more unused code

* fix: Unused imports

* feat: reorder Deploy.s.sol helpers

Cleans up import statments in all contracts within the src/
directory. We should probably add a semgrep rule or a check that
confirms that everything looks good. I'll leave that part for
another commit.

* feat: delete unused deploy functions

* feat: Delete some more unused code

* fix: Unused imports

Bumps [github.com/stretchr/testify](https://github.com/stretchr/testify) from 1.9.0 to 1.10.0.
- [Release notes](https://github.com/stretchr/testify/releases)
- [Commits](https://github.com/stretchr/testify/compare/v1.9.0...v1.10.0)

---
updated-dependencies:
- dependency-name: github.com/stretchr/testify
  dependency-type: direct:production
  update-type: version-update:semver-minor
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

* op-deployer: Add a proxy to handle RPC failures

Anvil is very finnicky when it comes to RPC responses. It'll blow up if it encounters any HTTP error, and does not support retries. This PR adds a locally-hosted proxy which sits between Anvil and the fork RPC source and handles retries automatically.

* lint

* lint again

* ignore lint

Signed-off-by: chuangjinglu <chuangjinglu@outlook.com>

* fix: use eth_getBlockReceipts for Infura RPC kind

* fix infura test

* ci: Convert all Go tests into a single job

Takes all the go-e2e and go-test jobs and combines them into one single super-job. This has the following benefits:

1. We can centralize all config for Go tests in one place and simplify our CI pipeline.
2. It leaves parallelism up to the Go runtime, which is more efficient.
4. It makes our test reporting more accurate.

As part of this PR, I also:

1. Created a new resource class just for Go tests. It runs on the large Latitude runner, but limits the number of parallel go test jobs to 3 to prevent resource starvation.
2. Enabled test retries on `gotestsum`. This should reduce the number of flakes we see from the Go tests by retrying them up to 2 times in the event that they fail. Failures are still captured by the test reporting, so we can address them.

* code review updates

* add LegacyMintableERC20 tests

* fixes

* add test for L1ChugSplashProxy

* fixes

* fixes

Most of the time there's no output from our Go tests, it means there's a bug. Lowering the timeout here will reduce the amount of time we spend waiting for jobs to fail. For long-running tests like fault proofs, the value is configured at 20m since those tests take longer.

* improve drippie & faucet tests and coverage

* Update packages/contracts-bedrock/test/periphery/faucet/Faucet.t.sol
Co-authored-by: smartcontracts <kelvin@optimism.io>

* Update packages/contracts-bedrock/test/periphery/faucet/Faucet.t.sol
Co-authored-by: smartcontracts <kelvin@optimism.io>

---------
Co-authored-by: smartcontracts <kelvin@optimism.io>

* ctb: Fix Holocene upgrade script

* improve README

* pass deploy-config as parameter to just run, other improvements

* remove commented-out variables

Co-authored-by: Michael Amadi <amadimichaeld@gmail.com>

Moves the task for linting test names to after the tests job so
that the contract artifacts actually exist. We don't compile
everything ahead of time and rely on forge test to do the
compilation in an effort to cut down total compile time.

lint-forge-tests-check-no-build was causing issues in CI because
it was running as part of contracts-bedrock-checks which depends
on contracts-bedrock-build. contracts-bedrock-build builds the
contract files EXCEPT for tests whenever contracts-bedrock has
changed. This means that contracts-bedrock-build would not build
any test files and therefore lint-forge-tests-check-no-build would
not find any test files to check and the job would do nothing.
However, when contracts-bedrock is NOT changed, the build job gets
the full set of artifacts INCLUDING test artifacts from the cache.
This would trigger the full check to actually be run and it would
properly fail.

This PR moves that particular check into contracts-bedrock-tests
so that it's guaranteed to have all of the compiled contracts.

The tests in the txmgr queue were asserting that calls complete within a certain time window. This was causing flakes, particularly when the CI runner was under load. Generally we should avoid tests that use fixed durations as a way of asserting that code functions correctly.

* op-node: log mgasps across block processing lifecycle

* op-node: add 'import_time' field to block processing log

* op-node: make log message more descriptive

* op-node: log legacy codepath for InsertUnsafePayload

* op-node: include final ForkchoiceUpdate in block insertion time

* op-node: remove unused BuildTime field from events

* op-node: encapsulate log logic in TryUpdateEngineEvent methods

* op-node: linter fix

* op-node: add comment and adjust sync log wording

* op-node: fix BlockHash, BlockNumber in log

* op-node: seq add BuildStarted to PayloadProcessEvent

* op-node: refactor getBlockProcessingMetrics, protect against divide-by-zero

This reverts commit 8f5c20ee.

* fix a deadlock

* add comment