To improve the concurrency of the Bedrock actor tests, I hooked up an ETH faucet. Now, the actors can use the faucet to fund themselves before each run.

* op-node: Add comments to pipeline stages

This explains each stage in a bit better detail.

* Update op-node/rollup/derive/attributes_queue.go
Co-authored-by: Javed Khan <javed@optimism.io>
Co-authored-by: Javed Khan <javed@optimism.io>

* state-surgery: add immutables package

Add a package for dealing with immutables so that
the deployed bytecode can be set directly in state.
Each contract from the `op-bindings` package is imported
and deployed to a simulated backend. The `op-bindings`
package is enforced to be up to date by CI. Each of the
contracts was double checked that the arguments passed are
not immutables, the existing immutable values are hardcoded
into the contract itself. To handle the case where immutables
are dynamic, we will need to read in config and pass that
through. The methodology on how to do that is documented
in a TODO message.

The build step runs in CI meaning that changes to the
contract constructor interfaces will result in this
package failing to build.

This code could be slightly modularized to be used in
`op-e2e` for creating the initial L2 state. Will leave
that for an additional follow up PR.

* state-surgery: additional sanity check
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

* state-surgery: hardhat utils

Add hardhat utils for reading hardhat artifacts
from disk. This can read hardhat artifacts, buildinfo
and hardhat deploy artifacts. This is necessary for
implementing the state surgery in go, because these
files must be consumed as part of the state surgery process.

* ci: run state-surgery tests in ci
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

Add the rest of the predeploy contracts to `op-bindings`
so that they can be used as part of the state surgery process.
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

bedrock: v0.4.0

geth to v1.10.21
go to 1.19
golangci-lint to 1.48.0

* op-batcher op-proposer op-service: Introduce op-service

op-service is a utilities library that automates much of the boilerplate setup we've been duplicating across different daemons. This PR adds the following to op-batcher:

- Common setup utilities for logging, metrics, pprof, and RPC.
- For each of the above functions, a set of CLI flags and configs that can be used to parse + validate CLI configuration for that module.
- A base RPC server that implementers can extend.
- HTTP metrics utilities.

* update dockerfiles

* code review updates

* op-service: temporary replace, until op-service go module is canonical and can be resolved
Co-authored-by: protolambda <proto@protolambda.com>

* feat(contracts-bedrock,sdk): deploy config typing, hivenet

* bedrock: op-batcher,op-node,op-proposer flags consistency

* contracts-bedrock: hardhat deploy config review feedback

* contracts-bedrock: lint fixes

* contracts-bedrock: fix chain IDs in deploy configs

* ops-bedrock: remove l1 genesis timestamp env var
Co-authored-by: Matthew Slipper <me@matthewslipper.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

Also fix the issue with the deployments being
added to the gitignore when they shouldn't be.

This includes the contracts for the latest goerli
deployment.

* op-node: go 1.19 linting, update libp2p deps

* specs/meta: update go linting instructions

Co-authored-by: Mark Tyneway <mark.tyneway@gmail.com>

Modularize into 2 hardhat tasks, make them
easier to debug.

deps: fix yarn.lock

Fixes whitespace issues by adding quotes around
volumes.
Co-authored-by: wsdt <kevin.riedl@wavect.io>
Co-authored-by: wsdt <kevin.riedl@wavect.io>

It got out of date, this updates it. The changes
are very minimal.

* op-node: l1 source cache metrics

* op-node: update L1 source cache defaults

* op-node: goimports fix

* op-node: cache wrapper - use any instead of interface{}
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

Withdrawals don't seem to update balances correctly
for some reason, still need to debug why that is.

chore: add changeset

Version Packages

Develop -> Master

op-node: separate sequencing and publishing errors. Old payloads fail to publish, but should not hold back catch-up work of sequencer (#3186)

* pkg derive: wrap errors by severity

* pkg derive: wrap errors by severity

* pkg derive: fix call with 2 return values

* ci: fix golangci-lint warnings

* ci: fix attributes test

* pkg derive: update tests, restore err types

* driver: update to use err severity from derive

* op-node: fix retry callers

* pkg derive: reimplement the custom error type

* pkg derive: error - nits

* pkg driver: update Step in event loop to backoff

* pkg derive: match epoch hash with l1 block hash

* pkg driver: handle backoff error

* pkg driver: handle temporary errs with backoff

also include additional if / else statements to log errors

* op-node: req next step, with backoff, without blocking other events

* pkg derive: update error types

* pkg derive: typo
Co-authored-by: protolambda <proto@protolambda.com>

* update wsteth bridge address

* changeset
Co-authored-by: Will Cory <willcory@Wills-MacBook-Pro.local>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

* op-node: fix invalid payload insertion error handling

* fix goimports
Co-authored-by: Matthew Slipper <me@matthewslipper.com>

* ci: only run changeset integrety on PRs

It will fail on merges to develop

* ci: better fix

* actor-tests: Bedrock actor tests

- Moves actors into their own package
- Minor updates to logging, etc.
- Adds Cloudbuild config so that we can get Docker images on every commit to Develop
- Updates Actors to support Bedrock

Currently supported actors:

- Deposits
- Contract deployments
- Simple sends

* doctoc only look at specs

* goals for fixed int

* op-node: Fixed integer sizes in the channel header

This mainly modifies the channel_frame, but has some knock on effects
as the width of some of the fields have been reduced. The channel
frame code is also changed more than I expected due to differences
in API of working with fixed int vs uvarints in go. Otherwise the
code reads very similarly with using Reader/Writer APIs.

* op-node: fix frame unmarshal func to return correct err

* specs: update frame format specs

* Update op-node/rollup/derive/channel_bank_test.go
Co-authored-by: protolambda <proto@protolambda.com>
Co-authored-by: Matthew Slipper <me@matthewslipper.com>