* ctb: fix more invariant flakes

* add invariant docs

* exclude deploy contracts

* remove deployconfig from invariant exclusion list

Bumps [github.com/jackc/pgx/v5](https://github.com/jackc/pgx) from 5.5.2 to 5.5.3.
- [Changelog](https://github.com/jackc/pgx/blob/master/CHANGELOG.md)
- [Commits](https://github.com/jackc/pgx/compare/v5.5.2...v5.5.3)

---
updated-dependencies:
- dependency-name: github.com/jackc/pgx/v5
  dependency-type: direct:production
  update-type: version-update:semver-patch
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

Updates contract deployments to include a new config value that
will deploy OptimismPortal2 over OptimismPortal. Updates devnet
tests to also include a new environment variable that will trigger
this deployment and run SDK tests for this deployment.

* Use engine API that matches HF

* Simplify engine client code by using EngineAPIVersion

* Return full method string from engine api version methods

* FCUVersion takes payload attributes

* Use a proper method

* Handle switch default case explicitly

* Return fcuV3 if attrs is nil

* Add test cases for engine api version methods

* Reduce the min size for large preimages in devnet to make it easier to reach the limit.
* Provide a buffer between target and max batch size to avoid the batcher splitting the large tx across multiple smaller frames.

Co-authored-by: Will Cory <willcory@Wills-MacBook-Pro.local>
Co-authored-by: Matthew Slipper <me@matthewslipper.com>

* ci: Slim down CI builder

The CI builder image is downloaded many hundreds of times per day. The larger it is, the longer it takes for each CI executor to spin up which increases overall CI runtime. It also causes a significant amount of bandwidth costs from GCP Artifact Registry. This PR reduces the size of CI builder by moving the Rust toolchain to a dedicated builder image which will only be used with the jobs that require it. The Rust toolchain accounts for more than 50% of the size of the builder image.

* add binutils

* fix path

* Update SVM

* have to source bash first

* bad copy

* add back elf tools

* ci: Refactor dependencies to reduce runtime

This PR makes the following changes in order to reduce CI runtime:

1. Move the long-running Cannon E2E tests into a scheduled job. This reduces E2E runtime by 8 minutes.
2. Updates the `devnet` task to leverage the contract artifacts and cannon pre-state created in order jobs. This job was also updated to use a Geth binary rather than compiling it from source. This reduces devnet runtime by 5 minutes.
3. Removes intermediate linting, allocs, and geth version check jobs by putting them all in `pnpm-monorepo`. `pnpm-monorepo` already builds the contracts, devnet allocs, and checks the Geth version in parallel so splitting these jobs out actually makes things slower due to CCI image download/environment spin-up overhead. The `pnpm-monorepo` job now plugs into a bunch of downstream jobs.
4. Refactors the action tests to take advantage of multiple test executors.
5. Swaps the `bedrock-go-tests` image to a simple CCI base image to reduce runtime (this job alone was taking ~1min due to spin up overhead).
6. Puts linting in `go-mod-download` to avoid spin-up overhead.

In sum, this PR reduces overall CI runtime from 22 minutes to ~13.

* Use develop as a trigger rather than schedule

* schedule fpp-verify

* Data Puller

* comment edits

* remove redundant getTx over RPC

* fix

* Add Write and Merge Functionality

* lint fix

* add gob tooling

* run timer

* batching ; fit and finish ; rebase

* drain results

* fit and finish 2

* error handle

* improved resiliency ; remove ENV

* filename changes

Removes the stateviz code from the repo as it has not
been used in quite some time. It was removed from the
docker compose setup. This is part of cleaning up code
in the monorepo periodically to make sure that we keep
the codebase clean.

stateviz was useful for visualizing the state of the rollup.

P2P tests were flaking sometimes due to the checks asserting that everything received was published. `require.ElementsMatch` requires that lists be equal, sans duplicates. Sometimes, however, messages would get sent multiple times/not be picked up in time for the assertions leading to errors like this following:

```
listA:
([]string) (len=29) {
 ... elided
 (string) (len=68) "0xd770621d3b0a196087ace7efcb0a8a61b353736526f6095946c714ec134b1648:1",
 (string) (len=69) "0x71a364bf88d4e7e049cfe1635e95c31c9d93435b03a4d20c8863637fc0eb319f:28",
 (string) (len=68) "0xd770621d3b0a196087ace7efcb0a8a61b353736526f6095946c714ec134b1648:1"
}

listB:
([]string) (len=29) {
 (string) (len=68) "0xd770621d3b0a196087ace7efcb0a8a61b353736526f6095946c714ec134b1648:1",
 ... elided
 (string) (len=69) "0xf84acde28fb7b630f9e5d91bddc7c2d4d2e1f4ce29a7d1a4551fee9ac828b10f:29"
}

extra elements in list A:
([]interface {}) (len=1) {
 (string) (len=68) "0xd770621d3b0a196087ace7efcb0a8a61b353736526f6095946c714ec134b1648:1"
}

extra elements in list B:
([]interface {}) (len=1) {
 (string) (len=69) "0xf84acde28fb7b630f9e5d91bddc7c2d4d2e1f4ce29a7d1a4551fee9ac828b10f:29"
}
```

This PR refactors to use `require.Subset` which handles this case while still asserting that everything received was in fact published.

* op-chain-ops: ecotone post-activation-checks

* op-chain-ops: Ecotone post-activation check fixes

* op-chain-ops: check-ecotone fix nonce source

* op-chain-ops: cancun functionality post-validation checks

* fix lint

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

* ctb: Add leftmost bond payout rule to FDG

* Update packages/contracts-bedrock/test/dispute/FaultDisputeGame.t.sol
Co-authored-by: clabby <ben@clab.by>

* address @clabby's review

---------
Co-authored-by: clabby <ben@clab.by>
Co-authored-by: Matthew Slipper <me@matthewslipper.com>

* contracts-bedrock, ci: remove slither

Remove the committed in `slither-report.json` since it is not being used
as part of code review. Also remove running it as part of CI. We can
choose to add it back when we have the time for somebody to own making
it a first class tool, but since we haven't triaged the database yet
it only adds noise to PRs by adding extra diff and extra overhead by
forcing a regeneration of the report.

It hasn't found any issues in our code and our other checks are much
better at finding issues in the code.

Leaving the slither commands in the `package.json` for now but if we
do not commit to using slither in a way that is very helpful for us
in 6 months or so, then we should just remove the commands as well.

Maybe we can adopt a way of using slither where it is used before
doing releases of contracts or before going to audit. Having the diff
as part of code review is not super helpful for code reviewers.
It just adds noise in the current form.

* Kick build

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

* indexer: Pad gas estimates

Indexer tests flake fairly often due to out of gas errors. Case in point below, from a recent test run. This PR updates the indexer tests to pad gas estimates by 20%. This technique is used elsewhere in the indexer E2E tests, but wasn't being applied consistently.

```json
{
  "from": "0x15d34aaf54267db7d7c367839aaf71a00a2c6a65",
  "gas": "0x1f3cc",
  "gasUsed": "0x1edb1",
  "to": "0x978e3286eb805934215a88694d80b09aded68d90",
  "input": "0x",
  "error": "execution reverted",
  "calls": [
    {
      "from": "0x978e3286eb805934215a88694d80b09aded68d90",
      "gas": "0x1889b",
      "gasUsed": "0x1889b",
      "to": "0x8887e7568e81405c4e0d4caaabdda949e3b9d4e4",
      "input": "0x",
      "error": "out of gas",
      "calls": [
        {
          "from": "0x978e3286eb805934215a88694d80b09aded68d90",
          "gas": "0x15560",
          "gasUsed": "0x1f37",
          "to": "0x1c23a6d89f95ef3148bcda8e242cab145bf9c0e4",
          "input": "0xcc731b02",
          "output": "0x0000000000000000000000000000000000000000000000000000000001312d00000000000000000000000000000000000000000000000000000000000000000a0000000000000000000000000000000000000000000000000000000000000008000000000000000000000000000000000000000000000000000000003b9aca0000000000000000000000000000000000000000000000000000000000000f424000000000000000000000000000000000ffffffffffffffffffffffffffffffff",
          "calls": [
            {
              "from": "0x1c23a6d89f95ef3148bcda8e242cab145bf9c0e4",
              "gas": "0x13d07",
              "gasUsed": "0xba8",
              "to": "0xffba8944650e26653823658d76a122946f27e2f2",
              "input": "0xcc731b02",
              "output": "0x0000000000000000000000000000000000000000000000000000000001312d00000000000000000000000000000000000000000000000000000000000000000a0000000000000000000000000000000000000000000000000000000000000008000000000000000000000000000000000000000000000000000000003b9aca0000000000000000000000000000000000000000000000000000000000000f424000000000000000000000000000000000ffffffffffffffffffffffffffffffff",
              "value": "0x0",
              "type": "DELEGATECALL"
            }
          ],
          "type": "STATICCALL"
        }
      ],
      "value": "0xde0b6b3a7640000",
      "type": "DELEGATECALL"
    }
  ],
  "value": "0xde0b6b3a7640000",
  "type": "CALL"
}
```

* Update indexer/e2e_tests/bridge_transactions_e2e_test.go
Co-authored-by: refcell <abigger87@gmail.com>

---------
Co-authored-by: refcell <abigger87@gmail.com>

Deduplicate code by using a common json write function from
`op-service` instead of using one off json write functions.
This reduces the overall code in the codebase and helps to
standardize the way that we write json to disk.

* op-chain-ops: remove post process step

Now the the L1 genesis is generated by foundry, there should be
no need to post process the L1 genesis state. Transactions would
fail because of the resource metering values in the optimism portal
contract being set as part of the genesis setup. The setup was
originally created when Go code was used to generate the L1 genesis.

There is a small chance that this fails in CI. Will close if it fails.

* op-chain-ops: fix build

* chain-ops: delete dead argument

The `postProcess` arg is no longer used so delete it

* op-node: fix build

* check prior fork

* move inside cfg.Rollup.Check

* reuse checkFork logic

* fix testcase

* move checkFork to top level

* add testcases

* op-challenger: Exclude length prefix when uploading large preimages.

Make OracleData field private to avoid unintentionally using the length-prefixed data.

* fix(op-e2e): upstream merge

---------
Co-authored-by: refcell <abigger87@gmail.com>

Implements items 1 and 2 on the [CI Fixes Plan](https://www.notion.so/oplabs/CI-Fixes-Plan-c7b3fee803c1480eb8e0512bff51e208):

1. Disables `op-e2e-ws-tests`
2. Fails faster on HTTP tests

* ci: update to ci-builder 0.41.0

Updates ci-builder to the latest release

PR to update to latest foundry: https://github.com/ethereum-optimism/optimism/pull/9311
Tag service: https://github.com/ethereum-optimism/optimism/actions/runs/7750682174/job/21137300601
Release in circleci: https://app.circleci.com/pipelines/github/ethereum-optimism/optimism/43501/workflows/b0cedd5b-d439-49b2-b568-b137ca02e23c

* op-chain-ops: hack fix for state serialization

An update to foundry must change the way that the state dump
is serialized as there is a panic with the genesis json parsing.
Adds a simple hack for handling it.

* typo: fix

The `Artifacts` contract can be inherited by any contract
that wants to operate with a deploy artifact JSON file.
It was originally created to work with hardhat deploy artifacts,
which no longer exist. This cleans up some of the code to make
it more obvious that it isn't for hardhat deploy anymore.

This has no functionality changes, just renames and cleans up.

* feat(op-e2e): large preimage e2e output cannon test

* fix: Attempt to wait until leaves are uploaded (finalized) and then time travel past the challenge period

* fix: remove args in callback constructor

* op-e2e: Make batch size configurable via an option instead of always using large batches.

* fix: defend step

* op-e2e: Wait for preimage instead of immediate check and some tidy ups.

* op-e2e: Reduce keccak challenge period in devnet

* op-e2e: Make timings self-consistent.

* Use errors.Is

* feat(op-challenger): simple clock with accessor and mutator

* fix(op-service): make simple clock concurrent

* fix(op-service): simple clock lock free atomics

* op-challenger: Use state matrix matching the prestate leaf instead of poststate when calling squeeze.

* fix(op-service): move simple clock interfaces to their usage

* op-e2e: Skip test until insertion is fixed.

* op-e2e: Fix typo

---------
Co-authored-by: Adrian Sutton <adrian@oplabs.co>

Updates foundry to the latest nightly release.

https://github.com/foundry-rs/foundry/releases/tag/nightly-84d98427e17370ff08ec34f00e3c7e539753a760

It includes https://github.com/foundry-rs/foundry/pull/6986 which will
help to make L2 genesis generation much more simple.

Reminder that we must always do releases of `ci-builder` using a nightly
foundry release that has not yet been pruned from github. When building
from source in `ci-builder`, there are issues at runtime. When
downloading the binary from github, there are no issues.

Will follow up with a release of `ci-builder` and then update the
version of `ci-builder` used in CI.

* Add `CANCUN_ACTIVATION` timestamp to preimage oracle

* chores

* fix: add the cancun activation timestamp

* fix: add cancun activation field to full deploy config testdata:

---------
Co-authored-by: refcell <abigger87@gmail.com>