contracts: add a method for building legacy bindings

feat(docs/op-stack): Show how to modify predeploys

fix(docs/op-stack): In superchain-diag, 2->3

feat(docs/op-stack): Add the getting in touch form

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

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

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

As requested by Conduit.

Closing: DEVRL-780

fix(ci): run tests for common-ts

feat(ctb): Add constructor to init the SD implementation

op-chain-ops: Extract concurrent state iterator into util

Tests for common-ts were not being executed in CI because the command
being triggered was "yarn test:coverage" and this command was a noop for
common-ts. Also fixes a bug in the tests that wasn't caught by CI
because of this.

fix(sdk): automatically create std/ETH bridges

fix(op-batcher): Fix PR #5186 Nits

fix(op-node): Fix Channel Out MaxSize Underflow

specs: Update spec channel handling to match the current implementation

fix: fix flag name typo for MaxStateRootElements

Pulls the concurrent state iterator into a re-usable library. Additional tests have been added to assert that the iterator touches every key in state at least once. This will allow us to perform a complete check of the OVM_ETH migration as the last step of the migration.

op-e2e: Adjust for updated op-geth API changes

op-chain-ops: Fix data race in error handling

fix log package in proposer/driver

Updates the SDK to automatically create the Standard and ETH bridges as
long as the L1StandardBridge provider is supplied.

The error channel in the OVM_ETH migration is buffered. In rare cases, errors were written to this channel in such a way that they were not processed by the collector goroutine. This meant that errors were not being caught by the `lastErr != nil` check, and were instead triggering the total supply check below.

This error reliably reproduces when running the new TestMigrateBalancesRandomMissing test. This test generates a random state, and randomly removes an address/allowance from the witness data. It runs this 100 times per test. We didn't catch this with the other random test because it doesn't exercise the error handling codepath.