op-node: sequencing better encapsulated, now with events

[protolambda]

Description

Depends on #11040

work in progess

Refactors the op-node sequencer code to:

• interface with engine through events, no synchronous shared engine state anymore
• encapsulate sequencer start/stop logic (move out of Driver)
• Update engine deriver to handle sequencer functionality:
  • Each engine block processing step is its own input-event + event handler function.
  • No more global engine block payload-building state. The legacy engine-controller is only used to write and flush forkchoice updates, and the old block-insertion handling (can be refactored in follow-up PR to reuse more of new code)
• Change metered-engine metrics/logging to be default, get rid of old MeteredEngine wrapper, since it wrapped the non-events engine interface.

Tests

• TODO sequencer tests need to be updated

Additional context

This is the last major component that synchronously interacts with the engine-controller state. After this we can remove a decent amount of legacy engine-controller code.

Metadata

Fix #10979

[semgrep-app]

Please create a GitHub ticket for this TODO.

_{Ignore this finding from todos_require_linear.}

[semgrep-app]

Semgrep found 28 golang_fmt_errorf_no_params findings:

• op-chain-ops/upgrades/l1.go
  • L116 - Triage
  • L120 - Triage
  • L198 - Triage
  • L202 - Triage
  • L282 - Triage
  • L286 - Triage
  • L392 - Triage
  • L396 - Triage
  • L400 - Triage
  • L404 - Triage
  • 17 more - Triage
• op-chain-ops/safe/batch.go
  • L200 - Triage

No fmt.Errorf invocations without fmt arguments allowed

_{Ignore this finding from golang_fmt_errorf_no_params.}

Semgrep found 2 marshal-json-pointer-receiver findings:

• op-chain-ops/safe/batch.go
  • L131 - Triage
  • L274 - Triage

MarshalJSON with a pointer receiver has surprising results: golang/go#22967

_{Ignore this finding from marshal-json-pointer-receiver.}