SuperSystem: Add Emitter Contract - Basic Message Passing

[axelKingsley]

This PR has two components

Emitter Contract in op-e2e/interop/contracts

Now there is a directory where custom contracts can be used for message passing or other test behaviors. At the moment there is only one: emit.go. The Emitter contract simply takes the caller's data and emits it as a log event. This allows us to generate content on the Supervisor, and we can also use this contract to test the receiving side.
To update the binding, generate.sh does a forge build, extracts the bin and abi, and abigen creates a go binding.

Custom Contracts for L2s are currently stored in a map[string]interface{} to store arbitrary contract bindings in the future. This one is keyed as "emitter".

The Emitter Contract adds two functions to the SuperSystem interface. I'd like to avoid scope-creeping the interface, but for now this provides good abstraction for tests to use this contract:

• DeployEmitterContract - puts the emitter contract on the specified L2 by the specified user
• EmitData - uses the Emitter contract keyed as "emitter" for the given L2, and calls it on behalf of the specified user
• Contracts - does a map lookup and returns the contract item

AddL2RPC without Restarting Supervisor

As soon as the log emissions started, I found another bug in the Supervisor. This one is related to file closing -- in AddL2RPC, we Stop the backend before adding the L2 and then we Start it back up. However, in underlying calls, a file object is closed which never gets opened again.

As a symptom, log emissions for any chain except the most recently added would fail because the underlying file was closed to writing!

I could have tired to make the logDB re-openable, but the real solution was just to make it possible to add new L2s without stopping the existing monitors. It was easier than I expected, the main change required was that the code used to add L2s needed to know whether or not to start the chain-monitor too.

Testing

As soon as log emission worked, the closed-file issue started. By implementing these changes, I see each Chain Monitor start exactly once, and I see them process (or attempt to process) chain updates.

fixes: #11959

[semgrep-app]

Semgrep found 14 sol-style-input-arg-fmt findings:

• op-e2e/interop/testdata/src/emit.sol
  • L9 - Triage
• op-e2e/interop/testdata/scripts/ScriptExample.s.sol
  • L6 - Triage
  • L7 - Triage
  • L8 - Triage
  • L9 - Triage
  • L12 - Triage
  • L13 - Triage
  • L30 - Triage
  • L34 - Triage
  • L44 - Triage
  • L48 - Triage
  • 3 more - Triage

Inputs to functions must be prepended with an underscore (_)

_{Ignore this finding from sol-style-input-arg-fmt.}