Refactor op-e2e to use the same deploy script that is used in
production. There is only 1 way to deploy contracts now and
it goes through foundry. This removes the need to maintain
multiple sets of contracts deployments/devnets. It does
add a new tooling dep to the monorepo which is `geth`.
A top level `make` command is provided as well as some
docs in the `op-e2e` package on its usage.

The `devnetL1.json` deployconfig is the source of truth
for deploy configs in `op-e2e` now. In the long term,
we should delineate which config values are safe to alter
in memory and which are not. Some things are still coupled
together in weird ways, but I tried to add assertions where
I could to ensure that future people will get better error
messages.

To install geth, run the command:

`` `
make install-geth
```

To create the genesis state, run the command:

```
make devnet-allocs
```

This will run the forge deploy script against an instance of
`geth --dev` and then call `debug_dumpBlock` which will dump
the state into a format that we can inject at runtime into
the genesis of both `op-e2e` as well as the L1 of the docker
compose setup.

I tried to use `anvil` instead of `geth` but it does not
support this endpoint. It supports a different endpoint
that uses a compressed binary serialization of state,
it is easier to work with the JSON.