protolambda

I think I addressed all open review, let me know if there are other things to test or change.

Note: the time related tests are flaky on a slower machine, since they rely on the existing auth code, which uses `time.Now`. Maybe we can refactor the existing auth code...

The auth header creation is already deferred to just before the `client.Do`, the timestamp creation is accurate. The problem with CI turned out to be different: when testing boundaries with...

Tests passed, but then got interrupted after I enthusiastically changed pushed a second commit to use `NumericDate` with truncated time based on desired precision, instead of changing the global precision...

While this PR was stuck the execution API specs changed the expiry time from 5 seconds to 60 seconds. So I'm going to update the tests + revert the seconds->milliseconds...

The bindings have a separate go module now, with a single makefile, which is checked in CI to be equal to the bindings that are generated from the latest contracts:...

> It currently has the same name as upstream Geth. Renaming it would allow us to more easily import it into projects. It sounds nice, and this is what `optimism`...

IMHO we should keep the L1 and L2 geth implementations compatible. Our goal is to make the diff as small as possible, the library should be interchangeable aside from a...

Starting with a feature flag in the config and the replace directive would improve our ability to contribute back and forth with L1 a lot, very valuable long-term. I'll rename...

In the new optimism upgrade (bedrock) we're adopting regular block based bundling of L2 txs, that should fix this