SUAVE dns

[jinmel]

📝 Summary

Introduces SuaveEthRemoteBackendEnpointV2 flag to support multi domain kettle. suave.eth.remote-endpoint now accepts rpc endpoint configuration formatted in chain_id=rpc_url

📚 References

Generalizing suave execution namespace for multiple chains: flashbots/suave-specs#79

---

• I have seen and agree to CONTRIBUTING.md

[jinmel]

Should I keep the backwards compatibility by introducing new precompiles with different signatures? (i.e. buildEthBlockV2, simulateBundleV2 ...) It seems like the integration tests are failing due to changes in precompiles @dmarzzz @ferranbt

[ferranbt]

Overall looks good, could you write a unit test for the contracts_suave.go change and then another e2e integration? (suave/e2e).

Another question, I think the kettle should also define the port of the service. WDYT?

[jinmel]

@ferranbt Yeah I was also wondering if the kettle should set the port as well but that wouldn't be dns anymore. And also, if we are thinking of usage patterns like:

Suavex.simulateTxn("goerlisuavex", ...)

we should call this endpoint aliasing rather than dns. How about we just map a nickname such as "goerli" to a full url like "http://goerli.my-rpc.com:8545" rather than mimicking dns?

[jinmel]

@ferranbt After giving it some thought, I believe it's a better practice for developers if we resolve an alias name directly to its full URL, port included. It simplifies things by hiding the complexities of the URL endpoint. Let me know if you think otherwise

[ferranbt]

Only one comment to resolve but it is minor since the tests work.