Extend Rosetta to work with MyCelo testnets

[eelanagaraj]

Description

• Adds geth.networkid flag
  • This is nessary when genesis ChainID != networkID, as is the case for some mycelo-based testnets
• Adds functionality to fetch necessary state for Rosetta (core contract addresses, GPM, CarbonOffsetPartner) from a static block when initializing the monitor & specifically prompted via --monitor.initcontracts; note this is not done by default.
  • This is necessary for mycelo testnets where this information is already configured at genesis. (Rosetta by default only listens for updates which is not sufficient if there is initial state at genesis.) While this is primarily useful for testing with mycelo-based testnets right now, this should also help with future L2 changes.
  • Design note: this follows the structure of block_processor. A larger refactor is possible here but I felt that to be overkill for now (especially since we will need to make more sweeping changes for L2 work in the future) and would unnecessarily introduce possible complexity/changes to a relatively stable part of the Rosetta service.

Tested

• Manually tested this:
  • using the globaltestnet (with longer reconciliation tests)
  • spun up a 1-validator mycelo testnet and ran a rosetta node against this

[palango]

The changes look good, I'm just wondering if we can remove the --monitor.initcontracts flag and instead somehow detect if the initial state is set in the genesis block. But maybe that's not worth the effort.

[eelanagaraj]

Thanks! and yeah I couldn't find a non-hacky way immediately (and wanted to make sure this minimally affects existing rosetta, i.e. needing to explicitly turn this on) but I imagine with L2 stuff in the future we can spend more time finding a way to do that! If any ideas occur to you later, happy to update!