[EVM] Deploy on FEVM and Hyperspace testnet

[ryscheng]

Describe the feature you'd like to request

We should deploy on FEVM, as requested by the Filecoin Green team.

That includes FVM mainnet and Hyperspace testnet.

Describe the solution you'd like

• Deploy on FVM mainnet
• Deploy on Hyperspace testnet
• Add these deployments to the documentation (e.g. README)
• Add these deployments to the SDK

One big pot-hole to watch for --- the Graph does not support FVM
https://thegraph.com/docs/en/developing/supported-networks/

So we'll need to gracefully warn users that are trying to access our indexing client. We'll need to work with the Filecoin Green team to come up with a temporary solution that will work for them

Describe alternatives you've considered

Do it

[ottpeter]

I think the current active testnet is Calibration testnet. Hyperspace is obsolete.
https://docs.filecoin.io/networks/calibration/details/

[bitbeckers]

For reference form TG chat:

HypercertMinter is live on testnet: https://calibration.filfox.info/en/address/0xa16DFb32Eb140a6f3F2AC68f41dAd8c7e83C4941

Verifying isn't pretty, couldn't get the proxy to verify but at least the implementation was verified so you can use filfox to interact with the minter.

Two last bits:

1. trying to assign the contract ownership to a multi-sig, this appears to be Glif.io. But they don't support the address format generated by Ledger for creating multisigs? I could use some guidance here @TomHihihi
2. indexing of state. We're currently moving away from the Graph in favor of our own indexing solution. That indexer should ship this week or early next week. You can track progress in this repo: https://github.com/hypercerts-org/hypercerts-indexer/issues
   (The Graph doesn't support Filecoin, which doesn't matter since we're building our alternative and should ship that soon)

Filecoin has like 5 different address types, curious to learn if and how viem and the indexer handle this. From what I can see the EAM should handle this on RPC level

[bitbeckers]

Parking this for now for the reasons shared on TG:

The limiting factors are the lack of infrastructure, the translation between typical EVM patterns and FVM specificity and honestly the statement that 'indexing is challenging'.