Test P<>K bridge-init calls using chopsticks

[serban300]

We should test the bridge-init calls using chopsticks before actually executing them

Submitting the calls and checking the effects on the storage would be a first step.

With chopsticks we won't be able to run the relayer, since chopsticks doesn't come with the grandpa subscriptions and many RPCs. So as a second step it would be nice if we could take the forked chains and run them in zombienet somehow.

[serban300]

We can simulate an approved preimage in chopsticks with something like this:

const number = (await api.rpc.chain.getHeader()).number.toNumber()
// TODO: fill in
const extrinsic: HexString = "0x...";
const data = hexToU8a(extrinsic)
const hash = blake2AsHex(data, 256)

await api.rpc('dev_setStorage', {
    Preimage: {
      PreimageFor: [
        [
          [[hash, data.byteLength]],
          u8aToHex(compactAddLength(data))
        ]
      ],
      StatusFor: [
        [
          [hash],
          {
            Requested: {
              count: 1,
              len: data.byteLength,
            },
          },
        ],
      ],
    },
    Scheduler: {
      Agenda: [
        [
          [number + 1],
          [
            {
              call: {
                Lookup: {
                  hash: hash,
                  len: data.byteLength,
                },
              },
              origin: {
                system: 'Root'
              },
            },
          ],
        ],
      ]
    },
  });

await api.rpc('dev_newBlock', {count: 1});

[serban300]

Looks like there is this tool that can be used for forking-off chains: https://github.com/hack-ink/subalfred
And here is an example on how to use it: https://github.com/hack-ink/subalfred/blob/main/example/cmd/state/src/main.rs

Trying to experiment with it a bit. The plan would be to:

• fork the chains with chopsticks
• execute the preimages
• fork-off the chopsticks chains with subalfred
• start the forked-off chains in zombienet

[xlc]

If you can list all the RPCs used by relayer, we maybe able to provide a mock implementation in Chopsticks

[serban300]

@xlc thank you, it would be great if you could provide mock implementations in chopsticks. As far as I can tell there are only 2 additional RPCs that we would need apart from what is already available (@svyatonik please correct me if I'm wrong):

• grandpa_subscribeJustifications
• state_getReadProof

Do you think this would be doable ?

[svyatonik]

The full set of RPC calls we use is quite large: https://github.com/paritytech/parity-bridges-common/blob/polkadot-staging/relays/client-substrate/src/rpc.rs#L37-L170 (everything except BEEFY).

[xlc]

I think it is doable. However, Chopsticks is only a mocking client that doesn't actually finalize blocks using grandpa/beefy protocol so that it cannot provide real justifications.

There are two possible solutions:

1. Add a new feature to specify signing keys to Chopsticks so it can generate valid justifications.
2. Modify relayer to accept mock justifications. This will be useful to fork mainnet and run tests with it. As it is not possible to get real justifications from mainnet fork.

Let me know what do you think.

Also opened a tracking issue at Chopsticks AcalaNetwork/chopsticks#681

[serban300]

I think it is doable. However, Chopsticks is only a mocking client that doesn't actually finalize blocks using grandpa/beefy protocol so that it cannot provide real justifications.

There are two possible solutions:

1. Add a new feature to specify signing keys to Chopsticks so it can generate valid justifications.

2. Modify relayer to accept mock justifications. This will be useful to fork mainnet and run tests with it. As it is not possible to get real justifications from mainnet fork.

Let me know what do you think.

Also opened a tracking issue at Chopsticks AcalaNetwork/chopsticks#681

Thanks for the answer ! Moved the discussion to AcalaNetwork/chopsticks#681

[serban300]

Closing this issue since the Polkadot <> Kusama bridge has been initialized