Add possibility of proving additional relay storage keys in set validation data

[girazoki]

Chains like Tanssi require collators to prove additional relay storage keys (in this case, the para head of a particular parachain-id). The way chopsticks works right now is that in the set_validation_data a new relay storage root is computed and injected based, among other things, on the incoming xcms and the relay slot number.

It could be good that we can add a set of additional key_values that we can inject in the generation of the proof, in a similar manner to how we did it with rust: paritytech/polkadot-sdk#1757. Tipically the relay storage root provided in set_validation_data is used to verify any other inherents proving relay-storage, so we need would need to take any additional key-value pair into account in the mocked version.

Additionally, it would be good if the trie-nodes generated by the proof were accessible for other inherents. Currently each inherent is quite isolated, and know from each other. Potentially there could be a way in which both set_validation_data and any other inherent use the same proofGenerator instance.

[ermalkaleci]

can you give an example how you intend to use it?

[girazoki]

yes we have our own sort of hacky implementation here: moondance-labs#6

Essentially we generate a proof for additional keys in set_validation_data, and then we use these trie nodes in a the new inherent where the new keys are proved

[ermalkaleci]

@girazoki is storage authorInherent.highestSlotSeen() best way to get slot from your chain?

[girazoki]

sorry @ermalkaleci I somehow missed this! yes it is!