Continuations of XCMP implementation

[coax1d]

• Use SubXT to formulate and submit proofs from Sender to receiver and verify they are accepted.
• After accepting a proof of the XCMP messages on the receiver process the Xcm messages upstream #103
  • Instead of concatenating all the XCMP shards into a single blob, structure them so they can be decoded correctly on the other side possibly using a couple custom data structures see XcmpDataProvider Impl
  • How to convert XCMP messages to XCMs and sent upstream to be processed in the executor, see how DMP, HRMP handle this
• XcmpChannelTrieRoot To ParaBlock header Via Digest #74
  • Add new cumulus primitives to create DigestItem for the XcmpChannelTrieRoot
  • Alter cumulus-pallet-parachain-system on_finalize() to deposit the new digest item
• Store All Mmr Roots on Sender to Receiver channel, paraheaders from sender, Beefy MMR leaves #110
• Generate Multistage proof from Sender to Receiver on Relayer #111
  • Generate Proof of Beefy MMR leaf inclusion in Beefy MMR
  • Generate Proof of Sender Paraheader inclusion in Paraheader merkle tree (root is stored in Beefy leaf)
  • Generate Proof of MMR root inclusion in xcmp tree(Root is in the Senders ParaHeader digest)
  • Generate Proof of XCMP message inclusion in Sender Mmr
• Verify Multistage proof from Relayer on Receiver #112
  • Verify stage 1 of proof, since Beefy is an MMR can use MMR module on Receiver to Verify the proof the first part of the proof statelessly
  • Verify stage 2 by extracting the ParaHeader Tree Root from stage 1 proof and comparing it against stage 2 proof
  • Verify stage 3 by extracting the XCMP Root from stage 2 proof and comparing it against stage 3 proof
  • Verify stage 4 by extracting the MMR root from stage 3 proof and comparing it against stage 4 proof
• Update verifying Relayer Proofs which check the XcmpChannelRoots instead of just the MmrRoot hack #101
• Test submitting Relayer proofs and sending messages from Parachain A to B and see that everything flowing correctly. Send Mocked XCM traffic #102
• Refactor approach to not instantiate multiple pallets and instead have one pallet access various Mmrs #94
• Add in trustless initialization of Xcmp Channels. Something similar to HrmpNewChannelOpenRequest and HrmpChannelAccepted #95