Below are some draft (and dirty) implementations that attempt to port "Ethereum 1.0" to Substrate. We list subgoals here and current status.
This first aims to be a demo for Substrate multi-trie usability. substrate-trie is not generic yet, so here we use Cargo's [patch] section to replace substrate-trie by a custom implementation.
One issue faced right now is that we cannot yet support multiple hash functions for multi-trie. The fact that we don't have generic Hasher in Substrate further complicates the issue. The current custom trie implementation uses a "bridge" that wraps an underlying hash db, and pretends that it is of the targeted hash (see trie/src/eth.rs). This solution (while works, and while doesn't incur much performance penalty) is dirty. In the future, this should be replaced by either a refactoring on generic hash function in Substrate, or a refactoring over HashDB.
We have dependencies in many places that either only works for Substrate or Parity Ethereum, while serve similar propose. This makes it not possible to directly pull in dependencies. Below is an incomplete list of crates that have this issue:
substrate-primitivesandethereum-types. This is mostly due toH256and affects many children crates such askeccak-hasherandrlp.paritytech/parity-common/trieandparitytech/trie.