I'd recommend to think about ways to split this up into multiple Pull Requests instead. At least two, maybe even three or more pull requests. Roughly this could look like the following approach:

1. one PR that (only) adds all LL specific fields / data structures to the consensus messages according to the spec (DA ready consensus #205)
2. another one that makes use of them, additionally (but optionally) to the existing tendermint logic (partly dependent on Implement reading from IPLD merkle dag #194)
3. one PR that deprecates the now obsolete tendermint fields: PartSetHeader, Parts etc (depends on Implement reading from IPLD merkle dag #194)
4. One PR that changes the hash in the BlockID to a simple hash (instead of merkelization)

The nice thing about this approach is that 1. and somewhat 2. as well are independent of #194, #85, and #182 etc. and can be reviewed and merged independently and faster. Only after 3. we will have parity with the Spec and can close #181. In between we would have data structures that kinda support both: vanilla tendermint and LazyLedger main chain logic.

This pull request has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

@Wondertan I'll close this PR but will leave the branch around.

@liamsi, ok. I was thinking about closing it as well at some point, as it does things in the wrong way.