draft: txe native world state

[sklppy88]

Pinging @Thunkar, the TXE (and generally)👨‍🔬.

The biggest difference between old world state and new world state is that during the synching of blocks, the changes made to the trees MUST match the block itself. This gave me a bit of pain because it now has to more accurately model the side effect emission / collection of what happens in a tx.

Particular things to pay attention to:

1. New world state assumes block numbers start at 1, hence we no longer start at 0. This means that we were off by 1 in a few places.
2. Is the current behavior of the replacement of "getLatest" okay ? Previously we were using a different model, but now we need to make a distinction between what is happening in "this" block for public state, and what will be collected to be applied for the next block, for private state (outside of the execution note cache).
3. Emitting arbitrary note hashes / nullifiers (not through notifying note) is a tricky flow. I'm not sure if my approach is robust enough here, we don't use the note cache because we don't have access to the full note.

[sklppy88]

• draft: txe native world state #10478 👈 (View in Graphite)
• feat: emulating blocks and correct (non-partial) note discovery in txe #10356
• master

This stack of pull requests is managed by Graphite. Learn more about stacking.

[Thunkar]

Careful!

We used to do this because this.blockNumber in TXE means "the block I'm currently building" and we set the private context to "-1" so that we could do historical proofs.

Will this change break that flow?

[Thunkar]

Nice job! However I think this shows we really need a refactor and have some sort of TXE state machine that keeps track of the (now a lot more realistic) execution model.