[Proposal] Spending notes which haven't-yet been inserted

Aztec
18

Doesn’t this suffer from the same issue as ordering public state updates, where A::foo could read C before even calling into A::bar?

Yes it does. I’ll be starting a new thread soon on ordering of private state.

Yeah, it’s the same issues.
This is what I was alluding to with this earlier comment: [Proposal] Spending notes which haven’t-yet been inserted - #6 by Mike - perhaps I could have been clearer.
We can hopefully solve both the private and public versions in a consistent way.

We have been discussing state ordering quite a bit. But @suyash67, @jean and I only recently realized it may require us to forward read requests to later kernel iterations (alongside new commitments and nullifiers which are already forwarded). I don’t think counters are sufficient without this. Or we could have kernel iterations per-frame like mentioned in the other thread.

@Mike and/or @spalladino do you agree that if we take the counter approach, read requests need to be forwarded to later kernel iterations too enable reading of pending commitments?

3 Likes