Fix interstitial sprout anchors check

[conradoplg]

Motivation

There was an issue with the interstitial Sprout anchor check, described in #3236

Specifications

https://zips.z.cash/protocol/protocol.pdf#joinsplit

Designs

Solution

• Change state to add the matching note commitment tree with each Sprout anchor
• Also add it to non-finalized state (Chain)
• When checking the anchor, build the interstitial trees starting from the tree that matches the anchor (and not from the tree of the finalized tip)
• Update state documentation

Closes #3236 and #3210.

Review

Not particularly urgent, but it's the final consensus check. Usually @dconnolly and/or @upbqdn could review but @teor2345 might want to take a look.

After accepting but before merging we need to update the cached state in CI (or disable the CI job until we do that).

There is no specific test - I tested by syncing to the tip:

Reviewer Checklist

• Code implements Specs and Designs
• Tests for Expected Behaviour
• Tests for Errors

Follow Up Work

[codecov]

Codecov Report

Merging #3283 (365e429) into main (6c787dd) will decrease coverage by 0.07%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main    #3283      +/-   ##
==========================================
- Coverage   77.83%   77.76%   -0.08%     
==========================================
  Files         266      266              
  Lines       31449    31471      +22     
==========================================
- Hits        24478    24472       -6     
- Misses       6971     6999      +28     

[upbqdn]

All looks good. It's handy that the commitment trees could be stored under sprout_anchors.
I left only a few minor suggestions.

[upbqdn]

I can regenerate the cached state once someone else approves this PR.

[teor2345]

Looks good to me, but it needs a cached state rebuild.

[teor2345]

@upbqdn Conrado isn't available this week, so can you push all the changes you think are needed to this PR's branch?

I can re-approve after you're done.

I can regenerate the cached state once someone else approves this PR.

I think it's ready to go, after you make whatever changes you think are needed.

Just checking, are you going to regenerate the cached state, update the image in this PR, then merge it to main?
(That way, the tests won't fail on main, because there's no gap between the version and cached state update.)

The hash in this line needs to be updated:

zebra/.github/workflows/test.yml

Line 57 in 454897b

--create-disk name="zebrad-cache-$SHORT_SHA-mainnet-canopy",image=zebrad-cache-13c6a826-mainnet-canopy \

[upbqdn]

Just checking, are you going to regenerate the cached state, update the image in this PR, then merge it to main?

Yes, exactly. I'm going to follow the steps in the onboarding doc. I refined them recently, so it should be smooth.

[daira]

Looks good. I apologize again for the hairiness of the spec here.

[upbqdn]

This PR also partially documents the consensus rules for #3211.

[upbqdn]

The following statement should not be a consensus rule

This PR documents all consensus rules mentioned in #3210 except for the statement above, so it also closes #3210.

See zcash/zips#582 for more details regarding the statement above.

[conradoplg]

I managed to recreate the cached state, updated it and it works 🎉

(But now CI is failing 😢 , see #3366)

[conradoplg]

@dconnolly I don't know why it dismissed your review, I just accepted your suggestion 😕

[conradoplg]

@dconnolly oops, fixed formatting 😅