[MIP-012] Add Bubblegum Instruction to close Merkle Trees

[nhanphan]

Submitter: imetandy
Email: robinson.andrewjames@gmail.com

Submission transaction: 2nM5worDUytxEK4rZqnsenPWewT5kkHEkuK6p9qkXEr3rfG79LZtNoQey3DEbEUc5VWfmdErEQwfpVQjXp2rGLaT
Title: Add functionality to close Merkle Trees
Links: https://github.com/solana-labs/solana-program-library/blob/a08ec509/account-compression/sdk/src/generated/instructions/closeEmptyTree.ts#L32

Compressed NFTs lifespans can be short lived depending on use cases.

We are using them to denote betting slips, which will be created and then burned at a later date once a market is resolved and users claim their winnings.

A market creator may choose a small merkle tree, and then want to expand the amount of cNFTs they can offer at a later date.

In this scenario, it would be useful to open a larger merkle tree account, migrate people to this tree, and close the previous one to reclaim the SOL.

It looks like this was planned / in place in the spl-account-compression program, but has not made its way into MPL libraries.

[MetaplexFoundation]

This Community MIP is now open for community feedback. The community feedback period will remain open for 14 days.

[MetaplexFoundation]

This proposal is now referred to at MIP-012. It is live for voting in the DAO for 14 days. If passed, it will be a Metaplex Foundation implementation using the proposed PR from the submitter.

https://dao.metaplex.com/dao/Metaplex/proposal/4KqC3Nww8Ync1mcFXnB8rqtbcbikGiZYeUxvk5XixuGA

[MetaplexFoundation]

This proposal passed. MIP-012 will be a Metaplex Foundation implementation.

[danenbm]

Agree there is an existing facility in the account compression programs to close an empty tree, which requires all leaf nodes to actually be zeroed out (never used or all leaves burned).

We need to also consider protections for not increasing cNFT spam. For example, if a V2 tree is created with permanent burn delegate is in place, then spam could be delivered to many wallets, and after a period of being viewable, burned by the spam creator (for the same cost as minting them), most of the original cost is reclaimed.

Do you have technical implementation ideas in mind for this proposal?

[imetandy]

Thanks @danenbm !

I take your point on the spam-ability but I don't think the current cost is prohibitive enough to be the sole deterrent.

We have also come up against the permanent burn delegate issue however, we build markets with cNFTs denoting positions.
If 'lost' positions are not cleaned up, then accounts can't be closed for the market to reclaim rent.

I'll admit I hadn't really scoped that into this request, but now you've brought it up I think it may have to be.
I will carve out some time towards the end of the week to rough out an implementation concept for review.

[MetaplexFoundation]

Hi @imetandy - have you had a chance to prepare an implementation concept for the Metaplex developer community to consider?

[imetandy]

See my PR here: LINK

I think this is enough to expose the correct ix to devs for this outlined use case.

I have not addressed @danenbm point for increased spam, I do believe its valid, but probably out of scope for this MIP.

See my note on the PR regarding a timed delegation strategy. I think this could be a neat solution to handle short-lived cNFTs, but will require a more thought out plan. Since this was a fairly trivial implementation, I'd be happy to scope a concept around a delegation timer, if we want to extend the scope?

[danenbm]

The PR looks good and I agree with the assessment that spam is actually probably not a huge issue given that you can make a tree with no canopy for less than 0.5 SOL so its probably not much of an issue to get that back.

We could add a delegation timer to the scope if you have a good idea around that, but I think we are also happy to move this to a vote with just the close tree if it helps move the MIP along faster

[imetandy]

Thanks @danenbm

Yes let's just stick with the tree closure for this MIP if that works for y'all.

I've not enough time to really jam on the delegation timer solution so I'll submit another MIP for that if I manage to rub the right braincells together.