Memo bundle design overview published
The requirements and a draft design for the NU6 memo bundle were presented for the first time and discussed.
Agenda summary 9
Zebra update: regtest and zcash_script
Zcash Foundation · Arya, Alfredo, Daira
- They published a new zcash_script version and updated dependencies, restoring Zebra's Windows support.
- They added PoW disabling, network name and genesis hash fields to the testnet parameters, and put in changes for NU5 regtest such as the median time past calculation on short chains.
- A minimal regtest works on the params regtest branch and will be merged soon. Alfredo confirmed that Zebra's end of support period was extended from 16 to 20 weeks.
- Daira said they need to know the exact value because it could affect the NU6 schedule.
ECC update: Zashi 1.0 Android launch and crate refactoring
Electric Coin Company · Kris, Str4d, Daira
- Zashi 1.0 Android launched and installs reached about 1,600. Among the bugs found by new users, they are first triaging the problems in the common wallet backend code.
- They are working with Arya on the NU6 memo change specification, and are also sorting out issues related to ZIP-315, the best practices standard for wallet developers.
- They are refactoring the Rust crates to separate the Sprout, Sapling, Orchard and transparent functionality from each other.
- Downstream users who use only one feature can drop the other dependencies, and it separates the transparent functionality for the zcashd wallet replacement from the purely shielded wallet, which also helps the pool deprecation path.
ECC and ZF zcashd deprecation work
Electric Coin Company / Zcash Foundation · Str4d, Daira, Alfredo
- They keep filling in the DAG of tasks to be done for the zcashd deprecation, and discussed the functions the wallet replacement architecture must have.
- Str4d is reviving the P2SH script distribution analysis script they built in 2019-2020 and gathering data on the items to implement on the transparent side.
- They also began reviewing the PR in which Sean started porting the Bitcoin script interpreter to Rust (librustzcash issue 1362). On the Zebra side they plan to move the scan crate inside Zebra to a separate repository.
- The upgrade of the ECC library dependencies is harder than expected and is likely to start after ZconV. They are discussing with ECC engineers where to put the indexer.
Zcash Shielded Assets (ZSA) update
QEDIT · Vivek
- The burn digest was missing from the transaction ID digest and sighash calculation, so they added it to the ZIP and the implementation. Daira asked them to check whether other affected data went into the hash.
- The backward compatibility work in Orchard and librustzcash is mostly done. In Orchard, the base Orchard and Orchard ZSA bundles are built at the same time, and librustzcash handles v6 transaction processing.
- What remains is restoring the test vectors and some Orchard tests. They are regularly discussing with ECC whether the generalization put into Halo 2 should stay in Halo 2 or be kept specific to Orchard.
- They started a testnet for ZSA and completed the first step of bringing up a Zebra node on an ECS instance and connecting to it locally. Next week at ZconV they will present the asset swap design in more detail.
FROST demo and the share refresh feature
Zcash Foundation · Natalie, Conrado, Daira, Str4d
- The FROST team is preparing its ZconV talk. They are working on a refresh share feature that makes it possible to remove a signer without changing the address; key generation is redone but the public key stays the same.
- They are about to release version 2, which has breaking changes from the Rust component changes, and will also add a no_std flag, though not as the default. Daira and Str4d checked the security model.
- A sufficient number of honest signers must delete the old shares to exclude a malicious signer. Changing the threshold is not in scope this time and can be supported later.
Trailing Finality Layer and ZKProof 6 announcement
Electric Coin Company / QEDIT · Daira, Jon, Natalie, Pili
- Daira is preparing their ZconV talk and said the TFL book update reflecting Crosslink 2 could come right after ZconV. There is no implementation work on TFL yet, only specification work.
- ZKProof 6 takes place May 22-24, and a kickoff to select verifiers for formal verification (Groth16, Bulletproof, PLONK and others) is planned.
- There are keynotes by Jens Groth, Alessandro Chiesa and Kostas Halkias. ZconV is next week, and they asked people to send session questions in advance.
ZSF discussion and the NU6 decision timing
Daira, Str4d
- Under ECC's roadmap, what items to put in NU6 must be decided soon.
- ZSF discussion took place mostly among the ZIP editors and the Equilibrium side did not attend, so there was no new discussion this week.
Zcash memo bundle: overview and requirements
Electric Coin Company · Str4d, Daira
- Currently, putting the sender address in a memo is not authenticated. The proof for an authenticated reply address does not fit in a single 512-byte memo, and the approach in which light clients send trimmed memos complicates the security argument.
- The requirements are receiving variable-length memos longer than 512 bytes, granting the same memo viewing access to multiple recipients across several shielded pools, and keeping the number and length of distinct memos private.
- Ywallet already copies the same memo into the Sapling and Orchard outputs, which causes duplicate memos to appear in Zashi.
- A recipient must be able to tell whether they have memo access just by decrypting the note ciphertext, and memo pieces must be individually prunable without breaking transaction verification.
- Prunability is meant to let nodes respond locally, as with the abuse of Bitcoin OP_RETURN data. The 512-byte memo field per output is replaced by a memo key of about 32 bytes, and each memo piece carries AEAD overhead.
- The overhead is at most about 48 bytes per output, and as a privacy and size trade-off a single chunk can also be used. The ZIP-317 fee can also count the outputs and memo data separately.
Memo bundle use cases and design discussion
Kris, Str4d, Daira, Pacu, Dodger, Alfredo
- Kris proposed a shielded Ordinals protocol using structured multi-piece memos and authenticated reply addresses, similar to Bitcoin Ordinals.
- Str4d suggested thinking about payment disclosure that allows partial disclosure, and about cases of replacing transparent functionality on the shielded side. Daira pointed out that long memos may need additional proofs for verifiable encryption.
- The memo key is a bearer capability symmetric key, and it is important to keep a symmetric key for post-quantum properties. It was found that a memo tree holding memo keys also allows partial viewing.
- Pacu said use has been limited by the limits of unauthenticated memos, citing wallet remote configuration and forced-upgrade channels as examples.
- Dodger asked about memo expiry and pruning. Kris suggested convention-based pruning instructions, while Str4d and Daira worried that an expiry height becomes a new identifier and harms privacy.
- Daira explained that the security of the Poly1305 authenticator is about 2^-98 for a single memo but only online attacks are possible, so it is sufficient. The next call is May 16 at 21:00 UTC.
This summary is based on the original meeting notes; check them for exact details.