ZIP-317 in zcashd 5.5.0
The zcashd 5.5.0 release candidate, which includes the ZIP 317 changes, was approved and is due for release next week.
Agenda summary 8
Zebra audit progress
Zcash Foundation · Dodger, Deirdre
- The Zebra audit was briefly paused for Real World Crypto and has since resumed, and the issues it flagged are being fixed. No major problems so far, and the goal is an official stable release around June.
- Deirdre said they had reserved the crate names on crates.io in advance and plan to publish many of them at the stable stage after the audit, making them more accessible to general developers.
Zcashd 5.5.0 preparation and the Light Client Working Group
Electric Coin Company · Str4d, Pacu, Dodger
- The core of 5.5.0 is the ZIP 317 change; the release candidate has been approved and is waiting on CI. If the regular network tests show no problems, a release is expected early next week.
- The team fixed an SDK view issue that conflicted with the preparatory work for fast sync, reflecting the fact that some information cannot be shown reliably without the full blockchain.
- The remaining dependency is the secp256k1 Rust crate, which will be bumped from 0.21 to 0.26.
- 5.6.0 will contain the changes needed for fast sync and is targeted for about six weeks out. Some of it is already part of the 5.5.0 backend work.
The Light Client Working Group and wallet developer collaboration
Electric Coin Company · Str4d, Pacu, Dodger
- Pacu said the Light Client Working Group meets every other Thursday with Zingo Labs, Nighthawk and others, and the notes from the last 47 meetings are public on GitHub.
- John from the Foundation provides community support and communicates one-on-one with Unstoppable.
- Dodger said the Foundation's ecosystem relations manager Dan can help reach other wallet developers, for example on ZIP 317 adoption.
FROST crate releases and the FROST Book
Zcash Foundation · Deirdre, Dodger, Nate
- FROST core, all cipher suites from the RFC draft, and re-randomized FROST were published on crates.io. Re-randomization is a Zcash variant for redjubjub and redpallas, enabled with a flag in reddsa.
- The latest Rust, stabilized today, exposed a problem with the zeroize trait, so a fix was released as FROST core 0.2 and others, and a build problem with pulling in reddsa was also resolved.
- The IRTF chair passed the specification, bringing it one step closer to an RFC.
- An auditor for FROST was selected. The audit covers the Python/Sage reference implementation and the Rust implementation. The security proof for re-randomized FROST is also being completed, incorporating the adjustment that re-randomizes each signature so they cannot be linked on-chain.
- The FROST Book was published. It covers deployment guides and tips, such as key generation with the trusted dealer and DKG approaches. Nate proposed a presentation on FROST's overview and uses, and Dodger said a demo is in progress.
Technical collaboration forum (Zcash R&D category)
Nate, Dodger, Str4d, Teor, Dan
- Nate proposed using the Zcash R&D category on the Zcash forum as a space for long-form asynchronous technical collaboration. For transparency and web-based collaboration, reading is public and only invited people can write.
- The aim is to separate it from general forum discussions such as whether Zcash should do X, and have the R&D category focus on concrete technical discussion of what can be done and how.
- Str4d saw the advantage that, unlike GitHub issues, it is not tied to a repository or organization and makes it easier for people used to discourse-style venues like ethresear.ch to take part. They were concerned about the cost of curation.
- Currently Adi, Daira, Deirdre, Dan, Dodger, Nathan, Kris, Hudson, Str4d, Teor and others have access. Permissions are divided into See, Reply and Create.
- Dodger said that Dodger and Dan are moderators, so cleanup after the fact such as moving posts is easy, and it does not need to be perfect from the start. It was decided to allow everyone to read, applied immediately, and its usefulness will be reviewed at a meeting in a few months.
- Str4d suggested adding a PoS subcategory at the start. It was decided to allow everyone to read and this was applied right away, with its usefulness to be reviewed at a meeting in a few months.
NU5 retrospective plan
Dodger, Nate
- The NU5 retrospective has been postponed for a long time. A successor needs to be named for Steven, who had been in charge, and Nate said they would do it themselves or have the engineering team decide.
- They said that since ECC is focused on the SDK update in early June, after June is appropriate for the retrospective. It would come 13 months after activation, and they also mentioned that in future, holding the retrospective right after activation is an improvement to make.
NU6 and revisiting the Network Upgrade Process (NUP)
Nate, Dodger, Deirdre, Teor, Str4d
- The NUP was designed when ECC was the only node implementation, which made it hard for other organizations to contribute and made ECC a bottleneck. There are now multiple organizations such as Zebra and QEDIT, so it needs an overhaul, and Nate informed several organizations of this through a private forum thread.
- Nate said they separate execution logistics from governance and wants a process in which no single organization is a single point of failure. Dodger said there is a need to learn from Ethereum's multi-client experience.
- Dodger said to settle a process for NU6 after the NU5 retrospective and iterate on it, and Deirdre said that with several implementations and a more complex protocol, the NUP needs a full overhaul.
- Teor said that although ZSA is conceptually one thing, in practice it is at least three or four ZIPs (protocol, fees, transaction format), so its complexity budget is large and old features to retire alongside it should be considered.
NU6 - testnets, complexity and upgrade cadence
Nate, Dodger, Deirdre, Teor, Str4d
- Nate said teams like QEDIT should first stand up independent testnets, and that upgrade bundles and their order should be set using a feature dependency graph. They also said work like Shielded Labs' sustainability fund should not be blocked by the upgrade process.
- Dodger said to consider smaller, more frequent upgrades.
- Str4d pointed out that the frequent transaction format changes in the four months between Overwinter and Sapling drew complaints, and that the support policy (16 weeks for zcashd) limits the cadence.
- The agreement was to proceed after the NU5 retrospective and the June SDK.
This summary is based on the original meeting notes; check them for exact details.