FROST 1.0 release and TEX addresses
The stable FROST 1.0.0 was released, and the Binance response was settled on a TEX address approach.
Agenda summary 5
ECC update: mobile SDK and ZIP-320 (TEX addresses)
Electric Coin Company · Kris, Daira, Jack Grigg
- The current focus is the wallet SDK release. A zcash_client_backend release with Orchard support is being prepared, and the goal is to connect it to zcash_client_sqlite and the SDK to ship Zashi 1.0 in late March.
- Security audits are underway from the iOS layer through the wallet SDK and the Rust libraries. The iOS release is expected before Android, and the schedule will change if the audit finds a serious problem.
- The direction of the Binance transparent-source exchange address work was settled through Binance feedback together with Jason McGee, Hanh, Pacu and others. ZIP-320, which defines it, is to be finalized using the TEX-encoded address approach.
- Longer term, a separate ZIP is also being considered that would use address metadata to express restrictions such as allowing only shielded sources. A community grant was also proposed for Java, Swift and TypeScript implementations like the zcash_address crate.
Zebra update: scanner gRPC, regtest, hackathon
Zcash Foundation · Alfredo, Jack Gavigan
- The Zebra scanner's gRPC interface was organized into 6 methods and documented. The core is the scan push interface, which provides registration, scan start and an asynchronous result stream.
- There is no Orchard scanning yet and bugs remain, so the team asked the community for testing and feedback. Zebra 1.6.0, worked on by Marek, will be released soon.
- QEDIT and Zingo! each proposed adding a regtest mode to Zebra, and the team is investigating which to choose. For the first time, the ECC and Zebra teams held a meeting to discuss the features Zebra lacks to replace zcashd.
- The Zebra team will hold a two-week hackathon sprint. Ideas include restoring Tor support (Arti integration), running Zebra with a Tauri-based GUI, and restoring Windows support.
Trailing Finality status
Electric Coin Company · Daira
- There is no new update. The team plans to document the Crosslink 2 changes.
- How to fit PoS into ECC's schedule has not yet been decided.
FROST 1.0 stable release and vulnerability disclosure
Zcash Foundation · Conrado, Daira, Jack Grigg
- The stable FROST crate 1.0.0 was released. Compared with the release candidate, it has a few methods exposed based on user feedback and a vulnerability fix. Trail of Bits reported a vulnerability in the distributed key generation (DKG) protocol in January.
- It was a denial-of-service vulnerability: because the length of the polynomial coefficient commitment was not checked, a malicious participant could create a larger polynomial, raise the threshold and make the wallet unusable.
- It was not key theft, and it was fixed by checking the length and included in 1.0.0. No deployed application is known to be affected, but crate users were advised to update. It was pointed out that DKG has no specification and the RFC draft covers only signing.
- An opinion was raised to publish an unofficial standard. Demo stage 1 was merged, so running the server yourself enables a FROST demo over the internet. Stage 2, separating the server and the coordinator, is half done.
Definition of experimental features and use of Rust feature flags
Jack Gavigan, Daira, Jack Grigg, Alfredo, Kris, Andrew Arnott
- With Zebra about to make viewing key scanning an experimental feature, someone asked whether there is an official definition. Daira answered that it means there is no compatibility guarantee and that it may change or be removed without notice.
- Jack Grigg explained that experimental features are conceptually unrelated to Rust feature flags. Feature flags should be additive-only, but large changes can break semver.
- As an alternative, Orchard functionality was placed behind a --cfg zcash_unstable=orchard flag so that compilation fails without it. It is similar to zcashd's experimental flag, but compiling with all features is not possible.
- In response to Andrew Arnott's question, they answered that adding trait methods behind a flag should be done only by providing a default implementation or in crates close to the application layer.
- zcash_client_backend is of an application-crate nature, so it is allowed. Conversely, experimental features in deeply depended-on crates such as zcash_primitives should be kept out of feature flags.
- Zebra's Alfredo said there is no distinction between features hidden behind flags that are permanent and those slated for removal, and proposed clarifying the definition in a readme; having a crate shared by the two teams was also discussed.
This summary is based on the original meeting notes; check them for exact details.