← All calls
ARBORIST#61

Mobile SDK 2.0 RC3 released

The iOS and Android SDKs reached the RC3 stage, and the official 2.0.0 release is set for Monday.

Agenda summary 7

  1. Zebra update: spend-before-sync done and wallet support plan

    Zcash Foundation · Marek

    • Spend-before-sync support for lightwalletd is effectively done, and merging one more test PR finishes the work.
    • Next, a new project to add wallet functionality to Zebra begins, with the first step being implementation of Sapling viewing key scanning.
    • Kris advised using the scanning code in zcash_client_backend, adapted to Zebra's use, rather than implementing it anew. That code does not yet support Orchard scanning, which is future work.
  2. ECC update: mobile SDK RC3 and Zcashd binary

    Electric Coin Company · Jack Grigg(Str4d), Kris, Daira

    • The iOS and Android wallet SDKs are at the release candidate stage; iOS RC3 is out, and Android RC3 is being deployed to Maven Central. If there are no problems, the official 2.0.0 release is planned for Monday.
    • Daira said this release does not include ZIP-317 fees and uses a fixed fee of 10,000 zatoshi. Daira expects later SDK releases to come out more often in smaller increments.
    • Existing Zcashd nodes are close to expiring, so an update will come within 1-2 weeks. It bundles the Sprout verification parameters and the Sapling proving and verification parameters in the binary, so the parameter files on disk are not needed.
    • Bundling makes the binary larger, but only by about 16MB over the existing release binary. Daira confirmed that the current main build, stripped, is about 73MB.
  3. FROST audit complete and community call plan

    Zcash Foundation · Conrado, Dan

    • The audit findings were fixed and sent back to NCC, and the final audit report saying all were resolved was received. It will be published soon after review. The basic byte serialization needed for transfer between participants was added.
    • The serde encoder postcard is designated as the default, so users do not need to import serde themselves. The refactoring to prepare the 1.0.0 release candidate is being wrapped up, and feedback on the API is being sought.
    • Communication between wallets is the hard part of FROST deployment. A community call will be held on October 5 at 18:00 UTC to gather requirements, such as synchronous and asynchronous usage patterns.
  4. Trailing Finality Layer research and the Snap-and-Chat design

    Electric Coin Company · Nate, Daira

    • The ebb and flow paper uses only a simple BFT analysis assuming fewer than one third of nodes are malicious, so it was judged not sufficient as a basis for the TFL proposal.
    • The paper's log sanitization prefers the earlier snapshot when transactions are duplicated, but if a PoW rollback creates a double-spend transaction, finality and the PoW chain diverge and reasoning about transactions becomes complicated.
    • Daira regarded this as an error in the paper and concluded it can be solved by adding a property called finalisation availability to the ebb and flow model.
    • Daira considered the approach feasible since Ethereum uses a similar scheme in production. In the current design, PoW produces blocks and PoS is responsible only for finality.
    • The plan is to release a version organized in the TFL book by the next Arborist Call, and to write code simulating attack scenarios in 4-6 weeks.
    • How to allocate the security budget between PoW and PoS is not yet known, and the idea that block producers could participate in both protocols was also raised.
  5. ZSA circuit optimization, v6 transactions, issuance signature change

    QEDIT · Vivek, Jon

    • Before optimization, the ZSA circuit exceeded 2^11 (2048) rows, which enlarged the proof-related parameters. With Daira and Str4d, the circuit was reduced to just under 2048 rows. Work is underway to get ZSA running on a Zebra node and connect it to wallets.
    • The functionality of the Zcashd prototype has not yet been fully reproduced in Zebra. A v6 transaction data structure is being created and the changes moved into it. ZSA details have begun to be filled into ZIP-230 (zips PR 687).
    • It was decided to change the issuance authorization signature from redpallas to BIP-340 Schnorr (secp256k1). The reasons are good HSM support and independence from the Orchard parameters, and the plan is to reuse the FROST crate's implementation.
    • For asset swaps, the preliminary ZIP design document is substantially complete. Nate suggested validating the swap product and referring to Penumbra's design, and recommended verifying usability with a wallet prototype.
  6. Zebra wallet support direction and ZIP editors update

    Nate, Jack Grigg, Daira, Pili

    • Nate pointed out that a choice is needed on whether Zebra wallets will develop their own approach or pool effort on the Zcash wallet backend and lightwalletd stack. Jack Grigg said there are options in between.
    • From a security standpoint, Daira said it would be good to have at least one full-node wallet implementation that does not depend on the light client security model. The Foundation said wallet work is at an early stage, nothing is decided, and it will start with small features.
    • The ZIPs under review include the Zenate proposal, ZSA, issuance curve smoothing, ZSF and the v6 transaction format. The ZIP editors are Daira, Str4d, Conrado, Adi and Teor.
    • Android wallet SDK RC3 fixed a problem where chain continuity errors were misinterpreted in the FFI, found by Nighthawk.
  7. Network upgrade process discussion

    Adi, Nate, Daira, Dan

    • Adi asked whether there is an upgrade timeline linking in-development ZIPs such as FROST and ZSA to the upcoming halving.
    • Nate answered that there is no fixed timeline, and judged that the network upgrade pipeline v2 led by ECC did not work well. Nate proposed settling the process first and then the schedule.
    • Daira said the old process was too fixated on a fixed timeline, and that safe implementation and design and community trust are key. Daira said constraints such as a testnet period before mainnet activation could still be set.
    • It was decided to add a network upgrade process discussion slot starting with the next Arborist Call, and Nate asked participants to bring their own proposals.

This summary is based on the original meeting notes; check them for exact details.