PDK Dev Call Updates – July 28th 2025

[arminsabouri]

PDK Dev Call Updates – July 28th 2025

Armin

• Coordinated with Liana developers. They will begin payjoin groundwork after their 0.12 release. Specifically they’ll look into enabling foreign inputs.
• Met with Bitcoin Design and Liana teams to review the current Payjoin UX. Yash from Bitcoin Design will create wireframes for further development. These wireframes will be published on the bitcoind design community gh repo.
• Completed the refactor for deriving payjoin status + Misc. code clean up with our Liana integration
• Reviewed the entire codebase implementation to determine next steps Payjoin v4 liana#3

Ben

• Merged the FnMut closure PR and refactored relevant sections in Liana Use FnMut closures for receiver typestate checks #883
• Enabled logic for Liana receiver typestates; they are now integrated and active Save psbt with fees applied liana#8
• Will document missed error-handling cases in Liana integration
• Updated the cargo-mutants package, which introduced new mutation patterns.
• Identified many trivial mutations. Considering whether to exclude some non trivial mutations.
• Currently, developers discover new mutations via a weekly job.
• A proposed CI job which could use diff-mutations for this would significantly slow down builds and most likely would not be considered.
• Discussed long-term reliance on mutation exclusions. Some exclusions are temporary, like the ones for process_response, which uses system time internally rather than as a parameter.
• Nothingmuch proposed introducing a trait for system dependencies (e.g., OS RNG, system time) to allow mocking in tests.
• Highlighted another class of exclusions: arithmetic with 0's. For example, a fee rate of zero causes mutations in all calculations. Are these currently being excluded (?).

Chavic

• Finished final touches on the test fixtures. Almost ready to merge.
• Needs to update the caching mechanism based on Spacebear’s suggestion.
• Plans to sync with Spacebear to review the priority list and get next steps.
• Noted that review is constrained due to the PR’s size. Strategy is to review at a high level and validate fixtures incrementally during test activation.
• Noted that the large PR just introduces the fixtures. The tests using them are mostly disabled.
• Raised the question of how to measure readiness for release of this dart-uniffi library?

Dan

• Posted claimed areas of responsibility for library maintainers Claimed Areas of Accountability #898
• The mission: Do more payjoins. Pushing for more payjoin integrations.
• Audited the sender module. Initially written as a hierarchy, but now requires flattening and cleanup Abstract common sender functionality, not heirarchy #884 .
• Refactor sender to validate PjParam by version #901
• Spent a day at MIT DCI with Armin to explore collaboration opportunities.
• Reported continued progress on the foundation. Organizational paperwork is advancing.
• Removed ffi super-module for simplification and maintainability Remove mod/uni payjoin-ffi distinction #899

Nothingmuch

• Continuing the directory refactor.
• Priority shifted: ACME work now runs on its own server.
• Redis refactor is now top priority since it blocks Dart-related CI tests.
• Performed trivial cleanup in lib.rs. This caused conflicts with other PRs, such as the Prometheus integration Introduce metrics Collection via prometheus for payjoin directory #848
• Plans to merge the in-memory DB first to unblock Dart efforts.
• Will document the resulting follow-up issues after the merge.

Spacebear

• Upgraded Bull Bitcoin Mobile to use UniFFI bindings from version 0.24. Migration was slow but completed without UniFFI issues.
• Migrated from a custom persistence system to the standardized SEL approach in 0.24. The BB Mobile architecture is domain-driven, so payjoin-cli patterns don’t apply directly.
• Identified missing UniFFI exports in the payjoin-ffi proc macros. Will follow up with a PR.
• Nothingmuch proposed using return-value types instead of callbacks. This enables a fully functional state machine that never blocks on I/O and runs in any context.
• Suggested scheduling a call to discuss this design direction.
• Highlighted a larger challenge: BB Mobile already uses bespoke persitance strategy. Migrating to the 0.24 model is complex.
• This week: continuing BB Mobile integration and preparing talk for Riga.

Blu

• Merged PR adding feature templates for “Feature Request,” “Good First Issue,” and “General.”
• Paired with Spacebear to start drafting the companion article for the BTC++ talk: Payjoin: Scaling is Privacy.
• Continued work on the article.
• Continued Python Nix flake edits and plans to deliver an update soon.

Zele

• Opened PR Replaces 'flock' with mkdir locking #886 to replace flock with mkdir.
• Opened follow-up PR Add macOS and Windows to CI test matrix #888 to add macOS and Windows to the test matrix.
• Active discussion ongoing in the PR thread.

Shehu

• Addressed change requests from Nothingmuch on the payjoin-cli Sled-to-SQLite migration. Ready for another round of review Migrate payjoin-cli from sled to rusqlite #873
• Planning to pick up issue std::time::SystemTime has nanosecond precision but Payjoin only sees u32 UnixTime second precision #893.
• Started implementing resume logic for max fee per session payjoin-cli resume should make max feerate sticky per session #897. But we dont have acceptance criteria yet.