🛠 PDK Dev Call Notes

[0xBEEFCAF3]

🛠 PDK Dev Call — May 19 Updates

The following are updates from Payjoin contributors. Contributors describes what they are currently working on and what is blocking their progress.

Armin

• Rebased the Session event log sketch branch against master.
• Aiming to get a stable branch of session event log (SEL) to spacebear and Artur as they are partially blocked on their intergrations and need this functionality.
• Improved visibility of persister functions -- only the functions required to be implemented are now exposed.
• Fixed test failures and removed boilerplate in both Rust and FFI layers.
• Unblocked. ready to start migrating persister code once Yuval lands the idiomatic typestate PR.

Ben

• Planning a branch to remove danger_eliminate_https flag. V2 logic is in place, but V1 handling needs clarification.

• Reviewing Shinghim’s PR to get the sender test suite working.

• Raised a question about short_id: is it being removed by SEL? Was originally used to capture errors(?).

  • Nothingmuch suggested making it pub(crate) instead of removing it entirely.

• Aiming to make progress on the FFI PSBT PR — needs to upstream some custom changes into bitcoin-ffi. Getting feedback from bitcoin-ffi maintainers and Dan. No blockers on this. Note: currently payjoin-ffi refrences Ben's branch.

Dan

• BIP77 received significant review over the weekend. Working on addressing comments and updating the spec.
• Focusing on fundraising and legal setup for the foundation -- main priority unless integration teams are blocked.
• Assisting Brandon with bringing payjoin-cli up to spec.
• Reviewing the payjoin-ffi PR.

Spacebear

• No blockers for payjoin-ffi — outstanding PR descriptions question have been resolved. PR is close to mergable.

• Progress on Boltz integration is blocked by a bad dev environment. An issue has been opened in their repo, and the Boltz team is assisting.

  • Raised the question: should we pause Boltz until Session Event Log (SEL) is ready?

• Traveling next week for Bitcoin Conference. Speaking on the "Privacy Without CoinJoin" panel (Tue–Thu).

Artur

• Blocked on Session Event Log for Liana integration.
• contribute_inputs and try_preserving_privacy now working, but Liana GUI flags PSBTs as invalid (fails sanity checks). Need to write this issue up more clearly before others can assist.
• Working on a stable final branch for Liana devs to start providing technical feedback.
• Started submitting small PRs to relax some of the strict GUI sanity checks (e.g., foreign inputs).

Nothingmuch

• Juggling multiple small tasks and cleanup PRs (e.g., padding in spec and code).

• Backporting parts of the persistence sketch branch to master. Three seperate PRs: receiver, sender, and multiparty flows.

• Concerned about potential conflicts with the FFI PRs.

• Review backlog piling up.

• Needs to implement Tor-based OHTTP bootstrap for Leo(?).

• Adding docstrings to all SEL-sketch branch PRs. This serves as a preliminary review.

• Self-blocked, mainly due to bandwidth.

• In the Bitcoin OHTTP crate: identified unnecessary mut borrowing of context for public keys.

  • Tried removing it — tests failed. Added it back — still failed.
  • The commit that introduced it lacks explanation.
  • Soft requirement for SEL; low priority since it's non-blocking and tests are already failing.

Yashraj

• Preparing for a June 4th meeting focused on UX issues around BIP78
• Doing a walkthrough of the Bull Bitcoin UX to identify pain points.

[nothingmuch]

• Needs to implement Tor-based OHTTP bootstrap for Leo(?).

Not implement, describe. I need to write a discussion topic.

For context, OHTTP config bootstrap (for the receiver at least) normally requires TLS. Using the OHTTP relay does too, affecting both receiver and sender.

Since Bitcoin Core already can use tor and connect to hidden services, if the directory and the relay are also available via hidden service then Tor can provide the encryption and authentication for this bootstrap.

[yashrajd]

Yashraj

Preparing for a June 4th meeting focused on UX issues around BIP78

• preparing for a June 4 meeting focussed on UX of using payjoin to refresh timelocks in Liana

Yashraj

Doing a walkthrough of the Bull Bitcoin UX to identify pain points.

This is not planned, it was suggested and is an open topic for the future.

Please add:

• Looking for feedback on his payjoin UX challenges doc

[0xBEEFCAF3]

PDK Dev Call — May 26 Updates

Armin

• Spent the end of last week cleaning up docs, writing unit tests and opening a PR for session persister trait and state transition objects.
• Got review from Dan, addressed the last few comments, and asked for a final re-review.
• This week’s focus: get the rest of the SEL-related PRs up. Waiting on Yuval’s idiomatic type-state PR to land so the transition objects can be wired in cleanly and FFI bindings can be updated.
• The order should be: Session persiter Trait Add SessionPersister trait #716 , idomatic type state refactor, Refactor type states s.t they return a state transition object, and FFI bindings. ** The first three Prs are blocking boltz and liana integrations**.

Ben

• Building out secondary prod monitoring. Wrote an end-to-end bash script to ensure session directory + replay are working.
• Removed _danger_local_https, which was only used in tests Adjust _danger-local-https to be the default to prevent dangerous cert vulnerability when building with --all-features #705 . Now handling that logic directly in test functions.
• Splitting up the second mutants PR. Starting with v2 code, then v1, then directory code. Eventually the whole mutants path will be deleted.
• Flagged that this whole strategy should be documented in an issue.
• Asked how best to build a PSBT for sender mutation testing, looking for suggestions. Answer: probably best to create bitcoin::Transactions and convert them into PSBTs.

Dan

• Spent most of the week on non-code tasks: locking down Bitcoin treasury setup, drafting foundation bylaws, and will be contact with a lawyer. Setting up a project timeline.
• Reviewed the session persister PR and will continue to review SEL related PRs as they are opened.
• Big win: Cake Wallet integration landed and got a great community response.
• Pairing with Brandon to refactor the Payjoin CLI to use clap_derive.
• BIP77 is ready to merge. Just needs final coordination with nothingmuch on a couple items.

Spacebear

• At the Bitcoin conference this week.
• Payjoin FFI PR Import payjoin-ffi #701 is now merged 🔥
• Good news: the boltz team fixed their Docker-compose Bitcoin core credential issue. Can now run master branch without the hackathon MacOS workarounds.
• Improving error transparency in payjoin-flutter. Right now FFI errors aren’t helpful. working on a stopgap so devs can either display error messages or introspect into them.
• No blockers, but someone could take over payjoin-flutter error handling if they’re interested.
• This is seemingly unrelated to the separate bug Brahm and Constantine are running into.
• Before heading to Vegas, will contact them and see if this fix helps with their issue.

Artur

• Got onboarded onto the SEL public interfaces. Has been spending the last couple days integrating SEL into Liana.
• Got try_preserving_privacy input selection working!

Zealsham

• Just finished the BOSS program.
• Reviewing PRs and getting familiar with the codebase.
• Taking notes, learning the payjoin system, and may pick up some tech debt issues from the board.

[DanGould]

Bitcoin OHTTP crate: identified unnecessary mut borrowing of context for public keys.

-- @nothingmuch

Might this be because the public key needs to ratchet for each request? If this is indeed a problem, I suppose we should move this conversation to an upstream issue.