fix: let force-closed channels claim pending payments on-chain

[ovitrif]

Twin: synonymdev/bitkit-ios#929
Refs: synonymdev/ldk-node#125

Description

• Updates LDK Node from 0.7.0-rc.72 to 0.7.0-rc.74, so force-closed channels can claim their pending payments (HTLCs) on-chain again.
• The wallet's PSBT signing failed whenever an input it does not own stayed unsigned, which broke the HTLC claim transactions LDK builds after a force-close. 0.7.0-rc.73 and 0.7.0-rc.74 come from fix: let force-closed channels claim pending payments on-chain ldk-node#125 on top of 0.7.0-rc.72 (release).

Design

N/A — no UI changes.

Preview

N/A — no user-visible changes.

QA Notes

Journeys

• updated local-paykit-fallback.md

Manual Tests

• Upgrade from a build on 0.7.0-rc.72 with an existing wallet and open channels → the node starts, channels reconnect, balances match — needs a wallet written by the previous build.
• Pay a hold invoice from Bitkit, force-close the channel while the HTLC is pending, then mine past its expiry → the payment fails and its amount comes back to Savings through an on-chain claim — the Capabilities table lists no way to keep an HTLC pending across a force-close.

Automated Checks

• com.synonym:ldk-node-android:0.7.0-rc.74 is published to the ldk-node Maven registry.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-13T18:18:24.577510Z	d14b4c2	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[github-actions]

Regtest APK

Built from 8c26687 (run).

Download bitkit-dev-debug universal APK (expires in 30 days).

[greptile-apps]

Confidence Score: 5/5

[High impact] The PR appears safe to merge; no blocking issues were found.

Summary

Updates LDK Node from 0.7.0-rc.72 to 0.7.0-rc.74 to restore on-chain claims for pending payments after a force-close.

• Android channels can claim pending payments on-chain after a force-close.

_{Reviews (2) · Last reviewed commit: "docs: limit the changelog entry to the f..." · Reviewed by Greptile}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: d14b4c22a9

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Add a journey that exercises the HTLC claim

Changing only this version pin in the Paykit fallback does not exercise a force-close or an on-chain HTLC claim, so the listed journey cannot catch a regression in the behavior this dependency bump is meant to fix. The journey environment already provides Lightning channels and regtest block mining, so this scenario should be covered by an added or updated journey rather than only the manual-test list.

AGENTS.md reference: AGENTS.md:L109-L113

Useful? React with 👍 / 👎.

[ovitrif]

The fix is a version pin; the behavior it restores is covered in synonymdev/ldk-node#125 by repro and integration tests that fail on 0.7.0-rc.72. A device journey would need an HTLC kept pending across a force-close and then mined past its expiry: the Capabilities table in journeys/README.md lists LSP force-close but no way to hold an HTLC (the LSP regtest API has no hold invoice), so per AGENTS.md this stays a manual test, now worded with the missing capability.

[ovitrif]

Pushed 8c26687: the changelog fragment now names only the force-close claim fix, since Bitkit does not start splices; the description and manual tests match. Answers the Codex threads on the fragment and the journey.

[talosmachina]

No findings. Bumps ldk-node-android from 0.7.0-rc.72 to 0.7.0-rc.74 for the PSBT signing fix in synonymdev/ldk-node#125, plus the changelog fragment and the journey's version line. Reviewed 8c26687, full tier (dependency manifest).

What I checked, and 2 candidates I ruled out

Pin: tag v0.7.0-rc.74 in synonymdev/ldk-node resolves to 190a7d98b24aa0f4b337b21d9cee6fb8b2986679.
Upstream delta rc.72 to rc.74: 6 commits, d1bd9fc...190a7d9, touching crates/bdk-wallet-aggregate (signing), src/wallet/mod.rs, src/event.rs, version files and tests. No .udl or binding change, so the Kotlin and Swift API the app calls is unchanged.
Other references to rc.72: none left in the tree (grepped the branch head).

Ruled out

• API break in the bump: no UniFFI interface file changed between the two tags.
• sign_owned_inputs signature change: it is pub(crate) inside ldk-node and gains a known_prevouts argument there only; nothing in the bindings exposes it.

CI: build, lint and detekt green; build-local still pending when reviewed. The build job resolved com.synonym:ldk-node-android:0.7.0-rc.74 from the registry.
Not built here: Bitkit Android branches are not built on this reviewer's box; this review is reasoned from the diff, the upstream compare and CI.

Merge confidence: 4/5, a dependency bump whose fixed path (on-chain HTLC claim after a force-close) has no test in this repo and rests on the manual test in the PR body; the rc.74 tag also sits on the head of synonymdev/ldk-node#125, which is still open.

[ovitrif]

Replaced by #1480: same change, on a branch named after the force-close claim fix.