[ovitrif]

Companion:

• feat: move LN channel balances to spark bitkit-android#1774
• feat: move LN channel balances to spark bitkit-ios#1212

Description

The apps now show a one-time Move to Spark intro, at the top timed-sheet priority, once a wallet's Lightning channels can move a balance to a ready Spark wallet. That turns true on state a test cannot see (a receive, Spark connecting), so the intro can cover home at any step of a suite that opens a channel and receives over it.

• Adds dismissMoveToSparkIntroIfShown(), which taps Later, and withMoveToSparkIntroRecovery(), which retries a failed wait once after dismissing the intro, as the ANR recovery does for system dialogs.
• tap, the received-payment wait and the HeaderMenu wait in doTriggerTimedSheet recover through it, and acknowledgeReceivedPayment dismisses the intro if it shows within 3 s, ahead of balance reads.
• The timed-sheet trigger treats the intro as an intervening sheet, as it does the USDT intro: after dismissing it, it lets home surface the next queued sheet instead of triggering over it.
• The probe is MoveToSparkIntroLater, queryable on both platforms.

On synonymdev/bitkit-android#1774, without this branch, lightning_security, lnurl_transfer, multi_address_local and send failed with the intro covering home (@lightning_1, @lnurl_1, @transfer_2, @multi_address_4, @send_2).

QA Notes

• The app PRs' e2e runs check out this branch by name (feat/move-to-spark, the companion default), so the Android e2e run on feat: move LN channel balances to spark bitkit-android#1774 proves it; iOS e2e does not run on feat: move LN channel balances to spark bitkit-ios#1212.
• Proven on feat: move LN channel balances to spark bitkit-android#1774 at 95fe19287: all 7 e2e shards pass against b004e5e, and the helper dismissed the intro in lightning_security, send, multi_address_local and lnurl_transfer.
• tsc --noEmit, eslint and prettier --check pass on the changed files.

[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 2026-11-06T18:45:47.855687Z	c436191	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.

[talosmachina]

No findings. Teaches the timed-sheet helpers to dismiss the new Move to Spark intro with Later, as they already do the USDT intro, and to re-trigger the queue when it surfaces in place of the expected sheet. Reviewed c436191, full tier, static.

What I checked, and 5 candidates I ruled out

Read in full: test/helpers/actions.ts (triggerTimedSheetUnlessPresent, dismissInterveningSheets, dismissMoveToSparkIntroIfShown, waitForWalletHome, restoreWallet, completeOnboarding), test/helpers/navigation.ts (doTriggerTimedSheet), test/specs/transfer.e2e.ts (dismissHomeSheetsIfPresent, @transfer_1, @transfer_2), and the call sites in lightning, lnurl, send and multiaddress.
Cross-repo contract: the ids MoveToSparkIntro and MoveToSparkIntroLater match the sheet in synonymdev/bitkit-android#1774 (MoveToSparkSheet.kt testTags) and synonymdev/bitkit-ios#1212 (MoveToSparkSheet.swift accessibility identifiers). On Android the sheet is MOVE_TO_SPARK(priority = 7), above every other timed sheet, and onShown sets hasSeenMoveToSparkIntro, so the extra doTriggerTimedSheet() after a dismiss cannot bring the intro back.
CI: this repo reports no checks on the branch. The failure the PR targets is real: the bitkit-android#1774 E2E run before this branch existed failed @lightning_1 on TotalBalance-primary right after BackgroundPaymentsIntro-later not displayed. The app PRs' E2E runs that pick this branch up by name were still in progress when I looked.
Build skipped: talos-pr-run could not check the author's permission on this repo and fails closed, so nothing was run here; tsc, eslint and prettier are the author's report.

Ruled out

• triggerTimedSheet: false callers now trigger the queue: dismissInterveningSheets() runs before the early return, but it triggers only after it dismissed an intervening sheet, which is what the USDT branch already did.
• Intro surfacing more than 3 s after the final trigger: the earlier checks wait for home chrome and settle 700 ms around two dismiss passes before the trigger; a later surfacing needs the sheet manager to delay past that, which I could not show from the app code.
• dismissHomeSheetsIfPresent in transfer.e2e.ts does not know the intro: @transfer_1 is skipped, and @transfer_2 goes through dismissBackgroundPaymentsTimedSheet, which now handles it.
• Intro after restoreWallet in @lightning_1: whether it reappears depends on whether the restored settings backup carries hasSeenMoveToSparkIntro, which I could not establish statically; the companion E2E run decides it.
• waitForWalletHome and completeOnboarding do not wait for the intro: a fresh wallet has no channel, and canMove needs at least 1,000 sats sendable over a channel, so it cannot show there.

Merge confidence: 4/5, no findings, but nothing exercised this head yet: the companion E2E runs that do were still in progress.

[ovitrif]

Pushed ec5958c, from the first run on synonymdev/bitkit-android#1774:

• After the Move to Spark intro is dismissed, home re-checks the queue about 2 s later and surfaces the next sheet by itself (QuickPay in @multi_address_4), which covered HeaderMenu before the helper could re-trigger. The helpers now wait for the expected sheet first and trigger only when it does not come.
• The QuickPay and Background Payments dismiss helpers wait briefly for the intro afterwards (@lightning_1, @send_2), since it needs a Lightning balance as they do.

Re-running the failed Android e2e jobs against this head.

[ovitrif]

Pushed e2030ac and 70d01c1, from the second run on synonymdev/bitkit-android#1774:

• The intro surfaced more than 4 s after QuickPay closed (@send_2, @lightning_1), so the QuickPay and Background Payments helpers now wait up to 10 s for it.
• After the intro is dismissed, the next queued sheet can be QuickPay rather than the Background Payments sheet the caller expects (@multi_address_4, @lnurl_1); the trigger helper now stops there, leaving the best-effort Background Payments wait and the following QuickPay helper to handle it.

Re-running the failed Android e2e jobs against 70d01c1.

[ovitrif]

Pushed b004e5e. The third run on synonymdev/bitkit-android#1774 showed the intro becomes due on state a test cannot see: in @lightning_1 it was not due at the trigger right after QuickPay and showed after the next receive. So instead of waiting at fixed points:

• tap, the received-payment wait and the HeaderMenu wait in doTriggerTimedSheet recover from it as they would from the ANR dialog: when the wait fails and the intro is showing, tap Later and retry once.
• acknowledgeReceivedPayment dismisses it if it shows within 3 s, ahead of balance reads.
• The probe is MoveToSparkIntroLater, queryable on both platforms.
• The forced trigger after QuickPay is gone.

Re-running the whole Android e2e workflow against b004e5e.

[ovitrif]

Proven on synonymdev/bitkit-android#1774 at 95fe19287: all 7 e2e shards pass against b004e5e (https://github.com/synonymdev/bitkit-android/actions/runs/42309414918), and the helper dismissed the Move to Spark intro in each suite that meets it (lightning_security, send, multi_address_local, lnurl_transfer).