You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
[Due for payment 2026-06-19] Expense disappears from the report view before the thread screen is closed #92036
Version Number: 9.3.89-1 Reproducible in staging?: Yes Reproducible in production?: No If this was caught during regression testing, add the test name, ID and link from BrowserStack:#88538 Email or phone of affected tester (no customers): N/A Issue reported by: Applause Internal Team Bug source: Pull Request QA execution Device used: MacBook Air 26.5 Chrome App Component: Money Requests
Action Performed:
Create a new expense report with two expenses
Navigate to the Spend/ Reports and open the report
Click on one of the expenses to enter its transaction thread
Inside the transaction thread, tap More → Delete
Observe the report view while the transaction thread screen closes
Create a new expense report with two expenses
Navigate to the Spend/ Reports and open the report
Tap the More button - Delete
Observe the screen while the report is closing
Expected Result:
Step 5 - The transaction thread screen closes first, and only then the transaction entry disappears from the report view (the entry should not vanish while the transaction thread is still animating closed)
Step 9 - The report entry in Spend is removed only after the report screen has fully closed (the entry should remain visible during the closing transition and disappear afterward)
Actual Result:
Step 5 -The transaction entry disappears while the thread is animating closed
Step 9 - The report disappears from the screen while the report view is closing
👋 Friendly reminder that deploy blockers are time-sensitive ⏱ issues! Check out the open `StagingDeployCash` deploy checklist to see the list of PRs included in this release, then work quickly to do one of the following:
Identify the pull request that introduced this issue and revert it.
This is a regression from #88538 (merged 2026-05-28), which migrated InteractionManager.runAfterInteractions cleanup patterns to Navigation.goBack({afterTransition: ...}) / TransitionTracker.runAfterTransitions. Two distinct bugs explain the two reported symptoms.
Step 5 — transaction-thread delete drops the row early (high confidence):
The DELETE handler in useExpenseActions was only half-migrated. The single-transaction branch still calls Navigation.goBack(...) synchronously and then wraps the Onyx mutation in the old InteractionManager.runAfterInteractions(...), so the close animation and the optimistic delete race (useExpenseActions.ts:521-538). The sibling whole-report branch in the same file was migrated correctly to Navigation.goBack(backToRoute, {afterTransition: () => deleteAppReport(...)}) (useExpenseActions.ts:557-572), so the new pattern is clearly what was intended. A telling smell: the file carries an // eslint-disable-next-line no-restricted-imports for InteractionManager (useExpenseActions.ts:5-6) — the lint rule installed by the migration is flagging this exact line.
Step 9 — whole-report delete drops the row early (medium confidence):
That branch already uses the new afterTransition pattern, so the bug is in TransitionTracker itself. Navigation.goBack schedules the callback via runAfterTransitions({waitForUpcomingTransition: true}) (Navigation.ts:511-538). waitForUpcomingTransition races React Navigation's transitionStart against a 1 s safety timer (MAX_TRANSITION_START_WAIT_MS); if no transitionStart event arrives, the callback fires immediately at the 1 s mark (TransitionTracker.ts:116-146). On the Search → MoneyRequestReportView pop in the reported scenario, the popping screen does not appear to emit transitionStart, so the fallback timer fires deleteAppReport while the close animation is still running.
What changes do you think we should make in order to solve the problem?
In the single-transaction delete branch of useExpenseActions, replace the synchronous navigateOnDeleteExpense(goBackRoute) + InteractionManager.runAfterInteractions(() => deleteTransactions(...)) pair with a single Navigation.goBack(goBackRoute, {afterTransition: () => deleteTransactions(...)}), mirroring the whole-report branch a few lines below.
Remove the now-unused InteractionManager import (the file's eslint-disable comment can come off too).
Investigate why the Search → MoneyRequestReportView pop does not produce a transitionStart event. Likely candidates: the popping screen unmounts before the listeners in ScreenLayout fire, or it is rendered outside the PlatformStackNavigation wrapper that attaches the listeners. Once identified, ensure the listeners are attached on the right screen so the 1 s fallback is not relied on for the happy path. As a defensive fallback, the MAX_TRANSITION_START_WAIT_MS value could be reduced or the strategy could be replaced with a deferred requestAnimationFrame chain when no transition is in flight — but the real fix is making the screen emit transition events.
What alternative solutions did you explore?
Revert PR [BATCH 2] Migrate Navigation usage, part 1 #88538 — guaranteed to unblock the deploy, but loses the migration's wins and reintroduces the original regressions it set out to fix. Recommended only if the per-screen fix in (3) cannot be landed in time.
Wrap the deletion in setTimeout(..., 350) to match the navigation animation duration — works but is brittle and is exactly the kind of pattern PR [BATCH 2] Migrate Navigation usage, part 1 #88538 was created to remove.
Related issues
The PR's own test plan (Test 2 and Test 3 on #88538) describes the exact two scenarios that Applause is now flagging — suggesting these passed in the author's environment but regressed in QA, consistent with a transitionStart event mismatch by platform/layout.
Next Steps for Contributor+ team: To accept:@MelvinBot implement [this](https://github.com/Expensify/App/issues/92036#issuecomment-PLACEHOLDER) to create a draft PR. To refine:@MelvinBot <your feedback> To reject: Explain why you are rejecting Melvin's proposal.
After investigating, I discovered that the goBack animation for RHPs (both wide and Super Wide) is now much faster. I tested these two flows using my reverted PR and an AdHoc build. The transaction/report does get deleted after the animation ends, exactly as intended. However, due to the increased speed of the transition, it creates a visual illusion that the deletion happens before the transition finishes. I've attached two videos demonstrating that the transaction correctly waits for the transition to end:
2.mov
Also, when the transition starts, an unrelated bug occurs (the overlay's opacity doesn't animate with the transition). This also contributes to the illusion that the transaction deletes before transitionEnd. As shown in the video, this is not the case, the transaction properly deletes after the transition:
Skipping the payment summary for this issue since all the assignees are employees or vendors. If this is incorrect, please manually add the payment summary SO.
However, after further investigation, the actual root cause was introduced in PR #90212, and it was subsequently fixed in PR #93020 (which I reviewed).
Given that the regression was not caused by the PR I reviewed and that I reviewed the fix PR, I believe I should still be eligible for the standard C+ payment of $250 here.
[Contributor] The offending PR and associated issue have been commented on, pointing out the bug it caused and why, so the author and reviewers can learn from the mistake.
Link to the comment on the PR: N/A
[Contributor] If the regression was CRITICAL (e.g. interrupts a core flow) A discussion in #expensify-open-source has been started about whether any other steps should be taken (e.g. updating the PR review checklist) in order to catch this type of bug sooner.
Link to discussion:
[Contributor] If it was decided to create a regression test for the bug, please propose the regression test steps using the template below to ensure the same bug will not reach production again.
[BugZero Assignee] Create a GH issue for creating/updating the regression test once above steps have been agreed upon.
Navigate to the Spend/ Reports and open the report
Click on one of the expenses to enter its transaction thread
Inside the transaction thread, tap More → Delete
Verify that: The transaction thread screen closes first, and only then the transaction entry disappears from the report view (the entry should not vanish while the transaction thread is still animating closed)
Create a new expense report with two expenses
Navigate to the Spend/ Reports and open the report
Tap the More button - Delete
Verify that: The report entry in Spend is removed only after the report screen has fully closed (the entry should remain visible during the closing transition and disappear afterward)
Activity
applause-bot commentedon May 29, 2026
While Applause doesn't embed attachments for this bug, here are the links for quick view:
https://platform.applause.com/services/links/v1/external/59025b2bf075abfd3ade0cdccbda39bcc036cbb58b8bf3b436e057cf88390e1b
melvin-bot commentedon May 29, 2026
You have been assigned to this deploy blocker because you recently merged this PR: #88538
melvin-bot commentedon May 29, 2026
💬 A slack conversation has been started in #expensify-open-source
github-actions commentedon May 29, 2026
👋 Friendly reminder that deploy blockers are time-sensitive ⏱ issues! Check out the open `StagingDeployCash` deploy checklist to see the list of PRs included in this release, then work quickly to do one of the following:
MelvinBot commentedon May 29, 2026
Proposal
What is the root cause of that problem?
This is a regression from #88538 (merged 2026-05-28), which migrated
InteractionManager.runAfterInteractionscleanup patterns toNavigation.goBack({afterTransition: ...})/TransitionTracker.runAfterTransitions. Two distinct bugs explain the two reported symptoms.Step 5 — transaction-thread delete drops the row early (high confidence):
The
DELETEhandler inuseExpenseActionswas only half-migrated. The single-transaction branch still callsNavigation.goBack(...)synchronously and then wraps the Onyx mutation in the oldInteractionManager.runAfterInteractions(...), so the close animation and the optimistic delete race (useExpenseActions.ts:521-538). The sibling whole-report branch in the same file was migrated correctly toNavigation.goBack(backToRoute, {afterTransition: () => deleteAppReport(...)})(useExpenseActions.ts:557-572), so the new pattern is clearly what was intended. A telling smell: the file carries an// eslint-disable-next-line no-restricted-importsforInteractionManager(useExpenseActions.ts:5-6) — the lint rule installed by the migration is flagging this exact line.Step 9 — whole-report delete drops the row early (medium confidence):
That branch already uses the new
afterTransitionpattern, so the bug is inTransitionTrackeritself.Navigation.goBackschedules the callback viarunAfterTransitions({waitForUpcomingTransition: true})(Navigation.ts:511-538).waitForUpcomingTransitionraces React Navigation'stransitionStartagainst a 1 s safety timer (MAX_TRANSITION_START_WAIT_MS); if notransitionStartevent arrives, the callback fires immediately at the 1 s mark (TransitionTracker.ts:116-146). On the Search →MoneyRequestReportViewpop in the reported scenario, the popping screen does not appear to emittransitionStart, so the fallback timer firesdeleteAppReportwhile the close animation is still running.Full code paths and evidence
useExpenseActions.ts:483-540navigateOnDeleteExpense→Navigation.goBack(backToRoute)with no options (ReportUtils.ts:6247-6261)deleteTransactions→deleteMoneyRequestsetspendingAction: DELETE(DeleteMoneyRequest.ts:693-704)MoneyReportHeaderSecondaryActions.tsx:293PlatformStackNavigation/ScreenLayout.tsx:24-34startTransition/endTransition+ 1 s safety timer:TransitionTracker.ts:67-104TransitionTracker.ts:148-151What changes do you think we should make in order to solve the problem?
useExpenseActions, replace the synchronousnavigateOnDeleteExpense(goBackRoute)+InteractionManager.runAfterInteractions(() => deleteTransactions(...))pair with a singleNavigation.goBack(goBackRoute, {afterTransition: () => deleteTransactions(...)}), mirroring the whole-report branch a few lines below.InteractionManagerimport (the file'seslint-disablecomment can come off too).MoneyRequestReportViewpop does not produce atransitionStartevent. Likely candidates: the popping screen unmounts before the listeners inScreenLayoutfire, or it is rendered outside thePlatformStackNavigationwrapper that attaches the listeners. Once identified, ensure the listeners are attached on the right screen so the 1 s fallback is not relied on for the happy path. As a defensive fallback, theMAX_TRANSITION_START_WAIT_MSvalue could be reduced or the strategy could be replaced with a deferredrequestAnimationFramechain when no transition is in flight — but the real fix is making the screen emit transition events.What alternative solutions did you explore?
setTimeout(..., 350)to match the navigation animation duration — works but is brittle and is exactly the kind of pattern PR [BATCH 2] Migrate Navigation usage, part 1 #88538 was created to remove.Related issues
transitionStartevent mismatch by platform/layout.Next Steps for Contributor+ team:
To accept:
@MelvinBot implement [this](https://github.com/Expensify/App/issues/92036#issuecomment-PLACEHOLDER)to create a draft PR.To refine:
@MelvinBot <your feedback>To reject: Explain why you are rejecting Melvin's proposal.
collectioneur commentedon May 29, 2026
After investigating, I discovered that the
goBackanimation for RHPs (both wide and Super Wide) is now much faster. I tested these two flows using my reverted PR and an AdHoc build. The transaction/report does get deleted after the animation ends, exactly as intended. However, due to the increased speed of the transition, it creates a visual illusion that the deletion happens before the transition finishes. I've attached two videos demonstrating that the transaction correctly waits for the transition to end:2.mov
Also, when the transition starts, an unrelated bug occurs (the overlay's opacity doesn't animate with the transition). This also contributes to the illusion that the transaction deletes before
transitionEnd. As shown in the video, this is not the case, the transaction properly deletes after the transition:1.mov
luacmartins commentedon May 29, 2026
@collectioneur maybe we just need to make the transition animation a bit slower? I feels kinda jerky now that it's too fast 😂
luacmartins commentedon May 29, 2026
I don't think this is a blocker, but let's address this
32 remaining items
melvin-bot commentedon Jun 18, 2026
Skipping the payment summary for this issue since all the assignees are employees or vendors. If this is incorrect, please manually add the payment summary SO.
melvin-bot commentedon Jun 19, 2026
@collectioneur Whoops! This issue is 2 days overdue. Let's get this updated quick!
mallenexpensify commentedon Jun 20, 2026
@daledah , or any other C+, due payment here? If so, how much? Plz link to PRs and address any regressions. Thx
daledah commentedon Jun 23, 2026
@mallenexpensify Initially, this issue was considered a regression from PR #88538.
However, after further investigation, the actual root cause was introduced in PR #90212, and it was subsequently fixed in PR #93020 (which I reviewed).
Given that the regression was not caused by the PR I reviewed and that I reviewed the fix PR, I believe I should still be eligible for the standard C+ payment of $250 here.
mallenexpensify commentedon Jun 23, 2026
Payment Summary
Contributor+: @daledah due $250 via NewDot
@daledah plz complete the BZ checklist and tag me in a post once you have. Thx
daledah commentedon Jun 24, 2026
BugZero Checklist:
[Contributor] The offending PR and associated issue have been commented on, pointing out the bug it caused and why, so the author and reviewers can learn from the mistake.
Link to the comment on the PR: N/A
[Contributor] If the regression was CRITICAL (e.g. interrupts a core flow) A discussion in #expensify-open-source has been started about whether any other steps should be taken (e.g. updating the PR review checklist) in order to catch this type of bug sooner.
Link to discussion:
[Contributor] If it was decided to create a regression test for the bug, please propose the regression test steps using the template below to ensure the same bug will not reach production again.
[BugZero Assignee] Create a GH issue for creating/updating the regression test once above steps have been agreed upon.
Link to issue: https://github.com/Expensify/Expensify/issues/652659
Regression Test Proposal
Test:
Do we agree 👍 or 👎
Zapier Logs
Run ID: 00040eee-ff6a-a840-bbbe-847fe7e07dd4daledah commentedon Jun 24, 2026
@mallenexpensify i added the BZ checklist
quinthar commentedon Jun 29, 2026
Approved $250 to @daledah