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
Applied Has filter not cleared when resetting filter from Display modal #107501
Version Number: v9.5.43-3 Reproducible in staging?: Yes Reproducible in production?: N/A - new feature, doesn't exist in prod If this was caught during regression testing, add the test name, ID and link from BrowserStack:#107141 Email or phone of affected tester (no customers):kiatest.78+ck111826@gmail.com Issue reported by: Applause Internal Team Bug source: Exploratory - Significant User Experience Deterioration Device used: Windows 11/Chrome, Pixel 8a/Android 16 App Component: Search
Action Performed:
Prerequisites:
Control WS with Rules feature enabled
Has some expenses with violations (Missing category, Potential Duplicate) and it's submitted/approved
Has some expenses without violations
Steps:
Navigate to staging.new.expensify.com
Go to Spend >> Expenses
Select Display > Group by >> Violation filter >> Apply
Expand and note expenses with violations displayed
Click Display again >> click Reset CTA
Note applied Has filter is not cleared and expenses with violations still displayed
Click the Reset pill next to Has filter
Note all expenses (with and without violations) displayed
Expected Result:
On step 6 - When clicking Reset CTA from Display modal should clear the applied Has filter and all expenses (with and without violations) should be displayed
Actual Result:
On step 6 - When clicking Reset CTA in Display modal, Has filter still persists and expenses with violations are displayed. While the filter get cleared only when clicking the Reset button next to Has filter
👋 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.
The regression was introduced by “[Insights: Violations][R3] Add group-by:violation”. Selecting violation grouping automatically adds submitted/approved violation filters, but Display → Reset removes the grouping without removing those filters. This explains both the remaining Has chip and the violation-only results.
A stale-results or backend-only explanation does not account for the restrictive Has filter remaining in the frontend query.
What changes do you think we should make in order to solve the problem?
Prefer a targeted roll-forward fix: make Display Reset remove the violation restriction introduced by violation grouping, while preserving unrelated filters.
One design decision needs explicit agreement: the current query does not distinguish automatically added violation filters from manually selected ones. Define how pre-existing violation selections and saved/default queries should behave rather than clearing Has indiscriminately.
Regression tests should cover:
The reported grouping → Display Reset sequence restores all otherwise-matching expenses.
Unrelated Has values and other filters survive.
Explicit violation selections and saved/default queries behave as agreed.
Resetting Has still clears its paired violation grouping.
What alternative solutions did you explore? (Optional)
Reverting the feature PR is a safe fallback if reset semantics cannot be resolved and verified promptly. For this issue alone, a targeted fix is narrower.
Verification limit: This conclusion is based on code and release-history evidence; browser reproduction and fix verification were not performed.
Next Steps for Contributor+ team: To accept:@MelvinBot implement with a link to this proposal after clarifying filter-preservation semantics. To refine:@MelvinBot <your feedback> To reject: Explain why you are rejecting Melvin's proposal.
Confirmed, but I don't think it should block the deploy. It's a small UX gap in a new feature. No results are wrong: the Has pill still shows the filter that's applied, and it has its own Reset. I recommend removing DeployBlockerCash and fixing this in a follow-up.
The same thing happens when you switch Group by from Violation to another option, like Category.
Fix: when the grouping changes away from Violation (Display's Reset, or picking another Group by), also remove submitted-violation and approved-violation from Has when both are there, since that pair is what the grouping added. Keep any other Has values. One trade-off: if someone picked both values by hand before grouping, they'd be removed too. I think that's fine. Add cases to useUpdateFilterQuery.test.ts and SearchQueryUtilsTest.ts.
Fix: when the grouping changes away from Violation (Display's Reset, or picking another Group by), also remove submitted-violation and approved-violation from Has when both are there, since that pair is what the grouping added.
Mhm, I dunno about this Neil. I think you should be able to see the expenses with violations grouped by From or whatever if you want.
Selects "Has: submitted-violation"
Chooses "Group-by:violation"
Chooses "Group-by:from"
"Has:submitted-violation is removed." ❌ that doesn't seem right to remove it.
I'm not entirely convinced we need to do anything here, tbh. This doesn't seem bad?
Select "Group-by: Violation"
Has:submitted-violation, approved-violation are applied automatically as they are required for that group-by.
Reset Display -> "Group-by: Violation" is removed because you've reset the display. Maybe you're going to pick another one, or just want the raw list of expenses with has submitted, approved violations.
"Has-submitted-violation, approved-violation" stays applied. The filter reset button removes filters if they want.
Activity
applause-bot commentedon Nov 20, 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/7dddb0f0862d0a866c3547d6b9e74fc63bb8a112d2e41846c39d97f61ce287b9
expensify-chore-bot commentedon Nov 20, 2026
You have been assigned to this deploy blocker because you recently merged this PR: #107141
expensify-chore-bot commentedon Nov 20, 2026
@neil-marcellini you have been assigned to this deploy blocker because you reviewed the offending PR: #107141
expensify-chore-bot commentedon Nov 20, 2026
@nkdengineer you have been assigned to this deploy blocker because you authored the offending PR: #107141
expensify-chore-bot commentedon Nov 20, 2026
💬 A slack conversation has been started in #expensify-open-source
github-actions commentedon Nov 20, 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 Nov 20, 2026
Proposal
What is the root cause of that problem?
The regression was introduced by “[Insights: Violations][R3] Add group-by:violation”. Selecting violation grouping automatically adds submitted/approved violation filters, but Display → Reset removes the grouping without removing those filters. This explains both the remaining Has chip and the violation-only results.
Code and release evidence
useUpdateFilterQuery.tsx:37–46adds the violation-stage Has filters.DisplayPopup.tsx:102–107invokes the display-reset helper.SearchQueryUtils.ts:1469–1489resets display settings while retaining Has filters.SearchQueryUtils.ts:1437–1462.A stale-results or backend-only explanation does not account for the restrictive Has filter remaining in the frontend query.
What changes do you think we should make in order to solve the problem?
Prefer a targeted roll-forward fix: make Display Reset remove the violation restriction introduced by violation grouping, while preserving unrelated filters.
One design decision needs explicit agreement: the current query does not distinguish automatically added violation filters from manually selected ones. Define how pre-existing violation selections and saved/default queries should behave rather than clearing Has indiscriminately.
Regression tests should cover:
What alternative solutions did you explore? (Optional)
Reverting the feature PR is a safe fallback if reset semantics cannot be resolved and verified promptly. For this issue alone, a targeted fix is narrower.
Verification limit: This conclusion is based on code and release-history evidence; browser reproduction and fix verification were not performed.
Next Steps for Contributor+ team:
To accept:
@MelvinBot implementwith a link to this proposal after clarifying filter-preservation semantics.To refine:
@MelvinBot <your feedback>To reject: Explain why you are rejecting Melvin's proposal.
view run
neil-marcellini commentedon Nov 20, 2026
(Neil's AI agent)
Confirmed, but I don't think it should block the deploy. It's a small UX gap in a new feature. No results are wrong: the Has pill still shows the filter that's applied, and it has its own Reset. I recommend removing
DeployBlockerCashand fixing this in a follow-up.Cause: choosing Group by → Violation adds both violation Has values. Display's Reset only resets display settings and keeps every filter, so the grouping is removed but the added
has:submitted-violation,approved-violationstays. The other direction is already handled: resetting the Has filter also removes the violation grouping.The same thing happens when you switch Group by from Violation to another option, like Category.
Fix: when the grouping changes away from Violation (Display's Reset, or picking another Group by), also remove
submitted-violationandapproved-violationfrom Has when both are there, since that pair is what the grouping added. Keep any other Has values. One trade-off: if someone picked both values by hand before grouping, they'd be removed too. I think that's fine. Add cases touseUpdateFilterQuery.test.tsandSearchQueryUtilsTest.ts.@nkdengineer could you take the follow-up?
nkdengineer commentedon Nov 20, 2026
I can take this
trjExpensify commentedon Nov 20, 2026
Mhm, I dunno about this Neil. I think you should be able to see the expenses with violations grouped by
Fromor whatever if you want.I'm not entirely convinced we need to do anything here, tbh. This doesn't seem bad?
neil-marcellini commentedon Nov 20, 2026
Yeah good point. Let's close it.