[$250] "Uh-oh, something went wrong!" when trying to access New Expensify or the dedicated chat with Concierge

[m-natarajan]

If you haven’t already, check out our contributing guidelines for onboarding and email contributors@expensify.com to request to join our Slack channel!

---

Version Number: v9.3.27-8
Reproducible in staging?: Needs Reproduction (Unable to reproduce)
Reproducible in production?: Needs Reproduction
If this was caught during regression testing, add the test name, ID and link from BrowserStack:
Email or phone of affected tester (no customers):
Logs: https://stackoverflow.com/c/expensify/questions/4856
Expensify/Expensify Issue URL:
Issue reported by: @abbagail
Slack conversation (hyperlinked to channel name): #Expensify Bugs

Action Performed:

1. Login In Classic, click "support."
2. Select "Concierge"
3. Get redirected to New Expensify

Expected Result:

New Expensify Load without issues.

Actual Result:

Page takes a few seconds to load, but fails to load and gets "Uh-Oh, something went wrong." (App crashes)

Workaround:

Unknown

Platforms:

Select the officially supported platforms where the issue was reproduced:

• Android: App
• Android: mWeb Chrome
• iOS: App
• iOS: mWeb Safari
• iOS: mWeb Chrome
• Windows: Chrome
• MacOS: Chrome / Safari

Screenshots/Videos

The video for the issue is in Slack thread

View all open jobs on GitHub

Issue Owner

Current Issue Owner: @aimane-chnaif

Upwork Automation - Do Not Edit

• Upwork Job URL: https://www.upwork.com/jobs/~022028954940716757609
• Upwork Job ID: 2028954940716757609
• Last Price Increase: 2026-04-14

[melvin-bot]

Triggered auto assignment to @stephanieelliott (Bug), see https://stackoverflow.com/c/expensify/questions/14418 for more details. Please add this bug to a GH project, as outlined in the SO.

[MelvinBot]

Proposal

Please re-state the problem that we are trying to solve in this issue.

When users click "Support" > "Concierge" in Classic Expensify and get redirected to New Expensify, the app crashes after a few seconds of loading, showing the "Uh-Oh, something went wrong!" error page (the GenericErrorPage rendered by the app's top-level ErrorBoundary). This affects web platforms (Windows Chrome, MacOS Chrome/Safari) on version v9.3.27-8.

What is the root cause of that problem?

Note: Without access to the actual Sentry error logs / stack trace referenced in the issue, the following is a high-confidence hypothesis based on code analysis. The actual error message from the ErrorBoundary would confirm the exact crash point.

The most likely root cause is an unhandled error during the report rendering flow after navigating from Classic to the Concierge chat. The crash flow is:

1. Classic redirects to new.expensify.com/concierge (or via /transition?exitTo=/concierge)
2. ConciergePage (src/pages/ConciergePage.tsx:29-43) mounts, waits for data, then calls navigateToConciergeChat() which navigates to ReportScreen
3. ReportScreen renders and subscribes to ONYXKEYS.DERIVED.VISIBLE_REPORT_ACTIONS via useOnyx (src/pages/inbox/ReportScreen.tsx)
4. After a few seconds (while data loads and derived values compute), a re-render triggers an uncaught error that the ErrorBoundary catches

There are several potential crash vectors identified in the code:

Primary suspect — VISIBLE_REPORT_ACTIONS derived value (PR #81280):

• This derived value (src/libs/actions/OnyxDerived/configs/visibleReportActions.ts) computes visibility for ALL report actions across ALL reports with no try-catch error handling in createOnyxDerivedValueConfig. If any computation throws, it propagates to the component.
• PR Report Actions Derived Value #81280 was a massive change (951 additions, 38 files) merged Feb 26, 2026, included in v9.3.27-0. It has a history of instability — the feature was introduced, reverted, and re-introduced multiple times (git log shows 3 reverts touching this file). PR [CP Staging] Fix LHN system message preview not matching report content #83742 explicitly acknowledges it caused a regression.

Secondary crash vectors:

• ReportActionsView uses actions.pop()! with a non-null assertion (src/pages/inbox/report/ReportActionsView.tsx:210) — if the actions array is empty, this returns undefined and accessing .pendingAction on line 212 throws a TypeError
• ReportActionCompose has explicit throw new Error(...) statements (src/pages/inbox/report/ReportActionCompose/ReportActionCompose.tsx:310-316,474-482) if composerRef.clearWorklet is not set when the composer initializes
• ConciergeThinkingMessage calls Parser.replace() inside RenderHTML (src/pages/home/report/ConciergeThinkingMessage.tsx:199) — malformed reasoning data could cause a render crash

Contributing factor — v9.3.27 staging instability:
The v9.3.27 staging cycle had 5 PRs reverted via cherry-picks, an unusually high number. Notably, the revert of useLoadReportActions changes (v9.3.27-3) directly affects how report actions load for ReportScreen, and the revert of InteractionManager migration (v9.3.27-2) touched Navigation.ts and Report/index.ts. These reverts may create subtle data-timing inconsistencies with PR #81280's derived value changes.

What changes do you think we should make in order to solve the problem?

Step 1: Get the actual error from logs — The Sentry/error logs referenced in the issue (internal SO link) should contain the exact JavaScript error message and component stack trace from the ErrorBoundary. This will pinpoint the exact crash location among the candidates identified above.

Step 2: Based on the most likely crash points, the fix should include:

1. Add error handling to the VISIBLE_REPORT_ACTIONS derived value computation in src/libs/actions/OnyxDerived/configs/visibleReportActions.ts — wrap the compute function in a try-catch to prevent computation errors from propagating to components.

2. Remove the non-null assertion actions.pop()! in src/pages/inbox/report/ReportActionsView.tsx:210 and add a null check before accessing createdAction.pendingAction on the next line.

3. Add defensive null checks in ReportScreen for the derived value data paths that could be undefined during initial load.

What alternative solutions did you explore? (Optional)

• Reverting PR Report Actions Derived Value #81280 entirely would likely fix the crash but would remove the performance optimization the derived value provides. A targeted fix (error handling + null checks) is preferable.
• Related issue [$250] Android app freezes and shows “Uh-oh, something went wrong!” when navigating #83886 reports the same "Uh-oh" crash on Android at v9.3.25-13, suggesting this may be a broader issue with the derived value system rather than specific to the Concierge flow.

---

Next Steps for Contributor+ team: Reply with @MelvinBot implement this to create a draft PR, or @MelvinBot <your feedback> to refine this analysis.

[melvin-bot]

Triggered auto assignment to Contributor-plus team member for initial proposal review - @aimane-chnaif (External)

[melvin-bot]

Job added to Upwork: https://www.upwork.com/jobs/~022028954940716757609

[melvin-bot]

@aimane-chnaif Eep! 4 days overdue now. Issues have feelings too...

[melvin-bot]

@aimane-chnaif Eep! 4 days overdue now. Issues have feelings too...