-
Type:
Bug
-
Resolution: Unresolved
-
Priority:
High
-
Affects Version/s: None
-
STOP TB Sprint 12
Summary
NullPointerException in getBinding() when UI updates run after the Activity/Fragment view has been destroyed (VolunteerActivity, NewBenRegFragment, TBScreeningFormFragment).
Background / Context
Source: StopTB Android Crashlytics RCA (Oct 15–21 2026, release 2.2.3). RCA ref RCA-05, status: Probable Root Cause. Code inspected: release-2.2.3, commit cd02b155 (only VolunteerActivity inspected). Priority: P1 (VolunteerActivity), P2/P3 for the others. 60 events across 3 issues.
Problem Statement
UI update runs after the activity or fragment view has been destroyed, so the _binding!! getter hits null.
Current Behaviour (Actual)
- VolunteerActivity.startCampHubReconnect posts {{runOnUiThread
{ refreshCampHubOfflineBanner() }
}} from an IO coroutine; _binding is nulled in onDestroy, so the getter throws (VolunteerActivity.kt line 129 in release-2.2.3; crash line 124).
- Binding is not guarded against lifecycle end; network callbacks and coroutines outlive the view.
Crashlytics issues:
- VolunteerActivity.getBinding (VolunteerActivity.kt:124) — 50 events / 44 users — repetitive, 1.1.4→2.2.3, 3 variants
- NewBenRegFragment.getBinding (NewBenRegFragment.kt:64) — 8 / 6 — 2.0.3→2.2.3, 2 variants (file not inspected)
- TBScreeningFormFragment.getBinding (TBScreeningFormFragment.kt:52) — 2 / 2 — seen only 2.2.0–2.2.1 (file not inspected)
Expected Behaviour / Requirement
No crash when async work completes after the view is destroyed; UI updates are skipped when the view/binding no longer exists.
User Story
As a volunteer using the app, I want screens to stay stable during rotation, backgrounding or reconnect, so that I do not lose my in-progress work.
Acceptance Criteria
- Given a camp hub reconnect in progress, When the device is rotated or the app is backgrounded/closed, Then no crash occurs (CAPA-12).
- Given VolunteerActivity, When _binding is null or the activity is finishing, Then UI updates are skipped; jobs are cancelled and callbacks unregistered in onDestroy.
- Given NewBenRegFragment and TBScreeningFormFragment, When inspected, Then the same pattern is fixed where present.
- (Preventive, CAPA-13) Lint rule against _binding!! getters passes; view work scoped to viewLifecycleOwner.
Business Rules / Validations
None identified.
Dependencies
None identified.
Impact / Affected Areas
VolunteerActivity.kt (binding getter, startCampHubReconnect, onDestroy), NewBenRegFragment.kt, TBScreeningFormFragment.kt; Camp Hub reconnect flow.
Out of Scope
App-wide refactor of all view bindings beyond the lint rule [TBD whether lint rule is in this ticket].
QA / Testing Considerations
Rotate and background the app during hub reconnect; repeat on New Beneficiary Registration and TB Screening form; confirm no new NPE events in Crashlytics post-release.
Open Questions / Clarifications
- NewBenRegFragment and TBScreeningFormFragment not inspected; same root cause assumed from listing only [TBD].
- Is CAPA-13 (lint rule) to be delivered in 2.5 or tracked separately?
Source: StopTB_Crashlytics_RCA.xlsx (RCA-05, CAPA-12, CAPA-13).
- relates to
-
STOP-785 [Crashlytics][2.4] VolunteerActivity.getBinding NullPointerException still crashing in 2.3 (370 events / 214 users in 30 days)
-
- Pending QA
-