move fcm token storage #1175
move fcm token storage #1175
Conversation
WalkthroughMoves FCM token persistence from the proving machine to a persisted settings store, emits a new PROVING_BEGIN_GENERATION event from the proving machine, and adds a provider listener that conditionally registers the device token via a dynamically imported notification service. Updates UI to store token in settings. Changes
Sequence Diagram(s)sequenceDiagram
autonumber
actor User
participant ConfirmBelongingScreen
participant SettingStore
participant ProvingMachine
participant SelfClientProvider
participant NotificationService
User->>ConfirmBelongingScreen: confirm device
ConfirmBelongingScreen->>ConfirmBelongingScreen: obtain FCM token
ConfirmBelongingScreen->>SettingStore: setFcmToken(token)
Note right of SettingStore #D3F9D8: persisted fcmToken
User->>ProvingMachine: start proving
ProvingMachine-->>SelfClientProvider: emit PROVING_BEGIN_GENERATION {uuid,isMock,context}
alt fcmToken exists in SettingStore
SelfClientProvider->>SelfClientProvider: dynamic import notificationService
SelfClientProvider->>NotificationService: registerDeviceToken(uuid, fcmToken, isMock)
NotificationService-->>SelfClientProvider: success / error
SelfClientProvider->>SelfClientProvider: log analytics & proof events
else no token present
SelfClientProvider->>SelfClientProvider: skip registration
end
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes Possibly related PRs
Suggested labels
Suggested reviewers
Poem
Pre-merge checks and finishing touches❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✨ Finishing touches
🧪 Generate unit tests
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
This reverts commit 4854d6f.
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Nitpick comments (2)
app/src/providers/selfClientProvider.tsx (2)
167-167: Consider logging when token is missing.The code silently skips token registration if
fcmTokenis null/empty. Consider adding a debug-level log to help troubleshoot cases where notifications aren't working because the token wasn't stored.if (fcmToken) { + } else { + logProofEvent('debug', 'Skipping device token registration - no token available', context); + }
162-189: Replace console.error in the PROVING_BEGIN_GENERATION listener with your centralized logging utility. The notification service’sregisterDeviceTokenonly logs trimmed tokens (first/last 5 chars) and does not expose full FCM tokens or UUIDs, so you can safely remove the rawconsole.errorto prevent unstructured stack traces in production.
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (5)
app/src/providers/selfClientProvider.tsx(2 hunks)app/src/screens/prove/ConfirmBelongingScreen.tsx(3 hunks)app/src/stores/settingStore.ts(2 hunks)packages/mobile-sdk-alpha/src/proving/provingMachine.ts(2 hunks)packages/mobile-sdk-alpha/src/types/events.ts(2 hunks)
🧰 Additional context used
📓 Path-based instructions (6)
packages/mobile-sdk-alpha/**/*.{ts,tsx}
📄 CodeRabbit inference engine (packages/mobile-sdk-alpha/AGENTS.md)
packages/mobile-sdk-alpha/**/*.{ts,tsx}: Use strict TypeScript type checking across the codebase
Follow ESLint TypeScript-specific rules
Avoid introducing circular dependencies
Files:
packages/mobile-sdk-alpha/src/types/events.tspackages/mobile-sdk-alpha/src/proving/provingMachine.ts
**/*.{js,ts,tsx,jsx,sol,nr}
📄 CodeRabbit inference engine (.cursorrules)
**/*.{js,ts,tsx,jsx,sol,nr}: NEVER log sensitive data including PII (names, DOB, passport numbers, addresses), credentials, tokens, API keys, private keys, or session identifiers.
ALWAYS redact/mask sensitive fields in logs using consistent patterns (e.g.,***-***-1234for passport numbers,J*** D***for names).
Files:
packages/mobile-sdk-alpha/src/types/events.tsapp/src/stores/settingStore.tspackages/mobile-sdk-alpha/src/proving/provingMachine.tsapp/src/providers/selfClientProvider.tsxapp/src/screens/prove/ConfirmBelongingScreen.tsx
packages/mobile-sdk-alpha/**/*.{ts,tsx,js,jsx}
⚙️ CodeRabbit configuration file
packages/mobile-sdk-alpha/**/*.{ts,tsx,js,jsx}: Review alpha mobile SDK code for:
- API consistency with core SDK
- Platform-neutral abstractions
- Performance considerations
- Clear experimental notes or TODOs
Files:
packages/mobile-sdk-alpha/src/types/events.tspackages/mobile-sdk-alpha/src/proving/provingMachine.ts
app/**/*.{ts,tsx}
📄 CodeRabbit inference engine (app/AGENTS.md)
Type checking must pass before PRs (yarn types)
Files:
app/src/stores/settingStore.tsapp/src/providers/selfClientProvider.tsxapp/src/screens/prove/ConfirmBelongingScreen.tsx
app/src/**/*.{ts,tsx,js,jsx}
⚙️ CodeRabbit configuration file
app/src/**/*.{ts,tsx,js,jsx}: Review React Native TypeScript code for:
- Component architecture and reusability
- State management patterns
- Performance optimizations
- TypeScript type safety
- React hooks usage and dependencies
- Navigation patterns
Files:
app/src/stores/settingStore.tsapp/src/providers/selfClientProvider.tsxapp/src/screens/prove/ConfirmBelongingScreen.tsx
packages/mobile-sdk-alpha/src/proving/**
📄 CodeRabbit inference engine (.cursor/rules/mobile-sdk-migration.mdc)
Place proof input generation in packages/mobile-sdk-alpha/src/proving/
Files:
packages/mobile-sdk-alpha/src/proving/provingMachine.ts
🧬 Code graph analysis (3)
packages/mobile-sdk-alpha/src/types/events.ts (1)
packages/mobile-sdk-alpha/src/proving/internal/logging.ts (1)
ProofContext(16-19)
app/src/providers/selfClientProvider.tsx (3)
app/src/stores/settingStore.ts (1)
useSettingStore(37-95)app/src/Sentry.ts (1)
logProofEvent(260-265)app/src/Sentry.web.ts (1)
logProofEvent(263-268)
app/src/screens/prove/ConfirmBelongingScreen.tsx (1)
app/src/stores/settingStore.ts (1)
useSettingStore(37-95)
⏰ Context from checks skipped due to timeout of 300000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (7)
- GitHub Check: Cursor Bugbot
- GitHub Check: android-build-test
- GitHub Check: e2e-ios
- GitHub Check: analyze-android
- GitHub Check: analyze-ios
- GitHub Check: build-deps
- GitHub Check: Cursor Bugbot
🔇 Additional comments (10)
packages/mobile-sdk-alpha/src/types/events.ts (2)
72-78: LGTM: Well-documented event addition.The new
PROVING_BEGIN_GENERATIONevent is clearly documented with appropriate guidance on its recommended usage. The naming convention aligns with existing SDK events.
108-112: Verified ProofContext safe
ProofContextextendsBaseContext, which only includessessionId, optionaluserId,platform, andstage; pluscircuitTypeandcurrentState. No PII, credentials, or sensitive fields are exposed.app/src/stores/settingStore.ts (2)
74-75: LGTM: Clean token state management.The
fcmTokeninitialization and setter implementation follow the existing pattern in the store. Nullable type is appropriate for optional FCM token.
23-24: Fix typo and ensure FCM token persistence hygiene.
- Change the comment in
app/src/stores/settingStore.ts:35from “sensative” to “sensitive.”- Confirm that storing
fcmTokenin plain AsyncStorage aligns with your security requirements (e.g., encryption) and implement clearing it (e.g.setFcmToken(null)) on logout or app reset.app/src/screens/prove/ConfirmBelongingScreen.tsx (3)
45-45: LGTM: Clean migration to settings store.The switch from
useProvingStoretouseSettingStorefor FCM token management correctly reflects the separation of concerns. The token is now stored in settings rather than proving state.
78-78: LGTM: Simplified token setter signature.The
setFcmToken(token)call correctly uses the new simplified signature that no longer requires theselfClientparameter. This is cleaner since the settings store doesn't need SDK client context.
67-102: Verify notification failure doesn't break proving flow.The
onOkPressflow stores the FCM token before starting the proving process. IfgetFCMToken()throws an exception, the token won't be stored, butsetUserConfirmedwill still be called (lines 84-87), allowing the proving flow to continue.Confirm this is intentional behavior - that users can proceed with identity verification even if notification registration fails. This seems reasonable, but worth documenting if not already clear.
packages/mobile-sdk-alpha/src/proving/provingMachine.ts (3)
1197-1197: LGTM: FCM token removed from proving state.The
fcmTokenhas been correctly removed from the proving state destructuring, aligning with the refactor to move token management to the settings store.
1219-1224: LGTM: Event emission correctly placed.The
PROVING_BEGIN_GENERATIONevent is emitted at the right time - after validation of prerequisites but before the try block that generates and sends the payload. This ensures:
- Listeners can register the FCM token before TEE processing begins
- Event fires with validated
uuidandpassportData- The context provides proper logging information
The event is fire-and-forget (not awaited), which is appropriate since token registration is a side effect that shouldn't block proof generation.
1222-1222: Validate mock flag assignment and fallback
EnsurepassportData.mockis explicitly set totruein all test and staging flows and thatpassportData?.mock ?? falsecorrectly defaults to production when the flag is absent.
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (1)
app/src/providers/selfClientProvider.tsx (1)
99-107: Consider refactoring to reuse the notification adapter.The notification adapter defined at lines 99-107 already performs the same dynamic import and call to
registerFirebaseDeviceToken. The new listener code (lines 172-174) duplicates this logic.Consider calling the adapter's
registerDeviceTokenmethod instead:- const { registerDeviceToken: registerFirebaseDeviceToken } = - await import('@/utils/notifications/notificationService'); - await registerFirebaseDeviceToken(uuid, fcmToken, isMock); + await adapters.notification.registerDeviceToken(uuid, fcmToken, isMock);However, note that
adaptersis defined in the outer scope and this listener is created in auseMemo, so you'd need to ensure proper dependency tracking or pass adapters into the listener factory. The current approach with separate imports is acceptable if you prefer to keep concerns separated.Also applies to: 172-174
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (2)
app/src/providers/selfClientProvider.tsx(2 hunks)sdk/core/src/SelfBackendVerifier.ts(1 hunks)
✅ Files skipped from review due to trivial changes (1)
- sdk/core/src/SelfBackendVerifier.ts
🧰 Additional context used
📓 Path-based instructions (3)
**/*.{js,ts,tsx,jsx,sol,nr}
📄 CodeRabbit inference engine (.cursorrules)
**/*.{js,ts,tsx,jsx,sol,nr}: NEVER log sensitive data including PII (names, DOB, passport numbers, addresses), credentials, tokens, API keys, private keys, or session identifiers.
ALWAYS redact/mask sensitive fields in logs using consistent patterns (e.g.,***-***-1234for passport numbers,J*** D***for names).
Files:
app/src/providers/selfClientProvider.tsx
app/**/*.{ts,tsx}
📄 CodeRabbit inference engine (app/AGENTS.md)
Type checking must pass before PRs (yarn types)
Files:
app/src/providers/selfClientProvider.tsx
app/src/**/*.{ts,tsx,js,jsx}
⚙️ CodeRabbit configuration file
app/src/**/*.{ts,tsx,js,jsx}: Review React Native TypeScript code for:
- Component architecture and reusability
- State management patterns
- Performance optimizations
- TypeScript type safety
- React hooks usage and dependencies
- Navigation patterns
Files:
app/src/providers/selfClientProvider.tsx
🧬 Code graph analysis (1)
app/src/providers/selfClientProvider.tsx (2)
app/src/stores/settingStore.ts (1)
useSettingStore(37-95)app/src/Sentry.ts (1)
logProofEvent(260-265)
⏰ Context from checks skipped due to timeout of 300000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (7)
- GitHub Check: Cursor Bugbot
- GitHub Check: build-deps
- GitHub Check: e2e-ios
- GitHub Check: android-build-test
- GitHub Check: analyze-android
- GitHub Check: analyze-ios
- GitHub Check: Cursor Bugbot
🔇 Additional comments (2)
app/src/providers/selfClientProvider.tsx (2)
23-23: LGTM - Import addition is appropriate.The
useSettingStoreimport is correctly added to support accessing the persisted FCM token.
164-189: No sensitive data exposed in ProofContext.ProofContext contains only sessionId, optional userId (non-PII identifier), platform, stage, circuitType, and currentState—none of which are sensitive credentials or PII. Logging this context is safe.
| } | ||
| }, | ||
| ); | ||
|
|
There was a problem hiding this comment.
Potential sensitive data exposure in error logging.
The console.error statement logs the raw error object, which could potentially contain sensitive information such as tokens, API responses, or PII depending on what the registerFirebaseDeviceToken function might throw.
Consider sanitizing the error before logging:
- console.error('Error registering device token:', error);
+ console.error('Error registering device token');The error details are already captured in the analytics event and logProofEvent call, so the console.error can be simplified to avoid potential data leakage.
As per coding guidelines: Never log sensitive data including credentials, tokens, API keys, or private keys.
📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| console.error('Error registering device token:', error); | |
| console.error('Error registering device token'); |
🤖 Prompt for AI Agents
In app/src/providers/selfClientProvider.tsx around line 182, the console.error
currently logs the raw error object which may contain sensitive data; replace it
with a non-sensitive message (e.g., a generic "Failed to register device token")
or log only a sanitized error identifier/code, and remove any token/PII from the
output; ensure analytics/logProofEvent still capture full details privately,
while the console log avoids printing sensitive fields.
There was a problem hiding this comment.
@aaronmgdr Understood, I'll leave the logging as-is. Thanks for clarifying!
🧠 Learnings used
Learnt from: CR
PR: selfxyz/self#0
File: .cursorrules:0-0
Timestamp: 2025-09-22T11:10:22.019Z
Learning: Applies to **/*.{js,ts,tsx,jsx,sol,nr} : NEVER log sensitive data including PII (names, DOB, passport numbers, addresses), credentials, tokens, API keys, private keys, or session identifiers.
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
| } | ||
| } | ||
| }, | ||
| ); |
There was a problem hiding this comment.
Bug: FCM Token Registration Race Conditions
The FCM token registration, now an async PROVING_BEGIN_GENERATION event listener, is prone to two race conditions. The proving flow doesn't await its completion, and the fcmToken can be null if the settings store hasn't rehydrated. This may cause users to miss push notifications for proof status updates.
Additional Locations (1)
There was a problem hiding this comment.
the proving flow doesnt need the notification for completion
There was a problem hiding this comment.
the token will be there.... trust me
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Nitpick comments (1)
app/src/providers/selfClientProvider.tsx (1)
153-180: Improve error handling, token freshness, and idempotency in PROVING_BEGIN_GENERATION listener.
- Wrap the dynamic
import('@/utils/notifications/notificationService')in its own try/catch so import-specific failures can be logged separately from registration errors.- Fetch the current FCM token at runtime (e.g. via
await getFCMToken()) instead of using the potentially stale value from the settings store.- Debounce or dedupe concurrent
registerDeviceTokencalls (or confirm the backend gracefully handles duplicates) to avoid race-condition spikes.
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (9)
app/src/providers/selfClientProvider.tsx(2 hunks)app/tests/src/utils/proving/validateDocument.test.ts(0 hunks)app/tests/utils/selfClientProvider.ts(0 hunks)packages/mobile-sdk-alpha/src/client.ts(1 hunks)packages/mobile-sdk-alpha/src/types/public.ts(0 hunks)packages/mobile-sdk-alpha/tests/client.test.ts(8 hunks)packages/mobile-sdk-alpha/tests/documents/utils.test.ts(0 hunks)packages/mobile-sdk-alpha/tests/utils/testHelpers.ts(0 hunks)packages/mobile-sdk-demo/src/providers/SelfClientProvider.tsx(0 hunks)
💤 Files with no reviewable changes (6)
- app/tests/src/utils/proving/validateDocument.test.ts
- packages/mobile-sdk-demo/src/providers/SelfClientProvider.tsx
- app/tests/utils/selfClientProvider.ts
- packages/mobile-sdk-alpha/src/types/public.ts
- packages/mobile-sdk-alpha/tests/documents/utils.test.ts
- packages/mobile-sdk-alpha/tests/utils/testHelpers.ts
🧰 Additional context used
📓 Path-based instructions (7)
**/*.{js,ts,tsx,jsx,sol,nr}
📄 CodeRabbit inference engine (.cursorrules)
**/*.{js,ts,tsx,jsx,sol,nr}: NEVER log sensitive data including PII (names, DOB, passport numbers, addresses), credentials, tokens, API keys, private keys, or session identifiers.
ALWAYS redact/mask sensitive fields in logs using consistent patterns (e.g.,***-***-1234for passport numbers,J*** D***for names).
Files:
app/src/providers/selfClientProvider.tsxpackages/mobile-sdk-alpha/src/client.tspackages/mobile-sdk-alpha/tests/client.test.ts
app/**/*.{ts,tsx}
📄 CodeRabbit inference engine (app/AGENTS.md)
Type checking must pass before PRs (yarn types)
Files:
app/src/providers/selfClientProvider.tsx
app/src/**/*.{ts,tsx,js,jsx}
⚙️ CodeRabbit configuration file
app/src/**/*.{ts,tsx,js,jsx}: Review React Native TypeScript code for:
- Component architecture and reusability
- State management patterns
- Performance optimizations
- TypeScript type safety
- React hooks usage and dependencies
- Navigation patterns
Files:
app/src/providers/selfClientProvider.tsx
packages/mobile-sdk-alpha/**/*.{ts,tsx}
📄 CodeRabbit inference engine (packages/mobile-sdk-alpha/AGENTS.md)
packages/mobile-sdk-alpha/**/*.{ts,tsx}: Use strict TypeScript type checking across the codebase
Follow ESLint TypeScript-specific rules
Avoid introducing circular dependencies
Files:
packages/mobile-sdk-alpha/src/client.tspackages/mobile-sdk-alpha/tests/client.test.ts
packages/mobile-sdk-alpha/**/*.{ts,tsx,js,jsx}
⚙️ CodeRabbit configuration file
packages/mobile-sdk-alpha/**/*.{ts,tsx,js,jsx}: Review alpha mobile SDK code for:
- API consistency with core SDK
- Platform-neutral abstractions
- Performance considerations
- Clear experimental notes or TODOs
Files:
packages/mobile-sdk-alpha/src/client.tspackages/mobile-sdk-alpha/tests/client.test.ts
packages/mobile-sdk-alpha/{**/*.test.{ts,tsx},**/__tests__/**/*.{ts,tsx}}
📄 CodeRabbit inference engine (packages/mobile-sdk-alpha/AGENTS.md)
packages/mobile-sdk-alpha/{**/*.test.{ts,tsx},**/__tests__/**/*.{ts,tsx}}: Do NOT mock @selfxyz/mobile-sdk-alpha in tests (avoid jest.mock('@selfxyz/mobile-sdk-alpha') and replacing real functions with mocks)
Use actual imports from @selfxyz/mobile-sdk-alpha in tests
Write integration tests that exercise the real validation logic (not mocks)
Test isPassportDataValid() with realistic synthetic passport data (never real user data)
Verify extractMRZInfo() using published sample MRZ strings (e.g., ICAO examples)
Ensure parseNFCResponse() works with representative, synthetic NFC data
Never use real user PII in tests; use only synthetic, anonymized, or approved test vectors
Files:
packages/mobile-sdk-alpha/tests/client.test.ts
**/*.{test,spec}.{ts,js,tsx,jsx}
⚙️ CodeRabbit configuration file
**/*.{test,spec}.{ts,js,tsx,jsx}: Review test files for:
- Test coverage completeness
- Test case quality and edge cases
- Mock usage appropriateness
- Test readability and maintainability
Files:
packages/mobile-sdk-alpha/tests/client.test.ts
🧠 Learnings (5)
📚 Learning: 2025-09-22T11:10:22.019Z
Learnt from: CR
PR: selfxyz/self#0
File: .cursorrules:0-0
Timestamp: 2025-09-22T11:10:22.019Z
Learning: Applies to **/*.{js,ts,tsx,jsx,sol,nr} : NEVER log sensitive data including PII (names, DOB, passport numbers, addresses), credentials, tokens, API keys, private keys, or session identifiers.
Applied to files:
app/src/providers/selfClientProvider.tsx
📚 Learning: 2025-08-29T15:31:15.924Z
Learnt from: CR
PR: selfxyz/self#0
File: packages/mobile-sdk-alpha/AGENTS.md:0-0
Timestamp: 2025-08-29T15:31:15.924Z
Learning: Applies to packages/mobile-sdk-alpha/{**/*.test.{ts,tsx},**/__tests__/**/*.{ts,tsx}} : Use actual imports from selfxyz/mobile-sdk-alpha in tests
Applied to files:
packages/mobile-sdk-alpha/tests/client.test.ts
📚 Learning: 2025-08-29T15:31:15.924Z
Learnt from: CR
PR: selfxyz/self#0
File: packages/mobile-sdk-alpha/AGENTS.md:0-0
Timestamp: 2025-08-29T15:31:15.924Z
Learning: Applies to packages/mobile-sdk-alpha/{**/*.test.{ts,tsx},**/__tests__/**/*.{ts,tsx}} : Do NOT mock selfxyz/mobile-sdk-alpha in tests (avoid jest.mock('selfxyz/mobile-sdk-alpha') and replacing real functions with mocks)
Applied to files:
packages/mobile-sdk-alpha/tests/client.test.ts
📚 Learning: 2025-08-29T15:31:15.924Z
Learnt from: CR
PR: selfxyz/self#0
File: packages/mobile-sdk-alpha/AGENTS.md:0-0
Timestamp: 2025-08-29T15:31:15.924Z
Learning: Applies to packages/mobile-sdk-alpha/{**/*.test.{ts,tsx},**/__tests__/**/*.{ts,tsx}} : Never use real user PII in tests; use only synthetic, anonymized, or approved test vectors
Applied to files:
packages/mobile-sdk-alpha/tests/client.test.ts
📚 Learning: 2025-08-29T15:31:15.924Z
Learnt from: CR
PR: selfxyz/self#0
File: packages/mobile-sdk-alpha/AGENTS.md:0-0
Timestamp: 2025-08-29T15:31:15.924Z
Learning: Applies to packages/mobile-sdk-alpha/{**/*.test.{ts,tsx},**/__tests__/**/*.{ts,tsx}} : Verify extractMRZInfo() using published sample MRZ strings (e.g., ICAO examples)
Applied to files:
packages/mobile-sdk-alpha/tests/client.test.ts
🧬 Code graph analysis (1)
app/src/providers/selfClientProvider.tsx (3)
app/src/stores/settingStore.ts (1)
useSettingStore(37-95)app/src/Sentry.ts (1)
logProofEvent(260-265)app/src/Sentry.web.ts (1)
logProofEvent(263-268)
⏰ Context from checks skipped due to timeout of 300000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (4)
- GitHub Check: build-deps
- GitHub Check: analyze-ios
- GitHub Check: analyze-android
- GitHub Check: e2e-ios
🔇 Additional comments (4)
packages/mobile-sdk-alpha/src/client.ts (1)
43-43: LGTM: Notification adapter correctly removed from required adapters.The removal of
'notification'fromREQUIRED_ADAPTERSis consistent with the PR's goal of moving FCM token management out of the SDK layer and into the app's settings store.app/src/providers/selfClientProvider.tsx (1)
23-23: LGTM: Settings store import aligns with token storage refactor.The import of
useSettingStorecorrectly supports the new flow wherefcmTokenis persisted in the app's settings store rather than the SDK's proving store.packages/mobile-sdk-alpha/tests/client.test.ts (2)
9-9: LGTM: Import updated to remove NotificationAdapter.The removal of
NotificationAdapterfrom the import statement correctly reflects the API surface change where notification handling has been removed from the SDK.
52-174: LGTM: Test fixtures updated to remove notification adapter.All test cases have been correctly updated to remove the
notificationadapter from theadaptersobject. The tests continue to validate required adapters (scanner, network, crypto, documents, auth) and cover the core SDK functionality without notification dependencies.
move fcm token from proving store to setting store for better separation of concerns.
having fcm token in the proving machine felt a bit out of place. its not core to proving in any way. moving it out means we dont need to expose it to the sdk. this is good because sdk users can setup notifications how they wish.
Note
Removes notification adapter from SDK, adds a PROVING_BEGIN_GENERATION event, moves FCM token to Settings store, and handles device token registration in the app via a listener.
notificationadapter andregisterNotificationsTokenfromSelfClient; drop fromAdapters/REQUIRED_ADAPTERS.SdkEvents.PROVING_BEGIN_GENERATIONwith payload (uuid,isMock,context).PROVING_BEGIN_GENERATIONat start; removefcmToken/setFcmTokenfrom proving state.fcmTokenandsetFcmTokentouseSettingStore(persisted).useSettingStoreinstead of proving store.PROVING_BEGIN_GENERATIONto register device token using storedfcmToken; remove SDK-level notification adapter usage.SelfBackendVerifier.Written by Cursor Bugbot for commit e5522cc. This will update automatically on new commits. Configure here.
Summary by CodeRabbit
New Features
Refactor
Breaking Changes
Tests