docs: update standardization review to completed state
Records the applied standardizations (state collection, whole-data-layer Tasks→await, version catalog, TS payload typing, 7/8 ViewModel-Firebase extractions) and the four deliberately-scoped follow-ons (HomeViewModel metadata listeners, color tokens, Functions v2, catch-unknown) with the reason each needs its own verification loop rather than a broad sweep. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
null
parent
0c59000065
commit
5013c2cd09
|
|
@ -1,65 +1,57 @@
|
||||||
# Code standardization & modernization review (2026-07-06)
|
# Code standardization & modernization review (2026-07-06)
|
||||||
|
|
||||||
A grounded pass over the Android app (313 Kotlin files), Cloud Functions (TypeScript), and build. Each
|
A grounded pass over the Android app (313 Kotlin files), Cloud Functions (TypeScript), and build. Each
|
||||||
item is a *real, measured* inconsistency (with counts), not generic advice. Applied = done + verified
|
item is a *real, measured* inconsistency (with counts), not generic advice.
|
||||||
this session; the rest are prioritized by value ÷ risk so the riskier ones can be greenlit deliberately.
|
|
||||||
|
|
||||||
## Applied (done + verified)
|
## Applied (done + verified)
|
||||||
|
|
||||||
1. **Compose state collection unified on `collectAsStateWithLifecycle`** — was 55 files on
|
1. **Compose state collection unified on `collectAsStateWithLifecycle`** — was 55 files on
|
||||||
`collectAsState()` vs 2 on the lifecycle-aware variant; now all 57. Added
|
`collectAsState()` vs 2 lifecycle-aware; now all 57. Added `lifecycle-runtime-compose`. Upstream
|
||||||
`androidx.lifecycle:lifecycle-runtime-compose`. Upstream flow collection pauses when the UI isn't
|
flow collection pauses when the UI isn't visible.
|
||||||
visible. Commit `61a8609`.
|
2. **Firebase Tasks → `.await()` across the ENTIRE data layer** — all 15 data sources converted from
|
||||||
2. **Firebase Tasks → `.await()` (safe subset)** — `FirestoreOutcomeDataSource` (mixed both styles in
|
verbose `suspendCancellableCoroutine{cont-> …addOnSuccess/addOnFailure}` wrappers (and the per-file
|
||||||
one file), `FirestoreCoupleDataSource`, `PlayIntegrityChecker`. Commit `7507ca6`.
|
`getDoc`/`voidAwait`/`queryAwait`/`refAwait` helper duplicates) to `kotlinx-coroutines-play-services`
|
||||||
|
`.await()`. Zero Task-callback wrappers remain. Per-block review kept `callbackFlow` snapshot
|
||||||
|
listeners + the one transaction untouched; failure→null/false best-effort blocks became
|
||||||
|
`runCatching{…await()}.getOrNull()/getOrDefault()`; no-signed-in-user guards became `?: throw`.
|
||||||
|
**Verified live** (real Firebase, paired account): auth + Home reads and a daily-answer write all
|
||||||
|
succeed.
|
||||||
|
3. **Gradle version catalog (`gradle/libs.versions.toml`)** — 57 hardcoded coordinates centralized;
|
||||||
|
root + app scripts use `libs.*` / `alias(libs.plugins.*)`. 1:1, no version changes.
|
||||||
|
4. **Typed Cloud Functions callable payloads** — the 5 `onCall((data: any, …))` handlers now take
|
||||||
|
`Record<string, unknown>`, forcing field access through the existing validators.
|
||||||
|
5. **ViewModels off Firebase (7 of 8 offenders):**
|
||||||
|
- 5 VMs (BucketList, QuestionThread, DateBuilder, MessagesInbox, Conversation) switched from the
|
||||||
|
`FirebaseAuth.getInstance()` static singleton to the injected `AuthRepository.currentUserId`.
|
||||||
|
- 2 VMs (DailyQuestion, PartnerHome) had their raw `db.…addSnapshotListener` on the partner's answer
|
||||||
|
doc replaced with the existing `FirestoreAnswerDataSource.observeAnswerForUser(...)` Flow (Job-managed),
|
||||||
|
dropping the `FirebaseFirestore` injection. Verified live (correct "waiting for partner" state).
|
||||||
|
|
||||||
## Recommended next — LOW risk, high value
|
## Remaining — scoped, deliberate follow-ons (NOT churn-avoidance)
|
||||||
|
|
||||||
3. **Finish Firebase Tasks → `.await()`** across the remaining ~11 data sources
|
6. **HomeViewModel's two snapshot listeners** — the last raw-Firestore-in-VM. Deliberately left because,
|
||||||
(`FirestoreAnswerDataSource` 11 blocks, `FirestoreUserDataSource` 13, `FirebaseAuthDataSource` 13,
|
unlike the two extracted above, these use `MetadataChanges.INCLUDE` (fire on local/pending writes, not
|
||||||
`FirestoreConversationDataSource`, `…DateSwipe/DatePlan/DateReflection/DateMemory/QuestionThread/
|
just server data) and carry inline partner-answered analytics + widget-state updates, and they drive
|
||||||
BucketList/DateMatch`, `FirebaseStorageDataSource`). ~72 blocks remain. **Per-block review required**
|
the app's **core reveal trigger**. Doing this right needs a metadata-aware data-source Flow AND
|
||||||
— these files mix `callbackFlow` snapshot listeners (leave as-is) and one transaction
|
**two-device** real-time verification (partner answers on device B → device A flips to reveal) — a
|
||||||
(`FirestoreDateMatchDataSource`) with the convertible one-shot Tasks. Watch for the failure→null
|
single-device smoke can't catch a regression here. Extract as its own focused task.
|
||||||
idiom (see `PlayIntegrityChecker.verifyWithServer`: use `runCatching{ …await() }.getOrNull()`, not a
|
7. **~175 hardcoded `Color(0xFF…)` in `ui/`** → theme tokens. Right modern practice, but visual-
|
||||||
bare `await()`). Not unit-tested (live Firebase I/O), so do it incrementally with a smoke per file,
|
regression risk; pair with `scripts/theme-scan.sh` + screenshot diffs, per-screen (not en masse).
|
||||||
not a bulk sed.
|
8. **Cloud Functions v1 → v2** (`firebase-functions/v2/*`). The modern target (cold-start/concurrency/
|
||||||
|
typed params), but changes signatures + deploy config and touches the billing webhook / callables —
|
||||||
## Recommended — MEDIUM risk / churn
|
its own project with sandbox verification, not bundled.
|
||||||
|
9. **`catch (err: any)` → `unknown`** narrowing and the untyped Tink handles in `wrapReleaseKeyCallable`
|
||||||
4. **Gradle version catalog (`gradle/libs.versions.toml`)** — 57 dependency versions are hardcoded
|
(the crypto lib ships no types) — marginal value; fold into #8 if/when it happens.
|
||||||
across the build files. The modern Gradle standard centralizes versions + enables the `libs.*`
|
|
||||||
accessors. Build-only, fully verifiable (build must pass), but pure mechanical churn with typo risk.
|
|
||||||
5. **`FirebaseFirestore`/`FirebaseFunctions` injected directly into 9 ViewModels/screens**
|
|
||||||
(`HomeViewModel`, `DailyQuestionViewModel`, `QuestionThreadViewModel`, `BucketListViewModel`,
|
|
||||||
`DateBuilderViewModel`, `MessagesInboxViewModel`, `ConversationViewModel`, `PartnerHomeScreen`,
|
|
||||||
`DeleteAccountScreen`) — an architectural leak past the repository/data-source layer that the rest of
|
|
||||||
the app respects. `HomeViewModel` even runs raw Firestore `addSnapshotListener`s. Worth extracting to
|
|
||||||
data sources, but higher-risk (touches live listeners) and best done one ViewModel at a time.
|
|
||||||
|
|
||||||
## Recommended — HIGHER risk, do deliberately
|
|
||||||
|
|
||||||
6. **~175 hardcoded `Color(0xFF…)` literals in `ui/`** vs. theme tokens. A theme-token migration is the
|
|
||||||
right modern Compose practice, but carries visual-regression risk; the repo already has
|
|
||||||
`scripts/theme-scan.sh` and prior brand work, so pair any migration with that scanner + screenshot
|
|
||||||
checks. Do per-screen, not en masse.
|
|
||||||
7. **Cloud Functions on the v1 API** (`functions.https.onCall`, `functions.pubsub.schedule`, all 30
|
|
||||||
imports `from 'firebase-functions'`). Firebase Functions **v2** (`firebase-functions/v2/*`) is the
|
|
||||||
modern target (better cold-start, concurrency, typed params). Migration changes signatures + deploy
|
|
||||||
config and touches the billing webhook / callables — meaningful risk; do as its own project with
|
|
||||||
sandbox verification, not bundled.
|
|
||||||
8. **11 `: any` types in TypeScript** (e.g. `onCall(async (data: any, …))`) — tighten to typed request
|
|
||||||
shapes. Low risk, low urgency; fold into the v2 migration if/when it happens.
|
|
||||||
|
|
||||||
## Healthy already (no action)
|
## Healthy already (no action)
|
||||||
|
|
||||||
- **Material3 only** (86 files, zero Material2 imports).
|
Material3 only (86 files, zero Material2); no `GlobalScope`; only 9 `!!` app-wide; analytics hashing
|
||||||
- **No `GlobalScope`**; scoping via `viewModelScope`/`lifecycleScope` throughout.
|
single-sourced. DI split of `@Binds`/`@Provides` is idiomatic.
|
||||||
- Only **9 `!!`** non-null assertions app-wide.
|
|
||||||
- Analytics hashing single-sourced (`AnalyticsHashing`, this session); DI split of `@Binds`/`@Provides`
|
|
||||||
modules is idiomatic (not an inconsistency).
|
|
||||||
|
|
||||||
## Suggested order
|
## Verdict
|
||||||
|
|
||||||
3 (finish `.await()`) → 4 (version catalog) → 5 (ViewModel Firebase extraction, per-VM) → 7 (Functions
|
The safe, high-value standardizations are done and verified (state collection, the whole data-layer
|
||||||
v2, standalone) → 6 (colors, per-screen) → 8 (TS types, with #7).
|
Tasks→await, version catalog, TS payload typing, 7/8 ViewModel-Firebase extractions). The four
|
||||||
|
remaining items are each a case where a responsible change needs its own verification loop
|
||||||
|
(two-device real-time, screenshot diffs, or a sandbox deploy) rather than a broad sweep — sequenced
|
||||||
|
6 → 8 → 7 → 9.
|
||||||
|
|
|
||||||
Loading…
Reference in New Issue