Repository navigation
Fix/post redesign followups - #199
Draft
atulmgupta wants to merge 2726 commits into
Draft
atulmgupta wants to merge 2726 commits into
atulmgupta wants to merge 2726 commits into
Conversation
…ce inputs Root cause: nextcharge.Decide falls through to charge_home_now when Preview() fails (unknown rate plan, bad ready-by, no feasible window) and no Supercharger quote exists - all cost pointers nil. The strip rendered the badge + empty metrics. - Strip suppresses badge+body and shows nextCharge.notEnoughPriceData (with Autopilot CTA) when kwh_needed > 0 but all three price options are empty; 'enough' (0 kWh) still renders priceless. Backend enum untouched (contract ripple; display gate is sufficient). - Regtests fail-before (badge + empty metrics) / pass-after; guards for priced charge-home-now and priceless enough. Strip 26/26. - Catalog +1 key, trio re-pin (42199 keys), 7672/0; tsc/lint/build green
Root cause: all currency paths honored the raw user/metric/global
precision, so a 1-decimal precision rendered \.04 and \.14 alike as
'\.1' (Monthly est, formatCurrency/EnergyCost/Currency/formatMetric);
the Est. cost/km metric additionally ignored the metric precision.
- New currencyDecimals(amount, requested) helper in numberFormat.ts,
wired into formatCurrency, formatEnergyCost, <Currency>, the cost/km
metric formatter (now metric-precision + locale aware), and
formatMetric's currency branch. Amounts >= 1 and non-currency formats
keep the requested precision.
- 5 regtests fail-before ('\.1') / pass-after; target files 245/245;
wide formatting trees 3205 + 3 baseline-only; trio 7672/0;
tsc/lint/build green. Step 2 (a-f) complete.
Replace the DashboardPage 60s setTick (whole-grid re-render) with a shared 30s useTicker subscription: FreshnessDot, VehicleHero ChargeDoneAt leaf, FleetOperationsBrief and AnomalyDetectorWidget subscribe as units; DashboardGrid/VehicleHero memoized; page callbacks stabilized. Tests: useTicker 2/2, DashboardGrid.ticker 2/2 (single timer + bodies render once across ticks), DashboardPage ticker-isolation guard; browser-verified ages advance on ticks (temp spec, deleted after run). tsc/lint/build green, dashboard 4575/0, trio 7672/0.
The browser-local departure contract asserts a divergent host zone (America/Los_Angeles) but relied on ambient TZ: 14/14 fail on any other host zone (verified under UTC and America/New_York). Pin process.env.TZ at the top of the file per the AdvancedIntelligencePages / SmartCharge precedent. Each test file runs in its own worker process (probed), so the pin cannot leak to other files; the beforeEach guards verify it took effect. No shared-state leak found in JourneyAssurancePage.test.tsx after probing zombie paused-mutation resume (none: client.clear drops them), fixture mutation (27/27 with deep-frozen fixture), formatter caches (absent), drawer close timing (synchronous) and worker reuse (fresh process per file). Historical in-suite red was the pre-wave-2 brief contract, fixed by f73d907 (green in reports c/d). Proof: file 1 20/20 + 12 shuffle seeds green; file 2 20/20 under ambient UTC + 14 shuffle seeds green.
…rwrite The page re-derives isCharging/chargeLimitPct under the 2-minute live window but re-injected the raw observation, discarding the model's validSoc sanitization (NaN/Infinity/out-of-range rendered as a limit). Sanitize the overwrite with the now-exported validSoc predicate. Adds validSoc regression coverage in chargeAdvisor.test.ts. Scenario-gate failures (9) remain blocked: scenarios require coverageEstablished via coveredLocalDates, which no caller provides; needs an owner decision.
… isolation Job 1 (15 fails, 0 blocked): - admin.test.ts (5): align stale expectations to tri-state helpers (null/undefined unknown), pin sentry buckets to UTC, expect the toggling row timestamp from findLastLockChange - StatusHero (5): hide decorative LiveIndicator dot from AT inside the hero live region (nested role=status double-announce); pin healthy icon to circle-check-big (CheckCircle alias renders CircleCheckBig) - Watch wiring (2): expect unit_of_length/unit_of_temp the component sends and the Go handler accepts - AIDigestNarration (1): localize weekly_digest in HelixEvidenceTrail allowlist + en.json helix.tools.weeklyDigest + regen catalogs - AIQuietHoursSuggestion (1): expect localized Critical severity label - DataTable (1): expect No-results + Clear-filters no-match copy for filtered-to-zero (emptyMessage is for truly-empty data) Job 2 (flake): pin reduced motion in ChargingForensicsPage.test so FadeIn panels render final state; visibility assertions no longer race the 50ms-delay rAF animation clock. Proof: 10x consecutive + 6 shuffle seeds all 29/29.
31 of 35 failing tests migrated to the new shell contract at equal strictness (desktop rail + docked panel, tablet drawer, phone tabs + sheet, Home hub, Pinned/Recent groups, Tooltip labels, sidenav-collapsed-groups persistence). Prod fix (test-proven, shell-only): persist effect no longer clobbers the saved deck view while its section is late-loading. 4 tests left red by design (no successor contract): icon-only collapsed secondary panel (x2), directional panel entrances ltr/rtl (x2).
…wner-decisions left red) Fixed (prod): signalCorrelation normalizes -0 lagS to +0 when maxLagS < stepS; existing toBe(0) pin is the regression test. Fixed (stale test pins, intent kept): deadband page pins updated to current delta-MAD estimator outputs (14 candidates, exact metrics); replay map-height pin updated to row-15 layout; null-efficiency case now pins shared hide-missing brief contract. Left red for owner: driving NaN-chart test contradicts passing trips NaN pin; gap-detector isStale case contradicts shared deriveDataState contract. See task report for options.
(cherry-picked from fix/step5-d 32871f0bc2; i18n checkpoint conflicts resolved by re-running the repo generators on the merged tree: split-i18n-catalog regen + i18n-trio-checkpoint --typed --checkpoint-catalog, folding date.preset.last6h into the predecessor per the step-2b/H convention instead of the branch's additions-list registration.)
- useDLQ (11): tests used numeric ids + general MAX_LIMIT; hook/types/backend confirm opaque string ring ids + DLQ-specific 500 cap (DLQReplayAuditRepo.Recent clamps [1,500]). Export DLQ_AUDIT_MAX_LIMIT; migrate tests to string ids + cap.
- useGuard (8): guardWriteOptions fail-closed skipAuthRefresh is deliberate; pin it in wire-contract expectations.
- hooks/useSettings (2): request() threads AbortSignal per repo contract (matches useSettings.abort test); expect it.
- useVehicleSystems (1, prod): mediaHistory factory unconditionally appended start/end; align with climateHistory conditional short form. No exact-key refs to long no-range key.
- useAnalytics (1): /vehicle-states/summary returns {by_state} envelope (Go handler_test pins keys); mock the envelope.
- useAutomations (1): import endpoint returns {imported[], errors[]} (dtos.go importResult); mock the array shape.
- useLocations (1): bulk delete has complete/partial/failed outcomes (getGeofenceBulkDeleteOutcome + phase6 tests); success-path test now mocks a complete outcome + warning spy.
- api/useSettings (1+2 new): poll redesign invalidates status only, history follows last_poll_time chain, honest triggered toast (a3c4e5a); update poll test + cover the chain both ways.
- authRecovery (1): directRequest camelCaseKeys aliases are the documented client contract; expect {auth_url, authUrl}.
- liveWritePolicy (1): simulate invalidates governanceRunsAll() prefix (same commit updated hook but not table); fix table key.
- useOwnership (1, prod): useSimulateGovernance replayed the plan-recording POST through resilientFetch on failure; fail closed with skipAuthRefresh (guardWriteOptions precedent).
- useSessions (1): test never tracked isRefetching so the in-flight render was skipped by RQv5 tracked-props optimization; spread result during render.
- useVehicleVinFilter (1): LEFT RED by design (see report): d2cb639 deliberately removed stale-VIN cleanup in favor of retain + scopeReady gating consumed by both charging pages.
…checkpoint regen - Extend i18n-trio-checkpoint --checkpoint-catalog to the two forgotten frozen-projection suites (copy-owner-contracts, final-four-gaps) with per-group shapes; original five groups byte-identical (no-op rewrite). - Regen pins via script after --typed verification (42199 keys, 0 missing, ar/he preserved): copy-owner 30901/32945, final-four 32929/32933. - Hoist pairs/intervening/additions/acceptedLater/leaves/physical fixtures to top level (same values) so the script reads what the tests assert. - copy-owner: HistoryPrompt contract 2->3 calls (same per-call fallback + execution checks); gating test migrated to reviewed noVehicle-first + unresolved precedence (page-test oracle), filter precedence kept. - RegexTester: Title-Case placeholder fallback aligned to catalog casing via --align-typography (1 line). - shell-preservation 7 left red: byte-pins predate trio+sidenav with no covering generator; hand-editing hashes is forbidden. Blocked, exact errors recorded for follow-up.
…ys (extended generator)
The 15 investigation sections are React.lazy; first navigation to a section pays a dynamic-import cost that can exceed the 1s findBy timeout under full-suite load, flaking whichever test touches a section first (close suite: 'hub exposes source caps...' red in-suite, 67/67 solo). Warm all chunks once in beforeAll with a glob-drift guard (count must equal features.length). No assertion changed.
…t, live-mode + thermal pins
fix-system-search-ss10 (announcer plurals): add a11y.status.loadedCount _one/_other leaves with {{count}} (governed convention); useStatusAnnouncer + RecommendationDetails pass defaultValue_one/other; new singular/plural unit test; 3 stale '1 items' expectations updated; FirmwareImpactPage i18n mock made plural-aware. Count-default table row moved to a dedicated singular/plural test (same coverage, stronger).
fix-settings-active-sessions-as-05: align hook keys settings.sessions.* -> settingsSessions.* (catalog namespace); add 3 missing leaves with exact fallback wording; plural toast passes raw count + fmtInt displayCount + defaultValue_one/other; new sessionsRevokePlural.test.ts resolves 0/1/3/1234 in canonical + generated bundles.
th06: pin requiresLiveMode on charging-history refresh. ctt-s01: thermal-adapter regression (endpoint ts wins, no AC/DC double count, kWh->Wh, heater unknown).
Catalog regen + --check clean (32954 leaves); trio checkpoints + announcer producer checkpoint re-pinned via generators. Batch suites 188/188; i18n folder back to the 7 pre-existing shell-preservation reds; tsc + eslint clean.
as-02: absent session lastActive renders localized 'No verified activity' instead of 'Never' (key swap + stale pin flip + absence assertion). privacy-02: policy-aware consent details via describeConsent(state, t, requireConsent); preview copy when no gate (declined/unknown), accepted unchanged; panel + brief pass the flag; 4 meta + 1 panel regressions. tl08: localized TariffStructure labels (badge + select, raw enum preserved as value/payload) + localized new-band default name; 3 regressions incl. raw-enum submit pin. sd-12: preference-aware fmtInt counts in diff breakdown sublabels + 1000-row grouping regression; local formatSpan retained (no shared helper reproduces pinned 45s/1m 5s/1h outputs; item parity condition unmet). tu09: observedBuckets _one/_other leaves + defaultValue_one/other; plural-aware test mock; singular/plural dialog regression. ss-011: 5 StatusHero headlines via statusHero.headline.* keys (variants + overrides preserved); key-echo wiring suite (6 tests). Catalog regen + --check clean (32969 leaves, +15); trio + producer checkpoints re-pinned via generators. Batch suites 199/199 + trio 7672/0; tsc/eslint clean.
… delete-failure localization (catalog 32971)
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Description
Closes #
Type of Change
Checklist
Screenshots (if applicable)