נתונים ודיווח
שילוב נתוני ייחוס במוצר
תיעוד שילוב גילוי עמדות ובחירת רכב בממשק.
The previous release exposed national data only at /network; the driver home and vehicle form still used the original eight fictional station records and local vehicle preferences. This change replaces those discovery entry points with the actual reference data and account-backed vehicle saving.
/and/favoritesuse/api/reference/map, which reads every location using bounded database pagination, validates the count/unique IDs and never falls back to the demo list. The public-only response can be cached for five minutes. Filtering by name/address/operator alias, operator and reported fast points affects the list and map together. The list paginates independently; all matching sites are present in the clustered GeoJSON map. Geolocation enables straight-line distance sorting. National favorites are explicitly local to the device./stations/mot-<id>opens the actual source location. Navigation uses its recorded coordinates. Availability, tariffs and connector identities remain unknown, and remote charging stays disabled. The government snapshot is partial and dated March 19, 2026.- The product header opens a global model picker. Its selection routes to
/vehicle. The same picker is available within the driver garage and fleet editor. Queries go to the official catalog with bounded pagination; Latin brand fallback is used only when the original query has no results. /vehiclesupports authenticated plate lookup, explicit model/variant selection, personal workspace provisioning with consent, and saving to an explicitly selected writable personal/business tenant. Plate lookup is informational and never establishes ownership. Saving is a separate user action.- The platform API re-fetches the selected official model ID on save; caller-provided catalog snapshots are rejected.
ev1_fleet_vehicles.catalog_vehiclestores technical catalog facts.model_idis nullable and a database constraint requires exactly one catalog snapshot or existing curated model. No battery, connector or efficiency is fabricated. Existing tenant RLS, immutable scope, same-tenant FKs and audit triggers remain in force. - Other pre-existing driver demo pages remain labelled as demonstrations. This release does not enable provider charging, card payments or settlement.
Validation: production Next build and ESLint pass; 98 unit tests and 28 PostgreSQL tenant/constraint checks pass. Browser checks cover the actual main map, filters/favorites, deep links, unavailable upstream behavior, global Latin-name search, model selection, mobile layouts and existing fleet flows. scripts/check-reference-live.mjs additionally supports LNE_QA_PRODUCT=1 to verify a temporary account's real provision/save/reload/fleet journey and records its fixture IDs in the explicitly provided private file for cleanup.
Rollback baseline: [operational-id-redacted], runtime e0d2a0c. The additive schema remains in place after app rollback. Newly saved catalog vehicles need the new UI to display their names; an app rollback does not delete or rewrite those records.
Released 2026-11-08: product deployment [operational-id-redacted], runtime 9587f46, is READY on https://product.linknodesenergy.com. All 22 relevant local browser scenarios passed; the five main discovery scenarios also passed on the production domain. The authenticated live suite passed 22 checks, including global search, explicit personal workspace provisioning, real catalog vehicle persistence, page reload, and fleet display. The temporary Clerk identity was deleted and its sessions revoked. Its single marked QA tenant, vehicle, policy, membership, test audits and quota were removed in a guarded transaction; the tenant audit trigger was restored and verified enabled. No customer records or provider charging/payment operations were changed. Production desktop and mobile screenshots were inspected.
Final production deployment: [operational-id-redacted], runtime fdb200f. Visual inspection found that the initial mobile flex layout gave the map container a zero-height percentage child. The mobile map now has an explicit 380px height; the regression test checks both the rendered canvas and its clipping container. The stronger test passed at 390px and 320px on the final production deployment, and a fresh mobile screenshot confirms the basemap and national clusters are visible. Backend behavior is unchanged from the authenticated verification above.