VEYNDO
Handbook
docs/evidence/configuration-summary-hosted-2026-10-16.mdView source ↗

English source · 4 min

Configuration details throughout purchase — hosted 16 October 2026

This follow-up closes the presentation gap discovered in the hardening acceptance. Exact build, test scope and capture hashes.

Final merged main publication

Final Core c70f246, image mirthful-zipper-2487, is online with one healthy instance and zero restarts. Its normal protected main workflow built the exact source (9m24s), took a fresh PostgreSQL snapshot (1m09s), ran the migration job successfully (22s) and waited for deployment readiness (1m35s). Separate system/tenant database connections and migration 103 remain enforced.

PRs 143, 144 and 147 passed their required checks and merged normally. PR 147 adds translated, explicit recovery for missing lazy modules; 618 frontend tests pass with the unchanged 100% workspace-boundary coverage gate. The first failed coverage run is retained and was fixed with an actual reload-click regression.

A public general-offline navigation produced the separate API/workspace message; returning online and explicitly reloading recovered the same signed-in shop and customer view. The attempted browser asset-blocking rule did not reject a lazy module, so this is not a public acceptance of the module-specific message. Actual React.lazy rejection and recovery-click controls pass in the frontend suite. Browser throttling was removed, blocking disabled and DevTools closed.

Customer view recovered on the final merged main

Configuration purchase on its observed image

Core image timely-zipper-1837, source 12b2a6b, started with one healthy instance, zero restarts and a successful public health response. No additional migration was needed. The PR head 4c1ee04 differs only in merged documentation.

A fresh public simulated order RAC-75415aa4 shows Eiche, 260 × 80 cm and EUR 48 surcharge, EUR 397 total, in native checkout, confirmation and merchant order details. The cart is empty afterwards. These are synthetic data and simulated payment; they do not establish provider settlement, invoice or inbox delivery.

Public native checkout retains material and dimensions

Saved public confirmation retains the accepted configuration

Native merchant order shows the same frozen configuration

The server freezes only declared public input labels/values, public material names, localized choices and the shop main-language fallback. Later app or library edits cannot rename a completed purchase. New reorder quotes use current rules and fresh labels. Existing purchases remain readable without fetching a newer package. One shared display handles checkout, confirmation, native order and app purchase history; no new app business rules, tables or authorization path were introduced.

Local actual PostgreSQL tests additionally cover independent B2B approval and current-rule reorder, changed material/field names and original-order immutability. All 612 frontend tests, two Rust regressions, workspace Clippy, structure/localization, frontend build and explicit formal source review pass. The full required follow-up GitHub repeat passed and PR 144 merged as deb452d. Public browser outcomes remain separately recorded above.

Original Storyfront and release-queue boundary

The earlier native Safari comparison completed original bag → embedded contact/address → review → native confirmation RAC-a4727b33, EUR 24.90, German product name, simulated authorization → return to original Experience → verified empty original bag. The iframe also loads in the in-app browser. A tab-bound Chrome repeat remained at about:blank despite a public HTTP 200 document with the exact parent framing policy; that controlled repeat is not claimed passed or silently substituted with direct checkout.

To distinguish that control path, a fresh tab in the same Chrome browser/profile was operated through native UI, without claiming it through tab debugging during load. The original embedded journey passed: RAC-0c3c77e8, EUR 24.90, German confirmation, return to Experience, original bag 0 and verified completion feedback. Only after the purchase was the tab attached for content-only screenshots. No server code, credential or browser security/framing setting changed between cases.

Original embedded confirmation completed in native Chrome

Original Chrome Storyfront bag after the confirmed purchase

Repeated original embedded confirmation in Safari

Original Storyfront bag after verified purchase

Final pinned-image repeat

The separate Northflank main-release workflow temporarily replaced the manual branch deployment while those comparisons ran. The queued main releases were allowed to finish, then timely-zipper-1837 was restored and the running source was observed again (vendune-core-75867998d-mhtcg, 1/1 passing). This distinction prevents the earlier comparisons from being silently attributed to another image.

On that pinned image, native Chrome completed RAC-08a6f8d0, EUR 24.90, German confirmation inside the original iframe, then returned to the original Storyfront with bag 0 and the verified completion feedback.

Final embedded confirmation on the observed pinned image

Original bag cleared after the final pinned-image purchase

Native Admin creation

On the restored pinned image, Chrome created a synthetic CRM-only profile and standard address, searched and selected it for a manual order, selected German order language, added one demo lamp, reviewed the native quote and saved RAC-41ad6146, EUR 59. The customer received no password or shopper session; no external payment provider was called.

Actual native manual order after customer lookup and quote review