mileztone17

This commit is contained in:
wolf-demon
2026-09-29 21:03:17 +01:00
parent 4dd33c90e1
commit aa3436fd1d
11 changed files with 263 additions and 11 deletions
+28
View File
@@ -0,0 +1,28 @@
# Desktop-parity acceptance
Run this supervised exercise against the exact HTTPS sandbox release using synthetic conversations and accounts. Compare the web workflow with the reviewed desktop baseline `18b983bf402ecdded6430fd40bc4d3320587595a`; this is workflow and safety parity, not a claim that the interfaces are visually identical.
Use the browser and workstation configuration intended for the pilot. Record its full browser version, operating system, viewport, scale factor and the hotel's saved IANA timezone. Keep screenshots and recordings in the restricted acceptance store because UI evidence may contain message text or account details. The operator and independent reviewer must be different people.
## Required scenarios
| Record ID | Exercise | Passing result |
| --- | --- | --- |
| `inbox-core-workflow` | Open a synthetic conversation, create and save a draft, use an approved answer, change status and reload. | The same conversation and saved state remain available, and no action sends mail. |
| `pagination-beyond-500` | Seed more than 500 tenant-scoped conversations and load successive 50-item pages while another conversation is inserted. | Every original record can be reached once without cross-hotel records, duplicates or skipped records. |
| `draft-conversation-guard` | Edit without saving, select another conversation, reject the discard prompt, then accept it. | Rejection retains the draft and selection; acceptance changes conversation and discards only the unsaved edit. |
| `draft-filter-search-guard` | Repeat the reject/accept checks while changing status filters and search text. | Rejection preserves the edit and prior view; acceptance applies the requested view change. |
| `draft-route-history-reload-guard` | With an unsaved edit, try sidebar navigation, browser Back/Forward and reload or tab close. | In-app navigation and browser history require confirmation; reload or close raises the browser's native unsaved-change warning. |
| `timezone-and-dst` | Choose a timezone different from the workstation and inspect inbox, activity, health, mailbox, automation, PMS, payment and team-link times, including values around a daylight-saving transition. | Every operational timestamp follows the saved hotel timezone and represents the same instant consistently. |
| `role-and-control-parity` | Exercise owner and staff accounts with provider writes and automation disabled. | Staff cannot access owner functions, and neither role can bypass feature, review or provider controls. |
| `desktop-layout` | Complete the workflow at the recorded desktop viewport and scale, including keyboard navigation and browser zoom checks agreed for the pilot. | Primary controls remain visible and usable without clipped dialogs, overlapping content or an inaccessible action. |
## Record and validation
Copy `deploy/desktop-acceptance.example.json` into the restricted acceptance store. Replace its release identifiers, environment, workstation details, timestamps and people. Mark each scenario `pass` only after the independent reviewer has checked its evidence. The example is intentionally invalid while scenarios are `not-run`.
```sh
python3 deploy/desktop_acceptance.py /secure/acceptance/desktop-acceptance.json
```
The validator checks structure, release and baseline binding, desktop dimensions, independent review, complete passing scenarios and opaque evidence references. It cannot inspect screenshots or prove browser behavior. Retain the validated record and its checksum, then reference it from `inbox-usability` in the pilot approval record.
+2
View File
@@ -48,3 +48,5 @@ python3 deploy/pilot_approval.py /secure/acceptance/pilot-approval.json
The validator requires the exact Gate B evidence set, or that set plus independently accepted PMS and payment-provider evidence for Gate C. It also verifies that observed concurrency meets the pre-agreed target and that p95 latency and error rate remain within their pre-agreed bounds. Structural validation does not inspect evidence or authorize rollout by itself.
Complete the [incident and rollback exercise](incident-exercise.md) before marking `incident-support` as passed. Its record must use the same release identifiers and target gate as this decision. Reference the retained exercise record and validator output; do not substitute a local automated-test result for the supervised exercise.
Complete the [desktop-parity acceptance exercise](desktop-acceptance.md) before marking `inbox-usability` as passed. Bind it to the same release identifiers and retain its independently reviewed record outside the repository.