VESPERANManual testing
Reliability and operator rehearsals / 44

Operator: reproduce the automated evidence

Existing suites support the manual plan without replacing human observations.

Prerequisites: an isolated developer checkout with requirements-dev.txt installed and Playwright Chromium available. Run tools from the project root. All laboratory output destinations must be new.

The recorded runtime 0.4.3 baseline reports 599 automated tests, 110 browser checks, eight launcher/restart checks and eleven selected quick-laboratory checks passing. These counts are historical verification evidence for that source build; they are not live availability probes or completed human playtests.

Step by step

Tick a step after observing the expected result. Saved on this device only.

01Run the six automated suites

Run .venv/bin/python tools/verify.py

EXPECTED RESULT

All expected suites run and the command exits successfully. Inspect actual failures, skips and test counts rather than accepting an empty run.

02Run native browser journeys

Run .venv/bin/python tools/verify.py --browser after installing Chromium.

EXPECTED RESULT

The report identifies native navigation and the four browser journeys. A transport bridge fallback, if needed, must be labeled with its cookie/origin/CSP verification limits.

03Run real-process persistence checks

Run .venv/bin/python tools_patch/verify_launcher.py

EXPECTED RESULT

The harness uses temporary accounts/saves and reports eight restart checks. Its operator clock or elapsed-time controls are explicit.

04Replay an earned economy chain

Run .venv/bin/python tools_patch/earned_journey.py and .venv/bin/python tools_patch02/earned_material_journey.py in the isolated checkout.

EXPECTED RESULT

These use legitimate player production and exchanges with explicit test identity/clock controls; the exact final state and inventory ledger replay correctly.

05Run the Patch03 quick laboratory

Run .venv/bin/python run_balance.py --out balance_runs/qa-new-01 --quick, using a fresh output directory.

EXPECTED RESULT

The experiment produces source-pinned reports from temporary worlds; scripted producer policies are simulation inputs, not human demand or live autonomous players.

06Review rather than certify automatically

Run .venv/bin/python tools_patch03/review_content.py --evidence balance_runs/qa-new-01/RUN_REPORT.json --out balance_runs/qa-new-01-review.json. Read the balance and editorial findings.

EXPECTED RESULT

Passing structural checks does not certify long-term pacing, economy equilibrium, complete player journeys or enjoyment.

Type two or more characters to search.