O8C Report 03 — Baseline test / smoke result (G2)

• macro: v0.6-o8c-real-db-rollback-proof-readiness
• date_utc: 2026-05-21 · host: vmi3080463 (Contabo) · Python 3.12.3
• gate covered: G2 baseline tests / smoke
• result: G2 PASS

1. py_compile — deployed tree

python3 -m py_compile <every .py in the promoted tree>
-> OK (67 source files compiled, 0 errors)

2. Full unittest suite — deployed tree

cd /opt/incomex/dot/iu-cutter-v0.6
python3 -m unittest discover -s tests
-> Ran 394 tests in 3.825s
-> OK            (exit 0, 0 failures, 0 errors)

3. Targeted live-wiring suite

python3 -m unittest tests.test_orchestrator_o8a_live_wiring -v
-> Ran 28 tests in 0.011s
-> OK            (exit 0)

4. Mode.LIVE kill-switch + no-simulator-fallback (deployed tree)

Re-proven first-hand against the deployed tree by the O8C proof harness (Report 05); summarised here:

kill_switch_off_refuses_mode_live:   YES — all 6 ProductionLiveExecutionAdapter
                                     methods raise ProductionExecutionNotAuthorized
                                     BEFORE any DB connection is opened (PROOF-2)
mode_live_no_simulator_fallback:     YES — require_live_adapter accepts only a
                                     real LiveExecutionAdapter; refuses a
                                     simulator-shaped object / None / str / int
                                     (PROOF-3)
mode_dryrun_simulator_intact:        YES — unchanged; 394/394 covers it

5. G2 verdict

py_compile:            OK (67 files)
full_suite:            394 / 394 PASS  (exit 0)
targeted_live_wiring:  28 / 28 PASS
mode_live_killswitch:  refused when OFF (proven)
simulator_fallback:    impossible by construction (proven)
baseline:              matches O8B-retry Report 07 (394/394) — no regression
g2:                    PASS