Retain aircraft through native landing and qualify bounded input diagnostics

2026-09-19 20:34:18 +02:00
parent f870443b9b
commit 2fba12f987
129 changed files with 50369 additions and 88 deletions
@@ -49,3 +49,24 @@ The first two native samples meeting the measured≤3° facing criterion were345
 A four-turn/five-second facing target is an optimization hypothesis, not an established runtime bound: this run's alignment started near real64.1s and reached its qualifying facing pair at69.768s. Keep all native response/settling requirements. Later genuine held-F work is conditional on independently measured remaining time, as described in [the held-action design](prison-held-action-design.md).

 This audit inspected installed files and preserved logs only. No display connection, button/key request, native launch, input-setting change or executable-source edit was made.
+
+
+## Implemented selection experiment, 19 September 2026
+
+The explicit `--exercise-prison-selection` runner path now requires the prior
+private input/look/approach/alignment and native-choice options. It first validates
+actual facing with disabled Open selected, then permits one logical button5 pair.
+It must observe the same native lock ID selected and bound in two fresh samples.
+No F or action completion is available. Independent validators re-read all raw
+intermediate observations and retain the original8s facing/selection and45s
+observer deadlines. The helper retains uncertain-button and paired-release
+records; the runner binds its exact two sent events and source hashes.
+
+Offline policy, helper and raw-evidence regression checks pass, with independent
+review. In-game selection remains unproven. The201104 trial failed its prerequisite
+camera return: approximately0.3915degrees rather than the allowed0.35, then timed
+out under the original4s limit. One unchanged repeat201352 refused original
+window/process identity before input. Both failures and verified cleanup are
+preserved in the acceptance ledger. Neither reached the wheel stage; do not infer
+a button mapping or selected-lock result from the simulated tests. No further
+unchanged repeat is justified by these two observations alone.