Start source-bound P23-03 acceptance audit after Opus sequencing decision

2026-10-15 07:36:46 +09:00
parent a2e739396b
commit e01e04c62d
6 changed files with 112 additions and 5 deletions
@@ -0,0 +1,62 @@
+{
+  "plan_id": "P23-03",
+  "state": "in_progress",
+  "stage": "requirements/evidence audit; no implementation",
+  "source_head": "a2e739396bb3a0064cbb299d379f43161bab5a9c",
+  "original_acceptance": "人物や設備の表示が実状態に対応し、同需要の動線改善を観察できる。",
+  "work": "部屋、行列、EV、職員、物資、営業状態が断面上で理解できる表示へ調整する。",
+  "denominator": {
+    "packages": 30,
+    "units": 88,
+    "credit": 0,
+    "partial_credit": false
+  },
+  "dependencies": {
+    "package_declared": [
+      "P08",
+      "P22"
+    ],
+    "current_reused_support": [
+      "P08-02 completed",
+      "P08-03 completed",
+      "existing WorldPreview draw/read paths"
+    ],
+    "to_resolve": "Map package dependencies to the exact display/route/save inputs before registering bounded contract. Do not assume unfinished P08-01/P22 work is delivered or silently omit required facility states."
+  },
+  "conditions_to_evidence": [
+    "All six requested display categories correspond to actual typed Runtime state, including occupancy vs queues, EV position/load/direction, staff job/position and carried goods, operating/closed/blackout/raid pause where applicable.",
+    "Same demand and fixed inputs with only layout route improvement changed; visible travel/queue change matches actual Runtime measurement. No invented production constants.",
+    "Actual ordinary OS observations, formal save and separate-process resume retain correct state/display mapping.",
+    "Exact final source/settings/environment evidence, mandatory full regression, independent Opus review, integration/final and ledger."
+  ],
+  "source_pins": {
+    "scenes/main/world_preview.gd": "07e9a36baaa5d0904693f968164101a3938248cfd2cb65b600b64d10eca961d8",
+    "scripts/ui/pixel_tower_art.gd": "71df9b8400cb883a5bb1c89729f80c77e734df503fc54154be3e486adb8391df",
+    "tests/ui/test_guest_observation.gd": "8f5af3d6ef367512b34d0cf35fd8c1d2848d020419dddfa234120206dbe4f6e4",
+    "tests/ui/test_pixel_tower.gd": "f149f4b5507247dadf623e777e90b9bdb81f54324b406a8f55adaf89b649866a",
+    "AGENTS.md": "8d8dc40211390275f858f322743309a0663c4f2d859780a8b8a08cea1450e8b3",
+    "CLAUDE.md": "336cc4fbf19beaada7ccf9986414fa91851a8d7a07dfb3ccbe800a69eed0ab49"
+  },
+  "existing_observations": [
+    "WorldPreview._draw_world_entity_icons uses actual facilities, room presentations and facility color/status.",
+    "_prepare_queues derives actual queue entries and elevator queued trips, hides overflow with count badge.",
+    "_draw_elevators uses transport report shafts/stops/boarded.",
+    "_draw_actors derives staff.current_cell/state/carried_inventory_id and room/rescue/queue placements.",
+    "Guest observation test covers live state/nonmutation/restore but not all six cross-section categories or same-demand layout comparison.",
+    "Pixel tower test covers atlas/camera/hit testing and simulation isolation; not whole P23-03 acceptance."
+  ],
+  "reuse_policy": "Reuse implementations, collection methods and matching source evidence only. No old source suite counts or P23-02 screenshots relabelled as current P23-03 acceptance. Current full176 is baseline, not future changed-source proof.",
+  "remaining": [
+    "Finish real-state/display mapping and same-demand comparison fixture contract.",
+    "Confirm only concrete required source changes; no Runtime/service/data-value changes without separately justified scope.",
+    "Register exact immutable contract/validation, existing unlimited individual repair policy; Opus start decision then one assigned worker.",
+    "Static/local/impact checks, then Root native+OS and one mandatory exact-final full run, independent review and closure."
+  ],
+  "deferred_global": [
+    "Windows same-distribution physical acceptance",
+    "P23-02 held Ctrl-wheel/left-drag human gestures",
+    "P25 finished asset/full motion requirements"
+  ],
+  "new_worker_started": false,
+  "updated_at_jst": "2026-10-15T07:36:34.134709+09:00"
+}
@@ -0,0 +1,11 @@
+{
+  "run": "71ce1c9e91e34df3aae826e3a86e8447",
+  "decision": {
+    "action": "wait",
+    "task_id": "P08-03",
+    "reason": "P08-03 is finished: the Runner shows it completed (submitted d369d1c, integrated as 027861b on main a025d77), with validation, review, integration and final checks all passing. The only valid action is wait. No dispatch, no new task and no duplicate worker.\n\n**Recommended next unit: P23-03 「生活の観察」** (proposal only; the owner decides whether to register it).\n- Work: make rooms, queues, EVs, staff, goods and business state readable on the cross-section.\n- Acceptance (all of it, nothing partial): 「人物や設備の表示が実状態に対応し、同需要の動線改善を観察できる。」\n- It stays one existing unit of 30 packages/88 units, with no partial credit.\n\n**Why this one.** It is a display-and-state check. It needs no new costs, staff counts, mana or other numbers, no P08-01 decisions (VIP high floor, exterior cannon) and no Windows or human gestures in the unit's own acceptance. Windows stays a global platform/release item, as it did for P08-03. The other candidates are blocked by unknowns or missing pieces:\n- **P12-02:** linen chute and hidden door are in the original (BASELINE:49) but have no definitions in facility_catalog.gd; only staff_door exists.\n- **P07-01/P07-03:** weather and trend effect sizes and timing are not decided.\n- **P14-03 and P06-02:** both need warp, which P14-02 has not delivered.\n- **P06-02 also:** it depends on the mana decisions in P06-01.\n- **P09-03, P16-01, P05-03:** they need level, hiring and bank values that are not decided (A04/D04).\n- **P20-01/P20-02:** they depend on D08.\n- **P11-03:** casino and arena are not defined.\n\n**Existing evidence to inspect and reuse, only after checking it applies to the exact current source:**\n- scenes/main/world_preview.gd draw paths: _draw_congestion_overlay :653, _draw_world_entity_icons :773, _draw_readable_badge :803, _draw_elevators :1041, _draw_actors :1133, _draw_pixel_actor :1283, _draw_facility_labels :1012\n- scripts/ui/pixel_tower_art.gd\n- tests/ui/test_guest_observation.gd and tests/ui/test_pixel_tower.gd\n- P08-03's integrated OS screenshot/input-receipt method and checker pattern (tools/run_p08_normal_placement.py, check_p08_normal_placement_evidence.py)\n- P08-02's mixed_world fixture\n- Path and congestion services, plus current full-regression results\n- Old 1273/168 results and P23-02 candidates are not to be carried over as PASS.\n\n**Missing conditions:**\n- (a) A mapping from each displayed element (room occupancy/business state, queue length, EV position/load/direction, staff job/position, goods carried, closed/blackout/raid pause) to the real typed Runtime state. It needs a checker that recomputes this from whole states, not from boolean claims.\n- (b) A same-demand comparison: two layouts with the same fixed seed and arrivals, one with the route improvement. Measured travel/queue differences must be visible on screen and match the Runtime numbers. No new production constants.\n- (c) Real OS screenshots and normal-UI observation, plus formal save and resume in a separate process, with the display still matching state.\n- (d) Clear deferral boundaries: P25-01/P25-02 own finished assets and full motions; P23-02 owns x3 performance and Windows scaling. The old 1.09 s stall must not be passed off as solved.\n\n**Bounded scope:**\n- Changes limited to the Main/WorldPreview/pixel_tower_art display layer and read-only Runtime queries. No Runtime, service or data value changes.\n- Fixtures under tests/fixtures/p23-03, one new strict marker, a Root driver and checker with meaningful negatives (state/display mismatch, stale source, fake OS provenance, unchanged comparison), and docs.\n- Worker does static and Python checks only.\n- Root runs native and OS collection, then one final full regression, an independent Opus review of the whole condition set, integration, final verification and the ledger.\n\n**Before registering:** the owner should confirm allowed_paths and the fixture choice.",
+    "codex_model": "gpt-6.1-sol",
+    "codex_model_reason": "No worker is being sent now (wait). If P23-03 is registered later, the default gpt-6.1-sol fits: it is display wiring over existing draw paths plus a checker and fixtures, with no Runtime or domain changes or cross-system debugging. That doesn't justify Astra's extra time and usage. Re-check this once the scope is set."
+  },
+  "root_action": "Read-only acceptance preflight only; no worker or task dispatched. Existing user authorization covers necessary next-unit contract preparation; no additional permission requested."
+}
@@ -0,0 +1,21 @@
+{
+  "measured_at_jst": "2026-10-15T07:35:34.949219+09:00",
+  "free_bytes": 5319131136,
+  "swapusage": "vm.swapusage: total = 11264.00M  used = 9749.31M  free = 1514.69M  (encrypted)",
+  "closed_profile_recovery_bytes": 1787904000,
+  "historical_20GB_exact_attribution": "Unavailable: no per-directory snapshot immediately after app removal.",
+  "measured_causes": [
+    "Repeated native snapshots/raw state, old release copies and Runner review/build copies (see storage-20261014/FOLLOWUP_CAUSE_MEASUREMENTS.json)",
+    "This closed full176 run produced ~1.94GB logical private profile bytes, now losslessly compressed to ~157MB allocated",
+    "macOS swap currently allocates11GiB; not included in artifact sums"
+  ],
+  "measurement_limits": [
+    "APFS du/shared blocks are not additive",
+    "Free space changes while OS and other apps run"
+  ],
+  "preserved": [
+    "all raw accepted evidence",
+    "source/spec/assets/user saves",
+    "active processes and installer data"
+  ]
+}