Garage vehicles — 2026-10-30
Menu configuration verification
The admin menu now exposes nearby count (0–24), garage selection chance (0–100%), radius (250–1500 m), and spawn delay (5–60 s). All apply next round; zero count or chance disables creation. Defaults preserve 16 / 70% / 1000 m / 5 s.
The native menu-settings fixture persisted 4 / 100% / 750 m / 8 s through the production settings owner. It verified those runtime values and the eight-second creation cadence, then saved 250 m during the round and confirmed the active radius remained 750 m. Travel, used-car retention and unclaimed retirement still passed. Menu source/routing, input bounds, defaults and disabled startup were checked separately.
Current: gradual population near players
The nearby fixture passed discovery and staged spawning in two parts of Everon. It created four vehicles near the first observer, retained a claimed car, waited through the retirement grace using an accelerated fixture clock, removed the three distant untouched vehicles, and created four near the new observer. All native checks completed with zero script errors; installed asset diagnostics remain in the logs.
Discovery, planning, resource loading and creation happen in different updates. The eight actual native creations were at least five seconds apart. The first cold creation took 71 ms locally; subsequent creations took 7–35 ms. These measurements do not promise zero hitches or establish dedicated-server/client FPS. The native totals line has an unsupported %10 formatting field; the report derives spawn timings from the individual spawn-step records instead.
See current native evidence and fixture. Policy tests additionally cover player-entry forwarding, claims, unknown/visible/near observer vetoes, grace resets, shutdown/client guards and prevention of same-frame work bursts.
Earlier one-pass placement comparison (superseded)
The final optional-addon run placed 12 S105 civilian cars and 4 unarmed WCS RX100 motorbikes into random existing small garages on Everon. The car-only comparison placed 16 S105s when the optional bike addon was absent.
Checks after settling: every vehicle remained alive within 0.5 m of its initial position, had fuel and an accessible empty driver seat, passed the native filled collision volume, and stayed at least 150 m from other selected garages. Repeated start/step/stop calls did not add vehicles. The final source matches the recorded production-file hash in bikes/report.json.
The car-only survey found 107 garages in 676 spatial cells, with 728 ms total measured work and a 65 ms maximum step, spread over roughly 70 seconds. This measures survey/spawn work, not ongoing server FPS or network cost.
The fixture bypasses unrelated scenario preparation and supplies clear observer positions. Production observer vetoes, garage-family filtering, bounds, admin routing and settings migration have separate policy tests. The fixture is local Workbench integration, not a multiplayer or manual driving test. Optional dependencies were added only to the isolated test snapshot; the production addon dependency list is unchanged.
Both runs completed their checks with zero script errors. The strict runner still exits nonzero on asset diagnostics: 13 engine/resource errors in the car-only run, and 34 with bikes loaded, including the installed motorbike texture GUID references. These are retained in the logs; this is not a clean-all-logs claim.