Add mission customization, HQ support and aggression-driven AI

Integrate faction and random settings menus, generated objectives, scouts, convoys, air support, mortar stations and body-intel recovery. Add identified-body aggression, finite hunter UAZ deployment, dedicated objective defence and maximum AI skill with separate reduced aim accuracy. Include policy tests, native probes, evidence and documentation.
2026-09-23 15:07:14 +02:00
parent 04bd19327b
commit 1b1a8b5173
242 changed files with 12153 additions and 875 deletions
@@ -0,0 +1,85 @@
+# Generated starts and radio camps
+
+Everon now chooses a new prison, radio camp and four mortar sites at preparation time. Other caches,
+outposts retain their authored positions. The extraction helicopter uses the validated
+Everon airfield pad (4800.953 / 11487.412), with no off-airfield fallback. The server
+resolves the prison first, then the radio, before creating any site entities.
+Joining clients receive the real replicated locations; intel keeps the same
+stable destination IDs.
+
+Site authors can use `RandomClear 1` on a prison, radio or mortar definition. `Heading`
+sets the authored fallback heading; `RootHeight` offsets a composition above its
+surveyed ground (the radio/table composition uses 0.8 m). Runtime placement never
+rewrites the configuration file. Existing definitions default to authored sites.
+
+Placement samples the island bounds and rejects water, obstructed footprints,
+terrain varying by more than 0.6 m, nearby roads (12 m), and unsafe guard footing.
+Prisons avoid physical runway/control-tower anchors by 1,500 m. The radio stays
+at least 1 km from the prison and 300 m from other sites. The extraction pickup
+must remain inside the prison-relative escape boundary with 100 m clearance.
+Sampling yields after an 8 ms work budget, with a 30-second overall deadline and
+12,000 candidate limit per site. The budget is checked between native terrain
+queries, so a single engine query can exceed it. Authored fallback positions
+must pass the same checks; otherwise preparation reports failure instead of
+spawning an incomplete camp or waiting forever.
+
+The existing prison composition, gate, backpack and guards move together.
+Player slots and escape-boundary tests use the same rotation. Admission still
+waits for physical guards, inventory and the client's nearby replicated scene.
+Spawn clearance uses a filled standing-volume trace for scenery and actual
+character positions for body spacing, avoiding false overlaps from rotated
+entity bounding boxes. Pending reservations remain part of the clearance check.
+The radio camp has a radio and table inside a full-size tent, with an antenna
+and sandbag cover outside. Its complete guard squad permanently defends a 20 m
+area around the radio and cannot be reassigned by virtual HQ. See
+[radio camp security](radio-camp-security.md).
+It retains the existing contact-extraction-HQ action and defence progression.
+
+Partial intel displays a question mark at the centre of the containing 500 m
+square. Exact identification changes it to a site icon at the real location.
+The prison is always hidden. Notifications use the original Intel wording.
+
+`/esc killais` is an administrator-only debugging command centred on the
+administrator's current living prisoner. It applies lethal damage to AI
+characters within 1,000 m in three dimensions, including friendly AI and seated
+crews. Player-controlled and retained player-roster bodies are excluded, as are
+vehicles and structures. Death callbacks run normally; the command does not
+refill AI forces or manufacture HQ reports. `/esc help killais` explains it.
+
+Four mortar sites each have a physical gun, four crew, finite shells and
+targeting through the HQ radio-report chain. Sites are at least 1 km apart and
+500 m from the prison. See [mortar intel and perimeter guards](mortar-intel-and-prison-map.md).
+
+## Validation
+
+`test_generated_sites.py` executes the actual bounded-search scheduler;
+`test_debug_ai.py` executes kill eligibility. Existing player-tool, admission,
+prison-slot and intel-feedback tests cover the changed boundaries.
+`--project core --smoke generated-sites` opts into real randomized placement and
+tests admission, hidden/identified markers, authenticated killais (including a
+seated AI), and the radio objective. The `opening-tools` and `airfield-markers` outcomes also use the live site
+definitions. Other Workbench smoke outcomes explicitly load
+`Configs/ERR/Tests/LegacySites.conf` to retain their historical deterministic
+fixtures, including the retired western cache and old extraction position.
+Ordinary Workbench play and dedicated servers use the live definitions.
+
+On 2026-09-22, the full `python3 tools/check.py` suite passed. The final native
+`generated-sites` fixture passed 39 assertions at a randomized Everon start:
+prison at approximately (5100, 9137), heading 89 degrees; radio at (5991, 5659).
+It placed twelve physical prisoners inside the rotated prison, verified that
+occupied body/wall positions still reject spawning, and checked admission with
+stocked backpack, hidden partial-intel markers, exact
+discovered coordinates, authenticated AI killing (including a seated passenger),
+and the existing radio transmission objective. The runner confirmed unchanged
+source inputs and complete process cleanup. Native diagnostics still include
+external RHS/WCS localization and stock resource-keyword errors; the fixture had
+no script/runtime failure. A captured camp view was inspected for radio/table,
+antenna and shelter placement.
+
+Local scripted fixtures do not establish ordinary input, remote-client
+replication, or complete multiplayer acceptance. Dedicated-server validation is
+still outstanding; this change has not been published or deployed.
+
+The western guarded vehicle/cache site has been removed. The medical stash now
+provides the second reachable source of radio intel; mortar logistics uses the
+existing eastern outpost. See [airfield and map changes](airfield-extraction-and-map-icons.md).