perf(ship): gate telemetry emission on listeners and headless mode

Ship._emit_telemetry_data() ran get_euler()+trig every physics tick
for every ship regardless of whether a HUD was watching, wasting work
on AI ships and every headless training instance. Disable
_physics_process outright when headless, and skip emission the rest
of the time unless a signal actually has a listener.

Game/scripts/ship.gd

@@ -83,6 +83,12 @@ func _ready():
 
 	_boundary = get_tree().get_first_node_in_group("arena_boundary")
 
+	# Telemetry emission is HUD-only work; headless (RL/CI) instances never
+	# have one, so skip the per-tick cost entirely rather than relying on
+	# the listener check below to no-op every frame.
+	if DisplayServer.get_name() == "headless":
+		set_physics_process(false)
+
 
 func _apply_team_color() -> void:
 	if not is_inside_tree():
@@ -112,7 +118,18 @@ func set_controller(new_controller: ShipController) -> void:
 
 
 func _physics_process(_delta):
-	_emit_telemetry_data()
+	if _has_telemetry_listeners():
+		_emit_telemetry_data()
+
+
+# Covers a second/AI ship with no HUD watching it - headless mode is already
+# handled by disabling _physics_process entirely in _ready.
+func _has_telemetry_listeners() -> bool:
+	return speed_changed.get_connections().size() > 0 \
+		or altitude_changed.get_connections().size() > 0 \
+		or attitude_changed.get_connections().size() > 0 \
+		or heading_changed.get_connections().size() > 0 \
+		or thrust_changed.get_connections().size() > 0
 
 
 func _integrate_forces(state):

TODO.md

@@ -23,18 +23,9 @@ Bugs found in an adversarial review. None are gameplay- or physics-affecting, so
 - [ ] Reuse a member `ShipAction` in `ship.gd` (controllerless path) and `player_ship_controller.gd` instead of allocating one per physics tick; `ai_ship_controller.gd` already does this correctly.
 - [ ] Delete the duplicate 1 MB texture — `assets/textures/planet_surface.png` and `assets/models/nebula_planet_planet_surface.png` are byte-identical.
 
-## DRY / structure
-
-- [x] `arena_base.tscn` + inherited themes. `arena_01/02/03.tscn` each restate ~40 identical lines (both lights, the reflection probe, goal transforms, ball/ship spawn markers); only sky, ambient, three glow numbers, tint and decoration differ. The `_elevated` variants already prove the inherited-scene pattern works here — `arena_01_elevated.tscn` is 20 lines to `arena_01.tscn`'s 100. Unblocks cheap new arenas.
-- [x] Hoist bot construction into `GameMode` — `match_mode._make_opponent_controller` and `spectate_mode._make_bot` are the same function (same existence check, same three fields, same warning + inert fallback).
-- [x] Hoist score-keeping into `GameMode` — both modes declare `score := {0:0, 1:0}`, a `score_changed` signal, and the identical increment/emit/print block.
-- [x] `HudInstrument extends Control` base for the three HUD widgets: each repeats `const SMOOTHING := 12.0` and the same `1.0 - exp(-SMOOTHING * delta)` lerp/redraw `_process`, and `hud_heading_tape.gd` reaches across to a static angle helper parked on `HudAttitudeIndicator`.
-- [x] `MAIN_MENU_SCENE_PATH` is declared in both `game_mode.gd` and `settings_menu.gd`.
-- [x] Comment `main_menu.gd`'s `DIFFICULTIES` to say the tiers deliberately share `easy.json` and differ only by handicap (superseded once real `medium`/`hard` tiers land — see the AI section above).
-
 ## Performance
 
-- [ ] Gate `Ship._emit_telemetry_data()` — it runs `get_euler()` + trig per ship per physics tick for every ship, including AI ships nobody displays and `--headless` training where no HUD exists. Gate on having signal connections, and `set_physics_process(false)` when headless (`arena_boundary.gd` already does this correctly for its `_process`).
+- [x] Gate `Ship._emit_telemetry_data()` — it runs `get_euler()` + trig per ship per physics tick for every ship, including AI ships nobody displays and `--headless` training where no HUD exists. Gate on having signal connections, and `set_physics_process(false)` when headless (`arena_boundary.gd` already does this correctly for its `_process`).
 - [ ] Stop the HUD instruments redrawing once settled: all five `queue_redraw()` every frame forever, re-recording canvas items with `draw_string` glyph work, and `HUD.tscn` sets `process_mode = 3` so it continues while paused.
 - [ ] Shared per-team materials instead of `Ship._apply_team_color()` allocating a fresh `StandardMaterial3D` and assigning it as `material_override` (and running at least twice per ship — once from `_ready`, once from the `team` setter). Removes the allocation and lets same-team ships batch.
 - [ ] Merge each goal's visuals into one `ArrayMesh` with a hull surface and a net surface — currently 11 `MeshInstance3D`s and 4 materials per goal, ~22 draw calls for a static prop. `arena_boundary.gd:333` already demonstrates the `SurfaceTool` technique in this codebase.