# Optimized draw of `model`, a part of `actor`, from the view of `camera`
	#
	# This drawing should only produce usable depth data. The default behavior,
	# uses `shadow_depth_program`.
	protected fun draw_depth(actor: Actor, model: LeafModel, camera: Camera)
	do
		var program = app.shadow_depth_program
		program.use
		program.mvp.uniform camera.mvp_matrix

		var mesh = model.mesh

		program.translation.uniform(actor.center.x, actor.center.y, actor.center.z, 0.0)
		program.scale.uniform actor.scale
		program.use_map_diffuse.uniform false

		program.tex_coord.array_enabled = true
		program.tex_coord.array(mesh.texture_coords, 2)

		program.coord.array_enabled = true
		program.coord.array(mesh.vertices, 3)

		program.rotation.uniform new Matrix.gamnit_euler_rotation(actor.pitch, actor.yaw, actor.roll)

		if mesh.indices.is_empty then
			glDrawArrays(mesh.draw_mode, 0, mesh.vertices.length/3)
		else
			glDrawElements(mesh.draw_mode, mesh.indices.length, gl_UNSIGNED_SHORT, mesh.indices_c.native_array)
		end
	end