update rbx_loader

make invalid states unrepresentable!!!

.cargo/config.toml

@@ -0,0 +1,2 @@
+[registries.strafesnet]
+index = "sparse+https://git.itzana.me/api/packages/strafesnet/cargo/"

Cargo.toml

@@ -1,23 +1,36 @@
 [package]
 name = "strafe-client"
-version = "0.2.0"
+version = "0.10.5"
 edition = "2021"
+repository = "https://git.itzana.me/StrafesNET/strafe-client"
+license = "Custom"
+description = "StrafesNET game client for bhop and surf."
+authors = ["Rhys Lloyd <krakow20@gmail.com>"]
 
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+[features]
+default = ["snf"]
+snf = ["dep:strafesnet_snf"]
+source = ["dep:strafesnet_deferred_loader", "dep:strafesnet_bsp_loader"]
+roblox = ["dep:strafesnet_deferred_loader", "dep:strafesnet_rbx_loader"]
 
 [dependencies]
-async-executor = "1.5.1"
 bytemuck = { version = "1.13.1", features = ["derive"] }
+configparser = "3.0.2"
 ddsfile = "0.5.1"
-env_logger = "0.10.0"
-glam = "0.24.1"
-log = "0.4.20"
-obj = "0.10.2"
+glam = "0.29.0"
+id = { version = "0.1.0", registry = "strafesnet" }
+parking_lot = "0.12.1"
 pollster = "0.3.0"
-wgpu = "0.17.0"
-winit = "0.28.6"
+strafesnet_bsp_loader = { version = "0.2.1", registry = "strafesnet", optional = true }
+strafesnet_common = { version = "0.5.2", registry = "strafesnet" }
+strafesnet_deferred_loader = { version = "0.4.0", features = ["legacy"], registry = "strafesnet", optional = true }
+strafesnet_rbx_loader = { version = "0.5.1", registry = "strafesnet", optional = true }
+strafesnet_snf = { version = "0.2.0", registry = "strafesnet", optional = true }
+wgpu = "22.1.0"
+winit = "0.30.5"
 
 [profile.release]
-lto = true
+#lto = true
 strip = true
 codegen-units = 1

LICENSE

@@ -1,5 +1,5 @@
 /*******************************************************
-* Copyright (C) 2023 Rhys Lloyd <krakow20@gmail.com>
+* Copyright (C) 2023-2024 Rhys Lloyd <krakow20@gmail.com>
 *
 * This file is part of the StrafesNET bhop/surf client.
 *

src/compat_worker.rs

@@ -0,0 +1,21 @@
+pub type QNWorker<'a,Task>=CompatNWorker<'a,Task>;
+pub type INWorker<'a,Task>=CompatNWorker<'a,Task>;
+
+pub struct CompatNWorker<'a,Task>{
+	data:std::marker::PhantomData<Task>,
+	f:Box<dyn FnMut(Task)+Send+'a>,
+}
+
+impl<'a,Task> CompatNWorker<'a,Task>{
+	pub fn new(f:impl FnMut(Task)+Send+'a)->CompatNWorker<'a,Task>{
+		Self{
+			data:std::marker::PhantomData,
+			f:Box::new(f),
+		}
+	}
+
+	pub fn send(&mut self,task:Task)->Result<(),()>{
+		(self.f)(task);
+		Ok(())
+	}
+}

src/face_crawler.rs

@@ -0,0 +1,127 @@
+use crate::physics::Body;
+use crate::model_physics::{GigaTime,FEV,MeshQuery,DirectedEdge,MinkowskiMesh,MinkowskiFace,MinkowskiDirectedEdge,MinkowskiVert};
+use strafesnet_common::integer::{Time,Fixed,Ratio};
+
+#[derive(Debug)]
+enum Transition<F,E:DirectedEdge,V>{
+	Miss,
+	Next(FEV<F,E,V>,GigaTime),
+	Hit(F,GigaTime),
+}
+
+type MinkowskiFEV=FEV<MinkowskiFace,MinkowskiDirectedEdge,MinkowskiVert>;
+type MinkowskiTransition=Transition<MinkowskiFace,MinkowskiDirectedEdge,MinkowskiVert>;
+
+	fn next_transition(fev:&MinkowskiFEV,body_time:GigaTime,mesh:&MinkowskiMesh,body:&Body,mut best_time:GigaTime)->MinkowskiTransition{
+		//conflicting derivative means it crosses in the wrong direction.
+		//if the transition time is equal to an already tested transition, do not replace the current best.
+		let mut best_transition=MinkowskiTransition::Miss;
+		match fev{
+			&MinkowskiFEV::Face(face_id)=>{
+				//test own face collision time, ignoring roots with zero or conflicting derivative
+				//n=face.normal d=face.dot
+				//n.a t^2+n.v t+n.p-d==0
+				let (n,d)=mesh.face_nd(face_id);
+				//TODO: use higher precision d value?
+				//use the mesh transform translation instead of baking it into the d value.
+				for dt in Fixed::<4,128>::zeroes2((n.dot(body.position)-d)*2,n.dot(body.velocity)*2,n.dot(body.acceleration)){
+					if body_time.le_ratio(dt)&&dt.lt_ratio(best_time)&&n.dot(body.extrapolated_velocity_ratio_dt(dt)).is_negative(){
+						best_time=dt;
+						best_transition=MinkowskiTransition::Hit(face_id,dt);
+						break;
+					}
+				}
+				//test each edge collision time, ignoring roots with zero or conflicting derivative
+				for &directed_edge_id in mesh.face_edges(face_id).iter(){
+					let edge_n=mesh.directed_edge_n(directed_edge_id);
+					let n=n.cross(edge_n);
+					let verts=mesh.edge_verts(directed_edge_id.as_undirected());
+					//WARNING: d is moved out of the *2 block because of adding two vertices!
+					//WARNING: precision is swept under the rug!
+					for dt in Fixed::<4,128>::zeroes2(n.dot(body.position*2-(mesh.vert(verts[0])+mesh.vert(verts[1]))).fix_4(),n.dot(body.velocity).fix_4()*2,n.dot(body.acceleration).fix_4()){
+						if body_time.le_ratio(dt)&&dt.lt_ratio(best_time)&&n.dot(body.extrapolated_velocity_ratio_dt(dt)).is_negative(){
+							best_time=dt;
+							best_transition=MinkowskiTransition::Next(MinkowskiFEV::Edge(directed_edge_id.as_undirected()),dt);
+							break;
+						}
+					}
+				}
+				//if none:
+			},
+			&MinkowskiFEV::Edge(edge_id)=>{
+				//test each face collision time, ignoring roots with zero or conflicting derivative
+				let edge_n=mesh.edge_n(edge_id);
+				let edge_verts=mesh.edge_verts(edge_id);
+				let delta_pos=body.position*2-(mesh.vert(edge_verts[0])+mesh.vert(edge_verts[1]));
+				for (i,&edge_face_id) in mesh.edge_faces(edge_id).iter().enumerate(){
+					let face_n=mesh.face_nd(edge_face_id).0;
+					//edge_n gets parity from the order of edge_faces
+					let n=face_n.cross(edge_n)*((i as i64)*2-1);
+					//WARNING yada yada d *2
+					for dt in Fixed::<4,128>::zeroes2(n.dot(delta_pos).fix_4(),n.dot(body.velocity).fix_4()*2,n.dot(body.acceleration).fix_4()){
+						if body_time.le_ratio(dt)&&dt.lt_ratio(best_time)&&n.dot(body.extrapolated_velocity_ratio_dt(dt)).is_negative(){
+							best_time=dt;
+							best_transition=MinkowskiTransition::Next(MinkowskiFEV::Face(edge_face_id),dt);
+							break;
+						}
+					}
+				}
+				//test each vertex collision time, ignoring roots with zero or conflicting derivative
+				for (i,&vert_id) in edge_verts.iter().enumerate(){
+					//vertex normal gets parity from vert index
+					let n=edge_n*(1-2*(i as i64));
+					for dt in Fixed::<2,64>::zeroes2((n.dot(body.position-mesh.vert(vert_id)))*2,n.dot(body.velocity)*2,n.dot(body.acceleration)){
+						if body_time.le_ratio(dt)&&dt.lt_ratio(best_time)&&n.dot(body.extrapolated_velocity_ratio_dt(dt)).is_negative(){
+							let dt=Ratio::new(dt.num.fix_4(),dt.den.fix_4());
+							best_time=dt;
+							best_transition=MinkowskiTransition::Next(MinkowskiFEV::Vert(vert_id),dt);
+							break;
+						}
+					}
+				}
+				//if none:
+			},
+			&MinkowskiFEV::Vert(vert_id)=>{
+				//test each edge collision time, ignoring roots with zero or conflicting derivative
+				for &directed_edge_id in mesh.vert_edges(vert_id).iter(){
+					//edge is directed away from vertex, but we want the dot product to turn out negative
+					let n=-mesh.directed_edge_n(directed_edge_id);
+					for dt in Fixed::<2,64>::zeroes2((n.dot(body.position-mesh.vert(vert_id)))*2,n.dot(body.velocity)*2,n.dot(body.acceleration)){
+						if body_time.le_ratio(dt)&&dt.lt_ratio(best_time)&&n.dot(body.extrapolated_velocity_ratio_dt(dt)).is_negative(){
+							let dt=Ratio::new(dt.num.fix_4(),dt.den.fix_4());
+							best_time=dt;
+							best_transition=MinkowskiTransition::Next(MinkowskiFEV::Edge(directed_edge_id.as_undirected()),dt);
+							break;
+						}
+					}
+				}
+				//if none:
+			},
+		}
+		best_transition
+	}
+pub enum CrawlResult<F,E:DirectedEdge,V>{
+	Miss(FEV<F,E,V>),
+	Hit(F,GigaTime),
+}
+type MinkowskiCrawlResult=CrawlResult<MinkowskiFace,MinkowskiDirectedEdge,MinkowskiVert>;
+pub fn crawl_fev(mut fev:MinkowskiFEV,mesh:&MinkowskiMesh,relative_body:&Body,start_time:Time,time_limit:Time)->MinkowskiCrawlResult{
+	let mut body_time={
+		let r=(start_time-relative_body.time).to_ratio();
+		Ratio::new(r.num.fix_4(),r.den.fix_4())
+	};
+	let time_limit={
+		let r=(time_limit-relative_body.time).to_ratio();
+		Ratio::new(r.num.fix_4(),r.den.fix_4())
+	};
+	for _ in 0..20{
+		match next_transition(&fev,body_time,mesh,relative_body,time_limit){
+			Transition::Miss=>return CrawlResult::Miss(fev),
+			Transition::Next(next_fev,next_time)=>(fev,body_time)=(next_fev,next_time),
+			Transition::Hit(face,time)=>return CrawlResult::Hit(face,time),
+		}
+	}
+	//TODO: fix all bugs
+	//println!("Too many iterations!  Using default behaviour instead of crashing...");
+	CrawlResult::Miss(fev)
+}

src/file.rs

@@ -0,0 +1,144 @@
+use std::io::Read;
+
+#[derive(Debug)]
+pub enum ReadError{
+	#[cfg(feature="roblox")]
+	Roblox(strafesnet_rbx_loader::ReadError),
+	#[cfg(feature="source")]
+	Source(strafesnet_bsp_loader::ReadError),
+	#[cfg(feature="snf")]
+	StrafesNET(strafesnet_snf::Error),
+	#[cfg(feature="snf")]
+	StrafesNETMap(strafesnet_snf::map::Error),
+	Io(std::io::Error),
+	UnknownFileFormat,
+}
+impl std::fmt::Display for ReadError{
+	fn fmt(&self,f:&mut std::fmt::Formatter<'_>)->std::fmt::Result{
+		write!(f,"{self:?}")
+	}
+}
+impl std::error::Error for ReadError{}
+
+pub enum DataStructure{
+	#[cfg(feature="roblox")]
+	Roblox(strafesnet_rbx_loader::Model),
+	#[cfg(feature="source")]
+	Source(strafesnet_bsp_loader::Bsp),
+	#[cfg(feature="snf")]
+	StrafesNET(strafesnet_common::map::CompleteMap),
+}
+
+pub fn read<R:Read+std::io::Seek>(input:R)->Result<DataStructure,ReadError>{
+	let mut buf=std::io::BufReader::new(input);
+	let peek=std::io::BufRead::fill_buf(&mut buf).map_err(ReadError::Io)?;
+	match &peek[0..4]{
+		#[cfg(feature="roblox")]
+		b"<rob"=>Ok(DataStructure::Roblox(strafesnet_rbx_loader::read(buf).map_err(ReadError::Roblox)?)),
+		#[cfg(feature="source")]
+		b"VBSP"=>Ok(DataStructure::Source(strafesnet_bsp_loader::read(buf).map_err(ReadError::Source)?)),
+		#[cfg(feature="snf")]
+		b"SNFM"=>Ok(DataStructure::StrafesNET(
+			strafesnet_snf::read_map(buf).map_err(ReadError::StrafesNET)?
+			.into_complete_map().map_err(ReadError::StrafesNETMap)?
+		)),
+		_=>Err(ReadError::UnknownFileFormat),
+	}
+}
+
+#[derive(Debug)]
+pub enum LoadError{
+	ReadError(ReadError),
+	File(std::io::Error),
+	Io(std::io::Error),
+}
+impl std::fmt::Display for LoadError{
+	fn fmt(&self,f:&mut std::fmt::Formatter<'_>)->std::fmt::Result{
+		write!(f,"{self:?}")
+	}
+}
+impl std::error::Error for LoadError{}
+
+pub fn load<P:AsRef<std::path::Path>>(path:P)->Result<strafesnet_common::map::CompleteMap,LoadError>{
+	//blocking because it's simpler...
+	let file=std::fs::File::open(path).map_err(LoadError::File)?;
+	match read(file).map_err(LoadError::ReadError)?{
+		#[cfg(feature="snf")]
+		DataStructure::StrafesNET(map)=>Ok(map),
+		#[cfg(feature="roblox")]
+		DataStructure::Roblox(model)=>{
+			let mut place=model.into_place();
+			place.run_scripts();
+
+			let mut loader=strafesnet_deferred_loader::roblox_legacy();
+
+			let (texture_loader,mesh_loader)=loader.get_inner_mut();
+
+			let map_step1=strafesnet_rbx_loader::convert(
+				&place,
+				|name|texture_loader.acquire_render_config_id(name),
+				|name|mesh_loader.acquire_mesh_id(name),
+			);
+
+			let meshpart_meshes=mesh_loader.load_meshes().map_err(LoadError::Io)?;
+
+			let map_step2=map_step1.add_meshpart_meshes_and_calculate_attributes(
+				meshpart_meshes.into_iter().map(|(mesh_id,loader_model)|
+					(mesh_id,strafesnet_rbx_loader::data::RobloxMeshBytes::new(loader_model.get()))
+				)
+			);
+
+			let (textures,render_configs)=loader.into_render_configs().map_err(LoadError::Io)?.consume();
+
+			let map=map_step2.add_render_configs_and_textures(
+				render_configs.into_iter(),
+				textures.into_iter().map(|(texture_id,texture)|
+					(texture_id,match texture{
+						strafesnet_deferred_loader::texture::Texture::ImageDDS(data)=>data,
+					})
+				)
+			);
+
+			Ok(map)
+		},
+		#[cfg(feature="source")]
+		DataStructure::Source(bsp)=>{
+			let mut loader=strafesnet_deferred_loader::source_legacy();
+
+			let (texture_loader,mesh_loader)=loader.get_inner_mut();
+
+			let map_step1=strafesnet_bsp_loader::convert(
+				&bsp,
+				|name|texture_loader.acquire_render_config_id(name),
+				|name|mesh_loader.acquire_mesh_id(name),
+			);
+
+			let prop_meshes=mesh_loader.load_meshes(bsp.as_ref());
+
+			let map_step2=map_step1.add_prop_meshes(
+				//the type conflagulator 9000
+				prop_meshes.into_iter().map(|(mesh_id,loader_model)|
+					(mesh_id,strafesnet_bsp_loader::data::ModelData{
+						mdl:strafesnet_bsp_loader::data::MdlData::new(loader_model.mdl.get()),
+						vtx:strafesnet_bsp_loader::data::VtxData::new(loader_model.vtx.get()),
+						vvd:strafesnet_bsp_loader::data::VvdData::new(loader_model.vvd.get()),
+					})
+				),
+				|name|texture_loader.acquire_render_config_id(name),
+			);
+
+			let (textures,render_configs)=loader.into_render_configs().map_err(LoadError::Io)?.consume();
+
+			let map=map_step2.add_render_configs_and_textures(
+				render_configs.into_iter(),
+				textures.into_iter().map(|(texture_id,texture)|
+					(texture_id,match texture{
+						strafesnet_deferred_loader::texture::Texture::ImageDDS(data)=>data,
+					})
+				),
+			);
+
+			Ok(map)
+		},
+	}
+}

src/framework.rs

@@ -1,516 +0,0 @@
-use std::future::Future;
-#[cfg(target_arch = "wasm32")]
-use std::str::FromStr;
-#[cfg(not(target_arch = "wasm32"))]
-use std::time::Instant;
-#[cfg(target_arch = "wasm32")]
-use web_sys::{ImageBitmapRenderingContext, OffscreenCanvas};
-use winit::{
-	event::{self, WindowEvent},
-	event_loop::{ControlFlow, EventLoop},
-};
-
-#[allow(dead_code)]
-pub fn cast_slice<T>(data: &[T]) -> &[u8] {
-	use std::{mem::size_of, slice::from_raw_parts};
-
-	unsafe { from_raw_parts(data.as_ptr() as *const u8, data.len() * size_of::<T>()) }
-}
-
-#[allow(dead_code)]
-pub enum ShaderStage {
-	Vertex,
-	Fragment,
-	Compute,
-}
-
-pub trait Example: 'static + Sized {
-	fn optional_features() -> wgpu::Features {
-		wgpu::Features::empty()
-	}
-	fn required_features() -> wgpu::Features {
-		wgpu::Features::empty()
-	}
-	fn required_downlevel_capabilities() -> wgpu::DownlevelCapabilities {
-		wgpu::DownlevelCapabilities {
-			flags: wgpu::DownlevelFlags::empty(),
-			shader_model: wgpu::ShaderModel::Sm5,
-			..wgpu::DownlevelCapabilities::default()
-		}
-	}
-	fn required_limits() -> wgpu::Limits {
-		wgpu::Limits::downlevel_webgl2_defaults() // These downlevel limits will allow the code to run on all possible hardware
-	}
-	fn init(
-		config: &wgpu::SurfaceConfiguration,
-		adapter: &wgpu::Adapter,
-		device: &wgpu::Device,
-		queue: &wgpu::Queue,
-	) -> Self;
-	fn resize(
-		&mut self,
-		config: &wgpu::SurfaceConfiguration,
-		device: &wgpu::Device,
-		queue: &wgpu::Queue,
-	);
-	fn update(&mut self, event: WindowEvent);
-	fn move_mouse(&mut self, delta: (f64,f64));
-	fn render(
-		&mut self,
-		view: &wgpu::TextureView,
-		device: &wgpu::Device,
-		queue: &wgpu::Queue,
-		spawner: &Spawner,
-	);
-}
-
-struct Setup {
-	window: winit::window::Window,
-	event_loop: EventLoop<()>,
-	instance: wgpu::Instance,
-	size: winit::dpi::PhysicalSize<u32>,
-	surface: wgpu::Surface,
-	adapter: wgpu::Adapter,
-	device: wgpu::Device,
-	queue: wgpu::Queue,
-	#[cfg(target_arch = "wasm32")]
-	offscreen_canvas_setup: Option<OffscreenCanvasSetup>,
-}
-
-#[cfg(target_arch = "wasm32")]
-struct OffscreenCanvasSetup {
-	offscreen_canvas: OffscreenCanvas,
-	bitmap_renderer: ImageBitmapRenderingContext,
-}
-
-async fn setup<E: Example>(title: &str) -> Setup {
-	#[cfg(not(target_arch = "wasm32"))]
-	{
-		env_logger::init();
-	};
-
-	let event_loop = EventLoop::new();
-	let mut builder = winit::window::WindowBuilder::new();
-	builder = builder.with_title(title);
-	#[cfg(windows_OFF)] // TODO
-	{
-		use winit::platform::windows::WindowBuilderExtWindows;
-		builder = builder.with_no_redirection_bitmap(true);
-	}
-	let window = builder.build(&event_loop).unwrap();
-
-	#[cfg(target_arch = "wasm32")]
-	{
-		use winit::platform::web::WindowExtWebSys;
-		let query_string = web_sys::window().unwrap().location().search().unwrap();
-		let level: log::Level = parse_url_query_string(&query_string, "RUST_LOG")
-			.and_then(|x| x.parse().ok())
-			.unwrap_or(log::Level::Error);
-		console_log::init_with_level(level).expect("could not initialize logger");
-		std::panic::set_hook(Box::new(console_error_panic_hook::hook));
-		// On wasm, append the canvas to the document body
-		web_sys::window()
-			.and_then(|win| win.document())
-			.and_then(|doc| doc.body())
-			.and_then(|body| {
-				body.append_child(&web_sys::Element::from(window.canvas()))
-					.ok()
-			})
-			.expect("couldn't append canvas to document body");
-	}
-
-	#[cfg(target_arch = "wasm32")]
-	let mut offscreen_canvas_setup: Option<OffscreenCanvasSetup> = None;
-	#[cfg(target_arch = "wasm32")]
-	{
-		use wasm_bindgen::JsCast;
-		use winit::platform::web::WindowExtWebSys;
-
-		let query_string = web_sys::window().unwrap().location().search().unwrap();
-		if let Some(offscreen_canvas_param) =
-			parse_url_query_string(&query_string, "offscreen_canvas")
-		{
-			if FromStr::from_str(offscreen_canvas_param) == Ok(true) {
-				log::info!("Creating OffscreenCanvasSetup");
-
-				let offscreen_canvas =
-					OffscreenCanvas::new(1024, 768).expect("couldn't create OffscreenCanvas");
-
-				let bitmap_renderer = window
-					.canvas()
-					.get_context("bitmaprenderer")
-					.expect("couldn't create ImageBitmapRenderingContext (Result)")
-					.expect("couldn't create ImageBitmapRenderingContext (Option)")
-					.dyn_into::<ImageBitmapRenderingContext>()
-					.expect("couldn't convert into ImageBitmapRenderingContext");
-
-				offscreen_canvas_setup = Some(OffscreenCanvasSetup {
-					offscreen_canvas,
-					bitmap_renderer,
-				})
-			}
-		}
-	};
-
-	log::info!("Initializing the surface...");
-
-	let backends = wgpu::util::backend_bits_from_env().unwrap_or_else(wgpu::Backends::all);
-	let dx12_shader_compiler = wgpu::util::dx12_shader_compiler_from_env().unwrap_or_default();
-
-	let instance = wgpu::Instance::new(wgpu::InstanceDescriptor {
-		backends,
-		dx12_shader_compiler,
-	});
-	let (size, surface) = unsafe {
-		let size = window.inner_size();
-
-		#[cfg(any(not(target_arch = "wasm32"), target_os = "emscripten"))]
-		let surface = instance.create_surface(&window).unwrap();
-		#[cfg(all(target_arch = "wasm32", not(target_os = "emscripten")))]
-		let surface = {
-			if let Some(offscreen_canvas_setup) = &offscreen_canvas_setup {
-				log::info!("Creating surface from OffscreenCanvas");
-				instance.create_surface_from_offscreen_canvas(
-					offscreen_canvas_setup.offscreen_canvas.clone(),
-				)
-			} else {
-				instance.create_surface(&window)
-			}
-		}
-		.unwrap();
-
-		(size, surface)
-	};
-	let adapter = wgpu::util::initialize_adapter_from_env_or_default(&instance, Some(&surface))
-		.await
-		.expect("No suitable GPU adapters found on the system!");
-
-	#[cfg(not(target_arch = "wasm32"))]
-	{
-		let adapter_info = adapter.get_info();
-		println!("Using {} ({:?})", adapter_info.name, adapter_info.backend);
-	}
-
-	let optional_features = E::optional_features();
-	let required_features = E::required_features();
-	let adapter_features = adapter.features();
-	assert!(
-		adapter_features.contains(required_features),
-		"Adapter does not support required features for this example: {:?}",
-		required_features - adapter_features
-	);
-
-	let required_downlevel_capabilities = E::required_downlevel_capabilities();
-	let downlevel_capabilities = adapter.get_downlevel_capabilities();
-	assert!(
-		downlevel_capabilities.shader_model >= required_downlevel_capabilities.shader_model,
-		"Adapter does not support the minimum shader model required to run this example: {:?}",
-		required_downlevel_capabilities.shader_model
-	);
-	assert!(
-		downlevel_capabilities
-			.flags
-			.contains(required_downlevel_capabilities.flags),
-		"Adapter does not support the downlevel capabilities required to run this example: {:?}",
-		required_downlevel_capabilities.flags - downlevel_capabilities.flags
-	);
-
-	// Make sure we use the texture resolution limits from the adapter, so we can support images the size of the surface.
-	let needed_limits = E::required_limits().using_resolution(adapter.limits());
-
-	let trace_dir = std::env::var("WGPU_TRACE");
-	let (device, queue) = adapter
-		.request_device(
-			&wgpu::DeviceDescriptor {
-				label: None,
-				features: (optional_features & adapter_features) | required_features,
-				limits: needed_limits,
-			},
-			trace_dir.ok().as_ref().map(std::path::Path::new),
-		)
-		.await
-		.expect("Unable to find a suitable GPU adapter!");
-
-	Setup {
-		window,
-		event_loop,
-		instance,
-		size,
-		surface,
-		adapter,
-		device,
-		queue,
-		#[cfg(target_arch = "wasm32")]
-		offscreen_canvas_setup,
-	}
-}
-
-fn start<E: Example>(
-	#[cfg(not(target_arch = "wasm32"))] Setup {
-		window,
-		event_loop,
-		instance,
-		size,
-		surface,
-		adapter,
-		device,
-		queue,
-	}: Setup,
-	#[cfg(target_arch = "wasm32")] Setup {
-		window,
-		event_loop,
-		instance,
-		size,
-		surface,
-		adapter,
-		device,
-		queue,
-		offscreen_canvas_setup,
-	}: Setup,
-) {
-	let spawner = Spawner::new();
-	let mut config = surface
-		.get_default_config(&adapter, size.width, size.height)
-		.expect("Surface isn't supported by the adapter.");
-	let surface_view_format = config.format.add_srgb_suffix();
-	config.view_formats.push(surface_view_format);
-	surface.configure(&device, &config);
-
-	log::info!("Initializing the example...");
-	let mut example = E::init(&config, &adapter, &device, &queue);
-
-	#[cfg(not(target_arch = "wasm32"))]
-	let mut last_frame_inst = Instant::now();
-	#[cfg(not(target_arch = "wasm32"))]
-	let (mut frame_count, mut accum_time) = (0, 0.0);
-
-	log::info!("Entering render loop...");
-	event_loop.run(move |event, _, control_flow| {
-		let _ = (&instance, &adapter); // force ownership by the closure
-		*control_flow = if cfg!(feature = "metal-auto-capture") {
-			ControlFlow::Exit
-		} else {
-			ControlFlow::Poll
-		};
-		match event {
-			event::Event::RedrawEventsCleared => {
-				#[cfg(not(target_arch = "wasm32"))]
-				spawner.run_until_stalled();
-
-				window.request_redraw();
-			}
-			event::Event::WindowEvent {
-				event:
-					WindowEvent::Resized(size)
-					| WindowEvent::ScaleFactorChanged {
-						new_inner_size: &mut size,
-						..
-					},
-				..
-			} => {
-				// Once winit is fixed, the detection conditions here can be removed.
-				// https://github.com/rust-windowing/winit/issues/2876
-				let max_dimension = adapter.limits().max_texture_dimension_2d;
-				if size.width > max_dimension || size.height > max_dimension {
-					log::warn!(
-						"The resizing size {:?} exceeds the limit of {}.",
-						size,
-						max_dimension
-					);
-				} else {
-					log::info!("Resizing to {:?}", size);
-					config.width = size.width.max(1);
-					config.height = size.height.max(1);
-					example.resize(&config, &device, &queue);
-					surface.configure(&device, &config);
-				}
-			}
-			event::Event::WindowEvent { event, .. } => match event {
-				WindowEvent::KeyboardInput {
-					input:
-						event::KeyboardInput {
-							virtual_keycode: Some(event::VirtualKeyCode::Escape),
-							state: event::ElementState::Pressed,
-							..
-						},
-					..
-				}
-				| WindowEvent::CloseRequested => {
-					*control_flow = ControlFlow::Exit;
-				}
-				#[cfg(not(target_arch = "wasm32"))]
-				WindowEvent::KeyboardInput {
-					input:
-						event::KeyboardInput {
-							virtual_keycode: Some(event::VirtualKeyCode::R),
-							state: event::ElementState::Pressed,
-							..
-						},
-					..
-				} => {
-					println!("{:#?}", instance.generate_report());
-				}
-				_ => {
-					example.update(event);
-				}
-			},
-			event::Event::DeviceEvent {
-				event:
-					winit::event::DeviceEvent::MouseMotion {
-						delta,
-					},
-				..
-			} => {
-				example.move_mouse(delta);
-			},
-			event::Event::RedrawRequested(_) => {
-				#[cfg(not(target_arch = "wasm32"))]
-				{
-					accum_time += last_frame_inst.elapsed().as_secs_f32();
-					last_frame_inst = Instant::now();
-					frame_count += 1;
-					if frame_count == 100 {
-						println!(
-							"Avg frame time {}ms",
-							accum_time * 1000.0 / frame_count as f32
-						);
-						accum_time = 0.0;
-						frame_count = 0;
-					}
-				}
-
-				let frame = match surface.get_current_texture() {
-					Ok(frame) => frame,
-					Err(_) => {
-						surface.configure(&device, &config);
-						surface
-							.get_current_texture()
-							.expect("Failed to acquire next surface texture!")
-					}
-				};
-				let view = frame.texture.create_view(&wgpu::TextureViewDescriptor {
-					format: Some(surface_view_format),
-					..wgpu::TextureViewDescriptor::default()
-				});
-
-				example.render(&view, &device, &queue, &spawner);
-
-				frame.present();
-
-				#[cfg(target_arch = "wasm32")]
-				{
-					if let Some(offscreen_canvas_setup) = &offscreen_canvas_setup {
-						let image_bitmap = offscreen_canvas_setup
-							.offscreen_canvas
-							.transfer_to_image_bitmap()
-							.expect("couldn't transfer offscreen canvas to image bitmap.");
-						offscreen_canvas_setup
-							.bitmap_renderer
-							.transfer_from_image_bitmap(&image_bitmap);
-
-						log::info!("Transferring OffscreenCanvas to ImageBitmapRenderer");
-					}
-				}
-			}
-			_ => {}
-		}
-	});
-}
-
-#[cfg(not(target_arch = "wasm32"))]
-pub struct Spawner<'a> {
-	executor: async_executor::LocalExecutor<'a>,
-}
-
-#[cfg(not(target_arch = "wasm32"))]
-impl<'a> Spawner<'a> {
-	fn new() -> Self {
-		Self {
-			executor: async_executor::LocalExecutor::new(),
-		}
-	}
-
-	#[allow(dead_code)]
-	pub fn spawn_local(&self, future: impl Future<Output = ()> + 'a) {
-		self.executor.spawn(future).detach();
-	}
-
-	fn run_until_stalled(&self) {
-		while self.executor.try_tick() {}
-	}
-}
-
-#[cfg(target_arch = "wasm32")]
-pub struct Spawner {}
-
-#[cfg(target_arch = "wasm32")]
-impl Spawner {
-	fn new() -> Self {
-		Self {}
-	}
-
-	#[allow(dead_code)]
-	pub fn spawn_local(&self, future: impl Future<Output = ()> + 'static) {
-		wasm_bindgen_futures::spawn_local(future);
-	}
-}
-
-#[cfg(not(target_arch = "wasm32"))]
-pub fn run<E: Example>(title: &str) {
-	let setup = pollster::block_on(setup::<E>(title));
-	start::<E>(setup);
-}
-
-#[cfg(target_arch = "wasm32")]
-pub fn run<E: Example>(title: &str) {
-	use wasm_bindgen::prelude::*;
-
-	let title = title.to_owned();
-	wasm_bindgen_futures::spawn_local(async move {
-		let setup = setup::<E>(&title).await;
-		let start_closure = Closure::once_into_js(move || start::<E>(setup));
-
-		// make sure to handle JS exceptions thrown inside start.
-		// Otherwise wasm_bindgen_futures Queue would break and never handle any tasks again.
-		// This is required, because winit uses JS exception for control flow to escape from `run`.
-		if let Err(error) = call_catch(&start_closure) {
-			let is_control_flow_exception = error.dyn_ref::<js_sys::Error>().map_or(false, |e| {
-				e.message().includes("Using exceptions for control flow", 0)
-			});
-
-			if !is_control_flow_exception {
-				web_sys::console::error_1(&error);
-			}
-		}
-
-		#[wasm_bindgen]
-		extern "C" {
-			#[wasm_bindgen(catch, js_namespace = Function, js_name = "prototype.call.call")]
-			fn call_catch(this: &JsValue) -> Result<(), JsValue>;
-		}
-	});
-}
-
-#[cfg(target_arch = "wasm32")]
-/// Parse the query string as returned by `web_sys::window()?.location().search()?` and get a
-/// specific key out of it.
-pub fn parse_url_query_string<'a>(query: &'a str, search_key: &str) -> Option<&'a str> {
-	let query_string = query.strip_prefix('?')?;
-
-	for pair in query_string.split('&') {
-		let mut pair = pair.split('=');
-		let key = pair.next()?;
-		let value = pair.next()?;
-
-		if key == search_key {
-			return Some(value);
-		}
-	}
-
-	None
-}
-
-// This allows treating the framework as a standalone example,
-// thus avoiding listing the example names in `Cargo.toml`.
-#[allow(dead_code)]
-fn main() {}

src/graphics.rs

@@ -0,0 +1,989 @@
+use std::borrow::Cow;
+use std::collections::{HashSet,HashMap};
+use strafesnet_common::map;
+use strafesnet_common::integer;
+use strafesnet_common::model::{self, ColorId, NormalId, PolygonIter, PositionId, RenderConfigId, TextureCoordinateId, VertexId};
+use wgpu::{util::DeviceExt,AstcBlock,AstcChannel};
+use crate::model_graphics::{self,IndexedGraphicsMeshOwnedRenderConfig,IndexedGraphicsMeshOwnedRenderConfigId,GraphicsMeshOwnedRenderConfig,GraphicsModelColor4,GraphicsModelOwned,GraphicsVertex};
+
+struct Indices{
+	count:u32,
+	buf:wgpu::Buffer,
+	format:wgpu::IndexFormat,
+}
+impl Indices{
+	fn new<T:bytemuck::Pod>(device:&wgpu::Device,indices:&Vec<T>,format:wgpu::IndexFormat)->Self{
+		Self{
+			buf:device.create_buffer_init(&wgpu::util::BufferInitDescriptor{
+				label:Some("Index"),
+				contents:bytemuck::cast_slice(indices),
+				usage:wgpu::BufferUsages::INDEX,
+			}),
+			count:indices.len() as u32,
+			format,
+		}
+	}
+}
+struct GraphicsModel{
+	indices:Indices,
+	vertex_buf:wgpu::Buffer,
+	bind_group:wgpu::BindGroup,
+	instance_count:u32,
+}
+
+struct GraphicsSamplers{
+	repeat:wgpu::Sampler,
+}
+
+struct GraphicsBindGroupLayouts{
+	model:wgpu::BindGroupLayout,
+}
+
+struct GraphicsBindGroups{
+	camera:wgpu::BindGroup,
+	skybox_texture:wgpu::BindGroup,
+}
+
+struct GraphicsPipelines{
+	skybox:wgpu::RenderPipeline,
+	model:wgpu::RenderPipeline,
+}
+
+struct GraphicsCamera{
+	screen_size:glam::UVec2,
+	fov:glam::Vec2,//slope
+	//camera angles and such are extrapolated and passed in every time
+}
+
+#[inline]
+fn perspective_rh(fov_x_slope:f32,fov_y_slope:f32,z_near:f32,z_far:f32)->glam::Mat4{
+	//glam_assert!(z_near > 0.0 && z_far > 0.0);
+	let r=z_far/(z_near-z_far);
+	glam::Mat4::from_cols(
+		glam::Vec4::new(1.0/fov_x_slope,0.0,0.0,0.0),
+		glam::Vec4::new(0.0,1.0/fov_y_slope,0.0,0.0),
+		glam::Vec4::new(0.0,0.0,r,-1.0),
+		glam::Vec4::new(0.0,0.0,r*z_near,0.0),
+	)
+}
+impl GraphicsCamera{
+	pub fn proj(&self)->glam::Mat4{
+		perspective_rh(self.fov.x,self.fov.y,0.4,4000.0)
+	}
+	pub fn world(&self,pos:glam::Vec3,angles:glam::Vec2)->glam::Mat4{
+		//f32 good enough for view matrix
+		glam::Mat4::from_translation(pos)*glam::Mat4::from_euler(glam::EulerRot::YXZ,angles.x,angles.y,0f32)
+	}
+
+	pub fn to_uniform_data(&self,pos:glam::Vec3,angles:glam::Vec2)->[f32;16*4]{
+		let proj=self.proj();
+		let proj_inv=proj.inverse();
+		let view_inv=self.world(pos,angles);
+		let view=view_inv.inverse();
+
+		let mut raw=[0f32; 16 * 4];
+		raw[..16].copy_from_slice(&AsRef::<[f32; 16]>::as_ref(&proj)[..]);
+		raw[16..32].copy_from_slice(&AsRef::<[f32; 16]>::as_ref(&proj_inv)[..]);
+		raw[32..48].copy_from_slice(&AsRef::<[f32; 16]>::as_ref(&view)[..]);
+		raw[48..64].copy_from_slice(&AsRef::<[f32; 16]>::as_ref(&view_inv)[..]);
+		raw
+	}
+}
+impl std::default::Default for GraphicsCamera{
+	fn default()->Self{
+		Self{
+			screen_size:glam::UVec2::ONE,
+			fov:glam::Vec2::ONE,
+		}
+	}
+}
+
+pub struct FrameState{
+	pub body:crate::physics::Body,
+	pub camera:crate::physics::PhysicsCamera,
+	pub time:integer::Time,
+}
+
+pub struct GraphicsState{
+	pipelines:GraphicsPipelines,
+	bind_groups:GraphicsBindGroups,
+	bind_group_layouts:GraphicsBindGroupLayouts,
+	samplers:GraphicsSamplers,
+	camera:GraphicsCamera,
+	camera_buf:wgpu::Buffer,
+	temp_squid_texture_view:wgpu::TextureView,
+	models:Vec<GraphicsModel>,
+	depth_view:wgpu::TextureView,
+	staging_belt:wgpu::util::StagingBelt,
+}
+
+impl GraphicsState{
+	const DEPTH_FORMAT:wgpu::TextureFormat=wgpu::TextureFormat::Depth24Plus;
+	fn create_depth_texture(
+		config:&wgpu::SurfaceConfiguration,
+		device:&wgpu::Device,
+	)->wgpu::TextureView{
+		let depth_texture=device.create_texture(&wgpu::TextureDescriptor{
+			size:wgpu::Extent3d{
+				width:config.width,
+				height:config.height,
+				depth_or_array_layers:1,
+			},
+			mip_level_count:1,
+			sample_count:1,
+			dimension:wgpu::TextureDimension::D2,
+			format:Self::DEPTH_FORMAT,
+			usage:wgpu::TextureUsages::RENDER_ATTACHMENT,
+			label:None,
+			view_formats:&[],
+		});
+
+		depth_texture.create_view(&wgpu::TextureViewDescriptor::default())
+	}
+	pub fn clear(&mut self){
+		self.models.clear();
+	}
+	pub fn load_user_settings(&mut self,user_settings:&crate::settings::UserSettings){
+		self.camera.fov=user_settings.calculate_fov(1.0,&self.camera.screen_size).as_vec2();
+	}
+	pub fn generate_models(&mut self,device:&wgpu::Device,queue:&wgpu::Queue,map:&map::CompleteMap){
+		//generate texture view per texture
+		let texture_views:HashMap<strafesnet_common::model::TextureId,wgpu::TextureView>=map.textures.iter().enumerate().filter_map(|(texture_id,texture_data)|{
+			let texture_id=model::TextureId::new(texture_id as u32);
+			let image=match ddsfile::Dds::read(std::io::Cursor::new(texture_data)){
+				Ok(image)=>image,
+				Err(e)=>{
+					println!("Error loading texture: {e}");
+					return None;
+				},
+			};
+
+			let (mut width,mut height)=(image.get_width(),image.get_height());
+
+			let format=match image.header10.unwrap().dxgi_format{
+				ddsfile::DxgiFormat::R8G8B8A8_UNorm_sRGB=>wgpu::TextureFormat::Rgba8UnormSrgb,
+				ddsfile::DxgiFormat::BC7_UNorm_sRGB =>{
+					//floor(w,4),should be ceil(w,4)
+					width=width/4*4;
+					height=height/4*4;
+					wgpu::TextureFormat::Bc7RgbaUnormSrgb
+				},
+				other=>{
+					println!("unsupported texture format{:?}",other);
+					return None;
+				},
+			};
+
+			let size=wgpu::Extent3d{
+				width,
+				height,
+				depth_or_array_layers:1,
+			};
+
+			let layer_size=wgpu::Extent3d{
+				depth_or_array_layers:1,
+				..size
+			};
+			let max_mips=layer_size.max_mips(wgpu::TextureDimension::D2);
+
+			let texture=device.create_texture_with_data(
+				queue,
+				&wgpu::TextureDescriptor{
+					size,
+					mip_level_count:max_mips,
+					sample_count:1,
+					dimension:wgpu::TextureDimension::D2,
+					format,
+					usage:wgpu::TextureUsages::TEXTURE_BINDING|wgpu::TextureUsages::COPY_DST,
+					label:Some(format!("Texture{}",texture_id.get()).as_str()),
+					view_formats:&[],
+				},
+				wgpu::util::TextureDataOrder::LayerMajor,
+				&image.data,
+			);
+			Some((texture_id,texture.create_view(&wgpu::TextureViewDescriptor{
+				label:Some(format!("Texture{} View",texture_id.get()).as_str()),
+				dimension:Some(wgpu::TextureViewDimension::D2),
+				..wgpu::TextureViewDescriptor::default()
+			})))
+		}).collect();
+		let num_textures=texture_views.len();
+
+		//split groups with different textures into separate models
+		//the models received here are supposed to be tightly packed,i.e. no code needs to check if two models are using the same groups.
+		let indexed_models_len=map.models.len();
+		//models split into graphics_group.RenderConfigId
+		let mut owned_mesh_id_from_mesh_id_render_config_id:HashMap<model::MeshId,HashMap<RenderConfigId,IndexedGraphicsMeshOwnedRenderConfigId>>=HashMap::new();
+		let mut unique_render_config_models:Vec<IndexedGraphicsMeshOwnedRenderConfig>=Vec::with_capacity(indexed_models_len);
+		for model in &map.models{
+			//wow
+			let instance=GraphicsModelOwned{
+				transform:model.transform.into(),
+				normal_transform:glam::Mat3::from_cols_array_2d(&model.transform.matrix3.to_array().map(|row|row.map(Into::into))).inverse().transpose(),
+				color:GraphicsModelColor4::new(model.color),
+			};
+			//get or create owned mesh map
+			let owned_mesh_map=owned_mesh_id_from_mesh_id_render_config_id
+			.entry(model.mesh).or_insert_with(||{
+				let mut owned_mesh_map=HashMap::new();
+				//add mesh if renderid never before seen for this model
+				//add instance
+				//convert Model into GraphicsModelOwned
+				//check each group, if it's using a new render config then make a new clone of the model
+				if let Some(mesh)=map.meshes.get(model.mesh.get() as usize){
+					for graphics_group in mesh.graphics_groups.iter(){
+						//get or create owned mesh
+						let owned_mesh_id=owned_mesh_map
+						.entry(graphics_group.render).or_insert_with(||{
+							//create
+							let owned_mesh_id=IndexedGraphicsMeshOwnedRenderConfigId::new(unique_render_config_models.len() as u32);
+							unique_render_config_models.push(IndexedGraphicsMeshOwnedRenderConfig{
+								unique_pos:mesh.unique_pos.iter().map(|v|v.to_array().map(Into::into)).collect(),
+								unique_tex:mesh.unique_tex.iter().map(|v|*v.as_ref()).collect(),
+								unique_normal:mesh.unique_normal.iter().map(|v|v.to_array().map(Into::into)).collect(),
+								unique_color:mesh.unique_color.iter().map(|v|*v.as_ref()).collect(),
+								unique_vertices:mesh.unique_vertices.clone(),
+								render_config:graphics_group.render,
+								polys:model::PolygonGroup::PolygonList(model::PolygonList::new(Vec::new())),
+								instances:Vec::new(),
+							});
+							owned_mesh_id
+						});
+						let owned_mesh=unique_render_config_models.get_mut(owned_mesh_id.get() as usize).unwrap();
+						match &mut owned_mesh.polys{
+							model::PolygonGroup::PolygonList(polygon_list)=>polygon_list.extend(
+								graphics_group.groups.iter().flat_map(|polygon_group_id|{
+									mesh.polygon_groups[polygon_group_id.get() as usize].polys()
+								})
+								.map(|vertex_id_slice|
+									vertex_id_slice.to_vec()
+								)
+							),
+						}
+					}
+				}
+				owned_mesh_map
+			});
+			for owned_mesh_id in owned_mesh_map.values(){
+				let owned_mesh=unique_render_config_models.get_mut(owned_mesh_id.get() as usize).unwrap();
+				let render_config=&map.render_configs[owned_mesh.render_config.get() as usize];
+				if model.color.w==0.0&&render_config.texture.is_none(){
+					continue;
+				}
+				owned_mesh.instances.push(instance.clone());
+			}
+		}
+		//check every model to see if it's using the same (texture,color) but has few instances,if it is combine it into one model
+		//1. collect unique instances of texture and color,note model id
+		//2. for each model id,check if removing it from the pool decreases both the model count and instance count by more than one
+		//3. transpose all models that stay in the set
+
+		//best plan:benchmark set_bind_group,set_vertex_buffer,set_index_buffer and draw_indexed
+		//check if the estimated render performance is better by transposing multiple model instances into one model instance
+
+		//for now:just deduplicate single models...
+		let mut deduplicated_models=Vec::with_capacity(indexed_models_len);//use indexed_models_len because the list will likely get smaller instead of bigger
+		let mut unique_texture_color=HashMap::new();//texture->color->vec![(model_id,instance_id)]
+		for (model_id,model) in unique_render_config_models.iter().enumerate(){
+			//for now:filter out models with more than one instance
+			if 1<model.instances.len(){
+				continue;
+			}
+			//populate hashmap
+			let unique_color=unique_texture_color
+			.entry(model.render_config)
+			.or_insert_with(||HashMap::new());
+			//separate instances by color
+			for (instance_id,instance) in model.instances.iter().enumerate(){
+				let model_instance_list=unique_color
+				.entry(instance.color)
+				.or_insert_with(||Vec::new());
+				//add model instance to list
+				model_instance_list.push((model_id,instance_id));
+			}
+		}
+		//populate a hashset of models selected for transposition
+		//construct transposed models
+		let mut selected_model_instances=HashSet::new();
+		for (render_config,unique_color) in unique_texture_color.into_iter(){
+			for (color,model_instance_list) in unique_color.into_iter(){
+				//world transforming one model does not meet the definition of deduplicaiton
+				if 1<model_instance_list.len(){
+					//create model
+					let mut unique_pos=Vec::new();
+					let mut pos_id_from=HashMap::new();
+					let mut unique_tex=Vec::new();
+					let mut tex_id_from=HashMap::new();
+					let mut unique_normal=Vec::new();
+					let mut normal_id_from=HashMap::new();
+					let mut unique_color=Vec::new();
+					let mut color_id_from=HashMap::new();
+					let mut unique_vertices=Vec::new();
+					let mut vertex_id_from=HashMap::new();
+
+					let mut polys=Vec::new();
+					//transform instance vertices
+					for (model_id,instance_id) in model_instance_list.into_iter(){
+						//populate hashset to prevent these models from being copied
+						selected_model_instances.insert(model_id);
+						//there is only one instance per model
+						let model=&unique_render_config_models[model_id];
+						let instance=&model.instances[instance_id];
+						//just hash word slices LOL
+						let map_pos_id:Vec<PositionId>=model.unique_pos.iter().map(|untransformed_pos|{
+							let pos=instance.transform.transform_point3(glam::Vec3::from_array(untransformed_pos.clone())).to_array();
+							let h=bytemuck::cast::<[f32;3],[u32;3]>(pos);
+							PositionId::new(*pos_id_from.entry(h).or_insert_with(||{
+								let pos_id=unique_pos.len();
+								unique_pos.push(pos);
+								pos_id
+							}) as u32)
+						}).collect();
+						let map_tex_id:Vec<TextureCoordinateId>=model.unique_tex.iter().map(|&tex|{
+							let h=bytemuck::cast::<[f32;2],[u32;2]>(tex);
+							TextureCoordinateId::new(*tex_id_from.entry(h).or_insert_with(||{
+								let tex_id=unique_tex.len();
+								unique_tex.push(tex);
+								tex_id
+							}) as u32)
+						}).collect();
+						let map_normal_id:Vec<NormalId>=model.unique_normal.iter().map(|untransformed_normal|{
+							let normal=(instance.normal_transform*glam::Vec3::from_array(untransformed_normal.clone())).to_array();
+							let h=bytemuck::cast::<[f32;3],[u32;3]>(normal);
+							NormalId::new(*normal_id_from.entry(h).or_insert_with(||{
+								let normal_id=unique_normal.len();
+								unique_normal.push(normal);
+								normal_id
+							}) as u32)
+						}).collect();
+						let map_color_id:Vec<ColorId>=model.unique_color.iter().map(|&color|{
+							let h=bytemuck::cast::<[f32;4],[u32;4]>(color);
+							ColorId::new(*color_id_from.entry(h).or_insert_with(||{
+								let color_id=unique_color.len();
+								unique_color.push(color);
+								color_id
+							}) as u32)
+						}).collect();
+						//map the indexed vertices onto new indices
+						//creating the vertex map is slightly different because the vertices are directly hashable
+						let map_vertex_id:Vec<VertexId>=model.unique_vertices.iter().map(|unmapped_vertex|{
+							let vertex=model::IndexedVertex{
+								pos:map_pos_id[unmapped_vertex.pos.get() as usize],
+								tex:map_tex_id[unmapped_vertex.tex.get() as usize],
+								normal:map_normal_id[unmapped_vertex.normal.get() as usize],
+								color:map_color_id[unmapped_vertex.color.get() as usize],
+							};
+							VertexId::new(*vertex_id_from.entry(vertex.clone()).or_insert_with(||{
+								let vertex_id=unique_vertices.len();
+								unique_vertices.push(vertex);
+								vertex_id
+							}) as u32)
+						}).collect();
+						polys.extend(model.polys.polys().map(|poly|
+							poly.iter().map(|vertex_id|
+								map_vertex_id[vertex_id.get() as usize]
+							).collect()
+						));
+					}
+					//push model into dedup
+					deduplicated_models.push(IndexedGraphicsMeshOwnedRenderConfig{
+						unique_pos,
+						unique_tex,
+						unique_normal,
+						unique_color,
+						unique_vertices,
+						render_config,
+						polys:model::PolygonGroup::PolygonList(model::PolygonList::new(polys)),
+						instances:vec![GraphicsModelOwned{
+							transform:glam::Mat4::IDENTITY,
+							normal_transform:glam::Mat3::IDENTITY,
+							color
+						}],
+					});
+				}
+			}
+		}
+		//fill untouched models
+		for (model_id,model) in unique_render_config_models.into_iter().enumerate(){
+			if !selected_model_instances.contains(&model_id){
+				deduplicated_models.push(model);
+			}
+		}
+
+		//de-index models
+		let deduplicated_models_len=deduplicated_models.len();
+		let models:Vec<GraphicsMeshOwnedRenderConfig>=deduplicated_models.into_iter().map(|model|{
+			let mut vertices=Vec::new();
+			let mut index_from_vertex=HashMap::new();//::<IndexedVertex,usize>
+			//this mut be combined in a more complex way if the models use different render patterns per group
+			let mut indices=Vec::new();
+			for poly in model.polys.polys(){
+				let mut poly_vertices=poly.iter()
+				.map(|&vertex_index|*index_from_vertex.entry(vertex_index).or_insert_with(||{
+					let i=vertices.len();
+					let vertex=&model.unique_vertices[vertex_index.get() as usize];
+					vertices.push(GraphicsVertex{
+						pos:model.unique_pos[vertex.pos.get() as usize],
+						tex:model.unique_tex[vertex.tex.get() as usize],
+						normal:model.unique_normal[vertex.normal.get() as usize],
+						color:model.unique_color[vertex.color.get() as usize],
+					});
+					i
+				}));
+
+				let a=poly_vertices.next().unwrap();
+				let mut b=poly_vertices.next().unwrap();
+
+				poly_vertices.for_each(|c|{
+					indices.extend([a,b,c]);
+					b=c;
+				});
+			}
+			GraphicsMeshOwnedRenderConfig{
+				instances:model.instances,
+				indices:if (u32::MAX as usize)<vertices.len(){
+					panic!("Model has too many vertices!")
+				}else if (u16::MAX as usize)<vertices.len(){
+					model_graphics::Indices::U32(indices.into_iter().map(|vertex_idx|vertex_idx as u32).collect())
+				}else{
+					model_graphics::Indices::U16(indices.into_iter().map(|vertex_idx|vertex_idx as u16).collect())
+				},
+				vertices,
+				render_config:model.render_config,
+			}
+		}).collect();
+		//.into_iter() the modeldata vec so entities can be /moved/ to models.entities
+		let mut model_count=0;
+		let mut instance_count=0;
+		let uniform_buffer_binding_size=crate::setup::required_limits().max_uniform_buffer_binding_size as usize;
+		let chunk_size=uniform_buffer_binding_size/MODEL_BUFFER_SIZE_BYTES;
+		self.models.reserve(models.len());
+		for model in models.into_iter(){
+			instance_count+=model.instances.len();
+			for instances_chunk in model.instances.rchunks(chunk_size){
+				model_count+=1;
+				let mut model_uniforms=get_instances_buffer_data(instances_chunk);
+				//TEMP: fill with zeroes to pass validation
+				model_uniforms.resize(MODEL_BUFFER_SIZE*512,0.0f32);
+				let model_buf=device.create_buffer_init(&wgpu::util::BufferInitDescriptor{
+					label:Some(format!("Model{} Buf",model_count).as_str()),
+					contents:bytemuck::cast_slice(&model_uniforms),
+					usage:wgpu::BufferUsages::UNIFORM|wgpu::BufferUsages::COPY_DST,
+				});
+				let render_config=&map.render_configs[model.render_config.get() as usize];
+				let texture_view=render_config.texture.and_then(|texture_id|
+					texture_views.get(&texture_id)
+				).unwrap_or(&self.temp_squid_texture_view);
+				let bind_group=device.create_bind_group(&wgpu::BindGroupDescriptor{
+					layout:&self.bind_group_layouts.model,
+					entries:&[
+						wgpu::BindGroupEntry{
+							binding:0,
+							resource:model_buf.as_entire_binding(),
+						},
+						wgpu::BindGroupEntry{
+							binding:1,
+							resource:wgpu::BindingResource::TextureView(texture_view),
+						},
+						wgpu::BindGroupEntry{
+							binding:2,
+							resource:wgpu::BindingResource::Sampler(&self.samplers.repeat),
+						},
+					],
+					label:Some(format!("Model{} Bind Group",model_count).as_str()),
+				});
+				let vertex_buf=device.create_buffer_init(&wgpu::util::BufferInitDescriptor{
+					label:Some("Vertex"),
+					contents:bytemuck::cast_slice(&model.vertices),
+					usage:wgpu::BufferUsages::VERTEX,
+				});
+				//all of these are being moved here
+				self.models.push(GraphicsModel{
+					instance_count:instances_chunk.len() as u32,
+					vertex_buf,
+					indices:match &model.indices{
+						model_graphics::Indices::U32(indices)=>Indices::new(device,indices,wgpu::IndexFormat::Uint32),
+						model_graphics::Indices::U16(indices)=>Indices::new(device,indices,wgpu::IndexFormat::Uint16),
+					},
+					bind_group,
+				});
+			}
+		}
+		println!("Texture References={}",num_textures);
+		println!("Textures Loaded={}",texture_views.len());
+		println!("Indexed Models={}",indexed_models_len);
+		println!("Deduplicated Models={}",deduplicated_models_len);
+		println!("Graphics Objects:{}",self.models.len());
+		println!("Graphics Instances:{}",instance_count);
+	}
+
+	pub fn new(
+		device:&wgpu::Device,
+		queue:&wgpu::Queue,
+		config:&wgpu::SurfaceConfiguration,
+	)->Self{
+		let camera_bind_group_layout=device.create_bind_group_layout(&wgpu::BindGroupLayoutDescriptor{
+			label:None,
+			entries:&[
+				wgpu::BindGroupLayoutEntry{
+					binding:0,
+					visibility:wgpu::ShaderStages::VERTEX,
+					ty:wgpu::BindingType::Buffer{
+						ty:wgpu::BufferBindingType::Uniform,
+						has_dynamic_offset:false,
+						min_binding_size:None,
+					},
+					count:None,
+				},
+			],
+		});
+		let skybox_texture_bind_group_layout=device.create_bind_group_layout(&wgpu::BindGroupLayoutDescriptor{
+			label:Some("Skybox Texture Bind Group Layout"),
+			entries:&[
+				wgpu::BindGroupLayoutEntry{
+					binding:0,
+					visibility:wgpu::ShaderStages::FRAGMENT,
+					ty:wgpu::BindingType::Texture{
+						sample_type:wgpu::TextureSampleType::Float{filterable:true},
+						multisampled:false,
+						view_dimension:wgpu::TextureViewDimension::Cube,
+					},
+					count:None,
+				},
+				wgpu::BindGroupLayoutEntry{
+					binding:1,
+					visibility:wgpu::ShaderStages::FRAGMENT,
+					ty:wgpu::BindingType::Sampler(wgpu::SamplerBindingType::Filtering),
+					count:None,
+				},
+			],
+		});
+		let model_bind_group_layout=device.create_bind_group_layout(&wgpu::BindGroupLayoutDescriptor{
+			label:Some("Model Bind Group Layout"),
+			entries:&[
+				wgpu::BindGroupLayoutEntry{
+					binding:0,
+					visibility:wgpu::ShaderStages::VERTEX,
+					ty:wgpu::BindingType::Buffer{
+						ty:wgpu::BufferBindingType::Uniform,
+						has_dynamic_offset:false,
+						min_binding_size:None,
+					},
+					count:None,
+				},
+				wgpu::BindGroupLayoutEntry{
+					binding:1,
+					visibility:wgpu::ShaderStages::FRAGMENT,
+					ty:wgpu::BindingType::Texture{
+						sample_type:wgpu::TextureSampleType::Float{filterable:true},
+						multisampled:false,
+						view_dimension:wgpu::TextureViewDimension::D2,
+					},
+					count:None,
+				},
+				wgpu::BindGroupLayoutEntry{
+					binding:2,
+					visibility:wgpu::ShaderStages::FRAGMENT,
+					ty:wgpu::BindingType::Sampler(wgpu::SamplerBindingType::Filtering),
+					count:None,
+				},
+			],
+		});
+
+		let clamp_sampler=device.create_sampler(&wgpu::SamplerDescriptor{
+			label:Some("Clamp Sampler"),
+			address_mode_u:wgpu::AddressMode::ClampToEdge,
+			address_mode_v:wgpu::AddressMode::ClampToEdge,
+			address_mode_w:wgpu::AddressMode::ClampToEdge,
+			mag_filter:wgpu::FilterMode::Linear,
+			min_filter:wgpu::FilterMode::Linear,
+			mipmap_filter:wgpu::FilterMode::Linear,
+			..Default::default()
+		});
+		let repeat_sampler=device.create_sampler(&wgpu::SamplerDescriptor{
+			label:Some("Repeat Sampler"),
+			address_mode_u:wgpu::AddressMode::Repeat,
+			address_mode_v:wgpu::AddressMode::Repeat,
+			address_mode_w:wgpu::AddressMode::Repeat,
+			mag_filter:wgpu::FilterMode::Linear,
+			min_filter:wgpu::FilterMode::Linear,
+			mipmap_filter:wgpu::FilterMode::Linear,
+			anisotropy_clamp:16,
+			..Default::default()
+		});
+
+		// Create the render pipeline
+		let shader=device.create_shader_module(wgpu::ShaderModuleDescriptor{
+			label:None,
+			source:wgpu::ShaderSource::Wgsl(Cow::Borrowed(include_str!("shader.wgsl"))),
+		});
+
+		//load textures
+		let device_features=device.features();
+
+		let skybox_texture_view={
+			let skybox_format=if device_features.contains(wgpu::Features::TEXTURE_COMPRESSION_ASTC){
+				println!("Using ASTC");
+				wgpu::TextureFormat::Astc{
+					block:AstcBlock::B4x4,
+					channel:AstcChannel::UnormSrgb,
+				}
+			}else if device_features.contains(wgpu::Features::TEXTURE_COMPRESSION_ETC2){
+				println!("Using ETC2");
+				wgpu::TextureFormat::Etc2Rgb8UnormSrgb
+			}else if device_features.contains(wgpu::Features::TEXTURE_COMPRESSION_BC){
+				println!("Using BC");
+				wgpu::TextureFormat::Bc1RgbaUnormSrgb
+			}else{
+				println!("Using plain");
+				wgpu::TextureFormat::Bgra8UnormSrgb
+			};
+
+			let bytes=match skybox_format{
+				wgpu::TextureFormat::Astc{
+					block:AstcBlock::B4x4,
+					channel:AstcChannel::UnormSrgb,
+				}=>&include_bytes!("../images/astc.dds")[..],
+				wgpu::TextureFormat::Etc2Rgb8UnormSrgb=>&include_bytes!("../images/etc2.dds")[..],
+				wgpu::TextureFormat::Bc1RgbaUnormSrgb=>&include_bytes!("../images/bc1.dds")[..],
+				wgpu::TextureFormat::Bgra8UnormSrgb=>&include_bytes!("../images/bgra.dds")[..],
+				_=>unreachable!(),
+			};
+
+			let skybox_image=ddsfile::Dds::read(&mut std::io::Cursor::new(bytes)).unwrap();
+
+			let size=wgpu::Extent3d{
+				width:skybox_image.get_width(),
+				height:skybox_image.get_height(),
+				depth_or_array_layers:6,
+			};
+
+			let layer_size=wgpu::Extent3d{
+				depth_or_array_layers:1,
+				..size
+			};
+			let max_mips=layer_size.max_mips(wgpu::TextureDimension::D2);
+
+			let skybox_texture=device.create_texture_with_data(
+				queue,
+				&wgpu::TextureDescriptor{
+					size,
+					mip_level_count:max_mips,
+					sample_count:1,
+					dimension:wgpu::TextureDimension::D2,
+					format:skybox_format,
+					usage:wgpu::TextureUsages::TEXTURE_BINDING|wgpu::TextureUsages::COPY_DST,
+					label:Some("Skybox Texture"),
+					view_formats:&[],
+				},
+				wgpu::util::TextureDataOrder::LayerMajor,
+				&skybox_image.data,
+			);
+
+			skybox_texture.create_view(&wgpu::TextureViewDescriptor{
+				label:Some("Skybox Texture View"),
+				dimension:Some(wgpu::TextureViewDimension::Cube),
+				..wgpu::TextureViewDescriptor::default()
+			})
+		};
+
+		//squid
+		let squid_texture_view={
+			let bytes=include_bytes!("../images/squid.dds");
+
+			let image=ddsfile::Dds::read(&mut std::io::Cursor::new(bytes)).unwrap();
+
+			let size=wgpu::Extent3d{
+				width:image.get_width(),
+				height:image.get_height(),
+				depth_or_array_layers:1,
+			};
+
+			let layer_size=wgpu::Extent3d{
+				depth_or_array_layers:1,
+				..size
+			};
+			let max_mips=layer_size.max_mips(wgpu::TextureDimension::D2);
+
+			let texture=device.create_texture_with_data(
+				queue,
+				&wgpu::TextureDescriptor{
+					size,
+					mip_level_count:max_mips,
+					sample_count:1,
+					dimension:wgpu::TextureDimension::D2,
+					format:wgpu::TextureFormat::Bc7RgbaUnorm,
+					usage:wgpu::TextureUsages::TEXTURE_BINDING|wgpu::TextureUsages::COPY_DST,
+					label:Some("Squid Texture"),
+					view_formats:&[],
+				},
+				wgpu::util::TextureDataOrder::LayerMajor,
+				&image.data,
+			);
+
+			texture.create_view(&wgpu::TextureViewDescriptor{
+				label:Some("Squid Texture View"),
+				dimension:Some(wgpu::TextureViewDimension::D2),
+				..wgpu::TextureViewDescriptor::default()
+			})
+		};
+
+		let model_pipeline_layout=device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor{
+			label:None,
+			bind_group_layouts:&[
+				&camera_bind_group_layout,
+				&skybox_texture_bind_group_layout,
+				&model_bind_group_layout,
+			],
+			push_constant_ranges:&[],
+		});
+		let sky_pipeline_layout=device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor{
+			label:None,
+			bind_group_layouts:&[
+				&camera_bind_group_layout,
+				&skybox_texture_bind_group_layout,
+			],
+			push_constant_ranges:&[],
+		});
+
+		// Create the render pipelines
+		let sky_pipeline=device.create_render_pipeline(&wgpu::RenderPipelineDescriptor{
+			label:Some("Sky Pipeline"),
+			layout:Some(&sky_pipeline_layout),
+			vertex:wgpu::VertexState{
+				module:&shader,
+				entry_point:"vs_sky",
+				buffers:&[],
+				compilation_options:wgpu::PipelineCompilationOptions::default(),
+			},
+			fragment:Some(wgpu::FragmentState{
+				module:&shader,
+				entry_point:"fs_sky",
+				targets:&[Some(config.view_formats[0].into())],
+				compilation_options:wgpu::PipelineCompilationOptions::default(),
+			}),
+			primitive:wgpu::PrimitiveState{
+				front_face:wgpu::FrontFace::Cw,
+				..Default::default()
+			},
+			depth_stencil:Some(wgpu::DepthStencilState{
+				format:Self::DEPTH_FORMAT,
+				depth_write_enabled:false,
+				depth_compare:wgpu::CompareFunction::LessEqual,
+				stencil:wgpu::StencilState::default(),
+				bias:wgpu::DepthBiasState::default(),
+			}),
+			multisample:wgpu::MultisampleState::default(),
+			multiview:None,
+			cache:None,
+		});
+		let model_pipeline=device.create_render_pipeline(&wgpu::RenderPipelineDescriptor{
+			label:Some("Model Pipeline"),
+			layout:Some(&model_pipeline_layout),
+			vertex:wgpu::VertexState{
+				module:&shader,
+				entry_point:"vs_entity_texture",
+				buffers:&[wgpu::VertexBufferLayout{
+					array_stride:std::mem::size_of::<GraphicsVertex>() as wgpu::BufferAddress,
+					step_mode:wgpu::VertexStepMode::Vertex,
+					attributes:&wgpu::vertex_attr_array![0=>Float32x3,1=>Float32x2,2=>Float32x3,3=>Float32x4],
+				}],
+				compilation_options:wgpu::PipelineCompilationOptions::default(),
+			},
+			fragment:Some(wgpu::FragmentState{
+				module:&shader,
+				entry_point:"fs_entity_texture",
+				targets:&[Some(config.view_formats[0].into())],
+				compilation_options:wgpu::PipelineCompilationOptions::default(),
+			}),
+			primitive:wgpu::PrimitiveState{
+				front_face:wgpu::FrontFace::Cw,
+				cull_mode:Some(wgpu::Face::Front),
+				..Default::default()
+			},
+			depth_stencil:Some(wgpu::DepthStencilState{
+				format:Self::DEPTH_FORMAT,
+				depth_write_enabled:true,
+				depth_compare:wgpu::CompareFunction::LessEqual,
+				stencil:wgpu::StencilState::default(),
+				bias:wgpu::DepthBiasState::default(),
+			}),
+			multisample:wgpu::MultisampleState::default(),
+			multiview:None,
+			cache:None,
+		});
+
+		let camera=GraphicsCamera::default();
+		let camera_uniforms=camera.to_uniform_data(glam::Vec3::ZERO,glam::Vec2::ZERO);
+		let camera_buf=device.create_buffer_init(&wgpu::util::BufferInitDescriptor{
+			label:Some("Camera"),
+			contents:bytemuck::cast_slice(&camera_uniforms),
+			usage:wgpu::BufferUsages::UNIFORM|wgpu::BufferUsages::COPY_DST,
+		});
+		let camera_bind_group=device.create_bind_group(&wgpu::BindGroupDescriptor{
+			layout:&camera_bind_group_layout,
+			entries:&[
+				wgpu::BindGroupEntry{
+					binding:0,
+					resource:camera_buf.as_entire_binding(),
+				},
+			],
+			label:Some("Camera"),
+		});
+
+		let skybox_texture_bind_group=device.create_bind_group(&wgpu::BindGroupDescriptor{
+			layout:&skybox_texture_bind_group_layout,
+			entries:&[
+				wgpu::BindGroupEntry{
+					binding:0,
+					resource:wgpu::BindingResource::TextureView(&skybox_texture_view),
+				},
+				wgpu::BindGroupEntry{
+					binding:1,
+					resource:wgpu::BindingResource::Sampler(&clamp_sampler),
+				},
+			],
+			label:Some("Sky Texture"),
+		});
+
+		let depth_view=Self::create_depth_texture(config,device);
+
+		Self{
+			pipelines:GraphicsPipelines{
+				skybox:sky_pipeline,
+				model:model_pipeline
+			},
+			bind_groups:GraphicsBindGroups{
+				camera:camera_bind_group,
+				skybox_texture:skybox_texture_bind_group,
+			},
+			camera,
+			camera_buf,
+			models:Vec::new(),
+			depth_view,
+			staging_belt:wgpu::util::StagingBelt::new(0x100),
+			bind_group_layouts:GraphicsBindGroupLayouts{model:model_bind_group_layout},
+			samplers:GraphicsSamplers{repeat:repeat_sampler},
+			temp_squid_texture_view:squid_texture_view,
+		}
+	}
+	pub fn resize(
+		&mut self,
+		device:&wgpu::Device,
+		config:&wgpu::SurfaceConfiguration,
+		user_settings:&crate::settings::UserSettings,
+	){
+		self.depth_view=Self::create_depth_texture(config,device);
+		self.camera.screen_size=glam::uvec2(config.width,config.height);
+		self.load_user_settings(user_settings);
+	}
+	pub fn render(
+		&mut self,
+		view:&wgpu::TextureView,
+		device:&wgpu::Device,
+		queue:&wgpu::Queue,
+		frame_state:FrameState,
+	){
+		//TODO:use scheduled frame times to create beautiful smoothing simulation physics extrapolation assuming no input
+
+		let mut encoder=device.create_command_encoder(&wgpu::CommandEncoderDescriptor{label:None});
+
+		// update rotation
+		let camera_uniforms=self.camera.to_uniform_data(
+			frame_state.body.extrapolated_position(frame_state.time).map(Into::<f32>::into).to_array().into(),
+			frame_state.camera.simulate_move_angles(glam::IVec2::ZERO)
+		);
+		self.staging_belt
+			.write_buffer(
+				&mut encoder,
+				&self.camera_buf,
+				0,
+				wgpu::BufferSize::new((camera_uniforms.len() * 4) as wgpu::BufferAddress).unwrap(),
+				device,
+			)
+			.copy_from_slice(bytemuck::cast_slice(&camera_uniforms));
+		//This code only needs to run when the uniforms change
+		/*
+		for model in self.models.iter(){
+			let model_uniforms=get_instances_buffer_data(&model.instances);
+			self.staging_belt
+				.write_buffer(
+					&mut encoder,
+					&model.model_buf,//description of where data will be written when command is executed
+					0,//offset in staging belt?
+					wgpu::BufferSize::new((model_uniforms.len() * 4) as wgpu::BufferAddress).unwrap(),
+					device,
+				)
+				.copy_from_slice(bytemuck::cast_slice(&model_uniforms));
+		}
+		*/
+		self.staging_belt.finish();
+
+		{
+			let mut rpass=encoder.begin_render_pass(&wgpu::RenderPassDescriptor{
+				label:None,
+				color_attachments:&[Some(wgpu::RenderPassColorAttachment{
+					view,
+					resolve_target:None,
+					ops:wgpu::Operations{
+						load:wgpu::LoadOp::Clear(wgpu::Color{
+							r:0.1,
+							g:0.2,
+							b:0.3,
+							a:1.0,
+						}),
+						store:wgpu::StoreOp::Store,
+					},
+				})],
+				depth_stencil_attachment:Some(wgpu::RenderPassDepthStencilAttachment{
+					view:&self.depth_view,
+					depth_ops:Some(wgpu::Operations{
+						load:wgpu::LoadOp::Clear(1.0),
+						store:wgpu::StoreOp::Discard,
+					}),
+					stencil_ops:None,
+				}),
+				timestamp_writes:Default::default(),
+				occlusion_query_set:Default::default(),
+			});
+
+			rpass.set_bind_group(0,&self.bind_groups.camera,&[]);
+			rpass.set_bind_group(1,&self.bind_groups.skybox_texture,&[]);
+
+			rpass.set_pipeline(&self.pipelines.model);
+			for model in &self.models{
+				rpass.set_bind_group(2,&model.bind_group,&[]);
+				rpass.set_vertex_buffer(0,model.vertex_buf.slice(..));
+				rpass.set_index_buffer(model.indices.buf.slice(..),model.indices.format);
+				//TODO: loop over triangle strips
+				rpass.draw_indexed(0..model.indices.count,0,0..model.instance_count);
+			}
+
+			rpass.set_pipeline(&self.pipelines.skybox);
+			rpass.draw(0..3,0..1);
+		}
+
+		queue.submit(std::iter::once(encoder.finish()));
+
+		self.staging_belt.recall();
+	}
+}
+const MODEL_BUFFER_SIZE:usize=4*4 + 12 + 4;//let size=std::mem::size_of::<ModelInstance>();
+const MODEL_BUFFER_SIZE_BYTES:usize=MODEL_BUFFER_SIZE*4;
+fn get_instances_buffer_data(instances:&[GraphicsModelOwned])->Vec<f32>{
+	let mut raw=Vec::with_capacity(MODEL_BUFFER_SIZE*instances.len());
+	for mi in instances{
+		//model transform
+		raw.extend_from_slice(&AsRef::<[f32; 4*4]>::as_ref(&mi.transform)[..]);
+		//normal transform
+		raw.extend_from_slice(AsRef::<[f32; 3]>::as_ref(&mi.normal_transform.x_axis));
+		raw.extend_from_slice(&[0.0]);
+		raw.extend_from_slice(AsRef::<[f32; 3]>::as_ref(&mi.normal_transform.y_axis));
+		raw.extend_from_slice(&[0.0]);
+		raw.extend_from_slice(AsRef::<[f32; 3]>::as_ref(&mi.normal_transform.z_axis));
+		raw.extend_from_slice(&[0.0]);
+		//color
+		raw.extend_from_slice(AsRef::<[f32; 4]>::as_ref(&mi.color.get()));
+	}
+	raw
+}

src/graphics_worker.rs

@@ -0,0 +1,65 @@
+pub enum Instruction{
+	Render(crate::graphics::FrameState),
+	//UpdateModel(crate::graphics::GraphicsModelUpdate),
+	Resize(winit::dpi::PhysicalSize<u32>,crate::settings::UserSettings),
+	ChangeMap(strafesnet_common::map::CompleteMap),
+}
+
+//Ideally the graphics thread worker description is:
+/*
+WorkerDescription{
+	input:Immediate,
+	output:Realtime(PoolOrdering::Ordered(3)),
+}
+*/
+//up to three frames in flight, dropping new frame requests when all three are busy, and dropping output frames when one renders out of order
+
+pub fn new<'a>(
+	mut graphics:crate::graphics::GraphicsState,
+	mut config:wgpu::SurfaceConfiguration,
+	surface:wgpu::Surface<'a>,
+	device:wgpu::Device,
+	queue:wgpu::Queue,
+)->crate::compat_worker::INWorker<'a,Instruction>{
+	let mut resize=None;
+	crate::compat_worker::INWorker::new(move |ins:Instruction|{
+		match ins{
+			Instruction::ChangeMap(map)=>{
+				graphics.clear();
+				graphics.generate_models(&device,&queue,&map);
+			},
+			Instruction::Resize(size,user_settings)=>{
+				resize=Some((size,user_settings));
+			}
+			Instruction::Render(frame_state)=>{
+				if let Some((size,user_settings))=resize.take(){
+					println!("Resizing to {:?}",size);
+					let t0=std::time::Instant::now();
+					config.width=size.width.max(1);
+					config.height=size.height.max(1);
+					surface.configure(&device,&config);
+					graphics.resize(&device,&config,&user_settings);
+					println!("Resize took {:?}",t0.elapsed());
+				}
+				//this has to go deeper somehow
+				let frame=match surface.get_current_texture(){
+					Ok(frame)=>frame,
+					Err(_)=>{
+						surface.configure(&device,&config);
+						surface
+							.get_current_texture()
+							.expect("Failed to acquire next surface texture!")
+					}
+				};
+				let view=frame.texture.create_view(&wgpu::TextureViewDescriptor{
+					format:Some(config.view_formats[0]),
+					..wgpu::TextureViewDescriptor::default()
+				});
+
+				graphics.render(&view,&device,&queue,frame_state);
+
+				frame.present();
+			}
+		}
+	})
+}

src/lib.rs

@@ -1 +0,0 @@
-pub mod framework;

src/main.rs

@@ -1,750 +1,17 @@
-use bytemuck::{Pod, Zeroable};
-use std::{borrow::Cow, time::Instant};
-use wgpu::{util::DeviceExt, AstcBlock, AstcChannel};
+mod file;
+mod setup;
+mod window;
+mod worker;
+mod physics;
+mod graphics;
+mod settings;
+mod face_crawler;
+mod compat_worker;
+mod model_physics;
+mod model_graphics;
+mod physics_worker;
+mod graphics_worker;
 
-const IMAGE_SIZE: u32 = 128;
-
-#[derive(Clone, Copy, Pod, Zeroable)]
-#[repr(C)]
-struct Vertex {
-	pos: [f32; 3],
-	texture: [f32; 2],
-	normal: [f32; 3],
-}
-
-struct Entity {
-	index_count: u32,
-	index_buf: wgpu::Buffer,
-}
-
-//temp?
-struct ModelData {
-	transform: glam::Mat4,
-	vertex_buf: wgpu::Buffer,
-	entities: Vec<Entity>,
-}
-
-struct Model {
-	transform: glam::Mat4,
-	vertex_buf: wgpu::Buffer,
-	entities: Vec<Entity>,
-	bind_group: wgpu::BindGroup,
-	model_buf: wgpu::Buffer,
-}
-
-// Note: we use the Y=up coordinate space in this example.
-struct Camera {
-	time: Instant,
-	pos: glam::Vec3,
-	vel: glam::Vec3,
-	gravity: glam::Vec3,
-	friction: f32,
-	screen_size: (u32, u32),
-	offset: glam::Vec3,
-	fov: f32,
-	yaw: f32,
-	pitch: f32,
-	controls: u32,
-	mv: f32,
-	grounded: bool,
-	walkspeed: f32,
-}
-
-const CONTROL_MOVEFORWARD:u32 = 0b00000001;
-const CONTROL_MOVEBACK:u32 = 0b00000010;
-const CONTROL_MOVERIGHT:u32 = 0b00000100;
-const CONTROL_MOVELEFT:u32 = 0b00001000;
-const CONTROL_MOVEUP:u32 = 0b00010000;
-const CONTROL_MOVEDOWN:u32 = 0b00100000;
-const CONTROL_JUMP:u32 = 0b01000000;
-const CONTROL_ZOOM:u32 = 0b10000000;
-
-const FORWARD_DIR:glam::Vec3 = glam::Vec3::new(0.0,0.0,-1.0);
-const RIGHT_DIR:glam::Vec3 = glam::Vec3::new(1.0,0.0,0.0);
-const UP_DIR:glam::Vec3 = glam::Vec3::new(0.0,1.0,0.0);
-
-fn get_control_dir(controls: u32) -> glam::Vec3{
-	//don't get fancy just do it
-	let mut control_dir:glam::Vec3 = glam::Vec3::new(0.0,0.0,0.0);
-	if controls & CONTROL_MOVEFORWARD == CONTROL_MOVEFORWARD {
-		control_dir+=FORWARD_DIR;
-	}
-	if controls & CONTROL_MOVEBACK == CONTROL_MOVEBACK {
-		control_dir+=-FORWARD_DIR;
-	}
-	if controls & CONTROL_MOVELEFT == CONTROL_MOVELEFT {
-		control_dir+=-RIGHT_DIR;
-	}
-	if controls & CONTROL_MOVERIGHT == CONTROL_MOVERIGHT {
-		control_dir+=RIGHT_DIR;
-	}
-	if controls & CONTROL_MOVEUP == CONTROL_MOVEUP {
-		control_dir+=UP_DIR;
-	}
-	if controls & CONTROL_MOVEDOWN == CONTROL_MOVEDOWN {
-		control_dir+=-UP_DIR;
-	}
-	return control_dir
-}
-
-	#[inline]
-	fn perspective_rh(fov_y_slope: f32, aspect_ratio: f32, z_near: f32, z_far: f32) -> glam::Mat4 {
-		//glam_assert!(z_near > 0.0 && z_far > 0.0);
-		let r = z_far / (z_near - z_far);
-		glam::Mat4::from_cols(
-			glam::Vec4::new(1.0/(fov_y_slope * aspect_ratio), 0.0, 0.0, 0.0),
-			glam::Vec4::new(0.0, 1.0/fov_y_slope, 0.0, 0.0),
-			glam::Vec4::new(0.0, 0.0, r, -1.0),
-			glam::Vec4::new(0.0, 0.0, r * z_near, 0.0),
-		)
-	}
-
-impl Camera {
-	fn to_uniform_data(&self) -> [f32; 16 * 3 + 4] {
-		let aspect = self.screen_size.0 as f32 / self.screen_size.1 as f32;
-		let fov = if self.controls&CONTROL_ZOOM==0 {
-			self.fov
-		}else{
-			self.fov/5.0
-		};
-		let proj = perspective_rh(fov, aspect, 0.5, 1000.0);
-		let proj_inv = proj.inverse();
-		let view = glam::Mat4::from_translation(self.pos+self.offset) * glam::Mat4::from_euler(glam::EulerRot::YXZ, self.yaw, self.pitch, 0f32);
-		let view_inv = view.inverse();
-
-		let mut raw = [0f32; 16 * 3 + 4];
-		raw[..16].copy_from_slice(&AsRef::<[f32; 16]>::as_ref(&proj)[..]);
-		raw[16..32].copy_from_slice(&AsRef::<[f32; 16]>::as_ref(&proj_inv)[..]);
-		raw[32..48].copy_from_slice(&AsRef::<[f32; 16]>::as_ref(&view_inv)[..]);
-		raw[48..52].copy_from_slice(AsRef::<[f32; 4]>::as_ref(&view.col(3)));
-		raw
-	}
-}
-
-pub struct Skybox {
-	camera: Camera,
-	sky_pipeline: wgpu::RenderPipeline,
-	entity_pipeline: wgpu::RenderPipeline,
-	ground_pipeline: wgpu::RenderPipeline,
-	main_bind_group: wgpu::BindGroup,
-	camera_buf: wgpu::Buffer,
-	models: Vec<Model>,
-	depth_view: wgpu::TextureView,
-	staging_belt: wgpu::util::StagingBelt,
-}
-
-impl Skybox {
-	const DEPTH_FORMAT: wgpu::TextureFormat = wgpu::TextureFormat::Depth24Plus;
-
-	fn create_depth_texture(
-		config: &wgpu::SurfaceConfiguration,
-		device: &wgpu::Device,
-	) -> wgpu::TextureView {
-		let depth_texture = device.create_texture(&wgpu::TextureDescriptor {
-			size: wgpu::Extent3d {
-				width: config.width,
-				height: config.height,
-				depth_or_array_layers: 1,
-			},
-			mip_level_count: 1,
-			sample_count: 1,
-			dimension: wgpu::TextureDimension::D2,
-			format: Self::DEPTH_FORMAT,
-			usage: wgpu::TextureUsages::RENDER_ATTACHMENT,
-			label: None,
-			view_formats: &[],
-		});
-
-		depth_texture.create_view(&wgpu::TextureViewDescriptor::default())
-	}
-}
-
-fn get_transform_uniform_data(transform:&glam::Mat4) -> [f32; 4*4] {
-	let mut raw = [0f32; 4*4];
-	raw[0..16].copy_from_slice(&AsRef::<[f32; 4*4]>::as_ref(transform)[..]);
-	raw
-}
-
-fn add_obj(device:&wgpu::Device,modeldatas:& mut Vec<ModelData>,source:&[u8]){
-	let data = obj::ObjData::load_buf(&source[..]).unwrap();
-	let mut vertices = Vec::new();
-	let mut vertex_index = std::collections::HashMap::<obj::IndexTuple,u16>::new();
-	for object in data.objects {
-		let mut entities = Vec::<Entity>::new();
-		for group in object.groups {
-			let mut indices = Vec::new();
-			for poly in group.polys {
-				for end_index in 2..poly.0.len() {
-					for &index in &[0, end_index - 1, end_index] {
-						let vert = poly.0[index];
-						if let Some(&i)=vertex_index.get(&vert){
-							indices.push(i as u16);
-						}else{
-							let i=vertices.len() as u16;
-							vertices.push(Vertex {
-								pos: data.position[vert.0],
-								texture: data.texture[vert.1.unwrap()],
-								normal: data.normal[vert.2.unwrap()],
-							});
-							vertex_index.insert(vert,i);
-							indices.push(i);
-						}
-					}
-				}
-			}
-			let index_buf = device.create_buffer_init(&wgpu::util::BufferInitDescriptor {
-				label: Some("Index"),
-				contents: bytemuck::cast_slice(&indices),
-				usage: wgpu::BufferUsages::INDEX,
-			});
-			entities.push(Entity {
-				index_buf,
-				index_count: indices.len() as u32,
-			});
-		}
-		let vertex_buf = device.create_buffer_init(&wgpu::util::BufferInitDescriptor {
-			label: Some("Vertex"),
-			contents: bytemuck::cast_slice(&vertices),
-			usage: wgpu::BufferUsages::VERTEX,
-		});
-		modeldatas.push(ModelData {
-			transform: glam::Mat4::default(),
-			vertex_buf,
-			entities,
-		})
-	}
-}
-
-impl strafe_client::framework::Example for Skybox {
-	fn optional_features() -> wgpu::Features {
-		wgpu::Features::TEXTURE_COMPRESSION_ASTC
-			| wgpu::Features::TEXTURE_COMPRESSION_ETC2
-			| wgpu::Features::TEXTURE_COMPRESSION_BC
-	}
-
-	fn init(
-		config: &wgpu::SurfaceConfiguration,
-		_adapter: &wgpu::Adapter,
-		device: &wgpu::Device,
-		queue: &wgpu::Queue,
-	) -> Self {
-		let mut modeldatas = Vec::<ModelData>::new();
-		add_obj(device,& mut modeldatas,include_bytes!("../models/teslacyberv3.0.obj"));
-		add_obj(device,& mut modeldatas,include_bytes!("../models/suzanne.obj"));
-		add_obj(device,& mut modeldatas,include_bytes!("../models/teapot.obj"));
-		println!("models.len = {:?}", modeldatas.len());
-		modeldatas[1].transform=glam::Mat4::from_translation(glam::vec3(10.,5.,10.));
-		modeldatas[2].transform=glam::Mat4::from_translation(glam::vec3(-10.,5.,10.));
-
-		let main_bind_group_layout = device.create_bind_group_layout(&wgpu::BindGroupLayoutDescriptor {
-			label: None,
-			entries: &[
-				wgpu::BindGroupLayoutEntry {
-					binding: 0,
-					visibility: wgpu::ShaderStages::VERTEX,
-					ty: wgpu::BindingType::Buffer {
-						ty: wgpu::BufferBindingType::Uniform,
-						has_dynamic_offset: false,
-						min_binding_size: None,
-					},
-					count: None,
-				},
-				wgpu::BindGroupLayoutEntry {
-					binding: 1,
-					visibility: wgpu::ShaderStages::FRAGMENT,
-					ty: wgpu::BindingType::Texture {
-						sample_type: wgpu::TextureSampleType::Float { filterable: true },
-						multisampled: false,
-						view_dimension: wgpu::TextureViewDimension::Cube,
-					},
-					count: None,
-				},
-				wgpu::BindGroupLayoutEntry {
-					binding: 2,
-					visibility: wgpu::ShaderStages::FRAGMENT,
-					ty: wgpu::BindingType::Sampler(wgpu::SamplerBindingType::Filtering),
-					count: None,
-				},
-			],
-		});
-		let model_bind_group_layout = device.create_bind_group_layout(&wgpu::BindGroupLayoutDescriptor {
-			label: None,
-			entries: &[
-				wgpu::BindGroupLayoutEntry {
-					binding: 0,
-					visibility: wgpu::ShaderStages::VERTEX,
-					ty: wgpu::BindingType::Buffer {
-						ty: wgpu::BufferBindingType::Uniform,
-						has_dynamic_offset: false,
-						min_binding_size: None,
-					},
-					count: None,
-				},
-			],
-		});
-
-		// Create the render pipeline
-		let shader = device.create_shader_module(wgpu::ShaderModuleDescriptor {
-			label: None,
-			source: wgpu::ShaderSource::Wgsl(Cow::Borrowed(include_str!("shader.wgsl"))),
-		});
-
-		let camera = Camera {
-			time: Instant::now(),
-			pos: glam::Vec3::new(5.0,0.0,5.0),
-			vel: glam::Vec3::new(0.0,0.0,0.0),
-			gravity: glam::Vec3::new(0.0,-100.0,0.0),
-			friction: 90.0,
-			screen_size: (config.width, config.height),
-			offset: glam::Vec3::new(0.0,4.5,0.0),
-			fov: 1.0, //fov_slope = tan(fov_y/2)
-			pitch: 0.0,
-			yaw: 0.0,
-			mv: 2.7,
-			controls:0,
-			grounded: true,
-			walkspeed: 18.0,
-		};
-		let camera_uniforms = camera.to_uniform_data();
-		let camera_buf = device.create_buffer_init(&wgpu::util::BufferInitDescriptor {
-			label: Some("Camera"),
-			contents: bytemuck::cast_slice(&camera_uniforms),
-			usage: wgpu::BufferUsages::UNIFORM | wgpu::BufferUsages::COPY_DST,
-		});
-
-		let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
-			label: None,
-			bind_group_layouts: &[&main_bind_group_layout, &model_bind_group_layout],
-			push_constant_ranges: &[],
-		});
-
-		// Create the render pipelines
-		let sky_pipeline = device.create_render_pipeline(&wgpu::RenderPipelineDescriptor {
-			label: Some("Sky"),
-			layout: Some(&pipeline_layout),
-			vertex: wgpu::VertexState {
-				module: &shader,
-				entry_point: "vs_sky",
-				buffers: &[],
-			},
-			fragment: Some(wgpu::FragmentState {
-				module: &shader,
-				entry_point: "fs_sky",
-				targets: &[Some(config.view_formats[0].into())],
-			}),
-			primitive: wgpu::PrimitiveState {
-				front_face: wgpu::FrontFace::Cw,
-				..Default::default()
-			},
-			depth_stencil: Some(wgpu::DepthStencilState {
-				format: Self::DEPTH_FORMAT,
-				depth_write_enabled: false,
-				depth_compare: wgpu::CompareFunction::LessEqual,
-				stencil: wgpu::StencilState::default(),
-				bias: wgpu::DepthBiasState::default(),
-			}),
-			multisample: wgpu::MultisampleState::default(),
-			multiview: None,
-		});
-		let entity_pipeline = device.create_render_pipeline(&wgpu::RenderPipelineDescriptor {
-			label: Some("Entity"),
-			layout: Some(&pipeline_layout),
-			vertex: wgpu::VertexState {
-				module: &shader,
-				entry_point: "vs_entity",
-				buffers: &[wgpu::VertexBufferLayout {
-					array_stride: std::mem::size_of::<Vertex>() as wgpu::BufferAddress,
-					step_mode: wgpu::VertexStepMode::Vertex,
-					attributes: &wgpu::vertex_attr_array![0 => Float32x3, 1 => Float32x2, 2 => Float32x3],
-				}],
-			},
-			fragment: Some(wgpu::FragmentState {
-				module: &shader,
-				entry_point: "fs_entity",
-				targets: &[Some(config.view_formats[0].into())],
-			}),
-			primitive: wgpu::PrimitiveState {
-				front_face: wgpu::FrontFace::Cw,
-				..Default::default()
-			},
-			depth_stencil: Some(wgpu::DepthStencilState {
-				format: Self::DEPTH_FORMAT,
-				depth_write_enabled: true,
-				depth_compare: wgpu::CompareFunction::LessEqual,
-				stencil: wgpu::StencilState::default(),
-				bias: wgpu::DepthBiasState::default(),
-			}),
-			multisample: wgpu::MultisampleState::default(),
-			multiview: None,
-		});
-		let ground_pipeline = device.create_render_pipeline(&wgpu::RenderPipelineDescriptor {
-			label: Some("Ground"),
-			layout: Some(&pipeline_layout),
-			vertex: wgpu::VertexState {
-				module: &shader,
-				entry_point: "vs_ground",
-				buffers: &[],
-			},
-			fragment: Some(wgpu::FragmentState {
-				module: &shader,
-				entry_point: "fs_ground",
-				targets: &[Some(config.view_formats[0].into())],
-			}),
-			primitive: wgpu::PrimitiveState {
-				front_face: wgpu::FrontFace::Cw,
-				..Default::default()
-			},
-			depth_stencil: Some(wgpu::DepthStencilState {
-				format: Self::DEPTH_FORMAT,
-				depth_write_enabled: true,
-				depth_compare: wgpu::CompareFunction::LessEqual,
-				stencil: wgpu::StencilState::default(),
-				bias: wgpu::DepthBiasState::default(),
-			}),
-			multisample: wgpu::MultisampleState::default(),
-			multiview: None,
-		});
-
-		let sampler = device.create_sampler(&wgpu::SamplerDescriptor {
-			label: None,
-			address_mode_u: wgpu::AddressMode::ClampToEdge,
-			address_mode_v: wgpu::AddressMode::ClampToEdge,
-			address_mode_w: wgpu::AddressMode::ClampToEdge,
-			mag_filter: wgpu::FilterMode::Linear,
-			min_filter: wgpu::FilterMode::Linear,
-			mipmap_filter: wgpu::FilterMode::Linear,
-			..Default::default()
-		});
-
-		let device_features = device.features();
-
-		let skybox_format = if device_features.contains(wgpu::Features::TEXTURE_COMPRESSION_ASTC) {
-			log::info!("Using ASTC");
-			wgpu::TextureFormat::Astc {
-				block: AstcBlock::B4x4,
-				channel: AstcChannel::UnormSrgb,
-			}
-		} else if device_features.contains(wgpu::Features::TEXTURE_COMPRESSION_ETC2) {
-			log::info!("Using ETC2");
-			wgpu::TextureFormat::Etc2Rgb8UnormSrgb
-		} else if device_features.contains(wgpu::Features::TEXTURE_COMPRESSION_BC) {
-			log::info!("Using BC");
-			wgpu::TextureFormat::Bc1RgbaUnormSrgb
-		} else {
-			log::info!("Using plain");
-			wgpu::TextureFormat::Bgra8UnormSrgb
-		};
-
-		let size = wgpu::Extent3d {
-			width: IMAGE_SIZE,
-			height: IMAGE_SIZE,
-			depth_or_array_layers: 6,
-		};
-
-		let layer_size = wgpu::Extent3d {
-			depth_or_array_layers: 1,
-			..size
-		};
-		let max_mips = layer_size.max_mips(wgpu::TextureDimension::D2);
-
-		log::debug!(
-			"Copying {:?} skybox images of size {}, {}, 6 with {} mips to gpu",
-			skybox_format,
-			IMAGE_SIZE,
-			IMAGE_SIZE,
-			max_mips,
-		);
-
-		let bytes = match skybox_format {
-			wgpu::TextureFormat::Astc {
-				block: AstcBlock::B4x4,
-				channel: AstcChannel::UnormSrgb,
-			} => &include_bytes!("../images/astc.dds")[..],
-			wgpu::TextureFormat::Etc2Rgb8UnormSrgb => &include_bytes!("../images/etc2.dds")[..],
-			wgpu::TextureFormat::Bc1RgbaUnormSrgb => &include_bytes!("../images/bc1.dds")[..],
-			wgpu::TextureFormat::Bgra8UnormSrgb => &include_bytes!("../images/bgra.dds")[..],
-			_ => unreachable!(),
-		};
-
-		let image = ddsfile::Dds::read(&mut std::io::Cursor::new(&bytes)).unwrap();
-
-		let texture = device.create_texture_with_data(
-			queue,
-			&wgpu::TextureDescriptor {
-				size,
-				mip_level_count: max_mips,
-				sample_count: 1,
-				dimension: wgpu::TextureDimension::D2,
-				format: skybox_format,
-				usage: wgpu::TextureUsages::TEXTURE_BINDING | wgpu::TextureUsages::COPY_DST,
-				label: None,
-				view_formats: &[],
-			},
-			&image.data,
-		);
-
-		let texture_view = texture.create_view(&wgpu::TextureViewDescriptor {
-			label: None,
-			dimension: Some(wgpu::TextureViewDimension::Cube),
-			..wgpu::TextureViewDescriptor::default()
-		});
-		let main_bind_group = device.create_bind_group(&wgpu::BindGroupDescriptor {
-			layout: &main_bind_group_layout,
-			entries: &[
-				wgpu::BindGroupEntry {
-					binding: 0,
-					resource: camera_buf.as_entire_binding(),
-				},
-				wgpu::BindGroupEntry {
-					binding: 1,
-					resource: wgpu::BindingResource::TextureView(&texture_view),
-				},
-				wgpu::BindGroupEntry {
-					binding: 2,
-					resource: wgpu::BindingResource::Sampler(&sampler),
-				},
-			],
-			label: Some("Camera"),
-		});
-
-		//drain the modeldata vec so entities can be /moved/ to models.entities
-		let mut models = Vec::<Model>::with_capacity(modeldatas.len());
-		for (i,modeldata) in modeldatas.drain(..).enumerate() {
-			let model_uniforms = get_transform_uniform_data(&modeldata.transform);
-			let model_buf = device.create_buffer_init(&wgpu::util::BufferInitDescriptor {
-				label: Some(format!("Model{}",i).as_str()),
-				contents: bytemuck::cast_slice(&model_uniforms),
-				usage: wgpu::BufferUsages::UNIFORM | wgpu::BufferUsages::COPY_DST,
-			});
-			let model_bind_group = device.create_bind_group(&wgpu::BindGroupDescriptor {
-				layout: &model_bind_group_layout,
-				entries: &[
-					wgpu::BindGroupEntry {
-						binding: 0,
-						resource: model_buf.as_entire_binding(),
-					},
-				],
-				label: Some(format!("Model{}",i).as_str()),
-			});
-			//all of these are being moved here
-			models.push(Model{
-				transform: modeldata.transform,
-				vertex_buf:modeldata.vertex_buf,
-				entities: modeldata.entities,
-				bind_group: model_bind_group,
-				model_buf,
-			})
-		}
-
-		let depth_view = Self::create_depth_texture(config, device);
-
-		Skybox {
-			camera,
-			sky_pipeline,
-			entity_pipeline,
-			ground_pipeline,
-			main_bind_group,
-			camera_buf,
-			models,
-			depth_view,
-			staging_belt: wgpu::util::StagingBelt::new(0x100),
-		}
-	}
-
-	#[allow(clippy::single_match)]
-	fn update(&mut self, event: winit::event::WindowEvent) {
-		match event {
-			winit::event::WindowEvent::KeyboardInput {
-				input:
-					winit::event::KeyboardInput {
-						state,
-						virtual_keycode: Some(keycode),
-						..
-					},
-				..
-			} => {
-				match (state,keycode) {
-					(k,winit::event::VirtualKeyCode::W) => match k {
-						winit::event::ElementState::Pressed => self.camera.controls|=CONTROL_MOVEFORWARD,
-						winit::event::ElementState::Released => self.camera.controls&=!CONTROL_MOVEFORWARD,
-					}
-					(k,winit::event::VirtualKeyCode::A) => match k {
-						winit::event::ElementState::Pressed => self.camera.controls|=CONTROL_MOVELEFT,
-						winit::event::ElementState::Released => self.camera.controls&=!CONTROL_MOVELEFT,
-					}
-					(k,winit::event::VirtualKeyCode::S) => match k {
-						winit::event::ElementState::Pressed => self.camera.controls|=CONTROL_MOVEBACK,
-						winit::event::ElementState::Released => self.camera.controls&=!CONTROL_MOVEBACK,
-					}
-					(k,winit::event::VirtualKeyCode::D) => match k {
-						winit::event::ElementState::Pressed => self.camera.controls|=CONTROL_MOVERIGHT,
-						winit::event::ElementState::Released => self.camera.controls&=!CONTROL_MOVERIGHT,
-					}
-					(k,winit::event::VirtualKeyCode::E) => match k {
-						winit::event::ElementState::Pressed => self.camera.controls|=CONTROL_MOVEUP,
-						winit::event::ElementState::Released => self.camera.controls&=!CONTROL_MOVEUP,
-					}
-					(k,winit::event::VirtualKeyCode::Q) => match k {
-						winit::event::ElementState::Pressed => self.camera.controls|=CONTROL_MOVEDOWN,
-						winit::event::ElementState::Released => self.camera.controls&=!CONTROL_MOVEDOWN,
-					}
-					(k,winit::event::VirtualKeyCode::Space) => match k {
-						winit::event::ElementState::Pressed => self.camera.controls|=CONTROL_JUMP,
-						winit::event::ElementState::Released => self.camera.controls&=!CONTROL_JUMP,
-					}
-					(k,winit::event::VirtualKeyCode::Z) => match k {
-						winit::event::ElementState::Pressed => self.camera.controls|=CONTROL_ZOOM,
-						winit::event::ElementState::Released => self.camera.controls&=!CONTROL_ZOOM,
-					}
-					_ => (),
-				}
-			}
-			_ => {}
-		}
-	}
-
-	fn move_mouse(&mut self, delta: (f64,f64)) {
-		self.camera.pitch=(self.camera.pitch as f64+delta.1/-2048.) as f32;
-		self.camera.yaw=(self.camera.yaw as f64+delta.0/-2048.) as f32;
-	}
-
-	fn resize(
-		&mut self,
-		config: &wgpu::SurfaceConfiguration,
-		device: &wgpu::Device,
-		_queue: &wgpu::Queue,
-	) {
-		self.depth_view = Self::create_depth_texture(config, device);
-		self.camera.screen_size = (config.width, config.height);
-	}
-
-	fn render(
-		&mut self,
-		view: &wgpu::TextureView,
-		device: &wgpu::Device,
-		queue: &wgpu::Queue,
-		_spawner: &strafe_client::framework::Spawner,
-	) {
-		let time = Instant::now();
-
-		//physique
-		let dt=(time-self.camera.time).as_secs_f32();
-		self.camera.time=time;
-		let camera_mat=glam::Mat3::from_euler(glam::EulerRot::YXZ,self.camera.yaw,0f32,0f32);
-		let control_dir=camera_mat*get_control_dir(self.camera.controls&(CONTROL_MOVELEFT|CONTROL_MOVERIGHT|CONTROL_MOVEFORWARD|CONTROL_MOVEBACK)).normalize_or_zero();
-		let d=self.camera.vel.dot(control_dir);
-		if d<self.camera.mv {
-			self.camera.vel+=(self.camera.mv-d)*control_dir;
-		}
-		self.camera.vel+=self.camera.gravity*dt;
-		self.camera.pos+=self.camera.vel*dt;
-		if self.camera.pos.y<0.0{
-			self.camera.pos.y=0.0;
-			self.camera.vel.y=0.0;
-			self.camera.grounded=true;
-		}
-		if self.camera.grounded&&(self.camera.controls&CONTROL_JUMP)!=0 {
-			self.camera.grounded=false;
-			self.camera.vel+=glam::Vec3::new(0.0,0.715588/2.0*100.0,0.0);
-		}
-		if self.camera.grounded {
-			let applied_friction=self.camera.friction*dt;
-			let targetv=control_dir*self.camera.walkspeed;
-			let diffv=targetv-self.camera.vel;
-			if applied_friction*applied_friction<diffv.length_squared() {
-				self.camera.vel+=applied_friction*diffv.normalize();
-			} else {
-				self.camera.vel=targetv;
-			}
-		}
-
-		let mut encoder =
-			device.create_command_encoder(&wgpu::CommandEncoderDescriptor { label: None });
-
-		// update rotation
-		let camera_uniforms = self.camera.to_uniform_data();
-		self.staging_belt
-			.write_buffer(
-				&mut encoder,
-				&self.camera_buf,
-				0,
-				wgpu::BufferSize::new((camera_uniforms.len() * 4) as wgpu::BufferAddress).unwrap(),
-				device,
-			)
-			.copy_from_slice(bytemuck::cast_slice(&camera_uniforms));
-		//This code only needs to run when the uniforms change
-		for model in self.models.iter() {
-			let model_uniforms = get_transform_uniform_data(&model.transform);
-			self.staging_belt
-				.write_buffer(
-					&mut encoder,
-					&model.model_buf,//description of where data will be written when command is executed
-					0,//offset in staging belt?
-					wgpu::BufferSize::new((model_uniforms.len() * 4) as wgpu::BufferAddress).unwrap(),
-					device,
-				)
-				.copy_from_slice(bytemuck::cast_slice(&model_uniforms));
-		}
-		self.staging_belt.finish();
-
-		{
-			let mut rpass = encoder.begin_render_pass(&wgpu::RenderPassDescriptor {
-				label: None,
-				color_attachments: &[Some(wgpu::RenderPassColorAttachment {
-					view,
-					resolve_target: None,
-					ops: wgpu::Operations {
-						load: wgpu::LoadOp::Clear(wgpu::Color {
-							r: 0.1,
-							g: 0.2,
-							b: 0.3,
-							a: 1.0,
-						}),
-						store: true,
-					},
-				})],
-				depth_stencil_attachment: Some(wgpu::RenderPassDepthStencilAttachment {
-					view: &self.depth_view,
-					depth_ops: Some(wgpu::Operations {
-						load: wgpu::LoadOp::Clear(1.0),
-						store: false,
-					}),
-					stencil_ops: None,
-				}),
-			});
-
-			rpass.set_bind_group(0, &self.main_bind_group, &[]);
-
-			rpass.set_pipeline(&self.entity_pipeline);
-			for model in self.models.iter() {
-				rpass.set_bind_group(1, &model.bind_group, &[]);
-				rpass.set_vertex_buffer(0, model.vertex_buf.slice(..));
-
-				for entity in model.entities.iter() {
-					rpass.set_index_buffer(entity.index_buf.slice(..), wgpu::IndexFormat::Uint16);
-					rpass.draw_indexed(0..entity.index_count, 0, 0..1);
-				}
-			}
-
-			rpass.set_pipeline(&self.ground_pipeline);
-			//rpass.set_index_buffer(&[0u16,1,2,1,2,3][..] as wgpu::BufferSlice, wgpu::IndexFormat::Uint16);
-			//rpass.draw_indexed(0..4, 0, 0..1);
-			rpass.draw(0..6, 0..1);
-
-			rpass.set_pipeline(&self.sky_pipeline);
-			rpass.draw(0..3, 0..1);
-		}
-
-		queue.submit(std::iter::once(encoder.finish()));
-
-		self.staging_belt.recall();
-	}
-}
-
-fn main() {
-	strafe_client::framework::run::<Skybox>(
-		format!("Strafe Client v{}",
-			env!("CARGO_PKG_VERSION")
-		).as_str()
-	);
+fn main(){
+	setup::setup_and_start(format!("Strafe Client v{}",env!("CARGO_PKG_VERSION")));
 }

src/model_graphics.rs

@@ -0,0 +1,48 @@
+use bytemuck::{Pod,Zeroable};
+use strafesnet_common::model::{IndexedVertex,PolygonGroup,RenderConfigId};
+#[derive(Clone,Copy,Pod,Zeroable)]
+#[repr(C)]
+pub struct GraphicsVertex{
+	pub pos:[f32;3],
+	pub tex:[f32;2],
+	pub normal:[f32;3],
+	pub color:[f32;4],
+}
+#[derive(Clone,Copy,id::Id)]
+pub struct IndexedGraphicsMeshOwnedRenderConfigId(u32);
+pub struct IndexedGraphicsMeshOwnedRenderConfig{
+	pub unique_pos:Vec<[f32;3]>,
+	pub unique_tex:Vec<[f32;2]>,
+	pub unique_normal:Vec<[f32;3]>,
+	pub unique_color:Vec<[f32;4]>,
+	pub unique_vertices:Vec<IndexedVertex>,
+	pub render_config:RenderConfigId,
+	pub polys:PolygonGroup,
+	pub instances:Vec<GraphicsModelOwned>,
+}
+pub enum Indices{
+	U32(Vec<u32>),
+	U16(Vec<u16>),
+}
+pub struct GraphicsMeshOwnedRenderConfig{
+	pub vertices:Vec<GraphicsVertex>,
+	pub indices:Indices,
+	pub render_config:RenderConfigId,
+	pub instances:Vec<GraphicsModelOwned>,
+}
+#[derive(Clone,Copy,PartialEq,id::Id)]
+pub struct GraphicsModelColor4(glam::Vec4);
+impl std::hash::Hash for GraphicsModelColor4{
+	fn hash<H:std::hash::Hasher>(&self,state:&mut H) {
+		for &f in self.0.as_ref(){
+			bytemuck::cast::<f32,u32>(f).hash(state);
+		}
+	}
+}
+impl Eq for GraphicsModelColor4{}
+#[derive(Clone)]
+pub struct GraphicsModelOwned{
+	pub transform:glam::Mat4,
+	pub normal_transform:glam::Mat3,
+	pub color:GraphicsModelColor4,
+}

src/physics_worker.rs

@@ -0,0 +1,242 @@
+use strafesnet_common::mouse::MouseState;
+use strafesnet_common::physics::Instruction as PhysicsInputInstruction;
+use strafesnet_common::integer::Time;
+use strafesnet_common::instruction::TimedInstruction;
+use strafesnet_common::timer::{Scaled,Timer,TimerState};
+use mouse_interpolator::MouseInterpolator;
+
+#[derive(Debug)]
+pub enum InputInstruction{
+	MoveMouse(glam::IVec2),
+	MoveRight(bool),
+	MoveUp(bool),
+	MoveBack(bool),
+	MoveLeft(bool),
+	MoveDown(bool),
+	MoveForward(bool),
+	Jump(bool),
+	Zoom(bool),
+	ResetAndRestart,
+	ResetAndSpawn(strafesnet_common::gameplay_modes::ModeId,strafesnet_common::gameplay_modes::StageId),
+	PracticeFly,
+}
+pub enum Instruction{
+	Input(InputInstruction),
+	Render,
+	Resize(winit::dpi::PhysicalSize<u32>),
+	ChangeMap(strafesnet_common::map::CompleteMap),
+	//SetPaused is not an InputInstruction: the physics doesn't know that it's paused.
+	SetPaused(bool),
+	//Graphics(crate::graphics_worker::Instruction),
+}
+mod mouse_interpolator{
+	use super::*;
+	//TODO: move this or tab
+pub struct MouseInterpolator{
+	//"PlayerController"
+	user_settings:crate::settings::UserSettings,
+	//"MouseInterpolator"
+	timeline:std::collections::VecDeque<TimedInstruction<PhysicsInputInstruction>>,
+	last_mouse_time:Time,//this value is pre-transformed to simulation time
+	mouse_blocking:bool,
+	//"Simulation"
+	timer:Timer<Scaled>,
+	physics:crate::physics::PhysicsContext,
+
+}
+impl MouseInterpolator{
+	pub fn new(
+		physics:crate::physics::PhysicsContext,
+		user_settings:crate::settings::UserSettings,
+	)->MouseInterpolator{
+		MouseInterpolator{
+			mouse_blocking:true,
+			last_mouse_time:physics.get_next_mouse().time,
+			timeline:std::collections::VecDeque::new(),
+			timer:Timer::from_state(Scaled::identity(),false),
+			physics,
+			user_settings,
+		}
+	}
+	fn push_mouse_instruction(&mut self,ins:&TimedInstruction<Instruction>,m:glam::IVec2){
+		if self.mouse_blocking{
+			//tell the game state which is living in the past about its future
+			self.timeline.push_front(TimedInstruction{
+				time:self.last_mouse_time,
+				instruction:PhysicsInputInstruction::SetNextMouse(MouseState{time:self.timer.time(ins.time),pos:m}),
+			});
+		}else{
+			//mouse has just started moving again after being still for longer than 10ms.
+			//replace the entire mouse interpolation state to avoid an intermediate state with identical m0.t m1.t timestamps which will divide by zero
+			self.timeline.push_front(TimedInstruction{
+				time:self.last_mouse_time,
+				instruction:PhysicsInputInstruction::ReplaceMouse(
+					MouseState{time:self.last_mouse_time,pos:self.physics.get_next_mouse().pos},
+					MouseState{time:self.timer.time(ins.time),pos:m}
+				),
+			});
+			//delay physics execution until we have an interpolation target
+			self.mouse_blocking=true;
+		}
+		self.last_mouse_time=self.timer.time(ins.time);
+	}
+	fn push(&mut self,time:Time,phys_input:PhysicsInputInstruction){
+		//This is always a non-mouse event
+		self.timeline.push_back(TimedInstruction{
+			time:self.timer.time(time),
+			instruction:phys_input,
+		});
+	}
+	/// returns should_empty_queue
+	/// may or may not mutate internal state XD!
+	fn map_instruction(&mut self,ins:&TimedInstruction<Instruction>)->bool{
+		let mut update_mouse_blocking=true;
+		match &ins.instruction{
+			Instruction::Input(input_instruction)=>match input_instruction{
+				&InputInstruction::MoveMouse(m)=>{
+					if !self.timer.is_paused(){
+						self.push_mouse_instruction(ins,m);
+					}
+					update_mouse_blocking=false;
+				},
+				&InputInstruction::MoveForward(s)=>self.push(ins.time,PhysicsInputInstruction::SetMoveForward(s)),
+				&InputInstruction::MoveLeft(s)=>self.push(ins.time,PhysicsInputInstruction::SetMoveLeft(s)),
+				&InputInstruction::MoveBack(s)=>self.push(ins.time,PhysicsInputInstruction::SetMoveBack(s)),
+				&InputInstruction::MoveRight(s)=>self.push(ins.time,PhysicsInputInstruction::SetMoveRight(s)),
+				&InputInstruction::MoveUp(s)=>self.push(ins.time,PhysicsInputInstruction::SetMoveUp(s)),
+				&InputInstruction::MoveDown(s)=>self.push(ins.time,PhysicsInputInstruction::SetMoveDown(s)),
+				&InputInstruction::Jump(s)=>self.push(ins.time,PhysicsInputInstruction::SetJump(s)),
+				&InputInstruction::Zoom(s)=>self.push(ins.time,PhysicsInputInstruction::SetZoom(s)),
+				&InputInstruction::ResetAndSpawn(mode_id,stage_id)=>{
+					self.push(ins.time,PhysicsInputInstruction::Reset);
+					self.push(ins.time,PhysicsInputInstruction::SetSensitivity(self.user_settings.calculate_sensitivity()));
+					self.push(ins.time,PhysicsInputInstruction::Spawn(mode_id,stage_id));
+				},
+				InputInstruction::ResetAndRestart=>{
+					self.push(ins.time,PhysicsInputInstruction::Reset);
+					self.push(ins.time,PhysicsInputInstruction::SetSensitivity(self.user_settings.calculate_sensitivity()));
+					self.push(ins.time,PhysicsInputInstruction::Restart);
+				},
+				InputInstruction::PracticeFly=>self.push(ins.time,PhysicsInputInstruction::PracticeFly),
+			},
+			//do these really need to idle the physics?
+			//sending None dumps the instruction queue
+			Instruction::ChangeMap(_)=>self.push(ins.time,PhysicsInputInstruction::Idle),
+			Instruction::Resize(_)=>self.push(ins.time,PhysicsInputInstruction::Idle),
+			Instruction::Render=>self.push(ins.time,PhysicsInputInstruction::Idle),
+			&Instruction::SetPaused(paused)=>{
+				if let Err(e)=self.timer.set_paused(ins.time,paused){
+					println!("Cannot pause: {e}");
+				}
+				self.push(ins.time,PhysicsInputInstruction::Idle);
+			},
+		}
+		if update_mouse_blocking{
+			//this returns the bool for us
+			self.update_mouse_blocking(ins.time)
+		}else{
+			//do flush that queue
+			true
+		}
+	}
+	/// must check if self.mouse_blocking==true before calling!
+	fn unblock_mouse(&mut self,time:Time){
+		//push an event to extrapolate no movement from
+		self.timeline.push_front(TimedInstruction{
+			time:self.last_mouse_time,
+			instruction:PhysicsInputInstruction::SetNextMouse(MouseState{time:self.timer.time(time),pos:self.physics.get_next_mouse().pos}),
+		});
+		self.last_mouse_time=self.timer.time(time);
+		//stop blocking. the mouse is not moving so the physics does not need to live in the past and wait for interpolation targets.
+		self.mouse_blocking=false;
+	}
+	fn update_mouse_blocking(&mut self,time:Time)->bool{
+		if self.mouse_blocking{
+			//assume the mouse has stopped moving after 10ms.
+			//shitty mice are 125Hz which is 8ms so this should cover that.
+			//setting this to 100us still doesn't print even though it's 10x lower than the polling rate,
+			//so mouse events are probably not handled separately from drawing and fire right before it :(
+			if Time::from_millis(10)<self.timer.time(time)-self.physics.get_next_mouse().time{
+				self.unblock_mouse(time);
+				true
+			}else{
+				false
+			}
+		}else{
+			//keep this up to date so that it can be used as a known-timestamp
+			//that the mouse was not moving when the mouse starts moving again
+			self.last_mouse_time=self.timer.time(time);
+			true
+		}
+	}
+	fn empty_queue(&mut self){
+		while let Some(instruction)=self.timeline.pop_front(){
+			self.physics.run_input_instruction(instruction);
+		}
+	}
+	pub fn handle_instruction(&mut self,ins:&TimedInstruction<Instruction>){
+		let should_empty_queue=self.map_instruction(ins);
+		if should_empty_queue{
+			self.empty_queue();
+		}
+	}
+	pub fn get_frame_state(&self,time:Time)->crate::graphics::FrameState{
+		crate::graphics::FrameState{
+			body:self.physics.camera_body(),
+			camera:self.physics.camera(),
+			time:self.timer.time(time),
+		}
+	}
+	pub fn change_map(&mut self,time:Time,map:&strafesnet_common::map::CompleteMap){
+		//dump any pending interpolation state
+		if self.mouse_blocking{
+			self.unblock_mouse(time);
+		}
+		self.empty_queue();
+
+		//doing it like this to avoid doing PhysicsInstruction::ChangeMap(Rc<CompleteMap>)
+		self.physics.generate_models(&map);
+
+		//use the standard input interface so the instructions are written out to bots
+		self.handle_instruction(&TimedInstruction{
+			time:self.timer.time(time),
+			instruction:Instruction::Input(InputInstruction::ResetAndSpawn(
+				strafesnet_common::gameplay_modes::ModeId::MAIN,
+				strafesnet_common::gameplay_modes::StageId::FIRST,
+			)),
+		});
+	}
+	pub const fn user_settings(&self)->&crate::settings::UserSettings{
+		&self.user_settings
+	}
+}
+}
+
+pub fn new<'a>(
+	mut graphics_worker:crate::compat_worker::INWorker<'a,crate::graphics_worker::Instruction>,
+	user_settings:crate::settings::UserSettings,
+)->crate::compat_worker::QNWorker<'a,TimedInstruction<Instruction>>{
+	let physics=crate::physics::PhysicsContext::default();
+	let mut interpolator=MouseInterpolator::new(
+		physics,
+		user_settings
+	);
+	crate::compat_worker::QNWorker::new(move |ins:TimedInstruction<Instruction>|{
+		interpolator.handle_instruction(&ins);
+		match ins.instruction{
+			Instruction::Render=>{
+				let frame_state=interpolator.get_frame_state(ins.time);
+				graphics_worker.send(crate::graphics_worker::Instruction::Render(frame_state)).unwrap();
+			},
+			Instruction::Resize(size)=>{
+				graphics_worker.send(crate::graphics_worker::Instruction::Resize(size,interpolator.user_settings().clone())).unwrap();
+			},
+			Instruction::ChangeMap(map)=>{
+				interpolator.change_map(ins.time,&map);
+				graphics_worker.send(crate::graphics_worker::Instruction::ChangeMap(map)).unwrap();
+			},
+			Instruction::Input(_)=>(),
+			Instruction::SetPaused(_)=>(),
+		}
+	})
+}

src/settings.rs

@@ -0,0 +1,139 @@
+use strafesnet_common::integer::{Ratio64,Ratio64Vec2};
+#[derive(Clone)]
+struct Ratio{
+	ratio:f64,
+}
+#[derive(Clone)]
+enum DerivedFov{
+	FromScreenAspect,
+	FromAspect(Ratio),
+}
+#[derive(Clone)]
+enum Fov{
+	Exactly{x:f64,y:f64},
+	SpecifyXDeriveY{x:f64,y:DerivedFov},
+	SpecifyYDeriveX{x:DerivedFov,y:f64},
+}
+impl Default for Fov{
+	fn default()->Self{
+		Fov::SpecifyYDeriveX{x:DerivedFov::FromScreenAspect,y:1.0}
+	}
+}
+#[derive(Clone)]
+enum DerivedSensitivity{
+	FromRatio(Ratio64),
+}
+#[derive(Clone)]
+enum Sensitivity{
+	Exactly{x:Ratio64,y:Ratio64},
+	SpecifyXDeriveY{x:Ratio64,y:DerivedSensitivity},
+	SpecifyYDeriveX{x:DerivedSensitivity,y:Ratio64},
+}
+impl Default for Sensitivity{
+	fn default()->Self{
+		Sensitivity::SpecifyXDeriveY{x:Ratio64::ONE*524288,y:DerivedSensitivity::FromRatio(Ratio64::ONE)}
+	}
+}
+
+#[derive(Default,Clone)]
+pub struct UserSettings{
+	fov:Fov,
+	sensitivity:Sensitivity,
+}
+impl UserSettings{
+	pub fn calculate_fov(&self,zoom:f64,screen_size:&glam::UVec2)->glam::DVec2{
+		zoom*match &self.fov{
+			&Fov::Exactly{x,y}=>glam::dvec2(x,y),
+			Fov::SpecifyXDeriveY{x,y}=>match y{
+				DerivedFov::FromScreenAspect=>glam::dvec2(*x,x*(screen_size.y as f64/screen_size.x as f64)),
+				DerivedFov::FromAspect(ratio)=>glam::dvec2(*x,x*ratio.ratio),
+			},
+			Fov::SpecifyYDeriveX{x,y}=>match x{
+				DerivedFov::FromScreenAspect=>glam::dvec2(y*(screen_size.x as f64/screen_size.y as f64),*y),
+				DerivedFov::FromAspect(ratio)=>glam::dvec2(y*ratio.ratio,*y),
+			},
+		}
+	}
+	pub fn calculate_sensitivity(&self)->Ratio64Vec2{
+		match &self.sensitivity{
+			Sensitivity::Exactly{x,y}=>Ratio64Vec2::new(x.clone(),y.clone()),
+			Sensitivity::SpecifyXDeriveY{x,y}=>match y{
+				DerivedSensitivity::FromRatio(ratio)=>Ratio64Vec2::new(x.clone(),x.mul_ref(ratio)),
+			}
+			Sensitivity::SpecifyYDeriveX{x,y}=>match x{
+				DerivedSensitivity::FromRatio(ratio)=>Ratio64Vec2::new(y.mul_ref(ratio),y.clone()),
+			}
+		}
+	}
+}
+
+/*
+//sensitivity is raw input dots (i.e. dpi = dots per inch) to radians conversion factor
+sensitivity_x=0.001
+sensitivity_y_from_x_ratio=1
+Sensitivity::DeriveY{x:0.0.001,y:DerivedSensitivity{ratio:1.0}}
+*/
+
+pub fn read_user_settings()->UserSettings{
+	let mut cfg=configparser::ini::Ini::new();
+	if let Ok(_)=cfg.load("settings.conf"){
+		let (cfg_fov_x,cfg_fov_y)=(cfg.getfloat("camera","fov_x"),cfg.getfloat("camera","fov_y"));
+		let fov=match(cfg_fov_x,cfg_fov_y){
+			(Ok(Some(fov_x)),Ok(Some(fov_y)))=>Fov::Exactly {
+				x:fov_x,
+				y:fov_y
+			},
+			(Ok(Some(fov_x)),Ok(None))=>Fov::SpecifyXDeriveY{
+				x:fov_x,
+				y:if let Ok(Some(fov_y_from_x_ratio))=cfg.getfloat("camera","fov_y_from_x_ratio"){
+					DerivedFov::FromAspect(Ratio{ratio:fov_y_from_x_ratio})
+				}else{
+					DerivedFov::FromScreenAspect
+				}
+			},
+			(Ok(None),Ok(Some(fov_y)))=>Fov::SpecifyYDeriveX{
+				x:if let Ok(Some(fov_x_from_y_ratio))=cfg.getfloat("camera","fov_x_from_y_ratio"){
+					DerivedFov::FromAspect(Ratio{ratio:fov_x_from_y_ratio})
+				}else{
+					DerivedFov::FromScreenAspect
+				},
+				y:fov_y,
+			},
+			_=>{
+				Fov::default()
+			},
+		};
+		let (cfg_sensitivity_x,cfg_sensitivity_y)=(cfg.getfloat("camera","sensitivity_x"),cfg.getfloat("camera","sensitivity_y"));
+		let sensitivity=match(cfg_sensitivity_x,cfg_sensitivity_y){
+			(Ok(Some(sensitivity_x)),Ok(Some(sensitivity_y)))=>Sensitivity::Exactly {
+				x:Ratio64::try_from(sensitivity_x).unwrap(),
+				y:Ratio64::try_from(sensitivity_y).unwrap(),
+			},
+			(Ok(Some(sensitivity_x)),Ok(None))=>Sensitivity::SpecifyXDeriveY{
+				x:Ratio64::try_from(sensitivity_x).unwrap(),
+				y:if let Ok(Some(sensitivity_y_from_x_ratio))=cfg.getfloat("camera","sensitivity_y_from_x_ratio"){
+					DerivedSensitivity::FromRatio(Ratio64::try_from(sensitivity_y_from_x_ratio).unwrap())
+				}else{
+					DerivedSensitivity::FromRatio(Ratio64::ONE)
+				},
+			},
+			(Ok(None),Ok(Some(sensitivity_y)))=>Sensitivity::SpecifyYDeriveX{
+				x:if let Ok(Some(sensitivity_x_from_y_ratio))=cfg.getfloat("camera","sensitivity_x_from_y_ratio"){
+					DerivedSensitivity::FromRatio(Ratio64::try_from(sensitivity_x_from_y_ratio).unwrap())
+				}else{
+					DerivedSensitivity::FromRatio(Ratio64::ONE)
+				},
+				y:Ratio64::try_from(sensitivity_y).unwrap(),
+			},
+			_=>{
+				Sensitivity::default()
+			},
+		};
+		UserSettings{
+			fov,
+			sensitivity,
+		}
+	}else{
+		UserSettings::default()
+	}
+}

src/setup.rs

@@ -0,0 +1,292 @@
+use crate::window::WindowInstruction;
+use strafesnet_common::instruction::TimedInstruction;
+use strafesnet_common::integer;
+
+fn optional_features()->wgpu::Features{
+	wgpu::Features::TEXTURE_COMPRESSION_ASTC
+	|wgpu::Features::TEXTURE_COMPRESSION_ETC2
+}
+fn required_features()->wgpu::Features{
+	wgpu::Features::TEXTURE_COMPRESSION_BC
+}
+fn required_downlevel_capabilities()->wgpu::DownlevelCapabilities{
+	wgpu::DownlevelCapabilities{
+		flags:wgpu::DownlevelFlags::empty(),
+		shader_model:wgpu::ShaderModel::Sm5,
+		..wgpu::DownlevelCapabilities::default()
+	}
+}
+pub fn required_limits()->wgpu::Limits{
+	wgpu::Limits::default()
+}
+
+struct SetupContextPartial1{
+	backends:wgpu::Backends,
+	instance:wgpu::Instance,
+}
+fn create_window(title:&str,event_loop:&winit::event_loop::EventLoop<()>)->Result<winit::window::Window,winit::error::OsError>{
+	let mut attr=winit::window::WindowAttributes::default();
+	attr=attr.with_title(title);
+	#[cfg(windows_OFF)] // TODO
+	{
+		use winit::platform::windows::WindowBuilderExtWindows;
+		builder=builder.with_no_redirection_bitmap(true);
+	}
+	event_loop.create_window(attr)
+}
+fn create_instance()->SetupContextPartial1{
+	let backends=wgpu::util::backend_bits_from_env().unwrap_or_else(wgpu::Backends::all);
+	let dx12_shader_compiler=wgpu::util::dx12_shader_compiler_from_env().unwrap_or_default();
+	SetupContextPartial1{
+		backends,
+		instance:wgpu::Instance::new(wgpu::InstanceDescriptor{
+			backends,
+			dx12_shader_compiler,
+			..Default::default()
+		}),
+	}
+}
+impl SetupContextPartial1{
+	fn create_surface<'a>(self,window:&'a winit::window::Window)->Result<SetupContextPartial2<'a>,wgpu::CreateSurfaceError>{
+		Ok(SetupContextPartial2{
+			backends:self.backends,
+			surface:self.instance.create_surface(window)?,
+			instance:self.instance,
+		})
+	}
+}
+struct SetupContextPartial2<'a>{
+	backends:wgpu::Backends,
+	instance:wgpu::Instance,
+	surface:wgpu::Surface<'a>,
+}
+impl<'a> SetupContextPartial2<'a>{
+	fn pick_adapter(self)->SetupContextPartial3<'a>{
+		let adapter;
+
+		//TODO: prefer adapter that implements optional features
+		//let optional_features=optional_features();
+		let required_features=required_features();
+
+		//no helper function smh gotta write it myself
+		let adapters=self.instance.enumerate_adapters(self.backends);
+
+		let mut chosen_adapter=None;
+		let mut chosen_adapter_score=0;
+		for adapter in adapters {
+			if !adapter.is_surface_supported(&self.surface) {
+				continue;
+			}
+
+			let score=match adapter.get_info().device_type{
+				wgpu::DeviceType::IntegratedGpu=>3,
+				wgpu::DeviceType::DiscreteGpu=>4,
+				wgpu::DeviceType::VirtualGpu=>2,
+				wgpu::DeviceType::Other|wgpu::DeviceType::Cpu=>1,
+			};
+
+			let adapter_features=adapter.features();
+			if chosen_adapter_score<score&&adapter_features.contains(required_features) {
+				chosen_adapter_score=score;
+				chosen_adapter=Some(adapter);
+			}
+		}
+
+		if let Some(maybe_chosen_adapter)=chosen_adapter{
+			adapter=maybe_chosen_adapter;
+		}else{
+			panic!("No suitable GPU adapters found on the system!");
+		}
+
+
+		let adapter_info=adapter.get_info();
+		println!("Using {} ({:?})", adapter_info.name, adapter_info.backend);
+
+		let required_downlevel_capabilities=required_downlevel_capabilities();
+		let downlevel_capabilities=adapter.get_downlevel_capabilities();
+		assert!(
+			downlevel_capabilities.shader_model >= required_downlevel_capabilities.shader_model,
+			"Adapter does not support the minimum shader model required to run this example: {:?}",
+			required_downlevel_capabilities.shader_model
+		);
+		assert!(
+			downlevel_capabilities
+				.flags
+				.contains(required_downlevel_capabilities.flags),
+			"Adapter does not support the downlevel capabilities required to run this example: {:?}",
+			required_downlevel_capabilities.flags - downlevel_capabilities.flags
+		);
+		SetupContextPartial3{
+			instance:self.instance,
+			surface:self.surface,
+			adapter,
+		}
+	}
+}
+struct SetupContextPartial3<'a>{
+	instance:wgpu::Instance,
+	surface:wgpu::Surface<'a>,
+	adapter:wgpu::Adapter,
+}
+impl<'a> SetupContextPartial3<'a>{
+	fn request_device(self)->SetupContextPartial4<'a>{
+		let optional_features=optional_features();
+		let required_features=required_features();
+
+		// Make sure we use the texture resolution limits from the adapter, so we can support images the size of the surface.
+		let needed_limits=required_limits().using_resolution(self.adapter.limits());
+
+		let trace_dir=std::env::var("WGPU_TRACE");
+		let (device, queue)=pollster::block_on(self.adapter
+			.request_device(
+				&wgpu::DeviceDescriptor {
+					label: None,
+					required_features: (optional_features & self.adapter.features()) | required_features,
+					required_limits: needed_limits,
+					memory_hints:wgpu::MemoryHints::Performance,
+				},
+				trace_dir.ok().as_ref().map(std::path::Path::new),
+			))
+			.expect("Unable to find a suitable GPU adapter!");
+
+		SetupContextPartial4{
+			instance:self.instance,
+			surface:self.surface,
+			adapter:self.adapter,
+			device,
+			queue,
+		}
+	}
+}
+struct SetupContextPartial4<'a>{
+	instance:wgpu::Instance,
+	surface:wgpu::Surface<'a>,
+	adapter:wgpu::Adapter,
+	device:wgpu::Device,
+	queue:wgpu::Queue,
+}
+impl<'a> SetupContextPartial4<'a>{
+	fn configure_surface(self,size:&'a winit::dpi::PhysicalSize<u32>)->SetupContext<'a>{
+		let mut config=self.surface
+			.get_default_config(&self.adapter, size.width, size.height)
+			.expect("Surface isn't supported by the adapter.");
+		let surface_view_format=config.format.add_srgb_suffix();
+		config.view_formats.push(surface_view_format);
+		config.present_mode=wgpu::PresentMode::AutoNoVsync;
+		self.surface.configure(&self.device, &config);
+
+		SetupContext{
+			instance:self.instance,
+			surface:self.surface,
+			device:self.device,
+			queue:self.queue,
+			config,
+		}
+	}
+}
+pub struct SetupContext<'a>{
+	pub instance:wgpu::Instance,
+	pub surface:wgpu::Surface<'a>,
+	pub device:wgpu::Device,
+	pub queue:wgpu::Queue,
+	pub config:wgpu::SurfaceConfiguration,
+}
+
+pub fn setup_and_start(title:String){
+	let event_loop=winit::event_loop::EventLoop::new().unwrap();
+
+	println!("Initializing the surface...");
+
+	let partial_1=create_instance();
+
+	let window=create_window(title.as_str(),&event_loop).unwrap();
+
+	let partial_2=partial_1.create_surface(&window).unwrap();
+
+	let partial_3=partial_2.pick_adapter();
+
+	let partial_4=partial_3.request_device();
+
+	let size=window.inner_size();
+
+	let setup_context=partial_4.configure_surface(&size);
+
+	//dedicated thread to ping request redraw back and resize the window doesn't seem logical
+
+	//the thread that spawns the physics thread
+	let mut window_thread=crate::window::worker(
+		&window,
+		setup_context,
+	);
+
+	if let Some(arg)=std::env::args().nth(1){
+		let path=std::path::PathBuf::from(arg);
+		window_thread.send(TimedInstruction{
+			time:integer::Time::ZERO,
+			instruction:WindowInstruction::WindowEvent(winit::event::WindowEvent::DroppedFile(path)),
+		}).unwrap();
+	};
+
+	println!("Entering event loop...");
+	let root_time=std::time::Instant::now();
+	run_event_loop(event_loop,window_thread,root_time).unwrap();
+}
+
+fn run_event_loop(
+	event_loop:winit::event_loop::EventLoop<()>,
+	mut window_thread:crate::compat_worker::QNWorker<TimedInstruction<WindowInstruction>>,
+	root_time:std::time::Instant
+	)->Result<(),winit::error::EventLoopError>{
+		event_loop.run(move |event,elwt|{
+			let time=integer::Time::from_nanos(root_time.elapsed().as_nanos() as i64);
+			// *control_flow=if cfg!(feature="metal-auto-capture"){
+			// 	winit::event_loop::ControlFlow::Exit
+			// }else{
+			// 	winit::event_loop::ControlFlow::Poll
+			// };
+			match event{
+				winit::event::Event::AboutToWait=>{
+					window_thread.send(TimedInstruction{time,instruction:WindowInstruction::RequestRedraw}).unwrap();
+				}
+				winit::event::Event::WindowEvent {
+					event:
+						// WindowEvent::Resized(size)
+						// | WindowEvent::ScaleFactorChanged {
+						// 	new_inner_size: &mut size,
+						// 	..
+						// },
+						winit::event::WindowEvent::Resized(size),//ignoring scale factor changed for now because mutex bruh
+					window_id:_,
+				} => {
+					window_thread.send(TimedInstruction{time,instruction:WindowInstruction::Resize(size)}).unwrap();
+				}
+				winit::event::Event::WindowEvent{event,..}=>match event{
+					winit::event::WindowEvent::KeyboardInput{
+						event:
+							winit::event::KeyEvent {
+							logical_key: winit::keyboard::Key::Named(winit::keyboard::NamedKey::Escape),
+								state: winit::event::ElementState::Pressed,
+								..
+							},
+						..
+					}
+					|winit::event::WindowEvent::CloseRequested=>{
+						elwt.exit();
+					}
+					winit::event::WindowEvent::RedrawRequested=>{
+						window_thread.send(TimedInstruction{time,instruction:WindowInstruction::Render}).unwrap();
+					}
+					_=>{
+						window_thread.send(TimedInstruction{time,instruction:WindowInstruction::WindowEvent(event)}).unwrap();
+					}
+				},
+				winit::event::Event::DeviceEvent{
+					event,
+					..
+				} => {
+					window_thread.send(TimedInstruction{time,instruction:WindowInstruction::DeviceEvent(event)}).unwrap();
+				},
+				_=>{}
+			}
+		})
+}

src/shader.wgsl

@@ -1,21 +1,23 @@
-struct SkyOutput {
-	@builtin(position) position: vec4<f32>,
-	@location(0) sampledir: vec3<f32>,
-};
-
-struct Data {
+struct Camera {
 	// from camera to screen
 	proj: mat4x4<f32>,
 	// from screen to camera
 	proj_inv: mat4x4<f32>,
 	// from world to camera
 	view: mat4x4<f32>,
-	// camera position
-	cam_pos: vec4<f32>,
+	// from camera to world
+	view_inv: mat4x4<f32>,
 };
+
+//group 0 is the camera
 @group(0)
 @binding(0)
-var<uniform> r_data: Data;
+var<uniform> camera: Camera;
+
+struct SkyOutput {
+	@builtin(position) position: vec4<f32>,
+	@location(0) sampledir: vec3<f32>,
+};
 
 @vertex
 fn vs_sky(@builtin(vertex_index) vertex_index: u32) -> SkyOutput {
@@ -29,9 +31,8 @@ fn vs_sky(@builtin(vertex_index) vertex_index: u32) -> SkyOutput {
 		1.0
 	);
 
-	// transposition = inversion for this orthonormal matrix
-	let inv_model_view = transpose(mat3x3<f32>(r_data.view[0].xyz, r_data.view[1].xyz, r_data.view[2].xyz));
-	let unprojected = r_data.proj_inv * pos;
+	let inv_model_view = mat3x3<f32>(camera.view_inv[0].xyz, camera.view_inv[1].xyz, camera.view_inv[2].xyz);
+	let unprojected = camera.proj_inv * pos;
 
 	var result: SkyOutput;
 	result.sampledir = inv_model_view * unprojected.xyz;
@@ -39,93 +40,73 @@ fn vs_sky(@builtin(vertex_index) vertex_index: u32) -> SkyOutput {
 	return result;
 }
 
-struct GroundOutput {
-	@builtin(position) position: vec4<f32>,
-	@location(4) pos: vec3<f32>,
-};
-
-@vertex
-fn vs_ground(@builtin(vertex_index) vertex_index: u32) -> GroundOutput {
-	// hacky way to draw two triangles that make a square
-	let tmp1 = i32(vertex_index)/2-i32(vertex_index)/3;
-	let tmp2 = i32(vertex_index)&1;
-	let pos = vec3<f32>(
-		f32(tmp1) * 2.0 - 1.0,
-		0.0,
-		f32(tmp2) * 2.0 - 1.0
-	) * 160.0;
-
-	var result: GroundOutput;
-	result.pos = pos;
-	result.position = r_data.proj * r_data.view * vec4<f32>(pos, 1.0);
-	return result;
+struct ModelInstance{
+	transform:mat4x4<f32>,
+	normal_transform:mat3x3<f32>,
+	color:vec4<f32>,
 }
+//my fancy idea is to create a megatexture for each model that includes all the textures each intance will need
+//the texture transform then maps the texture coordinates to the location of the specific texture
+//group 1 is the model
+const MAX_MODEL_INSTANCES=512;
+@group(2)
+@binding(0)
+var<uniform> model_instances: array<ModelInstance, MAX_MODEL_INSTANCES>;
+@group(2)
+@binding(1)
+var model_texture: texture_2d<f32>;
+@group(2)
+@binding(2)
+var model_sampler: sampler;
 
-struct EntityOutput {
+struct EntityOutputTexture {
 	@builtin(position) position: vec4<f32>,
 	@location(1) texture: vec2<f32>,
 	@location(2) normal: vec3<f32>,
 	@location(3) view: vec3<f32>,
+	@location(4) color: vec4<f32>,
+	@location(5) @interpolate(flat) model_color: vec4<f32>,
 };
-
-@group(1)
-@binding(0)
-var<uniform> r_EntityTransform: mat4x4<f32>;
-
 @vertex
-fn vs_entity(
+fn vs_entity_texture(
+	@builtin(instance_index) instance: u32,
 	@location(0) pos: vec3<f32>,
 	@location(1) texture: vec2<f32>,
 	@location(2) normal: vec3<f32>,
-) -> EntityOutput {
-	var position: vec4<f32> = r_EntityTransform * vec4<f32>(pos, 1.0);
-	var result: EntityOutput;
-	result.normal = (r_EntityTransform * vec4<f32>(normal, 0.0)).xyz;
-	result.texture=texture;
-	result.view = position.xyz - r_data.cam_pos.xyz;
-	result.position = r_data.proj * r_data.view * position;
+	@location(3) color: vec4<f32>,
+) -> EntityOutputTexture {
+	var position: vec4<f32> = model_instances[instance].transform * vec4<f32>(pos, 1.0);
+	var result: EntityOutputTexture;
+	result.normal = model_instances[instance].normal_transform * normal;
+	result.texture = texture;
+	result.color = color;
+	result.model_color = model_instances[instance].color;
+	result.view = position.xyz - camera.view_inv[3].xyz;//col(3)
+	result.position = camera.proj * camera.view * position;
 	return result;
 }
 
-@group(0)
+//group 2 is the skybox texture
+@group(1)
+@binding(0)
+var cube_texture: texture_cube<f32>;
+@group(1)
 @binding(1)
-var r_texture: texture_cube<f32>;
-@group(0)
-@binding(2)
-var r_sampler: sampler;
+var cube_sampler: sampler;
 
 @fragment
 fn fs_sky(vertex: SkyOutput) -> @location(0) vec4<f32> {
-	return textureSample(r_texture, r_sampler, vertex.sampledir);
+	return textureSample(cube_texture, cube_sampler, vertex.sampledir);
 }
 
 @fragment
-fn fs_entity(vertex: EntityOutput) -> @location(0) vec4<f32> {
+fn fs_entity_texture(vertex: EntityOutputTexture) -> @location(0) vec4<f32> {
 	let incident = normalize(vertex.view);
 	let normal = normalize(vertex.normal);
 	let d = dot(normal, incident);
 	let reflected = incident - 2.0 * d * normal;
 
-	let dir = vec3<f32>(-1.0)+2.0*vec3<f32>(vertex.texture.x,0.0,vertex.texture.y);
-	let texture_color = textureSample(r_texture, r_sampler, dir).rgb;
-	let reflected_color = textureSample(r_texture, r_sampler, reflected).rgb;
-	return vec4<f32>(mix(vec3<f32>(0.1) + 0.5 * reflected_color,texture_color,1.0-pow(1.0-abs(d),2.0)), 1.0);
-}
-
-fn modulo_euclidean (a: f32, b: f32) -> f32 {
-	var m = a % b;
-	if (m < 0.0) {
-		if (b < 0.0) {
-			m -= b;
-		} else {
-			m += b;
-		}
-	}
-	return m;
-}
-
-@fragment
-fn fs_ground(vertex: GroundOutput) -> @location(0) vec4<f32> {
-	let dir = vec3<f32>(-1.0)+vec3<f32>(modulo_euclidean(vertex.pos.x/16.,1.0),0.0,modulo_euclidean(vertex.pos.z/16.,1.0))*2.0;
-	return vec4<f32>(textureSample(r_texture, r_sampler, dir).rgb, 1.0);
+	let fragment_color = textureSample(model_texture, model_sampler, vertex.texture)*vertex.color;
+	let reflected_color = textureSample(cube_texture, cube_sampler, reflected).rgb;
+	return mix(vec4<f32>(vec3<f32>(0.05) + 0.2 * reflected_color,1.0),mix(vertex.model_color,vec4<f32>(fragment_color.rgb,1.0),fragment_color.a),0.5+0.5*abs(d));
 }

src/window.rs

@@ -0,0 +1,223 @@
+use crate::physics_worker::InputInstruction;
+use strafesnet_common::integer;
+use strafesnet_common::instruction::TimedInstruction;
+
+pub enum WindowInstruction{
+	Resize(winit::dpi::PhysicalSize<u32>),
+	WindowEvent(winit::event::WindowEvent),
+	DeviceEvent(winit::event::DeviceEvent),
+	RequestRedraw,
+	Render,
+}
+
+//holds thread handles to dispatch to
+struct WindowContext<'a>{
+	manual_mouse_lock:bool,
+	mouse:strafesnet_common::mouse::MouseState,//std::sync::Arc<std::sync::Mutex<>>
+	screen_size:glam::UVec2,
+	window:&'a winit::window::Window,
+	physics_thread:crate::compat_worker::QNWorker<'a, TimedInstruction<crate::physics_worker::Instruction>>,
+}
+
+impl WindowContext<'_>{
+	fn get_middle_of_screen(&self)->winit::dpi::PhysicalPosition<f32>{
+		winit::dpi::PhysicalPosition::new(self.screen_size.x as f32/2.0,self.screen_size.y as f32/2.0)
+	}
+	fn window_event(&mut self,time:integer::Time,event: winit::event::WindowEvent) {
+		match event {
+			winit::event::WindowEvent::DroppedFile(path)=>{
+				match crate::file::load(path.as_path()){
+					Ok(map)=>self.physics_thread.send(TimedInstruction{time,instruction:crate::physics_worker::Instruction::ChangeMap(map)}).unwrap(),
+					Err(e)=>println!("Failed to load map: {e}"),
+				}
+			},
+			winit::event::WindowEvent::Focused(state)=>{
+				//pause unpause
+				self.physics_thread.send(TimedInstruction{
+					time,
+					instruction:crate::physics_worker::Instruction::SetPaused(!state),
+				}).unwrap();
+				//recalculate pressed keys on focus
+			},
+			winit::event::WindowEvent::KeyboardInput{
+				event:winit::event::KeyEvent{state,logical_key,repeat:false,..},
+				..
+			}=>{
+				let s=match state{
+					winit::event::ElementState::Pressed=>true,
+					winit::event::ElementState::Released=>false,
+				};
+				match logical_key{
+					winit::keyboard::Key::Named(winit::keyboard::NamedKey::Tab)=>{
+						if s{
+							self.manual_mouse_lock=false;
+							match self.window.set_cursor_position(self.get_middle_of_screen()){
+								Ok(())=>(),
+								Err(e)=>println!("Could not set cursor position: {:?}",e),
+							}
+							match self.window.set_cursor_grab(winit::window::CursorGrabMode::None){
+								Ok(())=>(),
+								Err(e)=>println!("Could not release cursor: {:?}",e),
+							}
+						}else{
+							//if cursor is outside window don't lock but apparently there's no get pos function
+							//let pos=window.get_cursor_pos();
+							match self.window.set_cursor_grab(winit::window::CursorGrabMode::Locked){
+								Ok(())=>(),
+								Err(_)=>{
+									match self.window.set_cursor_grab(winit::window::CursorGrabMode::Confined){
+										Ok(())=>(),
+										Err(e)=>{
+											self.manual_mouse_lock=true;
+											println!("Could not confine cursor: {:?}",e)
+										},
+									}
+								}
+							}
+						}
+						self.window.set_cursor_visible(s);
+					},
+					winit::keyboard::Key::Named(winit::keyboard::NamedKey::F11)=>{
+						if s{
+							if self.window.fullscreen().is_some(){
+								self.window.set_fullscreen(None);
+							}else{
+								self.window.set_fullscreen(Some(winit::window::Fullscreen::Borderless(None)));
+							}
+						}
+					},
+					winit::keyboard::Key::Named(winit::keyboard::NamedKey::Escape)=>{
+						if s{
+							self.manual_mouse_lock=false;
+							match self.window.set_cursor_grab(winit::window::CursorGrabMode::None){
+								Ok(())=>(),
+								Err(e)=>println!("Could not release cursor: {:?}",e),
+							}
+							self.window.set_cursor_visible(true);
+						}
+					},
+					keycode=>{
+						if let Some(input_instruction)=match keycode{
+							winit::keyboard::Key::Named(winit::keyboard::NamedKey::Space)=>Some(InputInstruction::Jump(s)),
+							winit::keyboard::Key::Character(key)=>match key.as_str(){
+								"w"=>Some(InputInstruction::MoveForward(s)),
+								"a"=>Some(InputInstruction::MoveLeft(s)),
+								"s"=>Some(InputInstruction::MoveBack(s)),
+								"d"=>Some(InputInstruction::MoveRight(s)),
+								"e"=>Some(InputInstruction::MoveUp(s)),
+								"q"=>Some(InputInstruction::MoveDown(s)),
+								"z"=>Some(InputInstruction::Zoom(s)),
+								"r"=>if s{
+									//mouse needs to be reset since the position is absolute
+									self.mouse=strafesnet_common::mouse::MouseState::default();
+									Some(InputInstruction::ResetAndRestart)
+								}else{None},
+								"f"=>if s{Some(InputInstruction::PracticeFly)}else{None},
+								_=>None,
+							},
+							_=>None,
+						}{
+							self.physics_thread.send(TimedInstruction{
+								time,
+								instruction:crate::physics_worker::Instruction::Input(input_instruction),
+							}).unwrap();
+						}
+					},
+				}
+			},
+			_=>(),
+		}
+	}
+
+	fn device_event(&mut self,time:integer::Time,event: winit::event::DeviceEvent) {
+		match event {
+			winit::event::DeviceEvent::MouseMotion {
+			    delta,//these (f64,f64) are integers on my machine
+			} => {
+				if self.manual_mouse_lock{
+					match self.window.set_cursor_position(self.get_middle_of_screen()){
+						Ok(())=>(),
+						Err(e)=>println!("Could not set cursor position: {:?}",e),
+					}
+				}
+				//do not step the physics because the mouse polling rate is higher than the physics can run.
+				//essentially the previous input will be overwritten until a true step runs
+				//which is fine because they run all the time.
+				let delta=glam::ivec2(delta.0 as i32,delta.1 as i32);
+				self.mouse.pos+=delta;
+				self.physics_thread.send(TimedInstruction{
+					time,
+					instruction:crate::physics_worker::Instruction::Input(InputInstruction::MoveMouse(self.mouse.pos)),
+				}).unwrap();
+			},
+			winit::event::DeviceEvent::MouseWheel {
+			    delta,
+			} => {
+				println!("mousewheel {:?}",delta);
+				if false{//self.physics.style.use_scroll{
+					self.physics_thread.send(TimedInstruction{
+						time,
+						instruction:crate::physics_worker::Instruction::Input(InputInstruction::Jump(true)),//activates the immediate jump path, but the style modifier prevents controls&CONTROL_JUMP bit from being set to auto jump
+					}).unwrap();
+				}
+			}
+			_=>(),
+		}
+	}
+}
+pub fn worker<'a>(
+	window:&'a winit::window::Window,
+	setup_context:crate::setup::SetupContext<'a>,
+)->crate::compat_worker::QNWorker<'a,TimedInstruction<WindowInstruction>>{
+	// WindowContextSetup::new
+		let user_settings=crate::settings::read_user_settings();
+
+		let mut graphics=crate::graphics::GraphicsState::new(&setup_context.device,&setup_context.queue,&setup_context.config);
+		graphics.load_user_settings(&user_settings);
+
+	//WindowContextSetup::into_context
+		let screen_size=glam::uvec2(setup_context.config.width,setup_context.config.height);
+		let graphics_thread=crate::graphics_worker::new(graphics,setup_context.config,setup_context.surface,setup_context.device,setup_context.queue);
+		let mut window_context=WindowContext{
+			manual_mouse_lock:false,
+			mouse:strafesnet_common::mouse::MouseState::default(),
+			//make sure to update this!!!!!
+			screen_size,
+			window,
+			physics_thread:crate::physics_worker::new(
+				graphics_thread,
+				user_settings,
+			),
+		};
+
+	//WindowContextSetup::into_worker
+		crate::compat_worker::QNWorker::new(move |ins:TimedInstruction<WindowInstruction>|{
+			match ins.instruction{
+				WindowInstruction::RequestRedraw=>{
+					window_context.window.request_redraw();
+				}
+				WindowInstruction::WindowEvent(window_event)=>{
+					window_context.window_event(ins.time,window_event);
+				},
+				WindowInstruction::DeviceEvent(device_event)=>{
+					window_context.device_event(ins.time,device_event);
+				},
+				WindowInstruction::Resize(size)=>{
+					window_context.physics_thread.send(
+						TimedInstruction{
+							time:ins.time,
+							instruction:crate::physics_worker::Instruction::Resize(size)
+						}
+					).unwrap();
+				}
+				WindowInstruction::Render=>{
+					window_context.physics_thread.send(
+						TimedInstruction{
+							time:ins.time,
+							instruction:crate::physics_worker::Instruction::Render
+						}
+					).unwrap();
+				}
+			}
+		})
+}

src/worker.rs

@@ -0,0 +1,216 @@
+use std::thread;
+use std::sync::{mpsc,Arc};
+use parking_lot::Mutex;
+
+//WorkerPool
+struct Pool(u32);
+enum PoolOrdering{
+	Single,//single thread cannot get out of order
+	Ordered(u32),//order matters and should be buffered/dropped according to ControlFlow
+	Unordered(u32),//order does not matter
+}
+//WorkerInput
+enum Input{
+	//no input, workers have everything needed at creation
+	None,
+	//Immediate input to any available worker, dropped if they are overflowing (all workers are busy)
+	Immediate,
+	//Queued input is ordered, but serial jobs that mutate state (such as running physics) can only be done with a single worker
+	Queued,//"Fifo"
+	//Query a function to get next input when a thread becomes available
+	//worker stops querying when Query function returns None and dies after all threads complete
+	//lifetimes sound crazy on this one
+	Query,
+	//Queue of length one, the input is replaced if it is submitted twice before the current work finishes
+	Mailbox,
+}
+//WorkerOutput
+enum Output{
+	None(Pool),
+	Realtime(PoolOrdering),//outputs are dropped if they are out of order and order is demanded
+	Buffered(PoolOrdering),//outputs are held back internally if they are out of order and order is demanded
+}
+
+//It would be possible to implement all variants
+//with a query input function and callback output function but I'm not sure if that's worth it.
+//Immediate = Condvar
+//Queued = receiver.recv()
+//a callback function would need to use an async runtime!
+
+//realtime output is an arc mutex of the output value that is assigned every time a worker completes a job
+//buffered output produces a receiver object that can be passed to the creation of another worker
+//when ordering is requested, output is ordered by the order each thread is run
+//which is the same as the order that the input data is processed except for Input::None which has no input data
+//WorkerDescription
+struct Description{
+	input:Input,
+	output:Output,
+}
+
+//The goal here is to have a worker thread that parks itself when it runs out of work.
+//The worker thread publishes the result of its work back to the worker object for every item in the work queue.
+//Previous values do not matter as soon as a new value is produced, which is why it's called "Realtime"
+//The physics (target use case) knows when it has not changed the body, so not updating the value is also an option.
+
+/*
+QR = WorkerDescription{
+	input:Queued,
+	output:Realtime(Single),
+}
+*/
+pub struct QRWorker<Task:Send,Value:Clone>{
+	sender: mpsc::Sender<Task>,
+	value:Arc<Mutex<Value>>,
+}
+
+impl<Task:Send+'static,Value:Clone+Send+'static> QRWorker<Task,Value>{
+	pub fn new<F:FnMut(Task)->Value+Send+'static>(value:Value,mut f:F) -> Self {
+		let (sender, receiver) = mpsc::channel::<Task>();
+		let ret=Self {
+			sender,
+			value:Arc::new(Mutex::new(value)),
+		};
+		let value=ret.value.clone();
+		thread::spawn(move || {
+			loop {
+				match receiver.recv() {
+					Ok(task) => {
+						let v=f(task);//make sure function is evaluated before lock is acquired
+						*value.lock()=v;
+					}
+					Err(_) => {
+						println!("Worker stopping.",);
+						break;
+					}
+				}
+			}
+		});
+		ret
+	}
+
+	pub fn send(&self,task:Task)->Result<(), mpsc::SendError<Task>>{
+		self.sender.send(task)
+	}
+
+	pub fn grab_clone(&self)->Value{
+		self.value.lock().clone()
+	}
+}
+
+/*
+QN = WorkerDescription{
+	input:Queued,
+	output:None(Single),
+}
+*/
+//None Output Worker does all its work internally from the perspective of the work submitter
+pub struct QNWorker<'a,Task:Send>{
+	sender: mpsc::Sender<Task>,
+	handle:thread::ScopedJoinHandle<'a,()>,
+}
+
+impl<'a,Task:Send+'a> QNWorker<'a,Task>{
+	pub fn new<F:FnMut(Task)+Send+'a>(scope:&'a thread::Scope<'a,'_>,mut f:F)->QNWorker<'a,Task>{
+		let (sender,receiver)=mpsc::channel::<Task>();
+		let handle=scope.spawn(move ||{
+			loop {
+				match receiver.recv() {
+					Ok(task)=>f(task),
+					Err(_)=>{
+						println!("Worker stopping.",);
+						break;
+					}
+				}
+			}
+		});
+		Self{
+			sender,
+			handle,
+		}
+	}
+	pub fn send(&self,task:Task)->Result<(),mpsc::SendError<Task>>{
+		self.sender.send(task)
+	}
+}
+
+/*
+IN = WorkerDescription{
+	input:Immediate,
+	output:None(Single),
+}
+*/
+//Inputs are dropped if the worker is busy
+pub struct INWorker<'a,Task:Send>{
+	sender: mpsc::SyncSender<Task>,
+	handle:thread::ScopedJoinHandle<'a,()>,
+}
+
+impl<'a,Task:Send+'a> INWorker<'a,Task>{
+	pub fn new<F:FnMut(Task)+Send+'a>(scope:&'a thread::Scope<'a,'_>,mut f:F)->INWorker<'a,Task>{
+		let (sender,receiver)=mpsc::sync_channel::<Task>(1);
+		let handle=scope.spawn(move ||{
+			loop {
+				match receiver.recv() {
+					Ok(task)=>f(task),
+					Err(_)=>{
+						println!("Worker stopping.",);
+						break;
+					}
+				}
+			}
+		});
+		Self{
+			sender,
+			handle,
+		}
+	}
+	//blocking!
+	pub fn blocking_send(&self,task:Task)->Result<(), mpsc::SendError<Task>>{
+		self.sender.send(task)
+	}
+	pub fn send(&self,task:Task)->Result<(), mpsc::TrySendError<Task>>{
+		self.sender.try_send(task)
+	}
+}
+
+#[cfg(test)]
+mod test{
+	use super::{thread,QRWorker};
+	use crate::physics;
+	use strafesnet_common::{integer,instruction};
+	#[test]//How to run this test with printing: cargo test --release -- --nocapture
+	fn test_worker() {
+		// Create the worker thread
+		let test_body=physics::Body::new(integer::vec3::ONE,integer::vec3::ONE,integer::vec3::ONE,integer::Time::ZERO);
+		let worker=QRWorker::new(physics::Body::ZERO,
+			|_|physics::Body::new(integer::vec3::ONE,integer::vec3::ONE,integer::vec3::ONE,integer::Time::ZERO)
+		);
+
+		// Send tasks to the worker
+		for _ in 0..5 {
+			let task = instruction::TimedInstruction{
+				time:integer::Time::ZERO,
+				instruction:strafesnet_common::physics::Instruction::Idle,
+			};
+			worker.send(task).unwrap();
+		}
+
+		// Optional: Signal the worker to stop (in a real-world scenario)
+		// sender.send("STOP".to_string()).unwrap();
+
+		// Sleep to allow the worker thread to finish processing
+		thread::sleep(std::time::Duration::from_millis(10));
+
+		// Send a new task
+		let task = instruction::TimedInstruction{
+			time:integer::Time::ZERO,
+			instruction:strafesnet_common::physics::Instruction::Idle,
+		};
+		worker.send(task).unwrap();
+
+		//assert_eq!(test_body,worker.grab_clone());
+
+		// wait long enough to see print from final task
+		thread::sleep(std::time::Duration::from_millis(10));
+	}
+}

tools/arcane

@@ -0,0 +1 @@
+mangohud ../target/release/strafe-client bhop_maps/5692113331.snfm

tools/bhop_maps

@@ -0,0 +1 @@
+/run/media/quat/Files/Documents/map-files/verify-scripts/maps/bhop_snfm

tools/cross-compile.sh

@@ -0,0 +1 @@
+cargo build --release --target x86_64-pc-windows-gnu --all-features

tools/iso

@@ -0,0 +1 @@
+mangohud ../target/release/strafe-client bhop_maps/5692124338.snfm

tools/make-demo.sh

@@ -0,0 +1,4 @@
+mkdir -p ../target/demo
+mv ../target/x86_64-pc-windows-gnu/release/strafe-client.exe ../target/demo/strafe-client.exe
+rm ../target/demo.7z
+7z a -t7z -mx=9 -mfb=273 -ms -md=31 -myx=9 -mtm=- -mmt -mmtf -md=1536m -mmf=bt3 -mmc=10000 -mpb=0 -mlc=0 ../target/demo.7z ../target/demo

tools/run

@@ -0,0 +1 @@
+mangohud ../target/release/strafe-client "$1"

tools/settings.conf

@@ -0,0 +1,4 @@
+[camera]
+sensitivity_x=98384
+fov_y=1.0
+#fov_x_from_y_ratio=1.33333333333333333333333333333333

tools/surf_maps

@@ -0,0 +1 @@
+/run/media/quat/Files/Documents/map-files/verify-scripts/maps/surf_snfm

tools/toc

@@ -0,0 +1 @@
+mangohud ../target/release/strafe-client bhop_maps/5692152916.snfm

tools/utopia

@@ -0,0 +1 @@
+mangohud ../target/release/strafe-client surf_maps/5692145408.snfm