success

make invalid states unrepresentable!!!

.cargo/config.toml

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

.gitignore

@@ -1,2 +1 @@
 /target
-/textures

Cargo.toml

@@ -1,30 +1,36 @@
 [package]
 name = "strafe-client"
-version = "0.8.0"
+version = "0.10.3"
 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"
-lazy-regex = "3.0.2"
-log = "0.4.20"
-obj = "0.10.2"
+glam = "0.28.0"
+id = { version = "0.1.0", registry = "strafesnet" }
 parking_lot = "0.12.1"
 pollster = "0.3.0"
-rbx_binary = "0.7.1"
-rbx_dom_weak = "2.5.0"
-rbx_reflection_database = "0.2.7"
-rbx_xml = "0.13.1"
-wgpu = "0.17.0"
-winit = "0.28.6"
+strafesnet_bsp_loader = { version = "0.1.3", registry = "strafesnet", optional = true }
+strafesnet_common = { version = "0.4.0", registry = "strafesnet" }
+strafesnet_deferred_loader = { version = "0.3.1", features = ["legacy"], registry = "strafesnet", optional = true }
+strafesnet_rbx_loader = { version = "0.3.2", registry = "strafesnet", optional = true }
+strafesnet_snf = { version = "0.1.2", registry = "strafesnet", optional = true }
+wgpu = "22.0.0"
+winit = "0.30.4"
 
-#[profile.release]
+[profile.release]
 #lto = true
-#strip = true
-#codegen-units = 1
+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/aabb.rs

@@ -1,91 +0,0 @@
-#[derive(Debug,Clone,Copy,Hash,Eq,PartialEq)]
-pub enum AabbFace{
-	Right,//+X
-	Top,
-	Back,
-	Left,
-	Bottom,
-	Front,
-}
-#[derive(Clone)]
-pub struct Aabb {
-	pub min: glam::Vec3,
-	pub max: glam::Vec3,
-}
-
-impl Default for Aabb {
-	fn default() -> Self {
-		Aabb::new()
-	}
-}
-
-impl Aabb {
-	const VERTEX_DATA: [glam::Vec3; 8] = [
-		glam::vec3(1., -1., -1.),
-		glam::vec3(1., 1., -1.),
-		glam::vec3(1., 1., 1.),
-		glam::vec3(1., -1., 1.),
-		glam::vec3(-1., -1., 1.),
-		glam::vec3(-1., 1., 1.),
-		glam::vec3(-1., 1., -1.),
-		glam::vec3(-1., -1., -1.),
-	];
-
-	pub fn new() -> Self {
-		Self {min: glam::Vec3::INFINITY,max: glam::Vec3::NEG_INFINITY}
-	}
-
-	pub fn grow(&mut self, point:glam::Vec3){
-		self.min=self.min.min(point);
-		self.max=self.max.max(point);
-	}
-	pub fn join(&mut self, aabb:&Aabb){
-		self.min=self.min.min(aabb.min);
-		self.max=self.max.max(aabb.max);
-	}
-	pub fn inflate(&mut self, hs:glam::Vec3){
-		self.min-=hs;
-		self.max+=hs;
-	}
-	pub fn intersects(&self,aabb:&Aabb)->bool{
-		(self.min.cmplt(aabb.max)&aabb.min.cmplt(self.max)).all()
-	}
-	pub fn normal(face:AabbFace) -> glam::Vec3 {
-		match face {
-			AabbFace::Right => glam::vec3(1.,0.,0.),
-			AabbFace::Top => glam::vec3(0.,1.,0.),
-			AabbFace::Back => glam::vec3(0.,0.,1.),
-			AabbFace::Left => glam::vec3(-1.,0.,0.),
-			AabbFace::Bottom => glam::vec3(0.,-1.,0.),
-			AabbFace::Front => glam::vec3(0.,0.,-1.),
-		}
-	}
-	pub fn unit_vertices() -> [glam::Vec3;8] {
-		return Self::VERTEX_DATA;
-	}
-	pub fn face(&self,face:AabbFace) -> Aabb {
-		let mut aabb=self.clone();
-		//in this implementation face = worldspace aabb face
-		match face {
-			AabbFace::Right => aabb.min.x=aabb.max.x,
-			AabbFace::Top => aabb.min.y=aabb.max.y,
-			AabbFace::Back => aabb.min.z=aabb.max.z,
-			AabbFace::Left => aabb.max.x=aabb.min.x,
-			AabbFace::Bottom => aabb.max.y=aabb.min.y,
-			AabbFace::Front => aabb.max.z=aabb.min.z,
-		}
-		return aabb;
-	}
-	pub fn center(&self)->glam::Vec3{
-		return (self.min+self.max)/2.0
-	}
-	//probably use floats for area & volume because we don't care about precision
-	pub fn area_weight(&self)->f32{
-		let d=self.max-self.min;
-		d.x*d.y+d.y*d.z+d.z*d.x
-	}
-	pub fn volume(&self)->f32{
-		let d=self.max-self.min;
-		d.x*d.y*d.z
-	}
-}

src/bvh.rs

@@ -1,107 +0,0 @@
-use crate::aabb::Aabb;
-
-//da algaritum
-//lista boxens
-//sort by {minx,maxx,miny,maxy,minz,maxz} (6 lists)
-//find the sets that minimizes the sum of surface areas
-//splitting is done when the minimum split sum of surface areas is larger than the node's own surface area
-
-//start with bisection into octrees because a bad bvh is still 1000x better than no bvh
-//sort the centerpoints on each axis (3 lists)
-//bv is put into octant based on whether it is upper or lower in each list
-#[derive(Default)]
-pub struct BvhNode{
-	children:Vec<Self>,
-	models:Vec<u32>,
-	aabb:Aabb,
-}
-
-impl BvhNode{
-	pub fn the_tester<F:FnMut(u32)>(&self,aabb:&Aabb,f:&mut F){
-		for &model in &self.models{
-			f(model);
-		}
-		for child in &self.children{
-			if aabb.intersects(&child.aabb){
-				child.the_tester(aabb,f);
-			}
-		}
-	}
-}
-
-pub fn generate_bvh(boxen:Vec<Aabb>)->BvhNode{
-	generate_bvh_node(boxen.into_iter().enumerate().collect())
-}
-
-fn generate_bvh_node(boxen:Vec<(usize,Aabb)>)->BvhNode{
-	let n=boxen.len();
-	if n<20{
-		let mut aabb=Aabb::new();
-		let models=boxen.into_iter().map(|b|{aabb.join(&b.1);b.0 as u32}).collect();
-		BvhNode{
-			children:Vec::new(),
-			models,
-			aabb,
-		}
-	}else{
-		let mut octant=std::collections::HashMap::with_capacity(n);//this ids which octant the boxen is put in
-		let mut sort_x=Vec::with_capacity(n);
-		let mut sort_y=Vec::with_capacity(n);
-		let mut sort_z=Vec::with_capacity(n);
-		for (i,aabb) in boxen.iter(){
-			let center=aabb.center();
-			octant.insert(*i,0);
-			sort_x.push((*i,center.x));
-			sort_y.push((*i,center.y));
-			sort_z.push((*i,center.z));
-		}
-		sort_x.sort_by(|tup0,tup1|tup0.1.partial_cmp(&tup1.1).unwrap());
-		sort_y.sort_by(|tup0,tup1|tup0.1.partial_cmp(&tup1.1).unwrap());
-		sort_z.sort_by(|tup0,tup1|tup0.1.partial_cmp(&tup1.1).unwrap());
-		let h=n/2;
-		let median_x=sort_x[h].1;
-		let median_y=sort_y[h].1;
-		let median_z=sort_z[h].1;
-		for (i,c) in sort_x{
-			if median_x<c{
-				octant.insert(i,octant[&i]+1<<0);
-			}
-		}
-		for (i,c) in sort_y{
-			if median_y<c{
-				octant.insert(i,octant[&i]+1<<1);
-			}
-		}
-		for (i,c) in sort_z{
-			if median_z<c{
-				octant.insert(i,octant[&i]+1<<2);
-			}
-		}
-		//generate lists for unique octant values
-		let mut list_list=Vec::with_capacity(8);
-		let mut octant_list=Vec::with_capacity(8);
-		for (i,aabb) in boxen.into_iter(){
-			let octant_id=octant[&i];
-			let list_id=if let Some(list_id)=octant_list.iter().position(|&id|id==octant_id){
-				list_id
-			}else{
-				let list_id=list_list.len();
-				octant_list.push(octant_id);
-				list_list.push(Vec::new());
-				list_id
-			};
-			list_list[list_id].push((i,aabb));
-		}
-		let mut aabb=Aabb::new();
-		let children=list_list.into_iter().map(|b|{
-			let node=generate_bvh_node(b);
-			aabb.join(&node.aabb);
-			node
-		}).collect();
-		BvhNode{
-			children,
-			models:Vec::new(),
-			aabb,
-		}
-	}
-}

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 +1,119 @@
+use crate::physics::Body;
+use crate::model_physics::{FEV,MeshQuery,DirectedEdge};
+use strafesnet_common::integer::{Time,Planar64};
+use strafesnet_common::zeroes::zeroes2;
+
+enum Transition<F,E:DirectedEdge,V>{
+	Miss,
+	Next(FEV<F,E,V>,Time),
+	Hit(F,Time),
+}
+
+	fn next_transition<F:Copy,E:Copy+DirectedEdge,V:Copy>(fev:&FEV<F,E,V>,time:Time,mesh:&impl MeshQuery<F,E,V>,body:&Body,time_limit:Time)->Transition<F,E,V>{
+		//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_time=time_limit;
+		let mut best_transtition=Transition::Miss;
+		match fev{
+			&FEV::<F,E,V>::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 t in zeroes2((n.dot(body.position)-d)*2,n.dot(body.velocity)*2,n.dot(body.acceleration)){
+					let t=body.time+Time::from(t);
+					if time<=t&&t<best_time&&n.dot(body.extrapolated_velocity(t))<Planar64::ZERO{
+						best_time=t;
+						best_transtition=Transition::Hit(face_id,t);
+						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!
+					for t in zeroes2(n.dot(body.position*2-(mesh.vert(verts[0])+mesh.vert(verts[1]))),n.dot(body.velocity)*2,n.dot(body.acceleration)){
+						let t=body.time+Time::from(t);
+						if time<=t&&t<best_time&&n.dot(body.extrapolated_velocity(t))<Planar64::ZERO{
+							best_time=t;
+							best_transtition=Transition::Next(FEV::<F,E,V>::Edge(directed_edge_id.as_undirected()),t);
+							break;
+						}
+					}
+				}
+				//if none:
+			},
+			&FEV::<F,E,V>::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 t in zeroes2(n.dot(delta_pos),n.dot(body.velocity)*2,n.dot(body.acceleration)){
+						let t=body.time+Time::from(t);
+						if time<=t&&t<best_time&&n.dot(body.extrapolated_velocity(t))<Planar64::ZERO{
+							best_time=t;
+							best_transtition=Transition::Next(FEV::<F,E,V>::Face(edge_face_id),t);
+							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 t in zeroes2((n.dot(body.position-mesh.vert(vert_id)))*2,n.dot(body.velocity)*2,n.dot(body.acceleration)){
+						let t=body.time+Time::from(t);
+						if time<=t&&t<best_time&&n.dot(body.extrapolated_velocity(t))<Planar64::ZERO{
+							best_time=t;
+							best_transtition=Transition::Next(FEV::<F,E,V>::Vert(vert_id),t);
+							break;
+						}
+					}
+				}
+				//if none:
+			},
+			&FEV::<F,E,V>::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 t in zeroes2((n.dot(body.position-mesh.vert(vert_id)))*2,n.dot(body.velocity)*2,n.dot(body.acceleration)){
+						let t=body.time+Time::from(t);
+						if time<=t&&t<best_time&&n.dot(body.extrapolated_velocity(t))<Planar64::ZERO{
+							best_time=t;
+							best_transtition=Transition::Next(FEV::<F,E,V>::Edge(directed_edge_id.as_undirected()),t);
+							break;
+						}
+					}
+				}
+				//if none:
+			},
+		}
+		best_transtition
+	}
+pub enum CrawlResult<F,E:DirectedEdge,V>{
+	Miss(FEV<F,E,V>),
+	Hit(F,Time),
+}
+pub fn crawl_fev<F:Copy,E:Copy+DirectedEdge,V:Copy>(mut fev:FEV<F,E,V>,mesh:&impl MeshQuery<F,E,V>,relative_body:&Body,start_time:Time,time_limit:Time)->CrawlResult<F,E,V>{	
+	let mut time=start_time;
+	for _ in 0..20{
+		match next_transition(&fev,time,mesh,relative_body,time_limit){
+			Transition::Miss=>return CrawlResult::Miss(fev),
+			Transition::Next(next_fev,next_time)=>(fev,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 +1,141 @@
+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::Dom),
+	#[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(dom)=>{
+			let mut loader=strafesnet_deferred_loader::roblox_legacy();
+
+			let (texture_loader,mesh_loader)=loader.get_inner_mut();
+
+			let map_step1=strafesnet_rbx_loader::convert(
+				&dom,
+				|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(target_arch = "wasm32")]
-use web_sys::{ImageBitmapRenderingContext, OffscreenCanvas};
-use winit::{
-	event::{self, WindowEvent, DeviceEvent},
-	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, window: &winit::window::Window, device: &wgpu::Device, queue: &wgpu::Queue, event: WindowEvent);
-	fn device_event(&mut self, window: &winit::window::Window, event: DeviceEvent);
-	fn load_file(&mut self, path:std::path::PathBuf, device: &wgpu::Device, queue: &wgpu::Queue);
-	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;
-
-	let optional_features = E::optional_features();
-	let required_features = E::required_features();
-
-	//no helper function smh gotta write it myself
-	let adapters = instance.enumerate_adapters(backends);
-
-	let mut chosen_adapter = None;
-	let mut chosen_adapter_score=0;
-	for adapter in adapters {
-		if !adapter.is_surface_supported(&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!");
-	}
-
-	#[cfg(not(target_arch = "wasm32"))]
-	{
-		let adapter_info = adapter.get_info();
-		println!("Using {} ({:?})", adapter_info.name, adapter_info.backend);
-	}
-
-	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);
-
-	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::Scroll),
-							state: event::ElementState::Pressed,
-							..
-						},
-					..
-				} => {
-					println!("{:#?}", instance.generate_report());
-				}
-				_ => {
-					example.update(&window,&device,&queue,event);
-				}
-			},
-			event::Event::DeviceEvent {
-				event,
-				..
-			} => {
-				example.device_event(&window,event);
-			},
-			event::Event::RedrawRequested(_) => {
-
-				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 +1,990 @@
+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};
+
+#[derive(Clone)]
+pub struct GraphicsModelUpdate{
+	transform:Option<glam::Mat4>,
+	color:Option<glam::Vec4>,
+}
+
+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,
+	model_buf:wgpu::Buffer,
+	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,angles):(glam::Vec3,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 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:Into::<glam::Mat3>::into(model.transform.matrix3).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|*Into::<glam::Vec3>::into(v).as_ref()).collect(),
+								unique_tex:mesh.unique_tex.iter().map(|v|*v.as_ref()).collect(),
+								unique_normal:mesh.unique_normal.iter().map(|&v|*Into::<glam::Vec3>::into(v).as_ref()).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,
+					model_buf,
+				});
+			}
+		}
+		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(crate::physics::PhysicsOutputState::default().extrapolate(strafesnet_common::mouse::MouseState::default()));
+		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,
+		physics_output:crate::physics::PhysicsOutputState,
+		predicted_time:integer::Time,
+		mouse_pos:glam::IVec2,
+	){
+		//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(physics_output.extrapolate(strafesnet_common::mouse::MouseState{pos:mouse_pos,time:predicted_time}));
+		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 +1,69 @@
+use strafesnet_common::integer;
+
+pub enum Instruction{
+	Render(crate::physics::PhysicsOutputState,integer::Time,glam::IVec2),
+	//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(physics_output,predicted_time,mouse_pos)=>{
+				if let Some((size,user_settings))=&resize{
+					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());
+				}
+				//clear every time w/e
+				resize=None;
+				//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,physics_output,predicted_time,mouse_pos);
+
+				frame.present();
+			}
+		}
+	})
+}

src/instruction.rs

@@ -1,48 +0,0 @@
-#[derive(Debug)]
-pub struct TimedInstruction<I> {
-	pub time: crate::physics::TIME,
-	pub instruction: I,
-}
-
-pub trait InstructionEmitter<I> {
-	fn next_instruction(&self, time_limit:crate::physics::TIME) -> Option<TimedInstruction<I>>;
-}
-pub trait InstructionConsumer<I> {
-	fn process_instruction(&mut self, instruction:TimedInstruction<I>);
-}
-
-//PROPER PRIVATE FIELDS!!!
-pub struct InstructionCollector<I> {
-	time: crate::physics::TIME,
-	instruction: Option<I>,
-}
-impl<I> InstructionCollector<I> {
-	pub fn new(time:crate::physics::TIME) -> Self {
-		Self{
-			time,
-			instruction:None
-		}
-	}
-
-	pub fn collect(&mut self,instruction:Option<TimedInstruction<I>>){
-		match instruction {
-			Some(unwrap_instruction) => {
-				if unwrap_instruction.time<self.time {
-					self.time=unwrap_instruction.time;
-					self.instruction=Some(unwrap_instruction.instruction);
-				}
-			},
-			None => (),
-		}
-	}
-	pub fn instruction(self) -> Option<TimedInstruction<I>> {
-		//STEAL INSTRUCTION AND DESTROY INSTRUCTIONCOLLECTOR
-		match self.instruction {
-			Some(instruction)=>Some(TimedInstruction{
-				time:self.time,
-				instruction
-			}),
-			None => None,
-		}
-	}
-}

src/load_roblox.rs

@@ -1,478 +0,0 @@
-use crate::primitives;
-
-fn class_is_a(class: &str, superclass: &str) -> bool {
-	if class==superclass {
-		return true
-	}
-	let class_descriptor=rbx_reflection_database::get().classes.get(class);
-	if let Some(descriptor) = &class_descriptor {
-		if let Some(class_super) = &descriptor.superclass {
-			return class_is_a(&class_super, superclass)
-		}
-	}
-	return false
-}
-fn recursive_collect_superclass(objects: &mut std::vec::Vec<rbx_dom_weak::types::Ref>,dom: &rbx_dom_weak::WeakDom, instance: &rbx_dom_weak::Instance, superclass: &str){
-	for &referent in instance.children() {
-		if let Some(c) = dom.get_by_ref(referent) {
-			if class_is_a(c.class.as_str(), superclass) {
-				objects.push(c.referent());//copy ref
-			}
-			recursive_collect_superclass(objects,dom,c,superclass);
-		}
-	}
-}
-fn get_texture_refs(dom:&rbx_dom_weak::WeakDom) -> Vec<rbx_dom_weak::types::Ref>{
-	let mut objects = std::vec::Vec::new();
-	recursive_collect_superclass(&mut objects, dom, dom.root(),"Decal");
-	//get ids
-	//clear vec
-	//next class
-	objects
-}
-fn get_attributes(name:&str,can_collide:bool,velocity:glam::Vec3,force_intersecting:bool)->crate::model::CollisionAttributes{
-	let mut general=crate::model::GameMechanicAttributes::default();
-	let mut intersecting=crate::model::IntersectingAttributes::default();
-	let mut contacting=crate::model::ContactingAttributes::default();
-	let mut force_can_collide=can_collide;
-	match name{
-		//"Water"=>intersecting.water=Some(crate::model::IntersectingWater{density:1.0,drag:1.0}),
-		"Accelerator"=>{force_can_collide=false;intersecting.accelerator=Some(crate::model::IntersectingAccelerator{acceleration:velocity})},
-		"MapFinish"=>{force_can_collide=false;general.zone=Some(crate::model::GameMechanicZone{mode_id:0,behaviour:crate::model::ZoneBehaviour::Finish})},
-		"MapAnticheat"=>{force_can_collide=false;general.zone=Some(crate::model::GameMechanicZone{mode_id:0,behaviour:crate::model::ZoneBehaviour::Anitcheat})},
-		"Platform"=>general.stage_element=Some(crate::model::GameMechanicStageElement{
-			mode_id:0,
-			stage_id:0,
-			force:false,
-			behaviour:crate::model::StageElementBehaviour::Platform,
-		}),
-		other=>{
-			if let Some(captures)=lazy_regex::regex!(r"^(Force)?(Spawn|SpawnAt|Trigger|Teleport|Platform)(\d+)$")
-			.captures(other){
-				general.stage_element=Some(crate::model::GameMechanicStageElement{
-					mode_id:0,
-					stage_id:captures[3].parse::<u32>().unwrap(),
-					force:match captures.get(1){
-						Some(m)=>m.as_str()=="Force",
-						None=>false,
-					},
-					behaviour:match &captures[2]{
-						"Spawn"|"SpawnAt"=>crate::model::StageElementBehaviour::SpawnAt,
-						"Trigger"=>{force_can_collide=false;crate::model::StageElementBehaviour::Trigger},
-						"Teleport"=>{force_can_collide=false;crate::model::StageElementBehaviour::Teleport},
-						"Platform"=>crate::model::StageElementBehaviour::Platform,
-						_=>panic!("regex1[2] messed up bad"),
-					}
-				})
-			}else if let Some(captures)=lazy_regex::regex!(r"^Bonus(Finish|Anticheat)(\d+)$")
-			.captures(other){
-				force_can_collide=false;
-				match &captures[1]{
-					"Finish"=>general.zone=Some(crate::model::GameMechanicZone{mode_id:captures[2].parse::<u32>().unwrap(),behaviour:crate::model::ZoneBehaviour::Finish}),
-					"Anticheat"=>general.zone=Some(crate::model::GameMechanicZone{mode_id:captures[2].parse::<u32>().unwrap(),behaviour:crate::model::ZoneBehaviour::Anitcheat}),
-					_=>panic!("regex2[1] messed up bad"),
-				}
-			}
-		}
-	}
-	//need some way to skip this
-	if velocity!=glam::Vec3::ZERO{
-		general.booster=Some(crate::model::GameMechanicBooster{velocity});
-	}
-	match force_can_collide{
-		true=>{
-			match name{
-				//"Bounce"=>(),
-				"Surf"=>contacting.surf=Some(crate::model::ContactingSurf{}),
-				"Ladder"=>contacting.ladder=Some(crate::model::ContactingLadder{sticky:true}),
-				other=>{
-					//REGEX!!!!
-					//Jump#
-					//WormholeIn#
-				}
-			}
-			crate::model::CollisionAttributes::Contact{contacting,general}
-		},
-		false=>if force_intersecting
-		||general.jump_limit.is_some()
-		||general.booster.is_some()
-		||general.zone.is_some()
-		||general.stage_element.is_some()
-		||general.wormhole.is_some()
-		||intersecting.water.is_some()
-		||intersecting.accelerator.is_some()
-		{
-			crate::model::CollisionAttributes::Intersect{intersecting,general}
-		}else{
-			crate::model::CollisionAttributes::Decoration
-		},
-	}
-}
-
-struct RobloxAssetId(u64);
-struct RobloxAssetIdParseErr;
-impl std::str::FromStr for RobloxAssetId {
-	type Err=RobloxAssetIdParseErr;
-	fn from_str(s: &str) -> Result<Self, Self::Err>{
-		let regman=lazy_regex::regex!(r"(\d+)$");
-		if let Some(captures) = regman.captures(s) {
-			if captures.len()==2{//captures[0] is all captures concatenated, and then each individual capture
-				if let Ok(id) = captures[0].parse::<u64>() {
-					return Ok(Self(id));
-				}
-			}
-		}
-		Err(RobloxAssetIdParseErr)
-	}
-}
-#[derive(Clone,Copy,PartialEq)]
-struct RobloxTextureTransform{
-	offset_u:f32,
-	offset_v:f32,
-	scale_u:f32,
-	scale_v:f32,
-}
-impl std::cmp::Eq for RobloxTextureTransform{}//????
-impl std::default::Default for RobloxTextureTransform{
-    fn default() -> Self {
-        Self{offset_u:0.0,offset_v:0.0,scale_u:1.0,scale_v:1.0}
-    }
-}
-impl std::hash::Hash for RobloxTextureTransform {
-	fn hash<H: std::hash::Hasher>(&self, state: &mut H) {
-		self.offset_u.to_ne_bytes().hash(state);
-		self.offset_v.to_ne_bytes().hash(state);
-		self.scale_u.to_ne_bytes().hash(state);
-		self.scale_v.to_ne_bytes().hash(state);
-	}
-}
-#[derive(Clone,PartialEq)]
-struct RobloxFaceTextureDescription{
-	texture:u32,
-	color:glam::Vec4,
-	transform:RobloxTextureTransform,
-}
-impl std::cmp::Eq for RobloxFaceTextureDescription{}//????
-impl std::hash::Hash for RobloxFaceTextureDescription {
-	fn hash<H: std::hash::Hasher>(&self, state: &mut H) {
-		self.texture.hash(state);
-		self.transform.hash(state);
-        for &el in self.color.as_ref().iter() {
-            el.to_ne_bytes().hash(state);
-        }
-    }
-}
-impl RobloxFaceTextureDescription{
-	fn to_face_description(&self)->primitives::FaceDescription{
-		primitives::FaceDescription{
-			texture:Some(self.texture),
-			transform:glam::Affine2::from_translation(
-				glam::vec2(self.transform.offset_u,self.transform.offset_v)
-			)
-			*glam::Affine2::from_scale(
-				glam::vec2(self.transform.scale_u,self.transform.scale_v)
-			),
-			color:self.color,
-		}
-	}
-}
-type RobloxPartDescription=[Option<RobloxFaceTextureDescription>;6];
-type RobloxWedgeDescription=[Option<RobloxFaceTextureDescription>;5];
-type RobloxCornerWedgeDescription=[Option<RobloxFaceTextureDescription>;5];
-#[derive(Clone,Eq,Hash,PartialEq)]
-enum RobloxBasePartDescription{
-	Sphere,
-	Part(RobloxPartDescription),
-	Cylinder,
-	Wedge(RobloxWedgeDescription),
-	CornerWedge(RobloxCornerWedgeDescription),
-}
-pub fn generate_indexed_models(dom:rbx_dom_weak::WeakDom) -> crate::model::IndexedModelInstances{
-	//IndexedModelInstances includes textures
-	let mut spawn_point=glam::Vec3::ZERO;
-
-	let mut indexed_models=Vec::new();
-	let mut model_id_from_description=std::collections::HashMap::<RobloxBasePartDescription,usize>::new();
-
-	let mut texture_id_from_asset_id=std::collections::HashMap::<u64,u32>::new();
-	let mut asset_id_from_texture_id=Vec::new();
-
-	let mut object_refs=Vec::new();
-	let mut temp_objects=Vec::new();
-	recursive_collect_superclass(&mut object_refs, &dom, dom.root(),"BasePart");
-	for object_ref in object_refs {
-		if let Some(object)=dom.get_by_ref(object_ref){
-			if let (
-					Some(rbx_dom_weak::types::Variant::CFrame(cf)),
-					Some(rbx_dom_weak::types::Variant::Vector3(size)),
-					Some(rbx_dom_weak::types::Variant::Vector3(velocity)),
-					Some(rbx_dom_weak::types::Variant::Float32(transparency)),
-					Some(rbx_dom_weak::types::Variant::Color3uint8(color3)),
-					Some(rbx_dom_weak::types::Variant::Bool(can_collide)),
-				) = (
-					object.properties.get("CFrame"),
-					object.properties.get("Size"),
-					object.properties.get("Velocity"),
-					object.properties.get("Transparency"),
-					object.properties.get("Color"),
-					object.properties.get("CanCollide"),
-				)
-			{
-				let model_transform=glam::Affine3A::from_translation(
-					glam::Vec3::new(cf.position.x,cf.position.y,cf.position.z)
-				)
-				* glam::Affine3A::from_mat3(
-					glam::Mat3::from_cols(
-						glam::Vec3::new(cf.orientation.x.x,cf.orientation.y.x,cf.orientation.z.x),
-						glam::Vec3::new(cf.orientation.x.y,cf.orientation.y.y,cf.orientation.z.y),
-						glam::Vec3::new(cf.orientation.x.z,cf.orientation.y.z,cf.orientation.z.z),
-					),
-				)
-				* glam::Affine3A::from_scale(
-					glam::Vec3::new(size.x,size.y,size.z)/2.0
-				);
-
-				//push TempIndexedAttributes
-				let mut force_intersecting=false;
-				let mut temp_indexing_attributes=Vec::new();
-				if let Some(attr)=match &object.name[..]{
-					"MapStart"=>{
-						spawn_point=model_transform.transform_point3(glam::Vec3::ZERO)+glam::vec3(0.0,2.5,0.0);
-						Some(crate::model::TempIndexedAttributes::Start{mode_id:0})
-					},
-					"UnorderedCheckpoint"=>Some(crate::model::TempIndexedAttributes::UnorderedCheckpoint{mode_id:0}),
-					other=>{
-						let regman=lazy_regex::regex!(r"^(BonusStart|Spawn|ForceSpawn|OrderedCheckpoint)(\d+)$");
-						if let Some(captures) = regman.captures(other) {
-							match &captures[1]{
-								"BonusStart"=>Some(crate::model::TempIndexedAttributes::Start{mode_id:captures[2].parse::<u32>().unwrap()}),
-								"Spawn"|"ForceSpawn"=>Some(crate::model::TempIndexedAttributes::Spawn{mode_id:0,stage_id:captures[2].parse::<u32>().unwrap()}),
-								"OrderedCheckpoint"=>Some(crate::model::TempIndexedAttributes::OrderedCheckpoint{mode_id:0,checkpoint_id:captures[2].parse::<u32>().unwrap()}),
-								_=>None,
-							}
-						}else{
-							None
-						}
-					}
-				}{
-					force_intersecting=true;
-					temp_indexing_attributes.push(attr);
-				}
-
-				//TODO: also detect "CylinderMesh" etc here
-				let shape=match &object.class[..]{
-					"Part"=>{
-						if let Some(rbx_dom_weak::types::Variant::Enum(shape))=object.properties.get("Shape"){
-							match shape.to_u32(){
-								0=>primitives::Primitives::Sphere,
-								1=>primitives::Primitives::Cube,
-								2=>primitives::Primitives::Cylinder,
-								3=>primitives::Primitives::Wedge,
-								4=>primitives::Primitives::CornerWedge,
-								_=>panic!("Funky roblox PartType={};",shape.to_u32()),
-							}
-						}else{
-							panic!("Part has no Shape!");
-						}
-					},
-					"WedgePart"=>primitives::Primitives::Wedge,
-					"CornerWedgePart"=>primitives::Primitives::CornerWedge,
-					_=>{
-						println!("Unsupported BasePart ClassName={}; defaulting to cube",object.class);
-						primitives::Primitives::Cube
-					}
-				};
-
-				//use the biggest one and cut it down later...
-				let mut part_texture_description:RobloxPartDescription=[None,None,None,None,None,None];
-				temp_objects.clear();
-				recursive_collect_superclass(&mut temp_objects, &dom, object,"Decal");
-				for &decal_ref in &temp_objects{
-					if let Some(decal)=dom.get_by_ref(decal_ref){
-						if let (
-							Some(rbx_dom_weak::types::Variant::Content(content)),
-							Some(rbx_dom_weak::types::Variant::Enum(normalid)),
-							Some(rbx_dom_weak::types::Variant::Color3(decal_color3)),
-							Some(rbx_dom_weak::types::Variant::Float32(decal_transparency)),
-						) = (
-							decal.properties.get("Texture"),
-							decal.properties.get("Face"),
-							decal.properties.get("Color3"),
-							decal.properties.get("Transparency"),
-						) {
-							if let Ok(asset_id)=content.clone().into_string().parse::<RobloxAssetId>(){
-								let texture_id=if let Some(&texture_id)=texture_id_from_asset_id.get(&asset_id.0){
-									texture_id
-								}else{
-									let texture_id=asset_id_from_texture_id.len() as u32;
-									texture_id_from_asset_id.insert(asset_id.0,texture_id);
-									asset_id_from_texture_id.push(asset_id.0);
-									texture_id
-								};
-								let normal_id=normalid.to_u32();
-								if normal_id<6{
-									let (roblox_texture_color,roblox_texture_transform)=if decal.class=="Texture"{
-										//generate tranform
-										if let (
-												Some(rbx_dom_weak::types::Variant::Float32(ox)),
-												Some(rbx_dom_weak::types::Variant::Float32(oy)),
-												Some(rbx_dom_weak::types::Variant::Float32(sx)),
-												Some(rbx_dom_weak::types::Variant::Float32(sy)),
-											) = (
-												decal.properties.get("OffsetStudsU"),
-												decal.properties.get("OffsetStudsV"),
-												decal.properties.get("StudsPerTileU"),
-												decal.properties.get("StudsPerTileV"),
-											)
-										{
-											let (size_u,size_v)=match normal_id{
-												0=>(size.z,size.y),//right
-												1=>(size.x,size.z),//top
-												2=>(size.x,size.y),//back
-												3=>(size.z,size.y),//left
-												4=>(size.x,size.z),//bottom
-												5=>(size.x,size.y),//front
-												_=>panic!("unreachable"),
-											};
-											(
-												glam::vec4(decal_color3.r,decal_color3.g,decal_color3.b,1.0-*decal_transparency),
-												RobloxTextureTransform{
-													offset_u:*ox/(*sx),offset_v:*oy/(*sy),
-													scale_u:size_u/(*sx),scale_v:size_v/(*sy),
-												}
-											)
-										}else{
-											(glam::Vec4::ONE,RobloxTextureTransform::default())
-										}
-									}else{
-										(glam::Vec4::ONE,RobloxTextureTransform::default())
-									};
-									part_texture_description[normal_id as usize]=Some(RobloxFaceTextureDescription{
-										texture:texture_id,
-										color:roblox_texture_color,
-										transform:roblox_texture_transform,
-									});
-								}else{
-									println!("NormalId={} unsupported for shape={:?}",normal_id,shape);
-								}
-							}
-						}
-					}
-				}
-				//obscure rust syntax "slice pattern"
-				let [
-					f0,//Cube::Right
-					f1,//Cube::Top
-					f2,//Cube::Back
-					f3,//Cube::Left
-					f4,//Cube::Bottom
-					f5,//Cube::Front
-				]=part_texture_description;
-				let basepart_texture_description=match shape{
-					primitives::Primitives::Sphere=>RobloxBasePartDescription::Sphere,
-					primitives::Primitives::Cube=>RobloxBasePartDescription::Part([f0,f1,f2,f3,f4,f5]),
-					primitives::Primitives::Cylinder=>RobloxBasePartDescription::Cylinder,
-					//use front face texture first and use top face texture as a fallback
-					primitives::Primitives::Wedge=>RobloxBasePartDescription::Wedge([
-						f0,//Cube::Right->Wedge::Right
-						if f5.is_some(){f5}else{f1},//Cube::Front|Cube::Top->Wedge::TopFront
-						f2,//Cube::Back->Wedge::Back
-						f3,//Cube::Left->Wedge::Left
-						f4,//Cube::Bottom->Wedge::Bottom
-					]),
-					//TODO: fix Left+Back texture coordinates to match roblox when not overwridden by Top
-					primitives::Primitives::CornerWedge=>RobloxBasePartDescription::CornerWedge([
-						f0,//Cube::Right->CornerWedge::Right
-						if f2.is_some(){f2}else{f1.clone()},//Cube::Back|Cube::Top->CornerWedge::TopBack
-						if f3.is_some(){f3}else{f1},//Cube::Left|Cube::Top->CornerWedge::TopLeft
-						f4,//Cube::Bottom->CornerWedge::Bottom
-						f5,//Cube::Front->CornerWedge::Front
-					]),
-				};
-				//make new model if unit cube has not been created before
-				let model_id=if let Some(&model_id)=model_id_from_description.get(&basepart_texture_description){
-					//push to existing texture model
-					model_id
-				}else{
-					let model_id=indexed_models.len();
-					model_id_from_description.insert(basepart_texture_description.clone(),model_id);//borrow checker going crazy
-					indexed_models.push(match basepart_texture_description{
-						RobloxBasePartDescription::Sphere=>primitives::unit_sphere(),
-						RobloxBasePartDescription::Part(part_texture_description)=>{
-							let mut cube_face_description=primitives::CubeFaceDescription::new();
-							for (face_id,roblox_face_description) in part_texture_description.iter().enumerate(){
-								cube_face_description.insert(
-								match face_id{
-									0=>primitives::CubeFace::Right,
-									1=>primitives::CubeFace::Top,
-									2=>primitives::CubeFace::Back,
-									3=>primitives::CubeFace::Left,
-									4=>primitives::CubeFace::Bottom,
-									5=>primitives::CubeFace::Front,
-									_=>panic!("unreachable"),
-								},
-								match roblox_face_description{
-									Some(roblox_texture_transform)=>roblox_texture_transform.to_face_description(),
-									None=>primitives::FaceDescription::default(),
-								});
-							}
-							primitives::generate_partial_unit_cube(cube_face_description)
-						},
-						RobloxBasePartDescription::Cylinder=>primitives::unit_cylinder(),
-						RobloxBasePartDescription::Wedge(wedge_texture_description)=>{
-							let mut wedge_face_description=primitives::WedgeFaceDescription::new();
-							for (face_id,roblox_face_description) in wedge_texture_description.iter().enumerate(){
-								wedge_face_description.insert(
-								match face_id{
-									0=>primitives::WedgeFace::Right,
-									1=>primitives::WedgeFace::TopFront,
-									2=>primitives::WedgeFace::Back,
-									3=>primitives::WedgeFace::Left,
-									4=>primitives::WedgeFace::Bottom,
-									_=>panic!("unreachable"),
-								},
-								match roblox_face_description{
-									Some(roblox_texture_transform)=>roblox_texture_transform.to_face_description(),
-									None=>primitives::FaceDescription::default(),
-								});
-							}
-							primitives::generate_partial_unit_wedge(wedge_face_description)
-						},
-						RobloxBasePartDescription::CornerWedge(cornerwedge_texture_description)=>{
-							let mut cornerwedge_face_description=primitives::CornerWedgeFaceDescription::new();
-							for (face_id,roblox_face_description) in cornerwedge_texture_description.iter().enumerate(){
-								cornerwedge_face_description.insert(
-								match face_id{
-									0=>primitives::CornerWedgeFace::Right,
-									1=>primitives::CornerWedgeFace::TopBack,
-									2=>primitives::CornerWedgeFace::TopLeft,
-									3=>primitives::CornerWedgeFace::Bottom,
-									4=>primitives::CornerWedgeFace::Front,
-									_=>panic!("unreachable"),
-								},
-								match roblox_face_description{
-									Some(roblox_texture_transform)=>roblox_texture_transform.to_face_description(),
-									None=>primitives::FaceDescription::default(),
-								});
-							}
-							primitives::generate_partial_unit_cornerwedge(cornerwedge_face_description)
-						},
-					});
-					model_id
-				};
-				indexed_models[model_id].instances.push(crate::model::ModelInstance {
-					transform:model_transform,
-					color:glam::vec4(color3.r as f32/255f32, color3.g as f32/255f32, color3.b as f32/255f32, 1.0-*transparency),
-					attributes:get_attributes(&object.name,*can_collide,glam::vec3(velocity.x,velocity.y,velocity.z),force_intersecting),
-					temp_indexing:temp_indexing_attributes,
-				});
-			}
-		}
-	}
-	crate::model::IndexedModelInstances{
-		textures:asset_id_from_texture_id.iter().map(|t|t.to_string()).collect(),
-		models:indexed_models,
-		spawn_point,
-		modes:Vec::new(),
-	}
-}

src/model.rs

@@ -1,269 +0,0 @@
-use bytemuck::{Pod, Zeroable};
-#[derive(Clone, Copy, Pod, Zeroable)]
-#[repr(C)]
-pub struct Vertex {
-	pub pos: [f32; 3],
-	pub tex: [f32; 2],
-	pub normal: [f32; 3],
-	pub color: [f32; 4],
-}
-#[derive(Clone,Hash,PartialEq,Eq)]
-pub struct IndexedVertex{
-	pub pos:u32,
-	pub tex:u32,
-	pub normal:u32,
-	pub color:u32,
-}
-pub struct IndexedPolygon{
-	pub vertices:Vec<u32>,
-}
-pub struct IndexedGroup{
-	pub texture:Option<u32>,//RenderPattern? material/texture/shader/flat color
-	pub polys:Vec<IndexedPolygon>,
-}
-pub struct IndexedModel{
-	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 groups: Vec<IndexedGroup>,
-	pub instances:Vec<ModelInstance>,
-}
-pub struct IndexedGroupFixedTexture{
-	pub polys:Vec<IndexedPolygon>,
-}
-pub struct IndexedModelSingleTexture{
-	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 texture:Option<u32>,//RenderPattern? material/texture/shader/flat color
-	pub groups: Vec<IndexedGroupFixedTexture>,
-	pub instances:Vec<ModelGraphicsInstance>,
-}
-pub struct ModelSingleTexture{
-	pub instances: Vec<ModelGraphicsInstance>,
-	pub vertices: Vec<Vertex>,
-	pub entities: Vec<Vec<u16>>,
-	pub texture: Option<u32>,
-}
-#[derive(Clone)]
-pub struct ModelGraphicsInstance{
-	pub transform:glam::Mat4,
-	pub normal_transform:glam::Mat3,
-	pub color:glam::Vec4,
-}
-pub struct ModelInstance{
-	//pub id:u64,//this does not actually help with map fixes resimulating bots, they must always be resimulated
-	pub transform:glam::Affine3A,
-	pub color:glam::Vec4,//transparency is in here
-	pub attributes:CollisionAttributes,
-	pub temp_indexing:Vec<TempIndexedAttributes>,
-}
-impl std::default::Default for ModelInstance{
-	fn default() -> Self {
-		Self{
-			color:glam::Vec4::ONE,
-			transform:Default::default(),
-			attributes:Default::default(),
-			temp_indexing:Default::default(),
-		}
-	}
-}
-pub struct IndexedModelInstances{
-	pub textures:Vec<String>,//RenderPattern
-	pub models:Vec<IndexedModel>,
-	//may make this into an object later.
-	pub modes:Vec<ModeDescription>,
-	pub spawn_point:glam::Vec3,
-}
-//stage description referencing flattened ids is spooky, but the map loading is meant to be deterministic.
-pub struct ModeDescription{
-	pub start:u32,//start=model_id
-	pub spawns:Vec<u32>,//spawns[spawn_id]=model_id
-	pub ordered_checkpoints:Vec<u32>,//ordered_checkpoints[checkpoint_id]=model_id
-	pub unordered_checkpoints:Vec<u32>,//unordered_checkpoints[checkpoint_id]=model_id
-	pub spawn_from_stage_id:std::collections::HashMap::<u32,usize>,
-	pub ordered_checkpoint_from_checkpoint_id:std::collections::HashMap::<u32,usize>,
-}
-impl ModeDescription{
-	pub fn get_spawn_model_id(&self,stage_id:u32)->Option<&u32>{
-		if let Some(&spawn)=self.spawn_from_stage_id.get(&stage_id){
-			self.spawns.get(spawn)
-		}else{
-			None
-		}
-	}
-	pub fn get_ordered_checkpoint_model_id(&self,checkpoint_id:u32)->Option<&u32>{
-		if let Some(&checkpoint)=self.ordered_checkpoint_from_checkpoint_id.get(&checkpoint_id){
-			self.ordered_checkpoints.get(checkpoint)
-		}else{
-			None
-		}
-	}
-}
-pub enum TempIndexedAttributes{
-	Start{
-		mode_id:u32,
-	},
-	Spawn{
-		mode_id:u32,
-		stage_id:u32,
-	},
-	OrderedCheckpoint{
-		mode_id:u32,
-		checkpoint_id:u32,
-	},
-	UnorderedCheckpoint{
-		mode_id:u32,
-	},
-}
-
-//you have this effect while in contact
-#[derive(Clone)]
-pub struct ContactingSurf{}
-#[derive(Clone)]
-pub struct ContactingLadder{
-	pub sticky:bool
-}
-//you have this effect while intersecting
-#[derive(Clone)]
-pub struct IntersectingWater{
-	pub viscosity:i64,
-	pub density:i64,
-	pub current:glam::Vec3,
-}
-#[derive(Clone)]
-pub struct IntersectingAccelerator{
-	pub acceleration:glam::Vec3
-}
-//All models can be given these attributes
-#[derive(Clone)]
-pub struct GameMechanicJumpLimit{
-	pub count:u32,
-}
-#[derive(Clone)]
-pub struct GameMechanicBooster{
-	pub velocity:glam::Vec3,
-}
-#[derive(Clone)]
-pub enum ZoneBehaviour{
-	//Start is indexed
-	//Checkpoints are indexed
-	Finish,
-	Anitcheat,
-}
-#[derive(Clone)]
-pub struct GameMechanicZone{
-	pub mode_id:u32,
-	pub behaviour:ZoneBehaviour,
-}
-// enum TrapCondition{
-// 	FasterThan(i64),
-// 	SlowerThan(i64),
-// 	InRange(i64,i64),
-// 	OutsideRange(i64,i64),
-// }
-#[derive(Clone)]
-pub enum StageElementBehaviour{
- 	//Spawn,//The behaviour of stepping on a spawn setting the spawnid
- 	SpawnAt,
- 	Trigger,
- 	Teleport,
- 	Platform,
- 	//Speedtrap(TrapCondition),//Acts as a trigger with a speed condition
-}
-#[derive(Clone)]
-pub struct GameMechanicStageElement{
-	pub mode_id:u32,
-	pub stage_id:u32,//which spawn to send to
-	pub force:bool,//allow setting to lower spawn id i.e. 7->3
-	pub behaviour:StageElementBehaviour
-}
-#[derive(Clone)]
-pub struct GameMechanicWormhole{//(position,angles)*=origin.transform.inverse()*destination.transform
-	pub model_id:u32,
-}
-#[derive(Default,Clone)]
-pub struct GameMechanicAttributes{
-	pub jump_limit:Option<GameMechanicJumpLimit>,
-	pub booster:Option<GameMechanicBooster>,
-	pub zone:Option<GameMechanicZone>,
-	pub stage_element:Option<GameMechanicStageElement>,
-	pub wormhole:Option<GameMechanicWormhole>,//stage_element and wormhole are in conflict
-}
-#[derive(Default,Clone)]
-pub struct ContactingAttributes{
-	pub elasticity:Option<u32>,//[1/2^32,1] 0=None (elasticity+1)/2^32
-	//friction?
-	pub surf:Option<ContactingSurf>,
-	pub ladder:Option<ContactingLadder>,
-}
-#[derive(Default,Clone)]
-pub struct IntersectingAttributes{
-	pub water:Option<IntersectingWater>,
-	pub accelerator:Option<IntersectingAccelerator>,
-}
-//Spawn(u32) NO! spawns are indexed in the map header instead of marked with attibutes
-pub enum CollisionAttributes{
-	Decoration,//visual only
-	Contact{//track whether you are contacting the object
-		contacting:ContactingAttributes,
-		general:GameMechanicAttributes,
-	},
-	Intersect{//track whether you are intersecting the object
-		intersecting:IntersectingAttributes,
-		general:GameMechanicAttributes,
-	},
-}
-impl std::default::Default for CollisionAttributes{
-	fn default() -> Self {
-		Self::Contact{
-			contacting:ContactingAttributes::default(),
-			general:GameMechanicAttributes::default()
-		}
-	}
-}
-
-pub fn generate_indexed_model_list_from_obj(data:obj::ObjData,color:[f32;4]) -> Vec<IndexedModel>{
-	let mut unique_vertex_index = std::collections::HashMap::<obj::IndexTuple,u32>::new();
-	return data.objects.iter().map(|object|{
-		unique_vertex_index.clear();
-		let mut unique_vertices = Vec::new();
-		let groups = object.groups.iter().map(|group|{
-			IndexedGroup{
-				texture:None,
-				polys:group.polys.iter().map(|poly|{
-					IndexedPolygon{
-						vertices:poly.0.iter().map(|&tup|{
-							if let Some(&i)=unique_vertex_index.get(&tup){
-								i
-							}else{
-								let i=unique_vertices.len() as u32;
-								unique_vertices.push(IndexedVertex{
-									pos: tup.0 as u32,
-									tex: tup.1.unwrap() as u32,
-									normal: tup.2.unwrap() as u32,
-									color: 0,
-								});
-								unique_vertex_index.insert(tup,i);
-								i
-							}
-						}).collect()
-					}
-				}).collect()
-			}
-		}).collect();
-		IndexedModel{
-			unique_pos: data.position.clone(),
-			unique_tex: data.texture.clone(),
-			unique_normal: data.normal.clone(),
-			unique_color: vec![color],
-			unique_vertices,
-			groups,
-			instances:Vec::new(),
-		}
-	}).collect()
-}

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/model_physics.rs

@@ -0,0 +1,981 @@
+use std::borrow::{Borrow,Cow};
+use std::collections::{HashSet,HashMap};
+use strafesnet_common::model::{self,MeshId,PolygonIter};
+use strafesnet_common::zeroes;
+use strafesnet_common::integer::{self,Planar64,Planar64Vec3};
+
+pub trait UndirectedEdge{
+	type DirectedEdge:Copy+DirectedEdge;
+	fn as_directed(&self,parity:bool)->Self::DirectedEdge;
+}
+pub trait DirectedEdge{
+	type UndirectedEdge:Copy+UndirectedEdge;
+	fn as_undirected(&self)->Self::UndirectedEdge;
+	fn parity(&self)->bool;
+	//this is stupid but may work fine
+	fn reverse(&self)-><<Self as DirectedEdge>::UndirectedEdge as UndirectedEdge>::DirectedEdge{
+		self.as_undirected().as_directed(!self.parity())
+	}
+}
+
+#[derive(Debug,Clone,Copy,Hash,id::Id,Eq,PartialEq)]
+pub struct MeshVertId(u32);
+#[derive(Debug,Clone,Copy,Hash,id::Id,Eq,PartialEq)]
+pub struct MeshFaceId(u32);
+
+#[derive(Debug,Clone,Copy,Hash,id::Id,Eq,PartialEq)]
+pub struct SubmeshVertId(u32);
+#[derive(Debug,Clone,Copy,Hash,id::Id,Eq,PartialEq)]
+pub struct SubmeshEdgeId(u32);
+/// DirectedEdgeId refers to an EdgeId when undirected.
+#[derive(Debug,Clone,Copy,Hash,id::Id,Eq,PartialEq)]
+pub struct SubmeshDirectedEdgeId(u32);
+#[derive(Debug,Clone,Copy,Hash,id::Id,Eq,PartialEq)]
+pub struct SubmeshFaceId(u32);
+
+impl UndirectedEdge for SubmeshEdgeId{
+	type DirectedEdge=SubmeshDirectedEdgeId;
+	fn as_directed(&self,parity:bool)->SubmeshDirectedEdgeId{
+		SubmeshDirectedEdgeId(self.0|((parity as u32)<<(u32::BITS-1)))
+	}
+}
+impl DirectedEdge for SubmeshDirectedEdgeId{
+	type UndirectedEdge=SubmeshEdgeId;
+	fn as_undirected(&self)->SubmeshEdgeId{
+		SubmeshEdgeId(self.0&!(1<<(u32::BITS-1)))
+	}
+	fn parity(&self)->bool{
+		self.0&(1<<(u32::BITS-1))!=0
+	}
+}
+
+//Vertex <-> Edge <-> Face -> Collide
+pub enum FEV<F,E:DirectedEdge,V>{
+	Face(F),
+	Edge(E::UndirectedEdge),
+	Vert(V),
+}
+
+//use Unit32 #[repr(C)] for map files
+#[derive(Clone,Hash,Eq,PartialEq)]
+struct Face{
+	normal:Planar64Vec3,
+	dot:Planar64,
+}
+struct Vert(Planar64Vec3);
+pub trait MeshQuery<FACE:Clone,EDGE:Clone+DirectedEdge,VERT:Clone>{
+	fn edge_n(&self,edge_id:EDGE::UndirectedEdge)->Planar64Vec3{
+		let verts=self.edge_verts(edge_id);
+		self.vert(verts[1].clone())-self.vert(verts[0].clone())
+	}
+	fn directed_edge_n(&self,directed_edge_id:EDGE)->Planar64Vec3{
+		let verts=self.edge_verts(directed_edge_id.as_undirected());
+		(self.vert(verts[1].clone())-self.vert(verts[0].clone()))*((directed_edge_id.parity() as i64)*2-1)
+	}
+	fn vert(&self,vert_id:VERT)->Planar64Vec3;
+	fn face_nd(&self,face_id:FACE)->(Planar64Vec3,Planar64);
+	fn face_edges(&self,face_id:FACE)->Cow<Vec<EDGE>>;
+	fn edge_faces(&self,edge_id:EDGE::UndirectedEdge)->Cow<[FACE;2]>;
+	fn edge_verts(&self,edge_id:EDGE::UndirectedEdge)->Cow<[VERT;2]>;
+	fn vert_edges(&self,vert_id:VERT)->Cow<Vec<EDGE>>;
+	fn vert_faces(&self,vert_id:VERT)->Cow<Vec<FACE>>;
+}
+struct FaceRefs{
+	edges:Vec<SubmeshDirectedEdgeId>,
+	//verts:Vec<VertId>,
+}
+struct EdgeRefs{
+	faces:[SubmeshFaceId;2],//left, right
+	verts:[SubmeshVertId;2],//bottom, top
+}
+struct VertRefs{
+	faces:Vec<SubmeshFaceId>,
+	edges:Vec<SubmeshDirectedEdgeId>,
+}
+pub struct PhysicsMeshData{
+	//this contains all real and virtual faces used in both the complete mesh and convex submeshes
+	//faces are sorted such that all faces that belong to the complete mesh appear first, and then
+	//all remaining faces are virtual to operate internal logic of the face crawler
+	//and cannot be part of a physics collision
+	//virtual faces are only used in convex submeshes.
+	faces:Vec<Face>,//MeshFaceId indexes this list
+	verts:Vec<Vert>,//MeshVertId indexes this list
+}
+pub struct PhysicsMeshTopology{
+	//mapping of local ids to PhysicsMeshData ids
+	faces:Vec<MeshFaceId>,//SubmeshFaceId indexes this list
+	verts:Vec<MeshVertId>,//SubmeshVertId indexes this list
+	//all ids here are local to this object
+	face_topology:Vec<FaceRefs>,
+	edge_topology:Vec<EdgeRefs>,
+	vert_topology:Vec<VertRefs>,
+}
+#[derive(Clone,Copy,Hash,id::Id,Eq,PartialEq)]
+pub struct PhysicsMeshId(u32);
+impl Into<MeshId> for PhysicsMeshId{
+	fn into(self)->MeshId{
+		MeshId::new(self.0)
+	}
+}
+impl From<MeshId> for PhysicsMeshId{
+	fn from(value:MeshId)->Self{
+		Self::new(value.get())
+	}
+}
+#[derive(Debug,Default,Clone,Copy,Hash,id::Id,Eq,PartialEq)]
+pub struct PhysicsSubmeshId(u32);
+pub struct PhysicsMesh{
+	data:PhysicsMeshData,
+	complete_mesh:PhysicsMeshTopology,
+	//Most objects in roblox maps are already convex, so the list length is 0
+	//as soon as the mesh is divided into 2 submeshes, the list length jumps to 2.
+	//length 1 is unnecessary since the complete mesh would be a duplicate of the only submesh, but would still function properly
+	submeshes:Vec<PhysicsMeshTopology>,
+}
+impl PhysicsMesh{
+	pub fn unit_cube()->Self{
+		//go go gadget debug print mesh
+		let data=PhysicsMeshData{
+			faces:vec![
+				Face{normal:Planar64Vec3::raw_xyz( 4294967296, 0, 0),dot:Planar64::raw(4294967296)},
+				Face{normal:Planar64Vec3::raw_xyz( 0, 4294967296, 0),dot:Planar64::raw(4294967296)},
+				Face{normal:Planar64Vec3::raw_xyz( 0, 0, 4294967296),dot:Planar64::raw(4294967296)},
+				Face{normal:Planar64Vec3::raw_xyz(-4294967296, 0, 0),dot:Planar64::raw(4294967296)},
+				Face{normal:Planar64Vec3::raw_xyz( 0,-4294967296, 0),dot:Planar64::raw(4294967296)},
+				Face{normal:Planar64Vec3::raw_xyz( 0, 0,-4294967296),dot:Planar64::raw(4294967296)}
+			],
+			verts:vec![
+				Vert(Planar64Vec3::raw_xyz( 4294967296,-4294967296,-4294967296)),
+				Vert(Planar64Vec3::raw_xyz( 4294967296, 4294967296,-4294967296)),
+				Vert(Planar64Vec3::raw_xyz( 4294967296, 4294967296, 4294967296)),
+				Vert(Planar64Vec3::raw_xyz( 4294967296,-4294967296, 4294967296)),
+				Vert(Planar64Vec3::raw_xyz(-4294967296, 4294967296,-4294967296)),
+				Vert(Planar64Vec3::raw_xyz(-4294967296, 4294967296, 4294967296)),
+				Vert(Planar64Vec3::raw_xyz(-4294967296,-4294967296, 4294967296)),
+				Vert(Planar64Vec3::raw_xyz(-4294967296,-4294967296,-4294967296))
+			]
+		};
+		let mesh_topology=PhysicsMeshTopology{
+			faces:(0..data.faces.len() as u32).map(MeshFaceId::new).collect(),
+			verts:(0..data.verts.len() as u32).map(MeshVertId::new).collect(),
+			face_topology:vec![
+				FaceRefs{edges:vec![SubmeshDirectedEdgeId((9223372036854775808u64-(1<<63)+(1<<31)) as u32),SubmeshDirectedEdgeId((9223372036854775809u64-(1<<63)+(1<<31)) as u32),SubmeshDirectedEdgeId((9223372036854775810u64-(1<<63)+(1<<31)) as u32),SubmeshDirectedEdgeId(3)]},
+				FaceRefs{edges:vec![SubmeshDirectedEdgeId((9223372036854775812u64-(1<<63)+(1<<31)) as u32),SubmeshDirectedEdgeId((9223372036854775813u64-(1<<63)+(1<<31)) as u32),SubmeshDirectedEdgeId(6),SubmeshDirectedEdgeId(1)]},
+				FaceRefs{edges:vec![SubmeshDirectedEdgeId(7),SubmeshDirectedEdgeId(2),SubmeshDirectedEdgeId((9223372036854775814u64-(1<<63)+(1<<31)) as u32),SubmeshDirectedEdgeId((9223372036854775816u64-(1<<63)+(1<<31)) as u32)]},
+				FaceRefs{edges:vec![SubmeshDirectedEdgeId(8),SubmeshDirectedEdgeId(5),SubmeshDirectedEdgeId((9223372036854775817u64-(1<<63)+(1<<31)) as u32),SubmeshDirectedEdgeId(10)]},
+				FaceRefs{edges:vec![SubmeshDirectedEdgeId((9223372036854775815u64-(1<<63)+(1<<31)) as u32),SubmeshDirectedEdgeId((9223372036854775818u64-(1<<63)+(1<<31)) as u32),SubmeshDirectedEdgeId(11),SubmeshDirectedEdgeId((9223372036854775811u64-(1<<63)+(1<<31)) as u32)]},
+				FaceRefs{edges:vec![SubmeshDirectedEdgeId(4),SubmeshDirectedEdgeId(0),SubmeshDirectedEdgeId((9223372036854775819u64-(1<<63)+(1<<31)) as u32),SubmeshDirectedEdgeId(9)]}
+			],
+			edge_topology:vec![
+				EdgeRefs{faces:[SubmeshFaceId(0),SubmeshFaceId(5)],verts:[SubmeshVertId(0),SubmeshVertId(1)]},
+				EdgeRefs{faces:[SubmeshFaceId(0),SubmeshFaceId(1)],verts:[SubmeshVertId(1),SubmeshVertId(2)]},
+				EdgeRefs{faces:[SubmeshFaceId(0),SubmeshFaceId(2)],verts:[SubmeshVertId(2),SubmeshVertId(3)]},
+				EdgeRefs{faces:[SubmeshFaceId(4),SubmeshFaceId(0)],verts:[SubmeshVertId(0),SubmeshVertId(3)]},
+				EdgeRefs{faces:[SubmeshFaceId(1),SubmeshFaceId(5)],verts:[SubmeshVertId(1),SubmeshVertId(4)]},
+				EdgeRefs{faces:[SubmeshFaceId(1),SubmeshFaceId(3)],verts:[SubmeshVertId(4),SubmeshVertId(5)]},
+				EdgeRefs{faces:[SubmeshFaceId(2),SubmeshFaceId(1)],verts:[SubmeshVertId(2),SubmeshVertId(5)]},
+				EdgeRefs{faces:[SubmeshFaceId(4),SubmeshFaceId(2)],verts:[SubmeshVertId(3),SubmeshVertId(6)]},
+				EdgeRefs{faces:[SubmeshFaceId(2),SubmeshFaceId(3)],verts:[SubmeshVertId(5),SubmeshVertId(6)]},
+				EdgeRefs{faces:[SubmeshFaceId(3),SubmeshFaceId(5)],verts:[SubmeshVertId(4),SubmeshVertId(7)]},
+				EdgeRefs{faces:[SubmeshFaceId(4),SubmeshFaceId(3)],verts:[SubmeshVertId(6),SubmeshVertId(7)]},
+				EdgeRefs{faces:[SubmeshFaceId(5),SubmeshFaceId(4)],verts:[SubmeshVertId(0),SubmeshVertId(7)]}
+			],
+			vert_topology:vec![
+				VertRefs{faces:vec![SubmeshFaceId(0),SubmeshFaceId(4),SubmeshFaceId(5)],edges:vec![SubmeshDirectedEdgeId((9223372036854775811u64-(1<<63)+(1<<31)) as u32),SubmeshDirectedEdgeId((9223372036854775819u64-(1<<63)+(1<<31)) as u32),SubmeshDirectedEdgeId((9223372036854775808u64-(1<<63)+(1<<31)) as u32)]},
+				VertRefs{faces:vec![SubmeshFaceId(0),SubmeshFaceId(5),SubmeshFaceId(1)],edges:vec![SubmeshDirectedEdgeId((9223372036854775812u64-(1<<63)+(1<<31)) as u32),SubmeshDirectedEdgeId(0),SubmeshDirectedEdgeId((9223372036854775809u64-(1<<63)+(1<<31)) as u32)]},
+				VertRefs{faces:vec![SubmeshFaceId(0),SubmeshFaceId(2),SubmeshFaceId(1)],edges:vec![SubmeshDirectedEdgeId(1),SubmeshDirectedEdgeId((9223372036854775810u64-(1<<63)+(1<<31)) as u32),SubmeshDirectedEdgeId((9223372036854775814u64-(1<<63)+(1<<31)) as u32)]},
+				VertRefs{faces:vec![SubmeshFaceId(0),SubmeshFaceId(2),SubmeshFaceId(4)],edges:vec![SubmeshDirectedEdgeId(2),SubmeshDirectedEdgeId(3),SubmeshDirectedEdgeId((9223372036854775815u64-(1<<63)+(1<<31)) as u32)]},
+				VertRefs{faces:vec![SubmeshFaceId(3),SubmeshFaceId(5),SubmeshFaceId(1)],edges:vec![SubmeshDirectedEdgeId(4),SubmeshDirectedEdgeId((9223372036854775817u64-(1<<63)+(1<<31)) as u32),SubmeshDirectedEdgeId((9223372036854775813u64-(1<<63)+(1<<31)) as u32)]},
+				VertRefs{faces:vec![SubmeshFaceId(2),SubmeshFaceId(3),SubmeshFaceId(1)],edges:vec![SubmeshDirectedEdgeId(5),SubmeshDirectedEdgeId(6),SubmeshDirectedEdgeId((9223372036854775816u64-(1<<63)+(1<<31)) as u32)]},
+				VertRefs{faces:vec![SubmeshFaceId(2),SubmeshFaceId(3),SubmeshFaceId(4)],edges:vec![SubmeshDirectedEdgeId(7),SubmeshDirectedEdgeId(8),SubmeshDirectedEdgeId((9223372036854775818u64-(1<<63)+(1<<31)) as u32)]},
+				VertRefs{faces:vec![SubmeshFaceId(4),SubmeshFaceId(3),SubmeshFaceId(5)],edges:vec![SubmeshDirectedEdgeId(10),SubmeshDirectedEdgeId(11),SubmeshDirectedEdgeId(9)]}
+			]
+		};
+		Self{
+			data,
+			complete_mesh:mesh_topology,
+			submeshes:Vec::new(),
+		}
+	}
+	pub fn unit_cylinder()->Self{
+		Self::unit_cube()
+	}
+	#[inline]
+	pub const fn complete_mesh(&self)->&PhysicsMeshTopology{
+		&self.complete_mesh
+	}
+	#[inline]
+	pub const fn complete_mesh_view(&self)->PhysicsMeshView{
+		PhysicsMeshView{
+			data:&self.data,
+			topology:self.complete_mesh(),
+		}
+	}
+	#[inline]
+	pub fn submeshes(&self)->&[PhysicsMeshTopology]{
+		//the complete mesh is already a convex mesh when len()==0, len()==1 is invalid but will still work
+		if self.submeshes.len()==0{
+			std::slice::from_ref(&self.complete_mesh)
+		}else{
+			&self.submeshes.as_slice()
+		}
+	}
+	#[inline]
+	pub fn submesh_view(&self,submesh_id:PhysicsSubmeshId)->PhysicsMeshView{
+		PhysicsMeshView{
+			data:&self.data,
+			topology:&self.submeshes()[submesh_id.get() as usize],
+		}
+	}
+	pub fn submesh_views(&self)->impl Iterator<Item=PhysicsMeshView>{
+		self.submeshes().iter().map(|topology|PhysicsMeshView{
+			data:&self.data,
+			topology,
+		})
+	}
+}
+
+//mesh builder code
+#[derive(Default,Clone)]
+struct VertRefGuy{
+	edges:HashSet<SubmeshDirectedEdgeId>,
+	faces:HashSet<SubmeshFaceId>,
+}
+#[derive(Clone,Hash,Eq,PartialEq)]
+struct EdgeRefVerts([SubmeshVertId;2]);
+impl EdgeRefVerts{
+	const fn new(v0:SubmeshVertId,v1:SubmeshVertId)->(Self,bool){
+		(if v0.0<v1.0{
+			Self([v0,v1])
+		}else{
+			Self([v1,v0])
+		},v0.0<v1.0)
+	}
+}
+struct EdgeRefFaces([SubmeshFaceId;2]);
+impl EdgeRefFaces{
+	const fn new()->Self{
+		Self([SubmeshFaceId(0);2])
+	}
+	fn push(&mut self,i:usize,face_id:SubmeshFaceId){
+		self.0[i]=face_id;
+	}
+}
+struct FaceRefEdges(Vec<SubmeshDirectedEdgeId>);
+#[derive(Default)]
+struct EdgePool{
+	edge_guys:Vec<(EdgeRefVerts,EdgeRefFaces)>,
+	edge_id_from_guy:HashMap<EdgeRefVerts,SubmeshEdgeId>,
+}
+impl EdgePool{
+	fn push(&mut self,edge_ref_verts:EdgeRefVerts)->(&mut EdgeRefFaces,SubmeshEdgeId){
+		let edge_id=if let Some(&edge_id)=self.edge_id_from_guy.get(&edge_ref_verts){
+			edge_id
+		}else{
+			let edge_id=SubmeshEdgeId::new(self.edge_guys.len() as u32);
+			self.edge_guys.push((edge_ref_verts.clone(),EdgeRefFaces::new()));
+			self.edge_id_from_guy.insert(edge_ref_verts,edge_id);
+			edge_id
+		};
+		(&mut unsafe{self.edge_guys.get_unchecked_mut(edge_id.get() as usize)}.1,edge_id)
+	}
+}
+
+#[derive(Debug)]
+pub enum PhysicsMeshError{
+	ZeroVertices,
+	NoPhysicsGroups,
+}
+impl std::fmt::Display for PhysicsMeshError{
+	fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+		write!(f,"{self:?}")
+	}
+}
+impl std::error::Error for PhysicsMeshError{}
+
+impl TryFrom<&model::Mesh> for PhysicsMesh{
+	type Error=PhysicsMeshError;
+	fn try_from(mesh:&model::Mesh)->Result<Self,PhysicsMeshError>{
+		if mesh.unique_pos.len()==0{
+			return Err(PhysicsMeshError::ZeroVertices);
+		}
+		let verts=mesh.unique_pos.iter().copied().map(Vert).collect();
+		//TODO: fix submeshes
+		//flat map mesh.physics_groups[$1].groups.polys()[$2] as face_id
+		//lower face_id points to upper face_id
+		//the same face is not allowed to be in multiple polygon groups
+		let mut faces=Vec::new();
+		let mut face_id_from_face=HashMap::new();
+		let mut mesh_topologies:Vec<PhysicsMeshTopology>=mesh.physics_groups.iter().map(|physics_group|{
+			//construct submesh
+			let mut submesh_faces=Vec::new();//these contain a map from submeshId->meshId
+			let mut submesh_verts=Vec::new();
+			let mut submesh_vert_id_from_mesh_vert_id=HashMap::<MeshVertId,SubmeshVertId>::new();
+			//lazy closure
+			let mut get_submesh_vert_id=|vert_id:MeshVertId|{
+				if let Some(&submesh_vert_id)=submesh_vert_id_from_mesh_vert_id.get(&vert_id){
+					submesh_vert_id
+				}else{
+					let submesh_vert_id=SubmeshVertId::new(submesh_verts.len() as u32);
+					submesh_verts.push(vert_id);
+					submesh_vert_id_from_mesh_vert_id.insert(vert_id,submesh_vert_id);
+					submesh_vert_id
+				}
+			};
+			let mut edge_pool=EdgePool::default();
+			let mut vert_ref_guys=vec![VertRefGuy::default();mesh.unique_pos.len()];
+			let mut face_ref_guys=Vec::new();
+			for polygon_group_id in &physics_group.groups{
+				let polygon_group=&mesh.polygon_groups[polygon_group_id.get() as usize];
+				for poly_vertices in polygon_group.polys(){
+					let submesh_face_id=SubmeshFaceId::new(submesh_faces.len() as u32);
+					//one face per poly
+					let mut normal=Planar64Vec3::ZERO;
+					let len=poly_vertices.len();
+					let face_edges=poly_vertices.into_iter().enumerate().map(|(i,vert_id)|{
+						let vert0_id=MeshVertId::new(mesh.unique_vertices[vert_id.get() as usize].pos.get() as u32);
+						let vert1_id=MeshVertId::new(mesh.unique_vertices[poly_vertices[(i+1)%len].get() as usize].pos.get() as u32);
+						//index submesh verts
+						let submesh_vert0_id=get_submesh_vert_id(vert0_id);
+						let submesh_vert1_id=get_submesh_vert_id(vert1_id);
+						//https://www.khronos.org/opengl/wiki/Calculating_a_Surface_Normal (Newell's Method)
+						let v0=mesh.unique_pos[vert0_id.get() as usize];
+						let v1=mesh.unique_pos[vert1_id.get() as usize];
+						normal+=Planar64Vec3::new(
+							(v0.y()-v1.y())*(v0.z()+v1.z()),
+							(v0.z()-v1.z())*(v0.x()+v1.x()),
+							(v0.x()-v1.x())*(v0.y()+v1.y()),
+						);
+						//get/create edge and push face into it
+						let (edge_ref_verts,is_sorted)=EdgeRefVerts::new(submesh_vert0_id,submesh_vert1_id);
+						let (edge_ref_faces,edge_id)=edge_pool.push(edge_ref_verts);
+						//polygon vertices as assumed to be listed clockwise
+						//populate the edge face on the left or right depending on how the edge vertices got sorted
+						edge_ref_faces.push(!is_sorted as usize,submesh_face_id);
+						//index edges & face into vertices
+						{
+							let vert_ref_guy=unsafe{vert_ref_guys.get_unchecked_mut(submesh_vert0_id.get() as usize)};
+							vert_ref_guy.edges.insert(edge_id.as_directed(is_sorted));
+							vert_ref_guy.faces.insert(submesh_face_id);
+							unsafe{vert_ref_guys.get_unchecked_mut(submesh_vert1_id.get() as usize)}.edges.insert(edge_id.as_directed(!is_sorted));
+						}
+						//return directed_edge_id
+						edge_id.as_directed(is_sorted)
+					}).collect();
+					//choose precision loss randomly idk
+					normal=normal/len as i64;
+					let mut dot=Planar64::ZERO;
+					for &v in poly_vertices{
+						dot+=normal.dot(mesh.unique_pos[mesh.unique_vertices[v.get() as usize].pos.get() as usize]);
+					}
+					//assume face hash is stable, and there are no flush faces...
+					let face=Face{normal,dot:dot/len as i64};
+					let face_id=match face_id_from_face.get(&face){
+						Some(&face_id)=>face_id,
+						None=>{
+							let face_id=MeshFaceId::new(faces.len() as u32);
+							face_id_from_face.insert(face.clone(),face_id);
+							faces.push(face);
+							face_id
+						}
+					};
+					submesh_faces.push(face_id);
+					face_ref_guys.push(FaceRefEdges(face_edges));
+				}
+			}
+			PhysicsMeshTopology{
+				faces:submesh_faces,
+				verts:submesh_verts,
+				face_topology:face_ref_guys.into_iter().map(|face_ref_guy|{
+					FaceRefs{edges:face_ref_guy.0}
+				}).collect(),
+				edge_topology:edge_pool.edge_guys.into_iter().map(|(edge_ref_verts,edge_ref_faces)|
+					EdgeRefs{faces:edge_ref_faces.0,verts:edge_ref_verts.0}
+				).collect(),
+				vert_topology:vert_ref_guys.into_iter().map(|vert_ref_guy|
+					VertRefs{
+						edges:vert_ref_guy.edges.into_iter().collect(),
+						faces:vert_ref_guy.faces.into_iter().collect(),
+					}
+				).collect(),
+			}
+		}).collect();
+		Ok(Self{
+			data:PhysicsMeshData{
+				faces,
+				verts,
+			},
+			complete_mesh:mesh_topologies.pop().ok_or(PhysicsMeshError::NoPhysicsGroups)?,
+			submeshes:mesh_topologies,
+		})
+	}
+}
+
+pub struct PhysicsMeshView<'a>{
+	data:&'a PhysicsMeshData,
+	topology:&'a PhysicsMeshTopology,
+}
+impl MeshQuery<SubmeshFaceId,SubmeshDirectedEdgeId,SubmeshVertId> for PhysicsMeshView<'_>{
+	fn face_nd(&self,face_id:SubmeshFaceId)->(Planar64Vec3,Planar64){
+		let face_idx=self.topology.faces[face_id.get() as usize].get() as usize;
+		(self.data.faces[face_idx].normal,self.data.faces[face_idx].dot)
+	}
+	//ideally I never calculate the vertex position, but I have to for the graphical meshes...
+	fn vert(&self,vert_id:SubmeshVertId)->Planar64Vec3{
+		let vert_idx=self.topology.verts[vert_id.get() as usize].get() as usize;
+		self.data.verts[vert_idx].0
+	}
+	fn face_edges(&self,face_id:SubmeshFaceId)->Cow<Vec<SubmeshDirectedEdgeId>>{
+		Cow::Borrowed(&self.topology.face_topology[face_id.get() as usize].edges)
+	}
+	fn edge_faces(&self,edge_id:SubmeshEdgeId)->Cow<[SubmeshFaceId;2]>{
+		Cow::Borrowed(&self.topology.edge_topology[edge_id.get() as usize].faces)
+	}
+	fn edge_verts(&self,edge_id:SubmeshEdgeId)->Cow<[SubmeshVertId;2]>{
+		Cow::Borrowed(&self.topology.edge_topology[edge_id.get() as usize].verts)
+	}
+	fn vert_edges(&self,vert_id:SubmeshVertId)->Cow<Vec<SubmeshDirectedEdgeId>>{
+		Cow::Borrowed(&self.topology.vert_topology[vert_id.get() as usize].edges)
+	}
+	fn vert_faces(&self,vert_id:SubmeshVertId)->Cow<Vec<SubmeshFaceId>>{
+		Cow::Borrowed(&self.topology.vert_topology[vert_id.get() as usize].faces)
+	}
+}
+
+pub struct PhysicsMeshTransform{
+	pub vertex:integer::Planar64Affine3,
+	pub normal:integer::Planar64Mat3,
+	pub det:Planar64,
+}
+impl PhysicsMeshTransform{
+	pub const fn new(transform:integer::Planar64Affine3)->Self{
+		Self{
+			normal:transform.matrix3.inverse_times_det().transpose(),
+			det:transform.matrix3.determinant(),
+			vertex:transform,
+		}
+	}
+}
+
+pub struct TransformedMesh<'a>{
+	view:PhysicsMeshView<'a>,
+	transform:&'a PhysicsMeshTransform,
+}
+impl TransformedMesh<'_>{
+	pub fn new<'a>(
+		view:PhysicsMeshView<'a>,
+		transform:&'a PhysicsMeshTransform,
+	)->TransformedMesh<'a>{
+		TransformedMesh{
+			view,
+			transform,
+		}
+	}
+	pub fn verts<'a>(&'a self)->impl Iterator<Item=Planar64Vec3>+'a{
+		self.view.data.verts.iter().map(|&Vert(pos)|self.transform.vertex.transform_point3(pos))
+	}
+	fn farthest_vert(&self,dir:Planar64Vec3)->SubmeshVertId{
+		let mut best_dot=Planar64::MIN;
+		let mut best_vert=SubmeshVertId(0);
+		//this happens to be well-defined.  there are no virtual virtices
+		for (i,vert_id) in self.view.topology.verts.iter().enumerate(){
+			let p=self.transform.vertex.transform_point3(self.view.data.verts[vert_id.get() as usize].0);
+			let d=dir.dot(p);
+			if best_dot<d{
+				best_dot=d;
+				best_vert=SubmeshVertId::new(i as u32);
+			}
+		}
+		best_vert
+	}
+}
+impl MeshQuery<SubmeshFaceId,SubmeshDirectedEdgeId,SubmeshVertId> for TransformedMesh<'_>{
+	fn face_nd(&self,face_id:SubmeshFaceId)->(Planar64Vec3,Planar64){
+		let (n,d)=self.view.face_nd(face_id);
+		let transformed_n=self.transform.normal*n;
+		let transformed_d=d+transformed_n.dot(self.transform.vertex.translation)/self.transform.det;
+		(transformed_n/self.transform.det,transformed_d)
+	}
+	fn vert(&self,vert_id:SubmeshVertId)->Planar64Vec3{
+		self.transform.vertex.transform_point3(self.view.vert(vert_id))
+	}
+	#[inline]
+	fn face_edges(&self,face_id:SubmeshFaceId)->Cow<Vec<SubmeshDirectedEdgeId>>{
+		self.view.face_edges(face_id)
+	}
+	#[inline]
+	fn edge_faces(&self,edge_id:SubmeshEdgeId)->Cow<[SubmeshFaceId;2]>{
+		self.view.edge_faces(edge_id)
+	}
+	#[inline]
+	fn edge_verts(&self,edge_id:SubmeshEdgeId)->Cow<[SubmeshVertId;2]>{
+		self.view.edge_verts(edge_id)
+	}
+	#[inline]
+	fn vert_edges(&self,vert_id:SubmeshVertId)->Cow<Vec<SubmeshDirectedEdgeId>>{
+		self.view.vert_edges(vert_id)
+	}
+	#[inline]
+	fn vert_faces(&self,vert_id:SubmeshVertId)->Cow<Vec<SubmeshFaceId>>{
+		self.view.vert_faces(vert_id)
+	}
+}
+
+//Note that a face on a minkowski mesh refers to a pair of fevs on the meshes it's summed from
+//(face,vertex)
+//(edge,edge)
+//(vertex,face)
+#[derive(Clone,Copy)]
+pub enum MinkowskiVert{
+	VertVert(SubmeshVertId,SubmeshVertId),
+}
+#[derive(Clone,Copy)]
+pub enum MinkowskiEdge{
+	VertEdge(SubmeshVertId,SubmeshEdgeId),
+	EdgeVert(SubmeshEdgeId,SubmeshVertId),
+	//EdgeEdge when edges are parallel
+}
+impl UndirectedEdge for MinkowskiEdge{
+	type DirectedEdge=MinkowskiDirectedEdge;
+	fn as_directed(&self,parity:bool)->Self::DirectedEdge{
+		match self{
+			MinkowskiEdge::VertEdge(v0,e1)=>MinkowskiDirectedEdge::VertEdge(*v0,e1.as_directed(parity)),
+			MinkowskiEdge::EdgeVert(e0,v1)=>MinkowskiDirectedEdge::EdgeVert(e0.as_directed(parity),*v1),
+		}
+	}
+}
+#[derive(Clone,Copy)]
+pub enum MinkowskiDirectedEdge{
+	VertEdge(SubmeshVertId,SubmeshDirectedEdgeId),
+	EdgeVert(SubmeshDirectedEdgeId,SubmeshVertId),
+	//EdgeEdge when edges are parallel
+}
+impl DirectedEdge for MinkowskiDirectedEdge{
+	type UndirectedEdge=MinkowskiEdge;
+	fn as_undirected(&self)->Self::UndirectedEdge{
+		match self{
+			MinkowskiDirectedEdge::VertEdge(v0,e1)=>MinkowskiEdge::VertEdge(*v0,e1.as_undirected()),
+			MinkowskiDirectedEdge::EdgeVert(e0,v1)=>MinkowskiEdge::EdgeVert(e0.as_undirected(),*v1),
+		}
+	}
+	fn parity(&self)->bool{
+		match self{
+			MinkowskiDirectedEdge::VertEdge(_,e)
+			|MinkowskiDirectedEdge::EdgeVert(e,_)=>e.parity(),
+		}
+	}
+}
+#[derive(Debug,Clone,Copy,Hash,Eq,PartialEq)]
+pub enum MinkowskiFace{
+	VertFace(SubmeshVertId,SubmeshFaceId),
+	EdgeEdge(SubmeshEdgeId,SubmeshEdgeId,bool),
+	FaceVert(SubmeshFaceId,SubmeshVertId),
+	//EdgeFace
+	//FaceEdge
+	//FaceFace
+}
+
+pub struct MinkowskiMesh<'a>{
+	mesh0:TransformedMesh<'a>,
+	mesh1:TransformedMesh<'a>,
+}
+
+//infinity fev algorithm state transition
+enum Transition{
+	Done,//found closest vert, no edges are better
+	Vert(MinkowskiVert),//transition to vert
+}
+enum EV{
+	Vert(MinkowskiVert),
+	Edge(MinkowskiEdge),
+}
+
+impl MinkowskiMesh<'_>{
+	pub fn minkowski_sum<'a>(mesh0:TransformedMesh<'a>,mesh1:TransformedMesh<'a>)->MinkowskiMesh<'a>{
+		MinkowskiMesh{
+			mesh0,
+			mesh1,
+		}
+	}
+	fn farthest_vert(&self,dir:Planar64Vec3)->MinkowskiVert{
+		MinkowskiVert::VertVert(self.mesh0.farthest_vert(dir),self.mesh1.farthest_vert(-dir))
+	}
+	fn next_transition_vert(&self,vert_id:MinkowskiVert,best_distance_squared:&mut Planar64,infinity_dir:Planar64Vec3,point:Planar64Vec3)->Transition{
+		let mut best_transition=Transition::Done;
+		for &directed_edge_id in self.vert_edges(vert_id).iter(){
+			let edge_n=self.directed_edge_n(directed_edge_id);
+			//is boundary uncrossable by a crawl from infinity
+			let edge_verts=self.edge_verts(directed_edge_id.as_undirected());
+			//select opposite vertex
+			let test_vert_id=edge_verts[directed_edge_id.parity() as usize];
+			//test if it's closer
+			let diff=point-self.vert(test_vert_id);
+			if zeroes::zeroes1(edge_n.dot(diff),edge_n.dot(infinity_dir)).len()==0{
+				let distance_squared=diff.dot(diff);
+				if distance_squared<*best_distance_squared{
+					best_transition=Transition::Vert(test_vert_id);
+					*best_distance_squared=distance_squared;
+				}
+			}
+		}
+		best_transition
+	}
+	fn final_ev(&self,vert_id:MinkowskiVert,best_distance_squared:&mut Planar64,infinity_dir:Planar64Vec3,point:Planar64Vec3)->EV{
+		let mut best_transition=EV::Vert(vert_id);
+		let diff=point-self.vert(vert_id);
+		for &directed_edge_id in self.vert_edges(vert_id).iter(){
+			let edge_n=self.directed_edge_n(directed_edge_id);
+			//is boundary uncrossable by a crawl from infinity
+			//check if time of collision is outside Time::MIN..Time::MAX
+			let d=edge_n.dot(diff);
+			if zeroes::zeroes1(d,edge_n.dot(infinity_dir)).len()==0{
+				//test the edge
+				let edge_nn=edge_n.dot(edge_n);
+				if Planar64::ZERO<=d&&d<=edge_nn{
+					let distance_squared={
+						let c=diff.cross(edge_n);
+						c.dot(c)/edge_nn
+					};
+					if distance_squared<=*best_distance_squared{
+						best_transition=EV::Edge(directed_edge_id.as_undirected());
+						*best_distance_squared=distance_squared;
+					}
+				}
+			}
+		}
+		best_transition
+	}
+	fn crawl_boundaries(&self,mut vert_id:MinkowskiVert,infinity_dir:Planar64Vec3,point:Planar64Vec3)->EV{
+		let mut best_distance_squared={
+			let diff=point-self.vert(vert_id);
+			diff.dot(diff)
+		};
+		loop{
+			match self.next_transition_vert(vert_id,&mut best_distance_squared,infinity_dir,point){
+				Transition::Done=>return self.final_ev(vert_id,&mut best_distance_squared,infinity_dir,point),
+				Transition::Vert(new_vert_id)=>vert_id=new_vert_id,
+			}
+		}
+	}
+	/// This function drops a vertex down to an edge or a face if the path from infinity did not cross any vertex-edge boundaries but the point is supposed to have already crossed a boundary down from a vertex
+	fn infinity_fev(&self,infinity_dir:Planar64Vec3,point:Planar64Vec3)->FEV::<MinkowskiFace,MinkowskiDirectedEdge,MinkowskiVert>{
+		//start on any vertex
+		//cross uncrossable vertex-edge boundaries until you find the closest vertex or edge
+		//cross edge-face boundary if it's uncrossable
+		match self.crawl_boundaries(self.farthest_vert(infinity_dir),infinity_dir,point){
+			//if a vert is returned, it is the closest point to the infinity point
+			EV::Vert(vert_id)=>FEV::<MinkowskiFace,MinkowskiDirectedEdge,MinkowskiVert>::Vert(vert_id),
+			EV::Edge(edge_id)=>{
+				//cross to face if the boundary is not crossable and we are on the wrong side
+				let edge_n=self.edge_n(edge_id);
+				// point is multiplied by two because vert_sum sums two vertices.
+				let delta_pos=point*2-{
+					let &[v0,v1]=self.edge_verts(edge_id).borrow();
+					self.vert(v0)+self.vert(v1)
+				};
+				for (i,&face_id) in self.edge_faces(edge_id).iter().enumerate(){
+					let face_n=self.face_nd(face_id).0;
+					//edge-face boundary nd, n facing out of the face towards the edge
+					let boundary_n=face_n.cross(edge_n)*(i as i64*2-1);
+					let boundary_d=boundary_n.dot(delta_pos);
+					//check if time of collision is outside Time::MIN..Time::MAX
+					//infinity_dir can always be treated as a velocity
+					if (boundary_d)<=Planar64::ZERO&&zeroes::zeroes1(boundary_d,boundary_n.dot(infinity_dir)*2).len()==0{
+						//both faces cannot pass this condition, return early if one does.
+						return FEV::<MinkowskiFace,MinkowskiDirectedEdge,MinkowskiVert>::Face(face_id);
+					}
+				}
+				FEV::<MinkowskiFace,MinkowskiDirectedEdge,MinkowskiVert>::Edge(edge_id)
+			},
+		}
+	}
+	fn closest_fev_not_inside(&self,mut infinity_body:crate::physics::Body)->Option<FEV::<MinkowskiFace,MinkowskiDirectedEdge,MinkowskiVert>>{
+		infinity_body.infinity_dir().map_or(None,|dir|{
+			let infinity_fev=self.infinity_fev(-dir,infinity_body.position);
+			//a line is simpler to solve than a parabola
+			infinity_body.velocity=dir;
+			infinity_body.acceleration=Planar64Vec3::ZERO;
+			//crawl in from negative infinity along a tangent line to get the closest fev
+			match crate::face_crawler::crawl_fev(infinity_fev,self,&infinity_body,integer::Time::MIN,infinity_body.time){
+				crate::face_crawler::CrawlResult::Miss(fev)=>Some(fev),
+				crate::face_crawler::CrawlResult::Hit(_,_)=>None,
+			}
+		})
+	}
+	pub fn predict_collision_in(&self,relative_body:&crate::physics::Body,time_limit:integer::Time)->Option<(MinkowskiFace,integer::Time)>{
+		self.closest_fev_not_inside(relative_body.clone()).map_or(None,|fev|{
+			//continue forwards along the body parabola
+			match crate::face_crawler::crawl_fev(fev,self,relative_body,relative_body.time,time_limit){
+				crate::face_crawler::CrawlResult::Miss(_)=>None,
+				crate::face_crawler::CrawlResult::Hit(face,time)=>Some((face,time)),
+			}
+		})
+	}
+	pub fn predict_collision_out(&self,relative_body:&crate::physics::Body,time_limit:integer::Time)->Option<(MinkowskiFace,integer::Time)>{
+		//create an extrapolated body at time_limit
+		let infinity_body=crate::physics::Body::new(
+			relative_body.extrapolated_position(time_limit),
+			-relative_body.extrapolated_velocity(time_limit),
+			relative_body.acceleration,
+			-time_limit,
+		);
+		self.closest_fev_not_inside(infinity_body).map_or(None,|fev|{
+			//continue backwards along the body parabola
+			match crate::face_crawler::crawl_fev(fev,self,&-relative_body.clone(),-time_limit,-relative_body.time){
+				crate::face_crawler::CrawlResult::Miss(_)=>None,
+				crate::face_crawler::CrawlResult::Hit(face,time)=>Some((face,-time)),//no need to test -time<time_limit because of the first step
+			}
+		})
+	}
+	pub fn predict_collision_face_out(&self,relative_body:&crate::physics::Body,time_limit:integer::Time,contact_face_id:MinkowskiFace)->Option<(MinkowskiEdge,integer::Time)>{
+		//no algorithm needed, there is only one state and two cases (Edge,None)
+		//determine when it passes an edge ("sliding off" case)
+		let mut best_time=time_limit;
+		let mut best_edge=None;
+		let face_n=self.face_nd(contact_face_id).0;
+		for &directed_edge_id in self.face_edges(contact_face_id).iter(){
+			let edge_n=self.directed_edge_n(directed_edge_id);
+			//f x e points in
+			let n=face_n.cross(edge_n);
+			let verts=self.edge_verts(directed_edge_id.as_undirected());
+			let d=n.dot(self.vert(verts[0])+self.vert(verts[1]));
+			//WARNING! d outside of *2
+			for t in zeroes::zeroes2((n.dot(relative_body.position))*2-d,n.dot(relative_body.velocity)*2,n.dot(relative_body.acceleration)){
+				let t=relative_body.time+integer::Time::from(t);
+				if relative_body.time<t&&t<best_time&&n.dot(relative_body.extrapolated_velocity(t))<Planar64::ZERO{
+					best_time=t;
+					best_edge=Some(directed_edge_id);
+					break;
+				}
+			}
+		}
+		best_edge.map(|e|(e.as_undirected(),best_time))
+	}
+	fn infinity_in(&self,infinity_body:crate::physics::Body)->Option<(MinkowskiFace,integer::Time)>{
+		let infinity_fev=self.infinity_fev(-infinity_body.velocity,infinity_body.position);
+		match crate::face_crawler::crawl_fev(infinity_fev,self,&infinity_body,integer::Time::MIN,infinity_body.time){
+			crate::face_crawler::CrawlResult::Miss(_)=>None,
+			crate::face_crawler::CrawlResult::Hit(face,time)=>Some((face,time)),
+		}
+	}
+	pub fn is_point_in_mesh(&self,point:Planar64Vec3)->bool{
+		let infinity_body=crate::physics::Body::new(point,Planar64Vec3::Y,Planar64Vec3::ZERO,integer::Time::ZERO);
+		//movement must escape the mesh forwards and backwards in time,
+		//otherwise the point is not inside the mesh
+		self.infinity_in(infinity_body)
+    	.is_some_and(|_|
+			self.infinity_in(-infinity_body)
+    		.is_some()
+     	)
+	}
+}
+impl MeshQuery<MinkowskiFace,MinkowskiDirectedEdge,MinkowskiVert> for MinkowskiMesh<'_>{
+	fn face_nd(&self,face_id:MinkowskiFace)->(Planar64Vec3,Planar64){
+		match face_id{
+			MinkowskiFace::VertFace(v0,f1)=>{
+				let (n,d)=self.mesh1.face_nd(f1);
+				(-n,d-n.dot(self.mesh0.vert(v0)))
+			},
+			MinkowskiFace::EdgeEdge(e0,e1,parity)=>{
+				let edge0_n=self.mesh0.edge_n(e0);
+				let edge1_n=self.mesh1.edge_n(e1);
+				let &[e0v0,e0v1]=self.mesh0.edge_verts(e0).borrow();
+				let &[e1v0,e1v1]=self.mesh1.edge_verts(e1).borrow();
+				let n=edge0_n.cross(edge1_n);
+				let e0d=n.dot(self.mesh0.vert(e0v0)+self.mesh0.vert(e0v1));
+				let e1d=n.dot(self.mesh1.vert(e1v0)+self.mesh1.vert(e1v1));
+				(n*(parity as i64*4-2),(e0d-e1d)*(parity as i64*2-1))
+			},
+			MinkowskiFace::FaceVert(f0,v1)=>{
+				let (n,d)=self.mesh0.face_nd(f0);
+				(n,d-n.dot(self.mesh1.vert(v1)))
+			},
+		}
+	}
+	fn vert(&self,vert_id:MinkowskiVert)->Planar64Vec3{
+		match vert_id{
+			MinkowskiVert::VertVert(v0,v1)=>{
+				self.mesh0.vert(v0)-self.mesh1.vert(v1)
+			},
+		}
+	}
+	fn face_edges(&self,face_id:MinkowskiFace)->Cow<Vec<MinkowskiDirectedEdge>>{
+		match face_id{
+			MinkowskiFace::VertFace(v0,f1)=>{
+				Cow::Owned(self.mesh1.face_edges(f1).iter().map(|&edge_id1|{
+					MinkowskiDirectedEdge::VertEdge(v0,edge_id1.reverse())
+				}).collect())
+			},
+			MinkowskiFace::EdgeEdge(e0,e1,parity)=>{
+				let e0v=self.mesh0.edge_verts(e0);
+				let e1v=self.mesh1.edge_verts(e1);
+				//could sort this if ordered edges are needed
+				//probably just need to reverse this list according to parity
+				Cow::Owned(vec![
+					MinkowskiDirectedEdge::VertEdge(e0v[0],e1.as_directed(parity)),
+					MinkowskiDirectedEdge::EdgeVert(e0.as_directed(!parity),e1v[0]),
+					MinkowskiDirectedEdge::VertEdge(e0v[1],e1.as_directed(!parity)),
+					MinkowskiDirectedEdge::EdgeVert(e0.as_directed(parity),e1v[1]),
+				])
+			},
+			MinkowskiFace::FaceVert(f0,v1)=>{
+				Cow::Owned(self.mesh0.face_edges(f0).iter().map(|&edge_id0|{
+					MinkowskiDirectedEdge::EdgeVert(edge_id0,v1)
+				}).collect())
+			},
+		}
+	}
+	fn edge_faces(&self,edge_id:MinkowskiEdge)->Cow<[MinkowskiFace;2]>{
+		match edge_id{
+			MinkowskiEdge::VertEdge(v0,e1)=>{
+				//faces are listed backwards from the minkowski mesh
+				let v0e=self.mesh0.vert_edges(v0);
+				let &[e1f0,e1f1]=self.mesh1.edge_faces(e1).borrow();
+				Cow::Owned([(e1f1,false),(e1f0,true)].map(|(edge_face_id1,face_parity)|{
+					let mut best_edge=None;
+					let mut best_d=Planar64::ZERO;
+					let edge_face1_n=self.mesh1.face_nd(edge_face_id1).0;
+					let edge_face1_nn=edge_face1_n.dot(edge_face1_n);
+					for &directed_edge_id0 in v0e.iter(){
+						let edge0_n=self.mesh0.directed_edge_n(directed_edge_id0);
+						//must be behind other face.
+						let d=edge_face1_n.dot(edge0_n);
+						if d<Planar64::ZERO{
+							let edge0_nn=edge0_n.dot(edge0_n);
+							//divide by zero???
+							let dd=d*d/(edge_face1_nn*edge0_nn);
+							if best_d<dd{
+								best_d=dd;
+								best_edge=Some(directed_edge_id0);
+							}
+						}
+					}
+					best_edge.map_or(
+						MinkowskiFace::VertFace(v0,edge_face_id1),
+						|directed_edge_id0|MinkowskiFace::EdgeEdge(directed_edge_id0.as_undirected(),e1,directed_edge_id0.parity()^face_parity)
+					)
+				}))
+			},
+			MinkowskiEdge::EdgeVert(e0,v1)=>{
+				//tracking index with an external variable because .enumerate() is not available
+				let v1e=self.mesh1.vert_edges(v1);
+				let &[e0f0,e0f1]=self.mesh0.edge_faces(e0).borrow();
+				Cow::Owned([(e0f0,true),(e0f1,false)].map(|(edge_face_id0,face_parity)|{
+					let mut best_edge=None;
+					let mut best_d=Planar64::ZERO;
+					let edge_face0_n=self.mesh0.face_nd(edge_face_id0).0;
+					let edge_face0_nn=edge_face0_n.dot(edge_face0_n);
+					for &directed_edge_id1 in v1e.iter(){
+						let edge1_n=self.mesh1.directed_edge_n(directed_edge_id1);
+						let d=edge_face0_n.dot(edge1_n);
+						if d<Planar64::ZERO{
+							let edge1_nn=edge1_n.dot(edge1_n);
+							let dd=d*d/(edge_face0_nn*edge1_nn);
+							if best_d<dd{
+								best_d=dd;
+								best_edge=Some(directed_edge_id1);
+							}
+						}
+					}
+					best_edge.map_or(
+						MinkowskiFace::FaceVert(edge_face_id0,v1),
+						|directed_edge_id1|MinkowskiFace::EdgeEdge(e0,directed_edge_id1.as_undirected(),directed_edge_id1.parity()^face_parity)
+					)
+				}))
+			},
+		}
+	}
+	fn edge_verts(&self,edge_id:MinkowskiEdge)->Cow<[MinkowskiVert;2]>{
+		match edge_id{
+			MinkowskiEdge::VertEdge(v0,e1)=>{
+				Cow::Owned(self.mesh1.edge_verts(e1).map(|vert_id1|{
+					MinkowskiVert::VertVert(v0,vert_id1)
+				}))
+			},
+			MinkowskiEdge::EdgeVert(e0,v1)=>{
+				Cow::Owned(self.mesh0.edge_verts(e0).map(|vert_id0|{
+					MinkowskiVert::VertVert(vert_id0,v1)
+				}))
+			},
+		}
+	}
+	fn vert_edges(&self,vert_id:MinkowskiVert)->Cow<Vec<MinkowskiDirectedEdge>>{
+		match vert_id{
+			MinkowskiVert::VertVert(v0,v1)=>{
+				let mut edges=Vec::new();
+				//detect shared volume when the other mesh is mirrored along a test edge dir
+				let v0f=self.mesh0.vert_faces(v0);
+				let v1f=self.mesh1.vert_faces(v1);
+				let v0f_n:Vec<Planar64Vec3>=v0f.iter().map(|&face_id|self.mesh0.face_nd(face_id).0).collect();
+				let v1f_n:Vec<Planar64Vec3>=v1f.iter().map(|&face_id|self.mesh1.face_nd(face_id).0).collect();
+				let the_len=v0f.len()+v1f.len();
+				for &directed_edge_id in self.mesh0.vert_edges(v0).iter(){
+					let n=self.mesh0.directed_edge_n(directed_edge_id);
+					let nn=n.dot(n);
+					//make a set of faces
+					let mut face_normals=Vec::with_capacity(the_len);
+					//add mesh0 faces as-is
+					face_normals.clone_from(&v0f_n);
+					for face_n in &v1f_n{
+						//add reflected mesh1 faces
+						face_normals.push(*face_n-n*(face_n.dot(n)*2/nn));
+					}
+					if is_empty_volume(face_normals){
+						edges.push(MinkowskiDirectedEdge::EdgeVert(directed_edge_id,v1));
+					}
+				}
+				for &directed_edge_id in self.mesh1.vert_edges(v1).iter(){
+					let n=self.mesh1.directed_edge_n(directed_edge_id);
+					let nn=n.dot(n);
+					let mut face_normals=Vec::with_capacity(the_len);
+					face_normals.clone_from(&v1f_n);
+					for face_n in &v0f_n{
+						face_normals.push(*face_n-n*(face_n.dot(n)*2/nn));
+					}
+					if is_empty_volume(face_normals){
+						edges.push(MinkowskiDirectedEdge::VertEdge(v0,directed_edge_id));
+					}
+				}
+				Cow::Owned(edges)
+			},
+		}
+	}
+	fn vert_faces(&self,_vert_id:MinkowskiVert)->Cow<Vec<MinkowskiFace>>{
+		unimplemented!()
+	}
+}
+
+fn is_empty_volume(normals:Vec<Planar64Vec3>)->bool{
+	let len=normals.len();
+	for i in 0..len-1{
+		for j in i+1..len{
+			let n=normals[i].cross(normals[j]);
+			let mut d_comp=None;
+			for k in 0..len{
+				if k!=i&&k!=j{
+					let d=n.dot(normals[k]);
+					if let Some(comp)=&d_comp{
+						if *comp*d<Planar64::ZERO{
+							return true;
+						}
+					}else{
+						d_comp=Some(d);
+					}
+				}
+			}
+		}
+	}
+	return false;
+}
+
+#[test]
+fn test_is_empty_volume(){
+	assert!(!is_empty_volume([Planar64Vec3::X,Planar64Vec3::Y,Planar64Vec3::Z].to_vec()));
+	assert!(is_empty_volume([Planar64Vec3::X,Planar64Vec3::Y,Planar64Vec3::Z,Planar64Vec3::NEG_X].to_vec()));
+}
+
+#[test]
+fn build_me_a_cube(){
+	let mesh=PhysicsMesh::unit_cube();
+	//println!("mesh={:?}",mesh);
+}

src/physics_worker.rs

@@ -0,0 +1,238 @@
+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_render_stuff(&self,time:Time)->(crate::physics::PhysicsOutputState,Time,glam::IVec2){
+		(self.physics.output(),self.timer.time(time),self.physics.get_next_mouse().pos)
+	}
+	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 (physics_output,time,mouse_pos)=interpolator.get_render_stuff(ins.time);
+				graphics_worker.send(crate::graphics_worker::Instruction::Render(physics_output,time,mouse_pos)).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/primitives.rs

@@ -1,506 +0,0 @@
-use crate::model::{IndexedModel, IndexedPolygon, IndexedGroup, IndexedVertex};
-
-#[derive(Debug)]
-pub enum Primitives{
-	Sphere,
-	Cube,
-	Cylinder,
-	Wedge,
-	CornerWedge,
-}
-#[derive(Hash,PartialEq,Eq)]
-pub enum CubeFace{
-	Right,
-	Top,
-	Back,
-	Left,
-	Bottom,
-	Front,
-}
-const CUBE_DEFAULT_TEXTURE_COORDS:[[f32;2];4]=[[0.0,0.0],[1.0,0.0],[1.0,1.0],[0.0,1.0]];
-const CUBE_DEFAULT_VERTICES:[[f32;3];8]=[
-	[-1.,-1., 1.],//0 left bottom back
-	[ 1.,-1., 1.],//1 right bottom back
-	[ 1., 1., 1.],//2 right top back
-	[-1., 1., 1.],//3 left top back
-	[-1., 1.,-1.],//4 left top front
-	[ 1., 1.,-1.],//5 right top front
-	[ 1.,-1.,-1.],//6 right bottom front
-	[-1.,-1.,-1.],//7 left bottom front
-];
-const CUBE_DEFAULT_NORMALS:[[f32;3];6]=[
-	[ 1., 0., 0.],//CubeFace::Right
-	[ 0., 1., 0.],//CubeFace::Top
-	[ 0., 0., 1.],//CubeFace::Back
-	[-1., 0., 0.],//CubeFace::Left
-	[ 0.,-1., 0.],//CubeFace::Bottom
-	[ 0., 0.,-1.],//CubeFace::Front
-];
-const CUBE_DEFAULT_POLYS:[[[u32;3];4];6]=[
-	// right (1, 0, 0)
-	[
-		[6,2,0],//[vertex,tex,norm]
-		[5,1,0],
-		[2,0,0],
-		[1,3,0],
-	],
-	// top (0, 1, 0)
-	[
-		[5,3,1],
-		[4,2,1],
-		[3,1,1],
-		[2,0,1],
-	],
-	// back (0, 0, 1)
-	[
-		[0,3,2],
-		[1,2,2],
-		[2,1,2],
-		[3,0,2],
-	],
-	// left (-1, 0, 0)
-	[
-		[0,2,3],
-		[3,1,3],
-		[4,0,3],
-		[7,3,3],
-	],
-	// bottom (0,-1, 0)
-	[
-		[1,1,4],
-		[0,0,4],
-		[7,3,4],
-		[6,2,4],
-	],
-	// front (0, 0,-1)
-	[
-		[4,1,5],
-		[5,0,5],
-		[6,3,5],
-		[7,2,5],
-	],
-];
-
-#[derive(Hash,PartialEq,Eq)]
-pub enum WedgeFace{
-	Right,
-	TopFront,
-	Back,
-	Left,
-	Bottom,
-}
-const WEDGE_DEFAULT_NORMALS:[[f32;3];5]=[
-	[ 1., 0., 0.],//Wedge::Right
-	[ 0., 1.,-1.],//Wedge::TopFront
-	[ 0., 0., 1.],//Wedge::Back
-	[-1., 0., 0.],//Wedge::Left
-	[ 0.,-1., 0.],//Wedge::Bottom
-];
-/*
-local cornerWedgeVerticies = {
-	Vector3.new(-1/2,-1/2,-1/2),7
-	Vector3.new(-1/2,-1/2, 1/2),0
-	Vector3.new( 1/2,-1/2,-1/2),6
-	Vector3.new( 1/2,-1/2, 1/2),1
-	Vector3.new( 1/2, 1/2,-1/2),5
-}
-*/
-#[derive(Hash,PartialEq,Eq)]
-pub enum CornerWedgeFace{
-	Right,
-	TopBack,
-	TopLeft,
-	Bottom,
-	Front,
-}
-const CORNERWEDGE_DEFAULT_NORMALS:[[f32;3];5]=[
-	[ 1., 0., 0.],//CornerWedge::Right
-	[ 0., 1., 1.],//CornerWedge::BackTop
-	[-1., 1., 0.],//CornerWedge::LeftTop
-	[ 0.,-1., 0.],//CornerWedge::Bottom
-	[ 0., 0.,-1.],//CornerWedge::Front
-];
-//HashMap fits this use case perfectly but feels like using a sledgehammer to drive a nail
-pub fn unit_sphere()->crate::model::IndexedModel{
-	let mut indexed_model=crate::model::generate_indexed_model_list_from_obj(obj::ObjData::load_buf(&include_bytes!("../models/suzanne.obj")[..]).unwrap(),*glam::Vec4::ONE.as_ref()).remove(0);
-	for pos in indexed_model.unique_pos.iter_mut(){
-		pos[0]=pos[0]*0.5;
-		pos[1]=pos[1]*0.5;
-		pos[2]=pos[2]*0.5;
-	}
-	indexed_model
-}
-pub type CubeFaceDescription=std::collections::HashMap::<CubeFace,FaceDescription>;
-pub fn unit_cube()->crate::model::IndexedModel{
-	let mut t=CubeFaceDescription::new();
-	t.insert(CubeFace::Right,FaceDescription::default());
-	t.insert(CubeFace::Top,FaceDescription::default());
-	t.insert(CubeFace::Back,FaceDescription::default());
-	t.insert(CubeFace::Left,FaceDescription::default());
-	t.insert(CubeFace::Bottom,FaceDescription::default());
-	t.insert(CubeFace::Front,FaceDescription::default());
-	generate_partial_unit_cube(t)
-}
-const TEAPOT_TRANSFORM:glam::Mat3=glam::mat3(glam::vec3(0.0,0.1,0.0),glam::vec3(-0.1,0.0,0.0),glam::vec3(0.0,0.0,0.1));
-pub fn unit_cylinder()->crate::model::IndexedModel{
-	let mut indexed_model=crate::model::generate_indexed_model_list_from_obj(obj::ObjData::load_buf(&include_bytes!("../models/teapot.obj")[..]).unwrap(),*glam::Vec4::ONE.as_ref()).remove(0);
-	for pos in indexed_model.unique_pos.iter_mut(){
-		[pos[0],pos[1],pos[2]]=*(TEAPOT_TRANSFORM*glam::Vec3::from_array(*pos)).as_ref();
-	}
-	indexed_model
-}
-pub type WedgeFaceDescription=std::collections::HashMap::<WedgeFace,FaceDescription>;
-pub fn unit_wedge()->crate::model::IndexedModel{
-	let mut t=WedgeFaceDescription::new();
-	t.insert(WedgeFace::Right,FaceDescription::default());
-	t.insert(WedgeFace::TopFront,FaceDescription::default());
-	t.insert(WedgeFace::Back,FaceDescription::default());
-	t.insert(WedgeFace::Left,FaceDescription::default());
-	t.insert(WedgeFace::Bottom,FaceDescription::default());
-	generate_partial_unit_wedge(t)
-}
-pub type CornerWedgeFaceDescription=std::collections::HashMap::<CornerWedgeFace,FaceDescription>;
-pub fn unit_cornerwedge()->crate::model::IndexedModel{
-	let mut t=CornerWedgeFaceDescription::new();
-	t.insert(CornerWedgeFace::Right,FaceDescription::default());
-	t.insert(CornerWedgeFace::TopBack,FaceDescription::default());
-	t.insert(CornerWedgeFace::TopLeft,FaceDescription::default());
-	t.insert(CornerWedgeFace::Bottom,FaceDescription::default());
-	t.insert(CornerWedgeFace::Front,FaceDescription::default());
-	generate_partial_unit_cornerwedge(t)
-}
-
-#[derive(Copy,Clone)]
-pub struct FaceDescription{
-	pub texture:Option<u32>,
-	pub transform:glam::Affine2,
-	pub color:glam::Vec4,
-}
-impl std::default::Default for FaceDescription{
-	fn default()->Self {
-		Self{
-			texture:None,
-			transform:glam::Affine2::IDENTITY,
-			color:glam::vec4(1.0,1.0,1.0,0.0),//zero alpha to hide the default texture
-		}
-	}
-}
-impl FaceDescription{
-	pub fn new(texture:u32,transform:glam::Affine2,color:glam::Vec4)->Self{
-		Self{texture:Some(texture),transform,color}
-	}
-	pub fn from_texture(texture:u32)->Self{
-		Self{
-			texture:Some(texture),
-			transform:glam::Affine2::IDENTITY,
-			color:glam::Vec4::ONE,
-		}
-	}
-}
-//TODO: it's probably better to use a shared vertex buffer between all primitives and use indexed rendering instead of generating a unique vertex buffer for each primitive.
-//implementation: put all roblox primitives into one model.groups <- this won't work but I forget why
-pub fn generate_partial_unit_cube(face_descriptions:CubeFaceDescription)->crate::model::IndexedModel{
-	let mut generated_pos=Vec::<[f32;3]>::new();
-	let mut generated_tex=Vec::new();
-	let mut generated_normal=Vec::new();
-	let mut generated_color=Vec::new();
-	let mut generated_vertices=Vec::new();
-	let mut groups=Vec::new();
-	let mut transforms=Vec::new();
-	//note that on a cube every vertex is guaranteed to be unique, so there's no need to hash them against existing vertices.
-	for (face,face_description) in face_descriptions.into_iter(){
-		//assume that scanning short lists is faster than hashing.
-		let transform_index=if let Some(transform_index)=transforms.iter().position(|&transform|transform==face_description.transform){
-			transform_index
-		}else{
-			//create new transform_index
-			let transform_index=transforms.len();
-			transforms.push(face_description.transform);
-			for tex in CUBE_DEFAULT_TEXTURE_COORDS{
-				generated_tex.push(*face_description.transform.transform_point2(glam::Vec2::from_array(tex)).as_ref());
-			}
-			transform_index
-		} as u32;
-		let color_index=if let Some(color_index)=generated_color.iter().position(|color|color==face_description.color.as_ref()){
-			color_index
-		}else{
-			//create new color_index
-			let color_index=generated_color.len();
-			generated_color.push(*face_description.color.as_ref());
-			color_index
-		} as u32;
-		let face_id=match face{
-			CubeFace::Right => 0,
-			CubeFace::Top => 1,
-			CubeFace::Back => 2,
-			CubeFace::Left => 3,
-			CubeFace::Bottom => 4,
-			CubeFace::Front => 5,
-		};
-		//always push normal
-		let normal_index=generated_normal.len() as u32;
-		generated_normal.push(CUBE_DEFAULT_NORMALS[face_id]);
-		//push vertices as they are needed
-		groups.push(IndexedGroup{
-			texture:face_description.texture,
-			polys:vec![IndexedPolygon{
-				vertices:CUBE_DEFAULT_POLYS[face_id].map(|tup|{
-					let pos=CUBE_DEFAULT_VERTICES[tup[0] as usize];
-					let pos_index=if let Some(pos_index)=generated_pos.iter().position(|&p|p==pos){
-						pos_index
-					}else{
-						//create new pos_index
-						let pos_index=generated_pos.len();
-						generated_pos.push(pos);
-						pos_index
-					} as u32;
-					//always push vertex
-					let vertex=IndexedVertex{
-						pos:pos_index,
-						tex:tup[1]+4*transform_index,
-						normal:normal_index,
-						color:color_index,
-					};
-					let vert_index=generated_vertices.len();
-					generated_vertices.push(vertex);
-					vert_index as u32
-				}).to_vec(),
-			}],
-		});
-	}
-	IndexedModel{
-		unique_pos:generated_pos,
-		unique_tex:generated_tex,
-		unique_normal:generated_normal,
-		unique_color:generated_color,
-		unique_vertices:generated_vertices,
-		groups,
-		instances:Vec::new(),
-	}
-}
-//don't think too hard about the copy paste because this is all going into the map tool eventually...
-pub fn generate_partial_unit_wedge(face_descriptions:WedgeFaceDescription)->crate::model::IndexedModel{
-	let wedge_default_polys=vec![
-		// right (1, 0, 0)
-		vec![
-			[6,2,0],//[vertex,tex,norm]
-			[2,0,0],
-			[1,3,0],
-		],
-		// FrontTop (0, 1, -1)
-		vec![
-			[3,1,1],
-			[2,0,1],
-			[6,3,1],
-			[7,2,1],
-		],
-		// back (0, 0, 1)
-		vec![
-			[0,3,2],
-			[1,2,2],
-			[2,1,2],
-			[3,0,2],
-		],
-		// left (-1, 0, 0)
-		vec![
-			[0,2,3],
-			[3,1,3],
-			[7,3,3],
-		],
-		// bottom (0,-1, 0)
-		vec![
-			[1,1,4],
-			[0,0,4],
-			[7,3,4],
-			[6,2,4],
-		],
-	];
-	let mut generated_pos=Vec::<[f32;3]>::new();
-	let mut generated_tex=Vec::new();
-	let mut generated_normal=Vec::new();
-	let mut generated_color=Vec::new();
-	let mut generated_vertices=Vec::new();
-	let mut groups=Vec::new();
-	let mut transforms=Vec::new();
-	//note that on a cube every vertex is guaranteed to be unique, so there's no need to hash them against existing vertices.
-	for (face,face_description) in face_descriptions.into_iter(){
-		//assume that scanning short lists is faster than hashing.
-		let transform_index=if let Some(transform_index)=transforms.iter().position(|&transform|transform==face_description.transform){
-			transform_index
-		}else{
-			//create new transform_index
-			let transform_index=transforms.len();
-			transforms.push(face_description.transform);
-			for tex in CUBE_DEFAULT_TEXTURE_COORDS{
-				generated_tex.push(*face_description.transform.transform_point2(glam::Vec2::from_array(tex)).as_ref());
-			}
-			transform_index
-		} as u32;
-		let color_index=if let Some(color_index)=generated_color.iter().position(|color|color==face_description.color.as_ref()){
-			color_index
-		}else{
-			//create new color_index
-			let color_index=generated_color.len();
-			generated_color.push(*face_description.color.as_ref());
-			color_index
-		} as u32;
-		let face_id=match face{
-			WedgeFace::Right => 0,
-			WedgeFace::TopFront => 1,
-			WedgeFace::Back => 2,
-			WedgeFace::Left => 3,
-			WedgeFace::Bottom => 4,
-		};
-		//always push normal
-		let normal_index=generated_normal.len() as u32;
-		generated_normal.push(WEDGE_DEFAULT_NORMALS[face_id]);
-		//push vertices as they are needed
-		groups.push(IndexedGroup{
-			texture:face_description.texture,
-			polys:vec![IndexedPolygon{
-				vertices:wedge_default_polys[face_id].iter().map(|tup|{
-					let pos=CUBE_DEFAULT_VERTICES[tup[0] as usize];
-					let pos_index=if let Some(pos_index)=generated_pos.iter().position(|&p|p==pos){
-						pos_index
-					}else{
-						//create new pos_index
-						let pos_index=generated_pos.len();
-						generated_pos.push(pos);
-						pos_index
-					} as u32;
-					//always push vertex
-					let vertex=IndexedVertex{
-						pos:pos_index,
-						tex:tup[1]+4*transform_index,
-						normal:normal_index,
-						color:color_index,
-					};
-					let vert_index=generated_vertices.len();
-					generated_vertices.push(vertex);
-					vert_index as u32
-				}).collect(),
-			}],
-		});
-	}
-	IndexedModel{
-		unique_pos:generated_pos,
-		unique_tex:generated_tex,
-		unique_normal:generated_normal,
-		unique_color:generated_color,
-		unique_vertices:generated_vertices,
-		groups,
-		instances:Vec::new(),
-	}
-}
-
-pub fn generate_partial_unit_cornerwedge(face_descriptions:CornerWedgeFaceDescription)->crate::model::IndexedModel{
-	let cornerwedge_default_polys=vec![
-		// right (1, 0, 0)
-		vec![
-			[6,2,0],//[vertex,tex,norm]
-			[5,1,0],
-			[1,3,0],
-		],
-		// BackTop (0, 1, 1)
-		vec![
-			[5,3,1],
-			[0,1,1],
-			[1,0,1],
-		],
-		// LeftTop (-1, 1, 0)
-		vec![
-			[5,3,2],
-			[7,2,2],
-			[0,1,2],
-		],
-		// bottom (0,-1, 0)
-		vec![
-			[1,1,3],
-			[0,0,3],
-			[7,3,3],
-			[6,2,3],
-		],
-		// front (0, 0,-1)
-		vec![
-			[5,0,4],
-			[6,3,4],
-			[7,2,4],
-		],
-	];
-	let mut generated_pos=Vec::<[f32;3]>::new();
-	let mut generated_tex=Vec::new();
-	let mut generated_normal=Vec::new();
-	let mut generated_color=Vec::new();
-	let mut generated_vertices=Vec::new();
-	let mut groups=Vec::new();
-	let mut transforms=Vec::new();
-	//note that on a cube every vertex is guaranteed to be unique, so there's no need to hash them against existing vertices.
-	for (face,face_description) in face_descriptions.into_iter(){
-		//assume that scanning short lists is faster than hashing.
-		let transform_index=if let Some(transform_index)=transforms.iter().position(|&transform|transform==face_description.transform){
-			transform_index
-		}else{
-			//create new transform_index
-			let transform_index=transforms.len();
-			transforms.push(face_description.transform);
-			for tex in CUBE_DEFAULT_TEXTURE_COORDS{
-				generated_tex.push(*face_description.transform.transform_point2(glam::Vec2::from_array(tex)).as_ref());
-			}
-			transform_index
-		} as u32;
-		let color_index=if let Some(color_index)=generated_color.iter().position(|color|color==face_description.color.as_ref()){
-			color_index
-		}else{
-			//create new color_index
-			let color_index=generated_color.len();
-			generated_color.push(*face_description.color.as_ref());
-			color_index
-		} as u32;
-		let face_id=match face{
-			CornerWedgeFace::Right => 0,
-			CornerWedgeFace::TopBack => 1,
-			CornerWedgeFace::TopLeft => 2,
-			CornerWedgeFace::Bottom => 3,
-			CornerWedgeFace::Front => 4,
-		};
-		//always push normal
-		let normal_index=generated_normal.len() as u32;
-		generated_normal.push(CORNERWEDGE_DEFAULT_NORMALS[face_id]);
-		//push vertices as they are needed
-		groups.push(IndexedGroup{
-			texture:face_description.texture,
-			polys:vec![IndexedPolygon{
-				vertices:cornerwedge_default_polys[face_id].iter().map(|tup|{
-					let pos=CUBE_DEFAULT_VERTICES[tup[0] as usize];
-					let pos_index=if let Some(pos_index)=generated_pos.iter().position(|&p|p==pos){
-						pos_index
-					}else{
-						//create new pos_index
-						let pos_index=generated_pos.len();
-						generated_pos.push(pos);
-						pos_index
-					} as u32;
-					//always push vertex
-					let vertex=IndexedVertex{
-						pos:pos_index,
-						tex:tup[1]+4*transform_index,
-						normal:normal_index,
-						color:color_index,
-					};
-					let vert_index=generated_vertices.len();
-					generated_vertices.push(vertex);
-					vert_index as u32
-				}).collect(),
-			}],
-		});
-	}
-	IndexedModel{
-		unique_pos:generated_pos,
-		unique_tex:generated_tex,
-		unique_normal:generated_normal,
-		unique_color:generated_color,
-		unique_vertices:generated_vertices,
-		groups,
-		instances:Vec::new(),
-	}
-}

src/settings.rs

@@ -1,33 +1,41 @@
+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},
-	DeriveX{x:DerivedFov,y:f64},
-	DeriveY{x:f64,y:DerivedFov},
+	SpecifyXDeriveY{x:f64,y:DerivedFov},
+	SpecifyYDeriveX{x:DerivedFov,y:f64},
 }
 impl Default for Fov{
-	fn default() -> Self {
-		Fov::DeriveX{x:DerivedFov::FromScreenAspect,y:1.0}
+	fn default()->Self{
+		Fov::SpecifyYDeriveX{x:DerivedFov::FromScreenAspect,y:1.0}
 	}
 }
-
+#[derive(Clone)]
+enum DerivedSensitivity{
+	FromRatio(Ratio64),
+}
+#[derive(Clone)]
 enum Sensitivity{
-	Exactly{x:f64,y:f64},
-	DeriveX{x:Ratio,y:f64},
-	DeriveY{x:f64,y:Ratio},
+	Exactly{x:Ratio64,y:Ratio64},
+	SpecifyXDeriveY{x:Ratio64,y:DerivedSensitivity},
+	SpecifyYDeriveX{x:DerivedSensitivity,y:Ratio64},
 }
 impl Default for Sensitivity{
-	fn default() -> Self {
-		Sensitivity::DeriveY{x:0.001,y:Ratio{ratio:1.0}}
+	fn default()->Self{
+		Sensitivity::SpecifyXDeriveY{x:Ratio64::ONE*524288,y:DerivedSensitivity::FromRatio(Ratio64::ONE)}
 	}
 }
 
-#[derive(Default)]
+#[derive(Default,Clone)]
 pub struct UserSettings{
 	fov:Fov,
 	sensitivity:Sensitivity,
@@ -36,21 +44,25 @@ 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::DeriveX{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),
-			},
-			Fov::DeriveY{x,y}=>match 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)->glam::DVec2{
+	pub fn calculate_sensitivity(&self)->Ratio64Vec2{
 		match &self.sensitivity{
-			&Sensitivity::Exactly{x,y}=>glam::dvec2(x,y),
-			Sensitivity::DeriveX{x,y}=>glam::dvec2(y*x.ratio,*y),
-			Sensitivity::DeriveY{x,y}=>glam::dvec2(*x,x*y.ratio),
+			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()),
+			}
 		}
 	}
 }
@@ -71,7 +83,7 @@ pub fn read_user_settings()->UserSettings{
 				x:fov_x,
 				y:fov_y
 			},
-			(Ok(Some(fov_x)),Ok(None))=>Fov::DeriveY{
+			(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})
@@ -79,7 +91,7 @@ pub fn read_user_settings()->UserSettings{
 					DerivedFov::FromScreenAspect
 				}
 			},
-			(Ok(None),Ok(Some(fov_y)))=>Fov::DeriveX{
+			(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{
@@ -94,20 +106,24 @@ pub fn read_user_settings()->UserSettings{
 		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:sensitivity_x,
-				y:sensitivity_y
+				x:Ratio64::try_from(sensitivity_x).unwrap(),
+				y:Ratio64::try_from(sensitivity_y).unwrap(),
 			},
-			(Ok(Some(sensitivity_x)),Ok(None))=>Sensitivity::DeriveY{
-				x:sensitivity_x,
-				y:Ratio{
-					ratio:if let Ok(Some(sensitivity_y_from_x_ratio))=cfg.getfloat("camera","sensitivity_y_from_x_ratio"){sensitivity_y_from_x_ratio}else{1.0}
-				}
-			},
-			(Ok(None),Ok(Some(sensitivity_y)))=>Sensitivity::DeriveX{
-				x:Ratio{
-					ratio:if let Ok(Some(sensitivity_x_from_y_ratio))=cfg.getfloat("camera","sensitivity_x_from_y_ratio"){sensitivity_x_from_y_ratio}else{1.0}
+			(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)
 				},
-				y:sensitivity_y,
+			},
+			(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()

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

@@ -48,7 +48,7 @@ struct ModelInstance{
 //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=4096;
+const MAX_MODEL_INSTANCES=512;
 @group(2)
 @binding(0)
 var<uniform> model_instances: array<ModelInstance, MAX_MODEL_INSTANCES>;

src/sweep.rs

@@ -1,8 +0,0 @@
-
-//something that implements body + hitbox + transform can predict collision
-impl crate::sweep::PredictCollision for Model {
-	fn predict_collision(&self,other:&Model) -> Option<crate::event::EventStruct> {
-		//math!
-		None
-	}
-}

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

@@ -2,16 +2,68 @@ 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.
 
-pub struct Worker<Task:Send,Value:Clone> {
+/*
+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> Worker<Task,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 {
@@ -45,63 +97,120 @@ impl<Task:Send+'static,Value:Clone+Send+'static> Worker<Task,Value> {
 	}
 }
 
-pub struct CompatWorker<Task,Value:Clone,F>{
-	data:std::marker::PhantomData<Task>,
-	f:F,
-	value:Value,
+/*
+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<Task,Value:Clone,F:FnMut(Task)->Value> CompatWorker<Task,Value,F> {
-	pub fn new(value:Value,f:F) -> Self {
-		Self {
-			f,
-			value,
-			data:std::marker::PhantomData,
+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(&mut self,task:Task)->Result<(),()>{
-		self.value=(self.f)(task);
-		Ok(())
-	}
-
-	pub fn grab_clone(&self)->Value{
-		self.value.clone()
+	pub fn send(&self,task:Task)->Result<(),mpsc::SendError<Task>>{
+		self.sender.send(task)
 	}
 }
 
-#[test]//How to run this test with printing: cargo test --release -- --nocapture
-fn test_worker() {
-	println!("hiiiii");
-	// Create the worker thread
-	let worker = Worker::new(crate::physics::Body::with_pva(glam::Vec3::ZERO,glam::Vec3::ZERO,glam::Vec3::ZERO),
-		|_|crate::physics::Body::with_pva(glam::Vec3::ONE,glam::Vec3::ONE,glam::Vec3::ONE)
-	);
+/*
+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,()>,
+}
 
-	// Send tasks to the worker
-	for _ in 0..5 {
-		let task = crate::instruction::TimedInstruction{
-			time:0,
-			instruction:crate::physics::PhysicsInstruction::StrafeTick,
+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::Planar64Vec3::ONE,integer::Planar64Vec3::ONE,integer::Planar64Vec3::ONE,integer::Time::ZERO);
+		let worker=QRWorker::new(physics::Body::default(),
+			|_|physics::Body::new(integer::Planar64Vec3::ONE,integer::Planar64Vec3::ONE,integer::Planar64Vec3::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));
 	}
-
-	// 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_secs(2));
-
-	// Send a new task
-	let task = crate::instruction::TimedInstruction{
-		time:0,
-		instruction:crate::physics::PhysicsInstruction::StrafeTick,
-	};
-	worker.send(task).unwrap();
-
-	println!("value={:?}",worker.grab_clone());
-
-	// wait long enough to see print from final task
-	thread::sleep(std::time::Duration::from_secs(1));
 }

src/zeroes.rs

@@ -1,28 +0,0 @@
-//find roots of polynomials
-#[inline]
-pub fn zeroes2(a0:f32,a1:f32,a2:f32) -> Vec<f32>{
-	if a2==0f32{
-		return zeroes1(a0, a1);
-	}
-	let mut radicand=a1*a1-4f32*a2*a0;
-	if 0f32<radicand {
-		radicand=radicand.sqrt();
-		if 0f32<a2 {
-			return vec![(-a1-radicand)/(2f32*a2),(-a1+radicand)/(2f32*a2)];
-		} else {
-			return vec![(-a1+radicand)/(2f32*a2),(-a1-radicand)/(2f32*a2)];
-		}
-	} else if radicand==0f32 {
-		return vec![-a1/(2f32*a2)];
-	} else {
-		return vec![];
-	}
-}
-#[inline]
-pub fn zeroes1(a0:f32,a1:f32) -> Vec<f32> {
-	if a1==0f32{
-		return vec![];
-	} else {
-		return vec![-a0/a1];
-	}
-}

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