WIP: update gfx-hal

7 files changed, 352 insertions, 290 deletions

diff --git a/stockton-render/Cargo.toml b/stockton-render/Cargo.toml
index aafd43b..eb717c4 100644
--- a/stockton-render/Cargo.toml
+++ b/stockton-render/Cargo.toml
@@ -6,38 +6,16 @@ authors = ["Oscar <oscar.shrimpton.personal@gmail.com>"]
 [dependencies]
 stockton-types = { path = "../stockton-types" }
 winit = "0.19.1"
-gfx-hal = "^0.3"
+gfx-hal = "^0.5"
 arrayvec = "0.4.10"
 nalgebra-glm = "0.4.0"
 shaderc = "0.6.1"
 log = "0.4.0"
 
 [features]
-default = ["gl"]
-metal = ["gfx-backend-metal"]
-gl = ["gfx-backend-gl"]
-dx11 = ["gfx-backend-dx11"]
-dx12 = ["gfx-backend-dx12"]
+default = ["vulkan"]
 vulkan = ["gfx-backend-vulkan"]
 
-[dependencies.gfx-backend-gl]
-version = "0.3"
-features = ["glutin"]
-optional = true
-
 [dependencies.gfx-backend-vulkan]
-version = "0.3"
-features = ["winit"]
-optional = true
-
-[target.'cfg(any(target_os = "macos", all(target_os = "ios", target_arch = "aarch64")))'.dependencies.gfx-backend-metal]
-version = "0.3"
-optional = true
-
-[target.'cfg(windows)'.dependencies.gfx-backend-dx11]
-version = "0.3"
-optional = true
-
-[target.'cfg(windows)'.dependencies.gfx-backend-dx12]
-version = "0.3"
+version = "^0.5"
 optional = true
diff --git a/stockton-render/src/draw/buffer.rs b/stockton-render/src/draw/buffer.rs
new file mode 100644
index 0000000..d29f857
--- /dev/null
+++ b/stockton-render/src/draw/buffer.rs
@@ -0,0 +1,295 @@
+// Copyright (C) 2019 Oscar Shrimpton  
+
+// This program is free software: you can redistribute it and/or modify it
+// under the terms of the GNU General Public License as published by the Free
+// Software Foundation, either version 3 of the License, or (at your option)
+// any later version.
+
+// This program is distributed in the hope that it will be useful, but WITHOUT
+// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+// FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+// more details.
+
+// You should have received a copy of the GNU General Public License along
+// with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+use std::marker::PhantomData;
+use std::ops::{Index, IndexMut, Range};
+use std::convert::TryInto;
+use core::mem::{ManuallyDrop, size_of};
+use hal::memory::{Properties, Requirements, Segment};
+use hal::buffer::Usage;
+use hal::adapter::{Adapter, MemoryType, PhysicalDevice};
+use hal::device::Device;
+use hal::{MemoryTypeId, VertexCount, InstanceCount};
+use hal::Backend;
+use crate::error::CreationError;
+use super::RenderingContext;
+
+// TODO: Proper sizing of buffers
+const BUF_SIZE: u64 = 32;
+
+fn create_buffer(device: &mut <back::Backend as hal::Backend>::Device,
+	adapter: &Adapter<back::Backend>,
+	usage: Usage,
+	properties: Properties) -> Result<(<back::Backend as hal::Backend>::Buffer, <back::Backend as hal::Backend>::Memory), CreationError> {
+	let mut buffer = unsafe { device
+	        .create_buffer(BUF_SIZE, usage) }
+	    .map_err(|e| CreationError::BufferError (e))?;
+
+	let requirements = unsafe { device.get_buffer_requirements(&buffer) };
+	let memory_type_id = adapter.physical_device
+	    .memory_properties().memory_types
+	    .iter().enumerate()
+	    .find(|&(id, memory_type)| {
+	    	requirements.type_mask & (1 << id) != 0 && memory_type.properties.contains(properties)
+	    })
+	    .map(|(id, _)| MemoryTypeId(id))
+	    .ok_or(CreationError::BufferNoMemory)?;
+
+	let memory = unsafe {device
+		.allocate_memory(memory_type_id, requirements.size) }
+		.map_err(|_| CreationError::OutOfMemoryError)?;
+
+	unsafe { device
+		.bind_buffer_memory(&memory, 0, &mut buffer) }
+		.map_err(|_| CreationError::BufferNoMemory)?;
+
+	Ok((buffer, memory
+))
+}
+
+trait ModifiableBuffer: IndexMut<usize> {
+	fn commit<'a>(&'a self) -> &'a <back::Backend as hal::Backend>::Buffer;
+}
+
+pub struct StagedBuffer<'a> {
+	stagedBuffer: ManuallyDrop<<back::Backend as hal::Backend>::Buffer>,
+	stagedMemory: ManuallyDrop<<back::Backend as hal::Backend>::Memory>,
+	buffer: ManuallyDrop<<back::Backend as hal::Backend>::Buffer>,
+	memory: ManuallyDrop<<back::Backend as hal::Backend>::Memory>,
+	mappedStaged: &'a mut [f32],
+	stagedIsDirty: bool
+}
+
+
+impl<'a> ModifiableBuffer for StagedBuffer<'a> {
+	fn new(device: &mut <back::Backend as hal::Backend>::Device, adapter: &Adapter<back::Backend>, usage: Usage) -> Result<Self, CreationError> {
+
+		let (stagedBuffer, stagedMemory) = create_buffer(device, adapter, Usage::TRANSFER_SRC, Properties::CPU_VISIBLE)?;
+		let (buffer, memory) = create_buffer(device, adapter, Usage::TRANSFER_DST | usage, Properties::DEVICE_LOCAL)?;
+
+		// Map it somewhere and get a slice to that memory
+		let rawPtr = unsafe {
+			device.map_memory(&stagedMemory, Segment::ALL).unwrap() // TODO
+		};
+		let sliceSize: usize = (BUF_SIZE / 4).try_into().unwrap(); // size in f32s
+		let mappedStaged: &'a mut [f32] = std::slice::from_raw_parts_mut(rawPtr as *mut f32, sliceSize);
+
+		Ok(StagedBuffer {
+			stagedBuffer: ManuallyDrop::new(stagedBuffer),
+			stagedMemory: ManuallyDrop::new(stagedMemory),
+			buffer: ManuallyDrop::new(buffer),
+			memory: ManuallyDrop::new(memory),
+			mappedStaged: mappedStaged,
+			stagedIsDirty: false
+		})
+	}
+}
+
+impl<'a> Index<usize> for StagedBuffer<'a> {
+	type Output = f32;
+
+	fn index(&self, index: usize) -> &Self::Output {
+		&self.mappedStaged[index]
+	}
+}
+
+impl<'a> IndexMut<usize> for StagedBuffer<'a> {
+	fn index_mut(&mut self, index: usize) -> &mut Self::Output {
+		self.stagedIsDirty = true;
+		&mut self.mappedStaged[index]
+	}
+}
+
+// trait VertexLump {
+// 	pub fn new(device: &mut <back::Backend as hal::Backend>::Device, adapter: &Adapter<back::Backend>) -> Result<Self, CreationError> {
+// }
+
+// pub(crate) struct VertexLump<T: Into<X>, X: Pod> {
+// 	pub (crate) buffer: ManuallyDrop<<back::Backend as hal::Backend>::Buffer>,
+// 	memory: ManuallyDrop<<back::Backend as hal::Backend>::Memory>,
+// 	requirements: Requirements,
+
+// 	unit_size_bytes: u64,
+// 	unit_size_verts: u64,
+// 	batch_size: u64,
+
+// 	num_batches: usize,
+
+
+// 	/// An instance is active if it has been assigned to
+// 	pub active_instances: Range<InstanceCount>,
+// 	pub active_verts: Range<VertexCount>,
+
+// 	active: bool,
+
+// 	_t: PhantomData<T>,
+// 	_x: PhantomData<X>
+// }
+
+// const BATCH_SIZE: u64 = 3;
+
+// impl<T: Into<X>, X: Pod> VertexLump<T, X> {
+// 	pub fn new(device: &mut <back::Backend as hal::Backend>::Device, adapter: &Adapter<back::Backend>) -> Result<VertexLump<T, X>, CreationError> {
+// 		let unit_size_bytes = size_of::<X>()  as u64;
+// 		let unit_size_verts = unit_size_bytes / size_of::<f32>() as u64;
+
+// 		let mut buffer = unsafe { device
+// 		        .create_buffer(BATCH_SIZE * unit_size_bytes, Usage::VERTEX) }
+
+// 	        .map_err(|e| CreationError::BufferError (e))?;
+
+// 		let requirements = unsafe { device.get_buffer_requirements(&buffer) };
+// 		let memory_type_id = adapter.physical_device
+// 	        .memory_properties().memory_types
+// 	        .iter().enumerate()
+// 	        .find(|&(id, memory_type)| {
+// 	        	requirements.type_mask & (1 << id) != 0 && memory_type.properties.contains(Properties::CPU_VISIBLE)
+// 	        })
+// 	        .map(|(id, _)| MemoryTypeId(id))
+// 	        .ok_or(CreationError::BufferNoMemory)?;
+
+// 		let memory = unsafe {device
+// 			.allocate_memory(memory_type_id, requirements.size) }
+// 			.map_err(|_| CreationError::OutOfMemoryError)?;
+
+// 		unsafe { device
+// 			.bind_buffer_memory(&memory, 0, &mut buffer) }
+// 			.map_err(|_| CreationError::BufferNoMemory)?;
+
+// 		Ok(VertexLump {
+// 			buffer: ManuallyDrop::new(buffer),
+// 			memory: ManuallyDrop::new(memory),
+// 			requirements,
+// 			active_verts: 0..0,
+// 			active_instances: 0..0,
+// 			num_batches: 1,
+// 			unit_size_bytes,
+// 			unit_size_verts,
+// 			batch_size: BATCH_SIZE, // TODO
+// 			active: true,
+// 			_t: PhantomData,
+// 			_x: PhantomData
+// 		})
+// 	}
+
+// 	pub fn set_active_instances(&mut self, range: Range<InstanceCount>) {
+// 		let count: u64 = (range.end - range.start).into();
+// 		let size_verts: u32 = (count * self.unit_size_verts).try_into().unwrap();
+// 		self.active_verts = range.start * size_verts..range.end * size_verts;
+// 		self.active_instances = range;
+// 	}
+
+// 	pub fn add(&mut self, tri: T, ctx: &mut RenderingContext) -> Result<(), ()> { 
+
+// 		// figure out where to put it
+// 		let idx: usize = (self.active_instances.end).try_into().unwrap();
+// 		let batch_size: usize = self.batch_size.try_into().unwrap();
+// 		let max_size: usize = self.num_batches * batch_size;
+
+// 		// make sure correct size
+// 		if idx >= max_size {
+// 			self.num_batches += 1;
+
+// 			debug!("Reallocating Vertex buffer to {} batches ({} instances)", self.num_batches, self.num_batches as u64 * self.batch_size);
+// 			// get new buffer
+// 			let (new_buffer, new_requirements, new_memory) = {
+// 				let mut buffer = ManuallyDrop::new(unsafe { ctx.device
+// 				        .create_buffer(self.batch_size * self.unit_size_bytes * self.num_batches as u64, Usage::VERTEX) }
+// 			        	.map_err(|_| ())?
+// 	        	);
+// 				let requirements = unsafe { ctx.device.get_buffer_requirements(&buffer) };
+
+// 				let memory_type_id = ctx.adapter.physical_device
+// 			        .memory_properties().memory_types
+// 			        .iter().enumerate()
+// 			        .find(|&(id, memory_type)| {
+// 			        	requirements.type_mask & (1 << id) != 0 && memory_type.properties.contains(Properties::CPU_VISIBLE)
+// 			        })
+// 			        .map(|(id, _)| MemoryTypeId(id))
+// 			        .ok_or(())?;
+
+// 				let memory = ManuallyDrop::new(unsafe { ctx.device
+// 					.allocate_memory(memory_type_id, requirements.size) }
+// 					.map_err(|_| ())?);
+
+// 				unsafe { ctx.device
+// 					.bind_buffer_memory(&memory, 0, &mut buffer) }
+// 					.map_err(|_| ())?;
+				
+// 				(buffer, requirements, memory)
+// 			};
+
+// 			// copy vertices
+// 			unsafe {
+// 				let copy_range = 0..self.requirements.size;
+
+// 				trace!("Copying {:?} from old buffer to new buffer", copy_range);
+
+// 				let reader = ctx.device.acquire_mapping_reader::<u8>(&*(self.memory), copy_range.clone())
+// 					.map_err(|_| ())?;
+// 				let mut writer = ctx.device.acquire_mapping_writer::<u8>(&new_memory, copy_range.clone())
+// 					.map_err(|_| ())?;
+
+// 				let copy_range: Range<usize> = 0..self.requirements.size.try_into().unwrap();
+// 				writer[copy_range.clone()].copy_from_slice(&reader[copy_range.clone()]);
+
+// 				ctx.device.release_mapping_reader(reader);
+// 				ctx.device.release_mapping_writer(writer).map_err(|_| ())?;
+// 			};
+
+// 			// destroy old buffer
+// 			self.deactivate(ctx);
+
+// 			// use new one
+// 			self.buffer = new_buffer;
+// 			self.requirements = new_requirements;
+// 			self.memory = new_memory;
+// 			self.active = true;
+
+// 		}
+
+// 		{
+// 			// acquire writer
+// 			let mut writer = self.writer(ctx)?;
+
+// 			// write to it
+// 			writer[idx] = tri.into();
+// 		}
+
+// 		// activate new triangle
+// 		let new_range = self.active_instances.start..self.active_instances.end + 1;
+// 		self.set_active_instances(new_range);
+
+// 		Ok(())
+// 	}
+
+// 	pub(crate) fn writer<'a>(&'a mut self, ctx: &'a mut RenderingContext) -> Result<VertexWriter<'a, X>, ()> {
+// 		let mapping_writer = unsafe { ctx.device
+// 			.acquire_mapping_writer(&*(self.memory), 0..self.requirements.size)
+// 			.map_err(|_| ())? };
+		
+// 		Ok(VertexWriter {
+// 			mapping_writer: ManuallyDrop::new(mapping_writer),
+// 			ctx
+// 		})
+// 	}
+
+// 	pub(crate) fn deactivate(&mut self, ctx: &mut RenderingContext) {
+// 		unsafe { ctx.device.free_memory(ManuallyDrop::take(&mut self.memory)) };
+// 		unsafe { ctx.device.destroy_buffer(ManuallyDrop::take(&mut self.buffer)) };
+// 		self.active = false;
+// 	}
+// }
+
diff --git a/stockton-render/src/draw/context.rs b/stockton-render/src/draw/context.rs
index bc27395..53f92cf 100644
--- a/stockton-render/src/draw/context.rs
+++ b/stockton-render/src/draw/context.rs
@@ -16,7 +16,6 @@
 //! Deals with all the Vulkan/HAL details.
 use crate::error as error;
 use crate::error::{CreationError, FrameError};
-use super::VertexLump;
 
 use std::mem::{ManuallyDrop, size_of};
 use std::convert::TryInto;
@@ -28,6 +27,9 @@ use arrayvec::ArrayVec;
 use hal::*;
 use hal::device::Device;
 use hal::format::{AsFormat, Rgba8Srgb as ColorFormat, Format, ChannelType};
+use hal::pool::CommandPool;
+use hal::queue::{QueueGroup, Submission};
+use hal::window::SwapchainConfig;
 
 use hal::Instance as InstanceTrait;
 
@@ -35,6 +37,9 @@ use hal::Instance as InstanceTrait;
 use back::glutin as glutin;
 
 use stockton_types::Vector2;
+use super::buffer::StagedBuffer;
+
+type ModifiableBuffer<'a> = StagedBuffer<'a>;
 
 const ENTRY_NAME: &str = "main";
 const COLOR_RANGE: image::SubresourceRange = image::SubresourceRange {
@@ -74,7 +79,7 @@ type Instance = ();
 /// Contains all the hal related stuff.
 /// In the end, this takes some 3D points and puts it on the screen.
 // TODO: Settings for clear colour, buffer sizes, etc
-pub struct RenderingContext {
+pub struct RenderingContext<'a> {
 	pub events_loop: winit::EventsLoop,
 	surface: <back::Backend as hal::Backend>::Surface,
 
@@ -97,20 +102,20 @@ pub struct RenderingContext {
 	present_complete: Vec<<back::Backend as hal::Backend>::Fence>,
 
 	frames_in_flight: usize,
-	cmd_pools: Vec<ManuallyDrop<CommandPool<back::Backend, Graphics>>>,
-	cmd_buffers: Vec<command::CommandBuffer<back::Backend, Graphics, command::MultiShot>>,
-	queue_group: QueueGroup<back::Backend, Graphics>,
+	cmd_pools: Vec<ManuallyDrop<<back::Backend as hal::Backend>::CommandPool>>,
+	cmd_buffers: Vec<<back::Backend as hal::Backend>::CommandBuffer>,
+	queue_group: QueueGroup<back::Backend>,
 
-	map_verts: VertexLump<Tri2, [f32; 15]>,
+	vert_buffer: ModifiableBuffer<'a>,
+	index_buffer: ModifiableBuffer<'a>,
 
 	descriptor_set_layouts: <back::Backend as hal::Backend>::DescriptorSetLayout,
 	pipeline_layout: ManuallyDrop<<back::Backend as hal::Backend>::PipelineLayout>,
 	pipeline: ManuallyDrop<<back::Backend as hal::Backend>::GraphicsPipeline>,
 	pub (crate) adapter: adapter::Adapter<back::Backend>
-
 }
 
-impl RenderingContext {
+impl<'a> RenderingContext<'a> {
 	/// Create a new RenderingContext for the given window.
 	pub fn new() -> Result<Self, CreationError> {
 	    let events_loop = EventsLoop::new();
@@ -148,12 +153,30 @@ impl RenderingContext {
 	    let mut adapter = adapters.remove(0);
 
 	    // Device & Queue group
-	    let (mut device, queue_group) = adapter
-	    	.open_with::<_, Graphics>(1, |family| surface.supports_queue_family(family))
-	    	.map_err(|e| CreationError::DeviceError (e))?;
+	    let (mut device, queue_group) = {
+	    	// TODO
+	    	let family = adapter
+	            .queue_families
+	            .iter()
+	            .find(|family| {
+	                surface.supports_queue_family(family) && family.queue_type().supports_graphics()
+	            })
+	            .unwrap();
+
+	    	let mut gpu = unsafe {
+	            adapter
+	                .physical_device
+	                .open(&[(family, &[1.0])], hal::Features::empty())
+	                .unwrap()
+	        };
+
+	        (gpu.queue_groups.pop().unwrap(), gpu.device)
+	    };
 
 	    // Swapchain stuff
 	    let (format, viewport, extent, swapchain, backbuffer) = {
+            use hal::window::{PresentMode, CompositeAlphaMode};
+
 		    let (caps, formats, present_modes) = surface.compatibility(&mut adapter.physical_device);
 
 		    let format = formats.map_or(Format::Rgba8Srgb, |formats| {
@@ -164,15 +187,14 @@ impl RenderingContext {
 		    });
 
             let present_mode = {
-                use hal::window::PresentMode::*;
-                [Mailbox, Fifo, Relaxed, Immediate]
+                [PresentMode::Mailbox, PresentMode::Fifo, PresentMode::Relaxed, PresentMode::Immediate]
                     .iter()
                     .cloned()
                     .find(|pm| present_modes.contains(pm))
                     .ok_or(CreationError::BadSurface)?
             };
             let composite_alpha = {
-                [CompositeAlpha::OPAQUE, CompositeAlpha::INHERIT, CompositeAlpha::PREMULTIPLIED, CompositeAlpha::POSTMULTIPLIED]
+                [CompositeAlphaMode::OPAQUE, CompositeAlphaMode::INHERIT, CompositeAlphaMode::PREMULTIPLIED, CompositeAlphaMode::POSTMULTIPLIED]
                     .iter()
                     .cloned()
                     .find(|ca| caps.composite_alpha.contains(*ca))
@@ -243,7 +265,7 @@ impl RenderingContext {
 			};
 
 			let dependency = SubpassDependency {
-				passes: SubpassRef::External..SubpassRef::Pass(0),
+				passes: None..0,
 				stages: PipelineStage::COLOR_ATTACHMENT_OUTPUT..PipelineStage::COLOR_ATTACHMENT_OUTPUT,
 				accesses: Access::empty()
 	                ..(Access::COLOR_ATTACHMENT_READ | Access::COLOR_ATTACHMENT_WRITE)
@@ -259,8 +281,14 @@ impl RenderingContext {
 			main_pass: &renderpass
 		};
 
-	    // Vertex buffer
-	    let map_verts = VertexLump::new(&mut device, &adapter)?;
+	    // Vertex and index buffers
+	    let (vert_buffer, index_buffer) = {
+	    	use hal::buffer::Usage;
+	    	(
+				ModifiableBuffer::new(&mut device, &adapter, Usage::VERTEX | Usage::TRANSFER_DST),
+				ModifiableBuffer::new(&mut device, &adapter, Usage::TRANSFER_SRC)
+	    	)
+	    };
 
 	    // Command Pools, Buffers, imageviews, framebuffers & Sync objects
     	let frames_in_flight = backbuffer.len();
@@ -278,7 +306,7 @@ impl RenderingContext {
 			        device.create_command_pool_typed(&queue_group, pool::CommandPoolCreateFlags::empty())
 			    }.map_err(|_| CreationError::OutOfMemoryError)?));
 
-			    cmd_buffers.push((*cmd_pools[i]).acquire_command_buffer::<command::MultiShot>());
+			    cmd_buffers.push((*cmd_pools[i]).allocate_one(hal::command::Level::Primary));
 			    get_image.push(device.create_semaphore().map_err(|_| CreationError::SyncObjectError)?);
 			    render_complete.push(device.create_semaphore().map_err(|_| CreationError::SyncObjectError)?);
 			    present_complete.push(device.create_fence(true).map_err(|_| CreationError::SyncObjectError)?);
@@ -332,7 +360,8 @@ impl RenderingContext {
     		pipeline_layout: ManuallyDrop::new(pipeline_layout),
     		pipeline: ManuallyDrop::new(pipeline),
 
-    		map_verts,
+    		vert_buffer,
+    		index_buffer,
 
     		adapter
     	})
@@ -660,7 +689,7 @@ impl RenderingContext {
 	}
 }
 
-impl core::ops::Drop for RenderingContext {
+impl<'a> core::ops::Drop for RenderingContext<'a> {
 	fn drop(&mut self) {
 		// TODO: Probably missing some destroy stuff
 		self.device.wait_idle().unwrap();
@@ -687,7 +716,7 @@ impl core::ops::Drop for RenderingContext {
             use core::ptr::read;
             for cmd_pool in self.cmd_pools.drain(..) {
 	            self.device.destroy_command_pool(
-	                ManuallyDrop::into_inner(cmd_pool).into_raw(),
+	                ManuallyDrop::into_inner(cmd_pool),
 	            );
             }
             self.device
diff --git a/stockton-render/src/draw/mod.rs b/stockton-render/src/draw/mod.rs
index 0a4c17a..c506f7a 100644
--- a/stockton-render/src/draw/mod.rs
+++ b/stockton-render/src/draw/mod.rs
@@ -16,8 +16,7 @@
 //! Given 3D points and some camera information, renders to the screen.
 
 mod context;
-mod vertexlump;
+mod buffer;
 
 pub use self::context::RenderingContext;
 pub use self::context::Tri2;
-pub(crate) use self::vertexlump::VertexLump;
\ No newline at end of file
diff --git a/stockton-render/src/draw/vertexlump.rs b/stockton-render/src/draw/vertexlump.rs
deleted file mode 100644
index ad607ef..0000000
--- a/stockton-render/src/draw/vertexlump.rs
+++ /dev/null
@@ -1,239 +0,0 @@
-// Copyright (C) 2019 Oscar Shrimpton  
-
-// This program is free software: you can redistribute it and/or modify it
-// under the terms of the GNU General Public License as published by the Free
-// Software Foundation, either version 3 of the License, or (at your option)
-// any later version.
-
-// This program is distributed in the hope that it will be useful, but WITHOUT
-// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
-// FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
-// more details.
-
-// You should have received a copy of the GNU General Public License along
-// with this program.  If not, see <http://www.gnu.org/licenses/>.
-
-use std::marker::PhantomData;
-use std::ops::{Index, IndexMut, Range};
-use std::convert::TryInto;
-use core::mem::{ManuallyDrop, size_of};
-use hal::memory::{Pod, Properties, Requirements};
-use hal::buffer::Usage;
-use hal::adapter::MemoryTypeId;
-use hal::{VertexCount, InstanceCount, Adapter, Device, PhysicalDevice, mapping};
-use hal::Backend;
-use crate::error::CreationError;
-use super::RenderingContext;
-
-pub(crate) struct VertexLump<T: Into<X>, X: Pod> {
-	pub (crate) buffer: ManuallyDrop<<back::Backend as hal::Backend>::Buffer>,
-	memory: ManuallyDrop<<back::Backend as hal::Backend>::Memory>,
-	requirements: Requirements,
-
-	unit_size_bytes: u64,
-	unit_size_verts: u64,
-	batch_size: u64,
-
-	num_batches: usize,
-
-
-	/// An instance is active if it has been assigned to
-	pub active_instances: Range<InstanceCount>,
-	pub active_verts: Range<VertexCount>,
-
-	active: bool,
-
-	_t: PhantomData<T>,
-	_x: PhantomData<X>
-}
-
-const BATCH_SIZE: u64 = 3;
-
-impl<T: Into<X>, X: Pod> VertexLump<T, X> {
-	pub fn new(device: &mut <back::Backend as hal::Backend>::Device, adapter: &Adapter<back::Backend>) -> Result<VertexLump<T, X>, CreationError> {
-		let unit_size_bytes = size_of::<X>()  as u64;
-		let unit_size_verts = unit_size_bytes / size_of::<f32>() as u64;
-
-		let mut buffer = unsafe { device
-		        .create_buffer(BATCH_SIZE * unit_size_bytes, Usage::VERTEX) }
-
-	        .map_err(|e| CreationError::BufferError (e))?;
-
-		let requirements = unsafe { device.get_buffer_requirements(&buffer) };
-		let memory_type_id = adapter.physical_device
-	        .memory_properties().memory_types
-	        .iter().enumerate()
-	        .find(|&(id, memory_type)| {
-	        	requirements.type_mask & (1 << id) != 0 && memory_type.properties.contains(Properties::CPU_VISIBLE)
-	        })
-	        .map(|(id, _)| MemoryTypeId(id))
-	        .ok_or(CreationError::BufferNoMemory)?;
-
-		let memory = unsafe {device
-			.allocate_memory(memory_type_id, requirements.size) }
-			.map_err(|_| CreationError::OutOfMemoryError)?;
-
-		unsafe { device
-			.bind_buffer_memory(&memory, 0, &mut buffer) }
-			.map_err(|_| CreationError::BufferNoMemory)?;
-
-		Ok(VertexLump {
-			buffer: ManuallyDrop::new(buffer),
-			memory: ManuallyDrop::new(memory),
-			requirements,
-			active_verts: 0..0,
-			active_instances: 0..0,
-			num_batches: 1,
-			unit_size_bytes,
-			unit_size_verts,
-			batch_size: BATCH_SIZE, // TODO
-			active: true,
-			_t: PhantomData,
-			_x: PhantomData
-		})
-	}
-
-	pub fn set_active_instances(&mut self, range: Range<InstanceCount>) {
-		let count: u64 = (range.end - range.start).into();
-		let size_verts: u32 = (count * self.unit_size_verts).try_into().unwrap();
-		self.active_verts = range.start * size_verts..range.end * size_verts;
-		self.active_instances = range;
-	}
-
-	pub fn add(&mut self, tri: T, ctx: &mut RenderingContext) -> Result<(), ()> { 
-
-		// figure out where to put it
-		let idx: usize = (self.active_instances.end).try_into().unwrap();
-		let batch_size: usize = self.batch_size.try_into().unwrap();
-		let max_size: usize = self.num_batches * batch_size;
-
-		// make sure correct size
-		if idx >= max_size {
-			self.num_batches += 1;
-
-			debug!("Reallocating Vertex buffer to {} batches ({} instances)", self.num_batches, self.num_batches as u64 * self.batch_size);
-			// get new buffer
-			let (new_buffer, new_requirements, new_memory) = {
-				let mut buffer = ManuallyDrop::new(unsafe { ctx.device
-				        .create_buffer(self.batch_size * self.unit_size_bytes * self.num_batches as u64, Usage::VERTEX) }
-			        	.map_err(|_| ())?
-	        	);
-				let requirements = unsafe { ctx.device.get_buffer_requirements(&buffer) };
-
-				let memory_type_id = ctx.adapter.physical_device
-			        .memory_properties().memory_types
-			        .iter().enumerate()
-			        .find(|&(id, memory_type)| {
-			        	requirements.type_mask & (1 << id) != 0 && memory_type.properties.contains(Properties::CPU_VISIBLE)
-			        })
-			        .map(|(id, _)| MemoryTypeId(id))
-			        .ok_or(())?;
-
-				let memory = ManuallyDrop::new(unsafe { ctx.device
-					.allocate_memory(memory_type_id, requirements.size) }
-					.map_err(|_| ())?);
-
-				unsafe { ctx.device
-					.bind_buffer_memory(&memory, 0, &mut buffer) }
-					.map_err(|_| ())?;
-				
-				(buffer, requirements, memory)
-			};
-
-			// copy vertices
-			unsafe {
-				let copy_range = 0..self.requirements.size;
-
-				trace!("Copying {:?} from old buffer to new buffer", copy_range);
-
-				let reader = ctx.device.acquire_mapping_reader::<u8>(&*(self.memory), copy_range.clone())
-					.map_err(|_| ())?;
-				let mut writer = ctx.device.acquire_mapping_writer::<u8>(&new_memory, copy_range.clone())
-					.map_err(|_| ())?;
-
-				let copy_range: Range<usize> = 0..self.requirements.size.try_into().unwrap();
-				writer[copy_range.clone()].copy_from_slice(&reader[copy_range.clone()]);
-
-				ctx.device.release_mapping_reader(reader);
-				ctx.device.release_mapping_writer(writer).map_err(|_| ())?;
-			};
-
-			// destroy old buffer
-			self.deactivate(ctx);
-
-			// use new one
-			self.buffer = new_buffer;
-			self.requirements = new_requirements;
-			self.memory = new_memory;
-			self.active = true;
-
-		}
-
-		{
-			// acquire writer
-			let mut writer = self.writer(ctx)?;
-
-			// write to it
-			writer[idx] = tri.into();
-		}
-
-		// activate new triangle
-		let new_range = self.active_instances.start..self.active_instances.end + 1;
-		self.set_active_instances(new_range);
-
-		Ok(())
-	}
-
-	pub(crate) fn writer<'a>(&'a mut self, ctx: &'a mut RenderingContext) -> Result<VertexWriter<'a, X>, ()> {
-		let mapping_writer = unsafe { ctx.device
-			.acquire_mapping_writer(&*(self.memory), 0..self.requirements.size)
-			.map_err(|_| ())? };
-		
-		Ok(VertexWriter {
-			mapping_writer: ManuallyDrop::new(mapping_writer),
-			ctx
-		})
-	}
-
-	pub(crate) fn deactivate(&mut self, ctx: &mut RenderingContext) {
-		unsafe { ctx.device.free_memory(ManuallyDrop::take(&mut self.memory)) };
-		unsafe { ctx.device.destroy_buffer(ManuallyDrop::take(&mut self.buffer)) };
-		self.active = false;
-	}
-}
-
-pub struct VertexWriter<'a, X: Pod> {
-	mapping_writer: ManuallyDrop<mapping::Writer<'a, back::Backend, X>>,
-	ctx: &'a mut RenderingContext
-}
-
-impl<'a, X: Pod> Drop for VertexWriter<'a, X> {
-	fn drop(&mut self) {
-		unsafe {
-			self.ctx.device.release_mapping_writer(ManuallyDrop::take(&mut self.mapping_writer))
-		}.unwrap();
-	}
-}
-
-impl<'a, X: Pod> Index<usize> for VertexWriter<'a, X> {
-	type Output = X;
-
-	fn index(&self, index: usize) -> &Self::Output {
-		&self.mapping_writer[index]
-	}
-}
-
-impl<'a, X: Pod> IndexMut<usize> for VertexWriter<'a, X> {
-	fn index_mut(&mut self, index: usize) -> &mut Self::Output {
-		&mut self.mapping_writer[index]
-	}
-} 
-
-
-impl<T: Into<X>, X: Pod> Drop for VertexLump<T, X> {
-	fn drop(&mut self) {
-		if self.active {
-			panic!("VertexLump dropped without being deactivated");
-		}
-	}
-}
\ No newline at end of file
diff --git a/stockton-render/src/error.rs b/stockton-render/src/error.rs
index 7b2b3c0..485d1e7 100644
--- a/stockton-render/src/error.rs
+++ b/stockton-render/src/error.rs
@@ -27,7 +27,7 @@ pub enum CreationError {
 	WindowError,
 	BadSurface,
 	
-	DeviceError (hal::error::DeviceCreationError),
+	DeviceError (hal::device::CreationError),
 
 	OutOfMemoryError,
 
@@ -42,7 +42,7 @@ pub enum CreationError {
 	BufferNoMemory,
 	
 	SwapchainError (hal::window::CreationError),
-	ImageViewError (hal::image::ViewError)
+	ImageViewError (hal::image::ViewCreationError)
 }
 
 /// An error encountered when rendering.
diff --git a/stockton-render/src/lib.rs b/stockton-render/src/lib.rs
index 31cf57e..e51a582 100644
--- a/stockton-render/src/lib.rs
+++ b/stockton-render/src/lib.rs
@@ -60,7 +60,7 @@ use std::sync::{Arc, RwLock};
 
 pub struct Renderer<'a> {
 	world: Arc<RwLock<World<'a>>>,
-	pub context: RenderingContext
+	pub context: RenderingContext<'a>
 }