summaryrefslogtreecommitdiffstats
path: root/wgpu
diff options
context:
space:
mode:
authorLibravatar Malte Veerman <malte.veerman@gmail.com>2019-12-06 16:47:40 +0100
committerLibravatar Malte Veerman <malte.veerman@gmail.com>2019-12-11 20:02:43 +0100
commit80324284282f173e4d26e1f297daaf71a93f51a6 (patch)
treeafaf4816e7a9ae4f57616afc2386ff8695f3b3c4 /wgpu
parentc1b9f6652517dcbf5ffd83b5db4a624f9a5b0da4 (diff)
downloadiced-80324284282f173e4d26e1f297daaf71a93f51a6.tar.gz
iced-80324284282f173e4d26e1f297daaf71a93f51a6.tar.bz2
iced-80324284282f173e4d26e1f297daaf71a93f51a6.zip
Implemented SVG support in iced_wgpu.
Diffstat (limited to 'wgpu')
-rw-r--r--wgpu/Cargo.toml1
-rw-r--r--wgpu/src/lib.rs2
-rw-r--r--wgpu/src/primitive.rs9
-rw-r--r--wgpu/src/renderer.rs34
-rw-r--r--wgpu/src/svg.rs595
5 files changed, 640 insertions, 1 deletions
diff --git a/wgpu/Cargo.toml b/wgpu/Cargo.toml
index 47ec0537..e8e9c877 100644
--- a/wgpu/Cargo.toml
+++ b/wgpu/Cargo.toml
@@ -17,3 +17,4 @@ image = "0.22"
glam = "0.8"
font-kit = "0.4"
log = "0.4"
+resvg = { version = "0.8", features = ["raqote-backend"] }
diff --git a/wgpu/src/lib.rs b/wgpu/src/lib.rs
index 9f9ed8db..1f5794ee 100644
--- a/wgpu/src/lib.rs
+++ b/wgpu/src/lib.rs
@@ -28,11 +28,13 @@ mod image;
mod primitive;
mod quad;
mod renderer;
+mod svg;
mod text;
mod transformation;
pub(crate) use crate::image::Image;
pub(crate) use quad::Quad;
+pub(crate) use svg::Svg;
pub(crate) use transformation::Transformation;
pub use primitive::Primitive;
diff --git a/wgpu/src/primitive.rs b/wgpu/src/primitive.rs
index 04264e5d..c637626b 100644
--- a/wgpu/src/primitive.rs
+++ b/wgpu/src/primitive.rs
@@ -3,6 +3,8 @@ use iced_native::{
VerticalAlignment,
};
+use crate::svg;
+
/// A rendering primitive.
#[derive(Debug, Clone)]
pub enum Primitive {
@@ -46,6 +48,13 @@ pub enum Primitive {
/// The bounds of the image
bounds: Rectangle,
},
+ /// A svg icon primitive
+ Svg {
+ /// The handle of the icon
+ handle: svg::Handle,
+ /// The bounds of the icon
+ bounds: Rectangle,
+ },
/// A clip primitive
Clip {
/// The bounds of the clip
diff --git a/wgpu/src/renderer.rs b/wgpu/src/renderer.rs
index fa52bd96..9895e1c4 100644
--- a/wgpu/src/renderer.rs
+++ b/wgpu/src/renderer.rs
@@ -1,4 +1,4 @@
-use crate::{quad, text, Image, Primitive, Quad, Transformation};
+use crate::{quad, text, Image, Primitive, Quad, Svg, Transformation};
use iced_native::{
renderer::{Debugger, Windowed},
Background, Color, Layout, MouseCursor, Point, Rectangle, Vector, Widget,
@@ -23,6 +23,7 @@ pub struct Renderer {
queue: Queue,
quad_pipeline: quad::Pipeline,
image_pipeline: crate::image::Pipeline,
+ svg_pipeline: crate::svg::Pipeline,
text_pipeline: text::Pipeline,
}
@@ -31,6 +32,7 @@ struct Layer<'a> {
offset: Vector<u32>,
quads: Vec<Quad>,
images: Vec<Image>,
+ svgs: Vec<Svg>,
text: Vec<wgpu_glyph::Section<'a>>,
}
@@ -41,6 +43,7 @@ impl<'a> Layer<'a> {
offset,
quads: Vec::new(),
images: Vec::new(),
+ svgs: Vec::new(),
text: Vec::new(),
}
}
@@ -64,12 +67,14 @@ impl Renderer {
let text_pipeline = text::Pipeline::new(&mut device);
let quad_pipeline = quad::Pipeline::new(&mut device);
let image_pipeline = crate::image::Pipeline::new(&mut device);
+ let svg_pipeline = crate::svg::Pipeline::new(&mut device);
Self {
device,
queue,
quad_pipeline,
image_pipeline,
+ svg_pipeline,
text_pipeline,
}
}
@@ -128,6 +133,7 @@ impl Renderer {
self.queue.submit(&[encoder.finish()]);
self.image_pipeline.trim_cache();
+ self.svg_pipeline.trim_cache();
*mouse_cursor
}
@@ -237,6 +243,13 @@ impl Renderer {
scale: [bounds.width, bounds.height],
});
}
+ Primitive::Svg { handle, bounds } => {
+ layer.svgs.push(Svg {
+ handle: handle.clone(),
+ position: [bounds.x, bounds.y],
+ scale: [bounds.width, bounds.height],
+ })
+ },
Primitive::Clip {
bounds,
offset,
@@ -345,6 +358,25 @@ impl Renderer {
);
}
+ if layer.svgs.len() > 0 {
+ let translated = transformation
+ * Transformation::translate(
+ -(layer.offset.x as f32),
+ -(layer.offset.y as f32),
+ );
+
+ self.svg_pipeline.draw(
+ &mut self.device,
+ encoder,
+ &layer.svgs,
+ translated,
+ bounds,
+ target,
+ (dpi * 96.0) as u16,
+ (dpi * 20.0) as u16,
+ );
+ }
+
if layer.text.len() > 0 {
for text in layer.text.iter() {
// Target physical coordinates directly to avoid blurry text
diff --git a/wgpu/src/svg.rs b/wgpu/src/svg.rs
new file mode 100644
index 00000000..87394799
--- /dev/null
+++ b/wgpu/src/svg.rs
@@ -0,0 +1,595 @@
+use crate::Transformation;
+use iced_native::{Hasher, Rectangle};
+
+use std::{
+ cell::RefCell,
+ collections::{HashMap, HashSet},
+ hash::{Hash, Hasher as _},
+ mem,
+ path::PathBuf,
+ rc::Rc,
+};
+use std::fmt::Debug;
+
+
+#[derive(Debug)]
+pub struct Pipeline {
+ cache: RefCell<Cache>,
+
+ pipeline: wgpu::RenderPipeline,
+ uniforms: wgpu::Buffer,
+ vertices: wgpu::Buffer,
+ indices: wgpu::Buffer,
+ instances: wgpu::Buffer,
+ constants: wgpu::BindGroup,
+ texture_layout: wgpu::BindGroupLayout,
+}
+
+impl Pipeline {
+ pub fn new(device: &wgpu::Device) -> Self {
+ let sampler = device.create_sampler(&wgpu::SamplerDescriptor {
+ 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,
+ lod_min_clamp: -100.0,
+ lod_max_clamp: 100.0,
+ compare_function: wgpu::CompareFunction::Always,
+ });
+
+ let constant_layout =
+ device.create_bind_group_layout(&wgpu::BindGroupLayoutDescriptor {
+ bindings: &[
+ wgpu::BindGroupLayoutBinding {
+ binding: 0,
+ visibility: wgpu::ShaderStage::VERTEX,
+ ty: wgpu::BindingType::UniformBuffer { dynamic: false },
+ },
+ wgpu::BindGroupLayoutBinding {
+ binding: 1,
+ visibility: wgpu::ShaderStage::FRAGMENT,
+ ty: wgpu::BindingType::Sampler,
+ },
+ ],
+ });
+
+ let uniforms = Uniforms {
+ transform: Transformation::identity().into(),
+ };
+
+ let uniforms_buffer = device
+ .create_buffer_mapped(
+ 1,
+ wgpu::BufferUsage::UNIFORM | wgpu::BufferUsage::COPY_DST,
+ )
+ .fill_from_slice(&[uniforms]);
+
+ let constant_bind_group =
+ device.create_bind_group(&wgpu::BindGroupDescriptor {
+ layout: &constant_layout,
+ bindings: &[
+ wgpu::Binding {
+ binding: 0,
+ resource: wgpu::BindingResource::Buffer {
+ buffer: &uniforms_buffer,
+ range: 0..std::mem::size_of::<Uniforms>() as u64,
+ },
+ },
+ wgpu::Binding {
+ binding: 1,
+ resource: wgpu::BindingResource::Sampler(&sampler),
+ },
+ ],
+ });
+
+ let texture_layout =
+ device.create_bind_group_layout(&wgpu::BindGroupLayoutDescriptor {
+ bindings: &[wgpu::BindGroupLayoutBinding {
+ binding: 0,
+ visibility: wgpu::ShaderStage::FRAGMENT,
+ ty: wgpu::BindingType::SampledTexture {
+ multisampled: false,
+ dimension: wgpu::TextureViewDimension::D2,
+ },
+ }],
+ });
+
+ let layout =
+ device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
+ bind_group_layouts: &[&constant_layout, &texture_layout],
+ });
+
+ let vs = include_bytes!("shader/image.vert.spv");
+ let vs_module = device.create_shader_module(
+ &wgpu::read_spirv(std::io::Cursor::new(&vs[..]))
+ .expect("Read image vertex shader as SPIR-V"),
+ );
+
+ let fs = include_bytes!("shader/image.frag.spv");
+ let fs_module = device.create_shader_module(
+ &wgpu::read_spirv(std::io::Cursor::new(&fs[..]))
+ .expect("Read image fragment shader as SPIR-V"),
+ );
+
+ let pipeline =
+ device.create_render_pipeline(&wgpu::RenderPipelineDescriptor {
+ layout: &layout,
+ vertex_stage: wgpu::ProgrammableStageDescriptor {
+ module: &vs_module,
+ entry_point: "main",
+ },
+ fragment_stage: Some(wgpu::ProgrammableStageDescriptor {
+ module: &fs_module,
+ entry_point: "main",
+ }),
+ rasterization_state: Some(wgpu::RasterizationStateDescriptor {
+ front_face: wgpu::FrontFace::Cw,
+ cull_mode: wgpu::CullMode::None,
+ depth_bias: 0,
+ depth_bias_slope_scale: 0.0,
+ depth_bias_clamp: 0.0,
+ }),
+ primitive_topology: wgpu::PrimitiveTopology::TriangleList,
+ color_states: &[wgpu::ColorStateDescriptor {
+ format: wgpu::TextureFormat::Bgra8UnormSrgb,
+ color_blend: wgpu::BlendDescriptor {
+ src_factor: wgpu::BlendFactor::SrcAlpha,
+ dst_factor: wgpu::BlendFactor::OneMinusSrcAlpha,
+ operation: wgpu::BlendOperation::Add,
+ },
+ alpha_blend: wgpu::BlendDescriptor {
+ src_factor: wgpu::BlendFactor::One,
+ dst_factor: wgpu::BlendFactor::OneMinusSrcAlpha,
+ operation: wgpu::BlendOperation::Add,
+ },
+ write_mask: wgpu::ColorWrite::ALL,
+ }],
+ depth_stencil_state: None,
+ index_format: wgpu::IndexFormat::Uint16,
+ vertex_buffers: &[
+ wgpu::VertexBufferDescriptor {
+ stride: mem::size_of::<Vertex>() as u64,
+ step_mode: wgpu::InputStepMode::Vertex,
+ attributes: &[wgpu::VertexAttributeDescriptor {
+ shader_location: 0,
+ format: wgpu::VertexFormat::Float2,
+ offset: 0,
+ }],
+ },
+ wgpu::VertexBufferDescriptor {
+ stride: mem::size_of::<Instance>() as u64,
+ step_mode: wgpu::InputStepMode::Instance,
+ attributes: &[
+ wgpu::VertexAttributeDescriptor {
+ shader_location: 1,
+ format: wgpu::VertexFormat::Float2,
+ offset: 0,
+ },
+ wgpu::VertexAttributeDescriptor {
+ shader_location: 2,
+ format: wgpu::VertexFormat::Float2,
+ offset: 4 * 2,
+ },
+ ],
+ },
+ ],
+ sample_count: 1,
+ sample_mask: !0,
+ alpha_to_coverage_enabled: false,
+ });
+
+ let vertices = device
+ .create_buffer_mapped(QUAD_VERTS.len(), wgpu::BufferUsage::VERTEX)
+ .fill_from_slice(&QUAD_VERTS);
+
+ let indices = device
+ .create_buffer_mapped(QUAD_INDICES.len(), wgpu::BufferUsage::INDEX)
+ .fill_from_slice(&QUAD_INDICES);
+
+ let instances = device.create_buffer(&wgpu::BufferDescriptor {
+ size: mem::size_of::<Instance>() as u64,
+ usage: wgpu::BufferUsage::VERTEX | wgpu::BufferUsage::COPY_DST,
+ });
+
+ Pipeline {
+ cache: RefCell::new(Cache::new()),
+
+ pipeline,
+ uniforms: uniforms_buffer,
+ vertices,
+ indices,
+ instances,
+ constants: constant_bind_group,
+ texture_layout,
+ }
+ }
+
+ fn load(&self, handle: &Handle) {
+ if !self.cache.borrow().contains(&handle) {
+ if !handle.path.is_file() {
+ let mem = Memory::NotFound;
+
+ let _ = self.cache.borrow_mut().insert(&handle, mem);
+ }
+
+ let mut opt = resvg::Options::default();
+ opt.usvg.path = Some(handle.path.clone());
+ opt.usvg.dpi = handle.dpi as f64;
+ opt.usvg.font_size = handle.font_size as f64;
+
+ let mem = match resvg::usvg::Tree::from_file(&handle.path, &opt.usvg) {
+ Ok(tree) => Memory::Host { tree },
+ Err(_) => Memory::Invalid
+ };
+
+ let _ = self.cache.borrow_mut().insert(&handle, mem);
+ }
+ }
+
+ pub fn draw(
+ &mut self,
+ device: &mut wgpu::Device,
+ encoder: &mut wgpu::CommandEncoder,
+ svgs: &[Svg],
+ transformation: Transformation,
+ bounds: Rectangle<u32>,
+ target: &wgpu::TextureView,
+ dpi: u16,
+ font_size: u16,
+ ) {
+ let uniforms_buffer = device
+ .create_buffer_mapped(1, wgpu::BufferUsage::COPY_SRC)
+ .fill_from_slice(&[Uniforms {
+ transform: transformation.into(),
+ }]);
+
+ encoder.copy_buffer_to_buffer(
+ &uniforms_buffer,
+ 0,
+ &self.uniforms,
+ 0,
+ std::mem::size_of::<Uniforms>() as u64,
+ );
+
+ // TODO: Batch draw calls using a texture atlas
+ // Guillotière[1] by @nical can help us a lot here.
+ //
+ // [1]: https://github.com/nical/guillotiere
+ for svg in svgs {
+ let mut handle = svg.handle.clone();
+ handle.set_dpi(dpi);
+ handle.set_font_size(font_size);
+
+ self.load(&handle);
+
+ if let Some(texture) = self
+ .cache
+ .borrow_mut()
+ .get(&handle)
+ .unwrap()
+ .upload(device, encoder, &self.texture_layout, svg.scale[0] as u32, svg.scale[1] as u32)
+ {
+ let instance_buffer = device
+ .create_buffer_mapped(1, wgpu::BufferUsage::COPY_SRC)
+ .fill_from_slice(&[Instance {
+ _position: svg.position,
+ _scale: svg.scale,
+ }]);
+
+ encoder.copy_buffer_to_buffer(
+ &instance_buffer,
+ 0,
+ &self.instances,
+ 0,
+ mem::size_of::<Instance>() as u64,
+ );
+
+ {
+ let mut render_pass = encoder.begin_render_pass(
+ &wgpu::RenderPassDescriptor {
+ color_attachments: &[
+ wgpu::RenderPassColorAttachmentDescriptor {
+ attachment: target,
+ resolve_target: None,
+ load_op: wgpu::LoadOp::Load,
+ store_op: wgpu::StoreOp::Store,
+ clear_color: wgpu::Color::TRANSPARENT,
+ },
+ ],
+ depth_stencil_attachment: None,
+ },
+ );
+
+ render_pass.set_pipeline(&self.pipeline);
+ render_pass.set_bind_group(0, &self.constants, &[]);
+ render_pass.set_bind_group(1, &texture, &[]);
+ render_pass.set_index_buffer(&self.indices, 0);
+ render_pass.set_vertex_buffers(
+ 0,
+ &[(&self.vertices, 0), (&self.instances, 0)],
+ );
+ render_pass.set_scissor_rect(
+ bounds.x,
+ bounds.y,
+ bounds.width,
+ bounds.height,
+ );
+
+ render_pass.draw_indexed(
+ 0..QUAD_INDICES.len() as u32,
+ 0,
+ 0..1 as u32,
+ );
+ }
+ }
+ }
+ }
+
+ pub fn trim_cache(&mut self) {
+ self.cache.borrow_mut().trim();
+ }
+}
+
+#[derive(Debug, Clone, PartialEq)]
+pub struct Handle {
+ id: u64,
+ path: PathBuf,
+ dpi: u16,
+ font_size: u16,
+}
+
+impl Handle {
+ /// Returns the unique id of this [`Handle`]
+ ///
+ /// [`Handle`]: struct.Handle.html
+ pub fn id(&self) -> u64 {
+ self.id
+ }
+
+ /// Creates a svg [`Handle`] pointing to the svg icon of the given path.
+ ///
+ /// [`Handle`]: struct.Handle.html
+ pub fn from_path<T: Into<PathBuf>>(path: T) -> Handle {
+ let path = path.into();
+ let dpi = 96;
+ let font_size = 20;
+ let mut hasher = Hasher::default();
+ path.hash(&mut hasher);
+ dpi.hash(&mut hasher);
+ font_size.hash(&mut hasher);
+
+ Self {
+ id: hasher.finish(),
+ path,
+ dpi,
+ font_size,
+ }
+ }
+
+ fn set_dpi(&mut self, dpi: u16) {
+ if self.dpi == dpi {
+ return;
+ }
+
+ self.dpi = dpi;
+
+ let mut hasher = Hasher::default();
+ self.path.hash(&mut hasher);
+ self.dpi.hash(&mut hasher);
+ self.font_size.hash(&mut hasher);
+
+ self.id = hasher.finish();
+ }
+
+ fn set_font_size(&mut self, font_size: u16) {
+ if self.font_size == font_size {
+ return;
+ }
+
+ self.font_size = font_size;
+
+ let mut hasher = Hasher::default();
+ self.path.hash(&mut hasher);
+ self.dpi.hash(&mut hasher);
+ self.font_size.hash(&mut hasher);
+
+ self.id = hasher.finish();
+ }
+}
+
+impl Hash for Handle {
+ fn hash<H: std::hash::Hasher>(&self, state: &mut H) {
+ self.id.hash(state);
+ self.dpi.hash(state);
+ self.font_size.hash(state);
+ }
+}
+
+enum Memory {
+ Host {
+ tree: resvg::usvg::Tree,
+ },
+ Device {
+ bind_group: Rc<wgpu::BindGroup>,
+ },
+ NotFound,
+ Invalid,
+}
+
+impl Memory {
+ fn upload(
+ &mut self,
+ device: &wgpu::Device,
+ encoder: &mut wgpu::CommandEncoder,
+ texture_layout: &wgpu::BindGroupLayout,
+ width: u32,
+ height: u32
+ ) -> Option<Rc<wgpu::BindGroup>> {
+ match self {
+ Memory::Host { tree } => {
+ let extent = wgpu::Extent3d {
+ width,
+ height,
+ depth: 1,
+ };
+
+ let texture = device.create_texture(&wgpu::TextureDescriptor {
+ size: extent,
+ array_layer_count: 1,
+ mip_level_count: 1,
+ sample_count: 1,
+ dimension: wgpu::TextureDimension::D2,
+ format: wgpu::TextureFormat::Bgra8UnormSrgb,
+ usage: wgpu::TextureUsage::COPY_DST
+ | wgpu::TextureUsage::SAMPLED,
+ });
+
+ let mut canvas = resvg::raqote::DrawTarget::new(width as i32, height as i32);
+ let opt = resvg::Options::default();
+ let screen_size = resvg::ScreenSize::new(width, height).unwrap();
+ resvg::backend_raqote::render_to_canvas(tree, &opt, screen_size, &mut canvas);
+ let slice = canvas.get_data();
+ let temp_buf = device
+ .create_buffer_mapped(
+ slice.len(),
+ wgpu::BufferUsage::COPY_SRC,
+ )
+ .fill_from_slice(slice);
+
+ encoder.copy_buffer_to_texture(
+ wgpu::BufferCopyView {
+ buffer: &temp_buf,
+ offset: 0,
+ row_pitch: width * 4,
+ image_height: height,
+ },
+ wgpu::TextureCopyView {
+ texture: &texture,
+ array_layer: 0,
+ mip_level: 0,
+ origin: wgpu::Origin3d {
+ x: 0.0,
+ y: 0.0,
+ z: 0.0,
+ },
+ },
+ extent,
+ );
+
+ let bind_group =
+ device.create_bind_group(&wgpu::BindGroupDescriptor {
+ layout: texture_layout,
+ bindings: &[wgpu::Binding {
+ binding: 0,
+ resource: wgpu::BindingResource::TextureView(
+ &texture.create_default_view(),
+ ),
+ }],
+ });
+
+ let bind_group = Rc::new(bind_group);
+
+ *self = Memory::Device {
+ bind_group: bind_group.clone(),
+ };
+
+ Some(bind_group)
+ }
+ Memory::Device { bind_group, .. } => Some(bind_group.clone()),
+ Memory::NotFound => None,
+ Memory::Invalid => None,
+ }
+ }
+}
+
+impl Debug for Memory {
+ fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> Result<(), std::fmt::Error> {
+ match self {
+ Memory::Host { .. } => write!(f, "Memory::Host"),
+ Memory::Device { .. } => write!(f, "Memory::Device"),
+ Memory::NotFound => write!(f, "Memory::NotFound"),
+ Memory::Invalid => write!(f, "Memory::Invalid"),
+ }
+ }
+}
+
+#[derive(Debug)]
+struct Cache {
+ map: HashMap<u64, Memory>,
+ hits: HashSet<u64>,
+}
+
+impl Cache {
+ fn new() -> Self {
+ Self {
+ map: HashMap::new(),
+ hits: HashSet::new(),
+ }
+ }
+
+ fn contains(&self, handle: &Handle) -> bool {
+ self.map.contains_key(&handle.id())
+ }
+
+ fn get(&mut self, handle: &Handle) -> Option<&mut Memory> {
+ let _ = self.hits.insert(handle.id());
+
+ self.map.get_mut(&handle.id())
+ }
+
+ fn insert(&mut self, handle: &Handle, memory: Memory) {
+ let _ = self.map.insert(handle.id(), memory);
+ }
+
+ fn trim(&mut self) {
+ let hits = &self.hits;
+
+ self.map.retain(|k, _| hits.contains(k));
+ self.hits.clear();
+ }
+}
+
+#[derive(Debug)]
+pub struct Svg {
+ pub handle: Handle,
+ pub position: [f32; 2],
+ pub scale: [f32; 2],
+}
+
+#[repr(C)]
+#[derive(Clone, Copy)]
+pub struct Vertex {
+ _position: [f32; 2],
+}
+
+const QUAD_INDICES: [u16; 6] = [0, 1, 2, 0, 2, 3];
+
+const QUAD_VERTS: [Vertex; 4] = [
+ Vertex {
+ _position: [0.0, 0.0],
+ },
+ Vertex {
+ _position: [1.0, 0.0],
+ },
+ Vertex {
+ _position: [1.0, 1.0],
+ },
+ Vertex {
+ _position: [0.0, 1.0],
+ },
+];
+
+#[repr(C)]
+#[derive(Clone, Copy)]
+struct Instance {
+ _position: [f32; 2],
+ _scale: [f32; 2],
+}
+
+#[repr(C)]
+#[derive(Debug, Clone, Copy)]
+struct Uniforms {
+ transform: [f32; 16],
+} \ No newline at end of file