Resolve a primitive's position through a chain of move slots

The plumbing for O(1) subtree movement (LAYOUT.md §2), with every slot
still at zero, so this changes no pixels and the next commit can change
behaviour against a known-good picture.

Every active widget owns a slot in `UiData::moves`: a translation in
physical pixels and the slot it is relative to. A primitive instance and
a mask each name one, and `prelude.wgsl` walks the chain and adds the
accumulated delta. A mask resolves its own chain rather than the drawn
primitive's, so a stationary viewport can clip content that moves inside
it. `CHAIN_LIMIT` is stated on both sides; it bounds a malformed cycle
rather than any real tree.

A slot outlives any one `ActiveData`, because a redraw replaces that
while the widget's children go on pointing at the slot, so it lives in
`UiRenderState::moves` keyed by widget and is retired when the widget
stops being drawn. `MoveIdx` is its own type rather than another
`Id<u32>`: it sits beside `MaskIdx` in an instance and the two must not
be swappable.

`Vec2` is now `repr(align(8))`, which is WGSL's alignment for a
`vec2<f32>`, so a GPU struct holding one is laid out the way its shader
reads it without saying so itself -- `GlyphPrimitive` no longer states
its own alignment, and `MoveOffset` never has to. Both keep a manual
`unsafe impl Pod`, since the trailing padding that alignment introduces
is what `derive(Pod)` refuses. `WindowUniform` holds the `Vec2` its
shader has always called `dim` rather than two loose floats, which was
the last place the two sides described the same bytes differently.

Checked: fmt, clippy and 40 tests. `tabs` (with the image replay),
`view` and `minimal` render byte-identical to `upstream/main`, and
`text` is unchanged.

core/src/render/data.rs

@@ -1,11 +1,10 @@
-use crate::{UiRegion, util::Id};
+use crate::{UiRegion, util::Id, util::Vec2};
 use wgpu::*;
 
 #[repr(C)]
 #[derive(Copy, Clone, bytemuck::Pod, bytemuck::Zeroable, Default)]
 pub struct WindowUniform {
-    pub width: f32,
-    pub height: f32,
+    pub dim: Vec2,
 }
 
 #[repr(C)]
@@ -13,15 +12,17 @@ pub struct WindowUniform {
 pub struct PrimitiveInstance {
     pub region: UiRegion,
     pub mask_idx: MaskIdx,
+    pub move_idx: MoveIdx,
 }
 
 impl PrimitiveInstance {
-    const ATTRIBS: [VertexAttribute; 5] = vertex_attr_array![
+    const ATTRIBS: [VertexAttribute; 6] = vertex_attr_array![
         0 => Float32x2,
         1 => Float32x2,
         2 => Float32x2,
         3 => Float32x2,
         4 => Uint32,
+        5 => Uint32,
     ];
 
     pub fn desc() -> VertexBufferLayout<'static> {
@@ -43,4 +44,45 @@ impl MaskIdx {
 #[derive(Debug, Copy, Clone, bytemuck::Pod, bytemuck::Zeroable)]
 pub struct Mask {
     pub region: UiRegion,
+    pub move_idx: MoveIdx,
+}
+
+/// Its own type rather than another `Id<u32>`, because it sits beside
+/// `MaskIdx` in an instance and the two must not be swappable.
+#[repr(transparent)]
+#[derive(Debug, Copy, Clone, PartialEq, Eq, Hash, bytemuck::Pod, bytemuck::Zeroable)]
+pub struct MoveIdx(u32);
+
+impl MoveIdx {
+    pub const NONE: Self = Self(u32::MAX);
+
+    pub(crate) fn slot(idx: usize) -> Self {
+        Self(idx as u32)
+    }
+
+    pub(crate) fn idx(self) -> usize {
+        self.0 as usize
+    }
+}
+
+/// One link of the chain a primitive's position is resolved through: a
+/// translation in physical pixels, and the slot it is relative to. Moving a
+/// subtree writes its own slot and nothing else.
+#[repr(C)]
+#[derive(Debug, Copy, Clone)]
+pub struct MoveOffset {
+    pub delta: Vec2,
+    pub parent: MoveIdx,
+}
+
+unsafe impl bytemuck::Pod for MoveOffset {}
+unsafe impl bytemuck::Zeroable for MoveOffset {}
+
+impl MoveOffset {
+    pub fn root(parent: MoveIdx) -> Self {
+        Self {
+            delta: Vec2::ZERO,
+            parent,
+        }
+    }
 }

core/src/render/mod.rs

@@ -17,7 +17,7 @@ mod texture;
 mod util;
 
 pub use atlas::*;
-pub use data::{Mask, MaskIdx};
+pub use data::{Mask, MaskIdx, MoveIdx, MoveOffset};
 pub use primitive::*;
 
 const PRELUDE: &str = include_str!("./shader/prelude.wgsl");
@@ -34,6 +34,7 @@ pub struct UiRenderNode {
     active: Vec<usize>,
     window_buffer: Buffer,
     masks: ArrBuf<Mask>,
+    moves: ArrBuf<MoveOffset>,
 }
 
 struct RenderLayer {
@@ -127,32 +128,35 @@ impl UiRenderNode {
         for primitive in &mut self.primitives {
             primitive.render.update(ui);
         }
+        let mut regroup = false;
         if ui.masks.changed {
             ui.masks.changed = false;
-            if self.masks.update(device, queue, &ui.masks[..]) {
-                self.shared_group = Self::shared_group(
-                    device,
-                    &self.shared_layout,
-                    &self.window_buffer,
-                    &self.masks,
-                );
-            }
+            regroup |= self.masks.update(device, queue, &ui.masks[..]);
+        }
+        if ui.moves.changed {
+            ui.moves.changed = false;
+            regroup |= self.moves.update(device, queue, ui.moves.entries());
+        }
+        if regroup {
+            self.shared_group = Self::shared_group(
+                device,
+                &self.shared_layout,
+                &self.window_buffer,
+                &self.masks,
+                &self.moves,
+            );
         }
     }
 
     pub fn resize(&mut self, size: impl Into<Vec2>, queue: &Queue) {
         let size = size.into();
-        let slice = &[WindowUniform {
-            width: size.x,
-            height: size.y,
-        }];
+        let slice = &[WindowUniform { dim: size }];
         queue.write_buffer(&self.window_buffer, 0, bytemuck::cast_slice(slice));
     }
 
     pub fn new(device: &Device, config: &SurfaceConfiguration) -> Self {
         let window_uniform = WindowUniform {
-            width: config.width as f32,
-            height: config.height as f32,
+            dim: Vec2::new(config.width as f32, config.height as f32),
         };
         let window_buffer = device.create_buffer_init(&BufferInitDescriptor {
             label: Some("window"),
@@ -166,7 +170,13 @@ impl UiRenderNode {
             BufferUsages::STORAGE | BufferUsages::COPY_DST,
             "ui masks",
         );
-        let shared_group = Self::shared_group(device, &shared_layout, &window_buffer, &masks);
+        let moves = ArrBuf::new(
+            device,
+            BufferUsages::STORAGE | BufferUsages::COPY_DST,
+            "ui move offsets",
+        );
+        let shared_group =
+            Self::shared_group(device, &shared_layout, &window_buffer, &masks, &moves);
 
         Self {
             shared_layout,
@@ -177,6 +187,7 @@ impl UiRenderNode {
             layers: HashMap::default(),
             active: Vec::new(),
             masks,
+            moves,
         }
     }
 
@@ -252,7 +263,8 @@ impl UiRenderNode {
         })
     }
 
-    /// What every draw in the ui is given: the window and the masks.
+    /// What every draw in the ui is given: the window, the masks and the
+    /// move chain every position is resolved through.
     fn shared_layout(device: &Device) -> BindGroupLayout {
         device.create_bind_group_layout(&BindGroupLayoutDescriptor {
             entries: &[
@@ -276,6 +288,16 @@ impl UiRenderNode {
                     },
                     count: None,
                 },
+                BindGroupLayoutEntry {
+                    binding: 2,
+                    visibility: ShaderStages::VERTEX | ShaderStages::FRAGMENT,
+                    ty: BindingType::Buffer {
+                        ty: BufferBindingType::Storage { read_only: true },
+                        has_dynamic_offset: false,
+                        min_binding_size: BufferSize::new(size_of::<MoveOffset>() as u64),
+                    },
+                    count: None,
+                },
             ],
             label: Some("ui shared"),
         })
@@ -286,6 +308,7 @@ impl UiRenderNode {
         layout: &BindGroupLayout,
         window: &Buffer,
         masks: &ArrBuf<Mask>,
+        moves: &ArrBuf<MoveOffset>,
     ) -> BindGroup {
         device.create_bind_group(&BindGroupDescriptor {
             layout,
@@ -298,6 +321,10 @@ impl UiRenderNode {
                     binding: 1,
                     resource: masks.buffer.as_entire_binding(),
                 },
+                BindGroupEntry {
+                    binding: 2,
+                    resource: moves.buffer.as_entire_binding(),
+                },
             ],
             label: Some("ui shared"),
         })

core/src/render/primitive.rs

@@ -3,7 +3,7 @@ use std::{any::TypeId, marker::PhantomData};
 use crate::{
     Color, TextureHandle, UiData, UiRegion, WidgetId,
     render::{
-        data::{MaskIdx, PrimitiveInstance},
+        data::{MaskIdx, MoveIdx, PrimitiveInstance},
         page::GlyphRender,
         texture::ImageRender,
     },
@@ -246,6 +246,7 @@ impl LayerDraws {
             primitive,
             region,
             mask_idx,
+            move_idx,
         }: PrimitiveInst<P>,
     ) -> PrimitiveHandle {
         self.updated = true;
@@ -258,7 +259,11 @@ impl LayerDraws {
             .get_or_insert_with(InstanceList::new::<P>)
             .push(
                 id,
-                PrimitiveInstance { region, mask_idx },
+                PrimitiveInstance {
+                    region,
+                    mask_idx,
+                    move_idx,
+                },
                 bytemuck::bytes_of(&primitive),
             );
         PrimitiveHandle {
@@ -304,6 +309,7 @@ pub struct PrimitiveInst<P> {
     pub primitive: P,
     pub region: UiRegion,
     pub mask_idx: MaskIdx,
+    pub move_idx: MoveIdx,
 }
 
 pub struct PrimitiveChange {
@@ -347,7 +353,7 @@ impl RectPrimitive {
 
 /// `color` is multiplied by the atlas alpha for a mask glyph; a colour glyph
 /// takes the texel unchanged, which `GlyphEntry::IS_COLORED` selects.
-#[repr(C, align(8))]
+#[repr(C)]
 #[derive(Debug, Copy, Clone)]
 pub struct GlyphPrimitive {
     pub uv_min: Vec2,
@@ -358,8 +364,8 @@ pub struct GlyphPrimitive {
     pub flags: u32,
 }
 
-// Manual rather than derived: the align(8) leaves four bytes of padding, which
-// is how WGSL lays the struct out.
+// Manual rather than derived: `Vec2`'s alignment leaves four bytes of padding
+// here, which is how WGSL lays the struct out.
 unsafe impl bytemuck::Pod for GlyphPrimitive {}
 unsafe impl bytemuck::Zeroable for GlyphPrimitive {}
 impl Primitive for GlyphPrimitive {

core/src/render/shader/prelude.wgsl

@@ -7,6 +7,8 @@
 var<uniform> window: WindowUniform;
 @group(0) @binding(1)
 var<storage> masks: array<Mask>;
+@group(0) @binding(2)
+var<storage> move_offsets: array<MoveOffset>;
 
 struct WindowUniform {
     dim: vec2<f32>,
@@ -15,6 +17,32 @@ struct WindowUniform {
 struct Mask {
     x: UiSpan,
     y: UiSpan,
+    move_idx: u32,
+}
+
+struct MoveOffset {
+    delta: vec2<f32>,
+    parent: u32,
+}
+
+const MOVE_NONE: u32 = 4294967295u;
+// Keep in step with `iris_core::CHAIN_LIMIT`. It bounds a malformed cycle
+// rather than any real tree, and the CPU walk uses the same number so both
+// resolve a deep one the same way.
+const CHAIN_LIMIT: u32 = 64u;
+
+fn resolve_move(idx: u32) -> vec2<f32> {
+    var total = vec2<f32>(0.0, 0.0);
+    var at = idx;
+    for (var step = 0u; step < CHAIN_LIMIT; step++) {
+        if at == MOVE_NONE {
+            break;
+        }
+        let entry = move_offsets[at];
+        total += entry.delta;
+        at = entry.parent;
+    }
+    return total;
 }
 
 struct UiSpan {
@@ -33,6 +61,7 @@ struct InstanceInput {
     @location(2) y_start: vec2<f32>,
     @location(3) y_end: vec2<f32>,
     @location(4) mask_idx: u32,
+    @location(5) move_idx: u32,
 }
 
 struct VertexOutput {
@@ -57,8 +86,9 @@ fn vs_main(
     let bot_right_rel = vec2(in.x_end.x, in.y_end.x);
     let bot_right_abs = vec2(in.x_end.y, in.y_end.y);
 
-    let top_left = floor(top_left_rel * window.dim) + floor(top_left_abs);
-    let bot_right = floor(bot_right_rel * window.dim) + floor(bot_right_abs);
+    let moved = resolve_move(in.move_idx);
+    let top_left = floor(top_left_rel * window.dim) + floor(top_left_abs + moved);
+    let bot_right = floor(bot_right_rel * window.dim) + floor(bot_right_abs + moved);
     let size = bot_right - top_left;
 
     let uv = vec2<f32>(
@@ -86,8 +116,11 @@ fn masked(in: VertexOutput, color: vec4<f32>) -> vec4<f32> {
     let br = vec2(mask.x.end.rel, mask.y.end.rel);
     let br_abs = vec2(mask.x.end.abs, mask.y.end.abs);
 
-    let top_left = floor(tl * window.dim) + floor(tl_abs);
-    let bot_right = floor(br * window.dim) + floor(br_abs);
+    // Its own chain, not the drawn primitive's, so a stationary viewport
+    // clips content that moves inside it.
+    let moved = resolve_move(mask.move_idx);
+    let top_left = floor(tl * window.dim) + floor(tl_abs + moved);
+    let bot_right = floor(br * window.dim) + floor(br_abs + moved);
     let pos = in.clip_position.xy;
     if pos.x < top_left.x || pos.x > bot_right.x || pos.y < top_left.y || pos.y > bot_right.y {
         return color * 0.0;