feat: GPU scissor rect for paper-space viewport wires

Viewport-projected wires now clip at pixel level to their viewport boundary.

- WireModel.vp_scissor: Option<[f32; 4]> — paper-space bbox set on projection
- WireGpu.vp_scissor carries it through to the draw loop
- Pipeline.compute_wire_scissors() converts paper→NDC→pixel each frame
- render() calls set_scissor_rect() per wire, reset after viewport group
- INTEGRATION_GAPS.md: Viewport ⚠️ → ✅, all 51 gaps resolved

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

src/scene/complex_lt.rs

@@ -203,6 +203,7 @@ pub fn apply_along(
             key_vertices: vec![],
             aabb: WireModel::UNBOUNDED_AABB,
             plinegen: true,
+            vp_scissor: None,
         })
         .collect()
 }

src/scene/mod.rs

@@ -976,6 +976,7 @@ impl Scene {
                 out.points = clipped;
                 out.color = [r * 0.80, g * 0.80, b * 0.80, a * 0.85];
                 out.line_weight_px = wire.line_weight_px;
+                out.vp_scissor = Some([vp_x0, vp_y0, vp_x1, vp_y1]);
                 projected.push(out);
             }
 
@@ -3212,6 +3213,7 @@ fn tessellate_entity(
             key_vertices: vec![],
             aabb: WireModel::UNBOUNDED_AABB,
             plinegen: true,
+            vp_scissor: None,
         });
         return wires;
     }

src/scene/pipeline/mod.rs

@@ -52,6 +52,8 @@ pub struct Pipeline {
     /// Cached texture format (needed to recreate MSAA / depth textures on resize).
     surface_format: wgpu::TextureFormat,
     gpu_wires: Vec<WireGpu>,
+    /// Pixel scissor rects [x, y, w, h] for viewport-clipped wires. Recomputed each frame.
+    wire_pixel_scissors: Vec<Option<[u32; 4]>>,
     /// Ghost copies (25% alpha) of selected wires for the X-ray depth pass.
     gpu_selected_wires: Vec<WireGpu>,
     gpu_hatches: Vec<HatchGpu>,
@@ -584,6 +586,7 @@ impl Pipeline {
             blit_bind_group,
             surface_format: format,
             gpu_wires: vec![],
+            wire_pixel_scissors: vec![],
             gpu_selected_wires: vec![],
             gpu_hatches: vec![],
             gpu_wipeouts: vec![],
@@ -607,6 +610,30 @@ impl Pipeline {
             .collect();
     }
 
+    /// Recompute pixel scissor rects for viewport-clipped wires from the current view_proj.
+    /// Called every frame from prepare() because scissor pixels shift with pan/zoom.
+    pub fn compute_wire_scissors(&mut self, view_proj: glam::Mat4, clip_w: u32, clip_h: u32) {
+        let w = clip_w as f32;
+        let h = clip_h as f32;
+        self.wire_pixel_scissors = self.gpu_wires.iter().map(|wire| {
+            let [x0, y0, x1, y1] = wire.vp_scissor?;
+            let corners = [
+                view_proj.project_point3(glam::Vec3::new(x0, y0, 0.0)),
+                view_proj.project_point3(glam::Vec3::new(x1, y0, 0.0)),
+                view_proj.project_point3(glam::Vec3::new(x0, y1, 0.0)),
+                view_proj.project_point3(glam::Vec3::new(x1, y1, 0.0)),
+            ];
+            let px: Vec<f32> = corners.iter().map(|c| (c.x + 1.0) * 0.5 * w).collect();
+            let py: Vec<f32> = corners.iter().map(|c| (1.0 - c.y) * 0.5 * h).collect();
+            let sx0 = px.iter().cloned().fold(f32::INFINITY, f32::min).max(0.0) as u32;
+            let sy0 = py.iter().cloned().fold(f32::INFINITY, f32::min).max(0.0) as u32;
+            let sx1 = (px.iter().cloned().fold(f32::NEG_INFINITY, f32::max) as u32).min(clip_w);
+            let sy1 = (py.iter().cloned().fold(f32::NEG_INFINITY, f32::max) as u32).min(clip_h);
+            if sx1 <= sx0 || sy1 <= sy0 { return None; }
+            Some([sx0, sy0, sx1 - sx0, sy1 - sy0])
+        }).collect();
+    }
+
     /// Upload all 3DFACE entities as two batched GPU objects:
     /// - `gpu_face3d_fill`: filled triangles (1 buffer, 1 draw call)
     /// - `gpu_face3d_edges`: merged edge wires (1 buffer, 1 draw call)
@@ -875,11 +902,27 @@ impl Pipeline {
             pass.set_viewport(0.0, 0.0, vp.width as f32, vp.height as f32, 0.0, 1.0);
             pass.set_pipeline(&self.wire_pipeline);
             pass.set_bind_group(0, &self.uniform_bind_group, &[]);
-            for wire in &self.gpu_wires {
-                if wire.vertex_count >= 6 {
-                    pass.set_vertex_buffer(0, wire.vertex_buffer.slice(..));
-                    pass.draw(0..wire.vertex_count, 0..1);
+            let mut scissor_active = false;
+            for (i, wire) in self.gpu_wires.iter().enumerate() {
+                if wire.vertex_count < 6 {
+                    continue;
                 }
+                match self.wire_pixel_scissors.get(i) {
+                    Some(Some([x, y, w, h])) => {
+                        pass.set_scissor_rect(*x, *y, *w, *h);
+                        scissor_active = true;
+                    }
+                    _ if scissor_active => {
+                        pass.set_scissor_rect(0, 0, vp.width, vp.height);
+                        scissor_active = false;
+                    }
+                    _ => {}
+                }
+                pass.set_vertex_buffer(0, wire.vertex_buffer.slice(..));
+                pass.draw(0..wire.vertex_count, 0..1);
+            }
+            if scissor_active {
+                pass.set_scissor_rect(0, 0, vp.width, vp.height);
             }
         }
 

src/scene/pipeline/wire_gpu.rs

@@ -111,11 +111,16 @@ impl WireVertex {
 pub struct WireGpu {
     pub vertex_buffer: wgpu::Buffer,
     pub vertex_count: u32,
+    /// Paper-space bbox [x0, y0, x1, y1] for GPU scissor clipping.
+    /// Set only for viewport-projected wires; None for regular wires.
+    pub vp_scissor: Option<[f32; 4]>,
 }
 
 impl WireGpu {
     pub fn new(device: &wgpu::Device, wire: &WireModel) -> Self {
-        Self::build(device, wire, wire.color)
+        let mut g = Self::build(device, wire, wire.color);
+        g.vp_scissor = wire.vp_scissor;
+        g
     }
 
     /// Creates a ghost copy with `alpha` applied on top of the wire's own alpha.
@@ -209,6 +214,7 @@ impl WireGpu {
                 Self {
                     vertex_buffer,
                     vertex_count: chunk.len() as u32,
+                    vp_scissor: None,
                 }
             })
             .collect()
@@ -310,6 +316,7 @@ impl WireGpu {
         Self {
             vertex_buffer,
             vertex_count: vertices.len() as u32,
+            vp_scissor: None,
         }
     }
 }

src/scene/render.rs

@@ -131,6 +131,7 @@ impl shader::Primitive for Primitive {
             pipeline.upload_face3d(device, &self.face3d_wires[..]);
             pipeline.cached_epoch = self.geometry_epoch;
         }
+        pipeline.compute_wire_scissors(self.uniforms.view_proj, clip_size.width, clip_size.height);
         if self.show_viewcube {
             pipeline.viewcube.upload(
                 queue,

src/scene/tessellate.rs

@@ -105,6 +105,7 @@ pub fn tessellate(
             key_vertices: te.key_vertices,
             aabb: WireModel::UNBOUNDED_AABB,
             plinegen: true,
+            vp_scissor: None,
                 };
             }
 
@@ -134,6 +135,7 @@ pub fn tessellate(
                             key_vertices: te.key_vertices,
                             aabb: WireModel::UNBOUNDED_AABB,
             plinegen: true,
+            vp_scissor: None,
                         };
                     }
                     _ => {}
@@ -161,6 +163,7 @@ pub fn tessellate(
                         key_vertices,
                         aabb: WireModel::UNBOUNDED_AABB,
             plinegen: true,
+            vp_scissor: None,
                     };
                 }
             }
@@ -186,6 +189,7 @@ pub fn tessellate(
                         key_vertices,
                         aabb: WireModel::UNBOUNDED_AABB,
             plinegen: true,
+            vp_scissor: None,
                     };
                 }
             }
@@ -217,6 +221,7 @@ pub fn tessellate(
                     key_vertices,
                     aabb: WireModel::UNBOUNDED_AABB,
                     plinegen: true,
+            vp_scissor: None,
                 };
             }
 
@@ -243,6 +248,7 @@ pub fn tessellate(
                     aci: 0,
                     key_vertices,
                     plinegen: false,
+            vp_scissor: None,
                     aabb: WireModel::UNBOUNDED_AABB,
                 };
             }
@@ -274,6 +280,7 @@ pub fn tessellate(
         key_vertices,
         aabb: WireModel::UNBOUNDED_AABB,
             plinegen: true,
+            vp_scissor: None,
     }
 }
 
@@ -330,6 +337,7 @@ pub fn tessellate_dimension(
         key_vertices,
         aabb: WireModel::UNBOUNDED_AABB,
             plinegen: true,
+            vp_scissor: None,
     }];
 
     if let Some(text) = dimension_text_entity(dim) {
@@ -379,6 +387,7 @@ fn tessellate_leader(
             key_vertices: vec![],
             aabb: WireModel::UNBOUNDED_AABB,
             plinegen: true,
+            vp_scissor: None,
         }];
     }
 
@@ -448,6 +457,7 @@ fn tessellate_leader(
         key_vertices,
         aabb: WireModel::UNBOUNDED_AABB,
             plinegen: true,
+            vp_scissor: None,
     }]
 }
 
@@ -558,6 +568,7 @@ fn tessellate_multileader(
             key_vertices: key_verts,
             aabb: WireModel::UNBOUNDED_AABB,
             plinegen: true,
+            vp_scissor: None,
     }];
 
     // Render text content as MText wire

src/scene/wire_model.rs

@@ -63,6 +63,9 @@ pub struct WireModel {
     /// When false the linetype pattern restarts at each NaN-separated segment
     /// (DXF PLINEGEN=0).  When true the pattern runs continuously (PLINEGEN=1).
     pub plinegen: bool,
+    /// Paper-space bounding box [x0, y0, x1, y1] for GPU scissor clipping.
+    /// Set only for viewport-projected wires in paper-space layouts.
+    pub vp_scissor: Option<[f32; 4]>,
 }
 
 impl WireModel {
@@ -89,6 +92,7 @@ impl WireModel {
             key_vertices: vec![],
             aabb: Self::UNBOUNDED_AABB,
             plinegen: true,
+            vp_scissor: None,
         }
     }
 
@@ -189,6 +193,7 @@ impl Default for WireModel {
             key_vertices: Vec::new(),
             aabb: Self::UNBOUNDED_AABB,
             plinegen: true,
+            vp_scissor: None,
         }
     }
 }