feat: Implement workspace system (Task #77)

Self-contained workspaces for parallel demo development.

Structure:
- workspaces/main,test - Demo-specific resources
- assets/common - Shared resources
- workspace.cfg - Configuration per workspace

CMake integration:
- DEMO_WORKSPACE option (defaults to main)
- cmake/ParseWorkspace.cmake - Config parser
- Workspace-relative asset/timeline/music paths

Migration:
- Main demo: demo.seq to workspaces/main/timeline.seq
- Test demo: test_demo.seq to workspaces/test/timeline.seq
- Common shaders: assets/common/shaders
- Workspace shaders: workspaces/*/shaders

Build:
cmake -B build -DDEMO_WORKSPACE=main
cmake -B build_test -DDEMO_WORKSPACE=test

All tests passing (36/36).

handoff(Claude): Task #77 workspace system complete. Both main and test workspaces build and pass all tests.

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

1 files changed, 51 insertions, 0 deletions

diff --git a/workspaces/test/shaders/ellipse.wgsl b/workspaces/test/shaders/ellipse.wgsl
new file mode 100644
index 0000000..05dfcfc
--- /dev/null
+++ b/workspaces/test/shaders/ellipse.wgsl
@@ -0,0 +1,51 @@
+#include "common_uniforms"
+
+@group(0) @binding(0) var<uniform> uniforms: CommonUniforms;
+
+@vertex fn vs_main(@builtin(vertex_index) i: u32) -> @builtin(position) vec4<f32> {
+    var pos = array<vec2<f32>, 3>(
+        vec2<f32>(-1.0, -1.0),
+        vec2<f32>(3.0, -1.0),
+        vec2<f32>(-1.0, 3.0)
+    );
+    return vec4<f32>(pos[i], 0.0, 1.0);
+}
+
+fn sdEllipse(p: vec2<f32>, ab: vec2<f32>) -> f32 {
+    var p_abs = abs(p);
+    if (p_abs.x > p_abs.y) {
+        p_abs = vec2<f32>(p_abs.y, p_abs.x);
+    }
+    let l = ab.y * ab.y - ab.x * ab.x;
+    let m = ab.x * p_abs.x / l;
+    let n = ab.y * p_abs.y / l;
+    let m2 = m * m;
+    let n2 = n * n;
+    let c = (m2 + n2 - 1.0) / 3.0;
+    let c3 = c * c * c;
+    let d = c3 + m2 * n2;
+    let g = m + m * n2;
+    var co: f32;
+    if (d < 0.0) {
+        let h = acos((c3 + m2 * n2 * 2.0) / c3) / 3.0;
+        let s = cos(h);
+        let t = sin(h) * sqrt(3.0);
+        co = (sqrt(-c * (s + t * 2.0) + m2) + sign(l) * sqrt(-c * (s - t * 2.0) + m2) + abs(g) / (sqrt(-c * (s + t * 2.0) + m2) * sqrt(-c * (s - t * 2.0) + m2)) - m) / 2.0;
+    } else {
+        let h = 2.0 * m * n * sqrt(d);
+        let s = sign(c3 + m2 * n2 + h) * pow(abs(c3 + m2 * n2 + h), 1.0 / 3.0);
+        let u = sign(c3 + m2 * n2 - h) * pow(abs(c3 + m2 * n2 - h), 1.0 / 3.0);
+        let rx = -s - u + m2 * 2.0;
+        let ry = (s - u) * sqrt(3.0);
+        co = (ry / sqrt(sqrt(rx * rx + ry * ry) - rx) + 2.0 * g / sqrt(rx * rx + ry * ry) - m) / 2.0;
+    }
+    let si = sqrt(max(0.0, 1.0 - co * co));
+    return length(p_abs - vec2<f32>(ab.x * co, ab.y * si)) * sign(p_abs.y * ab.x * co - p_abs.x * ab.y * si);
+}
+
+@fragment fn fs_main(@builtin(position) p: vec4<f32>) -> @location(0) vec4<f32> {
+    let uv = (p.xy / uniforms.resolution - 0.5) * 2.0;
+    let movement = vec2<f32>(sin(uniforms.time * 0.7), cos(uniforms.time * 0.5));
+    let d = sdEllipse((uv * vec2<f32>(uniforms.aspect_ratio, 1.0)) - movement, vec2<f32>(0.5, 0.3) * (1.0 + uniforms.beat * 0.2));
+    return mix(vec4<f32>(0.2, 0.8, 0.4, 1.0), vec4<f32>(0.0), smoothstep(0.0, 0.01, d));
+}