RM-190: Finish Plotting Core Function Set

[nallana]

Note

High Risk
Touches core plotting render and serialization paths, adding per-axes uniform state and new scene schema fields; regressions could affect subplot rendering correctness and backward compatibility of figure snapshots/scenes.

Overview
Expands plotting capabilities and scene serialization. Adds new plot kinds and metadata to the figure scene/event schema (e.g. line3, quiver, pie, richer errorbar/stem styling, area lower curve, surface image_mode + optional per-cell colors), plus per-axes metadata (limits, log modes, view angles, labels/styles, legend style, annotations, active axes).

Reworks subplot rendering to be truly per-axes. The renderer now maintains per-axes cameras, applies stored 3D view angles, fits extents per-axes (including honoring axes-specific limits/axis_equal), and uses per-axes uniform/direct-uniform/grid-uniform bind groups to avoid state bleed between subplots; subplot passes also force msaa_samples=1 to prevent resolves from overwriting other panels.

Adds GPU compute “packers” for more primitives. Introduces new WGPU compute shaders and Rust wrappers for generating vertex buffers on-GPU for area, errorbar, image (truecolor), line3, quiver, and stem.

Also bumps the TS bindings package version to 0.3.4 and updates exports to include pie labels in image/SVG exports.

^{Written by Cursor Bugbot for commit b0df409. This will update automatically on new commits. Configure here.}

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
runmat	Ready	Preview, Comment, Open in v0	Mar 30, 2026 7:22pm

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

Bugbot Autofix prepared a fix for the issue found in the latest run.

• ✅ Fixed: Stem GPU buffer underallocated when baseline hidden
  • The bug was real, and I fixed it by making the stem vertex base offset conditional on baseline_visible so hidden-baseline draws no longer write past the allocated buffer.

Or push these changes by commenting:

@cursor push 6b597cb148

Preview (6b597cb148)

diff --git a/crates/runmat-plot/src/gpu/shaders/stem.rs b/crates/runmat-plot/src/gpu/shaders/stem.rs
--- a/crates/runmat-plot/src/gpu/shaders/stem.rs
+++ b/crates/runmat-plot/src/gpu/shaders/stem.rs
@@ -74,7 +74,7 @@
         return;
     }
 
-    let base = 2u + i * 2u;
+    let base = select(0u, 2u, params.baseline_visible != 0u) + i * 2u;
     write_vertex(base, vec3<f32>(buf_x[i], params.baseline, 0.0), params.color);
     write_vertex(base + 1u, vec3<f32>(buf_x[i], buf_y[i], 0.0), params.color);
 }
@@ -156,7 +156,7 @@
         return;
     }
 
-    let base = 2u + i * 2u;
+    let base = select(0u, 2u, params.baseline_visible != 0u) + i * 2u;
     write_vertex(base, vec3<f32>(f32(buf_x[i]), params.baseline, 0.0), params.color);
     write_vertex(base + 1u, vec3<f32>(f32(buf_x[i]), f32(buf_y[i]), 0.0), params.color);
 }

_{This Bugbot Autofix run was free. To enable autofix for future PRs, go to the Cursor dashboard.}

[cursor]

Stem GPU buffer underallocated when baseline hidden

High Severity

The GPU output buffer is too small when baseline_visible is false. The shader always writes stem vertices at base = 2u + i * 2u (reserving indices 0–1 for the baseline regardless of visibility), but the Rust allocator uses baseline_count + len * 2 which drops to len * 2 when the baseline is hidden. The last point writes at index 2*len + 1, exceeding the 2*len-sized buffer by 2 vertices — an out-of-bounds GPU storage write.

Additional Locations (1)

• crates/runmat-plot/src/gpu/shaders/stem.rs#L76-L79