[iwoplaza]

This lets the browser handle the error as if it was unhandled, yet we can still keep our code running.

[github-actions]

pkg.pr.new

No packages changed since release

[pullfrog]

✅ No new issues found.

Reviewed changes

• Example error reporting — in apps/typegpu-docs/src/utils/examples/currentSnackbarAtom.ts, runWithCatchAtom now calls window.reportError(err) instead of console.log(err), so the browser treats the caught error as unhandled while the example keeps running.

This is a correct use of reportError: it is typed via WindowOrWorkerGlobalScope in the DOM lib (the docs tsconfig already includes DOM), it is invoked only from browser event handlers, and its Baseline-2023 support (Chrome 95+, Firefox 100+, Safari 16.4+) is fine for a WebGPU docs site. No blocking concerns.

ℹ️ Same pattern still uses console.error in ExampleView

apps/typegpu-docs/src/components/ExampleView.tsx:75,78 catch example-load failures and call console.error directly. If the intent is consistent "unhandled" error signalling across the examples page, those two sites are candidates for the same treatment. Non-blocking and outside this diff.

  ^{｜ View workflow run ｜ Using deepseek-v4.1-flash (free via Pullfrog for OSS) ｜ 𝕏}

[github-actions]

Resolution Time Benchmark

---
config:
  themeVariables:
    xyChart:
      plotColorPalette: "#E63946, #3B82F6, #059669"
---
xychart
  title "Random Branching (🔴 PR | 🔵 main | 🟢 release)"
  x-axis "max depth" [1, 2, 3, 4, 5, 6, 7, 8]
  y-axis "time (ms)"
  line [0.97, 1.92, 4.25, 6.44, 7.42, 12.64, 22.58, 23.49]
  line [1.01, 1.95, 4.29, 5.86, 7.98, 12.66, 25.21, 26.78]
  line [0.96, 2.02, 4.03, 6.18, 7.84, 11.12, 23.92, 25.59]

Loading

---
config:
  themeVariables:
    xyChart:
      plotColorPalette: "#E63946, #3B82F6, #059669"
---
xychart
  title "Linear Recursion (🔴 PR | 🔵 main | 🟢 release)"
  x-axis "max depth" [1, 2, 3, 4, 5, 6, 7, 8]
  y-axis "time (ms)"
  line [0.28, 0.57, 0.69, 0.80, 1.12, 1.21, 1.41, 1.59]
  line [0.29, 0.48, 0.67, 0.90, 1.17, 1.23, 1.43, 1.60]
  line [0.29, 0.52, 0.76, 0.85, 1.20, 1.19, 1.45, 1.57]

Loading

---
config:
  themeVariables:
    xyChart:
      plotColorPalette: "#E63946, #3B82F6, #059669"
---
xychart
  title "Full Tree (🔴 PR | 🔵 main | 🟢 release)"
  x-axis "max depth" [1, 2, 3, 4, 5, 6, 7, 8]
  y-axis "time (ms)"
  line [0.97, 2.13, 4.17, 7.71, 12.37, 25.59, 55.83, 111.64]
  line [0.91, 2.15, 3.91, 6.97, 11.82, 25.87, 57.11, 112.23]
  line [0.85, 2.13, 3.79, 6.93, 12.43, 25.35, 57.78, 109.58]

Loading

[github-actions]

Bundle size comparison (import * as ... in PR vs import * as ... in target):

🟢 Decreased	➖ Unchanged	🔴 Increased	❔ Unknown
0	355	0	0

No notable changes.

If you wish to run a comparison for other, slower bundlers, run the 'Tree-shake test' from the GitHub Actions menu.

[aleksanderkatan]

Why do we want to treat the error like it was unhandled? I think it was handled pretty well: we caught it, logged it and displayed it in the snackbar.