plan(2026-06-01) + feat(telemetry): weekly plan & exportable engine session report (#699)

[ford442]

1. Weekly planner run — 2026-06-01 (User Idea mode)

Updates weekly_plan.md: knob render-pass plumbing moved to Done (landed on main, build passing); the idea's core drift-kill carried into today's focus (shared knobMaterial contract). Reconciled backlog — closed-unmerged PRs #506/#507 removed, 0-open-issues flagged, duplicate Copilot PRs #685/#686 + Jules PR #693 added.

2. feat(telemetry): exportable hybrid-engine session report — Closes #699

Adds a manual, privacy-preserving export of hybrid audio-engine telemetry.

Changes

• src/utils/engineTelemetry.ts — bounded per-subsystem resolutionHistory (cap 20); sampleCount + p95Reliable on snapshot() (threshold 128, consistent with the 256-sample ring cap — not 500); getBrowserCapabilities() (capability tier; no raw userAgent, optional low-entropy userAgentData); pure serializeEngineReport(); injectable DownloadProvider; generateReportJSON() + exportReport().
• src/main.tsx — registers window.__devtools.exportEngineReport at app bootstrap (not on HUD/component mount), gated on import.meta.env.DEV || ?devtools.
• src/components/EngineHUD.tsx — "Download Report" + "Copy JSON" buttons via event delegation (survives the 500ms innerHTML re-render); clears inert while visible so buttons are clickable.
• src/utils/__tests__/engineTelemetry.test.ts — 9 tests: serialization shape, no-userAgent-leak, filename safety, capability shape, p95Reliable boundary (127→128), history cap (25→20), ring cap (300→256), injected export provider.

Verification

• npx tsc -b → 0 errors
• npx vitest run → 934 passed, 1 skipped, 0 failed (new file: 9/9)
• eslint on touched files → clean
• Full npm run build not run here — needs the WASM toolchain (asc/wasm-pack/emscripten) absent from this container; covered by CI / the mid-day pipeline task.

Design notes (resolved open questions)

• Decoupled export from the HUD (registered at bootstrap) so it works even if the user never opened the HUD.
• Privacy: capability tier, no raw userAgent; manual-only; dev/?devtools gated.
• IndexedDB cross-session persistence explicitly rejected as over-engineered; in-memory ring buffer only.

https://claude.ai/code/session_01WGd5LsqEqQawHbQ4QKT4dd

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 3e3bf279-82b0-4d4c-84fa-b01c9110e8a8

📥 Commits

Reviewing files that changed from the base of the PR and between 2f6b9e4 and f31c414.

📒 Files selected for processing (1)

• weekly_plan.md

---

📝 Walkthrough

Walkthrough

Updated weekly_plan.md to reflect a shift in focus for the holographic-knob renderer work. The plan now identifies shared knobMaterial constants extraction as the concrete task for preventing visual drift between WGSL and 2D rendering paths, following completion of singleton plumbing infrastructure.

Changes

Weekly knob-drift-kill planning

Layer / File(s)	Summary
Plan update and focus refinement weekly_plan.md	Shifted focus to "knob drift-kill" task: extract shared knobMaterial constants module (palette, geometry, bloom, background) used by both KnobGPUContext WGSL uniform packing and HardwareModule.renderWith2D. Updated backlog, Done section (noting plumbing completion and that visual-drift goal remains open), and Last run section to clarify drift root cause as independently hardcoded palettes.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~2 minutes

Possibly related PRs

• ford442/web_sequencer#510: Both PRs update weekly_plan.md with overlapping holographic knob renderer work items—docs(plan): weekly plan update 2026-05-04 #510 sets up the knob perf audit/WGSL unification, while the main PR advances that same "knob material/WGSL vs renderWith2D contract" tracking.
• ford442/web_sequencer#618: Main PR's 2026-06-01 weekly plan builds on #618's WGSL render-pass/material-contract unification by adding the next specific step to extract shared knobMaterial constants between KnobGPUContext uniform packing and HardwareModule.renderWith2D, directly addressing the same WGSL-vs-2D "drift" issue.
• ford442/web_sequencer#530: Both PRs update weekly_plan.md around the holographic-knob unification work, specifically continuing/expanding the KnobGPUContext.ts singleton implementation focus into a shared knobMaterial contract.

Poem

🐰 A rabbit hops through plans so grand,
Where shaders drift across the land,
But now we'll bind with one shared hold—
A knobMaterial true and bold!
No more two palettes, side by side,
One truth shall be our trusty guide! 🎚️

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch claude/gallant-carson-8737B

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}