feat(effects): add HDR glass showcase and emissive packet lighting

[ibgreen-openai]

Goals

• Demonstrate reusable state-of-the-art glass effects through Multipath Reliable Connection (MRC) packet spraying inspired by OpenAI's supercomputer networking article.
• Add reusable, portable optical-material and postprocessing primitives that work across WebGPU and WebGL.
• Keep the network diagram readable while showing resilient multi-plane routing, alternating red/green traffic, physically motivated glass, and restrained emissive lighting.

Changes

• Add reusable emissiveMaterial and opticalPointLights shader modules and plugins, including matching WGSL/GLSL implementations and a fixed 16-light portable uniform layout.
• Extend reusable glass and reflective materials with optional nearby-light shading while preserving their existing helper signatures.
• Correct reflective environment sampling to follow the camera-reflected view ray and preserve the full opacity range for opaque metallic surfaces.
• Support configurable color attachment formats in A-buffer and weighted-blended transparency renderers so HDR values survive translucent compositing.
• Store exact A-buffer translucent fragments as two packed half-float pairs, preserving highlights above 1.0 through capture and resolve; update the 16-byte fragment allocation and bounded slice planning accordingly.
• Add an exported ACES tone-mapping shader pass and configurable linear samplers for named shader-pass render targets.
• Correct multiscale bloom filtering and avoid alpha-related blur amplification that previously produced square halos.
• Render small saturated red/green emissive packets, selective bloom, dynamic colored switch/link reflections, and muted metallic red/green endpoint servers.
• Preserve alternating packet traffic, a 4x4 server grid, independent network planes, optional orbit controls, hover picking, and WebGPU/WebGL fallbacks.
• Rename the example to Effects: Glass, display Network Packet Spraying as its visible subtitle, and add optional nonbreaking subtitle support to shared example headers.
• Reduce showcase A-buffer over-allocation, remove an unused shader-input allocation, and release metallic shader inputs during teardown.
• Document glass, transparency, reflective shading, HDR, selective bloom, tone mapping, and the MRC showcase.

Verification

• nvm use: unavailable in this shell; the active Node version is v22.22.1.
• yarn install: passed using env -u YARN_NO_PROXY corepack yarn install.
• yarn build: passed using env -u YARN_NO_PROXY corepack yarn build after the final code and formatting changes.
• yarn lint fix: passed using env -u YARN_NO_PROXY corepack yarn lint fix.
• yarn test-node: passed, 186 tests with two skipped, including regression coverage for camera-relative WGSL/GLSL reflections and opaque reflective materials.
• Focused Chromium GPU tests: passed, 27 tests covering actual translucent HDR A-buffer readback, WBOIT HDR output, bloom, tone mapping, and linear shader-pass target samplers.
• yarn examples:typecheck: passed for 32 example workspaces.
• Packet-spraying standalone production build: passed.
• yarn website:build: passed.
• (cd website && yarn build): passed.
• WebGPU and WebGL browser screenshots/canvas-pixel checks: passed with no page or console errors; hover explanations passed for servers, access switches, plane switches, and spine switches.
• GitHub Actions: all required checks passed on the final HDR-storage and showcase-title revision: test, examples, and website.
• yarn test: Node suite passed and the browser suite reported 1,281 passing, 25 skipped, and one existing unrelated failure in modules/arrow-layers/test/layers/arrow-layers.spec.ts (polygonViewportUniforms missing from the Arrow polygon WebGPU shader layout). The same Arrow failure was present before these changes.

Follow-up

• Resolve the pre-existing Arrow polygon WebGPU test failure independently of this optical-material/showcase change.
• More advanced depth-derived thickness, caustics, environment probes, and ray-traced transmission remain separate future work.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 9bba82bae4

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".