refactor(core): size and filter viewports by canvas

[ibgreen-openai]

Goal

Teach the existing ViewManager to lay out and select views by presentation canvas without introducing a second canvas-size tracking system.

Why this is needed

Deck.width and Deck.height describe the default canvas, but views assigned to different presentation canvases must use each canvas's own CSS-pixel dimensions. luma's existing CanvasContext / PresentationContext already owns resize observation, CSS dimensions, device-pixel ratio, and drawing-buffer dimensions.

This PR therefore adds only an optional internal getCanvasContext(canvasId) callback. ViewManager reads context.getCSSSize() when constructing viewports and controller viewports; it does not maintain a size map, size cache, resize observer, or duplicate drawing-buffer bookkeeping. The follow-up feature PR invalidates the existing viewport layout when luma reports a presentation-canvas resize.

Changes

• Resolve a view's canvas and read its authoritative CSS dimensions directly from the existing luma context.
• Fall back to existing ViewManager.width / height when no context lookup is supplied.
• Support both whole-canvas viewport queries ({canvasId}) and combined canvas/pixel queries ({x, y, canvasId}).
• Preserve first-canvas defaults, unknown-canvas filtering, unchanged viewport identity, and explicit resize invalidation.
• Add focused coverage for independent canvas dimensions, context-owned resize updates, whole-canvas filtering, and first-canvas fallback.

Review scope

Only modules/core/src/lib/view-manager.ts and its existing unit test file. No new public Deck props, canvas lifecycle changes, widgets, presentation code, or parallel size tracking.

Validation

• yarn test-headless test/modules/core/lib/view-manager.spec.ts test/modules/core/lib/deck.spec.ts --retry 1 --silent — 32 passed.
• Core TypeScript check passed.
• Repository lint, formatting, and 21 node smoke tests passed through the pre-commit hook.
• Downstream feature validation: 90 affected rendering, picking, widget, resize, and integration tests passed.

Stack

1. refactor(core): honor canvas contexts throughout render passes #10474: context-aware render/picking plumbing.
2. This PR: context-derived viewport sizing and canvas filtering.
3. RFC/POC: Extend View system to support multi-canvas rendering #10229: multi-canvas presentation and interaction.

[coveralls]

Coverage is 83.389% — codex/multi-canvas-view-layout into codex/multi-canvas-context-plumbing. No base build found for codex/multi-canvas-context-plumbing.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: b73e172efb

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".