Files

T

a0d3d9c701 feat(editor): door/window placement-feel polish + placement-state refactor (#411 )

Round 8 of the opening-placement UX work. Make door/window placement feel
physical and predictable, and unify the validity/placement logic behind one
shared decision.

UX:
- Window default sill 0.5 m (DEFAULT_WINDOW_SILL_M) so fresh windows float
  slightly above the floor; existing windows keep their own sill.
- Dev-only floor "shadow" projection for windows during placement/move
  (footprint + dashed drop-line) so an elevated window's plan spot is legible.
- Move SFX: one soft grid-snap click per grid step — identical free-following
  over floor or sliding on a wall (keyed on the raw cursor; per-frame + step
  dedup), no separate snap cue (that was a "double"). Mirrored into the 2D
  floorplan-move so 2D and 3D match.
- Shift = force-place over a collision (commit allowed; ghost stays a red
  warning) + free-place (lands at the raw cursor but keeps the alignment guides
  visible). Tint flips green/red live when Shift is pressed/released stationary.
- On-wall preview is now the tinted ghost (green placeable / red colliding),
  matching the free-follow ghost, instead of a pale solid mesh + thin wireframe.
- R-flip fixes: always toggles (no initial no-op needing a second press),
  e.repeat filtered, ghost rebuilds with the live `side`, and the ghost's
  on-wall world yaw uses `itemRotation - wallAngle` so it faces exactly what
  commit places (cursorRotation was π off for the asymmetric ghost). R ownership
  follows the current pointer pane (capture-phase + stopImmediatePropagation in
  the 2D overlay) so 3D and 2D never double-flip or go dead.

Refactor / quality:
- New `resolveOpeningPlacement({collides,forcePlace}) -> {placeable,tint}` in
  shared/wall-attach-target.ts — the single source of truth the ghost tint AND
  the commit gates both consume, so they can't disagree under Shift.
- Consolidated the byte-identical `hasWallChildOverlap` into one shared impl
  (door-math/window-math re-export it).
- applyGhost gained a green "valid" tint.
- Removed dead `cursorRotation` from the move-tool targets after the yaw fix.

Docs: "2D <-> 3D behavioral parity" principle in wiki/architecture/tools.md
(+ README + AGENTS.md) — applicable behaviors must exist in both views.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

2026-06-16 10:20:44 -04:00

3.5 KiB

Raw Blame History

Agent Instructions — `pascalorg/editor`

Public, open-source home of @pascal-app/{core,viewer,editor,mcp} and the standalone editor app. Consumed both as npm packages and (in pascalorg/private-editor) as a git submodule.

Repo Shape

Path	Purpose
`packages/core`	Scene graph, node schemas, stores, event bus, core systems — pure logic, no Three.js
`packages/viewer`	Standalone 3D canvas: renderers, viewer systems, presentation state
`packages/editor`	Editor UI components reused by the standalone app and embedders
`packages/mcp`	MCP server and scene storage adapters
`apps/editor`	Standalone editor app — composes `viewer` + `editor` + tools

Where to look

Architecture rules — wiki/architecture/ (read on demand; index in wiki/architecture/README.md).
Skills (ready workflows) — .agents/skills/<name>/SKILL.md. Same content is reachable as .claude/skills/, .cursor/skills/, .codex/skills/ (symlinks to .agents/skills/).
Repo orientation for humans — README.md, SETUP.md, CONTRIBUTING.md.

CLAUDE.md, GEMINI.md, and .github/copilot-instructions.md are symlinks to this file. Codex reads this file directly.

Layer Boundaries (read once, internalise)

packages/core owns domain data and pure logic. It must not import Three.js, packages/viewer, apps/editor, rendering/UI concepts, tools, modes, phases, or view-specific concepts such as floorplan or paint preview.
packages/viewer owns the standalone 3D canvas, renderers, viewer systems, and genuine presentation state. It must not know about useEditor, editor tools, phases, modes, paint mode, floorplan state, or editor-only presentation vocabulary.
apps/editor owns the editing experience: tools, useEditor, panels, floorplan helpers, paint mode, keyboard shortcuts, command palette, action menus, cursor badges, and editor-only overlays. Editor features are injected into <Viewer> via props and children.

Details, examples, and rationale live in wiki/architecture/layers.md, wiki/architecture/viewer-isolation.md, wiki/architecture/systems.md, wiki/architecture/renderers.md, wiki/architecture/tools.md.

When making architecture-sensitive changes

Read the relevant page in wiki/architecture/ before writing code. The page list lives in wiki/architecture/README.md. As a minimum:

Adding a node type → node-schemas.md, renderers.md, systems.md
Adding a tool → tools.md, spatial-queries.md, events.md
Adding / changing a placement or move interaction → tools.md ("2D ↔ 3D behavioral parity": applicable behaviors must exist in both views; port the change to the sibling 2D/3D file in the same PR)
Adding a system → systems.md, scene-registry.md
Anything in packages/viewer → viewer-isolation.md, layers.md
Anything touching selection → selection-managers.md, scene-registry.md, events.md

When reviewing a PR

Invoke the review-architecture skill (.agents/skills/review-architecture/SKILL.md). It loads the required architecture pages, fetches the diff, classifies each new file by layer, and reports findings grouped by severity.

Operating rules

Read the full file before editing. Plan all changes, then make one complete edit.
When the user corrects you, stop and re-read their message.
After two consecutive tool failures, stop and change approach.
Don't introduce backwards-compatibility shims, dead code, or speculative abstractions.
Don't write new comments unless they explain a non-obvious why.

3.5 KiB Raw Blame History

Agent Instructions — pascalorg/editor