Phase 5 Stage D: revert curve + whole-item move affordances to legacy

User-reported regressions made it clear that my D-ports of the curve
and whole-item move tools introduced more friction than they removed.
The legacy CurveFenceTool / CurveWallTool / MoveFenceTool /
MoveSlabTool / MoveCeilingTool ship more polish than the new ports do
right now:

  - Legacy curve tools pre-snap the pointer position to the 0.5m grid
    BEFORE projecting onto the chord normal, and they Shift-toggle to
    a free-place mode. The ports skipped both — finer math, but the
    user-visible UX regressed (laggy because the cascade resolver fires
    per move, history feels broken near the no-op threshold).
  - Legacy whole-item moves use scene.update per tick which keeps
    hosted children visually aligned. The live-drag mesh.position port
    cleared the offset before the GeometrySystem could rebuild,
    producing a one-frame teleport on commit.

Drop the affordance registrations for those tools — the ToolManager /
MoveTool dispatch falls back to the legacy per-kind tools when the
registry doesn't declare the affordance. Stage D progress preserved
for: fence move-endpoint (linked cascade + alt-detach), slab/ceiling
boundary + hole editors, fence placement, slab placement, ceiling
placement.

The kind-owned files (curve-tool.tsx, move-tool.tsx, actions/curve.ts,
actions/move.ts) stay on disk for the next iteration — when they
reach parity with legacy UX, re-add the affordance entries.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

packages/nodes/src/ceiling/definition.ts

@@ -58,13 +58,12 @@ export const ceilingDefinition: NodeDefinition<typeof CeilingNode> = {
   // with a vertical TSL-gradient connector + ground-shadow lines.
   tool: () => import('./tool'),
 
-  // Stage D: drag/edit affordances. Boundary editor + hole editor
-  // delegate to the shared `<PolygonEditor>`; move uses the single-
-  // undo dance.
+  // Stage D — boundary + hole editors ported. Whole-ceiling move kept
+  // on the legacy MoveCeilingTool for the same reason as slab (live-
+  // drag mesh.position one-frame teleport on commit).
   affordanceTools: {
     'boundary-edit': () => import('./boundary-editor'),
     'hole-edit': () => import('./hole-editor'),
-    move: () => import('./move-tool'),
   },
 
   renderer: {

packages/nodes/src/fence/definition.ts

@@ -74,26 +74,16 @@ export const fenceDefinition: NodeDefinition<typeof FenceNode> = {
   // Legacy `floorplanFenceEntries` short-circuits to [] when fence is
   // registered (see floorplan-panel.tsx).
   floorplan: buildFenceFloorplan,
-  // Stage D (in progress): drag-affordance components owned by the kind.
-  // ToolManager looks up these lazy modules at runtime when the matching
-  // editor state activates — no static import from editor → nodes
-  // (which would create a circular dep).
+  // Stage D — partial port. Endpoint drag (with linked-fence cascade,
+  // alt-detach, angle label) is ported here; curve + whole-fence move
+  // are intentionally kept on the legacy CurveFenceTool / MoveFenceTool
+  // because the legacy code is more polished than the ports were
+  // (cursor anchoring, snap step, performance, history). Those ports
+  // remain in `curve-tool.tsx` + `move-tool.tsx` + their `actions/`
+  // siblings for the next iteration; the legacy fallback runs until
+  // they reach parity.
   affordanceTools: {
-    // Triggered by useEditor.curvingFence. Pure DragAction logic in
-    // actions/curve.ts; this component is the React wrapper using
-    // useDragAction + the cursor visuals.
-    curve: () => import('./curve-tool'),
-    // Triggered by useEditor.movingFenceEndpoint. Pure logic in
-    // actions/move-endpoint.ts (linked-fence cascade, alt-detach,
-    // single-undo dance). Wrapper owns the angle label + detach badge.
     'move-endpoint': () => import('./move-endpoint-tool'),
-    // Triggered by useEditor.movingNode when the moving node is a
-    // fence. Whole-fence rigid translation with linked-fence cascade.
-    // Pure logic in actions/move.ts uses the live-drag exception
-    // (mesh.position + useLiveTransforms) to avoid rebuilding fence
-    // geometry every pointer tick; commits the final start/end with
-    // the single-undo dance.
-    move: () => import('./move-tool'),
   },
 
   toolHints: [

packages/nodes/src/slab/definition.ts

@@ -58,13 +58,15 @@ export const slabDefinition: NodeDefinition<typeof SlabNode> = {
   // with axis/45° snap (Shift to defeat).
   tool: () => import('./tool'),
 
-  // Stage D: drag/edit affordances. Boundary editor + hole editor
-  // delegate to the shared `<PolygonEditor>`; move uses the single-
-  // undo dance for the polygon-translate commit.
+  // Stage D — boundary + hole editors ported (thin <PolygonEditor>
+  // wrappers, behaviorally identical to legacy). Whole-slab move kept
+  // on the legacy MoveSlabTool: the live-drag mesh.position port
+  // introduced a one-frame teleport on commit (geometry rebuild lags
+  // the position clear), and the legacy already has acceptable perf
+  // via RAF-batched markDirty.
   affordanceTools: {
     'boundary-edit': () => import('./boundary-editor'),
     'hole-edit': () => import('./hole-editor'),
-    move: () => import('./move-tool'),
   },
 
   // Stage B: pure geometry function.

packages/nodes/src/wall/definition.ts

@@ -57,13 +57,14 @@ export const wallDefinition: NodeDefinition<typeof WallNode> = {
 
   parametrics: wallParametrics,
 
-  // Stage D (in progress): drag-affordance components owned by the kind.
-  // Only the curve affordance is ported today — the remaining wall
-  // tools (endpoint drag, whole-wall move, placement) are larger and
-  // queued for future sessions.
-  affordanceTools: {
-    curve: () => import('./curve-tool'),
-  },
+  // Stage D — deferred for wall. The curve port (`curve-tool.tsx` +
+  // `actions/curve.ts`) needs more work to match the legacy
+  // CurveWallTool's UX (pre-snap on pointer position, 0.5m grid step,
+  // Shift override, smooth scene.update without cascade overhead).
+  // Legacy fallback runs until that lands. Endpoint move / whole-wall
+  // move / placement are all still legacy too — they're the biggest
+  // tools and have linked-wall corner cascade logic that needs a
+  // careful port.
 
   renderer: {
     kind: 'parametric',