test: characterize inspector-vs-block placement semantics in translation tests

[tracygardner]

Motivation

• Lock in the current inspector-vs-block placement semantics before performing internal transform/origin refactors so future changes can be validated against the present behavior.
• Capture the existing (known-quirky) relationship between requested Y, mesh pivot/base and mesh.position.y so future work can be non‑destructive and well‑tested.
• Provide explicit coverage for both useY: true and useY: false code paths to avoid regressions when changing translation logic.

Description

• Added a new test suite Current inspector-vs-block placement semantics @translation in tests/transform.translate.test.js that creates a box and sets explicit pivot settings via flock.setAnchor.
• Added a test asserting that flock.positionAt(..., useY: true) aligns the base/pivot (boundingBox.minimumWorld.y and getAbsolutePivotPoint()) to the requested Y while mesh.position.y remains offset.
• Added a test asserting that calling flock.positionAt(..., useY: false) preserves the existing Y/base offset semantics while still updating X and Z.

Testing

• Ran the translation test target with npm run test:api translation as an automated smoke run to validate the added tests.
• The runner successfully started the dev server but the Playwright browser launch failed because Playwright browser binaries are not installed in this environment (Executable doesn't exist ... chrome-headless-shell).
• No test assertions failed; the automated run was blocked by the environment limitation described above rather than by test failures.

---

Codex Task