test: compare resolved paths where the product resolves

[Hasnain2430]

Related issue

None.

Summary

Two test assertions compared a product-returned path against an unresolved TemporaryDirectory path. The product resolves deliberately at both boundaries, so the comparisons hold only when no path component is rewritten during resolution:

• verified_local_path in src/vidxp/artifact_delivery.py calls path.resolve(strict=True) before an artifact path reaches the caller.
• _validated_marketplace_root in src/vidxp/codex_plugin.py calls .expanduser().resolve() before the Codex marketplace root is used.

On Windows, TMP commonly contains an 8.3 short component when the user profile name contains a space. Resolution expands it, and both assertions fail on the short-versus-long form of the same path:

AssertionError: WindowsPath('C:/Users/Hasnain Ibrar Butt/AppData/Local/Temp/tmpzyyjwosm/clip.mp4')
            != WindowsPath('C:/Users/HASNAI~1/AppData/Local/Temp/tmpzyyjwosm/clip.mp4')

AssertionError: assert 'C:\Users\HASNAI~1\...\.agents\plugins\marketplace.json'
             == 'C:\Users\Hasnain Ibrar Butt\...\.agents\plugins\marketplace.json'

The change compares the resolved form on both sides, which is what each boundary documents and returns.

Internal-only, test scope. No product behavior changes: both delivered paths were already correct and are still asserted, now in the form the boundaries actually return.

Why CI does not catch these

• The full suite runs on ubuntu-latest, and the windows-2025 provider job runs only test_models.py and test_indexing.py, so test_mcp.py is never executed on Windows.
• test_codex_plugin.py is never collected by CI at all. It defines module-level pytest functions, and CI discovers tests with python -m unittest discover -s tests, which collects only unittest.TestCase subclasses. test_native_ingestion.py (4 tests) is in the same position. That is a separate gap and I have not touched it here; happy to open an issue if it is not already known.

Validation

Windows 11, Python 3.12.14 in the project venv, TMP=C:\Users\HASNAI~1\AppData\Local\Temp.

• uv run --no-sync python -m pytest -q tests/test_codex_plugin.py tests/test_mcp.py — 60 passed, 7 subtests passed. Both assertions fail on upstream/main with the errors quoted above, so this exercises the real failure rather than a mock.
• uv run --no-sync python -m pytest -q on upstream/main before the change — 4 failed, 731 passed, 5 skipped. Two failures are the assertions fixed here.
• uvx "ruff~=0.16.1" check src tests — all checks passed.

The other two pre-existing failures on upstream/main are test_native_ingestion.py::test_local_ingestion_automatically_indexes_and_becomes_searchable and ::test_streamable_http_browser_upload_indexes_in_one_session, both AssertionError: Autonomous indexing did not finish without status polling. They are unrelated to this change and I have not prepared model artifacts locally, so I cannot tell whether they are environmental or a real defect. Flagging rather than claiming.

[Hasnain2430]

Adding a complete validation run, now that I have the scene model artifacts prepared locally.

Full suite on this branch, Windows 11, Python 3.12.14, scene models prepared:

uv run --no-sync python -m pytest -q
735 passed, 5 skipped, 106 subtests passed in 458.85s (0:07:38)

No failures. That covers both assertions changed here, and it includes the 85 tests in the eight files that CI's unittest discovery does not collect (reported separately as #156), so this is a wider run than the branch would otherwise receive.

For completeness on the two unrelated failures I flagged in the description: they were environmental, not defects. test_native_ingestion.py passes once scene models are prepared (4 passed in 104.31s), and I have withdrawn that part of #156. Nothing in this branch depends on it.