`execute_test.py` pollutes test output with error traces (MacOS)

Follow-up issue of https://github.com/tiny-pilot/tinypilot/issues/1642.

We were able to silence the undesired stacktrace printing via https://github.com/tiny-pilot/tinypilot/pull/1708 – this, however, only appears to work on Linux systems and not on MacOS, because on MacOS, the multiprocessing package behaves differently.

We weren’t able to find a suitable workaround for MacOS, except if we’d account for this issue inside the underlying execute.py module (e.g. by trying to silence stderr there, with a /dev/null redirect). That is probably not a good trade-off, because we’d potentially be risking to change production behaviour only to fix a minor, local testing issue. There might be other approaches to solve this problem, though.

Start with app/execute_test.py, especially the multiprocessing-related lines linked in the issue, then compare its behavior with app/execute.py on MacOS and Linux. Reproduce the unwanted error traces and investigate approaches that silence test-only output without changing production behavior; done means the MacOS test output is clean while Linux behavior remains intact.