fix: use utf-8 for test case json files

[Ghraven]

Summary

• read and write generated test-case JSON files with explicit UTF-8 encoding
• read dynamic pytest test-case JSON fixtures with the same encoding

Before / after

Before, the CLI test-case generator and load_test_cases() relied on the platform default text encoding for JSON fixture files.
After, both paths use UTF-8 explicitly, making generated fixtures portable across non-UTF-8 locales.

Verification

• python -m py_compile burr/cli/__main__.py burr/testing/__init__.py
• git diff --check

[andreahlert]

LGTM following #790