fix(testing): read test case json as utf-8

[Ghraven]

Problem

Burr's generated pytest test-case JSON path used the platform default encoding when appending fixtures from burr test-case create and when loading those fixtures through burr.testing.load_test_cases().

Those JSON files can contain serialized application state, prompts, responses, or user text. On systems whose default encoding is not UTF-8, fixtures with non-ASCII text can fail to load or be interpreted inconsistently.

Before / after

Before, the CLI writer and pytest helper opened test-case JSON files without specifying an encoding.

After, both the generated fixture writer and load_test_cases() read/write those JSON files with encoding="utf-8". A focused regression test writes a UTF-8 JSON fixture with literal non-ASCII text and verifies it loads with the expected test id and state payloads.

Verification

• python -m pytest tests/test_testing.py tests/integrations/test_burr_pydantic_future_annotations.py
• git diff --check