[KennyMcSimpson]

On non-UTF-8 locales, jsondiff and jsonpatch read UTF-8 JSON files using
the system encoding, which can silently change Unicode values. In-place
jsonpatch -i -u can also raise UnicodeEncodeError when writing characters
outside that encoding, even if the input files contain only ASCII escapes.

Explicitly use UTF-8 for JSON file arguments and both in-place output paths,
and document the file encoding. Standard input and standard output continue
to use their existing stream encodings.

The regression tests disable Python's UTF-8 mode and locale coercion. They
cover both programs' UTF-8 inputs and Unicode in-place output through the
atomic and fallback paths. All four cases fail against the original source
on a GBK Windows system.

Validation: Python 3.10 and 3.13 full unit suites each pass 217 tests with one
POSIX-only skip; all 20 Python 3.13 property tests pass. Critical Flake8
checks and git diff --check pass.

AI assistance: OpenAI Codex prepared the change, regression tests, and this description.

[Copilot]

🟢 Approval recommended

The focused changes cover all intended file I/O paths with targeted regression tests and no identified blocking issues.

No open issues 🎉

What changed in this PR

Makes JSON command-line file I/O consistently use UTF-8, preventing locale-dependent Unicode corruption while preserving existing standard input and output encodings.

Changes:

• Explicitly uses UTF-8 for file arguments and both in-place output paths.
• Adds regression tests for UTF-8 inputs and Unicode output with Python’s UTF-8 mode disabled.
• Documents the file-encoding behavior.

File	Description
tests.py	Adds encoding regression coverage and updates subprocess and output-mocking helpers.
jsonpatch_cli.py	Sets UTF-8 explicitly for JSON file reads and in-place writes.
doc/​commandline.rst	Documents UTF-8 file input and in-place output.

_{Review effort: Balanced}

[coveralls]

coverage: 100.0%. remained the same — KennyMcSimpson:fix/cli-utf8-json-files into stefankoegl:master