test: add mutation testing coverage for survived and untested mutants

[Copilot]

Adds targeted tests to kill mutants identified by mutation testing reports. All changes are test-only — no source code was modified.

Protocol & Models

• Direct tests for _decode_temperature, _check_length, _encode_temperature (mod-2 mutant), _decode_temp_unit, _decode_flame_effect (bitwise AND vs OR), _encode_heat_settings (boost boundary)
• kebab_name tests verifying lowercase + underscore-to-hyphen conversion

Auth

• Direct calls to _build_app() verifying each MSAL constructor kwarg individually (CLIENT_ID, AUTHORITY, validate_authority=False, token_cache)
• Direct calls to _save_cache() verifying mkdir(parents=True, exist_ok=True) and write_text content
• Logger format/arg verification via caplog

Client

• Parametrized _parse_fire_features tests covering all 24 feature fields individually (catches key-name and default-value mutations)
• get_fire_overview feature sourcing tests (FireDetails vs WifiFireOverview fallback)
• Field default tests, status boundary checks (199/200/299/300), write_parameters payload structure

CLI

• build_parser description/help text assertions
• Precise _display_* output verification (timer arithmetic // 60 / % 60, error hex/binary formatting, feature labels)
• _masked_input terminal handling (asterisks, backspace sequences, stdin.read(1) calls)

TUI Widgets

• Exact label string assertions for all _format_* functions (kills XX-prefix/suffix and case mutations)
• _build_fire_art structural verification (frame elements, flame chars, LED/media styles, heat wave chars)
• _expand_flame gap distribution logic

TUI App

• _resolve_version subprocess.run arg verification for all three git commands (describe, rev-parse, status) including kwargs (capture_output=True, text=True, timeout=2) and all branch paths

Remaining mutants

~3,400 mutants remain labeled "no tests" by mutmut due to its test selection algorithm failing to map integration-style tests to specific source functions. The code paths are exercised by existing tests but mutmut cannot associate them. Fixing this requires mutmut configuration changes or source modifications, not additional tests.

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[deviantintegral]

@copilot compare this PR against PR #78. Report on what the differences are, and suggest which one is the better version, if either.

[Copilot]

@copilot compare this PR against PR #78. Report on what the differences are, and suggest which one is the better version, if either.

Here's my comparison of this PR (#80) vs PR #78:

Size: PR #80 is larger (+3,114 lines across 7 files) vs PR #78 (+1,786/-190 across 9 files).

Overlapping coverage (both PRs target the same mutants):

• test_models.py — Both add kebab_name tests; PR test: kill mutants #78 adds 5 tests including an all_enum_members loop, PR feat: add --debug flag for HTTP request/response logging with credential redaction #80 adds 3 focused tests.
• test_protocol.py — Both add _decode_temperature, _check_length, and surviving mutant tests. PR test: kill mutants #78 adds +169 lines, PR feat: add --debug flag for HTTP request/response logging with credential redaction #80 adds +122 lines; similar scope.
• test_client.py — Both add _parse_fire_features, lifecycle, _request, turn_on/turn_off tests. Very similar coverage (~636 vs ~687 lines added). PR feat: add --debug flag for HTTP request/response logging with credential redaction #80 uses @pytest.mark.parametrize for 24 feature fields individually; PR test: kill mutants #78 uses a single loop assertion.
• test_cli_commands.py — PR feat: add --debug flag for HTTP request/response logging with credential redaction #80 is significantly larger (+1,155 vs +238/-83) with more granular label/help text checks.
• test_tui_actions.py — Both add _resolve_version subprocess.run arg verification. PR feat: add --debug flag for HTTP request/response logging with credential redaction #80 adds +137 vs +98.
• test_widgets_format.py — PR feat: add --debug flag for HTTP request/response logging with credential redaction #80 adds +718 lines (new classes for every _format_* function with exact label assertions); PR test: kill mutants #78 modifies existing tests (+109/-101) with tighter assertions.

Unique to PR #78:

• test_fireplace_visual.py (+491 lines, new file) — Dedicated tests for _expand_flame and _build_fire_art covering heat rows, styles, and flame geometry in a standalone file.
• test_b2c_login.py (+7) — Exact log record assertions.
• test_tui_screens.py (+2/-1) — Minor fix.

Unique to PR #80:

• test_auth.py (+275 lines) — TestBuildAppDirect, TestSaveCacheDirect, TestTokenAuthCallableMutant — direct unit tests calling _build_app() and _save_cache() without going through get_token().

Quality differences:

• PR test: kill mutants #78 is more surgical — it modifies existing tests to tighten assertions (e.g., replacing loose in caplog.text with exact record.msg checks) rather than adding entirely new test classes. This approach is cleaner since it doesn't duplicate test setup.
• PR feat: add --debug flag for HTTP request/response logging with credential redaction #80 tends to add new test classes alongside existing ones rather than tightening existing tests, which leads to more code but also more redundancy.
• PR test: kill mutants #78 creates a dedicated test_fireplace_visual.py for the fire art tests, which is better separation of concerns vs PR feat: add --debug flag for HTTP request/response logging with credential redaction #80 putting them in test_widgets_format.py.
• PR test: kill mutants #78 uses @pytest.mark.parametrize with itertools for the _parse_fire_features tests — elegant but PR feat: add --debug flag for HTTP request/response logging with credential redaction #80's explicit parametrize list is more readable.

CI status: PR #78 passes all CI checks. This PR (#80) also passes.

Recommendation: PR #78 is the better version. It's more concise (-1,328 fewer added lines), takes a surgical approach of tightening existing tests rather than duplicating setup, and covers test_fireplace_visual.py as a dedicated test file. PR #80's main advantage is the test_auth.py coverage (27 mutants) and more granular CLI help text tests, which could be cherry-picked into PR #78 if desired.