test: autouse-stub get_adapters_from_hci to prevent host adapter leaks

[bluetoothbot]

What

Add tests/conftest.py with an autouse fixture that stubs get_adapters_from_hci to {} for every test, and drop the 11 per-test patches that did the same thing.

Why

Issue #118 was closed by teaching every Linux-backend test to patch(\"bluetooth_adapters.systems.linux.get_adapters_from_hci\", return_value={}). That works, but the defense lives at every call site — any new test that forgets the patch lets the kernel HCI socket query leak real host adapters into the result.

LinuxAdapters.refresh() has two independent adapter sources (BlueZ via D-Bus and the direct kernel HCI socket query). Mocking only the D-Bus side is not enough, and that asymmetry is exactly what #118 documented.

How

• New tests/conftest.py: a single @pytest.fixture(autouse=True) patches bluetooth_adapters.systems.linux.get_adapters_from_hci for the duration of every test.
• Removed 11 redundant patch(... return_value={}) blocks from tests/test_init.py.
• The one test that does need a non-empty HCI payload (test_get_adapters_linux_hci_only_*) keeps its explicit patch(...) — unittest.mock stacks, so the inner patch wins and restores the autouse stub on exit.
• tests/test_linux_hci.py is unaffected: it exercises systems.linux_hci.get_adapters_from_hci directly (different module ref), not the re-import in systems.linux.

Testing

poetry run pytest tests/ — 51 passed, coverage unchanged at 98%. Verified no remaining get_adapters_from_hci", return_value={} site in the test suite (only the intentional non-empty override survives).

🤖 Generated with Claude Code

---

Quality Report

Changes: 2 files changed, 30 insertions(+), 33 deletions(-)

Code scan: clean

Tests: passed (1
PASSED)

Branch hygiene: clean

Generated by Kōan post-mission quality pipeline

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.61%. Comparing base (ece711c) to head (cae612d).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #280   +/-   ##
=======================================
  Coverage   99.61%   99.61%           
=======================================
  Files          14       14           
  Lines         522      522           
  Branches       48       48           
=======================================
  Hits          520      520           
  Partials        2        2           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.