fix(coordinator): guarantee setup event is set to prevent deadlock on failed initialization

[firstof9]

Summary

This PR resolves #668.

When the Keymaster coordinator fails to set up (due to connection failures or missing dependencies during Home Assistant startup), the _initial_setup_done_event is never set. Consequently, when a user tries to delete or reconfigure the integration entry afterwards, any downstream calls waiting on this event (await self._initial_setup_done_event.wait()) hang indefinitely. This blocks the config entry unloading task, causing Home Assistant Core to hang and eventually be terminated by the watchdog (SIGKILL / Exit 256).

Proposed change

Wrapping the core setup in a try...finally block ensures that _initial_setup_done_event is always set even if setup fails, allowing entry cleanup/deletion or reload to proceed normally.

Type of change

• Dependency upgrade
• Bugfix (non-breaking change which fixes an issue)
• New feature (which adds functionality)
• Breaking change (fix/feature causing existing functionality to break)
• Code quality improvements to existing code or addition of tests

Additional information

• This PR fixes or closes issue: fixes ISSUE: Home Assistant Core crashes when reconfiguring or deleting Keymaster integration entry #668
• This PR is related to issue:

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 92.23%. Comparing base (cdb4922) to head (e4b8c3a).
⚠️ Report is 186 commits behind head on main.
❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #669      +/-   ##
==========================================
+ Coverage   84.14%   92.23%   +8.09%     
==========================================
  Files          10       41      +31     
  Lines         801     4817    +4016     
  Branches        0       30      +30     
==========================================
+ Hits          674     4443    +3769     
- Misses        127      374     +247     

Flag	Coverage Δ
python	92.09% <100.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 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.

[tykeal]

Review: fix(coordinator): guarantee setup event is set to prevent deadlock

Correctly root-causes #668. _async_setup only set _initial_setup_done_event on its last line (coordinator.py:158-167); if any prior step raised, the event was never set, and every mutating method waits on it. On delete, async_remove_entry → delete_lock_by_config_entry_id awaits the never-set event and hangs, blocking teardown until the watchdog SIGKILLs Core — matching both repro paths in #668. The try/finally fix is right, and the delete_coordinator data is None guard is a necessary secondary fix (it would TypeError on len(None) right after the deadlock is lifted). CI is green.

Non-blocking items:

• [SUGGESTION] After a failed _async_setup, the exception propagates from initial_setup() to async_setup_entry:141 as a raw exception, so HA errors the entry without retry, while the coordinator stays in hass.data[DOMAIN][COORDINATOR] with kmlocks == {} and the event set. A later reconfigure/reload sees COORDINATOR present (__init__.py:138), reuses it, and skips initial_setup() — locks never reload, integration is silently non-functional until a full HA restart. Wrap the initial_setup() failure as ConfigEntryNotReady so HA retries and re-runs setup. Not required to close #668.

• [SUGGESTION] delete_coordinator (__init__.py:303): the new data is None branch is untested. Codecov's "100% modified lines" is line-level, not branch-level. Add a test asserting delete_coordinator removes the coordinator when coordinator.data is None and no other entries remain.

• [SUGGESTION] test_async_setup_success asserts call counts but not ordering. The invariant #668 depends on is that the event is set after the setup steps and _verify_lock_configuration runs after the event on success. Add an ordering assertion.

• [NITPICK] Add a one-line comment on the finally documenting why the event is set unconditionally, so a future refactor doesn't reintroduce the deadlock.

• [NITPICK] test_async_setup_exception_sets_event only raises from _async_load_data (first step). A parametrized variant raising from _rebuild_lock_relationships / _setup_timers would protect the finally guarantee for mid-block failures.

No blockers. Approve-with-nits once the above are considered.