* Convert persistent notification tests to async
* Create/dismiss persistent notifications in exposed functions, not service calls
* Fix notify persistent_notification
* Remove setting up persistent_notification
* Drop more setups
* Empty methods
* Undeprecate sync methods because too big task
* Fix setup clearing notifications
* Fix a bunch of tests
* Fix more tests
* Uno mas
* Test persistent notification events
* Clean up stale comment
Co-authored-by: Martin Hjelmare <marhje52@gmail.com>
* Avoid creating sockets in homekit port available tests
* prevent new bridge from being setup -- its too fast now that the executor job is gone and it revealed an unpatched setup
- Sometimes homekit will go unresponsive because a pairing for a specific
device is missing. To avoid deleting the config entry and recreating
it, which can be a painful process if there are many bridged entities,
the homekit.unpair service allows forceful removal of the pairings so
the accessory can be paired again.
* Fix homekit checking for port cleanup too many times
The loop should have terminated as soon as the port was available
* coverage
* tweak homekit shutdown wait
If an entity changes between restart the iOS/controller
device may have cached the old chars for the accessory.
To force the iOS/controller to reload the chars, we
increment the config version when Home Assistant restarts
When we set up HomeKit, we asked users if they wanted
to create an entry in bridge or accessory mode.
This approach required the user to understand how HomeKit works and
choose which type to create.
When the user includes the media player or camera domains,
we exclude them from the bridge and create the additional entries
for each entity in accessory mode.
HomeKit pairings and accessory ids from versions 0.109 and earlier are no longer
migrated on upgrade. Users upgrading directly to 2021.3 from 0.109
and older should upgrade to 2021.2 first if they wish to preserve
HomeKit configuration and avoid re-pairing the bridge.
This change does not affect upgrades from 0.110 and later.
Safe mode was added to work around a race condition where
the mdns announcment was sent too early and would cause
pairing to fail. Since this has been corrected in
HAP-python, there is no longer a need to have
safe mode.
* Make async_get_device connections Optional, default None
* Remove unnecessary async_get_device connections arg usages
Some of these were using an incorrect collection type, which didn't
cause issues mostly just due to luck.
* Refactor zeroconf setup to be async
Most of the setup was calling back to async because
we were setting up listeners. Since we only need
to jump into the executor to create the zeroconf
instance, its much faster to setup in async.
In testing this cut the setup time in half
or better.
* partial revert to after_deps
J. Nick Koston