* Improve Plex debounce/throttle logic
* Use Debouncer helper, rewrite affected tests
* Mock storage so files aren't left behind
* Don't bother with wrapper method, store debouncer call during init
* Test cleanup from review
* Don't patch own code in tests
* solve modbus issue #33872
CONF_DELAY was used in a serial connection, which is not
permitted.
Sometimes async_update is called after async_setup is completed,
but before event EVENT_HOMEASSISTANT_START is issued, leading to
a missing object.
* resolve review comment.
Do not wait for start event, but activate pymodbus directly in
async setup.
* review 2
Remark, this does not work, async_setup hangs.
clean start_modbus() from async calls, leaving only the pymodbus
setup.
* review 2a
Moved listen_once back to start_modbus, since it is sync.
* fix bug when search for string in dict fails when dict is null
* another bug fix that I only noticed because of this other bug
* add test to cover failure scenario
* update docstring
* add additional assertions to cover failure scenario that's being fixed
* Add a config flow for flume
* Sensors no longer block Home Assistant startup
since the flume api can take > 60s to respond on
the first poll
* Update to 0.4.0 to resolve the blocking startup issue
* Missed conversion to FlumeAuth
* FlumeAuth can do i/o if the token is expired, wrap it
* workaround async_add_entities updating disabled entities
* Fix conflict
Sometimes a modbus server do not respond to a transaction,
this is a contradiction to the modbus protocol specification,
but merely a matter of fact.
Use asynio.await_for() to provoke a timeout, and close the
transaction.
Some providers have set their wait_timeout to 60s
in order to pack as many users as they can on a machine.
The mysql default is 28800 seconds (8 hours)
Since mysql connection build and tear down is relativity
expensive, we want to avoid being disconnected.
We now accommodate this scenario with the following:
1. Raise the mysql session wait_timeout 28800 when we connect
2. The event session now does a 30 second keep alive to
ensure the connection stays open
* Speed up TP-Link lights
* Color temp kan be None
* hs as int, force color temp=0
* Fix color temp?
* Additional tplink cleanups to reduce api calls
* Update test to return state, remove Throttle
* Fix state restore on off/on
* Fix lights without hue/sat
Co-authored-by: J. Nick Koston <nick@koston.org>
* Load integrations with requirements in device_automation
- Split cached loader behavior out of async_get_integration
- Use cached loader for both async_get_integration and
async_get_integration_with_requirements
- Use async_get_integration_with_requirements for device_automation
resolves#33104
* Duplicate caching logic in requirements, remove loader mods
* Update homeassistant/requirements.py
Co-authored-by: Paulus Schoutsen <paulus@home-assistant.io>
* Use asynctest patch instead of mock_coro
* Add test for supervisor ip ban
* Do not ban supervisor ip if set
* Extract supervisor ip helper
* Check supervisor ip before banning
* Remove added blank line
* Clean up get supervisor ip
Co-Authored-By: Pascal Vizeli <pvizeli@syshack.ch>
Co-authored-by: Pascal Vizeli <pvizeli@syshack.ch>
* Convert nut to a multi step config flow
* Users can now choose the ups they want in
step 2 (or skipped if only one)
* Users can now select the resources they want
to monitor based on what is actually available
on the device
* CONF_NAME has been removed as we now get the
name from NUT
* Device classes have been added which allows
the battery charge state to be seen in the
devices UI
* Remove update_interval as its for a followup PR
* explict
* reduce
* fix bug, add tests for options flow
* Test for dupe import
* Test for dupe import
* Call step directly
* Update homeassistant/components/nut/config_flow.py
Co-Authored-By: Martin Hjelmare <marhje52@gmail.com>
Co-authored-by: Martin Hjelmare <marhje52@gmail.com>
jjlawren