The patch was still too late in #115442
There is no good candidate to patch here since the late operation
is the error log that is being tested.
Patching the logger did not seem like a good idea so I went with
patching to wait for the error to be emitted since emit is the public
API of the log handler and was less likely to change
* add re-auth-flow to fyta integration
* add strings for reauth-flow
* resolve typing error
* update based on review comments
* Update homeassistant/components/fyta/config_flow.py
Co-authored-by: G Johansson <goran.johansson@shiftit.se>
* add async_auth
* adjustment based on review commet
* Update test_config_flow.py
* remove credentials
* Update homeassistant/components/fyta/config_flow.py
Co-authored-by: G Johansson <goran.johansson@shiftit.se>
* Update tests/components/fyta/test_config_flow.py
Co-authored-by: G Johansson <goran.johansson@shiftit.se>
* Update tests/components/fyta/test_config_flow.py
Co-authored-by: G Johansson <goran.johansson@shiftit.se>
* Update conftest.py
* Update test_config_flow.py
* Aktualisieren von conftest.py
* Update test_config_flow.py
---------
Co-authored-by: G Johansson <goran.johansson@shiftit.se>
* Ignore coverage for aiohttp_resolver backport helper
* Adjust generate to sort core items
* Adjust validate to sort core items
* Split line
* Apply suggestion
Co-authored-by: Dave T <17680170+davet2001@users.noreply.github.com>
* Fix suggestion
---------
Co-authored-by: Dave T <17680170+davet2001@users.noreply.github.com>
When I turned on eager_start here the data would always end up being
None because _async_update_data always returned None. To fix this
it now returns the value from the push loop. It appears this race
would happen in production so this may be a bugfix but since
I do not use this integration it could use a second set of eyes
* Use enum device class in Netatmo wind direction
* Use enum device class in Netatmo wind direction
---------
Co-authored-by: J. Nick Koston <nick@koston.org>
ConfigEntries.async_initialize was trigger asyncio warnings because of the CPU time
to call __setattr__ for every variable for each ConfigEntry being loaded at startup
* Bump aiohttp to 3.9.4
This is rc0 for now but will be updated when the full release it out
* cleanup cruft
* regen
* fix tests (these changes are fine)
* chunk size is too small to read since boundry is now enforced
* chunk size is too small to read since boundry is now enforced
* Add notify entity component
* Device classes, restore state, icons
* Add icons file
* Add tests for kitchen_sink
* Remove notify from no_entity_platforms in hassfest icons, translation link
* ruff
* Remove `data` feature
* Only message support
* Complete initial device classes
* mypy pylint
* Remove device_class implementation
* format
* Follow up comments
* Remove _attr_supported_features
* Use setup_test_component_platform
* User helper at other places
* last comment
* Add entry unload test and non async test
* Avoid default mutable object in constructor
_async_attach_triggers was writing state, async_enable was
writing state, and all of them were called async_added_to_hass
After entity calls async_added_to_hass via add_to_platform_finish
it will also write state so there were some paths that
did it 3x
async_disable was also writing state when the entity was
removed
Sid
J. Nick Koston
Jan Bouwhuis
Bram Kragten
dontinelli
karwosts
Brett Adams
epenet
Robert Resch
Marc Mueller
Stefan Agner
Joost Lekkerkerker
Allen Porter
Luke Lashley
Santobert
TheJulianJES
Jessica Smith
Steven B
Mike Degatano