Write state if schedule update state from async context #31758
Conversation
|
Seems we should drop the force feature from the function and replace all references of force=false with just a write call? The force naming is a bit confusing anyway. |
|
That would be a too big of a breaking change for custom components I think. |
|
Suppose it would have to be done by marking it deprecated, and just having it call two new functions. But sure, probably not worth the hassle as a cleanup action. |
|
@SukramJ I am running into a failed test with Homematic IP cloud: This PR makes calls to |
|
The reason the other tests were failing is that we call 1 async function less. That 1 async function was waited for when we call |
| if force_refresh: | ||
| self.hass.async_create_task(self.async_update_ha_state(force_refresh)) | ||
| else: | ||
| self.async_write_ha_state() |
There was a problem hiding this comment.
The name of async_schedule_update_ha_state doesn't match what happens here now.
There was a problem hiding this comment.
True. But the intention of the method is still true: write the state to HA without a guarantee that it's available after the function call is done. I'll update the doc string
There was a problem hiding this comment.
We're now removing the choice from the developer if the state should be updated in this iteration or a later iteration of the event loop. That might be ok. 🤷♂️
There was a problem hiding this comment.
I don't think that's not something for the developer to decide. The developer can only indicate that it has new data and wants to be updated. I can't think of a reason for a developer wanting to let us know, but not write it to the state machine as soon as possible.
|
|
||
| @callback |
There was a problem hiding this comment.
update_list is not async safe. We're loading games which loads a json file.
|
Hi @balloob, BR |
|
I'm a bit confused why it needs to be forced now . So by writing a state faster to the state machine than before we can end up requiring to do a full update? Was there a race condition and the full update was always needed? |
|
Yes you are right |
|
10 Minutes too fast. I create a new PR to avoid the |
Breaking change
Proposed change
If we're in async land scheduling a state update and we know we're not going to trigger an update, write the state directly.
Type of change
Example entry for
configuration.yaml:# Example configuration.yamlAdditional information
Checklist
black --fast homeassistant tests)If user exposed functionality or configuration variables are added/changed:
If the code communicates with devices, web services, or third-party tools:
Updated and included derived files by running:
python3 -m script.hassfest.requirements_all.txt.Updated by running
python3 -m script.gen_requirements_all..coveragerc.The integration reached or maintains the following Integration Quality Scale: