* Switch async_track_same_state to use async_call_later
There was no need to use async_track_point_in_utc_time here since
we only need a delay
* update trigger tests
* remove some more utcnow patching
* remove some more utcnow patching
* remove some more utcnow patching
* Allow passing an optional name to async_track_time_interval
This is the same idea as passing a name to asyncio.create_task which
makes it easier to track down bugs
* more
* short
* still cannot find it
* add a few more
* test
When a template render raised an exception we would start listening for
all states until the template did not raise an exception anymore. This
was not needed since the entity that is causing the exception was
already in the tracker.
Re-rendering on all state changes can be extremely expensive and can
bring an instance into a sluggish or unresponsive state when updating
from a much older version that did not raise ValueError when a default
was missing.
* Improve event filters to reject earlier
- Avoid running the callbacks for state added/removed from
a domain if there are no listeners that care about the domain
- Remove some impossible checks in the listeners that will
never match since they were already rejected by the filter
* leave one guard since there is a race when we return control via await
* Fix high latency at 0 microseconds
fixes#82231
* fix async_track_utc_time_change alignment
* use replace to preserve fold
* naming
* tweak
* make async_fire_time_changed aware of the thundering heard issue
* Reduce overhead to refire events
- asyncio timers can fire early for a varity of reasons including
poor clock resolution and performance. To solve this problem
we re-arm async_track_point_in_utc_time and try again later
when this happens.
- On some platforms this means the async_track_point_in_utc_time can
end up trying many times to prevent firing the timer early since as
soon as it rearms it fires again and this repeats until we reach
the appointed time. While there is not much we can do to prevent
asyncio from firing the timer callback early, we can reduce the
overhead when this happens by using avoiding creating datetime
objects
* tweak mocking
* -vvv
* fix time freeze being too broad in litterrobot
* adjust
J. Nick Koston
Erik Montnemery
Marc Mueller
epenet
rlippmann
krahabb
Franck Nijhof
Michaël Arnauts
HarvsG