Yes, look at my example, then what you wrote. You’ll notice they don’t match. Hint: You’re missing characters.
Ah yes…thanks
I also replied on your github issue, if you don’t understand the fixes make a forum post and tag me.
Again, thanks for your help. I just tried the expand('group.echos')
in my config ( then rebooted HA). The template works, but still get the same log entry warning…Template variable warning: No first item, sequence was empty. when rendering
.
Then it’s coming from the | first
filter. You can probably safely ignore that warning without any repercussion. Otherwise you’ll have to check if there’s items in the list before collecting them.
I believe a fix is in the works, best to keep an eye on this thread -
Is everybody using MySQL (or maybe Maria DB) for the database? I thought since the whole Oracle thing a lot of people moved to Postgres.
When I read about something like 7 hours for a 1.x GB database I got really afraid. My database ist 54 GB (on a slow server with a Skylake Pentium but at least on a SSD) and I thougth “well one week without HA or a database reset”. But it looks like that (this time) the migration was something that was easy for PostgreSQL to do. Took only a few minutes.
Not sure how long it will take.
I had a mySQL version lower then 5.7, which turned out to take forever…(killed it after 17hrs)
well, tbh, I am not sure this is about the new template rendering, but Ive never seen this before, so please forgive me if this should go elsewhere:
Suddenly a few of my calibrated temp sensors (Hue core + an input_number) have huge precision, while the individual sensor and number have only 2 decimals. I know I can round(2) these, but should I really, or is this a bug:
{{states('sensor.driveway_buiten_sensor_temperature')|float +
states('input_number.temp_calibration')|float}}
{{states('sensor.driveway_buiten_sensor_temperature')|float}}
{{states('input_number.temp_calibration')|float}}
I believe what you are seeing there is a result of what is known as Machine Epsilon.
Someone else had encountered it and reported it as an Issue. It was also due to machine epsilon.
“input_number.increment” uses wrong step size · Issue #48770 · home-assistant/core · GitHub
I’m getting ‘The configuration check did not pass!’ with the same error prior to the upgrade when I use the check_config addon.
Not sure what I need to fix before upgrading.
Am currently running core-2021.3.4 / supervisor-2021.03.9 / HassOS 5.3
thanks Taras, I’ve added to that issue. As said there, because of the floats, I apparently am forced to use the |round(2)
(dont convert it to a string, because I need to build further logic on these numbers elsewhere in the system. I don’t yet understand when this happens…Only 2 or 3 out of 23 templates show this behavior (for as far as I have noticed)
I saw your posts and already more than one issue in GitHub on this.
Is the solved in the meantime on .4.4 or still open? I’m not able to update to 2021.4.x without being still able to edit on iOS devices.
It is and was already a pain that I was and am only able to copy&paste single yaml lines in iOS App, even if I select several ones, but not being able to edit at all is a showstopper.
today i wake up with 11gb of log, i’m tryingg to see what is inside.
I can see states being persisted in the recorder DB even though there are no changes.
This started with 2021.4 I think.
These are MQTT messages coming from Tasmota telemetry data published every 2 minutes.
A release for advanced users that needed zero advanced steps from me. Thank you for a great release.
My DB upgrade went flawlessly and took only 42s for a 1.32GB Postgres DB. This is 7 days’ history tracking about 400 states. I almost thought something was wrong!
This is a dual core 1.8GHz Intel Celeron system with 2GB memory running HA core in a Python venv.
Breaking changes for Broadlink custom component for the S1c alarm kit? Tried several times but it failed to work.
Rolling back to 2021.3.4 for now until I have more time to look at the issue.
Unless it will be fixed in the next release?
Problem is not fixed.
And I do not see any dev being assigned yet. I would be less concerned if someone had taken ownership of it - even if the forecast for a fix is in a month
This issue affects me as well on an iPhone 11Xs with 14.4.2. Bram checked in something CodeMirror related yesterday, for what it’s worth: