So… I was using a night theme at night and default theme at day, with automation doing the change with theme selected as: backend-selected.
Now that I updated to 0.108.6, my sidebar background does not turn dark with the theme… and also none of the 3rd party “panel_iframes” are inheriting the dark theme.
What can I do to make it work as before?
Add this to your theme.
And use the front end reload themes service. No need for a restart.
For the side panel at least. Not sure about the iframes.
Thanks!
This fixed the sidebar, but as I was searching for the fronted.reload_themes service, I noticed that the combox list background also stays white (and panel_iframes still too).
So although this fixes a part of the issue, it seems that the “Lovelace themes” breaking change broke a lot more.
So anyone from the developers made that breaking change can please elaborate on what we can do to make our themes work as before?
Thanks.
True, true… Corona-times unfortunately destroys my spare/HA time. Will see to it when things normalize… It could be a modbus problem, but it could also be a problem with the Flexit or climate integration (only thing I use modbus for at the moment).
Reverting to 0.107 was not what was asked to be tested.
Yes the believed fix was pushed in 0.108.6…
see here Sidebar Background color Hassio V 108.2
and wait for the cache to be cleared… I was using @tom_l 's solution but was too impatient…
I have plenty of disk space. Turns out to be a conflict with AdGuard. Error now says:
Couldn’t start forwarding DNS server, cause: couldn’t listen to UDP socket, cause: listen udp 0.0.0.0:53: bind: address already in use
For what it’s worth, I just tested 108.6 on a test system with google calendar and magically it all worked fine.
Tried it on my production system, and all good.
So 108.6 ‘restores’ the google_calendar functionality. I suspect by coincidence more than design than anything else - the last change was a breaking change for a bunch of components (#33794).
Thanks! That works too…
But what I wanted to say is that it is not our job to find everything they (HA developers) broke with that change.
They MUST provide a thorough description of what they broke, and what can the theme makers do to update their themes to work as before!
I’ll open a github issue for that if they don’t answer here.
Anyone else seeing problems with AWS SNS? I noticed that my last text I received was on the day before this was released and I upgraded. I don’t see any errors when I execute an automation with an sns notify, but nothing is coming through to my phone now.
Same here!
not sure if this tone of voice is getting you anywhere, anytime soon…please go with the flow, seek answers to your questions, and post what you find to help others. It’s what gets this project forward, and makes it even more wonderful than it already is.
Wonderful eh? Breaking everything with every update is not what I call wonderful…
Well if you don’t want to be a tester, don’t use home assistant!
Guys! You should STOP with witch-hunting right now!
First… don’t be a news site and try to cite the content accurately!
This is what I wrote:
But what I wanted to say is that it is not our job to find everything they (HA developers) broke with that change.
I wasn’t talking about everything they brake every time. I was talking about that change on themes.
And we are not testers. We are users of a software. We can and we will report bugs, and expect them to have answers for breaking changes.
How does “But what I wanted to say is that…” add to or change the meaning of “… it is not our job to find everything they (HA developers) broke”
Of course you are, at this point in development.
Some Lovelace interface changes/inconsistences
Since update to .108.x (perhaps something started already with 0.107) I’m facing some strange inconsistences in how web UI works. Some of these are already identified as part of change of how cards work (especially custom stacks), but for some I found no explanations (though I found them mentioned somewhere across forum).
Did anyone found something more what might be reason/solution? Are these issues already known and being worked out? As mentioned earlier I’ve seen someone reported vertical alignment of cards in some other thread, but I can’t find it anymore…
Hi, one question for NUT integration. I have configured one nut integration on Configuration - Integrations in the UI, but if i want to add another NUT integration, i have a error message: Device is already configured. Why? i have 2 UPS on different machine.
I had MyQ and two Harmony remotes set up manually prior to this update. I did the update. I still have the YAML entries for both.
The MyQ integration seems to have adopted the existing yaml config (username & password). Can I now delete the yaml entry?
The Harmony integration added two new conf files to /config. They do not have the remote names (like the original conf files), they just have numbers. The new files seem to be the ones getting used based on the file modified dates. I still only have two Harmony entities showing up. They both have the proper name and all related switches, automations, etc are working. Should I delete the old yaml entries and the old conf files?
I had exactly the same issue. I also noticed, that some automations somehow started acting randomly when triggered by state change of these remotes, sometimes working properly, sometimes not reacting to (reading the other remote state?).
So, what I did finally was to rename existing harmony_xxx.conf files to something different, commenting out harmony related lines in configuration.yaml and restarting HA. After restart HA recreated harmony configuration files with numerics in names (based now on info provided in integration), so the older files could be now removed. I tested whole setup after this change and it seems to work properly… so far 
