#KNX changes
1 messages · Page 1 of 1 (latest)
Hi 👋!
I did not really think that far. Currently there is nothing in a UI entity that can't be done in yaml - and vice versa.
I think it is way faster to create entities in bulk in yaml, so don't really see a need to deprecate that. Let's see what time brings.
@last moss Regarding the breaking changes of hvac modes... which key exactly was that? I'm not aware of any breaking change here...
I think there was somethig with aux_heat or so with climate and breaking. My entities doesn't have that and are working fine.
All things seem to work as before it looks at first glance, I'll post here if something is fishy over the day.
Oh, maybe it's the operation_modes: config. I wanted to remove that anyway... let's see maybe it can be done in beta - imho better than a breaking change.
And controller_modes:
Fan only now has to be fan_only and dry is now dehumidification I think
I have 216 entities in my KNX Config. Converting those to UI and then changing the entities in the visualizations and automations manually would be a royal pain in the b***
Ah another thing I’ve noticed when I included the KNX Integration into my test installation it did not give me the „Automatic“ Interface option. Manually setting an interface worked fine.
That points to a multicast issue in your test installation. Docker?
I think it would be best to have a migration once we support all platforms. Then one could just wipe the entire yaml instead of just parts.
No but Hyper V with that virtual switch thing, so multicast probably won’t work.
Maybe a routine which converts each yaml entry and on success comments it out so that if things go pearshaped there is a simple way to recover
There is no way to edit configuration.yaml programmatically. At least none that would be allowed in Core ... or preserve my sanity 🙃
Oh then this is of course more complicated because you would have to integrate some level of protection against duplicates
Usually such migrations work like
- migrate all yaml
- set a flag in the config entry that this was done
- don't load any yaml anymore when this flag is set
- open repair issue for the user to delete the yaml
That sounds like a smart way to do it. 👍 I have played around with the UI a bit more. I have found a little detail which makes it hard to use for me. When in the configuration for lets say creating a switch. You will see the KNX devices on right with Manufacturer and official name of the device. In a classical KNX installation you might have the same device a couple of times. I have the same lightswitch and the same blinds actor like 10 times each. So I would have to go by physical adress to distinguish them. I have used the description field in the ETS to describe where each is located. It would be cool if one could show the description as well. So there would be something like "Gira Giersiepen" "Tastsensor 3" "Kitchen"
Ah I see. Hm... most projects I have seen use the name field for this, but description may make sense too.
What xknxproject version was used to parse your project (you can see that in the info panel)?
There has been a change in name parsing in 3.7.0, but I don't remember what exactly it did.
It is 3.7.1
Alright. Will have to look how this can be done since description is a multiline string...
Would you mind creating an issue at the Knx-frontend repo so I don't forget 😬
Done. Maybe you could just show one line and then show the rest when hovering with the mouse over the line
Yeah let's see. I think ETS prepends the name with the first line in its topology view
Yes it does, first physical adress, then 1. line of description (tried adding second line to description, wasn´t shown), then name
this should fix it before the release hits https://github.com/home-assistant/core/pull/123043
You should see the description now too, with latest beta.
Wow thank you! This is so much better!