I asked the same question a while ago. As OH is now using Vue, why not use Vuex, or Pinia to store state?
The rules run on the server side and not in the frontend, so Vuex isn’t helpful IMHO.
Rich, Vue is definitely used in the MainUI (I know for sure)
To be more precise: It is not a real problem with rendering but a problem with some widgets like the thermostat widget from the market place which are designed a way that they need more space then available on a quite small phone display, leading to overlapping elements, elements partly outside of sight and text just cut off (especially when two cells are displayed in the same line and get too small for the full text). You can optimize everything for the phone in a way that there is never more then one cell/widget in a line, but such pages get very long on the PC screen. Main UI is nice but quite consuming in space on the screen and it is not easy to find the right compromise using it on the big screen and the mobile.
It would be cool and great if Thing actions were directly available from the GUI in the Thing page.
Explain more what you mean by this. That the Actions are listed somehow or documented somehow or that you want to be able to trigger/activate an Action from the Things page?
If the latter, wouldn’t it make more sense to be able to directly call Actions from UI rules directly without needing to create a Script Action? I’m not sure why someone would want to trigger an action from the Things page except for testing.
I also wish support for Matter.
I would like it to work in both ways.
First to use Matter things in OpeHAB, but i also that OpenHAB is an Matter Bridge with possibility to publish legacy things to Matter network,
Yes, the latter. Things actions can go unnoticed when people don’t dive in documentation. So I see two entry points :
- from the Thing page as an indication of their existence, eventually triggerable if no parameter is needed
- from UI rules directly (it did exist in PaperUI IIRW and disappeared in MainUI).
…and for sure, for testing because it’s easier than having to create a rule/script to trigger an action.
I can explain how I understood glhopitals suggestion, and what I think would be a good change.
Today you create a thing. It has several channels. But you have to create Items for each of them before you get any look&feel for what they contain and how they can be used. If you create items, it is tedious to create for everything, and you end up with lots of garbage items for channels you don’t need.
Being able to experiment with a thing before creating the channels would make it easier to start using the thing and understanding which channels you should use.
Think control page of OH2, but showing also unlinked channels.
(I understand that there are complicating factors here like channels that aren’t computed unless they are linked and defaults for which profile to use.)
It existed but didn’t work and was eventually removed in the 2.3 timeframe I think (
I filed the issue complaining that if they don’t work they shouldn’t be shown it wasn’t me but I participated in the discussion). There is an open issue I think to bring them back but I’m not sure of it’s status. I do see that the Cloud Connector notification Actions are available (you have to click on “show all” to see them though) but other Thing actions are not yet available there.
This is something completely different. Thing Actions are completely independent from and their use is completely different from Channels and Items. Actions are basically functions that can be called from Rules to do something/get information that the binding provides.
The best workflow for this today would be to use “Create Equipment from Thing”. You can select all the Channels or just pick the ones you may be interested in, and create all the Items in one go.
Later on, you can individually add/remove Items/Channels using “Create Points from Thing”.
If you need to start over from scratch, you can from the Thing’s Channels tab use “Unlink and remove Items” to get rid of all the Items linked to that Thing and then use “Create Equipment from Thing” again to recreate those you need with the changes you want.
Actually, many (most) bindings do not even collect/generate data for Channels that are not Linked so this would require a pretty big change across the board I think.
The REST endpoint for this was added in [rest] Allow invoking ThingActions via API by J-N-K · Pull Request #2810 · openhab/openhab-core · GitHub. Only the UI part is missing.
Can you please link me that issue?
After a few weeks of using openHAB 3.4, two things come to my mind:
Enforce user authentication, so that openHAB could be safely exposed to the Internet and accessed without having to go through the cloud service. The connection between the cloud service and my local OH is not very stable, which seems to be a known issue.
Support for iOS widgets in the iOS app - it would be neat to be able to use openHAB widgets as iOS widgets.
That would not be sufficient to safely expose OH to the Internet. Frankly, it’s not safe to expose anything in your LAN to the Internet.Those who choose to do so should have the skills, knowledge, and willingness to constantly review logs and monitor their systems for attack which is not something that the average user is capable of.
But, if you want to do something like this you have options.
- Use a VPN (Tailscale doesn’t require opening a port in your firewall).
- Deploy your own instance of the openHAB Cloud Server in a VPS.
If you insist on exposing your LAN to the Internet and are willing to take on the extra work entailed to do that safely:
- Deploy your own instance of openHAB Cloud Server on your LAN and expose that to the Internet.
- Use a reverse proxy as a to go between your OH instance and the Internet. See Running openHAB Behind a Reverse Proxy | openHAB for details and configure good protections there (e.g. client certificate authentication o TFA).
Just read that google opened the gates to add
IOt apps to android auto.
Now we need an openHAB app for our cars
Would be nice to have iOS open up carplay. If you use the homekit integration you can receive notifications in carplay…like garage door closed, etc…
My understanding is that only some stuff shows up on CarPlay through HomeKit (eg garage door openers when you are returning home) and only when ‘Siri’ decides to show it IF you are on the Siri start page . A CarPlay interface would be nice.
if I had a wish it would be around extending the functionality of the openhab mobile app specifically in regards to notifications and adding the possibility to send the device position to the server. Let me explain:
Currently there are a lot of bindings that can be used (Pushover, Telegram, etc.) which provide rich experience when sending notifications to mobile clients. I would love to see some of the functionality be moved into the native openhab app. I’m specifically thinking of ways to
a) send responses back to openhab (e.g. Doorbell rang, should the door be opened? - Yes; - No)
b) Send images as part of notifications. (e.g. a camera image or similar)
Also for this scenario there are 3rd party bindings (e.g. gpstracker) which can do this. If it would be possible to integrate the the possibility to update an openhab item with the current position of the mobile phone in regular intervals (or if the position significantly changes) it would eleviate the necessity to use 3rd party apps and services.
I think both of these features would greately extend the featureset of the mobile integration of openhab.
Thank you for the great work.
On Android, this is already possible with Tasker.
For IOS, you can use the iCloud binding to get the actual position of your device.