So the Zwave-JS integration ?
That’s we talk about.
So the Zwave-JS integration ?
That’s we talk about.
LOL. Sure, a survey will bring more precise information with no bias
I agree that a lot of the pain of updates is self-inflicted by the design architecture which bundles so much into each core update. Then when things designated as “core” break, remain unfixed or are deprecated, well, we’ve seen how that goes.
This (GPIO support) could be a good example of how core components could be broken out as optional integrations. I agree HACS isn’t the ideal tool. I’d envision something more like the community add-ons, but of course that’s just an uniformed observation, since I haven’t looked into the details of how it’s built.
You can use any integration as custom by copying the files and placing them into custom_components.
Not likely, people keep glancing over 2 of the 5 reasons why GPIO was specifically deprecated. It basically comes down to hardware if you don’t feel like reading those bullets I linked earlier.
Yep, that’s how we manage real products, we talk with customers.
…but as this is no product but a open source project, stuff is done differently.
Low usage proven wrong. Maintaner for rpi gpio was found in hours. Containers have nothing to do with rpi gpio. Rpi gpio on 64bit has been working great. Esphome can not be comparet to hardwired short distance connection.
Desison maker just thought wrong.
You’ve spent more time writing replies than the time it takes to switch to the custom integration, 10x over. Give it a rest.
Also, the container needs to talk to the hardware it runs on for gpio… you’re just showing your lack of knowledge on system with comments like this.
So I’m trying to find out how to do this, follow links which lead to loads of comments, follow through, find another link, follow through, find another link… Not a single instruction, not a single helpful post with actually what you have to do.
Then they wonder why people are here about the loss of GPIO, talk of hacs integration, no good if it was a part of the main Home Assistant then all good but it isn’t. Looked at installing hacs but couldn’t make head nor tail of instructions.
I run a clean install with only official add on, make the GPIO part of the official add on and no one would have a problem with it. I like my clean install, easy to look after, easy to update, and is reliable.
Your company do product management through surveys? Intriguing…
Sure it doesn’t prioritize what benefits most of its customers with the (always limited) workforce it has, including deprecating/retiring products?
That’s definitely what the software company I work for does…
It’s a really easy install… what can’t you make heads or tails of?
In general those are fair points.
But one can argue with them. Let me go through the list
What metrics are used to measure that? HA stats which are no mandatory so unreliable value-wise?
Or those few thousands of users really don’t count? How many integrations HA have being used by only such a small percentage of users?
Define ‘unmaintained’. If it works feature wise - there is no reason to exclude it.
If it has been included to the core, it shouldn’t be maintained by single “external” person. It becomes part of internal system and should be supported by core developers. Otherwise every component would be vulnerable to loosing main developer.
Cannot comment on it.
But HA core contains a lot of functionality not available for other users. Not sure what you mean. Is it the need of maintaining those differences in the integration? So it applies to to HACS integration too. What’s the difference then?
If it would had been true, most users would have migrated to ESPhome already. I have no impression they agree with you.
But again, thank you for providing more reasons behind the decision.
Definitely was not easy for me. Took 3 tries until I found good video how to. Again, first time ssh, github and magic commands on terminal.
Now it is easy to use.
Analytics have 100k+ sample size. Saying that 'it isn’t enough information is ridiculous.
That’s the thing, it doesn’t work on a bunch of RPi combos at the moment. All hardware that no one owns yet there was open issues for them.
No one maintains it. It doesn’t have a code owner. Every integration has a person who is the code owner. They approve other peoples PRs for their integration. You can see code-owners in the manifest file.
And ultimately they get to decide what they maintain. Currently they maintain over 200 integrations that are vital to home assistant. They don’t have the hardware or resrouces to spend extra time on GPIO, that much has been stated.
Yes, but then it’s up to whoever takes it over. In this case, it’s @thecode. So there’s literally no loss. All these complaints at this point are from people who are unwilling to move to the custom integration.
You don’t need to run CLI to install hacs… You found a poor video. There’s a bunch of installation videos on HACS community guides. You can do this all without CLI.
The main gist of the installation is really simple. So simple, i can basically write it from memory
This is probably the best english video without CLI
which is linked on HACS website under the installation community guides area
It is not about unwilling. For me it is now easy.
I am still here to make you understand that it is not easy for many.
Suprisingly many have contacted me here and my own channel about this problem. All hardware guys with no coding background.
I think I have to make a video about all this…
I bet it would be less work in future for all if just keep it running.
Like someone said earlier, would be nice to atleast have instructions how to keep using it. Right now on rpi gpio page it only says:
"This integration is deprecated and will be removed in Home Assistant Core 2022.6.
For more information see: Architectural Decision Record 0019
Most probably, I was unclear. The main point was that you cannot be sure to still use a core component 1:1 in the release, where it is not part of core anymore, because one single line/design change/… in core in major or dot release can break every component to work.
E.g. you cannot use 2011.12.10 sonos integration as custom component in 2022.2.0, because core introduced other dependencies for media player, …
Don’t need to discuss this here, but I only wanted to point in the direction, that it is not correct if the user think, that it is always enough to copy it to custom_components to kepp this running (without changes).
Ah yes, I follow your statement now. Core dependencies aside, you should be able to take any integration and make it custom.
I am still here to make you understand that it is not easy for many.
Oh I fully understand.
Folders in HA? Copy paste? Didn’t see them before i added file browser… didn’t need file browser before I had to install HACS
It’s a computer… what did you think when you made it. Keep moving those goalposts.