0.110 broke more stuff for me than anything for a while. Lots of custom cards icons broke. CH broken. Hopefully the devs fix the remaining ones before Thursday.
That’s not good to hear.
Yeah, another irony…
(see my other still answered points elsewhere about how the strong tie-in with Amazon and Google are meant to be reconciled with ‘local control and privacy first’)
I thought ‘Cloud’ was ‘bad’.
Why is it included by default?
(The answer of course is ‘money’ or ‘profit’, which is absolutely fine in my book but at least stop being duplicitous about it)
I think that’s harsh. I can understand why it’s on by default.
1 Like
Yeah I’d attribute it to demand more than anything.
1 Like
Not that it matters now, but I try and visit the forums 2-3 times a day, and usually try and look at at least the topic title of every topic since my last visit. Not once did I see this in my recent topics list, which given the activity on here I find utterly weird. So that’s why I never voted - I never saw it!
I did see your post in the original thread asking for it to be stickied, but must have forgotten to go back to the link when I finished catching up on that thread 
Sorry!
HA in venv here, if it matters 
Yes, I accept it is a bit harsh in the context of this thread (within a thread) but added together with the whole Amazon and Google thing I just find it a all bit hard to take.
Not that I object in the slightest to any of it (Cloud, Amazon, Google) I just think total transparency would be better.
talking about that…
how do we need to change the http: section:
base_url: !secret base_url
ssl_certificate: /ssl/fullchain.pem
ssl_key: /ssl/privkey.pem
ip_ban_enabled: true
login_attempts_threshold: 10
I can’t seem to find the documentation on the breaking changes for Http’s internal_url and external_url, (though my 110b4 seems to be fine…)
of course, that is without mentioning the horror of no more colored icons set by custom-ui…
Which makes me answer the OP here with: I haven’t on my production system, only on my test config… Hassio btw. Or what it’s called nowadays.
You don’t need to change the base URL. It will be migrated to the new options for you. Reverting to the previous version did leave this in place and caused issues. Not sure if it has been addressed or not with subsequent beta releases, I lost track.
even if I have this in my configuration.yaml?
Yes. From the RC notes:
After upgrading to 0.110, you can delete
base_urlfrom your configuration as Home Assistant will automatically migrate that setting for you on upgrade.
yes, I can see it now, in the hidden core.config… along with everything from:
homeassistant:
name: Ha Z-wave
latitude: !secret latitude
longitude: !secret longitude
elevation: !secret elevation
unit_system: metric
time_zone: !secret time_zone
meaning I can take these out of configuration.yaml too…? Do these entries still serve any purpose at all ?
sorry if this diverges too much from the OP topic.
there is still a http file in the hidden directory though, which still holds
{
"data": {
"base_url": "https://
etc
Again, I don’t understand why the poll is closed 
300 votes is still not much compared to the entire install base of HA…
I think they should have a long running poll and have as many people vote as possible so they have at least a good idea of their userbase usage…
1 Like
Only if you want to set all that using the UI.
No I dont, I really like to have some control. Yet, maybe later.
So I probably can copy the internal_url and external_url too, in to my configuration.yaml. That way I am always 100% sure what setting is used? (I have 3 instances all using duckdns.org, so the port is of utmost importance, and I love to see that in my config…)
this works:
homeassistant:
name: Ha Z-wave
external_url: !secret external_url
latitude: !secret latitude
longitude: !secret longitude
elevation: !secret elevation
unit_system: metric
time_zone: !secret time_zone
packages: !include_dir_named packages
whitelist_external_dirs: /config
auth_providers:
- type: homeassistant
and take out base_url from the http: section.
thanks, wont go on further in this topic.
Less than 1/2 percent of the community responded and you’re defending that paltry sample size as being statistically significant. It’s not.
A sample size of 0.5% is far from “paltry” from a “statistically significant” stand point. There are likely issues with the poll, but the sample size gotten here (for a forum pop between 50~100k?) isn’t it.
Compared to the entire true population of HA users? Now that’s a different story.
1 Like
- You live in a community of 60K people.
- 300 people who don’t own dogs protest to ban all dogs from the community.
You’re saying the opinion of this miniscule group adequately reflects the opinion of the entire community.
This poll was locked to a specific group of 300 people with a specific agenda? Huh. Who knew.
If 300 random people in your 60k community share their views on something, that is significant and statistically likely to be valuable.
1 Like
I’m sorry, but you’re both kind of wrong. 
You can’t take either data material for your comparison. 300 voters here in the forum are not enough people to get a reasonable and useful vote. It is a sign into a direction, but nothing more.
But, setting up 60000 members against that is unrealistic at best. Even I have a second user account here, because I lost the password to the email account registered… 
And as usage statistics from other forums show, if you have 60000 registered members, at least one third of these accounts is “not real”. Imo the percentage is even more, I’d say half of the accounts is dead or otherwise not necessary for our comparison…
Reasonable would be, to set the active users from, let’s say, the last three months. At least two log-ins in this timeframe are necessary and a minimum post count should also be considered. One or two time posters are not interesting for us.
But that doesn’t solve the problem that we have here. We, as a community, won’t be able to get transparent data.
I’d suggest a different look on all that. The people, that vote here, are the people, that are important for HA. Most users of a software are not present in community forums, if they step in, it is mostly because there are problems with their installation (not installation method!). Some stay afterwards, but most users just leave after solving their problem, or even worse, the problem doesn’t get solved and they move on to other software.
The people that bring a system forward, are the people that work with a software and are willing to go the extra mile. Why is not important. These small number of people, mostly these are also “the community people”, are the foundation of a software, that is community driven. And if these people don’t get an installation method, that is doable for them, they will as well move on. Maybe not at first, but after some time…This would mean a lot less reports, and the system will not grow and get better. You always need the “hard core” fans to bring this forward.
I for one would select the most active users from the last two months with a minimum post count of 20 posts and would set these users up, to vote for their installation method. That would be a reasonable data base to get some understanding, of what,where,when.
Just my two cents.
2 Likes
In defense of my original statement, I was merely arguing what is or isn’t paltry in terms of statistical significance. From a purely statistical point of view, 300+ on a 60k pop is great.
I don’t see anything I disagree with in what you’ve written. For whatever that’s worth 
1 Like