Jellyfin client on mobile and AndroidTV, and Strawberry on PC. All my music is on my NAS, which Jellyfin server and CIFS/SMB can access.

Keep meaning to look into Music Assistant for Home Assistant, as I have the latter.

If you’re not using a domain for email, the MX record should not be left empty, as it can still be used maliciously.

Following a guide like this (it’s one of many) will ensure it’s not used at all:

https://www.mailhardener.com/kb/hardening-unused-domains

It does. Very well. Comes with the GE (Glorious Eggroll) version of Proton by default.

I use it for all non-Steam/-native games.

On my desktop with a 2060 it’s zero issue, but on my 940M hybrid video laptop it’s been spicy at times.

Usually updates without an issue, and the most I need to do is delay the -settings package update until after the rest. (IIRC, see what I mean 😄).

I’m not familiar with what’s been said or done regarding bots, but I’d be surprised if they were planning to shutdown the self-hosting part? But I can see how they might shutdown the free cloud auth aspect.

As RD consists of the server/client software and the authentication software, the latter is also made available online to all with no reliability promises. But there’s nothing stopping people from hosting both parts locally, and is how I implemented it. No traffic to third parties, etc.

Can’t advise on domain-based setup, as I’ve not tried it, but depending on how you’re planning to use it, there may be no need for a domain. I only used mine locally (or via WireGuard when outside), with hbbr and hbbs hosted in Docker on my NAS, and it worked fine with my mobile devices and PCs.

I’ve been using Wallabag for years, after leaving Pocket, then just sharing links between devices using browser bookmarks or sync, and trying another solution that was shiny but lacking (can’t recall the name).

Wallabag has extensions for most browsers, apps for iOS and Android (that hook into the “Share with…” functionality), and runs in a low-resource Docker container. The web UI isn’t pretty, but it’s functional.

I have SWAG+Authelia (another Docker stack) and Cloudflare sitting in front of it, so has all the SSL and MFA needed.

At this point, I’m assuming OP wasn’t on Lemmy last week.

(Possibly wrong, but it’s a plausible explanation of why this kind of thing happens all the time here.)

After too many wild rides with Watchtower auto-nuking services, thanks to breaking changes (migrations, DB updates, deployment changes, etc), I switched to What’s Up Docker and pin the version for all of my containers.

WUD lets me know when something has an update, so I periodically go through their release notes and do the update(s) manually. Usually as simple as read the notes, changes version in compose, down (or pull), then “up -d”. But this approach has saved my bacon multiple times.

I’ve seen there are other solutions - of varying degrees of promises vs delivery - but most of my stuff is long term and stable. My approach maintains all that.

Makes my day when my weird sense of humour gets a reaction. Thanks. 😄

There’s something similar for Genshin Impact and Honkai: Star Rail. In their case it lets those games run on Linux without any other rain dancing needed. And both work very well, the last time I checked.

IIRC, they also disable or omit some or the telemetry and privacy-invasive stuff. (I mean, why not… if they get found, the full weight of the overpaid lawyers they sic on you will behave as if you did far worse).

No idea if this client does similar, though.

I mean, this is exactly why example.com exists. But I bet ICANN didn’t expect this level of meta abstraction to the absurdity. 😅

My recommendation is Debian for a server (real or virtual), or Proxmox. The former is perfectly reasonable and excellent experience; the latter is more flexible and more complex.

Debian is the parent distro of numerous Linux flavours (including *buntu, which aren’t suitable as a server OS, IMHO), so administration and services are all common (apt, etc). No need to learn dnf, pacman/yay, etc.

It’s still my preferred server OS, despite other options and being experienced.

Though I do also have a NUC running Proxmox (for VMs and LXCs), and both a NAS and RasPi running Docker. 🤷‍♂️ My Debian server is a VM inside one of them.

Can’t speak for the rest, but the “shrinking when replying” thing is a known issue and impacts iPhone and Android.

It is, I think, related to the iPhone theme in the app (was enabled by default on my Android, for some reason). Going into the settings and changing it from iPhone to Android fixed it for me.

As another user said, the dev is doing a great job on this app with limited resources, and does test and check on the devices they have available, but “works on my machine 🤷‍♂️” is a real thing. Frustrating, but valid.

Exactly the same with me. 😄

This is a known issue that the dev is working on. I think the current workaround is to change the app theme from “iOS” to “Android” (unsure about other theme options).

On Android it’s called Device Mode in the System section of Settings > Appearance > Themes.

Edit: I made this setting change earlier and the above was my first chance to test it. Can confirm: changing the theme to “Android” works on my Samsung phone.

It’s true that people on the internet can be dicks. Even more so technical people (and that’s not limited to online: those online dicks are usually IRL dicks when taking technical stuff). But that’s a hurdle, not a barrier.

There’s little anyone here can do to help OP, as they (if I understand it correctly) have already irreparably nuked their hardware. The current problem is significantly different and harder than the original problem. Asking randos on this community is unlikely to yield results. Hence the focus on variations of “Now… what did we learn? 🤨”

I’m not trying to help, as I’m not familiar enough with SAS nor the current problem. The same is likely true of others here.

Can you really blame anyone who turns to AI, because that garbage at least sounds like it tries to help you?

A comfortable lie is still a lie. Everything that comes out of an LLM is a lie until proven otherwise. (“Lie” is a bit misleading, though, as they don’t have agency or intent: they’re a variation of your phone keyboard’s next-word text prediction algorithm. With added flattery and confidence.)

There’s a reason experienced people stress hard to others about not using them as shortcuts to your own knowledge. This is the outcome.

Another way to look at it is “trust, but verify”. If you’re intent on relying on probabilistic text as an answer, instead of bothering to learn, then take what it’s given you and verify what that does before doing it. You could learn to be an effective sloperator with just that common sense.

But if you’re going to give an LLM root/admin access to a production environment, then expect to be laughed at, because you had plenty of opportunities to not destroy something and actively chose not to use them.

On a stock Samsung S22 with the latest security patch (November).

+1 for me on this behaviour, too. Usually after making a reply, but rarely (not never) at other times. Usually need to kill the app and reload to fix it.

It’s been happening for about a month now?

Edit: There it is again, after making this reply. 😅

You can still use tailscale and reverse proxy to allow remote streaming

I used to use Plex and when I discovered there was paid remote streaming function - that goes through their servers - my reactions were “Haha, no”* and checking whether my existing WireGuard setup would do it instead.

Whaddya know, remote streaming using Plex and PlexAmp at no cost.

*Not because I begrudge them recouping costs, but because it’s designed that way to justify charging for it, gives them whatever information they want from my viewing, and it’s not self-hosting if there’s any third party cloud/account component to it.