Debian has an effective Rolling distribution through testing than can get ahead of Arch.
I wouldn’t call a distro “branch” where maintainers say “don’t use this, it’s not officially supported and may even be insecure” an “effective” distribution. I’d consider it a test bed.
Debian tends to align its release with LTS Kernel and Mesa releases so there have been times the latest stable is running newer versions than Ubuntu
Ubuntu LTS.
Ubuntu’s regular channel releases every 6 months, similar to Fedora or NixOS. That in itself is already a “stable” distro, just not long-time stable (LTS).
So Debian can for a short span of time after release be about as fresh as stable distros which is …kinda obvious? I would not consider a month or so every 2 years to be significant to even mention though, especially if you consider that Debian users aren’t the kind to jump onto a new release early on.
For some the priority to run software that won’t have major bugs, that is what Debian, Ubuntu LTS and RHEL offer.
That’s not the point of those distros at all. The point is to have the same features aswell as bugs for longer periods of time. This is because some functionality the user wants could depend on such bugs/unintended behaviour to be present.
The fact that huge regressions have to be weeded out more carefully before release in LTS is obvious if you know that it’d be expected for those “bugs” to remain present throughout the release’s support window.
As an example, users of Debian are reporting tons of KDE Plasma bugs that was already fixed, but because they are running an ancient version, they still have the bugs.
The idea is that those bug fixes would be backported as patches; old feature version + new security/bug fixes.
In practice, that’s really expensive to do, so often times bug fixes simply aren’t backported and I don’t even want to know the story of security fixes though I’d hope they do better there.
And, even more importantly, search.nixos.org/options to figure out which options to set. Always search for options first. “Installing” something by just adding the package to systemPackages etc. is usually the correct thing to do for end-user applications but not for “system things” such as services.
I meant that as a reply to the second paragraph which generalised anarchism; including the non-Linux world.
I also disagree that this isn’t an issue in the broader Linux community however. See for example the loud minority with an irrational hate against quite obviously good software projects like systemd who got those ideas from charlatans or “experts”.
You gave them an irrevocable license to basically use your content in any way they see fit. Them not showing posts you deleted is just them being nice, not being obligated to do so. They could simply ignore your request or restore posts later.
You should have thought about that when you gave them that license to your content.
I was worried about possibly needing to change license.
I’d rather ask the contributors to consent to licensing their code under the new license. You don’t need the copyright in the hand of one entity to change license, it’s enough if all copyright holders agree.
The situation is made seemingly complicated by the possible need to use copylefted images
WDYM by “images”?
As in art assets? I’m not sure those would even be infectious. I think it’s possible to even use non-free assets in a GPL’d application. It may be better to treat them as such to keep the licensing simple though.
Even then, it’s usually possible to “upgrade” permissively licensed code (such as Apache 2.0) to a copyleft license as long as the original license’s conditions are still met which usually involves denoting which parts of the code is also available under the permissive license.
Why does it need to be public-facing? There may be solutions that don’t require exposing it to billions of people.
Security is always about layers. The more independent layers there are, the fewer the chances someone will break through all of them. There is no one technology that will make your hosting reasonably secure, it’s the combination of multiple.
You’ve already mentioned software ran inside an unprivileged sandbox.
There’s also:
Sandbox ran unprivileged inside a VM
VM ran inside unprivileged sandbox
Firewall only allowing applications to open certain ports
Server running all of that hosted by someone else on their network with their own abstractions
There’s the WIP NixOS-based SnowflakeOS that aims to make NixOS approachable for mere mortals but that’s still declarative configuration and of course still NixOS under the hood.
There’s a bunch of immutable distros out there that use OStree or some other imperatively managed snapshotting mechanism such as Fedora Silverblue or VanillaOS.