For what purpose though? All my servers and containers run debian. Everthing I care about publishes fresh packages for it, but on their own repos. My desktop and laptop run pop_os with a few additional repos.
Everything that needs to be bleeding edge can come from snap or flatpak these days anyway.
I use Fedora. I like the combination of recent, stable, up-to-date software, new releases every six months, and firmware updates for my ThinkPad direct from Lenovo.
Guix is a source based (rolling release) distro. Any package operation you do like like installing, updating, or removing, can be rolled back. So if an update ever breaks anything you can just roll-back and wait for the fix. You can even pin that specific package and continue to upgrade the rest of your system. And every state is saved in a generation, so you can go to any state your system has ever been in package/configuration wise.
No, I have no issue with pacman, it’s the “garuda-update” script I don’t care for. I see endeavour has eos-update which I haven’t really looked at much but in Garuda if use “pacman -Syu” it will interrupt with “Garuda uses garuda-update for updates” - I know it’s trivial and I don’t have to use it but I don’t like that. Don’t interrupt my workflow to try and coerce me to use your script. Yes, it’s a petty gripe but it feels very microsoft-like in the same way that Windows 11 will delay the launch of Firefox to tell you “Edge was built for Windows.”
Crux user here. I like the port tree system and simple package building recipes. It’s also a distro that kept things very simple over the years despite the rise of dbus and systems. Also the mascot.
Void Linux. A great compromise between being up to date and being stable AF. They’re not bleeding edge, but cutting edge, most definitely. For example, they only recently transfered to kernel 6.3, while Arch had it months ago… with instability issues I might add. Void maintainers would rather let these wrinckles get ironed out than implement the latest and greatest.
It is a rolling release distro, so nothing new there. Packages get regular updates, same as any other rolling release distro, except for the kernel packages which are carefully examined before being submitted in the repo. The number of precompiled packages is not huge, but the src templates are (you just have to compile them from source with xbps-src, which is a piece of cake when you already have the template file).
The good thing is that all package templates get checked for buildability (test) on GH. If the template passes all tests, it makes it in the repo, if not, it doesn’t, simple as that.
If you think you would be comfortable with Arch, you’d be comfortable with Void as well 😉.
Besides those builders that run the checks made by the developers of the apps (is simple running a make check or whatever the build system the package uses, void Linux does have some problems mostly because of the small team.
One of the biggest one is basically being stuck in 1.X series of musl forever until someone steps up and creates a solution that doesn’t require to rebuild all packages because of an ABI breaking change in armv6l systems at 2.X musl series .
I feel like something like Fedora fits the bill: great, reliable, well-maintained repositories, decently updated kernels, yet never faced any major issues, and access to quite updated packages. Only issue is Red Hat caused a stir recently, though I still believe Red Hat does more good than bad in the open source community.
Some are making the case that Fedora’s new telemetry integration isn’t like the bad telemetry like Google and others, it is ‘anonymised.’ Every corporation says this before they remove the username from the data collected and keep the unique user id. I don’t trust Red Hat…and now with this latest reveal, Fedora either. And privacy is all about trust.
I’m making that case. I trust Fedora and Red Hat to handle telemetry correctly, but I can verify it by looking at the source and I’ll give them constructive feedback if I have concerns. May I ask which distro you are planning to use where the source is NOT contributed primarily by engineers working for a corporation that puts dollars first?
I disagree that as the as the article states telemetry “contradicts open-source values”, nowhere is it said in the official definition that telemetry by itself is not ok and as long as it is opt-in and the handler makes clear reports on the data they gathered, I’d say it’s a good opportunity to give valuable insight to the developers on the use of their software, done in this manner it doesn’t trample over anyone’s choices either.
Notable examples of open source projects that implement telemetry are KDE and Mozilla, it’s not unheard of at all
While I admit that the timing with Red Hat’s closed-sourcing is really bad, and I’m also going to start avoiding Fedora for the same reason, saying that opt-in telemetry (that one can literally read the source code of) is “putting dollars first” is really dumb. Do you think the same about Debian’s popularity-contest, which has existed since 2004?
Some are making the case that Fedora’s new telemetry integration isn’t like the bad telemetry like Google and others, it is ‘anonymised.’ Every corporation says this before they remove the username from the data collected and keep the unique user id. I don’t trust Red Hat…and now with this latest reveal, Fedora either. And privacy is all about trust.
Please stop with the FUD about the Fedora telemetry. It is opt-in and is no different than popularity-contest on Debian.
@NoRecognition84@rodneyck its bad anyway. Why a opensource project will do something like that? Telemetry causes bad performance in production. If its opt in, no one will activate, and soon the business will force its use.
I use to be like you. I used Arch for a long time then tried everything else that was similar like tumbleweed etc. Then I used Fedora and forgot about distrohopping entirely. I still use Arch on my pi4 though because it works nicely for use cases like that.
However I will warn you anything can and will be unstable eventually. Its the nature of software, bugs will happen. For instance a package called ostree was pretty much broken on all distros even Fedora which is crazy.
They did mention stable, which is not something Manjaro can claim in my experience. They tend to hold back packages in the name of stability but it causes problems when using the AUR sometimes.
+1 for Tumbleweed, it works so incredibly well. In the very rare case where an update doesn’t work out for you, you can easily roll back to a previous btrfs snapshot.
Fedora is quite nice, too, but I’ve come to prefer rolling distros over a release based one.
Kalpa / Aeon might be interesting, too, if your use case fits an immutable distro.
After many years on Ubuntu I switched to a Tumbleweed and couldn’t be happier. Apparently a rolling distro can be more reliable than a traditional point-release one.
linux
Hot
This magazine is from a federated server and may be incomplete. Browse more on the original instance.