These packages IMO were great for noobs, upgrading the kernel automatically without user intervention. If there was a concern of breaking a noob’s system, I feel like the solution to that is make “linux-lts” the default. Then users who were knowledgeable enough to know what a kernel is would have to actively switch to “linux-latest”.
Having to manually upgrade the kernel when a new minor revision is released is a regression in the user experience. We shouldn’t be moving backwards, but moving forwards.
I don’t know - or can’t remember - why it was dropped back then.
It’s maybe coincidental, but I think it was at the same period of kernel 5.10.
For a LTS kernel, the first weeks of this kernel showed an awful bunch of bugs and instability. Many people that switched to early 5.10 complained.
Automatically switch to the next LTS kernel is not always the right move to do.
I remember someone saying LTS kernels are falsely considered as more stable than others (especially when they just arrive). It’s just they are supported way longer.
Early 5.12 was also flagged by Linus Torvalds as don't use, it could wipe your entire system.
Personnaly I consider kernel selection as a choice by user, rather than a choice by system.
That’s something that should be fixed by holding those problematic kernels back from the stable channel. That’s the whole point behind the update model Manjaro uses. Users trade in being behind a bit on updates for increased stability.
The solution isn’t push the decision to the user. Now, the user, who doesn’t know any better, will blindly update without knowing that the next kernel has some issues.
That was part of the beauty behind automating kernel updates. Manjaro could put it out when it was ready.
The Linux philosophy is “power in the hands of the user”, so if you see N00bs doing thing they shouldn’t be doing, simply refer them to the N00b Tutorial, which explains this to them:
And Manjaro’s philosophy is: Manjaro is the best user-friendly OS for very new hardware so the latest kernel is our standard and Manjaro will always work straight out of the box even where other Linux distro’s wont!
Although your point is well-taken, the user needs to learn that Linux is not a car like Windows any more: they’re riding a racing motorbike now! and though it’s much more fun riding a motorbike, they have to unlearn some skills they were used to under Windows and learn some new skills they didn’t even know were possible with a computer!
If your system breaks due to a package like linux-latest - who would you blame - I know who.
You would flame against the packager and the distribution which made your workstation stall due to a kernel replacement and you would have to spend several hours - if inexperienced, even days - to get it back online.
I think the idea was good - but it also created a heap of other issues - probably why the idea was abandoned.
That’s why I said the default out of the box would be ‘linux-lts’. And users who switched to ‘linux-latest’ would have to actively make that decision, knowing the potential risk. Though personally, I’ve never had an issue with the automatic upgrades, even on linux-latest.
Part of the point of Manjaro is that the user didn’t have to learn new skills, and they don’t have to use the terminal for basic tasks, unlike other distros, etc.
The way I see it, the solution to reducing problems with kernel updates, is keeping stable and LTS kernel updates in the unstable and testing branches for a bit longer, maybe 3 or 4 weeks (or more) if necessary.
Nope. The point is for it to work out of the box on hardware other Linux distros don’t even support.
The best analogy I’ve seen so far is:
Windows is a car. Linux is a motorbike. They both get you from point A to point B, but they do that radically differently. A motorbike needs you to learn new skills that you don’t have yet and unlearn some skills you already have…
Manjaro is just a racing motorbike… (rolling release) so you need to learn even more skills on top of the skills the other non-Arch Linux distros require. One of them is kernel management as the other distros do that for the end-user…
P.S. Your original question was already answered by @linux-aarhus so I’ve marked that as the solution to your question. Now we’re just divulging into a philosophical debate on the goals of Linux and Manjaro, which this forum wasn’t designed to do.