Hacker News Reader

Comment by Scramblejams | Hacker News Reader

Any indication whether this make it into Ubuntu 20.04?

They are staying with 5.4 but will back port WireGuard.

Yikes, that feels like a mistake. 5.4 has awful issues with intel integrated graphics crashing.

The timing is a little unfortunate. 5.6 will be too late to make it in for 20.04, but 5.5 isn't an LTS kernel and 20.04 is an LTS distribution, so I understand not wanting to use 5.5.

The issue of a rigid release schedule! April means April whether you miss out on something or not.

You still end up having to make compromises no matter what.

One method allows for you to determine which compromises are acceptable, while the other does not.

Not really. Let's consider an alternate flexible version that stretches up to 2-3 months outside of emergencies. Versus a non-flexible version that stretches 0 months outside of emergencies.

If you compare the flexible version targeted at June, and the non-flexible version targeted at August, you'll find that they're making almost the exact same compromises.

Nothing ever stops you from using an earlier version if it's more stable. So both schedules get to chose from multiple versions. Maybe flexible chooses from 6 month old to 3 months in the future code, via delaying. But non-flexible can choose from 9 month old to 0 month old code. It works out the same, and the only difference is how you label it.

I'm kind of curious why Canonical don't simultaneously release a non-LTS release at the same time as the LTS release, to continue the flow of non-LTS releases. It seems like it'd make gearing up for the "LTS + 6mo" non-LTS release easier. Plus, then there'd be a clear non-LTS release to attribute any backported update packages to (that show up before they cut the "LTS + 6mo" non-LTS release.)

They do keep latest kernel PPAs for those who want to experiment. https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.5/

When I was involved we did LTS, and then half the distro team was focused on the LTS .1 release which happens 3 months later. The reality of software is that no amount of internal testing is the same as the amount of bug reports you get after the release. And, for an LTS as it's going to be around for a long time getting fixes in is more important - with a standard release you can choose to leave some that you'll resolve in the next six month release.

The reality of doing LTS, plus LTS .1 mean that the next standard release didn't get as much attention. So in reality the first six month release was often very rough, with a lot of new things in it that would be straightened out over the course of the entire cycle.

As someone mentions later in the thread Canonical started doing HWE releases, which means you can run a later kernel on the LTS. So I'm running 18.04.4 (I think) which has a HWE kernel that is newer than the one that originally shipped - not just security fixes but newer hardware compatibility.

ITT: why Manjaro's nerfed rolling is best model while point releases are a server meme.

Eh, just as Canonical will backport Wireguard, I'd expect them to backport stabilization fixes to address issues like that.

Have the issues been fixed in 5.5 or 5.6?

Yes, it was fixed in 5.5. There's a few corners of the internet explaining the issue...this one is good enough https://bbs.archlinux.org/viewtopic.php?id=250765 .

I have a newish all intel setup, and 5.4 crashes to the point of insanity. I gave up and installed 5.5 and it went away. It's a serious issue for an LTS release.

First time I put Ubuntu on a laptop (2004 or so), there was an intel graphics regression so bad, I put windows back on it... This past October was the first time, in a very long time, Linux went on as my main OS. And running a new x570 board with an rx 5700xt has been painful until much later, when support got dramatically better.

Still have an issue with when I come back from the blank screen and login, the login screen won't go away... all the top/side bar shows, but it covers apps... I wind up ctrl+alt=f3 to login to a terminal and reboot. Should really figure out the commands to kill and restart gnome, but it's really been a pain. not sure if KDE might be better.

(sudo) pkill Xorg should likely do the trick, but check e.g. htop (hit F5 to get a tree view) for the name of the root process of the graphics session.

switched over to KDE after writing that... will see if it's still an issue in a day or two.

Really? Do you have a source for that? (I'm not doubting you, I just haven't heard anything about that, and would be very happy if this is true).

https://lists.ubuntu.com/archives/kernel-team/2020-February/...

Sorry, I meant the part about backporting wireguard. I did find a phoronix article on it: https://www.phoronix.com/scan.php?page=news_item&px=Ubuntu-2...

https://www.phoronix.com/scan.php?page=news_item&px=Ubuntu-2... is one source but I remember seeing it elsewhere.

After ubuntu 20.10 is released 20.04 will probably get a later version of the kernel in the linux-hwe package

It's gonna be 5.4

Nothing forbides you of installing 5.6 on Ubuntu 20.04 or other older version.

any reason for choosing ubuntu over arch for personal computers? Or any rolling release distro for that matter

Python packages don't install on /usr/local on Arch which makes it easy to wreck your installation, many projects and scripts out there only support Ubuntu and not Arch, etc.

And I say this as an Arch fan who would rather avoid Ubuntu...

virtualenv?

Before I try to respond, you have to ask yourself: what do you imagine is the reason why Ubuntu folks have gone out of their way to patch Python for /usr/local? Surely it's not out of ignorance for virtualenv?

interesting, I never looked where the python aur packages were installed

This isn't AUR-specific, it's an issue that comes up with sudo pip install.

pip with sudo is always discouraged. Did you encounter use cases that absolutely requires you to install packages with `sudo pip`?

I don't even encounter use cases that "absolutely require" me to use Arch in the first place.

I assume no then, so just stop misusing python package manager and stop blaming distros for it

You can assume whatever you want, and I'm not blaming anyone, just telling you what reasons people have for using one distro over the other. You're arguing with the wrong person for the wrong reason.

Sorry I came down to arguing like that, but I really honestly wanted to know what problems there were with the scenario you proposed. I am in devops and those are the kind of nuances I need to account for. That I why I asked for examples and felt your answer was condescending. But anyways, I can see from this thread that asking questions is somewhat frowned upon here, when the question itself is of a nature that displeases people, so I'll refrain from participating like this in the future.

Ubuntu takes all of about 10 minutes to install a full blown desktop and Just Works™ based on my recent experience. Not to take anything away from the amazing tweakability and customizability of Arch, but the target audience is completely different.

I use the anarchy distro to help me through the installation. It doesn't change the fact you need to understand the underlying installation process, but reduces the installation time to several minutes. There is also arco linux for a more user-friendly interface

The out of the box crowd can always use Manjaro.

really folks, I know the downvote system is useful, but on a question? I mean, I have to know the answer beforehand so I don't get downvoted? I am really tempted not to use this platform anymore

Downvoters are aggressive here yes. It's a shame, though your question isn't particularly pertinent to this article.

Kind of a double-edged sword that keeps the discussion on target.

Reply on news.ycombinator.com