Automated package testing before each update rolls out to users.

In the event that an update does break your system, you can roll back to the last snapshot from the grub menu (using the smart btrfs setup that is the installation default).

Also generally, maintained by very smart people, community is not toxic, corporate overlords more benign than most, IMO.

…spilled his arithmetic overflow on the ground.

Check out JOOQ.

Yes, it avoids the worst of stupid thumbnail patterns that the YouTube algorithm pressures people into. Yes, nobody is making an O face, and the title isn’t misleading. But this is still in a style shaped by those same pressures, and IS cringey to me. Titling something “I did such and such” and the visual style… does this not scream dumbed down for the algorithm to you? Five years ago you’d have wondered if this was targeted at children or something.

Listen, I understand why people who may have good quality content make compromises to reach more viewers. You might think i’m being excessively hard-ass about it, but in my opinion, playing ball with the algorithm just contributes to the problem. The fact that the style of this thumbnail has become so normalized that people can’t even see what I find objectionable, IMO, just demonstrates what a slippery slope this is.

I’m sorry I just can’t bring myself to watch a video whose thumbnail looks like that.

And further thanks for taking this much care wrt backwards compatibility and helping clients avoid breaking.

There is definitely a caveat with nvidia. The nvidia repo is managed external to the main repos, so it is possible for a new kernel to drop in the system repo and the nvidia repo not yet be updated with a compatible driver.

I always wait a few days on such updates and watch the mailing lists for problems especially from nvidia users. So far I’ve only experienced problems due to prime wonkiness that required re-running a couple of prime commands. I haven’t had to use the boot-from-btrfs-snapshot yet, but it’s a nice security blanket.

Tumbleweed? Could you have been looking at Leap?

I’d recommend openSUSE Tumbleweed instead. They originated the btrfs setup that lets you rollback in the grub menu, which has been copied by others. They are bleeding edge except that all packages go through an automated testing system before being rolled out so there’s much less breakage to start with.

I’d suggest checking out fish shell.

As someone who has always been cautious about SSD writes (possibly overcautious/ paranoid? Idk, some seem to think it’s not a concern with modern SSDs. But I haven’t really spent any time researching recently.) I always like to have a hard disk as well as an SSD and I put my writeback device and any swap partitions there.

Sorry this probably isn’t a helpful answer.

Sorry, it’s been a while since I read this stuff and I don’t have the links. The state of web searching these days sucks and I can’t easily find them.

One bit I remember was that a lot of the concern about LRU inversion in ZRAM that might make ZSWAP look preferable is out of date since the addition of a writeback option to ZRAM. I also remember people claiming that ZRAM had an advantage in being multithreaded.

FWIW I find this three year old answer saying the kswapd that ZSWAP uses is single threaded but there is a patch to make it multithreaded that significantly improves it’s performance. No idea if this is out of date.

I’m talking about ZRAM and I did mean writeback.

“With CONFIG_ZRAM_WRITEBACK, zram can write idle/incompressible page to backing storage rather than keeping it in memory.”

From https://www.kernel.org/doc/html/latest/admin-guide/blockdev/zram.html

Aren’t their embedded systems that run the Linux kernel without the core-utils (maybe with busybox instead) and would therefore be non-gnu linux variants?

I just want to comment on the Nvidia thing, since they are so common on gaming machines. And I have no opinion / data on the performance of nvidia vs other to gpus with linux.

With nvidia on linux you will be fine if you just do a couple of things: hang back a little on applying updates (specifically kernel and Nvidia driver updates) and watch the relevant forums / lists for problems from nvidia users. Only update after a few days have gone by without such reports or, if reports have surfaced, after they get fixed.

openSUSE Tumbleweed user here, and I’ve actually had very few problems, and they were specifically caused by prime. I may have dodged a problem or two with the above strategy though.

This is a theoretical advantage of ZSWAP over ZRAM, but when I researched it, every real world comparison I found seemed to find that ZRAM performed better even when this advantage should have come into play.

I’m not going to remember the right terminology, but you can also configure it with a chunk of disk to stick files that it can’t compress into so they don’t end up clogging up your swapspace.

I just bang my head on the table and save a step.

Tumbleweed is great! It’s close to bleeding edge with an automated testing system preventing most problems from ever getting to you. And if an update does break your system, if you installed withtheir btrfs default, you can just boot to a pre-update snapshot right from the grub menu, and roll back to it.

Tumbleweed is great! It’s close to bleeding edge with an automated testing system preventing many problems from ever getting to you. And if an update does break your system, if you installed work their btrfs default, you can just boot to a pre-update snapshot right from the grub menu, and roll back to it.