So I’ve been iso live testing Manjaro KDE Plasma lately and it looks very polished.
On the other hand, there is a negative vibe towards it.
Why the hate?
So I’ve been iso live testing Manjaro KDE Plasma lately and it looks very polished.
On the other hand, there is a negative vibe towards it.
Why the hate?
It’s not all “purists” and “tribalism”, Manjaro actually has issues. Besides the well known certificate issues and older packages, I have the following anecdote which made me really dislike it.
A friend has Manjaro and one day his nvidia drivers stopped working after an update. I helped troubleshoot over the phone, while looking over the wiki. For nvidia drivers they have their own wrapper around pacman.
Turns out there’s a different nvidia driver for each kernel version. Already a stupid design. So unlike arch where there’s 1 kernel package (the latest the distro offers) and 1 matching nvidia driver, Manjaro has dozens…
The wiki never mentions how to install or update the drivers manually with pacman or anything like that. It pushes their own tool, a stupid wrapper around pacman, which is supposed to manage this for you.
In my friend’s case, the tool failed. It was trying to run pacman but there was a conflict issue. But the tool didn’t show the pacman output, so we couldn’t figure out what the tool is trying to do, and why it doesn’t work. We tried removing the tool and re-installing, and all kinds of messing around with it. It failed to install the drivers, it failed to remove the drivers, it kept failing whatever we tried.
Eventually we figured out the naming convention they used for the packages (again not mentioned in the wiki), and manage to install the correct kernel - driver pair manually, using pacman.
Tl;dr: poor design, bad documentation, and they push their own crappy tools which hinder instead of helping
That’s not a stupid design at all. A nvidia kernel module artifact is only compatible with exactly one kernel ABI. Thus you need one binary nvidia package for each kernel you ship.
Arch also has one package for every kernel ABI they ship:
nvidiaandnvidia-lts.Though it should be noted that their design assumes that these two ABIs are the only possible ABIs which isn’t strictly the case as the zen, hardened or RT variants may sometimes lag behind their regular counterpart. That’s a stupid design if anything as it increases the friction of kernel ABI upgrades as a kernel package maintainer.
We at NixOS also ship the nvidia module for each of our ~50 kernel variants; all major versions of the Nvidia module compatible with that kernel in fact.
The only possible way to access these nvidia kernel modules is via a certain kernel’s
linuxPackagesattribute set that contains all packages that rely on a kernel ABI such as kernel modules or packages likeperf. That’s good design if you ask me but I’m obviously biased ;)I know you need a new nvidia driver every time the kernel updates, but why keep 50 kernel versions? My beef was them offering so many (outdated) versions instead of keeping the latest one which would make things very simple for users (imo).
These aren’t all versions per se but mostly variants, versions and versions of variants. For example, we have packaged the xanmod kernel which is a modified kernel optimised for desktop use but it has two variants: Main and LTS. We have packaged both.
Here are the names of all of our kernels currently to give you an idea (as a JSON list):
[ "linuxPackages", "linuxPackages-libre", "linuxPackages-rt", "linuxPackages-rt_latest", "linuxPackages_4_14", "linuxPackages_4_19", "linuxPackages_4_19_hardened", "linuxPackages_4_9", "linuxPackages_5_10", "linuxPackages_5_10_hardened", "linuxPackages_5_15", "linuxPackages_5_15_hardened", "linuxPackages_5_18", "linuxPackages_5_19", "linuxPackages_5_4", "linuxPackages_5_4_hardened", "linuxPackages_6_0", "linuxPackages_6_1", "linuxPackages_6_1_hardened", "linuxPackages_6_2", "linuxPackages_6_3", "linuxPackages_6_4", "linuxPackages_6_5", "linuxPackages_6_5_hardened", "linuxPackages_6_6", "linuxPackages_custom", "linuxPackages_custom_tinyconfig_kernel", "linuxPackages_hardened", "linuxPackages_latest", "linuxPackages_latest-libre", "linuxPackages_latest_hardened", "linuxPackages_latest_xen_dom0", "linuxPackages_latest_xen_dom0_hardened", "linuxPackages_lqx", "linuxPackages_rpi0", "linuxPackages_rpi02w", "linuxPackages_rpi1", "linuxPackages_rpi2", "linuxPackages_rpi3", "linuxPackages_rpi4", "linuxPackages_rt_5_10", "linuxPackages_rt_5_15", "linuxPackages_rt_5_4", "linuxPackages_rt_6_1", "linuxPackages_testing", "linuxPackages_testing_bcachefs", "linuxPackages_xanmod", "linuxPackages_xanmod_latest", "linuxPackages_xanmod_stable", "linuxPackages_xen_dom0", "linuxPackages_xen_dom0_hardened", "linuxPackages_zen" ](Note that some of these are aliases;
linuxPackages_latestis currentlylinuxPackages_6_6for example.)Each of these has the following
nvidiaPackages(modulo incompatibilities):[ "beta", "dc", "dc_520", "latest", "legacy_340", "legacy_390", "legacy_470", "production", "stable", "vulkan_beta" ](Again, some of these are aliases.)
This is useful to have because users might have hardware constraints. It’s not hard to imagine a scenario where a user might have a WiFi chip that only works with kernel ABIs < 5.4 and require the 470 nvidia driver for their old GPU. Packaging just the latest kernel and just the latest Nvidia driver would make this user unable to use their system.