The majority of other distros value package managers that allow for complex graph evaluation of dependencies, and the ability to roll back. This is granted with rpm and Deb, but not for pkgsource, which is a pretty lightweight format compared to those.
As for AUR, the major distros (Ubuntu, Debian, Fedora) support 3p repositories as well. The main concern is security. IIRC one of major complaints for AUR in the past was that it didn’t foresee a strongly secure distribution system.
It gives you a lot of convenience, auto updates, and dependencies. While it is nice being up to date by checking the git and making it by yourself it is much more convenient to have a package manager for it when you have many Make packages
Definitely the framework laptop, check it out, it’s completely modular, thin, light, performant, and insanely repairable, they even include qr codes on every part to help replace them and they will ship the device to you with no os for a discount and disassembled for a much bigger discount.
When does it hang? If it’s after you log into your user, you could try to instead when on the login screen open a shell by pressing Ctrl + Alt + F2, uninstall Nvidia drivers and install mesa drivers. Or maybe at least investigate further.
yeah it hangs way before the login screen. It basically hangs immediately after exiting GRUB and goes directly to a flashing cursor.
elltee explained it above.. MX Linux’s stable distribution is on kernel 5.10 and I need kernel 6.x (part of their advanced hardware support release) to get AMD RDNA3 drivers. So I’ll most likely be doing a full reinstall unless you know how to replace kernels in place.
I have no idea what MX is but maybe you could just compile your own? It’s not that hard. If you’re going to wipe the system anyway, it’s an opportunity to mess around a little and learn
Just remember to
backup old kernel AND initrd
have a rescue usb in case it completely stops booting up
choose modules that are used for your hard drive and filesystem as built into the kernel, just in case
you should be able to copy the CONFIG of the old kernel in the repo and use it (as the wiki proposes)
Not sure I’m feeling adventurous enough to be compiling my own kernel at this stage in the game. I might do that for some of my homelab machines as an experiment but my main computer really needs to be drop-in and go with full stability.
You should be able to boot into runlevel 1 easily and upgrade the kernel using the steps someone else already mentioned. No reason to reinstall the entire system.
Radeon 79XX drivers are integrated into the 6.X kernel. Kernels lower than that won’t really work. I don’t have a MX install any more, but I’d guess that’s the reason. You might try looking for the upgraded kernel. The nvidia drivers shouldn’t have anything to do with it.
Ahhhh fuck that’s it right there. MX’s normal distribution is Debian stable and baked onto the 5.10 kernel. I have to install MX’s AHS release for the 6.x kernel. I didn’t even think about that.
MX is fairly integrated and I’m not sure I could upgrade the kernel in place without destroying the install, assuming I even knew how to replace such a thing.
I’m too used to having all my old hardware lol. Thank you so much!
Just to clarify for anyone reading this good advice; you want to separate the root and home partition. That allows for reinstalling the OS in your root partition without losing data in your home partition.
Bringing this over from the site that shall not be named.
you probably need to update mx-packageintaller-pkglist, current version is 22.11.01mx21 if you don’t have that and it doesn’t upgrade you might want to change the repo (try MX Repo Manager) refresh and try to upgrade again, then after you update that it should be available in “Popular Applications” tab in MX Package Installer.
Dunno if it will help in your particular situation, but it might keep someone else from going to deddit in the future.
The short answer is that it's not anything like a trivial transition, and the best-case scenario appears to be precisely what you already have after you invest a substantial amount of time.
For ages, I thought it was actually outright broken, but it turns out that if your GPU is Nvidia, you need to pass a special argument and agree not to report issues. This is obvious on the CLI but invisible if you try to start an environment the standard way by selecting it in your display manager. Then you WLR_NO_HARDWARE_CURSORS=1 if you actually want to see a cursor. I'm sure proponents believe its Nvidia's problem that they haven't provided better support for a niche of a niche of a niche, but some of us would like to do actual work and don't care whose fault it is that something works shitty.
Then there is scaling for those of us with high/mixed DPI. Xwayland windows end up so blurry that you would be forgiven for imagining you had a head injury. I don't want to worry about replacing any app that doesn't support Wayland again to have no benefit of any kind. In fact since some apps are to interact with other people I can't just change them out regardless of wayland support. There is no plan to fix this.
Beyond that flameshot doesn't work correctly, replacing xcape with interception tools looks quite complicated, and depending on what app and version you use screen sharing may or may not break and other people give zero fucks about your niche environment.
ahh yeah, i heard nvidia can be a pain in the ass with sway/wayland. did you try sway when you still needed the –my-next-gpu-wont-be-nvidia flag lol?
but i remember reading about some new driver that made things much better… no idea though i’ve always been on intel-only laptops
mixes dpi
never had any issues with that. although i don’t think i even have anything going through xwayland at this point
apps to interact with ppl
zoom? or something electron-based?
xcape
i don’t know the tool, but the readme sounds like evdoublebind might work, it’s rock solid for me, replacing caps lock with escape on tap/alt om hold
screensharing
i feel you. when i started out with sway i had this horrible hack running a vnc server and a x11 vnc client on the same machine because pre-pipewire browsers could only share x11 windows haha
but nowadays that works really well. only zoom is still a little annoying ux wise. sometimes their screenshare popup just disappears leaving me with no way to stop a share gracefully.
They fixed a lot of load issues as of this morning so it might be okay. Lemmy.world itself already feels way more responsive. You have a point tho, might be better to host a Wine community on more FOSS centric instances.
Oh no doubt about that! I didn’t mean to imply lemmy.world was against foss. I just meant there’s already instances that are very FOSS focused that might be a better fit than lemmy.world which is already overloaded as it is.
I hope that solves lemmy.world essentially DoSing my instance at whatever intervals it would eventually take a turn to post all the stuff that had happened on it since the last flood.
How do you notice that you are not really awake yet? By thinking for several minutes about what LXD has to do with containers and then realising that you yourself had LXDE in mind.
Something to understand here, it is exactly the same with the automotive industry. It is almost never about the actual safety, let me explain.
If you work as a safety engineer in a company like Boeing the name of the game is to not be responsible for the safety of a component at all. You always hide behind some kind of certifications then always ask a contractor to do it. The contractor might be scared too so will ask for a subcontractor and so on until someone is in an obscure juridiction or brave enough to just develop the software like almost anyone else but just with someone rubber-stamping the paperwork.
The safety engineer will have the paperwork so for them, it is safe! If there is an issue this is not them.
So for them Linux is absolutely out of the question, who wants to sign a paper for it?
linux
Hot
This magazine is from a federated server and may be incomplete. Browse more on the original instance.