File: file.png (135 KB, 1476x244)

135 KB PNG

I want to pick a distro that will work the most seamlessly with KiCAD. KiCAD flat-out isn't fully supported by Wayland or it's X-Wayland compatibility layer, and Xorg is only getting more deprecated. KiCAD officially supported Debian, Ubuntu, and Fedora, but none of those are shipping with an option for Xlibre. As far as I can tell, I should be aiming for a derivative of one of these three distros, with the same package manager and all that, such that it's most likely to be bug-free. Is Devuan my best choice? There's been some rumours of Debian's apt and other core utilities getting rewritten in rust and still missing features that were in the original. Do I just use Fedora with third-party Xlibre, or are the team behind Fedora likely to end up making that incompatible?

I've been on Arch for a year, but there's like three somewhat different builds of it in across the aur and flatpak and support won't talk to you unless you're in the 3 supported distros.

Planning on setting up Gentoo soon and will likely change from Nvidia to AMD GPU some time after. Will there be any significant issues in regards to having to rebuild a whole load of shit related to the graphics drivers or am I just overthinking it? Don't have any experience in gentoo hence why I'm asking.

>>107726786
isn't it safe to always assume a change in graphics card, especially from N to AMD brings about complications ?
cant say i have any experience with gentoo but i always just backed up important files and do a clean OS install alongside swapping out the GPU for ease of setup.

>>107726804
On Arch when I had to switch before it was a matter of just updating the packages from what I remember. Just wondering if the process for gentoo is more involved due to the nature of the distro, compared to just deleting and installing a handful of packages.

>old thread is on page 11
What's the video editor situation look like in the year of our flying spaghetti monster 2026?
I need one program to add literal text^1 to like 30 frames of an existing MP4 video. I tried shotcut and it and doesn't work. I can't the video or something I don't understand what it's doing because it doesn't make a frame by frame preview. I tried avidemux and it's just ffmpeg with a GUI. Openshot has no installation candidate so it doesn't exist to me. And I tried to install kdenlive because that was the best in 2006, but it wants to install fucking wayland so I can't use it on my X and ALSA-only, no systemd builds.
^1 I tried doing frame by frame in the GIMP already, and then gluing it all back together with ffmpeg, and it came out like shit because I was fighting compression removing data from non keyframes.
If there's a way to do it with ffmpeg I could almost make do, but I need it shifted by like 20 pixels a frame in a random direction (following motions) which is why I was going to a GUI tool.

>>107726648 (OP)
computer related magazines still exist? i only bought them back in the day because they came with hundreds of megs of stuff on cd's which would have otherwise taken days to download

File: nvidia trixie.jpg (122 KB, 1280x720)

122 KB JPG

Why is the NVIDIA driver in Debian so outdated? Almost everything in Debian 13 is from 2025, including Mesa, so you'd think the NVIDIA drivers would be a little bit more recent; yet Debian 13 is using the 550 driver EVEN IN THE SID BRANCH, despite 550 being from god damn 2024...

>>107727335
Bro I think your solution is to install Windows XP and download Sony Vegas Pro 8.0 lol

>>107727474
Anon, that's a 1080. Pretty sure you have to use the older NIGVIDIA driver set.

>>107726786
it shouldn't be difficult, add in the support for your new card ahead of time, switch the cards, then remove the nvidia-specific stuff and it should be seamless.
i haven't switched from nvidia to amd in gentoo specifically, but it shouldn't be much different to switching amd cards which i've done, like moving from radeonsi to amdgpu, check your VIDEO_CARDS and LLVM_TARGETS/AMDGPU_TARGETS variables, adjust any relevant use flags (couldn't say which since i haven't use an nvidia card in nearly 15 years)
few packages are built for a specific gpu, so it's less of a deal vs. making a major cpu change (depending on how generic/specific you build your binaries, i think most people build their binaries specifically for their cpu given it's gentoo and you may as well)

>>107727335
>but it wants to install fucking wayland so I can't use it on my X and ALSA-only, no systemd builds.
couldn't you just use a nested compositor, i think valve's gamescope can do single window wayland applications on X
and then isolate systemd in a container or something
most of the container managers i've seen require systemd but i'm pretty sure all you actually need is kernel support for user namespaces so i think if necessary you could do it in bash if there isn't a non-systemd based one

>>107727487
Not my screenshot; I just went on Google Images and chose the most attention-grabbing result for "nvidia debian".

>>107727335
Blender.

File: CONFIG_LOCALVERSION.jpg (23 KB, 751x298)

23 KB JPG

>>107727474
Point releases are supposed to be outdated, that's "stability".
>>107726786
>having to rebuild a whole load of shit
idk about the rest of the graphics stack but what goes for the kernel build, it greatly depends on YOU. It's you who has to go thru the hundreds if not thousands of options and turn off anything that's useless. Compiling support for everything imaginable results in a kernel build that can take a whole day instead of doing a tailored build that compiles in 5-15 minutes.
But don't think of it as a Gentoo thing, think of it as a custom software part tailored for your specific hardware. I do these custom kernel configurations whenever I get a new PC, regardless the Linux distro.

>>107727541
Point releases are meant to BECOME outdated, and Debian 13 was otherwise full of relatively new software when it came out. If the AMD drivers (aka Mesa) were fairly new, and ARE very new if you use the backports, why are the NVIDIA drivers still so god damn old?

>>107727561
Maybe check the unstable repo?

File: Linux_Format_Annual_Volum(...).jpg (3.39 MB, 2000x2588)

3.39 MB JPG

>>107727460
>computer related magazines still exist? i only bought them back in the day because they came with hundreds of megs of stuff on cd's which would have otherwise taken days to download
They still exist. Recent publications too.

Reposting >>107726813 →
> Linux_Format_Annual_Volume_9_page_40__2026.jpg
> >>107726655 →
> Magazine from:
> https://web.archive.org/web/20251217225259/https://megalodon.jp/2025-1201-1116-51/https://archive.is:443/2025.11.30-050127/https://www.magazinesdirect.com/us/single-issues/linux-format
>
> Here's another image from it. What I learned from this print media: ChromeOS (proprietary Google software/OS) is based on FOSS ChromiumOS, which is based on Gentoo.
>
> (Photography info. Best to take photos of a magazine with either a scanner, which is cumbersome, or a camera which does great in low-light environments. Taking a photo with flash on results in pic related glare from the glossy paper; you can't see the text "Chevrolet" on the red car because the glossy glare is in the way. Editing: ratio is 8.5 by 11.)

That statement about Gentoo is on page 63. I wonder if some of the images in this magazine are AI-generated. (I was recently reading some other physical magazine; it had a story about a guy dying of ALS, and it had AI-generated images of his in it.)

>>107727594
Like I said, even in Debian Sid you're still only getting the 550 driver, just like Debian Stable. Debian Experimental does have the 555 driver but that's still old as hell.

>>107727474
>Almost everything in Debian 13 is from 2025
...Because debian 13 is from 2025?

File: 550 series.png (92 KB, 731x361)

92 KB PNG

>>107727662
Yeah, so why is it using the 550 series of drivers (which don't even work on the 50xx GPU's) when the 580 drivers already existed by the time Debian 13 launched? Maybe 550 is fine for the stable branch, but 580 REALLY needs to be in the backports and sid.

>>107727674
because to debian family distros stability means pick a package version irrespective of what the actual developers working on the project consider to be stable, bugfree, and atomically versioned or what users would need, and literally never update it aside from having volunteers who are at most vaguely familiar with the package codebase hackport security fixes back onto it
they're meant for the linux version of the retards who run decades EOL windows machines because they built their entire business on some shitty proprietary middleware or drivers by a company that went defunct in the 90s that only runs on 1 version of windows ME, they never had the source code, and refuse to even consider replacing it until it's too late and the computer finally dies or is somehow exposed to the network

I'm using proaudio I guess on pipewire, my problem is that I thought it was the same as alsa hw output (one audio source only, no resampling raw output etc) but I can play audio from multiple sources simultaneously so I know that's not what's being replicated. Am I doing something wrong? What even is proaudio then?

>>107727819
>irrespective of what the actual developers working on the project consider to be stable, bugfree, and atomically versioned or what users would need
The problem is that the software world jumped into the fucking void around debian. Not the other way around.
When you release a major point version you should consider it stable and backport to it for like 5 years. But node and shit broke that. That's why java software still targets 1.8, the actual stable baseline version.
But meanwhile shitters decided that every point version can break API and ABI compatibility breaking semver and and also meaning that each new version needs new debian release with a new build stack done for it.

>>107727918
>But meanwhile shitters decided that every point version can break API and ABI compatibility breaking semver and and also meaning that each new version needs new debian release with a new build stack done for it.
while arbitrary API breakages are annoying as fuck (fuck python in particular, even when they don't change anything it still breaks) i genuinely think ABI breakages at least should be encouraged
windows wouldn't be half the mess it is today if they hard broke and fucked over all the enterprise shitware using undefined/undocumented ABI dependent behavior
aside from one minor breakage in preview code from 2022 their C++ ABI literally doesn't change and hasn't since XP despite the library not even supporting it anymore
there's fuckloads of bugs that wouldn't break proper to specification API usage if the applications were just recompiled that cannot be fixed because muh proprietary enterprise shitware just has to keep running on the latest version
if they had intentionally broke their ABIs from the start it would have encouraged a lot better development and dependency management practice

if you can't recompile or have access to someone who can that indicates a potential long term problem

>>107727335

kdenlive transparency png other videostream

>>107727674
Because thats the version it had available in testing before the freeze
Why? I don't know. Was there a major change that happened in the nvidia driver recently? I know that some gpus got dropped recently

>>107727819
How does this apply to debian sid?

>>107728381
>Was there a major change that happened in the nvidia driver recently?
Yeah namely support for the graphics cards they released several months before Debian 13 released. But also 10+% higher framerates.

>>107721622 →
>>107726834 →
thanks dude
i was looking through other sites and couldn't find anything up to my standards of quality. as usual 4chan came through in the end

>>107727918
>When you release a major point version you should consider it stable and backport to it for like 5 years.
erm, nope. And if I'd do it, I would also backport bugfixes without a CVE, unlike debian that rather keeps the bug.
>node and shit broke that
That wasn't the case before those as well
>java software still targets 1.8
You get weird looks from me if you still require java 8.
>each new version needs new debian release
the fuck? debian not willing to updating the packages in their distro is debians own retardation, upstream is not at fault for that shit.

File: unstable.png (46 KB, 1613x48)

46 KB PNG

>>107728390
obviously their unstable branch isn't going to be out of date
and i was under the impression it's not really comparable to a real rolling release or even more properly maintained distro and was more like openSUSE factory i.e. potentially unstable (the real definition of unstable) packages so it has the opposite problem

>>107728429
>the fuck? debian not willing to updating the packages in their distro is debians own retardation, upstream is not at fault for that shit.
A new version, that requires regression testing, shouldn't come out more than once every 2 years. And with some exceptions that lines up with Debian release cadence. You have to think from the mindset of "integration testing shit is really fucking time consuming and expensive and error prone, so we want to do it as little as possible". If software has release cadences more frequently than every year with -1 version backports, it's a bad fit for Debian.
Debian's philosphy is all about shifting the testing burden to the upstreams and makes sense if you consider it as "Upstream should not release software more often than they can ensure it doesn't break this corner case of an edge case of some guys retarded workflow involving init.d" testing.

>>107728518
your ideas of how release schedules should be is rejected.

File: Start button.png (6 KB, 264x149)

6 KB PNG

How do I make the start button on Plasma wider? I'm trying to get an XP look going on but as you can see it's not working.

>>107728518
You are right. Sustainable update rate would be 3 to 6 months. Even 3 is too much. Just because git allows people to push a new version out (it works as their work drive... working 8h every day you'll get blind) doesn't mean it's ready.

>>107728518
Look mate, I'm fine with getting new packages every 2 years, but couldn't the NVIDIA driver I got THIS year have been a little more recent?

>>107728518
acquiescing to the demands of users like that makes things worse for everyone
more often than not in real life those faggots aren't some retarded shmuck but companies like google making inane demands of small projects
and if they are individual users they should be told to fuck off and kill themselves for demanding outsized support with a problem that's probably their fault

/fglt/ do you think snapshots are unnecessary in this day and age?
I've seen quite a few people stick to ext4 (or even switch to xfs) instead of btrfs on rolling release distros like arch or even derivatives that I feel are prone to weird changes more often.
I've been using btrfs with snapshots enabled for quite some time now and I'll be honest, the only reason I rolled back was to test out a different desktop environment and not have too many shits installed.
Have any of you been saved by snapshots? Do you think they're pointless as long as "you know what you're doing"?

>>107727335
I've seen a video just yesterday about some guy (I guess a videogame streamer? fuck if I know) struggling on his switch to linux and finding out that most video editors on linux are too basic or janky, davinci is unintuitive and made for professionals, and Blender is just the right thing in the middle.

>>107728941
You absolutely need them, bad upates are real as not every hardware configuation is tested. Just rawdogging updates is insane. Even better is not allowing updates to touch a running system.

>>107728941
I never was in a situation where snapshots would have helped me

>>107728941
i thought they were initially but they've saved my ass from serious system misconfigurations caused by attempts at experimentation on several occasions
being generally savy isn't necessarily going to help you and might even put you more at risk due to overconfidence if you go poking into strange places and get too handsy with the kernel, firmware, or drivers

>>107728941
i don't use snapshots to fix the system, at least i haven't needed to yet. i use snapshots to fix human mistakes. like recently i compressed my screenshots folder from png to jxl, and later realised that it doesn't compress png's to lossless jxl by default which i had expected. so i restored them from a snapshot
they're good for these kinds of "oh shit" moments, they effectively delay the permanence of anything you do

File: 1754890791605510.gif (992 KB, 498x280)

992 KB GIF

I'll create 2 VM's in virt manager and create a microk8s cluster out of them. I want to test persistent storage options for multi node clusters. Linux truly is the best because this is all free and open source.

>>107729203
Your relative performance is still tied with cpu versus vm bandwidth bottleneck.

>>107726786
It should be pretty seamless nowadays. In the old days you used to have different OpenGL providers and had to manually select which one to use (Mesa or NVIDIA) via an eselect module but now libglvnd takes care of that. The only issue you might have is if you use NVIDIA's xconfig tool then it may generate X11 specific configuration for NVIDIA that will break the Xorg Server if you don't undo that after switching.

For the most part you can just do what the other anons said and compile in the AMD support ahead of time before you make the switch and then afterwards uninstall the NVIDIA crap.

>>107727335
You can install Kdenlive via Flatpak or AppImage and that'll probably work on your special snowflake system.

>>107729240
Nvidia has always been for big boy software like Maya. If opengl works that was it for goyvidia. Ironically when first goyvidia cards came out they wouldn't support hardware overlay. That didn't last too long.
Only recently they have paid some attention to gaming and "Wayland".
Nvidia is ok but it's for workstations in linux.

>>107729228
Doesn't matter, I just want to test persistent storage options. How stable, easy and versatile they are.

>>107729228
Well I still agree with you that I have to keep in mind this bottleneck to not draw a wrong conclusion about performance. Good call anon.

>>107728587
You probably have to find the QML file for it and edit that. I think the size is hard coded.

>>107729260
Maybe there is relative difference but I think it won't apply too well. Sorry to be a bitch.

File: 1753094029642361.png (360 KB, 2608x922)

360 KB PNG

>>107729267
Well, it is what it is. Joining these 2 was easier than I expected, k8s is neat.

>>107729303
Kubernetes is pretty neat but unless you're a hyperscaler you don't need it.

I want to switch from Win to Linux. I've tried several Distros but my problem is that sooner or later a regular update breakes the distro. It's always either a problem with the bootloader or the kernel and won't let me boot into the system which is annoying because I also want my system to be encrypted.

What would be the correct way to backup both my system files and personal files so that I can idealy just roll back the last system backup (if even possible)?

Should I backup them seperately or together? What software/backup strategy would you reccomend?
(I can backup to a NAS on the network so an automated system would be best)

>>107729334
Why? I see the benefits in just single node "cluster". It keeps the things running by itself. And I feel like it isn't even that complex, it's pretty simple and the API (yaml/json resource specs) are very clearly and well designed IMO.

>>107729335
Rollbacks don't save you from bootloader corruption. And on many Distros they always keep the last kernel around in case the newly installed one breaks. On the Arch family you can get the same by installing the LTS kernel additionally.

That said bootloader problems mostly stem from installing Linux and Windows on the same harddrive.

so what would happen if I use the most recent version of yt-dlp on linux mint 22.2 instead of the official-package-repositories.list? would the videos I download also require a more updated version of mpv/vlc to play as well? I tested out a live USB of Mint to see what version each had.
-Live USB / mpv 0.37.0 / yt-dlp v.2024.04.09
-My Mint / mpv 0.41.0 / yt-dlp v.2025.12.08
My issue is that the video either does not play (VLC) or the video lags and my CPU threads go bonkers. Here is the video for reference:
https://youtu.be/co-TFLbaZAE?si=jMDhc1BkKUWG91R2 [Embed]
should I just use a distro with more updated packages? I was on Fedora before but I hated Gnome. I swapped a drive that had Fedora 42 and downloaded the same video using yt-dlp and the video played normally in VLC

>>107729335
>>107729374
>problem with the bootloader
Supergrub is good to have on a USB stick in case that happens

File: .png (17 KB, 1210x61)

17 KB PNG

>>107729382
can't mint use backports?
https://backports.debian.org/uploads/trixie-backports/

holy shit this captcha

>>107729382
yt-dlp is just a giant python script. As long as it downloads the video you don't need to worry about the version number. It seems however that there is something wrong with your codecs. Did you enable them during the mint installation?

File: 1766291000129120.jpg (2.6 MB, 3072x2304)

2.6 MB JPG

>>107726648 (OP)
My system locked up and I tried REISUB, but apparently Debian sets /proc/sys/kernel/sysrq to 438, which means no signals can be sent to processes. I guess this means the E/I don't work. Is there any security issue if I enable these? I was able to get the machine to restart, and fsck fixed a bunch of inodes, but I'm wondering what I should do for the future.

File: 86029901_p10.png (2.41 MB, 1736x2475)

2.41 MB PNG

Is there a decently affordable linux compatible printer out there somewhere? Preferably one I can buy used?
I want to hang memes and inappropriate images up in public.

>>107729343
But it gives you nothing you couldn't get with Docker or Podman, etc, and only brings you added complexity that's not needed or desirable.

>>107729540
Since CUPS is also used by Apple then if the printer supports Mac then Linux should probably work.

>>107729446
Potentially containers could send requests to /proc/sysrq if they're running as root. Besides that there's no real issue. I have sysrq always enabled on all my systems. The default most distros have of not enabling it is excessive in my mind. Yes, I get if you have some kiosk system it might be desirable to not enable it but most systems will want it to work properly. Sometimes there's no other way to reboot a system except to use the sysrq keys.

>>107729540
Any used HP printer will probably do.

>>107729446
Hej grønlandsanon

>>107729560
Docker and podman are not production ready.

>>107729641
They are if you know how to deploy them to production properly. They make a damn sight more sense than running fucking Kubernetes in a VM at least.

>>107729573
Can those use color?

>>107729771
If its specification sheet says so then sure.

This might sound crazy, but I don't think Linux runs any better or faster on my machine than windows. The CPU temps are the better on windows under light load. The RAM usage is slightly better on Linux, but both are low enough that it doesn't matter

>>107730047
Probably depends on what distro/DE you're using

>>107730064
Artix on OpenRC with Niri as the WM. No DE, only basic background processes, like bluetooth. I feel like that's pretty lightweight relative to other options

File: 2026-01-01_15-55_1.png (968 KB, 1278x1024)

968 KB PNG

lifelong wintoddler finally giving linux a proper chance. have to say im enjoying it so far. i wish i had a radeon card though.

>>107730047
CPU temps depends on the speed/frequency its running at. If Microsoft's power/energy settings aren't on performance mode then it may be running slower than Linux. Try playing with power profiles daemon or the cpupower tool.

>>107730047
how do you even measure this?

>>107730152
Feels.

>>107730176
felling what? Do you run something specific, is it resources used at idle or what?
Besides that, feels are not a good measure.

>>107730152
Wdym how? Just look at the temps and RAM usage on both OSs

>>107729700
>running a fucking kubernetes in a VM
I'm running it in VM just to test stuff locally. I don't get your problem, "fucking kubernetes" - I don't get this at all, there's nothing special about k8s, a single node cluster can be very lightweight and still offers a shitload more functionality than docker or podman AND it is not any more complex in fact it's simple.

>>107730196
To add to this: I develop on podman, then I deploy it to k8s, this works flawlessly for me.

>>107730196
>>107730204
Why would you not just deploy Podman if you're already using it for development?
You're obviously not getting any of the scaling benefits from a single node so what's it for?
It's API? Docker/Podman has an API too.

>>107730381
>muh scaling benefits
I'm getting other benefits and I'm also getting a single node "cluster" that I can expand to multiple nodes if I ever need more compute. The resource definitions for few deployments, services, ingress and pods are quite small and simple if you keep it simple and know what you're doing. Not to mention prometheus for log storage and search and cert-manager for automatic internal TLS certificate management and also external TLS via letsencrypt integration. These things are not available easily on podman or docker.

Lubuntu with "nothing running" (sddm, tor, GUI, etc. turned off): it uses 362 MB out of 3.70 GB of RAM. Pretty sure this is a better baseline memory usage than when Windws 10 was on this laptop years ago. I don't need any GUI or audio for what I use this black-colored laptop for. (You probably can't easily turn off the GUI and do the stuff I do in MS Wangblows 10.)

>>107730412
Docker also has different logging drivers available and you can absolutely use letsencrypt if you know what you're doing. You've got multiple options to make that work.
I guess being able to scale it later is useful if you want to but you're not Google. You're never going to need to do that and if you do then there are things like Docker Cluster that exist.

>>107730462
Why not just install the server edition of Ubuntu at that point?

>>107730462
Probably less if you sign in with x11 instead of wayland

>>107730471
Docker absolutely has features to make all that happen, but I find it more cumbersome than just deploying a production ready set of prometheus and cert-manager resources with few commands. For me k8s has been the obvious choice for my servers after I bothered trying it out, before that I used docker swarm because k8s seemed daunting - which it really is not, it is beautifully designed and simple. And I do have concerns of scaling in the future when it comes to my servers, the userbase is growing slowly but surely, sure you can just add more CPU and RAM to a single server but spreading those compute resources across multiple servers has redundancy benefits which I personally do like.

>>107730462
I didn't switch out NetworkManager for wpa_supplicant (yet?); that would use less memory, I think.

>>107730480
I could do something like that for headless usage. In the past I used the GUI, but no longer. All those gigabytes going towards GUI software could be used for a larger swapfile instead (if that helps and doesn't harm). Or, I could setup Arch Linux but quit before installing anything I don't need like a DE. Probably not much difference between server Ubuntu and said Arch setup.

>>107730494
>but spreading those compute resources across multiple servers has redundancy benefits which I personally do like.
But you're not doing that. You might never do that. That's a premature optimisation.

>>107730516
>That's a premature optimisation.
Which isn't a problem in this case because I personally prefer k8s in production anyways. You can keep using docker or podman, not my problem, but my own experience is exactly what I described in my earlier post: once I tried k8s I never looked back, single or multi node doesn't matter.

File: PGEE86bs_WM-maxresdefault.jpg (111 KB, 1280x720)

111 KB JPG

I have these files:
>5.0.0.10.qBittorrent-Enhanced-Edition-x86_64.AppImage
>qBittorrent-Enhanced-Edition-x86_64.AppImage

It's version 5.0.0.10 and some other version. Is there a way to run these qBittorrent AppImages with no GUI? Is there an AppImage for qbittorrent-nox?

Update - This could be the solution:
>https://github.com/userdocs/qbittorrent-nox-static
>https://web.archive.org/web/20260101155422/https://codeload.github.com/userdocs/qbittorrent-nox-static/legacy.zip/refs/heads/master
>A bash script which builds a fully static qbittorrent-nox binary with current dependencies to use on any Linux OS

BTW, today I reached a share ratio of 15.01 on 438.11 GiB downloaded of a folder. That's 6.422 TiB uploaded: maybe the most uploaded per single torrent I've ever done. More details in this TXT file (<s>I think I can delete one of the ~3 copies of this "[AOmundson] Fullmetal Alchemist Complete Series 1080p [Dual]" that I have</s>):
https://dymwizard.store/raw/jy-5aDiA49L-2ix9L75S5Gz7l1C0LlsmMzXO_Xmwq1c

File: VirtualBox_Ubuntu_31_12_2(...).png (1.64 MB, 2560x1282)

1.64 MB PNG

Figured I'd try trimming some of the fat off my Ubuntu VM.This folder seems like a good candidate, but I'm not 100% sure what exactly I can get rid of

There's also what may or may not be a bit of bloat in Firefox's cache2/entries folder, even though it's only about a GB or so.

If it helps, this used to be on 22.04 and maybe 20.04 before that, so there might be some leftover stuff from all that somewhere.

>>107731058
A /g/ anon said:
>you can always download https://github.com/userdocs/qbittorrent-nox-static/releases/latest straight to /usr/local/bin/

The latest one is qbittorrent 5.1.4 libtorrent 2.0.11 (64-bit one is "x86_64-qbittorrent-nox"):
https://github.com/userdocs/qbittorrent-nox-static/releases/tag/release-5.1.4_v2.0.11

Will probably try this out.

Migrating this question from /sqt/ to see if anyone here can figure this out:
I'm attempting to add the mozilla repos to apt (on Debian Trixie) and got confused when I reached the step on actually adding the repository to the sources file, where it has me using this command:

 
cat <<EOF | sudo tee /etc/apt/sources.list.d/mozilla.sources
Types: deb
URIs: https://packages.mozilla.org/apt
Suites: mozilla
Components: main
Signed-By: /etc/apt/keyrings/packages.mozilla.org.asc
EOF 

Is it asking me to run the whole prompt in the CL? Or only the first line, then input the rest of it? If it's the latter how do I save the changes? Attempting to input the info via nano and the prioritizing the repo/updating apt brings up the error that the file can't be read. Same applies to the step after, cause I think it's formatted the same way. Actual page for reference:
https://support.mozilla.org/en-US/kb/install-firefox-linux?utm_source=www.firefox.com&utm_medium=referral&utm_campaign=firefox-download-thanks#w_install-firefox-deb-package-for-debian-based-distributions

>>107731085
Hang on, the last portion of that question is a bit of a copy-paste abomination, what I meant to say is:
Is it asking me to run the whole prompt in the CL? Or only the first line, then input the rest of it? If it's the latter how do I save the changes? I attempted to input the info via nano but after prioritizing the repo and running apt update, it said the file cannot be read. The actual file that prioritizes the new repo:

echo '
Package: *
Pin: origin packages.mozilla.org
Pin-Priority: 1000
' | sudo tee /etc/apt/preferences.d/mozilla 

Ran fine and created the file.

>>107731121
>Is it asking me to run the whole prompt
The whole prompt, yes

>>107731068
Is this on top of Windows? If so then why not just use WSL

>>107731148
IIRC this was initially done to see how some programs run in Wine. I have found some issues with some programs, though part of it might be VM-specific. One program has at least one issue since at least version 9, but despite filing a report, I don't think anyone looked at it.

>>107731174
wine and wine inside a vm is not the same thing

>>107731182
Yeah, at this point I kinda get that. On the other hand, the bug I am referring to I have seen happen on real hardware as well.

FWIW, I have a laptop with not all that much on it. I've been floating cutting that in half and putting Mint on the remaining space. On the other hand, a weird bug I found with Ventoy when it comes to installing Linux alongside an existing OS has made me reconsider.

File: Screenshot at 2026-01-01 (...).png (66 KB, 822x366)

66 KB PNG

>>107731145
Sorry, but is there a way to save the changes made, I don't think I've ever actually used cat before. It leaves me on a hanging square.

>>107731208
Make sure there are no newlines or anything after the second EOF. Another way to do this is just to open an editor with root privileges and create /etc/apt/sources.list.d/mozilla.sources with the content

Types: deb
URIs: https://packages.mozilla.org/apt
Suites: mozilla
Components: main
Signed-By: /etc/apt/keyrings/packages.mozilla.org.asc

File: grew.gif (1.81 MB, 390x300)

1.81 MB GIF

How much performance loss and instability do you get out of using dash to panel and arcmenu on gnome?
t. reliability autist

>>107731245
Oooh, okay I was pasting EOF at the end of the file which caused it not to be read, Thank you so much.

>>107728429
Debkeks assume everyone wants to use Debian or enable Debian but can't because they're stupid or something. It really is a mental illness.

How do I frobnicate a list of numbered files sequentially, using list expansion in bash?
I wanted to do something like

$ frobnicate file{1..8}.frob

but frobnicate only works on the first filename argument and ignores the rest, so that won't work.

>>107731649
you can use xargs with a limit to number of arguments (-n)
like "echo file{1..8}.frob | xargs -n1 frobnicate"

ubuntu 25 fucking sucks.

>>107731809
I had imagined the solution to be a little bit simpler. But that works. Thanks. And now I finally have a real case for trying to understand what exactly it is xargs does.

>>107731985
Want to elaborate on that? Maybe contrast to 24?

>>107731985
it's ubuntu after all

Why does the Linux community hate Pop OS so much? It's a good distro, it works out of the box. It just works.
So why does everyone hate it? is that elitism bullshit where
>hrrrrr you didn't have to use a wiki a take 6 hours to install your distro like me!!!!
mindset?

Thoughts on OpenCode for the terminal?

>>107732292
It's like ubuntu, but even slower, when ubuntu is already too slow.

>>107732292
Everything is hated by somebody. I have never ever seen any particular hate towards Pop OS.

anybody has problems with kernel 6.18 and amdgpu?

>>107732387
No, but all I have is an RX580.

>>107732332
But it's not. I started using Linux with Slackware and moved through a ton of distros over the years to include SUSE, Red Hat, Fedora, Ubuntu, Debian, Mint, Arch, and now Pop.
Pop is not slow at all. It works right out of the box especially for gaming. I had my fun in the early days of compiling card and peripheral drivers, tracking down dependencies... I just want something that works now and using a VM machine for tinkering and fucking around.

SO I want a real answer by everyone hates Pop, but as someone who's used Linux for almost 2 decades, there is literally nothing wrong with it.

>>107732422
>It works right out of the box
I no longer trust anyone who says this about anything linux. At best it might "work right out of the box" in some special cases.

>>107726691
i don't use it that often and am not very good at it, but it seems to be working for me on Fedora with KDE and Wayland, what specifically is broken? not debating just curious

>>107728498
Somehow Debian sid still lag behinds compared to something like arch when it comes to rolling release. There's also a debian experimental branch which would be more closer to arch rolling release but arch does also have their own "testing" branch
>>107728518
This thing about debian testing every package for bugs in testing before it gets released into stable is starting to become a myth as i've experienced packages with bugs in the stable branch that never get fixed until the next new stable release.

>>107728941
Ive been using snapshots for my /home folder more than for system snapshots and never had to use system snapshots while /home snapshots have helped me out a few times.

>>107729334
>>107729343
>>107729560
Is there any advantage to kubernetes over regular podman?
>>107729641
>not production ready.
What's that even supposed to mean?

>>107732496
The thing I dislike about snapshots in /home is that a lot of software isn't suited it. Torrents, steam downloads, anything with a sqlite db in it. Like it won't kill you but its not ideal to be using CoW for a lot of user software. And I always forget to do chattr +C until after I started writing files there. Or, it's just unnecessary like in a git repo. It almost seems like it would be easier to make copy on write opt in. Once you have it turned off in a bunch of places snapshots lose some of their charm anyhow

>>107732387
system update running, let's see if I'll regret it

>>107732539
Honestly, k8s and docker cannot even be compared because they are different things. K8s is a container orchestration system while docker is a container runtime. You can configure k8s to use docker as its container runtime, so it's one abstraction layer above docker in that sense. TL;DR you tell k8s a state you want and it maintains that state across nodes (servers) and pods and containers and other resources within your cluster.

>>107732496
My problem with /home snapshots is the amount of shit I'd have to exclude (personal files, games, music, downloads, cache). But if I have to be honest, for snapshots to be fully effective in my case, it would have to take some /home folders, because I found more often than not that some config files would remain in there like a nuisance. I feel snapshots should be system + config. But I don't know how to do this in a way that doesn't involve creating 20 fucking subvolumes.

I keep getting weird files like .bash_history-01005.tmp in my homedir and just can't find a solution. What causes this and how can I fix this?

>>107732547
The irony is there is a certain filesystem with a form of opportunistic opt-in-ish CoW already.

NTFS.

When a volume shadow copy is created, the state of all files in the volume is frozen in the VSC, all file changes are redirected, and the filesystem goes back to being non-CoW when the VSC is destroyed. This is how backup programs are able to 'snapshot' a volume on Windows. I think you can access them on Linux, but there is no way to create them.

>>107731200
I'm using Linux right now which is installed on the same disk as Windows. It should work fine. I dunno about Ventoy though because I don't use it.

I think some people say that Windows updates can sometimes mess with the GRUB bootloader and you have to fix that to get back into Linux. I haven't had that. But if it happens I assume you can fix GRUB from a live USB or something like that.

>>107732617
You could just leave /home/ empty besides the configuration files and store everything else on 'work' partition.
As long as you don't go overboard with flatpaks /home/ can be quite small.

>>107732802
Yeah but I like having this /home place for everything. I like keeping things simple.
Also there's some other shit I'd rather not snapshot, like .mozilla (or .floorp or .zen or whatever the fuck a firefox based web browser is called).

>>107732649
I have this problem too. Not entirely sure I fully understand the details of how it works or why those .tmp files keep appearing, but apparently temporary command histories from your bash shell that get left behind when bash sessions are unable to write out histories to .bash_history for whatever reason.
Would be nice if I could figure out what makes my system unstable enough that those files keep appearing, but at least I'm pretty sure they're safe to delete.

Your .bash_history is the record of everything you type in the command line, which lets you repeat commands with with up and down arrows. You can show you command line history with the

history

command.

>>107732649
Look into .profile or .bashrc, maybe they load some weird stuff for your shell. Hard to say what's going on since it could be your shell customisation or the distro's, some like to do that.

>>107732952
I've read that far, but is also leaves me with no clue why the history might not be writable

>>107733001
doesn't seem to include anything weird, and should be the same as my previous system that didn't have that problem (but used different terminal etc.)

>waydroid shits itself for no apparent reason
>check and see an update is avaiable
>install update
>it didn't help
sometimes, I hate this shit

How does KMag KDE magnifier work? I click it and I just see a black window with a black background

>>107733079
>also leaves me with no clue why the history might not be writable
Like I said, I'm right there with you.

File: Screenshot_2026-01-01_14;22;25.png (216 KB, 1920x1080)

216 KB PNG

>>107733109

forgot pic

>>107732292
It had some stupid fuckups like uninstalling steam deleting everything. It's on Ubuntu 24.04 base despite 26.04 is right around the corner, plus cosmic is just another attempt of "Gnome but this time it's finally good".
Overall a pretty meh distro.

>>107733109
You will need to buy magnifier license.

>bought a laptop yesterday (picrel)
>install linux mint because its the only linux i have used before and have had zero issues albeit on a desktop
>plug in HDMI to my tv
>nothing shows
I just spent about 20 minutes running commands and it still doesnt work. xrandr shows "HDMI-0 disconnected (normal left inverted right x axis y axis)
" . the nvidia drivers are installed and updated and set to nvidia performance mode aka defaulting to the dGPU and not the iGPU. Can anyone help me here?


just in case laptop specs are:
>2560x1600 OLED 165HZ (but i currently have it set to 60 Hz)
>AMD Ryzen 7 260
>RTX 5060

my tv is 4K 60Hz

>>107733263
forgot pic lol but the specs are more important anyways

>>107732952
>>107732649
iirc this is done on debian by default when using bash. I forgot what the fix is, but it's easy to find with a simple web search.

>>107733263
>RTX 5060
Install something with newer software like Cachy or Garuda or Kubuntu. Mint with 2024 packages might be too old for your gpu.

>>107733285
>a simple web search
Yes because I definitely didn't think to do that already.

>>107733263
>>107733299

Yeah its either old drivers or xserver fucking up.

>>107733323
clearly not since I've literally found multiple sources on how to fix this in less than a few minutes

>>107733755
Then what keeps causing it?

>>107733360
>>107733299
i am about to try popos. is that what you mean by cachy?

File: Email_icon.png (25 KB, 512x512)

25 KB PNG

Im looking for an email client which will work with yahoo mail using oath what ever it is and imap. I have been using thunderbird for years but it currently won't work. I have K9 mail working on mobile with imap but want something for desktop.

I found this but wonder how long until it gets fixed: https://github.com/thunderbird/knowledgebase-issues/issues/117

>>107731283
On Debian 12 it did not crash at all, on Debian 13 it has so far not crashed at all. I didn't get accurate performance measurements with and without because in my opinion Gnome is fairly unusable without dash-to-panel, so I didn't even try.

>>107731809
That function is built into the shell

 for f in file{1..8}.frob; do whatever; done 

xargs is almost always the wrong solution.

if i want to back up my install with all my personal files can i just copy the root directory to an external drive

>>107734690
Right, I forgot the for loop syntax existed. It's so ugly it never sticks in my mind. Thanks.
When is xargs the right solution?

>>107729443
I did enable multimedia codecs during installation. I checked to see if I had mint-meta-codecs and I did. I'm gonna give up and say its probably an issue with my CPU not being able to handle these high bitrate videos. The videos only played smoothly in mpv with my laptop plugged in. VLC refused to play the video, only displaying a black screen, though the audio was playing. Though I could've sworn that Fedora was able to play the video in VLC without needing to be plugged in...I'll check. I appreciate you even taking the time to help.

>>107734721
When you're dealing with a very convoluted list generation that can't be done with find or a bash one liner, you either can't use parallel or don't care about parallelism or tracking errors, and you also have GNU xargs, because every other implementation is cancer.

File: 1766582612640645.jpg (37 KB, 397x400)

37 KB JPG

>>107726648 (OP)
Guys I'm retarded, I swear I turned on the "sound events" option for GTK programs somewhere within the KDE settings but I can't find it anymore. It's really annoying hearing a "pop" sound when I press buttons on a GTK program.

>>107734690
>xargs is almost always the wrong solution.
xargs can fork processes up to defined limits like the CPU count, and batch inputs for said processes so you can use it to create a work queue, good luck doing that with a for loop.

File: 1746299318005204.jpg (113 KB, 800x800)

113 KB JPG

I'd like to hear input from some people before I decide, I'm a noob.
I have a NAS on my home network that I use to store general media, and documents like medical stuff, taxes for backup and archival.

I'm currently on ext4 jbod 2x8tb, 2x16tb hdd. But, I want some redundancy in case a drive fails. Rollbacks in case of a mistake would be nice. I'm also clueless about the true usefulness of check sums for stuff like bitrot.

From what I've read, it seems like btrfs or ZFS would best suit me?
I'm on Debian 12, the system has a Intel i7-5930K (12) @ 3.700GHz, and 16gb of DDR4, 2400. 1gbe LAN, maybe 2.5gbe in the future.

It seems in ZFS, I would be able to raid1 2x8tb and 2x16tb and present it as one pool? I'd like to be able to add more disks naturally in the future too. Is there any hidden gotchas or somethig I should be aware of? It would seem ZFS is not within normal package distribution for Debian, I'm not sure if this is an issue. Would it be worth moving to another distro with native support for ZFS?

a few months ago I got my first linux distro, I somehow decided to go for debian 13, it ended up being a huge pain in the ass, shit would be complicated to install, and would regularly stop working

finally reached a breaking point today and wiped it out, and installed fedora, everything just works on fedora so far. it's amazing truly amazing compared to debian.

probably this comes at a cost of security/privacy or something, there has got to be a catch

>>107735542
>there has got to be a catch
The big one is that you're an unpaid beta tester for Red Hat's commercial product. I've been using it since FC-1 was released; and I've yet to have it bite me in the ass.

>>107735653
What are you really saying here? Maybe go back to windows or try Mac? serious question

>>107735477
btrfs, add all disks to the same filesystem, enable RAID1 (mirror) mode for data.
You lose 50% of the storage space since all data is duplicated, but this is the case with any RAID1-like setup.
LVM is the other option, again offering a mirroring option and the ability to mix and match disks as you wish (not as simple as btrfs but just as reliable).

>>107735653
You're helping Red Hat polish it's paid offering. Some people take offense at that, since they don't see a well managed and free - as in beer - operating system as a fair trade for doing that.
I don't freak out over it, so I use Fedora. I have since its first release. It hasn't let me down; I doubt it will mistreat you.

>>107735731
Was for >>107735699.

>>107735731
>>107735653

ok so what operating system should I use if I dont want to get cucked like that

>>107735739
Then you're only option is to write your own.

>>107735477
btrfs has a lot fewer restrictions about adding new disks and rotating out old ones out, and it's not out of tree. zfs enterprise larp is hugely overrated.

>>107735806

let me put it like this: the reason I have not yet switched from windows 10 to 11 is because I was hoping to find something that was doing less data mining on my data. Is the topic of beta testing for red hat relevant to that?

>>107735856
>Is the topic of beta testing for red hat relevant to that?
Not really. Linux distros are neither more nor less inherently trustworthy than Microsoft.
Finding replacements for the software you're used to using should be more of a concern.

>>107735960
>Linux distros are neither more nor less inherently trustworthy than Microsoft.


then why do people actually bother with linux? windows is not expensive considering it just works

>>107735975
I mean everything is harder on linux, I am now trying to install brave on fedora 43 and it's again a huge time sink, shit just does not work as the manual says it should

>>107735984
It's not harder. It just requires effort, which people today don't want to invest in. Their TikTok rotted brains can't handle anything that requires their attention for more than 30 seconds.
The sooner humaity is wiped out and we can start over the better

>>107735975
NTA, but inherent is a key word there. You can configure just about anything you like if you've got a problem with a piece of software. There's also a matter of the degree and amount of your data you give away. You won't (or at least haven't) seen any Linux distros recording screenshots of your activity or forcing your files into their company's cloud, for example. You can decide if details about program usage is an issue for you, and choose one that does less of that if it is.

Also, Linux runs way the fuck better on older hardware.

>>107736014
I'm not saying I am not willing to make an effort, I am just questioning if there is any upside to it all other than not having to pay a microsoft licence

>>107735975
>then why do people actually bother with linux?
In my case, the first computers I used for school and work were running Ultrix, DEC's version of Unix, so when when I could put an equivalent Linux based OS with most of the same tools on my own hardware I jumped at the chance.
Other people have other reasons.

>>107736021
What distro would you recommend if you are mainly concerned with privacy. I was led to believe Debian was really good for that but after several months of frustration with Debian and noticing on my second laptop that I never have those issues with Fedora, I changed debian to fedora, but now worried if it will come at a cost of privacy

>>107736061
Data privacy was my reason for switching from Windows and a concern moving forward for me, but not a chief concern. Anyone invading my privacy less than Windows 11 is potentially fine, so my bar is low. More than anything, other than not having cyberpunk dystopian monitoring and ad performance software on my machine, I want stability and to not have to screw with my system once it's set up, so I'm on Debian but probably would take a look at Fedora or Arch if I were inconvenienced by anything.

As for privacy... those concerns might be overblown. There's sometimes a spirit of privacy that bucks any and all monitoring which is admirable in a way, but also guaranteed to make your life difficult. If Red Hat is collecting some simple data to improve their product, then selling that improved product to someone else I might not have a problem with that depending on what they gather, but I'd have a problem with them selling that same data to someone else, as an example. It's worth researching what it is they actually take and what they can do with it; if you're low on the schizo scale it might not be important.

It's really surprising to hear that you had trouble with installation especially, Debian was dead nuts simple for me to get running and only gave me a bit of trouble when I changed some hardware. I have it on three computers and it just werks.

>>107732908
Mozilla moved to supporting the XDG standard last year (fucking finally) so that's not an issue now if you want to snapshot the config files in ~/. config/Mozilla but not the cache, etc.

Ironically they are better than Chrome now which stores its fucking databases in ~/.config like that's a totally normal thing it should be doing.

>>107736205
I think the issue was debian 13 being unstable and it kept acting up with proton vpn

>>107735725
>You lose 50% of the storage space since all data is duplicated

Yup that makes sense, I meant more like hypothetically if I had a pool of storage using RAID, the usage size is limited to the smallest drive in the pool. From what I've read so far that's only the case for stuff like RAID5, etc.? Like I said I only plan on using raid1 on identical drive sizes.

In regards to btrfs: I assume that'd it'd use the matching drives (8+8) (16+16) as raid 1 pairs, and present the pair as individual devices that are then added to the pool for 24tb total, but I figured I would ask to be 100% sure.

>>107735848
>btrfs has a lot fewer restrictions about adding new disks and rotating out old ones out
I think I'll do btrfs then.

>>107735848
>it's not out of tree
I didn't know what this meant so I googled it.
Is the issue here mainly that the out of tree module developers are at the whim of the upstream kernel developers, and shit can break at any time? So BTRFS would just be more reliable since Debian supports it naturally?

File: GigFV86awAEJ3J8 (1).jpg (2 KB, 40x44)

2 KB JPG

I recently bought a windows 8 era laptop with a pretty bad pentium and 4gb of ram. I was wondering what distro would be the best purely on the basis of getting the most performance out of the laptop. I'm basically a complete novice at using linux, but windows 10 ltsc isnt cutting it so I figured its time to branch out and try something new.

>>107736569
start with Mint XFCE

>>107736577
Don't do this, Mint is extremely cumbersome
>>107736569
You've got more than enough horsepower for any distro with the KDE Plasma 6 desktop environment, which will give you basically the Windows GUI except without eating up all your RAM. If you're not interested in tinkering, there's several choices! The question is what sort of update frequency do you want?
>I want updates very, very, very rarely
Go with Debian KDE
>I want updates maybe once a week or month
Go with Manjaro KDE
>I want updates ASAP
CachyOS if and only if your CPU supports x86_64_v3

I too had to install Linux on an old laptop and worried that I needed to install a lightweight distro. So I did, only to discover I had a comically large amount of headroom lol. Windows trains you into believing your hardware is inadequate when it is not. So I put a Plasma distro on this bitch and it's been smooth sailing ever since. In the Linux world, lightweight doesn't mean it'll run on a 10yo laptop, but a 20yo laptop, so I doubt you will find yourself stretched thin on the CPU side with Plasma. But if that does end up being the case, consider using the Xfce desktop environment instead.

File: brave.jpg (386 KB, 1079x1866)

386 KB JPG

>>107735984
>I mean everything is harder on linux
???
>I am now trying to install brave on fedora 43
There's multiple ways. First is this from the website; you paste this into the terminal and you are done:

sudo dnf install dnf-plugins-core

sudo dnf config-manager addrepo --from-repofile=https://brave-browser-rpm-release.s3.brave.com/brave-browser.repo

sudo dnf install brave-browser

The second way is you download the Brave Flatpak from your preferred app store. I believe there may also be an .rpm installers you can download from the GitHub, but don't quote me on that.

>>107736862
Thanks for the advice. Any considerations for running any of these on the same drive? As far as I've researched, all you have to do is partition out the drive for each and it all should be fine. I don't plan on updating my windows install ever so I don't think I will have problems with the boot managers fighting each other.

>>107732464
>what specifically is broken
Mainly cursor warping, that's a feature that Wayland doesn't care to implement, but also window positioning apparently:
https://www.kicad.org/blog/2025/06/KiCad-and-Wayland-Support/
Personally I run a relatively small 4K monitor set to 250% Scale in System Settings (also KDE and Wayland for the moment), and I frequently get issues with KiCAD and its cursor positioning (cursor appears to be in X position, is actually in Y position). Anything 200% or above causes this problem. When I put stock Fedora and its Gnome spin onto a flash drive and tested KiCAD, I got the same issue both times. Going to the flatpak fixed this for the most part, but I still have issues zooming to an undesired coordinate, and the render resolution is diminished with no openGL instance found. I'm not certain all of my problems are caused by Wayland not playing ball, but at the very least if I can remove it from the equation their support staff might actually pay attention to me. They also said "arch can be weird" so that's also a variable I'd like to be rid of by going to something closer to Ubuntu, Debian, or Fedora.

If you have a few minutes, set your monitor's scale to over 200% and see if you get any problems.

>>107737433
Why would you intentionally buy hardware that needs anything but 100 or 200% scaling?

>>107737573
If they have some 13" laptop I can see that (great little portable machines but always a nightmare with their displays).
They said monitor though. I agree with you. I don't know why you'd intentionally buy such a small 4K monitor like that. Must have been a mistake.

>>107735739
Kubuntu

>>107736444
>In regards to btrfs: I assume that'd it'd use the matching drives (8+8) (16+16) as raid 1 pairs
No. btrfs mirroring works at file level so your disks can be of arbitrary sizes and it will work optimally as long as no one disk is larger than the sum of all other disks.
>shit can break at any time?
Not with Debian, since it maintains the same kernel version throughout its lifetime. And not with Ubuntu since Canonical fully supports ZFS.
This would be a concern with Arch (since it's rolling and gets new kernels fast) or Fedora (new releases happen as frequently as Ubuntu but without the commitment to support ZFS).

>>107736569
Kubuntu 25.10
No matter what you choose, create a swap file 2-4 GB in size, you'll definitely need it.

>>107737647
btrfs raid works at the chunk level, which are 1G in size. regardless you can mix drives with 1G granularity, so like an 8+8+16TB raid 1 is fine, you'll get the full 16TB to use

>>107737573
>>107737618
I want high pixels per steradian, what's wrong with that? Exact integer multiples shouldn't matter, it's still dynamically rendering a 4K frame, just that the UI elements are scaled up to not be 2mm high. I'd be fine with 300%, though it is kinda goofy, but that causes the same problems. I like not being able to see individual pixels, maybe owning a macbook has poisoned my mind. I wasn't expecting to have to use 250%, but large UI elements are actually nice. You might even want more than 300% for raspberry pi projects.

My question is why my different monitors with different pixel densities have to be scaled to different levels in the first place, as if a pixel has some standard size with respect to UI elements. Instead everything should default to scaling to have pixels the same size with respect to that pixel's size on the monitor itself. There should be a slider entitled "pixels per millimetre", not some arbitrary scale percentage limited between 50-250%. Seems like a no-brainer, idk. Maybe there's a utility out there that does this, especially for people using linux on laptops with extra screens plugged in.

>>107738023
Additionally, some applications don't even respect this scale in the first place, and appear miniscule, as if set to 100% scale. If I knew I'd have this much trouble getting high-DPI monitors working I'd have bought a lower resolution monitor. I didn't even think 27" was that small. To get those 100%-only applications running at my preferred scale, I'd have to use a 1080p 32" monitor, which seems unreasonably grainy. But maybe there's a hacky workaround for scaling those uncooperative applications, or using a different window manager might do something. Suggestions welcome.

For reference, the Submit button here on 4chins is 9mm tall when set to the 250% scale. I sit between 40" and 28" from my screen.

>>107738023
>Exact integer multiples shouldn't matter
is / aught
>as if a pixel has some standard size with respect to UI elements
When you're talking about PC content it's supposed to be 1/96 ~ 1/110 linear inch square. Nobody since the early 90s steps outside this.

rocminfo shows nothing in chroot but works in the parent environment, why's that?

# rocminfo 
ROCk module is NOT loaded, possibly no GPU devices

Tried systemd-nspawn and conventional chroot but nothing.

>>107738225
did you bind the host /dev?

>>107738324
Obviously.
>/proc
>/sys
>/dev
>/dev/shm
>/dev/pts
>/run
There's more crap under /sys on the host though.
>/sys/kernel/security
>/sys/fs/cgroup
>/sys/fs/bpf
and more, idk if relevant or not. Didn't mount those.

>>107738374
i'm not too familiar with loading modules in a chroot. are they loaded on the host? are you loading them from the guest? if the latter, are the guest's modules built for the host kernel version? maybe check if dmesg in the guest has anything relevant

>>107738225
Do you have the kfd device in your chroot?

stat /dev/kfd

>>107738467
>>107738225
Also /dev/dri is probably needed

>>107738437
AFAIK kernel and modules are 100% the same. I guess the "module" in the error message means something else.
>>107738467
>>107738473
Yes, they exist. What are those even?

>>107738481
The kfd device is what's used to interface with the GPU for compute workloads and /dev/dri is for render nodes.

If you have that then I'm not sure why it doesn't work. I can use ROCm in an Archlinux chroot on my Gentoo host system. I would guess maybe your kernel is too old?

>>107738481
should double check they're the same. that is ensure the modules folder in (chroot) /var/lib/modules is the same name you see when running uname -r

How do i make my computer not flash black when booting up? I get the grub menu, then after that it does the loading... text and then it goes black, then kinda gray, then black again, then boot logo. Why can't it be a boot logo all the way through?

>>107733282
hehe

>>107737108
Most distro's installers will let you choose to have a particular drive as the boot drive (where GRUB will be), including an existing boot partition that already has GRUB in it. That way you can select any of your distros to boot into when turning your computer on.

>>107735731
>>107735653
Why do people deride Fedora as "RHEL beta testing"? You act like Fedora ships broken updates and makes you troubleshoot them lmao

With a good password, how safe is a LUKS-encrypted CD or USB against a robber doing brute force against it?

Not sure if this is the right thread but, I've had a screen flickering/stuttering problem on Ubuntu 24.04.13 Wayland (also happens on Xorg) for a while now. It mostly happens if I lock the screen and log back in after a few minutes, its stuttering really bad and goes back to normal in a minute. But now it just started happening as I'm typing this and it won't stop.

Kernel version: 6.14.0-37-generic
I'm using the integrated i7-12700 graphics, with the i915 driver.

>>107739087
You should start by telling which distribution you have because the solutions of some distributions will break others.

>>107739636
Fedora.

>>107739630
Tech support is /wsr/ but regarding your question you're going to have better luck by telling which distribution, which desktop environment, and lockscreen application (could be the default).

>>107739665
Ubuntu, Gnome, default.

>>107739654
Fedora uses Plymouth for boot logo afaik and your issue could be kernel parameters. You have to set it to include the "rhgb quiet" option if you don't want to see text. It could also be a plymouth theme issue and in that case you'd have to regenerate with "sudo plymouth-set-default-theme fedora -R".

However, note that distributions like Fedora, Ubuntu, etc. are not meant for you to go around making lots of changes. Some seem to work and then when you upgrade your OS it will get some unexpected configuration and fail the upgrade.

Hiding boot text might be safe idk but if you want to do lots of custom configs you would use something tailored for that like Arch Linux, NixOS, Gentoo, etc.

My PC uses CSM, if i disable it i cannot use any of my main drives with a few different OS's. Do i need to reinstall the OS's to get secure boot to work since i cannot use CSM with sec boot?

>>107739654
Also see this: https://wiki.archlinux.org/title/Silent_boot

It explains how all of this works. It's for Arch but it's more detailed than Fedora docs.

>>107738437
>>107738504
Kernel modules in a chroot are irrelevant. Only userspace stuff matters.

>>107738225
Do you have the rocm runtime installed in the chroot?

>>107739683
It's probably this problem: https://wiki.archlinux.org/title/Intel_graphics#Screen_flickering

Several actually. Intel graphics is sadly buggy even in windows.

>>107739738
>Do i need to reinstall the OS's to get secure boot to work
No, it should all just work. Are your disks using GPT or MBR partition tables?
What's the OS that doesn't work?

>>107739805
>Kernel modules in a chroot are irrelevant. Only userspace stuff matters.
I figured the same, damn those nonsense posts.
>Do you have the rocm runtime installed in the chroot?
Absolutely no idea lol

How do I train an ai from scratch I want to train it on only dictionary and grammar information.

>>107739919
Umm I posted this in the wrong thread by mistake

>>107739919
LLM AI doesn't work like you think. You'd have to make your own dictionary and grammar AI yourself. The LLM AIs you are out there are generic.

>>107735477
>distro with native support for ZFS?
I wish that were a thing.

>>107737647
>Canonical fully supports ZFS.
False.

Why is there so much hype for Bazzite right now? Isn't Nobara the same thing with more and not being immutable? Is it just the steam big screen thing that people are after?
Can't you make something like Mint function the same as Bazzite? Am I missing something? I'm new to Linux.

>>107740010
Same reason that made you ask. It's the current meme that generates interaction/revenue bait. Just like saying "Arch + Debian is all you need, also fuck Red Hat" did 2 years ago.

>>107740010
I tried it as one of my first Linux OSes last year (though I tried Ubuntu like 15 years ago to see what it's all about), and I just got so annoyed with everything. Nothing seemed to work the way I wanted it, nothing installed easily. I genuinely don't get what people like about it. Maybe if I tried now, knowing a few more things about Linux, I'd feel different though.

>>107740010
Bazzite follows the current best practices

>>107740010
Because people in general work with emotions are feeling, not brain power. How many have you heard saying they're waiting for SteamOS to switch to Linux? In people's logic they know SteamOS and trust Valve so that's what they want.

If you're going by that logic CachyOS is better than Nobara for gaming since it gives you more flexibility and performance. Bazzite advantage over Nobara is that it's a almost zero effort maintenance distribution, while Nobara is a tuned Fedora for gaming.

>>107739810
Unfortunately it didn't work, it's still flickering after reboot. Maybe I should update the kernel?

>>107740041
In terms of?

>>107740053
Yes. Updating everything is in general a good idea for any drivers. If that doesn't work your best bet is asking the i915 maintainers but you'll need very good technical knowledge because they'll ask if you used the lastest kernel available, debug logs, etc. I'm not sure that's going to be easy in Ubuntu desu since it's not really made for custom kernels.

>>107740055
Non immutable distros are just archaic and dangerous in 2026, especially in the hands of n00bs (99% of people). This simple change in how distro is managed forces them to act properly instead of messing with the system. They don't even try to reinvent the wheel as Bazzite isn't technically a distro, but a Fedora Atomic image. All the good choices of embracing new tech in Fedora trickles down to Bazzite.

>>107739941
I just want to feed the ai dictionaries and basic information and not shit like current pop culture or shit from reddit etc is that possible I don't have any clue how this stuff works

>>107740091
>>107740049
Thanks.

I have these things running ("ps aux"), don't think they need to be running all of the time:
>root /usr/bin/python3 /usr/share/unattended-upgrades/unattended-upgrade-shutdown --wait-for-signal
>user /usr/bin/python3 /usr/share/system-config-printer/applet.py
The printer one says 'cups.require ("1.9.42")' so that's about physical printers.

>>107731084
I'm using that software now. It's dumb that it doesn't use /etc/hosts like older versions of qBittorrent:
>"WebUI: Unable to bind to IP: b.lan, port: 7000. Reason: Unsupported socket operation"
>
>******** Information ********
>Unable to bind to IP: b.lan, port: 7000. Reason: Unsupported socket operation
>To fix the error, you may need to edit the config file manually.

>>107740094
Your post reminds me of this thread I was reading:
>https://bbs.archlinux.org/viewtopic.php?id=285819 - "[SOLVED] caca / aa support mplayer / mpv / vlc / Multimedia and Games / Arch Linux Forums"
>[second post in the thread,] 2023-05-15 15:12:19:
>Here's a pro-tip: stop asking ShitGPT stuff - it's full of chat and you'll only get garbage for answers.
>It's not intelligent in any way shape or form but a reddit driven bullshit generator.

>>107740094
No. Not with LLM AIs. Your AI would be very stupid since it wouldn't have enough data. The large in LLM has to be very large. You'd have to make an AI specifically for your use-case. The best you can do with popular LLM is exclude certain sites.

>>107740094
To get any practical use out of it you need to feed it large amounts of text so it "learns" how to form sentences. It is next to impossible with "ethical" data. Also you need data centers worth of computing power to train it. That's why most "new" models are just fine tuned off of existing ones.

Ubuntu.com is offline now:
https://www.isitdownrightnow.com/ubuntu.com.html

And it's back up soon after I saw that message from isitdownrightnow:
https://ubuntu.com/download/server
>The latest version of Ubuntu Server, including nine months of security and maintenance updates until July 2026.
>Download 25.10
>1.9GB

>>107730125
I wish I was using a Ubuntu-based system instead of what I'm using now. Come to think of it, I could run what I'm thinking of in a Virtual Machine with Ubuntu Server (or do some other thing)...

>>107740166
Fixed by changing "WebUI\Address=b.lan" to "WebUI\Address=10.0.0.232" in ~/.config/qBittorrent/qBittorrent.conf - had to restart it for the change to take effect though.

>>107729603
>Hej grønlandsanon
"Hello Greenlandanon"

ive been talking to gemini aI today..

she convinced me to try out AntiX...

im not a advanced user btw.. she said she can teach me gamedev, lmms

good bye

>>107740347
Noooo anon don't go! Dont you know all LLMs lie?

File: 4d2cb9d5010d4ae3baaf52337(...).png (14 KB, 360x360)

14 KB PNG

Ubuntu 25.10 live server amd64 ISO file (2.12 GiB)

magnet:?xt=urn:btih:ZS6UPIYKLIJ2KJQCSXSL2ZOAHASE5HPQ&dn=ubuntu-25.10-live-server-amd64.iso&xl=2283360256&tr=https%3A%2F%2Ftorrent.ubuntu.com%2Fannounce

magnet:?xt=urn:btih:ccbd47a30a5a13a5260295e4bd65c038244e9df0&dn=ubuntu-25.10-live-server-amd64.iso&tr=https%3a%2f%2ftorrent.ubuntu.com%2fannounce

https://releases.ubuntu.com/25.10/ubuntu-25.10-live-server-amd64.iso.torrent
https://nutbread.github.io/t2m/

>>107739829
I don't know what they use but GPT seems familiar.
OS is openSUSE and windows 8.

>>107739611
Nobody is recovering LUKS encrypted anything unless it has a weak password. Maybe the US government, but even they seem to have rigged everything so they never have to attack FDE directly.

>>107739271
They only use rinkydink hobby distros where the scale of RedHat is incomprehensible. It's like how Euros make fun of stick houses because they imagine living in them without climate control.

>>107737433
i see, i don't need scaling myself so that's why I've never encountered that

>>107735739
Rocky or Alma Linux. They're based on the final product of the Fedora - CentOS Stream - RHEL pipeline. You're doing netori with Red Hat this way and avoid being ntrd.

Ubuntu 24.04 LTS (Noble Numbat) Cloud Image:
https://portal.cloud.hashicorp.com/vagrant/discover/cloud-image/ubuntu-24.04

I think the usage of this Vagrant cloud / Vagrant Box thing is you can headless-ly run a virtual machine running whatever OS. Then you can ssh into or enter into that running OS. Install for Linux: https://developer.hashicorp.com/vagrant/install -> https://releases.hashicorp.com/vagrant/2.4.9/vagrant_2.4.9_linux_amd64.zip

>>107740426
- Downloaded and running that Ubuntu 25.10 Live Server OS in VM.AppImage. At the welcome screen, go to Help > Enter shell.
- (Made a user with a password: no wheel group, so run "usermod -aG sudo user". Made a /home/user folder owned by user.)

Oddly, after switching to the non-root user, there's no history. You can't press up or down to run previous commands. More importantly, trying to install this thing -- https://docs.arweave.org/developers/mining/getting-started/install -- and running "./bin/arweave wallet create rsa ." failed/crashed. It complained about the file descriptor limit and huge pages settings. Even if I changed those to what it wants I think it would still fail. Also it said it's for Ubuntu 24 and not Ubuntu 25. Lastly, I don't have much free storage space to contribute to this anyways: only have like 2300 GB free on other computers (only 500 GB of that is usable as I could move that SSD around).

Why a stable 6 month release schedule over a rolling one for regular users? Rolling ones with good quality checking are very reliable. The only advantage i can think of is that a release schedule allows is that it has a reduced download size of one update once or twice a year, so you probably shave off a lot of gb's.

>>107741000
I was able to use this vagrant software before (maybe in Ubuntu); it ain't working today:
>$ ./vagrant
>bash: symbol lookup error: /tmp/.mount_vagranedIFne/usr/lib/libreadline.so.8: undefined symbol: UP
>$ # ran on Arch Linux

All roads lead to running the Ubuntu 24 OS on real hardware (not a virtual machine).

>>107740805
good old 5$ wrench works for them as well

>>107732649
most file managers have ctrl+h as the toggle to hide hidden files

>>107732752
ntfs supports a lot of things that windows either doesn't support or doesn't expose. like it supports symlinks, but i have to use a third party explorer extension to actually use them

>>107741082
Why do people feel the compulsion to make those smart-ass statements over and over and over? Have you no imagination?

>>107741129
I wanted to post it too but I managed to control myself

>>107734077
i use claws-mail myself, though i use it with my own mail server. i couldn't say specifically if you can use it with yahoo/oauth2

The Mint forums now require login to even read the threads. Is that just me, am I missing something?

File: file.png (364 KB, 552x552)

364 KB PNG

been considering switching to linux however i do recall that intel CPU's and Nvidia GPU's sucks massive ass on linux

should i upgrade to a full AMD system or wait for steamOS to drop on pc?

>>107741211
Intel CPUs are just as good as AMD ones. Might have to deal with the whole E/P core debacle but there are already solutions for it.

Nvidia is "fine". Yes they currently face lower performance in certain DX12 titles but that will be fixed this year. The rest are some quirks that also AMD is not free of. There is no "just werks" for any GPU vendor for every use case.

File: 7a2835a8b8c96242b23863789(...).png (187 KB, 4500x4500)

187 KB PNG

It's cool that you can ssh into a VirtualBox runtime for Ubuntu 25.10:

>>107741000
>use case
I can use VirtualBox-*.AppImage and ssh in to it (don't need vagrant for that). Just open up the settings for that ISO file in VirtualBox, then go to Network > Adapter 1 - should be set to NAT > Port Forwarding > add a rule named "ssh" with "Host Port=2222" (or any port not in use) and "Guest Port=22" and everything else in the rule is blank. Then, from the non-VM computer:
>$ ssh user@localhost -p 2222
>[...]
>Welcome to Ubuntu 25.10 (GNU/Linux 6.17.0-5-generic x86_64)

>>107741129
Car battery attached to nipples. There. There's your imagination.

>>107741169
>The Mint forums now require login to even read the threads
Bad. Bad Mint. Linux Mint forums are worse because of that change. https://forums.linuxmint.com/

>>107741169
Guess that's their way of dealing with AI crawlers sucking their bandwidth.

>>107741309
I see, which Linux is the most retardproof/newbie proof? i dont want my computer to detonate like a middle easterner just because i fucked a command up.

>>107741339
Bazzite if you do the gayming, Mint otherwise.

>>107741000
>Oddly, after switching to the non-root user, there's no history. You can't press up or down to run previous commands.
I found out why. In /etc/passwd that user was set to
>user:x:1002:1003::/home/user:/bin/sh
Notice it's "/bin/sh" and not "/bin/bash". (Now fixed.)

Why does it initialize new users with a login to sh and not bash? Odd.

>>107741328
It would be better if whoever is running that site had the ability/resources and willingness to make everything public and open web data: like it was before.

>>107741373
Why would anyone be willing to pay for the bandwidth that AI(Actually Indians) is using to train their models? Especially if its like 90% of the total bandwidth on the site.

File: omarhee.jpg (409 KB, 1080x1054)

409 KB JPG

Linux bros this is not fair
I spent countless days and nights learning this obscure stuff so I can be different from mac and windows users, and now thiw happens?
How can we continue to gatekeep our secret club?

>>107741082
He said a thief not a kidnapper.

>>107741309
>but that will be fixed this year
lol

>>107741508
The cause was identified half a year ago. Even Nvidia isn't that slow.