File: Debian-OpenLogo.svg.png (70 KB, 960x1273)

70 KB PNG

It just works.

Does coding naked make me a powerperv?

File: 20260704_23h30m32s_grim.png (152 KB, 1600x900)

152 KB PNG

>install guix package manager
>download MATE-terminal
>it downloads a bunch of non-related packages
>MATE-terminal isn't on my system
>but MATE-calculator is
OK...

>>109204038
Best distro ever. Except when you try to uninstall something big like KDE. Totally trashed my system. I have to either reinstall or manually purge packages.

lame question perhaps but is it possible to disable tearfree option on xlibre? and if so, is it possible to disable tearfree say just before a game is launched, like you'd do the compositor?
i found this project https://github.com/gabrielcapilla/xtearfree that i will try later; right now i cannot test all this I am merely curious if anyone else have explored this

>>109204289
Fixed it by manually removing packages while referencing a dpkg selections export that I took before installing KDE. Is there a better way to do a clean DE uninstall on Debian?

can the noatime mount option actually cause any issues?

Does anyone have any good resources on making a custom kernel (on gentoo)? I have been trying to find some but unsuccessful and I don't trust AI with my own kernel

What do you guys think of Bluefin and immutable distros in general?

I recently tried it and I think it should absolutely be the defacto normgroid recommendation. Much better choice than stuff like Ubuntu and Mint for non-tech people.

How do I compile ZFS directly into the kernel? There was some anon here who said he has done it but I didn't ask at that time, now I really want to do this.

>>109204344
This is possible but trying to do it before/after a game automatically will be painful, but you could use that tool to do so manually. I would honestly just disable tearfree completely and then get a compositor which enables/disables if a game is running.

>>109205016
https://wiki.gentoo.org/wiki/Kernel

>>109205111
Yeah they're basically a go-to for people that don't truly want to learn Linux and have it feel more like macOS or something in the weeds. Like giving people Bazzite/Aurora who aren't true linuxheads is what I recommend these days.

>>109205016
Have you tried the Gentoo handbook?
https://wiki.gentoo.org/wiki/Handbook:AMD64/Installation/Kernel#Alternative:_Manual_configuration
Try using sys-kernel/modprobed-db for a starting config then modify it from there.
Also see:
https://wiki.gentoo.org/wiki/Kernel/Configuration#Configuration

>>109205111
They're objectively a better approach to OS design. They should become standards as soon as possible.

>>109205436
Kernel either has zfs support or it doesn't have it.

>>109204944
It shouldn't. I use it for root, work, and home.
Bigger impact for your drive's health is something like moving web browser's profile to ram.
>https://wiki.archlinux.org/title/Firefox/Profile_on_RAM
It's also an incentive to clean it up. It's only ~300Mb of extra ram overhead but Librewolf does zero disk writes unless I'm downloading something of course.

>>109204034
How retarded is uploading my system configs to github? Will glowies start rubbing hands?

>>109205016
Configuring a custom kernel is easy
>unset a kernel flag you think you don't need
>wait 30 minutes recompiling the kernel
>boot it
>fail
>put the feature back on
>repeat with another flag
There are hundreds of them, good luck

>>109205828
if it has passwords or API keys, very retarded
if it's literally just configs, no one will give two shits

File: Screenshot 2026-07-06 021002.png (62 KB, 1497x733)

62 KB PNG

>>109205632
https://blog.woralelandia.com/ten-1-reasons-to-avoid-the-immutable-desktop-en.html

>Lately, it seems the industry is aggressively pushing immutable, image-based operating systems for the desktop. Shipping a cryptographically sealed, read-only root filesystem sounds super fancy and elegant on paper, but in reality, it's a solution looking for a problem, unless you're running a smart fridge or a corporate kiosk.

>Let me be perfectly clear: immutable systems are the correct choice for single-purpose appliances, point-of-sale terminals, and sealed server nodes. If your hardware only needs to run a fixed set of remote directives without any user intervention, then yes, a mutable root filesystem is just begging for trouble.

>But trying to force this architecture onto a general-purpose desktop? That is a spectacular regression. It wraps the user in a straitjacket, offloads system complexity onto you, and strips away ownership, performance, and actual security in the process.

>Here is the engineering reality of why the immutable desktop fails the general user.

>The Decentralized Trust Fallacy
>In a traditional mutable distribution (like Fedora or Debian), the supply chain is centralized. A dedicated, paranoid security team audits and patches the core repositories. Immutable desktops, however, force you to rely on containerized app stores (Flatpak or Snap). This shifts your trust from a central team to thousands of random, unvetted package maintainers. If a developer gets tired of maintaining a popular app and a malicious actor takes over the namespace, they can silently push a credential stealer directly to your machine. It gets delivered instantly via background updates, completely bypassing distribution-level QA.

>>109205632
>>109206080
>Furthermore, standardizing the exact same OS image across millions of installations creates a beautiful, uniform monoculture. If a zero-day is found in that specific image, every single user on Earth gets compromised by the exact same payload. Congratulations! On a mutable system, your messy "state drift" (different library versions, weird compiler flags, custom packages) acts as a natural defense system, frequently causing exploit payloads to crash and burn. Diversity is security, compa! ;)

>Weaponized Negligence (Zombie Dependencies)
>A security disaster doesn't require active malice; simple developer laziness works just as well. If a developer using Flatpak to distribute an app bundles an outdated version of libwebp or openssl in their container and forgets about it, that vulnerability is now permanently camped out on your system. Because your host OS is read-only, your central package manager is completely powerless to patch it.

>While common base libraries might get updated in shared runtimes, modern desktop software is dominated by Electron, Node, Python, and Rust. These apps bundle their own massive piles of NPM packages, Cargo crates, and entire copies of Chromium inside their container. When a critical vulnerability drops in one of these nested dependencies, your shared runtime does absolutely nothing to save you. You are completely at the mercy of the upstream developer rebuilding their app. Until they do, you are left running a collection of isolated zero-day vulnerabilities while basking in a false sense of security. Sweet dreams!
The Sandbox Illusion

>>109205632
>>109206090
>The Sandbox Illusion
>Advocates love to claim containerized apps protect you via sandboxing. This is, plain and simple, an architectural lie. To make complex applications actually work, maintainers routinely ship containers with massive filesystem overrides (like --filesystem=home). The moment a compromised application updates, it has immediate access to your SSH keys, browser session cookies, and cloud tokens. Securing /usr/bin while leaving /home wide open to unvetted third-party containers is a spectacular failure in threat modeling.

>Besides, malware doesn't even need write access to /usr to stick around. It can easily drop a systemd user service (systemd --user) or a simple autostart entry in your home directory. Bam! Permanent persistence achieved without ever needing to touch root. The read-only system partition didn't even notice.
>Sure, you can install Flatseal to manually restrict these overrides. But honestly: who has the time to audit permissions for dozens of apps every time they update? Nobody.
Memory Exhaustion on Modest Hardware

>Memory Exhaustion on Modest Hardware
>Immutability decouples applications from the base system, which completely kills the memory-saving benefits of dynamic linking (shared libraries). Forcing a machine with 4GB or 8GB of RAM to load multiple massive runtimes (like GNOME 45, GNOME 46, and the Freedesktop SDK) into memory simultaneously is a crime against efficiency. The immutable model assumes your RAM and storage are infinite, actively punishing users who aren't running high-end rigs.

File: screenshot-20260705-191355.png (14 KB, 297x37)

14 KB PNG

>>109205828
It's just getting used to train future AI models.
I showed my whole polybar config to Claude because I leeched some guys power button setup (don't really need it because I use 'shutdown' command anyway).
Menu worked but it messed up the typography.

>>109206121
Here is the original. I use Apple backgrounds because they are mostly pleasant with your eyes.

>>109206121
>It's just getting used to train future AI models
they already have man pages which have examples inside, his little configs won't do shit here

>>109206080
skill issue from a filtered literally who.
just use nix/guix instead of fedorameme

>>109206080
>>109206090
>>109206104
>posts some random person's opinions which are filled with obvious FUD and lies
Irrelevant. It's a fundamentally better OS architecture design.

>>109206219
>random person
It's Rénich Bon Ćirić, a Fedora package maintainer and a Linux consultant.

File: 1640458180177.jpg (61 KB, 604x694)

61 KB JPG

anyone want to help a simple fella out who's just been doxxed and wants to do a bit of a house cleaning on my network to make sure there aren't actually any obvious holes. my media server logs which is where I'd expect the easiest route to break in doesn't have any weird IP connections and I don't have anything weird on my PC. maybe I'm just a bit shuck up. I thought those schizo posts about televisions reading your social security details were just shitposts. I'm on arch btw.

>>109206219
>fedora janny
xe has my consolences.

>>109206166
It's more about privacy. You as a former windows fag are living in a problem. Privacy means more like computing habits.

>>109206090
>acts as a natural defense system, frequently causing exploit payloads to crash and burn.
I certainly wouldn't count on this.
I pretty much agree with every thing else though.

>>109206104
>Forcing a machine with 4GB or 8GB of RAM to load multiple massive runtimes (like GNOME 45, GNOME 46, and the Freedesktop SDK) into memory simultaneously is a crime against efficiency.
this gets me thinking, would it be a very bad idea to install bluefin or another immutable os on a laptop with 8gb ram? is mutable the way to go on low-ram machines?

>>109206230
>package maintainer
>consultant
OMG! I didn't realize he was soooo smart and important!

>>109206402
>would it be a very bad idea to install bluefin or another immutable os on a laptop with 8gb ram
Not at all. The poster is a retard. Each Flatpak only adds 4-10MB of memory overhead. You'd need to have 400 Flatpak applications open for it to actually increase your memory usage by 4GB.

By his brainlet logic we should all abandon GUIs altogether and use 640x480 displays with TUI applications. You know, since displaying a GUI at 1080p requires 2x-4x the memory compared to using a 720p display. Imagine how much memory you'd be saving by not even using a GUI and also using a tiny display resolution. You could even use a 4GB laptop!

I'm currently running Fedora KDE and want to try out Budgie for a simpler, more lightweight DE. I used dnf install @budgie-desktop-environment and everything appeared to install properly. switched over to budgie and it's a black window. I can right click and get to system settings, but there's this constant XWayland window that is blocking just about everything I try to do. I can briefly see the actual desktop loading when logging in, but I have no clue how to prevent this xwayland pop-up. there's nothing under the autostart section of system settings.

>>109206456
>Each Flatpak only adds 4-10MB of memory overhead.
False. They load however many shared libraries are unique to the running application. Depending on the task it can be substantial.

>>109206702
Again, it's not substantial unless you're running 100s of Flatpaks in parallel.

>>109206691
Just install gnome with the panel extension, budgie is uglier and heavier than gnome.

Hey, does the nix ecosystem allow me to achieve permanence over my linux laptop if i were to install finix on it?

I've always been interested in NixOS but SystemD always pushed me off of it and so I've been using GoboLinux in the meantime but I've found that it still suffers from dependency hell, just compile-from-source flavored and comes with a version of glibc from 2019 which kneecaps my ability to just shove binaries in the /Programs directory to bypass the compiling from source issue
it lets me have multiple versions of a program through symlink abuse but for some reason only one can be active in path at a time and it's entirely source-based unless i go out and download portable binaries and throw them in /Programs/Programname/VersionNumber/ manually and SymlinkProgram to make it my primary program for the duration of its usage

anyway here's what i'm looking for: for example if i install it today and then leave it as is for about 4-8 years, i should be able to keep using everything as normal but still be able to install and update certain software without causing dependency hell and breaking everything just because i upgraded the penis library from 1.2 to 2.4 in order to use live_service_program_that_relies_on_the_internet at the expense of breaking pretty_important_program, right?

>>109206831
I want to have a daily driver laptop where I don't have to deal with sudden breakages because I want to use software on a spectrum of "wow this is really ancient and it's from like 2003" to "i like using this particular stable version of this software but it's not too old so it's pretty average" to "wow this was updated yesterday"

I have found this to be an incredibly difficult task on the vast amount of distributions i've used. Any time I want to do something new that requires installing some new software within this spectrum it usually turns into a dependency hell hunt and GoboLinux having to make me compile everything from source on a laptop that's slightly old and "business" oriented makes the process of compiling and waiting for a dependency to fail which feels even worse since I'm completely on my own in that regard when it comes to getting everything to compile myself
And most programs don't exactly provide portable binaries that I can just throw into a folder, either

My last resort of an idea was going to be installing Alpine and installing all of my software in separate Distroboxes but then I realize that it'd be hell to integrate all of that software cleanly within my system especially if I need software to intermingle with one-another, and the Linux FHS is a bit wonky for me which is definitely a big part of why I've stayed with GoboLinux for the better part of about a year now since it just cleanly organizes everything so much but it seems that Nix takes a similar program organization approach which seems appealing

I run FVWM2 and have plenty of GTK2 and GTK+/GTK1 programs lying around but I also have GTK3-GTK4 programs on here at the same time since alternatives don't exist or are actually functionally inferior, windowmaker dockapps and some ancient software that I was able to get running after spending a few weeks finding sources from dead links that moved to new links and juggling 4 different versions of GCC which wasn't pleasant

>>109206813
yea, I can see appreciate that. the setting was for xwaylandvideobridge in the /etc/xdg/ folder. once I got the desktop visible, the panel autohide is broken, with requests to fix it going to back 2017. I might also switch to >>109204038 while I'm at it to stabilize extension lifespan. thanks, anon.

>>109206837
've tried to cope by having an Arch chroot for any new software that I want to install but it likes to break every single time i go about a week or two without updating the repos and updating all of my software (I have to install the exact version of Mesa that's on my host system on Arch in order for graphics to work, I use Mesa v19.. Such a pain) and that chroot alone has an entire set of problems that are quite similar to the ones that caused me to seek other places in the first place

My current setup is not very cleanly setup in the slightest and it lacks clean reproducibility which I see is the top thing that NixOS sells themselves on. If I nuked this system right now and tasked myself with getting it all back to the way it was, I would fail immediately because I'd have to redo all of the hacks I did and rewrite all of my additions to the recipes directory, find all of the new links for all of the software I compiled, compile all of the software I compiled which took me weeks to do (including fixing all of the dependency errors and selectin the right GCCs) and all of that beautiful beautiful stuff

File: nasa space station fvwm.png (1.01 MB, 1024x768)

1.01 MB PNG

>>109206846
I'm fine with updating a system, but the part that I don't like is when updating something causes half of my workflow to fail and force me to spend my day cracking at it in order to fix it because someone decided to make arbitrary changes to a critical library or program that provide no perceived benefit to its functionality solely for the sake of making such a change. Most of my problems stem from userland software and dependency hell rather than the inner workings of what makes a distribution. I have avoided distros like Debian ever since I experienced things like apt's tendency to leave bits and pieces of software lying around the hundreds of system directories that exist on a regular Linux distribution (This is also one of the things that pushed me to try Gobo, too, since it also contains everything a software does either within it's folder or in your home directory's dots)

>>109206844
Budgie is remaking their desktop, it might be worth it when they are done but for now this is just quick last release of the old version.
Debian is gay and sucks, don't use it, use something up to day, I recommend Fedora, opensuse, solus, pikaos even ubuntu. Anything but dumbian really.

>>109206844
You can stay one version behind on fedora for extension lifespan.

What email client do you guys use? Thunderbird? Claws? Geary? Something else?
And why?

>>109207200
my web browser

Can someone explain to me why there seems to be this strange misconception that you cannot customize Omarchy? I've seen it in several desktop threads now where people seem to think that just because it comes with the dev's favorite software installed, you're not able to uninstall/replace that software (there's literally an option in the main menu to uninstall ALL of the pre-installed software) and that you're unable to use your own config files. None of which is true.

Also, only somewhat related - why are the the Artix users such rude people? They seem to think that Arch and Artix take some kind of skill or knowledge to install, and makes you superior to others by doing so. Artix literally has a graphical Calamares installer ffs, a mactoddler could install Artix. It holds your hand.

Will switching to AMD GPU improve my linux experience? I'm running Arch on old 1060 6GB and it works, but I sometimes notice graphical glitches on Plasma, and performance on Linux native games isn't as good as on Windows.
Is it because of Nvidia, old card, or DE?

>>109204034
How do I hide phantom or unused ports in Pavucontrol for Ubuntu? Google isn't really helping me out too much.

How do I get xdg-user-dirs-update respect my locale settings in alpine linux? It creates english name folders no matter what I do

>>109207449
>Will switching to AMD GPU improve my linux experience?
I don't know honestly
>I'm running Arch on old 1060 6GB and it works, but I sometimes notice graphical glitches on Plasma,
Sounds like a lack of ram or that Plasma might be too heavy for your old 1060, have you tried doing a dressed down minimal version and see if the graphical glitches persist? Have you looked at its efficiency rate at idle and when you're using it?
>Performance on Linux native games isn't as good as on Windows.
Again, hard to say definitively, could be a bad port to Linux, could be your machine, could be that they didn't bother optimizing the game assuming that the Windows optimization was enough, could be your old hardware. If you really want to know you might have to spend some money to find out.
>Nvidia, old card, or DE?
All of them >:3c

>>109207477
That's hard to do considering your system is immutable, if you can't do it through user configs you probably won't be able to do it because your system files aren't editable and "reset" everytime you shut off and log back on unless you have user configs or bash files that you prioritize early enough to tell your system on boot to respect your local language every time you boot.

>>109206230
Literally who?
Nobody gives a shit about some random maintainer's opinion. Hell, even if it was Stallman himself saying it, who cares?

>>109204034
Any recommendations on how to reduce ssd writes so that they will last longer? I already know about doing an fstrim weekly and setting noatime in fstab.
Which filesystem would be the best to use for that usecase? In the past cow filesystems used to be worse for ssds but i dont know if these days btrfs or xfs have gotten better when it comes to that.

>>109206219
If its obvious FUD and lies then you should be able to easily refute it.

>>109204944
I recall mutt or neomutt used to have issues with noatime in the past.

>>109206104
>Besides, malware doesn't even need write access to /usr to stick around. It can easily drop a systemd user service (systemd --user) or a simple autostart entry in your home directory. Bam! Permanent persistence achieved without ever needing to touch root. The read-only system partition didn't even notice.
Flatpaks by default are supposed to block access to ~/.config and keep all their files in ~/.var/long.ass.appName/

>>109207536
You're replying to someone who is either a troll, doesn't understand shit or isn't a "real" user.

>>109205111
The BSDs did it better. At least when it comes to separating the base system from extra package. Instead of how most of those immutable distros end up using two or three extra secondary package managers to make up for how terrible it is to install anything extra ontop of the base system.

>>109206080
>In a traditional mutable distribution (like Fedora or Debian), the supply chain is centralized. A dedicated, paranoid security team audits and patches the core repositories. Immutable desktops, however, force you to rely on containerized app stores
Immutable distros should've figured out a way to replicate the /usr/local/ thing bsd distros do with their ports system like a seamless distrobox or something.
>>109207553
They're just quoting a blog post from a fedora dev arent they?

>>109207560
isn't most of the rust shit replacing gnu coreutils and other related programs also using a cuck license?

>>109207490
>alpine
>immutable
Wtf are you talking about?

>>109207573
it's not even a dev, just a package maintainer. package maintainers are just bean counters.

Seriously, I could not use an immutable distro. That would not work for me. I could not rely on containerized apps. Immutable distros might be good for users who do not customize their distro much. I am all up in them system files, yo!

>>109205699
This doesn't answer my question
How can a compile a kernel with zfs support without dkms?
You know how when you do

make menuconfig

there are options to build a driver directly into the kernel or as a module? I want that but for zfs so I can build it directly into the kernel.

Also I found the post I was referring to:
https://desuarchive.org/g/thread/108429990/#108430778
I wish I could ask that anon how he builds ZFS directly into the kernel

>>109207913
Immutable distros are one of those ideas that sound good on paper but fail horribly in reality, I don't even get what's the point of immutability? who is this even for? it seems immutable distros are built under the assumption that the end-user is a total idiot, I don't really like such a system, might as well use MacOS if you want your OS to be a walled garden.

>>109207739
Have you been living under a rock?

>>109208211
Actually it's opt-in, not default. Most alpine users are not running immutable systems.
https://wiki.alpinelinux.org/wiki/Immutable_root_with_atomic_upgrades

>>109208090
The majority of PC users are total idiots. So immutable distros are there for the majority of people, not your average lazy tech nerd who prefers a specific way of tinkering with an OS and doesn't mind the potential instability in their daily-use OS.

In any case, the discussion started with Bluefin which isn't even an immutable distro according to Fedora devs. They abandoned the "immutable" label since Fedora Atomic isn't a truly immutable OS.

I get frequent driver timeouts on my RX 7700 XT on Windows during gaming. Are AMD drivers and API implementations better on Linux? Willing to try this out before I swap out the PSU, since I assume it's either software, or my power supply not giving enough juice anymore.

>>109208253
>doesn't mind the potential instability in their daily-use
Why does the immutable tranny always assume normal users have no file history or change management strategy at all? It's like Schrodinger's retard. Somehow they pass the 80 IQ check to install Arch but fail the 80 IQ check to plug in a USB drive and get a backup service running.

>>109208325
You're delusional as fuck if you think people keep backups or would be willing to fuck around reinstalling their OS every time it breaks (or recover a backup when their OS breaks). There's far less people willing to do that than there are Linux users.

Doing a regular factory reset would be more common. Not that you would know since a good factory reset feature is something that's only feasible to implement in immutable operating systems.

>Somehow they pass the 80 IQ check to install Arch
Arch isn't the only distro and it's not even the most popular one. Also not everyone is installing an OS themselves, not even Linux.

>>109208391
Whether most people do it wrong is irrelevant. There's a correct way which is universally applicable and an incorrect way which enables the kind of retard who's OK with universal ID verification. The latter deserves to be gatekept. Let them go be Valve or Google's problem.

>>109208391
so your argument for immutable is you, like most people, don’t know and don’t want to know how to backups and snapshots

>>109207462

sudo vi /etc/xdg/user-dirs.conf

# This controls the behaviour of xdg-user-dirs-update which is run on user login
# You can also have per-user config in ~/.config/user-dirs.conf, or specify
# the XDG_CONFIG_HOME and/or XDG_CONFIG_DIRS to override this
#

enabled=True

# This sets the filename encoding to use. You can specify an explicit
# encoding, or "locale" which means the encoding of the users locale
# will be used
filename_encoding=UTF-8

Have you exported your lang environment? In .profile or /etc/profile

>>109208505
for >>109207477

>>109208253
I feel like immutable distros are only there for people retarded enough to delete their whole kernel because some jackass told them it would mak their PC run faster

>>109206831
>>109206837
>>109206846
>>109206851
>discussing linux in the "linux" thread

>>109208535
are there even people retarded enough to do this? maybe some windows users really deleted their system32 folder, who knows
https://www.youtube.com/watch?v=dpjoUu_24ys

>>109208090
Bazzite is an example of an immutable distro done right; gamers tired of Windows install Bazzite, they use heroic launcher for their Amazon, gog, and epic games, lutris for battle net launcher games, and steam. Nearly impossible to fuck up your machine, but very easy to do the tasks it's meant for. Excellent for Windows users who don't want to learn Linux

>>109208535
or it could just be because .deb files and unknown scrips and repos are dangerous

>>109208895
just wait until you hear about npm and pip

File: cantread.png (194 KB, 735x451)

194 KB PNG

>>109208811
>I THOUGHT IT WOULD GET RID OF THE VIRUSES!!!

>>109206876
>Debian is gay and sucks
>use something up to day
the only things that matter to me, I'll compile from source. certainly better than Ubuntu and whatever the fuck they are doing with their frankendebian spinoff these days.

>>109209059
This is another gay myth, debian wouldn't be anywhere wothout ubuntu, canonical workers are the ones maintaining debian.

>>109209148
>debian wouldn't be anywhere wothout ubuntu
you've got it backwards, anon. Debian was already a major distribution associated with the FSF with many forks before Ubuntu came along, and Mark Shuttleworth had worked for Debian in the past which is why he used it for Ubuntu. Source: me, a fucking graybeard.

>>109209631
My point is it would have been irrelevant by now, in a similar situation to slackware.

>>109209631
lots of things ubuntu needs they make and wow look you can easily port it to debian. canonical is definitely doing a lot of the leg work on important things like apparmor

>>109207200
Thunderbird. Reality is that most emails are HTML pages so you kinda need a gigantic client such as Thunderbird.
>>109206846
>I have to install the exact version of Mesa that's on my host system on Arch in order for graphics to work
So. You got perfectly functioning Arch and then break it with manually stuffing a differing Mesa into it?

>>109209654
Maybe. Slackware wasn't forked much after SuSE but Debian was forked probably more than a dozen times before Ubuntu arrived, which seems to suggest it's a better OS.

so i guess the meta should be to install the most basic distro you can find, then get some smart frontier AI like claude to live in your terminal and let it recognize the hardware, investigate gaps in hardware support, reverse engineer proprietary drivers, etc?

>>109209702
I'd rather Qwen/Gemma or something instead of Claude. Use an open LLM agent.

>>109209702
>>109209941
none of them are anywhere near capable enough yet for this, not the frontier models and especially not local models

File: twitter-gif-2066532918453(...).gif (646 KB, 480x398)

646 KB GIF

Could someone recommend me a local ai watermark remover? Preferably native to arch if one exists. Tired of using browser ones and photopea inpainting sucks.

>>109210187
Two that I know off. IOPaint (formerly Lama Cleaner) for GUI and noai-watermark for CLI.
You run them both in local Python environments.

>>109210187
>native to arch
Perhaps go to a library and read a book about computers first, then come back asking more questions.

Wtf why can't I set GNOME to 90% scale? Shit's too big.

>>109210291
if it's just the fonts you can change it with tweaks or dconf/gsettings

>>109207200
Evolution since Thunderbird enshittificated itself by including the enigmal plugin and fucking it up by wanting to have their own keyring and accepting less keys than pgp would.

>>109209631
He's not wrong. Ubuntu is what is popular, not Debian. And Ubuntu is what brought most developers and maintainers to the Debian ecosystem. Had Ubuntu been based on Fedora or Arch, Debian would be a "literally who" distro like Gentoo or Slackware are.

>>109210458
There was nothing else to base Ubuntu on. Arch did not have the tooling. Fedora was under the control of a competitor. Debian would survive and arguably be better off without Ubuntu.

>>109210291
There's no usecase for scaling. It's a new design paradigm, forfeit typographical rules and usability. I don't think it's even justified to blame phones at this point because it has been 20 years since...
You can see the same issue on Steam even. It's just unusable trash.

>>109210484
>Debian would survive and arguably be better off without Ubuntu.
Them surviving is very arguable since it's already barely holding on as a project. They're lacking maintainers and contributors and from what I've heard they're lacking proper leadership too. Without Canonical it would be an absurdly niche project like Slackware is.
>Fedora was under the control of a competitor.
This is hardly relevant in the FOSS world. If anything making a distribution that is NOT based on what is the closest thing we have to a "standard linux" is a detriment to Linux as a whole. Canonical never achieved anything relevant on a technical level, it was always RHEL/Fedora who were deciding what the rest of the Linux ecosystem does. Imagine how far ahead of the competition Linux would be if Canonical didn't waste people's time on their failed gimmicks and if they had compatibility with RHEL/Fedora.

>>109210544
>Them surviving is very arguable since it's already barely holding on as a project.
How so? Trixie is arguably the best Debian release yet.
>They're lacking maintainers and contributors and from what I've heard they're lacking proper leadership too.
More manpower could be good but that's true for almost all free software projects.
>Without Canonical it would be an absurdly niche project like Slackware is.
I think you might be overestimating Canonical's actual contribution to Debian.
>This is hardly relevant in the FOSS world.
I think it's quite relevant. All else being equal, you don't want a hostile upstream if you can avoid it.
>If anything making a distribution that is NOT based on what is the closest thing we have to a "standard linux" is a detriment to Linux as a whole.
As long as there are multiple distributors, this point is moot.
>Canonical never achieved anything relevant on a technical level, it was always RHEL/Fedora who were deciding what the rest of the Linux ecosystem does.
And yet somehow Debian can't survive without Canonical?
>Imagine how far ahead of the competition Linux would be if Canonical didn't waste people's time on their failed gimmicks and if they had compatibility with RHEL/Fedora.
Probably not any further ahead than it currently is. There is nothing stopping people from avoiding Canonical.

>>109210544
>Them surviving is very arguable since it's already barely holding on as a project.
NTA. I think this statement falls into the myth category. Debian in a major part of some public sector and university / research institutions (I worked for one btw). I've been hearing about Debian dying for at least 15 years and it hasn't happened. They literally have 1000 developers, way more than Mint and no one ever thinks Mint is going to disappear.

>>109208067
I believe zfs is only supported as an out of tree module and cant be baked directly in the kernel without patching the kernel itself.

>>109210599
>I think you might be overestimating Canonical's actual contribution to Debian.
Canonical (Ubuntu and Ubuntu-based distributions) held 70% of the total Linux market share on desktops not so long ago, and probably around half of market share on servers. Their contribution doesn't have to be direct manpower or financial contribution. Their contribution was simply making Debian have an indirect user share dominance for an entire decade so that pretty much all Linux software is built with Debian in mind and often only Debian/Ubuntu in mind. Without Canonical, .deb packages wouldn't have been "the norm" in the 2010s. You're underestimating how relevant Canonical was in 2006-2016. Most people and institutions who started using Linux back then started with Ubuntu and often stayed in the Ubuntu/Debian ecosystem.

>And yet somehow Debian can't survive without Canonical?
Depends on how you look at it. We're kind of discussing 2 different realities here.
In a reality where Canonical didn't use Debian as a base, Debian wouldn't have half the contributors it currently does since no major player would be backing it. Again, most people (users and developers) are in the Debian ecosystem either directly or indirectly because of Canonical alone.
Sure, in the actual today's world Debian can survive. But again, mostly because of the massive development and user ecosystem that was primarily established by Canonical.
Also, I'm not saying it would be a completely dead/abandonware distribution. I'm just saying it would be obscure like Slackware, Gentoo or NixOS are (had Ubuntu never been Debian-based). Definitely not something that a large number of users, developers, institutions or companies care about.

>>109210642
>They literally have 1000 developers, way more than Mint and no one ever thinks Mint is going to disappear.
You're comparing apples to oranges here. Mint is not an upstream project. Mint is just the Cinnamon DE and a slightly modified Ubuntu. You don't need more than a handful of people working on it.

>>109208090
I'm pretty sure that for most users setting up btrfs snapshots for rootfs will do 99% of what people want from immutability while still being able to install stuff ontop of the base system

>>109210642
Debian requires more manpower to maintain than something like mint, with the massive repository of software and different architectures to maintain. They also would need to keep track of security fixes and roll those out asap. Repackage and rebuild entire toolchains when theres a major release version and check that all the packages can build with it etc.
Even voidlinux struggles to keep up with all of this, and they have like 10% of the contributors of debian, not to include voidlinux tranny devs alienating potentional contributors due to politics and whatever other dumb reason they get triggered by.

>>109210843
This is why updating all the time is bad. It's useless and volatile.

>>109210821
I hear what you're saying but I'm not sure that I agree. Debian was already huge when Ubuntu was born. You are speculating.

>>109208535
Immutable distros are for devs who want to jerk themselves off while believing they know what the average dumb user wants.

>>109210877
No shit I'm speculating, we're talking about a theoretical scenario. But it's extremely close to what would have happened. Back in 2006-2016, Ubuntu=Linux and Linux=Ubuntu for the overwhelming majority of users, developers, institutions and companies. It was literally THE mainstream Linux distro.
Whatever "base" they used for their distro would be the dominant one during that time. They simply opted for Debian because, as you've said, they wanted to be a competitor to RHEL and SUSE and didn't want to be in the same ecosystem. They wanted to create their own ecosystem and standards (most of which have failed miserably, but their failures don't suggest that Canonical wasn't relevant and influential).
But, Ubuntu wasn't good because it was based on Debian, it was good because of all the polish that went into Ubuntu as a whole. To most people at the time it was the easiest Linux distro to get into, similar to what SteamOS, Bazzite, Mint, Zorin, etc. are today. Everything else was computer science-tier. Ubuntu alone was probably the most important pillar of Linux growth at the time.
>Debian was already huge
Sure, but "was" is different than "is". Being "huge" back when there were 10000x less people in the Linux ecosystem is entirely different from being huge today. Debian being large in the 90s is the reason why Canonical went with it, but that's precisely the reason why Debian went and grew further rather than stagnating. Without an active and live ecosystem of developers and users, a distro becomes "irrelevant". Again, look at Slackware as an example. It was one of the most dominant distros at the time and yet effectively nobody uses it today or any distro based on it.

You're underestimating the importance of Canonical and the 2nd order effects of Ubuntu being as popular as it is/was.

I have been trying to find a program that fully replaces Fusion360.

File: 1753663334744418.png (967 KB, 1234x1115)

967 KB PNG

Hello, I've been using cachyos for a couple months now but in the last couple days I've had a bunch of crashes/issues after an update. I think I've fixed them but wanna ask if this will be a semi regular thing and if so is it worth switching to another distro that might be more stable?

>>109210876
why cant yall learn to code better and stop breaking shit?

>>109211206
hey look Linux moment, you can be using it and BAM it just becomes unstable. I bet you were just using your PC as normal and BAM it fucked itself up. I wish Linux weren't so awfu.

>>109210876
Updating kde from 6.6 to 6.7 isnt as big of a problem as updating toolchains like glibc to the next major revision or qt5 to qt6

>>109211206
Its a regular thing with anything arch-based like cachyos unless you know what you're doing.
Use fedora, mint, or whatever instead if you dont need bleeding edge.

>>109211200
Your best bets are shit like FreeCAD or browser-based ones like Onshape.

>>109211207
"Learning to code better" won't stop your shit from breaking but it'll at least make it easier for you to be able to identify what the thing is that actually broke and help you to report it to the right place.

At a minimum, everyone should learn basic things like how to downgrade packages, how to build packages from source, how to bisect regressions, etc.
(Although, I get that some people just won't be fucked with that and want something basic that caters to their every needs)

>>109211206
Bazzite or any other Universal Blue distribution

File: sheer_disgust.jpg (46 KB, 651x513)

46 KB JPG

>>109204230
>indicator is solid blocks of white instead of spaced bars

>>109204230
>kexec
>kmod
>guile
>eudev
>elfutils
why does it need all these?

>>109211206
probably amdgpu regressions or kde plasma regressions, or both

>>109211228
>I get that some people just won't be fucked with that and want something basic that caters to their every needs
This is exactly what is keeping Linux off of everyone's desktop. There isn't a single distro that does this. Ones with KDE cone close but I find myself adding a bunch of repositories and shit just to get it usable. I find distros who only put a handful of packages in their repositories fucking retarded. Everything about current Linux is to keep the mass from installing it, like somehow it's made shitty on purpose.

>>109211156
I'm not saying Canonical had zero effect. But Debian was always destined for greatness. And by the same token, Slackware was always destined to be niche. Just compare their governance models, policy and infrastructure. Ubuntu wasn't the first influential project to leverage Debian, and it won't be the last.

>>109211220
>>109211232
Yeah I might switch to bazzite then, at least when I can be bothered since switching systems again might be a bit tedious after I've mostly settled in

>>109211257
kwin + nvidia conflicts from the 6.7 update on kde is what I think it was

and instead of creating the best experience for teh end user yall wanna argue over init systems, sound systems, desktops, etc just make a KDE centric distro, stabalize it, test updates and have every app be available. why the fuck is this impossible to find, oh and make it update and all work with discover. that's all you gotta fucking do and yall act like this is impossible when I can set it up in 15 minutes. Drivers configured etc. just make it happen.

>>109211273
you would think with all the money they get from valve they could collectively afford a 5080 and test against it before releasing updates to kde plasma.

>>109211289
Most NVIDIA breakages are not caused by Plasma but by an issue in the driver and how it interacts with the compositor.
Same for AMDGPU.

When KDE "fixes" it what they are actually doing a lot of the time is disabling something that shouldn't be broken (but is or regressed for some reason) or changing their behaviour to workaround bugs in the driver.

It's not just NVIDIA either, AMDGPU has had a few regressions of late.

>>109211304
so buy a fucking gpu and test updates before releasing them faggot, stop programming it inside a fake virtual machine and expecting it to just work

>>109211313
They do test them. It's not their fault if 6.7 comes out and then kernel 7.0 comes out right after and breaks something though. KDE doesn't work on the GPU drivers. They test what's available at the time but have no control over other people's software.

i dont love gnome but kde plasma releases always being buggy made me switch to it
i will take gnomes limitations over annoying bugs

File: 1.png (747 KB, 1037x785)

747 KB PNG

>>109211289
Yes but at which developer's house does it get to stay?

>>109211322
sounds like an easy fix
>hey we are about to drop an update, oh look new kernel and new graphics drivers, let's test that before releasing this broken shit to the public.
yall are just lazy and I think its time for us at Valve to release our own desktop, one that actually works because it seems KDE devs are lazy pricks. wah wah wah excuse excuse excuse fucks sake.

>>109211338
we should get yall a bunch of 5060s its a good card, and would stop a lot of trouble.

>>109211313
>>109211342
It's not KDE's responsibility to fix driver issues in GPU drivers.
>let's test that before releasing this broken shit to the public.
This is the responsibility of the Linux distribution, not KDE or nVidia/AMD. It's the distribution that opts into packaging these versions and deploying them to people's hardware.

>>109211331
Bugs are temporary. They already released two patch releases for 6.7. Any major bugs get fixed or worked around almost immediately. Blame your distro for being slow to adopt them if you're stick waiting.

>>109211342
They do test new kernel releases. That's how they discover something is broken and then immediately fix it in a quick patch release. They are Linux users just like you and me, when it breaks for you it breaks for them too. They can't test unreleased software though.

I think people don't release just how many lower layers there are below the desktop that can cause the desktop to break. Too many bugs in these lower layers get attributed to the desktop being broken because that's the thing you see with your own eyes but it's not necessarily the cause of the problem.

>>109211348
see, point made for me. if you are going to be lazy I think its time to bring it up in a meeting today to see if the internal desktop is coming along and if its ready to deploy, cuz KDE keeps breaking things. on purpose it seems. This means no more money to KDE.

>>109211348
when you release an update fedora and arch push it out to it's users, you aren't very smart are you, you need to change how you internally handle things.

>>109211354
>waah waah
You're simply wrong. It's the distribution's responsibility to ship packages which don't have breaking conflicts. There's literally nobody else that is to blame other than the distribution since that's exactly what a distribution does. In this case it's either Cachy or Arch that broke something, nobody else.

If you want a distribution where KDE has full responsibility over the packages, then use KDE Linux. If it breaks only then is it KDE's fault.

>>109211361
>fedora and arch push it out to it's users
Again, it's on them to test it against the rest of the packages they deliver.

>>109208211
I am using Alpine for months and I can assure you it's not immutable.

it will look and act like KDE and Windows but it will be based on KDE big picture mode, and cef cuz kde is unusable.

>>109211368
passing the job onto others is going to keep you poor in life.

File: 1763799098227903.png (62 KB, 717x574)

62 KB PNG

>>109211354
KDE doesn't work like that because they're not responsible for the thing that's deployed. They are only responsible for the desktop.

If you want them to take on more of that responsibility then that's what KDE Linux will be but they're not there yet, there's still a lot of work to do before they have a usable Linux distro for everyone.

>>109211384
well yeah, isn't that what you were getting paid for? It is part of your goddamn contract. you're lucky valve just wants to pull out and doesn't sue for breach of contract because this is horseshit. Users complaining because kwin bricked steam machines. the nerve. It's ok the desktop should be out very soon.

>>109211395
No, KDE Linux is nothing to do with Valve. It's something they do on their own time.

What Valve pays them is to work on the compositor KWin and the desktop. Sometimes this work will lead to work at the lower levels of the stack like this:
https://zamundaaa.github.io/wayland,/graphics/drivers/2026/06/24/display-next-hackfest.html

The thing you have to understand though is that they don't control anything except KDE software so when some other Linux distro ships the latest kernel or latest Mesa driver, etc, and that has an issue. KDE can't do anything about that other than to work around it after the fact in a patch release.

Ultimately, these regressions will be irrelevant on KDE Linux because it will have multiple images available, much like Steam OS has multiple images. If there is a problem you rollback to the other image.

File: 1.png (1.01 MB, 1559x900)

1.01 MB PNG

Turns out that it's up to (You) to get NVIDIA working on KDE!

>>109211331
i dont know how people use kde

>>109211412
They want people to test it because most of the developers use Intel or AMD hardware. If you encounter a bug or issue caused by their distro that is their fault then they are generally very receptive to it if it's something they can fix and have control over.

Initially, KDE Linux didn't even have NVIDIA support at all, period, it just wasn't there but obviously that's a non-starter for the huge amounts of people that still like to give Jenson money.

It is using the open driver stack though which is another reason they want people to test it. Truth be told it's probably not ready yet but that's why it's still experimental. It will stable eventually.

>>109211379
Package maintenance is quite literally the job of distribution maintainers. You're the one who is in the wrong here, dummy.

>>109211412
All of those are things that every distro needs.

>>109211406
our bad for trusting you then

>>109211412
>less common
nigger, nvidia gpus are the most used gpus on earth

>>109211450
They never had control over the thing you're trusting them with in the first place about from the KDE software that they do control.

When AMDGPU or NVIDIA DRM breaks something that can never be KDE's fault. They're left to cleanup the mess distros should be doing though and they do that. If they can temporarily disable something or change their behaviour to workaround it then they do that.

>>109211458
Intel graphics is more widely used by far. NVIDIA is the most used dGPU though.

>>109211458
Not really. Intel holds around 66% of the GPU market because of their integrated graphics. AMD and nVidia both have around 15% GPU market share. nVidia is quite irrelevant (on desktops) in the grand scheme of things. It's only relevant to enthusiast gamers.

>>109211492
you will be replaced by AI, gamers are the only real market next

File: 6546654.png (53 KB, 1139x794)

53 KB PNG

Hey guys I am kinda retarded, what are my flags like, what should I change?

Using Ryzen 9 5900x with RX 5600XT, on niri, openrc, cachyos kernel 7.1.3, bash, idk what else I should mention, let me know

>>109211514
No, AI isn't going to replace all of the people that just want a laptop for browsing websites and watching YouTube.

No matter how much Jenson pays you to say AI is going to revolutionize the world there is always going to be a bunch of people that just want a fucking computer.

>>109211553
You probably don't need vdpau support.

>>109211514
It's the opposite. nVidia gamer rigs are being phased out in favor of cheaper alternatives that both intel and AMD provide, along with the future ARM devices.

>Mint
>4080
>doing AI image/video gens
>watching youtube on second monitor
>100% GPU utilization
>some days the youtube video stutters the entire time during generation
>other days it runs completely smooth the entire time
Any ideas how I can troubleshoot this and figure out what's happening? If it was stuttering 100% of the time I'd just chalk it up to the high GPU utilization, but since it runs totally fine other times I'd love to figure out if I can get it to run smoothly more often

>>109211648
Firstly try disabling hardware acceleration in your web browser.

>>109211371
>He's been living under a rock

hi i tried linux several times, mainly arch based distros. for me, gaming is non negotiable. I can give up kernel level anti cheat games, that's whatever. But in other pvp/comp games I've noticed that it just isn't as responsive and snappy & sharp as on win11 where everything feels super fast & sharp. This is not about fps/troughput but... input latency ? I guess ? I'm not sure. I monitored frame times and those seem to be normal. I don't exactly know what causes the slowed down feeling, I assume it's the overhead of the translation layer ? on cachyos id use the lavd scheduler with the the BORE kernel.
Anyways yeah it just feels bad to play with a huge disadvantage.

>>109212037
It would be the translation level, yes. You can always just make a new pc or reinstall Windows 11 (F in chat) to play your gaymens, then build or use your current computer as your personal computer. That way you are separating your digital life from Spycrosoft. I have a W11 system at home that is explicitly for incompatible games and media, I don't surf the internet with it outside of maybe music and troubleshooting.

>>109212037
Maybe Wayland adding a frame of delay? Is it noticable with vsync off? If you went with KDE or Gnome for desktop environment try something that uses X11 like Xfce or Cinnamon to check if it feels any better.

>>109212037
>I assume it's the overhead of the translation layer
No, that's not normal. WINE and Proton do not add any input latency. If there is any it should be less than a single frame, as in latency that's under 1 millisecond and not noticeable. That's the worst possible scenario.

Latency would be caused by something else.
>outdated proton version
Proton without esync/fsync/ntsync. Note that regular WINE doesn't have esync/fsync and only the latest WINE has ntsync.
>bad GPU or vulkan drivers
Check if other people using your GPU and the same driver version have similar issues and how they resolved it.
>your compositor
x11 vs wayland vs gamescope. You should generally use gamescope.
>vsync
vsync inherently adds latency due to how it works. You should enable VRR or disable vsync while gaming.

>>109212135
>Proton without esync/fsync/ntsync. Note that regular WINE doesn't have esync/fsync and only the latest WINE has ntsync.

I either use the steam experimental release or cachyos's own proton

>bad GPU or vulkan drivers
I simply use the latest nvidia drivers my package managers provides for my rtx 3060.

>x11 vs wayland vs gamescope
I use kde plasma which is only has wayland on cachyos at least, haven't tried gamescope yet but these guys https://discuss.cachyos.org/t/gamescope-do-you-guys-use-it/21660
say it doesn't provide any performance benefit ?

>vsync
always off, just like on windows.

>>109211959
see
>>109208232

>>109212188
>rtx
Check if there's a way to enable nVidia Reflex on Linux.
>doesn't provide any performance benefit
Performance no, but latency might improve. The reason Gamescope was made isn't really for performance, but to avoid the issues of your desktop compositor interfering with your game compositor.
>steam experimental release or cachyos's own proton
I'm not familiar with CachyOS so I don't know what's special about their version, but you could try ProtonGE (GE-Proton 11.1 currently). ProtonGE has many small tweaks and improvements over the regular version of Proton, some of which positively impact performance.

>>109212037
Probably Wayland, it forces 1-2 frames of lag in most scenarios.

Applications themselves cannot update while being drawn, and the current frame must be finished drawing before the screen can be updated.

In some fullscreen games this can be disabled, but it is an option that must be enabled in plasma settings, and even then in some games they won't properly fullscreen or be detected. And of course, for windowed games there is no change.

I'd try "cinnamon" if you aren't intent on changing OS's but this might also be overhead on Linux's end or maybe it just handles your mouse poorly, but this was something I noticed personally after using cachy on a gaming laptop.

>>109211384
Hopefully kde linux figures out a way for people to install packages ontop of the system in a seamless way without needing crap like containers or flatpak or having to reboot your system after waiting an hour every time with rpm-ostree

Do you have your home on your root partition or on a separate one?

>>109212232
How many times am I going to have to say the rock meme before you stop giving me Yous?

>>109211273
Don't bother with bazzite just use regular fedora.

>install Artix (installer)
>runit
>check htop
>log shows an runsvdir error
>current and supervise are down

Anyone else experience this problem?

File: linux_encryption.png (52 KB, 680x400)

52 KB PNG

>>109212452
Separate encrypted partition, the root filesystem itself is unencrypted for le simple boot process.

>>109212188
I'm on KDE Plasma with Nvidia (Fedora) and I don't notice any difference to Windows when playing in full screen with vsync off and variable refresh rate on (CachyOS fork of Proton). Very similar setup to you so it definitely feels like something is misbehaving. Did you go to display settings and select allow tearing in full screen? I think that shaves off a frame.
I guess there could always be a very very small difference across the board and I just can't tell because I'm not as much of a pr0 g4m3r as you but I feel more people would notice it, especially in the CachyOS crowd since that distro is made for maximum performance tweaks.

My old Thinkpad went from Mint Xfce->Arch KDE->various other DEs->Arch Cinnamon ->Debian Cinnamon
Now I think I don't care for Debian, it's just Mint with extra work to get it set up right, and uglier defaults. So should I just go back to Mint Cinnamon, which I already use on my main? Or try something different, Fedora, or something else?

I like trying new stuff but the DE is more impactful than the distro most of the time anyway, and I already tried basically all of them. I like both KDE and Cinnamon but don't care for GNOME, as for the more lightweight ones they're mostly OK but pointless.

>>109212452
Separate btrfs subvol

>>109212930
You could use the LMDE version of mint which is based on debian directly instead of being based on ubuntu

>>109212933
What's the advantage of a subvolume over a separate volume?

>>109212939
Yeah but why, I'm basically just back to normal Mint at that point.

Weird, I'm getting significantly higher AV1 speeds on Linux than I did on Windows. It's also faster than encoding x265 or even x264.

>>109213041
AMD?

>>109212954
Everything is kept on one partition, but your home folder is on a seperate subvol from the rest.

>>109213054
Yes. x265 is slow as shit, x264 is usable, but AV1 is somehow faster than both. Went from a usable preset 7-8 on Windows to 4-5 on Linux. That's pretty damn impressive.

>>109213041
>>109213069
might be licensing? Ikonw in fedora you have to install extra codecs for it to work right. AV1 is open

>>109213069
AMD GPU or amd CPU? I'm using CPU and I don't really see a difference to Windows. Are you sure you are using the same version of ffmpeg in case there was a change in the presets.
I use preset 6 mostly (reasonable speed, decent quality for size) and 4 for short videos I want to squeeze maximum quality into like when posting on platforms with video size limit (much slower but noticably less artifacts in certain places).

>>109213197
Handbrake, not ffmpeg as it's easier with HDR to SDR conversion. Just an AMD CPU, GPU is NVIDIA, which sucks because I had to AUR the drivers.

>>109212641
this but the opposite

>>109212954
They're basically just snapshot boundaries. It's the same as one big volume.

File: 1719279395763608.jpg (96 KB, 688x1165)

96 KB JPG

I was in here a year ago complaining about sdl2-compat so here i go again: I still cannot play quake 3 because of sdl2-compat the keyboard just won't register inputs
how is this shit compatible it straight up does not work and hasn't for years
lol

>>109210215
thanks for the rec this is perfect
>>109210259
go fuck yourself

>>109208895
How are .deb files dangerous? Because they don't sandbox like flatpak?

File: ts.png (82 KB, 1024x790)

82 KB PNG

Is this graph useful anymore or is it outdated? I think its a benchmark of how long it takes a terminal to execute some sort of bash command stress test.

>>109213853
1. Because they're not vetted
2. Because 3rd party .deb files are built against specific versions of packages and Debian, so upgrading the distro can cause issues
3. Yes, the lack of sandbox too

>>109213896
I could be wrong but .deb's are vetted though.

>>109213853
They're only potentially dangerous if you're installing them from a 3rd party repo, or manually installing them.

>>109213923
There's no "vetting" involved when a 3rd party .deb file is distributed. Anyone can make a .deb file. The best thing you can do is have signatures, but pretty much nobody adds signatures to .deb files nor does any distro warn you when you try to install an unsigned .deb file or a .deb file signed by a known malicious actor. So for all intents and purposes, .deb files are completely unvetted and installing them is the same if not worse than installing an AUR package.

>>109213923
>I could be wrong but .deb's are vetted though.
In the official repository sure, but anyone and make and distribute a .deb file. I've made and distributed them myself. Lot's of github projects just have build scripts that make .deb packages of their releases. It's a lot more sketchy than you would think.

>>109213896
Anyone with a github account can add a program to flathub. They don't have to be the original author. It's quite literally worse than AUR.
>muh sandbox
Parameters are set by the pak author.

>>109213896
>>109214012
>>109214076
>>109214080
>>109214151
So we are all literally just installing random shit hoping it doesn't fuck us?

>>109214175
Its the same as on windows, or fagos, or even android

>>109208895
>>109213853
>>109213896
>>109213923
It depends on where you're getting the .deb from. You can download Chrome as a .deb package straight from Google. I think that's a pretty trustworthy .deb package, compared to any random GitHub project with 4 stars which might distribute its own .deb package.

>>109214175
Just look at where it came from retard. Check up on the author for 5 seconds. Non-state actors aren't going to work a project for 3 years just so they can burn it in 2 weeks distributing an info stealer.

>>109214222
>Just look at where it came from retard.
That's not very friendly and I like how you just kind of assume that I know how to do that. Maybe I should look up how to download a .deb safely on youtube though.

>>109214175
Precisely. Even if it's open source you're not safe. Look into the xz supply chain attack. It was randomly found by a Microsoft employee who was benchmarking the package and we're lucky the guy had enough autism to care about benchmarks and was kind enough to report the backdoor that he found. Less popular packages have far less eyes on them, and you can safely assume that almost anything you're getting outside of your distribution's repository is not even checked by anyone (open source or not).

>>109214151
If you read the original reply chain you'll notice that the discussion between anons was related to immutable distros. It's easier to verify if your system is infected on one of those and it's trivial to do a total factory reset as a last resort. I believe that was the original point of the anon criticizing .debs.
>Parameters are set by the pak author
>Anyone with a github account can add a program to flathub
Okay? It has a permission system in place, it enforces signatures, and it's at least somewhat vetted by the Flathub team. So Flatpak is still better than both AUR and installing random .deb packages.
Also, you can easily set global permission overrides so that packages can't sneakily enable unwanted permissions.

>>109214287
>and it's at least somewhat vetted by the Flathub team
You hallucinated this. If you report something malicious they might take it down. That's moderation not vetting.
>So Flatpak is still better than both AUR
It isn't because it's a big retard friendly GUI which encourages idiots to blindly install things, and there's no single file the user can audit and see what changes were made for the flatpak version.

>>109214222
>Non-state actors aren't going to work a project for 3 years just so they can burn it in 2 weeks distributing an info stealer.
Remember when Jia Tan contributed to xz for years and then infected xz with a backdoor

>>109214151
The sandbox wont work properly for most usecases aside from isolating shit from ~/.config and any other related directories since for example someone who uses libreoffice or gimp would expect to be access and save files wherever they want to not wonder why half of their home directory cant be accessed or wont properly save

>>109214414
yeah good example of a state actor

>>109212037
same anon here, so far the best tweak yet was disabling window compositing while in game. Game feels much more responsive now.

What's the best way of doing frame generation on a 3090 for games without built-in FSR support? I found some guides using lsfg-vk tied to a paid DLL from Lossless Scaling. I don't mind dropping a few $ to try that but any anons here with experience or other methods? I know framegen works (despite being AMD FSR on an Nvidia card) for games that have it built-in.
Yes I am aware of the tradeoffs for framegen. I've already optimized graphics to get real frames, but the game I'm playing (Nuclear Option) gets shit frames from network/cpu bottleneck on high player counts.

File: annoying.png (965 KB, 1920x1080)

965 KB PNG

>move pointer to the top left to go back or click file
>this bullshit pops up

How do I disable this useless feature? I can't search it up online because I don't know what this shit is called. It's so annoying.

>>109214414
Yeah you're always going to get pwnd by state actors if you're on the internet. They hack each other all the time too. Computers are insecure by design to make jobs for those assholes.

>>109215371
you can definitely search for it
just write out what you're experiencing into google
(then look at the ai response)

File: ss_20260706_161531.png (82 KB, 1055x781)

82 KB PNG

>>109215371

>>109215384
I have AI disabled.

>>109215385
God bless. I don't know why couldn't find this. Still too used to being a windowsbab.

>>109215387
i'd highly recommend learning to use ai in a sensible manner

>>109215397
But it's demonic.

>>109215428
Respectfully, you're logged into Google watching Youtube, and probably something from your "algorithm" or "feed". You already sold your soul to the devil.

>>109215428
as long as you remember your humanity, you will not be taken over by the demons
you're merely harnessing their power

i have a mini pc with 16gb ddr4 that i use for hosting my game dedicated servers
i still want to use it with a de, because i'm a retard
is xfce the go-to lightweight option still? or is there something 'better'?

I am sure there are DEs lighter than XFCE but it simply has more popularity, which gives you more option of customization and easier troubleshooting. So just go with that. I had alpine with XFCE that I also used as a server and it ran well.

Why not DE, though?

>>109215446
I wish I could degoogle myself.
>>109215471
I want to destroy them with my own power, not harness theirs.

>>109215529
>Why not DE, though?
do you mean 'why not use something without a de?'
i don't want to rely on nerdtyping absolutely everything
i like dragging and dropping, actual text editors, steam gui, etc

a ~500mb tax on 16gb doesn't seem like a bad tradeoff for way more convenience

>>109215529
For me it's MATE, it's just as light as XFCE but is actually more polished and has nearly as much customization options

i'm tired of KDE makers trying to LARP like they're windows 98, gimme your weird/unique KDE themes
so far i've found:
>Xeno
>Nightglow Blue

>>109214848
should add that only worked on xorg. Wayland is absolutely dogshit and I can't believe people actually shill it to others migrating from windows.

Hey guys, i recently installed Linux Mint Xcde (or however it's written) in an oldish laptop. It went pretty well, save that I spent an hour without sound because apparently updating the kernel just nukes your audio. But it's been working great! The laptop couldn't even do youtube at 480 without shitting itself, now it can even do 720 or even 1080 if it's not 60fps. I plan on using it mainly for writing, but I have noticed a problem with it:
I can't use terminal. For some reason most commands aren't recognized. Any idea why?

File: Openlogo-debianV2.svg.png (152 KB, 1920x2378)

152 KB PNG

>latest Chromium update on Debian breaks it
>doesn't even launch any more

Nice "stable" operating system.

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1141488

My nomacs also broke.

File: 1728400548675872.jpg (16 KB, 430x319)

16 KB JPG

Every now and then when I use ffmpeg to try and compress videos my pc just shits itself. Even if I use "-hwaccell cuda", htop shows that only my CPU is being used, causing it to briefly overheat. CachyOS, RTX 3070. Any fixes?

>>109215932
no matter what, a cpu load shouldn’t be causing your system to overheat.
this is just too much incompetence to be a real stupid question.

I recently created a nfs share on a Debian SSH server and but on my client system, Dolphin fails to open when the server is turned off because its scanning for the nfs shared folder even though its not available. How do I set up the nfs folder within /etc/fstab so that it doesn't hang there on Dolphin and crash Ranger?
default,nofail?

I'm completely lost. My filesystem became read-only overnight, I assumed it was related to hardware failure (even though this is only a 3 year old NVMe drive) and btrfs had done that to prevent data loss, but "btrfs device stats" and "btrfs check" both showed zero errors. What do?

>>109215861
Gentoo wins again.

I’m hunting for knowledge in the linux kernel and have found LKML lore being a central part where seniors share knowledge and discuss/review implementation. I’ve found linux-mm, linux-arch kvm to be relevant subsystems.

According to lei/lore syntax;
>bs:”<keyword>
>bs:”caching”
>bs:”undefined behaviour”
But how do I query efficiently? This isn’t efficient; help me.

My new build was running fine (Mint 22.3 Cinnamon, 14700K, 6800 XT) for a few days but now it displays a blank screen after the Mint logo on start up unless I add "nomodeset" to the boot options. Obviously this means it's something related to my graphics driver, but every source I find tells me that the driver is automatically handled by Mint so I shouldn't worry about it. Clearly not. Any idea here? IDK how to fix something that Mint is supposed to handle.

>>109215861
Why the fuck would you still use Google Chrome in 2026?

>>109216147
Probably just a drive firmware hiccup or host memory error. btrfs goes read-only at the slightest provocation.

If it happens again check journalctl -b for errors.

what fonts do you use for asian characters support?

>>109215932
Are cuda packages installed? Check arch wiki.

>>109208535
Only advantage seems to be slight increase in security.