(Copypasted from a previous 4chin /g/ thread as a foundation to making these generals on leftypol)
Users of all levels are welcome to ask questions about GNU/Linux and share their experiences.
* Please be civil, notice the "Friendly" in every Friendly GNU/Linux Thread *
Before asking for help, please check our list of resources.
If you would like to try out GNU/Linux you can do one of the following:
0) Install a GNU/Linux distribution of your choice in a Virtual Machine.
1) Use a live image and to boot directly into the GNU/Linux distribution without installing anything.
2) Dual boot the GNU/Linux distribution of your choice along with Windows or macOS.
3) Go balls deep and replace everything with GNU/Linux.
Resources: Please spend at least a minute to check a web search engine with your question.
*Many free software projects have active mailing lists.
$ man %command%
$ info %command%
$ %command% -h/–help
$ help %builtin/keyword%
Don't know what to look for?
$ apropos %something%
Check the Wikis (most troubleshoots work for all distros):
https://wiki.archlinux.orghttps://wiki.gentoo.org/g/'s Wiki on GNU/Linux:
https://igwiki.lyci.de/wiki/Category:GNU/Linux
>What distro should I choose?https://igwiki.lyci.de/wiki/Babbies_First_Linux>What are some cool programs?https://wiki.archlinux.org/index.php/list_of_applicationshttps://directory.fsf.org/wiki/Main_Pagehttps://suckless.org/rocks/>What are some cool terminal commands?https://www.commandlinefu.com/commands/browsehttps://cheat.sh/>Where can I learn the command line?https://mywiki.wooledge.org/BashGuidehttps://www.grymoire.com/Unix/>Where can I learn more about Free Software?https://www.gnu.org/philosophy/philosophy.html>How to break out of the botnet?https://prism-break.org/en/categories/gnu-linuxWhat’s the most FreeBSD-like Linux distro that works on most hardware?
>>26557both BSD like stuff use gnu so just Linux
>>26557Void Linux I belive. I use it on my home server MAS and I hate runit. Just install Rocky or Alma if you want the LOOOOONG Term Support or Ubuntu for everything else
Thoughts on Deepin (Linux with Chinese Characteristics)?
>>26562Haven't used it myself but seems to be one of the prettier ones from a strictly visual point of view. Debian base most likely means it's solid.
The main thing you have to ask yourself is if you're comfortable with using an OS based in China. If it's not subverted right now, it can be tomorrow at the drop of a hat, and you won't know.
Best distro for a paranoid political dissident?
>>26571Look at heads (
https://heads.dyne.org) and roll your own release with more recent software.
why does the 'enterprise' world favor red hat so heavily and OSS/FLOSS users are obsessed with debian-* ??
>>26600The enterprise world favors RH because of legacy, they were the first ones to do it successfully w support contracts. From there a sense of "safe" carries.
End users like Debian because *checks notes* it's good. It's versatile. Seriously, half or more of all distros are Debian for a good reason. It's not completely brain dead easy like Ubuntu or Zorin but it's not far off
>>26616AFAIK the original point of ubuntu was to enterprise-ify debian and give it named/numbered releases and support contracts
>>26620I've never heard that but it makes sense. Not to mention, at least that's the impression I have, these days Ubuntu Server does see some good use. Still not as much as RH, but respectable still.
>>26571Qubes OS or Devuan/Artix/Void with full disk encryption (dm-crypt). Note that Artix has better support for OpenRC and probably S6. Their Apparmor packages don't install init script for Runit. The init script loads all profiles on boot. I recommend always using Apparmor because it adds extra security. Always also install firewall (like ufw).
More tips:
Do not make a swap partition (or encrypt it). Don't use suspend on disk (Sleep mode) or suspend on RAM. Have a separate computer for your personal activities and political activities. Don't use the same password for both computers. Lock the root account if you don't use it ( with sudo passwd –lock r00t ).. Be sure to always turn off the computer that you use for political activities when you don't use it. This way they can't extract your encryption keys from RAM. Prefer using Tor Browser whenever you can and use bridges. Encrypt your mail using GnuPG (don't trust your webmail to do it). I prefer distros that don't have SystemD because SystemD is so huge that it adds to the attack surface. Example: the old xz-utils backdoor required a custom downstream patch to OpenSSH daemon that enabled systemd-notify support. Red Hat, Fedora, Ubuntu and Debian were targeted because no one else had patched their OpenSSH to enable that functionality.
>>26711endeavour is for those who can't install arch by reading some wiki pages but still want arch
>mxI think it's terrible, one glance at the faq and they're already suggesting reinstalling the distro if the debian stable base changes
I recommend gentoo if you have the attention span to read the whole handbook
>>26571gentoo, you can make it as secure as humanly possible if you're a wizard at programming + it's a meta distro, meaning that no one will be able to assume what components you have in a gentoo system, like they won't be able to assume if you're using GNU or busybox, or systemd or openrc or s6 or etc, or even the kernel itself(in the future it may have hurd support or freebsd kernel support)
>>26729Cool off on the gentoo shilling a bit. I understand you may be excited about the options portage offers, but in practice gentoo doesn't come close to being a meta-distro. You have a mainline GNU configurations with openrc or systemd and a lot of commonly used linux software, that is the only one actively supported and tested. Then you have a slew of lesser used ebuilds and overlays, where the most support you will get is asking someone on the forums. Furthermore the project has a history of capitulating to the latest snakeoil and deprecating any sane alternatives, like what happened with udev, dbus and logind.
For someone wanting a systemd-free distro that doesn't use openrc and supports package source compilation, i would recommend both alpine and void over gentoo, maybe devuan and obarun as well.
>>26729is gentoo more related to red hat or debian?
>>26731Neither really, though maybe it's closer to debian in spirit. Gentoo was very clearly inspired by BSD ports.
>>26730>>26730>For someone wanting a systemd-free distro>that doesn't use openrcbruh
>Cool off on the gentoo shilling a bitnah
ok only a little bit>the project has a history of capitulating to the latest snakeoil and deprecating any sane alternatives, like what happened with udev, dbus and logind.I don't think they have many developers to keep maintaining too many components like what happened with eudev
>>26735>openrcYes it's a step forward from sysvinit+shellscripts, but it only supports rudimentary service supervision and paralell startup is experimental to this day.
>eudevEudev and elogind were straight up forks of the systemd code. I'm speaking about static /dev and consolekit.
>>26738>I'm speaking about static /dev and consolekit.but like, where would one find recent/still maintained documentation for static /dev and consolekit, even LFS doesn't teach it and on the gentoo forums only like
actually I'd rather not mention any names of the users, probably disrespectful and etc knows about this
>>26752>where would one find recent/still maintained documentation for static /dev and consolekitA long time ago both both were supported options for a gentoo installation, then a certain redhat project EEEd the linux desktop software ecosystem. Slackware never stopped using consolekit2 for example (though udev clearly won the battle decades ago because of plug-and-play fetishization). Logind is only a hard dependency for modern crapware like Skype and Wayland anyway.
What this historical precedent ultimately means is, that you can't necessarily trust the gentoo maintainers not to deprecate gtk+2 or a release of a compiler version you need. They've already started removing python2 ffs. The project just isn't committed to long term software stability in any way.
>>26756guess I'll just stick with normal gentoo, the most I wanna do of system administration 99% of the times after I build everything is just
emerge -avuDN @world
Why the FUCK did IBM/REDHAT kill centos? Now people are gonna switch off it even more. Do they really think people are gonna pay $350 USD to use a red hat OS with no support to do the same thing you can do with debian for free?
>>26762Their main market isn't desktop and not even small-scale orders (
https://it-notes.dragas.net/2023/08/27/that-old-netbsd-server-running-since-2010/). They're in a dominant market position, so they peddle unstable software to maximize profits from businesses with large-scale orders, that will ultimately buy their contracts regardless of software quality. Nobody ever got fired for buying IBM.
>>26752>even LFS doesn't teach it andActually BLFS briefly teaches how to bootstrap static /dev from a running udev. Thought you might want to see this
https://linuxfromscratch.org/blfs/view/svn/postlfs/devices.html >>26769>>26762well technically you can get up to 16 free copies of red hat os with a developer license. so actually you dont need to pay the 350
>>26802>beyond blfs knowledgeI think glownonymous is basically a god that has transcended wizardhood and knows everything about */linux
>>26568Because Xfce is abandonware
>>26983I agree with none of the placements except that of NixOS. Should've called it AutismOS
>>26983Whats your reasoning behind putting ubuntu at A while Mint on B?
>>27031Ubuntu can be used as a server OS and mint is desktop only
is it just me or debian based distros break more easily than arch/gentoo ? also when a debian based distro in fact breaks the lack of documentation is just annoying
>>27273what? do you expect me to spoonfeed you on how to update it yourself?
>>27272 (me)
nvm I didn't break the company's notebook, it's just ubuntu being ub
aduntu
>>27274Check the name of the thread, jackass
>>27273>latest is 6 years agooof, very spooked
>>27279>Please spend at least a minute to check a web search engine with your question.all necessary resources are in a git repository
https://github.com/headslive >>27293>latest was in 2019I think tails is safer
>>27294it's a devuan install with a custom kernel and some config files. devuan is still updated and more secure than system-dbus-starts-pulseaudio-automatically will ever be. why can't anyone on here read ffs?
>>27295And Devuan upgrades every year or at most every other year.
You good??
Just harden/configure a debian install for any third person reading
>>27297Devaun is nice too.
>>26752>static /dev and consolekit,Was looking into this early based on the conversation here and elsewhere. Am probably going to install a system like this on my new laptop, if ever get around to setting it up.
Posted links to the Gentoo forums just before this, but if you think it's disrespectful then will maintain the obscurity. >>27332What's your partition table look like? Have been saved and hurt by a lack of a separate /home, /tmp, and swap before but have never tried to move /var or /usr to their own partitions. Think install packages too often to exploit the benefits of a read-only /usr.
>>27332I did it just for the meme on gentoo on a virtual machine, it's not worth it
Just randomly remembered there being an FSF-inspired sim in Second Life around 2008 that had an "Ubuntu Pyramid", like a virtual FSF office and some massive box with a rotating Debian logo inside, literally four pillars of free software to find that described each pillar and random free scripts licensed with GPL offered.
Absolute fever dream stuff in hindsight, but I do recall that all of the explanations and assertions about free software these random structures contained were convincing to my early teen brain.
>>27332>>27342static /dev is totally worth it, i haven't had any device problems since setting it up. never used consolekit though. i'm not even sure which programs need a login manager except for gnome, kde and skype for linux????
all my boot bundle in s6-rc does is mount some filesystems, launch fdholder, set the hostname, keymap and a seed for /dev/random, as well as load everything from /etc/modules. the next step is starting some gettys. average linux users don't understand how needlessly complicated their setup is compared to this.
Don't plan on using /opt/ much but do plan on a split-usr.
EFI 512MIB
swap 8GiB
/ 16GiB
/opt/ 32GiB
/var/ 32GiB
/usr/ 64GiB
/home/ ~1TB
tmpfs 32GiB
Anything about this seem nuts?
>>27361>s6-rcWill look into this.
>>27361but after you set all this up, will it need to be maintained/constantly tinkered with or is the only thing necessary of system administration would be just to update the system like on
>>26758 ?
>>27365>/ 16GiBon openbsd i could get by with only 1G. you usually want root to be small, so no files get accidently dumped there. 4-8G would work for modern linux maybe.
>/var/ 32GiBseems way too large. i usually keep mine under 4G, except for /var/lib/flatpak that's a symlink to my massive /usr/local partition.
>/home/ ~1TBhaving /home be the partition for large data is a matter of taste. i like to have /usr/local fill that role. if i were you i would also symlink /opt to /usr/local/opt
>tmpfs 32GiBi can't tell anymore how much memory the average person has, i'm still at 16G on my desktop. i would advise against using your whole memory for /tmp. my system ground to a halt before while launching a game and having a few large files there.
>>27367on the contrary. nothing ever happens on static /dev, unless you create new devices yourself. on most distros there's a larger chance of messing something up by stupidly following standard upgrade procedures. after my first fedora update it refused to boot, pacman croaked once, leaving me with a damaged database and my current alpine setup has a kernel from 4 releases back, because current ones refuse to boot. this is precisely the reason my laptop has run dulap, a gentoo snapshot from circa 2017, for the last 2 years.
>>27368>/ 16GiB>on openbsd i could get by with only 1G. you usually want root to be small, so no files get accidently dumped there. 4-8G would work for modern linux maybe.You're probably right here. Have a bit of a fear of / not being writable.
>/var/ 32GiB>seems way too large. i usually keep mine under 4G, except for /var/lib/flatpak that's a symlink to my massive /usr/local partition.Apparently portage puts temporary files there so if wanted to install chromium this would be 16GiB plus whatever was already there.
>tmpfs 32GiB>i can't tell anymore how much memory the average person has, i'm still at 16G on my desktop. i would advise against using your whole memory for /tmp. my system ground to a halt before while launching a game and having a few large files there.My system has 64GB of ram, might make tmpfs smaller, like 8GiB. Haven't gotten to this step yet.
>>27372>My system has 64GB of ram, might make tmpfs smaller, like 8GiB. Haven't gotten to this step yet.fortunately you can change this at any later point with minimal effort.
>Apparently portage puts temporary files there so if wanted to install chromium this would be 16GiB plus whatever was already there.i say symlink the portage stuff somewhere else. keeping /var small and easily fsckable is important for recovering log files.
>You're probably right here. Have a bit of a fear of / not being writable./ not being writable would be a good thing in some cases. here's a reminder what the point of having multiple partitions is:
<most user accounts should not have the ability to write large filesif any unprivileged program goes rogue, you can still write to "full" partitions as root and clean up the mess.
<the base system should be on its own small partitionpower failure during lots of i/o can leave a partition crippled and you don't want that to be your root. fsck also won't take as long.
<you can set options on individual partitionshave a look at the mount manpage for restricting file types, disabling asynchronous i/o and performance options related to atime.
>>27374Huh, this is not what thought the benefit was. Anyway trying to install gentoo for the second time with musl, static, static-libs, split-usr, and static dev/ isn't quite turning out yet.
>>27369>gentoo snapshot from circa 2017but I wanna keep my system bleeding edge up to date, so I'll just stay with udev as I am not a demigod nor willing to put in the research tbh
>>27379Since there isn't a musl split-user stage3 looks like the key was to copy the list of programs in /bin/ of a split-usr stage3 (with minor adjustments as needed), copy over /lib/ from /usr/lib/ temporarily and then recompile everything with
split-usr static static-libs After this it should be possible to remove /lib/ because /bin/ is static, and remove those files in /usr/bin/ which are also in /bin/ though this last sentence is yet to be tested. Also had to create a new /dev/null for static-usr.
>>27379>this is not what thought the benefit was.with partitions you can also take advantage of the physical properties of hdds, letting you group related files into partitions. for ssds you instead want the simplest scheduler (deadline) and ideally jfs on the most accessed partitions.
>musl, statichuh, i tried this once but was hindered by the lackluster documentation and my cursory knowledge of portage, but it seems as though musl support is no longer experimental, providing stage3 tarballs. marxspeed anon!
static /dev would be interesting if I could keep it bleeding edge updated, like regular gentoo
>>27383Managed to get this thing to boot and with a clean split /usr/ (/lib/ is not totally empty) etc. Have no idea really how to use static-dev however. You can add some filesystems with mknod and a device MAJOR MINOR these seem to be available under /sys/dev/block (one for each partition plus one for the disk it seems) and /sys/dev/char (199 different devices on my system!) . How would one go about determining the MAJOR MINOR numbers for a device e.g. my wireless card, rather than by table (e.g.
https://lxr.linux.no/#linux+v2.6.31/Documentation/devices.txt). Or is this way off.
>>27397Ah, this is apparently way off, let me look into it more.
>>27398Got it, was just a matter of loading the driver. Am a bit worn out from the system administration now though, and didn't do as good a job in the end as started out.
>>27397>these seem to be available under /sys/dev/blockinteresting. i only knew what numbers they were assigned, depending on how you configured your per-device partition maximum. documentation specifically for device numbers is under
https://lxr.linux.no/linux+v6.7.1/Documentation/admin-guide/devices.txt>How would one go about determining the MAJOR MINOR numbers for a device e.g. my wireless cardnetwork devices don't require device files. you may need network_throughput (c10,61) and network_latency (c10,62) for DHCP though.
if at any point you feel like you're missing something, mount a tmpfs under /dev and have a look at what dbus does. 'busybox mdev -s' would also be a good start. the point of static /dev is understanding what character and block devices your system accesses and managing them yourself, which is usually a one-time job.
>>27400>if at any point you feel like you're missing something, mount a tmpfs under /dev and have a look at what dbus does. 'busybox mdev -s' would also be a good start.Have dbus masked, but could install busybox it seems. Might just watch out for changes in /sys/dev and mount accordingly though, if that would work.
>the point of static /dev is understanding what character and block devices your system accesses and managing them yourself, which is usually a one-time job.Have largely postponed this step for better or worse. Will not be able to stay in this state of ignorance because will be needing to connect flash drives etc. But have already questioned why there would be 97 TTY devices when only 6 of them are writable by the tty user.
Wasn't able to reasonably avoid installing llvm/clang. Could be done with the vesa driver, but considering that my monitor is already flickering half the time considered it a bad idea to experiment with drivers.
>>27434Welp, my graphics card isn't showing up under /sys/dev/ though it is present in /sys/devices/. Attempted to link to no avail. Presumably this is another driver issue, but it could just as well be a kernel issue. Tried to give up the static-dev in hopes it would make Xorg work but it turns out that there's a hard dependency against udev with static-libs and ugrd for installkernel to work with a separate /usr/ partition. No doubt this could be gotten around but was enough to discourage me from betrayal. Overall the install isn't fully in the spirit of what it represents because of my avoidance of configuration. Probably should have just used vesa too, not that it's too late.
>>27450>Presumably this is another driver issue, but it could just as well be a kernel issueSwitching to kernel modesetting seemed to solve things. Managed to write a couple custom ebuilds to fetch xf86-input-{keyboard,mouse} and these now work also. Might publish these to a git repository. The screen flicker seems to have been system instability caused by having an excess of ram installed (64GB in a system which officially only supports 32GB). Next up is migrating packages, config, and data files from two different systems and considering how to deal with the merger, and whether to try something new like Nyxt.
>>27396>static /dev would be interesting if I could keep it bleeding edge updated, like regular gentooThis is evidently possible, with the exception of some drivers,
if it really does interests anyone.
>>27450>XorgXorg's device probing depends on udev. There's a script that generates a starting xorg.conf, see the manpage for more information. You will also need to use evdev drivers as opposed to libinput and create the appropriate devices under /dev/input.
Flood detected; Post discarded. >>27460>see the manpage for more informationIt's actually in the Xorg manpage. Launching Xorg with the -configure option should auto-generate most things. The only settings i had to adjust were the keyboard device number and mouse protocol.
>>27463On that note, a good way of testing which input device corresponds to which character device is catting the files to the terminal. Don't forget to type reset afterwards.
>>27464>On that note, a good way of testing which input device corresponds to which character device is catting the files to the terminal. Don't forget to type reset afterwards.Intereesting
>>27463>It's actually in the Xorg manpage. Launching Xorg with the -configure option should auto-generate most things. The only settings i had to adjust were the keyboard device number and mouse protocol.Actually didn't need to do anything after setting up the old drivers it just worked!
Ended up abusing package.provided to mask at-spi2 (dbus), and librsvg (rust-bin). Assuming that works should be able to get everything up to a webrowser on the system without too much of whatever one is trying to avoid. Can apparently even do the same with virtual/tmpfiles though all of this has yet to be tested.
>>27469Guess am still going to need a browser with a dev toolkit for web development in addition to the one going to be using as my daily driver. Guess will just use palemoon for that unless anyone has a better idea.
>>27470Was wrong about the flickering being caused by the ram, and also had to find patches for GTK+ to work without DBus. :(
>>27470>Guess am still going to need a browserGet an alpine chroot. It has up-to-date versions of chrome and firefox.
>>27469>>27471Dbus and Gtk+3 are more trouble than they're worth. Masking them is just common sense. See Stanislav's /etc/portage/package.mask/crapolade:
sys-apps/systemd
sys-fs/udev
media-video/libav
gnome-base/gconf
>=app-crypt/gnupg-2.0.22
app-accessibility/at-spi2-atk
app-accessibility/at-spi2-core
sys-apps/dbus
sys-auth/consolekit
sys-auth/polkit
gnome-extra/polkit-gnome
dev-util/gdbus-codegen
gnome-base/dconf
>=x11-libs/gtk+-3.0.0
>>27473Wasn't looking forward to trying to keep up with patches for Gtk+3 let alone trying to maintaining them. Will give masking >=x11-libs/gtk+-3.0.0 a go also at your suggestion.
>Get an alpine chroot. It has up-to-date versions of chrome and firefox.A chroot to run the browser in sounds reasonable, but using alpine will be my last resort. Seems with either webkit-gtk2, webkit-qt, or palemoon a static build wouldn't need too much.
Unfortunately think palemoon is the only option in terms of development tooling which will build. >>27474>Will give masking >=x11-libs/gtk+-3.0.0 a go also at your suggestion.>Using alpine will be my last resort.These two did not go well together, basically wasted the whole day. No modern browser would compile either at all or without patching Gtk+3. Still no idea why Palemoon wouldn't compile (no real messages or hints), and didn't get around to compiling with debug before got too frustrated and tried to resume with the patched Gtk+3 and webkitgtk2 builds. Now my hope is that www-client/surf will work with gcr patched out. This aspect will be a worse environment for web development but it might still be tolerable.
>>27475Ended up nearly doing an exhaustive search, and doing certain things over and over again with slight alternation hoping they would work. Palemoon doesn't compile without glibc, and am not presently able to make the patches myself. Removing gcr from surf wasn't as easy as it looked. Am also not presently able to make ebuilds for anything more than dummy packages.
Nyxt requires a small patch so might give this a go at some point, put luakit compiled and ran with only the Gtk+3 patch.
The flickering was a hardware issue to do with Turbo Boost. But the trackpad still doesn't work (/dev/input/event* and /dev/input/mice exist) and because used genkernel suspend doesn't work.
>>27479>The flickering was a hardware issue to do with Turbo Boost. lol, no this wasn't it either…
>>27482Mesa failed to open iris, no wake from sleep, high low power screen flickering, failed to build with musl. What more could you want.
So far as can tell ACPI is broken on this machine, suspend doesn't even work with the livedisk. Updated mesa from i915 to intel at the expense of installing LLVM and this resolved the graphical errors was having. Disabled "additional" C-states and all the other extraneous power management options in BIOS, this might have helped with the flickering. Avoiding packages that won't build with musl. >>27521Still haven't migrated over my personal and development files; haven't quite got comfortable with it just yet. A number of compromises were necessary as can be seen from the screenshot. A problem with such compromises is that a number of these systems are interelated, and to remove one modification means removing others. So compromising on udev meant that through grub pciutils needed the -static-libs flag and my system isn't completely static (not sure if it was anyway). Switching to the modern intel drivers over i915 meant acceloration worked but dragged in LLVM (one of the more tollerable compromises). Was further suprised by the lack of modern webbrowsers (and only modern webbrowsers) which would build with the system.
If decided to do it over again would make sure to configure the kernel, and such that devtmpfs was disabled allowing for permissions to be maintained with a static-dev, or else would write a script to handle this for me. That should allow for shaving all but the Gtk+3 compromise. The suggestion to run the browser in a chroot does seem to be the only way to have a clean build environment. Seems like ideally there would be a WPE based browser available (and technically there is):
https://eleni.mutantstargoat.com/hikiko/wpe/ >>27365>32GB tmpfsJust
why >>27539>grubwhy are you using grub and not lilo or efistub
>>27548Lifo was left unmaintained like a decade ago
>>27548>why are you using grub and not lilo or efistubDidn't try LILO or Syslinux, but EFI stubs still required the -static-libs for pciutils.
>>27545Didn't end up setting this up, because it was a bad idea.
>>27551>Didn't end up setting this up, because it was a bad idea.Interestingly /dev/shm and /run are this size.
>>27549>le unmaintainedholy shit why do soydevs want every piece of software to expand like an infinitely growing cancer. old, feature-complete versions are optimal, because such software is stable and well-tested. omnipresent bitrot and instability are the unifying principles of most modern linux desktop software for a reason.
>>27555rob pike, chuck moore, rms, guy steele, alan kay, richard p. gabriel, djb, jwz, uriel, mircea popescu…
all of them were right
I'm trying to install this pirated version of ck3 but I get the following error back
Image mounted as /run/media/user/disk
binaries/ck3: /usr/lib64/libtinfo.so.6: no version information available (required by /run/media/user/disk/binaries/libdxcompiler.so.3.7)
Warning: 2 log categories not defined in log_settings_live.json
Game AI
Game Tests
Trying to Write minidump to dir: /home/user/.local/share/Paradox Interactive/Crusader Kings III/crashes/ck3_20241222_215346/
Created Minidump: /home/user/.local/share/Paradox Interactive/Crusader Kings III/crashes/ck3_20241222_215346//d90f73cc-dfe7-44ca-0bbd5589-d314a18c.dmp
Unmounting /dev/loop0... Error unmounting /dev/loop0: GDBus.Error:org.freedesktop.UDisks2.Error.DeviceBusy: Error unmounting /dev/loop0: target is busy
Failed to unmount /dev/loop0
CD from the mounted directory in another terminal/FM
CD from the mounted directory and press ENTER
Unmounted /dev/loop0.
>>27556>all of them were rightTrvthnvke.
>>27939I assume the minidump has the actual error and the unmounting happens after the game already crashed? Is there anything interesting in the minidump?
>>27943I thought the dump didn't contain anything I could make sense of, but I saw that it wanted vulkan libs. Which I did not have,installed vulkan stuff and now it works.
>get new job, pretty chill
>windows sysadmin is a reactionary
Like pottery
Flood detected; Post discarded.
question for the veteran users: does su mean "super user" or "switch user"?
>>28201i think its "substitute user"
>>27554>>27942one of them is not like the others…
I dunno where else to post this because it hardly warrants its own thread, but I updated an arch install that was last updated on 2022, and it went 100% smoothly. more than a 1000 packages and KDE migrated from Xorg to wayland and everything works as I last left it. i can't imagine a lot of things updating this smoothly nowadays. Arch Linux has really come a long long way, huh?
>>28526that's nice lol, i want to try hyprland or whatever but there's an entire page dedicated to nvidia that begins with "there's no official support" which makes me hesitate, usually what breaks between updates is something that doesn't jive with the proprietary nvidia blobs
im thinking about switching from 'buntu to debian, its increasingly clear to me that canonical really don't care about the desktop experience outside arbitrary changes to bring it in line with what they're doing on servers et al. debian just seems like a cleaner experiance, anything i should know or any tips from debian users?
>>28580There's probably some stuff, debian has quirks. Like you'll have to install sudo. But you're already running debian if you're using ubuntu. There shouldn't be a huge change.
That said, I've been trying fedora based distros and gnome is really nice… If you're ready for an actual switch, try smth fedora with gnome.
>>28582>you'll have to install sudoreally? that's surprising, even if you install it with a desktop like gnome?
>I've been trying fedora based distros and gnome is really nice… If you're ready for an actual switch, try smth fedora with gnome.i actually used fedora gnome a while back but it really didn't like my laptop and would spit out dozens of incomprehensible errors everytime i turned it on, one day an update seemed to fix it, but only a few days later another update brought it back, one of the reasons i was considering debian was that i felt the frequency of updates on other distros was too fast for me, and i also think a vanilla gnome would be better than ubuntu's odd custom gnome, although i like both.
in the meantime i had to reinstall anyway so i decided to try out linux mint for some reason, and i am not enjoying it, so i will still probably be switching to debian soon.
>>28580>its increasingly clear to me that canonical really don't care about the desktop experienceDebian cares even less. Switch to Pop, system76 sells laptops and workstations so they're deeply in touch with the desktop experience. Cosmic DE is what convinced me to finally stop changing window managers and ricing my shit because it met my needs. Comfy tiling with vim controls. It just works.
I have been using Fedora for 3 months and I have not encountered a single problem. I had to track down some printer drivers, but that is simply a proprietary software issue and not a distro specific one. Even then, getting the printer to work was seamless. Switching from Windows has been liberating.
My friend has some dogshit celeron laptops from 2020. What distro/DE copes best with e-waste laptops? Needs to be lightweight but also have good laptop integration.
>>29233Glad to hear anon.
Anons, I need distro help. Been an Arch user for eight years now (wow, I feel old). But recently, I've become very busy so system maintenance has become something that I can't always do. I've gotten a few partial upgrade issues that ended up being deprecation issues with configs on other packages when fixing my system with a full upgrade. I've also had a bunch of AMDGPU-related constant kernel panics on the latest version a few months ago so I had to downgrade once the LTS kernel got those issues too, and downgrading is hard on Arch by design.
So basically, I'm questioning whether it's time for me to change which distribution I'm using. I'm not sure where to go though. Ideally, I'd like a distro that allows me to pick whether to install packages from a stable repository with security updates always and feature updates every month or two, or from a bleeding edge repository that pulls in whatever dependencies are needed. Bonus points if dependency hell is avoided and I can have a stable and latest version of the same package at once for libraries for example. I'd also need as few abstractions as possible, no distro-specific config BS, and configuration without frameworks. I should be able to configure my system only using the software's specific documentation instead of a random, distro-specific meta-config tool or whatever. Also would be nice if the repositories had optimized binaries for newer architectures or LTO or whatever.
Gentoo seemed appealing and like the answer, but as I said, maintenance is becoming a chore and constantly checking USE flags and manually versioning is something that I have experience with but found to be a big time sink.
Any help? [spoiler]Yet another distro help post, a thread got unbumped for this, joy[/spoiler]
>>29350I should mention that Tumbleweed was my main candidate, but holy abstraction layers.
>be me, loonix user
>>try olive
>clunky qt
>>try openshot
>lol
>>try some gtk video editor that doesnt exist afaik
<gtk
<gtk
<gtk
>KEK
is it over? can I ever become le professional editor? or am I capped to these premade transitions on some online editor?
>inb4 4chinz larper
i am a refu/g/ee since 4chud is kil.
>>27473>masking dbus and gtk3elaborate. I am looking to get rid of the gtk3 trash.
Anyone know why my battery would charge in BIOS, unplugging and plugging the power cord not causing problems. But the state change does not occurring in Linux. i.e. plugging in the power adapter without having had it plugged in during boot does not result in the laptop charging.
ummm i am on a netbook
blender is overkill for this hardware.
>>29358can you check the BIOS settings?
which brand is your bios?
>>29359smh i forgot to add
>>29357 >>29359It might not be much different on the hardware considering the encoding and such is all done by a ffmpeg anyway.
>>29360>can you check the BIOS settings?>which brand is your bios?The machine is a Dell Latitude 5590, couldn't locate any such problem online. It seems so arbitrary. If am in BIOS it works, and if am logged in as root it works literally every other time, and otherwise it doesn't work at all.
>>29356Assuming you have cursory knowledge of portage, during dependency tracking it will read any file in /etc/portage/package.mask and skip ebuilds that match the rules, if it encounters 'sys-apps/dbus' for example, it will refuse to emerge any version of dbus and ebuilds depending on it as well. You can use /etc/portage/package.use similarly to add use flags to specific ebuilds.
I don't know if you can still build a useful desktop with the current state of gentoo patches though, i've been using dulap gentoo, a snapshot from circa 2017, for the last five years.
>>29358>>29362Seems like the charger only works sporadically in even the BIOS now.
Have placed an order for an alternative OEM cable.
Have my doubts that this will fix it because of screen flickering in low power.
Suspect it's actually a fault in the way power is handled internally.
>>29411>it actually works.Shocked.
Waiting for junk to install on my Mac, and playing with my gentoo system.
Think am going to try to cleanup my config, which means fattening up my build.
There is really no need for the static flag sense most packages don't build statically anyway.
Got rid of the static-dev so there is no need for the custom ebuilds for xf86-keyboard and xf86-mouse.
Am thinking about ditching the custom ebuild for Nyxt because the browsers too bugged and it's switching to electron.
Probably could replace it with a custom ebuild for Basilisk, this would let me continue to avoid dbus, and modern gtk3+.
>>29501This should actually reduce the amount of "revisionist" Linux on my system.
Mostly on account of discovering the i965 driver with mesa-amber.
>>29501>custom ebuild for BasiliskLooks like may also need to reinstall with glibc for this.
Also need to write a script to migrate my Obsidian to Denote (Emacs).
Sadly am not sure am up for this however.
>>29511Ran into other problems between musl and mesa-amber.
On account of this decided to take the nuclear option, reformating the disk.
Am shocked have the energy to do this, but it's going smoothly so far.
>>29515Even =x11-libs/gtk+2* pulls in rust now, and palemoon doesn't build with gtk+3 so this required the following ebuild:
https://github.com/stefan11111/stefan_overlay/blob/main/x11-libs/gtk%2B/gtk%2B-2.24.35.ebuild supposedly it's possible to use stefan11111's fake-gtk2 to build palemoon with >=x11-libs/gtk+3.0.0 think this would be my preference as it would allow for other graphical applications to be installed in the future but wasn't able to figure out how to get it to work, oh well.
There was a bug in media-libs/mesa-amber ebuild which required the following patch to be compatible with available versions of media-libs/mesa:
https://bugs.gentoo.org/945373#c3To have a functioning >=x11-libs/gtk+3.0.0 previously had to package.provided virtual/freedesktop-icon-theme-0-r4 after installing x11-themes/hicolor-icon-theme, and install a dummy app-accessibility/at-spi2-core from
https://github.com/stefan11111/stefan_overlay/blob/main/app-accessibility/at-spi2-core/at-spi2-core-9999.ebuild these changes are no longer needed as am not presently using >=x11-libs/gtk+3.0.0
So that's where we're at.
>>29526This is version two of my gentoo!
Near completely revisionism-free!
Can't make myself purge udev again.sys-apps/systemd
sys-auth/elogind
# sys-apps/systemd-utils
# virtual/udev
sys-fs/eudev
sys-auth/polkit
sys-auth/consolekit
media-sound/pulseaudio
net-dns/avahi
sys-apps/dbus
llvm-core/llvm
sys-devel/clang
dev-lang/rust
dev-lang/rust-bin
x11-libs/gtk+::gentoo
media-libs/mesa-amber::gentoo
>=gnome-base/librsvg-2.41.0
>>29529based! better make a snapshot for later use.
>>29530This is a great idea! But am going to wait for the system to stabilize first.
>>29529>systemd>pulseaudiowhy even bother with gentoo at this point
>>29535That's the package.mask file.
It prevents anything here from being installed.
>>29536so no librewolf/firefox?
>>29537No, that's Palemoon in the picture.
The alternatives are net-libs/webkit-gtk based.
>>29538Does anyone else have performance issues with palemoon?
Overall it's not bad but am seeing for example 50%+ CPU usage from just opening
https://deepseek.comFurther having lots of lag in the GUI for example moving tabs, or even just the loading graphic on the tabs.
Would try www-client/luakit instead but would need to write a patch for Emacs style bindings:
https://github.com/luakit/luakit/issues/557Don't think am up for that at the moment, even given the increased movement towards programming lately.
>>29526>virtual/freedesktop-icon-themeApparently if you unmask <gnome-base/librsvg-2.41.0 there is a versions of x11-themes/adwaita-icon-theme which can run with this meaning you don't have to provide virtual/freedesktop-icon-theme. The disadvantage of this is that there are known vulnerabilities with this package and no one (so far as know) seems to want to take up the mantle to maintain the <gnome-base/librsvg-2.41.0 version of the library. Not sure it's worth doing.
>>29529/leftypol/ was down for a split second, and my work isn't going well so decided to give removing virtual/udev another go, but this time just using devtmpfs, that is removing virtual/dev-manager. Haven't heard of anyone doing this but it's mentioned in the sys-fs/static-dev ebuild as an option. Think it should work.
>>29671Good news is the thing boots, and wifi and drm &c work. Bad news it only does so with my old dracut kernel and audio doesn't. Think ugrd might make an initramfs too smol, or else there was something in the kernel update. Trying to downgrade first.
>>29671>only does so with my old dracut kernelWould be open to suggestions if anyone has any ideas as to why this might be happening.
With ugrd error is that everything is missing its shared libraries including /bin/login.
Tried modifying the configuration to explicitly mount /usr/ and even /dev/ but the latter wasn't supported and the former did nothing.
Tried just removing the udev dependency from the dracut.ebuild and adding the option omit_dracutmodules+=" udev-rules ".
This results in an error which have now forgotten, but don't have much hope for this path.
>>29355I use kdenlive. it works good enough for my needs
making dumb memes >>29673>audio doesn'tThis was a trivial permission issue. It makes sense that there would be some of these because of udev not existing. Am just surprised it's so few (just one it seems).
>>29730Downgrading did no good, have to wait for the 2.0 release to get rid of the =sys-kernel/ugrd-9999 ** accept-package. In the mean time wrote a device-manager lol:
#!/sbin/openrc-run
start() {
chown root:audio /dev/snd/*
chmod 0660 /dev/snd/*
} >>29731>this is REAL device management done by REAL udev usersmaybe you should untar your device configuration on each boot instead of literally replicating filesystem functionality
>>29737Is the documentation complete and up to date?
https://www.kernel.org/doc/html/latest/admin-guide/devices.htmlWould probably rather avoid the manual approach if it turns out it's not.
>>29740>mknodno, i just think saving the udev-generated /dev and restoring it on each boot to preserve the files is conceptually cleaner than writing device management scripts only to set permissions
>>29741Well, will consider translating /sys/dev/ entries into /dev/ entries in the next iteration anyway.
Wonder how folks did this before sysfs? >>29744>Wonder how folks did this before sysfs?Apparently with MAKEDEV spam, which learned about earlier.
I'm tired of running my own matrix-synapse server. It's not much maintenance but it takes up too much space. Recommend me a fun homeserver
>>29745Had the "genius idea" to hand compress and parse the LAD table.
The idea is that devices are typically linked to the correctly named device in /sys/dev/.
However this doesn't give you the subdirectory in /dev/ to mount them.
So you just need to make a compressed and parsed table of subdirectories from the LAD documentation.
This way could have something like a minimal MAKEDEV before handling permissions.
Got until 116 (ALSA) before realized that the documentation is incomplete.
ALSA doesn't know to look outside of /dev/snd for these devices, but there's no indication.
Bummer.
>>29810Could just use MAKEDEV does it install all the special files for anything that might install?
>>29810Apparently the data used by udev to determine default mountpoints are stored in sysfs in the uevent file of the /sys/dev/*/*/ directory.
https://www.kernel.org/doc/Documentation/ABI/testing/sysfs-uevent it's stored as DEVNAME.
Of course this doesn't solve the permissions issue,
one could reimplement a parser for udev rules.
Then one could have a hand-crank udev/MAKEDEV replacement, with all the proper permissions applied.Then /dev/ is just a directory which can be wiped and replaced with a snapshot tar.
>>29817This is a bad joke for a number of reasons one of which is that -udev (supposedly) prevents udev rules from being installed, and removing the udev flag enables interaction with libudev.
>>29819>i didn't find that out until after wiping my perfectly good suse installfor the future keep in mind linux distros, especially the glibc ones like suse, are pretty much interchangeable for running third-party software. dynamic linking means every binary needs library dependencies installed as well, but other than that executing a binary only involves the kernel loading it and the program itself searching for files it needs, nothing as dramatic as the differences between windows versions. this is why they have steam in most repositories, despite it only being packaged for archlinux (ubuntu formerly).
It's late, was trying to get alpine chroot with tbb running but ran into issues when attempting to execute as non-root:
bwrap: Creating new namespace failed: Operation not permitted
error: ldconfig failed, exit status 256
The typical fixes didn't help:
https://github.com/flatpak/flatpak/issues/703#issuecomment-293250243https://github.com/containers/bubblewrap/issues/324#issuecomment-517378834Feel have closed in on an exhaustive search of the browser space with "clean" dependencies.
Seems like you can even use Musl if you don't go the Pale Moon route.
The options are Pale Moon (glibc), LuaKit, * in AppImage, or * in chroot.
LuaKit or TBB in chroot are probably the cleanest of these.
>>27473>>29821>>29824Cheaped out a little on the chroot, and just used debootstrap, but got this working.
The only thing that seems to not be working is the audio, but don't have it in me atm.
Have something like the following in my ,bashrc:
alias torbrowser-launcher=' \
sudo mount --bind /dev ${HOME}/jail/dev && \
sudo mount --bind /dev/pts ${HOME}/jail/dev/pts && \
sudo mount --bind /proc ${HOME}/jail/proc && \
sudo mount --bind /sys ${HOME}/jail/sys && \
sudo mount --bind /run ${HOME}/jail/run && \
sudo -E chroot ${HOME}/jail /bin/su ${CHROOTUSER} -c torbrowser-launcher' >>29824>* in AppImageShould have tested this before reporting.
They dynamically link to common libs.
But don't seem to have a list of what these are.
>>29828Needed to run with apulse and modify /etc/group so that the chroot and system have the same group number for audio.
Was considering replacing Pale Moon with LuaKit for webdev and public persona, but decided against it.
Guess it's all working now exactly as it should.
>>29895FSF Emacs for common lisp and emacs, MicroEMACS with some of my own hacks for everything else. I also hacked on qemacs a bit, but couldn't get font and default colors to work properly. I've also thought about writing my own emacs, yet the client-server architecture i had in mind, designed around an editing process controlling only buffers and various i/o services, eventually morphed into a network-transparent forth dialect with a message-based object system i have no hope of implementing within the near future.
>>29895Emacs is the goat program, I love it so much. I also love all the ancient quirky shit It has tbh. 10/10 super fun program.
Also, I'm forced to use windows at work and emacs on windows is actually good enough to use and the eshell gives me a shell which isn't powershell
Is there a distro (Preferably in the Debian family) with an "LTSC" style release pattern?
Finally got sick of W11 and tried linux again. First time in my life I felt comfortable enough with it to delete my windows partition. Everything just werks. Been really easy to google answers for the problems I do hit.
Bro why did nobody tell me about KDE Connect, I had to find out it existed from the comment section of fucking linus tich teps, it's fucking awesome. There's a nightly for macOS too, so everything is connected and cozy
>>30450If/when it gives you problems (which it did for me, very quickly actually) you can try LocalSend
>>30527im using endeavour. they say it is just arch with calamares installer. it uses dracut but idk what that means. it also has yay for pacman by default which i like. i just picked it cause i was gonna kde anyway and i liked the space stuff. i was using grub to dual boot until i deleted windows and reinstalled on defaults so now its systemd.
people say manjaro is bad now. i think the popular one is Cachy? the gamers like it better but the real gamer meme is Garuda. i wanted less bloat and only install gamer stuff as needed. endeavour works great zero issues for almost a year now.
>>30500LocalSend is insanely good for sharing files, definitely much better than KDE Connect but I'm mostly using KDEConnect as a KVM and it's actually pretty fucking good for that, I'm typing this form my macbook into an arch machine in another room, very fancy stuff
>>30620IIRC from configuring the kernel, there are generic usb ethernet and wifi drivers that include all devices, so you should be hard pressed to find any that don't work.
>>26557Definitely void. It uses the runit init system which is quite similar to BSD init scripts afaik. Also I don't think there's anything similar to jails on Linux so if that's important to you then stick to BSD. Although bare in mind I don't have much experience with BSD so I could just be talking out my ass right now
>>30708BSD Init has startup and shutdown scripts, which are kind of like the scripts runit uses to emulate sysv runlevels. Otherwise runit uses an entirely different service supervision model than BSD, which only has sysv-style init.d (rc.d in their case) and inittab.
AUR Chaos malware: an analysis>https://www.mh4ckt3mh4ckt1c4s.xyz/blog/aur-chaos-malware-analysis/* There were malicious packages in Arch User Repo that installed remote access tool (RAT).
* The known malware packages are: librewolf-fix-bin, firefox-patch-bin, zen-browser-patched-bin, minecraft-cracked, ttf-ms-fonts-all, vesktop-bin-patched, ttf-all-ms-fonts.
* There might be even more malware on the AUR!
* Be sure to always review all package PKGBUILD files, even when you are updating AUR packages that were previously to be known to be good.
* I suggest using rua or yay AUR helper because they can show diff views of AUR packages.
>>26557I would just run FreeBSD if you want to use it. You could also use Slackware but I think FreeBSD or some more mainstream GNU/Linux distro is better, since package management is painful on Slackware.
>>30936>Be sure to always review all package PKGBUILD files, even when you are updating AUR packages that were previously to be known to be good.If they can pull malware directly from the PKGBUILD, they might as well go all the way and pull a legit-looking source from a compromised mirror. At this point you're better off getting the upstream source and handpicking any patches you need yourself.
Given that the distro refers to many AUR packages in its wiki and those especially would be maintained by relatively trusted community members, they should have made a PKGBUILD repository with barebones vetting already. Arch retardation syndrome strikes again!
>>30936>>30937Having read the article, the perpetrator seems to have been discovered very much because they were an amateur. As i already mentioned, a smart attacker could have inserted the suspect code directly into the program source, which could either fork unprivileged malware, like a miner, or try to gain root privileges through an install wizard invoking sudo or any privilege escalation 0day.
>>26557It depends on what specific features of FreeBSD you are looking for in a Linux distro.
If you want a BSD init system, the only distro I know of that uses BSD init is CRUX; Arch used to use it, being originally based on CRUX, but has long since moved to systemd. The other init systems on Linux like SysV, openrc, and runit are all totally different systems from BSD init.
If you want ZFS, you can set up OpenZFS after the fact on any Linux distro, but if you want ZFS-on-root right out of the box then FreeBSD and Ubuntu are the only OSes I know of that provide that in their installers.
If you want jails, there's not really any equivalent to that on Linux, there's containerization and things like Docker but they work in a fundamentally different way than jails.
If you want a lightweight minimal OS that adheres to KISS principles and doesn't use systemd, there is Alpine Linux, Void, Gentoo, etc.
why are snaps so ass
>>31347Canonical forced it, every new install of Ubuntu installs snap-based versions of things like firefox etc etc.
Slower than native apps, adding unnecessary bloat and dividing the Linux package ecospace further.
Hard to get rid of, I've heard people bricking their systems over it.
What is happening on NixOS? Been living under a rock for some time and /g/ only talks about transhumanists bad.
>>31405Use snap install and do a lsblk, then stab out your eyes
So, i tried installing CachyOS today. Live CD looked like a standard "Arch for lazy and/or stupid people" kind of thing… At first glance. First problem i noticed, was that with a 2-and-something GB image here was no offline install option. Yeah, weird, but it's 2025, so i could theoretically forgive that. Well, it was actually a problem for me, since government is doing some weird shit with the internet, a lot of mirrors were inaccessible, but that's not the point. So, i was going to set up some VPN or proxy, but then… Something happened. Something truely, minblowingly terrible.
Don't open the spoilers less you lose your faith in humanity:
No, seriously, this is a real war crime, if you open next spoiler you're going to lose that last little shred of faith in Linux community you didn't even know you had.
"bash: man: command not found"
MAN, MOTHER OF GOD FUCKING HELL, MAN
Yeah, we've seen shitty distros, but THIS… I never would have imagine anyone could sink this low.
Curse upon anyone responsible for this abomination.
Am considering getting a pinetab2 8gb for tutoring work.
The idea is to use a patched build of chromium for the mpp. [^1]
This should alleviate the CPU pressure from recording the small built in.
It should be able to run an unrecorded 1080p separate monitor for references too.
My concern is running the un-GPU-accelerated whiteboard.
Maybe krita could work, which has partial GPU-acceleration?
But this would seem to be a little less a convenient usage of the platform.
And I'm not sure if it would work anyway…
Anyone with experience on a RK3566 think it could handle it?
[^1]:
https://github.com/JeffyCN/libv4l-rkmpp>>31595>The idea is to use a patched build of chromium for the mpp. [^1]>This should alleviate the CPU pressure from recording the small built in.There's apparently not really negotiation on the encoding.
It's entirely possible that wyzant doesn't support the hardware encoding.
>>31596>The idea is to use a patched build of chromium for the mpp. [^1]Hardware accelerated encode is a (non-ARM) ChromeOS, Windows, or Mac OS exclusive.
Was looking into "brunch" as a way around this, maybe with crostini.
And meanwhile learned that ChromeOS is being replaced by Android…
The cleanest solution found was to use the ChromeOS kernel in Gentoo.Use this Gentoo box to host LXC containers running the ChromeOS ROOT-A.After all this it's possible that hardware acceleration still wouldn't work with the tutoring platform.
This is because of "negotiations" on the WebRTC encoding.
But probably it would work…
Unique IPs: 65