/fglt/ - Friendly GNU/Linux Thread

Yeah, install x server (or Wayland if you're one of those guys) and pick a WM you like. If you're new to this stuff (sounds like you are) , gnome will probably be one of the easier ones & handle stuff for you. But it's your choice really

I'm trying to split a command in multiple lines but it doesn't work.
i3lock -i ~/.lock \ # background image
--clock \ # display clock
--indicator \ # always display indicator
--insidecolor=$color_circle \ # inside color of a circle
--insidevercolor=$color_circle_ver \ # inside color of a circle during v
--insidewrongcolor=$color_circle_wrong \ # inside color of a circle if user
--ringcolor=$color_ring \ # ring color
--ringvercolor=$color_ring_ver \ # ring color during verification
--ringwrongcolor=$color_ring_wrong \ # ring color if user entered wrong
/

Errors I get all say that --clock, --indicator, --insidecolor... are not commands, like the interpreter completely omits '\' at the end of each line.

Arguably if you pick a wm or login manager, xorg or wayland will be pulled in, it's not like they usually can compile without either and so portage has them as dependency.

You may have to supplement drivers for xorg though.

A graphical environment comprises a couple big pieces
-X, or Wayland if you're weird, or Mir if you're African
-the display manager (DM) to handle the paperwork of logging in and out
-a desktop environment (DE) to make something nice to look at

You're probably right, I don't use portage so I'm wasn't sure how much it handles for you

You can't use comments like that. Remove them and it will work.

thanks

Anarchy Linux confirmed to be Devil Linux

Attached: Screenshot from 2019-03-22 10-26-45.png (1024x768, 6K)

that's exactly what i thought, i went to sleep and when i woke up a few hours later the thread was dead.

>forking arch
>keeping systemd
>devils work
imagine my shock

Attached: 1539509220831.jpg (394x399, 36K)

I'm currently running Ubuntu LTS. Most of my problems stem from their shitty version of GNOME. Will it get better if I upgrade to a non-LTS version (i.e. is the up to date version any different?)? Or is my only choice to do things the hard way and abandon Ubuntu for something actually good?

has this now become a meme or are you just asking the same question over and over again?

Ubuntu's shitty GNOME is a bit of a meme, but my question still stands. Is the up to date version noticeably less shit?

i couldn't tell you but you could just switch repos and test it. if now my recommendation to switch to mint, lubuntu or xubuntu or even better debian still stands.
let us know if it is significantly better user

>or even better debian still stands
Actually, I had that in mind. I hear that Debian Sid is more stable than most distro's LTS versions, so that could be a way to get a decent DE without having to learn a fundamentally new distro.

Still, I'm mostly just asking because I can't be bothered to switch distro either. Upgrading my Ubtunu is all that I can really be bothered to do to fix my DE problems, but as I've said, I'm asking my question because I don't know if that'll just be a waste of time.

yes but switching the repository doesn't need that much time, i would just try it, tell us if it makes a difference

Sounds like a path to breaking shit. No thanks.

Budgie desktop is sooooooo disappointing. I can't interact with tray applets. I can't fucking turn redshift on and off from the icon and I can't even set my own keyboard shortcuts. And another thing, why ar desktop icons so fucking huge? Is this a piss take? All style and no substance.

Attached: vomit.jpg (680x680, 85K)

Hello friend, have you considered making your own perfect DE?

That won't be necessary becauz now I know MATE is the perfect desktop now that I know what a monstrosity budgie really is.

I'm working on a cron script to send me notifications about warnings and critical errors in a log file.
I'm using logtail to read and remember previous errors but I can't send strings longer than 2000 characters to my notification service.
To avoid ugly cuts I'm planning to break each line from the newly read logs, count them and make a loop out of it to send n-1 notifications.
I'm not quite sure how I'm going to deal with single lines containing more than 2000 characters (lines containing urls or uris for example) but I'd like to deal with this a step at a time so:

tldr; how do I count the total lines of a string (used cat + a bunch of greps to isolate warnings and errors) in bash? wc piping didn't work

Attached: 1533854289326.jpg (5500x2679, 806K)

I am using linux-hardened, but I can't seem to load acpi_call, Idk how to get it to run....

Is there a way to search locally via a hash?
Something like fsearch would work but with the capability to search based on a has(md5 for example)

>how do I count the total lines of a string
>total lines
>a string

find ./ -exec md5sum {} + | uniq -w32 -dD

how to make vim `e` stop at underscores?
I tried
:set isk+=_
but that didn't do shit. real annoying to edit names in php

isnt it just an installer for arch?

Sometimes I want to listen to something else and usually, I use YT in a browser to do that.
But doing this means I don't have media controls.
So I added that to my media buttons by sending the keys through xdotool.
Is there a faster way?
Right now, my screen flashes because I change desktop, do something to the browser and then returns.
It is purely visual, I mean it works fine, but is there a better method?

it is

Grep -c can count

send the youtube links to mpv instead

That wont work with multiple paths, i need a way to specify multiple root paths that span millions of files and have it search, similar to fsearch.

found a plugin
github.com/chaoren/vim-wordmotion
werks fine

>wont work with multiple paths
yes it will you just need to thonk a bit

$ ls tt ttt
tt:
1 2 3 mark

ttt:
1 4 5 mark uniq-file

$ find ./tt ./ttt ! -empty -type f -exec md5sum {} + | sort -k1 | uniq -w32 -dD
1d4171a91c31af2c59acb89a042836e1 ./tt/3
1d4171a91c31af2c59acb89a042836e1 ./ttt/5
51548e9396084cc0f79cbb718b49602d ./tt/mark
51548e9396084cc0f79cbb718b49602d ./ttt/mark
ba1f2511fc30423bdbb183fe33f3dd0f ./tt/1
ba1f2511fc30423bdbb183fe33f3dd0f ./tt/2
ba1f2511fc30423bdbb183fe33f3dd0f ./ttt/1
ba1f2511fc30423bdbb183fe33f3dd0f ./ttt/4

Attached: 41681241.jpg (221x228, 9K)

Actually that doesn't work for me in that case.
I use YT as a makeshift radio. I don't want to control what I am listening to, I want the algorithm to spin out based on a single starting point.
It already works fine just by opening YT in a private browser window. It takes a second to setup and then requires no interaction for hours.
I can just leave it on a desktop I don't use and it is perfect.
I don't think I can set this up the same way with mpv.

I just bought an old machine because mine died and it came with Win10. It turns out that disabling all animations in the os also disables Chrome tab animations. That's some bullshit I've always wanted that feature on linux.

If I mount my one drive cloud folder with rclone mount to have the drive accessible from my file manager, Does it download everything to my computer everytime I turn on my comp? Or does it just copy the indexes and download the files when I open them? I don't wanna rape my ssd and I probably don't have enough storage left anyways

Attached: proxy.duckduckgo.com.jpg (326x294, 34K)

it worked

Attached: cc893c9fcca956b4287fa59c2807318f.gif (500x281, 131K)

The wintoddlers stay up all night making tech request threads and /v/ shit

They arent all relative to the folder though
And i dont particularity feel like putting 20 full absolute paths.
A database is much simpler and easier to change. I dont buy that in 40 years time no one has created a searchable database via hash

how is void so based and comfy

Hey, I'm trying to set up dropbear to remotely decrypt my root partition. It seems to boot fine, but sending the unlock command does nothing... Even if I see into the machine and manually run cryptsetup to open the LUKS partition, it doesn't work. It accepts the passphrase, and it doesn't throw and error, but no device gets created under /dev/mapper

What gives? Have I cocked something up?

>see into the machine
ssh into the machine*

you replied to the wrong post

Just cache probably. If it is like mounting sshfs or s3fs, you don't download the whole thing.

What's so comfy about it?

I could have sworn that I heard something really interesting about the way its package manager works a long time ago, but I could never find that information again. Even on the void website I don't really see anything about how the package manager works internally.

Either I'm misremembering or they're foolish for not advertising what I remember possibly being the best feature of void.

the best feature of void is runit.
runit is fast and easy. i'm not the user you responded to but for me it's kinda comfy to install everything i need after a fresh install, it's kinda lovely to install bash because it doesn't come with a default shell. also it's a rolling release, a lot of anons here think it doesn't have enough packages but for my basic needs i rarely have to compile and install something from source. oh and the installation itself is super easy and fast.

kill me

I'm trying to install Mint onto a SSD with some HDDs in the machine, but I'm not sure what to do about swap, or what Ubuntu even does by default. Should I just make a swap partition on a HDD, and Ubuntu won't do anything fucky on the SDD like make a swap file? If I don't make a swap partition or file, will I really just not have swap at all? Are partitions even good instead of swap files? Does Linux not have dynamically sized swap files? (does Windows even have that; I never thought about it)

I've never even had Windows on a SSD, so I don't really know what I'm doing. Sorry for the stream of questions, please help.

Void does a lot to follow the unix philosophy. The components are all pretty simple. Runit, for example, is a very simple, elegant, and manageable init system that improves upon sysvinit without adding bloat like systemd. Xbps is a speedy package manager with a nice interface. A lot of the components are separate executables, xbps-install, xbps-query, etc. which keeps things lean and modular. It also has a nice build system with xbps-src. By doing a manual installation, you get a very lean, minimal, and understandable system. It feels a lot like what arch is trying to be, but without the autofellatio.

As far as advertising, that's probably my favorite thing about it. The people working on void are doing it because it's the OS they want to use. They don't really care about the rest of the linux-verse or if you use void. They're not hostile or elitist about it, they just don't care. As long as you're reasonably familiar with linux and want a binary distro, void is just a complete pleasure to use imo.

Attached: 1542829343708.jpg (300x296, 13K)

>bump.
id also like to know

How much RAM do you have? It's not advisable to put swap on an SSD obviously and you may not even need swap.

That said, Linux does have ways to use a swapfile. I don't know how Ubuntu handles this by default, unfortunately.

if you have 16GB or more of ram you probably dont even need a swap unless your doing some heavy multitasking

what are good programs on linux to test stability of an overclock?

emerge -1 gcc

Running now. I added a -B as well to that to not actually install it

>MAKEOPTS="-j4"
>i5 4590 cpu
is it normal for kde and its deps(about 220 packages) to take 2 and a half hours to compile on a 4 core cpu?

stress-ng
cpuburn
various tests in the phoronix test suite

Yep, on most settings it'll take longer.

Also 4 core cpu isn't very telling anyhow, that's anything from 10 year old machines and onboard ARM to current Ryzen/i7.

thats why i listed my cpu an i5 4590

I have 16GB of RAM. And looking into it more, it seems the latest versions of Ubuntu do use a swapfile by default at /swapfile

idk, if I probably won't ever use it anyway, maybe I should just leave it there?

and i read some where that the haswell architecture has some quirks that slow down compilation

you can disable it if you want just
>sudo swapoff /swap
and update your /etc/fstab file comment out
>#/swapfile

Well, even if you have a lot of RAM, the OS may choose to pre-emptively swap anyway depending on swapiness settings. You should either disable swap, or manually create the swapfile on your HDD array somewhere and bind mount it to /swapfile in fstab

Good point. I still have *no idea* how fast that CPU is at compiling relative to the current Ryzen/i7.

Also, RAM (and SSD or not) also plays a rather important role on top of that. And then compiler choices and settings. Even the linker choice can make quite a few minutes in difference over some hours of compiles.

But at 2h, you can assume it's completely normal, it might even take twice as much time on not particularly good hardware and average settings / tooling.

and if you want to turn it back on for some reason just
>sudo swapon /swap
and uncomment the line in you fstab file like so
>/swap

Well I don't want to disable it completely, but thanks.

I kind of like the idea of having faster swap if it doesn't get used that often though, maybe I should just change my swappiness setting instead?

Also, if I do put it on the HDD, then just moving it with swapoff/swapon is better than bind mounting it, right? I was actually kind of unsure of how that works with having to auto mount the HDD and stuff.

I'm kind of confused with this Gentoo wiki page on amdgpu
wiki.gentoo.org/wiki/AMDGPU
it says if you compile it as a module you can skip the "firmware selection" part, but what exactly is that? Does that mean the whole "Incorporating firmware" section can be skipped?

if you compile amdgpu into kernel then you need to compile in the firmware blobs as well. just compile it as a module and it'll look for the firmware blob on its own

so which part in the wiki page should I skip then?

Usually when I break my operating system it's because I did something stupid like half uninstall something or fuck with permissions but this time the only thing I can think of that I did was run the updater and all of a sudden Mint is shitting itself at all times. On boot, running, even on shut down. It took two whole minutes to shut down my PC. Everything is slow as shit and I can't figure out what the hell it is. Should I just nuke everything and restore an old system image?

Attached: URHGAAUGH.png (569x571, 43K)

>Mint
there's your problem

>Should I just nuke everything and restore an old system image?
yeah i would recommend that if you can't find out where the problem came from

the "Incorporating firmware" section

ok thanks

>the ol' nuke everything fix
Can't say I miss doing that myself. Maybe use this as an exercise in fixing your shit? Could be hard but ultimately worth it.

Anyone familiar with setting up WPA wireless for Gentoo? I just installed it for my new Thinkpad W520 and while the networking was perfectly fine (light on and everything) in the livecd it's not in my installation. I do HAVE an interface but dhcpcd times out trying to establish a connection, since the network light isn't on when I'm in my install I reckon I must've missed a kernel module but I know I've enabled basically all of them for Intel in the network device drivers section.

Hello Jow Forums&/fglt/I have always wanted to go balls deep into anonymity and start running some Linux distro and post images of my sweet riced up desktop but I never had the time but now I'm taking the weekend to start properly.

lurking I found this guide: anonguide.cyberguerrilla.org/
(for some reason you can no longer download it) but I ran into some problems and wanted to know if anyone was familiar and if its still viable. At this point fuck it I just want to dual boot Arch, what are the underwater stones?Recommendations?

Attached: 1489257705250.jpg (1000x963, 203K)

>arch
how much do you like sniffing your own farts?

>always wanted to go balls deep
>I just want to dual boot

Unless you're using linux as your main OS there's no reason to dual boot. Just install it in a VM to play around with until you're ready to ditch the diaper and out on your big boy pants.

Is there any limitation or why is that not a single DE offers bottom rounded corner on windows without borders?

Attached: 1553221533877.gif (283x292, 195K)

I'd suggest installing connman in live mode and then using that.

Also, certain more proprietary networking hardware needs firmware blobs that aren't in the kernel (most have a package, but you still need to install that). I'd suggest going back to live mode, mounting and chrooting everything, and then emerging connman + hwinfo

Hwinfo to figure out what wlan hardware/ kernel driver / device name you have very easily.

I thought all the cool kids use it

That would be a border...

I want it to be a mid-long term project where I rice it out and get familiar with the GNU/Linux way, once I am ready I'll switch.

I just reset my windows to factory settings, made a partition and put arch on a pendrive anything I need to know before I go forth?

Attached: 1508282245299.png (500x357, 153K)

I mean like this.

Attached: 2568kf5.png (200x202, 43K)

Yes, that is a border.
X11 renders squares. What you want is a border that rounds it

It probably doesn't exist just because it's kind of stupid since it could cut off stuff you actually want to see.