codygman has quit [Read error: Connection reset by peer]
codygman has joined #nixos
markus1189 has quit [Ping timeout: 246 seconds]
codygman has quit [Read error: Connection reset by peer]
proofofkeags has quit [Remote host closed the connection]
codygman has joined #nixos
sigmundv has quit [Ping timeout: 256 seconds]
proofofkeags has joined #nixos
as3 has joined #nixos
edwtjo has quit [Ping timeout: 256 seconds]
h0m1 has quit [Ping timeout: 244 seconds]
<as3>
What is the proper way (if any) of replacing usage of glibc with musl on my local machine? Something to the effect of nixpkgs.config.packageOverrides = pkgs: pkgs.pkgsMusl;
<raghavsood>
Perfect, that's what I figured, but sanity checks are easier before things break
<raghavsood>
Thanks!
boxscape has quit [Quit: Connection closed]
<hexa->
if some package does not work on python2 anymore set disabled conditionally
<raghavsood>
It still works with python2, but it just felt wrong having something use it when its being removed everywhere else - the current clickhouse derivation doesn't use python at all, so no real issues with breaking someone's set up by using python3 straight away
<raghavsood>
The clickhouse build process runs a number of shell scripts which use `git rev-parse` to get the git repo root directory - I don't want to `leaveDotGit` just for this - is there a better approach than having a patch file/substituteInPlace to update those to use `$src`?
<cole-h>
AFAIK, there is nothing better than `sed`ing/patching them out.
pjt_tmp has joined #nixos
<raghavsood>
`substituteInPlace` seems nicer than `sed`, the current clickhouse derivation already uses it for some `postInstall` fixes. A git patchfile might be more readable though, instead of semi-random shell commands replacing stuff
kleisli_ has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @thoughtpolice pushed commit from @cpu to master « honggfuzz: init at 2.2 »: https://git.io/JJqvY
tskc[m]1 has joined #nixos
as3 has quit [Quit: Connection closed]
<KarlJoad>
So, I'm having a strange issue with my GPUs and my monitor outputs. Does someone think they can help me?
<cole-h>
Try asking your question(s)/explaining your situation first, and if someone can help, they'll chime in.
<KarlJoad>
I have 2 GPUs, one for each monitor. Both are nVidia. I use KDE. I only get output on 1 monitor, from the main GPU.
<{^_^}>
[nixpkgs] @expipiplus1 closed pull request #91079 → liburing: bump version and fix cross compilation → https://git.io/JfFMm
proofofkeags has quit [Remote host closed the connection]
<KarlJoad>
The second GPU is visible to the system, but cannot have a virtual screen added to it, i.e. it is not present in KDE's System Settings.
proofofkeags has joined #nixos
<KarlJoad>
However, I know this setup works, because a booted Ubuntu ISO correctly set up the monitors to work the same way they work on Windows.
<KarlJoad>
So, I've been digging through my settings, but I don't know where to go from here.
<KarlJoad>
I think it's an X thing, because if I use the nVidia drivers, to get the nVidia X Control Center, I can add another virtual screen for the second GPU.
azure1 has quit [Read error: Connection reset by peer]
azure1 has joined #nixos
orivej_ has quit [Ping timeout: 264 seconds]
orivej has joined #nixos
<KarlJoad>
I've already tried messing with `xrandr --setprovideroutput x y`, which did something. It made both monitors go completely gray and the X session being useless.
proofofkeags has quit [Ping timeout: 240 seconds]
moet has quit [Ping timeout: 246 seconds]
proofofkeags has joined #nixos
KarlJoad` has joined #nixos
KarlJoad` has left #nixos [#nixos]
KarlJoad` has joined #nixos
KarlJoad has left #nixos ["ERC (IRC client for Emacs 26.3)"]
orivej has quit [Ping timeout: 272 seconds]
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<{^_^}>
[nixpkgs] @mweinelt opened pull request #92679 → wallabag: apply patches to add missing migrations in initial setup → https://git.io/JJqf8
kvda has joined #nixos
stree has joined #nixos
hke has quit [Read error: Connection reset by peer]
<gueorgui>
Hi, I've paired a bluetooth headset with my machine and connected to it/trusted it, but no pulseaudio sink/source appear for it. I'm using the pulseaudioFull package, and I can see the bluetooth pulseaudio plugins. Any ideas what could go wrong? I'm not using a DE and using bluetoothctl and the blueman applet.
<{^_^}>
[nixpkgs] @petabyteboy pushed commit from @r-ryantm to master « gortr: 0.14.4 -> 0.14.5 (#92689) »: https://git.io/JJqml
jengbers has joined #nixos
arahael1 has joined #nixos
Arahael has quit [Ping timeout: 272 seconds]
pjt_tmp has quit [Ping timeout: 256 seconds]
stree has joined #nixos
stree has quit [Client Quit]
stree has joined #nixos
stree has quit [Client Quit]
stree has joined #nixos
stree has quit [Client Quit]
sigmundv has joined #nixos
thc202 has joined #nixos
<patagonicus>
gueorgui: Not really sure what is happening, but for my BT headset I can right click on it in blueman and switch the profile between headphones, headset (mic + speakers) or off - maybe yours is set to off for some reason?
zanc has quit [Ping timeout: 272 seconds]
zanc has joined #nixos
domogled has joined #nixos
noudle has joined #nixos
orivej has quit [Ping timeout: 256 seconds]
orivej has joined #nixos
marsh has quit [Remote host closed the connection]
<gueorgui>
patagonicus: Oh I don't have that option, I wonder if I set it up wrong somehow
mounty has quit [Remote host closed the connection]
<patagonicus>
Hmm. I didn't really do much other than enabling bluetooth.
orcus has joined #nixos
orivej has quit [Ping timeout: 256 seconds]
orivej has joined #nixos
orcus has quit [Client Quit]
davidv7_ has joined #nixos
<gueorgui>
patagonicus: Oh I see it now. Does your headset also only have super bad sound quality in headset mode, and good sound quality in high fidelity mode?
<gueorgui>
Is there a way to have good sound quality and still have the bluetooth input available?
<gueorgui>
This seems like such a basic thing...
orcus has joined #nixos
me has joined #nixos
me is now known as Guest5357
davidv7 has quit [Ping timeout: 256 seconds]
fendor has quit [Remote host closed the connection]
kreyren has quit [Remote host closed the connection]
kreyren has joined #nixos
<gueorgui>
patagonicus: Thanks! This kind of sucks, but I guess that's how it is. Maybe using the laptop's internal mic instead of the headset's mic is ok.
<clever>
gueorgui: that lets you hear the cpu fan and every keystroke!
<raghavsood>
gueorgui: I ended up buying a USB dynamic mic to go along with my QC35 - linux audio management is a pain, bluetooth is also a pain, putting them together is unwise
<gueorgui>
raghavsood: That really sucks, I don't want to travel with a usb mic :(
bennofs_ has quit [Ping timeout: 256 seconds]
<clever>
gueorgui: main downside of the G533, is that you need the usb dongle to make it work, and the headset is a big large for travel
<{^_^}>
[nixpkgs] @xaverdh opened pull request #92697 → jabcode: init at git-2020-05-13 → https://git.io/JJqZO
<raghavsood>
gueorgui: I avoid that issue by never leaving home :) But really, when I do travel, it's easy enough to pack into my laptop back, uses a standard USB cable, unlike some XLR only options
<gueorgui>
I love Linux and I love working in i3 and I love Nix and everything, but shit like this makes me want to go back to macos. They might be locking the OS down a bit more every year, but at least things work more often than not. It's a frustrating state of things. Maybe I came to linux with too high expectations...
<clever>
gueorgui: its just vendor lockin, adding special features that only their hw can use
<gueorgui>
Yeah, I understand that.
kvda has joined #nixos
kvda has quit [Client Quit]
<ashkitten>
G533?
FRidh has quit [Quit: Konversation terminated!]
<clever>
ashkitten: logitech g533
* ashkitten
wrote a linux program for controlling the g933
<clever>
ashkitten: whats there to control? it has fewer options then the older headset i have
<ashkitten>
it works for g533 apparently, someone added the usb devire id
<ashkitten>
lighting, sidetone, eq profile
<ashkitten>
useful for getting g933 battery level
<clever>
battery level would be useful
<ashkitten>
isn't g533 wired?
<ashkitten>
i forget
<clever>
i tried to get that from the older headset, but all i could do was read the buttons
<l-as>
I'm going to make a PR to fix freeimage, but I am wondering about what the preferred way of including patches is. I need all the patches in https://src.fedoraproject.org/rpms/freeimage/tree/master, and am thinking that I could just do a `fetchpatch`, or is that discouraged? Is just including the files in the nixpkgs repo better?
<l-as>
FWIW I don't know what license those patches are under either
<adisbladis>
Those patches can very reasonably be assumed to be GPL.
meh` has quit [Ping timeout: 258 seconds]
cosimone has quit [Quit: Quit.]
meh` has joined #nixos
<l-as>
thanks
wnklmnn has joined #nixos
<l-as>
using fetchpatch then
<{^_^}>
[nixops] @adisbladis pushed to master « shell/default.nix: Use the flake-compat default attribute »: https://git.io/JJqcf
meh` has quit [Ping timeout: 256 seconds]
Guest5357 has quit [Ping timeout: 260 seconds]
FRidh has joined #nixos
aszlig_ is now known as aszlig
arahael1 has quit [Ping timeout: 272 seconds]
werner291 has joined #nixos
arahael1 has joined #nixos
armin has quit [Remote host closed the connection]
<patagonicus>
What's the format for boot.kernelPatches.patch? Is there an example somewhere I can look at? I have some changes I'd like to apply to the kernel before building.
<{^_^}>
[nixops] @adisbladis closed pull request #1318 → Add plugin hooks required for the encrypted links plugin → https://git.io/JfIDH
<{^_^}>
[nixpkgs] @iamthememory opened pull request #92704 → gcc: Fix manpage symlinks when architecture tuples are prefixed to the file name → https://git.io/JJqCb
domogled has quit [Remote host closed the connection]
domogled has joined #nixos
cjpbirkbeck has joined #nixos
knupfer has quit [Read error: Connection reset by peer]
<ikwildrpepper>
aszlig on frontpage of HN (ip2unix)
<maier>
Given a set containing some packages as attributes, say { inherit foo bar baz; }, can I use this as the basis to "generate" the lambda `p: [ p.foo p.bar p.baz ]`? I manage the set of packages at one place, but need this function at another place to yield the same packages as a list
<maier>
It's tedious to always have to think of inserting new packages in both places
<maier>
infinisil: I'll check it out, thanks. Background: I need the lambda for `pkgs.haskellPackages.shellFor`. Maybe there's another possibility to pass them there.
<{^_^}>
[nixpkgs] @B4dM4n opened pull request #92710 → wrapQtAppsHook: wrap binaries in all outputs → https://git.io/JJq86
<AlpineLlama>
Hey all! Just discovered NixOS and excited to try it ! I'm having trouble with the install though. when booting from usb it tells me /dev/sdb: open failed: No medium found
<AlpineLlama>
Not sure what to do with this
<infinisil>
maier: Yeah what I suggested should work then
<AlpineLlama>
I can press i to get into an interactive term. then fdisk tells me there are only sda and sdc
<clever>
AlpineLlama: can you try a different usb port? or a different usb stick?
<AlpineLlama>
@clever will do
<AlpineLlama>
@clever: different usb port gives the same error
<{^_^}>
[nixpkgs] @LEXUGE opened pull request #92711 → minecraft-server: add declarative setting for ops.json → https://git.io/JJq4f
alp has quit [Remote host closed the connection]
<AlpineLlama>
different usb stick also gives me the same error
alp has joined #nixos
<AlpineLlama>
must have something to do with my hw config
jasom has joined #nixos
<AlpineLlama>
I can try to install with the livecd from another distro right ?
<AlpineLlama>
was hoping to give it a spin first though
<clever>
AlpineLlama: you can also try the kexec route
<AlpineLlama>
I would like to understand why the live cd doesn't boot though
domogled has quit [Ping timeout: 272 seconds]
<clever>
AlpineLlama: if you boot with the kexec method, then you have a livecd-like env running entirely from ram, and can poke at the usb more with the benefit of the whole OS working
eoli3n_ has joined #nixos
<{^_^}>
[nix] @domenkozar pushed 2 commits to 2.3-maintenance: https://git.io/JJq4i
eoli3n__ has joined #nixos
<AlpineLlama>
@clever: running in debug mode It says that it can't open the /dev/loop0. any idea why ?
p_m[m] has joined #nixos
<clever>
AlpineLlama: it creates that after it mounts the main filesystem on the disk
<makefu>
hey all, i am currently stumbling about the issue that pkgs.fetchgit ignores the system certificates which results in "unable to get local issuer certificate". i am behind an intercepting proxy and i've set the proxy correctly, however this specific host is internal with its own tls certificate. what would be a good way to debug this?
<clever>
makefu: most of the fetch functions also ignore how valid ssl certs are
<clever>
since they are already validating the final hash of the product
<infinisil>
maier: There's also nixpkgs-fmt, which I think is used by more people
<infinisil>
And seems to be better maintained
stree has joined #nixos
<makefu>
clever: i made it a step further and it seems to be related to the proxy thing. i've set no_proxy and in the shell and with nix-prefetch-git the fetching works, however it behaves differently with pkgs.fetchgit
stree has quit [Client Quit]
lpsmith has quit [Max SendQ exceeded]
<infinisil>
(nixfmt on the other hand probably doesn't need much maintenance due to its simplicity, which is also a good thing)
<AlpineLlama>
@clever: so it seemed to be how I was flashing the hard drive that was problematic. I was using pv image.iso > /dev/sdd
<AlpineLlama>
after using balena etcher instead It ended up booting
<AlpineLlama>
that command has worked in the past though. guess it might be less stable ?
<clever>
AlpineLlama: weird, i would expect both of those to work the same way
<AlpineLlama>
Yeah I was surprised
<AlpineLlama>
but the first one doesn't perform any verification so idk
<makefu>
clever: okay i've actually got it working, (by setting sourceRoot="." and fetchurl { name = "repo.tar.gz" ...}, however i'd love to simply use fetchgit without all this hassle :)
iyzsong has quit [Read error: Connection reset by peer]
iyzsong has joined #nixos
VF has joined #nixos
<robodojo>
the installer doesn't do partitioning or formatting. but then later i read that configuration.nix contains that info. so is it possible to set up a system, so that data's in place, then reinstall it and in the installation shell prompt rather than configuring that stuff implicitly, somehow use the configuration.nix?
<gchristensen>
robodojo: no, NixOS's configuration.nix doesn't do any partitioning
7GHABVQQD has quit [Remote host closed the connection]
<gchristensen>
one of the goals of NixOS's options is you should be able to roll forward and roll back and forward again to any point, and get the same result -- and partitioning isn't really compatible with that
CMCDragonkai has joined #nixos
<fendor>
where can I find documentation of common functions? E.g. I want to know what parameters the functon haskell.packages.ghc883.mkDerivation accepts
alexherbo2 has quit [Read error: Connection reset by peer]
<robodojo>
gchristensen so what you have to set up manually (partitions/formatting) isn't the kind of fs related data put into configuration.nix?
<fendor>
infinisil, there definitely should be. Discovering the API in nixpkgs is a horrible experience. Isnt there currently an iniative to improve this?
<clever>
robodojo: that tends to go into hardware-configuration.nix, which is auto-generated by nixos-generate-config, based on what is mounted
<robodojo>
@fendor would be nice if pkgs were indexed as part of channel builds and the index was used to feed IDE extensions giving autocomplete
<infinisil>
fendor: Not directly related, but there's rfcs#67 which would help a bit
<robodojo>
fendor ya and the single page GIANT manual files too
<fendor>
robodojo, which still often do not document what the functions actually do
<robodojo>
so none of the initial `mkfs.ext4 -L nixos /dev/sda1` stuff could be put into configuration.nix?
alexherbo2 has joined #nixos
jengbers has quit [Remote host closed the connection]
<clever>
robodojo: correct, nixos can only begin acting on configuration.nix after the drive has been formatted and mounted
<robodojo>
seems like if symbols in the entire nix ecosystem were indexed not only could you feed that to IDE autocomplete but could throw a web gui at it too
jengbers has joined #nixos
<robodojo>
clever well no that's not what i'm asking
<robodojo>
what i'm wondering is if that file could hold info that could be used to perform the initial setup, separate from a rebuild
<robodojo>
like if partition/formatting were in a .nix, maybe you could do nixos-setup with it and not have to run really any system admin commands
<clever>
robodojo: have you seen my justdoit stuff?
<robodojo>
but maybe that's impossible for one of the reasons gchristensen said
<clever>
robodojo: this shell script will partition, format, mount, and install nixos, in a single command
<robodojo>
ya! and does it get its data driving that (sizes etc) from configuration-hardware.nix?
<gchristensen>
pamplemousse: I like your nickname, it makes me smile when I see you join / part
knupfer has quit [Ping timeout: 246 seconds]
<clever>
robodojo: it gets the sizes from another configuration.nix, that is used to build the livecd
jengbers has quit [Read error: Connection reset by peer]
<robodojo>
why isn't that in the base nixos?
<robodojo>
like nixos-init or whatever
<robodojo>
then if im not mistaken we could set up machines with like 2 commands and 0 direct admin of the system
archH has quit [Quit: Connection closed]
<gchristensen>
because it is not truly declarative
<gchristensen>
it is more chef/ansible/puppet's idea of declarative
jaewest304[m] has joined #nixos
<clever>
some one-time setup commands, that cant be changed after the fact
<gchristensen>
right
zebrag has quit [Ping timeout: 240 seconds]
<gchristensen>
I think there could be a neat project for partitioning and setting up disks, but we'd want to be really careful and probably would want it to be in addition to NixOS, not part of NixOS
<clever>
i was thinking it would be defined at the nixops layer
<robodojo>
ya
zebrag has joined #nixos
<robodojo>
makes sense
knupfer has joined #nixos
<{^_^}>
[nixos-search] @garbas merged pull request #116 → dont show browser autocomplete for query input → https://git.io/JJqRT
<{^_^}>
[nixos-search] @garbas pushed to master « dont show browser autocomplete for query input (#116) »: https://git.io/JJq2B
<{^_^}>
[nixos-search] @garbas pushed 0 commits to remove-autocomplete-from-input: https://git.io/JJq2R
asymptotically has quit [Quit: Leaving]
NeoCron has quit [Ping timeout: 260 seconds]
pamplemousse has quit [Ping timeout: 240 seconds]
<robodojo>
opinion of skipping a swap partition for ssd systems?
<robodojo>
apparently it wears the ssd out?
<gchristensen>
I wouldn't worry about wearing out the SSD, SSDs are good and a lot of that concern is FUD from when SSDs were still really bad and underprovisioned
<gchristensen>
that said, I find swap to be undesirable in almost every case, and I would almost always prefer some program be killed than make my computer unusable while it swaps
<robodojo>
so running out of mem and not having swap won't bring the system down just the program requesting the mem?
veleiro`` has quit [Remote host closed the connection]
<{^_^}>
[nixos-search] @garbas pushed to include-query-in-suggestions « include query item in search suggestions »: https://git.io/JJq27
<LnL>
yeah I don't use swap either, unless the system is tight on ram
<{^_^}>
[nixos-search] @garbas opened pull request #117 → include query item in search suggestions → https://git.io/JJq25
<robodojo>
ill start running without swap and see if i run into problems thanks
red[evilred] has joined #nixos
<red[evilred]>
yeah - memory is so cheap now there really shouldn't be any need for swap at all
<red[evilred]>
we're not running SunOS in 512M anymore
<{^_^}>
[nixpkgs] @jtojnar pushed commit from @dasj19 to master « gnome-shell: 3.36.3 -> 3.36.4 »: https://git.io/JJqaq
nixuser has quit [Read error: Connection reset by peer]
noudle has joined #nixos
<robodojo>
are there any special instructions to disabling swap or just don't create a swap partition and that's that?
<wrl>
robodojo: i haven't had a swap partition in the last... 6 or 7 years
<wrl>
RAM's cheap
<wrl>
buy more
<simpson>
Swap is an opt-in, not an opt-out. If you don't configure swap, then you won't have swap.
<robodojo>
ya i got 64gb
<wrl>
then you don't need swap
<wrl>
lol
<LnL>
that's it, or you don't run swapon or whatever it is
<robodojo>
thanks!
<wrl>
i don't have swap on my 24gb lappy and it's never caused problems
<wrl>
64gb is easy
<patagonicus>
You do need swap for suspend-to-disk/hibernation. Otherwise it depends on what you do and how much RAM you have. With 64GB I doubt you'll need it unless you want to s2disk (and then you'll need 64GB of swap …).
<red[evilred]>
the only place I have swap is in the cloud
<red[evilred]>
and that's because they cloud doesn'treally give me any choice
<red[evilred]>
well, affordable choice
nabataeus has joined #nixos
<simpson>
patagonicus++ excellent point about hibernation.
<{^_^}>
patagonicus's karma got increased to 1
<nabataeus>
Does anyone know how to enable completion on `nix-repl' in Emacs?
marsh has joined #nixos
<robodojo>
ya i read about that but don't need hibernation i don't have any laptops
<robodojo>
thanks for bringing it up
<robodojo>
red[evilred] why not just use vps so you can choose disk setup?
<{^_^}>
[nixpkgs] @vbgl pushed commit from @marsam to master « ocamlPackages.ppx_yojson_conv_lib: init at 0.14.0 »: https://git.io/JJqai
catern has joined #nixos
camsbury has joined #nixos
<{^_^}>
[nixpkgs] @KyleOndy opened pull request #92718 → spaceship-prompt: enable on darwin → https://git.io/JJqaS
maier has quit [Ping timeout: 256 seconds]
<red[evilred]>
all I'msaying is that if you want morethan 2G of RAM - cloud gets expensive
<robodojo>
anyone find 512mb root partition not big enough in practice?
<gchristensen>
will your /nix/store be on that root partition?
eoli3n_ has joined #nixos
<gchristensen>
if /nix and /home are not on the root partition, you're probably fine
<gustavderdrache>
what about /var?
<robodojo>
whatever's default. i just see the install guide recommending 512mb and was curious
<gchristensen>
I think it is suggesting 512m for the /boot partition
<robodojo>
ya i'm sorry, boot
<gchristensen>
512m should be plenty for /boot :) mine is 208M and has lots of boot history
<robodojo>
thanks!
pr06lefs has joined #nixos
<adisbladis>
For reference I'm right under 500M with 25 different initrds and 20 kernels
<clever>
64mb for /boot is a bit too low, i often have trouble getting 2 generations to fit
<adisbladis>
512M will be fine if you somewhat limit the number of generations
<adisbladis>
I GC as rarely as I possibly can
pamplemousse has joined #nixos
<robodojo>
why does the guide say to create boot partition after root and swap partitions? that gives it highest partition number but it's at the start of disk
<{^_^}>
[nixos-search] @garbas pushed to include-query-in-suggestions « when there is one result in suggestions and is the same as query we don't show results »: https://git.io/JJqV0
<clever>
robodojo: i would always create boot first, swap 2nd, and root last
<gchristensen>
robodojo: I don't know :)
<robodojo>
ya me too!
<{^_^}>
[nix] @edolstra pushed 5 commits to latest-release: https://git.io/JJqVg
<{^_^}>
[nix] @edolstra pushed 0 commits to refs/tags/2.3.7: https://git.io/JJqVa
<patagonicus>
robodojo: The order of mkfs doesn't matter.
<robodojo>
ok thanks
<clever>
only the order of mounting really matters
erasmas has joined #nixos
<patagonicus>
Hmm. I'm trying to use an LVM volume (nothing system relevant, just mounting to $HOME/foobar), but it looks like NixOS is not activating the volume. I use /dev/$vgname/$lvname as the device in my configuration.nix. My / is on a regular partition on a different physical device. This is on an ARM machine, in case that matters.
<patagonicus>
On boot systemd is timing out waiting for the device and now login also seems to hang.
<robodojo>
clever the order of mounting matters? should boot be mounted first?
<selfsymmetric-mu>
Ultimately I'm breaking on `libpq-fe.h: No such file or directory` b/c ostensibly at some point that was moved from "lib" to "dev". But now I can't get at it.
<clever>
robodojo: because when you mount to /mnt, it hides anything that was previously at /mnt, such as /mnt/boot/
<robodojo>
oh weird. so if you mount boot last it'll be the only thing showing?
pr06lefs has quit [Ping timeout: 244 seconds]
bahamas has joined #nixos
<clever>
robodojo: no, boot has to be mounted after root
<clever>
robodojo: basically, sort all things by the length of the path your mounting to (in terms of number of /'s), and then mount from shortest to longest
<clever>
so /mnt goes before /mnt/boot and /mnt/home
<selfsymmetric-mu>
oh dev got removed again??
<selfsymmetric-mu>
okay
<selfsymmetric-mu>
Now I don't know where to get a libpq-fe.h…
<clever>
selfsymmetric-mu: the stdenv should just auto-detect the right output when you put `builtInputs = [ postgresql ];`
<robodojo>
thanks clever
<slabity>
,locate libpq-fe.h
<selfsymmetric-mu>
clever: Thanks! Will that work in a nix-shell that needs it to build a haskell project with cabal?
VF has quit [Remote host closed the connection]
<clever>
selfsymmetric-mu: yeah, just add it to the inputs or the -p param
<slabity>
Is the locate command not working?
cole-h has joined #nixos
wfranzini has joined #nixos
<clever>
slabity: it only works if , is the first character of the msg
johnny101m has quit [Quit: -a- Connection Timed Out]
<selfsymmetric-mu>
clever: Thanks! I've added it, crossing my fingers on the build now.
johnny101m has joined #nixos
<slabity>
,locate libpq-fe.h
<{^_^}>
Found in packages: postgresql95, postgresql96, postgresql_10, postgresql_11, postgresql_12
<slabity>
I must have added a space or something?
<selfsymmetric-mu>
Oh interesting. It might not be in the unnumbered postgresql package...
<robodojo>
my system has 5 HDs and i only want to use 1 of them for nixos. can i hide the other drives from nixos somehow?
<selfsymmetric-mu>
Holy moly that worked. slabity++ what did we learn?
<{^_^}>
slabity's karma got increased to 9
<selfsymmetric-mu>
Okay I mean I learned to use `nix-locate`.
<selfsymmetric-mu>
Instead of just `fd` like a barbarian.
<slabity>
selfsymmetric-mu: No problem. Just run `nix-index` to gather information of what's in packages, and then you can 'nix-locate' whatever files you need
<slabity>
Super useful when trying to find dependencies or missing headers
<slabity>
I don't think a bug that would mount and write files to your disk is likely... Especially considering it requires root-access to do so
<robodojo>
what would blacklisting kernel module buy me?
<gchristensen>
yeah, I think your option is to use a hypervisor, though that is a pretty high degree of paranoia
<colemickens>
I think tails blocks internal HDDs from appearing somehow
<gchristensen>
the system would not know what to do with devices of that type
<slabity>
It would prevent any device that uses that kernel module from loading
<robodojo>
the drives i want to hide are all nvme fwiw
<simpson>
Removing kernel modules probably won't work, unless your drives are on different bridges with different chipsets.
<simpson>
robodojo: What have you done in the past with Linux to enforce this? Or is this a new desire?
<robodojo>
never used linux before, coming from macos
<robodojo>
fwiw apple had a bug that wiped attached drives on a system upgrade
<robodojo>
so it's not like i'm out of reality
eoli3n__ has joined #nixos
<simpson>
Don't mount the drives you don't want touched. Most folks are generally going to accidentally do things to those drives as root anyway; preventing *that* really does require unplugging the data cable.
<simpson>
(It's not like Apple's known for their code quality~)
<LnL>
there's nothing in nixos that messes with devices AFAIK
<slabity>
Apple had a bug that badly?
<robodojo>
yep
<slabity>
Jegus I never heard of that
<LnL>
all the partitioning, etc. is up to the user
<robodojo>
simpson so if i don't manually mount the other drives they won't be writable?
jakobrs has left #nixos ["WeeChat 2.7.1"]
eoli3n_ has quit [Ping timeout: 256 seconds]
<slabity>
robodojo: Pretty much...
<slabity>
I suppose it's possible to have a root-user program directly write to /dev/sdb or whatever
<robodojo>
:/ the unix sec model leaves a lot to be desired
<slabity>
But if a program is doing that then there's not really anything stopping it from doing anything else...
<simpson>
robodojo: If a drive is unmounted, and its device node is owned by root, then it's a *Linux* bug if your non-root user can write to it, yes.
<slabity>
robodojo: I don't think it's possible to have a security model that doesn't do that...
<robodojo>
i'll blacklist the nvme kern mod and never mount the drives, that's my plan
<slabity>
And still allow your hardware to be usable
<robodojo>
thanks everyone
<simpson>
slabity: seL4 and Genode might be worth looking into; in general, capability-oriented models can be *extremely* fine-grained and detailed about permissions.
<robodojo>
ya
<robodojo>
boot.blacklistedKernelModules nice
kalbasit has joined #nixos
<robodojo>
we don't change hardware-configuration.nix right? these kernel settings should be overrides to those defaults and put into configuration.nix iiuc
<slabity>
It says you shouldn't change it...
proofofkeags has joined #nixos
<simpson>
hardware-configuration.nix is definitely changeable, and you shouldn't feel strange about touching it. It's *autogenerated*, but not immutable.
<slabity>
But it's fine if you do
<slabity>
I do at least
<slabity>
Infact I just copy the hardware-config into my normal configuration.nix file
<robodojo>
any benefit to changing it? what's it autogenerated from? only if i run the config generate command?
<simpson>
It's hard to write by hand, that's all. I appreciate that we have a tool which automates that stuff; otherwise, we'd have to do it by hand like in Gentoo installation.
jakobrs has joined #nixos
<clever>
robodojo: any time you run nixos-generate-config, the file gets overwritten
<clever>
robodojo: you can choose to just never run nixos-generate-config again, and edit it all you want
<robodojo>
clever thanks. can i also leave it untouched but override it in configuration.nix?
<clever>
yep
<robodojo>
is that good/bad?
<clever>
mostly up to personal preference
drakonis has joined #nixos
<robodojo>
i don't see the same warning in configuration.nix so it's not volatile to running nixos-generate-config right?
<robodojo>
what's in hardware-configuration.nix boot.initrd.availableKernelModules is only what devices were detected, but NONE of those will be actually loaded until i add them to boot.initrd.kernelModules or boot.kernelModules (if it's not used to load filesystem) right?
<{^_^}>
[nixpkgs] @RaghavSood opened pull request #92724 → [20.03] rippled: update to 1.5.0 as 1.4.0 no longer syncs → https://git.io/JJqoy
justanotheruser has joined #nixos
KarlJoad` has joined #nixos
<Henson>
hi everyone, is there a way to find the closure of a shell.nix file to copy the files over to another computer to recreate the shell there? If I use mkShell in the shell.nix file, I can't run nix-build on it to get an output with which to find the closure. Any suggestions?
<jakobrs>
I believe I've found a privilege escalation attack in the libvirtd module. It allows a local (AV:L) attacker to access qemu:///system without authenticating.
<jakobrs>
(bypassing polkit)
KarlJoad` has quit [Remote host closed the connection]
NeoCron has joined #nixos
emilis has joined #nixos
<Henson>
jakobrs: that sounds serious. Is this a NixOS-specific problem, or a libvirtd problem
NeoCron has quit [Remote host closed the connection]
<jakobrs>
It's a simple oversight in the nixos module. So yes, it's nixos-specific.
<robodojo>
what's in hardware-configuration.nix boot.initrd.availableKernelModules is only what devices were detected, but NONE of those will be actually loaded until i add them to boot.initrd.kernelModules or boot.kernelModules (if it's not used to load filesystem) right?
bahamas has joined #nixos
bahamas has joined #nixos
bahamas has quit [Changing host]
KarlJoad` has joined #nixos
<clever>
robodojo: availableKernelModules can also be auto-loaded by udev, when the relevant device is detected
<KarlJoad`>
How does NixOS write the x11/xorg server configuration file? I'm having issues with a second GPU'
b1000101 has quit [Quit: Lost terminal]
sheeldotme has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<KarlJoad`>
second GPU's monitor not being used, but the GPU is detected.
<KarlJoad`>
I've already attempted using `xrandr --setprovideroutput x y` but that doesn't work.
<Henson>
robodojo: take a look at the NixOS options website: https://nixos.org/nixos/options.html under those option names. The availableKernelModules are the ones put into the initrd and are loaded on demand. The ones in kernelModules are always loaded by initrd
<KarlJoad`>
Henson: They are both nVidia GPUs, a 1080 and a 980Ti. Both using the Nouveau drivers (although that is not the problem).
alp has quit [Remote host closed the connection]
alp has joined #nixos
<slabity>
KarlJoad: Do both sets of connectors appear in your xrandr output?
<slabity>
Or just one?
<KarlJoad`>
I have just one X Screen showing up. That screen has all the correct outputs for the 1080. Nothing for the other.
<slabity>
KarlJoad: So if you just run `xrandr`, than only one set of connectors appear, correct?
<Henson>
KarlJoad`: what window manager or desktop environment are you using?
<robodojo>
enabling boot.loader.efi.canTouchEfiVariables means that nixos could modify the hardware for any OS using the comp right?
cript0nauta has joined #nixos
<KarlJoad`>
slabity: Correct. All outputs of the 1080 are shown. None of the 980Ti's are present. Although, the GPUs are present to the system.
<slabity>
robodojo: Yes. Specifically it will allow NixOS to add a boot-entry with `efibootmgr`
<KarlJoad`>
Henson: KDE/Plasma 5. Also did a quick test with i3 and got the same result.
<clever>
KarlJoad`: that implies xorg cant see the card, and nothing xrandr does can change things
<robodojo>
slabity oh so i wont have to choose which disk to boot using f8?
<slabity>
robodojo: If `f8` you mean going into your BIOS and selecting the boot-disk, then yes...
<KarlJoad`>
clever: That's what I was thinking. I don't know how to troubleshoot this problem, as Ubuntu had no issues figuring this out from a live boot disk.
jb55 has quit [Remote host closed the connection]
kalbasit has joined #nixos
matias_ has joined #nixos
<robodojo>
slabity right after POST f8 will let you choose between any connected storage device to boot from, don't have to go into bios
<clever>
KarlJoad`: read the xorg.conf file and logs that both nixos and ubuntu are using
cript0nauta has quit [Read error: Connection reset by peer]
<Henson>
if you're using KDE, can you use the display properties in system settings to see if it detects the other monitors. If you're using the prioprietary NVIDIA driver you should also be able to use the "nvidia-settings" GUI to do some investigation.
<slabity>
KarlJoad: So going by what I see on the freedesktop website, Nouveau only allows multiple-gpu-monitors using Xinerama
<KarlJoad`>
slabity: But, from your investigation, I either have to use the proprietary NVIDIA drivers, or use Xinerama with Nouveau, right?
<slabity>
When did {^_^} get sass?
<cole-h>
Note that I use the hooks because I only have a single GPU. May or may not be necessary if you have multiple.
user_0x58 has quit [Quit: Leaving]
<cole-h>
slabity: A while back :D
<robodojo>
"On 64-bit systems, if you want OpenGL for 32-bit programs such as in Wine" but isn't wine 64bit now?
<slabity>
KarlJoad: That is correct. If you want both GPUs to dispaly to a monitor then you gotta have Xinerama
<slabity>
Or prop-drivers
<slabity>
Assuming the freedesktop article is up-to-date
<slabity>
Which is like a 50/50 chance :/
<patagonicus>
How can I use ccache for kernels? Setting programs.ccache.packageNames = ["linuxPackages_latest"] gives me "error: attribute 'override' missing".
<worldofpeace>
slabity: it was diva'fied
<slabity>
worldofpeace: I don't know what that means...
werner291 has quit [Quit: werner291]
<KarlJoad`>
slabity: Gotcha. If I use the proprietary drivers, how can I let the nvidia-settings write to the xorg.conf file?
<slabity>
Oh, the bot
<worldofpeace>
the essence of having that sassy diva vibez
<KarlJoad`>
Or, how do I get the output of that into the NixOS-generated xorg.conf file?
<KarlJoad`>
Because that file has some default stuff in it by NixOS.
<clever>
patagonicus: linuxPackages_latest isnt a single package, its a set of every optional kernel module
ashesham` has joined #nixos
<slabity>
KarlJoad: `services.xserver.config` should contain your xorg.conf contents
<slabity>
If you do a `ps aux | grep "/bin/X"` you can probably look at the `-config` flag and see the path of your current xserver.conf
<KarlJoad`>
slabity: That's the setting I was working with, but it still left some default settings behind.
<slabity>
KarlJoad: Also take a look at `services.xserver.xrandrHeads` as that provides all the `Section "Monitor"` parts of the config...
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
<KarlJoad`>
I'll try with the `services.xserver.config` option again, and see if that works. That should just write to the output xorg.conf file directly, right?
<slabity>
KarlJoad: Basically, it will automatically generate the FontPaths and ModulePaths that you don't need to worry about... Then append your config to the end
<{^_^}>
[nixpkgs] @kalbasit pushed commit from @r-ryantm to master « atlantis: 0.13.0 -> 0.14.0 (#92674) »: https://git.io/JJqiA
<robodojo>
what's initrd mean? the rd part
<slabity>
robodojo: ramdisk
<robodojo>
oh
cole-h has quit [Quit: Goodbye]
<slabity>
robodojo: Before your OS gets fully booted, the kernel loads an initial ramdisk, which contains all the software required to mount and load your actual environment
<robodojo>
does the ramdisk go away after the OS is loaded?
<slabity>
robodojo: Yes. It switches to the new environment entirely
<robodojo>
ohh interesting
<slabity>
Your bootloader will say, "I'm going to run `/path/to/kernel/linux-bzimage` with `initrd=/path/to/initrd`", which your kernel accepts as an argument. The initrd contains the bare minimum to load the devices required to mount your actual disk (filesystem drivers, LUKS for encryption, LVM if you use that, etc)
red[evilred] has quit [Quit: Idle timeout reached: 10800s]
<robodojo>
so executing the kernel is just like invoking any other binary where you can pass arguments?
missionformilk has quit [Quit: leaving]
<slabity>
Pretty much. You can check `/proc/cmdline` to see what your kernel booted with
<patagonicus>
But now ccache is pulling in asciidoc, which pulls in dblatexFull and that wants to build latex, inkscape and a few other things. I doubt I'll end up saving time if I need to build all of these. Let's see if I can build ccache without asciiidoc …
growpotkin has joined #nixos
missionformilk has joined #nixos
ddellacosta has joined #nixos
fresheyeball has joined #nixos
<drakonis>
hmm, i'm having a bit of trouble figuring out whether i'm accessing the internet through a vpn
<fresheyeball>
So my struggles with node2nix are a thing of the past, sorta
<drakonis>
i need to make sure it works for developing a system
<drakonis>
nice.
<gchristensen>
it is a really hard question to answer, drakonis
<fresheyeball>
And I want to socialize what I learned.
<clever>
patagonicus: you usually dont use ccache with nix, since nix can already cache things
<drakonis>
i'll rephrase it then
<fresheyeball>
I submitted an issue with the expression updates that allow node2nix generated code to allow for source overrides
<drakonis>
i have connected to a vpn that's supposed to be authorized for a particular task
<fresheyeball>
however I think it would be even nicer if we could having something like overlays for node stuff
<drakonis>
but i can't tell whether it is going through there
<drakonis>
maybe wireshark would help?
<patagonicus>
clever: I know. But I plan on rebuilding a kernel a bunch of times for figuring out with commits from a fork I actually need. :)
<fresheyeball>
curious if anyone has thoughts
<Henson>
drakonis: you could check your routing table to see if the path to the IP you're working with are routed through your VPN.
justan0theruser has joined #nixos
bahamas has joined #nixos
bahamas has joined #nixos
bahamas has quit [Changing host]
<patagonicus>
clever: Upstream ccache docs mark asciidoc as optional and it built fine without it - currently running tests, though, so maybe it'll fail in the install phase.
<Henson>
drakonis: and yes, wireshark would help. If you monitor packets on that interface you'll be able to see
missionformilk has quit [Client Quit]
<Henson>
drakonis: you could also add something to your firewall to prevent packets to the target IP from leaving your internet-connected interface.
<fresheyeball>
also, I wonder if we can have a makeOverlayable function in stdenv, like there is makeOverridable
<fresheyeball>
is there a reason overlays are general like that?
Thra11 has quit [Quit: WeeChat 2.8]
<drakonis>
there's no dedicated vpn interface
<slabity>
fresheyeball: What do you mean? What would 'makeOverlayable' do?
<robodojo>
to make a new user with bash shell does users.users.<name> need to include anything or is bash the default?
<drakonis>
bash is the default, yes.
pr06lefs has joined #nixos
<drakonis>
if unset, bash is default
<robodojo>
thanks!
<Henson>
drakonis: hmm, I don't think I understand your question, then.
Mrmaxmeier has joined #nixos
<slabity>
Isn't `users.extraUsers` preferred?
justanotheruser has quit [Ping timeout: 265 seconds]
missionformilk has joined #nixos
knerten2 has quit [Quit: knerten2]
<fresheyeball>
slabity: makeOverlayable :: Set -> Set
<slabity>
I thought `users.users` were for system-level users
<drakonis>
hmmm, i had tried it on windows but couldnt tell whether it worked either
<patagonicus>
Hmm, yeah, fails during install because it tries to call asciidoc. :(
<fresheyeball>
adding in an overlay argument, where you can pass a list of overlays that update that set
<drakonis>
ah there is one
<fresheyeball>
kinda along the lines of how makeOverridable addis in an overrides member
<drakonis>
tun0
<robodojo>
slabity users.extraUsers Alias of users.users.
missionformilk has quit [Client Quit]
missionformilk has joined #nixos
<jakobrs>
gchristensen: Sorry for taking so long to respond
<drakonis>
wow this is pretty slow rip
<jakobrs>
I wanted to make sure I didn't say anything that could be considered incorrect
<drakonis>
or maybe it isnt possible to ping 8.8.8.8 with it
missionformilk has quit [Client Quit]
fendor has joined #nixos
alp has quit [Ping timeout: 265 seconds]
missionformilk has joined #nixos
<drakonis>
Henson: all is fine
<drakonis>
i'm just not sending any data through the vpn tunnel
<Henson>
drakonis: ok!
<slabity>
drakonis: Check your default route. Maybe only traffic going to your VPN network will go through the VPN
<drakonis>
i'm pinging using the tunnel interface and there's no reply
<drakonis>
how fun
<drakonis>
or maybe not, idgi
<drakonis>
i'm actually getting a reply here
sheeldotme has joined #nixos
<drakonis>
i havent done this kind of vpn stuff before, so pardon me
sheeldotme has quit [Client Quit]
plutes has quit [Ping timeout: 246 seconds]
<robodojo>
i'm sitting at installer shell prompt and after a while of being idle a console message starts being spammed every couple minutes, igb exceed max 2 second (sic) why's that?
<{^_^}>
[nixpkgs] @flokli pushed 2 commits to release-20.03: https://git.io/JJqPN
alexherbo2 has joined #nixos
pjt_tmp has joined #nixos
tazjin has joined #nixos
Ke has joined #nixos
<KarlJoad`>
slabity: Would it make more sense to try to translate the config that nvidia-settings generates into a `services.xserver.<thing>` nix expression?
<slabity>
KarlJoad: I'm not sure. I don't have any experience with the `nvidia-settings` tools or what they generate. I'm not sure what the difficulty of that would be
KarlJoad` has quit [Quit: ERC (IRC client for Emacs 26.3)]
<{^_^}>
[nix] @edolstra merged pull request #3790 → Cherry-pick "Use fragment size for autoGC capacity calculation" on 2.3 → https://git.io/JJtnN
<{^_^}>
[nix] @edolstra pushed 2 commits to 2.3-maintenance: https://git.io/JJqXy
<numkem>
ryantm: I just got pinged for one of my packages by your bot. It did what it was supposed to do, I've commented that it works just fine for me. Is there something else I need to do?
<gchristensen>
jakobrs: will do!
<ryantm>
numkem: Next time, please use the GitHub review interface to approve it, but other than that, no. I monitor all the r-ryantm PRs for maintainer approvals.
proofofkeags has quit [Remote host closed the connection]
eoli3n_ has quit [Ping timeout: 265 seconds]
eoli3n_ has joined #nixos
eoli3n_ has quit [Remote host closed the connection]
eoli3n_ has joined #nixos
cr4y1 has quit [Remote host closed the connection]
cr4y1 has joined #nixos
knupfer has quit [Remote host closed the connection]
knupfer has joined #nixos
<{^_^}>
[nix] @edolstra pushed to 2.3-maintenance « Bump version »: https://git.io/JJq1c
codygman has quit [Ping timeout: 272 seconds]
codygman has joined #nixos
eoli3n__ has joined #nixos
eoli3n_ has quit [Ping timeout: 240 seconds]
fresheyeball has quit [Quit: WeeChat 2.7.1]
<jlv>
Anyone know a good guide on running Sway in NixOS? I'm using Home Manager. Disabling `xession` and enabling `sway` in Home Manager resulted in a broken, small, white window. Disabling `xserver` in NixOS resulted in no login manager, and when I tried to run `sway` from the tty, I got error "[sway/server.c:47] Unable to create backendDRM deviceerer".
<jlv>
slabity: I didn't try that yet, because I wanted to manage Sway with Home Manger. Honestly, I'm still not clear on what is managed by root or user session in Wayland, as opposed to X11.
mdlayher has joined #nixos
alp has quit [Ping timeout: 272 seconds]
<MasseGuillaume>
Hey, I'm trying to build the telepresence derivation (a python package) from master, but I'm having issues with setuptools-scm, It says it cannot find the version. This is probably because it cannot get the git tag
<MasseGuillaume>
How can get a shell with the src fetched ?
<energizer>
MasseGuillaume: post your code and error
<slabity>
blv: It is a bit ambiguous... I typically just set `services.xserver.desktopManager.session = [ { name = "custom"; start = "" } ];`, then set `services.xserver.desktopManager.default = "custom";`
orivej has quit [Ping timeout: 272 seconds]
<slabity>
Then I just let home-manager handle it all
fendor has quit [Remote host closed the connection]
<slabity>
Oh, and let lightdm do the autologin stuff
<jlv>
slabity: would I use `services.xserver.desktopManager` even if I'm using Wayland?
<slabity>
I'm not sure... I haven't made the switch to sway because my laptop has an nvidia card :/
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
<slabity>
I wish we could take all the desktopManager stuff out of xserver
<slabity>
Actually I wish we could take everything out of xserver that isn't related to X11
<colemickens>
is there something like genAttrs that lets you rename the namevaluepair name?
<jlv>
slabity: are you saying `desktopManager` isn't a part of X11, even though it is in `xserver`?
<jtojnar>
colemickens mapAttrs'
alp has joined #nixos
<slabity>
jlv: Maybe? I'm not sure specifically. I know `services.xserver.libinput` and `services.xserver.displayManager` aren't necessarilly X11 related and can be used with Wayland
<MasseGuillaume>
energizer found it, it was missing ``deepClone = true;` for the fetchgit
<jtojnar>
yeah, the xserver should be renamed to session or something
<slabity>
jlv: And of course you need `videoDrivers` which is in `services.xserver`
fnords has joined #nixos
<jtojnar>
GNOME also can use wayland but you still enable it with services.xserver
c00w has joined #nixos
taktoa[c] has joined #nixos
<jlv>
Well, that's confusing.
sneakweb has joined #nixos
dvim has joined #nixos
acarrico has quit [Ping timeout: 265 seconds]
manveru has joined #nixos
aristid has joined #nixos
victorbjelkholm_ has joined #nixos
<jlv>
Do you know if I need `services.xserver.enable` to use `services.xserver.libinput` and `services.xserver.displayManager`? I'd rather not enable an X11 server if I don't need it.
<numkem>
ryantm: thanks for the info, I'll do that next time
mpickering has joined #nixos
omnigoat has joined #nixos
peelz has joined #nixos
domogled has joined #nixos
eacameron has joined #nixos
eyenx|m has quit [Quit: killed]
philipcristiano has joined #nixos
NekomimiScience has joined #nixos
eyenx|m has joined #nixos
stree has joined #nixos
jbetz has joined #nixos
tnks has joined #nixos
nand0p has joined #nixos
lightandlight has joined #nixos
etrepum has joined #nixos
pamplemousse has joined #nixos
kitemikaze_ has joined #nixos
mitchellh has joined #nixos
<slabity>
Does anyone know of any Github issue or Discourse discussion on making Wayland a first-class citizen in NixOS? From what I can tell it looks like it is impossible to configure a pure-Wayland system without also setting up a ton of X11 programs and configurations
<slabity>
I don't see anything on my initial search for any type of discussion like that
<Dandellion>
I tried adding them to the propgated buildInputs, but that didn't work
emilazy has joined #nixos
proofofkeags has joined #nixos
endformationage has joined #nixos
feepo has joined #nixos
chpatrick has joined #nixos
adamse has joined #nixos
<jtojnar>
Dandellion you need to add PATH using a wrapper, or patch the path in the sourcecode
sangoma has quit [Ping timeout: 240 seconds]
midchildan has joined #nixos
vdemeester has joined #nixos
raboof has joined #nixos
murmr has joined #nixos
alunduil has joined #nixos
pasukon has joined #nixos
coot has joined #nixos
steveeJ has joined #nixos
<{^_^}>
[nixpkgs] @domenkozar pushed to master « remove usage of patchelfUnstable now that it's released »: https://git.io/JJqMN
bitonic has joined #nixos
lally has joined #nixos
<jlv>
`exec sway` works now that I enabled `programs.sway` in NixOS. Guess I just won't use a display manager *shrug*.
<robodojo>
slabity i'd also like a pure wayland setup
TheNumb has joined #nixos
peel has joined #nixos
kalbasit has joined #nixos
<colemickens>
slabity: robodojo : there's no x11 required.
nlofaro has joined #nixos
d10n-work has joined #nixos
<colemickens>
per jlv 's comment, enabling 'programs.sway' in nixos is sufficient
<jlv>
robodojo: I second that. I think better Wayland support should be on the NixOS roadmap.
<colemickens>
What does that mean?
<colemickens>
What doesn't work?
<jlv>
colemickens: I think we're mostly talking about how thinks like `libinput` and `displayManager` require running an X11 server.
<jlv>
*things like
<colemickens>
Is it just a matter of renaming the attributes behind .xserver. for DMs? It is a bit confusing that you have to enable something with xserver in it to get GDM enabled.
<colemickens>
how does libinput require an x11 server? I'm using sway and libinput and there's definitely no xserver.
<colemickens>
(and I have no xserver related or named options enabled)
knupfer has quit [Ping timeout: 260 seconds]
<slabity>
colemickens: How do you enable libinput for example?
<slabity>
Or let's say lightdm?
<colemickens>
There is no such thing as "enable libinput" for Wayland. It's the only option.
<colemickens>
And yeah, the DM options should move out from behind the xserver option set, I agree with that.
<jlv>
colemickens: I'm not 100% certain, but it looks like `services.xserver.libinput` requires `services.xserver.enable`.
<colemickens>
jlv: that option has zero effect in any Wayland setup.
<slabity>
colemickens: Okay, so wayland pulls in the libinput dependencies... But how do you configure libinput?
<colemickens>
slabity: that is compositor dependent.
<slabity>
Using `services.xserver.libinput.*` options
<colemickens>
Someone would have to build an entire abstraction over every wayland compostiors configuration formats.
<jlv>
Or it should be moved out of `xserver` and have an effect, I mean.
<colemickens>
Someone could build it but it doesn't seem likely.
<colemickens>
Xserver and Wayland are pretty much fundamentally different in this regard.
kalbasit has quit [Ping timeout: 256 seconds]
<jlv>
colemickens: maybe if not a central option, we could have something like `programs.sway.libinput`, and use a template of `libinput` options. Kind of like how Home Manager uses the same options for `i3` and `sway` configuration.
kalbasit has joined #nixos
<slabity>
jlv: That would still require the backend to be written
<colemickens>
jlv: yeah, that's what I meant, you'd have to build an entire layer on top of that, that knows about how sway reads libinput config, along with any other compositor you awnt to support
<colemickens>
times every libinput option
<colemickens>
many years ago I had proposed that libinput (or wlroots, or someone) define a bit of json config that could be portable across Wayland DEs but... I don't think anyone liked that idea :(
<jlv>
colemickens: I figure we could generalize a lot of the configuration, so we minimize the code to support individual compositors.
<slabity>
colemickens: Heresay. How dare you suggest people standardize a complicated process
<colemickens>
jlv: maybe. I am pretty happy with the level of abstraction I get from HM now, personally. But I also just spent time making it so that alacritty/termite/kitty always have identical color schemes, so... I get it :)
<slabity>
But yea, if what you say is true, then it sounds like there are only a few options that would need to be removed from `services.xserver`
<colemickens>
(slabity: idk if it would even stand up over time, I could imagine it just being an annoying moving target over time)
<colemickens>
slabity: yeah, I think that's why no one ends up doing it
<colemickens>
slash many of us in sway just wind up doing 'exec sway'.
<colemickens>
(plus I don't like most of the DM options for one reason or another.)
kalbasit has quit [Ping timeout: 246 seconds]
<colemickens>
there's also the issue of naming, where would we move the DM options to?
<colemickens>
who wants to determine which ones only support X, do we move those too? etc? yuck
<slabity>
DM options... desktopManager or displayManager?
<jlv>
colemickens: I should probably be done eventually, but I get what you mean. The sway wiki even provides a few lines of code to automatically start sway from a tty login: https://github.com/swaywm/sway/wiki#login-managers
<jlv>
*it should probably
kalbasit has joined #nixos
<colemickens>
jlv: oh I think a sway.autologin for HM would be sweet
<slabity>
What about having each different desktopManager stand on its own and have the ones that require X11 or Wayland to automatically `xserver.enable = true;`?
<jlv>
slabity: I was just thinking about that. I think it's a good idea. Sway is even on it's own right now.
<jlv>
Could be something like `windowManager.*` and `desktopManager.*`.
shibboleth has joined #nixos
<slabity>
jlv: Well I'd assume windowManager would still be part of xserver since Wayland doesn't really have that concept
<jlv>
slabity: I though Sway was considered a window manager.
kalbasit has quit [Ping timeout: 240 seconds]
shibboleth has quit [Remote host closed the connection]
<slabity>
jlv: Not really... It's just a spiritually wayland-based successor to i3, which was a window manager
<bqv>
Man, I never thought id be arguing in favour of sway.
<bqv>
Its bloated sure, but its not like it comes with a settings dialog
<slabity>
It's... Bloated?
o1lo01ol1o has joined #nixos
<bqv>
Well what makes you think its a DE besides that?
<slabity>
Because it's a full environment on a desktop
<simpson>
I'm not sure why X11 terms should apply to Wayland. They're quite different things. The entire *concept* of a separate window-management process as distinct from the windowing server is because they communicate over the Network, which Wayland can't do.
jakobrs has joined #nixos
<bqv>
The thing is, I don't think WM vs DE is X11 specific, unless you literally shovel everything into one process
<bqv>
Cause I can still pick up a random wayland panel and use it with sway, right? So its build your own DE, like i3 and friends
<simpson>
(And we'll just ignore folks who don't use desktop environments, right~)
<slabity>
bqv: I'm not sure what you are saying in that case
<slabity>
i3 is an X11 based window manager, correct? And i3 + X11 is a desktop environment, correct?
<bqv>
simpson: well thats my point, in that scenario theyve built nothing
<bqv>
slabity: no?
<slabity>
bqv: Why not
<bqv>
Google DE :|
<qyliss>
what's the "desktop" part of i3 + X11?
<qyliss>
There is no desktop
dmj`_ is now known as dmj`
turion has joined #nixos
Extends has joined #nixos
Fare has joined #nixos
<slabity>
I'm not sure how that's not a desktop. If I throw X11 + i3 + Compton + A couple of panels is that a desktop?
alp has joined #nixos
<qyliss>
I've always interpreted a desktop to be a thing you could put files on
<qyliss>
Not sure I've ever heard the term refer to anything else in software
<robodojo>
personally i wanna use something like wayland because i don't need network gui redirect and i like the extra security
<clever>
this puts a haskell binary at /init inside the initrd
<robodojo>
i want more isolation of processes. the more the better
<jakobrs>
gchristensen: Sorry for pinging you, but
<jakobrs>
Have you received my last (fourth) email?
<bqv>
robodojo: sounds like a nice DE
m1cr0man has quit [Quit: G'luck]
o1lo01ol1o has quit [Remote host closed the connection]
<jakobrs>
I'm asking because I don't do email very often, and I wanted to make sure you actually got the emails.
<slabity>
bqv: If I put panels in my tty-console, is that a DE?
<bqv>
Now you're just being a troglodyte
<robodojo>
does tty mean teletype?
<eacameron>
I feel like such a noob but I have a symlink to a /nix/store file. I am not sudoer on this system but the symlink is owned by me. Yet when I do rm symlinkfile it gives me Permission denied. Even unlink symlinkfile gives me that. What??
<slabity>
Introducing kmscon + conky + tint2
<clever>
eacameron: `rm symlink` not `rm symlink/`
* bqv
will be back in half an hour, ideally much less annoyed
<slabity>
Sorry :P
<robodojo>
hope it wasn't me
<eacameron>
clever: Yeah I'm doing that !
<slabity>
I'm at least 85% certain it was me
<clever>
eacameron: what is the exact command you ran?
<clever>
eacameron: the symlink is inside a read-only directory, so you cant modify the list of filenames in that dir
<eacameron>
clever: Ohh... my
<clever>
eacameron: chmod o+w . to give yourself write
<clever>
or wait, was it u+w ?
<eacameron>
clever: had to do chmod +o but that worked! Thanks!
<eacameron>
I have never ever run into this before so I was so confused...
<eacameron>
+w I meant
fendor has quit [Read error: Connection reset by peer]
<eacameron>
I now realize why this happened. I `cp -r`ed a symlink to a nix path to somewhere else. So it copied the *contents* of the nix store path but kept the permissions read-only on the directory.
jakobrs has quit [Quit: WeeChat 2.7.1]
<clever>
ah
<eacameron>
TIL!
fendor has joined #nixos
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
m1cr0man has joined #nixos
<infinisil>
`cp --no-preserve=mode` is useful for scripts
<robodojo>
setting up an install, i got confused and rm'd /etc/nixos/hardware-configuration.nix and /mnt/etc/nixos/hardware-configuration.nix. after i generate it again what's the right way to get it into /mtn?
<Extends>
if you haven't installed nixos yet, you can regenerate the config onto /mnt with --root parameter and it'll work
<KarlJoad>
robodojo: If you're in a live environment, preparing for an install, you can just regenerate the config like you generated it the first time.
cript0nauta has joined #nixos
<robodojo>
KarlJoad ya but like i said i already did that, but that didn't put it into /mnt
<robodojo>
ill try --root
cript0nauta has quit [Remote host closed the connection]
<Extends>
nixos-generate-config --root /mnt should work
werner291 has joined #nixos
acarrico has quit [Ping timeout: 246 seconds]
<robodojo>
there we go thanks!
cript0nauta has joined #nixos
cript0nauta has quit [Remote host closed the connection]
<eyJhb>
Is bqv back yet?
<bqv>
Yeah
<bqv>
What ever happened to the better errors idea?
<bqv>
There isn't even anything self-referential at a glance, that's why I'm stumped
<robodojo>
why does the installer output 'setting up /etc...' twice?
<bqv>
Oh wait, maybe it's rec…
<KarlJoad>
Is there a way to configure things that are NOT the first part of a section in the xorg.conf file? All the settings (services.xserver.monitorSection and others) only refer to the "first" of that item.
lsix has quit [Quit: WeeChat 2.8]
<slabity>
KarlJoad: `services.xserver.config` will overload whatever gets generated
<simpson>
slabity: FWIW, "window manager", "display server", "compositing manager", all are very specific roles which X11 clients play; specific client processes have those powers.
<KarlJoad>
slabity: Overload everything? The entire file will reflect whatever I give it?
<slabity>
simpson: No worries. I'm (unfortunately) pretty familiar with X11 subsystems...
<KarlJoad>
slabity: That's what I was thinking too. But when I tried to generate it manually, I don't think I got a completely overridden file.
<slabity>
At least the ones that interact with DRM
<slabity>
simpson: I think my only complaint is that `services.xserver.desktopManager` is not necessarilly xserver specific
<slabity>
Some can be ran without xserver
Fare has quit [Ping timeout: 260 seconds]
<m1cr0man>
Anyone know a lib function to remove newlines from a string? (other than replaceStrings?)
<robodojo>
after a base system is installed can settings deep within programs be set in configuration.nix? like having plasma's night color already activated to the right kelvin
Fare has joined #nixos
<gchristensen>
jasom: Ihave, yes
<gchristensen>
oops
<jasom>
gchristensen: :)
<dminuoso>
robodojo: In principle, sure.
<dminuoso>
Often, its a bit out of the scope for nix because its more concerned about system wide, home-manager tries to close that gap a bit.
<dminuoso>
But its nothing fundamental about nixos that would prevent that.
<dminuoso>
:s/nix /nixos /
<KarlJoad>
slabity: I have just run a `sudo nixos-rebuild switch` after setting `services.xserver.config` manually and the generated output still keeps the defaults in the xserver.nix module.
<dminuoso>
robodojo: It largely depends on how well the nixos module is written. Sometimes you get options for settings inside the program, other times you get access to some `extraConfig` or `config` option allowing you take manual control over the programs config file.
<robodojo>
ohh
<robodojo>
ill start with just the system and applications then try settings later
<dminuoso>
robodojo: the key part to realize is use nixos modules rather than installing things by hand.
<robodojo>
ya im even using mutableUsers = false. i don't want to do any imperative system admin
<dminuoso>
So for instance, if you want dnsmasq installed, rather than adding it to your packages (or worse, installing it by nix-env), you'd set `services.dnsmasq.enable = true;`
<dminuoso>
Then you get to set things like `services.dnsmasq.servers = [ "8.8.8.8" "8.8.4.4" ];`
<robodojo>
got it
<robodojo>
good flexibility
qryptic[m] has joined #nixos
<dminuoso>
If the application has a granular configurable nixos module, yes.
<dminuoso>
Sadly, there's a large bunch of modules that dont have many options. :(
<robodojo>
what does networking.networkmanager.enable buy us? i read the doc on it but in practice as a plasma user i don't see any difference. settings still doesn't show details network status or transfer rate up/down
<robodojo>
would be neat if we could even set prefs, within games, within lutris, within nixos :)
<dminuoso>
NetworkManager is a networking daemon ontop of libudev to provide automatic networking features.
meh` has quit [Ping timeout: 256 seconds]
<dminuoso>
That is, the key of nm is not some graphical interface. The term "manager" rather refers to the fact that it "manages networking things for you"
<evanjs>
I guess I could just point all the other haskellPackages -> mine.xUserConfig.xsession.windowManager.xmonad.haskellPackages for now XD
<evanjs>
until I figure something better out
<evanjs>
lazy and using xmonad everywhere rn so I don't really care lol
<clever>
the nix file in the same dir will build it static, and build the initrd too
<robodojo>
vi doesn't have native module support so i'll add it to system packages :/
<bqv>
clever: did you ever flesh it out more?
<robodojo>
oh vi isn't even available for nixos?
<clever>
bqv: nope, it was mostly just a proof of concept, that you can make /init be pure haskell
<clever>
and as an example, if anybody wanted to take it further
<bqv>
i suppose it would be pretty insane to basically have the haskell runtime act as a second kernel
<clever>
bqv: there is also halvm....
__monty__ has quit [Quit: leaving]
<bqv>
clever: ok that answers that question
cole-h has joined #nixos
<{^_^}>
[nixpkgs] @utdemir opened pull request #92752 → nix-tree: Expose as a top-level attribute, add maintainer → https://git.io/JJqHy
justan0theruser has quit [Ping timeout: 240 seconds]
erasmas has quit [Quit: leaving]
<KarlJoad>
I can't seem to figure out why when I give something for the `services.xserver.config` option the text that is printed by default does not get overridden.
<clever>
KarlJoad: this shell script is used to generate the config file, lines 122-139 generate the `Section "Files"` part, and line 141 will print the config (from 118) which is your services.xserver.config
maxwell_lt has joined #nixos
<clever>
KarlJoad: but!, line 713-808 is also setting services.xserver.config, so nixos is merging the 2 strings
<clever>
KarlJoad: how are the defaults conflicting?
<colemickens>
cachix outages block nix builds by default still. the pain is real.
<clever>
colemickens: --option substituters "list goes here" to temporarily skip it
fendor has quit [Remote host closed the connection]
<KarlJoad>
clever: That first section is fine, because it's pulling in other files as libraries for X11.
<KarlJoad>
However, I have 2 NVIDIA GPUs, which I am attempting to configure with 1 monitor each. I have narrowed the problem down to the X server. I am using the NVIDIA drivers, so I used nvidia-settings to generate the xorg.conf file for me.
<clever>
KarlJoad: compare the generated xorg.conf to the one nixos is creating, how do they differ, what needs to be changed?
<KarlJoad>
But the portions of Lines 713-808 are always present with the string that I input, and I think it's confusing the server, because it crashed.
<KarlJoad>
I am unsure what portions need to change. I managed to have this working with defaults on the Ubuntu live image, so something is being specified that is breaking something in NixOS.
dansho has quit [Read error: Connection reset by peer]
dansho has joined #nixos
<colemickens>
Hm, I thought --fallback was supposed to do the right thing.
<robodojo>
when im doing an install and im doing the `parted` steps it says i might need to edit fstab. do i? or will nixos-install take care of that?
maxwell_lt has quit [Remote host closed the connection]
<KarlJoad>
clever: I have the nvidia-settings output on GitHub, with the rest of my config if that helps.
<clever>
robodojo: nixos-generate-config handles that step
<robodojo>
thanks!
<clever>
KarlJoad: you need to look at the string interpolation on 713-808, and set the right settings to make it generate the right config
<clever>
KarlJoad: for example, monitorSection for the first monitor, and xrandrHeads for the extra ones
<KarlJoad>
clever: So, you're recommending I translate the generated config to a Nix expression?
<clever>
yes
<KarlJoad>
Oof... Ok. I was hoping for a quick and dirty solution to test this hypothesis, but ok.
<clever>
KarlJoad: for a real dirty solution, `services.xserver.displayManager.xserverArgs = lib.mkForce [ ... ];`
<clever>
then you can just ignore nearly everything, but you must generate the arg list properly yourself (or copy/paste)
<KarlJoad>
That's a bit dirtier that I was really hoping for. I'll translate the generated config to Nix.
justan0theruser has joined #nixos
<{^_^}>
[hydra] @grahamc merged pull request #788 → updateDeclarativeJobset: only set the emailresponsible column when defined → https://git.io/JfpQW
<{^_^}>
[hydra] @grahamc pushed commit from @basvandijk to master « updateDeclarativeJobset: only set the emailresponsible column when defined (#788) »: https://git.io/JJqQL
alp has quit [Ping timeout: 272 seconds]
<robodojo>
the default configuration.nix shows installing vim using environment.systemPackages. why doesn't it use programs.vim.defaultEditor = true;?
shibboleth has quit [Quit: shibboleth]
proofofkeags has quit [Remote host closed the connection]
proofofkeags has joined #nixos
<{^_^}>
[nix] @matthewbauer opened pull request #3795 → Only store hash in OutputDerivationOutput → https://git.io/JJqQ2
<jneplokh>
Hey everyone, I am trying to setup Wireguard as a "server"/forward all traffic like a VPN on a NixOS machine using the instructions here: https://nixos.wiki/wiki/Wireguard but am having issues with reaching the internet.
<jneplokh>
When connected, I can ping `10.100.0.1` but cannot ping `1.1.1.1`, for instance. Is there a way I can diagnose if it is an issue on my "client" config, or the server config?
acarrico has joined #nixos
m0rphism has quit [Ping timeout: 265 seconds]
<Henson>
jneplokh: are you able to ping from the client to the server on the wireguard interface?
nikita` has quit [Quit: leaving]
<jneplokh>
Henson: Sorry, just a little confused. Do you mean from my "client" if I can ping the server? If so, I can ping the local IP for the wireguard interface
<jneplokh>
Which I define as `networking.wireguard.interfaces.wg0.ips = [ "10.100.0.1/24" ];`
<Henson>
jneplokh: can you describe what you're trying to achieve with wireguard?
MmeQuignon has quit [Ping timeout: 246 seconds]
<jneplokh>
Basically just use the server I already have running as a "VPN" server, routing my traffic through it
<jneplokh>
I can connect just fine, but the problem comes from then actually accessing the internet
<infinisil>
jneplokh: For your client config, use networking.wg-quick.* instead of networking.wireguard.*
<Henson>
jneplokh: so you've got a central server, connected to the Internet, running NixOS. You want to connect from a client, over the Internet using Wireguard as a VPN, and route traffic from your client, through the VPN, to the server and then out to the Internet?
<jneplokh>
infinisil: The client I am currently using is not running NixOS, so I am just configuring Wireguard through the app
<jneplokh>
Henson: Yes, that is what I am trying to achieve.
<Henson>
jneplokh: ok. You've got the networking.nat things set according to the wiki article? Are you also using similar IP addresses as the article, in the 10.100.0.1/24 subnet?
<{^_^}>
[nixpkgs] @grahamc pushed to master « libvirtd: don't start libvirtd-tcp.socket by default »: https://git.io/JJq7I
<Henson>
jneplokh: the thing you're trying to ping but can't, is it on the Internet, or on a network only visible to the server?
<jneplokh>
Henson: Yep, I have all of those networking.nat settings exactly as it is own that wiki article. Using the same IP addresses as well
<jneplokh>
Henson: I am trying to ping `1.1.1.1`
pamplemousse has joined #nixos
kvda has joined #nixos
<{^_^}>
[nixpkgs] @grahamc pushed to release-20.03 « libvirtd: don't start libvirtd-tcp.socket by default »: https://git.io/JJq73
<bqv>
oh my god
<bqv>
my bug was due to me using "prev" in a let
mbrgm_ has joined #nixos
<bqv>
i didn't realise prev (final, prev) was the same as "prev" (nRev, pRev)