<qq_pls>
If I set an attribute within a top level expression, how do I pass that attribute into another attribute set? I've tried `inherit` but it didn't seem to work.
missionformilk has quit [Ping timeout: 264 seconds]
ihsan has quit [Ping timeout: 240 seconds]
<dramforever>
What's the context? A NixOS module?
<dramforever>
If not, and it's just an attribute set, something like: rec { a = 3; b = ... use a here ...; }
<qq_pls>
Yes - I actually wasn't aware it being inside a nixos module mattered. My scenario is { module1.optionA = "value"; module2.optionB = module1.optionA; } I'm going to try to play around in repl if that doesn't make sense.
zebrag has quit [Quit: Konversation terminated!]
zebrag has joined #nixos
gustavderdrache has quit [Quit: Leaving.]
codyopel has joined #nixos
<{^_^}>
[rfc39-record] @grahamc pushed commit from rfc39 to main « Automated team sync results. »: https://git.io/JmrIQ
dalvescb has quit [Quit: Connection closed]
poscat1 has quit [Remote host closed the connection]
ihsan has joined #nixos
<cole-h>
qq_pls: If it's a module option, you'll need to access it with a `config.` prefix.
<cole-h>
e.g. `config.attr.path.to.module1.optionA`
poscat has joined #nixos
ihsan has quit [Client Quit]
SomeoneS1 has quit [Ping timeout: 276 seconds]
missionformilk has joined #nixos
<qq_pls>
I have seen that, that makes sense cole-h. I wonder if the inability to access it even with config. is related to this being a bit nested within another config = { config, ...}: expression.
sigmundv_ has quit [Read error: Connection reset by peer]
sigmundv_ has joined #nixos
<cole-h>
If you paste all relevant parts of your config, maybe I can help
h0m1 has quit [Ping timeout: 264 seconds]
<cole-h>
(note: NOT into here... paste it into a GitHub gist or some similar paste service)
<siraben>
doesn't freenode autokick you if you paste more than 5 lines
<dramforever>
qq_pls: The main thing about it mattering in a nixos module is that there's a better way in nixos modules :)
<dramforever>
Though a two-level nested module is something I have *not* seen before
<qq_pls>
Ok I'm probably doing that wrong! I'll try breaking out into separate modules then?
<cole-h>
qq_pls: Regardless, I don't think that's something that's going to work... The sops secret will be decrypted on the local machine. It won't be available in the container unless you bind mount it
<dramforever>
qq_pls: no no no that's fine
mcornick has joined #nixos
<dramforever>
It's just you're doing something more advanced than what I'm used to :)
<qq_pls>
cole-h ah I had figured that as well, so maybe I just go that way and bind to the path
* dramforever
hides
missionformilk has quit [Ping timeout: 246 seconds]
<cole-h>
qq_pls: Note that you cannot access the actual value of the secret with sops. You must read the file (from `config.sops.secrets.<name>.path`)
<dramforever>
The sops thing would still not work, right...?
<qq_pls>
I'm glad to see you suggest this, It makes me feel like I'm learning. I tried something similar but gave up on it too soon. Let me try your suggestion, thank you!
<cole-h>
dramforever: If it's properly configured, and not just stubbed, it'll roughly work
<dramforever>
Yeah, 'let' is what I was thinking too
<cole-h>
But qq_pls will need to have /run/secrets mounted into the container
<dramforever>
I have no experience with containers :(
<qq_pls>
@cole-h I assumed as much, if you have a better way of doing this feel free to let me know. Very new to it.
<cole-h>
A better way would be to not use containers ;)
<qq_pls>
cole-h that seemed to have worked, at least built so thank you. Because I want to understand this better, why is it we needed to let the value into a new variable? Is it something to do with the change in scope and if so why would rec and then calling it by it's original name not work?
<cole-h>
It's because of the `config` in `{ config, ... }:`
<cole-h>
If you removed the config there, you could get away with not needing to let-bind it
<qq_pls>
I copied it from an example of how to setup a container. Maybe it's extraneous or maybe it's not. I'll play around with it. thanks again.
<cole-h>
unless you're using it (which you're not) you don't need it
<{^_^}>
[rfc39-record] @grahamc pushed commit from rfc39 to main « Automated team sync results. »: https://git.io/JmrZX
acarrico has quit [Ping timeout: 260 seconds]
attila_lendvai_ has quit [Ping timeout: 260 seconds]
dvim has quit [Quit: Connection closed for inactivity]
<qq_pls>
cole-h mount point worked as expected, have access to the file from the container. Unfortunately echo $passwordFile comes back empty. Could this be the assignment to the empty attribute list? I just followed the sop-nix guide on that.
<qq_pls>
I do, that's how it gets read into the /run/secrets path. I can see the password from the container. I just think there's an error in the assignment of the environment variable.
worldeva has quit [Ping timeout: 264 seconds]
<cole-h>
well, the env var is only set for the service
<cole-h>
are you running `echo $PASSWORDFILE` as the service?
<qq_pls>
Oh good point thank you! I was echoing as the user I've set to run the service.
<matthewcroughan_>
the latest `nix`build seems to remove --update-input from `nix flake`
<matthewcroughan_>
Why? Seems silly.
<matthewcroughan_>
Now, `nix flake update` just updates all of your inputs.
everythingTaken has joined #nixos
<everythingTaken>
simpson: lovesegfault I don't know if you guys are up, but I downloaded nixos-gnome iso after being unable to solve the issue. Tested live gnome environment firefox in 3 different machines, my current old laptop, beefy gaming laptop, average desktop copmuter. All 3 have same problems with firefox as before. Live environment probably != full installation, but still I think in other distros you don't
<everythingTaken>
have that in even live environments.
<everythingTaken>
cole-h: oh hey cole, no will have to download another 1.5G on my slow internet to test it
<everythingTaken>
cole-h: maybe it's a kernel issue?
<everythingTaken>
cole-h: here is the thing though, one guys says gnome works, plasma doesn't, another guy plasma works gnome doesn't.
<cole-h>
¯\_(ツ)_/¯
<cole-h>
impurity is everywhere
<everythingTaken>
appropraite emoji lol
<everythingTaken>
man, this is so frustrating.
<everythingTaken>
cole-h: it rubs me the wrong way when I can't get basic things working without bugs, but this is NixOS I am willing to make the sacrifice :/
<everythingTaken>
other distros would get nuked so fast
bebarker has quit [Read error: Connection reset by peer]
<simpson>
everythingTaken: Do you have some sort of funky extensions, maybe!? I don't know what's the same across all of your test machines.
<everythingTaken>
simpson: oh hi, no no extensions, this was tested from fresh latest-gnome-nixos iso
redmp has quit [Quit: leaving]
<simpson>
If *the* only common factor is Firefox, then you should probably take this upstream.
ddellaco_ has joined #nixos
<everythingTaken>
simpson: none of them have AMD GPU? And Firefox 86 works fine on Xorg on every other distro?
Supersonic112 has joined #nixos
Supersonic has quit [Ping timeout: 265 seconds]
Supersonic112 is now known as Supersonic
ddellacosta has quit [Ping timeout: 256 seconds]
missionformilk has quit [Ping timeout: 264 seconds]
paddymahoney has quit [Ping timeout: 246 seconds]
missionformilk has joined #nixos
everythingTaken has quit [Quit: everythingTaken]
endformationage has quit [Quit: WeeChat 2.9]
<simpson>
Didn't we rule out GPU problems?
michael65 has joined #nixos
michael65 has quit [Client Quit]
michael88 has joined #nixos
<michael88>
Hello
<michael88>
I have a small problem and was wondering if somebody have faced it before
<michael88>
I'm trying out nixos nightingale plazma for the first time. and installed gcc and other c compilers. but when I try to configure any application it prints out the following:
<michael88>
checking whether build environment is sane... yes
<michael88>
checking for a thread-safe mkdir -p... /run/current-system/sw/bin/mkdir -p
<michael88>
checking for gawk... gawk
<michael88>
checking whether make sets $(MAKE)... no
<michael88>
checking whether make supports nested variables... no
<michael88>
checking for gcc... gcc
<michael88>
checking whether the C compiler works... no
zebrag has quit [Read error: Connection reset by peer]
<dramforever>
michael88: Please put the log in a paste bin
zebrag has joined #nixos
dsrt^ has quit []
tfl^ has joined #nixos
coderobe has quit [Ping timeout: 276 seconds]
dramforever has quit [Quit: Leaving]
simba1 has joined #nixos
coderobe has joined #nixos
addcn has quit [Ping timeout: 265 seconds]
ahmedelgabri has joined #nixos
Guest16 has joined #nixos
<Guest16>
simpson yes I think so, glxinfo | grep rendering shows direct rendering yes too
Guest16 has quit [Client Quit]
ahmedelgabri has quit [Ping timeout: 240 seconds]
stree has quit [Ping timeout: 246 seconds]
Guest96 has joined #nixos
<Guest96>
simpson tried firefox, firefox-bin, firefox-beta-bin, firefox-esr, same issue lol. Tried playing around with cpuFreqGoverner, nothing. Currently downloading plasma version, maaan. Sorry to bother you much
<Magic_RB>
hi guys! first time here, hate to be so answer seeking so soon, but I've got a Hydra question, I've managed to run my own Hydra instance and get builds working. I'm build OCI Images and I'm wondering whether it's possible to configure Hydra to somehow upload those images to an image registry, even if indirectly such as with a notification webhook
<Magic_RB>
so that some other piece of software could do the uploading part. Thanks for reading this and your answers :)
<Guest16>
Magic_RB I don't have relevant answer, but if you don't get answer here try Discord, seems it's more active
<Magic_RB>
I tried the Discord, seems people there don't really mess with Hydra :P I'm hoping to get in touch with the real Nix magicians here
<Guest16>
also could try forums
<Magic_RB>
I'll try the Discord again, stuff there gets buried rather quickly, then the forums, btw, by `forums` you mean Discourse?
<eyJhb>
Magic_RB: Might be relevant in #nixos-dev ?
Chiliparrot has joined #nixos
ahmedelgabri has quit [Ping timeout: 240 seconds]
Guest16 has quit [Ping timeout: 260 seconds]
<Magic_RB>
eyJhb: perhaps, though I think I've gotten somewhere, I think I'd have to write a module in perl
<Magic_RB>
I'll cross post my question over there, hoping more people will notice :)
<eyJhb>
^ Seems about right Magic_RB :)
<Magic_RB>
i've never written a line of perl in my life, if I really do have to write a perl function? module? then it'll be a real challenge
Boomerang has joined #nixos
<eyJhb>
Didn't you note there was webhooks for Hydra?
<eyJhb>
Or was that just "hooks" ?
<eyJhb>
Magic_RB *tagging*
Boomerang has quit [Client Quit]
konubinix has quit [Quit: Coyote finally caught me]
konubinix has joined #nixos
<Magic_RB>
eyJhb there are these https://github.com/NixOS/hydra/tree/master/src/lib/Hydra/Plugin but they're not generic webhook plugins, rater specific to one service, like Slack, so I could theoretically repurpose one of them, but perl is not a language I know
<Magic_RB>
It was used way before my time, Nix is barely younger than me, if we assume that Nix first appeared in around 2006
Guest16 has joined #nixos
<eyJhb>
Magic_RB: A youngling an IRC! - But yeah, I have not done much perl myself ... :p
<eyJhb>
But, learn by doing I guess
<Magic_RB>
yup, that's the way I learned Nix lol, just went in head first and managed to not drown somehow, came out a week later with a horrible NixOS config
ahmedelgabri has joined #nixos
<Guest16>
after upgrading to unstable I get libgl error loading nouveau?
fendor has joined #nixos
<Magic_RB>
Guest16 did you reboot?
<Guest16>
yes I did
<Guest16>
the reason why I upgraded to unstable is to see if the issue with firefox will be resolved
teto has joined #nixos
<Magic_RB>
can you remain here for a while? i've an online class rn, and I need to pay attention somewhat, I'll get back to you in 40 minutes if no one else steps up
<Guest14>
supersandro2000 I upgraded to unstable and tried to install something with home manager (added something to it), then now it's reinstalling/upgrading everything in home manager packages list
<Guest14>
Magic_RB you free?
lewo has quit [Quit: ERC (IRC client for Emacs 27.1)]
<aforemny>
Turns out idiotbox above works if I don't link statically (the Makefile does that and I missed it the first time). Still, I am not sure why it fails. -ltls pulled in libpthreads which pulled in the linker which does not resolve using -static. Whats the NixOS situation here regarding static linking?
<simonpe^^>
I'm looking to find some project that's been working on general wrapping of impure build systems that insists on downloading sources by themselves, is anyone sitting on some information that might be useful regarding this matter?
<simonpe^^>
And other impurities like maintaining their own build caches and so on
stree has joined #nixos
<Ke>
what would you think would be theoretically possible?
<Ke>
sppofing some common tools like wget, curl or git?
<ehmry>
simonpe^^: can you pre-download and link into the build tree? thats what I do
<simonpe^^>
ehmry: it's a bit too much, and bitbake relies heavily on its "sstate" and "download" caches, which are non relocatable
<simonpe^^>
or at least they used to be, IDK if they can be relocated these days
griff_ has quit [Quit: griff_]
<simonpe^^>
Ke: yes that would be the naive option but I think it's very hard to make something that considers every tool, every scm, and gets it right
puffnfresh has joined #nixos
<simonpe^^>
Another option would be to provide a "pure" proxy server that you can point your build system at
<simonpe^^>
But I didn't figure out how that would work, it might not be feasible
<ehmry>
build systems for embedded and baremetal stuff all seem to have this problem
<simonpe^^>
ehmry: yes, and built by neckbeards at <insert old megacorp chip manufacturer here>
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<ehmry>
unfortunately I think the people that build these thinks look at nix and think "my problem can't be that complicated"
<simonpe^^>
and then they build their own complicated mess lol
<yaymukund>
I'm getting both "JavaScript error: resource://gre/modules/XULStore.jsm, line 66: Error: Can't find profile directory." and "Unable to load hand2 from the cursor theme" on both firefox-wayland and firefox. I'm on swaywm, nixos unstable.
<yaymukund>
The 1st happens at startup and 2nd when I mouseover anything that would change the cursor to a hand pointer. (It does not change as expected.)
<Ke>
Guest14: does this freeze happen everywhere frequently, otherwise does not sound so bad to me
<Guest14>
Ke it's annoying, 3 places where happening, scrolling in preferences, opening new tab and trying to write in address bar 1-2 sec delay (1 out of 10 times). Creating new profile crashes with segfault
<yaymukund>
1s periodic freezes sounds miserable to me
<simonpe^^>
Guest14: change browser or change OS, that is the question :D
<Guest14>
simonpe^^ haha change os probably
<Guest14>
i got void in dual boot
shenbosi has quit [Quit: WeeChat 2.9]
leotaku has quit [Ping timeout: 260 seconds]
<simonpe^^>
Guest14: I mean, you don't use nixos without having to tinker and get very intimate with the nixpkgs github repo, if that doesn't sit well with you I think changing OS is the better option
sebbadk has joined #nixos
sebbadk is now known as shenbosi
<Guest14>
I like nixos, if firefox worked, I would switch over. But seems like these random bugs which can't be fixed will come up now and then.
leotaku has joined #nixos
<Guest14>
simonpe^^ ^
<yaymukund>
it sounds like it may be your WM, based on the end of that thread. (I am on sway and also do not experience this)
<Guest14>
yaymukund Xmonad + dwm + Plasma + Gnome all same
zupo has joined #nixos
<yaymukund>
ooo
<Guest14>
simonpe^^ I am ready tinker, actually been tinkering for 2-3 days :D
<Guest14>
I just lost hope now
<Guest14>
if it was any other distro it wouldnt be worth it
zupo has quit [Client Quit]
<Guest14>
simonpe^^ I tried everything, what to do now? Doesn't seems to be upstream issue. In other distros work fine.
<simonpe^^>
Guest14: when I've had problems and didn't feel like getting my hands dirty I've tried switching channel
<Guest14>
simonpe^^ I started with stable, changed to unstable same thing.
<Guest14>
one thing though unstable giving me errors opengl cant find nouveau
<Guest14>
newer mesa drive I think
<Guest14>
it says glibc 2.32 not found showing 2.31-74 path
<Guest14>
in unstable
<Guest14>
but that probably isn't the fix for firefox
<simonpe^^>
I don't use wayland so I can't really tell
<Guest14>
I use xorg primarily, tried wayland if it will fix the issue
<simonpe^^>
Aren't the new nvidia proprietary drivers fixed for wayland now?
<simonpe^^>
I read something about that on reddit but it might not be released yet
<Guest14>
simonpe^^ old hardware from 2010ish
puffnfresh has quit [Ping timeout: 256 seconds]
<Ke>
wayland compositor probably loads with mesa just like xorg
<simonpe^^>
ah
<Ke>
so the code should be the same
<Guest14>
yeah xorg gives the same error. Who are considered nixos wizards here? :D
<Guest14>
I have to get this fixed somehow. Issue on github is open for about 14 weeks
<Guest14>
I seem to be the minority having the issue with some other people
<Ke>
I would validate the drivers symlink and mesa for $your_program are the same
<Ke>
the thing is, I had your issue, like everyone else, but for every one the cause is different
<Guest14>
about:support shows graphics driver is working I think, checked yesterday
<Guest14>
Ke you mean exact same issue?
<Ke>
mesa driver not loading
<simonpe^^>
Guest14: if you look at the contributors in the PR you can maybe see if you can contact them personally either in this channel or in #nixos-dev
<Guest14>
Ke oh
<Ke>
the mesa plugin loading is a hack and requires your running system to be in sync with the symlink
<Ke>
this is guaranteed by having tmpfs rootfs and rebooting
<Guest14>
simonpe^^ thanks, but they are probably busy. Maintainer didn't have much time either
<Ke>
but even otherwise you can just see, which mesas you have installed and loaded
<Guest14>
mesa/nouveau was working ok in stable, even then firefox had the issue, so Iam not sure fixing mesa issue will fix firefox issue in unsatble thanks Ke
<Ke>
they should somehow be part of the same system closure as the program you are running
le0taku has joined #nixos
<Ke>
the symlink is in /run, so in theory reboot should fix it, if you are not using packages from different repo
<ToxicFrog>
Under what circumstances does something need to read /etc/nixos/configuration.nix? I was under the impression that it was only needed by `nixos-rebuild`, but I'm trying to run `nix-shell` here as a service user and it dies with "getting status of /etc/nixos/configuration.nix: permission denied"
<ToxicFrog>
It looks like it's dying in this bit of overlay code:
<ToxicFrog>
which is meant to make the overlays in /etc/nixos/overlays shared between configuration.nix and nix-shell, but I didn't think that needed runtime access to the configuration.nix source...
<Miyu-saki>
ToxicFrog: Yes, that reads `/etc/nixos/configuration.nix`.
s1341_ has quit [Quit: Connection closed for inactivity]
puffnfresh has quit [Ping timeout: 264 seconds]
<ToxicFrog>
Oh.
<ToxicFrog>
Well that's aggravating.
<ToxicFrog>
I assumed it was getting <nixpkgs> from NIX_PATH but I guess it's actually <nixos-config>? Or does that not use NIX_PATH at all?
<Miyu-saki>
ToxicFrog: One simple way to get around this is to have a separate file for your overlays, which you source both in your configuration.nix, and through your nix-shell.
<Miyu-saki>
dramforever: Yeah, but I think they specifically want to refer to /etc/nixos/configuration.nix's overlays.
<ToxicFrog>
Yes, that.
<ToxicFrog>
right now I have /etc/nixos/overlays/default.nix, which is imported from configuration.nix and contains a bunch of overlays as {...}: { nixpkgs.overlays = ... }
<ToxicFrog>
And then nixpkgs-overlays in NIX_PATH points to a file in the store that contains that fragment that loads the overlays from the system configuration.
<Miyu-saki>
ToxicFrog: Why not just import that directly?
puffnfresh has joined #nixos
<Miyu-saki>
Instead of using `overlays = (import <nixpkgs/nixos> { }).config.nixpkgs.overlays;`
<Miyu-saki>
Use `overlays = import /etc/nixos/overlays/default.nix;`
<dramforever>
Oh
<dramforever>
Yeah that's what I do too
<ToxicFrog>
Last time I tried that, it didn't work, and I was recommended here to use that import -- I'll give it a try
<ToxicFrog>
that said there are also a bunch of subsidiary files containing more overlays
<ToxicFrog>
like, /etc/nixos/overlays/default.nix has a bunch of overlays but also imports a bunch of other overlay files containing their own overlay definitions, which get merged together at nixos-rebuild time (AIUI)
<ToxicFrog>
I don't know if that interferes with things
<Miyu-saki>
ToxicFrog: Okay, funnily, I think dramforever's idea would actually work.
<supersandro2000>
Does nixos have a /etc/bash.bashrc symlink?
<supersandro2000>
also the default profile should contain both a .profile and .bashrc
<ToxicFrog>
Argh. Ok, so I can make it reference /etc/nixos/overlays instead of /etc/nixos/configuration.nix, but right now all of /etc/nixos is root-only because there's secrets in there, and I was hoping to do this without faffing about with making parts of it selectively available.
<ToxicFrog>
I'm able to copy /etc/nixos/overlays/nix-shell-compat.nix into the store, but the contents aren't evaluated except at nix-shell time, so I can't have it copy /etc/nixos/overlays into the store at nixos-rebuild time -- I have to either figure out some way of doing that outside nix-shell-compat, or completely rethink how I do this (like store the overlays outside /etc/nixos entirely, which may be the
<ToxicFrog>
easiest way to do this)
wallacer has joined #nixos
palo has quit [Ping timeout: 240 seconds]
saschagrunert has quit [Remote host closed the connection]
<pie_>
/run/current-system is symlinked to a build result in the nix store that has all the stuff and you should be able to find what you need in there in isolation, but otherwise also on your system
bb__ has joined #nixos
palo1 has joined #nixos
<pie_>
hm ok, so if you check man bash it says:
<pie_>
shell (see INVOCATION below).
<pie_>
Do not read either the system-wide startup file /etc/profile or any of the personal initialization files ~/.bash_profile, ~/.bash_login, or ~/.profile. By default, bash reads these files when it is invoked as a login
<pie_>
--noprofile
palo has quit [Ping timeout: 240 seconds]
palo1 is now known as palo
<pie_>
that does seem to be what then invokes /etc/bashrc and /etc/profile.local
bb_ has quit [Ping timeout: 264 seconds]
<pie_>
/etc/profile for me invokes /nix/store/*-set-environment which seems to be what actually sets things
<pie_>
so TIL where that comes from.
<pie_>
bb__:
dev_mohe has joined #nixos
<bb__>
hmm
stree has joined #nixos
aminechikhaoui has quit [Quit: Ping timeout (120 seconds)]
<bb__>
pie_, so source /etc/profile sould do it?
<pie_>
i guessssss
<pie_>
give it a shot
<pie_>
technically, from what little i looked at, /etc/bashrc will also work because it *looks like* it checks if /etc/profile has been run, and it if hasnt, it runs it
<bb__>
pie_, source /etc/profile && nix-shell '<home-manager>' -A install
<bb__>
error: file 'home-manager' was not found in the Nix search path (add it using $NIX_PATH or -I)
<bb__>
nope :(
<pie_>
check that it actually changed anything
<pie_>
and maybe the issue is somewhere else
<pie_>
did your rebuild even go thorough properly?
<pie_>
ok i took a quick peek up in scroll
<pie_>
you said somethin about adding a channel for home manager
<bb__>
yes
<pie_>
whats the value of the old and the new nix_path
<bb__>
so prelogout nixpkgs=/nix/var/nix/profiles/per-user/root/channels/nixos:nixos-config=/etc/nixos/configuration.nix:/nix/var/nix/profiles/per-user/root/channels
<pie_>
or more specifically, you should check if the nix_path in the set-environment script contains the new value
<pie_>
if it doesnt, something is wrong, probably in the configuration
<pie_>
also check su -l as well just in case
<pie_>
(check man su for what -l does)
<bb__>
post logout /home/bb/.nix-defexpr/channels:nixpkgs=/nix/var/nix/profiles/per-user/root/channels/nixos:nixos-config=/etc/nixos/configuration.nix:/nix/var/nix/profiles/per-user/root/channels
<bb__>
so it adds to PATH /home/bb/.nix-defexpr
<bb__>
hmm
<bb__>
wonder if I go PATH=blabla nix-env <home...>
<ToxicFrog>
Although this reminds me that I need to figure out what the deal is with the /tmp/env-vars errors when multiple users are using nix-shell at once
<ToxicFrog>
According to nix#262 this was fixed in 2014, but...
oxalica has quit [Remote host closed the connection]
acowley has joined #nixos
<acowley>
status.nixos.org has been stuck at "Loading data from Prometheus..." since yesterday for me. Is there anyone that should be notified?
<Reventlov>
Do nixos provide qcow2 images somewhere ?
<gchristensen>
mmm I forgot to fix it up, acowley
<acowley>
gchristensen: I've been hoping to see an unstable update that covers the git CVE from last week, so I was checking the page daily. I am a human cron job.
<gchristensen>
looking
vidbina has joined #nixos
<gchristensen>
using git on a case insensitiv efs?
stevenxl has joined #nixos
<Ke>
Reventlov: I guess you know qemu-img?
<acowley>
I do use it on macOS. My understanding is that it could be an issue there.
<acowley>
It is kind of worrying how long it takes for CVE fixes to get to an release. A couple days back, before the status page broke, it looked like there wasn't a patched version of git in either the stable or unstable channels.
<gchristensen>
we do pretty well generally
<acowley>
I wish I had something insightful to offer to help with the unstable breakage. Seeing so much red on the status page is unfortunate.
<gchristensen>
I think the motivation wasn't quite there to get this one through because of the case insensitivity of the problem. not saying that is great, though
<Ke>
are isos somehow different images?
<Ke>
ie. why not?
<acowley>
Yes indeed! That's why I was surprised this one was taking as long as it seems to have.
<Ke>
just making sure you know about a conversion tool
<gchristensen>
I'm trying to get the fix rolled out, but I have to compile nix first :P
<stteevveen>
Hi! I have a two questions about NixOps. If I have a server, with nixos installed, it has its own configuration.nix in its /etc/nixos. If NixOps deploys a config.nix there, will the entire /etc/nixos wiped out ? Do the config.nix that I plan to deploy need to include the generated hardware.nix like it was in the original nixos configuration of the
<stteevveen>
server (before deployemnt) ?
stree has joined #nixos
<stteevveen>
The second question is : Does NixOps manage the subscribed nix-channel a target machine is subscribed to ?
avaq_ has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<sterni>
stteevveen: the nix-channel is not managed by nixops, but you can use nix.nixPath to influence it
<stteevveen>
@qq_pls I think the best is to manage one's configuration.nix file in git. That's what I did from the start, and many would advise to do it.
zakame has joined #nixos
<{^_^}>
[nixpkgs] @sternenseemann opened pull request #116931 → haskellPackages.hedgehog-classes: unbreak by adjusting semirings bound → https://github.com/NixOS/nixpkgs/pull/116931
<gchristensen>
acowley: btw that is fixed
<sterni>
stteevveen: also I'm pretty sure the configuration.nix on the remote machine doesn't matter at all, but you'd need to try it, I don't use nixops personally
<ajs124>
noonien: rarely, but I have it installed. Why?
<noonien>
i want to use nix on my android tablet, and this seems like a good start, i'm wondering if there's some tips on using it
<noonien>
i'll try to get flakes working on it
afreak_ has quit [Remote host closed the connection]
<noonien>
i'm currently using `nix-shell -p nixFlakes` when i need flakes
<gchristensen>
I might accidentally spam this channel here in a second, hopefully not but I'll be watching for it if I do
<noonien>
and now, i'm trying to build something, it seems to take a LOT of time after downloading stuff
<ajs124>
I basically only used it so I have a bunch of network debugging tools on my phone, so I don't think I'll be able to give much advice.
zebrag has quit [Quit: Konversation terminated!]
zebrag has joined #nixos
DevilFather[m] has quit [Quit: Idle for 30+ days]
dominikh[m] has quit [Quit: Idle for 30+ days]
<qq_pls>
I have a service defined in a container, it's configured to use a particular User. I'm getting 217 errors on starting the service inside the container and I'm suspicious that this is some kind of dependency thing. Is it possible to ensure the user account is created before the service attempts to start?
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
dbmikus has joined #nixos
Qwerky has quit [Remote host closed the connection]
<Guest78>
lovesegfault simpson any of you guys up? I finally found something, firefox freeze issue resolved in XFCE!!! Now gotta figure out what exactly and port to other de/wms
<gchristensen>
I was worried that would happen ... I'll take a look
<{^_^}>
[nix] @edolstra pushed to master « Restore 'nix flake info' as a deprecated alias »: https://git.io/JmPAn
<supersandro2000>
only 3 extra messages.
<radvendii>
I am trying to set up distributed builds, and to see if it's working i tested a build with --max-jobs 0 and it worked (so far so good). but then I disabled the distributed build and ran again with --max-jobs 0 and it still worked.
<supersandro2000>
not like they bridged in a cycle got several thousand characters long and where spammed every other second
<radvendii>
am i understanding --max-jobs 0 correctly? without distributed building the command should fail, right?
<exarkun>
seems like it is the kind of url that really oughta not change its contents
<gchristensen>
I wonder if you can get the old file and diffoscope them
<Guest96>
cole-h hey cole, so I finally found something about firefox freezing. I am amazed. running xfce-session fixes it, I don't know why, need to find out. Any ideas?
<exarkun>
What if GitHub is fine and I'm the one who screwed up. What kind of thing could I have done to screw it up like this?
<Guest14>
this is increasingly getting more interesting
<exarkun>
Okay how about this for a hypothesis
mayhew` has quit [Ping timeout: 240 seconds]
<exarkun>
I fetched version X of the fixed-output derivation w/ the correct hash. Then I changed the expression to fetch version Y and forgot to update the hash. Then a long time passed with everything apparently working. Then I ran the garbage collector. Then I tried to re-build this expression.
<{^_^}>
[nixpkgs] @FRidh pushed 33 commits to staging-next: https://git.io/JmXZf
vidbina has quit [Ping timeout: 240 seconds]
<Guest14>
simpson if all fails I think I can run xfce-session command without any junk (panel etc) + wm
<exarkun>
Only problem is git history does not support it... last change is for both rev and hash. But I guess there could have been an intermediate, uncommitted version that fit the hypothesis and that I can't see.
hiro98 has quit [Ping timeout: 246 seconds]
qq_pls has quit [Quit: Connection closed]
<Guest14>
how often does unstable break?
<exarkun>
I almost want fixed-output derivation to be confirmed on every single build ... and also a 100x better cpu and network connection
<cole-h>
It doesn't break very often, but it has the potential to do so :)
aasg has joined #nixos
aleph- has quit [Ping timeout: 264 seconds]
<exarkun>
cole-h: I feel like how often it breaks is partially a function of the user ... and I seem to screw up this way more often than is comfortable
phalrax has joined #nixos
<phalrax>
Hi, is it possible to set ACLs in configuration.nix? I would like to give my personal user write access to it, so I can omit sudo/easier edit the file in vscode or whatever or check it into a git repo. Any suggestions?
Guest14 has joined #nixos
aleph- has joined #nixos
<{^_^}>
[nixpkgs] @SuperSandro2000 merged pull request #115121 → ardour: add a videoSupport option, harvid: init at 0.8.3, xjadeo: init at 0.8.10 → https://github.com/NixOS/nixpkgs/pull/115121
<supersandro2000>
Guest14: at least something in unstable breaks every day
<Ke>
you can run nixos-rebuild with NIXOS_CONFIG set to point directly to git repo
stteevveen has quit [Quit: Connection closed]
<Ke>
or set nixPath
<supersandro2000>
but with the amount of packages that is expected
<supersandro2000>
and usually it is a sign that you should add yourself as a maintainer to that package because it is not that widely used
<supersandro2000>
or someone somewhere overlooked something
<Guest14>
Ke simpson I uninstalled xfce and installed standalone xfce-session package. it works with it. something in xfce-session. checked processes, "xfconfd" xfconf-4.14.3 is running.
aasg has quit [Quit: Bridge terminating on SIGTERM]
<simpson>
Guest14: Curious. I have literally no idea what it might be doing which makes things work for you. Glad that you've got a solution, at least,
<flokli>
etu: back to the videoDrivers thing, it seems to confusingly not be enough to just set the option
<flokli>
maybe the order matters too (?)
lordcirth has joined #nixos
simba1 has quit [Ping timeout: 240 seconds]
simba1 has joined #nixos
<flokli>
ah, so radeon explicitly isn't allowed to be in the list I'm setting. mhhm
growpotkin has joined #nixos
pandas has joined #nixos
astrofog has joined #nixos
phalrax has quit [Quit: Connection closed]
missionformilk has quit [Ping timeout: 276 seconds]
<pandas>
hello! I'm trying to make a vim setup with home-manager but i can't get vim-which-key to work. The `register` function can't be found, despite my best efforts.
<pandas>
I installed the plugin by adding it to `programs.neovim.plugins` as it is in `nixpkgs`. Anyone got an idea?
<matthewcroughan_>
I want to override the package, but get error: attribute 'freecad-flatmesh' missing, at /home/julian/tunnelvr/shell.nix:8:45
<sterni>
matthewcroughan_: you define a pkgs set with just a freecad-flatmesh attribute which never gets evaluated because it is shadowed by the pkgs binding of the function below
<andoriyu>
I'm very confused by nixos xorg/wayland configuration. From what I understand "default gnome" is wayland all the way. How do i change main monitor before login on that one? Right now it chooses my horizontal monitor to display login form. it's bad. Second, how do I switch to xorg instead? I want to run wm that doesn't work in wayland.
<leex>
Another question I have is, how do I get pulse audio to start up on boot? It works if I run start-pulseaudio-x11 after booting, I tried adding "autospawn=yes" to .config/pulse/client.conf
<Guest43>
have dbus running leex?
<Guest43>
I think dbus starts it when needed
<leex>
dbus.service loaded active running D-Bus User Message Bus dbus.socket loaded active running D-Bus User Message Bus Socket
<leex>
looks like it :)
<supersandro2000>
Guest43: without encryption support in hardware definitely
ppppp has joined #nixos
alexherbo2 has joined #nixos
<gchristensen>
I have a problem with Firefox that I don't have in Chrome: I can't click and drag elements on websites. anyone seen this before? nobody else seems to have this issue with firefox. my only extensions are ublock origin and passff
<leex>
trace: warning: The option 'services.dbus.socketActivated' in `/etc/nixos/configuration.nix' no longer has any effect and can be safely removed: the user D-Bus session is now always socket activated.
<leex>
So yeah that won't change anything
jrm2 is now known as jrm
<leex>
Guest43: in the journal logs it says that it cannot find GLIBC_2.32
vidbina has joined #nixos
<Guest43>
leex hmm I don't know man I am new to nixos too, you can try removing other options of pulseaudio except enable in conf nix
lsix1 has quit [Ping timeout: 246 seconds]
<Guest43>
You have some other options enabled
<leex>
yeah this is a straight copy of what worked on my x230 (and arguably an older version of nixos) for my x1 nano, where I went with unstable
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « haskell-barbly: fix incorrect override that adds apple-sdk framework »: https://git.io/JmXhx
<Guest43>
leex tried rebooting?
<leex>
Guest43: multiple times :)
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « haskell-aws: add jailbreak to fix the build with base16-bytestring version 1.x »: https://git.io/JmXj3
<leex>
Guest43: this was low on my prio list, I first had to fix the trackpoint + wifi :) because it always worked when I started it by hand, so it's an annoyance but not that bad.
<pie_>
when you start pulse manually does it not complain in the logs about the same stuff thats in the journal logs?
<pie_>
(i think you can make it stay in the foreground though
<leex>
On the terminal I only get this error: pulseaudio
attila_lendvai_ has quit [Ping timeout: 240 seconds]
<leex>
E: [pulseaudio] backend-ofono.c: Failed to register as a handsfree audio agent with ofono: org.freedesktop.DBus.Error.ServiceUnknown: The name org.ofono was not provided by any .service files
<pie_>
you van pass a few -v i think, or try to dig out of the systemd service the commandline for it and see what happens when you run that
<pie_>
im not sure about the details either, just maybe trying to isolate the issue further
<{^_^}>
[nixpkgs] @peti opened pull request #116965 → Update Haskell package set to Stackage Nightly 2021-03-19 (plus other fixes) → https://github.com/NixOS/nixpkgs/pull/116965
<pie_>
actually, do that and also maybe paste the unit
superherointj has joined #nixos
<pie_>
im not sure how to find it, theres probably some systemd command for that
hiro98 has joined #nixos
fabian_a has joined #nixos
<leex>
pie_: :D I was just about to ask where to find it
<pie_>
leex: so this is a completely new system with just the config transferred?
<leex>
yes :)
<superherointj>
There must have some solution to the crazy amount of PRs for nixpkgs. What if there was a fund we could donate to hire people to review PRs? Is there such thing?
erasmas has quit [Quit: leaving]
<leex>
pie_: well I also went for unstable this time because I hoped it would fix the Fn keys and the wifi (it fixed the wifi)
<pie_>
leex: im asking to make sure there isnt some sort of mixing of channels going on
<pie_>
which might actually explain it if anything in your user environment depends on your user channel. maybe say,PATH hash stuff from the wrong channel , that could lead to the user servie resolving the wrong libc or something
<leex>
okay I am done rebuilding my user stuff with home-manager
<pie_>
aha
<pie_>
i should have asked about that...
<leex>
I will give it a reboot and report back in a second
<pie_>
i dont know anything specific relating to home manager but i could see that setting up something with the user services
<leex>
pie_: is there a problem with home-manager and pulse nowadays?
<PIE_>
leex: is home manager on nixpkgs-unstable while your system is on nixos-unstable or something?
<PIE_>
that could explain the version difference but i still dont really know whats going on there
<leex>
PIE_: yes, nixos-unstable for the system and nixpkgs-unstable for home-manager
<leex>
But that still doesn't explain why that one bt line breaks pulseaudio
philr_ has joined #nixos
<leex>
If I check the PA path as root I also get 14.2, I also don't have a PA version mentioned in my home.nix
<leex>
root should not be covered by home-manager
<PIE_>
ok so home manager _isnt_ the one giving the pulse
<PIE_>
yeah i didnt expect it ot be but i dunno
<PIE_>
err, i kinda misread
<PIE_>
letme try again
<PIE_>
ah wait no nevermind ok
<PIE_>
rootis also 14.2
<leex>
exactly
<PIE_>
well so theres tw oquestions still, why is 14 in PATH but 13 in the service, and why does bt break it
<PIE_>
*two questions
<leex>
100%
<leex>
I have to run soon, but I am gonna put that exact question in the issue, maybe someone more knowledgable than us stumbles across it and can us tell why
<PIE_>
what i find the most confusing is how there are two versions to begin with not so much why the different things have different versions
<supersandro2000>
leex: /nix/store/0c7c96gikmzv87i7lv3vq5s1cmfjd6zf-glibc-2.31-74/lib/libc.so.6: version `GLIBC_2.32' version mismatch
LilleCarl has joined #nixos
Brandon__ has joined #nixos
jhuizy9 has joined #nixos
shla0 has joined #nixos
dupon16 has joined #nixos
concept20 has joined #nixos
bpye9 has joined #nixos
jhuizy has quit [Read error: Connection reset by peer]
shla has quit [Read error: Connection reset by peer]
dongcarl has quit [Write error: Connection reset by peer]
concept2 has quit [Read error: Connection reset by peer]
abathur has quit [Read error: Connection reset by peer]
acowley_ has joined #nixos
pushqrdx has quit [Remote host closed the connection]
abathur has joined #nixos
Ankhers has quit [Ping timeout: 264 seconds]
shla0 is now known as shla
jhuizy9 is now known as jhuizy
concept20 is now known as concept2
tpw_rules has quit [Excess Flood]
rprosper- has joined #nixos
pushqrdx has joined #nixos
tpw_rules has joined #nixos
strobelight has quit [Remote host closed the connection]
bpye has quit [Read error: Connection reset by peer]
acowley has quit [Ping timeout: 246 seconds]
dupon1 has quit [Read error: Connection reset by peer]
rprospero has quit [Ping timeout: 246 seconds]
dupon16 is now known as dupon1
bpye9 is now known as bpye
<supersandro2000>
The last days I got into how to write bash-completion scripts. Anyone would be interested in more package completion for various nix-commands like nix edit oder nixpkgs-review -p? just wanted to know if there is a general interest or I would do it for myself.
<matthewcroughan_>
Does anyone know how you get hydra running on ipv6
dev_mohe has quit [Quit: dev_mohe]
apache8080 has joined #nixos
simba1 has quit [Ping timeout: 240 seconds]
<matthewcroughan_>
maybe gchristensen knows? :D
<gchristensen>
it should just work? :)
<matthewcroughan_>
gchristensen: lol, I had the firewall on, sigh.
<gchristensen>
phew
<matthewcroughan_>
Oh, no, that's not the case at all!
<gchristensen>
I'm glad that was it
LilleCarl has quit [Ping timeout: 256 seconds]
<gchristensen>
uh oh
<matthewcroughan_>
it wasn't it, false positive, it's not listening on ipv6
<matthewcroughan_>
I've tried setting the listenHost to '::' to no avail
LilleCarl has joined #nixos
<gchristensen>
are you exposing the perl server directly to the web? usually there is nginx or apache in front
<matthewcroughan_>
yeah, directly to the web
<gchristensen>
huh...
<gchristensen>
welp
<gchristensen>
"Catalyst does not have any problems with deployment of IPv6 aware applications, as long as it is done through an IPv6 aware HTTP engine (like Apache). However, the myapp_server script currently only listens on an IPv4 socket, and ignores any IPv6 addresses that your server may have." -- http://wiki.catalystframework.org/wiki/faq
<colemickens>
made the mistake of deleting /var/lib/postgres and /var/lib/hydra and can't figure out how to get hydra to start over initializing the db
<colemickens>
oh I see it, I rm'ed the contents of /var/lib/hydra, missing the dotfile it uses to sniff db init'd
dev_mohe has quit [Client Quit]
vidbina has quit [Ping timeout: 260 seconds]
<figgyc>
is there a function that converts \n to \r\n (crlf line endings)?
xO1_ has quit [Ping timeout: 240 seconds]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<figgyc>
actually that might not work if there's already \r\n in the file although i dont think that should be the case for what i'm doing hopefully, there might be a better way though
<supersandro2000>
abathur: there are plenty
<supersandro2000>
eg wal-g
luxemboye has quit [Remote host closed the connection]
<supersandro2000>
if the package contains C code you might need to add runVend and libraries go to buildInputs
<matthewcroughan_>
attribute 'tls.certresolver' already defined at /nix/store/96gwqzrvxb7qjig70g9f7siylc1wzxgz-source/hosts/fu-cluster/modules/hydra.nix:30:9
<abathur>
supersandro2000: my rough ~project is figuring out if/how I can identify commands that may execute other commands; false positives are fine, but false negatives aren't, so it also entails identifying when to back off or use alternate heuristics
<supersandro2000>
matthewcroughan_: that does not look like nginx...
fl0_id has joined #nixos
<supersandro2000>
abathur: not sure how you want to do that
<supersandro2000>
you can do lots of fancy tricks to hide executables in binaries
<abathur>
masochism + magic
<abathur>
I'm not looking for malware
<matthewcroughan_>
supersandro2000: I figured it out, traefik module sets tls = true if you use any member attribute of tls, it'll enable it
<supersandro2000>
well, some people could declare chrome as malware
dev_mohe has joined #nixos
<supersandro2000>
but any proprietary binary will be hard
dev_mohe has quit [Client Quit]
bulukucing has joined #nixos
missionformilk has quit [Ping timeout: 256 seconds]
<abathur>
my goal is to carve out slices (ideally just the big ones) of the broad category of executables that I feel like I've found a good heuristic for *actually* separating out sub-executing binaries for, and treating everything outside of the slices as if it does by default
codygman__ has joined #nixos
<viric>
clever: this explains all about the dynamic symbols resolution & dlopen. RTLD_DEEPBIND is specially interesting for my case. https://akkadia.org/drepper/dsohowto.pdf
<abathur>
so proprietary binaries aren't of particular interest (beyond the extent to which some of them may still walk and talk like other binaries...)
bulukucing has quit [Quit: Connection closed]
<abathur>
afk a bit
gustavderdrache has joined #nixos
civodul has quit [Ping timeout: 240 seconds]
<supersandro2000>
abathur: so you are starting with the easy ones that use standard methods and go from there. Not a bad idea to tackle any problem but it might require some refactoring down the line which is alright.