<{^_^}>
[nixpkgs] @mweinelt closed pull request #89136 → mini-graph-card: init at v0.9.3; mini-media-player: init at v1.8.1 → https://git.io/JfoSy
Thra11 has quit [Quit: WeeChat 2.8]
proofofkeags has quit [Ping timeout: 258 seconds]
slack1256 has quit [Remote host closed the connection]
dermetfan has quit [Ping timeout: 260 seconds]
dermetfan has joined #nixos
Dagger2 has quit [Excess Flood]
Dagger2 has joined #nixos
Dagger2 is now known as Dagger
<manveru>
energizer: with flakes my eval time is 0.177s :)
<energizer>
manveru: holy crap how?
<manveru>
magic :D
<simpson>
Because unchanged flakes don't need re-evaluation, hopefully.
<energizer>
manveru: what's your e2e nixos-rebuild switch time for a minor leaf change?
<manveru>
let's see
orivej has quit [Ping timeout: 256 seconds]
noudle has quit []
<manveru>
about 15 seconds
dermetfan has quit [Ping timeout: 246 seconds]
<energizer>
i see
Dagger has quit [Excess Flood]
<manveru>
it's a pretty huge config... atm i'm trying to get my raspberry running from a flake which should be a lot faster still
Dagger2 has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JfKax
jlv[m] has quit [Quit: authenticating]
jlv[m] has joined #nixos
<abathur>
anyone around using single-user macOS?
orivej has joined #nixos
Dagger2 has quit [Excess Flood]
Dagger2 has joined #nixos
pamplemousse has quit [Ping timeout: 272 seconds]
orivej has quit [Ping timeout: 272 seconds]
cr4y1 has quit [Ping timeout: 264 seconds]
mudtose has joined #nixos
<mudtose>
Hello again. I am trying to dual boot NixOS with Windows 10. I had a Fedora and boot (ext4) partition but I have deleted them. Neither the MBR nor the GPT instructions seem to work for me. I think I have UEFI with MBR, maybe that's the reason?
<mudtose>
I get an error each time I run nixos-install, mostly that I need GPT but I also got an error saying I don't have a boot partition.
<simpson>
mudtose: Are you mounting your new / and /boot before running nixos-install?
LevelChart8 has joined #nixos
<mudtose>
I've tried mounting one as well as mounting both, and neither seems to work
thc202 has quit [Ping timeout: 260 seconds]
<simpson>
Is your /boot formatted as UEFI-compatible? Like, vfat? (Not to make your day worse, but how did Windows boot? Are they going to share the UEFI boot filesystem?)
<simpson>
At this point, you may need to share *how* things aren't working.
<iqubic>
Will I have to reboot after enabling JACK audio in my configuration.nix and rebuilding?
<mudtose>
Yes, sorry. Give me a moment simpson
<iqubic>
I just enabled JACK, and rebuilt my configuration. The audio that was playing is still playing. I assume I'll need to reboot.
<mudtose>
My Windows partition booted on the same grub installation as Fedora, on an ext4 /boot partition. I deleted this so I could reinstall GRUB with Nix, which I have been unsuccessful so far
<simpson>
iqubic: I would assume a reboot is necessary. At a bare minimum, you will need to make PA and JACK aware of each other in the right order, or if you're disabling PA entirely; both PA and JACK have userspace daemon components which compete for the same audio resources.
<abathur>
iqubic: don't know that you have to, but I like to do a rebuild boot whenever I mess with hardware just to make sure I'm not either falsely assuming it works or the inverse
<iqubic>
I'm planning on using PA andn JACK together.
<mudtose>
According to gdisk, I have MBR and according to /sys/firmware/efi/efivars I have UEFI. There are no instructions that target this combination, it seems?
<simpson>
Going from ALSA to JACK, maybe more viable, not sure. Traditionally, anything PAM-related like login limits for low latency will require relogin, including tearing down X sessions.
<iqubic>
So, I should reboot to make sure this works properly?
<simpson>
mudtose: This combination is viable; your MBR record should have a vfat-ish boot partition, and both GRUB and Windows 10 will put their bootloaders in there. GRUB may be weirded out by the situation, but it ought to work.
li_matrix has joined #nixos
<li_matrix>
<energizer "i wish nix could rebuild so fast"> Been working on that, could be testable in weeks
<abathur>
have you already built it?
<iqubic>
abathur: Yeah. I just ran a rebuild just now.
<abathur>
I usually do a nixos-rebuild test, and then a nixos-rebuild boot in situations like that
<simpson>
iqubic: At a minimum you may need $(pulseaudio -k) to reboot the PA user daemon. After that, you can run your JACK patchbay of choice, e.g. qjackctl, and see what's working.
<iqubic>
I think I'm just going to reboot. If worst comes to worst I can easily roll back my system to when I just have PA.
iqubic has quit [Remote host closed the connection]
<mudtose>
I don't have a /boot partition currently, and when I did, I am pretty sure it was ext4
<{^_^}>
#87895 (by evils, 2 weeks ago, open): firefox-wayland cannot be started if it is already running
<kaychaks_riot>
coz am still facing it for every firefox version I try - beta,dev
<simpson>
mudtose: Interesting. Perhaps you chain-booted into Windows before? Anyway, I'll try to focus on NixOS first; you should just avoid changing the Windows partition if possible.
<simpson>
mudtose: If your beginning and end of disk have space... a Fun Fact about disk layouts is that GPT can co-exist with MBR, sort of, sometimes. So, if you're feeling adventurous (and I can't promise that this will work), you could read up on putting a GPT without removing the MBR.
orivej has joined #nixos
<simpson>
And then you can follow the standard UEFI setup recommendations without any tools erroring out on GPT vs MBR.
<mudtose>
I can't risk bricking my Windows partition, unfortunately
teto has quit [Ping timeout: 260 seconds]
<simpson>
No worries, totally understandable.
klntsky has quit [Ping timeout: 240 seconds]
orivej has quit [Ping timeout: 272 seconds]
klntsky has joined #nixos
<energizer>
li_matrix: sweet. is there a repo i should bookmark?
m0rphism has quit [Quit: WeeChat 2.7.1]
iqubic has joined #nixos
Dagger2 has quit [Excess Flood]
<iqubic>
so, I have qjackctl and pavucontrol running. I can't tell if JACK and PA are able to interface with eachother.
<mudtose>
simpson, when I mount boot, it complains that I don't have GPT. when I don't make a boot partition, it complains it can't find it
<iqubic>
So, how can I tell if JACK and PA are seeing each other?
Emantor has joined #nixos
<iqubic>
All I'm going to say is that something seems weird, becauase most of the tabs in pavucontrol are blank.
<emily>
mudtose: the archwiki has some docs on uefi mbr setups
<emily>
(but also, my condolences)
<simpson>
Try playing music?
<iqubic>
simpson: I tried playing music, and pavucontrol shows me that firefox is outputing sound. However, I can't play that sound from my laptop speakers, because that option is not available.
<iqubic>
Like the configuration tab of pavucontrol shows nothing, not even my built in sound card.
<simpson>
Check out qjackctl? Hopefully you'd have Firefox going to JACK, if that's a thing.
<iqubic>
I don't know what I should be looking for in qjackctl
drakonis has joined #nixos
<iqubic>
This is really strange
fusion809 has joined #nixos
<iqubic>
Oh. I see what I need to do here. Internet searching helps a lot.
h0m1 has quit [Ping timeout: 272 seconds]
slack1256 has joined #nixos
<DigitalKiwi>
quinn: i finally got a working and useable gui on my raspberry pi and my mlv-app works on it! :D
<DigitalKiwi>
well, works in the sense it launches and doesn't segfault, i haven't actually used it for post prod on anything lol
Maxdamantus has quit [Remote host closed the connection]
sigmundv has quit [Ping timeout: 256 seconds]
<DigitalKiwi>
if anyone would like to review that...grazie
nuncanada has quit [Quit: Leaving]
<energizer>
me and one other nixos user only get black video from webcam in Zoom. can anybody else report whether your webcam works in zoom and what kind of webcam it is?
raghavsood has quit [Ping timeout: 245 seconds]
alp has quit [Ping timeout: 265 seconds]
Maxdamantus has joined #nixos
<iqubic>
webcam works for me in Zoom. I'm just using whatever webcam is built into my laptop. Laptop is a Dell Latitude 5480.
<DigitalKiwi>
does your webcam only not work in zoom or does it not work in other aps
m0rphism has joined #nixos
<pie_>
iqubic: yeah uh jack and pulse are kind of screwy, if you figure anything not mentioned in the thread, do leave me a message or a post on https://github.com/NixOS/nixpkgs/issues/74742
<{^_^}>
#74742 (by bcaccia, 25 weeks ago, open): Jack Audio does not present any valid output devices
<iqubic>
What thread?
<pie_>
the one i linked
<pie_>
*tell me on irc or leave a post on the issue
moet has joined #nixos
<pie_>
ugh i really need to write a wiki document for it. theres a bunch of stuf fi dont understand either, and getting debug output is a bit of a pain
* pie_
goes back to other stuff
<moet>
hello! the default nixos console font is `consoleFont = "Lat2-Terminus16";` right now.. how can i select a size 14 version? i tried changing 16 to 14, but it seems to not exist
<iqubic>
I wish ldlework were aroundn.
proofofkeags has joined #nixos
<iqubic>
ldlework: You around?
<hpfr[m]>
Hey! If anyone’s moved to flakes I’d be interested in checking out their system repo
<moet>
another question: when you look at consolefont filenames, there's usually a four-charactercode at the end, like x14b or l16n or something.. what does the first character in this code mean? the middle two are the size, and the last one is (b)old or (n)ormal..
<energizer>
DigitalKiwi: only zoom. it's fine in google hangouts and guvcview
<ldlework>
hi iqubic
<ldlework>
sound not working?
<iqubic>
No it isn't.
<ldlework>
run
<ldlework>
pactl load-module module-jack-sink
<energizer>
iqubic: make sure it's plugged in all the way. learned that lesson the hard way before
<energizer>
goibhniu: you mentioned your webcam isn't working with zoom. what kind of camera is that?
<energizer>
mine is usb logitech C310
<iqubic>
ldlework: Now I see a source in pavucontrol.
<ldlework>
iqubic: try a youtube video or something now
<iqubic>
I see a sink rather.
<ldlework>
iqubic: you should see the same in jack
<LevelChart8>
I have nix on a usb but when i boot into bios on windows startup and select the usb, the screen just goes blank for a moment and stays there
<ldlework>
though in jack, it's a source
<iqubic>
Ah, there we go. That seems to have made it work.
<LevelChart8>
it does not boot into nix
<ldlework>
iirc
<ldlework>
iqubic: :)
<iqubic>
Is there a way to add a jack sourc to pavucontrol?
<ldlework>
sec
<iqubic>
pactl load-module module-jack-source works for me.
<iqubic>
This is being really screwy and not working properly.
<iqubic>
I'm going to stick with basic pulseaudio
codygman has quit [Read error: Connection reset by peer]
codygman has joined #nixos
spacefrogg has quit [Quit: Gone.]
aw has quit [Quit: Quitting.]
aw has joined #nixos
spacefrogg has joined #nixos
<iqubic>
Is there an irc channel where I can talk to people about advanced PA sound stuff?
felixfoertsch23 has joined #nixos
eyenx has quit [Quit: bye]
eyenx has joined #nixos
eyenx has joined #nixos
eyenx has quit [Changing host]
felixfoertsch has quit [Ping timeout: 246 seconds]
felixfoertsch23 is now known as felixfoertsch
<LevelChart8>
I understand that it is possible to use npm and do javascript development in nixos, but do you still benefit from the package management aspects of NixOS?
<iqubic>
Yes. Of course.
<iqubic>
You should learn how to use a shell.nix to create a development environment for JS.
<LevelChart8>
great thank you
<LevelChart8>
this is great news
<LevelChart8>
i want nixos to deliver me from environment hell
tureba has left #nixos [#nixos]
iqubic has quit [Remote host closed the connection]
iqubic has joined #nixos
<iqubic>
Well, I've somehow screwed up my sound system even worse than it was before.
<iqubic>
I just did a rollback.
<iqubic>
To a time when I had only Pulse Audio on my system.
philr_ has joined #nixos
<iqubic>
Now pavucontrol is claiming that it can't find any cards at all, when I open the configuration tab.
<ldlework>
...er wasn't it working?
<iqubic>
ldlework: I just rolled back to a NixOS generation from last niight when my audio was working.
<iqubic>
I'm on that generation now, and it isn't working.
<ldlework>
iqubic: i mean just an hour ago
philr has quit [Ping timeout: 264 seconds]
<ldlework>
you watched a youtube video after getting the sink to work
<iqubic>
Oh, yeah. It was working, but the sound was all distorted. I then realized that I don't actually want to use JACK at all.
<ldlework>
lol
cmk_zzz has quit [Ping timeout: 256 seconds]
cmk_zzz has joined #nixos
<jlv[m]>
iqubic: I guess that's one definition of working.
<iqubic>
Turns out I accidentally selected the latest generation in my bootloader, instead of the one from a last night.
jluttine has quit [Ping timeout: 246 seconds]
<iqubic>
So, instead what I did is pulled my configuration.nix down from github, and rebuilt using that.
<jlv[m]>
LevelChart8: What part isn't clear? The partitioning?
philr has joined #nixos
<LevelChart8>
I guess. I thihnk I am supposed to be on step 2.1.1?
<energizer>
simpson: what's your JACK config look like?
<simpson>
energizer: Ugly and not shareable.
<jlv[m]>
LevelChart8: You can mostly ignore 2.1.1. If you need to connect with wifi, you can use the typical GUI widget. If you have an ethernet connection, I don't think you have to do anything.
<energizer>
ok
philr_ has quit [Ping timeout: 256 seconds]
<LevelChart8>
so for 2.2.1 (e.g. “parted /dev/sda -- mklabel gpt”) i enter a terminal in the nixos gui?
<exarkun>
Why'd my fetchFromGithub sha256 hashes suddenly change
<jlv[m]>
LevelChart8: Ya. NixOS doesn't have a graphical installer yet, so it's mostly terminal commands.
<LevelChart8>
ok
<jlv[m]>
exarkun: Is your `rev` a branch that may have recieved new commits?
mudtose has quit [Remote host closed the connection]
<LevelChart8>
one more question: the side keypad on my E595 doesn’t work. I can fix this later, right? i dont need to get a different nixos kernel/iso?
mudtose has joined #nixos
<exarkun>
jlv[m]: nope
<mudtose>
since we're talking about installation right now, do any of you know of an option for computers with MBR and UEFI?
<jlv[m]>
LevelChart8: One of the best things about NixOS is that (just about) everything can be updated from your NixOS configuration. Even after installing. Even the kernal :)
<LevelChart8>
ok cool
<LevelChart8>
when I ran “parted /dev/sda -- mklabel gpt” it says the partition is already being used. i assume that’s not an issue
<LevelChart8>
because that is the USB presumably
<jlv[m]>
LevelChart8: You should replace `sda` with the device you want to install on.
<mudtose>
USB is sdb on my machine
<mudtose>
If you are dual booting, you don't want to rewrite the whole table
<LevelChart8>
yes i want to dual boot
<mudtose>
and you you should skip the parted commands, I think
<mudtose>
I haven't been able to finish the installation myself but look at this guide:
<jlv[m]>
Those instructions are for a blank drive. You could wipe your drive if you follow them to dual boot on a single drive.
slack1256 has quit [Remote host closed the connection]
hke has quit [Read error: Connection reset by peer]
orivej has quit [Ping timeout: 256 seconds]
hke has joined #nixos
kamid has joined #nixos
<jluttine[m]>
Is it really so that any normal user can just run `nix-collect-garbage -d` and make older system generations unbootable?
<jluttine[m]>
hmm.. maybe not
<jluttine[m]>
not sure.. a huge amount of stuff was deleted when i wasn't running it as root, but something more was deleted when running as sudo. so maybe without sudo it doesn't delete anything needed by system generations?
drakonis has quit [Read error: Connection reset by peer]
<ashkitten>
jluttine[m]: any user can run nix-collect-garbage, only trusted users can delete generations
<ashkitten>
or maybe it's just root that can? don't remember
LevelChart8 has quit [Quit: LevelChart8]
palo1 has joined #nixos
drakonis has joined #nixos
<jluttine[m]>
ashkitten: yeah, so when any user runs it, nothing that is needed by any system generation won't get deleted, right?
<ashkitten>
yes
<jluttine[m]>
👍️
palo has quit [Ping timeout: 258 seconds]
palo1 is now known as palo
Nixola has quit [Ping timeout: 258 seconds]
Dagger has quit [Excess Flood]
Dagger2 has joined #nixos
drakonis has quit [Read error: Connection reset by peer]
drakonis has joined #nixos
urkk_ has joined #nixos
urkk has quit [Ping timeout: 256 seconds]
noudle has joined #nixos
<clever>
jluttine[m]: nix-collect-garbage -d can only delete generations from profiles it has write access to
<elvishjerricco>
Anyone else get `Unable to open printcap file /etc/printcap for read!` from `smbd` frequently? I'm not sure what correlates with getting this in my journalctl, but I get it a lot
<clever>
drwxr-xr-x 4 root root 44 May 23 12:30 /nix/var/nix/profiles/
<clever>
jluttine[m]: system profiles are in this directory, so only root can delete its generations
<clever>
ashkitten: ^
<clever>
if its a multi-user system, then you are limited to your own nix-env generations
<ashkitten>
ah, ty
<elvishjerricco>
It seems weird that samba would care at all about printers...
<clever>
elvishjerricco: samba includes the windows printer sharing logic i believe
<elvishjerricco>
Oh
<elvishjerricco>
Lol that's a little weird
<elvishjerricco>
Anything I can do to shut this message up?
<clever>
blame MS for rolling file-sharing and printer-sharing into a single protocol
<clever>
does printcap exist?
<elvishjerricco>
It does not
<elvishjerricco>
I don't think I've used a printer in 5 years :P
<clever>
what happens if you simply create it with `touch` ?
<elvishjerricco>
I'd have a silly file that I don't want to have :P But seriously, I worry there's some format the file is supposed to conform to that an empty file would not conform to
<clever>
most stuff in /etc/ is text based, and an empty file is just 0 lines
<elvishjerricco>
true
philr has joined #nixos
jb55 has quit [Ping timeout: 240 seconds]
philr_ has quit [Ping timeout: 265 seconds]
lightbulbjim has quit [Quit: Connection closed for inactivity]
jb55 has joined #nixos
eoli3n_ has joined #nixos
DerGuteMoritz has quit [Read error: Connection reset by peer]
zupo_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
aij has quit [Ping timeout: 264 seconds]
aij has joined #nixos
meh` has joined #nixos
eoli3n_ has quit [Ping timeout: 256 seconds]
korner has joined #nixos
<korner>
hello, is it possible to switch to grub bootloader or make the sysd boot more sorted out? like entry to newest gen and hide all other ones in submenu
<{^_^}>
[nixpkgs] @robaca opened pull request #89203 → merge-request-notifier: init at 1.7.0 → https://git.io/JfKPV
migy__ has quit [Ping timeout: 260 seconds]
migy__ has joined #nixos
Dagger2 has quit [Max SendQ exceeded]
Dagger2 has joined #nixos
Dagger2 is now known as Dagger
sphalerite has quit [Ping timeout: 260 seconds]
sphalerite has joined #nixos
elibrokeit has quit [Quit: No Ping reply in 180 seconds.]
meh` has quit [Ping timeout: 246 seconds]
elibrokeit has joined #nixos
Nixola has joined #nixos
Nixola is now known as Guest29099
cocreature has quit [Quit: A monad is just a monoid in the category of endofunctors, what's the problem?]
cocreature has joined #nixos
hke has quit [Read error: Connection reset by peer]
<energizer>
hello i can't boot, init script not found
<energizer>
i can rollback, but. what's it saying, how did that happen, how do i fix it?
<betaboon>
energizer: whats your kernel cmdline ?
lvmond_ has quit [Ping timeout: 272 seconds]
lvmond_ has joined #nixos
<betaboon>
energizer: do you get into stage1? or is the kernel complaining even before that?
<betaboon>
energizer: i would suspect your kernel-cmdline is missing `init=` ?
<energizer>
betaboon: stage 2 init script not found
<chrismatheson>
morning, has something recently changed in the nix setup script? i ran it like a week ago on my laptop and everything is fine. but today i get
<chrismatheson>
> you do not have 'xz' installed, which I need to unpack the binary tarball
<{^_^}>
error: syntax error, unexpected $undefined, expecting ')', at (string):313:17
<energizer>
betaboon: i would look for the kernel cmdline but not sure how to get a shell
<betaboon>
energizer: the kernel logs the cmdline in the very beginning
iqubic` has joined #nixos
<betaboon>
energizer: look for a line starting with "Kernel command line: ...."
iqubic has quit [Ping timeout: 260 seconds]
<energizer>
betaboon: i dont see that
meh` has joined #nixos
<betaboon>
energizer: can you post your kernel log (atleast the first 20 lines starting with "Booting Linux") to a gist or pastebin ?
<energizer>
betaboon: how do i get a shell?
<betaboon>
energizer: you dont need a shell for that
noudle has quit []
<energizer>
betaboon: how do i use a pastebin?
<energizer>
if i dont have a shell
dermetfan has quit [Quit: WeeChat 2.7.1]
<betaboon>
energizer: ah you dont have any access? not even through a serial console? i see your problem then
<energizer>
my options are "reboot immediately" "ignore the error and continue"
<betaboon>
energizer: you could add `boot.debug1` to your cmdline to get a shell right at the beginning of stage1
<energizer>
but ignore the error gives kernel panic
<betaboon>
energizer: or you could add `boot.shell_on_fail` to cmdline to get an option to enter a shell in that situation
<kini>
I think I don't really get how overrideAttrs works -- I overrode an attribute of a package, but another attribute which depends on that attribute didn't change. Is there some way to make that happen?
<energizer>
kini: how are you referencing the attribute?
<chrismatheson>
> <@chrismatheson:matrix.org> morning, has something recently changed in the nix setup script? i ran it like a week ago on my laptop and everything is fine. but today i get
<chrismatheson>
> > you do not have 'xz' installed, which I need to unpack the binary tarball
<{^_^}>
error: syntax error, unexpected '<', at (string):313:1
<{^_^}>
error: syntax error, unexpected '>', at (string):313:1
<chrismatheson>
not entirely sure how to use the new one though ....
<kini>
energizer: Basically, I have one package looking like `mkDerivation rec { ... a = [ "foo" ]; b = "...${f a}"; ...}` and another one where I have `c = pkg1.overridetAttrs (old: { a = old.a ++ [ "bar" ]; })`. In `nix repl`, `c.a` does indeed show `[ "foo" "bar" ]`, but `c.b` shows `"...${f a}"` and not `"...${f (a ++ [ "bar" ])}"`.
<kini>
I can show you the concrete code if you like, gimme a minute...
<kini>
I see -- is there some other way to do something like what I want?
Diogo has joined #nixos
chaoflow_ has joined #nixos
yasuaki has joined #nixos
yasu has joined #nixos
<hyperfekt>
kini: these overrides happen after the references are resolved. if there's no way around this late binding you intend, you could try and use the fixpoint of the derivation inside the derivation's attributes instead of a recursive record.
marusich has quit [Quit: Leaving]
chaoflow has quit [Ping timeout: 272 seconds]
chaoflow__ has joined #nixos
<hyperfekt>
actually, you won't have access to the derivations attributes unless you duplicate them in the passthrough attribute, so you probably want to make it a fixpoint of the record passed to mkDerivation instead of one of the derivation
<kini>
I see... well, there are definitely workarounds. I was just curious whether such a thing was possible, but I guess not easily :) Thanks!
chaoflow_ has quit [Ping timeout: 256 seconds]
<energizer>
betaboon: the init= points to a file that doesn't exist
<betaboon>
energizer: you did a simple `nixos-rebuild switch` and that succeeded ?
<energizer>
betaboon: afaik yes
<energizer>
betaboon: actually i think i cancelled it early
<energizer>
ctrl-c
chaoflow has joined #nixos
cyphase has quit [Ping timeout: 265 seconds]
<betaboon>
energizer: does rolling-back then running `nixos-rebuild switch` again fix your problem ?
<energizer>
betaboon: no
<energizer>
it thinks it's done
justanotheruser has quit [Ping timeout: 260 seconds]
<betaboon>
energizer: kinda hard for me to trace-back what might have happened as i dont know your setup and what you did :/
chaoflow has quit [Client Quit]
<betaboon>
energizer: but you ran it as `sudo nixos-rebuild switch` ?
<energizer>
yes
chaoflow has joined #nixos
<betaboon>
energizer: i wonder how store-contents that are clearly still used by an older-generation get missing. they should have a gc-root and therefore not go missing.
<betaboon>
energizer: you're using grub? and it is managed by nixos?
<energizer>
betaboon: nixos uefi boot
<betaboon>
energizer: is your /boot full by any-chance ? oO
knupfer1 has joined #nixos
knupfer has quit [Quit: knupfer]
knupfer1 is now known as knupfer
thc202 has joined #nixos
<energizer>
betaboon: no
<betaboon>
energizer: no clue what happens there. rolling back and running `sudo nixos-rebuild switch` should lead to a running system. i dont know what kind of configuration changes you did between the working and the broken configuration that might lead to this kind of behavior
<energizer>
i wish i'd been using `nixos-rebuild boot` as a post-commit hook instead of running it manually
<energizer>
then i could figure it out
<betaboon>
do you remember the changes you did between the generations ?
<energizer>
no
<energizer>
the last working generation is from weeks ago
<betaboon>
you don't have your configuration in git ? oO
<energizer>
i do, but it's not tested in CI, and i dont commit between all my local changes
__monty__ has joined #nixos
<energizer>
oh
<betaboon>
energizer: is the diff between your last-working state and the current state that big ?
<d4rkshad0w>
Hi. I want to install nixos onto a server from the rescue system. I have installed nix in the rescue system but I do not understand how to install nixos-install and nixos-generate-config. Is this even possible?
<betaboon>
i need a second opinion on something: I'm trying to boot using a readonly nfsroot, i have an init-script that creates an overlayfs ontop of that nfsroot. now where i need an opinion on: should i try to write this init-script to the exec stage1, or should i completly skip stage-1 and go straight to stage-2 ?
<d4rkshad0w>
<Mic92 "fabian: if you talk about the he"> Almost. I'm working on hetzner cloud. I'll have a look at it
<Mic92>
d4rkshad0w: did they not had an option to upload nixos iso's. Or do I mix something up?
<d4rkshad0w>
You can boot into a nixos ISO that's correct. But you can only control it via their web console. My goal is completely automatic server deployment so using their console is not an option....
gerschtli has quit [Quit: WeeChat 2.7.1]
gerschtli has joined #nixos
fendor has joined #nixos
Thra11 has joined #nixos
<betaboon>
anyone knows how `@systemConfig@` in stage-2-init.sh gets populated ?
asbachb has joined #nixos
greizgh has quit [Quit: greizgh]
knupfer has quit [Quit: knupfer]
knupfer has joined #nixos
<{^_^}>
[nixpkgs] @doronbehar opened pull request #89208 → rl2009.xml: Mention wrapMpv's new interface → https://git.io/JfKyb
calbrecht has quit [Remote host closed the connection]
<infinisil>
cinimod: What's your nix version?
cinimod has quit [Ping timeout: 272 seconds]
ericsagnes has quit [Ping timeout: 260 seconds]
<betawaffle>
is it bad practice to have nix expressions generated from build output? for example, if i wanted to generate separate derivations for each go module my project is dependent on
<manveru>
cross-compilation seems pretty broken for networkmanager...
<cinimod>
Whatever I did seems to have killed my daemon
<betawaffle>
does anyone use nixpkgs-fmt from inside vim?
<cinimod>
error: cannot connect to daemon at '/nix/var/nix/daemon-socket/socket': Connection refused
<cinimod>
:(
chloekek_ has joined #nixos
doyougnu has joined #nixos
sigmundv_ has quit [Ping timeout: 256 seconds]
<{^_^}>
[nixpkgs] @piperswe opened pull request #89212 → keybase-gui: set platforms to x86_64-linux → https://git.io/JfKQt
o1lo01ol_ has quit [Remote host closed the connection]
<teto>
betawaffle: I run it with `:!nixpkgs-fmt %` yeha
alexherbo2 has quit [Ping timeout: 264 seconds]
<{^_^}>
[nixpkgs] @moritzschaefer opened pull request #89213 → Add python packages: pyensembl and its dependencies, gffutils → https://git.io/JfKQY
<{^_^}>
[nixpkgs] @moritzschaefer closed pull request #87551 → Add python packages: pyensembl and its dependencies, gffutils, pyBigWig → https://git.io/JfW7g
noudle has quit []
<betawaffle>
well, the problem i'm having is the vim-nix plugin i'm using likes to indent things slightly differently from that. i was hoping for something more integrated
acarrico has joined #nixos
<cinimod>
As my boss used to say: when you have cleared out the problems in front you can see the problems behind
<cinimod>
/nix/store/wa7ws1zxydsm19x1hlv09fj42rdv0k53-arpack-3.7.0.drv' failed with exit code 2
<symphorien>
betawaffle: yes I have a binding to format the whole file
<symphorien>
nothing fancier though
Avaq has joined #nixos
<{^_^}>
[nixos-homepage] @github-actions[bot] pushed commit from @edolstra to master « Update flake.lock and blogs.xml [ci skip] »: https://git.io/JfK7J
Dagger has quit [Excess Flood]
<Avaq>
Hi folks! What's the easiest way in NixOS to install a (bash) script as a command into the PATH? Just adding it as a function to my .bashrc almost does the trick, but I specifically want an executable in PATH.
<Avaq>
I was thinking to make it a derivation with the script inlined into it somehow, and then adding `(callPackage ./my-derivation.nix {})` to my user packages.
est31 has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @bjornfor pushed commit from @r-ryantm to master « solaar: 1.0.1 -> 1.0.2 »: https://git.io/JfK5s
alexherbo2 has joined #nixos
jb55 has joined #nixos
teto has joined #nixos
urkk has joined #nixos
civodul has joined #nixos
urkk_ has quit [Ping timeout: 265 seconds]
LevelChart8 has quit [Quit: LevelChart8]
<eyJhb>
What does collistion between actually mean?
orivej has joined #nixos
LevelChart8 has joined #nixos
whatisRT has joined #nixos
<{^_^}>
[nixos-homepage] @github-actions[bot] pushed commit from @edolstra to master « Update flake.lock and blogs.xml [ci skip] »: https://git.io/JfK5D
<LevelChart8>
I am 99.99% sure that my system is UEFI, but when i ran nixos-install, I got this error: File system /dev/block/259:11 has wrong type for an EFI system partitioin (ESP)
whatisRT has quit [Quit: ERC (IRC client for Emacs 27.0.90)]
whatisRT has joined #nixos
<{^_^}>
[nixpkgs] @vbgl opened pull request #89218 → ocamlPackages.pa_test: remove at 112.24.00 → https://git.io/JfK5j
<alexherbo2>
what is the recommended TMP path?
<alexherbo2>
in my default environement, it is not set, and Elvish uses /tmp
<alexherbo2>
but in rails environement, TMP is set to /run/user/1000/
<alexherbo2>
so Elvish does not find the socket path
<alexherbo2>
manveru: maybe you can tell me
<alexherbo2>
why TMP is set in rails env
<romdisc>
Hi, I try to upgrade ue4 to 4.25.0 the build scripts in ue4 use mono but I need to map this executable to /nix/store/mono/bin/mono how can I sed -i all occurences of mono in these shell scripts to use nix version?
<simpson>
Hm. Well, I got nothing. git's supposed to not have this problem.
<exarkun>
I feel like I should be able to dig into the store where the build has succeeded already and learn more, but I'm not exactly sure how to do that digging.
<exarkun>
Apart from "nix is broken" or "git is broken" the only explanation that I can think of is that I entered the wrong hash a month ago but it worked anyway because ... there was a similar-enough object with that hash in my local development machine's store already.
is_null has joined #nixos
<exarkun>
But it fails anywhere else where nix actually has to build the store object
<exarkun>
But inspecting the object in my local nix store doesn't seem to support this, since it appears to contain the source from the requested git revision
ramses_ has joined #nixos
<ramses_>
Hey guys, anyone using elm2nix here? How do you manage static assets in that case? They are not taken into account by elm2nix, it seems
<ramses_>
It would be nice to be able to have them copied into the resulting nix-path
Dagger2 has quit [Excess Flood]
<exarkun>
If I got the hash wrong and it referenced some other thing I already had so it seemed to be right then the other thing should be _different_ from the revision I'm requesting ... it's not really though (as far as I can tell; just comparing a checkout of that revision to the object in /nix/store)
fendor_ has joined #nixos
cosimone has quit [Quit: Quit.]
<exarkun>
There are actually some text differences because the source contains $Format:...$ stuff that 'git archive' rewrites I guess?
<exarkun>
But that's supposed to be different I guess...
Dagger2 has joined #nixos
<exarkun>
If I use nix-prefetch-github locally, though, this gives the same hash as the remote builder which is failing expects, though ...
asbachb has joined #nixos
fendor has quit [Ping timeout: 240 seconds]
<asbachb>
Hi. Do I need to add a package to `buildInputs` when I reference a file as default value in default module option?
<Izorkin>
talyz: etu: it is possible to fix the work of phpExtensions in nginx unit
<exarkun>
I wonder if there is some way to find the store object that has this _other_ hash that fetchFromGithub does not like
<exarkun>
Hmm but I wonder what file name I should supply.
artemist has joined #nixos
dingenskirchen has quit [Quit: dingenskirchen]
dingenskirchen has joined #nixos
<chaoflow>
Pandas can't find plotly backend when installed in same environment (https://pastebin.com/d0sBjM3F). Any thoughts on this? On an ubuntu system with pandas and plotly via pip, it works.
<contrun>
WARNING: Could not access 'pyvenv.cfg' despite a virtual environment being active. Assuming global site-packages is not accessible in this environment.
<contrun>
I am unable to install a python package with pip install --user. Can anyone reproduce this? pip install git+git://github.com/thann/play-with-mpv --user
<contrun>
ERROR: Can not perform a '--user' install. User site-packages are not visible in this virtualenv.
orivej has quit [Ping timeout: 258 seconds]
<chaoflow>
kai_w: That did it! Thanks a lot for your help! I like the overlay system. Nixpkgs has come a long way!
leo60228 has joined #nixos
shafox has joined #nixos
<leo60228>
i have a low-end laptop that needs a specific kernel option for the keyboard to work
ramses_ has quit [Ping timeout: 272 seconds]
<leo60228>
this option isn't available by default
<leo60228>
the laptop is wayyyyyyyy too slow to compile a kernel
<leo60228>
s/available/enabled
<leo60228>
what's the best way to handle this?
<leo60228>
build a kernel on my desktop, upload it to cachix, and PR nixpkgs to enable the option?
<leo60228>
there's a similar option for an adjacent device that nixpkgs does this for
<{^_^}>
#10416 (by puffnfresh, 4 years ago, closed): linux: add config options needed for a Bay Trail Chromebook
shafox has quit [Read error: Connection reset by peer]
<etu>
Izorkin: Ok, make a PR then?
LevelChart8 has joined #nixos
shafox has joined #nixos
<jlv[m]>
leo60228: You can send a Nix store closure to another machine. I don't remember the commands off the top of my head, but it is possible to build the configuration on your desktop and send it directly to the laptop.
<clever>
nix-copy-closure
romdisc has quit [Remote host closed the connection]
<cmcaine[m]>
Hi, I've just started using Nix, and my favourite programming language doesn't have an up to date package. I downloaded a pre-built binary from their site https://julialang.org/downloads/ and tried to run it, but I get the mystifying error "zsh: no such file or directory: ./julia-1.4.2/bin/julia"
<cmcaine[m]>
Obviously I've checked that the file does actually exist.
<{^_^}>
[nixpkgs] @eadwu opened pull request #89221 → nixos/kvmgt: allow multiple uuids on same mdev → https://git.io/JfKAb
<cmcaine[m]>
I'd appreciate some help with getting this binary package working.
<jlv[m]>
cmcaine: You probably need to change the interpreter `patchelf --set-interpreter "$(nix eval --raw nixos.glibc)"/lib/ld-linux-x86-64.so.2 path/to/binary`
<alexarice[m]>
cmcaine: you might have more success modifying the existing derivation rather than trying to use a prebuilt binary
<clever>
jlv[m]: you want to use $NIX_CC and nix-support, /lib/ld-linux-x86-64.so.2 will only work for 64bit x86
<clever>
jlv[m]: that file always has the right path for the stdenv your in (via nix-shell -p patchelf)
<cmcaine[m]>
The patchelf line gave me this: error: attribute 'nixos' in selection path 'nixos.glibc' not found
<cmcaine[m]>
nixpkgs instead worked
<jlv[m]>
cmcaine: alexarice is right. There is a Julia package in nixpkgs. If you want a newer version, check the unstable branch first, and update the package itself if that isn't up to date.
<cmcaine[m]>
patchelf worked, though!
<cmcaine[m]>
jlv: I've already cloned nixpkgs, but the nix expression is quite complicated.
<simpson>
cmcaine[m]: I'm looking at the Julia packaging right now. It looks like there's a shared description, and then hacks for each minor version.
<alexarice[m]>
You could make an issue on nixpkgs saying the new version is missing
chloekek_ has quit [Ping timeout: 272 seconds]
<cmcaine[m]>
I'm also going to want to run other binary packages sometimes without writing a whole nix expression, unless I get particularly good at the language.
<cmcaine[m]>
simpson: v1.3 doesn't follow that pattern
Jackneill has quit [Read error: Connection reset by peer]
<Izorkin>
etu: nor loaded defaultPhpExtensions
agjacome has joined #nixos
agjacome has left #nixos [#nixos]
agjacome has joined #nixos
<{^_^}>
[nixpkgs] @etu opened pull request #89224 → unit: Expose PHP expressions used so it can easily be accessed for configs → https://git.io/JfKxx
<etu>
Izorkin: It use the default extensions, otherwise the test wouldn't work?
<cmcaine[m]>
The destination of this symlink doesn't exist. That doesn't seem right to me?
<cmcaine[m]>
lrwxrwxrwx 1 colin users 45 May 30 05:30 /home/colin/.nix-defexpr/channels -> /nix/var/nix/profiles/per-user/colin/channels
<cmcaine[m]>
› ls -l ~/.nix-defexpr/channels
meh` has joined #nixos
<cmcaine[m]>
Discovered because I had to explicitly specify nixpkgs to use golint: ` nix run -f '<nixpkgs>' golint `
nikivi has quit [Ping timeout: 272 seconds]
<Izorkin>
If "file": "${pkgs.unit.usedPhp74}/lib/php.ini", all extensions load. If php74-test = pkgs.unit.usedPhp74.withExtensions ({ all, ... }: with all; [ opcache memcached ]); not loaded
<etu>
Izorkin: You have to override the php package sent to unit
<Izorkin>
etu: with php74-test = pkgs.unit.usedPhp74.withExtensions ({ all, ... }: with all; [ opcache memcached ]); and package = pkgs.unit.override { php74 = php74-test; }; not loaded opcache memcached
<etu>
Izorkin: why would php74-test be based of unit.usedPhp74? Just use php74 as base
i077 has quit [Ping timeout: 256 seconds]
<Izorkin>
etu: checking
<etu>
Izorkin: I did this override to unit php74.withExtensions({ enabled, all }: enabled ++ [ all.memcached ]); and added memcached to the unit test, and it found memcached in the phpinfo output
<cmcaine[m]>
In gparted, click on the unallocated space, then click new, then make a partition with the GUI
domogled has joined #nixos
<LevelChart8>
formatted for ext4
<cmcaine[m]>
Yep
<cmcaine[m]>
Other filesystems are available, but ext4 is fine.
alexherbo2 has joined #nixos
<LevelChart8>
i am pretty sure Lenovo is UEFI though
<LevelChart8>
from the docs
<cmcaine[m]>
so what?
<LevelChart8>
"File system /dev/block/259:11 has wrong type for an EFI system partition (ESP)"
<LevelChart8>
that happened last time i tried to run nixos-install
<LevelChart8>
when the partition was ext4
<cmcaine[m]>
But did you mount that first partition in the table to /boot?
<LevelChart8>
yes the last time
<cmcaine[m]>
You're not trying to boot from the new partition you're making
<LevelChart8>
but i will skip this time
<LevelChart8>
and just have 1 partition called nixos
<LevelChart8>
(and another for Windows)
<cmcaine[m]>
You need to be making a new partition for linux and writing the EFI stuff to the existing EFI partition.
<cmcaine[m]>
Then booting with UEFI
<LevelChart8>
right but i dont know why i got the “wrong type for EFI system partition last time"
<cmcaine[m]>
Anyway, that's all I did. I didn't make any edits to the configuration.nix to get it to work.
<cmcaine[m]>
Good luck!
<LevelChart8>
thanks
<cmcaine[m]>
Thing to be aware of: Ext2fsd for windows completely fucks ext4 partitions if you have it installed.
<puck>
LevelChart8: the ESP, which is usually mounted at /boot, needs to be FAT32, since it is read by the UEFI
<cmcaine[m]>
If you want to access your ext partition from windows then you need to use ext3 or (avoid certain options in ext4).
agjacome has quit [Quit: agjacome]
averell has quit [Quit: .]
fusion809 has quit [Remote host closed the connection]
<LevelChart8>
puck: so i do need that boot partition as FAT32 and nixos partition as ext4
gustavderdrache has quit [Quit: Leaving.]
eoli3n_ has joined #nixos
<puck>
LevelChart8: yeah. use the existing "SYSTEM" partition (/dev/nvme0n1p1) as your /boot, do not reformat it (it'd break Windows, can be fixed with a boot iso), then format your / as something else and set it up like that
<Izorkin>
etu: thanks for help. Worked.
<LevelChart8>
ok the “New Partition #1” is called nixos and using ext4
<LevelChart8>
I tried `mkdir /mnt/boot && mount /dev/disk/by-partlabel/nixos\\x20system\\x20partition /mnt/boot` as suggested above but it did not work
<LevelChart8>
“mount: /dev/disk/by-partlabel/nixos: can’t find in /etc/ fstab”
<puck>
the partlabel should be just "SYSTEM" i think
magnetophon has quit [Ping timeout: 264 seconds]
<LevelChart8>
same error
cinimod` has quit [Ping timeout: 240 seconds]
<puck>
hm. try mount /mnt/nvme0n1p1?
<puck>
(into /mnt/boot)
<LevelChart8>
what do yuo mean “into /mnt/boot”?
<puck>
mount /dev/nvme0n1p1 /mnt/boot, sorry
philr has quit [Ping timeout: 246 seconds]
<LevelChart8>
ok
<LevelChart8>
that worked
pamplemousse has quit [Ping timeout: 256 seconds]
domogled has quit [Quit: domogled]
<LevelChart8>
“nixos-generate-config --root /mnt” now this?
eoli3n_ has quit [Ping timeout: 256 seconds]
<puck>
something like that, i'd say
<etu>
Izorkin: Error or warning/notice?
<LevelChart8>
lol
shafox has quit [Read error: Connection reset by peer]
<etu>
Izorkin: I don't see this error and it seems to load for me
<puck>
LevelChart8: .. wait, did you mount your / first
<LevelChart8>
i suppose not - everything i did is in that screenshot
<puck>
right, so unmount /mnt/boot, then mount your nixos disk (check gparted or so for the path), then mount the boot into /mnt/boot. i need to go, so i can't really help any further
<etu>
LevelChart8: Which partition is it? root or efi?
<LevelChart8>
its the nixos partition in the screenshot
Guest29099 has quit [Quit: leaving]
<etu>
LevelChart8: Looking at the image tells me that you should do: mount /dev/nvme0n1p5 /mnt
<etu>
LevelChart8: And then: mount /dev/nvme0n1p1 /mnt/boot
drakonis has quit [Ping timeout: 260 seconds]
<etu>
LevelChart8: So the first one will mount your new root partition on /mnt and the second one will mount the efi partition, you may want to do "mkdir /mnt/boot" before the second mount line
<LevelChart8>
i can “cat” the file but I need to edit it
zupo has joined #nixos
<etu>
nano is preinstalled
magnetophon has joined #nixos
<LevelChart8>
the GUI is a environment.systemPackages item?
<etu>
Nah, environment.systemPackages is a way to install packages to the global user environment
<etu>
Graphical sessions are services, such as services.xserver.desktopManager.plasma5.enable
<etu>
There's examples of that further down in the file
<LevelChart8>
oh i see
<LevelChart8>
and i comment that in
<LevelChart8>
or “uncomment” iit
<etu>
yeah, you need the line about enabling the xserver as well
<m1cr0man>
Hey folks. Has anyone got experience using nixops and morph? I have some servers that I'm not using either with right now, and I'm trying to decide which one to invest in. Morph looks like the better choice from my birds eye view reading the docs.
<etu>
LevelChart8: And when you're happy with your changes, you can do a "nixos-rebuild" followed by "boot" to get a new boot generation to reboot to or "switch" to create a new generation and to register it as a boot generation
<LevelChart8>
not “nixos-rebuild switch”?
<LevelChart8>
ok
<LevelChart8>
yes
<etu>
The switch will also activate the new generation
delroth has quit [Ping timeout: 260 seconds]
<LevelChart8>
build failed. it was unable to download many things
<LevelChart8>
i guess i need to connect to wifi from the CLI somehow
vidbina has joined #nixos
delroth has joined #nixos
<etu>
LevelChart8: yeah... I don't know if you have the utils to handle wifi... That's kinda why people edit the config before reboot. Then one can enable things like... networkmanager etc.
<LevelChart8>
so am i screwed
<{^_^}>
[nixpkgs] @Mic92 merged pull request #89221 → nixos/kvmgt: allow multiple uuids on same mdev → https://git.io/JfKAb
<etu>
LevelChart8: you should look at /mnt/etc/nixos/configuration.nix
<etu>
LevelChart8: Otherwise it's from the USB
codygman has joined #nixos
Dagger2 has quit [Excess Flood]
Dagger2 has joined #nixos
<LevelChart8>
networking.wireless.enable = true;
<etu>
I would still enable networkmanager
<etu>
And the firmwares
shabius has joined #nixos
codygman has quit [Ping timeout: 260 seconds]
<etu>
LevelChart8: When you have done your edits you just run nixos-install again
<energizer>
LevelChart8: "networking.wireless.enable = true;" means wpa_supplicant not networkmanager
magnetophon has quit [Read error: Connection reset by peer]
<energizer>
and iirc those options conflict
codygman has joined #nixos
<etu>
I don't think they do... but not sure... but wpa_supplicant won't give as nice of a user interface as networkmanager
vidbina has quit [Ping timeout: 264 seconds]
<Avaq>
Hi. I have xrandr in my PATH, but there is no mention of it in my NixOS configuration. Is there a way I can query NixOS to find out why this is in my PATH?
<LevelChart8>
i fiinished making edits to /mnt/etc/nixos/configuration.nix but it would not let me write the file
<LevelChart8>
so how will i know if the GUI is properly installed
<LevelChart8>
and i wont end up in CLI again on rebooting
<Avaq>
energizer: Ah, yes. I remember seeing why-depends somewhere, but I couldn't remember the parent command, nor the actual exact wording. Anyway, for xrandr, this has told me: "nixos-system-20.03 -> activate (truncated WRAPER_PATH) => system-path -> bin/xrandr => xrandr-1.5.0", which hasn't given me much more to go by. :\
<energizer>
Avaq: yeah that's not much is it
<Avaq>
Seems like it's just telling me that "it's part of your system because it is part of your system" :P
<energizer>
Avaq: there's another command with nix-store -q but i forget exactly
<Avaq>
These packages all seem to "depend on" xrandr, explaining why xrandr is there at all, but not quite explaining why it is included in my PATH. I thought package dependencies are not usually exposed to the user.
<Avaq>
There's even some packages there that I have uninstalled in the meantime. I'm guessing they'll disappear after I collect garbage.
<Avaq>
Huh.. They didn't.
cinimod` has joined #nixos
<kai_w>
if I'm using a remote builder, does 'post-build-hook' run on the builder or on my local machine?
iqubic` has joined #nixos
iqubic` has quit [Remote host closed the connection]
<kai_w>
^ if I have the same hook configured on both machines
iqubic` has joined #nixos
iqubic` has quit [Client Quit]
iqubic has quit [Ping timeout: 260 seconds]
iqubic has joined #nixos
cinimod` has quit [Ping timeout: 272 seconds]
<Avaq>
energizer: There's a `--all` flag for `nix why-depends` which "shows all paths in the dependency graph" rather than just one. Some of the paths included now are more insightful. :)
NeoCron has joined #nixos
<energizer>
Avaq: ah nice
<etu>
LevelChart8: I spotted a syntax error where you defined your user
<etu>
LevelChart8: The closing }; is commented out
knupfer has quit [Ping timeout: 256 seconds]
<etu>
LevelChart8: And opening the ports ... won't probably work, you should put numbers there if you care
mallox has quit [Quit: WeeChat 2.8]
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
<LevelChart8>
the GUI seems to have installed properly except the password i set at the end of nixos-install does not seem to work on login
<LevelChart8>
leaving the password field empty doesnt work either
armin has joined #nixos
<eyJhb>
Does this even work?
<armin>
hellooooo :-) so how do i find out what kind of hash this here is? hashedPassword = "$6$/z4n8AQl6K$kiOkBTWlZfBd7PvF5GsJ8PmPgdZsFGN1jPGZufxxr60PoR0oUsrvzm2oQiflyz5ir9fFJ.d/zKm/NgLXNUsNX/";
<eyJhb>
Where does it get the profiles from, when it is not a user service?
<{^_^}>
[nixpkgs] @Ekleog merged pull request #89191 → nix-daemon module: do not rely on the daemon being running → https://git.io/JfKRP
<{^_^}>
[nixpkgs] @Ekleog pushed to release-20.03 « nix-daemon module: do not rely on the daemon being running (#89191) »: https://git.io/Jf6vj
adrielus has joined #nixos
<infinisil>
armin: man $(nix-build --no-out-link '<nixpkgs>' -A manpages)/share/man/man3/crypt.3.gz
Jonathan92 has joined #nixos
<adrielus>
Hey, how can I fetch a tar.gz file from mega.nz?
o1lo01ol1o has quit [Ping timeout: 246 seconds]
<{^_^}>
[nixpkgs] @Ekleog pushed commit from @notriddle to release-20.03 « buildRustCrate: Replace hyphen with underscore in env variables (#88054) »: https://git.io/Jf6fU
<armin>
infinisil: perfect, thanks a lot.
dottedmag has quit [Quit: QUIT]
<Jonathan92>
Hey everyone, I am new to nix and doing an install from the minimal image. I am having issues getting my network setup. There seems to be an issue with overly long names in wpa_supplicant https://github.com/benzea/gnome-network-displays/issues/70 . It looks like the solution is to enable net.ifnames = 0, but I am not sure which .nix file to add
<adrielus>
saying the name of the file cannot contain #
<adrielus>
the # is from the mega.nz link
dottedmag has joined #nixos
dottedmag has joined #nixos
dottedmag has quit [Changing host]
<catern>
should I be able to build static executables if I don't explicitly pass glibc.static in my buildInputs?
<catern>
if I do explicitly pass it, I'm able to build a static executable with ./configure LDFLAGS='-static' and some other stuff, but normal ./configure doesn't work
<catern>
(some segfault in the configure tests?)
<energizer>
Jonathan92: i dont know much about wpa_supplicant except that networkmanager now has an alternative backend available called iwd
<simpson>
catern: Are you using pkgsStatic?
<catern>
no, I don't want to build everything statically - I'm just working on a project (using autotools) that needs to build one static executable (to bootstrap in empty filesystems) but otherwise wants to use normal dynamic libraries
i077 has joined #nixos
i077 has quit [Client Quit]
<LevelChart8>
is there a reason why my GUI password would not be the password i set at the end of “nixos-install”?
<energizer>
LevelChart8: are they for the same user?
<LevelChart8>
i didn’t set a password for the user in configuration.nix and there is only one user in configuration.nix
<energizer>
log in as root and do `passwd levelchart8`
<LevelChart8>
thank you
<LevelChart8>
that let me reset the password
<LevelChart8>
not clear on what it thought the password was before
<LevelChart8>
but it’s fixed now
o1lo01ol1o has joined #nixos
drewr has joined #nixos
<energizer>
i think it doesn't have a password (disallowing all logins)
Soo_Slow has quit [Remote host closed the connection]
magnetophon has joined #nixos
pjt_014 has joined #nixos
<etu>
Izorkin: Do you mind adding a cc for @NixOS/php in that issue regarding mysqlnd?
<etu>
Izorkin: We have a maintainer team now :)
adrielus has quit [Remote host closed the connection]
<etu>
I'm pretty sure it will help, unless you got some super weird hardware
<LevelChart8>
I am on E595 Lenovo
<LevelChart8>
i rebuilt and switched. i will reboot now.
<LevelChart8>
it still does not show up. do i need to go all the way back to the USB?
Avaq has quit [Quit: Leaving.]
<etu>
Does it show up if you run "ip a"?
<etu>
I would still recommend networkmanager ;)
<catern>
hmmmmm so NixOS passes "--disable-static" to my configure script lol
<catern>
should I work around this in my configure script and really really force static? (since I do need static libraries internally) or should I just change my NixOS package to not pass --disable-static?
<{^_^}>
[nixos-homepage] @garbas pushed to landing-page « initial version of the video »: https://git.io/Jf6Uz
<{^_^}>
[nixos-homepage] @garbas pushed 22 commits to landing-page: https://git.io/Jf6Uw
is_null has joined #nixos
<{^_^}>
[nixos-homepage] @garbas pushed to landing-page « too long »: https://git.io/Jf6Ur
civodul has quit [Quit: ERC (IRC client for Emacs 26.3)]
globin_ has joined #nixos
<iqubic>
So, I'm trying to create a simple executable shell script in NixOS. I tried adding the shebang: "#!/bin/bash" but when I run the script I'm being told that /bin/bash doesn't exist. What is the issue here?
lunaa has joined #nixos
<{^_^}>
[nixos-homepage] @garbas pushed to landing-page « rebuild when demo.scenario changes »: https://git.io/Jf6UX
globin has quit [Ping timeout: 260 seconds]
<clever>
iqubic: /bin/bash doesnt exist, you can usually get away with /bin/sh
<ramses_>
iqubic: try `#! /usr/bin/env bash`
<iqubic>
I see. I'll try just /bin/sh.
<iqubic>
Basically I found this script on the arch wiki at the bottom of the page on Pulse Audio and I want to see if that will work for me on NixOS. http://dpaste.com/1X50853
<energizer>
error: 'flake' is not a recognised command
<cmcaine[m]>
iqubic: if you want to check if a script is posix sh compatible, you can also use shellcheck. Shellcheck is stricter than you need to be, almost all the time, but it's quite nice for avoiding the abundant sharp edges of shell languages.
chloekek_ has quit [Ping timeout: 265 seconds]
<gilligan>
energizer: flakes
<gilligan>
flake*s* plural
<cole-h>
energizer: You also need `nixFlakes`. There is no flakes code in neither nix nor nixUnstable.
<energizer>
can i just add it to my environment.systemPackages?
<gilligan>
possible but i wouldn't
<cole-h>
Sounds like trouble just waiting to happen. Unless you're actually switching to nixFlakes, using it in a shell should be good enough.
<cole-h>
And if you are switching to flakes, you should set `nix.package`, not add it to your `environment.systemPackages`.
mnoronha has joined #nixos
<mnoronha>
Hey, I think I'm running a old version of nixos, synced to an old nixpkgs channel. Could I get some help figuring out what (and how) I need to update things to access tensorflow 2.0?
<{^_^}>
[nixos-homepage] @garbas pushed to landing-page « adding some documentation »: https://git.io/Jf6T1
<infinisil>
mnoronha: Do you need that for a project you have a shell.nix for?
<mnoronha>
Currently the way I'm getting tensorflow is by grabbing `py.pkgs.tensorflowWithCuda` in a shell.nix
<infinisil>
Or just to try it out quickly?
<mnoronha>
Yea, ideally shell.nix
<infinisil>
mnoronha: I guess you're using `import <nixpkgs>` in your shell.nix?
<mnoronha>
Is it okay if I pin a newer version than my OS version?
<infinisil>
Yup, no problem at all
<infinisil>
(probably)
<infinisil>
(you could have problems if your NixOS is too old)
lunaa has quit [Remote host closed the connection]
lunaa has joined #nixos
proofofkeags has joined #nixos
m0rphism has joined #nixos
maddo has quit [Quit: See ya]
asbachb has quit [Remote host closed the connection]
<mnoronha>
infinisil: thanks again, I think it's working -- I switched from unstable to 20.03 to see if I can get a binary instead of building from source
meh` has quit [Ping timeout: 256 seconds]
<mnoronha>
jk looks like I'm going to be building some stuff from source either way :)
<cmcaine[m]>
infinisil's suggestion sounds easier (I think it will edit sudoers for you?)
<infinisil>
Yeah looks like it
magnetophon has joined #nixos
gilligan has quit [Quit: WeeChat 2.7.1]
justanotheruser has joined #nixos
reallymemorable has quit [Ping timeout: 240 seconds]
mnoronha has quit [Remote host closed the connection]
<energizer>
is there some reasonable threat model where security.sudo.wheelNeedsPassword=false is a bad idea?
<cmcaine[m]>
energizer: if you expect a bad actor to get onto a user account with sudo, but compromising that account itself isn't bad.
<pistache>
energizer: I don't think so
<energizer>
cmcaine[m]: i can't imagine that
<pistache>
cmcaine[m]: well unless the user's environment is truly immutable, it's not a very realistic threat model
<pistache>
I mean, compromising the account itself allows for aliasing sudo, for example
<pistache>
allows to alias*
mbrgm has quit [Ping timeout: 246 seconds]
mbrgm has joined #nixos
<cmcaine[m]>
Yes, obviously, they can alias sudo and mess with the shell and so on.
<cmcaine[m]>
Yeah, I can't think of any non-contrived scenarios for it ;)
<cmcaine[m]>
Except, as you say, the home dir itself is mounted read only.
<cmcaine[m]>
(and there are protections on the process space of running processes under the same user)
<cmcaine[m]>
But why would
<pistache>
anyone do that
<cmcaine[m]>
you bother with that if you're going to have the sudo password off ;)
<pistache>
exactly
urkk_ has joined #nixos
urkk has quit [Ping timeout: 256 seconds]
kamid has joined #nixos
<cmcaine[m]>
I think just not having a home dir might be more common in enterprise setups.
gustavderdrache has quit [Quit: Leaving.]
<emily>
energizer: yes: now malicious stuff can automate sudo without needing your interaction
<emily>
of course, you often have stuff you really don't want random scripts from the internet accessing as your user anyway
<emily>
but it doesn't change the fact that git clone ... && cd foo && nix-shell --run make becomes meaningfully more dangerous
<energizer>
emily: i guess that could work against automated attackers. i imagine any human attacker would get root easily from that point
<maurer>
Honestly I think that keeping the password on is mostly important as a guard against well-intentioned but bad-acting scripts
<maurer>
If someone has access to your account, but sudoing needs a password, they'll just hook some activity requiring root to instead authenticate a root shell, at which point they execute both their command and yours
<maurer>
but if, say, you did ./configure --prefix=~/some-local-thing; make; make install
<maurer>
some of these will in fact try to sudo and do things to the global system, even though they really shouldn't
<energizer>
if they added askpass to my zshrc i'm pretty sure i'd enter my password without thinking about it
growpotkin has joined #nixos
nikivi has quit [Remote host closed the connection]
matthiaskrgr has quit [Remote host closed the connection]
<maurer>
Another possibility is that if you have multiple valuable user accounts on your machine, and disk crypto, this increases your risk if the machine is physically stolen