<samueldr>
tobiasBora0: you can boot in the previous generations still, yes?
<samueldr>
tobiasBora0: what kernel are they using?
toastyavocado has joined #nixos
<samueldr>
IIRC older u-boots didn't know about the 3B+, so your u-boot reported as a 3B, and the mainline kernel is now incompatible with a 3B+ acting like a 3B
<jlv>
Is the minimal installer ISO supposed to boot into a GRUB menu with no kernel loaded? From the description, I assumed it at least booted into a shell.
<samueldr>
which might be the issue you're having
cosimone has quit [Quit: Quit.]
o1lo01ol1o has quit [Ping timeout: 256 seconds]
<tobiasBora0>
samueldr: I can boot indeed on an older generation with kernel 5.6
<samueldr>
tobiasBora0: can you tell me the output of `dmesg | grep 'Machine model:'` ?
<samueldr>
one of the first thing the kernel prints is the model name *as configured in the device tree*
<samueldr>
so if it's the situation where a 3B+ shows up as a 3B I think (not sure) that it would show up as a 3B there
lovek323 has joined #nixos
ericmoritz has quit [Remote host closed the connection]
<risson>
Hey everyone! I'm having trouble getting dns in a NixOS container. It seems the networking.nameservers I am setting inside the container configuration isn't used :/
<tobiasBora0>
samueldr: hum, it's written [ 0.000000] Machine model: Raspberry Pi 3 Model B+
<samueldr>
then I guess my assumption is not the right one
<samueldr>
I know other users have had issues with 3B+, though I don't have one to hel
<samueldr>
help*
<jlv>
nvm, iso wasn't copied correctly
<tobiasBora0>
samueldr: how is it possible to update u-boot? I need to reinstall the new image 20.03 and migrate all data from the old install to the new one?
<samueldr>
that's an unresolved question still ;/
<samueldr>
:/*
<lovek323>
Hey all. I've received a message that I'm no longer in the nixpkgs-committers group, just when I've come back to NixOS with my work system. Any chance of getting added back into that group?
<lovek323>
I have one PR open that needs some review or merging: https://github.com/NixOS/nixpkgs/pull/96728 and I will be opening another to add TablePlus to nixpkgs soon.
<tobiasBora0>
samueldr: I was reading the sd card on another computer when I answer your question. But I think the /boot mounts the partition 1 folder nixos, but I'll try to reboot to confirm
<samueldr>
that's almost a given
<samueldr>
there should at least be extlinux.conf in there
<{^_^}>
[nixpkgs] @marsam pushed commit from @Chili-Man to master « terraform_0_13: 0.13.1 -> 0.13.2 (#96971) »: https://git.io/JUOmh
<lovek323>
samueldr: thanks, yep, I saw that. just wondering if I can get back in now that I'm likely to be more active?
<samueldr>
lovek323: I make no decisions, just linked in case *that* particular issue escaped you
<lovek323>
yep, thanks!
<samueldr>
lovek323: I would simply ask in there
<samueldr>
the RFC does state:
<samueldr>
>> Previous committers are moved to a new Nixpkgs Committers Emeritus team, to honor their past contributions. Members of this team will remain in the GitHub organisation, and may regain push access at a later time.
<samueldr>
so you may regain push access at a later time, I guess ;)
spease has quit [Remote host closed the connection]
cosimone has joined #nixos
<tobiasBora0>
I think I'm going to reinstall a new version from scratch. Just, while I'm running the current version, is there a simple way to backup the data? For example, can I just copy/paste the /var folder?
<samueldr>
tobiasBora0: still two partitions, but the FAT32 partition for the raspberry pi is used as an opaque blob we don't care about in the running system
<samueldr>
and for data, it depends what you have been running
<samueldr>
nothing raspberry-pi specific
cosimone has quit [Client Quit]
<samueldr>
might want /etc/ssh too
<samueldr>
and /home
ris has quit [Ping timeout: 258 seconds]
<tobiasBora0>
samueldr: just some apps like gitea (based on postgresql)
<samueldr>
yeah, so whatever those apps add as state I guess :)
<cole-h>
jtojnar: Thanks, I figured it was something like that. I was just curious what the difference between `nix repl '<nixpkgs/nixos>'` and `nix repl '<nixos-config>'` should be.
palo has joined #nixos
<{^_^}>
[nixpkgs] @cyraxjoe closed pull request #49734 → nixos/dockd: add new service → https://git.io/fxhMn
<jtojnar>
I would not even expect `nix repl '<nixos-config>'` to work
orivej has quit [Ping timeout: 240 seconds]
xd1le has joined #nixos
<jtojnar>
cole-h the `configuration.nix` usually contains a function and nix repl would not know what to pass it
<tobiasBora0>
By the way, unrelated questions, but what is the usual way to create the secret files? Because for now I've a script that create the files, and give them the appropriate permissions (for example chmod gitea:gitea XXX). But the issue is that when gitea is not yet installed, this fails. So for now I build once the system, put the appropriate permissions, and rebuild again, but sometimes it corrupts
<tobiasBora0>
some states (got an issue with nextcloud recently), and it's not very practical
<cole-h>
So `nix repl '<nixpkgs/nixos>'` functions, and `nixos-option` does too.
<samueldr>
tobiasBora0: you could put back the same stateVersion that you had, since you're putting back the same state
<samueldr>
tobiasBora0: in fact you probably should
<tobiasBora0>
samueldr: ok, I'll do that then, thanks
orivej has quit [Remote host closed the connection]
orivej has joined #nixos
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
flythief has joined #nixos
knerten has joined #nixos
<flythief>
hi, my opengl program throws glfw error with " GLX: No GLXFBConfigs returned". I just add pkgs.glfw3 in my default.nix by direnv. And my glxgears works. glxinfo show GLXFBConfigs like this file. Anyone have some suggestions? thanks firstly. https://www.irccloud.com/pastebin/iBwWTeIR/
<ryantm>
Maybe I'm using my flake in the wrong way then because I'm getting:
<ryantm>
Oh... I misread the error message, sorry.
<cole-h>
:P
<cole-h>
ryantm: \For curiosity's sake, what was the error message?
<ryantm>
flake 'git+file:///home/ryantm/p/nixfiles' does not provide attribute 'packages.x86_64-linux.defaultPackage.x86_64-linux', 'legacyPackages.x86_64-linux.defaultPackage.x86_64-linux' or 'defaultPackage.x86_64-linux'
<ryantm>
cole-h: I was calling nix build, when I should have been calling nixos-rebuild switch --flake
<cole-h>
Yeah, that's what I was about to ask.
spease has joined #nixos
<spease>
I'm getting "Unable to use C math library functions (with or without -lm)" when trying to build a library that's using ios.toolchain.cmake - anybody have an idea why that might be, or something I could do to figure out what I need to do to fix it?
Yaniel has quit [Ping timeout: 246 seconds]
<spease>
Thanks
<cole-h>
ryantm: You can also set `defaultPackage.x86_64-linux = nixosConfigurations.hostname.config.system.build.toplevel;`, I think. Then `nix build` would at least build the system config (but not switch to it)
<ryantm>
cole-h: I don't actually know how to reference the switcher script, do you have an example?
Supersonic112 has joined #nixos
Supersonic has quit [Ping timeout: 260 seconds]
Supersonic112 is now known as Supersonic
<cole-h>
ryantm: I think it would be something like: `defaultApp.x86_64-linux = "${self.nixosConfigurations.home1.config.system.build.toplevel}/bin/switch-to-configuration switch";`?
<cole-h>
Untested, but that's the perl switcher script
flythief has quit [Quit: Connection closed for inactivity]
<fzakaria>
I set it up with TLS from ACME too on Quassel.
tldr32 has joined #nixos
<jtojnar>
cole-h that issue is something different (though it was kind of hijacked), emoji shrunk after it was opened
<ryantm>
cole-h: here we go this works: defaultApp.x86_64-linux = nixpkgs.legacyPackages.x86_64-linux.writeScriptBin "switch" "${self.nixosConfigurations.home1.config.system.build.toplevel}/bin/switch-to-configuration switch";
<samueldr>
jtojnar: shrunk after it was opened?
tobiasBora0 has quit [Ping timeout: 272 seconds]
<jtojnar>
samueldr we had working emoji before the fontconfig changes
<jtojnar>
merged in july
<samueldr>
yes
<jtojnar>
the issue was opened in may
<samueldr>
yes, I assumed the issue shouldn't have been hijacked
waleee-cl has quit [Quit: Connection closed for inactivity]
<jtojnar>
I am investigating it now
<samueldr>
OH, they started shrinking, after the point in time _the issue_ was opened
<samueldr>
I thought you meant emojis shrank after the file was opened or something like that
Maxdamantus has quit [Ping timeout: 265 seconds]
orivej has quit [Ping timeout: 240 seconds]
Maxdamantus has joined #nixos
veleiro has joined #nixos
hke has quit [Read error: Connection reset by peer]
<cole-h>
Isn't the answer just "update the package and config accordingly to avoid the backwards-incompatible change"?
Muukaayuu has quit [Quit: WeeChat 2.9]
orivej has quit [Quit: orivej]
ubergum has joined #nixos
orivej has joined #nixos
spudly- has joined #nixos
spudly- has joined #nixos
spudly- has quit [Changing host]
orivej has quit [Ping timeout: 240 seconds]
ubergum has quit [Remote host closed the connection]
spudly has quit [Ping timeout: 264 seconds]
spudly- is now known as spudly
orivej has joined #nixos
<jtojnar>
cole-h the issue he refers to occurs when mixing channels or using Linux distro with incompatible nixpkgs
<cole-h>
Ah, right.
<jtojnar>
but I wonder if it even was a config issue in the first place
<jtojnar>
the reason this mostrosity was introduced
nixos_newbie has joined #nixos
<jtojnar>
it might have just been a cache incompatibility which can be solved more easily
iqubic has joined #nixos
domogled has quit [Quit: domogled]
<jtojnar>
need to find a nixpkgs revision that still contains some gui software linked against fc 2.10
<jtojnar>
ideally in biinary cache
<jtojnar>
but too lazy for that
<iqubic>
So, I'm currently in the process of switching from an old laptop that runs NixOS to a new laptop. I'm going to wiping Window 10 off the new laptop and installing NixOS. What's the best method for migrating all of my configuration data, also all of my personal data too?
<iqubic>
I think rsync might work, but I haven't worked with SSH ever.
<pjt_tmp>
If you're anything like me all you need to worry about is /etc and ~. Be sure if that works for you though.
<iqubic>
I'm also going to need a refresher on how to set up ZFS. I have ZFS on my old laptop, but I don't recall how I set it up.
<pjt_tmp>
...or /etc and ~ without ~/.cache
<iqubic>
pjt_tmp: yeah. I think /etc and ~, but not ~/.cache will work for me.
sputny has joined #nixos
<iqubic>
But I have no idea how to transfer the data over.
<pjt_tmp>
iqubic: also a few other dot folders prolly. Wish I remembered how to display them >_<
<iqubic>
And actually, I might need to generate a new hardware-configuration.nix
<iqubic>
But the main issue here is getting zfs setup, and transferring files with rsync/scp
<energizer>
iqubic: i would use a usb drive
<iqubic>
Why would you use that over rsync?
<energizer>
faster
mariatsji has joined #nixos
<iqubic>
energizer: I actually want to use this as an excuse to learn how SSH works.
<pjt_tmp>
energizer: unless ya got usb 1.0 but 5ghz wifi :L
<iqubic>
I'm willing to wait as long as it takes for the data to transfer.
<pjt_tmp>
important consideration: how much you care about metadata?
<energizer>
iqubic: you'll also have to learn about how to make computers able to see each other over a network (local or internet)
<energizer>
otherwise they wont be able to make an ssh connection
<energizer>
that can be fiddly and have security implications for your systems
<iqubic>
energizer: Yeah. I know. I want to do this.
<energizer>
do you control the wifi network?
<energizer>
man i'm curious about your background. you use haskell and nixos, but haven't used ssh before, that's an unusual combination i think
<pjt_tmp>
if not, a good portable mini router is like ~$30 usd
<nixos_newbie>
hey all, maybe some can shed a light on some problem that I can't fix. I am trying to use zathura 0.4.6 instead of zathura 0.4.5, which is the version on the unstable channel. I've made a simple package override before, but the way that zathura is organized is confusing me
<nixos_newbie>
the version I want to override is defined in /core/default.nix
<iqubic>
energizer: My background is just that I'm a self-taught teenager that has interests in technology.
<iqubic>
And yes, I do control the wifi network here.
<energizer>
ah, teenager, that makes sense :)
<nixos_newbie>
I tried to add the following on my configuration.nix nixpkgs.overlays = [ (self: super: { zathura = super.zathura.overrideAttrs(_: { zathura_core.version = "0.4.6"; }); }) ];
<iqubic>
I'm 19.
<nixos_newbie>
I get the error: cannot coerce a set to a string, at /nix/store/j0mxbhf15s4fkm76k7kp71ra8yiwbb0l-nixos-20.09pre239318.c59ea8b8a0e/nixos/pkgs/build-support/trivial-builders.nix:7:7 on rebuild
<{^_^}>
[nixpkgs] @danieldk opened pull request #96996 → broot: 0.20.3 -> 1.0.0, fix Darwin build → https://git.io/JUOBJ
<energizer>
iqubic: turns out "which computers can see each other over my network" is a rapidly evolving right now due to things like wireguard replacing old tech
<iqubic>
And since I'm so much more technologically literate than my father, whom I live with, my father gave me complete control over the network. For better, or for worse.
<energizer>
and zerotier, tailscale, etc
<pjt_tmp>
nixos_newbie: just a tip if you wanna see that offline, something that might work: ls ~/.nix-defexpr/channels_root/nixos-unstable/nixpkgs/pkgs/applications[etc].
<energizer>
meanwhile there's new stuff going on with ssh (gravitational.com)
<pjt_tmp>
much easier w/ tab completion
<iqubic>
Right. So how would I go about seeing if my two machines can see eacho other?
ericsagnes has quit [Ping timeout: 264 seconds]
<energizer>
iqubic: on one of them run `ip addr`
<energizer>
see what the *local* ip address is for that machine
<pjt_tmp>
nixos_newbie: and yeah, that layout is weird
<energizer>
local meaning on your home network, not the internet
<pjt_tmp>
energizer: iqubic: could also learn nmap here. that's a nice tool to shove in the ol' back pocket
<energizer>
good idea
<pjt_tmp>
nmap -A 192.168.1.0/24
ericsagnes has joined #nixos
sangoma has joined #nixos
<pjt_tmp>
or something
<energizer>
the computation i was waiting for just finished, so i'll hand the baton over to whoever wants it
<pjt_tmp>
nixos_newbie: no, it opens the file in the store itself, which can't/shouldn't be modified manually. It's a weird command to have considering that.
<iqubic>
So, what should I pass into nmap?
<nixos_newbie>
yeah, just notices it was on read only, but a handy command to have
<pjt_tmp>
nixos_newbie: typing nix[space][tab] will show you what all the subcommands are
<nixos_newbie>
the first line of the derivation is { symlinkJoin, lib, makeWrapper, zathura_core, file, plugins ? [] }
<{^_^}>
[nixpkgs] @bzizou opened pull request #96998 → hp2p: init at 3.3 → https://git.io/JUOBF
<iqubic>
pjt_tmp: What nmap command should I run?
<pjt_tmp>
nixos_newbie: the reason for the whole layout, roughly, is for modularity. Hence the files for postscript/djvu etc. I'd have to stare at it for awhile to know exactly how though. I have a diff solution that I can explain in a sec
<pjt_tmp>
iqubic: do you know the subnet?
ubergum has joined #nixos
<iqubic>
I'm not sure. I assume looking at the output of `ip addr` can tell me, but I'm not sure how to parse that.
<pjt_tmp>
iqubic: so 1 is loopback. That one can only talk to itself (oversimplifiying), so it's not that one you want.
<pjt_tmp>
2 looks like a wired connection, and 3 a wireless one
<iqubic>
I want the wlp1... thing, as I want to connect over wifi.
<iqubic>
That's correct.
noudle has joined #nixos
<pjt_tmp>
also, the ifconfig is good here too. A bit nicer looking
<pjt_tmp>
* ifconfig command
<iqubic>
What am I looking for?
cole-h has quit [Quit: Goodbye]
<pjt_tmp>
in ifconfig, there are lines that start with 'RX packets' 'TX packets'
justanotheruser has quit [Ping timeout: 265 seconds]
<iqubic>
Correct.
<pjt_tmp>
and an amount indicating bytes. If the wired one was way higher you'd know your wireless connection was not being prefered.
<iqubic>
I'm only connected via wireless. The values for the wired connection are all zeros
<pjt_tmp>
just good to know
<iqubic>
*zeroes
<pjt_tmp>
iqubic: okay, what's the address? 192.something?
<iqubic>
Not sure.
<iqubic>
termbin.com/pyu01
<pjt_tmp>
it's the line that starts with 'inet'
<iqubic>
So, 10.0.0.10?
<iqubic>
Seems weird.
<pjt_tmp>
seems like it then yeah. and indeed a bit weird. 192.168.8.x or 172.20.0.x are more common
<pjt_tmp>
but anyway
<pjt_tmp>
10.0.0.0/24 should be your subnet, I think
<pjt_tmp>
that covers 10.0.0.[0-255]
<iqubic>
I guess this is an artifact of using a comcast wireless router.
<pjt_tmp>
could be I guess? hm
<pjt_tmp>
so doing nmap 10.0.0.0/24 would be the simplest way to scan. Add -v to be sure it's starting at .0 and going up.
xd1le has quit [Ping timeout: 240 seconds]
<pjt_tmp>
adding -T4 will make it faster though older devices may not like that.
<pjt_tmp>
same for -T5 but more so
<pjt_tmp>
-Pn will give more info, -A will give even more
nixos_newbie has quit [Ping timeout: 245 seconds]
alp has joined #nixos
wfranzini has joined #nixos
wfranzini has quit [Client Quit]
<iqubic>
So, what sort of output do I want from nmap?
<pjt_tmp>
iqubic: if you recognize the hostnames of results, then you have an easy way of identifying devices.
<pjt_tmp>
iqubic: just try the version without flags first.
<iqubic>
See, that's just it. I've got several devices that I know are turned on and connected to the router, but I can't see any of them in the nmap output.
<iqubic>
I tried with the -v flag.
<energizer>
some routers come configured to not let devices see each other by default
<iqubic>
Weird.
mananamenos has joined #nixos
<energizer>
security
<pjt_tmp>
iqubic: does -v say it's scanning 256 addrs?
<iqubic>
Yes.
<pjt_tmp>
hm
<pjt_tmp>
try with -A. It's a bit more comprehensive
<pjt_tmp>
also, this whole approach is most usefull when not all devices have a monitor
<iqubic>
How long is this going to take?
sascha has joined #nixos
<pjt_tmp>
hit enter to see an eta
<iqubic>
using -v took like 30 seconds.
sascha is now known as Guest64249
<iqubic>
Oh, and -A is 95% done.
<iqubic>
The last 5% is taking so long.
nature has quit [Ping timeout: 258 seconds]
<iqubic>
Well, It's at about 100% and it's still got 5 seconds to go.
<iqubic>
Ah. Now we have progress. NSE Timing is happening now.
<pjt_tmp>
hm. The router might be throttling the scan possibly.
<pjt_tmp>
you have to see it's settings to know
<iqubic>
Weird.
<pjt_tmp>
also, one really nice unrelated command is tldr
maxdevjs has joined #nixos
<iqubic>
Isn't that like `man` but more short?
<pjt_tmp>
shows some simple use cases for certain commands. tldr nmap is a nice example
<pjt_tmp>
yep
<pjt_tmp>
nice since not all commands bother to include examples
<iqubic>
nmap is done.
<iqubic>
nmap 10.0.0.0/24 -A
<iqubic>
No idea what I'm looking at.
<pjt_tmp>
each result should say something like 'scan results for [hostname] (ip address) at the top
knerten2 has joined #nixos
<iqubic>
It doesn't though.
<pjt_tmp>
is it an error or something?
<iqubic>
No.
<iqubic>
Nmap scan report for 10.0.0.179 Host is up (0.031s latency).
<iqubic>
And then it just goes on to list other stuff.
<pjt_tmp>
I guess that device doesn't have a hostname(?)
<pjt_tmp>
not sure
<iqubic>
That particular device is a Brother Printer.
mananamenos_ has joined #nixos
knerten1 has quit [Ping timeout: 256 seconds]
<iqubic>
But the issue is that my other laptop, and my other two machines that I know are connected to this wifi router aren't being found.
<iqubic>
In fact, none of the results have any hostname associated with them. Just ip addresses.
<pjt_tmp>
so all the devices you see aren't them? Because that's also odd. Even my most concealed devices still say "Host is up (x.xx latency)\nAll 100 scanned ports on _ are closed"
mananamenos has quit [Ping timeout: 258 seconds]
<iqubic>
Yeah, I'm still getting that message.
mallox has joined #nixos
<pjt_tmp>
but with no hostnames
<iqubic>
Nope.
<iqubic>
Not of that.
<iqubic>
*none of them.
civodul has joined #nixos
<pjt_tmp>
at this point I'd say go into the router's settings. If all of them are like that (and all of them *do* have hostnames), the router is the one thing they have in common.
cfricke has joined #nixos
<iqubic>
Yeah. Alright. I just went to my router's settings, which can be found by pointing a web browswer to 10.0.0.1
<iqubic>
And I can get the ip address that way.
<pjt_tmp>
yeah, prod aroind in there. do keep nmap handy though, it can typically give much more useful info than here
<iqubic>
And I can go on my old laptop and run `ping 10.0.0.10` which is the ip of my new laptop. That responds to pings in the same way as if I had pinged google.com
<iqubic>
Or if I pinged 127.0.0.1 (local host)
<{^_^}>
[nixpkgs] @danieldk opened pull request #97000 → stdenv: unset SDKROOT on Darwin → https://git.io/JUO09
mallox_ has joined #nixos
<iqubic>
Lets just try connecting with SSH.
<iqubic>
I tried running `ssh 10.0.0.10`. Recieved: ssh: connect to host 10.0.0.10 port 22: Connection refused
alp has joined #nixos
mallox has quit [Ping timeout: 256 seconds]
<energizer>
systemctl status sshd
<pjt_tmp>
iqubic: either that's the router at it again, or services.ssh.{etc} isn't set and so sshd isn't running on the other device
<pjt_tmp>
I gotta sleep tho. ##hardware might be of help with the router, ##linux for other things
<iqubic>
Yeah. I have no idea if services.ssh.{etc} is set up by default on the NixOS graphical installer. LOL.
<pjt_tmp>
unlikely
<iqubic>
So, how can I set up sshd?
<iqubic>
Is that possible to do?
<pjt_tmp>
if you have access to said device via kbd/monitor or something else sure.
<pjt_tmp>
also could make a custom iso
<pjt_tmp>
with it enabled by default.
<pjt_tmp>
some instructions on nixos.wiki
<{^_^}>
[nixpkgs] @vbgl opened pull request #97001 → coqPackages.paramcoq: init at 1.1.2 for Coq 8.12 → https://git.io/JUOEI
<iqubic>
It's a laptop. I have access to the machine's kbd and monitor
<iqubic>
So, how do I run sshd?
<iqubic>
Is there just some command that I can run?
<energizer>
enable it in configuration.nix
<pjt_tmp>
man configuration.nix, /-search services.ssh
<pjt_tmp>
or any old tutorial online to learn how to do it the imperative way
<energizer>
ew :)
<pjt_tmp>
hey, only gotta do it once most cases
<pjt_tmp>
I do gotta sleep tho. might be here again in ~20 hrs
<iqubic>
Well, that didn't work.
mmohammadi9812 has joined #nixos
o1lo01ol1o has joined #nixos
<iqubic>
Warning, do not know how to make this configuration bootable; please enable a boot loader.
tsrt^ has quit []
<iqubic>
I just want to run sshd, damnit.
xd1le has joined #nixos
<iqubic>
At this point I'm heavily considering abandoning the SSH route and just using a second USB flash drive to transfer my data over.
<srk>
iqubic: installer does have ssh but not running by default, start it with systemctl start sshd and set password for root
<{^_^}>
srk: 8 hours, 17 minutes ago <bqv> can I interest you in a vervis flake? or maybe you'd like to upstream it to nixpkgs? since i recall you were interested in it too. One problem with both of those is that fetchDarcs doesn't seem to work properly with a patch rev, and there's no darcs libfetcher yet, so it'd have to be cached or updated regularly
<iqubic>
Thanks.
<pjt_tmp>
could do, yeah. You could still learn some of this later if wanted.
<srk>
bqv: sure, I wanted to play with vervis!
gthm has quit [Quit: leaving]
gthm has joined #nixos
<iqubic>
srk: Do I need to set a root password on the live system?
__monty__ has joined #nixos
<srk>
iqubic: yes
<bqv>
srk: to which? :p
<srk>
yes. :D
* bqv
!
<srk>
bqv: when I've tried to package vervis even darcs was broken on nixos .. :D
<bqv>
Lmao
<srk>
and darcs repo hosted on darcs .. fun :) got resolved quickly over IRC
<iqubic>
Alright, I've got ssh set up.
<bqv>
the problem I have is that half the vervis repos are darcs, and hosted on vervis, so I can't get an archive
<bqv>
Or I don't think I can
pjt_tmp has quit [Ping timeout: 240 seconds]
<iqubic>
Now, first things first, partioning + zfs.
<srk>
maybe ask on #darcs if there's something like git archive
<bqv>
mm
<iqubic>
So, the nixos installation guide says I should start with partitioning. How does that work if I'm going to be using zfs?
<srk>
depends if you want separate boot or everything zfs
<iqubic>
What I need is a UEFI boot partition, a swap partition, and a large partition for my zpool
<iqubic>
large partition = remainder of the hard drive.
<iqubic>
I think I'm just going to stick with zfs, actually.
<iqubic>
So I'd like to have a UEFI boot partition, a swap partition, and leave the rest to ZFS.
<turlando>
btw from my experience nix is one of the best linux os regarding zfs compatibility
<turlando>
the other one being ubuntu
<iqubic>
I agree.
<turlando>
debian, redhat, arch all have some issues or it is not supported officially
<srk>
how much RAM do you have iqubic?
<etu>
nixos (other than ubuntu) is the only distro I would trust to have zfs as root on
<iqubic>
16gigs.
<srk>
that's fine
<iqubic>
So I don't need swap?
<srk>
turlando: often legal issues :)
<turlando>
srk eh...
<srk>
pretty sure you do, as arc will consume bunch of your memory
<turlando>
s/consume/use: the arc space will be freed as soon as the kernel is under memory pressure
<srk>
sometimes
<turlando>
which is the same behavior as the normal cache in other filesystems
<iqubic>
So, I want a swap partition?
<iqubic>
How do I list the sub-volumes in a zpool?
<energizer>
iqubic: controversial
<energizer>
that sounds like a question for the zfs documentation :)
<iqubic>
Yeah. I've figured it out.
<turlando>
srk I haven't carefully checked on linux but I'm pretty sure I've never ran into memory issue on freebsd because of arc eating all the memory
<srk>
turlando: sometimes it would take me number of failed nix-build attempts to trigger arc freeing space (large evaluation requiring like 16G mem)
<srk>
was fine after I've set arc_max
<turlando>
mh, that's not good at all
<iqubic>
How big is a swap partition supposed to be?
<energizer>
some people use a rule of thumb like "twice the size of your ram" but yknow, whatever
<turlando>
there is not a go-to answer
<turlando>
if you want to use hibernation at least as much as your ram
<iqubic>
lsblk on my new machines is confusing me so much.
<iqubic>
Here's the output: termbin.com/p82nd
<iqubic>
I'm not sure which one of those is the laptop's hard drive.
<energizer>
probably neither
<energizer>
need to set up a partition table etc
<energizer>
see the nixos manual for instructions
<iqubic>
Yeah. I'm sure I do.
<energizer>
(note that doing this will erase the drive)
<iqubic>
I'm aware.
arjen-jonathan has joined #nixos
<iqubic>
I want to go through with that.
<iqubic>
So, how shall I set up the partition table?
<iqubic>
I've opened gparted, but all it can see is the 8GB flash drive that I'm running the live system from.
<energizer>
try gdisk
<iqubic>
gdisk wants me to enter a device filename, but I'm not sure which name to give.
page has quit [Ping timeout: 240 seconds]
<bqv>
srk: I think the smart thing is just to fix fetchdarcs tbh, cause its a really simple fix, I've just been avoiding it cause I vowed never to PR nixpkgs again
<energizer>
`sudo blkid` sometimes helps
<bqv>
It just needs --to-hash as an option
<bqv>
As well as --tag
page has joined #nixos
<iqubic>
energizer: sudo blkid gives me the same output: termbin.com/p82nd
<energizer>
no, that's lsblk
<energizer>
i said blkid
<iqubic>
termbin.com/pvhql
<iqubic>
No idea what's going on now.
<ar>
,locate bin ntpclient
<{^_^}>
Couldn't find in any packages
<ar>
,locate bin ntpdate
<{^_^}>
Found in packages: ntp
<energizer>
idk iqubic, it looks like for some reason your drives aren't detected
<energizer>
maybe there's a driver issue or missing kernel param or broken drive
<srk>
bqv: what's the flag it needs? darcs get --to-patch?
<bqv>
srk: yup
<energizer>
or maybe you just need to reboot
<bqv>
srk: --to-hash might be better
<bqv>
Either works
thc202 has joined #nixos
<srk>
ok. can you link the vervis flake?
<iqubic>
Perhaps I should go through the Windows 10 setup stuff first, and then try installing NixOS.
<bqv>
srk: http://rc.fron.io/overlays/vervis.nix is what I have right now, I can lift that into a flake, I just wanna be sure it makes sense (I.e. The fetchers work)
<energizer>
oh sometimes i need to fiddle with the firmware settings
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<iqubic>
What sort of firmware settings? Like in the bios?
<energizer>
yeah
<energizer>
but this stuff is hard for me to understand, someone else would be better than me
<bqv>
srk: the revs are commented out cause fetchdarcs treats them as tags, and they aren't, there are no tags
<srk>
bqv: cool, no need to convert right now, I'm not a flake user yet
<mananamenos_>
maybe someone could point to some easy docs about this
<bqv>
srk: I fixed that by changing a version, those are just packaging fixes (except the middle two which do a version upgrade of pandoc)
<srk>
I see
<srk>
mananamenos_: nix-shell -p pkgs.leiningen (to override version and src you need something bit more complex, like nix-shell -E 'with (import <nixpkgs> {}); pkgs.leiningen.overrideAttrs (old: { version = ...; src = ...; })'
<mananamenos_>
thank you srk, i will try that
<{^_^}>
[nixpkgs] @FRidh pushed 84 commits to staging-next: https://git.io/JUOg4
<{^_^}>
[nixpkgs] @andersk opened pull request #97002 → cryptsetup: Build with an lvm2 that’s built with udev → https://git.io/JUOgB
<mananamenos_>
srk, nix-shell -E 'with (import <nixpkgs> {}); pkgs.leiningen.overrideAttrs (old: { version = 2.8.1; })' this gives me an error error: attempt to call something which is not a function but a float, at (string)
<srk>
you also need to override src as it contains the hash of the original version
page_ has joined #nixos
page has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @doronbehar opened pull request #97003 → gst_all_1.gst-rtsp-server: fix pkgconfig includedir and libdir → https://git.io/JUOgS
<mananamenos_>
srk, src attribute is a fetchurl function call. There is url attribute which already must get the correct version as it embed with ${version} variable. So i have to change the sha256 only, right? How do find out its value of my new version?
<iqubic>
Screw this. I'm going to quickly run through the Windows set-up stuff once, and then try installing NixOS again.
<srk>
mananamenos_: you either supply some wrong hash and fix it after you get hash mismatch error or you can use nix-prefetch-url
<srk>
,tofu mananamenos_
<{^_^}>
mananamenos_: To get a sha256 hash of a new source, you can use the Trust On First Use model: use probably-wrong hash (for example: 0000000000000000000000000000000000000000000000000000), then replace it with the correct hash Nix expected. For inserting 52 0's in vim: <esc>52i0<esc>
<srk>
mananamenos_: not sure if you can override just sha256 of the fetchurl call, probably
yutyo[m] has quit [Remote host closed the connection]
<srk>
yeah, looks like it's possible with src = old.src.override { inherit version; sha256 = ".."; };
cmcaine[m] has quit [Quit: Idle for 30+ days]
l-as has quit [Quit: Idle for 30+ days]
aniketd[m] has left #nixos ["Kicked by @appservice-irc:matrix.org : Idle for 30+ days"]
clandest[m] has left #nixos ["Kicked by @appservice-irc:matrix.org : Idle for 30+ days"]
porphyrogenetos has quit [Quit: Idle for 30+ days]
hr[m] has quit [Quit: Idle for 30+ days]
kaiha[m] has quit [Quit: Idle for 30+ days]
<mananamenos_>
srk, so, i've just tried to pass the only some dummy sha256 as you mentioned, but I fetchurl requieres url param too. So now I've ended up with `nix-shell -E 'with (import <nixpkgs> {}); pkgs.leiningen.overrideAttrs (old: { version = "2.8.1"; src = fetchurl {url = "https://raw.github.com/technomancy/leiningen/${version}/bin/lein-pkg"; sha256 = "934839";};})'` which gives an error: undefined variable version. Why doesn't it see the
<srk>
mananamenos_: yeah, it's not part of the fetchurl { <scope> }, try with fetchurl { inherit version; ...
<iqubic>
What the bloody hell is going on here?
<iqubic>
Why can't linux see my hard drive?
lord| has quit [Ping timeout: 256 seconds]
<iqubic>
Like, Windows works fine. But I don't give a hoot about that.
<iqubic>
I'm trying to install NixOS on a Dell Latitude 3500.
lord| has joined #nixos
<srk>
missing kernel module?
<srk>
is that nvme ssd?
<iqubic>
Hmm. It might be.
<iqubic>
How do I check?
<srk>
specs :)
<srk>
try
<srk>
modprobe nvme
<srk>
dmesg
<srk>
as root, to see if it finds the drive
<mananamenos_>
srk, `inherit version` hasn't helped. Anyway I hardcoded the version in the url, the expression succedded and im in the nix-shell now. However, `lein version` shows me 2.9.1 instead of 2.8.1. So it sees my system's lein instead the one in the nix shell. Why?
redkahuna has joined #nixos
<iqubic>
modprobe nvme fails because it couldn't insert nvme, as that operation isn't permitted.
<redkahuna>
I have the same error with the other example graphiz
<redkahuna>
Any idea ? Thanks
<iqubic>
`modprobe nvme` returns nothing at all.
benedikt-h has joined #nixos
benedikt-h has quit [Client Quit]
<srk>
mananamenos_: hmhm, not sure. maybe it's better to add that expression to nixpkgs.overlays now as something like lein_2_8_1 = and use that
<srk>
iqubic: that's correct, it retuns exit code 0 :)
<iqubic>
Yeah.
<srk>
,overlays mananamenos_
<{^_^}>
srk: Did you mean overlay?
<{^_^}>
mananamenos_: Overlays look like `self: super: { foo = ...; }`. Use the self argument to get dependencies, super for overriding things and library functions (including callPackage). More info: https://nixos.org/nixpkgs/manual/#sec-overlays-install
benedikt-h has joined #nixos
<iqubic>
So how do see if this is nvme?
<srk>
check the end of the kernel logs with dmesg
<srk>
or with lsblk again
<srk>
maybe lshw could tell if its nvme
benedikt-h has quit [Client Quit]
boxscape has joined #nixos
benedikt-h has joined #nixos
<iqubic>
I just ran `lshw | grep -i nvme` and got nothing.
<iqubic>
tried searching for sata, also got nothing..
<mananamenos_>
srk, i've just uninstalled leiningen from my systemPackages in configuration.nix, got into nix shell again, and find out lein is not even in scope in the nix-shell.
<mananamenos_>
so, yeah..:/ will try to with overlay in configuration.nix then
<srk>
mananamenos_: probably due to nix-shell -E not doing what we want here, I guess it would require additional wrapper to put it to path
sputny1 has joined #nixos
<srk>
iqubic: and the other two? does it find the drive now?
<iqubic>
What other two?
<srk>
[11:15] < srk> or with lsblk again
<srk>
[11:15] < srk> check the end of the kernel logs with dmesg
hiro99 has joined #nixos
<iqubic>
Ah. lsblk still return nothing.
<iqubic>
Or rather, it only finds the usb stick I've booted from.
<mananamenos_>
srk, I finally copied the default.nix to my /etc/nixos/, change the version in that file, and called from systemPackages in configuration.nix with (callPackage ./lein-2.8.1.nix { })
<iqubic>
Alright... This is running nvme
<mananamenos_>
for now this works :) thank you very much for your help!
benedikt-h has left #nixos [#nixos]
<iqubic>
I can tell because the drive names aren't sdX... but nvmeX...
<iqubic>
I assume I can just carry on with a NixOS installation like normal, yeah?
<iqubic>
I know this that the nvme drive is the correct one to use because 1. It's newly added to the lsblk list, and b) the number and fs types of the partitions match that of a Windows installation.
sputny1 has quit [Quit: sputny1]
<srk>
mananamenos_: yw!
<srk>
iqubic: yup
<iqubic>
srk: Thanks for all the help.
<srk>
yw! :)
<iqubic>
The rest of this should just be following the installation guide.
<iqubic>
I've done plenty of NixOS installs in the past. It usually consists of copying over my current configuration.nix, updating the stateVersion, and generating a new hardware-configuration.nix
<iqubic>
I assume I can address this the physical hard drive by using /dev/nvme...
<iqubic>
Where nvme... is what's displayed in lsblk. Is that right?
<srk>
yup, nothing special about that, just different naming scheme
<iqubic>
Cool.
<srk>
for VMs you sometimes get dev/vda instead of dev/sda depending on IO subsys. config
sascha has joined #nixos
sascha is now known as Guest43221
<iqubic>
So, this laptop has a fingerprint scanner built into it. Unfortunately it requires propritary Dell software that you can only get if you have a laptop that comes with Dell's own spin of Ubuntu.
<iqubic>
So, I'm not going to try making that work.
Ramon has joined #nixos
<iqubic>
Like even if I wanted to make that work, I can't. I don't have access to Dell's propritary linux drivers for that.
benedikt-h has joined #nixos
<iqubic>
But that wasn't a feature I was ever planning on using, so I'm totally fine with that.
gxt has joined #nixos
redkahuna has quit [Ping timeout: 240 seconds]
Ramon has quit [Remote host closed the connection]
benedikt-h has left #nixos ["WeeChat 2.7.1"]
<sphalerite>
What's the current preferred way to to package go stuff for nixpkgs? I know vgo2nix is a thing, but it's not mentioned in the nixpkgs manual — is that only suitable for out-of-nixpkgs go packages?
<sphalerite>
(cc adisbladis who maintains vgo2nix)
<q3k[m]>
buildGoModule is the new standard for nixpkgs
SanchayanMaity has quit [Ping timeout: 264 seconds]
damjan has quit [Quit: bye]
Ramon has joined #nixos
redkahuna has joined #nixos
<Ramon>
Does anyone know how to get Nix to fetch a tarball from URL that required Kerberos Auth? I am trying to do ```nix-shell -I nixpkgs=<URL> ...``` but I kept getting 401 unauthorised. When I curl the URL myself in command line, it works fine
damjan has joined #nixos
spudly has quit [Remote host closed the connection]
redkahuna has quit [Ping timeout: 246 seconds]
ubergum has quit []
redkahuna has joined #nixos
<immae>
Can I make a derivation forget that it has some runtime dependency? (namely, haskellPackages.postgresql-simple-migration thinks it needs ghc at runtime - so it downloads a bunch of additional derivations - but it’s wrong)
redkahuna has quit [Ping timeout: 240 seconds]
SanchayanMaity has joined #nixos
Sanchayan has quit [Ping timeout: 240 seconds]
redkahuna has joined #nixos
alp has quit [Ping timeout: 244 seconds]
<infinisil>
immae: Currently on my phone, but there are some helpers for that, grep for eeeeeeeeee in nixpkgs
<immae>
Found it thanks!
spudly has joined #nixos
spudly has quit [Changing host]
spudly has joined #nixos
redkahuna has quit [Ping timeout: 260 seconds]
<infinisil>
immae: What's it called for the future?
<immae>
remove-references-to
<immae>
But grepping for eeeee is far easier than remembering the name :p
civodul has quit [Quit: ERC (IRC client for Emacs 26.3)]
<infinisil>
Hehe
<immae>
(In my case I only had to copy bin/migrate to another derivation though, the dependency comes from elements of lib/ which I don’t need)
Sanchayan has joined #nixos
SanchayanMaity has quit [Ping timeout: 264 seconds]
<GolDDranks>
Hi, if there are any people with review/merge rights, consider helping us getting this PR over the goal line so that 20.09 wouldn't have major bugs with Raspberry Pi 4! https://github.com/NixOS/nixpkgs/pull/91295
<{^_^}>
#91295 (by jbaum98, 10 weeks ago, open): Rewrite raspberrypi-builder.sh based on extlinux
<GolDDranks>
It's the only major one I've encountered thus far. (often preventing nixos-rebuild because the cleanup of /boot isn't done properly). It would be great to have merged!
<clever>
sephii: now insert that new and correct hash into the expr
mmohammadi98123 has joined #nixos
noudle has quit []
mmohammadi9812 has quit [Ping timeout: 246 seconds]
mmohammadi98123 is now known as mmohammadi9812
<sephii>
mmmh it's in a weird format (`sha256-XVNM8cEQeSxDvfXLtF3FQnKi8qZvyH/ePrQO4x9+rus=`). Anyway, I think adding `--fetch-submodules` to `nix-prefetch-git` got me the correct hash :) thanks!
dansho has quit [Remote host closed the connection]
dansho has joined #nixos
waleee-cl has joined #nixos
<{^_^}>
[nixpkgs] @trepetti opened pull request #97021 → svls: init at 0.1.17 → https://git.io/JUOHi
mmohammadi9812 has quit [Ping timeout: 240 seconds]
GolDDranks has quit [Remote host closed the connection]
noreferences has joined #nixos
fendor has quit [Read error: Connection reset by peer]
<Baughn>
Is there any way to install NixOS on a laptop where I cannot disable secure boot?
<Baughn>
(If I do, it'll erase the SSD.)
Darkmatter66_ has joined #nixos
noreferences has joined #nixos
noreferences has quit [Client Quit]
<clever>
Baughn: if the secureboot is doing its job properly, not really
<Baughn>
Well, it certainly refuses to boot the default iso.
<clever>
Baughn: you would need to make a backup of the ssd to another machine, then turn off secureboot
noreferences has joined #nixos
noreferences has quit [Client Quit]
Darkmatter66 has quit [Ping timeout: 258 seconds]
<Baughn>
Can't it be booted with efistub?
<Baughn>
Or... whatever the shim is called?
<clever>
Baughn: only if that shim is trusted by the secureboot
<Baughn>
Hmm. Well, I can't really backup the SSD without booting linux.
<Baughn>
Maybe ubuntu will work
<benedikt-h>
All right, I have figured it out. There is eclipses.buildEclipse and eclipses.eclipseWithPlugins. Those did the trick
orivej has quit [Ping timeout: 258 seconds]
lotharn has quit [Ping timeout: 246 seconds]
MtotheM has joined #nixos
dyaso has quit [Quit: Konversation terminated!]
<dbe2>
What's the standard way to create a write-enabled directory for a specific user? I have a service running as a "isSystemUser" user, and that service needs some writeable directories.
lsix has joined #nixos
sangoma has quit [Ping timeout: 265 seconds]
<cransom>
dbe2: an entry in `systemd.tmpfiles.rules` if the user isn't going to get a typical home directory
<cransom>
the service could also use some of the DynamicUser options which will have systemd create some stuff too, but it depends on your use case.
zupo has joined #nixos
zupo has quit [Ping timeout: 260 seconds]
flythief has joined #nixos
benedikt-h has left #nixos ["WeeChat 2.7.1"]
<{^_^}>
[nixos-homepage] @github-actions[bot] pushed commit from GitHub Actions to master « Update flake.lock and blogs.xml [ci skip] »: https://git.io/JUOdP
nixos_newbie has joined #nixos
<clever>
dbe2: i prefer just doing createHome = true; on that user
<{^_^}>
[nixpkgs] @DamienCassou opened pull request #97026 → emacs: Add missing patches to the Emacs derivations → https://git.io/JUOFT
erasmas has joined #nixos
kreyren has joined #nixos
Maxdamantus has quit [Ping timeout: 265 seconds]
Maxdamantus has joined #nixos
jumper149 has joined #nixos
kreyren has quit [Remote host closed the connection]
kreyren has joined #nixos
<zebrag>
I've installed NixOS on a laptop a year ago. It was easy and fine, but because I needed npm and I had issues with it at the time, I haven't used that machine since...
<zebrag>
First question: how is the project faring since?
<simpson>
NixOS is fine. npm's still pretty bad. I think that we added a new nodePackages version for the latest stable series.
<zebrag>
This time I intend to use npm the most naive way
<gchristensen>
is there a way to disable nscd for a single terminal, like an env variable glibc looks at or something
cosimone has joined #nixos
xcmw has joined #nixos
<{^_^}>
[nixos-homepage] @garbas pushed 3 commits to demos: https://git.io/JUObd
<{^_^}>
[nixos-homepage] @garbas opened pull request #513 → Converting examples into asciinema videos → https://git.io/JUObA
<domenkozar[m]>
yeah, looking into it
xd1le has quit [Remote host closed the connection]
<asheshambasta>
also; I seem to have run into an issue (https://github.com/cachix/cachix/issues/267); but I've deleted the nix file in my case; and it still doesn't seem to make cachix not download from that particular cache. As a result; I cannot even nixos-rebuild anymore; and I was wondering if this was related to the outage.
<lordcirth_>
I'm confused. Trying to run 'ld' says to install nixos.ccacheWrapper. nix-env -qa says it doesn't exist. But -qaA finds "ccache-links-wrapper-". What am I supposed to put in systemPackages to get ld?
davidv7 has joined #nixos
thonkpod_ has quit [Quit: WeeChat 2.7.1]
thonkpod has joined #nixos
asheshambasta has quit [Ping timeout: 260 seconds]
<{^_^}>
[nixpkgs] @timokau merged pull request #97006 → fpc: 3.0.4 -> 3.2.0, add support for aarch64-linux → https://git.io/JUOPT
<inkbottle>
https://nixos.org/manual/nixos/stable/release-notes.html#sec-release-19.09; "NetworkManager systemd unit was renamed"; I do not have "networkmanager" per se in my configuration.nix; what I do have is: `networking.hostName = "nixos";`, `networking.networkmanager.enable = true;`
<inkbottle>
not sure what I should do in this concern
<inkbottle>
I will probably end up posting my configuration.nix
<aanderse>
does anyone know what all the weird vim errors (warnings?) are with nixos unstable right now when i rebuild my system? `....... vim.1.gz: whatis parse for rview(1) failed`
mariatsji has quit [Remote host closed the connection]
mariatsji has joined #nixos
eoli3n_ has joined #nixos
maier has joined #nixos
lsix has quit [Quit: WeeChat 2.9]
maier has quit [Ping timeout: 246 seconds]
cole-h has joined #nixos
mariatsji has quit [Remote host closed the connection]
<inkbottle>
It doesn't seem I really need to specify: `system.stateVersion` This seems outdated way of doing things?
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
xcmw has joined #nixos
xcmw has quit [Client Quit]
<inkbottle>
Is there a reasonable list of packages for KDE-Plasma to put in `environment.systemPackages = with pkgs; [`?
<clever>
inkbottle: you need to specify it if your using things where the state has changed, such as the ssh hostkey type, or the default postgres master user
<clever>
inkbottle: and youll only know the state has changed if you read the nixos release notes closely
<inkbottle>
clever: well, I did have quite a close look at them
<inkbottle>
Also I'm trying to simplify things before the upgrade
<clever>
inkbottle: thats what stateVersion is for, to minimize the breakage when you upgrade without reading the release notes
<clever>
so it continues to work as before
<clever>
it should just be left at the previous value
<inkbottle>
Do you mean that won't prevent installation of newer linux-kernel, and newer version of packages?
<clever>
inkbottle: the only case i can think of where it stops things from upgrading is the postgresql version
pjt_tmp has joined #nixos
<clever>
inkbottle: because the new version cant read the old db files, and upgrading would break postgresql hard
<inkbottle>
okay
<cole-h>
state version only affects NixOS modules, AFAIK
<V>
is there any good reason why modules don't each maintain a separate state version, and do so automatically
o1lo01ol_ has joined #nixos
<clever>
V: it has to be done from your configuration.nix, because you might be building for another machine (nixops)
<V>
like, what's stopping the postgres module from keeping a file in /var/lib/postgresql/version
<clever>
V: then you would have to ship 2 versions of the postgres binary to the end machine, and somehow swap which one is in the global $PATH based on the file
o1lo01ol1o has quit [Read error: Connection reset by peer]
<V>
no? can you not have it upgrade
<clever>
V: upgrading involves dumping the entire db to .sql format while the old daemon runs, then switching the daemon out, and re-importing the entire db
<V>
ugh
<clever>
yes
<clever>
so you need to be aware of those changes, and manually migrate the state before you change the stateVersion
<clever>
inkbottle: eval time warnings from the nix code arent held in the `nix-store -l` logs
cosimone has quit [Read error: Connection reset by peer]
<ivunchata>
Hey guys, I'm new to NixOS and have an issue with my NIC (probably hardware related), which is solved if I run ethtool on it, disabling TSO. My question is - how do I properly run a command after an interface comes up or after the system boots?
<inkbottle>
I don't really know what to do about it
<cole-h>
Ignore it. There's nothing you can do about it.
<cole-h>
Or rather
<colemickens>
Can't... they fix it?
<cole-h>
Do what it suggests
<colemickens>
I thought I did when I got that warning
<cole-h>
I thought it was complaining about loaOf in general
<inkbottle>
it is
<cole-h>
inkbottle: Follow the suggestion: ("Do ..... instead of ..... ")
<inkbottle>
and I don't understand the suggestions
<inkbottle>
are the "..." a placeholder?
<cole-h>
Instead of having a list of attrs like `[ { name = "root"; othersetting = "awsdf"; } ]`, change it to be an attr set like `{ root = { othersetting = "awsdf"; }; }`
<inkbottle>
no, that wouldn't do
<inkbottle>
reading
<cole-h>
Please paste your config if you want more specific help.
<superherointj>
I have seen many 'result' folders (as symlink) appear in several places in my system. Like: in `~/` and `/etc/nixos/`. Can I delete it and not worry about it? I am not sure what causes it to appear.
<Extends>
it appears when you try to build something manually (like a derivation for instance) so you can usually delete it if it is useless
inkbottle has quit [Ping timeout: 240 seconds]
<V>
yeah, it's just a symlink into the store
<V>
it will remove a gc root though
<V>
i.e. whatever it pointed can then be gc'd
<superherointj>
Should I garbage collect after removing it?
<V>
well, if you don't care about any of those results, gcing would free some space up
<superherointj>
Thanks.
<V>
if you have stuff in /etc/nixos it sounds like you did nix-build on nixos
<V>
so you probably have some fairly hefty space savings coming your way
<pjt_tmp>
hi! I'm building a thing with a painfully large set of deps, and want to pin the longest-taking ones to the store. How can I log that information?
<pjt_tmp>
Right now I just have something like "watch -n 10 nix-top -1 >> log.txt"
<pjt_tmp>
and will later do some sed magic to find the duration of each thing
<pjt_tmp>
there's probably a way to make that less painful, right?
<energizer>
pjt_tmp: there are some Nix github issues about logging for this purpose, idk if they made any progress
<energizer>
adisbladis: do i need to patch poetry2nix's pip to include poetry? https://bpa.st/FIGQ
mariatsji has quit [Ping timeout: 260 seconds]
<pjt_tmp>
energizer: that would be nice yeah. I don't suppose you know of some bash that'd make this less awkward?
fresheyeball has quit [Quit: WeeChat 2.7.1]
pkern has joined #nixos
<energizer>
,locate bin/loop
<{^_^}>
Found in packages: loop
<energizer>
pjt_tmp: ^
<pjt_tmp>
hm
<pjt_tmp>
one install later...
<pjt_tmp>
yeah that could help
<pjt_tmp>
oh, just thought of something! nix-top --json would be a nice thing.
<pjt_tmp>
time to file a feature request
<energizer>
structured logs should be built into nix, not a third party polling tool
<dminuoso>
What would that structured logs even contain, energizer?
<pkern>
And what I actually want to do is run a Terraria server with the latest build, but it also looks like there's no way to specify the package in the service defintion.
Extends has quit [Remote host closed the connection]
<dminuoso>
energizer: I have my doubts logging on the nix implementation itself is sensible.
noonien has quit [Quit: Connection closed for inactivity]
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #97047 → stdenv/adapters.nix: add isStatic when static stdenv adapter is used → https://git.io/JU3mN
<energizer>
dminuoso: daemons are the right place to log. any other place will miss short-lived events.
<cole-h>
pjt_tmp: You could try `nix-build-uncached`? Might help.
<pjt_tmp>
dminuoso: some nice json with a well-defined schema would be good. Easy to play with via jq.
<grfn>
I am led to believe that there exists somewhere a "what package has this file" search engine
<grfn>
does anyone know what that is?
<chkno>
toastyavocados: So the minimal CD doesn't have NetworkManager on by default, but the graphical CD does: nix-instantiate --eval -E 'with import <nixpkgs> {}; builtins.hasAttr "NetworkManager" (pkgs.nixos { imports = [<nixpkgs/nixos/modules/installer/cd-dvd/installation-cd-base.nix>];}).config.systemd.services' => false. But replace installation-cd-base.nix with installation-cd-graphical-base.nix & you get
<grfn>
hodapp: a bit, but mostly maintainer of tooling for machine learning engineers
<grfn>
hmmmm
<grfn>
what is this package expecting then
<hodapp>
grfn: I'm sorry to hear that.
<grfn>
hodapp: haha yeah. it's a living
<hodapp>
especially if you must touch TensorFlow.
<grfn>
I must.
<V>
UsePythonExtensions.cmake:25 gives find_package(PythonExtensions) under 'example usage'
<hodapp>
ow.
<V>
so
<grfn>
hmm V sshould I just patch this package to add `include(${scikit-build}/lib/...)` at the top of CMakeLists?
<grfn>
or is there a better way to do that
gustavderdrache has quit [Quit: Leaving.]
sbetrium has joined #nixos
<V>
Not sure, I'm not a cmake/python packaging expert
<grfn>
hah.
<V>
but if you grep through nixpkgs for find_package you do find a number of places where that kind of thing is patched
<grfn>
I'm almost one of those, I've just never run into scikit-build before
<grfn>
ah ok cool
<grfn>
I'll give that a shot
<V>
I would like to think there is a better way
<grfn>
I feel like I have that thought about 1 in 5 times I package something for nixpkgs
<V>
I have that thought over half the times I do anything in nixpkgs
arjen-jonathan has joined #nixos
<chkno>
How do I use builtins.filterSource (or lib.cleanSourceWith) inside a flake? Everything is already store paths during flake evaluation, and filterSource refuses to operate on store paths ("cannot refer to other paths"). builtins.unsafeDiscardStringContext makes it work. How scared of the "unsafe" should I be?
<grfn>
blergh, that doesn't work
<grfn>
still no FindPythonExtensions.cmake
<grfn>
oh I thin maybe I need to add it to CMAKE_MODULE_PATH
<zecnate>
Hello. I am interested in using nix as a reproducible or deterministic build system within Docker. The idea is that for users who already have docker (but not nix), they can produce bitwise identical binaries as any other users for statically compiled linux binaries. Is this a plausible use case?
eoli3n_ has quit [Ping timeout: 256 seconds]
<immae>
zecnate: if I understand correctly your question, using nix inside of docker requires --priviledged, which is a dangerous flag
<zecnate>
I'm defining a .nix file beginning with `{ pkgs ? import (fetchTarball …) {} }`... How can I tell nix to perform the fetch/unpacking that `fetchTarball` does without installing a derivation?
alp has joined #nixos
spudly has quit [Ping timeout: 256 seconds]
malook has quit [Quit: malook]
Church- has quit [Ping timeout: 258 seconds]
Church- has joined #nixos
noudle has quit []
<chkno>
zecnate: Nix only ever writes stuff to disk by building derivations. If you don't want it to show up in the main /nix/store, you can set NIX_STORE_DIR (and possibly other vars mentioned at the bottom of `man nix-store`) to have it write somewhere else.
<zecnate>
Because I'm running nix inside a nixos docker image, I want to fetch/build all dependencies of my project at one shell command, then after that build my project. I'm exploring doing a project build as a derivation.
<grfn>
zecnate: nix-build -A env is what I do
<zecnate>
(Splitting out those two steps is a caching convenience for devs who encounter build errors.)
<zecnate>
@grfn Thanks. I don't yet understand what `-A` means and what `env` is.
<grfn>
-A is short for "attribute"
<zecnate>
If you can't tell, I'm just now learning nix. ;-)
<grfn>
haha no worries
<colemickens>
(you could build the container itself with Nix, too, btw, maybe down the line...)
<zecnate>
The next issue I run into is that my derivation attempts to set `src` to a path to a directory, which is the source of my project I want to build. So the nix expr lives in the same repo with the source. It seems like `src` always expects to fetch and unpack a tarball, though.
<V>
zecnate: are you using quotes around the path
<grfn>
zecnate: doUnpack = false;
<grfn>
or
<grfn>
dontUnpack = true;
<grfn>
?
<grfn>
one of those
Church- has quit [Quit: WeeChat info:version]
<grfn>
or unpackPhase = "true"
<grfn>
whatever works
<chkno>
zecnate: A common idiom when the .nix file lives inside its own project's source is to use src = lib.cleanSource ./.;
zecnate has quit [Remote host closed the connection]
cosimone_ has joined #nixos
cosimone has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @dali99 opened pull request #97054 → vsci: init at 7.0.12 → https://git.io/JU3Zn
<{^_^}>
[nixpkgs] @tiagolobocastro opened pull request #97055 → p3x-onenote: init at 2020.10.111 → https://git.io/JU3ZW
alexherbo2 has quit [Ping timeout: 260 seconds]
cosimone_ has quit [Client Quit]
julm has joined #nixos
cosimone has joined #nixos
kreyren_ has quit [Ping timeout: 240 seconds]
xcmw has joined #nixos
Rusty1 has joined #nixos
ntwaaci^ has quit []
tsrt^ has joined #nixos
<bqv>
,locate lspci
<bqv>
,locate lsblk
<bqv>
,locate lsmod
<{^_^}>
Found in packages: toybox, busybox, pciutils
<{^_^}>
Found in packages: utillinux, libuuid.bin, zfsUnstable, utillinux.bin
<{^_^}>
Found in packages: kmod, toybox, busybox
iqubic has joined #nixos
<iqubic>
So, I'm trying to partition my SSD for a new NixOS installation. How shall I go about doing this? I want a uefi boot partition, a 32G swap and the rest should be allocated to zfs.
<jlv>
Anyone know a guide for setting up Kerberos in NixOS?
<jlv>
Although. I think I've had situations where I still got an alignment notice even while using that flag. It's been a while since I did a fresh install.