<clever>
pie__: nearly all programs just obey $HOME
schjetne has joined #nixos
<clever>
the few that dont, will look it up in /etc/passwd
<clever>
you can abuse the 1st fact to just change $HOME and have a 2nd set of config, without a 2nd user
Matthieu` has joined #nixos
arahael2 has quit [Ping timeout: 272 seconds]
Anton-Latukha has quit [Quit: Leaving.]
arahael2 has joined #nixos
schjetne has quit [Ping timeout: 248 seconds]
Matthieu` has quit [Ping timeout: 268 seconds]
arahael2 has quit [Excess Flood]
arahael2 has joined #nixos
<{^_^}>
[nixpkgs] @Ma27 opened pull request #62914 → doc/nixos-rebuild(8): add Nix options to summary → https://git.io/fjgUe
<pie__>
clever, ye
arahael2 has quit [Ping timeout: 246 seconds]
fusion809 has joined #nixos
dftxbs3e has quit [Quit: Leaving.]
Ariakenom_ has quit [Quit: Leaving]
o1lo01ol1o has joined #nixos
arahael2 has joined #nixos
o1lo01ol1o has quit [Ping timeout: 248 seconds]
Matthieu` has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/fjgU3
<zfnmxt>
What's the best way to get an xorg.conf that's more "custom" than configuration.nix allows for?
<pie__>
zfnmxt, sometimes theres osmetihng like an extraArgs or something that lets you put literal stuff in it
<zfnmxt>
Yeah, that doesn't work because I specify two drivers and it repeats the extra stuff for both...but they need to differ.
<pie__>
zfnmxt, i know it doesnt help much immediately, but have you tried looking at the source code to see if you can get it to do what you want?
clacke_movim has left #nixos [#nixos]
<zfnmxt>
Just started started doing that, but I'm new to nix and figured it'd be prudent to maybe get a nudge in the right direction.
<pie__>
yeah sure, just sayinm, sinc ei havent looked at that stuff
<pie__>
worst case file an issue if you think its something that should/could be handled reasonably, and you could look into bypassin the fancy configuratoin stuff and just generate one manually - though i dont know how that will interact with the rest of the system
<pie__>
maybe someone else has better ideas
<zfnmxt>
It's not clear to me how to modify an existing package to my needs without ending up with a whole different fork of nixpkgs.
clacke_movim has joined #nixos
<zfnmxt>
I want to just stay on the nixos-unstable channel...but, with the xserver package also modified :D
<zfnmxt>
Also...would it be really bad if I just modifed the generated xorg.conf in the store >.>?
psy3497 has joined #nixos
silver has quit [Read error: Connection reset by peer]
<zfnmxt>
clever: Oh, huh. Somehow I missed config and only saw extraConfig (which wasn't working).
<zfnmxt>
Thanks :D
<clever>
zfnmxt: youll also want to look at the config nixos generates after playing with those options
<clever>
zfnmxt: and there is things to insert things into the generated modules and display sections and stuff
<zfnmxt>
I was, that's how I figured it out was super broken :(
<zfnmxt>
I figured I'd just take the gneerated config that's broken, fix it, and then put that in for `xserver.config`
<clever>
but then it wont regenerate properly
<clever>
zfnmxt: what part of it is broken and needs to be changed
<zfnmxt>
Well, I'm trying to get nouveau + PRIME to work.
<zfnmxt>
Don't really know what I'm doing tbh. I'm just trying to get an external monitor to work, but my stupid HDMI port is only wired to the dedicated GPU =/
grubles has left #nixos [#nixos]
<zfnmxt>
It sorta worked under bumblebee with intel-virtual-output, but it was really laggy.
alex``` has quit [Quit: WeeChat 2.4]
Fzer0 has quit [Quit: Leaving]
arahael2 has quit [Ping timeout: 268 seconds]
Henson has joined #nixos
Glider_IRC_ has quit [Remote host closed the connection]
<Gopal[m]>
I see two different firefox dev edition packages. One of them is `firefox-devedition-bin` and one of them is `firefox-devedition-unwrapped`. What does _unwrapped_ mean?
<Gopal[m]>
I see two different firefox dev edition packages. One of them is `firefox-devedition-bin` and one of them is `firefox-devedition-bin-unwrapped`. What does _unwrapped_ mean?
<samueldr>
that would be added to the bootloader, e.g. using `e` on systemd-boot or grub to change settings
<pie_>
im going to have to actually remember these at some point, clever keeps telling me about that
<iqubic>
So I'd love to use my laptop as an external display so I can try setting up my Raspberry Pi before I place an order for a screen to use as a permenant display for this Raspberry Pi. But I'm not sure if my laptop is set-up in that way or not.
mac10688 has joined #nixos
<pie_>
samueldr, but it doesnt crash, boot continues, so iwll i be able to see anzthing
<samueldr>
it debug1devices will cause it to stop to investigate
<pie_>
samueldr, ok
arahael2 has quit [Ping timeout: 272 seconds]
<samueldr>
iqubic: most likely it isn't; if you shared the exact model number it should be trivial to confirm
<jackdk>
iqubic: it's an extra "u" key, to make it easier to write "colour", "flavour", etc.
<samueldr>
US has the ANSI layout, with a slim wide boi enter, while most everywhere in the world will use an ISO layout
<iqubic>
Well, my USA keyboard is an ANSI.
<pie_>
samueldr, funnily enough i think the error is from exactly line 220 beccause its right after the lvm message and right before the if then fail
<iqubic>
What is the extra key that is put on ISO keyboards?
<pie_>
so why is lvm still looking for sdb
<samueldr>
iqubic: you'd have to compare the software layouts, it depends on the language
<iqubic>
Next question: Is Nixos worth putting onto my Raspberry Pi 3?
<samueldr>
and not only language, but the specific locale (country's) layout
<samueldr>
iqubic: is NixOS worth puttin on a computer?
<samueldr>
though, for the raspberry pi, it all depends on what you intend to do with it, and your skill level
<samueldr>
if you intend to use the raspberry pi with external hardware made for raspberry pi, it will be harder to use with NixOS due to how the raspberry pi foundation implement their distros
marusich has joined #nixos
<samueldr>
(they build a 32 bit distro even on 64 bit hardware, NixOS builds a 64 bit distro)
<iqubic>
I have a fairly high skill level with Linux, but I have no idea what I'm planning on doing with my Pi.
<samueldr>
also, the raspberry pi foundation has a kernel fork which diverges a bit from the mainline kernel, but that's less of an issue as it can be used with NixOS if needed
<samueldr>
at worst, you can switch the SD cards around to test and play around
<iqubic>
I'll try that out. Thank you so much.
arahael2 has quit [Ping timeout: 248 seconds]
<samueldr>
pie_: add a sleep to `boot.initrd.preLVMCommands` just a hunch, unlikely to help
<samueldr>
especially since udevadm settle should... settle things down
<pie_>
samueldr, /dev/sdb exists
<pie_>
samueldr, im in the shell, running lvm lvscan gives the same error
wfranzini has quit [Remote host closed the connection]
<samueldr>
ah, much better results then
<samueldr>
but no idea! :/
wfranzini has joined #nixos
<pie_>
/dev/sdb: open failed> No medium found
<pie_>
ok thanks anyway, this helped too
<samueldr>
ooh
<samueldr>
do you have an sd card reader?
<samueldr>
or something like that
<pie_>
the question is why is it even trzing to do this, because i want to make sure i didnt misconfigure something
<pie_>
looks like it but ive never seen this error before
<samueldr>
wow is redhat scummy
<samueldr>
their access.redhat.com does the same ol' bamboozle as experts exchange did in the past
<pie_>
expertsexchange
Glider_IRC__ has joined #nixos
<samueldr>
so uh, what _is_ /dev/sdb in your computer?
<pie_>
samueldr, i have no idea
<pie_>
ah sorry, replied too fast, hold on
<samueldr>
because the message sounds like something that would happen if you had a drive without media in, like a floppy drive
<samueldr>
(but that wouldn't be /dev/sdX)
<pie_>
samueldr, the reason im a bit iffy is i set this up inside a VM with a raw disk, now im ooting the raw disk. but i cant remember the drive names so im not sure if it was running off sda or sdb in the vm
<pie_>
i used uuids
<pie_>
i think sdb might be my internal drive
<pie_>
ah no, nvme is my internal drive
<clever>
and nvme is always at a path like nvme0n1p?
<samueldr>
it's likely /dev/sdX are not affected from previously ran systems
<pie_>
samueldr, i was thiknig more along the lines of if my root was on sdb or something
<clever>
pie_: and for your home= problem, you need to use mountpoint=legacy and fileSystems.whatever.neededForBoot = true; to make nixos mount it early enough, or it may createHome before it mounts
<pie_>
i think i can look in by-id so see if it has some hardware specifier
<pie_>
clever, ok that makes perfect sense
<clever>
pie_: try blkid and lsblk
<pie_>
clever, but also unintuitive
marusich has quit [Remote host closed the connection]
<pie_>
clever, ive only got blkid in this shell right now
Glider_IRC_ has quit [Ping timeout: 272 seconds]
<pie_>
its not listed
<clever>
pie_: i would just keep home in /home, and swap things around when not booted into the os
<pie_>
ah wait i thikn udev had some command? thats what sets this up anyway right
<samueldr>
dmesg | grep /dev/sdb ?
<pie_>
samueldr, nothin useful
<samueldr>
:/
<pie_>
ok
<pie_>
problem solved
<pie_>
ls -l /dev/disk/by-id lists usb-Generid-_SD_MMCotherstuffhere -> /dev/sdb
<pie_>
samueldr was on point
<samueldr>
:)
<pie_>
apparently you can run udevadm info /dev/sdb
arahael2 has joined #nixos
pie__ has joined #nixos
schjetne has joined #nixos
MmeQuignon has joined #nixos
arahael2 has quit [Ping timeout: 272 seconds]
schjetne has quit [Ping timeout: 246 seconds]
Matthieu` has quit [Ping timeout: 246 seconds]
drakonis has joined #nixos
arahael2 has joined #nixos
drakonis_ has quit [Ping timeout: 272 seconds]
drakonis_ has joined #nixos
slack1256 has joined #nixos
arahael2 has quit [Ping timeout: 245 seconds]
drakonis has quit [Ping timeout: 252 seconds]
arahael2 has joined #nixos
wfranzini has quit [Remote host closed the connection]
teto has joined #nixos
dycan has joined #nixos
arahael2 has quit [Ping timeout: 248 seconds]
arahael2 has joined #nixos
arahael2 has quit [Ping timeout: 258 seconds]
arahael2 has joined #nixos
arahael2 has quit [Ping timeout: 245 seconds]
<teto>
is there a command to build and output the store path or should I compose 2 different ones ? nix-instantiate returns the drv
<slack1256>
nix-build is that command
emacsomancer has quit [Quit: WeeChat 2.4]
arahael2 has joined #nixos
mexisme has quit [Ping timeout: 248 seconds]
Glider_IRC_ has joined #nixos
emacsomancer has joined #nixos
zupo has joined #nixos
endformationage has quit [Ping timeout: 248 seconds]
Glider_IRC__ has quit [Ping timeout: 246 seconds]
mexisme has joined #nixos
emptyflask has quit [Quit: Lost terminal]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
arahael2 has quit [Ping timeout: 272 seconds]
<mac10688>
I am editing the xmonad.hs config file for xmonad. I am trying to use imported package haskellPackages.xmonad-contrib but when I import the modules and restart xmonad, it complains that it couldn't find the modules.
<mac10688>
How can I get xmonad.hs to find the imported modules
<slack1256>
mac10688: On your /etc/nixos/configuration.nix you can enable services.xserver.windowManager.xmonad = { enable = true ; exnableContribAndExtras = true ; } ; and will do what you want
<mac10688>
oh wow!
<slack1256>
But you actual question shows that you're trying to fit how other distros install packages into nixos, which doesn't allow that paradigm
<teto>
slack1256: thanks
<slack1256>
xmonad recompiles itself with modm-q, so it depends on a working copy of ghc plus xmonad as a library and contrib as a library. Nix install everything under /nix/store with symlink references to the dependencies, you have to declare that dependency on ghc when you build xmonad
<slack1256>
which is what the previous line on configuration.nix does
drakonis has joined #nixos
zupo has joined #nixos
<mac10688>
Sorry I'm trying to wrap my head around what you're saying
dckc has joined #nixos
<dckc>
I'm struggling to install idris on a mac:
<dckc>
dyld: lazy symbol binding failed: Symbol not found: _renameatx_np Referenced from: /nix/store/987ybb398qwmb69ai6i1nzbhx8z013c1-coreutils-8.31/bin/mv Expected in: /nix/store/mrrc440cshmc13rqqj4405jbqnklf4ja-Libsystem-osx-10.12.6/lib/libSystem.B.dylib
drakonis_ has quit [Ping timeout: 245 seconds]
mac10688 has quit [Quit: WeeChat 2.4]
<slack1256>
mac10688: it's okay, it is confusing at first
mac10688 has joined #nixos
<dckc>
Libsystem-osx-10.12.6 makes me wonder... i'm running 10.10 ; should I expect it to work? I don't see any documentation that says otherwise.
<slack1256>
what I am saying is this: what you were doing at first, installing xmonad-contrib after the declaration and install of xmonad would not work. Because xmonad needed to know where xmonad-contrib was building time for the declaration of its copy of ghc know where to look for.
<mac10688>
I was download xmonad-contrib in the configuration file like all my other packages. Your suggestion got it working.
<slack1256>
xmonad it's immutable once its built. To add xmonad-contrib to xmonad, you need to install a new derivation of xmonad with that dependency
<slack1256>
mac10688: nice ;-)
<dckc>
odd... `nix-shell -p idris` tries to install idris but `nix-env -i idris` does not
<gyroninja>
what about nix-env -iA nixos.idris?
<slack1256>
I don't think he is on nixos
<gyroninja>
oh
arahael2 has joined #nixos
<slack1256>
What is a good gopher client for linux?
palo1 has joined #nixos
<dckc>
`nix-env -iA nixpkgs.idris` did what I expected... except that it lost on _renameatx_np
palo has quit [Ping timeout: 272 seconds]
palo1 is now known as palo
mac10688 has quit [Quit: WeeChat 2.4]
arahael2 has quit [Ping timeout: 268 seconds]
<dckc>
so `nix-shell -p removeReferencesTo` triggers the _renameatx_np error. Anybody using a mac? care to try it?
veske has quit [Quit: This computer has gone to sleep]
schjetne has joined #nixos
MichaelRaskin has quit [Ping timeout: 258 seconds]
eal has left #nixos ["ERC (IRC client for Emacs 26.2)"]
linarcx has joined #nixos
Zer0xp has joined #nixos
<Zer0xp>
Hey does anyone here use polybar with NixOS ?
<Zer0xp>
I want to use it with my desktop manager instead of windows manager
<Zer0xp>
Does anyone know how I could set it up ?
arahael4 has quit [Ping timeout: 248 seconds]
<Athas>
I've 'nix build'ed a bunch of copies of the same derivation that I was working on. Now I have a dozen copies of it in /nix/store. How do I remove them? 'nix-collect-garbage' does not pick them up.
risson has joined #nixos
<cocreature>
Athas: nix-store --gc --print-roots
arahael4 has joined #nixos
<Athas>
cocreature: no obvious smoking guns. Everything looks reasonable, except for hundreds of 'memory' roots (but they are not pointing at my duplicate packages).
veske has joined #nixos
<Athas>
Unless nix-build creates new system links, I guess.
<cocreature>
nix-build creates a symlink to "result" by default that is a GC root
<cocreature>
if you run "nix-store --delete /path/to/store/path" it should tell you what is still referring to it
ee1943 has quit [Read error: Connection reset by peer]
jackdk has quit [Ping timeout: 248 seconds]
veske has quit [Client Quit]
<Athas>
Alright, I did that, and also deleted all previous NixOS generations. Looks like I only have two copies left of my package, which is explainable, I guess.
risson has quit [Quit: Pouet. WeeChat 2.4]
<hyper_ch>
Athas: how did you delete all previous ones?
ee1943 has joined #nixos
<Athas>
Speaking of, I'd like to test out a system using my own local changes to (parts of) nixpkgs. What's the best way of doing that? I know I can pass '-I nixpkgs=...' to nixos-rebuild, but then I'll also spend the entire day recompiling browsers and desktop environments.
<Athas>
hyper_ch: nix-collect-garbage -d.
<hyper_ch>
hmmm, that should have worked
<cocreature>
put your changes in an overlay
risson has joined #nixos
arahael4 has quit [Ping timeout: 246 seconds]
veske has joined #nixos
mexisme has quit [Ping timeout: 248 seconds]
arahael4 has joined #nixos
viric has quit [Read error: Connection reset by peer]
<teto>
Athas: make your changes on nixos-unstable, and avoid changing anything that could trigger a rebuild. This way you limit compilation to just your changes basically
viric has joined #nixos
<teto>
or for some big software, you can also pin them to a stable channel
<Zer0xp>
Hey does anyone here use polybar with NixOS ?
<Zer0xp>
I want to use it with my desktop manager instead of windows manager
<Zer0xp>
Does anyone know how I could set it up ?
asheshambasta_m has joined #nixos
asheshambasta_m has quit [Max SendQ exceeded]
asheshambasta_m has joined #nixos
jmeredith has quit [Quit: Connection closed for inactivity]
asheshambasta_m has quit [Quit: Leaving.]
<Athas>
teto: nixos-unstable corresponds to the master branch of nixpkgs?
<teto>
Athas: it's like the "master with a reasonable number of packages in cache"
<teto>
if you use git master, than you will have to compile everything
<teto>
(i.e., it's a bit behind master)
<teto>
,nixos-unstable
<yorick>
it looks like libtensorflow is built with -march=native on hydra
<teto>
,help
<{^_^}>
Use `,` to list all commands, `,foo = Foo!` to define foo as "Foo!", `,foo =` to undefine it, `,foo` to output "Foo!", `,foo somebody` to send "Foo!" to the nick somebody
<Athas>
teto: so what's the actual procedure? I can add the nixos-unstable channeland update my /etc/nixos/configuration.nix and do the upgrade. But then how should I test my local changes? If I modify my nixpkgs checkout and pass '-I nixpkgs=...' to nixos-rebuild, won't that still rebuild a lot?
<yorick>
wait, it's not built from source at all
arahael4 has quit [Ping timeout: 268 seconds]
<teto>
Athas: what I do is that I have a local git checkout that I regularly rebase on nixos-unstable. If you just modify some packages, then you won't compile much
<Athas>
teto: is there a nixos-unstable branch for nixpkgs? I can't find one.
<Athas>
Or do you mean you rebase on master?
arahael4 has joined #nixos
asheshambasta_m has joined #nixos
<{^_^}>
[nixpkgs] @dasJ opened pull request #62936 → nixos/memcached: Isolate the service → https://git.io/fjgI7
<teto>
Athas: yeah that's it, you setup several remotes, and you are good
<eyJhb>
Athas I normally checkout nixpkgs master, and rebuild it as is, with -I as you state. Then I make my changes to it, and test
<eyJhb>
Some things might rebuild, but most of the time it does not take THAT much time. Or you could grab a coffee in the meantime
mexisme has joined #nixos
<Athas>
Maybe I still have PTSD from running Gentoo as a teenager, where a full rebuild could take a weekend.
<clacke_movim>
Athas: if you want to rebuild your system without activating, so you can see first what came out
<clacke_movim>
You can `nix-build '<nixos/nixos>' -A system`
<clacke_movim>
There is probably a good nixos-rebuild command for this, but this works when playing with system definitions on a non-NixOS system too
<eyJhb>
Or just, nixos-rebuild boot
<clacke_movim>
Also a good way to get to the nixos-* tools when infecting a non-NixOS :-)
<clacke_movim>
Yeah, but that would still alter system configuration, just leaves the current running state alone
<clacke_movim>
There is a check or test, I think?
<clacke_movim>
But I think those don't print the path of the resulting system on stdout , but I.may be remembering wrong
<eyJhb>
nixos-rebuild --help -> dry-build, or dry-activate
<clacke_movim>
There we go, thanks!
<clacke_movim>
Anyway, that nix-build is what happens under the hood, and allows you to pass alternative configs and whatnot. Sooner or later the need arises. :-D
thc202 has joined #nixos
<eyJhb>
Athas you don't need to rebuild everything, it will use the cache when possible. Some stuff might need to be rebuild, but as long as your system is somewhat up to date, it shouldn't be that annoying. :) Plus, you can disable things that aren't important while testing.. E.g. Virtualbox with guest additions.....
<clacke_movim>
^^^^^
Anton-Latukha has joined #nixos
<etu>
clacke_movim: I use "nixos-rebuild test" quite often
<etu>
when testing changes to the config
<clacke_movim>
+1
<clacke_movim>
Does that just make sure everything evaluates?
<clacke_movim>
(so, runs nix-instantiate under the hood)
<eyJhb>
Activates the build, without adding it to grub
<{^_^}>
[nixpkgs] @abbradar pushed commit from @deadloko to master « qtcreator: add clang code model compilation »: https://git.io/fjgtz
cyris212 has joined #nixos
avn_ has quit [Remote host closed the connection]
<eyJhb>
Is it possible to configure networkmanager from .nix, like wpa_supplicant?
<teto1>
eyJhb: grep man configuration.nix for networkmanager (some of the config can be achieved through nix nut not everything)
<teto1>
at least not in a forward way
teto1 is now known as teto
<eyJhb>
teto the only place where it could be, might be extraconfig. But, it would be nice if you could configure wireless networks like i wpa_supplicant package.
civodul has joined #nixos
<eyJhb>
And also, for some reason it somewhat broke my entire system when I used wpa_supplicant
arahael4 has joined #nixos
Zer0xp has quit [Ping timeout: 256 seconds]
mexisme has quit [Ping timeout: 245 seconds]
mighty_vee has quit [Remote host closed the connection]
_cyris212 has joined #nixos
arahael4 has quit [Ping timeout: 248 seconds]
arahael4 has joined #nixos
MmeQuignon has quit [Ping timeout: 248 seconds]
MmeQuignon has joined #nixos
cyris212 has quit [Remote host closed the connection]
_cyris212 has quit [Quit: ZNC 1.6.5+deb1+deb9u1 - http://znc.in]
<clacke_movim>
li_matrix: You need to find the file defining that derivation, and feed it the right parameters. Luckily that's often (but not always) as easy as `nix-build -E 'with import <nixpkgs>; callPackage /path/to/file.nix { myattribute = "myvalue"; }`
<{^_^}>
[nixpkgs] @orivej-nixos pushed commit from @orivej to master « python: cx_Oracle: fix license »: https://git.io/fjgqO
toppler has joined #nixos
sigmundv has joined #nixos
<{^_^}>
[nixpkgs] @ar1a opened pull request #62939 → jost: init at 3.3 → https://git.io/fjgqs
Arahael has joined #nixos
<joepie91[m]>
<pie__ "joepie91, clever im opening more"> Good idea ☺️ I've added some links to related forum posts
<joepie91[m]>
Debugging is definitely also a topic that needs attention
mthst has joined #nixos
auri_ has joined #nixos
<auri_>
is this the correct place to ask questions about the nix package manager running on distros other than nixos?
timon37 has joined #nixos
psy3497 has quit [Ping timeout: 268 seconds]
Arahael has quit [Ping timeout: 268 seconds]
<joepie91[m]>
auri_: yep
<{^_^}>
[nixops] @Nekroze opened pull request #1162 → Add support for SSH jump hosts → https://git.io/fjgqC
<auri_>
alright. I'm currently trying to upgrade to nix 2.2.2 on void linux and it's giving me a bit of a headache - every single nix store executable segfaults
<auri_>
even the absolute simplest ones like "true" from coreutils
<auri_>
or "hello"
Arahael has joined #nixos
v88m has quit [Ping timeout: 248 seconds]
<joepie91[m]>
auri_: that sounds like a bug to me... Possibly a glibc mismatch?
mkoenig has joined #nixos
<joepie91[m]>
That's usually what's causing mystery segfaults when installing packages from unstable on stable NixOS as well... A runtime library version conflict
Arahael has quit [Ping timeout: 244 seconds]
<joepie91[m]>
Either way it's probably worth filling an issue about this on nixpkgs
<joepie91[m]>
If none exists yet
dansho has quit [Quit: Leaving]
Arahael has joined #nixos
<auri_>
joepie91[m]: indeed, it looks like a glibc mismatch. Void has 2.29 and nix has 2.27
<joepie91[m]>
Right. Unfortunately I have no idea how you would solve that, other than something like patching in a different glibc version in your nixpkgs. But I feel like there has to be a more streamlined option...
<auri_>
I could probably downgrade my Void glibc, though that's rather risky
<auri_>
I think the best option would be to patch nixpkgs
<joepie91[m]>
Also, changing the glibc version in nixpkgs means that you'll be locally building everything, since all the packages in the binary cache are built with standard glibc
<{^_^}>
[nixpkgs] @woffs opened pull request #62941 → openvpn: fix static key mini howto url → https://git.io/fjgqV
<auri_>
yep, that's not gonna be fun
<joepie91[m]>
But yeah, make sure to trawl the issue tracker for a bit first, to see if there are tricks for handling glibc incompatibility without rebuilds
<mthst>
did somebody manage to make boot.initrd.network work? can you share the config?
Arahael has quit [Ping timeout: 244 seconds]
<joepie91[m]>
Surely you're not the first person to run into this...
ericsagnes has quit [Ping timeout: 258 seconds]
<auri_>
you know what's interesting is that at some point (on a different Void install) I was able to set up Nix
drakonis has joined #nixos
<auri_>
and I'm 100% sure that machine had glibc 2.28
<auri_>
I guess it's 2.29 that broke it
<joepie91[m]>
Yeah, maybe that just happened to introduce an incompatibility
<auri_>
probably
drakonis_ has quit [Ping timeout: 248 seconds]
<tokudan>
can I use builtins.split to split multiple lines read from a file into a list? a regex "\n" or "\\n" does not seem to work
dycan has quit [Remote host closed the connection]
dashkal has joined #nixos
iyzsong has joined #nixos
<tokudan>
I'm probably missing something obvious right now: "df -h /" shows me 55% usage with ~11 GB free, but "mkdir /whatever" fails with "No space left on device"
<tokudan>
root is not mounted read-only
Arahael has quit [Ping timeout: 272 seconds]
Neo-- has joined #nixos
<iyzsong>
tokudan: maybe the inode is full? try 'df -i'
<tokudan>
iyzsong, bingo... that's it
<tokudan>
thanks
<tokudan>
so ext4 is not a good filesystem for nixos root
Arahael has joined #nixos
<hyper_ch>
who says so?
<tokudan>
well, if the nix store fills up the inodes...
<hyper_ch>
every file fills up the inodes
<hyper_ch>
I fail to see how that's related to ext4
<tokudan>
ext4 has a limited number of inodes. the nix store uses a high amount.
<hyper_ch>
tune2fs -l /dev/sdxy
<tokudan>
Inode count: 1572864
<hyper_ch>
so you can have 1.5 million files on that system
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<tokudan>
yeah. all used up by the nix store
<hyper_ch>
or set a different one with mke2fs -N
<hyper_ch>
tokudan: delete old generations then
<hyper_ch>
assign more inodes to the ext4
<hyper_ch>
use bigger partition that gets auto-assigned more inodes
<hyper_ch>
or use a different fs that has no inode limit per se
<tokudan>
the last one is what I'm going to do ;)
<hyper_ch>
but it's still far fetched to say ext4 is not a good fs for nixos store or root
<hyper_ch>
I'd recommend to upgrade to zfs :)
<tokudan>
RAM is at a premium for that system, it only has 2 GB RAM
<tokudan>
if that wouldn't be an issue, I'd have used zfs
<hyper_ch>
ieeeks :)
Anton-Latukha has quit [Ping timeout: 246 seconds]
<das_j>
Hey, is there some way to add a phase to a derivation?
<simpson>
das_j: Sure, I think so. For example, if you're developing a language-specific kind of derivation and need to do some specific sort of work in a phase. What are you working on?
<das_j>
simpson: I'd like to have the *last* command executed be some `find | br` to compress the web assets
<das_j>
simpson: Basically something (from within an overlay): myPackage = compressWeb super.myPackage;
<simpson>
das_j: Sounds like you want postFixup then. No need for a custom phase, just use that one!
<das_j>
I did it with pkgs.runCommand (and taking the original derivation as input), but as it turns out, stuff like .override stops working then (obviously)
<das_j>
simpson: Do I need to do postFixup = ''${old.postFixup}; mystuff'' or are the lines appended?
Arahael has quit [Ping timeout: 268 seconds]
ericsagnes has joined #nixos
<simpson>
das_j: I don't know! That's an interesting question. How many fixups do you have? Normally a derivation has just a single fixupPhase specified; the reason to use postFixup is already to try to avoid clobbering other fixups.
<das_j>
simpson: I now have ${old.postFixup or ""} and it seems to work. I'll just leave it as-is until something breaks ;) thank you
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<zfnmxt>
Which package provides `nvidia-settings`?
Arahael has quit [Ping timeout: 244 seconds]
<{^_^}>
[nixpkgs] @joachifm opened pull request #62946 → udisks1: mark as broken → https://git.io/fjgmS
<etu>
,locate nvidia-settings
<{^_^}>
Found in packages: zafiro-icons, maia-icon-theme, moka-icon-theme, paper-icon-theme, iconpack-obsidian, numix-icon-theme-circle, numix-icon-theme-square
<etu>
,locate bin nvidia-settings
<{^_^}>
Couldn't find in any packages
<zfnmxt>
etu: It seems like it's a part of nvidia_X11 and `locate nvidia-settings` finds a it in my /nix/store when I install nvidia_X11, but it doesn't actually get placed in my PATH o.O
alter-schjetne has joined #nixos
Arahael has joined #nixos
schjetne has quit [Ping timeout: 258 seconds]
Henson has joined #nixos
rauno has quit [Ping timeout: 244 seconds]
arahael1 has joined #nixos
Arahael has quit [Ping timeout: 244 seconds]
drakonis1 has joined #nixos
Ariakenom has joined #nixos
<mthst>
can i route services.transmission traffic through a wireguard interface?
kvaster has joined #nixos
veske has quit [Quit: This computer has gone to sleep]
arahael1 has quit [Ping timeout: 258 seconds]
MinceR_ is now known as MinceR
o1lo01ol1o has joined #nixos
Anton-Latukha has joined #nixos
phreedom has quit [Remote host closed the connection]
arahael1 has joined #nixos
phreedom has joined #nixos
<{^_^}>
[nixpkgs] @adisbladis merged pull request #62315 → nixos/pulseaudio: Set speex-float-5 as default resample-method → https://git.io/fj0jA
<{^_^}>
[nixpkgs] @ggPeti opened pull request #62950 → temporal_tables: init at 1.2.0 → https://git.io/fjgOa
o1lo01ol1o has joined #nixos
danielrf has quit [Quit: WeeChat 2.4]
wucke13_ is now known as wucke13
mexisme has joined #nixos
kvaster has quit [Ping timeout: 248 seconds]
swerlk has joined #nixos
mexisme has quit [Ping timeout: 244 seconds]
hoijui has joined #nixos
alter-schjetne has quit [Ping timeout: 252 seconds]
toppler has quit [Ping timeout: 248 seconds]
<Athas>
teto: I followed your advice and did 'sudo nixos-rebuild test -I nixpkgs=~/repos/nixpkgs/', where ~/repos/nixpigs is a checkout of the nixos-unstable branch from https://github.com/nixos/nixpkgs-channels (with local modifications to mesa and libclc). It looks like it's rebuilding from scratch things like GIMP and WebKit, for reasons I fail to understand.
<Gopal[m]>
timeouts also seem to be occuring with llvm 8.0
<Gopal[m]>
something on my side?
dwdr has joined #nixos
nikivi has quit [Read error: Connection reset by peer]
shibboleth has joined #nixos
arahael1 has joined #nixos
<Henson>
I'm having trouble enabling dhcpd on NixOS. I've specified services.dhcpd4.enable = true and some stuff in extraConfig, but it's not starting. Looking at the source it looks like dhcpd.nix checks for services.dhcpd.enable being true, but nixos-options says that services.dhcpd is just an alias for services.dhcpd4.
<Henson>
yeah, and renaming dhcpd4 to dhcpd causes nixos-rebuild to complain about it
<Henson>
clever: ugh, the problem is that I didn't specify the interfaces on which to listen, so it just uses "eth0" by default, which doesn't exist.
<clever>
Henson: ah, that will do it
<Henson>
clever: should that be changed? When starting dhcpd manually including eth0 it barfs. But when leaving eth0 off it does the right thing and figures out which devices to listen to. Should dhcpd.nix be changed to have that behaviour? If not, shouldn't it require that the interfaces be specified?
veske has quit [Quit: This computer has gone to sleep]
<Henson>
clever: ahh yes, I see
<clever>
Henson: yeah, you can try setting it to [] to just force it to figure it out, and maybe file a pr if you think thats worth being the real default
<clever>
you can just set interfaces = []; in configuration.nix for testing
<Henson>
clever: ok, I see that now
mthst has quit [Ping timeout: 248 seconds]
<m1crOman>
clever, your kexec worked perfectly for me :) Was able to use it on my Hetzner box. I didn't run justdoit though I wanted to do my own thing involving boot + root on a memory key but it's all running great now
<clever>
m1crOman: :D
mmercier has quit [Quit: mmercier]
<mpickering>
does dontBuild stop postUnpack working?
<clever>
mpickering: your thinking of dontUnpack i believe
<mpickering>
I'm trying to work out why postUnpack is not working
<mpickering>
it doesn't seem to run the phase
<Ralith>
does a nix-shell env not run setup hooks?
<clever>
mpickering: looks like unpack phase doesnt have a disable flag, you just want unpackPhase = ":";
<Gopal[m]>
@clever is how I got nixos running on all my servers too :D
<mpickering>
clever: I don't want to disable it
<mpickering>
I want to work out why it is disabled
<clever>
mpickering: did you set buildCommand?
<mpickering>
No
Matthieu_ has quit [Ping timeout: 244 seconds]
<clever>
mpickering: can you paste your full nix expr?
<mpickering>
It looks like configureFlags aren't applied even
<mpickering>
something looks very wrong
<clever>
mpickering: it looks like the configure script is using #!/bin/bash, which doesnt exist
<m1crOman>
The learning curve of nix is hitting me very hard though. I am now trying to figure out how best to store my dns configs. whilst browsing repos earlier I saw someone declare file contents where a path was expected. That was cool, but I can't find it again :P
<mpickering>
right but there is a patchShebangs which should fix it
<ToxicFrog>
Given an image on Hydra, is there any way to get the configuration.nix that produced it? I'm trying to debug an issue with the rpi3 where the install image boots fine but nothing (including the original generation baked into the install image) boots successfully after the first nixos-rebuild.
<clever>
mpickering: can you paste your full nix expr?
<mpickering>
I am running `nix-build . -A ghc802 --arg pkgs "import <nixpkgs> {}"`
<mpickering>
Which invokes `artifact.nix` via `default.nix`
arahael1 has quit [Ping timeout: 272 seconds]
arahael1 has joined #nixos
<mpickering>
hmm I am doing something a bit wrong I think because it does two builds
<mpickering>
The first build seems to use default settings
mthst has quit [Ping timeout: 246 seconds]
<mpickering>
the second does applying the patchShebangs and so on
fusion809 has quit [Remote host closed the connection]
<mpickering>
The name of the derivation it tries to build is "building '/nix/store/mxn6j4s49m7i1c0nb8vw851bxxrp3cs0-bindist-version.drv'..."
mthst has joined #nixos
<mpickering>
ah I see now
<mpickering>
thanks all
Izorkin_ has joined #nixos
<m1crOman>
ToxicFrog, that isn't what I'm looking for actually. I just want a way to load my blog somewhere onto the system. https://github.com/NixOS/nixops/blob/master/examples/nix-homepage.nix this seems like a good starting point - the fetchFromGithub does p. much exactly what I need. Then I just need to build the nodejs project
Izorkin has quit [Ping timeout: 248 seconds]
Izorkin_ is now known as Izorkin
<m1crOman>
they do that path-as-a-derivation magic too. That's really cool.
drakonis has quit [Ping timeout: 272 seconds]
<hyper_ch>
grahamc[m]: with yesterday's fix to wireguard service it seems now to work just fine
<m1crOman>
So I guess if I put my DNS records in a github repo and then made a derivation that pulls that repo, I could reference the files from there.
drakonis has joined #nixos
<hyper_ch>
what's the best way to find out if a a git commit to master is already in unstable?
selfsymmetric-pa has joined #nixos
<{^_^}>
[nixpkgs] @abbradar pushed to master « dwarf-fortress-packages.dwarf-therapist-original: expose to package set »: https://git.io/fjgsq
<eraserhd>
what's the process for a maintainer not responding on a PR? Specifically, how long do we wait?
kim0 has joined #nixos
<eraserhd>
(nixpkgs, I mean)
nD5Xjz has quit [Ping timeout: 258 seconds]
erasmas has joined #nixos
nD5Xjz_ has joined #nixos
bdesham[m] has joined #nixos
arahael1 has quit [Ping timeout: 268 seconds]
<ekleog>
FRidh: hey! got a second to talk about the python infrastructure? I've got an idea that I'd like to run by you before spending too much time on it
<ekleog>
FRidh: basically, the idea is to try and bring static linking to python, by (ab)using `sys.meta_path`, which would allow us to stop needing `propagatedBuildInputs` for python
karetsu has joined #nixos
<ekleog>
FRidh: (and `python.withPackages` too, by the way of static linking too)
<joepie91[m]>
selfsymmetric-pa: (it's a community fork of Font Awesome, since that has become more and more commercial and proprietary over time)
<joepie91[m]>
eraserhd: I don't think we have a formal process for that? I may be wrong though
xcthulhu has quit [Quit: Leaving]
<ekleog>
FRidh: as far as I can guess I think this could be done with a patch to python to get it to always consider an additional `sys.meta_path` element in addition to the ones actually in the list, that would check into the `nix-support/python-deps` file in the store (relatively from the calling file) to figure out where to import from, but I wonder if you can see any reason not to try and do it, or hurdles I might
<ekleog>
be missing?
arahael1 has joined #nixos
<eraserhd>
joepie91[m]: ok. It's reasonable to wait a bit longer, but I was curious.
Matthieu_ has joined #nixos
<joepie91[m]>
eraserhd: maintainer PR handling is generally a bit unreliable. there was a purge of maintainers and unmaintained packages a while ago iirc to at least partially address it, but perhaps some of the issues have persisted
<Taneb>
Does anyone have experience or know blog posts about using CI software other than Hydra in conjunction with nix builds?
<joepie91[m]>
that sounds like a question that gchristensen may have an answer to
<Taneb>
I'm exploring options
o1lo01ol1o has quit [Remote host closed the connection]
<ekleog>
FRidh: oh just saw you're on -dev too, if you want we can move the discussion over there :)
<karetsu>
Taneb: I'm in the midst of trying to get a GitLab CI runner sorted in work but with all the issues we have around data security its a chore convincing people atm
o1lo01ol1o has joined #nixos
<Taneb>
We're currently using Hydra at work with reasonable success but we've got some itches it doesn't quite scratch
<Taneb>
(mostly, we want to be able to do builds which aren't nix too)
<karetsu>
we have a self-hosted gitlab server but 90% of the features are turned off atm -_-
<__monty__>
joepie91[m]: Are we talking about the maintainer field of packages? Cause there's no automatic notification of such maintainers afaik?
o1lo01ol1o has quit [Remote host closed the connection]
<karetsu>
Taneb: is the biggest issue support for non-nix builds?
rfold has quit [Ping timeout: 258 seconds]
<karetsu>
because for all my non-nix things gitlab has been great
<Taneb>
karetsu: that and some reliability issues (it's not picked up changes before, or not started a build)
<Taneb>
Yeah, gitlab's one of the things on my list
<karetsu>
having CI/CD just built into the sidebar and options for a repo is nice, work are split across internal gitlab and external github+travis and gitlab is just less faff
<NYXT>
I'm trying to write a package that uses appimageTools.wrapType2 but it complains that meta and profile (and presumably other attributes if I kept commenting them out) are unexpected arguments despite being able to find a bunch of examples of this working on nixpkgs? Any ideas?
<clever>
> appimageTools.wrapType2
<{^_^}>
<LAMBDA>
<Henson>
how does one enable the package installation auto suggestion when you type a command that isn't recognized? It's enabled on one of my NixOS computers but not the other, and I don't know how it got enabled, but I want to enable it on the other.
<NYXT>
It's almost identical to a lot of the other packages I've found on NixOS/nixpkgs when looking for wrapType2 examples.
<clever>
NYXT: and what is line 56 of /nix/var/nix/profiles/per-user/root/channels/nixos/pkgs/build-support/appimage/default.nix ?
<NYXT>
I've added that as another comment.
* NYXT
It's different from what you linked on nixpkgs.
<clever>
NYXT: yeah, the ... is what allows it to ignore other arguments
<clever>
your nixpkgs is old and lacks that feature
arahael2 has quit [Ping timeout: 248 seconds]
* NYXT
Ah. I see.
<NYXT>
I need to learn how to use my IRC client better, I don't know why the messages are sending with "/me".
ambro718 has joined #nixos
drakonis_ has quit [Ping timeout: 252 seconds]
o1lo01ol1o has quit [Remote host closed the connection]
drakonis_ has joined #nixos
hoijui has quit [Quit: Leaving]
<Henson>
wow, I figured it out. You just have to do "nix-channel --update" and it will update the /nix/var/nix/profiles/per-user/root/channels/nixos/programs.sqlite file to make the "command-not-found" script work
drakonis_ has quit [Ping timeout: 252 seconds]
arahael2 has joined #nixos
asheshambasta_m has quit [Quit: Leaving.]
nschoe has joined #nixos
Anton-Latukha has joined #nixos
arahael2 has quit [Ping timeout: 245 seconds]
craigevil[m] has joined #nixos
Jackneill has quit [Remote host closed the connection]
arahael3 has joined #nixos
sigmundv has quit [Ping timeout: 246 seconds]
alex`` has quit [Ping timeout: 248 seconds]
shibboleth has quit [Quit: shibboleth]
Anton-Latukha has quit [Ping timeout: 246 seconds]
alex`` has joined #nixos
<{^_^}>
[nix] @grahamc opened pull request #2934 → libstore: consider a resolution error as a NotFound → https://git.io/fjgG4
judson has joined #nixos
o1lo01ol1o has joined #nixos
<nschoe>
Hi everyone. Anyone particular with `QtCore/QtCoremod.sip`? It seems `sip` cannot find it when building my derivation, I'm clueless as to hwo to proceed.
<lordcirth>
nschoe, could you please pastebin the exact error?
<Henson>
I've got a derivation that uses buildCommand to build it that references files in the source tree using ${./source.file.cpp} for example. The problem is this source file is copied to the nix store. If I don't use the ${} braces then the buildCommand isn't able to find the file. How can I prevent these source files from going into the nix store?
v88m has joined #nixos
acarrico has joined #nixos
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
Anton-Latukha has joined #nixos
o1lo01ol1o has joined #nixos
arahael4 has joined #nixos
<nschoe>
lordcirth, looks like the .sip module is somehow not in path
<Henson>
hmm, I see why it might need these files, as they are part of computing the hash of the derivation
<nschoe>
Henson, yes they are. Any input is part of the hash for a derivation.
arahael3 has quit [Ping timeout: 245 seconds]
karetsu has joined #nixos
<Henson>
nschoe: but is there some way to not include the source code for your program in your package output? I guess this is the difference between binary and source deployments. But for some reason the source seems to be making its way into binary deployments.
* Henson
thinks
<karetsu>
clever: my xmonad problem... my ~/.ghc/<blah>/package.conf.d had no packages in it -__-
<exarkun>
how do I set an environment variable in the environment I get from python.withPackages?
<karetsu>
one to remember for the future if people switch distro and retain /home
<nschoe>
Henson, I'm not super knowledgeable in nixOS, I'm a newbie as well. But if you use `mkDerivation` and set the `src` attribute, I beliebe your source code gets copied to `${out}/source/`.
<Henson>
nschoe: ok, thank you
<nschoe>
Henson, I'd recommence reading the nix-pills, it really helps make sense of certain things.
arahael4 has quit [Ping timeout: 248 seconds]
<{^_^}>
[nixpkgs] @jonringer opened pull request #62960 → vimPlugins.indentLine: init at 2019-02-22 → https://git.io/fjgGd
arahael4 has joined #nixos
freeman42y has joined #nixos
qoy has joined #nixos
<NYXT>
Do I need to do anything specific to have my udev rules apply to a FHS user env (in particular, a wrapped appimage)?
<nschoe>
selfsymmetric-pa, `nix searhc font-awesome` gives you the correct name
<karetsu>
yeah, or ttf, I assume ttf will be the latest one?
<selfsymmetric-pa>
I think something's wrong with my cache...
acarrico has quit [Ping timeout: 258 seconds]
<selfsymmetric-pa>
Yeah, I just want my icons to work in Firefox. Right now I'm getting too many boxes.
qoy has joined #nixos
qoy has quit [Client Quit]
<eyJhb>
Generally me and icons are a no go. I have loads of applications where I have to GUESS where the buttons are. - Which sites are you visiting? Haven't had problems with icons on pages
<karetsu>
surely the websites deal with that in most cases because of font APIs?
ks0ne has joined #nixos
arahael4 has joined #nixos
ks0ne has quit [Read error: Connection reset by peer]
<manveru>
at least font-awesome, nobody relies on that being installed
<m1crOman>
When using fetchFromGithub what is the sha256 based on?
<manveru>
selfsymmetric-pa: did you check what font the site actually wants?
<ToxicFrog>
m1crOman: the contents of the commit that's fetched.
<manveru>
m1crOman: nix-prefetch-url $url/archive/$rev.tar.gz i think
<nschoe>
m1crOman, you can compute the sha256 with `nix-prefetch-github <github-ownser> <repo> --rev specific-commit
<manveru>
or that :D
<manveru>
,tofu
<{^_^}>
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.
<ToxicFrog>
You can also just set it to something obviously wrong in the .nix and then see what error message you get~
<hyper_ch>
manveru: with 5G everything will be better - even in Germany ;)
* hyper_ch
huggles his symm. gigabit fiber
<eyJhb>
5 mins away I have access to 10 Gbit/s
<manveru>
we barely have 3G here :P
<m1crOman>
nschoe, manveru, ToxicFrog thanks for the info :)
<nschoe>
hyper_ch, scary :/ But again: it's true if you host it on gitlab or bitbucket. There's no better things that to make backups, host in several places :D
<eyJhb>
hyper_ch but, do you trust yourself to make good backups of your gitlab setup?
<hyper_ch>
zfs send/recv
<nschoe>
eyJhb, my thoughts exactly :)
<hyper_ch>
I have at least 2 offsite backups of everything I do
<nschoe>
hyper_ch, true, but your provider might also shutdown your server ^^
<hyper_ch>
why?
<nschoe>
hyper_ch, yeah that's the kind of paranoia I like :D
<eyJhb>
I don't trust myself still.. I need to make sure that the off-site backups works too, and that I can restore from them :|
<eyJhb>
The only thing that keeps my containers running etc. is my fatso container + pushbullet
<manveru>
eyJhb: i pay around 40 EUR for 10/2 Mbit/s conn :(
<hyper_ch>
I pay around 50 for symm. gigabit + 1 static ipv4
<eyJhb>
manveru this is shared where I live, but I am guessing that I am the one who uses the most.. We share 700 Mbit/s, and are all limited to 100 Mbit/s
<manveru>
and my landlord refused to let me install cable connection this year again...
drakonis_ has joined #nixos
<eyJhb>
manveru apt?
<manveru>
can't even dream of 4G or fibre
<eyJhb>
apartment*
<manveru>
yeah
<eyJhb>
Do like me, what the landlord doesn't know is none of his concern
<eyJhb>
I have like.. Setup cable canals, bunch of holes in the walls, moved most of the kitchen to install dishwasher etc..
ks0ne has joined #nixos
<eyJhb>
I litterally have a cable that goes through 3 walls, across my apartment, to get the connection into my router
<manveru>
lol
ks0ne has quit [Read error: Connection reset by peer]
<nschoe>
Argh did anymore ever managed to package something with sip and pyqt5? I'me constantly getting "sip: Unable to find file "QtCore/QtCoremod.sip" and I've tried about 73467 different things :/
<eyJhb>
manveru he has pretty much given up on me.. But student apartments, they are quite relaxed...
<manveru>
eyJhb: well, there's 3 more owners of this house and they live here, so doing stuff without them noticing (like digging up half our street) won't go over well
<clever>
error: path '/nix/store/1d2hm2z1922m7yah8m5bjjxyricpg5p5-hi' is not valid
<clever>
gchristensen: dang, it can eval an expr, but it wont build it!
<gchristensen>
hehe
<gchristensen>
it is fine, I can `rev` it :P
Anton-Latukha has joined #nixos
<clever>
gchristensen: let me check your original script closer...
<aanderse>
i've never packaged anything using nodejs before and the documentation seems to expect a higher level of understanding of nodejs/npm than i have... can anyone take a look and possibly help me out? http://ix.io/1Lqw
<aanderse>
the package i'm trying to build uses cmake, and it also uses cmake to run npm
<clever>
gchristensen: it deleted just fine, what does `nix-store --query --roots` say for your path?
<nschoe>
aanderse, I know what you're going through ^^ My advice is to try and see in nixpkgs how nodejs-packaged packages are build. Maybe try to reproduce?
<aanderse>
nschoe: yeah, i have been tinkering a bit... not fun
_ris has joined #nixos
ricekrispie2 has joined #nixos
<nschoe>
aanderse, I'm trying to do something similar, too (and also failing). So... have faith!
<aanderse>
nschoe: :)
shibboleth has joined #nixos
ris has quit [Ping timeout: 260 seconds]
weedloser has quit [Ping timeout: 248 seconds]
ricekrispie has quit [Ping timeout: 248 seconds]
<{^_^}>
[nixpkgs] @bjornfor opened pull request #62963 → nixos-generate-config: add size attribute for swap *files* → https://git.io/fjgnC
gratinb has joined #nixos
arahael1 has joined #nixos
<Athas>
I don't understand this example from the Nix manual: nix-shell '<nixpkgs>' -A pan
<Athas>
How can passing '<nixpkgs>' for the path work? According to the nix-shell manpage, that has to be a filename or an http/https address.
Arahael has quit [Ping timeout: 258 seconds]
gratin has quit [Ping timeout: 248 seconds]
<nschoe>
Athas, yeah, '<nixpkgs>' is sort of magic, it works there too.
<nschoe>
There's a nix pill about it, let me find it
<{^_^}>
[nixpkgs] @peti pushed 2 commits to haskell-updates: https://git.io/fjgnz
ks0ne has quit [Read error: Connection reset by peer]
ks0ne has joined #nixos
ks0ne has quit [Read error: Connection reset by peer]
arahael1 has joined #nixos
<clever>
gchristensen: did --query --roots reveal anything?
boomshroom has joined #nixos
<gchristensen>
will look later :)
<boomshroom>
Hello!
arahael2 has joined #nixos
<boomshroom>
I recently got an SSD and moved my EFI boot partition onto it. How can I get NixOS to reset my motherboard to boot from it rather than the main drive? I did a partition copy, so it has the same UUID as the old location if that changes anything.
<nschoe>
especially the cmakeFlags? I'm using this derivation (which is called `mkRosPackage`) to build a package, but when I set the cmakeFlags inside, they seem to be discarded.
<joepie91[m]>
boomshroom: I would strongly recommend changing the UUID if the old drive is still connected - duplicate UUIDs will break a ton of stuff in weird ways ☺️
<nschoe>
Is this derivation override the cmakeFlags and make me unable to specify additional ones?
<colemickens>
"reset motherboard" = new system efi manager entry?
<boomshroom>
joepie91[m]: I deleted the old partition in hopes of reducing confusion, but it seems like it just made things worse.
<colemickens>
some BIOS tools will let you delete entries that have been pushed in, otherwise you can use efibootmgr to edit them manually, or re-call `nixos-install` to reactivate the bootloader.
<colemickens>
I'd recommend: `efibootmgr` delete all nixos entries, boot USB media, set NIX_PATH, call `nixos-install` and let it re-ensure your bootloader is setup.
<boomshroom>
I did make a new entry from a live installer; it didn't work, which is why I want to get NixOS to set the entry because it's less likely to mess up.
<boomshroom>
I can do nixos-install into an existing NixOS installation
<boomshroom>
?
<colemickens>
I don't know what "into" means necessarily, but I think the answer is yes.
<colemickens>
`nixos-install` can be safely re-executed in virtually every case I can imagine.
<colemickens>
it basically builds the system config, activates it and calls the activation script that ensures the bootloader is configured per the system config.
<boomshroom>
Can I call that activation script manually?
<colemickens>
(which has left me actually wondering what the difference is between `nixos-install` and `nixos-rebuild switch --install-bootloader` but I'm not in the mood to go read perl right now)
Ariakenom has quit [Quit: Leaving]
<colemickens>
You could, but I would really commend just calling nixos-install
<boomshroom>
--install-bootloader I forgot about that when building the system!
<colemickens>
yeah, nixos-install more or less bottoms out on calling switch-to-configuration which is the same thing nixos-rebuild eventually calls.
<colemickens>
Anyway, I run a custom nixpkgs, so I basically mount my partitions, clone my nixpkgs, set `NIX_PATH=nixpkgs=/tmp/nixpkgs:nixos-config=/tmp/nixos-config.nix` and then call `nixos-install`.
ThatDocsLady has quit [Ping timeout: 252 seconds]
jboy has quit [Ping timeout: 246 seconds]
<boomshroom>
That script mentions nixos-enter, but I don't see a similar command
<boomshroom>
Yay for using an old installer that doesn't use nix 2.0
arahael2 has quit [Ping timeout: 248 seconds]
<aveltras>
is it possible to deploy from an ubuntu machine to a nixos server and inside a nixos container on that server ?
<aveltras>
(using nixops)
<adisbladis>
aveltras: Yes, it's possible.
<aveltras>
adisbladis: are containers the way to have multiple apps running on one server ?
<kraem>
i've been struggling with gpg-agent, ssh-agent, gnome-keyring and seahorse for the last day - i couldn't get gpg-agent use a pinentry which follows the set gtk theme so i switched to gnome-keyring + seahorse which solved the problem - now I don't get a graphical query for the ssh-key passphrase but only in the terminal. I've tried setting programs.ssh.startAgent to true but I can't figure out how to run
<kraem>
ssh-add on login though - right now I have to remember to do it before i SSH anywhere. Any thoughts? (I'm using fish shell btw) thanks!
<aveltras>
and, is there an example somewhere ? i'm stuck right now trying to make this work
<aveltras>
i already have a running nixos server
<adisbladis>
aveltras: Sure, it's a good way to run multiple instances of a service.
<adisbladis>
aveltras: And have you made a nixops deployment?
<aveltras>
ye
<aveltras>
it fails trying to connect to localhost as root
<pie__>
joepie91[m], extremely tangential to discussions of with; i wouldnt mind a non-recursive let construct :I
<adisbladis>
aveltras: Have you set up a key to access the root account?
<adisbladis>
aveltras: Have you copied your configuration.nix and hardware-configuration.nix off the host?
<aveltras>
does it use the keys set up on machine ? i can "ssh root@ip" in bash on the machine without entering password but i have not done anything specific for nixops
<Athas>
symphorien: thanks, that's useful. Although it looks like there are duplicates...
<aveltras>
adisbladis: no
<symphorien>
it is possible that not all software on your system uses the same mesa
<aveltras>
adisbladis: do i have to modify anything server side ?
<adisbladis>
aveltras: Right, so you are missing all your hardware configs
<adisbladis>
aveltras: No, you can just copy these files off the server
<aveltras>
you mean like, the physical config should be a merged configuration.nix + hardware-configuration.nix taken froim the server + deployment.targetHost = xxx ?
<boomshroom>
I probably could have used my hard disk to avoid downloading all of 18.04, but oh well
<joepie91[m]>
boomshroom: yes, afaik
<joepie91[m]>
pie__: hmm. have you run into issues with it?
<{^_^}>
[nixpkgs] @kalbasit pushed commit from @dtzWill to master « miniflux: 2.0.15 -> 2.0.16 (#62918) »: https://git.io/fjgCa
o1lo01ol1o has quit [Remote host closed the connection]
selfsymmetric-pa has joined #nixos
mexisme has joined #nixos
o1lo01ol1o has joined #nixos
<joepie91[m]>
I can't find any of my thumbdrives, and I want to install NixOS to a new SSD I got for this PC (that I'm talking from right now). is there a (reliable) way to do that, without needing any kind of thumbdrive or optical media?
o1lo01ol1o has quit [Ping timeout: 272 seconds]
<joepie91[m]>
netboot, installing from a live system, whatever
<clever>
joepie91[m]: add this to the imports section of a laptop's configuration.nix, point the wan option on line 54 to the wifi card, and the lan option to the ethernet
<clever>
joepie91[m]: anything you plug into the ethernet can now netboot
<clever>
and the laptop will share its wifi to that device over the ethernet
ryantrinkle has quit [Ping timeout: 248 seconds]
fendor has quit [Quit: Leaving]
<joepie91[m]>
clever: oh, huh, nice, that seems like what I need.
Lem_ has joined #nixos
<joepie91[m]>
clever: netboot should then work from my system's UEFI without issues?
<clever>
joepie91[m]: i have uefi netboot code on line 93, but ive not really had much luck getting it to work before
<Lem_>
hi, I'm looking for a way to set postgres listen_address to '', instead of '*' or 'localhost'.
selfsymmetric-pa has left #nixos ["gone to the land of dead hiccups and extinguished light bulbs"]
<clever>
joepie91[m]: you can also do a `boot.loader.grub.efiInstallAsRemovable = true;` after booting via legacy pxe
<clever>
joepie91[m]: thats how i setup my current nvme+efi laptop
<joepie91[m]>
clever: so how would I boot into this from my desktop?
<joepie91[m]>
given that I have no thumbdrive
<clever>
joepie91[m]: you would need to take an existing nixos laptop, clone my repo, and add /path/to/netboot_server.nix to the imports of configuration.nix
<clever>
joepie91[m]: and nixos-rebuild switch that laptop
Soo_Slow has quit [Remote host closed the connection]
<joepie91[m]>
right, I get the laptop side, my question is what do I do on the desktop side, ie. on the side of the system-to-be-installed
<clever>
joepie91[m]: just tell the bios to network boot, and plug it into the laptop's ethernet port
civodul has quit [Quit: ERC (IRC client for Emacs 26.2)]
ris has joined #nixos
<clever>
if i'm reading this right, you can just drop a .vmdk file onto a harddrive, and tell windows to boot from it
_ris has quit [Ping timeout: 258 seconds]
<joepie91[m]>
clever: ah right, lol, I didn't know that nixos-install existed outside of the live CD
<joepie91[m]>
will do that instead, then :P
<joepie91[m]>
but first, I need to actually install the SSD...
<clever>
yeah, if i'm reading the above page right, windows just natively supports running from disk images, neat
arahael2 has quit [Ping timeout: 248 seconds]
ddellacosta has joined #nixos
toppler has quit [Ping timeout: 272 seconds]
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
domogled has quit [Ping timeout: 244 seconds]
arahael2 has joined #nixos
qqlq has quit [Read error: Connection reset by peer]
qqlq has joined #nixos
toppler has joined #nixos
qqlq has quit [Client Quit]
ilya-fedin has joined #nixos
o1lo01ol1o has quit [Ping timeout: 245 seconds]
qqlq has joined #nixos
o1lo01ol1o has joined #nixos
mac10688 has joined #nixos
qubasa has joined #nixos
<mac10688>
hi, i have a xrandr command I want to run before login screen appears to position my secondary monitor above my laptop. Where is a good place for me to save that command?
fendor has joined #nixos
<m1crOman>
folks I'm trying to run bind on my server. I've added a very simple single zone master config, with the zone loaded from a local file. When I run nixos-rebuild switch it's just exiting with code 1 and not printing any messages. What could this be? I had no issues on a different server. it even told me when I had config issues
<THFKA4>
mac10688: isn't this something that goes into xorg.conf?
<mac10688>
I don't know. I never heard of a xorg.conf until about 10 minutes ago. I think the page was talking about how to get rid of it or something.
<m1crOman>
btw see when I install nixos, and then I want to create a local copy of the nixpkgs git, if I want to keep it inline with what I just installed is it correct to do nix-channel --list and clone the linked head of that channel?
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
Lears has joined #nixos
<matthewbauer>
it might, i'm not sure how ghc decides whether it can run th or not, you might need to override something
mexisme has quit [Ping timeout: 272 seconds]
[Leary] has quit [Ping timeout: 248 seconds]
<nh2>
matthewbauer: and do you know about the absence of the 32-bit musl entry?
<joepie91[m]>
clever: do you know of any gotchas or reasons *not* to use an LVM volume for a swap partition?
<matthewbauer>
probably just an oversight. you only need those if you set ```localSystem```. you can use crossSystem without those entries
mexisme has joined #nixos
Avaq has quit [Ping timeout: 250 seconds]
m1cr0man has joined #nixos
<nh2>
thanks!
o1lo01ol1o has quit [Remote host closed the connection]
ng0 has quit [Ping timeout: 256 seconds]
o1lo01ol1o has joined #nixos
Anton-Latukha has quit [Ping timeout: 258 seconds]
<clever>
joepie91[m]: ive never ran into problems with swap on lvm
<joepie91[m]>
great, thanks :)
<clever>
joepie91[m]: but about the only reason i can see to do that is to be able to resize it easily
<joepie91[m]>
right, that is the objective
<joepie91[m]>
clever: and probably the last question: if I want to use LVM for my main system disk, and I'm using GPT / UEFI boot, how do I set up the EFI partition? it's not clear to me whether it can be an LVM volume or not, etc.
<clever>
joepie91[m]: the ESP must be its own partition, with just the vfat in it
<clever>
joepie91[m]: so you would have 1 partition for ESP, and then 1 for the lvm PV
<joepie91[m]>
clever: aha. any recommendations on size? I'm not sure if all the NixOS kernels end up there
<joepie91[m]>
right
<clever>
joepie91[m]: ESP gets mounted to /boot/ so all kernels land in it
<joepie91[m]>
and I can just create a PV off a partition, without issues or caveats?
<clever>
joepie91[m]: yep
<clever>
joepie91[m]: i would make the ESP 512mb or maybe even 1g
<clever>
joepie91[m]: that will leave plenty of room for this if you choose to use it
<clever>
joepie91[m]: this module dumps an entire nixos installer in /boot/, and adds a grub option for it
o1lo01ol1o has quit [Ping timeout: 258 seconds]
<clever>
so you can then boot that any time you want, to perform major changes or repair-work
<joepie91[m]>
... you have a lot of useful configs
<joepie91[m]>
lol
<joepie91[m]>
this feels like something that should be upstreamed at some point?
<clever>
i used that on my laptop, to move /nix from the / dataset to its own zfs dataset
<joepie91[m]>
right
<clever>
so zfs would stop snapshotting /nix/store/
<joepie91[m]>
I'm thinking of giving /nix/store its own LV
<clever>
joepie91[m]: the grub config on line 17/18 assumes that you are mounting a partition directly to /boot/ and will misbave if you happen to be using /boot on / or other weird things
<clever>
joepie91[m]: i would just go with zfs on all installs
<clever>
joepie91[m]: then you dont have to staticly decide how much space to give to each FS
* joepie91[m]
is not currently planning on using ZFS
<joepie91[m]>
($reasons)
<clever>
joepie91[m]: you can basically think of it like, what if every single file, was an LV within lvm? lol
<clever>
the zfs pool is basically the lvm vg
ricekrispie has joined #nixos
<clever>
and then all filesystems share the free space within the pool
<joepie91[m]>
right, I get that, but a jump to ZFS requires a lot more consideration than I have the time and energy for right now :)
<joepie91[m]>
I have seen a few too many caveats over the years
<joepie91[m]>
hence sticking with ye olde LVM for now
<clever>
swap on zfs has issues, which is why i previously did zfs+swap on lvm on luks
<clever>
but with recent changes to nixos, you can now do (zfs on luks) + (swap on luks)
<clever>
(just 2 partitions)
<clever>
oh, biggest limitation, zfs cant shrink, so you cant reclaim space for things like swap if you wanted
<joepie91[m]>
right, so this is the sort of thing I don't want to have to think about right now :P
<joepie91[m]>
it's for future consideration
<joepie91[m]>
anyway, any issues you can see with a separate LV for /nix/store, other than that the size has to be set explicitly?
nckx has joined #nixos
<clever>
i would put the entire /nix onto that LV
<clever>
so the profiles/db are on the same FS as the stoe
<ashkitten>
lol, dammit. i didn't build my environment before rebooting so it's doing it all in the background where i can't monitor progress
<m1cr0man>
how can I test this change (aka run tests/bind.nix) without messing with my current nix store on my server?
acarrico has joined #nixos
wfranzini has joined #nixos
<clever>
m1cr0man: one min
arahael2 has quit [Ping timeout: 245 seconds]
fendor has quit [Read error: Connection reset by peer]
bennofs has quit [Quit: WeeChat 2.3]
<ashkitten>
me and my whole "hidden costs in getters" rant and im waiting around for my computer to fetch and build dependencies because i put a nix-shell invocation in the startup script