supersandro2000 has quit [Disconnected by services]
supersandro2000 has joined #nixos
<tnks>
I see that Haskell.nix has a documented way to map names to packages for non-Haskell dependencies. Does normal Haskell infrastructure do any intelligent guessing of attribute names? My gut feeling is it does, but I haven't found the code that does that.
<tnks>
dminuoso: yeah, I'm there now. I was looking at the wrong spot of the code before
detran has joined #nixos
<tnks>
dminuoso: that's great. thanks for finding that faster than I was finding it. I've been assuming this was the case, but never saw it documented.
<{^_^}>
[nixpkgs] @Gaelan opened pull request #103517 → <!-- To help with the large amounts of pull requests, we would appreciate your reviews of other pull requests, especiall… → https://git.io/JkLxb
<{^_^}>
[nixpkgs] @djanatyn opened pull request #103518 → xmountains: init at 2.1 → https://git.io/JkLxN
<zeta_0>
hello there, i'm trying to get haskell-language-server to work in emacs, but it's throwing this error? `Command "hie-wrapper --lsp -d -l /tmp/hie.log" is not present on the path`, do I need to manually set an environment variable in a shellHook in my shell.nix config, or does this issue have to do with emacs?
dansho has quit [Quit: Leaving]
<zeta_0>
if this issue is not nix releated, then let me know so that I can elsewhere for help!
<Wulfsta>
I am also looking for feedback on if this is the right way to expose something like this - I did something similar with gmsh but that has not been looked at either
Wulfsta has quit [Remote host closed the connection]
<ezemtsov>
Hi people. For anybody who's familiar with yarn2nix, it is possible to have two version of same library? I need protobufjs 6.8.8 while 5.0.3 is getting installed with grpc library
<ezemtsov>
yarn.lock refers to both, but I can't find the 6.8.8 one in node_modules during the build stage
orivej has quit [Ping timeout: 272 seconds]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<jdnixx[m]>
I mean isn't there some way to like hash the entire running /nix tree and get a sha256 out of it?
<energizer>
djahandarie: that's not what you want tho
<jdnixx[m]>
keyword there is running. already live system
<energizer>
what's the use case?
<jdnixx[m]>
I was just thinking from like a security/public trust perspective
<jdnixx[m]>
actually it was from some discussion about the voting machines
orivej has joined #nixos
<jdnixx[m]>
and someone said "yeah but how could you even be sure the machines are running the same code you're looking at even if it's open-source?"
<jdnixx[m]>
and I was like hm bet you could deploy them on nixos, and then anyone could walk up to a voting machine runnign the software and compute the hash for the running system
<energizer>
sure you can hash the filesystem or whatever but that's not gonna help if the problem is in memory
<jdnixx[m]>
make sure it matches the "official" one
<jdnixx[m]>
oh yeah good point
jonringer has quit [Ping timeout: 264 seconds]
<jdnixx[m]>
hmm
<jdnixx[m]>
what if could could observe it from boot
<energizer>
/run/current-system is a symlink to a place in the store, but that's pretty weak security
<jdnixx[m]>
would there be any way to get "bad" extra stuff in memory that wouldn't show up in the filetree once you're logged in?
<jdnixx[m]>
yee
gthm has quit [Remote host closed the connection]
sangoma has quit [Read error: Connection reset by peer]
coco has joined #nixos
werner291 has quit [Remote host closed the connection]
<dminuoso>
ezemtsov: Did you re-run yarn2nix after the modification?
werner291 has joined #nixos
<ezemtsov>
dminuoso: I run it imlicitly with mkYarnPackage
<dminuoso>
ezemtsov: Well, Id just dive into the source code of yarn2nix to see what it even does
teto has joined #nixos
<dminuoso>
Not much of a choice
<ezemtsov>
nix makes me read much more than I did before
<ezemtsov>
I guess it's a bonus
<dminuoso>
"bonus"
<dminuoso>
Im not convinced "use the source" is a good thing if its obligatory
<dminuoso>
There was a day, when your system shipped with a comprehensive manual that explained everything. Every command, every subsystem, every API function...
<simpson>
On one hand, that's an unfree license. (It clearly denies the Four Freedoms, particularly the freedom to run the code on the user's chosen platform.) OTOH, it might be unenforceable BS.
<{^_^}>
[nixos-homepage] @garbas pushed to redesign-features « make description pop out more »: https://git.io/JktD9
<JaakkoLuttinen[m>
Yeah, I feel reluctant to put software with such stupid license to nixpkgs. Perhaps I'll just package it temporarily for my personal use. 🤔
civodul has quit [Read error: Connection reset by peer]
addcninblue has joined #nixos
<jdnixx[m]>
<JaakkoLuttinen[m "Packaging ElectrumSV. However, t"> what........the
civodul has joined #nixos
<zeorin>
Heya, I'm trying to install several different versions of Firefox on my machine alongside one another and I'd like to be easily able to run these programs simultaneously. I'm using the Mozilla nixpkgs overlay.
<zeorin>
The problem is that there are collisions between the packages
<jdnixx[m]>
leave it to craigvin to turn whatever they're working on completely narcissistic and annoying
<zeorin>
e.g. `collision between `/nix/store/p0fxwqjkbhhgqcqcvfhbqjjcvgsjrmfw-firefox-bin-84.0a1/share/applications/firefox.desktop' and `/nix/store/cm986afw7707hd07dblj6bix0xqklg05-firefox-bin-82.0.3/share/applications/firefox.desktop'`
<betaboon>
i am booting via nfs (nix-store is mounted on nfs). when the network goes down, the nfs-share is down. when the network comes up again it seems like the system doesnt get an ip via dhcp. i guess this is due to the dhcpcd being on the nfs-share and therefore not available in the first place. any suggestions how to tackle this ?
<jonge>
hi! i am trying to get nixos run on a raspberry pi 4. i followed the recommendations in the wiki https://nixos.wiki/wiki/NixOS_on_ARM/Raspberry_Pi_4 but the GPU config there does not work because the config attributes are deprecated. the "fbdev" X config also does not work on nixpkgs unstable - i simply get no desktop, but TTYs work and journalctl
<jonge>
shows no display-manager.service errors. is there any more-up-to-date info?
<clever>
jonge: there has been recent work in the kernel to get proper kms support, and i think x11 just detects that automatically on startup
<clever>
but you need to first enable kms in config.txt
nek0 has joined #nixos
<jonge>
clever: thank you, that sounds helpful. is the change in config.txt something that i do manually or does this happen via /etc/nixos/configuration.nix ?
<{^_^}>
[nixos-homepage] @grahamc pushed 2 commits to master: https://git.io/JkqGx
<{^_^}>
[nixos-homepage] @grahamc pushed 0 commits to prefer-gnome-iso: https://git.io/JkqGp
acarrico has quit [Ping timeout: 240 seconds]
<jonge>
clever: i used this config line and rebuilt the system. i can confirm that this entry is now within config.txt on the boot partition, but it shows no effect after reboot. i was expecting a desktop to come up or get error messages in journacltl, but both doesnt happen.
<deni>
clever: I'm trying to install a module (sops-nix but that's not important)...and I have my niv/sources.json pinned to a version. I then use nixops modify and jq to pin the nixpkgs version in the nixops sqlite database to that exact version. I know we talked about this already but I'm again confused if this in fact makes sure to build everything with that pinned version and deploy to the remote
<deni>
machine?
<deni>
I'm also not too happy that I have to upgrade the entire machine to 20.09 to be able to build a thing I want to use. But I guess there's not way to just build package-A with nixos-20.09 and the rest with nixos-19.09 ?
<clever>
deni: when using something like `nixops modify -I nixpkgs=URL`, that gets baked into the nixops state
<clever>
deni: and then all future builds/deploys will act like you had `-I nixpkgs=URL` in the cmd line
<clever>
so the internal <nixpkgs> used by nixops, will use that version of nixpkgs
<clever>
if you wanted just a single package from 20.09, you can do the same thing you would do on normal nixos
<clever>
basically, just `(import (sources.nixpkgs) {}).hello`, though you need to be careful if darwin machines get involved
jonatanb has joined #nixos
<deni>
clever: back to the pinning...so this makes sure that nixops uses the pinned version and everything I'm referencing in my configuration.nix via <nixpkgs> will use the same pinned version? This is what I was going for but it's been a long time since I set it up so wasn't sure anymore.
<clever>
deni: if you use modify to set `-I nixpkgs=`, then all of nixops and the nixos being built, will use that version
<deni>
clever: as for other vesions....so this mean theoretically that I can have nixpkgs and nixpkgs-foo in my sources.json and then I can (import (sources.nixpkgs-foo) {}).hello right?
<clever>
yep
<clever>
that would let you pick and choose which nixpkgs to use for each program
<clever>
so they can differ from the main <nixpkgs> used to build nixos itself
<deni>
huh...unrelated but this would have made it easier to get an older version of bluez on my machine when I needed it...but at least I learned how to use overlays in the process! :D
<deni>
clever: thanks for your help! This time I'm going to jot down some notes so I don't forget again :D
<{^_^}>
[nixpkgs] @veprbl pushed commit from @bachp to master « k3b: remove qtwebkit dependency (#103164) »: https://git.io/Jkq2s
<{^_^}>
[nixpkgs] @yorickvP opened pull request #103615 → python3Packages.reedsolo: init at 1.5.4 → https://git.io/Jkq2G
cosimone has quit [Read error: Connection reset by peer]
<humancalico[m]>
What is the correct way to install npm packages in Nix apart from those which are in nixpkgs? Like one would do `npm install -g` on other distros.
<dminuoso>
humancalico[m]: How much headaches do you want?
<mindtree[m]>
Hi folks! I'm interested in trying out some new upstream firmware (the ath11k driver to get my xps 13 9310 wi-fi working). I think it's expected to be included in 5.9, however I I'd like to try to include the firmware I'm after now rather than waiting. Is there a central place I can search for kernel modules, similar to how one can search for pkgs in `search.nixos.org/packages`? I've not yet had a go at making a kernel
<mindtree[m]>
module for nix yet, so I'm hoping to find that someone has already beaten me to it for this particular piece of firmware :)
<dminuoso>
humancalico[m]: anyhow, you can use `nix-shell -p nodejs-...` and then use npm from there.
<humancalico[m]>
lol. As minimal as possible
<humancalico[m]>
> humancalico: How much headaches do you want?
<{^_^}>
error: syntax error, unexpected ')', expecting ID or OR_KW or DOLLAR_CURLY or '"', at (string):346:1
mallox has quit [Ping timeout: 256 seconds]
gustavderdrache has joined #nixos
<humancalico[m]>
> humancalico: anyhow, you can use `nix-shell -p nodejs-...` and then use npm from there.
<humancalico[m]>
iirc you can't do global installs using it.
<{^_^}>
error: syntax error, unexpected ',', expecting ')', at (string):345:20
superherointj has quit [Remote host closed the connection]
kalbasit has joined #nixos
<eyJhb>
Is there any way to point nix to a git dir, where it will use two different branches ? Like... `pkgsStable = .... { src = ./nixpkgs; brach = stable; }`?
<jlv[m]>
Anyone know why `systemd.services` appears to have no effect in `containers.*.config`?
superherointj has joined #nixos
endformationage has joined #nixos
hke has quit [Read error: Connection reset by peer]
<neonfuz2>
<eyJhb "Is there any way to point nix to"> I use multiple channels with an overlay, like `unstable = import <nixos-unstable> { config = config.nixpkgs.config; };`
<{^_^}>
[nixpkgs] @danieldk opened pull request #103622 → fedora-backgrounds.f33: init at 33.0.7 → https://git.io/Jkq1C
<neonfuz2>
and I do all my packages in environment.systemPackages, and can use an unstable package like `unstable.packagename`
<neonfuz2>
eyJhb: I don't know if you can add a local directory as a channel, but maybe replace <nixos-unstable> with a path to the git dir and it might work
<{^_^}>
[nixos-homepage] @github-actions[bot] pushed commit from GitHub Actions to master « Update flake.lock and blogs.xml [ci skip] »: https://git.io/JkqSJ
kiasdfasdd has joined #nixos
<neonfuz2>
So I have an open PR for a version bump, and the bot auto made an update for the next version. I need my manual changes but also the bot changes, should I merge the bot changes into my PR and close the bot issue?
kiasdfasdd is now known as pumpy_
gthm has quit [Ping timeout: 260 seconds]
pumpy has quit [Ping timeout: 256 seconds]
<jlv[m]>
This is really weird. Even `systemd = lib.mkOverride 0 false;` has no effect inside `containers.*.config`, and yet, services that change `systemd.services` and the like still work in a container. I can completely mangle the `systemd` attribute in `containers.*.config` and NixOS acts as if it doesn't exist. What is going on with `systemd` in containers?
<{^_^}>
[nixpkgs] @FRidh pushed to staging « buildPython*: remove pythonRecompileBytecodeHook as dependency »: https://git.io/JkqdK
<{^_^}>
[nixpkgs] @vcunat opened pull request #103633 → kresd service: switch .listenDoH to new implementation → https://git.io/Jkqdd
<tpw_rules>
does anybody use gimp on nix on ubuntu? i just installed it and it crashes with an X window error (not really any further information) when i try to copy some bitmap data
<lordcirth>
tpw_rules, what Ubuntu version and nixpkgs channel?
<tpw_rules>
lubuntu 20.04, 20.09 stable
expelledboy has joined #nixos
<lordcirth>
tpw_rules, and you haven't changed to wayland?
<tpw_rules>
i haven't tried into. i can crash it by starting a new gimp instance, creating a new blank image with file -> new, hitting ctrl-a to select all of it, then hitting ctrl-c (or selecting copy from the menus)
<lordcirth>
tpw_rules, your 20.04 is up to date? Rebooted recently?
<lordcirth>
tpw_rules, supposedly someone fixed a similar issue by deleting ~/.config/GIMP
<lordcirth>
You could try moving it
<tpw_rules>
earlier today and earlier today. just tried to delete it and that did not help
<FRidh>
ughh regexes
<ToxicFrog>
regexes \o/
<lordcirth>
FRidh, do you now have two problems?
* ToxicFrog
rolls around in a big pile of Kleene stars
growpotkin has joined #nixos
jonatanb has quit [Remote host closed the connection]
<tpw_rules>
ah hah, it does not crash with small images. like if i create a 16x16 one it's fine
bdju_ is now known as bdju
<lordcirth>
tpw_rules, what's your graphics card and driver?
brettgilio has joined #nixos
bdju has quit [Quit: Reconnecting]
bdju has joined #nixos
<tpw_rules>
intel iris plus 655, i915 kernel module
<FRidh>
you can say that
elibrokeit has quit [Quit: A random quit message]
elibrokeit has joined #nixos
vandenoever has quit [Read error: Connection reset by peer]
<tpw_rules>
iirc i'm using the iris userspace driver
<Siilwyn>
The rust binary in question is Volta, I asked the author too if he had any idea and he mentions that it might be because nixos uses a non-gnu version of libc:
<Siilwyn>
Honestly no idea if that's the case, I did try to run it after running `nix-env -i glibc`
smajdo has quit [Ping timeout: 245 seconds]
jonatanb has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @NinjaTrappeur pushed 2 commits to master: https://git.io/JkqpM
m4tsa_ has joined #nixos
m4ts has quit [Ping timeout: 256 seconds]
axxoldaccount[m] has joined #nixos
<axxoldaccount[m]>
hi everyone, i discovered my regular `nixos-rebuild switch` has maxed out my Raspberry Pi 4's tiny boot partition (100 M) on its sd card
<axxoldaccount[m]>
can't find a way to remove the old kernels and initrd's from there, although i tried quite a few garbage collection commands
<axxoldaccount[m]>
is this a known problem?
<Yaniel>
nix-collect-garbage --delete-old does that I think?
chkno has quit [Read error: Connection reset by peer]
chkno has joined #nixos
<Ke>
I think next bootloader install does the cleanup after gc
<Ke>
not sure
zupo has joined #nixos
<lordcirth>
Looks like it
<Ke>
also the bootloader install is arbitrary code, so some may behave differently
<lordcirth>
axxoldaccount[m], Not sure about Pi, but for grub, you can just run "sudo nixos-rebuild boot" after the GC
<lordcirth>
axxoldaccount[m], Also, you can set boot.loader.grub.configurationLimit in future to cap the usage
FRidh has quit [Quit: Konversation terminated!]
<axxoldaccount[m]>
in this case, i'm using boot.loader.raspberryPi
<axxoldaccount[m]>
<lordcirth "axx (old account), Not sure abou"> well that's sort of the problem: i can't clear the contents of /boot/old/, so nixos-rebaild boot fails
<lordcirth>
axxoldaccount[m], why not? Are they chattr +i?
<lordcirth>
Also, I don't have a /boot/old here?
<lordcirth>
Maybe it depends on bootloader, my kernels are in /boot/EFI/nixos
<axxoldaccount[m]>
no idea, they don't appear to be chattr +i no
<axxoldaccount[m]>
>cp: error writing '/boot/kernel.img.tmp': No space left on device
<axxoldaccount[m]>
lordcirth: you mean just run rm? I was worried that would mess the OS in some way
<colemickens>
for the rpi4, before I expanded /boot, I often had to clear /boot/old, do GC, then rebuild so that it doesn't just immediately restuff /boot too full
<lordcirth>
axxoldaccount[m], the kernels are copied from /nix/store into boot by nixos-rebuild. So long as you don't reboot before running "nixos-rebuild boot" (or switch), it should be fine
<axxoldaccount[m]>
makes sense, thanks lordcirth
<lordcirth>
Then set a low config limit for next time
<manveru>
axxoldaccount[m]: also see `boot.loader.raspberryPi.uboot.configurationLimit`
<manveru>
oh, i'm late :)
<lordcirth>
You were more specific
<axxoldaccount[m]>
thanks, i'd actually started looking at that, lordcirth and manveru :)
<{^_^}>
error: syntax error, unexpected $undefined, at (string):345:1
<{^_^}>
error: syntax error, unexpected $undefined, at (string):345:1
<lordcirth>
Don't use > for quotes :P
<axxoldaccount[m]>
triple backtick then?
<clever>
lordcirth: i tend to use >> when this bot is around
<axxoldaccount[m]>
reading https://nixos.wiki/wiki/NixOS_on_ARM, it says to avoid using a boot partition altogether and to just mark the ext4 partition as the boot one
sigmundv_ has quit [Read error: Connection reset by peer]
chkno has quit [Read error: Connection reset by peer]
<qyliss>
the more focused channels give you a more focused audience, because people who really care about, say, systemd are likely to read all the messages in #nixos-systemd but not all the messages in here
<cirno-999>
the nixos logo is really good for some reason
<gchristensen>
some people have the opinion that you're not in too many channels until you have multiple connections to freenode with different nicknames to get around freenode's joined channel limit
<cirno-999>
.shrug
<DigitalKiwi>
console irc client i only have so many keys! :D
<cirno-999>
what is the limit?
<DigitalKiwi>
heh you should see my znc output when i reconnect ;_;
mananamenos has quit [Ping timeout: 246 seconds]
<qyliss>
cirno-999: 120 I think
<qyliss>
I'm in about 110 atm
werner291 has joined #nixos
aidenholmes has joined #nixos
chkno has quit [Read error: Connection reset by peer]
chkno has joined #nixos
alp has joined #nixos
heyitsrama has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @orivej-nixos pushed 3 commits to master: https://git.io/JkmYU
pinkieval has joined #nixos
<{^_^}>
[nixpkgs] @sikmir opened pull request #103647 → qvge: init at 0.6.1 → https://git.io/JkmYk
komasa has joined #nixos
fendor_ is now known as fendor
drangon has joined #nixos
<drangon>
Is this the right place to ask a quick question?
<ivan>
sure
<drangon>
I am attempting to build a rust project where one of the dependencies links against glibc. The linker seems to be choosing glibc 2.31 even though the package requires 2.32 and I have 2.32 installed.
<drangon>
error: /nix/store/9df65igwjmf2wbw0gbrrgair6piqjgmi-glibc-2.31/lib/libc.so.6: version `GLIBC_2.32' not found (required by /home/kylerosenberg/repos/vulkan-graphics/target/debug/deps/libconst_fn-60e096b97088955b.so)
shibboleth has quit [Quit: shibboleth]
lunik17 has joined #nixos
<drangon>
I tried to run it in a nix-shell with nix-shell -p glibc but that didnt seem to help
lunik1 has quit [Ping timeout: 246 seconds]
lunik17 is now known as lunik1
<ivan>
my first instinct is to fix it to not require glibc 2.32
<ivan>
it looks like nixos 20.09 has 2.31 and master has 2.32
<ivan>
you can try the unstable channel or pointing nix-shell at a checkout of nixpkgs:master with -I or NIX_PATH
<drangon>
ivan it actually gives an error error: attribute 'glibc' in selection path 'glibc' not found
alp has quit [Ping timeout: 272 seconds]
<drangon>
which is strange because it is explicitly installed in both my user level env and my configuration.nix
<clever>
drangon: you want `nix-build '<nixpkgs>' -A glibc`
<clever>
,libraries ivan
<{^_^}>
ivan: Don't install libraries through nix-env or systemPackages, use nix-shell instead. See https://nixos.wiki/wiki/FAQ/Libraries for details.
johnnyfive has joined #nixos
<drangon>
with import <nixpkgs> {};stdenv.mkDerivation { name = "dev-environment"; # Probably put a more meaningful name here buildInputs = [ pkg-config glibc ];}this still did not help
<drangon>
spacing got eaten
<drangon>
does this have code formatting
<drangon>
```
zebrag has quit [Ping timeout: 272 seconds]
zebrag has joined #nixos
<colemickens>
is manually hashing a NAR documented?
<colemickens>
I guess probably in the paper. I am hoping to emulate `nix sign-paths` without having to write the key to disk again.
heyitsrama has quit [Remote host closed the connection]
Morfio has quit [Quit: This computer has gone to sleep]