<dxtr>
How do I generate the sha256 that pkgs.fetchpatch expects?
<clever>
dxtr: simplest option is to just have nix generate it
<clever>
,tofu
<cole-h>
,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. See: tofu-vim
<{^_^}>
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. See: tofu-vim
<cole-h>
clever: Too fast for me haha
chagra has quit [Ping timeout: 256 seconds]
<clever>
cole-h: :D
chagra has joined #nixos
butcher_ has joined #nixos
smatting has quit [Ping timeout: 246 seconds]
<dxtr>
hmm
<dxtr>
How do I clean up a failed nixos-rebuild switch --upgrade?
<clever>
dxtr: how did it fail?
<dxtr>
The patch that I am providing didn't apply to the new kernel
<dxtr>
And now it won't re-download the patch because it already did it once
<dxtr>
So I can't update the checksum and it's still using the old patch
<clever>
dxtr: if you claim the sha256 is the same, it will believe you, and not bother to re-download the file
<clever>
dxtr: use a different sha256
<dxtr>
oh
smatting has joined #nixos
<dxtr>
yay it's building
ex-parrot has joined #nixos
<dxtr>
Maybe I should run upgrade more often :P
<dxtr>
This is gonna take forever
<ex-parrot>
hello all. I'm looking at the existing powerdns derivation and I'd like it to run powerdns as a different user. currently it forces it to run as "nobody", what's the most elegant way to change that behaviour?
<ex-parrot>
I can override systemd.services.pdns.serviceConfig.ExecStart but I don't think I can access the value of the configDir variable that needs to be interpolated in to that
<clever>
ex-parrot: simplest option then is to edit the module
<clever>
monokrome: but with kexec, you just upload a tar, and run 2 shell commands, and boom, its now running nixos from a ramdisk
<clever>
monokrome: correct
<monokrome>
seems nice
<clever>
monokrome: and once your in that nixos ramdisk, you can just format the disk and install as normal
sigmundv_ has joined #nixos
<monokrome>
what's the limitation?
smatting has quit [Ping timeout: 264 seconds]
<monokrome>
clever: kexec will let me unlock the root disk and repartition it?
<clever>
monokrome: kexec can be tricky to use when the server cant dhcp, and if you make any mistakes during the install, it can leave the server unable to boot
<clever>
monokrome: yeah
<monokrome>
oh that's awesome
hlisp has quit [Remote host closed the connection]
<monokrome>
this reminds me that I need to get KVM working... DMESG says my bios won't allow it :D
<clever>
monokrome: kexec doesnt need kvm
<monokrome>
yeah, I know :D
<monokrome>
reminded me of it though
<clever>
monokrome: my kexec script also includes the justdoit script, which can install your entire os in 1 command, once configured right
<monokrome>
=^.^=
<monokrome>
w/ userdata scripts?
<clever>
monokrome: the entire config is baked into the tar file
<monokrome>
Yeah, so it can use cloud-config to do it then?
<clever>
monokrome: it could be modified to do that
<monokrome>
ah cool
<colemickens>
kexec seems easier than involving cloud-config, imo
<monokrome>
I think cloud-config can receive a tar file containing scripts
<monokrome>
colemickens: separate things
<colemickens>
I know
<colemickens>
I hate cloud-config.
<clever>
monokrome: in this case, the kexec tar contains the entire os, and a single shell command will reboot into it
<monokrome>
Like, w/ cloud-config you can launch a ubuntu server that initiates a nixos reinstall on boot w/out direct user access...
<colemickens>
I know. cloud-config is one of the worse things I've ever had to interact with.
<colemickens>
I had to support service outages because of it.
<monokrome>
ok?
<colemickens>
Sorry. I'm not contributing anything useful.
alp has quit [Ping timeout: 240 seconds]
<monokrome>
something has to run the kexec :P
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JfCt2
* colemickens
agrees
leo60228 has joined #nixos
<clever>
monokrome: for aws, i was thinking you just generate an AMI with the same kernel/initrd and grub
<leo60228>
is there a reason dpkg doesn't add to unpackCmdHooks like unzip and p7zip do?
<monokrome>
Doesn't AWS have official AMIs for NixOS?
<clever>
monokrome: slightly modified to get ssh keys from the aws api
<clever>
monokrome: my problem with the official nixos AMI, is that it is never the same nixpkgs rev that i am deploying with nixops
<clever>
monokrome: so it will always double in disk usage on the first deploy, and likely need a reboot to update the kernel
morgrimm has joined #nixos
<monokrome>
>.<
* monokrome
is just about to try to learn nixops again
<clever>
monokrome: run `nixops create -d house deployments/house.nix` to make a deployment, then `nixops deploy -d house` to deploy it
<clever>
monokrome: nixops will ssh into every machine (IP's on lines 12 and 16), and force it to run a new build of nixos, which comes from nas.nix and router.nix
<clever>
if you want to keep things simple, just treat nas.nix like a configuration.nix file, and your done
<monokrome>
clever: Ah, it doesn't launch machines for you?
<monokrome>
Just manages ones you already launched?
<clever>
monokrome: this is the "none" backend that uses pre-existing hw
<clever>
monokrome: but if you use another backend like ec2, it can launch machines for you
<monokrome>
oh cool
<leo60228>
if you don't want to use nixops it looks like you can import <nixpkgs/nixos/maintainers/scripts/ec2/amazon-image.nix> and get an AMI from config.system.build.amazonImage
<clever>
monokrome: if using a backend, you basically just set `deployment.targetEnv = "ec2";` and then give nixops some keys (many ways), and your done
fabianhjr has joined #nixos
fabianhjr has quit [Max SendQ exceeded]
m0rphism has joined #nixos
fabianhjr has joined #nixos
philr has joined #nixos
<monokrome>
clever: What kind of NAS do you use? I've been trying to find a case for one but they're expensive =^.^=
<monokrome>
My NAS is a Drobo5N but wanting to make one w/ NixOS
fabianhjr has quit [Ping timeout: 244 seconds]
<clever>
monokrome: my nas is just a random x86 motherboard in a case, with a heap of sata disks in it
Fare has quit [Ping timeout: 246 seconds]
<clever>
monokrome: the motherboard also came with hdmi out onboard, so its also acting as a plex frontend
ultranix` has quit [Ping timeout: 256 seconds]
<monokrome>
clever: so would your script let me install NixOS over my current Ubuntu?
<clever>
monokrome: yeah
<monokrome>
cool :o
<monokrome>
That's what I want my NAS to be too
<monokrome>
right now I have an intel NUC that uses my DROBO5N over NFS
<monokrome>
I don't think the Drobo has the hardware to do the transcoding live TV / DVR stuff though
<monokrome>
cases w/ a lot of HDD slots are expensive though
<clever>
monokrome: one min...
fabianhjr has joined #nixos
<exarkun>
I am so so so confused. I have a project where `src = ./.;` and ./. is a git checkout and ... some ... source files that end up in result are not from the git checkout, they're from an older revision of the repo
<exarkun>
I put a syntax error into test_controller.py but nix-build succeeded anyway
<exarkun>
before that, I was looking at coverage reports and they show nonsense that looks like it comes from applying coverage data to the wrong source file for presentation
<exarkun>
Then I diff'd my checkout version of controller.py and the one that ended up in ./result/ and they were pretty different
<clever>
exarkun: try adding postUnpack = "exit 1"; to the derivation, then build it with --keep-failed, and diff the src in the /tmp dir to the original
<quinn>
exarkun: you can `nix repl` and then `:l file.nix` and then use tab completion (iirc) to double check the phases are the way you think they are
<energizer>
what does an "env" refer to in nix?
<quinn>
and then i like to put random ls and echo statements in my phases if they aren't working right to make sure they're happening
Tndc has quit [Remote host closed the connection]
Emantor has joined #nixos
<exarkun>
I think the postUnpack and postCheck were the first nix expression changes I made since I started observing the problem
<exarkun>
So my hypothesis would be that some cache got corrupted somehow
<exarkun>
But I am now too fried and I no longer have a reproducer to dig into that
<clever>
exarkun: every time you run nix-build, it will hash ./. and copy it into /nix/store/
<exarkun>
clever: That's sort of what I thought
<clever>
exarkun: what does `nix-store --verify --check-contents` say?
<exarkun>
clever: So I wonder if such a failure would have to involve filesystem or disk failure of some sort
<spinlock[m]>
This might be a naive question, but how do you figure out what packages should go into `services.dbus.packages`?
<eadwu[m]>
I'm not understanding, the build only needs to be purely evaluated, you can set nix.nixPath to set <nixpkgs> to point to your nixpkgs flake
<eadwu[m]>
And that resolves the runtime issue
<hexa->
spinlock[m]: Packages whose D-Bus configuration files should be included in the configuration of the D-Bus system-wide or session-wide message bus. Specifically, files in the following directories will be included into their respective DBus configuration paths: pkg/etc/dbus-1/system.d pkg/share/dbus-1/system.d pkg/share/dbus-1/system-services pkg/etc/dbus-1/session.d pkg/share/dbus-1/session.d
<clever>
hexa-: usually, anything that is offering some kind of service via dbus
<hexa->
^ spinlock[m]
<spinlock[m]>
I guess I'm wondering why the configurations for those apps don't automatically add them. it feels like one of those things that you "just have to know"
<clever>
spinlock[m]: the nixos module for the app should be doing it
<clever>
when you do programs.foo.enable = true;
sigmundv_ has quit [Ping timeout: 256 seconds]
<spinlock[m]>
hum hum. probably a newbie error on my part, but it took me ages to get nextcloud to be able to store its secrets in gnome-keyring. I had to explicitly add `services.dbus.packages = [ pkgs.gnome3.gnome-keyring pkgs.gcr ]` to my config
<clever>
spinlock[m]: it also configures other things, like a setuid binary
<spinlock[m]>
hmm. I had `home-manger.users.spinlock.services.gnome-keyring.enable = true` guess that's different
<clever>
spinlock[m]: home-manager cant touch the system dbus config
<spinlock[m]>
thanks for the tip! I'm going to look into these packages and see if I can tidy up my config a bit. I was sort of enabling everything I could until it worked
<spinlock[m]>
ohhh
<euank>
Presumably that's a systemd --user thing, and nextcloud is running not as your user, but as the system I'd guess
<clever>
home-manager is limited to your home folder
<euank>
Meaning it wants to talk to stuff on the system bus
<euank>
It would be removing the current channel and adding a new one with the alias nixos, so i.e. `nix-channel --remove nixos; nix-channel --add https://nixos.org/channels/nixos-20.03`
<euank>
but if you're on unstable, there's not an equivilant because unstable's more like a rolling-release (where dist-upgrade doesn't make sense)
<eadwu[m]>
The file in `nix-packaging` can already be evaluated without any changes so it should build fine (which doesn't because it doesn't create a bin output but that's a problem originating from morph) while runtime would use <nixpkgs> unless I'm getting something wrong
pxc has joined #nixos
<pxc>
simple question about Nix flakes as currently implemented: is there any way to refer directly to paths (not necessarily packages or attributes) in Nix code comparable to the old way with Nix paths?
<pxc>
like the equivalent of <nixpkgs/path/to/some/file>?
<pxc>
what I'm wondering though is if I'm using flakes instead of channels for my local NixOS configuration, how do I refer to whatever copy of Nixpkgs my NixOS configuration does, if I want to just directly run a bit of Nix via like nix-instantiate --eval to count how many things I have in systemPackages, or something like that?
<kyren>
Are melpa updates regularly backported to stable releases? Is that something that I could potentially help make happen?
<bqv>
eadwu[m]: again, I use more than one nixpkgs in my flake, and I can't access the inputs without actually evaluating the flake. I don't understand how you think this can work
<bqv>
Have you tried it? Or have you even tried flakes?
<bqv>
Not trying to be rude, just it's frustrating me a bit
<bqv>
pxc: each input is just a source derivation, you cat browse into them as you would otherwise. And you can set your system flake registry's nixpkgs to the one used for the system, so you can use that directly or (getFlake "nixpkgs") it
<Miyu-saki>
buildGoModule does not support overriding
<Miyu-saki>
Is it simple enough to add overriding to it?
tsrt^ has quit []
<cole-h>
bqv: So would `<nixpkgs/pkgs/some/package/path>` be equivalent to `"${getFlake "nixpkgs"}/pkgs/some/package/path"`? If not, is there any equivalent? (A little slow today so apologies if you actually just answered this in your above response)
symphorien has quit [Ping timeout: 260 seconds]
<bqv>
cole-h: yup, should work
symphorien has joined #nixos
<pxc>
ah, ok. thanks, bqv!
<cole-h>
Cool, thanks. Is there any shorthand for that (yet)?
<bqv>
Well, in the context of a flake you'd presumably just have threaded the input down anyway, but otherwise no
<bqv>
It is a bit unergonomic
<cole-h>
I see. I hope somebody makes an "Idiot's Guide to Flakes" sometime soon :P
<bqv>
:p
symphorien has quit [Ping timeout: 260 seconds]
hsrt^ has joined #nixos
symphorien has joined #nixos
proofofkeags has quit [Remote host closed the connection]
<pxc>
hm. I see that adding a flake to root's registry doesn't make it available to all users. Is there any way to add to the global Nix flakes registry, or does it just mean whatever's on the one hosted at nixos.org?
<pxc>
I used to manage some channels in my NIX_PATH with `sudo -H nix-channel ...`, so they'd be available to all users on the system
<pxc>
I'm wondering if there's something equivalent here or not really
M3DPACTE[m] has joined #nixos
hlisp has quit [Remote host closed the connection]
bandali has quit [Quit: migrating servers]
hlisp has joined #nixos
shafox has quit [Remote host closed the connection]
shafox has joined #nixos
butcher_ has quit [Remote host closed the connection]
<bqv>
There's the system registry, pxc
<bqv>
nix.registry
<pxc>
aah! thanks
<bqv>
np
<pxc>
so we're losing a layer here
<bqv>
,locate libgccjit.so
<{^_^}>
Couldn't find in any packages
<pxc>
we used to have user, default, and system. Now just user and system
<bqv>
grr
<bqv>
pxc: there's also the online registry
<pxc>
which also means users on non-NixOS don't have a way to make flake registry entries available to all users
<bqv>
If nixpkgs isn't in user or system registries it will still resolve from online
<pxc>
It kinda seems like the flake registry layering should match the layering for Nix profiles
<pxc>
that's a neat failsafe
waleee-cl has quit [Quit: Connection closed for inactivity]
chagra_ has joined #nixos
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #87324 → [20.03] build-bazel-package: switch hash mode to “flat” → https://git.io/JfCcj
<pxc>
bqv: ah so you're also using a config based on nixflk? I'm currently reading through it and just playing with some things and trying to decide how closely I want to follow it
chagra has quit [Ping timeout: 260 seconds]
<bqv>
Use it as you like :p theres a few things I'd like to change about it but haven't got around to
luna has joined #nixos
<bqv>
I think I might need to make a PR for gcc
<bqv>
That's terrifying
<pxc>
what's the `flake = false;` stuff I see in your flake.nix? is that so you can use the flake machinery to refer to source code repos and other objects that don't have flake.nix inside them?
<luna>
Anyone else use GPG here?
<pxc>
bqv: hahaha why might you need to change gcc?
<pxc>
luna: I do :)
<bqv>
pxc: yep
<luna>
pxc: Have you successfully made a private key or imported one?
<pxc>
bqv: ah you mean a PR for the gcc package in Nixpkgs. Is this for the new AOT Elisp compilation stuff? I think there's some work on that already underway
<luna>
What settings did you need to do that in you nix config? I keep getting "no pinentry" errors.
<bqv>
pxc: ah, right!
<bqv>
Yeah that was my plan
<bqv>
I've already built emacs
<bqv>
Just looking for gccjit
<pxc>
luna: if you're on NixOS, you want to enable GPG via the NixOS module instead of manually installing it. That'll take care of pinentry for you: https://nixos.org/nixos/options.html#programs.gnupg.agent.enable
<pxc>
luna: once you do that, you'll want to gracefully shut down your GPG agent and then start the user-level systemd unit that NixOS enables for the GPG agent... or log out and back in again :)
<luna>
pxc: haha, I guess the logout/in is what I was missing. I'll give it a try. Thank you!
<bqv>
Ah
<pxc>
you're welcome! yeah, GPG aggressively starts agents all its own now, and then the systemd unit can't start because the one that the gpg cli tools started hogs the socket
<colemickens>
I had it jammed up the other day, had to restart things like 3 times, finally in the right sequence, before everything was just right.
<colemickens>
the gpg-agent daemon that was launching on its own was not configured the same as the user unit
<pxc>
it's definitely a source of a lot of confusion and frustration when you don't expect it. And I've heard from some people that try to use GPG in automated workflows that the presumption of a running agent can be very annoying there
<pxc>
colemickens: yep. A friend of mine did that same thing but killed the cli-tools-started agent ungracefully (like I told him not to, lol) and something weird happened with the file descriptors gpg-agent opens and long story short, we couldn't start either gpg-agent anymore after that
<pxc>
only figured out it was really that after a reboot
luna has quit [Ping timeout: 264 seconds]
<colemickens>
YES!
<colemickens>
that was it!
<colemickens>
I had to also delete stale socket files or something like that
<colemickens>
either way, one of many experiences that inspires me to try to get rid of gpg in my life if I can
<pxc>
yeah, you have to close the gpg agent with some special command or it'll leave open socket files behind and possibly file descriptors that idek how to close after that
<pxc>
I use GPG with pass, and I really like it for that
<pxc>
but it's not a very ergonomic tool at all
<hyper_ch>
pxc: pass <3
<pxc>
it's nice to have great password management with multiple frontends and no dependency on The Cloud(tm)
<bqv>
I still have pinentry issues over ssh
<hyper_ch>
what issues?
<bqv>
I don't even wanna try and solve them
<bqv>
hyper_ch: pinentry fails, presumably cause no gui or agent can't be reached
<bqv>
I don't know or care at this point
<hyper_ch>
you're just doing it wrong :)
<bqv>
hyper_ch: I believe you
<pxc>
it works on my system, using gpg as my SSH agent and ssh agent forwarding
<hyper_ch>
I have no issues only when Konsole is too small
<pxc>
I learned a little while ago that ssh agent forwarding might be kinda bad, though? lol
<hyper_ch>
then the ncurses (?) pinentry won't appear
<bqv>
I don't use ncurses pinentry
<bqv>
I use emacs or gtk
<bqv>
(somehow it varies)
<bqv>
And its all so fragile that I just go with it
<pxc>
bqv: where's the action for your bqv-flakes branch of home-manager? I want to add home-manager to my config but I don't really wanna have to use a special branch or fork. Are the flake changes required to make use of home-manager as a flake.nix input, or is it just convenience stuff that allow you to refer to flakes inside your home-manager config or something?
<hyper_ch>
I have pinentry, pinentry-qt and gnupg installed... no special setting
<bqv>
I'd like if it did fallback to ncurses but something something cake eating
<bqv>
pxc: I use the nixos module for hm, not the app
<bqv>
I prefer having unified representation of system state
<bqv>
Makes reasoning way easier
<monokrome>
whoa nixops is Python
<pxc>
bqv: ah. my old config gets reused on NixOS and on non-NixOS Linux and on Darwin. So I want to do the same here, and I'd rather the version of home-manager pulled in be the same for all of them
* monokrome
mind blown
<bqv>
pxc: oh, misread, sorry
<bqv>
Rycee has flakes on a backburner
<bqv>
So we community maintain a branch
<bqv>
monokrome: there's morph if you prefer :p
<pxc>
is the community for flakes users organized anywhere, or just kinda incidentally here and on the Discourse and so on
<{^_^}>
[nixpkgs] @vbgl opened pull request #87326 → ocamlPackages.ipaddr: remove at 2.6.1 → https://git.io/JfCCl
<bqv>
Yeah mostly here
WoC has quit [Remote host closed the connection]
<pxc>
what does the flake support in the home-manager repo add at this point? why not just use the flake machinery to get a reference I can refer to like with getFlake and then import the module manually?
<iqubic>
Are there any EXWM users around? I'm trying to set up a compositor on EXWM and I'm not sure how to make that work.
<pxc>
oh nvm, I'm looking at some of the commits and I get it
<bqv>
HM internally is impure, so references channels, and will fail without those patches
<pxc>
you had to remove NIX_PATH references
<bqv>
Yes
<bqv>
iqubic: never tried a compositor
<iqubic>
bqv: Ah, that's a shame.
<pxc>
wait, you mean you've never used a compositing window manager at all? or is there a sense of 'compositor' being used which I don't know here?
hlisp has quit [Remote host closed the connection]
<iqubic>
I think he means that he's never used EXWM with a compositor
<bqv>
Some window managers are so small that the compositing functionality is a separate app
<pxc>
when compositing was still new to Xorg, compositors were often separate too
<bqv>
Ahh the good old days of compiz
hlisp has quit [Ping timeout: 256 seconds]
<iqubic>
EXWM doesn't offer built in compositing, and I'm trying to get that working.
<monokrome>
people don't use compiz any more? O_o
<pxc>
I was like 13 when Compiz and Beryl came out and I thought it was the coolest shit ever
<monokrome>
lol
<monokrome>
compositing was all new and fancy
<iqubic>
so, should I use a compositor?
<monokrome>
do you care about widnow opacity, shadows, or other fancy effects?
<bqv>
I don't. It seems like a waste of time
<iqubic>
I don't really.
<monokrome>
then no :D
<bqv>
But people do like their shinies
<monokrome>
opacity can be useful sometimes
<iqubic>
I'm just getting these weird graphical issues that should up all some time.
<pxc>
my favorite fancy effect is built into KWin, and it's 'organize windows'. I like it because it presents all your windows visually and searches as you type. Sometimes it's the nicest way to switch to the desired window
<monokrome>
like I'd use it if I didn't use a tiled window manager so that I can read text behind terminals
<pxc>
but for it to be sane, the animation has to include moving windows around and scaling them and so on
<iqubic>
I hate these graphical issues I'm getting.
<monokrome>
what are they?
<iqubic>
I'm not sure of the best way to describe them.
<monokrome>
:o
<monokrome>
Is it ghosting?
<bqv>
Tearing?
<pxc>
one other useful WM compositing thing is thumbnail view, basically picture-in-picture for arbitrary windows. It can be kinda nice if you want to keep an eye on something (like maybe just to see that there's still new output in a terminal) but you don't need to see the whole thing on your screen
<pxc>
other than that most uses I've seen compositors put to is pretty cosmetic
<iqubic>
Yeah. It's screen tearing.
m0rphism has quit [Ping timeout: 256 seconds]
<monokrome>
I wonder if there's a way to compile Compton into DWM so that I don't need to run two apps
<monokrome>
cuz I just ran it and having semi-transparent terminal w text behind it is nice in fullscreen tiling mode
<iqubic>
I know this was is a graphical issue because it's happening in several different places.
<bqv>
iqubic: frobnicate your x drivers
<iqubic>
However, I was able to fix the issue in firefox by setting use WebGL renderer to true.
<iqubic>
But I'm still getting these issues in other applications.
<iqubic>
How do I frobnicate my x drivers?
<monokrome>
webgl renderer shouldn't effect every page
virus_dave has quit [Quit: virus_dave]
<pxc>
iqubic: does your window manager support vsync?
<monokrome>
oh, cool... New one for me to try =^.^=
<iqubic>
Only use it if you like Emacs.
<monokrome>
o :(
<iqubic>
It's the Emacs X Window Manager.
<monokrome>
oooh I've heard of that
<monokrome>
cool =^.^=
* monokrome
is a Vim girl >.<
<iqubic>
I didn't know you were a girl.
Maxdamantus has quit [Ping timeout: 256 seconds]
<aditsachde36>
Hey good morning. I'm attempting to build a set of nixos configurations via CI/CD using github actions. I have a couple of questions about nixpkgs and versions and stuff. Since I'm building everything on a single server and using nix-copy-closure to copy the build output to a nixos server, the nixpkgs channel on the nixos server shouldn't matter
<aditsachde36>
right?
<monokrome>
O_o
domogled has quit [Remote host closed the connection]
domogled has joined #nixos
Maxdamantus has joined #nixos
<aditsachde36>
Additionally, I have read that I should be pinning nixpkgs. Do I just choose an arbitrary nixpkgs hash on the channel that I want to track or is there any specific version to select?
palo1 has joined #nixos
alp has joined #nixos
<monokrome>
woot - got DWM to run initPrograms =^.^=
<pxc>
also bqv, monokrome: same-- I use both Vim and Emacs. My Emacs config is very vimmy (and very clunky), though
<bqv>
Building off it now, gcc is a beast though. I wonder if I can steal a build from borg
<iqubic>
So, should I reboot now?
<pxc>
alright it's time to stop letting myself procrastinate/wallow in decision fatigue at all the ways I _could_ reorganize my Nix configs
<pxc>
let this NixOS install inaugurate the new layout, based on nixflk with all the concrete examples taken out and maybe a few other changes
<pxc>
(I've been on IRC tonight from the NixOS install media lol)
alp has quit [Ping timeout: 265 seconds]
<iqubic>
Have you installed NixOS yet?
<pxc>
wait, since when do partitions have names as well as labels
<pxc>
iqubic: yeah, I've been running NixOS for a few years. This is a new computer and I decided to use the occasion to migrate my config to one of pinned sources and a different layout
<monokrome>
compiz became compiz-something which became compton which became picom
<monokrome>
TIL
<iqubic>
What options should I use?
<monokrome>
picom is the one that is still maintained
<monokrome>
I don't know what other compositors are out there
<iqubic>
picom looks great. What options should I try?
<monokrome>
not sure, mine isn't running w/ options tbh
<iqubic>
I see.
<monokrome>
trying that in a bit
<iqubic>
Should I try enabling vsync?
<monokrome>
in what?
<monokrome>
oh, does the compositor enable that?
<monokrome>
If your issue is tearing then you should enable vsync
<monokrome>
tearing is where the rendering pipeline flips your framebuffer mid-render, so you end up with a line through the screen for a frame that has half the old frame on one side of it or similar
<iqubic>
Yeah, there's an option to let picom enable vsync
<monokrome>
vsync solves that problem at the potential expense of lower performance depending on hardware
Yakulu[m] has joined #nixos
<iqubic>
I see.
<monokrome>
Is that the visual errors you're seeing?
<iqubic>
Yeah, actually.
<monokrome>
then try w/ vsync
<monokrome>
you don't need to restart your VM, you can just run it
<monokrome>
VM = WM
<monokrome>
sorry, sleepy lol
<iqubic>
I'm going to restart my WM anyways, because I want to get the latest kernel
<kaychaks_riot>
how to take care of the error `firefox is already running but not responding` in sway on nixos ? have put all necessary wayland specific env vars
<monokrome>
Is Firefox running on your desktop?
<kaychaks_riot>
yes
<kaychaks_riot>
it is running when I am trying to launch again either via rofi or in a diff workspace I get this error
<{^_^}>
[nixpkgs] @FRidh pushed commit from @r-ryantm to master « cutelyst: 2.9.0 -> 2.10.0 »: https://git.io/JfC4y
<kaychaks_riot>
so if I launch fresh firefox from rofi then it launches just fine and I could create new window. now I switch workspace and again try to launch firefox via rofi like first time hoping that it will just launch a new window for the existing process, but then I get the error
<{^_^}>
[nixpkgs] @FRidh pushed commit from @mmilata to release-20.03 « cups: 2.3.1 -> 2.3.3 »: https://git.io/JfC0o
<staydownchamp>
i'd be willing to backport the fix, but if it's not going to be supported, i guess it makes more sense for me to migrate to a status bar
<adisbladis>
I'm happy to hack on the automatic updater bits if I just have an expression to base things on
<bkv>
adisbladis: ah, sure, see bqv/nixos
<bkv>
it's not all merged/pr'd yet
<bkv>
i fetched bhipple's gccemacs expression, and the libgccjit from pr 87157
<bkv>
i only tested opening it with no init.el so far, cause i'm using exwm so using my full config would be complicated
justan0theruser has quit [Ping timeout: 240 seconds]
<bkv>
but i might restart now and see what happens
<adisbladis>
=)
<adisbladis>
Ok I'm gonna wait a bit, at least til 87157 is usable on the unstable channel
<niso>
bkv: just noticed that you switched your nixos configuration to flakes. by chance you outlined somewhere your experiences?
<adisbladis>
Once that's done I want to add gccemacs to the overlay
<bkv>
yup, that was my plan this morning, until i found out gccjit was a requirement
<{^_^}>
[nixpkgs] @andir opened pull request #87357 → nixos/tests/gitdaemon: fix spurious test failures due to flaky network → https://git.io/JfCVz
<bkv>
niso: bleh, i have done several times, it's pretty smooth and i wouldn't ever go back at this point, but at the time i migrated it was a little bit bumpy
<bkv>
others have migrated since, e.g. emily, drakonis, and a few others
<asbachb>
Mhh I'm not there yet: `${import "${toString modulesPath}/../lib/from-env.nix" "NIXOS_CONFIG" <nixos-config>}` basically I want the content of used config. To write it out via `environment.etc."nixos/lxd.nix".text = ''`.
<bdju>
anyone else run into some text fields in firefox having both dark text and background so that it's unreadable without highlighting it?
zarel has joined #nixos
<bdju>
I just tried changing from dark theme to default theme to light theme, all have the same problem as far as I can tell. I haven't restarted the browser between the theme changes, though.
<bdju>
Maybe linked to my GTK theme.
<{^_^}>
[nixpkgs] @SFrijters opened pull request #87369 → urserver: init at 3.6.0.745 → https://git.io/JfCou
NeoCron has joined #nixos
hlisp has quit [Ping timeout: 260 seconds]
o1lo01ol1o has quit [Remote host closed the connection]
asheshambasta has quit [Ping timeout: 252 seconds]
hlisp has quit [Remote host closed the connection]
<asbachb>
infinisil: If I get it right `config.nixpkgs.config` is the filled datastructure. I want to revwrite the config to a static destination like /etc/nixos/used-config.nix.
<infinisil>
asbachb: Not really possible with NixOS unfortunately
<infinisil>
E.g. think what should happen if somebody assigns `nixpkgs.config.packageOverrides = pkgs: {};`
<infinisil>
You'd have to turn a function into a string representing it, which is not possible in a purely functional language (I'm pretty sure)
<Athas>
For the last week or so, trying to install 'steam' fails with a bunch of 404 errors as it tries to download some .deb packages. Is anyone else seeing this? I couldn't find anything on the nixpkgs issue tracker.
<Athas>
Ah, wait, looks it was reported recently, but the fix hasn't percolated to unstable yet.
<infinisil>
I do like the idea of being able to serialize a whole NixOS configuration
<adisbladis>
asbachb: Ugh, that's a cursed option.
hlisp has quit [Remote host closed the connection]
<asbachb>
Basically I generate a system image for lxc, but it needs some configuration to be able to do nixos-rebuild. So I thought about copying the configuration used for generation and link it to the default `/etc/nixos/configuration.nix`.
<asbachb>
*lxd
torporish has quit [Remote host closed the connection]
<the_pumpkin_man[>
Hello, I am trying to create a nix derivation, I tested it using `nix-shell path/to/nixpkgs -A derivation --pure` and it works. Now I try `nix-build path/to/nixpkgs -A derivation` and it complains with "cannot open git-upload-pack". Why does it work in nix-shell and not in nix-build ?
ottidmes_ has quit [Quit: Configuring ZNC, sorry for the joins/quits!]
<{^_^}>
[nixpkgs] @FRidh pushed 5 commits to python-unstable: https://git.io/JfCKX
<the_pumpkin_man[>
It might be an issue with networking, is it allowed within nix-build ?
chagra_ has joined #nixos
Maxdamantus has quit [Ping timeout: 240 seconds]
zupo has joined #nixos
Maxdamantus has joined #nixos
alp has quit [Ping timeout: 246 seconds]
jonge[m] is now known as jongeOld[m]
o1lo01ol1o has joined #nixos
<the_pumpkin_man[>
Networking seem to not work, tried with a shitty wget and it can't resolve hostnames ... That's a problem. How can I enable networking while building ?
<lukego>
Hey does anyone fancy a chat about Common Lisp packaging? I'm using ql2nix - an abandonware (?) out of tree solution to packaging Quicklisp - and it's been good but doesn't support modifications to packages e.g. patches out of the box. I'm wondering whether to hack that or whether the in-tree quicklisp support is actually more suitable. maybe someone works with this stuff?
<balsoft>
the_pumpikin_man: You have to use a fixed-output derivation for that
<MichaelRaskin>
It's a wrapper for QL in the style of our cargo wrapper
<balsoft>
Nix builds are sandboxed by design
ottidmes has joined #nixos
<the_pumpkin_man[>
balsoft: how do I do that ?
<the_pumpkin_man[>
It is actually my first derivation, I'm discovering things
<maralorn>
I have an expression which I think is a derivation. I think that because I can build it with nix-build -E '<expr>'. But when I try to coerce it to a string in a bigger expression I get cannot cource set to a string.
<maralorn>
How can I get the output path of the derivation?
<balsoft>
the_pumpkin_man: ah, in this case you should probably just forget about networking in nix builds for now
<MichaelRaskin>
lukego: It might not even be abandonware, we just have a thing with different tradeoffs for putting stuff inside Nixpkgs
chloekek has quit [Ping timeout: 256 seconds]
<balsoft>
Well, if you want to, I can tell you how to wget, but it's not a recommended practive
<balsoft>
the_pumpkin_man: recommended practice is to use fetchurl or fetchTarball and then place the result of those in your build
<the_pumpkin_man[>
Yes I know
<the_pumpkin_man[>
But I'm packaging some software
<balsoft>
Both of them are implemented via fixed-output derivations (nixpkgs versions, that is, builtin versions are actually eval-time)
<balsoft>
the_pumpkin_man: ah, in that case please wait a second
iyzsong has quit [Ping timeout: 260 seconds]
<balsoft>
the_pumpkin_man: You have to specify outputHash = "sha256"; outputHashMode = "recursive"; outputHash = your_sha256;
<lukego>
MichaelRaskin: hm well if it's not abandoned then I have another patch I need to submit :)
<the_pumpkin_man[>
Ok, will try
<balsoft>
the_pumpkin_man: I'm not aware of any good way to get your_sha256 other than by building it with a fake sha256 first and replacing with the sha256 nix will spit out in its error message
<MichaelRaskin>
I mean, Common Lisp is not a high-churn ecosystem, if something works with it, it will probably continue to work
<the_pumpkin_man[>
I already built it with nix-shelll
<the_pumpkin_man[>
So it's already good
iyzsong has joined #nixos
<balsoft>
the_pumpkin_man: you might be able to nix-hash the result you've built with nix-shell and use that as your sha256 (there might be some issues with that approach)
<lukego>
MichaelRaskin: One other issue I have with ql2nix is having to constantly re-bootstrap it each time I need a new package. Could be nice to just include every package every time. That way I only need to rebootstrap when a new Quicklisp is released.
<balsoft>
the_pumpkin_man: Ah, it's basically what argument is passed to nix-hash
<{^_^}>
[nixpkgs] @petabyteboy pushed commit from @prusnak to master « giada: 0.16.1 -> 0.16.2.2 »: https://git.io/JfC6r
<{^_^}>
[nix-pills] @dmedinag opened pull request #142 → Paraphrase utility of strip → https://git.io/JfC6o
<MichaelRaskin>
lukego: well, in Nixpkgs we include a significant but not even close to full list of Quicklisp things, and you can include them without the expensive re-bootstrapping step
<the_pumpkin_man[>
But what does it changes ?
<balsoft>
the_pumpkin_man: (BTW there's a matrix room #nix:matrix.org)
<balsoft>
the_pumpkin_man: I haven't seen any other argument but "recursive" there
<the_pumpkin_man[>
Ok, thank you !
<balsoft>
the_pumpkin_man: I think it also accepts flat and then passes nix-hash --flat but I don't know how that works
<balsoft>
There are people here who know a lot more about this stuff, you might want to wait for them :)
<MichaelRaskin>
Recursive hashing means you are checksumming a directory
<the_pumpkin_man[>
Oh then that's not what I want
<betawaffle>
if i use a nixos module import (i.e. { imports = [ ./secrets.nix ]; }), is that going to be copied to the nix store, or can i keep it out of the store?
chaoflow has joined #nixos
<the_pumpkin_man[>
Just need that for a single file
<MichaelRaskin>
There is another mode that is only applicable to single plain files and is compatible with standard sha*sum
<adisbladis>
betawaffle: That's not copied to the store
<betawaffle>
ok great
<adisbladis>
Your secrets may be depending on how you treat them
tokudan has quit [Remote host closed the connection]
<MichaelRaskin>
Well, recursive is applicable to single files in principle
<adisbladis>
But `imports = []` doesn't inherently put things in the store
<betawaffle>
adisbladis: right, in my case i'm only using `users.users.<name>.hashedPassword = "..."`
<balsoft>
betawaffle: I think that it'd be very hard to keep the secrets you set in secrets.nix from appearing in nix store, though. If you use them in any file/script/whatever, they're automatically there.
<chaoflow>
I'd like to use a shell.nix file for multiple projects and use the directory basename as derivation name. Is there something like: stdenv.MkDerivation rec { name = basename { ./. }; }?
Noisytoot has quit [Ping timeout: 258 seconds]
<adisbladis>
I made a hack for that the other day when we had a discussion about secrets
<adisbladis>
They still go in the instantiated drv, but will fail at build time
<betawaffle>
doesn't nixos keep /etc/shadow things out of the store?
ircuser5678 has joined #nixos
<siers>
Where should I put "openvpn --mktun --dev"s, "brctl addbr"s and etc. for openvpn bridging? Should I make a systemd service that has the dependency? // see bridge-start in https://openvpn.net/community-resources/ethernet-bridging/
tokudan has joined #nixos
<adisbladis>
betawaffle: Yes and no, it ensures they are hashed iirc
<betawaffle>
well it's already a hash of my password...
<betawaffle>
but it's still strictly better to not leak it
<ircuser5678>
My issue is that I get an error message when trying to nixos-rebuild switch: error: undefined variable 'blk' at /etc/nixos/scripts-configuration.nix:17:18
* adisbladis
should implement his hack in that repo :3
<ircuser5678>
I can't just assign simple variables when using writeScriptBin? The script works fine if I just create it manually chmod +x it and execute it
<adisbladis>
ircuser5678: What exactly is the problem?
<adisbladis>
blk is using the nix string interpolation syntax, that's gonna fail
<ircuser5678>
It's what a normal shell does as well. How do I need to change this for nix?
<user51[m]>
I have set networking.proxy.httpProxy in my configuration.nix, but apparently the environment variable is only set for root. How can i set it for all users?
<betawaffle>
completely unrelated question. is `.override` the way i'm supposed to pass arguments to a package?
<simpson>
adisbladis++
<{^_^}>
adisbladis's karma got increased to 70
<simpson>
maralorn: TBH, though, I often find myself having to look up various stuff for various languages, and Nix isn't as annoying as, say, GLSL, or DOM-flavored ECMAScript.
<maralorn>
Oh, wow. Thanks
<adisbladis>
betawaffle: It's the way you amend arguments to the package function
<betawaffle>
is there a different way i can pass arguments to packages that are designed to be "configurable" ?
<maralorn>
simpson: I mostly write haskell and basically everything pales in contrast to hoogle.
morr has quit [Quit: WeeChat 2.7.1]
<adisbladis>
We should make nix-1p part of the official docs...
waleee-cl has quit [Quit: Connection closed for inactivity]
morr has joined #nixos
hlisp has quit [Ping timeout: 256 seconds]
<simpson>
maralorn: It's alright, but when reading Haskell I usually have to go look up instance definitions and ensure that the type's defined in the way that I want. TBF I'm usually porting *from* Haskell.
opthomasprime has quit [Remote host closed the connection]
<slack1256>
I got a binary (cave story) and I could modify on an stdenv.mkShell environment with patchelf to run. What is the correct setup to package such binary?
<slack1256>
Should it be a default.nix that has an empty build setup and just a postInstall that uses patchelf to overwrite the rpaths?
<slack1256>
(cave story) that I modified on a stdenv.mkShell..
<{^_^}>
[nixpkgs] @bjornfor pushed commit from @wamserma to master « mupdf: add MIME types to .desktop »: https://git.io/JfCP8
hlisp has quit [Remote host closed the connection]
<slack1256>
I also have to deal with save-files. On a FHS-distro, this game is put on /opt and it can write data on its own folder.
hlisp has joined #nixos
vidbina has joined #nixos
<{^_^}>
[nixpkgs] @stianlagstad opened pull request #87383 → dstask: init at 0.17 → https://git.io/JfCP7
<ircuser5678>
adisbladis: Thanks but I guess I'm too dumb. A very simple case: blk=abc. I can then use echo $blk. But if I need to use this variable like so (e.g. in a string) echo ${blk}. It fails. The bot said that I need to escape it this way: echo \${blk}. But leads (again) to the undefined variable error when switching the build
<balsoft>
ircuser5678: try passing inherit blk to your stdenv.mkDerivation call
rogue_koder_ has quit [Remote host closed the connection]
<hyper_ch>
I have a weird issue: have networking.nameservers = [ "127.0.0.1" "10.10.10.1" "8.8.8.8" "8.8.4.4" ]; but the /etc/resolv.conf only shows: nameserver 127.0.0.53 and options edns0
virus_dave has joined #nixos
<steevveen>
Hi! I'm learning Perl, version5, and I try to do some little scripts. For one I need a CPAN module. I have a default.nix that includes perl528Packages.CPAN. I launch the command line, but when installing the module install X11::FullScreen, at some moment there is an error. I suspect CPAN is conflicting with the nix way, as it obviously tries to do
<steevveen>
stuff inside the /nix/store. To perl5 programmers, how do you do with external CPAN module , when doing the nix way ?
fling has quit [Remote host closed the connection]
<adisbladis>
hyper_ch: Thats normal I think
<adisbladis>
hyper_ch: IIRC the nameservers populate your resolver which I'm gonna guess is either systemd-resolved or unbound ?
<hyper_ch>
adisbladis: If I add explizit nameserver in the nixos config and then they don't appear in resolv.conf is normal? resulting in not actually being able to resolve domain names?
<adisbladis>
Let me look into exactly what the module is doing, but I'm pretty sure that's the way things work.
<hyper_ch>
adisbladis: I had to manually add a namserver to /etc/resolv.conf because otherwise I couldn't update nixos because it couldn't resolve nixos.org
<adisbladis>
Indeed
<adisbladis>
hyper_ch: Are you using resolved or some other caching resolver like unbound or bind?
<adisbladis>
Are you using networkmanager?
<hyper_ch>
adisbladis: no
_d0t has joined #nixos
<_d0t>
ohai! How do I check if my derivation is built with sandboxing enabled?
<exarkun>
clever: "nix-store --verify --check-contents" exited with 0 and no further output fwiw
<tobiasBora>
I'm quite curious to know, the version of Ocaml that uses flambda seems pretty common and famous, but I just tried to install it and it compiles everything from scratch (and takes ages). Is there any reason the flambda version of ocaml is not compiled by Hydra?
codygman has quit [Read error: Connection reset by peer]
codygman has joined #nixos
<adisbladis>
tobiasBora: Just quickly eyeing the expression it's not enabled by default
<adisbladis>
And there isn't a derivation in nixpkgs setting it to true
<jakobrs>
Could somebody take a look at #86955? It makes it possible to perform a bind mount in initramfs without having to have the file systems ordered alphabetically.
<{^_^}>
https://github.com/NixOS/nixpkgs/issues/86955 (by jakobrs, 3 days ago, open): fsBefore does not recognise one filesystem as depending on another filesystem in the case of bind mounts where the source of the bind mount is not itself a mount point
<infinisil>
adisbladis: Unfortunately mypy still doesn't work in the checkPhase of a buildPythonApplication, but I'll probably just make a separate derivation instead :)
<infinisil>
I guess dependencies are installed differently in there
wavirc22_ has quit [Read error: Connection reset by peer]
wavirc22 has joined #nixos
<infinisil>
adisbladis: I wonder if it's reasonable to use buildEnv's for buildPythonApplication
hlisp has quit [Ping timeout: 272 seconds]
hlisp has joined #nixos
<{^_^}>
[nixpkgs] @talyz opened pull request #87406 → arcanist: Fix on darwin → https://git.io/JfCFY
alexherbo2 has joined #nixos
<{^_^}>
[nixpkgs] @bhipple opened pull request #87407 → doc: update python manual to reference python 3.8 instead of 3.5 → https://git.io/JfCFW
<{^_^}>
[nixpkgs] @bjornfor merged pull request #86436 → maintainers/scripts/build.nix: script to build all maintainer packages → https://git.io/JfOPz
<{^_^}>
[nixpkgs] @bjornfor pushed commit from @gnidorah to master « maintainers/scripts/build.nix: script to build all maintainer packages »: https://git.io/JfCFl
<ixxie>
I just want an editor which is not a project to configure
rock_rerrigan10 has joined #nixos
<ixxie>
but which I can declaratively configure :D
<cole-h>
Why not?
<cole-h>
That's the best part!
<etu>
I started configuring my Emacs 12 years ago, it works well enough now. It's quite rare that I change the config.
rock_rerrigan10 has quit [Remote host closed the connection]
<adisbladis>
Did someone say emacs? :)
<clever>
exarkun: weird
<adisbladis>
Tbh you probably save time setting up emacs properly :P
the_stanley_hote has joined #nixos
zeenk has quit [Ping timeout: 260 seconds]
<hyper_ch>
if emacs only had a decent editor
the_stanley_hote has quit [Remote host closed the connection]
<cole-h>
That's why I use evil-mode :^)
tandem20 has joined #nixos
tandem20 has quit [Remote host closed the connection]
<ixxie>
I don't know, I kinda wish I could have an editor without its own configuration DSL
<MichaelRaskin>
ixxie: so, Textadept?
<ixxie>
so lua configs
<ixxie>
I guess that's an improvement
<adisbladis>
Well... Everyone has their own thing I gues...
<{^_^}>
[nixpkgs] @bjornfor pushed to master « gnomeExtensions: add 'uuid' attr to all extensions »: https://git.io/JfCb9
<zeta_0>
hello there, i am finally able to get ihaskell to build in a shell.nix config using ghc883 without any errors, to be used with my (lorr+direnv+emacs-direnv) setup, after initializing everything and running `lorri shell` most of my shell.nix projects work out-of-the-box, but for some reason ihaskell is giving me some problems, when running `M-x ein:run`, emacs throws an error that it can't find jupyter? i'm thinking
<zeta_0>
it's because the emacs environment is not in sync with nix environment?
<adisbladis>
zeta_0: I have a hunch which may or may not be correct. Could you paste your shell.nix ?
shibboleth has joined #nixos
<zeta_0>
i tried placing `use_nix shell.nix` in the .envrc file in the root directory of the project where the shell.nix file is, it opens another emacs application with everything working, but that defeats the purpose of using this lorri setup, any tips on how i can get this working correctly?
<kimat>
I imagine OfBorg not being green for a week made people ignore this. Should I close & reopen the MR instead ?
<pistache>
iqubic: have you installed "dunst" using nix-env ?
<iqubic>
No. No I haven't.
fresheyeball has joined #nixos
<iqubic>
but the strange thing is that I didn't get this issue last night when I ran home-manager switch.
<adisbladis>
zeta_0: I need to eat something first
<pistache>
iqubic: I get this error when HM tries to install something that was already installed through nix-env
<adisbladis>
[picom + exwm]: Yes, it just works(tm)
<iqubic>
It's only today, after updating my channel that this fails.
hlisp has joined #nixos
<{^_^}>
[nixpkgs] @flokli pushed to master « Revert "nixos/resolved: Include dbus alias of resolved unit" »: https://git.io/JfCAv
<iqubic>
adisbladis: Good to hear.
oborot has quit [Ping timeout: 240 seconds]
<cole-h>
kimat: No need to close and open anew; we just get such a large volume of PRs it's impossible to get to every single one (and they are easily missed).
<cole-h>
Also, ofborg looks perfectly green to me on your PR :)
alexherbo2 has quit [Ping timeout: 256 seconds]
<kimat>
I had to rebase master 2 days ago and force push to get it green
<iqubic>
So, I'm not sure what's up with my `home-manager switch` not working.
b42 has joined #nixos
<zeta_0>
i'm going to go eat as well
hlisp has quit [Ping timeout: 246 seconds]
cosimone has quit [Remote host closed the connection]
cosimone has joined #nixos
Fare has joined #nixos
pxc has joined #nixos
<hpfr[m]>
hi, I'm trying to add nerd fonts locally because my connection is poor and it rebuilds whenever a dependency updates. with the following nix, I get `error: attribute 'requireFile' missing, at /nix/var/nix/profiles/per-user/root/channels/nixos/lib/modules.nix:221:28`: https://bin.privacytools.io/?09885406fe170af1#ihMjIEjXflpD8ndWIh2xbHJgTPZBySOtUTWVendDpS4=
<rycee>
iqubic: What does `nix-env -q` say?
<iqubic>
Ah. Oh. I see.
<iqubic>
Turns out I did actually have dunst installed locally.
<rycee>
Indeed :-)
hlisp has joined #nixos
<{^_^}>
[nixpkgs] @chkno opened pull request #87414 → nixos/test-driver: Specify /bin/sh shell when running a bourne shell script as the user → https://git.io/JfCAs
<balsoft>
kimat: Are you sure about platforms.all ?
morgrimm has joined #nixos
<cole-h>
balsoft: The software touts itself as cross-platform and has Windows, Linux, and macOS helpers, so I'd imagine it's fine.
<balsoft>
cole-h: I've seen the platforms being limited for software that _should_ have support for a certain platform but doesn't, like SDL
<cole-h>
I guess it could be -> platforms.x86_64 if aarch is untested/unknown
<balsoft>
cole-h: I mean it'll try to build it for e.g. Windows cross, and I'm not sure we even have Rust toolchain for Windows...
<emily>
does hydra even build much windows stuff?
<emily>
it's always been compiling a bunch manually whenever I've tried
<emily>
(I think platforms.all is fine)
domogled has joined #nixos
<emily>
(if there are portability issues in dependent packages then those should be fixed separately)
<balsoft>
Maybe I should rephrase my question: what does meta.platforms really mean?
<emily>
having to manually override platforms because they're too conservative because of past breakage is frequent and annoying
<balsoft>
Does it have the same semantics as broken or not?
<clever>
balsoft: nixpkgs will automatically block trying to build a package on an OS where meta.platforms doesnt list it
<balsoft>
Yeah, I guess you're right, so meta.platforms actually mean "what platforms upstream supports" and not "for what platforms this package actually builds"
<balsoft>
clever: yes, I understand that, and I think it'
<clever>
balsoft: hydra will also try to build it for all platforms, that are enabled within hydra
<emily>
balsoft: tbh, I'm not sure why we'd have platforms at all rather than broken = isPlatform
<{^_^}>
[nixpkgs] @matthewbauer closed pull request #43300 → [wip] haskell: add core libraries logic to remove need for global package db → https://git.io/fNTR7
<emily>
I guess because in a lot of cases it's intentional rather than "brokenness"
<balsoft>
kimat: Also, it looks like espanso might need a NixOS module to function properly, since it needs a systemd service?
<balsoft>
clever: emilazy: Do you think we should fix platforms where they don't match the definition of "does upstream support it?" (e.g. SDL)
<balsoft>
And ocaml
<emily>
balsoft: I wouldn't necessarily object, but I don't know if it's worth changing after the fact without actually fixing the build
<emily>
like, I'd rather ask "can we get SDL working for Windows"
alp has joined #nixos
<balsoft>
Oh, RE: OCaml, is there any work on making it cross-compile to windows? I tried to do that with 4.09 and failed miserably because apparently mingw doesn't have enough libraries or something
<balsoft>
I don't remember the exact issue, just that I couldn't make it work. Maybe another question would be whether anybody else cares about this?
<emily>
ocaml on windows has historically been flaky, right? I think it's gotten better recently, but "Nix Windows builds are flaky" + "OCaml on Windows has historically been flaky" makes me worry it might be a bit of a mess
chagra_ has joined #nixos
<balsoft>
I've heard that massive improvements are happening in 4.10 and 4.11
ircuser5678 has quit [Quit: Going offline, see ya! (www.adiirc.com)]
<balsoft>
I don't really care about it much, just would be nice to show off how cool nix is to a client by shipping windows executables...
<{^_^}>
[nixpkgs] @martinetd opened pull request #87418 → mpd service: remove user/group from conf → https://git.io/JfCxP
<{^_^}>
[nixpkgs] @thefloweringash opened pull request #87419 → compiler-rt: fix build on armv7l → https://git.io/JfCxX
hlisp has joined #nixos
<ixxie>
anybody familiar with the vscode extension system?
<balsoft>
ixxie: I'm slightly familiar with it, and it's a pain
<ixxie>
balsoft: I'm getting this weird situation where random extensions are not being discovered, including the nixfmt and the nix-env-selector extension
hlisp has quit [Ping timeout: 258 seconds]
<balsoft>
Are you installing them via the vscode's builtin extension manager or with nix?
<asymmetric>
is there a way to change the path of /etc/nixos/configuration.nix?
<ixxie>
balsoft: nix... via the home-manager specifically
<asymmetric>
i'd like to keep it together with the rest of my dotfiles
<lovesegfault>
there are multiple green evals but no release in 10 days
<lovesegfault>
c.f. status.nixos.org/
<lovesegfault>
cc. gchristensen
o1lo01ol1o has quit [Remote host closed the connection]
pxc has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @tbenst opened pull request #87421 → <!-- Nixpkgs has a lot of new incoming Pull Requests, but not enough people to review this constant stream. Even if you … → https://git.io/JfChe
vidbina has quit [Ping timeout: 264 seconds]
luna has quit [Ping timeout: 256 seconds]
<{^_^}>
[nixpkgs] @tbenst closed pull request #87421 → <!-- Nixpkgs has a lot of new incoming Pull Requests, but not enough people to review this constant stream. Even if you … → https://git.io/JfChe
<c4rc4s>
newb question: I'm on NixOS 20.03 and trying to install (what I believe to be) a package derivation, obs-v4l2sink, but it isn't found when I try to add it to systemPackages. It exists in the nixpkgs repo, under obs-studio, but when I query for it with nix-env I'm notified that "selector 'obs-v4l2sink' matches no derivations"
<c4rc4s>
Do I need to do something special to make that available..? Other derivations are found OK, e.g. obs-ndi
<MichaelRaskin>
I think it might be simply newer than the release branch branch-off
o1lo01ol1o has joined #nixos
morgrimm has quit [Ping timeout: 256 seconds]
<clever>
c4rc4s: id say start by greping nixpkgs, to see what is loading it
<c4rc4s>
Cool, made the unstable channel available ala https://stackoverflow.com/a/48838322 since I didn't want to switch everything to it and was able to install that package. Thanks for the pointers folks.
mcwitt has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
codygman has quit [Read error: Connection reset by peer]
<snicket>
any idea why 'nixpkgs.overlays = [ (self: super: { sdrangel = super.callPackage ./packages/sdrangel { libusb-compat-0_1 = self.libusb; }; });' fails with the "anonymous function called without required argument 'mkDerivation' error?
mac10688 has joined #nixos
<snicket>
sdrangel is just the file copied from nixpkgs
codygman has quit [Read error: Connection reset by peer]
codygman has joined #nixos
<numkem>
how do you generate the `mod256` for a go modules? I think it's `go.sum` it's looking at. Using nix-hash I can't reproduce the same hash.
<cole-h>
,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. See: tofu-vim
<numkem>
cole-h: what if I need to automate it and replace the value?
<cole-h>
I've seen scripts in nixpkgs that just use that trick and then look for the expected hash on stdout and replace it that way, I think.
<numkem>
cole-h: that's what I'd do but I figured there was a nicer way of doing it. If that's considered good enough, fine by me
<cole-h>
Curious what you're automating, though.
<numkem>
cole-h: maybe I'm approaching it the wrong way, still trying to wrap my head around it. I'd like to use Nix to build docker containers. In order to do that i need to do a `default.nix` for my package. It says in the docs you can set it to `null` but if you do it breaks sandboxing and reproducable builds which I'd like to avoid. So I was wondering how I could make a git hook that would just swap the values if my `go.sum` or `go.mod` changes
<numkem>
trying to make it as transparent as possible for developers to use
<bdju>
another new firefox issue. after manually quitting with ctrl-q to try out a new about:config setting, it won't relaunch until I pkill the process. I've waited whole minutes. No errors about it being open already either. Just nothing.
<bdju>
I have an issue with text input boxes being dark background and dark text or light bg/light text so it's unreadable, and I'm trying some suggestions related to overriding gtk theme settings, but here another issue slows me down
hlisp has joined #nixos
seku has joined #nixos
<numkem>
bdju: have you noticed too that screenshots doesn't work anymore? Wondering if it's just me
Maxdamantus has quit [Ping timeout: 272 seconds]
<fabianhjr>
Is it just me or has there been no updates to nixos-unstable channel in a while?
<numkem>
fabianhjr: status.nixos.org says it was last updated 10 days ago
<iqubic>
I ran `systemctl status display-manager.service` and that gave me the following output: https://dpaste.com/3J6RGGY
<iqubic>
I have no idea how to fix this issue.
<iqubic>
This issue happened after I updated my nixos config, and my home-manager config and tried rebooting.
<iqubic>
I'm currently running irssi in a tty to talk to you guys.
fendor has joined #nixos
punkj has joined #nixos
<lukegb>
hmm, dpaste doesn't like me
<cole-h>
Yeah, I get EOF
<iqubic>
Oh. I see.
<lukegb>
cole-h: Out of interest, what points towards a PHP thing contributing to channel-advancement-brokenness
<iqubic>
Might be the method I'm using to upload the file.
<cole-h>
lukegb: 08:37 <gchr*stensen> nixos-unstable is not updating because of this: error: in package ‘php’: this derivation has bad 'meta.outputsToInstall'
<iqubic>
How do I share the contents of that file with you?
<iqubic>
*command
<cole-h>
Well, ideally you would look at the contents yourself and figure out what went wrong :P You can redirect it to a file and upload the most interesting parts.
hlisp has joined #nixos
user_0x58 has joined #nixos
<monokrome>
there's a cool service called sprunge for uploading text files via curl
<bdju>
for some reason I'm still on firefox 75 even after running updates again. is something wrong with my setup?
<{^_^}>
[nixpkgs] @prusnak opened pull request #87434 → lite: init at 1.03 → https://git.io/JfWJQ
<iqubic>
Here's all the relevent stuff from journalctl
<iqubic>
I hope that paste looks right.
<cole-h>
Nice, actually 0 information in that paste as well :D Next step is to look at Xorg logs. On traditional systems, it's `/var/log/Xorg.0.log`. No idea if NixOS follows that.
jb55 has quit [Remote host closed the connection]
jb55 has joined #nixos
<iqubic>
cole-h: I'm just trying to pull up whatever logs I can.
o1lo01ol1o has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #87435 → treewide: strip more things in cross → https://git.io/JfWUk
CitadelCore has quit [Quit: Connection closed for inactivity]
iqubic has quit [Remote host closed the connection]
iqubic has joined #nixos
wykurz has joined #nixos
<wykurz>
hipple
Maxdamantus has quit [Ping timeout: 240 seconds]
<iqubic>
the xorg logs tells me that the extraConfig section has a parse error: "the Section keyword requires a quoted string to follow it"
<iqubic>
So, something there is screwing me up.
vika_nezrimaya has joined #nixos
Maxdamantus has joined #nixos
<cole-h>
iqubic: It's probably the indentation that's screwing you.
<iqubic>
Ah, yes. That would be the issue: In multiline strings in nix, I don't need to escape quotes. So the config file just has extra backslashes in it causing a parse error. X11 treats that parse error as a fatal error and crashes on me.
<cole-h>
The `Section` is not indented at all, yet the `EndSection` is, excessively so.
<infinisil>
> '' Section \"foo\" ''
<{^_^}>
"Section \\\"foo\\\" "
<iqubic>
But also, yeah. How do I ensure consistent spacing here?
<cole-h>
Newline after the first ''; instead of `''Section`, it would be `''\nSection`
<cole-h>
I only did \n because you can't type multiline in IRC. Expanding the \n into an actual enter key-press was left as an excercise to the reader...
<iqubic>
Ah. I see.
ixxie has joined #nixos
remirol has joined #nixos
<iqubic>
Yeah. That seems to have worked.
<iqubic>
And oddly enough once I did that rebuild, lightdm started up.
zupo has joined #nixos
<iqubic>
Why might that happen?
<cole-h>
Why might what happen/
<iqubic>
Why might LightDM start after I changed the extraConfig section of my xserver config and then do a rebuild?
numkem has quit [Ping timeout: 256 seconds]
lorimer has quit [Ping timeout: 240 seconds]
wykurz has quit [Read error: Connection reset by peer]
<cole-h>
If you `switch`ed to the new generation, I would imagine modified services get reloaded and restarted to take on these new changes.
<qyliss>
Can I have a NixOS option that will do a deep merge if specified multiple times?
<infinisil>
qyliss: By deep you mean that attribute sets get merged together recursively?
<qyliss>
Yeah
<qyliss>
Like, recursiveUpdate
<{^_^}>
[nixpkgs] @bhipple opened pull request #87440 → doc: use idiomatic overlay names in the python manual example → https://git.io/JfWTT
<infinisil>
qyliss: You can encode that in the type, e.g. `let atom = oneOf [ str int ]; value = attrsOf (either atom value); in value`
<qyliss>
_Technically_ it only has to go two layers deep
<balsoft>
yes, you can, but the way you implement that is not simple I don't think...
morgrimm has joined #nixos
<qyliss>
In fact, exactly two layers deep I guess
ixxie has quit [Ping timeout: 256 seconds]
<qyliss>
So I suppose attrsOf (attrsOf str) is probably what I want
<balsoft>
qyliss: I think you should go for a simple-minded two-layers-deep merge approach, because otherwise you'll have to write your own type that merges recursively (I don't think there's one OOTB)
rogue_koder_ has quit [Remote host closed the connection]
rogue_koder_ has joined #nixos
<morgrimm>
Anyone here using Lorri with home-manager?
<iqubic>
time to see if picom works as expected.
hlisp has joined #nixos
iqubic has quit [Quit: leaving]
<morgrimm>
I've just set lorri up with home manager and not specifically systemd, but it seems to not have the permissions it needs to run
Maxdamantus has joined #nixos
<morgrimm>
do I just have to elevate every lorri command? or is there a correct way to set it up?
mpiechotka has joined #nixos
iqubic has joined #nixos
<mpiechotka>
Hi. I want to dual boot NixOS on my machine which has Arch. Is it possible to configure nix to only create loader entries and not to install its own bootloader?
<iqubic>
So, not only did picom work out of the box, but enabling vsync via picom fixed the screen tearing issues I was having. Well, it was either that, or it was adding that services.xserver.extraConfig options that did it.
user_0x58 has quit [Quit: Leaving]
hlisp has quit [Ping timeout: 260 seconds]
chloekek has quit [Ping timeout: 272 seconds]
<morgrimm>
iqubic: I'm considering dropping picom into my NixOS flake - how does it perform?
<morgrimm>
Do you have a dGPU, and if so, any issues running it? :O
Maxdamantus has quit [Ping timeout: 256 seconds]
<iqubic>
I don't have a dedicated GPU, and I don't have enough experience with picom to tell you how it performs. Ask me again in a week or two.
Maxdamantus has joined #nixos
<morgrimm>
Fair enough - hope it works well :D
dnlkrgr has quit [Ping timeout: 256 seconds]
<iqubic>
I'm sure it will.
_kwstas has quit [Quit: _kwstas]
knupfer has quit [Ping timeout: 260 seconds]
Maxdamantus has quit [Ping timeout: 256 seconds]
Maxdamantus has joined #nixos
Maxdamantus has quit [Ping timeout: 272 seconds]
Maxdamantus has joined #nixos
dsg has quit [Remote host closed the connection]
dsg has joined #nixos
ixxie has quit [Ping timeout: 272 seconds]
ixxie has joined #nixos
hlisp has joined #nixos
mpiechotka has quit [Quit: mpiechotka]
mpiechotka has joined #nixos
mpiechotka has quit [Client Quit]
zupo has quit [Ping timeout: 272 seconds]
o1lo01ol1o has quit [Remote host closed the connection]
chagra_ has joined #nixos
hlisp has quit [Ping timeout: 246 seconds]
zupo has joined #nixos
morgrimm has quit [Ping timeout: 264 seconds]
dongcarl has joined #nixos
iqubic` has joined #nixos
iqubic` has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
iqubic has quit [Ping timeout: 240 seconds]
tobi_ has joined #nixos
hlisp has joined #nixos
<ldlework>
I just did `nix-env -iA qsynth` then ran qsynth and got
<ldlework>
qt.qpa.plugin: Could not find the Qt platform plugin "xcb" in ""
<ldlework>
This application failed to start because no Qt platform plugin could be initialized. Reinstalling the application may fix this problem.
<ldlework>
I thought everything in nixpkgs is supposed to work :P
<Yaniel>
qt is not very friendly to the nixos approach to things
<ldlework>
Yeah but how did qsynth make it into nixpkgs
<ldlework>
if it does not work
<ldlework>
I noticed the same for supercollider
<ldlework>
fails with same message
<Yaniel>
it probably works when installed in systemPackages
<cole-h>
ldlework: Did you try qsynth from master?
o1lo01ol1o has quit [Ping timeout: 246 seconds]
<ldlework>
master nixpkgs you mean?
<ldlework>
Is there some cli oneliner to nixpkgs from an arbitrary nixpkgs rev?
<tobi_>
hello, I have packaged https://github.com/pdbpp/pdbpp. before I can make a pull request I have a few questions: 1. in order to package pdbpp, I had to package a few dependencies as well. is it ok to create a single pull request for all of them? 2. how important are tests? I'm not that familiar with python unittesting and when I tried to make them run I ran into significant problems. 3. In the packaging
<tobi_>
process, I installed the packages to my machine, however I didn't have "useSandbox" enabled. Is there an easy way to completely remove the packages again, so they have to be rebuilt? I've tried removing them from system packages and running "nix-collect-garbage-d" but it didn't seem to work
<Yaniel>
tobi_: preferrably create separate pull requests for dependencies
<Yaniel>
as in one PR per package
<Yaniel>
tests are not strictly required but if you have any, do share
luna has joined #nixos
<Yaniel>
useSandbox is also not a strict requirement for non-essential userspace stuff
<Yaniel>
uninstalling depends on how you installed it
luna has quit [Client Quit]
<ldlework>
OK
<Yaniel>
if you used `nix-env -i` then `nix-env -e` and `nix-collect-garbage` will take care of it
cosimone has quit [Ping timeout: 260 seconds]
<ldlework>
I am officially considering this Nix to be shitting the bed.
<tobi_>
Yaniel: thanks for your quick answers! I added them to my systemPackages via an overlay
quinn has joined #nixos
<Yaniel>
tobi_: then remove them, nixos-rebuild switch, nix-collect-garbage, readd and nixos-rebuild switch
<ldlework>
Yaniel: what would logging out do?
<infinisil>
Yaniel: tobi_: Hm actually I'd much rather have a single PR for all dependencies
<Yaniel>
not logging out but I'm guessing there are some environment variables that are set on login
<Yaniel>
infinisil: noted
gustavderdrache has joined #nixos
<tobi_>
Yaniel: hm, I thought I did that, but I will try again. infinisil: will do
hlisp has quit [Ping timeout: 272 seconds]
<ldlework>
Yaniel: for what though? I don't run Qt or any kind of desktop environment.
<ldlework>
I just don't know what environment variable is supposed to magically appear.
iqubic has joined #nixos
fling has joined #nixos
<Yaniel>
idk how nix handles them
<Yaniel>
but I do recall there being some env variable(s) that Qt needs in order to work
<Yaniel>
namely QT_PLUGIN_PATH
<ldlework>
Yeah but I don't expect that to be magically setup by me logging out and back in.
<infinisil>
The fix is to add wrapQtAppsHook to nativeBuildInputs of the derivation
<ldlework>
Ah OK so if I use a newer nixpkgs for, say, QSynth, maybe it'll work.
<Yaniel>
which channel are you using?
* infinisil
checks master
mbrgm_ has joined #nixos
<Yaniel>
last summer isn't exactly what I'd call "new" anymore
<infinisil>
Doesn't seem to have wrapQtAppsHook in latest master
<Yaniel>
hmm
mbrgm has quit [Ping timeout: 240 seconds]
<infinisil>
So nope
mbrgm_ is now known as mbrgm
<bdju>
When I click on the output device for a specific program on the playback tab of pavucontrol, the menu doesn't appear, it's like it's invisible. Can't seem to select with arrows either, though.
<monokrome>
Is there a specific package I need in a nix-build for it to use HTTPS? I don't know why it can't pull from PyPi here? https://builds.sr.ht/~monokrome/job/205315
<ldlework>
well nix-channel --list doesn't return anything
<ldlework>
So I'm unsure.
<Yaniel>
bdju: worksforme
fdert has quit [Read error: Connection reset by peer]
<bdju>
Yaniel: pcmanfm's right click menu has never worked for me on NixOS either. I'm sure I've got all sorts of GTK brokenness or something
<aleph->
Hmm, is there a util to see if all reqs of a python package are in nixpackages?
<ircuser5678>
hi. the wiki entry for samba states that I shoud be able to e.g. restart the service with systemctl restart samba. Great, unfortunately I'm getting this (as root): Failed to restart samba.service: Unit samba.service not found.
alp has quit [Ping timeout: 260 seconds]
<Yaniel>
ldlework: again, worst case you have to reboot
<ircuser5678>
But it is currently running... I can see that when using: ps aux | grep samba
<ldlework>
even if it breaks the bootloader?
<Yaniel>
okay, absolute worst case you need to get a live USB'
<Yaniel>
and then reboot
<Yaniel>
if you have another computer at hand that you can use for writing the live USB then you have nothing to worry
<Yaniel>
if not, consider making one before updating
<cole-h>
infinisil: Why would wrapQtapps hook need to be added? It uses `libsForQt5.callPackage` already
hlisp has quit [Ping timeout: 256 seconds]
<ldlework>
I do have another machine.
<infinisil>
cole-h: Ah, does that do it automatically?
<Yaniel>
ldlework: then you will at most lose half an hour of time
<Ashy>
ldlework: nixos is amazing for updates, i have successfully upgraded and downgraded multiple times without issues
<ex-parrot>
oh hello Ashy
mpiechotka has joined #nixos
<Yaniel>
plus however long it takes to download the iso
<cole-h>
infinisil: IIRC
<infinisil>
cole-h: The error does still happen for ldlework, so they probably use a version where it doesn't do that yet
<ex-parrot>
<-- food_pics
<tobi_>
ircuser5678: I think you have to use smbd.service
<Ashy>
ex-parrot: ah i was wondering haha
<ex-parrot>
:)
<ircuser5678>
tobi_: Failed to restart smbd.service: Unit smbd.service not found.
<ldlework>
Wait
<ldlework>
After updating the channel
<ldlework>
Isn't there some VM fuckery
<ldlework>
I can use to test whether my configuration will work after switching to it / rebooting?
<ldlework>
I feel like I use Nix because I don't like this feeling.
<ldlework>
LOL
<mpiechotka>
How to skip installing bootloader? I'm trying to install dual boot but nixos-install seems to insist to install something.
<Ashy>
ex-parrot: dunno if you're interested in apl-j too, i'm working on getting all the addons included in nixpkgs atm, in #jsoftware if you're interested
<infinisil>
ldlework: `nixos-rebuild build-vm` :)
asymmetric has quit [Quit: Peace.]
<ex-parrot>
Ashy: I am academically interested in apl but never really done anything with it
<Ashy>
same, heh
<Ashy>
instead of learning it properly i'm doing this nixpkgs side quest...
<infinisil>
mpiechotka: I believe with boot.loader.grub.device = "nodev"
<ex-parrot>
Ashy: sounds good. I want to get an APL typeball for one of my Selectrics...
asymmetric has joined #nixos
<ldlework>
infinisil: how do I boot it after
<infinisil>
ldlework: result/bin/run-nixos-vm
<bdju>
how do I check a meta.maintainers list? trying to file an issue
<ldlework>
infinisil: btw, you still using that module system we made?
<infinisil>
ldlework: I'm not sure what you mean
<mpiechotka>
infinisil: "File system "/boot" is not a FAT EFI System Partition (ESP) file system."
<ex-parrot>
Ashy: I'm working on remaking the PowerDNS module again today
<mpiechotka>
infinisil: NM. Forgot to disable systemd
hlisp has joined #nixos
<ldlework>
infinisil: the "mine" options/config system
<infinisil>
bdju: Got more info on that?
<ldlework>
recImport
<ldlework>
etc
<ldlework>
maybe you evolved so far you don't even see the abstractions anymore
<infinisil>
Ah yeah, not really a module system on its own, just some definitions
<ldlework>
yeah, you're right
<Ashy>
ex-parrot: you have multiple selectrics? :O
<ex-parrot>
currently two, but one of them is technically only on loan from sharrow
<ex-parrot>
the other one is mine
<bdju>
infinisil: part of the issue template says to check the meta.maintainers list for a package and then mention the maintainer of the package with the bug on github, but I'm not sure how I do this
<infinisil>
bdju: In the package file you're changing, there should be a meta.maintainers section at the bottom
<bdju>
I'm not changing anything
<bdju>
how do I get the a package file?
<bdju>
s/the/to
<infinisil>
Ah
<bdju>
I'm just reporting a bug
<tobi_>
ircuser5678: just checked on my server, "samba-smbd" should work
<infinisil>
bdju: You can do `nix-instantiate --eval -A <package attribute>.meta.maintainers --strict`