<{^_^}>
[nixpkgs] @SuperSandro2000 merged pull request #110580 → python3Packages.wasmer: make cargoHash invariant to the Python version → https://git.io/Jtscu
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed 2 commits to master: https://git.io/JtsNu
<{^_^}>
[nixpkgs] @SuperSandro2000 merged pull request #110639 → mc: move dependency pruning from post to pre fixup → https://git.io/JtsQX
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed 2 commits to master: https://git.io/JtsAS
ahmedelgabri has quit [Remote host closed the connection]
civodul has quit [Quit: ERC (IRC client for Emacs 27.1)]
<bbigras>
Is anyone able to cross compile Rust to windows? I got either `linker `cc` not found` or `error: could not determine how to read list of mounted file systems`. the later is if I try to add gcc.
mbrgm_ has joined #nixos
mbrgm has quit [Ping timeout: 265 seconds]
mbrgm_ is now known as mbrgm
pjan has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @fabaff opened pull request #110646 → go-audit: init at 1.0.0 → https://git.io/JtsAF
<m3thos>
hey there, I'm having a hard time using `virtualization....` to configure my box to always run a couple of docker containers (which aren't managed by me and are already published on docker/quay). All the docs I find are to build containers, not to run them..
<bbigras>
m3thos: did you check virtualisation.oci-containers ?
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed 2 commits to master: https://git.io/JtspU
<rmcgibbo[m]>
Does anyone know the "right way" to compute the intersection between two lists in the nix language? I can only think of a way that's O(N^2), when it should be possible in O(N log N).
<rmcgibbo[m]>
Ah nevermind. I found intersectLists.
<matthewcroughan>
My first question is how/why we need to flash the bootloader at specific sectors on the SD Card?
<matthewcroughan>
The "why" is very important. I'm confused by it immediately. Though I find it cool that I'm about to do something esoteric.
<superloach>
i would assume the board has hardcoded sectors it boots from
<superloach>
similar to how x86 bios boots from a 512m section at the beginning of a drive
<matthewcroughan>
Sure, and that generic image we're getting for armv7 I assume does not contain u-boot?
<superloach>
*512k
<matthewcroughan>
If so, why not?
<superloach>
im not sure on that
<matthewcroughan>
Does this mean that armbian encodes, for each board, a section on the boot media specially?
<matthewcroughan>
What most people are used to is flashing a .img/iso and powering the board on. Here there's an extra step.
<matthewcroughan>
Now it's very cool that this is the case, because it teaches you what a bootloader is, (if what you're saying is true). But I'd love to know the specifics of it a bit more.
<superloach>
that sounds about right yeah - a full bootable image will have the bootloader in it - but nix requires a bit more manual setup
<matthewcroughan>
So, do you think it's possible to generate bootable machine images rather than having to use the installer?
<superloach>
im not too familiar with nix yet, but the concept should apply
<matthewcroughan>
I think it would be great if we could encourage the usage of generators, or some other method, to provision disk images for hardware, at least as an option in the wiki. Has anyone explored this?
<matthewcroughan>
Surely colemickens must automate this?
<superloach>
the installer itself is a full bootable image, though im not sure what it uses
<matthewcroughan>
"nixos-aarch64-images provides a mechanism to modify the official NixOS installer with custom uboot bootloader required for different boards. This method does not require qemu or native arm, since the existing hydra image/pre-build uboot binaries are used. "
<matthewcroughan>
oh thanks Mic92 :D
<matthewcroughan>
So it looks like this wiki article is a bit conflicting.
<{^_^}>
[nixpkgs] @SuperSandro2000 merged pull request #110439 → multimc: Update to latest development build → https://git.io/JtOH5
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed commit from @starcraft66 to master « multimc: Update to latest development build (#110439) »: https://git.io/JtGve
<matthewcroughan>
OH! I see. The rockchip boards all have basic metadata, like ubootOffset, idbloaderOffset, which you're making a function to deal with here. This is beautiful!
<matthewcroughan>
This is concerningly great. I mean, this is *your* creation right? You could have missed something here right?
<matthewcroughan>
There could be an odd bit of metadata that you're not thinking about that is necessary to have a board boot, that's not included in images/rockchip.nix right?
sigmundv_ has quit [Ping timeout: 272 seconds]
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #110650 → Fix libretro.play on gcc10 → https://git.io/JtGJO
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed 2 commits to master: https://git.io/JtGJ4
<{^_^}>
[nix-mode] @matthewbauer merged pull request #127 → nix-drv-mode.el: make regexp more selective → https://git.io/Jt3Xj
<{^_^}>
[nix-mode] @matthewbauer pushed 2 commits to master: https://git.io/JtGJz
jonringer has joined #nixos
<demize>
Well, it should have everything that is required to boot rockchip boards, but of course it's possible that there's something broken about one of the boards that haven't been tested, sure.
<samueldr>
matthewcroughan: I also invite you to (also) ask aarch64 boards questions in #nixos-aarch64
pjt_tmp has joined #nixos
<demize>
Since they were talking about the Orange Pi, they would need to make a custom version of the rockchip function and ./images/rockchip.nix for Allwinner boards.
<samueldr>
not that it's not welcome here, but you might end up with the relevant people seeing questions faster
<samueldr>
demize: ah, I didn't catch up to *all* the backlog
<demize>
:)
<samueldr>
I assumed matthewcroughan wanted to deal with a rockchip board
<samueldr>
so `u-boot-sunxi-with-spl.bin` from the u-boot output, at the given location (bs=1024 seek=8)
<samueldr>
I guess, matthewcroughan, another thing to keep in mind is: is the board already in u-boot mainline (extremely probable) and is it exposed in Nixpkgs? (possible, but trivial)
<donofrio>
pjt_tmp, ;( "/pkgs/tools/backup/partclone/default.nix:38 is not supported on ‘x86_64-darwin’, refusing to evaluate."
<tpw_rules>
donofrio: i don't think partclone can compile on macos even without nix
<tpw_rules>
e.g. it's not on their website or in homebrew
<donofrio>
I wonder why
Mr_Cue has quit [Remote host closed the connection]
Mr_Cue has joined #nixos
<tpw_rules>
because it's heavy on linux specific raw io probably
<tpw_rules>
i believe it also just links to the filesystem drivers for what it supports too. it doesn't reimplement them
<donofrio>
tpw_rules, any way to backup my working nix install completely yet?
<tpw_rules>
idk. i worked on that with you a while ag
<tpw_rules>
o
<donofrio>
yes but I never got anything like "here use gzip and then it is all backed up" I heard things about direvitives and etc?
rb2k has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
logand`` has joined #nixos
cr4y1 has quit [Ping timeout: 240 seconds]
logand` has quit [Ping timeout: 240 seconds]
steve-chavez has joined #nixos
<{^_^}>
[nixpkgs] @jonringer merged pull request #108481 → nixos/doc/manual: In the preface, add link to #chap-contributing → https://git.io/JLbde
<{^_^}>
[nixpkgs] @jonringer pushed commit from @attila-lendvai to master « nixos/doc/manual: In the preface, add link to #chap-contributing »: https://git.io/JtGYL
<superloach>
samueldr: unrelated to the sbc stuff, I'm still having issues with nix on CYAN
<superloach>
I'm running the 5.2 kernel from gallium's tree, but still no internal keyboard
<superloach>
I'll poke around later, see if there's anything I'm missing
<superloach>
the working version I was using on arch was the 4.16.18
steve-chavez has quit [Ping timeout: 265 seconds]
<superloach>
the latest release marked "stable" is 5.1.16, I'll try that before I go all the way down to 4.x
<samueldr>
I couldn't say, at that point you seem to have a better grasp on that situation :)
<samueldr>
my most recent x86_64 chromebook is a PEPPY
<superloach>
alright, haha
<samueldr>
which works with mainline stuff
<superloach>
lucky
<colemickens>
the internal kb worked in Arch mainline though, right?
<superloach>
it did, yes
<samueldr>
that's odd
<samueldr>
note that we default to "current LTS at release"
<superloach>
but audio still didn't work on arch mainline
<samueldr>
have you tried rebuilding with _latest?
<superloach>
so I'm aiming to get the gallium kernel working anyway
<samueldr>
right, I would assume that still will be an issue
<superloach>
and I mean, it's /working/, but not enough lol
<KarlJoad>
I am having a hard time getting a package to find a library. The library is detected with autoconf, I believe. The error is: "dolfin headers could not be found, som functionalities will be disabled, don't worry your package will still be working, though."
sangoma has joined #nixos
<superloach>
rebuilding with 5.1.16 now
K0kada has joined #nixos
<energizer>
what is the advantage of gcing very frequently?
<superloach>
well, less garbage
<energizer>
let's say i have a good amount of storage space
<superloach>
then there's not too much advantage
<superloach>
and there's advantage not to, if some of the garbage can be reused by later builds :P
aw has quit [Quit: Quitting.]
spacefrogg has quit [Quit: Gone.]
aw has joined #nixos
<superloach>
hello aw
<energizer>
gchristensen: i'm curious what's your reasoning for gcing frequently?
spacefrogg has joined #nixos
kalbasit_ has joined #nixos
<superloach>
the long wait of a kernel build on a celeron... mmm
<KarlJoad>
Hey, octaveFull is crashing on GUI startup. I have an strace of what happened. What's the best way to move forward with that?
<tpw_rules>
read it? strace probably won't be too useful though
<KarlJoad>
It seems to be related to QT stuff. A line: access("/nix/store/h102qc0imha9ckbgl4skyjmj1j8rii1q-plasma-workspace-5.18.5/lib/qt-5.15.2/plugins/platforms/.", F_OK) = -1 ENOENT (No such file or directory)
<{^_^}>
[nixpkgs] @Ericson2314 merged pull request #97070 → lowdown: Split out "bin" and "man" outputs → https://git.io/JU3B8
<KarlJoad>
energizer: I have used that. But, when I evaluate the set of packages, one of the packages is evaluated (It is a function with a required argument that isn't defined in nixpkgs yet.)
* energizer
waits for someone who knows the real answer
Hail_Spacecake has quit [Remote host closed the connection]
Hail_Spacecake has joined #nixos
nvmr has joined #nixos
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
donofrio has joined #nixos
<iqubic>
How hard is it to get a Rust programming environment when using NixOS?
<iqubic>
I'm looking to get a shell.nix create to make my environment for me.
<bbigras>
pretty easy with a shell.nix. It's a bit more complex if you want to build a rust app with nix.
<iqubic>
Really? How is making a shell.nix different from building a rust app with nix?
<bbigras>
with shell.nix you can add cargo as a dependency and then you can build. you might need llvm too. to build a rust program with nix you might need something like crate2nix
<iqubic>
Oh. I see.
<iqubic>
That's a bit of a shame.
<bbigras>
you know what I mean by "building with nix". when it ends up in the nix store
<iqubic>
I wish there was a way to have nix automate this more often.
<iqubic>
Yeah.
kalbasit has joined #nixos
<bbigras>
for development with shell.nix it's great. and there's overlays for rust nightly
<bbigras>
I'll show you one. for building with nix you can use the build rust or whatever thing and skip crate2nix. it's simple but it will build every crates every time.
<bbigras>
yes you can use that
<iqubic>
I'd like to see what you were going to show me.
<bbigras>
ok wait
<iqubic>
Thanks for helping me with this.
<iqubic>
I'm just now starting down my path to learn Rust, so I haven't even got a development environment set up yet.
<{^_^}>
[nixpkgs] @wahjava opened pull request #110663 → kakounePlugins.kak-lsp: init at 9.0.0 → https://git.io/JtGy2
jluttine has quit [Read error: Connection reset by peer]
est has quit [Remote host closed the connection]
est31 has joined #nixos
jluttine has joined #nixos
Mr_Cue has quit [Remote host closed the connection]
shiriru has quit [Quit: Leaving]
Mr_Cue has joined #nixos
maier has quit [Ping timeout: 260 seconds]
maier has joined #nixos
simba2 has quit [Ping timeout: 264 seconds]
woobilicious has quit [Ping timeout: 248 seconds]
simba2 has joined #nixos
aveeran[m] has quit [Quit: Idle for 30+ days]
simba2 has quit [Ping timeout: 265 seconds]
<maier>
It worked fine just using the Windows-created ESP!
<maier>
But now I'm having visual problems again, trying to fix that by using proprietary nvidia drivers. But I'm getting cannot download NVIDIA-Linux... from any mirror when nixos-install runs.
<maier>
Do I have to configure a special mirror for that or something?
<maier>
Could this be a networking problem somehow? I can download the file manually just fine.
<Ke>
maybe that nvidia license forbids redistribution and they have removed the package in questiob
<Ke>
not sure how nixos binary caches work there
simba2 has joined #nixos
<maier>
Ke: The drv seems to just try to download from the link I can DL manually though, but gets stuck there. It even says "trying https://download.nvidia.com/XFree86/...";
cfricke has joined #nixos
<maier>
I don't know how to proceed, since without this driver my system doesn't seem to work/start correctly. I didn't need that proprietary stuff on my old PC.
respawn_ has quit [Quit: Leaving]
jonringer has quit [Ping timeout: 264 seconds]
<maier>
I tried gdm and lightdm, but they just won't come up without this driver. Getting into the graphical installer is also a hassle because of it, so I need to sleep the PC and wake it. But I think it would all be fixed if I could just install the proprietary driver. Any ideas on how I can get this to download? I enabled allowUnfree.
garionion has quit [Read error: Connection reset by peer]
garionion has joined #nixos
<Hail_Spacecake>
I'm runn ing into an issue using nixops
<srk>
maier: depends on the channel / nixpkgs version you are using, you might need to bump that
<Hail_Spacecake>
I have a local version of my nixops config that should be the same as the version running on my nixos box
<Hail_Spacecake>
if I do nixos-rebuild switch on the box itself, everything works fine
hiro98 has joined #nixos
mallox has joined #nixos
<maier>
srk: I'm using the unstable installer, downloaded yesterday. Should be quite recent thus.
<Hail_Spacecake>
if I do a nixos deploy via ssh, the deploy fails on every one of my acme-<sitename> services
<srk>
maier: hmm, yea, i'm running that driver myself
<Hail_Spacecake>
"account <whatever> is not registered, use 'run' to register a new account"
<Hail_Spacecake>
this seems to be an error from the logo library for interacting with lets encrypt
<Hail_Spacecake>
I am not usre how to fix this
<Hail_Spacecake>
and I'm not sure why this only happens wiht nixops and not when I run nixos-rebuild on the box itself
<maier>
srk: It might be an issue with a second gpu being detected before that. It's so hard to check because I always have to reboot into the installer, then sleep, wake up and cross fingers :D
mallox has quit [Client Quit]
<maier>
srk: I mean, I am getting the graphical installer to show up, so having some display should be possible after the installation, too. I just don't get to that point.
<Hail_Spacecake>
it looks like nixos generates a script in the systemd service that referneces the `lego` binary by name
<srk>
maier: uh, graphical installer has just nouveau driver
<Hail_Spacecake>
so I can't run this script myself unless I do nix-env -iA lego
<maier>
srk: Do you think it's worth trying to specify just "nouveau" then? But shouldn't it be the default if I don't use "nvidia"? THing is with the default, after rebooting into the new system I just get visual artefacts from the console, and no DM starts.
<srk>
Hail_Spacecake: I doubt it's nixops vs nixos issue
<maier>
I just checked and lspci only finds one VGA, which is the correct NVidia Geforce GTX 1660 SUPER.
simba2 has quit [Ping timeout: 272 seconds]
<Hail_Spacecake>
srk: yeah it looks like on the box if I try to restart the systemd service, it fails
<Hail_Spacecake>
same error that nixops reports
<Hail_Spacecake>
but for whatever reason running nixos-rebuild switch on the box doesn't try to restart that service, so it doesn't fail
<Hail_Spacecake>
I guess a systemd timer is what is supposed to run this, and I haven't had this running long enough for lets encrypt to try to renew on its own
<srk>
maier: yup, it should be the default. installer works but not after installation?
<Hail_Spacecake>
I do successfully have ssl certs working
cfricke has quit [Ping timeout: 264 seconds]
<srk>
Hail_Spacecake: either timer or changes in configuration
<Hail_Spacecake>
I don't think there are differences in configuration
<Hail_Spacecake>
I copied the .nix files I'm trying to apply with nixos from the box itself
<srk>
I mean, nixos-rebuild switch will only restart services which conf. have changed (or package version changed)
<maier>
srk: yes, exactly. I tried with the default configuration.nix, but it gets stuck in the terminal with graphical blurry lines (so does the installer, but I can do the sleep/wake thing there after starting the display-manager service).
<srk>
Hail_Spacecake: only if service file differs from previous generation
<Hail_Spacecake>
srk: is that different from nixops' behavior?
<Hail_Spacecake>
in any case I don't think it would be a problem restarting the renewal services if they worked, the problem is they don't work
<{^_^}>
nix#4469 (by dhess, 1 day ago, open): etag assertions
<maier>
Then I tried adding the nvidia driver after finding someone with the same issue. But that driver doesn't download even though the link is valid. I updated the nix-channel now, and I'll try again. Then I'll check out "nvidiaBeta", which I read somewhere.
<nerdypepper>
deleting ~/.cache/nix does not seem to work
<srk>
maier: can you try with kernelPackages = pkgs.linuxPackages_latest;
<Hail_Spacecake>
seems like there's some lets encrypt state that is probably missing
<Hail_Spacecake>
but I was able to successfully create these certs at some point
<srk>
maybe nixpkgs.pkgs NixOS option could be used
saschagrunert has joined #nixos
<srk>
I'm pretty sure there was a way (or multiple) to do that with older nixops as well
<srk>
many ways to nix.. :)
<maier>
srk: I think your linuxPackages_latest made "nouveau" work. Once in the new system on the command line (still not lightdm working), I could switch to "nvidia" and the download would finally come through! Now it seems to all work!
<maier>
srk: thanks again
<invra>
hey guys, i am wanting to install nixos on a btrfs subvolume i got ready... i am using the sh <(curl -L https://nixos.org/nix/install) --daemon
<invra>
i am wondering if i can specify the actual location of the installù
<invra>
i dont want it to touch the current running system
<maier>
invra: This does not install NixOS, but rather Nix the package manager IIRC.
<invra>
ah
<Flakebi>
supersandro2000: I fixed the merge conflicts of #108363 (kakoune plugins)
<colemickens>
you're sure the error is because of the symlink?
dannyp has joined #nixos
<asymmetric>
weird, i can't repro anymore.. i git-added the files, wonder if there's some weird interaction between the two things
<asymmetric>
the flake.nix the symlink pointed to hadn't been staged yet
invra85 has joined #nixos
aexoxea has joined #nixos
<invra85>
would anyone be so kind to review my config before i try my first install? https://dpaste.com/67YMXJJUJ
<yorick>
invra85: I think this is fine, btrfs at your own risk :)
<invra85>
what you mean
<yorick>
enableAllFirmware is a lot of firmware that you probably don't need
<invra85>
hmm
<invra85>
what hosuld i put there?
<yorick>
you can try enableRedistributableFirmware unless you have weird devices
<invra85>
should
<yorick>
have you used btrfs before?
<invra85>
yes i ahve 2 OS's installed on btrfs since 5 years
<yorick>
okay, then it's fine :)
<invra85>
what about Allfirmware?
<invra85>
i think im missing grub or it will install on its own?
<yorick>
I think you can remove it and hardware-configuration.nix includes the redistributable firmware already via not-detected.nix
<yorick>
invra85: you have systemd-boot enabled, not grub
<invra85>
allright ill comment that line
<invra85>
will it mess with my other grub install?
<invra85>
or with windows perhaps
<yorick>
you may want to consider hardware.cpu.amd.updateMicrocode = true or hardware.cpu.intel.updateMicrocode = true
<yorick>
definitely, yes
<invra85>
i rather disable that then and jsut manage grub from my other OS
<yorick>
nixos really likes to manage grub, since it's used for booting and rollbacks
<yorick>
it creates a new entry every time you build a new config
<yorick>
you can try boot.loader.grub.useOSProber to use the os-prober (no idea how that works), or boot.loader.grub.extraEntries to specify entries for your other OS's
sangoma has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @sternenseemann opened pull request #110674 → solo5: add missing test dependencies → https://git.io/JtGdb
<yorick>
invra85: that should work, semicolon on line 7
<invra85>
aha ok
<invra85>
need to understand the syntax
simba2 has joined #nixos
<invra85>
You must set the option ‘boot.loader.grub.devices’ or 'boot.loader.grub.mirroredBoots' to make the system bootable.
<invra85>
- You cannot use both services.xserver.displayManager.defaultSession option and legacy options (services.xserver.desktopManager.default and services.xserver.windowManager.default).
<{^_^}>
[nixpkgs] @SuperSandro2000 merged pull request #109981 → lsyncd: build on darwin → https://git.io/Jtt1j
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed commit from @nmattia to master « lsyncd: build on darwin (#109981) »: https://git.io/JtGFe
hiro98 has quit [Ping timeout: 256 seconds]
<invra85>
seems im doing it wrong
fuiltilt has quit [Quit: WeeChat 2.9]
<yorick>
invra85: I see, add boot.loader.grub = { enable = true; efiSupport = true; }
<{^_^}>
[nixos-homepage] @github-actions[bot] pushed commit from GitHub Actions to master « Update flake.lock and blogs.xml [ci skip] »: https://git.io/JtGFE
<invra85>
i added this boot.loader.grub = { enable = true; efiSupport = true; useOSProber = true; };
gulplante has quit [Client Quit]
pjt_tmp has quit [Ping timeout: 256 seconds]
<yorick>
yeah, your paste was missing the final semicolon, sorry
<invra85>
- You cannot use both services.xserver.displayManager.defaultSession option and legacy options (services.xserver.desktopManager.default and services.xserver.windowManager.default).
<invra85>
this is the last error im getting
m0rphism1 has joined #nixos
<invra85>
should it jsut be : desktopManager = { pantheon.enable = true; default = "elementary"; lightdm.enable = true; };
_madalu has quit [Remote host closed the connection]
<clever>
invra85: if you want to boot via legacy bios on gpt, you must create a bios boot partition, 1mb, not formatted, not mounted
<invra85>
im on uefi
<clever>
invra85: then you must set boot.loader.grub.devices = [ "nodev" ];
<clever>
that disables the legacy path
zupo_ has joined #nixos
<JaakkoLuttinen[m>
Can I somehow see what is the current `stateVersion` in use in my running system? I don't have it specified at the moment in `configuration.nix` so I'd like to set it but so that it's the same that my system is currently using.
<invra85>
clever and remove device = "/dev/sda" ?
<clever>
invra85: yeah
<clever>
JaakkoLuttinen[m: run nixos-option on the option name
<asymmetric>
JaakkoLuttinen[m:
<asymmetric>
> nixos-option system.stateVersion
<{^_^}>
undefined variable 'nixos-option' at (string):471:1
<JaakkoLuttinen[m>
clever: asymmetric Thanks!
<invra85>
i dnt understand :(
<invra85>
where do i setup my fstab?
<clever>
invra85: the fileSystems key in (hardware-)configuration.nix
<invra85>
i ahve to do that manually?
<clever>
invra85: nixos-generate-config will pre-fill hardware-configuration.nix for you
zupo has quit [Ping timeout: 265 seconds]
<invra85>
grub-install: error: /boot doesn't look like an EFI partition.
<invra85>
install-grub.pl: installation of GRUB EFI into /boot failed: No such file or directory
<asymmetric>
JaakkoLuttinen[m: to know the full option name, `man configuration.nix`
<clever>
invra85: you must mount the efi system partition to /boot before you run nixos-install
<invra85>
where do i specify where my efi partition is
<invra85>
on boot or /boot/efi?
<clever>
invra85: and you must re-run nixos-generate-config if you change what is mounted
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed 2 commits to master: https://git.io/JtGAP
<invra>
how do i isntall a kernel?
<virkony>
lukegb: that's actually something that was bothering me. what is official way to have user scripts in nixos? (so far I was updating all my scripts with env so they work on any POSIX OS)
<JaakkoLuttinen[m>
Hmm.. `services.uwsgi` doesn't seem to create the directory `runDir` so my uwsgi systemd service kept failing until I manually created the directory and set its permissions properly. Is this how it's supposed to be? Couldn't that directory and its permissions be created automatically?
<invra>
i dont know how to troubleshoot this system its totally different from anything i used for the apst 10 years
<invra>
:/
<lukegb>
yeah, it's usually just in the nix store if you're using grub
<lukegb>
check your grub.cfg, what's in it?
ahmedelgabri has joined #nixos
<invra>
linux ($drive2)/@nixos/nix/store/2q66pa0wqaly3d6fg9d52wvjq21narjl-linux-5.9.16/bzImage
<lukegb>
invra: iirc you'll either need boot.loader.efi.canTouchEfiVariables or to manually use efibootmgr to point your system at grub
<{^_^}>
[nixpkgs] @erictapen pushed commit from @LeSuisse to release-20.09 « libgadu: 1.11.2 -> 1.12.2 »: https://git.io/JtGx2
<virkony>
lukegb: I have mine set to false and I still got it UEFI configured automatically. Though my BIOS before didn't used it
<lukegb>
I can't parse that sentence
<virkony>
lukegb: boot.loader.efi.canTouchEfiVariables is false for me (nixos-options say). But when were installing last week it automatically updated my BIOS with booting via EFI
<cloud>
It looks like libnvidia-container doesn't support the new cgroup hierarchy, which means both the docker and podman nvidia container runtime derviations will stop working without setting a specific kernel parameter to disable the hierarchy. Is it kosher to set a kernel parameter as part of a nixos module?
ahmed_elgabri has quit [Ping timeout: 240 seconds]
<JaakkoLuttinen[m>
If a package works only on linux, is it sufficient to put `meta.platform = platforms.linux;` or do I need to specify something else too like `broken` attribute?
<KarlJoad>
Is there a way to mark a package as broken so that those derivations are not evaluated at all? I have a function with an invalid argument and the function returns a derivation where meta.broken = true. BUt the function is still evaluated and fails.
<{^_^}>
[nixpkgs] @Infinisil opened pull request #110707 → Bring back `types.functionTo`, finally → https://git.io/JtZkZ
<ldlework>
DigitalKiwi: hmm so what do you suggest
rajivr has quit [Quit: Connection closed for inactivity]
<{^_^}>
[nixpkgs] @jonringer merged pull request #96037 → python38Packages.coloredlogs: disable tests on darwin → https://git.io/JUvrj
<{^_^}>
[nixpkgs] @jonringer pushed commit from @tobim to master « python38Packages.coloredlogs: disable tests on darwin »: https://git.io/JtZku
<DigitalKiwi>
ldlework: idk :( git bisecting nixpkgs is always `fun` but it's nice to be able to point to a specific commit as breaking it... if it finds the right commit :|
<ldlework>
it takes like an hour to build my machine
vidbina_ has quit [Ping timeout: 246 seconds]
nvmr has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Lord_of_Life has quit [Read error: Connection reset by peer]
zakame has joined #nixos
<DigitalKiwi>
and then i got distracted with different pattern matches... and the curious phenomena of it finding double the amount the next time i ran it
<tpw_rules>
i want to nix-store --export everything that's in there before i run a garbage collection that could potentially delete things i care about
Guest53963 is now known as d3od
noudle has quit [Client Quit]
ahmedelgabri has quit [Ping timeout: 265 seconds]
<immae>
tpw_rules: why does "ls -1 /nix/store" not do what you want?
<tpw_rules>
if i just pass that to nix-store --export it barfs on some things
<tpw_rules>
and says they are not valid store paths
<{^_^}>
[nixpkgs] @bjornfor pushed commit from @romildo to master « lxqt: fix oxygen-icons5 attribute path »: https://git.io/JtZqL
xcmw has joined #nixos
Darkmatter66 has quit [Ping timeout: 246 seconds]
<immae>
tpw_rules: you could try to use nix-store --dump-db
<immae>
(it needs some parsing though)
<tpw_rules>
oh i figured it out: nix-store --export $(nix path-info --all) > everything
endformationage has joined #nixos
kfiz has quit [Ping timeout: 240 seconds]
<tpw_rules>
so this leads into my next question: i've made a flake with a devShell in it. if i do nix develop --profile somewhere then it creates a gc root and everything is saved
kfiz has joined #nixos
<tpw_rules>
with the exception of the nixpkgs source used. so if i collect garbage and run nix develop again, it has to redownload the nixpkgs version but nothing else. how do i fix that
gustavderdrache1 has joined #nixos
<DavHau[m]>
If I'm using nix from `pkgsStatic.nix`, essential stuff like `builtins.fetchGit` or `builtins.unsafeDiscardStringContext` is missing and nix raises attribute errors as soon as I evaluate anything.
<DavHau[m]>
Any ideas what could be the reason? It only happens with the static nix build.
sss2 has joined #nixos
gustavderdrache has quit [Ping timeout: 265 seconds]
vidbina_ has joined #nixos
ahmed_elgabri has quit [Ping timeout: 272 seconds]
gustavderdrache1 is now known as gustavderdrache
zupo has joined #nixos
aveltras has quit [Quit: Connection closed for inactivity]
berberman has joined #nixos
berberman_ has quit [Ping timeout: 264 seconds]
<tpw_rules>
hm, nix develop does not run the shellHook
fendor_ has quit [Remote host closed the connection]
fendor has joined #nixos
fendor has quit [Remote host closed the connection]
<tpw_rules>
is this a bug? it seems to specifically unset it, then try to run it
fendor has joined #nixos
fendor has quit [Remote host closed the connection]
<fuzzypixelz>
I was able to get all the dependencies with a nix-shell and avoid the installer using pip because that didn't work
<virkony>
I think something changed with packages search on nixos packages search over past days. It looks cool, but I miss ability to check out sources
<fuzzypixelz>
now there is this binary dmg2img
<fuzzypixelz>
virkony: Yes! I too noticed it today, kodos to the people working on it!
cfricke has quit [Ping timeout: 244 seconds]
<virkony>
oh. found it. it is hidden in gray color :)
<fuzzypixelz>
anyway, opening this binary as text format, I see this path: (...)/lib64/ld-linux-x86-64.so.2
<fuzzypixelz>
I know I can patch it using patchelf
<fuzzypixelz>
but I'm not sure what library to provide it or where the library even is
justanotheruser has quit [Ping timeout: 260 seconds]
lsix1 has joined #nixos
<fuzzypixelz>
okay so this is apparently the dynamic linker
<energizer>
gchristensen: i noticed frequent gc in your config, wondering what the advantage is. are you conserving disk space, or some other reason?
<gchristensen>
it was to make lorri very much needed, by GC'ing my shells constantly to be super annoying
winternya has joined #nixos
<gchristensen>
I don't do that anymore
fendor has joined #nixos
<winternya>
I
attila_lendvai has joined #nixos
attila_lendvai has quit [Changing host]
attila_lendvai has joined #nixos
bitmapper has quit [Quit: Connection closed for inactivity]
<energizer>
it's still annoying in that it has to fetch sources again, no? i tried it for a while but it was constantly redownloading nixpkgs
<winternya>
I'm not sure exactly what I'm doing wrong in my udev rules list - but `chmod` in the generated file points to `/usr/bin/chmod`, which obviously doesn't work. I assume it has something to do with me having to add `coreutils` to `services.udev.path`, but already tried setting it to `[ pkgs.coreutils ]`, and that didn't work. What exactly am I doing
<turbo_MaCk>
Is it possible to override desktopItem definition in my overlay?
samhza has quit [Ping timeout: 260 seconds]
<Emantor>
Whats the policy on forked libraries? I have a pyserial PR open for 2 years now and have since been maintaining a fork since than. Should I package the fork for my other program?
* Emantor
strikes through the double since
winternya has joined #nixos
steve-chavez has quit [Ping timeout: 246 seconds]
<{^_^}>
[nixpkgs] @petabyteboy opened pull request #110716 → beamPackages: use callPackagesWith from nixpkgs lib → https://git.io/JtZ3L
<fuzzypixelz>
YES! I succesfully patched a binary! Only had to put in 3 so but still!
<{^_^}>
[nixpkgs] @Emantor opened pull request #110717 → python3Packages.labgrid: init at 0.3.0 → https://git.io/JtZ3Y
Mr_Cue has quit [Remote host closed the connection]
Mr_Cue has joined #nixos
zupo_ has joined #nixos
samhza has joined #nixos
zupo has quit [Ping timeout: 256 seconds]
<winternya>
nvm, got it working
lsix1 has quit [Ping timeout: 264 seconds]
jakubgramsz[m] has joined #nixos
<worldofpeace>
Emantor: pyserial appears to be maintained... hmm. IMHO forked libraries don't serve a huge purpose in nixpkgs, and I think other distro's don't really allow forks of the same software to be packaged. But there is no actual policy in nixpkgs.
<worldofpeace>
leo60228: I think the why is more important than the what in this situation?
<leo60228>
the specific error is "Could not load a pixbuf from icon theme. This may indicate that pixbuf loaders or the mime database could not be found."
<worldofpeace>
leo60228: I believe whatever application this is has a packaging issue? Is it in nixpkgs?
<leo60228>
it's an application i'm developing locally
<leo60228>
if i compile the binary within the nix-shell and then run it outside it's fine
<leo60228>
so i think some environment variable must be getting messed up?
<worldofpeace>
I actually know the exact issue, I think. once sec
<worldofpeace>
I'm also not too sure how to do this with a shell.nix when developing locally
<worldofpeace>
oh yeah that will run. So if you add gdk-pixbuf you should have that in the environment
<leo60228>
worldofpeace: adding gdk_pixbuf didn't help but adding wrapGAppsHook did, thanks!
<leo60228>
(i only needed wrapGAppsHook to be clear, not both that and gdk_pixbuf)
fuzzypixelz has quit [Ping timeout: 264 seconds]
<winternya>
On other Linux systems I've used, when starting `i3`, the initial workspace is `1`. But, in the NixPkgs build (on NixOS at least), the default workspace is `10`. Anyone know why that is?
<ldlework>
I have a speaker system which has bluetooth capability. My iphone can connect to it with a press of a button. When I use blueman or bluetoothctl I cannot seem to connect to the device.
<ldlework>
Anyone handy with bluetooth?
<worldofpeace>
leo60228: ... I'm not sure how. maybe it was find_gio_modules
<ldlework>
I can pair with it, but not connect to it.
<leo60228>
looking at the code, i think it might be bc wrapGAppsHook adds librsvg
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/JtZGU
<samueldr>
(with pulseaudio, I don't know for other sound servers)
<leo60228>
worldofpeace: gdk-pixbuf is in gtk3's propagatedBuildInputs
lsix1 has joined #nixos
turbo_MaCk has quit [Quit: WeeChat 3.0]
hiro98 has quit [Ping timeout: 265 seconds]
fuzzypixelz has joined #nixos
<viric>
meh. I just tried a cross-built binary for raspberryPi in a raspberryPi, and I get illegal instruction
<ldlework>
worldofpeace: samueldr I'm guessing the problem is that bluetooth wants to use A2DP, but I can't enable that as per https://nixos.wiki/wiki/Bluetooth#Enabling_A2DP_Sink because I'm using Jack with the pulseaudio sink
leotaku has quit [Ping timeout: 240 seconds]
<ldlework>
it wants me to run
<ldlework>
pacmd set-card-profile "$(pactl list cards short | egrep -o bluez_card[[:alnum:]._]+)" a2dp_sink
<samueldr>
viric: "raspberryPi" is an overloaded term :)
<ldlework>
but I doubt that's gonna work
<viric>
samueldr: I mean pkgsCross.raspberryPi
<KarlJoad>
I am trying to build an application written in Go. It needs a few Go packages, but I can't find them in nixpkgs. The compiler attempts to download them from GitHub (which is forbidden). How can I get the packages defined?
<samueldr>
ldlework: right, sorry I don't know outside of pulseaudio how to
<viric>
samueldr: It should be the pi 1
<samueldr>
viric: yep
<samueldr>
1 and zero
rb2k has quit [Ping timeout: 240 seconds]
<samueldr>
and since you're using pkgsCross, great, you're actually using cross-compilation
<samueldr>
(not that qemu-user thing)
<viric>
the illegal instruction is: vldr d18, [pc, #380]
<samueldr>
viric: tried the simple "hello" package?
<{^_^}>
[nixpkgs] @SuperSandro2000 merged pull request #110716 → beamPackages: use callPackagesWith from nixpkgs lib → https://git.io/JtZ3L
<worldofpeace>
KarlJoad: several ways. is it packaged using a go modules?
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed 2 commits to master: https://git.io/JtZGB
winternya has quit [Quit: Connection closed]
<samueldr>
viric: I can't really help you much more, but at least that should help point towards something in the base toolchain vs. something specific to what you tried to cross-compile
vidbina_ has quit [Ping timeout: 256 seconds]
<viric>
samueldr: the instruction is in x265... I'll check how it built. VFP should be fine, but there are versions of VFP I think
<samueldr>
viric: the way you speak you're more likely to find the trouble than I am :)
<viric>
:)
winternya has joined #nixos
<KarlJoad>
worldofpeace: I'm not sure. I haven't done any work with Go before. I just want to download a package to use real quick. The instructions are vague. They just say to use `make`.
<samueldr>
but yeah, also verifying if it's the package itself or something more basic that fails can help as a first step
<winternya>
KarlJoad: what's the package?
<KarlJoad>
In the go file, the other packages are imported with their URL go Github.
<jumper149>
How can I use a cachix binary cache system wide (in configuration.nix)? Just run cachix with sudo??
mmkarakaya has quit [Quit: Connection closed]
steve-chavez has joined #nixos
grimler32 has joined #nixos
<grimler32>
Hi Are There Any Al1ens Or T1me Travelers Who can HELP ME PLEASE?
<energizer>
jumper149: nix.binaryCaches maybe?
philr has joined #nixos
<jumper149>
energizer: That sounds better
<fuzzypixelz>
grimler32: pretty sure there are many Humans here willing to help, not sure if that's good enough
hnOsmium0001 has joined #nixos
<grimler32>
well about 25% of people are in disguise (really Aliens so I hope i trck down one of those willing to help or travelers visititing from places like D1267GT10
<ldlework>
I am able to run glxgears, and it runs on my GPU. However, when I run chrome or firefox they fail to utilize the gpu for webgl and the like.
<jumper149>
energizer: Doesn't seem to work, I'm still compiling..
<infinisil>
winternya: Not sure what you mean
<energizer>
winternya: afaik you cant
<winternya>
suspected so
<infinisil>
Ah yeah I get it
<infinisil>
Nope can't do that
<winternya>
alright i'll just redefine the defaults
<winternya>
cheers
<energizer>
i think that's a mistake in how nixos defines its options
<winternya>
yeah
<energizer>
not in the library/type/module design, just the namespacing
<tpw_rules>
fuzzypixelz: i'm getting an error about "attribute "nixosSystem" missing'. do i have to turn on an option when i import nixpkgs to get that to show up?
<energizer>
it would be more flexible to provide defaults as environment.systemPackages = nixosPolicyDefaults.environment.systemPackages - ["bashInteractive"]
<energizer>
er it would be more flexible to provide defaults as environment.systemPackages = nixosPolicyDefaults.environment.systemPackages - [bashInteractive]
<tpw_rules>
hm, it seems it's not reading my configuration from the place i expect
ahmedelgabri has joined #nixos
<tpw_rules>
oh. because it wasn't tracked by git yet
<infinisil>
energizer: At some point I'll push for making systemPackages be `types.attrsOf types.package`
<infinisil>
Or rather, `types.attrsOf (types.nullOr types.package)`
<infinisil>
Then you can specify packages with `systemPackages = { inherit (pkgs) ripgrep tmate; }`
rb2k has joined #nixos
<infinisil>
And you can remove ones with `systemPackages.ripgrep = mkForce null`
<infinisil>
Added benefit is that you get an error if you add multiple versions of the same package (-> you need to explicitly specify a different attribute name to do that)
<winternya>
Why is the GTK file picker full screen and "zoomed in" no matter what? Not really sure how I'd fix this...
fendor has quit [Read error: Connection reset by peer]
<winternya>
(not using GNOME)
jaexilgilroyPhil has joined #nixos
ahmed_elgabri has quit [Ping timeout: 265 seconds]
ahmed_elgabri has joined #nixos
ixxie has joined #nixos
Ok3be has joined #nixos
Ok3be has quit [Client Quit]
steve-chavez has quit [Ping timeout: 264 seconds]
attila_lendvai has quit [Ping timeout: 256 seconds]
<ixxie>
howdy folks :) I'm running into an issue [1] involving `gnome-session-restart-dbus.service` and I need to tweak it; how would I find the corresponding Nix expression so I could override it?
woobilicious has quit [Quit: Ping timeout (120 seconds)]
<tpw_rules>
would i be missing anything from nix-env if i switched over to nix profile for general use?
<tpw_rules>
on non-nixos specifically
kalbasit_ has joined #nixos
civodul has quit [Quit: ERC (IRC client for Emacs 27.1)]
<cole-h>
ldlework: How? For future generations who are led to your plight via search engines :)
<jgart[m]>
Is is possible to put my ip address directly in this option: `services.nginx.virtualHosts."8.8.8.8"`
<jgart[m]>
any help is greatly appreciated
domogled has quit [Quit: domogled]
<cole-h>
jgart[m]: Your world-visible IP, or your local IP?
<energizer>
,locate bin/blake2
<ldlework>
cole-h: I had created nvidia offload wrappers for chrome and chrome-dev so that they were gpu powered when I ran them, but as of latest unstable this seems to *break* GPU support for applications instead of enabling it for them.
<ldlework>
I apparently don't need any environment variables to get gpu support now.
<{^_^}>
Found in packages: nss.tools, nss_3_53.tools, nss_latest.tools
<ldlework>
Would be so nice if nixos documented this stuff..
<cole-h>
ldlework: Well, it is unstable :^)
<ldlework>
"Hey we completely inverted the semantics of Optimus in the latest changes."
<ldlework>
you know
<ldlework>
something
<cole-h>
:D
<jgart[m]>
<cole-h "jgart: Your world-visible IP, or"> I want my world-visible IP to be used in that option
<cole-h>
jgart[m]: Not possible, unfortunately.
<jumper149>
energizer++
<{^_^}>
energizer's karma got increased to 29.000000000000004
<jumper149>
energizer: Actually the caching is working now :)
<jgart[m]>
<cole-h "jgart: Not possible, unfortunate"> so should I point DNS then?
<energizer>
nice
<ldlework>
day of my life i'm never getting back
<jgart[m]>
that's the only option?
<cole-h>
jgart[m]: That would be my suggestion.
<jgart[m]>
cole-h: ok thank you!
<cole-h>
The thing is: the only way (I know of) to get your world-visible IP is with a `dig` command pointed at a DNS resolver (like Google's or OpenDNS's)
<cole-h>
Which Nix can't do for you
<cole-h>
Because it's both impure, and requires networking (both of which violate the sandbox)
<ldlework>
kinda nice that no environment variables are needed now
<infinisil>
jgart[m]: The problem is likely to the right
<infinisil>
You can scroll horizontally in that window
zeta_0 has quit [Quit: rcirc on GNU Emacs 27.1]
<{^_^}>
[nixpkgs] @alyssais opened pull request #110728 → autoreconfHook: make overridable → https://git.io/JtZRp
<tilcreator>
I'm trying to build https://github.com/NixOS/nixpkgs/blob/master/pkgs/tools/graphics/waifu2x-converter-cpp/default.nix with a custom version (the newest commit hash). So I cloned nixpkgs replaced `rev = "v${version}";` with `rev = "27ad20e6a2007d0415a7a1d7fc957e81395ab06b";` and tried building it with `nix-build -E 'with import <nixpkgs> {}; callPackage ./default.nix {}'`, I expected an error, because the sha256 is wrong,
<tilcreator>
but instead it just tried building without changing the version. How do I build my modified `default.nix`?
<{^_^}>
#109841 (by zhaofengli, 6 days ago, open): powerdns-admin: init at 0.2.3
<zhaofeng>
tilcreator: You need to change the sha256 to all zeroes or `lib.fakeSha256` so Nix will be forced to pull the sources
<tilcreator>
zhaofeng: Worked, thx!
<DigitalKiwi>
if i wasn't already so annoyed with these NOT A DATABASE errors and not being able to upgrade without deleting all of my messages and there not being a way to export/import them(?) :| ...this would be funny https://mostlyabsurd.com/files/2021-01-24-235627_2880x1800_scrot.png