<{^_^}>
[nixpkgs] @Radvendii reopened pull request #38632 → win-dll-link: Fix to look for DLLs in lib/ → https://git.io/vx7nB
<{^_^}>
[nixpkgs] @tesq0 opened pull request #91452 → unity3d: fix editor not launching due to missing libraries → https://git.io/JfhCE
hoverbear has joined #nixos
Rusty1 has joined #nixos
matthewcroughan has joined #nixos
<{^_^}>
[nixpkgs] @veprbl merged pull request #91452 → unity3d: fix editor not launching due to missing libraries → https://git.io/JfhCE
<{^_^}>
[nixpkgs] @veprbl pushed commit from @tesq0 to master « unity3d: fix editor not launching due to missing libraries (#91452) »: https://git.io/JfhWJ
quinn has quit [Ping timeout: 240 seconds]
quinn has joined #nixos
proofofkeags has quit [Remote host closed the connection]
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #nixos
boxscape has quit [Quit: Ping timeout (120 seconds)]
boxscape has joined #nixos
lunaa has joined #nixos
<{^_^}>
[nixpkgs] @sikmir opened pull request #91453 → asuka: init at 0.8.0 → https://git.io/JfhWB
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JfhWR
<bqv>
hey os
<bqv>
got a strange problem
<bqv>
nixos-rebuild now hangs, for me
atlas_ has joined #nixos
polypus74 has quit [Quit: WeeChat 2.8]
boxscape has quit [Client Quit]
<bqv>
i'm not sure what it's even restarting
<infinisil>
Random suggestion: bash -x nixos-rebuild switch
<bqv>
(would that even work?)
<infinisil>
(try it :))
jbox has quit [Ping timeout: 240 seconds]
<bqv>
hmm, works on the rebuild script, but not anything inside, it seems
atlas_ is now known as jbox
<bqv>
but no matter. if there's no immediately obvious solution i'll just dig into it when i have more time later
zebrag has joined #nixos
<infinisil>
Should at least be able to know what step inside hangs like that
<bqv>
it's hanging in part of the systemctl start stage
<bqv>
it's fine if i kill systemctl
<bqv>
but otherwise it's stayed up for hours
<bqv>
#justSystemdProblems
<gchristensen>
that isn't necessary
inkbottle has quit [Ping timeout: 256 seconds]
erasmas has joined #nixos
orivej_ has joined #nixos
orivej has quit [Ping timeout: 264 seconds]
erasmas has quit [Ping timeout: 264 seconds]
<pie_>
anyone know why pip just hangs in virtualenv now
<pie_>
ive had trouble with other things too before...
<Dagger>
test with `wget <whatever URL pip was pulling>`; that should give you a better idea of what was failing and where rather than just "it was IPv6"
<Dagger>
because it's rarely IPv6, but rather some breakage in something which should be fixed rather than throwing an entire protocol out
<{^_^}>
[nixpkgs] @ConradMearns opened pull request #91454 → obsidian: init at 0.7.3 → https://git.io/Jfhlk
<pie_>
i think i had issues with wget at some point but curl worked
<pie_>
thAT MIGHT HAVE BEEN UNRELATED THOUGH, was two weeks ago
h0m1 has quit [Ping timeout: 272 seconds]
h0m1 has joined #nixos
MtotheM has quit [Read error: Connection reset by peer]
MtotheM has joined #nixos
Rusty1 has joined #nixos
motesc has joined #nixos
anderslundstedt has quit [Ping timeout: 246 seconds]
dansho has quit [Ping timeout: 246 seconds]
kristofers has quit [Remote host closed the connection]
anderslundstedt has joined #nixos
orivej has quit [Ping timeout: 240 seconds]
miguel_clean has quit [Ping timeout: 240 seconds]
orivej has joined #nixos
<OmnipotentEntity>
If I have a cargo build that I'm attempting to perform in a nix shell, and cargo downloads a .so file for the purpose of linking. I would need to patchElf that .so file to make it work in this environment, right? Is there an easy way to get that to happen? (Should I make a nix build script, will nix force cargo to patch dependencies?)
<OmnipotentEntity>
For clarity, this is a project that I'm writing and changing, not a completely finished project, if that would change the approach.
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #91457 → Fixes from prefixed pkg config → https://git.io/Jfh8v
<aterius>
Is there a good way to determine why you aren't hitting the hydra cache? I'm overriding a package to not use unfree and still seem to be having to build from source
orivej has quit [Ping timeout: 246 seconds]
orivej has joined #nixos
krjt has joined #nixos
matthewcroughan has quit [Ping timeout: 240 seconds]
<simpson>
In general, overrides aren't going to be cached, since you're the only one with the override. nixpkgs would have to also carry that same package configuration and Hydra would have to attempt to build it.
krjt has quit [Client Quit]
<aterius>
Yes, I'm trying to override a global setting (allowUnfree) to match the hydra setting for a specific package
camsbury has joined #nixos
<simpson>
Oh, I see. Curious; that should be fine.
<aterius>
the former hits the cahce, the latter doesn't
<aterius>
* the former hits the cacge, the latter doesn't
felixfoertsch23 has joined #nixos
felixfoertsch has quit [Ping timeout: 265 seconds]
felixfoertsch23 is now known as felixfoertsch
erasmas has joined #nixos
orivej has quit [Ping timeout: 260 seconds]
aw has quit [Quit: Quitting.]
spacefrogg has quit [Quit: Gone.]
aw has joined #nixos
spacefrogg has joined #nixos
erasmas has quit [Ping timeout: 246 seconds]
motesc has quit [Quit: Leaving]
dansho has joined #nixos
cyphase has quit [Ping timeout: 256 seconds]
choward7 has joined #nixos
cyphase has joined #nixos
choward has quit [Ping timeout: 260 seconds]
choward7 is now known as choward
lunaa has quit [Remote host closed the connection]
apache8080 has joined #nixos
<apache8080>
I am trying to setup nixos on a raspberry pi and am running into issues. I can boot into the default aarch64 sd image from the pi but when I run nix-rebuild switch on my modified config and then reboot u-boot just gets stuck at Starting kernel. I followed this tutorial https://citizen428.net/blog/installing-nixos-raspberry-pi-3/ . I also found
<apache8080>
issues that said to try reverting to the 4.19 kernel and I still get the same error. One other thing that happens is when I run rebuild switch and then reboot and then try to boot into the original configuration from the original config from the default sd image
<apache8080>
Any ideas as to what could be happening?
dansho_ has joined #nixos
<samueldr>
3B or 3B+?
<samueldr>
there were reports that the 3B+ had an issue with display out and some versions of the kernel
dansho has quit [Ping timeout: 240 seconds]
<samueldr>
I don't know what's the state though
<samueldr>
though "stuck at Starting kernel..." doesn't mean that the kernel isn't booting... confusingly enough...
<{^_^}>
[nixpkgs] @jonringer pushed commit from @marsam to master « pythonPackages.internetarchive: 1.9.3 -> 1.9.4 »: https://git.io/Jfh0v
<apache8080>
samueldr whats weird is that I can't even boot into the previous working configuration
<samueldr>
the one person that reported issues directly to me at one point unplugged everything, plugged back everything and didn't have issues anymore :/
gxt_ has joined #nixos
<samueldr>
so it doesn't look like trivially reproducible
<evax>
when defining a flake input using git over ssh, is it possible to specify a tag for repo? The `ref` field in the input attributes set seems to only accept a branch name
<zangi>
ashkitten: end users don't have to clone the github repositories to compile packages, but putting nixos repositories in a non-free hosting service where nixos packages are stored is kinda... like you even need root access to be able to install the packages, and the package manager is accessing the non-free hosting service as root
alp has joined #nixos
<zangi>
on github at least you can clone it as a normal user
philr_ has quit [Ping timeout: 265 seconds]
<ashkitten>
i'm not sure i understand. are you suggesting that aws might be maliciously serving the wrong files?
<ashkitten>
iirc someone was talking about package signing a while ago
<ashkitten>
actually, the binary cache is signed, isn't it? i can't remember
<ashkitten>
if it is, then there's no problem
<regnat>
zangi ashkitten the packages are indeed signed when built, and the signature is checked when you fetch them from the cache
<ashkitten>
regnat++
<{^_^}>
regnat's karma got increased to 1
<ashkitten>
zangi: does that answer your concern?
alp has quit [Ping timeout: 272 seconds]
<zangi>
okay, but why did the nixos decided to use amazon? like most distros have their own hosting services?
<ashkitten>
presumably because it was the cheapest/best option
drakonis has joined #nixos
octe has joined #nixos
drakonis_ has quit [Read error: Connection reset by peer]
drakonis_ has joined #nixos
<energizer>
which part of nixos is on aws?
abbe_ has joined #nixos
<DigitalKiwi>
the 250+TB of package cache, i think
<DigitalKiwi>
solution: if it using S3 bothers you, the solution is don't use the cache, build everything yourself
<simpson>
zangi: You are under no obligation to use the cache. You can build everything as if you were on Gentoo.
lsix has joined #nixos
<zangi>
nixos is a source distro?
drakonis has quit [Ping timeout: 260 seconds]
<DigitalKiwi>
yes
<DigitalKiwi>
cache.nixos.org is a convenience but not necessary
<simpson>
I'm not the expert, but AIUI some folks donated the resources which power both the community Hydra, which builds packages, and the community cache, which serves packages.
<ashkitten>
hydra just builds everything exactly like your local machine can
<ashkitten>
hydra just tends to be many times faster and more powerful than your local machine
<radvendii>
What's with the animals on the shirts?
NickHu has joined #nixos
Guest98635 has joined #nixos
ZerataX has joined #nixos
<ashkitten>
radvendii: the mascots for nixos releases!
<radvendii>
ohh gotcha
nanashi0x74Old[m has joined #nixos
aquarial has joined #nixos
vtest[m] has joined #nixos
<simpson>
s1341: AFAICT, we don't configure the subnet at all. You could play with virtualisation.docker.extraOptions.
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
ky0ko has joined #nixos
felschr[m] has joined #nixos
<simpson>
Or maybe, I can't quite tell from what I'm reading, you can configure `docker0` as you like, and Docker will use whatever's got that name? I'm not sure at all.
toraritte has joined #nixos
watzon has joined #nixos
pitch has joined #nixos
philipp[m]2 has joined #nixos
eddyb has joined #nixos
worldofpeace has joined #nixos
MilkManzJourDadd has joined #nixos
Yakulu[m] has joined #nixos
Notkea[m] has joined #nixos
o1lo01ol1o has joined #nixos
Nacho[m]1 has joined #nixos
<zangi>
does nixos provide mirrors for the cache?
siraben has joined #nixos
spinlock[m] has joined #nixos
kai_w has joined #nixos
rycee has joined #nixos
mindtree[m] has joined #nixos
bqy has joined #nixos
iwvt[m] has joined #nixos
sonerhalis[m] has joined #nixos
JameySharp[m] has joined #nixos
gnxlxnxx[m] has joined #nixos
marius851000[m] has joined #nixos
nanashi0x74[m] has joined #nixos
leonardp has joined #nixos
jluttine[m] has joined #nixos
apple_at_cha[m] has joined #nixos
regivanx[m] has joined #nixos
thequux[m] has joined #nixos
roberth has joined #nixos
faya01[m] has joined #nixos
<ashkitten>
it's 250+TB, as stated before, so no.
<ashkitten>
but it's behind fastly's cdn so it is very fast
<energizer>
tbh you can buy 250TB HDD for like 2 grand
<dminuoso>
ashkitten: Im not so sure about that, actually.
<radvendii>
wow, so if I decide to run a really old version of nixos, I can still use the binary cache?
<ashkitten>
radvendii: yes
<dminuoso>
ashkitten: Especially some of the older distributions maintain long histories of packages, with many different flavours like stable, long term stable, experimental, power users, in different architectures.
<energizer>
yeah i'd be surprised if ubuntu wasn't storing petabytes
<dminuoso>
For each release...
<ashkitten>
hmm yeah, i wasn't thinking about that
<ashkitten>
but their archives are generally separate from their main repos
<DigitalKiwi>
arch doesn't store old versions
<ashkitten>
arch doesn't do a lot of things, tbf
<dminuoso>
arch has by far the best wiki for general linux knowledge.
<dminuoso>
For what its worth..
<ashkitten>
i agree
<ashkitten>
i used arch from 2012 to 2017 or so
<ashkitten>
but that doesn't change what i said
<zangi>
cool, does the cache keep track of everything since the beginning of nixos?
<energizer>
zangi: ya
<radvendii>
I should try out arch again...
<energizer>
at least that's what they say. i have no idea how to verify
<radvendii>
I've tried a couple of times, but i wasn't nearly as linux savvy then
<ashkitten>
arch is a good distro to learn deep with
NeoCron has joined #nixos
<bahamas>
is there a way for me to access the config of a service in configuration.nix? specifically, I want to enable backups with restic and I want to specify a path that is written to by another service
<lewo>
Only nixpkgs commits build by Hydra are pushed to the binary cache (6 per day)
<gthm>
does anyone here use spotifyd? I tried it in nix-shell yesterday but as far as I could tell it didn't register any dbus paths, so I couldn't do anything with it
<radvendii>
It has the stable channels and unstable
<DigitalKiwi>
nixos-unstable is rolling
<radvendii>
Oh, okay. I stand corrected. I thought there was some technical difference between them
<DigitalKiwi>
idk
<zangi>
is the stable the equivalent of debian stable (frozen packages)?
<DigitalKiwi>
seems to fit the definitions i know but i'm sure some rule lawyer could debate it
<radvendii>
Like it would be rolling release if we just used master or something
<radvendii>
I feel like I've seen people correct this before, but I didn't pay enough attention to actually remember the details
<radvendii>
does anyone have experience with objdump?
asymptotically has joined #nixos
<radvendii>
I'm trying to fix a script we have that find .dll dependencies. It uses objdump to find all the dependencies of a .exe/.dll file, but it seems to just not be finding all of them.
knupfer has joined #nixos
<{^_^}>
[nixos-homepage] @github-actions[bot] pushed commit from GitHub Actions to master « Update flake.lock and blogs.xml [ci skip] »: https://git.io/JfhVA
Chiliparrot has joined #nixos
<{^_^}>
[nixpkgs] @magnetophon opened pull request #91471 → industrializer: init at 0.2.6 → https://git.io/JfhwW
bahamas has quit [Ping timeout: 240 seconds]
<s1341>
simpson: thanks. fwiw the solution was to add virtualisation.docker.extraOptions="--bip=<my custom subnet>"
<simpson>
s1341: Rad.
__monty__ has joined #nixos
boxscape has joined #nixos
<fendor>
,locate pdflatex
<{^_^}>
Found in packages: tetex, texlive.combined.scheme-full, texlive.combined.scheme-basic, texlive.combined.scheme-small, texlive.combined.scheme-medium, texlive.combined.scheme-minimal
<{^_^}>
[nixpkgs] @peti pushed 2 commits to haskell-updates: https://git.io/JfhrW
ris has quit [Ping timeout: 272 seconds]
<{^_^}>
[nixpkgs] @peti pushed 836 commits to haskell-updates: https://git.io/JfhrK
erasmas has joined #nixos
<gregf>
Would someone help me get through making my first diff for a pkg upgrade? Stuck on the build process not sure where I'm going wrong. -vvvv hasn't been super helpful to me.
zupo_ has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed 2 commits to haskell-updates: https://git.io/JfhrP
<evils>
gregf: sure, paste your code (and maybe the errors too)
zupo__ has joined #nixos
<gregf>
evils: ty, 1 sec
<{^_^}>
[nixpkgs] @peti opened pull request #91474 → Update Haskell package set to LTS 16.2 (plus other fixes) → https://git.io/JfhrM
<KarlJoad>
In Nix, both the then and else bodies are required in an if expression. Must both options evaluate to the same type too, like in Haskell?
<gregf>
It mentions a sha256sum being wrong, but I can't tell where it's coming from. It's not the hash I put in the diff.
<zangi>
,help
<{^_^}>
Use `,` to list all commands, `,foo = Foo!` to define foo as "Foo!", `,foo =` to undefine it, `,foo` to output "Foo!", `,foo somebody` to send "Foo!" to the nick somebody
<zangi>
,
<{^_^}>
Special commands: find tell locate expand inclusive-language random-pr - Commands sorted by use count, page 0 (use ,<n> to view page <n>): help unstable libraries paste escape'' tofu declarative escape" library overlay ping -a pinning ifd unfree profiling ask botsnack which-channel channels xml pr tofu-vim xy wololo fancy-uninstall imperative nixlang++ cache pointers invite pills runtimedeps stateversion home-manager matrixbridge smart-questions
nixbitcoin has joined #nixos
FRidh has joined #nixos
<evils>
,tofu gregf
<{^_^}>
gregf: 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
<gregf>
evils: what if i'm using emacs?
<evils>
press `0` 52 times?
<evils>
or paste in the example above :P
sigmundv_ has joined #nixos
<gregf>
so your saying i put all 0's in for the hash and it's going to replace them on the first build?
<nixbitcoin>
No it will prompt you with the TOFU hash
bahamas has joined #nixos
<evils>
no, it will say it wanted 00000..., and got [the hash of what it got]
<gregf>
build-build isn't prompting me
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « haskell-repline: update override for ghc-8.10.x to the latest version »: https://git.io/JfhoO
<gregf>
i guess I gotta go back to the manual
<evils>
you want to change the fetchurl sha256 to `14c8js503pvj3yfqfscn5anznl005a9fhbwybbd3qdamrfjsw87z` i think
<evils>
you gave it a wrong hash, it tells you what it got instead, you're trusting what it got is right, hence trust on first use
<gregf>
k
dermetfan has joined #nixos
<evils>
you'll probably run into the same issue for the fetchFromGitHub
<gregf>
@evils the problme is the hash mismatch it's talking about doesn't even match the hash i put in there
<gregf>
and updating it to what it says gives the same error
* tnks
What makes a derivation require the "allowUnsupportedSystem" attribute? I'm on a Mac, and I figured out a simple fix for the ansifilter CLI tool. And I can modify the derivation to get it to build, but it still is marked as unsupported in the Nixpkgs expression.
<evils>
i think the hash it says it wanted is the hash you gave it, but using the custom encoding used here
* tnks
accidentally /me-prefixed that question. Tab-complete, I guess.
<gregf>
either way putting what it asks for in there gives the same error
bahamas has quit [Ping timeout: 260 seconds]
<tnks>
Nevermind... using client wrong...
<evils>
gregf: did you fill in what it got, or what it wanted? (`wanted` is the expected hash based on what you provided in the default.nix)
* tnks
Okay, I think I know... I need to mark the platforms explicitly in the metadata.
<evils>
gregf: and don't copy over the `sha256:` part
<gregf>
evils: yes
<gregf>
what part do i copy then? I guess tht's where i'm going wrong
<tnks>
damn... hard to train myself not to do a command-Enter, because that's how I have it set up on other clients.
<ldlework>
Ah screw it, Docker should be fine for this.
alp has joined #nixos
knupfer has quit [Read error: Connection reset by peer]
<gthm>
in case anyone is wondering about by q above about spotifyd: (a) we don't build it with dbus support (https://github.com/NixOS/nixpkgs/issues/71362) and (b) I can control it fine through the spotify app on my phone, so I guess it works as advertised
o1lo01ol1o has joined #nixos
<{^_^}>
#71362 (by NilsIrl, 35 weeks ago, open): Spotifyd service doesn't work
<radvendii>
does anyone know where fixupOutputHooks gets *used*? I'm looking in the github, and i see a bunch of `fixupOutputHooks+=(foo)` statements, but that's it...
<Taneb>
:( it seems like the GithubPulls set-up on my work's Hydra server has stopped making new jobsets since I upgraded the version?
<Taneb>
There's still .jobsets making the json fail as per usual
bahamas has joined #nixos
rembo10 has joined #nixos
<zangi>
,locate nix-index
<{^_^}>
Found in packages: nix-index
<zangi>
,locate bin/nix-locate
<{^_^}>
Found in packages: nix-index
proofofkeags has joined #nixos
Thra11 has quit [Ping timeout: 265 seconds]
pingiun has joined #nixos
fuzen3 has joined #nixos
<Taneb>
*json file
proofofkeags has quit [Ping timeout: 246 seconds]
jakobrs has left #nixos ["WeeChat 2.7.1"]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<{^_^}>
[nix] @rodarima opened pull request #3749 → Fall back to copyPath if link fails with EPERM → https://git.io/Jfhi0
rtb has joined #nixos
<rtb>
Error getting name to temp file from template /boot/kernels/initrd-secrets.XXXXXXXX: Parent directory (/boot/kernels/) does not exist at /nix/store/8d8gd2jrkfip2kryiy381jjbk5m8sqmg-install-grub.pl line 368.
<fps>
rtb: can you pastebin that pl script from your store?
<fps>
for some reason it seems that line 364: mkpath(dirname($initrdSecretsPath), 0, 0755); didn't result in the directory existing?
orivej has joined #nixos
<rtb>
fps: I just did: #mkdir /boot/kernels
<rtb>
fps: and now it works... but mkpath yes...
<rtb>
will reboot
<fps>
oh, the script uses $bootPath, not /boot/. hmmm
<fps>
nah, that should be ok
<rtb>
$bootPath can be boot or efi, or?
rtb has quit [Quit: WeeChat 2.8]
alp has quit [Ping timeout: 272 seconds]
rtb has joined #nixos
erasmas has joined #nixos
erasmas has quit [Ping timeout: 256 seconds]
<fps>
rtb: well, might be worth filing a bug :)
mananamenos has joined #nixos
<mananamenos>
Hi, I want to create a hakyll project using stack so I could also use it in windows with stack. I started adding `nix: enable: true` to ~/.stack/config.yaml. Then, `nix-shell -p stack haskellPackages.hakyll`. Then, `stack exec hakyll-init my-site` gives me an error *Executable named hakyll-init not found on path:*. What am i doing wrong?
<{^_^}>
[nixpkgs] @worldofpeace merged pull request #91459 → treewide: remove deprecated value `Application` from makeDesktopItem → https://git.io/JfhBt
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/JfhX1
ph88 has quit [Remote host closed the connection]
pingiun has joined #nixos
kleisli_ has joined #nixos
orivej has quit [Ping timeout: 258 seconds]
orivej has joined #nixos
alp has joined #nixos
kreyren has quit [Remote host closed the connection]
kreyren has joined #nixos
kleisli__ has quit [Ping timeout: 240 seconds]
<radvendii>
mananamenos: why do you need to use `stack exec`. can't you just run `hakyll-init my-site`? it's been a while since i've initialized a hakyll project though
D_ has quit [Ping timeout: 260 seconds]
<mananamenos>
radvendii, i have never used stack. I followed the tutorial of hakyll's website, where they do `stack exec..`
<radvendii>
oh.
<radvendii>
I'm tired
<radvendii>
I'm not using stack
D_ has joined #nixos
<radvendii>
I'm using cabal
<radvendii>
hahaha, I think I ran into a bunch of problems setting it up with stack so I switched to cabal
karetsu has joined #nixos
<{^_^}>
[nixpkgs] @sternenseemann opened pull request #91485 → dune-release: init at 1.3.3 → https://git.io/Jfh1z
orivej has quit [Ping timeout: 260 seconds]
orivej_ has joined #nixos
<duairc>
I'm trying to use the latest NixOps, but it uses flakes which I haven't used before
m0rphism has joined #nixos
<duairc>
I'm trying to create a nixpkgs overlay such that the nixops it provides will be the nixops from git master
<duairc>
Ah, I figured it out. I needed to add ".default" to the expression I was using
<{^_^}>
[nixpkgs] @Mic92 merged pull request #91479 → fix NUR by fixing python3.pkgs.{jaraco_itertools,jaraco_logging,tempora} → https://git.io/JfhKj
<{^_^}>
[nixpkgs] @primeos pushed 2 commits to release-20.03: https://git.io/Jfh9h
<stepcut>
I am using nixops to deploy to a server. Whenever a new version is deployed to the server, I'd like the server to send an email to my team. I'm thinking just a one-liner at the end of the activation script. That would also be run if the server got restarted -- which isn't a bad thing. I feel like there is any easy way to do this that I am overlooking
magnetophon has quit [Ping timeout: 256 seconds]
ebopp has joined #nixos
magnetophon has joined #nixos
<{^_^}>
[nixos-homepage] @github-actions[bot] pushed commit from GitHub Actions to master « Update flake.lock and blogs.xml [ci skip] »: https://git.io/JfhH3
<{^_^}>
[rfc-steering-committee] @worldofpeace pushed to worldofpeace-patch-1 « Some agenda updates »: https://git.io/JfhQk
<{^_^}>
[rfc-steering-committee] @worldofpeace opened pull request #40 → Some agenda updates → https://git.io/JfhQL
<regnat>
stepcut: Apparently there's a deps attribute that you can add to the activationScripts, maybe you can do something very dirty to make it depend on every other activationScript. But that would start to get nasty
<BugeyeD>
hi all. trying to develop with flask and ckeditor; flask-ckeditor package is not available, and i'm unable to install/build it via virtualenv/pip ... can anyone point to a resource which shows me how to get around this?
orivej has quit [Quit: No Ping reply in 180 seconds.]
o1lo01ol1o has quit [Read error: No route to host]
o1lo01ol1o has joined #nixos
o1lo01ol_ has quit [Ping timeout: 240 seconds]
fresheyeball has quit [Ping timeout: 260 seconds]
blisszs has joined #nixos
codygman has quit [Read error: Connection reset by peer]
fresheyeball has joined #nixos
codygman has joined #nixos
dermetfan has quit [Ping timeout: 240 seconds]
erasmas has joined #nixos
<patagonicus>
excelsiora: Isn't working how?
erasmas has quit [Ping timeout: 265 seconds]
codygman has quit [Read error: Connection reset by peer]
codygman has joined #nixos
<KarlJoad>
excelsiora: nix-collect-garbage will only delete things that have no symlinks pointing to it, roughly speaking. Also, if you want to remove packages installed system-wide, you need to run the command with sudo.
<Rusty1>
with -d ?
knerten1 has joined #nixos
growpotk- has quit [Ping timeout: 256 seconds]
<clever>
yeah
knerten2 has quit [Ping timeout: 264 seconds]
alexherbo2 has quit [Remote host closed the connection]
<clever>
excelsiora: youll need to manually delete something non-nix on the same FS
eri has joined #nixos
eri is now known as eri451
<{^_^}>
[nixpkgs] @mdlayher opened pull request #91498 → maintainers/teams: add mdlayher to golang team → https://git.io/Jfhbf
<patagonicus>
excelsiora: Any reason you're specifying --max-freed 1? I'm trying to find docs on it, but I think with that nix will only try to free a single byte.
<excelsiora>
I already deleted one project
<excelsiora>
I'll try some other files about 10 MB
<excelsiora>
`sudo nix-collect-garbage -d` seemed to make progress but then same error
<sephii>
I have a channel "foobar" that exposes 3 derivations (let's say "foo", "bar" and "baz"). How can I use a nix expression with one of these derivations with nix-env -i? I've tried "nix-env -i -f '<foobar>' --from-expression 'foobar: foobar.foo.withPlugins [ foobar.bar ]'" but I'm getting "value is a function while a set was expected". Any idea why?
<excelsiora>
I'm basically out of things to delete I think, maybe journals
<OmnipotentEntity>
excelsiora, are you actually having a disk io error? (Check dmesg)
<OmnipotentEntity>
oh out of disk space.
drakonis has joined #nixos
<pbogdan>
excelsiora: you don't have any root reserved space on your fs you could free up?
<{^_^}>
[nixpkgs] @AndersonTorres pushed 2 commits to master: https://git.io/JfhbO
<excelsiora>
not that I know of, aside from journals, I'm googling how to delete those (nginx for example)
<OmnipotentEntity>
if you have no need for them excelsiora, you can probably kill the logs safely, they're located in /nix/var/log/nix/drvs
drakonis_ has quit [Ping timeout: 256 seconds]
<OmnipotentEntity>
on my system they're another 38M
growpotkin has joined #nixos
<OmnipotentEntity>
which is way more than the db itself, at least on my box.
<excelsiora>
so rm -rf /nix/var/log/nix/drvs/
<excelsiora>
?
<OmnipotentEntity>
you'll probably want to not remove the folder itself, but the subfolders within. But I'm not an expert here. I don't *think* the logs are used by nix. But I can't say that categorically.
<bqv>
Oh, that reminds me, theres no nix-command to restore a NAR file
<bqv>
Is that intentional?
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<bqv>
You can dump one but not read one
erasmas has joined #nixos
<srk>
nix-store --restore ?
drakonis1 has joined #nixos
<excelsiora>
clever: I say that because I did `sudo nix-collect-garbage -d`
drakonis has quit [Ping timeout: 260 seconds]
<bqv>
srk: nix2
<srk>
right
<srk>
even --restore is not exactly what you asked for since you need --import as well
<bqv>
I noticed cause I've aliased away all the nix-* commands
o1lo01ol_ has joined #nixos
ckauhaus is now known as ckauhaus[afk]
o1lo01ol1o has quit [Ping timeout: 240 seconds]
dingenskirchen has quit [Quit: dingenskirchen]
dingenskirchen has joined #nixos
mallox has quit [Quit: WeeChat 2.8]
drakonis1 has quit [Quit: WeeChat 2.8]
<excelsiora>
clever: but I think I understand what you were suggesting now. So it could have worked after I deleted the 10 MBs of files if I had used that argument...
mallox has joined #nixos
<clever>
excelsiora: yeah
<clever>
then nix would delete a small amount, and gradualy work its way up as you increase that number
ckauhaus[afk] has quit [Quit: WeeChat 2.7.1]
<evanjs>
Might be missing something, but I'm confused by a snippet in the nixos manual
<evanjs>
"You must set the option boot.loader.systemd-boot.enable to true. nixos-generate-config should do this automatically for new configurations when booted in UEFI mode.
<evanjs>
I currently have all my UEFI systems running with grub... Am I misreading this snippet? It seems to imply that systemd-boot is required for UEFI
<KarlJoad>
evanjs: No, you're not reading it wrong. As far as I know, systemd-boot is the main/default boot option for UEFI, but you can still use grub if you want to.
sayanarijit[m] has joined #nixos
o1lo01ol1o has joined #nixos
<fresheyeball>
so I am serving my nix cache with serveSssh
<fresheyeball>
and everything is working
<evanjs>
KarlJoad: interesting. Thanks! I guess I was expecting it to read like "systemd-boot _or_ grub" if that was the case, etc
<fresheyeball>
but I have a friend and I want to serve it to them now
<fresheyeball>
they generated a new ssh key as root
<fresheyeball>
and I added it to my nixos
<fresheyeball>
if they run ssh nix-ssh@fresheyeball.com
<fresheyeball>
they get a connection, (run as root)
<fresheyeball>
but when doing a nix build they get permission denied
<fresheyeball>
I don't remember what to do here. How can I make sure this ssh key is supplied to the nix build, so that they can hit my cache?
zaeph has joined #nixos
o1lo01ol_ has quit [Ping timeout: 240 seconds]
<evanjs>
fresheyeball: I remember seeing something like adding a command + key to authorized_keys
<KarlJoad>
evanjs: If you think its warranted, you could always open an issue on GitHub to make it clear in the documentation that if you use UEFI, you can use both.
<patagonicus>
evanjs: If logging in as regular root with a shell works, it means that the key is in authorized keys without restrictions.
<patagonicus>
fresheyeball: I'm guessing that nix is downloading with the build users, not root and those can't find the key.
<evanjs>
patagonicus: *goes back* ah, root, yeah I see that now lol
<fresheyeball>
patagonicus: right
<fresheyeball>
but how to fix? I just don't recall
<evanjs>
I do remember going in circles with this as well :D
<evanjs>
KarlJoad: yeah, sort of a nitpick, as I'm not sure if it's really that much of an issue, just could be more explicit, as always :D
<BugeyeD>
is a more complete resource available to walk a nixos user through "making things work" where python virtualenv and pip is required due to lack of nix packages?
rajivr has quit [Quit: Connection closed for inactivity]
sangoma has joined #nixos
orivej has quit [Quit: No Ping reply in 180 seconds.]
<simpson>
BugeyeD: For example, one .envrc I have three lines, `use nix; layout python3; pip3 install litecli` and it does what I want, first loading a shell.nix and then using the CPython from the resulting environment to set up virtualenv and pip.
knerten has joined #nixos
<sephii>
You can even combine it with a shell.nix if there are some libraries you can't compile (like psycopg2, lxml, stuff like that), to use the Nixos' ones
<BugeyeD>
simpson / sephii - thank you very much, going to try to read through this and implement right now. i have wasted so much time on this ... fairly new to nixos, really want to use it. but so far it's really cramping my python. been using virtualenv for so long ...
<BugeyeD>
simpson++
<{^_^}>
simpson's karma got increased to 25
knerten1 has quit [Ping timeout: 240 seconds]
<BugeyeD>
sephii++
<{^_^}>
sephii's karma got increased to 3
<BugeyeD>
not sure exactly how that helps but there you go
<sephii>
I've read in various places "don't use nix-env, use configuration.nix". While I get the advantage of using the declarative way of installing packages, how do you do that on non-nixos systems (ie. when you just use the nix package manager)?
<simpson>
One can prefer nix-shell to nix-env.
<gustavderdrache>
i use nix-shell for projects and "nix run nixpkgs.<foo>" for one-offs
Chris82 has joined #nixos
<bbigras>
sephii: you can use home-manager if it's only for your user.
<Chris82>
Yo! Where is the cron log saved on nixos?
gilligan has quit [Quit: WeeChat 2.7.1]
<bbigras>
maybe try `journalctl _COMM=cron`
WIIIHEIIM has quit [Remote host closed the connection]
ris has joined #nixos
orivej has quit [Ping timeout: 265 seconds]
<bbigras>
found that with google. no idea tbh
orivej has joined #nixos
nabataeus has joined #nixos
<sephii>
(I'm asking because I'd like to add installation instructions to an app, inviting people to use Nix to install it, but I'm not sure what's the best way of explaining how to install it. I was thinking about `nix-channel --add` and then `nix-env -i`, but I was wondering if I should advise something else since people are advocating against `nix-env`)
<Chris82>
It provides a lot more information, thanks bbigras!
o1lo01ol_ has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
<duairc>
Basically I'm importing something else which imports <nixpkgs>, but I want the nixpkgs it imports to be a specific nixpkgs that's known to the other file, rather than the system nixpkgs
<mananamenos>
When i do `nix-shell`, i need to open emacs from the resulting shell for it be able to use the executables nix-shell has brought to scope. Is there a way to "hook" to shell's env from the already opened emacs?
<quinn>
does anyone have experience with netboot images? i'm trying to set up something like config.system.build.isoImage but for makeNetboot
<bbigras>
mananamenos: you can use direnv with emacs
MidHotaru has quit [Quit: Connection closed for inactivity]
<bbigras>
I use it with lorri to be able to use nix-shell with many projects.
<quinn>
bbigras: not ideal, but useful still. thank you!
<quinn>
bbigras++
<{^_^}>
bbigras's karma got increased to 1
<mananamenos>
bbigras, but isn't direnv for env variables (key/value string pairs)? I want to add brittany for example to mkShell and then use it from emacs
<bbigras>
quinn: you're welcome.
<bbigras>
mananamenos: I use direnv to have a different Rust version for projects. so they have different `cargo` and `rustc` packages. I also add node stuff for web projects.
<quiet_laika[m]>
i have an oddly large number of GHCs hanging around in my nix store, in spite of the fact that i've `rm` all the `result` links i can find that should anchor them... whats the best way to debug why a certain package is in my store?
<{^_^}>
[nixpkgs] @cdepillabout pushed 2 commits to haskell-updates: https://git.io/Jfjep
choward has quit [Ping timeout: 256 seconds]
cybrian has joined #nixos
<quiet_laika[m]>
i wonder if lorri does any cache cleanup on its own
<quiet_laika[m]>
maybe i should file an issue
<symphorien>
I think it keeps exactly one root per shell
<{^_^}>
[nixpkgs] @svmhdvn opened pull request #91508 → haskellPackages: cleaning up obsolete overrides (2/x) → https://git.io/Jfjvv
<quiet_laika[m]>
actually that makes sense
<quiet_laika[m]>
i think i had just unfortunately had 6 projects with a difference GHC pin each
<quiet_laika[m]>
different*
<symphorien>
Maybe a "lorri delete" command might make.sense
<quiet_laika[m]>
?yeah, `lorri cleanup` could be nice but im not sure if lorri has any way to check whether the projects themselves still exist on disk
choward has joined #nixos
<quiet_laika[m]>
the question mark was supposed to be at the end of that...
<symphorien>
Or lorri should remove the root for deleted shell.nix
seku has joined #nixos
<symphorien>
Maybe it does that, dunno
<quiet_laika[m]>
i guess that would be in the purview of the daemon?
<quiet_laika[m]>
yeah i didnt actually delete any of the projects themselves, just cleaned up the nix `result`s, so lorri is probably behaving sensibly here
erasmas has joined #nixos
<quiet_laika[m]>
heh, my `nix-collect-garbage` before blasting the lorri cache found like <10GB, and after >25GB
<BugeyeD>
simpson / sephii - looks like i'm off to the races now. plus i learned about litecli, which will be added to my list of useful bit as well. thanks again.
<bbigras>
`lorri cleanup` would be nice if it wouldn't touch recently used projects
zupo has quit [Ping timeout: 265 seconds]
zupo has joined #nixos
ashesham` has joined #nixos
lopsided98 has quit [Remote host closed the connection]
<quinn>
http://ix.io/2qaz okay so i'm evaluating this file with nix-build -A test, and it's giving me 'value is a path while a set was expected'. any advice?
lopsided98 has joined #nixos
<gustavderdrache>
> let nixpkgs = <nixpkgs>; in nixpkgs.pkgs
<{^_^}>
value is a path while a set was expected, at (string):317:29
<gustavderdrache>
if you want to use nixpkgs, you need import <nixpkgs> {}
fresheyeball has quit [Quit: WeeChat 2.7.1]
<quinn>
> let foo = <nixpkgs/nixos> {}; in foo
<{^_^}>
attempt to call something which is not a function but a path, at (string):317:11
<gustavderdrache>
don't forget the import keyword
<quinn>
> let foo = import <nixpkgs/nixos> {}; in foo
<quinn>
gustavderdrache: http://ix.io/2qaB so that snippet obviously worked, but i'm still getting the same error, very confused.
<gustavderdrache>
looks like it's due to your usage of nixpkgs.pkgs.symlinkJoin
knerten2 has joined #nixos
<whatisRT>
quinn: thanks!
Chris82 has quit [Remote host closed the connection]
andrewrk has left #nixos ["Leaving"]
<quinn>
gustavderdrache: you'd think, however nix-build -A netboot '<nixpkgs/nixos/release.nix>' works and that calls nixpkgs.pkgs.symlinkJoin so i'm wondering where i'm going wrong here
bahamas has quit [Read error: Connection reset by peer]
<gustavderdrache>
when you nix-build a file, it does an auto-call under the hood
proofofkeags has quit [Remote host closed the connection]
hoverbear has joined #nixos
proofofkeags has joined #nixos
<quinn>
sorry, i don't understand ;(
mdlayher has joined #nixos
knerten1 has quit [Ping timeout: 246 seconds]
<gustavderdrache>
one sec - i'll put a small example together
<gustavderdrache>
nix-build just fills in the defaults (while respecting --arg and --argstr options)
simba1 has joined #nixos
<gustavderdrache>
the import keyword doesn't have this behavior because it doesn't assume that it's importing a derivation or nixpkgs; it could just as easily be importing a small config file or utility function
<quinn>
gustavderdrache: so my error with pkgs.symlinkJoin is because i'm calling it with a malformed paths argument? and i need to import nixos differently to avoid that error?
drakonis_ has quit [Read error: Connection reset by peer]
<gustavderdrache>
ultimately, the issue is that writing <nixpkgs> in an expression is just the path to nixpkgs - on my system, it's in ~/.nix-defexpr/channels/nixpkgs
<gustavderdrache>
or <nixpkgs/nixos>, or similar
<gustavderdrache>
you need the import keyword to actually load the code from that path
<gustavderdrache>
and in the case of <nixpkgs>, you need to call the imported function with parameters
<quinn>
yeah, but my plainnixos is set to import "${nixpkgs}/nixos" { inherit system }; now, and it's still throwing that error.
<gustavderdrache>
let me try restructuring the paste a bit, one sec
<pie_>
it would be easier if i didnt have to figure out how to get reportlab to do what i want
<pie_>
which is what the label thingy uses
<philipp[m]>
I've got two PRs that have been basically ready since january but they were never merged (my fault). Anybody want to give them another look now?
<pie_>
adisbladis: can i pass a pypi package to poetry2nix and get nix files or do i need to run poetry myself to create the lock files and such, and then use that as input
whatisRT has quit [Ping timeout: 246 seconds]
<adisbladis>
pie_: The latter
sangoma has quit [Read error: Connection reset by peer]
drdee has quit [Ping timeout: 260 seconds]
<pie_>
ok
<adisbladis>
pie_: poetry2nix doesn't do code generation at all
fendor has quit [Remote host closed the connection]
<lucifer_alive>
Any rules that I should know about?
<adisbladis>
lucifer_alive: Be nice I guess?
<hoverbear>
lucifer_alive: Most folks treat this as very asyncronous so try to ask complete questions with all the information in one message and if folks have an idea how to help or want to resond they will send you a message with your name in it (that should highlight you)
<judson_>
Is there really not a ulogd package in nixpkgs?
<hoverbear>
lucifer_alive: Please respect all your fellow humans and bots. :) Be patient, have fun.
<qyliss>
Be nice, don't paste more than a few lines, it's okay to re-ask the same question if you don't get an answer, but give it at least a few hours because not everybody is looking at IRC all the time
<lucifer_alive>
Roger that! Thanks for the info hoverbear
<lucifer_alive>
qyliss: Got it thanks!
<qyliss>
welcome to IRC and NixOS :)
<adisbladis>
Remember that there are no stupid questions, we were all new once =)
<lucifer_alive>
adisbladis, thanks for the encouragement. This is already better than StackOverflow!
<{^_^}>
[nixpkgs] @bjornfor pushed commit from @neumantm to master « eclipse: add eclipse-committers »: https://git.io/JfjT3
stepcut has quit []
<pie_>
adisbladis: is it in scope for poetry2nix to have something that gives you the environment with a debuggable python and gdb with python extensions?
lsix has quit [Ping timeout: 244 seconds]
* pie_
tries to find his snippet
<adisbladis>
pie_: You can pass a custom python derivation to mkPoetryApplication/mkPoetryEnv
<pie_>
theres a lot of crap in the file that this is in but https://bpa.st/NJVQ
<pie_>
its not a big deal just kind of a pain in the ass to figure out
<pie_>
i guess these should be somewhere in nixpkgs
<pie_>
really the hard part is waiting for the first python build to fail because you didnt know to clear (or whatever) forbidden references, and finding where in the gdb repo the python stuff is
AdrianW has joined #nixos
whatisRT has joined #nixos
<adisbladis>
pie_: I don't really know what I'm looking at ^_^
<{^_^}>
#48809 (by CMCDragonkai, 1 year ago, open): Getting Python Extensions to GDB
<adisbladis>
pie_: But presumably you should be able to pass `debugpy` as `python` to mkPoetryApplication or mkPoetryEnv
<pie_>
yeah probably
<whatisRT>
has someone seen the error `ld: library not found for -lzzip` on macos? I'm trying to install texlive and get this error. I can't even find what library it's trying to link...
<pie_>
i dont actually need this right now, just throught it would be nice
<adisbladis>
pie_: I think it's out of scope,
<adisbladis>
But it could be documented how to do it
cr4y1 has quit [Remote host closed the connection]
<pie_>
yeah i think it makes sense that its out of scope, the pass a python func makes sense and this stuff is more generic anyway and should be part of pkgs
<pie_>
ok so im in a poetry2nix env but i cant import one of the dependencies in the shell
<pie_>
nix-shell default.nix -A shell --run "python -c 'import pylabels'"
<pie_>
so you can copy paste :P
Chiliparrot has joined #nixos
<pie_>
im probably rushing through and did something dumb
<pie_>
inb4 shell doesnt work properly or something
<pie_>
and i already made a bug report to poetry lol
<pie_>
(unrelated)
apache8080 has joined #nixos
lsix has quit [Ping timeout: 260 seconds]
cybrian has joined #nixos
<apache8080>
samueldr still running into the HDMI raspberry pi issues from last night. I was able to connect to the raspberry pi over serial and it is booting fine and with the 5.4 kernel the hdmi screen worked when I plugged it in after verifying the boot worked over serial. Then I tried to revert to the 4.19 kernel as the github issues said and I'm now seeing
<apache8080>
this error in the dmesg logs `[drm:drm_atomic_helper_wait_for_dependencies [drm_kms_helper]] *ERROR* [CONNECTOR:29:HDMI-A-1] flip_done timed out`
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<samueldr>
annoying :)
<samueldr>
sorry, don't really know much more what to do
<apache8080>
lol, literally just unplugged the hdmi and plugged it back in and it worked
orivej has quit [Ping timeout: 264 seconds]
<apache8080>
seems like I just need to keep serial connected to verify this stuff
lsix has joined #nixos
<samueldr>
I assume that other distros using the mainline kernel would have similar issues, unless there is one option not enabled in our kernel build that may help
orivej has joined #nixos
<apache8080>
whats the difference between the linuxPackages_rpi3 and linuxPackages_4_19
<apache8080>
they both seem to be the 4.19 kernel
<samueldr>
_rpi3 is from the raspberry pi foundation's fork
<samueldr>
_4_19 is mainline linux, but the older LTS release
<apache8080>
oh ok
justan0theruser has joined #nixos
WIIIHEIIM has quit [Remote host closed the connection]
marble_visions has joined #nixos
justanotheruser has quit [Ping timeout: 256 seconds]
knerten2 has joined #nixos
mv_test has joined #nixos
<mv_test>
test 1, 2
apache8080 has quit [Ping timeout: 245 seconds]
alp has quit [Ping timeout: 272 seconds]
<adisbladis>
pie_: `import labels` ^_^
<pie_>
doh figurtes
<pie_>
adisbladis: sorry for wasting your time 'xD
knerten1 has quit [Ping timeout: 246 seconds]
cybrian has quit [Read error: Connection reset by peer]
fendor has quit [Read error: Connection reset by peer]
virus_dave has joined #nixos
<virus_dave>
Heyas - i’m repeatedly seeing errors like this (with different files each time) that cause broken builds. These then succeed (or get farther) on retries. Any ideas? cannot link '/nix/store/.tmp-link-74253-935932891' to '/nix/store/.links/0dn7nr9j3hdqr32pd709y6imxmhi7d91iq8i481qq1j6f8vxgk7r': File exists
iceypoi has joined #nixos
<virus_dave>
googling (and last time this happened) seems to indicate that just clearing out the 90k+ files in my /nix/store/.links is OK, but it’s quite annoying to have to do this repeatedly
<gchristensen>
ooooh you hit the birthday problem with your filesystem
<infinisil>
gchristensen: There are 32^52 possible hashes though
<gchristensen>
you could run nix-collect-garbage moreo often to avoid hitting the birthday problem, or use a filesystem https://github.com/NixOS/nix/issues/1522
<{^_^}>
nix#1522 (by domenkozar, 2 years ago, open): Optimize store stops working at ~10M files
icey_ has quit [Ping timeout: 258 seconds]
<gchristensen>
infinisil: it is an issue with inode collisions, not filename collisions
asymptotically has quit [Ping timeout: 240 seconds]
<infinisil>
Ahh
dermetfan has quit [Quit: WeeChat 2.7.1]
dermetfan has joined #nixos
<infinisil>
> pow 32 52
<{^_^}>
0
<infinisil>
Not that many after all!
<gchristensen>
> ack 32 52
<gchristensen>
so few!
<infinisil>
:P
<{^_^}>
[nixpkgs] @kalekseev opened pull request #91512 → greg: add setuptools to propagatedBuildInputs → https://git.io/JfjLv
<virus_dave>
ackerman isn’t infinitely recursive. Work harder dammit, bot!
orivej has joined #nixos
<gchristensen>
> :v ack
<{^_^}>
ack = m: n: if m == 0 then n + 1 else if (m > 0 && n == 0) then (ack (m - 1) 1) else (ack (m - 1) (ack m (n - 1)))
<virus_dave>
Although i’m curious: 1522 doesn’t actually look to be the same problem, as far as i can see
<virus_dave>
i only have 90k files in .links, not 10M. And the error isn’t about “out of space” (inodes). Rather it seems race-condition-y, given that reruns then proceed just fine and either succeed or fail further on
LambdaDuck has joined #nixos
<virus_dave>
also, my FS isn’t ext4
apache8080 has joined #nixos
<qyliss>
What is it?
<apache8080>
samueldr I did some further testing and so if I let the boot complete and then login over serial then unplug and plug the HDMI back in then it works fine and this is reproducible
mv_test has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @vbgl pushed commit from @sternenseemann to master « ocamlPackages.{ipaddr,macaddr}: 3.1.0 → 5.0.0 and update all affected packages (tuntap, conduit, cohttp, …) (#91350) »: https://git.io/JfjL6
<tokudan>
whatisRT, yes, it will, provided that you have overwritten the channel and not added another channel
<energizer>
whatisRT: i prefer to manage my channels with root rather than my user account
<MtotheM>
is the container / virtualization part of nixOS rather than nix?
<sumina4>
Hi, very sorry if it's the stupidest question ever, but I've been introduced to NixOs and home-manager and got the advice "Except Xorg and stuff that can't work per user base, it's good to have most of your packages in `home.nix`. Is there any documentation to understand rational and good practice around what packages should be in `home.nix` and
<sumina4>
which should be in `configuration.nix` ?
<whatisRT>
tokudan: yeah, the unstable one doesn't appear anymore in `nix-channel --list`
<samueldr>
MtotheM: part of Nixpkgs and NixOS
<MtotheM>
allright
<mgdm>
sumina4: things like X have to start before you log in, so those kinds of things are better in configuration.nix. Same for any services you want to run. But any I put any user application (including Firefox etc) in home.nix
<tokudan>
whatisRT, then you should be fine. just watch out for the channels used per user, as mentioned by energizer, root channels are systemwide, but you can have user specific channels
<whatisRT>
tokudan: I'm using nix-darwin, which adds a channel. Can I somehow test if that influences my derivation?
<sumina4>
mgdm: so if I understand correctly, if I use `xfce`, I should put it in `home.nix` as it is past _login_ phase?
<tokudan>
whatisRT, I've got no experience with macos, sorry. what you can do is just rebuild something and see what you get
<qyliss>
You probably want to use the NixOS module for xfce, because you need it to be started automatically for you
orivej has quit [Ping timeout: 265 seconds]
orivej has joined #nixos
<mgdm>
sumina4: I have seen people put i3 in home-manager but personally I keep my desktop environment in configuration.nix
user_0x58 has joined #nixos
b42 has quit [Ping timeout: 256 seconds]
<sumina4>
qyliss, that's why I didn't find any xfce option on `xsession.windowManager` option I imagine (only stuff like i3, bspwm, xmonad and so on) ? I will look into NixOs modules then, thanks!
<sumina4>
mgdm: thanks
b42 has joined #nixos
<whatisRT>
tokudan: well, the issue is that it's not building in the first place, and I'd like to figure out if that that channel breaks something or if it was broken all along
<sumina4>
qyliss: I'm just trying to make sense of the rule of thumbs I've been told "if it's after login, it should be managed by home-manager" and if I understand correctly, after login, a `startx` with a `startxfce4` in the xinit would work in a normal system, so I would have put it naturally in home-manager. Am I missing an important concept here for the
<sumina4>
segregation between packages in `home-manager` and the ones in `nixos` ?
cybrian has joined #nixos
<quinn>
sumina4: there's no really sophisticated demarc. home-managner provides some functionality, nixpkgs/nixos provides some functionality. in some cases they overlap even]
kleisli has joined #nixos
<quinn>
if you're using both i would advise running ripgrep on both codebases and reading the output. if you can't read .nix files very well yet, check the manual/discourse/here/discord/git issues etc
proofofkeags has quit [Remote host closed the connection]
proofofkeags has joined #nixos
cr4y1 has joined #nixos
AdrianW has quit [Remote host closed the connection]
<apache8080>
samueldr So did timing thing (about 5.5 mins after boot + login) and then booted the pi with no peripherals waited 5.5 mins and then plugged the HDMI cable to the monitor and it says the `current input timing is not supported by the monitor display. Please change your input tiing to 3840x2160, 60hz `
<samueldr>
:/
<samueldr>
I really wonder what makes logging-in work better
<quinn>
apache8080: 4k on a raspberry pi? respect
<samueldr>
that should make no difference *system wise*
<samueldr>
3B+ even, so that doesn't even sound right?
<apache8080>
it should be able to drive the monitor based off of what I have seen online
<samueldr>
(I'm not sure whether the 3B+ can do more than 1080p. but the 3B can't)
<apache8080>
hmm ok
<samueldr>
(but it wouldn't matter, it just would do 1080p)
civodul has quit [Quit: ERC (IRC client for Emacs 26.3)]
<apache8080>
I'm going to try the same experiment except this time leave the hdmi plugged in then wait the 5.5 mins and then unplug and replug and see what happens
proofofkeags has quit [Ping timeout: 264 seconds]
simba1 has quit [Ping timeout: 260 seconds]
fabianhjr has quit [Quit: Leaving.]
Acou_Bass has quit [Ping timeout: 246 seconds]
Acou_Bass has joined #nixos
proofofkeags has joined #nixos
zangi has quit [Read error: Connection reset by peer]
zangi has joined #nixos
<apache8080>
samueldr so did the experiment I described above and after doing an unplug and replug the monitor just says no hdmi signal
<{^_^}>
[nixpkgs] @svmhdvn opened pull request #91519 → haskellPackages: cleaning up obsolete overrides (4/x) → https://git.io/JfjqE
<samueldr>
oh
<samueldr>
right
<samueldr>
the raspberry pi may be setting things up for the display earlier before the kernel and u-boot boots
<samueldr>
which then maybe the kernel's vc4 driver needs??
<duairc>
Is there any way to set NIX_PATH from inside Nix?
<duairc>
I have a nix file A, which imports B. B imports <nixpkgs>. But I really want it to import not the system nixpkgs, but a specific nixpkgs that's defined in A
<virus_dave>
B is behaving badly :| I don’t think there’s a way to do that from inside A.nix
<infinisil>
duairc: Then do `import expression-that-point-to-the-path-of-nixpkgs`
<duairc>
Currently I have a shell.nix that sets NIX_PATH and then I can nix-build A from inside that shell
<duairc>
infinisil: Unfortunately I can't edit B
zeta_0 has joined #nixos
m0rphism has quit [Quit: WeeChat 2.7.1]
fresheyeball has joined #nixos
<fresheyeball>
hey out there
<fresheyeball>
I am getting "hash mismatch in fixed-output derivation"
<quinn>
fresheyeball: how exatly are you getting this? nixos-rebuild, nix-shell, nix-build on a personal project, what?
alp has joined #nixos
<fresheyeball>
fetchGit in an application, with nix-build
<fresheyeball>
I am specifying the source with fetchGit
<quinn>
you mean in a build script for an application? like a default.nix you're calling with nix-build? if so, how about you post that somewhere and i'll look at it
<fresheyeball>
yes
<fresheyeball>
I have a default.nix file the builds a project
<fresheyeball>
and in there I set source for this package
<fresheyeball>
and now I get this weird error
<quinn>
what it probably is, is there's some derivation in nixpks that is getting code with a different checksum than it expects, so you're going to want to overlay the package until it gets fixed
<quinn>
fresheyeball: if you had a working version and then you changed something and it broke please post that too
<zeta_0>
what irc #channel should i go ask for help with setting up `keybase` ? searching `nixos keybase` in a web browser does not return anything useful, so i don't think my problem has anything to do with nix
<quinn>
zeta_0: nixos has keybase options under services.keybase and services.kbfs. if those aren't the issue or you don't know how to use those from looking at the module source, try looking for an irc channel on the keybase webpage. https://nixos.org/nixos/options.html#keybase
proofofkeags has quit [Remote host closed the connection]
proofofkeags has joined #nixos
<zeta_0>
quinn: thanks, i did not know that those nix options for keybase existed, the #keybase.io and ##keybase.io channels are empty though, i just installed the keybase nix pkg, that's probably why it's not working out of the box
<zeta_0>
i need to use keybase through my web browser through some website, so i hope this is not to painful to set up
<quinn>
remember that you can use that website i linked you to search for any option in the codebase, as well as nixos-option if you know the attr. already
<quinn>
you should probably read the manual before puzzling about the nixos options or package or anything if you're having trouble tho
wpcarro` has joined #nixos
ToxicFrog has joined #nixos
wpcarro` has quit [Remote host closed the connection]
orivej has quit [Ping timeout: 265 seconds]
o1lo01ol1o has quit [Remote host closed the connection]
<zeta_0>
quinn: agreed, thanks for the help, i'll let you guys know if i get stuck along the way
orivej has joined #nixos
proofofkeags has quit [Read error: Connection reset by peer]
o1lo01ol1o has joined #nixos
proofofkeags has joined #nixos
<{^_^}>
[nixpkgs] @RohanHart opened pull request #91524 → oracle-xe: make scripts executable → https://git.io/JfjmK
<betawaffle>
c00w: hey, you remembered me
karetsu has joined #nixos
<c00w>
betawaffle: yep :)
<betawaffle>
i'm gonna read through the PR comments, but let me know what you'd like help with
<aleph->
Hmm stupid question, why would I not be able to run binaries under ~/.nix-profile/bin/ in a shell script run by my user?
<c00w>
Uh - it's approved but about to get rejected *facepalm*
<c00w>
So if you could help argue that it's useful that would be awesome
nabataeus has quit [Remote host closed the connection]
<moet>
i have a trivial custom xsession `services.xserver.displayManager.session = [ { manage="desktop"; name="custom"; start=""; } ];` and this leaks processes.. it starts a `systemd --user` process which remains after logout and a `kwallet --pam-login` process which remains after login/logout
<moet>
if i login/logout three times, a singel `systemd --user` process remains, and three `kwallet --pam-login` processes remain
<moet>
i have no .xinitrc or .xsession (the homedir is empty)
mbrgm has quit [Ping timeout: 272 seconds]
mbrgm_ is now known as mbrgm
<moet>
i believe this is a bug in `nixpkgs/nixos/modules/services/x11/display-managers/default.nix` .. specifically in xsessionWrapper
<moet>
this is a bug because the session isn't ended even though no processes were started by user code (because the user code was empty)
<moet>
or rather, this is a bug because the session isn't ended by the wrappers around user code (what the user code does doesn't really matter)
knerten1 has joined #nixos
<moet>
if you then try to shutdown, you'll get errors that the user is still logged in, and the command `who` will identify the user as logged in without indicating any tty
<bqv>
betawaffle: excellent
<bqv>
(but the rice part is slightly separate)
knerten2 has quit [Ping timeout: 265 seconds]
whatisRT has quit [Ping timeout: 240 seconds]
orivej has quit [Quit: No Ping reply in 180 seconds.]
<moet>
does anyone here have experience with starting a custom xsession in nixos?
orivej has joined #nixos
<moet>
i've confirmed that this bug is present without any custom xsession
dermetfan has quit [Ping timeout: 260 seconds]
<moet>
if you log in with the showcased plasma environment, `services.xserver.desktopManager.plasma5.enable = true` using sddm `services.xserver.displayManager.sddm.enable = true;` using the built in "plasma5" session, and then log out, several times, you'll leak one `systemd --user` process and several `baloo_file` processes
Rusty1 has joined #nixos
justan0theruser has quit [Ping timeout: 246 seconds]
philr has joined #nixos
<moet>
cool, well, back to startx for me
moet has quit [Quit: leaving]
orivej has quit [Ping timeout: 265 seconds]
orivej_ has joined #nixos
karetsu has quit [Quit: WeeChat 2.8]
moet has joined #nixos
`slikts has quit [Ping timeout: 246 seconds]
`slikts has joined #nixos
`slikts has joined #nixos
`slikts has quit [Changing host]
thc202 has quit [Ping timeout: 260 seconds]
acarrico has quit [Ping timeout: 240 seconds]
codygman has quit [Read error: Connection reset by peer]
lawr3nce has joined #nixos
codygman has joined #nixos
cr4y1 has quit [Ping timeout: 264 seconds]
codygman has quit [Read error: Connection reset by peer]