<erictapen>
I already tried to tell the app via configureFlags that wayland-scanner is located elsewhere but they seem to ignore that
<jackdk>
monokrome: I'm sad you're having trouble with that thing. I was seriously considering a librem13 (though was holding out for new HW and 32GiB RAM)
<rycee>
infinisil: I added an edit in the description of home-manager#97 with a link to the manual.
<infinisil>
jackdk: You need 32GB? 16Gb + NVMe swap no good?
<colemickens>
"Librem comes with Coreboot + SeaBIOS payload. That means EFI boot is not possible. Use fdisk to partition hard drive, and GRUB as a bootloader"
<rycee>
infinisil: It's not very extensive manual but I think it shows an installation method that is a bit more intuitive 🙂
<samueldr>
I'm thinking there's "no issue" right now
<samueldr>
until monokrome can say how they asserted they are not booted in a 64 bit linux
<samueldr>
because yeah, bios boot wil install a grub that will have "i386-pc" stuff, but that doesn't preclude it being able to boot a 64 bit os
* colemickens
nods
hakujin has joined #nixos
<colemickens>
ah yeah and I think I've read the scripts where it picks grub if it's non-uefi booted
<jackdk>
infinisil: oh, probably, but would it stay enough over the next several years?
<infinisil>
Oh great, you can filter for Draft pull requests, but you can't filter them out :/
<rycee>
I guess pie___ isn't here but I've tried to set up a channel #home-manager now. ldlework had registered with chanserv and was nice enough to transfer it to me 🙂
<drakonis>
ah nice
<samueldr>
rycee: want it to be logged?
<pie___>
rycee, hullo
<rycee>
samueldr: Certainly, that would be very nice :-)
<pie___>
rycee, side note, i kind of just head-canoned home-manager as my user space config generator idea dump, i hope thats ok :p
slack1256 has joined #nixos
<pie___>
i havent seen anything other then the home manager ?alternatives? that deals with stuff like this
<drakonis>
there's alternatives?
<mekeor>
i set up gitlab and gitlab-runner on nixos and i'm using the shell executor. after working around an issue with the PATH variable as described by mightybyte at https://github.com/NixOS/nixpkgs/issues/53985, i'm now getting a strange error from the job involving gpg-connect-agent and a refused connection: http://sprunge.us/1qQhBE – any ideas what's going on? (PS: "computer" is resolved to 127.0.0.1)
<aanderse>
i should really review PRs more often...
<infinisil>
s/should/could
<aanderse>
s/could/will?
<drakonis>
infinisil++
<{^_^}>
infinisil's karma got increased to 63
<infinisil>
I'm taking it very casually with reviewing, yes I could review lots of PRs every day, but it's not something I get paid for and nobody expects me to do that, also I don't have time for that all the time, and it's also not very interesting work
jluttine has joined #nixos
<gchristensen>
it is important work, but it is also important to not do so much you burn out. better to do what feels good
thc202 has quit [Ping timeout: 246 seconds]
<infinisil>
^^
jluttine has quit [Ping timeout: 252 seconds]
hlolli has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/fjfph
<{^_^}>
[nixpkgs] @MasseGuillaume closed pull request #56864 → kubernetes: set current-context and context name correctly → https://git.io/fhxi6
vidbina has joined #nixos
<{^_^}>
[nixpkgs] @illegalprime opened pull request #58104 → nix: replace build utils with host utils in config.nix → https://git.io/fjfjq
buffet- has joined #nixos
adetokunbo has quit [Quit: This computer has gone to sleep]
buffet has quit [Ping timeout: 272 seconds]
o1lo01ol1o has joined #nixos
vincentpants has joined #nixos
adetokunbo has joined #nixos
o1lo01ol1o has quit [Ping timeout: 245 seconds]
MinceR_ has joined #nixos
MinceR has quit [Ping timeout: 246 seconds]
mekeor has quit [Ping timeout: 252 seconds]
hakujin has joined #nixos
ym555 has quit [Ping timeout: 244 seconds]
vincentpants has quit [Ping timeout: 268 seconds]
hio has joined #nixos
orivej has joined #nixos
vk3wtf has quit [Ping timeout: 252 seconds]
vincentpants has joined #nixos
justanotheruser has quit [Ping timeout: 245 seconds]
<jasom>
so I'm running into https://github.com/NixOS/nixpkgs/issues/47552 is there any better way to fix than a "nix-env -u --eq" ? This is a laptop so rebuilding my KDE apps is going to take some time...
<monokrome>
about to try networkmanager but I hate networkmanager :(
pbb has joined #nixos
<colemickens>
Didn't like the cli client for iwd, couldn't get libvirt networking working with networkd, soooo NM for me.
<colemickens>
aside: I think NM can use IWD as a backend now though, but I guess wpa_supplicant has never hurt me.
<monokrome>
IWD?
<monokrome>
I thought everything would work fine since this is a laptop meant for Linux w/ Atheros wifi
<monokrome>
but I guess not :'(
<monokrome>
as a separate note
<monokrome>
Is there a way to force it to reinstall the system as x86_64 instead of i386? O_o
<colemickens>
it sounds like a config issue, not that that's what you want to hear.
<colemickens>
Why do you think it installed i386?
<colemickens>
I'm trying to see how wireless is meant to work with services.wireless.enable + wpa by hand.
<colemickens>
It's possible that wpa_supplicant doesn't even read from the file you wrote.
<colemickens>
I think NixOS expects you to use `networking.wireless.networks.<name>.auth` etc, to configure that.
<monokrome>
that's not what all the docs I can find say
<monokrome>
and this used to work on my old laptop
<monokrome>
but I'd be ok w trying
<colemickens>
weird, the wiki page is pretty drab. I could well be wrong, again NM user :S
vidbina has quit [Quit: vidbina]
<colemickens>
oh yeah, the manual straight up disagrees with me, sorry.
<monokrome>
well, I would have assumed the same as you
<colemickens>
:( I wish I had a better suggestion for you.
vincentpants has quit [Ping timeout: 244 seconds]
<colemickens>
I guess the only thing would be try restarting? every once in a while I do wind up having to restart when I change enough networking type things.
mekeor has joined #nixos
<colemickens>
I think that's a known reason why there's some desire to move to networkd - more declarative and less state changes to manage?
<monokrome>
:(
<yl[m]>
is it ok to merge master into staging? There are currently some conflicts, not sure why. see #58105 for an example
<yl[m]>
monokrome: turned out to be a patch that was changed both sides. I fixed the conflict and I'm testing the package using this patch now before I push it out
mekeor has quit [Quit: digital revolution = cryptoanarchism + cybercommunism]
orivej has quit [Ping timeout: 272 seconds]
drakonis has joined #nixos
drakonis_ has quit [Ping timeout: 252 seconds]
<{^_^}>
[nixpkgs] @kalbasit opened pull request #58106 → Merge master into staging → https://git.io/fjfjQ
adetokunbo has quit [Quit: This computer has gone to sleep]
adetokunbo has joined #nixos
pie__ has joined #nixos
adetokunbo has quit [Client Quit]
adetokunbo has joined #nixos
<CMCDragonkai>
I'm trying to use the nvidia runtime for automatic mounting of GPU devices to the container. I'm using a container build via `dockerTools.buildImage`.
adetokunbo has quit [Client Quit]
adetokunbo has joined #nixos
<CMCDragonkai>
However I find that the containers doesn't find any GPU devices when using nvidia runtime.
<CMCDragonkai>
Except when I run the official containers from NVIDIA.
<CMCDragonkai>
Is there something special that needs to be done to the images that we build via `dockerTools.buildImage` so that it's compatible with NVIDIA docker runtime?
chreekat has joined #nixos
adetokunbo has quit [Client Quit]
adetokunbo has joined #nixos
adetokunbo has quit [Client Quit]
adetokunbo has joined #nixos
adetokunbo has quit [Client Quit]
vincentpants has quit [Ping timeout: 245 seconds]
adetokunbo has joined #nixos
pie___ has quit [Ping timeout: 246 seconds]
adetokunbo has quit [Client Quit]
adetokunbo has joined #nixos
teozkr has quit [Ping timeout: 240 seconds]
mpickering has quit [Ping timeout: 240 seconds]
pingveno has quit [Ping timeout: 240 seconds]
chrisaw has quit [Ping timeout: 240 seconds]
feepo has quit [Ping timeout: 250 seconds]
chrisaw has joined #nixos
teozkr has joined #nixos
mpickering has joined #nixos
sppky has quit [Ping timeout: 250 seconds]
zimbatm has quit [Ping timeout: 250 seconds]
typetetris has quit [Ping timeout: 250 seconds]
prooftechnique has quit [Ping timeout: 240 seconds]
sorear has quit [Ping timeout: 240 seconds]
gleber_ has quit [Ping timeout: 250 seconds]
pointfree has quit [Ping timeout: 250 seconds]
jw358 has quit [Ping timeout: 250 seconds]
mudri has quit [Ping timeout: 250 seconds]
parseval has quit [Ping timeout: 250 seconds]
georgew has quit [Ping timeout: 240 seconds]
wildsebastian has quit [Ping timeout: 240 seconds]
pasukon has quit [Ping timeout: 268 seconds]
vdemeester has quit [Ping timeout: 268 seconds]
sppky has joined #nixos
dunk_ has joined #nixos
NemesisD_ has joined #nixos
digitalgrease has quit [Ping timeout: 250 seconds]
joedevivo has quit [Ping timeout: 240 seconds]
adelbertc has quit [Ping timeout: 268 seconds]
duoi has quit [Ping timeout: 268 seconds]
grenade has quit [Ping timeout: 250 seconds]
carter has quit [Ping timeout: 250 seconds]
nimblepoultry has quit [Ping timeout: 250 seconds]
<MasseR>
I would like to build my home-manager *configuration* (for CI purposes), but the configuration is basically a function that takes `{lib, pkgs, config, ...}`, how can I write a nix-build that has access to those values (aside from pkgs which is just <nixpkgs>.pkgs)
<dminuoso>
How can I discover how to feed a pkgs set into a callCabal2nix generated derivation?
<dminuoso>
Oh wait it just hit me, nevermind.
<MasseR>
dminuoso: You don't need to :)
<{^_^}>
[nixpkgs] @cdepillabout opened pull request #58113 → Fix cudatoolkit building on non-sandboxed systems → https://git.io/fjJvG
<dminuoso>
MasseR: Yeah, it just came to me that Im using callCabal2nix as `pkgs.haskell.packages.${compiler}.callCabal2nix "compass-radius" ./. { }`, so modifying the pkgs set happens beforehand. :)
<{^_^}>
[nixpkgs] @Mic92 pushed 4 commits to release-18.09: https://git.io/fjJv0
<dminuoso>
I have a nix derivation that is supposed to pull a package from github, but it seems to not be using either the package from github or a wrong revision. Im left with no clues how to debug this.
<linarcx>
Hi guys. i'm creating a theme package for nixos. In `installPahse`, i have to do some sort fo delete unnecessary files from output directory. i'm using for in loop in bash like: ` for i in "${conflict_files[@]}"; ...`, but when i want to build the expression locally i'v get an error. undefined `@`. what should i do?
<johanot>
dminuoso: can you paste what you have somewhere?
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/fjJJh
<dminuoso>
johanot: For some reason that is beyond me, it's not using the specified rev from github (my patch fixes a particular failure. Building this expression however provokes that failure, indicating that its not using my hexquote version)
<linarcx>
clever: No. it's not working too.
<clever>
linarcx: can you pastebin the entire nix file?
<johanot>
dminuoso: I can confirm that building your derivation does _not_ trigger a build of ./hexquote.nix :) but when I build ./hexquote.nix alone it does in fact pull the revision you ask for. Looks like some issue with your haskellPackages override, but I'm don't know enough about that, sry.
<dminuoso>
johanot: Thank you for the confirmation. Mind my asking, how did you confirm that it does not trigger a build of ./hexquote.nix?
<johanot>
from looking at the output of "nix-build ./default.nix" and from searching my nix-store for hexquote. And further confirmed by "haskellPackages.callPackage ./hexquote.nix" which triggered a full build.
<linarcx>
Anyone sucessfully configure POWERLEVEL9k in nixos? I have .zpreztorc in home directory. but seems it can't be load properly.
random has joined #nixos
random has quit [Client Quit]
<{^_^}>
[nixops] @AmineChikhaoui merged pull request #1117 → ec2 backend: Remove support for AMI resizing <nixos 15.09 → https://git.io/fjfdE
<{^_^}>
[nixops] @AmineChikhaoui pushed 2 commits to master: https://git.io/fjJTD
<barrel>
Hey guys, I have an issue with the mesa drivers on my ThinkPad T430 Which uses HD 4000, some software such as compton can't find i915, i965 and swrast as an example, I'm currently not using the compton service in my configuration.nix but via my own launch script
<barrel>
I hate coming in to IRC Just to ask questions but I've not been able to fix it for quite a long time :P
<rycee>
Yep, that looks fine. The home-manager script does something similar. 😃
mariel__ has joined #nixos
<MasseR>
Yeah that's where I scraped it
barrel has quit [Ping timeout: 252 seconds]
gwen has joined #nixos
<gwen>
Hi, does any one of y'all know how to use ZFS on NixOS? I created an encrypted dataset and could mount it once, but not anymore and when I rebuild the system I'm dropped into an emergency shell and can't do anything.
i1nfusion has quit [Remote host closed the connection]
i1nfusion has joined #nixos
linarcx has quit [Quit: WeeChat 2.4]
linarcx has joined #nixos
linarcx has quit [Client Quit]
linarcx has joined #nixos
<dminuoso>
Nixos question, how/when are updates released? Does nix have the same concept of a stack resolver that is a list of curated nixpkgs (and identified by system.stateVersion)?
<__monty__>
dminuoso: ^ For info on how the channels move.
<jasongrossman>
dminuoso: system.stateVersion is a misnomer. Just ignore it.
<__monty__>
And no, nixpkgs is one static package set.
<dminuoso>
__monty__: Just like a stack resolver would give you a static package set though?
o1lo01ol1o has joined #nixos
linarcx has quit [Client Quit]
hakujin has joined #nixos
<__monty__>
haskellPackages specifically is a mix of stack (LTS I think), head.hackage and anything else from hackage that's not covered by either of those.
linarcx has joined #nixos
<__monty__>
Oh, I thought you were talking about stack's cabal feature.
<dminuoso>
Heh no. :)
<dminuoso>
Well yes.
vincentpants has joined #nixos
<dminuoso>
Trying to understand whether nixpkgs is similar to a stack resolver.
<__monty__>
In some ways yes, in other ways no.
arjen-jonathan has joined #nixos
<dminuoso>
__monty__: What I dont understand is, your website indicates that channels are not fixed. Where does determinism come from, when the same set of attributes could lead to different builds?
<dminuoso>
I mean if channels receive updates continuously, how can I do pining?
pita_ has quit [Quit: EliteBNC - http://elitebnc.org (Auto-Removal: idle account/not being used)]
<linarcx>
betaboon: Thank you.
amir has joined #nixos
endformationage has joined #nixos
Fare has joined #nixos
vincentpants has joined #nixos
<pbb>
Hi, I am running a git clone inside an stdenv.mkDerivation builder and I'm getting "SSL certificate problem: unable to get local issuer certificate", running the same command in a shell works. What is the package for the ca certificates called, or what else could be the problem here?
rycwo has quit [Ping timeout: 272 seconds]
<pbb>
(yes I know it's not nice to do it this way, but in this case it's no easily possible to clone all dependencies via nix, so I'm using a fixed-output derivation)
<betaboon>
pbb: why are you trying to run git clone inside a mkDerivation ? oO whats wrong with `src = fetchgit { ... };` ?
<pbb>
I'm actually running ruby bundler, which runs git
knupfer has joined #nixos
<pbb>
I guess git can not find the ca certificates in the build environment, how would I include them?
zupo has joined #nixos
<betaboon>
dont know what the right approach would be, guess someone else has to step in to help ;)
<pbb>
okay, I also tried to use bundix but that failed on multiple levels and since bundler worked fine in a shell I thought I could use that with a fixed-output derivation as a first step
crmlt has joined #nixos
<pbb>
(I'm actually trying to put a mastodon instance on my server)
gwen has quit [Ping timeout: 245 seconds]
reinhardt has quit [Ping timeout: 252 seconds]
<{^_^}>
[nixpkgs] @mwilsoninsight opened pull request #58125 → showoff: init at 0.20.1 → https://git.io/fjJLh
reinhardt has joined #nixos
gwen has joined #nixos
knupfer has quit [Ping timeout: 246 seconds]
<linarcx>
Gyus, i'm creating a package called: bicon. it's written in c. But when i want to build it locally i get an error: `make: *** No rule to make target 'install'. Stop.`
<slabity>
Is anyone else unable to install NixOps from nixos-unstable? Getting an error: `error: aiohttp-3.5.4 not supported for interpreter python2.7`
pie__ has quit [Read error: Connection reset by peer]
pie__ has joined #nixos
dnlkrgr has quit [Ping timeout: 244 seconds]
linarcx has quit [Quit: WeeChat 2.4]
linarcx has joined #nixos
pie_ has quit [Ping timeout: 250 seconds]
<linarcx>
simpson: Yessss:) `patchShebans` works really fine :). Thank you again.
jtojnar has joined #nixos
b1000101 has joined #nixos
Glider_IRC has joined #nixos
timor has joined #nixos
<timor>
Is there a way to remove the no-network-access restriction during checkPhase, when calling `nix-build` interactively? I tried `nix-build ... -o sandbox=false`, but that did not seem to have any effect.
slack1256 has joined #nixos
Fare has quit [Ping timeout: 246 seconds]
Makaveli7 has joined #nixos
dnlkrgr has joined #nixos
inquisitiv3 has joined #nixos
<{^_^}>
[hydra] @mray-infor opened pull request #646 → Microsoft Teams notification plugin → https://git.io/fjJtj
rauno has quit [Remote host closed the connection]
crmlt has quit [Ping timeout: 250 seconds]
pie___ has joined #nixos
<inquisitiv3>
I'm having trouble cloning a private git repo hosted @ GitLab using the SSH URL. When I try to clone the repo I get an error message (https://paste.debian.net/1074226/). I'm to clone the repo using HTTPS after authenticating myself. Verified that the SSH key was added correctly by executing the command `ssh -T git@gitlab.com` described in GitLab's documentation. The problem don't occur when I use Git from Fedora's repos. Nix (just
<inquisitiv3>
the package manager) is installed upon a Fedora 29 host.
<inquisitiv3>
Any idea what the problem could be?
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<{^_^}>
[nixpkgs] @LinArcX opened pull request #58128 → bicon: init at 38725c0 → https://git.io/fjJqq
<inquisitiv3>
vaibhavsagar: Seems there isn't any `Match`-blocks in /etc/ssh/sshd_config for me. Suse 11 SP4 seems to have been released in March last year. Fedora 29 was frozen at the end of August. Feels like the issue should've been fixed by then.
vincentpants has joined #nixos
<vaibhavsagar>
no idea
<nore>
Hi there, does anyone know how to make opam2 work with nixos? When I try to use it to install merlin, it crashes because it doesn't find "which"
o1lo01ol1o has quit [Ping timeout: 246 seconds]
Fare has joined #nixos
mariel__ has quit [Read error: Connection reset by peer]
ThatDocsLady_ has joined #nixos
o1lo01ol1o has joined #nixos
crmlt has joined #nixos
o1lo01ol1o has quit [Ping timeout: 246 seconds]
Fare has quit [Ping timeout: 250 seconds]
o1lo01ol1o has joined #nixos
<pie___>
im trying to use appimage-run, is that supposed to work? im getting a complaint but adding the library to the env doesnt work
<pie___>
* /home/nixos/.cache/appimage-run/5a5fe9ef9c555d086f3f6ecc2ed8649c4bd752be8232026f65ffa35c0338fe62//squashfs-root/usr/bin/icestudio: error while loading shared libraries: libnotify.so.4: cannot open shared object file: No such file or directory
o1lo01ol1o has quit [Ping timeout: 245 seconds]
Neo-- has joined #nixos
<b1000101>
Hi, is there a way to remove jobsets from Hydra project? I can delete the project, but not the jobset. One more thing - I set the evaluation to keep to 1 but it seems my disk is absolutely full from previous evals. How to garbage collect them on a regular basis? (2018-08-07 (using nix-2.1.3).)
<aminechikhaoui>
b1000101 deletes aren't really reliable, it would often fail due to constraints errors
<{^_^}>
#58037 (by Twey, 1 day ago, open): NeoComp: init at git-v0.6-17-g271e784-2019-03-12
<aminechikhaoui>
regarding garbage collection, as far as I know there is a hydra-update-gc-roots script which will determine what are the gc roots needed
<aminechikhaoui>
and then running nix-collect-garbage should remove anything extra
<b1000101>
aminechikhaoui: yeap, I already tried it but it freed 0 MB :)
<b1000101>
aminechikhaoui: aaah, right, forgot about that...
crmlt has quit [Ping timeout: 246 seconds]
<b1000101>
aminechikhaoui: 7GB freed, thanks!
hyper_ch2 has quit [Ping timeout: 256 seconds]
<b1000101>
aminechikhaoui: so if I want to delete a jobset should I delete the whole project then? If you say it often fails
<aminechikhaoui>
sure, it's better to have hydra-update-gc-roots and a nix-gc systemd services
<aminechikhaoui>
that run in timers
<b1000101>
aminechikhaoui: sure, I'll look into that
<aminechikhaoui>
b1000101 in our internal hydra we disable jobsets and hide them by default
<Twey>
Oh, I guess you're just supposed to expect those packages
<aminechikhaoui>
that way we usually don't need to delete
ym555 has quit [Ping timeout: 245 seconds]
arjen-jonathan has quit [Ping timeout: 272 seconds]
crmlt has joined #nixos
<aminechikhaoui>
although some people are requesting that but not sure if there is an easy solution with the current db schema
<b1000101>
aminechikhaoui: that's what I did but I want to get rid of some jobsets as they're not really needed and they're taking up extra space. Say I have 3 jobsets, one running against unstable, one against 18.09 and one 18.09-small
<b1000101>
aminechikhaoui: so I guess creating separate projects for that and not having different releases together would be the best solution in this case?
<b1000101>
aminechikhaoui: they all have the same channel, right? so I should not mix them together in the first place
<b1000101>
aminechikhaoui: but then, I was trying to mimick hydra @ nixos, where there's different releases under the same project
<pie___>
nevermind looks like icestorm exists in nixpkgs idk why i couldnt get it
mekeor has joined #nixos
<pie___>
ok it works off fresh nixpkgs
alex`` has joined #nixos
vincentpants has quit [Ping timeout: 255 seconds]
<pie___>
err nevermind
crmlt has quit [Ping timeout: 250 seconds]
Myhlamaeus has joined #nixos
ixxie has joined #nixos
Tucky has quit [Quit: WeeChat 2.2]
<b1000101>
hah, I can't even delete the project since there are some test releases - any idea how to delete those? :)
<b1000101>
ah got it
<linarcx>
What is it: `╰ nix-build -A saaghar `
<linarcx>
error: anonymous function at /home/linarcx/nixpkgs/pkgs/applications/misc/saaghar/default.nix:1:1 called without required argument 'qmake', at /home/linarcx/nixpkgs/lib/customisation.nix:69:12
<manveru>
pie___: yeah... that's not the proper way to build it, and yarn2nix apps aren't really well received in nixpkgs :)
<slabity>
Is nixops breaking for anyone else? I'd expect at least that package to be well tested
<slabity>
error: aiohttp-3.5.4 not supported for interpreter python2.7
rycwo has joined #nixos
<slabity>
nixos-unstable channel
i1nfusion has quit [Remote host closed the connection]
i1nfusion has joined #nixos
<pie___>
manveru, ok i cant actually get it to build though now
<pie___>
manveru, im sitting at a tutorial and it might be neat if i could get this to build
<pie___>
manveru, error Couldn't find a package.json file in "/home/nixos/stuff/clash/icebuild"
<pie___>
when i try to nix-shell it
<manveru>
pie___: this is meant for nix-shell in a clone of icestudio
rycwo has quit [Ping timeout: 250 seconds]
<manveru>
sorry, don't have a lot of time atm
<pie___>
manveru, an di dont have a lot of brain :)
<pie___>
oh ok
<pie___>
so the problem is probably just that im not in a repo
Ariakenom has quit [Ping timeout: 268 seconds]
<pie___>
ty
vincentpants has joined #nixos
<pie___>
manveru, ok it worked
<pie___>
manveru, i tried nix-shell -p stack yosys arachne-pnr icestorm "import ../icestudio-shell.nix" -v but I get variable $src or $srcs should point to the source
<pie___>
if i just do nix-shell icestudio-shell.nix it works
orivej_ has joined #nixos
orivej has quit [Ping timeout: 268 seconds]
dashkal has quit [Ping timeout: 268 seconds]
<manveru>
yeah :)
<manveru>
there's no `src` for the nix-shell, so it doesn't know where the source is
jtojnar has quit [Quit: jtojnar]
<pie___>
manveru, sorry, wat do?
<pie___>
do i just pass a src
<manveru>
no
<manveru>
you did the right thing already
<pie___>
how do i make it not break
<manveru>
it's just not something you could nix-build
<pie___>
manveru, the problem is i need these other packages in my env
<pie___>
(sorry very split brian right now)
<pie___>
*brain
<aminechikhaoui>
b1000101 itt seems fine to me to have them in the same project, so if you want to garbage collect maybe you can set number of evaluations to keep to 0 and run the updates gc-roots + collect garbage
<NemesisD>
i'm trying to load a haskell project into a docker image. i'd like to be able to select a subset of the executables from the build to put into the docker image so i can have an image per executable. is this possible?
<aminechikhaoui>
that should take care of freeing up space
<aminechikhaoui>
then you can delete the jobset from postgres db but it won't be the same as the action from the ui as it won't try to delete the builds from the db table
knupfer has joined #nixos
dashkal has joined #nixos
<NemesisD>
whew lad. actually i just looked at the docker image and the executables are quite small. the real problem is that the closure drags in 5.1GB of data, including all the dependencies and a 1.4gb expression for GHC which should not be needed at runtime
o1lo01ol1o has joined #nixos
sigmundv has quit [Ping timeout: 252 seconds]
<b1000101>
aminechikhaoui: hmmm, something is rotten in the state of Denmark... I've ran the garbage collector and even deleted the project and re-created it but there's still of space being used by a 19.09 (unstable) jobset that I previously tried to fetch and build. Any idea why it might not get deleted by GC ? I'm running the hydra on a 18.09 machine so the dependencies should not be needed anywhere in the system
<pie___>
oh i can just nest nix shells probably
<pie___>
doh
ggpeti has joined #nixos
<pie___>
...or not
alex`` has quit [Ping timeout: 246 seconds]
Ariakenom has joined #nixos
<ggpeti>
hello
<ggpeti>
does someone know how often the unstable channel merges nixpkgs/master?
<gchristensen>
hello, it is automatic
<aminechikhaoui>
b1000101 if you have a store path that you want to check why it's still not GCed you could run `nix-store -q --referrers <path>`
<gchristensen>
it isn't actually a merge, just an advancement. ggpeti do you mean nixos-unstable or nixpkgs-unstable?
alex`` has joined #nixos
<manveru>
pie___: just add any other packages to buildInputs
<pie___>
hmm, its probably something wrong with the app
<ggpeti>
nice resources, thx
o1lo01ol1o has quit [Ping timeout: 250 seconds]
<gchristensen>
yep!
linarcx_ has joined #nixos
crmlt has joined #nixos
technoidX has quit [Read error: Connection reset by peer]
Fare has joined #nixos
chreekat has quit [Quit: quitting]
ubert has quit [Quit: Leaving]
<monokrome>
Is there a way to manually override specific keycode mappings in the nix config?
tilpner has joined #nixos
crmlt has quit [Ping timeout: 250 seconds]
alex`` has quit [Ping timeout: 245 seconds]
dnlkrgr has quit [Ping timeout: 246 seconds]
<sphalerite>
Any tinc users who have tinc consistently working? For me, nixos never seems to set the network addresses consistently, which often leads to a device no longer being accessible via the VPN after a reboot.
<infinisil>
noonien: import (fetchwhatever { ... }) works
<infinisil>
,ifd
<{^_^}>
import-from-derivation (IFD) is when you evaluate nix from a derivation result, for example `import (pkgs.writeText "n" "1 + 1")` will evaluate to 2. This is sometimes problematic because it requires evaluating some, building some, and then evaluating the build result.
Fare has joined #nixos
<sphalerite>
gchristensen: apparently that one can't deal with the size of my graph :p
<sphalerite>
clever: I'll try that one, thanks
<gchristensen>
it is running on a teeny tiny heroku thing :)
<ggpeti>
git-minimal-2.19.2 on x86_64-darwin build is broken on nixpkgs-unstable since 3 days :(
<srhb>
ggpeti: Is there an issue on this?
<sphalerite>
clever: aww it doesn't want to load from non-regular file
Olgierd has joined #nixos
Olgierd has joined #nixos
Olgierd has quit [Changing host]
Jetien has quit [Ping timeout: 255 seconds]
<ggpeti>
srhb I don't see one
<noonien>
infinisil: hmm, what if i don't want to import qemu.nix, or /overlays/qemu, do i just do foo = fetchsomething {}; import $foo/qemu.nix?
<noonien>
s/don't //
<srhb>
ggpeti: Maybe you can open one? :)
<ggpeti>
srhb sure
<betaboon>
slabity: ah ok now i see :)
<infinisil>
noonien: import (fetchsomething {} + "/qemu.nix") should work
<betaboon>
slabity: weird tho that i can build nixops from nixpkgs-master no problem oO
<noonien>
infinisil: awesome, thanks!
<slabity>
betaboon: Looks like nixos-unstable hasn't been updated since yesterday. Maybe there was a change?
<ggpeti>
uhh missed the output ^_^ just a few more minutes
goibhniu has quit [Ping timeout: 268 seconds]
<srhb>
ggpeti: And are you on the tip of master right now?
<ggpeti>
yes. any way I can check the ongoing build log?
<srhb>
ggpeti: Not sure whether nix log works until it's done.
<srhb>
No rush though.
MiloIgnis[m] has joined #nixos
<betaboon>
infinisil: o/ your name popped up in a number of my PRs *wink wink* XD
<infinisil>
betaboon: Like as a reviewer?
<betaboon>
infinisil: si si
<infinisil>
Ah yeah, I get this for all PRs that touch NixOS modules, so I often have quite a list to go through when I'm not active for a bit. Got any specific PR's you want to have get looked at?
<NemesisD>
do any haskell nix users how i may avoid capturing .a and .o files in a haskell package expression's closure? they're enormous for large projects and i need to do better than a 5gb docker image for 100mb of executables
<NemesisD>
not sure if this would be a more appropriate question for #haskell
<srhb>
NemesisD: Seems fine here.
<srhb>
NemesisD: Sounds like you want haskell.lib.justStaticExecutables ?
<buffet>
weird question, but what exactly does nix do when i install another package manager using it? do they just install packages into the fs root like they normally would?
<NemesisD>
srhb: i've heard of that one. i'm not sure how to apply it. i've added my package to the haskellPackages set but it doesn't seem to work with justStaticExecutables hpkgs.my-package nor justStaticExecutables (hpkgs.callPackage hpkgs.my-package), both give me `cannot coerce a function to a string`
<infinisil>
NemesisD: You mean like when doing `src = ./.`?
<srhb>
buffet: Unless it were heavily patched, I would assume so. They will probably not work correctly, frankly.
<srhb>
NemesisD: You apply it to your expression.
knupfer has quit [Ping timeout: 268 seconds]
<buffet>
srhb: why are they packaged btw?
vincentpants has joined #nixos
<{^_^}>
[nixpkgs] @flokli merged pull request #57938 → network-manager: move para about service rename to 19.09 changelog → https://git.io/fjfGe
<srhb>
buffet: Some of them also get used internally in our expressions, where possible.
agander has joined #nixos
timor has quit [Ping timeout: 250 seconds]
noogie has quit [Ping timeout: 272 seconds]
vincentpants has quit [Ping timeout: 250 seconds]
bemeurer has joined #nixos
slack1256 has joined #nixos
glasserc has quit [Ping timeout: 252 seconds]
glasserc has joined #nixos
agander_ has joined #nixos
<slack1256>
Does somebody use texlive.combined.scheme-full ? what made you switch from scheme-medium to scheme-full?
<NemesisD>
infinisil: srhb i think i was holding it wrong. holy smokes, 768MB tar/gzipped image down to 36M!
<srhb>
NemesisD: Yeah, innit great ;-)
vincentpants has joined #nixos
<srhb>
slack1256: I do, in one expression
<srhb>
slack1256: I think I went "ok, chasing deps down takes too much of my time and space is cheap (in this case)"
<{^_^}>
[nixpkgs] @xeji pushed commit from @andir to release-19.03 « subversion_1_10: 1.10.3 -> 1.10.4 »: https://git.io/fjJOF
<srhb>
Nothing truly inspiring...
agander has quit [Ping timeout: 255 seconds]
<slack1256>
Any package you use that is on full but not on medium (I wanna know if maybe I will miss something obvious that is on full but not on medium).
glasserc has quit [Read error: Connection reset by peer]
<srhb>
slack1256: Hmm, honestly I can't recall.
m0rphism has quit [Quit: WeeChat 2.2]
<NemesisD>
srhb: it is! so my plan was to build an image per-component after this point: a webserver image that just has the web serve exe + copies in my compiled frontend assets, a worker image with just the worker exe and none of that frontend stuff. what's the path to that look like? would i write a vanilla derivation that takes that justStaticExecutables output as an input + some files and sets up the desired directory structure?
<srhb>
It's just this one project, I could check it later.
<aanderse-work>
infinisil: you published it!
<aanderse-work>
i have no time to read it now, but hopefully tonight i can :D
<aanderse-work>
thanks for pinging me on it
<srhb>
NemesisD: The components are all outputs of the same cabal project in this case?
agander__ has joined #nixos
<NemesisD>
srhb: yep
glasserc has joined #nixos
<infinisil>
aanderse-work: Ah yeah I pinged you when I wrote it, but you didn't see it then :P
<srhb>
NemesisD: Yeah, you could get by with runCommand, but at that size I'd honestly be wary of the cost savings compared to the manual dependency chasing that might go wrong.
<{^_^}>
[nixpkgs] @bendlas merged pull request #58113 → Fix cudatoolkit building on non-sandboxed systems → https://git.io/fjJvG
<{^_^}>
[nixpkgs] @bendlas pushed commit from @cdepillabout to master « Fix cudatoolkit building on non-sandboxed systems »: https://git.io/fjJOp
<aanderse-work>
eh? you did? sorry :\
<aanderse-work>
but thanks!
<infinisil>
aanderse-work: Oh wait did you see my ping in the PR I'm just reviewing?
<aanderse-work>
yeah its the ping in the PR i see
<srhb>
NemesisD: I'd probably prefer splitting it at the cabal package level, if possible.
<aanderse-work>
where was the other ping?
<srhb>
NemesisD: Or not at all. It's pretty small already :P
<infinisil>
aanderse-work: Oh wow, I haven't submitted that review, I'm surprised you see it already
<NemesisD>
srhb: hmm, so i guess you're saying is the alternative is 1 docker image with all the exes? the size is small enough that i could entertain that, but i still need to figure out how to inject my frontend assets into that image, ones that are not build via a nix expression right now (but rather just built with *tools* provisioned by nix)
<srhb>
NemesisD: Oh, okay, if you're already doing that, it's unsafe already, so manually copying things in place sounds reasonable.
<{^_^}>
[nixpkgs] @bendlas pushed 2 commits to release-19.03: https://git.io/fjJ3J
benpa[m] has left #nixos ["User left"]
<NemesisD>
srhb: so on the docker image i'd envision a `/myapp` dir or something with my static assets loaded in it. i use shake to call nix-build to build the docker image so i could guarantee the frontend assets are rebuilt before copying, i just don't know how i'd do it in nix-land
<srhb>
NemesisD: Can't you just invert the build and have nix outermost? Does frontend building use the network or something?
b1000101 has quit [Quit: Page closed]
pointfourone has joined #nixos
pointfourone has quit [Remote host closed the connection]
crmlt_ has quit [Read error: Connection reset by peer]
ixxie has joined #nixos
m0rphism has joined #nixos
slack1256 has quit [Remote host closed the connection]
<srhb>
If not, I'm sure what you proposed is quite reasonable.
m0rphism has quit [Client Quit]
lord| has quit [Quit: WeeChat 2.4]
m0rphism has joined #nixos
<NemesisD>
srhb: maybe. its just that the integration with nix on the frontend side is really shallow. its building purescript code with spago (which nix provisions). i don't know what work is involved to get nix to drive that process. its like i want nix to compute the hash of my compiled output files but not necessarily know how to build them, just know that if they change, the expression for my container must be rebuilt
<srhb>
NemesisD: I don't know about spago, but you probably don't care about the output hash, just the inputs.
<NemesisD>
its much more apparent to me how to teach shake how to build my frontend than nix
<srhb>
NemesisD: (That is, if you can run it via nix)
<srhb>
ok :)
<NemesisD>
it may be possible to run via nix. nix provisions spago, my minifier, etc. i guess its mainly accounting for the inputs, because all nix would have to do is run `build.sh`
<srhb>
Right.
<NemesisD>
i guess i should actually try my hand at writing a dang nix derivation
<NemesisD>
i've so far only interacted with it through tools built on top of that like the haskell tooling
lord| has joined #nixos
<srhb>
Took a quick look. If this involves calling out to npm, it's probably not going to be a fun first project
<NemesisD>
it currently doesn't, but this project is a pilot for one that will :(
arjen-jonathan has joined #nixos
<srhb>
It's doable, with npm2nix or whatnot, but not exactly trivial usually :)
<NemesisD>
i think i'm looking for an escape hatch where i can say if any of these files change, run this script. i don't know how compatible that is with the nix build environment, like it seems like all the relative paths my build scripting depends on may not be available
<srhb>
NemesisD: That is the default behaviour of Nix really.
<srhb>
NemesisD: Consider in the simple case where you have something like this in your installPhase: cp -r ${./somedir} $out/
<srhb>
NemesisD: In this case, you can consider ${.somedir} as its own derivation, with a hash and everything.
arjen-jonathan has quit [Client Quit]
<srhb>
The same logic applies when ./somedir is an input to some more complex derivation as well.
<NemesisD>
srhb: ok. well i'll just give this a shot since it sounds like i have to do it no matter what. i'll put aside the idea of trying to split images per exe since that seems to not get me much (and may be an overall net waste unless i get image layering working)
<NemesisD>
before using justStaticExecutables, building a layer image spawned some python script from nix which took up 22GB of memory and crashed :)
<srhb>
infinisil: And you can often (during EU daytime hours) find the authors adamt and johanot in the channel.
<slabity>
wolfram154: It might not be correct to put it in systemPackages. There's a few options like `services.xserver.xkb*`
<wolfman154>
Do I use xkbcomp-1.4.2 or xorg.xkbcomp to install xkbcomp in my configuration.nix : environment.systemPackages
<srhb>
For some reason none of them ever get around to creating irc bouncers.. And it seems like an easy job for morph!
<dsx>
infinisil: can it upload arbitrary files to arbitrary locations?
<infinisil>
adamt++ johanot++
<{^_^}>
adamt's karma got increased to 2, johanot's karma got increased to 1
<wolfman154>
I want to do this right so I don’t crash nixos again
<infinisil>
dsx: Check out the readme, I haven't fully read through it, but there is support for uploading private files via scp
<betaboon>
srhb: any opinions on max-line-length ?
<dsx>
infinisil: ok, thanks!
<srhb>
betaboon: I like pretty things.
<infinisil>
wolfman154: xorg.xkbcomp
<betaboon>
srhb: best answer :)
<dsx>
>Files can be uploaded without ever ending up in the nix store, by specifying each file as a secret. This will use scp for copying a local file to the remote host.
<infinisil>
wolfman154: Not sure how that could crash NixOS though
<dsx>
It can.
<dsx>
infinisil++
<{^_^}>
infinisil's karma got increased to 64
<slabity>
wolfram154: It sounds like you're crashing GNOME or the X11 service
<wolfman154>
infinisil: I think I need to install other dependencies that xkbcomp depends on, I don’t know why it does not do it automatically?
<betaboon>
infinisil: have you had a chance to test out morph ?
<infinisil>
wolfman154: Well do you have a crash log or something?
<infinisil>
"store your secrets in password store" :O
<infinisil>
Also interesting, might have to try both out at some point
<wolfman154>
slabity: gnome crashes and when I reboot it stops at services sign in gdm throws me to a terminal where I get stuck, I had to switch back to an older configuration.nix
<slabity>
wolfman154: I believe gdm logs get stored in /var/log/ if you can find it there
endformationage has joined #nixos
<slabity>
If not, the X.0.log should at least be there
<werner291>
nixos-in-place allows you to replace an existing Linux installation with NixOS. Yes, I'm using that because my service provider didn't an option to use a custom iso for the VPS>
<werner291>
nixos-in-place allows you to replace an existing Linux installation with NixOS. Yes, I'm using that because my service provider didn't an option to use a custom iso for the VPS.
<tilpner>
To upgrade to the latest release (18.09), you could try nix-channel --add https://nixos.org/channels/nixos-18.09 nixos && nix-channel --update
<tilpner>
(Try, because I don't use nix-channel and should probably stop helping whenever it's involved)
<srhb>
Sounds reasonable, though I think it might have to be --remove'd first
<{^_^}>
#46387 (by Infinisil, 27 weeks ago, open): Impossible to easily upgrade NixOS from <=17.09 to 18.09
<tilpner>
srhb: No, it overwrites them
<infinisil>
I'm not sure if this issue is resolved
<tilpner>
Oh no
<srhb>
tilpner: Neat :)
<infinisil>
"As far as I recall, it is also not possible to upgrade to 18.03 from 17.03 without going though 17.09 first, since the Nix 1.11 in 17.03 is too old to evaluate 18.03, so this isn't a completely unusual situation."
<infinisil>
Well this is shitty haha
<srhb>
I think... we fixed that... Somehow
<tilpner>
Couldn't you use Nix from 18.09 to skip versions?
<srhb>
But I don't really recall tbh :P
<srhb>
tilpner: Yes
grizwako has quit [Ping timeout: 268 seconds]
<werner291>
*Is doing the channel stuff.*
<srhb>
Last time I did it I hacked a rebuild script to point to the newer nix, I think..
<srhb>
Which is not a great end user experience, but alas... :P
zupo has joined #nixos
<werner291>
error: syntax error, unexpected $undefined, expecting IND_STR or DOLLAR_CURLY or IND_STRING_CLOSE, at /nix/store/5cyzdxqwax7hvh93b8a47jbsv6kf2nim-nixos-18.09.2375.680f9d7ea90/nixos/pkgs/build-support/fetchpatch/default.nix:42:40 (use ‘--show-trace’ to show detailed location information) building Nix...
<tilpner>
werner291: You can use nix-shell -I nixpkgs=https://nixos.org/channels/nixos-18.09 -p nix to obtain a Nix capable of evaluating newer nixpkgs
<srhb>
werner291: Yep, that's the issue we were just talking about
<werner291>
Ah
hoijui has quit [Remote host closed the connection]
<tilpner>
Uhh, maybe not
<srhb>
Alternatively, go to 17.09 first
<srhb>
That should be compatible with your nix
<srhb>
It's... Silly, sorry about that, but it should work
<werner291>
Heh. Will do, thanks for the suggestion for 17.09. Not sure what the rest means, but it's still going, it seems.
<srhb>
werner291: Should be out of syntax change land then :)
<LnL>
nix-store -r /nix/store/mzrh3y44sqrxrbb0xba5li9y838ndyxs-nix-2.1.3 followed by nix.package = /nix/store/mzrh3y44sqrxrbb0xba5li9y838ndyxs-nix-2.1.3; before the upgrade probably also works
<werner291>
I'm on 17.09 now... I think?
<srhb>
werner291: Cool, proceed to 18.09 :-P
<srhb>
werner291: (you can run nixos-version if you want certainty)
<werner291>
I'm gonna let it do its' thing for now.
<duairc>
Does anyone else get semi-regular "HTTP error 200 (curl error: Failure when receiving data from the peer); retrying in 339 ms" errors when using Nix over IPv6?
<duairc>
It only happens from my home connection and always works when I force the connection to use IPv4
<tilpner>
Wait, error 200?
<werner291>
Like... 200 ok?
<srhb>
duairc: Yes, it's seen before.
<LnL>
that's a curl error, not status code I presume
<duairc>
Haha, it's okay, I'm using irssi on my server anyway, I just got logged out of SSH is all
fragamus_ has joined #nixos
<{^_^}>
[nixpkgs] @marler8997 closed pull request #58148 → Added the google repo tool → https://git.io/fjJGM
orivej_ has quit [Ping timeout: 250 seconds]
agander has quit [Ping timeout: 245 seconds]
<duairc>
I don't know if this complicates things but my connection doesn't actually have real IPv4 at all, I use a tunnel from Hurricane Electric... maybe it's on that end? Or maybe I've set it up on my OpenWRT router wrong...
fragamus has quit [Ping timeout: 245 seconds]
fragamus_ is now known as fragamus
* gchristensen
has no idea how to help :|
<duairc>
Me neither! That's okay, at least I know roughly what the problem is
<duairc>
So I appreciate that
<LnL>
I find it a bit strange that I've not seen this at all yet, I'm also going over v6 by default
ThatDocsLady_ has quit [Ping timeout: 246 seconds]
adetokunbo has quit [Quit: This computer has gone to sleep]
<srhb>
LnL: Presumably your router behaves?
<andi->
I think I only see it very infrequently and I also have v6 everywhere these days..
<Dagger>
but I can't find a straight statement saying that it's that
<duairc>
Maybe I'll bug my ISP about giving me (back) real IPv6... I used to have real IPv6 but only fake NAT IPv4, I called them to get a real IPv4 and they said no problem, but whatever update they pushed to my modem took away IPv6!
<Dagger>
"At Fastly, this hashing is performed by an Ethernet switch ASIC, and to avoid breakage, the IPv6 hashing function must not include the flow label." okay, but does your Ethernet switch ASIC include the flow label or not? why doesn't this sentence actually say so, one way or the other?
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
<duairc>
Not that it matters but just in case anyone is ever reading the logs of this or whatever, I meant to say above that my connection doesn't have real *IPv6*
<Dagger>
I'm also behind HE and haven't seen this problem, but I'm not really a heavy Nix user
<duairc>
I mean it happens basically every time for me, so I think you'd have noticed it by now
<Dagger>
and my only NixOS container can't update because it can't fetch the channel list, since it only has v6 and the main site only has v4
rcshm has quit []
<duairc>
Haha
<tilpner>
!!!
<tilpner>
Have you opened an issue for that, Dagger?
<srhb>
Something went wrong then, because I am not in Hydra :P
<Cale>
Does anyone have handy an equivalent of something like nix-shell -p 'haskellPackages.ghcWithPackages (pkgs: with pkgs; [ hp2pretty text aeson cabal-install template-haskell constraints lens mtl ])' but which lets me specify which version of GHC it is that I want?
<srhb>
I'll figure it out..
<gchristensen>
(literally, the inHydra option)
<gchristensen>
srhb: do you see this output? Failed to evaluate ${attrs.name or "«name-missing»"}: «${reason}»: ${errormsg}
<srhb>
Probably some part of the haskell framework misusing settings
o1lo01ol1o has joined #nixos
<srhb>
gchristensen: error: Failed to evaluate hfsevents-0.1.6: «unsupported»: is not supported on ‘x86_64-apple-darwin’
<inquisitiv3>
Is it unusual for desktop applications creating .desktop files if only using Nix on a Linux system? Can only find a Firefox desktop-entry file in ~/.local/share/applications.
o1lo01ol1o has quit [Ping timeout: 250 seconds]
<srhb>
inquisitiv3: Not quite sure I parsed that correctly. Are you asking whether that's the normal location for desktop files for nix (not nixos) installed packages in a user profile? If so, yes.
NightTrain has joined #nixos
aw_ has quit [Quit: Quitting.]
spacefrogg has quit [Quit: Gone.]
aw has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<inquisitiv3>
srhb: I interpreted my question quite correctly. My question was if desktop applications (e.g. Firefox, Chromium, Libreoffice, etc.) usually create .desktop files somewhere so that the host system is able show the applications icons in the dashboard or similar.
spacefrogg has joined #nixos
<srhb>
Ah. I would expect not, but I don't actually know.
<inquisitiv3>
Thanks! When I'll stop looking. Couldn't find any documentation regarding that, neither on the Wiki or the manual. Guess it isn't so common to install desktop applications with Nix on another system?
asch has joined #nixos
<srhb>
inquisitiv3: Sure it is.
<srhb>
inquisitiv3: A lot of us never see those desktop files or use them for anything, though. So it's likely there's no real integration there, unless you manually do it.
asch has quit [Quit: WeeChat 2.4]
simukis_ has joined #nixos
<simukis_>
What could be reasons for $NIX_BUILD_CORES to be equal to 1 inside of plain `nix-shell all-packages.nix -A package`
<simukis_>
This is a machine with 72 cores.
o1lo01ol1o has joined #nixos
<srhb>
simukis_: A setting in /etc/nix/nix.conf
Zer000 has quit [Ping timeout: 246 seconds]
<srhb>
simukis_: namely cores
<simukis_>
a-ha, thanks!
<srhb>
simukis_: (if you set it to 0, it'll be the number of cores)
<srhb>
Smash that button, it'll eventually behave.
<gchristensen>
:D
<jpaus>
Hey guys, i just finished my first nixos install and was trying to install some packages. But nix-env spits out python2.7-caldavclientlibrary-asynk-asynkdev in top-level/python-packages is marked as broken. How would you handle that?
knupfer has quit [Ping timeout: 240 seconds]
<srhb>
jpaus: In general, this means that someone has actually tested the package and decided that it's actally broken. It probably needs some degree of rewriting in order to fix it.
<srhb>
(It might fail to build, or it might build and have possibly dangerous runtime problems...)
<jpaus>
but this completely stops me from installing or updating unrelated packages
<jpaus>
or at least it seems so to me ^^
<srhb>
jpaus: No, some package that you're (possibly indirectly...) depending on must be depending on the broken package in turn
<jpaus>
ok, so my best chance is waiting until it gets fixed or i gather enough knowledge to do it myself right?
<{^_^}>
[nixpkgs] @srhb opened pull request #58149 → haskellPackages.hfsevents: Fix eval on Darwin → https://git.io/fjJn3
<Yaniel>
uhh how do I make an application work with dconf on nixos
<srhb>
Actually, it wouldn't have saved us since it was the haskell set update..
<gchristensen>
oh?
<noonien>
i'm trying to get wifi on boot on a rPI 3B+, this is my config: http://vpaste.net/GrsFK, wpa_supplicant doesn't seem to start
<srhb>
gchristensen: iirc those are usually not run through a PR process.
Notkea[m] has joined #nixos
<noonien>
`systemctl list-units | grep wpa` also shows nothing
<srhb>
But still.
<gchristensen>
I think ofborg evaluates just like hydra does?
<srhb>
gchristensen: Right, my point is that ofborg doesn't catch it unless it's a PR, right? I mean, short of breaking subsequent evals.
<srhb>
It's still more visible on github though
<gchristensen>
much more
<srhb>
break alllll the things, yay
<srhb>
noonien: Does it even have a log? sudo journalctl -u wpa_supplicant
<srhb>
I'm guessing "no" because of list-units
<noonien>
it does not
<infinisil>
noonien: systemctl list-units --all ?
<srhb>
Maybe the sd image removes the wantedBy
<noonien>
infinisil: still nothing
<infinisil>
Yaniel: Have you tried programs.dconf.enable = true?
<srhb>
noonien: Well, I could explain it not autostarting because of: systemd.services.wpa_supplicant.wantedBy = mkOverride 50 []; -- in nixos/modules/profiles/installation-device.nix
<srhb>
noonien: But not sure if the unit is missing entirely..