<infinisil>
I think systemd.user.services.foo = { wantedBy = [ "graphical-session.target" ]; partOf = [ "graphical-session.target" ]; script = "echo foo"; };` should work then, there's a bunch of modules using this in nixpkgs
<infinisil>
First example for when the program doesn't need X, second for when it does
<qualiaqq>
I tried setting it up using systemd, but was getting an error message, which I found by greping for the name of the program on the output of "jouralctl".
<qualiaqq>
[FATAL] This application failed to start because it could not find or load the Qt platform plugin "xcb"
<infinisil>
qualiaqq: Use `journalctl -u <servicename> -e` to look at its logs/errors
<infinisil>
qualiaqq: Oh, yeah that's a popular error about QT, I think the solution is to install something about qt via nix-env
<qualiaqq>
Program I was trying to setup is called "albert" it works when I just launch it manually, but when I try to set it up as a systemd service it gets that error. I'm not sure what to do.
<infinisil>
Ah
<clever>
qualiaqq: it needs an env var related to QT to be set
<clever>
qualiaqq: $QT_PLUGIN_PATH
<infinisil>
clever: According to home-manager#269, setting PATH might work too
tenten8401 has quit [Remote host closed the connection]
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Taneb has quit [Quit: I seem to have stopped.]
b has quit [Quit: Lost terminal]
andymandias has joined #nixos
sir_guy_carleton has quit [Quit: WeeChat 2.0]
<waynr>
hmm i removed an argument from a nix expression i've written for a package i am trying to add to nixpkgs
<waynr>
but re-running `nix-build $PWD -A dnscrypt-proxy2` doesn't seem to result in a rebuild of the package
Guest29 has joined #nixos
<waynr>
i'm probably misunderstanding something fundamental
<waynr>
oh well, time to sleep
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
jperras has quit [Ping timeout: 244 seconds]
stepcut has quit [Ping timeout: 276 seconds]
spinus has quit [Ping timeout: 264 seconds]
<kalbasit>
I'm trying to `chmod 400 $out/path/to/file` in the `installPhase` but it does not seem to be working, the files has permissions `444`. What can I do to chmod it?
spinus has joined #nixos
sbdchd has quit [Remote host closed the connection]
<samueldr>
the store has strict restrictions to modes
* samueldr
is finding the right manual section
<samueldr>
last point in this section
<kalbasit>
samueldr: interesting, so I cannot install say SSH keys?
<samueldr>
» After the build, Nix sets the last-modified timestamp on all files in the build result to 1 (00:00:01 1/1/1970 UTC), sets the group to the default group, and sets the mode of the file to 0444 or 0555 (i.e., read-only, with execute permission enabled if the file was originally executable). Note that possible setuid and setgid bits are cleared. Setuid and setgid programs are not currently supported by Nix. This is because the Nix archives used in
<samueldr>
deployment have no concept of ownership information, and because it makes the build result dependent on the user performing the build.
<samueldr>
nope, secrets should stay out of the store
<kalbasit>
I see
<samueldr>
usually, you would reference an out of store path *somehow* in your configuration file or software
<kalbasit>
well I have private files that I'm trying to install (basically link to my home dir). I guess I'll have to manually manage these files
<rotaerk>
Is there some way to identify how much of a given nixpkgs commit is in the binary cache
<rotaerk>
I'm trying to bisect nixpkgs and running into lots of commits where it tries to build GHC or GCC
Guest29 has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<rotaerk>
Would be nice to know which commits to skip *before* attempting to build against them
spinus has quit [Ping timeout: 260 seconds]
<rotaerk>
(basically building my own project against various versions of nixpkgs to pinpoint which nixpkgs commit broke my project)
<rotaerk>
And really don't want to build against versions where I can't pull GHC and GCC and such from the cache
<rotaerk>
At least, until I get to the point where I'm really close to the breaking commit
Guest29 has joined #nixos
<worldofpeace>
Is there a distinct reason to, when building a go package, to turn the deps into an expression with say dep2nix when the deps are vendored in the source?
<qualiaqq>
clever, infinisil: however, now I'm not able to launch anything. It needs to be able to index and launch applications. I'll have to mess with this more. I guess it needs the PATH.
jperras has quit [Ping timeout: 248 seconds]
worldofpeace has quit [Quit: worldofpeace]
rschm has quit [Ping timeout: 264 seconds]
justbeingglad has joined #nixos
justbeingglad has left #nixos [#nixos]
phreedom has quit [Ping timeout: 250 seconds]
phreedom has joined #nixos
drakonis has quit [Read error: Connection reset by peer]
<colemickens>
When I enable a wireguard interface in configuration.nix, the right thing happens.
<colemickens>
When I remove it, the kernel module isn't included...
<colemickens>
But boot.kernelModules = ["wireguard"]; doesn't seem to help...
<colemickens>
any ideas? the wiki doesn't give any hints.
tadni_ has quit [Read error: Connection reset by peer]
tadni- has joined #nixos
wgas has quit [Quit: Leaving]
jperras has joined #nixos
slack1256 has quit [Ping timeout: 264 seconds]
jperras has quit [Ping timeout: 245 seconds]
winem_ has joined #nixos
hyper_ch2 has joined #nixos
freeman42x]NixOS has quit [Ping timeout: 240 seconds]
qualiaqq has quit [Remote host closed the connection]
<vaibhavsagar>
hi, how do I specify a nix-shell interpreter of more than one word?
<vaibhavsagar>
e.g. nix-shell -i "runhaskell -Wall"
knupfer has joined #nixos
<vaibhavsagar>
hmm, that seems to work, but not the single quotes I was using earlier
knupfer has quit [Ping timeout: 276 seconds]
<teto1>
vaibhavsagar: nix-shell can be capricious
<vaibhavsagar>
I've noticed :)
<vaibhavsagar>
I fixed the single quotes but my -p was using single quotes so then that failed
wgas has joined #nixos
rprije has quit [Ping timeout: 260 seconds]
testuser has joined #nixos
thc202 has joined #nixos
juhe has joined #nixos
andreabedini has joined #nixos
<adisbladis[m]>
worldofpeace: [go package source]: No. There is no reason to run dep2nix if all dependencies are vendored.
crmlt has joined #nixos
<cocreature>
I’m trying to get "nix-build" to work without an internet connection by passing --fallback but it just keeps trying to fetch a .narinfo for a while complaining that it can’t resolve the hostname and then stops
<cocreature>
am I misunderstanding what --fallback is supposed to do?
<Guanin>
Hi, I've written a long-running oneshot systemd service. Whenever I run nixos-rebuild, it seems to restart it. I've set restartIfChanged and stopIfChanged to false, but it still seems to run every time I rebuild the system?
init_6 has joined #nixos
rprije has joined #nixos
<LnL>
I think restartIfChanged is for running services
kyren has quit [Remote host closed the connection]
kyren has joined #nixos
yonk23 has joined #nixos
silver has quit [Read error: Connection reset by peer]
<infinisil>
Yeah, it chooses the most popular one by commit count
<worldofpeace>
I'll find some way to abuse it :P
<rschm>
hi, i am using nix for vim install on a macos. is there a way to pull down vim plugin which is a submodule of another project on github? thanks.
mariatsj_ has joined #nixos
MercurialAlchemi has quit [Ping timeout: 240 seconds]
Xiro` has quit [Quit: ERC (IRC client for Emacs 25.3.1)]
jensens has joined #nixos
mariatsji has quit [Ping timeout: 260 seconds]
<infinisil>
rschm: What are you using for managing vim plugins?
mariatsj_ has quit [Ping timeout: 240 seconds]
<rschm>
i am using the buildVimPlugin to bring down the plugin code.
<rschm>
but i am not sure how to do that with submodule.
<infinisil>
rschm: Submodules are just git repos, on github you can just click on the submodule to get to the other repo
<rschm>
you mean the parent repo?
<infinisil>
rschm: Then it's a normal src = fetchFromGitHub { owner = ... to get it to use that repo as a source
<infinisil>
Yeah, parent repo can have submodules as 'children'
<rschm>
how do i define the buildVimPlungin? from what i see i need the .git file.
vmandela has quit [Quit: Leaving]
camsbury has joined #nixos
<talyz>
adisbladis[m]: had any time to look at my PR again? :)
leo60228 has joined #nixos
<leo60228>
What's the best practice for a service that depends on another service enabling security options on that service?
jensens has quit [Ping timeout: 244 seconds]
<leo60228>
For example, I'm writing a custom module that provides a git-annex server service, which is effectively an ssh server with a user that has a public key and with the shell git-annex-shell.
<leo60228>
Should I: do nothing about password authentication, disable password authentication, or set it to mkDefault false?
<rschm>
infinisil: can you give an example how do i write the buildVimPlugin? thanks for your help.
<leo60228>
Setting a user whitelist seems intuitively overreaching.
<makefu>
gchristensen: i saw you implemented NIX_SSL_CERT_FILE, unfortunately it does not seem that it is enough to make a complete installation through an intercepting proxy. when running nix-channel --update nixpkgs (as part of the install script) it crashes with "peer certificate cannot be authenticated with given CA certificates"
<leo60228>
Also, IMO either ChallengeResponseAuthentication or PasswordAuthentication should be disabled. ChallengeResponseAuthentication defaults to a more secure implementation of PasswordAuthentication, so I don't think they should both be enabled by default.
<jophish>
I created a keypair using nix-store and have added the secret key to secret-key-files in /etc/nix/nix.conf and the public key to nix.binaryCachePublicKeys
<jophish>
however when I use 'nix copy --to ssh://blah /nix/store/blah' I get a 'lacks a valid signature'
<gchristensen>
makefu: hmmm.... works from here..?
<jophish>
I'm sure I've solved this issue a couple of times in the past, but I keep forgetting how to do it every time on a new machine
<timokau[m]>
I have a package that requires all its *reverse* dependencies to be build with `-std=c++11`. Is there a nice way to accomplish that? Also what is the default C stdlib and what is the best way to change it?
<makefu>
gchristensen: i've just read about the /etc/bashrc, i will try that as well
<gchristensen>
ok
jmiven has quit [Quit: pan]
<jophish>
hmm, setting secret-key-files to a non-existent file doesn't throw an erroir
nlytend has quit [Read error: Connection reset by peer]
<gchristensen>
https_proxy is set to literally the exact same contents? you're 100% sure?
<gchristensen>
(check carefully)
<makefu>
yep, also http://
<joko>
Hey, is anyone using pcmanfm? I have trouble using default applications, it appears MIME types are not working
<gchristensen>
I seem to remember a bug here
<makefu>
is there a way to get more logs from nix-channel?
<gchristensen>
"(FWIW the problem is with https_proxy="https://foobar:1111" old curl will ignore https and substitute in http, new curl can handle HTTPS proxies, and will try to negotiate TLS with the proxy server.)" -- not applicable :(
<makefu>
damn
<gchristensen>
makefu: try using Nix's curl instead of your system curl to fetch it
<{^_^}>
#43218 (by mpickering, open): whitebox-tools: init at 0.9.0
<infinisil>
,MCVE = Create a Minimal Complete and Verifiable Example when you have a problem, it helps a lot with troubleshooting. How to: https://stackoverflow.com/help/mcve
<makefu>
gchristensen: damn, same issue. i tried to build nix from source but it seems it desperately needs gcc with c++14 . not available on the centos7 i run on
<gchristensen>
NixOS works on a wide array of hardware, though x86 is best supported and should probably have 4+GB RAM and a nice sized disk (I'd hesitate before getting less than 128G)
zopsi has quit [Quit: Oops]
<gchristensen>
erm, less than 256G
mmercier_ has quit [Quit: mmercier_]
ThatDocsLady has quit [Remote host closed the connection]
ThatDocsLady has joined #nixos
zopsi has joined #nixos
Ariakenom has quit [Ping timeout: 244 seconds]
<justme_>
gchristensen: Probably, there may be problems with Wi-Fi drivers?
<hyper_ch2>
depends on the wifi
<gchristensen>
and nothing unique to nixos
<hyper_ch2>
using non-latest intel wifi usually works fine
andromeda-galaxy has quit [Ping timeout: 256 seconds]
rschm has quit [Read error: Connection reset by peer]
dan_ is now known as Guest38719
rschm has joined #nixos
Guest38719 has quit [Client Quit]
mroutis has quit [Ping timeout: 245 seconds]
hyper_ch2 has quit [Quit: Page closed]
<justme_>
clever: thank you, this may be helpful
johanot has quit [Quit: Leaving.]
justme_ has quit [Quit: Page closed]
Basic_Dud3 has joined #nixos
Basic_Dud3 has left #nixos [#nixos]
numkem has joined #nixos
nlytend has joined #nixos
Sonarpulse has joined #nixos
<mpickering>
Looks like rustc-1.27.0 is broken?
drakonis_ has joined #nixos
pie_ has quit [Remote host closed the connection]
pie_ has joined #nixos
knupfer has joined #nixos
<numkem>
Hello, I'm starting to get used to use nix-shell and whatnot but I can't seem to make protobuf work for one of my project. When launching the app I get this message: -glibc-2.27/lib/librt.so.1: symbol __close_nocancel, version GLIBC_PRIVATE not defined in file libc.so.6 with link time reference. Yet it works fine within a container
<Myrl-saki>
gchristensen: I've lived on a 32 GB /
<Myrl-saki>
0/10, would not recommend.
<Myrl-saki>
But it's possible
jperras has quit [Ping timeout: 276 seconds]
<gchristensen>
yeah it was possible but miserable
goibhniu has quit [Ping timeout: 244 seconds]
<andi->
I have a few servers running on 20G disk space.. garbage collection every day after update...
mroutis has joined #nixos
<andi->
works fine
numkem has quit [Quit: WeeChat 2.0]
Intensity has joined #nixos
numkem has joined #nixos
<gchristensen>
canyou keep >1 generation?
<andi->
yes
<andi->
it runs on defaults
<andi->
it also runs on 1G of RAM (with 1G of swap)
<gchristensen>
incredible
<andi->
right now just 8.1GB used.. and that thing stores graphite time series...
<gchristensen>
ahh nice
<gchristensen>
I don't have such small expectations of much of my hardware
<andi->
:)
<gchristensen>
except for my router, which is quite minimal
<andi->
I am still procrastinating migrating my home router to NixOS..
rschm has quit [Remote host closed the connection]
<gchristensen>
come on in, the waters fine
<andi->
hrhr
<infinisil>
Procrastination is my philosophy
<srhb>
andi-: Out of curiousity, did you forgo the nix 2 upgrade?
rschm has joined #nixos
<andi->
forgo?
<srhb>
Skip
goibhniu has joined #nixos
<andi->
no
<srhb>
Wow. They must be very minimal indeed to fit in 2G memory then.
<srhb>
andi-: I believe some/all of the memory leaky fixes are in there, but not stable.
<andi->
gchristensen: thanks, I'll definitly have a look... part of the upgrade will be a switch of hardware.. I did not manage to settle on anything yet
<gchristensen>
<gchristensen>
ah
<gchristensen>
I'm using a apu2c4 and a juniper switch
<{^_^}>
[nixpkgs] @samueldr opened pull request #43257 → Revert removal of `linux_testing_bcachefs` from "linux: remove versions unmaintained upstream" → https://git.io/fNJml
<andi->
I am doing a bit of bittorent on the router.. thus I need 20G+ of RAM..
rihards has joined #nixos
<andi->
so the choice isn't very easy there
<gchristensen>
I mean ... you could do it not on the router ...
<andi->
yes
<andi->
but then yet another machine..
blankhart has quit [Read error: Connection reset by peer]
<andi->
I have 1 x86 router that does (historically) the network foo + some torrenting, a server that serves as NAS/VM-Host with two (now ancient) Xenos, a Desktop and a bunch of notebooks..
<andi->
It's just me being lazy
<gchristensen>
how much torrenting can you do with just linux ISOs and other free/libre content
blankhart has joined #nixos
<andi->
alot!
<andi->
imagine every package being a torrent..
<gchristensen>
imagine having to establish peers for every single drv you build
<infinisil>
Ipfs!
<samueldr>
ipfs sounds like a painful onomatopeia
rschm has quit [Remote host closed the connection]
rschm has joined #nixos
<gchristensen>
oh look at the time, it is silly-convo-o'clock
* manveru
shoves people into #nixos-chat :)
<andi->
I think IPFS would be a nice thing.. if it works
<gchristensen>
manveru++
<{^_^}>
manveru's karma got increased to 1
<samueldr>
on a more serious note, other than the light guidelines for backporting, there are no backport team or anything more formal written down, right?
<gchristensen>
samueldr: that should be part of the responsibilities of the PR author (determine if it is appropriate) and merger -- (to agree or not if a backport is appropriate)
<andi->
samueldr: not that I know... It would proably be nice if we had a guideline (non binding, informal)
<gchristensen>
agreed
<samueldr>
r-ryantm seems to not care (understandably) for backports
<srk>
I'm now trying to run hydra slaves without disk, only 12G overlayfs in 24G ram
goibhniu1 has joined #nixos
<srk>
works fine if I garbage collect often enough
<clever>
srk: nix now has a min-free and max-free config flag, for auto-gc
<samueldr>
and I'm sure many point updates from unstable aren't even being considered by their authors for unstable
<srk>
clever: wow! that's handy for this use case
jperras has joined #nixos
<samueldr>
(I would assume the authors themselves use unstable)
<samueldr>
s/for unstable/for stable/
<clever>
srk: both are in bytes, in nix.conf
<srk>
will try for sure
goibhniu has quit [Ping timeout: 240 seconds]
goibhniu1 is now known as goibhniu
graphene has quit [Remote host closed the connection]
<andi->
samueldr: yes.. I see the same.. It makes me a bit angry from time to time
graphene has joined #nixos
<srk>
in fedora, you would push updates everywhere if it was just minor version bump, breaking changes only went to rawhide
<samueldr>
for 18.09, maybe it would be relevant to figure out a team or system to try and keep updates trickling in stable
<drakonis_>
unstable say what now
<srk>
you could run a bot, that would try to apply backports automatically where it makes sense
<gchristensen>
it is a very human decision
<srk>
like only backport if minor/micro part version changes
<srk>
*of the
<andi->
samueldr: well I guess tooling is the only thing the holds us back.. Provide people with a list of things that were (minor) bumped or a newer (in only minor versions) and someone could have a go at it
<gchristensen>
you have to read and understand the changelog to decide
<andi->
but yeah. changelogs become important (again)
<srk>
indeed
<samueldr>
there is software which don't do semver, which sometimes bump "major" but changes are minimal, and sometimes bump "minor" and changes are big :/
<samueldr>
so automatisation is hard
__monty__ has quit [Quit: leaving]
<andi->
well you need the diff of packages / upgrades over time in some sort of list. That automation is reasonably easy. I wouldn't want to work myself through the git log every other day...
<gchristensen>
that was the most miserable thing
<gchristensen>
looking through commit logs to send out advisory emails just about killed me
<{^_^}>
[nixpkgs] @vcunat pushed commit from Kim Ingmar Ferenc ALLAMANDOLA to master « nilfs-utils: force enable limbount support »: https://git.io/fNJO2
mmercier_ has joined #nixos
<{^_^}>
[nixpkgs] @vcunat pushed commit from Kim Ingmar Ferenc ALLAMANDOLA to release-18.03 « nilfs-utils: force enable limbount support »: https://git.io/fNJ3l
<makefu>
gchristensen: i think i have an idea what may be a problem .... the intercepting proxy downgrades all ssl stuff to dh1024, the nix curl is configured with cipher suites "!LOW" ... what a pain
<pie_>
what do you guys burn dvds with
<pie_>
more like intercepting proxy baad
<pie_>
:P
<{^_^}>
[nixpkgs] @Assassinkin opened pull request #43260 → pythonPackages.sumo: init at 1.0.9 ( And other packages that sumo depends on) → https://git.io/fNJ3D
spinus has quit [Ping timeout: 240 seconds]
iyzsong has quit [Read error: Connection reset by peer]
mmercier_ has quit [Ping timeout: 244 seconds]
pie__ has joined #nixos
pie_ has quit [Ping timeout: 244 seconds]
igo95862 has quit [Quit: igo95862]
ihar has joined #nixos
<jophish>
kk
doyougnu has joined #nixos
humanoyd has quit [Quit: WeeChat 2.1]
jb55 has joined #nixos
blankhart has quit [Ping timeout: 240 seconds]
__monty__ has joined #nixos
acarrico has joined #nixos
<{^_^}>
[nixpkgs] @Ericson2314 pushed to master « haskell generic-builder: Make test and benchmark tool depends `nativeBuildInputs` »: https://git.io/fNJnC
winem_ has quit [Ping timeout: 245 seconds]
<pie__>
unstable, or at least some applications seem to have issues with qt
<pie__>
QT_PLUGIN_PATH doesnt end up getting set or something and so programs fail with a slightly misleading error message about not finding xcb
<pie__>
wel at least i run into this when doing nix-shell -p k3b
<pie__>
and ivehad similar issues before with other things
sigmundv has quit [Ping timeout: 276 seconds]
<pie__>
hmm google never brought this page up for me oddly enough
blankhart has joined #nixos
nlytend has quit [Ping timeout: 264 seconds]
drakonis_ has quit [Remote host closed the connection]
hakujin has joined #nixos
civodul has quit [Quit: ERC (IRC client for Emacs 26.1)]
srl295 has joined #nixos
mightybyte has joined #nixos
rihards has quit [Quit: rihards]
<mightybyte>
Does nix have GHC 8.4.3 as a compiler yet? I'm only seeing haskell.compiler.ghc841.
<infinisil>
But HIE is working well for me, with stack projects at least
<tnks>
acowley: otherwise, I'm using Dante... but it's a bit tough to tell people that to develop like me they need exactly Nix+Emacs+Dante. It would be nice to have Nix+SomethingElse.
<acowley>
I think it's likely that patching intero's elisp to find the intero executable again won't be hard, but it's not something I want to do
Ariakenom has joined #nixos
<acowley>
I've always found dante to be just enough worse than intero to motivate using intero. Things like targets and completions from imported modules.
<elvishjerricco>
acowley: I think there's a fix for that hie thing lying around somewhere...
<tnks>
acowley: I see... I haven't worried as much about completions yet... maybe I don't even have that enabled correctly.
<acowley>
I really like lsp-ui in emacs for C++, so I'd be interested in hie, but it's such a sprawling project compared to intero and dante and never works as reliably as either of those. (I've never actually gotten hie to work, but I trust that others have.)
<acowley>
tnks: Completions are great for something like "import qualified Data.Text as T", and now you type "T." and get completions and types popping up for just that module.
<tnks>
acowley: that's a good point. I'm not opposed to it.
<acowley>
elvishjerricco: That was to address a problem mentioning package-id's?
<elvishjerricco>
acowley: Not sure anymore :P
<acowley>
hie can't find my hoogle db either, though it's built as part of the nix-shell, so I got the impression that maybe it wasn't entirely aware of shell.nix
<tnks>
acowley: cool, I just tried out Dante with completion of some random file. I get C-M-/ (dabbrev-completion)
<tnks>
I guess it works sometimes for you, but not all the time?
SuperKazuya has quit [Ping timeout: 252 seconds]
<acowley>
tnks: when I can get intero working with nix, it works great (much better than dabbrev) :D. The problem is a recent change to intero to avoid having to rebuild the intero executable quite so often by having stack cache binaries, but we've got an intero on $PATH from nix, so I think a patch is needed to skip using the cache.
rschm has quit [Remote host closed the connection]
rschm has joined #nixos
Thra11 has joined #nixos
xy2_ has quit [Ping timeout: 256 seconds]
xy2_ has joined #nixos
mariatsji has joined #nixos
Coyin has quit [Ping timeout: 244 seconds]
Coyin has joined #nixos
tzemanovic has joined #nixos
hakujin has quit [Quit: WeeChat 2.1]
fragamus has joined #nixos
tzemanovic has quit [Ping timeout: 244 seconds]
muzzy_ has joined #nixos
<muzzy_>
how can I enable proprietary amd drivers to play games?
Coyin has quit [Ping timeout: 248 seconds]
mariatsji has quit [Remote host closed the connection]
mariatsji has joined #nixos
mariatsji has quit [Read error: Connection reset by peer]
mariatsji has joined #nixos
mariatsji has quit [Remote host closed the connection]
mariatsji has joined #nixos
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
coot has quit [Quit: coot]
coot has joined #nixos
chimay has quit [Ping timeout: 244 seconds]
mariatsji has quit [Ping timeout: 268 seconds]
<acowley>
Removing the binary caching got intero working for me in emacs, btw.
<moet>
ok, that checks out .. thank you .. /nix/store/f0y3dzaz15cxc6m9m8nh3cglrmir9456-vulkan-loader-1.0.61.1/lib/libvulkan.so.1
crmlt has joined #nixos
coot has quit [Quit: coot]
coot has joined #nixos
<gchristensen>
man infinisil is really doing great cool work
<srk>
any idea why I'm getting not of type `list of submodules' when I add '.addresses' to get rid of the warning? such as networking.interfaces.eno1.ip4.addresses = [ { address = "172.17.4.5"; prefixLength = 23; } ];
<gchristensen>
18.03?
<srk>
master
<{^_^}>
[nixpkgs] @Ericson2314 opened pull request #43264 → haskell infra: Use splicing like normal `callPackage` → https://git.io/fNJwD
<gchristensen>
eno1.ipv4.addresses ...
<srk>
ah
<srk>
thanks!
pie___ has joined #nixos
yonk23 has quit [Quit: leaving]
<azazel>
I'm getting an evaluation error on unstable if I define "services.kubernetes.addons.dashboard.enable = true;":
<azazel>
error: The option `services.kubernetes.addons.dashboard.rbac' is used but not defined.
<azazel>
so my question: is this a bug or I'm supposed to specify the enable AND the rbac? (although what I specified is supposed to be the default)
<azazel>
I was thinking that in such a case just by definining sevices..enable would be merged with the default config
<azazel>
I've a wrong expectation?
moet has joined #nixos
twitchard has joined #nixos
tippenein has quit [Ping timeout: 260 seconds]
<moet>
how does nixos indicate the location of shared libraries within nix-shell?
tippenein has joined #nixos
<tobiasBora>
Hello,
<johnw>
moet: everything is referenced by absolute path, so whichever dependencies your nix-shell brings into scope, they know exactly what path their shared lib dependencies live at
<johnw>
there shouldn't be any need to depend on LD_LIBRARY_PATH or those sorts of mechanisms
rschm has quit [Remote host closed the connection]
rschm has joined #nixos
walter has quit [Quit: WeeChat 2.1]
<{^_^}>
[nixpkgs] @aszlig pushed to staging-18.03 « stdenv-setup: Remove superfluous check for /bin/sh »: https://git.io/fNJKE
<{^_^}>
[nixpkgs] @aszlig pushed to staging « stdenv-setup: Remove superfluous check for /bin/sh »: https://git.io/fNJKz
<tobiasBora>
I'm trying to understand what is the goal of all these tools: kubernetes, docker swarm, as well as nixOps (is there any other nix-based tool for deploying multiple servers?). How is nixOps compared to the two others tools?
<moet>
johnw: so if i look under /lib or /usr/lib i should find symlinks to the shared library located in /nix/store, right?
<johnw>
I don't think Nix uses /lib
<moet>
johnw: that was going to be my next question..
<moet>
i'm trying to find the absolute path so then i can find out if the language runtime the application is using (rust) looks in that path
foldingcookie has joined #nixos
<moet>
using `find` just churns up the path in /nix/store
<domenkozar>
tnks: well it literally translates stack build plan into nix expression tree
<srk>
tobiasBora: nixops is exactly for deploying multiple machines, either physical servers or cloud instances (multiple backends supported)
<srk>
not sure how does it compare to docker swarm or kubernetes but so far it feels quite superior to ansible/docker/salt (mainly due to nix/nixos)
muzzy_ has quit [Ping timeout: 240 seconds]
nlytend has quit [Quit: disconnected.]
<srk>
I wonder if enabling nix-serve on hydra is the right way to be able to use it as a cache
<tobiasBora>
srk: can nixops deal with private data, like salt does?
<tobiasBora>
srk: great thanks! And any idea if it can deal with "load balancing"? (I'm quite new to this field, so let me know if my question is not specific enough)
<srk>
a bit, I guess you want http balancing
<srk>
if so try haproxy
<tnks>
domenkozar: right. I've just been able to coast on the curated set in nixpkgs. Your build of HIE was the first time I saw a project so strongly delegating to a Stackage build plan.
<srk>
tobiasBora: you can use nix to create bunch of machines (workers) easily with map and lib.range 0 10
<tnks>
speaking of which, is it well documented how nixpkgs/master flows into nixpkgs/nixpkgs-unstable and nixpkgs/nixos-unstable?
aarvar has joined #nixos
<tnks>
I asked this question before a while ago, but don't remember getting a firm answer.
<srk>
tobiasBora: perftest_node(s) are generated this way
hotfuzz_ has quit [Ping timeout: 244 seconds]
* infinisil
gets on restoring the holy karma
viric has joined #nixos
<pie___>
(clever) so im stumped. i have the same dnsmasq config on two machines, but on one it succeeds at sending normal queries to normal dns, on the other it doesnt. i.e. /etc/dnsmasq-resolv.conf on one of the machines remains empty
<pie___>
hm thats an idea. no idea how to find out, unless dnsmasq itself creates it...
* pie___
checks something
<pie___>
yeah, i dont know how to find it
<gchristensen>
hmm interesting
<pie___>
google doesnt yield anything immediately useful
drakonis__ has joined #nixos
<LnL>
gchristensen: has some gdb instructions and the nix package has that enabled by default IIRC
drakonis__ is now known as Drakonis
Drakonis is now known as drakonis
<gchristensen>
ok, cool, wow!
<gchristensen>
now to tell `perf` to look at them :)
<pie___>
hmm on the working machine it seems it might be generated by networkmanager
* pie___
coincidentally ran into resolvconf -l
<pie___>
oh hey it says in the file "generated by resolvconf"
<pie___>
well no i have a starting point, yay...
<pie___>
andi-, so it looks like thie issue might be that the machine that doesnt work, doesnt run networkmanager and just doesnt have anythng generating the contents of the file
Drakonis[m] has quit [Changing host]
Drakonis[m] has joined #nixos
Drakonis[m] has joined #nixos
<andi->
there you go :)
<pie___>
man dns is pretty complicated
<gchristensen>
dns is simple
<pie___>
dont really know what i could replicate the behavirous with though :x
<gchristensen>
dhcp is complicated
<pie___>
gchristensen, client side dns seems like a clusterf
<andi->
now imagine having to merge both DNS settings from DHCP (IPv4) and IPv6 RAs (RDNSS) with multiple processes ;)
<{^_^}>
[nixpkgs] @xeji pushed commit from @r-ryantm to staging « shared-mime-info: 1.9 -> 1.10 (#42948) »: https://git.io/fNJDh
<betaboon>
so you want to put the dns-server propagated via dhcp as dns-server for dnsmasq ?
<pie___>
yes
i0-dfn has quit [Ping timeout: 240 seconds]
<pie___>
hmm dhcpcd might be more appropriate for querying, im not sure yet (its installed)
<betaboon>
pie___: i thought that is the default configuration of dnsmasq !?
i0-dfn has joined #nixos
<pie___>
betaboon, thats fine but something actually needs to get the upstream servers
wgas has joined #nixos
<pie___>
looks like its normally NetworkManager
<betaboon>
i was under the impression that dhcpcd writes dns-servers that are propagated via dhcp into resolv.conf, and dnsmasq by default uses the system-dns-server for dns-forwarding? are those assumptions wrong ? oO
<pie___>
so, i *think* by default dhcpcd writes them to resolv.conf via resolvconf
<pie___>
dnsmasq by default uses the system dns servers yes
<pie___>
however when you enable dnsmasq dhcpcd no longer does resolvconf i guess
<{^_^}>
[nixpkgs] @matthewbauer pushed 500 commits to staging: https://git.io/fNJ73
<betaboon>
pie___: services.dnsmasq.servers ?
<pie___>
oh right
* pie___
rolls eyes xD
<pie___>
when you're in deep
blankhart has quit [Ping timeout: 260 seconds]
<betaboon>
pie___: which is technically not writing dnsmasq-resolv.conf but calling dnsmasq with -S
<pie___>
yeah
<pie___>
well, lately im learning a lot more about linux client side dns config than i ever wanted...
<pie___>
the luckiest thins is that im at least doing it on nixos..
<clever>
ive tied dhcpcd into bind before on gentoo
<clever>
but i never recreated it on nixos
<pie___>
something for nixos.wiki
<betaboon>
can anyone suggest a good terminal emulator with truecolor support? i have been using urxvt for ages. just recently switch to a patched st. but I
<betaboon>
am not completly sattisfied
seafood has quit [Quit: seafood]
<pie___>
thanks for the input guys. ill try to do something better later, will stick with static config for now
<tobiasBora>
clever: but it means that you need to remove basically the current OS and replace it with nixos right? It's not possible to have in parallel both debian and nixos for example?
<clever>
tobiasBora: yeah, you have to replace the OS
<tobiasBora>
ok, thank you. The best way if you don't want to replace the OS is to install nixos in a container that runs on, say, debian ?
<clever>
and its difficult to run it under a container as well
trcc has joined #nixos
<srk>
manageable
<kalbasit>
,locate colemak
<{^_^}>
Found in packages: kbd, oh-my-zsh
<kalbasit>
it's weird that I can't do `loadkeys colemak` in the NixOS isntalled
<tobiasBora>
clever: really? What's the problem with containers? So there is no solution if I want to deploy some nix "receipes" on non-nixos systems?
<clever>
tobiasBora: the nixos container framework assumes the host also runs nixos, so its tricky to get the kernel to run
<{^_^}>
#9884 (by o11c, open): Installing NixOS in a container with a non-NixOS host is painful
<tobiasBora>
so I don't get it. Let's say that I'm on debian. Then on it I create a systemd/LXD container that runs nixos. Finally, I forward on debian the ports of this guest nixos, and provide it to nixops
<clever>
check the above issue
betaboon has quit [Quit: WeeChat 2.0]
<infinisil>
,dnw = "It doesn't work" literally only has 0 bits of information useful for debugging, please be more specific: What doesn't work? What's the error?
<{^_^}>
dnw defined
<tobiasBora>
ok thank you clever! Hope this bug is going to be fixed then ;)
<infinisil>
,dnw kalbasit
<{^_^}>
kalbasit: "It doesn't work" literally only has 0 bits of information useful for debugging, please be more specific: What doesn't work? What's the error?
<srk>
looks quite old, we are running nixos on lxc without much issues
pie___ has joined #nixos
swampmonster is now known as justanotheruser
erratic has quit [Ping timeout: 256 seconds]
__monty__ has quit [Quit: leaving]
<kalbasit>
infinisil: sorry for the vague message, `ioadkeys colemak` yields the error `cannot open file colemak` on the nixos-installer as of yesterday (not sure which channel but the most recent as of yesterday)
<kalbasit>
s/ioadkeys/loadkeys/
<infinisil>
kalbasit: Not sure why it doesn't work, never used loadkeys
<infinisil>
kalbasit: But there's a different way to set the layout in the installer
<infinisil>
It might generate a file usable for loadkeys
<kalbasit>
I see
<kalbasit>
can I change `/etc/nixos/configuration.nix` in the installer?
<infinisil>
Yeah
<infinisil>
Ah yeah, you can just set the layout there!
blonkhart has joined #nixos
<infinisil>
kalbasit: I prefer to use `services.xserver.layout = "<yourlayout>"; i18n.consoleUseXkbConfig = true;`, which is how the linked stuff even applies
<infinisil>
(And yes it's using xserver there, but it doesn't actually require X)
freeman42x[nix] has joined #nixos
<kalbasit>
cool, let me try that then, thx
<Ralith>
I'm passing `{ config = { overlays = [ overlay ]; }; }` to nixpkgs, where overlay sets `llvm_6 = super.llvm_6.override { debugVersion = true; }`. This does not appear to be affecting the llvm_6 defined in the resulting pkgs. What might I be missing?
<infinisil>
Ralith: { overlays = [ overlay ]; }, it's separate from config
phreedom has quit [Ping timeout: 250 seconds]
<clever>
Ralith: overlays is its own attribute, beside config, not inside
<Ralith>
well that's a delightfully silent error
<{^_^}>
[nixpkgs] @Ericson2314 merged pull request #43264 → haskell infra: Use splicing like normal `callPackage` → https://git.io/fNJwD
camsbury has quit [Remote host closed the connection]
camsbury has joined #nixos
i0-dfn has quit [Ping timeout: 256 seconds]
<kalbasit>
is `/etc/nixos/configuration.nix` meant to be edited directly, or via some nix command?
i0-dfn has joined #nixos
<clever>
kalbasit: its meant to be edited manually
<kalbasit>
thx
camsbury has quit [Ping timeout: 256 seconds]
<infinisil>
Until I get my Nix AST editor working at least!
* infinisil
hasn't started even, don't get your hopes up
lsyoyom has quit [Ping timeout: 265 seconds]
jperras has quit [Quit: WeeChat 2.1]
i0-dfn has quit [Ping timeout: 264 seconds]
i0-dfn has joined #nixos
wgas has left #nixos ["Leaving"]
emil1234[m] has joined #nixos
tzemanovic has joined #nixos
slack1256 has quit [Ping timeout: 264 seconds]
xy2_ has quit [Ping timeout: 264 seconds]
lsyoyom has joined #nixos
rprije has joined #nixos
alexteves has quit [Remote host closed the connection]
init_6 has joined #nixos
Guest53532 has quit [Ping timeout: 240 seconds]
sigmundv_ has quit [Read error: Connection reset by peer]
Mrmaxmeier has quit [Read error: Connection reset by peer]
graphene has quit [Ping timeout: 260 seconds]
Olgierd has quit [Ping timeout: 260 seconds]
Mateon3 has joined #nixos
sbdchd has joined #nixos
tadni- has quit [Read error: Connection reset by peer]
sigmundv- has joined #nixos
Mateon1 has quit [Read error: Connection reset by peer]
Mateon3 is now known as Mateon1
realrokka has quit [Quit: rip]
tnks has joined #nixos
Olgierd has joined #nixos
Mrmaxmeier has joined #nixos
Aexoden has quit [Ping timeout: 260 seconds]
realrokka has joined #nixos
Olgierd has joined #nixos
tadni_ has joined #nixos
seafood has quit [Quit: seafood]
init_6 has quit [Ping timeout: 240 seconds]
tnks has quit [Ping timeout: 260 seconds]
<trcc>
I am trying to add a haskell package to shell.nix, but I can simply not figure out how. Should I use nix-env -f "<nixpkgs>" -iA haskellPackages.Allure?
<trcc>
This seems to add it more globally, which I am not interested in
jluttine has quit [Ping timeout: 260 seconds]
Aexoden has joined #nixos
Sonarpulse has quit [Ping timeout: 264 seconds]
init_6 has joined #nixos
seafood has joined #nixos
jluttine has joined #nixos
dhess has joined #nixos
<dhess>
Anyone around who's running bird on NixOS?
<infinisil>
trcc: gist your shell.nix?
<trcc>
infinisil: Think I just solved it
<trcc>
ls
<{^_^}>
[nixpkgs] @reardencode opened pull request #43269 → Add libudev, which Ledger Live app requires → https://git.io/fNUec
ericsagnes has quit [Ping timeout: 244 seconds]
<pie___>
how do i sed up rsync on a nixos machine so that i can rsync over ssh?
<rprije>
For some reason my nix-shell isn't showing me a different prompt. I've set my own bash prompt_command so that probably broke it. What environment variables can I key off to build an explicitly different prompt when in nix-shell?
init_6 has quit []
logzet has quit [Remote host closed the connection]