<slack1256>
I got a shell.nix file with a mkShell expression inside. If I run nix-collect-garbage that environment will be have to created anew on the next nix-shell call
<slack1256>
How do I make them persist nix-collect-garbage?
thc202 has quit [Ping timeout: 252 seconds]
<hlolli>
What could explain missing sndfile.h in an environment `nix-shell -p libsndfile` are the source files omitted?
<fyber>
wpa-supplicant never works for me, is there a way to use networkmanager in the install ISO?
<fyber>
I tried enabling it in the config but it conflicts with a bunch of things
<samueldr>
fyber: the graphical iso (which by default starts in a virtual console) has network-manager available, and can be used in the console using nmtui
<samueldr>
I have to check if the "minimal" iso has it too
<fyber>
I'm on minimal right now, doesn't work :(
<samueldr>
aw
<fyber>
Should I just grab the graphical one? It's kind of a pain to write the ISO again
<fyber>
or maybe someone can help me figure out why wpa-supplicant doesn't work
<samueldr>
well, let's first figure out: what does "doesn't work" mean?
<fyber>
I did `wpa_passhrase SSID PASSWORD > /etc/wpa_supplicant.conf`
<fyber>
and then started wpa supplicant
<fyber>
didn't connect to the network, and there's nothing in the systemctl status
<fyber>
The network is WPA2 (Personal) so it's not too weird I don't think
<samueldr>
should be fine
<fyber>
I had the same problem on my old distro too
<samueldr>
are there interfaces when you use `iw dev`?
<samueldr>
how did you solve the problem on the other distro?
<fyber>
yes, wlp2s0
<fyber>
I just installed networkmanager
<fyber>
and everything magically worked
<samueldr>
good, at least it's not because it's an unsupported (by default) network card... it's a pain to get the firmwares on the iso sometimes :/
<fyber>
Does wpa_supplicant work for anyone?
<fyber>
I can't even get it to work on a raspberry pi/raspbian, which is as close to a common linux distro as it's going to get
<samueldr>
wifi is so bad in this overcrowded area that I always use USB ethernet on laptops
<fyber>
I'm convinced everyone who says they use it is lying
<fyber>
I just have to write the config and restart wpa_supplicant, right?
<samueldr>
are there any special characters, or spaces, in either of the SSID or passphrase?
<samueldr>
(I'm not sure, it's been years since I last used wpa_supplicant, and not with a systemd service)
<fyber>
nope, all of them are a-z 0-9
<samueldr>
and the capitalization is followed for the SSID too? (dumb question, but let's not assume anything)
<fyber>
yes
<samueldr>
:/
<fyber>
it's actually all lowercase too
wpcarro has joined #nixos
<fyber>
actually the simplest possible scenario
<samueldr>
and iw dev doesn't show the SSID after starting wpa_supplicant?
<fyber>
nope
<fyber>
I found something interesting
<fyber>
when I start wpa supplicant it says wlp2s0 and unnamed netdev
<fyber>
in `iw dev`
<fyber>
but only wlp2s0 when wpa supplicant is stopped
<fyber>
I plugged in an ethernet dongle but that only has drivers for Windows so it didn't work :(
<samueldr>
I've been looking at older IRC logs, there's a user that seems to have good troubleshooting steps
<fyber>
paste?
<samueldr>
`systemctl status wpa_supplicant` there's probably log entries at the bottom
JonReed has joined #nixos
<fyber>
tried that, no log entries other than "success"
<samueldr>
(I'm reading a couple and taking the most successful replies)
<samueldr>
:/
wpcarro has quit [Ping timeout: 240 seconds]
<samueldr>
(I know I would get the graphical iso just because nmtui makes me stop thinking about thinking)
silver has quit [Read error: Connection reset by peer]
wpcarro has joined #nixos
wpcarro has quit [Remote host closed the connection]
mounty has quit [Ping timeout: 252 seconds]
iyzsong has joined #nixos
mayhewluke has quit [Ping timeout: 252 seconds]
mayhewluke has joined #nixos
wpcarro has joined #nixos
wpcarro has quit [Remote host closed the connection]
acarrico has quit [Ping timeout: 246 seconds]
acarrico has joined #nixos
<fyber>
13h (estimated) to 9h with KDE
tzemanovic has quit [Remote host closed the connection]
<fyber>
that's more than I was expecting
tzemanovic has joined #nixos
slack1256 has joined #nixos
<slack1256>
I crafted a shell.nix file to use with nix-shell. When entering it takes 50 seconds. Is there anyway to see where the time is being spent?
tzemanovic has quit [Remote host closed the connection]
<infinisil>
slack1256: Add as many -vvvvv's as you like
<slack1256>
Oh, let me see
<ldlework>
infinisil: lol
hlolli_ has joined #nixos
sigmundv has quit [Ping timeout: 246 seconds]
tzemanovic has joined #nixos
andrewrk has joined #nixos
<andrewrk>
I upgraded nixos and now apps can't find opengl
<andrewrk>
did something change?
<slack1256>
infinisil: It works, thanks!
hlolli has quit [Ping timeout: 252 seconds]
fresheyeball has joined #nixos
<fresheyeball>
hey out there
<fresheyeball>
I have a list of drvs
<fresheyeball>
[ hpack stylish-haskell ]
<fresheyeball>
can I convert that to a string, of all the binaries that list exposes?
<fresheyeball>
"hpack stylish-haskell"
<fresheyeball>
in this case
<infinisil>
fresheyeball: Not on the Nix level, haskell packages don't expose which executables they contain
<infinisil>
Only whether or not they do have executables
<fresheyeball>
infinisil: so with buildInputs
<fresheyeball>
they just end up on the $PATH
<fresheyeball>
I think this might be doable with foldl'
<fresheyeball>
but I don't understand how they end up on the path with buildInputs
<fresheyeball>
here is the goal
<fresheyeball>
I have pimpped out nix-shell for my project with useful development tools
<fresheyeball>
and now everyone uses tools of the same version, which is awesome
<fresheyeball>
and I want to print a help message stating which thing on $PATH have been added
<fresheyeball>
turns out the doc gets stale fast as people add tools
<fresheyeball>
so it would be nice to have the [drv] and programmatically make the list
<fresheyeball>
my first thought, is that if I can understand how this is done in buildInputs, or makeCmdBin
<fresheyeball>
then I should be able to do this in a combo of nix and bash
wpcarro has joined #nixos
slack1256 has quit [Remote host closed the connection]
tzemanovic has quit [Remote host closed the connection]
tzemanovic has joined #nixos
<{^_^}>
[nixpkgs] @Infinisil opened pull request #46410 → Fix idris gcc builds on darwin and disable broken package → https://git.io/fA21x
ydlr has quit [Remote host closed the connection]
tomberek has joined #nixos
<tomberek>
Any have experience with cross-compiling for ARM? I'm getting some failures, but I'm not sure where to start fixing them. Here's a basic thing that fails: `nix build nixpkgs.pkgsCross.armv7l-hf-multiplatform.gdb`
Dedalo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<tomberek>
infinisil: do you get an error with that?
hlolli_ has quit [Ping timeout: 272 seconds]
<tomberek>
`builder for '/nix/store/gwjlxlzc53hk8ws5w4dqbyhwiy45gpll-gdb-8.1.1-armv7a-unknown-linux-gnueabihf.drv' failed with exit code 2; last 10 log lines: make[3]: *** [Makefile:871: bfdt.stamp] Error 127 /nix/store/czx8vkrb9jdgjyz8qfksh10vrnqa723l-bash-4.4-p23/bin/bash: ./chew: No such file or directory make[3]: *** [Makefile:878: cache.stamp] Error 127 make[3]: Leaving directory '/tmp/nix-build-gdb-8.1.1-armv7a-unknown-linux-gnuea
<infinisil>
tomberek: Assuming we're using both linux_x86-64 and considering that nix builds are reproducible and you *did* get an error suggests that I'll get the exact same error
<tomberek>
infinisil: that's the hope, but it's always good to confirm
<infinisil>
tomberek: regarding the error, it might have compiled a binary ./chew in the build with arm as a target, which it then tries to run in the build, which will fail with a "no such file" error because it can't find the libraries it needs, because it's still running on x86-64, not arm. So the solution might be to make it compile chew with the build host as a target
pie_ has joined #nixos
<tomberek>
infinisil: thanks, gdb being a relatively base package I had hoped I was just doing something wrong.
sbdchd has quit [Remote host closed the connection]
ilikeheaps has quit [Ping timeout: 272 seconds]
lawlesseel has quit [Remote host closed the connection]
lawlesseel has joined #nixos
Ariakenom has joined #nixos
ryantm has quit [Quit: Lost terminal]
endformationage has quit [Quit: WeeChat 1.9.1]
<fresheyeball>
so I had an interesting conversation on the postgresql channel
<fresheyeball>
and they brought this up in response to the way nix packages the posgresql service
<ixxie>
I'm actually trying to fix issues with home-manager
<ixxie>
so I want to remove it and reinstall it (the installation method seems to have changed
tzemanovic has joined #nixos
b has quit [Ping timeout: 245 seconds]
<etu>
Oh, I haven't used overlays for the user like that
kiloreux has quit [Ping timeout: 240 seconds]
<ixxie>
I removed the overlay but don't know how to rebuild the user environment
<etu>
I guess it's safe to remove the directory. And if something reference it at some point it will yell at you that it's missing and you'll probably get to know where.
tzemanovic has quit [Remote host closed the connection]
<ixxie>
etu: but how to I trigger a rebuild to clean up the user env and make a new generation?
<ixxie>
aha
<ixxie>
nix-env --upgrade
<ixxie>
very unintuitive
sbdchd has joined #nixos
kiloreux has joined #nixos
sbdchd has quit [Ping timeout: 246 seconds]
orivej has quit [Ping timeout: 252 seconds]
knupfer has joined #nixos
timon37 has quit [Remote host closed the connection]
tzemanovic has joined #nixos
jasongrossman has joined #nixos
<funkmybios>
running my first nixos install.. 18.3 did not do networking for me so using unstable 19.03 from USB unetbootin ... so far only complains about Bad initial heap size ...
<jasongrossman>
funkmybios: \o/
abueide has quit [Ping timeout: 240 seconds]
tzemanovic has quit [Remote host closed the connection]
tzemanovic has joined #nixos
tzemanovic has quit [Remote host closed the connection]
ryanartecona has quit [Quit: ryanartecona]
tzemanovic has joined #nixos
tzemanovic has quit [Remote host closed the connection]
tzemanovic has joined #nixos
tzemanovic has quit [Remote host closed the connection]
tzemanovic has joined #nixos
mounty has joined #nixos
tzemanovic has quit [Remote host closed the connection]
tzemanovic has joined #nixos
<{^_^}>
[nixpkgs] @mbode opened pull request #46412 → pythonPackages.CDDB: fix darwin build → https://git.io/fA2HN
tzemanovic has quit [Remote host closed the connection]
<symphorien>
xok: It stores files in the store, so likely on both machines
<srhb>
xok: With NixOps, everything is built on your machine (or whatever the builder is) and then transferred to the target
<typetetris>
Oh, the channel #nix is invite only?
<srhb>
typetetris: I'm not sure that channel has anything to do with us tbh.
<ixxie>
what is the type for a function?
<ixxie>
can't seem to find it in types.nix
<srhb>
ixxie: It doesn't have one.
<typetetris>
Is there a reason, why the nix multi user install script insists on running everything through sudo? Nearly everything it does needs root privileges, why not just run as root directly?
<ixxie>
srhb: so what do I specify?
<srhb>
ixxie: Where are you trying to use a function?
<{^_^}>
[nixpkgs] @xeji pushed commit from @mbode to master « pythonPackages.tifffile: fix python 2 build (#46415) »: https://git.io/fA2dr
<srhb>
ixxie: why not just pass (_: extraPythonPackages) then?
<srhb>
ixxie: (eg. leave the list type as is, just change the call)
<{^_^}>
[nixpkgs] @xeji pushed commit from @mbode to release-18.09 « pythonPackages.tifffile: fix python 2 build (#46415) »: https://git.io/fA2dK
<ixxie>
srhb: error: The option value `programs.neovim.extraPythonPackages' in `/home/ixxie/nixdev/home-manager/modules/programs/neovim.nix' is not of type `list of packages'.
<ixxie>
because I get this :P
<{^_^}>
[nixpkgs] @etu opened pull request #46416 → lidarr: init at 0.3.1.471 → https://git.io/fA2dX
<srhb>
ixxie: When you do what exactly?
sbdchd has quit [Ping timeout: 244 seconds]
<ixxie>
well the problem is the defaults in the hm module are wrong
<ixxie>
so the hm module isn't updated to match the neovim modules
<srhb>
ixxie: Not if you change the call to neovim.override
lawlesseel has quit [Remote host closed the connection]
<srhb>
To be clear, what I am suggesting is neovim.override { ...; extraPythonPackages = const extraPythonPackages; }
__lawlesseel__ has joined #nixos
<ixxie>
srhb: aah I see
<ixxie>
thanks
<ixxie>
that should do the trick
<srhb>
It looks to me like the neovim function should still be compatible with using a list though...
* srhb
shrugs
<ixxie>
well it complains about type when I tried it
<srhb>
Odd..
<ixxie>
(note in my hm config I merely enabled neovim... not even passing anything)
goibhniu has joined #nixos
<typetetris>
If I install nix in multi user mode, which command did install the systemd unit for nix-daemon?
<tilpner>
Does anyone know how to change the target to musl with carnix?
sbdchd has joined #nixos
silver has joined #nixos
philippD has joined #nixos
sbdchd has quit [Ping timeout: 252 seconds]
Mrmaxmeier has joined #nixos
Ariakenom has quit [Quit: Leaving]
joehh has joined #nixos
<joehh>
I'm struggling to use nix-prefetch-git - I've been using nix-prefetch-hg for ages and no problems
<joehh>
when I try nix-prefetch-git ./ I get the following error:
<joehh>
fatal: Couldn't find remote ref HEAD
<joehh>
fatal: The remote end hung up unexpectedly
<joehh>
warning: empty strings as pathspecs will be made invalid in upcoming releases. please use . instead if you meant to match all paths
<joehh>
fatal: '' is not a commit and a branch 'fetchgit' cannot be created from it
<joehh>
Unable to checkout HEAD from ./.
<joehh>
Is it something obvious I'm missing (my git knowledge is at a pretty basic level)
<joehh>
?
<sphalerite>
joehh: (disclaimer: I have basically no idea what I'm doing with it either) does nix-prefetch-git $PWD work?
<sphalerite>
Is there a way to "detach" an input device from the mouse pointer while still having it available as an X input device that sends events?
spear2 has joined #nixos
<joehh>
sphalerite: It does - I don't think I understand why, but it appears to
smolboye has quit [Ping timeout: 245 seconds]
<joehh>
is the issue that the git pull/clone is run by the nix-daemon and so needs a full path
<sphalerite>
joehh: I suspect that the nix-prefetch-git script changes to a temporary directory, making . no longer refer to the right directory
<sphalerite>
arguably a bug
srghma has joined #nixos
<srghma>
hi, guys, I can't pass `impureEnvVars` to builder, nor from current shell, nor from `/nix/configuration.nix` `nix.envVars`, can someone help me https://github.com/NixOS/nixpkgs/issues/45404
<{^_^}>
#45404 (by srghma, 2 weeks ago, open): nix-build --option builders '' doesn't build with impureEnvVars from current shell
<avn>
sphalerite: yes, it `pushd $(mktemp -d ...)` (just checked)
polyzen is now known as dmc
srghma has quit [Ping timeout: 252 seconds]
srghma has joined #nixos
<aanderse>
if i have multiple configuration files (imports) and i want each to *append* values to a list property, not *set* the value, what would be the best way to do that?
<sphalerite>
aanderse: just set them in the configuration files, default behaviour is to concat lists
<sphalerite>
aanderse: if you want to affect the order, use mkBefore/mkAfter
<sphalerite>
or if you want to overwrite after all, use mkForce
<sphalerite>
srhb: does running nixos-install from a recent installer not count as a "good upgrade path" for you?
<srhb>
sphalerite: Yes, it's fine with me.
<srhb>
I guess my comment was unclear. :/
<srghma>
re my issue with `impureEnvVars` - I want to implement option to fetch from private repo for fetchFromGitlab https://github.com/NixOS/nixpkgs/issues/43583, but stuck with stupid issue - cant pass `impureEnvVars` from current shell
<{^_^}>
[nixpkgs] @xeji pushed commit from @r-ryantm to master « nextcloud: 13.0.5 -> 13.0.6 (#46236) »: https://git.io/fA2AG
<srhb>
sphalerite: I think the best general recommendation is to walk through every version though and just nixos-rebuild (considering installs where getting an installer running might be problematic)
<sphalerite>
srhb: what's such an install?
<srhb>
sphalerite: Any remote install where our usual guide to getting an installer running (eg. usb key) doesn't work. Otherwise we should have a very good guide for spinning up an installer on an existing system and recommend that.
<avn>
srghma: other option -- deploying key to access github. (odd, it visible in /nix/store to everyone)
stanibanani has joined #nixos
<sphalerite>
srhb: yep definitely in favour of getting a wider range of installer options better documented!
stanibanani has left #nixos [#nixos]
<srhb>
:)
<sphalerite>
I mean, you can probably get an installer running by unpacking a netboot initrd into a tmpfs and switch_rooting into it :D
<sphalerite>
... damn. I need to try this.
<srhb>
sphalerite: Yep, I think you absolutely can do this. :) I just want to make sure we provide a good suggestion, and right now I think the least mental overhead is simply saying "if you're upgrading from an unsupported release, always upgrade to release+1 at a time"
iyzsong has quit [Read error: Connection reset by peer]
<srghma>
avn: yes, I dont want my key to be visible in /nix/store to everyone
<symphorien>
fixed-output derivation produced path '/nix/store/2...hp-ripgrep-0.10.0-vendor' with sha256 hash '1r4...s' instead of the expected hash '1k...gr'
<symphorien>
how can I compare the two ?
<symphorien>
with diffoscope for example
<sphalerite>
symphorien: you can't unlessy ou have a copy with the right hash
<avn>
srghma: then only true solution is implement proxying of ssh agent in nix.
<symphorien>
ok
<avn>
or use intermediate tarballs
<sphalerite>
srghma: I think you may actually want to use builtins.fetchgit
Dedalo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<avn>
sphalerite: builtins.fetchgit use non-sandboxed git?
<das_j>
Umm, as I have no idea how Haskell and Haskell packaging works, I also have no idea what jailbreaking does and how dependencies are lowered :D
<{^_^}>
[nixpkgs] @Mic92 pushed commit from @romildo to release-18.09 « xsettingsd: git-2015-06-14 -> 1.0.0 »: https://git.io/fA2xx
<srhb>
das_j: Jailbreaking simply removes all version constraints
<srhb>
Meaning, it will help where the constraints were overly conservative.
<das_j>
Ah, that makes sense. So do I jailbreak hoauth2 or gitit?
<srhb>
lowering dependencies is like any other package, pass that version explicitly as an argument to override.
<srhb>
das_j: Both
<srhb>
But like I said, it's not quite sufficient in this case.
tzemanovic has quit []
<das_j>
Why do I need to lower that version?
<srhb>
Because gitit depends on pandoc, but a much much older pandoc, and the current version does not have the right exports for gitit
<das_j>
Oh, okay. I thought it would just execute the pandoc binary, but adding it as a library makes more sense now
<srhb>
Yeah, it uses it as a library.
<srhb>
das_j: One of the errors is this: src/Network/Gitit/Export.hs:44:33: error: Module ‘Text.Pandoc.Writers.RTF’ does not export ‘writeRTFWithEmbeddedImages’
<bobvanderlinden>
I'm trying to resolve the failing tests on Hydra in preparation of 18.09, but I'm running into a test that seems to be failing randomly. It's about the bittorrent test: https://hydra.nixos.org/build/81000569
ilikeheaps has quit [Read error: Connection reset by peer]
<bobvanderlinden>
What should be done with randomly failing tests? It seemed like the clients weren't able to find eachother in time and thus the file was not downloaded and thus the test would fail. If the timeout was set higher it might've helped, but I'm not sure.
rct has joined #nixos
ixxie has quit [Ping timeout: 246 seconds]
endformationage has joined #nixos
sbdchd has joined #nixos
<das_j>
How can I set ghc options?
sbdchd has quit [Ping timeout: 245 seconds]
rct has quit [Quit: WeeChat 1.9.1]
slyfox has quit [Quit: :)]
<das_j>
Essentially, I need cabal install --reinstall gitit --ghc-options="-rtsopts"
vmandela_ has joined #nixos
vmandela has quit [Ping timeout: 240 seconds]
Acou_Bass has joined #nixos
jtojnar has quit [Read error: Connection reset by peer]
Acou_Bass has quit [Client Quit]
slyfox has joined #nixos
jtojnar has joined #nixos
Acou_Bass has joined #nixos
rrr has quit [Quit: WeeChat 2.1]
vmandela_ has quit [Client Quit]
siriobalmelli has joined #nixos
<srhb>
bobvanderlinden: Making them less non-deterministic is always appreciated
<srhb>
bobvanderlinden: It depends a lot on the test how to do that. Oftentimes it's about waiting at strategic points for certain conditions to be met before proceeding.
<siriobalmelli>
hi guys, I would love any pointers to documnetation: trying to undestand why certain derivation (attribute?) names which are valid input for `nix-env` are not valid input for `nix-env -A`
<siriobalmelli>
I'm all for using -A ... but where do I go query what attributes are legal for -A ?
<symphorien>
for nix expressions you always use attribute names
<symphorien>
-A is for Attribute
<symphorien>
always use attributes except for nix-env -e
<{^_^}>
[nixpkgs] @Mic92 pushed commit from @dasJ to release-18.09 « treewide: Use http/https for homepages »: https://git.io/fAaeU
<symphorien>
how did you come up with this name "python3.6-jsonschema" ?
<siriobalmelli>
to give context: I am writing shell.nix files that encapsulate project environments
orivej has quit [Ping timeout: 252 seconds]
<siriobalmelli>
currently, to find available packages I will do a `nix-env -qa` and it gives me "python3.6-jsonschema"
<symphorien>
use nix-env -qaP instead
<symphorien>
and the attribute name will be displayed
<symphorien>
or nix search
<siriobalmelli>
@symphorien GENIUS
<siriobalmelli>
this is precisely the key I was searching for :)
<siriobalmelli>
all making sense now
<symphorien>
:)
<siriobalmelli>
very much appreciated :))
<ldlework>
,locate xsel
<{^_^}>
Couldn't find any packages
<ldlework>
what what
<bobvanderlinden>
srhb: symphorien thanks for the pointers. I found some commits that are indeed good examples of 'waiting at stagegic points'. I'll give it a go
davidar_ has quit [Quit: Connection closed for inactivity]
spear2 has quit [Remote host closed the connection]
<infinisil>
It passes allowUnfree, allowBroken and allowUnsupporteSystem all true now
<sphalerite>
not sure allowBroken is desirable..?
lopsided98 has quit [Quit: Disconnected]
sbdchd has quit [Ping timeout: 272 seconds]
lopsided98 has joined #nixos
Ariakenom_ has quit [Quit: Leaving]
palo has joined #nixos
palo has joined #nixos
<infinisil>
sphalerite: Tbh, I should just abandon the nix-instantiate, I'm currently only using it to find the packages name, which could very well be derived from the nix-locate output
<infinisil>
Huh but wait a minute, why is kibana in the hydra caches when it's unfree?
<sphalerite>
hm, I'm confused. I ran `nix build -I nixpkgs=channel:nixos-18.03 -f '<nixpkgs/nixos/release.nix>' iso_minimal.x86_64-linux` and it's building stuff locally, shouldn't everything be in the cache?
<infinisil>
I don't think it should compile a lot, but some stuff won't be in the cache, like the iso filesystem itself and some other things that are just faster to create locally I think
MinceR has quit [Ping timeout: 240 seconds]
<Dezgeg>
there will probably be some very minor differences like the channel content not exactly being the same as the git checkout
bobvanderlinden has joined #nixos
<sphalerite>
Dezgeg: blarg, that's a pain
MinceR has joined #nixos
<sphalerite>
infinisil: it needs to build the squashfs which is CPU-intensive
<sphalerite>
there should be *a* minimal iso squashfs on cache.nixos.org since the iso is built by hydra too, but apparently there are minor differences :(
<sphalerite>
aaah probably the version number. It uses the dummy version 18.03pre56789.gfedcba
jbo2 has joined #nixos
<jbo2>
Hi there - I've an unbootable nixos machine - all of the configurations in grub don't work. I'd like to chroot into it to run nixos-rebuild from an ubuntu live-usb, but don't have a shell that I can run. I've tried https://nixos.org/nix-dev/2014-December/015253.html but /run/current-system/bin/bash doesn't exist. What should I do?
<wajsel>
Hi Does anyone have experience of docker daemon?
orivej has joined #nixos
<wajsel>
I'm trying to install it according to the wikipage https://nixos.wiki/wiki/Docker so that the server is started by systemd
<clever>
wajsel: services.docker.enable = true; and its done
<wajsel>
unfortunately systemd does not recognize the service docker and my user does not become a member of docker group
<wajsel>
aha I did not try that yet....
dmc has joined #nixos
<wajsel>
@clever sorry but that did not work, error: The option `services.docker' defined in `/etc/nixos/configuration.nix' does not exist. (use '--show-trace' to show detailed location information) building Nix...
<clever>
wajsel: oops, its virtualisation.docker.enable
<wajsel>
yes, that's what I've been trying but with no luck :(
<wajsel>
[wajsel@xps:~]$ sudo nixos-rebuild switch -I nixpkgs=/etc/nixos/nixpkgs/ building Nix... building the system configuration... activating the configuration... setting up /etc... setting up tmpfiles
<wajsel>
should it not complain about invalid options?
<clever>
it should complain
<wajsel>
sorry for the indentation in the chat
<wajsel>
I cloned the nixpkgs repo and use the -I flag
<{^_^}>
[nixpkgs] @edwtjo pushed to master « nixos/i2pd: Update options to encompass recent additions to the daemon »: https://git.io/fAaTq
funkmybios has joined #nixos
<wajsel>
it has been working with other rebuilds so I don't think that's the problem
<clever>
which revision of nixpkgs are you on? and why did yoy need to clone it?
<clever>
and what does `sudo systemctl status docker.service` say?
<wajsel>
``` [wajsel@xps:~]$ sudo nixos-rebuild switch -I nixpkgs=/etc/nixos/nixpkgs/ building Nix... building the system configuration... activating the configuration... setting up /etc... setting up tmpfiles [wajsel@xps:~]$ [wajsel@xps:~]$ groups users wheel disk audio video networkmanager systemd-journal [wajsel@xps:~]$ sudo systemctl status docker.service Unit docker.service could not be found.```
humanoyd has joined #nixos
<clever>
wajsel: can you pastebin the entire configuration.nix file?
georges-duperon has joined #nixos
<wajsel>
I cloned the repo because I needed to do some changes due to elixir and erlang otp21
<wajsel>
It seems that the webchat does not handle newline so it may be messy message pasting the whole config ....
<clever>
i dont see the startOnBoot you mentioned earlier
<clever>
wajsel: which revision of nixpkgs are you on?
<averell>
how do i generate a static file in /var/db/ in my configuration.nix? dhcpcd moved a file from etc there, and previously i just used environment.etc."file" = "bla"
<wajsel>
yes I tried again with the startOnBoot removed
sbdchd has joined #nixos
<clever>
averell: i would just have a cp or ln command in the prestart, or find a config flag to tell it where to look
<wajsel>
I'm branching from channels/nixos-18.03 8b92a4e600458c01ab0a72f2492eb7120e18f9bc
<clever>
wajsel: the option is called enableOnBoot in that version, and it defaults to true
sbdchd has quit [Ping timeout: 240 seconds]
<wajsel>
I checked out the channel/18.03 (fetched yesterday) and it did not provide any docker service
<sphalerite>
averell: you could put it next to configuration.nix and put cp ${./somefile} /var/db
sigmundv has joined #nixos
<wajsel>
such a silly mistake of me :) Thanks sphalerite and clever. The service is starting now
<sphalerite>
wajsel: also a silly bug :/
<sphalerite>
normally trying to set a nonexistent option would cause an error, there's just that exception
<wajsel>
I usually do a rebuild dry-build to check the configuration but this was silent.... Should I open an issue, and if so where? on the nixpkgs repo or nix itself?
periklis has quit [Quit: ERC (IRC client for Emacs 26.1)]
<clever>
that also explains why startOnBoot didnt fail
Ariakenom_ has joined #nixos
<clever>
sphalerite: i also just had an idea, what about a global option to just turn mkRenamedOptionModule into a no-op?
<clever>
sphalerite: both to avoid bugs it makes, and to force me to fix all renames
<sphalerite>
clever: do it! :p
<wajsel>
I got one more question, I was expecting my user to be in the docker group by adding it to the users.extraUsers.wajsel.extraGroups then starting a new shell and check the `groups` command
<clever>
wajsel: you have to logout and login again for any group changes to take effect on linux
<sphalerite>
wajsel: you need to do nixos-rebuild switch, and log out and back in
<wajsel>
ok I thought it was enough to open a new shell
<clever>
it inherits the groups from the parent process
<clever>
and the new shell is just your window manager, which already has the "wrong" groups
<clever>
it must be spawned by something root, like sshd, or the login screen
<wajsel>
ok, checked with ctrl-alt-f1 login. It's working :)
<clever>
which then sets the right groups as it drops root
<sphalerite>
clever: hmmm sudo -iu $USER might work too right?
<clever>
sphalerite: sudo is smart enough to know no changes take place, and do nothing
<sphalerite>
damn
<sphalerite>
sudo su - $USER then
<clever>
sphalerite: you need to `exec sudo -i` then `exec sudo -i -u clever`
<clever>
but, ive found that the above breaks pulseaudio
<sphalerite>
lol
<clever>
i was trying to go beyond the hard ulimits
<clever>
but then i lost all audio in chrome
<clever>
sphalerite: `ssh localhost` would probably be simpler, lol
<srhb>
xok: I never tried using anything but apache + mod_php for it..
<srhb>
Ah, that was actually merged in a different PR later
<srhb>
But the default wasn't changed.
<srhb>
So you can just use that :)
<xok>
srhb: any chance to use mysql instead of psql?..
<srhb>
Not without rewriting the module, it's not configurable.
jedahan has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<xok>
srhb: why did you give up upgrading it?..
<xok>
I heard of NixOS couple of days ago..
<xok>
have never played with it before...
<xok>
I can't do such things just yet... :D
<{^_^}>
[nixpkgs] @knedlsepp opened pull request #46436 → pythonPackages.cozy: Disable on python2 → https://git.io/fAaIK
<srhb>
xok: I just didn't care about zabbix anymore. :)
<srhb>
xok: But using that package, you get a working 3.4 at least.
<xok>
srhb: I am pretty new here, can you help me get it working?..
<xok>
I don't know how to do that...
<srhb>
xok: Have you used an overlay before?
<xok>
no...
<xok>
I head about NixOS couple of days ago, no more...
<xok>
heard*
<srhb>
No worries, I'm just trying to gauge how to start explaining.
<srhb>
So, the core of the problem is this
<srhb>
> nixpkgs.zabbix.server.name
<{^_^}>
undefined variable 'nixpkgs' at (string):192:1
<srhb>
zabbix.server.name
<srhb>
> zabbix.server.name
<{^_^}>
"zabbix-1.8.22"
nbardiuk has quit [Remote host closed the connection]
<srhb>
The default zabbix attribute is 1.8.22
<shiver>
neat bot
<xok>
yes, very old..
<srhb>
Hmm, maybe the server component wasn't actually upgraded
knupfer has quit [Remote host closed the connection]
<srhb>
I was thinking to just set zabbix = zabbix34, but that won't work then.
<xok>
I spent two or three days to get the damned web UI displayed...
<xok>
I finally got it but it was 1.8...
<xok>
terribly old one...
<srhb>
Yeah. Either we should remove it or find someone who cares to maintain it.
<xok>
I started to do a "docker" instead...
<xok>
to start a docker container with zabbix on NixOS...
<xok>
but that's not what we really want...
<xok>
it's just a workaround...
<srhb>
Well, my PR should still work, you can use that locally..
<{^_^}>
[nixpkgs] @matthewbauer merged pull request #43470 → dwarf-fortress: Refactor, round 2: let the user pick which version of the lazy pack to install → https://git.io/fNqeY
<{^_^}>
[nixpkgs] @matthewbauer pushed 9 commits to master: https://git.io/fAaIx
<{^_^}>
[nixpkgs] @matthewbauer pushed 2 commits to master: https://git.io/fAaIj
<srhb>
xok: I would suggest by starting with reading the overlays section in the nixpkgs manual. Once you know how to create overlays, it becomes much easier to make ad-hoc upgrades to packages.
<xok>
I need to get this thing done by tomorrow...
<sphalerite>
xok: if you need to get it done by tomorrow and don't know what you're doing with nixos, I suggest sticking with a traditional distro for now.
<sphalerite>
xok: or hiring someone to do it for you :)
<xok>
sphalerite: I've got great progress, I can do it with docker...
<xok>
but would be awesome if I can stick with nixops...
orivej has quit [Ping timeout: 252 seconds]
<avn>
disasm: nice example of nixops btw, I should switch my home machines to nixops as well
<disasm>
xok: I can help you with getting a simple PoC setup for handling custom modules and overlays
<xok>
disasm: thatnk you but I think that's a hard part for me at the moment...
<disasm>
avn: my talk at NixCon is all about the structure in that repo, although I'm planning a nice cleanup overhaul now that I have CI testing for it to make sure I don't break stuff in the process.
<avn>
disasm: I unable to attend this year, but I hope I will able watch all talks recordings
jedahan has joined #nixos
<disasm>
avn: I just got hydra building my 2 nixops deployments and my non-nixops laptop this past week, as well as a nix-darwin deployment: https://github.com/disassembler/network/blob/master/release.nix -- Feels so good to know things are going to break before I run nixops deploy and having to rollback, lol.
<avn>
disasm: I just curious, if it possible to have `nixops` deploy, which can be switched to "non-nixops" in single command (for example I usually deploy my travel notebook from main build machine, but sometimes I need add something when I am travelling)
<xok>
damn, how I want this thing to be working in my environment... ;-(
<philippD>
I have seen the nix function `callCabal2nix` pop up at some places but I can't find it's definition anywhere in nixpkgs. I currently manually call cabal2nix each time I change the cabal file. Does this function take care of this update automatically and if so how do I get it?
<srhb>
philippD: It's in make-package-set.nix in the haskell modules
<srhb>
And yes, it does that.
<srhb>
philippD: You should simply use it from haskellPackages, eg. nixpkgs.haskellPackages.callCabal2nix
<srhb>
Which uses the scope of the surrounding package set like callPackage would.
jedahan has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<{^_^}>
[nixpkgs] @knedlsepp opened pull request #46439 → pythonPackages.joblib: Disable flaky test → https://git.io/fAaLx
wpcarro has quit [Ping timeout: 250 seconds]
<philippD>
srhb: do you have an example of a project using callCabal2nix?
<clever>
callCabal2nix "name" ./. {}; i think was the syntax
<clever>
it will then expect a name.cabal in the dir, and load it as normal
<srhb>
clever: Yes, that's correct
<philippD>
clever: I should stop doubting the function definitions and just go the easiest route and just call the damn function.
<srhb>
name src and arguments to the mkderivation
<srhb>
(Well, the usual override stuff)
[Leary] has quit [Read error: Connection reset by peer]
<clever>
srhb: callPackage/override args, not mkDerivation args
<philippD>
cool you can even specify the concrete ghc version.
<srhb>
clever: Sorry, right :)
<srhb>
philippD: The only rule is that name must be the name in "${name}.cabal" in src :)
<rfold>
When I boot my computer I get a light blue screen with the NixOS logo and a password prompt. After logging in it opens my window manager. What keyword am I looking for if I want to customize this login screen? Is it part of X server, display manager, or something else?
trevthedev has joined #nixos
fragamus has joined #nixos
jedahan has joined #nixos
<tilpner>
That's slim, and it's your display manager
<rfold>
Thanks!
knupfer has joined #nixos
<{^_^}>
[nixpkgs] @vcunat pushed to master « racket: warn to avoid repeating #45952 »: https://git.io/fAaqW
<rfold>
I can customize it with services.xserver.displayManager.slim.theme
<siers>
How do I run a script and get its exit value? Only through mkDerivation?
<tilpner>
You may want to switch away from slim entirely, it's been abandoned for 5 years. I like sddm
<disasm>
tilpner: I've been wondering if we're ever going to change the default :)
<tilpner>
siers - You will need a derivation (or builtins.exec, but don't do that), but you can use the wrapper runCommand
<clever>
disasm: i still use slim on my desktop
<{^_^}>
[nixpkgs] @bjornfor pushed commit from R. RyanTM to master « osinfo-db: 20180531 -> 20180903 »: https://git.io/fAaqg
<Ralith>
I think the author got eaten by real life a while back
<bkchr[m]>
Hmm okay. So we need a different way to package the rust stuff in Nixos.. I need a newer version of cargo-vendor, but that requires carnix. I'm trapped :D
jbo2 has quit [Quit: Lost terminal]
jonreeve has quit [Quit: WeeChat 2.1]
<Ralith>
or you could fix whatever the problem is
<bkchr[m]>
Yeah, I could do that. But I'm already to deep in the rabbid hole..^^
<symphorien>
bkchr[m]: why do you need a newer version of cargo vendor ? I am curious
<{^_^}>
#30742 (by ljli, 45 weeks ago, open): buildRustPackage: found duplicate version of package
<{^_^}>
[nixpkgs] @bobvanderlinden opened pull request #46443 → [WIP] add test for upnp using miniupnpd and miniupnpc → https://git.io/fAaO2
<symphorien>
ah ok.
<symphorien>
you can probably use buildRustPackage to build cargo-vendor
<symphorien>
just leave the one currently in nixpkgs alone, for boostrapping
<bkchr[m]>
symphorien: hmm, could work
<sphalerite>
tnks: the default linuxPackages is always a version that's still maintained and should be a sensible choice to run unless you have hardware that only a newer kernel supports.
<das_j>
Um, I have found the ARM images online, but how do I build them myself?
<tnks>
sphalerite: there was an era when I liked to keep current because power management always seemed to suck, and there was hope that the later kernel would improve it.
<tnks>
but now, I probably keep current out of habit than need.
<sphalerite>
tnks: also nothing wrong with _latest though either :)
<sphalerite>
das_j: 1. get an arm system with nix running
<sphalerite>
you may run into some memory constraints, although that should improve with nix 2.1
<srhb>
das_j: Same problems.
<clever>
das_j: the OOM issues when copying closures have been improved in 2.1
<{^_^}>
[nixpkgs] @xeji pushed commit from @knedlsepp to release-18.09 « pythonPackages.joblib: Disable flaky test (#46439) »: https://git.io/fAaOb
<das_j>
Ah, that's awesome
<das_j>
Is 2.1 in 18.09? Or do I have to switch to nix unstable?
<sphalerite>
I find zram to be quite helpful for that on my 2GB ARM device (not a pi but yeah)
<sphalerite>
I think it's in 18.09
<das_j>
sphalerite: How does zram affect the performance?
<clever>
sphalerite: only problem ive had with zram, is that `nixos-rebuild switch` sometimes wants to turn the zram off and back on, to reload it
<sphalerite>
note though that 18.09 is not yet released and may murder your cat
<clever>
which involves 30mins of decompressing ram, lol
<sphalerite>
clever: I typically just set it up imperatively under nixos's nose
<clever>
and if it doesnt fit in real ram, it then has bigger problems
<clever>
nixos-generate-config also doesnt understand the ramdisks it uses
<sphalerite>
das_j: better than normal swap :p
<emily>
"nixos-rebuild switch" toggling stuff off and on is really annoying
<das_j>
sphalerite: Both my cats (as well as my 18.09 installations) are still alive so it doesn't look too bad
<emily>
I feel like the activation scripts could be better about this but maybe it'd require too deep hooks :(
<das_j>
sphalerite: Thanks, I might give it a try for my low-memory systems
bobvanderlinden has quit [Ping timeout: 252 seconds]
<sphalerite>
das_j: :) just saying it mostly to avoid giving the impression that it's released (since it's september 2018 and the channel does exist now)
<emily>
even if NixOS murders your cat, it'll reconstruct it again in exactly the same configuration immediately afterwards, so it's not too big a deal.
<das_j>
sphalerite: I'm aware of that. But coming from Arch Linux, my feeling of completeness and stability is lacking
<emily>
but it involves 30 minutes of decompressing cat.
<sphalerite>
emily: :D
<das_j>
emily: I can't wait that long to pet a cat
<sphalerite>
das_j: fair enough
<joepie91>
!hash
<emily>
das_j: :D
<joepie91>
meh
<joepie91>
how do I produce a hash from a source tarball again (not from a URL), so that it matches whatever goes in the expression?
<joepie91>
notably, without adding it to the store
<clever>
joepie91: nix-hash
<clever>
--flat i think
<das_j>
Ooh, that's cool. I ususally use a random hash and let it fail
<clever>
joepie91: oh, but if its flag, you can also just use sha256sum i believe
<clever>
flat*
<joepie91>
that produces the wrong encoding I think?
<sphalerite>
das_j: it's fiddlier than you'd think
<clever>
joepie91: nix accepts both base16 and base32
<sphalerite>
joepie91: nope nix supports base16 for it now
<joepie91>
clever: seems `nix-hash --flat --type sha256 --base32` produces the format used in nixpkgs
<clever>
yep
<clever>
when using flat hashing, then its just the raw hash of the file, so you can use sha256sum
<joepie91>
yep, seems to produce the same output
<joepie91>
alright, thanks
<clever>
`nix-hash --type sha256 --to-base32 <hash>` can be used to convert 16 to 32
<clever>
for recursive hashing, you need the hash of the NAR encoding of the directory
<clever>
nix-hash without --flat can do that
<clever>
nix-prefetch-url --unpack can also do it
<clever>
you can also `nix-store --dump | sha256sum`
<sphalerite>
or nix-store --dump | sha256sum
<sphalerite>
dammit clever stop ninja'ing me! :p
<clever>
lol
<clever>
and i had to lookup the exact flag in the man pages!
* joepie91
continues on his all-the-WINE-versions project
<joepie91>
gotta love playonlinux for providing easily-parseable metadata about all WINE releases...
<das_j>
Btw, are people happy with discourse?
<das_j>
Or is anyone else experiencing completely blank pages unless you activate javascript?
<elvishjerricco>
das_j: I doubt discourse is designed to work without javascript
<elvishjerricco>
Personally, I think it's a huge improvement over the mailing list
<das_j>
elvishjerricco: I expect that but I hate allowing third-party hosts
rain1 has joined #nixos
<das_j>
And yes, it's better than the mailing list but that still bugs me
<elvishjerricco>
Just the nature of the web unfortunately :/
<elvishjerricco>
But I think you can trust code from nixos.org
<elvishjerricco>
If not, then your whole NixOS configuration should be considered untrusted :P
<das_j>
elvishjerricco: So I don't dislike discourse, I dislike the web ;)
<das_j>
elvishjerricco: It's not about the NixOS servers, it's about discourse-cdn-sjc2.com
<elvishjerricco>
das_j: Is it possible to allow JS only from nixos.org on the page?
<elvishjerricco>
Oh. Lame
<das_j>
elvishjerricco: No, the page is just blank
<joepie91>
I believe the NixOS forums are on hosted Discourse
<elvishjerricco>
nixos.org should host all its own discourse assets
<das_j>
Not sure that's possible
<joepie91>
as in, not on infrastructure operated by NixOS people in the first place
<elvishjerricco>
Well that's unfortunate...
<emily>
I think you can use discourse via mail?
<das_j>
Well, I'll probably bite into the sour apple (is that even an english term) and allow the host for discourse.nixos.org
<joepie91>
das_j: fellow dutchman?
<joepie91>
:P
<das_j>
joepie91: german
<joepie91>
oh, huh, didn't know that that saying existed in Germany as well
<joepie91>
to be fair, the Dutch version is to bite *through* the sour apple
<das_j>
I didn't know about the saying in dutch either
<das_j>
english seems to be "to grin and bear it" or "to swallow the pill"
<Myrl-saki>
[myrl@myrl-lappy:/etc/nixos]$ ag currentTime
<Myrl-saki>
nixpkgs/pkgs/tools/nix/info/multiuser.nix:7: # no cache: ${toString builtins.currentTime}
<Myrl-saki>
Glad to know that I'm not the only one who thought of using currentTime to force builds.
<sphalerite>
oh wow that's nasty x)
<Myrl-saki>
:D
<clever>
:D
<srhb>
Myrl-saki: I think we even have a wrapper for that...
<clever>
Myrl-saki: oh, nix also has a dedicated option for this
<samueldr>
not nasty there, it's to query fresh information for nix-info :)
<sphalerite>
samueldr: I suppooooose, but still
<clever>
i cant remember the name of it though
<samueldr>
sphalerite: that file is not used inside nixpkgs, but is a file used by the info script
<xok>
when we provision a new system, is there a chance to get the IP of the newly created machine?..
<clever>
rsynnest: that gets the .drv for the current-system
<xok>
and if yes, what's the command for it?..
sir_guy_carleton has joined #nixos
<bigvalen>
Where does nixos store the linux console resolution ? Seems it's running at 1024x786 after boot, despite me setting a text-mode in Grub
<clever>
rsynnest: and `nix-instantiate '<nixpkgs/nixos>' -A system` gets you the drv for the current configuration.nix
<clever>
rsynnest: then nix-diff the 2 of them, edit configuration.nix to solve the difference, and repeat until it has no diff
<{^_^}>
[nixpkgs] @ryantm opened pull request #46445 → rPackages.RPostgres: fix configure paths and includes → https://git.io/fAash
xok has quit [Ping timeout: 252 seconds]
<rsynnest>
clever: thanks! I'll have to do a lot more reading/practicing before I can reverse engineer derivations, but this definitely points me in the right direction of what I broke in my new configuration
fragamus has joined #nixos
simukis has quit [Ping timeout: 245 seconds]
zemm has quit [Ping timeout: 240 seconds]
jtojnar has quit [Read error: Connection reset by peer]
georges-duperon has quit [Ping timeout: 252 seconds]
Ariakenom_ has quit [Quit: Leaving]
<bigvalen>
I tried to move my one-disk system to a four-disk ZFS system. Apparently, you can make a raidz2 array with three disks, and one sparsefile disk that you 'offline' before you populate. Then you can detach that sparsefile one the data is moved over, and yay! you have a four-disk raidz2. But when I rebooted, it failed to import my new pool...because one of the disks was offline.
<bigvalen>
Of course it's offline. It was never really working :(
<samueldr>
(which is summarised in CONTRIBUTING.md)
<__monty__>
Do I need to add myself to maintainers somewhere?
<__monty__>
And how do I get the proper hash for fethfromgithub? Is it just the commit hash?
<samueldr>
you can if you want to maintain, and should if you *add* a package, but not for a version bump
<samueldr>
rev = ""; is a commitish that git can handle (use a tag or a revision ID)
<samueldr>
sha256 = ""; use the good ol' tofu trick
<samueldr>
,tofu
<{^_^}>
To get a sha256 hash of a new source, you can use the Trust On First Use model: use probably-wrong hash (for example: 0000000000000000000000000000000000000000000000000000) then replace it with the correct hash Nix expected.
<__monty__>
Yes but how do I run the default.nix to get the proper hash?
Edes has joined #nixos
<clever>
__monty__: in the root dir of nixpkgs, `nix-build -A packagename`
<haslersn>
Is there a way to use fetchurl or fetchGit without specifying sha256 (such that it can always take the master branch)?
<elvishjerricco>
haslersn: `builtins.fetchGit "ssh://git@github.com/owner/repo"` will just get `master`, and it'll cache it for an hour in `~/.cache/nix` (remember `sudo` will make this `root`'s ~). You can also do `builtins.fetchTarball "https://github.com/owner/repo/archive/master.tar.gz"` and get the same effect. I believe you can also instruct fetchGit to use a different branch but I forget how
<haslersn>
elvishjerricco: That's actually much easier than I thought. I do '(import (fetchGit "https://github.com/haslersn/<reponame>"))' and it just works. Is it the case that import automatically looks for a 'default.nix' file in the tar and imports that?
<elvishjerricco>
haslersn: Yea. But you could string interpolate the result of `fetchGit` if you wanted to point to a different file. `import "${builtins.fetchGit ...}/foo.nix"`
<haslersn>
Thanks, I just learned stuff within 10 minutes which I spent the whole day on trying to figure out
<elvishjerricco>
haslersn: Happy to help :)
joehh has quit [Ping timeout: 240 seconds]
orivej has joined #nixos
cnidario has quit [Remote host closed the connection]
<haslersn>
in nikpkgs/.github/CONTRIBUTING.md I read about contributing. However, what are the requirements for a package to be accepted (appart from meeting the formal requirements of the pull request)?