<{^_^}>
nix#1681 (by ledettwy, open): nix copy uses too much memory
<{^_^}>
[nixpkgs] @dtzWill pushed to master « Revert "Merge pull request #43983 from viric/coreutils_ssl_master" »: https://git.io/fNgXe
<gchristensen>
yay dtz[m]
init_6 has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/fNgXk
<dtz[m]>
gchristensen: haha yeah that was an easy decision ;)
<dtz[m]>
did.. borg approve it? surely not, right? lol
aarvar has joined #nixos
<dtz[m]>
don't see borg saying anything other than that it was mass-rebuild
<dtz[m]>
which is kinda silly to do on master so some utilities are faster-- since AFAIK they're not used in-tree for much?
<dtz[m]>
wonder if it makes sense to bother with that in stdenv vs coreutils package having them
<dtz[m]>
(Currently we inherit it from a stage or two earlier)
<gchristensen>
dtz[m]: nobody actually asked borg to build anything
Mateon1 has quit [Remote host closed the connection]
Mateon1 has joined #nixos
sam__ has joined #nixos
Fare has joined #nixos
trylist has quit [Ping timeout: 268 seconds]
Fare has quit [Ping timeout: 256 seconds]
<ivan>
copying path '/nix/store/n4iwhzv6ay40v0jyi58crsxhrfa4gxda-linux-4.14.59.tar.xz' from 'https://cache.nixos.org'... <- this seems to be going at 60KB/s from California for some reason
Guanin has quit [Remote host closed the connection]
<gchristensen>
time=43.9 ms not the nicest but not the worst
<gchristensen>
I don't see anything wrong
<ivan>
I wonder what's up with cloudfront then?
<gchristensen>
might be a transient thing
<ivan>
yes, it's back to being fast
sam__ has quit [Ping timeout: 256 seconds]
<gchristensen>
good :)
MP2E has joined #nixos
<buckley310>
If i would like to install a package that only exists in the master nixos branch, not in any release, is there a short way to tell nixos to pull from master, or is the best way to just pull it down locally and include it?
<buckley310>
(just pull the one package in form master, not switch my whole system to it)
blonkhart has joined #nixos
n0qqe has joined #nixos
<gchristensen>
there is something on the FAQ on nixos.wiki about that
Mateon1 has quit [Remote host closed the connection]
Mateon1 has joined #nixos
layus has joined #nixos
<ivan>
is this a bug? (installing nixpkgs.xfceUnstable) error: packages '/nix/store/1y1h9nzlq5x2wlzy8ykhmqq8aqzp4cix-xfce4-power-manager-1.6.1/share/icons/hicolor/icon-theme.cache' and '/nix/store/0g5zff60s4klajjxf47w84d422arrfhh-xfdesktop-4.13.2/share/icons/hicolor/icon-theme.cache' have the same priority 5; use 'nix-env --set-flag priority NUMBER INSTAL
<ivan>
LED_PKGNAME' to change the priority of one of the conflicting packages (0 being the highest priority)
slack1256 has quit [Remote host closed the connection]
<d1rewolf>
is npm2nix necessary to use npm on nix?
ericsagnes has joined #nixos
trylist has joined #nixos
<Fare>
d1rewolf, no
<d1rewolf>
Fare, ah, so nodejs deriv includes npm. I see
<smolboye>
i will upload my findings and fixes in my config on github
<smolboye>
btw can i install unstable packages?
<smolboye>
on stable
<smolboye>
also how unstable is unstable?
init_6 has quit [Ping timeout: 264 seconds]
<drakonis>
its unstable as in it changes
<drakonis>
may change a lot
<joepie91>
smolboye: the difference as I understand it, is basically that on stable/release channels care is taken to not break applications (including integration between applications), and generally updates will only be applied if they don't significantly change the application and don't break APIs
<joepie91>
whereas `unstable` lands basically any change that gets submitted to nixpkgs and that doesn't break the build
init_6 has joined #nixos
init_6 has quit [Changing host]
init_6 has joined #nixos
Tiez has joined #nixos
<smolboye>
ah
<NegaNexus>
joepie91: maybe a stupid question but how do I use ,locate
<smolboye>
um ok so i would like to install one unstable package on my system, but use the system as a stable one
<joepie91>
NegaNexus: literally just type it into channel like gchristensen and I did, `,locate` followed by whatever library/binary file you're looking for
<smolboye>
so adding the unstable channel and putting the package in my config is the only way?
orivej has quit [Ping timeout: 256 seconds]
<{^_^}>
[nixpkgs] @jameysharp opened pull request #44187 → Fix gobject-introspection, and package gnome-builder and dependencies → https://git.io/fNgS5
phdoerfler has joined #nixos
ericsagnes has joined #nixos
simukis has joined #nixos
<smolboye>
ok can i add channels as a user and use it for userpackages?
revskill has joined #nixos
<clever>
smolboye: yes
<smolboye>
noice, thanks
pie_ has joined #nixos
drakonis has quit [Read error: Connection reset by peer]
blankhart has quit [Ping timeout: 260 seconds]
blankhart has joined #nixos
revskill has quit [Ping timeout: 260 seconds]
katyucha has joined #nixos
phreedom has quit [Ping timeout: 250 seconds]
revskill has joined #nixos
<smolboye>
uh another question
<smolboye>
how do i add pulseaudio support to polybar
<smolboye>
sorry if that sounds stupid
phreedom has joined #nixos
<manveru>
smolboye: what kind of support?
<clever>
i glanced at it nix expression and it currently does not depend on pulse
<ivan>
"This isn't enabled by default for performance"
<manveru>
afaicr it uses namespaces to make only parts of the filesystem visible, plus the networking thing...
<{^_^}>
[nixpkgs] @FRidh pushed to master « python.pkgs.scikitlearn: disable tests, not package »: https://git.io/fNg7T
<manveru>
but i'd never turn it off unless you're building on a super slow sytem...
sir_guy_carleton has quit [Quit: WeeChat 2.0]
<ivan>
maybe it should be enabled by default?
<clever>
ivan: the performance hit is per dependency in the closure, since it has to do a --bind mount for every single input path
<clever>
id keep it enabled 24/7
<clever>
and i think its mainly a hit during startup
<clever>
per-derivation
<andi->
I think on my 2nd day of nixos I enabled it since running builds without it sounded scary. Since then there was no reason to undo that.
<manveru>
it really should be on by default, imho...
<manveru>
but that breaks macos
<andi->
no, they don't use nixos modules?
<manveru>
they can't use sandboxing
<andi->
I just meant it wont break macos (or should not, not a mac user) since that code is not relevant for them.
<manveru>
ah, for the nixos option, yeah...
<manveru>
i meant for the nix option
<andi->
ok
<manveru>
`nix --help-config | grep sandbox`
<andi->
As a first step I'd go for defaulting it in nixos. For nix is probably comes with more issues since everyone then needs the kernel modules required for sandboxing?
jasongrossman has quit [Remote host closed the connection]
kiloreux has quit [Ping timeout: 260 seconds]
jasongrossman has joined #nixos
<typetetris>
Is there a simple walkthrough guide to configure openvpn with networkmanager with nixos? I only find old stuff, which doesn't seem to work for me.
aszlig has quit [Quit: Kerneling down for reboot NOW.]
<{^_^}>
[nixpkgs] @peti pushed 2 commits to haskell-updates: https://git.io/fNg5v
<{^_^}>
[nixpkgs] @peti pushed 365 commits to haskell-updates: https://git.io/fNg5T
kiloreux has joined #nixos
<andi->
re sandboxing: Going through some recent changes showed that LnL helped someone rephrased the description of the sandbox option. I wonder if there was also an discussion about defaulting it to enabled.
<typetetris>
hyper_ch: That is not what I am looking for, I think. I get an '.ovpn'-File from my admin and try to use networkmanager to import it, but it refuses to store the path of my private key file and doesn't ask for it, when I try to start the connection with nm-applet. I don't want to setup a server.
<hyper_ch>
this is not about a server
<typetetris>
ah ok, hmm. Doesn't that collide with networkmanager? I will try it, but I suspect it will not update /etc/resolv.conf and stuff?
<{^_^}>
[nixpkgs] @andir opened pull request #44190 → nixos/nix-daemon: default `nix.useSandbox` to `true`. → https://git.io/fNg5l
<{^_^}>
nix#179 (by edolstra, open): Chroot builds are slow
iyzsong has joined #nixos
<typetetris>
hyper_ch: ok that works sort of, need to start it as root with `systemctl start openvpn-mine.service` but it doesn't integrate with networkmanager.
<hyper_ch>
not sure what you did
<hyper_ch>
systemctl enable openvpn-mine.service
knupfer has quit [Ping timeout: 260 seconds]
<typetetris>
I just added the lines (regarding the vpn client section) from https://nixos.wiki/wiki/OpenVPN to my nixos configuration and adapted it to use the configuration file I got from my admin. Then I can start and stop it with `systemctl start/stop openvpn-mine.service`. I don't want to have it running all the time, as it is configured to route all my traffic over my employer.
<typetetris>
After a `nixos-rebuild switch` of course.
<typetetris>
clever: I don't understand, what that shall achieve?
mariatsji has joined #nixos
<clever>
typetetris: that will disable the auto-starting of the service, so it will only respond to start/stop commands via systemctl
<typetetris>
clever: Oh thanks. I already did this with `services.openvpn.servers.mine.autoStart = false;`. At least it didn't start directly after doing `nixos-rebuild switch` didn't check a reboot though.
<clever>
if autoStart is true, then wantedBy = [ "multi-user.target" ];
<clever>
and if autoStart is false, its [], so no need to mkForce
mariatsji has quit [Ping timeout: 260 seconds]
<typetetris>
I can live with how it is now. It would be really nice if using 'Add VPN connection` in `nm-applet` and starting it through `nm-applet` was working. (At least I didn't managed to make it work.)
ok9swl has joined #nixos
MP2E has quit [Remote host closed the connection]
agander has quit [Ping timeout: 260 seconds]
ThatDocsLady has quit [Remote host closed the connection]
ThatDocsLady has joined #nixos
NightTrain has joined #nixos
orivej has joined #nixos
alex`` has quit [Quit: WeeChat 2.2]
<Akii>
building a derivative is failing; where can I see the whole log?
<Akii>
not just the last 10 lines of it
<Akii>
passing -v wasn't as verbose as I had hoped either :D
ris has joined #nixos
<ris>
wonder if anyone's had any luck with java 11/12 & nix
crmlt has quit [Ping timeout: 260 seconds]
knupfer has joined #nixos
<andi->
Akii: nix-store --read-log <drv-path>
<Akii>
drv path is the path to default.nix?
<andi->
no to the .drv file that it is trying to build
<andi->
usually that is being noted at the beginning and also when a build fails
<Akii>
ah ok, thanks!
phdoerfler1 has joined #nixos
<Akii>
actually found out why it failed to build but good to know for the future
timon37 has quit []
phdoerfler has quit [Ping timeout: 265 seconds]
Guanin has joined #nixos
<hyper_ch>
hmm, something has blocked nixos-unstable and nixos-unstable-small for the last 2 weeks :(
<cocreature>
it was blocked by a wide variety of things :/ not sure why the successful build yesterday didn’t cause an update. I feel like channel updates will forever remain a mystery to me
FRidh has quit [Read error: Connection reset by peer]
<hyper_ch>
but when I click on the first failed job I get like: Aborted: cannot connect to ‘root@wendy.ewi.tudelft.nl’: ssh: connect to host wendy.ewi.tudelft.nl port 22: Connection timed out (log, raw, tail)
<hyper_ch>
first failed job (one with a red x)
ajs124 has joined #nixos
me123456 has joined #nixos
ThatDocsLady has quit [Quit: Leaving]
mariatsji has quit [Ping timeout: 260 seconds]
<clever>
hyper_ch: the build slave was offline
<hyper_ch>
clever: but it's back online now?
<clever>
probably
<hyper_ch>
how to find out why it's not advancing the channel?
<d1rewolf>
guys, I have vs code installed via nix, and it keeps telling me there's an update available. I suppose in-app updates like this are not possible to use?
<hyper_ch>
what's vs code?
<d1rewolf>
hyper_ch, visual studio code
* hyper_ch
shivers
<hyper_ch>
no, in-app updates won't work
* d1rewolf
nods
<d1rewolf>
what would happen in this case?
agander has joined #nixos
<hyper_ch>
the universe might implod
<hyper_ch>
implode
<hyper_ch>
(or it just fails because nix store is read-only)
<d1rewolf>
does nix-channel --update effectively pull down new upgrades, which are applied upon the next nixos-rebuild switch?
nschoe has joined #nixos
<joko>
d1rewolf: the latter
<d1rewolf>
joko: what I mean is: "is the appropriate way to update your system 'nix-channel --update && nixos-rebuild switch' ?"
<gchristensen>
sure, also nixos-rebuild switch --upgrade
<d1rewolf>
gchristensen, ok. but they're effectively the same thing?
nschoe has quit [Client Quit]
<gchristensen>
yes
<d1rewolf>
gchristensen, thank you
nschoe has joined #nixos
jtojnar has quit [Read error: Connection reset by peer]
jtojnar_ has joined #nixos
jtojnar_ is now known as jtojnar
nschoe has quit [Client Quit]
nschoe has joined #nixos
<aanderse-laptop>
how can i get nix-repl to show me the full output? when i run code that outputs an array of sets it truncates the output
<joko>
d1rewolf: as gchristensen said... Think of this as apt-get update && apt-get upgrade or pacman -Sy && pacman -Su (or sth like that, I don't remember anymore)
<{^_^}>
[nixpkgs] @Mic92 pushed 2 commits to release-18.03: https://git.io/fNgNC
<d1rewolf>
joko, yes, it's clear now. thank you ;)
<sphalerite>
aanderse-laptop: you mean with <CODE>?
luketopia has quit [Quit: Leaving]
<aanderse-laptop>
yeah i'm just typing some code into nix-repl to try and learn
<aanderse-laptop>
but when it spits out sets it truncates them
<aanderse-laptop>
i see { ... }
<aanderse-laptop>
instead of what the values are
<sphalerite>
aanderse-laptop: use ":p <expr>"
<aanderse-laptop>
so i'm not sure if my code worked as expected or not
<sphalerite>
see :? for help
<d1rewolf>
guys, the node version (6.14.3) installed by nixos is very old. Is there a way to get a more up-to-date one? looks like the LTS version is at 8.11.3
<d1rewolf>
oh, I see there are nodejs-8_x
<aanderse-laptop>
thanks sphalerite!
<d1rewolf>
will try that
nschoe has quit [Quit: Program. Terminated.]
toby1851 has quit [Ping timeout: 244 seconds]
mariatsji has joined #nixos
smolboye has quit [Ping timeout: 260 seconds]
mariatsji has quit [Ping timeout: 240 seconds]
stanibanani has quit [Ping timeout: 256 seconds]
crmlt has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @7c6f434c pushed 4 commits to release-18.03: https://git.io/fNgAY
revtintin has joined #nixos
blankhart has quit [Read error: Connection reset by peer]
stanibanani has quit [Ping timeout: 256 seconds]
<Akii>
seems to be impossible to configure worker_processes for nginx
stanibanani has joined #nixos
blankhart has joined #nixos
<LnL>
howso?
toby1851 has joined #nixos
<Akii>
there is no option and appendConfig does not result in a correct config
<Akii>
apparently the worker_processes setting must come at the beginning of the config
<Akii>
sounds like this should be an option
<LnL>
there's appendConfig and appendHttpConfig
<LnL>
depending on what section it needs to be in
aanderse-laptop has quit [Ping timeout: 245 seconds]
<Akii>
nah I'm just stupid
<Akii>
wrote `worker_processes = 3;` instead of `worker_processes 3;`
<Akii>
it even says wrong number of args
<LnL>
ah :)
blankhart has quit [Ping timeout: 260 seconds]
<Akii>
but so cool
<Akii>
nix-shell -p nginx and I got my own little nginx to play around with
<LnL>
depending on your usecase it might be nicer to just use services.nginx.config
phdoerfler has joined #nixos
<LnL>
the options are nice for simple usecases but it's probably not worth it if you have to reverse engineer a normal nginx config
<Akii>
it's fine
phdoerfler1 has quit [Ping timeout: 265 seconds]
orivej has quit [Ping timeout: 256 seconds]
<Thra11>
When using distributed builders, I'm assuming the derivation to be built is sent to the builder: i.e. it doesn't matter which channel the builder is on, it will always build whatever's in the host's nixpkgs. Is that correct?
<joko>
Thra11: I think so, too
<joko>
It is fully declarative
<joko>
E.g. on my Hydra instance I have builders in NixOS 18.03 which are building nixos-unstable packages and configurations
<Thra11>
Presumably if the builder did build a different derivation, the hash wouldn't match, so the host would reject it anyway
alex`` has joined #nixos
orivej has joined #nixos
toby1851 has quit [Ping timeout: 244 seconds]
civodul has joined #nixos
<gchristensen>
you are trusting the builder to not be malicious, of course
<gchristensen>
your assumption is correct though
nschoe has joined #nixos
siriobalmelli has quit [Ping timeout: 252 seconds]
nckx has quit [Quit: Updating my GNU GuixSD server — gnu.org/s/guix]
<{^_^}>
[nixpkgs] @FRidh pushed to master « remove file that was accidentally added »: https://git.io/fN2vJ
jtojnar has quit [Remote host closed the connection]
jtojnar has joined #nixos
FareTower has joined #nixos
<{^_^}>
[nixpkgs] @etu opened pull request #44197 → php: drop 5.6 and 7.0 → https://git.io/fN2vB
koisell_ has quit [Ping timeout: 276 seconds]
mconstant has joined #nixos
mekeor has joined #nixos
<mconstant>
just installed another Linux distro alongside NixOS. now it only boots into the other distro. tips?
<Yaniel>
tell the bootloader to scan for other operating systems
<Yaniel>
and hope you didn't actually overwrite the nixos boot partition
<etu>
mconstant: If it's EFI it probably just registred another bootloader so you can select which one to use with your "bios settings"
sbdchd has quit [Remote host closed the connection]
<Yaniel>
if you are lucky, ^
worldofpeace_ has quit [Ping timeout: 256 seconds]
mconstant has quit [Ping timeout: 256 seconds]
mariatsji has joined #nixos
mariatsji has quit [Ping timeout: 244 seconds]
mconstant has joined #nixos
<mconstant>
Yaniel: how do I tell the bootloader to scan for other operating systems? and no chance it overwrote the efi boot for NixOS. its on a whole other drive that the Pop!os install did not touch...
<Yaniel>
well try first what etu said
<Yaniel>
< etu> mconstant: If it's EFI it probably just registred another bootloader so you can select which one to use with your "bios settings"
<juhe>
mconstant: check with "efibootmgr -v"
<mconstant>
Yaniel: so just use the BIOS to boot? I bet that will work, but is there a way to make a menu to make things nicer
phdoerfler has joined #nixos
<Yaniel>
yes, by telling the bootloader to scan for other installations
<Yaniel>
which varies depending on which bootloader exactly you want to use as the primary one
<Yaniel>
fwiw there is a nixos option for enabling that functionality in the nixos bootloader
<mconstant>
Yaniel: I'm guessing it is in the manual or on the wiki... and I add it to the configuration.nix?
<tilpner>
Any ideas how I might generate/document a kernel .config? I'd like a way to express "enable anything required for CONFIG_...", without listing the required options
<Dezgeg>
just implement a SAT solver for solving Kconfig equations :P
raynold has joined #nixos
<tilpner>
scripts/config looks interesting
tusj has quit [Ping timeout: 264 seconds]
<alphor>
registering an overlay in a nixos file is done like this, right? pkgs = import <nixpkgs> { overlays = [myOverlay] } and then refer to pkgs in your nixos config
<srk>
Dezgeg: wow :D
mconstant has quit [Ping timeout: 264 seconds]
<srk>
I've had to run uboot build incrementally adding deps few times, lovely
<Dezgeg>
(not even kidding because config options can have dependencies like 'depends on FOO || (BAR=n && BAZ)')
endforma1 has quit [Read error: Connection reset by peer]
endformationage has quit [Ping timeout: 256 seconds]
pie_ has joined #nixos
pie__ has joined #nixos
pie_ has quit [Client Quit]
periklis has joined #nixos
<{^_^}>
[nixpkgs] @jerith666 opened pull request #44202 → journald: set rateLimitInterval and rateLimitBurst to upstream defaults → https://git.io/fN2Jb
<{^_^}>
[nixpkgs] @jerith666 opened pull request #44203 → Restic fixes: pruning, process substitution → https://git.io/fN2Jp
<hyper_ch>
hmmm, I need git help again: I have a repo with Master and 2 branches. Now I noticed an error in master and want to fix it. But that same fix should also be applied to the two branches since they have the same error. How to achieve that?
endforma1 has joined #nixos
davidwest has joined #nixos
endforma1 has quit [Ping timeout: 245 seconds]
<cocreature>
hyper_ch: commit it to one of the branches and cherry pick it to the others
<{^_^}>
[nixpkgs] @manveru opened pull request #44205 → golangci-lint: init at 1.9.2 → https://git.io/fN2Ti
<manveru>
i could use some help with this PR :)
<manveru>
not sure how to tell gcc to not fail on -Wformat
periklis has quit [Quit: ERC (IRC client for Emacs 26.1)]
endforma1 has joined #nixos
<samueldr>
NIX_CFLAGS_COMPILE = "-Wno-format"; ?
<manveru>
let's see :)
<samueldr>
you also are missing `lib` in golangci-lint/default.nix
<samueldr>
(ofborg ofbarfed)
<manveru>
i saw, didn't rebuild after adding the meta :P
<samueldr>
:)
<manveru>
hard to tell what's failing now... but i think it tries to compile the examples for the test... which it shouldn't
endforma1 has quit [Ping timeout: 264 seconds]
<samueldr>
I see you deleted the tests folder
<samueldr>
instead maybe look for a makeFlag you could use
<samueldr>
or even patching the makefile
<manveru>
there's no makefile...
<samueldr>
uh, scratch that, I haven no idea what buildGoPackage does
<samueldr>
:)
<manveru>
well, none i use anyway
<manveru>
:D
<samueldr>
(I was thinking about the generic build)
<manveru>
anw, it looks for directories with .go files recursively and runs `go test` for all of them, basically
<manveru>
even though some of those dirs only contain example data for the linting, not actual tests
<samueldr>
and they say go doesn't have generics: pkgs/development/go-modules/generic/default.nix <-- ;)
<manveru>
so i need to find a way to ignore those
Sonarpulse has joined #nixos
<manveru>
ah, maybe if i specify subPackages
<samueldr>
manveru: fails at checkPhase or buildPhase?
<manveru>
checkPhase
<samueldr>
you could always scrap it alltogether if no checks at all need to run
<manveru>
and no, i don't just want to turn off tests :P
<samueldr>
good :)
<manveru>
they should be running perfectly fine if it wasn't for our weird way of finding them
<samueldr>
yeah, looking at the source subPackages is used by getGoDirs which is in turn used by checkPhase
davidwest has quit [Remote host closed the connection]
<manveru>
oh, it actually has a makefile with a test section
<manveru>
i shall use that then
stanibanani has quit [Ping timeout: 264 seconds]
Sonarpulse has quit [Remote host closed the connection]
FareTower has quit [Ping timeout: 245 seconds]
endforma1 has joined #nixos
zduch4c has joined #nixos
<zduch4c>
hallo
<zduch4c>
what was the command to override a download path in NixOS?
<manveru>
updated the PR... seems the subPackages was the superior solution
<zduch4c>
im getting a 404 out of one of the downlados, and I need it fixed fast
<zduch4c>
thanks, but I'm sure there was a command-line thing
Sonarpulse has quit [Remote host closed the connection]
<zduch4c>
something with nix-env or something
<manveru>
uhm, if you thinnk so :)
<manveru>
can't say i've heard of it
endforma1 has quit [Ping timeout: 240 seconds]
<zduch4c>
you basically put in a path to some nix command, and it used it while downloading instead of the URL
<zduch4c>
ok
<zduch4c>
its nix-prefetch-url
the has quit [Ping timeout: 244 seconds]
stanibanani has joined #nixos
Izorkin has joined #nixos
<samueldr>
zduch4c: this shouldn't allow replacing something with something else, since it will hash the results, only the provenance of said thing
knupfer has joined #nixos
<samueldr>
but yeah, since you're getting 404s that's a way to go
<zduch4c>
yes, but if one URL gives me 404 and the other doesn't while both have the same hash it works perfectly
<samueldr>
zduch4c: if the URL changed upstream, either open an issue or make a PR if possible?
<zduch4c>
i dont know how to use git apart from git clone
<zduch4c>
anywho, thanks for help
zduch4c has quit [Quit: Page closed]
<samueldr>
no worries :)
<manveru>
well, that makes more sense :)
<samueldr>
manveru: I assumed it was ready to be built by ofborg
Lisanna has quit [Remote host closed the connection]
<manveru>
should be
<samueldr>
then at least it will already have been built :)
<manveru>
:)
<manveru>
thanks
<samueldr>
btw, I updated nix-top, received a tip to get a good stand-in path without root
alex`` has quit [Quit: WeeChat 2.2]
sbdchd_ has joined #nixos
sbdchd has quit [Ping timeout: 256 seconds]
<{^_^}>
[nixpkgs] @Ericson2314 opened pull request #44206 → llvm 3.8, 3.9, 4: WIP split out compiler-rt for other versions → https://git.io/fN2kF
phreedom_ has quit [Remote host closed the connection]
wpcarro has joined #nixos
phreedom has joined #nixos
endforma1 has joined #nixos
stanibanani has quit [Ping timeout: 264 seconds]
<Thra11>
Is there an option to disable local builds, such that it will either use existing binary substitutes or offload a build to another machine, but won't build things locally?
endforma1 has quit [Read error: Connection reset by peer]
<Thra11>
(even if the other builders are fully loaded or not contactable)
<manveru>
samueldr: looks really nice now
<manveru>
almost professional :)
<samueldr>
manveru: I assume you didn't peek at the code ;)
<infinisil>
Thra11: --max-jobs 0 should work for that
<manveru>
samueldr: what's wrong with it?
<samueldr>
nothing much, just a bit hacky still
<samueldr>
(at one specific location I'm using begin/rescue while I should probably check stuff instead)
<Thra11>
infinisil: Of course! How did I miss that. Thanks!
<manveru>
not sure why you'd shellescape a string you pass to File.read, but well :)
<samueldr>
hah, that's probably from something that was previously in a command
<{^_^}>
[nixpkgs] @Infinisil merged pull request #44061 → nixos/avahi: add support for extraConfig → https://git.io/fNBoD
stanibanani has joined #nixos
<infinisil>
etu: Tested and it works?
<{^_^}>
[nixpkgs] @markuskowa opened pull request #44207 → halide: init at 2018_02_15 → https://git.io/fN2Ic
zduch4c has joined #nixos
<etu>
infinisil: yeah
<zduch4c>
its me again. when installing emacs via services.emacs (aka the emacs daemon), does the installed emacs desktop application launch emacs client (e.g. the emacs icon in GNOME)?
<zduch4c>
from what I gathered, the .desktop file launches `emacs %F`
<{^_^}>
[nixpkgs] @xeji pushed commit from @r-ryantm to master « pspg: 1.1.1 -> 1.2.1 (#43916) »: https://git.io/fN2Lq
goibhniu has joined #nixos
d1rewolf has quit [Remote host closed the connection]
<juhe>
samueldr: Yes, I have a release-18.03 for normal stuff, + some minor changes to other packages via overlay, but for cura, I build it from master branch and tried to run it with "nix run ..." shell.
<{^_^}>
[nixpkgs] @xeji pushed commit from @r-ryantm to master « toot: 0.18.0 -> 0.19.0 (#43898) »: https://git.io/fN2tV
<juhe>
samueldr: I've cherrypicked your b0d74d8486af commit on top of master and building the stuff now - it will take some time though, not a big performer this machine is...
<samueldr>
ah, I'm doing this right now too
<samueldr>
well, it's been running for a probably 10-15 minutes
<juhe>
ok, I was trying to figure out how to get the patch, I first tried to merge the feature branch, but it was a no-no, too many changes, so did quick cherry pick.
<infinisil>
Test dependencies should be in nativeBuildInputs, right/
<infinisil>
?*
<samueldr>
juhe: in my case, I rebased on master, I generally open my branches opened on nixos-unstable to reduce the cost building :)
<{^_^}>
[nixpkgs] @xeji pushed commit from @r-ryantm to master « libextractor: 1.6 -> 1.7 (#43756) »: https://git.io/fN2q5
<camsbury>
boom, mode is working As far as I can tell
<camsbury>
thanks!
<infinisil>
Glad I could help
<jasongrossman>
infinisil++
<{^_^}>
infinisil's karma got increased to 17
<camsbury>
I have noticed in the past that you store snippets of would be ".el" files in your .nix expressions?
<infinisil>
^.^
tusj has quit [Ping timeout: 256 seconds]
<camsbury>
is that how you manage all your emacs stuff?
<infinisil>
camsbury: Pretty much, I don't really change it that often though, what I have now has been working well for a while
<camsbury>
++
mounty has joined #nixos
<infinisil>
Well I am using my own little emacs module system for it, which is working well too
<camsbury>
hmm
<camsbury>
okay cool... I'll check out your stuff and get back to you sometime. I want to try to move the configuration of my packages into the nix expressiosns containing them
<samueldr>
it really is neat when dealing with rebuilds
<samueldr>
you get to gauge what's going on
<michas_>
Hi, i created a new bridge interface using "brctl addbr" and it automatically gets some 169.254.x.x address assigned. As far as I understand this zeroconf behavior is normaly handled by avahi but there seems to be no avahi daemon running. - Any idea where exactly that address comes from?
sigmundv_ has quit [Ping timeout: 248 seconds]
<clever>
michas_: the dhcp server will also do that
<clever>
michas_: nixos also supports creating a bridge for you in configuration.nix
<samueldr>
infinisil: opened an issue to eventually work on it
<michas_>
clever: I tried to play locally with coreos matchbox. having a bridge, a dnsmasq pointing to matchbox and starting vm on that bridge to boot via matchbox. - not sure if that is the best way to do it.
<clever>
michas_: ah
wpcarro has joined #nixos
<clever>
michas_: if you define the bridge and its static ip in the nixos config, then i believe dhcpcd will ignore it
tusj has joined #nixos
<infinisil>
samueldr: Hmm, regarding the Nix there, I usually prefer to not use random env vars (ADDITIONAL_PATH) but a let in instead
<samueldr>
(I'm even not using it as an environment variable!)
<samueldr>
yeah, I think a style guide is required :/
<michas_>
clever: maybe a real nixos test spinning up different VMs would be cleaner. but I had the feeling the manual way would be a bit easier. - Now I'm not completely sure. ;)
<infinisil>
samueldr: Ah yeah that's what I mean, it doesn't have to be an env var
<samueldr>
while it is, it's basically not :)
goibhniu has joined #nixos
<samueldr>
I'm never entirely sure if I should `let` them, or make em a bit less clean by not putting it inside a variable
<samueldr>
and then `version` I've seen comments that it's preferred to `rec` instead of `let + inherit`
<infinisil>
samueldr: Oh also, some nits: makeWrapper should be in nativeBuildInputs, and the chmod isn't needed (i think your file in git already has +x)
wpcarro has quit [Ping timeout: 264 seconds]
<infinisil>
samueldr: Yeah, a style guide would be nice :)
<clever>
samueldr: i prefer to put version into a let block, using rec misleads people into thinking overrideAttrs can change the version
<samueldr>
clever: *flailing* a style guide is necessary then! /me braces for bikeshedding
<infinisil>
clever: Agreed, however, it's convenient, as the version gets passthru'd, which I think would be nice to have for all derivations
blankhart has joined #nixos
<infinisil>
clever: Instead of using builtins.parseDrvName..
<clever>
> builtins.parseDrvName hello.name
<{^_^}>
{ name = "hello"; version = "2.10"; }
<infinisil>
Just doesn't feel right!
<samueldr>
infinisil: I'm happy with the nitpicking, always glad to make things more correct :)
<infinisil>
samueldr: Well if you say so: The last "=" in meta isn't aligned with the rest :P
wpcarro has joined #nixos
asymmetric has quit [Ping timeout: 260 seconds]
<infinisil>
Literally unreadable
<samueldr>
it's fixed :)
<infinisil>
Heh
<samueldr>
(I obviously took the meta block from elsewhere)
<samueldr>
oh, and not based off what you're doing, my peeve with nitpicking is when it's done bit by bit *on a non-instantaneous medium* e.g. comments/mail
<Dezgeg>
I personally prefer version inside the mkDerivation since then lib.getVersion can return that directly
<samueldr>
aw, I don't want to write another RFC *now*
<Dezgeg>
(and not fall back to the parseDrvName one which would return 'ntfs-3g-2018.01' having { name = "ntfs"; version = "3g-2018.01" })
wpcarro has quit [Ping timeout: 244 seconds]
<infinisil>
Dezgeg: Should really be in meta or passthru though
<Dezgeg>
why?
<infinisil>
Because it's not needed as an env var for the build
<infinisil>
Seems cleaner to use passthru for that (or meta)
<clever>
infinisil: i have also heard that env vars have a measurable impact on performance
<clever>
infinisil: some of the ghc benchmarks that have been running for years, have been discovered to have major error, because the stack is offset by the size of the env vars, and those where not cleared prior to testing
<infinisil>
Lol
mahalel_ has joined #nixos
<clever>
infinisil: and that adjusted the alignment of the stack, and basically all benchmarking and tuning for the last decade is wrong :P
<Dezgeg>
well that will be affected by a ton of random and unpredictable things like the length of your username
<__monty__>
So haskell's gonna get even better? : )
<clever>
Dezgeg: yep
<clever>
__monty__: now that that has been identified, i would hope that they can solve it, and make the benchmarks more pure
<Dezgeg>
I think I once had some problem that happened only when running the binary with the full path and not with ./binary due to the shifted memory layout
jasongrossman has quit [Remote host closed the connection]
jasongrossman has joined #nixos
stanibanani has quit [Ping timeout: 260 seconds]
<Dezgeg>
there was a missing cache invalidation in some gpu testcase which would only happen if the stars were aligned just right
<juhe>
samueldr: Build finished, tried to run cura, but no success: https://ptpb.pw/6QHC , Next, tried it with strace with following result: https://ptpb.pw/LUrU .
mizu_no_oto has joined #nixos
wpcarro has quit [Ping timeout: 256 seconds]
<samueldr>
hmm, still building here
<juhe>
oh, ok.
lord| has quit [Quit: WeeChat 2.2]
lord| has joined #nixos
jmeredith has quit [Quit: Connection closed for inactivity]
Anton-Latukha has quit [Ping timeout: 260 seconds]
vcunat has quit [Quit: Leaving.]
<juhe>
samueldr: It is searching also home folder, so did temporary symlink to relevant lib folder in nix store, though it fails - the library is found but couldn't initialize or something: https://ptpb.pw/H0CR
<samueldr>
juhe: which branch have you used?
<samueldr>
qtbase-5.11.0 doesn't smell right
<samueldr>
it should be 5.11.1 on master IIRC
<samueldr>
I have a branch with 5.11.0 *to break stuff*
<samueldr>
(mixing patch-level versions of qt is known to break)
wpcarro has joined #nixos
<juhe>
samueldr: Hm, good catch. Though I built using "commit 285d7709494fdba72299a61b05f9becd76bc3519 (origin/master, origin/HEAD, master)" on top of which is your patch...
<samueldr>
hm, weird how you got qtbase-5.11.0 there
<juhe>
samueldr: yes
<samueldr>
juhe: in your user-profile (nix-env -q) do you have other Qt software installed, maybe installed from an older unstable/master checkout?
<juhe>
samueldr: Neither know I... Maybe it is from past? I built cura some time ago (don't remember when exactly)
<clever>
juhe: ls -lhd ~/.nix-profile/lib/qt*
<juhe>
samueldr: freecad-0.17, I think that one uses qt (and python...) too
wpcarro has quit [Ping timeout: 264 seconds]
<juhe>
But "ls -lhd ~/.nix-profile/lib/qt*" doesn't return anything.
<juhe>
There's lib, but no "qt*" in there
Anton-Latukha has joined #nixos
ambro718 has quit [Quit: Konversation terminated!]
iceychris has joined #nixos
<juhe>
samueldr: freecad-0.17 uses qt4.
<samueldr>
yeah, AFAIK the qt issue is qt5 only
<juhe>
samueldr: ok
iceychris has left #nixos [#nixos]
<juhe>
I did try to build cura this way some time ago, but didn't install it into environment. So I think it should not matter.
mizu_no_oto has quit [Quit: Computer has gone to sleep.]
<juhe>
samueldr: Cleaned local environment, there's only freecad now.
<samueldr>
(it's now building pyqt, so hopefully I'll be soon able to confirm)
<infinisil>
samueldr: I'll remember to always ping you when I see the xcb problem now :)
<juhe>
samueldr: but... I have cura from release-18.03 installed via systemPackages... Will kick it out.
<samueldr>
infinisil: let's hope the patch fixes it for good :3
<juhe>
samueldr: no change after kicking release-18.03/cura, unstable/cura still fails and strace shows 5.11.0. Bah...
<samueldr>
is your system up-to-date though?
<infinisil>
,qt = This is a very common error for using QT applications without running KDE, maybe `nix-env -iA nixos.qt5.qtbase` will fix that issue for you for now. See samueldr's https://github.com/NixOS/nixpkgs/pull/44047 for a hopefully persistent fix
<samueldr>
,qt = A common error, and known issue, is how the Qt integration with nixpkgs fails under some circumstances to provide proper plugin paths, leading to Qt failing to load Qt platform plugin "xcb". // See https://github.com/NixOS/nixpkgs/pull/44047 for a hopefully persistent fix. Meanwhile, install Qt apps from only one channel (and update accordingly)
<{^_^}>
qt redefined, was defined as: This is a very common error for using QT applications without running KDE, maybe `nix-env -iA nixos.qt5.qtbase` will fix that issue for you for now. See samueldr's https://github.com/NixOS/nixpkgs/pull/44047 for a hopefully persistent fix
<{^_^}>
[nixpkgs] @AmineChikhaoui opened pull request #44212 → libsemanage: fix bad meta.outputsToInstall (same as 57b9719e4cf601c617f02f59365e7218eefb9b0f) → https://git.io/fN2Gc
<infinisil>
:)
<infinisil>
Eventual new feature for the bot: Automatically detect people posting very common errors such as this one and replying with such info
<infinisil>
s/replying/reply
<juhe>
samueldr: I am up to date (release-18.03), but I need to check config if I am not bringing some unstable packages to system environment. In general, I try to run release-18.03 as base system and bring random stuff into user environment.
<samueldr>
yeah, if it wasn't for the fiddly Qt, it would be the best solution (stable + unstable pickings)
<samueldr>
(I personally run -stable)
tzemanovic has quit []
smolboye has quit [Ping timeout: 268 seconds]
<samueldr>
hm, juhe, I initially thought the build would have been shorter, it does fail (I half-expected it due to python)
<juhe>
samueldr: I had more stuff from unstable, but later I went to overlays (e.g. kicad) to update to newer version but still keep the older dependencies. Cura is exception, because stable is way too old, standard doesn't work (the right hand side bar with settings is missing). There was a small time window when cura from unstable (brought in systemPackages) worked, but then there were some changes to python which broke it. So I kicked it all
<juhe>
from system stuff and try to keep system as close to release-18.03 as possible and play with "nix build -f "path/to/master/nixpkgs/nixos" and having unstable in user environment instead.
<samueldr>
though this is probably a good software to use as a basis for python support, since it fails du to using thinkgs like qtquick
<juhe>
cura is nasty beast, I'm thinking to ditch it and use curagengine directly
<samueldr>
though, such a nasty beast could be my best test bed for actually fixing qt :)
<samueldr>
well mixed Qt
<juhe>
hehe, yes.
drakonis has joined #nixos
<juhe>
I need to dig in my nix store with "nix-store -q..." to find out why there's this qt-5.11.0
<{^_^}>
[nixpkgs] @Mic92 merged pull request #44118 → jetbrains.editors: 2018.1.x -> 2018.2 (upgrade using the script) → https://git.io/fNEBk
<mroutis>
I'm having a hard time making work any desktop/window manager on my netbook, is there a way I can debug this?
<mroutis>
I'm able to login, however, I only see a black screen, doesn't matter if using gnome/kde/bspwm etc.
<andi->
mroutis: whats the graphic stack like non that machine?
blankhart has quit [Ping timeout: 240 seconds]
<mroutis>
andi-: how can I find out my graphic stack?
<andi->
mroutis: well what chip is in there?
<mroutis>
ohh, right, let me see
<mroutis>
andi-: ATI Radeon 4225 graphics controller
<mroutis>
it's a quite old acer aspire one notebook; however, it was running ubuntu 14, with some issues but running
<andi->
I guess the GPU is not supported by "amdgpu" (the modern ATI driver)? Thats my best guess. There might be another one... Not really an ATI user for long myself only recently got an RX580 after being fedup with NVIDIA.
<juhe>
samueldr: well... I did "nix-env -f '/path/to/my/clone/of/nixpkgs/nixos' pkgs.cura" (not to get it garbage collected), then "nix-collect-garbage" and it ditched 10GiB of stuff... and then I wanted to look at the nix-store thing, but tried to run cura and it started... :-O
agander has quit [Ping timeout: 240 seconds]
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<samueldr>
yeah, it could have been any component in $PATH that still had reference somehow to qtbase 5.11.0 :/
<juhe>
right, I'll strace it to check 5.11.0 is out (just to be sure...)
<mightybyte>
How do you set up gzip compression for nginx with NixOS?
<mroutis>
andi-: could be, mhm, where should I look for the driver? are they on the nixpkgs?
<mightybyte>
I've found tutorials for doing it with plain nginx, but it doesn't seem to be working on my NixOS machine.
<juhe>
samueldr: yes, 5.11.0 is gone and cura works.
<juhe>
samueldr++
<{^_^}>
samueldr's karma got increased to 11
<andi->
mroutis: services.x11.videoDrivers IIRC is the option.
<andi->
servics.xserver.videoDriver
<mroutis>
<3 thanks!
fleaz has quit [Quit: fleaz]
__monty__ has quit [Quit: leaving]
blankhart has joined #nixos
jasongrossman has quit [Remote host closed the connection]
sigmundv_ has joined #nixos
wpcarro has joined #nixos
<worldofpeace>
Like how does nixops handle ssh? I'm trying to setup my home network and I'm fuzzy on that.
<mightybyte>
Never mind, I figured out my question. I had to use services.nginx.appendHttpConfig.
<makefu>
mightybyte: there is an option for enabling gzip compression
<mroutis>
andi-: sadly, I'm having the same results
<mroutis>
I can control the backlight, tho hahahah
wpcarro has quit [Ping timeout: 244 seconds]
hamishmack has joined #nixos
knupfer has quit [Remote host closed the connection]
<mroutis>
thanks, anyways!
tzemanovic has joined #nixos
wpcarro has joined #nixos
wpcarro has quit [Ping timeout: 264 seconds]
Ariakenom has quit [Quit: Leaving]
mkoenig has joined #nixos
sbdchd has joined #nixos
TrentP has quit [Ping timeout: 240 seconds]
luketopia has joined #nixos
TrentP has joined #nixos
aanderse-laptop has quit [Ping timeout: 265 seconds]
wpcarro has joined #nixos
sir_guy_carleton has joined #nixos
asymmetric has joined #nixos
m0rphism has quit [Quit: WeeChat 2.0]
sam__ has joined #nixos
<worldofpeace>
How am I supposed to handle ssh for existing machines for nixops?
<clever>
worldofpeace: when using the none backend, nixops will add its own keys to the config, so you just need an ssh agent or pw for the first deploy
<worldofpeace>
clever: ah ok, let's see if I understood that in a second
wpcarro has quit [Ping timeout: 248 seconds]
achristianson has joined #nixos
wpcarro has joined #nixos
mroutis has quit [Ping timeout: 256 seconds]
sbdchd has quit [Remote host closed the connection]
blonkhart has quit [Ping timeout: 244 seconds]
blonkhart has joined #nixos
wpcarro has quit [Ping timeout: 244 seconds]
wpcarro has joined #nixos
mroutis has joined #nixos
juhe has quit [Quit: Leaving.]
lambdamu has quit [Ping timeout: 260 seconds]
wpcarro has quit [Ping timeout: 256 seconds]
tzemanovic has quit [Remote host closed the connection]
tzemanovic has joined #nixos
wpcarro has joined #nixos
sw4n has quit [Ping timeout: 260 seconds]
achristianson has quit [Remote host closed the connection]