<ottidmes>
otherwise try just modesetting, that should always work on modern GPUs as I understand it (its actually my driver being used on my desktop right now)
<thedavidmeister>
x11 also hangs when i try to power down
<clever>
thedavidmeister: what does it mention in the journal?
<clever>
journalctl -u display-manager
<ottidmes>
thedavidmeister: not some rights issue? I tend to forget to chown my home folder after setting up my file systems, just because the user is only made when I run my actual NixOS config
<thedavidmeister>
i'm root
<thedavidmeister>
it says starting
<thedavidmeister>
started
<clever>
it should also include logs for the whole session
<thedavidmeister>
main process exited, code=exited, status=1/FAILURE
<thedavidmeister>
service rstartsec=200ms expired, scheduling restart
<thedavidmeister>
then it stops
<thedavidmeister>
and restarts
<thedavidmeister>
and loops til it fails
<ottidmes>
clever: BTW, I am now trying my last plan, create my own socket (/dev/log is a socket), and mount bind /dev/log to my own socket that filters the data before writing it to the actual /dev/log :P
<{^_^}>
#49840 (by markuskowa, 1 week ago, open): nixos/postgresqlBackup: set to umask to 0077
<clever>
ottidmes: :D
<thedavidmeister>
verify did not show problems
<thedavidmeister>
i'll try modesetting
<ottidmes>
thedavidmeister: about being root, I meant, if you have like user david, make absolutely sure your /home/david is completey owned by david, those .Xauthority files and such, can cause it to die
<clever>
thedavidmeister: what display manager did you configure?
<clever>
thedavidmeister: there should be a lot more in the journal then what you listed above
<{^_^}>
[nixpkgs] @samueldr merged pull request #49840 → nixos/postgresqlBackup: set to umask to 0077 → https://git.io/fpfFm
<ottidmes>
clever: I even tried checking what strings where in the files, maybe one contained /dev/log, did find some others, like /dev/zero and /dev/urandom, but no /dev/log
dycan has joined #nixos
lassulus_ has joined #nixos
<clever>
ottidmes: run nix-store -qR on the storepath, then grep ALL of those for /dev/log
<thedavidmeister>
display manager is lightdm
<clever>
thedavidmeister: was there anything else in the logs that you omited?
<thedavidmeister>
desktop manager is plasma5
<thedavidmeister>
nope
<thedavidmeister>
just what i put on repeat until it died
<{^_^}>
[nixpkgs] @samueldr pushed commit from @markuskowa to release-18.09 « nixos/postgresqlBackup: set to umask to 0077 »: https://git.io/fpOav
<clever>
thedavidmeister: can you pastebin the entire hardware-configuration and configuration.nix files?
<thedavidmeister>
how would i do that?
<aanderse>
samueldr: awesome!
<clever>
thedavidmeister: ssh into the laptop from another system, scp can copy files, gist can upload from the CLI
stepcut has quit [Read error: Connection reset by peer]
vidbina_ has quit [Ping timeout: 268 seconds]
stepcut has joined #nixos
<{^_^}>
[nixpkgs] @fpletz opened pull request #50340 → nixos/switch-to-configuration: restart changed socket units → https://git.io/fpOrL
vk3wtf has joined #nixos
drakonis has quit [Read error: Connection reset by peer]
<thedavidmeister>
samueldr++
<{^_^}>
samueldr's karma got increased to 38
<thedavidmeister>
ottidmes++
<{^_^}>
ottidmes's karma got increased to 1
<thedavidmeister>
clever++
<{^_^}>
clever's karma got increased to 47
<clever>
thedavidmeister: aha
<clever>
thedavidmeister: let me see what kernel i'm on...
<clever>
thedavidmeister: Linux system76 4.14.67 #1-NixOS SMP Fri Aug 24 11:09:23 UTC 2018 x86_64 GNU/Linux
ottidmes has quit [Ping timeout: 250 seconds]
Zer000 has joined #nixos
maximiliantagher has joined #nixos
<Zer000>
So I'm still trying to figure out why upgrading my server from 18.03 to 18.09 caused gitlab to wipe my database. The postgres version was not changed from pkgs.postgresql100. The databsase itself still has all my data from hydra, it was just the gitlab db that was seemingly dropped and re-created fresh. clever suggested that there is another directory that was being used before but I don't think that's it.
maximiliantagher has quit [Ping timeout: 240 seconds]
freeman42x]NixOS has quit [Remote host closed the connection]
<Zer000>
I have files in /var/lib/postgresql/10.0/base modified Nov12, /var/lib/postgresql/9.4/base modified Apr 18 (probably when I upgraded to 10 because i have a db dump dated Apr 17) but I also have files in /var/db/postgresql/base dated Apr 17
dycan has quit [Remote host closed the connection]
<Zer000>
I guess I'm going to just restore the dump from April since these are just personal projects I don't think there would be comments or anything missing. But this is a shitshow of an upgrade
<simpson>
Yeah, that sounds rough.
rprije has quit [Ping timeout: 252 seconds]
sirkha has joined #nixos
sirkha has quit [Client Quit]
sirkha has joined #nixos
<sirkha>
Hi, weird problem. just upgraded to 18.09 and in my user profile i try nix-env --upgrade and get error: anonymous function at /nix/store/43vrvm0isa550h06c43ckx935hzqd6mp-nixos-18.09.1228.a4c4cbb613c/nixos/pkgs/applications/misc/qpdfview/default.nix:1:1 called with unexpected argument 'qt4', at /nix/store/43vrvm0isa550h06c43ckx935hzqd6mp-nixos-18.09.1228.a4c4cbb613c/nixos/lib/customisation.nix:69:12
<sirkha>
quick google didn't help, don't know where to start troubleshooting
<lopsided98>
sirkha: do you override qpdfview anywhere in your config?
<sirkha>
hmm, its possible (been a while since I installed it)
<lopsided98>
qpdfview was moved from qt4 to qt5 which explains why the qt4 argument no longer exists
<sirkha>
i tried uninstalling qpdfview and that doesn't change anything... probably because it is overridden somewhere and this is keeping my nix-env from evaluating properly
<clever>
sirkha: do you have a ~/.nixpkgs/config.nix or ~/.config/nixpkgs/config.nix?
orivej has quit [Ping timeout: 268 seconds]
<sirkha>
clever, yes, thanks
<sirkha>
removed override, trying again
<sirkha>
yup, that worked, thanks
maximiliantagher has joined #nixos
freeman42x]NixOS has joined #nixos
iyzsong has joined #nixos
maximiliantagher has quit [Ping timeout: 244 seconds]
wykurz has joined #nixos
jperras has quit [Ping timeout: 272 seconds]
iyzsong has quit [Ping timeout: 250 seconds]
stepcut has quit [Remote host closed the connection]
maximiliantagher has joined #nixos
maximiliantagher has quit [Ping timeout: 240 seconds]
jperras has joined #nixos
vk3wtf has quit [Ping timeout: 252 seconds]
<thedavidmeister>
ok so next question...
<thedavidmeister>
what would cause my x11 to hang on shutdown?
<romanofskiWork>
thedavidmeister: some process still reading from a socket or whatever?
<romanofskiWork>
if the process is still around you could strace it and see what's happening
freeman42x]NixOS has quit [Ping timeout: 268 seconds]
<romanofskiWork>
strace -p <pid>
maximiliantagher has joined #nixos
<thedavidmeister>
how do i check that if it is hanging on shutdown?
<thedavidmeister>
how do i run commands at that point?
jperras has quit [Ping timeout: 240 seconds]
<Rusty1>
ctl-alt-f1 perhaps?
maximiliantagher has quit [Ping timeout: 246 seconds]
<fresheyeball>
that is what I get out to the command line when I try to run a game
<fresheyeball>
Looks more and more like I need to learn about this ELF thing
maximiliantagher has joined #nixos
jperras has quit [Ping timeout: 268 seconds]
maximiliantagher has quit [Ping timeout: 246 seconds]
endformationage has quit [Ping timeout: 252 seconds]
jperras has joined #nixos
reinhardt has quit [Quit: Leaving]
<romanofskiWork>
fresheyeball: is it possible that you created the steam client installation on a Ubuntu machine and switched distros?
vmandela has joined #nixos
<fresheyeball>
romanofskiWork: no, the client was installed via nix
<fresheyeball>
on a nixos install
<romanofskiWork>
did you start the client on a fresh home directory tho?
<romanofskiWork>
e.g. did the ~/.steam folder exist before you installed nixos
<fresheyeball>
I formatted the drive and installed nixos on it 2 years ago
fractal has quit [Ping timeout: 264 seconds]
<fresheyeball>
I have been running nixos on this drive ever since
freeman42x]NixOS has joined #nixos
<romanofskiWork>
hm odd.. you could try: unset LD_PRELOAD; <your steam client or path to game> from the terminal
<romanofskiWork>
something adds the libraries to be loaded during the start of the game via the LD_PRELOAD env variable... gotta find where
jperras has quit [Ping timeout: 240 seconds]
<romanofskiWork>
but might be on the wrong food there... so not sure
jperras has joined #nixos
Watcher7 has quit [Quit: h]
boegel has quit [Remote host closed the connection]
jasongrossman has quit [Ping timeout: 244 seconds]
boegel has joined #nixos
jperras has quit [Ping timeout: 268 seconds]
divansantana has joined #nixos
<divansantana>
Greetings all :)
jD91mZM2 has joined #nixos
maximiliantagher has joined #nixos
mkoenig has quit [Remote host closed the connection]
vk3wtf has joined #nixos
jordansinn has joined #nixos
maximiliantagher has quit [Ping timeout: 246 seconds]
Guanin_ has joined #nixos
wykurz has quit [Quit: Leaving]
Itkovian has joined #nixos
Lisanna has joined #nixos
<Lisanna>
So, nix-store has a --repair-path option, which attempts to substitute the path from a binary substituter. Is there a technical reason why it couldn't just rebuild the path as a 'fallback' repair method?
<Lisanna>
You'd think that would work if the build was deterministic
<Lisanna>
context: *attempts to substitute the path from a binary substituter if the existing one in the /nix/store is corrupted in some way (hash mismatch, size mismatch)
hyper_ch2 has joined #nixos
reinhardt has joined #nixos
<divansantana>
Nix is very cool. Was up and running with it on a foreign distro quickly. I've installed google-chrome and qutebrowser. google-chrome has a font issue. Wondering if I have to install some font pkg for it to work. qutebrowser crashes on launch too. https://paste.debian.net/1051652/ Any ideas?
<romanofskiWork>
divansantana: perhaps Xorg running without 3D accelleration?
Maxdamantus has quit [Ping timeout: 276 seconds]
<romanofskiWork>
not sure if it should run without
<divansantana>
romanofskiWork: Hmm. I'm on guixsd. Not sure. I can have a look into that. If so, would one have to modify the nix recipe/install to work on guixsd?
shabius has joined #nixos
<angerman>
how do you figure out why your download (only some) in a sandbox fail with `nix-build`?
<romanofskiWork>
divansantana: yeah not sure.. you can check if you have 3D accelleration in a terminal by running: glxinfo
shabius_ has quit [Ping timeout: 264 seconds]
<Lisanna>
I just realized... I can use my nix-shell "trick" I discovered yesterday to repair corrupted /nix/store paths that way
<romanofskiWork>
perhaps you need to install the Xorg driver package for your graphics card ...
<Lisanna>
screw it, I'll write a whole tool to do it
<divansantana>
romanofskiWork: Cool. Let me look into this. Will report back.
<divansantana>
romanofskiWork: thanks
lopsided98 has quit [Ping timeout: 252 seconds]
lopsided98 has joined #nixos
maximiliantagher has joined #nixos
<divansantana>
glxinfo | grep rendering, says direct rendering: Yes . glxgears works too. 60fps.
<divansantana>
gm152 has quit [Ping timeout: 246 seconds]
jasongrossman has joined #nixos
Ridout has quit [Ping timeout: 268 seconds]
maximiliantagher has quit [Ping timeout: 252 seconds]
<srhb>
yurb: (The more general answer is: It's fine to ping here if things get forgotten for a while) :-) Thanks for contributing
smolboye has quit [Ping timeout: 272 seconds]
tommyangelo has quit [Ping timeout: 252 seconds]
peacememories has joined #nixos
smolboye has joined #nixos
jperras has quit [Ping timeout: 272 seconds]
peacememories has quit [Client Quit]
maximiliantagher has joined #nixos
jperras has joined #nixos
Diagon has quit [Quit: Leaving]
maximiliantagher has quit [Ping timeout: 240 seconds]
tommyangelo has joined #nixos
carlosdagos has quit [Quit: Connection closed for inactivity]
<Lisanna>
ekleog I mean, since time is an impurity
<ekleog>
Lisanna: oh, you mean changing gettimeofday? it may be possible to do through seccomp&ptrace, ISTR there's a tool to run the subtool in time fixed to 0
<ekleog>
like “faketime my_command” or something like that
<ekleog>
now, I don't think linux disables rdtsc so it wouldn't be perfect
<ekleog>
(unless you start hot-patching rdtsc calls out of memory, but that starts getting *really* hacky)
<Lisanna>
I'm just looking for something to deal with annoying build scripts that feel that stamping the output of 'date' into a binary so that you can know when the build was done is more important than determinism
<Lisanna>
or equivalents of date
<NickHu>
What's the correct way to get into a nix-shell with a gcc arm cross compile toolchain?
<ekleog>
that doesn't it nix-sandbox-wide, though… maybe this could be requested as a feature, idk
<NickHu>
I tried nix-shell '<nixpkgs>' --arg crossSystem '(import <nixpkgs> {}).lib.systems.examples.armv7l-hf-multiplatform' -A stdenv.cc.cc but it seems to give me everything I need, including a linker, except for gcc itself
<ekleog>
Lisanna: now, I guess some compilers may react strangely if their time is frozen to 0
<ekleog>
see eg. pkgs/applications/audio/axoloti/default.nix for a use of it
<Lisanna>
it can be used to deal with errors like "hash mismatch importing path"
<Lisanna>
ekleog maybe a build hook
<ekleog>
Lisanna: totally possible too, indeed, most likely cleaner :)
<ekleog>
NickHu: for gcc there's gcc-arm-embedded
civodul has joined #nixos
<NickHu>
I figured out it's nix-shell '<nixpkgs>' --arg crossSystem '(import <nixpkgs> {}).lib.systems.examples.armv7l-hf-multiplatform' -A gcc.cc that seems to give me what I need
<symphorien>
NickHu: nix-shell -A foo gives you the dependies needed for A, not A itself
<symphorien>
for foo, not foo itself*
<NickHu>
symphorien: Ah, I see - so is there a better invocation?
<symphorien>
put what you want to be available in the (native)buildInputs
<symphorien>
the actual content of phases will be ignored
jperras has quit [Ping timeout: 250 seconds]
<ekleog>
NickHu: actually for a cross-compile toolchain you're going to want targetPlatform, not crossSystem
<ekleog>
crossSystem is “runs on crossSystem”
<ekleog>
the reason why it works with -A is that it requires a localSystem->crossSystem compiler to compile the one you asked
<NickHu>
I see
<NickHu>
When I actually tried to use the compiler I got an error: fatal error: linux/compiler-gcc7.h: No such file or directory
__monty__ has joined #nixos
<symphorien>
NickHu: you got the dependencies to build stdenv.cc
<ekleog>
NickHu: what exactly are you wanting to compile for?
<symphorien>
that is stdenv.cc.cc which is gcc-unwrapped
<symphorien>
so this gcc does not know where to find the libc and so on
<NickHu>
ekleog: I'm trying to compile an arm linux kernel for a raspberry-pi-eqsue device
<puffnfresh>
zimbatm: I might upgrade my VM to Mojave and give it a go
<ekleog>
NickHu: have you tried just `-p gcc-arm-embedded`? if it's going to run without runtime (as I think the arm linux kernel is going to), then it'd maybe do it
<ekleog>
gcc upstream is missing some stuff IIRC
jperras has joined #nixos
<NickHu>
ekleog: It gives a similar error: fatal error: linux/compiler-gcc6.h: No such file or directory
<ekleog>
NickHu: there's nix-shell -I nixpkgs=channel:nixos-unstable -p 'gdb.override {stdenv = stdenv.override { targetPlatform = lib.systems.examples.riscv32; };}' that I once used to build a gdb for riscv32
<ekleog>
(the -I nixpkgs=channel:nixos-unstable being most likely unnecessary if you're already on unstable)
<ekleog>
symphorien: it needs be targetPlatform, not crossSystem
<symphorien>
huh maybe, the point being using -A stdenv.cc won't work
<ekleog>
(or even on 18.09, at the time I was 18.03)
<ekleog>
NickHu: you may be more lucky by trying to directly ask nix to build your kernel with crossSystem=..., btw
<NickHu>
symphorien: nix-shell '<nixpkgs>' --arg targetSystem '(import <nixpkgs> {}).lib.systems.examples.armv7l-hf-multiplatform' -A linux.dev doesn't seem to work
<symphorien>
how does it not work ?
<NickHu>
symphorien: It doesn't give me any arm-*-eabi-* binaries
<NickHu>
The only thing that seemed to was stdenv.cc.cc or gcc.cc
<symphorien>
NickHu: with crossSystem instead of targetSytem, it downloads armv7a-unknown-linux-gnueabihf-gcc
tommyangelo has quit [Remote host closed the connection]
<symphorien>
I think you mixed targetSystem and targetPlatform, actually
<symphorien>
mh targetPlatform does not work better
maximiliantagher has quit [Ping timeout: 260 seconds]
vidbina_ has joined #nixos
<zimbatm>
puffnfresh: let me know how it goes. do you want access to the repo? the staging branch is built on hydra
<NickHu>
symphorien: Ah, I do indeed get armv7l-unknown-linux-gnueabihf-, but I still end up getting the fatal error: linux/compiler-gcc7.h: No such file or directory
<ekleog>
NickHu: did you try compiling with gcc6? maybe your kernel sources don't support gcc7 (I can never remember which kernels support which compilers)
jperras has quit [Ping timeout: 264 seconds]
alienpirate5 has quit [Ping timeout: 250 seconds]
<NickHu>
ekleog: Any idea how to get into a shell with arm gcc6 (preferably wtihout having to compile gcc myself)?
<puffnfresh>
zimbatm: oh yeah access to the repo would be good
<NickHu>
And when I used gcc-arm-embedded it gave me the same error anyway, with 6 instead of 7
<puffnfresh>
thanks for setting up Hydra
<NickHu>
Since then I did a nix-channel --update and it looks like gcc-arm-embedded is gone now...
<ekleog>
that's definitely not normal
spluko has joined #nixos
jperras has joined #nixos
<symphorien>
NickHu: to change the compiler you needto use overrideCC stdenv gcc6 instead of stdenv iirc
<zimbatm>
puffnfresh: invite sent!
<symphorien>
but with linux it is complicated because linuxPackages is a scope
<Taneb>
Why might my hydra server not be running builds even though the queue is not empty?
Mateon3 has joined #nixos
iyzsong has joined #nixos
Mateon1 has quit [Ping timeout: 272 seconds]
Mateon3 is now known as Mateon1
erictapen has joined #nixos
<spluko>
Can anyone tell me how I get avr-gcc now? package nixos.avrgcc has gone from unstable.
<symphorien>
spluko: the commit message suggests pkgsCross.avr.buildPackages.gcc
jperras has joined #nixos
<Taneb>
hydra-queue-runner says something like "priorty of build 54986 increased; checking the queue for builds > 56382" when I bump one of the not-building builds
<NickHu>
ekleog: how might I go about asking nix to build my kernel directly?
iyzsong has quit [Ping timeout: 268 seconds]
iyzsong has joined #nixos
jasongrossman has quit [Ping timeout: 244 seconds]
<flokli>
infinisil: is that nix further down the chain trying to add it to a path?
<flokli>
can we get it to work somehow? I don't see any usages of writeTextDir with slashes in nixpkgs, but can imagine some usecases for it
<NickHu>
ekleog: I tried nix-shell '<nixpkgs>' --arg crossSystem '(import <nixpkgs> {}).lib.systems.examples.armv7l-hf-multiplatform' -A linuxPackages.kernel but it still gives the gcc7 error
<NickHu>
fails to find #include gcc_header(__GNUC__)
<ekleog>
NickHu: did you try with nix-build? that'd help knowing whether the problem is with your sources or the crossSystem
<ekleog>
NickHu: so it's now time to read the source of the derivation for `linuxPackages.kernel` and figure out what changes from your own build :D now, the alternative if you just want to build a patched kernel once (and don't need incremental rebuilds) is to just add a linux kernel patch to linuxPackages.kernel and use nix-build only
__monty__ has quit [Read error: Connection reset by peer]
__monty__ has joined #nixos
<NickHu>
ekleog: It's some out of tree custom kernel because the board isn't well supported... I guess I'll just start up a ubuntu vm and compile it or something, should only need to do it once
jperras has quit [Ping timeout: 268 seconds]
<NickHu>
I'm fairly sure this used to work with no problem on nixos when I tried to use the codesourcery toolchain
<NickHu>
(Which seems to have disappeared)
alienpirate5 has joined #nixos
<ekleog>
NickHu: you can `diff` between the out-of-tree kernel and the linux sources by your linuxPackages
<ekleog>
and use the diff as a patch
<ekleog>
and yeah, cross- is in a big refactor since last year or so, so things are stabilizing currently
nuncanada has joined #nixos
<ekleog>
though your issue appears not really cross-related, tbh
iyzsong has quit [Read error: Connection reset by peer]
primeos has quit [Quit: WeeChat 2.1]
primeos has joined #nixos
iyzsong has joined #nixos
orivej has joined #nixos
<MasseR>
I know how to override haskell packages, but if I wanted to override the compiler (ghcjs from a different commit hash), how would I go about that?
<MasseR>
I would prefer staying in a self contained nix file (so no ~/.config/nixpkgs/config.nix)
ottidmes has joined #nixos
<{^_^}>
[nixpkgs] @DzmitrySudnik opened pull request #50348 → exhibitor: fix paths for zookeeper shell scripts → https://git.io/fp3ID
Lisanna has quit [Quit: Lisanna]
crmlt has joined #nixos
vidbina_ has quit [Ping timeout: 264 seconds]
drakonis has joined #nixos
maximiliantagher has joined #nixos
maximiliantagher has quit [Ping timeout: 252 seconds]
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
jD91mZM2 has quit [Ping timeout: 260 seconds]
jD91mZM2 has joined #nixos
<symphorien>
you can eschew using ~/.config/nixpkgs/config.nix by using import <nixpkgs> { config = content of config.nix; } instead of import <nixpkgs> {}
Itkovian has quit [Read error: Connection reset by peer]
hedning has joined #nixos
<ottidmes>
symphorien: I do it the other way around, I set a different nixpkgs/config.nix location via NIXPKGS_CONFIG that points to a .nix file in my nixos config directory, and then in my nixos config I do: nixpkgs.config = import /my/nixpkgs/config.nix;
maximiliantagher has joined #nixos
juliendehos has quit [Quit: Leaving]
__Sander__ has joined #nixos
dbmikus_ has joined #nixos
maximiliantagher has quit [Ping timeout: 264 seconds]
periklis has quit [Ping timeout: 244 seconds]
klntsky has joined #nixos
nschoe has joined #nixos
locallycompact has joined #nixos
<locallycompact>
ls
<locallycompact>
if I just have a loose stack script I'm running as an executable how do I avoid this error error: attribute 'ghc862' missing, at (string):1:43
maximiliantagher has joined #nixos
<ottidmes>
locallycompact: thats not enough context, but my guess would be that ghc862 is defined only in a nixpkgs channel/checkout thats not your default, so you probably have to make explicit to use a particular channel/checkout where it is defined
<infinisil>
Stack uses NIX_PATH, and uses the ghc version it gets from the resolver to determine the nix ghc to use
iyzsong has quit [Ping timeout: 250 seconds]
swflint has quit [Ping timeout: 252 seconds]
swflint has joined #nixos
<locallycompact>
Why does stack care what host I'm on
<locallycompact>
it doesn't use host ghc on other systems
<locallycompact>
I thought the whole point of stack was to sandbox ghc
<locallycompact>
and libraries
<locallycompact>
when I pull a resolver on ubuntu/arch it just downloads whatever ghc I need
<cocreature>
locallycompact: stack has nix integration that is enabled by default on nixos and that tries to get GHC via nix (since the regular bindists don’t work on nixos)
<dramforever>
I have a script on this non-NixOS box in /usr/lib/systemd/system-sleep that reloads my ethernet drivers on resume. What's the equivalent on NixOS?
<dramforever>
From the source I think the option I mentioned just creates a new service and orders it to run on resume
<azazel>
dramforever: also it may be worthwhile to check the nixos-hardware repo for a configuration that supports your specific hardware
<dramforever>
Wow
<dramforever>
No, nothing for me
<divansantana>
Next, how to get google chromes fonts to not look crazy. Any clues? :)
camsbury has quit [Quit: Leaving...]
o1lo01ol1o has joined #nixos
maximiliantagher has joined #nixos
<dramforever>
I have no idea. This box is currently running Arch, and it didn't need that reload driver thing last week
<dramforever>
So maybe the specific set of software in NixOS also works just fine
<dramforever>
In any case, it's nice to know that if nixos options doesn't cover my needs, someone here will
<dramforever>
:)
endformationage has joined #nixos
vmandela has quit [Quit: Leaving]
maximiliantagher has quit [Ping timeout: 272 seconds]
<{^_^}>
[nixpkgs] @DzmitrySudnik opened pull request #50352 → exhibitor: fix paths for zookeeper shell scripts to point to local fo… → https://git.io/fp3Zc
<manveru>
zimbatm: anyway, i made the new design as far as i could without the psd... lemme know when you finally get it :)
drakonis has quit [Quit: WeeChat 2.3]
<manveru>
zimbatm: and while it's not a 1:1 copy (i suck at that apparently), it's at least responsive and pretty minimal in dependencies
hyper_ch2 has quit [Quit: Page closed]
Ariakenom has quit [Ping timeout: 264 seconds]
cfricke has quit [Quit: WeeChat 2.3]
<{^_^}>
[nixpkgs] @copumpkin merged pull request #50352 → exhibitor: fix paths for zookeeper shell scripts to point to local fo… → https://git.io/fp3Zc
<{^_^}>
[nixpkgs] @copumpkin pushed 2 commits to release-18.09: https://git.io/fp38o
<mog>
im not sure what i did or what changed but a lot of things that use dbus seemed to stop working on my nixos box
<mog>
so i cant open blueman-manger, or issue a reboot, suspend etc. both reboot and suspend work via sudo though
<WilliamHamilton[>
does the package alsaLib expose or use the package portmidi?
<WilliamHamilton[>
I'm not sure which attribute of alsaLib I should look in `nix repl` to answer this question
<jabranham>
yurb: yes, all emacs packages on MELPA are available through nix
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<codygman>
How can I boot to /bin/sh in nixos? I tried just changing the init to `init=/bin/sh` but it booted normally into my "fsck needs to be run manually error". I need to boot into a shell and run fsck.
<yurb>
jabranham: thanks.
<NickHu>
So today I was trying to install nix on a fedora machine on which I don't have root access
<NickHu>
I learned about nix v2 which supports arbitrary store paths by user namespaces
<NickHu>
But unfortunately there's no nix binaries for fedora anymore
maximiliantagher has joined #nixos
<NickHu>
Is there any way to get a nix v2+ binary without much hassle?
shabius has joined #nixos
<qyliss^work>
NickHu: doesn't the install script do that?
shabius has quit [Max SendQ exceeded]
maximiliantagher has quit [Ping timeout: 272 seconds]
<clever>
if the cache claims a given file exists, nix will hard fail when you GC that file from the cache
<clever>
and you must wipe the local cache
<aminechikhaoui>
so you can disable that by setting it to 0
<aminechikhaoui>
just for one command for example
Ariakenom has quit [Ping timeout: 246 seconds]
<ottidmes>
clever: lame as it is, the only thing that ended up working for me is to just have a global mount --bind for /dev/log to a socket that filters out all lines of that spamming program
<coconnor>
morning all!
<ottidmes>
coconnor: good evening! :P
johanot has quit [Quit: WeeChat 2.2]
<clever>
ottidmes: you could probably also just overwrite the /dev/log symlink
<coconnor>
hehe I wonder what portion of NixOS is US vs EU? Seems EU is good 3/4
rihardsk has joined #nixos
<Taneb>
coconnor: I guess there's a reason NixCon has been in Germany twice and the UK once ;)
<coconnor>
haha aye
<ottidmes>
clever: I prefer the bound, less mutable, that way if I disable the program in NixOS, /dev/log remains functional
locallycompact has quit [Remote host closed the connection]
asymmetric has joined #nixos
<clever>
ottidmes: yeah, you would have to undo the symlink when turning the filter program off
<asymmetric>
is there any eta for when the staging branch will be merged into master?
<clever>
ottidmes: also of note, unix sockets allow you to query the PID of the remote process
<clever>
ottidmes: that both helps (you can verify the binary of a msg when filtering), and harms (journald now thinks your filter prog is the only thing logging)
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<ottidmes>
clever: Hmm, that is better than my regex approach right now, I will look into it, thanks :)
<ottidmes>
clever: the only other program so far writing to /dev/log is sudo, had not expected that
<NickHu>
qyliss^work: Yeah, it's a little bit painful because I'd have to build all the dependencies manually too
<clever>
NickHu: the NIX_REMOTE will tell the local nix, to chroot into /home/clever/rootfs
<clever>
NickHu: NIX_CONF_DIR, NIX_LOG_DIR, NIX_STORE, NIX_STATE_DIR, all mutate where the local nix thinks things are, for its next build
<clever>
so a standard nix build, can act on /home/clever/nix/store
<clever>
and due to the chroot, its actually /home/clever/rootfs/home/clever/nix/store/
<clever>
then i use that standard nix build, to build nix, with a .override, to bake those changes into the resulting binary
<NickHu>
wow
<clever>
so now you copy /home/clever/rootfs/home/clever/nix to /home/clever/nix on a remote machine
<clever>
and it will just work
<clever>
put the same override into your local config.nix, to ensure nix doesnt un-patch itself, and install anything else you want
<clever>
Unode: cc ^^^
<clever>
NickHu: oh, some of those clever's should be example's
<clever>
it expects the username to be clever on the build system, and example on the target
<clever>
and its a var (line 3) so it can target another user easily
<NickHu>
Ok, so I can build this on my existing nixos system and then transfer it over
Tucky has quit [Quit: WeeChat 2.2]
<clever>
exactly
<NickHu>
are the parameters for stateDir and confDir relatively standard?
<NickHu>
I don't know how it interacts with the --store flag of nix
<clever>
i believe they are all standard
<NickHu>
Ok, great, I'll try it out, thanks!
<NickHu>
While you're here, do you happen to know anything about the current state of the arm cross compile toolchain?
<NickHu>
I'm trying to compile the linux kernel for a raspberry pi-like device and can't seem to get a working toolchain
<NickHu>
The closest I got was with nix-shell '<nixpkgs>' --arg targetPlatform '(import <nixpkgs> {}).lib.systems.examples.armv7l-hf-multiplatform' -A linux.dev, which seemed to give me the arm gcc compiler but then it couldn't find some gcc headers
<NickHu>
erm, that should be nix-shell '<nixpkgs>' --arg crossSystem '(import <nixpkgs> {}).lib.systems.examples.armv7l-hf-multiplatform' -A linux.dev
<Mic92>
asymmetric: pavucontrol is not strictly required some users may use the plasma widget or pavucontrol-qt.
<NickHu>
And it gives include/linux/compiler-gcc.h:106:1: fatal error: linux/compiler-gcc7.h: No such file or directory
<clever>
NickHu: not sure on that
<asymmetric>
Mic92: makes sense. but would make for a good option, don't you think?
<NickHu>
clever: Thanks anyway, maybe I'll file an issue for help
c0bw3b_ has joined #nixos
hyperfekt has joined #nixos
johnw_ has joined #nixos
johnw has quit [Ping timeout: 252 seconds]
nschoe has quit [Ping timeout: 244 seconds]
<asymmetric>
hey, i'm trying to write a default.nix where i override a dependency that is too old in nixpkgs, can anyone help me figure out why it doesn't build?
<pie_>
reboots seem to hang at some point in shutdown
<pie_>
i just tried to switch users in kde and that seems stuck
<{^_^}>
[nixpkgs] @Mic92 pushed commit from @avnik to release-18.09 « openarena: fix loading libGL.so with libglvnd »: https://git.io/fp3a8
<ottidmes>
pie_: doubt it, I am using NixOS on an UEFI system without any problems
<fresheyeball>
\leave
<samueldr>
pie_: I only have used nixos on uefi systems and... there shouldn't be anything specific to UEFI in nixos :/
fresheyeball has left #nixos ["WeeChat 2.2"]
<qyliss>
NixOS works fine on the several UEFI systems I’ve used
<pie_>
ottidmes, ok thanks. ugh. idk why im running into so many issues
<samueldr>
at least, nothing not present in other distros, except the bootloady parts, but by the time the kernel is started, it should be the same
<ottidmes>
pie_: did you try the latest kernel
<pie_>
ottidmes, idk? im on 18.09
FRidh has quit [Quit: Konversation terminated!]
asymmetric has quit [Ping timeout: 264 seconds]
<samueldr>
defaults are always "current LTS at release", which was 4.14
<ottidmes>
pie_: some modern hardware need things that are put in after 4.14 (default of 18.09, the LTS version)
<samueldr>
you can try 4.18 (and keep being updated to the latest) using `boot.kernelPackages = pkgs.linuxPackages_latest;`
sanscoeur has joined #nixos
<pie_>
also i tried to install virtualbox with oracle extensions and somehow it keeps mismatching the kernel version and the userspace version so its not working for me
<pie_>
samueldr, ok ill try that
<samueldr>
(though there's no guarantees it'll fix anything in your case :/)
<pie_>
samueldr, any reason thats not the default?
<samueldr>
LTS releases are updated for longer and "stabler"
<samueldr>
as such, 4.18 will stop being updated during the lifetime of 18.09, keeping a stable kernel version is probably better
<ottidmes>
latest tends to be most stable on recent hardware, LTS is better on older hardware (in my experience)
<dtz>
I don't rem the details but FWIW I use nixos w/uefi on an xps 15 and works great :). So, have hope! I think info such as this[1] contains everything I encountered and more: [1] https://github.com/NixOS/nixos-hardware/issues/24
<{^_^}>
nixos-hardware#24 (by domenkozar, 1 year ago, open): dell/xps/15-9560: add
<dtz>
big show-stopper was needing to set kernel params, forget if that's still needed w/latest kernel or not :3 sorry
<samueldr>
don't know about "LTS better on older hardware", but definitely sure it's better not to change major versions in stable updates :)
<pie_>
dtz, thanks
<pie_>
dtz, time to test the reporducibility of nixos ;P
<pie_>
otoh idk about my bios settings, etf
<pie_>
*etc
orivej has quit [Ping timeout: 244 seconds]
<pie_>
well, technically uefi settings right?
<pie_>
what would i lose by going back to bios?
<samueldr>
not sure, but it may not be possible if it's recent enough
<samueldr>
(and I wouldn't bet on it fixing anything)
sirkha has quit [Remote host closed the connection]
jperras has quit [Quit: WeeChat 2.2]
<ottidmes>
samueldr: I have had to downgrade back to a LTS version multiple times on my 9 year old hardware, it was not that it broke completely, but things like suspend/hibernate did break a few times I tried, but that kind of breakage is often fixed quickly enough. Just recently it caused a less stable ethernet connection
<samueldr>
ottidmes: yeah, breakage can happen :) mostly the wholesale "better on older hardware" isn't necessarily right, though it's a good thing LTS releases happen for those cases; it hopefully can be relied on in those situations
polman has joined #nixos
<ottidmes>
samueldr: hence the disclaimer "in my experience" ;) but indeed, good thing they exist, I had to use 4.4 for quite some time due to a GPU driver that only worked max that version of Linux
<cocreature>
has anyone managed to launch a digitalocean droplet with nixos using custom images? I tried to create an image using the nova infrastructure in nixpkgs but my droplet seems to be stuck at "Booting the kernel"
<tilpner>
cocreature - Yes, stand by while I find it
<bpye>
I imagine it should be possible to have an image that lets you specify your configuration.nix as the cloud-data
<azazel>
is there a way to pass additional arguments to a configuration module other than using lib/eval-config.nix and its extraArgs parameter?
<infinisil>
azazel: The `_module.args` option
<azazel>
infinisil: I saw traces of it in lib/modules.nix, but it's not clear how it should be used, maibe I'm missing the docs somewhere?
<azazel>
s/maibe/maybe
<infinisil>
azazel: Just set the option, e.g. `_module.args.foo = "foo"`
<infinisil>
Which will pass foo as an argument to every module
<azazel>
infinisil: it is to be used in a call to eval-config too?
<infinisil>
No, it's a normal option
<infinisil>
can be set in any module
<azazel>
like if I set it in a module, all the modules required by it will receive such argument?
<infinisil>
No, allll modules
zolk3ri has quit [Remote host closed the connection]
freeman42x]NixOS has joined #nixos
zolk3ri has joined #nixos
asymmetric_ has quit [Ping timeout: 245 seconds]
<azazel>
infinisil: oh... not so useful then, I have this container's configuration that I split up in "container-types" and this indirection is managed my a specific enum option in a config set. Now, for ease of developing this container types I would prefer be able to manage it like a group of modules which all receive specific /per type/ additinal parameters and so I tried to use eval-config, but it seem to have some weird interaction with
<azazel>
nixops and/or the container management which both in turn use eval-config...
<infinisil>
Use submodules then
<neonfuz>
Hello
<infinisil>
azazel: Won't be able to set nixos options directly with submodules, will need an additional layer for that, but they seem like the thing you want
<neonfuz>
I want to do some web development that is gonna have chinese content, and I want to make sure I have the fonts I would otherwise have on a windows or mac machine
<neonfuz>
do you guys know if there are any font packages that might include OS default chinese fonts from windows or mac?
<azazel>
infinisil: i get " The option `containers.<name>.extraArgs' defined in `/home/azazel....nix' does not exist" even when I commented every instance of extraArgs in the code
<neonfuz>
is there a like "msfonts" package?
<neonfuz>
I looked in pkgs/data/fonts, I didn't see anything but I might have missed it
<infinisil>
azazel: I can't help very much on such error if I don't see the code
<azazel>
infinisil: of course, of course... maybe you saw it already ;-) thanks anyway
<azazel>
infinisil: you mean using some kind of "polymorphic" submodules? (Sorry for the OO idiom :-D)
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #50363 → Fallout fixes from #50293 → https://git.io/fp3Di
<samueldr>
(right now the results are woefully out of date, but yeah, that one probably still is accurate)
<symphorien>
this index is soo useful
<infinisil>
azazel: Okay, very simple solution: Just do your container specific files like this: `container-type: { pkgs, lib, ... }:` then import them with `imports = [ (import ./file.nix "footype") ];`
* infinisil
finally understands what you want, he thinks
<neonfuz>
alright that doesn't have any chinese fonts, thanks though infinisil
<neonfuz>
that was what I was looking for though
<infinisil>
Um
<infinisil>
That wasn't me :)
hedning has quit [Quit: hedning]
<{^_^}>
[nixpkgs] @Synthetica9 opened pull request #50364 → nixos-version: use version_ instead of version → https://git.io/fp3ye
<leotaku>
I would like to deploy a nixops machine using a specific nixpkgs rev (different from the systems nixpkgs). How can I best achive that without having to redefine "$NIX_PATH"?
<{^_^}>
[nixpkgs] @Infinisil opened pull request #50365 → xpra: Mark as broken → https://git.io/fp3yd
<infinisil>
leotaku: You can just set NIX_PATH for the nixops command
<infinisil>
Even an alias `alias nixops="NIX_PATH=... nixops"` should work
<leotaku>
infinisil: Well, the problem is that I would like to deploy my home network, which contains a raspberry pi as well as an x64 server. I need to use different nixpkgs checkouts, because only "unstable-aarch64" has the aarch64 binary cache.
<infinisil>
Nixops unfortunately doesn't support different nixpkgs for different machines in the same deployment
<leotaku>
infinisil: Thanks, that looks very helpful.
<cocreature>
tilpner: got it working, thanks again!
primeos has quit [Quit: WeeChat 2.3]
<infinisil>
leotaku: I actually tried to implement something once which could solve this problem, but it turned out to be rather hard and I couldn't do it
primeos has joined #nixos
<{^_^}>
[nixpkgs] @Infinisil merged pull request #50365 → xpra: Mark as broken → https://git.io/fp3yd
<azazel>
infinisil: oh, sorry, i was strangely convinced that the attr name for specifying other config modules was "requires", it sometimes happens... with nix only. I dunno whym sorry :-)
<symphorien>
(in the article it is <nixos-unstable>, sorry)
<kingemp>
thank you very helpful.
hamishmack has joined #nixos
kingemp has quit [Remote host closed the connection]
dbmikus_ has quit [Ping timeout: 240 seconds]
maximiliantagher has joined #nixos
hedning has joined #nixos
maximiliantagher has quit [Ping timeout: 252 seconds]
<PolarIntersect>
Not asking, saying
<PolarIntersect>
oops
vidbina_ has joined #nixos
<PolarIntersect>
Does anyone know why grub would install fine w/ nixos-rebuild but for some reason systemd-boot doesn't ever update my bootloader?
<PolarIntersect>
Can't access my machine because X keeps trhing to start and failing, but it locks the screen whenever it restarts >.<
<infinisil>
PolarIntersect: Ctrl-Alt-F1 to switch to a console window
jasongrossman has quit [Quit: ERC (IRC client for Emacs 26.1)]
<the-kenny>
Ctrl-Alt-F1?
<the-kenny>
Ah :)
hedning has quit [Ping timeout: 272 seconds]
hakujin has joined #nixos
<hakujin>
bumped our pinned nixpkgs to a recent rev and now get variants of `strict.pm did not return a true value at /nix/var/nix/profiles/system/bin/switch-to-configuration line 3.`
<hakujin>
`strict.pm did not return a true value at /nix/store/9g4wc31j7a2xp22xpgwr0qssfxahxdzl-builder.pl line 3.`
<hakujin>
happy to open an issue; is this currently known?
<PolarIntersect>
samueldr: It's 1070 Mobile though
<samueldr>
nah, I was "wrong", I thought 390 was older than that
<PolarIntersect>
I can try plugging in my exgernal GPU to see if it "fixes" it
<PolarIntersect>
external*
<PolarIntersect>
I got this laptop becauseI thought the hardware looked A+ for Linux
<PolarIntersect>
very wrong haha
<clever>
PolarIntersect: what happens if you just leave the videoDrivers field unset?
<PolarIntersect>
omfg it worked
<PolarIntersect>
:O
<ottidmes>
PolarIntersect: I bought a AMD card thinking there is good support for it on Linux, little did I know it happened to be one of the few cards that has had problems ever since its release on Linux...
<samueldr>
it could be using nouveau, PolarIntersect, check the logs
<clever>
PolarIntersect: what does the log say when it worked?
<PolarIntersect>
one sec, I got excited and rebooted to see if it still worekd
<clever>
ottidmes: ive had a wide range of problems with my amd card, one of them nixos specific
<PolarIntersect>
hmm
<PolarIntersect>
worked again after rebooting
<PolarIntersect>
this is exciting :D
<clever>
ottidmes: nixos xorg packages are made by a perl script that parses an http directory listing (yes, lol)
<clever>
ottidmes: and its auto-generating definitions for deprecated packages, that have since been merged into another repo
<clever>
ottidmes: that results in 3 year old code being used, rather then the maintained version in another repo!
<clever>
so bugs that have already been fixed occur only on nixos!
thedavidmeister has joined #nixos
<ottidmes>
clever: yikes!
<clever>
ottidmes: one of the bugs i saw a lot, was the gpu using a buffer, before the buffer was initialized
<clever>
ottidmes: EVERY single time chrome rendered a tooltip, it was total garbage, uninitialized buffer
<samueldr>
set back videoDrivers = [ "nvidia" ]; rebuild boot and reboot?
<ottidmes>
clever: hah, Ive seen that one too! so weird! but so far on modesetting it works (except regular errors in the log, but apparantly non-fatal)
<PolarIntersect>
I don't want nouveau if there's an option
<clever>
from what ive heard, the intel card controls the display and handles all 2d actions
<clever>
and the nvidia is 3d only, and renders directly into the 2d buffer of the intel card
romildo has joined #nixos
romildo has left #nixos [#nixos]
<clever>
PolarIntersect: hardware.bumblebee = { enable = true; driver = "nvidia"; }; i think this is how you enable that stuff
<clever>
while keeping videoDrivers = [ "intel" ]; i think
<clever>
the reason for that design, is that the nvidia card can just be 100% disabled, and then you have a very low-power 2d-only card, and better battery life
__monty__ has quit [Quit: leaving]
jabranham has quit [Quit: ERC (IRC client for Emacs 27.0.50)]
<emily>
I don't think it's about 2D vs. 3D.
<emily>
but yeah, the intel card is always the one driving the display
<emily>
also the modesetting DDX driver is often preferable to intel
<emily>
which is kinda unmaintained i think
civodul has quit [Quit: ERC (IRC client for Emacs 26.1)]
thedavidmeister has joined #nixos
<simpson>
I think that that Intel chipset is i915-ish, so it can technically do 3D but not very fast.
<simpson>
On the flip side, `nvidia` is well-known to be pitiful at 2D work.
<clever>
that reminds me, one of the bugs i had on my amd card (since fixed), is that 2d rendering was horid slow
<simpson>
I never really liked these dual-chipset setups; there's next-to-no documentation on how the OS is supposed to drive them.
<clever>
as in, `ls -l /nix/store` took 5 minutes slow
<clever>
but 3d was fine
<clever>
if i turned hw accel off, 2d was fast, and it lacked 3d entirely
<romanofskiWork>
would anyone be able to help me with what command line or expression I can install this package: https://github.com/kubek2k/chunkwm.nix ?
<ottidmes>
romanofskiWork: follow the instructions on that link and do nix-env -iA chunkwm or nix-env -iA nixos.chunkwm, I guess? Not sure about darwin
thedavidmeister has quit [Quit: WeeChat 2.3]
* romanofskiWork
curses the darn macbook
vk3wtf has quit [Ping timeout: 252 seconds]
<PolarIntersect>
clever: hmm... So, now random syscalls are hanging
<azazel>
romanofskiWork: you can just install Nixos into a partition on a macbook too
<clever>
PolarIntersect: if you dont care about 3d accel, you can just go back to pure intel
<PolarIntersect>
like, if I SSH then the machine hangs when strace says it's getting the system time O_o
<PolarIntersect>
clever: Sadly, I do care about that :(
<romanofskiWork>
ottidmes: hm.. no... just tried it again, but the package can't be found... I've also thought about putting the config in a separate e.g. release.nix file and try to install that, but no luck
<clever>
PolarIntersect: have you checked the nixos-hardware repo on github?
jtojnar has quit [Quit: jtojnar]
makefu has quit [Ping timeout: 246 seconds]
<PolarIntersect>
Yeah, but my laptop isn't there
jtojnar has joined #nixos
<PolarIntersect>
It's an MSI
makefu has joined #nixos
thedavidmeister has joined #nixos
<clever>
PolarIntersect: try the other driver, hardware.bumblebee = { enable = true; driver = "nouveau"; };
<clever>
PolarIntersect: also, reboot things between each test
<Church->
PolarIntersect: Graphics not working?
<thedavidmeister>
hey, what does it mean if enabling bumblebee gives me a black screen when i login?
c0bw3b_ has joined #nixos
jtojnar has quit [Ping timeout: 272 seconds]
<neonfuz>
infinisil: oh sry, names were same length and color on my irc client lol
polman has quit [Ping timeout: 272 seconds]
polman has joined #nixos
Ariakenom has quit [Quit: Leaving]
<PolarIntersect>
Church-: ya -.-
<PolarIntersect>
clever: I have been rebooting for each test to make sure kernel modules don't conflict or anything
<PolarIntersect>
Is there a way to use beta drivers or something?
<Ralith>
how do I force a cmake package to turn off -Werror?
<PolarIntersect>
I think you can use `cmake -D CFLAGS=whatever`
<PolarIntersect>
maybe check out the advanced sections of `ccmake`
<clever>
Ralith: there is also cmakeFlags = [ "-D" "CFLAGS=whatever" ];
<Ralith>
clever: are you sure those are supposed to be separate arguments?
<clever>
not entirely sure
<PolarIntersect>
they aren't, sorry
<PolarIntersect>
My bad
<PolarIntersect>
cmake gets mad if you put a new ar between -D and it's argument
pie_ has quit [Remote host closed the connection]
pie_ has joined #nixos
pie_ has quit [Remote host closed the connection]
pie_ has joined #nixos
<PolarIntersect>
clever: I don't think bumblebee will work since it's an MXM card?
<{^_^}>
[nixpkgs] @Ericson2314 opened pull request #50375 → haskellPackages: Add shell completion for 18.09 → https://git.io/fpsJn
vidbina_ has quit [Ping timeout: 240 seconds]
shabius_ has joined #nixos
<PolarIntersect>
hmm
<ottidmes>
Could it have to do with the kernel version? I assume nvidia is closed source? I had to use kernel 4.4 LTS with for my AMD GPU driver (now I just use modesetting with only 2D)
<PolarIntersect>
nvidiaBeta also broken
<kingemp>
For the build phase, anyone know how to get nix to run parallel jobs? I see it's -j 1 by default, and I can't seem to override with makeFlags
<clever>
kingemp: enableParallelBuilding = true;
<kingemp>
clever, already have that set
<clever>
kingemp: nix show-config | grep core
<kingemp>
clever, eek! cores = 1
shabius has quit [Ping timeout: 252 seconds]
<clever>
kingemp: nixos or other?
<kingemp>
clever, nixos
<clever>
kingemp: nix.buildCores = 10; in configuration.nix
<kingemp>
clever, awesome thx! I wonder if this should be done by default in hardware-configuration..?
<kingemp>
happy to make an issue if you think so, but maybe there's a good reason
<clever>
kingemp: nixos-generate-config should fill that in, check what your generated config has