<ottidmes>
joehh: so rewrite it as a for loop for example
<joehh>
samueldr, ottidmes - thanks for that,
florianjacob has joined #nixos
Radivarig has quit [Quit: Leaving]
bennofs[m] has joined #nixos
<bennofs[m]>
small cross-compilation question: why is there no pkgs.foobar.nativeDrv? (only the callPackage-filled args seem to have a nativeDrv)
<bennofs[m]>
example: nix build --arg crossSystem 'with (import <nixpkgs/lib>); systems.examples.aarch64-multiplatform' nixpkgs.hello.nativeDrv does not work
<bennofs[m]>
(no attribute nativeDrv)
<{^_^}>
[nixpkgs] @worldofpeace opened pull request #51054 → pythonPackages.pandas: don't run tests on aarch64 → https://git.io/fpERE
<bennofs[m]>
yeah that gives the same result, but the problem is that it breaks if I use pkgs.hello as nativeBuildInput since stdenv.mkDerivation expects the nativeDrv attribute
pie__ has quit [Ping timeout: 264 seconds]
<v0|d>
bennofs[m]: um, i just do [ hello ]; not [ buildPackages.hello ];
Ariakenom has quit [Quit: Leaving]
<v0|d>
what might be the reason for you doing that?
graphene has quit [Remote host closed the connection]
graphene has joined #nixos
<bennofs[m]>
if you try to cross-compile nixos modules, some modules pass `pkgs.foobar` directly via override etc and this won't work
<v0|d>
see.
goibhniu has quit [Ping timeout: 252 seconds]
<ottidmes>
bennofs[m]: can pkgs in your use case be completely replaced with buildPackages? if so, maybe tweak the module system to behave like that in your nixpkgs checkout?
<bennofs[m]>
yeah in this case I can edit the module to use buildPackages. just seems wrong in general that it doesn't get passed along automatically
johann__ has quit [Ping timeout: 260 seconds]
<v0|d>
bennofs[m]: open up an issue?
<{^_^}>
[nixpkgs] @ryan4729 opened pull request #51056 → pythonPackages.yattag: init at 1.10.1 → https://git.io/fpER9
asymmetric has quit [Ping timeout: 240 seconds]
bachp has joined #nixos
fgaz has joined #nixos
Purple-mx has joined #nixos
clacke[m] has joined #nixos
dtz has joined #nixos
Ericson2314 has joined #nixos
simbergm has joined #nixos
koschei[m] has joined #nixos
ejpcmac has joined #nixos
leons has joined #nixos
MarkOtaris has joined #nixos
ptotter[m] has joined #nixos
cornu has joined #nixos
nyanloutre[m] has joined #nixos
kaychaks[m] has joined #nixos
icetan has joined #nixos
roberth has joined #nixos
thefloweringash has joined #nixos
schmittlauch[m] has joined #nixos
yegortimoshenko has joined #nixos
tokudan[m] has joined #nixos
timclassic has joined #nixos
sphalerit has joined #nixos
yangm97 has joined #nixos
timokau[m] has joined #nixos
vaibhavsagar has joined #nixos
stammon has joined #nixos
rycee has joined #nixos
rnhmjoj has joined #nixos
ysndr_ has joined #nixos
abathur has quit [Quit: abathur]
rpg has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Mic92>
but msys2 is probably not a native windows port anymore then.
phreedom has quit [Ping timeout: 256 seconds]
vmandela_ has quit [Remote host closed the connection]
vmandela_ has joined #nixos
crmlt has joined #nixos
<jophish>
Well, binutils builds but doesn't run on Windows
<jophish>
No error message or anything
<jophish>
I might give this one a miss for now
klntsky has joined #nixos
<angerman>
jophish: I'm also not very happy with the whole way we build gcc in nix; I think Ericson2314 would prefer a more modular approach as well. However, *please* make this work :-)
vmandela_ has quit [Ping timeout: 268 seconds]
phreedom has joined #nixos
ckauhaus has joined #nixos
vmandela has joined #nixos
pkx3 has joined #nixos
phreedom has quit [Ping timeout: 256 seconds]
vmandela has quit [Ping timeout: 276 seconds]
slyfox_ is now known as slyfox
worldofpeace has joined #nixos
<umbaws>
Mic92: Thanks it worked with: pkgs.lib.Mkforce [];
phreedom has joined #nixos
mkoenig has quit [Read error: Connection reset by peer]
mkoenig has joined #nixos
EarlDeLaWarr has joined #nixos
day|flip is now known as day|flip-afk
Thra11 has quit [Ping timeout: 268 seconds]
nikal has quit [Quit: Konversation terminated!]
phreedom has quit [Ping timeout: 256 seconds]
johanot has joined #nixos
worldofpeace has quit [Ping timeout: 240 seconds]
Mateon3 has joined #nixos
Mateon1 has quit [Ping timeout: 245 seconds]
Mateon3 is now known as Mateon1
<sphalerite>
Anyone here know things about MoreTea's yarn2nix? (presumably MoreTea, but I'd appreciate help from anyone ;)) I'm trying to package loomio, which uses lodash 3.x in its frontend but also has 4.x in its frontend's dependency closure. The output of the mkYarnPackage call has libexec/Loomio/node_modules, which contains everything but lodash 3.x
klntsky has quit [Ping timeout: 256 seconds]
mayhewluke has quit [Ping timeout: 250 seconds]
<sphalerite>
however lodash 3.x can be found in libexec/Loomio/deps/Loomio/node_modules (and it's the only thing there)
<{^_^}>
[nixos-homepage] @domenkozar pushed commit from @ekalosak to master « updated sources of security information; removed a dead key (#252) »: https://git.io/fpEMu
<{^_^}>
[nixos-homepage] @domenkozar merged pull request #252 → updated sources of security information; removed a dead key → https://git.io/fpslT
<{^_^}>
[nixpkgs] @dotlambda pushed commit from @marsam to master « pythonPackages.manuel: 1.8.0 -> 1.10.1 (#51029) »: https://git.io/fpE7M
sigmundv has joined #nixos
umbaws has joined #nixos
xvnvx is now known as alphabeta
lnikkila has joined #nixos
klntsky has quit [Ping timeout: 256 seconds]
klntsky has joined #nixos
asymmetric has joined #nixos
<{^_^}>
[nixpkgs] @Mic92 merged pull request #51005 → doc: add example usage for breakpointHook → https://git.io/fp05i
<{^_^}>
[nixpkgs] @Mic92 pushed to master « doc: add example usage for breakpointHook (#51005) »: https://git.io/fpEdm
acarrico has joined #nixos
<Mic92>
bennofs[m]: sometimes you can take binaries from buildPackages.<package>, but xxx-config is indeed not easy to solve.
<Mic92>
Those will most likely contain absolute paths. If you are lucky it is just a script.
<bennofs[m]>
I think xxx-config should just be compiled for build arch maybe?
<bennofs[m]>
Or is this where target becomes relevant?
<Mic92>
bennofs[m]: well, than you cannot use this binary on the host target. Also you would need to convince the build system to just build this package for a different architecture.
day|flip has quit [Quit: Leaving]
<Mic92>
Target comes into play if you have compilers that compile for certain target architecture.
<Mic92>
s/host target/host system/
<Mic92>
and I didn't mean package, but binary
<Taneb>
I'm looking at the services.openvpn NixOS options
<Taneb>
Is there a way to do user and password authentication without storing the password in plaintext on the disk?
<Mic92>
the only one-size fits all solution for -config like binaries would be some kind of emulator
<clever>
and qemu-user can handle linux ones
<Mic92>
Taneb: you can only store the password in an external file, if you want to use openvpn from a service.
<Mic92>
otherwise you would need a script
<Taneb>
...yeah, that feels obvious now you've pointed it out to me
<Taneb>
Thanks, Mic92
Lisanna has joined #nixos
__Sander__ has joined #nixos
mayhewluke has quit [Ping timeout: 246 seconds]
mayhewluke has joined #nixos
woffs has joined #nixos
steshaw has quit [Quit: Connection closed for inactivity]
<srhb>
__monty__: The boxes at the top explain channel bumps.
<__monty__>
srhb: Thank you.
<Mic92>
Can we an official domain for that?
<srhb>
__monty__: Currently there are some blocking tests that rely on staging-next to be merged, which is unfortunate. Hopefully it'll be done this week.
otti0815 has joined #nixos
fendor has joined #nixos
betaboon has joined #nixos
benny has quit [Ping timeout: 245 seconds]
iMatejC has quit [Ping timeout: 245 seconds]
iMatejC has joined #nixos
Rusty1 has joined #nixos
ma27 has quit [Ping timeout: 245 seconds]
ma27 has joined #nixos
graphene has quit [Remote host closed the connection]
graphene has joined #nixos
init_6 has quit []
<{^_^}>
[nixops] @chawki008 opened pull request #1056 → Update old backups → https://git.io/fpue9
acarrico has quit [Ping timeout: 252 seconds]
<fendor>
i can not see the store page in steam, is that just the state or am i doing something wrong?
<__monty__>
srhb: I think that doesn't affect me since I'm on nixpkgs-unstable. Waiting for a change that happened only 6 hours ago though.
<fendor>
clever, same, csgo, mb warband do not work, others like dont starve and civ 5 do.
<fendor>
etu, ok, then it is probably not helpful
<clever>
fendor: happen to know if an issue is already open?
<fendor>
clever, i opened one
<etu>
I packaged it whenever I wanted to use it, and it seemed to work well at that time. But then I've stopped using it and now it doesn't seem to build. Don't think anyone have changed anything in that package so it must be some dep breaking something.
<{^_^}>
#31189 (by dezgeg, 1 year ago, open): Solve the libGL ABI problem
<etu>
And nowadays I don't know how actively maintained LSI is either... So maybe it should just be dropped.
crmlt has joined #nixos
graphene has joined #nixos
ivan_ is now known as ivan
<fendor>
Actually, I can not even open the friends tab on nixos
<fendor>
*on steam in nios
<ekleog>
etu: fyi, delroth is going to try to make PHP-with-intl build work on osx, if it succeeds it'll supersede my PRs :)
<etu>
ekleog: Oh, nice.
iyzsong has quit [Ping timeout: 252 seconds]
<etu>
ekleog: It's very hard for me to try to do that since I don't have access to any macOS machines... So it's very slow to let ofborg do every single testbuild that it would require...
<ekleog>
same here :/ (except I don't have access to ofborg's osx builder for the time being)
klntsky has joined #nixos
arjen-jonathan has joined #nixos
fusion809 has joined #nixos
<ottidmes>
how can I get nix repl to show me things in depth, getting things like: [ { ... } { ... } { ... } { ... } { ... } { ... } { ... } { ... } ] is useless to me
<sphalerite_>
ottidmes: :p
<sphalerite_>
ottidmes: see :? for help as it tells you at the beginning ;)
WhittlesJr has quit [Ping timeout: 256 seconds]
<ottidmes>
sphalerite_: thanks, I tried man and --help ;) but I got it now, I had to use :p <expr>
<sphalerite_>
yep
jperras has joined #nixos
sphalerite_ is now known as sphalerite
<sphalerite>
ottidmes: when I wrote ":p" I did mean it as an answer to your question, not an emoticon :p
elgoosy has joined #nixos
markus1189 has joined #nixos
<sphalerite>
oh dear. That wasn't even intentional, putting the emoticon there at the end
<etu>
haha
<ottidmes>
sphalerite: which I of course at the time did not know :p ;)
<ottidmes>
sphalerite: do you know of a way to have some file always loaded other than explicitly doing :l on every start?
<sphalerite>
ottidmes: nix repl $file
<sphalerite>
nix repl --help isn't very obvious about that, I must admit :D
<Taneb>
So, it looks like I might be writing myself a Hydra plugin
<Taneb>
I want to have a jobset triggered on a schedule
<ottidmes>
sphalerite: makes sense, making an alias right now :) so that I finally can copy paste snippets from my config as-is
<Taneb>
Which I don't think is currently possible
<srhb>
Taneb: I'd sooner just curl the API to do that.
<sphalerite>
^
<srhb>
hydra-notify and all the plugin stuff is a world of pain, and adding another plugin to the list will slow down things a lot. :)
<Taneb>
I don't know what you mean by that
<srhb>
Taneb: Schedule a job that accesses hydra via http to trigger eval.
<Taneb>
That feels ugly to me
<Taneb>
Somehow more ugly than having to write perl
<gchristensen>
scheduled? that is what Check Interval is for, right?
Rovanion has joined #nixos
<srhb>
gchristensen: That's more like frequency, but I'm sure it's good enough for many cases.
<srhb>
Taneb: The individual plugins in perl are actually mostly fine.
<Taneb>
gchristensen: that doesn't give you any guarantee of the time of day, for example
<srhb>
Taneb: The problem is with the infrastructure responsible for executing the plugins.
<srhb>
Taneb: Speaking of which, I don't think plugins can even do what you want.
<srhb>
Taneb: They generally trigger when builds finish.
<gchristensen>
Taneb: what are you using it for which depends upon such strict time?
<Taneb>
gchristensen: nightly builds which take a lot of (distributed to work machines) compute, which we don't want to run when people are in the office
<gchristensen>
oh _cool_
<sphalerite>
Taneb: there's nothing ugly about using an API.
<Taneb>
sphalerite: there's something ugly about using the web API in an automated manner from the machine serving the API
graphene has quit [Remote host closed the connection]
<sphalerite>
I disagree. It allows the fact that it's the same machine to be a mere coincidence :D
graphene has joined #nixos
random_yanek has quit [Ping timeout: 260 seconds]
<Taneb>
sphalerite: we have different sense of aesthetic
<Taneb>
Actually, now that I think about it, I can see some benefits
crmlt has quit [Ping timeout: 252 seconds]
<Taneb>
(I still think it feels ugly)
aKriJcz has joined #nixos
<Taneb>
So if I'm going to do it using the API, where I can I find anything approximating documention (as rough as "source code") for the API?
<clever>
,locate libGL.so
<{^_^}>
Found in packages: libglvnd, primusLib, libGLU_combined
<infandum>
Two questions: 1. How can I install infocmp? I installed ncurses but the command is still not known. 2. Emacs (with spacemacs) takes up an absurd amount of memory and hangs on exit so I need to kill it. The same configuration is fine on Arch. What could be the issue?
<{^_^}>
[nixpkgs] @Mic92 pushed commit from @dingxiangfei2009 to master « google-compute-image: make it a module and the size tuneable (#49854) »: https://git.io/fputv
<{^_^}>
[nixpkgs] @dotlambda pushed commit from @peterhoeg to master « pythonPackages.pyupdate: init at 0.2.16 (#50817) »: https://git.io/fputW
reinhardt has quit [Quit: Leaving]
<clever>
infandum: main difference is just the abiVersion
<clever>
it may also work with `nix-shell -p ncurses`
<infandum>
I have ncurses installed though
booglewoogle has joined #nixos
<clever>
installing it wont give you the .dev output
<clever>
only nix-shell gives .dev
booglewoogle has quit [Remote host closed the connection]
<infandum>
Can I have the dev in my configuration.nix
<clever>
ncurses.dev
<fendor>
clever, ok, i found my problem... it was my bad. I had a horrible fix, when maven downloads an in memory mongodb instance, the executable fails because it cant find ld in /lib64. So i symlinked /lib64/ld... to some ld i found somewhere. This made the executable work, but apparently broke steam.
<infandum>
Fantastic, thank you! Is there documentation for the .dev notation?
kreisys has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<clever>
fendor: aha, if i run the game under steam-run, i get a dialog very similiar to the nix expression in the gist, but now with text!
<clever>
its stating that it needs steam to run
* clever
shakes fist at DRM
<fendor>
clever, gog should make an linux client, that should remove some of the problems at least. I am downlaoding borderlands 2 for a second and try to see if i can reproduce your problem
<clever>
one anoying thing, steam wont let me run the windows version under proton
<clever>
so i have no way to test that route
graphene has quit [Remote host closed the connection]
<Rovanion>
Does anyone have anything interesting on provisioning nixos on bare metal?
<clever>
fendor: and now to cheat, i now have a new game, /run/current-system/sw/bin/xterm !
klntsky has quit [Ping timeout: 256 seconds]
<Rovanion>
The nixops manual doesn't really have anything on it, or if it does I'd gladly be pointed at it.
graphene has joined #nixos
<gchristensen>
Rovanion: can you say more about your use case? I've done this quite a bit in different ways.
<gchristensen>
nixops typically assumes nixos is already installed when it gains control, except in a few specific cases where provider-specific hacks are used to "take over" the system with nixos
<fendor>
clever, thats smart, so you know that drm is messing it up
<clever>
fendor: drm is only making it harder to debug, but is not the root problem
crmlt has joined #nixos
<clever>
fendor: if i manually launch the game from an "xterm game" or steam-run, it fails to detect steam and aborts
<clever>
fendor: if i properly launch it from steam, it goes fullscreen black window, then exits after a few seconds
<catern>
hey #nixos
<catern>
how do I configure fonts I've installed with Nix on a non-NixOS distro?
<Rovanion>
gchristensen: I'm going to build a cluster starting with perhaps 10 machines but expanding to 100. I'm unsure if any PXE or network boot infrastructure exists at the moment. It will be machines of three or four different types.
rpg has joined #nixos
ThatDocsLady has quit [Quit: Leaving]
<gchristensen>
Rovanion: nixos can be netbooted, sure, and there have been a few uses of the netboot infra to do unattended installations
<aminechikhaoui>
Mic92 which nixpkgs version is used in travis-ci when running the checks for nixops ? is it unstable ?
<clever>
justdoit is also included in the netboot images
<Rovanion>
Eh, what is it justdoit does? I'm reading the wiki page on kexec and intuiting that it consumes a running linux OS and replaces it with nix?
<Mic92>
aminechikhaoui: yes
<clever>
Rovanion: justdoit is a bash script that will partition, format, and install nixos
<clever>
Rovanion: kexec is one of many ways of booting the nixos installer, but you can also use netboot
<fendor>
clever, same happens on my machine with the same error, I think
booglewoogle has joined #nixos
<Rovanion>
And thanks by the way. Nix seems to me like the "right" approach to providing a homogenous cluster of machines as I look at it today. Right perhaps as opposed to easy.
erasmas has joined #nixos
umbaws has quit [Quit: WeeChat 2.3]
<clever>
Rovanion: with the netboot_server above, you can get any machine on the LAN to boot into a nixos installer when it network boots
<clever>
then youll likely want a custom util to run on bootup, that talks to a central box ( gchristensen, i'm guessing this is similar to packet.net stuff)
<clever>
and that util will confirm its save to wipe the current box, then run justdoit, and report progress, and reboot
<clever>
depending on your hardware control (ipmi and stuff) you can then either remotely mess with the boot order to stop network booting
<clever>
or you can tweak the netboot_server to force hdd booting, even when it netboots
odi__ has joined #nixos
<clever>
and that will manage remotely reinstalling nixos on machines
<clever>
then you just need to fill in an ip on a nixops deployment with targetEnv = "none"; and you can deploy to it as normal
<Rovanion>
Loose thought: I've used fog before to provision systems over netboot. And that worked quite well. You could schedule for new images to be put on machines, if there were no image scheduled for that machine it would boot to HDD.
<{^_^}>
#51027 (by basvandijk, 1 day ago, open): opencv4: init at 4.0.0
<gchristensen>
domenkozar: not sure anyone
<domenkozar>
I'm getting tired
infandum has quit [Read error: Connection reset by peer]
infandum has joined #nixos
<Rovanion>
clever: Yeah, essentially the logic is that of a hashmap of MAC-addr to OS image url. If there was an os image available for the machine with MAC X then download and flash it.
crmlt has quit [Ping timeout: 252 seconds]
<domenkozar>
I'm going to give him last warning
<clever>
Rovanion: but it could also just netboot into an installer that runs nixos-install, rather then copying an image
<gchristensen>
Rovanion: lots of ways to do it :)
<Rovanion>
clever: Hmm, interesting. For some reason I thought there was a limit to the size of a netboot image so that it could only be very tiny
<Rovanion>
No idea where I got that idea from.
<gchristensen>
netboot images can be large, pxe images need to bet tiny because TFTP is so slow, but iPXE allows HTTP and then they can be big :)
<gchristensen>
I've done the hashmap-to-os-image before for stateless services, like memcache
<gchristensen>
just running in-memory
<Rovanion>
Ah, okey. That explains it then. TFTP brings up memories, no way to secure that thing right?
rpg has quit [Ping timeout: 268 seconds]
<clever>
Rovanion: i did design not-os to be secure, but it relies on the initial ipxe binary being on the machine somehow
<clever>
Rovanion: either in a /boot partition you dont modify, the network card boot rom, or baked into the bios
<gchristensen>
Rovanion: tftp is just a bit of a disaster :)
<clever>
Rovanion: this build of ipxe has been modified to have a x509 cert baked into it, along with a mini script (the EMBED part), so it wont do the normal netboot stuff automatically
<gchristensen>
Rovanion: usually people will netboot by using pxe to load ipxe, then ipxe loads a bigger image
<clever>
Rovanion: line 67 locks signing requirements on, so all files it loads from then on must be signed
<andrewrk>
can anyone recommend a service to purchase to run nixos-mailserver on? e.g. aws ec2, digital ocean, etc. how much cpu / ram / disk would I need?
<andrewrk>
I started setting it up on my home server but local ISP isn't going to let me have the ports I need and won't do the reverse DNS thing
<andrewrk>
I'm trying to escape google without going broke
<tilpner>
andrewrk - Vultr, Hetzner Cloud, Digital Ocean
<tilpner>
I've used NixOS on all of those, and they're listed in order of difficulty
<gchristensen>
andrewrk: the problem you're going to have with most of these is IP reputation
<tilpner>
(ascending)
<andrewrk>
thanks. any suggestions on needed specs to host my own email server? (just for personal use)
<tilpner>
Smallest option should be enough, but you probably want an ipv4 address
<Rovanion>
clever: Interesting, and you put it in /boot because that's where you can write?
<andrewrk>
noted
<tilpner>
(Vultrs smallest is ipv6-only)
<andrewrk>
thanks tilpner & gchristensen
<tilpner>
gchristensen - Isn't it enough to check the ip you get on creation against online blacklists?
vidbina has quit [Ping timeout: 246 seconds]
<clever>
Rovanion: my initial plan was to bake it into the motherboard with coreboot, but thats not always an option
<clever>
Rovanion: /boot is another trusted way to do it
<the-kenny>
Is anyone else getting this annoying issue where Mouseover-Windows (like these when you hover over something for a second and it shows an explanation) cause rendering artifacts? From time to time one of these windows gets "stuck" on my display and stays as a non-updating rectangle until I find the same parent window and create another mouseover hover window there.
<the-kenny>
I *think* it's limited to stuff rendering via Webkit - I mostly notice it in Chrome itself
nika has joined #nixos
vasiliy_san has quit [Quit: Connection closed for inactivity]
<the-kenny>
This is on a x260 using intel grapics. They just started appearing a few weeks ago - I can't pinpoint it to a specific NixOS release :/
infandum has quit [Remote host closed the connection]
Ariakenom has quit [Read error: Connection reset by peer]
nikka has joined #nixos
nika has quit [Ping timeout: 240 seconds]
<ottidmes>
the-kenny: I too have experienced that, but haven't in quite some time, which probably has to do with me now using bspwm without a compositor (rather than gnome 3)
<ottidmes>
the-kenny: but I have had that for years, not recently, and only the getting stuck part
jmeredith has joined #nixos
<andrewrk>
I just booted up a nixos instance on vultr. how do I know whether it uses uefi or legacy boot?
basvandijk has quit [Remote host closed the connection]
<andrewrk>
I guess i can just try it both ways and see which one works
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/fpulY
johanot has quit [Quit: WeeChat 2.2]
Tucky has quit [Quit: WeeChat 2.2]
shabius has quit [Remote host closed the connection]
shabius has joined #nixos
orivej has quit [Ping timeout: 272 seconds]
arjen-jonathan has quit [Ping timeout: 244 seconds]
<Lisanna>
I have a fairly interesting predicament with enableParallelBuilding and runInLinuxVM - if I'm building a derivation in a linux VM, if I'm on a massively multicore system, the build can fail if I don't give it a ton of memory (with the memSize attribute). The amount of memory required varies depending on the number of cores.
<Lisanna>
is there a reasonably safe way to use enableParralelBuilding in the linux vm?
<Lisanna>
seems like there should be a memPerCore setting for runInLinuxVM
nikka has quit [Ping timeout: 250 seconds]
shabius has quit [Remote host closed the connection]
shabius has joined #nixos
graphene has quit [Remote host closed the connection]
__Sander__ has quit [Quit: Konversation terminated!]
<clever>
Lisanna: you could use preStart to mutate some params...
graphene has joined #nixos
<clever>
Lisanna: memSize lands in $QEMU_OPTS, and you could then mess with that in the preStart script which has access to $NIX_BUILD_CORES
<clever>
preVM = ''export QEMU_OPTS="$QEMU_OPTS -m $(($NIX_BUILD_CORES * 128))"''; for e
<clever>
x
civodul has quit [Quit: ERC (IRC client for Emacs 26.1)]
<Lisanna>
clever yeah, will have to check if qemu handles multiple -m args ok
<clever>
Lisanna: or sed to change the existing flag
nikka has joined #nixos
<{^_^}>
[nixpkgs] @NickHu opened pull request #51064 → quantomatic: init at 0.7 → https://git.io/fpuBA
<aminechikhaoui>
Hey, I have a c++ codebase where we used to build with nixpkgs 17.09, while I was trying to update to 18.03 or later I noticed that the build now fails with undefined reference to a method from libgcc
<aminechikhaoui>
comparing the output of NIX_DEBUG=1 make between 17.09 and 18.03, I see that the difference is one is automatically adding -lgcc and lgcc_s while invoking ld while the other (failing one) only adds -lgcc_s
<samueldr>
>> Note that only this single file is copied, even if it imports others
<aminechikhaoui>
any idea what can be the reason of that ^ ?
<aminechikhaoui>
adding -lgcc explicitly to the flags works but that doesn't sound like the right fix
nika has joined #nixos
cyounkins has joined #nixos
nikka has quit [Ping timeout: 268 seconds]
sir_guy_carleton has quit [Quit: WeeChat 2.2]
Lisanna has quit [Quit: Lisanna]
graphene has quit [Remote host closed the connection]
graphene has quit [Remote host closed the connection]
graphene has joined #nixos
mmlb has joined #nixos
<delroth>
fyi if you didn't see the nodejs lol of the day: https://github.com/dominictarr/event-stream/issues/116 looks like nix is only depending on 0.1.2 of the flatmap-stream package, 0.1.1 is the backdoored version
<{^_^}>
dominictarr/event-stream#116 (by FallingSnow, 5 days ago, open): I don't know what to say.
<delroth>
it would be nice to have someone double-check, but it doesn't seem like nixpks currently references the backdoored package. now the real question is did it ever reference it...
rpg has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<delroth>
looks like a9efcc85cbfc0a8ebf5ab8dbfa914e9447018923..03bcca7a457e60bbfd80272347a2a85ff821b832 had the backdoored flatmap-stream package in nixpkgs.
xelxebar has joined #nixos
worldofpeace has joined #nixos
<delroth>
the other way around: 03bcca7a457e60bbfd80272347a2a85ff821b832..a9efcc85cbfc0a8ebf5ab8dbfa914e9447018923
dbmikus_ has quit [Ping timeout: 245 seconds]
<sphalerite>
gchristensen: it could at least complain rather than busy-looping if that invariant is violated :D
<sphalerite>
or crash
<gchristensen>
yikes
Thra11 has joined #nixos
philipcristiano has joined #nixos
<avn_>
sphalerite: another fun fact btw, if cache unavailable (dns not resolving for example), it won't build package locally (assuming that all prerequisites are in local storage)
avn_ is now known as avn
worldofpeace has quit [Remote host closed the connection]
<clever>
yl[m]: this is an old module for iscsi root as well
<yl[m]>
I saw that as well (found a Gist of yours about 3 years ago :))
random_yanek has joined #nixos
random_yanek has quit [Max SendQ exceeded]
<asymmetric>
srhb: that works! <3
<asymmetric>
so basically `import nixpkgs` always needs a `config` attribute?
<gchristensen>
you don't need to pass config = {}
<gchristensen>
you do need to pass an attrset, even an empty one
<yl[m]>
clever: does it make sense to port this work to NixOS? I'm thinking of a service similar to yours but also a serivce allowing discovery and opening sessions
<srhb>
asymmetric: You weren't calling it before at all (the {} at the end)
<clever>
yl[m]: yeah, that is something that should be in nixpkgs itself
<srhb>
asymmetric: config = {} just prevents some of your config leaking into your expressions.
<clever>
yl[m]: the "new" dhcp in the initrd stuff makes it a lot simpler as well
<asymmetric>
but when i did the double import i was calling it, right?
<asymmetric>
fetchFromGitHub was returning an attrset, and i was importing it - no?
<srhb>
asymmetric: Yes :)
<srhb>
asymmetric: Which gives you a function bound to the name nixpkgs
<srhb>
asymmetric: Which you never call.
rpg has joined #nixos
mmlb has quit [Read error: Connection reset by peer]
<srhb>
asymmetric: fetchFromGitHub gives you a path, really.
<srhb>
asymmetric: (But that's not the issue)
mmlb has joined #nixos
<yl[m]>
clever: I'm going to focus on simply getting devices available. I don't have an environment to test out the initrd so I'd leave that part to you
<srhb>
asymmetric: If line 11 read `(nixpkgs {}).haskell.packages...` it'd work, to illustrate.
<cyounkins>
"pkgs/development/haskell-modules/hackage-packages.nix:/* hackage-packages.nix is an auto-generated file -- DO NOT EDIT! */" -- so uh what auto-generates it?
<srhb>
cyounkins: hackage2nix, part of cabal2nix
random_yanek has joined #nixos
<asymmetric>
right, makes sense
<clever>
yl[m]: you could do something as simple as adding a /data fs, that is iscsi mounted, on a normal nixos install
ng0 has quit [Quit: Alexa, when is the end of world?]
<cyounkins>
srhb: Ah! Thank you.
<clever>
yl[m]: and the module i linked will try to mount /data during the initrd
<yl[m]>
clever: that's true
<clever>
yl[m]: i had originally written that module to netboot my rpi's
<{^_^}>
[nixpkgs] @erikarvstedt opened pull request #51073 → Minor doc fixes → https://git.io/fpuo0
equivrel has joined #nixos
<asymmetric>
the outer ones are not necessary
arianvp has quit [Quit: WeeChat 2.2]
arianvp has joined #nixos
<cyounkins>
Was trying to fix https://github.com/NixOS/nixpkgs/issues/50325 but it seems there is no provision in hackage2nix to add arguments to `cabal configure` for a specific package?
<srhb>
cyounkins: We do that in overrides or by changing cabal2nix itself.
<srhb>
cyounkins: see configuration-common.nix for example
<srhb>
cyounkins: or grep for appendConfigureFlag in the same directory
<cyounkins>
srhb: Thank you! I'll see what I can do.
odi_ has quit [Remote host closed the connection]
odi has quit [Remote host closed the connection]
rpg has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<srhb>
cyounkins: (For some context: cabal2nix itself contains a bunch of logic for determining what flags to use based on the cabal file. So there's several avenues: Quickfixes are in the overrides in that directory. configuration-foo.nix. Missing information in the cabal file should go to the upstream package author. If the cabal file is correct AND we generate a wrong nix expression, fixes go to cabal2nix)
<srhb>
More or less.
<cyounkins>
I understand somewhat. I'm a newbie in haskell. Maybe there's a place in documentation to describe that?
perique has joined #nixos
<srhb>
cyounkins: The user's guide to the haskell infrastructure in the nixpkgs manual *should* contain this but does not. Contributions very welcome. :-)
fusion809 has quit [Remote host closed the connection]
<yl[m]>
I used `after` and `requires`
<clever>
yl[m]: and is the case right in the ssh.service ?
<yl[m]>
sshd service yea
<yl[m]>
it seems to me that the `services.openssh.listenAddresses` is useless without making sure the address exists
<yl[m]>
I don't see a one fix to fit all for it though nixpkgs-side
<clever>
yl[m]: oh, you can also mess with the restart rules, to make systemd keep trying
<yl[m]>
that's true. I can make it a Restart=Always
<yl[m]>
let me try that
xelxebar has quit [Ping timeout: 244 seconds]
rpg has joined #nixos
<yl[m]>
clever: it's actually already is `Restart=always`
arjen-jonathan has joined #nixos
<clever>
yl[m]: does the journal show signs of it restarting? and your not waiting RestartSec?
<clever>
Configures the time to sleep before restarting a service (as configured with Restart=). Takes a unit-less value in seconds, or a time span value such as "5min 20s". Defaults to 100ms.
<{^_^}>
dominictarr/event-stream#116 (by FallingSnow, 5 days ago, open): I don't know what to say.
<{^_^}>
[nixpkgs] @benley merged pull request #49668 → stern: Include bash & zsh completion support → https://git.io/fxpbY
<{^_^}>
[nixpkgs] @benley pushed to master « stern: Include bash & zsh completion support (#49668) »: https://git.io/fpuXB
<ottidmes>
hyper_ch: it has been talked about in #nixos-chat
<hyper_ch>
ottidmes: ah.... :)
civodul has joined #nixos
nika has joined #nixos
<yl[m]>
clever: the RestartSec worked well. Thx!
<clever>
yep
v0|d has quit [Ping timeout: 252 seconds]
nikka has quit [Ping timeout: 268 seconds]
bbarker has joined #nixos
<bbarker>
Trying out nixos containers, and I'm not sure what I'm doing wrong regarding logging into the container. The manual suggests I should be able to login using a system user. After I get passed the first auth phase, using "nixos-container login foo" I then get a login prompt for the foo container.
<bbarker>
I can't login using the same user, nor can I login as a different user that is declaratively specified in the container's configuration using hashedPassword (of course, it is possible I screwed that up somehow).
fusion809 has joined #nixos
<clever>
bbarker: there is a nixos-container root-login command
<bbarker>
clever, that works, looks like hashedPassowrd didn't have an effect - the password in /etc/shadow for the container user is still empty
<bbarker>
(note this user exists only in the container)
<clever>
bbarker: hashedPassword i think is only when mutable users is false
<bbarker>
ahh
<clever>
if mutable users is true, then you want initialHashedPassword
<clever>
and even then, it only takes effect when first creating the user
<bbarker>
so "nixos-rebuild switch" on the host won't update it?
<clever>
correct
<clever>
also, nixos-rebuild siwtch has a nasty habbit of not updating nixos containers
<bbarker>
oh :(
<clever>
you have to restart contains@name.service
<bbarker>
I guess as a workaround I can just "nixos-container destroy foo" before each "nixos-rebuild switch"
<bbarker>
planning to switch to mutableUsers=false anyway
<gchristensen>
make sure you either have a root password set, or sudo enabled for your user :)
<bbarker>
gchristensen - I have sudo enabled for my host user (otherwise I wouldn't be able to do most of this I gather), or did you mean for some other reason?
<gchristensen>
it is possible to lock yourself out by not having those setin the nixos config and then enabling mtable users
<bbarker>
I didn't want there to be any root access within the container, unless someone on the host users root-login, but I guess then they already have root/sudo on the host
<bbarker>
oh yeah, I already did this! luckily 1) this is a VM and 2) I backed up the configuraiton.nix
umbaws has quit [Quit: WeeChat 2.3]
<{^_^}>
[nixpkgs] @benley opened pull request #51075 → stern: build shell completion scripts even when cross-compiling → https://git.io/fpu17
mmlb has quit [Read error: Connection reset by peer]
mmlb has joined #nixos
umbaws has joined #nixos
umbaws has quit [Client Quit]
orivej has quit [Ping timeout: 250 seconds]
v0|d has joined #nixos
umbaws has joined #nixos
drakonis1 has joined #nixos
<{^_^}>
[nixpkgs] @arianvp opened pull request #51077 → nixos/nspawn: Fix small typo → https://git.io/fpuMc
<dmj`>
I have a custom nix expression for building a python package, it fails to build because it's trying to download something from pypi and cannot
klntsky has joined #nixos
<ottidmes>
I am curious, I just read somewhere someone stating that NixOS does not have that many packages available, that might have been the case, but is that really still the case right now? The only packages that I had to package lately were so obscure they weren't available in AUR (Arch Linux) either
<dmj`>
specifically the "expects" package, which upon further inspection shows that its fetched for all python packages, iiuc
<hodapp>
ottidmes: it's rare for me to need to make my own derivation
<lunik1>
ottidmes: in my experience nixos has better coverage of the packages I use than most distros
<hodapp>
I run into missing Python packages here and there, and obscure stuff mostly... depends though
<lunik1>
I run mergerfs and snapraid, nixos was one of the few distros packaging both which is part of the reason I was drawn to it
<hodapp>
huh, I know what neither of those are >_>
<ottidmes>
hodapp: yeah, some programming languages support could be improved, but other distro's package managers do not even attempt this or have clear limitations
<srl295>
hi, how do I best debug `networking.localCommands` ? should `nixos-rebuild switch` call them and can i then see the results?
<ottidmes>
srl295: systemctl status network-local-commands.service should give you the location of the script containing the commands
<clever>
and nixos-rebuild should run it at every switch
fendor has quit [Remote host closed the connection]
<hodapp>
ottidmes: the programming language support is a lot of why I switched in the first place
<hodapp>
I was doing contract work for which I had a lot of really ugly development environments with libraries that just refused to play nice together
<clever>
nix is perfect for that
fendor has joined #nixos
<srl295>
clever: ottidmes : i was trying those but didn't have quite the right thing. Hmm! Looks like those may not be called by switch… i did `systemctl restart network-local-commands.service` and got an err, now i can debug that. maybe the `test` failed
umbaws has joined #nixos
<clever>
srl295: oh
<clever>
srl295: its set to oneshot and remain after exit
<clever>
so systemd may not think it needs to restart
<srl295>
hm
<ottidmes>
hodapp: the main thing for me going to NixOS is that the effort you put into configuring your system is made concrete (it results in an actual artifact)
<srl295>
clever: nixos bug?
<{^_^}>
[nixpkgs] @costrouc opened pull request #51080 → pythonPackages.uarray: init at 0.4 → https://git.io/fpuDH
<hodapp>
ottidmes: that was a big part of it for me, not so much from the entire-system standpoint but for the fact that the fugly development environment I had just set up was described deterministically and declaratively in some shell.nix/default.nix and I could reproduce it on (for instance) my laptop quickly, and upgrades were greatly simplified
umbaws has quit [Client Quit]
umbaws has joined #nixos
<srl295>
clever: anyway, i'm trying to set an IP address and default route onto a bridge device, i should perhaps be using `networking.interfaces.<name?>.ipv4.addresses` instead anyway
<srl295>
^ Hmm `ifconfig: command not found` in networking.localCommands
Akii has joined #nixos
umbaws has quit [Client Quit]
crmlt has joined #nixos
umbaws has joined #nixos
<srl295>
clever: needed the full path `/run/current-system-sw/bin/ifconfig` in `networking.localCommands` - is that expected?
drk_lrd_877_ has quit [Remote host closed the connection]
<ottidmes>
hodapp: in a way you are growing your own specialized distro with NixOS, allowing you to incorporating all the tricks and tweaks your gathered over time, and allowing you to easily share it across your machines
<clever>
srl295: you need to add ifconfig to systemd.services.network-local-commands.path
<srl295>
ottidmes: hodapp i like the repeatability of nixos.
<srl295>
^ aha even better thx clever
umbaws has quit [Client Quit]
umbaws has joined #nixos
<ottidmes>
srl295: A while I was installing a server, managed to screw up the encryption after being done setting it all up correctly, those are the moments that the reproducability of NixOS really shines...
<Thra11>
If I have a local default.nix which imports nixpkgs and defines a new package, 'foo', can I run something to cross-compile that package, similar to how I could run 'nix build pkgsCross.armv7l-hf-multiplatform.foo' if the package was in nixpkgs?
danielrf has joined #nixos
xantoz has joined #nixos
perique has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<bbarker>
One thing I don't get about "nixos-container login" -why is it necessary to authenticate if you are already logged in as a user that is in the wheel group?
<bbarker>
As far as I can see, this forces users that authenticate with ssh keys and don't have a password set to use "sudo nixos-container root-login foo" instead
<etu>
clever: Do you have any experience with grub and encrypted lvm with /boot inside? My friend xantoz is new to NixOS and wants this and might need some pointers. I don't have any experience with that scheme of partitioning.
<arianvp>
is it possible in NixOS to instantiate a template systemd unit?
<arianvp>
basically I want to do
<xantoz>
"error: disk 'lvmid/<volume-groupens-UUID>/<lv:ns UUID>' not found"
<arianvp>
systemd.services."foo@bar" = {}; to use the ExecStart from the Template unit
<xantoz>
"error: disk 'lvmid/<UUID of the VG>/<UUID of the LV>' not found"
klntsky has quit [Ping timeout: 256 seconds]
<clever>
etu: i just have a plaintext /boot and an encrypted lvm, and then everything else lives in lvm
umbaws has quit [Quit: WeeChat 2.3]
crmlt has quit [Ping timeout: 250 seconds]
<__monty__>
Is there a way to find out whether a certain commit is in nixpkgs-unstable or not? Maybe even when it will likely be?
<arianvp>
__monty__: yes
<arianvp>
github.com/nixos/nixpkgs-channels
<arianvp>
this is the git repository containing all commits for each channel
<arianvp>
there's a branch per channel
<{^_^}>
[nixpkgs] @bennofs opened pull request #51081 → nbench: init at 2.2.3 → https://git.io/fpu9W
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<ottidmes>
xantoz: if you want more help, maybe show what you have configured so far? I have the same setup as clever (unencrypted /boot, encrypted LVM for the rest), just from your error, the disk you specify seems strange to me
graphene has quit [Remote host closed the connection]
<ottidmes>
xantoz: you have mounted this properly on the installer as well? Did you create /boot on the root subvol? and is that @ syntax correct, never used it myself with btrfs
<xantoz>
yes
<xantoz>
ottidmes: the @ isn't syntax, it's literal
<xantoz>
part of the subvol name
<xantoz>
it's a common scheme for btrfs subvol:s on some other distros
graphene has joined #nixos
endformationage has quit [Quit: WeeChat 2.3]
Izorkin has joined #nixos
<{^_^}>
[nixpkgs] @timokau opened pull request #51084 → pythonPackages.rpi2: various fixes → https://git.io/fpuQF
<{^_^}>
[nixpkgs] @erikarvstedt opened pull request #51085 → containers: simplify env var definition → https://git.io/fpu7v
<ottidmes>
xantoz: I am sorry, that was all I got after checking your configs, I have no experience with encrypted grub myself, which is where seems to be the problem
<{^_^}>
[nixpkgs] @alyssais closed pull request #51066 → valgrind: disable parallel building on Darwin → https://git.io/fpuEr
<xantoz>
I've managed to get it working on Gentoo (BIOS boot, even), as well as with OpenSUSE Tumbleweed (the installer set up everything for me)
<xantoz>
I do remember having to fight a bit with grub-install to get it working on Gentoo, though
<ottidmes>
xantoz: one thing I noticed is that lsblk returns "nazrin--vg-swap" why the extra "-" there I wonder
hamishmack has joined #nixos
<ottidmes>
xantoz: more importantly: /mnt/mnt/btrfs
<ottidmes>
xantoz: or did you make a mnt folder inside the mounted @ subvol at /mnt? I see /home being reported for me (which is just one of the mount points)
<{^_^}>
[nixpkgs] @timokau pushed commit from @vdemeester to master « gauge: 1.0.2 -> 1.0.3 (#51068) »: https://git.io/fpubp
alex`` has joined #nixos
vk3wtf has quit [Ping timeout: 252 seconds]
qyliss_ has joined #nixos
<ottidmes>
xantoz: its probably the encrypted grub that somehow does not work well with UEFI, at least the way it is configured in NixOS, I cannot help you futher along, the only question I have for you is, why does boot need to be encrypted? its not like I cannot put a USB stick in your laptop and boot from there
qyliss has quit [Ping timeout: 260 seconds]
qyliss_ is now known as qyliss
rprije has joined #nixos
<xantoz>
ottidmes: but you won't be able to fiddle with my kernel if you do
<xantoz>
the hurdle for messing with the bootloader is slightly higher
<xantoz>
also, theoretically, if combined with secure boot; it should provide protection against the evil maid
<xantoz>
although I don't have any systems set to that level of paranoia, yet
<xantoz>
ottidmes: I could try setting it up for BIOS boot, I guess. it doesn't matter that much
<xantoz>
if that makes it work
<ottidmes>
xantoz: worth a shot I guess
nDuff has joined #nixos
<nDuff>
What needs to be done to have out-of-tree kernel modules available in NixOS tests? I'm trying to use sysdig to debug a program that isn't behaving properly in such a test; however, even with boot.extraModulePackages = with config.boot.kernelPackages; [ pkgs.sysdig ]; environment.systemPackages = [ pkgs.sysdig ]; boot.initrd.kernelModules = [ "sysdig_probe" ]; -- the sysdig_probe module (and thus /dev/sysdig0) isn't available.
civodul has quit [Quit: ERC (IRC client for Emacs 26.1)]
<sphalerite>
nDuff: you might be able to insmod it
<makefu>
nDuff: not sure about it but you probably need to put the module into `boot.kernelModules` (not into initrd)
<nDuff>
makefu, hmm, that smells plausible -- I was looking at the boot-stage1.nix test, but that's a special case.
<nDuff>
...alas, still not available. And sure, if I knew where the right version of the module compiled for this system was located in the store, I probably *could* just insmod it, but that's a thing that needs to be determined.
__monty__ has quit [Quit: leaving]
<makefu>
i had a similar issue with dm_thin_pool, putting it into initrd.kernelModules didn't work for me there
<pbogdan>
nDuff: boot.extraModulePackages = with config.boot.kernelPackages; [ pkgs.sysdig ]; <- I'm guessing but you probably meant "sysdig" here and not pkgs.sysdig
<{^_^}>
[nixpkgs] @grahamc opened pull request #51090 → Revert "zfs cannot be distributed. Disabling it in the isos." → https://git.io/fpuxP
<nDuff>
Ooh.
<nDuff>
*facepalm*.
<nDuff>
Indeed.
<makefu>
di di diiii
Thra11 has quit [Ping timeout: 244 seconds]
<Ashy>
woah, i've got a nixos xorg background now in i3
nika has quit [Quit: Konversation terminated!]
<Ashy>
think i ran an upgrade yesterday
<gchristensen>
oh cool, what does it look like?
ajs124 has quit [Quit: Gateway shutdown]
* samueldr
wonders if the lightdm default got kept by not setting anything
ajs124 has joined #nixos
mayhewluke has quit [Ping timeout: 250 seconds]
mayhewluke has joined #nixos
sigmundv__ has joined #nixos
srk has quit [Quit: gitter]
<mbrgm>
fpletz: do you have a repo with the VFR related parts of your config you presented at nixcon?
jabranham has quit [Quit: ERC (IRC client for Emacs 27.0.50)]
Ariakenom_ has quit [Quit: Leaving]
lnikkila has quit [Ping timeout: 244 seconds]
Ariakenom has joined #nixos
Ariakenom has quit [Client Quit]
jackdk has joined #nixos
jackdk has quit [Remote host closed the connection]
abathur has joined #nixos
carlosdagos has joined #nixos
vidbina has quit [Ping timeout: 268 seconds]
jackdk has joined #nixos
<jackdk>
mightybyte: did you ever get postgresql to build on darwin? In trying to fix #50902 I think I ran into a ranlib issue you asked about on irc here: https://logs.nix.samueldr.com/nixos/2018-04-16 (search "ranlib"), but nobody answered.