<{^_^}>
[nixpkgs] @Mic92 pushed commit from @c00w to release-19.09 « linux_rpi: copy dtb so raspberry pi 3a+ boots »: https://git.io/Jes5l
zacts has joined #nixos
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
kvda has joined #nixos
mojjo has quit [Ping timeout: 260 seconds]
pie__ has quit [Ping timeout: 245 seconds]
jonringer has joined #nixos
gyroninja has joined #nixos
msafadieh has joined #nixos
<msafadieh>
hello!
<msafadieh>
I've been playing around with nixos for a bit. today I tried to install it on my physical machine but it keeps complaining about missing modules
mexisme has joined #nixos
<msafadieh>
it doesn't complain about the default XMonad modules but it won't import the ones I defined under windowManager.xmonad.extraPackages
drakonis_ has quit [Ping timeout: 276 seconds]
mcbits has quit [Read error: Connection reset by peer]
<ryantm>
OfBorg has a file called outpaths.nix which is used to help figure out how many packages are affected by a code change. I also run it in nixpkgs-update, but it seems like it got broken by something: https://gist.github.com/ryantm/d3131693f31f4fdbfc678613690c5d0d
Zer0xp has quit [Ping timeout: 268 seconds]
gentauro has quit [Read error: Connection reset by peer]
dustinm has quit [Quit: Leaving]
ivan has quit [Quit: lp0 on fire]
gentauro has joined #nixos
<ryantm>
Any suggestions for figuring out what is wrong with `amule` that causes it to break?
rouma7 has quit [Read error: Connection reset by peer]
rouma has joined #nixos
erasmas has joined #nixos
sellout-1 has quit [Read error: Connection reset by peer]
sellout- has joined #nixos
erasmas has quit [Ping timeout: 268 seconds]
mcbits has quit [Ping timeout: 245 seconds]
pie_ has quit [Ping timeout: 276 seconds]
hmpffff has joined #nixos
palo1 has joined #nixos
mcbits has joined #nixos
palo has quit [Ping timeout: 245 seconds]
palo1 is now known as palo
kleisli has quit [Ping timeout: 264 seconds]
nikola_i has quit [Ping timeout: 250 seconds]
truthadjustr has joined #nixos
<truthadjustr>
can linuxrc behave like init to respawn killed child processes?
niksnut has quit [Quit: Lost terminal]
zupo has joined #nixos
rauno has quit [Ping timeout: 264 seconds]
truthadjustr has quit [Quit: leaving]
erasmas has joined #nixos
mojjo has joined #nixos
<mojjo>
Does the Nix language support deep destructuring: like `{pkgs : {bash, git}, config } : ...` ?
erasmas has quit [Ping timeout: 252 seconds]
kleisli has joined #nixos
<danderson>
so, my `nixos-rebuild switch` activation failed (due to something buggy in my own configs). How do I find the store path for the failed system derivation, so I can poke at it?
LysergicDreams has quit [Ping timeout: 240 seconds]
LysergicDreams has joined #nixos
ng0 has joined #nixos
<clever>
danderson: /run/current-system will probably point to it
<clever>
mojjo: import
<clever>
mojjo: also, every time you do `import ./something { inherit pkgs; }` you can instead fo `pkgs.callPackage ./something {};` and it will pass `pkgs` for you
endformationage has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @vcunat pushed to master « Revert "awscli: Get rid of runtime -dev dependencies" »: https://git.io/JesdD
zupo_ has joined #nixos
<{^_^}>
[nixpkgs] @vcunat pushed 221 commits to staging-next: https://git.io/JesdH
mcbits has quit [Read error: Connection reset by peer]
evils has quit [Remote host closed the connection]
veske has quit [Quit: This computer has gone to sleep]
freeman42x has joined #nixos
erasmas has joined #nixos
zupo has joined #nixos
hmpffff has joined #nixos
zupo_ has quit [Ping timeout: 246 seconds]
erasmas has quit [Ping timeout: 268 seconds]
johanot has joined #nixos
<pbb>
Phillemann: I ran into the same issue, even though I don't have any zfs filesystems
<pbb>
the only solutions for now seem to be switching back to an old kernel, or disabling the zfs module entirely to prevent it from trying to evaluate
<Phillemann>
pbb: Doesn't seem to work unfortunately.
zupo_ has joined #nixos
hmpffff_ has joined #nixos
<andi->
Phillemann: Any reason you need the latest kernel right now?
zupo has quit [Ping timeout: 268 seconds]
<andi->
(I am in the same boat, there seems to be patch for 5.3 in the ZFS tree but nobody has applied it on our NixOS tree yet)
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
johanot has joined #nixos
johanot has quit [Client Quit]
c0c0 has joined #nixos
johanot has joined #nixos
hmpffff has quit [Ping timeout: 265 seconds]
syd has joined #nixos
<Phillemann>
andi-: Nah, just wanted to try it out.
mexisme has quit [Ping timeout: 245 seconds]
Ariakenom has quit [Ping timeout: 250 seconds]
LysergicDreams has quit [Ping timeout: 265 seconds]
<c0c0>
when i echo $PS1, i see that it has the value "\n\[\033[1;32m\][\[\e]0;\u@\h: \w\a\]\u@\h:\w]\$\[\033[0m\]". I want that different, so i tried to change it by adding a environment.variables.PS1 = "..."; line, but it gets overridden by the value i posted before. how do i change this?
johanot has quit [Client Quit]
<Phillemann>
I'm experiencing some battery drain with 4.x, not sure if that might be fixed with the latest.
<pbb>
that's the point: why is the assert in the zfs module failing, even when zfs support isn't enabled? this prevents people from using the latest kernel even when they don't use zfs.
<enteee>
clever: i see, but i don't understand why networkmanager suddenly started doing dhcp. The interface gets assigend a valid ip address right after the modem is connected (i do think modem manager sets that ip, not sure though). This ip works until dhcp fails and network manager adds this other ip.
cfricke has quit [Quit: WeeChat 2.6]
kleisli has quit [Ping timeout: 264 seconds]
veske has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
o1lo01ol1o has quit [Remote host closed the connection]
boolman has joined #nixos
<pbb>
Is anyone here familiar with buildBazelPackage and can tell me what I have to do to prevent bazel from downloading dependencies during the second build?
<pbb>
Some dependencies are downloaded in the -deps derivation, but others are not, for example some maven jars, which are included in the WORKSPACE file with maven_jar(
jgt1 has quit [Ping timeout: 245 seconds]
graf_blutwurst has quit [Remote host closed the connection]
<LnL>
gchristensen: I recall you did some work to refresh the manual styling but can't find a link anymore
<gchristensen>
#68020 :)
<{^_^}>
https://github.com/NixOS/nixpkgs/pull/68020 (by grahamc, 2 weeks ago, open): Vacation PR: multi-paged, styled, searchable docs with wrapper tooling and epubs for NixOS and Nixpkgs
<qyliss>
epubs! :o
<LnL>
gchristensen: yeah that, thanks!
<makefu>
exarkun: 14k /quality/ lines to build you a perfectly reproducible environment :)
<exarkun>
makefu: 42 fwiw
<exarkun>
or are only 14k of them quality and the rest are crap? ;)
<makefu>
:D
ee1943 has quit [Ping timeout: 246 seconds]
ee1943 has joined #nixos
<exarkun>
maybe this is fine but I wonder if there is a way to carefully select my stack resolver and my nixpkgs rev so that cabal2nix output (which is like 50 lines) is valid?
rouma has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @rileyinman opened pull request #69351 → nodePackages: Use repo tree as nixpkgs → https://git.io/Jesx2
<{^_^}>
[nixpkgs] @flokli pushed commit from @wkral to release-19.09 « virtualbox: Temporary fix for kernel >= 5.3 »: https://git.io/Jesxr
c0c0 has quit [Ping timeout: 268 seconds]
joehh has joined #nixos
<clever>
exarkun: the old stack2nix generated cabal2nix output only for the packages you depend on
<tokudan[m]>
i'm working on fixing some issues with kicad in master, 19.09 and optionally 19.03. I see that there's a staging-next branch that touches the kicad/default.nix due to some gappsWrapperArgs changes. Should I base my changes on staging-next or on master?
<clever>
exarkun: but at one point, it changed, and it now generates output for every single expr in the entire stackage snapshot
<qyliss>
tokudan[m]: guessing kicad isn't gonna cause mass rebuilds?
<qyliss>
if so, I'd only base on staging if your changes would conflict
erasmas has joined #nixos
<tokudan[m]>
qyliss: probably not causing a mass rebuilt, it's pretty much a standalone application. I'll base my work on staging then
<{^_^}>
[nixpkgs] @Mic92 pushed 2 commits to release-19.09: https://git.io/Jesx9
drakonis_ has joined #nixos
<gchristensen>
LnL: gladly! I reckon I should put together a todo list to merging and get it done before it is too late :x
drakonis_ has quit [Read error: Connection reset by peer]
drakonis_ has joined #nixos
erasmas has quit [Ping timeout: 245 seconds]
<exarkun>
clever: that seems kind of reasonable but it seems like it would be better to generate a separate expression for the snapshot that all your other projects could reference and share (just thinking out loud)
<qyliss>
tokudan[m]: sorry, I meant only base on staging if it's likely to conflict
<clever>
exarkun: which is what haskell.nix is now doing
drakonis has quit [Ping timeout: 240 seconds]
<qyliss>
I think "if so" was the opposite of what I wanted to say
<clever>
exarkun: stackage.nix is a git repo, that contains nix exprs, for every version, of every package, on all of stackage
<clever>
exarkun: and then stack-to-nix will just import the right one, rather then generating it again
<tokudan[m]>
qyliss: I have to touch the same wrapProgram line, so it will cause a conflict
selfsymmetric-mu has quit [Remote host closed the connection]
Tavi has quit [Remote host closed the connection]
kleisli has quit [Ping timeout: 265 seconds]
joehh has quit [Ping timeout: 268 seconds]
rauno has quit [Ping timeout: 245 seconds]
Guest47353 has quit [Ping timeout: 240 seconds]
Zer0xp has quit [Ping timeout: 245 seconds]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
eskytthe has quit [Quit: Leaving.]
<rihardsk[m]>
hi, i'm trying to package some software that's using CUDA. i'm using Nix on Ubuntu. can i just add `cudatoolkit` and `cudnn` to the `buildInputs` in `mkDerivation` and expect the build result to be able to use the GPU on my Ubuntu machine, or is it more complicated than that?
<clever>
rihardsk[m]: the version of the cuda libraries you use, must match the version of the kernel driver
<rihardsk[m]>
currently i've gotten to a point that the package build fine. it detects CUDA and links against it as far as i can tell. but it fails during runtime with some `curand`-related error
<clever>
rihardsk[m]: and the kernel driver must match up to the linux kernel your using
<ivan>
rihardsk[m]: it's more complicated than that on nixpkgs master where you'd have to patch things with addOpenGLRunpath (grep nixpkgs)
<clever>
rihardsk[m]: so its simplest if you just use nixos with cuda
freeman42x has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @alyssais pushed commit from @Taneb to release-19.09 « manticore: 2018.09.29 -> 2019.09.20 »: https://git.io/Jeshr
veske has quit [Ping timeout: 240 seconds]
freeman42x has joined #nixos
<{^_^}>
[nixpkgs] @kmein opened pull request #69356 → tamarin-prover: mark as broken because upstream is broken → https://git.io/Jesho
<ivan>
I didn't see the Ubuntu part so what I said is not particularly relevant
Zer0xp has joined #nixos
<ivan>
rihardsk[m]: nvidia-uvm kernel module should be loaded for CUDA to work
Zer0xp_ has joined #nixos
<rihardsk[m]>
ivan: oh, ok. so i would just have to ensure that the `cudatoolkit` version in `nixpkgs` matches the kernel driver on Ubuntu, as clever said, right?
<clever>
rihardsk[m]: you might need overrides to enforce that, and it will break when ubuntu updates things
graf_blutwurst has joined #nixos
<rsa>
what is the best practice for installing luarocks packages? just use luarocks or write nix-expressions?
<Mic92>
rsa: teto can probably give you the answer.
<rihardsk[m]>
clever: overrides, as in using the different versions of cuda - `cudatoolkit_10_0`, `cudatoolkit_9_2` and such?
<clever>
rihardsk[m]: the problem is that the version you need might not be in nixpkgs
<rsa>
ty Mic92
<clever>
rihardsk[m]: and it will break every time ubuntu updates things
<clever>
rihardsk[m]: nixos avoids that problem, by using the same file to configure both the driver and libraries, so they always match
dansho has quit [Ping timeout: 265 seconds]
<rihardsk[m]>
clever: yeah, that's what i might be seing on my system. there's cuda 10.1 installed on the system, so i guess the driver is compatible with that. the latest cuda version i'm seeing in `nixpkgs` is 10.0, so that might be the problem
<{^_^}>
[nixpkgs] @Mic92 merged pull request #69349 → jetbrains.rider: fix on darwin → https://git.io/Jesxq
<clever>
rihardsk[m]: hence, the need for an override, lookup overrides in the nixpkgs manual
<rihardsk[m]>
@clever but that kindof contradicts what i've seen on Ubuntu otherwise - there multiple different versions of cuda tend to work side by side. i can install cuda 9.2, 10.0 and 10.1 from the nvidia repos, and i think they can peacefuly coexist
freeman42x has quit [Ping timeout: 276 seconds]
erasmas has joined #nixos
<{^_^}>
[nixpkgs] @matthewbauer merged pull request #69309 → androidndk: get correct libs for x86_64 → https://git.io/Jes1q
<{^_^}>
[nixpkgs] @matthewbauer pushed 3 commits to master: https://git.io/JeshA
<clever>
rihardsk[m]: its not specifically the cuda libraries, but the nvidia libraries
<clever>
rihardsk[m]: those have a version# in the 300's
<rihardsk[m]>
clever: yes, i've already taken a look into packaging cuda 10.1, but, unfortunatelly, nvidia packages it differently than 10.0 and as a result it seems to be a bigger undertaking than i'd want. i guess that's why `cudatoolkit_10_1` isn't already available in nixpkgs
<rihardsk[m]>
but regardles, thank's for the information. i'll have more things to check
erasmas has quit [Ping timeout: 240 seconds]
<rihardsk[m]>
i'd be using nixos to avoid all these problems, but i'm not the only user of the system. it'd be relatively easy to convince the other users to try Nix alone. not so much to switch from using Ubuntu to an entirely unfamiliar environment
<{^_^}>
[nixpkgs] @globin pushed commit from @JohnAZoidberg to release-19.09 « pythonPackages.astropy: Disable tests »: https://git.io/JesjS
xwvvvvwx has joined #nixos
hyper_ch2 has joined #nixos
<werner291>
So... I'm back, having once again broken something (sorry about that). For some reason, my NixOS VPS won't boot, an error occurred in stage 1 of the boot process. If I mount /dev/sda2 I can see my files. It says it timed out while waiting for some device to appear.
<clever>
werner291: while waiting for which device to appear?
<werner291>
"Waiting for device /mnt-root/old-root/nixos"
<werner291>
I used NixOS inplace to install it.
<werner291>
Mount then fails with "no such file or directory"
<clever>
werner291: can you gist the hardware-configuration.nix and configuration.nix files?
<{^_^}>
[nixpkgs] @globin pushed commit from @danme to release-19.09 « gnuk: 1.0.4 -> 1.2.14 »: https://git.io/JeGet
<waleee-cl>
What have one done to incur the wrath of nix compiling qtbase 5.12.3 from source, even if you change all suspected packages to the nixos-channel?
<clever>
waleee-cl: overrides and overlays
<waleee-cl>
well, if 2bwm counts, but that doesn't have a whiff of qt in its dependencies
<clever>
waleee-cl: comment it out and see if any changes happen
<werner291>
clever: That last one is the only one that mentions filesystems and directories, whatever caused this was probably introduced earlier and only shows up now that I rebooted.
graf_blutwurst has quit [Remote host closed the connection]
orivej has quit [Ping timeout: 276 seconds]
<werner291>
old-root doesn't have a nixos directory in it
xkapastel has quit [Quit: Connection closed for inactivity]
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #69361 → systems/parse.nix: use split builtin if available → https://git.io/JeGve
<sonercirit[m]>
I'll wait thx a lot alexarice :)
civodul has joined #nixos
<werner291>
Ok, I think that it should be mounting sda2 instead of sda1. Everything it needs is in sda2. Is it possible for those two to get swapped somehow?
<{^_^}>
[nix] @edolstra pushed to flakes « Shut up some warnings »: https://git.io/JeGvv
zacts has joined #nixos
Zer0xp has joined #nixos
Tucky has quit [Ping timeout: 240 seconds]
jgt1 has joined #nixos
Zer0xp_ has joined #nixos
gxt has joined #nixos
<werner291>
Because the "nixos" directory lives on sda2 but it's trying to mount sda1
jgt1 has quit [Read error: Connection reset by peer]
justanotheruser has joined #nixos
<{^_^}>
[nixpkgs] @arianvp opened pull request #69363 → nixos/networkd: Show that issue #45934 is fixed → https://git.io/JeGvt
<qyliss>
werner291: yes, you shouldn't really on those names
<{^_^}>
[nixpkgs] @arianvp closed pull request #69363 → nixos/networkd: Show that issue #45934 is fixed → https://git.io/JeGvt
<qyliss>
use /dev/disk/ instead
<werner291>
I shouldn't what?
<qyliss>
/dev/sd* are assigned in the order they're initialized
<Taneb>
werner291: rely on them
<qyliss>
sorry, you shouldn't rely on those names
<werner291>
Oh those CAN get swapped?
<gchristensen>
sda{1,2,3..} will always keep that order but sda could be come sdb
<exarkun>
is it an error to run stack2nix from one version of nixpkgs and then evaluate the expression with a different version of nixpkgs?
erasmas has joined #nixos
<Taneb>
...what happens if you have more than 26 drives?
<{^_^}>
[cabal2nix] @peti pushed to master « Fix build with ghc 8.8.1. »: https://git.io/JeGv3
<exarkun>
Taneb: sdaa of course
<Taneb>
exarkun: I see
<clever>
Taneb: i saw a blog, where a guy tried making 20,000 drives, just to see what would break first
<clever>
and performance began to suffer massively, because one of the layers had a plain array, and searched it one by one
<exarkun>
The expression stack2nix generated is failing to evaluate because "called without required argument 'text'"
<{^_^}>
[cabal2nix] @peti pushed to master « Fix build with ghc 8.8.1. »: https://git.io/JeGvC
foxmean has joined #nixos
<foxmean>
Hello, I have a little problem to ask for help. The problem is I cannot initialise pdf-tools with: (pdf-tools-install). I am using Steve Purcell’s setup. My machine is install Void GNU/Linux with co-using nix & xbps package manager. This is compilation https://paste.debian.net/1102372/. And this is Backtrace https://paste.debian.net/1102374/. I think I need to do something to add Nix search path. And I alredy know that I have to "add it
<foxmean>
using $NIX_PATH or -I" but I don't know where to run this command.
erasmas has quit [Ping timeout: 240 seconds]
hmpffff has joined #nixos
mmlb has joined #nixos
<LnL>
gchristensen: yeah I like it, feels easier to navigate then the giant page
<gchristensen>
LnL: awesome :D
<gchristensen>
LnL: some minor problems (building docs takes too long long probably, for example) but hopefully we'll merge this release
<{^_^}>
[nixpkgs] @peti pushed 2 commits to haskell-updates: https://git.io/JeGvu
<mdash>
foxmean: i wouldn't expect any install function like that to work with nix-installed packages
erasmas has joined #nixos
zupo has joined #nixos
<LnL>
gchristensen: hmm, would it be possible to split up the pages into separate builds?
zupo_ has joined #nixos
<gchristensen>
yeah, but that isn't the source of the problem: calculating the ToC on all the pages is very lows
<gchristensen>
slow*
<Athas>
What's the best way of keeping my configuration.nix under version control? Just symlink it to something in my home directory?
<tilpner>
Athas: Checkout your repo anywhere, point nixos-rebuild at it via NIX_PATH
<tilpner>
NIX_PATH=nixos-config=/path/to/your/checkout is one way
zupo_ has quit [Client Quit]
zupo has quit [Ping timeout: 240 seconds]
wfranzini has quit [Remote host closed the connection]
civodul has quit [Read error: Connection reset by peer]
<Athas>
I'm not particularly enthusiastic about having to remember typing that for every nixos command, nor to define yet another set of aliases. ISn't there a smoother way?
<tilpner>
Athas: The symlink you mentioned will just work
<tilpner>
Athas: But you only have to set NIX_PATH once every time you change it
<tilpner>
Athas: The other times, you will have set it globally in the options nix.nixPath
<{^_^}>
[nixpkgs] @peti pushed 4 commits to haskell-updates: https://git.io/JeGv6
orivej has quit [Ping timeout: 240 seconds]
__Sander__ has quit [Quit: Konversation terminated!]
sigmundv has quit [Ping timeout: 245 seconds]
mmlb has quit [Remote host closed the connection]
growpotkin has joined #nixos
<Athas>
tilpner: I don't follow. NIX_PATH is an environment variable, so I will have to make sure it is set whenever I run a nixos-* command. nix.nixPath is a configuration for overlays to nixpkgs, right?
<tilpner>
Athas: nix.nixPath will set NIX_PATH globally (may require a re-login)
<tilpner>
Athas: So you don't need to manually provide NIX_PATH, except for the transitions
<{^_^}>
[nixpkgs] @peti pushed 9 commits to haskell-updates: https://git.io/JeGvD
<Athas>
Ah. I guess that may work. Any advantage over symlinking all of configuration.nix?
<tilpner>
(Because the config can only become active after building and switching, but you need it before)
hmpffff has quit [Quit: nchrrrr…]
<tilpner>
Athas: Oh, one more. You can also just git init in /etc/nixos
<Athas>
I can change the owner of the files in there without trouble?
<tilpner>
Sure, that should be fine
gxt has quit [Ping timeout: 260 seconds]
<gchristensen>
just make sure only users you want to be able to change it, can change it :)
<tilpner>
gchristensen: If they chown to their primary user, all three methods are about equally vulnerable
kleisli has quit [Ping timeout: 240 seconds]
<gchristensen>
heh
<{^_^}>
[cabal2nix] @peti pushed 0 commits to refs/tags/v2.15.0: https://git.io/JeGvH
<tilpner>
If someone can change the link (from presumably their home directory), they can probably also change NIX_PATH in the users shell
<Athas>
Hm, if I symlink just configuration.nix, then it looks like hardware-configuration.nix is looked up at the symlink destination, rather than in /etc/nixos.
<clever>
Athas: what i do, is add `imports = [ /home/clever/nixos-configs/whatever.nix ];` to configuration.nix
<clever>
and leave configuration.nix in /etc/nixos/
<clever>
inkbottle: this script supports /boot being vfat, ext4, or just a directory on / (which is zfs)
<clever>
/boot being on zfs "works" but all attempts to debug it claim it shouldnt
<clever>
which means that when it does fail, all attempts to debug it will fail
growpotkin has quit [Ping timeout: 245 seconds]
<inkbottle>
I was looking for on of the 2 strings: ef00, C12A7328-F81F-11D2-BA4B-00A0C93EC93B
screpie has joined #nixos
drakonis_ has quit [Ping timeout: 246 seconds]
patrl has joined #nixos
<clever>
inkbottle: ef00 is shorthand for a longer uuid
<inkbottle>
yes, but the second string is that longer uuid
<clever>
now that you mention it.....
<clever>
i'm using the wrong uuid type on /boot!
<clever>
both root and /boot have "Linux filesystem data"
zupo has quit [Ping timeout: 265 seconds]
zupo has joined #nixos
zupo has quit [Client Quit]
nikola_i has quit [Ping timeout: 245 seconds]
zupo has joined #nixos
romildo has joined #nixos
zupo has quit [Client Quit]
<{^_^}>
[nixpkgs] @peti pushed 2 commits to haskell-updates: https://git.io/JeGfE
drakonis_ has joined #nixos
zupo has joined #nixos
justan0theruser has joined #nixos
judson has joined #nixos
judson__ has joined #nixos
<werner291>
So... assuming that sda1 and sda2 device labels got swapped, is it possible to edit the init script just so I can get the system to boot once so I can fix my config?
drakonis has joined #nixos
justanotheruser has quit [Ping timeout: 245 seconds]
hmpffff has joined #nixos
<gchristensen>
sda1 and sda2 won't get swapped, as they are not separate devices
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<gchristensen>
(sda and sdb might swap)
hmpffff has quit [Client Quit]
<werner291>
Well, it does appear to somehow have happened. My system is complaining about not finding old-root/nixos after mounting sda1, but I can see that file in sda2.
jgt1 has quit [Ping timeout: 265 seconds]
<__monty__>
clever: Do any tools actually change behavior based on the fs type?
<clever>
__monty__: the bios is supposed to only boot from a correctly tagged esp partition...
<clever>
but with how sloppy every motherboard firmware is....
fusion809 has joined #nixos
<__monty__>
Yes, the abundance of fs types just seems much ado about nothing. Which is too bad, I'm sure there's efficiency to be had if you can identify a partition by its fs type.
<{^_^}>
[hydra] @edolstra merged pull request #432 → Only set buildMachinesFiles when nix.buildMachines is defined → https://git.io/JeGfy
<emily>
clever: btw, you're meant to use "Solaris root" for zfs.
Ariakenom has quit [Remote host closed the connection]
<clever>
emily: i mostly just ignore the partition type codes, and trust blkid instead
<gchristensen>
emily: would you be interested in helping ZoL do more frequent releases, to support more current kernels?
<emily>
doesn't sound inherently uninteresting, but between being completely unfamiliar with both mammoth C codebases involved (the kernel and OpenZFS) it sounds somewhat daunting
<gchristensen>
aye
<emily>
(ok, I guess I've read a few kernel drivers in my time, but it's mostly a mystery)
o1lo01ol1o has quit [Remote host closed the connection]
Stuck_ has joined #nixos
magnetophon has joined #nixos
justan0theruser has quit [Ping timeout: 265 seconds]
<Stuck_>
hi, I want to use a vim plugin that is not in nixos19.03 vimPlugins, how can I do it in a nix way?
sigmundv has joined #nixos
<waleee-cl>
Stuck_: search for buildVimPluginFrom2Nix
<exarkun>
is there smart merging for that value? or if I set it once do I have to make sure the value includes everything I want?
<clever>
its a list type, so if you set it several times, it will concat the lists
<exarkun>
interestingly the key is still there, only the substituter address itself is gone. not sure if that's because of something else I did or not.
<enteee>
Does somebody know how to disable APIPA (after failed dhcp) on a networkmanager / modemanager managed gsm interface (ip config happens over ppp)? Or why nixos tries dhcp on such an interface in the first place?
jgeerds has joined #nixos
<exarkun>
Okay stack-to-nix looks like it might be nice. it's certainly a lot faster and generates more succinct output than stack2nix. I still can't figure out how to use its output, though. What I think should be a derivation, nix-build complains is a set...
<clever>
exarkun: run `nix repl` on that file
<clever>
exarkun: oh, and first, `import ./that-file` in nix repl
<vaibhavsagar>
Is there a `nub` for lists in nixpkgs.lib?
<exarkun>
clever: it imports. The expression I thought would be a derivation is instead `{ cabalFile = null; components = { ... }; identifier = { ... }; isHaskell = true; setup = «derivation /nix/store/4h8h63a9zc1a5nzxjn8iv0p5r8zswjlk-default-Setup.drv»; }`. So ... something about the `setup` attribute?
<clever>
exarkun: thats just the Setup.hs file
<exarkun>
ah
<clever>
exarkun: what is in the components set?
ris has joined #nixos
<exarkun>
clever: some good looking stuff, I guess. like an `exes` set with the exe I was looking for as an attribute
<ilyagotfryd>
Have a Python and distutils in NixOS question. I have an application that I am building a package for in NixOS. When I create Python application package from a local folder it runs setup.py and creates a package perfectly fine. HOwever when I fetch the same package from GitLabb and use that retrieved source code as the source fo my package I get an issue during isntallation phase of package building
<ilyagotfryd>
it is unable to find suitable package version for my packages, as if it is not connecting to pypi
LysergicDreams has quit [Ping timeout: 265 seconds]
<ilyagotfryd>
I know I should ideally be using Nix to manage my Python packages, but I am trying to solve one problem at a time
mmlb0 has joined #nixos
<exarkun>
ilyagotfryd: does the python packaging look at .git to figure out the version? like, using setuptools_scm or something?
LysergicDreams has joined #nixos
<ilyagotfryd>
@exarkun: I am not sure. How would you verify that?
<ilyagotfryd>
I am pulling in all of the packages that are on Pypi
<ilyagotfryd>
none of them are installed through git themselves
<exarkun>
What do you mean by this, exactly? "HOwever when I fetch the same package from GitLabb and use that retrieved source code as the source fo my package"
zupo_ has joined #nixos
<ilyagotfryd>
So my Python app is in GitLab repository.
zupo has quit [Ping timeout: 265 seconds]
psyanticy has quit [Quit: Connection closed for inactivity]
<ilyagotfryd>
To create app packaging initially I used local copy of it in a folder and used that folder as a `src`
<ilyagotfryd>
once I got that packaging to work I switch the `src` over to fetch data from GitLab (I had to create my own modification to fetch from a private repo)
<ilyagotfryd>
so once I started getting this from my private repo it would build OK but then it would run install to verify package contents and it would fail with the following error:
<ilyagotfryd>
```
mmlb0 is now known as mmlb
<ilyagotfryd>
ERROR: Could not find a version that satisfies the requirement ultra-config (from bvla-engine==0.0.7) (from versions: none)
<ilyagotfryd>
ERROR: No matching distribution found for ultra-config (from bvla-engine==0.0.7)
<ilyagotfryd>
ultra-config is just a package in pypi, I also tried to remove that package to see if it is that one in particular, it simply moves on to fail on another package
<exarkun>
ultra-config is not packaged in nixos. I would expect that you would need to package it. And whichever other python packages this project has as a dependency which aren't packaged in nixos.
<ilyagotfryd>
so nix-build sourcing a local folder goes over swimingly with all of these packages being pulled from pypi and not nixpkgs
<ilyagotfryd>
also when I am removing ultra-config it fails on the next one.
<ilyagotfryd>
It looks like the error is coming from pip or easyinstall and not nix per se
growpotkin has quit [Ping timeout: 240 seconds]
<exarkun>
I don't understand how or why it would be downloading packages from pypi and succeeding.
<exarkun>
That is not compatible with my understanding of how the Python support in Nixpkgs works.
<{^_^}>
[hydra] @edolstra pushed to flake « Build the container in Hydra »: https://git.io/JeGTn
<exarkun>
Maybe you can paste your setup.py and .nix file somewhere
<arianvp>
has anybody else had nix copy hang before?
<arianvp>
I can download and upload just fine to my server but nix copy hangs after copying around 5 MB
arianvp has quit [Quit: WeeChat 2.6]
arianvp has joined #nixos
o1lo01ol1o has joined #nixos
<ilyagotfryd>
exarkun you can either pull packages into environmet by importing them into build configuration and passing them through to buildPythonApplication or you can include them into setup.py as you normally would and it will just fetch them for you at install time.
<ilyagotfryd>
At least that is what it does for me when installed from local folder.
<exarkun>
Yes. I don't think that's true.
<ilyagotfryd>
It is recommended to do through nixpkgs but it is not mandatory
<exarkun>
I do not believe you can package Python software for NixOS by just letting pip download a raft of crap from PyPI.
<{^_^}>
[nixpkgs] @d-goldin opened pull request #69376 → epson-escpr2: bumping version to 1.1.1 → https://git.io/JeGTa
<exarkun>
I don't know where the contradiction comes from. Possibly I could tell by looking at your nix expressions, though I'm far from an expert so there's hardly a guarantee there.
<ivan>
I set INET_DIAG INET_TCP_DIAG INET_UDP_DIAG INET_RAW_DIAG INET_DIAG_DESTROY to yes
o1lo01ol1o has quit [Remote host closed the connection]
<ivan>
oh maybe I need = module;
<{^_^}>
[nixpkgs] @grahamc opened pull request #69379 → perlPackages.NetPrometheus: init at 0.07 (and StructDumb at 0.09) → https://git.io/JeGT7
tobiasBora has quit [Read error: Connection reset by peer]
<ilyagotfryd>
emily: NP, thanks.
growpotkin has quit [Ping timeout: 245 seconds]
<samueldr>
I have this systemd unit that I want to disable/mask, but can't just disable the option, and setting wantedBy to [] is not enough since other units wants = [ it.service ]
<samueldr>
do we have a fancy way to disable or "mask" a service that will exist?
<tobiasBora>
I'd like to know, does nix provides a way to setup discourse on a server?
<tobiasBora>
(I can see that the community uses discourse, so I'd expect to find this kind of integration inside nix, but I can't find any, even by searching nixpkgs)
aleph- is now known as Judge_Aleph-
<{^_^}>
[nixpkgs] @peterhoeg opened pull request #69380 → dxx-rebirth: build with gcc6 as gcc8 is not supported → https://git.io/JeGk7
<tilpner>
$ dig +short discourse.nixos.org CNAME
<tilpner>
nixos1.hosted-by-discourse.com.
<tilpner>
It's not selfhosted
<{^_^}>
[hydra] @grahamc opened pull request #680 → export a /prometheus endpoint → https://git.io/JeGIk
<{^_^}>
[nixpkgs] @peterhoeg reopened pull request #69380 → dxx-rebirth: build with gcc6 as gcc8 is not supported → https://git.io/JeGk7
<esclear>
Can anyone here confirm that the vlc from nixos-unstable is working with UPNP? I hoped that I could work around the missing sftp support that way.
<tobiasBora>
infinisil: oh, that' great!
o1lo01ol1o has quit [Remote host closed the connection]
<andi->
gchristensen: probably better then nothing :-)
asymptotically2 is now known as asymptotically
<LnL>
gchristensen: I case you didn't know yet, hydra does export some metrics to statsd
<gchristensen>
LnL: yeah, what do you think about re-exporting those to prom?
o1lo01ol1o has joined #nixos
<{^_^}>
[nixpkgs] @grahamc opened pull request #69381 → Backport: perlPackages.NetPrometheus: init at 0.07 (and StructDumb at 0.09) #69379 → https://git.io/JeGIB
<LnL>
not sure what's in there exactly but might be nice to include (some) of those
<tobiasBora>
infinisil: if I wanted to self host it without this PR, is it still possible to rely on docker directly?(I'm asking because I may be interested by Talkyard, which does not have any module, even in PR)
<{^_^}>
[nixpkgs] @peterhoeg merged pull request #69300 → nixos/home-assistant: set bluetooth perms → https://git.io/Jes6Y
<DigitalKiwi>
aanderse: which makes sense because adding the new package was the last thing i did and it didn't work before that so i don't know why it'd fix
sigmundv has joined #nixos
acarrico has quit [Ping timeout: 245 seconds]
Ariakenom has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @peterhoeg opened pull request #69387 → nixos/nix-optimise: be smarter about when we run the store optimiser → https://git.io/JeGqe
<DigitalKiwi>
gchristensen: is there a template/guide for how to make a nixops modulee?
<DigitalKiwi>
by template i mean minimal example
<{^_^}>
[nixpkgs] @peterhoeg merged pull request #69383 → dxx-rebirth: build with gcc6 as gcc8 is not supported → https://git.io/JeGtB
<{^_^}>
[nixpkgs] @peterhoeg pushed 2 commits to release-19.09: https://git.io/JeGqf
<{^_^}>
[nixpkgs] @ivan opened pull request #69388 → kernel/common-config: enable INET_{TCP,UDP,RAW}_DIAG and INET_DIAG_DE… → https://git.io/JeGqJ
<{^_^}>
[nixpkgs] @jonringer opened pull request #69392 → dotnet-sdk_3: init at 3.0.100 → https://git.io/JeGqu
Tucky has quit [Quit: WeeChat 2.6]
ilyagotfryd has quit [Remote host closed the connection]
<pikajude>
how often does nixpkgs-unstable actually update? seems pretty stable to me
ilyagotfryd has joined #nixos
nikivi has quit [Ping timeout: 276 seconds]
alanz has quit [Ping timeout: 264 seconds]
<samueldr>
as often as it will, when tests pass
syd has joined #nixos
<samueldr>
though nixpkgs-stable needs both linux and darwin tests to pass
<samueldr>
if you're using Linux only, even non-nixos, you may prefer to use nixos-unstable
ilyagotfryd has quit [Ping timeout: 240 seconds]
syd has quit [Remote host closed the connection]
<DigitalKiwi>
aanderse: yeah i just don't think you're allowed to use it there, nobody else has it in docs either, and another package fails when i modify it to have fpm.socket in a description
jgeerds has quit [Ping timeout: 265 seconds]
freeman42x has joined #nixos
is_null has joined #nixos
is_null_ has joined #nixos
jgt1 has joined #nixos
joehh has joined #nixos
noudle has quit [Ping timeout: 245 seconds]
growpotk- has joined #nixos
armin has quit [Ping timeout: 245 seconds]
jgt1 has quit [Ping timeout: 246 seconds]
* justan0theruser
jgt1 has joined #nixos
justan0theruser has quit [Quit: WeeChat 2.4]
justanotheruser has joined #nixos
<{^_^}>
[nixpkgs] @hlolli opened pull request #69393 → yakyak init at 1.5.4-beta-rolling and (patched) electron-packager init at 14.0.6 → https://git.io/JeGmO
joehh has quit [Ping timeout: 268 seconds]
joehh has joined #nixos
energizer has joined #nixos
dermetfan has quit [Ping timeout: 240 seconds]
<energizer>
is there somewhere that old versions of packages are kept? for example, if i wanted to get CPython 3.4
is_null has quit [Ping timeout: 265 seconds]
dansho has joined #nixos
is_null has joined #nixos
<ivan>
there is no CPython 3.4 in nixpkgs 19.09 but you could get it from an older release
<energizer>
in that case, why remove old packages?
<samueldr>
the *package set* itself is what is currently maintained
<samueldr>
old gcc versions aren't maintained in the current package set
<samueldr>
what is in nixpkgs is what is needed and used by nxpkgs and its users, and maintainers
<samueldr>
its goal is not to become a library of all versions of all software at all times :)
<energizer>
yeah i guess i'm asking "why not" :-)
<samueldr>
though nix being stable, and working hard in staying compatible, makes it so it's easy to use older versions
<{^_^}>
[nixpkgs] @esclear opened pull request #69394 → vlc: Add libssh2 as dependency to support sftp inputs → https://git.io/JeGmu
mexisme has quit [Quit: WeeChat 2.6]
mexisme has joined #nixos
<DigitalKiwi>
aanderse: i don't even think it needs to be there anyway and what i have isn't necessarily the right path either since it could be changed so
o1lo01ol1o has quit [Remote host closed the connection]