<haisheng>
I got error with `nix-shell`: error: value is a function while a set was expected, at /Users/haishengwu/Downloads/haskell/haste-compiler/shell.nix:1:1
betawaffle has joined #nixos
lopsided98 has joined #nixos
jluttine has quit [Ping timeout: 240 seconds]
<red[m]>
QuickQ: I'm opening a PR against a module and afaict modules don't have maintainers? Who should I tag (other than the person who did the previous commit?)
<{^_^}>
[nixpkgs] @redvers opened pull request #73086 → nixos/httpd: module fixes enableUserDir (attendum to #72789) → https://git.io/JeV0T
wildtrees has quit [Quit: Leaving]
detran has quit [Ping timeout: 240 seconds]
<samueldr>
looking at the previous committers is the right thing to do, red[m]
detran has joined #nixos
<samueldr>
maybe look at the relevant packages too
Fare has quit [Ping timeout: 240 seconds]
alexherbo2 has quit [Ping timeout: 246 seconds]
jgt has quit [Ping timeout: 252 seconds]
Fare has joined #nixos
<red[m]>
samueldr: the other biggest contributor to the package is edolstra so I don't really want to tag him ;-)
<samueldr>
depends on if it was recent, I guess, maybe it is fine if it is still something he's using
<__red__>
well, it's apache so not exactly the most trivial of packages ;-)
<evanjs>
How do I get outPaths for a script? e.g. zlibPath=$(nix-instantiate --eval -E 'with import <nixpkgs> {}; builtins.toPath (zlib.outPath + "/lib/libz.so.1")') returns something but passing it directly to patchelf doesn't seem to work
<evanjs>
i.e. it's a string
<infinisil>
__red__: Oh tag aanderse, I think he's interested in making the httpd module better
<__red__>
yup - he was who I tagged
<evanjs>
And gives me something like "/nix/store/gpszqcy0xi0lavbbjdq82zkkjp3jbp2a-bzip2-1.0.6.0.1-bin/lib/libbz2.so.1" => not found
<__red__>
I also want to talk to him about something else he's suggested doing that I want to make sure he makes optional :-)
<evanjs>
I do wonder if I should just put it in a shell.nix or something
o1lo01ol1o has joined #nixos
<clever>
eyJhb: 90% of the time, you dont want toPath
o1lo01ol1o has quit [Remote host closed the connection]
<lordcirth>
samueldr, I started building aarch64 using your cross-system repo
<samueldr>
and? :)
o1lo01ol1o has joined #nixos
evil-olive has quit [Quit: Ping timeout (120 seconds)]
<lordcirth>
And it's still compiling :P
<lordcirth>
2-core VM
<samueldr>
:) could take some time
tdeo has quit [Ping timeout: 268 seconds]
<lordcirth>
It's been an hour now, I will reboot the VM with 4 cores
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
slack1256 has quit [Remote host closed the connection]
jgt has joined #nixos
jluttine has joined #nixos
evil-olive has joined #nixos
tdeo has joined #nixos
tdeo has joined #nixos
tdeo has quit [Changing host]
<evils>
anyone here particularly familiar with WrapGAppsHook? i need to wrap 1 of several binaries
jgt has quit [Ping timeout: 252 seconds]
Ariakenom has quit [Quit: Leaving]
iyzsong has joined #nixos
jluttine has quit [Ping timeout: 265 seconds]
Ariakenom has joined #nixos
<DigitalKiwi>
paging worldofpeace
Ariakenom has quit [Remote host closed the connection]
thc202 has quit [Ping timeout: 250 seconds]
<{^_^}>
[nixpkgs] @MetaDark opened pull request #73087 → textlab: init at 1.6.0 → https://git.io/JeV0M
<lordcirth>
samueldr, I believe you said earlier that I would need a mainline kernel?
<samueldr>
either of the BSP kernel (booh) or a mainline kernel with appropriate patches (to be found)
ris has quit [Ping timeout: 276 seconds]
jluttine has joined #nixos
<samueldr>
likely you could track the source of the manjaro one
<samueldr>
(which is based on mainline)
dsx has quit [Quit: dsx]
<lordcirth>
samueldr, hmm. If I do get a kernel derivation working, say based on manjaro, how would one help push it upstream? A patched derivation in nixpkgs, or just wait for the patches to land in mainline?
<samueldr>
it will never be accepted upstream in nixos, the patches are likely to land soon in mainline once written up :)
dsx has joined #nixos
gentauro has quit [Read error: Connection reset by peer]
<samueldr>
but said patches can be documented on the NixOS on ARM pages on the wiki
<samueldr>
(can and will)
gentauro has joined #nixos
selfsymmetric-pa has joined #nixos
mexisme has joined #nixos
h0m1 has quit [Ping timeout: 276 seconds]
<lordcirth>
Ok, cool
<{^_^}>
[nixpkgs] @nixos-channel-bot pushed 2 commits to nixos-19.03-small: https://git.io/JeV4p
<lordcirth>
samueldr, the uboot build outputs: "CONFIG_OF_EMBED is enabled. This option should only be used for debugging purposes."... etc. Does that matter?
<lordcirth>
Also, if I use nixpkgs master, bash doesn't build
xkapastel has quit [Quit: Connection closed for inactivity]
o1lo01ol1o has quit [Remote host closed the connection]
<samueldr>
been a couple months since I last checked whether cross-system works
<lordcirth>
Actually no, it's "/nix/store/0mjqs0y4af5fnhy1cb4bnh1b1spxnvk8-groff-1.22.4-aarch64-unknown-linux-gnu.drv"
<lordcirth>
The error is /nix/store/zavn4np1jvm79f0rafkv0p1mrag09qkz-bash-4.4-p23/bin/bash: ar: command not found
<samueldr>
ah :)
<lordcirth>
Just a missing dependency?
<samueldr>
sounds more like something is not aware of cross-compilation
<samueldr>
ar would not be named "ar" but something else I think with cross-compilation
<samueldr>
something like aarhc64-linux-gnu-ar maybe
<lordcirth>
ah, I see
<lordcirth>
Perhaps I should try unstable instead of master?
<samueldr>
yeah, so you don't have to care too much about CONFIG_OF_EMBED for now, but it looks like the clean way is using CONFIG_OF_SEPARATE, which is the only way accepted for mainline
<samueldr>
I always go with unstable rather than master, if only for more artifacts in the cache
<samueldr>
but here not sure it will matter that much
* samueldr
looks
evil-olive has joined #nixos
tdeo has joined #nixos
tdeo has joined #nixos
tdeo has quit [Changing host]
o1lo01ol1o has joined #nixos
davidak has quit [Quit: Leaving]
jgt has joined #nixos
<lordcirth>
Still trying to wrap my head around sd-image.nix
jgt has quit [Ping timeout: 252 seconds]
<{^_^}>
[nixpkgs] @nixos-channel-bot pushed 15 commits to nixos-19.09: https://git.io/JeVEL
o1lo01ol1o has quit [Remote host closed the connection]
<samueldr>
lordcirth: looks like XML-Parser is the only other issue
<lordcirth>
samueldr, Would it be practical to add cross-compile tests for packages that are likely to break?
<samueldr>
it sure would be useful
<samueldr>
there was high level talks lately about maybe doing such a thing
<samueldr>
(and other related things
<rotaerk>
hmm, I've never used UEFI before. I'm trying to understand how to setup dual-boot between nixos and windows 10. windows 10 is installed, and it created a boot partition. should I ... overwrite that from linux, or create a second boot partition? when I used BIOS, I basically *replaced* whatever boot partition windows created and put GRUB in the new one ...
<samueldr>
lordcirth: looks like you can git revert 076dc17149af60635d52621e3a2a341af046af2e to unstick yourself until a real fix is found
<lordcirth>
Cool, thanks
<samueldr>
not 100% confirmed to go to the end with that revert, but it unstuck that issue
<lordcirth>
Well, I'll leave the -k build running for now
<lordcirth>
Any idea why a minor version bump broke perl5? I guess I might need to dig into the source
<samueldr>
perl is perfectly fine
<samueldr>
it's XML-Parser that's broken :)
<samueldr>
a perl... package? well, whatever name they use for their thingies
<{^_^}>
[nixpkgs] @jonringer pushed commit from @groodt to master « fluxctl: 1.13.3 -> 1.15.0 »: https://git.io/JeVuT
<evils>
worldofpeace: maybe i'm overlooking something, but that says nothing about only wrapping 1 of several binaries
lovesegfault has quit [Quit: WeeChat 2.6]
<worldofpeace>
evils: it doesn't, I believe we overlooked documenting manually wrapping.
<evils>
i've seen explicit use of WrapGApps, but that seems to not work as intended for kicad
jb55 has quit [Ping timeout: 260 seconds]
drakonis has quit [Quit: WeeChat 2.6]
jgt has quit [Ping timeout: 246 seconds]
iyzsong has joined #nixos
<evils>
actually, kicad as is on master uses wrapGApp, but with #72813 it breaks, setting `dontWrapGapps = false;` fixes the launcher and breaks everything else, manually adding XDG_DATA_DIR to that command fixes it again, but adds the path twice
<samueldr>
hmmm... gobject-introspection being brought into polkit and/or udisks is going to be an issue for cross-compilation :/
<samueldr>
gobject-introspection seems to not play nicely with the cross-compilation stuff currently, will need to circle back later
<evils>
worldofpeace: jup, that's basically what kicad currently does
<samueldr>
likely an issue for mobile-nixos too
<samueldr>
it needs both a build-time python and run-time python AFAICT
<worldofpeace>
evils: this is because it's missing things buildInputs
<evils>
worldofpeace: yea so i recently fixed this issue in mattermost-desktop, where `gtk` is in buildInputs, but with kicad, `wxGTK` is already an input, which i would think includes `gtk`
<worldofpeace>
If you read https://nixos.org/nixpkgs/manual/#ssec-gnome-hooks, it explains that all wrapGAppsHook does it set environment variables with values collected by other setup hooks. in particular, with gsettings, the glib setup hook needs to be ran to populate the environment variable GSETTINGS_SCHEMAS_PATH so it can prepend it to XDG_DATA_DIRS.
drakonis has joined #nixos
<worldofpeace>
wxGTK is complied with gtk, but it needs glib and gtk in buildInputs because it depends on their setup-hooks being ran
<worldofpeace>
So the solution is to add wxGTK.gtk and glib to buildInputs
<worldofpeace>
we use wxGTK.gtk attribute because it'll point to the correct gtk version, either gtk2 or gtk3, depending on if wxGTK was overriden with withGtk2 = true or false.
<{^_^}>
[nixpkgs] @jonringer pushed commit from @lodi to master « libco-canonical: v19.1 -> v20 »: https://git.io/JeVuO
<slack1256>
evils: Read the commit that introduced that change?
<slack1256>
It can something that made sense long ago.
<slack1256>
can be
<worldofpeace>
no idea, I've reviewed a PR today having an issue with that. gtk2 needs to go so I'd love a PR to change it to gtk3. not familar with wx widgets though.
bvdw has quit [Read error: Connection reset by peer]
bvdw has joined #nixos
waleee-cl has quit [Quit: Connection closed for inactivity]
<lopsided98>
wxPython also needed to be configured to use GTK3
Thra11 has joined #nixos
abathur has quit [Ping timeout: 276 seconds]
<lopsided98>
I tried running nix-review, but ran into issues - I might try it again now
<lopsided98>
There might also be runtime/UI breakage for certain applications
mexisme has quit [Ping timeout: 246 seconds]
ddellacosta has joined #nixos
jedai42 has joined #nixos
Ninlives has joined #nixos
<Ninlives>
Can I override a drv's stdenv with a stdenv produced by override the original stdenv? i.e. `self: super: { pkg = super.pkg.override { stdenv = super.pkg.stdenv.overrde { ... }; } ;}`, but this will not work because the `super.pkgs.stdenv` is not overrideable
dsx has quit [Quit: dsx]
ddellacosta has quit [Ping timeout: 240 seconds]
jedai42 has quit [Ping timeout: 265 seconds]
joehh has quit [Remote host closed the connection]
slack1256 has quit [Remote host closed the connection]
felixfoertsch23 has quit [Quit: ZNC 1.7.3 - https://znc.in]
<Ninelives>
Is there any example of using `extraSandboxProfile` ?
leungbk has joined #nixos
Rusty1 has quit [Quit: Konversation terminated!]
<evils>
worldofpeace: seems like i don't need to pass glib and gtk, `gsettings-desktop-schemas-3.34.0` doesn't get added to XDG_DATA_DIRS but everything still works. regardless, `gtk+3-3.24.12` gets added twice, is that normal?
slack1256 has quit [Remote host closed the connection]
slack1256 has joined #nixos
slack1256 has quit [Remote host closed the connection]
Ninelives has quit [Remote host closed the connection]
<leungbk>
I'm trying to write derivations for some emacs packages. I am trying to obtain a sha256 hash for a package with an existing declaration to make sure I understand how to do everything. The command `nix-hash --type sha256 --flat --base32 railgun.el` gives me a different hash from what exists in the nixpkgs repo. Is that not the right command? I've also tried hashing the git repo for that existing package, but the hash I produced had more
<leungbk>
characters than what is in the repo.
ddellacosta has joined #nixos
fusion809 has joined #nixos
ddellacosta has quit [Ping timeout: 268 seconds]
<wedens[m]>
leungbk: I think the hash is computed over nar archive
<wedens[m]>
in fetchgit
<leungbk>
wedens[m]: I see, thanks.
leungbk has quit [Quit: ERC (IRC client for Emacs 26.3)]
<rotaerk>
I've just been using xterm; wonder if I should try something else
<sphalerite>
Only the guys? In any case, I use alacritty+tmux
<rotaerk>
"guys" is gender neutral in this case :P
<rotaerk>
thanks
<sphalerite>
I'd like to have a setup based on dtach or similar one day, where I have multiple alacrittys (alacritties?) open and use my window manager for the tiling
<sphalerite>
and ideally, the layout gets saved Somehow™ and restored if my graphical session exits
<sphalerite>
but for now tmux is good enough
Heirlung has joined #nixos
<gyroninja>
Is there a way to make a service depend on wireguard being up?
<sphalerite>
gyroninja: yes, but that depends on your definition of "up"
<sphalerite>
you can make it depend on the interface being configured, but not on it being connected
<sphalerite>
since wireguard is stateless
mexisme has joined #nixos
<gyroninja>
Well right now the bind is failing with "Cannot assign requested address"
<gyroninja>
and then right after it there are logs about wireguard being loaded
<gyroninja>
right now my server has a 0/2 success rate in starting correctly due to that
wikiemol has joined #nixos
<gyroninja>
luckily the machine does not need to restart often, but it still would be nice to not have to manually restart the service
idf has joined #nixos
<wikiemol>
Hello, I am going through the Nix Pills tutorial here https://nixos.org/nixos/nix-pills/enter-environment.html, and I tried to follow the step of running `nix-env -i man` but I get the error `error: selector 'man' matches no derivations`, what could be going on?
<gyroninja>
adisbladis: Would you happen to know how to just concat that, since after is already defined from the module in nixpkgs?
<gyroninja>
or cons it
<adisbladis>
gyroninja: You can just set that in your configuration.nix :)
<adisbladis>
gyroninja: The NixOS module system merges all the options
<gyroninja>
Oh, it will automatically merge mine and the module?
<adisbladis>
Yes
<gyroninja>
thanks
<raboof>
when I want to start a shell.nix with 'python with pyserial' but also some other mkShell features, how do I do that? https://nixos.org/nixpkgs/manual/#python shows how to create a shell.nix with *just* python, but my nix-fu isn't strong enough to combine that with the other styles ;)
<raboof>
(specifically I also want some additional dependencies to be installed and add a shellHook that patchelf's some binaries)
<raboof>
adisbladis: wow so simple :D - I'll PR that to the docs I guess :D
wikiemol has quit [Remote host closed the connection]
<eyJhb>
ANy way to repair/supress `collision between `/nix/store/w16n1hxldjwv501vvlgd0i9q9ljlqkgz-gnupg-2.0.30/share/gnupg/help.cs.txt' and `/nix/store/2055v2i8i2q2i3yij8gy5ga1bx9mw2gj-gnupg-2.2.17/share/gnupg/help.cs.txt'`?
<sphalerite>
eyJhb: remove the old one if possible, otherwise give one of the two higher priority
knupfer has quit [Ping timeout: 250 seconds]
mexisme has joined #nixos
<eyJhb>
sphalerite: now I just need to figure out which module causes this - ``/nix/store/7x486sv5zmvpn2mqqi3jlxv5cp7km9pm-python3-3.7.5-env/include/python3.7/structmember.h' and `/nix/store/0sjparzwnm0d0cgxqyd5ylzp56hhqrjc-python3-3.7.5-env/include/python3.7/structmember.h'` :(
<eyJhb>
(gnu is fixed)
<{^_^}>
[nixpkgs] @bjornfor merged pull request #72981 → Fix allowUnfreePredicate example following the introduction of pname. → https://git.io/Jea5c
<{^_^}>
[nixpkgs] @bjornfor pushed commit from @Zimmi48 to master « doc: fix allowUnfreePredicate example following the introduction of pname »: https://git.io/JeVgh
<eyJhb>
Not sure if it is Python, or Python package
<kuznero[m]>
Hi All!
knupfer has joined #nixos
<kuznero[m]>
Is there any way in the context of a single derivation to change current users path? Is it a good idea?
<{^_^}>
[nixpkgs] @phi-gamma opened pull request #73098 → ncmpc: optionally use pcre → https://git.io/JeV2m
<{^_^}>
[nixpkgs] @vcunat pushed 2 commits to staging-19.09: https://git.io/JeV23
<eyJhb>
Precise same md5sum of them both...
jgt has joined #nixos
zupo has joined #nixos
<{^_^}>
[nixpkgs] @vcunat pushed commit from @JohnAZoidberg to staging-19.03 « libtiff: Patch CVE-2019-7663 (PR #72092) »: https://git.io/JeV2n
<eyJhb>
Found the issue sphalerite. Is there any good way, to have two separate .nix file, which both does python.withPackages... ? So that the two merge together
lorimer has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @nixos-channel-bot pushed 2 commits to nixos-unstable-small: https://git.io/JeV2c
jgt has quit [Ping timeout: 250 seconds]
lorimer has joined #nixos
wikiemol has quit [Remote host closed the connection]
<raboof>
adisbladis: aah now I get it - I tried this before but put the env directly into the `buildInputs`, which would have worked if I had put parentheses in the right place :D
jgt1 has joined #nixos
jgt has quit [Ping timeout: 252 seconds]
MichaelRaskin has joined #nixos
idf has quit [Ping timeout: 265 seconds]
<{^_^}>
[nixpkgs] @vcunat pushed commit from @raboof to master « documentation: drop double 'the' »: https://git.io/JeV2i
<colemickens>
Does someone have a parameterized systemd example they like?
<colemickens>
I have an rclone mount driven with a systemd unit, but now I want to have multiple.
<colemickens>
I could do this myself by looping with Nix, but maybe instead I should use systemd's method for making a unit reusable multiple times instead? I'm guessing someone might have an example nixos module to point me at for inspiration.
vidbina_ has joined #nixos
mexisme has joined #nixos
glasserc has quit [Ping timeout: 240 seconds]
mexisme has quit [Ping timeout: 246 seconds]
mexisme has joined #nixos
UndefinedIsNotAF has quit [Ping timeout: 240 seconds]
genesis has joined #nixos
mexisme has quit [Ping timeout: 250 seconds]
jedai42 has joined #nixos
<notgne2>
I can't seem to get Vulkan working on my R7 260x
<notgne2>
it works fine with the proprietary NVidia drivers, and I don't have any other AMD cards to test on (besides another R7 260x variation, which also doesn't work of course)
bvdw has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @contrun opened pull request #73107 → dnsproxy: init at 0.20.0 → https://git.io/JeVad
bvdw has joined #nixos
<adisbladis>
Also pushed the same commts cherry-picked on master to my nixpkgs fork in the branch mesa-19_2_3-master
<adisbladis>
Warning: You need to build _a lot_ of packages to try this out.
<notgne2>
adisbladis: I can give that a try, is there a proper/simple way of applying this? currently I'm on 19.09 but can switch to unstable if required
GiGa has joined #nixos
GiGa has joined #nixos
GiGa has quit [Changing host]
<notgne2>
I assume setting the fork as my nix channel probably a bad idea as I'll lose the sortof "protection" the nix channels offer of not updating until builds pass iirc
<GiGa>
Howdy folks.
<adisbladis>
Unless you need it short-term I suggest waiting
knupfer has joined #nixos
<GiGa>
I've been having problems in the unstable branch for some time - don't ever seem to be able to do an update without some dependencies failing to build.
<GiGa>
Any tips?
<GiGa>
I can grab log output
<adisbladis>
notgne2: Even on a massive build server this takes hours to build
<adisbladis>
And my system closure is not _that_ big
<adisbladis>
A full nox-review took a bit more than a full night
<adisbladis>
GiGa: Please pastebin your log output
<adisbladis>
notgne2: Btw, how are you testing vulkan? I can try the same thing on my machine.
mexisme has quit [Ping timeout: 276 seconds]
<GiGa>
adisbladis: one sec
<notgne2>
vulkaninfo prints `/build/source/vulkaninfo/vulkaninfo.c:5884: failed with VK_ERROR_INITIALIZATION_FAILED`, vkcube SIGSEGVs, and DX games on Wine complain about missing GPU/drivers
chloekek has quit [Ping timeout: 265 seconds]
<notgne2>
mass rebuilds probably not the best idea for me atm, I have 5 pretty capable machines online + my PC itself, but my HDD and internet are really slow so large builds still can be a nightmare
<adisbladis>
notgne2: At least vulkaninfo works on my gpu (I'm on mesa 19.2.3, RX 5700XT)
<adisbladis>
notgne2: Have you enabled firmware?
<{^_^}>
[nixpkgs] @baldo opened pull request #73108 → steam: Add libva to chrootenv → https://git.io/JeVaA
<GiGa>
Miyu-saki: Was wondering if that was "software defined environment" or something :)
<Miyu-saki>
:P
mexisme has quit [Ping timeout: 246 seconds]
<GiGa>
adisbladis: Interestingly I've just run "nixos-rebuild boot" and it has built OK. I made a few tweaks to the config, just looking at the differences.
<GiGa>
Stopped installing packages gnome3.five-or-more, jetbrains.idea-community, smartmontools and set gnome-documents.enable = false
<GiGa>
I wonder if the system was having problems building ideacommunity and Android Studio, as they're both practically the same.
mexisme has joined #nixos
jgt1 has quit [Ping timeout: 245 seconds]
<GiGa>
Going to reboot and see how this update really went.
mexisme has quit [Ping timeout: 250 seconds]
<GiGa>
Oh, I took out XFCE too, as I prefer Gnome 3
<asymmetric>
is there a way to override a module's attribute? e.g. modifying the ExecStart ?
mexisme has joined #nixos
chloekek has joined #nixos
mananamenos has joined #nixos
<{^_^}>
[nixpkgs] @nixos-channel-bot pushed commit from @risicle to nixpkgs-19.03-darwin « keepalived: patch CVE-2018-19115 (PR #72278) »: https://git.io/JeV2b
<kumikumi>
I'm confused by this whole "Using nix in c/c++ development" thing. If I use nix-shell to manage my dependencies, how do I get my IDE to recognise them and stop complaining about my imports?
<asymmetric>
ar: the problem is that the attribute is then defined twice, once by SERVICE_NAME.enable and once by that line, i think
mexisme has quit [Ping timeout: 252 seconds]
ddellacosta has joined #nixos
<Scriptkiddi>
Hey anybody an idea on fixin this issue "error: getting attributes of path '/nix/store/ajjdzjn42p0mr5b0m6fbx60mskxc70cv-unpack-channel.sh': No such file or directory", running nix-store --verify --repair produces "path '/nix/store/ajjdzjn42p0mr5b0m6fbx60mskxc70cv-unpack-channel.sh' disappeared, but it still has valid referrers!
<{^_^}>
[nixpkgs] @bjornfor pushed commit from @lschuermann to master « tpm2-tools: 4.0 -> 4.0.1 »: https://git.io/JeVVi
<{^_^}>
[nixpkgs] @bjornfor merged pull request #72318 → fmsynth: change build so it produces sound again → https://git.io/Jeznw
<{^_^}>
[nixpkgs] @bjornfor pushed commit from @magnetophon to master « fmsynth: change build so it produces sound again »: https://git.io/JeVVX
ddellacosta has quit [Ping timeout: 240 seconds]
GiGa has joined #nixos
GiGa has joined #nixos
GiGa has quit [Changing host]
knupfer has quit [Ping timeout: 245 seconds]
<GiGa>
Everything seems to be working so far :)
mexisme has quit [Ping timeout: 276 seconds]
asymmetric has quit [Ping timeout: 240 seconds]
<notgne2>
I seem to have met with a problem, I configured a remote buildmachine, but now rebuilds crash with an `FD_SETSIZE` related error
mexisme has joined #nixos
<notgne2>
removing the buildmachine would require a rebuild
<notgne2>
I thought I was being clever trying to rebuild with `--build-host` set to local, but that errors as the primary shell is fish, so it errors with `fish: Unsupported use of '='. To run 'nix-store' ...etc...`
<{^_^}>
[nixpkgs] @lejonet opened pull request #73109 → nixos/ceph: Clarify comments in tests and add source reference for the bootstraping method → https://git.io/JeVw3
ris has joined #nixos
mexisme has joined #nixos
<notgne2>
DigitalKiwi: looks like that worked, thanks
<DigitalKiwi>
Nice
shabius has quit [Ping timeout: 276 seconds]
chloekek has quit [Ping timeout: 240 seconds]
Soo_Slow has joined #nixos
__monty__ has joined #nixos
mexisme has quit [Ping timeout: 245 seconds]
<{^_^}>
[nixpkgs] @joachifm pushed 2 commits to release-19.09: https://git.io/JeVwW
<chris__>
Oh right! Great thats all good then. Thanks srhb, tilpner. When you said "depends on how you got that nixpkgs" - anything worth mentioning here?
mexisme has quit [Ping timeout: 252 seconds]
<chris__>
My aim is to just know that I could build a nix expression with that commit and have it "work" as it does now, in the "future".
<tilpner>
You may want to pin your nixpkgs manually, instead of using channels
<srhb>
chris__: say you just had a flat tarball unpacked somewhere. Nixos would have no way to know what commit was responsible for that.
mexisme has joined #nixos
<tilpner>
That way, you can use git to manage the nixpkgs version, and not rely on having the old generations still present
<srhb>
And yes, pinning is great. :)
mibbliosa has joined #nixos
<chris__>
Makes sense, that should be fine then. And yes, I am looking to pin it - just needed to know which commit to reference.
<mibbliosa>
Hi. I have a shell.nix for haskell https://pastebin.com/6JVRbMTg But I cannot get cabal new-repl to work because it hangs up on "<command line>: can't load .so/.DLL for: libz.so (libz.so: cannot open shared object file: No such file or directory)"
jgt1 has joined #nixos
mexisme has quit [Ping timeout: 245 seconds]
<mibbliosa>
Any clues on how to go about fixing this would be appreciated
<chris__>
mibbliosa: in your nix shell what is the output of `ldd $(which cabal)`?
aveltras has joined #nixos
<mibbliosa>
I see a `libz.so.1 => /nix/store/3bfzi6w3067hbdyl15vmypjn4q6az9j6-zlib-1.2.11/lib/libz.so.1`
mexisme has joined #nixos
Kritnich has quit [Quit: Bye bye.]
Kritnich has joined #nixos
<tilpner>
monotux: Hey, you can replace your entire systemd.timers dance with systemd.services.<name>.startAt
<mibbliosa>
The error used to come up when building digest. But that is going after I added zlib to build inputs. But package called `JuicyPixels-3.3.4..` is giving me trouble
<mibbliosa>
chris__: `--pure` didn't seem to fix it either. I think it has something to do with my hakyll dependency. Since stack has atleast tried to solve thise issue, I'll setup stack and give another shot.
<mibbliosa>
thanks for your help chris__
<{^_^}>
[nixpkgs] @aanderse closed pull request #72626 → nixos/httpd: rewrite module to support overridable virtual hosts → https://git.io/Jeg1x
<chris__>
mibbliosa: no worries. good luck!
lunik1213 has joined #nixos
<ivegotasthma>
how do I develop with golang on nixos?
domogled has joined #nixos
szicari has quit [Ping timeout: 268 seconds]
Unode has joined #nixos
mexisme has quit [Ping timeout: 250 seconds]
mexisme has joined #nixos
Kritnich has quit [Quit: Bye bye.]
Kritnich has joined #nixos
GiGa has quit [Quit: Leaving]
mibbliosa has quit [Remote host closed the connection]
iyzsong has quit [Read error: Connection reset by peer]
jmeredith has quit [Quit: Connection closed for inactivity]
<{^_^}>
[nixpkgs] @aanderse opened pull request #73113 → nixos/httpd: support overridable virtual hosts → https://git.io/JeVwj
wikiemol has joined #nixos
<wikiemol>
Hello, I am going through the Nix Pills tutorial here https://nixos.org/nixos/nix-pills/enter-environment.html, and I tried to follow the step of running `nix-env -i man` but I get the error `error: selector 'man' matches no derivations`, what could be going on?
<wikiemol>
Does man not exist in the default channel anymore?
<joko>
Hey, is there a way to persist (as in write in configuration.nix) an instantiated systemd service? So, there is a myservice@.service file and I can do "systemctl enable myservice@myarg.service", but I would like to persist also
<chris__>
,locate bin/man
<{^_^}>
Found in packages: man_db, mandoc, busybox, man-old, plan9port
<chris__>
wikiemol: try one of the above :)
<wikiemol>
oh! Cool thanks! Is there a command line equivalent of that locate command?
<chris__>
joko: what do you mean by instantiated systemd service?
<chris__>
Is there a way I can reference an executable path from nixpkgs package? Currently I have BSD netcat installed as well as GNU netcat. How can I reference either of them from the command line (currently it seems to default to BSD netcat)?
blahaj[m] has joined #nixos
<chris__>
It seems I could use `nix repl` for this - but I'm not sure how to use nix repl un-interactively?
<infinisil>
chris__: netcat has the binary "nc", whereas netcat-gnu has both nc and netcat
<infinisil>
So you should be able to do `nix run nixpkgs.netcat -c nc`
<chris__>
infinisil: Ahh great that works, thanks!
andymandias has quit [Ping timeout: 246 seconds]
andymandias has joined #nixos
m0rphism has quit [Ping timeout: 268 seconds]
mexisme has quit [Ping timeout: 276 seconds]
mexisme has joined #nixos
m0rphism has joined #nixos
<red[m]>
This is perhaps a silly question - but can hydra be used as a build system to build software for "normal"™ Linux distributions, windows et al?
<red[m]>
ie - could it replace circleci or other build systems like that?
<chris__>
red[m]: There is also https://hercules-ci.com/ though I'm not sure if it's related to hydra
<skrzyp>
Anyway... I'm considering doing a clean NixOS install as Dom0-esque KVM hypervisor on my desktop, with several VMs having PCIe resources passed through. Has anyone tried that before?
<infinisil>
asymmetric: Oh you could use ExecStartPre
<infinisil>
asymmetric: Or easier with systemd.services.foo.preStart
<asymmetric>
infinisil: i actually need to run whatever the ExecStart does in a different network namespace
<infinisil>
Not exactly the same, but might be good enough
drakonis has quit [Ping timeout: 276 seconds]
<asymmetric>
so i need to patch/override the ExecStart pre
<asymmetric>
*without the pre
<skrzyp>
(I've got the PCIe passthrough part already running on this machine, but I'm more interested in NixOS "fitting" in that case)
<clever>
dongcarl: and impure.nix defines a function
mexisme has joined #nixos
<clever>
which accepts localSystem, system, platform, crossSystem, config, overlays, and crossOverlays
<dongcarl>
okay so the `import` statement works with both paths and attrs, and if it's a path, it'll transform it into an attr by reading <path>/default.nix?
<clever>
dongcarl: import just parses whatever was in the given file, and returns that
<blaggacao>
I'm planning to switch to nixos to manage all my company's local machines and servers. Anyone has experience with it?
<blaggacao>
- I can't really map out the implications of LDAP user Management vs git-bound declarative user management...
<blaggacao>
- How would I manage ssh keys? Also for example to access a company cache server...
<blaggacao>
How is nix support for Wayland / sway?
<blaggacao>
As of to day, is it easy - or at least - feasible to manage a Raspberry Pi 3+ thin client farm (Call Center)?
<clever>
blaggacao: 3 main ways you could manage a farm of those
<clever>
blaggacao: 1: just re-image the sd card when changes happen, nix can build the whole .img file
<clever>
blaggacao: 2: use nixops to deploy changes to each pi over the network
<clever>
blaggacao: 3: network boot them all from a single image!
mexisme has joined #nixos
<clever>
then you can just swap the image file out, hit reboot, and your updated
chiefgoat has quit [Quit: ZNC 1.7.5-rc1 - https://znc.in]
chiefgoat has joined #nixos
<blaggacao>
clever: network boot is kind of slow and might churn the network. I like the nixops approach!
<clever>
blaggacao: the network boot i was thinking of, downloads the entire image on bootup, and then runs from ram
<clever>
so it only churns on startup or reboot
<clever>
you could also have a variant, that will re-flash its own SD card, over the network, on bootup, so the card acts like a cache
<blaggacao>
Yep I've played with it. It takes some extra seconds to boot and - I couldn't test - but might be problematic if 200 people all startup at once...
<clever>
ive also had some issues with the boot rom netboot code, i suspect it doesnt check checksums
<lordcirth>
If you do go with network boot, I would experiment with trimming the image
<lordcirth>
Every package removal could add up at scale
<blaggacao>
Reflashing over the network is an interesting idea :-)
<clever>
blaggacao: oh yeah, and there is not-os, ive gotten the image down to ~40mb
<kumikumi>
18:02 < blaggacao> Ok, one thing is true: network boot might erase the SD degrading problem...
<kumikumi>
whoopsie
<kumikumi>
(mouse copy paste accident)
<clever>
blaggacao: with some models of rpi, the boot rom cant fully network boot on its own, but if you make an SD card with ONLY bootcode.bin and nothing else, that can patch the boot rom bugs, and the card never accepts any writes
<clever>
the rpi4 also added an SPI flash chip, to solve that problem for good
<dongcarl>
So after reading the nix-1p... I have more questions: I understand now that in https://github.com/fort-nix/nix-bitcoin/blob/master/shell.nix#L4, we brought all of `((import <nixpkgs>) {})` into scope... So the usage of `stdenv` and `pkgs` later in this file is because we did this?
<clever>
so you can now update the boot code, without having to keep an SD card around
<blaggacao>
I don't have the exact reference, but I still got 20 Pi3 in stock which do network boot as they should setting the flag on the SOC
<clever>
blaggacao: this will build whatever is in .rpi_image of release.nix, and then set rpi3-netboot to point to the result
<clever>
blaggacao: it basically behaves the same as `nixos-rebuild switch`
<clever>
blaggacao: i then have a symlink in my tftp root dir, that points to rpi3-netboot
<ilya-fedin>
Is it possible to make package builded with crystal.buildCrystalPackage don't depend on crystal? It is just a compiler and doesn't need after packgae compilation
<clever>
so when an rpi netboots, it grabs from the last build
<clever>
blaggacao: now, the real magic, is that `nix-env -p /nix/var/nix/profiles/per-user/root/rpi3-netboot --rollback` can do a rollback!
<clever>
blaggacao: so if i brick the netboot image, its a single command to instantly un-brick it
<clever>
ilya-fedin: use `nix-store -qR ./result` and `nix why-depends ./result /nix/store/something` to figure out what it depends on, and why
<blaggacao>
Oh, nice... I don't fully understand yet. But that's queued for an in-depth look.
mexisme has quit [Ping timeout: 276 seconds]
<blaggacao>
clever: Do you serve just one image or can you serve on a per Mac basis?
<clever>
blaggacao: behind the scenes, nixos-rebuild is just a shell script that runs `nix-env -p ... foo.nix -A bar --set`
<clever>
blaggacao: the boot rom in the rpi, will search for a directory named after its mac addr
<clever>
and i have symlinks for some macs, pointing to the rpi3-netboot profile
<clever>
so i could just make a 2nd profile, and point other macs to that other profile
LysergicDreams has quit [Ping timeout: 240 seconds]
<clever>
or, the nix expresion in the profile, could generate a directory for every mac in the setup, and manage them all as one unit
<blaggacao>
Top
<blaggacao>
Before getting deeper, I have to cache up on this in a little exercise later
<clever>
another neat thing, is that if the rpi boot rom tries to netboot, it will query the tftp server for all the files
<clever>
that reveals, what files its searching for
<clever>
and some of those files also work on the sd card, and arent documented
<clever>
so you can abuse the tftp boot, as a way to see what the bootrom is thinking, and find hidden features
<blaggacao>
:-)
LysergicDreams has joined #nixos
drakonis_ has quit [Read error: Connection reset by peer]
<blaggacao>
That would be revers engineering broadcom's SoC? :-
<clever>
[root@router:/tftproot]# ls -ltrh 9080d9b6/root.squashfs -L
<clever>
-r--r--r-- 4 root root 41M Jan 1 1970 9080d9b6/root.squashfs
<clever>
blaggacao: partially, your only getting a trace of what filenames it tries to read
drakonis_ has joined #nixos
<clever>
blaggacao: i also have a PR open to add the entire VC4 toolchain (the gpu side of things) to nixpkgs, which can disassembly the gpu firmware
<clever>
the rules say you cant share the boot rom, but nothing stops you from reading it, so you could always dump your own boot rom, and dis-assemble it
<blaggacao>
clever: VC4 for RPi3? Or 4 only?
<clever>
VC4 is the GPU side for all modems
<clever>
models*
<blaggacao>
Ah but there was some driver development by anholt for Wayland support
mexisme has joined #nixos
<clever>
wayland could be done entirely on the arm side, using the dispmanx api
<clever>
VC4 is more if you want to replace the gpu firmware entire with opensource
<blaggacao>
Have you seen the broadcom's effort done by Anholt?
<clever>
nope
<clever>
when you first power on, the pi, it begins running the boot rom on the VC4 cpu
<clever>
ilya-fedin: somewhere inside the binary of that file, is a path to the library, youll want to run strings on the binary to get the full path, crystal (the compiler) may need to be modified to be split output, so the libs are in their own output
noudle has joined #nixos
<clever>
blaggacao: i was once writing v3d drivers for the rpi, they ran entirely from the arm side of things
<clever>
blaggacao: it was mostly just a matter of setting a config.txt flag to disable start.elf from touching v3d, and then writing linux drivers to poke the right registers
mexisme has quit [Ping timeout: 245 seconds]
<clever>
blaggacao: vc4 is a whole new beast, that ive only just recently started to play with
<blaggacao>
That goes well beyond my understanding. Urgs.
<clever>
blaggacao: basically, v3d is just a 3d rendering core, its highly specialized at turning textures and vertexes into 3d renderings
<blaggacao>
Just to confirm: currently sway/Wayland wouldn't work on a nixos Pi3 using standard packages?
<clever>
blaggacao: but vc4 is a dual-core cpu, that is heavily specialized into vector operations
<clever>
blaggacao: ive not looked into the wayland stuff, so i dont know that
<blaggacao>
I can tell:
<blaggacao>
- I made it run on raspbian
<blaggacao>
- I failed on arch linux
<blaggacao>
Not sure why? and why?
acarrico has quit [Ping timeout: 245 seconds]
<ilya-fedin>
clever: As far as I can see in strings, in the binary, path to the source files of the crystal libraries. It seems to be something like debug symbols. Can I somehow make nix ignore this?
<clever>
ilya-fedin: you could try running strip on the binary, in a postBuild hook
<clever>
blaggacao: there is a hello world example in the comments, which will print "hello world!" out the serial port, without any closed-source firmware (other then the boot rom)
skrzyp has quit [Remote host closed the connection]
skrzyp has joined #nixos
Acou_Bass has joined #nixos
mexisme has joined #nixos
<{^_^}>
[nixpkgs] @marsam opened pull request #73121 → nushell: init at 0.5.0 → https://git.io/JeV6u
Supersonic has quit [Disconnected by services]
Supersonic112 has joined #nixos
Supersonic112 is now known as Supersonic
ng0 has joined #nixos
mexisme has quit [Ping timeout: 252 seconds]
acarrico has joined #nixos
<blaggacao>
On another note: I've been poking around on a possible git layout for my local machines project, I've had a closer look at https://github.com/barrucadu/nixfiles Are there any recommendations?
<clever>
blaggacao: personally, i keep configuration.nix rather bare, only the stuff related to the hardware (like fs uuid's, and bootloader cfg)
<tilpner>
blaggacao: I disagree with the "files" directory in there, but a separation into shared "profiles" and per-device settings has worked well for me
<clever>
blaggacao: and then i do imports = [ ./machine1.nix ];
<clever>
blaggacao: so i dont need to make a host symlink, the imports entry handles that
<tilpner>
blaggacao: Though I have far fewer than 200 machines
<clever>
machine1.nix is per-device, and each has its own entry-point
<{^_^}>
[nixpkgs] @matthiasbeyer closed pull request #59105 → linux: kernel: Add linux-ck for 4.14, 4.19 and 5.0 → https://git.io/fjtvo
<blaggacao>
Probably a trade-off decision: ladp vs some gitops driven workflow for user and profile management. Reason: cloud and email services probably rather talk to the ldap service...
<blaggacao>
But declarative gitops driven is just so much better ;-
Supersonic has quit [Ping timeout: 246 seconds]
Supersonic112 has joined #nixos
adisbladis has quit [Quit: WeeChat 2.4]
mexisme has quit [Ping timeout: 245 seconds]
<blaggacao>
Guys, nix is exciting :-) I just learned about it yesterday and I'm sold. It simplifies my life, as it seems. :-)
<infinisil>
ilya-fedin: Oh are you packaging invidious? I have #67664 for that
<clever>
check the nuke-refs docs (or source) and try to nuke only references to crystal
bastion-tester has joined #nixos
<{^_^}>
[nixpkgs] @Ericson2314 pushed 3 commits to structured-attrs: https://git.io/JeVim
bastion-tester has quit [Client Quit]
<ilya-fedin>
But I have one more question, by the way ... I tried to override crystal to a newer version, but it doesn’t work out for me. Temporarily, I copied the functions that builds crystal, but I want to override the normal way (or at least somehow use the functions from let block without copying)
<{^_^}>
[nixpkgs] @zimbatm pushed commit from @filalex77 to master « epr: init at 2.3.0b (#73038) »: https://git.io/JeVPE
<mojjo>
does import-from-derivation not work with nixos configuration? the following works `{} // { normal config..... }` but this not: `import (pkgs.writeTextFile "my.nix" "{}") // {}` I'm getting 'infinite recursion...'
<clever>
mojjo: it does work, but nixos must know the content of your nixos config, to know what the value of pkgs is
<clever>
mojjo: you need to get writeTextFile from somewhere else, (import <nixpkgs> {}).writeTextFile
<worldofpeace>
hey gchristensen, remember when op me to change the topic. guess matrix messed that up and the topic is cut off at `For best support,`. Can we complete that to, `For best support, please paste your configuration.nix to a pasting service when asking your question` or something shorter?
<lucus16>
clever: pavucontrol is unchanged, the process is dead, presumably everything I start now will work
<clever>
mojjo: also, using // on nxios config will cause more problems, your better off putting my.nix into the imports list, it will merge things a lot better
<lucus16>
clever: I just wish I knew what started it
<clever>
lucus16: to start with, does `ps -eH x` show which process is the parrent of pulseaudio?
<worldofpeace>
clever: that combination just seemed to trigger a lot of bugs, and with systemd activation it doesn't make sense to be there.
<lucus16>
analog is card 0 device 0 and hdmi is card 0 device 3
mexisme has joined #nixos
<clever>
worldofpeace: ive never had any issues with slim + xfce
<clever>
worldofpeace: does the hdmi appear on the output devices tab of pavucontrol?
<lucus16>
assuming you mean me, no.
<clever>
oops
<clever>
that can also happen if something was using the alsa device when pulse started
<worldofpeace>
clever: I believe the issues were mainly in Gnome with gdm, and after some research, distros that patched this away couldn't reproduce the problems. A similar situation happened with flat-volumes, so I just decided it would be best.
<worldofpeace>
clever: yep it does. as long as the device has the correct configuration...
pamplemousse has joined #nixos
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<pamplemousse>
hi there :)
<pamplemousse>
I am giving a go at NUR, but it refuses to work, and I don't get what I am missing...
<lucus16>
clever: Apparently this works: pactl set-card-profile 0 output:hdmi-stereo
<lucus16>
\o/
<clever>
lucus16: ah, you probably wanted to go to the configuration tab of pavucontrol, and select a profile there
<babariviere[m]>
<pamplemousse "I am giving a go at NUR, but it "> hi, can you describe your issue ?
<clever>
lucus16: that changes how many inputs and outputs a card has
mexisme has quit [Ping timeout: 250 seconds]
<clever>
for example, my desktop soundcard has 3 physical inputs, 3 muxes to select any input, for 3 full capture channels
<clever>
(mic, rear-line-in, front-line-in)
<lucus16>
nice
<clever>
but by default, pulseaudio only supports 1 channel of capture
drakonis1 has joined #nixos
<clever>
i had to write my own config for the card, and even patch pulseaudio itself (to support multiple muxes), to make full use of it
<clever>
then i had no use for it, disabled the module, and never did anything with it, lol
<pamplemousse>
babariviere[m]: I have updated my `config.nix` as in https://github.com/nix-community/NUR#installation, but hit a `error: while evaluating the attribute 'buildInputs' of the derivation 'shell'` when I try to `nix-shell -p nur.repos.mic92.inxi`
<babariviere[m]>
pamplemousse how did you install nur ?
<babariviere[m]>
* pamplemousse how did you install nur ?
<rotaerk>
sondr3, if you leave out gcc9, do you still have gcc8 available? what happens if you run `which gcc`
pamplemousse has quit [Ping timeout: 268 seconds]
o1lo01ol1o has quit [Remote host closed the connection]
ng0 has joined #nixos
<sondr3>
rotaerk: yeah, I get the same GCC version if I leave out gcc9: `which gcc: /nix/store/5yyx688q9qxhb4ypawq7v80fm3ix27dm-gcc-wrapper-8.3.0/bin/gcc`
<{^_^}>
[nixpkgs] @1000101 opened pull request #73131 → Port tests to python → https://git.io/JeV1D
<fresheyeball>
anyone using nix-shell of python development?
knupfer has joined #nixos
<samueldr>
no one at all, it's been forbidden by dark forces since the dawn of time
<samueldr>
(sorry, less rude answer) probably, likely yes, but that's likely not the answer you're looking for
<samueldr>
anything in particular you're having questions or trouble about?
<jackdk>
Also there is a section on using Python in the nixpkgs manual; have you read it, and if so, did it help? Also also, http://sol.gfxile.net/dontask.html
<{^_^}>
[nixpkgs] @worldofpeace opened pull request #73138 → Port xfce tests to python → https://git.io/JeVM1
<sshow>
exarkun++
<{^_^}>
exarkun's karma got increased to 6
andymandias_ has joined #nixos
andymandias has quit [Ping timeout: 240 seconds]
work_ has joined #nixos
domogled has joined #nixos
NoctisLabs has joined #nixos
<fresheyeball>
samueldr: lol
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
ozychhi has joined #nixos
<shyim>
Hey, have someone an example of a nixos module in a nur repository?
<{^_^}>
[nixpkgs] @worldofpeace opened pull request #73140 → nixosTests.pantheon: port to python → https://git.io/JeVys
lane-s has quit [Ping timeout: 260 seconds]
philr has joined #nixos
knupfer has quit [Quit: knupfer]
knupfer has joined #nixos
ornx has joined #nixos
numinit_ has quit [Remote host closed the connection]
pamplemousse has quit [Ping timeout: 265 seconds]
knupfer has quit [Client Quit]
ornx is now known as ornxka
knupfer has joined #nixos
<KarelWDingeldey[>
Is anyone running 389-ds on NixOS? I'm having trouble setting it up.
lane-s has joined #nixos
<{^_^}>
[nixpkgs] @globin pushed to structured-attrs « treewide: fix nested lists in *buildInputs »: https://git.io/JeVyz
<lane-s>
So now it seems like I can access `nixos.org`, but now I'm getting other errors when trying to rebuild my config. I get `search path entry ..../root/channels/nixos does not exist`
<lane-s>
When I do `nix-channel --list` the output is blank
omnipotententity has joined #nixos
<omnipotententity>
If I'm trying to detect sqlite using cmake, and I have sqlite in my buildInputs of my shell.nix, and it still doesn't find it, then what stupid thing am I doing wrong?
<babariviere[m]>
and with `sudo nix-channel --list` ?
<babariviere[m]>
lane-s
<babariviere[m]>
<omnipotententity "If I'm trying to detect sqlite u"> do you have pkg-config in your buildInputs ?
<slack1256>
Is there a package for emacs that provides type-info or code-completion for nix files?
abathur has quit [Ping timeout: 265 seconds]
<slack1256>
I am using nix-mode but it only seems to provide syntax highlighting.
<clever>
omnipotententity: is pkgconfig in nativeBuildInputs ?
<fresheyeball>
any advice on adding a python dep with nix via github?
<babariviere[m]>
<slack1256 "Is there a package for emacs tha"> there is only syntax highlighting in my knowledge with `nix-mode`
<sshow>
fresheyeball: builtins.fetchGit
<fresheyeball>
sshow: I have no idea how python deps work
<lane-s>
babariviere[m] Sigh, that lists the channels but then I'm back to the problem I was having before. The same error with the search path occurs, and if I try to do `nix-channel --update nixos` then I get `unable to download. Couldn't resolve host name`
<fresheyeball>
can I just add the src cretaed by fetchGit as a build input?
<omnipotententity>
I now have pkgconfig in my nativeBuildInputs, but it seems to not be finding it still?
<clever>
omnipotententity: how is cmake being told to search for sqlite?
<babariviere[m]>
<lane-s "babariviere Sigh, that lists the"> it seems like you don't have a dns resolver, can you try adding `nameserver 8.8.8.8` in `/etc/resolv.conf` ?
<clever>
omnipotententity: and where if find_package defined?
<omnipotententity>
and I have a file under a directory named cmake called FindSqlite3.cmake that I copy-pasted from the CMake Gitlab repo
<lane-s>
babariviere[m] Ok, did that. Do I need to run any command to apply that change? Also it seems like I have some kind of dns resolver because I can browse to most websites.
<clever>
omnipotententity: what do the docs say find_path and find_library are doing, to actually find things?
GlennS has quit [Quit: No Ping reply in 180 seconds.]
<babariviere[m]>
<lane-s "babariviere Ok, did that. Do I n"> it will be applied immediately, I don't what you are using for networking but normally you can configure it in your configuration.nix
GlennS has joined #nixos
<lane-s>
babariviere[m] Using NetworkManager. The only network related thing in my config is currently `services.networkmanager.enable = true;`. The issue is that I can't update my config without a working connection to `nixos.org` in this case right?
<omnipotententity>
the error message is just a link failure during compile. If you inspect the variables CMake is generating, SQLite3_INCLUDE_DIR and SQLite3_LIBRARY are both set to FOO-NOTFOUND where FOO is the variable name
<lane-s>
It's currently just hanging when I try to update the `nixos` channel. Will wait a while and see what happens.
<clever>
omnipotententity: nix itself, would have already put sqlite into the linkers search path
<clever>
omnipotententity: so you dont have to search for sqlite, just blindly -lsqlite3
<babariviere[m]>
<lane-s "nixos-hardware https://github.co"> `sudo nix-channel --update` works ?
<lane-s>
@clever I'd say not very. Now it seems like they are hanging as well
fresheyeball has quit [Ping timeout: 268 seconds]
<babariviere[m]>
are you using WiFi or lan ?
<babariviere[m]>
DHCP or static IP ?
<lane-s>
USB wifi with dhcp. DHCP is the default right?
<babariviere[m]>
you may need to configure a DNS with network manager
<babariviere[m]>
yes it's the default
<Sisyphe[m]>
Hi, is this the right place for noob questions ?
<babariviere[m]>
USB with your phone ?
<babariviere[m]>
<Sisyphe[m] "Hi, is this the right place for "> hi, yes you can ask your question :)
<lane-s>
Sisyphe[m] If it's not, it's not stopping me from asking them.
<lane-s>
It's a USB wifi dongle
<leo_>
Are there any disadvantages to dynamic imports?
<lane-s>
I don't have a driver for the laptop's built-in wifi yet
<leo_>
i.e builtin.import?
<Sisyphe[m]>
ok thanks, I was trying to replace mkShell with buildFHSUserEnv since nix-shell takes a long time (~15s)
<babariviere[m]>
<lane-s "I don't have a driver for the la"> oh, you don't have any connection issue with your browser? I am sorry for those dumb questions, I am just trying to understand where the issue can be
<Sisyphe[m]>
but every python package I use is missing some dependencies (ImportError: No module named...) while no dependencies were missing when using mkShell and nix-shell --pure
<Sisyphe[m]>
so does someone know how to properly include python packages with buildFHSUserEnv ?
<lane-s>
No worries I'm glad to have your help. It's weird, sometimes the browser works super fast, and then sometimes it looks like it's also having DNS issues and the page never loads.
<babariviere[m]>
<Sisyphe[m] "so does someone know how to prop"> can you show your $PYTHONPATH in your buildFHSUserEnv ?
<babariviere[m]>
<lane-s "No worries I'm glad to have your"> do you know your DNS provider? it can be provided by your internet box
<clever>
babariviere[m]: previously, the resolv.conf has 192.168.1.1
<clever>
had*
<Sisyphe[m]>
<babariviere[m] "can you show your $PYTHONPATH in"> I've manually set it to "/usr/lib/python2.7/site-packages/" since buildFHSUserEnv doesn't seem to update it at all
<babariviere[m]>
ah yes, so it's the internet box, DNS queries are resolved by your box
prusnak has quit [Quit: leaving]
<babariviere[m]>
<Sisyphe[m] "I've manually set it to "/usr/li"> aren't you using python 3 ?
<lane-s>
It's all working fine on my other machine connected to the same router though
<babariviere[m]>
did you check that your packages are in this directory ?
<Sisyphe[m]>
<babariviere[m] "aren't you using python 3 ?"> nop python 2
<lane-s>
It also seems like my computer is getting weirdly hot for not really doing anything
<Sisyphe[m]>
<babariviere[m] "did you check that your packages"> yes and the packages are missing
<babariviere[m]>
<lane-s "It also seems like my computer i"> if you run htop, what programs are the most resources consuming ? I know there is some issue with docker if you are using it
<omnipotententity>
I also needed cmake as a build input, derp
<babariviere[m]>
<Sisyphe[m] "yes and the packages are missing"> can you show me how are you adding python packages ?
<{^_^}>
[nixpkgs] @flokli closed pull request #73109 → nixos/ceph: Clarify comments in tests and add source reference for the bootstraping method → https://git.io/JeVw3
<{^_^}>
[nixpkgs] @chessai opened pull request #73144 → vim-lsp and async.vim: init → https://git.io/JeVSm
<lane-s>
My config is pretty minimal. Don't actually even have htop. I'm going to boot from usb to try to add support for my laptop's built in wifi, add htop, dnsutils, and dns config
<omnipotententity>
thanks babariviere[m] and clever you two are always helping me out and I appreciate it.
bvdw has quit [Read error: Connection reset by peer]
<babariviere[m]>
no problem !
<Sisyphe[m]>
<babariviere[m] "try using `python2.withPackages`"> with (python27.withPackages (ps: with ps; [ my_python_pkg ])) the executables in my_python_pkg are not found
<babariviere[m]>
<leo_ "Are there any disadvantages to d"> I don't think there is any
bvdw has joined #nixos
<babariviere[m]>
<Sisyphe[m] "with (python27.withPackages (ps:"> and if you remove your manually set PYTHON PATH ?
ng0 has quit [Ping timeout: 260 seconds]
<Sisyphe[m]>
<babariviere[m] "and if you remove your manually "> same
<lane-s>
clever: Is your wifi device showing up with `ip a`? My usb wifi device shows up there, but all this wonky stuff still happens. My built-in wifi adapter doesn't show up at all.
<babariviere[m]>
<lane-s "clever: Is your wifi device show"> do you have any error in `dmesg` ?
<Sisyphe[m]>
What I cannot understand yet is why python packages listed as propagatedBuildInputs in included packages doesn't appear in /usr/lib/...
__monty__ has quit [Quit: leaving]
pie_ has joined #nixos
<babariviere[m]>
lane-s : does your laptop have an Nvidia card ? (was an issue for me)
<babariviere[m]>
and are you using `networking.<interface>.useDHCP` ?
<babariviere[m]>
Sisyphe I may say something wrong but for your buildFHSUserEnv, I don't think propagated build inputs works, you have to use target packages
Yaniel has quit [Ping timeout: 276 seconds]
<Sisyphe[m]>
<babariviere[m] "Sisyphe I may say something wron"> Oh that would explain. But retrieving all propagatedBuildInputs to put them as target inputs seems intractable, at least if I need to it manually.
<clever>
lane-s: ah, thats the exact same pid as the last guy with this issue
<clever>
lane-s: you need the 5.3 kernel for iwlwifi to support it
<clever>
lane-s: boot.kernelPackages = pkgs.linuxPackages_latest; will get you 5.3, if your on unstable
lane-s has quit [Remote host closed the connection]
lane-s has joined #nixos
JonReed has quit [Remote host closed the connection]
<lane-s>
Hmm now I'm unable to update `nixos-unstable`. Getting `unable to download /nixeprs.tar.bz2 URL using bad/illegal format or missing URL` and also a bunch of `couldn't connect to server`
<lane-s>
Oh it's actually happening with the other channels too...
<atlas_>
I'm trying to install nix on macOS Catalina. Catalina moved system-critical directories to a read-only filesystem. How can I install nix as a single-user installation at ~/nix?
<clever>
atlas_: you would need to re-compile every single package from scratch
<clever>
atlas_: there are directions somewhere on how to make / writable again
<atlas_>
clever: really? I can't just specify a new nix location?
<clever>
atlas_: all binaries have /nix/store hard-coded into them
<atlas_>
clever: I can probably fix it, but it will be difficult.
<atlas_>
clever: huh, ok. I'm sure that's done intentionally, right?
<babariviere[m]>
you can patchelf with a classic interpreter path
<Xal>
That seems... hacky?
<babariviere[m]>
that's hacky :p
<clever>
it relies on the ld.so from the target being compatible
<Xal>
Is there a way to get a dynamic linker in the standard location without patching it?
<clever>
thats not much different from using patchelf
<Xal>
honestly that's true
<Xal>
I will probably go with musl
<lane-s>
i think i'll have to give up on nixos for the time being. I will try again on my desktop later. It's just way too difficult with a computer that only has wifi, and the chicken/egg problem of needing wifi to get proper wifi drivers. Also need this computer to work by Monday...
<Sisyphe[m]>
<babariviere[m] "it contains all propagated build"> Nice! I'll try that
<{^_^}>
[nixpkgs] @globin pushed to structured-attrs « treewide: *inputs are lists »: https://git.io/JeV9Y
Xal has quit [Remote host closed the connection]
<clever>
lane-s: you can get an iso image, that is using 5.3 on bootup, and then use that to fix things
<babariviere[m]>
<lane-s "i think i'll have to give up on "> if time is your limit, I think it's your best bet to use another OS, but you can still use nix in your other OS
m0rphism has quit [Ping timeout: 265 seconds]
<niso>
lane-s: do yourselfe a favor (also for other distros) and just get a usb-ethernet dongle (saved me quite a lot of hours so far)
<clever>
2019-11-07 01:13:37< clever> rotaerk: that is the linux 5.3 iso image, for the current build of nixos-unstable
<clever>
lane-s: thats the iso with 5.3, and using kernelPackages = pkgs.linuxPackages_latest; will keep you on 5.3
<lane-s>
I do use nix in mint, I was just hoping to switch to nixOs. I'm just getting worried now that I could spend the rest of the weekend configuring my system as still not be ready for work. Still tempted to try though
oscarvarto has joined #nixos
<babariviere[m]>
maybe try to install it on a VM first
<babariviere[m]>
after that you will be able to find a solution for your network issues
<lane-s>
Yeah once I get my desktop set up I'll just install it there. Should have no network issues with a generic ethernet device and no stress about getting it set up in time for work
<maralorn>
The example from "nix why-depends --help" says "nix why-depends nixpkgs.hello nixpkgs.glibc" that doesn‘t work for me "error: attribute 'nixpkgs' in selection path 'nixpkgs.hello' not found". Why? Where does it take the variable nixpkgs? I have <nixpkgs> in my NIX_PATH but nixpkgs != <nixpkgs>.
<clever>
maralorn: you can also give it store paths, nix why-depends ./result /nix/store/fat-thing