<{^_^}>
[nixpkgs] @disassembler opened pull request #37289 → dnsdist: init at 1.2.1 → https://git.io/vxO6G
<clever>
disasm: heh, we opened PR's within 7 seconds of eachother
KABA has joined #nixos
<disasm>
clever: haha, nice! review swap? :)
<clever>
already skimmed over your description
<clever>
i'll have a deeper look
<KABA>
jtojnar or someone working on gnome3.28, I just tried installing gnome-music on 2fabd17 (latest hydra build) and it fails with searchbar.py:301: Warning: cannot retrieve class for invalid (unclassed) type 'void'
<clever>
disasm: one piece of input i'm looking for, is if i should keep the default at xz, or change the default to pixz
<clever>
the biggest difference is runtime (5m vs 1m20) and output size (118mb vs 129mb)
<KABA>
jtojnar: Oh, I see. Just wanted to help testing as I can't really (due to lack of knowledge/time) help with other things, but would really like to see gnome3.28
<clever>
but i have no idea how deterministic pixz is, and i havent been able to get build-repeat to work on nix
nD5Xjz has joined #nixos
<jtojnar>
KABA: hopefully it will not take long now, we already know the cause of the issue, someone just needs to update the regex pattern
eliaslfox has joined #nixos
<clever>
disasm: only thing i can see as an issue in your PR is the lack of enableParallelBuilding = true;, its only using 1 core when i test the build
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
logzet has quit [Remote host closed the connection]
noqqe has quit [Ping timeout: 264 seconds]
n0qqe is now known as noqqe
astrofog has joined #nixos
ottidmes has quit [Ping timeout: 264 seconds]
<{^_^}>
[nixpkgs] @coreyoconnor opened pull request #37292 → wine: unstable to 3.4 and make "unsafe" GStreamer plugins non-default → https://git.io/vxO13
astrofog has quit [Quit: Quite]
michas__ has quit [Ping timeout: 260 seconds]
variable has quit [Quit: Found 1 in /dev/zero]
<shlevy>
aszlig: Around?
ashgillman has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @shlevy pushed to binutils-2.30 « binutils: Bump default to 2.30 »: https://git.io/vxO1x
mbrgm has quit [Ping timeout: 256 seconds]
mbrgm has joined #nixos
mounty has quit [Ping timeout: 265 seconds]
Supersonic112 has joined #nixos
Supersonic has quit [Disconnected by services]
markus1199 has joined #nixos
Supersonic112 is now known as Supersonic
Fare has joined #nixos
markus1189 has quit [Ping timeout: 246 seconds]
loonquawl has quit [Quit: loonquawl]
Fare has quit [Ping timeout: 240 seconds]
<pjan>
any home-manager users in the room? rycee? I'm currently configuring home-manager by setting the HOME_MANAGER_CONFIG environment variable. Can this be done through the nixPath as well?
jason has joined #nixos
jason is now known as Guest40848
Guest40848 has left #nixos [#nixos]
Guest40848 has joined #nixos
Guest40848 has left #nixos [#nixos]
_null_ has joined #nixos
Fare has joined #nixos
variable has joined #nixos
zie[m] has joined #nixos
katona has quit [Quit: Connection closed for inactivity]
<boomshroom>
I found my brainfuck interpreter in Nix.
<boomshroom>
Is there any way to convert an integer to it's unicode or ascii character?
eliaslfox has quit [Read error: Connection reset by peer]
s2mitrov has joined #nixos
<s2mitrov>
heyo, is there a safe/good way to delete bad, recent nixos configs?
<boomshroom>
s2mitrov: You mean that you know a generation won't work so you want to remove it from your boot menu?
<s2mitrov>
ex. my like 4 or 5 most recent configs give me a 'failed to execute login command' thing, so i'd prefer to just delete them from the boot list so that when i hit default boot it picks my last working one
<s2mitrov>
yeah
robstr has joined #nixos
<boomshroom>
s2mitrov: I don't know if it works for the nixos generations, but `nix-env --delete-generations` can do that for nix-env generations.
<boomshroom>
Maybe `sudo nix-env -p /nix/var/nix/profiles/system --delete-generations ${generations_to_delete}` would work?
jsoo has joined #nixos
jsoo has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #37296 → Fix fallout from #36275 → https://git.io/vxO98
<s2mitrov>
boomshroom: is there some way to reorder generations instead?
robstr has quit [Ping timeout: 260 seconds]
<{^_^}>
[nixpkgs] @coreyoconnor opened pull request #37297 → Add Deja-Vu fonts to jenkins environment and run as AWT headless → https://git.io/vxO92
<s2mitrov>
ex. i've booted into my working generation (from 3/17), made changes and nixos-rebuild switch'd but then when i default boot i get failed to login, i'm assuming because there's some shitty config/state lurking around from a previous config?
<s2mitrov>
but i'm also kinda confused since if i select the 3/17 config again, the changes i made are there
<boomshroom>
s2mitrov: Not that I know of. As far as I know, you can delete arbitrary generations and change the current environment to arbitrary generations. (rollback is just a shorthand to switch to the second most recent generation)
<s2mitrov>
boomshroom: is what i'm describing expected behavior? i would've assumed that doing a rebuild on a previous generation would create an entirely new one based off of that generation
<yrashk>
On that note, I noticed that it's getting harder to get regular PRs (inits and updates) merged in, I have three collecting dust for varying amounts of time (https://github.com/NixOS/nixpkgs/pulls/yrashk). Is there anything can help the core team with on that end?
<{^_^}>
[nixpkgs] @veprbl opened pull request #37298 → pyslurm: 20171102 -> 20170302 to fix after slurm bump → https://git.io/vxO9o
<boomshroom>
s2mitrov: If you rebuild, nothing in the current or other generations should matter. Only what's currently in the configuration.nix and the files it imports.
<s2mitrov>
thats what I thought, but it doesn't seem to be behaving that way
<boomshroom>
I should mention that neither a rebuild nor a switch should impact the configuration.nix. Maybe you didn't quite remove what's causing it to fail?
<s2mitrov>
the whole thing started when i added lightdm and feh to set my background on startup; it worked, but for whatever reason i couldn't launch a terminal after logging in, so i got rid of lightdm from my config, rebuilt and then i was getting 'failed to excecute login command' or whatever, so i just removed all the feh stuff from the config and that worked in that it's no longer in the config, but i can't use any of my recent configs to
<s2mitrov>
login
<s2mitrov>
i have to go back to the 3/17 config which for some reason has the changes i've made (which I would expect to exist only in the 3/18 configs)
<coconnor>
No timezone weirdness between those two dates?
<s2mitrov>
shouldn't be, i don't think
<s2mitrov>
i didn't play with it, at least
<s2mitrov>
https://i.imgur.com/pbrseZG.png so the bottom config there is the good one; i can boot into it, make changes, nixos-rebuild switch and everything is 'fine'
<s2mitrov>
but when i reboot and select default bot (or choose any of the more recent configs) it fails
<s2mitrov>
so i go back to the 3/17 one at the bottom there and it works and for whatever reason also has the changes to config i've made previously
<s2mitrov>
oh
<s2mitrov>
is messing with user groups stateful or something?
<coconnor>
there is a part of switching NixOS configs that is stateful.
<coconnor>
essentially modifies some stuff in /etc that otherwise be moved. And misc stuff like changing hte MBR
<boomshroom>
When I specify multiple sources to use, how can I tell where they get extracted to in relation to eachother?
<coconnor>
there are two modes for user/groups: if "mutableUsers" is enabled or not
<s2mitrov>
the only other change i made was reordering stuff under my user (which I don't think could matter?) and changing the extraGroups, where it was previously `["wheel" "video" "audio" "disk"]` and i changed it to just `[ "wheel" ]`
<s2mitrov>
mutableUsers is enabled by default, no?
<coconnor>
I don't recall... checking
<coconnor>
yes
<coconnor>
huh... It's what I use but I expected the default to be declarative users by now haha
<coconnor>
what does the "disk" user grant?
<s2mitrov>
is there a more strict nix-collect-garbage i can run? i'm thinking lighdm maybe left some crap littered around that might be messing with boot
_null_ has quit [Remote host closed the connection]
<s2mitrov>
oh
<s2mitrov>
hmm...
<s2mitrov>
running nix-collect-garbage on the working config cleaned up nothing, but i just ran it on the 'broken' config where i'm logged in as root and it cleaned up 600 megs of crap
<s2mitrov>
buncha lightdm stuff in the output too, hopefully this fixes it
orivej has quit [Ping timeout: 260 seconds]
<s2mitrov>
alright nvm lmao
rogue_koder has quit [Quit: Konversation terminated!]
<s2mitrov>
it just deleted all my profiles, rip
ok2`` has quit [Ping timeout: 260 seconds]
newhoggy has joined #nixos
<s2mitrov>
:"^) so the issue was in my reordering of stuff for my user, i managed to accidentally delete `isNormalUser = true;`
Guanin has quit [Ping timeout: 264 seconds]
<hyper_ch>
hmmm, something is blocking nixos-unstable and nixos-unstable-small
<s2mitrov>
anyone using feh to set their wallpaper on startup?
eliaslfox has joined #nixos
_null_ has joined #nixos
<_null_>
has anyone had problems with using spectacle (kde screenshot tool). I think kde can't find it? there's a default shortcut that's there it just sorta doesn't happen. I've tried putting spectacle in services.dbus.packages and it doesn't seem to help (do I need to restart my display manager?)
<boomshroom>
s2mitrov: I was. I think I stopped after switching to lightdm
<boomshroom>
Let's see if my jiri derivation builds.
<boomshroom>
Let's see if my jiri derivation builds.
<boomshroom>
Whoops, thought this was my build terminal.
Guanin has joined #nixos
<etu[m]>
hyper_ch: Yeah... The NixOS unstable age is starting to get annoying...
<boomshroom>
I'm making a non-standard derivation. I don't want to bother with all of the phases, so I'm trying to set the builder manually, however toFile doesn't let the output refer to derivation outputs. How can I get around this?
<boomshroom>
More specifically, I have a list of git locations and a desired path for each. I want to copy the git repositories to their coresponding path and make that the input to the actual derivation.
newhoggy has quit [Remote host closed the connection]
<s2mitrov>
it seems like installing rust using mozillas overlay does/should include rust-src, and i'm pretty sure i saw it in the output while nix was copying things over, but after installing i can't actually find where it is
<_null_>
where would you expect rust-src to be?
<_null_>
is it possible it's getting deleted after the build?
<{^_^}>
→ 3315a549 by @yrashk: xml2rfc: init at 2.9.6
<{^_^}>
→ 71574041 by @FRidh: Merge pull request #37304 from yrashk/xml2rfc
<s2mitrov>
garbas: heyo, you around? :P i'm trying to use your rust-overlay but i'm having trouble finding where rust-src is so that i can set $RUST_SRC_PATH for racer
<tilpner>
You could set RUST_SRC_PATH = latest.rustChannels.stable.rust-src in your shell.nix
<MichaelRaskin>
boomshroom: you want runCommand, I think
troydm1 has joined #nixos
<s2mitrov>
tilpner:how does shell.nix work? i just make a shell.nix in /etc/nixos and any configs in there would apply globally to nix-shells?
ma27 has joined #nixos
<tilpner>
No, you would put that into every shell.nix :/
<s2mitrov>
:x that uh, sounds tedious to say the least
<tilpner>
Or you can set it globally via environment.variables
<tilpner>
But that causes problems when you work with different Rust channels
<s2mitrov>
i guess i could write a little script to piggy back on cargo new and spit out a shell.nix as well
alhariel has quit [Ping timeout: 260 seconds]
<tilpner>
That would be pretty much useless
<tilpner>
cargo has no information over what Nix dependencies to add, so all it could fill in is the crate name
<s2mitrov>
i'm kinda confused now haha
<s2mitrov>
am i supposed to ignore cargo entirely for rust dev on nix?
<tilpner>
No
<tilpner>
(You can even use rustup now IIRC)
<s2mitrov>
i tried using rustup initially but without success, which is why i'm using this overlay :(
<s2mitrov>
so every time you make a new project you have to also write a shell.nix for it?
<tilpner>
I don't currently create many new Rust projects. I use "nix-shell -p cargo rustc" for the ones I still have around
<_null_>
the pill only mentions definiing it ourselves, and some nixpkgs reference it but I can't find from where. Using it as my input fails and it's not in stdenv
<tilpner>
Yes, it exists. pkgs.callPackage
<_null_>
okay, hm I was screwing up something else and it was giving me "attribute ‘callPackage’ missing," as the error
<_null_>
but I fixed that much, it seems I need to do something about the hash?
<_null_>
now I just get "opening file ‘/nix/store/[the-nix-path]/default.nix" no such file or directory
<clever>
_null_: how did you run callPackage?
<_null_>
I'm trying to use it on a package I'm downloading, so I'm calling it on pkgs.the-pkg
<clever>
thats already been callPackage'd
<_null_>
uhhhhh, okay so is there a way I can supply different arguments to it then?
<clever>
.override
<clever>
callPackage adds an .override function to everything it loads
<_null_>
okay so I use "the-pkg.override { my-arg = true }" ?
<clever>
yep
orivej has quit [Ping timeout: 268 seconds]
<_null_>
awesome, thanks so much for your patience with me
<{^_^}>
[nixpkgs] @rycee pushed commit from @limeytexan to master « perlPackages.PkgConfig: init at 0.19026 »: https://git.io/vxObk
<clefru>
tilpner: nix-build -A python36Packages.pycurl '<nixpkgs>' fails for me with errors in the test suit that runs after the main build
freeman42x]NixOS has quit [Ping timeout: 240 seconds]
<tilpner>
What's the content of /nix/var/nix/profiles/per-user/root/channels/nixos/.git-revision ?
<tilpner>
(Assuming <nixpkgs> refers to that for you)
<clefru>
tilpner: I don't pull from the channel, I have a local git checkout. 557aa6e7d7a6bc03b845098d518725c1d77af5d4 is the latest commit as of a few minutes ago
mounty has joined #nixos
<clefru>
(but my other system that's pulling from the channel fails also at pycurl)
Ariakenom has joined #nixos
<clefru>
looks like https://github.com/NixOS/nixpkgs/pull/37038 would fix the issue, but I am just surprised that the channel release-18.03 moved forward despite the build failure
<{^_^}>
[nixpkgs] @Mic92 pushed to master « python.pkgs.pushover: tests exists, but require network »: https://git.io/vxOxw
<clever>
tilpner: there is also nix-instantiate --eval '<nixpkgs>' -A lib.nixpkgsVersion
<tilpner>
Oh
<clever>
[clever@amd-nixos:~]$ nix-instantiate ~/apps/nixpkgs --eval -A lib.nixpkgsVersion
<clefru>
regardless of the flakyness of the tests, I don't understand why there is a channel release, but not all packages in that release a precompiled on hydra
<tilpner>
That works here, but it's also a different hash. I'm on x86_64-linux
<{^_^}>
→ 33b0ad83 by @teto: qemu-vm: passing QEMU_NET_OPTS would fail
<{^_^}>
→ ac9818fa by @Mic92: Merge pull request #37216 from teto/qemu_comma_fix
<clefru>
huch, how can this be a different hash?
<clefru>
(x86_64-linux as well)
<tilpner>
Differences in sandboxing and nixpkgs configuration
<clefru>
I have no package overrides or similar.
<tilpner>
Other than that, no idea
rumble has joined #nixos
grumble is now known as Guest75212
Guest75212 has quit [Killed (tolkien.freenode.net (Nickname regained by services))]
rumble is now known as grumble
newhoggy has joined #nixos
<jluttine>
has anyone set up jekyll locally for creating sites for github pages? if someone has a working configuration and/or instructions, that'd be great. at the moment, bundix is raising weird errors so i'm quite stuck
<{^_^}>
→ 34682ddc by @ryantm: exempi: 2.4.4 -> 2.4.5
<{^_^}>
→ 9bb679aa by @gebner: Merge pull request #37286 from ryantm/auto-update/exempi
<clefru>
clever: if that's a question for me, I don't understand it, sorry.
<tilpner>
~ > nix show-config | grep 'sandbox '
<tilpner>
sandbox = true
<clefru>
btw: "git cherry-pick 99131b4fc20f1256bdfc4c5237392434c157d234" + "nix-build '<nixpkgs>' -A python36Packages.pycurl" fetched me a prebuild pycurl from the cache
<robstr_>
Maybe not a nixos question, is it possible to drop all incoming packages to a specific host / port via the cmd line ? Want to test my application if it handles timeout correct
<clever>
robstr_: iptables
<robstr_>
clever: is networking.firewall handled with iptables ?
<clever>
robstr_: yes, but you can also manually change the state at runtime
<{^_^}>
→ 72742bae by @ryantm: xpra: 2.1.3 -> 2.2.4
<{^_^}>
→ edb0d094 by @gebner: Merge pull request #36723 from ryantm/auto-update/xpra
<clever>
and some inputs got changed
<clefru>
clever & tilpner: thanks for the interesting discussion and your help! I suggested a cherry-pick in https://github.com/NixOS/nixpkgs/pull/37038 and did the same locally, which fixes the problem for me.
<{^_^}>
→ c7e1dff9 by Andreas Wiese: nixos/window-managers: actually add evilwm to default.nix
<{^_^}>
→ 5c1c4fbb by @Mic92: Merge pull request #37084 from herrwiese/evilwm
<{^_^}>
[nixpkgs] @Mic92 pushed commit from Andreas Wiese to release-18.03 « nixos/window-managers: actually add evilwm to default.nix »: https://git.io/vxOjW
<kandinski>
I just installed Nixos (18.03) and I want to run my work virtualbox vms as before. Is it enought o configure virtualisation.virtualbox.host.enable = true?
<clever>
kandinski: yes
leotaku has quit [Ping timeout: 252 seconds]
<unlmtd>
kandinski: yes. im curious what you use virtualbox for
<{^_^}>
[nixpkgs] @Mic92 merged pull request #37068 → pkgs: some speech-related things → https://git.io/vxLOt
<{^_^}>
→ e6c63e9f by SLNOS: flite: 2.0 -> 2.1, switch to fetchFromGitHub, enable ALSA
<{^_^}>
→ 507d96d3 by SLNOS: speech-tools: 2.1 -> 2.4, unbreak
<{^_^}>
→ 77864f73 by SLNOS: eflite: cleanup expression, switch to fetchpatching from debian
kerrhau has quit [Ping timeout: 260 seconds]
<kandinski>
unlmtd: work. We develop an intranet that has to work on a given version of a given OS, so we develop/test on a VM with that OS and version.
<unlmtd>
I tried it once it was really heavy
<unlmtd>
maybe its easier to setup, but nix changes the meaning of 'setup'
<kandinski>
I know
<kandinski>
that's why I'm installing Nixos on my laptop, and then I'll do the same on my personal server, etc.
<clever>
i just virtualisation.virtualbox.host.enable = true; and then i can run VirtualBox and launch things
<{^_^}>
→ 66fd7d2b by @tsaeger: pythonPackages.ofxhome: 0.3.1 -> 0.3.3
<{^_^}>
→ 9a49811f by @gebner: Merge pull request #30263 from tsaeger/tsaeger/ofxhome
<clever>
that can be applied to almost any derivation, as long as you understand the subset of files you need to keep
John882 has joined #nixos
<clever>
so without any recompile of openssl, i can just ditch the junk i dont need
<clever>
and its still dynamic and depending on the right glibc
<unlmtd>
were even working on something to ditch SSL
<clever>
in this case, i need to generate a self-signed cert at runtime
<unlmtd>
a simple HIP implementation
<unlmtd>
we might want to contract you to work on the nix code, if youre interested
<unlmtd>
I love nix but im not an expert :P
<clever>
heh, already been hired for my nix skillz :P
<unlmtd>
yea its in high demand
<unlmtd>
ill come back to it when the tme comes
Fare has joined #nixos
<{^_^}>
[nixpkgs] @dotlambda merged pull request #36338 → nixos home-assistant: a couple of fixes → https://git.io/vAQz7
<{^_^}>
[nixpkgs] @dotlambda pushed commit from @peterhoeg to master « nixos home-assistant: a couple of fixes (#36338) »: https://git.io/vx3vn
John882 has quit [Ping timeout: 240 seconds]
<jluttine>
manveru: i tried your jekyll solution now but i still get the same error as before when running bundix: "/nix/store/dns77vf9frhd2s2xvcdd47rs2l08nqin-ruby-2.4.3/lib/ruby/2.4.0/json/common.rb:156:in `parse': 409: unexpected token at ';1mwarning: unknown setting 'signed-binary-caches' (JSON::ParserError)"
<{^_^}>
[nixpkgs] @lprndn opened pull request #37312 → luakit: init at 2018.08.10 → https://git.io/vx3va
<jtojnar>
nioncode: yes, LnL is right, I will test it now
<LnL>
jtojnar: I'm looking at it right now :)
<jtojnar>
cool, LnL, I would just run `env XDG_DATA_DIRS= $(nix-build -A spaceFM)/bin/spacefm` or something
<jluttine>
manveru: ah, now it seems to work after i deleted the lock file and gemset.nix. so nvm. thanks for the gist!
<mpickering>
Hmm, this Qt application builds successfully but when run "This application failed to start because it could not find or load the Qt platform plugin "cocoa""
<LnL>
jtojnar: it's rebuilding anyway
John882 has joined #nixos
<hyper_ch>
hmmm, just tried to install nixos on a notebook... in the live iso (self generated) ethernet worked fine... after installation - no ethernet connection
la_putin has joined #nixos
la_putin_ has quit [Read error: Connection reset by peer]
<{^_^}>
→ 77d1fdb2 by Timo Kaufmann: ecm: 6.4.4 -> 7.0.4
<{^_^}>
→ faf094ad by @gebner: Merge pull request #36435 from timokau/origin/ecm-update
<{^_^}>
[nixpkgs] @jtojnar pushed commit from @nioncode to master « spacefm: fix segfault by exporting shared-mime-info/share in XDG_DATA_DIRS »: https://git.io/vx3fZ
<{^_^}>
[nixpkgs] @gebner merged pull request #36797 → spacefm: fix segfault by exporting shared-mime-info/share in XDG_DATA_DIRS → https://git.io/vxesI
<dbe>
You still have to do "let nixpkgs = (import <nixpkgs> {})" etc in standalone default.nix-packages, right? Actually, maybe I'm thinking about this all wrong. I have a piece of software. I want to develop this software on a nix machine, and ultimately build a package for it. What do you guys do in this situation?
<clever>
dbe: i usually start my project as: with import <nixpkgs> {}; stdenv.mkDerivation { name = "name"; src = ./.; }
<clever>
and if it has several components, i may switch over to newScope
<dbe>
that's basically what I'm doing now. What's newScope?
<{^_^}>
→ b6ca5726 by @xeji: epoxy: fix darwin build
<{^_^}>
→ aa878c2d by @Mic92: Merge pull request #37315 from xeji/epoxy-darwin-fix
<sphalerite>
chisui: cool! Also for your prompt you could use $name (the shell derivation's name) rather than fiddling with $NIX_SHELL_FILE_PATH
<chisui>
sphalerite: what do you mean by that?
<abcdw>
Hi guys. I'm pretty new to nixos and trying to play with rust on it. I installed latest version of rust using mozilla overlay, it works, but emacs spamming with an error:eldoc error: (user-error No such directory: /nix/store/0n7zmyas042kxl5q72ikdrc4gqjcr8rj-rust-1.24.1-2018-02-27-d3ae9a9e0/lib/rustlib/src/rust/src. Please set ‘racer-rust-src-path’ or ‘RUST_SRC_PATH’)
<sphalerite>
samueldr: do you now get a highlight whenever someone refers to the logs? x)
humanoyd has joined #nixos
nD5Xjz has quit [Ping timeout: 256 seconds]
<abcdw>
sphalerite: thanks a lot, going to try second option
leat has quit [Ping timeout: 246 seconds]
<Lisanna>
Okay, here's an interesting meta-nix project: a function which gets the build log of a derivation
<Lisanna>
...I'm guessing no one has tried to implement this yet?
<sphalerite>
Lisanna: in nix?
<Lisanna>
yea
<Lisanna>
i.e. the build logs ends up in $out
<chisui>
sphalerite: cool, I'll try using $name
<clever>
Lisanna: there is already `nix-store -l` and `nix log`
<sphalerite>
Lisanna: but why x)
<sphalerite>
Lisanna: I suppose you could just edit mkDerivation to add a "buildlog" output by default, and add `exec &> >(tee $buildlog)` to stdenv's setup.sh
<{^_^}>
[nixpkgs] @peterhoeg pushed 4 commits to release-18.03: https://git.io/vx3Uj
<{^_^}>
→ 3ae19a36 by @Mic92: python.pkgs.pushover: there are no tests
<{^_^}>
→ c68ed708 by @peterhoeg: pythonPackages.pushover: init at 0.3
<{^_^}>
→ 0b3170b8 by @Mic92: python.pkgs.pushover: tests exists, but require network
tertle||eltret has quit [Quit: Connection closed for inactivity]
<clever>
sphalerite: oh, that may help with the trouble i had doing exec 2>&1 | tee logfile
<Lisanna>
sphalerite 1) because the more things you can do in nix the better 2) you want to return the result of running a test without being responsible for the contents of the log
<clever>
Lisanna: that will very likely break --repeat
<Lisanna>
clever only if the build log isn't deterministic
<clever>
parallism and timestamps
<symphorien>
Also it will add a dependency on the build tools because their hash might appear in it
<clever>
though its its own output, so nix can gc that
leat has joined #nixos
<Lisanna>
cool thing about this (if it's possible) is maybe you have a test that is a fixed-output derivation, and you want to get the test run log without screwing with the fixed-output-ness of the derivation
<symphorien>
Well as long as you keep the log it will keep the build tool, won't it ?
<clever>
symphorien: yep
<clever>
Lisanna: fixed-output derivations can only have 1 output
<sphalerite>
I have a build (rust on 32-bit arm) that fails when cores is set too high, because it tries to allocate more memory than you can allocate on 32-bit arm. I'd like to limit the number of cores it runs the tests with in order to fix that, what would the neatest way of doing this be?
nD5Xjz has joined #nixos
<Lisanna>
clever I know... this would be a function which takes a derivation (FOD or not) as input and produces a new derivation which is just the build log
magnetophon has joined #nixos
Fare has joined #nixos
<clever>
sphalerite: --cores or nix.conf build-cores?, or patch it to scale-back $NIX_BUILD_CORES based on ram?
<Lisanna>
now, this is only possible if it's possible to run nix log in the nix sandbox... that likely lives in /nix/var. does /nix/var info get copied into runInLinuxVM? ah, probably not, right?
<clever>
yeah, no recursive nix
Xal has quit [Ping timeout: 240 seconds]
<sphalerite>
right now I'm thinking something like preCheck = ''min() { echo $(( $1<$2 ? $1 : $2)) ; } ; export NIX_BUILD_CORES=$(min $NIX_BUILD_CORES 8)'';
<clever>
sphalerite: that sounds good, maybe replace the 8 with something based on free -g
Xal has joined #nixos
<sphalerite>
it's not a memory issue, it's an address space issue
davidlt_ has joined #nixos
<clever>
is this process or thread parallism?
<sphalerite>
thread apparently :p
<clever>
ah
<clever>
then a hard-coded limit for 32bit makes sense
<clever>
i also just had a random thought
nD5Xjz has quit [Ping timeout: 240 seconds]
<clever>
the spectre/meltdown changes, require that the 3-4gig range isnt mapped anymore
<clever>
could that allow that range to be used by userland??
<clever>
4gig of userland memory in 32bit!
davidlt has quit [Ping timeout: 240 seconds]
<clever>
hmmm, my desktop locked up in the middle of a ghc rebuild, and after reset, it locked up in the middle of nixos-rebuild
<clever>
hdd light is solid on, no reply to arp
<sphalerite>
is there a way to check word size in nixpkgs?
<sphalerite>
because i'm guessing the same problem will occur on i686
<clever>
i think so
<clever>
where was lib.platforms defined...
<sphalerite>
oh derp it was is32bit not is32Bit
<clever>
ah
<clever>
error: attribute 'is32bit' missing, at (string):1:1
<abcdw>
sphalerite: I installed rust overlay in ~/.config/nixpkgs... , but I trying to set environment variable in /etc/nixos/configuration.nix and it fails. How to bypass it?
<clever>
abcdw: configuration.nix doesnt load those overlays by default
<clever>
you have to nixpkgs.overlays = [ (import /path/to/overlay) ];
<ottidmes>
weird, I am getting infinite recursion with nix-env, but only when I set `-I nixpkgs-overlays=`, how could my overlays prevent recursion :S (my NIX_PATH has a nixpkgs-overlays entry)
<sphalerite>
ottidmes: by using self when you should be using super
<clever>
amdgpu doesnt even change the res in text mode
qknight has joined #nixos
<ottidmes>
sphalerite: That would only make sense when I would have set my overlays to use overlays, but I set it to empty, i.e. no overlays, but it might be defaulting to some other location with outdated overlays, I will check that
<sphalerite>
ottidmes: yeah it uses ~/.config/nixpkgs/overlays by default
qknight has quit [Client Quit]
<sphalerite>
where "by defaults" includes "if the NIX_PATH entry doesn't exist"
<sphalerite>
ottidmes: try --arg overlays []
vaninwagen has quit [Quit: Connection closed for inactivity]
<clever>
sphalerite: ahh, fixed
abcdw_ has quit [Quit: Page closed]
<ottidmes>
sphalerite: I would have thought `pathOverlays = try <nixpkgs-overlays> "";` would have failed when I set it like that and I have nothing in ~/.config/nixpkgs, but your suggestion is better (does not depend on any defaulting behavior) and works, so I will be using that instead, thanks
zzamboni has quit [Quit: Leaving.]
<{^_^}>
[nixpkgs] @shlevy pushed to binutils-2.30 « Disable tests on llvm. »: https://git.io/vx3kZ
<{^_^}>
[nixpkgs] @jtojnar pushed commit from @nioncode to master « conky: provide option to build with pulseaudio support »: https://git.io/vx3kC
<sphalerite>
ottidmes: if you want to avoid all defaulting behaviour, consider importing pkgs/top-level/default.nix rather than default.nix which passes on to pkgs/top-level/impure.nix
xcmw has joined #nixos
abcdw has joined #nixos
xcmw has quit [Client Quit]
<{^_^}>
[nixpkgs] @jtojnar merged pull request #36387 → conky: provide option to build with pulseaudio support → https://git.io/vAdrZ
ryantm has joined #nixos
<abcdw>
clever, sphalerite thanks a lot for help. solved the problem by nixpkgs.overlays = [ (import /home/abcdw/.config/nixpkgs/overlays/rust-overlay.nix) ]; and environment.variables.RUST_SRC_PATH = "${pkgs.latest.rustChannels.stable.rust-src}/lib/rustlib/src/rust/src";
<ottidmes>
sphalerite: I know, it is just strange, I did not get this loop with nixos-rebuild with the same argument, I cannot explain what is the cause at this point. I actually only need to overwrite overlays with nixos-rebuild
aanderse has quit [Ping timeout: 240 seconds]
<clever>
ottidmes: nixos-rebuild uses the overlays set in nixpkgs.overlays = [ ... ];
<clever>
ottidmes: there is also a bug where nixos will fail due to problems in the default overlays, but never actually obey them
<clever>
its loading them somewhere it shouldnt, and not really using the result
humanoyd has quit [Quit: WeeChat 2.0.1]
<ottidmes>
clever: there is another where the manual says it supports overlays.nix, but when you use it as nixpkgs-overlays it has to be directory, which I will double check as well, if it still does that I will check if there is already an issue out for that
<jtojnar>
nioncode: patchelf can be buggy, modifying ELF files is not really possible in general, fortunately most ELF files are reasonable enough
nD5Xjz has quit [Ping timeout: 260 seconds]
<jtojnar>
sometimes it will work if you disable stripping or use patchelfUnstable
<jtojnar>
nioncode: direct download should be fine, if the URL is public and does not change
<{^_^}>
[nixpkgs] @lheckemann closed pull request #34019 → dnsmasq module: test config on system build → https://git.io/vN06L
<nioncode>
jtojnar: the link should be static imo since the version info is part of the link. only question is if it's allowed to directly download, for other packages I saw that the user must use prefetch-url manually
<jtojnar>
nioncode: I think it is used if the link requires registration or the site explicitly forbids it
<ottidmes>
clever: I had the bug with nixpkgs-overlays where it would fail with nixos-rebuild when I removed one my overlay files in my config, it would complain on rebuild that it was no longer available, even though my new config never referenced the removed file any more, which explicitly setting overlays to empty fixed for me, but I cannot reproduce it anymore (maybe because I am now on Nix 2.0)
abcdw has quit [Ping timeout: 260 seconds]
Ariakenom has quit [Ping timeout: 240 seconds]
<nioncode>
jtojnar: ok, I rebased and pushed, could you please review again?
<jtojnar>
nioncode: you are setting interpreter only for p4merge.bin but wrapping *.bin, why is that?
nD5Xjz has joined #nixos
<nioncode>
jtojnar: yep, that's wrong, let me fix that
<robstr_>
ottidmes: is it possible to do this inside my stdenv.mkDerivation block ?
<symphorien>
I doubt that
<ottidmes>
robstr_: Not that I am aware of, that would basically force others to use unfree, breaking the whole point of that flag
Fare has joined #nixos
<jtojnar>
nioncode: I am not sure if it will work though, AFAIK QT does not have stable ABI
<robstr_>
ottidmes: Ok, I'm trying to build my dev env for a specific project, is there another recommended way ?
<ottidmes>
robstr_: Ah, I just thought of a hacky way to do it, but the recommend way is to just setup your nixpkgs configuration correctly with allowUnfree = true;
<symphorien>
let pkgs = import <nixpkgs> { config = { allowUnfree = true; }; }; in
<symphorien>
When you are the user, you can do whatever you want
<ottidmes>
symphorien: Yeah, was about to paste that, but I would consider that the hacky way, not recommended if you want to do it properly
<symphorien>
Well for a shell.nix if find it perfectly legitimate
Ariakenom has joined #nixos
<nioncode>
jtojnar: Qt has strict ABI compatibility for major versions, so it should work imo. I spent a few hours trying to get it to run with the p4v provided Qt libraries, but I failed
<clever>
nioncode: if you install anything qt based with nix-env, its libraries wind up in ~/.nix-profile/lib/ and can break other versions
<clever>
nioncode: i have an issue open about that
<symphorien>
For inclusion in a larger scheme, when you are not user but provide a facility, this is problematic, granted
<{^_^}>
→ cef5ed1b by @Ralith: matrix-synapse: 0.26.0 -> 0.26.1
<{^_^}>
→ bcbe9e98 by @Mic92: Merge pull request #37159 from Ralith/matrix-synapse
<{^_^}>
[nixpkgs] @Mic92 pushed commit from @Ralith to release-18.03 « matrix-synapse: 0.26.0 -> 0.26.1 »: https://git.io/vx3tZ
<robstr_>
clever: thank you for checking
freeman42x]NixOS has joined #nixos
<Lisanna>
...thinking about fixed-output derivations, and had an idea: it would be nice if you could have a FOD with multiple possible outputHashes (As opposed to currently where you just get one). This is still safe since you still have to prove that you know about all of the possible outcomes.
nD5Xjz has quit [Ping timeout: 246 seconds]
<clever>
Lisanna: it would be easyer to make a fixed-output derivation that returns a single directory, then have a non-fixed derivation that splits it into many outputs
<clever>
then if you only depend on an output of the 2nd, it will be able to gc the unused parts
<Lisanna>
clever the intended usecase is more "the result could be one of these possibilities, but I'm not sure which"
<clever>
ah, yeah, nix isnt really capable of that
<Lisanna>
I know, this is a theoretical feature idea :)
<Lisanna>
I think that having it would not violate any of the core principles of nix though
<{^_^}>
→ e33c221c by @ryantm: ipset: 6.35 -> 6.36
<{^_^}>
→ df538c82 by @Mic92: Merge pull request #37027 from ryantm/auto-update/ipset
<ottidmes>
clever: NIX_PATH contains nixpkgs-overlays=/etc/nixpkgs/overlays.nix and gives error: opening directory '/etc/nixpkgs/overlays.nix': Not a directory, which does not match the description found here: https://nixos.org/nixpkgs/manual/#sec-overlays-install
<ottidmes>
I guess I will just have to make an issue for that, but have to go now
<clever>
if pathOverlays != "" && pathExists pathOverlays then overlays pathOverlays
<clever>
# it's a file, so the result is the contents of the file itself
<clever>
import path;
<ottidmes>
clever: yeah, I know, I know, I have checked the source code before, yet why then does it fail?
<clever>
ottidmes: what line#'s does it give with --show-trace?
justanotheruser has joined #nixos
<ottidmes>
clever: Last line before error: while evaluating 'overlays' at /wheel/fork/nixpkgs/pkgs/top-level/impure.nix:47:18, called from /wheel/fork/nixpkgs/pkgs/top-level/impure.nix:59:61:
<Lisanna>
You call builtins.pathExists on the outPath of the FOD. Of course, maybe it just hasn't been attempted to be built yet. So, you also need an indicator derivation that you only build when attempting to build the FOD.
<clever>
arg 0 is a program
<clever>
Lisanna: aha, builtins.exec must print a nix expression to stdout
<Lisanna>
possible return values of this theoretical function are "success", "failure", and "unknown" (hasn't been attempted yet)
<clever>
which nix will parse, and return as-if you did import on it
<Lisanna>
(of course you have to unsafeDiscardStringContext it)
<clever>
then nix wont even bother trying to build it
<Lisanna>
well, that's the idea: trying ot build a FOD that fails ruins your entire day. the assumption is that the build attempt of the FOD happens somewhere else / sometime else
<clever>
Lisanna: oh, also, when a derivation fails to build, nix leaves the failed output in /nix/store/!
<clever>
its just not registered as valid
<clever>
so no proper nix tool will read it
<Lisanna>
clever ...would pathExists find it?
<clever>
and it gets deleted automatically if its in the way
<clever>
GC also eats those first
<clever>
yeah, pathExist would find it
<Lisanna>
...oh. darn.
loonquawl has joined #nixos
<clever>
nix leaves those behind, so you can inspect why it failed
<Lisanna>
hey wait a second...
<clever>
a fixed-output derivation may leave html in the .tar.gz
<MichaelRaskin>
Really, if you can attempt the build elsewhere, just write the outcome somewhere outside the store
<Lisanna>
MichaelRaskin ugh, that's extra infrastructure
<clever>
Lisanna: importnative may also be of interest to you
<clever>
Lisanna: i think that lets you write your own primops in c++
<MichaelRaskin>
Triggering the build is extra infrastructure anyway, no?
<clever>
which can then do anything you want
<Lisanna>
clever for failed FODs nix will also create an additional output with the actual hash that it ended up being, so that if you fix the expression and build again it'll already be there
<Lisanna>
so, if you know what both of those are, you can call pathExists on both
<clever>
Lisanna: only in certain modes, there is a special prefetch flag you have to set
<Lisanna>
oh, darn
<{^_^}>
[nixpkgs] @mpickering opened pull request #37326 → open-orienteering-mapper: init at 0.8.1 → https://git.io/vx3YB
<clever>
Lisanna: but you could maybe use readFile on the potentially invalid output of a FOD
<clever>
to just read it, wether it worked or not
<Lisanna>
clever so IFD from the outPath of the derivation, with unsafeDIscardStringContext?
<Lisanna>
another possibility: construct multiple FODs each with a different candidate outputHash. Then pathExists on all of them - if one of them exists, you know which one it is. If none of them exists, then assume that it hasn't been attempted yet.
<Lisanna>
if you write your code bad then it's possible that there's an outputHash that you missed, and you won't be able to distinguish that case from "not attempted"
justanotheruser has quit [Ping timeout: 260 seconds]
<Lisanna>
wait, nvm, my idea won't work for the reason you gave earlier - the path is still created even if the FOD fails. so I would need to also insert an indicator in the outPath
<Lisanna>
unless there's a way to also check if the path is valid
<{^_^}>
[nixpkgs] @dotlambda pushed 2 commits to release-18.03: https://git.io/vx3O3
<{^_^}>
→ 6daee8c3 by @dotlambda: pythonPackages.tmdb3: 0.6.17 -> 0.7.2
<{^_^}>
→ 5e4411f4 by @dotlambda: pythonPackages.tmdb3: move expression
<{^_^}>
→ b549f18b by @dotlambda: pythonPackages.CommonMark: move expression
<{^_^}>
→ d7d4ce73 by @dotlambda: pythonPackages.CommonMark: 0.6.3 -> 0.7.5
<{^_^}>
→ 9f276ccd by @dotlambda: pythonPackages.bleach: 2.1.2 -> 2.1.3
risciii has quit [Ping timeout: 248 seconds]
<testuser>
hi guys, is there a simple way to update a variable (precisely the version) of a certain package when doing nixos-rebuild? I would like to rebuild the system with an updated version of the package?
<clever>
testuser: you must override the src, and its recomended to also override the name
<clever>
setting the version variable will have no real effect
newhoggy has quit [Ping timeout: 240 seconds]
<testuser>
clever, okay, so do I need to write a new .nix package with updated version, then link it into the configuration.nix ?
<clever>
you can use overrides to change it without a new file
<{^_^}>
[nixpkgs] @peterhoeg pushed to release-18.03 « nixos home-assistant: a couple of fixes (#36338) »: https://git.io/vx3O5
<{^_^}>
[nixpkgs] @dotlambda pushed 24 commits to release-18.03: https://git.io/vx3OF
<{^_^}>
→ 9dfac949 by @dotlambda: pythonPackages.CommonMark: move expression
<{^_^}>
→ 090ca7e4 by @dotlambda: pythonPackages.CommonMark: 0.6.3 -> 0.7.5
<{^_^}>
→ f9def2f4 by @dotlambda: pythonPackages.readme_renderer: 17.2 -> 17.4
zzamboni has quit [Quit: Leaving.]
<Lisanna>
MichaelRaskin actually your suggestions sounds easier at this point tbh
larsvm has quit [Ping timeout: 264 seconds]
leotaku has joined #nixos
nD5Xjz has joined #nixos
<Lisanna>
I'll just have the build threads dump the evaluation result (success or fail) to some state directory, and import the results from that out-of-band
stanibanani has joined #nixos
risciii has joined #nixos
<Lisanna>
build result*
zzamboni has joined #nixos
<testuser>
clever, http://dpaste.com/0YR2CEX, I'm getting "error: undefined variable ‘name’ at /etc/nixos/configuration.nix:271:51", why isn't the name previously set recognized?
<clever>
testuser: you need to use drv.name there
<clever>
oh
<clever>
or use rec { on line 2
<clever>
rec allows you to refer to other keys in the same set
<tobiasBora>
Hello, (sorry, if I don't answer you, that's just because I'm in a place with bad internet connexion, but I'll read your messages anyway latter)
<leotaku>
is it possible to override a services attributes in nix similar to how it works with packages?
risciii has quit [Client Quit]
<testuser>
clever, ok the config is valid now, but the system is not picking the override up; it's still building the old version, the environment.systemPackages = with pkgs; [ bacula ... ]; is used.
<clever>
testuser: can you pastebin the entire configuration.nix file?
rauno has joined #nixos
<testuser>
clever, I have secrets in there, so I would rather not, can I copy-paste whatever is needed?
<clever>
testuser: you could also just delete the lines that have the secrets
<leotaku>
I just need to make some slight edits to a service and do not want to copy+paste the definition from nixpkgs.
<tobiasBora>
I'd like to know, does nix provide a way to create a binary cache between several computers? I'd be curious to know how difficult it would be to create a "community binary cache" for armv7.
<clever>
tobiasBora: currently, you need to setup every machine as a seperate binary cache, and configure every machine to ask every other machine
<tobiasBora>
clever: hum… and would it be possible to create a central binary cache that then asks to other machines, to centralize the list of servers?
<testuser>
clever, it appears that if I change the ${drv.name} into the "bacula-9.0.6", everything is fine, the variable is not resolved properly.
<clever>
testuser: ah yeah, drv.name refers to the old name
rauno has quit [Ping timeout: 256 seconds]
<clever>
so you need rec { and ${name} to refer to the name you just set
<testuser>
clever, why does the drv.name refer to the old name?
chrios_ has quit [Ping timeout: 240 seconds]
<clever>
testuser: the drv attrset is all of the old values
stanibanani has quit [Ping timeout: 256 seconds]
<testuser>
clever, is this just for the drv attrset or any name I chose ?
<tobiasBora>
clever: oh nice thank you for the sharing. And meanwhile, is it possible to automatically send the builds to a centralized binary cache at each new compilation?
<fearlessKim[m]>
wideland url is broken on master. I just updated the url with the correct one url = "http://launchpad.net/widelands/build${version}/build${version}/+download/widelands-build${version}-src-gcc7.tar.bz2"; with wget from my shell it works but not from nix-build, any idea why ?
<fearlessKim[m]>
(download stays stuck at 0%)
<clever>
tobiasBora: not really, you should probably just use hydra for this kind of thing
<clever>
tobiasBora: overrideAttrs takes a function that accepts 1 argument (typically called drv), and it will call that function with the original attributes
<clever>
testuser: oops, ^
<clever>
testuser: and it will then use the result of that function to modify some attrs
<tobiasBora>
clever: hydra's code you mean? I'll give it a look thanks. Does hydra provide any way to "check" that the binaries are honnestly build, for example by trying to send the same build on several clients and see the result?
<clever>
tobiasBora: not currently
<clever>
tobiasBora: hydra is more used to build things automatically, on an array of trusted machines, and then pull the results back to 1 central place
<clever>
then the slaves can GC safely and you dont loose anything
<clever>
tobiasBora: also, not a lot of the compilers are able to produce bit-perfect results, so such checks can falsely fail
<testuser>
clever, I have one more question that I always wanted to know: when debugging the package installation and entering nix-shell, after manually executing unpackPhase/patchPhase (which is fine), before executing the next script configurePhase, I need to cd into the directory to where the package was download. Why is that, why doesn't unpachScript already cd into the dir ?
chrios_ has joined #nixos
<clever>
tobiasBora: but you can review things with `nix-build foo.nix -A bar --repeat 2`, as root, it will build it 3 times, and tell you if it differs
<clever>
that will forciably disable mysql, even if other things wanted it on
<rauno>
hey, does nixos have some support for OCFS2 filesystem ?
<manveru>
rauno: is that in a kernel driver?
<tobiasBora>
clever: ok thank you. Just, not sure to see what you mean by "slaves can gc safely". and it's too bad that two compilers send different outputs on the same input, even with the same compiler
<clever>
tobiasBora: the hydra will copy all outputs to itself once built, so there is no need to keep the results on the build slaves
<ottidmes>
clever: I am using a gitignore with a set of secret files, how do you deal with files? secrets folder or something?
<rauno>
manveru, wiki says OCFS2 (version 2) was integrated into the version 2.6.16 of Linux kernel.
y0no has joined #nixos
kitemikaze has quit [Read error: Connection reset by peer]
<clever>
ottidmes: i just keep all the secrets in secrets.nix, which is in .gitignore
<clever>
ottidmes: and then i just let secrets = import ./secrets.nix; in everywhere
<manveru>
rauno: then it should work :)
<testuser>
clever, thank you that explains it fully. One more thing, when manually executing the phases in nix-shell, when doing installPhase, the script will return an error since it doesn't have access to the /nix store: http://dpaste.com/0P3FTGM ? Any ideas how to be able to manually execute installScript to see why it's failing?
kitemikaze has joined #nixos
<ottidmes>
clever: so you inline all your secret files? I guess I do not actually have non-text secret files, so it should be possible
<{^_^}>
→ fb6c5352 by @romildo: deepin-gtk-theme: 17.10.5 -> 17.10.6
<{^_^}>
→ 250527c0 by @Mic92: Merge pull request #37322 from romildo/upd.deepin-gtk-theme
zzamboni has quit [Quit: Leaving.]
<manveru>
rauno: you can add it to `boot.kernelModules` to make sure it works after reboot as well, or `boot.initrd.extraKernelModules` if you need to boot from it...
<rauno>
yep, but still, it probably needs some extra package also
<michas__>
Hi I tried to use the jetbrains-toolbox (https://www.jetbrains.com/toolbox/app/) on nixos. I used `patchelf --set-interpreter /nix/store/2kcrj1ksd2a14bm5sky182fv2xwfhfap-glibc-2.26-131/lib/ld-linux-x86-64.so.2 --set-rpath /nix/store/y76fs08y8wais97jjrcphdw2rcaka1qa-fuse-2.9.7/lib:/nix/store/4csy6xvbrqxkp3mk6ngxp199xkr476lj-glib-2.54.3/lib:/nix/store/r43dk927l97n78rff7hnvsq49f3szkg6-zlib-1.2.11/lib jetbrains-toolbox` to fix the binary. It
<michas__>
now runs, but it now complains because it is not able to write to some temporyary directory it creates. Did I miss something obvious?
spear2 has quit [Ping timeout: 246 seconds]
<{^_^}>
[nixpkgs] @coreyoconnor opened pull request #37329 → transmission: refactor preStart into script. → https://git.io/vx3Gp
nD5Xjz has quit [Ping timeout: 256 seconds]
ryanartecona has quit [Client Quit]
<michas__>
(Not sure if it is a good idea to us that toolbox. It will mainly download other application, but those should be plain java, so there is probably no need to "fix" them.)
<tobiasBora>
clever: ok thank you very much, I'll give a look to hydra, thanks
knupfer has quit [Remote host closed the connection]
endformationage has joined #nixos
jensens has joined #nixos
<foldingcookie>
what controls laptop hibernate-on-lid on nixOS? I didn't expect it to be a thing without installing acpid but perhaps systemd has agglomerated that functionality now?
<{^_^}>
→ b506cdbf by @mbode: spass: make gcc dependency explicit
<{^_^}>
→ dda33d69 by @mbode: samtools: make gcc dependency explicit
boomshroom has joined #nixos
<boomshroom>
Hello!
<coconnor>
howdy
oida has joined #nixos
<boomshroom>
If I have a derivation with multiple sources, how do I configure them all to be in the right path?
<boomshroom>
Ideally, I'd like it to be reusible for different inputs and different paths.
justbeingglad has left #nixos [#nixos]
<testuser>
how can I reference $out in configureFlags, I'm getting the following when I build the package: "configure flags: --disable-static --prefix=/nix/store/icv4akj4q27jjimq89ill93wl0rrqfhd-bacula-9.0.6 --with-sqlite3=/nix/store/vk9416l2cv1xy1vckm6ygq0f717kgb5i-sqlite-3.22.0-dev --with-postgresql=/nix/store/khz84zphhqyszfjwqdqmlja6s4g8klpv-postgresql-9.6.7 --with-logdir=\$out/log --with_working_dir=\$out/working
<{^_^}>
→ 4c52e34c by @Ericson2314: stdenv: Clean up check meta args
<{^_^}>
→ c26252af by @Ericson2314: lib, stdenv: Check `meta.platforms` against host platform and be open world
<{^_^}>
→ eae19f3c by @Ericson2314: release-lib: Adapt to work with new meta.platforms
coot has joined #nixos
<boomshroom>
I just made the builder call a build function, and now I have to fix the shas
<fendor>
someone here having experience with haskell-ide-engine and nix?
dan_b has quit [Ping timeout: 264 seconds]
<{^_^}>
[nixpkgs] @Ericson2314 opened pull request #37337 → Meta check 18.03 → https://git.io/vx3lC
<mpickering>
fendor: I have
<mpickering>
but always ask your question
<mpickering>
boomshroom: So it works fine for you? I'm not sure I see the problem .. :)
ma27 has quit [Ping timeout: 265 seconds]
<boomshroom>
The main problem was that builtins.toFile doesn't like derivation outputs in its source, so I moved the real code to a build value and called that from the builder.
<boomshroom>
Now however, I have a different problem.
sasoiliev has joined #nixos
<sasoiliev>
Hi all
<sasoiliev>
I'm trying to set up NixOS on my box and I'm having a little trouble with the 'nixos-rebuild switch' command
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<sasoiliev>
Is my understanding that 'nixos-rebuild switch' is supposed to 1) take immediate effect and 2) remain in effect after rebooting?
<boomshroom>
sasoiliev: Correct. That said there are somethings which are difficult to swap post-boot which is why current-system and booted-system are both used.
<symphorien>
Yes, except for environnement variables and settings related to xorg which need a login/logout/restart at least
<boomshroom>
Also, applications that are already running will keep running until exited and relaunched.
<symphorien>
The list of non restarted services is printed by nixos rebuild
<boomshroom>
Found my problem: libgit has a folder with a dummy submodule and it's confusing nix.
<sasoiliev>
Right, so I wanted to set up NetworkManager (I'm setting up a laptop) as per https://nixos.org/nixos/manual/index.html#sec-networkmanager, so I added the 'networking.networkmanager.enabled = true;' in configuration.nix and then did a 'nixos-rebuid switch'
vaninwagen has joined #nixos
<{^_^}>
[nixpkgs] @jtojnar pushed 97 commits to gnome-3.28: https://git.io/vx3lh
<{^_^}>
→ 566e08ba by @jtojnar: deja-dup: remove unneeded rpath fixup
<{^_^}>
→ d37c7ff8 by @jtojnar: meson: fix rpath fixup
<{^_^}>
→ fe835eca by @jtojnar: epiphany: remove unneeded rpath fixup
ma27 has joined #nixos
mightybyte has joined #nixos
<boomshroom>
It should be safe to just disable submodules entirely for that repo
<sasoiliev>
The command succeeds, but I have two issues - 1) after switching the network-manager service is not running and 2) after reboot it is not even available (systemctl start network-manager says 'unit not found')
<mightybyte>
Would it be possible to make nix commands (mainly I'm thinking of nix-build, nix-shell, and nix-env) default to not trying to connect to the internet if all the dependencies are available locally?
dnovosel has joined #nixos
muzzy has joined #nixos
<boomshroom>
mightybyte: I don't see anything in nix-build --help that would suggest so, but I agree that it could be useful. Maybe post an issue on the github repo?
<symphorien>
sasoiliev: does nixos rebuild says it will build network manager or something related ?
<symphorien>
Maybe you miswrote your configuration.nix
<symphorien>
You can use nixos-option to actually query how your configuration.nix is parsed
<boomshroom>
`mkdir: command not found` Gosh darnit!
<mightybyte>
boomshroom: Ok, I'll do that.
zzamboni has quit [Quit: Leaving.]
<sasoiliev>
symphorien: I don't think it builds anything, it downloads the bits and finishes successfully.
<mightybyte>
boomshroom: Any idea what repo I should open that issue on?
nD5Xjz has joined #nixos
<symphorien>
Does it download nm then ?
<sasoiliev>
it does, yes
xcmw has joined #nixos
<boomshroom>
mightybyte: That's a nix specific issue rather than one for the package repo, so https://github.com/NixOS/nix
<sasoiliev>
if I say 'systemctl start network-manager' it does start and works
<sasoiliev>
(at least until I reboot the machine)
<symphorien>
Troubling
<symphorien>
Can you paste your configuration.nix ?
<guillaum1>
can I replace `nix-shell -A something` with `nix run ??? something`?
<guillaum1>
Same question for stuff packages with python / haskell. `nix run 'python27.withPackages (pkgs : with pkgs; [pkgs.pillow])'` fails with "error: don't know what to do with argument 'python27.withPackages (pkgs : with pkgs; [pkgs.pillow])'"
zzamboni has quit [Read error: Connection reset by peer]
zzamboni has joined #nixos
<boomshroom>
For some reason, my build function gets smushed to a single line, despite being made with a concat seperator of a newline.
dnovosel has quit [Ping timeout: 240 seconds]
<symphorien>
sasoiliev: so if you switch to this conf, the unit is found, but when you reboot it is not found anymore
<symphorien>
Do I understand correctly ?
<sasoiliev>
That's right. After a reboot I need to do 'nixos-rebuild switch' again to have the nm available again (plus the need to manually start it).
<symphorien>
This means that your bootloader configuration is wrong
<sasoiliev>
And that goes for all (or most) of the confgiurations I did in configuration.nix
<symphorien>
You boot into the wrong generatio
<sasoiliev>
OK, how can I boot into the correct one?
<symphorien>
Compare /run/current-system after switching and after rebooting (these are symlinks) to be sure
<symphorien>
nixos-rebuild switch should ensure that, but if you misconfigured it it may have no effect
<symphorien>
Legacy/uefi problems and so on
<leotaku>
When I define a new service unit/copy and edit one from nixpkgs, should I just be able to put it into "imports = [ ... ]" and have it work?
<leotaku>
Or is there something more complicated I have to be aware of?
<symphorien>
No idea if this could mess things up but you have both grub conf and systemd-boot conf
<sasoiliev>
Ah, right, I will look into this, thanks
rihards has joined #nixos
<symphorien>
If you use grub, the menu will conveniently show you which generation you boot into
<symphorien>
For systemd-boot, no idea
Sonarpulse has quit [Ping timeout: 276 seconds]
Arcaelyx has joined #nixos
newhoggy has joined #nixos
sasoiliev has quit [Ping timeout: 260 seconds]
<leotaku>
most of the stuff seems to work, but nix is unable to find pkgs.runtimeShell which seems to work with the service definitions from nixpkgs
<leotaku>
What could I be doing wrong?
zzamboni has quit [Read error: Connection reset by peer]
newhoggy has quit [Ping timeout: 256 seconds]
zzamboni has joined #nixos
xeji has joined #nixos
troydm1 has quit [Quit: What is Hope? That all of your wishes and all of your dreams come true? To turn back time because things were not supposed to happen like that (C) Rau Le Creuset]
troydm has joined #nixos
<symphorien>
leotaku: did you take the module from a more recent version of nixpkgs ?
<{^_^}>
[nixpkgs] @jtojnar pushed 44 commits to gnome-3.28: https://git.io/vx3B8
<{^_^}>
→ 8ed3a8e8 by @hedning: gvfs: fix build
<{^_^}>
→ 18fc5c09 by @hedning: gnome3.gnome-characters: fix build
<{^_^}>
→ 3ac5e247 by @jtojnar: gnome3.gucharmap: clean up
<leotaku>
symphorien: oh, I've found the problem, Indeed i copied from the master branch. my checkout uses stdenv.shell
<leotaku>
symphorien: thanks
sigmundv has joined #nixos
romildo has joined #nixos
<xeji>
anyone know what's wrong with hydra? error 500 ...
romildo has quit [Client Quit]
<{^_^}>
[nixpkgs] @obadz pushed commit from @ryantm to master « spaceFM: 1.0.5 -> 1.0.6 »: https://git.io/vx3B1
<infinisil>
In my case that would be some custom routes with `ip route`
shout-user93 has joined #nixos
<shout-user93>
hey, coming back to nix on macOS after a while - I understand that nix-darwin is a thing, but more a general question. when I run nix --version all I get is 2.0 - is there a way to see a more specific version?
<shout-user93>
nix-env --version gives me the same message
<shout-user93>
then later /nix/store/isrf97h8jqyx9p51frjdmz2609a88252-nix-2.0 - but still says unknown-deriver...
<clever>
but we know your running whats in nixpkgs, so thats simpler
<clever>
one sec
<clever>
nix-instantiate --eval '<nixpkgs>' -A nixStable.src.urls
<clever>
shout-user93: that gives you the url to the exact source you just installed
<shout-user93>
awesome, thanks
<disasm>
fpletz: I think I remember you mentioning you had a good way of sharing nixops state between members of a team. If so, could you give me some advice?
nD5Xjz has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @obadz pushed commit from Moritz Küttel to master « projectlibre: init at 1.7.0 »: https://git.io/vx3EC
alex`` has quit [Quit: WeeChat 2.0.1]
<{^_^}>
[nixpkgs] @jlesquembre opened pull request #37341 → terminus: init at 1.0.0-alpha.41 → https://git.io/vx3EB
<boomshroom>
I'm trying to unpack a set of sources into specific paths to prepare for a build. Currently, my build function gets squashed into a single line when I try to call it from the builder.
<clever>
boomshroom: can you gist your entire nix file?
<clever>
runCommand is just a helper to set buildCommand for you
<NickHu_>
Hmm, I'm noticing an issue whereby I do a nix-channel --update, and it doesn't actually seem to update the channel despite giving me the downloading and unpacking channels output...
<clever>
NickHu_: each user has his own channels, and root's are available to all users
<clever>
NickHu_: check `nix-channel --list` as each user
<NickHu_>
I understand that clever; I am trying to install something from nixpkgs-unstable
rardiol1 has joined #nixos
<NickHu_>
I even removed it, added it under a different alias, and then did a nix-channel --update (all as my local user)
<NickHu_>
But nix-env -iA nixpkgs-unstable.blah doesn't recognise the attribute path
<clever>
NickHu_: how are you installing it?
<NickHu_>
nix-env -ivA nixpkgs-unstable.pandoc --- error: attribute 'nixpkgs-unstable' in selection path 'nixpkgs-unstable.pandoc' not found
<clever>
what does nix-channel --list say?
<NickHu_>
ah it says it's at the alias nixpkgs
<clever>
the man page tells you how to set the name when you add
<NickHu_>
But still, it tries to install 2.0.6, which isn't the latest version on nixpkgs-unstable
<NickHu_>
So I am still suspicious it isn't updating the channel properly
<boomshroom>
Now I'm seeing `cp: cannot create {regular file,directory} '...' Permission denied` About once per top level file in git2go.
<NickHu_>
I have tried issuing nix-channel --update as both my user and as root
<NickHu_>
Got confused because in master, pandoc really is at version 2.1.2
<NickHu_>
I see that
<clever>
NickHu_: the branches on this repo coresponde to the channels
<clever>
NickHu_: and all of the master based channels are 2 days or older, QT broke things
vaninwagen has quit [Ping timeout: 256 seconds]
<NickHu_>
I see
nD5Xjz has quit [Ping timeout: 264 seconds]
<boomshroom>
When build command is run, where is the pwd relative to the (copied) source directory.
<NickHu_>
Was it a big problem? I'm interested in how Qt broke things
tos9 has joined #nixos
<clever>
boomshroom: you can just insert a pwd and ls -ltrh into there to find that
<clever>
boomshroom: it should be a temp dir under /tmp, which nix will nuke when its done
<tos9>
Hi (first time using nix in quite a long while). If I want to have some perl modules available for irssi (for plugins), does that mean I have to override its definition
<clever>
tos9: yeah, thats how i would handle things
<tos9>
It appears to not be able to see them if I just `nix-env -i` them (which I guess makes sense)
<boomshroom>
I ask because it can't find the directory jiri wants to build under.
<clever>
tos9: wrapProgram or makeWrapper can be used to modify the PERLPATH when starting irssi
<tos9>
clever: Cool -- any chance yo... aha, ok, will look for an example of those
justbeingglad has joined #nixos
<boomshroom>
Ah, I need to copy the source directory to the build directory.
<clever>
boomshroom: the unpackPhase handles that for you, but the buildCommand disables all of that
fendor has quit [Ping timeout: 246 seconds]
srdqty has quit [Ping timeout: 265 seconds]
<boomshroom>
`python: command not found` You didn't say I needed python! You specifically asked for cmake, go, ninja, and git.
srdqty has joined #nixos
<clever>
and thats why nix exists, to find the un-documented deps!
<clever>
and then clearly document them in the nix expression
nix-noob has joined #nixos
<[Leary]>
I've got pkgs.rxvt_unicode in my environment.systemPackages and services.urxvtd.enable = true. ps -e | grep urxvtd shows a running .urxvtd-wrapped instance, yet urxvtc fails with "unable to connect to the rxvt-unicode daemon: Connection refused".
<[Leary]>
urxvtc starts working if I manually start a second urxvtd instance from terminal, but that really shouldn't be necessary. Is there something else I need to do in my config?
hyperbolic_ has joined #nixos
<MichaelRaskin>
Who is the owner of urxvtd instance?
<hyperbolic_>
what's this "signed-binary-caches" warning I keep getting, and how do I get rid of it?
<[Leary]>
My non-root user in both cases, MichaelRaskin.
<boomshroom>
Does nix-shell set the $src variable?
<clever>
boomshroom: if src is set on the derivation, yes
<clever>
hyperbolic_: are you on nixos?
<boomshroom>
derp, I was running the wrong derivation.
<ottidmes>
srid: If you can configure i3 to look for its config in /etc then you could, or if you use something like home-manager, or you could even just write your own systemd service / activation script linking your /nix/store located config to the needed location, so there are plenty of ways
muzzy has joined #nixos
newhoggy has quit [Remote host closed the connection]
zzamboni has quit [Quit: Leaving.]
newhoggy has joined #nixos
<srid>
Interesting, I didn't know about `environment.etc`
roberth has quit [Ping timeout: 264 seconds]
zzamboni has joined #nixos
newhoggy has quit [Ping timeout: 256 seconds]
loonquawl has quit [Quit: loonquawl]
<{^_^}>
[nixpkgs] @vcunat pushed to master « projectlibre: meta fixes »: https://git.io/vx3VL
reinzelmann has quit [Quit: Leaving]
jensens has quit [Ping timeout: 240 seconds]
blankhart has joined #nixos
<kandinski>
I just left my computer installing virtualbox overnight (I wanted the Oracle extensions), and it seems it worked but I don't have a virtualbox binary, though the VBox* set of binaries are there. How can I access the virtualbox GUI?
<the-kenny>
kandinski: VirtualBox ist the name of the binary
<kandinski>
of course, thanks
blankhart has quit [Ping timeout: 256 seconds]
newhoggy has joined #nixos
clefru has quit [Quit: Leaving]
paraseba_ has quit [Quit: EliteBNC - http://elitebnc.org (Auto-Removal: idle account/not being used)]
nD5Xjz has joined #nixos
blankhart has joined #nixos
zzamboni has quit [Quit: Leaving.]
sigmundv has joined #nixos
dan_b has joined #nixos
nD5Xjz has quit [Ping timeout: 260 seconds]
fendor has joined #nixos
fendor has quit [Client Quit]
<{^_^}>
[nixpkgs] @LnL7 pushed to master « guile-lint: disable tests on darwin »: https://git.io/vx3wh
fendor has joined #nixos
<{^_^}>
[nixpkgs] @LnL7 pushed to release-18.03 « guile-lint: disable tests on darwin »: https://git.io/vx3rt
<fendor>
hi, when using stack with the `--nix` flag, stack downloads some indices and fails with a rather non-descriptive error msg, any ideas why?
<infinisil>
fendor: can't help much without the error
<tos9>
I'm trying to install a package which exists on nixos but which appears not to when using nix on darwin -- how do I see why that is?
<infinisil>
tos9: Packages meta.platforms attribute usually does that
<fyuuri>
When I install google-chrome using nix-env from nixpkgs it does not work (there is no error message, the command just returns immediately). When I install it from nixos using the declarative way it work. Does anybody have the same problem?
<ottidmes>
fyuuri: Nothing in the journal as well?
<pcarrier>
what's the closest to a manpage experience for nixos config? everything has types and descriptions and the online manual is fine but I'd love a CLI to explore, grep, etc.
<Guanin>
Thanks, ottidmes :)
<pcarrier>
I remember seeing some CLI tools but I can't find anything anymore
<dhess>
How do I get nix-build to fetch from a private GitHub repo? I've created an SSH keypair for this purpose, added the public one to my GitHub account, and then put the private one in ~root/.ssh/id_ed25519 on the host where I'm running nix-build, but I'm getting "Host key verification failed" when nix-build tries to fetch the URL
<dhess>
My Hydra can fetch private GitHub repos using a similar setup just fine
mguex has left #nixos ["WeeChat 2.0"]
nD5Xjz has quit [Ping timeout: 240 seconds]
<symphorien>
pcarrier: man configuration.nix
<ottidmes>
Guanin: I double checked and indeed, they talk about fstab options, which is what fileSystems.<name>.options are, which in turn are dynamically converted to mount units by systemd
<Guanin>
The service file looks good, too - So I'll notice the next time my server will be rebooted :)
<pcarrier>
@symphorien do you know if that's functionaly derived from the checkout I nixos-install switch to?
<pcarrier>
OK looks like there's some docbook I'd need to track
<ottidmes>
dhess: Someone (I guess avnik given the username) posted this on the IRC some time ago, I haven't had the need yet so I haven't tried it for myself, but it might help you out: https://gist.github.com/avnik/ef7dd2f420733834abd026d3dd70db7f
<[Leary]>
Does anyone have a usage example for nix-prefetch-git? Can't find any documentation for it. I'm using fetchgit in packageOverrides to pull source for something from a repo, but it's a pain to have to grab the hash from the error every time.
tg has quit [Excess Flood]
mk-fg has joined #nixos
mk-fg has quit [Changing host]
mk-fg has joined #nixos
<dhess>
oh yay! using an ssh://git@github.com/... URL worked!
<dhess>
that is sweet
tg has joined #nixos
acarrico has quit [Ping timeout: 256 seconds]
<dhess>
one drawback is that builtins.fetchGit does not take a sha256 argument.