<Acou_Bass>
hey folks, im installing nixos 17.09 from the iso, and its telling me its refusing to evaluate an insecure samba version - ive added to the allowed insecure packages list but im wondering why its trying to use an insecure samba> is there a not-insecure version it should be pulling?
orivej has joined #nixos
jtojnar has quit [(Remote host closed the connection)]
fendor has quit [(Read error: Connection reset by peer)]
jtojnar has quit [(Remote host closed the connection)]
jtojnar has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] grahamc pushed 2 new commits to master: https://git.io/vFxF6
<NixOS_GitHub>
nixpkgs/master 1200f9b Bastian Köcher: Adds missing documentation for copyFileFromHost in NixOS tests
<NixOS_GitHub>
nixpkgs/master 0726cd6 Graham Christensen: Merge pull request #31972 from bkchr/copyfilefromhost_missing_documentation...
NixOS_GitHub has left #nixos []
jtojnar has quit [(Remote host closed the connection)]
jtojnar has joined #nixos
marusich has quit [(Quit: Leaving)]
ssmike has quit [(Ping timeout: 258 seconds)]
acarrico has joined #nixos
leat has quit [(Ping timeout: 240 seconds)]
fragamus has joined #nixos
jtojnar has quit [(Remote host closed the connection)]
jtojnar has joined #nixos
jtojnar has quit [(Remote host closed the connection)]
ris has quit [(Ping timeout: 255 seconds)]
jtojnar has joined #nixos
Wizek has quit [(Quit: Connection closed for inactivity)]
Cheng has joined #nixos
thc202 has quit [(Ping timeout: 268 seconds)]
leat has joined #nixos
deepfire has quit [(Ping timeout: 260 seconds)]
<pcarrier>
gah I don't understand how the imports stuff works at all
zraexy has quit [(Ping timeout: 268 seconds)]
<clever>
pcarrier: what dont you understand about it?
<pcarrier>
well for starters where is that <foo> syntax defined?
<clever>
that searches for 'foo' within NIX_PATH
<clever>
just like #include <stdint.h> searches in the -I paths of gcc
<clever>
nix also has a -I that prepends things to $NIX_PATH
<pcarrier>
and in `import <nixpkgs> {}` I'm calling the resulting expression on an empty value?
<clever>
yeah
<clever>
your importing the path to nixpkgs, and calling that function with an empty set, which leaves everything at the defaults
<clever>
but that user.nix you linked is actually a config.nix file
<clever>
so you want import <nixpkgs> { config = import ./user.nix; }
<pcarrier>
and it's a "config.nix" in that it expects to be passed as config okay dokay
<pcarrier>
to nixpkgs.
<clever>
if you dont pass a config argument to nixpkgs, it will internally import ~/.config/nixpkgs/config.nix to get the config
<pcarrier>
is there some kind of tracing debugging for expression evaluation? I'm not 100% sure what I'm asking for but it'd be useful to understand what goes on when I evaluate expressions
<pcarrier>
given it's lazy, it shouldn't be insanely large should it?
<clever>
nix-repl> builtins.trace "printme" "return me"
<clever>
trace: printme
<clever>
"return me"
<clever>
pcarrier: this function will print the 1st argument when evaluated, then return the 2nd argument
<pcarrier>
I was more hoping to understand the general path in nixpkgs, but I guess I can traverse by hand
jb55 has joined #nixos
<clever>
pcarrier: to start with, install nix-repl and then try to eval just <nixpkgs> alone
<pcarrier>
And my recent builds don't even boot into anything anymore. How do I troubleshoot systemd hanging in the boot sequence if I don't have a shell?
<pcarrier>
Rebooting into a working config and looking at journald -b -1 doesn't tell me anything I'm afraid
athan has quit [(Ping timeout: 240 seconds)]
<pcarrier>
wow got it. broken filesystem, it went into maintenance eventually
fiveht has joined #nixos
fiveht has quit [(Client Quit)]
<gchristensen>
hmm when did nixpkgs get to 2,000 stars
<gchristensen>
it felt like just yesterday we were at 1,000 (2016-03-29)
lambdamu_ has joined #nixos
<samueldr>
hmm, yesterday we were in 2017, just beginning, not in early 2016.
<samueldr>
both graphs are pull requests, green is open, red+black is closed, orange is the difference
<samueldr>
first is log scale, second is linear
<rcshm>
hi i am trying to configure nix vim plugin. i followed this blog https://beyermatthias.de/blog/2015/11/25/how-to-setup-neovim-on-nixos/. I do have to pass in params like Plug 'junegunn/fzf', { 'dir': '~/.fzf', 'do': './install --all' }. how that is done in nix vim setup? appreciate any help.
<samueldr>
if anyone's better at using the data: stuff.samueldr.com/nixos/nixos-pull-requests-2017-11-23.gnumeric
<pcarrier>
@adisbladis yeah I mean my point is that it doesn't pass through nix-shell.
<pcarrier>
:)
<clever>
pcarrier: nix-shell does nothing to actually block the agent, just re-set that env variable with export
<clever>
pcarrier: and i believe buildFHSUserEnv is to blame
<pcarrier>
can I declare it as something to passthrough?
<clever>
buildFHSUserEnv is generally the wrong answer to most problems
hiratara has joined #nixos
<adisbladis>
pcarrier: I was just about to write the same as clever :>
<pcarrier>
my problem here is that I have a system that relies on being a boring OSX or Linux system
<pcarrier>
and I'm not about to change how our tools compile go binaries or invoke shell scripts so they can fit into nixos, that's not a reasonable expectation from my team to maintain those invariants
<clever>
it will need to be patched
<pcarrier>
if it needs to be patched I can't realistically use nix
<pcarrier>
changing #!/bin/bash to #!/usr/bin/env bash is one thing
<pcarrier>
changing how we build go binaries we run on our production systems is another.
<pcarrier>
AFAICT all I need at this point is a /lib64/ld-linux-x86-64.so.2 and commands like `which gcc` to return something
<pcarrier>
I might try to migrate everything to nix down the road, because quite frankly it would do a much better job than our hack
<pcarrier>
but I need to find myself productive with it for a while at the very least before I suggest we replace everything with it :)
<pcarrier>
this is definitely a case where I want to use buildFHSUserEnv because this project expects something akin to FHS
<pcarrier>
is there a less impactful way to put a few things in a chroot?
<MichaelRaskin>
Well, buildFHSUserEnv is not that high-impact.
acarrico has quit [(Ping timeout: 248 seconds)]
<pcarrier>
it seems to hardcode a bunch of mountpoints. hmmm.
<vaibhavsagar>
what happens when you do `nix-shell -p rust`?
<Ralith>
personally I usually use the mozilla overlay
<Ralith>
not at my computer right now regardless
<jeaye>
vaibhavsagar: error: cannot coerce a set to a string, at /nix/store/zb1lzdmjsk0j9289dap9z8w9inpd2x71-nixos-17.09.2138.48bf1f43cd/nixos/pkgs/stdenv/generic/make-derivation.nix:98:1
<adisbladis>
vaibhavsagar: error: cannot coerce a set to a string ...
<vaibhavsagar>
exactly
<vaibhavsagar>
at least it's not just me
<adisbladis>
As expected. But "nix-shell -p rustc cargo" works :)
<adisbladis>
I think the docs might just be outdated
<vaibhavsagar>
yup, I submitted a PR to correct this just now
<jeaye>
Good man.
jb55 has quit [(Ping timeout: 240 seconds)]
<adisbladis>
vaibhavsagar: Gave you a small feedback on your PR
rcshm has quit [(Ping timeout: 248 seconds)]
ChongLi_ has joined #nixos
miko__ has quit [(Ping timeout: 258 seconds)]
ChongLi has quit [(Read error: Connection reset by peer)]
jb55 has joined #nixos
erictapen has joined #nixos
ma27 has joined #nixos
rtjure has joined #nixos
erictapen has quit [(Read error: Connection reset by peer)]
erictapen has joined #nixos
justan0theruser has quit [(Ping timeout: 240 seconds)]
<jluttine>
building of teamname python package fails
ma27 has quit [(Ping timeout: 246 seconds)]
jiffle has quit [(Quit: leaving)]
<jluttine>
i don't know if there is a better way to create a simple python module in configuration.nix
<fearlessKim[m]>
you src for teamname looks wrong
<jluttine>
now i wrote the module file, a setup.py file and then a python package derivation
<fearlessKim[m]>
it s not a folder
<fearlessKim[m]>
well it doesn4t need to be afolder but having the content of the setup.py won't work, it needs to be the parent folder of the setup.py I would say
<jluttine>
ah, it's a full path to the text file itself? how can i get the parent directory from a path to a file
reinzelmann has joined #nixos
reinzelmann has quit [(Client Quit)]
<hyper_ch>
oh, there are python packaging gurus awake?
<adisbladis>
vaibhavsagar: "Update xyz" doesnt tell me anything about why something was updated. I'll happily accept something like "rust is an attribute set, not a package" or "rustc and cargo are separate packages".
<adisbladis>
hyper_ch: What do you need help with
<vaibhavsagar>
adisbladis: there is a description in the commit message
reinzelmann has joined #nixos
<adisbladis>
vaibhavsagar: Ahh sorry. My bad.
<vaibhavsagar>
I've updated it now
<hyper_ch>
adisbladis: well, I'm trying to package OCRmyPDF. it requires several modules, 2 of them aren't packages yet. One of the modules I could package - ruffus. However the second module, img2pdf requires another module - pdfrw. And I can't get pdfrw to pack properly - https://gist.github.com/sjau/d541329af380b543d39061e432e3f1bc
<adisbladis>
vaibhavsagar: I missed out on some lines. The last message was actually fine. Sorry about that
<vaibhavsagar>
no worries :)
<hyper_ch>
adisbladis: I think I found the problem... it seems to want to use another module...
reinzelmann has quit [(Quit: Leaving)]
<avn>
hyper_ch: I seen pdfrw couple weeks ago on github, and fetchFromGitHub works well for buildPythonLibrary
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] adisbladis pushed 2 new commits to master: https://git.io/vFpIs
<NixOS_GitHub>
nixpkgs/master 1349174 Vaibhav Sagar: doc: rust is an attribute set, not a package...
<NixOS_GitHub>
nixpkgs/master 162c92d adisbladis: Merge pull request #31998 from vaibhavsagar/patch-1...
NixOS_GitHub has left #nixos []
reinzelmann has joined #nixos
<hyper_ch>
avn: well, I just haven't been able to package it yet
<hyper_ch>
avn: also, OCRmyPDF works really nice... I tested it i a debian vm....
<hyper_ch>
also, fetchPypi seems to be better for python modules
<avn>
hyper_ch: yes, but sometimes we neeed more recent versions ;)
<avn>
hyper_ch: I try to help bit later, when I come back
<hyper_ch>
avn: maybe I have figured it out by then - but I'm not too optimistic
<hyper_ch>
more recent versions? oO
<hyper_ch>
also, why aren't any samsung 960 pro 2TB on 50% sale today.... I was really hoping for that....
<orivej>
fearlessKim[m]: nativeDrv and crossDrv are automatically added to every derivation
<jluttine>
fearlessKim[m]: solved my problem: writeTextFile takes an attribute set as an argument, i should've used writeText which takes name and content as arguments
<fearlessKim[m]>
jluttine: arf
<rsa>
been re-compiling everything x for the last three days, when continuing today i forgot to plug in power adapter to the laptop... >.<
<rsa>
also can no longer 'nixos-rebuild switch' due to things not being found in the nix search path for some reason
<fearlessKim[m]>
orivej: thanks. I am trying to allow for more flexible kernel compilation. I think I got it now but understanding the manual-config/generic.nix was not fun. Looks a bit convoluted
<fearlessKim[m]>
rsa: any reason for recompiling X ? you can set an alternative path with rebuild -I nixpkgs=...
deepfire has joined #nixos
Tucky has joined #nixos
<rsa>
custom keyboard layout to xkeyboardconfig (xkb/symbols)
<rsa>
i would like to use the channel configured with nixos-channel rather than setting the path explicitly
<fearlessKim[m]>
orivej: mind explaining the 'configfile = configfile.nativeDrv or configfile ' and why it's not inherit configfile please ? I can't find what the nativeDrv does. I suppose it's the derivation before any override but I don't see the point there ?
Isorkin has joined #nixos
<orivej>
fearlessKim[m]: I'm not familiar with the crosscompilation of the kernel, but during the native compilation pkg.nativeDrv = pkg
<Isorkin>
Hi. How to configure zsh plugin - antigen - http://pastebin.ru/Y6RmN6fx - not normal work config. Error - -antigen-cache-generate:zcompile:54: can't write zwc file: /nix/store/...-antigen-2.2.1/share/antigen/antigen.zsh.zwc
<fearlessKim[m]>
orivej: thanks
<rsa>
even if i set nixpkgs explicitly there are other things missing, e.g.
<rsa>
error: file ‘nixos-config’ was not found in the Nix search path
<rsa>
something broke when the power was cut
<orivej>
rsa: could you try "nixos-rebuild switch --rollback"? or booting an older generation of NixOS from the boot menu?
ChongLi has joined #nixos
<pcarrier>
so it might sound weird but I'm thinking of using nix for the whole repo. is there a way to use directories as sources and have nix efficiently hash them to know whether they changed as inputs or not?
<rsa>
yup
Wizek has joined #nixos
<rsa>
same error
ChongLi_ has quit [(Read error: Connection reset by peer)]
<rsa>
i tried both nix-env -p $mysystemprofile --switch-generation and rebooting
<rsa>
(selecting the same previous profile in the grub menu)
mrkgnao has joined #nixos
vcunat has joined #nixos
vcunat has quit [(Client Quit)]
freusque has joined #nixos
<manveru>
rsa: sounds like your NIX_PATH for root is broken
<manveru>
it should look like `nixpkgs=/nix/var/nix/profiles/per-user/root/channels/nixos/nixpkgs:nixos-config=/etc/nixos/configuration.nix:/nix/var/nix/profiles/per-user/root/channels`
<rsa>
thanks
asuryawanshi has joined #nixos
roberth has quit [(Ping timeout: 260 seconds)]
bkchr has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 2 new commits to master: https://git.io/vFpqb
<NixOS_GitHub>
nixpkgs/master 05761e9 Mathijs Kwik: firewall: fix rpfilter blocking dhcp offers when no ip was bound yet
<NixOS_GitHub>
nixpkgs/master a59864c Franz Pletz: Merge pull request #31839 from bluescreen303/fix-rpfilter...
NixOS_GitHub has left #nixos []
<bkchr>
Hi, if I have a project that is included as a subproject and I want to apply a patch to this subproject, can I somehow tell nix to apply a patch to the sub project?
deepfire has quit [(Ping timeout: 255 seconds)]
toppler has quit [(Remote host closed the connection)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 1 new commit to release-17.09: https://git.io/vFpm3
<NixOS_GitHub>
nixpkgs/release-17.09 87c9b01 Mathijs Kwik: firewall: fix rpfilter blocking dhcp offers when no ip was bound yet...
<NixOS_GitHub>
nixpkgs/master 3ad0845 Franz Pletz: Merge pull request #31371 from gnidorah/sway...
NixOS_GitHub has left #nixos []
asuryawanshi has quit [(Remote host closed the connection)]
davidak has joined #nixos
chrishill has quit [(Remote host closed the connection)]
davidak has quit [(Client Quit)]
asuryawanshi has joined #nixos
<stranger__>
fpletz: thank you!
mrkgnao has quit [(Ping timeout: 240 seconds)]
sigmundv has joined #nixos
<petsa>
New with Nixos. Trying to install dropbox. The version under 17.09 channel does not work. In master there is supposedly a fix, but i was not able to install that.
<petsa>
Ran this: nix-env -if ~/nixpkgs/pkgs/applications/networking/dropbox/default.nix
Sigma has quit [(Remote host closed the connection)]
<petsa>
got this: error: cannot auto-call a function that has an argument without a default value (‘stdenv’)
drever has joined #nixos
roberth has joined #nixos
Lisanna has quit [(Ping timeout: 260 seconds)]
Sigma has joined #nixos
<fpletz>
petsa: try the following: nix-env -f ~/nixpkgs -iA dropbox
davidak has joined #nixos
mahalel_ has joined #nixos
<petsa>
installing, looks promising
thc202 has joined #nixos
roberth has quit [(Remote host closed the connection)]
asuryawanshi has quit [(Remote host closed the connection)]
asuryawanshi has joined #nixos
orbekk has quit [(Quit: WeeChat 1.9.1)]
Wizek_ has joined #nixos
<tommyangelo>
is there any recommended practice for managing backups for multiple nixops hosts databases? I'm building a network with several instances of the same webapp, and would like to automate offhost mysql (well, mariadb) backups.
drever has quit [(Quit: leaving)]
drever has joined #nixos
asuryawanshi has quit [(Remote host closed the connection)]
<srhb>
None specific to Nix that I know of.
dan_b has joined #nixos
<joko>
Do environment.sessionVariables or .variables affect systemd services or should I rewrite the target services?
<LnL>
no, services don't have any environment variables by default
<joko>
ok, I thought so...
<joko>
So if I have to add an env var to a service, I should copy the target NixOS service module and edit it, right?
<joko>
Only package modules are overridable
<joko>
?
chrishill has joined #nixos
<chrishill>
Hey, does anyone know of an alternative to Unetbootin in Nixpkgs? I have searched for “usb live” on the online search but only Unetbootin was the result.
<chrishill>
Unetbootin seems very unreliable and I am unable to get a Clonezilla live USB to work with it.
<chrishill>
Dezgeg: dd didn’t work, some missing binary or something. Maybe the Clonezilla ISO can’t be dded or something
<hyper_ch>
makefu: hmmm, let me try
townsend has quit [(Ping timeout: 248 seconds)]
<makefu>
hyper_ch: however, due to the fact that you only need these pdfs for testing i'd rather not copy them to $out but into the build directory
<hyper_ch>
makefu: I'm open to suggestions... I'm trying to get OCRmyPDF to run
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
erictapen has joined #nixos
<srhb>
chrishill: What's the problem with dd?
<chrishill>
srhb: I tried `sudo dd if=/home/kristoffer/clonezilla-live-2.4.2-61-amd64.iso of=/dev/sdc1 bs=2048 conv=sync,notrunc` but I couldn’t boot into that
<srhb>
chrishill: Did you mean to of=/dev/sdc ?
<chrishill>
The Clonezilla has some instructions on using the zip file so I’m trying that instead ☺
<chrishill>
Oh, it’s not on the partition but directly like that? Huh ☺
<srhb>
chrishill: Yeah.
<srhb>
(Note that it'll nuke the partition table etc., but that should work)
<chrishill>
srhb: Thanks a lot ☺ easy for me to miss such things
<srhb>
Sure :)
chrishill has quit [(Quit: Leaving)]
ssmike has quit [(Remote host closed the connection)]
ssmike has joined #nixos
pie_ has joined #nixos
iyzsong has joined #nixos
Lisanna has joined #nixos
<Lisanna>
uh oh, I installed nautilus (only nautilus, running it inside dwm), and discovered that the MIME types are all messed up - it's showing all my picutures as text files. Anyone know a way to have correct MIME types for files without having to install a whole desktop manager?
<pcarrier>
when I run `nixos-rebuild switch`, what passes pkgs to my configuration expression?
<pcarrier>
to configuration.nix I mean
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] shlevy force-pushed fetchGit-fast-revision-update from cfbb3e7 to eedbc4e: https://git.io/vFp4o
<NixOS_GitHub>
nix/fetchGit-fast-revision-update eedbc4e Shea Levy: fetchGit: Ignore tarballTtl if rev is set and not in the repo....
NixOS_GitHub has left #nixos []
<Isorkin>
How to configure zsh plugin - antigen - http://pastebin.ru/Y6RmN6fx - not normal work config. Error - -antigen-cache-generate:zcompile:54: can't write zwc file: /nix/store/...-antigen-2.2.1/share/antigen/antigen.zsh.zwc
<hedning[m]>
pcarrier: it will use `<nixpkgs>` by default, which you can override with `-I nixpkgs=some/path` (if I understood the question correctly)
hoverbear has quit [(Changing host)]
hoverbear has joined #nixos
hoverbear has quit [(Changing host)]
hoverbear has joined #nixos
<fpletz>
pcarrier: that is coming from the NIX_PATH environment variable
<pcarrier>
OK, but what's "it"?
<pcarrier>
I mean something chose that `<nixpkgs>` would be imported and passed as an argument to that function
<pcarrier>
I'm just wondering where that stuff lives
pie_ has quit [(Ping timeout: 248 seconds)]
<pcarrier>
the code that made that decision I mean
<pcarrier>
if that's not abstracted at all that's fine, I'm just curious.
<pcarrier>
also trying to pass my own pkgs that can pass through to nixpkgs
<srhb>
pcarrier: nixos-rebuild itself. :)
<srhb>
pcarrier: less `which nixos-rebuild`
<hedning[m]>
pcarrier: you can do `nix eval '(<nixpkgs>)' to see the path `<nixpkgs>` resolve to
<pcarrier>
hunky dory.
<pcarrier>
so I guess I could just never use this `pkgs` and everything would work fine right?
mg_ has quit [(Ping timeout: 264 seconds)]
<srhb>
pcarrier: The argument?
<sphalerite>
Any shell wizards here who know if it's possible to connect two programs in a sort of loop, i.e. stdin<>stdout in both directions?
<pcarrier>
srhb: yeah I'd rather have stuff go through "my repo"
<pcarrier>
if that makes any sense
<srhb>
pcarrier: but <nixpkgs> is shorthand for exactly that, whatever nixpkgs is set to in NIX_PATH
<pcarrier>
yeah no sure exactly
Lisanna has quit [(Remote host closed the connection)]
<sphalerite>
err wait no, I don't think you need the callPackage bit
<sphalerite>
yay!
<pcarrier>
well systemd existed but wasn't in unstable
<pcarrier>
coming back, everything is just sooo exciting it's crazy
<pcarrier>
anybody started replacing gradle with nix for modular JVM codebases by any chance? :D
<pcarrier>
I want this awesomesauce on everything.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 1 new commit to master: https://git.io/vFpBP
<NixOS_GitHub>
nixpkgs/master 04a462b Orivej Desh: zcash: does not build on darwin
NixOS_GitHub has left #nixos []
<pcarrier>
this does a ridiculous lot of the stuff I've been wanting to do muuuch better
<pcarrier>
can I run nix environments in Linux xhyve VMs from nix on darwin?
<sphalerite>
yeah, that's the feeling I got as well. SSHing into an old debian server that i'm still in the process of replacing always makes me sad
<sphalerite>
It's certainly possible in theory but in practice I think a lot of tooling you'd need for it doesn't exist. Particularly because you'd need the ability to build for linux on OSX
<pcarrier>
I replaced my 12-core, 64-GB-of-RAM workstation running a reasonably maintained and tweaked Ubuntu desktop because I started running a fine NixOS setup on my laptop. I'm gonna want to wipe it to move to nix+zfs soon.
taktoa has joined #nixos
<srhb>
12 core o_o
<sphalerite>
♥
<pcarrier>
@sphalerite well the point is that I wouldn't if I have nix. I should be able to get cached artifacts for everything, or even dispatch the build to a Linux server over SSH if needed
ThatDocsLady_ has quit [(Ping timeout: 240 seconds)]
<pcarrier>
I'm happy to provision SSH keys on build workers from nix files living in the repo as a way to onboard developers. we're a small shop.
<pcarrier>
but we have decent automation on top of kubernetes.
<pcarrier>
so provisioning Linux build workers accessible by all laptops is _not_ a problem.
<hyper_ch>
makefu: ok, got the pdfs now but next problem has arrived
<pcarrier>
I do have to support a majority of Mac developers though, and I am perfectly content as long as they are
<hyper_ch>
pcarrier: why zfs?
<pcarrier>
hyper_ch: yeah so after years of pain with laptops and workstations and [...] I'm investing in having a proper system now
<hyper_ch>
how does that relate to zfs?
<pcarrier>
so nix recipes for all the software, zfs for all the data, tons of snapshots and remote backups over ssh
<hyper_ch>
I see
<hyper_ch>
zfs claims one more vicim - yey :)
<pcarrier>
worst case scenario I get a new laptop and restore the full system from the nixos livecd after adding zfs boot support to the config and switching
<pcarrier>
hoping to generate my own rescue disk from nixos, which seems like the next logical step
<pcarrier>
should be easy to tweak the ISO config :D
<hyper_ch>
making own nixos iso is simple
erictapen has quit [(Ping timeout: 248 seconds)]
<pcarrier>
yeah, I'm investing in my mental health here.
<pcarrier>
I will not suffer software and hardware bullshit much longer now.
<pcarrier>
so can I declare the overlays in <nixos-config> somehow, or do I need a new -I in the "bootstrap" nixos-rebuild then I set it through nix.nixPath?
<pcarrier>
set <nixpkgs-overlays> that is
<pcarrier>
I do want to maintain them in the repo
<srhb>
pcarrier: nixpkgs.overlays is a list of overlays
<srhb>
pcarrier: (used in the system profile)
<pcarrier>
LOL this is too awesome
stranger__ has quit [(Quit: Connection closed for inactivity)]
<makefu>
then cd tests; /usr/bin/env PYTHONPATH=. py.test
periklis has joined #nixos
<hyper_ch>
makefu: if I copy the static_pdfs to test/static_pdfs then I get the same error as when I don't download them at all
<hyper_ch>
if I copy the static_pdfs just to ./static_pdfs then I get the error from the pastebin
<hyper_ch>
no idea which one is correct
<hyper_ch>
can't you skip all that testing?
<makefu>
yes you could but wouldn't it be so much cooler if it would work? :D
<hyper_ch>
no
<vaibhavsagar>
is GHC 8.2.2 going to make it into nixos-17.09?
<hyper_ch>
as I can't figure out where the problem is
<gchristensen>
vaibhavsagar: what is 17.09 at now?
<vaibhavsagar>
very far behind
<vaibhavsagar>
GHC 8.2.1
<gchristensen>
~maybe?
<vaibhavsagar>
but GHC 8.2.2 was only released a couple of days ago
<vaibhavsagar>
I tried to do a cherry-pick of all the commits that touched the pkgs/development/haskell-modules directory but I couldn't get it to work in the presence of merge commits
<gchristensen>
though it'd be a hoot to see what ORCmyPDF would do
deepfire has joined #nixos
<snajpa>
guys, how do I find what's creating entries in /var/lib/libvirt? I want to mount ZFS dataset there, but can't, something always populates it well before even libvirtd service starts
<gchristensen>
you should be able to mount over it even if there is something there
<snajpa>
but for the sake of what I've been trying to solve - how to dig for what's creating the entries in that dir before zfs mounts there?
<clever>
snajpa: what are the files called?
<hyper_ch>
"As of 2014-03-04, you should set the mountpoint property of your ZFS filesystems to be legacy and let NixOS mount them like any other filesystem (such as ext4 or btrfs), otherwise some filesystems may fail to mount due to ordering issues"
<snajpa>
hyper_ch: rpool yes, dpool not
<snajpa>
'dpool' is my 2 disk 'large data pool'
<hyper_ch>
snajpa: so just set it to legacy and put it into your hardware-configuration.nix
<snajpa>
I'm going to try legacy mount for /var/lib/libvirt
<snajpa>
brb
<hyper_ch>
hmmm, should I use fetchFromGitHub because I get this error: "Make sure you're either building from a fully intact git repository or PyPI tarballs. Most other sources (such as GitHub's tarballs, a git checkout without the .git folder) don't contain the necessary metadata and will not work."
Cheng has joined #nixos
<clever>
hyper_ch: the tarballs probably contain some special files that include the metadata
<clever>
and you could just create those yourself
<hyper_ch>
clever: well, I have no real knowledge about python and I'm just getting that error now... not sure how to use pip either to get things installed
<gchristensen>
anyone want to make a .drv parser in rust? I'll pay $50 :)
<clever>
gchristensen: haskell and node already have them
Cheng has quit [(Client Quit)]
<hyper_ch>
does buildPythonApplication even use pip install?
<clever>
no idea
<gchristensen>
clever: should be easy to follow that example then
<hyper_ch>
it worked for the other dependencya modules
<snajpa>
mission was a success, thx everyone!
<gchristensen>
w00t
<snajpa>
srk: legacy mount...
<gchristensen>
clever: can you link me to those implementations?
<pcarrier>
OK so builtins.trace inserted in .nixpkgs/config.nix does get picked up, but the overlay doesn't get traced
<pcarrier>
despite being imported from there into overlays!?
mahalel_ has quit [(Ping timeout: 255 seconds)]
<infinisil>
Pinging wizzup[m] vaibhavsagar and elvishjerricco for my question
mahalel_ has joined #nixos
ssmike has quit [(Ping timeout: 240 seconds)]
nixos1022 has joined #nixos
bennofs has quit [(Quit: WeeChat 1.9.1)]
<nixos1022>
When building switching to a new homemanager config (which for testing purposes only contains a config for git), I get the following error: Wide character in die at /nix/store/64jc9gd2rkbgdb4yjx3nrgc91bpjj5ky-buildenv.pl line 79.
<nixos1022>
collision between ‘/nix/store/l7zddbrfv6qb607lmfn16bw4iqv5i8y2-git-2.14.1/bin/git’ and ‘/nix/store/ihrjcy5km8vh98ba35v2a82zgaldfcji-home-manager-path/bin/git’; use ‘nix-env --set-flag priority NUMBER PKGNAME’ to change the priority of one of the conflicting packages
<nixos1022>
how can this happen?
mbrgm has quit [(Ping timeout: 264 seconds)]
<MoreTea>
you've tried to install git via two mechanisms
<MoreTea>
probably via nix-env -i git as well
<nixos1022>
yes
<MoreTea>
or via /etc/nixos/configuration.nix
<nixos1022>
but why is this problematic?
<MoreTea>
it's not, but now nix doesn't know which binary to put in your $PATH, basically.
<MoreTea>
I'd remove git installed via nix-env in this case (via nix-env -e git), and manage it via your home manager.
mbrgm has joined #nixos
<elvishjerricco>
infinisil: That probably has something to do with the fact that Stack has to fetch an lts. I’ve only ever used it for blank GHC “resolvers” so that I could just use Nix’s GHC and package set
<nixos1022>
Yes that works. So for all my programs I can have them either installed by nix-env or homemanager? How could I have both and specify which one I want to be in my $PATH?
Cheng has quit [(Remote host closed the connection)]
<nixos1022>
Also thx for your help
asuryawanshi has joined #nixos
<elvishjerricco>
infinisil: I wouldn’t be terribly surprised if you need to add some networking component to the definition of buildStackProject
<infinisil>
(Code is ugly as hell, it's my first Haskell project ever :P)
acarrico has joined #nixos
mbrgm_ has joined #nixos
mbrgm has quit [(Ping timeout: 255 seconds)]
mbrgm_ is now known as mbrgm
mbrgm has quit [(Client Quit)]
mizu_no_oto has quit [(Client Quit)]
mbrgm has joined #nixos
<FRidh>
Building Nix 1.11.15 from a tarball, setting LIBSECCOMP_CFLAGS and LIBSECCOMP_LIBS to find respectively the headers and the libs. Unfortunately, the linker fails. Any clue?
ssmike has joined #nixos
Isorkin has quit [(Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)]
chreekat has quit [(Quit: quitting)]
<infinisil>
vaibhavsagar: (if you happen to try it out, i just committed a version with latest nixos-unstable pinned)
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
erictapen has joined #nixos
<das-g[m]>
> das-g[m]: not unless you add your own channels, this is fine :) can you file a bug though? http://github.com/nixos/nix-pills NixOS/nix-pills - GitHub Contribute to nix-pills development by creating an account on GitHub.
<joko>
Is it possible to run commands in configuration.nix? E.g., I am already setting statically the IPv4 address of an interface and I would like not to set it if it doesn't exist (so "ip link show | blabla" should run and store its results on a list, so that I can do the check in nix)
erictapen has quit [(Ping timeout: 240 seconds)]
<hyper_ch>
I don't quite understand what you try to achieve
goibhniu1 is now known as goibhniu
<clever>
joko: make a systemd unit to run it at the right time
<joko>
hyper_ch: I have something like networking.interfaces = { eth0 = { ip4 = [ { address = "131.211.84.78"; prefixLength = 25; } ]; }; } in my setup, I would like to check if eth0 exists before setting it.
<pcarrier>
hyper_ch: could it possibly explain the crazy I seem to run into?
<pcarrier>
or do I understand nothing?
<hyper_ch>
pcarrier: overlays hate me :)
<pcarrier>
as in, I should give up? :D
cacatoes has joined #nixos
<hyper_ch>
not you
<hyper_ch>
I said they hate me
periklis has quit [(Ping timeout: 240 seconds)]
dywedir has joined #nixos
jb55 has joined #nixos
atractiva-y-cabr has joined #nixos
bfrog has joined #nixos
orbekk has joined #nixos
jb55 has quit [(Ping timeout: 248 seconds)]
freusque has quit [(Quit: WeeChat 1.9.1)]
<pcarrier>
I mean at this point I have no idea how I can start using them. this infinite recursion is such a weird thing to run into, and I have literally no idea what could cause it
<infinisil>
Would be nice if somebody could add the Topic: Haskell label to it ^^
<pierron>
This application failed to start because it could not find or load the Qt platform plugin "xcb", I've got this issue with openshot-qt and kdenlive, anybody got an idea?
ssmike has quit [(Ping timeout: 240 seconds)]
<elvishjerricco>
infinisil: Actually I think I might know the cause of that.
<elvishjerricco>
Let me see if I can find it again...
<goibhniu>
pierron is your channel up to date? There was a problem like this fairly recently.
<pierron>
goibhniu: probably not
<pierron>
I will check
<vaibhavsagar>
how do I create a source-only derivation?
<elvishjerricco>
infinisil: Left a comment pointing at a similar thing. I believe it's just that the environment doesn't have that `/etc/protocols` thing.
<atractiva-y-cabr>
hi
<pierron>
goibhniu: no, I tried with the nixos-unstable channel downloaded directly from github
<infinisil>
elvishjerricco: Hmm, I don't know how to do that in nix though
darlan has quit [(Client Quit)]
atractiva-y-cabr has left #nixos []
<elvishjerricco>
infinisil: Me either. I'm actually surprised it was an error. It's probably because you used the Nix sandboxing. Not sure how to fix it in the sandbox...
<Isorkin>
Please help to build package - http://pastebin.ru/egm8Ijk0 - error - Checking for C library ssl... no - not found
<infinisil>
elvishjerricco: Yeah, turning off sandbox makes it work
<elvishjerricco>
infinisil: Ok. Still, probably needs to be fixed to make it work in the sandbox
<infinisil>
elvishjerricco: Hmm right, and you can't download stuff with non-fixed-output derivations
<pcarrier>
`let noop = import /tmp/what; in [ noop ]` instead of `[ import /tmp/what; ]`
<pcarrier>
I'm not 100% sure I understand why, but yay magic.
<elvishjerricco>
infinisil: I think the right solution is that the stack home needs to be a fixed output derivation, and then stack builds depend on the home. This would also help by caching all of stack's Haskell builds
<sphalerite>
The rust overlay seems to be broken on nixUnstable, but fine on 1.11
chreekat has quit [(Ping timeout: 240 seconds)]
hotfuzz_ has joined #nixos
chreekat has joined #nixos
chreekat has quit [(Client Quit)]
hotfuzz has quit [(Ping timeout: 248 seconds)]
chreekat has joined #nixos
dannyg has joined #nixos
<hyper_ch>
I don't feel like reading code that's all one line
<hyper_ch>
joko: I don't feel like reading code that's all one line
ssmike has joined #nixos
endformationage has joined #nixos
<sphalerite>
pcarrier: [ import /foo ] is a list containing the function import, and the path /foo. You probably meant [ (import foo) ]
<pcarrier>
oh wow OK thanks
chreekat has quit [(Client Quit)]
<gchristensen>
^ a common thing that bites new people
<pcarrier>
yeah that problem makes sense now :)
<sphalerite>
pcarrier: as for the installation command, you probably want -A. -i without -A searches all of nixpkgs by package name, which is less predictable, slower, and more memory-hungry.
<pcarrier>
oooh thank you
<joko>
hyper_ch: sorry, I don't understand, the code has multiple lines via let, right?
jb55 has quit [(Quit: WeeChat 1.9)]
<dannyg>
When I need to run multiple instances of a module, e.g. tomcat, until now I have been rewriting the module definition, making the options explicit arguments to the nix file, and including it as a job with systemd.services.2ndmoduleinstance = import ./converted-module.nix {...}. Is this still a good way to solve the problem of having multiple instances of a module, or is there a more general facility for handling this?
<hyper_ch>
joko: in your paste it's all one line
<pcarrier>
@sphalerite and nix-env -iA foo should look for an attribute foo from my overlay right?
<joko>
hyper_ch: you mean eth0.ip4 = ... ?
<pcarrier>
cause that doesn't seem to work :(
ssmike has quit [(Remote host closed the connection)]
<sphalerite>
pcarrier: I'd recommend nix-env -f '<nixpkgs>' -riA foo, again for consistency. One day we we will have a nix install command which will be much nicer.
ssmike has joined #nixos
<joko>
hyper_ch: I just wanted to focus on the availableInterfaces part in let, not the attribute set, but anyway, let me transform it a bit
<sphalerite>
pcarrier: you could probably also do nix-env -riA nixos.foo, but that's less reliable because it uses ~/.nix-defexpr which may or may not be at its default, and won't work on non-nixos systems with nix.
<pcarrier>
@sphalerite jesus now it works.
<sphalerite>
pcarrier: I personally have nix-env aliased to nix-env -f '<nixpkgs>'
JosW has joined #nixos
<pcarrier>
@sphalerite do you do that by creating a wrapper through an overlay? :)
<sphalerite>
no, I just put it in my zshrc :p
<sphalerite>
(which I manage through nix though!)
<sphalerite>
the nice thing about having it as a shell alias is that zsh then still works out the tab completion correctly
<pcarrier>
can overlays override the config?
<pcarrier>
the nixpkgs config attributes?
<joko>
hyper_ch: is it better now? http://nixpaste.lbr.uno/00ccC-L0?nix I have also added a second interface and my point is that on some setups there is eth0 and eth1 and on some others not. So, on the latter I would like networking.interfaces to be evaluated as { }
<hyper_ch>
joko: still all one line in that paste
<snajpa>
woohoo finally got everything working with my GPUs inside KVM under Libvirt's control
<snajpa>
starting to love nixos, thx guys for making it
<snajpa>
though you didn't want to hear what I had to say during the process, was it a pain... but finally arrived at a pretty nice configuration.nix, which I'm happy with
<joko>
But if anyone else has an idea, please hop on
<sphalerite>
pcarrier: yeah I think so, but I think the config.nix mechanism is generally not well-liked
<sphalerite>
that is, prefer other options (overrides)
<hyper_ch>
joko: one line here
<hyper_ch>
or rather one long line broken up into 4 lines because it's too long
<joko>
hyper_ch: if you are referring to { eth0 = {...}; eth1 = {...}; }, there is absolutely no need to focus there, I just need to fix availableInterfaces if possible
<pcarrier>
@sphalerite for allowUnfree? :)
<sphalerite>
eh, I guess that's a reasonable application for it :P
<pcarrier>
I'm up for alternatives
<sphalerite>
well you can use an env var too, but as a permanent solution I'd say that's much worse :p
<sphalerite>
(although it is very handy for testing)
bfrog has quit [(Ping timeout: 264 seconds)]
acarrico has joined #nixos
<Isorkin>
Please help to build package - http://pastebin.ru/egm8Ijk0 - error - Checking for C library ssl... no - not found
FRidh has quit [(Quit: Konversation terminated!)]
miko__ has quit [(Ping timeout: 264 seconds)]
stphrolland has joined #nixos
<stphrolland>
Hi. I'm still evolving from the first install. 9th configuration for now. I get addict to the workflow of changing configuration this way.
<stphrolland>
I have remarked a little something: my /tmp folder is not emptied between reboots
<stphrolland>
is it normal ?
Sonarpulse has joined #nixos
<stphrolland>
sorry, was lazy: boot.cleanTmpDir found in search nixos option
<hyper_ch>
(or tmpfs for /tmp)
<Orbstheorem>
Hi, I have this "blob" in my configuration.nix: https://paste.gnugen.ch/raw/LgMp As you can see is petty much the same, which differences being the name and the 'subvolume' property. Is there anyway to write this more compactly? or avoid the copy/paste?
<pcarrier>
can I somehow keep the work dir around when a nix derivation I'm working on fails?
<Orbstheorem>
So, back at the time I was using debian I had my rootfs snapshoted every hour or so. Nixos guarantees my system state on every boot, do you see any reason why someone would want to snapshot a rootfs with nixos?
<hyper_ch>
yes
<Orbstheorem>
hyper_ch: why?
<hyper_ch>
quick restoration
<hyper_ch>
logs
<hyper_ch>
because you can
<hyper_ch>
snapshotting is a cool thing to do
<hiroshi>
haha
<hyper_ch>
there's probably even better reasons
<hiroshi>
Orbstheorem: do you use zfs ?
<Orbstheorem>
hiroshi: btrfs
<hiroshi>
Orbstheorem: Ouch.
<hyper_ch>
oh.... btrfs
<hiroshi>
xD
<Orbstheorem>
:P
<hyper_ch>
why?
<hiroshi>
snapshots are cheap and fast in zfs. you can test it in a freeBSD without issue and out of the box
phreedom has quit [(Quit: No Ping reply in 180 seconds.)]
<hiroshi>
it is important that all the OS is in zfs including /boot. nixOs hasn't that but eventualy linux distros will have it
asuryawanshi has joined #nixos
<sphalerite>
hiroshi: in fact I think it is possible to have /boot on zfs in nixos
<Orbstheorem>
I can't give you an argument on why do I use btrfs over zfs. I just remember my friend complaining about succesively destroying his data because of false zsh manipulations
<sphalerite>
I don't think it's possible with EFI though, only legacy (BIOS) boot
<Orbstheorem>
so, I gave btrfs a shoot
<sphalerite>
Orbstheorem: still better than the filesystem destroying the data for you without you telling it to ;)
<hiroshi>
what has that to do with nixOS. well you have snapshots with nixOS but you can have an other layer of verification and 'backups' .
<hyper_ch>
btrfs destroys your data even without false manipulations ;)
<sphalerite>
(never happened to me personally, but I've seen a number of people complain about it)
<Orbstheorem>
idk, I've been on btrfs for about a year, and so far, so good :)
<hiroshi>
sphalerite: do you have gpt with full zfs disk ? [boot included]
phreedom has joined #nixos
<Orbstheorem>
specially since I do btrfs send every morning :)
<sphalerite>
hiroshi: I don't have it myself, but it should be possible. Not with EFI though.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] edolstra pushed 2 new commits to master: https://git.io/vFppf
<NixOS_GitHub>
nix/master 90948a4 Eelco Dolstra: nix-shell/nix-build: Support .drv files again...
<NixOS_GitHub>
nix/master 549c370 Eelco Dolstra: nix path-info: Show URL of NARs in binary caches
NixOS_GitHub has left #nixos []
goibhniu has quit [(Ping timeout: 268 seconds)]
<Orbstheorem>
and also I didn't know zfs send was a thing
<hyper_ch>
and there's also zfs receive ;)
<sphalerite>
hiroshi: AFAIU EFI inherently conflicts with it since you need to have the EFI System Partition which is specified to be FAT.
<sphalerite>
I think GPT with BIOS boot would work? Not sure though
<hiroshi>
bacula is an option [with zfs of course]
<hiroshi>
haha
asuryawanshi has joined #nixos
<hyper_ch>
stphrolland: probably the two most important things in my config are how to have a seperate file with your secrets and how to get the yoyo list directly into your hosts file for adserver blocking :)
<pcarrier>
trying to package minijail0, which is a pretty fun exercise. the produced binary wants to know the path of a build artifact to put that path in LD_PRELOAD
<pcarrier>
kinda smells like I would have to build in 2 phases?
<pcarrier>
as in, have a derivation for the preloaded library, and have another derivation for the binary that contains that path
<stphrolland>
just to know: anyone here has an example of nix configuration file with FLUXBOX as a window manager ? I have my old style fluxbox files, but I wonder how it looks in nix, and if it can be entirely configured with nix.
<pcarrier>
oh no I'm building everything @hyper_ch
<hyper_ch>
pcarrier: can you pastebin it?
<pcarrier>
@hyper_ch the problem is that the binary calls other binaries with a modified LD_PRELOAD, as in, it wants to add its library in there
<hyper_ch>
you need to include those packages so that nix correctly links it...
<sphalerite>
pcarrier: that shouldn't be necessary AFAIU. If you get the library installed into $out/lib/libwhatever.so.bla and get the binary to add that path to LD_PRELOAD I think it should work
<pcarrier>
I mean sure but it's broken because of that
<pcarrier>
to add which path @sphalerite?
<sphalerite>
pcarrier: $out/lib/libwhatever.so
<pcarrier>
oh wait $out is literally gonna be something like /nix/store/4380924732897423-foo-1.23?
<pcarrier>
@sphalerite thank you. so obvious in retrospect :D
<sphalerite>
pcarrier: looks to me like it would work unmodified in fact
<sphalerite>
as long as DESTDIR is set appropriately
<pcarrier>
really? there's no DESTDIR in the code
<sphalerite>
or LIBDIR, rather.
<pcarrier>
yes indeed. I'm on that :)
<sphalerite>
So if you add makeFlags = ["LIBDIR=$out/lib"]; to your mkDerivation call I think it should work. Not 100% sure though, you may need to do it in one of the phases so that the $out gets substituted right
<sphalerite>
It's always a bit confusing to work out which variables are escaped and which aren't
<pcarrier>
now I should try to clean that up and submit it to nixpkgs I guess?
<sphalerite>
Yay! Yes that would be cool!
orivej has quit [(Remote host closed the connection)]
stphrolland has quit [(Quit: Page closed)]
orivej has joined #nixos
ertes has quit [(Quit: Bye!)]
cromachina has joined #nixos
<sphalerite>
Feel free to ping me on the PR, I'm lheckemann on github
<pcarrier>
it's almost 1pm and I still haven't slept, so probably another time :)
<Baughn>
...and no, that's not what this error is saying.
ertes has joined #nixos
FRidh has joined #nixos
<sphalerite>
pcarrier: 1pm seems like a reasonable time to be awake to me. 1am maybe? x)
<pcarrier>
no 1pm running from yesterday.
<sphalerite>
oh.
<sphalerite>
OH.
<pcarrier>
I get excited.
<sphalerite>
How do you manage that o.O
<pcarrier>
oh I don't know how long I'm gonna sleep for
<sphalerite>
I mean, sure, nixos is cool. But I couldn't lose that much sleep getting excited over it x)
<sphalerite>
Well enjoy your sleep
<hiroshi>
haha
<cromachina>
hello all, i am using an old laptop that uses iwlwifi firmware, but whatever the latest is for my device is buggy as hell (from the linux nonfree firmware package). i've seen other complains about this particular firmware version searching around. i would like to test out older versions of the firmware (a .ucode file from intel one usually drops in /lib/firmware), so how can i specify this in my configuration?
civodul has quit [(Quit: ERC (IRC client for Emacs 25.3.1))]
pie_ has quit [(Quit: Leaving)]
orbekk has quit [(Ping timeout: 240 seconds)]
stphrolland has joined #nixos
cromachina has quit [(Ping timeout: 260 seconds)]
<stphrolland>
I need some help about how to find the documentation. I only know how to use Search Nixos Options for now. I want to use fluxbox, the package is already in my nix config, and runs as my desktop manager. The traditionnal fluxbox is using configuration files, located in the user home directory. But I don't know if the fluxbox nixos package allows to configure fluxbox with nix. What should I look at ? What should I search for ? Whic
<samueldr>
(your message stopes at "What should I search for ? Whic")
<samueldr>
stops*
cromachina has joined #nixos
<cromachina>
of course I disconnect immediately after asking my question...
<hyper_ch>
stphrolland: nixos doesn't handle files in user homes...
<stphrolland>
hyper_ch: same for .bashrc and .profile ?
<hyper_ch>
stphrolland: copy over your your old fluxbox config files
dannyg has quit [(Quit: dannyg)]
<hyper_ch>
stphrolland: there's some defaults when a new user is created (probably from /etc/skel or wherever nixos has it) but besides that, nothing is managed in user home
<hyper_ch>
(ok, there's is home-manager which does handle files in user home)
bennofs has joined #nixos
<sphalerite>
stphrolland: yeah there's no built-in dotfile management. You can use home-manager, or wrapper scripts, but your old config should also just work out of the box if you copy it over.
<ertes>
on unstable i just downgraded from firefox to firefox-esr, but it still comes with the new noscript… i'd like to downgrade that one as well… any options without installing it through firefox?
ssmike has quit [(Quit: ssmike)]
sigmundv has quit [(Ping timeout: 248 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] pcarrier opened pull request #32010: minijail: init at android-8.0.0_r34 (master...pcarrier/minijail) https://git.io/vFheA
NixOS_GitHub has left #nixos []
<pcarrier>
@sphalerite here comes :)
<sphalerite>
pcarrier: go sleep!! :p
<pcarrier>
@sphalerite meh, I'm excited :)
jensens has quit [(Ping timeout: 260 seconds)]
<pcarrier>
hmmm. looks like it worked in 17.09 but doesn't in master?
<pcarrier>
nix-shell -A minijail is happy but nix-build -A minijail isn't
ssmike has joined #nixos
orbekk has joined #nixos
<sphalerite>
See my comment on the PR - you need to add it to all-packages.nix
<sphalerite>
and the hash doesn't seem to match the one I get
<sphalerite>
It may be that whatever serves archives on android.googlesource.com isn't deterministic
<sphalerite>
fetchgit may work better.
ma27 has joined #nixos
<pcarrier>
yeah happy to switch :)
tester123 has joined #nixos
<tester123>
How do I upgrade haskell stack to 1.5.1 in NixOS? Right now it says I'm running 1.4.
jb55 has quit [(Ping timeout: 248 seconds)]
<Turion>
Does anyone successfully develop with GHCJS here?
<sphalerite>
Turion: yes. Not sure who though, sorry
stphrolland has quit [(Ping timeout: 260 seconds)]
<Turion>
sphalerite: Haha so I'm assuming it's not you j)
<sphalerite>
indeed. I haven't tried though either :p
<sphalerite>
are you encountering specific issues or is it more of a "has it been done" question?
<Turion>
sphalerite: No, specific issues. cabal says that a dependency ghcjs-base is missing
<Turion>
It seems like this package is not properly installed with ghcjs, or I don't know how to bring it into the scope
<sphalerite>
if you have a specific issue, describe the specific issue right away. Reduces turnaround time significantly! :)
<Turion>
Will keep that in mind for the future ;)
<Turion>
So do I need to start a nix-shell somehow to make cabal find the package?
asuryawanshi has joined #nixos
<hyper_ch>
ok, a step further and now next problem with python....
taktoa has quit [(Remote host closed the connection)]
pie_ has joined #nixos
<sphalerite>
Turion: as far as I can tell, yes. I think what you'd want is to use is something like ghc.withPackages, except for ghcjs. Not quite sure how to do that though.
petsa has joined #nixos
asuryawanshi has quit [(Ping timeout: 240 seconds)]
townsend has quit [(Remote host closed the connection)]
<Turion>
sphalerite: Oh ok! I can't find the attribute that corresponds to ghcjs-base, though :(
<sphalerite>
yeah, I haven't been able to find it either
<pcarrier>
@sphalerite did the switch
<sphalerite>
oh you mean the ghcjs-base bit
<sphalerite>
I don't think you need to, that should just be included, I believe
<Turion>
sphalerite: I started a "nix-shell -p haskell.compilers.ghcjs" and tried "cabal install --only-dependencies" in there, to no avail :( is that what you mean?
<pcarrier>
@sphalerite so the header removal is needed in master but breaks the build in 17.09
<pcarrier>
@sphalerite I'm not 100% sure I want to figure out exactly which glibc upgrade broke it...
JosW has quit [(Quit: Konversation terminated!)]
mbrgm has quit [(Ping timeout: 240 seconds)]
mbrgm has joined #nixos
<sphalerite>
pcarrier: I'd say it's okay if you just remove it rather than delving into the details. I don't think we backport new packages to the stable release so the discrepancy won't appear in nixpkgs itself anyway, only in your overla
<sphalerite>
y
<pcarrier>
I know how to overlay the fix for 17.09 now :)
MichaelRaskin has joined #nixos
<pcarrier>
Though I don't import this channel anyway
Slabity35 has quit [(Remote host closed the connection)]
<sphalerite>
gchristensen: did you give me ofborg access?
<sphalerite>
(besides, eventually, the bot reporting success/failure of a build)
<gchristensen>
oh I didn't give you access
ambro718 has joined #nixos
<sphalerite>
aah ok, that explains it :D
<gchristensen>
gimme a few minutes!
<sphalerite>
no worries. Also I just saw all the PRs with their tags for number of rebuilds. It's wonderful!
<gchristensen>
:D
tester123 has quit [(Ping timeout: 260 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to staging: https://git.io/vFhUq
<NixOS_GitHub>
nixpkgs/staging 7a13e5a Frederik Rietdijk: Merge remote-tracking branch 'upstream/python-unstable' into HEAD
NixOS_GitHub has left #nixos []
<qmm>
hi, i have run "nix-build" and i receive an error: error: attribute ‘applicative-quoters’ missing, at /nix/store/5mjppffgvxpnndcf9nv4jsgjamrwx4jk-61fbdb47a69f78998f55207e64122a0798047b5d.tar.gz-unpacked/pkgs/development/haskell-modules/configuration-ghc-8.0.x.nix:42:37
<qmm>
i have run nix-channel update, but that didn't fix it
<qmm>
specifically, nix-channell --update
<qmm>
without the extra l
<pcarrier>
Are there docs about the rebuild GitHub labels in nixpkgs?
cromachina has quit [(Ping timeout: 260 seconds)]
fendor has joined #nixos
<sphalerite>
pcarrier: it's just the number of packages that need to be rebuilt as a result of the changes
<qmm>
figured it out
<pcarrier>
oooh nice thanks
vcunat has joined #nixos
<hyper_ch>
pcarrier: not testing nixos unstable yet? ;)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to staging: https://git.io/vFhTI
<NixOS_GitHub>
nixpkgs/staging 0f50d5a Frederik Rietdijk: python.pkgs.trollius: use optionals instead of optional, fixes eval for neovim...
NixOS_GitHub has left #nixos []
Neo-- has quit [(Ping timeout: 246 seconds)]
Turion has quit [(Read error: Connection reset by peer)]
<pcarrier>
@hyper_ch not really what I'm looking for
<gchristensen>
sphalerite: ok try that comment again (you can just edit it and resave it)
rosa has joined #nixos
<sphalerite>
cool, thanks
<gchristensen>
Mic92: ping?
<sphalerite>
"only fetch onee jbo at a time" keyboard playing up again? x)
<gchristensen>
ugh
rosa has quit [(Quit: rosa)]
ssmike has quit [(Remote host closed the connection)]
ssmike has joined #nixos
petsa has quit [(Ping timeout: 260 seconds)]
asuryawanshi has joined #nixos
<Mic92>
gchristensen: pong
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to python-unstable: https://git.io/vFhkQ
<NixOS_GitHub>
nixpkgs/python-unstable 44e98de Frederik Rietdijk: python.pkgs: maintenance updates
NixOS_GitHub has left #nixos []
asuryawanshi has quit [(Ping timeout: 258 seconds)]
<pcarrier>
alrighty. but it should be in all-packages not conditioned by anything right?
<gchristensen>
right
<sphalerite>
Putting meta.platforms already takes care of it
<benley>
anyone know how to write nixos modules for packages that ship with their own systemd .service/.socket files?
ssmike has joined #nixos
<benley>
I'm trying to decipher the pulseaudio module to see how it gets /etc/systemd/user/pulseaudio.service to be a symlink to ${pkgs.pulseaudio}/lib/systemd/user/pulseaudio.service
<benley>
thanks. I was so close, too - I saw systemd.packages and mistakenly read it as environment.systemPackages
fragamus has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
jsgrant_ has joined #nixos
stphrolland has joined #nixos
<stphrolland>
I was trying a little bash script, with the common place #!/bin/bash at the start when the .sh file is set to executable... but it tells me that/bin/bash doesn't exists. I guess it is located elsewhere ?
<benley>
stphrolland: replace the first line with #!/usr/bin/env bash
jsgrant has quit [(Ping timeout: 248 seconds)]
<benley>
Or /bin/sh, if the script will work with plain sh
<benley>
symphorien: how about dbus? I added my package to services.dbus.packages and now it's mentioned in /etc/dbus/session.conf, so that's good, but it doesn't get a symlink in /run/current-system/sw/share/dbus-1/services the way other things appear to
<symphorien>
the only time I touched to dbus I had to write that :https://github.com/symphorien/nixpkgs/blob/12ed7f153475b1db2ca219f07296b53ece63b591/nixos/modules/services/networking/networkmanager.nix#L317
fragamus has joined #nixos
Ralith_ has joined #nixos
fendor has quit [(Ping timeout: 252 seconds)]
Garmy has joined #nixos
jtojnar has quit [(Remote host closed the connection)]
<benley>
That's pretty much the same as what I just did
jtojnar has joined #nixos
metaphysician has quit [(Ping timeout: 240 seconds)]
<symphorien>
I don't know then... sorry
<benley>
I'll figure it out eventually :)
metaphysician has joined #nixos
hamishmack has joined #nixos
jb55 has joined #nixos
goibhniu has joined #nixos
oida has quit [(Ping timeout: 240 seconds)]
Oida has joined #nixos
ottidmes has joined #nixos
<stphrolland>
I used to augment the PATH variable in my .bashrc file. With Nixos, do you advise, or normally, do you do it in programs.bash.interactiveShellInit, programs.bash.loginShellInit, or programs.bash.shellInit, or is there a dedicated nix feature for ENVIRONMENT variables ?
<edef>
environment.variables.FOO
<stphrolland>
edef: do you this personnally to enrich the PATH variable with new folders ?
<edef>
no
<edef>
environment.interactiveShellInit would do okay for that, I guess
<edef>
oh huh, environment.profileRelativeEnvVars is a thing
<edef>
that's new
<edef>
that just goes relative to per-user-pkgs anyhow
<FRidh>
ugh, made a new store, out link to Nix, ran nix-collect-garbage and poof, goodbye store
ambro718 has quit [(Quit: Konversation terminated!)]
jb55 has quit [(Ping timeout: 258 seconds)]
<ottidmes>
I want to use an USB device during boot.initrd.preLVMCommands so I thought about giving it a persistent name through an extra udev rule, but apparently the udev rule does not get applied yet at this point, since it is unavailable, but it available once fully booted
<ottidmes>
Any idea how to fix this?
stphrolland has quit [(Quit: Page closed)]
pie_ has quit [(Read error: Connection reset by peer)]
pie_ has joined #nixos
LexLeoGryfon has joined #nixos
petsa has quit [(Ping timeout: 248 seconds)]
civodul has joined #nixos
Garmy has quit [(Quit: Bye!)]
<LexLeoGryfon>
HELLO GENTLEMEN!
<LexLeoGryfon>
please give me example how to declare boot.kernelParams values in configuration.nix. Thnx
srk has quit [(Ping timeout: 240 seconds)]
<LexLeoGryfon>
let's say I want to include "cryptkey=device:offset:size " parameter
mbrgm has quit [(Ping timeout: 255 seconds)]
dywedir has quit [(Ping timeout: 255 seconds)]
Mateon1 has quit [(Ping timeout: 248 seconds)]
<LexLeoGryfon>
will boot.kernelParams = [ "cryptkey=device:offset:size " ]; do the job? or it will be grammatic\syntax error?
Mateon1 has joined #nixos
srk has joined #nixos
<Moredread>
what is the best place to discuss changes to config option semantics?
FRidh has quit [(Quit: Konversation terminated!)]
<Moredread>
I think that it be better if instead of security.acme.*.production there was an option to set the server directly.
mbrgm has joined #nixos
<Moredread>
similarly to services.caddy.ca
jsgrant has joined #nixos
<Mic92>
vcunat: I can recommend gdbgui, which was recently merged in nixpkgs.
hamishmack has quit [(Quit: hamishmack)]
jb55 has joined #nixos
<vcunat>
Mic92: I feel like I'm missing some context
justan0theruser has quit [(Ping timeout: 240 seconds)]
<LexLeoGryfon>
gentz, plz, halp
<Mic92>
vcunat: no, you don't. This is just a tool tip
digitalmentat has joined #nixos
<vcunat>
oh, right
<digitalmentat>
has anyone successfully setup luksSuspend and luksResume on their NixOS machine?
<vcunat>
I probably will try it out
ottidmes has quit [(Quit: Page closed)]
<digitalmentat>
I've got pretty far basing off of other distro's work but I'm stuck right now with luksResume hanging and it's hard to tell, from a chroot'ed initrd, what is causing this issue (seems like it's blocked on IO?)
<symphorien>
what is luksResume ?
vcunat has quit [(Quit: Leaving.)]
<digitalmentat>
symphorien, if you encrypt your device with LUKS (using cryptsetup) you can safely freeze IO and purge the encryption key from memory with `cryptsetup luksSuspend` and resume IO (after unlocking) with `cryptsetup luksResume`
<symphorien>
ah ok
<digitalmentat>
which is an improvement over vanilla suspend-to-ram
<digitalmentat>
but it requires some gymnastics because you need the utility to unfreeze it, which means you have to create a temporary initramfs with cryptsetup in it
bennofs has quit [(Ping timeout: 264 seconds)]
<Moredread>
digitalmentat: I'd be very interested in your work when you get it running :)
ma27 has quit [(Quit: WeeChat 1.9.1)]
<LexLeoGryfon>
I use plain dm-crypt, meanwhile wondering where to specify 5GB size in crypto=<hash>:<cipher>:<keysize>:<offset>:<skip>
<digitalmentat>
Moredread, I plan on writing a blogpost and adding an article to the user wiki when I am done. I just need to figure out why luksResume is getting blocked
asuryawanshi has joined #nixos
ma27 has joined #nixos
justan0theruser has joined #nixos
digitalmentat_ has joined #nixos
asuryawanshi has quit [(Ping timeout: 255 seconds)]
digitalmentat has quit [(Ping timeout: 240 seconds)]
<LexLeoGryfon>
I always plan the same, but soon as I'm done I forgot that, my bad
<LexLeoGryfon>
few times I tried to contribute post knowledge, but soon as I see website that requires me to REGISTERACCOUNT or prevent me to post via Tor , I wish luck to operator\society and backing to my anonymous world with no login required
hiratara has quit [(Ping timeout: 240 seconds)]
hiratara has joined #nixos
hamishmack has joined #nixos
ssmike1 has joined #nixos
ssmike has quit [(Ping timeout: 240 seconds)]
ssmike1 is now known as ssmike
jb55 has quit [(Ping timeout: 252 seconds)]
<LexLeoGryfon>
I tried nix-env -i emacs-nox-25.3 and it completely hanged up for 10 mins with 100% cpu usage, what happens when I invoke this statement and why it always take long, menawhile why pacman -Syu happens faster?
nixos1022 has quit [(Ping timeout: 248 seconds)]
<LexLeoGryfon>
and is there ways to get faster? like compiling nix-env binary with native optimizations or something
<joepie91>
manveru: I can't for the life of me figure out why - the expression looks correct - but it seems to be defaulting to wine32 even on 64-bits systems
<joepie91>
manveru: as evidenced by the path being wine-2.18, not wine-wow-2.18 as it should be for a WoW build
<joepie91>
now running a build with (nixpkgs.wineUnstable.override { wineBuild = "wineWow"; })
<joepie91>
the build path at least has wine-wow in it now, and it's not in the binary cache
<tanonym>
current install of Telegram has issues with the version of qt. I haven't been able to figure out how to fix the issue.
JagaJaga_ is now known as JagaJaga
civodul has quit [(Quit: ERC (IRC client for Emacs 25.3.1))]
<tanonym>
Running telegram-desktop from console gives the following error:
<tanonym>
QApplication: invalid style override passed, ignoring it.
<tanonym>
Cannot mix incompatible Qt library (version 0x50901) with this library (version 0x50902)
<tanonym>
Aborted
<tanonym>
I have searched online, Arch Wiki did suggest downgrading, but since Nix is built against the required version, it shouldn't be anything in my KDE install that's messing with it. Even with a system upgrade.
<tanonym>
So, seems that something in the expression is pulling a version that telegram doesn't like.
<manveru>
looks like it tries to build with two incompatible dependencies?
deepfire` has joined #nixos
<tanonym>
manveru: That makes sense to me.
deepfire` has left #nixos []
<manveru>
well, take a look at the derivation and see how it gets them
<tanonym>
So I would have to have a look at the expression to see which dependency is which.
deepfire` has joined #nixos
<sphalerite>
Iirc telegram uses a weird patched qt of its own
<manveru>
:)
<tanonym>
Hrmm.
Neo-- has joined #nixos
<tanonym>
I couldn't install this with nix-env -i tdesktop, I had to use nix-env -i telegram-desktop. Looking at the packages list for NixOS it does list the identifier as tdesktop.
<joepie91>
manveru: anything wrong you can see about the wine expression that might cause it to default to wine32?
<tanonym>
cutegram was also broken, it brings up the window but there's nothing inside the border. I just see the widgets.
<sphalerite>
tanonym: needs more -A :)
<tanonym>
Okay, I'll try with nix-env -iA nixos.tdesktop
<tanonym>
See if that works, I might have gotten a quirky build.
<tanonym>
Trying that now.
<tanonym>
I did have a look at the derivation, but still not grokking how everything works yet. Slowly learning.
<tanonym>
Okay, now it works. It still gives a warning about an invalid style override passed, but it ignores it and launches the application
<tanonym>
Okay, client is functional at least. It does throw a bunch of message about unexpected reply signatures: got \"\" expected \"a{sv}\"", so some regex is out of whack as well.
<tanonym>
Guess I need to go digging in their git sources and see what the hell is going on.
hamishmack has joined #nixos
<tanonym>
Still doesn't like being launched from the KDE menu, but I can run it from a terminal.
<tanonym>
It's kludgey but it works now.
<sphalerite>
tanonym: that looks like a dbus type, not a regex, to me
<tanonym>
I stand corrected. I know nothing about dbus and only a little about regex so that's where that guess came from :)
<alphor>
I've just upgraded from 17.03 to 09. There are a couple of modules that aren't defined when building but are defined before the 17.09 tag on GH. What am I missing?
babyflakes has quit [(Quit: Connection closed for inactivity)]
ssmike has joined #nixos
deepfire` has quit [(Remote host closed the connection)]
deepfire has quit [(Remote host closed the connection)]
deepfire has joined #nixos
<orivej>
alphor: what modules?
elurin has joined #nixos
<alphor>
nullmailer
<alphor>
orivej: the module was added with the build, and I definitely can install it interactively
<alphor>
ie in the same commit
JagaJaga has quit [(Ping timeout: 260 seconds)]
<ekleog>
hmm, anyone knows how to not log the “accepted connection from pid X, user Y”? I'm trying nox-update, and this logging makes iotop go crazy :/
<alphor>
orivej: nevermind, it's something with my config. sorry!
<srk>
if I want to package supercollider-sc3-plugins and these depend on supercollider itself should I create a wrapper like supercollider-with-plugins? or should I just callPackage sc3-plugins with self (supercollider?)
<srk>
#recursion
ssmike has quit [(Ping timeout: 264 seconds)]
jsgrant has quit [(Ping timeout: 240 seconds)]
Sonarpulse has quit [(Ping timeout: 252 seconds)]
dan_b has joined #nixos
fiveht has quit [(Quit: WeeChat 1.9.1)]
<srk>
hm I understand how to create a wrapper with plugins but supercollider doesn't seem to allow passing additional paths to it, it expects plugins in its own folder under subdirs