<qwerty_asdf>
I can't find anything in the manual .. unless I missed it
<samueldr>
kiloreux: I'm doubting it's that line specifically, just reading the cmakelists, if you're on nixos... it checks for a path (in the if) that's probably not present on nixos
worldofpeace has quit [Quit: worldofpeace]
worldofpeace has joined #nixos
<samueldr>
though, probably something like `substituteInPlace src/CMakeLists.txt --replace "/lib/systemd/system" "$out/lib/systemd/system"` in postPatch
<samueldr>
though it MAY break the if in creating a directory that *definitely doesn't exist yet*
<samueldr>
so either replace beforehand the if condition (--replace /lib/systemd/system /) or create $out/lib/systemd/system before the check is done?
<samueldr>
(untested)
<samueldr>
qwerty_asdf: doesn't look like there is anything about qutebrowser outside of its derivation, so looks like there isn't a better way (yet)
ajs124 has left #nixos ["Stream closed by us: Timed out waiting for stream resumption (connection-timeout)"]
<kiloreux>
samueldr, the patch way perfectly solved my problem. Thank you very much <3
<qwerty_asdf>
ooo okay, I'll just run the script inside the derivation dir. I thought maybe there was a general config package option to run scripts in derivations. Thanks
<qwerty_asdf>
:)
<samueldr>
kiloreux: did you do two replaces? create the systemd folder? did the install install the files as expected (or was it then a no-op?)
<kiloreux>
I just deleted that part. Since I will need to start it manually anyway. So I created a patch that deletes that part in CMakeLists
<samueldr>
good to know
<kalbasit>
in `~/.config/nixpkgs` the recommended way of adding a new package (that does not exist in nixpkgs) is through an overlay at `~/.config/nixpkgs/overlays/<name>/default.nix` ?
knupfer1 has joined #nixos
<samueldr>
» If the path is a directory, then we take the content of the directory, order it lexicographically, and attempt to interpret each as an overlay by:
<samueldr>
» Importing the file, if it is a .nix file. Importing a top-level default.nix file, if it is a directory.
<samueldr>
so yeah, it should work, kalbasit
<samueldr>
but that default.nix file should *contain an overlay* and not the derivation
<dhess>
Ugh, now I'm running into the problem where you can't build a container with NixOps on macOS
<dhess>
even though literally every other type of deployment.targetEnv works fine from macOS
tertle||eltret has quit [Quit: Connection closed for inactivity]
<kalbasit>
samueldr: I see, I'll give it a shot, thx
andymandias has joined #nixos
<kalbasit>
samueldr: so if I have a new pkg, not an overlay, should it go in `.config/nixpkgs/pkgs/<name>/default.nix` instead?
<kalbasit>
or an overlay is always necessary to add it there?
<clever>
kalbasit: an overlay is still needed to make nixpkgs actually read the new package
<samueldr>
what you could do, kalbasit, is have .config/nixpkgs/overlays/my-awesome-stuff/package-name/default.nix, and then in .config/nixpkgs/overlays/my-awesome-stuff/default.nix you `self: super: { package-name = self.callPackage ./package-name; }`
<clever>
samueldr: missing a {} on the callPackage
<kalbasit>
I do not have to `super.package-name =` ? super is inferred in `package-name =`
<clever>
super is an argument being passed in, and it cant be modified
<qwerty_asdf>
?https
<clever>
you are returning a new set of packages to override
<kalbasit>
ohhh got it
<kalbasit>
thx clever
sigmundv has joined #nixos
thc202 has quit [Ping timeout: 244 seconds]
cryptomonoid has joined #nixos
Have-Quick has joined #nixos
fragamus has joined #nixos
<dhess>
clever: got any NixOps configs with container targetEnv's ?
<clever>
dhess: only ec2 and none
<dhess>
ok thanks
Guest3332 has quit [Ping timeout: 240 seconds]
<tobiasBora>
Hello,
<tobiasBora>
I'm not sure to see why, but when I do "environment.systemPackages = [ pkgs.sslh ]; services.sslh.enable = true;", I don't have sslh running (don't appear on netstat), and "systemctl status sslh" returns that the service does not exist...
<tobiasBora>
Note that on the last run, ./disabled_modules.nix was enabled, but I checked and sslh was not part of the disabled modules
<tobiasBora>
ohhh actually I'm very stupid
<tobiasBora>
it was actually disabled
fragamus has quit [Client Quit]
<tobiasBora>
(I messed things up between commented and uncommented)
<tobiasBora>
looks better now
<kalbasit>
clever: samueldr: so my `all` overlay now looks like this https://gist.github.com/kalbasit/b5d34e1f63286e202306da1ac881bb8c is there a better way of doing the `all = with super.pkgs; buildEnv {`, which has the purpose of installing all the packages I need in my profile?
<clever>
tobiasBora: you also dont need to put it into the systemPackages
<tobiasBora>
ok, I need to enable the firewall also I guess
<clever>
tobiasBora: systemd actively ignores all things in systemPackages
<clever>
kalbasit: you can just do `with super;` to start with
<clever>
kalbasit: super and self are both package sets, nearly identical in function to pkgs itself
<tobiasBora>
clever: I putted it here because initially it was not working, so I thought it was the reason. But I don't understand, why don't I have an error when I run "services.sslh.enable = true;"?
Have-Quick has quit [Quit: Have-Quick]
<kalbasit>
clever: cool!
<clever>
tobiasBora: the lack of an error is odd, can you double-check the files on the rpi and see if rsync is to blame?
<clever>
kalbasit: id also recomend using self, not super in line 32
<hodapp>
bleh. anybody know offhand if it's possible to build Boost with both Python 2 and Python 3 bindings?
cryptomonoid has quit [Ping timeout: 264 seconds]
<clever>
kalbasit: super will give you the version without all the changes this overlay is doing
<tobiasBora>
clever: I checked, and the rsync did the job
<kalbasit>
clever: I found that out the hard way lol, I just changed it to `super` and it failed, but worked with `self`
<clever>
tobiasBora: what exactly did you do to disable the modules?
<samueldr>
pkgs is an attribute of pkgs, so pkgs.pkgs.pkgs.pkgs works (don't know it works though)
<tobiasBora>
clever: I imported the file ./disabled_modules.nix, that contains {...}{disabledModules = [ ... ]}. When sslh was still disabled, I had in the list one line containing "services/networking/sslh.nix".
<kalbasit>
clever: samueldr: so pkgs is defined in itself? cool
Have-Quick has joined #nixos
<clever>
tobiasBora: when disabledModules works correctly, it will fail claiming services.sslh.enable is not a valid option
<samueldr>
I wanted to check if they were the same, but
<samueldr>
> pkgs == pkgs
<{^_^}>
error: syntax error, unexpected '=', at (string):93:9
<samueldr>
uh
<samueldr>
> (pkgs == pkgs)
<{^_^}>
assertion failed at /var/lib/nixbot/.cache/nixbot/nixpkgs/nixpkgs/pkgs/top-level/all-packages.nix:8784:12
<samueldr>
can't compare them du to assertions
<clever>
samueldr: i think thats recursively resolving every single attribute in the set
<samueldr>
probably!
<tobiasBora>
clever: Hum... I should have a pretty recent version (I installed the rasperry pi something like 5 days ago)
<tobiasBora>
I tried one more time to disable it, and indeed, no error, and sslh was not running anymore.
<clever>
tobiasBora: it might be that your own sslh.nix is being disabled by the same flag?
<clever>
rename it?
<kalbasit>
can I make `/etc/nixos/configuration.nix` load the overlays I have in `~/.config/nixpkgs/overlays`?
<clever>
kalbasit: you have to put a list of (already imported) overlays in nixpkgs.overlays
<jtojnar>
how do I use overrideScope? With https://paste.gnome.org/phczjjsym I get elementary-greeter/default.nix:1:1 called with unexpected argument 'stdenv'
<clever>
jtojnar: i think your using enableDebugging wrong, not overrideScope
ericsagnes has joined #nixos
revskill has joined #nixos
<clever>
368 # intended to be used like nix-build -E 'with <nixpkgs> {}; enableDebugging fooPackage'
<clever>
nvm, that is the correct way to use enableDebugging
<clever>
i was thinking of keepDebugInfo
<revskill>
Hi, i gove the issue: waiting for locks or build slots...
<revskill>
what to do now ?
<clever>
jtojnar: try removing eself:
<clever>
revskill: what command did you run?
<jtojnar>
clever: oh right, the expression does not accept stdenv
jackdk has joined #nixos
<jtojnar>
worldofpeace: ^
<revskill>
i build Haskell project with stack
<infinisil>
> 1 == 2
<{^_^}>
false
<infinisil>
> "foo" == "bar"
<{^_^}>
false
<jtojnar>
worldofpeace: that is another issue with custom mkDerivation helpers
<tobiasBora>
clever: it was a good idea, but nope... If it can help, it says "stopping the following units: sslh.service [...] removing user ‘sslh’" when I disable it
<revskill>
clever: the problem is with packages blas and liblapack :|
<clever>
revskill: what command did you run?
<revskill>
stack build
<clever>
revskill: do you have build slaves configured?
<revskill>
no, i just use stack integration with nix
<clever>
revskill: nixos or normal linux?
<revskill>
because the library requires lapack to build
<revskill>
on Macosx
Forkk has joined #nixos
<clever>
revskill: can you pastebin the output of `ps aux` ?
tzemanovic has quit [Remote host closed the connection]
<clever>
revskill: it appears to be currently building atlas, and it should start the next build when that finishes
Have-Quick has quit [Quit: Have-Quick]
<revskill>
clever: Thanks for the info, i'm making my coffee
<jtojnar>
I just do not undestand why the function accepted by makeScope's overrideScope has the arguments swapped (compared to overlays and haskell's overrideScope)
<kalbasit>
clever: how to change the location of `/etc/nixos/configuration.nix`?
<clever>
kalbasit: nixos will search for <nixos-config> in $NIX_PATH
<acowley>
jtojnar: That has also infected emacs configuration, iirc. It's terrible :()
wpcarro has quit [Ping timeout: 240 seconds]
<kalbasit>
oh I see it now `nixos-config=/etc/nixos/configuration.nix`
<kalbasit>
not sure what's wrong, but see https://gist.github.com/kalbasit/493201e0cda1c431d284bcae64703b35 (../.. is actually located at `/home/kalbasit/.nixpkgs`), I have manually exported NIX_PATH so nixos-rebuild finds the right path, `nixos-rebuild test` fails with the error `error: illegal name: '.nixpkgs'`
<kalbasit>
I'm trying to avoid hardcoding the path as I plan to move it to `.config/nixpkgs` and my username could change as well
<clever>
kalbasit: the path refered to by ../.. must not begin with a .
<dhess>
Hmm I wonder if somewhere in my config nixpkgs.pkgs is getting set, which makes nixpkgs ignore nixpkgs.localSystem.system and nixpkgs.system
<clever>
rotaerk: for gui programs, the version of nixpkgs you built the project from has to be fairly similar to the nixpkgs the host OS was built from
<dhess>
maybe overlays?
<clever>
dhess: overlays dont set nixpkgs.pkgs
Mr_Keyser_Soze has joined #nixos
<dhess>
clever: ok that's good to know
jperras has joined #nixos
<rotaerk>
clever, hmm interesting; I'll try upgrading nixos and see if that helps at all
<dhess>
There's this bit in nixos/modules/misc/nixpkgs.nix, which is probably where the assert is coming from: pkgs_i686 = cfg.pkgs.pkgsi686Linux;
<clever>
dhess: that only gets ran if you refer to the pkgs_i686 argument in any nixos module
<dhess>
oh because of laziness?
<clever>
dhess: can you pastebin the backtrace from --show-trace ?
<clever>
yeah
<dhess>
see that's what's odd, because I'm not explicitly calling that and the container literally has no services configured at all. It's just the bare minimum container configuration for NixOps.
<dhess>
let me check the backtrce
<dhess>
looks like it's coming from the "bonds" attribute in networking
<clever>
the state of this file, at the commit you linked
<clever>
now just change the branch in the UI
<clever>
nixos-18.03 lacks it
<infinisil>
Or: git fetch channels (assuming you have that as a remote), `git branch --contains 6bf1421f13d667c2997b67728cf777c6a70716a5 -r | rg 'channels/.*'`
<rotaerk>
hmm I see.
<infinisil>
Shows nixos-unstable, nixos-unstable-small and nixpkgs-unstable for me
<clever>
infinisil: still has the same problem that backport commits have a different hash
<rotaerk>
think I will need to switch to the unstable channel, then
<infinisil>
clever: Oh yeah
<rotaerk>
if I want this to work
<infinisil>
Damnit git
<infinisil>
clever: pijul doesn't have that problem
<rotaerk>
never heard of that *googles*
<rotaerk>
dammit, don't introduce me to something that's going to make me hate what I'm currently using
<infinisil>
Hehe
<rotaerk>
do you guys usually use unstable or stable, on machines where you do development?
<infinisil>
With pijul a branch is just a *set* of changes, and you can add the same change (identified by hash) to different branches
<infinisil>
clever: ^
<rotaerk>
I'm thinking unstable isn't going to be THAT bad, since it's not arch and you can roll back if you need to
lassulus_ has joined #nixos
<dhess>
rotaerk: I use nixos-unstable-small typically, and I run Hydra builds against that channel on as much of my software as possible to see if it's "safe" before deploying it
<dhess>
except on macOS where I use nixpkgs-unstable
<clever>
dhess: something i was thinking of today, is to make a nix expression so hydra builds my entire deployment
<clever>
dhess: let me throw one together...
<dhess>
clever: somewhere I have some notes about Hydra'ing my NixOps deployments; not actually deploying them, just building the closures
<rotaerk>
hmm k
<dhess>
it might be something you said a few months back
<dhess>
I've also recently refactored all of my NixOps networks so that I can swap in a different "state" overlay. Eventually I want to use that to deploy to VirtualBox networks before deploying to production hosts, at least as much as is feasible.
jperras has quit [Quit: WeeChat 2.1]
fragamus has joined #nixos
<dhess>
well now I have a choice to make: start deploying containers, which would make a lot of my future work easier, but lose the ability to deploy from macOS; or don't deploy containers, replace some modules from NixOS with my own versions to support multi-instance services (like unbound), and retain the ability to deploy from macOS.
Have-Quick has joined #nixos
<dhess>
I guess I should do the container thing and meanwhile try to figure out why the macOS stuff isn't working.
<rotaerk>
"Nix Hydra is the world's only purveyor of eggcentric apps backed by both teenage girls and venture capitalists. Proud creators of Egg Baby and Hot Guy Alarm Clock."
<rotaerk>
lol
<dhess>
or option 3 is to switch to Kubernetes :)
jackdk has joined #nixos
<rotaerk>
dunno if I'll do the hydra testing; this is just my PC; I'm not a server admin
<Lisanna>
hydra isn't updating my declarative jobsets in a very timely manner. the .jobsets build finished 14 minutes ago, but the displayed jobsets still have the old config
<iqubic>
rotaerk, get back to #haskell where you belong. LOL
<dhess>
won't some crazy NixOS person please start a VC-backed company to integrate NixOS, Terraform, and Kubernetes please.
<rotaerk>
I will, once I can get my nix package to actually work :P
<iqubic>
Whst package?
<dhess>
rotaerk: yeah if you're not comfortable with server admin you should probably stay away from Hydra :)
<iqubic>
Why are all the #haskell folks flocking here?
<rotaerk>
but watch this upgrade to unstable magically fix all that
<iqubic>
Do you know how to do that?
<dhess>
iqubic: there was a decent amount of Nix interest at ZuriHac this year.
<dhess>
from people who haven't used it (much) yet
<iqubic>
What is ZuriHac? I don't know that convention.
<dhess>
A Haskell conference in Zurich each summer.
<iqubic>
Ah. I see. I really need to attend some conferences. However, I live in Seattle, Washington, USA. Most conferences happen in Europe as far as I can tell.
jackdk has quit [Ping timeout: 260 seconds]
halfbit has quit [Ping timeout: 245 seconds]
<rotaerk>
Hmm, well I'm on my phone now... The upgrade to unstable just closed X and put me on the console saying <<< NixOS Stage 1 >>>
<rotaerk>
seems unresponsive
<dhess>
There are indeed a lot of haskell conferences in Europe.
<iqubic>
I am however attending a summer camp focused around AI, ML and the Python Tensorflow library
<clever>
rotaerk: same thing happened to me, it killed something dbus related and broke the entire gui
<rotaerk>
Guess I'll hard reboot
<clever>
rotaerk: pressing the power button normally triggers a normal reboot
<iqubic>
dhess: and plenty of other coding things.
<clever>
rotaerk: no need for hard
<rotaerk>
Oh k
<iqubic>
clever: can I make it so that pressing power button does nothing.
<iqubic>
I have accidentally hit that on my laptop more times than I care to admit.
<rotaerk>
Sounds like something that would create more problems than it would solve
<iqubic>
I lose far too much work to that. I wish I could change that.
<clever>
iqubic: i have done that on gentoo before, but i dont remember how
<iqubic>
rotaerk: what would?
<iqubic>
clever: I was hoping there would be a simple configuration.nix option to do it.
<rotaerk>
Oh you want to still be able to hold it to force it to power off
<clever>
iqubic: i use nixops to manage my NAS and router, both are within the house
<clever>
[clever@system76:~/nixos-configs]$ nixops deploy -d house
<dhess>
clever: do you have any clever tricks for switching a laptop between a config that assumes you have remote builders available, and one that assumes you're, say, on a plane with no remote access?
<dhess>
I mean, there's the obvious "rebuild --switch" but I have to remember to do that before I get on the plane :)
<clever>
dhess: sudo rm /etc/nix/machines
<dhess>
heheheh
<dhess>
ok fair enough
<clever>
dhess: nixos will automatically repair it at the next rebuild-switch or boot
<dhess>
oh interesting, I would have thought that nix-daemon would read and cache that
<clever>
thats one file it doesnt seem to cache
<dhess>
I guess you could HUP it in any case
<rotaerk>
the fact that there was something in nixpkgs that was impacted by the environment bugs me
<rotaerk>
though I'm guessing it's a necessary compromise for some reason
<clever>
rotaerk: nixpkgs has to load the libGL in /run/opengl-drivers/ to match what your gpu needs
sbdchd has quit [Remote host closed the connection]
sbdchd has joined #nixos
<rotaerk>
hmm I wonder if I should have my project reference nixpkgs-channels rather than nixpkgs
<clever>
rotaerk: that only matters if you reference a branch
<clever>
you can just pick the rev from a channel, and fetch it from nixpkgs
<rotaerk>
ah
<clever>
dhess: hydra has been re-deployed, now eval'ing
<clever>
dhess: still failed?!
<dhess>
:(
<clever>
Jul 17 00:19:00 nas hydra-evaluator[13174]: /nix/store/fxlvm8c0dv46ciq6cbfrm6kkg12c2vwk-hydra-2017-11-21/bin/.nix-prefetch-git-wrapped: line 187: awk: command not found
<clever>
dhess: ah yeah
<dhess>
is there some length limit to a container name in NixOS? My container "test-container" is running as container@test-co.service :\
<clever>
dhess: does it have the same shortened name in /etc/systemd/system/ ?
<dhess>
On the host? It doesn't show up there at all.
<dhess>
Only container@.service
<clever>
dhess: read that .service file
<clever>
it should contain a script that refers to another dir
<clever>
and uses the value after the @ to compute the full path
<dhess>
clever: "test" doesn't show up in that container@.service file at all. It's my understanding that systemd will automatically apply the name as an argument for a script like that
<dhess>
/script/service/
<dhess>
Jul 17 03:20:24 hoover systemd[1]: Started Container 'test-co'.
<dhess>
it's clearly running it as 'test-co' rather than 'test-container'
<clever>
dhess: yeah, it looks like it is being truncated
<dhess>
wtf
<dhess>
that is bizarre
<clever>
EnvironmentFile=-/etc/containers/%i.conf
<dhess>
DOS all over again
<clever>
dhess: and if you check in this dir?
<dhess>
yeah test-co.conf
<jackdk>
7.4 instead of 8.3. progress
<clever>
lol
<dhess>
what on earth is going on
<{^_^}>
[nixos-channel-scripts] @matthewbauer opened pull request #20 → Add netboot to mirror-nixos-branch.pl → https://git.io/fN34S
<dhess>
Does anyone actually use containers in NixOS? :D
freeman42x]NixOS has quit [Ping timeout: 244 seconds]
<clever>
dhess: i sometimes do when i run services for a friend and dont fully trust the service's security
<dhess>
ohh it might be a systemd unit name limitation
<dhess>
hmm not likely
<bongsun>
i also use containers but not declaratively with systemd-nspawn. its nice to have a debootstrapped folder to test what isnt available yet in nixpkgs.
jackdk has quit [Ping timeout: 264 seconds]
<clever>
dhess: i cant see where that name truncation is coming from
<dhess>
yeah I guess I should have asked, more specifically, does anyone use *declarative* containers with NixOS
<clever>
dhess: this maps over the config.containers set, and creates a containers/${name}.conf for each key in the set
<dtz>
can we not completely break staging please? >:| :( :'(
<rotaerk>
lol
<clever>
dhess: the nas has successfully built itself on hydra, no job took over 5 seconds, and it was mostly recreating config files without the secrets
wpcarro has joined #nixos
<samueldr>
dhess: I tried to use declarative containers at first, but hit issues with services not reliably restarting on config change :/
<clever>
dhess: it has also done the same with the router
<samueldr>
maybe it's my custom service
<dhess>
clever: awesome
tll has quit [Read error: Connection reset by peer]
<dhess>
samueldr: it seems there are other issues as well, like "nixops reboot <container>"
<samueldr>
(though I don't use nixops at all)
<dhess>
I would have thought this stuff had been more plumbed out by now
<clever>
dhess: so it will now be pre-building (and blowing all my free space) both machines every time nixpkgs updates
<dhess>
clever: sweet
tll has joined #nixos
<dhess>
I'll try to do the same with my configs over the next few days
tll has quit [Client Quit]
<bongsun>
so i picked up an odroid xu4 and im now realizing that booting arm machines isnt as straightforward as intel ones with blobs etc.
<bongsun>
does anyone have a working configuration.nix for an odroid xu4 with hdmi working?
jackdk has joined #nixos
srdqty has joined #nixos
<dhess>
it's not the hyphen -- calling my container "blahblahblahblah" also gets shortened to "blahbla"
<dhess>
the hostname on the container is set properly. It appears this is just some kind of "machine ID" for systemd-nspawn to use
<srk>
bongsun: I only run it headless. does it run with mainline kernel btw?
<dhess>
Unfortunately, using NixOps to deploy a new container with a new name did not stop the old container
<dhess>
ugh
<clever>
dhess: nixos wont restart containers upon making declarative changes
wpcarro has quit [Ping timeout: 256 seconds]
<bongsun>
srk: hey thats cool, im only able to run ubuntu with kernel 4.14 with the stock images. did you get far without a UART cable? i just ordered one and could >week to arrive.
<dhess>
I guess that makes sense. I should probably have to tell NixOps to destroy it.
Have-Quick has quit [Quit: Have-Quick]
<srk>
bongsun: no, I only run these boards headless so I have plenty of USB->UARTS :D I've promised to update pull request adding support for it - now it looks like only kernel is needed, uboot is stock - Linux nixodroid 4.14.29 #1-NixOS SMP PREEMPT Thu Jan 1 00:00:01 UTC 1970 armv7l GNU/Linux
<dhess>
clever: is this filed as an issue somewhere?
<clever>
dhess: not that i know of
<dhess>
well it looks like there are a lot of declarative container issues :)
orivej has quit [Ping timeout: 240 seconds]
<dhess>
clever: played around with Kubernetes at all?
jackdk has quit [Ping timeout: 256 seconds]
<clever>
dhess: nope
justbeingglad has joined #nixos
justbeingglad has left #nixos [#nixos]
<takeda>
is there an equivalend of fetchTarball function that similarly to fetchurl checks provided hash (of the archive not its contents) and works with cleanSource? I basically want a command that would extract archive to a directory, because before I start the actual build I just wanted to analyze the contents of the archive to extract some information about it
marusich has joined #nixos
<kalbasit>
so in `/etc/nixos/configuration.nix` I import another file that sets `security.pki.certificates`, How can I append to the existing value after the import? What I tried is this: https://gist.github.com/9debee38e1d577df27dfa8dabaae1624 but I get infinite recursion error
<ghasshee>
Yes, for that command, I think I should pass some argument to kernel or initramfs ... but I do not know how to ..
ntqz has joined #nixos
<adisbladis[m]>
ghasshee: You shouldn't have to.
<adisbladis[m]>
ghasshee: And btw, if you close your laptop lid systemd will put the laptop to sleep for you
<adisbladis[m]>
So if everything works as intended you shouldn't really have to do much
<etu>
(That's suspend though)
<ghasshee>
sleep is working fine, but not hibernate in my laptop..
<adisbladis[m]>
etu: Oh. Sorry, I mixed them up.
<etu>
Well, these two things have a history of being buggy. Suspend is quite fine on most laptops nowadays. I haven't used hibernate since I've started with Full Disk Encryption.
<ghasshee>
In archlinux I remember I passed 'resume=/dev/myswap' to kernel via GRUB and it didn't consume battery quite at all.
<etu>
ghasshee: For instance for hibernate you have to have an unencrypted swap on your drive etc.
<iqubic>
LnL: I need an ssh connection even if I'm going to be using local host as the build server for hydra?
<adisbladis[m]>
iqubic: Yes.
<iqubic>
What benefits does hydra give me that straight up NixOS doesn't?
<ghasshee>
yes.. encryption is now too high level for me now, I will try later .. :-)
<ghasshee>
@etu
<iqubic>
Like why would I go through the pain of setting up Hydra if I'm just an "average" nixos user?
hamishmack has joined #nixos
<adisbladis[m]>
ghasshee: Actually encryption is very easy on nixos :) If the config generator detects an encrypted partition that's saved to your config
* etu
has used LUKS since adisbladis[m] told him to about 10 years ago
<adisbladis[m]>
:)
<iqubic>
Woah, you guys have known each other for 10 years?
<adisbladis[m]>
iqubic: Just about I think
<iqubic>
Now I feel like the new kid on the block, having only been using IRC for the past 2 years.
ntqz has quit [Read error: Connection reset by peer]
<ghasshee>
adisbladis[m]: Oh, really, that sounds greate :-)
<ghasshee>
great
<etu>
iqubic: yeah, something like that. I don't keep count of when I first met everyone I know :p
<iqubic>
How long has IRC been around? Have you guys all been on IRC for like a long time?
ntqz has joined #nixos
<dramforever>
IRC must have been around since like forever ago
<etu>
The first standard for IRC seems to be 30 years old this year
ntqz has quit [Read error: Connection reset by peer]
<dramforever>
I mean, how could we be still talking in plaintext in 2018 if it weren't because of something that old?
<dramforever>
(Similar arguments goes for programming)
<adisbladis[m]>
iqubic: 15-20 years I think?
<adisbladis[m]>
Something like that
<iqubic>
Well now I feel really young.
<iqubic>
I mean, I am only 18...
<etu>
iqubic: But me and adisbladis[m] kinda hanged out in person for several years as well on irc. Not much in person lately due to locality :)
<iqubic>
I really did miss the glory days of computing. Darn.
<adisbladis[m]>
iqubic: 10 years difference then. I'm 28
ntqz has joined #nixos
<Ashy>
for a while i was worried irc wasnt getting enough new blood joining anymore
coot has joined #nixos
<iqubic>
It isn't. At least I don't think it is.
cryptomonoid has quit [Ping timeout: 248 seconds]
<adisbladis[m]>
iqubic: These are the glory days of computing :)
ntqz has quit [Read error: Connection reset by peer]
<andi->
IRC is my only mainstream messenger.. Its great and for the last 15ish years it never let me down :)
<adisbladis[m]>
Things are mostly getting better
<etu>
iqubic: No worries, time flies, you'll learn things and contribute to things you like and enjoy and before you know it there will be other people feeling young :)
<sphalerite>
adisbladis[m]: I like matrix in principle but not so much in practice. Search doesn't really work at all, performance is awful, and there's still no decent plain-text client AFAIK
<iqubic>
etu: I'm not planning on leaving IRC behind for a long time.
<iqubic>
adisbladis[m]: What is matrix?
<andi->
I have issues with matrix in regards to logging and how much it stores... Basically every few seconds when you retrieve new data it writes to the database. Collects all your IPs for no reason etc..
<sphalerite>
adisbladis[m]: not including encryption.
<adisbladis[m]>
sphalerite: What. I didn't know that.
ntqz has joined #nixos
<sphalerite>
idk maybe that's changed. But last I checked it didn't support it.
<adisbladis[m]>
iqubic: It's like XMPP(jabber) for the new generation
<iqubic>
bluetooth headsets are still being a bit of a pain.
<andi->
As an operator of an IRC service it is easy to not have any user data. As soon as you have matrix (or a bridge to IRC for matrix) you are deep into the logging all conversations business..
<ghasshee>
andi-: lol
<sphalerite>
oh and rooms aren't encrypted by default
<adisbladis[m]>
andi-: Yeah I know :) Me and etu used to host quite a large one back in they day
<adisbladis[m]>
Though most servers back then was pretty bad about leaking user info all over the place
<adisbladis[m]>
Ours was heavily patched
<etu>
yeah, that was great fun
<andi->
I am still somewhat involved in one and the matrix bridge is a headache..
dramforever has quit [Ping timeout: 252 seconds]
Izorkin has joined #nixos
<etu>
It was fun when we hit a bug in the software that we couldn't have more than 255 people (I think it was) online at once, so we reached out to upstream and they had no idea that any network "that big" actually used their software :D
<iqubic>
what software was it?
<adisbladis[m]>
etu: Was it ngircd?
<etu>
adisbladis[m]: I think it was, but it was also like 8 years ago we had that running
<adisbladis[m]>
Yeah I don't really remember exactly
<etu>
Some of the privacy patches reached upstream iirc
<{^_^}>
mozilla/nixpkgs-mozilla#101 (by bkchr, closed): VidyoDesktop: Fixes missing libidn by using 1.34
<garbas>
first*
<mpickering>
I am trying to use buildPythonPackage, the error is "No matching distribution found for six==1.9.0"
alexteves has joined #nixos
<infinisil>
mpickering: The package seems to be having a dependency on *exactly* version 1.9.0, which is not good for nixpkgs. The way to resolve this is to patch the source to use `six>=1.9.0`
<infinisil>
Upstream if possible
<mpickering>
is there an equivalent of jailbreaking for python?
ntqz has joined #nixos
<bkchr[m]>
garbas: yeah np :)
<infinisil>
mpickering: Not that I know of
fenedor has joined #nixos
ntqz has quit [Read error: Connection reset by peer]
<garbas>
bkchr[m]: i just noticed that you are also from berlin. did we meet already? /me hopes this is not already another embarrasing situation that my memory put me in
<qyliss>
Has anybody got a working development environment for a ruby gem using nix-shell?
<bkchr[m]>
garbas: No, I don't think that we ever met ^^
<infinisil>
,dnw qyliss
<{^_^}>
qyliss: "It doesn't work" literally only has 0 bits of information useful for debugging, please be more specific: What doesn't work? What's the error?
<infinisil>
:P
<qyliss>
I’m typing!! :P
<infinisil>
np, just wanted to get a chance to use dnw lol
<qyliss>
hehe
<infinisil>
,dnw = "Does not work" isn't very helpful: What doesn't work? What's the error?
<{^_^}>
dnw defined
<qyliss>
anyway, doing that tries (and fails) to install the gem I’m working on, which isn’t what I want — I want a development environment for the gem, so I think I want all the dependencies, but not the gem itself
<{^_^}>
[nixpkgs] @pSub pushed commit from @zgrannan to master « markdown-pp: init at 1.4 (#43525) »: https://git.io/fN3dW
johanot has joined #nixos
ntqz has quit [Read error: Connection reset by peer]
ntqz has joined #nixos
<infinisil>
qyliss: Have you read the last section of the linked chapter? " The first one allows one to quickly drop into nix-shell with the specified environment present. E.g. nix-shell -A sensu.env would give you an environment with Ruby preset so it has all the libraries necessary for sensu in its paths"
ntqz has quit [Read error: Connection reset by peer]
<infinisil>
Where are data files supposed to go? $out/share/???
<infinisil>
Used exclusively by the program in $out/bin
<etu>
fish puts file in $out/share so it seems reasonable
rui has joined #nixos
ntqz has joined #nixos
<rui>
I use i3, and also do 'gnome3.enable = true' to use stuff like gnome-control-center
<rui>
But now gnome-control-center doesn't work. After nixos-rebuilding yesterday
<infinisil>
etu: Oh yeah and haskell apparently does too, but with a derivation name postfix, so $out/share/<name>/
<rui>
It opens an almost empty window
<qyliss^work>
infinisil: I still get errors
<qyliss^work>
one is that the entry for my gem in gemset.nix doesn't have a source
ntqz has quit [Read error: Connection reset by peer]
cryptomonoid has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixos-hardware] @michaelpj opened pull request #65 → Add readme section on using fetchgit → https://git.io/fN3NV
ntqz has joined #nixos
<tfc[m]>
hey there. does anybody use nix-build in the gitlab CI? it is very easy and straight-forward to use the nixos docker image and run nix-build on every pull request, but i am worrying about caching the /nix/store folder. does anybody have some thoughts and/or experience with that?
<srhb>
tfc[m]: Maybe bind mounting in the nix daemon socket and the store itself?
<srhb>
Building sensible gc roots may be tricky.
<{^_^}>
[nixpkgs] @Ekleog opened pull request #43658 → chromium: do jumbo builds → https://git.io/fN3Nx
<tfc[m]>
srhb: not sure if that is even possible with gitlab ci... i mean bind mounting stuff into them
ntqz has quit [Read error: Connection reset by peer]
Neo-- has joined #nixos
<ij>
gchristensen, I want to use the type difference to treat data differently, but if someone had a string path, one couldn't cast it :(
<gchristensen>
ij: what is your goal?
ntqz has joined #nixos
<ij>
I want to pass either a string or a list of (string or path), so I could strings read from files (i.e. get a [string]) and concat those.
<Dezgeg>
something like the readPathsFromFile function does?
tzemanovic has joined #nixos
<ij>
Dezgeg, not quite
johanot has quit [Ping timeout: 268 seconds]
johanot has joined #nixos
the_real_plumps has quit [Quit: No Ping reply in 180 seconds.]
ntqz has quit [Read error: Connection reset by peer]
<{^_^}>
[nixos-hardware] @yegortimoshenko pushed to yegortimoshenko-patch-2 « common/cpu/intel: modesetting is the default »: https://git.io/fN3A5
<{^_^}>
[nixos-hardware] @yegortimoshenko pushed 2 commits to master: https://git.io/fN3pH
ericsagnes has joined #nixos
<{^_^}>
[nixos-hardware] @michaelpj opened pull request #67 → Some suggestions for CONTRIBUTING → https://git.io/fN3p5
ntqz has quit [Read error: Connection reset by peer]
<elvishjerricco>
nh2: It can take either
spacefrogg has joined #nixos
<nh2>
elvishjerricco: hmm I get `error: value is a function while a set was expected`
the_real_plumps has joined #nixos
rprije has quit [Ping timeout: 240 seconds]
ntqz has joined #nixos
ntqz has quit [Read error: Connection reset by peer]
aw has joined #nixos
<nh2>
elvishjerricco: btw I'm using your snippet *in an overlay definition*, should that be OK/
nwspk has joined #nixos
Mic92 has joined #nixos
<nh2>
elvishjerricco: never mind, I had a typo: "self: super:\nself: super"
<elvishjerricco>
nh2: Ah the woes of dynamic typing :P
<nh2>
elvishjerricco: yeah it's super counter-productive, especially because it says just `error: value is a function while a set was expected` without any error location.
<joepie91>
doesn't look like a dynamic typing issue to me?
<elvishjerricco>
joepie91: They passed a function of four args where a function of two args was expected and got a runtime error. Furthermore, override uses dynamic typing to allow different types to be passed to it, which can make situations like these hard to reason about (although in this case it was a red herring)
ntqz has joined #nixos
ntqz has quit [Read error: Connection reset by peer]
<joepie91>
that seems like an error reporting problem, not a dynamic typing problem
<joepie91>
it would have still been just as wrong without dynamic typing
<{^_^}>
[nixpkgs] @dotlambda opened pull request #43660 → [18.03] home-assistant: mark as insecure → https://git.io/fN3jW
<joepie91>
and it failed loudly, so aside from the questionable error reporting (which is a broader issue in Nix, and not really related to the typing model) I really don't see how dynamic typing introduces any 'woes' there
ntqz has quit [Read error: Connection reset by peer]
<nh2>
elvishjerricco: is this how you do it?
<elvishjerricco>
nh2: When I'm trying to be conscious of the extend problem yea. If I'm just doing things quickly and carelessly I just use extend since it's more concise, works fine often, and fails loudly.
ntqz has joined #nixos
<elvishjerricco>
nh2: Worth noting you can feel free to use final.lib.composeExtensions rather than previous.lib.composeExtensions. Shouldn't matter at all, but it's a cheap way to potentially be more correct.
ntqz has quit [Read error: Connection reset by peer]
<srhb>
gchristensen: Careful, liquid nitrogen in a home setting can be quite dangerous.
ntqz has joined #nixos
<d1rewolf_>
hi guys. Trying to get full disk encryption going following https://nixos.wiki/wiki/Full_Disk_Encryption. I can follow the steps and generate hardware-configuration.nix, but when rebooting I am not prompted for encryption phrase and filesystem isn't mounted. any ideas what I'm missing?
<nh2>
elvishjerricco: OK, no I just need to fix an
<nh2>
error: cannot coerce null to a string, at /nix/store/w1fndb8swaamrf7amgjjvxfp12nifl2d-2c07921cff84dfb0b9e0f6c2d10ee2bfee6a85ac.tar.gz/pkgs/development/haskell-modules/make-package-set.nix:131:71
<nh2>
I'm suspecting I'm using `callCabal2nix` wrong.
<nh2>
What is its third argument, usually passed as {} ?
<Raybih>
, nontrinitarian denominations comprise a minority of modern Christianity.
<d1rewolf_>
gchristensen, that looks good...I'll try it. thanks!
Guest50115 has quit [Quit: WeeChat 1.4]
<Taneb>
Running Hydra with "spike" as a build machine, I get the error "possibly transient failure building '/nix/store/jw1525wgk8pky990ylrgpiqd0ixwnjli-bash43-029.drv' on 'nix@spike': opening lock file '/nix/var/nix/temproots/1856': Permission denie"
Raybih has left #nixos [requested by gchristensen (not here)]
<Taneb>
*denied
<Taneb>
I don't think that file exists on either the VM running hydra or the build machine
<Taneb>
So I think it's something trying to make that file can't
<elvishjerricco>
nh2: Its third argument is just there in case you want to override the arguments to the generated default.nix
cryptomonoid has joined #nixos
<elvishjerricco>
What version of nixpkgs are you on?
<nh2>
elvishjerricco: the thing above I wrote seems to cause the issue. Note for me `pkgs = (import <nixpkgs> {}).pkgsMusl`, as I'm working on the musl stuff
<elvishjerricco>
Oh that could do it
<elvishjerricco>
> pkgs.pkgsMusl.glibcLocales
<nh2>
so somewhere in the combination of musl + overlays + haskell .override there seems to be an issue that it goes via this glibc thing. But I don't know where
<elvishjerricco>
nh2: Yea the problem is that callCabal2nix relies on the glibcLocales thing. Not sure if it should... But if the pkgsMusl stuff doesn't provide that derivation, that would break callCabal2nix
ntqz has joined #nixos
<nh2>
elvishjerricco: do you know what it uses this locales stuff for? And can I comment it out or override something so that it doesn't?
<elvishjerricco>
nh2: No, I don't really know much about the LOCALE_ARCHIVE environment variable or why cabal2nix would care about it
<elvishjerricco>
Or if musl provides an alternative
graphene has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @nlewo opened pull request #43664 → goBuildPackage: Add -x to the go build command if NIX_DEBUG >= 1 → https://git.io/fNsJL
cryptomonoid has quit [Ping timeout: 240 seconds]
ntqz has quit [Read error: Connection reset by peer]
graphene has joined #nixos
ntqz has joined #nixos
dbmikus has joined #nixos
ntqz has quit [Read error: Connection reset by peer]
drewr has quit [Remote host closed the connection]
<{^_^}>
#43665 (by nh2, open): callCabal2nix doesn't work with musl-native
<nh2>
dtz: ^
__monty__ has quit [Quit: leaving]
ntqz has joined #nixos
<nh2>
elvishjerricco: but `callCabal2nix` is pure in the sense that it only generates a literal nix expression, right? So I could take callCabal2nix from a glibc package set instead of the musl one, use it to generate the expression, and then build that expression Haskell package with musl?
<Taneb>
Separate to my hydra woes: writeShellScriptBin (from trivial-builders.nix in nixpkgs) fails for me with an error about something being a directory inexplicably
<dtz>
nh2: I think I just never got that far previously--other nixpkgs references to using glibcLocales w/haskell on musl are conditional but this isn't
<dtz>
might be as easy as ... well, 1 sec
dbmikus has quit [Ping timeout: 256 seconds]
ntqz has quit [Read error: Connection reset by peer]
<nh2>
dtz: haha you have set yourself a high bar of expectations, because so far every time you said "1 sec" I you got the solution shortly afterwards indeed ^^
<Baughn>
Well, via Crostini. It was entirely painless.
<gchristensen>
wow, cool!
<nh2>
dtz elvishjerricco: so I suspect that this is very related to some other thing I'm working on in GHC & co: That many Haskell build tools break when you don't force the encoding: https://github.com/commercialhaskell/stack/issues/3988
<{^_^}>
commercialhaskell/stack#3988 (by nh2, open): stack and ghc don't work well on Windows when the user name has non-ASCII chars in it
<{^_^}>
[nixpkgs] @dtzWill opened pull request #43666 → haskell-modules: only include glibc locales when using glibc → https://git.io/fNsT8
<dtz>
nh2: interesting
<Baughn>
gchristensen: No customization needed; this Just Works. Though it does still require a Pixelbook on the dev channel. Give it a couple months.
<dtz>
nh2: well musl is utf8 by default :3 lol
<Baughn>
I'll probably send a few PRs to make the graphics work, once virtio-gpu... works.
stepcut has joined #nixos
ntqz has joined #nixos
<dtz>
(nh2: please take a look at the PR when you can, I haven't tried with your example or anything)
knupfer has joined #nixos
jperras has joined #nixos
<nh2>
dtz: will try now
ntqz has quit [Read error: Connection reset by peer]
<pie_>
pretty sure im not in a nix shell right now
<johanot>
pie_: All of the above alternatives can be used... Alternatively, a fetchGit url like : file:///path/to/stuff is supported
<gchristensen>
b/c if pkgs.lib.inNixShell then drv.env else drv <- detects you're in a nix shell and gives you a thing you can't build
<gchristensen>
and that thing you can't build will be a thing without a $src, and will result in that build error
<gchristensen>
I fought this fight for like an hour just yesterday '(
<pie_>
thats seems kind of horrible
ntqz has joined #nixos
<gchristensen>
maximum agreement
<pie_>
well it looks like it gives drv.env
<gchristensen>
yeah so ... is the env var IN_NIX_SHELL set?
ntqz has quit [Read error: Connection reset by peer]
Sonarpulse has joined #nixos
endformationage has joined #nixos
<pie_>
gchristensen, doesnt look like it
graphene has quit [Read error: Connection reset by peer]
<pie_>
ugh
<pie_>
Setup: This package description follows version 2.2 of the Cabal
<pie_>
specification. This tool only supports up to version 2.0.1.0.
ntqz has joined #nixos
<pie_>
halp
ntqz has quit [Read error: Connection reset by peer]
graphene has joined #nixos
betaboon has joined #nixos
<gchristensen>
unfortunately you've exhausted my knowledge of haskell
<gchristensen>
+ nix
<gchristensen>
maybe try deleting "if pkgs.lib.inNixShell then drv.env else drv" and adding back in just "drv"
<pie_>
x)
<pie_>
thats how i got this far
wpcarro has quit [Ping timeout: 244 seconds]
<pie_>
aaaand now it built, wtf? xD
<pie_>
(i just ran it again, nothing else)
<betaboon>
hello #nixos. i got a question regarding nixops. is it possible to set a route53-record to a s3bucket ?
<pie_>
oh i cd-d into the dir
<pie_>
so, it works if i run it directly but fails if i use callpackage
ntqz has joined #nixos
<hyper_ch2>
WD is closing factory.. but I do like WD-spinning rust for mass storage
ntqz has quit [Read error: Connection reset by peer]
<slabity[m]>
I'm trying to learn NixOps, but I'm a bit confused. In section 3.1 of the manual (Deploying to VirtualBox) there are two separate files, `trivial.nix` and `trivial-vbox.nix`. How exactly does NixOps handle this? Does it just merge the two `webserver` sets together? Could I use one file with all the information? Can I use more than two?
<slabity[m]>
I understand one is the logical configuration and the other is the physical deployment, but is this enforced?
<gchristensen>
slabity[m]: those two configurations merge, yes
wpcarro has joined #nixos
<gchristensen>
you could use as many files as you want :) as long as you create them like "nixops create ./trivial.nix ./trivial-vbox.nix ./your-other-files.nix -d trivial"
<slabity[m]>
Thanks. That makes a lot more sense.
knupfer has quit [Ping timeout: 264 seconds]
<clever>
gchristensen: nixops also allows you to make a tree of files, similar to imports in nixos
<gchristensen>
clever: I know, but not trying to overwhelm them :P
<manveru>
:D
<nh2>
dtz: OK cool thanks for clearing it up
hyper_ch2 has quit [Quit: Page closed]
<slabity[m]>
A tree of files? Like you just do `nixops create ./my-dir -d system` and it will load all the files in? Or just using `imports = [ ... ]` like normal?
<{^_^}>
[nixpkgs] @7c6f434c merged pull request #43624 → treewide: either fix or disable tests → https://git.io/fNOS8
<clever>
slabity[m]: you can put require = [ ./trivial.nix ./trivial-vbox.nix ./your-other-files.nix ]; into a single nix file, and then run `nixops create` on that single file
wpcarro has quit [Ping timeout: 268 seconds]
<clever>
and its recursive like imports, so you can make a tree of files, that require more
<slabity[m]>
Oh cool. I can definitely incorporate that.
ntqz has quit [Read error: Connection reset by peer]
spear2 has quit [Remote host closed the connection]
Guest84592 has joined #nixos
<slabity[m]>
Now to just... Slowly switch every machine I have to nix
<slabity[m]>
Probably take a few weeks
<gchristensen>
:D
<clever>
i still have a gentoo in the house, and a few ubuntu in the cloud
<Guest84592>
Hey all, I am attempting to use nix with buildkite for CI, but I am getting this error: error: file 'nixpkgs' was not found in the Nix search path
Guest84592 is now known as drbrule
<clever>
drbrule: NIX_PATH is blank by default, you need to set it yourself
<drbrule>
hmm
<drbrule>
not sure what to set it to/
<clever>
drbrule: and you really want to set it to a specific revision, so the nixpkgs doesnt change without warning
<pie_>
the one thing i dont like about nix is it kind of seems like lots of magic incantations
<clever>
drbrule: your bootstrap nixpkgs needs a NIX_PATH
<d1rewolf_>
pie_, as someone on the verge of switching over to nix, could you give a for example?
<drbrule>
what does that do?
<d1rewolf_>
I'd like to go in eyes wide open ;)
<clever>
drbrule: it loads nixpkgs from $NIX_PATH
<clever>
drbrule: then uses that to download the nixpkgs defined in the json
<nh2>
dtz: for the couple Haskell exes I'm trying to build statically now, the only missing piece seems to be linking sqlite statically. The nix package doesn't have libsqlite3.a, if I build it manually with autoconf on Ubuntu, it's in `lib/`. Any idea what disables it or skips copying it?
<drbrule>
but nixpkgs isn't downloaded yet, right?
<drbrule>
so why would specifying a path help?
<pie_>
d1rewolf_, i just feel like that sometimes :p
Tobba has quit [Remote host closed the connection]
<clever>
drbrule: if you do the export i gave above, then it will find the bootstrap nixpkgs
<d1rewolf_>
pie_, what do you find most magic about it?
ntqz has joined #nixos
tzemanovic has quit [Remote host closed the connection]
Tobba has joined #nixos
tzemanovic has joined #nixos
<logzet>
It is possible to enable TRIM for swaps by adding the "discard" option to the fstab, how do I do this on nixos? SwapDevices does not have a field for mount options…
<d1rewolf_>
I really wish that nixos used something aside from Haskell. I understand the functional nature of haskell probably helps, but learning haskell just to hack on nixos is meh :-/
<drbrule>
ok so I set an env var NIX_PATH to what?
<gchristensen>
d1rewolf_: nixos doesn't use haskell
<clever>
2018-07-17 11:53:19 < clever> drbrule: something like this would be a good start: export NIX_PATH=nixpkgs=https://github.com/nixos/nixpkgs/archive/dae9cf6106d.tar.gz
tzemanovic has quit [Remote host closed the connection]
<drbrule>
ahhhh
<drbrule>
ok
<clever>
nh2: by default, nixos doesnt do very much static linking
tzemanovic has joined #nixos
<d1rewolf_>
gchristensen, really? I'm sorely mistaken then
<drbrule>
so can I set this to rev?
<d1rewolf_>
I could've sworn this was the case
<drbrule>
or master
<slabity[m]>
logzet: There's a variable called `services.fstrim.enable`
ntqz has quit [Read error: Connection reset by peer]
<clever>
drbrule: its best to set it to a single rev, so it doesnt change
<drbrule>
hmm ok
tzemanovic has quit [Read error: Connection reset by peer]
<slabity[m]>
logzet: It will automatically take care of trimming SSDs
<nh2>
clever: but there must be something that turns the generation of the `.a` file off manually, because it's not even in the .libs build directory after `make` on Ubuntu
tzemanovic has joined #nixos
<nh2>
sorry I mean on nix
<nh2>
and it is there on Ubuntu
<drbrule>
One last question, should I pick a rev through github?
<drbrule>
Not sure where to choose that link
tzemanovic has quit [Remote host closed the connection]
<clever>
nh2: let me find it...
tzemanovic has joined #nixos
<clever>
drbrule: i put the latest rev from nixos-unstable in the url
<d1rewolf_>
gchristensen, I'm not sure where it happened, but somehow on my journey to nixos I developed the notion that it used haskell. I guess I was completely wrong :)
<drbrule>
yeah but I need to know how to update in the future
<nh2>
the configure flags don't really look static/dynamic realted
<nh2>
related
tzemanovic has quit [Remote host closed the connection]
tzemanovic has joined #nixos
<clever>
nh2: default flags in the stdenv
<gchristensen>
d1rewolf_: :) nope, no haskell! also -- just like I tell the haskell users about learning C++, you, too, can learn Haskell! you are smart enough and capable enough.
tzemanovic has quit [Remote host closed the connection]
<pie_>
gchristensen, so... nix-shell -v -I "nixpkgs=https://nixos.org/channels/nixos-unstable/nixexprs.tar.xz" -p "with import <nixpkgs> {}; pkgs.callPackage ./hashcons/shell.nix {}" fails, but running the shell.nix directly works fine??
<drbrule>
just trying to fiigure out where to find a listing of tarballs to set NIX_PATH to
<clever>
drbrule: the download link in the top right corner
<nh2>
clever: what you said should certainly work on the stdenv level (makeStaticLibraries env) but that would make it build *only* static libs. I'd expect I could also just set `dontDisableStatic = true;`
<drbrule>
got it, thank you I really appreciate it
<gchristensen>
yep
<clever>
i find it "simpler" to just memorize every type of URL github can do, and shove revs in as needed, lol
<gchristensen>
and if you ever want to update that URL, you can follow the directions and they'll work later :) (though they'll change every 6mo as new stable releases are made, so in a few months it'll be 18.09 etc.
stepcut has quit [Remote host closed the connection]
ntqz has quit [Read error: Connection reset by peer]
<rydnr>
I have an old box (nix 1.10). I don't know how to upgrade. nixos-rebuild switch fails with error: current Nix store schema is version 10, but I only support 7
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #43670 → Recurse on hydra → https://git.io/fNsZ0
ntqz has joined #nixos
ntqz has quit [Read error: Connection reset by peer]
<betaboon>
does anyone know how to use a s3bucket in a route53-recordset using nixops ?
Neo-- has joined #nixos
ntqz has joined #nixos
ntqz has quit [Read error: Connection reset by peer]
udkyo has quit [Quit: udkyo]
justbeingglad has joined #nixos
justbeingglad has left #nixos [#nixos]
ntqz has joined #nixos
ntqz has quit [Read error: Connection reset by peer]
ntqz has joined #nixos
ntqz has quit [Read error: Connection reset by peer]
ntqz has quit [Read error: Connection reset by peer]
Tucky has quit [Quit: WeeChat 2.1]
<rydnr>
clever: it finished with no error. However, running nixos-rebuild switch gives me the error "current Nix store schema is version 10, but I only support 7". I'm running now "_NIXOS_REBUILD_REEXEC=1 nixos-rebuild switch".
<clever>
rydnr: `nixos-rebuidl boot` sets up grub to run the new version when you reboot
<clever>
so you need to reboot next
<rydnr>
ah ok
<clever>
switch can get funky with such major version changes
<rydnr>
clever: another question, if I may. I've been using nixos from some time now, but I still find intimidating using nix-shell and building custom packages. I read some blog posts some time ago. Any suggestions on where to start (again)?
<clever>
rydnr: it may help to read something like <setup.sh> and learn how the stdenv works
tzemanovic has joined #nixos
<rydnr>
clever: setup.sh ?
<infinisil>
clever: That seems to be broken whoops
<{^_^}>
[nixpkgs] @benley pushed commit from @brainrape to master « keybase-gui: fix kbfs mount path check (#40080) »: https://git.io/fNsCH
<rizary>
anyone use nix on bash on windows ever occured the "warning: SQLite database '/nix/var/nix/db/db.sqlite' is busy (SQLITE_PROTOCOL)" loop?
ntqz has joined #nixos
ntqz has quit [Read error: Connection reset by peer]
<benley>
oooh does nix-on-bash-on-windows actually work at all now?
<rizary>
idk, i just remember i have installed it on my windows os. I remember back then I can install some packages (but i'm not sure), so I try it again.
<gchristensen>
apparently so!
<rizary>
and got that loop.
<gchristensen>
is there a emacs nix-mode which works in let blocks?
<dtz>
nh2: if you didn't find already a) you can set "dontDisableStatic = true;" to disable the generic builder's --disable-static , and b)there might be some helper that does that for you although honestly might as well just overrideDerivation yourself
<gchristensen>
right now I just use text-mode whenever I'm in a let block and it isn't ideal
ntqz has joined #nixos
adamCS has joined #nixos
<nh2>
dtz clever: No I already succeeded, used `sqlite_static = (pkgs.sqlite.overrideAttrs (old: { dontDisableStatic = true; })).out;` in this case and managed to build another huge Haskell app statically :D day success
Neo-- has quit [Ping timeout: 240 seconds]
<dtz>
dtz<3 yep
* dtz
can never remember which of overrideDerivation / overrideAttrs is the deprecated one lol
<dtz>
errr nh2 <3 yep
<dtz>
lol
<dtz>
one of them is I thought :3
<dtz>
anyway \o/
Havvy has joined #nixos
<rydnr>
infinisil: thank you. Annotated.
<dtz>
oh maybe it's overrideDerivation :3. idk.
isHavvy has quit [Ping timeout: 256 seconds]
<nh2>
dtz: "overrideAttrs" is the new good one as far as I remember
<dtz>
okay :D ty
ntqz has quit [Read error: Connection reset by peer]
<dtz>
:P (by `rg overrideDerivation |wc -l` and such)
<gchristensen>
oh heck yes!
sigmundv has quit [Ping timeout: 240 seconds]
<iqubic>
What benefits would the average user gain from Hydra?
__Sander__ has quit [Quit: Konversation terminated!]
<gchristensen>
none
<benley>
none unless you want your own CI server
<benley>
like if you would consider running Jenkins for your own use, you might have a use for Hydra
ntqz has joined #nixos
<iqubic>
Why would someone decide to use Hydra in the first place? What benefits does it give you?
<gchristensen>
Hydra is a Nix-based continuous build system, released under the terms of the GNU GPLv3 or (at your option) any later version. It continuously checks out sources of software projects from version management systems to build, test and release them. The build tasks are described using Nix expressions. This allows a Hydra build task to specify all the dependencies needed to build or test a project. It supports a
<gchristensen>
number of operating systems, such as various GNU/Linux flavours, Mac OS X, and Windows.
qwerty_asdf has joined #nixos
<iqubic>
Why would one want a continuous build system. This confuses me.
<gchristensen>
if you build 5,000 packages of your own every day you probably don't want to run nix-build yourself
<iqubic>
So as an average user I don't need Hydra?
ntqz has quit [Client Quit]
<gchristensen>
15:58 <iqubic> What benefits would the average user gain from Hydra?
<gchristensen>
15:58 <gchristensen> none
<gchristensen>
:)
<iqubic>
cool. Thanks.
kim0 has quit [Quit: Connection closed for inactivity]
<{^_^}>
[nixpkgs] @benley opened pull request #43677 → keybase-gui: fix kbfs mount path check (#40080) → https://git.io/fNslr
<dhess>
Anyone around who's using ZFS+encryption on NixOS?
<infinisil>
,ask
<{^_^}>
Just ask your question. It's the best way to know if anybody can help.
<dhess>
That's a bit patronizing. :\
<gchristensen>
dhess: hopefully not ZFS's encryption :)
<benley>
dhess: I am
<benley>
dhess: seems to work!
<dhess>
gchristensen: well that's one of the reasons I'm asking. Not sure how stable it is.
<gchristensen>
not stable:)
<infinisil>
dhess: sorry but it's a very common thing on IRC :)
<dhess>
benley: have you used LUKS before with NixOS?
<gchristensen>
from #zfsonlinux's topic: "Native encryption is not production ready, keep backups (but it works great)"
<benley>
yes, I used to use LUKS with ZFS on top of it
<dhess>
benley: cool and do you use NixOps as well?
<benley>
For the past few months I've switched to native zfs encryption on my laptop for the root filesystem and so far it's going well
dbmikus has quit [Ping timeout: 265 seconds]
<benley>
Nope sorry, I don't use NixOps
<dhess>
benley: OK, thanks. I want to use it on a fileserver, so it needs to boot unattended and then mount the encrypted ZFS filesystem when the key(s) are deployed to /run/keys, like autoluks does with NixOps.
<gchristensen>
dhess: will you have good backups?
<dhess>
gchristensen: yes sir.
<gchristensen>
good :)
<benley>
I would be cautious about using it for a fileserver unless there's a _real good_ backup plan :-)
<qwerty_asdf>
This may be real dumb, but is there is a difference between nixos.unstable & nixpkgs.unstable?
<benley>
on my laptop I just accept the risk that it may explode :-P
<benley>
qwerty_asdf: different set of tests required to pass before the channel gets updated
<gchristensen>
,which channel
<Taneb>
iqubic: as an example of where someone would use Hydra, at work we use Hydra as our build server. It automatically builds pull requests and runs tests, and GitHub waits for it to approve stuff before it gets merged. It also builds and runs a hoogle server
<benley>
the one big downside for me at this point is I can't just use zfs send | ssh ... zfs receive to do backups, because my fileserver is not running the unstable zfs with encryption support
wchresta has joined #nixos
<qwerty_asdf>
thanks a ton benley gchristensen :)
<gchristensen>
benley: not to mention the part that sort of backup wouldn't be sufficient, probably, for a not production ready FS feature
<infinisil>
gchristensen: Huh, either I didn't port it from factoids or I overwrote it with ,channels or something else
<benley>
gchristensen: oh yeah, that's a very good point
<dhess>
benley: oh, so with zfs send/receive, if the source filesystem is encrypted then the destination one must also be?
<benley>
dhess: yeah, I believe so.
wpcarro has joined #nixos
<dhess>
huh, that's surprising.
<benley>
dhess: it doesn't decrypt during transmission, it just sends the blocks, and the receiving end needs to know what to do with it
d1rewolf_ has joined #nixos
<dhess>
I would expect it should be an option to decrypt during transmission, anyway.
<dhess>
maybe that's in the works.
<gchristensen>
I doubt it
<benley>
It's been a couple of months since I read up on this stuff - maybe
<gchristensen>
the reason send/receive is fast and nice is its just a diff of trees
<d1rewolf_>
gchristensen, following your "nixos on dell" to get luks working. When I list /dev/mapper, I only see "control"...no other files/drives. Any idea why?
<benley>
anyway at the moment I mostly rely on dropbox / gdrive / github to sync data off of my laptop instead of having traditional Proper Backups of it
<dhess>
gchristensen: ok that's a good point, but it should still be available an option, IMO.
xy2_ has joined #nixos
<gchristensen>
d1rewolf_: sounds like you didn't implement this: $ cryptsetup luksFormat /dev/nvme0n1p2
<infinisil>
dhess: Maybe ask #zfsonlinux, they know stuff!
humanoyd has quit [Quit: WeeChat 2.1]
<dhess>
infinisil: Thanks. Most of my immediate questions are regarding the integration with NixOS and NixOps, specifically, so this seemed like a better place to start.
xy2_ has quit [Client Quit]
<d1rewolf_>
gchristensen, ah, yes. luksOpen does make my encrypted partition show up there. thanks.
<d1rewolf_>
but only that partition and control. no other drives. perhaps that's no big deal
* d1rewolf_
continues to read
wpcarro has quit [Ping timeout: 265 seconds]
wchresta has quit [Ping timeout: 240 seconds]
<infinisil>
dhess: Yeah, I mean more in regards to what stuff is in development
<d1rewolf_>
gchristensen, out of curiosity, what's nix-shell -p emacs do versus just running emacs directly?
civodul has quit [Quit: ERC (IRC client for Emacs 26.1)]
xy2_ has joined #nixos
<gchristensen>
I don't have emacs installed in that article
<d1rewolf_>
gchristensen, where would be the best to go to learn the differences of using nix-shell vs nix-env?
stepcut has joined #nixos
<gchristensen>
probably the nix manual
<benley>
the change is merged to master and I'm pretty sure it belongs in release-18.03, I'm just not sure whose approval I ought to seek before merging it there.
<d1rewolf_>
gchristensen, probably right ;-) sorry for the silly question
<milanos>
I a bit confused on how to configure sudoers.d. I'm looking at the configuration.nix manual and I dont know how to implement the following commands:
<milanos>
<user> ALL = NOPASSWD: /path/to/alcove/priv/alcove and Defaults!/path/to/alcove/priv/alcove !requiretty
wpcarro has quit [Ping timeout: 240 seconds]
cxkzm has quit [Ping timeout: 260 seconds]
<aminechikhaoui>
milanos: you should be able to set that in security.sudo.configFile
<d1rewolf_>
gchristensen, so when you say you don't use nix-env, you're saying just at that point in the article you don't have emacs installed, so instead of doing an nix-env -i emacs, you use nix-shell -p emacs to download a binary cache of emac and run temporarily.
<d1rewolf_>
is that correct?
Mateon3 has joined #nixos
<gchristensen>
right
<gchristensen>
but also, I don't use nix-env ever, really :)
Coyin has quit [Ping timeout: 256 seconds]
<LnL>
I bet you do, you just don't realize ;)
<d1rewolf_>
gchristensen, really? that's interesting. so what...you just manipulate your config and rebuild every time you want to have a package stay permanently, but otherwise you use nix-shell -p to run it temporarily?
<gchristensen>
exactly, d1rewolf_ :)
wpcarro has joined #nixos
<d1rewolf_>
is there a benefit of operating in this way or a drawback to using nix-env?
Mateon1 has quit [Ping timeout: 268 seconds]
Mateon3 is now known as Mateon1
<gchristensen>
I find nix-env confusing and frustrating
<d1rewolf_>
if you install a package with nix-env, does it handle manipulating your configuration.nix for you so that it will be reproduced along with your environment, or is it just installed locally with no changes to configuration.nix?
<gchristensen>
just local, no change
<LnL>
the only things in my user profile are packages that are in the 'trial period'
<d1rewolf_>
ok, so the benefit of doing it your way, gchristensen, is that you always know you can reproduce your environment with just your configuration.nix
<Dezgeg>
dhess: yes
<gchristensen>
exactly, Dezgeg
<gchristensen>
d1rewolf_:
<d1rewolf_>
LnL, by profile, you mean packages installed through nix-env and not part of your more permanent config? so the workflow on your end is to nix-env install them, but if you continue to use, add them to your configuration.nix and rebuild?
<d1rewolf_>
gchristensen, cool. that's exactly what I want, so I will travel down your path ;)
<clever>
d1rewolf_: and also remove them from nix-env
<betaboon>
hm. nixops is not currently capable of setting route53recordsets to point to a s3 bucket
<d1rewolf_>
clever, ah...so one you add them to your configuration.nix, you should nix-env -e them before rebuild?
<clever>
d1rewolf_: you can also `nix-env -q` to list what is currently installed in the user profile
<clever>
d1rewolf_: and roots profile is also available to all users
<d1rewolf_>
clever, ok, great. thanks. is there a drawback/risk of adding it to the configuration.nix, rebuilding, yet leaving it also installed with nix-env?
<LnL>
d1rewolf: yep, and I don't really install project specific packages either
wpcarro has quit [Ping timeout: 244 seconds]
<clever>
d1rewolf_: the one in nix-env has priority, so you wont be getting any updates configuration.nix brings in
<d1rewolf_>
ah, k.
<d1rewolf_>
sweetness. I can't wait until I have my configurations figured out so I can install ;-)
<d1rewolf_>
thanks guys
<d1rewolf_>
(and gals)
<gchristensen>
d1rewolf_++ have fun!
<d1rewolf_>
gchristensen, already am ;)
sbdchd has joined #nixos
<nek0>
infinisil: Do you know some c#?
<gchristensen>
clever: remember that weird mystery problem yesterday with the hashes?
<clever>
gchristensen: yeah
<gchristensen>
clever: my host and the build host disagree about the hash, but neither have an all-zero hash
wpcarro has joined #nixos
<clever>
gchristensen: can you run `nix-store --query --hash` on the problem storepath, on both machines?
<gchristensen>
clever: but it is building a FS image for OpenStack with what seems to be a VM, and the mismatch error comes when copying from the remote builder to the VM inside of it
<gchristensen>
copying to "local" is happening inside a nix-build on the remote linuxkit builder
<infinisil>
nek0: You could open an issue for it
<clever>
gchristensen: and if you run this on the linuxkit machine: nix-store --query --hash /nix/store/2kcrj1ksd2a14bm5sky182fv2xwfhfap-glibc-2.26-131
<clever>
gchristensen: what does it return?
<nek0>
infinisil: yes, but the maintainer of the package is unwilling/unable to do something about it.
<infinisil>
Is there an issue already?
<gchristensen>
clever: on the remote, sha256:1advm2fxdvi... on my mac machine, sha256:1l5kld8dcqjw...
<{^_^}>
[nixpkgs] @leo60228 opened pull request #43680 → msbuild: init at 15.6 → https://git.io/fNsuZ
<clever>
gchristensen: i suspect that either the remote slave has a corrupt copy of glibc, or you built your own before hydra, and the build is not deterministic
<dhess>
gchristensen: when linuxkit runs a VM, I assume that VM is software-only, i.e., no vmx or similar support from the CPU?
<clever>
gchristensen: try to rsync them into the same machine, and then run `diff -ru` on the 2 to see how they differ
<infinisil>
nek0: I see, well let's just hope somebody who knows how to solve the error finds it :)
infandum has joined #nixos
<d1rewolf_>
gchristensen, as your kernel boots, you get prompted for your luks passphrase? I have followed your instructions, but I don't get prompted and the filesystem never gets mounted, even though it's in hardware-configuration.nix
<infandum>
I'm having some trouble setting up xmonad as a window manager in nixos. I'm new to nixos. I have my ~/.xmonad/xmonad.hs set up and the xmonad.enable and xmonad.enableContribAndExtras set to true, but when I try to compile my xmonad config it says that it could not find the module XMonad etc (all of the other modules). I've looked everywhere online and haven't found a working solution.
klntsky has quit [Remote host closed the connection]
<srhb>
infandum: less $(which xmonad) might be enlightening. It's a wrapper that has its own ghc with its own pkgdb
<infandum>
infinisil, srhb: I'm doing xmonad --recompile and thats the error that pops up (Could not find module 'XMonad', could not find module 'XMonad.Action.CopyWindow, etc.)
<davidak>
should i do a nix-channel --update before nixos-install?
<srhb>
infandum: which xmonad, please
halfbit has left #nixos ["WeeChat 2.1"]
<srhb>
infandum: Sounds like you may have installed a different xmonad.
<infinisil>
infandum: And you did a nixos-rebuild switch?
<infandum>
Yeah
<cocreature>
nek0: no idea about that sry, I don’t use mono
<infandum>
I want to reiterate that those two values are the only xmonad and haskell values in my configuration.nix
<qwerty_asdf>
infandum: did you define your extraPackages in your configuration.nix? ie. windowManager.extraPackages = haskellPackages: [ ... <packages> ];
<qwerty_asdf>
ill share a pastebin of what that piece of configuration looks like if you want
<clever>
gchristensen: any update?
<infinisil>
qwerty_asdf: It shouldn't be necessary to import the xmonad module, should be included by default
<gchristensen>
clever: blocked on two things: my hands ache, and the remote doesn't have `scp` . working on it :)
wpcarro has quit [Ping timeout: 256 seconds]
<infandum>
qwerty_asdf: No I don't have that. I tried that with the xmonad packages in the past according to the internet but that did not work either
<clever>
gchristensen: ah
<srhb>
infandum: Just to confirm, that path is not actually a binary, right?
<d1rewolf_>
gchristensen, never mind. think i've got it
<srhb>
infandum: It should be a small wrapper script
<srhb>
infandum: bash, that is
<infandum>
srhb: let me check
<clever>
gchristensen: one option is to tar it up on the remote end, then `ssh remote 'cat foo.tar' > foo.bar`
<infandum>
srhb: It's a bash script
<infandum>
Also, I'm on the stable channel
<qwerty_asdf>
hmm, i had to I think ... I'm going to check if I actually need the explicit imports
<infandum>
The latest one
<srhb>
infandum: Puzzling. The extraContrib... option ought to make sure that ghc pkgdb has those packages.
<infinisil>
infandum: What's in the script? Can you gist it?
<srhb>
And of course, xmonad should definitely be there.
<infandum>
srhb: It's not just extra, though it can't even find the XMonad module
<d1rewolf_>
hmm...now I get prompted for the luks key and it takes it, but then the boot stops at "Started Hardware RNG Entropy Gatherer Daemon". Any ideas?
<infandum>
infinisil: It would be a little difficult to copy from the computer, but it has two exports with NIX_GHC and XMONAD_XMESSAGE with an exec in a nix store bin xmonad
<gchristensen>
clever: only diff is the copy has locales and the remote doesn't
<clever>
gchristensen: id run `nix-store --delete` on the remote one, and then try the build again
<infinisil>
infandum: Well it would be possible with the gist script, but that won't help much. See the path in NIX_GHC, run the command `/nix/store/<thehash>/bin/ghc-pkg list`
<infandum>
Wait a second, I have a build script in my .xmonad for when I have it on arch, it isn't looking at that, is it?
winem_ has quit [Ping timeout: 244 seconds]
wpcarro has joined #nixos
<infinisil>
Well it depends on xmonad, nix doesn't do anything special
<infandum>
Omg that was it
reinzelmann has joined #nixos
<infandum>
I forgot about that script, didn't realize xmonad automatically looks at it
<srhb>
It was trying to run ghc from within the program itself? ah
<srhb>
That makes sense.
<infandum>
In my arch installation I have a build script that xmonad can use to rebuild itself with stack
sw4n has quit [Ping timeout: 276 seconds]
<infandum>
I forgot it uses it automatically if it's in .xmonad, I guess
<infinisil>
srhb: That could probably be patched in xmonad to work with nix's ghc
<infandum>
so it preferred that over nix
<srhb>
infinisil: Yes, but why tough
<srhb>
infinisil: :)
<infinisil>
srhb: Why what?
<srhb>
Why patch it.
<iqubic>
What are we trying to do here?
<srhb>
It doesn't add anything to the nix functionality.
<infinisil>
srhb: It just removes a hurdle from other distros, and xmonad is patched to work with nix already :)
<srhb>
infinisil: Is it? I didn't realize.
<iqubic>
Why did you have to patch xmonad to work with nix?
<infandum>
Nice, everything works now, thanks everyone!
<infinisil>
Hmm okay I guess it might be harder to patch that build script
<infinisil>
:)
<{^_^}>
[nixpkgs] @dtzWill opened pull request #43683 → Revert "unixtools: link instead of copying" → https://git.io/fNsz5
wpcarro has quit [Ping timeout: 244 seconds]
<infinisil>
I should finish my nix builder for xmonad configs, then you won't have to invoke --recompile and you'll get reproducible builds
<infinisil>
And no dependency on ghc at runtime
<d1rewolf_>
I'm trying to mount a luks partition at boot. I have it working, and it prompts me for the passphrase, but then the boot doesn't progress past "Started Hardware RNG Entropy Gatherer Daemon". how should I troubleshoot this?
wpcarro has joined #nixos
<d1rewolf_>
ctrl+alt+fx to a tty doesn't seem to be working
<infinisil>
Lol, this code triage thing gives a 404 when you click on the "Send me a new issue" link and have javascript disabled
<infinisil>
And it sends an email instead of just redirecting you to the issue..
logzet has quit [Ping timeout: 265 seconds]
logzet_ has joined #nixos
<d1rewolf_>
although I'm in a virtualbox vm, so it may mucking with the ctrl+alt+fx
stepcut has joined #nixos
<d1rewolf_>
booting a previous configuration which doesn't mount the partition works fine
<clever>
d1rewolf_: virtualbox may have checkboxes in the menu to hold ctrl and alt
<clever>
d1rewolf_: dang, i checked and it doesnt, only options to insert some combos
<d1rewolf_>
clever, i booted up in another configuration. right "ctrl", which in virtualbox represents "ctrl+alt", plus a fn key, successfully switches to a tty when booted into the working configu
<clever>
d1rewolf_: but if you have a dedicated usb keyboard, you could forward that in
<d1rewolf_>
in the config which hangs on boot though, it doesn't, which leads to be believe the vm may be frozen in that state
<clever>
d1rewolf_: ah
<d1rewolf_>
what would be the best way to get error messages? doesn't journalctl give you some sort of way to see errors from previous boots?
<infandum>
Do you recommend running on the stable or unstable in configuration.nix
<d1rewolf_>
or is there somethign nix-specific I could run to determine what's going on?
<clever>
d1rewolf_: `journalctl -b -1` will filter it to the prebious boot
<d1rewolf_>
root
<d1rewolf_>
clever, I'll try that. thanks!
<samueldr>
for ctrl+alt+Fx in virtualbox, old the modifier key (default to right alt) and Fx, so RALT+F1
<d1rewolf_>
booted up again. it made it past RNG, and froze after "Reached target Swap"
<d1rewolf_>
samueldr, on mine that's right ctrl
<samueldr>
to see if the modifier works, RALT+HOME should show the menu (unless it's only when the menubar is hidden?)
<samueldr>
oh, rightctrl I mean duh :)
<samueldr>
it's that on mine too
<d1rewolf_>
it will work in the properly booting config, but not in the borked one :-/
<clever>
d1rewolf_: what changes where made to the config?
tzemanovic has joined #nixos
<d1rewolf_>
clever, just adding the luks partition. I'm tracking down the old config to diff
<clever>
d1rewolf_: can you screenshot it when its hung and upload that somewhere?
<d1rewolf_>
clever, sure
<d1rewolf_>
where are old configs stored?
<clever>
d1rewolf_: they arent, thats why i keep my config in a git repo and also have zfs snapshots
<d1rewolf_>
clever, thanks very much. Have a 2PM meeting but will try right after
<{^_^}>
[nixpkgs] @matthewbauer merged pull request #43161 → config.skipAliases: also disable vim, git, xfce and gnome3 aliases → https://git.io/fNvKw
<{^_^}>
[nixpkgs] @matthewbauer pushed 4 commits to master: https://git.io/fNsa3
xy2_ has quit [Ping timeout: 260 seconds]
<acowley>
clever: Thank you! Can I list that in a patches attribute?
neeasade has quit [Ping timeout: 240 seconds]
cxkzm has joined #nixos
<clever>
acowley: you probably want to run fetchpatch over that url first
johnw has quit [Remote host closed the connection]
<clever>
grep nixpkgs for fetchpatch to find examples
Coyin has joined #nixos
<d1rewolf_>
clever, ah,...getting somewhere (had to try before meeting). I'm off now but will report later. thanks!
<gchristensen>
clever: however! nix-build is still failing with the 00's error on the vm inside the vm
<clever>
gchristensen: do you know if its running qemu inside linuxkit, to generate the nova image?
<gchristensen>
clever: it is
<clever>
gchristensen: the qemu uses 9plan to mount the host nix store, and a closure export to limit its db.sqlite to just the closure of that build
<clever>
gchristensen: and that closure export lacks hashes
<clever>
so it has to nix-store --verify --check-contents inside the build every single time
<gchristensen>
yeah that seems totally reasonable, but this same build works when I run it on a linux machine
<acowley>
clever: Thanks again, that's just what I needed!
<clever>
gchristensen: i have also run into 9plan bugs, where a directory has totally wrong contents, try to ls glibc inside the derivation, inside qemu, if you can
xy2_ has joined #nixos
Coyin has quit [Ping timeout: 244 seconds]
<gchristensen>
ehhhh
<gchristensen>
unfortunately I've run out of time on this on
<dhess>
oh god, there's another project named linuxkit?
<gchristensen>
dhess: huh?
Sonarpulse has quit [Ping timeout: 276 seconds]
<dhess>
gchristensen: uhh.. never mind maybe? I think I'm confused. I thought that puffnfresh's linuxkit thing was different
<gchristensen>
dhess: he started it, I picked it up, we pushed it to the finish line together :)
<dhess>
gchristensen: I was confusing linuxkit with hyperkit
<gchristensen>
ah
<dhess>
maybe. I'm not sure actually.
derped has joined #nixos
<dhess>
gchristensen: so maybe you can clarify for me. This linuxkit-builder thing is intended to be used as a remote builder for Nix/nixpkgs on Darwin, right?
<gchristensen>
yeah
<dhess>
and it uses hyperkit for virtualization, right?
<orbekk>
pstn: replace makes it idempotent. the routes are persistent across service restarts, so if you use ip route add it will fail the second time it runs
<gchristensen>
linuxkit calls hyperkit and vpnkit
<sphalerite>
ooooh, I didn't know about ip route replace
<sphalerite>
good to know. I always found it annoying using ip route add and then not being able to repeat the command and having to change it to "change" (which of course doesn't work if it doesn't already exist)
<orbekk>
sphalerite: yeah... ditto :)
<gchristensen>
dhess: with linuxkit and a few more LoC I can take this linuxkit builder and make it work on a windows machine
<gchristensen>
dhess: b/c linuxkit makes it easy to do the same thing on windows
<infinisil>
Hey, does anybody know how hackage-packages.nix gets its specific version numbers for certain packages?
<Pneumaticat>
would it be a good idea to run NixOps for a laptop? I find myself enticed by the idea of unifying my configuration, but also worried about if I need to install something while away from my home network, for example
<d1rewolf_>
clever, it's interesting that the error wouldn't show up on the main boot screen tho
<clever>
Pneumaticat: i run nixops on my laptop, and i have a VPN setup, so it can still reach all the machines in the cluster
<clever>
d1rewolf_: yeah, that is rather strange
<Pneumaticat>
clever: oh that's a good idea! Do you run the VPN on the laptop all the time, or only when you deploy config?
bennofs has joined #nixos
<clever>
Pneumaticat: at all times
<clever>
Pneumaticat: the vpn doesnt change the default gateway, so i can choose to use it or not on a per-connection basis
<clever>
Pneumaticat: i simply have 2 IP's that work for reaching some machines, and 1 IP always works
<Pneumaticat>
clever: oh, that's nice! I may try to replicate that ;)
<clever>
Pneumaticat: services.toxvpn = { enable = true; localip = "10.x.y.z"; }; and then run toxvpn-remote and its help sub-command to link the peers up
<Pneumaticat>
clever: oh wow, toxvpn looks cool! :B Thanks for the info
xy2_ has joined #nixos
<Pneumaticat>
I have a long way to ascend on the nix ladder ;)
ersran9 has quit [Ping timeout: 256 seconds]
<dhess>
gchristensen: in that case, may I humbly suggest that the project is poorly named ;)
<gchristensen>
dhess: 2018-05-29 16:50:57 gchristensen Nix friends, I have a project to create a small Linux VM on your macOS machine very simply. Right now it has a terrible name, "nix-linuxkit-builder". I need a snappy name. any suggesntions?
<gchristensen>
:P
Guest64931 has joined #nixos
<dhess>
hehe
freeman42x]NixOS has joined #nixos
<tilpner>
gchristensen - Did you not like the perfectly non-confusing lunixkit?
<dhess>
hyperkit-builder is better, though it is hard to come up with something that catpures exactly what it does, succinctly.
<gchristensen>
yea
<samueldr>
hyperkit ties it into the underlying tech, *-for-hyperkit and I want he shed to be purple
<samueldr>
the shed*
<samueldr>
(always depending if it's intended to be a more generic API that could be applied to other techs)
wpcarro has joined #nixos
Guest64931 has quit [Ping timeout: 276 seconds]
uwap has quit [Remote host closed the connection]
<d1rewolf_>
g0!!0l3!!m1ss
uwap has joined #nixos
<gchristensen>
uh oh
<samueldr>
d1rewolf_: need a couple password resets now probably
<d1rewolf_>
lol
<tilpner>
Mic92 - If nix-community/NUR#39 is merged, can I no longer include a GPLv3 script in my repo?
wpcarro has quit [Remote host closed the connection]
wpcarro has joined #nixos
knupfer has joined #nixos
<kalbasit>
how to work with nodejs on Nix? for instance I need to `npm install -g doctoc` but of course I get permission denied since the node_modules is located in the `/nix-store`
<d1rewolf_>
kalbasit, that must be an faq around here...i've see it a good bit
<d1rewolf_>
(don't know how to fix tho, sorry)
<samueldr>
kalbasit: either you configure npm so -g installs to your home or avoir -g
<samueldr>
avoid*
<slabity[m]>
If I have a file 'network.nix' that contains just '{ network.description = "My network"; }', and then run 'nixops create network.nix -d network', then why does it say the description is "Unnamed NixOps network"?
juhe has joined #nixos
<kalbasit>
I see
<samueldr>
I personally avoid -g, which creates a node_modules folder, within it there's the .bin folder, so for webpack it becomes ./node_modules/.bin/webpack
<samueldr>
(I'm not following node conventions, I add a "bin" directory to my projects with "binstubs" that redirect to those tools)
<cxkzm>
iset
<cxkzm>
woops ignore that
<kalbasit>
samueldr: I see, I'll probably do the same
<kalbasit>
thx
<samueldr>
(I'm tracking down the doc for -g in your home, for completion's sake)
<samueldr>
oh, and kalbasit, I'm always doing that inside a nix-shell
<juhe>
\o/, running nix-repl gives me "error: Nix database directory '/nix/var/nix/db' is not writable: Permission denied", should I run the nix-repl as root?
<{^_^}>
juhe: To use nix-repl with Nix 2.0, either use the new `nix repl` or `NIX_REMOTE=daemon nix-repl`. Just using `nix-repl` gives an error because it is linked to Nix 1.x which requires NIX-REMOTE to be set correctly, while Nix 2.0 doesn't (and unfortunately it wasn't kept for backwards compatibility)
<samueldr>
it was less verbose I think
wpcarro has joined #nixos
<juhe>
tilpner: thx!
<kalbasit>
samueldr: I need access to `doctoc` in my dotfiles to generate the table of contents, I'll probably write a script that uses `nix-shell` to pull node and install doctoc before using it
graphene has quit [Remote host closed the connection]
sir_guy_carleton has quit [Quit: WeeChat 2.0]
<samueldr>
oh, if it's not for a project, but for a tool to use in your day-to-day life, there's something else that should be done
* samueldr
will create the appropriate wiki page with the information
<kalbasit>
samueldr: true, but I only need it in my dotfiles, so it should be probably there. For my day to day, I'll probably follow the link I pasted above. I use yarn though so I still have some R&D for that
<kalbasit>
samueldr: that'd be great, it's a bit confusing for noobs like me
<samueldr>
oh, kalbasit, I usually use yarn too :)
<samueldr>
release.nix is for building using yarn2nix and nix
<d1rewolf_>
when I run "nixos-rebuild switch" I'm dropped to a prompt stating to enter my root password for maintenance...as if a boot had failed. How can i track down the error that's causing this? where does nixos-rebuild log to?
<kalbasit>
samueldr: awesome! thank you.
<slabity[m]>
Okay, NixOps does not seem to be reading any files I pass into it. Anyone run into this before?
<Mic92>
tilpner: Is the script part of the build description or is it going to be deployed?
<tilpner>
Mic92 - It is used as src, and then copied to $out/bin (and wrapped)
<Mic92>
tilpner: that should be ok.
<Mic92>
we have the same policy in nixpkgs
xy2_ has quit [Ping timeout: 260 seconds]
<tilpner>
Mic92 - In what case would it not have been okay? What if it was used by the builder?
<infinisil>
Mic92: You mean for building packages in ones repo without NUR? Then you probably want to do { pkgs ? import <nixpkgs> {}, lib ? pkgs.lib }
<hodapp>
huh, so it *is* possible to build Boost with multiple Python versions at once, but I'd have to hack up my own derivation to do it
<gchristensen>
clever: more weirdness!
<clever>
gchristensen: ?
<flexw>
:q
flexw has quit [Quit: leaving]
<gchristensen>
clever: I added a real linux machine as a target remote-builder, where I've run nix-build on the project locally andit worked.
<gchristensen>
clever: this time, iana-... failed with a mismatch. I copied all the source to the remote builder and ran nix-build there, and it worked fine.
<hodapp>
I guess I can modify pkgs/development/libraries/boost/generic.nix to do so and make it configurable, but I feel like this is a file that hundreds or thousands of things rely on so I'd have to be very careful
<clever>
gchristensen: weird
<hodapp>
anyone know of anything else in Nixpkgs that can use multiple Python versions simultaneously?
<gchristensen>
clever: any ideas? :$
<sphalerite>
hodapp: neovim I think?
<hodapp>
sphalerite: thanks, will take a look
<sphalerite>
not sure though
<clever>
gchristensen: is any of this building against master?
<Lisanna>
srhb wait... you don't need to put the script in $out/bin?
derped has quit [Ping timeout: 260 seconds]
magnetophon has joined #nixos
freeman42x[nix] has joined #nixos
<infinisil>
When I create a systemd timer that runs *hourly*, will it run right when it's installed the first time?
<infinisil>
Or will it wait for e.g. 10:00 oclock?
<clever>
infinisil: after setting it up, run `systemctl list-timers`
<infinisil>
clever: I'm not actually setting one up, just wanting to know when it will run
<gchristensen>
clever: you jinxed it and it works now :P
<clever>
infinisil: i think it will wait
<infinisil>
Oh yeah, hourly is just an alias for *:00:00 after all
<magnetophon>
I keep getting "error: while setting up the build environment: getting attributes of path '/etc/hosts': Permission denied", when I try to do a "nixos-rebuild" including certain packages, even with a minimal config. Full details here: https://github.com/NixOS/nixpkgs/issues/41651#issuecomment-405702448 Any ideas how to troubleshoot further?
<benley>
Welp, now I see why nixos-rebuild isn't using the new `nix build` command instead of nix-build
<sphalerite>
why is that?
<benley>
I went to try and convert it and discovered that it uses like _every possible feature_ of nix-build, and many of them are not available in `nix build`
<sphalerite>
like what?
<LnL>
benley: it's also not stable yet so scripts shouldn't use it
<gchristensen>
well that and also it ins't good to make...whatlnlsaid
<sphalerite>
LnL: but nixos-install uses it
<benley>
--add-root --indirect --no-out-link ... that might be it
<benley>
LnL: oh that's a good reason too
<benley>
anyway it will be cool when nixos-rebuild can use the new shiny
<gchristensen>
sphalerite: we know for certain which nix nixos-install is using
<sphalerite>
benley: --add-root --indirect is default behaviour for both nix build and nix-build
<gchristensen>
uglier with the nixos-rebuild I think
<LnL>
it is?
<LnL>
:/
<sphalerite>
gchristensen: nixos-rebuild builds its own nix…
* gchristensen
has no idea what he's talking about
<benley>
huh ok. Well, perhaps the real reason is for compatibility with nix 1.x -> 2.x transitions?
<gchristensen>
clever: it hasn't moved to any path at all
<clever>
gchristensen: it may have moved to a path not being copied
<gchristensen>
hrm. what is "it" that we're talking about?
<clever>
gchristensen: the problem
<{^_^}>
[nixpkgs] @pSub pushed to master « jdepend: add myself (pSub) as maintainer »: https://git.io/fNsDW
mkoenig has quit [Ping timeout: 240 seconds]
<ma27>
does anybody want to test this (https://github.com/NixOS/nixpkgs/pull/43690) with OSX and Nix? Upstream claims that they support XCode now, but I don't have the hardware to confirm this with Nix %)
mkoenig has joined #nixos
<infinisil>
Great, my link matcher doesn't work with ) after the link..
drakonis has joined #nixos
cryptomonoid has quit [Ping timeout: 240 seconds]
Guest64931 has quit [Ping timeout: 276 seconds]
<gchristensen>
clever: do you think lots of GC's everywhere might re-expose it?
Guest64931 has joined #nixos
<clever>
gchristensen: possibly
<gchristensen>
:(
<gchristensen>
this is a frustrating thing
<sphalerite>
benley: oh and nix build has --no-link too
<Dezgeg>
the all-zeroes hash?
<sphalerite>
but it's not useful because you can't get the path out of it
<sphalerite>
just using the result symlink is better anyway because of GC safety :)
<benley>
sphalerite: ha. I guess there's another command you could run to get the store path for the thing you built?
<gchristensen>
Dezgeg: yeah
<sphalerite>
benley: `readlink result`
<sphalerite>
:)
<Dezgeg>
it's probably coming from an old EC2 image or something
<benley>
lol ok yes
<juhe>
To fix kicad-unstable's ngspice simulation (libngspice.so library is not found), I see two possibilities, one is using wrapper and LD_LIBRARY_PATH, another one is to make a patch to kicad's sources to specify where to look for the library. Which one is better? (I tend to like the latter more.)
<gchristensen>
Dezgeg: I don't have any EC2 systems here, would that still be possible?
Tobba has quit [Read error: Connection reset by peer]
<Dezgeg>
some other vm image?
<gchristensen>
the remote builder, the openstack image, and my local machine are all built from nixos 18.03 v411cc559c5052fe...
<gchristensen>
minus that first v
drakonis has quit [Remote host closed the connection]
<infinisil>
Alright so I reverted nixbot to a state where it wasn't as fragile lol (the only feature it doesn't have anymore is the channels and automatic updates of nixpkgs), things shouldn't be laggy now, should respond rather quick. And I fixed the url thing as well. I'll let nixbot be like this for a while until I can figure out how to properly test and deploy software with haskell :)
* infinisil
will shut up about that bot now
crmlt has joined #nixos
<Dezgeg>
you should check the initial openstack image
drakonis has joined #nixos
<samueldr>
,runtimeDeps juhe
<{^_^}>
juhe: In order of preference: Patch source OR ((if it uses PATH -> wrap with new $PATH) AND (if it uses dlopen, (patchelf --set-rpath in postFixup OR wrap with new LD_LIBRARY_PATH)))
<samueldr>
so yeah, patch it if you can :)
<sphalerite>
manveru: I just figured… if you put the chrootDeps thing in <nix/config.nix> in your derivation somewhere it'll probably pull in the deps so you can use them for the botostraping
crmlt has quit [Client Quit]
kyren has quit [Ping timeout: 276 seconds]
kyren has joined #nixos
<sphalerite>
manveru: and maybe `builtins.storePath config.shell` would work as well, not sure though
<BlessJah>
is there good way to provide users PATH to systemd user unit?
flexw has joined #nixos
<andi->
BlessJah: $HOME/.nix-profile/bin ?
<BlessJah>
user unit that is a file dumped into .config/systemd/user, not created via configuration.nix
<andi->
not sure if that solves all the edge cases
<benley>
BlessJah: iirc you'll want to run `systemctl --user import-environment` somewhere
<BlessJah>
~> type cat
<BlessJah>
cat is /run/current-system/sw/bin/cat
<puffnfresh>
dhess: I originally wrote the builder to use HyperKIt but it was tricky to execute (e.g. you had to pass in the correct modules, do process management, open VPNKit, etc.)
<benley>
BlessJah: assuming you want to have $PATH set in the environment when the unit starts up
<andi->
benley is prbably right
<puffnfresh>
dhess: all LinuxKit does, is makes this easy
<flexw>
hello again, if I create a derivation with stdenv.mkDerivation in environment.systemPackages the derivation gets build but the reulsting executable is not in my path so that I can execute it in the shell. How can I get this to work?
<benley>
I've started putting things in services.xserver.displaymanager.sessionCommands to have commands run before systemd user units come up, for gui sessoin stuff
<manveru>
sphalerite: i think i'm finally grasping the low-level details of how nix works, after over 2 years or using it... :P
<sphalerite>
manveru: there is a lot to it :D I think I'm nearing my 2-year nix anniversary myself!
<BlessJah>
benley: import might have one big disadvantage: doesn't get populated until I log in (in case I wanted to run units after reboot)
<juhe>
samueldr: ok, thx!
davidak has quit [Quit: Leaving]
<manveru>
sphalerite: turns out teaching nix to other people actually made me look into the details for the first time... but after all i'm still fascinated by the complexity of nixpkgs
<sphalerite>
:)
kyren_ has joined #nixos
<benley>
BlessJah: very true
<benley>
BlessJah: one possibility could be installing the user units globally, but only activating them per user, like you can do with the nixos emacs module?
<BlessJah>
I'd preffer to keep it in $HOME
<BlessJah>
time to figure out how $PATH is populated in NixOS maybe?
<benley>
hm ... use one of the various nix homedir manager tools to generate your systemd unit file via nix so you can populate $PATH at that point?
<benley>
(or not need to populate it maybe)
kyren__ has joined #nixos
kyren has quit [Ping timeout: 276 seconds]
testuser has joined #nixos
<testuser>
hi, how can I run the nix-shell in order to get into the awesome-3-5 environment, so I can run lua, them do "local awful = require("awful")" ?
<manveru>
testuser: `nix-shell '<nixpkgs>' -A awesome-3-5`
<manveru>
sorry :)
<dhess>
oooh NixOS boots on my Jetson TX2 now.
<{^_^}>
[nixpkgs] @Ma27 opened pull request #43691 → termtie: factor wrapper out into its own file → https://git.io/fNsHm
<dhess>
puffnfresh: ok cool, thanks for the clarification.
<manveru>
hmm
logzet_ has quit [Remote host closed the connection]
magnetophon has quit [Ping timeout: 248 seconds]
<manveru>
testuser: doesn't seem to have awful
<andi->
dhess: cool, how much hackery was required? :-)
<dhess>
andi-: so far none, but I've only just gotten to the installer prompt.
<dhess>
let's connect ethernet and see if that comes up
<BlessJah>
nvm, prefixing commands with /bin/sh -c 'exec ...' should solve that
<manveru>
testuser: are you sure it's part of awesome?
<gchristensen>
Dezgeg: I made my own openstack image
rcshm has joined #nixos
<gchristensen>
Dezgeg: did something with these 0's end up in the cache?
doyougnu has joined #nixos
magnetophon has joined #nixos
coot has left #nixos ["Good Bye"]
ckauhaus has quit [Quit: Leaving.]
kerrhau has joined #nixos
kerrhau has quit [Changing host]
kerrhau has joined #nixos
<testuser>
manveru, yes it is
<manveru>
neither awful.lua nor awful.so are in the nix store
<testuser>
manveru, basically it's a folder inside /nix/*-awesome-3.5.9/, but the rc.lua seems to import it as I specified.
<testuser>
manveru, that much I'm aware, but I'm not sure how it's possible that the rc.lua is importing it with "local awful = require("awful")"
logzet has joined #nixos
mightybyte has joined #nixos
<manveru>
hmm
<manveru>
good question :)
tenten8401 has joined #nixos
<tenten8401>
Can't tell if this is NixOS specific or not, but is anyone else having issues creating 64-bit wineprefixes with wine staging? I feel like the verison packaged might be 32-bit only or something because with WINEARCH=win64 it still continues to create 32-bit prefixes
__monty__ has joined #nixos
wpcarro has quit [Remote host closed the connection]
<testuser>
manveru, I've extracted that tarball, but awful.so is not in it?
<manveru>
no
<sphalerite>
testuser: you may want to have a look at the awesome executable. It's probably a wrapper script that sets LUA_PATH or similar
<manveru>
i think it's the init.lua.in
<manveru>
maybe that gets compiled into something?
<sphalerite>
testuser: if you copy that you should probably be able to get awful in a lua REPL
<manveru>
i'm not very familiar with lua
<testuser>
manveru, me neither
<Dezgeg>
gchristensen: yes, check the nix database contents of the image, they are probably coming from there
<Dezgeg>
initial nix database contents, that is
<Dezgeg>
deploy a new vm from that image or something
wpcarro has quit [Ping timeout: 244 seconds]
<gchristensen>
hrm
<gchristensen>
seems like it should be impossible
<Dezgeg>
or mount the disk image
<gchristensen>
Dezgeg: does it make sense that it would sometimes work?
<gchristensen>
like it worked on machine A when I built directly on machine A, but didn't work on machine A when I used A as a remote builder for machine B
<andi->
With carnix is there a way to run tests of my project? I can't find a references to tests in buildRustCrate :/
<Dezgeg>
B could have had one of those affected store paths so A didn't have to build it
<tenten8401>
I have a sneaking suspicion that the NixOS wine version is 32-bit only
<tenten8401>
I've tried about everything I can find and it's refusing to make a 64-bit prefix
<manveru>
tenten8401: you used wineWow?
<tenten8401>
what's that?
<gchristensen>
Dezgeg: machine B is a macos machine :/
<tenten8401>
I'm using wineStaging currently
<Dezgeg>
the checksums might correct themselves when copying from A to B
<manveru>
testuser: and you're on a x86_64 system?
<Dezgeg>
not sure on that one
<mightybyte>
disasm: In the example you gave for ssh caching, why did you use "binaryCaches"? I'm doing "man nix.conf" and don't see that option listed.
<sphalerite>
testuser: as far as I can see the lua modules are in /nix/store/<hash>-awesome-4.2/share/awesome/lib
<sphalerite>
testuser: so if you run lua with LUA_PATH set to that maybe it will work
<tenten8401>
I feel like the wine package should really come with builds for a 64-bit system and a 32-bit system by default instead of having to override it
<tenten8401>
since the 64-bit build comes with 32-bit support as well
jmeredith has quit [Quit: Connection closed for inactivity]
<testuser>
sphalerite, I'm doing: "LUA_PATH="/nix/store/4ifqpn2c7g450h2kam5k33jj832s46ih-awesome-3.5.9/share/awesome/lib:$LUA_PATH" lua", however the require("awful") is still not found
<sphalerite>
testuser: I think lua uses a semicolon ; rather than : as the path separator, does it work with that instead?
dbmikus_ has quit [Quit: WeeChat 2.1]
aaron has joined #nixos
<testuser>
spacefrogg, yeah I figured, just testing now, it seems that it finds the path, but fails regardless: http://dpaste.com/3ESGBQK
aaron is now known as Guest75251
<infinisil>
testuser: Agreed, not sure why they don't
Guest75251 is now known as aanderse-laptop
<testuser>
sphalerite, seems like it didn't recognize init.lua, but finds the directory instead
<testuser>
sphalerite, seems like the path requieres "?.lua" at the end - weird
<infinisil>
tenten8401: Whoops that was meant for you ^^
tzemanovic has joined #nixos
<testuser>
sphalerite, that doesn't work either
<sphalerite>
testuser: the mysteries of lua!
<aanderse-laptop>
has anyone used mod_perl with apache on nixos? i'm trying to figure out how to properly set the environment apache will have with respect to perl includes
<manveru>
is there any easy way of getting the .drv corresponding to a nix path other than grepping all .drvs?
wpcarro has quit [Read error: Connection reset by peer]
wpcarro has joined #nixos
simukis has quit [Ping timeout: 244 seconds]
magnetophon has joined #nixos
kerrhau has quit [Ping timeout: 276 seconds]
<wpcarro>
I'm having some difficulty following along with the Nix Pills tutorial. Is it likely that some of the APIs have changed since that was written?
<clever>
manveru: nix-store --query --deriver
<manveru>
thanks :)
<clever>
wpcarro: what errors are you getting?
<sphalerite>
wpcarro: we try to keep it up to date. Are you following the version on the nixos website?
<LnL>
it has a tendency to retain references to the compiler, keeping everything in the runtime closure
<sphalerite>
wpcarro: right yeah that's the right one. What's not working?
<sphalerite>
clever: btw --max-depth can be abbreviated to -d as well. I usually write it as du -chd0 :)
<wpcarro>
sphalerite: let me try and reproduce some of the issues so I can be more specific. Most of the stuff has worked, but I'd say I've encountered ~5+ errors when trying to follow along
<wpcarro>
Apologies for not having more information at hand. Wanted to first verify that I should *expect* to be able to follow along
<sphalerite>
wpcarro: it's great to get feedback for this sort of thing so we can fix it :)
<sphalerite>
no worries
<manveru>
uh... wtf
<manveru>
yeah
<wpcarro>
sphalerite: what version of nix etc is that tutorial asserting on? Does it mention that somewhere that I missed?
<sphalerite>
manveru: yeah this is why buildGoPackage (iirc) uses removeReferencesTo to get rid of the reference to the go compiler
<wpcarro>
Looking like I'm on 2.0.4
<elvishjerricco>
I think Nix should have a concept of content hash dependencies. i.e. a store path should be able to declare that some of its runtime dependencies must have certain content hashes. If you try to substitute a store path but you have the wrong hashes on those dependencies, the path will be rejected, and you'll have to build from source.
<manveru>
so it includes not only the go compiler, but also gcc, perl, and python?
<LnL>
manveru: its the pkg stuff in GOPATH
<sphalerite>
wpcarro: it was written for a 1.x version but it should be updated for anything that's changed
<wpcarro>
sphalerite: ok nice... will report back
<elvishjerricco>
This would be useful for Haskell derivations, where building a package locally causes dependents substituted from a cache to break
<LnL>
manveru: and every version control system you can imagine for go get
<manveru>
fun
<manveru>
lemme see how this removeReferencesTo works
magnetophon has quit [Remote host closed the connection]
magnetophon has joined #nixos
<wpcarro>
Also had trouble running the `su - nix`. That part seemed unclear to me
<{^_^}>
[nixpkgs] @shlevy pushed to master « pythonPackages.simpy: init at 3.0.11. »: https://git.io/fNsNd
<sphalerite>
wpcarro: the `su - nix` is for logging in as the test user which… I think was meant to be set up in pill 2 but maybe that got edited out at some point or something
<wpcarro>
sphalerite: yea I couldn't run the `source ...` command that pill 2 recommended
<wpcarro>
bc of the missing `nix.sh` file
<sphalerite>
wpcarro: as for nix.sh, this should exist if you've installed nix. If you're on nixos the environment for using nix is already set up so you don't need to source it
<wpcarro>
and I didn't know what to put as the password when prompted in the `su - nix` command
<sphalerite>
yeah that's normal if you haven't actually set up a nix user
<clever>
wpcarro: su always asks for the root password
<sphalerite>
this is plainly the pills' fault
<sphalerite>
clever: no it doesn't
<sphalerite>
clever: it asks for the password of the user you're logging in as
<clever>
ah
<wpcarro>
sphalerite: I'm on OSX - not NixOS
<wpcarro>
assuming those are mutually exclusive...
rihards has quit [Quit: rihards]
<sphalerite>
aaah I think the installer defaults to a multi-user install nowadays which means… I don't know exaclty what that means
<sphalerite>
can you run `nix-store`?
<wpcarro>
mhmm
<wpcarro>
all of the nix-commands we're available to me so far
<wpcarro>
nix-env, nix-store, etc
<sphalerite>
alright then you can just skip that step
<sphalerite>
the source nix.sh one
<wpcarro>
ok... was a little concerned to not see nix.sh in the profile though
<wpcarro>
I was under the impression that nix was somehow bootstrapping itself or something
<sphalerite>
yeah if it's a multi-user setup it lives… somewhere else. Not sure how this actually works on non-nixos
<wpcarro>
It's ok that that part is opaque though... I'm getting off of OSX asap
<sphalerite>
in any case you don't need to source anything if it's already in your PATH
johanot has quit [Quit: Leaving.]
<wpcarro>
ok nice... I'll search around for some other stuff that tripped me up
<wpcarro>
was afraid I had a bad install or something
orivej has joined #nixos
<wpcarro>
I'm afraid I won't have the time to run through the steps (1-9) today that tripped me up. Perhaps later this week, in which I'll post the findings here
<{^_^}>
[nixpkgs] @shlevy pushed to release-18.03 « pythonPackages.simpy: init at 3.0.11. »: https://git.io/fNsAQ
<sphalerite>
wpcarro: that's great! If you have a github account maybe reporting them against the repo https://github.com/nixos/nix-pills would be better so that it's persistently and obviously documented and it doesn't just disappear into the depths of my passive memory and the IRC logs ;)
<wpcarro>
sphalerite: even better!
<LnL>
wpcarro: what I've noticed is really hard to get right with the nix-pills is a good balance between examples to explain what's going on and showing stuff that actually works everywhere
<wpcarro>
LnL: it might be helpful to have nix-pills be pinned to nix versions and OSs. Seems like some MacOS users are experiencing similar issues to me RE: pill 7
<wpcarro>
But maybe branches of the nix-pills repo could be dedicated for docs that certainly work with given nix versions on certain OSes
<wpcarro>
idk... or maybe even a Docker image that the NixPills can be run in for non-NixOS users
<__monty__>
I'm not really a fan of the nix pills. They're a fine rough overview of nix but they were mostly useless to me to actually do anything with nix.
<wpcarro>
__monty__: they're helping me gain confidence in Nix's inner-workings. Thankfully my company is looking to adopt Nix and we have an experienced user already trailblazing. I'm playing catchup and pills are serving me alright
<LnL>
wpcarro: yeah, 7 is a great example
magnetophon has quit [Ping timeout: 248 seconds]
derped has joined #nixos
<__monty__>
I go by the manual, code in nixpkgs and asking lots of questions here. But if it works for you all the better.
<sphalerite>
♥ #nixos
<iqubic>
Why is it that a single nix-build might create 3 different result symlinks?
<LnL>
__monty__: that's not the goal, I think they are great but what you're talking about is something that's a pretty weak spot in terms of documentation
<samueldr>
multiple output derivations
<iqubic>
result, result-2, result-3
<iqubic>
samueldr: I see. I'm actually quite confused.
tzemanovic has joined #nixos
<clever>
iqubic: you pointed nix-build to an attribute set containing 3 attributes
tzemanovic has quit [Client Quit]
<samueldr>
oh, that's not what I thought
<LnL>
iqubic: that are not outputs but 3 different derivations
<sphalerite>
yeah if you give it a multiple-output thing it gives you result-dev and similar
<iqubic>
I just tried using node2nix to try and build an NPM thing, and I got three results
<LnL>
if you nix-build a list or a set it will build everything in there
<iqubic>
I have very little clue what node2nix actually just did.
<clever>
iqubic: what do the 3 symlinks point to?
cryptomonoid has joined #nixos
<iqubic>
node-http-server, node-shell-http-server, and node-tarball-http-server
<clever>
iqubic: and can you gist the nix file you ran nix-build on?
lor|d has joined #nixos
lord| has quit [Disconnected by services]
lor|d is now known as lord|
doyougnu has quit [Ping timeout: 256 seconds]
aanderse-laptop has quit [Ping timeout: 265 seconds]
<iqubic>
Feel free to look at the *.nix files here.
<clever>
iqubic: you need to pass one of: -A tarball, -A package, or -A shell
<clever>
depending on what you want to build
leo60228 has joined #nixos
<iqubic>
I'm not sure what the differences are.
<iqubic>
I'm not even sure how node2nix found three different things to build.
<clever>
tarball, is the tarball containing your source
<clever>
shell is only for nix-shell use
<clever>
package is your package
<clever>
all 3 of them are the same thing
<leo60228>
Hello, I'm trying to update Eagle. It has not been updated in several years, and there have been 2 new major versions, and an ownership change, since then. It appears to have switched to Qt; I'm getting "Could not initialize GLX" whenever I try to start it, though. Here's my expression: https://hastebin.com/qikiwaqofe.nix
<iqubic>
I guess I just want the package version.
<iqubic>
Because I want to be able to use this inside and outside of nix-shell.
<iqubic>
Does node2nix always create three different derivations to build?
<clever>
iqubic: yes
Guest64931 has quit [Quit: WeeChat 1.4]
<iqubic>
How do you make a package that is only useable in nix-shell?
bennofs has quit [Ping timeout: 256 seconds]
<clever>
iqubic: it likely lacks a build phase
<leo60228>
It appears to be looking for libGLX_indirect.so.0, that's not in any package, but libGLX.so.0 is
<iqubic>
So I haven't actually installed anything have I?
<iqubic>
Like I need to use nix-env to install this to my store, right?
<clever>
iqubic: correct
<iqubic>
nix-env -i -f 'default.nix' -A package
<iqubic>
I just ran that and it seems to have worked correctly.
<leo60228>
Wait, it finds libGLX already...
cryptomonoid has quit [Ping timeout: 244 seconds]
<leo60228>
Why is it looking for libGLX_indirect?
<iqubic>
clever: Is that all I need to do to install this package?
<rizary>
Hey guys, does nix.conf has the same behavior with configuration.nix for non-nixos user? I mean I try to install docker in my non-nixos machine, and when I see how easy the configuration in nixos, I try to configure it in my non-nixos machine. But I can't get it working.
<sphalerite>
rizary: no, nix.conf is for configuring nix
<clever>
iqubic: you can also: nix-env -f default.nix -iA package
<sphalerite>
rizary: configuration.nix is for configuring nixos, and requires the rest of nixos to be built essenitally
<rizary>
Hmm I see..
wpcarro has quit [Read error: Connection reset by peer]
<iqubic>
clever: that's the same thing, just re-arranging the position of the flags from what I did: nix-env -i -f 'default.nix' -A package
<elvishjerricco>
sphalerite: I'd love it if we could use the nixos modules on other systemd linuxes
<leo60228>
Is there anyone who works on the OpenGL packages here? It looks like one of them should be providing libGLX_indirect.so.0, but none is, according to nix-locate.
wpcarro has joined #nixos
<elvishjerricco>
Just the bare minimum for creating systemd units with nix
<sphalerite>
elvishjerricco: yeah that would be cool, although I use nixos on all my machines so I don't personally need it. And I'm still looking for a job, so I don't have a place where I might need that professionally :p
<sphalerite>
otherwise I might have done that by now, because it would indeed be nice
<clever>
iqubic: pretty much
wpcarro has quit [Remote host closed the connection]
<iqubic>
well... "which https-server" works, and gives me a valid path. I assume that I have installed this properly.
Tobba has joined #nixos
wpcarro has joined #nixos
zduch4c has joined #nixos
<zduch4c>
hello
<iqubic>
Hey there.
igo95862 has joined #nixos
<juhe>
How to reference attribute of function patermeter from overlay? E.g. kicad-unstable has ngspiceSupport, how to access it from overlay?
<zduch4c>
on my project, i do `nix-shell -p SDL2 SDL2_image' to launch my environment... after i try to compile it (of course with -lSDL2 and -lSDL2_image) the compiler starts spouting this bs: https://p.teknik.io/Raw/h4Jz0
<sphalerite>
juhe: you mean how to change it? Or how to get its value?
sbdchd has joined #nixos
<zduch4c>
any idea how to fix it up?
<sphalerite>
juhe: you can change it using override, e.g. kicad-unstable.override { ngSpiceSupport = true; }
<juhe>
sphalerite: how to check it's value from overlay
<sphalerite>
juhe: as for reading it, I don't think you can
<clever>
zduch4c: SDL is being stupid, the headers are in $out/include/SDL2/ and you must manually add -I${sdl2}/include/SDL2 (or use pkgconfig)
<juhe>
sphalerite: oh, ok...
<clever>
zduch4c: they are trying to fix the problem of sdl versions conflicting with eachother in the same /usr/include, which nix entirely prevents
<zduch4c>
mhm
<zduch4c>
thanks
<rizary>
sphalerite: so you're saying that in order to make it possible, we somehow need to make nix able to access non-nixos systemd?? Would you show me how to see code implementation on nixos systemd?
<juhe>
sphalerite: I wanted to add a patch depending on that value. If that is not possible I need to find a different way.
<sphalerite>
rizary: no, nix doesn't do that stuff
__monty__ has quit [Quit: leaving]
* rizary
nod
<sphalerite>
rizary: nixos-rebuild does most of the work, calling nix to build the system config
<sphalerite>
then it calls the activation script and such
<iqubic>
nixos systemd is the same as regular systemd. There is no distinction between "nixos systemd" and "non-nixos systemd"
<sphalerite>
juhe: it might be possible actually, I just don't know how and haven't ever seen it done before
<maurer>
I mean, there's a teeny bit of difference because in non-nixos systemd, you're expected to be able to perform some configuration operations by removing and adding symlinks in etc I think...
<juhe>
sphalerite: ok
<maurer>
but the daemons should be compatible
<iqubic>
So the default port of http-server is already being used by the haskell hoogle documentation server. What other ports should I use?
<iqubic>
I'm thinking I want to use 3030, but I'm not sure if a regular non-root user can use that port number.
<sphalerite>
iqubic: completely up to you
<clever>
iqubic: non-root can use any port over 1024
<sphalerite>
iqubic: iirc ports above 1024 are available to non-root
<iqubic>
Cool
<zduch4c>
ok itw orks thanks clever
zduch4c has quit [Quit: Page closed]
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
johnw_ has joined #nixos
<{^_^}>
[nixpkgs] @xeji pushed commit from @flokli to release-18.03 « virtualbox: 5.2.12 -> 5.2.14 »: https://git.io/fNsjV
Coyin has quit [Ping timeout: 265 seconds]
Coyin has joined #nixos
johnw has quit [Ping timeout: 276 seconds]
wpcarro has quit [Remote host closed the connection]
doyougnu has joined #nixos
wpcarro has joined #nixos
tzemanovic has joined #nixos
SuperKazuya has quit [Ping timeout: 252 seconds]
<iqubic>
If I edit the code in a directory and then re-run nix-build and nix-env, will I get the new version of the executable?
<sphalerite>
yes
<sphalerite>
you don't need to run nix-build separately
<sphalerite>
nix-env will build as necessary
<iqubic>
I know.
<iqubic>
nix-env -f 'default.nix' -iA package
sbdchd has quit [Remote host closed the connection]
<iqubic>
Will that pick up new changes to the code in that directory?
sbdchd has joined #nixos
<sphalerite>
if the nix expression refers to that directory
<sphalerite>
yes
Supersonic112 has joined #nixos
Supersonic has quit [Disconnected by services]
<sphalerite>
if it fetches it from github or something then no
Supersonic112 is now known as Supersonic
igo95862 has quit [Quit: igo95862]
<iqubic>
It isn't fetching anything from github.
igo95862 has joined #nixos
sbdchd has quit [Remote host closed the connection]
<iqubic>
I took an existing node package, forked it, built it with node2nix, then built it, then made some minor changes to the code (changing default port) and then built again.
<iqubic>
I was wondering if the second build would have the default port changed, and it looks like the answer is yes.
kerrhau has joined #nixos
kerrhau has quit [Changing host]
kerrhau has joined #nixos
vandenoever has joined #nixos
<sphalerite>
it does depend on the generated expressions. I'm not sure what node2nix does
<sphalerite>
but the easiest way to find out is to try I guess :)
sbdchd has joined #nixos
<logzet>
Is anyone running Virtualbox with the extension Pack? I enabled enableExtensionPack via an override and it seems to be contained in the result (libexec/applications/virtualbox/ExtensionPacks/Oracle_VBoxDTrace_Extension_Pack). But VirtualBox does not detect the pack.
<clever>
logzet: when i hit the gui button to install the extensions, it just downloads an iso at runtime
<clever>
logzet: i never enabled any overrides for it
wpcarro has quit [Ping timeout: 264 seconds]
<iqubic>
sphalerite: node2nix is a way to deploy node packages on nix. I'm using it as an alternative to "npm install -g https-server"
igo95862 has quit [Client Quit]
<iqubic>
I have to go now.
<logzet>
clever: I'm trying Oracle's Extension Pack (mainly for USB 2.0), is there a chance that you are talking about the guest additions?
<iqubic>
clever: thank you for helpping me get node2nix working.
<logzet>
clever: I'm trying to install the Oracle VM VirtualBox Extension Pack (it is needed for USB2/3 and some other features)
<logzet>
(The guest additions iso is something else)
<iqubic>
I thought there were configuration.nix options for this kind of thing.
<clever>
logzet: ah, checking closer, my VM is configured with a 1.0 controller
<clever>
logzet: and it correctly tells me that it needs an extension pack if i select 2.0 or 3.0
<logzet>
iqubic: There is an enableExtensionPack flag, I enabled it with override. VBox does not detect it though…
<Lisanna>
erhm... I have the meta.schedulingPriority of one of my jobs set to 200 (all the other ones are 100), but it feels like hydra started it *last*
puckipedia is now known as puck
<Lisanna>
...do I have schedulingPriority backwards
<clever>
Lisanna: i dont think so
<Lisanna>
So, bigger number = starts sooner?
<Lisanna>
time to read the source code...
<clever>
Lisanna: i believe its tied into the scheduling shares
<Lisanna>
clever my jobset has fewer shares than jobs - about 600 jobs (300 of them will be cached results), and 100 shares
<logzet>
clever, iqubic: Since I don't see a way in the gui to point VBox to a pack, I'll open an issue on gh
<{^_^}>
[nixpkgs] @xeji pushed commit from @volth to master « xfce4-13.xfce4-power-manager: 1.6.0 -> 1.6.1 (#43626) »: https://git.io/fNGvS
<clever>
logzet: what config are you using to try to enable the extension pack?
<dhess>
ugh, kernel panic on the Nvidia TX2
<dhess>
oh well, back to the TX1
lord| has joined #nixos
<dhess>
the fan never came on, that might be why
<logzet>
clever: (virtualbox.override { enableExtensionPack = true; pulseSupport = true; }) in user packages in addition to virtualisation.virtualbox.host.enable = true;
<clever>
logzet: can you gist the full configuration.nix?
<romildo>
What is the link to the page that lists the latest releases (commits) of nixos-unstable and other nixos versions? I had it on my bookmarks, but it has been lost.
doyougnu has quit [Ping timeout: 248 seconds]
freeman42x[nix] has quit [Ping timeout: 256 seconds]