<kandinski>
I have restarted it with `sudo systemctl restart jack.service` and `sudo systemctl restart jack-session.service`, but it's either the "Failed to open server" error or the new "Cannot initialize driver": https://paste.debian.net/1116481/
<colemickens>
Is grub the only way to have multiple bootable "profiles" per generation?
<clever>
slack1256: basically, every time you call setsid (like, opening a new shell in ssh or a term emulator), it creates a new autogroup
<clever>
slack1256: and each autogroup gets its own fair share, and then nice levels only work within that autogroup
<clever>
so `make -j300` gets an equal share to the browser
m0rphism has quit [Ping timeout: 276 seconds]
drakonis__ has quit [Quit: Leaving]
kecia[m] has joined #nixos
mexisme has quit [Ping timeout: 245 seconds]
<slack1256>
Yeah I knew about that behaviour with respect to autogroup. But now I am using cgroup-v2 actual tree, controlling memory/disk/cpu (nix-shell/ghc compiling taffybar) and another cgroup with cpu weight high for mpv. I was wondering if the nice behaviour of autogroups translate to cgroups.
<clever>
ive not really played with cgroups yet
<slack1256>
I played with them on 2012 (matlab class). But now haskellPackages and watching movies made it neccesary.
<slack1256>
It's not hard if you know in abstract how it ought to work
<clever>
slack1256: you can also tweak the nicelevel of an entire autogroup, the reddit thread explains that
<slack1256>
Yeah, writing to the /proc/<pid>/autogroup file.
<slack1256>
There is a cpu.weight.nice file set to 5 on the cgroup, but the docs don't say what it does.
<slack1256>
Ohh well.
<slack1256>
clever: If you want to play with cgroups some day, run something like "systemd-run --user --scope -p IOAccounting=true -p IOReadBandwidthMax="/dev/sda 2M" -p IOWriteBandwidthMax="/dev/sda 2M" -p MemoryHigh=1800M -p MemoryMax=2G -p MemorySwapMax=40M -p CPUWeight=30 zsh"
<clever>
ah, ive messed with systemd-run before
<clever>
i think it was pam related issues
<slack1256>
It will give you a shell on a new cgroup (it will print it on screen). Then you can migrate nix-daemon to such cgroup and never fear nix-shell again!
<clever>
when you login, pam needs a special flag to set who initially signed in
<clever>
and sudo cant clear that flag
<clever>
testing things, requires a shell, that has not yet signed in!
<slack1256>
Oh that was bug on systemd-run?
<clever>
nope
<clever>
systemd-run lets you spawn new things from pid1, and they arent tagged to your user
<clever>
so you can escape the tracking metadata, and experiment with setting it again
<lovesegfault>
clever: I thought it updated when nixos:trunk-combined:tested passed?
<clever>
lovesegfault: the main channels wait for hydra to finish building everything, plus testing a few select things (the tested job)
<clever>
lovesegfault: the -small channels dont want for hydra to build everything, so they update faster (but then you dont have good coverage on the cache)
<lovesegfault>
clever: Do users of the small channel experience more breakages?
<clever>
lovesegfault: -small also tests fewer things
<clever>
so some things may slip thru
<lovesegfault>
Aha, I see
<lovesegfault>
Got it
<lovesegfault>
clever: If the CI tests stuff so thoroughly how can there be so many failing jobs?
<lovesegfault>
Also, how can there ever be a channel release with all those failing jobs
<gwen>
Hi, I've got trouble running matplotlib with Qt -- installed matplotlib.override {enableQt = true;} via python3.buildEnv, but when running I get the "plugin xcb not found" error
<{^_^}>
Found in packages: framac, uclibc, cdrtools, musl.dev, glibc.dev, emscripten, postgresql94, postgresql95, postgresql96, postgresql_10, postgresql_11, glibc_multi.dev, glibc_memusage.dev
<gwen>
lovesegfault: Could it be that matplotlib does not use wrapQtAppsHook? On a quick look, I don't find anything related in python-modules/matplotlib/default.nix
<lovesegfault>
gwen: Yes, that's what I'm thinking
<{^_^}>
#56921 (by romildo, 36 weeks ago, open): multibootusb can not find the Qt platform plugin "xcb"
<clever>
gwen: the issue with python stuff, is that you tend to have to wrap the final script, that is using the python library, that is using the qt libraries
<clever>
lovesegfault: these are all semi-optional things, that are grouped up for easy on/off
<lovesegfault>
What I'm trying to make more manageable is change propagation
<clever>
take media-center.nix for example, that deals with running plex on bootup, and configuring auto-login
<clever>
so the hdmi out behaves like a dumb device, like a chromecast
<lovesegfault>
I have a lot of machines with very similar configs, and I end up making changes on my personal laptop that I have to then manually propagate to other laptops etc, which is what annoys me
<clever>
any machine that should be doing that duty, just has ./media-server.nix added to its ${hostname}.nix file
<lovesegfault>
I see
<clever>
lovesegfault: thats solved by just putting it all into a single git repo, so you can share a file between them
<lovesegfault>
Yeah, I already have that
<clever>
and i do also have the issue that i dont push often enough
<clever>
so each machines copies of nixos-configs begin to diverge
<clever>
and i have to commit and pull/push to sync them up every now and then
justan0theruser has quit [Ping timeout: 240 seconds]
<lovesegfault>
Yeah, I have that too
<srhb>
I find that pushing from a central location mostly gets rid of that (in my case, my primary laptop)
<lovesegfault>
I need a cron job or something :P
<lovesegfault>
srhb: Of what?
<clever>
srhb: its more that i need to edit and test things on one end, and then i forget to push
<srhb>
Of diverging git repos.
<srhb>
clever: Sorry, I meant pushing system closures
<srhb>
git only lives in one place for me.
<clever>
srhb: i have been switching over to nixops, the nas and router are now managed via that
<clever>
but the desktop/laptop are still normal nixos
<srhb>
Right. :)
<srhb>
I guess I might be tempted to do the same if I had a desktop.
<lovesegfault>
Since the machines don't have fixed IPs
<clever>
lovesegfault: and because i avoid using nixops features, i'm able to still build the config outside nixops
<lovesegfault>
but I could just use subdomains of lovesegfault.com
alanz has joined #nixos
<clever>
lovesegfault: there is an uncommited change, to swap my router over to a vpn ip
<clever>
lovesegfault: because i had to deploy while at nixcon, lol
<lovesegfault>
My nix-config repo is private because of secrets, I might do what my coworker does of having a private repo as a subtree on a public repo
<lovesegfault>
clever: Hehehe
<lovesegfault>
All my coworkers were at nixcon but me
<clever>
lovesegfault: look at load-secrets.nix in my repo
<lovesegfault>
clever: You've thought of everything!
<clever>
lovesegfault: the trick, is that nix just lets storepaths from other arches co-exist in /nix/store/
mexisme has joined #nixos
<lovesegfault>
Yeah, which is cool
<clever>
lovesegfault: so you just get 2 builds of nixos in using nix-copy-closure, and then configure 2 different bootloaders, to start at different paths
<srhb>
lovesegfault: Yeah, doesn't seem too bad..
<clever>
lovesegfault: when you nixos-rebuild under arm, it only updates the arm bootloader config
<clever>
so grub still points to an x86 nixos
<lovesegfault>
clever: How do you manage your weechat config?
<lovesegfault>
I've been meaning to put that in git forever
<gjabell>
clever: thanks! wish github's search worked better, finding nix examples would be so much easier :P
<clever>
lovesegfault: but i only use it for the slack plugin and nothing else
<clever>
gjabell: i often find it simpler to just clone and grep, but that doesnt work when you dont know what to clone
<lovesegfault>
Hmmm
<lovesegfault>
I have an idea
<lovesegfault>
Oh, speaking of
<lovesegfault>
if foo.nix sets environment.systemPackages = [ a b c ]; and bar.nix sets environment.systemPackages = [ d e f ]; what happens when baz.nix does imports = [ foo.nix bar.nix ];?
<srhb>
lovesegfault: They merge according to type
<srhb>
lovesegfault: For lists, that means concat.
<lovesegfault>
srhb: What about attr sets?
<srhb>
lovesegfault: They merge via deepUpdate, where each attribute with duplicate attrs merge according to type
<clever>
lovesegfault: depends on if the set cointains more options or is an option itself
<srhb>
lovesegfault: And so on down.
<lovesegfault>
if I import foo = { a = true; }; and foo = { a = false; };
<srhb>
It depends on the type of foo
<clever>
and the type of a
<srhb>
lovesegfault: Try checking out the NixOS manual section on option types.
<lovesegfault>
looking
<xd1le>
how do I find out where to access the "shutdown" command? Would it just be something like "${pkgs.systemd}/bin/shutdown"
<lovesegfault>
thanks srhb & clever :)
<xd1le>
or is it under utillinux?
<clever>
srhb: i was also horified to discover, that if you dont set a type on something, the default merge rule inspects the types of the 2 args, and for bools, will just logical or them
<clever>
,locate bin shutdown
<{^_^}>
Found in packages: systemd, sysvinit, molly-guard, systemd_with_lvm2
<srhb>
clever: uh what :D
<srhb>
I would expect simple bools to merge on equal. yikes.
<xd1le>
clever: oh cool ty
<clever>
srhb: if you dont set the type of something, then set the value to both true and false, it will silently true
<clever>
srhb: but if you set type = types.bool;, it will silently merge true and true, but complain about true and false
iyzsong has joined #nixos
<srhb>
Yeah, that's the behaviour one would assume..
aveltras has joined #nixos
<clever>
srhb: another weird thing, is that the string type, complains if you set it to both "foo" and "foo"
<srhb>
Yeah, it's mergeOnEqual i think?
<srhb>
Er
<srhb>
MergeUniq
<lovesegfault>
Hmmm
<srhb>
Or whatever it's called
<lovesegfault>
so what to call the dir with the joined modules
<srhb>
lovesegfault: combo!
<lovesegfault>
I thought of `top-level`
<srhb>
lovesegfault: transform!
<lovesegfault>
srhb: Ha! I like combo :D
<clever>
lovesegfault: i just spew things all over the root dir :P
<lovesegfault>
clever: I saw it :P
<srhb>
Yeah, mine's in root of "modules" and from there there's a lib/foo.nix
<srhb>
Which is technically bad use of terminology but meh
tmaekawa has quit [Ping timeout: 276 seconds]
<{^_^}>
[nixpkgs] @c0bw3b pushed to master « gzrt: fix homepage and src url redirect »: https://git.io/JeolY
kai_w has quit [Quit: Konversation terminated!]
kai_w has joined #nixos
<lovesegfault>
clever, srhb: if I import an attr set that says i18n.consoleKeyMap = "us", and then after in the file I say i18n.consoleKeyMap = "uk";, what will happen?
<lovesegfault>
Is it the same behaviour as conflicting imports?
<clever>
i see over 6 PR's involved in adding libva to steam, in recent irc logs, lol
<otwieracz>
How can I use fuse in container?
<otwieracz>
I have enabled fuse on host, but in cotainer it still does not work.
<clever>
otwieracz: does /dev/fuse exist in the container?
<boogiewoogie>
DigitalKiwi: yeah, tried all the channels for just steam
<otwieracz>
clever: obviously no, that's the problem :)
<clever>
otwieracz: you may need to configure nixos-container to bind mount /dev/fuse into the guest
<otwieracz>
let me try...
<DigitalKiwi>
i'm fairly happy with running nixos-unstable as my primary channel and sometimes use an old channel version of a program if it's broken and i don't want to wait...or you fix it and submit a PR and be a hero :D
<FRidh>
MichaelRaskin: would be nice if we had impure derivations. Then we could use Nix in CI not for builds but also for all other steps.
<MichaelRaskin>
I think the problems were also about caching needs, though
<clever>
FRidh: i just use shellHook in nix-shell
<clever>
FRidh: then you can just `nix-shell -A action` and the shellHook will run things, with the buildInputs already in PATH, and then shellHook can either `exit 0` or `exit 1`
<clever>
FRidh: there is also the escape-hatch of builtins.exec
<xwvvvvwx>
How does nix determine if a file has changed? I'm importing nixpkgs from a local copy, but when I make changes in nixpkgs it doesn't seem to update the path to the nixpkgs source in the store
turion has quit [Ping timeout: 240 seconds]
<clever>
xwvvvvwx: how are you importing nixpkgs from a local copy?
<xwvvvvwx>
clever: in a let block: `nixpkgs = import ../../nixpkgs {};`
<clever>
xwvvvvwx: that should detect changes every time you re-run the nix commands
ebzzry has joined #nixos
jgt has joined #nixos
<xwvvvvwx>
hmmmmm
<tilpner>
gjabell: Can you paste what you meant by "adding to user path"?
o1lo01ol1o has quit [Ping timeout: 265 seconds]
<clever>
xwvvvvwx: what file are you changing, what command are you running?
pareidolia has quit [Ping timeout: 240 seconds]
<gjabell>
tilpner: I just added an line to zsh to export PATH=${weechat-matrix}/bin:$PATH
<{^_^}>
[nixpkgs] @joachifm pushed 2 commits to release-19.09: https://git.io/Jeo8d
<tilpner>
gjabell: Where did you add that?
zupo has joined #nixos
zupo has quit [Client Quit]
gwen has joined #nixos
genesis has joined #nixos
lucasvo_3 has joined #nixos
<timokau[m]>
Is the nixos option `nix.nixPath` not equivalent to setting the `NIX_PATH` environment variable? Because I have set the nixos option, and it doesn't seem to have an impact on home-manager. When I set `NIX_PATH` to the same value however, home-manager uses the right version of nixpkgs.
lucasvo_3 has quit [Ping timeout: 240 seconds]
thc202 has joined #nixos
<clever>
timokau[m]: nix.nixPath only takes effect after nixos-rebuild completes, and wont affect the current build
<clever>
timokau[m]: it also only takes effect after you relaunch the shell
<timokau[m]>
clever: Yes, I know but the system has already rebuilt
<timokau[m]>
clever: Oh, that might be it
mexisme has joined #nixos
<timokau[m]>
clever: Okay so I had to reboot for it to take effect. Simply opening a new shell instance didn't do the trick
<clever>
timokau[m]: logging out and back in may also do it
<timokau[m]>
clever: I assumed that `nix.nixPath` was an actual option of the nix daemon, and setting it would cause the daemon to reload its config. But turns out it actually just sets `NIX_PATH` in `environment.sessionVariables`. Confusing.
<timokau[m]>
clever: Thanks! That trick seems to be exactly what I'm looking for.
<betawaffle>
what's the right way to read a file into a config value as a string?
<clever>
timokau[m]: yeah, the nix exprs and evaluation happens entirely on the user side, and nix-daemon doesnt know how to parse nix files
<clever>
betawaffle: builtins.readFile ?
<betawaffle>
ideally with a path relative to the current file. oh, thanks
<clever>
betawaffle: builtins.readFile ./foo.txt
<betawaffle>
perfect
<betawaffle>
(i've finally decided to move my nftables rules into a separate file :P)
<timokau[m]>
clever: Only ugly thing about that trick is that it will still require one initial reboot/relogin
mexisme has joined #nixos
* clever
heads off to bed
knupfer has quit [Ping timeout: 245 seconds]
leo_ has joined #nixos
mananamenos has joined #nixos
magneticduck has joined #nixos
mexisme has quit [Ping timeout: 245 seconds]
gwen has quit [Ping timeout: 276 seconds]
tilpner has quit [Quit: tilpner]
tilpner has joined #nixos
mexisme has joined #nixos
xkapastel has quit [Quit: Connection closed for inactivity]
<asymmetric>
has anyone managed to send a notification via libnotify from a systemd system service?
cosimone has joined #nixos
mexisme has quit [Ping timeout: 276 seconds]
__monty__ has joined #nixos
mexisme has joined #nixos
exfalso has joined #nixos
zupo has joined #nixos
mexisme has quit [Ping timeout: 252 seconds]
mexisme has joined #nixos
gwen has joined #nixos
<wedens[m]>
why do I get this error "internal error: End of file from qemu monitor" on VM shutdown? 19.09 with libvirt 5.4.0 and qemu 4.0.1. also default libvirt network doesn't work after VM restart
mexisme has quit [Ping timeout: 276 seconds]
cosimone has quit [Remote host closed the connection]
<asymmetric>
even though the DBUS_SESSION_BUS_ADDRESS variable has the same value for root and my user
<otwieracz>
I am trying to run sshfs and encfs from systemd unit.
o1lo01ol1o has joined #nixos
<otwieracz>
However, sshfs fails with error: Nov 16 13:37:37 bacula-sd kn7bi59p53d4q8n4qg5lhqzaw1md25m2-unit-script-arsnap-start[345]: fuse: failed to open /dev/fuse: Operation not permitted
<asymmetric>
is there a way to configure dbus to use `unix:path` instead of `unix:abstract`?
mexisme has quit [Ping timeout: 276 seconds]
mexisme has joined #nixos
<asymmetric>
as i think that will be more permanent than the unix:abstract=/tmp/dbus-cvbmvEi4fR,guid=419870d45e1b0b6ad43f06b65dce60ec stuff i have now, allowing me to create notifications from another user
vesper11 has joined #nixos
<{^_^}>
[nixpkgs] @FRidh opened pull request #73496 → NixOS Python test driver: Python package and Sphinx docs → https://git.io/JeoBS
<{^_^}>
[nixpkgs] @bjornfor pushed commit from @livnev to master « john: copy rules »: https://git.io/JeoRr
vika_nezrimaya has joined #nixos
mexisme has quit [Ping timeout: 250 seconds]
mananamenos has quit [Remote host closed the connection]
mexisme has joined #nixos
<{^_^}>
[nixpkgs] @FRidh opened pull request #73501 → CF: don't build swift-corefoundation anymore on Hydra → https://git.io/JeoRX
Rusty1 has joined #nixos
mexisme has quit [Ping timeout: 276 seconds]
reallymemorable has joined #nixos
xkapastel has joined #nixos
vika_nezrimaya has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @c0bw3b opened pull request #73503 → efitools: fix build with gnu-efi >= 3.0.11 → https://git.io/JeoRD
mexisme has joined #nixos
chris___ has joined #nixos
<chris___>
Just curious. If I do the following in the nix repl `pkgs.haskell.packages.ghc865.callCabal2nix` it shows an output of `.../nixos/pkgs/development/haskell-modules/make-package-set.nix...` however this path does not exactly match the nixpkgs repo? (pkgs is in the root rather than /nixos/pkgs). Why is this different?
<__monty__>
Is your channel called `nixos` perchance?
mexisme has quit [Ping timeout: 246 seconds]
mexisme has joined #nixos
<chris___>
It's using the `nixos-19.09` branch if that is what you're asking? But yes my hmm "system" channel is named nixos iirc.
cosimone has quit [Remote host closed the connection]
cosimone has joined #nixos
civodul has joined #nixos
mexisme has quit [Ping timeout: 245 seconds]
<{^_^}>
[nixpkgs] @c0bw3b pushed to master « gaia: restrict to X86 platforms »: https://git.io/JeoRx
mexisme has joined #nixos
MichaelRaskin has quit [Ping timeout: 250 seconds]
fendor has quit [Read error: Connection reset by peer]
boogiewoogie has quit [Quit: Leaving]
<__monty__>
chris___: Then I'm gonna guess and say that the "nixos" in that path is the channel name.
<{^_^}>
#73444 (by ryantm, 1 day ago, merged): Add configuration for GitHub stale bot
<{^_^}>
[nixpkgs] @tomfitzhenry opened pull request #73505 → systemd-networkd: add DHCPServer test → https://git.io/Jeo0c
mexisme has quit [Ping timeout: 252 seconds]
domogled has joined #nixos
mexisme has joined #nixos
Thra11_ has joined #nixos
mexisme has quit [Ping timeout: 250 seconds]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
bvdw has quit [Read error: Connection reset by peer]
exfalso has joined #nixos
<chris___>
Where can I read about variable substition in a string with nix? I can't seem to find anything on google :/. I'm trying to understand how a set gets substituted in a string.
<wedens[m]>
are there any known libvirt networking with systemd-networkd quirks ?
magnetophon has joined #nixos
<eyJhb>
hyper_ch: is it working for you without anything ?
<hyper_ch>
eyJhb: for some reason kde/kwin couldn't boot into it... it stopped at the splash screen... so I went back to previous generation. Not sure why that is
zupo has joined #nixos
<eyJhb>
hyper_ch: DisplayLink works in mysterious ways, most of them are trying to screw up over
<sondr3>
is there a way to import <lib> in a nix repl? Getting tired of fixing all the `undefined variable` when copying expressions from a nix file to the repl to work on them
kleisli has joined #nixos
<tilpner>
:l <lib>
chris___ has quit [Ping timeout: 240 seconds]
mexisme has quit [Ping timeout: 252 seconds]
fragamus has joined #nixos
emily has quit [Remote host closed the connection]
emily has joined #nixos
mexisme has joined #nixos
reallymemorable has quit [Quit: reallymemorable]
<sondr3>
tilpner: I just get `error: file 'lib' was not found in the Nix search path`, I've attempted to import `<nixos>` and `<nixpkgs>` but neither makes :l <lib> work
kenshinCH has joined #nixos
kenshinCH has quit [Client Quit]
<betaboon>
is it just me or is qemu-vm.nix somewhat of a mess ?
<wedens[m]>
what's <lib>? there is no such thinkg in defaule NIX_PATH
<tilpner>
sondr3: Well, what do you want to import?
<tilpner>
sondr3: :l <nixpkgs/lib> or :l <nixos/lib> may be what you meant
<tilpner>
sondr3: But there's a better alternative: :a lib
<tilpner>
(Assuming you have imported a nixpkgs into scope)
<sondr3>
tilpner: functions like mapAttrs etc, `:l <nixpkgs/lib>` worked
<sondr3>
tilpner: you mean with `nix repl -I`?
<tilpner>
sondr3: I use: nir is an alias for nix repl '<nixpkgs>'
<sondr3>
tilpner: thanks, that final alias worked, I'll add it to my own aliases, thanks!
<sondr3>
tilpner++
<{^_^}>
tilpner's karma got increased to 53
mexisme has quit [Ping timeout: 276 seconds]
<{^_^}>
[rfcs] @d-goldin opened pull request #59 → rfc-0057: systemd service secrets → https://git.io/Jeouv
<wedens[m]>
hmm. after updating from 19.09 channel from 1w old to the current revision, usb sound card is not being detected
mexisme has joined #nixos
<sondr3>
wedens[m]: which version? I'm on `19.09.1208.ef8c34c4721` and my JSD Atom sound card works fine
<wedens[m]>
I tried upgrading to nixos-19.09.1254.9104be2ee08 from nixos-19.09.1172.2d896998dc9
<ddima>
wedens[m]: I'll have to check when Im back at home, I'm on the go atm.
<ddima>
I'll poke you later then.
<adisbladis>
wedens[m]: What problems are you seeing exactly?
<magneticduck>
when I want to include a resource (fetchUrl, fetchPyPi, etc.) into something I'm building, how should I get the hash? up to now, my solution has been to use a fake hash, run the builder to fail, and then enter in the hash that the builder expected.
<magneticduck>
obviously this is stupid. what should I do instead?
<adisbladis>
,tofu magneticduck
<{^_^}>
magneticduck: To get a sha256 hash of a new source, you can use the Trust On First Use model: use probably-wrong hash (for example: 0000000000000000000000000000000000000000000000000000) then replace it with the correct hash Nix expected. See: tofu-vim
<adisbladis>
magneticduck: Also there are tools like nix-prefetch-git and nix-prefetch-url
alexherbo2 has joined #nixos
cosimone_ has quit [Remote host closed the connection]
<magneticduck>
what's tofu-vim?
cosimone_ has joined #nixos
grumble has quit [Read error: Connection reset by peer]
mexisme has quit [Ping timeout: 245 seconds]
grumble has joined #nixos
mexisme has joined #nixos
<wedens[m]>
adisbladis: usb sound card appears in lsusb, but not visible as sound device in alsa/pulseaudio. problem started after upgrading to 19.09.1254.9104be2ee08 from 19.09.1254.9104be2ee08. it works, if I rollback
<wedens[m]>
* from 19.09.1172.2d896998dc9
cosimone_ has quit [Remote host closed the connection]
cosimone_ has joined #nixos
mexisme has quit [Ping timeout: 252 seconds]
blumenkranz has joined #nixos
<blumenkranz>
Hello. I'm having trouble launching the latest Steam update. It complains about a missing libva.so and libva-x11.so. Is anyone else having the same issue?
<adisbladis>
blumenkranz: It's been fixed on master, but the fix has not yet hit unstable.
dckc has joined #nixos
<blumenkranz>
Well damn.
<adisbladis>
blumenkranz: Or it seems like it should have by now. Try updating your channels
<andi->
wasn't that in some deprecated.nix file ? :-)
<magneticduck>
this is the pinnacle of silly software engineering
<adisbladis>
andi-: The deprecated.nix file that doesn't have any explanations why anything was deprecated or even when it's gonna be removed :P
<andi->
adisbladis: exactly.
<Neo-->
Hey hey! Is there anything out there that takes a look at the changelog and user-specified configuration.nix and compares it to specified version and highlight if there are any packages / services mentioned that are affected by the upgrade?
* andi-
removes the fake sha
tsrt^ has joined #nixos
MarvelousWololo_ has quit [Ping timeout: 276 seconds]
<{^_^}>
[nixpkgs] @c0bw3b pushed to release-19.09 « electrum-dash: mark as vulnerable »: https://git.io/JeoaT
Thra11_ has quit [Ping timeout: 265 seconds]
<blumenkranz>
adisbladis: Ah fug, it works now. Many thanks!
<{^_^}>
[nixpkgs] @c0bw3b pushed to release-19.03 « electrum-dash: mark as vulnerable »: https://git.io/JeoaG
blumenkranz has quit [Remote host closed the connection]
<magneticduck>
I'm trying to build a python 3 package, but when I try to build this file, I get an error that looks like the python 3 interpreter is being pointed as a special nix-related build script written in python 2!
<magneticduck>
what's going on?
cosimone_ has quit [Remote host closed the connection]
<ddima>
wedens[m]: so, upgraded to the latest version (nixos-19.09.1258.07e66484e67) and didnt notice any changed behaviour with my audio interface
magnetophon has joined #nixos
<Ralith_>
how would I get a stdenv with a glibc built optimized with debug symbols?
reallymemorable has quit [Quit: reallymemorable]
<magnetophon>
I just upgraded unstable and in the course of that changed from slim to lightdm. Now my displayManager.sessionCommands are not ran anymore. How do I troubleshoot that?
cosimone has joined #nixos
cosimone has quit [Client Quit]
cosimone has joined #nixos
bvdw has quit [Quit: bvdw]
bvdw has joined #nixos
alex``` has quit [Ping timeout: 240 seconds]
gwen has joined #nixos
hmpffff has quit [Quit: nchrrrr…]
<hyper_ch>
eyJhb: I don't think that kde problem was becaus of displaylink... after first reboot in old generation I had same issue and second reboot in it worked again
alex``` has joined #nixos
oxford has joined #nixos
cosimone has quit [Remote host closed the connection]
cosimone has joined #nixos
cartwright has quit [Ping timeout: 260 seconds]
magnetophon has quit [Remote host closed the connection]
turion has joined #nixos
turion has quit [Client Quit]
cosimone_ has joined #nixos
cosimone has quit [Read error: Connection reset by peer]
siraben has left #nixos ["Kicked by @appservice-irc:matrix.org : issued !quit command"]
cosimone has quit [Remote host closed the connection]
hmpffff has quit [Ping timeout: 276 seconds]
cosimone has joined #nixos
LysergicDreams has quit [Ping timeout: 240 seconds]
hmpffff has joined #nixos
lovesegfault has quit [Client Quit]
<duairc>
Does anybody know how to tell NixOS to automatically decrypt a LUKS partition on boot? It's not the root partition, and it has to be done *after* the root filesystem is mounted, because the key is on the root filesystem. The LUKS-related NixOS options that I can find all seem intended for before the root filesystem is mounted
<duairc>
I guess I can just crypttab?
exfalso has joined #nixos
cosimone has quit [Remote host closed the connection]
cosimone has joined #nixos
exfalso has quit [Ping timeout: 240 seconds]
cosimone has quit [Excess Flood]
cosimone has joined #nixos
rardiol has quit [Read error: Connection reset by peer]
<random_user>
When I connect a SATA drive to my system and run lsblk, none of my drives are recognised, I am only able to see my boot drive (NVMe). What have I done wrong?
<tilpner>
jco: That was the wrong hash
<clever>
random_user: did you plug it in while the machine was off?
rnhmjoj has quit [Changing host]
rnhmjoj has joined #nixos
rnhmjoj has joined #nixos
Thra11 has joined #nixos
<sondr3>
I'm modularizing my nix configuration and have looked at how others do it, is there a difference between having a `options.mine.blah.enable = ...` vs `let cfg = config.mine.blah in {`?
<clever>
sondr3: one is using options which isnt the right way to read the config
<rnhmjoj>
do you think issue #8247 could be pinned in the security project?
<Thra11>
If you're not on your most recent NixOS 'generation' (e.g. you've gone back and forth a bit and ended up on generation 6 of 10), and you do a nixos-rebuild switch, resulting in generation 11, does NixOS 'remember' which generation you were on before (because `nixos-rebuild --rollback switch` will take you to current - 1, i.e. 10, not where you were before, 6)?
<clever>
Thra11: rollback goes to current-1, not previous, ive ran into that problem before
<Thra11>
clever: I know. I'm wondering if there's a solution to the problem...
<clever>
Thra11: you could delete the bad generations after you rollback, so they are skipped
<{^_^}>
dansho: Don't install libraries through nix-env or systemPackages, use nix-shell instead. See https://nixos.wiki/wiki/FAQ/Libraries for details.
<evanjs>
Wait why am I using cachix worried about secrets when `nix.sshServe.enable` is a thing 🤔 is there really anything complex about the setup required compared to cachix?
<clever>
evanjs: you could garbage collect things with sshServe, and you need enough storage and bandwidth to host the things you want to share
<clever>
evanjs: but if its only for personal use, sshServe seems simpler
<clever>
dansho: and how did you load that nix file?
knupfer has joined #nixos
<dansho>
$ nix-shell in the current directory
<dansho>
shell.nix
<evanjs>
Ahhh wait then I won’t have to push, either, right? And yeah it’s not really a concern with storage atm. Basically I have three machines and I was hoping I could let the other machines use the results from the first for things like linux kernel
<evanjs>
My only problem atm is timing, then
<Thra11>
I was hoping maybe there was something simple, like a log of nixos switching history
<clever>
evanjs: yeah, sshServe just serves whatever is in /nix/store/
jco has quit [Quit: WeeChat 2.6]
<evanjs>
Currently all machines are set to auto upgrade every six hours. So I would have to offset that somehow
<clever>
Thra11: not really
<evanjs>
Like “wait until rig finishes upgrading and then try” hrm
<clever>
evanjs: i try to avoid auto-upgrade, since it can lead to unexpected breakage
<evanjs>
Especially when I use IFD 🤪
<evanjs>
I’ve had that multiple times now. `nixos-option` stops working, I have to rebuild, etc ...
<evanjs>
Wait autoUpgrade breaks things? Not auto collect?
<clever>
evanjs: not directly, but the changes it brings in may
<evanjs>
More than manually rebuilding?
<clever>
evanjs: for example, slim was recently deprecated, so my config no longer even builds on the latest nixos-unstable
<clever>
if i want to keep using it as is, i need to avoid updating channels, until i fix that
<clever>
autoUpgrade would keep updating the channel on me, and cause Thra11's issue, where i have to issue 20 rollbacks to get to a working revision
<evanjs>
Ohhh like that. Yeah like my config stopped updating because angular broke (#73225), qdirstat, too, but that’s fixed now
<evanjs>
Argh that sounds nice. So I wouldn’t even have my machines doing the work all the time
<clever>
and hydra will distribute the work between things, and avoid duplicating it
<evanjs>
Yeah that sounds awesome. Guess I need another box for hydra, then..
<clever>
evanjs: initially, i ran hydra on the router, but the hdd was too io bound
<clever>
evanjs: so i moved it to my nas
<evanjs>
That would solve my linux + VirtualBox upgrade issues, too.
<evanjs>
I could also just set some rules up to tell stuff to go ahead with the upgrades for me if certain conditions are met, otherwise notify me, etc
<evanjs>
I imagine that sort of thing might be more feasible with hydra, etc?
<clever>
evanjs: thats basically how nix-channel's are managed
<evanjs>
Vs a basic auto-upgrade setup
<evanjs>
Ah alright
<clever>
evanjs: you could even just have your own custom channel, that only updates when all machines build, then point auto-upgrade to that!
<{^_^}>
jkff/digest#5 (by bgamari, 1 year ago, open): Use pkgconfig to find zlib
<evanjs>
eyJhb: yah exactly. I could finally just not install it everywhere if VBox still needs to be built. And etc
<clever>
i havent changed any special flags, so my vbox comes from the binary cache
<dansho>
it seems to be a transitive dependency, not even sure where its coming from
<evanjs>
clever: “router”. Right but what’s your kernel? I try to stay on latest, just because.
<clever>
evanjs: 4.19.37 on the router
<eyJhb>
Yeah, vbox only needs to rebuild locally if guest additions are active
Rusty1 has joined #nixos
<clever>
evanjs: guest additions work on my windows guest, without having to build the host stuff
<evanjs>
eyJhb: clever: right right. That makes sense.
<clever>
i think virtualbox only needs to be compiled specially for usb3 support to the guest and some of the more advanced features
<clever>
screen recording and rdp dont work for me
random_user has quit [Remote host closed the connection]
Fare has quit [Ping timeout: 240 seconds]
_0bitcount has quit [Quit: Leaving]
MichaelRaskin has quit [Ping timeout: 250 seconds]
<evanjs>
clever: right that was my understanding. Hrm. Maybe I could just disable that for now as I haven’t really needed any of those for some time. Definitely prioritizing setting up a hydra instance, though
Fare has joined #nixos
FRidh has quit [Quit: Konversation terminated!]
<evanjs>
Then I’ll just need a Mac Mini or something so I can do iOS dev and test all my PRs on Darwin and start to make more packages compatible in nixpkgs and etc. (it’s hard waiting on stuff on my 2011 MBP)