<zeta_0>
clever: ok, so i undid all the changes i made in ghc.nix's default.nix file, bringing it back to it's default state and renamed it to ghc.nix(https://dpaste.org/SDyv), and then i ran `lorri init` and wrote the callPackage function you recommended in lorri's generated shell.nix file(https://dpaste.org/ONCQ)
<zeta_0>
clever: does my custom configured ghc.nix setup look correct? there's no nix syntax errors being thrown, so i think that the syntax is correct.
<{^_^}>
[nix-mode] @yilinwei opened pull request #103 → Add autoloads for eshell functions → https://git.io/JJR8T
<installingNIX>
Well goodbye guys thanks for help especially thanks to clever
installingNIX has quit [Remote host closed the connection]
Rusty1 has joined #nixos
sigmundv_ has joined #nixos
nij has joined #nixos
<nij>
Hello.. I'm learning nixOS by following nix pill. I'm stuck at the 10th chapter. I wonder if anyone could help..thank you.
<nij>
I'm stuck in the second command: `[nix-shell]$ source builder.sh
orivej has quit [Ping timeout: 264 seconds]
<nij>
I have entered nix-shell, and followed the instruction to source my `builder.sh`..
orivej has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JJR8P
<nij>
Unfortunately, that's a builder introduced in the earlier chapters
<nij>
The error I got from `[nix-shell]$ source builder.sh` is that `./configure: No such file or directory`.
<nij>
It seems that my nix-shell does not know what `configure` is.
<zeta_0>
in what part of the nixos wiki is the `callPackage` explained? https://nixos.wiki/
<nij>
I have included "gnutar gzip gnumake gcc binutils-unwrapped coreutils gawk gnused gnugrep findutils patchelf
<nij>
Aren't those not enough for the nix-shell to know `configure`?
<zeta_0>
clever: nope, the callPackage function needs to be placed somewhere else: `error: attempt to call something which is not a function but a set, at /home/zeta/ghc/ghc.nix/shell.nix:4:1
<zeta_0>
`
<zeta_0>
hmm, how strange, there was no syntax errors thrown
orivej has quit [Ping timeout: 264 seconds]
orivej has joined #nixos
<infinisil>
nij: What's your current Nix file?
johabi has quit [Quit: Leaving]
<nij>
infinisil: thanks for reaching out, it's `hello.nix`, an example given in nixpill as well.
<nij>
Lets see if I can fetch its content.. it's in a virtual machine.
<zeta_0>
does the placement of this callPackage function look correct? when i placed it there, emacs stopped throwing syntax errors, and the nix-shell works now
<infinisil>
nij: With default.nix as you showed, autotools.nix from that section and builder.sh as above (plus the tarball downloaded), I can successfully do `nix-shell` then `source builder.sh` to build the package
<infinisil>
Maybe you didn't download the tarball?
<nij>
Hmm.. with nix-build it worked.
<nij>
And in chap 10 we started using `nix-shell`. And it didn't work.
<nij>
Instead of `for p in $buildInputs $baseInputs`, I wrote two for loops.. I think they are equivalent.
<infinisil>
nij: Oh, do you perhaps have another directory in your pwd/
<infinisil>
?
<nij>
And.. I do have the tarball
<infinisil>
Because the builder.sh enters any directory it can find, this might be the one it just unpacked with the tar command, but it might be another one you have lying around
<infinisil>
And if it's another one, it wouldn't be able to find the configure file from the hello source
<nij>
lemme see
<nij>
I'm in my home dir
<nij>
there'are lots of dirs
<infinisil>
Yeah then that's it :)
<nij>
.cache .. .config ..
<infinisil>
I suggest creating a new empty directory for doing this instead then
<nij>
thank you :)
<infinisil>
Np :)
<infinisil>
(note that bash's * doesn't expand to dotfiles by default, so it shouldn't be those dirs)
<nij>
infinisil: YAY it worked .. thank you
<infinisil>
Awesome
<nij>
Yes..I read bash.. I'm overwhelmed by nixos.
<nij>
Fun indeed though. Appreciated :)
<tinamus[m]>
Hey guys, I just installed nixos and I am trying to setup everything. I use zsh, oh-my-zsh, and powerlevel10k on arch, I don't know how to change the theme of oh-my-zsh to powerlevel10k in nixos
<nij>
tinamus[m]: Me too at a similar stage. I reckon it's very different way. Don't remember how but I'm persuaded to patiently go through nixpill in order to understand.
<nij>
I wish there's another tutorial that teaches directly how to setup and compile on our own, and explain why it works.
<infinisil>
nij: Tbh, I don't think the pills are very useful in practice
<infinisil>
They teach you how Nix works underneath, but help very little in understanding that is the monstrosity called nixpkgs
<evanjs>
it's a little weird (pulled mostly from infinisil :P) but some commands are defined in the nix expression and then folded into the haskell file
<tinamus[m]>
What a life saver, thank you!
<infinisil>
Hehe
nij has quit [Ping timeout: 260 seconds]
joehh1_ has quit [Ping timeout: 264 seconds]
<infinisil>
Man, my nixos config is super messy
<infinisil>
Maybe I should set a goal of improving it a tiny bit every day
zeta_0 has joined #nixos
<tinamus[m]>
<infinisil "Maybe I should set a goal of imp"> always a good goal haha
<zeta_0>
i'm able to go into a nix-shell without any errors, but this callPackage function is not working? https://dpaste.org/u3HE
jmeredith has quit [Quit: Connection closed for inactivity]
jdawg has joined #nixos
MarcWebe1 has joined #nixos
MarcWeber has quit [Ping timeout: 240 seconds]
aw has quit [Quit: Quitting.]
spacefrogg has quit [Quit: Gone.]
aw has joined #nixos
spacefrogg has joined #nixos
<energizer>
ssh is rejecting public key that's in /etc/ssh/authorized_keys.d/energizer, generated by users.users.energizer.openssh.authorizedKeys.keys. how to diagnose?
<ornxka>
how do i turn a file into a string in nix?
<ornxka>
ive got a yuge file that would be really annoying to modify in any way and id like it to be a string that i can reference in my configuration.nix
<clever>
ornxka: builtins.readFile ./file
<ornxka>
ahhh ty
<{^_^}>
[nixpkgs] @alyssais opened pull request #93888 → libhugetlbfs: init at 2.22 → https://git.io/JJRgi
<ornxka>
is that meaningfully different from lib.strings.fileContents
<ornxka>
which i just found out about 10 seconds before you said that
<clever>
ornxka: fileContents just deletes the trailing \n at the end of the string
palo1 has joined #nixos
<ornxka>
ahhh
<clever>
so if you read a "12\n" into nix, it becomes "12"
<clever>
editors and echo tend to add that junk at the end
<ornxka>
readFile is what i want then, ty
* ornxka
has emacs configured to not put trailing newlines
<clever>
and for multi-line files, it matters less
palo has quit [Ping timeout: 246 seconds]
palo1 is now known as palo
drakonis has quit [Quit: WeeChat 2.8]
Fulgen has quit [Ping timeout: 260 seconds]
<{^_^}>
[nixpkgs] @marsam opened pull request #93889 → python38Packages.pandas: fix build on darwin → https://git.io/JJR2f
ardumont has quit [Ping timeout: 272 seconds]
ardumont_ has joined #nixos
inf has quit [Ping timeout: 258 seconds]
<kini>
I have a couple of LUKS volumes whose keyfiles are in `/boot`. I'd like them to be decrypted at some point in the boot process, but using `boot.initrd.luks.devices` makes the system try to decrypt the volumes before the `/boot` fs is loaded. Is there some way to get those volumes to be decrypted at a later point?
<energizer>
kini: there's a thing called fileSystems.foo.encrypted but idk what it does
prehonor has joined #nixos
<kini>
energizer: looking at the implementation in `nixos/modules/tasks/encrypted-devices.nix`, it seems like it just maps those things into `boot.initrd.luks.devices` :(
<kini>
oh wait, no -- it excludes the ones that don't have keyfiles, so maybe this will work...
<{^_^}>
[nixpkgs] @FRidh pushed commit from @r-ryantm to master « python37Packages.bravia-tv: 1.0.5 -> 1.0.6 »: https://git.io/JJRV0
<kini>
Yup, it sets `boot.initrd.postMountCommands` to add `cryptsetup open` there. Perfect! Not sure how it will then mount the decrypted `/dev/mapper` volume after that, but that's somewhere to start...
steshaw has quit [Remote host closed the connection]
<galagora>
Hey, how do libraries end up in /run/current-system/sw/lib/ ? Adoptopenjdk11 ends up there while oraclejdk11 (from master) seemingly does not, despite having essentially identical nix expressions
<edef>
hrm, that shouldn't really be happening at all
<{^_^}>
nix#2377 (by Valknjoggr, 1 year ago, open): Installer assumes liberal umask
ManiacOfMadness has joined #nixos
<steshaw>
I set the umask back to 022 and re-ran the installer. This required me to delete the old installation and re-run the installer. Folks here probably know the drill.
<steshaw>
After 2 install, I got a successful one with "Alright! We're done!"
<steshaw>
However, when I open a new shell, nix-shell is not found in my PATH. So the recommended test command fails (nix-shell -p nix-info --run "nix-info -m").
<{^_^}>
[nixos-homepage] @github-actions[bot] pushed commit from GitHub Actions to master « Update flake.lock and blogs.xml [ci skip] »: https://git.io/JJRKb
<nij>
==============
orivej has joined #nixos
<nij>
Could someone help me on configuring xmonad? I have @evanjs's config in my /home/me/xmonad/xmonad.hs
<nij>
I have `services.xserver.windowManager.xmonad.enable = true` in `/etc/nixos/configuration.nix`, and
<nij>
have 'services.xserver.windowManager.xmonad.config = /home/me/xmonad/xmonad.hs` in `/etc/nixos/configuration.nix` too.
<nij>
I've run `nixos-rebuild switch. No errors. However
<TheSirC[m]>
Let say that you have a bunch of overlays in the `overlays` directory (sometime in nested folders), how would you import them as overlay without breaking you whole system whit infinite recursion ?
<nij>
1. `nix-env --list-generations` doesn't suggest that I created a new generation.
codezero has joined #nixos
<nij>
2. `Even after reboot, it doesn't seem that my xmonad has changed.
<nij>
Maybe I missed some step?
<nij>
Btw, I just installed my nixOS, and have almost nothing else configured.
<nij>
(So perhaps I need to configure my haskell env?)
<lassulus>
nij: these steps should normally be enough, can you try setting some other setting and see if it has any effect?
<nij>
lassulus: Thanks for reaching out! What do you mean by switching settings?
nature has joined #nixos
<nij>
*other settings
<lassulus>
do some other settings inside your configuration.nix take effect? for example if you change your default text edit, or enable another service
<nij>
lassulus: I have set xmonad to disable. Ran `nixos-rebuild switch`
<nij>
nix-env --list-generations doesn't suggest any new generation is created. However, lemme first reboot and see what happens.
<{^_^}>
[nixpkgs] @danieldk opened pull request #93896 → rocm-opencl-runtime: use unaltered stdenv → https://git.io/JJR6u
<nij>
Weirdo! It did take away my xmonad. So I'd say the change is set.
<nij>
Also, I just noticed that `list-generations` suggests new gens are created in my normal user. I couldn't see any change because I ran it as root. (dunno why such difference)
<lassulus>
yeah, there are different generations for root and each user
<nij>
lassulus: hmm.. I didn't know that I was configuring for the normal users.
<lassulus>
that seems indeed a bit weird
<nij>
oh
<nij>
I switched it back, and ran xmonad without startx
<lassulus>
ah, what display-manager are you using?
<nij>
of course xmonad did not start, but I can now see its error message
asymptotically has joined #nixos
<nij>
It's in a virtual machine, and I haven't learned how to copy and paste.. lemme try
nature has quit [Ping timeout: 272 seconds]
<nij>
Lemme type it out manually..
nickeraqua has quit [Ping timeout: 245 seconds]
<nij>
XMonad is recompiling and replacing itself another XMonad process berrent prcess is called "xmonad" but the compiled config should be called "xmonad-x86_64-linux"
<nij>
XMonad will use ghc to recompile, because "/home/me/.xmonad/build" does not exist.
<nij>
XMOnad skipping recompile b/c it is not forced (e.g. via --recompile),and neither xmonad.hs nor any *hs *lhs *hsc files in /lib have been changed.
<nij>
/home/me/.xmonad/xmonad-x86_64-linux: executeFile:does not exist (No such file or directory)
<nij>
(( that's all ))
afreak has quit [Remote host closed the connection]
<lassulus>
what command did produce that error?
<nij>
xmonad
<nij>
`xmonad`
<lassulus>
ah, well it doesn't take your configuration into your account, which you have configured with services.xserver.windowManager.xmonad.config
<nij>
seems so..
nixbitcoin_ is now known as nixbitcoin
<lassulus>
yeah, the default xmonad wrapper tries to run from $HOME/.xmonad/xmonad.hs
<lassulus>
but thats not the same xmonad which is running if you configure services.xserver.windowManager.xmonad.config
bukkitgerman8608 has quit [Ping timeout: 256 seconds]
<lassulus>
but you could try moving your xmonad-config to the expected location and see what fails?
ngerstle[m] has joined #nixos
<nij>
Lemme do that
<nij>
I'll do it by backing up the original ~/.xmonad
<nij>
and symlink
bukkitgerman8608 has joined #nixos
WilliButz has quit [Quit: bye]
WilliButz has joined #nixos
philr_ has joined #nixos
<nij>
OK. Now `xmonad` tries to compile, but gave error:executeFile does not exist.
<nij>
Lemme dirty fix it by `nix-env -iA nixos.executeFile`
<nij>
Rebooting makes `startx` working again. However, there's still seem to be an undefined stuff in haskell.
<nij>
xmonad.hs:325:55: error:
<nij>
Not in scope: 'Search.promptSearchBrowser''
<nij>
Perhaps you meant one of these:
<nij>
'Search.promptSearchBrowser' (imported from XMonad.Actions.Search),
<nij>
----
<lassulus>
can't see the difference?
<lassulus>
maybe the extra ' ?
<nij>
... Eww this seems to be that evanjs[m] has redefined Search.promptSearchBrowser'
<nij>
yes
<nij>
lemme find where that is in the repo
knupfer has joined #nixos
tmaekawa has joined #nixos
newman has quit [Remote host closed the connection]
aflatter[m] has joined #nixos
tmaekawa has quit [Client Quit]
alp_ has joined #nixos
malook has quit [Ping timeout: 256 seconds]
<nij>
That's the only instance of `Search.promptSearchBrowser'` in the repo I believe.. I take the `'` manually away, and now it's not complaining about that.
<nij>
New complaints are "autorandr: command not found", "@xmobar@": command not found.
<nij>
lassulus: But this I should spend some time working through. There has been great process. Thank you so much.
<nij>
progress&
<nij>
progress*
<lassulus>
nij: yw, have fun with NixOS!
<nij>
lassulus: Thanks :D
knupfer has quit [Ping timeout: 264 seconds]
alp_ has quit [Ping timeout: 246 seconds]
cap has quit [Ping timeout: 244 seconds]
orivej has quit [Ping timeout: 264 seconds]
orivej has joined #nixos
civodul has joined #nixos
<{^_^}>
[nixpkgs] @petabyteboy merged pull request #93854 → tor: build with standard libevent → https://git.io/JJRna
<{^_^}>
[nixpkgs] @petabyteboy pushed to master « tor: build with standard libevent (#93854) »: https://git.io/JJRXl
<koozz>
hi! I have NixOS running for just one month, but I'm really liking it. They only thing that isn't working on my laptop is the microphone. The whole audio-in is not working (also with headphones that have a mic). Sound is working but 'Audio Volume Settings' state: 'No Input Devices Available'. I thought I might be able to contribute a hardware setup
<koozz>
for 'nixos-hardware/hp/elitebook/840-g6' but I don't actually know what soundcard or modprobe to configure. Are there any pointers to what I can try?
<Miyu-saki>
koozz: Does it appear on Alsa or nay?
<{^_^}>
[nixpkgs] @freezeboy opened pull request #93902 → kwave: init at 20.04.3 → https://git.io/JJRXS
orivej has quit [Quit: No Ping reply in 180 seconds.]
work_ has joined #nixos
orivej has joined #nixos
<koozz>
Miyu-saki How do you check? The alsamixer shows a capture volumebar.
inf has joined #nixos
zupo has joined #nixos
installingNIX has joined #nixos
<{^_^}>
[nixos-hardware] @afreakk opened pull request #180 → dell/xps/13-9360: conflicting font fix → https://git.io/JJR1T
<installingNIX>
Hello
<installingNIX>
I am installing nix and happy to chat from my first time booted nixos. it seems I have a problem with my boot configuration
<{^_^}>
[nixpkgs] @vcunat pushed to master « pantheon: more build fixes by using older vala »: https://git.io/JJR9a
<immae>
I have several packages in system path which adds elements to /run/current-system/sw/share, and vim in particular does not (in spite of the derivation having a /share/vim), what can I do to have vim put it to this place?
<immae>
It work! How does nix decide to put it or not though? I have for instance liquidprompt which has the path automatically even though it doesn’t have this pathsToLink anywhere
dsiypl4_ has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @Infinisil merged pull request #93813 → Ensure wpa_supplicant.conf is written when userControlled and extraConfig are used → https://git.io/JJBqi
<simpson>
nixpkgs has thousands more fresh packages than any other distro, and the next-closest distro is the stable snapshot of nixpkgs. Those thousands of open PRs are *extra* freshness, on top of our already-excellent fresh package count.
ravndal has joined #nixos
<simpson>
That said, the main things that could be done are already underway. For example, maintainers could use more powerful tools to test and accept more PRs. CI could lead to merge-on-green for some selected packages. And unused or defunct packages can be removed.
<simpson>
If you see a PR that interests you, but it's stale and nothing's happened for a while, test it out and add a comment saying how your test went. End-user feedback is *very* important for knowing whether a PR is good to merge.
<omasanori[m]>
OK, now I understand. Thank you for your explanation and letting me know repology.org. I guess it should be useful to determine some aspects of distros' health.
igghibu has joined #nixos
wolfshappen has joined #nixos
sangoma has quit [Ping timeout: 240 seconds]
zupo has joined #nixos
igghibu has quit [Client Quit]
xeijin has joined #nixos
<xeijin>
hello - having some trouble mounting disks on startup
<xeijin>
it appears the UUIDs for each disk in `hardware-configuration.nix`, generated in the installer using `nixos-install` are not the same as the ones the disks are actually identified by on first boot?
<xeijin>
can anyone tell me why this would be?
<srhb>
xeijin: Are you sure it's by-uuid paths in hardware-configuration?
<xeijin>
@srhb yes, but the UUIDs in `/etc/nixos/hardware-configuration.nix` are different to the ones that were generated in `/mnt/etc/nixos/hardware-configuration.nix` when I ran `nixos-install`
<xeijin>
If it makes any difference, my root and home partitions are both zfs stores
wolfshappen has quit [Ping timeout: 240 seconds]
<xeijin>
though I had no trouble decrypting them
<srhb>
xeijin: I'm confused. There should only be one relevant hardware-configuration.nix for your new system, and that's the one in /mnt/etc/nixos
<srhb>
xeijin: Which _becomes_ your final /etc/nixos/hardware-configuration.nix file in the booted system
<srhb>
xeijin: If those differ, it means you accidentally overwrote one or didn't generate it in the place you thought you did.
<{^_^}>
[nixpkgs] @FRidh pushed commit from @jonringer to staging « python3Packages.pybind11: 2.4.3 -> 2.5.0 »: https://git.io/JJRxa
<{^_^}>
[nixpkgs] @FRidh merged pull request #93911 → libuv: disable flaky test "udp_multicast_join" on darwin → https://git.io/JJRyA
<{^_^}>
[nixpkgs] @FRidh pushed commit from @thefloweringash to staging « libuv: disable flaky test "udp_multicast_join" on darwin »: https://git.io/JJRx6
<xeijin>
srhb without re-installing? How do I do that?
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<ornxka>
dear #nixos this is your daily reminder to clean out your /nix/var/nix/gcroots/auto folder, it is probably full of junk preventing gigabytes of store paths from being properly garbage collected
<hexa->
awesome, I can save 96K
<ornxka>
thirty years ago you would be thrilled to have saved that much space!
<hexa->
indeed
<zangi>
why does nixos use `users` with gid `100` instead of `1000`?
<ornxka>
arent we all daemons in a sense
shibboleth has quit [Quit: shibboleth]
reanimus has joined #nixos
doyougnu has quit [Quit: ERC (IRC client for Emacs 26.3)]
<shapr>
ornxka: how do I clean my gcroots?
<hexa->
zangi: so does debian
<hexa->
I don't have any other machines I can quickly check, but feels standard
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
orivej has quit [Ping timeout: 260 seconds]
orivej has joined #nixos
<zangi>
no, debian and other major distros use alice:1000 for first user, bob:1001 for second user, john:1002 for the third user
reanimus has quit [Quit: Logging off...]
<zangi>
on nixos though, it's alice:100, 100 is `users`
<zangi>
so normally, it's 1000:1000, 1001:1001, 1002:1002, 1003:1003, while nixos: 1000:100, 1001:100, 1002:100
andreas303 has quit [Ping timeout: 240 seconds]
sangoma has joined #nixos
<zangi>
I've noticed that crux distro uses this as well (100?:100)
knupfer has quit [Ping timeout: 246 seconds]
<hexa->
zangi: there are those distros that put all users in the same group and those that create per user groups
<hexa->
neither is uncommon
<zangi>
oh okay, didn't know about that
<zangi>
is there a documentation about this?
<hexa->
not sure
<hexa->
but remember that groups are about sharing permissions
reanimus has joined #nixos
<hexa->
so the users group is a sensible thing in that people can share files in a group by assigning group permission
andreas303 has joined #nixos
noudle has joined #nixos
orivej_ has joined #nixos
<shapr>
After upgrading to nixOS 20 from nixOS 19, I now have two sound cards, and the default is my nvidia HDMI port. Is there a way I can set the other sound card to be default?
orivej has quit [Ping timeout: 246 seconds]
zupo has joined #nixos
Fare has joined #nixos
<buffet>
does nix-shell use <nixpkgs>?
<buffet>
sure seems like it, interesting
wnklmnn has quit [Ping timeout: 256 seconds]
noudle has quit []
<DigitalKiwi>
you can use -I to change it
Fare has quit [Ping timeout: 246 seconds]
knupfer has joined #nixos
malook has quit [Ping timeout: 240 seconds]
<M0-[m]>
is there any way to replace a single member of a list with nix? I'm trying to override a package attr but the attr is a list that includes the package which I want to replace
orivej_ has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #nixos
<ornxka>
shapr: you just look in /nix/var/nix/gcroots/auto and it will have a bunch of symlinks to stuff in your filesystem (typically a bunch of symlinks called "result" that point to store paths, which are the result of running nix-build), you see which ones you want to keep and then you simply delete the ones you dont
<{^_^}>
[nixpkgs] @worldofpeace pushed 4 commits to master: https://git.io/JJ0Je
<ornxka>
then you do a garbage collect and the nix store paths those symlinks you deleted pointed to will be garbage collected
<shapr>
Anyone know how to tell mpd to use a particular audio device? Right now I can play notification sounds from pavucontrol, but mpd won't talk to my sound device.
<eyJhb>
MichaelRaskin: Clean install, so it should be default everything
zebrag has quit [Read error: Connection reset by peer]
<eyJhb>
Good place to set it?
<MichaelRaskin>
I would expect that unset goes to /tmp!
zebrag has joined #nixos
<MichaelRaskin>
Maybe you used an RPi profile that overrides
<eyJhb>
Using no profiles, only a default config
<eyJhb>
Also, I think the same goes for a normal NixOS system, that it will use /run/user
<MichaelRaskin>
How long I have not used NixOS!
<eyJhb>
:D
<eyJhb>
I can't find anything in the dock
<eyJhb>
doc**
<eyJhb>
I think it is basically this `boot.runSize` and then it uses the run dir
<LouisDK>
I'm about to make a new package for NixOS, however the make-command and I want to enter the proper /nix/store-location to view the build-files - however the build directory are removed after build fails. Can I preserve it in some way?
noudle has joined #nixos
malook has joined #nixos
fendor has joined #nixos
dongcarl has joined #nixos
<eyJhb>
MichaelRaskin: nix-shell -p xxx sets TMPDIR to /run/user/x...
zangi has quit [Ping timeout: 246 seconds]
<pie_>
,locate gdk_pixbuf-2.0.so
<{^_^}>
Couldn't find in any packages
orivej has quit [Quit: No Ping reply in 180 seconds.]
davidv7 has quit [Ping timeout: 264 seconds]
orivej has joined #nixos
<Ke>
protip, use btrfs for /nix and reinstall on every release and never collect cruft
<hexa->
btrfs for /nix with zstd compression <3
<Ke>
since reinstall is basically free anyway, there is no real effort there
<hexa->
there is some state you really want to keep
tst60 has joined #nixos
<Ke>
I did this even on debian, where reinstall has slight cost
o1lo01ol1o has quit [Remote host closed the connection]
magthe has quit [Quit: WeeChat 2.9]
<energizer>
Ke: isn't `nix-collect-garbage -d` sufficient to avoid collecting cruft in /nix?
orivej has quit [Quit: No Ping reply in 180 seconds.]
<LouisDK>
I'm building a new package for nix, but the build fails and the /nix/store/..pkg.. is not preserved afterwards sp I can expect the builddir and debug. Any better way of debug than putting things like ls etc. inside default.nix to debug?
tst60 has left #nixos [#nixos]
orivej has joined #nixos
ninjin has quit [Remote host closed the connection]
<hexa->
karetsu: the general answer is: it depends.
<hexa->
generally on a set of tests succeeding
<karetsu>
s
<karetsu>
sorry, ty*
<xacktm>
MichaelRaskin: fair enough
<archH>
Doubt removing modesetting will fix my issue though. When I don't specify nvidia as video driver I get 11 fps in a game where I would get 150 fps on windows and that is probably capped. But with it I get Assertion Failed: Fatal Error: OpenGL GLX extension not supported by display
<rjmill>
does "nix repl" heed any user config files? I tried inputrc and editrc, and neither seemed to work. (specifically looking for vi keybindings in the repl. If I can get that, I'm happy)
<karetsu>
I'm getting a 'discord-haskell-1.6.0 in /nix/store/.../hackage-packages.nix' is marked broken, but on github the file says 1.6.1 and I have no 1.6.0 marked *anywhere* - can't seem to get it to update
<ivan>
rjmill: rlwrap nix repl?
knupfer has quit [Ping timeout: 244 seconds]
arjen-jonathan has quit [Ping timeout: 260 seconds]
<{^_^}>
[nixpkgs] @FRidh pushed commit from @r-ryantm to master « kapacitor: 1.5.5 -> 1.5.6 »: https://git.io/JJ0Tp
malook has quit [Quit: malook]
<shapr>
Does the order of my config matter? I have mpd starting before jack, and mpd refuses to start.
<rjmill>
The source code for "nix repl" has an "#ifdef READLINE" block, so I'm guessing it's possible to build it with readline instead of editline? https://git.io/JJ0ke
<rjmill>
I'm not sure how I'd go about accomplishing that though.
<karetsu>
hexa-: so why would my nix store be using a version of a file in /nix/store/*nixos-20.09pre*/nixos/pkgs/development/haskell-modules/hackage-packages which is older than the one on github when I have not long run (20 minutes ago) a `nixos-rebuild switch --upgrade`?
<shapr>
Is there some way to see the logs from mpd when it fails to start? I don't see anything useful from "journalctl -xe"
<DigitalKiwi>
shapr: maybe; systemctl status mpd
<DigitalKiwi>
it should at least give you a line you can run manually
<hexa->
karetsu: on github is a vage statement
<hexa->
if you mean nixos/nixpkgs master, that is because you're probably living on a channel
<karetsu>
I'm living on nixos-unstable
<hexa->
a channel only gets published when a certain set of packages build and their tests succeeed
<hexa->
that is a channel
maier has quit [Ping timeout: 265 seconds]
<karetsu>
which that link you gave said was published 4 hours ago>?
<hexa->
10 hours ago
Nerdmaxx has joined #nixos
bpye has joined #nixos
<karetsu>
sorry yes, that's still 9 hours before my upgrade and I still have the the file which doesn't match the one on github
<karetsu>
i.e. older
<srhb>
karetsu: nixos-unstable has not updated in more than five days.
<DigitalKiwi>
if you really want you can use nixos-rebuild <whatever> -I nixpkgs=https://github.org/NixOS/nixpkgs/archive/master.tar.gz
<bpye>
Hey, I’m trying to setup cups for network printing however nixos seems to be making this hard
<hexa->
oh, so status.nixos.org is being silly again :<
<srhb>
karetsu: You may be looking at nixpkgs-unstable
<srhb>
karetsu: Not nixos-unstable.
<bpye>
I want to override extraConf because the default is restricting access, do I need to override the default value?
<hexa->
oh, misread as well
<karetsu>
ah yes, I am
<karetsu>
lol
<karetsu>
my b
<srhb>
karetsu: It'll updated once the remaining blockers pass testing, as hexa explained. :)
<DigitalKiwi>
to answer your earlier question: sometimes a few days sometimes a week or more
archH has quit [Ping timeout: 265 seconds]
<hexa->
oh, that was that macvlan thingy
<srhb>
Yes, but that's been fixed.
<karetsu>
ty, just it seems it broke my nix-shelling and I really don't want to go throwing allowBrokens around
cosimone has joined #nixos
<srhb>
Now there's a new blocker.
<hexa->
yup, I merged that
orivej has quit [Ping timeout: 256 seconds]
<hexa->
ah lovely
<srhb>
Ah, good :)
orivej has joined #nixos
<srhb>
karetsu: Usually things are indeed marked broken for a reason and just unmaking them will not improve the situation anyway.
<srhb>
karetsu: Especially haskell packages are usually marked broken because their builds fail automatic testing, so all you'll get is a still-broken build, but you'll spend more time getting there. :)
<karetsu>
yeah, I just seem to have fallen into a middle ground with nixpkgs-unstable vs nixos-unstable and made a mess somewhere
<eyJhb>
Ohh srhb , you came as called! - Does MOPRH support deploying from a x64 host to aarch64 ?
<hexa->
eyJhb: it does
<srhb>
eyJhb: morph doesn't care about those details much. :)
<eyJhb>
hexa-: anything I need to know ?
<srhb>
Anymoo, off to bed. o/
<hexa->
\o
<eyJhb>
Does it build locally or on the target?
<eyJhb>
Night srhb :)
<hexa->
locally or remote-builder
<hexa->
the target only gets the results
<rjmill>
ivan: `rlwrap -a nix repl` does the trick though. Thanks a bunch for the suggestion!!
<eyJhb>
What do I need to setup, for this to work?
<eyJhb>
Building on my machine, and then pushing to the target?
<ivan>
rjmill: cool
<hexa->
eyJhb: basically you have to inherit the correct system for your packages
<hexa->
eyJhb: yup
<hexa->
just have ssh working
<eyJhb>
inherit the correct system for your packages?
nikita` has quit [Quit: leaving]
<hexa->
yeah, like import packages with system set to aarch64-linux
<eyJhb>
So, if I just set morph up, then it should build on my own machine and push the result to the RPi?
<hexa->
hm, I don't think I have cross compilation working with morph
<eyJhb>
Any public config?
<hexa->
I did build in qemu for a while, now on an rpi4
<Ke>
energizer: in the past we have seen pretty mangled sqlite databases, in general I like to be sure
<Ke>
I plan to also roll out the tmpfs rootfs solution to only have well documented persistent state
<energizer>
is there a way to rebuild the database directly from the store?
<eyJhb>
hexa-: lets see, I have never used anything like MORPH before, but want to use it where I can
<eyJhb>
Wonder if I can use it to build my local machine as well :p
<Ke>
unlike the tutorial, also put the persistent state in a separate volume so you really can reinstall that nixos on the side and just reboot to a new system like nothing changed
<eyJhb>
Ke: I would suggest gchristensen version, with rollback of ZFS. tmpfs can really take some memory if some program writes to a place you did not intend
<Ke>
shapr: I normally follow systemctl show mpd.service | grep Exec to find the configs
<Ke>
or something similar
<shapr>
I found the configs, but I still can't get mpd to talk to pulse or also, not sure why.
o1lo01ol1o has joined #nixos
alp_ has joined #nixos
orivej has quit [Ping timeout: 264 seconds]
orivej_ has joined #nixos
smatting_ has joined #nixos
<Ke>
shapr: journalctl -u mpd.service
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<eyJhb>
lovesegfault: ping, are there some files missing from your nix-config?
<Ke>
and then scroll to end or similar
meh` has joined #nixos
<Ke>
maybe option like -b
<Ke>
shapr: if you think nixos dependency ordering is wrong, you can try that by creating a systemd service that goes after jack and before mpd
<Ke>
this allows you to not fix nixos and solve the issue for yourself once and for all, maybe make a pull request, if you are not lazy
codygman has quit [Read error: Connection reset by peer]
codygman has joined #nixos
leira has joined #nixos
<leira>
Trying to define a systemd user service for [neuron](https://neuron.zettel.page/). Normally, in ExecStart, we refer to a command using `${pkgs.openssh}/bin/ssh-agent`, But in this case, neuron is a custom package I defined in `packages` field. How do I refer to it?
<leira>
I tried ${pkgs.neuron}, but it refers to a different neuron package
<lassulus>
Leira: you can either define the packge in an overlay/packageOverride or reference the file with /run/current-system/sw/bin/neuron in ExecStart
<srid>
Basically you pass that custom package directly.
<leira>
srid, I used that and it worked
codygman has quit [Ping timeout: 240 seconds]
<srid>
Yea I get notified when somebody writes `srid`
<leira>
But I did defined it in home.packages too.
<{^_^}>
[nixpkgs] @ggreif opened pull request #93940 → [20.03] ghc: add new version 8.8.4 → https://git.io/JJ0IB
codygman has joined #nixos
<leira>
So I have this neuronSrc section defined twice in both home.packages and user service sections.
<{^_^}>
[nixpkgs] @marsam merged pull request #93889 → python38Packages.pandas: fix build on darwin → https://git.io/JJR2f
<lovesegfault>
eyJhb: I liked it's way of doing deployments better
<lovesegfault>
no cli tool, just nix-build
<leira>
👍
c0c0 has joined #nixos
karetsu has joined #nixos
<c0c0>
i'm trying to install nixos with UEFI boot. when i try "sudo nixos-install", it says: "You must set the option 'boot.loader.grub.devices' or 'boot.loader.grub.mirroredBoots' to make the system bootable." It looks like it assumes I want to boot in BIOS mode. How can I fix this?
<colemickens>
c0c0: do you have a uefi bootloader enabled?
Leira[m] has joined #nixos
<colemickens>
does it stop complaining if you enable systemd-boot?
<c0c0>
I've set 'boot.loader.systemd-boot.enable = true;'
orivej_ has quit [Quit: No Ping reply in 180 seconds.]
codygman has quit [Read error: Connection reset by peer]
<c0c0>
How do I enable a UEFI bootloader?
codygman has joined #nixos
pjt_014 has joined #nixos
<c0c0>
Should I set 'boot.loader.grub.device = nodev;'?
xkapastel has quit [Quit: Connection closed for inactivity]
<lovesegfault>
I think this is what I need
* lovesegfault
tries
<lovesegfault>
brb
<lovesegfault>
It works!!
<lovesegfault>
You can find the correct mapping number with cat /proc/fb
<eyJhb>
I however have this problem with xrandr, where I need to run this command `xrandr --output LVDS1 --rotation normal --reflection normal` (the reflection part), before I can run any other command with xrandr. And if my laptop is closed, it will simply not work, and force me to run it in my DP2 display...
<colemickens>
grfn: whatcha trying to do? maybe there's another way to slice that bread?
<grfn>
colemickens: connect to a postgres database as part of a test suite
noudle has quit []
<grfn>
nixos tests are currently my best option but that's so slow
<grfn>
this is cursed but I wonder if it's possible to run postgres in a sandbox
<grfn>
like
<grfn>
data dir in a sandbox, listening on a unix socket also in the sandbox
zbozbo[m] has joined #nixos
tilcreator has quit [Ping timeout: 260 seconds]
<samueldr>
I believe it should be possible
<samueldr>
grfn: this is running postgresql as not-a-system-daemon, locally, in a way similar to what you want, but not in a nix build https://github.com/NixOS/hydra/pull/759/files
<samueldr>
I guess the same approach to starting hydra yourself in the sandbox could work
<grfn>
samueldr: oh interesting!
<grfn>
this looks like it might work
orivej has quit [Ping timeout: 260 seconds]
john1 has quit [Ping timeout: 240 seconds]
orivej has joined #nixos
tilcreator has joined #nixos
MichaelRaskin has quit [Quit: MichaelRaskin]
afreak has quit [Ping timeout: 265 seconds]
cosimone_ has joined #nixos
john1 has joined #nixos
cosimone has quit [Ping timeout: 260 seconds]
cosimone_ is now known as cosimone
mbrgm_ has joined #nixos
kir0ul26 has quit [Ping timeout: 260 seconds]
mbrgm has quit [Ping timeout: 272 seconds]
mbrgm_ is now known as mbrgm
alp_ has quit [Ping timeout: 256 seconds]
zbozbo[m] has left #nixos ["User left"]
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #nixos
orivej has quit [Ping timeout: 246 seconds]
Rusty1 has joined #nixos
orivej has joined #nixos
knerten has quit [Quit: knerten]
bukkitgerman8608 has quit [Ping timeout: 256 seconds]