<ToxicFrog>
In configuration.nix, I can just do whatever = ./foo/bar; and once built that will refer to an in-store copy of ./foo/bar
<ToxicFrog>
How do I get a trailing slash on that?
<ToxicFrog>
Because if I do `whatever = ./foo/bar/;` the build fails complaining about the trailing slash, but for nginx, rsync, and some other things the presence or absence of a trailing slash is significant.
<infinisil>
ToxicFrog: Whether ./path imports the path into the store depends on how it's used
<infinisil>
By the module that reads the variable
<infinisil>
E.g. if it does `toString` on it, it turns into the path directly:
<infinisil>
> toString ./.
<{^_^}>
"/var/lib/nixbot/state/nixpkgs"
<infinisil>
While if it is interpolated, it's copied to the store:
<infinisil>
> "${./lib}"
<{^_^}>
access to path '/var/lib/nixbot/state/nixpkgs/lib' is forbidden in restricted mode
rajivr has joined #nixos
<infinisil>
Well, it would be anyways
<infinisil>
ToxicFrog: So what you can do is to explicitly copy it to the store by interpolating
<infinisil>
And then it's just a normal string
<ToxicFrog>
Aha
<ToxicFrog>
So I can do "${./foo/bar}/" to get the trailing slash
<ToxicFrog>
Thank you!
<{^_^}>
[nixpkgs] @mweinelt opened pull request #98244 → nixosTests.magnetico: wait for open port and make curl actually fail → https://git.io/JUEvT
<infinisil>
:)
codygman has quit [Read error: Connection reset by peer]
codygman has joined #nixos
<{^_^}>
[nixpkgs] @hongchangwu opened pull request #98246 → tmuxPlugins.nord: init at 0.3.0 → https://git.io/JUEvs
<lejonet>
Hmm, any reason why I'm getting "failed to create initrd secrets: No such file or directory" when I have no initrd secrets defined? doing a lib.mkForce {} doesn't help either
<lejonet>
(nixos-option boot.initrd.secrets also return {})
<lejonet>
it is seemingly blocking me from switching to a new generation
<lejonet>
I am just going to blame my RAM, I'm so sick and tired of it lol, I want to nuke it from orbit just to be sure that its BS doesn't haunt anyone else... I am gonna order new RAM next week
nullheroes has quit [Quit: WeeChat 2.9]
<lejonet>
because apparently file is just consistently segfaulting
<{^_^}>
[nixpkgs] @risicle pushed 2 commits to release-20.09: https://git.io/JUEf4
<matthewcroughan>
clever: oh.. dolphinEmuMaster seems to be disgustingly out of date :D
<matthewcroughan>
It's a lie! Damnit
<matthewcroughan>
"replacing old dolphin-emu-5.0-11824" "installing dolphin-emu-5.0-11824" noooooo
<matthewcroughan>
it's not the master, at all
<matthewcroughan>
That is months out of date in fact.
<clever>
matthewcroughan: nothing in nix can ever track master fully
<matthewcroughan>
so unlike the AUR which has `something-git` packages, nix cannot do this?
<matthewcroughan>
Why?
<clever>
matthewcroughan: because nix wants to know the hash of the source
<clever>
you would need to update the nix expr on every push, or use the import builtins.fetchGit
<clever>
dolphinEmu just uses a branch called "5.0" while dolphinEmuMaster uses 1b97f081b8eff9012132a4124537968bdb0e03e0
h0m1 has quit [Ping timeout: 244 seconds]
<clever>
`nix edit -f '<nixpkgs>' dolphinEmu` vs `nix edit -f '<nixpkgs>' dolphinEmuMaster` to see the difference
<matthewcroughan>
clever: Is there a way I can do this in a one-shot command?
<matthewcroughan>
like replace its hash
<clever>
matthewcroughan: you also have to update the rev as well
Fare has quit [Ping timeout: 244 seconds]
<matthewcroughan>
1. Go to GH and find hash 2. nix-env -iA nixpkgs.dolphinEmu --hash <updated_hash>
<matthewcroughan>
that would be my desired workflow
<matthewcroughan>
substitute or override the hash
<clever>
matthewcroughan: you need both the sha1 of a commit, and the sha256 of its contents
h0m1 has joined #nixos
<matthewcroughan>
how do I get the sha256 of contents?
<clever>
matthewcroughan: nix-prefetch-github
<clever>
,tofu matthewcroughan
<{^_^}>
matthewcroughan: 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. For inserting 52 0's in vim: <esc>52i0<esc>
<clever>
or that
<matthewcroughan>
I'm a bit confused
Fare has joined #nixos
<matthewcroughan>
I am just going to git clone, compile, and be done with it
<matthewcroughan>
then remove the directory later
<matthewcroughan>
nix-env is not useful in this context, I think
<clever>
matthewcroughan: you still want nix to compile it, `make install` is just wrong on nix
<matthewcroughan>
?
<matthewcroughan>
I simply want to run dolphin-emu, yet nix-env has complicated matters.
<clever>
this will fetch the default branch from github, and built it
<matthewcroughan>
So why doesn't the official nixpkg do this?
<clever>
because thats not pure, and youll get a different thing every time you do it
<simpson>
matthewcroughan: I strongly doubt that; I'm not an expert on Dolphin in particular, but emulators tend to be stable WRT their past history; if it builds, it should run. So "useless" is probably wrong.
<clever>
the whole point of nix, is to make things predictable
<simpson>
Indeed part of the power of Nix is that it should still build, for years to decades, as long as upstream tarballs are not changed.
<matthewcroughan>
simpson: nobody wants, in any case, a version older than the latest commit of Dolphin
<matthewcroughan>
The project is structured such that the latest commit *is* the desirable state.
<colemickens>
lol
<simpson>
matthewcroughan: That is a common sentiment/boast of many projects; nonetheless it is not part of our ethos. Instead, we value *reproducibility*.
boxscape has joined #nixos
<matthewcroughan>
I do appreciate that command clever
<matthewcroughan>
it is very good, and it proves anything can be overriden at the cost of what you're calling purity, which I haven't read enough about yet
<matthewcroughan>
If I want to play with my friends, I have to have the latest commit of dolphin simpson
<matthewcroughan>
The reason for this is the determinism of Dolphin's JIT
<clever>
matthewcroughan: if you want multi-player, i think its more important that you both run the same version
<matthewcroughan>
We all have to be running the same, latest version of Dolphin.
<clever>
matthewcroughan: and thats where reproducibility comes into advantage
<matthewcroughan>
latest, because my friend devs, so he's always on latest, forcing me to be on latest
<simpson>
matthewcroughan: Or, rather, the *indeterminism*. (I understand the latency constraints and the resulting preference to version netcode based on the JIT's output, but let's be clear about it.)
<clever>
matthewcroughan: so you can both run the same version, without the risk of one of you being 5 minutes newer
bvdw has joined #nixos
<matthewcroughan>
additionally, the dolphin website gives you the latest version, and encourages you to download the latest
<matthewcroughan>
within 1 week, everyone would or could have latest versions, so you just get the latest and play
<matthewcroughan>
We can do this any number of ways, but the point is that it's easiest and best to be on latest. Which the nixpkg itself can't do.
<simpson>
Just do it yourself. nixpkgs isn't magic; it's a git repository. I bump my local nixpkgs in order to pick up same-day releases of youtube-dl, for example. It's a common thing to do.
<matthewcroughan>
Such as an unofficial client for a service like youtube or discord.
meh` has quit [Ping timeout: 260 seconds]
<matthewcroughan>
you're saying this goes against Nix' ethos?
<simpson>
matthewcroughan: I've already used all my backchannels at YT and Discord to argue that they should open up the APIs. I tried, but they wouldn't budge.
<matthewcroughan>
So Nix doesn't support things that work or are architected like this? They are anti-nix?
<matthewcroughan>
Nix rubs against "latest" ?
<simpson>
Yes. Or, rather, "latest" rubs up against capability-security theory, and Nix treats packages as capabilities.
<matthewcroughan>
I'm going to have to get used to that, like I got used to bash syntax and pipes and stuff of that nature
<matthewcroughan>
but it is very different
<clever>
it gets much simpler if you use a config.nix or overlays
<simpson>
Why would you ever actually run that, though? It's how to do what you asked, but what you asked for is probably not what you wanted.
<matthewcroughan>
simpson: Why would I ever run it?
<simpson>
With emulators or tools like youtube-dl, I use nix-shell to summon the desired packages to a temporary environment, and direnv for environment management.
<matthewcroughan>
simpson: Well, I want Dolphin, as a one-off.
LnL has quit [Ping timeout: 260 seconds]
<matthewcroughan>
docker would be simpler in this context
<matthewcroughan>
for example, I run kubectl like this
<matthewcroughan>
alias kubectl='docker run --rm --name kubectl bitnami/kubectl:1.17'
<infinisil>
Or there's things like `niv` or `flakes` which can update rev and sha256's via a CLI, which you can then call automatically
<infinisil>
(rev/sha256's make Nix expressions pure)
<simpson>
matthewcroughan: We already have both; I'm currently on a NixOS laptop. There's no real problems here; everything you're struggling against is imagined barriers.
<simpson>
Well, no, I suppose that there's the real problem of `dolphinEmuMaster` being a poor name. But aside from that, I'm not sure what's wrong, exactly. The power here is that `nix-shell -p /nix/store/6px17dxc49zqp1qdm40xbjipm5aax706-dolphin-emu-5.0` isn't going to change which package is being referred to.
<matthewcroughan>
simpson: yes, with the downside of there being no nixic methodology for tracking master git repos officially without a verbose command the user must learn
<matthewcroughan>
`yay -S dolphin-emu-git` on Arch will dot his
<matthewcroughan>
do this*
<matthewcroughan>
I want -git style packages for nix, officially, I thought that's what dolphinEmuMaster was going to be
<matthewcroughan>
the details don't matter, it can be as separate from nixpkgs as it wants to be, I just want to be able to track master without replacing attributes
<matthewcroughan>
aka packages that have done it for me
<simpson>
matthewcroughan: That's not a downside to me. The main use of dolphinEmuMaster is likely that it can be run on a local git checkout of Dolphin itself, for *developing* Dolphin and having easy access to Nix's build environment.
<simpson>
As a package maintainer, I do *not* want users to be empowered to choose arbitrary versions from Last Thursday without some hassle; this sort of thing bogs down the issue trackers of Gentoo and Arch.
<energizer>
i'm on nixos-unstable, and nixpkgs.julia tries to build rather than downloading from cache. how can i diagnose?
<clever>
energizer: does the build pass or fail?
<{^_^}>
[nixpkgs] @mweinelt opened pull request #98255 → pythonPackages.sslyze: downgrade cryptography to 2.9 → https://git.io/JUEJC
* energizer
waits to find out
<clever>
energizer: its possible that the build fails, so hydra cant cache it
ydlr has quit [Read error: Connection reset by peer]
<matthewcroughan>
So I have to just git clone, and make it myself instead of using this convoluted package?
<clever>
you have to reproduce the cmakeFlags if you want to override it
patagonicus13 has joined #nixos
<{^_^}>
[nix-mode] @rossabaker closed pull request #108 → Check for liveness of process on nix-repl completions → https://git.io/JUCb9
knerten2 has joined #nixos
<clever>
2020-09-18 22:49:36 < clever> it gets much simpler if you use a config.nix or overlays
<matthewcroughan>
clever: Is that cmd going to be multithreaded?
<matthewcroughan>
No cache either btw?
<clever>
33 enableParallelBuilding = true;
<clever>
the expression already said to use multiple cores for building
<nbathum>
anyone have experience or thoughts deploying private packages?
<clever>
and it wont be cached, because your making changed hydra hasnt pre-built
<matthewcroughan>
clever: I mean cached from my last build I did with nix-env
<clever>
matthewcroughan: if you changed the cmakeFlags, then it has to rebuild the whole thing
jumper149 has quit [Quit: WeeChat 2.9]
patagonicus13 is now known as patagonicus1
patagonicus1 has quit [Ping timeout: 272 seconds]
knerten1 has quit [Ping timeout: 260 seconds]
<matthewcroughan>
clever: I really do appreciate you, a lot :d
<matthewcroughan>
:D
<matthewcroughan>
clever: What will that nix-shell command do exactly?
<matthewcroughan>
It will build and **run** the latest dolphin?
<matthewcroughan>
Or it will provide it to me in the current shell only on my path?
<clever>
it will build dolphin from master, and put it into the $PATH of a shell
ScottPuopolo has joined #nixos
ydlr has quit [Remote host closed the connection]
ydlr has joined #nixos
Pidgeotto has joined #nixos
Pidgeotto has quit [Excess Flood]
Pidgeotto has joined #nixos
Pidgeotto has quit [Excess Flood]
Pidgeotto has joined #nixos
Pidgeotto has quit [Excess Flood]
<matthewcroughan>
clever: Looks like dolphin can't get at any of my hardware devices
<matthewcroughan>
unsure why
<matthewcroughan>
running as root fixes, but then no audio for obvious reasons
<clever>
which device is it failing to access?
<matthewcroughan>
anything but the nipple on my thinkpad
<matthewcroughan>
[ 5173.423445] hid-generic 0003:057E:0337.0003: hiddev96,hidraw0: USB HID v1.10 Device [Nintendo WUP-028] on usb-0000:00:14.0-2/input0
<matthewcroughan>
permission denied
<clever>
you must be in the input group to open things in /dev/input/
spacefrogg has quit [Quit: Gone.]
aw has quit [Quit: Quitting.]
<energizer>
error: Package ‘qtwebkit-5.212.0-alpha4’ in /nix/store/6h4j4nh2z6ibl6cx5flnyqj00r0cb1ip-nixos-21.03pre243690.441a7da8080/nixos/pkgs/development/libraries/qt-5/qtModule.nix:54 is marked as broken, refusing to evaluate.
aw has joined #nixos
<energizer>
from nixos-rebuild. how do i find out why that package is trying to be built?
<clever>
energizer: --show-trace
spacefrogg has joined #nixos
<matthewcroughan>
clever: there is no input group
<matthewcroughan>
clever never mind it does exist
<matthewcroughan>
but sudo usermod -a -G input username DOESN@T WORK
<energizer>
+1 clever
<clever>
matthewcroughan: you must change groups with configuration.nix
ScottPuopolo has quit [Quit: Vision[0.10.3]: i've been blurred!]
<matthewcroughan>
why
<clever>
matthewcroughan: did you change something involving that?
<matthewcroughan>
My friends are just gonna have to play without me
<simpson>
I bet something else was changed. Maybe channels got updated.
<matthewcroughan>
Maybe it was
<matthewcroughan>
simpson: the channels did get updated
<matthewcroughan>
but I didn't say update
<ToxicFrog>
matthewcroughan: `nixos-rebuild` rebuilds every part of the system configuration that has changed. If you've been updating the channels, or if you have them configured to automatically update, that means it's going to rebuild all your updated packages.
<matthewcroughan>
I told you, I didn't say --upgrade
<simpson>
Okay, we found the root of the problem! Next time make sure your workstation's set up before committing to multiplayer gaming.
<matthewcroughan>
The channels updated by themselves, and started building
<clever>
matthewcroughan: one sec
<ToxicFrog>
--upgrade just says "automatically run nix-channel --update first"
<clever>
matthewcroughan: what does `echo $NIX_PATH` say?
<matthewcroughan>
it completed, I don't care
<matthewcroughan>
but my user is NOT in the input group
nckx has joined #nixos
sputny has quit [Quit: sputny]
sputny1 is now known as sputny
<clever>
matthewcroughan: you have to logout and login for any group changes to take effect, same as every other linux distro
sputny1 has joined #nixos
<matthewcroughan>
which means I need to rebuild dolphin clever ?
<matthewcroughan>
I need to rebuild dolphin right? because that was done in a nix-shell
<clever>
matthewcroughan: if master has changed, yes
<matthewcroughan>
master hasn't changed
<clever>
matthewcroughan: if master hasnt changed, it will reuse the old build
<simpson>
matthewcroughan: You'll have to build up a sense of when you've changed configuration.nix or when you've updated channels. I'm confident, though, that you requested the change which resulted in the mass rebuild.
<matthewcroughan>
gr8
<matthewcroughan>
clever: I still have insufficient perms
<matthewcroughan>
again, this works fine in sudo, but not as a user
<matthewcroughan>
I am a member of the input group, yet dolphin cannot access my hardware outside of sudo
<clever>
matthewcroughan: what device is it trying to open? strace -f -e open,openat dolphin-emu
<matthewcroughan>
I gave you the dmesg of the device
<clever>
but dmesg doesnt say what dolphin tried to actually open
<matthewcroughan>
[ 6214.866550] hid-generic 0003:057E:0337.0004: hiddev96,hidraw0: USB HID v1.10 Device [Nintendo WUP-028] on usb-0000:00:14.0-2/input0
<matthewcroughan>
[ 6214.868926] usb 1-2: usbfs: process 7172 (GC Adapter Scan) did not claim interface 0 before use
<clever>
but dmesg doesnt say what dolphin tried to actually open
aleph- has quit [Quit: WeeChat info:version]
<clever>
strace -f -e open,openat dolphin-emu
<energizer>
i just rebuilt on nixos-unstable, and it disconnected my wifi. nmcli d wifi c mywifi ... says Connection mywifi is not availalbe on device wlan0 because device is strictly unmanaged
waleee-cl has quit [Quit: Connection closed for inactivity]
<matthewcroughan>
clever: /dev/bus/usb/001/008
ericsagnes has joined #nixos
<clever>
matthewcroughan: ls -l /dev/bus/usb/001/008
<clever>
internally, dolphin bakes DOLPHIN_WC_REVISION into its binary, and then shares that with the other multi-player members
<clever>
and if that string doesnt match, it assumes your not compatible
ydlr has joined #nixos
<clever>
so you must set that to the revision your actually building from
<kini>
It sounds like it's probably a good idea to make a derivation that's parametrized by a git commit hash, and plug that into the expression wherever appropriate
<clever>
kini: yeah, that would make the override much simpler then what i gave above
<matthewcroughan>
damnit, I have to rebuild the whole thing again now? clever
<matthewcroughan>
no caching?
<clever>
matthewcroughan: correct
<matthewcroughan>
I value reproducibility too, but this is all very very very stupid
<clever>
matthewcroughan: nix only has derivation level caching
<matthewcroughan>
have you ever used yocto?
<matthewcroughan>
Yocto's SSTATE cache blows this silly process that you've just given me out of the water :D
<matthewcroughan>
Not that I would want to use yocto for other reasons, but the lack of caching here is really annoying me
<clever>
matthewcroughan: nix's purity is based on hashing every input to a build, and redoing the build if any input has changed
<clever>
matthewcroughan: having something like ccache involved would poison it all
<clever>
matthewcroughan: if the inputs to a derivation havent changed, nix can reuse the entire output, as a cached build
<clever>
but your changing the cmakeFlags input
<matthewcroughan>
What is the benefit of this purity?
<simpson>
(Languages with proper modules *could* integrate on an incremental per-module basis with Nix; I've done it out-of-tree for Monte. But the effort-to-payoff ratio doesn't seem attractive to folks.)
<matthewcroughan>
How is yocto *not* deterministic?
<matthewcroughan>
I do not understand how yocto is "poisoning" its cache
<clever>
matthewcroughan: it solves the nasty problem of things like `docker build` can just grab whatever it wants over the network, and you get a different version every time you build
<matthewcroughan>
Yocto and Nix are comparable, they essentially do mostly the same thing, yet one has caching that works like this, without the downsides you're proposing
<clever>
matthewcroughan: nix disables the network, and hashes every single input, so if any input changes, nix redoes the build
<{^_^}>
[nixpkgs] @AndersonTorres pushed 2 commits to release-20.09: https://git.io/JUETG
<kini>
matthewcroughan said the friend he plays with is a dolphin dev and insists on using master, so I guess it's a kind of special case
<clever>
drakonis: the developers encourage you to constantly update to latest, and the emulator itself enforces that everybody in multiplayer has a matching gitrev
<drakonis>
i see
boxscape has quit [Quit: Connection closed]
<kini>
Maybe the right approach for something like this is to have an overlay repo somewhere which has an always up to date version of `dolphin-emu/master.nix`? Someone would have to set up a cron job somewhere to update it, though.
kenran has joined #nixos
Miyu-saki has joined #nixos
Miyu-saki is now known as Guest82457
<kini>
clever: why is `-DDOLPHIN_WC_DESCRIBE=really-master` necessary in your expression? Could you get away with not overriding `cmakeFlags` at all? That might make the expression a little easier to understand.
Guest82457 has quit [Changing host]
Guest82457 has joined #nixos
Guest82457 is now known as Miyu-saki
<matthewcroughan>
clever: When I restart my laptop, nix-shell and nix-env stuff is going to be available still
<clever>
kini: because the original expression is using cmakeFlags to set the version based on the original git rev
<matthewcroughan>
how do I make it so my entire laptop is *ONLY* what is in configuration.conf
<clever>
kini: and the override can only update that if it changes all of the cmakeFlags
<clever>
matthewcroughan: nix-shell based stuff wont persist across a reboot
<matthewcroughan>
alright, well nix-env definitely does
<clever>
matthewcroughan: then use nix-env -e to uninstall things, and `nix-env -q` to list them
<Miyu-saki>
I didn't actually think about using ssh-copy-id lmao
<clever>
matthewcroughan: that will install dolphin from master, but every time you nixos-rebuild, it will check if master has changed, and may cause a surprise rebuild if master has changed
<Miyu-saki>
Yeah, I think that's a much more robust solution than extracting the NixOps key. Thanks.
<clever>
Miyu-saki: depending on which backend your using, the nixops key itself may also be in /root/.ssh/authorized_keys
<matthewcroughan>
clever: page not found
<Miyu-saki>
I'm using DO, because it's straight up the most accessible one
<matthewcroughan>
We're all on thinkpads here huh? :D
<Miyu-saki>
I'm stupid, I pasted that on a scratch buffer so that I can delete the IP, then did everything except delete the IP.
zebrag has quit [Ping timeout: 264 seconds]
<Miyu-saki>
Welps, /shrug
<clever>
Miyu-saki: ssh -v, to see what pubkeys it tried
<Miyu-saki>
Oh cool, didn't know you can do that.
<Miyu-saki>
Interesting, if I use `nixops ssh -d deployment machine -v` it doesn't mention "Trying private key"
<Miyu-saki>
Oh
rprije has joined #nixos
<Miyu-saki>
> debug1: Will attempt key: /tmp/nixops-tmpEuDKQc/id_nixops-midori
<{^_^}>
error: syntax error, unexpected ':', expecting ')', at (string):323:25
<clever>
Miyu-saki: you can maybe also steal the privkey from that path, while the ssh session is open
mir100 has quit [Ping timeout: 272 seconds]
<{^_^}>
[nixpkgs] @ryantm opened pull request #98257 → nixos/activation: fix type for activationScripts and userActivationScripts → https://git.io/JUETr
<colemickens>
I have a few repos now that basically just ver bump + lock for use with flakes. I should maybe genericize it (for a certain emulator, for examle)
<colemickens>
yes, I said genericize instead of generalize. it's been a day.
<Miyu-saki>
clever: Ohhh, nice idea. Thanks.
stiell has quit [Ping timeout: 265 seconds]
heavengin7 has joined #nixos
<clever>
Miyu-saki: oh, and nixops export dumps the full state including privkey as json
<clever>
and just a bit of tab-completing in `nix repl '<nixpkgs>'` and i find this
<hart111>
Ahhhh that makes sense
<hart111>
I didn't realize that you could get nixpkgs in repl
<hart111>
That's actually really nice to know!
<hart111>
Thank you :)
<clever>
yep
<hart111>
TIL that repl is not just for nix pills and actually has a purpose :)
<clever>
i use it to both experiment with new exprs, and to verify a nix file is doing the right thing
<clever>
and also to just search packages
<hart111>
Yeah all of the above are things I could definitely use it for
<hart111>
I think I'm going to write up a little bit in the doc about this after I get it working, it is mentioned but the actual process to do it is not written down
domogled has joined #nixos
mmohammadi98126 has joined #nixos
sputny1 has joined #nixos
sputny1 has quit [Remote host closed the connection]
sputny has quit [Ping timeout: 260 seconds]
<{^_^}>
[nixpkgs] @ju1m opened pull request #98262 → Fix gitolite updating of gitweb access list and description → https://git.io/JUEIc
palo1 has joined #nixos
<{^_^}>
[nixos-homepage] @samueldr pushed 4 commits to feature/2020-redesign: https://git.io/JUEIC
<{^_^}>
[nixpkgs] @asdf8dfafjk opened pull request #98263 → nixos/onedrive: Remove verbose flag → https://git.io/JUEI1
sangoma has joined #nixos
<cole-h>
LnL: OK, I think I know where the `to_string` is needed, but I think I need to impl `to_string` on the enum itself, so I can change the `defaultValue` to be `defaultValue.to_string()`
pf has quit [Ping timeout: 256 seconds]
pf has joined #nixos
noudle has quit []
<LnL>
cole-h: yeah it's the the toJSON path, it's probably nicer to define that for enum settings and have to/from string use it
<LnL>
SandboxMode has the same issue btw
<hyper_ch>
clever: you use nfs to make shares available in a vm? If so, do you use the /etc/exports file or do you zfs nfs export?
<clever>
hyper_ch: nfs.server = { enable = true; exports = ''.....''; }; is what ive been using
<clever>
hyper_ch: but the clients are just more nixos machines on the LAN, not VM's
<hyper_ch>
clever: so not use zfs exports
<clever>
hyper_ch: correct, i'm not sure if `zfs nfs export` even works on linux, it might be a solaris-only thing
<hyper_ch>
well, I have a debian mail server in a vm... becuase of the control panel software... and I thought I could just put the mails into an own dataset and then mount it from the host
<cole-h>
LnL: Argh. Looks like I need to do more investigations lol
<clever>
hyper_ch: you can also use a zvol to give the VM its own dedicated disk as a dataset
<hyper_ch>
to my knowledge it should work... enable nfs server and then just set the zfs export dataset .... I'll try
<hyper_ch>
clever: it has own dataset and I used zvol.....
<hyper_ch>
however I just want to have a large number of snapshots for the mail
<clever>
hyper_ch: ahh, a zfs filesystem is better for that
<hyper_ch>
bascially I want to have at least 1 year worth of snapshots every hour
<hyper_ch>
and after 9 months, it grew to over 1 TB usage
<hyper_ch>
clever: yes.... it's a bit more mail than I thought though
la-jesystani has joined #nixos
<clever>
hyper_ch: can you pastebin this? zfs list -t volume -o name,used,referenced,logicalused,logicalreferenced,written,usedbysnapshots,usedbydataset,refcompressratio,compressratio,compression,mountpoint,sync
<la-jesystani>
what do i need to read in order to package something from github? i want to use a fork of picom thats not in pkgs
<clever>
la-jesystani: that should be in the nixpkgs manual
<clever>
hyper_ch: yeah, you definitely want some trim, your only using 67gig, but zfs is keeping the unused blocks, so its using a total of 374gig for the current version
<hyper_ch>
as I said, on the debian side, things don't change often... so I don't need a large number of snapshots there.... hence seperating mails from the rest of the OS
mmohammadi981261 has joined #nixos
mmohammadi98126 has quit [Ping timeout: 272 seconds]
mmohammadi981261 is now known as mmohammadi98126
<clever>
hyper_ch: and its keeping that extra 307gig in every snapshot
<LnL>
cole-h: guess that means you got it working? :)
<cole-h>
LnL: Literally just that macro and it works good.
<clever>
hyper_ch: what is within tankHB/encZFS/VMs/Mail? mbr? ext(2|3|4) ?
<hyper_ch>
ext4
<cole-h>
Now to revert the stringify stuff (and also make a todo to PR the same macro for SandboxMode)
<clever>
hyper_ch: so if you run `blkid /dev/tankHB/encZFS/VMs/Mail` it reports ext4 directly?
<hyper_ch>
yeah
<clever>
hyper_ch: umount it within the vm (or shutdown), then mount it anywhere on the host, and run fstrim against it
<hyper_ch>
will do so... thanks
<clever>
hyper_ch: that will reduce your usage by 307gig immediately, and make any more snapshots you create a bit more optimal
<hyper_ch>
but I will still put the data on the host and mount it in the vm
<clever>
hyper_ch: but you would still need to repeat that dance regularly, or setup trim forwarding in the vm, or use nfs
<hyper_ch>
even if I remove old snapshots?
kleisli_ has quit [Quit: Leaving]
Pidgeotto has joined #nixos
Pidgeotto has quit [Excess Flood]
<clever>
hyper_ch: basically, when you delete a file in the guest, zfs doesnt know those blocks are unused
<hyper_ch>
ok
<clever>
hyper_ch: so the snapshots are holding onto every single deleted file as well
orivej has joined #nixos
Pidgeotto has joined #nixos
Pidgeotto has quit [Excess Flood]
<clever>
hyper_ch: nfs would let zfs know exactly what is happening, so the problem just vanishes
<clever>
hyper_ch: the manual trim tells zfs to forget about the unused blocks, and trim forwarding in the vm would automate it
Pidgeotto has joined #nixos
Pidgeotto has quit [Excess Flood]
<la-jesystani>
i have a derivation im just not sure how to get a shell that has it installed, or build it. i havent found that in the manual yet, is there a quick and easy one liner to do that?
<hyper_ch>
clever: thx
<clever>
la-jesystani: nix-build can build it
bourbon has quit [Quit: authenticating]
bourbon has joined #nixos
<la-jesystani>
i tried that its telling me it cant auto-call a function that has an argument without a default value (stdenv)
<cole-h>
LnL: Some day I'll be a functional polyglot like you -- pro in Rust and C++ (and probably more) :D
sangoma has joined #nixos
<hyper_ch>
clever: I can just run fstrim on the zvol?
<LnL>
cole-h: I don't really know c++, just know the nix codebase a bit
<cole-h>
Fair enough.
<clever>
hyper_ch: the zvol has to be mounted only on the host for you to fstrim it
<cole-h>
LnL: I think the only outstanding issue is the race between registering all the loggers causing bar and bar-with-logs to not be registered (at least at the point the manpage is generated...)
<clever>
hyper_ch: if you can configure the VM to forward trim from guest->host, then you can run fstrim inside the guest, or just mount it with auto-trim
<hyper_ch>
thanks
kleisli has joined #nixos
kleisli is now known as dillon
dillon is now known as endorphin
konobi has quit [Remote host closed the connection]
Darkmatter66_ has quit [Ping timeout: 260 seconds]
bahamas has quit [Ping timeout: 258 seconds]
<cole-h>
regnat: Going to bed, but: do you know *why* progress-bar.{hh,cc} are in libmain? I just tested moving it to libutil (along with names.{hh,cc}) and it seems to work (and also lets the loggers be registered properly for manpage generation)
<cole-h>
(Please ,tell me, so I can see your answer when I wake up... :P)
cole-h has quit [Quit: Goodbye]
cantstanya has quit [Remote host closed the connection]
mmohammadi98126 has quit [Ping timeout: 256 seconds]
mmohammadi98126 has quit [Quit: I quit (╯°□°）╯︵ ┻━┻]
malook has quit [Remote host closed the connection]
dermetfan has joined #nixos
tnias has quit [Quit: leaving]
dermetfan has quit [Remote host closed the connection]
cosimone has quit [Quit: Quit.]
dermetfan has joined #nixos
<sxiii>
Hi everybody :) Could someone hint me with my current quest. I am creating a Live ISO, and I have Openbox working there with default config (I can't find rc.xml file for it). How could I create it and apply so openbox eats it?
<sxiii>
I know it should be either /etc/openbox/rc.xml or /etc/xdg/openbox/rc.xml probably, but I can't manage to ask the "configuration.nix" config to create this file for me at this path
thc202 has joined #nixos
m1cr0m4n has joined #nixos
<{^_^}>
[nixpkgs] @raboof opened pull request #98275 → openjdk: add jdk11 jre and headless jre → https://git.io/JUEZy
<sterni>
sxiii: so you'd be using environment.etc."openbox/rc.xml" = "…";
<sterni>
openbox is a bit of a untypical window manager, usually you can configure a window manager in nixos using some attributes in the same place as the enable option…
<MichaelRaskin>
More like «only minimally supported»
fendor has quit [Remote host closed the connection]
<sxiii>
yep guys from another channel had given me nix config with firefox kiosk autostart. So now I am trying to do openbox with another program (sent) without any window borders
amir has quit [Quit: Have to disappear. Read you later guys!]
orivej has quit [Ping timeout: 246 seconds]
abathur has quit [Quit: abathur]
astro has quit [Read error: Connection reset by peer]
astro has joined #nixos
<infinisil>
Not really, but it shouldn't be any different from stdenv usage
balodja has joined #nixos
Fare has quit [Ping timeout: 244 seconds]
<mrSpec>
Hi guys! I'm preparing for my first nixos upgrade (to 20.03) and I've notice one inconsistency (not sure If I should be worried). nix-channel --list shows nixos [htt....]/nixos-19.09 however in configuration.nix system.stateVersion = "19.03". Is this a problem?
<adam_>
stdenv = clangStdenv;
<adam_>
mkShell {
<adam_>
I wrote that as my stdenv
<adam_>
and I have clang_11 in my build dependencies
<adam_>
now when I am building I use -DCMAKE_CXX_COMPILER=clang++
<adam_>
home/adam/research/adam/third_party/libtorch/include/pybind11/pybind11.h:1000:9: error: no matching function for call to 'operator delete' ::operator delete(p, s, std::align_val_t(a));
<adam_>
but when I build with g++ everything is good
<symphorien[m]>
mrSpec nope
<infinisil>
adam_: Don't use mkShell, because that uses stdenv underneath
<infinisil>
adam_: Use clangStdenv.mkDerivation instead
<sterni>
sxiii: you don't necessarily need homemanager, no
<sterni>
but then also you won't get the benefit of reproducibility
<sterni>
sxiii: can you check if the file /etc/openbox/rc.xml is created properly?
<adam_>
Okay I used that instead and then gave it a name as well
<adam_>
still same error
<balodja>
Hi all. Guys, I'm new to Nix, help me please a bit in diving. At the moment I'm trying to fix some deficiencies in the Gnucash package. And I'm stuck reproducing the build.
<adam_>
/nix/store/rclksjxdjgp6y6qkxyl9m4dx4b9d45zk-gcc-9.3.0/include/c++/9.3.0/new:154:6: note: candidate function not viable: no known conversion from 'pybind11::size_t' (aka 'unsigned long') to 'std::align_val_t' for 2nd argument
<adam_>
it still seems to be using gcc
<symphorien[m]>
yes linux's stdenv use libstdc++
<adam_>
yeah but I put llvmPackages_11.stdenv.mkDerivation
<adam_>
I assumed it would use clang
<adam_>
and libc++ rather than libstdc++
<symphorien[m]>
yes but this is a clang configured to use libstdc++
<adam_>
what would you recommend doing instead then?
<sterni>
sxiii: the openbox package distributes its own etc/openbox/rc.xml which is somewhere in /nix/store, it could be that it prefers that one over the one in /etc
<sterni>
sxiii: is there an command line option to tell openbox which config file to read?
<adam_>
CMake Error at CMakeLists.txt:4 (message): libc++abi now requires being built in a monorepo layout with libcxx available
<adam_>
dang
<adam_>
I can't use llvmPackages_11.libcxxStdenv
<symphorien[m]>
it does not build ?
mvnetbiz_ has quit [Quit: Bye!]
<adam_>
nope
<adam_>
like when I type nix-shell it failed
philr has quit [Ping timeout: 264 seconds]
<symphorien[m]>
I guess you should open an issue
<adam_>
with 10 it seemed to like it
mvnetbiz_ has joined #nixos
<adam_>
okay I'm getting a new set of errors!
<adam_>
../third_party/libtorch/include/pybind11/cast.h:305:19: error: use of overloaded operator '!=' is ambiguous (with operand types 'pybind11::detail::values_and_holders::iterator' and 'pybind11::detail::values_and_holders::iterator')
<adam_>
while (it != endit && it->type != find_type) ++it;
<adam_>
~~ ^ ~~~~~
mariatsji has quit [Remote host closed the connection]
mariatsji has joined #nixos
<adam_>
okay if I use c++17 instead of c++20 I can get past that error
<adam_>
but now I have a new one
<adam_>
libtorch/lib/libtorch_cpu.so: undefined reference to `std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >::npos@GLIBCXX_3.4.21'
<adam_>
why does gcc not have any of these problems
<adam_>
is nix not meant for it?
mariatsji has quit [Ping timeout: 246 seconds]
meh` has joined #nixos
<etu>
hexa-: You've asked me to update the miniflux pr to the even newer version released, I've pushed that now :)
hoxtonhopper has quit [Ping timeout: 272 seconds]
<hexa->
etu: cool, I'll check that the test run and the we merge that. It has been sitting long enough!"
<etu>
Indeed, I had forgotten about it :D
<hexa->
it's waiting for the darwin tests … rip :D
ManiacOfMadness` has quit [Ping timeout: 246 seconds]
adam_ has quit [Remote host closed the connection]
domogled has joined #nixos
Fare has joined #nixos
gv has quit [Remote host closed the connection]
gv has joined #nixos
<symphorien[m]>
> libtorch/lib/libtorch_cpu.so: undefined reference to `std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >::npos@GLIBCXX_3.4.21'
<symphorien[m]>
you can't use a libtorch compiled with libstdc++ with a clang using libc++
<{^_^}>
error: syntax error, unexpected ':', expecting ')', at (string):323:29
<symphorien[m]>
> is nix not meant for it?
<symphorien[m]>
yes it is, it is the default on darwin :)
<{^_^}>
error: syntax error, unexpected ')', expecting ID or OR_KW or DOLLAR_CURLY or '"', at (string):324:1
tim123123 has joined #nixos
domogled has quit [Quit: domogled]
<tim123123>
Hello People, is anybody using their mobile phone as webcam replacement? I found an app called DroidCam but I would need to make this application work in NixOS in order to use it https://github.com/aramg/droidcam/tree/master/linux are there already existing approaches? Many Thanks!
<sxiii>
sterni: you think I should try "openbox --config-file /etc/openbox/rc.xml" ?
sangoma has joined #nixos
andymandias has joined #nixos
bahamas has joined #nixos
ManiacOfMadness` has joined #nixos
tomberek has joined #nixos
polarfire has quit [Quit: WeeChat 2.7.1]
<{^_^}>
[nixpkgs] @worldofpeace pushed 0 commits to gnome-3.38: https://git.io/JUERz
polarfire has joined #nixos
<sterni>
sxiii: yeah or you could use writeText similarly to writeScript and then do openbox --config-file ${configFile} (and define configfile using let)
ddellacosta has joined #nixos
shibboleth has joined #nixos
stree has quit [Quit: Caught exception]
stree has joined #nixos
ManiacOfMadness` has quit [Ping timeout: 258 seconds]
<sxiii>
Seems like openbox is blaming my rc.xml file on being wrong. I'll try to recheck
orivej has joined #nixos
<sxiii>
After I tried openbox --restart --config-file /etc/openbox/rc.xml :) It throw error. Should have debug the openbox config first ;)
cole-h has joined #nixos
<{^_^}>
[nixpkgs] @worldofpeace merged pull request #98260 → [20.09] Revert "nixos/monit: Allow splitting the config in multiple files" → https://git.io/JUEkd
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to release-20.09: https://git.io/JUERp
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to release-20.09: https://git.io/JUERh
<cole-h>
regnat: I think it might be because, at one time, it depended on store-api.hh (and thus had to be in libmain or libstore), but it doesn't anymore.
<cole-h>
I'll ask Eelco in -dev :P
<cole-h>
(Maybe all these questions should have gone there in the first place... lol)
andymandias has quit [Ping timeout: 246 seconds]
Thra11 has joined #nixos
noudle has joined #nixos
fendor has quit [Ping timeout: 272 seconds]
<Uma>
"bluetooth hci0: Direct firmware load for rtl_bt/rtl8761a_config.bin" does anyone know where I can find these drivers?
werner291 has quit [Quit: werner291]
ransom has quit [Ping timeout: 244 seconds]
komik_mi has joined #nixos
<ToxicFrog>
Argh, I'm getting "store path contains illegal character <newline>" again and I don't remember how I fixed it the last time
<ToxicFrog>
I did not think that three lines were long enough to be worth pastebinning.
<cole-h>
I thought you were going to paste more 😅
dansho has quit [Quit: Leaving]
bk1603 has joined #nixos
<ToxicFrog>
I've now rolled back the configuration.nix to a known good one (passes `nixos-rebuild` on the host system) except that the disk configuration is different and I'm still getting the same issue, so I don't think it's a configuration.nix issue
<ToxicFrog>
And I vaguely remember having this issue once before but I don't remember how I fixed it, and I can't find any discussion of it in the logs here
sangoma has quit [Quit: WeeChat 2.9]
<bk1603>
Hey everyone! This is my first time on NixOS and I've been trying to set dwm up on it. But I've been running into a few issues. First one being that the colors that I set in `config.def.h` aren't actually there in the dwm binary, I've tried rebuilding and it doesn't work. Secondly a lot of patches don't work with overlays but they work when I use it separately on the dwm source
sangoma has joined #nixos
adelbertc has joined #nixos
<bk1603>
Also since I was doing a lot of rebuilding and rebooting, I found that sometimes ifconfig doesn't report my wireless interface, almost as if the drivers weren't loaded. I haven't yet dived deeper into this issue since I was so busy setting dwm up
<bk1603>
It's on sme reboots that the wireless card doesn't work, but on some reboots it does
<bk1603>
I'm using unstable 21.03
ManiacOfMadness` has joined #nixos
shibboleth has quit [Quit: shibboleth]
<bk1603>
Also setting `fonts.fonts = with pkgs; [ powerline-fonts ]` results in an error complaining about the variable `powerline-fonts` not being defined
cosimone has quit [Quit: Quit.]
<ToxicFrog>
Aha! The solution was to run nixos-install with --no-channel-copy, for some reason. I should probably figure out how to repro and file a bug.
arjen-jonathan has quit [Ping timeout: 244 seconds]
<ToxicFrog>
For the overlays vs source patch issue, make sure it's the same version? If the dwm package definition hasn't been updated in a while you might be patching a more recent version than nix is building.
meh` has quit [Ping timeout: 256 seconds]
<bk1603>
I forked 6.2 and have been using that for src. Patches work on the local repo fine, but when I try and add them to the overlay they report chunks failing. :(
codygman has quit [Read error: Connection reset by peer]
<bbigras>
Should I use the "PRs ready for review _" forum post for backports?
codygman has joined #nixos
<bk1603>
Also about the fonts issue, at first I thought moving it to `environment.SystemPackages` since it was under `nixos.powerline-fonts` but then I read a documentation saying that's not how it's done
<bk1603>
Installing it with `nix-env` works thugh
<bk1603>
I just wanted to add it to the configuration.nix file but that isn't working somehow :/
<bk1603>
ToxicFrog: (I forgot tagging you in the previous messages :P)
<cole-h>
bbigras: No, ping the RM team
<ToxicFrog>
Are you sure you don't just have a typo or something? Because that is how it's done and that line looks like it should work
<ToxicFrog>
Re the patches -- ok, you have a local copy of 6.2, your overlay points at that as src? Have you manually applied the patches to that directory? If so it won't be able to re-apply them
<bk1603>
Nono, I have a local repo and a fork on github. I use the github repo as src. I have no patches applied on the github repo
karantan has joined #nixos
<bk1603>
I tried applying patches on the local repo when they failed on the github repo, expecting to see the error. But then it just ran fine
<cole-h>
Probably means your repo has diverged.
<bk1603>
Also I can't change the colors by changing the value at `config.def.h` if that's related somehow. I was wondering if using `oa.patches ++ [ list of patches]` causes nix to include the patches that were applied earlier and theo nes defined in the list
<bk1603>
cole-h: by diverged do you mean that the repo is no longer the same as the original dwm one? Couldn't understand you there
<cole-h>
Well, if the patches apply locally but not to your fetchFromGitHub'd repo, it means there's some difference. That's all I mean by diverged
zakame has quit [Ping timeout: 256 seconds]
<ToxicFrog>
Yes, that's exactly what it does -- that's list concatenation
<ToxicFrog>
So if `oa.patches` is the list of patches present in the original (non-overlaid) build configuration, what you're doing there is adding all of your patches to be applied after the othres.
rajivr has quit [Quit: Connection closed for inactivity]
<bk1603>
Aha then maybe I understand why this happens. I tried adding and removing patches, so somewhere down the line I got to a point where I was trying to add a patch that was already there in my local install. Does it make sense?
sangoma has quit [Ping timeout: 240 seconds]
<ToxicFrog>
(assuming I'm understanding what you're doing in your overlay correctly without context -- maybe you should pastebin the whole thing)
<{^_^}>
[nix] @cole-h opened pull request #4041 → Serialize SandboxMode enum to string for JSON → https://git.io/JUEEk
cjpbirkbeck has quit [Ping timeout: 260 seconds]
<bk1603>
cole-h: I don't think so, because `git diff` shows nothing is different :/ It might be the list concatenation thing maybe. Not sure.
<cole-h>
bk1603: If you `git am`'d or othewise committed the patch, it won't show up in the diff.
andymandias has joined #nixos
Thra11 has quit [Ping timeout: 256 seconds]
<bk1603>
Umm, I don't think I commited any patches upstream, but I'll give it a look anyways. My memory often fails me :P
<ToxicFrog>
bk1603: yeah, that's going to apply all the patches that are configured in the nixpkgs definition + your own. *That said*, there are currently no patches in the nixpkgs definition, so it should just be your own.
D_ has joined #nixos
<bk1603>
ToxicFrog: Oh, that would mean that the patches mentioned in the list only are applied right? Wierd, I'll give my repo a look then, maybe it's me after all :P
<bk1603>
Also apart from that, I still don't get the colors issue. I changed the color values in `config.def.h` but nothing changed in the binary after I rebuilt.
tim123123 has quit [Remote host closed the connection]
sangoma has joined #nixos
<ToxicFrog>
You changed the colours in e348957bd7d289f5125f9669df631ca4f7ea1132 ?
<ToxicFrog>
Your configuration.nix still points at 7be69c06f61a2452fa271e498e7a00d003375bc0
lunik1 has quit [Quit: :x]
<bk1603>
Oh I changed the colors in the commit right after the inital commit but even they aren't coming up
<ToxicFrog>
When doing this kind of experimentation, it is often very useful to work with a local checkout of the source, and just do: `src = /path/to/my/local/checkout;` in configuration.nix
<ToxicFrog>
This means you're always getting the source as it exists on disk and you don't need to worry about updating the rev or the expected hash
<ToxicFrog>
Then, once you're done experimenting, you commit and publish your changes and point src at that
<bk1603>
I have the default color config, even though I changed to some colors I copied from Nord :/
ThaEwat has quit [Remote host closed the connection]
<ToxicFrog>
might want to try using that mechanism rather than patching config.def.h by hand, and/or disable that mechanism in the overlay to make sure it's not interfering.
zakame has joined #nixos
lunik1 has joined #nixos
<bk1603>
Oh okay I'll take a look into that then
Fare has joined #nixos
bk1603 has quit [Quit: Lost terminal]
silver has joined #nixos
<sxiii>
sterni: I think I need to ditch openbox and replace with some WM that has no windows borders by default and which will allow me to run apps on fullscreen by default ideally with 0 configuration
joko has quit [Quit: WeeChat 2.7.1]
Anonymyous11 has joined #nixos
<hart111>
i3 time :)
knupfer1 has joined #nixos
<sxiii>
Doesn't i3 includes some kind of panel by default? As I want just fullscreen...
tomberek has quit [Remote host closed the connection]
<sxiii>
With no panels no decorations anything...
<hart111>
I mean it has a bar at the top
malook has quit [Quit: malook]
<hart111>
But it's 1 line to disable it
<hart111>
And 1 more line to make all things fullscreen
<sxiii>
hart111: do you have an example config.nix file with i3?
<hart111>
I'm a bit cramped for time atm but it's available there
knupfer1 is now known as knupfer
<hart111>
And the two lines would be: in modules/home/i3.nix make the bars list empty to disable the bar, the other stuff is set in `window = { titlebar = false; border = 0; hideEdgeBorders = "none"; }`
<nbathum>
it seems like the other option (for this use-case of deploying private packages), is to serve the output tarball locally, and write a drv for it using fetchurl, and then rely on using caches/substitution during deployment so that systems in the field don't try to reach your private source server
<sterni>
Kloenk: on your system or in nixpkgs as a whole?
<bk1603>
ToxicFrog: Thanks a lot for your advice, dwm finally has the colors. And the patches I think were in conflict with each other, reordering them caused the other patch to break. And it wasn't that necessary either.
<kloenk>
sterni: in nixpkgs in whole. But I think I found my problem. I think I have a recursion because of utillinux
mallox has joined #nixos
karantan has joined #nixos
<bk1603>
One final question, I wrote a script for the status monitor, and it basically runs forever, which is why I think I can't use it `displayManager.sessionCommands`. How can I execute such a script everytime the window manager starts from `configuration.nix`?
<sterni>
Kloenk: then there's no easy way definitely, nix after all is a programming language and not a data format after all
<bk1603>
If I try `displayManager.sessionCommands = '' ./.xsetroot &'';` (with the new lines) dwm can't open up and I just return to lightdm.
<kloenk>
bk1603: '' try ~/.xsetroot & '';. not sure, but it could help
<cole-h>
Even better would be `${./.xsetroot} &`
<kloenk>
cole-h++
<{^_^}>
cole-h's karma got increased to 92
werner292 has joined #nixos
<bk1603>
Would there be a way of not using the `.xsetroot` file at all? And simply writing out the lines in the configuration.nix file itself? Would `${while \ script stuff \ done} &` work?
<kloenk>
how does callPackage work? when my package has the attribute `utillinux` and I overrwite it, does every other dependencie also gets overwritten?
gv has quit [Remote host closed the connection]
<sterni>
Kloenk: callPackage gives a package all its inputs from its current context, so no if you override one package to use a different utillinux not all packages it depends on will also use it
<sterni>
Kloenk: but you can do that by overriding the entirity of pkgs or probably better using an overlay
<kloenk>
sterni: yes, I know, thanks :-). I'm just still confused that systemd builds, but my dbus does not build, because of a recursion of lvm with utillinux (as far as I can tell)
bk1603 has quit [Remote host closed the connection]
<kloenk>
also confused that dbus seems to depend on dbus?
<sterni>
as far as I've read bootstrapping everything systemd is a bit of a nightmare
<mrSpec>
I added 20.03 stable channel, to upgrade from 19.03→20.03, then I called: nixos-rebuild switch --upgrade however it failed while building derivation: https://pastebin.com/56tbmVVc Could someone tell me where to look at and what should I fix?
<{^_^}>
[nixpkgs] @doronbehar pushed 2 commits to release-20.09: https://git.io/JUEa3
cjpbirkb1 has quit [Ping timeout: 258 seconds]
lordcirth has joined #nixos
lordcirth has quit [Remote host closed the connection]
lordcirth has joined #nixos
<{^_^}>
[nixpkgs] @samuela opened pull request #98291 → ###### Motivation for this change Upgrades vscodium from 1.48.2 to 1.49.1 → https://git.io/JUEaR
lordcirth has quit [Remote host closed the connection]
lordcirth has joined #nixos
<{^_^}>
[nixpkgs] @samuela closed pull request #98291 → ###### Motivation for this change Upgrades vscodium from 1.48.2 to 1.49.1 → https://git.io/JUEaR
<balodja>
Hello again guys. I'm still fighting with the Gnucash package. So help me kindly to understand how to debug an issue.
<balodja>
During the installation phase Gnucash copies some documentation files to $out/share/doc, meanwhile it should copy them to $out/share/doc/gnucash.
<balodja>
The code seems correct to me, but it doesn't work as intended.
orivej has quit [Ping timeout: 272 seconds]
<balodja>
And I don't know how to debug it. Because you know nix begins to rebuild nearly the whole world when I add a few print statements to that script.
<kloenk>
bbigras: what should play well with nix-community/impermanence? I missed the context
<bbigras>
Kloenk: systemd-home. That's the new thing using 1 file or something for home I think.
<bbigras>
systemd-homed*
balodja has left #nixos [#nixos]
nek0 has joined #nixos
<kloenk>
bbigras: oh yes. also curios about that. I not even understood how to use systemd-homed. (but I got it to compile xP)
<bbigras>
Kloenk: I have no idea how it works. it seems it can lock /home on suspend when encryption is used but I have no idea if it makes sense with full disk encryption.
leah2 has joined #nixos
<kloenk>
I think you also somehow can have your home on an usb stick, and just log out, take your usb stick to another pc, log in again, and continue where you left
<{^_^}>
[nixpkgs] @jonringer opened pull request #98305 → python3Packages.lightparam: fix source and deps → https://git.io/JUEKJ
<colecf>
Hi, I'm trying to use a swapfile, and adding it to my hardware-configuration.nix work fine (it shows up in swapon -s), but nixos-generate-config always erases it. Is there a way to get nixos-generate-config to recognize it?
<clever>
colecf: add it to configuration.nix instead
<colecf>
ah, ok. Wasn't sure if that was a valid thing to do, thanks
<clever>
colecf: nixos-generate-config will scan things like what is currently mounted, and then just blow away the whole hardware-configuration.nix and replace it
<clever>
colecf: so if you want changes in hardware-configuration.nix to stick, you have to stop running nixos-generate-config
<clever>
or just edit configuration.nix instead, since nixos merges them together for you
<colecf>
Good to know. I thought it would start detecting it and writing it into the file like how it detects newly-added harddrives
Fare has joined #nixos
adam_ has joined #nixos
mbrgm_ has joined #nixos
mbrgm has quit [Ping timeout: 244 seconds]
mbrgm_ is now known as mbrgm
eoli3n_ has quit [Ping timeout: 256 seconds]
werner292 has quit [Quit: werner292]
gentauro has quit [Read error: Connection reset by peer]
gentauro has joined #nixos
philr has joined #nixos
justanotheruser has joined #nixos
<{^_^}>
[nixpkgs] @marsam opened pull request #98306 → libavif: init at 0.8.1 → https://git.io/JUEKN
<{^_^}>
[nixpkgs] @mkg20001 opened pull request #98307 → libavif: init at 0.8.1 → https://git.io/JUEKA