<pickfire>
How come nixos is modiying the shell scripts to use /nix/store such that one of them is pointed at the wrong file?
<pickfire>
I added two packages, lxde-common and lxsession.
<pickfire>
But the startlxde in lxde-common is pointing to itself when calling lxsession, but not to lxsession/bin/lxsession.
<pickfire>
Maybe the fixup phrase is wrongly done?
<nesevis>
Excuse my ignorance — let's stay I've created a declarative container in my configuration.nix that uses a specific package. This package should be run as a service with provided parameters when the container boots. How would I declare the execution of the service?
<samueldr>
BUT if I exec that shell script using sudo it works!
<samueldr>
I wouldn't think echoing 0 would work if echoing 1 hangs
<evanjs>
infinisil: but like -- that's the package I am building lol
<evanjs>
e.g. Hydrus fails to find Hydrus during Hydrus' checkPhase :P
<infinisil>
Ah
<infinisil>
Should probably also mention that it's about python
<evanjs>
oh good point lol
<infinisil>
But I think you do this with PYTHON_PATH usually
<lopsided98>
samueldr: It does that even when there is no BindsTo/WantedBy?
<infinisil>
PYTHONPATH*
<samueldr>
just now rebuilding to test without it lopsided98
<samueldr>
the pinephone isn't a powerhouse :)
nesevis has quit [Ping timeout: 245 seconds]
<samueldr>
coupled with bad wifi
<samueldr>
bad wifi generally here, and the pinephone having not the best wifi hardware
<samueldr>
okay no, works just fine without the after/bindsTo/wantedBy
<lopsided98>
Why did you want BindsTo in the first place?
<samueldr>
I'll reexplain the goal
<samueldr>
(will help situate everyone)
<infinisil>
evanjs: E.g. see development/tools/detect-secrets/default.nix
<samueldr>
the modem in the pinephone is controlled via a /sys path, this is equivalent to turning it on/off completely
<samueldr>
I want to make a more ergonomic interface, hopefully using systemd since systemd is right htere, where you can just `systemctl start $modem` and is starts, and `systemctl stop $modem` and it turns off
<samueldr>
when it turns on it's a usb device being connected, and well, disconnected when turned off
<pumpy>
nixos running on pinephone?
<samueldr>
though if I do a dumb oneshot service, sure it works, but the status of the service doesn't reflect the actual status of the modem!
<samueldr>
it could be turned on/off manually via the sys fs path, or even turned off by other means (IIRC)
<samueldr>
lopsided98: does that help understanding the problem?
<samueldr>
pumpy: well, it _is_ just an A64 board with more fancyness around :)
<lopsided98>
yeah, thanks
<samueldr>
lopsided98: so basically I want the status of the service to reflect the presence of the hardware
<pumpy>
samueldr ya i'd run it.
<pumpy>
what about the power mgmt of pinephone that article said was sloppy and might end up with explosions and shit?
<lopsided98>
samueldr: powering it off causes it to drop off the bus, right?
<samueldr>
lopsided98: exactly
<samueldr>
lopsided98: note that start/stop may take some time
<samueldr>
but that's like not really an issue, once it's showing up it's started, once it disappears, it's gone
<samueldr>
[about the pinephone] NixOS (and Mobile NixOS) isn't yet "useful" on it, but progressing, and mostly the usefulness will soon be about general "PE" (phone environment, like desktop environment) porting wokr
<samueldr>
work*
<samueldr>
though the power management and stuff article was seen, and I was already working on using megi's (author of said article) linux fork
<samueldr>
so it's not really a NixOS issue at that point
<pickfire>
Why is nix fixup not checking if the output binary exists?
<evanjs>
infinisil: ah, I needed $src instead of $PWD
<evanjs>
same syntax you linked tho
orivej has quit [Ping timeout: 246 seconds]
<lopsided98>
samueldr: Is is always powered on at boot? So you need the service to start itself to reflect that state?
<pumpy>
samueldr huh interesting. can't wait to try it out whenever it's ready
<samueldr>
lopsided98: it could be, not a guarantee
<pumpy>
would be nice to "provision" a system to a new pinephone
<samueldr>
lopsided98: and the more important bit to me is that it could be started/stopped outside of the purview of systemd
<pumpy>
right from a desktop that's nixos
<pickfire>
startlxde modify /usr/bin/lxsession to /nix/store/...lxde-common.../bin/lxsession in a shell script when building lxde-common but that does not exist, it exists in lxsession.
<pumpy>
maybe nixops could be expanded to not just datacenter/server, but general devices
<pumpy>
deviceOps
orivej has joined #nixos
<lordcirth>
pumpy, well, there are switches that run Debian
<samueldr>
lopsided98: well, that's basically the same answer I think, at any point in time it might have been started or stopped, and systemd would need to reflect that or I need to not-systemd
<pumpy>
like apple integrates everything, but not opaque and tightly coupled
justanotheruser has quit [Ping timeout: 240 seconds]
<samueldr>
pumpy: many end-users use nixops to manage their multiple machines
<lopsided98>
It really does seem like WantedBy and BindsTo should be the right solution. If the modem was started or stopped outside of systemd, then the ExecStart or ExecStop scripts should be no-ops.
mschwaig has joined #nixos
<samueldr>
lopsided98: I also had After, will try without
<samueldr>
oh
<samueldr>
if it's after the device unit shows up... it can't start it, thus hangs?
<samueldr>
(duh)
<evanjs>
infinisil: "# we run the tests on the Qt thread atm # keep a window alive the whole time so the app doesn't finish its mainloop"
<evanjs>
lol nvm
<evanjs>
it uses a window so.... ok nope :P
ryjm_matrix has quit [Quit: killed]
<lopsided98>
samueldr: I'm not sure I follow. If the dependent unit is already started it should just continue.
<samueldr>
lopsided98: the modem is off, the device unit is not present, so trying to start modem-control (the service that starts the modem) would hang until the modem is present, which it won't because the script never started
justanotheruser has joined #nixos
<lopsided98>
oh I see
<lopsided98>
Its not clear to me from the documentation that removing After will fix that, but we'll see
<samueldr>
well, it worked in that now it doesn't hang and starts the service, which starts the modem...
<samueldr>
... but the modem is stopping... which I'm not sure is relevant to systemd services stuff
halfbit has quit [Quit: WeeChat 2.9]
sss2 has quit [Quit: Leaving]
<pickfire>
Can I know where does nix build keep the failed files at?
jkachmar has quit [Quit: Connection closed for inactivity]
<samueldr>
lopsided98++ thanks it now works infinisil++ for earlier rubber ducking
<{^_^}>
lopsided98's karma got increased to 0xf
<{^_^}>
infinisil's karma got increased to 368
<pickfire>
Also, how do I know what variables are available at a certain point of time during the build?
<samueldr>
will clean and show soon
<samueldr>
both ways work; controlling outside systemd marks the modem-control correctly, controlling with systemd (obviously?) works
<pickfire>
I don't want all users to use my config which they don't know how.
<pickfire>
> 11:01 <pumpy> is ~/.config/nixpkgs/home.nix in exactly the same format that configuration.nix is? where you start with {config, pkgs, ...}: etc
<{^_^}>
error: syntax error, unexpected ':', expecting ')', at (string):345:3
<pumpy>
pickfire so when you use homemgr you use the home.nix in ~?
<pumpy>
why isn't home.nix same format as configuration.nix?
<evanjs>
pickfire: pumpy: yeah if you mean secrets then sure, but that's the reason using home-manager from nixos is helpful lol
<pickfire>
pumpy: ~/.config/nixpkgs/home.nix
<pumpy>
not even just secrets, but logical separation
<pickfire>
evanjs: Not secrets.
iAetosDios has quit [Remote host closed the connection]
<evanjs>
like, you can assign -- home-manager.users.programs.lsd.enable = true
<pickfire>
But no one in my house can use dvorak.
<pickfire>
And dwm.
<evanjs>
right and that's what home-manager helps _solve_
<pickfire>
If they log in to their account, they cannot use then.
<evanjs>
you'd assign that to your user lol
<evanjs>
doesn't matter if it's in the system config if you assign it through the nixos module for home-manager
<pickfire>
Yes, that's why I feel sad to hear "drop multi-user support".
<pumpy>
if i put my user stuff in ~/.config/nixpkgs/home.nix, do i have to add anything to configuration.nix to make nixos look in my ~ for the home manager home.nix?
<pickfire>
evanjs: But I don't think the users know how to edit the files.
<pickfire>
I mean most of them are normal users.
<evanjs>
pickfire: right, and the defaults would be the system level stuff
<pumpy>
pickfire home manager's dropping multi user support?
<pickfire>
pumpy: No, I saw in one of the talk saying to drop multi user support in nixos (just wishing).
<pumpy>
wtf
<pickfire>
But yeah, hopefully it never happened.
<pumpy>
seems like a weird idea
<pickfire>
Exactly.
<pickfire>
There is also someone else beside me that agrees.
<lordcirth>
Sounds like some people only use NixOS in server containers and don't care about other use cases
<pumpy>
i know that most servers don't have 1000 shell accounts on them anymore, but the multi user/system sharing model is obviously useful
<pickfire>
True.
<pumpy>
even phones and shit should have it, for session isolation
<pickfire>
My home laptops are usually multi-users.
<pumpy>
lordcirth hopefully those narcissists never control nixos's future
<pickfire>
Sometimes they just want to use a laptop but I cannot afford to let them use my user since all the keys are changed.
<pumpy>
ya mine too
<pickfire>
pumpy: But yeah, nixos is already pretty tweak in such a way.
<pumpy>
like i have user accounts that run automated stuff that has different sudoers rules than the user account i use
<pickfire>
Like xscreensaver built in?
<pickfire>
I don't know why is that.
<pumpy>
to catch fat fingers etc
<pumpy>
pickfire ya it's tweak but it doesn't REMOVE any fundamentally useful tech
<pickfire>
But it ADDS fundamentally unuseful tech.
<pumpy>
what does?
<pickfire>
Screens does not need to be save.
<pickfire>
It's just wasting my CPU cycles doing the weird rotations or matrix.
<pickfire>
Making my CPU hot, reducing its lifetime.
<pumpy>
ya no clue why a screensaver would be bundled with a bare nixos
<pickfire>
I wish I could throw away the screensaver.
<DigitalKiwi>
anyway unless you run off of a git checkout it takes a while for changes to get to the channel
amfl_ has joined #nixos
amfl has quit [Ping timeout: 258 seconds]
amfl_ is now known as amfl
sss2 has joined #nixos
Rusty1 has quit [Quit: WeeChat 2.3]
pjt_tmp has joined #nixos
slack1256 has quit [Remote host closed the connection]
cdepillabout has quit [Ping timeout: 265 seconds]
<dsal>
Is this a reasonable way to do what I'm trying to do here? nix-shell -p openscad --run 'open -W $buildInputs/Applications/OpenSCAD.app'
res0nat0r has quit [Remote host closed the connection]
<dsal>
or: How do I resolve the installed path for the thing I specifie dwith the -p there?
koz_ has joined #nixos
koz_ has left #nixos ["WeeChat 2.9"]
<pjt_tmp>
dsal: do you know that $buildInputs is available at the time? does the env command show it?
<dsal>
The command works. I just looked through the env to find something that at least made a connection. I feel like there's a more right way to do this.
<dsal>
The problem is that it has an executable, but not like, a unix executable. It's a mac executable which works differently and isn't placed on the path.
xd1le has quit [Ping timeout: 240 seconds]
<dsal>
Woah, I just found out it *does* do some magic.
<pjt_tmp>
and here I was just suggesting something because it _sounded_ right-ish. Huh.
blackriversoftwa has joined #nixos
<dsal>
I don't know nix well enough to understand how this magic is meant to work. What I really want is something like nix-shell -p openscad --run 'open -W ${pkgs.openscad}/Appliations/Openscad.app' -- but that's not valid.
<pjt_tmp>
dsal: yeah, that doesn't look right. Nix is meant to be as deterministic as possible and and */Applications/* sounds very mac-ish
<pjt_tmp>
I mean I guess you already know that if you're using one
justanotheruser has quit [Ping timeout: 240 seconds]
<pjt_tmp>
still, this is part of the ol' rubber duck method or whatever it's called
<dsal>
heh. nix-shell -p [somepackage] --run 'commandfrompackage' works because it puts the bin from the package in the path. .app is no less deterministic than any other artifact. It's just not obvious how to get to it.
cole-h has joined #nixos
icey has joined #nixos
icey is now known as Guest88671
<pjt_tmp>
dsal: Hm. could you just drop into a shell and poke around to find it? use silver searcher or some other bespoke tool?
<dsal>
I've already found it as mentioned above. I just want a less dumb way to launch an app like I do with CLI apps.
<typetetris>
Seems like I want a builtin. Is fetchGit or fetchurl "better"? I darkly remember , there have been issues with fetchGit cloning the whole depot needlessly? Was there something like that and is it still relevant?
<flip[m]>
I'm trying to move /var/log to a different partition. I set up a bind mount which works until i reboot. It seems initramfs only mounts root partition so the bind mount fails. Any ideas of the best way to solve this?
cantaloupe has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @pickfire opened pull request #100947 → gtk+: refactor name → https://git.io/JTWiT
mschwaig has quit [Quit: WeeChat 2.7.1]
<DigitalKiwi>
flip[m]: did you add the mount to your configuration?
<pingiun>
I am deploying a django package to my server via nixops. I built the derivation on my local machine, which has macos, and deploying on a linux machine. how is it possible that nixops is just copying my macos closure to my server? obviously the python package built on macos won't run on linux
phr0m3indh0v3n has joined #nixos
<pingiun>
and is there a way I could circumvent this?
fendor_ has quit [Remote host closed the connection]
<rnhmjoj>
pkgs.python3 should definitely be different because darwin and linux have different binary formats. how do you know it's being copied? did you see inside the /nix/store on the remote machine?
<pingiun>
I see "blazeit> copying closure..." in nixops output
<pingiun>
and when I try to run it, it complains about the exec format
<pingiun>
and when I check the format, it tells me it's a Mach binary
<pingiun>
oh wait copying closure is the system closure
<pingiun>
I'm not actually sure then
<pingiun>
but I'm wondering how a Mach binary ends up on my linux server
<pickfire>
Ah, nixpkgs-update is different from what I thought.
<pickfire>
I thought it automatically finds newer versions for all the packages.
<Orbstheorem>
Btw, has anyone tried using nix with binary bsps in embedded systems?
<Orbstheorem>
(samueldr maybe ^^")
<pickfire>
How come nix does not have gtk2 for vte?
<pickfire>
Is gtk2 being destroyed for nix?
<pickfire>
Or maybe there isn't gtk2 at all?
<rnhmjoj>
pingiun: i'm not a nixops user but i skimmed through the issues and noticed a few about deploying from darwin. it looks like you have to tell nixops the architecture of the remote
<pickfire>
Ah, there isn't anymore.
<pickfire>
No wonder.
<simpson>
pickfire: I'm not an expert, but it looks like even the "old" VTE 2.90 is built with GTK+3. What were you hoping to do?
<pingiun>
rnhmjoj: I do have nixpkgs.localSystem.system = "x86_64-linux";
<pickfire>
simpson: No, even arch does not have lxterminal-gtk3.
<pickfire>
I guess lxterminal only have gtk3.
<pickfire>
Since vte have gtk3, not gtk2.
cfinch has joined #nixos
<pickfire>
Just wondering, how is the nix packages usually being updated? Is it automated?
<simpson>
pickfire: Don't worry too much about Arch or other distros. We have a lot of packages, but the tradeoff is that we typically only have one or two versions of each package.
<pickfire>
simpson: But arch is also the same right?
<simpson>
A massive number of packages are automatically updated using a couple bots, but many packages are also updated by hand.
<pickfire>
Except there is archlinuxcn which provides a whole tons of packages.
<pingiun>
rnhmjoj: that issue is very helpful, thanks
<pickfire>
Does the bot covers all the packages or which part are not covered?
<pingiun>
because I am importing nixpkgs manually
<simpson>
Arch is extremely different from NixOS. (NixOS is extremely different from most other distros.) I don't know how Arch packages VTE, but it's probably not the same way that we package it.
jabster28_ has joined #nixos
jabster28 has quit [Ping timeout: 260 seconds]
<simpson>
It varies by language and ecosystem. For example, we import Stackage, a collection of Haskell packages, fully automatically; meanwhile, most of the Python tree is hand-maintained.
<rnhmjoj>
pingiun: ah, you have to set the `system` for the imported nixpkgs as well, right?
<pingiun>
rnhmjoj: yes, that seems to be my problem
vldn has quit [Ping timeout: 272 seconds]
<pingiun>
so now I'm figuring out how to string this value through so it keeps working on macos too
<pickfire>
simpson: How about rust packages or random packages like x11 stuff?
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
vldn has joined #nixos
zupo has quit [Ping timeout: 240 seconds]
<simpson>
pickfire: I think that we update Xorg's package collection all at once, semi-automatically. Rust is happening, but there's still a lot of churn due to compiler versions; I keep putting off learning more about it, and maybe I should stop doing that.
<gchristensen>
it is a bit annoying having irc bots join / part a lot during dev, which is sort of the norm for bots, but you can also get the same by handing FDs around during re-execs, and also through something like znc as a proxy
<gchristensen>
but again, the bot has several pieces of software serving different functions... maybe znc would still have worked fine... but it is also nice to be able to send messages with simple POSTs
<Raito_Bezarius>
is there instructions to redeploy the same setup?
<gchristensen>
no
<Raito_Bezarius>
is this something that I can easily reproduce using NixOS?
<Raito_Bezarius>
e.g. take a RabbitMQ broker then link it to something something with IRC
<gchristensen>
not even sure the current source is pushed anywhere :x ... that said, I'd like to hand it off and have {^_^} run by a group of people so it isn't just me
<Raito_Bezarius>
I heard today was a hackday :-°
<Raito_Bezarius>
If you could push any source somewhere or even a nix expr that you use for this, please ping me, I'm super interested in
<Raito_Bezarius>
As for running {^_^} — I'd be interested, but I don't know what it entail and as I'm not in the NixOS Foundation or org, I'm not sure if I can qualify
cole-h has quit [Ping timeout: 258 seconds]
davidv7 has joined #nixos
alp has joined #nixos
<gchristensen>
I've got baby-arms but let me see what I can do
<tobiasBora>
Hello, I'm trying to run a software that requires opencl in a steam-run environment, but it fails: clinfo works outside steam-run but fails inside. Any idea how to solve that issue? For poeple interested, I enabled hardware.opengl.extraPackages = with pkgs; [ intel-ocl ]; and I run steam-run using nix-build -E "with import <nixpkgs> {}; (steam.override {extraPkgs = pkgs: [ocl-icd gnome3.librsvg
<tobiasBora>
clinfo]; }).run", then I run steam-run bash, then clinfo.
<infinisil>
jophish: Note that noXLibs is just a manual blacklist
<infinisil>
So it doesn't prevent all Xlib references
<jophish>
Yeah, I get that :)
ManiacOfMadness has joined #nixos
<jophish>
I guess pycairo needs to do the same thing as Cairo does
vldn has quit [Read error: Connection reset by peer]
vldn has joined #nixos
<infinisil>
Oh yeah looks like it
<pumpy>
if i put my user stuff in ~/.config/nixpkgs/home.nix, do i have to add anything to configuration.nix to make nixos look in my ~ for the home manager home.nix?
ilmu has quit [Remote host closed the connection]
ilmu has joined #nixos
<infinisil>
pumpy: By default, NixOS doesn't interact with home-manager at all
knupfer has quit [Ping timeout: 256 seconds]
waleee-cl has joined #nixos
<pumpy>
oh
<pumpy>
so they run independently
<pumpy>
is there a command like nixos-rebuild like home-rebuild?
<{^_^}>
[rfcs] @7c6f434c opened pull request #78 → Initial version of RFC proposal → https://git.io/JTlgT
<__red__>
When doing a nixos-rebuild, it is ignoring my users.users.red.uid = 1000;
<a-h>
Hey folks, I'm trying to add plugins to tmux using `environment.systemPackages = [ pkgs.tmux pkgs.tmuxPlugins.open pkgs.tmuxPlugins.yank ]` but it makes no difference to the configuration at all, I'm still seeing standard tmux. I tried deleting my tmux configuration to see if that was overriding anything, but no change. There's a PR that suggests
<a-h>
that all that's required is to install the packages, as I've done, and it talks about adding a line to the tmux conf, but not where that would be done and what that would look like at https://github.com/NixOS/nixpkgs/pull/36790 - does anyone have any ideas or examples?
<{^_^}>
#36790 (by pjan, 2 years ago, merged): adds tmuxPlugins
cfinch has quit [Read error: No route to host]
cfinch has joined #nixos
<__red__>
I get: "warning: not applying UID change of user 'red' (1002 -> 1000)
<__red__>
"
<__red__>
But that is really what I want to do
<__red__>
I don't see a way to override that - anyone have any idea?
saschagrunert has joined #nixos
<__red__>
and looking at the source for update-users-groups.pl, I don't see an override
<__red__>
I wonder if anyone in the hackday is working on smashing the new release bugs
acarrico has joined #nixos
cantaloupe has joined #nixos
lsix has quit [Ping timeout: 260 seconds]
<{^_^}>
[nixos-homepage] @github-actions[bot] pushed commit from GitHub Actions to master « Update flake.lock and blogs.xml [ci skip] »: https://git.io/JTlaa
<tokudan>
does virtualbox work on 20.09? I always get the following error: Qt FATAL: Cannot mix incompatible Qt library (version 0x50c07) with this library (version 0x50e02)
jabster28 has joined #nixos
beww has joined #nixos
<clever>
tokudan: that happens if you have QT programs in nix-env, and you upgrade nixos
<clever>
tokudan: you must either remove (or upgrade) every QT based program in `nix-env -q`
<tokudan>
clever, nix-env -q is completly empty
<clever>
tokudan: what about root's `nix-env -q` ?
<tokudan>
same
<clever>
tokudan: are you pinning anything to old versions?
<beww>
/join #nixcon
<tokudan>
i only use a single nixpkgs, as far as i know, but i'll double-check
<clever>
tokudan: not sure then, ive got nixpkgs rev e2bb73ce5f7 on my laptop, from nixos-unstable, and its running virtualbox just fine
dbmikus has quit [Quit: WeeChat 2.7]
<bqv>
Had a strange error with fetchpatch in a build. It refused to finish, citing "sed: couldn't flush stdout". when I did the same fetchpatch in the repl, it worked, so now I can't reproduce it
nschoe has joined #nixos
<bqv>
Just noting it for any future log-crawlers
<clever>
bqv: you can run nix-store --delete on the output, then you can re-run it
<bqv>
I really, really don't want to, because the dependent build takes half an hour
ddellacosta has quit [Ping timeout: 260 seconds]
phr0m3indh0v3n has quit [Read error: Connection reset by peer]
phr0m3indh0v3n has joined #nixos
cfinch has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @jfroche opened pull request #101037 → python3Package.python-markdown-math: remove support for python 2.7 → https://git.io/JT8v7
<nix_installer>
Hi! we tried installing with the UEFI route and since it's an old motherboard we decided UEFI wasn't supported (it's not listed in the manual for the motherboard) so we're trying to go the legacy boot route now, but we're getting an error: "Creating journal (262144 blocks): doneWriting superblocks and filesystem accounting information: mkfs.ext4:
<nix_installer>
Input/output error while writing out and closing file system" - any thoughts on how to solve?
supercoven has quit [Ping timeout: 260 seconds]
<nix_installer>
I was thinking maybe we could wipe the hard drive and try again?
sarcasticadmin has quit [Ping timeout: 272 seconds]
sigmundv has quit [Read error: Connection reset by peer]
sigmundv has joined #nixos
noonien3 has joined #nixos
noonien has quit [Ping timeout: 260 seconds]
noonien3 is now known as noonien
<nix_installer>
we have run smartctl -a /dev/sda
<m3thos>
ensure the drive is formatted with old style partition table and see if you're using a block size compatible with your drive, if it is a recent spinning hdd, you'll need to use 4kb blocks
<m3thos>
anyone using zfs knows if nixos uses the cachefile?
<nix_installer>
m3thos: why can't we use ext4?
<nix_installer>
(didn't see anything obvious from smartctl, but I'm a noob - currently running `badblocks -v /dev/sda`)
<m3thos>
you can use ext4, just check if the partitioning of the drive is compatible with normal bios and if the formatting is compatible w/ the drive itself. I never seen that but it might be that you're trying to write to a 512b offset on 4kb block drive
<nix_installer>
could we try a different file system than ext4, or does the install instructions assume ext4?
<nix_installer>
if we try a different one, what command would we use?
<m3thos>
yes, try btrfs, it's great
<m3thos>
mkfs.btrfs
<m3thos>
or xfs: mkfs.xfs
<m3thos>
btrfs has lots of features like native compression, checksumming to detect bitrot, snapshots..
<m3thos>
xfs is more similar to ext4, it doesn't have those advanced features
ransom has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
spease has joined #nixos
<nix_installer>
m3thos: we did `mkfs.btrfs -L nixos /dev/sda1` which seemed to succeed aside from listing two incompatible features: `Incompat features: extref, skinny-metadata`
<nix_installer>
should we just continue from here, do you suppose?
<m3thos>
no problem w/ that, btrfs evolves by using optional features, it isnt a blocker or a big issue, carry on your instalation normally, everytime the docs mention `ext4` check here for what you should use, but `btrfs` is the name of the filesystem
MmeQuignon has quit [Quit: WeeChat 2.9]
<nix_installer>
ok, thanks
<evanjs>
catern: thanks for the review. going over the test weirdness now lol
hnOsmium0001 has quit [Quit: Connection closed for inactivity]
alp has quit [Ping timeout: 272 seconds]
beww has joined #nixos
m3thos has quit [Quit: WeeChat 2.7.1]
<evanjs>
wait I'm using nose, which I guess that's normal for :P
<nix_installer>
ok, we are now somewhat stumped at nixos-install, we have this error message: `error (ignored): aborting transaction: SQL logic error (in '/mnt/nix/var/nix/db/db.sqlite')error: committing transaction: disk I/O error (in '/mnt/nix/var/nix/db/db.sqlite')`