<Henson>
infinisil: shouldn't the python27Packages list contain the LTS version of scikit-image, so it can be installed, instead of a version that's incompatible?
<Henson>
for those poor folks who want to, or have to, keep using python27?
bvdw has quit [Read error: Connection reset by peer]
<MichaelRaskin>
Just wait until upstream Python2 downloads disappear
<infinisil>
We have our own caches of upstream downloads in most cases, so that hopefully won't be a problem :)
<Henson>
infinisil: until the downloads to build the packages from source disappear from the web. I tried building a very old version of a package once (I can't remember which one) and the original downloads were no longer available.
<simpson>
infinisil: I love your optimism; you're clearly not a Python user~
<infinisil>
Hm yeah that is a problem
pie_ has joined #nixos
mounty has joined #nixos
<infinisil>
Python 2 source is cached on cache.nixos.org anyways though
<infinisil>
(verified with `nix-build '<nixpkgs>' -A python2.src`)
<{^_^}>
[nixpkgs] @sikmir opened pull request #76627 → gpxpy: init at 1.3.5 → https://git.io/JeNwt
<infinisil>
aanderse: Hm, can't you use `alias` to point it to the store path?
<aanderse>
in httpd you have directory directives and location directives
<aanderse>
sometimes you'll have a directory directive which you don't explicitly expose
<infinisil>
Ah, the difference between them would've been my other question
<aanderse>
ie `<Directory /var/www>`
<aanderse>
no alias there
<aanderse>
you don't always need an alias
<aanderse>
directory directives point to actual file system paths, location directives point to web url paths
<infinisil>
Hm I see
<aanderse>
you should (almost) never do permissions in a location directive
<infinisil>
Okay idea: Don't use the key to set the path, but an attribute instead
<infinisil>
Like directories.*.path = ...
<zeta_0>
for some reason jupyter is not able to detect the haskell kernel, i install ihaskell separately in home.nix, but for some reason nix is looking for the ihaskell binary in ghcWithHoogle instead?
<aanderse>
that is a pretty simple directory directive
<aanderse>
i want to append to it, or overwrite it
mbrgm has quit [Ping timeout: 258 seconds]
mbrgm_ is now known as mbrgm
drakonis has quit [Ping timeout: 268 seconds]
<infinisil>
Yeah that would be nice
<betaboon>
argh. i thought `dm-tool lock` would actually lock the user session. it does not ...
<zeta_0>
as you can see ghcWithHoogle and ihaskell are installed separately, so i don't know how to explicitly tell nix to search elsewhere for the ihaskell binary? https://hastebin.com/yipuzefehi.py
<infinisil>
The context-discarding approach doesn't seem too terrible I guess
<infinisil>
It's a valid use of that builtin imo
<aanderse>
infinisil: yeah, ok
alex`` has quit [Quit: WeeChat 2.7]
alexherbo2 has quit [Remote host closed the connection]
<aanderse>
thanks
<aanderse>
i was trying to avoid it
<aanderse>
but if not avoidable... wanted to see how "acceptable" it would be considered
<DigitalKiwi>
infinisil: yay it's fixed (well, allegedly, i still haven't gotten to test iet yet)
<infinisil>
Hmm..
<aanderse>
most (all) modules that access a pkgs.someWebApp are going to reference pkgs.someWebApp somewhere else to make it not "dangle"
<infinisil>
aanderse: I'm worrying about it being changed to something else at some place but not at others
<infinisil>
DigitalKiwi: Cool :)
<aanderse>
infinisil: random made up example for context?
<infinisil>
But no override at the unsafeDicardContext
<infinisil>
aanderse: How about this: Symlink /run/foo -> /nix/store/the-web-app, then use "/run/foo" as the key
<aanderse>
infinisil: ah i see, yeah ok
<infinisil>
It's not really nice either, but at least it's safe and allows for easy merging
<infinisil>
Also means you won't have to use the same pkgs.someWebApp.override { .. }; in all modules that potentially want to assign to it
<DigitalKiwi>
error: cannot connect to daemon at '/nix/var/nix/daemon-socket/socket': Connection refused
<infinisil>
(which can change in only one of them too)
<DigitalKiwi>
is that normal
<infinisil>
DigitalKiwi: I'd search through IRC logs, there's a bunch of mentions of it. Maybe it's a new issue though, in which case I'd open a new issue in the nix repo
<aanderse>
infinisil: are you suggesting i take care of that in the httpd module, or let people take care of that in their own module when they use httpd module?
<infinisil>
Ah hm, good question..
<infinisil>
Would probably have to be done by the people
<aanderse>
sounds good to me in that case
<aanderse>
also
<DigitalKiwi>
infinisil: lol i see a ton of people have asked but i haven't found one that had an answer :|
<aanderse>
using the `directories` option is entirely opt in... so not imposing
zeta_0 has quit [Quit: rcirc on GNU Emacs 26.3]
<infinisil>
aanderse: What do you mean by opt-in? Do you have to enable something to be able to use it?
Rusty1 has quit [Remote host closed the connection]
<infinisil>
DigitalKiwi: Hehe that's unfortunate, so maybe a new issue in the nix repo would be appropriate (or is there one already?)
mounty has quit [Ping timeout: 265 seconds]
<aanderse>
i mean module authors don't have to use it, they can continue to manually write `<Directory` in `extraConfig` if they think it is too dangerous
<aanderse>
are you familiar with `system.extraDependencies` at all?
mtjmullen has joined #nixos
<infinisil>
Not familiar but seen it
<aanderse>
if it operates as expected... how hack is it to use that instead of symlinking to a directory?
<infinisil>
Sounds like a good way to make unsafeDiscardContext safe again if you can ensure the same derivation is used
<infinisil>
But that can be hard to do, I'd personally use the symlinking in my config
<infinisil>
Well not hard, but you never know what happens
horner has quit [Quit: zzZzZzZzZZZzzz…]
<infinisil>
Though in the worst case your httpd will just fail to start
<aanderse>
i think with `httpd` the problem of replacing package isn't new...
<infinisil>
Which I guess could be pretty bad depending on what it's for
<aanderse>
people can still do crazy things to screw things up
<infinisil>
Yeah
<infinisil>
unsafeDiscardContext should probably be avoided in nixpkgs itself
<aanderse>
i know some of the web application services modify the package in the module to add plugins and stuff. that was a mistake (by me). the plugins should be included in the package, not the module
<aanderse>
infinisil: well... ideally i'd like to get this upstream for the `zabbixWeb` module
<infinisil>
Hm, I'd go for the symlinking thing
<infinisil>
Seems like a cleaner solution overall
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JeNrA
<infinisil>
Probably
<infinisil>
Hm, though doing that change just because nix doesn't support string context in attribute keys is a bit weird
<aanderse>
i assume string context not supported in attribute keys because performance?
<infinisil>
I have no idea, just wondering about that too
<guill>
Has anyone tried compiling with libc++ instead of libstdc++? Based on the existence of the libcxx/libcxxabi packages I assume it's supposed to be supported, but 'features.h' is missing. Is there some other arcane package I need to import to get that?
<supermathieu>
Howdy, I am on nixos and I use an overlay to modify a package. The main purpose of this overlay is to use a different version of the package.I am trying to understand how it works. So as of now it seems to work because when I use nixos-rebuild switch. Everything goes through without warnings or errors. However, my application still return the same version number. I was wondering how can I debug this. I'd like to confirm that at least the right
<supermathieu>
git tag is used
<tomberek>
where is the overlay? how are you integrating it into your configuration
<Dandellion>
is there some way to pass feature flags to `rustPlatform.buildRustPackage`
<supermathieu>
tomberek: In my configuration.nix :nixpkgs.overlays = [ (import ./nixpkgs_overlays.nix) ];
<supermathieu>
and it seems to at least get evaluated because, it was complaining about some typos earlier
<tomberek>
okay, that seems reasonable. did you run a "which" to see if you have a nix-env installed versions shadowing it
<supermathieu>
Hmm, that's odd. $ which /run/current-system/sw/bin/watson
<isd>
Hey all. Does nix have anything equivalent to guix's `pack` command? i.e. I have a profile with some program I want to run, and I'd like to just tar up the transitive closure so I can extract it on another box and run it.
<tomberek>
there are a few ways to obtain the same thing, of various maturity
Soo_Slow has joined #nixos
<guill>
I normally wouldn't think this was Nix-related, but it doesn't happen on any other dist I've tried... I have an application that's hitting a segmentation fault very early on -- within __libc_start_main. I've tracked it down to an instruction where it's essentially doing '*environ = *environ'. Any ideas why this might be happening in NixOS?
<tomberek>
I've used nix-bundle and liked both the arx and AppImage versions. You can also get the equivalent to `pack` with a tarball fairly easily. The Docker version would take a bit more effort with dockerTools in nixpkgs
<tomberek>
the squashfs version looks nice to. similar to the AppImage in nix-bundle
shajra has joined #nixos
<tomberek>
isd: So it's pretty easy to do if you dig into some Nix expressions. But there isn't a blessed official CLI tool.
<infinisil>
And for completeness, if you *do* have nix on the other machine, You can create a single file containing a dependency closure with `nix-store --export $(nix-store -qR paths) > out` (see `man nix-store` for more info)
<infinisil>
And import it with `nix-store --import out`
<tomberek>
infinisil: to contiue our last discussion about deployment. I've spent some time on exporting various formats of my nix projects, tar, rpm, deb, docker, nix-bundle, nix-rewrite in .zip, etc. nix-store --format <SOMETHING> --export PATH where <SOMETHING> could be tar,squashfs,docker,nar in a manner similar to `pack` would be awesome
fragamus has quit [Remote host closed the connection]
<mog>
i have a callpackage, and i want to use it as a dependency for another callpackage package how do i do that?
<isd>
My resaon for asking is actually that I was researching packaging sandstorm apps using nix. so was looking for building blocks to get at the bits of a filesystem image that I need. Thanks for the pointers.
bhipple has quit [Remote host closed the connection]
mexisme has joined #nixos
gyroninja__ has quit [Read error: Connection reset by peer]
gyroninja___ has joined #nixos
jbit has joined #nixos
<jbit>
Does anybody know why NixOS has stuck with glibc 2.27? (I just spent some time trying to figure out why directories sort differently on ubuntu and nixos, and it seems to be because ubuntu is using a newer glibc which has a newer iso14651_t1 definintion)
<MichaelRaskin>
Indeed it is Juniper kernel that is failing
<cinimod>
Adding buildInputs = [ pkgs.which ] gives me a different error - lol
o1lo01ol1o has joined #nixos
<cinimod>
I needed buildInputs = [ pkgs.which pkgs.R pkgs.libintl ];
<cinimod>
MichaelRaskin: thanks for your help
patagonicus has joined #nixos
patagonicus is now known as Guest38788
<Guest38788>
Hey all. I'm trying to set up a tagged VLAN in addition to the untagged default on a network interface. I've found the networking.vlans options, but by themselves they don't seem to do much. Is there any docs on how to set up VLANs that are no for containers/vms?
<{^_^}>
[nixpkgs] @McSinyx opened pull request #76635 → alure2: init at unstable-2019-09-21 → https://git.io/JeNMb
<{^_^}>
[nixpkgs] @romildo opened pull request #76636 → [WIP] warsaw: init at 1.13.1.2 → https://git.io/JeNMj
kamadorueda has quit [Remote host closed the connection]
luc65r has joined #nixos
<luc65r>
Hi!
<luc65r>
Can you guys help me to set a lightdm theme ?
luc65r has quit [Remote host closed the connection]
o1lo01ol1o has quit [Remote host closed the connection]
<Guest38788>
Is networking.vlans not doing what I want (seemingly nothing instead of setting up the vlan dev) something I should ask about on Discours? Or is the Github bug tracker the right place for it?
o1lo01ol1o has joined #nixos
<{^_^}>
[nixpkgs] @offlinehacker opened pull request #76637 → trash-cli: fix listing files over mount points → https://git.io/JeNDs
<{^_^}>
[nixpkgs] @offlinehacker pushed 3 commits to master: https://git.io/JeNDc
ambroisie has joined #nixos
Chiliparrot has joined #nixos
cosimone has quit [Quit: Terminated!]
dingenskirchen has quit [Quit: dingenskirchen]
dingenskirchen has joined #nixos
liif is now known as liff
<jD91mZM2>
Hi! Today I decided to ditch all the LTS stuff and use nixos-unstable+latest kernel, mainly because of my new computer's AMD Radeon RX 5700 graphics card which isn't supported by the LTS server as far as I understand it. But I recalled that there might be some trouble with the latest kernel and ZFS? It works currently, but is there anything I should beware of, or do I just need to downgrade whenever
<jD91mZM2>
it refuses to boot up?
akegalj has joined #nixos
akegalj has quit [Remote host closed the connection]
syd has joined #nixos
<syd>
Hi peeps! I have two external harddrives and I'd like to keep them in sync via something raid1 like
<syd>
is there a ready-made solution for that?
mexisme has joined #nixos
equivrel has joined #nixos
chloekek has joined #nixos
<simpson>
If you are feeling bold, it ought to be possible to put external disks under md, and then use software RAID.
<syd>
and if I'm feeling conservative? :)
<syd>
I'm fine with one being a master and the other being a slave. I'm not doing it for read performance
<simpson>
Well, software RAID needs assembly at boot, and you want to wait for all of your drives to be present before assembling. I don't know if we have anything that helps with that specific niche need. You may have to instead have a script which you run to explicitly assemble the array after you've connected your disks.
mexisme has quit [Ping timeout: 260 seconds]
<syd>
I could have a script that syncs the two directories. I'd be ok with that.
<syd>
Is there such a script?
<simpson>
And of course I wouldn't blame you if mdadm simply is scary. I'm a professional and I still need the man pages to write mdadm commands. Triple-check everything before running it, cross the fingers, etc.
<syd>
rsync would work, except it leaves deleted files
<syd>
aha
<syd>
`rsync --delete` :P
<simpson>
There are flags for rsync to fix that. But also, NixOS runs a Linux kernel, so mdadm is a real possibility.
<syd>
I'm a bit scared of that sort of stuff
<syd>
I'll get backto it later
bvdw has quit [Read error: Connection reset by peer]
bvdw has joined #nixos
UndefinedIsNotAF has joined #nixos
<syd>
and stick to rsync for now
<__monty__>
syd: git-annex might help keep track of which things are and are not on those disks.
<DigitalKiwi>
syd: i use zfs on external disks for that
corsair has joined #nixos
<DigitalKiwi>
jD91mZM2: i haven't had problems yet
<DigitalKiwi>
zfs native encryption for over a year now heh
chloekek has quit [Ping timeout: 258 seconds]
<DigitalKiwi>
which means i've been on nixos full time for over a year now ahhhh
<lirzhv>
looks relevant, will watch the vid thanks
<simpson>
lirzhv: Guix started as a fork of Nix. That's about it; there's not that much cross-pollenation, AFAICT. There's a healthy dialogue, and folks are friendly about it.
<jd823592>
how is it resolved within a nix env when two packages are installed at once but provide the same binary? say I have multiple channels and install possibly diffently named packages that however provide the binary `ghc`, what determines which of them will be picked up when i type ghc into the terminal
Chiliparrot has joined #nixos
<jd823592>
and how does it translate to packages being installed via environment.systemPackages in configuration.nix
<clever>
jd823592: nix-env has higher priority then systemPackages, and thats based just on the order of them in $PATH
<clever>
jd823592: nix-env will refuse to install duplicates by default, and just reject the entire package
<infinisil>
clever: jd823592: To be more exact, it's based on meta.priority of a package. It only gives a conflict if both have the same priority, but if that's not the case, the binary from the higher priority (meaning lower number) gets chone
<infinisil>
s/chone/chosen
<clever>
ah, i tend to just never allow conflicts
<infinisil>
You can use lib.lowPrio and lib.hiPrio to change priorities ad-hoc to some fixed lower/higher values
<{^_^}>
[nixpkgs] @veprbl pushed to master « unityhub: switch name to "unityhub" »: https://git.io/JeNS5
misterwhatever has quit [Remote host closed the connection]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
ng0_ is now known as ng0
mexisme has joined #nixos
zupo has joined #nixos
<noonien>
hello folks
<noonien>
can i get rid of .nix-profile and .nix-defexpr from my $HOME?
<noonien>
preferably move them somewhere XDG-compliant, if possible
<noonien>
.nix-profile is a symlink to a non-existant directory anyway
mexisme has quit [Ping timeout: 260 seconds]
<noonien>
path*
jgeerds has joined #nixos
LLED2_ has joined #nixos
jakobrs has quit [Remote host closed the connection]
<clever>
noonien: .nix-defexpr must exist for nix-env to find your channels, and .nix-profile must exist if your installing anything with `nix-env -i`
cosimone has joined #nixos
<noonien>
i don't use nix-env at all, but still those files seems to have been created. is it possible to move them to somwhere else? perhaps by settings some environment variables?
<clever>
noonien: one of the nix files in bashrc is creating them
<clever>
cant remember which one though
<noonien>
hmm, ok, i'm not using bash, and don't have any profilerc file apart from zshrc, i'm guessing zsh also loads a systam global shellrc file from somewhere
<clever>
yeah, likely /etc/zshrc
<clever>
pam may also be involved
<noonien>
i'm trying to clean up my $HOME, and there's so many dotfiles that seem to appear randomly, trying to get rid of even a few of them makes me feel like sisyphus
<{^_^}>
[nixpkgs] @tobim opened pull request #76644 → nixUnstable, nixFlakes: update to current → https://git.io/JeN9B
Synthetica has joined #nixos
<Synthetica>
I'm getting frequent freezes under Sway that aren't kernel panics (the capslock light isn't blinking and when I press the power-off button, it still powers off normally). However, I cannot switch to different ttys with ctrl-alt-fN... Anyone encounter this before?
gustavderdrache has joined #nixos
gxt has quit [Ping timeout: 240 seconds]
<chagra>
Synthetica: yeah I actually had that exact same issue quite a few times from september to begiining of this month when I used sway. But I did use sway from the nixpkgs-wayland overlay and they were quite rare. I'm no longer on sway so I don't know
<chagra>
Synthetica: but as you said nothing works but pressing the power off button shutsdown normally with no issues
<Synthetica>
chagra: Yeah, exactly!
jmeredith has quit [Quit: Connection closed for inactivity]
<Synthetica>
chagra: But no fix?
<joko>
Hey, is it possible to escape dots in Nix variable names?
<{^_^}>
[nixpkgs] @risicle opened pull request #76645 → aflplusplus: init at 2.59c → https://git.io/JeN9V
vidbina has joined #nixos
<immae>
joko: I don’t know if you can do that in variable names, but maybe you meant attrs keys?
<infinisil>
joko: Technically yes, but practically no
<infinisil>
> let "." = 10; in "foo"
<{^_^}>
"foo"
<infinisil>
It works, but you can't ever use the value
<immae>
infinisil: I had the same idea, but how do you use the variable after that? :D
<immae>
ha
<MichaelRaskin>
I think you can
<MichaelRaskin>
Let me check
<infinisil>
It works with attrsets
<infinisil>
> let attrs = { "." = 10; }; in attrs."."
<{^_^}>
10
<MichaelRaskin>
Ah right, the ${} works better than it should with attrsets, but not variables
Chiliparrot has joined #nixos
horner has joined #nixos
<{^_^}>
[nixpkgs] @adisbladis opened pull request #76646 → WIP: poetry2nix: init at 1.0.0 → https://git.io/JeN9M
<adisbladis>
martyet-o: ^
fenedor has joined #nixos
<chagra>
Synthetica: I have not managed to fix it because I couldn't recreate it
<chagra>
Synthetica: I don't know what caused it, it felt random
<notgne2>
Synthetica: are you using radeon graphics?
<Synthetica>
notgne2: No, intel integrated
<notgne2>
not sure then, I think I've seen that happen before with radeon drivers (but not amdgpu)
<Synthetica>
(oh, maybe also a good datapoint: music keeps playing when it happens)
<chagra>
yeah same with the intel integrated graphocs
<chagra>
oh yeah youtube videos will still work (audio only)
<notgne2>
might be worthwhile trying to ssh in and poke around with services and logs while it's frozen
<chagra>
the freeze is purely graphical but all the keybindings don't work either
drakonis has joined #nixos
<chagra>
which makes sense, because sway controls both keybindings and graphical output
<MichaelRaskin>
And, separately, all the extra policy stuff like window placement
<MichaelRaskin>
Let's just put everything in the same process, why not
<notgne2>
I was thinking more along the lines of kernel modesetting breaking independantly, one time with a faulty video driver config I had a machine that booted up until TTYs should be ready, then the screen would revert to the stage-1 boot, but ssh'ing in revealed everything was in working order besides KMS
horner has quit [Quit: zzZzZzZzZZZzzz…]
<chagra>
I really wish I could recreate it, it really bugged me that I didn't know how to trigger it
<chagra>
I know for sure it's independant of the load
orivej has quit [Ping timeout: 260 seconds]
<MichaelRaskin>
Close/open the lid?
<chagra>
doesn't affect anything
<chagra>
in fact the system does not go to sleep
<chagra>
when sway has froze
<MichaelRaskin>
I thought it could _trigger_ the freeze
<chagra>
oh no it doesn't for sure
<chagra>
I wrote a script with swayidle to do a lockscreen
<chagra>
and I put the computer to sleep multiple times in a row to test it
<chagra>
I've tried every situation with lid close/open. That does not trigger it
<chagra>
This may not be a nixos issue I think but just sway
<notgne2>
but I haven't spotted anything and don't really understand the priority system in this context
<infinisil>
Well you can use `environment.systemPackages = mkForce []` to empty the list completely
jmeredith has joined #nixos
gustavderdrache has quit [Quit: Leaving.]
<infinisil>
But then you also have to use mkForce in all other places to put more things inside
<chagra>
Maybe you can use the nixpkgs.pkgs option
<infinisil>
There's no other way to clear entries in a list in the current module system
<chagra>
If set, the pkgs argument to all NixOS modules is the value of this option, extended with nixpkgs.overlays, if that is also set. Either nixpkgs.crossSystem or nixpkgs.localSystem will be used in an assertion to check that the NixOS and Nixpkgs architectures match. Any other options in nixpkgs.*, notably config, will be ignored. If unset, the pkgs argument to all NixOS modules is determined as shown
<chagra>
in the default value for this option. The default value imports the Nixpkgs source files relative to the location of this NixOS module, because NixOS and Nixpkgs are distributed together for consistency, so the nixos in the default value is in fact a relative path. The config, overlays, localSystem, and crossSystem come from this option's siblings. This option can be used by applications like NixOps
<chagra>
to increase the performance of evaluation, or to create packages that depend on a container that should be built with the exact same evaluation of Nixpkgs, for example. Applications like this should set their default value using lib.mkDefault, so user-provided configuration can override it without using lib. Note that using a distinct version of Nixpkgs with NixOS may be an unexpected source of
<chagra>
problems. Use this option with care.
<infinisil>
That would work, with an overlay that changes nano to `null` or so
<chagra>
Perhaps you can set the nixpkgs.pkgs to an attibute set that's nixpkgs but nano is overriden to null
<chagra>
yeah exactly!
<infinisil>
Though this might cause many rebuilds depending on what else depends on nano
<infinisil>
Doesn't look like much though
drakonis has quit [Ping timeout: 258 seconds]
<{^_^}>
[nixpkgs] @alexfmpe opened pull request #76648 → Add missing space → https://git.io/JeNHy
<chagra>
what about nixpkgs.config.packageOverrides = pkgs: { nano = null; };
<chagra>
would that work
<chagra>
of course you would run nix why-depends --all /run/current-system nixpkgs.nano
<chagra>
before to check dependencies
<chagra>
nevermind it gives error: value is null while a set was expected
mexisme has joined #nixos
cosimone has quit [Quit: Terminated!]
mac10688 has joined #nixos
horner has joined #nixos
mexisme has quit [Ping timeout: 252 seconds]
drakonis has joined #nixos
cosimone has joined #nixos
Heirlung has joined #nixos
bvdw has quit [Write error: Connection reset by peer]
vika_nezrimaya has joined #nixos
bvdw has joined #nixos
<{^_^}>
[nixpkgs] @AndersonTorres pushed 0 commits to update/bochs: https://git.io/JeNHF
corsair has quit [Ping timeout: 258 seconds]
corsair has joined #nixos
<{^_^}>
[nixpkgs] @teto pushed commit from @filalex77 to master « starship: 0.32.1 -> 0.32.2 »: https://git.io/JeNQt
<{^_^}>
[nixpkgs] @vbgl pushed commit from @alexfmpe to master « ocamlPackages.conduit: 1.0.0 -> 1.4.0 and dependencies (#74821) »: https://git.io/JeN7x
horner has quit [Read error: Connection reset by peer]
horner has joined #nixos
vidbina has quit [Ping timeout: 258 seconds]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zupo has joined #nixos
<wucke13>
How can I import a nix file named as the current value of networking.hostName?
lovesegfault has joined #nixos
corsair has quit [Ping timeout: 260 seconds]
<elvishjerricco>
wucke13: `import "${config.networking.hostName}"` But beware that there are lots of ways to create infinite recursion this way
<lovesegfault>
colemickens: is colebot okay?
<lovesegfault>
Just noticed the bot didn't push in 6 days
<wucke13>
elvishjerricco: Ok, is there a more cleverish way of using one configuration on multiple systems?
<wucke13>
(Probably manually symlinking a specific nix file for each machine to /etc/nixos/configuration.nix is fine anyways, but I'm curious of alternatives)
<elvishjerricco>
wucke13: The complicated answer is that I use the module system (https://nixos.org/nixos/manual/index.html#sec-writing-modules) to define options that allow a config to concisely enable and configure things that matter to that config.
<elvishjerricco>
But I'll admit that is complicated
<chiefgoat>
I have a little headless machine that runs NixOS. Sometimes the power goes out. When the power comes back up, the little guy doesn't power on. Is there something I can set to true in my configuration.nix to have it boot when powered?
<multun>
chiefgoat: I'm pretty sure you can't, these are firmware / hardware settings
<multun>
chiefgoat: there might be a bios option for that
<multun>
or EFI, whatever
horner has quit [Quit: zzZzZzZzZZZzzz…]
<wucke13>
Pretty much everything which has an accesible bios configuration allows you set what to do on powerloss: On, Off, or Restore (whatever the case was before the powerloss).
<redd_at_36c3>
thanks multun, in case i tested it an it was added correctly. However yesterday on my installation via an NixOS USB stick it was not added
<multun>
that's odd
<pistache>
multun: it's the second choice, in case "ondemand" is not available in scaling_available_governors
Bryophyllum has quit [Quit: Bryophyllum]
<pistache>
(ondemand governor is not available on sandy bridge or later Intel CPUs)
<multun>
pistache: gotcha
<redd_at_36c3>
I cannot say from what kind of stick i booted, because it was provided to me by the people of the NixOS aseembly at the 36c3. So maybe this was not in "original" state...
<chagra>
if you have more issues with cpu or power management you might also be intrested in setting services.tlp in the future
<{^_^}>
[nixpkgs] @veprbl opened pull request #76659 → pkgsStatic: set BUILD_SHARED_LIBS=OFF for cmake → https://git.io/JeNdK
<multun>
redd_at_36c3: a bit old maybe
<chagra>
one of these days I need to go to a nixos assembly or nixcon
<chagra>
but alas visa issues and where I live make it hard and expensive
<chagra>
maybe I'll try to fit it in a visit one time
ng0 has joined #nixos
ng0 has joined #nixos
ng0 has quit [Changing host]
mexisme has joined #nixos
<redd_at_36c3>
chagra, It would be definitely worth it :)
<multun>
chagra: why would you pick tlp over powertop ?
<multun>
(I actually have no clue)
<chagra>
I don't know the difference either, I just know how to configure tlp and don't wanna bother changing over
<chagra>
does powertop offer better optimisations?
<chagra>
oh also it seems nixos has a config module for tlp only not powertop which is just autotuning
<chagra>
maybe that's why I chose, I can't remember
tokudan has quit [Quit: Dunno.]
<lukego>
How does gstreamer work with nixpkgs? Seems if I install 'gstreamer' I get an ancient legacy version, while if I install 'gst_all_1' I seem to get everything except the gst-launch frontend program? I'm trying to run examples from the internet that use gst-launch. If I find a copy in the nix store and run that directly then it tells me I'm lacking the v4l (video4linux) plugin too..
tokudan has joined #nixos
<p01ar>
hhhh
<p01ar>
whos at c3?
<redd_at_36c3>
me =)
misuzu has quit [Remote host closed the connection]
<p01ar>
h
<qyliss>
me
<pistache>
multun: tlp has thinkpad-specific features (battery start/stop thresholds, notably)
<pistache>
actually, tlp
misuzu has joined #nixos
<pistache>
scrape that^
ng0 has quit [Quit: leaving]
work_ has joined #nixos
<lukego>
(oh I think I'm supposed to use nix-shell...)
orivej has quit [Quit: No Ping reply in 180 seconds.]
<clever>
,libraries lukego
<{^_^}>
lukego: Don't install libraries through nix-env or systemPackages, use nix-shell instead. See https://nixos.wiki/wiki/FAQ/Libraries for details.
orivej has joined #nixos
<chagra>
lukego: gst_all_1 is an attribute set containing other packages. It seems that the latest gstreamer is the package with the attribute gst_all_1.gstreamer
ng0 has joined #nixos
ng0 has joined #nixos
<lukego>
thanks chagra. I'm onto new and exciting errors now that may not be related to the nix packaging :)
<chagra>
well that's life! going from one error to another
o1lo01ol_ has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
drakonis has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
jmeredith has quit [Quit: Connection closed for inactivity]
<chagra>
What's a good sales pitch for nixos and nix?
orivej has quit [Quit: No Ping reply in 180 seconds.]
<chagra>
Every time I get asked why do you use it, there are so many things to say I get confused and don't know how to summarize
<chagra>
it seems so hard to quickly explain to people in what ways it's great
orivej has joined #nixos
<HedgeMage>
chagra: When asked, I usually say that it gives me good visibility into what I'm installing and how, with the option to pull binary packages or compile from source, and a decent mix of package availability for the various hats I wear (software engineer, infosec geek, executive), without a ton of maintenance overhead.
<HedgeMage>
One of the hardest things for me, personally, was becoming a CISO and having to suddenly maintain a bunch of office-y things for my executive roles alongside dev and infosec tools. I can't sanely do that on anything but Gentoo or NixOS. Most distros are made for *a* use case, not that weird intersection of use cases.
<chagra>
I don't you have to have a certain job to benefit from nixos. I'm a medical student yet I find myself absolutely needing it to maintain my sanity
<drakonis>
^
<drakonis>
compsci instead
<chagra>
It's the only way I can know for sure that I'm using exactly what I want
<chagra>
otherwiser in the back my head an itch keeps wondering what's happening that I don't know about
<drakonis>
all my issues with other distros vanish under NixOS, hell, the issues other folks complain about, arent a thing
<drakonis>
except for the ones inherent to nixos's inner workings
<drakonis>
can always be worked around
<drakonis>
project request: find out which things alter what
<drakonis>
ie: find out all the options that modify a given option
<colemickens>
lovesegfault: no. It needs an overall rework. Gonna take a few more days probably
<drakonis>
a program for processing a configuration file that returns all the options that were modified by the config file and which options modified another option
<drakonis>
not exactly a stellar description of what i want to say
mexisme has quit [Ping timeout: 268 seconds]
equivrel has quit [Remote host closed the connection]
<chagra>
drakonis: so just looking up the definition of each option used and printing what they change when enabled?
<drakonis>
hmm kinda
<drakonis>
it should run deep and not at some surface level
<chagra>
the project is already written kinda I think. because that's Nix's job to find out what options change what other options
<drakonis>
i thought about it after the person from earlier asked why nano and w3m was available
o1lo01ol1o has quit [Ping timeout: 260 seconds]
<{^_^}>
[nixpkgs] @thoughtpolice merged pull request #76548 → monitoring plugins: fix path to sudo and mailq → https://git.io/JebMf
<{^_^}>
[nixpkgs] @thoughtpolice pushed commit from symphorien+git@xlumurb.eu to master « monitoring plugins: fix path to sudo and mailq »: https://git.io/JeNbI
<Guanin>
y log - anyone here who knows whats wrong?
<Guanin>
Hi, I'm currently trying to setup openldap for user management (first time working with it at all). It seems to work partially, as a wrong password gets rejected, while the valid password just restarts getty (with an error message in the journal). My configuration and the resulting files can be found here: https://gist.github.com/Amarandus/532641e29d92779f9d34f61cee76fbcb the eror message is in line 746 of m
cab404` has joined #nixos
cab404 has quit [Ping timeout: 248 seconds]
philr has joined #nixos
chagra has quit [Ping timeout: 240 seconds]
<das_j>
gchristensen: Is there anything to make ofborg evals PRs with an old broken evaluation like #76506?
<o1lo01ol1o>
I'm still trying to figure out how to reconstruct the path to cabal's data-dir. I given a cabal project called `src`, I can get figure out ${src.data}/share/${base-compiler.ghc.name}/x86_64-osx-ghc-8.8.1/${src.name}/foo, but where does "x86_64-osx-ghc-8.8.1" come from?
chloekek has quit [Ping timeout: 260 seconds]
<{^_^}>
[nixpkgs] @flokli pushed commit from @matthewbauer to release-19.09 « john: use proper configure flags »: https://git.io/JeNAu
h0m1 has quit [Quit: WeeChat 2.7]
h0m1 has joined #nixos
Synthetica has quit [Quit: Connection closed for inactivity]
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
_0mp has joined #nixos
<{^_^}>
[nixpkgs] @jtojnar opened pull request #76668 → sman: init at 0.2.0 → https://git.io/JeNAd
erictapen has quit [Ping timeout: 248 seconds]
<_0mp>
qyliss: hi there, I've heard rumors that you are working with FreeBSD. Would you like to help me finish the porting of Nix to FreeBSD, maybe?
o1lo01ol1o has quit [Remote host closed the connection]