<infty>
I'm thinking of dual-booting NixOS with Ubuntu. I'd have a shared home partition and a shared store partition (so I can use nix on Ubuntu). How terrible is this idea?
rager has quit [Ping timeout: 256 seconds]
rajivr has joined #nixos
<jybs>
ld.lld says unknown file type
rager has joined #nixos
infty has quit [Ping timeout: 260 seconds]
<{^_^}>
[nixpkgs] @NeQuissimus pushed 5 commits to release-20.09: https://git.io/JTpou
CyberManifest has joined #nixos
<jybs>
This is a Go binary if that helps
infty has joined #nixos
m0rphism has quit [Ping timeout: 260 seconds]
<jybs>
Oh did you mean ldd?
<jybs>
if so that's fine
<{^_^}>
[nixpkgs] @NeQuissimus opened pull request #102977 → linux-hardened: Do not block channel → https://git.io/JTpoS
codygman has quit [Read error: Connection reset by peer]
codygman has joined #nixos
thelounge812 has joined #nixos
heywoodlh has quit [Ping timeout: 265 seconds]
xelxebar has quit [Remote host closed the connection]
<__red__>
So - I'm actually writing a lot of security PRs for 20.03. Since it's freezing on the 27th am I just wasting my time or is this something that is actually useful?
<bqv>
oh wonderful, nrdxp kept nixflk up to date. guess i'll delete my template branch then
<colemickens>
I am trying to figure out if it makes sense to keep mine around or not
<colemickens>
I feel like mine isn't trying to be a template as much as ... a blog post about flakes along with a small code proof that it's just syntax.
<bqv>
heh
<veleiro>
colemickens: your repo was definitely an runner up for me
<veleiro>
but i think when i was looking at it, you were making a lot of changes
<colemickens>
I should maybe update nixcfg to point people at better resources in terms of templates/explanations. It's useful for a few specific oddball things I do, or have "mixins" written for... but it's a bit strict than what everyone else does. (I mix nixos/home-manager config freely, instead of keeping them separate/layered).
<colemickens>
s/a bit strict/a bit less strict/g
<veleiro>
i think i remember you using nixops, dont see any use of that here
<bqv>
if you're looking for something nixopsy, consider 'nixus'
<colemickens>
I used to use it with Azure 5 years ago or so I guess, but I manually deploy everything basically (copy, set new generation, activate)
<bqv>
it's extremely pleasant, and i ended up absorbing it into my flake
<colemickens>
(I want a nixus-lite, with a super constrained focus, I might make it)
<veleiro>
ah, i'm probably thinking of another i looked at
<bqv>
colemickens: lite how?! it's the lightest thing it could be!
thelounge812 has joined #nixos
<colemickens>
I don't want the health checking / multi host stuff. Honestly what I'm thinking wouldn't even be worth externalizing from my own config.
<colemickens>
Or the other potentially planned stuff.
<veleiro>
i havent seen nrd around but he did make a PR for being more arch agnostic
<veleiro>
in that template
<bqv>
oh. i see.
<{^_^}>
[nixpkgs] @jonringer pushed 9 commits to python-unstable: https://git.io/JTpNU
<colemickens>
Normally I wouldn't care and would use the pre-made thing, but my "manual" activation is like 5 lines, so I could wrap that into a module easily.
<colemickens>
bqv: we'll see, I may well wind up using nixus. it is popular with people with those folks who's style I like (that's a hard sentence to form well)
<bqv>
i don't see how the path could be any clearer
<veleiro>
git* repo :)
<bqv>
ok fine
<veleiro>
what is vervis
<bqv>
suffering
<{^_^}>
[nixpkgs] @lovesegfault opened pull request #102995 → plater: init at 2020-07-30 → https://git.io/JTpxW
<veleiro>
lol
<veleiro>
i feel that containers finally have proper use within nixos
<veleiro>
everywhere else they're being used to work around what nix provides
<{^_^}>
[nixpkgs] @jonringer pushed 3 commits to python-unstable: https://git.io/JTpxg
ilmu has quit [Ping timeout: 244 seconds]
bqv has quit [Quit: WeeChat 2.9]
<veleiro>
i was having a talk with a kubernetes guy the other day
bqv has joined #nixos
<dminuoso>
Im not seeing containers as that valuable inside nixos.
<veleiro>
its apparent that nixos's immutable generations are separate
<veleiro>
from an "image" or whatever you think of a container as
<veleiro>
I consider nixos immutable but i guess he thought differently as explained
<dminuoso>
well, they are orthogonal in the sense that a nixos generation doesn't compare to an image, it rather compares to a particular deployment of say docker-compose.
<dminuoso>
It makes more sense to compare nix derivations to images
<dminuoso>
(Which still is a stretch)
Fare has quit [Ping timeout: 264 seconds]
<veleiro>
containers in nixos could use some work i think, docker-compose is easier currently
explledboy has joined #nixos
<jybs>
It's not too bad!
<jybs>
extra-container helps a lot
ilmu has joined #nixos
<veleiro>
extra-container! yes
<dminuoso>
veleiro: The problem is that it steps back from many of the benefits of nix.
<jybs>
networking is the only nightmare
<dminuoso>
docker images are inherently non-deterministi
<dminuoso>
And it's rather tough to even specify docker containers inside nixos sensibly
<jybs>
I *still* can't get it to give the containers DHCP addresses and let them see each other / route out
<dminuoso>
jybs: arguably, networking is the nastiest part of docker
<veleiro>
i dont want this project's so and so dependencies in my nixos config
noudle has joined #nixos
<dminuoso>
Im willing to argue over 95% of docker users have bare to no understanding what happens beneath with iptables.
<veleiro>
but nixos containers are great for project-agnostic services
<veleiro>
just need a bit of friendlyness that docker-compose has, i think
<veleiro>
yes, what jybs says
captn3m0- has joined #nixos
<veleiro>
docker is a mess though, persistence works or sometimes doesnt
explledboy_ has joined #nixos
<veleiro>
i usually need to prune some volumes or containers to get definitions to update
explledboy has quit [Ping timeout: 264 seconds]
<veleiro>
i think that the purity goals like moving more to flakes will help solve
<veleiro>
these issues
captn3m0- has quit [Read error: Connection reset by peer]
<veleiro>
i have not read through the flakes threads to see the downsides however
<jybs>
I use nixos containers to run many, many stacked versions of Postgres, with various patches
<jybs>
it's great
<bqv>
that sounds terrifying
<veleiro>
but its good to have some standard way to approach configs
captn3m0- has joined #nixos
<jybs>
What are flakes? I see people talking about them a lot - but I don't really get it
ddellacosta has joined #nixos
<veleiro>
well per project services like postgres or any db is where things are interesting
sss2 has quit [Quit: Leaving]
<veleiro>
for example its really incredible in docker-compose to define a dir where a
<veleiro>
default sql file is always imported, i have a mariadb-init/db.sql that is auto
<veleiro>
imported if there is no db usually
sangoma has joined #nixos
<veleiro>
and per project the db exists within the project dir itself
<veleiro>
and not trashing my OS
<veleiro>
jybs: flakes are an RFC to package nix derivations themselves
<veleiro>
afaict its been everyone for themselves before the flake idea
<veleiro>
pinning git commits for example, instead of <nixpkgs> channels
Rusty1 has quit [Quit: WeeChat 2.3]
<jybs>
hmm
<jybs>
cool
domogled has quit [Ping timeout: 272 seconds]
<veleiro>
<nixpkgs> usually refers to a channel and that's not reproducible
ddellaco1 has joined #nixos
charukiewicz has quit [Quit: leaving]
<jybs>
While we are on the subject I don't suppose you know the magic incantation to get containers to be on the host network :-/
cantstanya has quit [Remote host closed the connection]
cantstanya has joined #nixos
<sphalerite>
,tell bitmapper (re bcm4331) I was afk, but I think it was mostly working towards the end of my time with it, with the b43 driver and networking.enableB43Firmware
<{^_^}>
sphalerite: I'll pass that on to bitmapper
hyper_ch5 has joined #nixos
domogled has quit [Ping timeout: 260 seconds]
waleee-cl has quit [Quit: Connection closed for inactivity]
<{^_^}>
[nixpkgs] @vbgl opened pull request #103001 → ocamlPackages.mirage-logs: init at 1.2.0 → https://git.io/JThII
gothnbass has joined #nixos
mallox has joined #nixos
shibboleth has quit [Remote host closed the connection]
Chiliparrot has joined #nixos
shibboleth has joined #nixos
jabster28 has joined #nixos
Chiliparrot has quit [Ping timeout: 244 seconds]
otulp has quit [Ping timeout: 246 seconds]
nschoe has joined #nixos
<{^_^}>
[nixpkgs] @Mic92 opened pull request #103002 → Various fixes for apple frameworks → https://git.io/JThLa
emmanuel_erc has quit [Ping timeout: 268 seconds]
jabster28 has quit [Ping timeout: 272 seconds]
<cole-h>
bqv: Not asleep, at work :P What's up?
gothnbass has quit [Ping timeout: 260 seconds]
MmeQuignon has joined #nixos
saschagrunert has joined #nixos
oxalica has quit [Quit: oxalica]
oxalica has joined #nixos
ncm[m] has joined #nixos
alp has joined #nixos
codygman has quit [Read error: Connection reset by peer]
codygman has joined #nixos
<Hm7000>
malorie: the command on the site is just a example
<Hm7000>
Home-manager has a zplug module
domogled has joined #nixos
sangoma has joined #nixos
<eeva>
Can someone tell me the difference between the release-20.09 and nixos-20.09 branches on nixpkgs github?
<cole-h>
release-20.09 is where all the release-targeted commits are merged to
<eeva>
Is any of these two in sync with a successful hydra build?
lsix has joined #nixos
<cole-h>
nixos-20.09 is the nixos-20.09 channel, which is updated after release-20.09 successfully evals and builds "some" important tests/packages/etc
<malorie>
Hm7000: I see. any idea why putting it into my configuration.nix wouldn't make it available, though?
<eeva>
cole-h: alright, thanks
<eeva>
I had trouble finding this information in the manual of nixpkgs
<benny>
Ke: there are a lot of pull requests, you can try to opt into marvin (which is a new bot that tries to help out with closing PRs) https://github.com/apps/marvin-mk2
felix[m]2 has joined #nixos
werner291 has quit [Ping timeout: 272 seconds]
werner292 is now known as werner291
<felix[m]2>
Hi everyone! First-time `NixOS` user here 👋 What exactly is the difference between the `python3`, `python3Full` and `python3Minimal` packages? They all seem to have the same description and link to the same nix expression 😕
<clever>
felix[m]2: python3Full is a variant with x11 support enabled
<clever>
felix[m]2: python3Minimal is a version without openssl, readlink, ncurses, sqlite, and several things disabled at compile time
civodul has joined #nixos
<felix[m]2>
Ah, ok. Makes sense. But that probably should be noted somewhere, because `Full` could also imply that it somes with a set of packages (similar to `texlive`). 🙂
<felix[m]2>
Thank you clever!
<Hm7000>
malorie: yeah look at the package in nix store
fuzen[m] has joined #nixos
<malorie>
it's in there, but I'd prefer not to source any paths with hashes in my zsh config, if possible
<Hm7000>
Thats why you use the ${}
FelixWittwer[m] has left #nixos ["User left"]
<raboof>
do I understand correctly that https://r13y.com/ reports on all packages that are needed to build the iso, not just the packages that actually end up in the iso?
<etu>
raboof: It's just the minimal iso it tracks
<raboof>
etu: right, but it's reporting a bunch of packages that I don't see in `nix-store --query --tree /nix/store/6vz21vkxs9jk6m0xa03yp73jry36mqxd-nixos-minimal-21.03pre56789.gfedcba-x86_64-linux.iso`
<cole-h>
clever: Kinda off-topic, but when you paste lines (like `python3Full = ...`), how are you doing so? Are you just opening the file in an editor and selecting the entire line?
<clever>
cole-h: triple-click in most programs, selects the entire line
<clever>
and middle-click in linux, pastes whatever you last selected
<malorie>
Hm7000: thanks. I'll read up on that
<cole-h>
OK, figured it was like that. Thanks
<clever>
cole-h: double and triple click add a lot of neat features that more people should use
<clever>
cole-h: if you double-click on text, but hold the 2nd click, and drag, it will always select whole words as you drag
<cole-h>
:d
<clever>
cole-h: and if you tripple click, but hold the 3rd and drag, it always selects whole lines
<cole-h>
Learning basic text manipulation with clever++ :P
<{^_^}>
clever's karma got increased to 539
<bqv>
cole-h: ah, missed your ping. been having some minor font weirdness with emacs for quite a while, was wondering if you'd have any insights
<cole-h>
Unfortunately not. I haven't seen anything of the sort, tbh. Probably because I use an Emacs "distro" (doom-emacs) -- a lot less fiddling I have to do :P
<bqv>
heh, fair.
codygman has quit [Ping timeout: 256 seconds]
<bqv>
ah well, i'll keep ignoring it
codygman has joined #nixos
<benny>
cole-h: what emacs version are you using with doom?
Sanchayan has quit [Ping timeout: 246 seconds]
<cole-h>
pgtk from the nixpkgs-wayland overlay
<cole-h>
Which is currently 28.0.50
<benny>
I wanted to try emacsGcc (which is also 28.0.50) but I don't know how to get vterm and sqlite working with that
<benny>
my nix-fu is not strong enough to get the proper environment variables set up for the dynamic module creation and loading
<bqv>
you don't need to
<bqv>
there's derivations for it, in places
<bqv>
search github for emacs-overlay
<bqv>
there's also derivations for pgtk+nativecomp
<bqv>
my flake has one, there's another in mjlbach(?)'s repo
<{^_^}>
[nixpkgs] @freezeboy opened pull request #103003 → k9copy: remove for being broken since 2014 → https://git.io/JTh3e
<benny>
I think that's what I mean by my lack of nix-fu, because this is what I have with 27.1 from nixpkgs: (emacsWithPackages (epkgs: (with epkgs; [ emacsql-sqlite vterm ])))
<benny>
but just using emacsGccWithPackages doesn't exist
<{^_^}>
[nixpkgs] @ehmry pushed commit from @vbgl to master « ocamlPackages.mirage-console: init at 3.0.2 »: https://git.io/JThsm
<mschwaig>
I'm running the nixos docker container in CI and 'nix develop --command whatever' does not print any output there. Any ideas why that might be?
<fuzen[m]>
omasanori[m], hopefully not on old hardware, compiled a "quick application" on my macbook forgetting its not quick when its not on thread-ripper ...
Sanchayan has joined #nixos
<dminuoso>
What's the protocol for adopting a machine with nixops in a team?
<dminuoso>
Im not really fond of permitting root password login
<dminuoso>
(nixops 2.0 seems to address this, but that one looks far from a release)
hyper_ch2 has quit [Ping timeout: 272 seconds]
<dminuoso>
Right now, if I dont want password based logins, Im forced to enable password based logins, let every team member create and run a deployment to generate keypairs, and then disable password login.
la-jesystani has quit [Ping timeout: 272 seconds]
<omasanori[m]>
fuzen: I'm in front of a 6xxx-era Core i5 laptop, so, yeah, not for compiling ex. Chromium or LibreOffice for sure
magnetophon has joined #nixos
Latr_work has joined #nixos
bqv has joined #nixos
<{^_^}>
[nixpkgs] @roberth merged pull request #103007 → ghc*: Increase build timeout to 1 day → https://git.io/JThnv
<{^_^}>
[nixpkgs] @roberth pushed 2 commits to staging-20.09: https://git.io/JThC9
sangoma has quit [Ping timeout: 260 seconds]
thibm has joined #nixos
<bqv>
something's very wrong
<bqv>
colemickens: i'd love to try, but my system's a bit unbootable
<{^_^}>
[nixpkgs] @zimbatm pushed commit from @leenaars to master « Sylk: 2.9.1 -> 2.9.2 (#102945) »: https://git.io/JThWG
thelounge812 has quit [Ping timeout: 256 seconds]
bqv has quit [Quit: WeeChat 2.9]
bqv has joined #nixos
<jkt>
is there some reasonable fix for Qt theme plugins? I'm on an unstable channel, I'm running plasma desktop and Trojita MUA, and after like 50% updates my styling for Trojita is totally broken with visual glitches
codygman has quit [Ping timeout: 256 seconds]
sangoma has joined #nixos
<{^_^}>
[nixpkgs] @roberth opened pull request #103008 → ghc*: Increase build timeout to 1 day → https://git.io/JThlT
codygman has joined #nixos
__monty__ has joined #nixos
werner292 has joined #nixos
<inquisitiv3>
How do I enable services on non-NixOS installations of Nix? Can't find any information in the Nix manual, and all research when googling talks about NixOS.
werner291 has quit [Ping timeout: 246 seconds]
werner292 is now known as werner291
<inquisitiv3>
More specifically I want to use Postgresql on my laptop where I'm running a Linux distro with Nix
<typetetris>
inquisitiv3: To my knowledge that is just not easily possible. Services from NixOS modules are meant for nixos. If you have some derivation containing systemd unit files, you probably enable them in the usual way.
hnOsmium0001 has quit [Quit: Connection closed for inactivity]
<typetetris>
inquisitiv3: For the special use case of postgresql on a non-nixos host I use shell scripts to create local databases. Can make you a gist, if you wish.
<inquisitiv3>
typetetris: If you would be so kind I would love it!
<{^_^}>
[nixpkgs] @peti pushed 992 commits to haskell-updates: https://git.io/JTh8D
zakame has joined #nixos
Chiliparrot has joined #nixos
<bqv>
Ok what might break the creation of system profiles?
<bqv>
I run switch-to-configuration boot, nothing happens
<thibm>
typetetris: the ./start command just ends with "Success. You can now start the database server using: /nix/store/…pg_ctl -D build/db/state -l logfile start"
<bqv>
I need someone clever
<typetetris>
thibm: That ./start is from IHP?
veleiro has quit [Ping timeout: 256 seconds]
<thibm>
typetetris: yes. You do "ihp-new blog", (then wait a few hours if you didn't configure cachix :p), then you get a blog directory with, among other, a start script
AmandaC has quit [Quit: Toodles]
spease has quit [Ping timeout: 265 seconds]
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « haskell-gi-base: update override for the new version »: https://git.io/JThBq
<thibm>
Is there any way to pretty print a string containing \n in the repl? (for instance various configurations files from of NixOS modules)
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JThBs
<thibm>
I end up by ":b writeText (expr)" and reading the resulting file in a text editor, it's annoying
<eyJhb>
Not sure how to phrase this best, but if I have a config in a .nix file, that I use in a mkDerivation, and I maybe want to override it, how would I best go about that? The output from the file should be a mkDerivation (path). If I do `let config = {}; in stdenv.mkDerivation { buildPhase = '# use config'; }`, how would I be able to override it? If that makes any sense.
rprije has quit [Ping timeout: 258 seconds]
<eyJhb>
I would also like to be able to have access to this config, so that I can override specific things. I am unsure if there is any examples/what the best way is to go about it?
<thibm>
eyJhb: if I understood correctly: use a function `{ config } : stdenv.mkDerivation {...}` and use `callPackage theFunction { config = defaultConfig; }`
ManiacOfMadness has quit [Ping timeout: 272 seconds]
<thibm>
eyJhb: better use makeOverridable instead of callPackage
AmandaC has joined #nixos
domogled has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed 2 commits to haskell-updates: https://git.io/JThRL
<bqv>
I don't get it
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « all-cabal-hashes: update to Hackage at 2020-11-06T01:07:40Z »: https://git.io/JThRs
<bqv>
Why is this so broken
<eyJhb>
thibm: Hmm, I think that makes sense to be. But If I do the let config in syntax, then I cannot get access to the config from outside the `in`, right?
<bqv>
I don't even know how to debug this
bqv has quit [Quit: WeeChat 2.9]
<thibm>
eyJhb: yes, for the second part of the question (access it), you need to expose it
<thibm>
you could try to use a mechanism like overrideAttrs to override only parts of config
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage2nix: update list of broken builds to fix evaluation errors on Hydra »: https://git.io/JThRV
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JThRr
bqv has joined #nixos
yinfeng has quit [Ping timeout: 260 seconds]
bqv has quit [Client Quit]
<{^_^}>
[nix] @zimbatm opened pull request #4224 → installer: simplify the per-build installation → https://git.io/JThRF
<roberth>
inquisitiv3: services are not a Nix feature, but rather NixOS. It's possible to reuse information from NixOS to generate systemd unit files, but that's not something many people do afaik
Darkmatter66 has quit [Read error: Connection reset by peer]
<eyJhb>
thibm: I am not sure I can use overrideAttrs
<{^_^}>
[nixpkgs] @FRidh merged pull request #103003 → k9copy: remove for being broken since 2014 → https://git.io/JTh3e
<{^_^}>
[nixpkgs] @FRidh pushed commit from @freezeboy to master « k9copy: remove for being broken since 2014 »: https://git.io/JTh04
Darkmatter66 has joined #nixos
<eyJhb>
thibm: this is what I have - https://gist.github.com/eyJhb/9caf7ac4af6dbb3c070e49d56d613393, and I get the thisismyconfig as output. But I am guessing it is because I can only override things that stdenv.mkDerivation passes to it, and not all walues.
<eyJhb>
Maybe your first suggestion would work better, however I really like this for some reason. But I am not sure, what is best, if you eg. has 50 pkg.nix, that needs a change
<thibm>
eyJhb: for overrideAttrs, you can't, not directly
<eyJhb>
So exposing the config + a default build function, is the ideal way to go?
<thibm>
I prefer the first solution and it's widely used in nixpkgs (through callPackge)
<eyJhb>
Let me try that real quick, I appreciate your help thibm ! :)
<thibm>
The only thing is that the "default config" would not be defined in the package file, unless you move the callPackage/makeOverridable call
<thibm>
eyJhb: BTW, you don't need { pkgs ? import <nixpkgs> {} } in pkg.nix, pkgs is already available in default.nix and used package in pkg.nix can be passed to mkDerivation through argument and callPackage mechanism
<thibm>
eyJhb: what are trying to do? Give us the big picture if possible
<eyJhb>
So basically, I have a lot of CTF challenges, such as here - https://github.com/fbsamples/fbctf-2019-challenges , for each one of these, I want to build them in some way, where each challenge has a config + a build derivation. Which could both e.g. writing the config to a file, adding some files, etc. - but it would be nice to be able to change all the configs, if you wanted to override some
<eyJhb>
specific setting using nix
<eyJhb>
Does that make sense?
<fkv>
Ill ask again, I have a problem where running switch-to-configuration boot does nothing at all, and switch puts my system in emergency mode. Attempting to monkeypatch the bootloader config and boot the config directly fails too. What ths flying fsck is going on?
<clever>
fkv: what happens if you run `bash -v result/bin/switch-to-configuration boot` ?
<fkv>
I did that
<fkv>
nothing happens
<fkv>
No new profile
<clever>
fkv: did it output anything?
<fkv>
No /boot changes
<fkv>
Nope
<clever>
fkv: what are the contents of switch-to-configuration?
<fkv>
Then I tried with NIXOS_INSTALL_BOOTLOADER, it reinstalled the bootloader, but didn't add the new system
<fkv>
Its got perl in it
<thibm>
eyJhb: the build derivation build the config or the challenge itself?
magnetophon has quit [Ping timeout: 256 seconds]
<clever>
fkv: what happens when you run `result/bin/switch-to-configuration boot ; echo $?` ?
werner292 has joined #nixos
<eyJhb>
Well, it basically is a attrset, which will be build in the derivation.
<fkv>
clever: 0
<eyJhb>
But it does not build the config itself, it builds like the challenge itself + adding the config in
kreyren has quit [Remote host closed the connection]
<teto>
I try to enter a nix-shell to develop on my python software with matplotlib but at runtime it gives me a `Namespace Gtk not available`. I have gobjectIntrospection pycairo pygobject in the env I am not sure what to try next
<eyJhb>
Does that make more sense thibm ?
<clever>
fkv: `strace -o logfiles -ff result/bin/switch-to-configuration boot`, and then pastebin all of the logfiles it made
<fkv>
Alright, sec
kreyren has joined #nixos
<fkv>
(I did check that, but it didn't look interesting)
<thibm>
eyJhb: the config argument is exposed in challenge1.nix and can be override thanks to callPackage. No need to add an extra function "cfg:" (like in your test2-pkg.nix)
<clever>
bqv: looks like your using systemd-boot
<fkv>
Yes
magnetophon has joined #nixos
<clever>
fkv: logfiles.2764 is when it ran the systemd-boot-builder.py script, which should copy your generations into the systemd-boot cfg files
<eyJhb>
thibm: But if I want the config to live in the challenge1.nix file, is that possible as well?
kreyren has quit [Remote host closed the connection]
<thibm>
eyJhb: if you want the default config in the challenge1.nix and not the default.nix, you can use default argument
<eyJhb>
I am guessing just setting config = { #config } in the file?
<eyJhb>
Hmm...
<FRidh>
teto: wrapping with wrapGAppsHook does not happen when you use python3.withPackages
<eyJhb>
thibm: I might have forgotten one last case, is when the config lives in the src of the challenge, and I will parse it from there, and then expose it. I am guessing that a default argument will not work
<teto>
FRidh: thanks, so I add it to the buildInputs of my mkShell ?
<clever>
fkv: then at 2246, i can see that its going over generations 296-309 of your system, and generating config
<clever>
fkv: let me double-check the source...
<eyJhb>
e.g. let src = ...; config = parseConfig src/config.yml; in { config = config; ...} ?
<fkv>
clever: none of those 296-309 are the system in question
<fkv>
I expect system-310 to have been created but it hasn't...
<clever>
fkv: i think thats the problem, i'm confirming it in the source now
<clever>
fkv: why do you expect system-310 to be created?
<clever>
fkv: switch-to-configuration never creates a generation
<fkv>
What does?
<clever>
fkv: nixos-rebuild handles the generations
<clever>
but you can do it manually as well
<fkv>
strange, ok
<inquisitiv3>
typetetris: How should I connect to the DB using psql? I've created the DB and started it using your scripts.
<clever>
fkv: i think the problem, is that grub's config builder, allows you to boot from a random storepath, that isnt in any generation
<clever>
fkv: so result/bin/switch-to-configuration boot, works with grub
<clever>
fkv: but systemd-boot's builder, is dumber, and only ever puts the generations into the cfg
<clever>
fkv: so you must run something like `nix-env --profile /nix/var/nix/profiles/system --set -i /nix/store/foo` to update the profile first
<fkv>
Gotcha.
<clever>
15 # system.build.toplevel path
<clever>
16 my $defaultConfig = $ARGV[1] or die;
<clever>
yeah, there it is, grub will default to the path switch-to-configuration came from
<thibm>
eyJhb: the easiest way is to expose the config by defining it in the mkDerivation argument. If it's an attrset, put it in passthru (because it cannot be coerced to string)
<inquisitiv3>
typetetris: I'm new to databases, and your script contains commands that isn't mentioned on Postgresql's getting started guide
<thibm>
eyJhb: (BTW, I changed "config" to "configuration" in the last example, because config already exist in pkgs. This break the default argument thing because pkgs.config will be given as argument by callPackage)
<typetetris>
inquisitiv3: I updated the gist of startdb.sh. Its psql -h localhost -p 5431 -U dev dev if you didn't changed the port number.
<clever>
208 for gen in gens:
<FRidh>
teto: probably yes
<clever>
210 if os.readlink(system_dir(*gen)) == args.default_config:
<clever>
fkv: yep bingo, systemd-boot figures out which one is default, by iterating over every GENERATE, and comparing it to where switch-to-configuration came from
<clever>
generation*
<clever>
fkv: so grub forces the build to be default, even if it wasnt a choice, while systemd-boot only makes it the default if its already a generation
<bqv>
awesome
<bqv>
so now it at least has a boot entry
<bqv>
now, my problem is that "switch" will crash my system
<inquisitiv3>
typetetris: Thanks! <3
<clever>
bqv: but you can now try to reboot into that new entry
<bqv>
i can try booting directly into it again, but last time that failed, and i can no longer reboot ita
<bqv>
yeah, i will do, i'll have to wait an hour to try that though
<eyJhb>
hmm thibm , that seems like it could work! :D Thanks!
<clever>
bqv: which nixpkgs channel did you build from?
<bqv>
unstable-small and master
magnetophon has quit [Ping timeout: 246 seconds]
<eyJhb>
thibm++
<{^_^}>
thibm's karma got increased to 0o5
<clever>
bqv: master goes thru less testing, and could have problems that make it unable to boot
<thibm>
eyJhb: cool :)
<bqv>
understandable
<eyJhb>
Thanks for your help! Hopefully I can get further now, after uni at least :p
<{^_^}>
[nixpkgs] @FRidh pushed 143 commits to staging-next: https://git.io/JThzD
<{^_^}>
[nixpkgs] @roberth opened pull request #103010 → Staging next tzdata revert to fat format → https://git.io/JThzS
<bqv>
clever: if it rings any bells, the issues i had on boot last time was that it could not find the disk /dev/vda (i don't even normally have a vda?) and could not mount /nix/.ro-store
<bqv>
former was when i tried using the new system's initrd, latter was when i used the old system's initrd to see how far i'd get
<clever>
bqv: double-check what you have in the fileSystems for the config you just built
<clever>
bqv: and why is it trying to touch .ro-store, what are happening with /nix/store?
<bqv>
there was a systemd service to mount it
<bqv>
that failed, and then it dropped me into an emergency shell
<clever>
bqv: what was the configuration.nix you used to build the system?
<bqv>
that will not be simple. it's my flake
magnetophon has joined #nixos
euandreh has quit [Remote host closed the connection]
<clever>
bqv: it sounds like you have something very wrong with the fileSystems. setup in it
<bqv>
it does, however i haven't touched that for months
<clever>
bqv: that looks like you included qemu-vm.nix in your imports
<bqv>
that doesn't look right...
<clever>
bqv: so its expecting you to boot under qemu
<bqv>
oh god fscking dammit
<clever>
i dont think /bin/fsck can fix that :P
<bqv>
:D
<thibm>
clever: OK, it make sense. Thanks a lot
alp has quit [Ping timeout: 272 seconds]
<thibm>
makes* (english conjugaison is so hard :p)
<clever>
qemu-vm.nix has a lot of mkForce commands to ignore your cfg, and then make it work under qemu
<bqv>
this explains so much
<bqv>
ok, i'll build this and switch, hope for the best. if it breaks and i vanish, i'll be back in around an hour
<clever>
bqv: this also explains why switch broke things so badly
<clever>
bqv: the new config said to mount vda, and you where missing vda, so crash
<bqv>
yeah :/
euandreh has joined #nixos
cosimone has joined #nixos
<thibm>
clever: one extra question: what happen if we bump /nix/var/nix/profiles/system generation with --set but do not "switch/boot" to the configuration (i.e do not upgrade bootloader menu)?
<thibm>
In other words, is /nix/var/nix/profiles/system just here to maintain history of generations, or is it involved in the boot/activation process?
<clever>
thibm: then the bootloader will continue to boot whatever the old config said to boot
<clever>
thibm: the main use of the generations, is so nix-collect-garbage knows what to keep
<clever>
and to act as input when building the bootloader config
cosimone has quit [Remote host closed the connection]
cosimone has joined #nixos
Swant has quit [Quit: gone to IKEA]
Swant has joined #nixos
cosimone has quit [Remote host closed the connection]
cosimone has joined #nixos
<thibm>
Well, tinydns nixos test pass on 20.09 and is broken on 20.03, but actual server works on 20.03 and is broken on 20.09. Neither the package, the test, or the nixos module changed.
cosimone has quit [Remote host closed the connection]
fendor has joined #nixos
fendor has quit [Remote host closed the connection]
<clever>
cinimod: you likely dont even need shell.nix, try just running nix-shell on the same attr you would nix-build
<cinimod>
I don't want to build via nix - I want to have an environment where I can build
<ncm[m]>
Packaging Go programs... seems like the way to handle static dependenies / "vendoring" is to always build Go programs and libraries from source at install time. I.e. pretend it is Python. Dependencies' representation on host are git repositories with revisions actually used fetched.
<clever>
cinimod: which is exactly what nix-shell is for, you point it to a derivation that would have normally built it for you
<clever>
cinimod: and you get the exact environment nix-build would have been using
<cinimod>
Well blow me down - that works
<bqv>
clever: so it seems my problem is that it's impossible to get below min-free by collecting garbage :D
<bqv>
gonna have to go spelunking ...
<clever>
bqv: you can `--option min-free 1234` to lower the limit more
<clever>
cinimod: ive always considered mkShell as a useless function, you can do the exact same things with stdenv.mkDerivation
<cinimod>
clever: thank you very much - I did not know that was what nix-shell was for
<clever>
bqv: find the pid of the main nix-daemon (the one without a pid in its args list), and run `strace -f -p PID` on it, then repeat `nix-build '<nixpkgs>' -A hello` as non-root
<clever>
that will capture what nix-daemon is up to when you try the build
<clever>
alexfmpe: nativeBuildInputs are only for tools ran at build time, not libraries/headers used in the build
<clever>
alexfmpe: you want to add it to the buildInputs instead
<ncm[m]>
* Packaging Go programs... seems like the way to handle static dependencies / "vendoring" is to always build Go programs and libraries from source at install time. I.e. pretend it is Python. Dependencies' representation on host are git repositories with revisions actually used fetched.
<nahamu>
Is there an easy way to review a change that says "This PR is no longer needed and can be closed"?
<gchristensen>
srhb: I don't suppose morph can build on the remote host?
alp has quit [Ping timeout: 272 seconds]
<srhb>
gchristensen: There's no magic involved, it's regular substitution mechanics. We tend to substitute on the hosts (but have ci build things in advance) -- we haven't set up any special mechanism for building. Not sure if the ssh-ng substituter could do it.
<gchristensen>
that answer doesn't make sense yet, I think I need to read the code :)
<srhb>
gchristensen: Simplified: morph can only nix copy
<thibm>
"--option extra-substituters" is cool to use an extra cache on command line and avoid shadowing cache.nixos.org, but there's no "extra-trusted-public-keys" option, so it becomes kind of useless if we still need to put cache.nixos.org's public key on the command line. Or am I missing something?
<gchristensen>
okay, so I can't ask morph to please build <<weird architecture>> on the other end, I'd need to set it up in the machines file
bkv has joined #nixos
<srhb>
gchristensen: Yes indeed.
<gchristensen>
dag! okay, though
<srhb>
We're wistfully discussing morph2 (like you would some mythical creature) and it would be interesting to make the mechanism for producing the remote path more generic.
<gchristensen>
that way might be madness
sss2 has joined #nixos
<srhb>
"We're all mad here"
saschagrunert has quit [Quit: Leaving]
zakame has quit [Read error: Connection reset by peer]
<gchristensen>
ain't that the truth
bqv has quit [Ping timeout: 244 seconds]
bkv is now known as bqv
waleee-cl has joined #nixos
<bqv>
it just wouldn't be the Linux Experience if everything didn't break every fortnight like clockwork
sangoma has quit [Ping timeout: 260 seconds]
<{^_^}>
[nixpkgs] @flokli merged pull request #102157 → wire-desktop: linux 3.20.2934 -> 3.21.2936, mac 3.20.3912 -> 3.21.3959 → https://git.io/JTS8j
<{^_^}>
[nixpkgs] @cheriimoya opened pull request #103023 → bluetooth_battery: init at 1.2.0 → https://git.io/JThjz
abathur has quit [Quit: abathur]
<exarkun>
in `nix-shell -p python2Packages.virtualenv`, `pip download ...` downloaded some manylinux1 and manylinux2010 wheels. is nixpkgs Python now manylinux compatible?
<flokli>
hmmh, is there a simple way to include /etc/passwd or /usr/bin/env in buildLayeredImage?
<gchristensen>
make a buildEnv with $out/etc/passwd and $out/usr/bin/env and put it in contents, I think
<gustavderdrache>
i used runCommand to symlink "${busybox}/bin/env" to $out/usr/bin/env
<flokli>
yeah, did that too
<flokli>
but nss is gnarly
<gchristensen>
yeah it is
<flokli>
I mean, sure I could add this all manually
thelounge812 has quit [Ping timeout: 240 seconds]
<flokli>
but what about just having one derivation with the usr/bin/env symlink, and another one with nss for root?
<flokli>
buildImage has some code for this, but I don't really want to move back to that, just for it.
bennofs1 has joined #nixos
<gchristensen>
yeah as long as it is in `contents` it should be linked at the root
<cole-h>
notgne2: How do you compare the commits? The sha changes upon merge/rebase, no?
<notgne2>
shit that's probably true
<notgne2>
you could also just kindof cheat and take advantage of the fact that any commit alert directly after a merge alert is probably not relevant
<simpson>
I wonder whether folks have considered having the automatic PRs be *batched*, rather than per-package. I don't want to ping the bot authors right now, because they've probably already considered this, but...
<simpson>
cole-h: I don't see this particular detail addressed. And I'd imagine that the discussion starts with statistics; batching up package changes is going to require an estimate of whether a particular package is likely to break on update.
<supersandro2000>
shouldn't this be enough to detect?
<notgne2>
supersandro2000: yeah it's definitely possible to do correctly, but adding more api things is more of a pain in the ass than a 1 line patch
<cole-h>
simpson: Not in the RFC itself, but in the discussion on the RFC
<supersandro2000>
you could probably fit that into one very long line
<supersandro2000>
Isn't rule number one when using Git: Never (force) push to master?
<cole-h>
Not necessarily
<supersandro2000>
*if you are not working alone on a repo
<cole-h>
That might be guideline number one in large projects with multiple collaborators, but not for Git itself ;)
<notgne2>
yeah now sometimes it's "Never (force) push to main"
<supersandro2000>
main master trunk whatever. Doesn't matter.
pushqrdx has joined #nixos
<simpson>
There are no rules when it comes to git. That's kind of the point. Most tree states are legal, and many of those can even be recovered somewhat. People have abused the format to store all sorts of things in all sorts of ways.
<simpson>
What's going on here is not a git integration, but a GitHub integration.
dbmikus__ has joined #nixos
urkk has left #nixos [#nixos]
<meatcar>
My first PR has been hanging in nixpkgs for over a month, what can I do to help to move it forward without creating unnescessary noise? Do I need to chase down both of the code owners? #99359
<turlando>
eyJhb I though NixOps was for managing multiple NixOS systems, while I'd like to manage VMs with other OSs inside a NixOS host
<tpw_rules>
so i'm trying to install tensorflow with cuda for python, and it appears nix is building and running all 16,296 of LLVMs tests? it's taking a bit
<eyJhb>
turlando: oh sorry, that is what I thought you meant, managing nixos + vms
<turlando>
no worries
<tpw_rules>
oh boy the number of tests is increasing
<primeos>
LnL: I know, but in this case it actually wasn't the old(?) GitHub bug for once :D No big deal though
<tpw_rules>
maybe this is just compiling tensorflow?
zupo has joined #nixos
<LnL>
primeos: oh, what happened then?
mallox has quit [Quit: WeeChat 2.9]
mallox has joined #nixos
werner292 has joined #nixos
davidv7 has quit [Ping timeout: 256 seconds]
magnetophon has quit [Read error: Connection reset by peer]
sangoma has joined #nixos
werner291 has quit [Ping timeout: 272 seconds]
werner292 is now known as werner291
magnetophon has joined #nixos
Izorkin has joined #nixos
ris has joined #nixos
<{^_^}>
[nixpkgs] @jonringer merged pull request #100602 → python3Packages.pyelftools: disable tests on Darwin → https://git.io/JT3Hu
<primeos>
LnL: Not sure, happened during the last force-push, some Git "accident" I guess :)
<{^_^}>
[nixpkgs] @jonringer pushed commit from @siriobalmelli to master « python3Packages.pyelftools: disable tests on Darwin »: https://git.io/JTj39
thc202 has quit [Ping timeout: 244 seconds]
jabster28 has joined #nixos
thelounge812 has quit [Ping timeout: 256 seconds]
jabster28 has quit [Ping timeout: 272 seconds]
jabster28 has joined #nixos
<{^_^}>
[nixpkgs] @tadfisher opened pull request #103030 → remarkable2-toolchain: init at 2.5.2 → https://git.io/JTjsa
<supersandro2000>
primeos: just always use --force-with-lease
<kini>
hmm, does hydra only build python 3.8 packages and not 3.6 or 3.7?
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « haskell-time-compat: update to latest version so that binary-instances compiles »: https://git.io/JTj8Q
<kini>
looks like it does build 3.7 but not 3.6, I see -- why is that?
Izorkin has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « haskell-time-compat: update to latest version so that binary-instances compiles »: https://git.io/JTj8N
Darkmatter66_ has quit [Ping timeout: 260 seconds]
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « python27Packages.objgraph: 3.4.1 -> 3.5.0 »: https://git.io/JTjR3
<kini>
normally I'd just use python3WithPackages which works fine and points to 3.8, but I'm working on something that needs to be deployed to RHEL8 which still only has python 3.6, so I switched to python36WithPackages in my shell.nix and was surprised to find that suddenly I was compiling fortran code, lol (building scipy)
jabster28 has quit [Ping timeout: 272 seconds]
<supersandro2000>
kini: thats the feel of enterprise software aka yesterday
<infinisil>
kini: You could use an older nixpkgs version when 3.6 was the default to get caches
<infinisil>
Though then you'd have outdated packages
pushqrdx has quit [Remote host closed the connection]