<{^_^}>
[nixpkgs] @thoughtpolice merged pull request #65005 → all-packages: add static exe alias for haskell 'shake' package → https://git.io/fj1go
<{^_^}>
[nixpkgs] @thoughtpolice pushed to master « all-packages: add static exe alias for haskell 'shake' package »: https://git.io/fj1az
doyougnu has quit [Ping timeout: 245 seconds]
<exarkun>
gchristensen: I thought it was lots but looks like it's one package getting bumped a couple times, 9354bc967d381c4e27db94b19b9c9a54f400130b and 6c233963dccc95dfd25f027ea84d8656eff303a0
<gchristensen>
ah, you're a knot user?
<{^_^}>
[nixpkgs] @dtzWill opened pull request #65008 → victor-mono: (re)init at 1.2.1 → https://git.io/fj1ag
recj has quit [Ping timeout: 245 seconds]
<exarkun>
gchristensen: nope! I was just reading the release-19.03 branch history to try to get an idea of why I'm fetching so many things
<gchristensen>
gotcha
<gchristensen>
probably a mass rebuild
<exarkun>
it might be cool to have a tool that takes a nixpkgs attr path and gives you a git log of commits that made changes that force a rebuild of that thing
<infinisil>
exarkun: Yeah, I really wish we had more such high-level tools
<exarkun>
if I succeed in convincing my employer to use nix, possibly
<infinisil>
Well, the easy way to implement this tool would be to just go through every commit, evaluate the attribute to find the drv path, and check if it changed between two commits
<infinisil>
But, this would be hella slow
<gchristensen>
`git bisect` could work to find the first
<infinisil>
Ah yeah
<infinisil>
And also I'm thinking, if we remember the paths in nixpkgs that can even influence that derivation, we can just ignore all commits that don't modify those paths
<infinisil>
Lorri does something similar probably
<gchristensen>
:)
thonkpod has joined #nixos
_pash has joined #nixos
<_pash>
im considering leaving arch for nixos+zfs, should i go for it?
<gchristensen>
sure
<infinisil>
And now I'm thinking, it's a waste that everybody has to evaluate all those drv's
<_pash>
what am i missing out on?
<infinisil>
There could be a nixpkgs-drv repo which contains all drv paths preevaluated
<infinisil>
Which gets automatically updated as nixpkgs gets commits
<gchristensen>
_pash: by sticking with arch, or switching to nixos?
<infinisil>
That would be neat
<_pash>
gchristensen: by sticking with arch, what am i missing out on?
<gchristensen>
nixos's rollbacks and safe upgrades are pretty weet
justanotheruser has quit [Ping timeout: 245 seconds]
xkapastel has quit [Quit: Connection closed for inactivity]
<infinisil>
_pash: Declarative system configuration, so if you happen to get a new computer, you can effortlessly make it have the same config. No messing with 100 commands to set everything up that you need
<exarkun>
infinisil: what about my konsole customizations
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/fj1a1
<exarkun>
and my i3 config
<infinisil>
exarkun: Well, some configs are stateful by default, but you can make them be declarative
<gchristensen>
my i3 config is contained within my nixos setup
<infinisil>
As I'm doing it with almost everything
<exarkun>
gchristensen: I guess you can have one system-wide config in the store? and since computers are basically single-user now that's fine ...
<gchristensen>
yes
<infinisil>
exarkun: home-manager is very nice for that, also supports multi-user setups if that's needed
<exarkun>
infinisil: indeed. also, runs on arch.
<infinisil>
And macOS even :)
<exarkun>
(still downloadin')
<gchristensen>
yeah, that is the nix model
<gchristensen>
updates deep in the tree make for big updates
<_pash>
if setting nixos with zfs, i dont need to partition for multiple zfs mounts, zfs does that with a pool, correct?
<clever>
_pash: yeah
<_pash>
is it possible to resize zfs partitions from the head of it?
<infinisil>
exarkun: gchristensen: Intensional store would help a lot though .. rfcs#17
<gchristensen>
infinisil: I'm actually not sure it would help a lot
<exarkun>
my brain can't cope with the spelling of that word
stepcut has quit [Remote host closed the connection]
<gchristensen>
infinisil: it would be interesting to determine exactly how much a lot is
<infinisil>
gchristensen: Assuming the bootstrap builds are deterministic, and you made some deep change that only modifies the build a bit, without changing the output, it should make it a lot quicker, right?
<_pash>
clever: soz, don't know much about zfs, is it possible to resize it from head?
<infinisil>
gchristensen: I guess the catch is that changes that don't modify the output aren't very useful..
<gchristensen>
exactly
stepcut has joined #nixos
<gchristensen>
and most of the builds are not "noops"
<clever>
_pash: you can only extend the tail
<_pash>
clever: gotcha thanks
chpatrick has quit [Quit: Connection closed for inactivity]
<adamantium>
Any zfs'ers here? In a rescue situation, i've imported my pool and trying to manually mount the datasets to /mnt, e.g. mount -t zfs rpool/ROOT/nixos /mnt
<adamantium>
But i'm unable to mount rpool/NIX/nix to /mnt/nix with:
moet has quit [Ping timeout: 272 seconds]
<adamantium>
oh wait ! I got it... had a silly typo
lovesegfault has quit [Quit: WeeChat 2.5]
jluttine has joined #nixos
<adamantium>
help: running "nixos-rebuild switch --install-bootloader" from within a chroot (nixos-enter) is giving me "error: unable to start build process"
<adamantium>
and error: cloning build process: operation not permitted
<DigitalKiwi>
for emergency after hours support please subscribe to our gold plan
<adamantium>
;)
<adamantium>
Hi kiwi
<hpfr[m]>
What do module enable options do by default? Install packages
<adamantium>
wanna tmux into it and fix it for meh
<DigitalKiwi>
unfortunatley i don't even know tmux let alone how to fix your problem :(
<DigitalKiwi>
what'd you do to get in this situation?
<{^_^}>
simonmichael/hledger#1030 (by simonmichael, 8 weeks ago, open): nix install on linux can fail with "cloning builder process: Operation not permitted"
<{^_^}>
#53576 (by alexandergall, 27 weeks ago, open): nix-env fails in chroot if nix.useSandbox = true
<DigitalKiwi>
isn't this when you just run themelios on it ;p
<adamantium>
i wonder if sandboxing is enabled on my usb stick could be the problem
<adamantium>
DigitalKiwi: yeah but thats too easy
<adamantium>
;)
<DigitalKiwi>
i think that's the point :D
<adamantium>
yeah okay screw it
<DigitalKiwi>
though you'll learn more if you figure it out...
<adamantium>
meh
<adamantium>
it's already doin its thang
<DigitalKiwi>
i have those letters in a box i used keep at my desk at work
<adamantium>
i'm deleting useSandbox = true, though. for sure
<adamantium>
i think that was the problem
<adamantium>
but i will have to rebuild my usb rescue iso
<DigitalKiwi>
that was part of my initiation into themelios making a new live iso from a live iso
<DigitalKiwi>
twice
<DigitalKiwi>
because the first time i messed up
<adamantium>
nice
<adamantium>
well time to watch netflix and chill
<adamantium>
lol o/
<DigitalKiwi>
*snicker*
<adamantium>
lol
Myhlamaeus has quit [Read error: Connection reset by peer]
moet has quit [Ping timeout: 245 seconds]
fresheyeball has joined #nixos
drakonis1 has joined #nixos
<fresheyeball>
I am interesting in learning to use a synthesizer for music
<fresheyeball>
anyone making music with nixos?
<fresheyeball>
any synth you can recommend?
mexisme has quit [Ping timeout: 246 seconds]
drakonis_ has quit [Ping timeout: 245 seconds]
<hpfr[m]>
How can I have home manager manage system packages? Does it just work, or if I enable firefox in home.nix will it install another Firefox for the user?
<fresheyeball>
hpfr[m]: what do you mean by system packages?
<hpfr[m]>
Packages declared in configuration. If
<hpfr[m]>
.nix*
<fresheyeball>
hpfr[m]: do you mean environment packages?
Rusty1 has quit [Quit: Konversation terminated!]
<hpfr[m]>
I want to use home manager to manage the configurations for system packages, sorry
jasongrossman has quit [Ping timeout: 245 seconds]
<hpfr[m]>
Not install another version for the user because I set enable=true;
<fresheyeball>
hpfr[m]: so why do you have environment.systemPackages in the first place?
<hpfr[m]>
To declare system packages?
<fresheyeball>
why do you need system packages
<fresheyeball>
let me ask another way
<fresheyeball>
since you manage your environment with home manager, why have a system environment in addition?
<fresheyeball>
for example I use it to provide tools to the root user
Ridout has quit [Quit: Lost terminal]
hedgie has joined #nixos
<fresheyeball>
Idk I think that would be the simplest
<fresheyeball>
next would be to just not have overlapping packages in your $PATH
wfranzini has quit [Remote host closed the connection]
<fresheyeball>
like if you have git in environment.systemPackages
<fresheyeball>
just don't have git in your users environment, it's already provided.
<fresheyeball>
if for some reason you absolutely have to have the variable in both places
<fresheyeball>
you can ensure they are the same by importing your home manager package set in configuration.nix via an `import` statement
<fresheyeball>
and just passing the same package by reference.
<hpfr[m]>
I’m referring to cases where the package is system wide and I want to configure per user. If git is in my system list, can i configure it with home manager without home manager trying to install it to my user env?
<fresheyeball>
hpfr[m]: does home manager install git by default or something?
<fresheyeball>
can't you ask home manager not to install it?
<hpfr[m]>
That’s what I don’t know. Does programs.git.enable = true try to install a version to your user env, or does it just enable configuration of the systemwide one? I’m new to this
<fresheyeball>
oic
<fresheyeball>
it's home manager
<fresheyeball>
so I would expect you are managing an additional user environment with it
<fresheyeball>
so it should not impact your core system configuration
<fresheyeball>
if you want to use systemwide git, instead of some local user environment git
<fresheyeball>
then I would set programs.git.enable = true in configuration.nix
<fresheyeball>
and remove that code in home mananger
<fresheyeball>
you don't have to set enable = false; its the default, so you can just remove it
<fresheyeball>
if you want a user git and a system git
<fresheyeball>
they can certainly diverge in version
<fresheyeball>
hpfr[m]: if you are new to nix, I kinda think home manager might be making your life harder
<fresheyeball>
personally I don't see the value, there is something nice about having everything in a single expression
m0rphism has joined #nixos
palo1 has joined #nixos
mexisme has joined #nixos
slack1256 has quit [Ping timeout: 246 seconds]
<hpfr[m]>
I just want to use home manager to manage the configuration of system packages. Like if multiple users on my system are using git, they could both use home manager to configure the systemwide git installation
<adamantium>
the value is the added features for managing dotfiles in home
<adamantium>
it's not immaculate but maybe a few nuggets in it for u
<hpfr[m]>
I found that wiki link which gives basically what I want, but I was hoping I could both just manage system packages and install and manage a few user packages
<fresheyeball>
oh interesting
<fresheyeball>
adamantium: how does it accomplish that?
<colemickens>
wow, looking at Guix, they even have an ungoogled+wayland chromium build already!
<fresheyeball>
adamantium: is it symlinking to dot files?
abathur has quit [Ping timeout: 248 seconds]
<adamantium>
fresheyeball: yes it is
<adamantium>
it also manages its own dotfiles for which it has configuration options
<adamantium>
so that i can configure programs like e.g. dunst in nix lang
<fresheyeball>
adamantium: right
<fresheyeball>
hmm
<adamantium>
the few programs it doesnt support you can just symlink, which is what i do in the above example
<adamantium>
the above example shows how to install programs in the global environment and the user environment, and it is all managed with the nixos-rebuild switch commands. i don't use the home-manager cli commands
<adamantium>
it's pretty slick
<hpfr[m]>
Why do you have everything in a modules folder
<fresheyeball>
oh hmmm
<fresheyeball>
I do like that
<fresheyeball>
and a slock user! wonderful
<adamantium>
hpfr[m]: everything in my modules folder is recursively auto loaded
<adamantium>
they're all toggle-able modules in my setup
<fresheyeball>
adamantium: I still don't see how you can set symlinks per user with nix
<adamantium>
hehe
<fresheyeball>
do you know where in the code it's doing that?
<adamantium>
fresheyeball: and that is all within my home-manager.users.adam = { section
abbiya has joined #nixos
<fresheyeball>
adamantium: I mean inside of home manager
<fresheyeball>
it has some code that takes your configuration
<adamantium>
it runs home-manager as a per user service
<adamantium>
and builds your dotfiles and symlinks and things as nix pkgs
<fresheyeball>
oic
<fresheyeball>
home manager IS the side effect
<adamantium>
and checks/recreates all of the symlinks as necessary everytime you restart the home-manager-userNAME.service
<fresheyeball>
there isn't a way to do it with nix, so you have to have a service that does the replace
<fresheyeball>
and that service can be nix
<fresheyeball>
so you get the one big expression, as well as home manager which is slick
<adamantium>
i dont really do internals
<adamantium>
i just enjoy how easy it is to administrate
<fresheyeball>
for sure
<hpfr[m]>
Why do you have a bunch of random gnome themes in system packages and home packages
<hpfr[m]>
Looks like most of the rest of your packages are handled by home manager
mexisme has quit [Read error: Connection reset by peer]
<fresheyeball>
is there a way to modify the user environment actually?
<adamantium>
hpfr[m]: that's OCD at work you see there, and that's because home-manager doesn't have a firejail option, and so my themes aren't accessible by firejail apps unless they're system-wide, too
<hpfr[m]>
So you don’t make many packages system wide
<adamantium>
in my case, didn't see the need to. now if i wanted i could have multiple users with entirely different themes and desktop systems etc
<hpfr[m]>
Yeah
<adamantium>
i do agree home-manager can be overkill
<adamantium>
but i like using it
<adamantium>
i can never get it to work correctly on bootstrap/initial chroot
<adamantium>
so i always end up commenting out home manager on a new install
abbiya has quit [Quit: abbiya]
<hpfr[m]>
I think I’m gonna end up using the wiki tweak, home manager just seems too patched-on to me. I guess it’s because it’s for nix, not nixos. It’s too bad something more integrated like that NixUP PR was never finished
<hpfr[m]>
Oh so home manager won’t see the root nix channel? Didn’t think of that
abbiya has joined #nixos
zupo has joined #nixos
<dminuoso>
hpfr[m]: In some sense that'd be a blessing
Yaniel has quit [Ping timeout: 248 seconds]
Edes has joined #nixos
<Edes>
does anyone know why gnome-shell doesn't let me adjust brightness?
<hpfr[m]>
dminuoso: not from a config standpoint
<Edes>
I tried using xbrightness and the media keys on my keyboard
mexisme has joined #nixos
<dminuoso>
hpfr[m]: Why not? It'd allow for remote impact where, if the root user was someone else, it would reduce the determinism of your own derivations.
<dminuoso>
Since the channel could be sneakily changed.
<dminuoso>
I dont actually know whether it's the case though, mind you.
schneid3306 has quit [Quit: ZZZzzz…]
<hedgie>
is talk about nix on $fruitOS completely OT in this channel?
abbiya has quit [Quit: abbiya]
endformationage has quit [Quit: WeeChat 2.5]
hyper_ch2 has joined #nixos
<hpfr[m]>
dminuoso: That’s true I suppose
Ariakenom has joined #nixos
ashesham` has joined #nixos
ashesham` has quit [Remote host closed the connection]
<DigitalKiwi>
hedgie: i'm not an op but i think it's fine
<jackdk>
hedgie: if you mean iMac OS X then I think it's fine
<jackdk>
nix-darwin is a thing
<hedgie>
yeah. well, macOS. I'm trying to get things to build, but for every package I try, I get an error where some part complains that there's an endianness is not defined in the header
<sphalerite>
hedgie: yes, this channel si for nix stuff in general including macos. We do also have #nix-darwin for osx-specific stuff though.
<Thra11>
I've added a remote builder to nix.buildMachines in /etc/nixos/configuration.nix. How can I make it add the builder as a substituter as well?
ng0 has joined #nixos
ng0 has quit [Client Quit]
<srhb>
rsoeldner: Usually if it's just a version bump you can use influxdb = super.influxdb.overrideAttrs (oa: { src = ...; name = ...; }) -- but I see you want at least version and buildFlagsArray here too.
<srhb>
rsoeldner: So yeah, in this case copying the entire thing might be fine..
mvnetbiz_ has joined #nixos
thc202 has joined #nixos
phreedom has quit [Remote host closed the connection]
phreedom has joined #nixos
___laika has quit [Ping timeout: 272 seconds]
___laika has joined #nixos
erik-4l has joined #nixos
<erik-4l>
Hi all, is there a way to tell nixops to disable sandboxing without modifying the deploying machine's nix config? So that it eg. can be adjusted between deployments?
<rsoeldner>
srhb, thank you, unfortunatly after nixos-rebuild switch I do not find the new influxdb version, any idea ?
o1lo01ol1o has joined #nixos
<erik-4l>
The only thing I can find is to add it to nix.conf, but this sets it for each deployment
<erik-4l>
I can't seem to find a way to override this for a specific deployment, any suggestions?
Shoubit has joined #nixos
johanot has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<srhb>
rsoeldner: In the paste you're importing the overlay default.nix, but your overlay is called overlay.nix
<srhb>
rsoeldner: Not sure if that's also the error on your system.
<rsoeldner>
srhb, pasting error, while running nixos-rebuild I do not receive any error\issue
<srhb>
rsoeldner: And you're installing influxdb via some nixos service option?
___laika has quit [Ping timeout: 245 seconds]
<srhb>
rsoeldner: or environment.systemPackages
shandogs has joined #nixos
gratto has quit [Ping timeout: 248 seconds]
___laika has joined #nixos
<srhb>
rsoeldner: overlays defined via that nixos option only take effect for the nixpkgs that nixos uses internally, not for eg. nix-env
<rsoeldner>
srhb I tried using `nix-env -iA nixos.influxdb`
<srhb>
^
gratto has joined #nixos
<srhb>
That'll be it then.
o1lo01ol1o has quit [Remote host closed the connection]
<rsoeldner>
srhb, Ok - What would be the way to allow `nix-env` usage ?
<srhb>
rsoeldner: It's a little messy to set up so you get a consistent experience across all the tools, but it's doable.
<rsoeldner>
srhb, thank you for the help :)
<{^_^}>
[nixpkgs] @FRidh merged pull request #65014 → remove myself since I don't use this anymore, and cannot maintain it. → https://git.io/fj1rP
<{^_^}>
[nixpkgs] @FRidh pushed commit from @s1lvester to master « remove myself since I don't use this anymore, and cannot maintain it. #65009 »: https://git.io/fj1Km
<abbec>
I am trying to use nix copy to achieve caching. However, I get the "dumping very large path" (I think from llvm) which runs out of memory on CircleCI. Are there any workarounds to this behavior?
knupfer has quit [Remote host closed the connection]
___laika has joined #nixos
knupfer has joined #nixos
kiloreux has joined #nixos
<kiloreux>
How can I override ld for a nix expression from a different revision?
chagra has joined #nixos
<Shoubit>
Zer0xp, IRC channel? Yes, you're in it :)
jb55 has quit [Ping timeout: 260 seconds]
<chagra>
if a package has ++ optional line in its definition, how do i make use of that in my nixos configuration? i can't seem to find the answer in the manuals
orivej has joined #nixos
v88m has quit [Ping timeout: 268 seconds]
___laika has quit [Ping timeout: 272 seconds]
<adisbladis>
chagra: That would depend on the bool expression the optional is depending on. Which one in particular are you thinking of?
nakkle has quit [Ping timeout: 264 seconds]
jackdk has joined #nixos
nakkle has joined #nixos
v88m has joined #nixos
knupfer has quit [Ping timeout: 245 seconds]
Neo-- has quit [Ping timeout: 248 seconds]
risson has joined #nixos
_cyris212 has joined #nixos
___laika has joined #nixos
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
cyris212 has quit [Quit: ZNC 1.6.5+deb1+deb9u2 - http://znc.in]
hyper_ch2 has quit [Remote host closed the connection]
mniip has quit [Quit: This page is intentionally left blank.]
mniip has joined #nixos
bakakuna has joined #nixos
<no-n>
which package can I get 'make' from?
<etu>
,locate bin makee
<etu>
,locate bin make
<{^_^}>
Found in packages: gnumake, gnumake42, netbsd.makeMinimal
<etu>
no-n: Probably gnumake
<{^_^}>
Couldn't find in any packages
<talyz>
no-n: gnumake
<no-n>
ahh, thx
<chaker>
no-n: If you're trying to build using nix, I think you don't need to add it to the buildInputs. It's available through the stdenv
___laika has joined #nixos
hyper_ch2 has joined #nixos
<eyJhb>
Is there like any good guides at making services config files? E.g. I want to make a service config for rotate-backups, which accepts a list of derivatives like e.g. acme.certs, and then writes it to a config file, which will run periodically with systemd, crontab, etc.
<eyJhb>
But I don't know where to start. Should I just write a rotate-backups.nix, and use `nix-shell -p..`, noo, can't do that.. How would one go about it!
<eyJhb>
s/!/?/
bakakuna has quit [Quit: bakakuna]
bakakuna has joined #nixos
<chaker>
eyJhb: systemd services?
<no-n>
chaker: ah, right. thx. :)
___laika has quit [Ping timeout: 246 seconds]
ericsagnes has joined #nixos
<eyJhb>
chaker: I need something like our NixOS options, so that I can define e.g. `services.rotate-backups.<dir?>.{daily,weekly,monthly,yearly,prefer-recent,relaxed}`, which would generate a config file, which I can then run peridocally
<eyJhb>
And just wondered if anybody had some good tips/tricks. Like I have no clue how to even write a module, and test it without mangling my whole system
tmaekawa has quit [Quit: tmaekawa]
<chaker>
eyJhb: Well, I don't recall seeing any tips/tricks about this. What I usually do is create a virtualbox backed deployment using nixops and play with the module.
marko[m]1 has left #nixos ["User left"]
<jonge[m]>
hey there. for clang-tooling i need system include folders for libc and so on but also C++ headers. so of course i can use `${stdenv.cc}/nix-support/cc-cflags` and so on, but none of these files contain the C++ headers. those are only located within the g++/clang++ wrapper and i have a hard time extracting these. does anyone have an idea?
<chaker>
If's a simple one ( less chances that it will break my nixos configuration ) I will just add it to my configurations.nix
<tilpner>
eyJhb: Look at existing modules for inspiration, use nix-instantiate or nix eval to test it quickly
Thra11 has joined #nixos
___laika has joined #nixos
<tilpner>
eyJhb: E.g. nix-instantiate '<nixpkgs/nixos>' --eval --strict -A config.boot.zfs.extraPools
<tilpner>
That would evaluate your config just to get that value and then print it
<chaker>
tilpner: That's more to test that syntax is valid, right?
<infinisil>
philipp[m]: Well, the RFC isn't accepted yet, but most people agree on it. And in any case, more options can always be added later, but you can never remove options, so it's good to start small. That RFC argues for having the module focus on the most useful/important options to keep the module reviewable/maintainable
<infinisil>
And the structured config thing (the first part of the RFC) is very much uncontroversial
___laika has quit [Ping timeout: 272 seconds]
<infinisil>
It's definitely the way forward to get away from the stringly typed/hardcoded configs
civodul has joined #nixos
<yorick>
infinisil: oh no, I *really* want that configText exposed as a nixos option
<yorick>
I mean, environment.etc."foo.json".text will work here, but people are going to just put it in an ExecStart="thing ${pkgs.writeText configText}"
<yorick>
infinisil: my use-case is stealing config from nixos when building docker images
<infinisil>
Ah, configText or configFile as a *full* replacement to the other options sounds good
<infinisil>
(I prefer configFile, as this allows you to even specify a file outside of /nix/store, while you can make it a string with `configFile = pkgs.writeText "file" "<the contents>"`
kfound has joined #nixos
<infinisil>
yorick: Hm I see
<yorick>
infinisil: I'd like to get all the nginx config in a docker image, but there's no nginx.configText that allows me to read the generated config
doyougnu has joined #nixos
<infinisil>
yorick: A configFile option would also work for this right?
xelxebar_ has joined #nixos
<yorick>
infinisil: yes
xelxebar has quit [Ping timeout: 260 seconds]
<infinisil>
I'll make a comment on this in the RFC
<yorick>
thanks!
bakakuna has joined #nixos
<tilpner>
Can we have both?
Mateon1 has quit [Quit: Mateon1]
<tilpner>
readFile on configFile seems more expensive than just using configText
Mateon1 has joined #nixos
<tilpner>
But I'm not sure how to make that easy to use
<infinisil>
tilpner: readFile won't always work because it's purely an evaluation time thing
turion has joined #nixos
<tilpner>
infinisil: For the user it's better IMO if there are configFile and configText, with configFile defaulting to a derivation that checks and writes configText to a file
<infinisil>
But reading the config at runtime should work
<turion>
Why is openvpn so out of date?
<tilpner>
infinisil: But I'm unsure how to make that easier to use for module developers
<turion>
We have openvpn-2.4.7, but there should be 2.7 out now
<{^_^}>
[nixpkgs] @TomSmeets opened pull request #65024 → lib/modules.nix: provide detailed error locations → https://git.io/fj11g
virus_dave has joined #nixos
<infinisil>
tilpner: I don't think it's worth dealing with the trouble of a stringly typed option and two options that do almost the same, when a simple `Use configFile = pkgs.writeText "config" "< the config >"` in the option's description should be enough to inform people how to use it
<tilpner>
infinisil: Except that configFile = pkgs.writeText "config" "< the config >" does no checking of the content
<tilpner>
Whereas configText = "..."; could do that
<infinisil>
tilpner: It can, if the checking is done in the options apply
kfound has quit [Ping timeout: 260 seconds]
<tilpner>
Oh, is that possible?
<infinisil>
Yup :)
<tilpner>
That wasn't one of the options I considered, huh
cfricke has quit [Quit: WeeChat 2.5]
<tilpner>
Can you comment with an example of checking in apply?
<tilpner>
(Or point me to one?)
<tilpner>
Another valid usecase is having a large config already in text form
<infinisil>
Hm, I can't think of an example, but it would make sense to include this in the config checking section of the RFC
<srhb>
Is there a simple http store server that just forwards the existing signatures? nix-serve apparently rips them off.
<tilpner>
And then doing configText = builtins.readFile ./their.conf;
<srhb>
I suppose I could just resign everything, but...
<tilpner>
Which should give them content checking, and would not be possible with just configFile (unless apply does magic?(
<infinisil>
tilpner: `configFile = ./their.conf`?
<tilpner>
*)
<tilpner>
infinisil: No checking of the content of their.conf
<mkaito>
does anyone here actually use Hail for deployment? how does the whole "put in profile" thing work? I know you can push a closure and install it in an environment with nix-env, but I'm not sure how to make that get anywhere into, say, a systemd unit's path.
rsoeldner has quit [Remote host closed the connection]
drakonis2 has quit [Ping timeout: 264 seconds]
drakonis_ has joined #nixos
<tilpner>
mkaito: I haven't used hail specifically, but I imagine it uses profiles to allow for rollbacks and history. AFAICT the activation script of a NixOS system derivation would take care of systemd
Drakonis has quit [Ping timeout: 250 seconds]
<{^_^}>
[nixpkgs] @costrouc opened pull request #65026 → [WIP] pythonPackages.apache-airflow: init at 1.10.3 → https://git.io/fj1Mm
<mkaito>
yeah I figure it's similar to a user profile, but that one has an etc/profile.d file that's sourced to set things up. if I `nix-env` something into an empty profile, that's obviously not there. so would I just use the "absolute" path in that profile to access things from a unit, for example?
<mkaito>
system users (or worse, DynamicUser) don't get their own profiles by default, so I'd have to create one myself.
<tilpner>
mkaito: That's not how I interpreted it
cinimod has joined #nixos
<tilpner>
mkaito: My interpretation: you wouldn't ever reference the profile manually (unless you want to do rollbacks)
<mkaito>
so how does the profile end up in the PATH for the service?
<tilpner>
And hail itself will activate the new system configuration
<cinimod>
error: Package ‘atlas-3.10.3’ in /nix/store/3r16s3jrv26l0panr38llhglrm901qwi-nixpkgs-19.03pre154023.89b618771ad/nixpkgs/pkgs/development/libraries/science/math/atlas/default.nix:111 is marked as broken, refusing to evaluate.
bakakuna has quit [Read error: Connection reset by peer]
<cinimod>
But it built with nix-env
bakakuna has joined #nixos
<cinimod>
I suppose I can allowBroken = true
<turion>
I have a local clone of nixpkgs where I've bumped the openssl version in NetworkManager's openvpn. What's the best way to rebuild my system with this running, without changing too much else about my system? Can I somehow say in my config "Get the network manager service from this file"?
<clever>
cinimod: what does `nix-instantiate --find-file nixpkgs` return?
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<clever>
turion: look up disabledModules in the nixos manual
<tilpner>
infinisil: If you already check for paths outside the store, you don't need to ln, I only proposed that because it would handle both cases. For other files, you can cp just as well
<turion>
Wait... instead of hacking around on my own branch of nixpkgs, could I also bump the openvpn version directly in my config somehow?
<turion>
*openssl version
<turion>
Given that openssl_1_1 is packaged, but not selected for openvpn
<infinisil>
tilpner: Ah, I'm thinking that if the file is outside the store, it would get copied to the store already with a "${file}"
<marek>
when and how is staging-19.03 merged into release-19.03?
<cinimod>
But it seems to be using python 2.7 (2!)
<clever>
cinimod: you should probably be using nix-build for testing, not nix-env
<clever>
cinimod: and you want to apply it to the right pythonPackages, there is also a python3Packages
___laika has quit [Ping timeout: 248 seconds]
kiloreux has quit [Quit: Leaving]
doyougnu has quit [Ping timeout: 264 seconds]
<tilpner>
infinisil: That "already gets copied" doesn't seem relevant. Symlinks could cause problems in some cases, so it doesn't matter where you symlink to. Just s/ln -s/cp/ again, it's little benefit for potential pain
<infinisil>
tilpner: I added an edit to the comment suggesting cp instead
<cinimod>
clever: I can nix-env with python37 but the nix-shell still fails with
<cinimod>
error: Package ‘atlas-3.10.3’ in /nix/store/3r16s3jrv26l0panr38llhglrm901qwi-nixpkgs-19.03pre154023.89b618771ad/nixpkgs/pkgs/development/libraries/science/math/atlas/default.nix:111 is marked as broken, refusing to evaluate.
<{^_^}>
[nixpkgs] @costrouc opened pull request #65027 → [WIP] pythonPackages.zarr: init at 2.3.2 → https://git.io/fj1M7
<clever>
cinimod: not sure whats up with that
<cinimod>
clever: is my nix/store screwed somehow? But I thought nix was supposed to avoid such problems
<clever>
cinimod: nope
<clever>
cinimod: what happens if you nix-build '<nixpkgs>' -A python37Packages.myScikits-odes
gratto has quit [Read error: Connection reset by peer]
gratto has joined #nixos
<cinimod>
Same error
<clever>
cinimod: then its nix-env being weird, its likely loading from a diff nixpkgs
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<clever>
cinimod: i prefer testing things with nix-build
doyougnu has joined #nixos
<turion>
clever: If I do the approach with disabledModules, what will be the build inputs to the new module I import? E.g. I set that openvpn in all-packages.nix has openssl_1_1. If I import the networkmanager module from the branch where I did this modification, which version will it pull?
<cinimod>
Ok it fails with my commented out code with nix-build
vmandela has quit [Quit: Leaving]
<clever>
turion: you will need a nixpkgs.config.packageOverrides also, to set `openvpn = (import /path/to/nixpkgs {}).openvpn;`
<clever>
turion: that will force openvpn to get its deps from the 2nd nixpkgs
bakakuna has quit [Quit: bakakuna]
bakakuna has joined #nixos
<turion>
clever: Ok, and where do I set that? In my configuration.nix?
<cinimod>
At least it is consistent now
<clever>
turion: yes
<clever>
cinimod: i would start by looking at the atlas-3.10.3 expr in nixpkgs, to see why its broken
hyper_ch2 has quit [Remote host closed the connection]
hyper_ch2 has joined #nixos
magnetophon has joined #nixos
<cinimod>
:( I am far from confident that with my skill level in nix that I will be able to diagnose the problem
veske has quit [Quit: This computer has gone to sleep]
<cinimod>
How do I know which commit of nixpkgs I am using?
<clever>
cinimod: pkgs/development/libraries/science/math/atlas/default.nix: name = "atlas${optionalString withLapack "-with-lapack"}-${version}";
<clever>
cinimod: nix eval nixpkgs.lib.version
<clever>
cinimod: you can also use `nix-instantiate --find-file nixpkgs` to just find the exact nixpkgs its currently using
matt` has joined #nixos
inquisitiv3 has joined #nixos
___laika has joined #nixos
fusion809 has quit [Remote host closed the connection]
<magnetophon>
Hi, I'm trying to update a rust pkg, but I'm getting: "error: the lock file /build/source/Cargo.lock needs to be updated but --frozen was passed to prevent this" I've found that I need to add "cargoPatches = [ ./Cargo.lock.patch ];", but I haven't found out how to determine the content of the .patch file.
<matt`>
I'm trying to preserve package source files in the nix store under a subdir called `src`. I've tried adding it to a postfixup phase in an overlay (mkdir -p $out/src, cp -r ${self.src} $out/src) but I'm still not getting the source dir even though the build seems to complete correctly. Anyone know of a way to do this?
<tilpner>
Gah! infinisil-- :c
<magnetophon>
infinisil: Thanks! ;)
<infinisil>
Sorry I couldn't resist xD
<tilpner>
magnetophon: Could try building without Nix, cargo might update the file itself. Or try deleting the lock file and building then
bakakuna has quit [Read error: Connection reset by peer]
<magnetophon>
After I hit enter, I saw it coming. Didn't know it was automated though! :)
bakakuna has joined #nixos
<magnetophon>
tilpner: Well, I want to update the pkg, so that doesn't help much...
<tilpner>
magnetophon: Why not? After cargo adjusts the lock file, you can generate a patch to feed to Nix
<clever>
matt`: can you gist the overlay?
___laika has quit [Ping timeout: 245 seconds]
<magnetophon>
tilpner: ah. lemme see if I can find out how to do that. Thanks!
<matt`>
ignore the unpackPhase... shouldn't be there
<clever>
matt`: if you nix-instantiate the drv, then `nix-store --query --binding postFixup /nix/store/foo.drv` does it show the changes?
Drakonis has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
hyper_ch2 has joined #nixos
jb55 has joined #nixos
o1lo01ol1o has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
sauyon has joined #nixos
drakonis_ has quit [Ping timeout: 246 seconds]
virus_dave has quit [Quit: virus_dave]
kfound has quit [Ping timeout: 260 seconds]
drakonis_ has joined #nixos
bakakuna has quit [Quit: bakakuna]
bakakuna has joined #nixos
<cinimod>
clever: I am in nixpkgs/pkgs/development/libraries/science/math and I can't see atlas? I am 2d575e31b8d which was updated 45 minutes ago - was it removed?
<matt`>
clever: it does, but when I follow the src it copies, it's not copying the right one
srid6390803 has quit [Ping timeout: 245 seconds]
<clever>
matt`: you want to use $src, no {
<clever>
matt`: i see, line 8 is doing self.src, and self is line 1, the nixpkgs self
<cinimod>
clever: I have no idea what it going on here
<clever>
cinimod: what does `nix-build '<nixpkgs>' -A atlas` say?
bakakuna has quit [Quit: bakakuna]
drakonis_ has quit [Ping timeout: 252 seconds]
bakakuna has joined #nixos
<cinimod>
nix-build '<nixpkgs>' -A atlas
<cinimod>
error: Package ‘atlas-3.10.3’ in /nix/store/3r16s3jrv26l0panr38llhglrm901qwi-nixpkgs-19.03pre154023.89b618771ad/nixpkgs/pkgs/development/libraries/science/math/atlas/default.nix:111 is marked as broken, refusing to evaluate.
<clever>
cinimod: thats in nixpkgs rev 89b618771ad
<clever>
`sudo -i` then `nix-channel --update` may help
hyper_ch2 has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
<{^_^}>
[nixos-org-configurations] @edolstra pushed 2 commits to master: https://git.io/fj1DA
drakonis1 has joined #nixos
hyper_ch2 has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
johanot has quit [Quit: WeeChat 2.4]
Drakonis has quit [Ping timeout: 276 seconds]
hyper_ch2 has joined #nixos
<cinimod>
nix-shell -I nixpkgs=/Users/dom/nixpkgs -p python37Packages.myScikits-odes -p python37Packages.nose --run "python3" seems to be building
<magnetophon>
tilpner: OK, I built a working binary, and I have a "Cargo.lock". How do I generate a .patch?
hyper_ch2 has quit [Remote host closed the connection]
drakonis_ has quit [Ping timeout: 250 seconds]
<anhduy>
Hi, I try to use overlay to override bundler to newer version with details in this gist. The result is bundle is upgraded to newer 2.0.2 version but other tool like bundix stop working with the error attach in gist. I thought if i override bundler, bundix need to be rebuilt to use new bundler but look like it confuses somehow. I am not sure how to
<manveru>
anhduy: you'll need a newer version of bundix as well
zupo has quit [Client Quit]
seppellll has quit [Ping timeout: 248 seconds]
<manveru>
i guess i didn't push a new gem in a while...
Drakonis has quit [Read error: Connection reset by peer]
Drakonis has joined #nixos
<manveru>
anyway, the versions of your bundler and bundix' bundler don't have to be the same
<manveru>
just run 'bundle lock && bundix' and things should be fine
<tilpner>
magnetophon: git diff should do it
drakonis1 has quit [Ping timeout: 264 seconds]
bakakuna has quit [Read error: Connection reset by peer]
bakakuna has joined #nixos
<anhduy>
@manveru I have tried override bundix in the overlay as well but it still give same problem. `bundle lock && bundix` doesn't work. Main problem is bundix itself coun't run because it couldn't find correct bundle version
<manveru>
well, my advice is to not touch bundix with your overlay
<manveru>
i can maybe push a new version that allows this, but only later today
<manveru>
sorry for the inconvenience
drakonis_ has joined #nixos
<anhduy>
thanks, i am appriciate that. But the original version of mine doesn't modify bundix, so i am not sure why bundix is not working
apaul1729 has joined #nixos
___laika has quit [Ping timeout: 246 seconds]
<magnetophon>
tilpner: hmm. not sure what I'm doing wrong. I cloned a second copy of the repo, and it turns out the Cargo.lock didn't change...
Drakonis has quit [Ping timeout: 246 seconds]
<magnetophon>
tilpner: I did: "diff broot/Cargo.lock broot2/Cargo.lock" and got no output
Zer0xp has quit [Ping timeout: 245 seconds]
<anhduy>
@manveru change overlay from `bundix = super.bundix.override { bundler = bundler; };` to `bundix = super.bundix.override { bundler = super.bundler; };`. I guess because i override bundler so bundix couldn't find it dependencies. So use super fixes it
bakakuna has quit [Ping timeout: 268 seconds]
hyper_ch2 has joined #nixos
nfip^ has quit []
tsrt^ has joined #nixos
ardumont has joined #nixos
nakkle has quit [Ping timeout: 258 seconds]
hyper_ch2 has quit [Remote host closed the connection]
npmccallum has quit [Quit: npmccallum]
Tucky has quit [Quit: WeeChat 2.5]
<anhduy>
actually it doesn't work correctly, bundix still use older bundler version so run `bundle lock && bundix` will fail, even though `bundix --version` will run successful now
tput has joined #nixos
mmlb has joined #nixos
v88m has joined #nixos
jabranham has joined #nixos
pamplemo1sse has joined #nixos
cinimod has quit [Ping timeout: 245 seconds]
iyzsong has quit [Ping timeout: 252 seconds]
meck has quit [Quit: WeeChat 2.5]
meck has joined #nixos
gratto has quit [Read error: Connection reset by peer]
gratto has joined #nixos
hyper_ch2 has joined #nixos
cinimod has joined #nixos
<manveru>
anhduy: what failure?
hyper_ch2 has quit [Remote host closed the connection]
<anhduy>
same error in the gist `/nix/store/phsfraakzj9wssr6qjgjkj4lgnw98k06-ruby-2.5.5/lib/ruby/2.5.0/rubygems/dependency.rb:312:in `to_specs': Could not find 'bundler' (~> 1.11) - did find: [bundler-2.0.2] (Gem::MissingSpecVersionError)`
endformationage has joined #nixos
hyper_ch2 has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
Thra11 has quit [Quit: WeeChat 2.5]
<manveru>
anhduy: well, if you can wait a bit longer i'll get some time to help you :)
<manveru>
still at work in some nasty js mess here :(
hyper_ch2 has joined #nixos
MmeQuignon has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
hyper_ch2 has joined #nixos
erasmas has joined #nixos
xkapastel has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
hyper_ch2 has joined #nixos
magnetophon has quit [Ping timeout: 268 seconds]
ubert has quit [Quit: Leaving]
hyper_ch2 has quit [Remote host closed the connection]
magnetophon has joined #nixos
tqbl has joined #nixos
fragamus has joined #nixos
inquisitiv3 has quit [Ping timeout: 245 seconds]
wildtrees has joined #nixos
jabranham has quit [Remote host closed the connection]
bakakuna has joined #nixos
Yaniel has joined #nixos
<mdash>
manveru: did you ever succeed in packaging an elixir app?
hmpffff has joined #nixos
vmandela has joined #nixos
bakakuna has quit [Read error: Connection reset by peer]
matt` has quit [Remote host closed the connection]
bakakuna has joined #nixos
wfranzini has quit [Remote host closed the connection]
fendor has joined #nixos
inquisitiv3 has joined #nixos
<tqbl>
i'm having a problem with the vimtex plugin being loaded too late. it was installed through vimrcConfig.packages.myVimPackage.start.
<tqbl>
what would be the best way to ensure it's loaded before the default tex ftplugin?
<anhduy>
@manveru no problem. Take your time, i am still be able to work around by using older bundler version at the moment.
<gchristensen>
19.03 is maintained for roughly 7 months. we releaese every 6 months
<carter>
ok
<gchristensen>
we could maintain longer, but would need a lot of community commitment (and probably $$$$$ :))
v88m has joined #nixos
Drakonis has joined #nixos
v88m has quit [Read error: Connection reset by peer]
v88m has joined #nixos
drakonis_ has quit [Ping timeout: 276 seconds]
turion has quit [Quit: Leaving.]
<pie_>
i put a thing in passthru and made the args to mkDerivation rec, because i dont like let expressions
<pie_>
am i bad
<Taneb>
Incorrigible
<pie_>
you cant access the contents of a let expression outside of it, i try to make my code as overridable as possible
<pie_>
*gracefully overridable
<Taneb>
Have you achieved that?
<pie_>
well, im still working on it
<pie_>
i havent done this specific thing yet
v88m has quit [Ping timeout: 248 seconds]
<pie_>
but it probably beats a let expression
<pie_>
on the other hand i dont like rec sets
<eacameron>
I'm trying to splice in ${../.hlint.yaml} in a drv but it says invalid file name... I remember nix paths can't support this or something. Is there a way to rewrite what the store path becomes from a raw path?
<das_j>
pie_: What aspects of nix do you like, then?
<pie_>
and its probably disproportionate to make the whole argument set rec, because that means you have a lot more stuff in scope than necessary
pmiddend has joined #nixos
<pie_>
let expressions > rec set > let expression
<pie_>
im in a bit of a bind :D
<pmiddend>
I'd like to write a derviation for a single text file, basically. I want to augment a Docker container with a file from the filesystem.
<mdash>
pmiddend: you shouldn't have to do anything unusual, evaluating a path in a nix expr copies it into the store
<srhb>
eacameron: builtins.path :)
seppellll has joined #nixos
<srhb>
eacameron: It takes an attrset { path = the-thing; name = "but-with-legal-name"; }
<pie_>
the idea is that rec { passthru = { whatever = ...;}; stuff = dosomething whatever; } lets someone access whatever if they want to do something to `stuff`
<pie_>
i mean if they want to rebuild stuff
<pie_>
you cant change passthru.whatever to change `dosomething whatever`
<pie_>
if you wanted that to be a thing you would need to use something like fix and self: {...} which is overkill
arjen-jonathan has quit [Ping timeout: 246 seconds]
<exarkun>
What's best practice for selecting stable revisions to which to upgrade?
<pie_>
but on the other hand using rec { ... } when you have more than a few attrs, and you dont _need_ to use rec, seems unnecessarily confusing, because then you now have to think about possible references
<carter>
exarkun: me too, wondering that
<pie_>
clever, any input on the above?
<srhb>
exarkun: What do you mean by stable revisions exactly?
<exarkun>
git revisions of a stable channel
<srhb>
exarkun: How about the latest channel release?
<srhb>
exarkun: As in head of the relevant nixpkgs-channel branch.
<pie_>
infinisil, its likely you might end up reviewing some of my code so ... ^
<exarkun>
Doesn't that just reformulate the question to be "... selecting a point in time to upgrade?"
<srhb>
There's never a reason to think that an old commit has somehow matured through aging. :-)
<exarkun>
I suppose I beg to differ
<exarkun>
The more people use a certain commit the more chances there are to find regressions introduced by that commit
<srhb>
But OK, the spectrum is somewhere from: Pick the latest commit from the relevant channel to: run CI on a bunch of NixOS tests that simulate your system of interest.
<srhb>
exarkun: Right, I was merely pointing out that the commit in question won't ever get better..
<srhb>
I'm sort of limiting myself to "what can you automate" here.
<exarkun>
Yes, and I agree with that. But also wanted to point out that time may reveal that a commit in question may be worse than you thought. :)
<srhb>
Absolutely.
<exarkun>
How does the state of cache.nixos.org relate? Is it populated before the channel is updated or after?
<infinisil>
pie_: Well, rec works very similar to a let in
<infinisil>
PRetty much the same actually
<srhb>
exarkun: It's continuously fed all the builds.
<infinisil>
pie_: I don't think there's much harm in using it, it can make things shorter
<pie_>
infinisil, right, i wanted to say they basically work the same, but theres some slightly orthogonal things
<infinisil>
It can be confusing if it shadows bindings though
<exarkun>
But do builds happen before or after the channel updates? :)
<pie_>
hm i didnt even think of shadowing
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<srhb>
exarkun: Before.
<pie_>
i mean ideally you arent shadowing, but see now its a question you have to ask yourself
<exarkun>
cool
<pie_>
"is anything being hsadowed"
<srhb>
exarkun: But not all packages have to succeed, far from it.
<pie_>
infinisil, ill just paste some code
<exarkun>
o
shomodj has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
MmeQuignon has quit [Ping timeout: 248 seconds]
<srhb>
exarkun: Essentially, for the nixos-* channels, the channel is barred from advancing until certain tests pass
<srhb>
exarkun: Those tests transiently pull in a ton of packages
<srhb>
exarkun: But every package _not_ touched by those tests is allowed to fail.
<exarkun>
got it
<infinisil>
pie_: Ah yeah, that is a big scope, I'd probably use a let in there
<infinisil>
Or not
<pie_>
infinisil, the choice seems to be between putting allIdrisDeps in a let expression, where you cant access its value from outside the structure | the alternative is this, but here the scope is big
<exarkun>
I should educate myself about what tests exist, I suppose
MmeQuignon has joined #nixos
<srhb>
exarkun: And which are included in the channel mechanism :)
<infinisil>
pie_: Oh wait, you want to use passthru just to get around having a let in?
<musicmatze>
Does someone in here use spacevim on nixos and has a module written for it? :-)
<pie_>
infinisil, i dont actually expect to use that value anywhere in my code
<infinisil>
pie_: Why export it with passthru then?
<pie_>
infinisil, but everything else can be accessed through a .overrideAttrs yeah?
<exarkun>
I imagine the criteria for what tests exist is "the ones someone felt like writing"? What's the criteria for tests that are included in the stable channel job?
vmandela has left #nixos ["Leaving"]
<pie_>
infinisil, because i get peeved when i want to override something and i cant or have to copy paste a bunch of code
<infinisil>
Hm alright
<pie_>
infinisil, i think ideally "patches" should be "local" and only semantically contain what i want changed - this is me trying to go in that direction
<pie_>
does it make sense at least?
<srhb>
exarkun: Sort of. Most tests that are 1) fast and 2) stable (as in, deterministic) should probably be included imo.
<srhb>
exarkun: Unfortunately the vm test framework is notoriously finicky.
<infinisil>
I guess haskell also exports all haskell deps in a similar way
<infinisil>
Which can be useful
<infinisil>
So sure, why not
<exarkun>
certainly non-deterministic tests should not be included (and in general should be deleted or made deterministic)
<psyanticy>
Hi @lnl7 . is there any reason hasicorp vault was packaged without the UI
<pie_>
infinisil, ok so given that its not fundamentally flawed now im just unhappy the rec has a large scope.
<exarkun>
"Hey computer does my software work?" "YES! NO!" "thanks computer."
<pie_>
infinisil, hm i had another ugly idea, i could just duplicate the references in passthru and avoid the rec
<srhb>
exarkun: Some of our very basic tests (as in, the base of your system) have had a tendency to be nondeterministic but are included because of their importance. We suffer false negatives on channel updates because of that.
<srhb>
exarkun: But yes, it's painful :P
<exarkun>
if you point me at some of those I might fix them (probably not but who knows)
<srhb>
exarkun: I've been out of the loop for a bit, but go visit http://howoldis.herokuapp.com/ for say 14 days, take note of the failing tests on the ones held back. You'll get a feel for it very fast :)
<exarkun>
cool, thanks
<srhb>
exarkun: That is, click the test set of one of the non-advancing ones, say unstable.
gerschtli has joined #nixos
<pie_>
infinisil, so now i can have my access and also no overscoped let. i guess its not perfect, but perfect seems unattainable for the moment so i guess im happy enough. thanks forlooking / rubber duck :D
<LnL7>
psyanticy: I think that I still have a branch with my attempt if you want to give it a shot
<carter>
srhb: i kinda agree with what exarkun is saying: TIME is hard, and validating changes may be a high review process in certain high assurance settings
<carter>
valdiating a update to a stable branch / release for code that handles lots of money or lives for example?
LnL7 is now known as LnL
<gchristensen>
LnL: now you're the right color
<LnL>
yeah :)
<gchristensen>
back to green
<srhb>
carter: I don't disagree. What I've done before is exactly run my own CI. I don't know how to actually, meaningfully, do any kind of vetting of an old commit otherwise.
<arianvp>
Maybe I should add that vault ui stuff to the wiki
<arianvp>
As it keeps getting asked for
o1lo01ol1o has quit [Remote host closed the connection]
npmccallum has joined #nixos
<Thra11>
These curl errors when downloading from cache.nixos.org are very frustrating -_- Most downloads succeed after a few retries, but mesa_drivers is 341MB, so it seems there's no chance of it finishing before an error occurs. (I've got it repeatedly trying again `while [ $? -ne 0 ]`)
MmeQuignon has quit [Ping timeout: 258 seconds]
MmeQuignon has joined #nixos
<srhb>
Thra11: Seems OK here.
<Thra11>
The farthest I've seen it get is 295MB, but mostly it doesn't get beyond 50
<Thra11>
thonkpod: Now that I've successfully run nix-prefetch-url, how do I make it notice the result? (I ran nix-prefetch-url on a different machine and used nix-copy-closure to copy the resultant store path to this machine)
srhb has joined #nixos
<thonkpod>
Thra11: You should just be able to run the build like you were earlier and it will see that the file is in the nix store
judson has joined #nixos
<Thra11>
thonkpod: It doesn't :( It just starts trying to download it again
<asymmetric>
i guess it's not possible, and should be configured manually in a nixos module
<Thra11>
The store path is /nix/store/4sgsya7i10rwfy760slvjky05davrpas-1hafm20qylsll52ysr0jwdpb4yvkx5c5y3m44svz7v89ir9jf2li.nar.xz I'm not sure if the build knows that that will be the result of its download
ThatDocsLady_ has quit [Quit: Leaving]
anhduy has quit [Ping timeout: 260 seconds]
noonien has quit [Quit: Connection closed for inactivity]
<gchristensen>
tilpner: https://nixos.org/nixos/foundation.html you can still send money with SEPA / IBAN (I don't know what these terms mean exactly, so excuse me if I'm using them wrong)
<Thra11>
Maybe I need to get it to unpack the nar.xz
<thonkpod>
Thra11: Did you make sure to use `--include-outputs` in your copy-closure command?
justanotheruser has quit [Ping timeout: 245 seconds]
<Thra11>
thonkpod: I don't think there are any outputs? It's just a single .nar.xz file as far as I can tell.
lunik1 has quit [Quit: :x]
___laika has quit [Ping timeout: 248 seconds]
<thonkpod>
Thra11: Sorry, this goes a bit beyond my knowledge. I need to read more into how nix does this
lunik1 has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Thra11>
I suspect if it was unpacked, the build would see the store paths and not try to build them, but I'm guessing that it doesn't know the exact hash of the .nar.xz download, so it doesn't recognise it...
<Thra11>
I'm just guessing though
Edes has joined #nixos
Edes has quit [Client Quit]
rfold has joined #nixos
<clever>
gchristensen: have you seen my recent stuff with runit and docker?
<Thra11>
clever: <hash> is the hash of the thing we're trying to build/fetch, right?
<clever>
yep
<clever>
the storepath hash, not the sha256 hash
o1lo01ol1o has joined #nixos
<Thra11>
clever: Ok. I'll try that next time (This time I've worked around it by putting the builder on the same nixpkgs version, 'building' it on the builder, then copying the closure --to the machine which requires it)
<clever>
Thra11: you can also just `nix-store -r /nix/store/foo` to download something from the binary cache
<clever>
Thra11: and `nix-copy-closure` is around
<clever>
Thra11: oh, you can also `nix copy --from https://cache.nixos.org --store file:///home/clever/cache /nix/store/foo`
<clever>
to download foo from the cache, directly to the cache dir
<clever>
then tar that up, and move it to another machine, however you can
<clever>
strap the USB stick to a carrier pidgeon!
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
<Thra11>
clever: so is nix-store -r/--realise like nix build, except you request the output, not the derivation which produces the output?
<clever>
Thra11: it will also try to build it, if it cant find it on a cache
<clever>
it just skips the nix expression half of things
<Miyu-chan>
I just used nixos-option, and whoever decided to put `This attribute set contains:` as stderr is amazing, in a non-sarcastic way.
<Miyu-chan>
It's nice that stdout is only contains actual pipable data.
gleber_ has joined #nixos
<clever>
Miyu-chan: nice
pmiddend has joined #nixos
Ariakenom_ has joined #nixos
<Thra11>
clever: If I run nixos-rebuild, get a transient failure (e.g. curl error), can I skip the 'nix expression half of things' and jump straight into fetching/building stuff when I re-try?
o1lo01ol1o has quit [Ping timeout: 245 seconds]
<clever>
Thra11: yeah, just run `nix-store -r /nix/store/foo.drv` on the top-level drv
<clever>
Thra11: nixos-rebuild dry-run will only do the nixexpr half, and print that drv
dsx has joined #nixos
___laika has quit [Ping timeout: 246 seconds]
mmlb has quit [Ping timeout: 245 seconds]
thc202 has quit [Ping timeout: 250 seconds]
<Miyu-chan>
for i in $(nixos-option services.postgresql); do printf '%s\n' "$i"; nixos-option "services.postgresql.$i"; done
<Miyu-chan>
No need to do `tail -n +1`
bakakuna has joined #nixos
<Thra11>
clever: This will save a lot of time on this rather slow machine! Thanks!
<clever>
Thra11: you can even nix-copy-closure a .drv file to another box
<clever>
Thra11: then nix-store -r to build it there, and nix-copy-closure the product back
<gchristensen>
xorAxAx: all the binary packages (.nar and .narinfo files) are signed, so despite the privacy violation, using the binary cache in Kazakhstan is not a risk. However, all is not well: you must not use channels to fetch nixpkgs, as the .tar.xz file fetched for the channel is not signed.
<xorAxAx>
gchristensen, yep, no channels for them
<gchristensen>
xorAxAx: for example: using SSH to fetch the nixpkgs repository
<xorAxAx>
neither git via https
<Miyu-chan>
Any idea on how to run a systemd service on rebuild and not on boot? I'm considering early returning on the systemd process, but how would I detect "first run since boot"?
bakakuna has quit [Ping timeout: 246 seconds]
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
Jackneill has quit [Remote host closed the connection]
<clever>
Miyu-chan: create a file in a tmpfs, which will be lost on reboot
qqlq has joined #nixos
ThatOtherGuy has joined #nixos
<Miyu-chan>
Yeah, /var/run is tmpfs on default, right?
<clever>
Miyu-chan: yep
o1lo01ol1o has joined #nixos
ris has joined #nixos
webster has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Neo-- has quit [Ping timeout: 264 seconds]
<gchristensen>
clever: I did see your docker stuff! looks great
pamplemo1sse has quit [Ping timeout: 246 seconds]
<clever>
gchristensen: main issue that needs solving, is creation of users
gratto has quit [Read error: Connection reset by peer]
<gchristensen>
you can make users no problem
<gchristensen>
just put a etc/passwd in the contents list
<clever>
gchristensen: and there was a recent nixpkgs pr that actually helps, something about if mutableUsers = false;, it can pre-generate passwd
<clever>
gchristensen: already doing that, but generating passwd was a pain
<{^_^}>
[nixpkgs] @svalaskevicius opened pull request #65036 → electron: fix electron 5.x crash when using file chooser dialog → https://git.io/fj1Qw
___laika has quit [Ping timeout: 244 seconds]
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
selfsymmetric-pa has joined #nixos
ambro718 has joined #nixos
shandogs has quit [Remote host closed the connection]
Chiliparrot has joined #nixos
anhduy has joined #nixos
o1lo01ol1o has quit [Ping timeout: 245 seconds]
jakkn has joined #nixos
ThatOtherGuy is now known as LogicAside
___laika has joined #nixos
seppellll has quit [Ping timeout: 248 seconds]
<mightybyte>
If I have two different derivations defined in default.nix that I build with "nix-build -A foo" and "nix-build -A bar", how can I create another nix expression both.nix such that "nix-build both.nix" builds both of them?
<mightybyte>
I know that "(import ./. {}).foo" is an expression that builds one of them.
___laika has quit [Ping timeout: 268 seconds]
<tilpner>
gchristensen: I saw that, but presumed it would mean the same administrative overhead that justied the switch to OC
<gchristensen>
tilpner: I think it will be somewhat less
detran has joined #nixos
<tilpner>
Oh, why so?
* tilpner
means a time overhead, not money
pamplemo1sse has joined #nixos
<gchristensen>
well the vast majority will be through O.C. so we'll be able to do almost all of it automatically.
<gchristensen>
the remaining few (I think we only got like a dozen over iban this year?) will have a dedicated place to put them
<gchristensen>
instead of hving to start transparency from scratch
<jakkn>
Hello, I have trouble getting bluetooth working on a Lenovo Yoga 710 and I need some help to debug further. bluetoothctl reports "No default controller available". I have followed the wiki page on bluetooth and enabled the modules listed, but something's not right. dmesg | ack 'Blue|Firmware' reports this https://termbin.com/mb2p
srid6390803 has quit [Ping timeout: 244 seconds]
anhduy has quit [Ping timeout: 260 seconds]
<eraserhd>
Can overlays access `config`?
<colemickens>
don't want to sound ungrateful (or /too/ whiney), but I would contribute to the wiki more if I didn't have to write wiki syntax. At the very least it could have a better set of buttons or quick syntax tips on the page. :/
<tilpner>
eraserhd: Not really
<tilpner>
gchristensen: If you could decide between less financial overhead (SEPA, I think) or less time overhead (OC), which would you prefer I use? It's just a few cents one hand, and an unknown amount of paperwork on the other
<gchristensen>
tilpner: in general we'd prefer to have as much go through OC as possible, although if someone wanted to make a significant contribution, we might want to discuss it first
xelxebar_ has quit [Remote host closed the connection]
<tilpner>
gchristensen: No, it's not significant. Will there be a breakdown of operating costs on OC?
xelxebar has joined #nixos
<gchristensen>
the plan is to post expenses there, yeah
<pie_>
infinisil, so yeah continuing here, i think it might be nice to have this in lib: wrapPhase = phaseName: phase: "runHook pre${phaseName}\n${phase}\npost${phaseName}"
<pie_>
except maybe call it wrapHook or hookWrap or something
<tilpner>
Awesome! Anything is better than the outdated-by-years graph on the old foundation page :)
<gchristensen>
however, historically the foundation has not received a lot of contributions, and doesn't have a lot of expenses -- as key infrastructure is contributed directly - like the s3 bucket, the fastly cdn, and the packet.com builders
<gchristensen>
historically, major expenses have been things like mac builders
<infinisil>
pie_: After another thought, I'm not sure if it's a good idea because the phases stuff is already confusing enough
<infinisil>
What runs when
<infinisil>
What phases are there
<infinisil>
And with this there's another indirection
MmeQuignon has quit [Ping timeout: 244 seconds]
<gchristensen>
I think that (but don't quote me, as I am not on the board of the foundation) by having OC, we will be able to spend money a bit easier on interesting projects
psyanticy has quit [Quit: Connection closed for inactivity]
<pie_>
infinisil, i _could_ argue this would improve on that given that people dont add this stuff half the time and now its just a single prefix call
<tilpner>
gchristensen: I'm sure a purpose for any leftover money can be found. If nothing else, you can always buy/rent hardware to expand ofborg power
<pie_>
infinisil, i think it would have been nicer to have the pre/post wrappers implicit but this is how stdenv ended up...
<gchristensen>
true :)
<infinisil>
pie_: Hm but when is wrapPhase actually needed?
<pie_>
infinisil, but yeah...unsure. i basically always use this
jakkn has quit [Remote host closed the connection]
<pie_>
infinisil, for what meaning of needed?
MmeQuignon has joined #nixos
inquisitiv3 has joined #nixos
<infinisil>
pie_: It seems that these are only needed for having a "base" layer of phases in mkDerivation
<infinisil>
for things like buildFooPackage
jakkn has joined #nixos
<pie_>
well, what im _really_ trying to argue is that hacing it would improve quality of life, but im open to being convinced that it would not
<pie_>
infinisil, any time you override a phase
<pie_>
overriding a phase means you lose the pre/post wrappers
<infinisil>
Well, with package builds you usually don't want to override *the* phase
<pie_>
i think people override phases "all the time" for *handwave* various reasons
<infinisil>
It's really only for like, different build systems where you need to
<pie_>
should i do a line count on pre* post* in nixpkgs?
<infinisil>
pie_: The standard mkDerivation (and buildPythonPackage, etc.) are designed so you shouldn't ever need to override the phases
<pie_>
i think it would be fine for that too though
<pie_>
i figure pre/post hooks are also kind of cluttery, what i want to see is "this is wrapped in the standard hooks"
<infinisil>
Okay let's maybe look at what derivations do override phases to see what they override it for
mmlb has joined #nixos
<pie_>
to be clear i think "oh no moar indirection" inst an unfair criticism
<pie_>
*isnt
seppellll has joined #nixos
<pie_>
* its easier to grep for runhook
<infinisil>
I think the real clue is that the main phases seem to be build-system dependent. So for a standard C project it's always going to be `make` as the buildPhase
<infinisil>
(plus the hooks)
<infinisil>
For rust projects `cargo build` or so
<infinisil>
Additional requirements can be had with `makeFlags`
<infinisil>
Or pre/postBuild
<pie_>
(685 lined for grep -ir "runhook" | wc -l , which is mor than what would be relevant)
<infinisil>
Hm..
<pie_>
infinisil, i dont follow
<infinisil>
Like, ideally we have abstractions for all build systems which set the phases. And ideally no package would need additional settings, just `buildFooProject` and done
<infinisil>
This is the case for mkDerivation when it's a standard C build with `make` and `make install` for example
<infinisil>
No need to set anything except src and name attributes
<pie_>
i think wrapHook would be fine for internal stuff
<pie_>
i dont see any difference for using it for internal and more leaf-like stuff
<infinisil>
Okay well I was going to say some more but whatever
bakakuna has joined #nixos
<pie_>
please do?
o1lo01ol1o has joined #nixos
<infinisil>
What I just realized: Whether you use wrapHooks or put the runHooks there manually, people still need to remember to do it
<pie_>
sure
<infinisil>
And the difference between wrapHooks and the two runHook lines is very small and not worth it imo
aristid has quit [Quit: WeeChat 2.4]
<infinisil>
What I'd rather like is to have abstractions that automatically add runHook's when needed
___laika has joined #nixos
<pie_>
id like that too but we dont have it :(
<infinisil>
Yeah, but I still don't think it's worth having such a function that saves like 1 line and makes people have to look up its definition to see what it's doing
<infinisil>
Hm yeah so I still don't think it's worth it. Imo time is better spent thinking about when the runHook's are needed, or why can't we just add them by default unconditionally (and transparently)
<pie_>
well..i dont have the willpower right now to start crusading for that :(
<infinisil>
pie_: The first one :) I think it's fine to use locally tbh
<infinisil>
Eh but I guess over all it doesn't matter much
<pie_>
infinisil, well if you like the first one better then? :P
<infinisil>
If you made a PR to add wrapHook I probably wouldn't oppose it, but I wouldn't agree with it either
<pie_>
reasonable
<infinisil>
(like a lib.wrapHook)
<pie_>
people seem generally very reluctant to add stuff to lib though
<pie_>
ok well ill add it to my pr bucket list
<infinisil>
Hm yeah..
<infinisil>
I guess it makes sense because functions in lib are supposed to be general and useful for a lot of things
<infinisil>
So they need to carry their weight
<pie_>
sure
<infinisil>
Whereas packages are just single-use, nobody cares much about a package having a bit of a wrong definition
<pie_>
to be fair i guess i do think ill actually get mapDir in (whatever it was renamed to last time)
<pie_>
even that got a bit of pushback though and yet later its already part all of the other two function in that file, just hasnt been factored out :P
<infinisil>
I hope to see some more plentiful nix libraries with Flakes :D
<pie_>
im writing nix libraries all the time, i just think noone will use them unless its in nixpkgs :p
<pie_>
though i do write rather complicated stuff...
<infinisil>
It is kind of a pain to use external nix projects without something like flakes tbh..
<infinisil>
Pinning manually and such
<pie_>
yea, but also unaudited code
<gchristensen>
yup
inquisitiv3 has quit [Ping timeout: 244 seconds]
<infinisil>
Well, if we can somehow get --pure-eval practical, that shouldn't be a worry
<pie_>
how does that help
<gchristensen>
the current flakes RFC forces pure evaluation
<infinisil>
Ohh nice
<gchristensen>
indeed :)
<pie_>
ok i guess it limits you to running code with user interaction on the produced result but ehh
<infinisil>
pie_: I was thinking that with --pure-eval nobody can read your files, because they'd need the sha256 for that. Whereas without it they can just do `curl PUT ${/home/infinisil/.ssh/id_rsa}` during a build
<pie_>
i could probably just dynamically construct a src or something and use that to exfiltrate data
<pie_>
oh
<pie_>
right we an run bash and stuff
<pie_>
i was thining of something else
<pie_>
thinking
<gchristensen>
infinisil: I bet you could make that evaluation-time too
<wildtrees>
any guides to getting bspwm working on nixos?
<infinisil>
gchristensen: How so?
<wildtrees>
I think I have it started but its not responding, so more of like how to configure bspwm on nixos
<pie_>
i dont understand why thinking about this is useful though?
<infinisil>
alexarice[m]: If it's an error I'd suggest posting a comment there
<gchristensen>
what do you mean?
<pie_>
re: just make the generated executables do bad things?
<alexarice[m]>
infinisil: said error but it seemed to keep on running so who knows. I'll take a look at the link, thanks
<gchristensen>
pie_: because evaluation should be safe
inquisitiv3 has joined #nixos
<infinisil>
We should make Nix wrap all executables in a sandbox/VM xD
<pie_>
^
<pie_>
moar selinux or whatever
nekopep has joined #nixos
<gchristensen>
like snaps
<nekopep>
hi there
<gchristensen>
you wouldn't want every executable in a sandbox/vm
<gchristensen>
hey nekopep !
<colemickens>
I keep thinking about if it would be possible to reuse xdg-portals and other flatpak tech, while still retaining the build-everything-from-source/drvs nature of NixOS.
<clever>
worldofpeace: still bisecting, lol
<colemickens>
but also I use like 3 desktop apps, so it might not matter.
<clever>
worldofpeace: down to 13739 commits
<gchristensen>
colemickens: have you seen the snapcraft work?
<nekopep>
I m a nix french nix noob seeking for a little bit of help :-)
<colemickens>
gchristensen: no, I've sort of avoided it because it seems to be another one of Canonicals antics (the snap server is closed source, you can only use one snap server [so I've heard] at a time)
<colemickens>
is it good? (or better than flatpaks?(
<gchristensen>
nekopep: if you need a French-language channel, #nixos-fr exists (but is small) :)
<nekopep>
in fact i m very enjoyed so far BUT i just got a #$%^$%^ azerty keyboard issue under KDE ....
<nekopep>
whatever i ve tryed so far i still am on a qwerty keyboard. I must miss something...
<worldofpeace>
clever: bisects like this are usually aided by an intevenous drip and maybe a mechanical arm :P
<nekopep>
for such a easy issue
<gchristensen>
nekopep: have you tried setting i18n.consoleKeymap = "azerty"; and services.xserver.layout = "azerty"; then run "nixos-rebuild switch" an log out and then back in? (I think azerty is the name here)
inquisitiv3 has quit [Remote host closed the connection]
bakakuna has quit [Read error: Connection reset by peer]
inquisitiv3 has joined #nixos
___laika has joined #nixos
Drakonis has quit [Ping timeout: 276 seconds]
selfsymmetric-pa has quit [Remote host closed the connection]
selfsymmetric-pa has joined #nixos
___laika has quit [Ping timeout: 246 seconds]
___laika has joined #nixos
pmiddend has quit [Ping timeout: 276 seconds]
fragamus has quit [Read error: Connection reset by peer]
psyanticy has joined #nixos
Jackneill has joined #nixos
nekopep has quit [Remote host closed the connection]
<psyanticy>
@arianvp please do add the vault ui stuff u mentioned to the wiki it would be helpful.
Drakonis has joined #nixos
drakonis_ has joined #nixos
<gchristensen>
anyone tried using virtualbox from within a nixos container? I'm getting annoying errors like Stderr: VBoxManage: error: The virtual machine 'tmpRBDvDLKgT6_default_1563482973306_60584' has terminated unexpectedly during startup with exit code 1 (0x1) -- VBoxManage: error: Details: code NS_ERROR_FAILURE (0x80004005), component MachineWrap, interface IMachine
<flavio_>
I believe i found a typo in nixos options... there is one option vor virtualisation written like virtualization, but it is the only one... all others are written virtualisation
cinimod has joined #nixos
<flavio_>
it's necessary to open a bug report? don't know if this is really a bug
<gchristensen>
actually, not a problem. here is what I have: systemd.services.buildkite-agent-1 = lib.mkAliasDefinitions { definitions = lib.catAttrs "buildkite-agent" options.services.buildkite-agent.definitions; }; --- but it doesn't end up duplicating the service
MmeQuignon has quit [Ping timeout: 246 seconds]
Shouou has quit [Ping timeout: 250 seconds]
<gchristensen>
ah. systemd.services
<gchristensen>
no luck.
MmeQuignon has joined #nixos
<{^_^}>
[nixpkgs] @matthewbauer pushed 0 commits to add-emacs-zmq: https://git.io/fj1d0
<infinisil>
Ahhh
<infinisil>
Um hold on
rfold has quit [Ping timeout: 245 seconds]
<viric>
gchristensen: it was about the lid, that the computer fell aslept
<viric>
gchristensen: logind changes in configration.nix
<infinisil>
(this is the same as above, but mkAliasDefinitions inlined)
<gchristensen>
boy, I am glad to have you around :)
<gchristensen>
(testing)
hmpffff has quit [Quit: Bye…]
<gchristensen>
the following new units were started: buildkite-agent-2.service, buildkite-agent-3.service, buildkite-agent-4.service, buildkite-agent-5.service, buildkite-agent-6.service, buildkite-agent-7.service, buildkite-agent-8.service, is-nix-channel-up-to-date.service
<gchristensen>
also, these tests are intended to be descriptive not prescriptive
<flavio_>
the only thing i noticed it is it doesn't show when you query options for configuration... the "offending" config option is security.virtualization.flushL1DataCache ... No other option appears when searching for virtualization, but a lot appears when searching for virtualisation
<gchristensen>
interesting, yeah, virtualisation should be used in nixos modules
<flavio_>
i think people will find it because they will search for security first, but the names should be uniform imho
<flavio_>
think its just a typo but dont know if needed to open a bug report
MmeQuignon has quit [Ping timeout: 245 seconds]
<gchristensen>
please do open a bug report :)
<gchristensen>
infinisil: a fun thing about this table is it shows browsers have improved significantly in the last year
<gchristensen>
last August, chrome crashed under its weight, and firefox barely managed.
Diagon has quit [Quit: Leaving]
mbrgm_ has joined #nixos
mbrgm has quit [Ping timeout: 248 seconds]
mbrgm_ is now known as mbrgm
fusion809 has joined #nixos
Garbanzo has joined #nixos
Garbanzo has quit [Remote host closed the connection]
Garbanzo has joined #nixos
<infinisil>
Damn
<{^_^}>
[nix] @grahamc pushed 3 commits to docs-cores-max-jobs: https://git.io/fj1bE
<{^_^}>
[nix] @grahamc pushed 0 commits to docs-cores-max-jobs: https://git.io/fj1bu