<Boarders>
I guess I need a nix file which gives me the environment variable that points at the conf rather than the one that is getting defaulted to
noudle has quit []
<Boarders>
that command and the related command I tried: ghci -package-db [path-to-nix-package.conf]
<Boarders>
No Module named System.IO is imported.
energizer` has joined #nixos
<Boarders>
and ghci doesn't even load
vykook has joined #nixos
Fare has quit [Ping timeout: 240 seconds]
pnobre has joined #nixos
boxscape has quit [Remote host closed the connection]
bvdw has quit [Ping timeout: 265 seconds]
Fare has joined #nixos
mbrgm has quit [Ping timeout: 240 seconds]
dsx has quit [Quit: dsx]
mbrgm has joined #nixos
<{^_^}>
[nixpkgs] @thoughtpolice pushed 2 commits to master: https://git.io/JvujY
<colemickens>
I was able to type that nix-shell line, and the ghci -> import Control.Lens worked fine
Synthetica has quit [Quit: Connection closed for inactivity]
vykook has quit [Ping timeout: 265 seconds]
<colemickens>
Not that it's super helpful to you, but even stranger that it's not working. Maybe you got nixpkgs-unstable and something is wonky there?
* colemickens
is stuck in a haskell repl now, this is fun
Boarders has quit [Remote host closed the connection]
mexisme_ has joined #nixos
Rusty1 has quit [Remote host closed the connection]
<evils>
i'm trying to package something that instead of shipping a big file, downloads it at build time (via a shell script), how can i run this script with network access?
<mjrosenb>
for some reason, that machine says that super doesn't have an attribute 'adoptopenjdk-bin', despite the fact that I have installed it with nix
jerryaldrichiii_ has quit [Ping timeout: 260 seconds]
drakonis has joined #nixos
Supersonic112 has joined #nixos
Supersonic has quit [Disconnected by services]
Supersonic112 is now known as Supersonic
<mjrosenb>
somewhat unrelatedly, how do I prevent nix-env -u from attempting to install python3-3.9.0a3? it fails to build, and I don't actually want a version of python that is that new.
<evanjs>
clever: so what’s your first guess when I can’t type into the console on real hardware with that barebones initrd setup? No console directive for the cmdline
vykook has joined #nixos
evils has joined #nixos
lovesegfault has joined #nixos
<lovesegfault>
worldofpeace: I am back
<lovesegfault>
TLP hacking time!
vykook has quit [Ping timeout: 260 seconds]
nuncanada has quit [Quit: Leaving]
Thra11 has quit [Ping timeout: 258 seconds]
Thra11 has joined #nixos
dbmikus has joined #nixos
Fare has quit [Ping timeout: 260 seconds]
vykook has joined #nixos
lovesegfault has quit [Quit: WeeChat 2.7.1]
h0m1 has quit [Ping timeout: 252 seconds]
vykook has quit [Ping timeout: 240 seconds]
h0m1 has joined #nixos
maxdevjs has quit [Remote host closed the connection]
<benley>
I have a oneshot systemd service that's "before" and "requiredBy" several other services that stay running, and I was hoping that when the oneshot service unit changes, it gets started (or restarted?) so the services that depend on it also get updated
<benley>
but that's not happening
cyris212 has joined #nixos
<benley>
(as part of the nixos activation script, that is)
<lovesegfault>
benley: does the unit have RemainAfterExit=yes ?
<benley>
nope, it does not.
* benley
looks up what that does exactly
vykook has quit [Ping timeout: 255 seconds]
<benley>
oh interesting, that might just do it. Let's find out
<benley>
lovesegfault: the only gotcha I noticed is that switching from the config prior to adding RemainAfterExit to the one with it does not cause the oneshot unit to get started
<lovesegfault>
Yeah, there's no way around that I think
mexisme_ has quit [Ping timeout: 258 seconds]
<benley>
but if I start it once (which would happen when any of the downstream services got started/restarted, or when it's first deployed) it works as I was hoping
<benley>
so I can live with that
<lovesegfault>
maybe if you enable reloadIfChanged?
<benley>
what's that do differently than restartIfChanged? (aside from calling systemd reload, of course)
<benley>
err, systemctl reload
<lovesegfault>
I'm taking the flip-knobs-till-it-works approach here
<benley>
haha ok
<benley>
I'll give it a try
<benley>
that's what I usually end up doing when it comes to systemd unit options
<evils>
i don't suppose anyone knows how to add a package to pkgs.vmTools.diskImageFuns image? getting `cannot touch '/mnt/var/lib/dpkg/status': No such file or directory` if i just pass `packages = [ "hello" ]` in
Heirlung has quit [Ping timeout: 240 seconds]
<drakonis>
energizer`: you declare it
Heirlung has joined #nixos
<energizer`>
drakonis: is there any documentation about this?
<drakonis>
not a lot of it atm
fresheyeball has joined #nixos
<drakonis>
there's the flake rfc, the repo one guy cobbled in order to create an reference configuration
<drakonis>
the code for the flakes branch might not be very useful for you
<energizer`>
alright. i guess it's still new so i can wait :-)
<energizer`>
my other question is how do i install a python package from github? pypi2nix seems to only support pypi (not github). I can do fetchFromGitHub but that doesn't pull any of the dependencies. I could package them all manually I guess, but that's too much work. So I'm using virtualenv, which gets the job done but perhaps there's a more nixy way?
<evils>
o hi clever, i'm finally getting around to using runCommand as you recommended 4 months ago..., my script is trying to create a temporary file and i'm getting permission denied...
<lovesegfault>
clever: Oh, thank the lord you are here
<CMCDragonkai>
found lib.fileContents
<lovesegfault>
clever: How the hell to I get my package to set up a dir in `/var/lib`
AmandaC has quit [Ping timeout: 272 seconds]
<lovesegfault>
`tlp` needs `/var/lib/tlp` to exist
vykook has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @c00w opened pull request #81159 → [go] Undo some 1.12 overrides that now build with 1.13 → https://git.io/JvzTU
<evanjsx[m]>
<clever "evanjs: usb keyboard, drivers pr"> Usb keyboard was missing but it’s there now. Come to think of it though, it didn’t seem to work in qemu graphical either but I’ll have to check tomorrow
vykook has joined #nixos
leotaku has joined #nixos
<clever>
lovesegfault: preStart on the systemd service
<mjrosenb>
so, ghcjs, anyone want to take a crack at it?
colemickens_ has quit [Quit: Connection closed for inactivity]
chagra_ has joined #nixos
chagra has quit [Ping timeout: 240 seconds]
endformationage has quit [Ping timeout: 255 seconds]
fresheyeball has joined #nixos
<fresheyeball>
anyone know how nixpkgs get's it's copy of the "nix" package?
<fresheyeball>
as in nix.package option pkgs.nix
<fresheyeball>
I still can't figure out how to update my nix.package to a specific version from the repo
<clever>
fresheyeball: why not just set nix.package in a module?
<fresheyeball>
I tried setting nix.package in my configuration.nix to the latest rev from nixpkgs, which worked, but didn't include the update I need
<lovesegfault>
clever: That's a solution, but the pkg itself, regardless of systemd unit, can't work without `/var/lib/tlp/`
<fresheyeball>
clever: what do you mean?
<lovesegfault>
Like, this should be done by the pkg install, not the systemd unit
<clever>
fresheyeball: how did it not work when you set nix.package?
<clever>
lovesegfault: there is no way for the installing of a thing, to make stuff outside of /nix/store
<clever>
lovesegfault: look into things like programs.screen, that is a module that will pre-create things (like a config file) and install the package
<fresheyeball>
but that errors out with: while evaluating the attribute 'nix.package.out' at undefined position:
<risson>
Hi! I'm working on a PR to update yubioath-desktop from 4.3.6 to 5.0.2. It seems the maintainer isn't updating that package anymore
<lovesegfault>
clever: meh, I will just do the systemd thing for now
<clever>
fresheyeball: what is the actual error, one line away from what you pasted?
<risson>
Is it good practice to submit my PR, ping the maintainer and suggest I take over?
mexisme_ has joined #nixos
<fresheyeball>
while evaluating anonymous function at /nix/var/nix/profiles/per-user/root/channels/nixos/lib/modules.nix:75:45, called from undefined position:
<clever>
fresheyeball: release.nix contains a function, so `import /path/to/release.nix` is of type function
<clever>
fresheyeball: you must call it with something like {}
<fresheyeball>
well I'm a dumbass
<fresheyeball>
omg
<fresheyeball>
thanks clever
<cole-h>
risson: If you want to take over, I believe the way to do it is just add yourself as a maintainer. No need to remove them.
<cole-h>
(and don't forget to add yourself to `maintainers/maintainers-list.nix`)
<lovesegfault>
clever: testing, brb
lovesegfault has quit [Quit: WeeChat 2.7.1]
chagra_ has quit [Ping timeout: 258 seconds]
<mjrosenb>
ahh, I believe the issue is in fact in the nixpkg tree. haskell-packages.nix has a definition for entropy which includes its dependencies for *normal* ghc, but it has different deps when being built under ghcjs
<risson>
cole-h: great! thanks for the answer :D
<clever>
mjrosenb: thats a limitation of cabal2nix, it cant process the if statements in the cabal file fully
<cole-h>
risson: If that's not the best course of action, one of the people in-the-know will let you know in your PR :P
<clever>
mjrosenb: you have to run cabal2nix --ghcjs, to tell it to assume its a ghcjs build
<mjrosenb>
right, but I never ran cabal2nix, I can only guess someone ran it, and committed all of that to the tree.
<clever>
mjrosenb: yep, but they ran it in the non-ghcjs mode
<clever>
mjrosenb: the only real solution is to re-run cabal2nix again, with --ghcjs, and callPackage it in as part of an overlay
<cole-h>
risson: Best practice is to have your modification to `maintainers/maintainer-list.nix` be a separate commit with message `maintainers: add risson`, and then have everything else in your update commit
<cole-h>
(including adding yourself to the maintainers list in that specific package)
<risson>
cole-h: i'll do that then
<cole-h>
Log would then be `maintainers: add risson`, followed by `yubioath-desktop: 4.3.6 -> 5.0.2`
lovesegfault has quit [Ping timeout: 272 seconds]
<cole-h>
also, it's not `long-description`
<cole-h>
It's `longDescription`
<cole-h>
(line 52)
lovesegfault has joined #nixos
<risson>
cole-h damn it, forgot to modify it, thanks for the review
zupo has joined #nixos
<cole-h>
risson: Got another one coming (leaving it on GH itself this time)
<aiverson>
I'm trying to use pipewire as my audio system, but with only an "enable" option and sparse documentation, I'm not sure how to get it working when just enabling it doesn't produce a working audio system.
bvdw has quit [Quit: bvdw]
bvdw has joined #nixos
vk3wtf has quit [Quit: WeeChat 2.4]
vk3wtf has joined #nixos
logzet has joined #nixos
vk3wtf has quit [Client Quit]
vk3wtf has joined #nixos
m0rphism has joined #nixos
drakonis has quit [Quit: WeeChat 2.7.1]
aiverson has quit [Remote host closed the connection]
<immae>
If you put just ./default.nix then you’ll get a derivation containing the content of default.nix, which is not what you want: you want the derivation contained in that file :)
Thra11 has joined #nixos
<energizer`>
ah
hmpffff has joined #nixos
gagbo has joined #nixos
klntsky has quit [Remote host closed the connection]
<genesis>
ho activate a dbus service file in a derivation ?
<immae>
energizer: you modified your default.nix file, which is now a function (taking an attrset with an element "pkgs"). When you "import ./default.nix" you now obtain a function, which you call with {inherit pkgs;}
<immae>
does it make sense to you?
o1lo01ol1o has joined #nixos
<energizer`>
i mean to ask, what is the difference between import and callPackage in this context
<immae>
"import foo.nix" merely imports the file, so "import foo.nix {something}" calls the function with {something}. `callPackage foo.nix {something}` will import the file but it will also prefill all the arguments which it can import by itself. In your case, `import ./default.nix {inherit pkgs;}` is equivalent to `callPackage ./default.nix {}` since callPackage will fill the "pkgs" attribute himself
<clever>
energizer`: callPackage will inspect the arguments the file wants, and pull the answers from pkgs automatically
<clever>
energizer`: so if your file begins with { stdenv }: then it gets a copy of pkgs.stdenv
<manveru>
i'm about to try vault, and thought this might come in handy, just wanted to know any gotchas :)
afics has joined #nixos
veske2 has quit [Quit: This computer has gone to sleep]
veske has quit [Quit: This computer has gone to sleep]
linarcx has quit [Quit: WeeChat 2.7.1]
o1lo01ol1o has joined #nixos
sigmundv has joined #nixos
erictapen has joined #nixos
<{^_^}>
[nixpkgs] @kampka opened pull request #81170 → kops_1_16: init at 1.16.0 → https://git.io/JvzYn
markus1189 has joined #nixos
pmyjavec has joined #nixos
cosimone has joined #nixos
<pmyjavec>
Hey, I'm trying to submit a package to nix-pkg. Before it's accepted, I wanted to install it. When I install it I get the error :error: cannot auto-call a function that has an argument without a default value ('lib')
<pmyjavec>
I'm so lost as to why this happens, what it means and how I can fix it.
<pmyjavec>
Any clues? I'm running it with: nix-env -f pkgs/top-level/all-packages.nix -i aws-mfa
<__monty__>
pmyjavec: Sounds like you're not using callPackage.
maxdevjs has quit [Remote host closed the connection]
zupo_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
maxdevjs has joined #nixos
zupo has joined #nixos
max_devjs has joined #nixos
max_devjs has quit [Remote host closed the connection]
hmpffff_ has quit [Read error: Connection reset by peer]
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
teto has joined #nixos
opticnerve has quit [Quit: Leaving]
TCZ has left #nixos ["Leaving"]
hmpffff has joined #nixos
o1lo01ol1o has quit [Ping timeout: 240 seconds]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
slyfox has quit [Ping timeout: 265 seconds]
slyfox has joined #nixos
<coderobe>
>/nix/store/qwlwalmadpd4zflxczd34nlfjah9ddgg-grub-2.04/sbin/grub-install: warning: your core.img is unusually large. It won't fit in the embedding area.
<coderobe>
How do i, uh, reduce the size of my grub core.img?
revtintin has joined #nixos
o1lo01ol1o has joined #nixos
<{^_^}>
[nixpkgs] @bignaux opened pull request #81174 → [wip] indicator-cpufreq: init at bzr 98 → https://git.io/JvzsJ
<coderobe>
clever: neither, it's embedding into a btrfs device without partition table
<clever>
coderobe: ah, i would just switch over to GPT, and make a bios boot partition, no fs, not formatted, not mounted
<clever>
coderobe: the bios boot partition must be big enough to hold core.img plus any grub modules needed to mount whatever /boot is
<coderobe>
i do not have spare space, the entire drive is owned by btrfs
<coderobe>
:P
<Shell>
clever: relevant quote from Arch wiki, GRUB can boot Btrfs partitions, however the module may be larger than other file systems. And the core.img file made by grub-install may not fit in the first 63 sectors (31.5KiB) of the drive between the MBR and the first partition. Up-to-date partitioning tools such as fdisk and gdisk avoid this issue by offsetting the first partition by roughly 1MiB or 2MiB.
<coderobe>
note that this used to work until i just rebuilt, i'm trying to figure out what changed
cfricke has joined #nixos
linarcx has quit [Ping timeout: 260 seconds]
iyzsong has quit [Ping timeout: 245 seconds]
clkamp_ has joined #nixos
<coderobe>
rolled back to 19.09 and everything works again...
<enamax>
Buon giorno! How would I go about using a specific (newer) version of a kernel module (specifically ENA)? I know about Developing Kernel Modules from https://nixos.wiki/wiki/Linux_kernel, and I can build the module, how do I "install" it?
rogue_koder_ has quit [Ping timeout: 258 seconds]
erictapen has quit [Ping timeout: 255 seconds]
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « ghc-8.10.x: add overrides to fix regex-* packages »: https://git.io/JvzZU
rogue_koder_ has joined #nixos
zupo has joined #nixos
Arahael has quit [Ping timeout: 258 seconds]
<tilpner>
enamax: Put it into boot.extraModulePackages
<{^_^}>
[nixpkgs] @stigtsp opened pull request #81180 → perlPackages.TextLorem: init at 0.3 → https://git.io/JvzZ3
<tilpner>
At least, that's what worked for a module I needed, that didn't ship with the kernel at all
<tilpner>
enamax: You generally want overrideAttrs instead of overrideDerivation
LysergicDreams has quit [Ping timeout: 265 seconds]
<{^_^}>
[nixpkgs] @marsam opened pull request #81183 → talloc: fix build on darwin → https://git.io/JvznH
hmpffff_ has joined #nixos
hmpffff has quit [Read error: Connection reset by peer]
<evanjs>
tilpner: what's the rule of thumb for those vs override?
<evanjs>
Not sure I ever remember :P
<tilpner>
evanjs: .override is for function arguments, .overrideAttrs is for attributes
<tilpner>
evanjs: In { foo }: mkDerivation { quux = foo; bar = 42; } you can alter quux with override, and both with overrideAttrs
<tilpner>
evanjs: override becomes useful when the value is used while constructing something else
<clever>
.override can only alter the foo in that
<tilpner>
evanjs: E.g. in { version }: mkDerivation { src = fetcher { inherit version; }; }, an override is dry-er and shorter than the overrideAttrs would be
<tilpner>
clever: And foo is only used in quux, so overrie can only affect quux there
<clever>
yeah
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage2nix: stick to older hackage-security version to make cabal-install happy »: https://git.io/JvzcU
KeiraT has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/Jvzck
<{^_^}>
[nixpkgs] @MrMebelMan opened pull request #81184 → python3Packages.sapi-python-client: init at 0.13.1 → https://git.io/JvzcL
<evanjs>
clever: though tbh input isn't as important as mounting the drive. Was only trying to enable input to debug that
<evanjs>
for some reason that's failing, too :D
<clever>
evanjs: how is that failing?
zupo has joined #nixos
<evanjs>
clever: lemme boot up the real hardware again and see
Neo-- has joined #nixos
<Cadey>
how would i make a nix package that adds a TLS certificate authority to the system trust store?
hmpffff has joined #nixos
hmpffff_ has quit [Read error: Connection reset by peer]
<tilpner>
You can't, you need a module for that
<enamax>
tilpner: any idea why the module isn’t built though?
<tilpner>
enamax: Have you changed it?
<clever>
security.pki.certificateFiles
<clever>
A list of files containing trusted root certificates in PEM format. These are concatenated to form /etc/ssl/certs/ca-certificates.crt, which is used by many programs that use OpenSSL, such as
<tilpner>
... to overrideAttrs?
<enamax>
Yep
<clever>
Cadey: you want to add the file to this list
<tilpner>
And then you did a nixos-rebuild switch?
<tilpner>
How do you test if it has worked?
<Cadey>
clever: will that work for a non-nixos system?
<tilpner>
Wait, what's that in your function argument pattern...
<tilpner>
How are you taking fetchFromGitHub up there?
<enamax>
Uh, bad copy paste from ena/default.nix? I was wondering where I am supposed to get fetchFromGitHub from...
<tilpner>
From pkgs
<tilpner>
But this makes me wonder if you've really been testing the right thing
<evanjs>
clever: well now I'm getting no filesystem type specified, which I don't think I was getting before. Suppose I'll specify fat, just don't usually need to do that IIRC
<clever>
enamax: now that i look at your gist, i instantly know the problem
<clever>
enamax: nixos will call boot.kernelPackages.extend to change some kernel flags, which undoes your // { ena = ...; }
<tilpner>
I was wondering about extend, but couldn't explain why that might help!
<tilpner>
:(
<clever>
enamax: you must use pkgs.linuxPackages_latest.extend (self: super: { ena = ...}) or it just gets undone
<evanjs>
clever: I keep forgetting stuff gets redirected to /dev/null by default >_>
hmpffff has joined #nixos
<clever>
and as a result of it ignoring the ena=, it lazily never tried to eval fetchFromGitHub
<clever>
so that error is silently ignored
<tilpner>
But then why did the manual nix-build build to the right version?
<evanjs>
clever: ah. nvm. coworker had a bright idea. check and see if fat support is compiled in....
hmpffff has quit [Client Quit]
<evanjs>
CONFIG_VFAT_FS=m :D oops
<clever>
evanjs: not using an initrd i guess?
<tilpner>
Assuming "the sock version" is the right version
<evanjs>
I am
dm9 has joined #nixos
<tilpner>
Ahh, it uses boot.kernelPackages.extend, and never exposes that
<evanjs>
We basically mount the ESP drive to /boot, and extract "update" files there (bzImage and initrd)
<clever>
tilpner: yep, i traced this down a few months ago when somebody else had the same issue
<evanjs>
but otherwise only really run off the initr(d/ramfs)
<enamax>
clever: ok that causes the module to be built, but it builds the stock version (/nix/store/012a7crf13dcyl8r873vahbhni76dfxb-ena-2.1.2-5.4.22.drv)?
zebrag has joined #nixos
<enamax>
possibly because the real source comes from the kernel tree? O_O
<JonReed>
I've found this command `nix-build '<nixpkgs/nixos>' -A vm -I nixos-config=./configuration.nix` that should build a configuration.nix into a runnable virtual machine. But I'm confused about how to set it up. What device should I specificy as a filesystem in configuration? What that virtual machine will use as a disk for storing things while its
<JonReed>
running (/nix/store/ is obviously readonly)?
vandenoever has quit [Read error: Connection reset by peer]
<clever>
JonReed: it will create a qcow2 file in the working directory, matching the hostname of the machine, as set by networking.hostName within configuration.nix
gustavderdrache has joined #nixos
<JonReed>
clever: Oh, that's cool.
vandenoever has joined #nixos
vandenoever has quit [Changing host]
vandenoever has joined #nixos
<enamax>
clever: thanks a bunch so far, I can confirm that it builds the right version however it doesn’t seem to use it. modinfo still shows the stock version
<CRTified>
JonReed: There are also some settings that allow to e.g. set custom qemu arguments or changing stuff like available memory
<JonReed>
So, I should just specify a standard `{ "/" = { device = "/dev/disk/sda" }; }` in that configuration.nix when using a virtual machine? What if I want to also pass a path from the host (like /mnt/my-storage) to that virtual machine while its running so that it can access it (for, let's say, a database storage)?
<enamax>
(I’d expect version to be 2.2.3 or something
<clever>
JonReed: nixos will override the definition of / when using a vm, and ignore whatever you specify
<clever>
JonReed: but nixos will also setup a 9plan mount of an xchg dir, so you can get files in/out
<clever>
JonReed: check the args to qemu (via ps aux), and the output of `mount` inside the guest, to see where xchg is
fresheyeball has quit [Quit: WeeChat 2.6]
<enamax>
clever: I’m thinking maybe I have to update boot.extraModulePackages as maybe its set to the stock version by amazon blurbs
<enamax>
tbh I have no idea what imports = [ <nixpkgs/nixos/modules/virtualisation/amazon-image.nix> ]; does
<JonReed>
clever: Ok, huge thanks!
<clever>
enamax: what did you set extraModulePackages to?
<clever>
enamax: just read that path, within the nixpkgs repo
veske2 has quit [Quit: This computer has gone to sleep]
veske has quit [Quit: This computer has gone to sleep]
<evanjs>
anybody know if there's something in nixpkgs/nixos like gentoo's "offensive" flag? :D I miss my potty-mouthed terminal - I think this --- OH
peanutbutter144 has quit [Quit: No Ping reply in 180 seconds.]
<evanjs>
withInsults! yessss. thank you nixpkgs maintainers :D
peanutbutter144 has joined #nixos
Boarders has joined #nixos
<Boarders>
Hi, I want someone to be able to install a cabal project on their machine without a local cabal or ghc, is it possible to do this with some variant on "nix-shell --packages ghc --run 'cabal install'
<Boarders>
I need it to use Cabal-3.0 though as the package uses features from that
<enamax>
I have repeated that in my configuration.nix (with the new ena module version), but *still* I’m being booted into the 2.1.0 version arghh ^^
<enamax>
Theory: the old version is still in the initrd? I can confirm that my build did regenerate the initrd, did it use the new version though? I dunno.
<enamax>
boot.initrd.availableKernelModules takes strings so.. I dunno
<evanjs>
is there any way to query current package arguments via nixos-option / nix eval? Like, to see if my setting for sudo's withInsults argument is in effect
<clever>
enamax: boot.extraModulePackages = [ ena ];
<clever>
enamax: how is this getting a reference to ena?
<JonReed>
Boarders: I don't remember specifics for Haskell / cabal, but I think if you `nix-build` the file that `cabal2nix` outputs, it should build the binary as well, which then can be installed using `nix-env -i`
<Boarders>
JonReed: Thanks for the pointer. Unfortunately this project doesn't build with just that as some of the dependencies are marked as broken and some have failing tests
<fresheyeball>
Boarders: nix-env -if .
<Boarders>
but I am hoping to try to use overlays to fix the problems
<fresheyeball>
will install the binary in your env
<fresheyeball>
but it sounds like you still need to do some things to get it building
<fresheyeball>
I still can't seem to import a newer version of the nix package
<Boarders>
Yeah, we have it so that it can build with cabal so long as you have the correct version of ghc on your machine and the correct version of cabal, I want to be able to easily have someone install the project without having to worry about those details so I want to use nix to grab the correct versions
<enamax>
clever: yeah it’s definitely getting its ena from somewhere else, maybe the kernel source tree? the loaded module doesn’t match any of the versions I have built...
<Boarders>
this seems to work for me: nix-shell --packages haskell.compiler.ghc881 --run 'cabal install'
<fresheyeball>
when I point it at a local checkout it works fine. But when using fetchFromGitHub I get a weird error
<Boarders>
separately I would like to build the whole project with nix, but that will take more work from me to figure out how you do it
<fresheyeball>
It says I don't have access to a public repo
<fresheyeball>
which makes no sense
<evanjs>
clever: hrm alright. now it's "special device /dev/sda1 does not exist" :P maybe i'll `ls /dev > /dev/console` in sysinit as well
<clever>
enamax: try running realpath on the ena.ko.gz, and see which storepath it came from
<Boarders>
dhess: thanks that will be really helpful
<dhess>
it's pretty complicated but also includes most of the things you'd ever need to do, I think.
<fresheyeball>
I checked the source directory mentioned in the error and it's the nix repo as expected
wfranzini has quit [Quit: wfranzini]
<clever>
evanjs: `exec > /dev/console` will change stdout for the entire shell script, so you dont have to redir everything
<enamax>
clever: yeah it comes from the linux-5.4.22 package
<fresheyeball>
clever: hopefully I forgot {} again or something
<clever>
evanjs: also double-check where 0,1,2 lead, in `ls -l /proc/self/fd/`
<dhess>
Boarders: feel free to copy it and adapt it to your own uses
<evanjs>
clever: ooooo good call
<clever>
enamax: ah, so youll need to somehow pick the ena from outside, rather then the one within linux
<aranea>
Hrm, pcscd doesn't even come with any rules, it's just run as root.
<aranea>
Well, I guess I'll just write my own udev rules then.
<enamax>
clever: guess I could rebuild the kernel with CONFIG_ENA_ETHERNET=n
<fresheyeball>
fatal: '/nix/store/8nnbnhj32z33gr7a4ix3z6fqw7xh1nwc-source' does not appear to be a git repository
<enamax>
but will it then automatically find the new ena or will I have an unbootable system...
<fresheyeball>
is definately the git repo
<dhess>
Boarders: you may run into problems trying to override Cabal, FYI. There are some packages in haskellPackages that are difficult to override because they're also part of GHC's built-in packages
<dhess>
I'm not saying it won't work, I just don't know for sure.
<clever>
enamax: thats what extraModulePackages does
<clever>
enamax: you can also use `nixos-rebuild test` and `modinfo` to check if the new module is present, before you make it permanent
<clever>
enamax: and confirm its using the new path
<fresheyeball>
oh I wonder if the nix repo needs it's .git folder, and fetchFromGitHub is removing it
<karantan>
I have this derivation: https://pastebin.com/F0SJDd3H. how can I add this derivation to my configuration.nix and use it in `environment.systemPackages` ?
<karantan>
I've try with adding `imports = [ ./hello.nix ...];` and then `environment.systemPackages = [ helloWorld ]; ` but it complains "undefined variable 'helloWorld'"
endformationage has joined #nixos
<{^_^}>
[nixpkgs] @Mic92 closed pull request #81187 → knot-resolver: fix cpu usage in kres-gc-cache → https://git.io/JvzWA
<clever>
karantan: but test-environment itself, cant never be built
<karantan>
so I first need to build the derivation?
<clever>
karantan: its myScript that you want in systemPackages, not test-environment
<fresheyeball>
ok that was weird, but I seem to have fixed it
<clever>
karantan: this is where it helps to have seperate shell.nix and default.nix files, shell.nix contains test-environment, default.nix contains myScript
JonReed has quit [Remote host closed the connection]
philr has quit [Ping timeout: 240 seconds]
<evanjs>
clever: okay maybe ls’ing _all_ of /dev wasn’t a great idea 😝 as for fd, 0,1, and 2 go to /dev/console
<evanjs>
3 cannot read link: No such file or directory
<clever>
evanjs: then io should work fine, and yeah 3 is a bit special with ls, thats the fd dir itself
<clever>
which it then closed, before trying to stat itself
<evanjs>
Alright, I’ll make my ls dev a bit more specific anyhow
ddellacosta has joined #nixos
<fresheyeball>
well I though it was all better
<fresheyeball>
I got the update nix package
<fresheyeball>
but `nix repl` says nix-command is disabled
<clever>
fresheyeball: ive been getting the same error in a nix-shell build of nix
<fresheyeball>
clever: let me know if you find out, it's odd
<fresheyeball>
oddly I have to set officialRelease = true as well
<evanjs>
infinisil: do you ever get nix eval errors with resulting from config/new-modules/default.nix:10:6 or similar? in this case, it can't find new-modules/dev, which will likely be resolved by a(n even partial) rebuild
<evanjs>
clever: okay, so nothing for /dev/sd* at all :D
<evanjs>
Wonder if I'm missing another kernel option...
<clever>
evanjs: dmesg would have shown the device anyways, what about `lspci` and `lsblk` ?
<clever>
evanjs: just dump the whole thing to the screen?
<clever>
evanjs: also if your in qemu, try /dev/ttyS0 and link the serial to stdio
<evanjs>
Yeah ttyS0 seems to work well with qemu, just wasn't sure how much I could emulate there that would apply to the other system
<karantan>
@JonReed looks nice. thanks, I'll use it
<evanjs>
and as for the real hardware, trying to be cautious with what dumps out, as it's a pretty "low-res" framebuffer -- i.e. not much fits on there :D
<JonReed>
karantan: So `;` though in the first expression. I wrote it from head and already see this mistake :D
chloekek has quit [Quit: WeeChat 2.7]
<karantan>
yeah np. I just needed to know the concept
<karantan>
can I also use any of this to actually create a nix package?
<karantan>
or can only I use this in my config
<evanjs>
clever: mmkay which "Kernel driver in use" am I looking for? I only see xhci_hcd
JonReed has quit [Remote host closed the connection]
<clever>
evanjs: the driver in use, for the line following your sata controller
<evanjs>
and confirmed no sda in /dev
<evanjs>
oh right sata duh
<clever>
07:00.0 SATA controller: ASMedia Technology Inc. ASM1062 Serial ATA Controller (rev 01) (prog-if 01 [AHCI 1.0])
<clever>
Kernel driver in use: ahci
JonReed has joined #nixos
<clever>
also, qemu may use virtio by default
<JonReed>
karantan: I have an alias for this `alias nix-build-call='nix-build -E "with import <nixpkgs> {}; callPackage ./default.nix {}"'` It builds the `default.nix` within a directory.
<clever>
,callPackage
<{^_^}>
If a Nix file foo.nix starts with something like `{ stdenv, cmake }:`, you can build it with `nix-build -E '(import <nixpkgs> {}).callPackage ./foo.nix {}'`
<evanjs>
clever: ....... good call, once again :D
<evanjs>
CONFIG_SATA_AHCI=m
<evanjs>
well then XD
<evanjs>
(note I'm not linking all modules into lib/modules, just a few for a minimal system)
<evanjs>
e.g. SATA_AHCI wouldn't work until I set ATA to no (or presumably yes)
<Valodim[m]>
Jan Tojnar: I did, but github is being annoying
<Valodim[m]>
the force-push succeeded, and git says upstream is up to date. it just doesn't show up in the github web interface. but github's been weird all day, so *shrug :)
<pikajude>
anyone here an expert on the rust crate builder?
cole-h has joined #nixos
<{^_^}>
[nixpkgs] @jtojnar closed pull request #79032 → sphinxsearch: Add support for MySQL & xmlpipe2 → https://git.io/Jv3AL
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « git-annex: update sha256 hash for version 8.20200226 »: https://git.io/Jvz8c
<teto>
,query {^_^}
<{^_^}>
[nixos-org-configurations] @edolstra pushed to master « Make the releases bucket listable again »: https://git.io/Jvz88
kleisli has joined #nixos
zupo has joined #nixos
waleee-cl has quit [Quit: Connection closed for inactivity]
sgrunert has quit [Remote host closed the connection]
<Baughn>
nixos-generate-config really needs to stop setting maxJobs equal to # of CPU threads.
<Baughn>
...that's my job, I guess.
<gchristensen>
oh?
smatting has quit [Ping timeout: 272 seconds]
<Valodim[m]>
Jan Tojnar: I amended the commit without changes and pushed again, seems it went through this time. No idea what happened there :) but commit is up now
<Baughn>
gchristensen: It defaults to 48 on my system, but I have only 16GB of ram.
o1lo01ol1o has quit [Remote host closed the connection]
<Baughn>
Which is silly, I know. Amazon is being slow. But still.
o1lo01ol1o has joined #nixos
<Baughn>
I threatened a year ago to rewrite it in Rust. Since then I've actually learned Rust. P
<evanjs>
mmmmm Rust. any progress with the rewrite so far? :P
<Baughn>
As I said, I have since learned Rust. :P
<evanjs>
right but the program lol
<Baughn>
To my satisfaction, at any rate. I could probably have bungled through back then, but it wouldn't have been readable.
<Baughn>
Well, I forgot about it until *this* happened.
<evanjs>
ahhh
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<softinio[m]>
So I am trying to understand how to get the most out of using niv, home manager & nix-darwin on my laptop
<softinio[m]>
so far I have only used nix-darwin
dozn has joined #nixos
<softinio[m]>
all the global apps I want installed from nixpkgs I have in my config for nix-darwin
<softinio[m]>
where does niv fit into my workflow? How should I be using it?
erasmas has joined #nixos
<pikajude>
niv is just a dependency manager for stuff you're fetching from github
<pikajude>
or stuff that isn't already in nixpkgs
<pikajude>
it automates the nix-prefetch-git => copy the hash into a source file workflow
<jared-w>
softinio[m]: niv makes pinning sources easy and ergonomic, that's all. I use it to pin down nixpkgs, all my overlays, and anything I would otherwise "fetchTarball/url/Github/etc".
<softinio[m]>
jared-w: so would you switch to using niv for your global packages in place of nix-darwin?
<{^_^}>
[nixos-org-configurations] @edolstra pushed to master « Add channels.nixos.org to hold the redirects to releases.nixos.org »: https://git.io/Jvz8D
<pikajude>
IME niv is good for managing dependencies of a package you're working on
<pikajude>
actually no, sorry, misunderstood the question
<jared-w>
softinio[m]: niv and nix-darwin are used for different things. nix-darwin is essentially a large collection of modules that automate the setup of a bunch of functionality for macos.
<pikajude>
nix-darwin is a bunch of nix expressions
<pikajude>
niv is *only* a manager that helps you keep your git and other dependency urls updated
<softinio[m]>
think I need to play with niv to understand its use more .... what you install with niv will it be global or just for the project you are working on ? or can it be both?
<pikajude>
oh yeah, nix-darwin and home-manager are basically the same thing
<pikajude>
`niv` is a command line tool that you can use to produce a json file containing a list of sources
cosimone has quit [Read error: Connection reset by peer]
<pikajude>
and a nix file for importing tarballs from that file
<softinio[m]>
oh didn't realize nix-darwin and home manager same
<pikajude>
they do the same thing, different projects though
<LnL>
some people use them both
<pikajude>
nix-darwin is obviously more darwin specific
<Baughn>
Things which I am trying: `stdenv = super.impureUseNativeOptimizations super.stdenv;`
<Baughn>
My poor CPU...
<LnL>
depends how much stuff you want to configure globally vs user local
<Baughn>
This caused it to download 8.3GB of source code. :X
<softinio[m]>
sorry for all the newby questions just trying to see how each tool is used and what I need for a perfect setup
<DigitalKiwi>
hey pedants; if a package is the attribute python{,2,3}Packages.digital-ocean, the directory is .../python-modules/digitalocean, and the pname is python-digitalocean,...are any of these named wrong, and which name do i use in a PR to update it (for unrelated reasons not the name being inconcsistent) >.>
<pikajude>
i use home-manager because i want to share the exact same config between nixos and macos
cosimone has joined #nixos
karantan has joined #nixos
karantan has quit [Client Quit]
<pikajude>
python packages are named manually aren't they?
<DigitalKiwi>
i know how to spell inconsistent i don't know how to type
waleee-cl has joined #nixos
<softinio[m]>
as I am happy with nix-darwin thinking I stick with it and not use home manager for now and maybe investigate further if niv is a benefit to my workflow and setup
<DigitalKiwi>
like should the directory be python-modules/digital-ocean or should it be changed in pkgs/top-level/python-packages.nix or does not not matter
nbathum[m] has joined #nixos
<DigitalKiwi>
pikajude: yeah the attribute comes from top-level/python-packages.nix
<pikajude>
that file isn't autogenerated i don't think
<pikajude>
if someone wrote it manually, i'd say fixing the attrname is just fine
<DigitalKiwi>
right, someone put it there like that, but renaming it would be breaking, but it annoys me that it's named different and it's named different from upstream :((
<pikajude>
you could just add an alias
<pikajude>
in aliases.nix
<pikajude>
oh actually, i don't know if those can be applied recursively
<Boarders>
if I wanted to install a bunch of packages then do I use nix-env -qa [list of packages]?
<Boarders>
the nix-env help page is very thorough but that makes it hard to find the default solution
<pikajude>
-q is query
<cole-h>
Pretty sure that won't install them. You need `-i`
<Boarders>
ah ok thank you
<cole-h>
(or `--install` if you want to be verbose)
<evanjs>
clever: woo! HID_GENERIC was indeed one of the missing pieces
<evanjs>
clever: I see no kernel driver in use for SATA, however. ... I didn’t need ATA... did I..?
<clever>
evanjs: which transport is qemu using?
gustavderdrache has quit [Quit: Leaving.]
<evanjs>
Oh I’m on the physical hardware atm
<evanjs>
I could link all the modules into the initrd and see what autoloads, or at least what appears under kernel driver in use for SATA?
c0c0 has quit [Quit: WeeChat 2.6]
genesis has quit [Quit: Leaving]
benpicco has joined #nixos
nuncanada has joined #nixos
<benpicco>
Hi, if in a script I don't want to call a shell build-in version of a command, e.g. echo, but the binary version, I'd usually call it with it's full path (/usr/bin/echo) - how do I get that path on NixOS?
<pikajude>
it's usually in /run/current-system/sw/bin
genesis has joined #nixos
<pikajude>
you can also find it with `type -p` in bash i think
<cole-h>
If you're creating this script in a Nix expression, you can use string interpolation like so: "${pkgs.package-with-binary}/bin/echo"
<benpicco>
well the script should be portable
<pikajude>
/usr/bin/env exists on nixos
<pikajude>
scripts that use #!/usr/bin/env python work, for example
<pikajude>
but echo isn't a shell so you shouldn't be using a shebang for it
<pikajude>
but a script would be using a specific shell anyway
<mdash>
yeah i was gonna say, a script probably uses bash
o1lo01ol1o has quit [Remote host closed the connection]
<mdash>
in zsh it's `disable echo`
zupo has joined #nixos
JonReed has joined #nixos
phreedom_ has joined #nixos
phreedom has quit [Ping timeout: 240 seconds]
<JonReed>
When I run a vm created with `nix-build '<nixpkgs/nixos>' -A vm -I nixos-config=./configuration.nix` the nixos console there is smaller than the QEMU window How can I can make it resize to all available window space?
<benpicco>
Thank you, I think /usr/bin/env echo will do the trick
<{^_^}>
[nixpkgs] @Tomahna reopened pull request #80000 → mopidy-mpd: init at version 3.0.0 → https://git.io/JvWwj
boxscape has quit [Remote host closed the connection]
leotaku has joined #nixos
boxscape has joined #nixos
<genesis>
if i remove a package, say random , nix remove so ~/.nix-profile/share/dbus-1/services/random.service , but random service is still activate on dbus
<ShaRose>
mdash not sure if you saw it, but the reason I was looking for command execution was because I was hoping to use a script that can decrypt information using private host info like the uuid and that
o1lo01ol1o has joined #nixos
<ShaRose>
I've already got a tool I can use for it (generates a keypair using host info, then uses that to decrypt info)
<jared-w>
nix copy --from "local?root=some_closure.tar.gz" would just make my day
<emily>
mdash: nice, thanks
<emily>
mdash: so the docs are wrong :(
<pikajude>
they are incomplete :)
<tilpner>
jared-w: Is a Nar not acceptable?
<jared-w>
I don't think nix copy v2 makes nars?
<xa0[m]>
the docs in nix really should be updated, it's not very newbie friendly
johnw has joined #nixos
rogue_koder_ has quit [Quit: Konversation terminated!]
rogue_koder_ has joined #nixos
<tilpner>
jared-w: The "nix" CLI is still experimental, there's nothing wrong with using the stable tooling
<emily>
pikajude: disagree; "fetchTarball url, builtins.fetchTarball url" implies to me that passing a non-URL is a type error
<pikajude>
it implies that because you assume nix has a type system
<pikajude>
it doesn't
<emily>
:(
<pikajude>
i tried to write a typed nix implementation once...
<emily>
well it says "Download the specified URL"!
<jared-w>
Currently I'm doing `nix copy --to "local?root=store"` and then tarring that up and extracting in a sandbox'd environment and copying it over with `nix copy --from "local?root=untarred_dir" --all`. Works much more reliably for me (for some obscure reason) than nix-store export/import was
<pikajude>
never try that, if you value your sanity
<mdash>
pikajude: sanity's just a one trick pony anyhow
vandenoever has quit [Read error: Connection reset by peer]
<jared-w>
(although I suspect the nix-store import/export issues were really something else that's subsequently been mostly fixed. But still)
<emily>
pikajude: kind of already signed up for that ;w; (though I'd never try to reimplement Nix-as-it-is since there's a bunch of obvious design issues you can fix if you're going to do something as drastic as adding a type system)
<xa0[m]>
i've always dreamed of just having a language that compiles to nix that's a bit more sane
vandenoever has joined #nixos
vandenoever has joined #nixos
vandenoever has quit [Changing host]
<pikajude>
honestly, what i actually tried was to make a typed nix
<mdash>
xa0[m]: dhall is one
<pikajude>
but i got caught up on making an untyped one
<pikajude>
dhall doesn't support dynamic key/values, which makes sense but it's a bit annoying
<emily>
pikajude: re "why are these paths in your system config?", nixpkgs-overlays=https://clbin.com/nsfTw, source-paths.json is https://clbin.com/Yc0bU
<xa0[m]>
hmm
<xa0[m]>
cool
<pikajude>
oh yeah, no, it's kind of an antipattern to "import" store paths
<pikajude>
because as you see, it's plainly unreadable
rogue_koder_ has quit [Read error: Connection reset by peer]
<emily>
this is a generated file
<emily>
my infrastructure is more powerful than you can possibly imagine, clearly :P
<pikajude>
oh really?
rogue_koder_ has joined #nixos
ris has joined #nixos
<emily>
I do this to get consistent overlays between my NixOS config and the nix commandline tools
<pikajude>
oh are you generating it?
<emily>
without requiring to import/eval the nixos config like https://nixos.wiki/wiki/Overlays#Using_nixpkgs.overlays_from_configuration.nix_as_.3Cnixpkgs-overlays.3E_in_your_NIX_PATH does
<emily>
Q: what's with that weird first overlay? A: /nix/store/l371qjv6jwn4slxa9mzfckagqc2jgzaw-lorri.nix is https://clbin.com/G9YZL
<pikajude>
so you're trying to make it possible to remember which package is what
<pikajude>
which is understandable
<emily>
yeah
<emily>
e.g. let's say nixpkgs-mozilla breaks horribly
<pikajude>
if ambitious
rogue_koder has joined #nixos
<emily>
it sucks to see a "-source" path in --show-trace
rogue_koder_ has quit [Client Quit]
<emily>
and people's nix code breaks randomly in ways that requires source-diving all the time, because it isn't typed :P
<pikajude>
is guile typed?
<emily>
also it sure is wild how there's builtins.{from,to}JSON but no {from,to}Nix
<pikajude>
genuine question, i have no idea
<emily>
don't think so
<pikajude>
classic
<emily>
racket has a pretty extensive contract system and also Typed Racket
<emily>
time for guix-racket clearly
<pikajude>
but racket doesn't have a package manager language does it
<pikajude>
oh
<pikajude>
racket? i barely know it!
<emily>
i don't think that exists, but it could, i guess
<jared-w>
emily: do you have the repository where you're doing the overlay stuff in its full? I'm currently doing the overlays in your nix path solution but I don't particularly enjoy having to eval the nixos config and your solution looks nice
<emily>
honestly racket seems like a much better choice for making DSLs than guile too, I assume it's just the embeddability and also GNU solidarity that made them pick Guile
<pikajude>
is there typed brainfuck
<emily>
did you know NixOS has an option to automatically use nonsense like GNU lsh instead of OpenSSH
<{^_^}>
[nixpkgs] @flokli opened pull request #81200 → edid-generator: init at unstable-2018-03-15 → https://git.io/JvzE5
<emily>
that sounds hard to do while retaining any of the characteristics of brainfuck whatsoever
<emily>
but I'd have said the same about javascript
<xa0[m]>
if i had a choice i'd probably be using guix over nix just because the language is so much more flexible and nix has quirks that i don't know i'll ever get used to, but guix is also even more gnu and has far fewer packages without just importing the nix ones
<JonReed>
I enjoy the Nix syntax. What I miss the most is a type system and good unit testing. A nice debugger with breakpoints would be nice to have too. For NixOS, I wish for it to be written with security and isolation in mind, so that everything only has access to resources it needs. It is an opportunity wasted, IMO, as resource access can be viewed as
<JonReed>
another form of dependency. So, if you have to think about every dependency a package has you might as well think about what resources it needs to run.
Henson has joined #nixos
<{^_^}>
[nixpkgs] @vcunat pushed commit from @Ma27 to release-20.03 « Merge #81155: wireshark: 3.2.1 -> 3.2.2 (security) »: https://git.io/JvzuG
<Henson>
is there a way to get nix-copy-closure to use an SSH port number other than 22?
cosimone has joined #nixos
<emily>
I like some things about the NixOS syntax but not so much forgetting ; constnatly
linarcx has joined #nixos
Shouou has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Henson>
setting up a different port for the host in ~/.ssh/config seems to do the trick
bvdw has quit [Read error: Connection reset by peer]
<worldofpeace>
lovesegfault: woah, it looks pretty confusing.
bvdw has joined #nixos
<zeta_0>
i am trying to install this live server extension in my home.nix, when i do a rebuild switch, nothing happens no errors or anything, and then when i reboot the extension does not seem to be there
<lovesegfault>
worldofpeace: Yeah
<worldofpeace>
lovesegfault: looking at the debug logs, I guess it creates a /run/tlp directory imperatively?
<zeta_0>
does anyone here have experience with installing extensions for vscodium?
* lovesegfault
nods at worldofpeace
<lovesegfault>
I didn't need to create that one
<xa0[m]>
presumably it's gonna create a new derivative of vscodium, not add the extension to the one you're running. i think if you restart it it should work?
<lovesegfault>
It's `TLP_RUN` var in the Makefile
<lovesegfault>
it gets seded to /run/tlp at make-time
<worldofpeace>
lovesegfault: Right, I guess that means it really would be overworking to add a RuntimeDirectory = "tlp";
<lovesegfault>
I think so, although I am not sure whether that is a lingering dir from some other version of tlp
<lovesegfault>
examining the source I don't see runtime mkdir of /run/tlp
<lovesegfault>
worldofpeace: Submitting patches upstream now
<zeta_0>
how strange, after adding a vscodium extension to my home.nix file and doing a rebuild switch, nothing happened no error or anything, i am not sure what is going on here? https://hastebin.com/hawajunewu.py
<zeta_0>
i am trying to install the liveserver extension
<worldofpeace>
lovesegfault: awesome, cc me on them if you can.
rogue_koder has quit [Ping timeout: 260 seconds]
<lovesegfault>
will do
<xa0[m]>
zeta_0: home.nix should require a "home-manager switch", it won't be read by global nixos config
<xa0[m]>
they're separate systems
noudle has quit []
rogue_koder has joined #nixos
<zeta_0>
xa0[m]: i've had home.nix integrated as a submodule for configuration.nix for the longest time, so that is not an the issue that i am having, maybe i named the vscodium extensions incorrectly, i'm not sure at this point
<xa0[m]>
hm, no idea then
<xa0[m]>
i don't use it :)
drakonis has joined #nixos
<evanjs>
bah, can I not skip make modules easily?
<evanjs>
Tried MODULES = no; and MODULES n but the build fails
<evanjs>
after the typical The present kernel configuration has modules disabled. and blah
<zeta_0>
xa0[m]: well, i can download extensions directly from vscodium, but i wanted to do everything the nix way, oh well, sometimes you have to do workarounds
<evanjs>
qyliss: figured as much. saw some custom configurePhases earlier but not feeling like doing that atm :D
<qyliss>
I've got it to work before but don't have it to hand
Thra11 has quit [Ping timeout: 258 seconds]
nuncanada has quit [Read error: Connection reset by peer]
nuncanada has joined #nixos
<{^_^}>
[nixpkgs] @worldofpeace merged pull request #80746 → openssl_1_0_2: mark as insecure; fixes #77503 (kinda) → https://git.io/JvR1f
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/Jvzg2
fuzzy_id has quit [Ping timeout: 240 seconds]
lovesegfault has quit [Ping timeout: 240 seconds]
lovesegfault has joined #nixos
<o1lo01ol1o>
clever: We had a life-cycle policy on S3 kick in and now I'm dealing with a number of hydra errors about paths to derivations not being valid. Do you know what I can do the purge the offending derivations from hydra's knowledge?
<NoctisLabs>
,locate bin telnet
<{^_^}>
Found in packages: busybox, inetutils
<cole-h>
o1lo01ol1o: Unfamiliar with Hydra, but would `nix-store --delete <path>` suffice?
<o1lo01ol1o>
cole-h: the paths are still alive
<cole-h>
`--ignore-liveness`?
<o1lo01ol1o>
cole-h: that deleted quite a bit from the hydra machine; I'll see if the jobs want to proceed now
<cole-h>
I take no responsibility if things end up breaking due to that
* cole-h
hides
zupo has joined #nixos
Thra11 has joined #nixos
rail has joined #nixos
lovesegfault has quit [Quit: WeeChat 2.7.1]
<Henson>
there's an old software package for Ubuntu called "rpncalc" that is no longer maintained and was removed from Ubuntu packages. However I like the program and would like to add it to Nix packages. I could take the source and add it to my own Github repository, but there's someone else who has also taken it and added it to their own Github repository. Should I contact them and see if they are....
rail has quit [Ping timeout: 240 seconds]
<Henson>
okay with me using their repository for the derivation, or should I just fork their repository and add it to mine? OR, should I make a nix derivation that downloads the source code from Ubuntu and compiles it directly? This would make any software modifications and improvements somewhat kludgy.
<Yaniel>
contacting them is generally a good idea
<Yaniel>
also you could just make your overlay for it
<Yaniel>
your own*
rogue_koder_ has joined #nixos
rogue_koder has quit [Ping timeout: 258 seconds]
bukkitgerman8 has quit [Quit: Ping timeout (120 seconds)]
bukkitgerman8 has joined #nixos
lovesegfault has joined #nixos
realrokka has quit [Ping timeout: 265 seconds]
<Henson>
Yaniel: right, I could just make my own overlay for my personal use, but perhaps I should make it so it can be added to nixpkgs? There aren't many ronsole reverse polish calculators in nixpkgs.
malSet has quit [Read error: Connection reset by peer]
<emily>
jared-w: sorry, missed this -- it's just my private config repo, sadly
malSet has quit [Read error: Connection reset by peer]
<jared-w>
emily: sadness. Make your configs public. #itIsTheWay :p
<emily>
jared-w: I just use niv to manage dependencies and pipe it through to my NixOS configs, then it's just setting overlay-paths and everything works from there
<jared-w>
(I can figure it out from the files you linked, I was just surprised to see an overlay method that wasn't documented in the wiki and had less downsides)
<emily>
but then all my modules can just take sources as an argument
<emily>
I should clean it up and make it public, tbh
<emily>
the problem is the "clean it up" part
realrokka has joined #nixos
<evanjs>
clever: think I found my (next) problem :D CONFIG_FAT_DEFAULT_CODEPAGE=437 CONFIG_NLS_CODEPAGE_437=m
<evanjs>
well then lol
lovesegfault has joined #nixos
<emily>
jared-w: fwiw even the method on the wiki could be simpler: if <nixpkgs-overlays> points to a file then it can evaluate to a list of overlays, whereas the wiki snippet duplicates the overlay application logic in a single overlay
<emily>
but I avoided it because I'd need to evaluate my NixOS config which sounded slow, and plumbing in the niv sources looked annoying
<evanjs>
Okay... Codepages are being finicky; is there something specific I need to do to force a specific NLS_ISO* option on?
<evanjs>
trying to set CONFIG_NLS_ISO8859_1=m to y and it doesn't seem to be listening to be >_>
drakonis_ has joined #nixos
drakonis_ has quit [Read error: Connection reset by peer]
drakonis has joined #nixos
chagra_ has joined #nixos
philr has joined #nixos
<evanjs>
it looks like it's in commonStructuredConfig as m. Is there anything special I need to do to override commonStructuredConfig items?
<o1lo01ol1o>
Ola clever, I was wondering if you knew how I could best address some objects that were recently life-cycled in S3. I'm getting hydra errors such as: Aborted: path '/nix/store/1h3b6a1884y16bib8xsf5wl0s3fvgwz2-gogol-gmail-0.4.0-doc' is not valid
drakonis has quit [Read error: Connection reset by peer]
drakonis has joined #nixos
Flambe has joined #nixos
Thra11 has quit [Quit: WeeChat 2.7]
drakonis has quit [Read error: Connection reset by peer]
drakonis has joined #nixos
JonReed has quit [Ping timeout: 260 seconds]
<evanjs>
.... found it ...
<evanjs>
to override anything from commonStructuredConfig, you must override extraStructuredConfig; any other options can be overridden via structuredExtraConfig x_x
vandenoever has joined #nixos
vandenoever has joined #nixos
vandenoever has quit [Changing host]
dingenskirchen has quit [Remote host closed the connection]
dingenskirchen has joined #nixos
pnobre_ has quit [Ping timeout: 258 seconds]
<evanjs>
Though this seems to not always be the case, either x_x
drakonis has quit [Read error: Connection reset by peer]
<kini>
Any fcitx users around? I can't seem to get fcitx to respond to the IME trigger key (which I've set to shift+esc) -- it's worked out of the box on a different NixOS machine, so I have no idea what could be different...
<iqubic>
cole-h: Yeah. A system rebuild fixed it.
xkapastel has joined #nixos
drakonis has quit [Read error: Connection reset by peer]
ixxie has quit [Ping timeout: 272 seconds]
drakonis_ has joined #nixos
<cole-h>
lovesegfault: With what? Probably just forgot to give a function an argument (or didn't give enough)
<lovesegfault>
cole-h: I figured it out
<lovesegfault>
I was:
<lovesegfault>
1. missing {} in import
kristian__ has joined #nixos
<lovesegfault>
2. using `pkgs` instead of `pkgs.path`
<theduke_>
is there a way to set the system-wide channel to unstable in configuration.nix ?
<kini>
hmm, looks like things are somewhat screwed up with fcitx -- looking at `journalctl -f` while changing settings in the config dialog, it seems to be failing to run something called `fcitx-remote` because it's looking for it in a relative path to the `fcitx-configtool` binary, whereas nixpkgs has it in a different store path
zupo has quit [Ping timeout: 265 seconds]
<samueldr>
channels are not generally configured in configuration.nix
<kini>
(and then it's also failing to even print that error message properly, generating another error message about writing to unitialized memory (!))
Boarders has quit [Remote host closed the connection]
lovesegfault has joined #nixos
<evanjs>
clever: alright, looks like it can mount the ESP now... though, it fails to do so automatically at start. Hrm. sleep until /dev/sda1 is there?
drakonis_ has quit [Read error: Connection reset by peer]
drakonis_ has joined #nixos
<cole-h>
lovesegfault: Yep, you forgot to give it an argument (the empty set). Glad you got it fixed!
<lovesegfault>
:D
karantan has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
gustavderdrache has joined #nixos
kapil_ has quit [Quit: Connection closed for inactivity]
lovesegfault has quit [Ping timeout: 240 seconds]
karantan has joined #nixos
chloekek has quit [Ping timeout: 258 seconds]
<{^_^}>
[nixpkgs] @jonringer opened pull request #81216 → nixos/plasma5: Fix activation script when XDG_CONFIG_HOME is unset → https://git.io/Jvzrd
philr has quit [Ping timeout: 258 seconds]
<kalbasit>
is there a way to get the download size from cache of a Nix store path?
linarcx has quit [Ping timeout: 240 seconds]
nerdmaxx has quit [Ping timeout: 265 seconds]
<evanjs>
how might one make an empty directory with makeInitrd?
<jared-w>
emily: What do you use instead of nixos-rebuild then?
<emily>
jared-w: with that system.nix you can do e.g. nix run -f system.nix -c sudo switch
<emily>
for other stuff generally you can just look at the nixos-rebuild.sh source code and replicate what it does... it's a pretty simple script
drakonis_ has quit [Read error: Connection reset by peer]
drakonis1 has joined #nixos
<jared-w>
https://clbin.com/JhIHA being the system.nix file you're talking about? Makes sense
drakonis1 has quit [Client Quit]
* jared-w
wonders how much simpler everything will be once flakes
drakonis has quit [Read error: Connection reset by peer]
maddo has quit [Quit: See ya]
vidbina has quit [Ping timeout: 260 seconds]
<maurer>
Out of curiosity, are we any closer to a solution for the openGL desync problem?
<maurer>
I know it's tricky because it kind of breaks fundamental assumptions about userspace/driver separation, but it is still an occasional annoyance
bvdw has quit [Quit: bvdw]
bvdw has joined #nixos
vidbina has joined #nixos
kleisli has joined #nixos
m4ts has quit [Ping timeout: 265 seconds]
drakonis has joined #nixos
andreas303 has quit [Remote host closed the connection]