<otavio>
go: inconsistent vendoring in /build/source/agent:
<{^_^}>
[nixpkgs] @grahamc opened pull request #104335 → nixos: release-combined: only build the amazon image for x86_64,aarch… → https://git.io/JkzJq
<{^_^}>
[nixpkgs] @grahamc opened pull request #104336 → [20.09] nixos: release-combined: only build the amazon image for x86_64,aarch… → https://git.io/JkzJs
<{^_^}>
[nixpkgs] @zowoq pushed to master « mmv-go: init at 0.1.2 »: https://git.io/JkzLU
page has quit [Ping timeout: 256 seconds]
Morfio has quit [Quit: This computer has gone to sleep]
<gchristensen>
clever: what do you mean the actual install?
<{^_^}>
[nixpkgs] @veprbl pushed commit from @bgamari to release-20.09 « mathematica: Prefer building locally (#100856) »: https://git.io/JkzLG
<clever>
gchristensen: on the machine where you where running nixos-install, check in its /mnt/
Morfio has joined #nixos
page has joined #nixos
hlolli_ has quit [Ping timeout: 246 seconds]
<clever>
gchristensen: i think i found the problem
<clever>
waiting on a nix-build to confirm
lovesegfault has quit [Quit: WeeChat 2.9]
<pie_>
clever: meh, more of the same; a package is built with qt514.callPackage, but the binary gives: `qt.qpa.plugin: Could not find the Qt platform plugin "xcb" in ""` , with nix-shell -I nixpkgs=channel:nixos-unstable --pure -p qt514.full gecode -v
<clever>
gchristensen: so depending on how old/new your nixos-install script is, it may behave differently, when installing the exact same nixpkgs
<clever>
[clever@amd-nixos:~/apps/nixpkgs-hasura]$ nix-build nixos --arg configuration '{users.mutableUsers=false; users.users.root.password="hunter2";}' -A vm -Q
<gchristensen>
I'm not actually calling nixos-install, just nix-build ./nixos -I nixos-config=./etc/nixos/configuration.nix --builders '' -A system then test -L result/sw/bin/passwd; echo $?
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « python27Packages.eradicate: 1.0 -> 2.0.0 »: https://git.io/JkzWq
growpotk- has joined #nixos
bbarker has quit [Remote host closed the connection]
bbarker has joined #nixos
<rogerr>
gchristensen ya i found that but why's it being created and how can i disable that?
<jackdk>
When you use an S3 bucket as a nix cache, is there any additional metadata stored there? i.e., if I put a lifecycle policy on the cache to delete old files, will I corrupt some metadata structure in the bucket?
<clever>
jackdk: a given path must have its entire closure on the bucket, or its invalid
<clever>
jackdk: there can also be multiple .narinfo files that depend on a single .nar.xz
<pie_>
its still telling me stuff like: -- Could NOT find Gecode, use Gecode_ROOT to hint its location (missing: GECODE_LIBRARY Driver Float Int Kernel Minimodel Search Set Support) (found suitable version "6.2.0", minimum required is "6.0")
<pie_>
oh wait what
<pie_>
"(found suitable version "6.2.0", minimum required is "6.0")" 0_o
<pie_>
aha ok
<pie_>
so i managed to pass it the includes i think but it cant find the libs
<pie_>
which is what i expected
sangoma has joined #nixos
arinov has joined #nixos
iH8c0ff33 has quit [Ping timeout: 260 seconds]
waleee-cl has quit [Quit: Connection closed for inactivity]
<pie_>
its definitely finding the libraries, so what the heck does `-- Could NOT find Gecode, use Gecode_ROOT to hint its location (missing: GECODE_LIBRARY) (found suitable version "6.2.0", minimum required is "6.0")` mean...
<{^_^}>
[nixpkgs] @FRidh pushed commit from @thoughtpolice to staging « pythonPackages: add new 'seccomp' library »: https://git.io/JkziX
simonpe^^ has joined #nixos
<simonpe^^>
I'm trying to use builtins.fetchGit to fetch the latest version of a repo for a specific scenario, how can I get it to force re-download when there are new commits on the branch I'm getting?
iH8c0ff33 has joined #nixos
<simonpe^^>
rn I have to manually enter the sha of the latest commit in the derivation, evaluate the derivation, then remove the sha again for it to "update"
roconnor has quit [Remote host closed the connection]
lordcirth has joined #nixos
roconnor has joined #nixos
lordcirth__ has quit [Ping timeout: 264 seconds]
<{^_^}>
[nixpkgs] @FRidh opened pull request #104362 → Merge staging 20.03 into master 20.03 → https://git.io/JkzPX
iH8c0ff33 has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @FRidh pushed 63 commits to staging-next: https://git.io/JkzX8
<chreekat[m]>
simonpe^^: there's a nix option that controls the timeout for fetchGit, but I'm on my phone and can't remember it off hand.
<chreekat[m]>
I think it's pretty ironic that a Nix built-in function is so dependent on state as arbitrary as "when was the last time this function was evaluated" :P
<pie_>
sphalerite: ah. I guess I should have actually tried. did end up by hand writing the 3 links
<simonpe^^>
sphalerite: thx
<pie_>
sphalerite: thanks for the linkfarm suggestion though, it made me clean up my stuff a bit and in the meantime i managed to think of ways to avoid all the recompiles
<angerman>
why can't my preStart part of my service create /var/log/<name> and keeps getting permission denied? I see other services using the same mkdir approach.
zupo has joined #nixos
<typetetris>
After adding an enty to my user flake registry, using this flake it gets downloaded from github once in a while. Used a branch in the flake registry entry. Will that stop, If I use a commit sha? (As it should be unchanging then?)
thc202 has joined #nixos
s34m has joined #nixos
s34m has quit [Client Quit]
<niksnut>
typetetris: yes
<niksnut>
well apart from garbage collection
<typetetris>
What about that `nix registry pin` thing? It forced me to call `nix registry pin github:NixOS/nixpkgs/nixos-20.09` instead of `nix registry pin flake:<my-symbolic-name-here>` that is kind of ... distrurbing? Can I only have on entry referring to nixpkgs/nixos-20.09 in my flake registry?
<typetetris>
Pinned I mean.
joko has joined #nixos
jonatanb has quit [Remote host closed the connection]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<etu>
jakobrs: If it's a set of packages it serves a purpose to use callPackage
<jakobrs>
I could use `makeScope`, but none of the packages depend on each other, so I'm wondering if I should just use callPackage directly?
as3ii has quit [Ping timeout: 264 seconds]
<jakobrs>
... I'll try to explain that again
werner292 has joined #nixos
<jakobrs>
`libretro` is a scope
<jakobrs>
As in, it's a set of packages
<jakobrs>
And I'm just wondering if it would be better to pass callPackage directly to `libretro` instead of using makeScope since makeScope might be more than is necessary for this
asbachb has joined #nixos
<Boomerang>
`with` has other footguns. Anyone tried `with pythonPackages`? unfortunately there is a `nixpkgs python package in there :face-palm:
werner292 is now known as werner291
werner291 has quit [Ping timeout: 272 seconds]
<jakobrs>
Boomerang: It's why I'm never quite sure if `.withPackages (pkgs: with pkgs; [ ... ])` is safe
<jakobrs>
Because one of the packages in `pkgs` might already be in scope, but with a different definition
<asbachb>
Hi. I've a laptop with full disk encryption. But my usb keyboard is not availble when I need to type in the password.
<asbachb>
I guess I need to add some additonal kenel module to initrd right?
<{^_^}>
[nix] @edolstra pushed 0 commits to refs/tags/2.3.9: https://git.io/JkgY9
<{^_^}>
[nix] @edolstra pushed 4 commits to latest-release: https://git.io/JkgYH
<etu>
asbachb: My USB-keyboard usually works at the bootloader but not right away when I should type the password, but after a second or two it wakes up and starts working. So, have you tried waiting a second and see if you get some kernel log and then it works?
vidbina has joined #nixos
mallox has quit [Quit: WeeChat 2.9]
<asbachb>
etu: I guess I already waited several seconds. Could you check your `boot.initrd.availableKernelModules` in hardware.nix?
<asbachb>
I guess the problem is that I installed NixOS without usb keyboard and maybe that's the reason it's not listed in availableKernelModules
mallox has joined #nixos
<etu>
asbachb: [ "xhci_pci" "nvme" "usb_storage" "sd_mod" ] on one laptop and [ "nvme" "ehci_pci" "xhci_pci" "usb_storage" "sd_mod" ] on the other one
<asbachb>
mh
<asbachb>
I've the same.
<asbachb>
Maybe it's because I'm using an usb hub.
<etu>
asbachb: A trick that might help you is to boot the computer without the keyboard, run lsmod and dump it to a file, plug in, run lsmod and dump to a second file to then diff the files.
gxt has quit [Remote host closed the connection]
mallox has quit [Quit: WeeChat 2.9]
<magnetophon>
I'm trying to make room on my hd. When I run nix-store --gc --print-roots I get a lot of paths like: /proc/976/maps What are they, and how do I get rid of them?
asbachb has left #nixos [#nixos]
gxt has joined #nixos
<{^_^}>
[nixpkgs] @adisbladis opened pull request #104374 → dockerTools: Always cross compile for another arch in the cross example → https://git.io/Jkg3S
Jackneill has quit [Ping timeout: 256 seconds]
<{^_^}>
[nixpkgs] @vaibhavsagar opened pull request #104375 → haskellPackages.hakyll: make buildable by Hydra → https://git.io/Jkg3F
<thibm>
magnetophon: if guess it is the programs currently running
<thibm>
they should not be gc'ed if they are running :)
<magnetophon>
thibm, right. thanks.
<thibm>
(more precisely, it's the memory maps of the program currently running, so all binaries, shared libs, etc)
<axx>
hi, regarding NFS, does one have to bind mount the directories one wants to export to /exports, as the wiki says? can't one just just export directories directly? I'm trying to do just that
<axx>
but seeing permission errors when starting the service
<axx>
i though the service was started as root, and therefore had access to whatever directories are being exported, but apparently not?
<gchristensen>
rogerr: I dunno, I don't have .pcsc10 so something you've configured or executed has created it
<thibm>
I'd say to use pkg.something for style. passthru attrs is merged into pkg
<thibm>
s/merged/updated (with the // operator) to be precise
zupo has joined #nixos
iH8c0ff33 has joined #nixos
<jakobrs>
I just ran into the issue with `with` scopes being weird
<jakobrs>
So it does happen in practice
Mateon2 has joined #nixos
Mateon1 has quit [Ping timeout: 264 seconds]
Mateon2 has quit [Remote host closed the connection]
Mateon1 has joined #nixos
orivej has quit [Ping timeout: 260 seconds]
iH8c0ff33 has quit [Ping timeout: 256 seconds]
werner292 has joined #nixos
jonatanb has joined #nixos
werner292 is now known as werner291
<evils>
weird, `nix-channel --update && nix-env -iA nixpkgs.firefox-wayland && nix-env -u firefox` results in 2 versions being installed, (firefox-82.0.3 and then "upgraded" to firefox-83.0)
<jakobrs>
evils: It's presumably "upgrading" to nixpkgs.firefox instead
<jakobrs>
which has a newer version and the same name
<jakobrs>
This is why you shouldn't use "nix-env -u" in general
kc[m] has left #nixos ["User left"]
fendor_ is now known as fendor
<evils>
jakobrs: i guess it should be nix-env -u firefox-wayland (that's not an option); any idea why nixpkgs.firefox-wayland isn't 83?
<jakobrs>
> firefox-wayland.version
<{^_^}>
"83.0"
<evils>
well, it should just be firefox :P
<jakobrs>
What does firefox-wayland.version say?
<jakobrs>
in nix repl
<evils>
on nixpkgs-unstable, firefox, firefox-wayland and firefox-unwrapped all are 82.0.3
hlisp has joined #nixos
<evils>
so why is the channel giving me 83...
hlisp has quit [Remote host closed the connection]
<jakobrs>
Do you perhaps have multiple channels installed?
<jakobrs>
nix-channel --list
<evils>
i do
<evils>
but i installed firefox-wayland from nixpkgs-unstable
<jakobrs>
maybe nix-channel --update is changing to a different channel
cfricke has joined #nixos
<jakobrs>
nix-channel --update doesn't care about where the package is from, it just looks for packages with the same name and a newer version
<evils>
yea, i'm getting it from nixos-unstable-small
<evils>
jakobrs++
<{^_^}>
jakobrs's karma got increased to 3
<jakobrs>
The solution is just not to use nix-env -u
<jakobrs>
thank you
<evils>
i started using nix-env for firefox when i was adviced to update to the latest firefox version because of a security release...
<evils>
which is also why i have the unstable-small channel xD
<evils>
i guess the right solution would have been to pin firefox to the newer channel in my configuration.nix until that version hit my usual channel?
<jakobrs>
Yes I think so
<jakobrs>
Or you could just do `nix-env -iA nixos-unstable-small`
<evils>
also, not everyone is on nixos so nix-env should still be able to deal with this...
<jakobrs>
Or you could just do `nix-env -iA nixos-unstable-small.firefox-wayland`
bbarker has joined #nixos
<evils>
i should just remove that channel, thought there was no harm in keeping it
<evils>
interesting that you could make your own channel that just claims to have version 99.9999 for every package and really screw someone :P
jakobrs has left #nixos ["WeeChat 2.9"]
<{^_^}>
[nixpkgs] @peti pushed 1000 commits to haskell-updates: https://git.io/Jkg4D
<{^_^}>
[nixpkgs] @zimbatm pushed commit from @raboof to staging « selinux: 2.9 -> 3.0 (#104087) »: https://git.io/JkguU
<thibm>
CptCaptain: Apparently it has a few years ago: systemd-boot do not seem to support specialisation :(
<CptCaptain>
Ok, thanks. I'll probably will have to switch to grub. I don't remember why I'm not using it as it seems to be the default. Maybe it's because of my luks setup
<thibm>
So you have 3 options: 1) use specialisation and switch to grub 2) use profiles 3) use specialisation and implement the support in systemd-boot ;)
<thibm>
I think it's only the default on BIOS systems
<{^_^}>
[nixpkgs] @adisbladis merged pull request #104374 → dockerTools: Always cross compile for another arch in the cross example → https://git.io/Jkg3S
<{^_^}>
[nixpkgs] @adisbladis pushed to release-20.09 « dockerTools: Always cross compile for another arch in the cross example »: https://git.io/Jkgwx
<rogerr>
how can i resolve a package to see where on the FS it's installed pls? wanna inspect its dir structure
Jackneill has quit [Ping timeout: 264 seconds]
<thibm>
rogerr: do you a program in your $PATH?
bbarker has joined #nixos
<rogerr>
?
<thibm>
What did you mean by "resolve" and "package"?
karuk has quit [Remote host closed the connection]
<rogerr>
i wanna see where pkgs.gnome3.adwaita-icon-theme is installed
<exarkun>
pretty sure I have seen this before but I can't remember what caused it or how to fix it. "error: cannot add path '/nix/store/...' because it lacks a valid signature"
Ariakenom has quit [Ping timeout: 246 seconds]
<thibm>
rogerr: The easiest is to do in the repl (nix repl '<nixpkgs>') `:b gnome3.adwaita-icon-theme` (ensure that <nixpkgs> points to the right nixpkgs from which the theme was installed
<thibm>
rogerr: you can also expect ~/.nix-profile and look for reference to it (like ls -lR | grep adwaita) (if you installed it with nix-env)
<thibm>
same for /run/current-system/sw/ if you installed it within nixos configuration
vandenoever has quit [Read error: Connection reset by peer]
<rogerr>
ty
vandenoever has joined #nixos
vandenoever has joined #nixos
vandenoever has quit [Changing host]
jonringer has joined #nixos
<thibm>
the latter option is better if you're not sure from which nixpkgs revision you installed the theme
jonatanb has joined #nixos
hnOsmium0001 has joined #nixos
saschagrunert has quit [Quit: Leaving]
<rogerr>
thibm++
<{^_^}>
thibm's karma got increased to 6
<rogerr>
using home manager when i comment out `home.packages = [pkgs.adwaita-icon-theme];` and add `gtk.iconTheme = { package = "pkgs.gnome3.adwaita-icon-theme"; name = "Adwaita"; };` the cursor doesn't work what i doing wrong pls?
<thibm>
exarkun: using nix-copy-closure with a non trusted user?
Jackneill has joined #nixos
<exarkun>
ah right. and setting SSH_USER=root until I can upgrade morph to a version that supports targetUser is the fix
<exarkun>
ty
* exarkun
writes it down this time
<thibm>
rogerr: you did you put quotes around the pkg?
<NobbZ[m]>
omnipotententity you could try to :b it and then manually browse the resulting store path
sangoma has quit [Read error: Connection reset by peer]
growpotk- has quit [Ping timeout: 240 seconds]
cosimone has joined #nixos
orivej has joined #nixos
kim0 has joined #nixos
<omnipotententity>
I guess I'm not really sure what I'm expecting, I'm trying to override this package, using overrideAttr, and I'm trying to figure out what I need to override, which isn't straight forward given the structure of the input code.
<whatisRT>
I'm having a weird problem: when I run `nixos-rebuild switch` I get the warning `warning: the following units failed: NetworkManager-wait-online.service` which seems to prevent the switch. The full error is this: https://dpaste.org/E3ck
hlisp has quit [Ping timeout: 240 seconds]
bbarker has quit [Ping timeout: 256 seconds]
<whatisRT>
I can work around this by setting `networking.networkmanager.enable = false;`, rebuild, and set it to true again, with another rebuild
<porphyrogenetos>
I'm maintaining a library in nixpkgs for telescope/astrophotography gear control (indilib), and I'm trying to add support for optional libraries contained in a different repo (same project but they split the repo off that has stuff like DSLR/GPS control). Repo 2 needs repo 1 to already be installed so I tried doing seperate packages. That didnt work because the way repo 2 hooks into repo 1 is by adding its entries to
<porphyrogenetos>
an XML file in /nix/store/repo1path/share/. Obviosuly repo 2's derivation cant modify another nix derivations output, so im guessing i'll need to incorporate indi-3rdparty into the indilib derivation as optional features. Anyone know where to point me for documentation of similar situations, or have tips?
<jasom>
Seems to be limited to emoji, it renders e.g. this just fine: ᚻᛖ ᚳᚹᚫᚦ ᚦᚫᛏ ᚻᛖ ᛒᚢᛞᛖ ᚩᚾ ᚦᚫᛗ ᛚᚪᚾᛞᛖ ᚾᚩᚱᚦᚹᛖᚪᚱᛞᚢᛗ ᚹᛁᚦ ᚦᚪ ᚹᛖᛥᚫ
Acou_Bass has quit [Ping timeout: 260 seconds]
kenshinCH has quit [Ping timeout: 260 seconds]
Acou_Bass has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « haskell: hls-brittany and brittany are the same package in our environment »: https://git.io/Jk2qQ
<tobiasBora>
Hello,
jonatanb has joined #nixos
<jasom>
And it's not a font fallback issue either; I can set it to a font that has only a very small number of characters and yet it still renders CJK just fine
hopheynananey has quit [Quit: WeeChat 2.9]
hlisp has joined #nixos
<jasom>
perhaps it won't render non-b&w characters?
ml| has quit [Ping timeout: 246 seconds]
stree has quit [Quit: Caught exception]
stree has joined #nixos
cosimone has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage2nix: update list of broken packages »: https://git.io/Jk2ms
s34m has quit [Quit: s34m]
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/Jk2mC
meh` has joined #nixos
hlisp has quit [Ping timeout: 246 seconds]
bbarker has quit [Remote host closed the connection]
<ajs124>
Izorkin: I'm not sure. If they actually disable building it, then we should probably not patch to re-enable.
<ajs124>
It just wasn't clear to me, from their documentation.
<ajs124>
That just says "not included in the package", which can mean a lot of things.
<tobiasBora>
I'm patching binaries using autopatchelfhook, but I get an error about libgomp library missing. Any idea how to get it? It seems to appear in gcc, but adding gcc in the input is not enough apparently
<jasom>
It is indeed the fact that only the text presentation shows up; the second not the first of this: 🕹️ 🕹. Note that on Ubuntu 18, these characters are displayed (in e.g. evolution but both with the text presentation)
<{^_^}>
[nixpkgs] @juselius opened pull request #104408 → nixos/kubernetes: add extraSANs to kubelet TLS cert. → https://git.io/Jk2md
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage2nix: update list of broken packages »: https://git.io/Jk2mN
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/Jk2Ye
konobi has quit [Remote host closed the connection]
diamondbond has joined #nixos
<tobiasBora>
cole-h: seems to remove the error, thanks! Also, I've the feeling that my solution (trying to patch every single library) is not the good one. Indeed, the archive is supposed to be pretty self contained (following what happens with appImages), but the input file is a bash file, so I can't use appImage-run. Any idea if it's possible to put at the same time all the stuff of appimage?
<diamondbond>
plasma or gnome live image? (ive only seen people using plasma online)
cosimone has quit [Quit: cosimone]
ml| has joined #nixos
hlisp has joined #nixos
supersandro2000 has quit [Quit: Ping timeout (120 seconds)]
supersandro2000 has joined #nixos
<tobiasBora>
oh, just adding appimage-run as input seems to give a not too bad result, let's wait and see
<tobiasBora>
not good enough missing again the libXrender.so.1-like stuff :-(
acarrico has joined #nixos
Morfio has quit [Quit: This computer has gone to sleep]
kenshinCH has joined #nixos
hlisp has quit [Ping timeout: 240 seconds]
Morfio has joined #nixos
<{^_^}>
[nixpkgs] @jonringer opened pull request #104411 → [staging] python3.pkgs.pipInstallHook: don't warn on script installation → https://git.io/Jk2Y5
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « haskell: hls-brittany and brittany are the same package in our environment »: https://git.io/Jk2Oe
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/Jk2Oq
meh` has quit [Ping timeout: 265 seconds]
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage2nix: update the list of broken packages »: https://git.io/Jk2O5
<turion>
How would I make a GHC/Haskell in my shell.nix aware of a custom unlit preprocessor like markdown-unlit?
hlisp has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/Jk23J
<turion>
If I just make overrideAttrs on the shell expression and add markdown-unlit as buildInputs, I have markdown-unlit in the shell, but GHC doesn't find it
<turion>
When I do haskellPackages.override, and in there ghc = super.ghc.overrideAttrs ... and add markdown-unlit there, then it says that there is infinite recursion
<{^_^}>
[nixpkgs] @peti pushed 42 commits to haskell-updates: https://git.io/Jk23c
<{^_^}>
[nixpkgs] @jonringer pushed 2 commits to release-20.09: https://git.io/Jk2sz
captn3m0- has quit [Ping timeout: 256 seconds]
hlisp has quit [Ping timeout: 265 seconds]
FRidh has quit [Quit: Konversation terminated!]
lsix has quit [Ping timeout: 260 seconds]
heywoodlh has joined #nixos
heyitsrama has joined #nixos
<turion>
Lol my infinite recursion came because I wrote super: self: and not self: super:, it seems
meh` has joined #nixos
<lordcirth>
lol oops
<lordcirth>
That overlay syntax is very confusing to me
werner291 has joined #nixos
<andi->
it is just a function with two arguments ;-)
<andi->
how hard can it be... but I've been there..
<turion>
It is confusing. I can never remember the order
<andi->
it is like python, self comes first
<lordcirth>
It's not really that part that is confusing, it's how to return stuff. I did finally get a package in an overlay, but I have no idea how to add in my service module now
<turion>
Can you change services in a overlay?
hlisp has joined #nixos
<turion>
Well, I'd like to know how to change the GHC derivation in a cabal2nix
<ajs124>
lordcirth: you just use the package
<ajs124>
it's overlayed. or you give it a new name, then use that name.
<ajs124>
turion: no, but you can use import and disabledModules
<lordcirth>
ajs124, I don't know what you mean by "just use the package" for a service?
<lordcirth>
I *think* you can add services in a overlay, as supposedly it's a change to the base nixpkgs
<ajs124>
what kind of service are we talking about? are you writing your own module? are you using a module from nixpkgs? neither? both?
bbarker has joined #nixos
<lordcirth>
ajs124, I made a package for Foundry, and then I wrote a service module to set it up as a service
<lordcirth>
I had it working on my VPS by importing it in configuration.nix, but then I switched to flake, and I haven't managed to get the service added properly yet.
<ajs124>
you can just put a module in the import = [] array of any file that makes up the configuration of the relevant system
<lordcirth>
So modules are in nixpkgs but overlays on nixpkgs can't overlay modules. This is why overlays are confusing
<lordcirth>
ajs124, thanks, I will try that
<turion>
TIL There is a section build-tool-depends in the .cabal file and when I add markdown-unlit there then cabal2nix will make markdown-unlit available at the right stage
<CitadelCore2>
Hi all. Trying to use the master branch of home-manager with a flake nixos configuration and I'm getting this error:
<CitadelCore2>
"The option `darwinModules' defined in `<unknown-file>' does not exist."
<CitadelCore2>
can see that it was only introduced on a commit not on the bqv-flakes branch, but I don't understand why it causes the expression to fail
<pushqrdx>
the mismatch between installing and uninstalling packages with nix-env is probably the biggest pain the butt aspect of nix xD
<pushqrdx>
like this can be perhaps fixed if each package had a metadata property that specifies -e name, for instance if a package is installed as _foo but needs to be erased with just foo
<pushqrdx>
package metadata migh contain "erase-name": "foo" or something along these lines
bostjan has joined #nixos
<ryantrinkle>
pushqrdx: i think it might actually make sense to not even have the name-based interface
davidv7 has joined #nixos
<{^_^}>
[nixpkgs] @Ma27 pushed commit from @magnetophon to release-20.09 « neomutt: 20200925 -> 20201120 »: https://git.io/Jk24R
<ryantrinkle>
most of the folks I know use attribute paths exclusively
<ryantrinkle>
which is also what you've gotta do if you're using, e.g. home manager
<ryantrinkle>
(i'd highly recommend home manager, btw!)
<pushqrdx>
ryantrinkle haven't yet felt the need for home-manager but it's on my checkout list. the list of packages i have in nix-env is < 3 items
<pushqrdx>
i mainly use nix-env as a test drive, but have all peristant user packages in configuration.nix under my user
<samueldr>
and ZFS (among others) can't do swap files
<srid>
Well, root is ext4 par, on a SSD in the thinkpad carbon.
<samueldr>
"but wouldn't a swap file be slower?" nah, Linux has a fast path for that
<gchristensen>
(nor can zfs hibernate)
<lordcirth>
Hibernation is obsolete anyway
<srid>
I don't think I care much about hibernation; suspend is fine.
hlisp has quit [Ping timeout: 260 seconds]
<srid>
So - swap file or partition? Does nixos have declarative configuration for swapfile?
<samueldr>
the advantage of swap files is they're quite easier to grow and shrink
<srid>
Is there a downside to using swap file?
<samueldr>
yes, declarative swap files
<samueldr>
AFAIK no downside, except for hibernation
<samueldr>
in fact I know of an upside: LUKS encrypted ext4 means your swap file is encrypted
<samueldr>
(without extra work)
<srid>
16G is all I got in RAM on this baby laptop; I was surprised I did not enable swap in last install; might explain ghcide thrashing.
<srid>
(P71 on the other hand has 64G, so I never enabled swap on that)
<samueldr>
swap is good even with high amount of RAM [citation needed]
<srid>
Oh, during this reinstall - I'd like to use the opportunity to use full disk encryption. Is that straightforward today in nixos?
<samueldr>
but not the cargo-culted "2 times or 1.5 times"
<samueldr>
srid: from the moment I started using NixOS it was already straight forward enough...
<samueldr>
... but I already knew about LUKS from archlinux
<srid>
I think I rarely saw myself using >50% of that 64G ram
<srid>
I know nothing about LUKS. /me peeks nixos docs
<samueldr>
yeah, AFAIUI even with that low use, there are things that can't happen without swap [citation still required :)]
<gchristensen>
samueldr: I think the 1.5x is so you have some usable swap while still having enough space to fully fit all your ram to disk
<samueldr>
yeah
<samueldr>
srid: the worst you can do with a wrong LUKS setup in your case, since you're starting fresh, is making a system that you can't boot I guess
<samueldr>
when I install a fresh NixOS I still look at this section because it doesn't happen often enough for me to remember
rprije has joined #nixos
<samueldr>
and "Are you sure? (Type uppercase yes):" don't be a smartbutt, and don't input "uppercase yes" ;)
<srid>
samueldr: when LUKS encrypting root, do you enter encryption password everytime on boot?
<samueldr>
yes
<samueldr>
I mean, *I* do
<samueldr>
there are other ways to decrypt which I've not explored
<srid>
And I'll choose ext4, unless there is a reaso to try something new ...
<lordcirth>
Password is simplest
zupo has joined #nixos
<srid>
I got the google titan key https://cloud.google.com/titan-security-key/ - which works well on nixos, fwiw - though not sure if that's gonna work with LUKS during boot time
<Yaniel>
I'd expect it to
<samueldr>
you can enroll additional methods to encrypt/decrypt after the fact with LUKS
<Yaniel>
AFAIK yubikeys do work too so
Jackneill has quit [Ping timeout: 256 seconds]
<samueldr>
(and change/remove the passphrase without needing a reencrypt, and still, you can reencrypt if needed)
<samueldr>
the last parameter is the name it will be mounted as, here it's `crypted`
Jackneill has quit [Ping timeout: 240 seconds]
<samueldr>
yeah, you'll need to `luksOpen` that container
<srid>
My bad; I really should be paying close attention
<samueldr>
no worries
<srid>
(The docs should really have the colors flipped; gray for output; and black for commands)
<samueldr>
the "commented" styling of those lines is pretty bad
<srid>
gray indicates code comments, by habit fo rme
<samueldr>
it's because it's using a scripting styling for that part of the docs, while in reality it's a command line session, which AFAIK there's no sytling for
<srid>
Okay, so from now - I should be using /dev/mapper/crypted instead of /dev/nvm... for commands like `mount` and `dd`
<samueldr>
yeah
meh` has quit [Ping timeout: 240 seconds]
<samueldr>
the partition holds the encrypted data container; luksOpen presents it as a block device under /dev/mapper/ (the filesystem would also be available as usual under /dev/disks/by-*)
astylian has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @jonringer pushed commit from @OmnipotentEntity to master « python3Packages.tensorflow_2: 2.3.0 -> 2.3.1 »: https://git.io/Jk2zn
<samueldr>
srid: yeah, the mapper name is from the "devices mapper" thing from the kernel, it maps to a less user-friendly name that could change (dm-0)
<samueldr>
you should probably change it to the mapper name
<samueldr>
otherwise it becomes dependent on ordering
<srid>
Esp if I'm to to create a second crypted partition, okay
<srid>
As an aside, the X1 Carbon is great for NixOS except for this annoying bug: https://bugzilla.kernel.org/show_bug.cgi?id=203709 - I think it might just be in the latest kernel, so I'm gonna use the 5.4 kernel in 20.09 this time (at the expense of losing support for 5k monitor resolution via thunderbolt)
LilleCarl has joined #nixos
growpotkin has joined #nixos
<fzakaria>
What's the solution for a JRE for Java11 ? The full JDK is a bit too big for just the runtime environment.
lsix has quit [Quit: WeeChat 3.0]
<LilleCarl>
Q: How soon in your Nix journey did you move to NixOS as your primary OS?
<fzakaria>
ah the java documentation has been updated to explain.
gustavderdrache has quit [Quit: Leaving.]
<{^_^}>
[nixpkgs] @LaloHao opened pull request #104426 → haskellPackages.binary-search: unbreak and patch → https://git.io/Jk2gH
<{^_^}>
[nixpkgs] @freezeboy opened pull request #104427 → obs-ndi: migrate to Qt5.15 → https://git.io/Jk2gN
Morfio has quit [Quit: This computer has gone to sleep]
pittma has quit [Ping timeout: 260 seconds]
evck_ has quit [Ping timeout: 260 seconds]
evck_ has joined #nixos
pittma has joined #nixos
jbetz has quit [Ping timeout: 260 seconds]
trevorriles has quit [Ping timeout: 260 seconds]
illustris has quit [Ping timeout: 260 seconds]
trevorriles has joined #nixos
jbetz has joined #nixos
illustris has joined #nixos
Morfio has joined #nixos
bbarker has joined #nixos
<{^_^}>
[nixpkgs] @freezeboy opened pull request #104428 → massif-visualizer: migrate to Qt5.15 → https://git.io/Jk22h
civodul has quit [Quit: ERC (IRC client for Emacs 27.1)]
growpotk- has joined #nixos
<DigitalKiwi>
how do i make a haskell program in nixpkgs use a different ghc (or whatever i need to do to fix it) now that we use the latest and greatest and least supported ghc by default
<{^_^}>
[nixpkgs] @rycee pushed commit from @jerith666 to master « eclipses: 2020-06 -> 2020-09 »: https://git.io/Jk2am
<DigitalKiwi>
oh do i just put it in the versioned configuration file instead of the common one...?
<infinisil>
DigitalKiwi: pkgs.haskell.packages.ghcXXX.<package> can be used
philr has joined #nixos
bbarker has quit [Remote host closed the connection]
bbarker has joined #nixos
hlisp has joined #nixos
<DigitalKiwi>
nobody else has used that in nixpkgs
sss2 has quit [Quit: Leaving]
growpotk- has quit [Ping timeout: 260 seconds]
bbarker has quit [Ping timeout: 260 seconds]
<infinisil>
Ah yeah I guess for nixpkgs it's no good
hlisp has quit [Ping timeout: 265 seconds]
<infinisil>
DigitalKiwi: Oh actually elm uses that
mananamenos has quit [Ping timeout: 256 seconds]
<tokudan>
if I have a list in an attribute, is there a way to dump that list as a nix-compatible string? reason: I want to build a configuration.nix for a system that can be later copied to /etc/nixos/configuration.nix that has network already configured.
<DigitalKiwi>
matterhorn complains about this base >=4.8 && <4.13 and glirc complains about happy (well, a dependency of glirc, config-value)
hlisp has joined #nixos
zecnate has joined #nixos
<zecnate>
Is there a better way to do this that doesn't involve a temporary file? `echo 'import (builtins.fetchTarball "https://github.com/timbertson/nix-wrangle/archive/v1.tar.gz")' > nix-wrangle-install.nix && nix-env -if ./nix-wrangle-install.nix && rm ./nix-wrangle-install.nix`
<zecnate>
Why doesn't `nix-env --install` support `--expr`?
<{^_^}>
[nixpkgs] @freezeboy opened pull request #104432 → partition-manager: migrate to Qt5.15 → https://git.io/Jk2Vp
<{^_^}>
[nixpkgs] @primeos pushed to master « chromiumDev: Unmark the build as broken »: https://git.io/Jk2wC
jonatanb has joined #nixos
<rogerr>
i add keybase-gui to home.packages then set services.kbfs.enable true. but i notice kbfsfuse and keybase run whether keybase-gui is running or not. can i make the daemons run on-demand instead pls?
erasmas has quit [Quit: leaving]
cr4y1_ has quit [Read error: Connection reset by peer]
jonatanb has quit [Ping timeout: 246 seconds]
neiluj has quit [Ping timeout: 265 seconds]
hlisp has joined #nixos
<hexa->
what's the difference between fetchurl and fetchpatch?
<hexa->
I mean … I can use both to fetch a patch, right?
<infinisil>
(I wouldn't have known without checking the source btw)
<hexa->
but you knew to check :)
cr4y1 has joined #nixos
<hexa->
sorry for being lazy, I guess :D
jbox has joined #nixos
konobi has quit [Client Quit]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
konobi has joined #nixos
<infinisil>
No problem!
hlisp has joined #nixos
noudle has joined #nixos
<rogerr>
i use i3 and keybase-gui. how can i make the keybase and kbfsfuse daemons only run when i run keybase-gui? right now they run always. i think it needs an option for "keybase ctl autostart --disable"?