<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JfnsF
<colemickens>
I love it.
<Raito_Bezarius>
Anyone tried to iPXE install NixOS on Scaleway?
proofofkeags has quit [Remote host closed the connection]
<Raito_Bezarius>
The image data was downloaded, but it seems stuck with no output now :/
proofofkeags has joined #nixos
<Raito_Bezarius>
I can re-attach the serial console, but there is still no output, I don't know if there are some ways of debugging that
u_u has quit [Ping timeout: 272 seconds]
<samueldr>
Cadey++ this is so dumb this is such a great example of fun shenanigans Nix and NixOS allows you to do :)
<{^_^}>
Cadey's karma got increased to 4
<Cadey>
samueldr: IKR, this is crazy, i love it
<Ashy>
my boot process is taking a long time because it's waiting for network-setup.service to complete fully before continuing, is there a way to control that?
<colemickens>
its a real thing though! see kioskix. I think these examples of how to use nixos like a little os toolkit are super powerful demos
vidbina has quit [Ping timeout: 246 seconds]
proofofkeags has quit [Ping timeout: 260 seconds]
<Ashy>
actually, `systemd-analyze critical-chain` shows that it's docker.service as the root cause of the delay
<Cadey>
colemickens: i was actually going to make a somewhat normal kiosk demo, but i think this way is better
<colemickens>
sorry, that's what I meant
<Cadey>
boot2mario
<colemickens>
:)
<colemickens>
I wound up just hawking some stuff from nixiosk. It requires custom nixpkgs last time I checked. Or enough commits I didn't want to carry them all
<pistache>
Raito_Bezarius: I haven't used PXE install on Scaleway, but I have installed NixOS successfully from the Debian in rescue mode
morgrimm has joined #nixos
<colemickens>
I'm doing sway+plex-mpv-shim but I'm blocked on some rpi4 stuffs :S
<Raito_Bezarius>
pistache: I can do this but I want to have some kind of reproducible way to spawn NixOS instances on Scaleway
<Raito_Bezarius>
Manually doing the kexec trick is nice
sigmundv__ has quit [Ping timeout: 246 seconds]
<Raito_Bezarius>
But I'd like some automation
<Raito_Bezarius>
with justdoit
<Raito_Bezarius>
Ideally, having BTRFS would be nice, apparently, the Scaleway's kernels support it
<{^_^}>
[nixpkgs] @utdemir opened pull request #87154 → Some performance optimizations to dockerTools.build{,Layered}Image → https://git.io/JfnGZ
m0rphism has quit [Ping timeout: 260 seconds]
thc202 has quit [Ping timeout: 244 seconds]
morgrimm has quit [Ping timeout: 246 seconds]
chagra_ has quit [Ping timeout: 272 seconds]
Acou_Bass has quit [Ping timeout: 264 seconds]
chagra_ has joined #nixos
lemsip has quit [Ping timeout: 604 seconds]
magnetophon has quit [Ping timeout: 272 seconds]
emilsp has joined #nixos
virus_dave has quit [Ping timeout: 264 seconds]
Acou_Bass has joined #nixos
hlisp has joined #nixos
`_ has quit [Ping timeout: 264 seconds]
numkem has joined #nixos
`_ has joined #nixos
is_null has quit [Remote host closed the connection]
philr has joined #nixos
proofofkeags has joined #nixos
hlisp has quit [Ping timeout: 256 seconds]
jluttine has quit [Ping timeout: 256 seconds]
<{^_^}>
[nixpkgs] @bhipple opened pull request #87155 → rustPlatform: fix bug with ambiguous diff tool → https://git.io/JfnGx
philr__ has quit [Ping timeout: 264 seconds]
proofofkeags has quit [Ping timeout: 256 seconds]
hlisp has joined #nixos
jluttine has joined #nixos
morgrimm has joined #nixos
<charukiewicz>
Is there a builtins.* that you can use to easily filter the newline character produced by using builtins.readFile?
<charukiewicz>
that is, I read a single line file containing "dog" and readFile produces "dog\n", I want to filter the \n out
<Raito_Bezarius>
armin: use a suid on a script or backlight tool
<unclechu>
hey, about vim/neovim, following this wiki page https://nixos.wiki/wiki/Vim i realized that plugins are loaded after my `customRC`, can i somehow change this order?
<Raito_Bezarius>
clever: justdoit with zfs requires a specific kernel?
<unclechu>
to load plugins first and only then my `customRC`?
<Raito_Bezarius>
i'm trying to automatic justdoit on scaleway
<srhb>
pjt_014: So it looks like nixpkgs is coming from there, and is a duplicate of root's nixos channel
<srhb>
pjt_014: The other two are from symlins in ~/.nix-defexpr (probably a channels_root symlink, joining both of root's channels by name)
cole-h has quit [Quit: Goodbye]
<pjt_014>
so what'd be the ideal way to regenerate it (manual editing seems like a bad idea)?
<srhb>
No idea. I don't use channels. They're messy. I'd just point it at the channel you want to be your default one unless you explicitly manipulate NIX_PATH.
palo has quit [Ping timeout: 272 seconds]
palo1 is now known as palo
fabianhjr has quit [Quit: Leaving.]
butcher has quit [Ping timeout: 260 seconds]
<pjt_014>
you use flakes then? also I'm not sure I get it. Which of these would be redundant (?):
<srhb>
pjt_014: Take a look in that root channels path
<srhb>
You should see both of roots channels
<srhb>
pjt_014: Er, none of those two are redundant. :) It's the nixpkgs one and the unqualified defexpr one
<srhb>
I don't use flakes system-wide yet. I just have NIX_PATH=nixpkgs=~/src/nixpkgs -- which is a git checkout
<pjt_014>
okay, still not sure which one. Which line would it be here? http://0x0.st/i2Cc.txt
<srhb>
pjt_014: Oh, I didn't notice the last one before.
<srhb>
pjt_014: ls -l /home/josh/.nix-defexpr/ /nix/var/nix/profiles/per-user/root/channels
<srhb>
pjt_014: One of those duplicates /nix/var/nix/profiles/per-user/root/channels/nixos
<srhb>
Or both, probably.
<srhb>
pjt_014: It's just files and symlinks, you can explore them easily.
<pjt_014>
this what I got here: http://0x0.st/i2Cm.txt I'm guessing that the channels one in defexpr I don't want?
oborot has quit [Ping timeout: 272 seconds]
<srhb>
pjt_014: Up to you. /nix/var/nix/profiles/per-user/root/channels and channels_root are interchangeable.
drakonis has quit [Quit: WeeChat 2.8]
butcher has joined #nixos
<srhb>
pjt_014: But removing the defexpr one will affect nix-env, since that's the source of truth it uses, not NIX_PATH
<pjt_014>
hmmm...
<pjt_014>
I might just wait till I reinstall. Might be too much effort for now
<srhb>
:P
<pjt_014>
especially for something I can fix with a jq filter
<srhb>
Again, I'm a proponent of not using channels for anything but the simplest of basicaly one-channel cases. And in that case, both NIX_PATH and ~/.nix-defexpr should probably point ot root's channels for simplicity.
<pjt_014>
So in the future should I just only do package stuff w/ sudo? That's looking like the ideal answer here
<srhb>
pjt_014: No, but you should maybe only manipulate channels with sudo.
<srhb>
_using_ root's channels is no problem.
<srhb>
With the setup outlined above.
linarcx has joined #nixos
<spinlock[m]>
I have a pretty odd error with home-manager. Using `services.picom.enable = true;` results in this error: `The option `services.picom' defined in `/etc/nixos/user/home/picom.nix' does not exist.`
<spinlock[m]>
* I have a pretty odd error with home-manager. Using `services.picom.enable = true;` results in this error: `The option 'services.picom' defined in '/etc/nixos/user/home/picom.nix' does not exist.`
<spinlock[m]>
Is there a proper way to specify additional channels in your `configuration.nix`? I'm apparently too dumb to remember that I have to manually add the channels before `nixos-install` and then once again once I first login to my live system.
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « openfst: 1.7.4 -> 1.7.6 »: https://git.io/Jfn0a
<drozdziak1[m]>
Git is unable to sign commits on my machine. gpg is visible from my shell and can be used for manual operations, it is enabled in my `configuration.nix`. How can I debug this, or what may I be missing?
orivej has quit [Ping timeout: 264 seconds]
<srhb>
drozdziak1[m]: We probably need at least an error message, or anything really, to start having an idea what's going on.
<{^_^}>
[nixpkgs] @martinbaillie opened pull request #87165 → ssm-session-manager-plugin: init at 1.1.61.0 → https://git.io/JfnEv
mallox has joined #nixos
eoli3n_ has joined #nixos
smatting has joined #nixos
<drozdziak1[m]>
@srhb it's just "error: gpg failed to sign the data" and "error: failed to write commit object"
butcher has quit [Ping timeout: 256 seconds]
<srhb>
drozdziak1[m]: Wonder if there's some way to get more verbosity there.
<fps>
if my configuration.nix has a line like boot.kernelPackages = pkgs.linuxPackages_rpi4; how would i go about applying patches to that kernel package?
<drozdziak1[m]>
@srhb --verbose doesn't do anything for this purspose unfortunately
<srhb>
fps: boot.kernelPatches iirc
cr4y1 has joined #nixos
opthomasprime has joined #nixos
rardiol has joined #nixos
<hyper_ch>
gchristensen: did you know that datasets in zol also can start with a ":"? I just tested it so that I can reflect the root fs accordingly: https://paste.simplylinux.ch/view/raw/a74005e4
asheshambasta has joined #nixos
<pjt_014>
zol?
proofofkeags has joined #nixos
<pjt_014>
nvm googled it
phreedom has quit [Ping timeout: 240 seconds]
proofofkeags has quit [Ping timeout: 256 seconds]
dermetfan has joined #nixos
zakkor has joined #nixos
<zakkor>
how come the unrar package exists but the rar package doesnt?
<makefu>
Zakkor: i will try with the new derivation
<makefu>
i think you need to callPackage { enableUnfree = true; }
<fgaz>
Anyone knows what happens if I make a link to somewhere else in the sistem in /nix/var/nix/gcroots/per-user/$(whoami) ?
<makefu>
you also need to enable insecure packages for p7zip
petrkr has quit [Quit: Leaving]
<fgaz>
Or if I put there arbitrary stuff for that matter
<energizer>
,locate bin/rar
Mic92 has quit [Quit: WeeChat 2.7.1]
<{^_^}>
Found in packages: john, iputils, radare2, nettools, argus-clients, gnome2.rarian, r2-for-cutter, nettools_mptcp
<fgaz>
I'd try, but to test it I'd have to do a GC...
<energizer>
,locate bin/rar$
Mic92 has joined #nixos
<{^_^}>
Couldn't find in any packages
<fgaz>
energizer: it's unfree
<fgaz>
Hydra doesn't build it so the binary isn't in the index
<makefu>
fgaz: is it actually available?
lukash_|away is now known as lukash_
<fgaz>
unrar is
<fgaz>
Not sure about rar
<energizer>
unrar doesnt provide the rar binary
gagbo has joined #nixos
<pjt_014>
,locate p7z
<{^_^}>
Found in packages: p7zip, firejail, zafiro-icons, maia-icon-theme, moka-icon-theme, paper-icon-theme, qogir-icon-theme, iconpack-obsidian, papirus-icon-theme, flat-remix-icon-theme, numix-icon-theme-circle, numix-icon-theme-square
<spinlock[m]>
pretty funny that it has rar2fs but not rar
<spinlock[m]>
also, there is a patch to remove rar support from calibre. I wonder what the story is about rar 🤔
<pjt_014>
you gotta admit tho 7z is so much nicer to work with than tar/tar+gz. Really wish 7z could back up user/group data etc
knupfer has joined #nixos
<danderson>
spinlock[m]: rar is a proprietary file format, with no OSS encoders afaik
kvda has joined #nixos
<danderson>
there are some OSS decoders, but they don't support all of the format's features or versions
<makefu>
i tried to build the p7zip from master, however when i try to create a rar archive i get "E_NOTIMPL"
<spinlock[m]>
fair enough. I suppose it's only really used for newsgroups these days
<danderson>
meanwhile, there has been a lot of great compression research in the past several years, so there's not much advantage to using a proprietary format+program when there's plenty of OSS compressors that can beat it
hio has joined #nixos
<pjt_014>
yep. zpaq is crazy cool
hlisp has joined #nixos
hio is now known as hooo
<pjt_014>
backwards AND forwrads version compatibility is a neat trick
<danderson>
rar is weirdly popular with pirates, yeah. I don't understand why
<makefu>
danderson: maybe it is close to "ARR!"? (sorry, bad joke)
<danderson>
heh
<pjt_014>
maybe it's the meme
<fgaz>
How can it have forward compat without duplicating data?
<danderson>
I suspect it's just inertia. At some point in the distant past, rar was a great format to use.
<danderson>
now it's obsolete and weirdly proprietary in a world where that's unheard of.
<pjt_014>
there's an entire subreddit for madlads that have paid for winrar
<pjt_014>
doubt that's entirely where it comes from but its an idea
<pjt_014>
fgaz: zpaq has a language (paql) for speicifying compression algoritims
<danderson>
fgaz: guessing the decompressor is a fairly simple virtual machine, like most compression formats. If you make that VM flexible enough, you can do the hard stuff in compression and just generate a more fancy decompression program
<makefu>
Zakkor: so in short, you cannot pack rar with p7zip even when unfree is enabled
<pjt_014>
danderson: yep. the comp. language was frozen at version 1.0, so backwards AND forwards compat is guaranteed.
<fgaz>
danderson: almost like one of those self-extracting archives
<danderson>
fgaz: pretty much! It's a very interesting mind-twist to think of compressed files as a program for a specialized virtual machine
eoli3n__ has joined #nixos
<danderson>
at its core, it's a VM with 2 commands: "output the next N bytes" and "output the N bytes at file offset X"
<danderson>
with the second opcode, you can re-output the same bytes multiple times, and voila! You can "re-duplicate" stuff that was deduplicated by the compressor
<pjt_014>
Zpaq also has some other cool stuff going on, like transactional version managment and offline indexing
<danderson>
(not quite as trivial in practice, and the _compressor_ can get really complex... But at its core, that's all it is)
<fgaz>
Cool
<danderson>
but that also explains "zipbombs"
<pjt_014>
If only it could backup all the metadata that tar can then I'd replace tar with it like 2 years ago
Neo--- has quit [Read error: Connection reset by peer]
<wedens[m]>
at some point I started getting `Settings schema 'org.gtk.Settings.FileChooser' does not contain a key named 'show-type-column'` error in rambox when file chooser is being summoned. is it some nix specific issue?
maddo has joined #nixos
lycium[m] has quit [*.net *.split]
denni9th[m] has quit [*.net *.split]
boogiewoogie[m] has quit [*.net *.split]
kl3 has quit [*.net *.split]
M0x0120f[m] has quit [*.net *.split]
kdedev[m] has quit [*.net *.split]
faya[m] has quit [*.net *.split]
chip[m] has quit [*.net *.split]
kazzimazzi[m] has quit [*.net *.split]
f0i[m] has quit [*.net *.split]
pingiun[m] has quit [*.net *.split]
petersjt014[m] has quit [*.net *.split]
imj[m] has quit [*.net *.split]
ub[m] has quit [*.net *.split]
Smith[m] has quit [*.net *.split]
arianvp[m] has quit [*.net *.split]
brococoli[m] has quit [*.net *.split]
Florian[m]1 has quit [*.net *.split]
morr[m] has quit [*.net *.split]
OJ[m] has quit [*.net *.split]
chvp has quit [*.net *.split]
sonercirit[m] has quit [*.net *.split]
infinisil-m has quit [*.net *.split]
alexgood has quit [*.net *.split]
nolan_d has quit [*.net *.split]
icetan has quit [*.net *.split]
gsals[m] has quit [*.net *.split]
Albakham[m] has quit [*.net *.split]
JJJollyjim has quit [*.net *.split]
the_pumpkin_man[ has quit [*.net *.split]
mkg20001 has quit [*.net *.split]
AmitLevy[m] has quit [*.net *.split]
aterius has quit [*.net *.split]
denbrahe[m] has quit [*.net *.split]
dpc has quit [*.net *.split]
mica[m] has quit [*.net *.split]
edrex has quit [*.net *.split]
i_run_out_of_use has quit [*.net *.split]
watzon has quit [*.net *.split]
wak-work has quit [*.net *.split]
NickHu has quit [*.net *.split]
philipp[m] has quit [*.net *.split]
shizonic has quit [*.net *.split]
sigwinch28 has quit [*.net *.split]
Dandellion has quit [*.net *.split]
Ox4A6F has quit [*.net *.split]
dukedave has quit [*.net *.split]
higherorder has quit [*.net *.split]
Fulgen has quit [*.net *.split]
eddyb[legacy] has quit [*.net *.split]
sethetter_ has quit [*.net *.split]
chrisaw has quit [*.net *.split]
cbarrett has quit [*.net *.split]
joshmeredith has quit [*.net *.split]
noj has quit [*.net *.split]
Dagger has quit [*.net *.split]
Shados has quit [*.net *.split]
zgrep has quit [*.net *.split]
alanz_ has quit [*.net *.split]
bandali has quit [*.net *.split]
edcragg has quit [*.net *.split]
turq has quit [*.net *.split]
smatting has joined #nixos
<fgaz>
To answer my previous question, yes, it seems like I can store stuff in /nix/var/nix/gcroots. I wonder if I can rely on that property...
shah^ has quit [Ping timeout: 246 seconds]
<fgaz>
For anyone wondering, no, I'm not backing up my home in gcroots ;)
<fgaz>
My objective is to store some metadata together with the roots
<fgaz>
does `nix-store --realise --add-root` do anything besides creating the link in the case of an already realised path?
<lightandlight>
I find myself wishing for a command that evaluates a nix expression of type `path` and executes the referenced file. Does this exist? I can't get the behaviour I want with `nix-shell` or `nix run`
zupo has quit [Client Quit]
sigmundv__ has joined #nixos
<manveru>
lightandlight: what do you mean with execute?
<iqubic>
Does anyone know what kernel version the NixOS 20.03 installer ISO will give you on a fresh install? Is there some stable kernel I can be using?
Bryan[m] has joined #nixos
<lightandlight>
Given an arbitrary nix expression which will evaluate to a `path`, evaluate it and get the path and exec the path
<iqubic>
I'm currently using 4.19, and I'm wondering if there's a more recent stable kernel I could be using.
<manveru>
exec within or outside the nix sandbox?
<Bryan[m]>
I'd really like it if I could disallow building certain packages rather than pulling them from cache. E.g. every time ghc tries to build itself locally (like right now on the 20.03 channel for some reason), I'd rather Nix just gave up. I can go back and use an older version
<fgaz>
iqubic I think on 20.03 it's 5.x
<lightandlight>
I suppose within? Because, assuming the path refers to an executable in the nix store, it's possible that executable will look for dynamic links in the store too?
jumper149 has joined #nixos
<iqubic>
fgaz: I'd like to know the exact version, if possible.
<fgaz>
iqubic: check out the nixpkgs commit specified in the iso name and check
hmpffff_ has joined #nixos
<lightandlight>
I can "just use a shellHook in nix-shell", so I'm not blocked. But I don't want the shell. `nix-shell --run` could work, but I haven't figured out what the command should be
<iqubic>
fgaz: I'm not sure how to do that.
thc202 has joined #nixos
<fgaz>
Ah wait they don't put the commit in the name anymore
<fgaz>
There's even a link to the commit on github there
<iqubic>
Does that tell us which kernel version they're using?
__monty__ has joined #nixos
iqubic has quit [Remote host closed the connection]
iqubic has joined #nixos
<Bryan[m]>
lightandlight: given a foo.nix that generates an executable, `$ $(nix-build foo.nix)` works, since nix-build prints out the path that gets built
nabzi has quit [Quit: -a- IRC for Android 2.1.55]
<iqubic>
I love how quick it is to reboot a machine running on an SSD.
<fgaz>
iqubic: yes, search for "linuxPackages =" in pkgs/top-level/all-packages.nix
<iqubic>
I will do that.
gilligan has joined #nixos
<lightandlight>
Thanks Bryan[m], that's a little too specialised for my case though
<jonge[m]>
hi there. if i want to create a .nar archive from a derivation *in* a derivation (i.e. i get a NAR file as the derivation's output so i can download it from hydra), i get weird errors when running `nix-store --export $(list of files)`, as it tries to access not-existing paths like /nix/var etc.
<jonge[m]>
is there some builtin way to create NARs?
<bew>
Hello, what _really_ happens when I do `src = ./.` in a derivation ? does it computes the hash of every files recursively and combine them with the buildInputs to determine the final hash?
<{^_^}>
[nixpkgs] @peti pushed 351 commits to haskell-updates: https://git.io/Jfnoj
<brightone_>
Hi! I'm having trouble with Hydra and declarative projects - after .jobsets gets evaluated (where branches and PRs are), nothing appears in project's jobsets list.
<{^_^}>
hydra#751 (by filalex77, 1 day ago, open): Jobsets from new GitHub PRs are not appearing in jobsets list
<brightone_>
I would really appreciate any help, as it's very hard to work with Hydra by myself, having little documentation and almost nobody to ask help from. Thanks!
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « git-annex: update sha256 hash for new version 8.20200501 »: https://git.io/JfnKL
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<jumper149>
ToxicFrog: You don't have to keep your fork locally, as it's still on GitHub. Regarding merging, how long has it been since your PR has been approved?
dingenskirchen has quit [Quit: dingenskirchen]
dingenskirchen has joined #nixos
<ToxicFrog>
jumper149: 11 days and 4 days. And I do have to keep it locally because I'm using it in my configuration.nix.
<ToxicFrog>
Once it's upstream I can stop rebasing it against nixpkgs.
dingenskirchen has quit [Client Quit]
dingenskirchen has joined #nixos
<jumper149>
Are these just packages or some more internal changes?
<jumper149>
You could ask a maintainer that usually merges PRs speficically to review your PR. You could just go through some recent commits to find someone. I would advise you to make some changes to your configuration.nix probably though.
AlexOrder has quit [Remote host closed the connection]
<jumper149>
ToxicFrog: Have you considered using overlays?
<ToxicFrog>
I am well aware of overlays, yes, but I want these in nixpkgs and they need to be periodically updated to keep pace with it anyways, so
<ToxicFrog>
Also, AIUI, overlays can only add or modify packages, not add new nixos options.
hooo has quit [Quit: Connection closed for inactivity]
<exarkun>
on CI-as-a-service (eg circleci) I sometimes have nix builds get killed because (as far as I can tell) resources for the job (memory I guess) have been exhausted
<exarkun>
are there any better options for dealing with this situation than playing with --jobs and --cores?
<exarkun>
It seems like a shame to turn jobs down to 1 when 90% of builds could run 2 or 4 in parallel and there are just one or two that need the whole system to themselves
<mojjoo>
After switching to 20.03: `nginx: [emerg] cannot load certificate "/etc/letsencrypt/live/foo/fullchain.pem": BIO_new_file() failed (SSL: error:0200100D:system library:fopen:Permission denied:fopen('/etc/letsencrypt/live/foo/fullchain.pem','r') error:2006D002:BIO routines:BIO_new_file:system lib)`Rights seem to be correct. The file is a world
<mojjoo>
readble symlink to a world readable file.. Any ideas?
<exarkun>
What about permissions on the parent directories?
<mojjoo>
exarkun: nice, that worked. but now it complains about 'privkey.pem' which currently has only 'rw' for root which feels weird to change
<exarkun>
not sure about that, I haven't actually used nginx on nixos... I guess it sounds like something changed about who nginx runs as between 20.03 and whatever you upgraded from?
<exarkun>
could be a nginx change or a nixos change I guess
<exarkun>
is it correct that if I want to override a Python derivation to not run the check I have to pass `doInstallCheck = false;` to overrideAttrs?
splintah has joined #nixos
<exarkun>
I hope it is. It makes sense to me, I think. buildPythonPackage translates `checkPhase` and `doCheck` into `installCheckPhase` (if that's the right name) and `doInstallCheck` but if I want to override then I don't get that logic applied because it happened already
<kraem>
found weechat-matrix is packaged in nixpkgs. since the attribute doesn't exist in all-package.nix, do i need to pass it in as `extraBuildInputs` to weechat/default.nix in an overlay to activate the plugin?
KeiraT has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @romildo merged pull request #86519 → enlightenment.efl: 1.23.3 -> 1.24.0; new test module → https://git.io/Jf3l5
<emily>
(you want it in your systemPackages standalone as well for the matrix_* scripts, and the python plugin needs weechat-matrix in its path to pick up the dependencies)
<kraem>
emily: thanks! i was just wondering how i'd gain access to pkgs when having a standalone overlay file for weechat. ended up inserting your snippet in a module instead, works flawlessly.
<kraem>
emily: ah alright! am i understanding it correctly if self and super _are_ the pkgs structure so i could just as well access `self.tmux` in my weechat overlay?
butcher_ has quit [Ping timeout: 260 seconds]
<emily>
yes
<kraem>
jumper149: no i don't. i really should though, as i seem to be keen on making typos :P
numkem has joined #nixos
Fare has quit [Ping timeout: 272 seconds]
<kraem>
thanks emily for clearing that up :)
<azazel>
wasn't better to do a `super.tmux`?
jophish has quit [Quit: WeeChat 2.4]
<bdju>
ahhh I had "netsurf" as a package instead of "netsurf.browser". I wish the error had been more clear
<jumper149>
azazel: I agree. Use super when you want the no-overlay version and self only when you want the one from your overlay. Keeping self references to a minimum.
<jumper149>
Doesn't really matter for the output though ^^
<ryantm>
bdju: Usually packages are marked as broken because they don't build succesfully. If it builds okay or you fix the build, please open a PR to remove the broken flag.
linarcx has quit [Remote host closed the connection]
<bdju>
ryantm: it was just my first time encountering something marked broken rather than just happening to be broken
butcher_ has joined #nixos
<ryantm>
bdju: As part of doing a NixOS release, there is a pass done to mark everything broken that isn't building on Hydra. That's usually when things get marked broken.
Maxdamantus has quit [Ping timeout: 260 seconds]
eyenx is now known as tntr
Maxdamantus has joined #nixos
butcher_ has quit [Client Quit]
butcher has joined #nixos
tntr is now known as eyenx_m
eyenx_m is now known as eyenx
<jumper149>
bdju: Here you can look at the hydra builds. nixpkgs:trunk regards to nixpkgs master afaik.
<exarkun>
Maybe I should stop trying to integrate with circleci so closely and just write a shell script that gets invoked
bachp has quit [Quit: Idle for 30+ days]
<cole-h>
Oh, I think you need to specify an interpreter with -i (e.g. is the file a python script? `-i python3`. Is it bash? `-i bash`.)
<cole-h>
And sorry, I meant the FIRST shebang
<cole-h>
`#!nix-shell`, not the next one with the python packages
alienpirate5 has left #nixos ["Kicked by @appservice-irc:matrix.org : Idle for 30+ days"]
jgarte[m] has left #nixos ["Kicked by @appservice-irc:matrix.org : Idle for 30+ days"]
<cole-h>
exarkun: So line 2 of the script would be `#! nix-shell -p 'python.withPackages (ps: [ ps.codecov ])' -i bash` in your case (I think)
<exarkun>
I think circleci is taking the circleci-configured shell and splatting a #!<something> at the top of this script and then running it
<exarkun>
so `shell: "nix-shell"` makes it write "#!nix-shell" on top of whatever is in the command section
<cole-h>
Oh, alright -- that should be fine then.
<cole-h>
But you'll still want your shebang with the python stuff to include the `-i bash`
<exarkun>
is that not just the default?
butcher has quit [Ping timeout: 260 seconds]
fasd has quit [Quit: Idle for 30+ days]
Henson has joined #nixos
kraem has quit [Quit: outta here]
kraem has joined #nixos
<Henson>
I know with nixops it's possible to overlay the nixpkgs that you're using to define your network, but say I wanted to define a machine that had nixos containers inside it. I'm having trouble getting the overlaid nixos options that appear in the network definition scope to appear in the configuration scope of the container that's being defined.
<cole-h>
exarkun: No, it's not. If you run that script locally, it'll drop you into a `nix-shell`. If you add `-i bash`, it'll actually run the codecov stuff.
<simpson>
That is so confusing. Like, where is even the `-c` anywhere in that file?
<cole-h>
It's part of how nix-shell gets invoked as a shebang, I'd bet
<simpson>
That's what I suspect, too. Something about the newlines?
chloekek has joined #nixos
<CRTified[m]>
Hi, is anyone here who as an SDRPlay running on NixOS? I was given one (SDRplay RSPdx) and I'm looking for a place to start software-wise. I was able to start gnu radio after fixing the PYTHONPATH, but I guess that I won't get the SDRPlay running with it (as it looks like the api/driver is closed source)
<cole-h>
Oh
<cole-h>
lmao
<cole-h>
exarkun: Is a multiline command treated as a file in circleci?
lsix has joined #nixos
<exarkun>
what do you mean by "treated as"? I assume this stuff is written to a file and then executed from a shell but I don't really know
<exarkun>
(docs are not exactly written by geniuses)
<cole-h>
I think simpson is right on the money: I think it's treating that entire block as a single command, and the newlines are messing stuff up. I bet if you were to use `--run 'cp ....; python -m coverage combine; python -m coverage report; codecov --file .coverage`, things would work just fine
<cole-h>
(This is one of many reasons why I hate yaml)
<evanjs>
icetan: but yeah it seems to be a mix of misconfigured remotes/repos and jq dying because mavenix ended up pulling in a bunch of 404 pages (i.e. html). The "content" attributes under "metas" aren't properly escaped (?) and jq just can't
codygman has quit [Read error: Connection reset by peer]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<duairc>
Is it possible to override a whole set of configuration from a NixOS module? Let's there's an attrset at foo.bar with a bunch of configuration that's set by another module that I don't have control over. In my module, I want to override that and say "foo.baz = config.foo.bar" and then completely unset foo.bar. Can you do stuff like that?
smatting has joined #nixos
knupfer has quit [Remote host closed the connection]
knupfer has joined #nixos
<maddo>
so uh, what is the difference between hplip and hplipWithPlugin?
<maddo>
I mean, can my printer work without the hplipWithPlugin and hp-setup?
<typetetris>
There was a tool to keep nix-shells around and organize env vars with direnv, could someone remind me of its name?
<niso>
typetetris: lorry
<niso>
s/lorry/lorri
<typetetris>
thanks
zupo has joined #nixos
<dxtr>
Haha, crap. The configuration doesn't happen to be stored in /nix somewhere, right?
m0rphism has quit [Ping timeout: 240 seconds]
<qyliss>
Not by default
<dxtr>
Darn. Accidentally removed it :D
<dxtr>
Lost some work!
<dxtr>
Oh well, no biggie.
<dxtr>
I have backups from yesterday
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
dadada_ has joined #nixos
rogue_koder has quit [Ping timeout: 260 seconds]
zupo has joined #nixos
<Henson>
does anybody have any suggestions on how I can combine a NixOS container deployment with a NixOS real computer deployment in a Nixops network?
opticnerve has joined #nixos
<Henson>
to define a real computer with a NixOS container within it, all under the same Nixops network definition?
morgrimm has joined #nixos
rogue_koder has joined #nixos
<{^_^}>
[nixpkgs] @bjornfor opened pull request #87203 → [20.03] conan: unbreak by backporting from master → https://git.io/JfnN0
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « phoronix-test-suite: 9.4.1 -> 9.6.0 »: https://git.io/JfnpZ
<kronicmage>
is there any way to reset the nix store on darwin? I've messed with a few permissions that I shouldn't have and now a lot of things are broken
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
fresheyeball has joined #nixos
ajs124 has quit [Client Quit]
<magnetophon>
anyone using jitsi with an usb interface for sound? I keep getting permission for mic denied, even though I enabled it in the settings of firefox. tried chromium too.
ajs124 has joined #nixos
domogled has joined #nixos
nschoe has quit [Ping timeout: 272 seconds]
iqubic has joined #nixos
seku has joined #nixos
<adisbladis>
magnetophon: Yep, no issues here.
<fresheyeball>
so I have two overlays the effect haskell packages
<adisbladis>
That should have very little to do with jitsi though?
<fresheyeball>
it appears one just write over the other
<fresheyeball>
is there a right way to compose overlays of haskell?
codygman has quit [Read error: Connection reset by peer]
codygman has joined #nixos
xfix[m] has joined #nixos
KeiraT- has joined #nixos
KeiraT has quit [Ping timeout: 240 seconds]
xfix[m] is now known as xfix
xfix has quit [Client Quit]
xfix has joined #nixos
xfix is now known as Guest82248
dramaturg has quit [Quit: dramaturg]
dramaturg has joined #nixos
KeiraT- is now known as KeiraT
Guest82248 has quit [Changing host]
Guest82248 has joined #nixos
Guest82248 has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
knupfer has quit [Quit: knupfer]
knupfer has joined #nixos
knupfer has quit [Client Quit]
knupfer has joined #nixos
knupfer has quit [Client Quit]
o1lo01ol1o has joined #nixos
dingenskirchen has quit [Remote host closed the connection]
knupfer has joined #nixos
dingenskirchen has joined #nixos
knupfer has quit [Client Quit]
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #87212 → nixos/gdk-pixbuf.nix: don’t set GDK_PIXBUF_MODULE_FILE in cross → https://git.io/Jfnj7
zupo has joined #nixos
knupfer has joined #nixos
knupfer has quit [Client Quit]
knupfer has joined #nixos
knupfer has quit [Client Quit]
knupfer has joined #nixos
knupfer has quit [Client Quit]
knupfer has joined #nixos
knupfer has quit [Client Quit]
Guest82248 has quit [Quit: Trying to configure Freenode bridge]
Guest82248 has joined #nixos
knupfer has joined #nixos
knupfer has quit [Client Quit]
knupfer has joined #nixos
knupfer has quit [Client Quit]
knupfer has joined #nixos
knupfer has quit [Client Quit]
Guest82248 has quit [Client Quit]
Guest82248 has joined #nixos
Guest82248 has quit [Client Quit]
Guest82248 has joined #nixos
o1lo01ol1o has quit [Ping timeout: 256 seconds]
Guest82248 has quit [Changing host]
Guest82248 has joined #nixos
Guest82248 has joined #nixos
knupfer has joined #nixos
knupfer has quit [Read error: Connection reset by peer]
knupfer has joined #nixos
knupfer has quit [Client Quit]
knupfer has joined #nixos
knupfer has quit [Client Quit]
knupfer has joined #nixos
knupfer has quit [Client Quit]
Woutifier has joined #nixos
Guest82248 is now known as xfix
knupfer has joined #nixos
knupfer has quit [Client Quit]
knupfer has joined #nixos
knupfer has quit [Client Quit]
<Woutifier>
hi
knupfer has joined #nixos
<Woutifier>
I'm having a quite basic issue I believe. I have spun up a NixOS, and now for the first act I would like to add Grafana to it. Bear in mind that this is the first time I touch Nix/NixOS.
<Woutifier>
I'm looking at the grafana.nix
knupfer has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @ryantm pushed commit from @Sohalt to master « mari0: init at 1.6.2 »: https://git.io/JfceI
<Woutifier>
and for the life of me I cannot figure out how the parameters end up anywhere
knupfer has joined #nixos
<Woutifier>
they are not referenced anywhere in any file
knupfer has quit [Client Quit]
<Woutifier>
I can for example set a "security.adminUser" for grafana, makes sense, but this value is not referenced in any config file, nix file, whatsoever
<thoughtpolice>
Woutifier: I'm not sure how experienced you are, but. When you write a NixOS module, it takes a 'config' parameter (first line of file), and one of the useful things this offers is access to the settings defined in that module, that you can access.
<Woutifier>
ah yes
<Woutifier>
but how do these config parameters end up in grafana's configuration
<thoughtpolice>
So you could ask for something like, `config.options.grafana.security.adminUser` in your module
<thoughtpolice>
A very common idiom is to write a NixOS module so it says something like `let cfg = config.options.grafana; in ...` and you can reference `cfg` from there on out. That's just a shortened form
o1lo01ol1o has joined #nixos
<thoughtpolice>
Woutifier: So, it is used and put into the attribute set named `envOptions` on line 9.
<thoughtpolice>
envOptions is basically a key value pair set. It says things like `FOO = "bar"`. We utilize systemd's features to set environments for a service to set all of the environment variables in envOptions to what they specify
ixxie has quit [Quit: Lost terminal]
<Woutifier>
right, so it is then up to Grafana to take the environment variables and actually do something with them
<thoughtpolice>
So, `cfg.security.adminUser` is used to define the environment variable `SECURITY_ADMIN_USER`, which is set by systemd service
<thoughtpolice>
You got it
<Woutifier>
thanks, that makes a lot more sense
<Woutifier>
I thought it was somehow used to write the config file itself or something like that
<{^_^}>
[nixpkgs] @DamienCassou merged pull request #86294 → Fix man pages in i3-gaps → https://git.io/Jfmjx
<{^_^}>
[nixpkgs] @DamienCassou pushed 3 commits to master: https://git.io/JfceC
<thoughtpolice>
Ugh, bad grammar but I'm glad you understand.
<thoughtpolice>
You can also do that! It depends on the service, tbh
<simpson>
For some services, yes. Indeed, systemd units are done that way.
<thoughtpolice>
And also how the author structures the module, basically. There are no real hard rules about that.
fabianhjr has joined #nixos
<Woutifier>
I was pretty impressed by the install, a few partitioning steps and that was basically all there was to it
<thoughtpolice>
The even better part is when you write a configuration and reformat your machine. Because it comes down to "a few partition steps and copy one file into place" and you basically have the exact same system. In other words, it takes 5 more seconds to get a carbon copy and resume where you left off.
<thoughtpolice>
That sold me on NixOS the first time I did it many years ago.
<Woutifier>
yes that is indeed the part that is very likeable. I have struggled with keeping my systems clean in the past, as I like to experiment with different things.
<thoughtpolice>
As an example of this I recently switched from an nvidia card (and the nvidia driver + many x11 packages built for that nvidia configuration) to an amd gpu using the upstream amdgpu driver in my headless machine. This was a one line change and worked instantly on reboot by just swapping out the cards.
<Yaniel>
yea, that was impressive
<Yaniel>
did the same around new year's
<thoughtpolice>
(You can tell NixOS to both "switch to a new version of the system" right now, or "switch to a new version on boot". So you just move to amdgpu "for the next boot" and then shutdown, swap cards, and turn everything back on.)
<Woutifier>
thanks for the quick help, it's appreciated
<Woutifier>
looks like I know what I"ll be spending my time on for a while
freeman42x[m]_ has quit [Quit: Connection closed for inactivity]
<Henson>
is there some way I can get Nix to import every *.nix file in a directory without explicitly specifying each file? What I'm currently doing is importing a default.nix file which explicitly imports every file in the directory.
<numkem>
is it possible that a new version of stdenv broken go 1.14.2? I have an override for it and it was running well for weeks than I did a channel update yesterday which triggered the rebuild and ever since I get weird behaviors
mallox has quit [Quit: WeeChat 2.8]
<cole-h>
Henson: AFAIK, that's the only way
<Henson>
cole-h: ok, thank you.
<numkem>
the package g_1_14 from stable (go 1.14.1) isn't using the same stdenv as it just downloaded another one I reverted back to it
<asbachb>
Is there any reason sudo 1.8.31p1 is not backported to 20.03?
<cole-h>
numkem: No idea, but sounds like an issue should be opened. Are you sure the "weird behaviors" are not related to Go itself?
<numkem>
cole-h: doubtful as if I run it as root it works and it work work with every packages
<cole-h>
That more sounds like an issue with your profile/a specific package than Go itself
<numkem>
it's my own code tho, a `go test` or `go run` on something that used to work suddenlty doesn't
<cole-h>
Well, if it works as root...
<numkem>
I'm trying to rebuild using unstable as source, see if that changes things
staydownchamp has joined #nixos
<cole-h>
asbachb: Most likely because that patch was released after branch off
mojjoo has quit [Ping timeout: 245 seconds]
<asbachb>
cole-h: Is there a guideline what to backport?
<asbachb>
After branch off.
<cole-h>
The changelog doesn't appear to be an "impending doom" change, so I don't think it's necessary. Do you need 1.8.31p1?
<cole-h>
asbachb: If you have a real reason for needing 1.8.31p1, feel free to open a backport PR and see what the people who can actually merge it say. If you explain your position, it might be more likely to be accepted.
<cole-h>
(That said, I don't think we really have a backport "standard" -- up to individual committers whether it gets merged or not.)
<cole-h>
s/it/things/
<aleph->
Trying to remember can I keep common nix files for nixos in say /etc/nixos/common/ and import them as ./common/security.nix and so on?
<asbachb>
cole-h: There's an error message. I don't know exactly how important it is. But I guess that the version number is indicating that it's a candidate for backporting.
<cole-h>
If you're importing it from /etc/nixos/somefile.nix, sure.
<cole-h>
asbachb: Well, we don't usually backport version bumps, AFAICT.
<aleph->
Yeah wanted to do something like imports = [ ./common/security.nix ... ];
<cole-h>
Either way, if it's important to you, open it; worst case, it doesn't get merged. :^)
<cole-h>
aleph-: As long as the relative path is correct, I don't see why not.
<aleph->
Got it, just wasn't sure. Thanks
<asbachb>
cole-h: I'll dive a little bit more into that issue. For now it seems just like a warning.
cr4y1 has quit [Remote host closed the connection]
splintah has joined #nixos
rogue_koder has quit [Ping timeout: 260 seconds]
cr4y1 has joined #nixos
<asbachb>
Is there a possibility to add some persistent system configuration besides configuration.nix? e.g. I create a system image and I want to have some configuration baked into that system.
rogue_koder has joined #nixos
<cole-h>
What's the aversion to using configuration.nix?
<numkem>
cole-h: same problem when building from unstable
<cole-h>
numkem: Considering you've said the problem doesn't manifest under root, I'd experiment in that direction to find out why.
<asbachb>
cole-h: I just thought if there's another solution. I wonder if a container os should explicit configure it's an container image for example.
* xeu
is away, 10 minute autoaway (l!on) : os/bx
endformationage has joined #nixos
<asbachb>
When I generate a lxc container image. The configuration.nix is empty and I need to setup several thing for a rebuild to work.
m0rphism has quit [Quit: WeeChat 2.7.1]
<cole-h>
You can write a config for it that's baked in when building the system with `-I nixos-config=/path/to/new/config.nix`
<cole-h>
That's how I test NixOS modules while I'm not on NixOS
<unclechu>
hey, how do i remove old system builds?
<cole-h>
unclechu: All of them?
<numkem>
cole-h: I'm not super familiar with the profile thing, what could be different between root and my user that could cause this. I do know that `go run` or `go test` writes in `/tmp` if that gives you any ideas
<unclechu>
cole-h: well, i'd prefer to keep latest 5 ones for instaces
staydownchamp has quit [Quit: staydownchamp]
<asbachb>
cole-h_ And that's persistent duriang a rebuild?
<unclechu>
for instance*
<cole-h>
asbachb: As long as you provide that parameter, the VM's config will always match it, AFAIK.
<unclechu>
cole-h: but you know, not latest hundred of them
<cole-h>
If you make a change, you'll need to rebuild the VM with `-I nixos-config=...`
<cole-h>
unclechu: Check out `man nix-collect-garbage`
<unclechu>
cole-h: i didn't know `nix-collect-garbage` can also handle system builds, thanks
<cole-h>
numkem: The fact that it differs between root and your unprivileged user makes it obvious to me that *something* is different. You could try `strace`-ing it.
m0rphism has joined #nixos
<numkem>
cole-h: it's obvious to me as well, I did try strace but didn't show anything obvious, I'll try again to see
<cole-h>
numkem: Compare the outputs. Either way, I don't believe this is a problem with Go itself.
<cole-h>
Might even try it in a pure shell
<unclechu>
cole-h: well `man` of `nix-collect-garbage` is pretty small and tells nothing about system builds
<unclechu>
and as i thought before it cleans only store paths
<cole-h>
"--delete-old), which deletes all old generations of all profiles in /nix/var/nix/profiles"
<cole-h>
A system build is a generation
nuncanada2 has joined #nixos
nuncanada has quit [Read error: Connection reset by peer]
<cole-h>
Therefore, deleting a generation is equivalent to deleting a system build
rogue_koder has quit [Ping timeout: 240 seconds]
<cole-h>
Therefore, the answer you seek indeed lies in that manpage :P
<cole-h>
Also note the `--delete-other-than` flag
<cole-h>
Err, `--delete-older-tan`
<cole-h>
...
<unclechu>
oh, then it makes sense, thanks, cole-h
<cole-h>
Fix spelling where appropriate
evanjs has quit [Ping timeout: 272 seconds]
<numkem>
cole-h: we'll I've learned something new, `go run` and `go test` writes to `~/.cache/go-build/` deleting that folder solved it...
<cole-h>
numkem: :)
<numkem>
I kept thinking about permission but environment is quite differnt with NixOS, something to keep in mind
nuncanada2 has quit [Ping timeout: 246 seconds]
nuncanada has joined #nixos
<{^_^}>
[nixpkgs] @kirelagin opened pull request #87219 → postgres: Do not log timestamp → https://git.io/JfcJT
chagra_ has joined #nixos
knupfer has quit [Ping timeout: 252 seconds]
evanjs has joined #nixos
zakkor has quit [Quit: Connection closed for inactivity]
dermetfan has joined #nixos
rogue_koder has joined #nixos
user_0x58 has quit [Ping timeout: 256 seconds]
user_0x58 has joined #nixos
chris_ has joined #nixos
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #87220 → numpy: set lapack, blas sections → https://git.io/JfcJV
gustavderdrache has quit [Quit: Leaving.]
dhess has quit [Remote host closed the connection]
splintah has quit [Ping timeout: 240 seconds]
lukash_ is now known as lukash_|away
<asbachb>
cole-h: Anything without passing an extra argument? :>
sigtrm has quit [Remote host closed the connection]
<cole-h>
None that I know of. If you do this in a specific directory, you could use direnv and put `NIX_PATH=nixos-config=./config.nix:$NIX_PATH` in your .envrc
matthiaskrgr has quit [Quit: Free ZNC ~ Powered by LunarBNC: https://LunarBNC.net - currently broken?]
<cole-h>
But then keep in mind that will be set for ALL subdirectories
user_0x58 has quit [Ping timeout: 246 seconds]
<chris_>
I kinda have a question (I'm new to NixOS, but not to Linux): I've installed NixOS, but I made a typo in the configuration file fileSystems."/boot". I fixed it and mounted it manually again, but whenever I try to run 'nixos-rebuild switch`, grub-install complains about /boot not being a efi partition and '/boot/' is unmounted once again. Am I doing something wrong or is there a way to prevent the unmounting of '/boot' while rebuilding the configuration?
inkbottle has joined #nixos
dermetfan has quit [Ping timeout: 260 seconds]
zebrag has quit [Ping timeout: 260 seconds]
user_0x58 has joined #nixos
matthiaskrgr has joined #nixos
morgrimm has quit [Ping timeout: 260 seconds]
magnetophon has quit [Read error: Connection reset by peer]
magnetophon has joined #nixos
user_0x58 has quit [Ping timeout: 246 seconds]
user_0x58 has joined #nixos
rogue_koder has quit [Ping timeout: 264 seconds]
rogue_koder has joined #nixos
punkj has joined #nixos
rogue_koder has quit [Excess Flood]
rogue_koder has joined #nixos
hlisp has joined #nixos
EdLin has joined #nixos
<lovesegfault>
Has anyone had any success running NixOS 19.09 or 20.03 on a packet host?
<lovesegfault>
Struggling really hard to get networking up and running
<andi->
ha, I bet that is their bonding again
<xfix>
chris_: is /boot an ESP partition?
splintah has joined #nixos
<xfix>
also, you may want to run nixos-generate-config after doing changes to mounts
sigtrm has joined #nixos
astrofog has joined #nixos
<numkem>
can't find it in the manual, what's the right way of saying to use the current directory and all subdirectories as `src`?
<simpson>
./.;
<simpson>
As in `src = ./.;`
freeman42x[m]_ has joined #nixos
<cole-h>
lovesegfault: Might want to ask gchristensen -- some/most of ofborg's stuff runs on packet
<Ashy>
ok this is annoying now, gnome screenlock is not working for me on nixos-unstable with gdm as my login manager, is nobody else getting this issue?
<Ashy>
on multiple machines for me (with a somewhat share config repo)
<cole-h>
Maybe a paste of gnome-related configs would help
brotherdust has quit [Remote host closed the connection]
freeman42x has joined #nixos
<Ashy>
i had seen this previously when i attempted to use sddm with gnome and found out that the screenlock functionality now requires gdm for gnome
<Ashy>
but now i am definitely using gdm...
<hr[m]>
Does anyone have experience setting up wireguard on NixOS with all traffic routed through the vpn? I attempted it a few months ago but remember there was an issue with routing all traffic.
o1lo01ol1o has quit [Remote host closed the connection]
Fare has joined #nixos
fusion809 has joined #nixos
zupo has quit [Ping timeout: 272 seconds]
o1lo01ol1o has joined #nixos
<pistache>
hr[m]: I'm using such a tunnel right now. there are multiple ways to do it
<pistache>
what I usually do is setup a specific routing table for the Wireguard traffic
zupo has joined #nixos
kim0 has quit [Quit: Connection closed for inactivity]
fendor has quit [Quit: Leaving]
<hr[m]>
pistache great to hear that it can be done. Do you know if there's any example configuration available online that I could look at? I just remember looking over a bunch of examples a few months ago and none of them seemed to work for routing all traffic. Thanks
iqubic has quit [Ping timeout: 260 seconds]
virus_dave has joined #nixos
<pistache>
hr[m]: I can show you my configuration, just give me a few seconds so that I can strip out the IPs and add some comments
<Fare>
Why is it that my layered image has a 1.7GB layer and plenty of small layers? Can I inspect what the layers are about somehow, or get debugging information as it creates them?
vidbina has quit [Ping timeout: 258 seconds]
<pistache>
hr[m]: I'm not using it here, but it can be done with NixOS's own networking abstraction
freeman42x has quit [Quit: Leaving]
<pistache>
hr[m]: ah, I'm not so sure it can be done with NixOS's own networking abstraction, as I don't see where we'd specifying the routing policy rules
<pistache>
hr[m]: however, if you want a simpler setup, I've just seen that "wg-quick" is integrated in NixOS, and it will build a routed VPN for you
cosimone_ has quit [Quit: Quit.]
<pistache>
wg-quick also uses routing table to achieve that result, but it handles it all for you
<hr[m]>
Ok, i'll take a look at those
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<pistache>
it's in networking.wg-quick
anderslu1dstedt has joined #nixos
<hr[m]>
I think I'll try use wg-quick as last time I was using `networking.wireguard`. Thanks for your help
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
lc has joined #nixos
anderslundstedt has quit [Ping timeout: 260 seconds]
phreedom has quit [Ping timeout: 240 seconds]
<lc>
So... I'm trying to install an unstable package, and it keeps downloading the (currently broken) stable package, which hash mismatches with the site it gets the .deb file from
<lc>
It's the rescuetime packages (nonfree)
anderslundstedt has joined #nixos
anderslu1dstedt has quit [Ping timeout: 272 seconds]
mbrgm_ has joined #nixos
itorres has quit [Ping timeout: 265 seconds]
hlisp has quit [Ping timeout: 240 seconds]
<lc>
Can someone else try to install the rescuetime package from stable to confirm this is something that happens elsewhere
<chris_>
xfix: It is, I god it fixed right now, by booting into the installer and reinstalling the system with the fixed config
<chris_>
I meant, I've got it fixed right now. I really need some sleep.
chris_ has quit [Quit: Konversation terminated!]
m0rphism has quit [Quit: WeeChat 2.7.1]
chloekek has quit [Ping timeout: 246 seconds]
<adisbladis>
Fare: The layer distribution is calculated only by "popularity" and doesn't take size into account
itorres has quit [Ping timeout: 272 seconds]
itorres has joined #nixos
<Fare>
emily, that's a great article (that I read months ago), but it doesn't explain how to inspect the size of the layers and tweak the build expression to improve sharing.
<emily>
right
<lc>
plz2hyelp
<emily>
I think source-diving might be your best bet there, but I dunno
<adisbladis>
There isn't really an interface to "improve" sharing
<adisbladis>
Fare: `nix-store --query --graph` will help you understand dependency graphs better
maddo has quit [Quit: See ya]
<Fare>
adisbladis, I understand the graph enough for my purposes (a few things I'm writing, and all the rest I'd like to put in a layer and share)
<Fare>
what I don't understand is the docker layer side of it.
wolfshappen has quit [Quit: bye]
o1lo01ol1o has quit [Remote host closed the connection]
<adisbladis>
Fare: Basically each store path gets a "weight" where the weight is the number of other paths refering to it
<adisbladis>
That's then sorted
<adisbladis>
And smushed to fit in to the number of max layers configured
<adisbladis>
You can't really customise this as things are right now.
<lc>
Is there any way I can install a debian package manually, because this the rescuetime package is broken and none of the tutorials that teach you how to install a single unstable package don't work either, and instead try to install the old package
<pistache>
lc: are you saying it has been fixed in unstable ?
<lc>
pistache: I'm saying I have yet to try it in unstable because my home-manager refuses to download the unstable version :(. I assume that unstable might fix it because the problem is that the site has changed versions since the hash was written down
<jackdk>
On current nixos-20.03 channel, `nix-build -E '(import <nixpkgs> {}).cabal2nix' tries to build ghc 8.8.1 from source, but `nix-build -E '(import <nixpkgs> {}).haskellPackages.cabal2nix'` hits the cache. why is this?
<pistache>
lc: indeed the hash in unstable seems to be correct
<lc>
If only there were a way for me to selectively install the unstable version, but no matter what tutorial I follow, it downloads 2.14.5.2
hlisp has quit [Remote host closed the connection]
<pistache>
assuming you have an "unstable" channel setup, "nix-env -f .nix-defexpr/channels/unstable -iA rescuetime" should work
<adisbladis>
jackdk: They are not the same derivation and haskellPackages is not built on Hydra
<adisbladis>
Except if they're dependent upon by other packages that are
<pistache>
if you want to do this in your HM config, use "let" to define "unstable" using builtins.fetchTarball, then reference the packageas "unstable.rescuetime" instead of "pkgs.rescuetime"
<jackdk>
adisbladis: okay, but it's the other way around. the haskellPacakges one is a cache hit, the toplevel one is a cache miss
oborot has joined #nixos
<lc>
pistache: I've tried a couple of those configurations in HM: Please show me exactly what would work
<adisbladis>
jackdk: That could be explained by something else depending on the haskellPackages derivation
<asbachb>
Is there a way to add some configuration somewhere that it's not needed to be added to configuration.nix?
justanotheruser has quit [Ping timeout: 246 seconds]
<asbachb>
e.g. when I' creating an image for a lxc I don't want the container image needs to define isContainer. It should like somehow be baked into that system.
<lc>
termbin.com/wohy is the output of my home-manager switch command @lc
<pistache>
lc: I'll try it on my system
MmeQuignon has joined #nixos
<pistache>
lc: with "unstable = import <nixos-unstable> { config = { allowUnfree = true; }; };", it seems to be building the updated package for me
<lc>
I just tried that exact line in place of what I have, didnt work
<lc>
Same output
<pistache>
lc: ah, so it seems like the fix has not reached unstable yet
justanotheruser has joined #nixos
<pistache>
it is in master though
<lc>
No, I mean that it still tries to download the stable version
<lc>
2.14.5.2
<lc>
hash mismatch, same hash, same everything
Arahael has quit [Ping timeout: 264 seconds]
<pistache>
lc: I think the version in unstable is still 2.14.5.2
user_0x58 has quit [Ping timeout: 256 seconds]
<lc>
Is that the version you downloaded? it's not what shows up in the nix package search under unstable
Arahael has joined #nixos
<Henson>
when I'm adding custom filewall rules via networking.firewall.extraCommands, should I use the nixos-fw and similar chains? I noticed that rules I just add to the INPUT and OUTPUT chains don't get deleted when I reload the system, they just keep piling up. If so, I noticed that the filter OUTPUT chain doesn't have an equivalent nixos-output chain.
marcusr has quit [Remote host closed the connection]
user_0x58 has joined #nixos
<clever>
Henson: i do use nixos-fw for my stuff
marcusr has joined #nixos
codygman has quit [Ping timeout: 260 seconds]
<spinlock[m]>
I have a strange issue where none of the autostart items added by packages to `/etc/xdg/autostart/` are being executed by xsession. For sure it's happening with at least `ibus` and `lightlocker`. Makes me feel that I'm missing some key option somewhere
codygman has joined #nixos
codygman has quit [Read error: Connection reset by peer]
<pistache>
lc: that's weird, I'm not sure what's up with that. I can confirm that the version in nixos-unstable is 2.14.5.2