<colemickens>
codygman: no, stateVersion is important stil
<colemickens>
codygman: stateVersion is used to get backward incompatible changes
<colemickens>
gate*
<colemickens>
So, Jellyfin, for example, had a majorly breaking change in it's latest version, so much so that you may be "stuck" with the old version of Jellyfin according ot your stateVersion.
<codygman>
I'm not sure why my state version is older than 20.09 when I'm on unstable. Also isn't this the home manager specific state version?
<colemickens>
codygman: the point is that stateVersion is independent from the nixpkgs version.
<adisbladis>
codygman: The stateVersion is fixed at config generation time
<adisbladis>
When you did `nixos-generate-config` a stateVersion was pinned
<infinisil>
This means you can't manually do the state migration for just one module without also doing the migration for all others
<infinisil>
adisbladis: ^
<gchristensen>
infinisil: not conceptually, it sounds like codygman thinks their stateversion needs "fixing"
supersandro2000 has quit [Disconnected by services]
<adisbladis>
codygman: pkgs.path has been around since basically forever I think?
<infinisil>
Ah
supersandro2000 has joined #nixos
<adisbladis>
So you should never have to fall back to <nixpkgs>
g-w1 has joined #nixos
<infinisil>
If I had enough motivation I'd do an rfc for making stateVersion per-module though :)
<codygman>
gchristensen: I didn't originally think my state version needed fixing until after hours banging my head on trying to use that flake despite all logic ;)
<adisbladis>
infinisil: Can you think of a good UX for it?
<kini>
nixpkgs has a fork of openconnect called openconnect_pa. But as of version 8.0, openconnect seems to have reabsorbed the fork. What's the best way to remove openconnect_pa while also offering users a seamless upgrade path to mainline openconnect?
<adisbladis>
kini: Remove the unmaintained fork and add an alias
<infinisil>
adisbladis: For one, just have e.g. `services.foo.stateVersion = mkOption { type = enum [ 0 1 2 ... ]; }`, without a default, forcing the user to set one
<infinisil>
An error would show why it's needed, and what the latest stateVersion is
<adisbladis>
kini: Aliases are defined in pkgs/top-level/aliases.nix
<kini>
adisbladis: They have different pname values ("openconnect" and "openconnect_pa"); doesn't `nix-env -u` work based on pnames rather than attribute names, which is what aliases do?
<g-w1>
is there a way for nix to show me what would be upgraded without actually upgrading?
<infinisil>
adisbladis: For another, I'd try to implement state version runtime checking, such that e.g. the state version would be recorded in /var/lib/foo/.nixos-state-version, such that in case of the evaluation value not matching the runtime value, it would complain, and tell you how to migrate
<infinisil>
And ideally you could migrate just by doing `nixos-state-version-migrate fooModule 3`, which then either does it automatically, or gives you instructions for how to do it
<infinisil>
With support for checking that it was done successfully, etc.
<codygman>
Thanks adisbladis, I'm posting an issue now
<infinisil>
And each module should be able to specify a migration for every state version update
<gchristensen>
kini: unfortunately we don't consider that often
<adisbladis>
Hmm, if it's in home-manager I feel like there should be a reason for it..
kolaente_ has quit [Quit: WeeChat 2.9]
<adisbladis>
codygman: Anyway, I think posting an issue is the right thing to do, at least to get some clarification
<kini>
gchristensen: OK, if not supporting `nix-env -u` in cases like this is standard practice, I'll just go with that :)
erictapen has quit [Ping timeout: 256 seconds]
<infinisil>
adisbladis: That would be really neat ux imo
<adisbladis>
infinisil: It sounds pretty dirty though... Eval with side effects ..
<adisbladis>
Why did I think you meant eval-time when you said runtime...
<{^_^}>
[nixpkgs] @github-actions[bot] pushed 19 commits to staging-next: https://git.io/JLsTX
* adisbladis
should go to sleep
<{^_^}>
[nixpkgs] @github-actions[bot] pushed 20 commits to staging: https://git.io/JLsT1
<infinisil>
Hehe
<adisbladis>
o/
<infinisil>
o/
<gchristensen>
\o
<infinisil>
Oh
<infinisil>
\o
<codygman>
I guess if I apply that patch and try building again I'll find out what feature home-manager uses that I have too old of a state version for.
<ronthecookie>
does the home-manager xsession.windowManager.i3 stuff replace the native nixos services.xserver.windowManager.i3?
<infinisil>
Heck, should I just draft up a quick RFC for this?
<infinisil>
Hmm but I don't really have a lot of time, neither to maintain the rfc nor to implement it
<gchristensen>
maybe not an rfc yet, but an experiment to see how it works
<gchristensen>
the best RFCs work already in poc form
<infinisil>
Yea
SomeoneSerge has quit [Ping timeout: 264 seconds]
* infinisil
does aoc instead hah
<genevino>
ronthecookie: i would assume it "overrides" the system-wide installed one because your home-manager stuff occurs earlier in your $PATH - am i wrong here?
<ronthecookie>
if you're right, does that mean i need to configure my display manager to work with home-manager's i3?
<ronthecookie>
...well i guess i could just try so ill do that
erictapen has joined #nixos
<{^_^}>
[nixpkgs] @kini opened pull request #106989 → openconnect_pa: remove from nixpkgs → https://git.io/JLsII
g-w1 has left #nixos ["WeeChat 3.1-dev"]
<codygman>
Okay so my understanding is: 1) I can't build home-manager as a flake purely unless my state version is newer 2) I shouldn't change my state version so 3) I can't build flakes using home manager without --impure.
<codygman>
Reinstalling isn't the answer, but I think it would give me a newer state version?
<codygman>
That's kinda frustrating if I have everything straight
<colemickens>
I guess you have it right, I haven't looked closely at the piece of HM code in question though.
<colemickens>
I tend to use HM as a NixOS module and it just inherits my nixpkgs being used by the nixos configuration.
<gchristensen>
tbh if you don't mind reinstalling, just change your state version
<gchristensen>
you might have something wierd going on but you can fix whatever things are weird
<qyliss>
pushqrdx: you can also use ,locate in PMs with {^_^}
<pushqrdx>
oh
<qyliss>
:)
<pushqrdx>
where is that gio dependency though 150 results isn't any helpful
<tpw_rules>
so i guess one might say home-manager is sort of a a way to generate application config files with nix?
<tpw_rules>
i'm trying to understand what it does since everybody seems to love it
<pushqrdx>
i am trying to build something and these deps are missing
<pushqrdx>
-- 'xrandr'
<pushqrdx>
<pushqrdx>
-- 'sm'
<pushqrdx>
-- 'ice'
<pushqrdx>
-- 'xi'
<pushqrdx>
tpw_rules homemanager to your home directory is what configuration.nix to your nixos
<ronthecookie>
tpw_rules: yeah thats it for the most part, theres a bit more to it that i dont understand either (glue code to run programs with the configs)
<tpw_rules>
i guess i'm not really familiar with what's in my home directory
<pushqrdx>
essentially it adds nixos capabilities to your home dir
<tpw_rules>
that statement doesn't mean anything to me
<ronthecookie>
it lets you roll back to old versions of all your configs and you can manage it all in the nix lang
<simpson>
tpw_rules: It does to ~/.vim/ what NixOS does to /boot/ or what Nix does to /usr/.
<ronthecookie>
im still a noob but home-manager has been quite pleasant even with the lack of docs
CMCDragonkai2 has quit [Ping timeout: 268 seconds]
medvid has quit [Ping timeout: 264 seconds]
medvid has joined #nixos
Lord_of_Life_ has joined #nixos
sigmundv__ has quit [Ping timeout: 272 seconds]
Lord_of_Life has quit [Ping timeout: 260 seconds]
Lord_of_Life_ is now known as Lord_of_Life
_deepfire has joined #nixos
dycan has quit [Quit: dycan]
iH8c0ff33 has joined #nixos
<energizer>
nixos-infect doc says `cat customConfig.optional nixos-infect | ssh root@targethost` what does `customConfig.optional` mean? is that metasyntax or what
KaboomShebang[m] has quit [Ping timeout: 268 seconds]
berberman[T] has quit [Ping timeout: 268 seconds]
demostanis[m] has quit [Ping timeout: 268 seconds]
manveru[m] has quit [Ping timeout: 268 seconds]
eddyb has quit [Ping timeout: 268 seconds]
grin[m] has quit [Ping timeout: 268 seconds]
srenatus[m] has quit [Ping timeout: 268 seconds]
tejing[m] has quit [Ping timeout: 268 seconds]
spacetato has quit [Ping timeout: 268 seconds]
DavHau[m] has quit [Ping timeout: 268 seconds]
emily has quit [Ping timeout: 268 seconds]
tmeres[m] has quit [Ping timeout: 268 seconds]
JaakkoLuttinen[m has quit [Ping timeout: 268 seconds]
beandipper[m] has quit [Ping timeout: 268 seconds]
redcedar[m] has quit [Ping timeout: 268 seconds]
Piece_Maker has quit [Ping timeout: 268 seconds]
ryantm has quit [Ping timeout: 268 seconds]
michaelpj has quit [Ping timeout: 268 seconds]
aanderse has quit [Ping timeout: 268 seconds]
DavHau[m] has joined #nixos
manveru[m] has joined #nixos
beandipper[m] has joined #nixos
demostanis[m] has joined #nixos
eddyb has joined #nixos
aanderse has joined #nixos
ryantm has joined #nixos
tmeres[m] has joined #nixos
NomisIV[m] has joined #nixos
Piece_Maker has joined #nixos
emily has joined #nixos
KaboomShebang[m] has joined #nixos
srenatus[m] has joined #nixos
berberman[T] has joined #nixos
Ken[m] has joined #nixos
axx has joined #nixos
<{^_^}>
[nixpkgs] @andir opened pull request #106995 → [PATCH] nixos/users-groups: createHome: Ensure HOME permissions, fix description → https://git.io/JLs4v
l33[m] has joined #nixos
redcedar[m] has joined #nixos
michaelpj has joined #nixos
spacetato has joined #nixos
ph_0x17[m] has joined #nixos
JaakkoLuttinen[m has joined #nixos
Skallwar has joined #nixos
tejing[m] has joined #nixos
DamienCassou has joined #nixos
grin[m] has joined #nixos
mkaito has joined #nixos
mkaito has joined #nixos
mkaito has quit [Changing host]
ZeloZhang has joined #nixos
pjt_tmp has joined #nixos
KarlJoad` has quit [Ping timeout: 260 seconds]
<{^_^}>
[nixpkgs] @andir opened pull request #106996 → wireshark: Do $PATH lookup in wireshark.desktop instead of hardcoding derivation → https://git.io/JLsRI
S0rin has quit [Ping timeout: 260 seconds]
vonfry has joined #nixos
Shados_ has joined #nixos
orivej has joined #nixos
Shados has quit [Ping timeout: 260 seconds]
S0rin has joined #nixos
<{^_^}>
[nixpkgs] @samuela opened pull request #106997 → vscodium: update script artifact names have changed → https://git.io/JLsR9
seanparsons has quit [Read error: Connection reset by peer]
<notgne2[m]>
then presumably it will sit in `/run/wrappers/bin/slock`
<ronthecookie>
i already saw the module but anyway thanks for the path, that works
seanparsons has joined #nixos
KarlJoad has joined #nixos
<KarlJoad>
Can someone check the work I've done so far, to make sure I am working on this the way nixpkgs would like, before I start getting to far into this little project of mine? https://github.com/KarlJoad/nixpkgs/tree/octave-modules
<{^_^}>
[darwin-stubs] @thefloweringash pushed 2 commits to main: https://git.io/JLsPa
<KarlJoad>
Which is preferred in nixpkgs for fetching a tarball: fetchurl or fetchTarball?
<clever>
KarlJoad: pkgs.fetchurl keeps the disk usage down more
<clever>
builtins.fetchTarball isnt allowed in nixpkgs itself, since its IFD
<KarlJoad>
clever: IFD?
<clever>
,ifd
<{^_^}>
import-from-derivation (IFD) is when you evaluate nix from a derivation result, for example `import (pkgs.writeText "n" "1 + 1")` will evaluate to 2. This is sometimes problematic because it requires evaluating some, building some, and then evaluating the build result. It has been described as "such a nice footgun."
evils has joined #nixos
<KarlJoad>
So, in short, an IFD is when a derivation ends up making a call to Nix in some way?
<clever>
when it winds up causing a build at eval time
codygman has quit [Read error: Connection reset by peer]
codygman has joined #nixos
<KarlJoad>
clever: I guess I'm not following them. Would a build occur _at_ eval time?
<clever>
KarlJoad: builtins.fetchurl and builtins.fetchTarball are both eval time fetches, so that forces it to download things in series, and to stop evaling while it waits on network
<clever>
KarlJoad: but pkgs.fetchurl and pkgs.fetchzip are derivations, it just describes how to fetch something later, once the eval is done, nix can do all of the fetching in parallel
<KarlJoad>
clever: I guess I have a worse understanding of Nix than I thought.
<clever>
and rebuild top-level, it should be much smaller now
mkaito has quit [Quit: WeeChat 2.9]
<clever>
hardware.enableAllFirmware = false; may also help
<{^_^}>
[nixpkgs] @Atemu opened pull request #107000 → waf: explicityly use native versions of packages run during build → https://git.io/JLsSy
<KarlJoad>
When unpacking a tarball, what is Nix's default way of naming the unpacked directory? If it has one?
<clever>
KarlJoad: it just unpacks to the current dir, and assumes the tar has a root directory for all files
<clever>
the default unpackPhase will compare the dir listing before and after, and cd into whatever it made
<clever>
it will complain and fail if more then 1 thing was made
<KarlJoad>
clever: That's what I thought. Now it's behaving differently than what I had before with builtins.fetchTarball.
<clever>
KarlJoad: builtins.fetchTarball is special, it will strip that common dir, and put it in /nix/store/hash-source/
ddellacosta has quit [Ping timeout: 264 seconds]
<KarlJoad>
clever: That's what I seem to be figuring out now.
<clever>
pkgs.fetchzip does the same (and can be used on tar files)
<clever>
but the unpackPhase and src= can operate on tar and zip without any need to pre-unpack
<KarlJoad>
It's not the unpacking stuff that's failing now. It's the installPhase.
<clever>
by default, it just runs `make install` and expects the package to respect the `--prefix=` that was given to configure
<KarlJoad>
I know. For these packages the installPhase just consists of copying things out of untar-ed/inst/* to $out/.
<clever>
KarlJoad: is it even building anything?
<KarlJoad>
clever: As in compiling? Yes. There is a bunch of fortran stuff right now. I am not sure if I need it at all. There is little information about how to manually build these packages.
<clever>
ah
<clever>
bbl
<KarlJoad>
I do run the build, and that works. The Fortran compiles correctly. The compiled binaries are copied to $out correctly. Then it fails in the installPhase.
<clever>
you can add `installPhase = "echo foo";` to the drv to make it skip that
<clever>
and then add mkdir and cp commands to it as needed
<KarlJoad>
clever: It's just me being stupid right now. The actual error I'm getting is "cp: missing destination file operand after '/nix/store/if33fmsr1k5iw6j47fljixn5nbjppy51-control-3.2.0/'"
MidAutumnHotaru has quit [Quit: Ping timeout (120 seconds)]
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « python37Packages.bugwarrior: 1.7.0 -> 1.8.0 »: https://git.io/JLGew
LnL has joined #nixos
<tnks>
I forget. Can a Nix build prevent arbitrary internet access?
<tnks>
I always think of sanboxing more about filesystems than networking. Just not sure.
<clever>
tnks: the sandboxing blocks both network and filesystem access
KarlJoad has quit [Quit: ERC (IRC client for Emacs 27.1)]
<tnks>
clever: okay, that's great... means I won't accidentally use a tool that downloads something I don't want it to.
<tnks>
clever: how is sandboxing implemented?
<clever>
tnks: namespacing, same tools as docker and systemd isolation
<tnks>
Okay, that makes sense then.
<tnks>
arduindo-cli is used by this build I'm tinkering with, and it seems hellbent on downloading things. I couldn't figure out how to get it to use tools provided by Nix. So I feel I'm left with the lesser option of downloading what it expects and putting them in the right directory to trick it to believe it's already downloaded it.
LnL has quit [Quit: exit 1]
LnL has joined #nixos
LnL has joined #nixos
LnL has quit [Changing host]
<tnks>
But with that approach, I think I really luck out that whatever's downloaded doesn't rely on shared libraries that it can't find.
orivej has quit [Ping timeout: 256 seconds]
<clever>
tnks: you can also do AVR stuff without the arduino tooling
<pjt_tmp>
If I'm in a local copy of nixpkgs and am trying to fix a pkg definition, is NIXPKGS=. NIX_PATH=. nix-shell -A pkgname sufficient (or on the right track)? I do that and it seems to ignore . in favor of the system's path.
<clever>
pjt_tmp: that says to look for <nixpkgs> as ./nixpkgs first, which likely doesnt exist
<clever>
however, you never said to actually load <nixpkgs>
<clever>
so it loads the default.nix in the current dir, and entirely ignores the NIX_PATH
<clever>
so just `nix-shell -A pkgname` would work
<arushi>
Hi , I am trying to use the tensorflow_2 nixpkg. For GPU I did python.pkgs.tensorflow_2.override {
<arushi>
}; ...but I kept getting this error : CUDA runtime implicit initialization on GPU:0 failed. Status: device kernel image is invalid .
<pjt_tmp>
clever: I'm not sure it's working with that. I have an "echo MMMMMMMMMM" in buildPhase which does not show up.
<tnks>
clever: I'm dealing with this Model 01 keyboard that seems to be piling a few repositories together. Their Makefiles seem to couple directly to arduino-cli. Maybe if I could somehow get a log of everything arduino-cli does at a lower level.
<clever>
pjt_tmp: `nix-shell -A pkgname` gives you an env suitable for building the pkg, not an env with a built copy of the pkg
<clever>
pjt_tmp: i think you want `nix-shell -p pkgname -I nixpkgs=.`
LnL has quit [Quit: exit 1]
LnL has joined #nixos
LnL has quit [Changing host]
<tnks>
Maybe I'll just brute-force wrap every executable with a script to log how it was executed... until I know what's all called in what order.
<pjt_tmp>
clever: ah, now it graces me with the MMMMMMMMMMMMM. Thank
<pjt_tmp>
s
themalik[m] has joined #nixos
palo has quit [Ping timeout: 256 seconds]
LnL has joined #nixos
<arushi>
I added cudaCapabilities = [ "sm_35" "sm_50" "sm_60" "sm_70" "sm_75" "compute_80" ] in my shell-nix to remove this issue (CUDA runtime implicit initialization on GPU:0 failed. Status: device kernel image is invalid )...but this is already there in default.nix for tensorflow_2 ..so why do we need to add this when using this package?
<regnat>
viric: Is this a multiple-outputs derivation? It's possible that one version only builds the default out path and the other one tries to build everything
<viric>
no, single
<viric>
I'm looking at the -vvvvv
faffolter has joined #nixos
<clever>
viric: you could reduce the logging, by running `nix-store --dry-run -r` on the .drv directly
<clever>
then you skip the eval
tcunha has quit [Ping timeout: 240 seconds]
<viric>
1st line is already querying info about missing paths...
<viric>
then start thread pool, and next checking substituter 'https://cache.nixos.org' for path '/nix/store/4s5qcjwd5xl02x4m57m2n02sm69aar1v-boost-1.66.0-vc14win64'
<viric>
which is the path that is ok according to nix 2.0.4
<clever>
does the path actually exist?
<clever>
if you run sqlite3 /nix/var/nix/db/db.sqlite, can you find it in the ValidPaths table?
<{^_^}>
[nix] @regnat pushed 10 commits to ca/drvoutputs-commands: https://git.io/JLGXv
<viric>
ARGH. The path doesn't exist
<viric>
likely nix 2.3.9 destroyed it?
<angerman>
is there some altenrative way for acme if you can't have port 80 working?
<pinpox>
Hi, I've run into what I think is a bug with nixos and systemd. After deploying my configuration, systemd and other stuff like sshd stops working. I can't even reboot!
<pinpox>
Can anyone help me debug this? I don't know what the cause is
<pinpox>
Failed to reboot system via logind: Connection timed out
<pinpox>
journalctl is spammed with messages like this (not sure if that's related): Dec 16 12:32:33 kartoffel kernel: refused connection: IN=enp0s20u2u1u1 OUT= MAC=18:65:71:e5:b1:e2:44:d9:e7:f2:b0:ed:08:00 SRC=192.168.2.110 DST=192.168.2.118 LEN=60 TOS=0x00 PREC=0x00 TTL=64 ID=58743 PROTO=TCP SPT=34308 DPT=8080 WINDOW=29200 RES=0x00 SYN URGP=0
<{^_^}>
[nixos-search] @garbas pushed 2 commits to implement-toNix-in-python: https://git.io/JLGA5
<pinpox>
Seems the network stuff goes away when I unplug that network adapter, but I still can't do anything with `systemctl` or `journalctl`
<{^_^}>
#106791 (by petabyteboy, 3 days ago, closed): system instability after updating systems
<colemickens>
I hit it on my rpi I think
<colemickens>
`sudo reboot -f` did the trick for me, rebooting it obviously
thematten[m] has quit [Ping timeout: 268 seconds]
<pinpox>
colemickens: I'll try the reboot -f, but does that fix the prolem?
<pinpox>
Im reading the issue, but that problem seemes to have been fixed if I read correctly?
thematten[m] has joined #nixos
<colemickens>
I don't know that that change has hit your channel, etc
respawn_ has joined #nixos
<colemickens>
if you're on the buggy version of systemd, you could rollback to an old generation, move your channel back, or wait and upgrade to the newer fixed systemd version
<pinpox>
colemickens: the reboot -f worked, but I'm pretty sure the problem will come again as soon as I rebuild the config, at least it has been like this till now
<pinpox>
colemickens: How can I verify if I'm on "the buggy version"?
<colemickens>
check the systemd version?
<pinpox>
systemctl --version
<pinpox>
systemd 247 (247)
<pinpox>
seems I am!
<pinpox>
colemickens: Did the change hit the channel you are using?
<{^_^}>
[nixos-search] @garbas merged pull request #252 → Pretty print nix examples and default fields in options → https://git.io/JLGoi
<{^_^}>
[nixos-search] @garbas pushed to master « Pretty print nix examples and default fields in options (#252) »: https://git.io/JLGhI
<colemickens>
pinpox: eh I'm just going to ride it out, if it causes me a headache again I'll revert to an old generation until the commit hits master. Or more likely, I'll pull it right now, queue re-builds for x86/aarch64 and then go back to bed
<{^_^}>
[nixos-search] @garbas pushed 0 commits to implement-toNix-in-python: https://git.io/JLGhL
WilliButz has quit [Quit: bye]
<colemickens>
it's hard to say how long it will sit in staging
WilliButz has joined #nixos
<pinpox>
I see. Well, I'll look if I can figure out how to pull and build it, havn't used anything else than the nixpkgs/nixos-unstable yet, if not I'll just wait a few days..
<Athas>
When using 'pip install --user' to install packages, I get an error that the setuptools module cannot be found. I understand that there are Nix ways of making Python modules available, but is there a non-Nix way to use pip for user-local package installation on NixOS?
<{^_^}>
[nix] @edolstra pushed 0 commits to ca/more-precise-build-noop: https://git.io/JLZu6
mkaito has joined #nixos
<pumpy>
anyone know why python3 -m http.server --directory .foo/ would start up fine but return 404 for everything? the file perms are 755 so should be fine
<{^_^}>
[nixpkgs] @adisbladis pushed to master « maintainers: Update my email address »: https://git.io/JLZzq
<infinisil>
pumpy: Well the . is suspicios, how about renaming that to a non-.-prefixed dir and trying that
zupo has joined #nixos
jd823592 has joined #nixos
hugolgst has quit [Ping timeout: 268 seconds]
berberman has joined #nixos
acarrico has joined #nixos
berberman_ has quit [Ping timeout: 260 seconds]
<pumpy>
ya that was it
<pumpy>
took it out worked ifne
<jd823592>
Hi, i am trying to add an optional structured parameter to an existing derivation to allow overwriting some hardcoded constants in the package but since I always forget most of the nix language whenever I try to modify something in nix I cannot get it working... How would one pass an optional set called vmoptions into a function providing a
<jd823592>
derivation that in its install step probes vmoptions and if there is an attribute Xmx it does something with the value of that attribute in shell and otherwise does nothing?
<stites[m]1>
has anyone gotten pip to work with conda-shell?
<{^_^}>
[nixpkgs] @andir merged pull request #106996 → wireshark: Do $PATH lookup in wireshark.desktop instead of hardcoding derivation → https://git.io/JLsRI
sigmundv__ has quit [Remote host closed the connection]
Ariakenom has joined #nixos
sigmundv has joined #nixos
<vldn>
Does somebody has any idea where i could find resources how the nix language is embedded at core? like nix version 0.0001 or something? or maybe guix version 0.0001 or anything else?
<hugolgst>
I have no idea but if you find I could be interested also ahah
|ando has joined #nixos
|ando has quit [Ping timeout: 256 seconds]
<vldn>
ftp://alpha.gnu.org/gnu/guix/guix-0.1.tar.gz maybe this could be intresting
xelxebar has quit [Remote host closed the connection]
fendor has quit [Remote host closed the connection]
iH8c0ff33-matrix has quit [Quit: Idle for 30+ days]
erasmas has joined #nixos
tester12345 has joined #nixos
tester12345 has quit [Remote host closed the connection]
sangoma has quit [Ping timeout: 264 seconds]
tlaxkit has joined #nixos
<eacameron>
I just built a NixOS desktop and it was booting fine until recently. It now says "A start job is running for /sys/subsystem/net/devices/wlp4s0" and it waits 1.5 minutes before moving on. Anyone know why this would "suddenly" start causing an issue or how to get around it? This desktop does have wifi but I'm not using it.
Ariakenom has quit [Remote host closed the connection]
Ariakenom has joined #nixos
<tejing>
first thing I'd do is check the log output for what's going wrong
growpotkin has joined #nixos
werner292 has joined #nixos
werner291 has quit [Ping timeout: 272 seconds]
werner292 is now known as werner291
M3nternamehere[m has joined #nixos
pie_ has quit [Quit: No Ping reply in 180 seconds.]
devalot has quit [Ping timeout: 256 seconds]
devalot has joined #nixos
pie_ has joined #nixos
<axx>
i was wondering how people managing ssh-agent, do you just have a snippet in your .bash_profile, or is there an option to run it as daemon via systemd?
<dutchie>
i have `programs.ssh.startAgent = true`
<eacameron>
tejing: How do I do that?
<tejing>
eacameron: journalctl is the comand to look at the logs. I usually use it as 'journalctl -eb' so it only shows me stuff from this boot and it starts with the pager at the end of the logs
<axx>
<dutchie "i have `programs.ssh.startAgent "> oh, that sounds… obvious, and great
<pushqrdx>
i am trying to build compiz and these packages are missing but i can't find them, can someone point me to them?
gustavderdrache has quit [Ping timeout: 260 seconds]
werner291 has quit [Ping timeout: 268 seconds]
werner292 is now known as werner291
<pushqrdx>
xrandr, xi, ice, sm
<kenran>
pushqrdx: xorg.libXrandr, xorg.libXi
<kenran>
don't know the others
<pushqrdx>
kenran nice thanks, now only ice and sm are missing :D
<{^_^}>
[nixpkgs] @fruit-in opened pull request #107060 → vimPlugins.brainfuck-vim: init at 2020-12-16 → https://git.io/JLndM
<kenran>
pushqrdx: I guessed that they're in the xorg attribute as well, so I did `nix search xorg.*ice` and found xorg.libICE, as well as xorg.libSM
<kenran>
pushqrdx: that looks promising
gxt has joined #nixos
<bqv>
one issue i tend do have, is that I would always love to know what part of my eval caused a build to happen
zupo has joined #nixos
<bqv>
but i'm not sure how to solve that
<bqv>
it's easy to identify a build, it's a derivation/storepath
<bqv>
wait, is that nix why-depends
<bqv>
did i just answer my own question..
<pushqrdx>
kenran nice! i think these were the missing ones, i am facing other cmake issues now but i think they are past these deps, thanks alot for help!
<kenran>
pushqrdx: no problem, I'm glad I could help someone and not just ask questions. I think this is the first time I knew how to look for an answer here xD
<kenran>
pushqrdx: in general, I use a combo of `nix search` and `nix-locate` (which comes with nixpkgs.nix-index) to find such stuff
tlaxkit has quit [Ping timeout: 268 seconds]
<pushqrdx>
kenran thanks for pointing this out, i still used a combination of website and github for search till now lol
<lordcirth>
jasom, so long as the program is pure / deterministic.
<jasom>
lordcirth: it is; just a data transform from text to a database format
<jasom>
And it only needs to be semantically deterministic; i.e. as long as two runs generate interchangably usable values then they don't need to be byte-for-byte identical, no?
<evanjs>
bah. there it is -- /build/broadcom-sta/src/wl/sys/wl_linux.c:1667:6: error: implicit declaration of function 'segment_eq' [-Werror=implicit-function-declaration]
<ixxie>
Anybody here acutally doing some web dev with JS on NixOS, as opposed to merely packaging existing JS modules? I'm looking for some sane development workflow
<ixxie>
I'm trying out node2nix of course, but its a little bit unclear to be how it is meant to be used in the course of developing on NixOS
<ixxie>
Then there seem to be two distinct yarn2nix implementations
<evanjs>
oooo sweet we already 106760 lol
<evanjs>
guess I can try to test that
<{^_^}>
[nixpkgs] @hjones2199 opened pull request #107062 → Indilib: add third party driver support → https://git.io/JLcv6
<notgne2>
ixxie (chat.freenode.net): my personal and biased opinion is to use pnpm, and package with pnpm2nix
<buffet>
how would i package a haskell package (library and executable) as part of my overlay?
<buffet>
just copy pasted a package definition from hackage-packages.nix, but that doesnt seem to work
lunik1 has quit [Quit: :x]
lunik1 has joined #nixos
classicc has joined #nixos
<adisbladis>
energizer: On a very high level, yes
<ixxie>
notgne2, adisbladis - thanks for those tips. Do you happen to know if there are posts somewhere documenting the workflow entailed with these?
<adisbladis>
ixxie: Nope, there isn't
<adisbladis>
ixxie: I'm the author of pnpm2nix
<adisbladis>
And was somewhat involved in planning npmlock2nix
<adisbladis>
So if you have some questions you can ask me
<notgne2[m]>
I also submitted a bunch of patches to it and I think I absolutely ruined the evaluation time of it too, so I can also answer question
nek0 has quit [Remote host closed the connection]
<notgne2[m]>
I might try to fix that eventually, I think I broke it when I was fixing some circular dependency things
<ixxie>
good to know adisbladis :D in that case, I will give these tools a go and try and make notes in a gist which we can then dump in the nixos.wiki for the next victim
<notgne2[m]>
though I'm actually not sure how my patches fixed the errors at all, so it may take me some time to figure out how to do it correctly
<ixxie>
to be continued in the coming weeks; expect much harasment :P
<notgne2[m]>
I think pnpm2nix actually crashes on some unstable versions of Nix too, though I haven't checked recently
<ixxie>
in that case, would it be correct to interpret the workflow as follows: for setup, I clone some project and run npmlock2nix to create the nix files
<adisbladis>
No, you don't need to create nix files
<adisbladis>
npmlock2nix is pure nix
<adisbladis>
It reads package.json / package-lock.json
<ixxie>
ooh
<ixxie>
so it just automagically launches a shell for me and runs builds based on those files?
<ixxie>
and to add dependencies I run `npm add x` as usual?
<adisbladis>
Yes, exactly
<DigitalKiwi>
ohhi just the two i wanted to talk to
<genevino>
is there a way to *NOT* let "nix-shell -p <package>" write to the hard drive, at all? e.g. when you fear your SSD wearing out and you have plenty of ram?
<ixxie>
what about all the stuff in the node_modules folder which npm creates?
<notgne2[m]>
ixxie: I think that should all be ignored
<notgne2[m]>
that's what traditional usage of NPM uses, and if you run your program without Nix that will get used, if you use Nix it will use Nix to pull your dependencies, not whatever is in that directory
<ixxie>
so I would typically idea stop npm from creating that folder somehow? Or perhaps use something like the add-dependency module which adds stuff to the package/-lock.json without installing it?
<evanjs>
hrm, can't seem to get the patch for #106760 working. Trying to build with the PR checked out, no...
<{^_^}>
[nixpkgs] @jonringer pushed commit from @jbalme to master « steam/fhsenv: Add LD_LIBRARY_PATH (and hence OpenGL driver library path) to output of --print-steam-runtime-library-path… »: https://git.io/JLcTb
<notgne2[m]>
ixxie: there's probably some kind of npm command that adds a dependency to the lockfile without actually installing it, though I'm not sure what that is
<numkem>
I'm running into a wall trying to get solaar to work with i3... the GUI works just fine under plasma but it seems to be missing something that I can't figure out on i3.
ransom_ has joined #nixos
eyenx has quit [Remote host closed the connection]
justanotheruser has quit [Read error: Connection reset by peer]
<pushqrdx>
can someone please tell me where is this python3-gi package, it isn't anywhere and i keep getting ModuleNotFoundError: No module named 'gi' in python
werner291 has quit [Quit: werner291]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<pushqrdx>
it's not python gobject as i already have that installed