<pushqrdx>
i tried to change the installation usb nix channel to unstable expecting that it would use it during installation however it used the default 20 channel instead
<pushqrdx>
so i wonder is it possible to install using the unstable channel directly instead of installing, then switch --upgrade
magnetophon has joined #nixos
Fare has quit [Ping timeout: 264 seconds]
<colemickens>
pushqrdx how did you change the channel?
<{^_^}>
[nixpkgs] @orivej-nixos pushed commit from @orivej to master « convertall: init at 0.8.0 (#103108) »: https://git.io/JkJZJ
yeongsheng_tan[m has joined #nixos
jonatanb has quit [Ping timeout: 260 seconds]
Rusty1 has joined #nixos
sulu has joined #nixos
teto has quit [Quit: WeeChat 2.9]
<pushqrdx>
colemickens doing so doesn't appear to have any effect though
magnetophon has joined #nixos
Chiliparrot has quit [Ping timeout: 246 seconds]
<danderson>
anyone done declarative nix-channel management before? I'm picturing adding stuff to my nixos config to manage ~/.nix-defexpr/channels, so that the channels match what the system was build with. Anyone done something like that before?
<danderson>
My experiment with flakes isn't going _super_ well, so I'm toying with how to get the nice things from flakes back in a non-flake world, while the flake world sorts itself out a bit more.
<gchristensen>
I do something you might be in to ...
<colemickens>
I just build '${thing}.drvPath' which gives me the derivation
<danderson>
I see you filed it, but the hacks described in the bug only seem to work for legacyPackages, not nixos configs.
<colemickens>
nixosConfigurations are just nix packages though
<danderson>
I was getting failures. Sec, reproducing
<colemickens>
nixosConfiguration.hostname.config.system.build.toplevel is the derivation for the entire system
<colemickens>
you can build that, copy derivs to remote, build, copy back build output. It's what I do for remote builds.
<danderson>
if I give nixosConfiguration.hostname.config.system.build.toplevel to `nix-build`, I end up with the entire build done locally. What do you build to just get the .drv closure?
<danderson>
...toplevel.drvPath doesn't seem to exist.
<danderson>
This is with: `nix --experimental-features 'nix-command flakes' build ".#nixosConfigurations.vega.config.system.build.toplevel" -o result`
<danderson>
eval of drvPath produces something of use. Presumably I can then nix-copy-closure that, and then nix-build it
<colemickens>
yup
<danderson>
thanks much. Maybe I can stick with flakes then
<colemickens>
np, glad it was helpful
<danderson>
although as of some recent update to nixUnstable, eval seems to have slowed down enough that I'm not sure I'm getting any benefits from the cached evals
<danderson>
but, one problem at a time. And "as slow as pre-flakes, but with flakes" is still an improvement
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to staging-next: https://git.io/JkJn7
<danderson>
the joys of running unstable code :)
<gchristensen>
I've had so many "joys" with nix unstable :').
yinfeng has joined #nixos
Fare has joined #nixos
<pumpy>
on 20.09 i get "lvmconfig failed" error on console. lvm2-activation-generator is the source i guess. how i can fix? i found https://www.mail-archive.com/debian-bugs-dist@lists.debian.org/msg1650584.html and i ran lvmconfig --typeconfig default --withcomments with a grep for event_activation and it's =1 which was the fix they talked about so dunno
<pumpy>
what to do :(
<pumpy>
any idea gchristensen? :/
<colemickens>
(I went ahead and updated the OP of that bug with an explicit example usage of the workaround too)
<danderson>
colemickens: why are you copying to file://, rsyncing, and copying into the store from file:// ? Does ssh:// not work?
<gchristensen>
pumpy: do you use lvm?
<colemickens>
danderson: it's faster than letting nix copy the drvs :(
<pushqrdx>
colemickens i believe you forgot about my question :')
<colemickens>
pushqrdx: I'm sorry, I just don't know. Sorry to leave you hanging like that though.
<pushqrdx>
ah np at all i just thought you forgot to answer :D
<danderson>
colemickens: ouch. I guess I believe that :(
<colemickens>
pushqrdx: if *I* were you, I would build the system config manually with whatever channel you want, and then call `nixos-install --system /nix/store/...` and let it install the build you produced.
<pumpy>
gchristensen nope
<pumpy>
nothing about it in any of my configs
pushqrdx has quit [Quit: Connection closed]
<gchristensen>
ignore it
<pumpy>
"lvm*"
<colemickens>
danderson: fwiw, I recently rewrote this, had reverted back to just naively `nix copy --derivation` the drvs and noticed how slow it was, re-implemented the extra copy/rsync and again noticed that it was faster. So I stand by it (without hard data)
<gchristensen>
or file an issue
<pumpy>
:/ ok ty i guess
<colemickens>
danderson: I didn't try `nix-copy-closure` though, maybe it's better.
<danderson>
colemickens: oh no, I believe it. rsync's reeeally good at copying stuff fast.
<danderson>
good tip, thanks!
<danderson>
seems like I just need to steal most of this shell script.
<gchristensen>
pumpy: please open an issue :)
Fare has quit [Ping timeout: 264 seconds]
<gchristensen>
pumpy: so I can "me also" it
h0m2 has joined #nixos
<pumpy>
gchristensen well i woulda but i dont have github account :/
<gchristensen>
ah bummer
<pumpy>
ya sorry to punt, not being lazy
h0m1 has quit [Ping timeout: 265 seconds]
bn_work has quit [Quit: Connection closed for inactivity]
spease has quit [Remote host closed the connection]
<danderson>
just above, you realize on the remote host without cachix, push to cachix (assuming keys etc.), then realize again... Isn't that always going to be a no-op?
<pumpy>
omasanori[m] you mind making github issue like gchristensen asking?
<pumpy>
short version...
<pumpy>
on 20.09 i get "lvmconfig failed" error on console. lvm2-activation-generator is the source i guess. how i can fix? i found https://www.mail-archive.com/debian-bugs-dist@lists.debian.org/msg1650584.html and i ran lvmconfig --typeconfig default --withcomments with a grep for event_activation and it's =1 which was the fix they talked about so dunno
<pumpy>
what to do :(
<colemickens>
danderson: Hm. 118 is the first time I realize on $target (note $target != $remote), on 79-82, I do a trial realization on $target, but that only works if $out already exists in cachix or on the $target.
<danderson>
ooooh, remote != target
<danderson>
I see. So, build on a build host, push to cachix, then deploy to target by way of cachix
leungbk has joined #nixos
vidbina_ has joined #nixos
<colemickens>
danderson: yes, and then there's untested support for not using cachix
<danderson>
Got it. Thanks for helping my reading comprehension :)
<colemickens>
np! Another thing of note about this script, (that took extra care due to bashism, etc) is the usage of nix-shell on the remote too (hence, this works on old Packet images, newer EC2 instances, and custom Azure instances with unstable enabled, etc)
m1cr0man has quit [Ping timeout: 240 seconds]
<colemickens>
(I am wondering if it's worth extracting this into a little go app or something? but also might be overkill for what this is.)
<danderson>
hmm, interesting
<gchristensen>
go app to do what colemickens?
<colemickens>
(and the shuttling around of the buildArgs to ensure we always get cache hits, even if the remote builder's nix-daemon doesn't already know about all of the binary caches we might want)
<danderson>
That does resolve something that was eating at me, how to bootstrap a machine for the first time into this nixUnstable world
<danderson>
that printf incantation does look quite scary :)
<colemickens>
I've seen a number of less comprehensive versions of this script around.
<gchristensen>
mm
<colemickens>
danderson: yes. it wasn't fun, and I think it doesn't work if you need to use a pipe in the remote command, but I'd had enough at that point :)
<danderson>
annoyingly, making that part Go wouldn't help much, SSH's protocol for executing commands is "gimme a string that I'll pass to the local shell"
mmohammadi9812 has quit [Ping timeout: 272 seconds]
<danderson>
I would certainly use such a tool if it existed. As it is, I'm going to crib bits and pieces of colemickens's shell script and adjust it to my subset of use cases
<danderson>
extra points for just making it part of `nix` or `nixos-rebuild` ;)
<omasanori[m]>
pumpy: OK, I will do. Thank you for your suggestion.
ransom has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<yjftsjthsd>
I would like to have a CI job to check my nixos configuration.nix; on my laptop I can just do `nixos-rebuild dry-build`, so I thought this would be easy, but the docker image nixos/nix doesn't seem to include nixos-rebuild, so I'm not sure how to get gitlab to run what I need. Does anyone happen to know how to do this (or better yet, am I being
<yjftsjthsd>
dumb and there's a better way)?
leungbk has joined #nixos
<colemickens>
yjftsjthsd: you can do one of two things, you can try to start a nix-shell that contains the nixos-rebuild script, or you can use `nix-build` to build the same thing that `nixos-rebuild` builds.
stephank1 has joined #nixos
<steell>
has anyone else been experiencing extremely slow + janky page loads with Chrome after a recent nixos-unstable update?
<{^_^}>
[nixpkgs] @jonringer pushed commit from @freezeboy to master « wordpress: 5.5.1 -> 5.5.3 »: https://git.io/JkJRf
stephank has quit [Ping timeout: 264 seconds]
stephank1 is now known as stephank
Taneb has quit [Quit: I seem to have stopped.]
domogled has quit [Ping timeout: 258 seconds]
Taneb has joined #nixos
iqubic has left #nixos ["ERC (IRC client for Emacs 28.0.50)"]
yinfeng has quit [Ping timeout: 246 seconds]
<yjftsjthsd>
Okay, newbie follow-up: How do I find the package containing nixos-rebuild? Following symlinks to the one in my path on nixos gives me /nix/store/[hash]-nixos-rebuild but `nix-shell -p nixos-rebuild` claims that that's not it, and `nix search nixos-rebuild` gives me nothing.
<yjftsjthsd>
I can't seem to find an equivalent to `yum whatprovides '*bin/programname` for nix
lejonet has quit [Ping timeout: 240 seconds]
<colemickens>
yjftsjthsd: it's sort of a weird package, it's a bit unique
<colemickens>
I think that'll maybe give you what you want?
<yjftsjthsd>
Yes, that does seem to give me nixos-rebuild, although running in docker it seems upset that there is no systemctl (?), which is annoying since I just want to dry-run, but I guess I get it.
<yjftsjthsd>
Is that documented, or ... honestly, what does "(nixos{}).nixos-rebuild" even mean? I assume "give me the nixos-rebuild from the nixos module(?)"?
jonatanb has joined #nixos
<colemickens>
yjftsjthsd: nixos is a special function that composes a nixos system, and then it just happens that it has an attribute hanging off of it that is a derivation tht provides a script nixos-rebuild
jonatanb has quit [Ping timeout: 240 seconds]
<colemickens>
I'm not sure how you'd know this without reading source, playing with a repl, or reading a Discourse thread (which is where I saw it)
<yjftsjthsd>
Okay, so very special case. Thx
plakband has joined #nixos
<yjftsjthsd>
That's fair:) While I have some opinions about nix being hard to learn, this one honestly seems pretty reasonable
<danderson>
on nix unstable, is there some way of forcing it to assume a color/etc. capable terminal? When I realize over ssh, I get none of the pretty progress info.
astronavt has quit [Quit: Leaving]
<energizer>
do not know how to unpack source archive, but it isn't an archive, why is it tryign to unpack
<energizer>
ssh -t might help danderson
jonatanb has joined #nixos
zupo has joined #nixos
jonatanb has quit [Ping timeout: 246 seconds]
domogled has joined #nixos
jonatanb has joined #nixos
<Ke>
on reboot all my nixos containers always timeout on shutdown, any idea what could be going on
yinfeng has joined #nixos
jonatanb has quit [Ping timeout: 256 seconds]
waleee-cl has quit [Quit: Connection closed for inactivity]
<yjftsjthsd>
Are they equally slow if you just try to stop them normally, or is it only on shutdown?
<Ke>
if I stop them manually they shutdown in a couple of seconds
<Ke>
ie machinectl login and shutdown -h now
<yjftsjthsd>
How about from the outside? Looks like `machinectl poweroff foo`, although I don't really know machinectl
<{^_^}>
[nix] @dependabot[bot] pushed to dependabot/github_actions/cachix/install-nix-action-v12 « Bump cachix/install-nix-action from v11 to v12 »: https://git.io/JkJ2U
<{^_^}>
[nix] @dependabot[bot] opened pull request #4237 → Bump cachix/install-nix-action from v11 to v12 → https://git.io/JkJ2T
<{^_^}>
[hydra] @dependabot[bot] pushed to dependabot/github_actions/cachix/install-nix-action-v12 « Bump cachix/install-nix-action from v10 to v12 »: https://git.io/JkJ2I
<{^_^}>
[hydra] @dependabot[bot] opened pull request #831 → Bump cachix/install-nix-action from v10 to v12 → https://git.io/JkJ2L
<{^_^}>
[hydra] @dependabot[bot] closed pull request #819 → Bump cachix/install-nix-action from v10 to v11 → https://git.io/JUiwu
<{^_^}>
[hydra] @dependabot[bot] pushed 0 commits to dependabot/github_actions/cachix/install-nix-action-v11: https://git.io/JkJ2q
sangoma has joined #nixos
cmk_zzz has quit [Ping timeout: 246 seconds]
cmk_zzz has joined #nixos
sangoma has quit [Ping timeout: 272 seconds]
captn3m0 has joined #nixos
plakband has quit [Quit: WeeChat 2.9]
captn3m0- has quit [Ping timeout: 265 seconds]
sangoma has joined #nixos
kalbasit has quit [Ping timeout: 256 seconds]
danderson has quit [Remote host closed the connection]
<yjftsjthsd>
Then I'm out of good ideas, sorry. If I were a nix expert I'd probably suggest looking at what the system does to tell containers to shut down when it's shutting down, but I have no clue how to track that down
<Ke>
that would be container@*.service
<yjftsjthsd>
Can you verify that the stop action there is `machinectl poweroff` or equivalent?
<Ke>
yes it is
Fare has quit [Ping timeout: 244 seconds]
<yjftsjthsd>
huh
<yjftsjthsd>
Then I am *definitely* out of ideas, sorry
Guest71661 is now known as Theuni
nicoo has quit [Remote host closed the connection]
zupo has quit [Ping timeout: 256 seconds]
danderson[m] has joined #nixos
zupo has joined #nixos
nicoo has joined #nixos
<Ke>
well I guess I'll work around this with a wrapper script, like I deal with most problems
<Ke>
not that waiting until timeout was a big issue, at least reboot is now fast
<{^_^}>
[nixpkgs] @expipiplus1 opened pull request #103183 → Fix GHC bootstrap in pkgsMusl and include patch for binutils/16177 → https://git.io/JkJod
eoli3n has joined #nixos
zupo has joined #nixos
lejonet has joined #nixos
gxt has quit [Ping timeout: 240 seconds]
oida has quit [Ping timeout: 240 seconds]
est31 has quit [Ping timeout: 240 seconds]
gxt has joined #nixos
oida has joined #nixos
est31 has joined #nixos
cjpbirkbeck has joined #nixos
cosimone has quit [Remote host closed the connection]
cosimone has joined #nixos
jonringer has quit [Ping timeout: 264 seconds]
<bk1603[m]>
How can I install mkosi on nixos?
justanotheruser has quit [Ping timeout: 268 seconds]
<Ke>
,locate bin/mkosi
<{^_^}>
Couldn't find in any packages
<Ke>
I guess you need to package it yourself or something
spudly1 has quit [Quit: killed]
<bk1603[m]>
Oh I see, I might try that sometime later :) I just wanted to check if I was missing something for now. I think I'll look for alternatives.
<{^_^}>
[nixpkgs] @mmahut opened pull request #103184 → asciigraph: init at 0.5.1 → https://git.io/JkJPz
growpotkin has joined #nixos
<bqv>
what needs to happen for there to be better support for musl
<bqv>
just curious
sangoma has quit [Read error: Connection reset by peer]
<bqv>
pkgsStatic.brotli is broken on my nixpkgs, so nix-static won't build. it's just not currently reasonable to depend on pkgsStatic
growpotkin has quit [Client Quit]
<{^_^}>
[nixpkgs] @ilmari-lauhakangas opened pull request #103185 → Use --without-system-libnumbertext instead of --disable-libnumbertext → https://git.io/JkJXq
niksnut has joined #nixos
cr4y1__ has joined #nixos
<{^_^}>
[nixpkgs] @vbgl opened pull request #103186 → ocamlPackages.mstruct: remove at 1.4.0 → https://git.io/JkJXr
mallox has joined #nixos
hyper_ch5 has joined #nixos
sangoma has joined #nixos
hyper_ch4 has quit [Ping timeout: 240 seconds]
jonatanb has joined #nixos
jonatanb has quit [Ping timeout: 265 seconds]
saschagrunert has joined #nixos
s34m has joined #nixos
<{^_^}>
[nixpkgs] @roberth merged pull request #103175 → Add utdemir to CODEOWNERS for dockerTools → https://git.io/JkJv0
euandreh has quit [Remote host closed the connection]
<colemickens>
castingmalloc: do you still need help with the last question? You'd just add it something like this: `nixpkgs.overlays = [ (import "${inputs.yourOverlayInput}/default.nix") ];`
h0m1 has joined #nixos
Chiliparrot has quit [Ping timeout: 244 seconds]
Chiliparrot has joined #nixos
castingmalloc has quit [Ping timeout: 245 seconds]
noonien has quit [Quit: Ping timeout (120 seconds)]
noonien has joined #nixos
Ariakenom has quit [Ping timeout: 272 seconds]
s34m has quit [Quit: s34m]
Maxdamantus has quit [Ping timeout: 258 seconds]
<{^_^}>
[nixpkgs] @dasJ opened pull request #103196 → nixos/plasma5: Fix when running without aliases → https://git.io/JkJFA
Maxdamantus has joined #nixos
s34m has joined #nixos
s34m has quit [Client Quit]
nonfinito has joined #nixos
zupo has joined #nixos
s34m has joined #nixos
s34m has quit [Client Quit]
<castingmalloc>
colemickens: could not figure it out, I get an error `attribute 'currentSystem' missing` and when I have system as parameter like this `(import "${neovim-nightly}/default.nix" { inherit system; })`, I get `attempt to call something which is not a function but a set`
<colemickens>
castingmalloc: when you're using flakes, 'currentSystem' is not allowed
<colemickens>
I don't see currentSystem in that overlay though?
<maralorn>
When I add a new dependency to my flake, which then is missing in my flake.lock, will the next run update my whole flake.lock or just add that dep?
<colemickens>
sorry that was meant for castingmalloc ^
FRidh has joined #nixos
werner292 has joined #nixos
<NieDzejkob>
maralorn: value is a function while a set was expected, at (string):1:1
werner291 has quit [Ping timeout: 268 seconds]
werner292 is now known as werner291
<NieDzejkob>
I also suspect that it is closer to opening a shell with deps of foo than a shell with foo itself
castingmalloc has quit [Ping timeout: 245 seconds]
Jackneilll has joined #nixos
Jackneill has quit [Read error: Connection reset by peer]
jonatanb has quit [Ping timeout: 265 seconds]
castingmalloc has joined #nixos
rprije has quit [Ping timeout: 260 seconds]
kreyren_ has quit [Remote host closed the connection]
kreyren_ has joined #nixos
Chiliparrot has quit [Ping timeout: 260 seconds]
hyper_ch4 has joined #nixos
<maralorn>
NieDzejkob: Yeah, tbh I can never figure out, when nix-shell does what … I now it does what you fear for -A, but thought maybe not for -E? Anyways I think in situations like this I often try "nix run" instead, but it's not much less arcane to me.
simonpe^^ has joined #nixos
<maralorn>
But maybe try "(import ./.).foo {}"
hyper_ch5 has quit [Ping timeout: 268 seconds]
<maralorn>
Or "(import ./. {}).foo" that makes probably more sense <- NieDzejkob
thibm has joined #nixos
<simonpe^^>
hey, why is it that in nixos if I create a file as a regular user and chmod it 777, then `sudo -u nixbld1 cat file` fails with permission denied?
castingmalloc has quit [Ping timeout: 245 seconds]
<simonpe^^>
I've checked `sestatus` and it tells me selinux is disabled
gthm has quit [Quit: leaving]
castingmalloc has joined #nixos
gthm has joined #nixos
<regnat>
simonpe^^: A parent directory that doesn't have a+x set?
griff_ has quit [Quit: griff_]
<patagonicus>
simonpe^^: Yeah, that's how I was able to reproduce it. Try with a file in /tmp, which should be 1777. In my $HOME it fails since $HOME is 700.
castingmalloc has quit [Ping timeout: 245 seconds]
<{^_^}>
[nixpkgs] @andir pushed 2 commits to release-20.03: https://git.io/JkJp2
castingmalloc has joined #nixos
<castingmalloc>
colemickens: I might just drop it wait for the official release in december
griff_ has quit [Ping timeout: 260 seconds]
<thibm>
Hey, I updated nix from 2.2.2 to 2.3.8 on a machine in single user-mode and store at a custom location, set with "store = /some/prefix" in ~/.config/nix/nix.conf
pta2002 has joined #nixos
<pta2002>
Hello
<pta2002>
I'm having some issues with bundlerEnv
<pta2002>
I'm trying to install something which depends on libv8 and mini_racer
<thibm>
The new nix ignore this conf and try to use /nix/store, for instance if I do "nix-build '<nixpkgs>' -A hello" it fetches hello from the cache and create a bogus symlink "result" to "/nix/store/…". Is the "store" option not supported anymore?
<thibm>
Can't find it in the manuals and I don't find useful clues in the github issues
<pta2002>
I got libv8 to work, but now mini_racer seems to be trying to install libv8 again, instead of taking into account the things i passed into it
davidv7_ has quit [Quit: Konversation terminated!]
__monty__ has joined #nixos
<pta2002>
has anyone had this issue? I know mini_racer depends on libv8, but I don't know why it's trying to do libv8 from scratch instead of using the one that's already there
<pta2002>
it's the same version too
iqubic has joined #nixos
griff_ has joined #nixos
<simonpe^^>
regnat: patagonicus: yeah I verified it works in /tmp
<iqubic>
I'm having issues with aspell on my system. Whenever I try to check the spelling of a file, I get the following error: The file "/run/current-system/sw/lib/aspell/en_US-science.rws" can not be opened for reading.
<iqubic>
And the reason this file can't be opened is because it doesn't exist.
<iqubic>
I don't know what's going on, but it's not working.
berberman has quit [Ping timeout: 260 seconds]
berberman has joined #nixos
jonatanb has joined #nixos
<iqubic>
Anyone know what to do to fix this?
<thibm>
OK, the "store" option seems to be defined in the unstable 2.4 manual from the last Hydra build (BTW store is a terrible choice for this option as it's difficult to search/grep) but not in 2.3
jonatanb has quit [Ping timeout: 256 seconds]
<thibm>
Even setting "--option store /tmp" on the nix-build command line does create /tmp/nix/{sore,var} BUT the symlink still points to /nix/store/…hello. Strange. Custom store seems really broken in 2.3.8
davidv7 has joined #nixos
<thibm>
iqubic: did you installed aspellDicts.en-science?
<iqubic>
No.
<iqubic>
I actually don't have that one installed, because it has issues.
<thibm>
lib/aspell/en_US-science.rws belongs to this package
fling has quit [Ping timeout: 260 seconds]
<thibm>
Try to figure out why aspell looks for it if you don't need it
<{^_^}>
[nixpkgs] @thefloweringash closed pull request #98541 → darwin (Big Sur): tbd based stdenv → https://git.io/JUg8z
<{^_^}>
[nixpkgs] @thefloweringash reopened pull request #98541 → darwin (Big Sur): tbd based stdenv → https://git.io/JUg8z
<{^_^}>
[nixpkgs] @mickours opened pull request #103200 → docker: Fix broken pipe on docker layer creation → https://git.io/JkJj6
castingmalloc has quit [Remote host closed the connection]
fling has joined #nixos
Dotz0cat has joined #nixos
superherointj has joined #nixos
Dotz0cat_ has quit [Ping timeout: 256 seconds]
cfricke has quit [Quit: WeeChat 2.9]
werner291 has quit [Ping timeout: 246 seconds]
vuko has joined #nixos
andreas31 has quit [Write error: Connection reset by peer]
kreyren_ has quit [Write error: Connection reset by peer]
gxt has quit [Write error: Connection reset by peer]
est31 has quit [Read error: Connection reset by peer]
joey has quit [Remote host closed the connection]
rojiro has quit [Write error: Broken pipe]
oida has quit [Read error: Connection reset by peer]
nicoo has quit [Read error: Connection reset by peer]
jb55 has quit [Write error: Connection reset by peer]
est31 has joined #nixos
<pta2002>
Well I give up on this ruby thing
rojiro has joined #nixos
joey has joined #nixos
<pta2002>
Is it possible to define a dockerfile that builds a docker image from a dockerfile?
oida has joined #nixos
<pta2002>
Was trying to port said dockerfile to a nix derivation but i couldn't get libv8 to build properly so i'm just using an ubuntu image now
jb55 has joined #nixos
<dminuoso>
12:31:05 pta2002 | Is it possible to define a dockerfile that builds a docker image from a dockerfile?
<dminuoso>
That's precisely what dockerfiles are for..
<dminuoso>
Building images. :)
<pta2002>
uh
gxt has joined #nixos
<pta2002>
a nix expression*
<pta2002>
lol
<pta2002>
that builds a dockerfile
<dminuoso>
Yes
<pta2002>
what should I do? Can't find any examples of it
<pta2002>
dminuoso: i know there's virtualisation.oci-containers
<pta2002>
but it also happens that one of those images was built by me, with my own dockerfile, so i was porting it to nix, but then I couldn't get it to build libv8 with the flags i passed it
<dminuoso>
pta2002: What exactly doesnt build with libv8? What's the error?
<{^_^}>
[nixpkgs] @alyssais pushed commit from @seppeljordan to master « nix-prefetch-github: 4.0 -> 4.0.1 »: https://git.io/JkUJ1
<dminuoso>
Can you perhaps share your derivation and the error?
<pta2002>
i sent the gist link above
<pta2002>
but basically it says this
<pta2002>
/nix/store/qlfvsqk17in6s3fhaf5an9q7063i8aih-ruby2.6.6-libv8-8.4.255.0/lib/ruby/gems/2.6.0/gems/libv8-8.4.255.0/ext/libv8/location.rb:57:in `configure': By using --with-system-v8, you have chosen to use the version (Libv8::Location::System::NotFoundError)
<pta2002>
of V8 found on your system and *not* the one that is bundled with
<pta2002>
oops
<pta2002>
the libv8 rubygem.
<pta2002>
However, your system version of v8 could not be located.
<{^_^}>
[nixpkgs] @alyssais pushed commit from @r-ryantm to master « groovy: 3.0.3 -> 3.0.6 »: https://git.io/JkUTF
<eyJhb>
If I have a intermediate option, e.g. lab, which has two submodules as such `lab = { networks = mkOption { ... }; virtuals = mkOption { ... }; }`, because then how can I set a default value for lab?
<jordandoyle>
hey, wondering if anyone knows why iptables/bash/systemd would start building from scratch if I `nix-shell -p hugo` or `nix-shell -p nodejs`? `nixos-rebuild switch --upgrade` pulls from the binary cache like normal (on the unstable channel)
<typetetris>
jordandoyle: May the user issuing the `nix-shell` has different channels set? Maybe even some overlays in `~/.config/nixpkgs/config.nix` or so?
werner292 has joined #nixos
werner291 has quit [Ping timeout: 246 seconds]
werner292 is now known as werner291
<{^_^}>
[nixpkgs] @marsam merged pull request #103131 → dogdns: fix build on darwin → https://git.io/JkfGY
<jordandoyle>
user channel list is empty & user nixpkgs/config.nix only has `allowUnfree = true` unfortunately - could the allow unfree be modifying something that's a dependency of _everything_? :thinking:
<{^_^}>
[nixpkgs] @raboof opened pull request #103209 → mailutils: update guile to current version → https://git.io/JkUnv
<{^_^}>
[nix] @edolstra pushed to master « Fix appending to Setting<StringSet> »: https://git.io/JkUnJ
hyper_ch4 has joined #nixos
<AmandaC>
Isn't there a nixpkgs helper for generating .desktop files?
<xineel>
hey there, I upgraded from macOS 10.15.6 to .7 and I can't use nix-env &co anymore. Is there a known fix procedure, without removing and reinstalling Nix entirely?
<pta2002>
I'm trying to use virtualisation.oci-containers to manage a podman container and I want to make it connect to the system's postgresql instance
<pta2002>
but i can't figure out how to do that
<simpson>
What have you tried, and how is it not working?
<{^_^}>
[nixpkgs] @peterhoeg opened pull request #103214 → plasma-thunderbolt: init at 5.18.5 → https://git.io/JkURs
<pta2002>
really just creating the container, don't even know where to start with this
<pta2002>
it just can't connect to postgres
<pta2002>
and im trying to connect to localhost:5432 from the container
<pta2002>
i think i'll just run psql on another container
<pta2002>
gotta figure out how to run them on the same pod though
<{^_^}>
[nixos-homepage] @github-actions[bot] pushed commit from GitHub Actions to master « Update flake.lock and blogs.xml [ci skip] »: https://git.io/JkURA
justanotheruser has quit [Ping timeout: 264 seconds]
eahlberg has quit [Quit: Lost terminal]
stree has quit [Quit: Caught exception]
stree has joined #nixos
<gchristensen>
anyone on 20.09 + sway + waybar with working icons?
<thibm>
(For the record, I solved my custom store problem I talked about earlier today with overriding nix storeDir and stateDir derivation arguments. I completely forgot I had to do that, but to honest it's not really documented. But to be honest, again, custom stores are not common)
<thibm>
And it has nothing to do with nix.conf's store option
<Cadey>
in a nixos config, can you conditionally import a folder if the config is being built by nixops?
unrooted[m] has quit [Quit: Idle for 30+ days]
<dminuoso>
Cadey: That sounds a bit strange. What are you trying to do?
FRidh has quit [Quit: Konversation terminated!]
<Cadey>
import service modules that need nixops features (keys/secrets) only if nixops is being used, so I can import my services tree in the default/happy path of my configs and it will only be used when Nixops is deploying a config
<clever>
Cadey: only put those into the imports list when using a nixops deployment.nix file
<srhb>
Cadey: The canonical way would be two different entry points.
<gchristensen>
I'm still adding some bits to it, but my i3bar is generated by nix too, to cover timezones based on where I am. I'd love to see yours
<V>
screenshot or nix config?
<gchristensen>
first one then the other? :)
<V>
heh
aveltras has quit [Quit: Connection closed for inactivity]
magnetophon has joined #nixos
bitmapper has joined #nixos
otavio__ has joined #nixos
otavio has quit [Read error: Connection reset by peer]
astro has quit [Read error: Connection reset by peer]
hnOsmium0001 has joined #nixos
astro has joined #nixos
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
Chiliparrot has joined #nixos
heyitsrama has joined #nixos
astro has quit [Ping timeout: 240 seconds]
kalbasit has joined #nixos
leungbk has joined #nixos
justanotheruser has joined #nixos
thibm has quit [Quit: WeeChat 2.6]
<Taneb>
I'm confused by something on the ZFS on Linux wiki page
<Taneb>
"If the key is not on the root filesystem, you will also need to set zfs-import-poolname.serviceConfig.RequiresMountsFor=/path/to/key, where poolname is the name of the data pool. This makes sure that systemd will mount the filesystem for /path/to/key first before importing the zfs pool."
leungbk has quit [Ping timeout: 264 seconds]
<Taneb>
Hang on, I'm confused why I'm confused
domogled has quit [Ping timeout: 264 seconds]
<Taneb>
OK, I don't seem to have that service
<srhb>
Taneb: iirc that only get generated for supplementary zpools
<srhb>
Taneb: ie. not that one with root on it.
<Taneb>
I don't have / in ZFS
<Taneb>
This is for a zpool which has things for /nix and a couple of other mount points
<Taneb>
Unless I've misunderstood what root means in this context
<tpw_rules>
is there any way to set up a remote desktop of some kind that lets me view display 0 remotely? i don't particularly care what DE or WM or protocol. i just need it to be GPU accelerated
<srhb>
I should clarify: Pools that are not required for building /
<srhb>
Taneb: So, if it's needed for boot, it's all stage 1, no import unit.
<Taneb>
Boot is failing, I have to go into the rescue console thing and do "zfs load-key -a"
<srhb>
Taneb: It's not prompting you automatically?
<Taneb>
It's prompting me, accepting what I enter, and then that's happening
<srhb>
That rings a bell... hmm.
<srhb>
Taneb: Is this a VM?
<Taneb>
No
jonatanb has quit [Remote host closed the connection]
jonatanb has joined #nixos
jonatanb has quit [Remote host closed the connection]
jonatanb has joined #nixos
<srhb>
All I remember is something like different devNodes for zpool discovery behaving.. weird. But if you get the prompt I am a bit confussled for sure. stage1 log is probably useful
<srhb>
Taneb: I think it'll be in journalctl -k if all is well
<srhb>
Taneb: Headered stage-1-init
jonatanb has quit [Ping timeout: 256 seconds]
<Taneb>
Ah, so it is
werner291 has quit [Ping timeout: 246 seconds]
werner292 is now known as werner291
pushqrdx has joined #nixos
<srhb>
But sometimes I find I need to run stage 1 by hand to really understand what's going on.
cheriimoya has joined #nixos
ajs124 has joined #nixos
mrpi has joined #nixos
Scriptkiddi has joined #nixos
das_j has joined #nixos
<pushqrdx>
i think some kind of update to nixos broke my ability to switch, i keep getting `The unique option `console.keyMap' is defined multiple times. Definition values:
<pushqrdx>
- In `/etc/nixos/configuration.nix': "us"
<pushqrdx>
- In `/nix/var/nix/profiles/per-user/root/channels/nixos/nixos/modules/config/console.nix': <derivation /nix/store/61c29pn1wj6c8xrsw30ywbwicmr5mgrw-xkb-console-keymap.drv>
<pushqrdx>
`
<pushqrdx>
i even commented out those in my nixos configuration and still getting the error
<clever>
pushqrdx: are you sure you commented it out? can you pastebin the new configuration.nix?
<pushqrdx>
clever by bad sudoedit didn't save the file, however i never had that problem before i want to specify the laytout and useXkbConfig in my configuration
rajivr has quit [Quit: Connection closed for inactivity]
<clever>
pushqrdx: have a look at /nix/var/nix/profiles/per-user/root/channels/nixos/nixos/modules/config/console.nix, what is it doing with console.keyMap?
<Taneb>
I think it's because it's trying to read a key-file that hasn't been mounted yet?
sangoma has quit [Ping timeout: 260 seconds]
<srhb>
Taneb: Wait, is this two layers of encryption?
<pushqrdx>
clever it appears to be the normal console.nix you can find on github, it has some default documented stuff for console, and #implemenentation section
pushqrdx has quit [Quit: Connection closed]
<Taneb>
srhb: not intentionally!
<srhb>
Taneb: I'm confused. How's your setup supposed to work? What's encrypted? :)
cfricke has quit [Quit: WeeChat 2.9]
<lordcirth>
I have my PC configured by a flake. I want to destroy a container that is not configured in that flake, or anywhere else that I can find. But nixos-container destroy says "cannot destroy declarative container (remove it from your configuration.nix instead)"
<Taneb>
One of the zfs things, which is mounted at /storage/encrypted
<Taneb>
Possibly also /, with LUKS, which is ext4
<srhb>
Taneb: Ok, right, so it looks like LUKS is working.
<srhb>
Taneb: Is "storage" needed for boot?
<Taneb>
/nix is a ZFS thing (I need to learn the terminology... "disk"?) in the storage zpool
<srhb>
dataset
<srhb>
And storage/zfs is not encrypted?
<Taneb>
What's storage/zfs here? I don't have anything with that name...
<srhb>
Sorry, storage/nix
<srhb>
Brainfart on my part.
<Taneb>
Yeah, storage/nix is not encrypted
<srhb>
Taneb: Do you have legacy mount set up on each dataset?
<Taneb>
Yes
<srhb>
Taneb: Positively? Because it looks like it wants to try and decrypt your encrypted dataset even though I don't think it should even be attempted mounted yet (since it isn't needed yet) but I may need to check the code..
<Taneb>
Positively. If I do "zfs list" the "MOUNTPOINT" colum is all "legacy"
<Taneb>
I think it tries to do the whole pool at once
<srhb>
That may be a misfeature in the zpool stage1 loader then..
<srhb>
Taneb: And you can also verify that that's the only encrypted dataset with `zfs get encryption storage` ?
<srhb>
Taneb: Say, if it's on the LUKS partition, you would need to do luks decrypting earlier... Which _may_ be possible using preLVM, though that's a bit of a hack
<Taneb>
srhb: I don't have any moral objections to make it a passphrase instead, if that's technically possible
<srhb>
I think it may still be technically possible, but I think you may need to tweak postDeviceCommands by hand
<srhb>
We need a proper DAG for these stage1 shenanigans :-)
<Taneb>
The key file is indeed on the LUKS drive
<srhb>
Ok, yeah, then I think your diagnosis is spot on, sorry it took me a while to catch up :P
<iqubic>
Something very strange is going on with my system. I just logged out of my graphical session, and started a new one via LightDM. Now, when I open a terminal and try to run "which nix-env" I get the following error:
<iqubic>
which: no nix-env in (/run/wrappers/bin:/home/avi/.local/share/flatpak/exports/bin:/var/lib/flatpak/exports/bin:/home/avi/.nix-profile/bin:/etc/profiles/per-user/avi/bin:/nix/var/nix/profiles/default/bin:/run/current-system/sw/bin)
evanjs has joined #nixos
<ldlework>
I have a package that ships a cli that calls nix-build. It also ships some nix. How can I have my cli call nix-build on a user nix expression, such that I am passing the package nix to the function in the user nix file?
kaivai has joined #nixos
<iqubic>
What the hell is going on there.
domogled has quit [Ping timeout: 272 seconds]
<iqubic>
Basically if I try running "which" on any program that I have installed on my computer, I get told that it doesn't exist.
<m1cr0man>
Can someone please enlighten me on how to fix "gpg: agent_genkey failed: No pinentry"? I have a brand new install, trying to generate a key, and it just won't work. Existing tickets haven't been any help.
<m1cr0man>
There's like 3 tickets and 2 discourse discussions on this
<{^_^}>
[nixpkgs] @gebner opened pull request #103225 → pulseaudio: add hsphfpd support → https://git.io/JkUSu
mmohammadi9812 has quit [Ping timeout: 256 seconds]
<srhb>
m1cr0man: Is this with the agent enabled through programs.gnupg.enable?
<m1cr0man>
srhb: Yes and no. Tried both
<cole-h>
Have you tried `gpg-connect-agent updatestartuptty /bye`?
<cole-h>
Sometimes I have to do that.
<m1cr0man>
no I will try that
<m1cr0man>
No dice. cole-h do you have programs.gnupg.enable? I will try with that
pbb has quit [Ping timeout: 240 seconds]
<cole-h>
I have the agent enabled
freezeboy has joined #nixos
kloenk has quit [Ping timeout: 272 seconds]
<m1cr0man>
Ok cool. I also noticed gpgconf --check-programs says pinentry isn't installed, and it's looking in the gnupg derivation
<cole-h>
Yeah, enabling the agent should do that automagically IIRC
<freezeboy>
Hi, can someone that can run nixpkgs-review help me to know if a package can be removed (I will definitely buy a new computer)
kloenk has joined #nixos
<m1cr0man>
Nah still no luck. Screw it, im going to override gnupg and add pinentry
mmohammadi9812 has joined #nixos
<srhb>
m1cr0man: fwiw manipulating these settings might require logging out and in.
<srhb>
m1cr0man: (At least for sockets)
<m1cr0man>
srhb: Yeah I've cycled my ssh session and killed/restarted gpg-agent each time
<srhb>
m1cr0man: ok :)
pbb has joined #nixos
<m1cr0man>
Really annoying haha because you fill out the gpg --full-generate-key form and it errors at the end
sangoma has joined #nixos
<jasom>
Is there a Right Way(tm) to deal with secrets in configuration files on nixos? I tend to just keep the secrets in /etc/nixos, but I do se a few modules that use other tricks to get the files in /etc without putting them in the world-readable nix store.
<dminuoso>
jasom: no.
<dminuoso>
jasom: nixops + deployment keys is one trick
domogled has quit [Ping timeout: 260 seconds]
<sss2>
hi all, how to install fresh version of firefox-devedition ? preferable with alsa instead of pulseaudio
<m1cr0man>
Personally I copy any secrets I need into /var/secrets, and where possible make the application read from there directly. I do also have a nix file in there for secrets that _have_ to be text in the config, but the side effect is they end up plain text in the store somewhere
cosimone has quit [Remote host closed the connection]
jonatanb has joined #nixos
cosimone has joined #nixos
<jasom>
sss2: alsa or just apulse?
<{^_^}>
[nixpkgs] @Mazurel opened pull request #103227 → devour: init at 12 → https://git.io/JkU91
<freezeboy>
by the way, how much ram is needed to use nixpkgs-review ?
<sss2>
jasom, better alsa, i guess pulse wrapper is also acceptable, never tested it
<jasom>
sss2: If firefox even has alsa-only as an option, it will require compiling rather than starting from a mozilla binary, so you'll probably want to start from the firefox nix expression and make your changes there. I think all the devedition firefox in nixpkgs are binary, so those won't be of help.
cosimone has quit [Remote host closed the connection]
<pumpy>
like 10 lan machines/vms and a few machines in each of like 5 different clients i manage servers for
rnea has joined #nixos
<lordcirth>
pumpy, and are they all NixOS, or not?
<{^_^}>
[nixpkgs] @Ekleog merged pull request #95879 → arcanist: copy directly from $PWD to pick up any applied patches → https://git.io/JJhAK
<{^_^}>
[nixpkgs] @Ekleog pushed commit from @cpcloud to master « arcanist: copy directly from $PWD to pick up any applied patches (#95879) »: https://git.io/JkU7h
chour has quit [Quit: WeeChat 2.9]
<pumpy>
mix of nixos and openbsd
<neonfuz2>
the thought of using ansible to install nixos stuff is funny
<neonfuz2>
I mean whatever works though lol
jonatanb has quit [Ping timeout: 272 seconds]
<neonfuz2>
man now I have the G'MIC plugin in gimp but the specific option I'm trying to use isn't here
<neonfuz2>
I hope it wasn't removed, wonder if it was renamed?
<pumpy>
well ansible just syncs my .nix files up then runs nixos-rebuild switch
<pumpy>
need to expand on that to get config files in for home manager to use
astronavt has quit [Quit: Leaving]
<pumpy>
when i put home.packages = [i3blocks]; in home.nix it errors "error: undefined variable 'i3blocks' at ..." but why?
<cransom>
guessing that pkgs isn't in scope, you'd need to `pkgs.i3blocks`
<lordcirth>
Yeah
<lordcirth>
Or home.packages = with pkgs; [ i3blocks ]; which is shorter once you have more packages
<neonfuz2>
so yeah "unstrip" is missing from the gmic plugin in gimp it seems
<neonfuz2>
I see other distros have a patch called "gmic-unstrip.diff", not sure if that is needed to enable it or something
lunik1 has quit [Quit: :x]
<neonfuz2>
lmao ok nevermind, the gmic-unstrip diff in suse means "don't strip the gmic binary"
<{^_^}>
[nixpkgs] @AndersonTorres pushed 9 commits to master: https://git.io/JkUNv
<pta2002>
Now I have feedbin running on my own server :D
<pta2002>
Saving myself roughly... 40 cents a month
<pta2002>
ah well
sangoma has quit [Ping timeout: 272 seconds]
<iqubic>
srh
pn has joined #nixos
pn has quit [Client Quit]
<iqubic>
srhb: I figured out the issue with my emacs. Turns out for some reason the emacs varible of "exec-path" was set incorrectly, so it just wasn't searching "/run/current-system/sw/bin"
pta2002 has quit [Quit: leaving]
<Ankhers>
How can I get glibc 2.32 in a nix shell? I seem to have 2.31 on my system.
alp has quit [Ping timeout: 246 seconds]
<txt-file>
How to get obs-studio with plugins? I installed obs-studio and obs-v4l2sink but the plugin is not shown in obs-studio.
<pushqrdx>
i am facing a strange behavior, if i copy a derivation from the nixpkgs store (only), as discussed earlier i can't nix-shell that because it needs to know where to get it's "input" from, so the normal fix would be to add with import <nixpkgs> {}; at the beginning
<pushqrdx>
however that doesn't work but what works though is creating another file with `with import <nixpkgs> {}; callPackage ./app.nix {}` that calls the derivation
<pushqrdx>
now i can't see why would that work and just adding `with import` directly worn't
ManiacOfMadness has joined #nixos
<pushqrdx>
s/with import/with import ...
pbb has quit [Quit: No Ping reply in 180 seconds.]