<tbenst>
jtojnar, good to know. However I get the same problem with `nix-build -I nixpkgs=. -A python3Packages.mlflow && result/bin/mlflow server`
jluttine has joined #nixos
<jtojnar>
tbenst if you want to use that, you either need to wrap the script and set PATH in the wrapper, or use the python through absolute pah
<jtojnar>
maybe through something like sys.executable
<tbenst>
jtojnar, the script has the correct PATH. it's just this subprocess call to gunicorn. If I call gunicorn from nix-shell all is good, but as soon as I call it through subprocess it no longer has correct PATH
<tbenst>
interesting I'll take a look at sys.executable
<jtojnar>
tbenst well, then it does not have the correct PATH
<jtojnar>
tbenst for that you would need the subprocess to preserve PYTHONPATH env var (or possibly extend it with the value of `site.getsitepackages`)
mexisme has joined #nixos
<palo>
hmm trying to make this SideQuest package work, but I can't connect to the device, although the quest asked me for permission :/
<blackriversoftwa>
Is there any reference somewhere for packaging software with CMake builds that have builtin fetching of external dependencies (via cmake's `ExternalProject_Add`) function?
<blackriversoftwa>
I haven't found any good way to have cmake just pick up the dependency from a nix build input instead of trying to run the fetcher (and failing) from within the build sandbox
<jtojnar>
blackriversoftwa if the project cannot find the dependency, the CMake files usually need to be patched
<jtojnar>
unfortunately, it seems that no one really knows how to write portable CMakeLists.txt
<blackriversoftwa>
jtojnar: that's what I was afraid of. There are a lot of dependencies, also built using CMake, and it doesn't just import includes and libraries but CMake settings from them
<blackriversoftwa>
I am quickly forming that impression of CMake
<tbenst>
jtojnar, not sure I understand the question, or at least not sure how to answer it. Right now I'm trying to just call gunicorn directly by overriding the mlflow script, but finding it mighty challenging
<blackriversoftwa>
jtojnar: Interesting...wonder if that would work in my case or cause linking issues
<blackriversoftwa>
going to give it a try
<jtojnar>
tbenst it would be nice to see if the env var/sitepackages are getting through the subprocess
felixfoertsch23 has quit [Ping timeout: 245 seconds]
<zeta_0>
clever: i haven't spent that much time learning the language for linux shell scripts, am i going to be using shell scripts a lot, like in the nix pills example?
<clever>
zeta_0: and here is an example of it in use, you execute runHaskell, pass it a name, a set of haskell packages to get things from, a list of haskell deps, an attribute set, and a string containing haskell code
cyphase has quit [Ping timeout: 240 seconds]
<clever>
zeta_0: this will grab a ghc that has cardano-sl-util, cardano-sl-infra, and split in scope, then compile lines 25-55, then it will execute the compiled haskell code, with the 2 env vars from lines 8&15
<zeta_0>
clever: cool, that stuff is a little over my head
<clever>
zeta_0: 35/36 will then do something with those variables, and in this case, the result is a binary "it worked" or "it failed", so it just always writes "done" to $out
<clever>
exceptions lead to it not creating $out
sudoforge has joined #nixos
* clever
heads off to bed
felixfoertsch23 has joined #nixos
felixfoertsch has quit [Ping timeout: 250 seconds]
endformationage has quit [Ping timeout: 265 seconds]
supermathieu has joined #nixos
bhipple has quit [Remote host closed the connection]
zeta_0 has joined #nixos
<zeta_0>
i am currently using nixos-unstable and would like to use the nix package ihaskell but it is only supported in nixos-stable, is there a simple and easy way for me to install ihaskell from the stable channel while still being in the unstable channel?
<zeta_0>
i am not sure how to go about this?
doyougnu has quit [Quit: WeeChat 2.6]
<zeta_0>
nevermind, i will come and ask this question tomorrow
zeta_0 has quit [Quit: rcirc on GNU Emacs 26.3]
dckc has quit [Ping timeout: 250 seconds]
<yorick>
???
felixfoertsch23 has quit [Quit: ZNC 1.7.3 - https://znc.in]
<raboof>
what would be the typical place for a nixos package to store some files 'for the convenience of the user', such as example configuration files? (that would go for example to /usr/share/doc on FHS-style systems)
<raboof>
I see there's for example /nix/store/c7pa5pdffi65ibkxzzx6d2c8jbxvn0hc-wpa_supplicant-2.9/share/doc/wpa_supplicant/wpa_supplicant.conf.example - would the recommended way be to find such files there 'manually' or is there a more convenient way?
joehhhh has quit [Remote host closed the connection]
<lovesegfault>
And if it worked well over a couple days I was going to submit it
<colemickens>
cool! I don't think I ever sat down and wrote a config but I'd like to setup a 3/4-finger swipe to change sway worksapces and see how it goes
<colemickens>
I had pretty mixed experiences with libinput-gestures but don't think I tried gebaar yet
<colemickens>
I'm afraid to look for fear of losing hours to finally setting up home-manager.
<lovesegfault>
colemickens: :D
<lovesegfault>
it's freetext
zupo has quit [Ping timeout: 250 seconds]
zupo has joined #nixos
<lovesegfault>
colemickens: Can I add nixpkgs-wayland as a channel instead of an overlay?
<lovesegfault>
Also, can I add you as a maintainer to my gebaar submission?
<colemickens>
No. I was going to look into being able to use it as an additional package set but never got around to it.
<colemickens>
Yes, you can add me as a maintainer.
<lovesegfault>
Alright, standby for a PR
zupo has quit [Client Quit]
<colemickens>
It would help me to understand why you would prefer the channel over an overlay?
<colemickens>
I'm not sure what it would mean to provide a channel. Effectively the channel would be nixos-unstable+overlay, which is just how you use the overlay anyway. Does that make sense?
<lovesegfault>
colemickens: TL;DR: I have a modular config. I want to use gebaar in i3. I don't want to put the wayland overlay in the i3 config module.
<colemickens>
Ah, I understand. I was just telling my partner that I suspect you take the nix and purity to another level. :)
<colemickens>
My nixcfg is not quite that pure - and yeah, my overlay leaks into my other non-sway configs that I build sometimes... Hm...
<lovesegfault>
colemickens: :D
cransom has joined #nixos
<colemickens>
You could just import it as a channel...
<colemickens>
Which is basically what I do when I build it.
mexisme has quit [Ping timeout: 250 seconds]
aveltras has joined #nixos
<colemickens>
build.nix from the repo should give you an idea, but you can just re-import your nixpkgs, with my overlay, as a new package set, and then pull from there.
<lovesegfault>
colemickens: I think I will leak for now and upstream gebaar
<{^_^}>
[nixpkgs] @tfc opened pull request #74898 → nixos/test: Use retry() in all looping functions that need timeouts → https://git.io/JeMHc
Ariakenom has joined #nixos
Jackneill has joined #nixos
smatting has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
tbenst has quit [Ping timeout: 268 seconds]
smatting has quit [Ping timeout: 268 seconds]
<{^_^}>
[nixpkgs] @jonringer merged pull request #74897 → python3Packages.nose-exclude: disable tests for darwin → https://git.io/JeMHT
<{^_^}>
[nixpkgs] @jonringer pushed to master « python3Packages.nose-exclude: disable tests for darwin »: https://git.io/JeMHQ
Makaveli7 has joined #nixos
tjg1 has joined #nixos
oscarvarto has joined #nixos
<oscarvarto>
How could I install libreadline-dev on Nixos?
<oscarvarto>
I am trying to compile a tutorial for Postgresql
<oscarvarto>
make is complaining about readline stuff
<etu>
,libs oscarvarto
<etu>
,library oscarvarto
<{^_^}>
oscarvarto: Don't install libraries through nix-env or systemPackages, use nix-shell instead. See https://nixos.wiki/wiki/FAQ/Libraries for details.
<oscarvarto>
so, nix-shell -p libreadline-dev postgres
<oscarvarto>
?
<etu>
yeah, something along those lines :)
<etu>
but probably readline instead of libreadline-dev
zupo has joined #nixos
<oscarvarto>
OK. I got it.
<oscarvarto>
Thanks a lot!
johnnyfive has quit [Ping timeout: 252 seconds]
<ronny>
again - how can i build a vm via nixos-rebuild with just nix as a starting pint on a non-nixos system
johnnyfive has joined #nixos
thc202 has joined #nixos
oscarvarto has quit [Quit: oscarvarto]
smatting has joined #nixos
CMCDragonkai has quit [Quit: Connection closed for inactivity]
<raboof>
what software would you suggest for burning an installation ISO to CD/DVD?
sudoforge has quit [Ping timeout: 240 seconds]
<raboof>
ronny: I'm not familiar with nix on non-nixos systems - do you have nix-build there?
<ajs124>
If someone could backport fda32e99e0d20e2d68ed47b15be31412982cac1c, so my kde systems build again, that would be really cool. I can open a PR if that helps.
<{^_^}>
[nixpkgs] @peti pushed to master « haskell-haddock-library: jailbreak version 1.8.0 so that we can build the test suite »: https://git.io/JeMbJ
rardiol has joined #nixos
<{^_^}>
[nixpkgs] @ajs124 opened pull request #74910 → spectacle: add new kwayland dependency → https://git.io/JeMbU
philr has joined #nixos
kckasumi has joined #nixos
<adisbladis>
ajs124: Thanks
erictapen has quit [Ping timeout: 276 seconds]
<emmanuelrosa>
I have a build of vifm which fixes `vifm-media` so that removal media can be mounted/unmounted using udisks2 as the back-end. However, adding the necessary dependencies (Python, dbus, and udisks2) increased the closure size from 41,164,432 to 382,642,536. In such a case, is it better to add a second package (ex. vifm-full) instead of changing the existing one?
<ronny>
raboof: nix-build is availiable, its just not a nix system yet - im taking a look at rebuilding my system and want to test my old setups before putting them back in a real box
<ronny>
that script is indeed looking like what i want, how do i install it in a nix env where global system is not availiable
turboMaCk has joined #nixos
__Sander__ has quit [Quit: Konversation terminated!]
<mananamenos>
hi, i've set services.postgresql.enable = true; services.postgresql.package = pkgs.postgresql_11; However, `psql -U postgres -l` fails, with psql: FATAL: Peer authentication failed for user "postgres". How can i use postgres with local nixos user, without having to create a role in postgres?
<clever>
mananamenos: i think you have to run that as the linux user postgres
<{^_^}>
[nixpkgs] @teto pushed commit from @equirosa to master « lazygit 0.8 -> 0.11.3 »: https://git.io/JeMxt
coderobe8 is now known as coderobe
mexisme has joined #nixos
<{^_^}>
[nixpkgs] @Lassulus merged pull request #73441 → prusa-slicer: use wxGTK31 instead of wxGTK30 → https://git.io/Jerb0
<{^_^}>
[nixpkgs] @Lassulus pushed commit from @Moredread to master « prusa-slicer: use wxGTK31 instead of wxGTK30 »: https://git.io/JeMxl
<ddima>
does anyone know, is there some way to retrieve a commit as a patch/diff from sourceforge for direct use with fetchpatch like is possible with github, gitlab and others?
mexisme has quit [Ping timeout: 246 seconds]
<{^_^}>
[nixpkgs] @edolstra merged pull request #74850 → nixos/version: PRETTY_NAME in /etc/os-release uses the release now → https://git.io/JeMrE
selfsymmetric-mu has quit [Remote host closed the connection]
bahamas has quit [Quit: leaving]
<invokesus>
Hey! Any suggestions on how to get ofborg building again? I see from https://hydra.nixos.org/eval/1539490#tabs-inputs that the switch to openssl_1_1_0 broke it. So, I pinned nixpkgs to e712667ffd386, but I get `error: use of deprecated item 'try': use the `?` operator instead`. Do I need to switch to rust 1.38.0? Thanks!
iqubic` has joined #nixos
endformationage has joined #nixos
xensky has quit [Ping timeout: 252 seconds]
mexisme has quit [Ping timeout: 252 seconds]
philr has quit [Ping timeout: 252 seconds]
xensky has joined #nixos
iqubic has quit [Ping timeout: 252 seconds]
mexisme has joined #nixos
<sondr3>
invokesus: the try! operator has been deprecated in 1.39, so any release before that should work
cosimone has quit [Quit: Terminated!]
<sondr3>
it should just be a deprecation warning, but I guess ofborg has all warnings as errors?
werner291 has joined #nixos
mexisme has quit [Ping timeout: 250 seconds]
mexisme has joined #nixos
mexisme has quit [Ping timeout: 246 seconds]
ng0_ has joined #nixos
cosimone has joined #nixos
dansho has quit [Ping timeout: 276 seconds]
ng0 has quit [Ping timeout: 260 seconds]
mexisme has joined #nixos
<invokesus>
How do I get rust 1.38.0? Not familiar with nixpkgs-mozilla. Tried this
<{^_^}>
[nixpkgs] @adisbladis pushed 2 commits to release-19.09: https://git.io/JeMp7
ng0_ is now known as ng0
cosimone has quit [Client Quit]
cosimone has joined #nixos
mexisme has quit [Ping timeout: 252 seconds]
mexisme has joined #nixos
<evanjs>
Hrm. Going to see how easy evcxr-jupyter is to nixify, but wondering where the best place for the kernel is in the nixpkgs tree. Any thoughts? We don't have a jupyter kernel or rustPackages area, etc
<sondr3>
invokesus: I think you can specify the version in the channel too, so {channel = `1.38.0`} should work
<{^_^}>
#74858 (by x123, 16 hours ago, open): nixos/tests/timezone: port to python
<{^_^}>
#74866 (by x123, 15 hours ago, open): nixos/tests/tiddlywiki: port to python
chloekek has joined #nixos
mananamenos has quit [Remote host closed the connection]
mexisme has quit [Ping timeout: 250 seconds]
cosimone has joined #nixos
mexisme has joined #nixos
xensky has quit [Remote host closed the connection]
sweep has joined #nixos
sweep is now known as genesis
genesis has joined #nixos
genesis has quit [Changing host]
kreisys has joined #nixos
mexisme has quit [Ping timeout: 276 seconds]
mexisme has joined #nixos
<evanjs>
Wait whaaat. evcxr_jupyter comes with evcxr, I just need to run `evcxr_install`? Hrm. Wonder how I could incorporate this into my jupyter module config...
rardiol has quit [Ping timeout: 250 seconds]
dansho has quit [Quit: Leaving]
rardiol has joined #nixos
<evanjs>
Or just switch to jupyterWith as a module, somehow... Which is what I've been meaning to do
Shoubit has quit [Remote host closed the connection]
<mananamenos>
hi, not nixos related related, but cannot figure out how to formulate my question in google. If i take a hdd with windows or linux in it, and connect it to a computer via usb, select this hdd in boot menu when turning the computer, would it actually load my os from hdd as it were connected via sata3?
knupfer has joined #nixos
genesis has quit [Ping timeout: 268 seconds]
Shoubit has joined #nixos
mexisme has joined #nixos
lopsided98 has quit [Quit: Disconnected]
lopsided98 has joined #nixos
knupfer has quit [Ping timeout: 268 seconds]
vaibhavsagar has joined #nixos
<vaibhavsagar>
mananamenos: yes, I believe so
mexisme has quit [Ping timeout: 250 seconds]
<vaibhavsagar>
I wouldn't expect it to be any different from a second internal drive (except for transfer speed)
dbmikus_ has joined #nixos
<vaibhavsagar>
sometimes your OS has hardware-specific drivers that won't play well with a different computer, and Windows in particular has a license key that's tied to the specific computer it's installed on
<mananamenos>
vaibhavsagar, thanks, i've just tried it. Once I select the external, usb connected hdd, i can choose windows/nixos from boot loader, but then loading the os it crashes, something is wrong.
avn has quit [Remote host closed the connection]
<mananamenos>
which is weird, cause after selecting hhd, my boot partition was recognized the same way as when hdd was connected to sata3, however, neither nixos nor windows finishes loading successfuly
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #74922 → ghc: don’t use gold when useLLVM = true → https://git.io/JeDeC
mexisme has joined #nixos
sweep has joined #nixos
<vaibhavsagar>
mananamenos: is this the same computer the HDD came from internally?
gxt has quit [Ping timeout: 260 seconds]
sweep is now known as genesis
genesis has joined #nixos
genesis has quit [Changing host]
emptyflask has joined #nixos
<mananamenos>
vaibhavsagar, yes
<vaibhavsagar>
hmm, that's very weird
<vaibhavsagar>
some bootloader thing maybe?
drakonis has joined #nixos
mexisme has quit [Ping timeout: 246 seconds]
<mananamenos>
yeah, maybe, gonna try check that. Thank you again
mananamenos has left #nixos ["Killed buffer"]
<emptyflask>
I was looking for some way to install https://github.com/RPTools/maptool on NixOS, but it says it requires Oracle Java SE 10, which isn't in nixpkgs. is there any reason it's not available?
<tilpner>
Licensing perhaps?
<srhb>
I think it's no worse than the existing oracle jdk. They have admittedly draconic licenses already and are a bother to maintain
<srhb>
So at a guess: It's cumbersome, no one bothered.
<srhb>
I'd probably try coercing the maptool to use a less evil jre.
<drakonis>
ah, rpg player.
mexisme has joined #nixos
<drakonis>
a rpg player.
<drakonis>
from what i remember, you don't actually have to use oracle's jdk, there are other builds that provide the exact same featureset under a different license
ambro718 has joined #nixos
<drakonis>
emptyflask: it requires java, not specifically oracle's build.
andrewzah has left #nixos ["WeeChat 2.5"]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<drakonis>
on windows, oracle jre is the most common option
<chloekek>
You could just give it a try with JDK from Nixpkgs.
<drakonis>
do that
<chloekek>
I doubt those are actually hard requirements, but rather there to reduce bug reports involving Java version issues.
mexisme has quit [Ping timeout: 276 seconds]
Havvy has joined #nixos
<emptyflask>
yeah I attempted using jdk11 and while it started up, it immediately crashes with `Exception in thread "main" java.lang.NoClassDefFoundError: sun/reflect/Reflection`
<drakonis>
dont actually use the wrong jdk version
<drakonis>
it wants jdk 10
<drakonis>
run jdk 10
isHavvy has quit [Ping timeout: 245 seconds]
mexisme has joined #nixos
<chloekek>
XD maybe sun apis aren’t in the free jdk afterall
knupfer has joined #nixos
<drakonis>
they arent?
<drakonis>
they appaer to be
<drakonis>
i think its due to jdk11 doing some changes
<drakonis>
newer jdk releases broke older code due to some classes getting shuffled around and consequentially no longer being part of the default classpath
mexisme has quit [Ping timeout: 246 seconds]
<drakonis>
so you had to manually include them
mexisme has joined #nixos
cosimone has quit [Quit: Terminated!]
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
xkapastel has quit [Quit: Connection closed for inactivity]
LLED2_3 has quit [Ping timeout: 268 seconds]
<Gopal-M>
hello!
<Gopal-M>
how do I pin down the name of my network device?
<srhb>
Gopal-M: Do you just want to disable the reproducible naming entirely?
<Gopal-M>
no, rather the opposite
<Gopal-M>
my interface is named `eth0`
<Gopal-M>
I have no idea
nuncanada has joined #nixos
<srhb>
Gopal-M: Which version of NixOS are you on?
<Gopal-M>
because I haven't disabled `networking.usePredictableInterfaceNames`
<Gopal-M>
19.09
<Gopal-M>
I am not able to configure `eth0` through `networking.interfaces`
halfbit has quit [Ping timeout: 246 seconds]
<srhb>
What's the device?
<srhb>
Um. :)
<Gopal-M>
srhb: wdym by the device?
<Gopal-M>
I have an encrypted zpool which I log into at stage-1, dropbear ssh, and unlock. At that time, my interface is named eth0.
<srhb>
aaah.
mexisme has quit [Ping timeout: 252 seconds]
<Gopal-M>
But during boot, if I don't configure anything, it acquires the primary IP by DHCP with the interface name `eth0`.
<Gopal-M>
but my problem is I want to add the secondary IP too
<Gopal-M>
anything I do with `networking.interfaces` results in a failure, causing `network-setup.service` wait for 1 mins 30 seconds at boot with it eventually failing
<Gopal-M>
is it normal behaviour to have the interface named `eth0` in this case?
<srhb>
I don't believe it is, but it may be related to the early configuration of the interface.
<srhb>
I don't know for sure though. You could try ensuring that it's dropped before passing on from the initrd.
<Gopal-M>
how can I ensure that?
<srhb>
Probably the systemd docs on the renaming will be useful here.
<srhb>
Probably deleting all IPs and downing the interface. It would be an interesting test at least. But the docs might be faster than guesswork :)
dbmikus_ has joined #nixos
<srhb>
I don't _think_ we actively do anything in NixOS to prevent the renaming of such early devices, but that might be an avenue to check too
mexisme has joined #nixos
halfbit has joined #nixos
<Gopal-M>
I also get a lot of `kernel: refused connection IN=eth0` if that's relevant
<srhb>
That just sounds like the firewall.
ninegrid has joined #nixos
mexisme has quit [Ping timeout: 245 seconds]
<Gopal-M>
but I don't think I have anything blocked in my firewall. Docker handles iptables completely. I've added only 4 SNAT rules.
<srhb>
I tried setting the test for predictable interfaces to enable boot.initrd.network and that appears to cause the interface to not be renamed, just as you see.
<Gopal-M>
I say this because on another nixos system, I don't get any of those kernel refused connection messages
<Gopal-M>
srhb: thank you
<Gopal-M>
could you try `networking.interfaces` to then manually configure that network?
<Gopal-M>
and see if that works?
<srhb>
Flushing the ips from the device before exiting initrd does not, however, fix the issue.
<srhb>
This feels like a regression..
<Gopal-M>
this is all part of a bigger issue for me, though
<Gopal-M>
I was using `networking.localCommands` or `systemd.network` to work around htis
<Gopal-M>
but the bigger issue is that `docker` containers fail DNS resolution after a random period of creation
<Gopal-M>
both internal and external. it works for a certain period of time but after that, they're no longer able to resolve domains.
<srhb>
Sorry, I can't stack pop until I understand this. :P
<Gopal-M>
oh, np, let me read
<Gopal-M>
thank you
<srhb>
I'm happy that my gut feeling was exactly what the problem was, but... :P
<Gopal-M>
can you point me towards how I'd go applying this fix?
demfloro_ has joined #nixos
<srhb>
I'm not sure you should apply it just yet..
<Gopal-M>
I see
Jackneill has quit [Read error: Connection reset by peer]
<Gopal-M>
then I'll wait for 20.03 :)
<srhb>
I meant, I'm still reading.
<Gopal-M>
ah, sorry about that :P
<srhb>
boot.initrd.network.postCommands = "${pkgs.iproute}/bin/ip link set dev eth0 down"; -- fixes it in the test. So, maybe it would suffice for you to somehow ensure that the link is downed like that at the end of your initrd session
<srhb>
not sure how the ssh stuff works, haven't used it in years.
<srhb>
So don't use the above, it was just to verify.
<srhb>
(The interface stops being busy, causing stage2 to be able to rename correctly)
<Gopal-M>
I thought of that but won't it make ssh inaccessible since it's taking `eth0` before stage 1 is over?
<Gopal-M>
you know, what I can do?
<srhb>
Yes, so don't use the exact command. Instead, somehow, make it the final command of your initrd session, so to speak
<Gopal-M>
I can just run that command myself
<srhb>
Presumably you can run arbitrary commands in the initrd.
<Gopal-M>
like: `zfs load-key -a && ip link set dev eth0 down && killall zfs`
<Gopal-M>
do you think this will work?
<srhb>
Is the killall what causes boot to progress?
<Gopal-M>
yes
<srhb>
Then you'll want that in a script that is not interrupted by your disconnecting somehow
<srhb>
But yes, that sounds plausible.
<Gopal-M>
wait, there was something in the NixOS on ZFS wiki
<{^_^}>
[nixpkgs] @waiting-for-dev opened pull request #74925 → rails: init at 6.0.1 → https://git.io/JeDfX
<srhb>
Not sure what happens to the profile script once you disconnect.
<Gopal-M>
echo "zfs load-key -a && ip link set dev eth0 down && killall zfs" >> /root/.profile
<Gopal-M>
damn
<srhb>
Might need some gracious nohupping
<srhb>
You're on the right track though
<Gopal-M>
yeah
<srhb>
So you'll probably get it to work with a few tests :)
<Gopal-M>
thank you :)
<srhb>
Welcome.
<Gopal-M>
if only i could run tmux or screen in the initrd :P
<srhb>
You can, of course, but that seems a bit heavyweight.
<Gopal-M>
yeah
<srhb>
if dropbear behaves like openssh it'll send SIGHUP once it realizes you're disconnected
<Gopal-M>
it worked
<srhb>
woo :)
<Gopal-M>
echo "zfs load-key -a && ip link set dev eth0 down && killall zfs" >> /root/.profile
<Gopal-M>
helll yeahhh!
<srhb>
I think you should still nohup guard it
<Gopal-M>
now it's eno1
<Gopal-M>
thank you so much, srhb
<srhb>
In case dropbear is sometimes faster at realizing you're offline
justanotheruser has quit [Ping timeout: 250 seconds]
<srhb>
No problem :)
<Gopal-M>
I'll search what nohup is :P
ddellacosta has quit [Read error: Connection reset by peer]
<srhb>
Basically `nohup foo` will intercept SIGHUP so foo never sees it
<acarrico>
I noticed several bug reports on man and info pages, but I don't understand the status of these. Even with 'nix-shell -p gcc gcc.info gcc.man', I don't get any docs. Does anyone know how this got resolved?
mexisme has joined #nixos
ddellacosta has joined #nixos
smatting has quit [Ping timeout: 268 seconds]
lsix has quit [Quit: Leaving]
<infinisil>
acarrico: What docs are you looking for?
<infinisil>
If you're looking for the developer man pages, they're in pkgs.manpages
<acarrico>
infinisil: for example: man gcc --> No manual entry for gcc.
<infinisil>
Hm yeah doesn't work for me either
<acarrico>
infinisil: nix-shell -p manpages --> No manual entry for gcc.
<acarrico>
There are a bunch of bugreports about multiple output derivations.
<acarrico>
It seems like there was some change which makes some outputs optional, including info/man/debug/etc., but I really don't understand how it works.
<o1lo01ol1o>
Can an arbitrary environment variable be read in nix?
iqubic` has quit [Quit: ERC (IRC client for Emacs 27.0.50)]
<ottidmes>
Is there a way to boot into NixOS forcing CPU rendering or some such? Because I changed GPUs and NixOS is still configured for the old GPU. Unfortunately it then crashes the display while booting into NixOS, so I cannot get access to the console either.
domogled has joined #nixos
<ottidmes>
plan B would be to use a NixOS liveCD and try nixos-enter, but then I would have to setup my file systems, which is a bit of a pain
<ottidmes>
plan C is to plugin my old GPU and fix my config before trying my new GPU again
<gchristensen>
I think you could blacklist the modules or something via kernel params?
mounty has joined #nixos
Soo_Slow has joined #nixos
<gchristensen>
,locate grunt-cli
<gchristensen>
,locate grunt
<{^_^}>
Found in packages: brackets
<{^_^}>
Found in packages: pig, brackets, vscodium, oh-my-zsh, python37Packages.jupyterhub, vimPlugins.vim-addon-errorformats
<tilpner>
ottidmes: Try with LIBGL_ALWAYS_SOFTWARE=1
<tilpner>
Although you seem to want something system-wide and for more than opengl? :/
lunik10 has joined #nixos
<Yaniel>
ottidmes: you can change the config and rebuild from the livecd I think
<Yaniel>
no need to touch filesystems
<Yaniel>
IIRC I had to do that once for some reason
selfsymmetric-mu has joined #nixos
<tilpner>
ottidmes: When exactly does it crash? Perhaps you can disable the unit by editing the kernel command line
<ottidmes>
gchristensen: I just cloned my config on this machine, checked my hardware config, it is configured to use modesetting, which I believe is using some basic standard that most GPUs will suppport, so strange that it crashes, but good suggestion otherwise
<tilpner>
ottidmes: You can always blacklist the kernel modules your GPU uses from the kernel command line
<tilpner>
ottidmes: Press e while booting with grub or systemd-boot, add modprobe.blacklist=<name>
<ottidmes>
tilpner: thanks, guess I will just try modprobe.blacklist=modesetting and see what happens
<tilpner>
Hmm, I don't think that will do anything
<tilpner>
That would be nomodeset
<ottidmes>
tilpner: awesome, that did the trick, now X11 just fails (well, obviously), but at least I have console access now!
nexgen has quit [Quit: Leaving]
<x123_>
So, I am trying to port some more tests from perl -> python, but I am running into a test that doesn't even work at all in the current branch, `nixos/tests/munin.nix`, but that seems to be becaues munin-node doesn't even load or have a single plugin enabled out of the box. Is that the test's responsibility to set that up or the modules responsbility to at least enable a basic monitoring plugin like load or
<x123_>
something else? It seems like the module wants to copy the plugins over
fendor has quit [Ping timeout: 268 seconds]
<erictapen>
I'm trying to build my system with flakes. Does anybody know wether I can reference another flake at any place in the code like I was able to do with '<nixpkgs>' ? Would be really handy here.
zeta_0 has joined #nixos
<x123_>
possible solutions that I see are just port the tests to python, and just check that the service starts but don't worry about checking munin-cron since it would need to have a munin-node with at least one plugin enabled
<zeta_0>
clever: uhh, i am currently running nixos-unstable and would like to use the nix package ihaskell, but ihaskell is only supported in nixos-stable, is there a workaround that i could do to install ihaskell? i have no idea what to do?
dbmikus_ has joined #nixos
<srhb>
x123_: Or drop it entirely. I'm not sure it's used. Might need a big of git log digging
<srhb>
x123_: Hmm, looks like it's mostly the test that has been abandoned.
<srhb>
x123_: In that case, assuming people use the module, it should probably be the responsibility of the test to set it up.
<srhb>
x123_: However, if no one is looking at the test anyway (since it's broken...) maybe it's a bit wasted effort
<srhb>
x123_: I'm basing this mostly on there having been module updates in january 2019.
ninegrid has quit [Ping timeout: 250 seconds]
<x123_>
I Mean, I can't even get a super basic plugin to load properly
cosimone has quit [Read error: Connection reset by peer]
cosimone_ has joined #nixos
<zeta_0>
clever: well, i am using nixos-unstable, and i thought it uses nixpkgs-unstable by default?
<srhb>
x123_: I'm tempted to say that your efforts are probably better spent elsewhere :P
<srhb>
We should probably delete the test entirely if it's useless and no one cares, for now.
<srhb>
It's not like git history is destroyed.
<clever>
zeta_0: just swap the names in the directions
<clever>
zeta_0: it still explains how to run software from a channel your not on
<x123_>
The thing is, I cna make the test at least make sure the service starts
<x123_>
I just can't do all these other silly things like checking that munin-cron can connect to a munin-node and produce expected output
<srhb>
x123_: OK, if you think that adds value that seems sensible. :)
<x123_>
should I juts do that and leave some comments that this should be extended further or... yea i dunno
<srhb>
I think it has approximately zero users, and quite certainly none that check whether the test succeeds.
<x123_>
I am just trying to help out with some of this test porting, if there's elsewhere that would be more helpful I'm happy to do it, I've been using NixOS for all of 4 days
<srhb>
So scale your effort accordingly.
<srhb>
I know :) And it's appreciated. I'm just trying to say there's some cruft around that's hard to gauge.
<srhb>
x123_: And I don't know whether a working test that no one looks at is much better than no test at all. :P
<x123_>
yeah, fair enough, I am just knocking out ones I see listed in the test port issue, I'll have in mind that if it looks crufty I should try to simplify/remove useless parts/or whatever I guess
fusion809 has quit [Remote host closed the connection]
<zeta_0>
clever: so if i understand correctly, i just need to swap `unstable` with `stable` in the instructions link you sent me?
<clever>
zeta_0: pretty much
<zeta_0>
clever: ok, understood
cosimone_ has quit [Remote host closed the connection]
cosimone_ has joined #nixos
<dahirsch>
im using the mozilla-nixpkgs overlay and trying to crosscompile a rust project to muslpi. with >cargo build --target=<right target< it cant find the std crate. any ideas?
<thoughtpolice>
^^ i love pushing broken commits because i forgot to amend something :( sigh, fix incoming, though if mimalloc has been busted for a month i doubt anyone was using it but me. my fualt
<thoughtpolice>
*fault
rardiol has quit [Ping timeout: 268 seconds]
ottidmes has quit [Ping timeout: 240 seconds]
fendor has quit [Ping timeout: 250 seconds]
cosimone_ has quit [Remote host closed the connection]
cosimone_ has joined #nixos
zupo has joined #nixos
<{^_^}>
[nixpkgs] @acowley opened pull request #74936 → emacs: do not use seq package from elpa for emacs >= 27 → https://git.io/JeDkZ
rardiol has joined #nixos
zeta_0 has quit [Quit: rcirc on GNU Emacs 26.3]
fendor has joined #nixos
cosimone_ has quit [Excess Flood]
cosimone_ has joined #nixos
pie_ has quit [Ping timeout: 265 seconds]
fendor__ has joined #nixos
ottidmes has joined #nixos
fendor has quit [Ping timeout: 240 seconds]
ottidmes has quit [Read error: Connection reset by peer]
knupfer has joined #nixos
pie_ has joined #nixos
cosimone_ has quit [Remote host closed the connection]
<zeta_0>
so which channel is right a the bleeding edge?
vesper11 has joined #nixos
<lassulus>
I guess nixos-unstable
<lassulus>
but stuff breaks sometimes
<zeta_0>
Yaniel: `yolo channel`?
chiefgoat has quit [Client Quit]
<{^_^}>
[nixpkgs] @x123 opened pull request #74938 → nixos/tests/i3wm: port to python → https://git.io/JeDIc
smatting has quit [Ping timeout: 240 seconds]
chiefgoat has joined #nixos
ottidmes has joined #nixos
<Yaniel>
zeta_0: "expect frequent breakage for most things"
<Yaniel>
nixos-unstable has a lot more tests so if something breaks it is more likely to get fixed before the channel updates
sweep is now known as genesis
genesis has joined #nixos
genesis has quit [Changing host]
dahirsch has quit [Quit: Leaving]
<zeta_0>
Yaniel: i plan on using nixos-unstable for work, so i guess i am better using nixos-unstable instead of nixpkgs-unstable
<Yaniel>
yes
<zeta_0>
Yaniel: thanks for the clarification
<gchristensen>
if you're using NixOS, you should use a nixos-* channel
<zeta_0>
Yaniel: typo: `off` after `better`
<zeta_0>
gchristensen: `nixos-* channel`?
<ottidmes>
Ah, that was confusing, so "vesa" includes amdgpu, and "amdgpu" is something different, like some older version? Since when I try videoDrivers = [ "amdgpu" ]; it fails, videoDrivers = [ "vesa" ]; it succeeds, while lspci gives me kernel modules: amdgpu when it succeeds
<ottidmes>
zeta_0: any of the channels with the nixos- prefix, so nixos-unstable nixos-19.09 etc
<clever>
zeta_0: you dont have a channel with a name of "nixos" anymore, so its failing to find "nixos"
<clever>
zeta_0: one of them must be named "nixos" in the left column, and nixos-rebuild will use that
<tilpner>
It can also be named nixpkgs
<clever>
tilpner: the default $NIX_PATH assumes its named nixos, and will complain if it isnt
<zeta_0>
i removed it, i did not know that i needed it, how do i fix this
<Thra11>
How do I search for a file (to work out what package it's in)?
<clever>
zeta_0: use --add again, the man page says how to specify a name
<clever>
Thra11: nix-locate and the ,locate command on {^_^}
<noneucat>
i usually just did "find /nix/store -name" but it looks like i've found a better way haha :)
<zeta_0>
clever: `nix-channel --add nixos` ?
<clever>
zeta_0: you must give both a name and url, the man page says which order they go in
<Thra11>
clever: How do I work out which package nix-locate is in? :p
<clever>
,locate bin/nix-locate
<clever>
,locate
<{^_^}>
Use ,locate <filename> to find packages containing such a file. Powered by nix-index (local installation recommended) https://github.com/bennofs/nix-index
<fuzen>
Trying to package a theme, I am running `nix-env -f $NIXPKGS -qaP 'sweet'` but it isn't showing the package. The repo is available at https://github.com/Fuzen-py/nixpkgs
<clever>
siraben: probably fine
<siraben>
Alright.
<siraben>
I'll raise it to 60 to be safe.
<tilpner>
fuzen: You didn't add it to all-packages.nix
<siraben>
How did you know that it was due to timeout?
<tilpner>
It failed during check, and the manpage for timeout documents the exit code 124
<siraben>
Ah. I see.
<{^_^}>
[nixpkgs] @burke opened pull request #74942 → itstool: fix double-shebang issue on macOS → https://git.io/JeDLW
<{^_^}>
[nixpkgs] @veprbl merged pull request #74911 → Update qtbase mac patches for 5.12.6 → https://git.io/JeMNG
<{^_^}>
[nixpkgs] @veprbl pushed commit from @lavoiesl to master « qt5.qtbase: fix darwin patches for 5.12.6 (#74911) »: https://git.io/JeDLl
drakonis has quit [Ping timeout: 240 seconds]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
justanotheruser has joined #nixos
zeta_0 has joined #nixos
justan0theruser has joined #nixos
justanotheruser has quit [Read error: Connection reset by peer]
drakonis has joined #nixos
mexisme has joined #nixos
<zeta_0>
clever: can i change the piece of code so that i can place it in home.nix or maybe in my user(wolf) packages in configuration.nix, or does it have to be placed in environment.systemPackages ?
<{^_^}>
[mobile-nixos] @samueldr opened pull request #60 → demo: Rely on lightdm now that SLiM is gone → https://git.io/JeDLy
waleee-cl has quit [Quit: Connection closed for inactivity]
<zeta_0>
clever: i am almost there, just have to fix a few more things
<noneucat>
zeta_0: no semicolon on the closing '}'
<noneucat>
you probably also want to rename pkgs to stable
<noneucat>
or stable to pkgs
<zeta_0>
noneucat: oh thanks for the tips, one of the packages is unstable and the other is unstable, so i think i am going to need to explicitly prefix them with pkgs and stable
<sondr3>
I think adding `lib` to `{ config, pkgs, ... }:` ought to fix that error
<zeta_0>
sondr3: like this: `{ config, pkgs, lib, ... }:` , i don't remember having to do this in home.nix?
doublex_ is now known as doublex
<sondr3>
zeta_0: sorry, not sure, for the most part those errors simply are because you are either not passing them down when requiring files or because you haven't explicitly imported them
<sondr3>
AFAIK, I'm not an expert at Nix
<zeta_0>
sondr3: also, i am not sure if i need to replace `pkgs` with `pkgsStable`, ihaskell is only supported in nixos-unstable
<zeta_0>
sondr3: its ok, thanks for trying to help me
<{^_^}>
[nixpkgs] @flokli merged pull request #74938 → nixos/tests/i3wm: port to python → https://git.io/JeDIc