<trevthedev>
Anyone know if this is just some issue with Nix or something wrong with my environment? I assume it's user error, but I can't seem to figure out how to get a functioning environment for GL
<trevthedev>
Any help would be appreciated
<colemickens>
exarkun: maybe not helpful, but anecdotally I've had pretty rough experiences with NixOS's built in acme stuff. It's mostly only succeeded in getting me rate limited
drakonis has quit [Quit: WeeChat 2.6]
<exarkun>
colemickens: I'm giving up and writing my own certbot-based thing I guess :(
<exarkun>
except not tonight, monday I guess.
<colemickens>
There's enough different tools out there to do acme stuff, I wouldn't think you'd need to author another, but then again I don't know youre requirements
<exarkun>
is there something that glues a working acme client into systemd on nixos? that's the only reason I tried to use security.acme.
<colemickens>
I can't speak to that, sorry, I wound up moving the cert stuff fully into k8s and letting a controller there manage it.
__marlene__ has quit [Ping timeout: 240 seconds]
<trevthedev>
oh and could it be poor drivers on my end or something? doesn't seem likely
<colemickens>
It doesn't work yet, I'm going to spend another 10-15 minutes on it and then give up, I don't know meson well enough. I'll post more info though depending on where I end up on that PR.
v0|d has joined #nixos
sevcsik is now known as sevcsik-
<trevthedev>
ha, was just me being dumb. works just fine.
<Church->
Hey is the unifi ap controller service broken for anyone else?
<Church->
I go to configure and run and it fails over when starting saying that "user unifi doesn't have perms inside /var/lib/unifi/..." So I make a local copy of the service, set it to run as root and disable the nixokg version. Now when I hit the webapi at port 8443 it's just displaying some broken glyphs
<Church_>
When I try and hit the port at 8443 this is what I get: https://ibb.co/zncXF5s
toppler has joined #nixos
<colemickens>
This seems bad. I can't connect to new networks either, or my tether whose password is cached. :/. I might have to walk home to try to update to nixos-unstable...
<Church_>
When I hit with curl it seems to says that it's binary data.
<colemickens>
Church I don't think you want to hit 8443. Try 8080 and see if it redirects you.
<colemickens>
I had similar problems with Unifi doing weird stuff when receiving https requests on it's http endpoint or vice versa.
artemist has quit [Ping timeout: 260 seconds]
orivej has quit [Ping timeout: 240 seconds]
<hpfr[m]>
Anyone have a link to a derivation that just installs a .deb? I want to try that
<Church_>
colemickens Yeah just trying to get port 8080 to ssh forward.
<Church_>
When I hit it via my box's lan ip it fails to rediect like so
<colemickens>
(oh, don't forget, I had to allowedTCPPorts = [ 8080 8443 ] in my unifi mixin.
<Church_>
And now it works...
<Church_>
I didn't change anything
<Church_>
Stupid gremlins
m0rphism has quit [Ping timeout: 265 seconds]
cartwright has quit [Remote host closed the connection]
fendor has quit [Read error: Connection reset by peer]
<mishac>
Hello everyone. I have this, what seems to be a tricky problem. I installed clpru compiler, but during installation its path is automatically added to the NIX_LDFLAGS and NIX_CFLAGS_COMPILE, so whenever i make a call to ggc, it searches clpru path. Does anyone know how can I remove reference to clpru from these two environmental variables?
<mishac>
Or alternatively, when create a nix expresion, how can I add bin to cpath so that package is visible to another nix expression?
mexisme has joined #nixos
waleee-cl has quit [Quit: Connection closed for inactivity]
iyzsong has joined #nixos
MmeQuignon has quit [Ping timeout: 276 seconds]
dreverri has joined #nixos
soju has joined #nixos
<bdju>
I'm on a fresh unstable install and I can't seem to get wifi with wpa_supplicant working. I tried declaring my home network in my system config, but also wpa_cli can't connect to wpa_supplicant so I can't try it manually either
<bdju>
ethernet is working fine but after unplugging it, the wifi doesn't seem to kick in
<bdju>
and wifi worked during the install so hardware and drivers should be good I would think
toppler has quit [Ping timeout: 268 seconds]
<Jonathan66>
@bdju I have never actually used wpa_supplicant (directly at least). I use network-manager for my WiFi. Do you need wpa_supplicant specifically?
fusion809 has joined #nixos
<Jonathan66>
@bdju I can help you setup network-manager if all you are looking for is internet. For wpa_supplicant someone else will need to advise.
aveltras has quit [Quit: Connection closed for inactivity]
<pie_>
is there any way to make nix verbosely eval with nix.conf
<pie_>
there seems to be a show-trace
<pie_>
we did poke at the code with clever a while back and didnt seem to find anything
toppler has joined #nixos
ng0_ has joined #nixos
<bdju>
Jonathan66: I thought it would be a bit lighter. I may end up just giving up and using networkmanager
* colemickens
also keeps having nix-builds stall out in weird unexpected places with idle disk/cpu
johnnyfive has quit [Ping timeout: 245 seconds]
* colemickens
can't wait until systemd-networkd plays nice with IWD. Leave wpa_supplicant behind us where it belongs
ng0 has quit [Ping timeout: 260 seconds]
<bdju>
okay... changed to networkmanager and it really just works with no hassle
<bdju>
is there a way to declare networkmanager info in my configuration.nix or do I just always use nmtui?
<bdju>
will it remember my connections later?
reallymemorable has quit [Quit: reallymemorable]
<colemickens>
Yes. Your connections made with nmtui/network-manager are stored in /etc/
<colemickens>
`sudo ls /etc/NetworkManager/system-connections/ | wc -l` = 41
<colemickens>
I currently use NM and configure it to use the IWD backend and have good luck.
<colemickens>
(But IWD is not at 1.0 in nixpkgs yet, and the nm integration is new, so YMMV)
mjsir911 has joined #nixos
satoshi has joined #nixos
<colemickens>
bdju: ^
<bdju>
alright, thanks
<Jonathan66>
Well I made progress, but now I could use help again. I am trying to run a Haskell project of my on my newly NixOS rig.
<colemickens>
I'd like to eventually migrate to systemd-networkd and iwd, but systemd needs some work, there's an open issue upstream somewhere.
<colemickens>
for now, NM is fine, I don't notice it too much and I try to keep my system light
<pie_>
colemickens: iwd?
<colemickens>
I don't remember the acronym but Intel devs have been working on a comprehensive IWD (and then some) replacement.
<colemickens>
It just hit 1.0
satoshi has quit [Client Quit]
<colemickens>
It is SO much faster for me (with NM, to be clear, which could be part of it) for scanning/connecting and roaming.
mishac has quit [Remote host closed the connection]
soju has quit [Read error: Connection reset by peer]
<pie_>
im guessing its some wifi thing
<colemickens>
(Of course, when I first said IWD there, I meant wpa_supplicant.)
<colemickens>
Yes.
<pie_>
aha
soju has joined #nixos
<pie_>
(obviously intewebs daemon)
o1lo01ol1o has quit [Remote host closed the connection]
<Jonathan66>
The most headway I made is by just using cabal-install (v2-commands). But it fails when looking for lib:druntime. I am linking D code and it has a runtime. Before why my rig was running arch it found it from the `dmd` package I believe, which I have installed now. Why can it not find it?
<pie_>
*interwebs
racoonslug has joined #nixos
soju has quit [Read error: Connection reset by peer]
soju has joined #nixos
<Jonathan66>
Is it because nix does not show libs to everything?
<Jonathan66>
Can I expose druntime in a shell or something?
soju has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @jonringer pushed commit from Averell Dalton to master « yacreader: 9.5.0 -> 9.6.2 »: https://git.io/Jeg4v
xelxebar has joined #nixos
<{^_^}>
[nixpkgs] @bendlas pushed commit from @obadz to release-19.03 « chromium: 78.0.3904.70 -> 78.0.3904.87 »: https://git.io/Jeg4f
silver has quit [Read error: Connection reset by peer]
GlennS has quit [Quit: GlennS]
GlennS has joined #nixos
UltraPhil has left #nixos ["Leaving"]
<colemickens>
worldofpeace: is there data on how many users on a stable release vs unstable release?
<colemickens>
On the one hand, I would be annoyed to be on ESR on a stable nixos release, but I'm on unstable for so many reasons otherwise that I feel like if I valued stability for the sake of stability, I would move to a release channel and could live with ESR?
racoonslug has quit [Ping timeout: 268 seconds]
dreverri has quit [Remote host closed the connection]
<pie_>
Jonathan66: on nix you can assume things cant see eachother by default unless something specifically enables them to
<pie_>
though that doesnt help too much in actually figuriing out if thats true
<pie_>
i guess what youd want is to figure out what the relevant environment variables or whatever is,and whether its being set
__marlene__ has quit [Ping timeout: 240 seconds]
<Jonathan66>
Turns out it was complaining about an internal library called druntime. But that library includes the lib files, which may be why it is failing.
<pie_>
im not actually sure offhand how to debug environment variables, i dont know if you can just check that with strace or similar?
<Jonathan66>
All it said is that it failed in the configure step.
<pie_>
Jonathan66: aha...
<Jonathan66>
Um, me neither.
GlennS has quit [Quit: GlennS]
<pie_>
is this stuff from nixpkgs?
<pie_>
ok youre using cabal to do stuff that calls d and you installed d somehow
<pie_>
hm
<Jonathan66>
Right it builds the d library locally.
<Jonathan66>
I have tried using stack--rather than cabal-install, but it is failing on "while building package Cabal" or while building "mono-traversable"
aminechikhaoui has joined #nixos
GlennS has joined #nixos
jedai has quit [Ping timeout: 240 seconds]
__marlene__ has joined #nixos
<Jonathan66>
pie_ interesting
<Jonathan66>
It is fetching it.
<Jonathan66>
Wouldn't it make more sense for the druntime to be a separate package?
felixfoertsch has quit [Ping timeout: 264 seconds]
o1lo01ol1o has joined #nixos
<Jonathan66>
Hum, I may have gotten past that pie_.
<Jonathan66>
I used cabal2nix to create a shell and ran cabal within it (after fixing the generated .nix file).
<Jonathan66>
Not it is failing the same way but with Haskell bindings (internal library) which is supposed to link the locally compiled d library.
o1lo01ol1o has quit [Ping timeout: 276 seconds]
<Jonathan66>
Oh. nevermind.
maralorn has quit [Quit: WeeChat 2.6]
maralorn has joined #nixos
<Jonathan66>
I confirmed it is in-fact crashing because of the druntime lib. I gave cabal the place to find the D library lib and it went back to failing of the druntim package.
<hpfr[m]>
Trying to nixos-rebuild on laptop with remote builds using -j0. I get building of drv: goal destroyed for about a dozen derivations, and then error: unable to start any build; either increase —max-jobs or enable remote builds
<hpfr[m]>
But in the output from nixos-rebuild, remote builds work fine on a bunch of things, or it appears they do
<hpfr[m]>
But it fails with “goal destroyed” for the Linux kernel and a dozen or so related derivations
mumuluxi has quit [Ping timeout: 252 seconds]
wucke13 has quit [Ping timeout: 252 seconds]
<hpfr[m]>
Oh, I do see “got 2 remote builders” and then “considering” both of them and finally “hook reply is ‘decline’”. How do I fix this?
<Jonathan66>
pie_: I succeeded in getting it to build. It is not pretty but it builds and runs. I just extracted to druntime lib from the old arch install.
<Jonathan66>
Now that I have *something* working I will be able to try to gradually improve it.
<wedens[m]>
bsima: no. if you want to specify an attribute from specific nix file, you have to use `nix build -f x.nix a.b`. if you don't specify a file explicitly, it'll look at NIX_PATH
o1lo01ol1o has joined #nixos
mexisme has joined #nixos
dreverri has quit [Ping timeout: 268 seconds]
domogled1 has joined #nixos
o1lo01ol1o has quit [Ping timeout: 246 seconds]
domogled has quit [Ping timeout: 276 seconds]
domogled1 is now known as domogled
mexisme has quit [Ping timeout: 246 seconds]
mexisme has joined #nixos
mumuluxi has joined #nixos
<bdju>
I have kernel 4.19.81 on nixos unstable. is this normal or am I just out of date?
<cransom>
bdju: you can switch to the linuxPackages_latest kernel for 5.3.8. believe 4.19 is the current extended support kernel that everything defaults to.
<bsima>
wedens[m]: oh thanks, i didn't know about NIX_PATH
work_ has joined #nixos
<bdju>
cransom: cool, that sounds like what I want. how do I switch? just add it to system packages list or something?
domogled1 has joined #nixos
<bdju>
oh wait do I use the boot.kernelPackages option?
jedai has joined #nixos
<bdju>
the example uses 3.10 so I thought I needed a number. okay I'll give this a shot
domogled has quit [Ping timeout: 240 seconds]
domogled1 is now known as domogled
toppler has joined #nixos
jedai has quit [Ping timeout: 268 seconds]
Scriptkiddi has quit [Remote host closed the connection]
das_j has quit [Remote host closed the connection]
mexisme has quit [Ping timeout: 245 seconds]
das_j has joined #nixos
Scriptkiddi has joined #nixos
endformationage has quit [Ping timeout: 268 seconds]
felixfoertsch23 has quit [Quit: ZNC 1.7.3 - https://znc.in]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<{^_^}>
[nixos-homepage] @dependabot[bot] pushed to dependabot/npm_and_yarn/packages-explorer/eslint-utils-1.4.3 « build(deps): bump eslint-utils from 1.3.1 to 1.4.3 in /packages-explorer »: https://git.io/JegR9
<{^_^}>
[nixos-homepage] @dependabot[bot] opened pull request #317 → build(deps): bump eslint-utils from 1.3.1 to 1.4.3 in /packages-explorer → https://git.io/JegRH
<ar>
there are also bluetooth modules with more codecs support for pulseaudio
zupo has quit [Ping timeout: 240 seconds]
o1lo01ol1o has quit [Ping timeout: 265 seconds]
toppler has joined #nixos
<equivrel>
How does system.stateVersion work? I thought I bump the number before I want to upgrade, but a comment in the 19.09 release notes makes it sounds like I want to keep it at 19.03 until after the upgrade? The combination of nix-channel and stateVersion is a bit confusing...
<wedens[m]>
it's not nixpkgs version. it's a state version which typically corresponds to nixpkgs version at the moment of nixos installation
<manveru>
i somewhat wish it was called `nixpkgsVersionAtTimeOfInstallation` or the like, nobody seems to read the comment :|
<wedens[m]>
it looks the same because state changes happen when nixpkgs version changes and it's just convenient for them to look the same
<bdju>
okay got dvorak in sway figured out now. it seems to care that you specify layout and variant, not just variant
<gchristensen>
pretty sure that isn't correct, bdju, since I only set layout and not variant
<LnL>
manveru: yeah, maybe something like that isn't a bad idea
<equivrel>
wedens: oh I see, indeed that was very confusing. So typically I would leave this to be whatever it was when I first installed the system? That does seem kind of weird and impure though to have my system depend on when I installed it...
<equivrel>
I guess I can see why it's there. anyway, thanks a lot for the clarification!
<foxmean>
Hello, I'm confusing how to install python with python packages as described in https://nixos.wiki/wiki/Python . I've successfully install emacs as described in https://nixos.org/nixos/manual/index.html#module-services-emacs . But as the wiki said that, new package must be defined before I can put python-with-my-packages into my environment.systemPackages. But I don't know where to put the code that define a new package in.
<foxmean>
I've try to put it in my configuration.nix but it didn't work. and I still wonder what the "..." in the source that presented in "https://nixos.wiki/wiki/Python" .
<foxmean>
Or I need to create a python.nix and import it like a emacs.nix?
<rkrishnan>
I am trying to build a haskell package with `stack' which is set to use nix packages (stack.yaml has nix enabled). When I do `stack build', it fails with: "building '/nix/store/nlf7qf8raz7f7g1am90xwrq6763988v7-source.drv'... /nix/store/v6za7w484cg7f1ppvziidvpdvmrqyzpx-stdenv-linux/setup: line 813: /build/env-vars: No such file or directory
<rkrishnan>
It is a Debian system with nix installed. nix version is 2.3.1
<{^_^}>
[nixpkgs] @fpletz pushed commit from @globin to gcc-9 « gcc: switch default to gcc9 »: https://git.io/JegEb
m0rphism has joined #nixos
__monty__ has joined #nixos
<contrun[m]>
how can i build a new nixos qemu image which is ready to use in another machine? Nixos-build-vms contains many scripts in the nix store. I want them packed together
<contrun[m]>
infinisil (@freenode_infinisil:matrix.org): unfortunately not. it is a x86 openwrt machine. There must be many quirks even if I can install nix.
johnnywho has joined #nixos
orivej has quit [Ping timeout: 240 seconds]
<infinisil>
contrun[m]: Unless you know something concrete that doesn't work I'd still give it a go. Having Nix available will make things a lot easier
<contrun[m]>
infinisil (@freenode_infinisil:matrix.org): yeah. i will try it.
<nschoe>
Hi everyone :-), where can I find information about glib "schemas", and how they integrate with nixOS? I'm having issues with glib-networking and TLS/SSL support for gstreamer. And I really hve no clue what is wrong. I'd like to read to understand more.
<nschoe>
On #gstreamer, I've been told that I need a "gnutls enabled glib-networking", but looking at glib-networking/default.nix I clearly see a gnutls dependency, and something abour wrapProgram in the postFixup phase. I don't understand what it does. Someone here mentioned to me wrapGAppsHook and schemas, but I can't figure out what this means.
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
chris__ has joined #nixos
<yorick>
reptyr test doesn't finish in nixos-unstable-small
dansho has quit [Remote host closed the connection]
<chris__>
When flushing my iptable rules, and doing a rebuild - it does not seem to reset it to the 'original' configuration. Is there a way I can do this?
<jonge[m]>
does anyone else have the problem that nixos 19.09 can't be upgraded because some osxupd package (dependency from facetimehd package) cannot be downloaded? this is part of the unfree redistributable firmware things
<symphorien>
I switched from all firware to redistributed firmware and the problem vanished
<Guest84>
I'm having trroubles starting a particular game with wine. Wine itself works fine and installed the said game. Are there any tricks as to where one should get started with debugging such issues? It seems that a related error is `0009:err:wgl:X11DRV_WineGL_InitOpenglInfo couldn't initialize OpenGL, expect problems`. But I have explicitly enabled
<Guest84>
opengl support with ` But I have explicitly enabled opengl support
<Guest84>
My environment.systemPackages has an entry: `( pkgs.unstable.wine.override { wineRelease = "unstable"; openglSupport = true;pulseaudioSupport = true;} )` (Where `pkgs.unstable` is the unstable channel. It would seem that this compiles wine with OpenGL support.
<Guest84>
(I did have to build wine locally)
<Guest84>
So the question is- how does one get started with this?
<wedens[m]>
do you have 32bit opengl libraries installed?
johnnywho has quit [Remote host closed the connection]
<nh2>
can anybody explain me the use of cgroups-v1 vs -v2 in systemd on NixOS? Is this a setting I have to enable to choose between the two, or does it happen automatically depending on what I write into my systemd units?
<{^_^}>
[nixpkgs] @andir merged pull request #68483 → nixos/shells-environment: Make ~/bin/ in $PATH optional → https://git.io/JemCQ
<{^_^}>
[nixos-hardware] @ehmry pushed commit from @filalex77 to master « Add Lenovo ThinkPad T420 profile »: https://git.io/Jegaa
<nh2>
symphorien: that brought me a bit further but I'm not sure it's right. `mount | grep cgroup` on NixOS shows that there are 2 types of mounts active, some of type `cgroup` and some of `cgroup2` (at `/sys/fs/cgroup/unified`), which suggests that you can have both at the same time
<{^_^}>
[nixpkgs] @flokli pushed 2 commits to master:
<edef>
cgroups v2 essentially require a single "manager" process to be in control
<edef>
i'm not sure how the two versions interact but Documentation/ in the kernel tree might provide enlightenment
bvdw has quit [Read error: Connection reset by peer]
<nh2>
I'm trying to read out the new `memory.max` v2 attribute. But it's not there in the file system. I suspect it might be that in contrast to v1, v2 spawns such attributes in the file system only if they are actually set explicitly for the cgroup
o1lo01ol1o has quit [Remote host closed the connection]
<nschoe>
Hi again, for information, I finally managed to solve my gstreamer TLS/SSL support issue. Once inside my nix-shell, I ran `export GIO_MODULE_DIR=/nix/store/6r6b6p3c7hyvgidlnlyqa1zrvh9ajm9d-glib-networking-2.62.1/lib/gio/modules/` and then it worked.
<nschoe>
I'm not sure how to "port" this in glib-networking/default.nix, but after installation, it appears we need to do something with the location of the GIO modules.
<clever>
nschoe: at build and shell time, you can use a setup hook to set that var, then something else needs to wrapProgram the final binaries
<clever>
nschoe: for setup hooks, look at things like pkgconfig
<nschoe>
clever, yep I'm writing my shellHook here for my shell.nix. But I suppose this is a hack and should be doen in the default.nix for glib-networking, right? Unfortunately, I'm not sure what "wrapProgram" actually does.
o1lo01ol1o has joined #nixos
<clever>
wrapProgram just creates a shell script to wrap a program, 100's of examples are in nixpkgs
<clever>
a setup hook behaves similar to a shellHook, but works one derivation later
<clever>
so if you put glib-networking into the buildInputs, then the glib-networking setup hook gets ran
<clever>
nschoe: thats only wrapping the programs under libexec, for glib-networking itself
<clever>
but nothing that uses glib-networking later on
<nschoe>
Oh I see, yes.
<nschoe>
clever, so you're telling me I should adda setupHook inside glib-networking/default.nix that does more or less this?
<clever>
yeah
<clever>
pkgconfig has an example of a setup hook
<wedens[m]>
is there IFD or something that makes derivation from a file where dependencies are inferred from referenced packages? for example, `@gawk@/bin/awk` in the file adds gawk dependecy.
<nschoe>
clever, I'll look into pkgconfig then, thanks!
<clever>
wedens[m]: not currently, but it could be written, callCabal2nix basically does the same thing, it just uses cabal2nix to generate an expr, then callPackage to load it, and whatever it referenced
CMCDragonkai has quit [Quit: Connection closed for inactivity]
o1lo01ol1o has quit [Remote host closed the connection]
<nschoe>
clever, quick question: I have added the setupHook, and in my setup-hook.sh, I tested a very simple myfunc() which echo something, and then added "addEnvHooks "$targetOffset" myfunc". When I test this with `nix-shell -I nixpkgs=/home/nschoe/nixpkgs -p glib-networking` it works, but the myfunc() gets called lots of time. Am I missing something?
<infinisil>
clever: Would be cool if you could just do `mkDerivation ({ ... } // pkgs)`, then substituteAll within it, and somehow all environment variables are evaluated lazily
<clever>
nschoe: addEnvHooks will run your thing for every buildInput, but you likely want to just export a var and not define any function
<clever>
infinisil: there is a pkgs.substituteAll that does almost that
<nschoe>
clever, hum okay, this is why it runs several times, I see. Thanks.
<clever>
nschoe: your setup hook likely just needs to add @out@ to GIO_MODULE_DIR
<clever>
nschoe: then something else will just use wrapProgram to persist the current value of GIO_MODULE_DIR
<clever>
nschoe: and there are helper functions in the stdenv to add to a : seperated list
<nschoe>
clever, okay I was going to ask how to get the out, because $out did not give me this. @out@ does. It looks obscure, where can I get more information about this?
<nschoe>
clever, when you say "something else will just use wrapProgram to persist the current value of GIO_MODULE_DIR". What does it mean? Do I have to do it? Or do you mena other programs will do it, when depending on glib-networking?
<magnetophon>
when I run "zbarimg --raw qr-x00.txt.png > temp" I get: "Name Error (Connection ":1.43" is not allowed to own the service "org.linuxtv.Zbar" due to security policies in the configuration file)". How can I troubleshoot that?
<clever>
magnetophon: sounds dbus related, any differences if ran as root?
<magnetophon>
clever: nope, same thing
<magnetophon>
A web search seems to suffest it's relevant that I'm not running a full desktop, but just i3.
<magnetophon>
s/suffest/suggest/
<jtojnar>
magnetophon how did you install it?
<infinisil>
magnetophon: You should be able to view dbus logs with `journalctl --user -u dbus -f` I think
<jtojnar>
the package has a dbus config, maybe you need to add it to `services.dbus.packages`
<magnetophon>
jtojnar: nix-env -iA nixos.zbar
<jtojnar>
magnetophon you can use `zbarimg --nodbus --raw ~/qr.png`
<magnetophon>
infinisil that command just gives me "-- Logs begin at Tue 2019-05-28 16:13:45 CEST. --"
<jtojnar>
if you do not want to install the dbus config
<infinisil>
I see, maybe logs would be at `journalctl -u dbus -f` then
<magnetophon>
jtojnar: yup, that seems to work, great!
<magnetophon>
thanks everybody!
nschoe has joined #nixos
aveltras has joined #nixos
<infinisil>
magnetophon: Nice, can you tell me whether you find any dbus logs at that command? For future troubleshooting
magnetophon has quit [Remote host closed the connection]
<jtojnar>
in this case, zbar tries to attach itself to system bus, but since its dbus config was not installed globally, it fails
nobbo has joined #nixos
<jtojnar>
I do not understand why they went with a system bus instead of session bus, and why they do that by default
<nh2>
symphorien: I think you were right the following way: Even though `/sys/fs/cgroup/unified` has `cgroup2` mounted, systemd's cgroup2 support works only if it's mounted at the parent dir `/sys/fs/cgroup`. So that needs `boot.kernelParams = [ "systemd.unified_cgroup_hierarchy=1" ];` (e.g. from https://github.com/NixOS/nixpkgs/issues/48749). Without that, things like `MemoryMax=` will have no effect.
<{^_^}>
#48749 (by uvNikita, 1 year ago, open): Containers with unified cgroup hierarchy are broken on 18.09
toppler has joined #nixos
<jtojnar>
anyone else not running gpg-agent and getting `error: gpg failed to sign the data` on latest unstable?
<magnetophon>
any gpg command that needs a PW gives me: "error receiving key from agent: No pinentry - skipped" It was working just fine earlier. I tried putting the full path to pinentry in ~/.gnupg/gpg-agent.conf but no luck. what's going on?
<clever>
magnetophon: and if you restart the agent with gpg-agent /bye ?
<magnetophon>
clever: that command gives me "gpg-agent[3913]: gpg-agent running and available"
<clever>
magnetophon: and then try signing again?
<magnetophon>
clever: nope, same result
<magnetophon>
also tried "gpg-agent -q"
<magnetophon>
to be clear: it stopped working after a reboot. no upgrade or anything. oddly running "pinentry" says: "OK Pleased to meet you" and waits for further input. Oddly, "which pinentry" says "pinentry not found"
<magnetophon>
fwiw: When it was working, I'd get a graphical window to enter my PW, not a terminal thing/
chloekek has quit [Ping timeout: 265 seconds]
<clever>
magnetophon: what about `type pinentry` ?
<magnetophon>
clever: pinentry not found
<clever>
no idea then
<magnetophon>
when I run "nix-env -iA nixos.pinentry" and then "which pinentry", I get "/home/bart/.nix-profile/bin/pinentry", but still no luck with pgp
<magnetophon>
"type pinentry" now also gives "pinentry is /home/bart/.nix-profile/bin/pinentry"
<magnetophon>
clever: hmmm, I manually killed gpg-agent, and now when I run "gpg-agent" I get "gpg-agent[4547]: no gpg-agent running in this session"... :(
<magnetophon>
ah, but now when I run a gpg command, it works! ¯\_(ツ)_/¯
<magnetophon>
clever: thanks for your suggestions!
<{^_^}>
[nixpkgs] @flokli opened pull request #72603 → nixos/ceph: run unprivileged, use StateDirectory and tmpfiles, don't pass extraServiceConfig → https://git.io/JegKZ
cosimone has joined #nixos
<Ericson2314>
clever: ok!
cosimone has quit [Client Quit]
endformationage has joined #nixos
Heirlung has joined #nixos
cosimone has joined #nixos
<clever>
[clever@amd-nixos:~/apps/nixpkgs-vc4]$ nix-build test.nix -A vc4-binutils
<clever>
error: Package ‘binutils-vc4’ in /home/clever/apps/nixpkgs-vc4/pkgs/development/misc/vc4/binutils.nix:4 is not supported on ‘vc4-none’, refusing to evaluate.
<clever>
Ericson2314: i think its trying to cross-compile the binutils to the target?
<Ericson2314>
clever: looks like it
<clever>
[clever@amd-nixos:~/apps/nixpkgs-vc4]$ nix-build -A vc4-gcc-stage1
<clever>
this seems to at least start to eval, and now its building something
toppler has quit [Ping timeout: 240 seconds]
jonringer has joined #nixos
<clever>
Ericson2314: the problems i can remember having, is getting c++ support in the gcc, and getting the gcc to be able to find crt0 and newlib
reallymemorable has quit [Quit: reallymemorable]
<hpfr[m]>
Trying to get a dpkg called OnlyKey to work with NixOS. After autoPatchElf, `ldd opt/OnlyKey/nw | grep 'not found'` returns `libnw.so`. How can I patch this in after autoPatchElf? Gitter in nixpkgs also uses nwjs but this is all a bit over my head right now :/. I tried `patchelf --set-rpath /nix/store/hash-nwjs-version/lib opt/OnlyKey/nw`, but trying the executable after gave `error while loading shared libraries: libnw.so:
<hpfr[m]>
cannot open shared object file: No such file or directory`
cosimone has quit [Quit: Quit.]
<symphorien>
is the libnw.so you found for the same arch (not 64 bits for 32 or something like that) ?
cosimone has joined #nixos
<Ericson2314>
clever: I'll scroll up to look at yoru code
<Ericson2314>
ugh matrix lousy scrollback doesn't help
<hpfr[m]>
symphorien: nix-locate gave nwjs-sdk.out, nwjs.out, gitter.out, and betaflight-configurator.out. I tried running makeLibraryPath on nwjs and used the resulting store path in the patchelf rpath. That’s all I know
<clever>
hpfr[m]: is the binary your patching 32bit or 64bit?
<hpfr>
clever: I tried patchelf --set-interpreter "$(cat $NIX_CC/nix-support/dynamic-linker)" --set-rpath /nix/store/hash-nwjs-version/share/nwjs/lib opt/OnlyKey/nw, but I'm getting the same error when trying to run it
<hpfr>
I'm doing this from a nix-shell with patchelf and autoPatchElf, not trying to make a derivation yet
<clever>
hpfr: what error?
<hpfr>
clever: `./opt/OnlyKey/nw: error while loading shared libraries: libnw.so: cannot open shared object file: No such file or directory`
<clever>
hpfr: what does patchelf --print-rpath say?
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « devpi-server: 5.1.0 -> 5.2.0 »: https://git.io/Jeg6G
<hpfr>
clever: progress! now I get a fontconfig warning and after a slight pause... a segfault. oof haha
<clever>
hpfr: gdb time!
<hpfr>
maybe I need to try a FHS method?
<hpfr>
clever: gdb run returned `0x00007ffff3c07da0 in ?? () from /nix/store/k6s14ppqyvcnlkzzxw7xzhg2lf8cpwl5-nwjs-0.33.4/share/nwjs/lib/libnw.so`, so that's not good.
<dhess>
gchristensen: your immutable image builder thingy, do any of your images need secrets? i.e., things that would normally be handled by `nixops deploy`?
<clever>
dhess: the ssh host keys already have an escape hatch for that, which iohk is using for secrets
<gchristensen>
oh no The option `security.acme.directory' can no longer be used since it's been removed. ACME Directory is now hardcoded to /var/lib/acme and its permisisons are managed by systemd. See https://github.com/NixOS/nixpkgs/issues/53852 for more info.
<gchristensen>
on Packet for the aarch64 machine I have a tiny partition that I loaded up with some secrets, and the immutable images mount it at startup
<dhess>
clever: hmm so if you were to do this in CI, then that persistentConfigDir would need to be on the CI server
<clever>
dhess: oh, are you talking about pure nixos images, or darwin images?
<dhess>
clever: cool thanks. I need to look more carefully at what IOHK is doing with Macs.
<dhess>
clever: NixOS images
<clever>
ah, i was thinkign about the darwin images, since they are also pure, lol
<clever>
(as pure as darwin can be)
<dhess>
It just occurred to me after I saw gchristensen's tweet a few days ago that most of my NixOS hosts could boot from an image, and rather than nixops deploy'ing them every few weeks, I could just reboot them from a new image
<dhess>
but ssh host keys would be an issue, at the very least.
<clever>
dhess: i think gchristensen solves host keys by just publishing the public on bootup
<dhess>
I have a few that participate in a WireGuard network, and those need to have secrets deployed.
<gchristensen>
yeah
<dhess>
clever: yeah I'm just thinking of SSH clients complaining all the time. It's not the kind of thing you want to get used to ignoring. :)
<gchristensen>
dhess: right so with Packet machines I publish the host key to their event API, and I look for a reboot event followed by a ssh host key publish "event". I use that for trusting the key, instead of being blind about errors
<gchristensen>
you could maybe do the same, but a bit harder to do the initial trust
<dhess>
gchristensen: you have some kind of service running that updates your .ssh/known_hosts?
<evanjs>
Bah. Confused about where I should add cargo-geiger (https://github.com/anderejd/cargo-geiger). What do you guys think? development/tools/rust or tools/package-management?
<evanjs>
Can't recall where/if the docs mention categories
<__monty__>
evanjs: Hmm, isn't there a rustPackages? cabal-install (comparable to cargo for haskell, kind of) is just in haskellPackages and aliased at the top level.
<inkbottle>
In a context where nodejs is installed I've tried `npm install --save-dev bs-platform`, but it failed; and sorry to say, but I have no clue why (I was just following https://ylixir.io/2019/02/08/reasonml-and-nix/)
<inkbottle>
I do not have OCaml installed or anything...
<inkbottle>
Just basic NixOS system
<__monty__>
I'd say it belongs in tols/rust though. Package-management sounds like end-user package managers not language-specific package managers.
<evanjs>
__monty__: nah not atm there isn't
mrCyborg has joined #nixos
toppler has quit [Ping timeout: 240 seconds]
<evanjs>
Right, the reason I asked is because I saw stuff like cargo-license in package-management, and cargo-license prints the licenses of your crate and all dependent crates. geiger prints out unsafe usages in your crate and dependencies
<evanjs>
Basically there's a bit of overlap between the two categories. But if it's primarily about checking for unsafe code, it seems okay to put it into tools/rust
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « python38Packages.elementpath: 1.3.0 -> 1.3.1 »: https://git.io/Jegyn
<hyperfekt>
samueldr: It got updates, but it's possible noone remembered to do the change. I guess I'll fix it and if there was a good reason I'll know because every single test will start failing :b