<Deiru>
I'll be sure to take note of this, the more you know.
pushqrdx has quit [Remote host closed the connection]
tboston has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Deiru>
Aaaand I've realized that whatever I was doing was utterly pointless. I've fiddled around with usermod on my server and removed myself from wheel, and all of this trouble of chrooting into nix to correct that when I had mutableUsers as false and I could have just... rebooted.
<Deiru>
😭
aasg has quit [Quit: killed]
<ToxicFrog>
VAAPI mystery solved: hardware.opengl.enable needs to be set to true for any of this to work at all
tboston has joined #nixos
<{^_^}>
[nixpkgs] @mweinelt opened pull request #103851 → tts: init at unstable-2020-06-17 → https://git.io/Jksw4
pushqrdx has joined #nixos
<Deiru>
Well, at least I have learned two important lessons today. Thanks again, clever.
<{^_^}>
[nixpkgs] @maralorn pushed 2 commits to haskell-updates: https://git.io/JksrH
<growpotkin>
Is there a standard way to use `gdb` on wrapped programs? I usually have to dig through wrapper scripts to find the real executable whenever I want to run `gdb` on installs from the store
<clever>
growpotkin: you can also run `gdb /proc/123/exe 123` to attach to a process thats already running
<clever>
growpotkin: the exe symlink always points to the real binary
<growpotkin>
Yeah in my case I need to launch from `gdb` to trace program's startup, so attaching isn't always the best fit.
<clever>
growpotkin: for basic tracing, you can still use strace instead
<growpotkin>
Symlinked binaries arent my issue, I'm trying to deal with wrapped programs.
<infinisil>
It would be nice if wrappers had a standard way for extracting the unwrapped version
<growpotkin>
Emacs is an example where `cat $( which emacs )` is actually a bash script that sets up a minimal environment before launch the real `emacs`
<infinisil>
Not only for gdb, but also for flattening doubly-wrapped binaries
<clever>
growpotkin: strace can work on those bash scripts
<growpotkin>
What I've ended up doing is rewriting the wrappers, but it would be nice if there was a standard way to do it
<growpotkin>
clever: That's a good point, I could use `strace`.
<clever>
growpotkin: i tend to just read the wrapper, export the vars in the current shell, and then run gdb on the original binary
<growpotkin>
clever: yeah that's exactly what I was doing so far.
<otavio>
I am having problem in getting xserver.xkbOptions applied. I am using 'startx' as display manager but my 'caps:super' does not apply
<otavio>
checking the localectl output it looks fine
<otavio>
and the xorg.conf.d file as well
johnw has quit [Client Quit]
andreas31 has quit [Remote host closed the connection]
jb55 has joined #nixos
andreas31 has joined #nixos
cryptomonad has joined #nixos
<cryptomonad>
I want to build a script that creates a nix-shell from default.nix (as if nix-shell was run), adds another dependency (through -p) and runs some bash commands, including buildPhase. I created a #!env nix-shell script and succeeded in adding another dependendy, but I'm stuck because I don't seem to have access to any of the phase functions like buildPhase. How do I access buildPhases from a script?
<{^_^}>
[nixpkgs] @maralorn pushed 3 commits to haskell-updates: https://git.io/JksDP
<clever>
cryptomonad: that will give you a shell suitable to build hello, but with one more pkg added to the inputs
h0m1 has quit [Ping timeout: 260 seconds]
h0m1 has joined #nixos
sss2 has joined #nixos
<growpotkin>
otavio: Are you restarting your machine between changes? In a perfect world you wouldn't need to, but I wouldn't be surprised if a reboot applied the changes.
<matthewcroughan_>
clever: :O You're alive
<{^_^}>
[nixpkgs] @enolan opened pull request #103857 → nixos/xidlehook: new service → https://git.io/Jksyl
<{^_^}>
[nixpkgs] @znewman01 opened pull request #103858 → kenlm: init at unstable-2020-11-14 → https://git.io/JksyR
<cryptomonad>
clever: Thanks. My problem is not getting additional dependencies into nix-shell, my problem is that I don't have access to "buildPhase" if I try to use nix-shell as a script interpreter. https://gist.github.com/wchresta/572b4e82b1799d019d3a49eafc2e64ae
<cryptomonad>
clever: Great, that worked. I moved my commands into a .nix file with shellHook instead of a script and can now use nix-shell myscript.nix. Thanks!
atlasznctestf has quit [Quit: ZNC 1.8.2 - https://znc.in]
atlasznctestf has joined #nixos
<matthewcroughan_>
raghavsood: thanks so much for commenting on that issue :)
<raghavsood>
I was dealing with that yesterday, took a while to figure out how to get it to run reliably
medvid has quit [Ping timeout: 265 seconds]
zeta_0 has joined #nixos
medvid has joined #nixos
<{^_^}>
[nixpkgs] @Mic92 opened pull request #103860 → Openafs next → https://git.io/JksS7
rajivr has joined #nixos
Deiru has quit [Remote host closed the connection]
<Reiser>
How do I capture docker logs in 20.09, they're no longer caught by journalctl it seems
<zeta_0>
hello there, I setup the emacs-overlay in a emacs.nix file, as explained by the repo: https://github.com/nix-community/emacs-overlay , then I imported the file in home.nix in home.packages as explained by the emacs instructions in the nixos manual( https://nixos.org/manual/nixos/stable/index.html#tip ): `(import ./emacs.nix { inherit pkgs; })` , but when I ran `nixos-rebuild switch --upgrade`, the terminal returned
<zeta_0>
the error: `undefined variable: emacsWithPackagesFromUsePackage`, is there an easy way to fix this so that the emacs-overlay specific variables are recognized, thus fixing the error returned by the terminal?
<hexa->
call docker logs?
<hexa->
and maybe configure logDriver journald
mmohammadi9812 has quit [Ping timeout: 256 seconds]
mmohammadi9812 has joined #nixos
<Reiser>
hexa-, tried, problem is, container is failing immediately and the container is purged
<Reiser>
so unless I imte my docker logs to some insane precision
<Reiser>
what I want to know is gone by the time I run it
<Reiser>
journalctl used to capture everything in 20.03
<Reiser>
hexa-, that looks super promising, I'll dig, thank you so much
<zeta_0>
I could easily just setup emacs-overly in my home.nix(home.packages) file like I did before, but i'm trying to keep my config files simple and organized, that's why I i'm trying to setup emacs-overlay in an emacs.nix file, and then importing it into home.nix(home.packages)!
<raghavsood>
So at least Parsec should support it, no idea what you'd need to get it to support Wayland on NixOS though
<matthewcroughan_>
raghavsood: it totally doesn't, I think they misunderstood what Wayland was.
<matthewcroughan_>
I think they tested that it works in Xwayland, then though that meant Wayland support, that changelog is erroneous.
Yaourt has joined #nixos
<matthewcroughan_>
The reason I think this is because 1. It doesn't work in Wayland, 2. They say "Wayland in Ubuntu" which doesn't make any sense 3. Wayland/Mir in Gnome was a thing that was developing back then in 2018, and there was a lot of people getting confused about what Wayland even was.
Supersonic112 has joined #nixos
Supersonic has quit [Ping timeout: 272 seconds]
Supersonic112 is now known as Supersonic
b has quit [Quit: Lost terminal]
kalbasit has joined #nixos
Yaourt has quit [Remote host closed the connection]
bn_work has quit [Quit: Connection closed for inactivity]
euandreh has joined #nixos
<leonardp>
i should add: i want to use podman and want to change the location of the local image store
<leonardp>
or should this be done with a bind mount?
orivej has joined #nixos
Darkmatter66 has joined #nixos
asheshambasta has joined #nixos
<asheshambasta>
So, I've been using `niv` lately and I'm trying to use its overrides to use a local dependency. It seems to work pretty well, but I'm having trouble getting local overrides to work. This is when you tell niv to use a local path for a particular override in its sources.json. Niv does this with environment variables.
<asheshambasta>
In my case, I seem to be confusing nix; in my output I see "unpacking sources ... do not know how to unpack source archive /some/override/path".
<asheshambasta>
Has anyone else encountered this, or has tips on where/what I can do to make it work?
revtintin has quit [Quit: WeeChat 1.9.1]
<bqv>
Is there a thing to shuffletidy nix attrs, or best I do a json dance?
<bqv>
I got a deeply nested one, wanna dedup all the keys
<danderson>
Is there a nixos option to add more things to $PATH globally? Similar to environment.systemPackages, but I need to just alter $PATH.
<danderson>
ah, environment.pathsToLink I think
<bqv>
danderson: no, and also, what
<danderson>
hrm, no.
<danderson>
So. For $day_job I develop a thing that runs as a system daemon. So I set up a systemd unit to run the binary out of my development git repo, but I want to also add that binary to $PATH (for the CLI tools).
<etu>
danderson: Not sure what happens if you put it in environment.variables
<danderson>
the obvious way seemed to be "get it dumped into /run/current-system/sw/bin somehow" :)
alp has quit [Ping timeout: 272 seconds]
<bqv>
can't that be a systemd variable?
<bqv>
Or a shell variable?
<bqv>
Or even both
<etu>
Sounds like a better idea
<bqv>
Considerably
<danderson>
For running the service, sure. But I want to run this stuff from my terminals also, not just via systemd. So yes, I'm trying to alter $PATH
<raghavsood>
environment.interactiveShellInit should be able to let you add an alias to the fully qualified binary, or update your $PATH
<colemickens>
but why not just install it as a system package?
<danderson>
because while I'm iterating during development, it's not a nix derivation
<danderson>
and building a nix derivation is at least 10x slower than the incremental rebuild via redo :)
<colemickens>
Oh, I see, I wasn't parsing the question correctly.
<danderson>
but, having poked around, I think I want home-manager's home.sessionPath
<raghavsood>
danderson: Just buy a faster computer /s
<danderson>
which adds stuff to $PATH
<raghavsood>
You could also just set up a nix-shell that you can enter while you're doing your development which adds it onto the PATH
<danderson>
I also run the daemon in question outside working hours for other purposes, so I want it in my general path, not an env-specific one
<danderson>
otherwise yes, I'd have tweaked my .envrc
<danderson>
effectively, I want to get as close to "enable this program in nixos, only running my bleeding edge mostly broken binaries out of ~" as I can :)
<raghavsood>
Since the channel seems more alive, anyone know what the process for proposing some structural changes is? I'd like to redo most of the blockchain stuff to be more consistent, and have a clearer separation between blockchains vs blockchain tools/companion programs. This would involve updating the nixpkg directory layout for blockchains, as well as
<raghavsood>
adding/rewriting some modules to help with consistency
<etu>
danderson: I usually have ~/bin linked in my path
<raghavsood>
Do I just open a github issue? Post on discourse? Some mailing list?
<etu>
danderson: There's an option for that
<etu>
danderson: Couldn't you create symlinks from there to your scripts/binaries
<etu>
Then you don't have to alter the PATH :)
m1cr0man has quit [Ping timeout: 256 seconds]
<etu>
danderson: environment.homeBinInPath
Izorkin has quit [Ping timeout: 272 seconds]
<danderson>
sure, but then I have things that aren't DECLARATIVE, DAMNIT
<danderson>
... but yeah, that also works :P
<etu>
Use home-manager to create the symlinks then :D
<raghavsood>
Do that, and then have a oneshot systemd script or home-manager to symlink it there for you
<MichaelRaskin>
raghavsood: I think reorganising the Nixpkgs structure might either go easily or be bogged down for the same reason of many people considering the current organisation an unsalvageable mess anyway
mallox has joined #nixos
<MichaelRaskin>
For modules — just open PRs I guess?
<raghavsood>
MichaelRaskin: Current organisation of the entire package tree, or just blockchains? I think doing it just for the blockchain stuff is approachable, we don't package a lot of them, and even fewer have accompanying modules
konobi has quit [Remote host closed the connection]
<MichaelRaskin>
Entire tree
FRidh has quit [Ping timeout: 272 seconds]
konobi has joined #nixos
<MichaelRaskin>
I guess if you create a tracking issue for cleaning up everything blockchain, and start with adding small modules, you have a good chances. I guess if anyone cares enough to oppose, you might at least find the other people who care about the blockchains to plan a cleanup together!
<evils>
raghavsood: what labels do you want on that?
<raghavsood>
evils: I was thinking of 0.kind enhancement, 6.topic module system, 9.needs: community feedback, but feel free to add/remove if you think others fit better, my first time approaching an issue like this
delan has joined #nixos
<evils>
raghavsood: done, want the 0.kind: bug removed?
<bqv>
ah, nix eval will output pure nix. I can eval the attr to format it!
<{^_^}>
[nixpkgs] @primeos opened pull request #103894 → iproute: Rename the attribute to iproute2 → https://git.io/JkGoK
<bqv>
Now just the guids
werner291 has quit [Quit: werner291]
werner292 has joined #nixos
ericsagn1 has quit [Ping timeout: 260 seconds]
werner292 is now known as werner291
Izorkin has quit [Ping timeout: 260 seconds]
vsiddharth has joined #nixos
raghavsood has quit [Ping timeout: 245 seconds]
<vsiddharth>
Hey, I was wondering what's the best way to deal with an authorization key in Nix, especially Nix Flakes. I tried including the key in the flake inputs but it errors saying that file:///.../key.txt is unsupported.
<patagonicus>
psibi[m]: Pretty sure the config file should have system-features = kvm without the quotes. That would also explain the extra quotes in the error message.
<raghavsood>
Anyone have any insight on how to backport a package update to 20.03 if the changes between the last commit for that package ion 20.03 and master contain stuff that's not available in 20.03 (such as runVend for go packges)?
chloekek has joined #nixos
<chloekek>
Is it possible to customize the suggested command template for missing commands? Instead of “nix-env -iA <expr>” I’d like it to print “nix run <expr> -c <command>” so I can easily copy–paste that into the prompt.
* chloekek
takes look at source code of command-not-found. Seems to be a hardcoded message. Guess I can write a patch file and apply it with nixpkgs.config.packageOverrides.
<bqv>
Hey whats something that uses nix but not nixpkgs?
<dxtr>
Hmm, so it is possible to use Type=notify for services created in systemd.services? The issue right now is that systemd is expecting the pid of the shell to sd_notify
<freezeboy>
hi there ! Anyone has a good and fast config using nix in gitlab-ci?
werner291 has joined #nixos
cosimone has quit [Quit: cosimone]
wnklmnn has joined #nixos
<catern>
adisbladis: hm I just noticed an issue with the emacs site-start.el; it calculates and sets find-function-C-source-directory based on its own path - but it's installed into ~/.nix-profile and other places, which are early on load-path, so running a different version of Emacs will get an incorrect site-start.el and therefore an incorrect find-function-C-source-directory
<catern>
that variable really needs to be set in some file actually tied to the Emacs version... maybe by patching the .el defining it
FRidh has quit [Ping timeout: 272 seconds]
FRidh has joined #nixos
<dminuoso>
freezeboy: shell runner and do whatever you want.
hammer has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
raghavsood has quit [Remote host closed the connection]
mallox has joined #nixos
iH8c0ff33 is now known as iH8c0ff33-matrix
FRidh has quit [Ping timeout: 264 seconds]
iH8c0ff33 has joined #nixos
FRidh has joined #nixos
<catern>
adisbladis: and find-function-C-source-directory can also be confused by running emacs from emacs built with emacsWithPackages - in that case EMACSLOADPATH will be inherited to the sub-emacs. hmm..
o1lo01ol1o has quit [Ping timeout: 240 seconds]
mallox has quit [Client Quit]
FRidh has quit [Remote host closed the connection]
<__monty__>
Hmm, I have a nix-shell containing a python37.withPackages ... [ pillow ] but I cannot `import PIL`. The nixpkgs manual seems to suggest I should be able to.
<FRidh>
__monty__: can't do much without knowing the nixpkgs version
<FRidh>
otherwise that should be fine
<__monty__>
nix-info reports nixpkgs-21.03pre251181.dd1b7e377f6 should be a fairly recent unstable revision. I'm on darwin btw.
ManiacOfMadness has joined #nixos
bn_work has joined #nixos
<DavHau[m]>
If I build `pkgs.staticPkgs.python3`, modules from the standard library, like for example `math`, are missing. Does anyone know a fix for this?
rajivr has quit [Quit: Connection closed for inactivity]
<DavHau[m]>
I cannot find any Setup.* files in the cpython source tree
pushqrdx has joined #nixos
<pushqrdx>
so pantheon desktop on latest nix-unstable channel is completely broken
<pushqrdx>
and i don't have a backup point before it broke
<pushqrdx>
can someone help me identify and perhaps fix the issue at least temporarily until some fix is applied upstream?
<__monty__>
Hmm, if I add numpy to the packages I can't import that either. (Figured I'd test one of the modules used in the manual examples.) Should I avoid using mkShell maybe?
nextloop has joined #nixos
<__monty__>
pushqrdx: Easiest would be to roll back a generation.
<pushqrdx>
note: it's broken even on a fresh install from the unstable channel
<pushqrdx>
__monty__ i don't have a generation, it's a fresh uninstall
o1lo01ol1o has quit [Ping timeout: 256 seconds]
<pushqrdx>
install*
<__monty__>
Then I don't have good advice. Overlay patheon with a pinned version and try older and older commits.
<srhb>
pushqrdx: You could also try jumping to 20.09 for a while.
<srhb>
Or pointing rebuild at an earlier nixos-unstable.
<pushqrdx>
srhb most of the stuff i work with and need aren't there
<pushqrdx>
srhb yes that, how can i do that
FRidh has quit [Ping timeout: 260 seconds]
FRidh has joined #nixos
<srhb>
pushqrdx: I think -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/<sha1>.tar.gz might work
<srhb>
pushqrdx: With <sha1> being a commit that works for you.
<__monty__>
Uhm, actually, I'm getting a python 3.*8*.6 in my nix-shell. Even though I'm using python37.withPackages. Anyone know why this might happen?
<pushqrdx>
srhb where that argument is supposed to be passed to?
<srhb>
pushqrdx: To nixos-rebuild :)
<{^_^}>
[nixpkgs] @lopsided98 opened pull request #103920 → pkgs-lib: allow paths in TOML, YAML and JSON → https://git.io/JkZ4r
<__monty__>
Hmm, nix-shell -p 'python37.withPackages (ps: [ ps.pillow ])' --run 'python -c "import pillow; print(PIL.__version__)"' works fine. But I still haven't a clue as to why the shell.nix gives me python 3.8.6 rather than the 3.7 with packages.
<FRidh>
__monty__: ahh some other buildInput is propagating Python
<__monty__>
Is it black? Why would it do so?
<pushqrdx>
srhb i don't know why i have the feeling that running that command is installing everything from nixpkgs :D
<srhb>
pushqrdx: Well it will, especially if you have crossed a glibc version or some other mass-rebuilding thing, since that changes everything.
<__monty__>
pushqrdx: It shouldn't be. But it is pulling down an entire nixpkgs checkout.
<FRidh>
because buildPython* does so
<pushqrdx>
it's building some haskell stuff, i don't have any thing todo with haskell in my config, and i don't remember seeing that being built as a dependency before
<srhb>
pushqrdx: Picking a sensible sha1 that actually has a functional binary cache is also important
<FRidh>
just add black to your python env
<srhb>
pushqrdx: But that's hard right now, with hydra down, if you don't know what a functional version was before.
<pushqrdx>
srhb i just picked one 2 days ago..
<FRidh>
ps.black that is
<srhb>
oh, hydra is back.
<pushqrdx>
i just hope it isn't downloading the whole nixpkgs repo and building stuff i don't need for some reason :D
<__monty__>
FRidh: Ah, yep. That was the problem. Is that really a sensible design choice for the buildPython* functions?
<pushqrdx>
srhb so how can i use hydra to pick correct hash?
<{^_^}>
#102613 (by FRidh, 1 week ago, open): WIP: Python: wrap and patch using `requiredPythonModules` instead of `propagatedBuildInputs`
<FRidh>
good luck fixing several thousand expressions
<srhb>
pushqrdx: Pick a commit from one of the green checkmarks -- the commit is the last part of the name
<srhb>
pushqrdx: That should have most things cached.
<srhb>
pushqrdx: But unless you know what broke pantheon, it's hard to give better advice here. Perhaps more debugging is in order :)
<__monty__>
FRidh: I'll support your good work silently from the sidelines.
<pushqrdx>
srhb can i just use the truncated part of the hash on hydra or do i need to hunt for it on github?
<srhb>
pushqrdx: Try?
<pushqrdx>
hh ok i tried says unpacking and seems to be going on so i guess i can :D
<pushqrdx>
picked the one from 5 days ago, at least if it worked i will have a range to look for what broke everything and hopefully find something and file a bug report
<pushqrdx>
so way back earlier i was asking if i can pin my configuration.nix to an exact time/commit and no one seemed to offer that -I thing
<pushqrdx>
from what i can see i can specify nixpkgs=http://git.... and my config .nix would use that everytime right?
FRidh has quit [Ping timeout: 256 seconds]
FRidh has joined #nixos
<srhb>
pushqrdx: It's not (really) possible from _within_ the config, not without flakes.
<srhb>
pushqrdx: There are workarounds, but none of them are particularly pretty.
<pushqrdx>
haven't felt the need to checkout the flake stuff yet so i don't know what they are about yet
<srhb>
It's also very experimental. Just pointing out that it's not pretty without it, yet.
<srhb>
You could build your own nixos-rebuild wrapper that does it with eg. niv or something like it. That would be pretty clean.
acarrico has joined #nixos
growpotkin has joined #nixos
boxscape has joined #nixos
shabius has joined #nixos
o1lo01ol1o has joined #nixos
asheshambasta has quit [Ping timeout: 272 seconds]
shabius has quit [Client Quit]
FRidh has quit [Ping timeout: 256 seconds]
FRidh has joined #nixos
o1lo01ol1o has quit [Ping timeout: 256 seconds]
hnOsmium0001 has joined #nixos
<Reventlov>
Hey there
<Reventlov>
trying to launch nix-build on some derivation, on debian (I just installed nix), I get:
<Reventlov>
file 'nixos-config' was not found in the Nix search path (add it using $NIX_PATH or -I), at /home/remy/.nix-defexpr/channels/nixpkgs/nixos/modules/system/activation/top-level.nix:289:68
KarlJoad has joined #nixos
<das_j>
Reventlov: You want `nix-store -r /path/to/drv`
<KarlJoad>
I am having issues getting rpcbind and rpcinfo to work. I can run `sudo rpcbind`, and I see lock and sock files show up in /var/run, but I cannot get any information from them using rpcinfo. Does anyone have any insights? Do I need to add the program to my system-level configuration?
<{^_^}>
[nixpkgs] @lopsided98 opened pull request #103925 → coreutils: fix build on 32-bit ARM → https://git.io/JkZza
<{^_^}>
[nixpkgs] @soulsoiledit opened pull request #103927 → multimc: remove root directory flag → https://git.io/JkZgS
nek0 has joined #nixos
<pushqrdx>
good god, is there a way i can tell nixos-rebuild to use all cpu cores when it encounters a build.. i have been waiting for kernel rebuild for almost an hour
pushqrdx has quit [Remote host closed the connection]
<joepie91>
I'm confused. 20.09 is out, but nixpkgs-channels does not have a nixos-20.09 branch? did something change in how channels are made available/
<srhb>
joepie91: It was even in the release notes ^^
nek0 has quit [Remote host closed the connection]
arianvp has quit [Quit: WeeChat 2.7.1]
<joepie91>
srhb: but hold on, the benefit of nixpkgs-channels was supposed to be that it only updates once all of the packages in it have built on Hydra
<connelly>
and I got steam installed and downloaded two games (shenzhen I/O and among us) and neither runs properly. I tried running them with steam-run and haven't gotten that working either
arianvp has joined #nixos
<srhb>
joepie91: Same for the ones in nixpkgs.
<joepie91>
the regular nixpkgs branches don't provide this, do they?
FRidh has joined #nixos
<srhb>
joepie91: Yes, exact same behaviour.
<srhb>
release-20.09 is before hydra, nixos-20.09 is after.
Poumpaloumpa has joined #nixos
<connelly>
it says to launch a shell script for your steam game from ~/.local/share/applicaitons/gamename.deskop on the Exec line, but neither of those games points to a shell script, just launching the main steam app with a steam:// URL... I tried creating a shell script that does that and launching it in the FHS but that failed for both games
<srhb>
connelly: I think you best bet is to pastebin some examples of the failures, such that people who know about steam can try and help to debug exactly what's going wrong. Also there's #nixos-gaming which may bea good place.
<joepie91>
srhb: but don't backports get merged into branches, and only trigger a build when they are?
<joepie91>
into those branches*
<srhb>
Reventlov: Do you have config.copySystemConfiguration set to true in your VMs?
<joepie91>
blah, damn lag
<joepie91>
srhb: ah right, I didn't realize there were two separate branches for each, thanks
ManiacOfMadness has quit [Ping timeout: 240 seconds]
vandenoever has quit [Read error: Connection reset by peer]
vandenoever has joined #nixos
<connelly>
srhb: I'm not sure where to find any failures - it's not obvious where the logging is happening - it's not in the systemd journal and the command line responses are minimal and launching from the GUI just silently fails
<connelly>
srhb: no response in there
<srhb>
connelly: It' low traffic, so immediate responses are not that common. I don't have time to fully guide you, but if you're having trouble launching via steam-run, you could still get some debug logs by launching steam itself from a terminal.
Duponin has joined #nixos
o1lo01ol1o has joined #nixos
<Duponin>
Hi, I'm currently learning Nix (using NixOS) and I don't understand a behavior. I create a custom module, import from `configuration.nix` in to module, import and enable it. When I run `nixos-rebuild dry-build` there are a few derivations, is it normal?
<Duponin>
(I migrate tasks from the configuration.nix to the module, typo)
<KarlJoad>
I am having issues getting rpcbind and rpcinfo to work. I can see lock and sock files in /var/run, but I cannot get any information from them using rpcinfo. Do I need to add the program to my system-level configuration?
<Ke>
seeing derivations is normal, not sure what is the mystery
<Reventlov>
so now I get a build error: / a 'x86_64-linux' with features {kvm} is required to build […] but I am a 'x86_64-linux' with features {benchmark, big-parallel, nixos-test}
<Reventlov>
/dev/kvm exists, can I tell nix I actually support kvm ?
<Yaniel>
Duponin: note that it only builds the derivations that changed
<Yaniel>
and dry-build just lists what it would build
<Duponin>
Ke: I would expect to see no new derivations since I just tasks from a module to another
<connelly>
srhb: thank you - here's the paste of the console out when launching steam that way: https://pastebin.com/2Q141yf7
berberman_ has joined #nixos
<srhb>
connelly: Yes, that didn't look very useful. What's your problem with steam-run?
berberman has quit [Ping timeout: 260 seconds]
<connelly>
srhb: well it says to grab the shell script to launch the game from the Exec= line in the .desktop file, but that line just points to steam steam://... - if I create a shell script that calls that and run it from steam-run it pops up steam but doesn't launch the game and doesn't have a lot of output
o1lo01ol1o has quit [Ping timeout: 260 seconds]
<srhb>
connelly: It's usually pretty easy to find the executable in ~/local/share/Steam/steamapps/common/$yourgame and then steam-run that :)
<connelly>
srhb: it's a .exe file there tho
<srhb>
connelly: Is it a proton title?
<srhb>
connelly: As in, not native Linux?
<connelly>
srhb: yes, I tried it with launch options for proton and it didn't change much --- it's "Among Us"
<srhb>
Sorry, I have no experience debugging Proton games.
<{^_^}>
[nixpkgs] @Ericson2314 pushed 2 commits to staging: https://git.io/JkZwG
bbigras has quit [Ping timeout: 260 seconds]
kraem has quit [Ping timeout: 260 seconds]
Alex[m] has quit [Ping timeout: 260 seconds]
m_hackerfoo has quit [Ping timeout: 260 seconds]
kamid has quit [Ping timeout: 260 seconds]
Alex[m] has joined #nixos
m_hackerfoo has joined #nixos
Synth[m] has quit [Ping timeout: 260 seconds]
jbal[m] has quit [Ping timeout: 260 seconds]
DarkTrancer[m] has quit [Ping timeout: 260 seconds]
Hm7000 has quit [Ping timeout: 260 seconds]
Jezhehim[m] has quit [Ping timeout: 260 seconds]
puzzlewolf has quit [Ping timeout: 260 seconds]
domenkozar[m] has quit [Ping timeout: 260 seconds]
witchof0x20[m] has quit [Ping timeout: 260 seconds]
tlater[m] has quit [Ping timeout: 260 seconds]
govanify has quit [Ping timeout: 260 seconds]
tet[m] has quit [Ping timeout: 260 seconds]
SpaghettiCthulhu has quit [Ping timeout: 260 seconds]
MinceR has quit [Ping timeout: 260 seconds]
DarkTrancer[m] has joined #nixos
jbal[m] has joined #nixos
Hm7000 has joined #nixos
Synth[m] has joined #nixos
Jezhehim[m] has joined #nixos
witchof0x20[m] has joined #nixos
kamid has joined #nixos
MinceR has joined #nixos
SpaghettiCthulhu has joined #nixos
tet[m] has joined #nixos
govanify has joined #nixos
domenkozar[m] has joined #nixos
tlater[m] has joined #nixos
lsix has quit [Ping timeout: 260 seconds]
kraem has joined #nixos
puzzlewolf has joined #nixos
bbigras has joined #nixos
zupo has joined #nixos
alp has joined #nixos
werner291 has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @yvesf opened pull request #103930 → firefox: fix wayland-lto issue reference number → https://git.io/JkZwN
cantstanya has quit [Remote host closed the connection]
cantstanya has joined #nixos
mohe2015 has joined #nixos
<mohe2015>
Hi, is there any reason for "error: --- Interrupted ---------------------------------------------------------------------------------------------------------- nixinterrupted by the user" when running `sudo nixos-rebuild --max-jobs 4 boot` and not watching the build? I assume this may happen in an out of memory situation but I don't understand how
<mohe2015>
it claims to be interrupted by the user
<srhb>
mohe2015: interrupted does sound a bit weird.
<clever>
KarlJoad: you can run `iptables-save` to print the current tables
fendor has quit [Remote host closed the connection]
fendor has joined #nixos
mohe2015 has joined #nixos
lsix has joined #nixos
malorie has joined #nixos
Poumpaloumpa has quit [Quit: Konversation terminated!]
<malorie>
I just tried configuring ibus without success, and now ^F isn't working anymore in my terminal. any idea how to fix this?
<malorie>
I've already tried removing ibus again, but the problem persists
<KarlJoad>
Ok. I figured out the rpcbind issue, for real this time. You should add it as a system service from /etc/nixos/configuration.nix, instead of nix-env -i it.
zupo has quit [Ping timeout: 272 seconds]
<{^_^}>
[nixpkgs] @roberth pushed 65 commits to staging-20.09: https://git.io/JkZ6p
<{^_^}>
[nixpkgs] @nuance opened pull request #103933 → emacs: Link native-lisp directory into Mac app → https://git.io/JkZiK
gxt has quit [Remote host closed the connection]
<hexagoxel>
mostly out of curiosity: is there a way to nix-instantiate --eval an expression that is a string without escaping it? Fine if it fails if encountering a non-string..
gxt has joined #nixos
<malorie>
loading one of the generations before I added ibus makes ^F work again, but booting into my current one (with the same configuration.nix) doesn't :\
<malorie>
is there a way to effectively revert what I did, and branch off the generation that is fine?
heyitsrama has joined #nixos
werner291 has joined #nixos
o1lo01ol1o has joined #nixos
ece has quit [Ping timeout: 260 seconds]
alp has joined #nixos
heyitsrama has quit [Ping timeout: 240 seconds]
o1lo01ol1o has quit [Ping timeout: 256 seconds]
domogled has joined #nixos
vandenoever has quit [Read error: Connection reset by peer]
vandenoever has joined #nixos
Darkmatter66 has quit [Quit: ZNC 1.7.5 - https://znc.in]
<notgne2>
hexagoxel (chat.freenode.net): I don't think so, your best bet is probably `... --json | jq -r`
<joepie91>
srhb: das_j: I'm running into an issue where most of my repo pushes in gitea are broken, and this thread suggests that the problem is the INTERNAL_TOKEN being re-generated on each start: https://github.com/go-gitea/gitea/issues/8251 -- now Gitea has been patched to generate one on first boot only and then store it in the app.ini for future use... which, of course doesn't work on NixOS because it re-copies the app.ini on every
<joepie91>
service start :) any ideas on how this might be solved in the nixpkgs module itself, conceptually?
<{^_^}>
go-gitea/gitea#8251 (by randyesq, 1 year ago, closed): Gitea: Internal error retrieve protected branches information failed: EOF
<joepie91>
I can work around it locally by hardcoding a token in my config, but this is obviously not going to help any other users of that module
<das_j>
ummmmm
<joepie91>
(always fun when an application writing back into config files ruins everyone's day... :P)
MarcWeber has quit [Remote host closed the connection]
oxalica1 has joined #nixos
<joepie91>
I don't know if there's a way to obtain that INTERNAL_TOKEN outside of the "start the daemon the first time and let it write into the config" process
mananamenos_ has quit [Read error: Connection reset by peer]
<joepie91>
(I'm actually about to test whether this even fixes my issue right now)
oxalica has quit [Ping timeout: 256 seconds]
oxalica1 is now known as oxalica
<das_j>
can you generate the token yourself? It's probably just a random string of hexadecimal characters, right?
<joepie91>
not sure. it starts with eyJhb
<joepie91>
wasn't that JWT?
<joepie91>
(hi eyJhb :P)
<joepie91>
meh okay apparently it doesn't even fix my issue :(
<das_j>
oof
<joepie91>
though it still looks like a break-y thing to me
pn has joined #nixos
pn has quit [Client Quit]
<eyJhb>
joepie91: you summoned me?
<eyJhb>
:p
<joepie91>
as in, it's behaving differently than Gitea expects and nuking its state, and so that might cause issues in the future
<{^_^}>
[nixpkgs] @AndersonTorres pushed 3 commits to master: https://git.io/JkZ9C
civodul has quit [Quit: ERC (IRC client for Emacs 27.1)]
pistache has joined #nixos
heyitsrama has quit [Ping timeout: 256 seconds]
<pistache>
hi! should I avoid circular references when using the Nix langage ?
<pistache>
language*
<colemickens>
I think you will indeed need to, for the most part. Did you have something in mind?
<pistache>
colemickens: I'm using Nix to specify and build the configuration for a deployment system, using nested attrsets to represent domain/host/services
<pistache>
I've made it possible for domain/hosts to specify default values for services, where the default values may use values defined in the services (using lib.fix)
<pistache>
I think it could be useful to provide a reference to the domain and the host in the service attrset, but that would imply a circular reference
sss2 has quit [Quit: Leaving]
<{^_^}>
[nixpkgs] @leo60228 opened pull request #103940 → netcoredbg: init at 1.2.0-635 → https://git.io/JkZ9x
<colemickens>
oh that's definitely a bit over my head then, sorry I can't advise further.
KarlJoad has quit [Quit: ERC (IRC client for Emacs 27.1)]
Acou_Bass has joined #nixos
cjpbirkbeck has joined #nixos
lord| has joined #nixos
turlando has quit [Remote host closed the connection]
turlando has joined #nixos
vidbina_ has quit [Quit: vidbina_]
vandenoever has quit [Quit: Konversation terminated!]
moet has joined #nixos
<moet>
hi! i've been having a recent problem with tmux in urxvt where panes i haven't viewed in awhile are partially undrawn.. it doesn't happened with bash, but does happen with terminal applications that take the whole character buffer such as vim, htop, and others.. it's become very bad after upgrading to 20.09 .. i suspect this is a niche issue that very few people have, so i'm not here to complain.. i'm just
<moet>
here to ask if there are others using tmux on urxvt as their daily driver, and whether they've encountered this
<cjpbirkbeck>
okay, so did some testing, and i cannot get any display or usb keyboard/mouse to work - although the power seems fine, the ethernet light is working and when i plug a usb hard drive, it seems to light up for at least a little while
shibboleth has joined #nixos
<cjpbirkbeck>
this occurs regardless of what drive i have plugin
<cjpbirkbeck>
it just seems very wierd that an os upgrade could do that
<simpson>
cjpbirkbeck: That could be your motherboard firmware, though. What you'll probably want to do is boot using your original NixOS install media, mount your system and nixos-enter it, and then nixos-rebuild to switch to a working generation, including the older boot loader.
<simpson>
You're dual-booting, you changed the bootloader on the Linux side, and you upgraded Windows; either of the latter two could have disturbed the careful balance required for dual-boot.
<cjpbirkbeck>
simpson: but how i do that without a monitor?
<cjpbirkbeck>
having said that, i'm guess that's the problem here
o1lo01ol1o has joined #nixos
<simpson>
cjpbirkbeck: Your NixOS install media doesn't boot anymore!? That *is* a problem. Can you get into the motherboard firmware?a
vidbina has joined #nixos
lord| has quit [Ping timeout: 272 seconds]
vidbina has quit [Client Quit]
<ivan>
cjpbirkbeck: check your motherboard manual to see how to do a CMOS reset / Reset BIOS settings
o1lo01ol1o has quit [Ping timeout: 256 seconds]
<ivan>
cjpbirkbeck: UEFI can sometimes just break itself from updating e.g. Boot variables
<ivan>
if you get it working again, see if there is a BIOS update
vidbina has joined #nixos
<ivan>
if that doesn't work there may be another procedure for flashing the BIOS from USB e.g. ASUS calls it FlashBack
<ivan>
a dumb thing you can try before CMOS reset or flash is to drain all the power properly and wait a minute
vidbina has quit [Client Quit]
vidbina has joined #nixos
vidbina has quit [Client Quit]
vidbina has joined #nixos
<ivan>
if you decide your motherboard is broken because of UEFI variable updates and there is no vendor fix, I believe you can configure nixos to not touch your boot variables and just boot from the UEFI fallback boot location
DigitalKiwi has quit [Quit: quite.]
DigitalKiwi has joined #nixos
ericsagn1 has quit [Ping timeout: 260 seconds]
<samueldr>
doesn't windows now sometimes do bios upgrades through their updates too?
vidbina has quit [Client Quit]
vidbina has joined #nixos
vidbina has quit [Client Quit]
vidbina has joined #nixos
pushqrdx has quit [Remote host closed the connection]
vidbina has quit [Client Quit]
vidbina has joined #nixos
werner291 has quit [Quit: werner291]
vidbina has quit [Client Quit]
malorie has quit [Remote host closed the connection]