<s1341_>
good morning. I'm using fetchgit to grab a bunch of dependencies that my derivation needs. How can I tell nix that I want a subdirectory of the fetched git repository to be the root for that source?
<jboyens[m]>
IIRC, you can use sourceRoot = "source/directory_you_want";
<dasj19>
morning, I have been tried yesterday to package a python program which has 3 binaries that should communicate with eachother via dbus. I can't seem to get the dbus working. Anyone that can help?
griff_ has joined #nixos
evanjs has quit [Read error: Connection reset by peer]
<LouisDK>
I'm trying to make this package: https://github.com/LouisDK1/nixpkgs/commit/1a44d79efb032440c676253d783933718f4179b6 and when I start mate-terminal or another GTK app I get this error: "Settings schema 'org.appmenu.gtk-module' is not installed". Debugging with strace it seems like the mate-terminal is not looking in the right directory and I don't
<LouisDK>
know how to fix that and whatever it can be done by setting variable.
<azazel>
adisbladis: the compilation of the latest pgtkgcc variant of the emacs overlay fails here. I see that the source is pinned to 4 days ago. There's any particular reason to not pin a more recent version?
<emilsp>
what should I do to not encounter the "nix-env: src/libexpr/attr-set.hh:54: void nix::Bindings::push_back(const nix::Attr&): Assertion `size_ < capacity_' failed." bug?
ahmed_elgabri has quit [Ping timeout: 250 seconds]
HenrikK has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<pennae>
hm. wondering whether we should abandon #121888 and redo the mosquitto module completely to use freeform config, but it doesn't lend itself to that very well with some options that repeated and have local suboptions that can also be applied globally
<eyJhb>
Anyone that has fixed this error with arc-theme? https://termbin.com/o732 it is really annoying me. And the current patches doesn't seem to be updated or apply. Cannot get them to work...
stree has joined #nixos
ManiacOfMadness` has joined #nixos
alexherbo25 has joined #nixos
griff_ has joined #nixos
alexherbo2 has quit [Ping timeout: 245 seconds]
alexherbo25 is now known as alexherbo2
ezemtsov has quit [Ping timeout: 246 seconds]
ajs124 has quit [Quit: Bridge terminating on SIGTERM]
Scriptkiddi has quit [Quit: Bridge terminating on SIGTERM]
mrpi has quit [Quit: Bridge terminating on SIGTERM]
cheriimoya has quit [Quit: Bridge terminating on SIGTERM]
marsh has quit [Ping timeout: 260 seconds]
ajs124 has joined #nixos
cheriimoya has joined #nixos
mrpi has joined #nixos
Scriptkiddi has joined #nixos
rubm has joined #nixos
ddellac__ has joined #nixos
ezemtsov has joined #nixos
sdfsefvcsvesavev has joined #nixos
marsh has joined #nixos
ezemtsov has quit [Read error: Connection reset by peer]
ddellac__ has quit [Ping timeout: 268 seconds]
sdfsefvcsvesavev has quit [Quit: Connection closed]
lordcirth_ has quit [Remote host closed the connection]
lordcirth_ has joined #nixos
<djahandarie>
In particular I want it to grab kernelPackages.mwprocature from unstable (which I have imported in my configuration.nix under `unstable')
tomasj has quit [Quit: Ping timeout (120 seconds)]
tomasj has joined #nixos
tomasj has left #nixos [#nixos]
<djahandarie>
Here's a related issue I just filed, maybe someone here could help me figure out how to get a local development version of a package in my kernelPackages..
<{^_^}>
[nixpkgs] @cole-h merged pull request #122746 → telfhash: add support for new TLSH hash format, which has been the default since TLSH 4.0.0. Deprecate support for the o… → https://github.com/NixOS/nixpkgs/pull/122746
<cole-h>
at least on nixUnstable, `--option` is "option"al now (sorry, couldn't resist :D)
<joedevivo>
haha
<joedevivo>
oh, completely off topic, do you think we should enable ssm-agent by default in the 21.05 AMI? it's easy enough to turn off, but if you need it, it's quite a pain to build your own ami just for that one thing
<cole-h>
gchristensen: ^ ?
rubm has quit [Ping timeout: 252 seconds]
<Ankhers>
joedevivo: I'm trying to build R23 right now on current master. Unfortunately I do not have a mac to test this on. I will in a week or two though.
rubm has joined #nixos
<gchristensen>
joedevivo: you don't need to build an AMI for it, you could pass configuration.nix info in userdata
shibboleth has joined #nixos
<joedevivo>
gchristensen: ahh right... all my drama was because it wasn't working, so doing that still couldn't get me on the instance. Now that it's fixed, I totally can do that :D
<gchristensen>
nice
<gchristensen>
that said, I'm open to the idea that it should be enabled by default!
<gustavderdrache>
as a heavy user of SSM i would *love* to see the agent enabled by default
<gustavderdrache>
it makes it easier to securely log in, too, since you can disable sshd in configuration.nix
<joedevivo>
Ankhers: ohai! It's looking like nix-darwin set my /etc/nix/nix.conf `cores = 0` which was too many cores, and parts of hipe were building out of order
<Ankhers>
hipe also hasn't really been maintained for a while. I believe they will be removing it at some point. Maybe they did already? I don't really know.
<Ankhers>
Also, if you can move to R24, the JIT is AMAZING.
<joedevivo>
yeah, hipe is missing from 24
<Ankhers>
Oh okay good.
<joedevivo>
I've got services to maintain on 22 for the time being
<Ankhers>
Fair enough.
<joedevivo>
A benefit here is that once it's built, it's built. so even if I have to spend a half hour building it with a single core, at least it's only the once
<Ankhers>
That is very true.
sangoma has joined #nixos
<joedevivo>
gchristensen: gustavderdrache: I'd say since the ssm-agent is enabled by default for both the standard ubuntu AMI and the Amazon Linux AMI, it's what I expected to be happening with the NixOS one as well
<gchristensen>
yeah
<gchristensen>
strong argument for enabling by default
lordcirth_ has quit [Read error: Connection reset by peer]
lordcirth__ has joined #nixos
stree has quit [Quit: Caught exception]
stree has joined #nixos
jmeredith has joined #nixos
Cyrinux3301386 has joined #nixos
<busti>
How should I setup a development environment for C programming if I want to use an IDE that requires project libraries installed? Should I launch the ide from my dev environment nix-shell instane, or should I just install all deps with nix-env?
ddellaco_ has joined #nixos
<adisbladis>
nix-env is not for development environments
<adisbladis>
Don't do that, it won't work for various reasons
<adisbladis>
Ideally your IDE can load enviroments using nix somehow
<adisbladis>
Maybe your IDE of choice has a direnv integration
<busti>
I would like to use clion since I have used other jetbrains stuff before.
<busti>
I haven't done much C programming though and most other buildtool ecosystems I have used before usually download dependencies into the local build environment
<adisbladis>
IME jetbrains are some of the most terrible to integrate Nix with
growpotkin has joined #nixos
<adisbladis>
They have a very opinionated view of what environments are and where libs should be found
<busti>
I might be a little locked in to the jetbrains ecosystem tbh. I have used it for the past 8 years or so and VSC looks and feels totally alien to me.
lordcirth_ has joined #nixos
MmeQuignon has quit [Ping timeout: 265 seconds]
ddellaco_ has quit [Ping timeout: 245 seconds]
<adisbladis>
At my previous job the jetbrains users created a bunch of wrapper scripts for each tool that the IDE had to be aware of
<adisbladis>
It wasn't a great workflow..
meh` has joined #nixos
lordcirth__ has quit [Ping timeout: 240 seconds]
philr has quit [Ping timeout: 246 seconds]
<pennae>
maybe something can be built with buildFHSUserEnv? (never had to, so no idea whether it'll actually work)
<adisbladis>
pennae: That means starting your IDE from within the env, which I think is pretty bad UX
<busti>
Looks like I was almost right. Other people seem to be successful with adding the IDE to `buildInputs` and running it from nix-shell.
<phalrax>
this will install the plugin on my system, but it's looks like it's not available inside grafana: declarativePlugins = with pkgs; [ grafana-image-renderer ];
<simpson>
zeta_0: It's probably not a loop. Login is a series of handoffs between processes. Your user session is probably crashing, and once it's no longer running, the login prompt refreshes.
<simpson>
If this is a non-graphical session, examine your .bashrc or other login-shell profiles. If this is a graphical session, examine your .xsession or whatever desktop-environment session you've chosen.
rail has quit [Quit: gone...]
<busti>
Yeah, opening the ide from nix-shell is bad UX. Ideally I would like to install it using either my system config or home manager and then have it load the required env when I select a project from the list in the launcher.
BlessJah has quit [Remote host closed the connection]
lordcirth_ has quit [Remote host closed the connection]
lordcirth_ has joined #nixos
<srid>
Can `nix run` be told to spit out output line by line instead of doing fancy one-line UI?
shla has joined #nixos
<cole-h>
-L
<cole-h>
maybe
<zeta_0>
simpson: I'm currently using plasma with lightdm, so ok, I'll take a look at the .xsession logs, also, whenever I login as root, I get passed the login prompt and the desktop env works fine, why does root login work, but not my user login?
slack1256 has quit [Remote host closed the connection]
<srid>
`nix run -L` has no effect
ddellac__ has quit [Ping timeout: 245 seconds]
<srid>
Oh, it is `nix -L run`
<simpson>
zeta_0: No idea. Also be aware that running graphical logins as root is a massive security risk.
lordcirth_ has quit [Remote host closed the connection]
<zeta_0>
I thought that you can tinker with configuration.nix as much as you want, and if something breaks, just go back to the generation before the change, it's strange that my issue is still persisting after new rebuilds.
eoli3n_ has quit [Remote host closed the connection]
<zeta_0>
pennae: I haven't set that immutableUsers option, but rolling back fix the issue, so I was able sign into the login prompt with my user, but now after doing new rebuilds, that login issue comes back.
waleee-cl has joined #nixos
<zeta_0>
I'm guessing it has to do with the uid option?
<zeta_0>
so what is the default uid?
<pennae>
zeta_0: that was just meant to say that "tinker with configuration.nix all you want" does have some exceptions, not that it's the cause of your issues
<zeta_0>
Do any of you recommended that I just reflash my microsd card and reinstall nixos on my pi from scratch, or is there a way to fix this issue?
<pennae>
cole-h: tail -n+3? :D
<romildo>
cole-h, probably I have now just find out how to get the date from the unix epoch
<pennae>
romildo: date -d @$epoch in scripts
<cole-h>
^
<simpson>
zeta_0: I recommend *in general* that, for embedded devices which are reliant on prepared boot images, one should have a separate machine which can reproduce those boot images. I also recommend not using such a fancy DE on low-horsepower devices; the Pi perenially struggles with memory bandwidth and its GPU is weak.
<simpson>
zeta_0: Of course you can fix this issue if you can login as root. Switch generations so that *no* user is configured in configuration.nix (and users are mutable), and then use the standard sysadmin techniques (useradd, etc.) to configure the system's users as you see fit.
<zeta_0>
simpson: ok, thanks, and ya, I'll switch from plasma to xmonad, a window-tiler should be much more light-weight correct?
<samueldr>
zeta_0: your issue is 99.99999% certainly not a Raspberry Pi specific issue, just saying
<samueldr>
I strongly assume that if you had done all the same operations on an x86_64 machine it would have done the same
<zeta_0>
samueldr: ya, I shouldn't have changed the uid option, I'm used to changing whatever I want, and going back a generation if something breaks, I guess I need to be more careful.
<exarkun>
how do I make the systemd/acme integration actually renew certs periodically, on its own? Sure, acme-<hostname>.timer fires every day but nothing ever *happens*. acme-<hostname>.service never /runs/ as a result. does no one use this? or is there some secret I just don't know about?
vidbina has quit [Ping timeout: 268 seconds]
<hyper_ch>
ma27[m]: online?
<simpson>
exarkun: I haven't used this specific integration, but ACME integrations usually have some sort of local check for certificate freshness, and perhaps somebody helpfully added something like that to this integration.
<simpson>
There's almost certainly some secret that we don't know about~
ddellac__ has joined #nixos
<odiug>
exarkun: I used certbot with a cron job.
sangoma has quit [Ping timeout: 245 seconds]
shibboleth has quit [Ping timeout: 240 seconds]
sangoma has joined #nixos
<exarkun>
simpson: It doesn't seem to be a cert freshness check. As far as I can tell nothing ever runs at all. If there is a hidden freshness check then it still hasn't triggered as of 9-days-and-counting.
<joedevivo>
Where's the code in the nix daemon that hands it's env vars off to nixbld users?
rajivr has quit [Quit: Connection closed for inactivity]
<simpson>
exarkun: Irritating. What happens if you manually start the service?
<exarkun>
simpson: I've done that in the past and it renewed the cert
<exarkun>
I haven't done it on this exact system...
<exarkun>
I feel like the problem is some basic misunderstanding of how systemd timers work
<zeta_0>
it's very frustrating installing nixos for the 1st time on a machine, as well as setting things up, but after that it gets better, definitely less frusting than other os's.
<exarkun>
But I lack a basic understanding of how systemd timers work ... so ...
ddellac__ has quit [Ping timeout: 260 seconds]
<hyper_ch>
why doesn't nginx work with acme? May 14 20:43:30 vpn acme-...-start[23690]: ....] acme: error: 403 :: urn:ietf:params:acme:error:unauthorized :: Invalid response from http://vpn.uniestro.com/.well-known/acme-challenge/AYCCd0CLoGwuTcYzPeqyszQtPYs3pyR_RcTatqhyVr0 [46.140.15.198]: "<html>\r\n<head><title>403 Forbidden</title></head>\r\n<body>\r\n<center><h1>403 Forbidden</h1></center>\r\n<hr><center>nginx</center>\r\n"...
<exarkun>
hyper_ch: Lots of reasons!
<hyper_ch>
I just get 403
disasm_ has quit [Quit: WeeChat 2.0]
<exarkun>
hyper_ch: Make sure DNS is pointed at the right host.
disasm has joined #nixos
<exarkun>
hyper_ch: Beyond that, it depends on how you configured acme and nginx.
<hyper_ch>
exarkun: next-cloud is the problem again
<hyper_ch>
grabbing it all with
eoli3n_ has quit [Remote host closed the connection]
sangoma has quit [Read error: Connection reset by peer]
<madonius[m]>
Hey, I am currently having a very specific issue. I have two laptops (Work/Private) that I ideally would like to manage from one single repository. Is there any example/recommendation on how to best achieve this?
<justanotheruser>
I've run the script as a standalone and it adds the dependencies properly
<odiug>
madonius[m]: For laptops, I'd opt to manage these independently.
<odiug>
Add some user directory sharing via samba.
<odiug>
and I would keep work/private separate for privacy reasons. In case a boss wants access to your work-laptop.
<typetetris>
madonius[m]: flakes offer that to some degree, you can have multiple nixos system definitions in one flake and you can specify to the cli, which one to build and switch (or if you don't choose, it will take a spec according to hostname). But you would have to use flakes for that.
<typetetris>
I would guess it takes a bit more than an evening. You have to learn the flakes stuff and the cli-ui changed a bit. (For me it is a progress still ongoing.)
<typetetris>
But you can get rid of niv and the likes and you get cached nix evaluations for super fast nix-shells .... (I still have to see if it was worth it)
jjakob has joined #nixos
jimmie_ has quit [Remote host closed the connection]
<lalaland>
I see. so what would you say is the most common use case of the nix os tech? on the website, it seems that it lets you spin up an environment with python installed, which is cool. Is it's purpose similar to npm?
<legendofmiracles>
Sure, you can use nix as a node package manager, but it's really much more than just that
<MichaelRaskin>
nix-shell could be interpreted as virtualenv but for _any_ kind of packages, at once
<legendofmiracles>
Yeah, for instance i use nix-shell for having rust installed inside my projects and not outside. It also helps with having different rust versions (nightly and stable)
<MichaelRaskin>
(with sharing between virtualenvs but without conflicts)
aiken has quit [Disconnected by services]
aiken has joined #nixos
<MichaelRaskin>
NixOS is basically saying that your entire system configuration is a package, and you can again have multiple installed (no conflicts but sharing what can be shared), then choose any of those on boot
aiken has quit [Disconnected by services]
<lalaland>
very cool, thanks. Is the website the best resource to learn to use nix os? Or do you use another resource?
aiken has joined #nixos
aiken has quit [Disconnected by services]
aiken has joined #nixos
<MichaelRaskin>
I, personally, started by reading the three manuals. But that was almost 14 years ago.
<legendofmiracles>
woah
ddellac__ has joined #nixos
aiken is now known as Guest738
Guest738 has quit [Killed (karatkievich.freenode.net (Nickname regained by services))]
<pennae>
they're still good today
<MichaelRaskin>
(and, uhm, I have also learnt Scheme by reading MIT Scheme reference manual — de facto, R5RS — and Bash scripting by reading man bash, so I have a bit of a tendency to just read the manual)
Guest738 has joined #nixos
lsix has joined #nixos
<MichaelRaskin>
Manuals are still good, but now there is some more choice for people with different style preferences, but I do not know which are good for whom
<joedevivo>
I see a lot of reference to not using single-user installs, but I don't see much explanation as to why not. Anyone know where the why might be hiding?