<samueldr>
some of it was verified to be useful on gru-dumo, a gru-scarlet variant (chromebook tablet)
<samueldr>
(as in, keyboardless tablet)
<hanetzer>
relatively. I'm familiar with alarm, but I'm one of those anti-systemd wierdos :P
<samueldr>
it's not based on ALARM, only using the work they did on getting a kernel going :)
<hanetzer>
I mean, I'll use it, but not if I can avoid it. and in this case, 'normal' gentoo would be hell on the chromebook.
<hanetzer>
yeah.
<hanetzer>
but if its the alarm kernel its not useful to me either, as they build without kexec support
<samueldr>
though NuxOS uses systemd, so I guess you're right in saying "bastard [...] distro" I guess :)
<samueldr>
you're free to change the configuration
<hanetzer>
but yeah. the idea is a gentoo binary dist cooked in a manner similar to chromeos, with nixos in writable home dir :)
camsbury has quit [Remote host closed the connection]
<samueldr>
in any ways, that repository from thefloweringash may be helpful as it declaratively describes some of the processes to make a bootable image, whether you want to use it or not, it's another person's notes :)
<hanetzer>
perhaps. does nixos cross-build good?
<clever>
hanetzer: i have had some success in cross-compiling nixos to my rpi3, but a surprising number of #!'s reference the x86 bash, yet somehow still work
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #nixos
slack1256 has joined #nixos
cosimone has quit [Quit: Quit.]
cr4y1 has quit [Ping timeout: 260 seconds]
<unclechu>
hi. about nix-shell scripts. i pick specific nixpkgs version like this: `-I nixpkgs=https://github.com/NixOS/nixpkgs/archive/5272327b81ed355bbed5659b8d303cf2979b6953.tar.gz`
<unclechu>
but this commit in the URL doesn't have anything to do with actual hashsum validation right?
xkapastel has quit [Quit: Connection closed for inactivity]
<unclechu>
so if there's MITM at the github.com i can get attacked, right?
<unclechu>
is there any way to validate sha-sum of that pick in a nix-shell script?
<clever>
unclechu: when used with -I, the https in the url will be respected, so the attacker will need to break https first
lunatera has joined #nixos
<unclechu>
<clever "unclechu: when used with -I, the"> what if someone get into the github servers, without a need of messing with https certs?
<samueldr>
but assuming they do, when used that way the archive is not validated against a hash
DerSaidin has joined #nixos
<clever>
unclechu: yeah, the attacker would need to control a CA or be within github's CDN
<clever>
unclechu: using something like niv, you can easily specify sha256's and then be protected from that
<unclechu>
clever: what is niv and how do i use it in a nix-shell script?
<cole-h>
I'm building the entire universe after bumping my nixos-unstable... RIP.
<unclechu>
clever: are you sure it can be used with nix-shell scripts?
<clever>
unclechu: basically, you just `nix-env -iA nixos.niv` then `niv init` and `niv add nixos/nixpkgs`
<clever>
unclechu: then you can just do `(import ./nix/sources.nix).nixpkgs` to get the path to nixpkgs
<unclechu>
clever: what it has to do with nix-shel lscripts?
<clever>
unclechu: that goes into your shell.nix file
<unclechu>
clever: shell.nix is not a nix-shell script
<clever>
thats exactly what shell.nix is for
<unclechu>
clever: i'm talking about shebang `#! /usr/bin/env nix-shell`
<unclechu>
unclechu: with all the `#! nix-shell -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/5272327b81ed355bbed5659b8d303cf2979b6953.tar.gz`
<unclechu>
clever: if i'm writing a script in a single file shell.nix doesn't exists
<clever>
yeah, those ones cant easily use niv, i tend to avoid them
<clever>
it tends to have worse performance, because it has to download the deps when you try to run it
<unclechu>
clever: performance is not what i care about when i need a single-file script. actually it's not that bad when the packages are already downloaded
<unclechu>
not very different from manually running `nix-shell`
<unclechu>
so i don't see the point about performance here. anyway, performance isn't what i have in my head
<unclechu>
i wan't be able to sometimes use nix-shell scripts
<unclechu>
but i need to avoid any malicious injection due to the tool eats whatever a url feeds it
<unclechu>
simple sha-sum-validation would help, as it works with `fetchGit`
<clever>
it would only work if it accepts -E
<clever>
and it would still get pretty ugly, having to put the entire nix expr on one line
<unclechu>
clever: i put safety here above non-uglyness
<AmandaC>
How hard would it be to generate an options.json which includes my own options as well? Right now I've got `nixosOptions = (nixosFuncUnstable { configuration = ./nixos-common.nix; }).config.system.build.manual.optionsJSON;` in my release.nix -- but afaict that doesn't contain any of my own options?
orivej has quit [Quit: No Ping reply in 180 seconds.]
<cmk_zzz>
does anyone have a configuration sample to setup DNS for an internal network using nixos? I am trying with tinydns but as soon as I enable it the local DNS resolution fails and it can't complete the configuration
gxt has quit [Client Quit]
jasongrossman has joined #nixos
jasongrossman has quit [Remote host closed the connection]
<neonfuz2>
do separate channels all use the same nixpkgs config?
<neonfuz2>
like do they all read .config/nixpkgs/config.nix
<srhb>
neonfuz2: a basic import of nixpkgs, no matter its source, channel or otherwise, has logic in pkgs/top-level/impure.nix which imports config if not explicitly given. In practice this means that all channels uses the same config (as long as the same user is evaluating it)
<srhb>
neonfuz2: For more details, it's better to study that fil to understand the config logic.
<srhb>
Or better yet, not use packageOverrides and config for this purpose, but (system-wide) overlays and forgo multiple channels.
<neonfuz2>
oh cool thanks
evanjs has quit [Read error: Connection reset by peer]
<cidkidnix>
is there a way to use mingw as a dependency for a build without having to use pkgs.crossSystem? as I need to compile wine with "--with-mingw" enabled
AluisioASG has quit [Read error: Connection reset by peer]
AluisioASG has joined #nixos
orivej has quit [Ping timeout: 258 seconds]
orivej has joined #nixos
turlando has joined #nixos
h0m1 has quit [Quit: WeeChat 2.8]
h0m1 has joined #nixos
orivej_ has joined #nixos
<murmr>
Hi, I am trying to create a shell.nix for a C++ project. It uses qt5. Once built the executable crashes with the `Could not find the Qt platform plugin "xcb"` error. Calling wrapQtApp interactively does not seem to fix things. Has anyone else ran into a similar issue?
cidkidnix has quit [Remote host closed the connection]
orivej has quit [Ping timeout: 265 seconds]
reivilibre has joined #nixos
<Guest60204>
where can i find log files for services (e.g ssh, nginx)
lsix has joined #nixos
chimay has joined #nixos
dermetfan has joined #nixos
asheshambasta has joined #nixos
orivej_ has quit [Ping timeout: 240 seconds]
orivej has joined #nixos
zupo has joined #nixos
<srhb>
Guest60204: Depends on the service in question. openssh only logs to the journal, for instance (so journalctl -u sshd)
<srhb>
This is the case for most services, but iirc nginx is particularly troublesome to set up in such a way, and will usually log at least access to somewhere else in /var/log/nginx perhaps -- but that depends on your configuration.
<Guest60204>
i see; does `services` in nix config files always correspond to systemd services?
<srhb>
Guest60204: Almost always, but not always.
<srhb>
Guest60204: (For instance, some of them might just further configure an existing service, say nginx, but not correspond to their own systemd service)
<srhb>
Guest60204: Why the sigh? :) fwiw, all the systemd ones are implemented in terms of systemd.services.${foo} in turn, which _always_ corresponds to a systemd service.
orivej has quit [Ping timeout: 256 seconds]
orivej has joined #nixos
<Guest60204>
i'm trying to run gitea which, despite being under the main services config, is NOT running as a systemd service (or at least can't find it immediately)
<karantan>
is there any way for e.g. haproxy service not to restart when I do nixos-rebuild switch? For example, even if I don't make any change to haproxy config?
<srhb>
karantan: If you don't make any change to it, it should not restart at all.
<srhb>
karantan: There is a way to disable restart _despite_ changes, but I would suggest you find out why you're seeing restarts when you think nothing has changed first.
<Guest60204>
karantan: maybe u edited an indirect dependent?
fendor has joined #nixos
ardumont_ is now known as ardumont
<Guest60204>
srhb: do you know if i can load the entire system-wide nix config into the repl and play around with it?
<Guest60204>
i'm trying to do `nix repl /etc/nixos/configuration.nix` right now but it's telling me i can't auto-call a function with an arg without giving "pkgs" a default value
orivej has quit [Ping timeout: 264 seconds]
eoli3n has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @vcunat pushed commit from @aszlig to release-20.03 « nixos/wireguard: Fix mismatched XML tag »: https://git.io/JJc9I
<{^_^}>
[nixpkgs] @vcunat merged pull request #93493 → nixos/wireguard: Fix mismatched XML tag → https://git.io/JJc2T
<gueorgui>
NobbZ[m]: I'm using nix-env. Does this mean I have to roll back the whole environment, not just the package in question?
m0rph has joined #nixos
<NobbZ[m]>
Yes.
m1sosoup has quit [Read error: Connection reset by peer]
meh` has joined #nixos
morph000 has joined #nixos
<gueorgui>
Thanks!
<NobbZ[m]>
Nix-env doesn't allow for rolling back individual packages. You might be able to simulate something similar by using multiple channels pointing to different versions of nixpkgs where you install your software from.
orivej has quit [Ping timeout: 246 seconds]
orivej has joined #nixos
<NobbZ[m]>
Though the power of nix is in its declarative package management.
m0rph has quit [Ping timeout: 258 seconds]
<gueorgui>
Yep, I get it :) I managed to use nix-env --rollback to get where I wanted
<gueorgui>
Wish there was a summary of changes between each generation though.
<{^_^}>
[nixpkgs] @c0deaddict opened pull request #93522 → erlangR22 and R23: upgrade wxGTK to 3.1 → https://git.io/JJcNW
knupfer has joined #nixos
knupfer has joined #nixos
<gthm_>
gueorgui: I think you need version control software, nix env tagging by commit IDs / commit tagging on nix updates, and discipline to get that
<gthm_>
like, I have a wrapper around nixos-rebuild switch that checks that my changes are committed and does that tagging
zupo has joined #nixos
<gthm_>
which I setup after I nuked /etc/nixos for the first time :)
orivej has quit [Quit: No Ping reply in 180 seconds.]
rardiol has quit [Read error: Connection reset by peer]
orivej has joined #nixos
rardiol has joined #nixos
nixuser has joined #nixos
<karantan>
which nix package contains `nixos-rebuild` command? I need to put it in Path when defining a custom "systemd.services"
lhurttila has joined #nixos
<karantan>
otherwise I get "/bin/sh: nixos-rebuild: command not found"
<manveru>
karantan: it's in `config.system.build.nixos-rebuild` if i remember right
jjakob_ has joined #nixos
<manveru>
of course you only get that when you have a nixos config...
<Taneb>
karantan: can you use /run/current-system/sw/bin/nixos-rebuild? I think nixos-rebuild is special
jjakob has quit [Read error: Connection reset by peer]
<karantan>
cool, looks like it works. now I just need to add nixpkgs/nixos to env
<gueorgui>
gthm_: That sounds at the same time super cool and like I don't really have time to think about how to set it up properly :D
delan has quit [Ping timeout: 258 seconds]
ArdaXi has quit [Quit: WeeChat 2.8]
<Aleksejs>
hello, has anyone tried to add a IKEv2 type of VPN connection? I googled it and couldn't find a NM addon for that. I saw that there is a way to add the VPN of this type via .nix options. Is this the only way?
orivej has quit [Read error: Connection reset by peer]
orivej_ has joined #nixos
heinrich5991_ is now known as heinrich5991
<fps>
hmm, just adding meson as nativeBuildInput seems to magically make use of it
<fps>
how can i control arguments passed to meson though? something like --verbose?
<gthm_>
gueorgui: no problem. I blatantly stole most of that from some forum post :)
sangoma has joined #nixos
lhurttila has quit [Ping timeout: 245 seconds]
sangoma has quit [Read error: Connection reset by peer]
c_wraith has joined #nixos
knupfer1 has joined #nixos
knupfer has quit [Quit: knupfer]
knupfer1 is now known as knupfer
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zjgkkn has joined #nixos
noudle has quit []
zupo has joined #nixos
onny[m] has joined #nixos
<onny[m]>
Hey, I searched yesterday for some hours but wasn't able to find a good example. I want to modify an existing wrapper, for example for the program "atom" or "signal-desktop" and want to include my own environment variables into the existing wrapper. I guess this is possible with derivations or overlays but I wasn't able to figure out how to include it into my configuration.nix :,(
CMCDragonkai1 has quit [Ping timeout: 246 seconds]
freezeboy has joined #nixos
stree has quit [Excess Flood]
stree has joined #nixos
c0c0 has quit [Quit: WeeChat 2.6]
gxt has quit [Ping timeout: 240 seconds]
knupfer has quit [Ping timeout: 256 seconds]
cosimone has joined #nixos
cosimone has quit [Client Quit]
<chiiba>
I'm using k3s from nixos-unstable (dc80d7bc4a). CoreDNS deployment isn't working. As per logs it seems that CoreDNS is trying to reach Kubernetes API at https://192.168.128.1:443, but judging by `route -n` there
<chiiba>
is no such route. It's probably going to the default gateway (i.e. totally useless). Am I missing something obvious?
cosimone has joined #nixos
freezeboy has quit [Remote host closed the connection]
<chiiba>
Ah, nvm, I'm seeing a lot of clues in k3s logs, will try to fix those first.
orivej_ has quit [Ping timeout: 256 seconds]
orivej has joined #nixos
a-h has joined #nixos
asymptotically has quit [Ping timeout: 240 seconds]
zupo has joined #nixos
asymptotically has joined #nixos
<symphorien>
buildRustPackage is not running checkPhase any more ? The logs of nix-du do not show that tests ran, while it has doCheck = true
<{^_^}>
[nixpkgs] @zimbatm pushed commit from @max-wittig to master « gitlab-runner: 13.1.0 -> 13.2.0 (#93514) »: https://git.io/JJcpi
zupo has quit [Ping timeout: 264 seconds]
orivej has quit [Ping timeout: 265 seconds]
orivej_ has joined #nixos
a-h has quit [Remote host closed the connection]
reivilibre has quit [Ping timeout: 256 seconds]
alp has joined #nixos
cosimone has quit [Quit: Quit.]
est31 has quit [Quit: No Ping reply in 180 seconds.]
stigo is now known as sgo
est31 has joined #nixos
freezeboy has joined #nixos
kreyren has quit [Ping timeout: 240 seconds]
<freezeboy>
how is it possible to help reducing the lag in PR merging ? most PRs are just sha256 updates, should we simply check the validity and post a comment ? or more work is required ? The "approved" flag doesn't seem to be useful also
sgo is now known as stigo
<bqv>
Whatever do you mean? PRs are merged very quickly. Any allegations of a broken process are fake news and we will sue you for slander if you continue to parade them.
arianvp_ is now known as arianvp
<freezeboy>
Oh didn't know Donald was a nixos contributor
<gchristensen>
bqv: that isn't appropriate or needed
sangoma has quit [Read error: Connection reset by peer]
zupo has joined #nixos
<{^_^}>
[nixpkgs] @basvandijk merged pull request #85635 → Support more than 99 nodes in NixOS tests & export toHex and toBase → https://git.io/JfTRI
<gchristensen>
freezeboy: reviewing the change, checking the build, confirming the results -- these are all useful. timokau[m] has been working on https://github.com/timokau/marvin-mk2/blob/master/USAGE.md lately which we're hoping will improve things. that said, it is important to remember we already merge a high number of PRs, merging about 60/day
<{^_^}>
[nixpkgs] @c0deaddict closed pull request #93522 → erlangR22 and R23: upgrade wxGTK to 3.1 → https://git.io/JJcNW
kreyren has joined #nixos
sangoma has joined #nixos
ericsagnes has quit [Ping timeout: 260 seconds]
<freezeboy>
gchristensen thank you, I will read the USAGE file then to get more inputs,and try to give a hand when my computer is idle. I see the commit merged quickly, but as I recently sent some PR I also saw the PR count growing, without knowing how to help no sure if it was a temporary problem due to summer time, or a more general problem as I used to only
<freezeboy>
watch the commit page, not the PR one
<freezeboy>
anyway, thank you all for the merging work, the content available in this distribution is really amazing
freezeboy has quit [Quit: Ping timeout (120 seconds)]
<karantan>
which nix package contains `sudo` command? I need to put it in Path when defining a custom "systemd.services" otherwise I get `/bin/sh: sudo: command not found`
<{^_^}>
Use ,locate <filename> to find packages containing such a file. Powered by nix-index (local installation recommended) https://github.com/bennofs/nix-index
<infinisil>
karantan: nixos-rebuild is available in NixOS as `config.system.build.nixos-rebuild`
<karantan>
great, thanks
<atemu12[m]>
infinisil: Why is that though? The same is true for nixos-install which would be very useful to have on non-NixOS systems to bootstrap and I couldn't for the life of me figure out where it is a while back
rardiol has quit [Ping timeout: 265 seconds]
<infinisil>
atemu12[m]: nixos-rebuild at least depends on the config.nix.package option in NixOS, so it's not a fully independent package
<infinisil>
But that could probably be worked around
<infinisil>
atemu12[m]: You can still get nixos-rebuild with e.g. `nix-build '<nixpkgs/nixos>' --arg configuration '{}' -A config.system.build.nixos-rebuild` though
<{^_^}>
[nixpkgs] @offlinehacker opened pull request #93526 → kata-containers: init at 1.11.2 → https://git.io/JJCfg
<aminechikhaoui>
Hey everybody, do people run javascript end to end tests (e.g using Cypress) on hydra/Nix builds with sandboxing ?
<aminechikhaoui>
I can't think of a clean way to run tests that open up a browser that pulls stuff from CDNs
<aminechikhaoui>
other than having separate entry-point html files that point to local files instead of CDN but just for testing
<{^_^}>
#17616 (by obadz, 3 years ago, closed): Expose nixos-tools (nixos-{generate-config, install, option, rebuild}) into nixpkgs
<adisbladis>
But I'd rather not use CDNs
<aminechikhaoui>
adisbladis yeah thought about that as well, so you mean fetchurl the needed files and serve them with nginx for example ?
<adisbladis>
Yes, exactly
<bennofs[m]>
you could also use some form of record/replay proxy for network requests
<{^_^}>
[nixpkgs] @dasj19 opened pull request #93527 → mod_cspnonce: init at 1.3 → https://git.io/JJCJv
<aminechikhaoui>
"But I'd rather not use CDNs" woah, is there something better nowadays ? :)
<bennofs[m]>
then run the tests once to trace all needed resources from the cdn and run them against a proxy in hydra
<adisbladis>
Serve the files yourself ?
<aminechikhaoui>
yeah I guess we could do that, latency isn't that much of a problem in our case since we deploy in the regions where the client exists when I think of it
<simonpe^^>
I'm setting up a new development environment for my customer (they make embedded software) and I'm using Nix to distribute the developer tools. What would be the best way to distribute the local tools? Adding a channel seems complicated since a very specific URL is expected and I can't just point `nix-channel --add` to the bitbucket repo I'm using for the derivations.
<simonpe^^>
I want it to be as easy as possible, just configure and forget - no manual updating of git repos etc.
<infinisil>
simonpe^^: What local tools?
<simonpe^^>
things like deployment scripts, serial number generators, etc
<infinisil>
If you package everything with Nix, all you need to distribute are the Nix expressions
<simonpe^^>
Yes, but how do I do that while preserving the simplicity of `nix-env -i the-deployment-scripts`
<infinisil>
simonpe^^: This is project-specific I assume?
<simonpe^^>
Also I want to have a default.nix in each project repo to bring up a nix-shell, and the shell should be able to refer to applications distributed internally
<infinisil>
If this is project specific you should avoid nix-env to install dependencies
<infinisil>
And instead use nix-shell to make them available
<infinisil>
And Nix doesn't mind something being internal or not, as long as you can package it with Nix there won't be any problems
<simonpe^^>
okay, that sounds reasonable but how do I share stuff between different repos?
<simonpe^^>
say I have a 3rd party utility that is required in several projects
lawr3nce has joined #nixos
<{^_^}>
[nixpkgs] @offlinehacker opened pull request #93529 → cloud-hypervisor: init at 0.8.0 → https://git.io/JJCU3
<infinisil>
simonpe^^: Could make a shared repo
<infinisil>
Any way to store the shared stuff work
<simonpe^^>
infinisil: the customer have this horrible horrible VPN solution called sonicwall which in some cases requires a socks proxy (when you're not on site). Is there a way to deal with all the fuzz caused by this inconvenience?
<infinisil>
Nix is a way to package stuff
asymptotically has quit [Quit: Leaving]
<infinisil>
You can of course use it to package scripts that do something
<simonpe^^>
I guess I was thinking about some global socks proxy setting or something like that, or just general 'aah, been there done that' advice
jbgi has joined #nixos
<infinisil>
Oh you mean that the Nix expressions should know about the proxy and use that for evaluation?
joem86 has joined #nixos
<simonpe^^>
for example: currently on my machine (NixOS) I run a centos VM with the VPN client, and do ssh dynamic forwarding through the VM to reach Bitbucket. So in each git repo I have `sshCommand = ssh -o ProxyCommand=\"connect -S 127.0.0.1:30000 %h %p\"`.
<simonpe^^>
But most developers will use ubuntu, where I managed to get the vpn software to work properly, they won't need that
<simonpe^^>
so if I distribute shared stuff through a git repo it will have to detect whether it should use the custom sshCommand setting or not
<simonpe^^>
more clearly, the nix-shell of the importing repo will have to know if it should set the sshCommand when fetching the dependencies from bitbucket
<joem86>
Hi everyone. I'm evaluating nixos and have a question. I'm on Fedora now and I need swayw 1.5, but the latest in nixos is sway 1.4. Would I need to write a nixpkg and build it from source?
<infinisil>
Oof
<simonpe^^>
I guess that means: write your own custom fetchgit :P
<{^_^}>
[nixpkgs] @peterhoeg pushed 0 commits to f/mediainfo: https://git.io/JJCIv
lawr3nce has joined #nixos
dozn has joined #nixos
<tskc[m]1>
simonpe^^I expected because there was one in video on NixOS front(Home) page. Then how do I create one?
<simonpe^^>
What are you trying to do?
<manveru>
tskc[m]1: you'll have to copy the one from the video then :)
<manveru>
(it's not actually video, it's just text)
<manveru>
though i'd argue it would be much better with `bat -p`
<simonpe^^>
manveru: you don't have it aliased to cat?
<simonpe^^>
rookie mistake :P
<manveru>
:D
<manveru>
it's not a drop-in replacement for cat
<manveru>
i use both...
<raboof>
I'd like to build openscad from source, which is also already in nixpkgs. when I run `buildPhase` in `nix-shell '<nixpkgs>' -A openscad` I get `no Makefile, doing nothing`, and cmake is not available in the nix-shell. Am I misunderstanding how this should work?
<simonpe^^>
I only use bat, but sometimes I miss cat
<simonpe^^>
like that old friend you're not really compatible with but can't separate from
<{^_^}>
[nixpkgs] @davidak opened pull request #93533 → Add buildFontPackage with test → https://git.io/JJCLz
<dutchie>
`nix-build` looks for default.nix only, `nix-shell` looks for shell.nix then default.nix (if you don't tell them explicitly)
hydridity[m] has joined #nixos
mariatsji has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @ajs124 opened pull request #93534 → python-gyp: update with python3 support + drop python2 dependency in nss build → https://git.io/JJCLS
fling has quit [Ping timeout: 260 seconds]
hoijui has quit [Remote host closed the connection]
maier has quit [Ping timeout: 256 seconds]
<{^_^}>
[nixpkgs] @erictapen opened pull request #93535 → fontpreview: init at 1.0.6 → https://git.io/JJCtI
<symphorien>
why does git say "invalid object" with builtins.fetchGit ?
<raboof>
when I `nix-build '<nixpkgs>' -A openscad --check` I see it does qmake things in 'configure', but when I `nix-shell '<nixpkgs>' -A openscad`, go to the source directory and try to run `configurePhase` I just get `no configure script, doing nothing`. What am I missing?
maier has quit [Ping timeout: 240 seconds]
<simpson>
raboof: Are there other phases?
leftfold[m] is now known as ice7[m]
<raboof>
simpson: patchPhase doesn't appear to do anything. I'm not sure about $prePhases and $preConfigurePhases, there are no environment variables with those names but I'm not sure how that works
karantan has quit [Ping timeout: 240 seconds]
davidv7 has joined #nixos
<simpson>
Hm. Yeah, I'm not seeing any interesting hooks. qmake has a hook which replaces configurePhase, qmakeConfigurePhase, but it is very plain and shouldn't be a problem.
aleph- has joined #nixos
<raboof>
oh! I need to `$configurePhase`, not `configurePhase`
<raboof>
then it correctly seem to run `qmakeConfigurePhase`
davidv7_ has quit [Ping timeout: 256 seconds]
<raboof>
but it seems I also need `buildPhase` and not `$buildPhase`? I'd like to understand this better :D
fling has joined #nixos
<infinisil>
raboof: You can run `genericBuild` which executes all phases like a standard nix-build would
sangoma has quit [Read error: Connection reset by peer]
oida has joined #nixos
sangoma has joined #nixos
<evanjs>
in writing a VNC server service for system for a NixOS system, is ther anything Nix-y that should be handling Xauthority?
<evanjs>
Or rather: is this usually handled in a certain manner on NixOS? ".Xauthority does not exist"
<evanjs>
(from xauth)
<kuramashinobi13[>
hello guys
<kuramashinobi13[>
can you please help me out?
<kuramashinobi13[>
i can't set cursor theme in bspwm
devalot has joined #nixos
<kuramashinobi13[>
i have xsetroot -cursor_name left_ptr in my config file
<kuramashinobi13[>
but it doesn't seem to work
<kuramashinobi13[>
as i only see my cursor in applications like chromium
leotaku_ has quit [Ping timeout: 256 seconds]
<kuramashinobi13[>
or here in element
<simpson>
evanjs: X authority is delegated to the user in control of the login session. You'll have to look at what other distros do for this, to get an idea of what kind of cross-user cooperation you'll need.
<simpson>
Like, literally, go to your logged-in X user and try `xauth` at a terminal. You'll get a prompt. ? for help. `info` and `list` will show you the authorities that you currently wield, and the files that delegate those cookies to you.
erasmas has joined #nixos
<evanjs>
simpson: yeah, looking at the arch wiki rn
<simpson>
The entire design is so that the capability to control the session is delegated from root to the logged-in user, in a way that even root finds hard to tamper with. This is because the GPU historically was both accessed from userspace, and also was only accessible from one controller at a time.
<tpw_rules>
so i'm trying to use nixpkgs-review wip to review a package in which i've included a patch. i wrote it to a file in the same directory as default.nix then included it with patches = [ ./thepatch.patch ]; but the review fails because nixpkgs-review doesn't seem to copy the patch file into the review environment?
orivej has joined #nixos
Rurd2di_ has joined #nixos
Rurd2di has quit [Ping timeout: 256 seconds]
ajs124 has quit [Quit: killed]
mrpi has quit [Quit: killed]
das_j has quit [Quit: killed]
Scriptkiddi has quit [Quit: killed]
<{^_^}>
[nixpkgs] @7c6f434c opened pull request #93540 → fsg: use a build of wxGTK without gstreamer plugins → https://git.io/JJCOy
fmzakari_ has joined #nixos
justanotheruser has quit [Ping timeout: 272 seconds]
das_j has joined #nixos
Scriptkiddi has joined #nixos
mrpi has joined #nixos
ajs124 has joined #nixos
<tpw_rules>
hm it works if i git add the patch (and my changes) then use wip --staged
mrpi has quit [Remote host closed the connection]
ajs124 has quit [Remote host closed the connection]
das_j has quit [Remote host closed the connection]
Scriptkiddi has quit [Remote host closed the connection]
orivej has quit [Ping timeout: 265 seconds]
ajs124 has joined #nixos
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #93541 → ios-deploy: move out of nodePackages → https://git.io/JJCOA
orivej has joined #nixos
civodul has quit [Quit: ERC (IRC client for Emacs 26.3)]
mrpi has joined #nixos
Scriptkiddi has joined #nixos
das_j has joined #nixos
fresheyeball has quit [Quit: WeeChat 2.7.1]
user_0x58 has joined #nixos
<{^_^}>
[nixpkgs] @danieldk merged pull request #93536 → mongodb: use pname instead of name → https://git.io/JJCq3
<simpson>
evanjs: Well, but where is "there"? It's in the logged-in user's homedir.
ddellacosta has joined #nixos
<evanjs>
yeah, that's what I mean
<evanjs>
I figure we're just using the wrong requires/after/etc if our service can't find that on start
sangoma has joined #nixos
philr has quit [Ping timeout: 258 seconds]
<evanjs>
ahhhh
<evanjs>
simpson: we had it set to multi-user :P
<evanjs>
Yeah I figure WantedBy to graphical-session might fix that haha
byrn has quit [Quit: byrn]
<simpson>
Aha.
orivej has quit [Ping timeout: 258 seconds]
byrn has joined #nixos
byrn has quit [Client Quit]
cjpbirkbeck has joined #nixos
orivej has joined #nixos
elux has joined #nixos
<elux>
hi there
<elux>
im trying to setup wifi on my desktop, but 'nmcli' says my wlp7s0 card (iwlwifi) is "sw disabled". im checking all my nix configs and not sure why its disabled
<elux>
lsmod | grep wifi shows that the module has been loaded too
<simpson>
Check kernel logs (dmesg) and see if your driver has something preventing it from going. Common culprits are RF kill switches (including software-driven ones, check rfkill) and missing firmware.
<simpson>
rfkill can undo soft blocks; `rfkill unblock wlp7s0` or so. You'll likely need root. Be sure to follow up by examining your kernel logs and seeing what may have triggered that switch.
proofofkeags has quit [Remote host closed the connection]
dbmikus has joined #nixos
proofofkeags has joined #nixos
<elux>
hmm.. what a pain / strange thing to get wifi working
<elux>
ive unblocked on rfkill the wlan, but "nmcli" is still saying its "sw disabled". i tried to rmmod iwlwifi and modeprobe it again,
<elux>
looking at dmesg, it loads the card, it does return the error... "Direct firmware load for iwl-debug-yoyo.bin failed with error -2", but im not sure if that is a critical error?
<{^_^}>
[nixpkgs] @7c6f434c merged pull request #93540 → fsg: use a build of wxGTK without gstreamer plugins → https://git.io/JJCOy
<neonfuz2>
okay thanks. I might bump this package, was thinking of removing the comment
<neonfuz2>
citra doesn't run on my computer for some reason though so I'll have to test on another machine
asheshambasta has quit [Ping timeout: 272 seconds]
<neonfuz2>
why didn't they enable fetchSubmodules I wonder
<hexa->
good question
xd1le has joined #nixos
<hexa->
the legacy mount option was the way to go
<{^_^}>
[nixpkgs] @ajs124 opened pull request #93545 → gnupg20: drop → https://git.io/JJCn0
<elux>
i figured out my desktop issue. i need to rebuild with hardware.enableRedistributableFirmware .. but since i dont have an internet connection, 'sudo nixos-rebuild switch' is complaining, anyway to run that without internet?
vykook has joined #nixos
drakonis1 has quit [Ping timeout: 260 seconds]
morph000 has quit [Ping timeout: 264 seconds]
<simpson>
Usually, no; in this particular case, the firmware that you want enabled may simply not be present on your disk.
ixxie has quit [Ping timeout: 256 seconds]
<o1lo01ol1o>
Are there any recent materials for getting nixos working on a recent macbook pro
<o1lo01ol1o>
?
knupfer has quit [Quit: knupfer]
knupfer has joined #nixos
knupfer has quit [Client Quit]
knupfer has joined #nixos
drakonis has joined #nixos
chrmb has quit [Read error: Connection reset by peer]
<o1lo01ol1o>
nickeraqua: that's for `nix` and not `nixos`? I'm looking to use nixos on mbp hardware
simonpe^^ has quit [Remote host closed the connection]
<lordcirth>
o1lo01ol1o, have you tried just installing it yet?
<notgne2>
o1lo01ol1o (chat.freenode.net): I'm not familiar with macbooks, but there is an entry under https://github.com/NixOS/nixos-hardware so I presume importing that should be all that is necessary to make stuff work
<o1lo01ol1o>
lordcirth: yes, I know nix works on (mojave) darwin.
zaeph has joined #nixos
<{^_^}>
[nixpkgs] @timstott opened pull request #93546 → lokalise2: init at 2.5.1 → https://git.io/JJCcz
<o1lo01ol1o>
notgne2: Apple MacBook Pro 12,1 is from early 2015; there's some google results to suggest the keyboard and touchpads require special drivers.
<o1lo01ol1o>
*recent mbps
hayman has joined #nixos
<lordcirth>
o1lo01ol1o, no, I mean, have you tried just running the nixos installer and seeing what happens?
<energizer>
i have a project with a default.nix and i want to use nix-bundle (https://github.com/matthewbauer/nix-bundle) to give my project a static executable. but nix-bundle seems to want a whole nixpkgs, not just an expression or path/to/default.nix. how can i satisfy it?
johnny101 has quit [Quit: Konversation terminated!]
<archH>
Anyone have experience with setting experimental settings with gsettings (gnome)? I want to enable fractional scaling, the command for it is:
<archH>
$ gsettings set org.gnome.mutter experimental-features "['scale-monitor-framebuffer']", but I kinda want to do it the nix way
<neonfuz2>
can someone explain why nix-env -u --leq rebuilds a bunch of packages?
<archH>
Im a bit too new to nixos to fully understand how to use that information to reach my goal. Does this look correct? https://kopy.io/Uil7l
<srhb>
neonfuz2: Not without more information. Presumably, something changed (eg. you updated nixpkgs)
<srhb>
neonfuz2: Assuming you mean that rebuild is performing actual building.
vykook has quit [Ping timeout: 246 seconds]
<lordcirth>
archh, I'm no expert either, but it looks reasonable to me. Try it and see?
<neonfuz2>
so, I have many packages installed in my nix-env. I run nix-env -u, no updates. I run nix-env -u --leq and several packages "update" (no version string change) and it compiles them
<neonfuz2>
I'm guessing something like a hash is different for some reason than the hydra package?
<lordcirth>
neonfuz2, why do --leq?
<srhb>
neonfuz2: The default (--lt) actually does a version comparison iirc. That is, there has to be a numerical, positive difference between the new version and the old version
<energizer>
how do i fetch a derivation from github that's called something other than default.nix?
<neonfuz2>
I'm just curious
<srhb>
neonfuz2: When --leq you lift that requirement and the package will "update" even if the numerical version is the same.
archH has quit [Quit: Connection closed]
<srhb>
neonfuz2: Whether that leads to an actual build (or new substitution) depnds on other changes in the dependency graph
stree has quit [Excess Flood]
<infinisil>
energizer: Is this still about nix-bundle?
<srhb>
neonfuz2: Yes, though that's probably not the question you really mean to ask
archH has joined #nixos
<energizer>
infinisil: thanks
<srhb>
neonfuz2: hydra builds (essentially, leaving out some details) all packages in nixpkgs for each commit of nixpkgs it is asked to build, and saves them (forever-ish) in cache.nixos.org
<neonfuz2>
I understand that
<srhb>
neonfuz2: So if I upgrade package bar, which is a dependency of package foo, but I don't change the version of foo, foo still gets a new hash
<neonfuz2>
its just interesting to me, I have 19 packages upgrading without a version string bump
<srhb>
Because its dependencies changed.
vykook has joined #nixos
<neonfuz2>
srhb: so each time this happens, does hydra rebuild that package again?
<neonfuz2>
like, rebuild foo
<srhb>
It's essentially an entirely "new" "package" from that perspective, despite its "version" being the same.
simba2 has joined #nixos
<srhb>
neonfuz2: Yes.
<neonfuz2>
so, my system has 19 packages upgrading with --leq, does this mean I have 19 packages with "outdated dependencies"?
<srhb>
Likely, yes.
<neonfuz2>
possibly
<neonfuz2>
so is it common for old libraries to just stick around? If I updated my system completely with --leq would that save space after a garbage collect?
<srhb>
neonfuz2: Only if you have no more gc roots pointing to the old paths
<srhb>
Like, old profiles.
<neonfuz2>
yeah I really mean like, the new profile will be smaller
<neonfuz2>
I understand the GC
<srhb>
Oh, I understand what you mean. Because it's not pointing to "multiple old dependency trees" ?
<neonfuz2>
it just makes sense to me that updating many programs to the latest libraries might save space
<energizer>
infinisil: fwiw `nix-bundle my.nix /bin/mybin` made an executable but apparently is missing the dependencies specified by poetry/poetry2nix as i'm getting ImportErrors on startup
rajivr has quit [Quit: Connection closed for inactivity]
<neonfuz2>
hardlinking aside too
<srhb>
neonfuz2: You can check this experimentally by querying the total closure path before and after
<infinisil>
energizer: Maybe you could make a minimal reproducible example and file it as an issue
<neonfuz2>
yeah
<energizer>
infinisil: yep
<neonfuz2>
srhb: cool thanks
<srhb>
neonfuz2: nix path-info -Sh /nix/var/nix/profiles/per-user/sarah/profile-266-link vs some other generation
<evanjs>
holy crap prometheus is nice. Now I just need to get exporters set up in a reasonable manner across devices...
<srhb>
neonfuz2: That will traverse the entire closure of said profile and compute its size
stree has quit [Excess Flood]
<neonfuz2>
sweet
stree has joined #nixos
vykook has quit [Ping timeout: 256 seconds]
EsperLily has joined #nixos
<neonfuz2>
also sometimes builds fail with nix-env -u --leq for me
<srhb>
neonfuz2: Presumably that's either impure or broken on Hydra too, for whatever commit of nixpkgs you're on
<thomassgn>
how does nixos deal with mdadm raids? I can't see it mentioned in the manual
<srhb>
thomassgn: iirc you'd just partition as normally and set it up during install, generate a hardware conf and it should pick up the raid
<srhb>
thomassgn: There's no automatic provisioning, if that's what you're asking
<srhb>
(Same goes for luks, zfs, ...)
<neonfuz2>
does hydra serve the latest non broken package?
<thomassgn>
oh, no. I have a raid setup, but how do I let nixos know about it in config?
<srhb>
neonfuz2: No, that's not quite how it works
<srhb>
neonfuz2: A hash is computed from your version of nixpkgs
<srhb>
neonfuz2: That hash may or may not exist. If it's a broken build, it does not exist.
orivej has quit [Ping timeout: 256 seconds]
<srhb>
neonfuz2: Then, your system tries to build it. If the build is pure, it also fails.
orivej has joined #nixos
<simpson>
thomassgn: To build on what srhb is saying, what I had to do last time was take the auto-detected md device and put its /dev/disk/by-uuid/ entry into my fileSystems configuration.
<srhb>
simpson: Oh, that's an annoying caveat. :/ Thanks.
<neonfuz2>
okay
seku has joined #nixos
<srhb>
neonfuz2: (Or to put it a different way. Yes, Hydra does, but your system doesn't know how to ask for it.)
stree has quit [Remote host closed the connection]
<thomassgn>
mm, ok thanks
<simpson>
The RAID array itself should be auto-detected, if you're using modern disk headers. If not, then yeah, we *don't* have tooling for assembling md by hand. You'll have to write a systemd unit which does the assembly by invoking mdadm. Yes, lots of work; but no, this shouldn't happen unless your array is ancient or weird.
<simpson>
srhb: It's not that big of a deal. Honestly it might be the bare minimum, given that it's only like four lines of config. I *think* that, if one mounts the array at the desired spot, then nixos-generate-config can do it, but not sure.
stree has joined #nixos
<srhb>
simpson: I would have hoped so, but better safe than sorry for sure. :)
<archH>
For fractional scaling (Nvidia Gnome3) this didn't do the trick, however this didnt work either, am I missing something? $ gsettings set org.gnome.mutter experimental-features "['scale-monitor-framebuffer']"
<cransom>
that it only creates linux binaries? unless you can cross compile. or have said person install nix and see if they can build your derivation.
<clever>
energizer: nix-bundle also relies on linux namespacing, so it wouldnt work on a mac
pamplemousse has quit [Ping timeout: 258 seconds]
stree has joined #nixos
vykook has joined #nixos
proofofkeags has quit [Ping timeout: 258 seconds]
<numkem>
I'm trying to get awscli2 and credstash to play ball and it seems like when I have both added to a nix-shell, credstash overrides botocore for awscli2 which just crashes awscli2... I checked the wrapper scripts and nothing jumps out as doing it. What else could I do?
<energizer>
cool "What I have so far is a working proof of concept which shows that is possible for nix-bundle to produce a runnable bundle for a different target architecture" https://github.com/matthewbauer/nix-bundle/issues/45
<{^_^}>
[nixpkgs] @worldofpeace merged pull request #89758 → electron: mark insecure versions (3,4,5) and require explicit version of electron in packages → https://git.io/JfMNf
<{^_^}>
[nixpkgs] @worldofpeace pushed 3 commits to master: https://git.io/JJC8f
knupfer has quit [Remote host closed the connection]
knupfer has joined #nixos
orivej_ has joined #nixos
vykook has joined #nixos
orivej has quit [Ping timeout: 258 seconds]
pamplemousse has joined #nixos
<{^_^}>
[nixpkgs] @vcunat pushed 5 commits to staging-next: https://git.io/JJC8Y
alp has quit [Ping timeout: 260 seconds]
archH has quit [Quit: Connection closed]
zaeph has quit [Ping timeout: 260 seconds]
vykook has quit [Ping timeout: 240 seconds]
archH has joined #nixos
mvnetbiz_ has quit [Read error: Connection reset by peer]
mvnetbiz_8 has joined #nixos
zaeph has joined #nixos
vykook has joined #nixos
archH has quit [Client Quit]
<{^_^}>
[nixpkgs] @tpwrules opened pull request #93552 → librealsense: add pyrealsense2 Python bindings + command line tools → https://git.io/JJC4s
<Rian[m]>
I noticed that when I do nixos-rebuild switch manually, afterwards my wireguard link doesn't work any more. restarting the wg0 interface fixes that
elux has quit [Ping timeout: 264 seconds]
vykook has quit [Ping timeout: 258 seconds]
knupfer has quit [Ping timeout: 260 seconds]
<clever>
Rian[m]: it would be better to fix the service definition, so it restarts correctly
jbgi has joined #nixos
<Rian[m]>
which service definition? that of the wireguard interface?
pamplemousse has joined #nixos
<clever>
Rian[m]: yeah
lsix has joined #nixos
multun has quit [Ping timeout: 258 seconds]
vykook has joined #nixos
pamplemousse has quit [Ping timeout: 246 seconds]
<yurb>
`nixos-rebulid switch --upgrade` seems to want to download a lot of packages, and it looks like they have different hashes, but some have the same versions. Does it mean there was a massive rebuild recently (for 20.03)?
<Rian[m]>
<clever "Rian: yeah"> so you mean I should fix it in nixpkgs?
<clever>
Rian[m]: yeah
<clever>
yurb: if you use nix-diff, you should be able to identify the root cause
<yurb>
haven't used it yet. I'm curious though - can this somehow happen due to me doing `nix-env --delete-generations +5 --profile /nix/var/nix/profiles/system` recently?
<yurb>
clever: sorry, brief loss of network messed up the order of messages for me. Thanks, good to know I didn't mess it up *that way*. The upgrade just finished. Hope I didn't mess it up some other way:)
<numkem>
is there a way to isolate python packages so their dependencies ton interfere with each other?
jbgi has quit [Ping timeout: 256 seconds]
vykook has quit [Ping timeout: 256 seconds]
secretspecter has joined #nixos
<simpson>
numkem: Fundamentally, not really; Python processes have a single global module table, and doing things the approved way means not being able to load multiple modules/packages with the same name.
pamplemo1sse has quit [Ping timeout: 256 seconds]
<simpson>
(There is PEP 302, and there exist libraries like Exocet which can do PEP 302 loading in an isolated manner. However, it's provably impossible to do this in a correct way, and loaded modules need to cooperate.)
<numkem>
simpson: I have to use awscli2 and credstash and botocore is in 2 seperate version. Meaning awscli2 is broken the second credstash in added to a nix-shell
pamplemousse has joined #nixos
jbgi has joined #nixos
<numkem>
simpson: I was under the impression that nix would isolate each cli app with it's own deps
<qyliss>
It should
orivej has quit [Quit: No Ping reply in 180 seconds.]
<secretspecter>
I'm trying to update the protonmail-bridge nixpkg to the latest version 1.2.7, however I've run into a problem where that version wants Qt 5.13 and my current config installs Qt 5.12. How can I install the correct version of Qt?
orivej has joined #nixos
<cole-h>
Pretty sure it's just credstash that's broken. `nix-shell -p python3Packages.credstash` fails because of a botocore version restraint.
<numkem>
so doing `nix-shell -I nixpkgs=/home/numkem/src/nixpkgs/ -p credstash awscli2` with that fix being in my nixpkgs fork breaks awscli2 because it's using credstash' botocore
<numkem>
I'll try changing the constraint on credstash and see what happens
proofofkeags has quit [Ping timeout: 260 seconds]
acowley has joined #nixos
<numkem>
cole-h: credstash depends on boto3 which depends on botocore which has a restriction for < 1.18.0
vykook has quit [Ping timeout: 265 seconds]
<Rian[m]>
<Rian[m] "so you mean I should fix it in n"> for now, I just defined another service that wants nixos-upgrade 🤷♂️ might look into it in the coming days though
<secretspecter>
I just wanna know how to get Qt 5.13 or higher :3
<cole-h>
You can't. The PR bumping to 5.14 is still WIP.
meh` has quit [Quit: I don't want to live on this planet anymore.]
<cole-h>
fmzakari_: Most of us want the packages in Nixpkgs to be high-quality, which is why some PRs go through extensive rounds of discussion/suggestion. What kind of things would you expect to be automated?
vykook has joined #nixos
<cole-h>
However, I don't agree that some things should be "fixed after" -- if the problem is apparent, and has a solution, why wait?
<fmzakari_>
I dunno mostly just venting --
<cole-h>
I sympathize. Others have also expressed displeasure at the high amount of back-and-forth some PRs can spawn, or how they're not quick enough to be merged.
<fmzakari_>
I'm writing a little article on Java + Nix specifically Maven; it's a pretty fragmented ecosystem. Kind of surprised given how prevalent Java is. I thought it would have had more first-class solution
<eacameron>
fmzakari_: We wrote a tool called "haven" that helps generate stuff for maven deps.
<fmzakari_>
(it uses the underlying Maven libraries rather than calling out to Maven to determine the repository + closure)
<Graypup_>
I am having a really strange docker build failure with nix's docker support
<Graypup_>
tar: temp/layer.tar: Cannot write: No space left on device
<Graypup_>
so I went looking at all my devices and they have 10+GB of space available
jbgi has quit [Ping timeout: 240 seconds]
<fmzakari_>
@Graypup I remember typing `docker info` to see available space
<Graypup_>
however! I looked at the output from the runAsRoot part, which appears to say that it is creating some sort of filesystem with 65536 inodes, which seems uh low
codygman has quit [Ping timeout: 240 seconds]
codygman has joined #nixos
<M0-[m]>
How does one use overrideAttrs to override an attribute like extraPkgs or targetPkgs which seem to be functions?
orivej has quit [Read error: Connection reset by peer]
vykook has joined #nixos
<Thra11>
alias what_the_hell_did_i_do="nix-diff $(nix-store -qd $(readlink -f /nix/var/nix/profiles/system)) $(nix-instantiate '<nixpkgs/nixos>' -A system)" :p
<cole-h>
fmzakari_++ eacameron++ A while ago, I wanted to build a Java project that used maven. It was a huge PITA, and I gave up.
<{^_^}>
fmzakari_'s karma got increased to 1
<{^_^}>
eacameron's karma got increased to 2
orivej has joined #nixos
<Graypup_>
M0-[m], make another function that wraps the existing function ^^
<fmzakari_>
cole-h I'll send you what i write; i'm documenting the "double maven process" solution. It works but it's a bit non-Nix since it doesn't take advantage of sharing the same dependencies across the store
drakonis has quit [Ping timeout: 260 seconds]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<fmzakari_>
I'm also looking to collaborate on https://github.com/fzakaria/mvn2nix ; it was annoyingly challenging to find useful Java libraries to use the underlying Maven libraries to calculate the transitive closure + resolve artifacts to the remote repositories
kenzie[m] has left #nixos ["User left"]
<M0-[m]>
Graypup_, any examples? I'm not sure as to how to accomplish that
<Graypup_>
Here's the output from my weird build failure. I'd try it on another computer but I am fresh out of fast computers :( my desktop runs Windows and can't do the nested virt to run the runAsRoot part, and my laptop is having the problem. https://gist.github.com/lf-/3965e2fee0ccec77917432ce406b24df
vykook has quit [Ping timeout: 258 seconds]
xd1le has quit [Quit: Quit]
emmanuel` has joined #nixos
drakonis has joined #nixos
<Graypup_>
M0-[m], it took me a second to find the source or docs for overrideAttrs: overrideAttrs = f: mkDerivation (attrs // (f attrs));
<Graypup_>
so I think you'd want something like `old: { extraPkgs = args: old.extraPkgs args ++ [abcde]; }`
<Graypup_>
assuming I have the types right
<Graypup_>
might also need some parens
vykook has joined #nixos
zupo has joined #nixos
fendor_ has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @flokli merged pull request #93073 → nixos/systemd: Implement a packages option for tmpfiles → https://git.io/JJOhS
<M0-[m]>
man, that seems like it would work but I get `error: cannot coerce a function to a string`
<samueldr>
is there a trick to use `nix log` for whichever derivation failed in the last call to nix-build? I'm kinda annoyed at having to find the derivation name in the log, copy/paste it
shibboleth has quit [Quit: shibboleth]
<Graypup_>
*sob* that kind of error makes me so sad when it happens :((((( so hard to debug
<Graypup_>
samueldr, I wonder if there's a log of built derivations that perhaps you could grep for fails?
<samueldr>
the most likely culprit for "cannot coerce a function to ____" is using a function in a list without using the proper parens
dbmikus has quit [Ping timeout: 240 seconds]
<Graypup_>
yup that would do it!
<samueldr>
[ fn arg ]; <- a list with a function, and an arg
<{^_^}>
[nixpkgs] @twhitehead opened pull request #93560 → Purity checking should accept $TMP and not just /tmp → https://git.io/JJCuw
<Graypup_>
I have done that to myself before lol
<samueldr>
[ (fn arg) ]; <- a list with the result of that function with arg
<Graypup_>
samueldr, I wonder if nix path-info --all could be of use
<Graypup_>
jk nope that's a lot of stuff
<Graypup_>
oh geez wow, the way the cachix github action finds what was just built was dumping all the paths in the nix store before and after the build and giving them to sort | uniq -u
alp has quit [Ping timeout: 240 seconds]
<samueldr>
one issue is that there is no concept of "previous failed build", or even "ordering" of builds, and definitely not "for this shell"
seku has quit [Quit: Connection closed]
<samueldr>
I'm thinking the solution would be to add a flag, --error-log-on-fail or something along the line, which would use $PAGER for the error that aborted the build
<samueldr>
I guess it could also be useful for non-console uses, in automated pipelines, $PAGER could be set to e.g. `gist -p` and send that out
<M0-[m]>
does irc handle multiline code blocks well? I want to share something that's around 10 lines but I'm not sure how it looks on the irc end
<M0-[m]>
The commented part works, but I'm trying to also modify nativeBuildInputs, but I can't figure out what function is being coerced
<samueldr>
just make sure to send your question / comment about the issue separately or else matrix will bundle it into the same link, and people *may* not look at long messages
civodul has quit [Quit: ERC (IRC client for Emacs 26.3)]
domogled has quit [Quit: domogled]
<samueldr>
(oldAttrs.extraPkgs pkgs) # M0-[m]
<samueldr>
(I think)
<M0-[m]>
unfortunately that didn't work
<clever>
i dont think anything overrideAttrs acts on, can be a function
<samueldr>
.override changes what is dependency injected into the "package call" (callPackage)
orivej has joined #nixos
<samueldr>
.overrideAttrs changes attributes passed to stdenv.mkDerivation
<samueldr>
(I know it took me a while to grok the difference between the two)
dbmikus has joined #nixos
hplar has joined #nixos
<Graypup_>
I have some code that does callPackage <nixpkgs/pkgs/build-support/docker> { ... } and I'd like to make it point to my pinned nixpkgs. How would I do that?
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #nixos
cosimone_ has joined #nixos
<{^_^}>
[nixpkgs] @edef1c merged pull request #93451 → nixos/users-groups: don't consider a system with Google OS Login inaccessible → https://git.io/JJnS6