<notgne2>
that sounds about the same as what I'm doing, I'm grabbing systemd service definitions from the evaled config then turning them into bash scripts, only difference is at the end I just filter out all the services which aren't oneshots (or the final service in the list)
<clever>
notgne2: i now have a nix expression, that generates a docker image, that runs postgresql, prometheus, grafana, nginx, oauth2_proxy, prometheus-postgres-exporter, and 2 custom services
<clever>
notgne2: and nearly all of those are just running the existing nixos modules, that run the same services on nixos
<simpson>
red[m]1: Nah, we're mostly a downstream. If we wrote our own init system from scratch, then maybe, but let's not do that.
chrishill has quit [Remote host closed the connection]
<clever>
notgne2: ah, have you seen the .runner attr?
judson__ has quit [Remote host closed the connection]
shabius_ has joined #nixos
<clever>
notgne2: but i ran into 2 bugs when using it, 1st: it doesnt support .User and .PermissionStartOnly, 2nd: it just does $var = "${foo}"; and if your ExecStart contains a ", all hell breaks loose :P
<clever>
notgne2: oh yeah, and it turns out, dropping root is a lot harder then it needs to be, lol
lucasvo has joined #nixos
<clever>
notgne2: su will reset PATH, even if you tell it not to
<clever>
sudo will reset PATH, even if you tell it not to
<notgne2>
oh I haven't even gotten to that yet, but I used gosu for user-specified entrypoints in continix, so maybe that will work here too?
<clever>
sudo needs valid pam config files, or it wont do anything
<clever>
the pam in nixpkgs lacks @include support, debian patched pam to have @include support, so the sudo in nixpkgs doesnt work in a docker based on ubuntu!!
<clever>
i had to write my own deroot binary, then i noticed 3 setuid binaries already in nixpkgs (nix-index :D), and runit already has its own util for just that
qubasa_ has joined #nixos
jluttine has quit [Ping timeout: 245 seconds]
<red[m]1>
@simpson if you wanted nix with a custom init there's always guix
qubasa has quit [Ping timeout: 240 seconds]
<red[m]1>
honestly though, 95% of my complaints arent arou d
<red[m]1>
the init function anyways - but the other projects and dependancies it ate
<simpson>
red[m]1: No, I want the tooling that comes after Nix, and I'd like a kernel less like Linux and more like seL4. While we're at it, I'd like for the linguafranca to be less like C and more like Smalltalk, and I'd like the Internet to be less like IP and more like NDN.
lucasvo has quit [Ping timeout: 245 seconds]
<clever>
simpson: seen my haskell-init stuff?
<gchristensen>
<3 simpson
<{^_^}>
simpson's karma got increased to 12
<notgne2>
clever: mine takes a single specified service by name (if I decide to include runit or something I may let it take a list), and it only itterates for the purpose of finding oneshots which want to run before the specified service
<clever>
notgne2: ahh, so your dealing with dependencies in the code, rather then requiring the user to do that
<red[m]1>
damn simpson - marry me - lol
<simpson>
clever: Yeah. I'm excited, but also not holding my breath. My current plan is to get Monte so self-hosting that I can replace not just my usage of Python, but all the Python. And then all the Ruby... Soon, all will be Monte~
<clever>
notgne2: currently, i dont have dependencies handled at all, several services fail to start because postgresql hasnt ran yet, but they just auto-restart until things work
<simpson>
clever: I really would like to see a plurality of init systems for NixOS. I also want Monte to be a good PID 1 candidate, because containers and such.
<red[m]1>
seL4 though - i have a new reference
<{^_^}>
[nixpkgs] @jglukasik opened pull request #70143 → perlPackages.Tirex: init at 0.6.1 → https://git.io/JenuQ
<simpson>
l33_: A language I work on. It is related to E.
<red[m]1>
i do have a smalltalk story you might enjoy
<clever>
simpson: i think if you want sel4 under nix, you would need to first look into getting the kernel to build with nix, and then booting it under qemu
<clever>
simpson: once a basic userland is up, you would need to port nix itself to that kernel, so it can build itself
<simpson>
clever: Ha, I hadn't even thought about that! That might be interesting. I hear that typical Linux can be hosted in seL4, so maybe it wouldn't be too bad to have a transitional path where NixOS boots everything within a minimally-configured seL4 context.
<__red__>
simpson: I was doing a penetration test in a live environment and found a smalltalk service running, doing billions of dollars of credit card transactions - ... and it was in a Java shop.
<clever>
simpson: ah, that would be even simpler!
<simpson>
__red__: Rad.
<__red__>
I asked about it, and they said: "We don't have anything but java here - you must be mistaken - where is it? We'll need to re-write it in Java"
<__red__>
Me? "YOu didn't know it existed because in the 10+ years it's been there it's never broken. I'm not telling you where it is - good luck"
<__red__>
:-)
jluttine has joined #nixos
<simpson>
That's like when you see a family accountant business, and all of the grandkids are studious well-trimmed hard-working yuppies, and then the granddad is sitting in his pajamas watching Wheel of Fortune and one-handedly running the whole business~
<__red__>
pretty much
<__red__>
it's still superiour to Java in every way
<l33_>
java, write once, debug everywhere.
<__red__>
and, unlike java it is actually an OO language
<simpson>
clever: Yeah, I'll /join #osdev once I'm at that point. Right now, I'm enjoying learning OCaml and writing more low-level Monte interpreters. How are you feeling about haskell-init, BTW?
<l33_>
haskell scares me..
<__red__>
Why the enterprise "standardized" on a language that was designed to compromise on every single point is beyond me
<l33_>
machine code's where it's at.
<__red__>
I'm auditioning new languages right now for my next language
<simpson>
__red__: Java's history is pretty cool. It was designed to expand upon the performance opportunities developed for Self, Java's most relevant ancestor. Self pioneered JIT concepts, and Java was designed to be JIT'd.
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JenzU
<simpson>
clever: It's funny that Go and Rust ended up like they have. In an alternate universe, the Gophers would have already rewritten coreutils to be statically-linkable Go. Such is life. It does make one wonder: Which parts should be rewritten, and which should simply be discarded entirely?
<__red__>
clever: nerves on elixir has similar issues.
<__red__>
it basically becomes a busybox initrd with elixir as the "init"
<clever>
simpson: i like the idea of keeping things dynamic (so you can share code between binaries), but using patchelf to reduce closure sizes
<__red__>
I wish we could just go "kernel" -> init=/whatever/iex
<clever>
simpson: but if you can jam every binary into a single executable (like busybox), that could be even better (as long as it doesnt harm performance)
<__red__>
but it needs userspace stuff :-(
<clever>
simpson: haskell lets you go one step further, and just use forkIO and pure code, rather then child procs!
jluttine has quit [Ping timeout: 240 seconds]
judson has quit [Ping timeout: 240 seconds]
judson__ has quit [Ping timeout: 245 seconds]
<simpson>
clever: Sure. Haskell, like all MLs, is well-suited to this task. If only the ecosystem could produce the required toolchains and platforms.
Rusty1 has joined #nixos
lucasvo has joined #nixos
logzet_ has quit [Ping timeout: 264 seconds]
<lucasvo>
clever: when I define to run mullvad-daemon as a service, it complains about not being able to find `ip` though the executable `ip` is definitely available. When I just launch it from a regular shell, it does not. How do I give it the right path when executing it as a service?
logzet has joined #nixos
<clever>
lucasvo: services have their own PATH, you have to add things to systemd.services.foo.path = [ pkgs.bar ]; to be able to run them
<notgne2>
clever: well if you want to try to work in dependent services to your code feel free to copy mine, there's already a whole separate function just dedicated to listing all the dependencies of any given service, but for a system with concurrent services it may need to be modified to output in layers rather than a simple list, so some can be launched
<notgne2>
async
<notgne2>
(the ordering may also be backwards in some places, I wrote the code "up-front" since no services I'm working with require it yet, so its untested)
ilyagotfryd has joined #nixos
leo_ has joined #nixos
boxscape has quit [Remote host closed the connection]
doublex has quit [Ping timeout: 245 seconds]
ilyagotfryd has quit [Ping timeout: 265 seconds]
philr has joined #nixos
drakonis_ has quit [Ping timeout: 245 seconds]
selfsymmetric-mu has quit [Remote host closed the connection]
selfsymmetric-mu has joined #nixos
rsoeldner has joined #nixos
doublex has joined #nixos
selfsymmetric-mu has quit [Remote host closed the connection]
matthuszagh has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @worldofpeace opened pull request #70149 → blueman: no optional networkmanager → https://git.io/Jengc
sellout-1 has quit [Read error: Connection reset by peer]
drakonis_ has joined #nixos
<clever>
notgne2: main problem i can forsee, is just making one wait for another, i dont know if runit supports that, so i may need to implement it in bash
sellout- has joined #nixos
drakonis_ has quit [Read error: Connection reset by peer]
<notgne2>
clever: I'm not really familiar with runit, but my idea was to modify my dependencies function to generate a set of layers (00, 01, etc) which will contain all the services that can start at once, then put them into another init system that way. I don't remember "life before SystemD" all the well, but I'm relatively sure existing init systems
<notgne2>
commonly worked using ordered services, before SystemD made them concurrent, something that can instead by replicated in a Nix expression
<clever>
notgne2: init.d didnt really support concurrent startup, and you just had to number everything so starting things in order worked
<clever>
notgne2: it also relied on things just blocking until it was fully started
rsoeldner has quit [Remote host closed the connection]
rsoeldner has joined #nixos
<notgne2>
yeah that must be what I'm thinking of, you could use a Nix expression to generate numbered scripts, which will run multiple services, obviously at the cost of some startup performance as some layer+1 services should already be able to start, but are waiting on another service for another layer+1 service
<clever>
notgne2: another option, is to just create a file when the service is started, then everything just waits in a while loop, until said file exists
zacts has quit [Quit: WeeChat 2.4]
<notgne2>
polling always upsets me, but that's probably the best/easiest way to go without making a proper init daemon
<notgne2>
oh that's true, something about it still bugs me though, but I guess using files is just as good as any other method of internal signaling in a "real" init daemon
<clever>
notgne2: the only way to make it any better, is to generate a dep tree, and fill the nodes with function pointers to initiate starting more things
<clever>
then when you started x, you run a function pointer, that runs y
<clever>
but that needs a proper language, not bash
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
rsoeldner has quit [Ping timeout: 276 seconds]
ilyagotfryd has quit [Ping timeout: 245 seconds]
rsoeldner has joined #nixos
<notgne2>
well I think the fair balance between easy and "right" is to generate bash scripts for each service with a header to wait on files being created for each of the dependencies, and a footer to write a file for themselves, then start all those bash scripts at once, I suppose all a "real" daemon would do, is essentially the same thing, except with
<notgne2>
waiting handled internally, and signaling done by actually waiting on the script iself
<notgne2>
I think I might give that method a try in Continix soon
orivej has joined #nixos
kvda has joined #nixos
<clever>
notgne2: the main issue i would run into, is that the perl runner script, does pre-start, start, post-start, and stop all for you
<clever>
notgne2: but, i want to wait until post-start has finished, then start things that depended on it
<clever>
so i would need to patch it into the perl code
<iqubic>
Anyone know what's up with my nix-shell not working?
<notgne2>
I'll probably continue to avoid .runner anyway since most of the work I've already replicated in nix+bash, and it brings in a pretty big perl dependency
<clever>
notgne2: yeah, ewww perl! :D
<drakonis>
aw ye, perl?!
<notgne2>
I mean, I'm building this system for a company that is mostly going to be running perl code in it, but I don't feel like having the ~90MB dependency in everything else too
boxscape has joined #nixos
Guest97776 has joined #nixos
<clever>
notgne2: have you seen what i did with haskell-init?
<clever>
iqubic: find out what package provides disp, and what version its in
<iqubic>
I know what disp is.
<iqubic>
It's in brick.
<clever>
does disp take a PackageName and return a Doc?
<iqubic>
Not sure.
<clever>
check on hackage
<iqubic>
Wait? It's not in brick.
Supersonic has quit [Disconnected by services]
Supersonic112 has joined #nixos
Supersonic112 is now known as Supersonic
<disasm>
doesn't seem to have any negative effect but I get `[/etc/tmpfiles.d/var.conf:19] Duplicate line for path "/var/cache", ignoring.` every time I rebuild switch on 19.09. Anyone else seeing this? Trying to determine if this is my machine or something affecting everyone.
<iqubic>
I have no idea where disp is coming from.
<iqubic>
clever: How do you know that's the issue?
<clever>
iqubic: because the error said so
boxscape has joined #nixos
<iqubic>
I have no idea where disp is coming from.
<clever>
iqubic: check the git history for cabal2nix, and the file that is failing to build
<clever>
when was disp added? do the commits say anythign?
rsoeldner has quit [Remote host closed the connection]
<iqubic>
I'm not sure which git repo's history I'm supposed to be checking.
rsoeldner has joined #nixos
<clever>
builder for '/nix/store/z4176nbyjj2x4rnrbapvicfd5cimjqx5-cabal2nix-2.14.4.drv' failed with exit code 1
<clever>
iqubic: it clearly says cabal2nix itself is failing to build
<iqubic>
Ah. I see.
<iqubic>
So how do I go ahead and search all the commits of the main nixpkg repo to find the commit that introduced these bugs?
<clever>
iqubic: or use git blame within cabal2nix, to see when the problem was introduced
<iqubic>
I don't have cabal2nix installed.
<iqubic>
I'm just pulling from the repos and calling it right away.
<clever>
iqubic: nix is build it for you, but that build is failing
<iqubic>
Right. I see.
<iqubic>
Why does nix have to build it for me?
<clever>
iqubic: because callCabal2nix needs a copy of cabal2nix, to call cabal2nix
<clever>
iqubic: you can also call it yourself
<iqubic>
I se.
<iqubic>
That makes sense.
<iqubic>
So how could I go about fixing this issue? What options do I have?
<clever>
start by figuring out when it was broken, by reading the git history of cabal2nix
<iqubic>
I will.
<iqubic>
First I have to figure out where the derivation of cabal2nix is located.
<clever>
iqubic: and hackage has a link to the github page
<iqubic>
Correct.
<iqubic>
So I have no idea what's going on here.
<clever>
iqubic: what file did the error say the problem was in?
drakonis has quit [Quit: WeeChat 2.6]
<iqubic>
derivation.hs
<clever>
iqubic: now open derivation.hs, on the github page
<iqubic>
I'll look at the history for that.
justanotheruser has quit [Ping timeout: 276 seconds]
lord| has quit [Read error: Connection reset by peer]
<iqubic>
There are a few commits modifing this file: one on Dec 17, 2018, and then another on Sept 21, 2019. Those are the latest two commits modifying this file.
<clever>
iqubic: do any commits add or remove a disp on line 115?
<iqubic>
Makes sense, as I haven't touched this project in about a week or so.
dansho has quit [Remote host closed the connection]
<iqubic>
When I run `nix-edit nixpkgs.cabal2nix` I see that the version on my machine is 2.14.2
dansho has joined #nixos
<clever>
then you need to update the nixpkgs
<iqubic>
I know.
<iqubic>
I don't recall how to do that though. Can you help me please?
<clever>
nix-channel --upda
<clever>
te
<iqubic>
Thanks.
<iqubic>
I guess I should have done that when 19.09 came out.
<iqubic>
But thank you for your help with this.
<iqubic>
Just running a `nixos-rebuild` now.
<iqubic>
Yeah. This is going to take a while.
lovesegfault has joined #nixos
<lovesegfault>
Can someone build an aarch64 image for my raspberrypi with ssh enabled?
<lovesegfault>
the default one doesn't have ssh and I don't have a monitor
<lovesegfault>
Makes it literally impossible to get NixOS working on my Pi3
<iqubic>
clever: Actually, something weird is happening here. I just ran `nix-channel --update` followed by `nixos-rebuild switch` Both as root. I then checked what version of cabal2nix I had, with `nix edit nixpkgs.cabal2nix` It was still at 2.14.4
felixfoertsch has quit [Quit: ZNC 1.7.3 - https://znc.in]
felixfoertsch has joined #nixos
<clever>
iqubic: what does `nix-channel --list` say?
<iqubic>
I just used the package searching function on the nixos website to find the derivation in nixpkgs. That confirms that unstable is still on 2.14.4
justanotheruser has joined #nixos
<iqubic>
Clever: Thank you for all your help with this.
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to gnome-3.34: https://git.io/Jen2B
<iqubic>
What does the `Hydra job for tests` mean?
<iqubic>
What does that column mean?
<clever>
iqubic: click that link
<clever>
every job on the left side must be green, before the channel will update
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to release-19.09: https://git.io/Jen2u
<iqubic>
Why have many jobs been cancelled?
Rusty1 has quit [Quit: Konversation terminated!]
rsoeldner has quit [Ping timeout: 246 seconds]
<clever>
iqubic: because those are from a week ago, and arent important
rsoeldner has joined #nixos
<iqubic>
The latest aggregate test of NixOS-unstable has, mostly green checks, 15 failed tests, 4 with dependencies that have failed, and a ton of cancelled tests.
<aria>
hi! i just updated my system after a few months and as i was trying to compile a local project that uses `pkgs.haskellPackages.callCabal2nix` it's giving me a strange error trying to compile something nix related. src/Distribution/Nixpkgs/Haskell/Derivation.hs:115:41: error Variable not in scope: disp :: PackageName -> Doc full log: https://gist.github.com/ar1a/979682ce95f8de2a09feb6f1ae5d2d0c
<iqubic>
I know the answer to that.
<iqubic>
LOL.
<clever>
aria: refer to the above conversation with iqubic, its already fixed in nixpkgs master, just need to wait for the channels to update
* clever
heads off to bed
<iqubic>
aria: clever and I just finished figuring out what the cause of that issue was.
knupfer has joined #nixos
<aria>
huh, good timing
<aria>
do i just coment out those lines for a few days then?
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « flatpak-builder: 1.0.8 -> 1.0.9 »: https://git.io/JenaJ
<{^_^}>
[nixpkgs] @mujx opened pull request #70152 → haskellPackages.proteaaudio: link with the AudioToolbox framework on Darwin → https://git.io/JenaL
<{^_^}>
[nixpkgs] @worldofpeace pushed 3 commits to master: https://git.io/JenaY
<ar>
jluttine: exec is a shell-builtin and a family of syscalls. if it's returning "No such file or directory", it means that the binary it tried to exec to either doesn't exist, or its interpreter doesn't exist, or one of the required dependencies doesn't (if it's a binary and not a script)
<jluttine>
ar: thanks. i guess it cannot find "sh" when the command is executed in systemd service. i wonder if it would be a correct patch to use `/bin/sh` instead of `sh` in general (not just NixOS)
abathur has quit [Ping timeout: 240 seconds]
dermetfan has joined #nixos
ilyagotfryd has joined #nixos
hyper_ch2 has joined #nixos
dermetfan has quit [Ping timeout: 265 seconds]
ilyagotfryd has quit [Ping timeout: 276 seconds]
m0rphism has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
dansho has quit [Remote host closed the connection]
<bahamas>
is there a way to check what channel is currently active? `nix-channel --list` shows nixos-19.03-small, but when I load <nixpkgs> in nix repl, I don't have access to postgresql_11
<eyJhb>
bahamas: what does it list (everything)?
<bahamas>
eyJhb: sorry, I don't understand the question
Chiliparrot has joined #nixos
<eyJhb>
bahamas: What is the output from `nix-channel --list`? Because when you use <nixpkgs> you are using the "nixpkgs" channel (the name you have given it in your channels)
<eyJhb>
So if the name is "nixos-19.03-small", you would use <nixos-19.03-small>
<bahamas>
eyJhb: oh, I see.
<{^_^}>
[nixpkgs] @cdepillabout opened pull request #70154 → linkFarmFromDrvs: Add a trivial builder to create a linkFarm from a list of derivations → https://git.io/JenV3
<etu>
bahamas: That site only show current stable channel and only free packages. So if you have enabled unfree on your system or are on another channel. It's better to do a local search.
<eyJhb>
Is there actually any reason for not having unfree packages? Or broken even? Just mark them as unfree/broken?
wfranzini has joined #nixos
<bahamas>
I find it strange that postgres is not in small considering that small is recommended for servers
<bahamas>
so I guess I should stick with the main nixos channel for development
hmpffff has joined #nixos
dermetfan has joined #nixos
Jackneill has joined #nixos
<{^_^}>
[nixpkgs] @woffs opened pull request #70156 → stenc: init at 1.0.7 → https://git.io/JenVi
nikola_i has joined #nixos
Heirlung has quit [Quit: ZNC 1.6.5+deb1+deb9u2 - http://znc.in]
ng0 has joined #nixos
<Ariakenom>
does nix-shell -p python 27 create a gc root for that python?
<Ariakenom>
nix-shell -p python27
veske has joined #nixos
<notgne2>
Ariakenom: I don't believe so but I may be wrong, my understanding of gc-roots from nix tools is they generally rest on generated `result`s that will appear in your working directory, so nix-shell (maybe not the more complex usages) probably wont
Heirlung has joined #nixos
<Ariakenom>
so a GC might break it?
<notgne2>
it will just mean nix-shell will have to re-download python27 the next time you execute
<Ariakenom>
but what if the GC runs while the shell is doing stuff?
<notgne2>
I believe nix has locks to avoid that kind of thing, what's in your nix store shouldn't really matter as nix will always pull/build anything that's required
hyper_ch2 has quit [Remote host closed the connection]
<notgne2>
I bet there's a temporary gc-root created somewhere while the shell is running, but I don't think there's anything that will remain afterwards
<notgne2>
so if you don't want it to re-pull python27 between every gc execution, you'll have to manually create a gc-root somehow
<eyJhb>
ma27[m]: Sorry for ping, but is there any preference regarding ZHF, regarding e.g. a package which was last updated February 2018, and the last update is 2019-01-03, which doesn't support OpenSSL 1.1.1, but needs to be compiled using 1.0.2? Should that be marked broken, or just give it the older version of OpenSSL?
Chiliparrot has joined #nixos
o1lo01ol1o has joined #nixos
Tucky has joined #nixos
<Yaniel>
the older version of openssl is packaged separately IIRC?
<Yaniel>
so just change the package to use that explicitly
aveltras has joined #nixos
<eyJhb>
`openssl_1_0_2`, but still, does it make sense to do so? I can do that of course, but it just seems like a one-off-package
ilyagotfryd has joined #nixos
<bahamas>
is there a tool that allows you to search for a package and see what versions are available in the different channels?
Palpares has quit [Remote host closed the connection]
Palpares has joined #nixos
bahamas has quit [Ping timeout: 252 seconds]
cfricke has quit [Quit: WeeChat 2.6]
nikola_i has quit [Read error: Connection reset by peer]
fendor has joined #nixos
alex`` has joined #nixos
hyper_ch2 has joined #nixos
lopsided98 has quit [Ping timeout: 250 seconds]
lopsided98 has joined #nixos
MarcWeber has joined #nixos
cfricke has joined #nixos
cfricke has quit [Read error: Connection reset by peer]
iyzsong has joined #nixos
shreyansh_k65 has joined #nixos
graf_blutwurst has joined #nixos
iqubic` has joined #nixos
FRidh has joined #nixos
mexisme has quit [Ping timeout: 276 seconds]
shreyansh_k65 has left #nixos [#nixos]
iqubic has quit [Ping timeout: 245 seconds]
<graf_blutwurst>
Good morning. Does anyone have experience with nested virtualization with nixos as a host? I get vt-x/amd-v errors when the nested VMs start. It does work on VMPlayer in my windows install though (after enabling vt-x/amd-v). enabling this in virtualbox however doesn't do a thing, it's grayed out in the GUI and forcing it through cli doesn't change a thing
o1lo01ol1o has quit [Remote host closed the connection]
shreyansh_k60 is now known as shreyansh_k
<teto>
anyone using coc-python ? executing .config/coc/extensions/coc-python-data/languageServer.0.4.24/Microsoft.Python.LanguageServer says the file doesn't exist though it exists. "file" says its interpreter is /lib64/ld-linux-x86-64.so.2. How can I rewrite this to something sensible ? (even if hackish)
o1lo01ol1o has joined #nixos
shreyansh_k has left #nixos [#nixos]
shreyansh_k has joined #nixos
shreyansh_k has quit [Remote host closed the connection]
yumasi has quit [Ping timeout: 268 seconds]
ilyagotfryd has joined #nixos
<teto>
seems like patchelf --interpreter should do the trick
<ma27[m]>
eyJhb: I guess it depends on how important the package is. We use older openssl for several packages atm where it's not possible to build against 1.1.1, but probably globin has a different opinion here :)
<bahamas>
in nixos, why doesn't nix-channel --list display the current channel? right now, it doesn't display anything
<bahamas>
this is a fresh install, btw
hmpffff has joined #nixos
ubert has joined #nixos
cfricke has joined #nixos
<eyJhb>
ma27[m]: will try to ask globin then. It seems like it isn't that important of a package
veske has joined #nixos
gxt has joined #nixos
<manveru>
does anyone know of a way to configure a systemd unit to restart when it becomes unresponsive? (i need to check that through checking its http response)
<manveru>
or do i need a separate unit for that? searching for this only gives me results about restarts on crashes... which isn't helpful :P
freeman42x has joined #nixos
<__monty__>
systemd watchdog?
<freeman42x>
I have installed Nix on Ubuntu KDE. But when I install a program it does not appear in the KDE menu
<freeman42x>
for example I installed Google Chrome but it is nowhere to be found
<manveru>
__monty__: that needs the service itself to send the signal, no? plus it reboots the whole system on failure :P
<freeman42x>
I can only launch it from the terminal
<manveru>
freeman42x: i think you'll have to login again for it to show up?
<freeman42x>
manveru, I rebooted and no luck...
<manveru>
you installed in systemPackages?
<__monty__>
manveru: Systemd watchdog can restart services afaik.
<freeman42x>
manveru, I installed at user level in ~/.nixpkgs/config.nix and then nix-env -i all
<__monty__>
But you're right, it does seem to require the service to send heartbeats.
sigmundv_ has joined #nixos
hmpffff has quit [Quit: nchrrrr…]
<manveru>
freeman42x: sorry, i'm not sure how it works, i never do installs via nix-env directly
<__monty__>
freeman42x: It's possible desktop items are a seperate target that doesn't get installed by default like that. Search your nix-profile for chrome.desktop.
<freeman42x>
__monty__, sorry, I do not understand how to do that and how it could help
<Twey>
Hey! I'm using the emacs nix-mode, but it seems to default to four-space indentation everywhere and I can't seem to figure out how to make it do two-space (my company's standard). Setting smie-indent-basic doesn't seem to help
<Twey>
I guess I could turn off smie, but I quite like it
phreedom_ has joined #nixos
<niksnut>
Twey: strange, nix-mode should use 2 spaces by default
<cinimod>
I am building haskell packages where the LICENSE file is a link. nix-build says (presumably via cabal) that there is no LICENSE file and fails :(
<{^_^}>
[nixpkgs] @NeQuissimus pushed 2 commits to release-19.03: https://git.io/Jenig
<{^_^}>
[nixpkgs] @NeQuissimus pushed 3 commits to release-19.09: https://git.io/Jeni2
<cinimod>
I have temporarily created real LICENSE files but there must be a better way
knupfer has joined #nixos
armin has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @schmittlauch opened pull request #70168 → cawbird: init at 1.0.1 → https://git.io/Jenio
hyper_ch2 has quit [Remote host closed the connection]
<gchristensen>
etu: what is an extra 0 between friends
jmeredith has joined #nixos
<etu>
gchristensen: [answering if it was a riddle or trick question] I don't know, what is it?
<gchristensen>
just a joke about my own accuracy, like an order of magnitude is no big deal :)
<bendlas>
is there a way to distinguish unrealized derivations in `nix-store -q`?
bahamas has joined #nixos
ddellacosta has joined #nixos
dm9 has joined #nixos
drakonis has joined #nixos
<ddellacosta>
hey folks, my laptop doesn't seem to go to sleep when it gets low on power, and just dies, which is pretty frustrating. Can anyone point me at where I should look to debug this more? E.g. which configuration directives in the nixos docs I may be missing
dansho has joined #nixos
Ariakenom has joined #nixos
<Ariakenom>
sddm keeps a separate pswd entry for each monitor on login. which is #1 strange. but #2 smears my pswd if I just connected it to monitors.
<ajs124>
I'm getting "infinite recursion encountered, at /nix/var/nix/profiles/per-user/root/channels/nixos-19.09-small/lib/modules.nix:336:9" with a module that worked in 19.03 :(
civodul has quit [Quit: ERC (IRC client for Emacs 26.3)]
<fgaz>
grahamc: What is the process/requirements to move myself from ofborg's known user to trusted? I'd like to speed up darwin builds on my prs
nschoe has joined #nixos
nschoe has quit [Ping timeout: 245 seconds]
justanotheruser has quit [Ping timeout: 240 seconds]
nakkle has joined #nixos
<infinisil>
ajs124: Oh that's unfortunate, what's the module?
<nakkle>
Hey there! I'm having trouble getting nix's --max-jobs to work. Even if I do a nix-build --max-jobs 0, it still builds what it's told to. What am I doing wrong?
<ajs124>
abbradar[m]: So... there's this commit you made in July, which now results in my module running into an infinite recursion. Actually, my module is based off of systemd-confinement, so that might be broken as well.
<Twey>
TheBrodmann: Uh, not specifically, but if you know the answer please feel free to reply :)
bahamas has quit [Ping timeout: 240 seconds]
mjsir911 has quit [Client Quit]
andi- has joined #nixos
mmlb7 is now known as mmlb
phreedom_ has quit [Ping timeout: 260 seconds]
<TheBrodmann>
Twey: Ah, I asked a question and I thought you are giving me a solution. I don't know the answer. sorry.
<Ericson2314>
tetdim: ping! did you ever push your meson work?
<{^_^}>
[nixpkgs] @basvandijk opened pull request #70180 → cargo-audit: init at 0.9.1 → https://git.io/Jen1E
mjsir911 has joined #nixos
mjsir911 has quit [Remote host closed the connection]
<abbradar[m]>
ajs124: ouch, can you give me more context?
<abbradar[m]>
ajs124: interesting. i'll take a look
is_null has quit [Read error: Connection reset by peer]
gila has quit [Quit: My Mac Pro has gone to sleep. ZZZzzz…]
mjsir911 has quit [Quit: Goodbye, World!]
Tucky has quit [Quit: WeeChat 2.6]
<qubasa>
Does someone here have an idea why `boot.kernelPackages = unstable.pkgs.linuxPackages_5_2;` loads the wifi driver rtw88 automatically and the stable kernel 5_2 does not?
freeman42x has joined #nixos
iqubic` has quit [Quit: ERC (IRC client for Emacs 27.0.50)]
iqubic has joined #nixos
<freeman42x>
I used: https://nixos.wiki/wiki/Wine but I am getting following error: `64-bit Windows is required` when I launch my program
veske has quit [Quit: This computer has gone to sleep]
<qubasa>
freeman42x: Wine runs normally in x86 mode to change that you have to open the wine settings menu and change the appropriate option to x64
wing8 has joined #nixos
zupo has joined #nixos
fenedor has joined #nixos
<ajs124>
abbradar[m]: Thanks! It seems like an interaction between nixos/modules/config/resolvconf.nix:115 and 128 (enable = ![...] and mkIf cfg.enable)
fendor__ has joined #nixos
<freeman42x>
qubasa, how do I open the wine settings menu?
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « imagemagick7: 7.0.8-58 -> 7.0.8-66 »: https://git.io/JenMO
mjsir911 has quit [Client Quit]
mjsir911 has joined #nixos
mjsir911 has quit [Remote host closed the connection]
mjsir911 has joined #nixos
<domenkozar[m]>
TheBrodmann: wait until tomorrow for haskell updates to land
Okinan has joined #nixos
Bruno[m] has joined #nixos
fendor__ is now known as fendor
mjsir911 has quit [Client Quit]
<{^_^}>
[nixpkgs] @kmein opened pull request #70181 → vimPlugins.vim-fetch: init at 2019-04-03 → https://git.io/JenMl
mjsir911 has joined #nixos
<cinimod>
I am building haskell packages which have their LICENSE file as a link. These fail but when I replace the link by a real file, the build succeeds
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #70182 → Revert "nixos/desktop-managers/xterm: Disable by default" → https://git.io/JenMz
<TheBrodmann>
domenkozar[m]: thank you
<shyim>
I am getting sometime the warning "warning: you don't have Internet access; disabling some network-dependent features" while doing nixos-rebuild. Is there any known fix?
<evanjs>
Bah. Just hit a wall. Trying to install multiple rust channels via my rust dev module and the moz overlay. But now they conflict, which I guess I should've expected...
<eyJhb>
wucke13: cp -r include/ $out ???
boolman has quit [Ping timeout: 240 seconds]
<evanjs>
Though I'm wondering, how does rustup handle this? Is there a way I can have them handled so I can call them like "cargo +stable run", etc?
<eyJhb>
evanjs: nix shells?
<wucke13>
thanks
<{^_^}>
[nixpkgs] @kampka opened pull request #70189 → kops_1_14: init at 1.14.0 → https://git.io/JenDg
<evanjs>
Eh, I have these installed on a system level, but derp. Yeah wow. I'm using nix-shells for these anyway... >.> so I just have to... change the shells to say stable...
<evanjs>
Okay, going to shelve what I'm currently doing and just do that for now. Thanks for answering my derp haha
<eyJhb>
Seems like it would be the best option. You can even do aliases, so that you can easily change between them
<eyJhb>
but you might also be able to "rename" the binaries, but yeah :p np
<eyJhb>
Just sitting looking at my code hating my life, so glad I can help
<evanjs>
Yeah I'm sorta wondering how NixOS rustup does it so I could potentially mimic that
<boxscape>
does nix-collect-garbage delete all packages that could be deleted with nix-store --delete without -f?
<{^_^}>
[nixpkgs] @Ericson2314 opened pull request #70195 → lib: Add armv7a-linux to doubles.nix → https://git.io/Jenyq
ntqz has quit []
zupo has joined #nixos
nschoe has quit [Ping timeout: 276 seconds]
bahamas has joined #nixos
<bahamas>
I have a nix package that takes a long time to build. can I copy it from the nix store and save it somewhere? I'm asking because I need to delete the VM where I built it
<clever>
bahamas: nix-store -r /nix/store/foo --add-root result --indirect, will create a result symlink pointing to it, which prevents it from being GC'd
<floscr>
Can somebody help me, I need an exact ruby version (2.4.6) but the stable channel has 2.4.5 and unstable 2.4.7
<bahamas>
clever: where foo is the name of the library, I assume
<clever>
bahamas: yep
<fuzen>
Hello, I created a project that builds with a default.nix, is there any way I can use this default.nix as a binary in docker.nix without adding it to a store or running two separate commands? Or should I just build default.nix then docker.nix
<bahamas>
clever: ok, thanks.
<clever>
fuzen: you should be able to just do `import ./default.nix` inside `docker.nix`
<fuzen>
clever: thanks
floscr has quit [Remote host closed the connection]
<Ericson2314>
tetdim: volth was just hammering through
<drakonis>
plus all the awkward xen specific code
<Ericson2314>
if you could merge that and get it building with meson (!!!!!!!!)
<drakonis>
(awkward not quite but still)
<drakonis>
a thing i've heard tho
<gnidorah>
infinisil: as you recommended I've switched from system.activationScripts to systemd service. howerver it doesn't start after config file in restartTriggers is regenerated :( https://www.irccloud.com/pastebin/gOuZlOSM/
<drakonis>
qubes people wanted to use nixos as a base for qubes instead of fedora
<tetdim>
Ericson2314, ahhhh
<tetdim>
it build all the way to main.cc
<tetdim>
failing on socket stuff
<tetdim>
hrm probably a freebsd specific link
<Ericson2314>
gotcha
gxt has quit [Ping timeout: 260 seconds]
<Ericson2314>
yeah the long tail of os-specific stuff
<gnidorah>
infinisil: I don't need it started or restarted in any other case. only when 00_site_settings is generated by my module
<infinisil>
gnidorah: What does `systemctl cat dconf-update` say?
<infinisil>
There should be some RestartTriggers field
johnw has joined #nixos
chloekek has joined #nixos
chloekek has quit [Ping timeout: 240 seconds]
fresheyeball has quit [Quit: WeeChat 2.4]
rotaerk has quit [Read error: Connection reset by peer]
<evils>
i read at least some of that, if there was syntax available to mark that this can be removed if i don't maintain it, i'd use that...
<gnidorah>
infinisil: thank you much, now it works! though it starts every time nixos-rebuild switch is done (even if configuration haven't been changed), is it ok?
<clever>
gnidorah: does the value in the restart triggers field change?
<infinisil>
gnidorah: Try RemainAfterExit = true
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<infinisil>
serviceConfig.RemainAfterExit = true
<gnidorah>
clever: no
<clever>
probably the remain after exit then
<infinisil>
evils: There isn't anything for that
<infinisil>
But nixpkgs is getting really big
<infinisil>
Too big imo
<infinisil>
Just look at the number of open PRs
mexisme has joined #nixos
<gnidorah>
infinisil: clever: that works, thank!
<infinisil>
And the more PRs for new packages we merge the more PRs will be opened to update the new packages
<evils>
infinisil: the impression i got from your comment is that you're trying to set a bias for exclusion because some aspect of nix aren't as optimized as they should be
zupo has joined #nixos
<evils>
infinisil: am i right in understanding that flakes are (partially) intended to mittigate the "too many PRs to handle" issue?
<infinisil>
Well if we ignore the NixOS module problem (that it slows down everybody's nixos-rebuild), this isn't really a problem with Nix itself
<infinisil>
evils: While I haven't thought about it this way, this sure does sound like a good argument for flakes :)
<infinisil>
flakes will be a better NUR
<evils>
on the basis that there's too many PRs, and that i'm still working on this tool, i can see how you wouldn't want to get stuck with future updates for v0.7...v0.99
<evils>
i'm fairly new to nix, i looked and NUR and it kinda seemed like a hack that could be very useful for advanced users
mexisme has quit [Ping timeout: 245 seconds]
<mdash>
the main thing is that not everyone wants to tie themselves to the nixpkgs review/release cycle
<infinisil>
I guess NUR's readme doesn't explain it too well imo, but ideally it should be usable by beginners too
<mdash>
right. and i maintain software that's built with nix outside of nixpkgs
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
termita has left #nixos [#nixos]
zupo has joined #nixos
<evils>
is there an expected timeline for flakes?
gxt has joined #nixos
<tetdim>
Ericson2314, got past some errors on freebsd :^)
<tetdim>
lets see what happens now
zupo_ has joined #nixos
<Ericson2314>
tetdim: noice noice noice!
<tetdim>
i might be adding more config options than whats already there
<tetdim>
and removing some that arent needed anymore
<tetdim>
theres a bunch of stuff missing and still gotta fix tests n such after
<evils>
infinisil: i'll probably close my PR, any suggestions on how i can help with the "too many PRs" issue?
zupo has quit [Ping timeout: 240 seconds]
jd823592 has joined #nixos
glasserc has quit [Read error: Connection reset by peer]
<jd823592>
hi, so i just created chroot dir within my home, generated config, nixos-installed it, nixos-entered and now i would like to install via nix-env and I get 'error: changing ownership of path '/nix/store': Invalid argument', is there anything to be done about it?
selfsymmetric-pa has joined #nixos
<infinisil>
evils: Reviewing and testing PRs helps a lot
<infinisil>
Since you're new to Nix it's probably hard to review the code properly though, so maybe more testing them
<evils>
infinisil: is there any documentation on the workflow to that?
<infinisil>
I'd recommend searching for packages you use yourself, whether there's open PR's for them
<kumikumi>
I'm thinking about writing my first nix package declaration, from this: https://paper.readthedocs.io/en/stable/server/getting-started.html I'm confused though because they seem to be offering "latest stable builds", which might change every few days. So how can I use fetchurl if I don't know the hash in advance? Or should I take a different route and try to build it myself?
o1lo01ol1o has quit [Remote host closed the connection]
selfsymmetric-pa has quit [Remote host closed the connection]
<evils>
kumikumi: the idea with nix is to have repeatable builds, so you pick a specific release, and set the hash for that
<{^_^}>
[nixpkgs] @esclear opened pull request #70203 → pgmodeler: init at 0.9.1 → https://git.io/JenHS
<setthemfree[m]>
I'm having a fonts issue with one package on non-nixos (the package is supercollider). In the in-app documentation browser some fonts seem to be missing and chunks of documentation are not rendered. There are errors logged by the process: https://bpaste.net/show/OhBF
o1lo01ol1o has quit [Remote host closed the connection]
<setthemfree[m]>
l33_: I assume this means you are on nixos?
<l33_>
LOL
<l33_>
you assume correctly
<l33_>
what are you running nix on?
<setthemfree[m]>
Fedora.
shabius has joined #nixos
<setthemfree[m]>
Maybe there is the possibility to set `FONTCONFIG_FILE` to point it to a nix one instead of the /etc/... one from Fedora, but I have no idea where to point it to (since there is no etc/fonts/ in my ~/.nix-profile)
<alienpirate5>
I'm having trouble cross-compiling libX11 for armv7l-hf
<alienpirate5>
you can test this by running `nix-build -E '(import <nixpkgs> {}).pkgsCross.armv7l-hf-multiplatform.xorg.libX11'`
<alienpirate5>
here is the error I get:
<alienpirate5>
../../include/X11/Xlib.h:44:10: fatal error: X11/X.h: No such file or directory
<alienpirate5>
#include <X11/X.h>
<alienpirate5>
i looked in the definition of libX11 and noticed that xorgproto is already in the build inputs
<c00w>
This might be a weird thought, but has anyone heard of having nix inject a compiler so each call produces a nix derivation build? (i.e. replacing the default cc compiler with one that pulls the inputs, and runs the build inside of nix so the output is cached for repeated calls?)
hmpffff has quit [Quit: nchrrrr…]
hoppfull has joined #nixos
hink_ornl has quit [Remote host closed the connection]
hmpffff has joined #nixos
hmpffff has quit [Client Quit]
ris has joined #nixos
Lears has joined #nixos
hmpffff has joined #nixos
hmpffff has quit [Client Quit]
<freeman42x>
stable gives cabal-install package v 3.0.0.0, how can I install an older version of cabal-install using the per-user configuration?
nexgen has joined #nixos
nexgen has quit [Remote host closed the connection]
gnidorah has quit [Quit: Connection closed for inactivity]
nexgen has joined #nixos
<hoppfull>
Hello, I'm setting up the configuration.nix for the first time. I understand there is some mapping from the possible settings to some applications. I don't have an intuition for this mapping. For example 'services.xserver.layout = "en"' maps somehow to the command setxkbmap -layout "en". How could I know this if I wasn't familiar with xserver? Is
<hoppfull>
there somewhere where this mapping is defined?
<hoppfull>
Also how do I list legal values for a setting?
<{^_^}>
[nixpkgs] @Lassulus pushed commit from @woffs to master « stenc: init at 1.0.7 »: https://git.io/JenQ9
<hoppfull>
So lets say I set up KDE Plasma 5 in my configuration.nix file, I can just later change it to OpenBox for example and run nixos-rebuild and I should now have another desktop environment set up as if I installed it from the beginning?
<PyroLagus>
it seems like it's going to take a while before fontconfig gets updated to 2.13.1
<SniffleBuzz>
I guess I need to override the nixpkgs compton derivation somehow, but how can I make the `services.compton` refer to this build instead?
gila has quit [Quit: My Mac Pro has gone to sleep. ZZZzzz…]
<SniffleBuzz>
oh okay. Maybe its a perspective coming from guix, but Id rather not replace anything, I just want to use a 'new' package. Cant I make and use a new derivation in the configuration?
<waleee-cl>
I don't use service.compton myself, but It looks like you'd have to replace the service module with one of your own to do that (some services do have a package option, but this one wasn't one of them)
<{^_^}>
[nixpkgs] @veprbl pushed commit from @tobiasBora to release-19.09 « owncloud-client: Use qt5's own mkDerivation (#70187) »: https://git.io/Jen7N
MrCCoffee has joined #nixos
ng0 has quit [Quit: Alexa, when is the end of world?]
<witchof0x20>
I'm looking to encapsulate a service (transmission) so that it uses a vpn (wireguard) for communication. LXC seems like a good way to accomplish this. Is there a nice way to integrate the NixOS config of a LXC guest with the NixOS config of the host?
<Capisce68>
Why am I getting the error `error: The option value `services.compton.vSync' in `/etc/nixos/configuration.nix' is not of type `one of "none", "drm", "opengl", "opengl-oml", "opengl-swc", "opengl-mswc"'.` from ` services.compton = { enable = true; vSync = true; };`
hoppfull has quit [Remote host closed the connection]
<drakonis>
no signed bootloader equipped isom
domogled has quit [Ping timeout: 240 seconds]
<drakonis>
isos yet?
o1lo01ol1o has quit [Ping timeout: 245 seconds]
<clever>
drakonis: it would have to be signed with microsoft's private key
<clever>
drakonis: and if they give that out, then secure boot is basically pointless
<waleee-cl>
if you want to just replace windows, wouldn't eg ubuntu work? Canonical is supposed to have an a deal with MS that allows them so sign their isos
<clever>
for my laptop, the firmware lets my enroll my own public keys, or just whitelist efi binaries by hash
<clever>
so i can have secureboot on, without dealing with actual signing
<drakonis>
handy.
<clever>
it just means that it wont boot if grub ever updates
<clever>
and grub itself isnt locked down, so you could just use the grub UI to hax the machine as hard as you want
domogled has joined #nixos
<clever>
and the boot dir isnt encrypted, so you could just edit grub.cfg after yanking the drive, to backdoor it all you want
<clever>
proper secure boot requires multiple phases of signing
mbrgm_ has joined #nixos
mbrgm has quit [Ping timeout: 264 seconds]
mbrgm_ is now known as mbrgm
drakonis_ has joined #nixos
Chargen has joined #nixos
hoppfull has joined #nixos
<hoppfull>
I installed the os according to the manual with KDE Plasma 5 but I don't seem to have a way of accessing my wireless network. I have configuration networking.wireless.enable = true What am I missing?
domogled has quit [Quit: domogled]
<ajs124>
abbradar[m]: no problem. Thanks in advance!
<tyrion-mx>
hoppfull: It should be explained in the Nixos Manual (there is a section about networkmanager and desktop environments). I cannot reach nixos.org at the moment so I cannot check it
<hoppfull>
tyrion-mx: Yeah, just found it. I'm going to try and add "networkmanager" to my user groups and see if that works.
<tyrion-mx>
hoppfull: yes, I think you should also disable networking.wireless and enable networking.networkmanager
<manveru>
hoppfull: you probably want `networking.networkmanager.enable` and install `networkmanager` in `environment.systemPackages` as well
<manveru>
i also use `networkmanagerapplet`, but not sure if kde needs that
<manveru>
seems like there's `plasma-nm`
<manveru>
and that's added automatically, so don't worry about it :)
<hoppfull>
plasma-nm: command not found
<manveru>
well, not sure what the executable is called, that's just the package name
<hoppfull>
manveru: ok cool
musteresel has joined #nixos
<hoppfull>
the site seems down
<hoppfull>
*snooooore*
<manveru>
it is :|
abathur has joined #nixos
<hoppfull>
Is this common?
<manveru>
nope, first time for me
cjpbirkbeck has joined #nixos
<musteresel>
Hi ... I just noticed that nix-build can build a list of derivations. What flags do I need so that it (tries to build) each derivation in that list, regardless of failures of single list items?
<drakonis_>
did nixos just go down?
<drakonis_>
oh it did
<hoppfull>
yep
<drakonis_>
a rare sighting
<musteresel>
nevermind it's --keep-going / -k .. I'm just blind
<cjpbirkbeck>
hello, what is the the status of 19.09?