<{^_^}>
[nixpkgs] @orivej-nixos pushed commit from @orivej to staging « zstd: 1.4.0 -> 1.4.1 »: https://git.io/fjMul
<iqubic>
No. I don't know where to find that.
aanderse has quit [Ping timeout: 252 seconds]
aanderse has joined #nixos
qubasa has quit [Ping timeout: 272 seconds]
<iqubic>
Hang on a sec. Let me connect my new printer to my WIFI router.
shomodj has joined #nixos
<iqubic>
I think a printer that isn't connected to my computer with a cord, or via the WIFI Router isn't going to work at all. Even if I give CUPS the right PPD.
<iqubic>
DigitalKiwi: How long is that supposed to take to run?
<DigitalKiwi>
mine took 3 seconds
<DigitalKiwi>
after you find the printer you can even figure out the web interface port
<iqubic>
Well, nmap is refusing to work.
<DigitalKiwi>
weird
thc202 has quit [Ping timeout: 245 seconds]
<iqubic>
yeah. It just stalls out for me.
<iqubic>
Why are Canon Printers a pain?
<iqubic>
And did I buy the wrong printer?
<iqubic>
I just purchased this earlier today, and it isn't working.
tnks has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/fjMuu
<gchristensen>
I can barely help with printer stuff. when I asked the internet for recommendations, the answers were HP is easiest and brother is second
<iqubic>
Well, this sucks.
<gchristensen>
yeah :(
<clever>
gchristensen: my dad found a used printer/scanner at a yard sale, and the http server within it let me link it to a google account
mix has joined #nixos
<jackdk>
that's horrifying
<clever>
gchristensen: i can now google cloud-print (bypassing cups entirely) from anywhere, and it comes out of his printer, despite the fact that he ran off with it :P
<gchristensen>
clever: oh my goodness
<gchristensen>
:o
<DigitalKiwi>
the nmap might not be the right command for your network if it's not fast
<iqubic>
And I didn't do my research before buying this printer.
<DigitalKiwi>
what's your ifconfig -a
<gchristensen>
iqubic: I didn't, either. I use my wife's mac to print to my canon.
<ajs124>
I always just export to pdf, put that on a usb stick, put that in the printer and print from usb
<ajs124>
it's a very advanced workflow, I know
<clever>
ajs124: at least your not printing pdf's to pdf
mix has quit [Read error: Connection reset by peer]
<iqubic>
I have a raspberry pi with Raspian on it. If worst comes to wort I can rsync the files over there then print from that. Because I hope Canon's Debian drivers work better than what ever the hell I'm trying to do now.
ris has quit [Ping timeout: 272 seconds]
<DigitalKiwi>
did you do nmap -sn 10.0.0.1-255 or what i gave?
<iqubic>
I did. And it just stalled out doing nothing.
<DigitalKiwi>
:(
<iqubic>
But I only ran it for like 30 seconds.
<ajs124>
iqubic: you can press any key to get progress updates
<iqubic>
I didn't know that.
<ajs124>
ok, maybe not any key. some keys do stuff. in doubt, just press space
<ajs124>
OT, but I hope nmap never removes the "In the Movies" section of its website
<iqubic>
So why is this printer being a pain?
<iqubic>
And this printer is now connected to the internet.
<DigitalKiwi>
in the future when buying a printer it's a good idea to see if other people got it to work with linux...like gchristensen said hp are usually more likely
<iqubic>
I know. I know. I screwed up.
stepcut has quit [Remote host closed the connection]
<DigitalKiwi>
now that it's on the wifi see if cups detects it
hyper_ch has quit [Ping timeout: 248 seconds]
<iqubic>
If all else fails, I can use google cloud print with this thing.
<clever>
iqubic: and what does a screenshot of the router status page say?
<iqubic>
What part in particular do you need?
<clever>
iqubic: the general status page to start with
<iqubic>
Here's localhost:10.0.0.1
<tA->
whats the easiest way to go about stopping screen tearing? compton has always been a bit of a pain in the ass to set up, wondering if theres some x setting for it or something
<{^_^}>
[nixpkgs] @worldofpeace pushed 4 commits to master: https://git.io/fjMuA
<ivan>
file 'nar/2l38315v05wz32virhmdwf3qd556vcrf.nar' does not exist in binary cache 'http://finssd1.wg:5000'
<clever>
ivan: is that nix-serve? was garbage collection ran?
<ivan>
yeah, a nix-serve is on the other end (actually running on the same machine), and I do run GC
<clever>
ivan: once nix discovers that a binary cache has something, it will assume it still has it for up to 30 days
<clever>
ivan: and a GC will hard-break things
mexisme has joined #nixos
<clever>
ivan: as root, run `rm -rf ~/.cache/nix/binary-cache*` to make it forget
<ivan>
interesting
<ivan>
I just did that and I still get file 'nar/2l38315v05wz32virhmdwf3qd556vcrf.nar' does not exist in binary cache 'http://finssd1.wg:5000' in a loop
<clever>
ivan: you may need to delete that file on other users too
<clever>
and `sudo rm -rf ~/.cache/nix/binary-cache*` wont delete it from roots home
<nh2>
clever: do you know what the current best way to iterate on failed builds with `--keep-failed` is for Haskell projects? Am I supposed to run `nix-shell` on the Haskell .drv?
mexisme has quit [Ping timeout: 250 seconds]
<clever>
nh2: nix-shell on the .drv wont give a usable shell, ive not yet looked into why
<clever>
nh2: you need to figure out the right attrpath in the nixexpr that made it, and add .env to the end, and shell into that
<nh2>
clever: Is there a better way involving just cd'ing into the kept directory, and running stuff in there?
<clever>
nh2: if the build fails under normal nix-shell, then you can also use unpackPhase to make the source, and then just `runhaskell Setup.hs configure` then build
<nh2>
clever: my problem though is a failure compiling ./Setup, I think using runhaskell here would not reproduce that the same way
<clever>
nh2: what is the exact error? which package is it?
<nh2>
clever: working on static-haskell-nix, another instance of the fixed-Cabal-splicing-into-package-database not working, with the classic "cannot satisfy -package-id Cabal-2.4.1.0-9MZFDeNrcJI10bcroa6pq8" message
<nh2>
clever: I know how to debug it but it involves sourcing env-vars in the --keep-failed dir and then running select parts of the environment variables to resume the build -- a big pain. I'd like it to just run the `buildPhase` or whatnot
<nh2>
clever: after `nix-shell`ing the `.env`, what should I do next?
lambda-11235 has joined #nixos
lambda-11235 has quit [Client Quit]
<ornx>
i am having an extremely strange problem where when i play games in wine with dxvk in a container, my system freezes in such a way that all read calls outside the container block indefinitely but the rest of the system runs "fine" and inside the container i can do reads/writes as normal
<ornx>
nothing gets reported to the system log and the actual game runs fine
<ornx>
its such a bizarre problem that i have no idea where to report it or how to isolate it
<ornx>
has anyone ever had anything remotely similar to this happen to them?
<ornx>
i did strace ls while it was in this state, and it hung on the getdents system call
Guanin has quit [Remote host closed the connection]
<clever>
nh2: when i nix-shell into .env, i typically do `cabal configure ; cabal build`, and since cabal-install isnt present, i replace cabal with `runhaskell Setup.hs`, nixpkgs builds it with normal ghc though
<nh2>
clever: I just found that when I do `eval "$compileBuildDriverPhase"` instead of `eval $compileBuildDriverPhase` it actually runs
aw has quit [Quit: Quitting.]
spacefrogg has quit [Quit: Gone.]
<nh2>
still says `runHook: command not found`, not sure why
aw has joined #nixos
<nh2>
where is that supposed to come from?
spacefrogg has joined #nixos
szicari has quit [Ping timeout: 250 seconds]
<gyroninja__>
If you do nixos-rebuild build-vm, how are you suppossed to login to the vm? The only information I found online was to add an ssh key for a user and then make qemu forward sshd from the vm.
<gyroninja__>
*supposed
<clever>
nh2: runHook is part of setup.sh
<clever>
nh2: and compileBuildDriverPhase has newlines in it, `eval $compileBuildDriverPhase` strips all of them and ruins it
<khumba>
Is there an easy way to get the attribute set A that an overlay evaluates to, after self and super get applied but prior to (nixpkgs // A)?
<clever>
nh2: you can also do `phases="compileBuildDriver" genericBuild` to run just one phase
<clever>
gyroninja__: you must set a password in the configuration.nix file, either with non-mutable users, or initial(hashed)password
jasongrossman has quit [Ping timeout: 245 seconds]
<clever>
gyroninja__: if using the initial stuff, you need to delete the old qemu disk image in the dir you launched it from
slack1256 has joined #nixos
<nh2>
clever: where is setup.sh, how can I run it in my env? I also don't have `genericBuild` in scope either
<clever>
nh2: nix-shell should source setup.sh for you
<nh2>
in this case, running nix-shell directly on the failed .drv
<nh2>
and it repros the error message I get from the nix build, which is great!
andgate has joined #nixos
<clever>
nh2: ah
<nh2>
should be `phases="compileBuildDriverPhase" genericBuild` though (missing "Phase")
<nh2>
clever: what's the difference between `phases="compileBuildDriverPhase" genericBuild` and `eval "$compileBuildDriverPhase"`, is any of them better?
<clever>
nh2: phases can be both bash functions, or env vars of identical name
<clever>
nh2: genericBuild deals with choosing which one to run, based on some internal priorities
<clever>
so `eval "$unpackPhase"` will fail if the user didnt override it
<clever>
and `unpackPhase` will fail to behave correctly if the user DID override it
<clever>
nh2: this line of code will either run $curPhase (a bash function by the name in the var), or eval $curPhase, (a bash var, named after the contents of a bash var)
InFlames has joined #nixos
<nh2>
clever: ok that's great to know! Also, how does one know one can run nix-shell on the .drv? I couldn't find that in any manual
<nh2>
is the answer "one reads all the source" ? :)
<clever>
nh2: i dont know how i discovered that option, lol
<nh2>
then I file a ticket now
<jackdk>
nh2++; there's a lot of great stuff in nix that is sadly just folklore
<{^_^}>
nh2's karma got increased to 13
<nh2>
where folklore == clever's brain
<nh2>
one day I'll just build a clever brain transfer apparatus. That will instantly solve 30% of my weekly problems
<clever>
nh2: youll also inherit my in-ability to close tabs
<clever>
nh2: i have 32gig of ram, 64gig of swap, and i'm running out of both
<nh2>
clever: I also cycle around 500-1000 tabs
<DigitalKiwi>
how do you find any of them
<clever>
nh2: 2092 open right now
<DigitalKiwi>
and why
<clever>
DigitalKiwi: a plugin called vimium. T will let you search by the title of tabs
<DigitalKiwi>
why don't you use bookmarks
<clever>
DigitalKiwi: because i dont want to save X forever, i want to read it and close it
<clever>
then i run out of time and leave it for later
<clever>
then i open another 10 tabs
<clever>
repeat
<nh2>
many people (at least 2) use tabs as a simple TODO list
<DigitalKiwi>
i do too
<DigitalKiwi>
but then i make them todones and close them or bookmark and come back you can always delete bookmarks :P
Netsu has joined #nixos
<DigitalKiwi>
or i have a text file with links
<Netsu>
so that is the question: how can I install non-trivial paths from derivation?
<clever>
Netsu: for example?
<Netsu>
for example /home
szicari has quit [Ping timeout: 250 seconds]
<Netsu>
or /some/random/path
<Netsu>
I have derivation with pretty custom paths and tryping to `dockerTools.buildLayeredImage` it
<Netsu>
*trying
<clever>
Netsu: if you make a $out/home/foo and add it to the contents of the docker image, it will wind up in /home/foo
<Netsu>
I have derivation that have $out/home and $out/share, but at docker I see /share only.
<Netsu>
clever: ^^^
<clever>
Netsu: can you pastebin the nix exprs?
<Netsu>
maybe part of it, ok
<Netsu>
however I got invalid reference format trying to load docker image, and for some reason it have no label
<{^_^}>
#64746 (by greydot, 6 days ago, open): stack is broken on Nixos
<gyroninja__>
Do I just need to wait for haskellPackages to be updated?
___laika has quit [Ping timeout: 245 seconds]
<DigitalKiwi>
there's always the old fashioned not using stak since being broken is a feature of it (and not just on nixos)
<gyroninja__>
Well, I would like to use intero-mode
<gyroninja__>
afaik that requires stack
<DigitalKiwi>
dante if you're using emacs, or haskell-ide-engine
day|flip has quit [Quit: Using Circe, the loveliest of all IRC clients]
<DigitalKiwi>
i was using dante i switched to hie it's nice
<gyroninja__>
oh thanks for the pointer
<DigitalKiwi>
tbf they are both nice
<DigitalKiwi>
and it's nice for contributers if you don't force them to use stack since it's often a huge pain point
<DigitalKiwi>
just in general don't use stack, this is not the advice you asked for, but the advice you need :P
<gyroninja__>
Thanks
<gyroninja__>
So what's an advantage to hie?
<mac10688>
That's weird. I go to `cd /var/spool/nginx` and I get permission denied. So i `sudo cd /var/spool/nginx/` and it says "sudo: cd: command not fount" lol
<mac10688>
this seems like a bug
<mac10688>
found*
<gyroninja__>
mac10688: cd is a builtin
<gyroninja__>
not a command
<mac10688>
oh
endformationage has quit [Quit: WeeChat 2.5]
<DigitalKiwi>
i think it has more features than dante it's a bit more helpful
<DigitalKiwi>
possible advantage of dante is i don't think it's tied as closely to ghc version
<DigitalKiwi>
(not sure though)
palo1 has joined #nixos
johnny101 has quit [Quit: Konversation terminated!]
palo has quit [Ping timeout: 272 seconds]
palo1 is now known as palo
day|flip has joined #nixos
init_6 has joined #nixos
<{^_^}>
[nixpkgs] @thefloweringash opened pull request #65212 → nixos-test-driver: allow configuration of net frontend and backend → https://git.io/fjMg2
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
pamplemousse has joined #nixos
jasongrossman has quit [Ping timeout: 248 seconds]
<mac10688>
It seems that when I configure nginx with a root. It takes the root path I gave it and copies it into a nix/store. The issue is when I update my site and want to see my changes, they aren't there
<mac10688>
How do I get nixos to refresh root?
<init_6>
Hi all!
<day|flip>
so. if i login into 6697. that means im using tls?
<Ralith>
consult the documentation for your IRC client
<Ashy>
day|flip: usually yes, check your server output tab for tls info
<tilpner>
day|flip: If you are referring to IRCs: that is the conventional port for it, so probably yes, but nothing stops people from running listening on 6697 without TLS
<enuiarte>
I'm currently trying to create a (local, for now) package for a binary. I have successfully patched the rpath, but I am getting the following error:
<{^_^}>
[nixpkgs] @marsam opened pull request #65215 → buildRustPackage: by default use Rust's platforms → https://git.io/fjM2Y
o1lo01ol1o has joined #nixos
mix has joined #nixos
ericsagnes has joined #nixos
<tilpner>
enuiarte: Which operating system are you using?
veverak has quit [Ping timeout: 248 seconds]
<enuiarte>
tilpner: NixOS (19.03). I'm exploring another solution atm, I think I can circumvent the problem
stepcut has joined #nixos
wfranzini has quit [Remote host closed the connection]
<tilpner>
No idea then. You could share the expression to verify it also occurs elsewhere
wfranzini has joined #nixos
veverak has joined #nixos
<enuiarte>
I was trying to patch a bundled version of nwjs and the contained libnw.so throws this error. nwjs is essentially chromium, so my guess is that this binary does some fancy safety shenanigans … My workaround is an attempt to just use nwjs from nixpkgs and a stitched together build of the application
<enuiarte>
(If you want to call throwing JS into the right places a build process)
o1lo01ol1o has quit [Remote host closed the connection]
mix has quit [Read error: Connection reset by peer]
vonfry has joined #nixos
vonfry has quit [Client Quit]
Chiliparrot has joined #nixos
pmiddend has joined #nixos
shabius has quit [Quit: Leaving]
shomodj has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Miyu-chan>
Yeah, this is weird. I'm having slow connection to cache.nixos.org on AWS/ap-southeast-1
<Miyu-chan>
The ping is amazing, but the bandwidth is a bit ???
<Miyu-chan>
curl: Expire in 1000 ms for 7 (transfer 0x7f1848002020)
<Miyu-chan>
download thread waiting for 100 ms
<Miyu-chan>
curl: Expire in 1000 ms for 7 (transfer 0x7f1848002020)
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
rfold has joined #nixos
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
ng0 has joined #nixos
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
bakakuna has joined #nixos
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
mexisme has joined #nixos
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
mix has quit [Read error: Connection reset by peer]
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
rfold has quit [Quit: WeeChat 2.4]
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
rfold has joined #nixos
stepcut has joined #nixos
linarcx has quit [Quit: WeeChat 2.5]
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
zupo has joined #nixos
stepcut has quit [Remote host closed the connection]
mexisme has quit [Ping timeout: 264 seconds]
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
freddie-freeload has joined #nixos
Thra11 has joined #nixos
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
<freddie-freeload>
Hi there! :) I have a fresh NixOS installation and when trying to start xfce, I get "exec: xinit: not found". Does this mean xserver is not installed?
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
mix has joined #nixos
stepcut has quit [Remote host closed the connection]
<jackdk>
there is a section in the manual about setting up xfce, did you have any luck that? if so, post non-sensitive parts of configuration.nix so we have some idea, as your message doesn't give us much to go on https://nixos.org/nixos/manual/#sec-xfce
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
stepcut has quit [Remote host closed the connection]
mexisme has joined #nixos
magnetophon has quit [Ping timeout: 246 seconds]
stepcut has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
stepcut has quit [Remote host closed the connection]
<delan>
how do you use wrapProgram in a custom installPhase? I get command not found unless I move it to postInstall like other packages I’ve seen, but then it seems to have no effect, which might be because I’ve still got code in my own installPhase?
<Miyu-chan>
`nix copy` should technically support multihops, right?
<cinimod>
jackdk: will do so if and when I get it working
<rnhmjoj>
do you know why qemu started complaining about a lock when launching a nixos vm? i have to delete .qcow2 image everytime. it wasn't like this before
domogled has quit [Ping timeout: 272 seconds]
<marek>
how do you disable tests for a package in hackage-packages.nix?
<Miyu-chan>
So.
<Miyu-chan>
nix copy --to ssh://myrl@192.168.2.102?remote-store=ssh://192.168.2.101 seems to actually work.
<jackdk>
marek: look at the configuration-whatever.nix files in that directory, in particular for calls to `dontCheck`
<rnhmjoj>
marek: look in configuration-common.nix, there's a dontCheck function
<Miyu-chan>
Actually, scratch that. It says "too many authentication failures."
<__monty__>
Is there a way to get rid of this? "warning: Nix search path entry '/nix/var/nix/profiles/per-user/root/channels' does not exist, ignoring"
<__monty__>
I'm on a single-user install so I know root doesn't have any channels.
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zupo has joined #nixos
sach has joined #nixos
sach has quit [Client Quit]
Havvy has quit [Read error: Connection reset by peer]
sach has joined #nixos
fendor has joined #nixos
Thra11 has quit [Ping timeout: 246 seconds]
kumikumi has quit [Ping timeout: 245 seconds]
mix has joined #nixos
trfl has joined #nixos
aveltras_ has joined #nixos
hoijui has quit [Ping timeout: 264 seconds]
alex`` has quit [Ping timeout: 246 seconds]
bakakuna has joined #nixos
alex`` has joined #nixos
mix has quit [Read error: Connection reset by peer]
tonic1 has joined #nixos
<NfixEstrada[m]>
How can I specify a remote builder in a NIxOps deployment? I'm trying to use NixOps to manage a RPI, which has an aarch64 architecture
jasongrossman has quit [Quit: ERC (IRC client for Emacs 26.2)]
<ma27[m]>
`nixops deploy --builders 'aarch64-builder'` with `aarch64-builder` being a remote builder defined in `/etc/nix/machines`?
<ma27[m]>
first of all, how do you specify a remote builder in the Nix config?
<ma27[m]>
and are there any errors when running that command?
<Miyu-chan>
Actually yeah, I'm probably correct. Nix by default does distributed building, and upon distributed building, it will look for a machine with the suitable architecture.
<Miyu-chan>
So what you need to do is override the machine's `system/localSystem`, forgot the exact option.
<NfixEstrada[m]>
ma27: if I execute `nix ping-store --store ssh://ip
betawaffle has joined #nixos
<NfixEstrada[m]>
I have both `nixpkgs.system = "aarch64-linux"` and `nixpkgs.crossSystem.system = "aarch64-linux"` set in the deployment file, but it doesn't work neither
<Miyu-chan>
What's the error?
<NfixEstrada[m]>
`error: a 'aarch64-linux' with features {} is required to build '/nix/store/fhy67ly3jj2wdhs2kyb25amz7cbw4lqi-offline.drv', but I am a 'x86_64-linux' with features {benchmark, big-parallel, kvm, nixos-test}
<Miyu-chan>
cinimod: overlays are too general. packageOverrides are there specifically for this.
<NfixEstrada[m]>
How can I SSH to it as with my user's keys?
<ma27[m]>
cinimod: IIRC we use stackage packages and import those regularly with an automated tool. Until that time it's probably the best to override the package in an overlay
mexisme has quit [Ping timeout: 250 seconds]
<Miyu-chan>
NfixEstrada[m]: I'd suggest against that.
<NfixEstrada[m]>
Miyu-chan: that's what I thought xD
<Miyu-chan>
The nice thing about SSH keys is that being asymmetric is that you can have one for each user/machine.
<Miyu-chan>
Just keygen for root, and be done with it. :)
<NfixEstrada[m]>
Miyu-chan: It seems to be working! At least it doesn't show any error! Is there any way to write this option into the NixOps deployment file?
<cinimod>
nix picks up random-fu-multivariate 0.1.2.0
<cinimod>
I'd like 0.1.2.1
<ma27[m]>
so, you're now able to build aarch64 stuff with the builder which is explicitly specified in the CLI?
<cinimod>
Can I do this on the command line or do I have to have a shell.nix or similar?
<NfixEstrada[m]>
ma27: It seems to. It's not throwing the error that was throwing before
<ma27[m]>
not sure if this was asked before, but what's in `/etc/nix/machines`?
<NfixEstrada[m]>
It doesn't exist
<ma27[m]>
because after skimming through the nixops sources, I'm pretty sure that Miyu-chan is right, Nix should be able to automatically use a builder for the appropriate arch
<cinimod>
But if I say nix-build -I nixpkgs=https://github.com/NixOS/nixpkgs-channels/archive/nixos-19.03.tar.gz release-8.6.nix --arg packages "haskellPackages: [ haskellPackages.random-fu haskellPackages.my-random-fu-multivariate ]"
<cinimod>
then I get
<cinimod>
error: attribute 'my-random-fu-multivariate' missing, at (string):1:46
<cinimod>
<cinimod>
(use '--show-trace' to show detailed location information)
<cinimod>
Why is nix-build not picking up my overlay?
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
hamishmack has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
bakakuna has joined #nixos
zupo has joined #nixos
<infinisil>
cinimod: Depends on what release-8.6.nix contains
<cinimod>
Ah ok - I had been using something like nix-shell -I nixpkgs=/Users/dom/nixpkgs -p python37Packages.myScikits-odes -p python37Packages.nose --run "python3" and then my overlays got picked up
<cinimod>
I don't really want to monkey with release-8.6.nix
<cinimod>
How would I use my overlays and release-8.6.nix?
<mkaito>
tazjin andi- : never got that email, but I managed to get diesel_cli to compile. turns out, I'm just too dumb to interpret linker errors. it wanted zlib and openssl libs and is now happy. it also required triple-wrapping the derivation to remove the package from the workspace and provide a custom Cargo.lock, but that's par for the course. I'll put it on github later.
sach has quit [Quit: Quitting]
sach has joined #nixos
aexoxea has quit [Ping timeout: 264 seconds]
<cinimod>
I thought overlays were supposed to be composable?
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
aexoxea has joined #nixos
bakakuna has quit [Ping timeout: 246 seconds]
Shouou has quit [Ping timeout: 276 seconds]
<mkaito>
aren't they?
<cinimod>
If I say But if I say nix-build -I nixpkgs=https://github.com/NixOS/nixpkgs-channels/archive/nixos-19.03.tar.gz release-8.6.nix --arg packages "haskellPackages: [ haskellPackages.random-fu haskellPackages.my-random-fu-multivariate ]" then they don't seem to be used at all
revtintin has quit [Quit: WeeChat 1.9.1]
bakakuna has joined #nixos
<cinimod>
OTOH if I say nix-shell -I nixpkgs=https://github.com/NixOS/nixpkgs-channels/archive/nixos-19.03.tar.gz -p python37Packages.myScikits-odes -p python37Packages.nose -p haskellPackages.random-fu -p haskellPackages.my-random-fu-multivariate then the overlays *are* used
Soo_Slow has joined #nixos
<cinimod>
Is there a way of using release-8.6 *and* my overlays?
<infinisil>
cinimod: I still have no idea what release-8.6.nix is. Whether overlays get used or not depends on your nix expressions
<cinimod>
infinisil: I apologise - my skill level with nix is very low so I am painfully aware my questions may not make sense
<mkaito>
doesn't help that overlays aren't exactly intuitive
<cinimod>
I think that just calls something else though
<cinimod>
I suppose I thought overlays *always got used before anything else happened
<Mateon1>
Hi, is it possible to pass an optional build flag with `nix-env`? I want to install ripgrep with PCRE support, the package has a `withPCRE2` boolean flag which defaults to false.
betaboon has joined #nixos
<infinisil>
cinimod: Ah I see the problem now that I know what's in the file: Your overlay changes haskellPackages, but the file uses (indirectly) haskell.packages.ghc864, which doesn't have the overlay
<infinisil>
cinimod: You can apply your overlay to haskell.packages.ghc864 with `self: super: { haskell = super.haskell // { packages = super.haskell.packages // { ghc864 = super.haskell.packages.ghc864.override { overrides = <...>; }; }; }; }`
<infinisil>
(A bit annoying but that's how it works)
<__monty__>
I'm not sure if it works, it does with nix-shell though.
<infinisil>
__monty__: Unfortunately that won't persist for updates
<__monty__>
Ah, then I guess an overlay would be a better solution?
<__monty__>
,overlays
<{^_^}>
__monty__: Did you mean overlay?
<{^_^}>
Overlays look like `self: super: { foo = ...; }`. Use the self argument to get dependencies, super for overriding things and library functions (including callPackage). More info: https://nixos.org/nixpkgs/manual/#sec-overlays-install
<Mateon1>
__monty__: Doesn't seem that nix-env -A accepts arbitrary nix expressions sadly
<Mateon1>
error: attribute 'override { withPCRE2 = true; }' in selection path 'nixos.ripgrep.override { withPCRE2 = true; }' not found
<__monty__>
Mateon1: Yeah, just go with an overlay.
<infinisil>
It would be -E (instead of -A) to accept an expression, but I recommend an overlay instead
<__monty__>
,declarative
<{^_^}>
There are multiple ways of managing declarative profiles. 1) Attrset, compatible with imperative use of nix-env https://git.io/fAQHW ; 2) buildEnv, providing more control over the paths that are linked into the profile https://git.io/fp0aU ; 3) home-manager, providing nixos-like config for your ~ https://github.com/rycee/home-manager
<__monty__>
Mateon1: I also recommend this ^ approach over nix-env -i.
<__monty__>
Well, one of these approaches : )
<Mic92>
Can I override the default value for a NixOS module option? I want to a fixed value for services.nginx.virtualHosts.*.listen.addr
<Mic92>
I suppose I need to iterate over all virtualHosts instead.
bakakuna has quit [Ping timeout: 248 seconds]
<infinisil>
Mic92: Yeah you can but it's a bit cumbersome. You need to redeclare the option in your module with the submodule type, and then set the default in that submodule
rnhmjoj has joined #nixos
rnhmjoj has quit [Changing host]
rnhmjoj has joined #nixos
<infinisil>
I think I have an example lying around somewhere
<Mateon1>
I'm probably doing something stupid, but why is the stuff installed via overlay using the root user's nixos channel, instead of my user's nixos channel (unstable)?
<cinimod>
error: attempt to call something which is not a function but a set, at /Users/dom/.config/nixpkgs/overlays/default.nix:22:13
<Mic92>
infinisil: actually, when I try this, then it conflicts with the other definition.
<infinisil>
Mic92: You need to use the same type, so `attrsOf (submodule` in case of virtualHosts
<Mateon1>
I ran the `nix-env -f '<nixpkgs>' -iA userPackages` command, where userPackages is the name I defined in the overlay file, as my normal user, but it started installing old packages from the nixos-19.03 channel (which my root user uses as the main `nixos` channel)
<tilpner>
Mateon1: Your NIX_PATH may define nixpkgs to point to roots channel entry nixpkgs
<tilpner>
echo $NIX_PATH
<infinisil>
cinimod: You forgot the //
<infinisil>
cinimod: Take a look at the overlay i wrote earlier again
<Mateon1>
tilpner: That returns "/nix/var/nix/profiles/per-user/root/channels/nixos/nixpkgs", my NIX_PATH does in fact point there. Can I configure this per-user from the configuration.nix, or should I add a line to .profile/.bashrc?
<cinimod>
infinisil: oops - I thought they were line breaks
<tilpner>
Mateon1: You can set nix.nixPath, but it's not that great
<infinisil>
Hehe
<tilpner>
Mateon1: Setting that option will still require you to login again for changes to NIX_PATH to take effect
<cinimod>
Some sort of concatenation operator?
<tilpner>
Mateon1: But you can set NIX_PATH anywhere you like, including the bash files you mentioned
<tilpner>
Mateon1: You can also avoid mentioning <nixpkgs> entirely and continue using your own users channels
<infinisil>
> { a = 10; } // { b = 20; } # cinimod
<{^_^}>
{ a = 10; b = 20; }
<infinisil>
It overrides attributes in an attribute set
<Mic92>
infinisil: error: The option value `services.nginx.virtualHosts.thalheim.io.listen' in `/home/joerg/.homesick/repos/dotfiles/nixos/vms/eve/modules/sslh.nix' is not of type `list of submodules'.
<infinisil>
Mic92: I said you need to set the default in the config section of the submodule. Assigning the options default won't work
<Mic92>
ah
<infinisil>
config.listen = [ { addr = ...
<Mic92>
so config.listen?
<infinisil>
Yup
<Mic92>
I was not aware of such a thing
<Mic92>
nifty
<cinimod>
infinisil: great - I now have my fixed package working in IHaskell via nix
<infinisil>
Well actually it might work, I'm not entirely sure
<Mateon1>
I get everything the same up to "post-installation fixup", and then I get: cp: missing destination file operand after '/nix/store/489yclal5bn9z8y4l85i6qxc22cliwjz-ripgrep-11.0.1/man/man1/'
<{^_^}>
#48762 (by unode, 39 weeks ago, open): Add support for burp client/server services
freedan42x has quit [Ping timeout: 260 seconds]
pie_ has quit [Ping timeout: 276 seconds]
pie_ has joined #nixos
pie_ has quit [Remote host closed the connection]
pie_ has joined #nixos
pie_ has quit [Remote host closed the connection]
pie_ has joined #nixos
bgamari has quit [Ping timeout: 250 seconds]
hoijui has quit [Quit: Leaving]
__monty__ has quit [Ping timeout: 248 seconds]
magnetophon has joined #nixos
freedan42x has joined #nixos
aveltras_ has quit []
<freedan42x>
I have problems with wireless connection. Can someone help?
aveltras_ has joined #nixos
<rnhmjoj>
freedan42x: what's the problem?
<freedan42x>
I can't connect to my wifi. I also have "boot.extraModulePackages = [ config.boot.kernelPackages.broadcom_sta ];", but it does not seem to help
<rnhmjoj>
i don't know exactly how network manager works (never used it), but you should use a gui or nmcli to add a wireless network and connect to it. otherwise, which is what i normally do, you can add the networks ine the nixos configuration, like this: file:///nix/store/lb0rlr5pybvsfsyarissxglv1h4zs9w8-nixos-manual-html/share/doc/nixos/index.html
<symphorien>
oh indeed it had already been asked, sorry
<freedan42x>
interesting thing I can see other wifi networks, but can't connect to mine. now they said "secrets were required, but not provided" even though if I typed the passwrod
<symphorien>
do you have a gui ?
<symphorien>
or are you stuck in a tty ?
<freedan42x>
I have a gui, yes
<freedan42x>
(no idea what is tty)
<rnhmjoj>
freedan42x: any particular reason why you installed network manager?
<freedan42x>
someone said to do that
<symphorien>
then make sure you have something like gnome-keyring running
<symphorien>
if you run gnome or kde this should be the case
<freedan42x>
could it be the problem that I don't have "wireless.enable = true" at configuration file?
<freedan42x>
I run xfce
<jonge[m]>
hi there. does anybody know how to solve this: i have a nix expression that is evaluated on hydra (https://github.com/tfc/nur-packages). This does not only contain derivations but also a function (wrapCC) that i want to provide to users of this expression. now i got my hydra run on this and it build everything nicely, but chokes on that wrapCC function because it is obviously a <LAMBDA> and not a derivation. how can i make
<jonge[m]>
hydra simply ignore non-derivations in order to be able to provide functions and derivations?
<symphorien>
freedan42x: wireless.enable is incomatible with network-manager
<freedan42x>
yeah, noticed that
<symphorien>
it should be called wpa_supplicant.enable
<symphorien>
anything relevant in sudo journalctl -u network-manager ?
<freedan42x>
"receive_packet failed on enp0s16f0u1: Network is down"
<symphorien>
this is for your ethernet connection, not the wireless one
<rnhmjoj>
if nm and wpa_supplicant are incompatible enabling both should raise an error, we could add an assertion maybe. i'll open an issue
<symphorien>
I thought there was one ?
<rnhmjoj>
well, freedan42x could build a configuration like this, so apparently no
<freedan42x>
I noticed that I have wifi issues on Ubuntu too, but fixed it by installing recommended drivers from Software & Updates. Feeling like my I would have errors on every distro, but on Windows it's all fine
mexisme has joined #nixos
<rnhmjoj>
freedan42x: to rule out a driver issue i would, for the moment, disable network manager and set 'configuration.wireless.networks.<yournetworkname>.psk="<yourpassword>";' and 'networking.wireless.enable = true;'
<nschoe>
Hi everyone, I have a quick question: in nixOS, should we have python3 installed in the system's profile?
<nschoe>
Or is it like libraries, where they should never be installed "system-wide", only used as dependencies for building derivations. Is it the same with python3 or not?
<freedan42x>
rnhmjoj: "The option `configuration' defined in `/etc/nixos/configuration.nix' does not exist."
<freedan42x>
rnhmjoj: is `"free.wifi" = {};` necessary to add?
<rnhmjoj>
no, that's an example for a network without password
<symphorien>
nschoe: system-wide python will work (contrary to c libraries)
<nschoe>
symphorien, thanks, so it's "okay" to have it.
<symphorien>
yes
<nschoe>
Is it "the nixOS way", though?
<nschoe>
Should I do it?
<symphorien>
why not ?
<nschoe>
I realize it's a broad question which depends on my use-case, but what I mean is: do msot nixOS users have it? Is it "normal" to have it system-wide, or does it mean there's something fishy?
<symphorien>
I have a python with bpython installed globally
<symphorien>
fwiw
<freedan42x>
rnhmjoj: so what's next? I changed the config file and switched to new configuration
<nschoe>
symphorien, okay, thanks for your answer.
Guest3466 has quit [Quit: WeeChat 2.4]
<symphorien>
nschoe: note that installing python packages globally does not work
anon has joined #nixos
anon is now known as Guest86285
Cale_ is now known as Cale
<nschoe>
symphorien, okay good to know.
<nschoe>
symphorien, Actually I don't think I'll really need python globally, but I was asking as a "last resort". I'm tryign this i3wn themer: https://github.com/unix121/i3wm-themer and it requires python3. But from what I can read from the docs, its _looks_ like python is only used to swithc between themes, if that's the case, I just have to fire a python env from nix-shell when/if needed.
<boogiewoogie>
hey, is anyone else getting address boundary errors after starting the teamspeak3 client?
<nschoe>
But in case I'm wrong and I definitely needs python I wanted to know if it was okay to have it installed .
<nschoe>
symphorien, wow awesome, thanks a lot for that. I'm going to try this then!
<fendor>
are nix pills still up to date and good to learn how to package projects?
kumikumi has joined #nixos
<symphorien>
they are good to learn what happens under the hood
<m1cr0man>
Hey folks. Anyone used moretea/yarn2nix recently? I'm getting an error "mv: cannot copy a directory, '/build', into itself, '/build/temp'" at the configure stage. Full error: https://u.m1cr0man.com/sd/nx4
<symphorien>
but for state of the art packaging in nixpkgs, less so, I would say
<boogiewoogie>
or, different question: is teamspeak3 working fine for anyone here? or is it a problem with the package?
<fendor>
symphorien, what would be more suited?
<rnhmjoj>
freedan42x: that's it. you should be connected
<symphorien>
fendor: I don't know, maybe looking in nixpkgs for similar applications ?
<nschoe>
Sorry to bother again, in a default.nix file, it fails to get the source source (forever hangs). I successfully `wget` the .tar.gz file, is there a way to add is to the nix store so that when I try to install the package (or nix-shell -p) it it detects it and won't try to fetch it from network?
<fendor>
symphorien, those are normally quite convoluted, imo not suitable for learning if I have almost no idea about it
<dmj`>
whats the best way to conditionally combine sets in nix
<symphorien>
fendor: what I meant is: nix pills will teach you how to package things. But nixpkgs has all sorts of niceties which are not covered by nix pills
<symphorien>
for example nix pills introduce custom builders, but they are nearly never necessary
<symphorien>
so nix pills are a good teaching material worth reading, but not representative of how you packages things for nixpkgs
<dmj`>
optionalAttrs
<hodapp>
how can I get a CMake build which needs Qt 5.x to stop trying to use FindQt4.cmake? In their 'official' Travis build they set CMAKE_PREFIX_PATH=/usr/include/x86_64-linux-gnu/qt5 but I'm not having any success setting it to ${qt5.full}, or that plus /lib, plus /lib/cmake...
<fendor>
symphorien, is there something more representative?
<symphorien>
not that I know
<symphorien>
the nixpkgs manual has language specific sections though
wfranzini has quit [Remote host closed the connection]
<symphorien>
maybe go there first
wfranzini has joined #nixos
<fendor>
I am there, it still is hard for me to see how to combine them, into overlays for example. Will try though, thanks!
gentauro has quit [Ping timeout: 244 seconds]
boogiewoogie has quit [Quit: Leaving]
mac10688 has joined #nixos
gentauro has joined #nixos
kumikumi has quit [Ping timeout: 272 seconds]
fusion809 has quit [Remote host closed the connection]
kl3 has joined #nixos
gentauro has quit [Ping timeout: 258 seconds]
<kl3>
hi. running unstable, added a patch to a package by amending default.nix with "patches = [ ./foo.patch ];" and dropping the file there. 'nixos-rebuild -I nixpkgs=~/nixpkgs switch' picks it up, it applies, but the compiler now fails to find <sys/capability.h> which I am including in my patch. searched the docs but cannot find out how to get the compiler see this header
<kl3>
talking tools/networking/openfortivpn here, is there some option i have to set so kernel headers are included in the build environment?
GiGa has joined #nixos
gentauro has joined #nixos
shomodj has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<{^_^}>
[nixpkgs] @hauleth opened pull request #65233 → Allow overriding SSL library in Erlang → https://git.io/fjMoN
noudle has quit []
mkoenig has quit [Remote host closed the connection]
sonofentropy has joined #nixos
hmpffff has joined #nixos
mkoenig has joined #nixos
<sonofentropy>
hi, is there a way to use environment.systemPackages with nixos/nix docker image? Better yet, check this: https://gist.github.com/0o153/09cacad3c90f75090b5c690e02d1e862 — what would be the best way to achieve extended nixos/nix image with preinstalled packages?
stepcut has joined #nixos
mmlb has joined #nixos
justanotheruser has quit [Ping timeout: 248 seconds]
<aminechikhaoui>
tilpner I asked whether one could subscribe to get notifications about issues with security label. Could be easy to setup, just didn't see how to do that from GH
b has joined #nixos
MmeQuignon has joined #nixos
astrofog has joined #nixos
<tilpner>
aminechikhaoui: I commented on that during the call, but there was no response. How would you expect to get notifications?
<tilpner>
An RSS feed perhaps?
shomodj has joined #nixos
<aminechikhaoui>
ideally through email, but yeah rss could be useful as well
<tilpner>
Scraping nixpkgs issues and gathering feeds for each label should be doable, and scales without being abusable
<tilpner>
Email is a few more moving parts, not sure I'd want that
<samueldr>
users that want e-mail could use rss2email themselves I guess
astrofog has quit [Remote host closed the connection]
phreedom has quit [Ping timeout: 260 seconds]
markus1189 has joined #nixos
<Miyu-chan>
I'm thinking NixOS people might know something like this.
__monty__ has quit [Ping timeout: 245 seconds]
<Miyu-chan>
Vim ex mode basically allows you to run vim commands(except it's line-oriented) in batch.
<Miyu-chan>
I need something like ex mode, but is character-oriented.
enuiarte has quit [Ping timeout: 264 seconds]
<Miyu-chan>
I want to make a script along the lines of `/\[foo\]; /key; df=; s/^ *//; w`
<Miyu-chan>
That parses an ini file. Look for a section, goto a key, delete the key, clean up the whitespace.
<samueldr>
I might be wrong, but isn't ex mode mostly ed, and sed scripted ed? maybe sed doesn't have the features you need though
justanotheruser has joined #nixos
<Miyu-chan>
samueldr: sed is more like stream ed.
<samueldr>
right
<Miyu-chan>
And yeah, awk, sed, are the first things I actually considered. :)
<samueldr>
though, what I meant from that is I think sed's commands are based on ed's
<samueldr>
and ex's on ed's
o1lo01ol1o has quit [Remote host closed the connection]
<Miyu-chan>
I'd say they're pretty close. But sed is more of a huge-ass switch/case statement.
<samueldr>
though I can see how diverging early one vim's ex mode could have additional features
<Miyu-chan>
That's why I want something like ex, because it's simply textual.
<Miyu-chan>
While sed is more like "if the current line matches the pattern, then run this command."
abathur has joined #nixos
<samueldr>
and using vim itself to do it wouldn't work?
<samueldr>
I think vim's ex mode can be used unattended
<Miyu-chan>
Yeah, ex mode can be used in batch, and it's actually pretty close to what I want.
<samueldr>
oh, re-read you and yeah, "but is character-oriented"
__monty__ has joined #nixos
<Miyu-chan>
It's a pretty cool idea, don't you thinK?
<Miyu-chan>
nix run nixpkgs.vim -c ex +'/\[miyu\]' +':mark a' +'/^aws_access_key_id' +'s/aws_access_key_id *= *//' +'.w! /dev/stderr' +'q' -R ~/dummy_credentials 2>&1 1>/dev/null
<Miyu-chan>
miyu
day|flip has joined #nixos
Cale has quit [Ping timeout: 276 seconds]
<{^_^}>
[nixpkgs] @suhr opened pull request #65236 → klystrack: init at 1.7.6 → https://git.io/fjMKX
<{^_^}>
[nixpkgs] @danbst merged pull request #61581 → postgresql: Make systemd support configurable → https://git.io/fjlXO
<{^_^}>
[nixpkgs] @danbst pushed commit from @nh2 to master « postgresql: Make systemd support configurable. See #61580. (#61581) »: https://git.io/fjMi8
pheoxy has quit [Quit: Leaving]
sonofentropy has quit [Quit: sonofentropy]
o1lo01ol1o has joined #nixos
Jackneill has joined #nixos
mexisme has joined #nixos
o1lo01ol1o has quit [Ping timeout: 245 seconds]
<{^_^}>
[nixpkgs] @suhr opened pull request #65238 → helio-workstation: init at 2.2 → https://git.io/fjMi1
psyanticy has quit [Quit: Connection closed for inactivity]
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
slack1256 has joined #nixos
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
bakakuna has joined #nixos
sushi_ is now known as ilikeheaps
alex`` has quit [Ping timeout: 258 seconds]
phreedom has joined #nixos
alex`` has joined #nixos
elux has joined #nixos
enuiarte has joined #nixos
<elux>
hey all
<elux>
does anyone here use nixos with kubernetes? id be interested in using it to config the cluster. ive been looking at a project called 'gitops' as well, but its so marketing heavy and light on the examples, seems salesy
<simpson>
It's alright. Definitely better than hand-writing YAML as soon as any templating is required. Not sure if I'd prefer it over jq.
<elux>
neat this looks good. i completely agree nix should have a significant place in this cloud cluster world of orchestration.
<elux>
jq? as in the json tool..? or something else
rsoeldner has quit [Remote host closed the connection]
<simpson>
Yeah, the JSON tool. k8s uses JSON under the hood, not YAML, and there's nothing inherently wrong about submitting JSON to kubectl.
<gchristensen>
(since json is yaml..)
<elux>
im trying to find the ideal setup for a CD process to declare the cluster and the apps within it, and just have it declaratively get the cluster to that state. similar to what nix does for my local system, i want to do for a network.
<elux>
this is why im looking at GKE+linkerd2+gitops, but -- i still wonder a place for nix as its so powerful
<simpson>
elux: Not gonna lie, I've been lately researching Spinnaker for my dayjob. Very complex, but I haven't seen anything that's simpler and still does what we need.
<simpson>
Also not gonna lie that k8s does the job *very* well when it comes to declaring the cluster's state.
bakakuna has quit [Ping timeout: 248 seconds]
waleee-cl has quit [Quit: Connection closed for inactivity]
___laika has joined #nixos
<hpfr[m]>
How can I reference $HOME within a path type in Nix
<clever>
hpfr[m]: ~/foo
<elux>
simpson: what about deploying apps to k8s..? what do you use
phreedom has quit [Ping timeout: 260 seconds]
knupfer has joined #nixos
<gchristensen>
does anyone setup wireguard to forward 0.0.0.0/0 ? I understand it doesn't quite work right, and would be intersted in workarounds
<hpfr[m]>
clever: ... oh. Was trying all kinds of “${$HOME}/foo”
<exarkun>
aanderse: maybe that means the module system should work differently :)
<clever>
gchristensen: i need to go over your wireguard changes, and update tgt_service, it has similar bugs to what your wireguard recently had (which got fixed)
<aanderse>
exarkun: or it means containers are awesome :shrugs:
<exarkun>
aanderse: there are lots of things that nixos containers aren't quite the right fit for
<aanderse>
i don't generally see arch or debian offering multi instance services, not sure why nixos should be so different in this regards
<gchristensen>
aanderse++
<{^_^}>
aanderse's karma got increased to 11
justanotheruser has joined #nixos
<gchristensen>
and you could reimplement the module to do whatever you want, without much fuss
<exarkun>
arch and debian don't offer functional, declarative system configuration either :p maybe nixos doesn't need to bother with that anymore
<gchristensen>
the build it modules cover the 80%. for the other 20%, it is time to go it alone
<gchristensen>
s/the build it//
<exarkun>
look my ideas might be stupid and my thinking totally out of whack. I'm not interested in arguing about whether nixos should be different or not. I'm interested in understanding the pros and cons of the current system.
<gchristensen>
clever: did you see my questions about the 0.0.0.0/0 forward?
<clever>
gchristensen: yep, havent done much with changing the default route
<gchristensen>
I don't think anyone thinks what you're saying is stupid
<clever>
gchristensen: the main prolem is adding a route override, to force the wg peer to not go over wf
<clever>
wg*
<aanderse>
exarkun: i've made a module which is multi instanced and it is nixos upstream... so i'm not arguing from a "you can't do that!" perspective ;
<gchristensen>
the modules shipped with nixos cover a lot of use cases, but not all of them
<aanderse>
i'm just saying... since i have done it i saw how much complexity it did add, so i wonder if containers are a *better* solution
<gchristensen>
clever: hmmm and I think it is a challenge because of policy routing?
<aanderse>
i've never really looked at the performance impact of nixos containers in detail, but i have run 2 containers + a host os on a bottom of the line linode server and it performs fine for a modest site
<aanderse>
so it makes me think nixos container overhead is relatively minimal
<gchristensen>
I'd like to see a NixOS ModuleExchange thing lik ethere are for chef cooboks
<exarkun>
aanderse: Containers are probably better sometimes. But if a module needs the host network and the host filesystem and the host process table and the host xyz ...?
<clever>
gchristensen: wireguard would need to dynamically figure out the remote ip, and add a route entry at runtime, to ensure it still goes via the gateway
<aanderse>
exarkun: making those available is possible. but yes, you need to balance what is "better" and "easier"
<exarkun>
aanderse: The abstraction you get with a container can reduce complexity and that's great but having to re-glue all of the pieces back to the host can increase complexity and that's not so great.
<gchristensen>
an important thitng is that the base modules shipped with nixos cannot be all things to all people
<exarkun>
gchristensen: Okay but I'm not arguing for that
<gchristensen>
I know :) so it is expected to need to "do it yourself" for some things
<infinisil>
Though I really wonder. What if we had a sandbox abstraction in NixOS that prevents service modules from modifying anything outside their playground
<exarkun>
aanderse: I see how allowing multi-instances might be complicated _now_, considering the way modules are defined (where they write into a system-global namespace). I don't think that means supporting multiple instances is inherently more complex.
<infinisil>
So they couldn't write files to /etc/
<infinisil>
And they couldn't choose their own port probably
<gchristensen>
infinisil: adopting POLA for modules would be cool
<exarkun>
infinisil: right. they would need to accept parameters for everything.
<aanderse>
exarkun: totally agreed
<infinisil>
exarkun: Well hopefully not, it would be nice to have in a transparent way, while getting an error if it would violate that
<exarkun>
"it's a big change from what exists now" might be a good reason not to pursue it. I wonder if there are other good reasons, though
<exarkun>
infinisil: hmm maybe. "transparent" sometimes looks a lot like "implicit" and that's not always better
mexisme has joined #nixos
<exarkun>
unfortunately, gotta go
<infinisil>
exarkun: I think implicitly having secure defaults should be alright
<gchristensen>
infinisil: isn't talking about parameterization, but privilege
<gchristensen>
clever: I think ther eis a problem with policy routing which breaks it, even if I know the IP at run time
<gchristensen>
clever: but I'm quite novice with network :)
<clever>
gchristensen: there is also a kernel feature android uses, but ive not looked into yet, it can support multiple routing tables
<gchristensen>
:o
<clever>
gchristensen: so you can have several sets of routes, and multiple default gateways, then flag a socket to use a certain one
<clever>
the android vpn interface, requires that you run a special function on the FD# for the upstream connection
<clever>
and the OS will specially flag it to ignore the default route, which goes into the /dev/tun handle you are given
<clever>
then you just read packets from that /dev/tun handle, and write them out the FD# you specially flagged
justanotheruser has quit [Ping timeout: 264 seconds]
<clever>
and at no point do you get root, so you can run any vpn you want on android, securely
<clever>
basically, nothing refers to Paths_HTTP.HTTP_bindir, so just drop the entire value when linking
chris| has joined #nixos
<clever>
samueldr: how did that work again...
<samueldr>
clever: that's... what I would like to know :)
<samueldr>
(trying to slim my closure size)
fendor has quit [Read error: Connection reset by peer]
johnny101m has joined #nixos
dansho has joined #nixos
tonic1 has quit [Ping timeout: 252 seconds]
mexisme has joined #nixos
mkoenig has quit [Remote host closed the connection]
shomodj has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
shomodj has joined #nixos
shomodj has quit [Client Quit]
mkoenig has joined #nixos
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
logzet__ has quit [Remote host closed the connection]
Cale has quit [Remote host closed the connection]
mexisme has quit [Ping timeout: 258 seconds]
shomodj has joined #nixos
Guest86285 has quit [Quit: WeeChat 2.4]
mexisme has joined #nixos
webster23 has quit [Ping timeout: 250 seconds]
Cale has joined #nixos
maralorn has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
rprije has joined #nixos
gratto has quit [Read error: Connection reset by peer]
gratto has joined #nixos
<maralorn>
Hey everyone: I have a question regarding the haskell infrastructure in nixpkgs. From using haskell with nixpkgs I get the impression, that nixpkgs just includes a hackage snapshot of all packages and lets them depend on each other. So it "prevents" cabal-hell by not even checking dependencies. So when a package works with all dependencys being the latest version on hackage its fine. Otherwise it's broken. Did
<maralorn>
I understand that correctly? If I'm wrong can you correct me or point me to some resources?
o1lo01ol1o has quit [Ping timeout: 248 seconds]
<jackdk>
I think it takes as a base snapshot one of the stackage snapshots, but then has to add packages on hackage that aren't part of any stackage snapshot, plus a bunch of other tweaks in the configuration-*.nix files
<maralorn>
aaah
<maralorn>
That sounds roughly like I understood it but not nearly as insane as I explained.
<maralorn>
Does hydra try to build packages marked es broken?
selfsymmetric-mu has joined #nixos
<maralorn>
I am trying to figure out why haskellPackages.shh is broken but I can‘t find the build job on hydra.
<samueldr>
it does not
<samueldr>
to figure that out git blame is likely to help
<samueldr>
at least it'll show when
<samueldr>
I think that haskellPackages though the brokens are added automatically, not sure what are the conditions
<aveltras>
anyone using the lightdm enso greeter ?
<jackdk>
maralorn: if i find something that works, I usually make a PR like #61536 ; some additional context in #61593 about why things are the way they are
<jackdk>
is this in your nixos config, or in configuration for a single project?
<maralorn>
jackdk: It's in a home-manager config.
<jackdk>
maralorn: so you have some binary from haskellPackages you want in $PATH?
ng0 has quit [Quit: Alexa, when is the end of world?]
Rusty1 has joined #nixos
mbrgm_ has joined #nixos
<maralorn>
jackdk: No, I am playing around with using the writeHaskellBin writer and want to use haskellPackages.shh as a dependency.
mbrgm has quit [Ping timeout: 264 seconds]
mbrgm_ is now known as mbrgm
realrokka has joined #nixos
abathur has quit [Ping timeout: 245 seconds]
<jackdk>
maralorn: and you are passing in haskellPackages.shh to the libraries list that is passed to writeHaskellBin? Then I think you could probably call haskell.lib.overrideCabal on and unmark it as broken. If you need to do this in a bunch places, I'd construct a modified haskellPackages set like I did in that link I sent you.
<maralorn>
jackdk: Yeah, thanks. I am doing exactly what you're suggesting right now.