srl295 has quit [Quit: Connection closed for inactivity]
<inkbottle>
I don't understand that line: services.emacs.package = import /home/cassou/.emacs.d { pkgs = pkgs; }; It seems to say that the emacs user-directory on one specific user will be used for each and every users?
<inkbottle>
It is only a guess though; but 'cassou' seems like a specific user
<inkbottle>
but services.emacs.package doesn't seem to target one specific user.
qubasa has quit [Ping timeout: 265 seconds]
<samueldr>
import is the nix construct; /home/cassou/.emacs.d is "just a folder", in a way, where there is a default.nix
<samueldr>
and in that default.nix there is a nix expression that gets evaluated
timtam has joined #nixos
sigmundv has quit [Ping timeout: 265 seconds]
_reanimus_ has joined #nixos
reanimus has quit [Ping timeout: 276 seconds]
domogled has quit [Quit: domogled]
<inkbottle>
okay, I understand
<inkbottle>
thanks
o1lo01ol1o has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @veprbl pushed commit from @scalavision to master « delly: init at 0.8.1 (#70871) »: https://git.io/JelGl
<samueldr>
pkgs.runCommandNoCC "that-file" {} "cp ${./that-file} $out" would copy it directly
<mog>
i want to make use of the gpgCard feature thats part of 19.09, so it does a cp encryptedPass
<mog>
but if i put a file in it isnt visible when it tries to build
<mog>
so i am imagining i could fix if i just added the gpg encoded secret into the nix store and pointed to it in the config
dbmikus has quit [Ping timeout: 240 seconds]
drakonis1 has joined #nixos
drakonis_ has quit [Ping timeout: 250 seconds]
felixfoertsch has quit [Ping timeout: 250 seconds]
felixfoertsch has joined #nixos
durka has joined #nixos
work_ has joined #nixos
<durka>
Hey, I'm trying to figure out where NixOS stores the name of the root partition for booting. My root partition is encrypted, so it must be specified somewhere else. But I've mounted the EFI boot partition and I can't find it -- the /efi/loader/entries/nixos-*.conf file just gives paths in /nix/store
<iqubic>
That's what I'd expect.
<durka>
Is there an fstab in the initrd perhaps?
<durka>
I tried to mount the initrd, but I don't think I did it right
<durka>
I'm speaking to you from a live USB because I moved a partition which apparently caused renumbering and now I can't boot :p
kshi has joined #nixos
<mog>
doing ${path} solved my problem
<mog>
i think
<durka>
aha I found out how to open the initrd -- it's a gzipped cpio archive
<durka>
now the question is whether I can repack it as such after editing :p
boxscape has quit [Remote host closed the connection]
<selfsymmetric-mu>
durka: This is like watching Houdini trying to escape from a USB stick.
andi- has quit [Remote host closed the connection]
aw has quit [Quit: Quitting.]
spacefrogg has quit [Quit: Gone.]
mjsir911 has joined #nixos
aw has joined #nixos
spacefrogg has joined #nixos
andi- has joined #nixos
dansho_ has joined #nixos
minall has joined #nixos
<minall>
Hello nixos!
dansho has quit [Ping timeout: 265 seconds]
<durka>
/nix/store/daasdf908734643nasdlk/hello
<minall>
I'm on a new nixos installation, when starting the system, I get to the final part which is: -Starting Xorg Server-, it shows it as -succesful-
<minall>
But I don't seem to have a xorg server, since all I have next after that is a terminal, I'm on emacs -erc-
<minall>
But in terminal mode
ebzzry has quit [Ping timeout: 250 seconds]
ebzzry has joined #nixos
<minall>
When launching -ssdm- I'm stuck on -Logind interface found-
waleee-cl has quit [Quit: Connection closed for inactivity]
durka has quit [Remote host closed the connection]
<minall>
how can I debug this?
ebzzry has quit [Ping timeout: 268 seconds]
ebzzry has joined #nixos
<{^_^}>
[nixpkgs] @veprbl merged pull request #70841 → mariadb: fix builds on darwin → https://git.io/JeWrL
<GlennS>
Has something changed about where journald stores its logs from NixOs 19.03 to 19.09? `services.journalbeat` has stopped working, and I think it is because it can't find the journal files in the default place.
<ebzzry>
Is there a way to install nixpkgs on Ubuntu, unattended?
<tilpner>
ebzzry: What do you mean by that? nixpkgs is just a collection of packages and modules, it can be installed by downloading a single tarball to anywhere on your system
<tobi_>
hello, I'm a nixos newbie and I'm having trouble setting up pre-commit. In my environment.systemPackages I have gitAndTools.git and gitAndTools.pre-commit. installing pre-commit works fine, when I try to install hooks (pre-commit install-hooks) for a repository which uses python based hooks, pre-commit complains that it can't find virtualenv ("/nix/store/.../bin/python3.7: No module named virtualenv"). I
<tobi_>
have taken a look at the default.nix of pre-commit and it mentions virtualenv in the imports and in the propagateBuildInputs, so it looks like it should be available. Many thanks in advance
toppler has quit [Read error: Connection reset by peer]
Tucky has joined #nixos
toppler has joined #nixos
tobi_ is now known as tobi1234
hmpffff has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
asymptotically has joined #nixos
orivej has joined #nixos
ng0 has quit [Ping timeout: 260 seconds]
Boomerang has joined #nixos
bahamas has joined #nixos
ng0 has joined #nixos
Makaveli7 has joined #nixos
Diagon has quit [Quit: Leaving]
hyper_ch2 has joined #nixos
<bahamas>
does it make sense to run a development server with systemd? I'm talking about the development server you use when you develop a web app with a framework like Django or Yesod.
<bahamas>
I'm asking because I'm wondering whether I should make my dev setup match my prod setup as much as possible
<meatcar>
tobi_: newbie here as well, my understanding is that each package that uses python has a different set of python packages that need to be explicitly overridden. In my case weechat's python plugin needed an extra python package to function, and I had to throw together an overlay to import that in. https://github.com/meatcar/dots/blob/master/nix/.config/nixpkgs/home.nix#L32
<bahamas>
ivan: I'm assuming this setting will tell systemd to not do anything when the process restarts. just leave it be and don't try to restart it. right?
<mg->
Hey, anyone else have issues with light-locker after the 19.09 update? I get this padlock icon and "you will be redirected in a few seconds", but nothing happens. Have to kill -9 from a different tty :S
veske has quit [Client Quit]
jedai has quit [Client Quit]
<asymptotically>
mg-: did not get it before too? i get this issue on debian too
<asymptotically>
usually mashing keys and switching between ttys fixes it
<mg->
asymptotically: no before I had another issue where light-locker turned off all the screens and would sometimes refuse to turn back on. Always had to input pw "blind".
gila has quit [Quit: My Mac Pro has gone to sleep. ZZZzzz…]
<ptotter[m]>
level=error ts=2019-10-11T09:59:48.135Z caller=main.go:216 msg="Error loading config" err="error reading config file: open /nix/store/dwhypd9kgv96wbr1bc6sfgz735caph40-checked-blackbox-exporter.conf: no such file or directory"
<{^_^}>
[nixpkgs] @primeos pushed commit from @alexbakker to master « android-studio: add new emulator dependencies (#70857) »: https://git.io/Jel84
<__monty__>
Just grep `mkDefault =` : )
BinkyTheClown has quit [Quit: 30 day idle timeout.]
<meatcar>
thanks : )
eqyiel has quit [Quit: 30 day idle timeout.]
RobinKroonen[m] has quit [Quit: 30 day idle timeout.]
grahamc[m] has quit [Quit: 30 day idle timeout.]
b1000101 has joined #nixos
gspia[m] has left #nixos ["Kicked by @appservice-irc:matrix.org : 30 day idle timeout."]
hr[m] has left #nixos ["Kicked by @appservice-irc:matrix.org : 30 day idle timeout."]
iceychris[m] has quit [Quit: 30 day idle timeout.]
info_supplicant[ has quit [Quit: 30 day idle timeout.]
<{^_^}>
[nixpkgs] @BrianHicks opened pull request #70962 → Update elm nodepackages → https://git.io/Jel8g
<b1000101>
does someone know how precisely the helper program, that activates when you type a command which is not on your system, works? (The program ‘javac’ is currently not installed. It is provided by several packages. You can install it by typing one of the following: )
<b1000101>
I mean, how are the packages (which would javac provide) identified. The problem is, I get a list, but not a full one and it's a bit weird
klbrn[m] has quit [Quit: 30 day idle timeout.]
kyleondy[m] has quit [Quit: 30 day idle timeout.]
Lisanna[m] has quit [Quit: 30 day idle timeout.]
matthewbauer has quit [Quit: 30 day idle timeout.]
<__monty__>
Probably similar to
<__monty__>
,locate bin javac
<{^_^}>
Found in packages: zulu, zulu8, jetbrains.jdk, adoptopenjdk-bin, adoptopenjdk-openj9-bin-11
Moredread[m] has quit [Quit: 30 day idle timeout.]
gagbo has quit [Remote host closed the connection]
<infinisil>
And that only generates an index for binaries, not other files
gagbo has joined #nixos
Synthetica has joined #nixos
<infinisil>
The ,locate command uses https://github.com/bennofs/nix-index which can work with any nixpkgs version (not only the channels), and also with any type of file (not only binaries), but you'll have to generate the index yourself first (or use the ,locate command from the bot)
schmittlauch[m]1 has left #nixos ["Kicked by @appservice-irc:matrix.org : 30 day idle timeout."]
<b1000101>
infinisil: Yeah, I've tried looking at the code but I wasn't sure where it gets the data from. Tanks a million. Reason I am asking is because for example with javac, the most obvious package, nixos.jdk is missing from the list. So I was wondering why.
Shyim[m] has quit [Quit: 30 day idle timeout.]
<infinisil>
> jdk.meta.license
<{^_^}>
{ fullName = "GNU General Public License v2.0 only"; shortName = "gpl2"; spdxId = "GPL-2.0-only"; url = <CODE>; }
<infinisil>
Hm It's also missing from ,locate
<infinisil>
License isn't unfree, which could've explained it
<b1000101>
Yeah, I know about the unfree missing from the list, which is fine. It just struck me that javac helper program prints really odd options :)
<__monty__>
That *is* odd.
steshaw[m] has quit [Quit: 30 day idle timeout.]
<{^_^}>
[nixpkgs] @jtojnar merged pull request #70734 → nixos-rebuild: Fix target-host using fish → https://git.io/JeWYR
<freeman42x>
simpson, but what I am building is an executable. And it has all the other dependencies mentioned in `executableHaskellDepends`. I tried adding zlib there but it did not help
feep_ has joined #nixos
feep has quit [Disconnected by services]
hke has quit [Remote host closed the connection]
hke has joined #nixos
<freeman42x>
simpson, I tried: librarySystemDepends and it still gives missing zlib error
<freeman42x>
if I do nix-shell -p zlib then I can build successfully. I just need to know the equivalent shell.nix change to do that
<__monty__>
freeman42x: I'm almost certain you can just add `buildInputs = [ zlib ];` to the mkDerivation argument.
schjetne has quit [Ping timeout: 265 seconds]
<freeman42x>
__monty__, I will do it again just for you! :)
hyper_ch2 has joined #nixos
<freeman42x>
__monty__, here's the error for doing that:
<freeman42x>
error: anonymous function at /nix/var/nix/profiles/per-user/root/channels/nixos/pkgs/development/haskell-modules/generic-builder.nix:12:1 called with unexpected argument 'buildInputs', at /nix/var/nix/profiles/per-user/root/channels/nixos/lib/customisation.nix:69:12
<simpson>
freeman42x: I tried my modification locally and it "worked"; the entire build command doesn't work, but that's to be expected since you're not really building with Nix.
cinimod has joined #nixos
<__monty__>
freeman42x: Did you add it somewhere in the set passed to mkDerivation? Because it sounds like you put it in the argument list for the f function.
<das_j>
is there a convenient way to find the path of nixpkgs?
<__monty__>
No, that's correct.
<das_j>
because pkgs.path results in an infinite recursion
<freeman42x>
__monty__, then, it gives that error I mentioned above
<__monty__>
Are you passing the haskellPackages mkDerivation into f?
<__monty__>
Ah, you are.
wfranzini has quit [Ping timeout: 250 seconds]
alex`` has quit [Ping timeout: 240 seconds]
<ajs124>
anyone here doing typescript dev on nixos? tsc doesn't seem to take NODE_PATH, what's the best workaround for that?
<simpson>
freeman42x: https://gist.github.com/MostAwesomeDude/103666f1e38c4e7c9963698226379fb3 Next time please give actual instructions for reproduction. I ended up just copying stuff out of the manual. How does the build actually need zlib? I don't want to mess with Stack or other Haskell crap.
<simpson>
(Also I am laughing my ass off at "AGI". Hilarious.)
alex`` has joined #nixos
wfranzini has joined #nixos
<freeman42x>
simpson, <<I am laughing my ass off at "AGI". Hilarious.>> Why thank you! Aren't you a nice, likeable and open minded fellow!
<niksnut>
wedens: basically never since it doesn't capture anything other than configuration.nix
<niksnut>
so you can't use it to reproduce a system
b101010 has joined #nixos
<wedens>
btw, how do I use nixos-* commands from live usb on existing installation when configuration.nix is symlinked?
smatting has joined #nixos
<wedens>
I tried using nixos-rebuild from nixos-enter, but it seems to have some issues with overlays (even though I have overlays-compat hack in nix.nixPath)
<wedens>
can I use nixops stateless if I want to provision only local machines? (ideally I should be able to provision any machine from any other machine)
hyper_ch2 has quit [Remote host closed the connection]
Makaveli7 has quit [Quit: WeeChat 2.6]
nDuff has quit [Quit: zzz]
<wedens>
makefu: thanks. I'll have a look
dm9 has quit [Quit: WeeChat 2.6]
bahamas has quit [Ping timeout: 264 seconds]
<Philonous_>
I have a zip file at an url that I'd like to fetch and extract with fetchzip. However, I am only interested in subdirectory inside the zipfile. How can I tell fetchzip to discard everything else?
<Philonous_>
The manual doesn't seem to give a list of parameters that fetchzip accepts
gentauro has quit [Read error: Connection reset by peer]
gentauro has joined #nixos
<hpfr[m]>
tilpner: how do i check the logs of display-manager? Systemctl
<infinisil>
Philonous_: Or you could just use "${fetchzip { ... }}/the/subdirectory" to refer to it
<arianvp>
wedens: my own take on this stuff: github.com/arianvp/nixos-stuff
idf has quit [Quit: WeeChat 2.6]
<Philonous_>
Oh, I guess that's easier. But it's good to have a way to prune the result anyway to save space (even though it's not necessary in this case)
<clever>
hpfr[m]: journalctl -t xsession, i think
idf has joined #nixos
<hpfr[m]>
Thanks!
<wedens>
arianvp: nice. looks very similar to what I'd like to implement
<feffe>
Anyone know how to force AptX connections? It says I've got an aptx point registered if i run `systemctl status bluetooth` but any time I connect it just uses SBC. Otherwise I haven't done anything to tell it to prefer aptx if available but I haven't been able to find anything that specifies how to do that either. Any ideas?
<wedens>
arianvp: can you explain a bit how your install.sh works, please?
<__monty__>
clever: Do you remember my problem from yesterday? http://ix.io/1Y75 After updating the remote nix I got a different error: http://ix.io/1Ya1 But now I'm back to the former o.O
<clever>
__monty__: --option sandbox false
kshi has joined #nixos
<Kim>
So I'm poking around at getting flickerfree startup working, but I'm running into some difficulties determining how units are prioritized in order to replace `systemd-ask-password-console` with `systemd-ask-password-plymouth`
<__monty__>
clever: : / More errors related to /build, does this mean I have to disable sandboxing for all my home-manager switches from now on?
<clever>
__monty__: you can also add sandbox = false to /etc/nix/nix.conf
<__monty__>
Sounds related to the lack of permissions to create /build from yesterday.
<clever>
__monty__: its a bug in nix, that seems to only occur on non-nixos machines
wfranzini has quit [Remote host closed the connection]
<clever>
__monty__: the problem is that an env var is incorrectly set to /build, when ti shouldnt be
<__monty__>
Oh, ok. So disabling sandboxing on the remote, right?
<clever>
yeah
Ariakenom has quit [Remote host closed the connection]
<red[m]>
maybe I need to include it in systemPackages as well (and not just rely on the service?)
<cransom>
it's already in systemPackages
<cransom>
if you enable the service, that is.
<gchristensen>
red[m]: you need to access the package on your own, it isn't magical string replacement -- it is just like a variable (which needs to be in scope) in any other programming language
wildtrees has quit [Quit: Leaving]
<red[m]>
which I guess is why I thought that nixpkgs.bacula would be correct
<{^_^}>
[nixpkgs] @flokli merged pull request #70861 → nixos/gitlab: Don't print sensitive data to log on startup → https://git.io/JeW1O
<red[m]>
but it doesn't know where to find nixpkgs either
hmpffff has quit [Quit: nchrrrr…]
<{^_^}>
[nixpkgs] @flokli pushed commit from @talyz to release-19.09 « nixos/gitlab: Don't print sensitive data to log on startup »: https://git.io/Jelzp
<red[m]>
(this is in my configuration.nix to be clear - I'm not editing the package itself)
<gchristensen>
so if you're in a NixOS configuration, make sure it starts with { pkgs, .... } (the "...." is probabl ystuff like 'config, lib' or something) and then do ${pkgs.whatever}... in your string blog
<gchristensen>
block(
<gchristensen>
gosh I can't type at all today!
<red[m]>
ah - pkgs, not nixpkgs ;-)
<red[m]>
now it compiles :-D
<gchristensen>
woot!
<__monty__>
Hmm, is that ssl error local or something up with the cache? Updating channels and rebuilding didn't help : /
<Miyu-saki>
This causes a regression of `nix-shell --run` not being pipeable.
idf has joined #nixos
<red[m]>
Oct 11 12:13:13 apophenia r8pbkwigfi84jfqkk98fn77bj8fqw1rq-unit-script-bacula-dir-pre-start[14556]: /nix/store/ylc0giqfzvcwkriz090cvw9nfhdzha25-bacula-9.4.4/etc/update_bacula_tables: line 50: su: command not found
<red[m]>
^^^ this is fine
* red[m]
sighs
* http
looks around
gxt has quit [Ping timeout: 260 seconds]
Ariakenom has joined #nixos
gxt has joined #nixos
<teto>
yep the new python infra is a bit verbose, we should have some shell logging functions: nix_debug/nix_info etc
<o1lo01ol1o>
Using `haskell.nix`, does anyone know the recommended way to make `pkgs.llvm` available to a a `stack-to-nix` project so that all dependencies with -fllvm flags can find llvm when they need it?
dm9 has quit [Quit: WeeChat 2.6]
<wedens>
should I use something like `linuxPackages_latest = super.linuxPackages_latest.extend(sf: sp: { mymodule = ... });` in overlay to add the module?
drakonis has joined #nixos
<{^_^}>
[nixpkgs] @Pamplemousse opened pull request #70982 → python3Packages.pwntools: init at dev3 → https://git.io/Jelg1
kleisli has quit [Ping timeout: 240 seconds]
chloekek has joined #nixos
dbmikus has joined #nixos
nicks0n has quit []
drakonis1 has quit [Ping timeout: 240 seconds]
gila has quit [Quit: My Mac Pro has gone to sleep. ZZZzzz…]
<wrl>
hey what's a quick command to display all of the files contained in a package?
<{^_^}>
[nix] @edolstra pushed to master « ssh-ng: Don't set CPU affinity on the remote »: https://git.io/JelgM
<cjpbirkbeck>
hello, do think 250 GB is good enough to store everything with nixos?
<multun>
sure
<multun>
you can even store multiple times eveything you need
<simpson>
Or it's not even close to enough. Depends on what "everything" is.
<multun>
If everything is your 2Tb of movies, it won't fit
<multun>
but if it's just your typical linux desktop install it will
<cjpbirkbeck>
true, i wanted get a 500gb ssd, but i couldn't pass up the deal on the 250 gb
<teto>
250 can be tight in my experience. If a linux kernel compilation fails and you don't clean /tmp you can laready have ~10 GB taken. Also depends if you stay on stable or not (more churn if you follow unstable etc).
<simpson>
Nix uses however much extra disk you have. 100 GB is feasible. 50 GB is feasible. 10 GB is feasible.
<{^_^}>
[nixpkgs] @lsix opened pull request #70984 → postgresql_12: init at 12.0 → https://git.io/Jel2D
<cjpbirkbeck>
right now i've used a desktop with / on a ~120 gb ssd parition and a /home parition on a ~1.5 tb hdd parition, and it's worked okay thus far
<cjpbirkbeck>
but now i've bought a used laptop, i've got a 250 gb ssd in it.
<cjpbirkbeck>
teto: i don't think i've ever had nix compile a kernel on me (it has compiled virtualbox, which took a bit of time on my aged hardware)
<multun>
teto: you can run the garbage collector from time to time
<cjpbirkbeck>
also, what are your recomendations on disk encryption?
<teto>
multun: but then you lose one of the benefits of nixos, falling back to previous generations
<ottidmes>
cjpbirkbeck: depends on your treat model, if you don't want you data easily accessible in case of theft for example, hardware encryption can be good enough, otherwise I tend to use LUKS, but you might look at ZFS encryption too
smatting has quit [Ping timeout: 245 seconds]
<teto>
if you track unstable do some haskell and kernel developments (with a few VMs), it may lead you to be thoughful of memory. In a way similar to any distro, just with nix you hope to keep a few generations, else it's frustrating :)
smatting has joined #nixos
<cjpbirkbeck>
i don't want to mess around with zfs, at least at this time. i was thinking of having full disk encycption with lucks (used it in virutalbox) along some kind of software encryption for protecting the system while booted.
doyougnu has joined #nixos
<simpson>
Like, in-memory encryption? Are you worried about Evil Maids?
avn has quit [Ping timeout: 268 seconds]
<infinisil>
ottidmes: hardware encryption? I've never seen that
ariutta has joined #nixos
greizgh has joined #nixos
<simpson>
infinisil: It's hard to remember the times before software block management, but there was a time when disk encryption, striping, RAID, etc. were managed by specialized hard-drive management controllers. You may hear graybeards discuss "RAID cards".
matthuszagh has joined #nixos
<infinisil>
I have heard about RAID controllers, never about hardware encryption though
<simpson>
Linux's md ended up being competitive with those hardware components, and they've slowly phased out of the market.
<ottidmes>
infinisil: meant Full Disk Encryption (FDE), which is often just called hardware encryption
sigmundv has quit [Ping timeout: 265 seconds]
Havvy has quit [Ping timeout: 240 seconds]
<simpson>
ottidmes: Right, because if you want to encrypt the *full* full disk, then you need a piece of HW that can decrypt the bootloader, or the bootloader with the key needs to be not-on-disk.
<aanderse>
anyone in here ever get pam to authenticate against ldap on nixos?
Havvy has joined #nixos
<aanderse>
a few different errors like "In order to perform this operation a successful bind must be completed on the connection."
<aanderse>
"cannot initgroups("nslcd",58) (ignored): Operation not permitted"
<aanderse>
etc...
<aanderse>
if anyone has a working config i would appreciate sharing, or any help
<aanderse>
for reference, trying to adapt configuration which is working on an ubuntu box
<infinisil>
ottidmes: Hm not entirely sure now
ris has joined #nixos
<infinisil>
ottidmes: I think nix-prefetch-git and fetchgit need to know the name of the derivation to work, and those should stay in sync, or so
<infinisil>
And it's "source" or something by default?
<infinisil>
Or is this only with fetchFromGitHub
smatting has quit [Ping timeout: 245 seconds]
smatting has joined #nixos
Jackneill has quit [Remote host closed the connection]
<ottidmes>
infinisil: no clue myself, just found that conversation after checking any references to nix-prefetch in the logs. Wish there was a good example, because from what I have read so far, seems already doable with an overrideDerivation
<infinisil>
What is doable specifically?
<ottidmes>
"nix-prefetch would have to be modified, to use the same name", and shouldn't overrideDerivation allow you to override the name to make it the same?
doyougnu has quit [Ping timeout: 245 seconds]
cinimod has joined #nixos
<infinisil>
Hm yeah
doublex has quit [Ping timeout: 240 seconds]
cjpbirkbeck has quit [Quit: Quitting now.]
<matthuszagh>
hi. i'm using symlinkJoin to wrap emacs and a collection of other packages I only intend to use with emacs. this works, but anytime a package changes it triggers a rebuild of emacs as well. is there a way to do this that doesn't require a rebuild of emacs each time? happy to post the config if you need more info
<wrl>
why is nano everywhere
<wrl>
why is nano *everywhere*
<wrl>
it's the default in `nixos-rebuild edit` if there's no EDITOR defined
<simpson>
wrl: What editor would you rather have as the system fallback?
fresheyeball has joined #nixos
<ottidmes>
wrl: what is the alternative, wouldn't people claim exactly the same for any other editor, to each their own, just set EDITOR if you prefer one
<infinisil>
matthuszagh: It shouldn't do that, nix files would be helpful to see
<wrl>
ottidmes: i did, and now i'm trying to figure out where it actually gets set and why `sudo` isn't picking it up
<wrl>
simpson: i just really don't like nano and i dislike the fact that i can't uninstall it
<fresheyeball>
what is this thing doing NIX_GHC_LIBDIR?
<infinisil>
matthuszagh: Are you perhaps including the current directory in the emacs build, which gets changed every time because the ./result link changes when you nix-build? That's a common cause of unintentional rebuilds
<wrl>
simpson: it propagates *some*, but i feel like propagating EDITOR is a hacky way about this
<simpson>
wrl: It's the standard way. What did you want to do instead?
<matthuszagh>
let me know if you'd like emacs overlay as well
cinimod has quit [Ping timeout: 246 seconds]
<infinisil>
matthuszagh: Can't see anything suspicious in that. An emacs overlay that changes the emacs build? That could be reltaed
<wrl>
simpson: there's stuff that ends up in `sudo env` and i'm not sure how it gets there. stuff like GTK_PATH, and i don't know where that's coming from
<wrl>
so, why does *that* end up in my env, but EDITOR (which i'm defining in configuration.nix) doesn't?
<infinisil>
matthuszagh: Ah, the src line looks very suspicious
<infinisil>
This means every time that directory changes, in any way, it will cause a rebuild
stampe has joined #nixos
stampe has quit [Client Quit]
NoctisLabs has joined #nixos
rsoeldner has quit [Ping timeout: 240 seconds]
<matthuszagh>
infinisil: that was intentional since it allows me to use the git master and incorporate any personal changes etc
<matthuszagh>
and it seems to rebuild even when I don't make a change to that directory
<infinisil>
Even .gitignored things are a change
<simpson>
wrl: No idea, sorry. I'm pretty skeptical of trying to set environment variables globally like that, since there isn't exactly a portable way to reliably establish such a setting.
<infinisil>
matthuszagh: Double/triple check that this isn't the cause, because that's almost surely the reason
doyougnu has joined #nixos
<infinisil>
matthuszagh: You could check the hash of a directory like nix computes it with `nix-hash /path/to/directory` to debug
<matthuszagh>
infinisil: that's a good idea, i'll try that
greizgh has quit [Ping timeout: 245 seconds]
greizgh has joined #nixos
domogled has quit [Ping timeout: 240 seconds]
<matthuszagh>
infinisil: it doesn't appear to be a change in the directory. the hash stays the same
<infinisil>
matthuszagh: Hm okay, then I have two things you could try
<infinisil>
The first is to use `nix-store -q --tree /nix/store/...-emacs.drv` to see the build time dependencies of the emacs derivation
<infinisil>
And compare this between two drvs that are apparently different
<infinisil>
The second is to just remove the `src` line to see if this is causing it
drakonis_ has joined #nixos
kshi has joined #nixos
<__monty__>
Any reason nix-shell'd complain about SSL certs but not my browser? http://ix.io/1YkV
<__monty__>
Debian's solution to the editor thing is to patch `sensible-editor` in everywhere. Then users have a way to change even hardcoded default editors.
MichaelRaskin has joined #nixos
drakonis_ has quit [Ping timeout: 250 seconds]
freepotion has quit [Remote host closed the connection]
<matthuszagh>
infinisil: thanks for the help! i think i found the issue. looks like I was mistaken about it rebuilding emacs itself. but i was using emacs with home-manager and every time the wrapped version changed all of the elpa/melpa packages were getting rebuilt
selfsymmetric-pa has joined #nixos
codygman[m] has joined #nixos
smatting has quit [Ping timeout: 265 seconds]
doyougnu has quit [Quit: WeeChat 2.6]
<wrl>
what's the preferred way to install home-manager these days? globally in configuration.nix (since there's a nixpkg for it now), or?
domogled has joined #nixos
<clever>
infinisil: nix-diff will diff drvs
<fresheyeball>
infinisil: I got hie to find packages
<__monty__>
wrl: rycee recommends just following the instructions in the manual. The nixpkg lags a bit. But you might not care.
<fresheyeball>
I got it!
<fresheyeball>
but fuck!
<fresheyeball>
2 things
<fresheyeball>
1. ghc must be provided by the nix-shell
<wrl>
__monty__: aha, got it now
<fresheyeball>
2. ghc must be configured with ghcWithPackages, and then ghcmod will find the packages
<wrl>
__monty__: is there a way to add channels inside of configuration.nix? i'm trying to keep the number of imperative commands i have to run low
knupfer has joined #nixos
<__monty__>
I'd just import a tarball or use fetchFromGithub in that case.
<{^_^}>
[nixpkgs] @flokli pushed 3 commits to release-19.03: https://git.io/JelVg
<clever>
fresheyeball: the .env attr will run ghcWithPackages for you
<fresheyeball>
does this mean that to access ghcmod tooling I need to maintain the same package list again in ghcWithPackages?
<slyfox>
clever: do i need to patch .nix file or there is a way to pass it via --arg?
<clever>
fresheyeball: just open a shell on the .env attr, and it will do the ghcWithPackages for you
<clever>
slyfox: simplest to just add doCheck = true; to the derivation in the nix file
<clever>
slyfox: --arg cant override that param
<fresheyeball>
clever: that is what I was doing
<fresheyeball>
and it did not have the effect
<slyfox>
aha, and if the tests require more depends is it customary to extend a derivation to include those conditionally? pointing me to a .nix example would be perfect
<fresheyeball>
clever: does it do this by shellHook?
<fresheyeball>
I am overriding
<clever>
fresheyeball: what are you overriding?
<fresheyeball>
the shellHook of my drv.env
<clever>
fresheyeball: its not using shellHook
yosemitesam has joined #nixos
<clever>
fresheyeball: what does `which --all ghc` return?
<catern>
nix run doesn't have the same kind of development environment stuff as nix-shell, right?
<clever>
correct
Squarism has joined #nixos
<Squarism>
Hey
<Squarism>
I need to troubleshoot this:
<Squarism>
error: anonymous function at /nix/store/m9hg0rx6hn47wv4lqghss3g77kiipai5-cabal2nix-LambdaForms/default.nix:1:1 called without required argument 'servant-client-ghcjs', at /nix/store/9r1vhvri4ba913g4vyg5jqrhg6pb9amn-source/pkgs/development/haskell-modules/make-package-set.nix:87:27
<clever>
Squarism: servant-client-ghcjs doesnt exist in haskellPackages
<Squarism>
oh ok
<clever>
Squarism: you need to add it with cabal2nix and an override
<Squarism>
hmm.. im super green on nix. But that would be in my project default.nix file?
<{^_^}>
cleverca22/nixos-configs#6 (by AmandaCameron, 21 hours ago, open): qemu module is broken on nixpkgs-unstable/nixos-unstable
<clever>
Squarism: can you pastebin it?
<Squarism>
sure
ravloony has quit [Quit: Leaving]
<clever>
AmandaC: havent gotten that one solved yet, but i have seen signs that something related may already be in nixpkgs, need to look into that when i get a chance
<fresheyeball>
domenkozar[m]: so now why ghcide instead of hie?
<fresheyeball>
both work on my end now
<wucke13_>
Whoever uses whatever tool to distribute `broken = true;` to broken packages: It would be very nice to check wether there is a PR open to fix that! I have a PR for a broken package open since a week or so, and had to rebase and resolve a merge conflict caused by some bot inserting `broken = true;`. Like instead of merging the broken marker, one could have simply merged the fix...
<ivan>
maybe broken = true; is a bad idea given how things break and unbreak
<gchristensen>
I don't think it is a bot that marks them broken. I'm sorry we missed your PR, though, wucke13_
<gchristensen>
marking packages as broken, when they're known to be broken, is good for users as it prevents them from wasting time trying to build it. also, it is good for hydra and build success metrics -- # of expected broken vs. actual broken
<ivan>
ah
<ivan>
cached failures could solve the wasted time
<ivan>
I guess the hard part is deciding whether it failed spuriously or not
<wucke13_>
No offense, but from my personal view that is somewhat pervert. So maybe have a look #70160.
<wucke13_>
(Even if some user reports it doesn't work for his device, it works for mine and having a package which at least compiles fine is better than havin a broken one)
selfsymmetric-mu has quit [Remote host closed the connection]
<infinisil>
fresheyeball: Personally I'll stick with haskell-ide-engine for a couple reasons, one of which is that there are a couple people actively developing it (again)
wedens has quit [Quit: Connection closed for inactivity]
toppler has quit [Read error: Connection reset by peer]
Jonathan45 has quit [Remote host closed the connection]
smatting has quit [Ping timeout: 276 seconds]
toppler has joined #nixos
smatting has joined #nixos
kshi has quit [Ping timeout: 250 seconds]
<AmandaC>
Any idea what's up with this? `/nix/store/l6h4ya0wzb4b8mr0y58k2gh2nhfql4sn-bash-4.4-p23/bin/bash: /nix/store/l6h4ya0wzb4b8mr0y58k2gh2nhfql4sn-bash-4.4-p23/bin/bash: cannot execute binary file`
<AmandaC>
I'm getting it when I try and offload some programs to my linux vm for distributed building from my mac.
<AmandaC>
seems to happen from random derevations
<wucke13_>
AmandaC: any way you can find out which binaries cause this?
<ivan>
run `file` on said binary and see if it can even run on your platform
<c00w>
AmandaC, I'd try sshing into the linux vm and running the binary. readelf may also tell you more information.
<wucke13_>
If you know them, check wether all of theire shared lirbaries are found (using ldd) and `file` to get more information about the interpreter
<clever>
AmandaC: it sounds like your mixing up linux and darwin binaries
<AmandaC>
clever: not sure how? on both machines it reads as an ELF 64-bit binary: `/nix/store/l6h4ya0wzb4b8mr0y58k2gh2nhfql4sn-bash-4.4-p23/bin/bash: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), dynamically linked, interpreter /nix/store/6yaj6n8l925xxfbcd65gzqx3dz7idrnn-glibc-2.27/lib/ld-linux-x86-64.so.2, for GNU/Linux 2.6.32, not stripped`
<AmandaC>
and when I pass -j0 it's not supposed to try and build anything locally, right?
<AmandaC>
I'm running `nix-build ./release.nix -A lilac --option system x86_64-linux -j0` on my mac. The same command minus --option system and -j0 runs fine on the remote machine
kvda has joined #nixos
<clever>
AmandaC: your trying to run a linux binary on darwin, i think
<{^_^}>
[nixpkgs] @scalavision opened pull request #70995 → deeptools: init at 3.3.1 → https://git.io/JelrX
<AmandaC>
thanks, clever, that actually gives me the snippet of knowledge I'll need for when the qemu-user stuff works for my raspberry pi experiments. :3
kleisli has quit [Ping timeout: 240 seconds]
oscarvarto has quit [Quit: Konversation terminated!]
wfranzini has quit [Remote host closed the connection]
<spiderbit>
hi, I just upgraded my nixos (19.03 -> 19.09) with a declerative container that uses openvpn. Which stopped work anymore:
is_null has quit [Ping timeout: 268 seconds]
<spiderbit>
system-getty.slice: unit configures an IP firewall, but the local system does not support BPF/cgroup firewalling.
doublex has quit [Ping timeout: 250 seconds]
<Yaniel>
,locate libnvidia-glvkspirv.so
<Yaniel>
,where libnvidia-glvkspirv.so
<{^_^}>
Couldn't find in any packages
<spiderbit>
my first instinct was to try to pin that vm back to 19.03 which would be possible over the "path" option but didn't understand how to make that work then I tried to create a imperative container but there I think I have no tun based network option and also not sure how to bind some paths inside that container... but is there a easy way to fix that without going back to 19.03 inside the container?
gnidorah has quit [Quit: Connection closed for inactivity]
doublex has joined #nixos
<spiderbit>
I think some sort of pinning would be useful so I can decide when I upgrade the container but if there is a easy solution to fix that network upgrade problem that would be nice too.
is_null has joined #nixos
gxt has quit [Remote host closed the connection]
rsoeldner has quit [Remote host closed the connection]
rsoeldner has joined #nixos
spiderbit has quit [Remote host closed the connection]
spiderbit has joined #nixos
justanotheruser has quit [Ping timeout: 240 seconds]
Synthetica has quit [Quit: Connection closed for inactivity]
m15k has quit [Ping timeout: 260 seconds]
rsoeldner has quit [Ping timeout: 246 seconds]
orbekk has quit [Quit: WeeChat 2.4]
<worldofpeace>
spiderbit: I think the release notes maybe have something about this
<clever>
,profile
<{^_^}>
clever: Did you mean profiling?
<{^_^}>
Use NIX_COUNT_CALLS=1 and/or NIX_SHOW_STATS=1 to profile Nix evaluation
<worldofpeace>
only because reading "unit configures an IP firewall, but the local system does not support BPF/cgroup firewalling." reminds me we use upstream sysctl snippets from systemd. but it may be unrelated
<{^_^}>
[nixpkgs] @LouisDK1 closed pull request #67128 → All available translations added for libreoffice-fresh and libreoffice-still → https://git.io/fjbRa
rsoeldner has joined #nixos
toppler has quit [Read error: Connection reset by peer]
toppler has joined #nixos
Church- is now known as in_hiding
fendor has quit [Quit: Leaving]
in_hiding is now known as Church-
gxt has joined #nixos
rsoeldner has quit [Remote host closed the connection]
rsoeldner has joined #nixos
asymptotically has quit [Quit: Leaving]
philr has joined #nixos
<danderson>
Hi! I'm getting my server onto 19.09, and I'm confused: there doesn't seem to be any linux 5.* kernel packages in 19.09. Am I reading nixpkgs wrong?
orbekk has quit [Quit: WeeChat 2.4]
<samueldr>
a change in the way versioned kernel attribute names are handled has been made starting with 19.09
<samueldr>
_latest still points to the latest kernel, currently 5.4 IIRC; but versioned names for non-LTS kernels have been dropped
<samueldr>
when the next LTS is announced, I figure it will be added into the list of attrnames available
<samueldr>
the principle behind this is to not ship with, let's say, the attrname linuxPackages_5_3, and then it gets removed, thus not having a stable "API" for the packages list
o1lo01ol1o has joined #nixos
<danderson>
Hmm, I see. I must have missed that in the release notes
<danderson>
I'll have to check if the ZFS performance patches got merged back into LTS, or if I need to run latest. Thanks!
orbekk has joined #nixos
rsoeldner has quit [Ping timeout: 276 seconds]
<infinisil>
samueldr: I wish we'd get into the habit of using `throw` for dropped attributes :/
<infinisil>
This isn't the first time I've seen people ask why attributes aren't there anymore
<infinisil>
With a throw the reason along with a workaround can be given
<hpfr[m]>
Now when I run nixos-rebuild switch, it hangs during "building the system configuration"
rsoeldner has quit [Remote host closed the connection]
rsoeldner_ has joined #nixos
is_null_ has joined #nixos
<danderson>
for the record, if anyone goes digging for this information: the nixos LTS kernel versions have the FPU function export patch applied, so SIMD checksumming/encryption for ZFS works on those kernels.
is_null has quit [Ping timeout: 240 seconds]
<danderson>
so I can downgrade from 5.2 to some LTS version, yay!
glenndavy has joined #nixos
<hpfr[m]>
What am I doing wrong? "sudo ssh builder" works with no interaction, and I think that's what I need since I'm on NixOS on the laptop, so root does builds, right? I ssh into my user account on builder, which is all that's necessary right? Don't need to ssh into the root account on the builder
is_null_ has quit [Ping timeout: 240 seconds]
is_null has joined #nixos
<hpfr[m]>
"sudo nix ping-store --store ssh://builder" returns nothing, which I believe is a success as well
kleisli has quit [Remote host closed the connection]
kleisli has joined #nixos
<kalbasit>
c00w: yes, I'm giving it one more ofborg then merge
<danderson>
so, when do I change system.stateVersion in my nixos configs? The comment in the installer-made file says to only do so after the release notes tell me to, but the release notes don't.
noudle has quit []
<danderson>
Parts of the release note mention having to stay at 19.03 for the first activation, because some services need porting of their state. Once that's done, is it safe to set the stateVersion to 19.09?
<danderson>
well, changing the state version seems to be a no-op from dry-build and dry-activate's POV, so... Yay? :)
<qyliss>
,stateVersion
<{^_^}>
Setting stateVersion to the latest release doesn't upgrade anything and can only break your setup at best. To actually upgrade NixOS see https://nixos.org/nixos/manual/#sec-upgrading. If you want to update the stateVersion option regardless, Ctrl-F for "stateVersion" in https://nixos.org/nixos/manual/release-notes.html to see things that need to be manually migrated with the new value.
kleisli has quit [Ping timeout: 268 seconds]
<fuzen>
clever: that looks good not sure how to use it. Is this part of nix-ops?
<clever>
fuzen: just run `nix-build -A go` and then `./result` to boot it
<clever>
fuzen: note, that you need a linux build slave to build the linux build slave...
<{^_^}>
[nixpkgs] @mweinelt opened pull request #71006 → babeld: add test → https://git.io/Jel6W
<AmandaC>
clever: ... I may have stumbled upon how to make your qemu stuff work again. I did the wrong flag when I did the `-D...` change -- it wants `-Ddefault_library` instead of `-Ddefault-library` -- it seems to have gotten much farther than before, at least.
<danderson>
qyliss: that sounds like stateVersion is never meant to be updated, even though the config comment implies it should be.
<AmandaC>
Yup, it's building qemu now
<qyliss>
danderson: it's only meant to be updated if you're told to update it by release notes and stuff
<qyliss>
which is, in practice, never
<clever>
AmandaC: oh, nice, thats much simpler then i thought
<AmandaC>
clever: It remains to be seen if the changes will be executable, but I'll test that next
<clever>
danderson: in general, you need to figure out what services you use, that care about stateVersion, and then investigate what changes it will cause
<clever>
danderson: sshd for example, at one point, it changed the type for ssh hostkeys, the upgrade is seamless, but it creates new hostkeys, causing mitm alerts for every single client, and nixos prefers to avoid causing such errors
<clever>
danderson: postgresql uses stateVersion to control the version of psql, so you can access your db, and you must manually export it to .sql, upgrade, then re-import
<clever>
danderson: the release notes should have details on that kind of thing
matira has joined #nixos
rsoeldner__ has quit [Ping timeout: 246 seconds]
<AmandaC>
clever: the "correct" option for future-proofing would probably be to filter out the one that gets added by makeStaticLibrary as well, but this'll do for now, assuming it works.
<clever>
danderson: if you remove or change the stateVersion without reading them, then you may break the things its meant to not break, but you could just go ahead, and then fix whatever is broken
<danderson>
I see. That said, all those things should show up in dry-build and dry-activate as changes. There was one change in 19.09 that cared about stateVersion, and it was a one-time thing
<AmandaC>
( And I'm not skilled enough with nix to do that filtering )
<clever>
danderson: ah yeah, thats a good catch
<danderson>
so, based on my reading of the release notes, and the lack of diff from build/activate... seems fine?
<red[m]>
So - bacula has issues.
<red[m]>
things like this: /nix/store/ylc0giqfzvcwkriz090cvw9nfhdzha25-bacula-9.4.4/etc/mtx-changer: line 336: awk: command not found
<clever>
danderson: what was the one thing that changed?
<danderson>
I guess I'll find out if all my data gets eaten when I'm not looking :P
Rusty1 has joined #nixos
<red[m]>
any idea why their shell scripts aren't able to find basic things like awk, su et al?
<danderson>
clever: systemd-timesyncd's state directory moved
<fuzen>
clever: Oh, ill keep looking into making a docker container to build it, im just trying to get my default.nix go-lang package to target linux-x86_64 only in docker.nix
<clever>
AmandaC: i'll take another look at my code and see about fixign that
<danderson>
but it has an activation shim to move the state around if stateVersion <19.09
<clever>
danderson: ahh, it doesnt really keep much state either, so not much lost
<red[m]>
This is from the systemd startup - so I'm sure that's the root of the problem ;-)
* red[m]
ducks
<danderson>
there were other changes in relnotes that depend on stateVersion, but not for software I'm running
<danderson>
so that's all fine.
<red[m]>
(when I run it from the command line it works fine... it's just when it's running as a daemon
<red[m]>
)
<spiderbit>
So I asked that before but I guess not much people were there before... I updated my server to 19.09 and I have a declaritive nixos container with openvpn setup... that did not survive the update
<clever>
red[m]: systemd has a different default for PATH
<clever>
red[m]: you must set systemd.services.foo.path = [ pkgs.gawk ];
<spiderbit>
system-getty.slice: unit configures an IP firewall, but the local system does not support BPF/cgroup firewalling.
<spiderbit>
cgroup compatibility translation between legacy and unified hierarchy settings activated. See cgroup-compat debug messages for details.
<spiderbit>
I would be fine to pin that back to 19.03 if somebody can tell me how to do that
<red[m]>
ah ha!
<red[m]>
is there a default or something for it?
<spiderbit>
I saw the "path" variable but that for some reason replaces the config variable...
<red[m]>
(ie - if i define it am I risking it unsetting some other paths that are needed?)
<clever>
red[m]: systemd will merge it with the other values
<red[m]>
clever: do you know what package su would be in?
<clever>
red[m]: a: su, b: why does the service need su?
<qyliss>
spiderbit: just wanted to say that i'm sorry you're not getting an answer, and that's it's likely because people here right now don't know, rather than you being ignored :)
drakonis has quit [Ping timeout: 246 seconds]
mbrgm_ has joined #nixos
<red[m]>
clever: I can show you - one moment
waleee-cl has quit [Quit: Connection closed for inactivity]
mbrgm has quit [Ping timeout: 240 seconds]
mbrgm_ is now known as mbrgm
rsoeldner__ has joined #nixos
<red[m]>
clever: having problems finding it in the sources - but this is the file that gets installed:
<clever>
red[m]: nixos has options to do that for you
work_ has quit [Quit: Connection closed for inactivity]
<hpfr[m]>
clever: I'm the guy with the nixos-rebuild distributed build issue, ran `nix-build '<nixpkgs/nixos>' -A system -v` and it's been sitting there for at least five minutes after printing out a bunch of "evaluating file" messages. The last one was for "nixos/pkgs/tools/X11/xkbvalidate/default.nix". I don't know if that was the last file to evaluate and it's now hanging on whatever the next step in the build process is, or if it's
<hpfr[m]>
actually hanging on this file. I'm guessing the former, but I don't know where to go from here then