<ottidmes>
lol, guess I have to filter nix.buildMachines on host, cause now it is delegating the building... to itself, causing a lock
sleepingforest has quit [Client Quit]
<{^_^}>
[nixpkgs] @c00w opened pull request #83253 → treewide Drop unneeded go 1.12 overrides → https://git.io/Jvyuk
sleepingforest has joined #nixos
cr4y1 has quit [Ping timeout: 258 seconds]
mbrgm has quit [Ping timeout: 272 seconds]
mbrgm has joined #nixos
maddo has quit [Quit: See ya]
sleepingforest has quit [Client Quit]
cosimone has joined #nixos
sleepingforest has joined #nixos
organixpear has joined #nixos
ixxie has quit [Ping timeout: 264 seconds]
sleepingforest has quit [Client Quit]
sleepingforest has joined #nixos
zeta_0 has joined #nixos
hmpffff has quit [Quit: nchrrrr…]
<zeta_0>
does nix-shell inherit environment variables automatically from home.nix, or will i need set them manually in default.nix?
<clever>
zeta_0: nix-shell will inherit vars from whatever shell its ran in, plus the shell.nix/default.nix file
<clever>
zeta_0: --pure will stop that inheriting
sleepingforest has quit [Quit: bye!]
<{^_^}>
[nixpkgs] @aanderse opened pull request #83254 → nixos/mysql: fix service so it works with mysql80 package [20.03 backport] → https://git.io/Jvyuw
sleepingforest has joined #nixos
<zeta_0>
clever: oh, i've always wondered what --pure meant
<energizer>
when they say nix is pure, which parts are they talking about?
<zeta_0>
clever: i just start contributing to the darcs open source project, but i am having problems configuring darcs in a default.nix file to run in a nix-shell, here's what my default.nix file looks like so far: https://hastebin.com/uturerewam.bash
sleepingforest has quit [Quit: bye!]
<energizer>
cuz it seems like i can import arbitrary paths, make network requests, and write to places on disk
<ottidmes>
there is still no setting to say, do not error if substitutor is missing, right?
<zeta_0>
clever: the people in #darcs don't understand nix to well, that's why i'm asking here in #nixos
<Ilya_G>
I am trying to run multiple instances of a service in systemd essentially from template. In my configuration.nix I have `systemd.services.my_service_name`
<clever>
zeta_0: what are you trying to do?
<Ilya_G>
I found recommendation in regular linux to create a `my_service@.service` file as configuration
<clever>
Ilya_G: basically, systemd.services."foo@" = ...; to define the template, and systemd.services."foo@bar".wantedBy to spawn an instance, i think
<zeta_0>
clever: i want to build darcs with cabal in a nix-shell but, in order to contribute to the darcs open source project, to build darcs from scratch with cabal, it requires some dependencies(haskell and non-haskell dependencies) and an environment variable setup for it to work: http://darcs.net/Development/GettingStarted
rardiol has quit [Ping timeout: 256 seconds]
<zeta_0>
clever: i think i just need to finish setting up the environment variable, then i should be good to go
<clever>
zeta_0: what happens if you simply `nix-shell '<nixpkgs>' -A haskellPackages.darcs.env` and try to `cabal configure` ?
sleepingforest has quit [Quit: bye!]
captjakk has quit [Remote host closed the connection]
sleepingforest has joined #nixos
<ottidmes>
clever: I was thinking just now, the one thing I hate about substitutors is that they do not allow any failure. In the past I patched most commands to dynamically set them, which works, but that's less than ideal. How difficult would it be to make a custom cache on my router that then redirects to substitutors that are live?
<zeta_0>
clever: let me know if i'm missing anything or doing something incorrect in my default.nix file, your the nix expert, i'm the layman
<zeta_0>
clever: hold on, give me a second to run that command
<Ilya_G>
clever: would I be able to read bar as a parameter inside the configuration?
<clever>
ottidmes: cachecache could be modified to dynamically change the list of upstream caches
<Ilya_G>
I want to be able to send port address which this service instance is going to be listening to.
ddellacosta has joined #nixos
<clever>
Ilya_G: systemd will replace %i with bar within the .service files
<clever>
Ilya_G: you can see containers.nix using %i everywhere
captjakk has joined #nixos
<Ilya_G>
Very cool, so it maps from regular Linux systemd. Ok I'll give it a whirl.
<ottidmes>
clever: yeah, I was thinking that, just wondering how hard that would be, or is any machine running nix-daemon already a cache if you want it to be? (as in, how difficult would it be to upstream to say my desktop)
<energizer>
has anybody compared the collection of packages in nixpkgs to those in other package repos to see how many things debian/arch/conda/... have that nixpkgs doesn't?
<clever>
ottidmes: you can add ssh://user@host as a substitutors
<ldlework>
energizer: it's not the size of the package repository, it's how you use it
<ottidmes>
clever: I am using ng-ssh:// (from the wiki), or is that now outdated, but would using that just work with cachecache if I make it work with a dynamic list?
<clever>
ottidmes: currently, cachecache only supports the http protocol
sleepingforest has quit [Client Quit]
<energizer>
ldlework: sure, but i just used conda to get something and i wish i didn't have to
sleepingforest has joined #nixos
<ottidmes>
clever: guess I could use nix-serve for that though, right?
<clever>
ottidmes: nix-serve would convert a normal /nix/store into an http cache
<ottidmes>
clever: which then would make it easier to tweak cachecache to get the behavior I want, right? If so, I will give it a try, if not, getting other protocols to work seems a bit to daunting for a nice to have feature
KeiraT has quit [Ping timeout: 240 seconds]
<clever>
ottidmes: currently, cachecache is close to just a dumb http cache, with a little bit of inteligence around the http/narinfo protocol
<clever>
ottidmes: nar.xz files are cached to a dir, and it can query multiple upstream sites to find one
<clever>
ottidmes: narinfo files are cached to ram, for faster lookup
<ottidmes>
clever: right, fair enough, I rarely consider such edge cases
<ottidmes>
kalbasit: is now status.nixos.org
<kalbasit>
🤩
<kalbasit>
nice!
<ottidmes>
whoever maintains it should really make a redirect
<kalbasit>
yea
sleepingforest has quit [Quit: bye!]
sleepingforest has joined #nixos
<ottidmes>
clever: just for my understanding the rename trick makes it work because it will have a different inode, right?
<clever>
ottidmes: it will have a different inode and last-mod timestamp
<ottidmes>
thinking of atomic, I always think of atomic commits, but what you mean is more like immutability, where your data you received does not change due to something else
<clever>
ottidmes: i'm thinking more in terms of the change either didnt happen, or fully happened
<clever>
ottidmes: no middle ground where you see half a change
srid has joined #nixos
mbrgm_ has joined #nixos
mbrgm_ has joined #nixos
mbrgm_ has quit [Changing host]
mniip has quit [Ping timeout: 600 seconds]
<srid>
`callPackage foo { }` <- why does this not respect the default values for arguments in foo/default.nix?
<clever>
srid: the defaults are for when they cant be found under pkgs
<ottidmes>
srid: could you eleborate, what do you mean with does not respect the default values, if it refers to anything not in the package set, only then will defaults be used
<srid>
`callPackage neuron { }` doesn't seem to set `name` (to "neuron"), for example.
<zeta_0>
clever: ok it seems to work, does haskellPackages.darcs.env include everything that i need or will i still need a default.nix file for the nix-shell? https://hastebin.com/ixubokeliy.sql
<clever>
zeta_0: it probably includes everything you need
<clever>
zeta_0: thats what nixpkgs is using to compile darcs
<srid>
clever: not sure what the relevance of `pkgs` here is. The default.nix link I provided above has arbitrarily named arguments
<clever>
srid: when you use pkgs.callPackage, it will search for each of those args, in pkgs
<srid>
or maybe I should be using something *other than* `callPackage` here? After all, `nix-env -fi` works, and I want to install this thing using home-manager.
<clever>
srid: so root is getting replaced with this
<clever>
srid: and what does rib do when you give it the wrong root?
<srid>
Well, since there is no `pkgs.rib`, I suppose the default is being used here for `rib`?
<clever>
yeah
<colemickens>
how do I determine what kernel module to force include in the initrd to get networking in initrd on my rpi4? I havne't managed to figure it out yet AFAICT
<srid>
(typically I override rib using `nix-shell --arg rib ../rib`)
<srid>
clever++
<{^_^}>
clever's karma got increased to 363
<clever>
colemickens: do you see genet in lsmod?
<ottidmes>
On one of my Nix projects I have been told to improve it by making a Makefile, but if it is something like I have now: https://gist.github.com/msteen/5ebe0cdda8ac7520b0d4934b6b3ce277 how would you even go about making it into a Makefile, other than the mkdir/chmod, the rest is stdenv/makeWrapper specific
<colemickens>
clever: when its booted normally in nixos, with working ethernet, "lsmod | rg genet" returns nothing. I had tried this last night as well.
mniip has joined #nixos
<colemickens>
Which made me think maybe support is built in? But if that were the case, I should've seen it attempt to configure on boot, there is definitely an explicit "echo"
<clever>
colemickens: what about `modinfo genet` ?
<colemickens>
modinfo: ERROR: Module genet not found.
<colemickens>
[cole@raspberry:~]$ ls /sys/class/net/eth0/device/driver
<colemickens>
bind fd580000.genet uevent unbind
<colemickens>
?
<colemickens>
This is a bit deeper in Linux than I normally go. I'm a bit turned around.
<clever>
colemickens: and what about ls /sys/module/genet/ ?
<colemickens>
[cole@raspberry:~]$ ls /sys/module/genet/
<colemickens>
parameters uevent
<colemickens>
maybe this is an rpi_kernel oddity?
<clever>
colemickens: so the genet module is present, but if modinfo cant find it, it must be compiled into the kernel
<colemickens>
okay, I see, thanks for that detail
<`_>
I've identified a very peculiar bug.
<clever>
colemickens: no need to do anything special then, it should just work in the initrd
<clever>
colemickens: ive also looked into the hw of it some, and the genet doesnt have any firmware
<{^_^}>
[nixpkgs] @bhipple opened pull request #83255 → [20.03] pythonPackages.flake8-future-import: 0.4.5 -> 0.4.6 and fix build → https://git.io/Jvyz9
* colemickens
is increasingly concerned he's making a more basic mistake somewhere
<clever>
colemickens: try getting a shell in the initrd, and then do `ip link`
growpotk- has joined #nixos
sigmundv_ has quit [Ping timeout: 250 seconds]
<colemickens>
yeah, I wish the stage-1 trap function gave an option to drop to shell instead of just continue/reboot
<colemickens>
would that be a sensible thing to add while I'm at this?
<clever>
colemickens: it can already do that, but you have to authorize it with a kernel cmdline flag
<colemickens>
Got it. Okay.
<clever>
151 boot.shell_on_fail)
<clever>
colemickens: add boot.shell_on_fail to the cmdline, and then a shell will be added to the menu
<clever>
colemickens: boot.debug1 and other debug ones (read the script) will also force a false failure at various points, letting you get a shell early
<bqv[m]>
ottidmes: gh:bqv/nixflk
cosimone has quit [Quit: Quit.]
<`_>
Perhsp I haven't identified a peculiar bug
<`_>
But I've identified a peculiarity.
<`_>
Nope, I've identified that I'm doing something wrong
<bqv[m]>
Oh shit it worked, I now have my full system as a flake, with nothing broken
<ottidmes>
bqv[m]: I got sent that link in #nixos-chat, is indeed the best resource I have seen so far
<zeta_0>
clever: that's great, can i place haskellPackages.darcs.env in a generic haskell default.nix file like this(line 17)? https://hastebin.com/oradolahud.bash
<ottidmes>
bqv[m]: I am going offline now, but thanks for letting me know flakes worked out for you! Gives me some hope I will be able to get them to work for me too
<zeta_0>
clever: i don't want to change anything in my default.nix, i just want to be able to place haskellPackages.darc.env in my default.nix, is there an easy way to do this?
<clever>
zeta_0: you could change only the shell.nix file
organixpear has joined #nixos
<zeta_0>
clever: could you pastebin a link? i'm not sure what you mean(also, which shell.nix file are you talking about?)
<clever>
zeta_0: create a shell.nix that has only: with import <nixpkgs> {}; haskellPackages.darcs.env
<clever>
zeta_0: then just run nix-shell
<zeta_0>
clever: pastebin how you would change the default.nix file
<clever>
zeta_0: ignore default.nix entirely
<organixpear>
i have install nix on an lfs build and it works. however, everyone can use commands like 'nix-env -i foo' and it will work. i would prefer to have this functionality for root only. presently i have nix-daemon running along with the nixbld users added. i have run chown -R root:root /nix (the store) along with /var/nix. what am i doing wrong?
<clever>
organixpear: allowed-users in nix.conf can be used to restrict it
<OmnipotentEntity>
organixpear, jfyi, nix-env works for local users, but it will only modify the /nix/store and the user's profile. Are you worried about users filling up your disk?
<organixpear>
OmnipotentEntity: disk space is some concern. i had just read that in the actual nixos nix-env -i operations are restricted to the root user only, so i wanted to have my system following that as well.
<OmnipotentEntity>
That's not actually true, I can nix-env -i foo without being root
<OmnipotentEntity>
it will install foo just for me though
<OmnipotentEntity>
In fact, that's how I test my nixpkg contributions
<zeta_0>
clever: does haskellPackages.darcs.env automatically come with tools like ghc865 and cabal-install, i thought i would have to set them up manually like in my default.nix file?
<organixpear>
that's good to know. just as a side not i think i read that tidbit in the nix manual
<clever>
zeta_0: it will come with whatever ghc is default for haskellPackages
<OmnipotentEntity>
If you can find a link, I can file a bug on the documentation
<clever>
zeta_0: if you want to force 865, use haskell.packages.ghc865.dars.env
<organixpear>
i'll go through my history
<clever>
zeta_0: for cabal-install, use .overrideAttrs to add cabal-install to the buildInputs
<OmnipotentEntity>
thanks organixpear
wiml has joined #nixos
<OmnipotentEntity>
averell, please forgive the ping if I have the wrong person. But are you the package maintainer of Leela Zero?
butterthebuddha has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
butterthebuddha has joined #nixos
captjakk has joined #nixos
butterthebuddha has quit [Client Quit]
wiml has quit [Quit: wiml]
drakonis has quit [Quit: WeeChat 2.7.1]
growpotk- has quit [Ping timeout: 264 seconds]
justanotheruser has quit [Ping timeout: 246 seconds]
wiml has joined #nixos
<zeta_0>
clever: i am not sure how to do this with .overrideAttrs? here's what i have so far in the shell.nix: https://hastebin.com/ocesagodig.js
<{^_^}>
[nixpkgs] @jonringer pushed commit from @bhipple to release-20.03 « pythonPackages.mkl-service: 2.1.0 -> 2.3.0 »: https://git.io/Jvygh
dtz1 has joined #nixos
<dtz1>
sometimes `buildInputs = old.buildInputs or [] ++ [ something ];` which is noteworthy because operators work out the way you'd want but might defensively add parens for, so maybe is a useful pattern
leonardp has quit [Ping timeout: 240 seconds]
<zeta_0>
clever: ok, but can i place the with haskellPackages somewhere so i don't have to write it out over and over again? https://hastebin.com/dirazahisa.js
<clever>
zeta_0: you need to merge line 1 and 3, you duplicated the darcs.env
<clever>
zeta_0: and `with haskellPackages;` can go in several places, after the nixpkgs with, after the old:, or after the buildInputs =
<clever>
zeta_0: basically, anywhere you put a value, you can put a `with expr; value`
<Henson>
does anybody have a suggestion on how I can get the closure of a shell environment to be preserved with a garbage collector root? I tried following the advice here https://nixos.wiki/wiki/Storage_optimization but it just creates an indirect root for the derivation of the shell, and doesn't actually prevent any of the outputs from being garbage collected
<Henson>
unless I'm doing it wrong
<glittershark>
is there something built in to nixpkgs to get the first non-null element of a list?
<glittershark>
I keep expecting || to be dynamically typed but it's not
<glittershark>
or like
<glittershark>
something equivalent to "false || 1 = 1"
<cole-h>
You might be able to do something with the `filter` builtin, or `lib.forEach`, or both
<glittershark>
yeah
<glittershark>
I could write this myself for sure
<glittershark>
just wondering if someone already has :)
iqubic` has quit [Quit: ERC (IRC client for Emacs 28.0.50)]
iqubic has joined #nixos
<{^_^}>
[nix] @gnprice opened pull request #3441 → doc: Files in the store have modes 444/555, not 644/755 → https://git.io/JvyaS
<leonardp>
but binfmt wrapper is slow, so if you can, you should cross compile
<bqv[m]>
somewhat bemused at how switching to flakes has broken my exwm config
<leonardp>
took ~2 days to build chrome on my desktop machine
<pjt_014>
on that note, do you know what the main diffis between using pkgsCross and, say --option system aarch64-linux? I tend to use both because it feels right
<leonardp>
bqv[m]: years ago ppl were joking how emacs, was a whole OS in itself, and now you're telling me there is an emacs WM? ha!
<bqv[m]>
lol
<bqv[m]>
i've been using it for more than a year now
butterthebuddha has joined #nixos
<bqv[m]>
also, there is actually a project that has emacs running as PID2
<leonardp>
pjt_014: if you have your host configured as crossSystem i *think* they are the same
<bqv[m]>
we joke but emacsOS is quickly gonna be a reality..
<leonardp>
haha!
cole-h has quit [Quit: Goodbye]
<pjt_014>
I just looked in the configuration.nix man page--how does that not mess up rebuilding? do nixpkgs attrs work differently?
<pjt_014>
the phrasing is throwing me off I guess
<srhb>
Does anyone know of an out-of-the-box distributed locking/master election/quorum library? I need one process amongst many to be the only one executing certain actions, across a network.
Andrey has joined #nixos
<pjt_014>
srhb: I remember stubling across one (I *think*)
<pjt_014>
'twas writtin in Go
Andrey has quit [Remote host closed the connection]
<srhb>
pjt_014: That sounds like Raft.
Andrey has joined #nixos
<pjt_014>
raft is the algorithm, not the implementation
<leonardp>
pjt_014 i only have binfmt active atm and no crossSystem configured and i do not know if they interfere with each other
<srhb>
pjt_014: Right. But I think it's much like GHC is not Haskell. :-)
veske2 has joined #nixos
veske has joined #nixos
lovesegfault has joined #nixos
<pjt_014>
leonardp: guess I'll just try it out, see what happens
<lovesegfault>
having a hard time using an s3 substituter :/ warning: AWS error fetching 'nix-cache-info': Access denied.
<pjt_014>
havent used it before but all the shiny text seems to be pointing in the right direction
<Andrey>
Hello, I am trying to run the application using buildFHSUserEnv shell, it tries to create directory and files inside /var, I am getting "Permission denied". How can I fix that?
<srhb>
pjt_014: Ah, yes. I was hoping for a drop-in library, but if I have to use etcd or some database that can provide locking, I might go with that..
<pjt_014>
there are some much simpler ones too
gentauro has quit [Read error: Connection reset by peer]
<pjt_014>
bit more specialized though
gentauro has joined #nixos
<pjt_014>
dqlite is the only one I can think of right now
<{^_^}>
[nixpkgs] @worldofpeace pushed 3 commits to master: https://git.io/JvyiA
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to release-20.03: https://git.io/Jvyih
Heirlung has joined #nixos
ZaraChimera has quit [Quit: ZaraChimera]
mexisme_ has joined #nixos
orivej has joined #nixos
vandenoever has joined #nixos
vandenoever has joined #nixos
vandenoever has quit [Changing host]
mexisme_ has quit [Ping timeout: 260 seconds]
c4droid has joined #nixos
c4droid has left #nixos [#nixos]
c4droid has joined #nixos
butterthebuddha has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
butterthebuddha has joined #nixos
butterthebuddha has quit [Client Quit]
<c4droid>
Hi, I just downloaded the nixOS installation image, when I have configured /mnt/etc/nixos/configuration.nix, and run nixos-install command, the installation process is very very slow. It nixOS have offical mirror at Asia? Thanks.
<bqv[m]>
yeah i found that shortly after i sent that
<bqv[m]>
not sure it's working, but meh
lsix has joined #nixos
veske2 has joined #nixos
veske has joined #nixos
Ralith_ has quit [Ping timeout: 258 seconds]
c4droid has left #nixos ["ERC (IRC client for Emacs 26.3)"]
civodul has joined #nixos
Ralith_ has joined #nixos
opthomasprime has joined #nixos
shafox has joined #nixos
veske has quit [Quit: This computer has gone to sleep]
veske2 has quit [Quit: This computer has gone to sleep]
opthomasprime has left #nixos [#nixos]
hmpffff__ has joined #nixos
hmpffff has quit [Ping timeout: 246 seconds]
maddo has joined #nixos
veske2 has joined #nixos
veske has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
orivej has joined #nixos
noudle has joined #nixos
<linarcx>
Guys, how knows informatino about installed packages with nix-env?(like version, descriptions,...)
abathur has joined #nixos
<symphorien>
nix-env -q
<symphorien>
this gives version
<linarcx>
symphorien: NO
<linarcx>
lorr for instance don't show the version.
<symphorien>
it depends on how the derivation was written actually
Andrey has quit [Remote host closed the connection]
<Valodim[m]>
NO?
<symphorien>
if they set name="lorri" you don't get the version
<linarcx>
hmm..
<symphorien>
how about... lorri --version ?
<linarcx>
Yeah, it prints the version
<klys>
is there a simple way with nixos as root to gain access to all installed packages within the various sub-environments?
<NobbZ[m]>
Will show the version of `lorri` thats currently in scope, not necessarily the one installed via `nix-env`
mexisme_ has quit [Ping timeout: 258 seconds]
jfroche has joined #nixos
Ariakenom has joined #nixos
<bqv[m]>
klys: ls /nix/store
<bqv[m]>
:p
abathur has quit [Ping timeout: 256 seconds]
<klys>
bqv[m], are you a seasoned user
<klys>
I figure this might count as a frequently asked question
<NobbZ[m]>
I've not seen this question before in the last 6 months, so it does depend on your definition of "frequent"…
<klys>
though, yes, one might point the output of ls -b /nix/store to the command to create a new environment
<bqv[m]>
nix has a very vague definition of "installed", i don't think your question is very useful
<bqv[m]>
what are you trying to achieve?
mexisme_ has joined #nixos
bahamas has quit [Ping timeout: 264 seconds]
<klys>
along the lines of reclaiming a disk and avoiding deletion of useful material when backing things up.
<NobbZ[m]>
Just do `nix-collect-garbage`
<bqv[m]>
a better approach might be to delete every generation except the current one, check gc roots for useless stuff, then do a garbage collect
<bqv[m]>
then the store is everything necessary and only necessary things
<NobbZ[m]>
As a regular user it will only collect your own environments, as root it will collect from all. It won't delete pathes referenced from any profile.
zupo has joined #nixos
<klys>
paths? patches?
akegalj has joined #nixos
<NobbZ[m]>
For this sake, just another word for "package".
<klys>
ah thank you
<bqv[m]>
is there any reason you're backing up your store though? the majority of things in the store are just copied from hydra, and the rest can be recreated from your configuration and sometimes a few /etc/ files
<bqv[m]>
that's kinda the point
<domenkozar[m]>
https://opencollective.com/nix-errors-enhancement has reached phase 1, Ben has already started work. Second phase crowdfunding is about plumbing existing errors to use the error reporting in existing codebase. If your company wants to support improving Nix ergonomics let me know if you have any questions.
<klys>
now, I am not a regular nixos user, and am merely asking questions to evaluate installing it alongside debian.
hmpffff__ has quit [Remote host closed the connection]
<bqv[m]>
ok, then i would advise you don't bother backing up the store
<klys>
however, the proposition I was making is supposing you inherit a nixos partition from someone who is no longer around.
<bqv[m]>
realistically, i would keep the store OS agnostic. for example, i use the same store for nixos and gentoo, by having it as a separate btrfs subvolume
hmpffff has joined #nixos
<bqv[m]>
if you inherit it, there's not much benefit beyond that it might save you some bandwidth
<bqv[m]>
likely half the things you build will not match that store, and most of those will just be fetched from hydra fresh anyway
<bqv[m]>
it's an exercise in squandering your hard drive space, in my opinion
<klys>
yes, primarily for the purpose of freeing up space and only keeping meaningful materials.
<bqv[m]>
my point is, anything you would want to keep, is not in the store
<NobbZ[m]>
bqv: Avoid BTRFS… It killed 2 machines of mine, because of beeing RO all of a sudden…
<bqv[m]>
i've been using btrfs for best part of a decade, i've never had all these apocalyptic scenarios everyone does on about...
bahamas has joined #nixos
<bqv[m]>
i'm happy with it
mounty has quit [Ping timeout: 240 seconds]
bahamas has quit [Changing host]
bahamas has joined #nixos
Makaveli7 has joined #nixos
<NobbZ[m]>
I was as well, until the point that I was unable to use my main dev machine anymore. Though thanks to nixOS, I was able to replace it within 30 minutes. Setting it up as an Arch again would have cost me days…
o1lo01ol1o has joined #nixos
mojjo has joined #nixos
<bqv[m]>
heh
<bqv[m]>
honestly i've had fewer disk issues on btrfs than any other filesystem, in my history
<mojjo>
Has anyone used nixos-infect recently directly on a digital ocean droplet? It works but it leaves the machine with no network connectivity.
<symphorien>
klys: if your question is "what can I remove from the nix store to reclaim some space" there is nix-du
hmpffff has quit [Ping timeout: 264 seconds]
<NobbZ[m]>
I had less issues with BTRFS than with ext4, but those I had were major and unfixable.
<NobbZ[m]>
Hope ZFS will play better.
<klys>
symphorien, the main reason I want to do what I asked in the first place is to try out or examine documentation for all installed packages.
<symphorien>
I don't really understand
<NobbZ[m]>
Though I need to learn how to transfer some BTRFS things over to ZFS.
<klys>
symphorien, if someone was doing good work and no one is left to tell the tale, then priority one is to identify what works that is present.
<symphorien>
do you still have the original /etc/nixos/configuration.nix ?
<klys>
supposing you do, does that make it easier? This is so hypothetical.
<symphorien>
yes. This is the source code, and the store ic compiled...
bvdw has quit [Read error: Connection reset by peer]
<symphorien>
you only have to read configuration.nix, and the output of nix-env -q and sudo nix-env -q
<klys>
the main thing I know you can do with nixos as a user is to create sub-environments with their own packages. however, I just want to make sure that isn't as messy as it sounds.
bvdw has joined #nixos
<symphorien>
they are usually defined in a file called shell.nix or default.nix in the corresponding folder.
<{^_^}>
[nix] @domenkozar merged pull request #3441 → doc: Files in the store have modes 444/555, not 644/755 → https://git.io/JvyaS
<bqv[m]>
sure :p i don't use snapper though, i was always too lazy to configure it in a way that it doesn't slurp up all of my disk space indiscriminately
<bqv[m]>
i switched to btrfs-snap, and then to nixos..
anders^ has quit [Client Quit]
fendor has joined #nixos
anders^ has joined #nixos
<srk>
,tofu
<{^_^}>
To get a sha256 hash of a new source, you can use the Trust On First Use model: use probably-wrong hash (for example: 0000000000000000000000000000000000000000000000000000) then replace it with the correct hash Nix expected. See: tofu-vim
<symphorien>
heh, I might have a look at btrfs-snap then, thanks for the pointer
linarcx has quit [Quit: WeeChat 2.7.1]
kai_w has quit [Quit: Konversation terminated!]
<Taneb>
I'm trying to add a new jobset to a declerative Hydra project, but after re-evaluating the .jobsets jobset and watching the build succeed, the new jobset doesn't seem to have appeared. How can I debug this?
veske has quit [Quit: This computer has gone to sleep]
veske2 has quit [Quit: This computer has gone to sleep]
veske has joined #nixos
veske2 has joined #nixos
gfixler has quit [Quit: WeeChat 1.9.1]
anders^ has joined #nixos
bahamas has quit [Ping timeout: 258 seconds]
m0rphism has joined #nixos
zakkor has joined #nixos
GDiaX has joined #nixos
<zakkor>
hi guys, in my derivation I'm trying to copy some source files to $out, and then copy some other source files to some sub-directory of $out, overwriting some of the files
<zakkor>
but this doesn't work, and I can't figure out how I can do this
mojjo has quit [Remote host closed the connection]
<zakkor>
LOL, that works
<zakkor>
I didn't think you could do that for some reason
ottidmes has joined #nixos
<bqv[m]>
it gets undone anyway
<bqv[m]>
when nix finishes building a derivation it neuters the permissions
<bqv[m]>
that's the cause of your problem, too
asheshambasta has joined #nixos
<zakkor>
I see, thanks
<asheshambasta>
what does an error message " called without required argument 'mkDerivation' " mean exactly? I'm trying to use callPackage with a custom nix expression.
<bqv[m]>
it means you should probably be using qt5.callPackage, or python3Packages.callPackage, or something similar that provides the expression with the mkDerivation arg
jasom has quit [Ping timeout: 272 seconds]
<asheshambasta>
the "expression" in this case is an anonymous function, right?
kai_w has joined #nixos
<bqv[m]>
yup
<{^_^}>
[nixpkgs] @orivej-nixos pushed commit from @orivej to master « python: pip-tools: 4.3.0 -> 4.5.1 »: https://git.io/JvyDj
jasom has joined #nixos
<asheshambasta>
bqv[m]
mounty has joined #nixos
<asheshambasta>
I'm still finding this hard to understand, though. The callPackages function to use here then depends on one being that supplies these defaults to the anon. function. But how is one supposed to know which callPackage function to use?
mniip has quit [Ping timeout: 606 seconds]
<vandenoever>
mumble on master is currently broken: ALSA lib conf.c:3558:(snd_config_hooks_call) Cannot open shared library (null) ((null): /nix/store/8f9b59c6ppg7vgskdkp37vllgw6xa8l3-qtdeclarative-5.12.3/lib/libQt5Quick.so.5: undefined symbol: _ZN14QDistanceFieldC1ERKS_, version Qt_5_PRIVATE_API)
<ottidmes>
mine is config is multi project with secret files not part of the repo (i.e. not put in the repo encrypted), so thats not going to work for me
<bqv[m]>
well, i'm thinking i'll write a rebuild script that uses git directly as well anyway
<ottidmes>
also, won't you now always have to commit/push even if you just want to try things out?
<ottidmes>
ah, that would counter that
<bqv[m]>
yes
<bqv[m]>
but with flakes you have to anyway
<bqv[m]>
your flake will not build unless the directory is not dirty
<bqv[m]>
so if you have to commit anyway, might as well push
bahamas has quit [Ping timeout: 250 seconds]
<ottidmes>
you sure? should it not work with git:file:// ?
<ottidmes>
"git: A Git repository or dirty local working tree."
<bqv[m]>
it works if you commit locally, yes, but at that point you might as well push anyway
<bqv[m]>
plus, this way i can have my system config available, but also keep my wip content in my $HOME
<bqv[m]>
zimbatm: hey, i assume you're not trying to take the piss or anything, but you do realise you've basically taken my example and added a couple of irrelevant lines...
<ottidmes>
hmm, I have to experiment with this, if I really always have to commit, it would break my workflow, and I might just as well tweak my current system to only use flakes as external dependencies
<bqv[m]>
i've just been leaving WIP commits all over the place
<bqv[m]>
it's actually quite useful
<bqv[m]>
you can see every avenue you've tried
<{^_^}>
[nixpkgs] @Rakesh4G opened pull request #83276 → pythonPackages.py-multibase: init at 1.0.1 → https://git.io/Jvy9k
<bqv[m]>
Mic92: hi, i'm really not sure what you mean, and it'd be faster if we talk here
<Mic92>
bqv[m]: the example shows how one would use NixOS modules from NUR in a flake file. What is the equivalent for packages.
<Mic92>
?
<Mic92>
I have not used flakes at all so far. But I think it should be mentioned in the documentation.
jfroche has quit [Remote host closed the connection]
chagra has joined #nixos
sigmundv_ has joined #nixos
phreedom_ has quit [Remote host closed the connection]
phreedom has joined #nixos
jfroche has joined #nixos
knupfer has joined #nixos
jfroche has quit [Remote host closed the connection]
<bqv[m]>
Mic92: the overlay works, so pkgs.nur is a thing, although i seem to have botched my fork
<bqv[m]>
but as an example i added pkgs.nur.repos.mic92.pkgs.hello-nur to environment.systemPackages, and it fetched the my NUR flake
<mojjo>
hi!! maybe someone has an advice for me: I'm setting up a CI pipeline with nixops, travis and digital ocean. It actually works. My only problem is that I cannot persist the nixops state on travis, so everytime the network is created anew. not so ideal :) So there is a way to export the nixops state to a file and I could theoretically save that in
<mojjo>
the travis cache. but not really, as the cache expires after a month plus the state file contains private keys. what would you do? recommendations of different services welcome as well ...
<{^_^}>
#83135 (by mmahut, 2 days ago, merged): tessera: init at 0.10.2
<evils>
2 days ain't instant..., it does mean ofborg can check it before merging
<Valodim[m]>
no review, no time passed between push and merge, didn't even bother to fill in the "Things done" template. that feels very sloppy
<Valodim[m]>
oh, sorry, I misread the timestamp
<evils>
even if it was instant, still better to have ofborg check it rather than just merging one's work directly i suppose
mallox has joined #nixos
<Valodim[m]>
I'm surprised there's no policy against merges by the committer with no second pair of eyes at all
<ottidmes>
bqv[m]: About the dirty working directories: https://github.com/NixOS/nix/issues/3231 so I am not going to use flakes for now, and just continue using my own system until that is addressed
<{^_^}>
nix#3231 (by Ma27, 18 weeks ago, open): Using dirty git checkouts as flakes
<evils>
the merged version is 2 releases and 3 months behind though
<gchristensen>
that is a bit interesting indeed
<gchristensen>
but eh
<evils>
gchristensen: aren't you the lead security person 'round here?
<{^_^}>
[nixpkgs] @hax404 opened pull request #83285 → teeworlds: python2 -> python3; add menu item → https://git.io/Jvy7O
<gchristensen>
hi :) I'm involved
bahamas has joined #nixos
<gchristensen>
I'd say it is weird that he didn't package the latest, and probably should have -- and it'd be good to talk to tem about it
<Valodim[m]>
So, it's a PR that was created and merged by the same person, without so much as a bot build or reactions by anyone. I guess my question is where this falls in the existing PR culture, between "violating policy", "not violating policy but kind of not how things should be", "fine for now but maybe not ideal on the long run", and "the way it's supposed to be"
<gchristensen>
Valodim[m]: the bot did build it on 3 platforms
<MichaelRaskin>
There was a bot build
<gchristensen>
and not violating policy and fine or now, and in the future it'd be nice to have 2 sets of eyeballs
<gchristensen>
imo anyway
<MichaelRaskin>
Definitely not violating policy, fine until no non-committer PRs get stuck in limbo
cosimone has quit [Quit: Terminated!]
cjpbirkbeck has joined #nixos
<Valodim[m]>
I kind of assumed that merging one's own PRs would be frowned upon (unless absolutely trivial). I admit this does slightly reduce my confidence in the package quality of the nixpkgs tree. :\ although I do understand it's a very practical tradeoff
bahamas has quit [Ping timeout: 256 seconds]
<gchristensen>
understood
<Valodim[m]>
ftr I didn't want to accuse anyone of violating policy, I was just curious what the policies and lived culture was :)
abathur has joined #nixos
<simpson>
Valodim[m]: Looking at the PR, there *were* bot builds. 16 checks passed.
<simpson>
Ah, I am late.
<Valodim[m]>
simpson: yeah I saw. I had assumed that the bot would only built if instructed to. I figure that's only the case for PRs of users not in the trust list of the bot?
<Valodim[m]>
guess I'll go read the relevant RFC processes now, I feel underinformed to really partake in the discussion :)
<gchristensen>
Valodim[m]: a few people push direct to master, this has come down significantly over the last 3 years. sometimes PR authors, if they're maintainers, merge their own PRs. this isn't abnormal.
<gchristensen>
the bot builds every PR now by default
<evils>
gchristensen: any idea when it'll start checking for unquoted URLs?
CMCDragonkai1 has quit [Ping timeout: 256 seconds]
<gchristensen>
hmm I'm not sure how I would make it do that
<{^_^}>
[nixos-homepage] @garbas pushed 0 commits to common-learn-page: https://git.io/Jvy7i
<evils>
nice
<Valodim[m]>
thanks for the explanations everyone
<simpson>
Extremely automated workflows are necessary at scale. I'd hope that, in the future, we can have more PRs that not only are automatically reviewed like this, but also automatically merged and automatically reverted upon breakage.
<gchristensen>
Valodim[m]: it is nice to have people care :)
<MichaelRaskin>
Also: in full accordance with the RFC motivation, we do have a couple of bugs where Nix and hnix parse unquoted URLs differently
abathur has quit [Ping timeout: 256 seconds]
<gchristensen>
MichaelRaskin: sounds like a bug in hnix :P
<Valodim[m]>
I guess this also has to do with a very fuzzy feeling of "fairness" from a non-committer's perspective
<srk>
would be cool to have some stats about PRs flowing into nixpkgs
abathur has joined #nixos
zupo_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<gchristensen>
Valodim[m]: do you play?
<Valodim[m]>
gchristensen: no, but my girlfriend does :)
<gchristensen>
nice, my wife does too
<Valodim[m]>
for the record btw, I very much enjoyed being invited to the nixos-maintainers team after my first PR. I believe stuff like that can be very motivating for new maintainers. big thanks for RFC-0039 👍️
<q3k>
no, i'm mostly familiar with go itself, and its interaction with bazel if anything
<xfix>
all its deps is mostly Rust compiler
<xfix>
to be fair
<srk>
evils: yeah, will check it out when I'm done working :)
zupo has joined #nixos
<flokli>
for each terraform module. but it seems some of them fail to build. and I'm not sure if it's missing vendor files, or if we're building them wrongly
<srk>
that sounds scary
<xfix>
and of course, Rust compiler requires LLVM
<srk>
xfix: that means it does a bunch of bootstrapping itself at build time right
<q3k>
flokli: is tht package (google.golang.org/api/sql/v1beta4) in goDeps?
<arianvp>
is buildGoPackage vendoring dependencies? Or does it do the weird lockfile sha256 hack?
<xfix>
not quite yet
bennofs1 has joined #nixos
<xfix>
currently nixpkgs bootstraps Rust compiler from binary
<xfix>
but I suppose in the future it may use mrustc, who knows
<arianvp>
xfix: we should! the guix people even did all the homework for us :)
<q3k>
flokli: oh, you should be using buildGoModule
<q3k>
flokli: buildGoPackage is the old method
<gchristensen>
arianvp: doesn't mrustc build an early rust, and then you have to build 100 versions of rust sequentially?
<q3k>
flokli: that pulls in all gomod defined deps
<arianvp>
gchristensen: that's the name of the game yes
<gchristensen>
yeah no thanks
<xfix>
mrustc can build 1.29, and then you have to build Rust compilers version by version
<srk>
xfix: I meant firefox, rustc is not that bad deps wise
<flokli>
q3k: but this means I need to pass another hash, as it's a fixed-output derivation? why is this even necessary if a vendor folder is provided?
<arianvp>
ah but buildGoModule uses fixed output derivations hack which is nit great
<q3k>
flokli: go vendor support is only via gomod iirc
<arianvp>
gchristensen: i'm not sure how friendly rust is to bootstrapping; it might be less bad than that
<q3k>
flokli: go build without gomod does not look at the vendor directory
<q3k>
flokli: go build without gomod is basically super old and super basic, and relies on all deps being in GOPATH
<tilpner>
mrust is a massive shortcut over bootstrapping via ocaml
<srk>
:D cool
<q3k>
flokli: go build with gomod is the new shit, even if you don't pull stuff from outside but just use a vendor subdir
<gchristensen>
Valodim[m]: I just left a couple comments :/
<tilpner>
It might still not please hydra though
<flokli>
q3k: so I need to check why `modSha256` is required, and if we can make it optional if a vendor folder is part of the source.
<arianvp>
welll then we need to change hydra :P
<q3k>
flokli: but you will still indeed need a hash for fixed-output, but that output should be stable
<q3k>
no, i think you need it regardless
<q3k>
again, don't remember
<q3k>
because this just invokes go with gomod, and doesn't know if it's actually going to pull anything or just use the vendor directory
<flokli>
q3k: yeah, it just makes automatic updating very painful and error-prone, as you don't notice things changed
<q3k>
maybe you could hack gobuildmodule to have a 'no external deps' flag that would prohibit pulling from the internet, but also allow not having to have a fixed output
<q3k>
not sure, `git blame` that part of nixpkgs and poke the authors ^^
<flokli>
yeah, that could work :-)
<flokli>
thanks :-)
<q3k>
for now just try it with the modsha256
<q3k>
if that works, then figure out how to get rid of that
<flokli>
Aye
<q3k>
but regardless, the legacy non-go-mod toolchain will not by itself use the vendor directory without some ugly-ass hacks like pointing GOPATH to some synthetic path with the vendor directory carefully copied there
<adisbladis>
Tbh I wanna rewrite vgo2nix and eliminate buildGoModule completely
<q3k>
(and there's some legacy tools that help with that, but you almost definitely shouldn't use them)
<q3k>
sure, that's another approach
<q3k>
(vgo == go modules)
opthomasprime has joined #nixos
<q3k>
(for flokli )
<xfix>
Go is much easier to bootstrap, the compiler is designed to be possible to compile with Go 1.4 (Go 1.4 being the last release to not be written in Go itself)
<q3k>
(vgo is how it was called experimentally before it landed in go mainline)
domogled has joined #nixos
<xfix>
(or you can use gccgo)
<adisbladis>
xfix: I guess you already know this, but that's exactly how we bootstrap go in nixpkgs.
Shoubit has quit [Remote host closed the connection]
waleee-cl has joined #nixos
cosimone has joined #nixos
<{^_^}>
[nixpkgs] @romildo merged pull request #83232 → ant themes: should not use a fixed output path → https://git.io/JvyCE
<jasom>
I'm getting some errors trying to run a program under wine, and some googling tells me that I may need 32-bit graphics drivers installed. Is there a way to do this under nixos?
<tilpner>
hardware.opengl.driSupport32Bit
<tilpner>
And maybe hardware.pulseaudio.support32Bit
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<nuriaion>
i would like to upgrad nodePackages.prettier. How can i do that? (I was able to update a normal package in nixpkgs but i didn't understood the nodePackags and haven't found a tutorial)
<nobrak>
x509: certificate signed by unknown authority
ottidmes has joined #nixos
<nobrak>
In troubleshooting I made sure that I had ca-certificates installed and updated
c0c0 has quit [Ping timeout: 260 seconds]
xkapastel has joined #nixos
fenedor is now known as fendor
oever has joined #nixos
zupo has joined #nixos
vandenoever has quit [Ping timeout: 246 seconds]
ottidmes has quit [Ping timeout: 256 seconds]
ottidmes has joined #nixos
<{^_^}>
[nixpkgs] @matthiasbeyer opened pull request #83299 → tut: init at 0.0.2 → https://git.io/Jvyhm
absoluutely_ has joined #nixos
leonardp has joined #nixos
<{^_^}>
[nixpkgs] @vcunat pushed commit from @marsam to release-20.03 « Merge #82081: swiProlog: openssl 1.0.2 -> openssl 1.1 »: https://git.io/JvyhG
c0c0 has joined #nixos
teto has quit [Ping timeout: 246 seconds]
KeiraT has quit [Ping timeout: 240 seconds]
absoluutely_ has quit [Client Quit]
ianmjones[m] has joined #nixos
<leonardp>
hello, i configured one of my hosts as binary-cache/distributed builder. now when doing a build on a client i get a lot of warnings like: `warning: substituter 'https://cache.nixos.org' does not have a valid signature for path '/nix/store....'`
KeiraT has joined #nixos
orivej has joined #nixos
<leonardp>
is this expected bahaviour could this indicate a miscofiguration with the substituters?
<clever>
leonardp: the default value of nix.binaryCachePublicKeys was lost
<{^_^}>
[nixpkgs] @thoughtpolice pushed commit from @emilazy to master « openresty: 1.15.8.2 -> 1.15.8.3 »: https://git.io/JvSee
<glittershark>
"garbage collection" isnt' the right term here really since I'm not registering gc roots, I just want to delete all store paths that haven't been read for some amount of time
hmpffff has joined #nixos
hmpffff_ has quit [Ping timeout: 256 seconds]
<bqv[m]>
glittershark: that sounds like a terrible idea
<clever>
glittershark: it is also invalid for a thing to depend on something that doesnt exist
<clever>
glittershark: so you would need to recursively delete many things
<tilpner>
Ehh, it sounds like "I have limited storage space and want to maximise cache hits"
<tilpner>
They're not asking about store GC, just cache GC
<bqv[m]>
imagine if you left your machine on too long and it deleted the "poweroff" binary
<clever>
bqv[m]: nix will respect the gc roots by default
captjakk has joined #nixos
<tilpner>
glittershark: You'll probably have to build it yourself, I haven't seen anything like that
<glittershark>
yeah the idea here is only to delete leaves in the store
<bqv[m]>
clever: but then it's just a partial gc, so i assumed he meant not doing that
<{^_^}>
[nixpkgs] @hedning opened pull request #83303 → gtk3: fix clipboard related crashes in wayland → https://git.io/JvSen
<bqv[m]>
because a partial gc would be entirely pointless
<glittershark>
not to invalidate any cross-references
<bqv[m]>
even if the alternative is recursively horrifying
<glittershark>
so I guess the more precise description would be: I want to delete all store paths that both aren't depended on by another store path, and also haven't been read for some amount of time
<glittershark>
tilpner is on the money here, this is primarily to save on cloud bills for our private binary cache
<glittershark>
because storage may be cheap, but it isn't free
<clever>
glittershark: the default algo for `nix-collect-garbage --max-freed 10g` is to delete random un-rooted paths, until they total to 10gig
<glittershark>
"recursively horrifying" is also probably an accurate description of what I'm looking for
<glittershark>
hmm ok
<bqv[m]>
so, stuff that isn't rooted, will get cleaned up by a gc anyway, the only thing i can think of that might otherwise be useful is something to point out gcroots that aren't your system and that haven't been touched for a long time, but even deleting those randomly and automatically sounds like such a bad plan
<glittershark>
if I could replace "random" with "LRU" there that'd be awesome
<bqv[m]>
oh
<bqv[m]>
lmao
<clever>
glittershark: its random, so it doesnt pick the same thing multiple times
<{^_^}>
[nixpkgs] @nh2 opened pull request #83304 → chromium: Enable swiftshader by default. → https://git.io/JvSez
<glittershark>
are there really that few people that are running a private binary cache for a team of developers that nobody's wanted to do this yet? this seems pretty obvious to me
<tilpner>
glittershark: How is your cache stored and served?
<glittershark>
well, right now it's just `nix-serve` serving off of a GCE persistentdisk, but I'm gonna swap it out for a GCS bucket with tweag/nix-store-gcs-proxy very soon
<glittershark>
because someone pointed out very rightly to me that nix-serve is a little silly at the scale I'm using it
<tilpner>
That's very important
cosimone has quit [Read error: Connection reset by peer]
<tilpner>
With nix-serve, you indeed want to maintain the "if I have X, I have the entire closure of X" property
<bqv[m]>
most people here just use hydra...
<tilpner>
But you don't really need to do that if you store it "cold" (not in an active store)
<clever>
hydra does have the ability to upload to an S3 bucket
cosimone has joined #nixos
<clever>
but there isnt much to GC the bucket
<glittershark>
right, I'm still assuming that S3 bucket gradually fills up over time
<clever>
and the bucket has no size limit, so we just ignore it for the most part
<tilpner>
And now you can just iterate through your objects, find the ones with the earliest access time, and delete until your target cache size is met
<captjakk>
anyone have advice on how to build android projects on nixos?
<glittershark>
sure, I was thinking I could just do this by disk atime
<tilpner>
I wonder when that's going to break...
<glittershark>
this is really mostly about money
<tilpner>
Yes, ideally you can rebuild missing parts of the closure, but it doesn't always work perfectly
<glittershark>
> The operation --delete deletes the store paths paths from the Nix store, but only if it is safe to do so; that is, when the path is not reachable from a root of the garbage collector
<{^_^}>
error: syntax error, unexpected ',', expecting ')', at (string):292:72
<glittershark>
so yeah `nix-store --delete` appears to do what I'm looking for here
<clever>
glittershark: and add `--store gsc://something` to reference that bucket
<glittershark>
ok sounds like I'm gonna have to roll this myself
<glittershark>
good excuse for a blog post I guess :P
<clever>
glittershark: i think i saw a script for this in hydra...
nuriaion has quit []
<glittershark>
oh hmm
<glittershark>
I'm not using hydra but I wonder if I could rip it out
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<infandum>
Does anyone use static-haskell-nix? I'm trying to build a binary with static-haskell-nix from a stack project but I'm running into issues with p11-kit.
<mjrosenb>
morning, all. I have a project where I unfortunately need two java jdks installed. a) is there a way for my .nix file to depend on two different jdks?
<mjrosenb>
b) I'm currently referencing JAVA_HOME directly in a script for a different derivation which I'm overriding. Will I have to thread the two different java homes through the layers between my project's .nix file, and that script?
akegalj has joined #nixos
cosimone has quit [Quit: Terminated!]
zaeph has quit [*.net *.split]
swapgs has quit [*.net *.split]
v0|d has quit [*.net *.split]
tilpner has quit [*.net *.split]
nek0 has quit [*.net *.split]
zemm has quit [*.net *.split]
Nafai has quit [*.net *.split]
b42 has quit [*.net *.split]
aw has quit [*.net *.split]
MichaelRaskin has quit [*.net *.split]
aither has quit [*.net *.split]
m1cr0man has quit [*.net *.split]
vesper11 has quit [*.net *.split]
cross_ has quit [*.net *.split]
silver_hook has quit [*.net *.split]
Diogo has quit [*.net *.split]
hexagoxel has quit [*.net *.split]
never_released has quit [*.net *.split]
cyphase has quit [*.net *.split]
wolke has quit [*.net *.split]
nckx has quit [*.net *.split]
jaeckel has quit [*.net *.split]
cptchaos83 has quit [*.net *.split]
pistache has quit [*.net *.split]
mzan has quit [*.net *.split]
kisoneca- has quit [*.net *.split]
pistache has joined #nixos
zemm has joined #nixos
v88m has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « elementary-xfce-icon-theme: 0.14 -> 0.15 »: https://git.io/JvSfX
<jluttine>
how to access the build directory in installPhase? i mean, the directory that i get if i run nix-build with --keep-failed and then go to that directory. i can see something that i'd like to use in the install phase i'm writing..
jaeckel has joined #nixos
<clever>
jluttine: that should still be the working directory
<clever>
jluttine: and $NIX_BUILD_TOP always points to it
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « nss_wrapper: 1.1.8 -> 1.1.10 »: https://git.io/JvSfM
<infandum>
jluttine: Is it like $out ?
o1lo01ol1o has joined #nixos
<Guest16427>
what can explain 32-bits application not running ("exec format error") from a FHS environment (that does not have "multiPkgs = null")
<Guest16427>
all 32-bit libraries seem to be available, including "ld-linux.so.2", but even "ld-linux.so.2" fails to be executed with "exec format error"
wiml has joined #nixos
hlavaty` has quit [Remote host closed the connection]
o1lo01ol1o has quit [Ping timeout: 240 seconds]
<jluttine>
clever: they are quite different. i get significantly different output for these two commands:
<jluttine>
ls -l
<jluttine>
ls -l $NIX_BUILD_TOP/linphone-desktop-*/
<clever>
jluttine: what does `pwd` print out in the installPhase?
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « opencl-clhpp: 2.0.10 -> 2.0.11 »: https://git.io/JvSJw
kfound has quit [Ping timeout: 246 seconds]
kfound has joined #nixos
<martin__>
I'm trying to create a nix shell for a c++ codebase (this one https://github.com/ethereum/solidity). I've managed to get it to compile inside the shell successfully on my nixbook, but running the program yields the error:
<martin__>
`Unknown exception during compilation: std::bad_cast`
<{^_^}>
[nixpkgs] @savoisn opened pull request #83308 → aws-sam-cli: update cookiecutter version to 1.7.0 since 1.6.0 doesn't exists anymore → https://git.io/JvSJS
hmpffff has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « namecoind: nc0.15.99-name-tab-beta2 -> nc0.19.1 »: https://git.io/JvSUh
<glittershark>
after using nix for development and in production for a little over 6 months, I think today is finally the day I install nixos on a laptop
<glittershark>
it's TIME
<glittershark>
have a garbage macbook air lying around that's perfect for this
cizra has joined #nixos
<cole-h>
ma27[m]: Are you planning on addressing all the packages that use `target/release` directly in your PR?
bahamas has joined #nixos
<cole-h>
Or the real question: should I close my PR that was supposed to help with that and let you deal with it
<{^_^}>
[nixpkgs] @jonringer pushed commit from @poscat0x04 to master « vimPlugins.ats-vim: init at 2020-03-03 »: https://git.io/JvSTt
<DamienCassou>
hi everyone. I would like to install nixos on an ARM 32bits. Everything I find on https://nixos.wiki/wiki/NixOS_on_ARM seems outdated (e.g., downloads from 2018 and mentions of bugs fixed one year ago). Should I follow instructions there anyway or is there a better way?
<Fare>
and where is the documentation / tutorial for it?
cizra has joined #nixos
<marek>
Valodim[m]: well, I would be more than happy to get more eyes on it! I hope you are doing at least few reviews a day to help with the current situation :)
shibboleth has quit [Remote host closed the connection]
<terlar>
Any ideas how to debug a slow nix-shell, all my nix-shells have become slow, even the simplest `nix-shell -p nix --pure --run exit` takes 8 seconds. I tried running with the latest release-19.09 and the latest nixos-unstable nixpkgs by passing -I, still the same result. I don't remember this being this slow earlier...
shibboleth has joined #nixos
<clever>
terlar: what about `time nix-shell -p --run exit` ?
<clever>
terlar: try this, then pastebin the entire profile.json file
<energizer>
terlar: if you do `time nix-instantiate '<nixpkgs>' -A hello --arg config '{}' --arg overlays '[]'` a couple times in a row does it remain slow?
<clever>
jonreeve[m]: is it nixos or another distro?
kleisli has joined #nixos
<terlar>
Is that supposed to return a derivation? That was actually fast
<jonreeve[m]>
NixOS
<clever>
terlar: yeah, it should return a drv file
<jonreeve[m]>
I haven't changed any of my permissions or anything
<clever>
jonreeve[m]: cabal2nix is trying to write to /nix/var while inside a nix sandbox, so there is a problem with the build process
sleepingforest has quit [Quit: bye!]
m1sosoup has quit [Ping timeout: 256 seconds]
<clever>
jonreeve[m]: can you pastebin `nix show-derivation /nix/store/02szzrd1w7wqmnkh9is5qlx64m7r6f1z-cabal2nix-color-word-analyzer.drv` ?
sleepingforest has joined #nixos
<jonreeve[m]>
clever: Sure, one sec
drakonis has joined #nixos
wavirc22 has quit [Quit: ZNC 1.7.2+deb3 - https://znc.in]
sleepingforest has quit [Client Quit]
nixproblemsooof has joined #nixos
wavirc22 has joined #nixos
opthomasprime has left #nixos [#nixos]
o1lo01ol1o has joined #nixos
<nixproblemsooof>
I keep getting hangs in "nixos-20.09pre217537.d96bd3394b7" and tailing the system log I see "nixos xserver-wrapper[1150]: (EE) client bug: timer event1 debounce short: scheduled expiry is in the past (-14ms), your system is too slow" even though I'm on a very fast system. How should I proceed?
<ma27[m]>
cole-h: ah no, I just wanted to wait for a few more reviews :)
<clever>
terlar: thats reporting only 0.175 sec, which is pretty speedy, weird
<{^_^}>
[nixpkgs] @orivej-nixos pushed commit from @orivej to master « gucharmap: do not depend on libtool after the switch to meson »: https://git.io/JvSkk
<terlar>
yes, now it is fast... I ran it a couple of times as energizer suggested as well, all execs up until that dump was slow
wavirc22 has quit [Client Quit]
<clever>
jonreeve[m]: its a json output, so if you rename the file in the gist, the syntax highlighting will be a lot more pretty
<terlar>
But with the overlays enabled again, it is slow, so I guess I have some bad overlays
<clever>
jonreeve[m]: what does the cabal file in /nix/store/fvka3nd7slnqgjvfkjg649v77y10lc89-color-word-analyzer look like? can you add that file to the gist?
<clever>
hpfr[m]: and behind the scenes, nixos-rebuild is just using nix-env to manage the system profile, so you can use this to list/delete generations
<jonreeve[m]>
Awesome, that did the trick, thanks
ChimeraZara has joined #nixos
abathur has joined #nixos
levdub has quit [Ping timeout: 258 seconds]
levdub5 is now known as levdub
zeta_0 has joined #nixos
ZaraChimera has quit [Ping timeout: 265 seconds]
<hpfr[m]>
Oh ok thanks
ChimeraZara is now known as ZaraChimera
<hpfr[m]>
Yup
<hpfr[m]>
Oh, ok
wiml has quit [Quit: wiml]
o1lo01ol1o has joined #nixos
<Valodim[m]>
marek: can't say it's "a few a day", but yes I am! :)
<{^_^}>
[nixpkgs] @hedning merged pull request #83303 → gtk3: fix clipboard related crashes in wayland → https://git.io/JvSen
<{^_^}>
[nixpkgs] @hedning pushed to staging « gtk3: fix clipboard related crashes in wayland (#83303) »: https://git.io/JvSI8
<Tomas69>
hey everyone what are the system reqierements for barebones nix? how would it run on an old laptop witha singlecore cpu 1.7ghz and 1gb ram with openbox?
<clever>
Tomas69: without -A, nix-env takes a lot more ram
<tilpner>
I guess it might use slightly more, because there can be multiple versions of the same dependency loaded into memory
srid has joined #nixos
<Tomas69>
what does -A do?
<energizer>
Maximum resident set size (kbytes): 580892
<clever>
Tomas69: lets you select an attribute from a set
<srid>
When using `developPackage`, is it possible to configure to add custom nix packages to the resultant Haskell executable derivation? So that those packages are installed when doing `nix-env -i`?
<{^_^}>
[nixpkgs] @drewrisinger opened pull request #83320 → pythonPackages.qiskit-ibmq-provider: init at 0.4.5 → https://git.io/JvSLR
ixxie has quit [Ping timeout: 240 seconds]
c0c0 has quit [Ping timeout: 240 seconds]
<zeta_0>
clever: is there a way for me to see what's inside? haskell.packages.ghc865.darcs.env , i want to take a look at all of darcs's build dependencies inside of nix's darcs.env?
ixxie has joined #nixos
<clever>
zeta_0: vi ~/apps/nixpkgs/pkgs/development/haskell-modules/hackage-packages.nix and go down to line 64228
<atemu12[m]>
Tomas69: Try adding zram to your config, it has allowed me to build a ton of packages on my 1GB raspberry pi 2 and allowed my 8GB laptop to do a 12GB git gc on Nixpkgs once
<Tomas69>
atemu12[m] il look into what zram is thank you
<atemu12[m]>
Tomas69: You might also want to considere doing a 32bit installation because pointers are half the size.
<Tomas69>
atemu12[m]its a 32bit cpu anyway
<zeta_0>
clever: i don't have those directories, do i need to clone nixpkgs from github?
<clever>
zeta_0: you will also find nixpkgs in ~/.nix-defexpr/
<atemu12[m]>
zeta_0: You have a checkout of Nixpkgs in your Nix store already
c0c0 has joined #nixos
<atemu12[m]>
zeta_0: It should be at /nix/var/nix/profiles/per-user/root/channels/nixos/ IIRC
T_S_ has joined #nixos
<zeta_0>
thanks, i'll take a look now
<zeta_0>
also, wow, nixpkgs has over 200,000 commits, that's insane
<T_S_>
To test a large package set, should I use `nix build` or `nix-build`? One seems newer and cooler, the other seems like it has some documentation. Thoughts?
lovesegfault has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to master « chromium: I accidentally added the webrtc patch into the wrong section »: https://git.io/JvSLN
pistache has quit [Remote host closed the connection]
<zeta_0>
clever: channels just has a manifest.nix file with [ ] only, so everything must be in channels_root
pie_[bnc] has quit [Quit: No Ping reply in 180 seconds.]
Tomas69 has quit [Remote host closed the connection]
pie_[bnc] has joined #nixos
pie_[bnc] has joined #nixos
pie_[bnc] has quit [Changing host]
shibboleth has quit [Remote host closed the connection]
shibboleth has joined #nixos
gustavderdrache has joined #nixos
<atemu12[m]>
<zeta_0 "also, wow, nixpkgs has over 200,"> Nixpkgs has slightly more activity than the Linux kernel currently
<atemu12[m]>
It's insane
frihetskampen has joined #nixos
<zeta_0>
clever atemu12[m] : ok, i finally found hackages-packages in: /home/zeta/.nix-defexpr/channels_root/nixos/nixpkgs/pkgs/development/haskell-modules
<zeta_0>
typo hackage-packages.nix
<frihetskampen>
Hello. I have been trying to make nixOS use the nixpkgs-unstable branch for packages, and now while upgrading I'm getting this error "error: creating symlink from '/nix/var/nix/profiles/system-83-link.tmp-11689-826462435' to '/nix/store/v5qbnijrnnwbbdp77q1367vrgc15anf2-nixos-system-yttrerymden-20.09pre2175configuration.nix [RO] 79,3
<clever>
frihetskampen: does dmesg say anything near the end?
philr has joined #nixos
<zeta_0>
atemu12[m]: more activity than the linux kernel, damn, that is crazy, but ya, nix manages a lot of packages
<frihetskampen>
@clever I see nothing out of the ordinary
<clever>
frihetskampen: can you screenshot the terminal window where you get the error?
<zeta_0>
clever: in hackage-packages.nix i went to line 64228, but i don't see darcs and 'anywhere near there?
ChimeraZara has joined #nixos
<clever>
zeta_0: it may be on a different line for you, search for darcs
ZaraChimera has quit [Disconnected by services]
<frihetskampen>
clever I had messed up my NIX-PATH, and it turned out I now did NOT run as sudo...I can't believe I missed that. Problem solved. However. I can't seem to make nix use the unstable channel for packages.
ChimeraZara has quit [Client Quit]
<clever>
frihetskampen: use `nix-channel --add` as root, to point the nixos name to nixos-unstable
<clever>
frihetskampen: if you want unstable to be the default for everything
<clever>
zeta_0: the one that is assigning a value to darcs
linarcx has quit [Quit: WeeChat 2.7.1]
<frihetskampen>
clever thank you, that's what I've been trying. I've been at it for a couple of hours. It seems to work now finally. Thanks again
<clever>
zeta_0: on line 62255
chagra has quit [Ping timeout: 240 seconds]
chagra has joined #nixos
opthomasprime has joined #nixos
<zeta_0>
clever: thanks, darcs has a lot of dependencies, no wonder i was not able to build it
ixxie has quit [Ping timeout: 256 seconds]
ZaraChimera has joined #nixos
opthomasprime has quit [Remote host closed the connection]
<zeta_0>
clever: would you recommend that i use, cabal build or nix-build, to build darcs, the instructions on the darcs website only covers cabal, but i'm using a shell.nix in a nix-shell, so i'm not sure if i should use nix instead? a long time ago set a setting to integrate nix with cabal, so i don't think it should matter, hopefully
ZaraChimera has left #nixos [#nixos]
<{^_^}>
[nixpkgs] @doronbehar opened pull request #83322 → pythonPackages.beets: add unconditionally deps needed for checks → https://git.io/JvSqZ
<clever>
zeta_0: `cabal build` inside that shell should do, if you are trying to develop darcs
<clever>
zeta_0: nix-build with an override against src, if you want to use it long-term
frihetskampen has quit [Remote host closed the connection]
<zeta_0>
clever: what do you mean by override against src?
<lovesegfault>
Can I make nix-copy-closure or nix copy handle drvs?
<lovesegfault>
like, I want to nix copy the drv and all it's deps
<clever>
lovesegfault: nix-copy-closure can already handle drv files
<clever>
lovesegfault: you can also use `--option substituters ssh://10.0.5.155` to make nix treat the remote machine as a binary cache
<lovesegfault>
:O
<lovesegfault>
that's neat
<clever>
zeta_0: once your in the nix-shell, you can mostly use cabal as normal
<zeta_0>
clever: sorry for all the questions, i'm just trying to get the correct nix workflow
<zeta_0>
clever: ok, i'm good to go, now that i build darcs, the only thing that i need to do next after building darcs is getting ghcide to work with it, i was having problems last time, but i guess the people from ghcide are going to fix it soon which is great
<zeta_0>
clever++
<{^_^}>
clever's karma got increased to 368
<zeta_0>
thanks for the help everyone, i'm going to head out to the #darcs channel now
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<{^_^}>
[nixpkgs] @flokli merged pull request #83254 → nixos/mysql: fix service so it works with mysql80 package [20.03 backport] → https://git.io/Jvyuw
<{^_^}>
[nixpkgs] @flokli pushed 3 commits to release-20.03: https://git.io/JvSqM
<{^_^}>
[nixpkgs] @edolstra pushed 3 commits to release-19.09: https://git.io/JvSqD
zeta_0 has left #nixos ["rcirc on GNU Emacs 26.3"]
<{^_^}>
[nixpkgs] @edolstra pushed 3 commits to release-20.03: https://git.io/JvSqH
lsix has quit [Quit: WeeChat 2.7.1]
erasmas has quit [Quit: leaving]
zupo has joined #nixos
shibboleth has quit [Remote host closed the connection]
shibboleth has joined #nixos
knupfer has quit [Remote host closed the connection]
shibboleth has quit [Remote host closed the connection]
civodul has quit [Quit: ERC (IRC client for Emacs 26.3)]
growpotkin has joined #nixos
aveltras has quit [Quit: Connection closed for inactivity]
o1lo01ol1o has quit [Remote host closed the connection]
toppler has joined #nixos
<{^_^}>
[nixops] @grahamc opened pull request #1264 → Example NixOps State Backends → https://git.io/JvSm3
toppler has quit [Quit: ERC (IRC client for Emacs 26.1)]
maddo has quit [Quit: See ya]
cosimone has quit [Ping timeout: 246 seconds]
growpotkin has quit [Ping timeout: 246 seconds]
tokudan has quit [Remote host closed the connection]
astrofog has joined #nixos
Jonathan44 has joined #nixos
zeta_0 has joined #nixos
nobrak has quit [Remote host closed the connection]
tokudan has joined #nixos
__monty__ has quit [Quit: leaving]
tokudan has quit [Remote host closed the connection]
<zeta_0>
clever: i almost forgot to ask, is there an xmonad.env option similar to the darcs env.option that you showed me earlier? https://hastebin.com/taletebebo.js
<zeta_0>
clever: i'm also planning on contributing to the xmonad open src project as well
tokudan has joined #nixos
<{^_^}>
[nixpkgs] @jonringer opened pull request #83323 → pythonPackages.azure*: bump many → https://git.io/JvSmi
<zeta_0>
clever: i've already built ghc with ghc.nix, so i'm good to go with building ghc with nix
<clever>
zeta_0: every single package in haskellPackages has a .env on it
<Jonathan44>
Anyone around for some assistance? I am trying to build a C library which depends on another C library (both build with cmake).
o1lo01ol1o has joined #nixos
<Jonathan44>
I got the dependency built but the main library is claiming to not find dependency.
<Jonathan44>
Am I correct that `make install` will not work with nix?
<zeta_0>
clever: that's great, so basically i just need to substitute darcs with xmonad in the shell.nix script i just paste binned?
<simpson>
Can you share what you've tried so far? Is the failure at configure time or build time?
<Jonathan44>
Configure time (cmake)
<clever>
zeta_0: yeah
growpotkin has joined #nixos
<atemu12[m]>
Jonathan44: Stupid question but did you add the dependency to the buildInputs of the main lib?
<simpson>
Usually, for cmake, we don't need to explicitly incant $(make install); that's close enough to default that it shouldn't need changing.
<Jonathan44>
When running cmake for the "library" I get `CCD is required by FCL`.
justanotheruser has quit [Ping timeout: 264 seconds]
<Jonathan44>
I build the "dependency" (CCD) and copied it's `.a` file to the local path for the "library" (FCL).
<clever>
Jonathan44: the "proper" way to do it, is to build the dependencies with nix, and add them to the buildInputs
<Jonathan44>
atemu12: no, should I?
<atemu12[m]>
Jonathan44: Yes.
<atemu12[m]>
Otherwise it's not going to see the dependency
<Jonathan44>
atemu12[m]: the "library" (FCL) is not mine so it would already be there right?
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<simpson>
Yes, Nix arranges for each declared build input to be in the environment when the builder is run.
<atemu12[m]>
Be where?
<zeta_0>
clever: that's great, i'm starting to see the benefits of using nix, if i was still using debian/apt, i'd still be very frustrated
johnny5 has quit [Ping timeout: 246 seconds]
johnnyfive has joined #nixos
<Jonathan44>
clever: I figured the right way to do it is with nix/nix-shell. But I am totally lost on how to do that--I would love to figure it out but not even sure where to start.
<simpson>
atemu12[m]: The builder environment is usually temporary. When Nix sets up the builder's environment, it sets up the inputs.
cosimone has joined #nixos
<clever>
Jonathan44: start by getting the dependencies to build with nix-build, not nix-shell
<Jonathan44>
clever: That helps.
abathur has quit [Ping timeout: 264 seconds]
<Jonathan44>
(That sounded sarcastic--it was not--let me look up nix-build in the manual)
<atemu12[m]>
simpson: I know, just wanted to make sure Jonathan44 is trying to get it into the right place the right way
<simpson>
Ah, sure.
<simpson>
Jonathan44: nix-build operates on a default.nix, which describes how your system is built. Then, you'll proceed to nix-shell, which can either work with your default.nix or optionally a shell.nix which contains selected build environment tools.
<atemu12[m]>
Is there a function in Nix/Nixpkgs that works like Python's `filter()` function?
<clever>
atemu12[m]: lib.filter
<atemu12[m]>
clever: Thanks!
<atemu12[m]>
oh wow, it's a builtin
* atemu12[m]
should RTFM
justanotheruser has joined #nixos
<evils>
odd, my systemd timer which is partOf a unit, gets stopped if that unit gets stopped, but not started if that gets started; am i missing something or is this discrepency due to it being a timer?
pistache has joined #nixos
GrimSleepless has joined #nixos
gustavderdrache has quit [Quit: Leaving.]
abathur has joined #nixos
abathur has quit [Client Quit]
logand`` has quit [Ping timeout: 256 seconds]
<pistache>
when I use boot.kernelPatches to enable config options for the kernel, ZFS modules refuse to load (because of "exec format error")
<clever>
pistache: can you pastebin your configuration.nix file?
<clever>
Jonathan44: now just reference that expresion in the buildInputs of the next thing, maybe using `rec {` or a let block, and nix-build the next thing in line
<pistache>
I've commented the offending lines, they are at the beginning of section 1.2.
<Jonathan44>
clever: Okeydok.
hio has quit [Quit: Connection closed for inactivity]
<clever>
pistache: what if you only use the kernelPackages change, and not the kernelPatches half?
arahael1 has joined #nixos
<pistache>
clever: it works in that case, but I cannot run any 32-bit application (because of the hardening)
<pistache>
I wanted to keep using the hardened packages, but tweak them to enable CONFIG_IA32_EMULATION
<GrimSleepless>
Hey guys! Can someone merget the following PR please. It has been 20 days since it was opened https://github.com/NixOS/nixpkgs/pull/81735 Thank you guys! :D
<{^_^}>
#81735 (by JeffLabonte, 2 weeks ago, open): Protonvpn ng update 2.2.0 to 2.2.2
<clever>
pistache: as an example, my nas starts from nas.nix, and you recursively follow everything listed in imports, to see the full config
leonardp has quit [Remote host closed the connection]
<pistache>
oh right, so you can just comment some of the imports and test with build-vm
<gchristensen>
any python people 'round? :)
craige has joined #nixos
<gchristensen>
its-a-trap.jpg
<clever>
pistache: yep
<adisbladis>
gchristensen: I hesitantly raise my hand ;)
<clever>
pistache: but you can also make a smaller module, that only turns on the hardening, and then make a completely seperate "machine" for build-vm, which wont try to mount disks that wont exist
<clever>
pistache: this builds a machine with zfs support, and a hardened kernel that has 32bit support
<clever>
pistache: and then you can just `imports = [ ./example.nix ];` to enable it on the main one
<gchristensen>
adisbladis: hoping for a review on my big-but-little PR to make the interesting PR easier to read :)
<Jonathan44>
Okay, I got `fcl` (library) to find `ccd` (dependency). Thanks for your help. I am now getting a strange cmake error `install EXPORT given no DESTINATION`. Looking at the CMakeList.txt it appears to come from `${CMAKE_INSTALL_LIBDIR}/cmake/${PROJECT_NAME}`. Does nix-build set those parameters differently than `cmake` from the commandline?
<clever>
pistache: though, the hostid part would probably be best to remove from the file when you confirm it works
<clever>
Jonathan44: you may need to set CMAKE_INSTALL_LIBDIR yourself, or see what sets it normally
<clever>
Jonathan44: search for CMAKE_INSTALL_LIBDIR in the files...
o1lo01ol1o has quit [Remote host closed the connection]
<energizer>
gchristensen: a bunch of functions are missing docstrings
<gchristensen>
good point, I should look up how to do those
<energizer>
looks like they're using the regular :param: format, imo the google/numpy format is much easier to read and it has built-in support in sphinx (sphinx.ext.napoleon)