<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JfBbY
jgeerds has joined #nixos
Darkmatter66_ has joined #nixos
Darkmatter66 has quit [Ping timeout: 265 seconds]
hlisp has quit [Ping timeout: 246 seconds]
hlisp has joined #nixos
noudle has quit []
LysergicDreams has quit [Ping timeout: 265 seconds]
proofofkeags has quit [Remote host closed the connection]
LysergicDreams has joined #nixos
ozychhi has quit [Ping timeout: 244 seconds]
proofofkeags has joined #nixos
mocker has quit [Quit: Connection closed for inactivity]
jkarni_ has joined #nixos
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
kvda has joined #nixos
FRidh has quit [Quit: Konversation terminated!]
jgeerds has quit [Ping timeout: 265 seconds]
ozychhi has joined #nixos
jkarni_ has quit [Ping timeout: 256 seconds]
LysergicDreams has quit [Ping timeout: 256 seconds]
LysergicDreams has joined #nixos
hlisp_ has joined #nixos
codygman has quit [Read error: Connection reset by peer]
hlisp has quit [Ping timeout: 240 seconds]
codygman has joined #nixos
<evanjs>
icetan: hey do you have any clue why I might be getting symlinks that um.. link to themselves? After building my mavenix project?
<evanjs>
The dependencies for the “recursive” symlinks do appear in my mavenix.lock
selfsymmetric-mu has joined #nixos
thc202 has quit [Ping timeout: 244 seconds]
codygman has quit [Read error: Connection reset by peer]
codygman has joined #nixos
<selfsymmetric-mu>
I have two files called `mermaid.css` and `mermaid.json`. I want to wrap up `mmdc` and pass those two files as arguments. But I need to get them into the nix store first. How do I get them there? https://hastebin.com/raw/hexelokaye
<c00w>
The simple answer is that their go.mod file isn't fully correct - right now the best way to fix this is to clone the repo, run `go mod tidy` then `go mod install -mod=vendor` (to confirm it works), and then git diff --patch and check the patch into nixpkgs and in the derivation.
<c00w>
But I'm cloning now to make sure this works :)
<bhipple>
c00w: on the subject of proxying in buildGoModule, is there a way I can inject my own Artifactory mirror without changing the hashes?
<c00w>
sure, just change the go.mod
<bhipple>
e.g. adding it in mirrors.nix or something
<c00w>
Oh wait
<c00w>
Are you trying to substitute the source?
<bhipple>
I mean across all of Nixpkgs: I'm building with a bunch of air-gapped builders that can hit Artifactory, but not the standard gopkgs endpoints; Artifactory has those mirrored, but I'd like to tell the builders to talk to Artifactory instead of golang.org et all
<c00w>
or set GOPROXY=whatever.your.mirror.is.that.supports.the.go.protocol.local
choward1 has joined #nixos
hlisp_ has quit [Remote host closed the connection]
<c00w>
Go has builtin support for this, and the go modules respects it.
m0rphism has quit [Ping timeout: 256 seconds]
choward has quit [Ping timeout: 246 seconds]
choward1 is now known as choward
<bhipple>
Right, that's what I've done, but we turn it off in NixPkgs -- so when I do it, it blows the hashes on everything. Altho maybe I need to just look at it more closely, because we should be building FODs ...
<c00w>
I think you can just do something like overrideModAttrs = ( _: {
<c00w>
GOPROXY="address.local";
<c00w>
});
<c00w>
Ah - you just want to do it for the module phase
<bhipple>
That'd make my overlay a little less invasive, because I could just add my mirror to mirrors.nix without hacking up the golang builder
<quinn>
so i'm using pulseaudio with sway, and though my DBUS_SESSION_ADDRESS is set inside of sway, it's not accessible to my pulseaudio service (/proc/$pulse_pid/environ). any thoughts on why that is/how i might go about fixing it
hlisp has joined #nixos
<c00w>
bhipple: You don't need to hack the file, overrideModAttrs lets you do it.
sarcasticadmin has quit [Ping timeout: 256 seconds]
<bhipple>
That's a good point, but I need to do it for every possible golang app that I build, so it'd be nice to do it just once (either hacking the infra, or mirrors.nix)
<{^_^}>
[nixpkgs] @maralorn opened pull request #87854 → go-neb: Init at unstable-2020-04-09 → https://git.io/JfBxT
<c00w>
Ah fair. We could probably put a flag there, but the default is the empty proxy :) (which means proxy.golang.org,direct)
proofofkeags has quit [Remote host closed the connection]
<c00w>
hexa-: ^^ lets you just sets regenGoMod = true to fix the go.mod file.
<hexa->
maralorn: ^
felixfoertsch has quit [Ping timeout: 244 seconds]
felixfoertsch23 is now known as felixfoertsch
proofofkeags has joined #nixos
<evanjs>
icetan: it looks like it’s only happening with SNAPSHOT deps? Idk. That might just be a coincidence
bhipple has quit [Ping timeout: 256 seconds]
proofofkeags has quit [Ping timeout: 256 seconds]
lightbulbjim has joined #nixos
<{^_^}>
[nixpkgs] @eadwu closed pull request #86835 → modules/filesystems/zfs: use current kernel → https://git.io/JfGcb
bhipple has joined #nixos
<quinn>
i am trying to override 'systemd.user.services.pulseaudio.serviceConfig.PassEnvironment', but when i do this i get a conflicting definitions error. any suggestions on how to fix this?
<{^_^}>
[nixpkgs] @eadwu opened pull request #87856 → nixos/kernel.nix: ensure same kernel is used → https://git.io/JfBxF
<c00w>
quinn: You can also probably fiddle the generated service file via the systemd interfaces.
hlisp has joined #nixos
<c00w>
i.e. systemd.services.pulseaudio.environment.DBUS_SESSION_ADDRESS = "the address";
gustavderdrache has quit [Quit: Leaving.]
<quinn>
c00w: dbus session address is (i believe) random, and is definitely not accessible from configuration.nix. and it's not a package so i can't override it
<quinn>
(cause packages expose the override attr which isn't present)
<c00w>
Uh - the whole point of systemd user services is that they don't read the login shell AFAIK. Is there a program you can run to get the address?
<c00w>
You could also configure your shell to write it to a file, and then read it from your pulseaudio config
<quinn>
assuming pulseaudio config supports that. i will consider doing so if i can't fix it with PassEnvironment
<quinn>
i mean it's exported in the environment, it's just a matter of fixing the serviceConfig
<c00w>
Oh wait is PassEnvironment a systemd option?
<c00w>
You can just set it in ExtraConfig on the service if that's the case.
turona has joined #nixos
selfsymmetric-mu has joined #nixos
<selfsymmetric-mu>
I realize I may have made a mistake with my config.
<selfsymmetric-mu>
I thought this was clever, since I could use multiple "channels" at once declaratively http://hastebin.com/raw/oyolayomik
<selfsymmetric-mu>
But now I realize that without the commit…I think these are going stale. When something get committed to those branches, NixOS doesn't know.
<selfsymmetric-mu>
So my unstable might be quite out of date. I'm not sure though. Anyone know?
<selfsymmetric-mu>
What should I do instead for this usecase? Is this what `niv` is for?
<cole-h>
Without the commit, it will get the current HEAD
<cole-h>
What makes you think your unstable is out of date?
<quinn>
c00w: sorry, i'm really stupid. do you mean the extraConfig for hardware.pulseaudio ? i thought that was appended to the config file
<selfsymmetric-mu>
cole-h: Because when I push to my fork on `master`, NixOS doesn't seem to pull down any new changes.
<c00w>
Nah - the extraconfig for systemd.services.pulseaudio
<c00w>
it lets you pass arbitrary systemd options.
<selfsymmetric-mu>
cole-h: Unless maybe GitHub has a lag time for updating the `master.tar.gz`?
hlisp has quit [Ping timeout: 265 seconds]
<quinn>
c00w: it's systemd.user.services.pulseaudio, and nixos-option says that systemd.services.pulseaudio.extraConfig doesn't exist
<cole-h>
selfsymmetric-mu: And you're rebuilding your NixOS config, right?
<selfsymmetric-mu>
cole-h: With a `sudo nixos-rebuild switch`.
<quinn>
c00w: well, pulseaudio is ran as a user service. but i don't see extraConfig in that list. and the problem with just setting serviceConfig is that i get conflicting definitions
andi- has quit [Remote host closed the connection]
<cole-h>
selfsymmetric-mu: I'm not familiar with packageOverrides, so I'm out of ideas. Sorry.
proofofkeags has joined #nixos
<c00w>
quinn: Ah - I think you've hit the end of my ideas.
<quinn>
selfsymmetric-mu: sorry, i'm only looking at the end of this. if you're fetchUrling your <head> and you have the sha256 of the old head, nix will just used the cached fetchUrl with the old sha256
<quinn>
without reading through that quirk may be relevant, sorry if it isn't
<quinn>
c00w: that's fine. i was really excited when you mentioned extraConfig though (:
<quinn>
c00w++
<{^_^}>
c00w's karma got increased to 2
<cole-h>
quinn: They're using fetchTarball, which doesn't require a sha (and which they're not providing)
Neo-- has quit [Ping timeout: 260 seconds]
<c00w>
selfsymmetric-mu: Wait, are you updating the hash? If it's not changed, it won't try to redownload.
<cole-h>
c00w: Read my previous message :P
<c00w>
Ah - derp
<selfsymmetric-mu>
Aha. "The fetched tarball is cached for a certain amount of time"
<energizer>
> After a number of rounds of review the discussion should settle and a general consensus should emerge. This bit is left intentionally vague and should be refined in the future. We don't have a technical commitee so controversial changes will be rejected by default.
<{^_^}>
error: syntax error, unexpected IN, expecting ')', at (string):310:163
<quinn>
,overlays
<{^_^}>
quinn: Did you mean overlay?
<{^_^}>
Overlays look like `self: super: { foo = ...; }`. Use the self argument to get dependencies, super for overriding things and library functions (including callPackage). More info: https://nixos.org/nixpkgs/manual/#sec-overlays-install
<{^_^}>
[nix-mode] @matthewbauer pushed 0 commits to refs/tags/v1.4.2: https://git.io/JfRv5
<emily>
energizer: steam-run is probably the lowest-effort thing
<emily>
if you just want to get an existing binary working and it doesn't work out of the box
hlisp has quit [Ping timeout: 272 seconds]
<emily>
there are ascendingly fancier things you can do if you want to do things the "right way", but for electron i don't know what they are beyond writing a derivation that patches the binary
cyphase has quit [Ping timeout: 272 seconds]
mir100 has joined #nixos
hlisp_ has quit [Ping timeout: 256 seconds]
proofofkeags has quit [Remote host closed the connection]
lightbulbjim has quit [Quit: Connection closed for inactivity]
drakonis_ has quit [Read error: Connection reset by peer]
drakonis_ has joined #nixos
<{^_^}>
[nixpkgs] @doronbehar opened pull request #87864 → todoman: Make it not reference glibLocale which is ~210Mb in closure size → https://git.io/JfRJ2
alexherbo2 has joined #nixos
palo1 has joined #nixos
drakonis_ has quit [Read error: Connection reset by peer]
drakonis_ has joined #nixos
palo has quit [Ping timeout: 240 seconds]
palo1 is now known as palo
philr_ has joined #nixos
philr has quit [Ping timeout: 258 seconds]
quinn has quit [Read error: Connection reset by peer]
<energizer>
https://bpa.st/W72Q builder for '/nix/store/g53pmmlg9x0bnjq06wpi9hl37cpcias4-shell.drv' failed to produce output path '/nix/store/2gyxdyqi7ppqnssnx9yrmdlxq8nd714m-shell'
<energizer>
what's wrong here?
reallymemorable has quit [Quit: reallymemorable]
agsdheidjd has joined #nixos
LysergicDreams has quit [Ping timeout: 260 seconds]
LysergicDreams has joined #nixos
LysergicDreams has quit [Ping timeout: 240 seconds]
LysergicDreams has joined #nixos
user_0x58 has quit [Quit: Leaving]
oborot has quit [Ping timeout: 265 seconds]
rauno has joined #nixos
<{^_^}>
[nixpkgs] @doronbehar opened pull request #87865 → vdirsyncer: Convert to a python module → https://git.io/JfRks
<cizra>
(sorry - asking again because I lost the answer due to limited scrollback) Hello! I'd like to fetch a hosts file in my nixos/configuration.nix from a remote source and stuff it into networking.extraHosts - is that possible? The closest I could find in Google is commented out (https://github.com/bjornfor/nixos-config/blob/master/cfg/base-small.nix#L77)
<srhb>
cizra: Yeah, you can use readFile to read the contents of a file.
<simpson>
energizer: Line endings; I ran dos2unix on it and got /nix/store/qwv020ij6il0fk8khgacc0529xkrmss7-orca-1.3.1
<srhb>
cizra: So it becomes: `networking.extraHosts = readFile (fetchurl ...);`
codygman has quit [Read error: Connection reset by peer]
<energizer>
simpson: you ran that on the downloaded .AppImage file?
codygman has joined #nixos
<cizra>
srhb: Thanks! Where do I get readFile? error: attribute 'readFile' missing
<srhb>
cizra: builtins.readFile, or lib.readFile, sorry :)
<srhb>
cizra: Same for fetchurl.
<srhb>
cizra: Check the nixpkgs manual for details on both functions.
<srhb>
(At least, I think the lib versions are documented there)
<simpson>
energizer: On the .nix file from the paste site. Did the paste site change the line endings? Specifically this is what I ran, in my local nixpkgs checkout: https://bpa.st/HNIQ
alp has joined #nixos
<srhb>
cizra: If not, it's in the Nix manual :)
lord| has quit [Ping timeout: 260 seconds]
<energizer>
simpson: that's interesting
<energizer>
it does make an empty result dir now
<energizer>
that's progres
<energizer>
s
<cizra>
srhb: Guess what, it works! Thank you so much!
<mananamenos>
hi, in xfce i have 7z installed and each time i have to extract stuff i need to go to konsole. How do make thunar integrate right click on zip file and "Extract here" option?
iceypoi has joined #nixos
hlisp has quit [Ping timeout: 265 seconds]
<{^_^}>
[nixpkgs] @nlewo merged pull request #87253 → Preserve environment variables from the parent image on dockerTools.buildImage → https://git.io/Jfcz7
<DamienCassou>
We are 2 users of NixOS on 2 recent Lenovo Thinkpad laptops. In both cases, our battery performance as decreased massively in NixOS. My own laptop is 3 weeks old and the battery lasts 2h instead of 6h.
<{^_^}>
[nix] @domenkozar pushed to master « actions: use latest OS »: https://git.io/JfRmK
<clever>
DamienCassou: you could try comparing the output of powertop as a start
o1lo01ol1o has quit [Ping timeout: 265 seconds]
kvda has joined #nixos
mallox has quit [Quit: WeeChat 2.8]
eoli3n_ has joined #nixos
eoli3n_ has quit [Remote host closed the connection]
<tom39291>
Hey, this'll sound ridiculous, but I've been a happy NixOS user for 5+ machines for 2+ years, and I have no idea how to evaluate/build/instantiate a nix expression/derivation if it's not part of nixops or nix-shell. I'm ~ok with Nix the language, but unfamiliar with conventions around importing lib/pkgs/nixpkgs and attrPaths. Can anyone suggest a resource for understanding these?
<{^_^}>
clever's karma got increased to -2147483648
<clever>
bqv: i keep my configuration.nix thin, just filesystem stuff, and imports = [ /path/to/real/config.nix ];
fendor_ has joined #nixos
<clever>
bqv: the filesystem stuff isnt worth keeping in git, because if i loose the root disk, the rest are likely also gone, and i have to redo the filesystems anyways
<bqv>
Makes sense
<bqv>
My /etc/nixos is just a symlink to the current flake in the store at this point, I have my WIP copy in ~
fendor has quit [Read error: Connection reset by peer]
<clever>
but, it also adds a .overrideScope' function, that can take an overlay, and insert it between the return-value and the `self` given to packages
<clever>
so you can make a set of inter-connected packages, that can callpackage eachother
<clever>
and then insert an overlay at a later time
__monty_1 has quit [Read error: Connection reset by peer]
hlisp_ has quit [Ping timeout: 260 seconds]
__monty__ has joined #nixos
hlisp has joined #nixos
<fendor__>
Hi! Installed nixos, booted into the usb via nomodeset since the gfx is not supported by default, installed nixos and then installed the graphics drivers via videoDrivers = ["nvidia"];
<fendor__>
but I get random freezes in KDE and the terminal, until keyboard and mouse become unresponsive altogether
<fendor__>
how can I debug that? I can boot with the initial boot config and nomodeset just fine and there are not freezes there
<{^_^}>
[nixpkgs] @adisbladis pushed 3 commits to release-20.03: https://git.io/JfRGI
vidbina has quit [Ping timeout: 256 seconds]
<fps>
fendor__: the lame answer: that's hard to debug since nvidia is closed source ;)
<fps>
fendor__: maybe take a look at the X logs
fendor__ is now known as fendor
<{^_^}>
[nixpkgs] @monacoremo opened pull request #87879 → openssl: split the (mostly empty) runtime dependencies of static builds into a separate output → https://git.io/JfRG0
<fendor>
fps, with `journalctl -xb`?
never_released has quit [Ping timeout: 256 seconds]
never_released has joined #nixos
<{^_^}>
[nixpkgs] @roberth opened pull request #87881 → haskell/generic-builder.nix: Fix C lib multiple inclusions → https://git.io/JfRG6
noudle has quit []
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<tnias>
evils: if you are using firejail, it can cause issues where you can't just "click on stuff and it works"
<evils>
tnias: i've never heard of firejail, i switched from `firefox` to `firefox-wayland` yesterday and now xdg-open ends with an error about firefox already running
marsh has quit [Ping timeout: 240 seconds]
coco has quit [Ping timeout: 258 seconds]
<{^_^}>
[nix] @edolstra pushed to flakes « Move registry-related commands from 'nix flake' to 'nix registry' »: https://git.io/JfRWW
marsh has joined #nixos
<{^_^}>
[nixpkgs] @pbogdan opened pull request #87889 → pulseaudio: enable soxr (resampler) support → https://git.io/JfRWa
<evils>
can someone clarify the criteria for using the staging branch?
<bqv>
cause if it's a very simple change you might get away with it
<bqv>
i think common sense is the discriminator, generally, is my point
<evils>
it's removing a dangling pointer, but causes 119 rebuilds, combined with another codebase that affects the same packaging, i've switched to staging and recommended they do the same
<evils>
s/codebase/PR/
<bqv>
fair enough
<evils>
what the!, bot now says rebuild 5001+ xD
<bqv>
you did rebase, right?
<bqv>
i hope you rebased...
<evils>
i switched my branch to staging
<evils>
and cherrypicked my commit on top of it
<bqv>
ok but github is still wanting to merge into master
<Raito_Bezarius>
Is it possible to install Nix on a phone with no support for user namespace in a non-root directory with --store ?
<bqv>
so now you're trying to merge staging into master
<bqv>
that sounds like a bad idea
<evils>
no, i switched the base of the PR
<bqv>
..interesting
<evils>
odd, labels are now back to normal
<evils>
but the log still shows "added 5001+" as the most recent change to labels
hmpffff_ has joined #nixos
<bqv>
yeah, i think i was slightly right, borg caught the state between you doing those two things
<bqv>
so for a little while it thought you were trying to merge staging into master
<infinisil>
evils: I'd say only to use staging for ~1000+ rebuilds
<bqv>
do a 180!
<evils>
infinisil: well, i was a bit hesitant to do a staging PR in the first place because i've never done that, and nix-review took less than an hour, but with the second PR, wouldn't using staging save a bunch of rebuilding?
gustavderdrache has joined #nixos
hmpffff has quit [Ping timeout: 272 seconds]
markus1189 has joined #nixos
orivej has quit [Ping timeout: 264 seconds]
mthst has joined #nixos
hmpffff has joined #nixos
markus1189 is now known as predator117
predator117 is now known as markus1189
<infinisil>
I'm not too involved with the staging process, but from what I understand, it's a pain to merge it into master. There can be conflicts, build problems, etc. And the more commits went to staging, the more difficult it gets to track down how to solve it
<evils>
ah, i thought it was close to master but built less often
remirol has joined #nixos
lorimer has quit [Ping timeout: 256 seconds]
remirol is now known as lorimer
<infinisil>
The reason for staging (again from what I understand) is to not cause stdenv rebuilds on master (meaning on master most things should be cached), and to bunch together changes that cause mass rebuilds (so multiple mass rebuild changes only cause one)
<evils>
well, bunch together mass rebuilds is the reason i thought it'd be a good idea to do the two asciidoctor PRs to staging, but apparently "the biggest rebuild i've done" doesn't count as a mass rebuild xD
<infinisil>
Considering there are like 20k-30k packages in nixpkgs, with a couple PR's that change 100s of packages, the overlap between them is probably rather slim, meaning no rebuilds would be saved
<evils>
well, the overlap is 100%
<evils>
they're both asciidoctor PRs
<bqv>
maybe just merge the PRs then?
<infinisil>
That's still rather small
<evils>
bqv: only one of the PRs is mine
nbrspo^ has quit []
<infinisil>
I'm talking about stdenv changes that would each cause 30k packages to be rebuilt. Doing these together means you save 30k rebuilds
<infinisil>
(per additional change)
drakonis_ has joined #nixos
drakonis1 has joined #nixos
<bqv>
evils: you can collaborate :p
hmpffff_ has joined #nixos
tsrt^ has joined #nixos
turion has joined #nixos
<evils>
bqv: but my PR is a small fix, and the other one increases the asciidoctor closure size by 40M; one's very likely to be merged, the other needs some review i think
<turion>
How do I start a nix repl in a nixpkgs tree such that I get all the packages in scope?
<infinisil>
I've had this idea before: We could have multiple levels of staging branches, and each level gets a different amount of rebuilds. E.g. 10k+ rebuilds go to the first level, 1k+ to the second one, 100+ to the third one
<infinisil>
Then you could save even more rebuilds, probably
<infinisil>
turion: nix repl .
<turion>
infinisil++
<{^_^}>
infinisil's karma got increased to 290
matthewcroughan has joined #nixos
<evils>
infinisil++
<{^_^}>
infinisil's karma got increased to 291
hmpffff has quit [Ping timeout: 272 seconds]
<evils>
infinisil: that's gonna need documentation xD
<bqv>
evils: if yours is a small fix, might as well get it in now, then by the time the next one gets in something else will probably have changed anyway so who cares
drakonis has quit [Ping timeout: 260 seconds]
hlisp has joined #nixos
<evils>
bqv: unfortunately, the size of the change does not affect the size of the rebuild
<bqv>
obviously not but there's no point postponing yours for the sake of a tiny gain
<evils>
hmm, would recursive nix make that fix not rebuild anything? (it has no practical impact)
<bqv>
if a package is different, it will get rebuilt. doesn't matter how it's written
<bqv>
evils: i think you're trying to overoptimize/overbeaurocratize this :p
orivej has joined #nixos
<turion>
Given a derivation, how do I create a nix-shell with it in path?
<evils>
there was some breaking change a while ago, maybe because of a server provider change, status.nixos.org got fixed howoldis got deprecated i suppose
m0rphism has quit [Quit: WeeChat 2.7.1]
<evils>
turion: and i think nixpkgs-unstable was considered close enough to nixos-unstable for the purposes of package search
shibboleth has joined #nixos
<supermarin>
Hi there, a somewhat noob question: yesterday I wanted to build neovim from source. Instead of installing all the deps on my main OS (macOS), I wanted to make a shell.nix file and include everything needed to build neovim in there, then use nix-shell --pure. I think using --pure is what I want, since I don't want any macOS-provided stuff involved
<supermarin>
in the build. Now the problem: I was expecting that a simple package override would work (use the same formula as nixpkgs uses, and override sources). I've also tried with stdenv.lib.overrideDerivation. My question is: Am I doing something horribly wrong, or neovim's formula is just written in a non-standard way so the standard nix practices
<iclanzan>
supermarin: Is there an error? Does the build work without --pure?
chris_ has joined #nixos
rogue_koder has joined #nixos
<supermarin>
iclanzan eventually i was able to build it without --pure, but when I launch it there's a bunch of lua errors (some lua deps are missing). but definitely not satisfied with how much i was ductaping.
alp has quit [Ping timeout: 265 seconds]
<supermarin>
the problem with --pure i didn't know how to solve is, some of the lua makefiles are using `sw_vers` which is a macOS binary
<supermarin>
and doesn't have a nix package you could include as a buildInput
<chris_>
Is it possible to mount a direct partition / drive to qemu using the `vm` module (I assume)? It seems qemu needs a `-hda` parameter passed to it, but I don't see it mentioned anywhere in the nix file.
alp has joined #nixos
<bqv>
supermarin: given neovim involves like seven different languages, i would be surprised if it wasn't nonstandard
<turion>
How do I fix this: nix-env -iA "nixpkgs.ghcWithPackages (p: [])"
<bqv>
but a question in that same vein, is there a straightforward way to override a buildRustPackage to add a certain feature?
<bqv>
or should i go to naersk
<turion>
I want to be able to install an expression, but use the installed channels
<supermarin>
iclanzan bqv just to follow up on my previous message above: is there a preferred, or any? way to build software like neovim from a local checkout?
<turion>
pbogdan: Oh wow. This is so insanely complicated.
<bqv>
turion: does the last command work if you just use -iE instead?
<bqv>
because that's a self contained expression, should be fine with a bit of massaging
<bqv>
turion: i mean yes, but you're not on a flake setup so i don't think it would work
<bqv>
manveru: nevermind, i misinterpreted your comment
<turion>
manveru: In this case it was a bit of an X-Y-question because I wanted to ask about a just-merged feature (agda support), which works similar to Haskell
<manveru>
ah, ok
<manveru>
i was mostly joking anyway
<manveru>
i modeled `ruby.withPackages` in a similar way
<turion>
I just wanted to write to the Agda mailing list that packaging Agda has just become easier because there is now good Agda support in nixpkgs. But I can't even find a good oneliner to simply install agda together with a standard library
<turion>
I mean, a oneliner so that people won't think I'm making a bad joke when I write "Installing Agda and some libraries has just become simpler!"
reallymemorable has joined #nixos
<bqv>
excellent, i'm now so many distractions deep that i've forgotten what i was doing on at least 3 levels
joko has quit [Quit: WeeChat 2.6]
mthst has quit [Quit: Leaving]
<manveru>
turion: haven't heard of it yet, but let's see...
<turion>
Many libraries don't work yet because they haven't updated to 2.6.1 yet
<manveru>
okk
<turion>
Agda is a fantastic programming language from a theoretical perspective, but unfortunately all the boring software engineering aspects like package management weren't the focus when the language was developed. Afaikt, there is no real package manager, and you have to do that by hand usually. I was hoping that the agda infrastructure in nixpkgs could step in to serve as the same role that stackage plays in the Haskell ecosystem.
<turion>
So if there is an ergonomic oneliner how people can install agda with some libraries using nix, that's an advancement.
hooo has joined #nixos
<manveru>
`nix-env -f channel:nixpkgs-unstable -i 'agda.withPackages (p: with p; [standard-library])'` something like that?
<manveru>
hm, no
<manveru>
i never use nix-env... :P
supermarin has quit [Remote host closed the connection]
<turion>
manveru: I'm at nix-env -f "<nixpkgs>" -iE "nixpkgs: (nixpkgs {}).agda.withPackages (p: with p; [ standard-library ])"
<turion>
It's clunky, but well...
<manveru>
that also won't work for most
<turion>
Well, you have to get the nix-channel first
<manveru>
yeah
<turion>
Or is there another issue?
Darkmatter66_ has joined #nixos
<manveru>
wondering if nix-shell might be nicer
lord| has quit [Read error: Connection reset by peer]
<turion>
I think I'm not going to try and talk people into creating a .nix file where they have to pin commit hashes
<turion>
manveru: Maybe, but then Agda breathes with interactive editors, and they need the executable available in the path
<manveru>
just start the editor from the shell...
<turion>
I think it's too much to ask if everyone has to start their favourite editor from within a nix-shell
<manveru>
oh well :)
lord| has joined #nixos
<turion>
I'd find that annoying as well, I don't want to know how annoying an average agda user will find it
Darkmatter66 has quit [Ping timeout: 272 seconds]
<manveru>
well, installing globally ad-hoc is the equivalent of installing npm packages globally... :|
<manveru>
might feel nice at first, but just asking for trouble later
<turion>
Yes :(
<turion>
$ nix-env --upgrade
<turion>
error: cannot auto-call a function that has an argument without a default value ('config')
<turion>
I never used that feature before. Is it supposed to work
hlisp has quit [Ping timeout: 260 seconds]
<manveru>
no
<manveru>
not with expressions like this
<manveru>
so i'd recommend to make a .nix file with a pinned nixpkgs and the agda expression and let them -i that
<manveru>
at least then they can update by bumping the hash and installing it again
<turion>
So I can break nix-env --upgrade without a helpful error message by installing something?
<manveru>
sure
Cale has joined #nixos
wolfshappen has joined #nixos
<manveru>
i just hope it gets deprecated soon :)
<manveru>
i mean, nixos is basically based on nix-env, but in a way more controlled fashion
hlisp has joined #nixos
<turion>
I hope so too. This is really unusable
wnklmnn has quit [Read error: Connection reset by peer]
fendor has quit [Ping timeout: 258 seconds]
<turion>
Thanks for guiding me out of that path... I still feel like I have no recommendation that I can give to non-Nix-diehards
<turion>
That's great and sounds simple. Now I want to go and also install a library, and suddenly I'm in a mess
<manveru>
yep
<manveru>
until you add `agda-with-stdlib` to nixpkgs or something...
<turion>
And then that again for every other library combination in Agda...
wolfshappen has quit [Ping timeout: 265 seconds]
wolfshappen has joined #nixos
<turion>
Agda libraries were hard before because it required people to keep a text file with paths of downloaded repos up to date. Now we're doing a bit better in nixpkgs by automating the downloaded repo part.
<turion>
Maybe tell them to put {pkgs}: { environment.systemPackages = [ pkgs.agda.withPackages (p: with p; [standard-library]) ] } in their ~/.config/nix/config.nix?
<cole-h>
*Was* a user :D But I still have it installed -- bbigras Is there any way you could wait until tomorrow? I have an essay I still need to finish, due tonight... 😬
<bbigras>
cole-h: no hurries. I have been waiting for weeks. Thank you very much.
<cole-h>
Hehe. Added to my TODO list; feel free to ping if you don't hear anything from me
<turion>
I guess I haven't given https://github.com/agda/agda-pkg enough credit in the whole Agda package management rant. So maybe nixpkgs should first interoperate with that before announcing the solution to all problems
gentauro has quit [Read error: Connection reset by peer]
<lunik1>
afaict megasync doesn't build on 20.03, but this has been fixed on unstable. what would be the process to request a backport, cherry-pick the relevant commits and open a PR?
<teto>
any advice on prioritizing nixos' pandas in a pipenv environment (pipenv's pandas complains about the missing libc)
<kaychaks_riot>
I come back again to ask for any solution to launch multiple firefox via rofi in sway without getting the "firefox is already running" error
msirabella has joined #nixos
<kaychaks_riot>
I'm running firefox-wayland package with all required env vars
justanotheruser has quit [Ping timeout: 244 seconds]
<colemickens>
what does "multiple firefox" mean?
<colemickens>
also, you don't want to use firefox-wayland if that's really still a thing, the regular firefox package runs on wayland
<evils>
colemickens: because `firefox` runs via xwayland and mouse movement events don't get detected outside of it; so i get tooltips on top of other wayland apps if i switch workspace
<colemickens>
man, this is also so jacked up
<colemickens>
if you use MOZ_USE_WAYLAND, stable firefox does not use xwayland
<colemickens>
this sets GDK_WAYLAND in a wrapper, which it shouldn't. includes libglvnd only if gdkWayland is true, even though it should always be included.
<colemickens>
I'll have to take a look at this later.
<evils>
colemickens: thanks
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
magnetophon has quit [Read error: Connection reset by peer]
<colemickens>
to reiterate, `nixos.firefox` + MOZ_USE_WAYLAND = success on wayland (including webrender)
magnetophon has joined #nixos
<gordon>
i’m building a derivation for a qt app, it builds, it runs in the nix-shell, but outside, in my standard environment, it fails with `qt.qpa.plugin: Could not find the Qt platform plugin "xcb" in ""`
<gordon>
but my QT_PLUGIN_PATH is the same in both environments :|
<gordon>
(i’m using wayland, that could be a useful info)
amf has quit [Quit: WeeChat 2.7.1]
<lunik1>
colemickens: so that should be used instead of nixos.firefox-wayland?
<qyliss>
isn't that all firefox-wayland is?
<colemickens>
lunik1: I can only say, without testing more which I'll do in a few hours, is that `firefox` hasn't suffered from the bug @evils is mentioning for some time.
<colemickens>
qyliss: no
morgrimm has joined #nixos
<colemickens>
Setting GDK_BACKEND is a Bad Thing To Do TM.
johnjay has quit [Ping timeout: 240 seconds]
<qyliss>
then should firefox-wayland be changed to set MOZ_USE_WAYLAND?
<colemickens>
I think. And we should look at libglvnd again.
dingenskirchen has quit [Read error: Connection reset by peer]
dingenskirchen has joined #nixos
sarcasticadmin has joined #nixos
<kaychaks_riot>
I just used nixos.firefox and MOZ_USE_WAYLAND=1 but still getting the same error in sway
<colemickens>
of course, s/USE/ENABLE/g because I always get this wrong.
<colemickens>
I set MOZ_ENABLE_WAYLAND for my entire session, that might have something to do with it.
<colemickens>
oh interesting.
<colemickens>
evils: maybe we should do that in the firefox-wayland wrapper then too?
* colemickens
isn't sure why he doesn't get that
<colemickens>
Firefox 74 ships new MOZ_DBUS_REMOTE env variable and it forces Firefox on X11 to use D-Bus remote protocol as well as Wayland version does. So when X11 Firefox is launched after Wayland one with the same profile, already running Wayland instance is reused to open the link instead of the “Close Firefox” greetings.
<colemickens>
That sounds like something else is still wrong then.
<colemickens>
You can't mix and match firefox sessions.
<evils>
well, i did that via wrapper.nix, i don't think i actually had the running instance use either of those variables... (unless dmenu uses the .desktop file)
dermetfan has joined #nixos
johnjay has joined #nixos
<evils>
running this twice, fails the second time (when the first instance is kept running) `env MOZ_DBUS_REMOTE=1 MOZ_ENABLE_WAYLAND=1 ./result/bin/firefox `
<{^_^}>
[nixpkgs] @zimbatm pushed to master « Revert "Get rid of go 1.13 (#86036)" »: https://git.io/JfRai
eoli3n___ has quit [Quit: eoli3n___]
bhipple has joined #nixos
LysergicDreams has quit [Ping timeout: 258 seconds]
LysergicDreams has joined #nixos
user_0x58 has joined #nixos
T0pH4t has joined #nixos
<Baughn>
gchristensen: I think you said you were using Sway... mind if I pick your brain for a second? Firefox is lagging out, and I wonder if that's a *known* problem. It does seem to be sway-specific, but I can't find references.
shafox has quit [Remote host closed the connection]
<T0pH4t>
is it possible to modify the packages passed to a module in a container? For example I might have containers.foo = { config = {config, pkgs, ...}: services.foo = {...} }. foo is a standard custom module, is it possible to modify the pkgs it uses to include some of my own?
matthewcroughan has quit [Quit: Connection closed]
selfsymmetric-mu has quit [Ping timeout: 260 seconds]
<Baughn>
T0pH4t: The preferred way to do that is to pass a package as a config option.
<T0pH4t>
Baughn: how would i do that?
fendor_ has joined #nixos
<cole-h>
Baughn: How do you mean "lagging out"?
<Baughn>
T0pH4t: Depends on whether you're writing the config code or not.
<Baughn>
cole-h: Every once in a while it stops responding to input for a couple of seconds.
<T0pH4t>
i'm wring the config code, but i didn't think I had access to the params to the module itself
selfsymmetric-mu has joined #nixos
<cole-h>
I don't have that issue, personally -- you're using `MOZ_ENABLE_WAYLAND=1`, yes?
<Baughn>
So you're not writing the module.
<Baughn>
I've tried it with and without.
Soo_Slow has quit [Quit: Soo_Slow]
<Baughn>
Not sure how to verify if Firefox is using it or not.
<T0pH4t>
i'm writing the module, but its part of a custom packages repo (eg we have a company repo with nixpkgs and modules)
<T0pH4t>
the idea is a developer can use a developer version of a service, rather than the prod version of a service
<T0pH4t>
so they can swap out the package used
<cole-h>
Baughn: If it's set in your environment, it's probably using it. If not, it probably isn't and you're using the X11 version.
<Baughn>
T0pH4t: That should work. Note the .package option.
<T0pH4t>
where the package is changable, but has a default value. The default value in my case was going to be from our companies private packages
<Baughn>
cole-h: I tried with and without, as I said. Doesn't seem to make a difference.
endformationage has joined #nixos
<T0pH4t>
Baughn: I see the package option and i was wondering about the default value. Eg i'd like to have pkgs.foo where foo is from our privage pkg repo
morgrimm has joined #nixos
<T0pH4t>
thats why i was asking how i could modify the pkgs passed into the module when configuring a container
<cole-h>
Baughn: What does your `about:support` say? `Window Protocol: wayland`?
<cole-h>
(It's somewhat far down the page, so you're better served Ctrl+F'ing it)
<Baughn>
I'll check next time I'm in Linux. Had to open windows to get some work done.
<Baughn>
T0pH4t: That should be a config option like any other... can I see how you're trying to define the container?
morgrimm has quit [Ping timeout: 246 seconds]
<T0pH4t>
yup 1 sec
codygman has quit [Read error: Connection reset by peer]
<hyper_ch>
43167 store paths deleted, 83348.54 MiB freed
<{^_^}>
[nixpkgs] @primeos pushed 3 commits to release-20.03: https://git.io/JfRV9
<T0pH4t>
I could explicitly define package inthat config, but i was hoping to default the value in the foo module to the default prod version
<Baughn>
I mean, yeah, both of those would work.
LysergicDreams has quit [Quit: leaving]
<{^_^}>
[nixpkgs] @peti pushed 35 commits to haskell-updates: https://git.io/JfRVd
<Baughn>
T0pH4t: So you put a mkDefault in the module, then reassign it in your dev cluster or whatever.
iclanzan has quit [Quit: leaving]
<T0pH4t>
a mkDefault in the container config or the module itself?
<Baughn>
Wherever you prefer, really.
<hyper_ch>
I still wonder how a few generations generate so much disk usage
<Baughn>
hyper_ch: Do you have auto-optimise-store enabled?
<T0pH4t>
Baughn can you maybe point me at an example of this?
<Baughn>
T0pH4t: The elasticsearch module *is* an example.
<hyper_ch>
Baughn: not per se but I do when when I clean 2272.57 MiB freed by hard-linking 162894 files
<Baughn>
hyper_ch: Perhaps consider enabling it. :)
<T0pH4t>
i didn't see a mkDefault in there
<Baughn>
T0pH4t: There's a default. It amounts to the same thing.
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JfRVp
<hyper_ch>
Baughn: hmmm.....
<hyper_ch>
Baughn: how to enable it?
<T0pH4t>
Baughn: If you are refering to the default option value then yes i see that, but I was asking about how i get my company pkgs pass thru to the module at configure time
<Baughn>
T0pH4t: Well, edit the default?
<Baughn>
Hmm.
<T0pH4t>
mm I mean the function arguments at top of module
<Baughn>
No, don't do that. Use the options system.
<Baughn>
Do you know how the options system works, in general?
<T0pH4t>
ig if iwant { config, lib, pkgs, company-pkgs, ...}
zupo has joined #nixos
<T0pH4t>
eg*
zupo has quit [Client Quit]
<T0pH4t>
if thats not doable, then thats fine
<Baughn>
Those packages are from the scope in which callPackage was invoked.
<Baughn>
You can't add something to it without first adding it to said scope.
<T0pH4t>
make sense, and I'm guessing I don't have access to that scope since this is invoked by nix
<Baughn>
But you *can* e.g. stick an "import /etc/company-pkgs" in there.
<T0pH4t>
right
<T0pH4t>
ok thx
<Baughn>
Or you can override .package elsewhere, e.g. from an import in the container.
<emily>
badly named, it has nothing to do with import :(
<symphorien>
oh nice
<evils>
colemickens: no offense taken, i was under the impression i tried setting the env var (by way of `env ... firefox`) (i'm not very familiar with setting stuff like this) which is why i'm trying to make sure by messing with the wrapper
<emily>
wonder what a better term would be. "build-in-(instantiation|evaluation)"?
lucksus has quit [Remote host closed the connection]
<cole-h>
`--no-allow-import-from-derivation` should work, I think.
<cole-h>
Maybe?
reallymemorable has quit [Quit: reallymemorable]
<emily>
huh, that's not documented in the manual but seems to exist, yes
<emily>
I wonder if it differs from --optino allow-import-from-derivation false at all
<cole-h>
Which is the only time I've seen it used :P
<srhb>
emily: Though I think v2 is the only one with proper completions for non --option style
<Fare>
Hi. I have an "interesting" problem with #!-interpreted #!-interpreters -- this works on Linux, but not on macOS. Using env as a shim works on macOS but fails on Linux in Docker because of a 128-char #!-line limitation. Using /usr/bin/env generally works, but apparently not if the binary is used during the nix build. Sigh.
<srhb>
Fare: Did you hear about that time Linux broke all of NixOS? :-)
<Fare>
Is it possible to for /usr/bin/env to exist while building? After all, it's a standard path defined in POSIX that even NixOS provides.
<emily>
srhb: v2 as in nix(1) or the nix- commands in 2.something?
<Fare>
srhb: which time?
<srhb>
emily: As nix commands without the dash
<evils>
colemickens: ok, i can't reproduce my issue from a `nix-shell --pure`
<srhb>
Fare: Ah nevermind, just the time where the kernel started not passing long shebangs. Anyway, are you saying this is specifically a problem in docker?
alp has joined #nixos
<bdju>
So, p7zip was marked insecure... is there a replacement? mainly I used it because it was free software and could unpack a rar file
<lunik1>
bsdtar (from libarchive) can unpack rars
<bdju>
ah, neat
dtz has joined #nixos
<dtz>
oh?? neat!
lucksus has joined #nixos
<bdju>
ugh home-manager keeps spitting out errors about p7zip even after taking it out of my home.nix. what a pain. how do I remove this thing?
<lunik1>
it can also pack and unpack .7z
<Fare>
how do I tweak my script to behave differently on BSD than Linux, already ?
<rycee>
bdju: `home-manager uninstall` should work if you use the standalone setup.
<Fare>
srhb, looks like the Docker kernel indeed is stripping the #! line to 128 characters or some such.
<Fare>
or are you saying the bad kernel comes from my own nix-produced docker image.
<ajs124>
bdju: you might have something that depends (as of now) on p7zip as a build or runtime dependency
MmeQuignon has joined #nixos
<ajs124>
like unetbooting (last time I checked, at least)
<bdju>
ajs124: oh, you're probably right. I think multibootusb needs it. I'll try that
<lunik1>
Fare: in bash you can check $OSTYPE?
<bdju>
rycee: is that to remove p7zip or home-manager itself? I don't want to remove home-manager
<bdju>
ajs124: there a way to locate what may be depending on it? commenting out multibootusb did not do the trick
<bqv>
lunik1: but not gnu tar, right?
<Fare>
yes, the kernel inside docker was made by NixOS
<lunik1>
not gnu tar afaik
<rycee>
bdju: Ah, right. That command uninstalls Home Manager :-)
<bqv>
Interesting
<ajs124>
bdju: you can use `nix why-depends` or some tool like that
<lunik1>
and I don't know if libarchive is used as the default tar in any other BSD than Free, so maybe not those either
<bdju>
ajs124: how do I find out how to use this? `nix why-depends p7zip` doesn't work, doing pkgs.p7zip didn't work either. Says more arguments are required, but can't think of what else I'd add.
<ajs124>
you need to pass it two... drvs? I think. someone here probably knows how to use it, but I'll go see if I can find some documentation.
<bdju>
as in long file paths to the nix store?
<bdju>
also, I don't know what two to pass it, I just want p7zip removed, I *think* mulitbootusb needs it, but commenting out both doesn't get me anywhere, so then I don't know how I would then go on to see if anything else needed it if I need two files listed
<ajs124>
ah, so wait. you have p7zip in you system or home-manager or where?
gordon has quit [Ping timeout: 246 seconds]
hlolli has joined #nixos
<bdju>
ajs124: p7zip is in my home.nix for home-manager, I can't update because of it, but removing it from my file also isn't working.
<bdju>
I do everything declaratively so it should just look at the list, see it's gone, and move on... not so simple, I guess.
<Fare>
Aha, Docker inherits the kernel from outside, so on my machine I have the NixOS kernel and everything's fine, but on Gitlab, they use one of these kernels with tighter #!-line limitations and I can't #!/...coreutils...env /full/path/to/nix/stuff
<ajs124>
bdju: then you can probably do nix-store -q --tree $HOME/.nix-profile
<lucksus>
Hey folks, I'm trying to cross-compile my Rust project to Android using a nix-shell (on Ubuntu host) with `crossSystem`, and I'm running into the exact same problem described and patched here: https://github.com/NixOS/nixpkgs/issues/56156. I tried to apply the same patch using an overly but no matter what I do, it seems the overlay doesn't apply to
<lucksus>
the LLVM pulled in by crossSystem. And I can't find comprehensive documentation for crossSystem, like how to set the LLVM version used etc. Any help or link to docs would be very appreciated!
<{^_^}>
#56156 (by nomeata, 1 year ago, closed): Cannot build LLVM-6 for Android
ATuin has joined #nixos
erasmas has quit [Read error: Connection reset by peer]
<bdju>
ajs124: aha! dtrx was the other thing. thanks so much!
lucksus has quit [Remote host closed the connection]
wnklmnn has quit [Client Quit]
<bqv>
Weird question, do you guys think it'd be possible to make a docker quine? :D I.e. A docker image that when spun up, starts another docker daemon inside and spins itself up as an image on it
<{^_^}>
[nixpkgs] @vbgl pushed to master « ocamlPackages.ipaddr: remove at 2.6.1 »: https://git.io/JfR6Y
o1lo01ol1o has joined #nixos
zaeph1 has joined #nixos
zaeph has quit [Disconnected by services]
zaeph1 is now known as zaeph
<balsoft>
bqv: If it crashes for reasons other than OOM, I bet it will crash on some power of 2 :)
<{^_^}>
[nixpkgs] @peti pushed 5 commits to haskell-updates: https://git.io/JfR64
<balsoft>
I have an intuition that after 256 nested containers, some word will overflow and cause a crash
o1lo01ol1o has quit [Ping timeout: 246 seconds]
<simpson>
balsoft: Due to *interpretative overhead*, I would be surprised if you can get 256 layers deep. Empirically, we know that each time we nest an interpreter, we get a multiplicative factor of a slowdown, and that factor is roughly constant, so that the overall slowdown is exponential.
<lassulus>
how do I get python2Packages.matplotlib and python2Packages.wxPython into the same env? if I run: nix-shell -p '(python2.withPackages (pp: [ pp.wxPython pp.matplotlib ]))' I get: collision between `/nix/store/riixmkhlx39af12libb31154p19h4f81-python2.7-backports.ssl_match_hostname-3.7.0.1/lib/python2.7/site-packages/backports/__init__.py' and
<simpson>
balsoft: And practically, the slowdown is something like 2x (V8) or 3x (Dolphin), so you're going to run out of compute time far before you run out of memory.
<simpson>
balsoft: Practical experiments suggest that a full VM, even with hardware passthrough, can only nest maybe a dozen times. Here's some folks trying for a four-layer-deep nesting: https://www.pcsteps.com/508-nested-virtualization/
<simpson>
And, to be on-topic, this is why IFD is so practically useful compared to recursive Nix: The latter is *way* harder to get right, because of the arbitrary nesting of calling Nix while calling Nix.
<balsoft>
simpson: I think recursive nix just calls the host's nix daemon though?
<{^_^}>
[nixpkgs] @FRidh pushed commit from @bhipple to master « doc: Rewrite python onboarding tutorials in manual »: https://git.io/JfRPF
aswanson has quit [Quit: WeeChat 2.7.1]
proofofkeags has quit [Remote host closed the connection]
aswanson has joined #nixos
fendor has joined #nixos
chloekek has quit [Ping timeout: 264 seconds]
lopsided98 has quit [Remote host closed the connection]
lopsided98 has joined #nixos
<aswanson>
I just switched my nix-channel over to 20.03, and now when I try to `nixos-rebuild` I'm getting errors about the perl interpreter missing a `perl530` attribute and the python default.nix not knowing what `callPackage` is
<aswanson>
what is going on?
Pot2Dash has quit [Quit: Leaving]
fabianhjr has joined #nixos
<cole-h>
Could you be more descriptive? What are the exact error messages?
<aswanson>
the exact message is "error: attribute 'perl530' missing, at /nix/store/b41b7jkbdr3jx9nfdvnhsg81jyz39j7j-nixos-20.03.1866.a7c70f2e10b/nixos/pkgs/development/interpreters/perl/default.nix:215:12"
<bqv>
energizer: you want makeWrapper, not wrapProgram, since ${appimage-run}/bin is not writeable
<T0pH4t>
how can i append another nix path entry in the configuration.nix without blowing away the defaults?
<energizer>
bqv: is my second paste ok?
<bqv>
that depends strongly on your definition of "ok"
<bqv>
but i don't see why it wouldn't work, yeah
iceypoi has quit [Ping timeout: 260 seconds]
<energizer>
bqv: is there any reason i should be using makeWrapper instead of writeShellScriptBin?
<bqv>
imo you should be using neither but ymmv
gustavderdrache has joined #nixos
<bqv>
to be honest, your entire derivation could be replaced by writeShellScriptBin
iceypoi has joined #nixos
<aswanson>
thanks for the tips cole-h: I'll have to keep poking at it to see what I can come up with. Out of curiousity, what in those logs tipped you off to look for help2man?
<cole-h>
1while evaluating the attribute 'buildInputs' of the derivation 'help2man-1.47.12' at /nix/store/b41b7jkbdr3jx9nfdvnhsg81jyz39j7j-nixos-20.03.1866.a7c70f2e10b/nixos/pkgs/development/tools/misc/help2man/default.nix:4:3:
<cole-h>
On a brief skim, I think that was the only one that pointed to an actual derivation
<cole-h>
and not the stdenv or other helpers
<cole-h>
Let me know if it works
<energizer>
bqv: is one of these more idiomatic than the other? https://bpa.st/6GSQ
hlolli has joined #nixos
icey_ has joined #nixos
Qfwfq has left #nixos ["Using Circe, the loveliest of all IRC clients"]
<cole-h>
energizer: btw you can put that `writeShellScriptBin` directly into your systemPackages (or user packages) I believe. No need to make a derivation that consists of... a derivation.
<energizer>
T0pH4t: i'd guess the defaults will stay added. are you seeing something different?
<bqv>
both are problematic (the former could be replaced by writeShellScriptBin and the latter is layers of superfluous stuff)
<aswanson>
cole-h: I've tried disabling my install of inetutils since a grep of nixpkgs showed that it had a depency on help2man, and the wireguard kernel module I had enabled, but no dice so far
<T0pH4t>
$energizer: yes, i blow away the nixos paths
<bqv>
take the second one and have that `run-orca` be your derivation
<bqv>
that's it
<cole-h>
aswanson: Do you have any overlays? I'd disable those to try and narrow it down.
iceypoi has quit [Ping timeout: 260 seconds]
<bqv>
> pkgs.rocksdb.version
<{^_^}>
"6.4.6"
justanotheruser has quit [Ping timeout: 260 seconds]
__monty__ has quit [Quit: leaving]
aswanson has quit [Quit: WeeChat 2.7.1]
<energizer>
bqv: ok cool that works. it gives a few noisy messages at the beginning, but they dont seem to matter https://bpa.st/5IUA
<energizer>
i dont even know which layer those messages are coming from but whatever
aswanson has joined #nixos
o1lo01ol1o has joined #nixos
teto has quit [Quit: WeeChat 2.8]
<aswanson>
cole-h: I don't think I have any overlays on this system, but I also have half a dozen abandoned projects or tweaks lying around so who knows. Thanks for the assist, but I'll revisit this another day
<cole-h>
aswanson: Sorry I couldn't be of more help.
proofofkeags has quit [Remote host closed the connection]
proofofkeags has joined #nixos
dermetfan has quit [Ping timeout: 260 seconds]
zeenk has quit [Quit: Konversation terminated!]
proofofkeags has quit [Ping timeout: 256 seconds]
gustavderdrache has quit [Quit: Leaving.]
fresheyeball has joined #nixos
Darkmatter66 has joined #nixos
Darkmatter66_ has quit [Ping timeout: 256 seconds]
daGrevis has quit [Remote host closed the connection]
shibboleth has quit [Quit: shibboleth]
daGrevis has joined #nixos
civodul has quit [Quit: ERC (IRC client for Emacs 26.3)]
oborot has joined #nixos
<palo>
I'm updating to 20.03 (from 19.09) and I get a messagt that my cargoSha is not correct (for some of my packages) but when I try to build (nix-build) I always get the same sha, also when I set it to 1111.. and run nixos-rebuild switch I get the same sha as 'got: sha256'
hlolli has quit [Ping timeout: 272 seconds]
<palo>
So I can't really update :/
<clever>
palo: it should be printing 2 hashes, is the 2nd one the one from your expr?
<clever>
the wanted one
asymptotically has quit [Quit: Leaving]
<palo>
when I set it to 11111 I get a set vs get listing, when I set the get (which was the original one) I don't get a set vs get listing, I gett a message : https://paste.ingolf-wagner.de/aci2mjEy/+inline?