<gchristensen>
colemickens: check PMs on the Packet slack?
<energizer`>
Is running `lorri shell` supposed to give the same environment as running `nix-shell`?
<AmandaC>
colemickens: wait Pluto has morethen Charon?
<energizer`>
I'm seeing commands available on PATH in `nix-shell` that aren't available in `lorri shell`. Not sure if I'm doing something wrong or misunderstanding lorri.
<cole-h>
What is the best way to have both `pinentry-gnome3` and `pinentry-curses` binaries in my path? If I put both `pinentry.gnome3` and `pinentry.curses` in my packages list, they conflict.
<{^_^}>
[nixpkgs] @orivej-nixos pushed commit from @tollb to staging « gdb: Fix abort w/NIX_DEBUG_INFO_DIRS patch (#80929) »: https://git.io/JvEPE
<AmandaC>
samueldr: was it you I was talking briefly about linux kernels on i915 with a couple days go? If so: 5.5.4 might be a viable recommendation as wellfor 20.03 if it's packaged for there. I've not had any hangs, or the other soft-crash issue since switching to it
<samueldr>
last time someone said something similar, ~15 minutes later their system locked-up :)
drakonis_ has quit [Read error: Connection reset by peer]
<samueldr>
this must make the maintainers lose their hair, no clear way to reproduce the (possibly multiple) issue!
<AmandaC>
haha
<AmandaC>
Well I've been running it since that talk, multiple suspend/resume cycles with no issue
<AmandaC>
I've hung awesomewm a couple of times, but that was my own stupidity, not i915
<AmandaC>
( and switching to a tty worked )
<samueldr>
once I'm done with my current work for the end of the month, I'll be dogfooding 20.03, including the default 5.5 kernel on i915 hardware
<samueldr>
then I'll be able to better recommend :)
* AmandaC
nods
<AmandaC>
Just depositing my $0.02
<samueldr>
sure, and thanks :)
jb55 has joined #nixos
<cole-h>
If I reference a package in a script with `pkgs.writeShellScriptBin`, will that package be added to the store?
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JvEPM
<cole-h>
My use case is basically a custom `pinentry` script that uses the gnome3 prompt if an env var is set, otherwise uses curses
<cole-h>
Because when I would ssh in, it would fail to authenticate because -- shock -- there was no graphical environment available to spawn the gnome3 pinentry on
stree has quit [Quit: Caught exception]
stree has joined #nixos
chagra has joined #nixos
drakonis has joined #nixos
tjg1 has quit [Quit: DON'T KEEP CALM AND FUCK THIS. FUCK THAT. FUCK YOU. FUCK OFF.]
drakonis_ has quit [Read error: Connection reset by peer]
<jonathan58>
I could use some help. I am trying to build an C++ application with CMake.
<jonathan58>
I have built it on my old Arch rig but I am struggling with this NixOS one.
chagra has joined #nixos
<jonathan58>
The application is in nixpkgs but I am trying to build the source (from my live code).
<cole-h>
What's the problem, and what have you tried?
drakonis has quit [Ping timeout: 240 seconds]
<jonathan58>
cole-h: I tryied to use the .nix from nixpkgs as shell.nix but as you probably know that did not work.
erictapen has quit [Ping timeout: 258 seconds]
<jonathan58>
cole-h: I tryied just pulling the list of dependencies from the .nix on nixpkgs to use in the dummy shell shown in the manual.
vladk has joined #nixos
<jonathan58>
`make` throughs this error make: /usr/bin/cmake: Command not foundmake: *** [Makefile:8514: cmake_check_build_system] Error 127
<jonathan58>
Which does not help...
<qyliss>
That /usr/bin/cmake is suspicous
<qyliss>
*suspcicious
<qyliss>
/usr/bin/cmake won't be available in the Nix sandbox.
<cole-h>
Probably a default path, and uses whatever cmake it finds in PATH before checking that
<qyliss>
You'll need to figure out where that's coming from, and see if you can override it.
<qyliss>
But what cole-h says may also be true, so you could try just adding cmake to nativeBuildInputs and see if that helps.
<jonathan58>
In is from a nix-shell
drakonis has joined #nixos
drakonis1 has quit [Read error: Connection reset by peer]
<jonathan58>
nativeBuildInputs? what is different between that and buildInputs?
<qyliss>
jonathan58: check the manual :)
<qyliss>
(the Nixpkgs manual)
<jonathan58>
Oh, right.
nuncanada has quit [Quit: Leaving]
<jonathan58>
The .nix on nixpkgs as libX11 as a buildInput (I see now) but when I but it as a build input in the shell script it says `undefined variable 'libX11'...`
<qyliss>
Does it have a "with" anywhere?
<qyliss>
It might be brining it in from another scope.
<cole-h>
Make sure you have nixpkgs in scope.
<jonathan58>
The libX11 comes from on of the arguments at the top.
<jonathan58>
cole-h: it is (using a with import)
<qyliss>
jonathan58: okay, check where that Nix function is called in all-packages.nix (or similar file in pkgs/top-level), and see if it's getting libX11 from somewhere special.
<jonathan58>
Interesting, okay.
<cole-h>
I think the problem might be that they are calling the file without supplying libX11 as an argument e.g. copy-pasting the file to a shell.nix outside of nixpkgs and then running `nix-shell` with it
<qyliss>
cole-h: right, but finding out where the libX11 comes from will reveal where they should get it from inside shell.nix
<qyliss>
jonathan58: for what it's worth, I think I know the answer here, but I'm trying to help you discover it for yourself
stree has quit [Quit: Caught exception]
<qyliss>
If you're in a hurry and want me to just tell you, let me know :)
stree has joined #nixos
<jonathan58>
Okay (:
<jonathan58>
I have been trying to do this off and on for a couple months, what is another few minutes...
kahiru has quit [Ping timeout: 260 seconds]
justanotheruser has joined #nixos
<cole-h>
I applaud that attitude.
drakonis_ has joined #nixos
drakonis_ has quit [Read error: Connection reset by peer]
drakonis_ has joined #nixos
drakonis has quit [Ping timeout: 240 seconds]
kahiru has joined #nixos
<jonathan58>
cole-h: It does not look like it has any special argument?!
<jonathan58>
in `all-packages.nix`
<DigitalKiwi>
took me about 7 months to package ocrmypdf
<cole-h>
Try looking for `libX11` in that file and see what's what... ;)
mounty has quit [Remote host closed the connection]
mounty has joined #nixos
<DigitalKiwi>
i've only ever needed it for one pdf and i'd already converted it on macos before it was in nixpkgs :|
<jonathan58>
cole-h: qyliss: Finally! It is compiling. Turns out the local clone of the git repo ao was corrupted and needed recloned...
<qyliss>
nice!
shafox has quit [Remote host closed the connection]
rogue_koder_ has joined #nixos
rogue_koder has quit [Remote host closed the connection]
<fusion809>
Is it possible to disable tests run when I built math packages as part of my system config? Some of these test suites chew up so much CPU and RAM that my PC frequently freezes and they take quite a while, so it would be nice to disable them, if possible.
rogue_koder_ has quit [Remote host closed the connection]
rogue_koder_ has joined #nixos
waleee-cl has quit [Quit: Connection closed for inactivity]
<{^_^}>
cole-h: I'll think about it..
<{^_^}>
[nixpkgs] @lopsided98 opened pull request #81004 → libyamlcpp: don't use multiple outputs → https://git.io/JvE1R
Wizek has quit [Quit: Connection closed for inactivity]
<NoctisLabs>
I switched to Wayland (sway) and now when I run a flatpak I get the error: "The name org.freedesktop.Accounts was not provided by any .service files", what am I missing?
<fusion809>
Noctislabs: I get that same error under KDE (with Xorg, or so I think), and I don't know of any fix.
<fusion809>
It doesn't bother me too much as what apps I ordinarily run with Flatpak I can run using a chroot.
<fusion809>
Maybe you can too?
rail has quit [Read error: Connection reset by peer]
<worldofpeace>
fusion809: I'm assuming plasma5, sddm, and flatpak enabled?
<fusion809>
Yep, all those goodies.
<NoctisLabs>
unstable here
jonathan58 has quit [Remote host closed the connection]
<worldofpeace>
that particular issue had a strange minimal setup, your setup sounds like something I can test in a vm. Will look into it.
<NoctisLabs>
I'm using sway, no dm, flatpak enabled
<colemickens>
TIL of xdg.portal.enable = true; && xdg.portal.extraPortals = [];
<fusion809>
Thanks for showing an interest, worldofpeace.
<worldofpeace>
fusion809: hmm, I don't think plasma5 enables accountsservice
<fusion809>
Guessing sway doesn't either.
<worldofpeace>
fusion809: could you try services.accounts-daemon.enable = true;
chagra has quit [Ping timeout: 258 seconds]
<worldofpeace>
that would obviously fix it because it will bring the interface up. but I was doubtful at first, because that discourse post said it didn't fix it for them
<fusion809>
Right in the middle of a rebuild that is *very* tedious (as it involves several math packages with large test suites, which is why I asked my earlier question). When this rebuild is finished I'll do that and rebuild.
<worldofpeace>
NoctisLabs: I'm assuming this isn't a critical error
<NoctisLabs>
services.accounts-daemon.enable = true; seems to have worked
<worldofpeace>
NoctisLabs: applications start without it enabled right?
<NoctisLabs>
No, they don't
vidbina has quit [Ping timeout: 260 seconds]
leungbk has quit [Ping timeout: 240 seconds]
rail has quit [Quit: gone...]
<worldofpeace>
NoctisLabs: hmm, that doesn't seem right to me.
jybs has joined #nixos
<worldofpeace>
can you run an app when it's in the enviroment it doens't work with G_MESSAGES_DEBUG=all ?
spacefrogg has quit [Quit: Gone.]
aw has quit [Quit: Quitting.]
aw has joined #nixos
felixfoertsch has joined #nixos
<jybs>
Hi all. I'm trying to get clang working (from Rust) with the libc stuff available
felixfoertsch23 has quit [Ping timeout: 240 seconds]
spacefrogg has joined #nixos
<jybs>
I'm pulling in llvmPackages.libcxxStdenv, should I need to do anything else?
<NoctisLabs>
So with accounts-daemon enabled Discord is now launching but it's giving the error "[WebContents] did-fail-load -202 ERR_CERT_AUTHORITY_INVALID retry in 20000 ms"
<hpfr[m]>
anyone using riot-desktop? I try to set dark theme and it immediately switches back to light themej
notgne3 has joined #nixos
tylerjl has joined #nixos
puck has joined #nixos
choward9 has joined #nixos
minicom1 has joined #nixos
mlen has quit [Ping timeout: 260 seconds]
ravndal has quit [Ping timeout: 260 seconds]
infinee has quit [Ping timeout: 260 seconds]
fre2 has quit [Ping timeout: 260 seconds]
emacsomancer has quit [Ping timeout: 260 seconds]
Lears has quit [Ping timeout: 260 seconds]
craige has quit [Ping timeout: 260 seconds]
multun has quit [Ping timeout: 260 seconds]
gyroninja has quit [Ping timeout: 260 seconds]
mir100 has quit [Ping timeout: 260 seconds]
mrus has joined #nixos
notgne2 has quit [Disconnected by services]
notgne3 is now known as notgne2
<colemickens>
hpfr: I'm like 99% sure they have a bug in their front end
vykook has quit [Ping timeout: 260 seconds]
marcusr has quit [Ping timeout: 260 seconds]
leothrix has quit [Ping timeout: 260 seconds]
rivercat has quit [Ping timeout: 260 seconds]
realrokka_ has quit [Ping timeout: 260 seconds]
dongcarl has quit [Ping timeout: 260 seconds]
medvid__ has quit [Ping timeout: 260 seconds]
Cale has quit [Ping timeout: 260 seconds]
samueldr has quit [Ping timeout: 260 seconds]
dmj` has quit [Ping timeout: 260 seconds]
dongcarl8 is now known as dongcarl
<colemickens>
hpfr: set it, restart riot, worked for me. Or you can use kb/mouse to try to re-trigger the onchange for the form, one time I got it to go that way.
wucke13 has joined #nixos
<colemickens>
I had the same thing happen in browser on riot.im on a new firefox profile too though, hence my suspicion
Church- has quit [Ping timeout: 260 seconds]
choward has quit [Ping timeout: 260 seconds]
sigtrm has quit [Ping timeout: 260 seconds]
choward9 is now known as choward
realrokka has joined #nixos
alexbakker has joined #nixos
sigtrm has joined #nixos
rivercat has joined #nixos
gyroninja has joined #nixos
Gohla has joined #nixos
petar has joined #nixos
fre2 has joined #nixos
medvid__ has joined #nixos
<hpfr[m]>
colemickens: thanks. restarting worked
mutantmell has joined #nixos
Church- has joined #nixos
dmj` has joined #nixos
mir100 has joined #nixos
samueldr has joined #nixos
craige has joined #nixos
emacsomancer has joined #nixos
Cale has joined #nixos
greymalkin has joined #nixos
kcalvinalvin has joined #nixos
mlen has joined #nixos
mkaito_ has joined #nixos
multun has joined #nixos
vykook has joined #nixos
myme has joined #nixos
ravndal has joined #nixos
marcusr has joined #nixos
drakonis_ has quit [Read error: Connection reset by peer]
drakonis_ has joined #nixos
chagra has joined #nixos
wiml has quit [Quit: wiml]
Gohla has quit [Ping timeout: 260 seconds]
Gohla has joined #nixos
mutantmell has quit [Ping timeout: 260 seconds]
mutantmell has joined #nixos
LnL has joined #nixos
LnL has joined #nixos
LnL has quit [Changing host]
drakonis_ has quit [Read error: Connection reset by peer]
drakonis_ has joined #nixos
chagra has quit [Ping timeout: 240 seconds]
<hpfr[m]>
when trying to run flatpaks, I'm getting this error since recently:
<hpfr[m]>
`error: The name org.freedesktop.Accounts was not provided by any .service files`
<hpfr[m]>
couldn't find anything relevant on the web
<hpfr[m]>
NoctisLabs: thanks, progress. now I get unable to get session bus?
growpotk- has joined #nixos
drakonis has joined #nixos
learnday20 has quit [Quit: Leaving]
lovesegfault has joined #nixos
<hpfr[m]>
do I need to give the flatpak apps that use accounts-daemon unrestricted perms to the session bus?
tomberek has quit [Ping timeout: 240 seconds]
<hpfr[m]>
got spotify working with that perm, but Discord says `[WebContents] did-fail-load -202 ERR_CERT_AUTHORITY_INVALID retry in 9182.537240017531 ms`
<ashkitten>
right, i don't mind installing additional targets through rustup since that doesn't seem to break
<vandenoever>
except for the cargo-web thing which got me confused because i'm also running nixos stable rust and using cargo binaries in different environments gives hard errors
<vandenoever>
ashkitten: which rustup do you use for that?
<chvp>
ah great, didn't find it when I looked, but I must have missed something
<chvp>
LnL++
<{^_^}>
LnL's karma got increased to 25
dm9 has joined #nixos
<jassob>
Anybody here using Tectonic (https://tectonic-typesetting.github.io/en-US/)? It sets (maybe interesting/strange, or intended behaviour?) the date in documents to 1970 and I don't know how to actually get it to use the provided date instead
<__monty__>
jassob: Hmm, the site says "*completely reproducible document compiles*" so that's a feature, not a bug.
<jassob>
__monty__: Yes, that's why I thought it might be intended behaviour.. But it also completely ignores a provided date (e.g. \date{2020-20-25}) and not just sets it to 1970 when provided \today..
<jassob>
Also, running tectonic on another OS (like inside an Ubuntu docker container) does not reproduce this issue
Sigma has joined #nixos
<Taneb>
A coworker tried to modify a derivation by changing its attributes with //, and I can't well explain why it doesn't work
<__monty__>
Oh, that sounds weird. An explicit date is already reproducible.
orivej has joined #nixos
<__monty__>
Taneb: `(stdenv.mkDerivation ...) // { some = attr; }`?
veske2 has quit [Quit: This computer has gone to sleep]
veske has quit [Quit: This computer has gone to sleep]
<wedens[m]>
Taneb: proper way is to use `overrideAttrs`
<Taneb>
__monty__: yes
<Taneb>
wedens[m]: I know, but I want to know why that didn't work
<wedens[m]>
also note that `//` is not "merge" when it comes to nesting
zupo has quit [Ping timeout: 260 seconds]
<__monty__>
Maybe they turned the derivation into something that's not quite a derivation?
<jassob>
__monty__: Yes, that's what I thought too.. Oh well, I'll keep lurking and see if anyone knows anything about it. Thanks for your time!
<hyper_ch2>
__monty__: I'll just test it without and will add it if needed :) thx
<__monty__>
hyper_ch2: Oh, and SLiM was my favorite display manager. It just works and is simple as promised. Was removed from nixpkgs for not being maintained. Probably still works if you dig it up though.
<hyper_ch2>
:) I have no idea what a displaymanager's job is anways :)
<__monty__>
IME, hogging memory until you log in.
<hyper_ch2>
unused memora == wasted memors :)
<__monty__>
Which may not matter on newer systems but it's what drove me to SLiM on older boxes (which includes all my boxes).
<hyper_ch2>
(typing fails, please autocorrect)
<__monty__>
No no, even when logged in via ssh I could barely do anything on the system because of the display manager.
<hyper_ch2>
hehehe :)
drgbr has joined #nixos
pauldub has joined #nixos
kiara01 has joined #nixos
<kiara01>
hi. I would like to use my nix environment in CI. locally I use `nix-shell`. does that have an equivalent to like `bash -c "echo 123; echo 456"` to pass it commands up-front?
malSet has joined #nixos
<iqubic>
kiara01: Just use --command
chiefgoat has quit [Ping timeout: 268 seconds]
<kiara01>
@iqubic thanks!
philr has joined #nixos
<iqubic>
No problems.
o1lo01ol1o has joined #nixos
oever is now known as vandenoever
chiefgoat has joined #nixos
kiara01 has quit [Remote host closed the connection]
peanutbutter144 has quit [Ping timeout: 265 seconds]
cr4y1_ has joined #nixos
o1lo01ol1o has quit [Ping timeout: 260 seconds]
cr4y1 has quit [Ping timeout: 240 seconds]
kahiru has quit [Ping timeout: 258 seconds]
<{^_^}>
[nixpkgs] @rnhmjoj opened pull request #81013 → nixosTests.networking: disable virtual test with networkd → https://git.io/JvEbO
kahiru has joined #nixos
o1lo01ol1o has joined #nixos
peanutbutter144 has joined #nixos
dingenskirchen has quit [Quit: dingenskirchen]
<{^_^}>
[nixpkgs] @fgaz opened pull request #81014 → platforms: add bigEndian and littleEndian → https://git.io/JvEbP
zupo has quit [Ping timeout: 260 seconds]
zupo has joined #nixos
<Taneb>
How does Hydra work out if a jobset has been unchanged on re-evaluation?
<Taneb>
Is it based on the out paths?
<CMCDragonkai>
is there an alternative to squashfs?
<CMCDragonkai>
it's so slow
<CMCDragonkai>
that is `mksquashfs` is so slow
qasaur has joined #nixos
<sphalerite>
CMCDragonkai: you can reduce the compression level to make mksquashfs faster
<LinuxHackerman>
CMCDragonkai: you can reduce the compression level to make mksquashfs faster
<__monty__>
Taneb: Re the attribute added to a derivation. Was this an attribute that can be part of a derivation? Maybe evaluation just drops all the extraneous stuff when creating a derivation?
<sphalerite>
(the matrix bridge is also slow…)
<etu>
sphalerite: one might think that you and LinuxHackerman is the same person ;) (I know you are)
<__monty__>
Wow, saying exactly the same thing at exactly the same time with the exact same lack of capitalization and punctuation.
<sphalerite>
well actually I said it 10 seconds earlier on matrix. But it took 13 seconds longer to get through.
<sphalerite>
anyway.
<jophish>
For building pkgsCross for building armv7l derivations using my x86 machine from my x86 machine. However is there also an easy way of building for armv7l using x86 *from* the armv7l machine
<jophish>
i.e. pkgsCross doesn't allow me to change system
<Taneb>
__monty__: I think all attributes can be part of a derivation?
<sphalerite>
jophish: pass --argstr system x86_64-linux
<jophish>
at the moment I'm doing: import pkgs.path { system = "x86...", crossSystem = { system = armv7l...; }; }
<sphalerite>
jophish: and set the x86 machine up as a remote builder
<jophish>
sphalerite: suppose this is in configuration.nix
<jophish>
I'd like to have an overlay saying: self: super: { hello = self.pkgsCrossFrom.x86.armv7l.hello; }
<sphalerite>
yes, using nix.buildMachines _or_ ad-hoc using --option builders or --builders (depending on the command)
<jophish>
sorry, I explained poorly. That part is already working well
<sphalerite>
aah ok
<jophish>
it's that I'd like to do this in an overlay (or similar) and not by calling: import pkgs.path { system... crossSystem};
<sphalerite>
I don't know of an existing way, but maybe you can work something out by looking at how pkgsCross is defined
<jophish>
that's a good idea
<jophish>
looking at that it shouldn't be too hard
<jophish>
haha, they all depend on 'nixpkgsFun' and the comment says: Use to reevaluate Nixpkgs; a dirty hack that should be removed
bvdw has quit [Read error: Connection reset by peer]
<jophish>
which I suspect is the very thing I'm hoping to avoid
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
iyzsong has quit [Ping timeout: 240 seconds]
iyzsong has joined #nixos
magnetophon has quit [Remote host closed the connection]
c0c0 has quit [Ping timeout: 272 seconds]
tmaekawa has quit [Quit: tmaekawa]
<fgaz>
What is best, $(out) or ${placeholder "out"}?
<fgaz>
What is best between $(out) and ${placeholder "out"}?
vykook has joined #nixos
<sphalerite>
fgaz: not $(out) if you're working with stdenv, $out
<fgaz>
Sorry for the duplicate... buggy client. Also, what if a command builds and installs at the same time? Should I put it in the build or install phase?
<fgaz>
sphalerite: sorry, could you rephrase?
<sphalerite>
fgaz: $(out) expands to the output of running the command "out", in bash and posix sh at least
<sphalerite>
fgaz: I'm guessing you mean $out
<fgaz>
Ah that's because I'm in makeFlags
c0c0 has joined #nixos
<fgaz>
So $out gets escaped (?)
<fgaz>
sphalerite: ^
vykook has quit [Ping timeout: 255 seconds]
tmaekawa has joined #nixos
dansho has quit [Quit: Leaving]
<fgaz>
I'm not sure by which black magic $(out) turns into the correct thing, but if you grep nixpkgs for "PREFIX=", that's what half of the packages does
<fgaz>
The other half uses placeholder
veske2 has quit [Quit: This computer has gone to sleep]
veske has quit [Quit: This computer has gone to sleep]
waleee-cl has joined #nixos
LouisDK has quit [Ping timeout: 240 seconds]
LouisDK has joined #nixos
vykook has joined #nixos
veske has joined #nixos
veske2 has joined #nixos
<{^_^}>
[nixpkgs] @thefloweringash opened pull request #81015 → liblinear: use absolute install name on Darwin → https://git.io/JvENA
<fgaz>
Oh right, $() is make's syntax. Still, which one should be preferred?
tmaekawa has quit [Quit: tmaekawa]
LouisDK has quit [Ping timeout: 258 seconds]
malSet has quit [Quit: Quit.]
vykook has quit [Ping timeout: 265 seconds]
qasaur has quit [Ping timeout: 255 seconds]
<sphalerite>
fgaz: yeah, that's right. I usually prefer the environment variable, because that's the common existing convention
<sphalerite>
fgaz: there was also something about greppability or something why I found placeholders annoying at one point. But they can certainly be useful when an environment variable is impractical.
vykook has joined #nixos
<Dandellion>
what is Security in nixpkgs, it's passed to a lot of packages but I dont really know what it does. Nor can I find it in any manual
veske has quit [Quit: This computer has gone to sleep]
veske2 has quit [Quit: This computer has gone to sleep]
<LnL>
Dandellion: yeah, Security.framework is for macOS
vykook has quit [Ping timeout: 260 seconds]
noudle has joined #nixos
<Dandellion>
it gives you access to IOKit? that's it?
Neo-- has joined #nixos
<fgaz>
sphalerite: thank you. Any opinion on build vs install?
<sphalerite>
fgaz: not particularly, it doesn't really make a difference. Maybe worth considering is what seems to make more sense when overriding it with preBuild/postBuild/preInstall/postInstall hooks.
<dhess>
Here's something that's bothered me for awhile, and it seems I must be missing something obvious. Assume you have some remote builders that you want to take advantage of, and they have the same architecture as your personal machine. How can I get nix-build and friends to use my host *as well as* the remote builders? Right now it seems like if I have any remote builders with the same arch as my personal machine in /etc/nix/machines,
<dhess>
nix-build always uses those and never my machine.
delli3 has joined #nixos
<dhess>
I tried adding a localhost entry to /etc/nix/machines, but that causes deadlocks.
<dhess>
I tried the localhost trick awhile ago, maybe this is something that's been fixed in more recent versions of Nix?
<{^_^}>
[nixos-homepage] @edolstra pushed 2 commits to flake: https://git.io/JvEx7
<evanjs>
dhess: also interested in this, but moreso just setting it up in the first place :D haven't gotten around to it yet but I've had mixed results with distcc and icecream (https://github.com/icecc/icecream) before
<evanjs>
I think one of my requirements is conditionally building on remote machines _if_ they're available.
<evanjs>
Oh yeah, anybody else having issues with neovim + youcompleteme on nixos-unstable? Haven't narrowed the issue down to nix, nvim, or ycm, yet
<{^_^}>
[nixpkgs] @vanschelven opened pull request #81020 → python.pkgs.httpretty: disable flaky test → https://git.io/JvExx
vykook has quit [Ping timeout: 260 seconds]
<evanjs>
I need to get a _recent_ Mac system so I can do iOS stuff again... *sigh* still no news on the Nix/Catalina story, I figure?
<dhess>
evanjs: it's working fine, LnL just pushed the new installer to master
<evanjs>
ahhhh sweeeeeet
Rusty1 has joined #nixos
<dhess>
the only catch is if you want to encrypt the /nix volume, in which case you need to do a few extra steps and you might have to restart Terminal after a reboot
<dhess>
the Terminal thing is a PITA but it's not a show-stopper.
<evanjs>
meh still haven't messed with encryption but that doesn't sound bad at all
<dhess>
if you don't encrypt /nix then it all works fine, assuming you use the new Nix installer.
drakonis_ has quit [Ping timeout: 258 seconds]
<evanjs>
yeah that sounds good. ah dumb right now, but where is the new installer? Not like I have a Catalina machine to worry about atm :P
<evanjs>
*am dumb right now
<dhess>
the new installer just got pushed to the nix master branch, I believe. I'm not sure how to get an installer from master but someone in #nix-darwin can probably help you with that.
<{^_^}>
[nixpkgs] @mmilata opened pull request #81022 → [20.03] nixos/parsoid: fix service by reinstating nodePackages.parsoid → https://git.io/JvEpT
<Yaniel>
a shame mesa 20 didn't quite make it
<{^_^}>
[nixpkgs] @bignaux opened pull request #81023 → quickbms: init at 0.10.1 → https://git.io/JvEpI
<evanjs>
oh okay I wasn't seeing it there and was just curious what commit or etc
<LnL>
dhess: master?
vykook has joined #nixos
<evanjs>
(ahhh perhaps I'm not crazy :D )
<dhess>
LnL: did you not push it? I thought I saw messages to that effect
<evanjs>
oh yeah! anyway my YCM problem: "YouCompleteMe unavailable: No module named 'future'"
<{^_^}>
nix#3212 (by LnL7, 15 weeks ago, open): install: configure and bootstrap synthetic.conf on darwin
<dhess>
evanjs: ok so the Nix installer doesn't support Catalina just yet but should soon. There are numerous scripts out there to work around it for now.
<AmandaC>
samueldr: the devil must have been stuck in traffic. Was browsing my stories for updates this morning -> gpu hang. :D
<evanjs>
dhess: right right which is totally fine for me. It's already on my late 2011 mbp which totally doesn't support Catalina so I'll probably be fine by the time I upgrade :P
vykook has quit [Ping timeout: 240 seconds]
<LnL>
evanjs: you don't know how good I am at procrastinating :p
<rnhmjoj>
ghcristensen: python2 is hidden even in critical infrastructure like hydra. i uncovered this dirty secret a few days ago and i'm trying to fix it
<clever>
[clever@amd-nixos:~]$ ls /nix/store/*python2* | wc -l
<clever>
5098
<clever>
fgaz: note that your counting every lib and bin dir within each python2
<etu>
$ ls /nix/store/*python2* | wc -l
<etu>
244
wavirc22 has quit [Ping timeout: 255 seconds]
<gchristensen>
rnhmjoj: of course it is :) thank you a lot for your work. there is definitely work ongoing to remove python2 overall, but many places to do it. for example: nix-store -qR /run/current-system | grep python2 | wc -l
<{^_^}>
nixops#1227 (by grahamc, 3 days ago, open): Python 2to3, too
<emily>
eon: yeah
wavirc22 has quit [Ping timeout: 258 seconds]
wavirc22 has joined #nixos
<hyper_ch2>
gchristensen: why do I keep getting those "join the nixos organisation" invites on github? it's so annoying because I don't want to use 2fa
waleee-cl has quit [Quit: Connection closed for inactivity]
<gchristensen>
hyper_ch2: it shouldn't be inviting you multiple times. ping me in 1.5h?
<dhess>
gchristensen: oh nice! I'll try to build that later today and test it
<gchristensen>
hyper_ch2: please ping in 1.5h -- gotta go -- on a call (also want to talk to you more then, dhess)
JonReed has quit [Ping timeout: 260 seconds]
<hyper_ch2>
I'll try to remember
wavirc22 has quit [Ping timeout: 255 seconds]
<dhess>
gchristensen: ok I should be lurking
wavirc22 has joined #nixos
rail has joined #nixos
<{^_^}>
[nixos-org-configurations] @garbas opened pull request #104 → adding garbas home ip address to the list which can ssh to nixos.org → https://git.io/Jvuef
<__monty__>
dhess: Fwiw find (or fd) probably handles that better than globbing.
t420babe has quit [Quit: Ping timeout (120 seconds)]
exarkun has quit [Excess Flood]
t420babe has joined #nixos
exarkun has joined #nixos
emilsp has quit [Quit: bye]
unacceptable has quit [Quit: ZNC 1.7.1 - https://znc.in]
swapgs has joined #nixos
swapgs has joined #nixos
swapgs has quit [Changing host]
emilsp has joined #nixos
unacceptable has joined #nixos
Acilim has quit [Remote host closed the connection]
chagra_ has joined #nixos
<{^_^}>
[nixpkgs] @veprbl merged pull request #81015 → liblinear: use absolute install name on Darwin → https://git.io/JvENA
<{^_^}>
[nixpkgs] @veprbl pushed commit from @thefloweringash to master « liblinear: use absolute install name on Darwin (#81015) »: https://git.io/JvueH
iqubic` has quit [Quit: ERC (IRC client for Emacs 28.0.50)]
<{^_^}>
[nixpkgs] @veprbl pushed commit from @thefloweringash to release-20.03 « liblinear: use absolute install name on Darwin (#81015) »: https://git.io/JvueF
<dvn>
o/ - trying to use an overlay for the first time... getting "error: infinite recursion encountered, at undefined position"
<{^_^}>
[nixpkgs] @KoviRobi opened pull request #81030 → polyml: fix with new libffi → https://git.io/JvueN
hyper_ch2 has quit [Remote host closed the connection]
<gchristensen>
dvn: you are probably using "self" somewhere where "final" would be better
<gchristensen>
or "final" when "self" would be better
<gchristensen>
s/final/super/
<gchristensen>
man I am just a mess. I need to step making such a mess right now. sorry.
<dvn>
put it in ~/.config/nixpkgs/overlays/wayland.nix
wavirc22 has joined #nixos
vandenoever has joined #nixos
vandenoever has joined #nixos
vandenoever has quit [Changing host]
<Taneb>
dvn: what you've linked isn't itself an overlay, it's a NixOS configuration.nix file which uses an overlay
c0c0 has quit [Quit: WeeChat 2.6]
<dvn>
ah, ok.. was beginning to suspect this the more i read
waleee-cl has joined #nixos
orivej has quit [Ping timeout: 258 seconds]
<Taneb>
dvn: those instructions are getting you to mix that into your /etc/configuration.nix file
<Taneb>
I think they're not very clear instructions :(
<colemickens>
:(
<dvn>
but at least it's consistent nix style. you must know how things work in order to use anything ;)
<colemickens>
PRs or issues welcome. I'm happy to fix. Code is no good if it's hard to use!
<dvn>
i'm teasing. <3 nix -- i come from guix, but the overlay concept is quite different there, so i was lost
<dvn>
colemickens: once i figure out how to use the overlay, then i will make a PR
<Taneb>
colemickens: my issue is there's no instructions about what to do with that exerpt
<Taneb>
If a new user, like dvn, sees that, the only clue they have is it uses overlays (there's a broken link there as well)
<Taneb>
If they google how to use overlays they're going to get the wrong thing
JonReed has joined #nixos
<JonReed>
Hi! Why is `buildFHSUserEnv.targetPkgs` and `multiPkgs` a function that accepts `pkgs`? In other words, why not simply have `buildFHSUserEnv.targetPkgs = [ foopkg ]`, compared to what it is right now `buildFHSUserEnv.targetPkgs = (pkgs: [ pkgs.foopkg ])`? What if I supply a package there not from `pkgs` that the function accepts but from my own
<JonReed>
custom pkgs? Will something go wrong?
<abathur>
anyone aware of a link/post that directly addresses this question (from nixpkgs issue 2925 on fixing install for macos Catalina): "Where can I find more info about why Nix uses /nix (and not /opt/nix for example), and why it's so difficult to change?"
<simpson>
JonReed: Some wrongness might occur if there is runtime linkage. But likely the worst that will happen is two copies of glibc, that sort of thing.
drakonis has quit [Read error: Connection reset by peer]
drakonis has joined #nixos
<JonReed>
simpson: Why would it copy glibc?
<simpson>
abathur: I just checked and I don't see anything obvious. (Nix ought to have an FAQ page!) The reasoning behind using a top-level directory IIRC is to keep paths as short as possible.
<simpson>
JonReed: If you happen to mix two different pkgs. Why would it be a problem to put a pkgs there?
wavirc22 has quit [Ping timeout: 240 seconds]
<__monty__>
abathur: It's hard to change because paths get hardcoded into binaries.
<__monty__>
abathur: You could change the store path fairly easily but then you can't use any caches.
vykook has quit [Ping timeout: 252 seconds]
<abathur>
simpson: that's my inference as well; I know I've seen comments around where people mention some of the difficulties of changing it, but I don't recall seeing a comprehensive thing I can point the asker to
<{^_^}>
[nixpkgs] @mjlbach opened pull request #81033 → Add bazel 1 for backwards compatibilitiy → https://git.io/JvuvH
growpotkin has joined #nixos
<simpson>
gchristensen, infinisil: ^^^ Is there a singular FAQ entry that we can use for this?
<JonReed>
simpson: What puzzles me is why a normal derivation has `mkDerivation { buildInputs = [] ;}`, but buildFHSUserEnv has `buildFHSUserEnv { targetPkgs = (pkgs: []); }`.
<abathur>
__monty__: nod; I've seen all of these come up; fwiw I think the asker is asking why it can't be changed *for everyone* and not just why it sucks for an individual to do it
<simpson>
JonReed: Perhaps FHS envs need to do something to the pkgs, so that bringing your own unprepared pkgs would not work.
<simpson>
abathur: Oh, sure. Changed to what? Most paths would be longer, and folks asking for this usually want /home/etc or something else non-root; that is, what they're *actually* asking for is Nix installation without root privs.
<JonReed>
simpson: Ok, thanks!
<__monty__>
abathur: The reasoning there is about the length of the path. Patchelf can't easily patch paths that are shorter then what it needs to substitute or something.
<abathur>
__monty__: /opt/nix, as the question said; this has been a recurring question in the macos Catalina discussion, since I gather it also fixes some other platforms or that some other platforms are also discussing making root / readonly
<abathur>
simpson: oops, wrong name :)
wavirc22 has joined #nixos
<abathur>
simpson: also, I think the question is global, not about individual overrides
vykook has joined #nixos
<clever>
abathur: nix can be compiled to use /opt/nix/store instead, but you cant mix things, so you can never have a /opt/nix/store path on a /nix/store machine
<clever>
abathur: that lack of mixing, means hydra.nixos.org can never touch mac stuff again, and cache.nixos.org would be useless for mac's
<abathur>
again for clarity, I think the questioner is asking about a global change: "Where can I find more info about why Nix uses /nix (and not /opt/nix for example), and why it's so difficult to change?"
<clever>
abathur: so now you have to duplicate all of the infra, to support macos, and nixops wont work on a mac either
<__monty__>
Switching everyone over at this point would be pretty unfortunate. There's TBs of cached files.
<gchristensen>
pretty much 100% certain we're not switching, so not much use worrying about it
<simpson>
abathur: The global rationale is what I said before: Shorter paths are better.
<simpson>
Like, we don't really *care* about Fedora, other than that it's the best example of an SELinux distro, and that it thus makes a good barometer for our SELinux support. I
<dhess>
gchristensen: if it comes to that, I'll just bite the bullet and switch all of my infrastructure to /opt/nix.
<__monty__>
gchristensen: What are you basing that on? IIRC for Catalina apple devs actually helped in finding a solution.
<gchristensen>
on their general direction of becoming increasingly hostile to third party software overall
<simpson>
*On Linux, we *have* a preferred userland already: NixOS. And AFAICT OSX is the *only* problematic kernel, which might have something to do with its publisher~ It's hard to rationalize embracing a vendor that actively is hostile towards us, *especially* when that vendor is a luxury fashion brand.
<gchristensen>
simpson: in the future, consider deleting the part after the last ","
<{^_^}>
[nixpkgs] @mjlbach opened pull request #81035 → Update tensorflow to work with new glibc → https://git.io/Jvuf3
<simpson>
__monty__: People don't get credit for "nice job breaking it, hero"; they don't get rewarded for first creating massive problems for the community at large, and then making token efforts to undo damage that they've caused.
<simpson>
gchristensen: I understand your point, but I hope that you understand mine: Folks who put down money on Apple hardware are spending a lot of cash in order to be locked out of functionality.
<gchristensen>
I understand your perspective, but that certainly isn't why I spent that money
<__monty__>
simpson: It's not always a choice. Some software is only support on macOS and not all people are capable of writing the software they'd personally have to miss out on.
mstruebing has quit [Quit: WeeChat 2.5]
abathur has quit [Read error: Connection reset by peer]
<simpson>
__monty__: We'll have to agree to disagree about what computers are for. Don't worry about it.
<gchristensen>
true enough
<dhess>
abathur: look the answer to that poster's question is that Nix uses /nix because that's what eelco wanted to use when he wrote it, and derivation hashes include the full store path as part of the hash and therefore you couldn't build packages for a system that uses /nix on a system that used, say, /opt/nix
<dhess>
that is the short answer. An even shorter answer is, "It doesn't matter because the people who run Nix and nixpkgs are not going to change it for macOS."
zupo has joined #nixos
<abathur>
FWIW, @lilyball noted at one point in the issue that `/opt/nix/s` would maintain the same character count, if not the same transparent naming
* emily
thinks there's no good reason to not change it
<emily>
it also breaks the FHS and stuff
<emily>
sure those aren't set in stone, but there's no good reason to use /nix in the first place, so.
<emily>
nix store paths are already incredibly long.
<simpson>
dhess: Neither of those answers are very satisfactory, though; they wouldn't explain why we won't change it for FreeBSD either. I think that it's worth *eating* the cost of fixing /nix as our path, and telling folks that we expect kernels to be able to deal with this just like they deal with /var, /opt, /sbin, /users, or any other "new" top-level directory.
<emily>
+1 for /opt/nix/s, you'd need to symlink /opt/nix/store to it for backcompat anyway
<emily>
which gives it more discoverability
<dhess>
simpson: I agree that they're not satisfactory but it's a complete waste of everybody's time to keep legislating it.
<emily>
(I don't use macOS so I am free from sin^Wbias)
<emily>
(except perhaps the bias of caring about vast majority of people who run non-free kernels)
<simpson>
dhess: Oh, definitely. Somebody should make an FAQ page that covers this sort of adversarial question; it would allow us to centralize and refine our reasoning.
<emily>
there's nothing adversarial about asking why we do something that breaks on multiple systems and gratuitously goes against standards, good grief
<emily>
it's a totally legitimate question
sigmundv has quit [Ping timeout: 260 seconds]
<simpson>
Adversarial questions are legitimate. My main technique for discovering whether software is any good is to take it outside and hit it against the wall a bunch.
<jared-w>
I do think there is an argument to be made for questioning why the full store path is part of the hash
<emily>
that one is harder to fix since store paths appear in the content itself
<dhess>
jared-w: I agree with that and I thought that was one of the things this intentional store RFC was trying to address, but I might have misunderstood that aspect of it.
<jared-w>
simpson: and I do think there's merit for asking adversarial questions in a non aggressive way. Phrasing helps a lot
<{^_^}>
[nixpkgs] @cillianderoiste pushed 2 commits to master: https://git.io/JvufX
<jared-w>
dhess: If the intensional store RFC addresses the nix store in the hash, it does so purely as an implementation detail or as a "let's get this out of the way while we're at it" sort of thing, iirc
<simpson>
jared-w: Sure. I'm not sure what's aggressive about the asker on GH, though; their phrasing seems quite reasonable. Additionally, I don't think that we should police the tone of folks on GH simply because they can't find links to docs that we haven't written.
ixxie has joined #nixos
<jared-w>
dhess: The intensional store is really about turning nix from a suspending deep constructive trace build system to a suspending constructive trace build system (that is, adding support for incremental builds in nix)
JonReed has quit [Remote host closed the connection]
<dhess>
oh ok. When the /nix on Catalina thing first came up and people started talking about changing the store path, I thought someone pointed to the intentional store RFC, but I might be mixing up issues.
Ianulus has joined #nixos
<jared-w>
It probably addresses that somewhere. Having the nix store location change the hash is incompatible with the idea of a content addressable store
wavirc22 has quit [Quit: ZNC 1.6.6+deb1ubuntu0.2 - http://znc.in]
<jared-w>
(but the explicit goal of "being able to change the store prefix arbitrarily" isn't necessarily addressed, it's more a implementation detail that needs to change in implementing an intensional store)
<jared-w>
s/intensional/intentional/
<dhess>
haha sorry I've been learning constructive mathematics recently and got "intensional" stuck in my head :)
<Ianulus>
I'm not there yet, a bash command will be fine for now :)
<jared-w>
simpson: that's fair; I'm not a fan of tone policing. I thought you were talking adversarial questions that you were posing
<simpson>
jared-w: Oh, sure; like?
<jared-w>
simpson: I don't have any examples, I thought you were talking in a more hypothetical nature and missed that there was a github thread being referenced
<jared-w>
simpson: you're fine, I haven't seen you write anything aggressive, but maybe I haven't looked hard enough :p
drakonis has quit [Read error: Connection reset by peer]
drakonis has joined #nixos
<Ianulus>
`nix-shell --command export PATH=$PATH:/my-path` fails with `error: getting status of '/home/ben/PATH=/home/ben/bin/[…]'`
hyper_ch2 has joined #nixos
<hyper_ch2>
gchristensen: I got another of those emails
hyper_ch2 has quit [Remote host closed the connection]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<jtojnar>
oof, partial github outage
hyper_ch2 has joined #nixos
<hyper_ch2>
gchristensen: bad connection on the train right now
<gchristensen>
hyper_ch2: for how long have you been receiving these repeated notifications?
<hyper_ch2>
I got a bunch of them a while back
<hyper_ch2>
would need to check the irc log
LouisDK has joined #nixos
<hyper_ch2>
now I haven't gotten any of them for some time and now today 3 again
<hyper_ch2>
I cancelled the 3rd invitation just now
<gchristensen>
hyper_ch2: how many did you get from before 2 days ago?
<hyper_ch2>
I didn't get any 2 days ago
<gchristensen>
okay, then it is related to a github database outage
<hyper_ch2>
it was january 28 where I got a bunch of those
smatting has quit [Ping timeout: 265 seconds]
<hyper_ch2>
[Dienstag, 28. Januar 2020] [09:07:36 CET] <hyper_ch2> why do I keep getting invites into the nixos organization when I decline one?
drakonis has quit [Read error: Connection reset by peer]
vandenoever has joined #nixos
vandenoever has joined #nixos
vandenoever has quit [Changing host]
drakonis has joined #nixos
kleisli has quit [Ping timeout: 255 seconds]
wavirc22 has quit [Ping timeout: 260 seconds]
<gchristensen>
possibly another database outage then, though I can't seem to load the dashboard for then
tazjin has quit [Read error: Connection reset by peer]
tazjin has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
veske has joined #nixos
hyper_ch2 has joined #nixos
veske2 has joined #nixos
<hyper_ch2>
I just find it weird when I decline it and get new invites
werner291 has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
<gchristensen>
yeah, definitely. on one hand, we could record who we've invited instead of trusting github to do that -- on the other hand, maybe it is good to recreate the invitation so you can join if you choose. on our end, we don't know you declined -- just that you haven't accepted it.
<Raito_Bezarius>
I'm unsure on how to package the Python app, I see how to do the config part though
cosimone has joined #nixos
<Raito_Bezarius>
I tried pypi2nix, it failed
<Raito_Bezarius>
(some error regarding a dependency)
abathur has quit [Read error: Connection reset by peer]
<Raito_Bezarius>
I cannot use pythonPackages.* because those are not necessarily compatible
<Raito_Bezarius>
reqtxt2nix does not work because it's relying on python 3.4 which is deprecated now and I tried to fix it but didn't find anything on the breaking changes of the pip module
<Raito_Bezarius>
python2nix seems to be less preferable than pypi2nix
wavirc22 has joined #nixos
<evanjs>
but yeah I'm using copy, probably wrongly; wanting it to push buildInputs recursively as well, I think
`mrleaW_ has quit [Quit: `mrleaW_]
cosimone has quit [Remote host closed the connection]
<Raito_Bezarius>
Just gave a try to pip2nix, it failed at installation
justanotheruser has joined #nixos
cr4y1__ has quit [Ping timeout: 260 seconds]
veske has quit [Quit: This computer has gone to sleep]
veske2 has quit [Quit: This computer has gone to sleep]
<leungbk>
I've been trying to add an Emacs package to Nixpkgs, rebuilding my installation of NixOS to use my modified branch of the Nixpkgs repo, but my installation doesn't seem to detect the package. Would anybody be able to help me diagnose this? http://paste.debian.net/1132159/
<nilsirl[m]>
For some reason, the bash session that I get when logging in with a tty is not a login shell which means that `.profile` isn't loaded
linarcx has joined #nixos
wavirc22 has joined #nixos
<nilsirl[m]>
This has only started happening recently on nixos-unstable. AFAIK I haven't done anything that would change the previous behaviour so I suspect it has to do with nixos-unstable
gagbo has quit [Quit: I'm out !]
<evanjs>
clever: switched to the new cachix I made for work stuff and it does everything for me and seems to be doing so recursively :D
<nilsirl[m]>
I can't find anyone reporting the same thing on github which makes we wander as to whether It's related to my system or nixos
<nilsirl[m]>
*wonder
wavirc22 has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @Ma27 pushed to release-20.03 « nixos/networkd: disable vrf sub-test which tests the behavior of tcp-packets »: https://git.io/JvuIK
wavirc22 has joined #nixos
<JonReed>
leungbk: Can't really help you, but one suggestion would be trying having `pkgs = import /home/brian/mygit/nixpkgs {};` in `test-emacs.nix`, instead of importing it from environment. This would close at least one vector of possibly somehow messing up and not switching to that custom branch
<hyper_ch>
gchristensen: getting more and more join invitations now
<leungbk>
JonReed: Thanks, it seems to detect the package now.
<evanjs>
clever: is this the best (or even proper) way to push everything I need for cache deployments? `nix-store -qR $(readlink ./result*) --include-outputs | xargs cachix push <cache>`
<gchristensen>
hi miah :)
<evanjs>
Either my build is messing things up and making it harder to see what dependencies are needed, or I'm querying such deps wrong
<gchristensen>
miah: nice to see you around these parts!
<miah>
indeed
<miah>
realized i wasn't idling in this channel and i could be :P
<miah>
(and maybe i could chat too)
<gchristensen>
=)
<gchristensen>
maybe even take a chance and get sucked in :P
<miah>
lol yes
<miah>
its been a bit since i tried nix, and tbh last time was nix on mac.. which was..
<miah>
(its been a few years)
<gchristensen>
I'd recommend giving NixOS a go in a VM or something for the full experience
<miah>
ya, i have a spare pc i'll throw it on
<gchristensen>
nice
mexisme_ has joined #nixos
<miah>
usually i throw arch on everything. but the thing that i tire of is reconfiguring everything
evanjs has joined #nixos
mexisme has quit [Ping timeout: 258 seconds]
<DigitalKiwi>
20:16 gchristensen: I'd recommend giving NixOS a go in a VM or something for the full experience
<DigitalKiwi>
this is how i went from macos to nixos full time
wavirc22 has quit [Ping timeout: 255 seconds]
<energizer`>
gchristensen: if I have a command available on PATH when running `nix-shell`, should it also be available on PATH when running `lorri shell`? That's what I expected but it's not what I'm seeing.
<gchristensen>
energizer`: I would think so, yeah
Fare has joined #nixos
acarrico has joined #nixos
<miah>
DigitalKiwi: ya i run arch on my desktop. but work gave me a macbook for work. wish i could have a linux laptop.
* DigitalKiwi
really likes parallels desktop on mac
<jared-w>
miah: having used nix on macos in anger lately, I'd say that it's much better now. Well, Catalina has kind of thrown things for a loop as per usual, but I think that seems to be mostly fixed now?
<cole-h>
Is it possible to override `rev` and `sha256` in `src = fetchgit { ... };`? I tried `old.src // { rev = "..."; sha256 = "..."; };` but that didn't behave as I expected
<drakonis>
doesn't catalina pose an interesting problem for nix?
<drakonis>
because soon there'll be another mac release later this year
iqubic` has joined #nixos
<drakonis>
which will retain this roadblock
werner291 has joined #nixos
<drakonis>
maybe it'll add another one
jgeerds has joined #nixos
iqubic has quit [Ping timeout: 260 seconds]
leungbk has quit [Ping timeout: 240 seconds]
philr has joined #nixos
<{^_^}>
[nixpkgs] @primeos pushed to master « gns3-{gui,server}: Fix the build by disabling the psutil tests »: https://git.io/JvuYK
<DigitalKiwi>
drakonis: clearly we just need to show them how many people move to nixos rather than deal with nix on macos so they give us a break
<DigitalKiwi>
there are dozens of us!
<drakonis>
there's a reasonably large number of users that run nix instead of homebrew on macs
<drakonis>
this will probably hurt them quite a lot.
<miah>
jared-w: interesting. i only recently upgraded to catalina.. haven't used my macbook since. try to use it minimally =)
infinisil has quit [Quit: Configuring ZNC, sorry for the joins/quits!]
<miah>
apple feels more and more hostile to me every year, definitely some interesting problems on that platform. but i really feel like i'm not the target.
infinisil has joined #nixos
nerdmaxx has joined #nixos
dingenskirchen has joined #nixos
<AmandaC>
Those hostilities, and the increasing realisation I'd probably be unable to afford any physical damage that happen to my macbook, is wht pushed me to buy a System76 laptop in 2017
<AmandaC>
I need to get around to nuking that thing and selling it offf sometime
<johnnywho>
I need some help installing jdk10 with an overlay
Thra11 has joined #nixos
vandenoever has quit [Ping timeout: 260 seconds]
<johnnywho>
I need to run a prebuilt astronomy app that needs the deprecated corba but won 't run with jdk8 that is available because it is built with a later version of java
<kini>
I'm trying to use `nix copy` to get a closure I build with `nix-build` on another machine, but I'm getting the error "cannot add path <whatever> because it lacks a valid signature". After some googling I added myself to the nix trusted users by setting nix.trustedUsers in configuration.nix and ran `nixos-rebuild switch`, but that doesn't seem to resolve the issue.
<kini>
Am I misunderstanding what I should have done?
<colemickens>
Oh my, this is rather confusing now that I look at this from different eyes. Thanks for the feedback, I can definitely make this easier to understand.
<colemickens>
Actually, I'm going to cleanup the Wiki page and then link out to that.
mexisme_ has quit [Ping timeout: 260 seconds]
<colemickens>
We should just have good examples on the Wiki. And I need to put my money where my mouth is. I think more content on the wiki will be good as nixos seems to be increasingly getting attention (or maybe my perception is wrong)
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
rajivr___ has quit [Quit: Connection closed for inactivity]
clodeindustrie has joined #nixos
<clodeindustrie>
hi there, I am new to this so bear with me
mexisme_ has joined #nixos
<clodeindustrie>
trying to create a private/local package for an older version of an app, I have copied over the default.nix from the current version that's in the repository
<clodeindustrie>
and thought I could just update the version in the file and nix-shell it
<clodeindustrie>
however it's not that easy :)
tilpner has quit [Ping timeout: 258 seconds]
<clodeindustrie>
the file is listing a bunch of makeLibrary from stdenv.cc.cc.lib
tilpner_ has joined #nixos
<clodeindustrie>
however one seems to be missing from my system libxcb
<clodeindustrie>
do I need to declare it as some sort of dependency?
<{^_^}>
[nixpkgs] @flokli pushed 2 commits to release-20.03: https://git.io/JvuGG
<ajs124>
was the version of that app ever in nixpkgs? if so, you copied the current version, not that one, right?
<ashkitten>
i don't use either gnome or kde and decided that kde wallet was the lesser of two evils
<hexa->
emily: have you experienced that the acme systemd unit is not being started directly after the rebuild? it didn't for me on two systems now and the timer says it will in 5 days :)
<ashkitten>
but if i try to use `kwallet-query -l kdewallet` it says `Couldn't start kwalletd: QDBusError("org.freedesktop.DBus.Error.Spawn.ChildSignaled", "Process org.kde.kwalletd5 received signal 6")`
<clodeindustrie>
mm it's not working, I have no idea of what I'm doing
<colemickens>
clodeindustrie: whats the pkg?
<emily>
hexa-: yeah sounds about right, because it's just the timer that gets restarted I think. it also won't actually renew your certificate until it's too old unless you set that variable really high
<emily>
the whole UX could use some work but I don't know exactly how to make it work out nicely
<colemickens>
clodeindustrie: I can try my hand at it and see if we can make a wiki page out of it or something like that :)
<hexa->
emily: ok, same here. think merging into 20.03 was a bit premature then :)
<emily>
hexa-: the lego rewrite you mean? maybe
<hexa->
yep
<emily>
I don't think the unit taking a few more days to run really matters honestly
<clodeindustrie>
colemickens:mongodb-compass
<hexa->
huh? I don't have certificates right now after rebuilding
<clodeindustrie>
that would be awesome
<emily>
since the task is run quite a bit more often than is needed anyway
<hexa->
have to poke the systemd unit manually
<colemickens>
clodeindustrie: you are trying to package a newer version?
<clodeindustrie>
colemickens:https://gist.github.com/clodeindustrie/1723252cbf02a0d14620403f592fd202 that's what my default.nix look like after copying from the repo, updating the version and adjusting the top line what was giving grief
<clodeindustrie>
yes
<colemickens>
clodeindustrie: if that's the case, you might just want to update it in the tree and send a PR to nixpkgs :)
Thra11 has quit [Quit: WeeChat 2.7]
jonten has quit [Ping timeout: 272 seconds]
<colemickens>
usually updates to latest version are generally accepted
<clodeindustrie>
or an older one, I need to test which one doesn't segfault on my machine :(
<colemickens>
clodeindustrie: okay, I see. I'll leave that to you then.
<{^_^}>
[nixpkgs] @teto opened pull request #81067 → allow to run luarocks test → https://git.io/JvuGQ
<colemickens>
I'll see if we can at least get a shell+out-of-tree example going, since that's what you asked for and seems like a good thing to have an example of anyway
<clodeindustrie>
that would be amazing
erasmas has quit [Quit: leaving]
bvdw has quit [Read error: Connection reset by peer]
<clodeindustrie>
I'm getting "build of .....drv failed"
kleisli has quit [Ping timeout: 240 seconds]
<clodeindustrie>
but at least it tried
<clodeindustrie>
what's in cmpkgs.tar.gz ?
<Guest_>
Is there a way I can control which python is running setup.py in a buildPythonPackage derivation? My best guess as to why my derivation is failing is that the python running setup.py isn't properly configured but setup.py pulls data from that install
<{^_^}>
[nixpkgs] @jtojnar pushed to netpbm « netpbm: split into multiple outputs »: https://git.io/JvuZe
<Guest_>
In particular, I need it to properly respond to torch.cuda.is_available() and torch.utils.cpp_extension.CUDA_HOME
<Guest_>
those work if I load python from a :s shell
<Guest_>
but not in the builder
<Guest_>
(which also makes me think that cudatoolkit isn't landing properly, but I can't figure that out either)
<Guest_>
(I've tried it in nativeBuildInputs, buildInputs, and propagatedBuildInputs)
<colemickens>
clodeindustrie: that's just my fork/branch of nixpkgs
<colemickens>
clodeindustrie: you can change that to <nixpkgs> to use your system channel
smatting has quit [Ping timeout: 255 seconds]
<clodeindustrie>
alright so it's telling me a dependency couldn't be built
<clodeindustrie>
ah silly me
<clodeindustrie>
the sha is not the same
noudle has quit []
<colemickens>
Um, come to think of it, you might want to make sure you're NOT using my branch.
<{^_^}>
[nixpkgs] @Infinisil merged pull request #80861 → nixos/acme: move the crt to fullchain.pem → https://git.io/Jv0aP