<jaym>
anybody having issues with gtk programs... i ahve multiple ones of them segfaulting: eclipse, chromium, terminator with: segfault at 10 ip 00007f5cf7ddd500 sp 00007ffc28004930 error 4 in libgdk-x11-2.0.so.0.2400.31[7f5cf7d6a000+b0000]
markus1189 has joined #nixos
magnetophon has quit [(Ping timeout: 258 seconds)]
markus1199 has quit [(Ping timeout: 265 seconds)]
oida has joined #nixos
ambro718 has quit [(Ping timeout: 260 seconds)]
goibhniu has quit [(Ping timeout: 245 seconds)]
Shou has joined #nixos
Wizek has quit [(Ping timeout: 240 seconds)]
glines has quit [(Ping timeout: 250 seconds)]
<jaym>
now i cant launch firefox :(
<clever>
jeaye: you there?
Sonarpulse has joined #nixos
<jeaye>
clever: Yes.
<clever>
jeaye: ive got some other ideas on how to do things like nixos-in-place, have you seen the current netboot stuff?
<jeaye>
I haven't installed NixOS since I put it on my VPS a year or so ago.
<clever>
jeaye: this creates a directory containing netboot.ipxe, and if you point ipxe at it, you get a nixos env
<clever>
jeaye: it works, by putting the entire root.squashfs from the install cd into the initrd
<jeaye>
Seems like it could be a simpler approach.
<jeaye>
Simpler than what nixos-in-place is doing, I mean.
<clever>
jeaye: i have an idea on how to abuse that for even more powerfull stuff, one sec
<eacameron>
What's the difference between || and or?
<Sonarpulse>
or is a mess
<Sonarpulse>
or is a function
<Sonarpulse>
that is basically fold (||)
<Sonarpulse>
list: (fold (a: b: a || b) list
<Sonarpulse>
or is also used for undefined fields
<Sonarpulse>
a.b or foo
<Sonarpulse>
uses foo if a has no b
<eacameron>
Sonarpulse: Ah so "or" is like Python's "or" whereas || is a true boolean operator
<jaym>
thoughts on what to do when a vm based on your configuration works fine, but your machine itself does not... what are places besides the home directory i should be looking for gtk related state
<Sonarpulse>
eacameron: I don't konw python
<eacameron>
Sonarpulse: Well it sounds like Python: None or 3 = 3
<clever>
jeaye: basicaly, my idea is to download the kernel+initrd for netboot, then just kexec them from the current OS
AllanEspinosa has quit [(Ping timeout: 258 seconds)]
<eacameron>
Sonarpulse: LOL i was like O_o
<clever>
jeaye: and if the netboot image is configured correctly, you will get a sshd running purely from ram, that is capable of partitioning and doing nixos-install
slack1256 has joined #nixos
<eacameron>
How do I bring "or" into scope?
<clever>
jeaye: but it fails for the same reason boot.crashdump.enable is broken
<clever>
jeaye: if this kexec bug can be fixed (does it even happen on another distro?) you can get full control of the partitioning
<jeaye>
It's a neat idea.
joelmo has quit [(Quit: Connection closed for inactivity)]
<clever>
and the kexec gets rid of the need to control the network boot stuff
<clever>
so it could be used in any environment, but if you brick the machine, there is no recovering it (but cloud providers usualy have a button to re-image a base OS)
glines has joined #nixos
<eacameron>
Well this sucks: error: a ‘x86_64-linux’ is required to build ‘/nix/store/8zc4snq7pcha87nhbmdvfpnnn8660alx-users-groups.json.drv’, but I am a ‘x86_64-darwin’
<clever>
"This kinds of errors usually caused by gcc unnecesarrily add options like -fexception, -fPIC, -fstack-protetor-* for building kexec purgatory which runs in kernel mode."
<clever>
jeaye: ah, it might be some of the hardening stuff
<clever>
eacameron: nixops?
<eacameron>
clever: Yah. Apparently it can't build a user/groups json file on darwin (my host) for a vbox deploy??
<clever>
eacameron: yeah, it needs a linux build of "echo" to create files, its a bit silly
<eacameron>
clever: :/
<clever>
eacameron: if you already have a virtualbox with nixos, you should be able to make it a build slave
<eacameron>
clever: Yes I do. You've told me about this option 5 times now I still haven't actually used it because it scares me ;)
<clever>
lol
<clever>
why does it scare you?
<eacameron>
clever: I don't quite understand it...
<clever>
eacameron: you put a line like this into a text file, and any time it has to build something for "x86_64-linux" it will just "ssh clever@c2d.localnet" and use that to do the build
Rotaerk has joined #nixos
<clever>
eacameron: and it will use nix-copy-closure to push the build inputs out, and pull the result bac
<clever>
k
<NixOS_GitHub>
[nixpkgs] acowley opened pull request #20847: ignition: init messages, tools, and transport2 (master...ignition) https://git.io/v1naQ
angerman has joined #nixos
<eacameron>
clever: Ah so I need to have ssh keys installed for that?
<eacameron>
The other machines I have were also deployed with nixops so I would have to copy the ssh keys somehow
<clever>
eacameron: yep, in my case /etc/nixos/keys/distro is a private key without any pw
<clever>
which nix will use to get access to c2d
<clever>
the matching public has to be in the ~/.ssh/authorized_keys of c2d
<clever>
and the slave doesnt have to be nixos, i have 2 gentoo slaves, and 1 mac slave
eacamero_ has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/v1nVO
<NixOS_GitHub>
nixpkgs/master 5afc6b5 Tim Steinbach: linux: 4.1.35 -> 4.1.36
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to release-16.09: https://git.io/v1nVc
<NixOS_GitHub>
nixpkgs/release-16.09 0dcdb9b Tim Steinbach: linux: 4.1.35 -> 4.1.36...
eacameron has quit [(Ping timeout: 244 seconds)]
<clever>
jeaye: aha, the problem is in the build of kexec, not the kernel
<clever>
though it just hangs upon executing
<clever>
and the laptop even shut itself off
<jeaye>
hah
Wizek has joined #nixos
slack1256 has quit [(Remote host closed the connection)]
eacamero_ has quit [(Remote host closed the connection)]
<clever>
angerman: it will merge 2 attribute sets, overwriting upon collisions
<clever>
angerman: and it is a shallow merge, so { a={b=2;}; } and { a={c=3;}; d=4; } will overwrite the whole .a attribute
<angerman>
clever, I think the operators sections that comes later should come earlier, or a note should say that "//" is the merge operator, (see Operators below) or something.
georges-duperon has quit [(Ping timeout: 260 seconds)]
mkoenig has joined #nixos
danharaj has quit [(Ping timeout: 250 seconds)]
uralbash has joined #nixos
marusich has joined #nixos
ugjka has quit [(Ping timeout: 260 seconds)]
NEo4v8_ has quit [(Ping timeout: 268 seconds)]
ifur_ is now known as ifur
glines has quit [(Ping timeout: 260 seconds)]
<eacameron>
clever: WOW. It works. Distributed builds work. Floored.
<clever>
main limitation, is that it splits it up at a derivation level, so the entire package gets built on 1 machine
<clever>
but if you have 2 slaves, and try to build 10 packages at once, it will send 5 to each
<clever>
so things like gcc and the kernel wont go any faster, but it can do the kernel at the same time as other stuff
<clever>
and they wont compete for resources
AllanEspinosa has quit [(Ping timeout: 260 seconds)]
<eacameron>
clever: Heck I don't care if it goes slower right now. I can actually deploy which all that matters ATM ;)
<eacameron>
clever: Yah so VirtualBox assigned an IP to one deploy that had historically belonged to another deploy. They weren't both running at once. But once I tried to run the older one it broke. And now the new one can't be accessed.
<eacameron>
Because I changed the known_hosts.
<eacameron>
ATM it's no sweat: destroy/deploy but that's a pretty annoying "gotcha" I'll have to look out for.
<clever>
static IP's can prevent issues, but not sure how to mix those into nixops+vbox
<eacameron>
clever: Yah...I'm not sure how to tell nixops that it should use a different IP
<eacameron>
clever: I think I'll have to write a blog post about this soon. With a couple scripts it could be entirely automated to bootstrap a virtualbox buildmachine for nixops to use...by using nixops to build the bootstrap
lverns has quit [(Ping timeout: 260 seconds)]
mkoenig has quit [(Read error: Connection reset by peer)]
redmq has quit [(Ping timeout: 250 seconds)]
mkoenig has joined #nixos
mkoenig has quit [(Remote host closed the connection)]
jgertm has joined #nixos
7ITAAUKMB has joined #nixos
17WAAOKPN has joined #nixos
RayNbow`TU has joined #nixos
Wizek has quit [(Ping timeout: 265 seconds)]
proteusguy has quit [(Ping timeout: 250 seconds)]
marusich has quit [(Ping timeout: 260 seconds)]
zagy has joined #nixos
Hismione is now known as justanotheruser
slack1256 has quit [(Ping timeout: 250 seconds)]
<pikajude>
hey, I'm having some trouble with bower2nix, or more specifically fetchbower
<pikajude>
it doesn't actually seem to be fetching packages afaict
<pikajude>
it just produces a cache entry for the requested package
<pikajude>
can anyone try to verify?
amarsman has joined #nixos
newhoggy2 has joined #nixos
7ITAAUKMB has quit [(Ping timeout: 250 seconds)]
17WAAOKPN has quit [(Ping timeout: 265 seconds)]
marusich has joined #nixos
zraexy has quit [(Ping timeout: 248 seconds)]
newhoggy2 has quit [(Remote host closed the connection)]
<jophish>
Is it going to be a problem if I run nix-build inside of buildPhase
<jophish>
The result shouldn't depend on the result of that derivation
<clever>
jophish: when the sandbox is doing its job, that wont work
<jophish>
clever: sandbox builds are out of the window anyway here I'm afraid
<jophish>
:D
jacob_ has joined #nixos
<jophish>
This is running some tests on an external bit of hardware, but to build the software for this hardware I'd like to use nix
<jophish>
I wonder if I can split it up into two stages
jacob_ is now known as Guest92738
<clever>
splitting it up is best
<jophish>
I think for now I'll keep this program calling nix, but not run this program from inside the derivation
<jophish>
and have the user run these tests manually
<jophish>
they take hours anyway, so shouldn't go on hydra as is
<eacameron>
clever: In my case, as you saw earlier probably, it was trying to create users for the target but for some reason (echo?) it can't do that on macOS...
<clever>
eacameron: i believe the user-groups.json eventualy gets written to disk in this derivation
<clever>
which uses the linux builds of mkdir, mv, and bash
<clever>
it doesnt have the ability to throw in a different arch for such basic tasks
<clever>
and due to how the rest of nix works, everything depending on this file would have to re-compile, so mac users cant make use of hydra builds (because hydra happened to use linux, and mac users want to use mac)
<lonokhov>
ok, found `substring 0 1`
<eacameron>
clever: Yah and that's not a problem as long as the target works...but for things like user config it doesn't work (unless you already deployed earlier with a different setup)
<clever>
eacameron: one key part though, to deploy, you must be able to ssh into the machine
<clever>
and being able to ssh into it, is also what you need to make it a build slave
<eacameron>
clever: Right. Which is why But I've found that I can have nixops just deploy an "empty" vbox server (only has sshd) and then use your trick to make it a slave.
ebzzry has quit [(Ping timeout: 245 seconds)]
<clever>
thats one way to fix things
<eacameron>
clever: :D
<eacameron>
I consider myself very clever
<eacameron>
clever: but certainly not as clever as you
<clever>
when the correct nix-build command is ran, it will create a bash script, you can then nix-copy-closure that script to a target server, and run it
<clever>
the current OS will immediately die, and nixos will rise from the ashes
<eacameron>
clever: ... clients don't grow on trees.
<jophish>
That's the recommended way to get buildPhase to not buffer stdout quite as much?
<clever>
i wound up taking over for the client guy, and redid the entire thing in QT
<clever>
and QT has support for windows, linux, mac, ios, android, windows phone, and blackberry
<eacameron>
clever: wow...NICE
<clever>
the ios stuff was a bit tricky, apple does not allow openssl in the apps
<clever>
you must use the ssl that they provide
<clever>
so i had to use an alpha version of QT that implemented QSslSocket on the apply ssl libs
<clever>
which also returned different error codes and made new bugs
chattered has joined #nixos
<eacameron>
clever: anytime you support that many devices your bound to wind up in junk like that.
<clever>
yeah
<clever>
another major problem, the QT android stuff, behaves more like QT desktop
<clever>
so instead of scrolling by flinging the list up/down, you have to drag the scroll bar, in the oposite direction (exactly like you do on a desktop)
<signal4>
as far as i understand i only changed (NIX_)CFLAGS to include "-march=native -mtune=native", everything else remains the same. and yet it can't build the derivation for some reason
FRidh has joined #nixos
takle has quit [(Remote host closed the connection)]
takle has joined #nixos
eacameron has joined #nixos
johann__ has quit [(Quit: Leaving.)]
takle has quit [(Ping timeout: 265 seconds)]
aboseley has joined #nixos
aboseley_ has joined #nixos
aboseley has quit [(Client Quit)]
jgertm_ has quit [(Ping timeout: 240 seconds)]
eacameron has quit [(Ping timeout: 265 seconds)]
johann__ has joined #nixos
derjohn_mobi has quit [(Ping timeout: 260 seconds)]
eacameron has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh force-pushed python-wip from 7a49b43 to 7f80872: https://git.io/vzaOS
<clever>
ah, and that explains why i couldnt recreate the issue with dry-run, only switch/boot/test run that
bin8me has quit [(Ping timeout: 250 seconds)]
<signal4>
clever: o/ do you know what "cycle detected in the references of..." error means? I followed your advice about stdenv but now i get that error
<clever>
signal4: you need to grab the 2 outputs it mentions, for the other output, and find out why there is a cycle
<clever>
viric: i think part of the problem is in git, when you ask it what files have been modified, it updates the .git/index to speed up future operations
johann__ has quit [(Quit: Leaving.)]
<viric>
clever: maybe git can be told not to update the index with some flag
<viric>
clever: moreover! I have the file with mode 600, so root cannot write it - then git overrides the ownership
Guest92738 has quit [(Ping timeout: 260 seconds)]
jacob_ has joined #nixos
roxma has quit [(Ping timeout: 246 seconds)]
oleks has quit [(Ping timeout: 246 seconds)]
oleks has joined #nixos
<clever>
viric: some programs dont open the file for writing, because thats not atomic
jacob_ is now known as Guest69519
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to master: https://git.io/v1cvc
<NixOS_GitHub>
nixpkgs/master da3d8d0 Frederik Rietdijk: pythonPackages.django_guardian: add missing pytest
<clever>
viric: so they typicaly open a file like index.tmp, write to it, then use the rename() syscall to atomicly update it
<viric>
ah maybe that
<viric>
makes sense, because I tried the permissions thing before without success
roxma has joined #nixos
<clever>
removing root's write to the directory may stop that
<viric>
root can't write to it
<viric>
it's owned by viric.viric
<clever>
but root also has the magic ability to just ignore chmod values entirely
<viric>
well, it obviously can
<viric>
exactly
<clever>
root is all powerfull
<clever>
it may also help to read the git source code
<clever>
and its getting pretty late here, i should get off to bed
<viric>
damn it...
<clever>
good luck :)
<viric>
I'm doomed :)
<viric>
good night, clever
<the-kenny>
Anyone have an idea why my synaptics touchpad (on a lenovo x260) is in absolute mode?
<clever>
the-kenny: mine is doing the same thing, i have caused that problem before when i was messing with custom kernel builds and turned on a synaptics driver, but i havent looked into fixing nixos yet
<clever>
dell d630 in my case
<clever>
and i'm off to bed, goodnight
<the-kenny>
clever: hm, but I haven't even touched the touchpad. I had it disabled completely until now
<the-kenny>
Maybe I need to restart X with synaptics.enable = true;
johann__ has joined #nixos
<the-kenny>
okay that did it :)
hlavaty has quit [(Read error: Connection reset by peer)]
johann__ has quit [(Quit: Leaving.)]
phanimahesh has joined #nixos
johann__ has joined #nixos
civodul has joined #nixos
eacameron has joined #nixos
derjohn_mobi has joined #nixos
eacameron has quit [(Ping timeout: 260 seconds)]
Itkovian_ has joined #nixos
pstn_ has quit [(Quit: leaving)]
Itkovian has quit [(Ping timeout: 250 seconds)]
georges-duperon has joined #nixos
christine_ has quit [(Ping timeout: 245 seconds)]
johann__ has quit [(Quit: Leaving.)]
arjen-jonathan has joined #nixos
LysergicDreams has quit [(Ping timeout: 256 seconds)]
<arjen-jonathan>
Hi all
NEo4v8_ has quit [(Ping timeout: 258 seconds)]
<signal4>
hi arjen-jonathan
LysergicDreams has joined #nixos
<NixOS_GitHub>
[nixpkgs] proteansec opened pull request #20854: rox-filer: init at 2.11 (master...pkgs/rox-filer) https://git.io/v1cIA
edvorg has joined #nixos
<goibhniu>
hrm, both screen and tmux fail to start in a fresh 16.09 container
<goibhniu>
I found some things online about dev/pts configuration on other distros, but it looks the same (via `mount`) as in an older working container
kier has quit [(Remote host closed the connection)]
newhoggy2 has quit [(Remote host closed the connection)]
phanimahesh has quit [(Ping timeout: 240 seconds)]
phanimahesh has joined #nixos
bigvalen has left #nixos ["Leaving"]
Shou has joined #nixos
Shoue has joined #nixos
herzmeister has quit [(Quit: Leaving)]
herzmeister has joined #nixos
Shou has quit [(Ping timeout: 248 seconds)]
jensens has quit [(Ping timeout: 250 seconds)]
<goibhniu>
FTR other weird things were broken in this container too (I can't see a pattern), e.g. installing a python package in a virtualenv failed (complaining about absolute paths in setup.py!?), luckily switching to 16.03 works!
Shoue has quit [(Ping timeout: 245 seconds)]
newhoggy2 has joined #nixos
stepcut has joined #nixos
zagy has quit [(Quit: Leaving.)]
zagy has joined #nixos
newhoggy2 has quit [(Ping timeout: 260 seconds)]
wkennington has quit [(Read error: Connection reset by peer)]
stepcut has quit [(Ping timeout: 258 seconds)]
edvorg has quit [(Ping timeout: 256 seconds)]
<LnL>
goibhniu: how did you create it?
<goibhniu>
imperatively, with nixos-container create
<LnL>
do you see the same if create a new container when 16.09 is active?
<goibhniu>
I noticed there was also a fairly recent change which stopped openvpn working from my containers, not sure if that's related
<goibhniu>
sorry, I don't understand, do you mean when the host system is using 16.09? ... my host system is on nixos-unstable
<LnL>
I have a feeling that the container did not upgrade
<goibhniu>
ah, so my procedure was to create the container, copy over configuration.nix, add the 16.09 channel, use nixos-container root-login to get into the container and then run nixos-rebuild switch --upgrade
<goibhniu>
and now, I've changed the channel to 16.03, and run nixos-rebuild switch --upgrade to switch the container to 16.03
<goibhniu>
I hope that makes sense
<goibhniu>
I've also upgraded the 16.09 container (there was a new channel since I created it yesterday)
<goibhniu>
nixos-rebuild switch completed without errors
<NixOS_GitHub>
[nixpkgs] viric pushed 1 new commit to master: https://git.io/v1cCo
<NixOS_GitHub>
nixpkgs/master 19fd547 Lluís Batlle i Rossell: Fixing the patch for icu for ARM....
<LnL>
goibhniu: did you run nixos-container update after the system upgrade?
<goibhniu>
ah, nope
<goibhniu>
I've never used nixos-container update actually
<LnL>
that's what I was talking about
<LnL>
you are essentially trying to run a 16.03 container in a 16.09 system
<goibhniu>
yeah, that's the idea ... well, I wanted to run a 16.09 container in a nixos-unstable system
<goibhniu>
running a 16.03 container in a nixos-unstable system works great
<goibhniu>
does that sound like trouble?
Shou has joined #nixos
* goibhniu
hopes this is actually supported, and didn't just work by accident until now
<LnL>
nixos conainers are very light weight, I've seen issues before if your host system diverges to much
<LnL>
but I'm not sure if this is supposed to work or not
<goibhniu>
ok, thanks for the pointer
<goibhniu>
I'll also have a closer look at the change mentioned in the 16.09 release notes:
<goibhniu>
Special filesystems, like /proc, /run and others, now have the same mount options as recommended by systemd and are unified across different places in NixOS. Mount options are updated during nixos-rebuild switch if possible. One benefit from this is improved security — most such filesystems are now mounted with noexec, nodev and/or nosuid options.
<goibhniu>
at least it sounds similar to reports I've read about screen not working on other distros
kier has joined #nixos
<christine_>
If I'm just using the Nix package manager and not NixOS: Is there a way to enable and configure a database?
<goibhniu>
Sorry christine_, I don't think there's a convenient way to do that
<goibhniu>
hrm, maybe you can use nix to create a docker container, or something like that
pi3r has quit [(Ping timeout: 260 seconds)]
<christine_>
goibhniu: yeah, Docker containers are very convenient for development if you quickly need a database or multiple ones in multiple versions. Just hoped there would be a Nix way to do the same.
<msd>
For example: I want a repo checked out with tag version "OTP_R16B02_basho8" . Is there a way to pipe that straight into fetchFromGitHub? Or must I use the rev?
<LnL>
fetchFromGitHub does not use git, it downloads a tarball
<msd>
jophish: now I see my mistake :P
<jophish>
:)
teoma has joined #nixos
<avn>
msd: for fetchgit you need rev="refs/tags/$tagname";
<msd>
avn: for packages, is fetchgit preferred over fetchFromGitHub?
<msd>
Is either preferred over the other?
<peti>
fetchFromGitHub is preferrable since it doesn't depend on git.
amyers has joined #nixos
<teoma>
After upgrading to 16.09, xinit gives this error: Couldn't open libEGL.so.1: libEGL.so.1: cannot open shared object file: No such file or directory
<teoma>
How would I go about fixing this?
zagy has quit [(Quit: Leaving.)]
<teoma>
Xorg.0.log says it happens while loading sub module glamoregl.
posco_ has joined #nixos
<msd>
peti: alright, switching over some packages to fetchFromGitHub
<msd>
peti: thanks
bin7me has joined #nixos
Mercuria1Alchemi has quit [(Ping timeout: 246 seconds)]
<msd>
LnL: could I please get a review on this PR (https://github.com/NixOS/nixpkgs/pull/20860)? Just a quick upgrade, and you have the background context for this package/why it's necessary for Riak CS/Stanchion.
posco_ has quit [(Ping timeout: 268 seconds)]
<LnL>
sure, I'll take a look when I get home
johann__ has quit [(Quit: Leaving.)]
<teoma>
Looks like libEGL.so.1 is produced by mesa_noglu, so I guess mesa_noglu should be added to buildInputs of some X server related package -- any idea which package?
posco__ has quit [(Ping timeout: 243 seconds)]
posco____ has joined #nixos
johann__ has joined #nixos
Damas has joined #nixos
posco___ has quit [(Ping timeout: 266 seconds)]
posco_____ has joined #nixos
<Damas>
hey everyone? Is there an idiot install guide for OSX with Fish? I would like to use also Nix versions of bash and fish. But the part with setting paths + fish is kinda confusing
uralbash has quit [(Quit: Leaving.)]
jensens has quit [(Ping timeout: 265 seconds)]
<msd>
LnL: thanks
<LnL>
Damas: I don't think our installer supports fish
<Damas>
well, what about zsh?
<Damas>
LnL: the way, that terminal will be fired with bash version installed by Nix - is it possible? Last time I tried I broke OSX
posco______ has joined #nixos
posco____ has quit [(Ping timeout: 268 seconds)]
<LnL>
you changed your login shell to the one from nix?
johann__ has quit [(Quit: Leaving.)]
<Damas>
not the login shell, just the one behind Terminal app (was easy to fix, but still)
<Damas>
just wanted Terminal to use as a default bash4 from Nix
acertainkind has quit [(Quit: Leaving)]
posco_______ has joined #nixos
posco_____ has quit [(Ping timeout: 268 seconds)]
georges-duperon has quit [(Ping timeout: 260 seconds)]
amyers has quit [(Ping timeout: 250 seconds)]
<teoma>
Apparently the epoxy library "picks the OpenGL library to load dynamically", but adding mesa to propagatedBuildInputs of epoxy triggers recompiling tons of packages including libreoffice. Does this look like the right thing to try?
katyucha_ has joined #nixos
katyucha1 has joined #nixos
Shou has quit [(Ping timeout: 265 seconds)]
derjohn_mobi has quit [(Ping timeout: 256 seconds)]
posco______ has quit [(Ping timeout: 264 seconds)]
posco________ has joined #nixos
posco_______ has quit [(Ping timeout: 273 seconds)]
posco________ is now known as posco_______
ryanartecona has joined #nixos
danielrf has joined #nixos
amyers has joined #nixos
nh2_ has joined #nixos
<LnL>
Damas: if you want to do that you should use /nix/var/nix/profiles/default/bin/<shell>
<LnL>
that way it won't break if you update the version
<LnL>
(unless you remove it from your profile)
Damas has quit [(Quit: Page closed)]
johann__ has joined #nixos
katyucha1 has quit [(Ping timeout: 250 seconds)]
katyucha_ has quit [(Ping timeout: 245 seconds)]
damas has joined #nixos
bin8me_ has quit [(Remote host closed the connection)]
roconnor has joined #nixos
pi3r has quit [(Ping timeout: 250 seconds)]
nh2_ has quit [(Quit: Leaving.)]
cpennington has joined #nixos
roconnor has quit [(Ping timeout: 246 seconds)]
jgertm_ has quit [(Read error: Connection reset by peer)]
jgertm__ has joined #nixos
digitalmentat has joined #nixos
mde_ has joined #nixos
<digitalmentat>
I've put a fresh install of NixOS onto my laptop but installing KDE5 is giving me a collision with phonon-backend-gstreamer-qt and the qt4 version of the same package; has anyone run into this?
glines has quit [(Ping timeout: 265 seconds)]
<digitalmentat>
I've gone through my configuration.nix mulitple times to make sure I'm not by accident installing something else that might depend on the older package but that is to no avail
<goibhniu>
can you post your configuration.nix?
<mde_>
hello guys, i have ntfs partitions and i cant copy anything from nixos to ntfs partitions, what i need to do ? i ve installed ntfs3g
<digitalmentat>
it's a little tough to copy it over right now since it's on the other laptop
<digitalmentat>
it's a *very* basic config
<digitalmentat>
like, the stock config + xserver + a new user + kde5 + networkmanager + a few editor tools and that's it
<digitalmentat>
how do I figure out what package is causing the phonon qt4 dependency?
<digitalmentat>
err "has"
damas has quit [(Quit: Page closed)]
<goibhniu>
digitalmentat: how are you installing kde5?
<goibhniu>
can you post the error? ... maybe use sprunge or something like that
<digitalmentat>
goibhniu, (thanks for helping me) I have this line in my configuration.nix: `services.xserver.desktopManager.kde5.enable = true;`
<digitalmentat>
`collision between `/nix/store/<hash>-phonon-backend-gstreamer-qt5-4.9.0/share/icons/hicolor/<resolution>/apps/phonon-gstreamer.png` and `/nix/store/<hash>-phonon-backend-gstreamer...`
<digitalmentat>
also - other laptop is booted up now into KDE5, nice and smooth
<goibhniu>
nice
amarsman has joined #nixos
christine_ has quit [(Ping timeout: 260 seconds)]
jgertm__ has quit [(Ping timeout: 240 seconds)]
seanz has joined #nixos
Kamil__ has quit [(Quit: Page closed)]
digitalmentat has quit [(Ping timeout: 250 seconds)]
magnetophon has joined #nixos
teoma has quit [(Remote host closed the connection)]
johnsonav has quit [(Ping timeout: 240 seconds)]
georges-duperon has quit [(Ping timeout: 260 seconds)]
ambro718 has joined #nixos
Judson1 has quit [(Ping timeout: 265 seconds)]
fritschy has quit [(Quit: Leaving.)]
johnsonav has joined #nixos
<magnetophon>
Hi. yesterday I spent a lot of time trying to install nixos on my parents imac, but was not able to run X. I just finished installing debian on it, and everything works now. I'd prefer nixos though. Can the debian installation give me cues as to what to do different on nixos? Ralith, avn.
<Ralith>
interesting!
ryanartecona has quit [(Quit: ryanartecona)]
<Ralith>
check the kernel and video driver and versions thereof, and maybe see if debian has relevant patches in their versions of those
<Ralith>
or just run nix on debian and call it a day
Shou has joined #nixos
georges-duperon has joined #nixos
<magnetophon>
Ralith: nix on debian doesn't help me much. my parents don't need multiple versions of programs or anything. In this case I want nixos cause it's a rolling release with rollback.
<magnetophon>
how do I check the videodriver?
<magnetophon>
nm I can google that
sigmundv_ has joined #nixos
<Ralith>
if it's for nontechnical users, do they even need rolling release?
<magnetophon>
Ralith no they don't, but it would save me a reinstall in a few years. no biggie
noexcept_ has joined #nixos
<magnetophon>
All they need is a browser and a VM for their bridge game. Probably the wise thing to do is to call it a day...
Dezgeg has quit [(Ping timeout: 250 seconds)]
civodul has quit [(Quit: ERC (IRC client for Emacs 25.1.1))]
<noexcept_>
hi, can anyone comment on nix vs guix? I like the idea of using Guile for the package definitions but I would prefer if GuixSD used systemd instead of some relatively unknown init system
<simpson>
noexcept_: I use NixOS because I want nixops.
<Ralith>
I use NixOS because it's more active
__Sander__ has quit [(Quit: Konversation terminated!)]
<noexcept_>
also I currently use Gentoo and the (broken) dependency resolution is driving me crazy
<magnetophon>
noexcept_: I use nixos because there's more pkgs
<Ralith>
if I'd known guix didn't use systemd I'd use NixOS for that reason too though
<Ralith>
systemd is the best
<noexcept_>
like portage (the gentoo package manager) will take 5-10 minutes to resolve dependencies for a system upgrade and still require me to manually resolve dependencies from use flags and generally takes a lot of tinkering to solve conflicts
<Ralith>
o_O
<Ralith>
yeah there's none of that
<Ralith>
the package search is a little slow at worst
ryanartecona has joined #nixos
schneefux has quit [(Quit: ZNC 1.6.3 - http://znc.in)]
Judson has joined #nixos
<Judson>
Is there a tidy way to determine the current "ld-linux.so"?
<Judson>
i.e. I want to be able to use patchelf to fix downloaded exes without needing to add them to an overlay necessarily
schneefux has joined #nixos
georges-duperon has quit [(Ping timeout: 260 seconds)]
Itkovian_ has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<noexcept_>
Ralith: the only thing that's worrying me a bit about NixOS is that there seems to be a lot of focus on making everything functional even when that does not necessarily make sense. The reason NixOS looks appealing to me is because every package is slotted and there are no conflicts, so dependency resolution is polynomial-time and not NP-hard.
rly has quit [(Ping timeout: 246 seconds)]
<gchristensen>
noexcept_: how do you mean?
<Ralith>
well, don't worry, it turns out to make sense
<noexcept_>
Ralith: and not because they use monads to express mutable computations as an immutable program description
<noexcept_>
I don't know if this is really the case for NixOS
<noexcept_>
just an impression I get from the description on the home page :)
<simpson>
noexcept_: The Nix language doesn't have monads. I'm not sure what you're getting at.
<noexcept_>
simpson: "Purely Functional Linux Distribution" makes it sound like it's a project by some haskell zealots. Again, I may be getting the wrong impression :)
<magnetophon>
noexcept_: a pkg can only run after it has been "nixified", if that's what you mean. That is the biggest advantage and disadvantage at the same time, imho.
<simpson>
noexcept_: Thankfully, "purely functional" doesn't mean anything. NixOS is just a distro that uses Nix as its package manager. Nix itself enforces purity in builds.
roconnor_ has quit [(Ping timeout: 248 seconds)]
eacameron has joined #nixos
<noexcept_>
simpson: If all that is meant by purity here is that each version of a package gets put into its own (immutable) directory, then that's perfect. :)
<simpson>
noexcept_: Slightly stronger: A package's output artifacts are parameterized solely on the build inputs.
<kmicu>
noexcept_: taking into account your requirements you could use NixOS (systemd) as an OS and Guix with Guile as a package manager.
octalsrc has joined #nixos
<noexcept_>
kmicu: interesting
MarcelineVQ has quit [(Ping timeout: 265 seconds)]
<octalsrc>
has anyone used the beamer package in texlive recently?
<octalsrc>
"texlive.combine { inherit (texlive) beamer; }" fails to build
<octalsrc>
on nixos-unstable
<noexcept_>
Do I understand it correctly that the reason it's impractical to install nix entirely without root is because the /nix/store prefix is hardcoded into the checksums?
<simpson>
noexcept_: The paths are also baked into the artifacts, so they're not relocatable.
<noexcept_>
if yes, why is there not a way to specify a custom prefix that is not included in checksum calculation?
<octalsrc>
am I using it wrong or should I report an issue?
<shlevy>
You can in fact do it in another store path and most of nixpkgs should work
<shlevy>
But you don't get any binaries
<shlevy>
the store should be included in the checksum
<shlevy>
as it affects the results, significantly
<shlevy>
as simpson says they're not reolcatable
<noexcept_>
ah, I assume that's because of hard coded runtime linker paths, etc?
newhoggy_ has joined #nixos
<kmicu>
octalsrc: what’s the error message? ‘nix-shell --pure -p "(texlive.combine { inherit (texlive) scheme-basic beamer; })"’ works correctly.
<shlevy>
noexcept_: Yeah
<octalsrc>
ah I wasn't including scheme-basic
<octalsrc>
that seems to make the difference
newhoggy2 has quit [(Ping timeout: 265 seconds)]
<octalsrc>
without it I get the error: "/nix/store/1rfr9q6xly91i2jjsf3s3lrl39a7a8yh-stdenv/setup: line 864: cd: /nix/store/ml5n8mx9msg5q5mgj31kqz35zg2za5fd-texlive-combined-2016/share/texmf/scripts/texlive: No such file or directory"
georges-duperon has joined #nixos
<octalsrc>
thanks! should scheme-basic more or less always be included?
<kmicu>
What packages are required depends on your *.tex file.
<kmicu>
(You could use scheme-full to install almost all TexLive packages.)
<octalsrc>
yes, I was trying to avoid that :P
eacamero_ has joined #nixos
jgertm__ has joined #nixos
eacameron has quit [(Ping timeout: 245 seconds)]
evangeline has joined #nixos
eacamero_ has quit [(Remote host closed the connection)]
eacameron has joined #nixos
<evangeline>
hi, I've added a new package to my own nixpkgs and installed it with nix-env -i while testing; after everything was working I removed it with nix-env -e, but after redoing nix-env -i the package is installed instantly, so it must be somewhere in the cache; is there a flag to nix-env, which would force the rebuilding of the package ?
<simpson>
evangeline: Why would you need a rebuild?
<simpson>
(You can remove the package from the store with nix-store.)
<evangeline>
simpson: to test if everything still works, since I've updated the package definition ...
<simpson>
evangeline: Huh, but the hashes didn't change? That's a little strange
<evangeline>
I've modified the license field.
<kmicu>
You could change anything (not meta) in package’s recipe or remove the package from the store with ‘nix-store --delete’ command or ‘nix-collect-garbage’.
newhoggy2 has joined #nixos
<magnetophon>
evangeline: the package was not cashed, it was just never actually removed from the store. installing/uninstalling is separate from actualy removing it fro disk. that's a feature, not aq bug. :)
<kmicu>
Sometimes you can lost build logs and you want to rebuild the same thing with the same hash.
<rardiol>
can't "nix-store --read-log" solve that problem?
newhoggy_ has quit [(Ping timeout: 248 seconds)]
iyzsong has quit [(Quit: bye.)]
<evangeline>
simpson: magnetophon kmicu hmm, the package is not being recompiled and I've ran both commands: ‘nix-store --delete’ command or ‘nix-collect-garbage’ command.
<evangeline>
I'm running this command: nix-env -f $NIXPKGS -i rox-filer
<kmicu>
Oh I’m sorry, those are not exact commands, only pointers.
MVQq has joined #nixos
<magnetophon>
evangeline: are you sure the pkg was not installed at the time you ran nix-collect-garbage?
<evangeline>
yeah, it wasn't, I did: 'nix-env -f $NIXPKGS -e rox-filer' prior to that; the 'nix-collect-garbage -d' also doesn't solve this
jgrasser has quit [(Ping timeout: 245 seconds)]
msd has quit [()]
<evangeline>
after doing the 'nix-env -f $NIXPKGS -e rox-filer' the binary is still in nix store, so I guess the packet wasn't removed at all ...
<kmicu>
You could try ‘nix-store --delete /nix/store/*-rox-filer*’
<evangeline>
error: cannot delete path ‘/nix/store/1512wjx8vnbcxyf71yfmp62c7jm7fcn8-rox-filer-2.11’ since it is still alive
<magnetophon>
evangeline: mayb it's still installed in another profile?
pi3r has joined #nixos
<magnetophon>
or in sytem pkgs?
jgertm__ has quit [(Ping timeout: 240 seconds)]
<evangeline>
magnetophon: it's not in the system and it shouldn't be installed as another profile, since I'm only using the system as my current user.
<simpson>
evangeline: I bet that your user profile has a generation with it.
<evangeline>
simpson: how can I verify that?
redmq has joined #nixos
<kmicu>
evangeline: could you check with ‘nix-store --gc --print-roots’ if there is a ‘.result’ symlink somewhere?
<NixOS_GitHub>
[nixpkgs] montag451 opened pull request #20862: gnome-control-center: provides access to the tracker gsettings schemas (master...gnome-control-center-tracker) https://git.io/v1CIl
Seichi_ has joined #nixos
Seichi has quit [(Ping timeout: 250 seconds)]
<kmicu>
(You can safely remove any symlink from ‘/somewhere/result -> /nix/store/HASH-rox-filler’ and after that try ‘nix-store --delete …’ command on rox-filler once again.)
<evangeline>
kmicu: there is a result yeah; how can I remove that?
<kmicu>
Plain old ‘rm /path/to/result’ is ok.
Seichi_ has quit [(Client Quit)]
danielrf has quit [(Quit: WeeChat 1.6)]
<evangeline>
kmicu: I'm still getting the same error about the package being active, but I've remove the result symlink
<evangeline>
however, now I'm getting this about a derivation:
<evangeline>
error: cannot delete path ‘/nix/store/6jrl2w5lnd1qw9rnasym548r0pg0lxw9-rox-filer-2.11.tar.bz2.drv’ since it is still alive
<kmicu>
Could you check that command linked by simpson: ‘nix-store -q --roots /nix/store/1512wjx8vnbcxyf71yfmp62c7jm7fcn8-rox-filer-2.11’?
<evangeline>
kmicu: sure, it returns empty result and status code 0
Shou has quit [(Ping timeout: 250 seconds)]
<kmicu>
Hmm, maybe it will be easier if you go to rox-filer’s recipe and change sequence of buildInputs’ entries. It will trigger the rebuild and give you the same package.
<evangeline>
kmicu: that does it; so if I change the meta tags, the hash of the package won't change meaning it won't be reinstalled?
<kmicu>
Or we could investigate further with ‘nix-store -q --references …’. I’m not sure what could keep it alive and it ‘sucks’ that ‘nix-store --delete …’ does not print that info ( ͡~ ͜ʖ ͡°)
LysergicDreams has quit [(Ping timeout: 260 seconds)]
<kmicu>
Iirc, yep, meta.* is ignored.
<evangeline>
kmicu: what do I need to pass to nix-store -q --references ?
<kmicu>
evangeline: could you do the same for ‘nix-store -q --referers …’?
<noexcept_>
how often does the configuration.nix
<noexcept_>
how often does the configuration.nix format change and does nix warn about outdated entries?
<kmicu>
configuration.nix is a function and Nix, the expression language, has almost no changes. Some packages have a proper deprecation warning, but some of them not; depends on maintainers.
<magnetophon>
kmicu: interesting. coud you point me to a pkg that has such a warning?
<evangeline>
those changes are not yet in the nixpkgs-17.03pre96585.fa4212e/nixexprs.tar.xz archive, so I probably have to pull directly from github. I have my own nixpkgs on developement machine, but this is server machine! What would be the best way to pull in those changes?
pi3r has quit [(Ping timeout: 265 seconds)]
<kmicu>
Watch out for ‘nixpkgs-unstable’ as NixOS channel, b/c it can brick NixOS. You should use ‘nixos-unstable’ for NixOS’ unstable channel.
<gchristensen>
kmicu: I wonder if we could add a check in there somehow
<kmicu>
evangeline: ‘kube2sky’ is in nixos-unstable.
<kmicu>
evangeline: ‘sudo nix-channel --add https://nixos.org/channels/nixos-unstable nixos’, ‘sudo nix-channel --update’ and ‘sudo nixos-rebuild boot’, and finally a reboot if you want to switch system to the nixos-unstable channel.
<kmicu>
‘nixos.’ prefix if you are on NixOS and ‘nixpkgs.’ prefix if you are on a standalone Nix (or ‘-f '<nixpkgs>"’ to avoid prefixing). ⸮ Idiosyncrasies FTW!
<kmicu>
(‘myphpstorm’ is an ‘attribute path’ and you need to use ‘nix-env -iA …’ to install it.)
tv has quit [(Ping timeout: 245 seconds)]
eacamero_ has joined #nixos
<kmicu>
evangeline: I’m sorry, but I’ve misread your situation. ‘kube2sky’ is available in nixos-16.09. It was removed from nixos-unstable or nixpkgs-unstable.
eacameron has quit [(Read error: Connection reset by peer)]
evangeline has quit [(Quit: ZNC 1.6.3 - http://znc.in)]
georges-duperon has quit [(Quit: Leaving)]
<kmicu>
(5bc7ae7 ‘kubernetes module: support for kubernetes 1.4’ redesigned that module and dropped ‘kube2sky’ in favor of ‘dns’ suboption.)
markus1189 has quit [(Remote host closed the connection)]
agjacome has joined #nixos
<sphalerite>
why does `nix-env -q --profile /run/current-system` not output anything?
ryantm____ has quit [(Remote host closed the connection)]
cstrahan has quit [(Remote host closed the connection)]
monad_cat has quit [(Remote host closed the connection)]
SovereignBleak has quit [(Remote host closed the connection)]
jmeredith has quit [(Remote host closed the connection)]
sjanssen has quit [(Remote host closed the connection)]
jml has quit [(Remote host closed the connection)]
ericbmerritt_ has quit [(Remote host closed the connection)]
manveru has quit [(Remote host closed the connection)]
ocharles_ has quit [(Remote host closed the connection)]
mbrock has quit [(Remote host closed the connection)]
<kmicu>
sphalerite: ‘/run/current-system’ is not a profile. Try e.g. ‘nix-env -q --profile ~/.nix-profile’
<sphalerite>
so what is /run/current-system, and how can I list the packages linked into it? Or is that a question that doesn't make sense for whatever /run/current-system is?
markus1189 has quit [(Remote host closed the connection)]
seanz has quit [(Quit: Leaving.)]
ryanartecona has quit [(Quit: ryanartecona)]
<dtzWill>
sphalerite: well something like 'nix-store -qR /run/current-system' lists the derivations
ryanartecona has joined #nixos
manveru has joined #nixos
ocharles_ has joined #nixos
cstrahan has joined #nixos
MarcelineVQ has quit [(Ping timeout: 250 seconds)]
sjanssen has joined #nixos
ixxie has joined #nixos
kiki9922 has quit [(Read error: No route to host)]
ryantm____ has joined #nixos
<dtzWill>
sphalerite: bit of a kludge, but the wiki suggests the following (or something like it): nix-store -q --references /var/run/current-system/sw | cut -d'-' -f2-
<sphalerite>
ah yes, that looks like what I wanted
<sphalerite>
Thanks!
<dtzWill>
\o/
ericbmerritt_ has joined #nixos
Itkovian has joined #nixos
mbrock has joined #nixos
mguentner has quit [(Ping timeout: 250 seconds)]
<NixOS_GitHub>
[nixpkgs] vbgl opened pull request #20867: omake: update to 0.10.1 (master...omake-0.10.1) https://git.io/v1CEk
MarcelineVQ has joined #nixos
AllanEspinosa has joined #nixos
redmq has quit [(Ping timeout: 246 seconds)]
SovereignBleak has joined #nixos
monad_cat has joined #nixos
jml has joined #nixos
<NixOS_GitHub>
[nixpkgs] viric pushed 1 new commit to release-16.09: https://git.io/v1Czo
<NixOS_GitHub>
nixpkgs/release-16.09 7015964 Lluís Batlle i Rossell: Fixing the patch for icu for ARM....
seanz has joined #nixos
jmeredith has joined #nixos
markus1189 has joined #nixos
mguentner has joined #nixos
seanz has quit [(Ping timeout: 260 seconds)]
redmq has joined #nixos
<viric>
I get a segfault reported in dmesg if I run "upower -d"
<viric>
[ 8110.233653] upowerd[13764]: segfault at 20 ip 00007f2ed634e350 sp 00007ffc3d00b258 error 4 in libglib-2.0.so.0.4800.2[7f2ed62ca000+10d000]
mguentner has quit [(Ping timeout: 246 seconds)]
lverns has joined #nixos
<gchristensen>
viric: hrmm nice
<gchristensen>
I think icu_54 is broken too :(
<gchristensen>
:o
danielrf has joined #nixos
seanz has joined #nixos
mguentner has joined #nixos
newhoggy2 has quit [(Remote host closed the connection)]
akaWolf has joined #nixos
newhoggy2 has joined #nixos
<LnL>
is there some way to get a passhru variable into the output of a derivation
<LnL>
I want need the path of the buildEnv that the derivation will end up in
ixxie has quit [(Ping timeout: 250 seconds)]
slack1256 has joined #nixos
civodul has joined #nixos
<michalrus>
Hey, how can I set a profile picture for a regular user in GNOME?
Guest69519 has quit [(Remote host closed the connection)]
AllanEspinosa has quit [(Ping timeout: 260 seconds)]
amyers has quit [(Ping timeout: 260 seconds)]
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
evangeline has joined #nixos
<evangeline>
hi, I turned to nixos-unstable branch, which contains the latest version of kubernetes, but while building it with "nixos-rebuild switch", the build failed: http://dpaste.com/3AANFEK : does anybody know how to go about it? I realize I'm running the unstable branch and this is bound to happen, but even so there are many people out there running nixos on unstable branch.
goibhniu has joined #nixos
<clever>
evangeline try running it again with -j1 and see what happens
evangeline has quit [(Quit: Quit)]
aboseley_ has quit [(Remote host closed the connection)]
<Ralith>
wow, that looks really broken
AllanEspinosa has joined #nixos
<NixOS_GitHub>
[nixpkgs] e-user opened pull request #20868: sisco.lv2: init at 0.7.0 (master...pkgs/sisco.lv2) https://git.io/v1C6e
glines has joined #nixos
<glines>
clever: my teensyduino patch was merged, just fyi
<clever>
glines: ah, nice
<viric>
gchristensen: who needs it?
<NixOS_GitHub>
[nixpkgs] wavewave opened pull request #20869: nixos-container: support bridge and port forwarding for imperative nixos-container (master...nixos-container) https://git.io/v1C6R
<gchristensen>
viric: zoom-us apparently? will test later if it actually needs it.
<clever>
gchristensen: ive tried patching zoom-us before, and it was a nightmare, lol
<gchristensen>
well
<gchristensen>
damn :(
<gchristensen>
would you like to try instead applying sec patches to icu 54.1? :)
<gchristensen>
I need a todo list, or a secretary, heh
evangeline has joined #nixos
<evangeline>
clever: yeah the -j1 works, why is that?
<clever>
evangeline: i think it tried to tar up files inside /nix/store, while other jobs modified /nix/store
<clever>
evangeline: hence the error: tar: /nix/store: file changed as we read it
<evangeline>
ah, that explains it
systemfault has joined #nixos
lverns has quit [(Ping timeout: 258 seconds)]
<viric>
gchristensen: do you run ARM?
hiratara has quit [(Ping timeout: 240 seconds)]
<viric>
hm I should get an hdmi screen
hiratara has joined #nixos
octalsrc has quit [(Quit: WeeChat 1.6)]
<clever>
viric: ive got 4 rpi's, and an HDMI on one of them, along with network boot configured
<viric>
brave.
<viric>
There is something I want since years and nixos lacks it
<clever>
?
<viric>
I want a Karaoke program
<clever>
ah
<viric>
there was one thing in freepascal but that never worked iirc
<clever>
i know somebody that had one in java
<viric>
"ultrastar deluxe"
Wizek has joined #nixos
<viric>
that's not only karaoke... It evaluates the singing and makes it a competition with scores :)
<clever>
seanparsons: order of operations, environment.systemPackages = [ foo bar function baz ]; is a list with 4 items, one of them a function
<clever>
it didnt run the function
<seanparsons>
But it works if I use nix-shell and the same expression.
justanotheruser has quit [(Ping timeout: 268 seconds)]
justan0theruser has joined #nixos
<seanparsons>
Ah...
<clever>
environment.systemPackages = [ foo bar (function baz) ];
<clever>
this will run the function, and put the result in the list
<seanparsons>
clever: Yes, that's it, you took like 3 seconds to find the thing I've been staring at for 2 hours... :/
<clever>
and i didnt even read the error or see the source
<clever>
its a common issue that comes up in here atleast twice a month
<seanparsons>
I wonder if there's a way to make nix highlight it in some slightly more useful way, as I was staring up and down the chain picking through nixpkgs to no avail.
<clever>
cannot coerce a function to a string, at /nix/var/nix/profiles/per-user/root/channels/nixos/nixpkgs/lib/strings.nix:442:44
<clever>
one of the elements in systemPackages was a function
<clever>
but there error messages are a bit difficult to read
<seanparsons>
Yeah, but I mean I was seeing references to config.system.path in the gnome3 config in the trace, which made me think I'd done something super crazy somehow.
<clever>
thats because gnome3 wants the absolute path for the systemPackages buildEnv
<clever>
so it depends on systemPackages being valid