<ij>
If my binary needs uname, I just add coreutils to deps?
<clever>
at least under root, the channel name should be nixos
<clever>
ij: at runtime or build time?
<ij>
runtime
<clever>
ij: you can either rely on the fact that coreutils is always installed, and just expect it to be in $PATH
<clever>
ij: or you need to bake the path of coreutils into your build, by either adding ${coreutils}/bin to $PATH, or adding $(which uname) in where you call uname
<clever>
the result of running which uname at build-time
newhoggy has joined #nixos
<ij>
Hm. What's the PATH for a docker build?
<clever>
dont know, havent done that much with docker
newhoggy has quit [(Ping timeout: 240 seconds)]
spinus has quit [(Ping timeout: 258 seconds)]
newhoggy has joined #nixos
newhoggy has quit [(Ping timeout: 255 seconds)]
<dmj`>
ij: asking about docker in the #nixos channel, risque
<dmj`>
sounds like you want to be persuaded
nh2 has quit [(Ping timeout: 255 seconds)]
newhoggy has joined #nixos
mudri has joined #nixos
<Infinisil>
Ah, I got it, made a systemd service to run that mpc idleloop player thing that outputs to a fifo, so that xmobar can use that instead of continuous polling (which is quite cpu intensive because it needs to connect to the mpd server every time)
pxc has joined #nixos
<Infinisil>
It's not quite stateless and needs to run as my user (so it shouldn't really be in configuration.nix), but it works, so whatever
<kiloreux>
I am getting a very weird error when installing ffmpeg-full on my VM "error while loading shared libraries: libbz2.so.1.0: cannot open shared object file: No such file or directory"
<kiloreux>
even though installing and using ffmpeg on my machine directly works just fine
<kiloreux>
any idea how I could debug this issue ?
<matumental>
so my bootloader got hosed but all the partitions are fine, will booting the nixos installer, mounting the existing partitions and running nixos-install fix it using the existing /nix/store or is there some incantation of nixos-rebuild I should use?
<clever>
matumental: nixos-install by itself, will use the installers copy of nixpkgs to fix it, which may upgrade or downgrade things temporarily
<clever>
matumental: you can also "nixos-install --chroot" and then "nixos-rebuild boot" to use the nixpkgs inside the install
<clever>
matumental: and make sure you fix the uuid for /boot in /etc/nixos/, so it knows which one to mount
<ertes>
while we're on the subject: i've so far used systemd-boot on EFI systems, but apparently GRUB supports EFI as well… can i use GRUB? reason i ask is that with GRUB i can specify multiple boot devices, while with systemd-boot i had to cheat by creating an MD array with metadata=1.0
darlan has joined #nixos
darlan has quit [(Client Quit)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peterhoeg pushed 1 new commit to master: https://git.io/vHhyi
<NixOS_GitHub>
nixpkgs/master 63b8d65 Peter Hoeg: syncthing-inotify: 0.8.5 -> 0.8.6
NixOS_GitHub has left #nixos []
ison111 has joined #nixos
MP2E has quit [(Quit: brb)]
MP2E has joined #nixos
hellrazor has joined #nixos
<matumental>
clever: thanks everything I needed : )
jake_ has joined #nixos
<jake_>
nix-env -iA nixos.hello doesn't work from a non-root, says attribute ‘nixos’ in selection path ‘nixos.hello’ not found
<NixOS_GitHub>
[nixpkgs] peterhoeg pushed 1 new commit to master: https://git.io/vHhpY
<NixOS_GitHub>
nixpkgs/master 4eeefad Peter Hoeg: krename: 20161228 -> 20170610
NixOS_GitHub has left #nixos []
<stanibanani>
so, i just noticed that immutable user
<stanibanani>
so, i just noticed that immutable user passwords aren't changed when rolling back
<stanibanani>
which makes sence securitywise, but is a major pain when i mess up the hash :/
simendsjo has joined #nixos
pie_ has quit [(Ping timeout: 240 seconds)]
arianvp2 has quit [(Quit: arianvp2)]
snikkers has quit [(Ping timeout: 260 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 1 new commit to master: https://git.io/vHhpr
<NixOS_GitHub>
nixpkgs/master ecc34cb Jörg Thalheim: glusterfs: do not set setuid in install...
NixOS_GitHub has left #nixos []
justbeingglad has joined #nixos
bennofs has quit [(Ping timeout: 240 seconds)]
ambro718 has joined #nixos
b has joined #nixos
romildo has joined #nixos
<romildo>
Vivaldi has suspended support for Linux 32-bit for Vivaldi 1.10. Unfortunately, this is due to Chromium suspending support for it and maintaining it themselves would take too much resources.
<qknight>
now i wonder if that is an important requirement or if nixos-generate-config.pl should be extended to have support for /nix/store being now subvolume
grumble has quit [(Read error: Connection reset by peer)]
newhoggy has quit [(Ping timeout: 255 seconds)]
grumble has joined #nixos
<gchristensen>
I don't think I follow, bennofs?
<gchristensen>
the one I pasted lets you sort a list but keep the header in place
<bennofs>
my paste was just to show some surprising behaviour with bash (the "body" line is missing from the output)
newhoggy has joined #nixos
<gchristensen>
ohh yes
<gchristensen>
yes, lines without a following newlines can bite you :(
<sphalerite>
yeah, encountered that while trying to work with PATH using printf "%s" "$PATH" | while read -d : element ; do ...
<bennofs>
how do you solve this even
<Mic92>
portable shell shell scripting, which can handles errors well, is pretty hard.
<sphalerite>
add another echo at the end :p
<sphalerite>
in the case of PATH, I just did
<sphalerite>
printf "%s" "$PATH:" | while read -d : element ; do ...
<sphalerite>
instead
<gchristensen>
gotta use shellcheck to even get a chance, Mic92
wubalubadubdub has left #nixos ["ERC (IRC client for Emacs 25.1.1)"]
newhoggy has quit [(Ping timeout: 240 seconds)]
arianvp2 has joined #nixos
python476 has quit [(Ping timeout: 260 seconds)]
peacememories has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
newhoggy has joined #nixos
arianvp2 has quit [(Client Quit)]
arianvp2 has joined #nixos
romildo has joined #nixos
vandenoever has joined #nixos
arianvp2 has quit [(Client Quit)]
newhoggy has quit [(Ping timeout: 240 seconds)]
arianvp2 has joined #nixos
peacememories has joined #nixos
jluttine has joined #nixos
SvenDowideit has quit [(Ping timeout: 268 seconds)]
matumental has quit [(Ping timeout: 260 seconds)]
newhoggy has joined #nixos
ebzzry has joined #nixos
leat has quit [(Ping timeout: 268 seconds)]
jluttine has quit [(Quit: WeeChat 1.7.1)]
zeus_ has joined #nixos
jluttine has joined #nixos
newhoggy has quit [(Ping timeout: 240 seconds)]
carlosda1 has joined #nixos
zeus_ has quit [(Ping timeout: 246 seconds)]
mudri has quit [(Ping timeout: 240 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] ttuegel pushed 46 new commits to master: https://git.io/vHjJy
<NixOS_GitHub>
nixpkgs/master 3f3d33a Thomas Tuegel: qt5: use distinct `lib/qt-5.x` prefix for each minor version...
<NixOS_GitHub>
nixpkgs/master 210f688 Thomas Tuegel: qt5: rename qmakeHook to qmake
<NixOS_GitHub>
nixpkgs/master 7b6f8dc Thomas Tuegel: qt5: move kdeDerivation to qt5.mkDerivation
NixOS_GitHub has left #nixos []
peacememories has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
carlosda1 has quit [(Ping timeout: 255 seconds)]
peacememories has joined #nixos
newhoggy has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 1 new commit to master: https://git.io/vHjUI
<NixOS_GitHub>
nixpkgs/master 92fc4a3 Jörg Thalheim: Revert "gdbGuile: mark as broken"...
NixOS_GitHub has left #nixos []
kezepema has joined #nixos
<qknight>
regarding the /nix/store i now just created a btrfs subvol, moved contents from /nix/store into it and named it /nix/store, after a reboot from the rescue system it worked
vanschelven has quit [(Ping timeout: 260 seconds)]
mudri has joined #nixos
newhoggy has quit [(Ping timeout: 255 seconds)]
<qknight>
how can i extend nixpkgs easily from configuration.nix itself? i have to use `fetchfromgit` and also add a nixos module (with a systemd service)
newhoggy has joined #nixos
newhoggy has quit [(Ping timeout: 240 seconds)]
newhoggy has joined #nixos
<DIzFer[m]>
systemd services go in systemd.services.<name>, packages in nixpkgs.config,<name>. But that's in theory, can you give more detail? Maybe there's a more correct solution
<DIzFer[m]>
rather, packages _can_ go in nixpkgs.config, but I haven't figured out overlays yet
<catern>
are there any other projects trying to solve the reproducibly-build-a-system-from-scratch-and-configure-it problem in the correct way other than Nix? (Docker is not correct for reasons I am sure we all know)
<dash>
catern: guix is a nix spinoff
<catern>
bundling libraries is obviously wrong but I don't see any other projects that aren't bundling libraries other than Nix? (and traditional distro package managers, but those are out of date big-time)
<dash>
catern: OSTree is vaguely aimed that way I think
<catern>
dash: oh sure, okay, besides Guix too :)
<catern>
oh yeah OSTree
<catern>
others?
<qknight>
DIzFer[m]: i know about these things but i simply want a oneliner which fetches stuff from a git repo, and does an imports on the module
<catern>
OSTree is kind of okay at this but it's still based on RPM which is suboptimal
<dash>
buildroot?
<catern>
true, buildroot does this but it's very embedded-focused
<DIzFer[m]>
so, the git repo already has the nix module in it? maybe `import pkgs.fetchgit { url = foo, rev = bar, sha256 = foobar };` works?
<qknight>
DIzFer[m]: that gives me infinite recursion
leat has joined #nixos
newhoggy has quit [(Ping timeout: 240 seconds)]
kezepema has quit [(Quit: Page closed)]
newhoggy has joined #nixos
jonte_ has joined #nixos
<DIzFer[m]>
Sorry to have to pastebin pieces of config, but I don't have it available publicly yet. Anyways, I have this working on my system. Big maybe, but maybe you can recycle it? https://pastebin.com/ezXQEJ7Y
<qknight>
DIzFer[m]: i've not put that to test yet, will be majure in a few days
<qknight>
DIzFer[m]: then i will write an announcement
<qknight>
DIzFer[m]: if you were just refering to the pattern of deployment, feel free ;-)
Isorkin has joined #nixos
newhoggy has joined #nixos
pie__ has quit [(Ping timeout: 260 seconds)]
newhoggy has quit [(Ping timeout: 260 seconds)]
<peti>
niksnut, ikwildrpepper: Hydry says that the evaluation of http://hydra.nixos.org/jobset/nixpkgs/haskell-updates has been been running for >1 hour. Normally, this takes 2 minutes or so. Is it possible that the machine is in trouble?
romildo has quit [(Quit: Leaving)]
carlosda1 has joined #nixos
revtintin has joined #nixos
arianvp2 has joined #nixos
peacememories has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
seanparsons has quit [(Ping timeout: 246 seconds)]
peacememories has joined #nixos
konfou[m] has joined #nixos
carlosda1 has quit [(Ping timeout: 240 seconds)]
vaibhavsagar has quit [(Ping timeout: 255 seconds)]
newhoggy has joined #nixos
cpennington has quit [(Remote host closed the connection)]
<catern>
dash: none of those are really any good though. (well, guix is good) OSTree just versions packages which come from RPM which has all the downsides of traditional imperative package managers (actually, is OSTree really any better than containers??) and buildroot is equivalent to portage or other source-based distribution package managers, which have their own problems
<gleber_>
when running nox-review, I'm getting:
<gleber_>
anonymous function at /home/gleber/.nox/nixpkgs/pkgs/applications/networking/feedreaders/rssguard/default.nix:1:1 called without required argument ‘qmakeHook’, at /home/gleber/.nox/nixpkgs/lib/customisation.nix:56:12
<catern>
so yeah
ryantrinkle has quit [(Remote host closed the connection)]
peacememories has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<Mic92>
buildroot is so hard to debug
newhoggy has joined #nixos
jonte_ has quit [(Remote host closed the connection)]
<Mic92>
Makefiles which call Makefiles which call Makefiles
<catern>
I don't think there are any interesting whole-system-build-systems around other than Nix, the only other popular "solution" is containers which are awful
<catern>
(Nix + Guix)
<Mic92>
only cross-compiling needs to get better
<catern>
big companies like Google do all their own in-house stuff and effectively build their own distros
<catern>
small companies use containers
<Mic92>
at facebook everthing builds with buck
<catern>
embedded developers use buildroot, sysadmins use apt/yum/etc.
<catern>
Mic92: yeah but how does that translate into a system which can actually boot on hardware?
<catern>
Mic92: or be actually deployed
<Mic92>
catern: no this is what they use in 2-3 mono repos, apart from that they use chef.
<catern>
yeah, but these bazel-clone build systems (such as buck) don't actually have a deployment story, right?
peacememories has joined #nixos
<catern>
they just build binaries, and you're screwed if you want to, say, invoke one binary from another
<catern>
that's my understanding anyway, maybe someone knows this is wrong?
newhoggy has quit [(Ping timeout: 240 seconds)]
newhoggy has joined #nixos
<Mic92>
Afaik nix/guix are the only systems combining package and configuration management.
<catern>
well that's not true
<catern>
traditional distro package managers do it too
<turion>
where in /sys do I find the CPU temperature?
<catern>
so does Docker, in a very hipster way, as it wraps those traditional package managers
<turion>
It's not at /sys/class/hwmon/hwmon*/temp*
<catern>
but docker and distro package managers both, unfortunately, suck
<Mic92>
But not really in depth. You have to merge your config files, which the upstream on your own.
<catern>
with the upstream*?
<Mic92>
*with the upstream
<catern>
no you don't :) you can use debconf in Debian
<catern>
which is a central place to configure things
arianvp2 has quit [(Remote host closed the connection)]
<Mic92>
debconf only touches the surface
<catern>
what do you mean?
newhoggy has quit [(Ping timeout: 240 seconds)]
<MichaelRaskin>
turion: head -n 999 /sys/class/thermal/thermal_zone*/*
<Mic92>
how you add entries to /etc/hosts with debconf?
<dash>
catern: debconf covers much much less configuration than nixos does
<catern>
dash: Mic92: sure sure I didn't say debconf was good
arianvp2 has joined #nixos
cfricke has quit [(Quit: WeeChat 1.8)]
<catern>
but it's purpose is to be a central way to configure everything
<catern>
its*
<turion>
MichaelRaskin, thanks!
<dash>
catern: anyway, yeah, if there was something i thought was better than nix i'd be in its irc channel instead of this one ;-)
<catern>
oh, I don't mean better than Nix :)
<catern>
I just mean anything even remotely comparable
<MichaelRaskin>
maybe OpenWRT uci is a more succesful configure-everything attempt; smaller scope helps
<dash>
my primary interest in nix is that i'm building a new programming language and I didn't want to invent my own package manager
<catern>
oh I've heard good things about opkg and OpenWRT
<catern>
their stuff might be good
<catern>
though smaller in scope, yeah
<catern>
dash: yes! excellent! :)
<Mic92>
dash: in a language package manager you probably want to handle version constraints, which is not handled by nix.
<dash>
Mic92: yes we've got a prototype for that layer
<catern>
oh right I talked to you in erights about it :)
magnetophon1 has joined #nixos
<MichaelRaskin>
Actually, if you looks at Nix as package manager, you could also consider making it possible to load different versions of the same library into the same running program without going crazy
jonte_ has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 1 new commit to release-17.03: https://git.io/vHjYR
<dash>
MichaelRaskin: yes that's feature #1 for me
<dash>
#2 is being able to sanely depend on binaries not written in our language
<MichaelRaskin>
As in libraries?
<dash>
no, subprocesses
<dash>
(we're not going to implement FFI, for security reasons)
<MichaelRaskin>
Ah, that should not be too hard unless you want strong guarantees about subprocess behaviours
newhoggy has joined #nixos
<dash>
well who doesn't want that, really
lambdael has quit [(Quit: WeeChat 1.7.1)]
<Mic92>
rust handles multiple versions sane, way better then node/npm
<MichaelRaskin>
Well, strong as in stronger than default OS guaranteses
<MichaelRaskin>
guarantees
<dash>
Mic92: it's not hard to be saner than javascript anything
carlosda1 has quit [(Ping timeout: 260 seconds)]
<MichaelRaskin>
It should help that you have access control built in: if you already hide whether a library is loaded, it should be easy to hide whether it is loaded once or twice
zeus_ has joined #nixos
newhoggy has quit [(Ping timeout: 240 seconds)]
hke has quit [(Quit: bye)]
zeus_ has quit [(Ping timeout: 246 seconds)]
newhoggy has joined #nixos
<gchristensen>
anyone on a laptop find it takes ~30s+ to reconnect to wifi after opening thte elid?
<gchristensen>
"thte elid" <- my laptop's keyboard bug out in full force :(
<MichaelRaskin>
For some networks this is how much time wpa_supplicant takes to connect
<ben>
do i put gnome3.adwaita-icon-theme into propagatedUserEnvPkgs if an app needs icons?
darlan has joined #nixos
<ben>
or is buildInputs enough?
<MichaelRaskin>
To be more precise, to do a fresh scan and connect
<gchristensen>
interesting, I wonder how apple laptops reconnect so dang fast
<MichaelRaskin>
They don't
<MichaelRaskin>
I mean, first this is to the same network, skipping the scan and caching the bssid
<gchristensen>
sorry?
<MichaelRaskin>
Second, they _report_ connecting and then spend multiple seconds after checking that the AP is present actually connecting
<gchristensen>
but if I'm on the same network as before, I'm able to immediately use the network
<MichaelRaskin>
wpa_supplicant does an honest scan, chooses the AP according to policy, connects and report success
<ben>
Does fetchgit do shallow clones/fetches?
<gchristensen>
I see
hke has joined #nixos
<MichaelRaskin>
Apple checks cached BSSID seems to be still visible, reports success, then connects
hke has quit [(Client Quit)]
<MichaelRaskin>
For brief suspends I often succeed letting wpa_supplicant reaunthicate to the same network
<Mic92>
gchristensen: iwd will be faster then wpa_supplicant. Unfortunaly it is not ready yet
newhoggy has quit [(Ping timeout: 240 seconds)]
<gchristensen>
interesting
<Mic92>
I already package it in nixos and add a module. But it currently lacks of clients to control it.
<MichaelRaskin>
Hm, written by intel and lives on Kernel.org, looks promising
<Mic92>
there are some interesting talks on the daemon
<MichaelRaskin>
Mic92: does it have its own cli client, though?
<MichaelRaskin>
On my not-exactly-NixOS system I use wpa_supplicant by manually editing config and restarting the daemon or using wpa_cli
<Mic92>
MichaelRaskin: it has, but to set wifi passwords at the moment
<Mic92>
*but it can't set wifi passwords
<MichaelRaskin>
These I can put into config, though
<MichaelRaskin>
(I hope)
stanibanani has quit [(Ping timeout: 240 seconds)]
<Mic92>
I think this is up to the client to do this.
<MichaelRaskin>
Wait a minute
<Mic92>
So it can be integrated with any desktop environment/embedded system
<MichaelRaskin>
Does this _require_ DBus?
newhoggy has joined #nixos
<MichaelRaskin>
(I mean, running DBus, libdbus is fine)
Tarinaky has joined #nixos
<turion>
Hm this cpu frequency scaling issue is really annoying. I'm still running on 800MHz and not getting anywhere. What place would be best to go for support?
<Mic92>
it is the otherway around. dbus server, but not libdbus
<MichaelRaskin>
That is somewhat sad
hke has joined #nixos
<Mic92>
well better then crap wpa_supplicant calls an api.
jluttine has joined #nixos
<MichaelRaskin>
It has wpa_cli that connects to a unix domain socket, and wpa_cli works fine
<Mic92>
I am speaking of accessing programatically.
<MichaelRaskin>
Well, spawn a wpa_cli subprocess
<catern>
sounds like a dbus debate is about to happen
<catern>
maybe you shouldn't?
<MichaelRaskin>
Actually, LibreOffice requires DBus to draw menus
<gchristensen>
O.O
<MichaelRaskin>
But it seems to be fine to live with an empty DBus session
<MichaelRaskin>
Which is OK as LO is big and heavy anyway
newhoggy has quit [(Ping timeout: 268 seconds)]
<MichaelRaskin>
But now I have a script to create isolated quasi-«system» DBus sessions because of LO, and iwd should be even less picky about DBus session
<jluttine>
how can i enable a user systemd service? i've written a service under systemd.user.services. i tried "systemctl --user enable servicename" but i got warnings that probably this wasn't the right way
Infinisil has joined #nixos
<Mic92>
in configuration.nix: systemd.user.services.<service>.enable = true:
<nh2>
is there a way to pass arguments (like --args) to `nixops deploy` or do I have to use the `set-args` thing?
marsel has joined #nixos
<jluttine>
Mic92: does it then enable it for every user?
<jluttine>
i want it only for some users..
<Mic92>
jluttine: yes, we do not have per user services at the moment
<jluttine>
:/
<jluttine>
i'm trying to get my tmux+weechat running automatically
mizu_no_oto has joined #nixos
<Infinisil>
Like on startup?
<lassulus>
jluttine: just write a systemwide service which runs as your user?
newhoggy has joined #nixos
<turion>
How do I install a 4.1 kernel? It doesn't show up in the list of packages...
<sphalerite>
ison111: yeah, I'm surprised too. But usually you can find this sort of stuff out by looking at all-packages.nix
fnljk has quit [(Read error: Connection reset by peer)]
marsel has quit [(Ping timeout: 240 seconds)]
fnljk has joined #nixos
myguidingstar has joined #nixos
jluttine has quit [(Quit: WeeChat 1.7.1)]
jluttine has joined #nixos
newhoggy has quit [(Ping timeout: 260 seconds)]
<Infinisil>
gleber_: There is `throw "error!"` for that I believe
newhoggy has joined #nixos
<gleber_>
Infinisil: I have found "abort" which is similar. Do you know when to use which?
<jluttine>
Infinisil: thanks. i wrote my service under ~/.config/systemd/user/irc.service
<Infinisil>
Ahh abort yeah
<Infinisil>
Hmm
<gleber_>
Infinisil: ah! The manual explains why throw is better
cpennington has joined #nixos
<gleber_>
Infinisil: Throw an error message s. This usually aborts Nix expression evaluation, but in nix-env -qa and other commands that try to evaluate a set of derivations to get information about those derivations, a derivation that throws an error is silently skipped (which is not the case for abort).
<LnL>
gleber_: yeah, you probably want builtins.throw instead of abort
ryanartecona has quit [(Quit: ryanartecona)]
newhoggy has joined #nixos
<Infinisil>
Does anybody know why compton is running under systemd --user? I'm not using any DE, just X11 + xmonad, services.compton is disabled in configuration.nix
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] gleber opened pull request #26700: qt: fix qmakeHook -> qmake in few packages. (master...fix-qmakeHook) https://git.io/vHjsu
NixOS_GitHub has left #nixos []
<sphalerite>
gchristensen: so since the hardware replacement didn't help I've been asked to send a video of the issue occurring in the firmware interface (because they don't support linux as they've said many a time), which they'll forward onto the product engineering team or something. Maybe one day this issue will be resolved :D
<Acou_Bass>
eey folks im trying to do an upgrade ubt i keep getting this error, and it seems to be relatively random which package it happens on (which then causes the derivation build to fail): /nix/store/8pcap19p6qwf06ra4iaja3n6k6p2jzwg-xz-5.2.2-bin/bin/xz: (stdin): File format not recognized
<MichaelRaskin>
Maybe network problem lead to download cancellation for binary cache entries?
arjen-jonathan has quit [(Client Quit)]
zeus_ has quit [(Ping timeout: 246 seconds)]
newhoggy has joined #nixos
turion has quit [(Remote host closed the connection)]
<Acou_Bass>
it seems to get stuck downloading 'breeze-plymouth' which isnt actually anywhere im my configuration so i have no idea why its downloading it..
<Acou_Bass>
XD
<Acou_Bass>
(i remember breeze-plymouth having this problem a while ago so removed it)
<LnL>
gleber_: I'll take a look
newhoggy has quit [(Ping timeout: 240 seconds)]
newhoggy has joined #nixos
turion has joined #nixos
marusich has joined #nixos
mellowmaroon has joined #nixos
newhoggy has quit [(Ping timeout: 255 seconds)]
<turion>
Yay, solved the cpufreq issue by downgrading the kernel to 3.10! Thanks to everyone who helped me
<mellowmaroon>
I keep getting errors from Haskell's stack when I try to build a new Hakyll project
<nh2>
mellowmaroon: which
<mellowmaroon>
It says Zlib isn't found, but I'm in a nix-shell with zlib
<nh2>
mellowmaroon: do you link it dynamically or statically
<mellowmaroon>
nh2: Not sure, whatever is default
<nh2>
mellowmaroon: because `nix-shell -p zlib` seems to provide only `/nix/store/....-zlib-1.2.11-static/lib/libz.a`
<mellowmaroon>
nh2: So do I need to change an option in the cabal file or somewhere?
<Guest43087>
sphalerite (IRC):
sean has joined #nixos
newhoggy has joined #nixos
sean is now known as Guest23783
<nh2>
who's building the Haskell code, the generic Haskell builder in nix?
<gchristensen>
sphalerite: and they want us to do that in the place to set the "bios" passsword?
<nh2>
mellowmaroon: because I think that one tries to link dynamically
<mellowmaroon>
nh2: I think it's stack in the nix-shell
<mellowmaroon>
My steps are nix-shell -p abcd, hakyll-init test, stack init, and stack build
<mellowmaroon>
and it chokes on that step
<nh2>
mellowmaroon: and if you put zlib as a `-p zlib`, does that help or do you do that already?
<mellowmaroon>
nh2: Tried it, same error
<nh2>
mellowmaroon: try zlibStatic
<nh2>
(I have no idea why `zlib` only downloads the package with the `.a` but `zlibStatic` also downloads the one with the `.so`)
newhoggy has quit [(Ping timeout: 240 seconds)]
<mellowmaroon>
nh2: Ok, I'll try that
<Guest43087>
sphalerite ok I replicated it in the bios.
newhoggy has joined #nixos
<mellowmaroon>
nh2: rats....same error
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] gebner pushed 1 new commit to master: https://git.io/vHjZU
<NixOS_GitHub>
nixpkgs/master 89e02c7 Gabriel Ebner: lean: 3.1.0 -> 3.2.0
NixOS_GitHub has left #nixos []
<nh2>
mellowmaroon: can you paste the full error?
<mellowmaroon>
nh2: It mentions the options --extra-include-dirs and --extra-lib-dirs, would those need to change? And sure, one moment
<sphalerite>
gchristensen: I used the boot option editor instead because it actually shows the characters
<sphalerite>
gchristensen: Of course now I can't actually reproduce it in the bios anymore -_-
<sphalerite>
gchristensen: I thought I'd reproduced it in the bios at some point after the motherboard replacement but maybe it came with a newer firmware that fixes the issue, at least in the BIOS :/
<LnL>
mellowmaroon: any particular reason why you want to use stack instead of the native haskell infrastructure?
<mellowmaroon>
LnL: What do you mean? I've only used stack before, so I don't really know what the native haskell infrastructure is
newhoggy has quit [(Ping timeout: 268 seconds)]
<mellowmaroon>
LnL: If there's a more NixOS-kosher way, I'm happy to use that instead
jgertm has joined #nixos
<Guest43087>
I've replicated it 3x in th boot editor
<LnL>
oh, you're not using the nix integration with stack?
<mellowmaroon>
LnL: I don't think so...like I said, just used `nix-shell stack hakyll` and took off
cfricke has quit [(Ping timeout: 240 seconds)]
<mellowmaroon>
LnL: woops that should be nix-shell -p
<sphalerite>
Guest43087: I presume you'll get as far as I did then. Replacement motherboard that fixes the issue in the BIOS but not in linux.
<sphalerite>
Guest43087: which firmware version do you have?
<nh2>
mellowmaroon: that won't work, without the integration, stack will only look for `.h` files in standard global locations like "/usr/include" and that kind of stuff. But that doesn't exist in nix
<LnL>
I'm not surprised that doesn't find system libraries
<mellowmaroon>
nh2: Good to know. I've been having a tough time wrapping my head around how Nix uses libraries
<nh2>
mellowmaroon: you need to either call stack with `--extra-include-dirs=/nix/store/...zlib-.../include` (the latter path can be evaluated by nix for you), or use the stack-nix integration, which does exactly that for you (creating on --extra-include/lib-dirs flag for each of your nix `buildInputs` dependencies)
newhoggy has joined #nixos
<LnL>
something like nix-shell -p haskellPackages.hakyll does work if you just want to use the binary
carlosda1 has joined #nixos
<mellowmaroon>
I'll most likely go for the stack-nix integration, that sounds like what I want
<mellowmaroon>
Off-topic, is there a way to set Chromium as the default browser? When I do it through Gnome/Chromium, it doesn't stick
<mellowmaroon>
So I assume there's a NixOS way to do that
khebbie has joined #nixos
<LnL>
I think this is the nix-shell equivalent of what you wanted to do nix-shell -p "haskellPackages.ghcWithPackages (hp: [ hp.hakyll ])"
<sphalerite>
mellowmaroon: idk, for firefox its built-in default setting mechanism works for me
newhoggy has quit [(Ping timeout: 268 seconds)]
Guest43087 is now known as grahamc
grahamc has quit [(Changing host)]
grahamc has joined #nixos
grahamc has quit [(Changing host)]
grahamc has joined #nixos
Infinisil has quit [(Quit: Lost terminal)]
yegods has joined #nixos
<gchristensen>
1.1.3, you?ePSA Build 4304.17
<gchristensen>
sphalerite: ^
<mellowmaroon>
LnL: That also gave me the zlib error
<sphalerite>
gchristensen: 1.2.2 according to dmidecode
<gchristensen>
interesting ...
<sphalerite>
can't see a build number
<sphalerite>
I also don't know what the old one was :/
newhoggy has joined #nixos
<sphalerite>
I'm guessing I'd have to boot into the bios to see the build number
myguidingstar has quit [(Ping timeout: 240 seconds)]
k2s has joined #nixos
<LnL>
mellowmaroon: hmm, what platform?
k2s has quit [(Client Quit)]
<mellowmaroon>
LnL: NixOS
newhoggy has quit [(Ping timeout: 260 seconds)]
<LnL>
17.03 or unstable
<mellowmaroon>
nix-channel --list says 17.03
<nh2>
mellowmaroon: if the `haskellPackages.ghcWithPackages` way, without any stack involved, gives you the zlib error, then something is wrong
<LnL>
yeah, let me check
<mellowmaroon>
nh2: Yeah, I copy+pasted what LnL gave me verbatim, and ran stack build after
newhoggy has joined #nixos
<sphalerite>
gchristensen: considering writing a kludge to filter them using raw device access and uinput
<gchristensen>
sphalerite: oh my word.
<sphalerite>
it could work, if they're easy enough to tell apart from intentional repeats.
<turion>
How do I find out what the current value of an attribute is, say services.xserver.synaptics.enable
<mellowmaroon>
turion: Thanks for asking, that's good to know :)
<turion>
sphalerite: thanks!
<sphalerite>
also tells you where it's defined and shows docs for the option (if they exist) :)
<turion>
Yes, that's helpful
<turion>
I wonder why my touchpad works although the value is false
<sphalerite>
probably using libinput
<turion>
so will anything change if I switch synaptics on?
newhoggy has quit [(Ping timeout: 240 seconds)]
<turion>
Could this also explain why it moves in a buggy way?
jluttine has quit [(Quit: WeeChat 1.7.1)]
<sphalerite>
Yes, it may stop working. If it continues to work it's likely to behave somewhat differently
<sphalerite>
Basically the best way to find out is to try :P
jluttine has joined #nixos
<mellowmaroon>
sphalerite: That reminds me, I've had issues on my desktop NixOS machine with libinput's naturalScrolling option
acertain has quit [(Remote host closed the connection)]
mizu_no_oto has joined #nixos
<mellowmaroon>
sphalerite: It will be enabled, but scrolling on my mouse remains unaffected. Is that option only for touchpads?
<sphalerite>
gchristensen: systemwide keylogger for collecting stats on the key-repeat issue, yay!
<gchristensen>
sphalerite: I just messaged them, and sent them 5 videos of me getting the problem. one was _really_ bad: I got asdsdfdf then asddf
<gchristensen>
(I just typed asdf over and over)
Baughn has quit [(Quit: ZNC 1.6.2+deb1 - http://znc.in)]
<sphalerite>
gchristensen: haha, cool. Hope you don't end up with the issue still occurring in linux but no longer occurring in the BIOS like me :/
<gchristensen>
I'd bet you'd still get it if you tried for awhile
Baughn has joined #nixos
<gchristensen>
it takes my system a few minutes of typing before it starts screwing up
peacememories has joined #nixos
newhoggy has joined #nixos
<turion>
What I don't really understand is when I have to reboot or restart certain services after nixos-rebuild switch.
<gchristensen>
sphalerite: also when I was describing the issue in a message to them I couldn't even type out "for example" without getting like 5 typos.
<turion>
Clearly I have to reboot for a kernel version switch, but otherwise, it's sometimes unclear
<sphalerite>
gchristensen: That sucks, the issue's never been that bad for me.
<turion>
It seems that X isn't automatically restarted, instead there is a warning. That's great. But some other time I had to reboot in order to get the new system, without any warnings before
ryanartecona has joined #nixos
* turion
restarts X quickly
turion has quit [(Remote host closed the connection)]
<gchristensen>
turion: sounds like a good bug to report :)
<gchristensen>
oh.
<sphalerite>
too quickly for you :p
<MichaelRaskin>
gchristensen: but «expected behaviour» part is really hard to describe
turion has joined #nixos
zeus_ has joined #nixos
<mellowmaroon>
I'll see if I can figure out this Haskell stuff a bit more later. Thanks for the help!
<gchristensen>
MichaelRaskin: yeah ... no kiddingg.
mellowmaroon has quit [(Quit: Leaving)]
spinus has quit [(Read error: Connection reset by peer)]
<sphalerite>
gchristensen: I typed for about 5 minutes (with the concession that I kept closing and reopening the boot option editor to clear the text field) and didn't get the issue once
<gchristensen>
huh.
newhoggy has quit [(Ping timeout: 268 seconds)]
<sphalerite>
even in linux the issue occurs only rarely
<MichaelRaskin>
Ah.
jbgi has joined #nixos
<gchristensen>
oh well that iss a huge improvement over what I have now
spinus has joined #nixos
<gchristensen>
I come of as an incompetent typist over here :Pf
yenzenz has quit [(Ping timeout: 240 seconds)]
zeus_ has quit [(Ping timeout: 246 seconds)]
<sphalerite>
http://ix.io/xEh I typed all of this and the bug *may* have occurred twice, but on both counts I'm not entirely sure it wasn't just me hitting the keys wrong.
<peti>
niksnut, ikwildrpepper: the evaluator on hydra.nixos.org is stuck.
<MichaelRaskin>
It's hard to ignore content when it is so great!
<sphalerite>
turion: you may need to set security.chromiumSuidSandbox.enable to true
<sphalerite>
MichaelRaskin: hahaha
<gchristensen>
sphalerite: say .. what key do you get a repeat on _most_?
<sphalerite>
hard to say, maybe a or '
<gchristensen>
what key layout
<sphalerite>
UK
newhoggy has joined #nixos
tmaekawa has joined #nixos
<gchristensen>
hmm mine is on the D key (but I use dvorak, so it comes out as e)
peacememories has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<sphalerite>
Hm, looking at the key log I can see that I often press the next key before releasing the previous one. But that's not helpful :p
newhoggy has quit [(Ping timeout: 255 seconds)]
<turion>
sphalerite: yes, security.chromiumSuidSandbox.enable fixes it. Strange that I only have this problem with the 3.10 kernel
<MichaelRaskin>
Unprivileged user namespaces may be the difference
<gchristensen>
sphalerite: dellcares' advertisement of being "24x7" is shockingly inaccurate seeming :P
<turion>
ah ok
<MichaelRaskin>
gchristensen: what if it is 24×7, and has a huge backlog?
<sphalerite>
gchristensen: If you'd care to run evtest on the keyboard device for a while (a while during which you're not typing anything secret of course) and share the results with me maybe I could see if the unexpected repeats have any sort of pattern to them
<sphalerite>
24h x 7 days, but how many people? ;)
peacememories has joined #nixos
newhoggy has joined #nixos
<makefu>
hey, i'd like to run a test for a module (namely ./nixos/tests/logstash.nix ) what do i need to start it? nixpkgs manual does not seem to contain the info how to build it
<sphalerite>
They do usually respond to me within 2 or 3 hours, which isn't too bah IMO
<turion>
MichaelRaskin, sphalerite: would it make sense to include an if-statement in the chromium package that activates the SuidSandbox whenever the kernel version is smaller than 3.something?
<turion>
If it makes sense, I'm happy to implement it, but I'd need some guidance possibly *^_^*
<MichaelRaskin>
The check has to be done in runtime, I guess
<MichaelRaskin>
And 3.x kernels are rare nowadays
<MichaelRaskin>
I am actually surprised that cpufreq support for your CPU has been dropped mere 10 years after its release.
<turion>
(Although it's a problem with the kernel, not ubuntu specific)
<turion>
(It was introduced with kernel version 4.1.something, but I didn't use this machine for 7 years)
newhoggy has quit [(Ping timeout: 258 seconds)]
<turion>
why does the check have to happen during runtime? Because people could boot different kernels? (One could just activate the SuidSandbox if they have a lower kernel installed, but yes, that's a lot of effort for little gain I guess)
<sphalerite>
turion: a problem there is that chromium the package can't really set systemwide options
<MichaelRaskin>
The bug linked is about 3.5, no?
<sphalerite>
you could nix-env -iA nixos.chromium and it would simply not be possible for that to change the system option
<turion>
sphalerite: Ah ok... it could at most issue a warning or refuse to be built
<MichaelRaskin>
And the machine where you build is not always the same as the one where you use
<sphalerite>
turion: it doesn't even have a way of determining if the option is set at build time
<MichaelRaskin>
Hydra, but you can have a local build server
newhoggy has joined #nixos
fnljk has quit [(Read error: Connection reset by peer)]
<turion>
sphalerite: But you just taught me about nixos-option..?
<MichaelRaskin>
turion: acpi=noirq: WORKS — did you try that?
<sphalerite>
turion: yes, so?
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
<MichaelRaskin>
Basically assume that the build is performed on a different machine
newhoggy has quit [(Ping timeout: 240 seconds)]
Mateon1 has quit [(Ping timeout: 255 seconds)]
Mateon3 is now known as Mateon1
<turion>
MichaelRaskin: Not yet, but I will, as soon as I understand what it does (it sounds like shutting off a lot of stuff to me right now)
<MichaelRaskin>
Which happens to run a different kernel
<turion>
But it's a runtime error
<MichaelRaskin>
Ah
<sphalerite>
turion: builds run in an environment defined entirely by what's included in the derivation. If you include the nixos config in the derivation, then the resulting derivation will be different on every machine and you need to build it separately on each machine (and each time a machine's config is updated)
<MichaelRaskin>
Then yes, you can edit wrapper, check uname + sandbox combo and complain
<MichaelRaskin>
Not sure if worth the effort
<sphalerite>
Ah right, yes, checking at runtime is of course possible
<sphalerite>
Maybe patching the error message :D
<turion>
Ah I begin to understand
<turion>
Or if we had a forum or something just discuss it once
<turion>
Well I guess the mailing list would do
* turion
seems to be too impatient for mailing lists ;)
<MichaelRaskin>
The story in the kernel Bugzilla about cpufreq bug looks interesting
newhoggy has joined #nixos
jbgi has quit [(Ping timeout: 246 seconds)]
<sphalerite>
turion: a forum is just a mailing list with a web interface :p
<LnL>
gleber_: ^, building the qt stuff takes a while :)
newhoggy has quit [(Ping timeout: 240 seconds)]
newhoggy has joined #nixos
<sphalerite>
turion: it's not like Linus's Nvidia rant really moved things forward...
<simpson>
sphalerite: It surprisingly didn't really hurt relations there, AFAICT; there's been a light animosity between Linus and the DRM maintainers for a while, so it didn't really make the relationship worse.
metaphysician has left #nixos []
<simpson>
What *is* surprising is that nVidia's ARM CPU efforts are continuing to cause them to open up more and more hardware spec sheets.
<sphalerite>
that's nice, maybe one day we'll have graphics drivers that don't crash the kernel.
metaphysician has joined #nixos
<simpson>
Get rid of PCIe first, or otherwise fix the design of the typical PC. GPUs can lock up and take the PCI bus with them.
<gchristensen>
anyone eusing Elm? :)
newhoggy has quit [(Ping timeout: 268 seconds)]
<catern>
okay, one example of doing better than Docker but not as good as Nix: unikernels
newhoggy has joined #nixos
<catern>
anyone have any other examples? :)
Isorkin has quit [(Ping timeout: 246 seconds)]
zeus_ has joined #nixos
Isorkin has joined #nixos
<sphalerite>
simpson: that would also be nice
<sphalerite>
simpson: maybe that's what's happening in my case
carlosda1 has joined #nixos
<sphalerite>
who knows, is there even a way to tell without attaching magical equipment to choice spots on the motherboard?
newhoggy has quit [(Ping timeout: 268 seconds)]
newhoggy has joined #nixos
zeus_ has quit [(Ping timeout: 240 seconds)]
<turion>
gchristensen: Not really "using" it, you probably need someone with programming experience in it?
<gchristensen>
turion: yeah, I'm curious in people's take an experience
<turion>
fwiw, I've graded a presentation and essay on it
roconnor has joined #nixos
<turion>
To me it seems like it's a great frontend library, but as a language I'd rather see GHCJS mature more and use Haskell library
fnljk has joined #nixos
newhoggy has quit [(Ping timeout: 240 seconds)]
<gchristensen>
oh interesting, fair point
Itkovian has joined #nixos
<turion>
I mean, there wasn't even typeclass polymorphism last time I checked
<turion>
And there is no code erasure in the created .js, so it's always I believe >100kb or so
<turion>
(Basically pastes the whole elm-stdlib into the .js)
<turion>
It seems more like a "I'm a maths bachelor doing a cool web startup"-thing than a "Let's do FP for really good websites"-thing
pie__ has joined #nixos
newhoggy has joined #nixos
<sphalerite>
Is there a way to get nix-store -q --tree to only show dependencies up to a certain depth in the tree? I can probably think of some grep hacks to do it but maybe there's a "right way"
ris has joined #nixos
<sphalerite>
nvm just gone and done it with grep -E '^(....){0,$depth}\+'
newhoggy has quit [(Ping timeout: 240 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 1 new commit to master: https://git.io/vHj8O
<NixOS_GitHub>
nixpkgs/master f8d7942 Franz Pletz: youtubeDL: 2017.05.29 -> 2017.06.18
<sphalerite>
huh, why does firefox have no override attribute?
<sphalerite>
actually I don't need to override it after all. But I'm guessing I'd have to do 'with import <nixpkgs> {}; wrapFirefox (firefoxUnwrapped.override {stuff})'
newhoggy has joined #nixos
phreedom_ has quit [(Remote host closed the connection)]
newhoggy has quit [(Ping timeout: 240 seconds)]
roconnor has quit [(Remote host closed the connection)]
roconnor has joined #nixos
newhoggy has joined #nixos
Isorkin has quit [(Ping timeout: 240 seconds)]
Isorkin has joined #nixos
pie__ has quit [(Remote host closed the connection)]
Filystyn has quit [(Quit: Konversation terminated!)]
newhoggy has quit [(Ping timeout: 268 seconds)]
newhoggy has joined #nixos
<turion>
How can it be that thunderbird uses 40% CPU when idling?
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<gchristensen>
lots of syncing or message indexing maybe
<sphalerite>
turion: doesn't for me
Itkovian has joined #nixos
newhoggy has quit [(Ping timeout: 260 seconds)]
newhoggy has joined #nixos
<marusich>
Does mkDerivation create a derivation (which can then be later built), or does it actually execute (build) a derivation? The Nix manual and the Ph. D. thesis are a little unclear about this.
<turion>
Ok, then I'll let it be and hope that it will be done syncing and indexing at some point :) thanks
<dash>
marusich: the former
Infinisil has joined #nixos
<marusich>
Is there an error on p. 27 of the Ph. D. thesis, then? It says: "The result of the function in this case is a derivation. This is Nix-speak for a component build action, which derives the component from its inputs. We perform a derivation by calling stdenv.mkDerivation." It suggests that one uses mkDerivation to both return a derivation and also to execute it, which I don't understand.
newhoggy has quit [(Ping timeout: 240 seconds)]
<MichaelRaskin>
We-ell
<MichaelRaskin>
It returns a derivation which is an action-as-object
<marusich>
I know that there are "store derivations" which are the serialized form of a derivation that totally describes how to build a software component, so I expected there to be some kind of Nix-expression-level representation of a "derivation"... I thought mkDerivation might create it, but that seems not to be the case.
<MichaelRaskin>
If you actually invoke this action by nix-build or nix-store -r or nix-env -iA…
<MichaelRaskin>
mkDerivation calls derivation
<MichaelRaskin>
(builtin)
<marusich>
I did look at that code (derivation), but to be honest I couldn't understand it, so I came here.
<MichaelRaskin>
It creates the derivation which is serialised as .drv
oever has quit [(Ping timeout: 240 seconds)]
<marusich>
What's confusing me is that in some places, the manual says mkDerivation executes a derivation, but in some places, it says stuff that seems to indicate that it *returns* a derivation.
<gchristensen>
sphalerite: hmm you had 1.2.x bios? 1.3.3 is available now
<MichaelRaskin>
execute vs prepare is the question of invokation
<sphalerite>
gchristensen: might be worth a try before trying to filter it :p
<MichaelRaskin>
In most cases Nix expressions do not care
newhoggy has joined #nixos
<marusich>
But section 14.3 implies that it returns a derivation: see comment "The result of this function call is an actual derivation that can be built by Nix" in item 3 under section 14.3.
<marusich>
So which is it?
<dash>
marusich: yeah it just returns one, e.g. calling mkDerivation in nix-repl will not build stuff
<marusich>
OK
<marusich>
That's kind of what I thought
<LnL>
creating the drv (nix-instantiate) just creates a file with the build instructions
<dash>
nix-repl has ':b' to build a derivation
<marusich>
So, is there a Nix-expression-level representation of a "derivation"? I know there is a C++ class defined to represent derivations, but I do not know about the Nix expression language.
pie_ has joined #nixos
<marusich>
Or, is it basically just treated like a function (a delayed evaluation of a nix expression)?
<MichaelRaskin>
on the Nix level there is an object representing derivation
<MichaelRaskin>
You can ask for its output path(s) for example
<marusich>
Hm, I see. What file will I find that in in the Nix source?
<sphalerite>
gchristensen: actually dell's site says that update is urgent because it fixes the ME vuln that everyone was talking about briefly recently
<gchristensen>
:o
<MichaelRaskin>
something in libexpr/primops.cc something in libstore/derivations.cc
<dash>
marusich: it's a basically an attribute set
<tilpner>
Infinisil - Just put them into systemPackages and use lxappearance to switch between them (tell me if you find a way to set a theme from nixos-config)
newhoggy has joined #nixos
<sphalerite>
grahamc: works just fine — stick the exe on a USB stick, press F12 to get to the boot menu, you get a GUI where you can select the path to the update file
<grahamc>
Nice...!
<sphalerite>
grahamc: also while it was updating the EC firmware it spun up the fans just in case. Perfectly sensible behaviour, but it amused me
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
<grahamc>
Haha. Yeah. Fair enough. I just got to that part.
magnetophon1 has quit [(Ping timeout: 260 seconds)]
<grahamc>
The EC update might be really stressful. Really generates a lot of heat writing a few kb of data.
<Infinisil>
tilpner: Ahh, I'll try that, thanks
<Infinisil>
It works, nice :D
newhoggy has quit [(Ping timeout: 240 seconds)]
<sphalerite>
grahamc: well given that the EC is what makes the fans respond to temperature changes I think it's very sensible to spin the fans up all the way while it's happening in case something does go wrong, the EC is broken, and the hardware is at genuine risk of ovoerheating
newhoggy has joined #nixos
<grahamc>
Yes of course :)
<sphalerite>
surprisingly slow to update though
<sphalerite>
What did they use, flash memory from 2005?
newhoggy has quit [(Ping timeout: 240 seconds)]
<Infinisil>
tilpner: Actually no it doesn't work :( Changing the theme in lxappearance doesn't persist over even restarting lxappearince
orivej has quit [(Ping timeout: 255 seconds)]
<tilpner>
Infinisil - After clicking "Apply", that works for me :/
simendsjo has quit [(Read error: Connection reset by peer)]
<tilpner>
Infinisil - I also have tx0.co/v9_H in my config
<Infinisil>
What's tx0.co/v9_H?
<sphalerite>
grahamc: let me know if the update helps with the keyboard issue! I haven't encountered it at all since booting after the update, but that could be a fluke
<sphalerite>
Infinisil: preusmably a paste URL
<tilpner>
Infinisil - Prepend https:// and it's a URL
<Infinisil>
Ah lol
<grahamc>
I got it already in the bios
peacememories has joined #nixos
orivej has joined #nixos
newhoggy has joined #nixos
<Infinisil>
tilpner: That does indeed look like something i should put in my configuration.nix
<tilpner>
Infinisil - You could try to do that temporarily in a shell, to test if it would fix your issue
<sphalerite>
grahamc: :(
<Infinisil>
tilpner: system.path is just $PATH?
<tilpner>
Infinisil - I just tried to get that value in nix-repl, and I have no idea what it is
<grahamc>
sphalerite do people tell you that you type very fast?
<tilpner>
Infinisil - config is an argument to my configuration.nix
<Infinisil>
Trying to do it in the command line
<Infinisil>
I need to figure out how my WM xmonad can accept custom session commands
<Infinisil>
Else I can't put it in configuration.nix
newhoggy has quit [(Ping timeout: 240 seconds)]
mudri has joined #nixos
<tilpner>
Infinisil - You could try setting services.xserver.displayManager.sessionCommands
<Infinisil>
tilpner: Just found that as well, gonna try
<turion>
It complains that it cannot find kdeinit5 and that klauncher isn't running
<turion>
I'm running xmonad, not plasma/KDE
<sphalerite>
gchristensen: of course, we can't really tell because it's not really documented properly publicly, and it's good old proprietary softwaare. ♥
<turion>
So are there missing dependencies in the okular package or so?
<sphalerite>
That aa was definitely my keyboard :(
<gchristensen>
:P
<sphalerite>
turion: I think it's just not designed to run outside KDE. My guess is that it wants to ask kuiserver for a file save dialog
<turion>
So it's not conceivable to let it start all the services it needs or so? (I don't really understand how KDE works)
<marusich>
dash, MichaelRaskin, thank you for the help! Also: LnL, thank you for the link. This sentence makes sense to me: "A derivation from a Nix language view point is simply a set, with some attributes." So it seems like mkDerivation does the following two things when it is invoked: (1) write a serialization of "the derivation" to the store (i.e., create the store derivation) and (2) return an attribute set representing the same "derivation", for
<marusich>
further processing by other Nix expressions. Is that correct?
<MichaelRaskin>
Yes
magnetophon1 has joined #nixos
simukis has quit [(Quit: simukis)]
<turion>
Is it just the problem that kuiserver isn't running or that it isn't installed at all?
<sphalerite>
turion: well it's loose coupling basically. It wants the service and doesn't care what it's provided by, which is a good thing basically, but means that with nix you'll have to make sure it can find a provider for it yourself
newhoggy has joined #nixos
ison111 has quit [(Ping timeout: 240 seconds)]
<sphalerite>
turion: you could try install kdeFrameworks.kinit and trying again, that's where kdeinit5 lives apparently
cpennington has quit [(Ping timeout: 240 seconds)]
<turion>
ok, will do!
<LnL>
marusich: yep and for nix expressions a derivation is an attrset with { type = "derivation"; }
<marusich>
Cool, that makes sense.
<LnL>
nix-repl will print that as <derivation ???>
ison111 has joined #nixos
<marusich>
I see. It seems to me like the lack of a clear definition for "derivation" can make it difficult to explain Nix to people who are not already familiar with it. A "derivation" specified by a store derivation is explicit, unique, and un-changing; it will always do the same thing, every time, when its build actions are executed. But a "derivation" specified by Nix expressions is subject to many possible interpretations, and thus it will create
<marusich>
many possible store derivations, depending on many factors (e.g., what version of the Nix packages collection you're using). Is that right?
<sphalerite>
turion: kuiserver is in plasma-workspace, you may need to install that too. I don't really know the details, I'm only finding this stuff by searching /run/current-system/sw on my laptop which happens to have kde enabled (even though I use i3)
Infinisil has quit [(Quit: Lost terminal)]
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
mudri has quit [(Ping timeout: 246 seconds)]
<sphalerite>
I also need to go to bed. Gnight folks!
<simpson>
marusich: That sounds about right, yes.
<turion>
sphalerite, good night, thanks!
<hodapp>
haah, "tensorflow is built from a downloaded wheel, because the upstream project's build system is an arcane beast based on bazel. Untangling it and building the wheel from source is an open problem."
Infinisil has joined #nixos
<hodapp>
yeah, I still have punch marks in my wall from the time I tried to run TensorFlow on Arch Linux :(
Infinisil has quit [(Client Quit)]
<turion>
kde is a general thing that didn't work well for me on nixos so far. Everything else, great or non-nixos-issue, but kde seems to be missing some important hints in the manual... I'll ask about all that the next days ;) good night!
roconnor has quit [(Remote host closed the connection)]
newhoggy has quit [(Remote host closed the connection)]
roconnor has joined #nixos
<marusich>
dash, MichaelRaskin, LnL, and simpson: thank you for the help. I was confused due to the lack of a clear definition for a "derivation," but what you've said makes sense. I understand better now.
turion has quit [(Ping timeout: 240 seconds)]
arianvp2 has quit [(Quit: arianvp2)]
Infinisil has joined #nixos
<Infinisil>
tilpner: It's working :D Somehow lxappearance is the only gtk application that doesn't use the set theme
<tilpner>
Infinisil - It does for me. Did you need to change anything else than adding the pasted variables?
<Infinisil>
I'm not sure, I set the variables and changed the theme with lxappearance, nothing else
<Infinisil>
Ohh
<Infinisil>
And I installed missing gtk engines
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 1 new commit to master: https://git.io/vHj0k
<Infinisil>
Ohh my next problem arises: I have a systemd service with Restart=always, how can it be that it's not running right now?
darlan has joined #nixos
<Infinisil>
Oh wait never mind, there's no wantedBy
darlan has quit [(Client Quit)]
cpennington has quit [(Ping timeout: 258 seconds)]
Roliga has left #nixos ["WeeChat 1.8"]
hiratara has quit [(Ping timeout: 246 seconds)]
hiratara has joined #nixos
<Infinisil>
Happy New Day!
xcb has joined #nixos
chris| has quit [(Quit: Quit)]
jgertm has quit [(Ping timeout: 240 seconds)]
Infinisil has quit [(Remote host closed the connection)]
chris| has joined #nixos
sary has joined #nixos
thc202 has quit [(Ping timeout: 246 seconds)]
<xcb>
I am probably making a beginner's mistake, but why does my cargo build fail on ld: cannot find -lxcb when it is installed (xorg.libxcb), I presume it is because it is not available at /lib. How can I make it available to the build?
ilyaigpetrov has quit [(Quit: Connection closed for inactivity)]
<c74d>
xcb: One can't, or at least shouldn't, install libraries like that. You can make it available by opening a Nix shell with `nix-shell -p xorg.libxcb` and running the Cargo build inside that shell
<c74d>
xcb: equivalently, you can run `nix-shell -p xorg.libxcb --run 'cargo build'`, and you can add more `-p foo` options to add more packages
<c74d>
if you have several libraries or other packages to include in the Nix shell, it may be better to write a `shell.nix` file that lists them (I can give you examples), and then just run `nix-shell --run
<c74d>
`nix-shell --run 'cargo build`, without needing to specify packages on the command line
<xcb>
c74d, that is what I tried: `nix-shell -p xorg.libxcb` and then running `cargo build`, but it still fails on "cannot find -lxcb"
<c74d>
oh
<c74d>
xcb: what crate are you trying to build?
<c74d>
(or, what crate is failing to build?)
<xcb>
it did build both crate dependencies xcb and i3ipc, but then I copied some of the example code and now it fails on that
<c74d>
xcb: what's the code you're trying to build? (paste in e.g. <https://dpaste.de>)
<xcb>
Basically a slimmed down version of what is the example of the documentation page of the xcb crate: https://github.com/rtbo/rust-xcb
newhoggy has joined #nixos
<ij>
Is there a package that builds locale files/sets LANG environment?
JagaJaga has quit [(Ping timeout: 260 seconds)]
<xcb>
c74d: Maybe the problem lies in me installing just rustStable.rustc and rustStable.cargo, and started using cargo right away. I find it unclear from the documentation on nixpkgs whether that is a viable approach or whether some other utilities are necessary to build a project correctly under nixpkgs
yegods_ has joined #nixos
yegods has quit [(Ping timeout: 260 seconds)]
peacememories has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<pikajude>
i forget; what do you do when "hook reply is decline" when trying to do a remote build
<pikajude>
dumb thing
<c74d>
xcb: I get a different error trying to build your code: "thread 'main' panicked at 'called `Result::unwrap()` on an `Err` value: Error { repr: Os { code: 2, message: "No such file or directory" } }'", while the xcb crate's build.rs is trying to look through its `xml` directory
<gchristensen>
pikajude: decline means the build hook can't build it
<xcb>
c74d: Ah, got that one myself too, python3 is a dependency that needs to be present for it to build
<xcb>
c74d: It is disappointing that they do not report what is missing by default
<pikajude>
gchristensen: but it never says why
<gchristensen>
and isn't that the pits
<pikajude>
you could say that
<pikajude>
i asked once about this and it couldn't build the linux kernel because i didn't specify big-parallel
<pikajude>
unnecessarily complicated way to find out a very simple problem
<pikajude>
but now it's happening for every derivation
<pikajude>
including ones that are text files
Jackneill has quit [(Remote host closed the connection)]
<gchristensen>
pikajude: as root can you ssh to the build machine?
<gchristensen>
using exactly thee config described in your machines fil
jluttine_ has joined #nixos
<xcb>
c74d: If I do ld --verbose and look at the SEARCH_DIR, I do not see libxcb in any of the paths present, so I it is likely that therein lies the problem. But I am obviously inexperienced with linkers, so I might be misunderstanding this
carlosda1 has joined #nixos
<pikajude>
gchristensen: yup
<pikajude>
i removed the "1 kvm big-parallel" from the end of the remote-systems config file
<pikajude>
i don't know what 1 is
<pikajude>
so
<pikajude>
but that machine does have kvm and big-parallel so i won't be able to use it to build linux
<pikajude>
i think it was the 1 that caused the issue
<gchristensen>
pikajude: do the build machines have listed system matching what you're trying to build?
<gchristensen>
oh was that fixing it?
<pikajude>
it did indeed
<pikajude>
yeah i have i686-linux,x86_64-linux in the systems line
<pikajude>
i mean
newhoggy has quit [(Remote host closed the connection)]
<pikajude>
building a text file shouldn't need too many advanced features
<gchristensen>
so that 1 was probably a speed factor
<xcb>
c74d: Awesome! Thank you so much for your help! I will try it out right away :)
yegods_ has quit [(Remote host closed the connection)]
carlosda1 has quit [(Ping timeout: 240 seconds)]
<xcb>
c74d: Yay, it builds :D, when I run it, it does not log any of my key presses, but that is likely just me not understanding xcb yet. Thanks so much!
<c74d>
xcb: the program didn't seem to do anything for me either
<xcb>
c74d: I just tried a complete example of one of the listed examples and it worked as intended. Thanks again for helping me figuring this out!
newhoggy has joined #nixos
<c74d>
ah, good
jgertm has joined #nixos
newhoggy has quit [(Remote host closed the connection)]