<petersjt014>
hi: I'm trying to make a 32 bit installation iso to burn to a physical cd (for an old machine)--this Is what I ended up writing: http://ix.io/1wUQ
<petersjt014>
does anything seem way off? It seems to have eval'd fine, but I only have a few cds to burn--just wanna be sure
<petersjt014>
I can paste the build log too if needed
<gchristensen>
can you test it on a usb disk?
<Ankhers>
Does nixpkgs have f.lux, the colour temperature changing application? I'm not sure if I'm just not finding it.
<simpson>
Ankhers: I don't, sorry. Haven't really looked into it. An easy fix would be to hop into a nix-shell first.
<pbb>
Can anyone tell me why the network-setup.service is not setting my defaultGateway6 correctly, when manually executing the commands in its ExecStart script does?
<pbb>
I get the following route:
<pbb>
default via fe80::214:f1ff:fea7:2400 dev br0 proto ra metric 220 mtu 1500 pref medium
<pbb>
even though I specified this: 2001:41d0:2:d0ff:ff:ff:ff:ff
<pbb>
and after running the commands I also get the route I specified:
<pbb>
default via 2001:41d0:2:d0ff:ff:ff:ff:ff dev br0 proto static metric 1024 pref medium
<pbb>
but when the service gets run at boot or after a rebuild, it somehow doesn't create this route
<{^_^}>
[nixpkgs] @veprbl merged pull request #54611 → libsForQt5.poppler: fix on darwin → https://git.io/fhKuW
<pbb>
also my ssh connection breaks when I restart network-setup.service even though it should only set the IPv6 route and I'm connected over IPv4. What the actual hell is going on here..
aleph- is now known as Church_
<clever>
pbb: is ssh using v4 or v6? does it recover or stay broken?
<pbb>
I am explicitly using ssh -4 to connect. A few seconds after I restart networkh-setup.service the connection recovers.
<clever>
pbb: sounds like its bringing the entire network down, then back up
<clever>
so you have momentary packet loss
<pbb>
yeah but the ExecStart script only writes a openresolv config file and sets the IPv6 default route. Nothing else.
<clever>
restarting network-setup.service may restart other units
<{^_^}>
[nixpkgs] @NeQuissimus pushed 0 commits to hardened_structure: https://git.io/fhiIt
<pbb>
clever++
<{^_^}>
clever's karma got increased to 85
<pbb>
it does say more
<pbb>
I think dhcpcd is messing with me
o1lo01ol1o has joined #nixos
<clever>
that reminds me, when my isp was down a month ago, i told dhcpcd to renew the lease with upstream
<clever>
when it failed, it deleted the internal LAN ip!!
<clever>
so i lost all control of the router
<clever>
why is dhcpcd messing with the static ip of an iface it doesnt manage?
<pbb>
I discovered that my provider supported IPv4 setup via DHCP by accident - I already thought I had lost that server, when it came back up because it got the correct IPv4 configuration via DHCP.
<pbb>
So I just left it enabled
maximiliantagher has joined #nixos
<pbb>
I should really disable dhcpcd and do that manually
<clever>
ah
<clever>
in my case, the router needs 2 IP's, WAN ip (via dhcp) and LAN ip (static, it IS the dhcp server!)
ddellacosta has joined #nixos
<pbb>
It is trying to do IPv6 autoconf even when I disable it specifically via sysctl, that is so wrong
<pbb>
like, I think dhcpcd just enables it again
<pbb>
so bye bye dhcpcd
<pbb>
because IPv6 autoconf results in a bad config in my case
tdbgamer has joined #nixos
<clever>
ah
<pbb>
at first it works but after a few minutes the link-local IPv6 default route disappears
<clever>
ive actually had ipv6 problems lately
<clever>
i turned on v6 forwarding, but the router doesnt route anymore
<clever>
so i had to just turn off radvd, or i have broken v6 on all machines
<{^_^}>
[nixpkgs] @NeQuissimus pushed to master « linux: Tag hardened kernels 2/2 »: https://git.io/fhiIn
maximiliantagher has quit [Ping timeout: 240 seconds]
<pbb>
wow I just broke my host's network config, but I could still ssh into a VM running on it and use that as a jump host using the IPv6 link-local IP :D
<drakonis>
infinisil: the attempts to run lutris seem to have died down, right?
<Yaniel>
doesn't nix have a builtin way to do that though
<petersjt014>
dunno
<petersjt014>
but mine accepts any number of paths and handles shared paths, so there's that
<petersjt014>
also no variables because I am terrified of them
<iqubic>
but it's also a really large PITA to try and understand.
<Yaniel>
I'm not even trying
<petersjt014>
takes in package names, then for each of them finds the path in the store, returns the paths, sorts them and eliminates duplicates, finds sizes of the paths, puts them all on one line with '+' as a separator, pipes in basic calculator and divides by 1024 (for MB) with 4 significant figures
<Church->
Yaniel: Weechats JS support is horrifying
<Church->
Just absolutely awful. They need to fix it I think.
<clever>
steam depends on the other mono too
<clever>
likely 32bit and 64bit mono's
<petersjt014>
JS has awful Js support
<Yaniel>
consider this though: P H P
<petersjt014>
AAAAAAAAHHH pls no
<Yaniel>
the thought that someone actually wants to use it for scripting
<Yaniel>
*shudder*
<clever>
Yaniel: many years ago, i made a php script, to defrag lvm
<petersjt014>
W A T
<Church->
Yaniel: Okay pointers when running just scripts aren't validated
<Church->
So many possible bugs
<Church->
So many
<Church->
And they know about it, that's on purpose.
<iqubic>
GHC and NerdFonts are among my biggest space hogs. So are two versions of mono.
<clever>
petersjt014: the backup files in /etc/lvm/backup/tank are a textual form of the lvm metadata, which fully describes what byte ranges you need to piece together, to make every LV
<Church->
Not sure if that's for all scripts(python, etc) or not though
fusion809 has quit [Remote host closed the connection]
<clever>
petersjt014: purely by reading that, i can find the fragments of a given lv, and then call pvmove, to safely move the fragments around to make them be in order
<clever>
petersjt014: so all writes to disk are done by pvmove, and there should be zero risk of corruption
fusion809 has joined #nixos
<clever>
one day, i was doing that while an xfs was mounted on the lv, and i suspect it ignored lvm via caching, and it corrupted an ext3 partition
<iqubic>
Oh, I'm sorry.
<clever>
only lost about 20gig of data at the time
<iqubic>
not that bad than.
<petersjt014>
I ran that thing from earlier and have spidermonkey on my system apparently
reinhardt has joined #nixos
<petersjt014>
it's a bare-bones vps
<clever>
petersjt014: spidermonkey is part of policykit
<clever>
petersjt014: polkit uses JS to define the rules of who can do what
<petersjt014>
js specifically...not json?
<clever>
and polkit is what allows you to do root-like things (rebooting) if you have physical access (logged in on a terminal or X11)
<clever>
petersjt014: yeah, it has JS functions, that are ran with your current login details, and return a bool to allow/deny access
<clever>
and then an admin can just write whatever logic they want to make up the ruls
<clever>
rules*
<clever>
petersjt014: did you disable man/info/documentation?
maximiliantagher has joined #nixos
<petersjt014>
no, I just never had a need to look at polkit's
<clever>
man/info/docs take up space, and if you dont need them, you can turn them off
PeterHK has joined #nixos
<PeterHK>
my nixos builder starter to fail yesterday with 'a 'x86_64-linux' with features {kvm} is required to build' (im using 'dockerTools.buildImage')
maximiliantagher has quit [Ping timeout: 244 seconds]
<clever>
petersjt014: also, wpad, the stuff to auto-configure proxy's on LAN's, uses JS
endformationage has quit [Quit: WeeChat 2.3]
<PeterHK>
clever: no /dev/kcm (its a AWS instance)
<PeterHK>
*kvm
<clever>
PeterHK: does dockerTools.buildImage have runAsRoot set?
<PeterHK>
yes
<clever>
PeterHK: runAsRoot requires spinning up an entire qemu VM, and without kvm, the performance is going to suck massively
<clever>
PeterHK: you can either fake it by setting `system-features = kvm` in nix.conv (via the nix.extraConfig option i think)
<PeterHK>
so i just need to add qemu in config.nix?
<clever>
PeterHK: or you can do things without runAsRoot
<clever>
PeterHK: it brings qemu in automatically, so you dont have to add it anywhere
<PeterHK>
ok cool. thanks a lot!
<clever>
if you fake having kvm support, the image build will be fairly slow
<clever>
so just not using runAsRoot would be the best option
<PeterHK>
we setup system user to run the app inside the container `runAsRoot = '' #!${stdenv.shell} ${dockerTools.shadowSetup} groupadd -r ${appUser} useradd -r -M -g ${appUser} ${appUser} '';`
<clever>
PeterHK: i think if you just add a derivation with $out/etc/{passwd,shadow,group} to the contents, you can skip runAsRoot entirely
systemfault has left #nixos [#nixos]
_kwstas has joined #nixos
_kwstas has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @abhi18av opened pull request #54761 → updated the text message and configs → https://git.io/fhim3
xkapastel has joined #nixos
<Church->
Hey how well is Steam's proton working? At all?
<clever>
Church-: works great for some games
maximiliantagher has joined #nixos
abathur has quit [Ping timeout: 250 seconds]
<clever>
Church-: stationeers only has 2 bugs, that are present on all linux distros (one font missing, minor, and a mouse moving glitch, patch available, but not easy on nixos)
<Church->
clever: Cool was just curious if it was working at all.
<clever>
Church-: megaton rainfall starts, but ive not tried to play yet, and the window size gets wonky if it looses focus
<Church->
Of if there was nixos packaging that needed doing
<clever>
dont have any other windows-only things installed currently
<clever>
Church-: it does need python3, nixpkgs was fixed for that a few months back
<clever>
it will clearly say "python3 not found" in stdout, if your behind
<clever>
.net based games (space engineers) dont work at all
<clever>
neither does borderlands 2, for an unknown reason
<clever>
anything unity based should be perfect
maximiliantagher has quit [Ping timeout: 240 seconds]
<clever>
wedens: before i switched to nixos, i had briefly gotten pci-passthru (but it depended on a monitor) to work on gentoo, with a win7 guest (using xen, rather then kvm)
<clever>
the main issue, is the my motherboard doesnt have per-slot reset lines, so i cant "reboot" the video card only
<wedens>
clever: without multiple monitors?
<clever>
and the linux gpu drivers mess with the gpu enough to break the windows drivers
<clever>
so, i had to 100% blacklist the gpu on the linux commandline
<clever>
which results in zero video output when linux boots
palo has quit [Ping timeout: 240 seconds]
palo1 is now known as palo
<clever>
single gpu, single monitor at the time
<clever>
then, via ssh, i can boot win7, and grant it control of the GPU
<clever>
and because linux has never touched it, the GPU is in the same state it would be in, when chain-loading win7 from grub
<clever>
and the drivers are happy
<wedens>
doesn't sound much better that dual-boot, tbh :)
<clever>
wedens: the worst part, is that if windows is shut down, it doesnt reset the state of the GPU
<clever>
so the windows drivers refuse to work on the next boot (of the guest)
maximiliantagher has quit [Ping timeout: 244 seconds]
<clever>
so, you must reboot the host to fix the gpu state
<clever>
and now, your linux host, has to live with the uptime stability of windows
<_deepfire>
or, maybe if Linux drivers could be used to do the resetting for Windows..
<clever>
_deepfire: i think that reset, is half done in the card itself (when the reset line on the pci slot is driven), and half done in the bios
<_deepfire>
as it, do just enough to bring the GPU state back to Windows-usable state
<clever>
i did mess with some utils, that try to save the gpu state, and restore it, but it didnt work
<wedens>
I think the main appeal of this looking-glass thing is that you can use you system as normal. and dedicated gpu output is just forwarded to your monitor
<clever>
yeah, with the mess i was doing, i would have needed a 2nd gpu, dedicated to linux, with its own monitors
<clever>
looking-glass just gets rid of the need for a 2nd monitor, but you still need 2 GPU's
<_deepfire>
GPU passthrough is black magic, extremely hardware specific : -/
<_deepfire>
(well, used to be a couple years ago)
<clever>
_deepfire: oddly, i had far far more success, with usb, ethernet, and soundcard passthru
<clever>
the gpu is just the weird duckling on the pci bus
<_deepfire>
there are VGA-specific bits in the PCI spec : -D
<clever>
the other anoying fact, is that once i got the gpu fully working, i realized i only gave it a 15gig disk image
<clever>
and i didnt have room to install a single game
<clever>
and after expanding the image, it just refused to work ever again
<Church_>
Reading through Writing an Interpreter in Go
<clever>
the above, is using the llvm libraries in haskell
<clever>
and then using standard haskell parsing libraries, to turn your language into an ast, into llvm, and then into binary code
<Church_>
Nice should show that to pushcx
<clever>
you can then either emit an executable, emit llvm bitcode, or directly write compiled code into ram and get a function-pointer you could FFI into
<clever>
so your code could load a script at runtime, compile it to native, and then run it, without ever writing to disk
<Church_>
Man LLVM is neat.
PeterHK has quit [Ping timeout: 256 seconds]
fusion809 has quit [Remote host closed the connection]
<bgamari>
but then if I invoke it again with nix build -f expr.nix --arg tarballPath $incorrect_path_to_tarball it fails
wigust has quit [Ping timeout: 245 seconds]
<domenkozar>
it will always need to be evaluated, but it would use derivation cached by the fixed-output hash
<clever>
bgamari: builtins.fetchurl or pkgs.fetchurl?
<bgamari>
clever, both
<clever>
bgamari: pkgs.fetchurl computes the $out based on the hash you declared, and will check /nix/store for it, and should only download once
<bgamari>
it doesn't appear to be doing so
<clever>
bgamari: builtins.fetchurl (when the hash is missing) will have to re-query every hour (configurable in nix.conf), and if the e-tag differs, re-download
<bgamari>
that's what I was hoping it would do
<clever>
bgamari: can you reproduce it with a simple example.nix?
<domenkozar>
bgamari: can you provide an sscce?
<bgamari>
I suspect so
<bgamari>
one moment
wigust has joined #nixos
marusich has quit [Remote host closed the connection]
<bgamari>
ahhh
<bgamari>
I think I see the reason
<bgamari>
I need to override the name attribute of the fetchurl derivation
<bgamari>
otherwise it will change with the url
slyfox has joined #nixos
<clever>
bgamari: ah, that can do it, the $out is based on both name and hash
<clever>
bgamari: so if you give it a different url, with the same contents, the name can sometimes differ, and then $out differs
<domenkozar>
yeah I always wished name wouldn't change the derivation hash
<clever>
bgamari: also, i try to avoid `(import <nixpkgs>{}).fetchurl` whenever possible
<bgamari>
clever, I don't do so in my application
<clever>
bgamari: one trick i use, to make sure i didnt miss any <nixpkgs> in the codebase, is to just unset NIX_PATH, but that makes it imposible
<clever>
ah
<bgamari>
I was just too lazy to freeze nixpkgs in the repro
<clever>
and builtins.fetchTarball gets you the same effect
<bgamari>
I generally include nixpkgs as a submodule
<bgamari>
in my project repository
<bgamari>
since I generally find I need to fiddle with it anyways
<bgamari>
at least in this particular project
<nimblepoultry>
yes got my first nixos desktop booting into kde ԅ(≖‿≖ԅ)
brejoc has joined #nixos
<_deepfire>
Is there a way to (easily) have a gcroot with "extended" semantics, where "extended" means "keep build dependencies too"?
<clever>
_deepfire: i think you want `nix-instantiate --add-root result --indrect` in combination with 2 nix.conf fields, keep-env-derivations and gc-check-reachability
brejoc has quit [Ping timeout: 240 seconds]
<_deepfire>
clever: ah, so it is a global config affair, I see..
hHhHhHh has joined #nixos
hHhHhHh has quit [Client Quit]
<clever>
_deepfire: the only other option, is to do what hydra does, traverse the build-time graph yourself, and add an indirect root for every drv
<_deepfire>
clever: I see, thank you!
<clever>
_deepfire: the `hydra-eval-jobs` binary does that, and warns you if you forgot the dir: warning: `--gc-roots-dir' not specified
<clever>
hydra-eval-jobs takes a release.nix, and spits out json describing every drv in the file, and creates the roots for you
<clever>
[root@nas:~]# ls -l /nix/var/nix/gcroots/hydra/
<clever>
it expects a path within the magic gcroots folder, since its not making indirect roots
<clever>
on its own, you can (ab)use that binary without any hydra setup
<clever>
but you wont have any way to prune the roots automatically, thats handled by hydra-update-gc-roots.service, which uses the postgres DB
<_deepfire>
Interesting.. I was wondering if the concept of keeping build deps could perhaps be sufficiently useful for people to warrant a higher-level concept.
<_deepfire>
*a higher-level abstraction
<clever>
_deepfire: also, where do you draw the line, do you keep the gcc source code?
<clever>
because the gcc source is in the build-time dep tree
<etu>
,locate pangoft2
<etu>
,locate ps2pdf
<{^_^}>
Found in packages: ghostscriptX, texlive.combined.scheme-medium
<{^_^}>
Found in packages: gnome2.pango.dev
<_deepfire>
clever: good question!
<clever>
_deepfire: one trick ive used, just leave a nix-shell open
<clever>
all paths in env vars are "in use"
<_deepfire>
clever: oh I do that!
<_deepfire>
it's fragile, as human memory is..
<clever>
that will draw the line for you, only the immediate deps of that shell
<_deepfire>
clever: perhaps cut-off points can be specified, like "please stop looking downwards at this .drv", which can be your ghc
<_deepfire>
ofc that might require specification of a set of cut-off points..
<_deepfire>
..at which point one might want a UI to do specify it comfortably (in general)..
<clever>
_deepfire: it might be simpler to just root the immediate deps of a given .drv, and not be recursive in any way
<clever>
_deepfire: nix-store --query --roots will be of use, to see if a given path has roots or not, and which ones
<greatscottttt>
hi there, I'm still having a bit of trouble patching an external binary. I'm pretty new to all of this, what's the easiest way to write a derivation that processes a local binary?
<greatscottttt>
it seems like I can give a .zip to src of mkDerivation and I _think_ that's working
<Taneb>
greatscottttt: what's your end goal here?
<wedens>
greatscottttt: you can set `src` to local path
<greatscottttt>
even if it's not a zip? it always complains that it can't unpack the src
<symphorien>
greatscottttt: the easiest way might be to attempt to compile it from source, if possible
<greatscottttt>
right, unfortuantely it's a version of clang-format that's in our repo, which we're using to keep things in sync for everyone
<greatscottttt>
it's kind of annoying
<symphorien>
what about compiling the right version of clang-format from source ?
<symphorien>
you could even reuse what is in nixpkgs
<greatscottttt>
right that's an option...i just need to keep it in sync and make sure I update accordingly when it changes in the repo
kai_w has joined #nixos
<symphorien>
is the source tree for this clang-format also in the repo ?
<greatscottttt>
no, that's the thing
<greatscottttt>
i can work around this, I just thought it might be easier to have a script or .nix in the repo that automatically patches it would be easier
<clever>
greatscottttt: src = ./foo; where foo is a directory containing the ELF binary
<clever>
then the binary will be in the working directory for the bulk of the phases
<reinhardt>
Hi! On current nixos-unstable, profanity complains "configure: error: C compiler cannot create executables". Any hints?
vidbina has quit [Ping timeout: 244 seconds]
__monty__ has joined #nixos
<siraben>
Ah found it, it needed to be displayManager.lightdm.greeters.gtk.extraConfig
fragamus has joined #nixos
<teto>
any advice for a very simple repo with default.nix/shell.nix using callCabal2nix ?
<siraben>
I'm trying to figure out how to build https://github.com/borgbase/vorta apparently it already is in pip? "pip3 install vorta"
<siraben>
but I'm not sure how to do this declaratively
<srhb>
teto: Not sure what you're looking for. callCabal2nix "yourprojectname" {} should work just fine in most cases.
__monty__ has quit [Quit: leaving]
gagbo has joined #nixos
fragamus has quit [Ping timeout: 250 seconds]
<teto>
I have a default.nix with "haskellPackages.callCabal2nix "netlink-pm" ./. {}" and shell.nix that does `(import ./. {}).env ` which triggers `error: attempt to call something which is not a function but a set, at /home/teto/mptcpnetlink/hs/shell.nix:2:2`
<srhb>
reinhardt: try building it with --keep-failed and checking config.log
<PeterHK>
dockerTools.buildImage runAsRoot suddenly requires 'kvm' (after update channel as part of deploying changes) which it did not before :( i prefer slower container build over running them as root... do i need to change channel back to old version or can i dissable this kvm requirement somewhere?
<srhb>
PeterHK: I thought runAsRoot always required KVM?
<srhb>
Since it spins up a VM to do the work.
amosbird has joined #nixos
<PeterHK>
it did not... this ci machine was building container with runAsRoot for almost a year without kvm
<PeterHK>
its an aws instance so getting /ev/kvm is a bit difficult :(
dermetfan1 has joined #nixos
<PeterHK>
error: a 'x86_64-linux' with features {kvm} is required to build '/nix/store/m44jn6dlppyyrmad292dinz2khqykq3w-docker-layer-app-15b798d.drv', but I am a 'x86_64-linux' with features {benchmark, big-parallel, nixos-test}
dermetfan has quit [Ping timeout: 250 seconds]
<symphorien>
you can probably add the feature kvm in ~/.config/nix/nix.conf
<srhb>
PeterHK: I use it on ec2 and have supportedFeatures = [ "kvm" ... ]
<srhb>
PeterHK: (It really just needs to be able to spin up a qemu machine)
<teto>
anyone using linuxManualConfig ? for some reason it doesn't seem to install modules, like my /nix/store/hjxp4dbgzzhqdxb9kncd15rxwn51f637-linux-4.19.0-mptcp_v0.94.1/lib/modules/4.19.0/kernel is empty while I do have modules setup in the config
<Guanin>
Hi, I have a problem with string escaping. I want to escape double quotes, but `escape ["\""] v` does not seem to do anything in the resulting file, while it works as intended in the repl
clefru has quit [Remote host closed the connection]
<PeterHK>
srhb: i can not find anything about supportedFeatures the only thing i found was `nix.buildMachines` in /etc/nixos/configuration.nix which does not work (same error). /etc/nix/nix.conf is auto generated and it does not have any supportedFeatures key documented :( i found a related commit (https://github.com/NixOS/nix/commit/1e7b8deea7e052ed9ebf47d1411bcaf542054b41) which mentions system-features
<srhb>
Guanin: Can you provide the actual code that uses escape and doesn't work like you expect it to?
<PeterHK>
you have it working, so i guess im just too bad in finding how and where :(
<manveru>
so you want quotes inside of `v` escaped?
<Guanin>
manveru, yes
<srhb>
Guanin: Looks right to me.
<PeterHK>
error: a 'x86_64-linux' with features {kvm} is required to build '/nix/store/m44jn6dlppyyrmad292dinz2khqykq3w-docker-layer-app-15b798d.drv', but I am a 'x86_64-linux' with features {benchmark, big-parallel, nixos-test}
<srhb>
PeterHK: Sanity check, did the daemon restart properly when you rebuilt?
<srhb>
PeterHK: And where did you previously define benchmark, big-parallel and nixos-test? (I'm unsure whether that's actually some kind of default)
<srhb>
PeterHK: Also, is this a hydra instance?
<PeterHK>
previously i did not define anything... these are defaults
<srhb>
PeterHK: Ah, you don't have distributedbuilds on
<srhb>
I think your machine definitions won't be used at all in that case.
<PeterHK>
makes sense :D let me check
<srhb>
So either enable that or make the change in nix.conf instead (using eg. nix.extraOptions)
<Guanin>
Oh man, I just mixed something completely up and tried to escape the wrong string... escape worked flawless, sorry for stealing your time, srhb and manveru
<srhb>
Guanin: No worries. :)
<PeterHK>
still the error :( no its a gitlab-runner :D
<srhb>
PeterHK: Did the daemon restart properly?
erictapen has joined #nixos
<PeterHK>
starting the following units: nix-daemon.socket
<PeterHK>
i think so :D
<PeterHK>
let me try nix.extraOptions
<srhb>
PeterHK: Could you try ensuring it's actually been killed first? Otherwise, try setting the system-features, and then I'm out of ideas :)
<PeterHK>
stopping the following units: nix-daemon.service, nix-daemon.socket
<PeterHK>
i guess i can reboot just in case
<srhb>
That should be good enough. Odd. Pretty sure I had that working.
<PeterHK>
nix.extraOptions in configuration.nix? that then generates the nix.conf?
<srhb>
PeterHK: Yeah. Using ''system-features = ...''
[Leary] has quit [Remote host closed the connection]
Lears has joined #nixos
<Taneb>
Is there a way to tell nix "If you need to do builds you can spin up a machine on AWS"
sinner has joined #nixos
sinner is now known as Guest16529
<PeterHK>
yay containers building again! thanks a lot! btw i am not sure how you define a set of multiple values... current i have just `system-features = kvm`
<srhb>
PeterHK: Usually they're whitespace separated
<srhb>
Taneb: iirc the hydra nixos org config used to do that manually, so you might glean some information from checking the repo, but I'm unsure what the mechanism is these days.
<__monty__>
Make has the same relation to gcc a hadrian has to ghc.
<__monty__>
*as
o1lo01ol1o has quit [Ping timeout: 246 seconds]
<Taneb>
infinisil: it'll be the default but not necessarily the only
<Taneb>
And also hopefully it won't require GHC 8.8 to build, so bootstrapping is still possible
<__monty__>
It doesn't, I built it with 8.4.3 a few days ago.
<Phillemann>
I have a Python application that needs to execute "./setup.py foobar" before building (with setup.py) begins. I think the "configurePhase" might be the right phase to do that?
init_6_ has joined #nixos
init_6 has quit [Ping timeout: 246 seconds]
init_6_ is now known as init_6
<infinisil>
__monty__: Hmm yeah you're right
<wedens>
is there some simple example of cross-compiling haskell? (linux -> windows)
<__monty__>
I doubt it, cross-compiling is never simple : >
<Jonas|2>
Hey there, quick question: How does nixops manage the nixpkgs channel for the deployments? Does it use the channel configuration of the host or are these channels managed separately?
erictapen has quit [Ping timeout: 250 seconds]
<gchristensen>
it uses the NIX_PATH of when nixops is run
<Phillemann>
I'd like to append a command to the installPhase. Can I somehow call the "parent" install phase in my Nix expression (it's a python application)?
<symphorien>
use postInstall instead
<Phillemann>
Yep, that works, thanks symphorien
hyper_ch2 has joined #nixos
<{^_^}>
[nixops] @PsyanticY opened pull request #1085 → aws[s3]: Add versioning support → https://git.io/fhi6n
<{^_^}>
[nixpkgs] @dotlambda pushed commit from @r-ryantm to master « python37Packages.fonttools: 3.34.2 -> 3.36.0 (#54722) »: https://git.io/fhiPJ
<siraben>
What are your strategies for managing dotfiles like .Xresources? GNU Stow doesn't seem to be able to handle them
<siraben>
I want to version control them in my ~/dotfiles folder
<wedens>
angerman: siraben home-manager
<wedens>
* siraben
<siraben>
how about outside of Nix?
<siraben>
I might need something more generic
<siraben>
wedens: i'll take a look at home-manager
<simpson>
siraben: Memorization and inner peace with not being able to keep every dotfile with me at all times.
<siraben>
simpson: What do you mean?
<gchristensen>
simpson thinks about dotfiles like zen gardeners feel about their sand. the meditation is in the recreation.
<wedens>
angerman: do you have any specific pointers on where to start with cross-compiling haskell?
<simpson>
siraben: I mean that I should find a way to be comfortable sitting down at a machine, even if the machine is not comfortable for me.
<simpson>
...And as long as it has a vi. But I'll put up with nano.
<siraben>
Sure, of course I can work with standard bash, vim, etc.
o1lo01ol1o has quit [Ping timeout: 240 seconds]
<simpson>
I suppose that I was very much into power-user customization when I was younger. I even ran KDE. But these days I only usually set a .inputrc and .vimrc of a few lines.
<siraben>
simpson: what window system are you using these days?
<gchristensen>
siraben: I hear stow is nice?
<angerman>
wedens: github.com/angerman/nix-tools
<gchristensen>
but I use nix :)
<siraben>
gchristensen: yeah it is, although it doesn't seem to be able to handle dotfiles that aren't in a folder (i.e. they're directly in ~)
fendor has quit [Remote host closed the connection]
<simpson>
siraben: I'm using Awesome. It's alright. Mostly I've memorized some key combinations; the cost of maintaining an rc.lua is too high.
<siraben>
I can just run "stow emacs" after I git clone and it handles the symlink for me
fendor has joined #nixos
<siraben>
I see. My favorite is i3 + XFCE
<gchristensen>
I can recommend against making your ~ a git repo
<siraben>
Yes, of course :P
<siraben>
Perhaps I could just concoct a simple shell script and store all the .foo files in one folder, to be symlinked into ~
<gchristensen>
how often do you do this?
<__monty__>
siraben: I've use a Makefile for this before.
eadwu has quit [Quit: WeeChat 2.3]
<qyliss^work>
My experimental approach to dotfiles is writing nix derivations that wrap the programs and inject their config files using whatever mechanism they provide.
<qyliss^work>
It works surprisingly well.
<wedens>
angerman: sorry, I don't understand how it helps with cross-compilation :) does it generate nix expressions for cross compiling?
<gchristensen>
qyliss^work: nice!
<angerman>
wedens: it helps with generating nix expression that form the basis for cross compilation. (they retain flags, conditionals, ...)
<{^_^}>
[nixpkgs] @NeQuissimus pushed commit from @r-ryantm to master « kotlin: 1.3.11 -> 1.3.20 »: https://git.io/fhiD2
alex`` has joined #nixos
raduom has quit [Read error: Connection reset by peer]
silver has joined #nixos
o1lo01ol1o has quit [Ping timeout: 268 seconds]
<teto>
I would like for HM to generate RW files instead of just read-only, like a sandboxed mode
<teto>
sandbox* . right now .config/git/config is RO so git config fails :/
tmaekawa has quit [Quit: tmaekawa]
<srhb>
teto: That sounds sort of purpose defying.
<Myrl-saki>
jD91mZM2: I forgot what I pinged you for. Giff me a few.
<etu>
teto: Well, since HM uses the nix-store as storage... The files will always be RO.
<Myrl-saki>
jD91mZM2: Oh, nix-lsp.
<etu>
teto: I don't think that will change ever
<ottidmes>
jD91mZM2: Thank you for replying, I am actually the person who made the issue, so all is good, thanks again for fixing and thanks making it in the first place!
o1lo01ol1o has joined #nixos
<teto>
srhb: there should be ways for more flexibility when you are working on your config
<Myrl-saki>
jD91mZM2: I can't seem to build it on 18.09.
<teto>
etu: it could create a copy in the store
Jackneill has quit [Ping timeout: 240 seconds]
<srhb>
teto: What would home-manager switch do to your manual changes?
<etu>
That's not how the store works
<teto>
or create RO files aside, like you have your own .config/git/config that includes .config/git/config.generated
<teto>
so that you can experiment and once you are confident put the setting in HM
<etu>
teto: The purpose of the store is to build it from declarative configurations. If you do interactive changes how would it keep track of that for updates etc?
<srhb>
teto: I understand the desite, but allowing any kind of mutability opens a huge can of worms. If you want, you can write activation scripts for home manager or something that copy the generated file out of the store or something, but I'd really consider this a misfeature if it were included.
<adisbladis>
locallycompact: azure-cli is already packaged in nixpkgs
<srhb>
s/desite/desire
<teto>
etu: it doesn't need to be 100% declarative. Some settings I want always there (declarative), some I just experiment with, sometimes I feel constrained by HM
<adisbladis>
teto: You can always create a symlink to some mutable thingy outside of the store
<teto>
to take the git instance, I would like HM to generate a file in .config/git next to ./config/git/config
o1lo01ol1o has quit [Ping timeout: 245 seconds]
<locallycompact>
adisbladis, the one in nixpkgs is old
<locallycompact>
adisbladis, the one in nixpkgs is in node and this one is in python
<teto>
(btw I understand the concerns but I feel some users could appreciate this kind of behavior )
<srhb>
Surely. But I feel they may not realize that they lose a ton of reasoning power over their system.
<sphalerite>
srhb: you mean like nixos does when you set modes other than 0644 in environment.etc? ;)
<countingsort>
symphorien: legacy doenst require a /boot partition, does it?
erictapen has quit [Ping timeout: 246 seconds]
jasongrossman has quit [Quit: ERC (IRC client for Emacs 26.1)]
Soo_Slow has joined #nixos
<sphalerite>
countingsort: no, it doesn't
jomik has joined #nixos
<countingsort>
there doesnt seem to be a /boot
<countingsort>
oof
<sphalerite>
countingsort: yeah so it should definitely be in /boot if you don't have boot.loader.grub.mirroredBoots set to anything other than the default value
<sphalerite>
countingsort: but it boots fine?
<jomik>
Anyone know why I still get the error from this issue https://github.com/NixOS/nixpkgs/issues/16327? I have checked and NO_AT_BRIDGE is set to 1, so I would assume that it would be disabled, yet Taffybar still complains.
<{^_^}>
#16327 (by Pitometsu, 2 years ago, closed): WARNING **: Error retrieving accessibility bus address: org.freedesktop.DBus.Error.ServiceUnknown: The name org.a11y.Bus was not provided by any .service files
<countingsort>
(i only wanted to ask that one question, im currently helping another person so information is pretty slow, i should get them here0
o1lo01ol1o has quit [Remote host closed the connection]
<countingsort>
sphalerite: boots into grub rescue
<sphalerite>
ah
o1lo01ol1o has joined #nixos
<sphalerite>
well booting into a nixos installer, mounting everything, then rerunning nixos-install should probably fix it
<countingsort>
so theres only etc, lost+found, nix on /
<countingsort>
thats what they did to get to this state
jomik has quit [Client Quit]
samrose has joined #nixos
<{^_^}>
[nixpkgs] @nlewo opened pull request #54800 → Remove cloud-init from the Openstack image configuration → https://git.io/fhiQQ
<sphalerite>
that seems even stranger
<countingsort>
indeed
<sphalerite>
and boot.loader.grub.enable is true?
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #54801 → make-derivation: fix position in trace → https://git.io/fhiQp
<jomik>
Hey, I overrode Taffybar to get v3.1.1, but now it cannot recompile the configuration. It does not find any of the System.Taffybar modules. Any idea why? I used `taffybar = unstable.haskell.lib.overrideCabal unstable.haskellPackages.taffybar` to apply a patch.
<jomik>
So I made an overlay that sets haskellPackages.taffybar to that.
<wedens>
jomik: `appendPatch unstable.haskellPackages.taffybar (pkgs.fetchpatch { .... })` or something like that
<tobiasBora>
I'd like to know, on KDE the "print" key is bound to "Start Screenshot Tool", but when I press the print key I don't have any screenshot and I need to start spectacle manually to take a screenshot. Any idea?
<wedens>
jomik: also, consider using `extend`, if you dp the override in overlay
<jomik>
That is in an overlay, yeah.
<jomik>
so haskellPackages = super.haskellPackages.extend ?
<ckauhaus>
if you're not doing that on a regular basis
<jomik>
Ah, results in the exact same taffybar derivation though :D But looks a bit cleaner.
<jomik>
Thanks :D
<ckauhaus>
sphalerite: I think all us of 3 got it (Wed 2000 CET)
<ckauhaus>
everyone should write it now into his own calendar :)
<gchristensen>
are there any nifty CLI tools which can do symmetric encryption in a pipeline? ie: my-thing | symmetric-encrypt | ssh -c "symmetric-decrypt > my-thing.log" ?
<ckauhaus>
gchristensen: aespipe
drakonis_ has joined #nixos
vidbina has joined #nixos
<wedens>
maybe `gpg --encrypt --symmetric`
<gchristensen>
I don't have GPG on the other side :|
<jomik>
I tried navigating to the nix ghc that taffybar creates in taffybar-with-packages. And running `./ghc ~/.config/taffybar/taffybar.hs` only searches for modules in my ~/.config/taffybar dir.. That seems wrong, doesn't it?
<gchristensen>
I guess I might as well set up GPG. guh, I hate GPG.
<sphalerite>
gchristensen: what's the -c for?
<qyliss^work>
It shouldn't need any setup for symmetric, should it?
<sphalerite>
^
<wedens>
yeah, you just need the binary, I think
<gchristensen>
sphalerite: typos :)
drakonis has quit [Ping timeout: 250 seconds]
dbmikus has joined #nixos
<gchristensen>
really? I didn't know gpg could do that
<gchristensen>
also, could we please invent something better than gpg
<sphalerite>
it might create a ~/.gnupg and stuff, but yeah
<wedens>
jomik: you've overwritten haskellPackages, but taffybar (taffybar-with-packages) doesn't use it
<sphalerite>
gchristensen: for the symmetric case I think it's actually not too bad. gpg -c asks you for a passphrase, then reads on stdin and writes to stdout. gpg -d reads the file, determines that it's symmetrically encrypted from some sort of header, asks you for the passphrase, and writes to stdout
<gchristensen>
nice
drakonis_ has quit [Ping timeout: 252 seconds]
<jomik>
It does though? Normally it is version 3.0.0, but if I run `./taffybar --version` then it is version 3.1.1
<evck>
hi, i'm trying to crosscompile a Golang application for ARM. building fails because GCC doesn't support the "-marm" flag, and I can't seem to find any packages for arm-linux-gnueabihf or similar... is there another approach I should be using for crosscompliation?
<wedens>
jomik: well, that's just a guess :) try `taffybar.override { ghcWithPackages = unstableHaskellPackages.ghcWithPackages }`
<jomik>
wedens: When I check the taffybar version in the taffybar-with-packages path in nix/store, then it says that it is version 3.1.1 though?
<wedens>
if THAT doesn't work, then I have absolutely no idea :D
<jomik>
Oh, it compiles :D
<jomik>
rebooting, cuz, service is acting up :D
jomik has quit [Quit: WeeChat 2.2]
das_j has quit [Remote host closed the connection]
das_j has joined #nixos
das_j has quit [Remote host closed the connection]
das_j has joined #nixos
<gchristensen>
anyone seen a bug in electron applications where they only redraw the screen when the screen resizes? I'm having trouble with Slack on my dell xps 9380, but no such problem on my dell xps 9560.
<gchristensen>
I was thinking maybe it was a video driver problem, but it seems intel is in the default video driver list, so not sure
<wedens>
I had something similar with workspace switching
<ar>
i've seen that before, but about 2 years ago
<wedens>
some electron apps were showing loading screen until I switch to a workspace with the app
<gchristensen>
ls
<gchristensen>
hrm
jomik has joined #nixos
<jomik>
wedens: nice!
<gchristensen>
I'm not sure what to do with that :P
<wedens>
I had this problem on Arch, so it may not be nixos specific (if it's the same problem)
<wedens>
it disappeared just recently (electron update?)
<gchristensen>
ok, so a channel update later and a reboot and still no luck
<jomik>
I was just confused as to why taffybar --version output 3.1.1, when it could not compile properly, but I guess it was because the taffybar bin got updated, but it lost its ghc environment or something..
<florianjacob>
ToxicFrog: I'd sync & resolve conflicts and go explicitly ping somebody for the review. e.g. FRidh seemed to have interest in reviewing when trying to start the ofborg tests, but probably just forgot about it.
samrose has quit [Ping timeout: 240 seconds]
EffSquared has joined #nixos
dermetfan1 has quit [Ping timeout: 240 seconds]
<ToxicFrog>
florianjacob: alright. I'll see about doing that later this week; I'm currently in the middle of some major computer renovations, and the repo I do this development in is also my <nixpkgs>, so I don't want to start syncing that again until the hardware shuffle is done.
<Baughn>
My server is possessed by demons.
<ToxicFrog>
Baughn: aah, the eternal mood
<gchristensen>
sounds spirited, Baughn
<Baughn>
Remember how I was having trouble with nixos tests?
Lears has quit [Remote host closed the connection]
<Baughn>
Well, I still am. I've been running `nix-build ~/dev/nix/system/nixos/tests/gnome3.nix` repeatedly, and getting different results basically every time.
<florianjacob>
ToxicFrog: if he has no time, I'd ping the maintainers of the munin package (the munin service seems to have no maintainers), but it's usually better to not ping too many people at once. It happens that a PR falls through the cracks, and maintainers are usually happy if you remind them if they don't show activity for ~1 month that you're still alive and interested in the PR.
<Baughn>
clever: I tried replacing /nix/store with an ext4 filesystem, by the way. It made no difference.
<Baughn>
But I'm not sure what might. This is clearly demons.
<clever>
Baughn: what about a memtest86 on the host?
<Baughn>
Tried that. memtest86 doesn't start; the machine just reboots if I try it.
<Baughn>
This goes for both my threadrippers, by the way.
<Baughn>
And it doesn't misbehave otherwise.
<Baughn>
Running memtester, though. Because why not.
<ToxicFrog>
florianjacob: alright, thanks.
ryantrinkle has quit [Ping timeout: 240 seconds]
<ToxicFrog>
clever: so, update on my adventures with ZFS -- putting a short script to import -a and mount -a in postBootCommands mostly worked.
<ToxicFrog>
There's some other stuff in activation (apart from createHome!) that creates directories inside what are meant to be empty mountpoints, so I had to add some extra code to delete those
Lears has joined #nixos
<ToxicFrog>
It doesn't look like there's a good place to insert commands before activation, unless I want to move them into the stage1 before pivot-root, so I'm sticking with this for now.
eadwu has quit [Quit: WeeChat 2.3]
<clever>
ToxicFrog: you would need to edit stage-2-init.sh in nixpkgs, and add a new option
ddellacosta has joined #nixos
<ToxicFrog>
clever: yeah, that seems like too much work compared to just putting `${pkgs.findutils}/bin/find /backup/borg-repo -maxdepth 1 -empty -delete` in there. :P
<{^_^}>
[nixpkgs] @zimbatm pushed commit from @yorickvP to master « buildkite-agent3: 3.0.1 -> 3.8.4 (#54430) »: https://git.io/fhP34
<disasm>
One thing I did (although never finished) was setting up dual boot for the factory ubuntu image on zfs with luks so I could debug factory vs nixos issues.
<Jonas|2>
Hey there, I was the confused dude from yesterday or so that had problems with cache.nixos.org: It works now after nix-channel --update. I guess that specific file in the cache that I needed was corrupt or some other problem.
<ddellaco_>
I'm seeing a black screen with a cursor only when I start X, completely unresponsive. Logged in via ssh from my other machine I see no errors in the logs, lightdm (I've tried slim too) doesn't have any errors, neither does X11, display-manager...etc. I've tried setting plasma5 as well as lumina as my desktop manager, no changes. Any ideas? Seems like something simple...
strelok7 has joined #nixos
<ddellaco_>
most of my setup is pretty vanilla, but I've got a X1 Extreme that took a bit of configuration to get the hybrid card working, but that does seem to be up and running fine now
noam has joined #nixos
<ddellaco_>
thinkpad x1 extreme
<infinisil>
Could be a driver issue, maybe try the nouveau one
ibanja has joined #nixos
dbmikus has quit [Ping timeout: 245 seconds]
samrose_ has quit [Ping timeout: 250 seconds]
<{^_^}>
[nixpkgs] @ptrhlm opened pull request #54831 → zsh: fix for infinite recursion in VCS_INFO_detect_p4 → https://git.io/fhPBq
agander_ has joined #nixos
Guest16529 has quit [Ping timeout: 268 seconds]
<ibanja>
I have nix package manager installed on Void Linux. I added a channel but I am getting an error on update: https://paste.debian.net/1062953/
agander_ is now known as agander
<ibanja>
I am a member of the group.
<gchristensen>
wow, nix on void, cool :D
<ibanja>
gchristensen: would be if I could get it working :)
<symphorien>
Has someone compiled nix from source recently ? I am in nix-shell, and make complains that g++: error: EDITLINE_LIBS@: No such file or directory
<{^_^}>
[nixpkgs] @worldofpeace pushed to master « feedreader: use more stable url for patch »: https://git.io/fhPRy
<sphalerite>
symphorien: did you rerun the configure script?
<symphorien>
ibanja: what does not work ? empty sandbox-paths or sandbox-paths = /bin/sh=/path/to/a/statically/linked/sh ?
<symphorien>
sphalerite: yes, multiple times
<symphorien>
ah maybe I must rerun ./bootstrap.sh as well
<{^_^}>
[nixpkgs] @worldofpeace pushed commit from @r-ryantm to master « gnome3.evolution: 3.30.3 -> 3.30.4 »: https://git.io/fhP0G
<ibanja>
symphorien: neither works. Is "/path/to/a/statically/linked/sh" the actual string I should use or am I supposed to put a "path to a statically linked sh"?
<symphorien>
replace /nix/store/3jh2d752bf78b9j9f72qcq952azj2ffb-busybox-1.29.3/bin/busybox by /bin/bash
<symphorien>
is /bin/bash statically linked ?
<ibanja>
oh... got it
ambro718 has joined #nixos
shibboleth has quit [Quit: shibboleth]
<ottidmes>
are the abbreviated commits reported by GitHub future proof, I know it will increase the size of the abbreviation if it can conflict with some new hash, but does it ensure the old one stays pointed at the same old thing? (asking this for versioning of a bleeding edge nixpkgs, I know it does not matter in Nix, but still would prefer using unique versions)
ckauhaus has quit [Quit: WeeChat 2.2]
<ibanja>
symphorien: /bin/bash is probably not statically linked... I only have 24 files in /nix/store and none of them are shells https://paste.debian.net/1062960/
<symphorien>
then disable the sandbox:
<symphorien>
sandbox = false
<symphorien>
To use the sandbox you need a shell to populate said sandbox
<symphorien>
and this shell must be statically linked
retr has joined #nixos
<symphorien>
sphalerite: oh after ./bootstrap.sh it worked! thankgs
<sphalerite>
\o/
retr has left #nixos [#nixos]
<ibanja>
I installed this via the void linux package manager. Maybe I should uninstall it and install directly from the nix installation guide.
<ibanja>
I keep getting a "$HOME/.nix-defexpr/channels' does not exist" message
<symphorien>
you can ignore this one
<ibanja>
ok
<symphorien>
but please, when you say "it does not work" be more precise
<spacekookie>
I'm kinda wondering if there's some docs on how to mount a zfs pool (thats not the root). https://pastebin.com/KEUEh4Q5 doesn't seem to work where `zpool import DataCore` does exactly what I want it to
<symphorien>
you must install xz
<symphorien>
and tar
<spacekookie>
(the posted snippet kinda crashes the whole system because systemd freaks out)
<sphalerite>
spacekookie: huh, how does it freak out?
<{^_^}>
[nixpkgs] @worldofpeace pushed commit from @r-ryantm to master « python37Packages.pysam: 0.15.1 -> 0.15.2 »: https://git.io/fhPz7
<elvishjerricco>
spacekookie: You can in theory manually edit hardware-configuration.nix to remove that dataset. But that's generally frowned upon
maximiliantagher has quit [Remote host closed the connection]
<gchristensen>
heh, yeah, I just realized I have some non-legacy ZFS datasets which aren't referenced by the configuration.nix and just depend on zfs auto-mounting them and they work okay
<gchristensen>
but that is probably gross and I should probably fix that
maximiliantagher has joined #nixos
<ToxicFrog>
If you do want to use zfs auto-mounting, in order for it to work reliably you pretty much have to put it in postBootCommands so it happens before systemd startup
alex`` has joined #nixos
strelok7 has quit [Remote host closed the connection]
<ToxicFrog>
Otherwise, neither NixOS nor systemd know about the mountpoints and there's no guarantee zpool import/zfs mount run before services trying to use those mountpoints start up.
<{^_^}>
[nixpkgs] @artemist opened pull request #54835 → one_gadget: init at 1.6.2 → https://git.io/fhPgT
<elvishjerricco>
ToxicFrog: Works fine if you don't have any services using the dataset that launch at boot :P
<elvishjerricco>
But is there no way to declare a systemd dependency on the mountpoint being mounted?
tmaekawa has joined #nixos
<ToxicFrog>
elvishjerricco: if you're not using mountpoint=legacy, the mounts are handled entirely via `zfs mount`, not fstab or systemd mount units
tmaekawa has quit [Client Quit]
<ToxicFrog>
NixOS doesn't have a systemd mount generator for ZFS datasets yet, so systemd has no idea anything is meant to be mounted there.
<ToxicFrog>
I had constant issues with systemd units starting up before /srv was mounted and user units starting up before /home was
doyougnu has joined #nixos
<ToxicFrog>
If you're ok using mountpoint=legacy, doing that and setting the mountpoints in configuration.nix works fine.
<ToxicFrog>
And even then you need to watch out for the activation script (which runs before postBoot) creating home directories and service directories in unmounted filesystems.
<devalot>
Sometime after 2019-01-13 a commit in nixpkgs-channels/18.09 completely breaks all my qemu-kvm virtual machines. Rolling back to Jan. 13 fixes my issues. I can't afford the downtime to figure out which commit causes the issue. How should I report this?
<ToxicFrog>
elvishjerricco: you could try making zfs import a prerequisite of local-fs.target and then making sure your services start up after local-fs.target, but I struggled to get the dependencies for that correct for ages and never got it working reliably; it's easier to just import and mount all the datasets before systemd startup.
maximiliantagher has quit [Ping timeout: 268 seconds]
<Thra11>
What's the quickest way to get started on a new haskell project using Nix? Is there a `cabal init` style utility? or a ready made .nix file I can drop in with minimal modifications?
EffSquared has joined #nixos
<elvishjerricco>
Thra11: `with import <nixpkgs> {}; haskellPackages.callCabal2nix "foo" ./. {}` is a minimum that works like 80% of the time. You can `nix-build` it, and `nix-shell -A env` for local development.
<elvishjerricco>
Thra11: Of course that assumes you've already done a `cabal init` to create the cabal project
<Thra11>
elvishjerricco Cool, thanks. I'll try that. Does the file name (default.nix?) matter?
<infinisil>
gchristensen: Lol nice
<elvishjerricco>
Thra11: Yea you'd want to use `default.nix` Otherwise you'll have to specify the file you put that in on the command line (just a positional argument for nix-build and nix-shell, so no biggy)
Ariakenom has quit [Read error: Connection reset by peer]
<jomik>
I don't think it is an issue in your case infinisil , because you are not installing from nixpkgs, you are building your own ?
<jomik>
IIUC
<spacekookie>
gchristensen: that's how world domination starts!
<infinisil>
jomik: Ah, that's just a wrapper for taffybar, not the taffybar build itself, but the fix should still work I think. Maybe verify that "${pkgs.librsvg.out}/lib/gdk-pixbuf-2.0/2.10.0/loaders.cache" actually exists, these hardcoded version numbers look suspicious
Ariakenom has joined #nixos
<infinisil>
jomik: Yeah I built my own taffybar there
<jomik>
Yup
<jomik>
infinisil: I verified that before I added it, exactly for that reason
<jomik>
And also after.
<Unode>
hi all, is it common practice to package software that ships as .jar files in nixpkgs or is it preferred to compile from source if source is available?
<jomik>
And it does indeed exist
<Twey>
What's the difference between buildEnv and symlinkJoin?
Havvy has quit [Read error: Connection reset by peer]
<elvishjerricco>
Unode: I've never done any Java with Nix. But at least with other languages, it's preferred to build from source when reasonable.
Soo_Slow has quit [Read error: Connection reset by peer]
<jomik>
Tho, I don't know how it uses that loaders.cahe.
<jomik>
Don't get why it causes that error :(
<elvishjerricco>
Actually I did do a little Java with Nix a while ago, and we just downloaded a bunch of jar files from maven repos with Nix. But we weren't trying particularly hard to do the right thing at the time :P
<jomik>
elvishjerricco: the little Java I have done with Nix was attempting to fix the android dev environment. But nixpkgs insists on trying to accept actual licenses, which the jars have to download... So, that can not happen, since it is impure to allow them to do stuff like that! :D
<Unode>
elvishjerricco: my question was more from a compatibility point of view. With binary software one has to use patchelf or wrappers, however jars are mostly system agnostic so I wondered...
<elvishjerricco>
Unode: Yea, we were able to get by without any kind of binary patching by using jars from maven repos. Should be fine. JVM is pretty good about that kind of stuff
<Unode>
elvishjerricco: great, giving it a try
<Unode>
by the way, for the people with commit/merge access in the channel, what's the general procedure to request a merge? I have a PR ready pending for a merge for months now. Do I need to call out to specific users?
<jackjennings>
I'm seeing this today in my nixos install… not sure what could have caused it: `/nix/store/15kgcm8hnd99p7plqzx7p4lcr2jni4df-set-source-date-epoch-to-latest.sh: line 13: [: : integer expression expected`
<jackjennings>
Anyone seen this before and know anything about it?
<ddellaco_>
folks, I've been hung up on getting X11 running for days now. I'm trying to get my 1050Ti card working with my thinkpad x1 Extreme per the instructions here: https://nixos.wiki/wiki/Nvidia
ryantrinkle has quit [Ping timeout: 245 seconds]
<ddellaco_>
I've tried the static configuration which just gives me a black screen and cursor, and the dynamic mode, wherein the primerun script fails telling me there's an unfree package and that I need to enable it in configuration.nix...well, it's already enabled
<ddellaco_>
...and I'm not really interested in using bumblebee
<ddellaco_>
not sure what to do at this point, any suggestions would be most appreciated
<symphorien>
how did you enable unfree packages ?
civodul has quit [Quit: ERC (IRC client for Emacs 26.1)]
<ddellaco_>
symphorien: nixpkgs.config.allowUnfree = true in my /etc/nixos/configuration.nix
alex`` has quit [Read error: Connection reset by peer]
<symphorien>
you need to enable unfree package for nix-env as well
<ddellaco_>
symphorien: that's not what the script is telling me
<symphorien>
the configuration for this goes in ~/.config/nixpkgs/config.nix
<symphorien>
what does it tell you then ?
<ddellaco_>
I mean, it's giving me two options, I assumed it was one or the other, but it sounds like you're suggesting I need both of those (configuration.nix vs. ~/.config/nixpkgs/config.nix)?
<ddellaco_>
honestly I don't even understand that wiki page, it and the script talk about running a game--it doesn't make sense to me
<symphorien>
you need both
<ddellaco_>
this all seems so weird and hacky
<ddellaco_>
at least, the dynamic option does
<symphorien>
one enables unfree package installed with nix-env and nix-build
<symphorien>
the other enables unfree for things declared in configuration.nix
<ddellaco_>
okay guess I'll try that, since I have no other options at this point
<symphorien>
they speak about games because those are usually the most picky software, I guess
<ddellaco_>
and now it says "you need to run ./primerun.sh from another console"
<ddellaco_>
ah well
teehemkay[away] has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
sicklork1n has joined #nixos
<symphorien>
I think they mean a tty
alex`` has joined #nixos
<ddellaco_>
I don't understand why I would want to run this script manually in any case when I'm just trying to set up X11 for general use
dermetfan1 has quit [Ping timeout: 245 seconds]
<symphorien>
the static mode is said to drain the battery
<ddellaco_>
like, when do I run it? In what context? How do I set it up so that nix starts X11 up when I boot, etc.?
m15k has joined #nixos
<infinisil>
ddellaco_: A simple `services.xserver.enable = true;` should do it
<symphorien>
the nvidia card will only be used for programs lauched by ./primerun.sh
<symphorien>
so: if you don't run the script, you get low performance, but goot battery life
jackdk has joined #nixos
<symphorien>
if you run the script, you get a new tty with high performance graphics
<symphorien>
hence their example about games, because usually you only do this for games
<ddellaco_>
symphorien: so I run the script manually for specific applications every time?! So how do I start X11 just for like, day-to-day use? Like with a desktop manager?
<ddellaco_>
infinisil: yeah that was the very first thing I tried back four days ago believe it or not
<symphorien>
ddellaco_: that's what they call static mode
<symphorien>
no script, low battery life, high performance for exverything
<pbb>
nixos-upgrade is failing for me with `error: executing 'git': No such file or directory`. I use the builtin fetchGit in my system configuration.
<yl[m]>
LnL: is there a know issue on Mac? I'm running OSX 10.14.2 fresh install, I installed nix multi-user `sh <(curl https://nixos.org/nix/install) --daemon` and I get `error: unexpected end-of-file` when I run `nix-shell -p hello --run hello` from unprivileged used
<pbb>
rebuilding the system as a user works fine, even when git is not in the path
<yl[m]>
s/used/user/
<symphorien>
pbb: you need git in path
ryantrinkle has joined #nixos
<ddellaco_>
symphorien: I understand all of this conceptual stuff. My problem is getting it to work.
<yl[m]>
LnL: it works if I use root `sudo -i nix-shell -I nixpkgs=/nix/var/nix/profiles/per-user/root/channels/nixpkgs -p hello --run 'hello'`
<pbb>
shouldn't fetchGit pull in git automatically?
<pbb>
and how do I get git in the path of nixos-upgrade?
ToxicFrog has quit [Quit: WeeChat 2.3]
<ddellaco_>
symphorien: the instructions in the wiki page for dynamic have hardware.nvidiaOptimus.disable = true;, the static config has hardware.nvidia.optimus_prime.enable = true;
<ddellaco_>
okay clearly we are talking past each other
<ddellaco_>
yep see ya
<yl[m]>
LnL: yep that looks like the issue, I'll read the issue
<ddellaco_>
alright, anyways, if anyone has gotten X11 working with an nvidia card via https://nixos.wiki/wiki/Nvidia I'd love to get a peek at your config
thc202 has quit [Ping timeout: 244 seconds]
<pbb>
what solved my problem was changing just fetchGit to builtins.fetchGit
<yl[m]>
LnL: what's the reservation behind not using OBJC_DISABLE_INITIALIZE_FORK_SAFETY in the plist file?
dsiypl4_ has joined #nixos
q6AA4FD has joined #nixos
<yl[m]>
LnL: I'm wondering because it did trigger a rebuild of many dependencies and I'm wondering if it's worth it
dsiypl4 has quit [Ping timeout: 246 seconds]
eadwu has quit [Quit: WeeChat 2.3]
<LnL>
the reason it fails by default now is that it's potentially dangerous and might break stuff, also apple might start ignoring the variable at any point
<yl[m]>
I see, it looks like it's going to be a complex fix
<LnL>
yeah, but you can set the variable instead if you want
<LnL>
it's just a workaround and not a valid solution for the issue IMHO
<gchristensen>
seems weird that our memtest86+ binary doesn't support checking more than 256M of RAM
sicklork1n has quit [Ping timeout: 246 seconds]
<yl[m]>
LnL: I'll probably do that for now, thanks!
<LnL>
I should make some time and go through a dtrace again with that custom build and see where things go wrong
Ariakenom has quit [Read error: Connection reset by peer]
<jackjennings>
Is it possible to get the name of the nixops deployment in the machine declaration? Does this work? `webserver = { config, name, deploymentName, ... }: {}`. I wish that it was documents somewhere what this attribute set contained
<elvishjerricco>
How can I update the version of steam in nixpkgs