o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
jgeerds__ has quit [Ping timeout: 265 seconds]
cap_sensitive has joined #nixos
o1lo01ol1o has quit [Ping timeout: 260 seconds]
janneke has quit [Ping timeout: 260 seconds]
<cap_sensitive>
I'm running nixos-unstable on a Raspberry Pi 3. Initially I used linuxPackages_latest and everything works. If I switch to linuxPackages_rpi3, my pi won't boot. How can I use the rpi-specific kernel then?
<mjsir911>
infinisil: regarding rfc0042, do you have any suggestions for allowing types of settings to be specified in the configuration file?
<mjsir911>
Use case: I have an application who's configuration file differentiates between bytes & strings. I can have nix strings default to one of those, but the other has to be specified explicity somehow...
<infinisil>
I haven't really thought of a good way to do that no
acarrico has joined #nixos
<infinisil>
Which is why the rfc is only for configuration formats that have a clear mapping to nix
takeda has joined #nixos
<infinisil>
You could probably do something like `{ byte-setting = { _type = "bytes"; _value = "..."; }; }` and then handle this specially in the generator
<mjsir911>
Yeah that is currently what I'm doing with `AuthNtLmSecureHash = {type="byte"; value="aaaaaaaaaaaaaaaaaaaaaa==";};` minus the leading underscores, but it's reassuring that you came up with the same thing
* colemickens
had to learn/use filterdiff, but he got libpng patched inside freeimage.
<colemickens>
aka imv builds on aarch64 now!
<mjsir911>
Thank you infinisil!
srid has quit [Quit: Connection closed for inactivity]
takeda has quit [Ping timeout: 268 seconds]
drakonis has joined #nixos
h0m1 has quit [Ping timeout: 245 seconds]
kvda has joined #nixos
tomberek has joined #nixos
h0m1 has joined #nixos
viric has quit [Ping timeout: 260 seconds]
<pie_[bnc]>
apparently you can just stop the socket
<{^_^}>
[nixpkgs] @colemickens opened pull request #77655 → imagefree: (embedded libpng): apng support → https://git.io/JvJVQ
<qyliss>
it'll start again as soon as something interacts with the socket
<qyliss>
but yes
<pie_[bnc]>
oh wow i just accidentally found out only my left speaker is absolutely messed up
<pie_[bnc]>
while plugging stuff around in JACK
<gchristensen>
pie_[bnc]: I think if yo ustop the socket, it deletes the socket from disk
<gchristensen>
qyliss: I, mean ^
<kalbasit>
I'm trying to setup Unifi Controller on my Rpi, but I'm doing it through Nixops from my x86_64 nixos machine. I did setup nixpkgs.system and it's getting correctly installed, but the mongo package taking from another nixpkgs is not getting the correct system. What am I doing wrong? https://www.irccloud.com/pastebin/bZTh16yq/configuration.nix
<day|flip>
pie_[bnc]: im keeping thing simple and using sndio. sndiod simi config?: sndiod -f rsnd/DAC -e s32le -r 48000
<qyliss>
gchristensen: it's socket-activated though
<qyliss>
apparently
<gchristensen>
I know but the .socket unit creates the socket file on start, and deletes it on stop, and it only starts the target service on open() of the socket file as long as the socket file exists
<gchristensen>
no?
<qyliss>
Oh, right
<qyliss>
I misread and didn't see that pie_[bnc] stopped the socket
<gchristensen>
ah :)
<qyliss>
I thought they meant the service
<qyliss>
I really should stop reading IRC this late at night
<gchristensen>
you: 02:27 probably, me: 21:27: same
<qyliss>
lol
<pie_[bnc]>
day|flip: i have no idea sorry :V
<pie_[bnc]>
day|flip: i know nothing about sndio
<day|flip>
it made by openbsd
<day|flip>
think of it like a lighter version of pusleaudio
LarryTheCow has joined #nixos
<day|flip>
used aplay -l to find out what card your using. card X will be rsnd/X
<LarryTheCow>
Hey, I wrote a small script and I want to add it to my bin path. What's the best way I can make it available from $PATH?
<pie_[bnc]>
LarryTheCow: put it in your systempackages or use nix-env -i
<pie_[bnc]>
or put another directory in your PATH for scripts, theres probably a few other options, but thats what i can think of off the top of my head
johnny101 has joined #nixos
<pie_[bnc]>
there's also home-manager
<LarryTheCow>
Oh wow. I didn't think it would be that simple. nix-env -i worked
<pie_[bnc]>
so what you want depends on how heavily you use configuration.nix
<pie_[bnc]>
:D nice
<LarryTheCow>
Thanks a lot
<pie_[bnc]>
its nice when things just work
<pie_[bnc]>
sadly nixos isnt so magic yet that things always just magically do what you want x3
<pie_[bnc]>
have fun
<LarryTheCow>
nix is magic
chenhq2005 has joined #nixos
<nornagon>
Has anyone had success installing Ghost on nixos? there doesn't seem to be a nixpkgs entry for it
<nornagon>
pie_[bnc]: yeah someone's maintaining a docker image for ghost, maybe i could use that
halfbit has quit [Ping timeout: 245 seconds]
bhipple has quit [Ping timeout: 260 seconds]
<kalbasit>
is there a way to change the directory where Nix is building? I'm not sure why but my rpi can't even extract nixpkgs archive due to space even though I still have 40G left
<pie_[bnc]>
kalbasit: i think you can set TMPDIR or something like that
<kalbasit>
pie_[bnc]: I'll try
<pie_[bnc]>
kalbasit: this is one of those errors you can google and should end up with results
<pie_[bnc]>
(or someone else will just say exactly what needs to be done)
<kalbasit>
pie_[bnc]: was not able to find something, perhaps I did not look deep enough
felixfoertsch23 has quit [Ping timeout: 265 seconds]
<selfsymmetric-mu>
gchristensen: I set hard nofile to 1048577 in `security.pam.loginLimits` but that number did not appear in `sysctl -a`
<selfsymmetric-mu>
Is there any way to get a `ulimit` binary in NixOS?
<selfsymmetric-mu>
Wait I might have a problem with my PATH, I found my `ulimit`.
<selfsymmetric-mu>
Unfortunately after setting `security.pam.loginLimits` to `[{domain = "*"; type = "hard"; item = "nofile"; value = "1048576";}]`, `ulimit -Hn` still reports `524288`. So that does not work.
<selfsymmetric-mu>
I'm going to try and put DefaultLimitNOFILE=1048576 in my systemd.extraConfig and reboot. See you in a bit.
selfsymmetric-mu has quit [Remote host closed the connection]
[Leary] has joined #nixos
Lears has quit [Read error: No route to host]
selfsymmetric-mu has joined #nixos
<selfsymmetric-mu>
That did it. Now to see if it helps Proton to stop crashing so much.
<{^_^}>
[nixpkgs] @jtojnar merged pull request #73548 → pavucontrol, waybar: Use system style → https://git.io/JeoSM
<selfsymmetric-mu>
Well, I managed to lose my game before the game crashed. So far so good, unfortunately with the nature of these things, and the poor logging available, I won't know for a while whether or not I just got lucky this time.
<cap_sensitive>
When I put 'nix.extraOptions = "keep-outputs = true"; and nixos-rebuild switch, I got an error message: warning: unknown setting 'experimental-feature' and then error: cannot connect to daemon at /nix/var/nix/daemon-socket/socket
<cap_sensitive>
I'm adding this option because I want to keep the output of nix-shell so that they won't be garbage-collected
<monsieurp>
there's a reason why btrfs isn't mainstream yet unfortunately
<etu>
eoli3n: The same arguments can be used to not use the nvidia driver :p
<day|flip>
ya. a 10 year old bug still plage the darn thing
<etu>
Or virtualbox
<etu>
or any out of tree module
<etu>
especially the ones that isn't free software licensed
<monsieurp>
day|flip: really?
<monsieurp>
:O
<eoli3n>
he says what he want, zfs is by reputation the best filesystem ever, and will still for a while, its just licensing war
<wedens[m]>
there is no "the best" filesystem...
<monsieurp>
day|flip: can you share a link about that 10 yo bug?
<eoli3n>
wedens[m]: i think there actually is, when a filesystem has tons of stable futuristic features
<{^_^}>
[nixpkgs] @marijanp opened pull request #77662 → nixos/test: added verbose output for failed tests → https://git.io/JvJXd
<day|flip>
monsieurp: it mostly have to do with raid 5/6 on btrfs. it been on it longest unuseble. it claim it mostly ok for used. witch still mean it not really fix or stable for full time used
<wedens[m]>
eoli3n: convince me to use it on a desktop or a laptop without raid ;)
<eoli3n>
snapshots + native encryption + pool storage + data integrity check / automatic repair + copy on write...
<eoli3n>
what do you want more :)
<eoli3n>
wq
<day|flip>
monsieurp: im looking forward for ReiserFS 5 to come out.
NoctisLabs has quit [Read error: Connection reset by peer]
Naptra has quit [Remote host closed the connection]
chloekek has joined #nixos
<{^_^}>
[nixpkgs] @zaninime opened pull request #77665 → systemd-networkd: add configuration for XFRM interfaces → https://git.io/JvJ1h
<Christian[m]1>
does anyone else have the issue that pcscd needs a restart in order to get the yubikey working with gpg-agent ?
<lewo>
Christian[m]1: yes! but no idea on the why :/
Chiliparrot has joined #nixos
<lewo>
It only appears _sometimes_ when my laptop wakes up
eoli3n_ has joined #nixos
Tucky has quit [Quit: WeeChat 2.7]
<Christian[m]1>
yeah, the thing is pcscd prints a bunch of error messages that it cannot add the yubikey after restart, but then it just works, so not much to go on
<cinimod>
If I try to use `python37.withPackages(ps: with ps; [ numpy matplotlib seaborn google.cloud.storage ])` then the "." is interpreted as a nix separator(?) How do I tell nix that "." is part of the name?
tmaekawa has quit [Client Quit]
<simpson>
cinimod: Are you sure that that's the name? I see a different name:
<rauno>
When defining a new service, how to configre so it won't be automatically restarted by nixops deployment? Or is there some why to have a policy on it ?
<lewo>
cinimod: ps."google.cloud.storage"
<cinimod>
simpson: thanks when you use it in python, it called google.cloud.storage so I naively assumed it would be the same in nix
Acilim has joined #nixos
eoli3n_ has joined #nixos
<simpson>
cinimod: That's not even the case without Nix; for example, the popular package "Pillow" gives the module name "PIL".
<hyper_ch2>
so, notebook still can't boot even when using latest git master
<denbrahe[m]>
I have a secondary source for my derivation, fetched through fetchTarball, in which I need to modify a file, before referring to it in the main derivation. Using sed in `postPatch` of the derivation gives me `PermissionDenied`, as the nix store is read-only. Are there better ways to do this then to manually curl the 2nd source?
<wedens[m]>
make this secondary source a derivation
<denbrahe[m]>
a derivation that internally uses fetchTarball and has a postUnpack hook?
<wedens[m]>
fetchurl and postUnpack hook (or patch in patches or whatever)
chagra_ has quit [Ping timeout: 260 seconds]
<denbrahe[m]>
Ok, will try. Thanks!
chagra has joined #nixos
FelschR has quit [Remote host closed the connection]
<ajs124>
yorick: I don't remember and I don't really want to look at the code right now. But I'm sure it broke something. I would have tried to upstream it (at least as a draft/wip) if I didn't believe it's bad and shouldn't be used.
<dalto5>
Sorry for the beginner question but what is the difference between the unstable-nixos and the unstable-nixpkgs channels? Based some random searching, they always seem to return the same results but I am guessing they are different in ways I am not searching for.
<clever>
dalto5: nixos-unstable goes thru extra testing, to make sure it wont brick a nixos machine
<dalto5>
ah....thanks!
<yorick>
ajs124: shallow clones seem worthwhile for most usecases
<clever>
dalto5: few years ago, grub began corrupting the grub config file, and nixos-unstable correctly stopped updating
<clever>
dalto5: for the next 3 weeks, every nixpkgs-unstable user came out of the wood-work, asking why the machine stopped booting :P
chiefgoat has quit [Ping timeout: 268 seconds]
<dalto5>
So, do the packages just move there after a certain period of time or is the process manual?
<clever>
dalto5: both nixos-unstable and nixpkgs-unstable point to whatever version of master has most recently passed a certain set of tests
wavirc22 has quit [Ping timeout: 260 seconds]
<clever>
the difference is just which tests they check for
<ajs124>
yorick: true. Idk why upstream hasn't implememted it. I sometimes get the impression @niksnut might not like fetchGit that much. Maybe because it allows for impure git inputs, idk.
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<dalto5>
thanks clever. This is a pretty interesting distro but the learning curve is substantial even though I have been using linux for 25 years. On the other hand, getting to learn new things it what keeps it fun for me.
<clever>
dalto5: one of my most recent projects, is making heavy use of the cross-compile stuff in nixpkgs
<{^_^}>
[nixpkgs] @mbbx6spp opened pull request #77674 → nixos/doc+manual: update copyright year range end 2019->2020 → https://git.io/JvJ9X
<clever>
dalto5: for example, it took ~2 days to add a new cpu arch to nixpkgs, and then i was able to pull together 3 different cross-compilers (vc4, arm-embeded, arm-linux) trivially, and start building stuff for the rpi
<das_j>
clever: Have you found a solution for cross-building derivations with the same hash as they would have when building natively?
<clever>
das_j: that not really feasible, because the hash depends on which compiler you used
<dalto5>
Hmm....so I have a laptop that needs a newer kernel. After a little research, I tried setting boot.kernelPackages = pkgs.linux_latest; This yielded the result: error: attribute 'extend' missing, at /nix/var/nix/profiles/per-user/root/channels/nixos/nixos/modules/system/boot/kernel.nix:40:31 so it looks like I missed something. Can anyone point
<dalto5>
me in the right direction?
o1lo01ol1o has joined #nixos
nexgen has joined #nixos
<andi->
dalto5: you want to use pkgs.linuxPackages_latest;
chenhq2005 has joined #nixos
tabasko has quit [Remote host closed the connection]
<{^_^}>
#48740 (by midchildan, 1 year ago, open): mirakurun: init at 2.13.0
<{^_^}>
#49413 (by midchildan, 1 year ago, open): epgstation: init at 1.6.2
turion has quit [Quit: Leaving.]
Jarva[m] has joined #nixos
nexgen has quit [Ping timeout: 272 seconds]
<Jarva[m]>
Has anyone experienced their network device ID changing before?
<Jarva[m]>
My network device has seemingly randomly changed from wlp1s0 to wlp2s0
<clever>
Jarva[m]: what does lspci say about that card?
<clever>
for me, wlp3s0 is 03:00.0 Network controller: Intel Corporation Wireless 8265 / 8275 (rev 78)
<clever>
so the 3, means its on pci bus #3, and the 0 means its device zero on that bus, 03:00
<Jarva[m]>
What provides lspci?
<pie_[bnc]>
reverbtank: oh, which work infra is that then?
<clever>
,locate bin/lspci
<{^_^}>
Found in packages: toybox, busybox, pciutils
<clever>
Jarva[m]: pciutils, and it would help to have an lspci from both wlp1s0 and wlp2s0
<Jarva[m]>
Which is preferred? pciutils?
<pie_[bnc]>
reverbtank: anyway, we could try to see if the ad-hoc set up works if you want
<Jarva[m]>
How can I search for files within packages on local?
<clever>
,locate
<{^_^}>
Use ,locate <filename> to find packages containing such a file. Powered by nix-index (local installation recommended) https://github.com/bennofs/nix-index
<pie_[bnc]>
did you try the last thing i said in the thread, or did it even make any sense?
<Jarva[m]>
Cheers
<Jarva[m]>
Network controller is 02:00.0 but previously it was wlp1s0
<reverbtank>
pie_[bnc]It is used for our infrastructure that runs a pretty sizeable web app. Nixops is used for all the machines which are hosted on AWS. I did not try the last steps you listed out in the bug as I was unsure what the specific pactl commands were that you were running
<Jarva[m]>
Now 01:00.0 is an Unassigned class [ff00]: Realtek Semiconductor Co., Ltd. RTS525A PCI Express Card Reader (rev 01)
<Jarva[m]>
Not sure how it would change
<clever>
Jarva[m]: it might be a firmware problem with the motherboard, re-allocating bus's
<Jarva[m]>
Running on a dell xps 7390
<Jarva[m]>
I have firmware updates on, could that change it?
<reverbtank>
pie_[bnc] I'm going to try it now. First i'll kill the jack, jack-session, and then pulseaudio processes
<clever>
Jarva[m]: updating the firmware could change hte order of devices
<reverbtank>
pie_[bnc] I've confirmed all those services are not running by checking with ps aux | grep SERVICENAME. Is the next step to start jack in non realtime mode with qjacktl OR run the pactl commands?
alex`` has joined #nixos
chiefgoat has quit [Client Quit]
<pie_[bnc]>
reverbtank: i *think* it didnt really matter, but id try starting jack first
chiefgoat has joined #nixos
<pie_[bnc]>
reverbtank: this is what my qjackctl log looks like in the beginning https://bpaste.net/ISRQ
<pie_[bnc]>
reverbtank: you can look at the systemd logs for pulse with sudo journalctl _SYSTEMD_USER_UNIT=pulseaudio.service
gustavderdrache has joined #nixos
<pie_[bnc]>
reverbtank: for the record, even though we're starting stuff manually i bet theres some configuration files that are still being used, so the contents of the system configuration probably matter
chenhq2005 has joined #nixos
<reverbtank>
pie_[bnc] There are no entries shown when checking the logs for the pulse audio service. My qjackctl log looks like this. It appears there is a failure. https://bpaste.net/V3NQ
<pie_[bnc]>
reverbtank: looks fine to me
<pie_[bnc]>
from line 17
<pie_[bnc]>
im not really sure about the memory lock error thingies, but i also get them
reallymemorable has joined #nixos
<pie_[bnc]>
so its at least not a fatal error
<reverbtank>
pie_[bnc] seems to get past whatever the socket error is. Any suggestions for config files to nuke
nexgen2 has joined #nixos
<pie_[bnc]>
i didnt mention the config files because they should be nuked, just that those also influence what we're doing probably
<pie_[bnc]>
i barely have any idea how most of this works
reallymemorable has quit [Client Quit]
<pie_[bnc]>
reverbtank: so now try starting pulse, and run the script that adds the pulse source and sink dvices
reallymemorable has joined #nixos
<reverbtank>
pie_[bnc] Hmmm, should I run that yet? At this point in the setup, i think an application should be able to connect to Jack as an output device without having to setup pulse sources and sink.
<pie_[bnc]>
reverbtank: oh right
<pie_[bnc]>
yeah i guess
<reverbtank>
pie_[bnc] Just tried with VCV Rack and still doesn't show a valid output device
nexgen has quit [Ping timeout: 272 seconds]
<pie_[bnc]>
reverbtank: i mean...if pulse works than at least we know it works, otoh vcv rack might be simpler
<reverbtank>
pie_[bnc] Ran that command and it appears to have succeeded. However if I ps aux | grep pulseaudio I still see processes running
<pie_[bnc]>
well you still have to kill it once
<pie_[bnc]>
or something
<pie_[bnc]>
systemctl stop pulseaudio.service
<pie_[bnc]>
or pulseaudio --kill
<pie_[bnc]>
*shrug*
<pie_[bnc]>
add --user to the systemctl command
<reverbtank>
pie_[bnc] Ok, that successfully killed Pulse. I re ran Rack and got the same error. Checked again with ps aux | grep pulseaudio and it was back. Looks like Rack is causing pulseaudio to start up again. I'm thinking I'll temporarily remove pulseaudio from my config, rebuild, and see if that resolves anything. Doing now....
<pie_[bnc]>
reverbtank: if ou stopped the ssocket it shouldnt be coming back up >_>
<pie_[bnc]>
weird
<reverbtank>
pie_[bnc] Yea, and I removed all the script commands from qjackctl
selfsymmetric-mu has quit [Remote host closed the connection]
<pie_[bnc]>
ive been running those manually so idk about that part
xkapastel has joined #nixos
<pie_[bnc]>
reverbtank: are you on an oooold version of nixos by any chance
nexgen2 has joined #nixos
<reverbtank>
pie_[bnc] 19.03
<pie_[bnc]>
reverbtank: what version is your vcvrack
<pie_[bnc]>
well, tell me if you manage to fight through the pulse stuff
tomberek has joined #nixos
<pie_[bnc]>
reverbtank: well, so the current hypothesis is that the file descriptor error is because pulse was grabbing the sound device, but idk
<pie_[bnc]>
also thats kind of weird becuase then i think jack should have complained about it too
<pie_[bnc]>
*shrug*
nexgen2 has quit [Ping timeout: 260 seconds]
<pie_[bnc]>
one way to confirm that would be the fuser command, which i dont know how to use off the top of my head
<reverbtank>
pie_[bnc] So I set the pulseaudio config to false in my nix config file and rebuilt. Rebooted the system. Confirmed that there are no pulseaudio processes running. Started jack with qjackctl. Booted up VCV Rack, got the same error message. At this point, no pulseuadio on the system. So we are getting closer. Seems to be a strictly jack issue at the
<reverbtank>
moment
philr has quit [Ping timeout: 258 seconds]
<{^_^}>
[nixpkgs] @Mic92 opened pull request #77677 → openssl: fix build linux with clangStdenv → https://git.io/JvJQr
<jared-w>
gchristensen: Noice. That's my favorite kinda evil
<gchristensen>
:)
<gchristensen>
I don't want to do evil, but there isn't really a better way afaik
<infinisil>
gchristensen: How about declaring options and deciding what to do based on those
<infinisil>
Or using networking.hostName
<pie_[bnc]>
reverbtank: might be easier to debug this with some other app but i dont really have ideas
drakonis has joined #nixos
<infinisil>
`config = mkIf (networking.hostName == "foo") { <some config for that host here> }`
<gchristensen>
no-can-do
Synthetica has joined #nixos
<jared-w>
You can also grab environment variables or files or some other stateful-system-thingy and use that to check stuff right? Might be possible to find something that's already on the system somehow
<gchristensen>
this is generating the host's configuration before it is handed off to a customer
<gchristensen>
just about everything is dynamic, and then I must hand back a single .nix file which describes how to manage the system
<reverbtank>
pie_[bnc] Upgrading my system to 19.09 right now. Then I'm going to remove libjack2 and jack2 seeing as how you don't have them in your config. Perhaps that is causing a conflict
<infinisil>
Huh, a single nix file :/
<infinisil>
That's disappointing
<gchristensen>
it is instead of the hardware-configuration.nix
<jared-w>
and all of it is dynamic in a "I don't know what their configuration is before I give them the nix file" way? Or in a "I get their configuration and have to give them a single nix file" sort of way?
<pie_[bnc]>
reverbtank: i kind of doubt it, more like i probably need to add them to my config (or not)
<pie_[bnc]>
i mean its not like that starts any more services or something
nexgen2 has joined #nixos
<pie_[bnc]>
but yeah upgrade might help..
<jared-w>
gchristensen: comedy option: Write a CLI tool that generates the hardware-nix file :p
<gchristensen>
I am!
<infinisil>
gchristensen: This might also be problematic, because the module system imports is a DAG. You can introduce duplicates unless you remove those
<gchristensen>
that is what this is :P
<gchristensen>
infinisil: how do you mean?
<jared-w>
> mfw bash
<infinisil>
gchristensen: The same file/module can be imported from multiple different modules
<infinisil>
I guess in most cases it won't be a problem though
<gchristensen>
right, this will never be imported more than once
<infinisil>
You can put all imports into a single flat list if you want to
<infinisil>
That shouldn't be a problem as long as you deduplicated multiple imports
<gchristensen>
when a customer requests a server, this tool runs and determines the class of the hardware, the disk layout, networking requirements, boot / hardware requirement, etc. some of them are dynamically generated .nix expressions, some of them come from pre-written versions. it then writes them out to a directory and flattens them in to a system.nix file
<gchristensen>
when the server is ready, nixops copies the system.nix file off and records it as the physical spec
<{^_^}>
[nixpkgs] @tobim opened pull request #77682 → glog: remove static flag → https://git.io/JvJ5I
virus_dave has joined #nixos
Shell has joined #nixos
qyliss has joined #nixos
<pie_[bnc]>
reverbtank: you could try running jack_simple_client and see if anything happens, it just seems to generate a tone but be careful it might be loud
drewr has joined #nixos
kenjis1 has joined #nixos
<reverbtank>
pie_[bnc] Yup, that generates a tone
nexgen2 has quit [Ping timeout: 268 seconds]
dalto5 has quit [Ping timeout: 240 seconds]
domogled has joined #nixos
nexgen has joined #nixos
nexgen has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @jtojnar opened pull request #77683 → doc: Make callout marks & prompts unselectable → https://git.io/JvJ5C
<infinisil>
So you can define `mine.binalias.foo = "echo foo"` in any module and it adds an executable called foo that does that
<infinisil>
(plus a shell alias)
sigmundv has joined #nixos
<{^_^}>
[nixpkgs] @ruuda opened pull request #77687 → python3Packages.sentry-sdk: add missing test dependency → https://git.io/JvJFt
dm9 has joined #nixos
Lears has joined #nixos
[Leary] has quit [Read error: Connection reset by peer]
gxt has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @danielstaleiny opened pull request #77688 → Add Robomongo keyword so users can find it by searching 'robomongo' → https://git.io/JvJFR
<{^_^}>
[nixpkgs] @jtojnar merged pull request #77674 → nixos/doc+manual: update copyright year range end 2019->2020 → https://git.io/JvJ9X
<mananamenos>
hi, in my postgresql service i have initialScript. I want to add more instructions there, but after rebuilding it does not take effect (meaning initialScript has just executed once ever). How do i achieve this via configuration.nix?
<reverbtank>
pie_[bnc] Ok, small bit of progress. In my nix config, I set all my audio dependencies to pull the unstable versions. These were unstable.qjackctl, unstable.libjack2, and unstable.jack2. Rebuilt my config, now jack starts correctly and VCV rack can connect and output audio.
<reverbtank>
Using pasuspender, pulse audio gets properly suspended, but doesn't recover after I stop jack via qjackctl
qyliss has quit [Remote host closed the connection]
Shell has quit [Client Quit]
<reverbtank>
Interestingly, Reaper and Helm still do not see valid Jack outputs. Wondering if this has to do with those programs not being built with the correct jack dependencies
xkapastel has quit [Quit: Connection closed for inactivity]
Shell has joined #nixos
qyliss has joined #nixos
chloekek has quit [Quit: WeeChat 2.6]
<evanjs>
Drawing a blank. Is there an easy way to reference rustPlatform, or more specifically, buildRustPackage via the moz overlay?
nakkle has quit [Ping timeout: 246 seconds]
<infinisil>
mananamenos: initialScript isn't intended for that. You can instead add your own service that gets executed as many times and whenever you need it to
<nDuff>
Hmm. I'm accustomed to NixOS systems having `/usr/bin/env`, but with an installer SD image built off current unstable, I'm not seeing it.
<clever>
nDuff: nixos should create that, on bootup
<nDuff>
clever, thank you; knowing I need to look at the activation script to figure out what's wrong is itself helpful.
Neo-- has quit [Ping timeout: 258 seconds]
kenjis1 has quit [Remote host closed the connection]
Naptra has quit [Remote host closed the connection]
kenjis1 has joined #nixos
<nDuff>
...okay, running ``/run/current-system/activate``, everything comes up just as it should; the question is why that didn't happen automatically on boot.
kenjis1 has quit [Remote host closed the connection]
<judson__>
Can I succeed here? Should I instead e.g. add an unstable channel tarball to my configure.nix and get the alacritty from there?
<nDuff>
clever: ahh, yes, indeed. This happened on first boot from freshly-written SD card media; it looks like somewhere there's logic that resizes the filesystem to fill the media (which is a cute trick), but that hadn't happened yet when the activation script was running and it was hitting errors related to the filesystem being full. Reboot and everything's happy.
<nDuff>
judson__: I'm not familiar with alacrity in particular, but in general, it's very possible to pull a single derivation (and its dependencies) from a non-default channel.
<lewo>
judson__: (sorry for the stupid question) did you override both src.sha256 and cargoSha256?
drakonis1 has joined #nixos
<judson__>
lewo, totally reasonable question: I did override both.
<{^_^}>
[nixpkgs] @c0bw3b pushed commit from @mweinelt to release-19.09 « piper: 0.2.903 -> 0.2.905 (#77551) »: https://git.io/JvJb1
<jared-w>
Current dilemma: I want to install vagrant + a plugin inside a nix shell, but that vagrant plugin should be from git rather than just "vagrant plugin install X"
<aswanson>
does anyone know of a good article or reference for how to override a python package within a shell.nix file? I'm looking to use a specific version of cx_oracle and my assumption is that I should be able to override the src attribute of the cx_oracle default.nix, right?
<jared-w>
Have yet to figure this out and it's driving me absolutely nuts.
<jared-w>
In theory all I need is a *.gem file and then vagrant plugin install foo.gem (specifically, the plugin is vagrant-ovirt4)
hyper_ch has quit [Ping timeout: 268 seconds]
domogled has joined #nixos
mananamenos has quit [Ping timeout: 272 seconds]
<lewo>
judson__: it seems you are right :(
janneke has joined #nixos
arno01 has joined #nixos
<judson__>
lewo, Presumably, rBP emits an overrideable package? It's just that the attribute to override isn't obvious?
hlavaty has quit [Remote host closed the connection]
takeda has joined #nixos
<judson__>
(one of these days, I'm going to add tofu = "<sha256 length worth of 0>" to a nix config and forget about it.)
<EsperLily>
fwiw you can use `lib.fakeSha256` for that
<EsperLily>
which looks visibly different than embedding an actual sha256
<judson__>
That's worth a lot!
sigmundv has quit [Ping timeout: 258 seconds]
<judson__>
What's it look like?
<srhb>
judson__: 52a0 is my goto :P
<EsperLily>
it looks like `lib.fakeSha256`
<EsperLily>
I meant in source it looks different
<judson__>
Oh, I see what you mean.
<EsperLily>
it's just a value set to a string full of zeroes
<EsperLily>
there's also constants for the other hash types too
<Taneb>
judson__: my mnemonic is it's the same number as the number of cards in a deck
<judson__>
Taneb, that's handy
opthomasprime has joined #nixos
<Taneb>
It certainly suits some types of brain
<jD91mZM2>
Hi! nixos-rebuild switch --upgrade just killed my X server, AMA... In all seriousness though, does anybody know if there's something I can do to help debugging this issue?
arno01 has quit [Remote host closed the connection]
<buckley310>
jD91mZM2 what DE are you using?
<buckley310>
this happens to me occasionally on gnome. i have theories.
<jD91mZM2>
buckley310: I'm using the bspwm window manager, no full DE.
<jD91mZM2>
"Process 6514 (X) of user 0 dumped core."
<jD91mZM2>
- systemctl status display-manager
<jD91mZM2>
maybe it wasn't restarted then, like I thought it was
<buckley310>
using a login manager?
<jD91mZM2>
LightDM
<buckley310>
:/ not sure then. ive never seen a WM under lightdm get killed by a switch
<buckley310>
i would suggest running your rebuilds inside of a tmux session, and then if the X crash happens, re-attach to tmux and see which services were restarted, and then debug from there
<jD91mZM2>
Yeah, I suppose. I was quite taken by surprise
<opticnerve>
maybe biased answers anyway: is it NixOS ok for daily usage as general purpose desktop os?
<samueldr>
maybe biased answer: yes
<{^_^}>
[nixpkgs] @vcunat pushed 5 commits to release-19.03: https://git.io/JvJAj
lsix has quit [Quit: WeeChat 2.6]
<HedgeMage>
opticnerve: Probably, depending somewhat on the user.
<opticnerve>
which is the biggest change for a debian user?
<HedgeMage>
opticnerve: I have a particularly varied set of use cases in that I'm a software engineer and infosec engineer/analyst turned CISO, so I need to simultaneously support a wide range of technical tools and a set of executive tools that were designed for people who don't really understand their computers...I can do both well on NixOS, which is why I migrated from Gentoo in the end. However, I know
<HedgeMage>
nothing of the use case from a gamer's or designer's perspective, etc.
<aswanson>
opticnerve: you're going to spend a lot of time learning how to do things you already know how to do the Nix way
<aswanson>
opticnerve: the upside is that you'll suddenly find yourself making major changes to your system state that just kind of work, with a few minor edits to configuration.nix
<HedgeMage>
opticnerve: The configuration is going to be a big change for you... Debian strongarms some things where NixOS expects you to make choices. Also, depending on whether you've worked with any declarative-syntax DSLs before or not, you may find "the Nix way" of doing things foreign at first.
captn3m0 has joined #nixos
Jackneill has quit [Remote host closed the connection]
tertl3 has quit [Quit: Connection closed for inactivity]
<opticnerve>
clear. Last question: can u suggest me some study materials for Nix begginers so that if I install it i know where to put my hands?
<keithy[m]>
Are there any tutorials on how to create a yaml/json file with nix.
<infinisil>
keithy[m]: What's your use-case?
<infinisil>
Ah
<infinisil>
You can just use `builtins.toJSON`
<infinisil>
> builtins.toJSON { foo = "bar"; }
<{^_^}>
"{\"foo\":\"bar\"}"
<keithy[m]>
creating ignition files to instanciate fedora core os and load nix into it
wildtrees has quit [Max SendQ exceeded]
wildtrees has joined #nixos
<aswanson>
what's the major difference between python.withPackages and python.buildEnv? So far the nixpkgs documentation seems to treat them as interchangeable
maurer is now known as maruer
<infinisil>
aswanson: withPackages is just a small convenience wrapper around buildEnv
<{^_^}>
[nixpkgs] @erikarvstedt opened pull request #77693 → Fix Python test driver → https://git.io/JvJx1
<infinisil>
aswanson: This is one way to figure out where things come from ^
<infinisil>
That line then contains `withPackages = import ./with-packages.nix { inherit buildEnv pythonPackages;};`
<infinisil>
And `nix repl '<nixpkgs>'` can be used to call the unsafeGetAttrPos there
<aswanson>
infinisil: thanks for the thorough answer. I have a feeling you might have a good answer for this: I'm attempting to install python package cx_oracle, but the build is currently failing. I'm in no position to troubleshoot what is going wrong with the package, so I thought it would be simpler for me to override the version number and use an older working version. Can I do an override like that within a
<aswanson>
nix-shell expression?
mallox has quit [Read error: Connection reset by peer]
<infinisil>
aswanson: I'd start by writing a shell.nix file. Expressions can become ugly quick, and having that all in a command makes mistakes easy
<chrisaw>
This is a long shot but is anyone using Mosquitto Broker on NixOS right now? I cannot get it to work properly and I've checked and double checked all the configs - all look fine but messages are never delivered!
lordcirth has quit [Ping timeout: 268 seconds]
zeta_0 has quit [Quit: rcirc on GNU Emacs 26.3]
<cransom>
i am. no recent issues as far as i know as home-assistant is still assiting.
lordcirth__ has joined #nixos
dalto5 has joined #nixos
<{^_^}>
[nixpkgs] @grahamc merged pull request #77683 → doc: Make callout marks & prompts unselectable → https://git.io/JvJ5C
<chrisaw>
cransom: That's exactly what I want to use it for. Strange thing is - using "mosquitto_sub" and "mosquitto_pub" on the same box and the same topic and the message is seemingly "sent" but it never actually arrives. :S
<chrisaw>
Completely confused with it!
lordcirth_ has joined #nixos
<cransom>
are you watching for the exact same topic that you pub to? including a possible leading /?
lordcirth__ has quit [Ping timeout: 260 seconds]
dsx has quit [Quit: dsx]
<cransom>
there's also a wildcard you can use... i think it's #, so subbing to /# will give you eeeeverything.
<fresheyeball>
I am trying to get this page to work but for some reason nginx serves it with weird chars
<fresheyeball>
everything looks great if I serve it locally with nodejs
growpotk- has quit [Ping timeout: 268 seconds]
<{^_^}>
[nixops] @pSub opened pull request #1219 → Fix Travis build: Travis creates /etc/nix for us → https://git.io/JvJjh
<gustavderdrache>
firefox is warning me that there's no character set declared
<gustavderdrache>
what happens if you add `<meta charset="utf-8">` to the top of the `<head>` tag?
hax404 has quit [Quit: WeeChat 2.6]
<chrisaw>
cransom: I have indeed, you can just use '#' for the topic too but still nothing. It's very strange. Like it's going through the motions but then just throwing it in /dev/null or something!
hax404 has joined #nixos
dsx has joined #nixos
<fresheyeball>
gustavderdrache++
<{^_^}>
gustavderdrache's karma got increased to 3
<chrisaw>
I can use mosquitto_sub with a public broker and get messages in and out of that j ust fine just seems to be my install which is silently discarding them for some reason.
<fresheyeball>
that did it
<fresheyeball>
weird
lovesegfault has joined #nixos
<gustavderdrache>
my guess is the node module was defaulting to utf-8 in the headers, whereas nginx isn't
<cransom>
chrisaw: possible that you ahve an acl set, or not enough acl set? i don't know what the reponse is if you try to send to a topic you don't have permissions for.
zeta_0 has joined #nixos
<chrisaw>
cransom: I've tried the most open configs I could find to the most restrictive. Doesn't seem to make any difference strangely. :/
<lordcirth_>
chrisaw, sounds like a question for the mosquito community
chagra has quit [Ping timeout: 260 seconds]
felixfoertsch has quit [Ping timeout: 246 seconds]
<{^_^}>
noonien: Provide as much info as possible on problems. What specifically doesn't work? What command did you run? What's the error? This makes debugging easier and helps getting faster responses
<noonien>
a friend is trying to build a nix package to be included into his configuration.nix as an overlay, i want to test that .nix file
pikajude3 has joined #nixos
<noonien>
and i want to use the same format that packages in nixos uses, this is why i need callPackage
pikajude3 is now known as pikajude0
<noonien>
i would also like to just run `nix-build`, so that i don't have to change my configuration.nix
<pikajude0>
so, what does nixos-rebuild do if you don't have a nixos channel set?
<pikajude0>
which one does it pick?
<noonien>
pkgs.callPackage seems to have worked, i just had a typo
<pikajude0>
i logged in as root and added the 19.09 channel and suddenly everything on my system is getting rebuilt
ambro718 has quit [Quit: Konversation terminated!]
<wiml>
Is there documentation on `stdenv.mkDerivation` or `buildEnv` ? I can find some examples of their use in the docs, but nothing that lists all of the attributes they look at and what they do with them. I feel like I'm looking in the wrong docs or something
<chloekek>
I sometimes get Firefox being unable to load pages after resuming from stand-by.
<chloekek>
Other apps can connect to internet just fine without restarting them e.g. IRC client.
elenapan has quit [Quit:]
elenapan has joined #nixos
elenapan has quit [Remote host closed the connection]
pikajude0 has quit [Remote host closed the connection]
zeta_0 has quit [Quit: rcirc on GNU Emacs 26.3]
<philipp[m]2>
Has anybody gotten gnunet in a state where you could at least get your peerinfo from cli?
o1lo01ol1o has quit [Remote host closed the connection]
jonten has joined #nixos
alexherbo20 has joined #nixos
kenjis1 has quit [Remote host closed the connection]
zeta_0 has joined #nixos
alexherbo2 has quit [Ping timeout: 265 seconds]
alexherbo20 is now known as alexherbo2
Guest25175 has quit [Ping timeout: 268 seconds]
alexherbo2 has quit [Client Quit]
<zeta_0>
i am currently in the process of installing python in home.nix, is there a nix python `anaconda` package that i can install or do i have to install all of those packages separately?
<ddima>
medard: hm, more of a hunch, but have you tried the nomodeset kernel parameter during boot to see if that helps?
opthomasprime has quit [Quit: Leaving.]
OmnipotentEntity has joined #nixos
mehlon has joined #nixos
<mehlon>
so I just moved between partitions, can I somehow copy over my existing /nix/store from my old partition to my new one?
<mehlon>
or at least, use it as a binary cache so I can reinstall my stuff
mexisme has joined #nixos
<medard>
ddima, yes, didnt work :-/
<{^_^}>
[nixpkgs] @Alan01252 opened pull request #77701 → gns3: Fix python package not being pinned → https://git.io/JvUUg
erasmas has quit [Read error: Connection reset by peer]
<OmnipotentEntity>
There's a lot of symlinks going in and out of that directory, mehlon, I'm not sure it'll be that easy. But the good news is if you try it and it doesn't work, you can probably just undo it.
<mehlon>
hmm
<OmnipotentEntity>
I'm trying out updating to unstable, and I'm having a bit of trouble with the openshot-audio, it's failing to compile, and I believe it has something to do with the newer compiler. Is there a way I can instruct nix to use an older compiler from nixos-rebuild?
<ddima>
medard: from the top of my head: assuming the GUI installer generally works, which I havent verified; has any other linux booted, like tried arch or ubuntu to double-check? is it by any chance a rel new AMD cpu or so?
<medard>
ddima, ot
<medard>
ddima, it's first generation ryzen cpu, i have many different VMs running, some with graphical interface
Soo_Slow has joined #nixos
<ddima>
medard: wait, are you getting black screen trying to boot nixos and not on the physical machine itself?
<medard>
ddima, no i boot properly. Then i do sudo systemctl start display-manager and then screen goes black (on the VM)
philr has joined #nixos
<ddima>
sorry, dropped words in the sentence before "trying to boot nixos installer in a VM"
o1lo01ol1o has quit [Remote host closed the connection]
<ddima>
medard: do the logs say anything about the failed display-manager start say anything (if you maybe switch back to console)? I vaguely remember that with virtualbox for instance I had to pick some specific GPU device to get GUI to work, maybe thats sth to look into to, iirc there were not too many options.
Wizek has joined #nixos
<clever>
medard: there are also pre-made virtualbox images, with nixos installed already
<{^_^}>
[nixpkgs] @jwiegley merged pull request #77634 → Disable further pass tests that end up relying on gpg-agent on… → https://git.io/JvJCF
<{^_^}>
[nixpkgs] @jwiegley pushed 0 commits to johnw/58975: https://git.io/JvUUx
<{^_^}>
[nixpkgs] @jonringer pushed commit from @ruuda to master « python3Packages.sentry-sdk: add missing test dep »: https://git.io/JvUkm
<nDuff>
Hmm; this is bizarre. I'm building an image with ``services.sshd.passwordAuthentication = false``, but it's generating a sshd_config with ``PasswordAuthentication true``.
Jackneill has joined #nixos
<nDuff>
Running ``nix show-derivation -r`` on the resulting built shows me the generated code copying that string to a file, but it doesn't really do anything to explain *why* that's what's being generated.
<nDuff>
(err, s/built/build/)
<noonien>
can i pass packages from different channels to -p in `nix-shell`?
<nDuff>
oh. Needs to be `services.openssh`, not `services.sshd`. Surprised that that didn't get caught in validation.
<{^_^}>
[nixpkgs] @tobim opened pull request #77705 → nodePackages.clubhouse-cli: init at 2.1.0 → https://git.io/JvUk0
<nDuff>
cap_sens1: I happen to be logged into a Raspberry Pi 4 built against nixos unstable right now, and /dev/gpiomem exists. That said, it's using linuxPackages_rpi4, so I suppose that doesn't answer your "mainline kernel" question.
<cap_sens1>
nDuff: I see. Yesterday I was trying linuxPackages_rpi3, and I was unable to boot
<nDuff>
Ahh. I don't have any pi 3 hardware, so can't test personally.
<cap_sens1>
nDuff: But that was good info! I also have a rpi4, so maybe I can switch to that
medard has quit [Quit: Leaving]
kvda has joined #nixos
<cap_sens1>
nDuff: BTW, in rpi4, all 4GB of RAM are avaliable right?
<nDuff>
yup.
<nDuff>
well, the one I'm logged into right now is a 1GB unit, but tested the SD image previously in a 4GB one, and /proc/meminfo looked right.
<cap_sens1>
nDuff: Thanks!
m0rphism has quit [Ping timeout: 240 seconds]
<cap_sens1>
Another noob question: It seems that i3lock and xscreensaver are installed by default. Can I remove them/How to remove them declaratively in /etc/configuration.nix?
mexisme has joined #nixos
<noonien>
nixos on rpi??
<cap_sens1>
noonien: yes
<noonien>
woah, nice!
<noonien>
i gave it a try some time ago, i remember having some issues making it run, really nice that it works now!
<noonien>
did you just follow the wiki?
<cap_sens1>
Yes
iqubic has joined #nixos
<noonien>
great!
<cap_sens1>
But on rpi3, I can't use rpi foundation's kernel so the functionalityh is limited
<iqubic>
So, I did a silly thing. I just ran: `nix-env -iA nixos.gtk3-x11`
<cap_sens1>
I have to use linuxPackages_latest or linuxPackages
<noonien>
i've got rpi2, rpi3, and rpi4. i'm only interested in the rpi4 atm
<iqubic>
No I can't remove this package.
<iqubic>
*now
<cap_sens1>
Aha, nDull claimed that linuxPackages_rpi4 is working, so I guess it's good time
<iqubic>
And I wonder why that is.
<noonien>
is anyone able to run qtcreator?
pbb has quit [Remote host closed the connection]
<noonien>
i get this error when trying to run it: qt.qpa.plugin: Could not find the Qt platform plugin "xcb" in ""
Soo_Slow has quit [Quit: Soo_Slow]
pbb has joined #nixos
<tbenst[m]>
noonien: are you in a shell? If so can add `QT_PLUGIN_PATH = "${qt5}/${qt5.qtPluginPrefix}";`
<tbenst[m]>
but in general need `wrapQtAppsHook` in `nativeBuildInputs`
<noonien>
hmmm, i added `qtcreator` to systemPackages. should that not make it work?
<iqubic>
No. Not with this package.
<noonien>
i see, is there a specific reason? and maybe a wrapper?
<tbenst[m]>
on another note, I accidentally put a "[" in my nix.conf via nix.extraOptions on NixOS...now I can't do anything involving nix-store, and can't sudo to fix the read-only /etc/ filesystem...is there a way to fix this other than a live CD?
<nDuff>
tbenst[m]: consider bind-mounting a fixed file over the real one, then restarting the nix-daemon service.
<nDuff>
tbenst[m]: or you can boot into an older configuration.
<samueldr>
/etc/ is not actually read-only, tbenst[m], you could sudo mv /etc/nix /etc/_nix; sudo mkdir /etc/nix/; cat /etc/_nix/nix.conf | sudo tee /etc/nix/nix.conf; sudo vim /etc/nix/nix.conf IIRC
<samueldr>
(unless you made /etc/ read-only by some means :))
<nDuff>
that too, though the bind-mount is safer; nothing you need to undo later, just reboot and it's gone. :)
boolman has quit [Ping timeout: 258 seconds]
boolman has joined #nixos
<samueldr>
ah, I thought the symlink was on /etc/nix but it's directly on /etc/nix/nix.conf, so even easier
<samueldr>
and when you rebuild it will overwrite that custom nix.conf iirc
<samueldr>
since I believe I had to fix a boo-boo like that in the past
<iqubic>
so, I cannot for the life of me figure out how to undo the `nix-env -iA nixos.gtk3-x11` comman I ran.
felixfoertsch has quit [Quit: ZNC 1.7.4 - https://znc.in]
<iqubic>
Whatever I do to try and remove gtk3 doesn't work. I can't seem to remove it from the list that `nix-env -q` shows.
<tbenst[m]>
nDuff: neat! I had no idea about bind mounting..I just tried `sudo mount --bind /tmp/nix.conf /etc/nix/nix.conf`, and the file indeed changed per `cat`, but I still see `error: opening file '[': No such file or directory` on nixos-rebuild
das_j has quit [Remote host closed the connection]
ajs124 has quit [Remote host closed the connection]
Scriptkiddi has quit [Remote host closed the connection]
o1lo01ol1o has quit [Remote host closed the connection]
<iqubic>
I'm so pissed that I can't easily remove this package.
<simpson>
iqubic: If you have the /nix/store path, you can pass /nix/store paths directly to nix-env as if they were package names, including for removal.