<ShaRose>
Huh. At one point I thought I tried interfaces = []; and it failed, but now it's working. w/e, seems I can create a bridge and configure it using networking.interfaces.<bridge name>
<ShaRose>
that works for me.
ddellacosta has joined #nixos
_ris has quit [Ping timeout: 258 seconds]
selfsymmetric-mu has joined #nixos
jgt has quit [Ping timeout: 276 seconds]
vika_nezrimaya has quit [Ping timeout: 276 seconds]
justanotheruser has joined #nixos
jgt has joined #nixos
<ShaRose>
also, TIL if you define something more than once (as a list) it autocombines. That's very nice.
jgt has quit [Ping timeout: 246 seconds]
noudle has quit []
<notgne2>
clever: I still can't figure out how to test why these extra packages are depended upon, but I've narrowed it down to `overlays = [ (self: super: { glibcLocales = super.glibcLocales.override { allLocales = false; }; }) ];` which seems to make my docker image install some random packages I can only guess it's grabbing from my actual PC's NixOS
<notgne2>
install, I assume there's something I'm not understanding in regards to overlays
<notgne2>
but even if I did something wrong, it seems very weird for this to happen as an image changing based off what I have installed seems unusual/impure
<clever>
notgne2: if you build the top-level drv, with nix-build, can you then compare the size of 2 of them with nix-store -qR ?
dbyte has quit [Read error: Connection reset by peer]
dbyte has joined #nixos
kleisli has joined #nixos
<notgne2>
if you mean to run that on the `.build.toplevel` directory, all the store paths there seem correct
<jb55>
is there any way to get a list of already built dependencies for a failing build (one that hasn't produced an output yet) since I can't use nix-store -q --references ...
<ShaRose>
it combines strings too: very helpful for config like quagga and dhcpd
<clever>
notgne2: if you build the toplevel twice, with and without that overlay, how do the paths differ?
<notgne2>
it seems dbus-1.12.12-lib, util-linux-2.33.1-bin, perl5.28.1-Net-DBus-1.1.0, nixos-system-nixos-19.03.173495.7339bd47600, and glibc-locales-2.27 have changed hashes, and nothing else has changed (but its a bit hard to tell since |sort doesn't really help due to hashes coming first)
<clever>
notgne2: that would imply its not adding anything new, which you have claimed it was
freeman42[NixOS] has joined #nixos
<notgne2>
yeah and if I look at the actual paths inside the resulting docker image, there is nothing new, but the docker image itself is way bigger, and untarring it shows a bunch more nix store paths
kleisli has quit [Ping timeout: 268 seconds]
<clever>
notgne2: can you pastebin the nix expr?
jgt has joined #nixos
jgt has quit [Ping timeout: 276 seconds]
jluttine has quit [Ping timeout: 276 seconds]
<worldofpeace>
what's hydra's log limit size?
<worldofpeace>
or perhaps where can this be found?
slack1256 has joined #nixos
freeman42[NixOS] has quit [Remote host closed the connection]
andi- has quit [Remote host closed the connection]
spacefrogg has quit [Quit: Gone.]
aw has quit [Quit: Quitting.]
aw has joined #nixos
jgt has quit [Ping timeout: 276 seconds]
wfranzini has quit [Remote host closed the connection]
spacefrogg has joined #nixos
wfranzini has joined #nixos
Ralith_ has joined #nixos
dansho has joined #nixos
andi- has joined #nixos
<dansho>
i'm trying to run nixos-rebuild boot --install-bootloader from nixos-enter, but it says "System has not been booted with systemd as init system (PID 1). Can't operate."
<clever>
dansho: can you pastebin the output of `fdisk -l /dev/sda` ?
<dansho>
blkid shows /dev/sda4 is PTTYPE="dos"
<dansho>
it's not the graphical live boot unfortunately
<dansho>
i only have console
<clever>
dansho: if you mount sda4 to the nixos boot folder (/boot if your still inside enter, /mnt/boot if your outside), what files are inside that dir?
Supersonic112 has joined #nixos
Supersonic has quit [Disconnected by services]
Supersonic112 is now known as Supersonic
<dansho>
mount: /mnt/boot: wrong fs type
<clever>
dansho: what exact args did you use with mount?
<dansho>
# mkdir /mnt/boot; mount /dev/sda4 /mnt/boot
<clever>
dansho: when you do `fdisk -l /dev/sda`, what is the disk label type?
<dansho>
'Extended'
<clever>
dansho: disk label type, not partition type
<dansho>
oh /dev/sda1 (swap) has a '*' in the Boot column
<dansho>
Disklabel type: dos
<clever>
dos partition types cant UEFI boot
<clever>
so /boot is optional, and sda4 is an extended partition, you cant mount sda4
freeman42x has joined #nixos
mexisme has quit [Read error: Connection reset by peer]
<clever>
freeman42x: how did you write the iso to the usb stick?
<freeman42x>
clever, Unetbootin
MinceR_ has joined #nixos
<clever>
freeman42x: that breaks everything, use plain dd
<iqubic>
Is there anything I should know about doing Java development on Nixos? Can I just install Intellij and use that?!?!
<ivan>
iqubic: yes
<freeman42x>
clever, are you sure?
<clever>
freeman42x: yes
<iqubic>
ivan: Thanks.
<clever>
freeman42x: the iso is already a bootable usb image, just dd it directly to the root device of the usb stick
<freeman42x>
clever, ok, I'll try that ty
drakonis has quit [Quit: WeeChat 2.4]
MinceR has quit [Ping timeout: 244 seconds]
<ShaRose>
oh my god. I just wasted like an hour of time trying to figure out why radvd wasn't working before I noticed my test VM was connected to my testing network that I thought I'd gotten rid of
<ShaRose>
fml
<clever>
dansho: with the current partition layout, you cant EFI boot, and must do legacy boot, which the configuration.nix is already setup for, you just need to make sure the firwmare is going to legacy boot, not efi boot
abathur has joined #nixos
freeman42x has quit [Read error: Connection reset by peer]
dbyte has quit [Read error: Connection reset by peer]
jgt has joined #nixos
<dansho>
clever: hmm, in my BIOS it shows boot is set to [Legacy Only]
growpotk- has joined #nixos
nash8192[m] has joined #nixos
jgt has quit [Ping timeout: 246 seconds]
<mdash>
Hmm. Is there a reasonable way to grant the chroot created by `nixos-enter` access to my X display?
<mdash>
I guess I could do it over TCP but I don't feel like reconfiguring my X server if I don't gotta.
pi4 has joined #nixos
pi2 has quit [Ping timeout: 245 seconds]
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
<ShaRose>
ok, not quite done with my config but hopefully it shows what I was looking for...
<ShaRose>
a while ago I was asking about passing arguments or something to an import so it could modify based on it. I'm not sure what the nix-like way to do it is, but here's what I mean.
<ShaRose>
yeah, but almost all the options require some modification between hosts
<ShaRose>
like for the qemu one, there are 8 lines that could be 'common'. all of the extraconfig for dhcpv6 and the interfaces are unique based on what IP range I want to use.
* mdash
seriously contemplating patching nixos-enter now
<eyJhb>
Is the build bot down as well, or just {^_^} ? (GrahamcOfBorg)
user____ has quit [Ping timeout: 258 seconds]
<etu>
> "a"
<etu>
{^_^} is in here, so I'm testing :p
<eyJhb>
{^_^}: is 100% down for a couple of days now :p
<etu>
yeah, I haven't seen any action from the bot. But I think the irc client was missing as well before
<etu>
It probably queues everything up and will respond when someone found where things got stuck ;)
ng0 has joined #nixos
LnL has joined #nixos
rauno has joined #nixos
LnL is now known as Guest55267
jgt has joined #nixos
<eyJhb>
Infinite loop? :D
<eyJhb>
Who operates {^_^} ? gchrholiday ?
Guest16 has joined #nixos
aszlig has quit [Quit: Kerneling down for reboot NOW.]
Guest55267 has quit [Quit: exit 1]
<FireFly>
etu: idly reminds me of a friend's IRC bot, which had some kind of eval functionality. someone issued it something a bit mean... and then it sort of stopped responding. then, a month later when everyone had forgotten all about that incident, it started replying to month-old invocations...
<eyJhb>
If it responds to everything, then I think I have given two people karma just to show it doesn't work :p
LnL7 has joined #nixos
aszlig has joined #nixos
dansho has quit [Ping timeout: 276 seconds]
Guest16 has quit [Remote host closed the connection]
pikajude has joined #nixos
snajpa has quit [Ping timeout: 246 seconds]
snajpa has joined #nixos
aither has quit [Read error: Connection reset by peer]
Tucky has joined #nixos
aither has joined #nixos
Jackneill has joined #nixos
qqlq has joined #nixos
dermetfan has joined #nixos
lewo` has joined #nixos
pie_ has quit [Ping timeout: 245 seconds]
sigmundv has joined #nixos
freeman42x has quit [Ping timeout: 276 seconds]
lewo` has quit [Remote host closed the connection]
karetsu has joined #nixos
civodul has joined #nixos
maralorn has quit [Quit: WeeChat 2.4]
maralorn has joined #nixos
boxdot has left #nixos [#nixos]
veske has joined #nixos
graf_blutwurst has joined #nixos
<fgaz>
Hi, I'm trying to run nix on a nfs4 filesystem, but I keep getting "error: preallocating file of 12935052 bytes: Not supported". Is there a way to tell nix not to use preallocation?
hyper_ch2 has quit [Remote host closed the connection]
karetsu has quit [Quit: WeeChat 2.4]
phreedom_ has joined #nixos
karetsu has joined #nixos
domogled has quit [Quit: domogled]
karetsu has quit [Client Quit]
sigmundv has quit [Remote host closed the connection]
phreedom has quit [Ping timeout: 260 seconds]
lewo` has joined #nixos
mexisme has joined #nixos
lewo` has quit [Remote host closed the connection]
graf_blutwurst has quit [Remote host closed the connection]
o1lo01ol1o has quit [Ping timeout: 240 seconds]
o1lo01ol1o has joined #nixos
zupo has joined #nixos
<hyperfekt>
How do I get the path of a derivation as a path instead of a string? I want to check whether a file is contained in another derivation but pathExists is impure and I can't construct a string to pass to it anyway because any string I construct is not allowed to refer to store paths.
Myhlamaeus3 has quit [Ping timeout: 246 seconds]
nakkle has quit [Quit: Leaving.]
<immae>
hyperfekt: `builtins.pathExists (your_derivation + "/file/in/derivation")`? There is nothing impure in here, it’s evaluated at compile time
nakkle has joined #nixos
<immae>
(but you have a kind of 'recurrence' in your derivation where you need to build a derivation before evaluating another one, I don’t remember the exact term)
o1lo01ol1o has quit [Remote host closed the connection]
<sphalerite>
since the bot is currently broken :(
o1lo01ol1o has joined #nixos
<immae>
Thanks sphalerite
<hyperfekt>
immae: that just gives me "string '[storepath]/file/in/derivation' cannot refer to other paths"
<immae>
hmm
<immae>
I use it with ./. instead of your_derivation, I don’t see how it could be different?
<hyperfekt>
If I prepend with ./. + I get "a string that refers to a store path cannot be appended to a path"
<hyperfekt>
The problem is that it becomes a string when I append one instead of staying a path :/
ubert has joined #nixos
<sphalerite>
hyperfekt: what are you actually trying to achieve?
<hyperfekt>
sphalerite: IFD conditioned on the file actually existing.
<sphalerite>
hyperfekt: no, on a broader scale
alex`` has joined #nixos
<hyperfekt>
sphalerite: build hie for a specific nixpkgs version. currently trying to do it with all-hies.
<hyperfekt>
sphalerite: all-hie defines custom overrides for some but not all ghc versions, using such a conditional import.
lord| has quit [Read error: Connection reset by peer]
lord| has joined #nixos
user____ has joined #nixos
logzet has joined #nixos
alex`` has quit [Quit: WeeChat 2.6]
alex`` has joined #nixos
veske has joined #nixos
dycan has quit [Read error: Connection reset by peer]
ajirx__ has joined #nixos
ajirx has joined #nixos
ajirx_ has quit [Ping timeout: 240 seconds]
<amitshah>
hey folks. I'm wondering if there are custom patches in the kernel builds to make NixOS run on EC2. We at AWS are rolling out new Xen versions to some of our fleet, and are noticing NixOS failing during early boot, whereas other flavours of Linux distros work just fine.
<amitshah>
I'm about to look at the patches in the git repos, but if someone has handy knowledge of any custom patches in place specifically for EC2, that'll be great.
<ikwildrpepper>
amitshah: as far as I know we do not use ec2 related kernel patches. the nixos image are generated from: nixpkgs/nixos/maintainers/scripts/ec2/amazon-image.nix
<ikwildrpepper>
that would be a good starting point to look at what is in the image. If you need any help, let me know.
elgoosy has joined #nixos
veske2 has joined #nixos
veske has quit [Ping timeout: 245 seconds]
orivej has joined #nixos
rsoeldner has joined #nixos
<ikwildrpepper>
amitshah: looking at the patches that are applied by default, there seem to be some xen netfront related patches
__monty__ has joined #nixos
hyper_ch2 has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
mexisme has quit [Ping timeout: 245 seconds]
<ikwildrpepper>
amitshah: my previous comment only applies to kernel 4.4, all newer ones should not have those patches
<amitshah>
ikwildrpepper: OK, thanks!
<amitshah>
what's the newer kernel that's in use?
iyzsong has joined #nixos
<ikwildrpepper>
nixos 19.03 uses 4.19, I believe
orivej has quit [Ping timeout: 245 seconds]
<amitshah>
got it, thanks!
orivej has joined #nixos
palo has quit [Quit: WeeChat 2.4]
meck has quit [Quit: WeeChat 2.5]
<amitshah>
ikwildrpepper: so if there's nothing ec2-specific, it could also be some generic kernel patches that get applied. Any place to look for those?
meck has joined #nixos
palo has joined #nixos
<ikwildrpepper>
amitshah: in pkgs/top-level/all-packages.nix, search for linux_4_19
evanjs- has joined #nixos
evanjs- has quit [Client Quit]
user____ has quit [Quit: Lost terminal]
orivej has quit [Ping timeout: 276 seconds]
evanjs- has joined #nixos
gxt has joined #nixos
<ikwildrpepper>
amitshah: any log you can share, or more information on where it fails?
hyper_ch2 has joined #nixos
MmeQuignon has joined #nixos
<Taneb>
Any idea what could cause "nix-env -iA foo" to differ from "nix-env -f '<nixpkgs>' -A foo"? The latter seems to be using the system nixpkgs channel (ie. root's view of the world) rather than the user's
<evanjs>
So I wasn't sure if it wasn't being parsed correctly, or if repology just wasn't picking it up, but according to them, there are some issues on the nix side of things
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
b1000101 has joined #nixos
<amitshah>
ikwildrpepper: we don't have logs yet. The guess curerntly is that it's in early boot, and there's probably a triple fault which causes the isntance to crash.
<b1000101>
does anyone have any experience with maven projects? I'm trying to package https://github.com/torakiki/pdfsam but there's no documentation. I've tried mavenix with no luck so far.
orivej has joined #nixos
Vikingman has quit [Remote host closed the connection]
wfranzini has quit [Remote host closed the connection]
zupo has joined #nixos
orivej has quit [Ping timeout: 240 seconds]
zupo has quit [Client Quit]
orivej has joined #nixos
asymptotically has joined #nixos
Makaveli7 has joined #nixos
samrose has quit [Ping timeout: 276 seconds]
<hc>
win 32
johanot has quit [Quit: WeeChat 2.4]
<etu>
hc: yuck ;)
<CMCDragonkai>
Has anybody managed to get Davinci Resolve on NixOS?
<hc>
etu: too many irssi windows + not paying close attention to what I'm typing ;)
arahael2 has joined #nixos
arahael1 has quit [Ping timeout: 268 seconds]
Makaveli7 has quit [Quit: WeeChat 2.6]
orivej has quit [Ping timeout: 246 seconds]
noudle has joined #nixos
orivej has joined #nixos
wfranzini has joined #nixos
<eyJhb>
hc: del win 32 :D
veske2 has quit [Quit: This computer has gone to sleep]
<hc>
win 32 is #guix, btw ;)
<mpickering>
is there some testing framework where I can use nix to set up a different environment for each test?
<mpickering>
Using nix derivations is not that useful because `cabal` tries to write to ~/home
iqubic has quit [Remote host closed the connection]
ajirx has quit [Quit: Leaving]
<eyJhb>
hc: is there any advantages of Guix over Nix?
hyper_ch2 has joined #nixos
nspin has joined #nixos
<hc>
eyJhb: hmm, I haven't tested guix much yet. It feels less well maintained than nixos. I like that it's using scheme for everything and it feels pretty hackable. :) I mean to look more into it once I have hardware that is more compatible with linux-libre, which it uses
psyanticy has joined #nixos
<hc>
s/well maintained/mature/
Myhlamaeus3 has joined #nixos
rauno has quit [Ping timeout: 276 seconds]
veske has joined #nixos
emiller88 has joined #nixos
alex`` has quit [Ping timeout: 240 seconds]
alex`` has joined #nixos
<ng0>
on the OS side the biggest difference and disadvantages can be found in the init managers once you have to use either of them as a company
<ng0>
speaking from years of experience, but I never bothered to write the systemd vs shepherd comparison.
<ng0>
I'm tired of discussions which might lead nowhere, but shepherd feels neglected. if more than 1 or 2 people would work on it sporadically, it might get better than "rc.d, but worse". just my personal opinion
xkapastel has joined #nixos
abrar has quit [Quit: WeeChat 2.4]
aswanson has joined #nixos
<ng0>
on topic: is there any reason to believe we could get maintained support (in Nix) for systems supported by pkgsrc (FreeBSD, NetBSD, illumos family,...) if developers in these systems maintain functionality? I have interest on the NetBSD side, and I know someone in FreeBSD is working on getting support back. I know FreeBSD support existed once and was dropped. which is understandable, but does anyone remember
<ng0>
the timespan it was non-functional? For example, in another project we dropped win32 after 10 years of no developer stepping up to patch it back to functionality or report bugs. I'm hessitating to work on something where I don't know how upstream handles systems they don't target for themselves.
hyper_ch2 has quit [Remote host closed the connection]
<mdash>
b1000101: i packaged a maven thing recently
nakkle has quit [Ping timeout: 276 seconds]
nakkle has joined #nixos
<b1000101>
mdash: which one? have you used mavenix or some other tools?
endformationage has joined #nixos
<mdash>
wait... was it maven? No, it was gradle.
<mdash>
Not sure how I got those confused. Blame my hangover, I guess.
<b1000101>
mdash: No worries :)
<mdash>
I touch Java stuff infrequently enough that it all blurs together apparently
<mdash>
ng0: sounds like a good question to ask on the forum
<ng0>
mdash: ok
bbarker has joined #nixos
pie_ has joined #nixos
emacsomancer has quit [Read error: Connection reset by peer]
veske has quit [Quit: This computer has gone to sleep]
alex`` has quit [Quit: WeeChat 2.6]
hyper_ch2 has joined #nixos
<emily>
I've narrowed down my predictable device naming issue...
<emily>
Sep 16 14:04:39 patchouli systemd-udevd[1680]: eth0: Failed to rename network interface 2 from 'eth0' to 'enp4s0': Device or resource busy
<emily>
unfortunately, I'm not sure how on earth I can fix this ^^;
wfranzini has quit [Remote host closed the connection]
<emily>
I presume it's because I run sshd in the initrd, but I'm not sure why the initrd isn't using predictable device names from udev already.
ddellacosta has joined #nixos
Soo_Slow has joined #nixos
wagnerf has joined #nixos
<wagnerf>
hi, i've change channel recently and lost opengl in the process
<wagnerf>
all processes failed finding libGLX_indirect.so.0
<wagnerf>
pb is I have no idea how to get it back
<wagnerf>
which package should I install ?
<wagnerf>
I currently use nvidia legacy drivers
das_j has quit [Remote host closed the connection]
das_j has joined #nixos
<manveru>
wagnerf: did you reboot?
<wagnerf>
yes
<manveru>
what channel did you change to?
<manveru>
unstable is not working right yet with nvidia, at least last i tried...
<wagnerf>
I switched from nixpkgs-unstable to nixos-unstable
<wagnerf>
then back to nixpgks-unstable but did not get my drivers back
<wagnerf>
oh ok
emacsomancer has joined #nixos
das_j has quit [Remote host closed the connection]
<manveru>
personally i'm sticking with 19.03 still, so not sure
das_j has joined #nixos
Myhlamaeus3 has quit [Ping timeout: 240 seconds]
<wagnerf>
manveru: well, that's maybe a good idea
<manveru>
you can still use any software you want from unstable, but the OS itself should be stable imho :)
eyJhb has quit [Quit: Clever message]
<manveru>
esp for proprietary stuff, which isn't covered by any tests...
<wagnerf>
I will give it a try thanks
eyJhb has joined #nixos
iqubic has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
hyper_ch2 has joined #nixos
<wagnerf>
ok, reboot time
<wagnerf>
bye bye
cransom has joined #nixos
<hyper_ch2>
you're going to reboot the internet?
drakonis has joined #nixos
emiller88 has quit [Remote host closed the connection]
emiller88 has joined #nixos
<Ariakenom>
manveru: the drivers aren't covered by tests or all proprietary stuff isn't covered by tests?
wagnerf has quit [Ping timeout: 246 seconds]
<manveru>
Ariakenom: both, in my experience
<Ariakenom>
any particular reason?
<manveru>
because they're not built by hydra
zacts has joined #nixos
<manveru>
maybe not _all_ proprietary stuff, didn't check everything, but pretty sure nothing that can't be redistributed anyway will be built on hydra
<manveru>
like, we have chrome and the steam chroot, but not spotify or vbox guest addons
<mdash>
tyrion-mx: it means bind the name "args" to the attribute set passed to this function, but also bind the names "name", "runScript", etc to the values of the attributes in the set
vika_nezrimaya has joined #nixos
Fendor has quit [Ping timeout: 265 seconds]
<asymmetric>
is there some consensus on what to do when one doesn't want to maintain a package anymore?
shazow has quit [Read error: Connection reset by peer]
<asymmetric>
the package has no other maintainer unfortunately
Chiliparrot has joined #nixos
<manveru>
asymmetric: remove yourself from maintainers... maybe write on the forum about it? :)
<manveru>
at least that's what i did
<asymmetric>
manveru: yeah.. i think the package should at that point be removed, but i guess someone could chime in on the PR and decide to step in
pi2 has joined #nixos
<asymmetric>
what happened with your package? was it left with zero maintainers?
<manveru>
no idea, don't even remember which it was :P
<tilpner>
asymmetric: I don't think it should just be removed if other packages depend on it
<asymmetric>
tilpner: it's an end user app, so i don't think that's the case (electrum-ltc)
<suzanne_>
Hi! I installed Nix on top of a 'buntu machine and cloned github.com/NixOS/nixpkgs .
<suzanne_>
Where can I find a minimal ~/.nixpkgs/config.nix and how do I "enable" it (nix-env [something] ~/.nixpkgs/config.nix I guess)?
<manveru>
should probably remove myself from some other stuff i never use anymore...
<suzanne_>
Preferably my config.nix would use the local clone of nixpkgs. The manual seems mostly about a standalone NixOS install, not much about a user/guest install.
<devalot>
suzanne_: One way to control where nixpkgs lives is via the NIX_PATH environment variable. In this case, something like: NIX_PATH=nixpkgs=/path/to/nixpkgs
<tilpner>
suzanne_: I haven't had a config.nix in a long time, but: "{}" (without quotes) is a minimal config.nix, and would be active by default
<suzanne_>
devalot, tilpner: thanks!
Ariakenom_ has joined #nixos
wfranzini has joined #nixos
Vikingman has joined #nixos
Ariakenom_ has quit [Remote host closed the connection]
Ariakenom has quit [Ping timeout: 265 seconds]
<Vikingman>
someone can help i need to set custom resolution for my screen i managed to find the xrandr command i need to input
<Vikingman>
but i need to make it permanent
<tilpner>
vikingman: I have no idea if it'll work, but there's services.xserver.resolutions
<eyJhb>
Thinking about adding `extraCommands` to the ZNC module, as I am finding it quite useful, if one wants to configure additional modules. Any thoughts?
<tilpner>
infinisil: ^
<Vikingman>
all this because the auto config does not work properly with my monitor
<infinisil>
eyJhb: What's that?
<infinisil>
Like, what would that option do
<Vikingman>
anything like it for refresh rate
<eyJhb>
infinisil: the same as e.g. `networking.firewall.extraCommands`, just a bunch of shell commands
<infinisil>
eyJhb: Okay but where would they run? What for?
<tilpner>
vikingman: Oh, services.xserver.xrandrHeads might be relevant to you too
<eyJhb>
infinisil: run at the end of the `preStart` script, e.g. in my case, I would use them to configure my NickSrv module, with the correct command+password
<Vikingman>
will try 1st with resolution if it fixe
<infinisil>
Ah
<infinisil>
eyJhb: I have planned to extend the znc module with support for per-module settings
sigmundv has quit [Ping timeout: 276 seconds]
<eyJhb>
I cannot think of a better way of doing this, except making a specific extra option, that takes "a lot" more into account, and that would ruin the beauty of the simple `.config`, as it would need to not just parse/translate, but also if I hit THIS key after X, then I should try to do THIS
iqubic has quit [Remote host closed the connection]
<eyJhb>
infinisil: Do it, do it, do it!
<eyJhb>
:D
<eyJhb>
This is the simple and lazy way of doing it. How would you extend it further?
<infinisil>
eyJhb: For now, if you can really set up the module configs with a shell script, you can do `systemd.services.znc.preStart = mkAfter "the shell commands";`
<eyJhb>
I have very very mixed feelings about what you are saying infinisil
<infinisil>
eyJhb: znc module config is done with a key-value-like file in the modules folder. So there could be an option for setting values of these files
<eyJhb>
Because if that works, I will throw myself out my apartment, as I have been sitting thinking about solutions for ... three hours I guess
<infinisil>
Haha
sk8forether has joined #nixos
<eyJhb>
So mkAfter will append it to the preStart script, right?
<infinisil>
Yup
<eyJhb>
I will make a list soon, and if I cannot come up with a somewhat good solution in 15 minutes, I will contact every name on that list. You are at the top now. :p
<tilpner>
infinisil: Wait, why are you copying it in the first place?
<tilpner>
infinisil: My mumble module just sets group and allowKeysForGroup
<infinisil>
So I can change the permissions, a symbolic link wouldn't allow that
<eyJhb>
I cannot link it to you, but I what I have started to do is create a seperate group, and then in my acme config allow that group to read the cert. And then just point it directly to it
<tilpner>
(I did steal your postRun just now though)
<infinisil>
Hm maybe that's smarter :)
<eyJhb>
Then you can have a cert and certGroup for each thing, and only allow specific users/groups to read it (just make them a member of the group) :)
<eyJhb>
And then just symbolic link them around, or point directly at it! Not sure what is best..
<eyJhb>
For znc.pem I think ln -s would be the better choice :/
xkapastel has quit [Quit: Connection closed for inactivity]
<tilpner>
eyJhb: You too, why copy or link at all?
<tilpner>
I don't use znc anymore, but my old module does the same (group and allowKeysForGroup)
<tilpner>
Just point znc at it: SSLCertFile = /var/lib/acme/znc.${config.networking.primaryDomain}/fullchain.pem
<eyJhb>
tilpner: avoids the confusion of the znc.pem, that will always be there
emiller88 has quit [Ping timeout: 245 seconds]
<eyJhb>
As the module creates it if it does not exists
<eyJhb>
:/
<tilpner>
Ehh, okay
<eyJhb>
Best reason I can give you, but I like just pointing to it
<eyJhb>
LIke your example tilpner
<eyJhb>
Any preference in how to create a multiline file?
<eyJhb>
WOndering if the modules will crash if the .registry files are readonly
wfranzini has quit [Remote host closed the connection]
<Soo_Slow>
can nix package manager be installed on distro without systemd? (devuan/void/etc)
<leah2>
we have nix on void
gxt has quit [Ping timeout: 260 seconds]
<infinisil>
Soo_Slow: Yeah, Nix doesn't depend on systemd at all
<Soo_Slow>
nice, thanks
zacts has quit [Ping timeout: 258 seconds]
civodul has quit [Quit: ERC (IRC client for Emacs 26.3)]
growpotkin has joined #nixos
Fendor has joined #nixos
<eyJhb>
tilpner: you need to specify three params, for it to read the other cert, rather than znc.pem, as far as I understand from here - https://wiki.znc.in/Signed_SSL_certificate
zacts has joined #nixos
elgoosy has quit [Remote host closed the connection]
zacts has quit [Client Quit]
<tilpner>
You can, don't have to
zacts has joined #nixos
zacts has quit [Quit: WeeChat 2.4]
eyJhb has quit [Quit: Clever message]
eyJhb has joined #nixos
ubert has quit [Quit: Leaving]
eyJhb has quit [Client Quit]
knupfer has joined #nixos
emacsomancer has quit [Read error: Connection reset by peer]
rauno has joined #nixos
cransom has quit [Quit: WeeChat 2.4]
knupfer has quit [Quit: knupfer]
knupfer has joined #nixos
<asymmetric>
i have a systemd unit that does a wget, and it always fails with "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... failed: Name or service not known."
<asymmetric>
and the same wget invocation from the shell succeeds, so it seems to be some kind of network isolation for the systemd unit (which is activated by a timer)
nexgen has joined #nixos
drakonis has quit [Quit: WeeChat 2.4]
mexisme has joined #nixos
emiller88 has joined #nixos
<asymmetric>
am i on the right track?
inkbottle has joined #nixos
<emily>
does anyone here have the ability to approve spamfiltered posts on the discourse...?
<tyrion-mx>
How do I find out if a package contains `libpangocairo-1.0.so.0`
<tyrion-mx>
is there some way?
<tilpner>
tyrion-mx: nix-index/nix-locate
<tilpner>
Suggests gnome2.pango.out for you
<tdeo>
,locate libpangocairo-1.0.so.0
cransom has joined #nixos
<tdeo>
well, it should be in pango
nspin has quit [Ping timeout: 260 seconds]
mexisme has quit [Ping timeout: 250 seconds]
Ariakenom has joined #nixos
dermetfan has joined #nixos
zacts has joined #nixos
<pikajude>
how do I disable the use of `builders` in my nix conf?
<tilpner>
pikajude: Try --option builders ""
<pikajude>
yeah, i tried that
<pikajude>
i get an error as follows:
<pikajude>
error: getting status of '/home/jude/confPath': No such file or directory
<tilpner>
nix show-config | grep confPath
<tilpner>
Does that have any results?
<pikajude>
no results
<pikajude>
show-config is cool though, i didn't know about that
<tilpner>
Is it actually confPath, or did you alter that?
<pikajude>
you mean the string confPath?
<tilpner>
Yes
<pikajude>
i typed what you typed as-is
jb55 has quit [Remote host closed the connection]
<pikajude>
nothing in nix show-config matches confPath
<pikajude>
oh
jb55 has joined #nixos
<tilpner>
Did you obfuscate it in the error, or is that what Nix actually complained with?
<pikajude>
that's what nix actually complained with
<pikajude>
if i was obfuscating it i'd put in a dollar sign or something to indicate a variable :D
<pikajude>
but this is from home-manager, and i checked the source code, so i think what it's doing is just screwing up the option passing
<pikajude>
home-manager is failing to pass the empty string arg to nix
<tilpner>
tyrion-mx: You will first need to populate the database
<tilpner>
tyrion-mx: I do that with: NIX_PATH=nixpkgs=channel:nixos-unstable nix-index
<tyrion-mx>
ok, I was missing the second step
<tyrion-mx>
thanks :)
<tyrion-mx>
out of memory :S
eyJhb has joined #nixos
<tilpner>
I would give you mine, but... the server is down for maintenance
johanot has quit [Quit: WeeChat 2.4]
eyJhb has quit [Client Quit]
<Soo_Slow>
does nix packages consume more space than more old (apt, pacman and such) package management systems? Coz I've heard it doesnt delete old versions of applications by default
<pikajude>
you can enable a service that auto deletes them
eyJhb has joined #nixos
<tilpner>
Soo_Slow: Yes, expect more disk usage (and more memory usage)
<tilpner>
Compared to apt, at least
<Soo_Slow>
tilpner, but for how much more?
<tilpner>
I hear pacman doesn't separate outputs, so nixpkgs might win there
<pikajude>
how much more what? space?
<Soo_Slow>
yeah
<slabity>
Soo_Slow: The reason for this is because Nix will not delete packages being used by any profile. Even profiles that are not in use
<pikajude>
for the same package, the same amount of space
<slabity>
The more profiles you have, the more space will be used
<pikajude>
but there will be multiple versions of that package in the store unless you garbage-collect
<tilpner>
One issue is that you need enough space to contain the entire system closure twice
<tilpner>
If you update infrequently, the updated closure can be entirely disjoint from your currrent one
<Soo_Slow>
but you can easily wipe all older versions of packages at once, if you need so, right?
<pikajude>
yes
<pikajude>
as long as nothing is using them
hc has quit [Ping timeout: 265 seconds]
<tilpner>
Yes, you can just run nix-collect-garbage
<pikajude>
nix-collect-garbage -d will delete old generations
<tilpner>
And that will clean up all currently unreferenced store paths
<Soo_Slow>
sounds kinda like ability to delete orphaned packages on pacman
<pikajude>
or you can manually pick which generations to delete, but i forget which
<tyrion-mx>
tilpner: what do you mean do give you yours? the index?
<pikajude>
how*
<tilpner>
tyrion-mx: Yes, it's just one file. I would expect it to be portable
<tyrion-mx>
ah, ok
hc has joined #nixos
<Soo_Slow>
another question. Does nix packages support third party repositories (like, you know - there are plenty of these for ubuntu, in case you want fancy themes or some niche software that isnt in mainline repo)? Or everything should be uploaded into the main one
<tilpner>
Soo_Slow: It's very easy to add things to nixpkgs, so even relatively niche packages live there
<tilpner>
Soo_Slow: But there is also the Nix User Repository, for package outside nixpkgs
boxdot has joined #nixos
<tilpner>
(We'll see how that survives flakes)
hax404 has quit [Ping timeout: 264 seconds]
<pikajude>
yeah, it's quite trivial to write your own packages and include them in your system config
<pikajude>
but you can also PR them in the github repository
<Soo_Slow>
tilpner, "very easy" - but do these get verified, in case person who added something had some shady intentions and modified software to perform unwanted actions?
<tilpner>
Soo_Slow: There is some reviewing for nixpkgs, none for the NUR
<pikajude>
well, someone has to read the PR
pi2 has left #nixos [#nixos]
<Soo_Slow>
tilpner, but does NUR feature score or something like that? Like, you know, AUR does (so packages, people upvote often, tend to be good. While newly added has score of zero and may worth double-checking before installation)
boxdot has left #nixos [#nixos]
<eyJhb>
tilpner: if you get the time, #68850 should be good-to-go :)
<tilpner>
Soo_Slow: There is currently no score, it wasn't considered AFAIK
domogled has joined #nixos
<Soo_Slow>
tilpner, eeeh, okay then. But is there, at least, commentaries, to see if some people has caught package-specific problems or something like that?
<tilpner>
eyJhb: There's not much I can do after my initial pass
<tilpner>
Soo_Slow: No, the NUR is still fairly new. It would probably be accepted if you implemented it. Also see #nixos-nur
<tilpner>
Soo_Slow: I share your concerns, and don't consume software from the NUR myself. It is still useful, even with that restriction, but I agree a reputation system for packages would be helpful
domogled has quit [Remote host closed the connection]
<tilpner>
(Though anything with larger adoption should probably live in nixpkgs for now)
<__monty__>
The NUR isn't popular enough to justify investing time in a UI yet imo.
<tilpner>
__monty__: Chicken and egg, with that attitude it's never going to become popular
<Soo_Slow>
tilpner, I see, thanks. Last question - I didnt quite understand about "some reviewing for nixpkgs". Is it mandatory, or unreviewed packages may still get there?
<__monty__>
Everything's chicken-and-egg if you look at it closely.
<tyrion-mx>
tilpner: what is the server that is down that you were talking about before? It would really help if I could use that tool. Now I am missing libatk-bridge-2.0.so.0 :D
<tilpner>
Soo_Slow: Trusted users have the ability to commit directly to nixpkgs without prior review. Pull requests (sent from trusted or untrusted users) are reviewed, and then merged by a trusted user
<Soo_Slow>
tilpner, but is it hard to become a trusted user?
<tilpner>
Soo_Slow: So for every PR, at least one trusted user looked at it, often more
<tyrion-mx>
I will try building the index, without even starting the graphical environment
jonaswouters is now known as zertox
<tilpner>
tyrion-mx: The server I would use to share the file with you. If you have another way of getting the 24MB file from here to you, I can do that
<tyrion-mx>
tilpnet, do you maybe have telegram?
hax404 has joined #nixos
_ris has joined #nixos
<tyrion-mx>
(I could give you my username there)
<tilpner>
I don't
<__monty__>
croc sounds like a great usecase for this tilpner, tyrion-mx.
<tilpner>
Soo_Slow: Hard is subjective. What would qualify as hard?
erasmas has joined #nixos
<tilpner>
__monty__: I'm not going to package that just for this
<__monty__>
And do you know how hard/easy it is to become an arch maintainer? (If that's what you're comparing to?)
<tyrion-mx>
lol
<tyrion-mx>
tilpner: maybe by email?
<__monty__>
Magic-wormhole then?
<tilpner>
tyrion-mx: Sure, send me a PM
<Soo_Slow>
tilpner, existance in community for a while, some background of bugfixes and other successfull contributions. Maybe known relations between account and real person who uses it (so if something will go wrong - there will be real person to blame, instead of just abandoned account)
<slabity>
Depends if you consider an AUR contributor an 'Arch maintainer'?
<tilpner>
Soo_Slow: If that is hard enough, then I guess it is hard. I can't seem to find the process description though
<tilpner>
infinisil probably knows what I'm talking about
<Soo_Slow>
I see, thanks
iqubic has joined #nixos
<tilpner>
You can't just say "hey, trust me", and get commit rights
<qyliss>
Nixpkgs is unusual in that being a committer doesn't give you the power to actually build packages, too
<tilpner>
qyliss: Do other distros do that? :o
<qyliss>
So, like, a Nixpkgs committer has waaaaaay less power than a Debian developer for example
<qyliss>
tilpner: Debian
<tilpner>
That is terrifying
<niksnut>
I think Debian is the exception
<qyliss>
I guess I could technically have done it in Homebrew too
<aanderse>
didn't debian just fix that recently?
<aanderse>
i thought i heard some chatter about them no longer allowing maintainers to upload binaries
jgeerds has joined #nixos
<niksnut>
could be
<aanderse>
but yeah... thats nuts, and if it was fixed, it was just fixed recently i'm pretty sure
<aanderse>
i think there are a number of people who want to remove the ability to commit directly to master and leave that up to PRs and bots, more than the number who explicitly don't want that
<aanderse>
maybe one day
<qyliss>
-1 from me on that
<tilpner>
Yes, I should really get to the RFC again
iqubic has quit [Ping timeout: 265 seconds]
rawtaz has quit [Read error: Connection reset by peer]
<aanderse>
qyliss: which bit? removing the ability to directly commit to master?
<qyliss>
yes
rawtaz has joined #nixos
growpotkin has quit [Ping timeout: 265 seconds]
<aanderse>
as someone who almost entirely commits via PRs i'm always interested in opinions that differ. would you care to share your use cases or ideas on that?
<qyliss>
Sure
m0rphism has quit [Quit: WeeChat 2.4]
<qyliss>
First, we nominally support patch review on Discourse, as a mailing list alternative
<qyliss>
Although in practice people submitting there are usually pressured into submitting on GitHub anyway, which is a big shame
<__monty__>
slabity: AUR maintainers aren't equivalent to nixpkgs committers.
<slabity>
Removing the ability to commit to master - Does that happen even locally? Can you set up a git repo to require a new branch even if it won't get pushed?
_ris is now known as ris
<qyliss>
slabity: no, you can do whatever you want with your own git repo
<slabity>
I would love to be able to set up a repo that denies local commits to the master branch
<qyliss>
why?
<__monty__>
qyliss: Yes, keep voicing my thoughts quicker than I can type. : >
<qyliss>
aanderse: second, committing to master is useful for stuff like fixing up PRs before merge, or making quick fixes
<aanderse>
qyliss: yeah, i sorta agree it is too bad that we require a github account... but i've seen patches come in on discourse where someone created a PR for the person who didn't have a github account
<__monty__>
qyliss: Committers can get automatic access to PR branches. Which is the proper way to do that imo.
<qyliss>
aanderse: and then that person can't participate in the PR review
<slabity>
qyliss: Stop my coworkers from accidentally commiting to master branch instead of a develop branch. It takes some time to fix and would be nice to be proactive in that regard
<qyliss>
__monty__: only if the person ticks a box when they open the PR, now?
Fendor has quit [Remote host closed the connection]
<__monty__>
slabity: Just lock down the remote branch. No need to do that locally?
<qyliss>
*no?
Fendor has joined #nixos
<averell>
you can have separate fetch and push remotes, and that way you could have different refs, so the push has no master branch to push to maybe?
<qyliss>
I would like to stop committers being able to make arbitrary new branches in the main repo
<slabity>
__monty__: Yea, but they need to fix it locally before it can be pushed.
m0rphism has joined #nixos
<aanderse>
qyliss: given that patches through mailing list is not the norm generally i think it is a small price to pay that all the work could happen over mailing list, then the PR could just be filled out instead of a direct commit (and benefit from ofborg tooling, etc...), but i haven't thought about that opinion too much, so it isn't a strong opinion
<slabity>
averell: I did not know you could do that. That might be a better solution
<qyliss>
aanderse: also, what if (like now!) ofborg is down?
<aanderse>
and as far as fixing up PRs before merge - if someone has a branch they want to merge into master do we lose the ability to modify that branch before it gets merged into master by the github merge button? i didn't think that was an issue
<__monty__>
qyliss: I guess. We have that ticked by default in ranger/ranger though. Ofborg could automatically request that permission if people don't tick the box, maybe?
<qyliss>
aanderse: they have to opt into that
<aanderse>
qyliss: mhm mhm, very valid point! (re ofborg)
<aanderse>
qyliss: right, i forgot about that
zacts has quit [Quit: WeeChat 2.4]
<qyliss>
Also, people will often make PRs from their master branches or something. I'm not comfortable pushing to those just because I want to modify commits for Nixpkgs
erba has joined #nixos
<aanderse>
qyliss: cool, i'm glad to hear these points as for the most part had never thought of them
<qyliss>
aanderse: finally, committers will definitely have to be able to push directly to release branches (which should be what most people are actually using), so what value does restricting pushes only to master even have?
<__monty__>
qyliss: I'm 100% certain you can make it opt-out at least.
<qyliss>
not good enough, especially given what I said about master branches
<__monty__>
qyliss: Why do you feel uncomfortable pushing to a master branch?
<erba>
Does anyone here have experience with working with projects that requires node/npm plus something like elm? There is elm2nix, node2nix but they both want to write to default.nix, ive been creating seperate .nix files for node and elm in the same directory but I donno if its the correct workflow?
davidak has joined #nixos
Jackneill has quit [Remote host closed the connection]
<tyrion-mx>
libatk-bridge-2.0.so.0 does not seem to be provided by any libs, but anyway some packages include it, how should I proceed? Still trying to make mailspring run :D
<aanderse>
qyliss: well my github-fu isn't strong enough to provide counter points to everything you said, but yeah i'm curious to see how things play out over the next few months with more and more discussion about bots, automation, and security. i hope you'll be available to provide opinions and ideas as those conversations happen :D
<__monty__>
,locate libatk-bridge-2.0.so.0
<tilpner>
tyrion-mx: at_spi2_atk
<tyrion-mx>
hah, I was looking only at the names between ()
zupo has joined #nixos
emiller88 is now known as Emiller88
eraserhd has quit [Ping timeout: 240 seconds]
zacts has joined #nixos
eraserhd has joined #nixos
dansho has joined #nixos
zacts has quit [Quit: WeeChat 2.4]
pikajude has quit [Quit: Leaving]
orivej_ has quit [Ping timeout: 265 seconds]
pikajude has joined #nixos
<cransom>
when updating a package, is there a preference to either mark as broken for a platform versus setting platforms.linux? systemd is a buildInput for datadog and that won't fly on darwin. i wouldn't mind looking into it, but i didn't want to scope creep a version bump at the moment.
Myhlamaeus3 has joined #nixos
pikajude has quit [Client Quit]
pikajude has joined #nixos
pikajude has quit [Client Quit]
pikajude has joined #nixos
<infinisil>
cransom: meta.platforms is for the platforms it's supposed to run on, broken for when the build doesn't succeed in certain conditions even though it should
<infinisil>
(well I haven't seen this written down anywhere, but this seems most reasonable)
Myhlamaeus3 has quit [Remote host closed the connection]
rsoeldner has quit [Remote host closed the connection]
rsoeldner has joined #nixos
<cransom>
then i will lean towards broken then i suppose.
lopsided98 has quit [Remote host closed the connection]
<PyroLagus>
you can still install it though, you'll probably have some duplicated unstable deps, which will take up some disk space
<rawtaz>
unstable stuff works great, ive done it many times, as a newbie
<zeta_0>
that's cool, i am currently running stable, what are the main differences between unstable and stable?
<PyroLagus>
sometimes stable updates more frequently than unstable, ironically
psyanticy has quit [Quit: Connection closed for inactivity]
zeta_0 has quit [Quit: `]
nexgen has quit [Remote host closed the connection]
tilpner has joined #nixos
waleee-cl has quit [Quit: Connection closed for inactivity]
mexisme has joined #nixos
jbgi has quit [Ping timeout: 265 seconds]
jgt has joined #nixos
knupfer has quit [Ping timeout: 250 seconds]
ToxicFrog has quit [Ping timeout: 276 seconds]
wfranzini has joined #nixos
jgt has quit [Ping timeout: 276 seconds]
drakonis has joined #nixos
Mateon1 has quit [Ping timeout: 240 seconds]
Mateon2 has joined #nixos
waleee-cl has joined #nixos
Mateon2 is now known as Mateon1
ToxicFrog has joined #nixos
knupfer has joined #nixos
Myhlamaeus3 has quit [Ping timeout: 240 seconds]
<tyrion-mx>
I am really getting crazy. I set up a chroot with all the required dependencies using buildFHSUserEnv, then I npm install stuff, it tries to run node-gyp, it runs g++ and it crashes saying that ld cannot find crti.o ... LD_LIBRARY_PATH contains it though
freeman42x has joined #nixos
<tyrion-mx>
any hints, what could it be?
ToxicFrog has quit [Ping timeout: 276 seconds]
ToxicFrog has joined #nixos
gchrholiday is now known as gchristensen
<devalot>
tyrion-mx: IIRC: when linking .o files, ld doesn't use LD_LIBRARY_PATH.
ee1943 has quit [Read error: Connection reset by peer]
<devalot>
That variable is used by the run time linker to find .so files.
<iqubic>
Let's say I have an IDE that I want to run which isn't in the Nixpkg. All I have is an executable file that needs the proper compiler in the path when run.
<iqubic>
How do I make sure that the compiler is in my path when I start the IDE.
<iqubic>
???
<tyrion-mx>
tilpner: I was confused by the .dev :)
<tilpner>
tyrion-mx: "package" is not really a concept Nix knows (for now)
<PyroLagus>
iqubic: the simplest one is to just install the compiler with nix-env or in your packages in configuration.nix
<PyroLagus>
if that'll be enough is a different question
<tilpner>
xorg.libX11 is a derivation with the outputs xorg.libX11.outputs ([ "out" "dev" "man" ])
<PyroLagus>
that is, if there's any library deps you need
justanotheruser has quit [Ping timeout: 265 seconds]
Jackneill has joined #nixos
davidak has quit [Quit: Leaving]
<Vikingman>
somebody can help me with services.xserver.xrandrHeads
<Vikingman>
i need this xrandr --newmode "1440x900_60.00" 106.50 1440 1528 1672 1904 900 903 909 934 -hsync +vsync
<Vikingman>
but im not sure how to achieve it
evanjs- has quit [Ping timeout: 258 seconds]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<zeta_0>
when is the new stable and unstable versions of nixos coming out?
Jackneill has quit [Remote host closed the connection]
iqubic has quit [Ping timeout: 258 seconds]
<qyliss>
zeta_0: new unstable as soon as master passes all its tests. On the order of a few days.
<qyliss>
zeta_0: by stable, do you mean the next set of 19.03 updates, or 19.09?
lejonet has quit [Ping timeout: 245 seconds]
__monty__ has quit [Quit: leaving]
inkbottle has joined #nixos
<zeta_0>
qyliss: the next set of stable and unstable versions, i heard a while back they are not going to be released until this october?
<qyliss>
I'm not sure I understand the questions, then. Do you understand what stable NixOS and unstable NixOS are?
<zeta_0>
yes i was just wondering the date of when the new versions are going to be released?
<qyliss>
Well, as I said, unstable will come up whenever the tests pass again
lejonet has joined #nixos
<qyliss>
I don't know about stable, but I think it'll just come out when the release managers decide it's ready
<zeta_0>
qyliss: sounds good, thanks for the heads up
<zeta_0>
qyliss++
turtlerabbit has joined #nixos
<qyliss>
zeta_0: the bot that keeps track of ++ is taking a break, unfortunately :)
<zeta_0>
oh
<tilpner>
tv: Hmm, now it's more trouble. Is there some escape hatch for disabling gixy locally?
<turtlerabbit>
Anyone using the "xmonad-with-packages" package? It's supposed to include xmonad-contrib but isn't found by my xmonad.hs
jgt has joined #nixos
alex`` has quit [Quit: WeeChat 2.6]
jgt has quit [Ping timeout: 245 seconds]
<tv>
tilpner: you could override pkgs.writeres.writeNginxConfig with e.g. pkgs.writeText
<tilpner>
tv: Hmm, that might work. I was hoping for a magic comment that tells gixy to ignore this location block, I do like the checking
noudle has quit []
<tv>
I don't think gixy supports selectively disabling some checks. If you do have legitimate configuration that doesn't work, I think it's best to create an issue upstream, and disable gixy until it gets fixed.
zeta_0 has quit [Quit: rcirc on GNU Emacs 26.1]
<tilpner>
tv: I'm probably just doing it wrong, I'm absolutely no expert with nginx
* tilpner
just can't figure out how to keep the rewrite without using $uri
drakonis has quit [Ping timeout: 258 seconds]
<tv>
I'm by no means an expert either, but you could paste the problematic part of your configuration, and I could give it a shot :)
<tilpner>
tv: Yes, that's what I tried first, but it breaks the rewrite
Ariakenom has quit [Quit: Leaving]
jlv has joined #nixos
nspin has quit [Ping timeout: 260 seconds]
<jlv>
If I am writing a systemd service, is it preferable to use `.serviceConfig` or `.script`, `.preStart`, etc? It looks like `.script` and the like serve the same purpose as `.serviceConfig`.
<qyliss>
Use the Nix wrapped ones (so not .serviceConfig) as much as possible
<qyliss>
they provide nice convencienes
<qyliss>
use serviceConfig only where you have to (which will be in quite a lot of places, so it's not a bad thing to do)
inkbottle has quit [Read error: Connection reset by peer]
fusion809 has quit [Remote host closed the connection]
<jlv>
Ok, makes sense.
<jlv>
Btw, is there a guide for writing NixOS services? I can't find anything in the manual.
Okinan has quit [Quit: Konversation terminated!]
mexisme has quit [Ping timeout: 245 seconds]
{^_^} has quit [Remote host closed the connection]
justbeingglad has joined #nixos
justbeingglad has left #nixos [#nixos]
Fendor has quit [Quit: Leaving]
<bsima>
jlv: i think your best bet is to find a similar service and copy stuff
jgt has joined #nixos
<tokudan>
i'm trying to fix kicad failing with a dynamically loaded lib: libngspice. if I set LD_LIBRARY_PATH to the correct path for that lib it works. can I easily add it to the wrapper created by this line? wrapProgram "$out/bin/kicad" "''${gappsWrapperArgs[@]}"
<jlv>
Can you refer to the name of a service when defining it? Like ${name}?
<tokudan>
jlv, what do you consider to be the "name"?
<jlv>
tokudan: `systemd.services.<name>`. The `name` part there.
<tokudan>
don't know
<tv>
tilpner: still ignoring the index.html rewrite, I think I have something that might work: replace your proxy_pass with $uri by rewrite ^(/.*)$ /paste$1 break; proxy_pass http://127.0.0.1:7000;
<tokudan>
jlv, maybe through some magic that I don't fully understand: nixos-option systemd.services.cjdns._module.args
<tilpner>
\o/ <3 tv
<{^_^}>
tv's karma got increased to 1
<tilpner>
rewrite ^/$ /paste/index.html break;
<tilpner>
rewrite ^(/[^/\s]+)$ /paste$1 break;
<tilpner>
That makes everything work, and should prevent the http splitting, probably
freeman42x has quit [Read error: Connection reset by peer]
<tv>
I did some testing using following server at port 7000: socat -vv TCP-LISTEN:7000,crlf,reuseaddr,fork SYSTEM:"echo HTTP/1.0 200; echo Content-Type\: text/plain; echo; echo test"