<leo60228>
ignore the collisions, i think i caused them by failing to install a qt4 library i was working on patching in a local nixpkgs
<clever>
leo60228: and since you did a channel update, what does `systemctl status docker` say now?
<leo60228>
Unit docker.service could not be found.
<clever>
leo60228: try turning docker back on by uncommenting the old config, and nixos-rebuild switch
<wilornel>
how do you extend a disk size on an aws ec2 with nixos? I'm trying growpart but it seems like the Filesystems listed are quite different from those in the aws ec2 docs examples
<clever>
wilornel: nixos will automatically growpart and resize for you on bootup, just reboot the machine after growing the ELB
<wilornel>
darn, then there's something wrong because I did the following: Shut down instance. Resize volume in the aws console. Boot instance. And then, when I run `df -h`, I see
<{^_^}>
→ 1b4027c6 by @flokli: abootimg: init at 0.6
<wilornel>
I'm not sure what nvme2n1 is and nvme1n1 and nvme0n1
<{^_^}>
→ ac716155 by @xeji: Merge pull request #39362 from flokli/abootimg
<clever>
wilornel: ah, i dont think nixos supports resizing the nvme drives currently, those are pretty new
<wilornel>
what are those
<clever>
wilornel: your hard-drives
<clever>
wilornel: oh, i think i see whats going on
<wilornel>
It's a kind of physical drive. I see. How come I am using that one? I guess changing it would be too much trouble? How could I get nixos to grow the filesystem?
<clever>
wilornel: you didnt resize it, you added another disk, you have 3 hard-drives, with sizes of 40gig, 80gig, and 8gig
<clever>
nixos is installed to the 40gig drive
<clever>
and the 80gig drive is unused
<clever>
you cant grow the nixos, because the drive its on is already full
<wilornel>
I had one with 40Gb, one with 50Gb, one with 8gb. I resized the 50gb to 80gb.
<wilornel>
Alright, I need to resize the 40gb one
<wilornel>
whoops
<clever>
the 80gig drive has no partitions on it
<clever>
its entirely unused
<wilornel>
how do you know?
<wilornel>
(darn )
<wilornel>
I think I had some work on there before aaah...
<wilornel>
now I need to mount it and the 8gb one
<clever>
wilornel: fdisk doesnt show any partitions on the 80gig one
<wilornel>
Oh that would be listed under "Device"
<wilornel>
?
<clever>
yeah
<wilornel>
ok resizing the 40gb one to 80gb
<clever>
nvme1n1 doesnt even have a disk label type
<clever>
so it has never been partitioned
<wilornel>
and after that's done I'll try to mount some things around..
<wilornel>
not been partitioned even before the resizing?
<clever>
correct
<wilornel>
Is it possible that it's because I booted nixos before the volume was done resizing?
<clever>
no
<PolarIntersect>
Okay, so I think the issue was that I was installing my EFI data onto a non-GPT partition?
<leo60228>
clever: do you know what happened?
<PolarIntersect>
Will that prevent booting the EFI disk?
<clever>
wilornel: ive had zero problems resizing things while the machine was running
<wilornel>
I mean I spun up the instance before the volume was done resizing
<wilornel>
Ah, I'll have to reboot it again.
<Dezgeg>
there were some quite recent fixes to ec2 disk image resizing
<Dezgeg>
maybe the AMI is older than that
<wilornel>
Why does it take so long to resize the disks?
freeman42x]NixOS has joined #nixos
<wilornel>
alright I see clever
Fare has joined #nixos
<clever>
PolarIntersect: yeah, efi firmware can only boot in efi mode from a gpt disk
rosa has quit [Quit: rosa]
xeji has quit [Quit: WeeChat 2.0]
<clever>
leo60228: i'm out of ideas
<PolarIntersect>
Got it
<PolarIntersect>
I wonder if nixos-install could warn about that to prevent this issue for others
<PolarIntersect>
Would it be a difficult addition?
<clever>
PolarIntersect: i think systemd-boot and grub should warn you
<PolarIntersect>
Yeah, they didn't
<PolarIntersect>
Well, systemd-boot did
<PolarIntersect>
but not grub :)
<Dezgeg>
I think some firmwares allow MBR booting as long as the partition type byte is set correctly
<PolarIntersect>
oic
<PolarIntersect>
Guessing mine isn't one of those, because it never showed up in the boot list
<Dezgeg>
but of course that's risking the luck
<PolarIntersect>
Although I did get into grub once and it just said "unknown filesystem", but I think that Grub was maybe a separate one on the MBR?
<PolarIntersect>
I'm guessing encryption.nix is the one you'll want to look at
pareidolia has quit [Ping timeout: 255 seconds]
<clever>
PolarIntersect: hardware.nix is also of interest
<PolarIntersect>
ok :)
<clever>
PolarIntersect: also line 28, the ssh service automatically allows port 22 in the firewall
<PolarIntersect>
ah, good to know
<clever>
you cant lock yourself out easily
<PolarIntersect>
Even if I explicitly set it to []?
<PolarIntersect>
It'll end up [ 22 ]?
<clever>
yeah
<PolarIntersect>
nice
<PolarIntersect>
thanks :)
<clever>
it concats the value of that option from every module
<clever>
so you actually set it to [ 22 22 ];
<PolarIntersect>
haha
pareidolia has joined #nixos
<clever>
PolarIntersect: everything looks fine, what does efibootmgr -v say?
<PolarIntersect>
clever: This same config is from a template that worked on 4 other machines I use as well :o
<PolarIntersect>
Added the efibootmgr one
Guest10968 has quit [Ping timeout: 260 seconds]
<clever>
PolarIntersect: Boot0001 doesnt look right, and why do i see a microsoft thing in Boot0000?
<PolarIntersect>
That's a separate disk w/ Windows installed on it
<PolarIntersect>
Boot0001 is a weird disk... I have no idea why it is even there, but it boots to Grub somehow
<PolarIntersect>
kinda annoying tbh lol
<PolarIntersect>
That's the one that I think I need to wipe the MBR on or something
<PolarIntersect>
but the nixos one isn't in the efibootmgr list
<clever>
PolarIntersect: if you do `blkid /dev/sd??` do you see the uuid from Boot0001?
<PolarIntersect>
clever: No, it doesn't even exist
<PolarIntersect>
oh woops I do!
<clever>
it looks like the efi vars are not updated correctly, so the firmware doesnt know about the new install
<clever>
which one is it?
<PolarIntersect>
/dev/sdb2 I guess
<ghostyy>
i tried downloading a nixos iso but when i mount it i just see a bunch of random garbage, how do i mount it so that i can chroot into it and complete my installation? i dont want to reboot into it
<clever>
PolarIntersect: can you add the `fdisk -l /dev/sdb` to the output?
<PolarIntersect>
seems like another Windows disk or something
<ghostyy>
oh its the squashfs thing
<clever>
ghostyy: the iso includes an MBR partition header, so you have to use the losetup flag for obeying partition tables, installing from a squashfs is also tricky
<PolarIntersect>
clever: done
<ghostyy>
this is so absolutely frustrating, why cant nixos installer just take a flag for "download everything instead of copying it from /nix/store"
<clever>
ghostyy: because the installer itself is a program that has to exist in /nix/store/
<clever>
PolarIntersect: that appears to be your windows drive
<PolarIntersect>
clever: Oh, yep
<PolarIntersect>
clever: So, I think the MBR for my windows drive just has an old grub written to it
* PolarIntersect
shrugs
<PolarIntersect>
shouldn't cause issues afaict
<clever>
PolarIntersect: id recomend you set boot.loader.grub.efiInstallAsRemovable = true; and then re-run nixos-install
aarvar is now known as typeastronaut
<clever>
PolarIntersect: and see what happens then
typeastronaut is now known as aarvar
<PolarIntersect>
clever: Shouldn't `boot.loader.efi.canTouchEfiVariables = true;` have let this work though?
lopsided98_ has quit [Remote host closed the connection]
<clever>
PolarIntersect: your firmware could be buggy
<PolarIntersect>
oic
<PolarIntersect>
okie
<PolarIntersect>
so that install finished
<PolarIntersect>
Should I reboot or does something else help?
<clever>
and /mnt/boot was mounted during that install?
<PolarIntersect>
oh hmm - it shows up in efibootmgr now
<clever>
and yeah, id try rebooting now
<PolarIntersect>
okie
fragamus has quit [Ping timeout: 260 seconds]
<PolarIntersect>
hmm ok
<PolarIntersect>
It showed up now but the root disk didn't work
<PolarIntersect>
but that at least is enough for me to test w/
<PolarIntersect>
thanks clever :o
<PolarIntersect>
I don't get why it works if I pretend it's a USB disk tho
* PolarIntersect
feels like you could probably hack a grub install to chainload Windows on a luks disk
<clever>
PolarIntersect: something i have been wanting to do, is to modify boot.ini to be able to chainload grub
<clever>
PolarIntersect: and i believe grub can be configured to support an ntfs "/boot" partition
<daveo>
clever: read the readme. That is neat.
* samueldr
shudders
<clever>
which means i could take ~4 files from the rescue_boot system, stick them in c:\, modify boot.ini, and then f8 my way to the nixos installer
<clever>
and nuke windows from orbit with wipefs!
<PolarIntersect>
haha
<{^_^}>
[nixpkgs] @grahamc pushed commit from @ryantm to release-18.03 « lib/eval-config: document NIXOS_EXTRA_MODULE_PATH »: https://git.io/vp4t7
<PolarIntersect>
Sounds pretty dope, but also means I won't have my games any more! :D
<clever>
xorg and gparted could be included in that "rescue" blob, and then you have gparted and the ability to shrink windows filesystems
mkoenig has quit [Ping timeout: 256 seconds]
pirateking has joined #nixos
<PolarIntersect>
that'd be awesome too
<daveo>
PolarIntersect: I don't have my games any more and now I am totally fine with that
<Fare>
Mic92_, thanks again!
<daveo>
PolarIntersect: Steam does actually work... but....
Supersonic112 has joined #nixos
Supersonic has quit [Disconnected by services]
migy has quit [Quit: migy]
<pirateking>
i changed my efi partition and now it has a new uuid. i can't boot through nixos anymore. ubuntu is broken too. is there a simple way to get nixos to understand the new uuid?
<clever>
pirateking: where does it fail when booting nixos?
<pirateking>
udev something
<pirateking>
some check
Supersonic112 is now known as Supersonic
<pirateking>
and that's after changing my hardware-configuration.nix to be correct*
<clever>
pirateking: you can probably change the init= in grub to init=/bin/sh to get a very basic single-user shell and attempt repairs, but it would probably be simpler to boot into the original installer ISO and repair it from there
<daveo>
I'm going through the Nix Pills and one of the commands is not a thing in NixOS 18.03
<justan0theruser>
why can environment.systemPackages use pkgs without "with pkgs"
pxc has joined #nixos
<daveo>
clever: why are some binaries out of ~/.nix-profile and others out of /run/current-system ?
<clever>
daveo: some are installed with environment.systemPackages and others with nix-env
<clever>
justan0theruser: line 11
<justan0theruser>
ty
<clever>
justan0theruser: which itself gets the default value from line 41
<clever>
justan0theruser: also, all of those options lack a type, which means it will get very buggy if 2 modules try to set the same option
rosa has joined #nixos
<daveo>
clever: some scripts expect things, like bash to be in certain places, for instance /bin/. If bash is in /run/current-system/bin that is close. What is the NixOS way to handle such situations.
<clever>
daveo: for the #! line, #!/usr/bin/env bash, for everything else, just do "bash" and let $PATH do its job
pxc has quit [Ping timeout: 240 seconds]
<daveo>
clever: sounds good to me. I suppose there is a certain amount of adaptation you have to do when people just assume #!/bin/bash etc will work. I've heard people say to do that so many times that it seems like people think that is the right way to do shebangs. I'm not here to tell them it is not, but this makes me wonder...
<clever>
daveo: /bin/sh and /usr/bin/env are the only 2 executables that are at "standard" paths
<daveo>
clever: it would be amazing if `nix-os --list-generations` could show details about the changes between generations, i.e. added [pkg a], deleted [pkg b]
<daveo>
clever: I've symlinked bash to a location inside /bin . probably discouraged I would guess
<acowley>
It's not obvious if it creates its own set of problems, but boy is the current situation a nuisance.
leothrix has joined #nixos
Piece_Maker has joined #nixos
Acou_Bass has quit [Ping timeout: 260 seconds]
Piece_Maker is now known as Acou_Bass
Fare has quit [Ping timeout: 240 seconds]
johnw has joined #nixos
acarrico has quit [Ping timeout: 264 seconds]
Sonarpulse has quit [Ping timeout: 240 seconds]
griff_ has joined #nixos
acarrico has joined #nixos
jgt1 has quit [Ping timeout: 264 seconds]
schoppenhauer has quit [Ping timeout: 264 seconds]
schoppenhauer has joined #nixos
johnw has quit [Ping timeout: 240 seconds]
jmeredith has quit [Quit: Connection closed for inactivity]
johnw has joined #nixos
rosa has quit [Quit: rosa]
hello_FRIEND has joined #nixos
hello_FRIEND has quit [Read error: Connection reset by peer]
acarrico has quit [Ping timeout: 248 seconds]
<unlmtd>
from nvim, the command `:! st` will launch an st term, but `:! st & disown` does nothing. The command works with termite and gnome-terminal. it also works from a plain st terminal window. what the heck!?
<unlmtd>
sorry, this probably has nothing to do with nix
<{^_^}>
[nixpkgs] @matthewbauer pushed 9 commits to master: https://git.io/vp43R
<{^_^}>
→ 8ffaeecf by @jokogr: clion: 2018.1 -> 2018.1.1
<{^_^}>
→ a52b5be4 by @jokogr: goland: 2018.1 -> 2018.1.1
<{^_^}>
→ e74d28ef by @jokogr: idea-community: 2018.1 -> 2018.1.2
klntsky has joined #nixos
oida has joined #nixos
<pirateking>
clever: not sure how to set init but i ended up using grep on everything to find where the actual value that's used for the uuid in boot. now i'm in
<colemickens>
I'm executing this. How can I get the hash of the attribute? (and what is the right terminology to ask this question)? nix-build '<nixpkgs/nixos>' --attr config.system.build.azureImage --argstr system x86_64-linux --out-link azure --max-jobs 10
pcr has quit [Ping timeout: 260 seconds]
<colemickens>
at the end the out-link 'azure' is a symlink to '<hash>-azure-image'. I want <hash> without using `readlink` or something on the out-link. I assume there's a `nix-something` command?
griff_ has joined #nixos
jaeckel has joined #nixos
MichaelRaskin has joined #nixos
simukis has joined #nixos
dx_ has quit [Quit: leaving]
dx_ has joined #nixos
orivej has joined #nixos
<Myrl-saki>
clever: I'm starting to get some segfaults on my build server.
<Myrl-saki>
clever: No signal 6 so far yet htough.
pxc has joined #nixos
dx_ has quit [Ping timeout: 264 seconds]
pxc has quit [Ping timeout: 260 seconds]
<kuri0>
Mic92_, when building the vm is there a way to use ccache ?
<kuri0>
so that when i rebuild its faster
griff_ has quit [Ping timeout: 240 seconds]
Huck777 has left #nixos ["WeeChat 2.1"]
dx_ has joined #nixos
rosa has quit [Quit: rosa]
jgt1 has joined #nixos
lonokhov has joined #nixos
stew has joined #nixos
nD5Xjz has joined #nixos
xy2_ has joined #nixos
dx_ has quit [Ping timeout: 240 seconds]
sigmundv__ has joined #nixos
nixname has joined #nixos
<sphalerite>
clever: so what do you want me to try?
jgt1 has quit [Quit: WeeChat 1.9.1]
<sphalerite>
Myrl-saki: fwiw probably a bit late now, but pkill -G nixbld
<sphalerite>
I suspect this is just another of the things that crop up when you run out of space though
nixname has quit [Quit: Page closed]
nixname has joined #nixos
<nixname>
can anyone tell me how long outputs remain in the binary cache, or what that depends on?
viric_ is now known as viric
<sphalerite>
nixname: in the official binary cache? So far, forever
liif has joined #nixos
nixname_ has joined #nixos
liif is now known as liff
<sphalerite>
nixname_: in the official binary cache, so far, forever
<nixname_>
that's great news! thanks for the info.
<sphalerite>
iirc it might get gc'd eventually, but for now there's no reason to since the bandwidth costs far exceed the storage costs, or something like that
<nixname_>
sphalerite: makes sense
nixname has quit [Ping timeout: 260 seconds]
sigmundv__ has quit [Ping timeout: 256 seconds]
<Mic92_>
kuri0: ccache for what?
<sphalerite>
ugh, stupid perl non-determinism
page has quit [Quit: leaving]
<sphalerite>
Also, /j #workingset
<sphalerite>
oops
furon has quit [Quit: Connection closed for inactivity]
NightTrain has quit [Quit: Lost terminal]
<nixname_>
Mic92_: i may not be using the right terminology, but i meant cache.nixos.org
<Mic92_>
nixname_: it uses the binary cache and your local downloaded assets when building vms
page has joined #nixos
<sphalerite>
nixname_: Mic92 was referring to an earlier question from kuri0, about ccache, not the nix binary cache :)
<nixname_>
ah - my apologies. sorry for the confusion
<sphalerite>
:)
<sphalerite>
So I'm getting path hash mismatches due to the appearance of files with names like _exit.2.gz~nix~case~hack~1
<sphalerite>
What causes the appearance of these files and how can I stop it from happening?
<sphalerite>
Remounting the store rw and changing the names back to what they're supposed to be fixes the hash mismatches
<Dezgeg>
fixed in 2.0.1 I think
<sphalerite>
ah, perfect
ihar has quit [Quit: WeeChat 2.0]
Harpalus has joined #nixos
Harpalus has left #nixos [#nixos]
<{^_^}>
[nixpkgs] @ookhoi opened pull request #39642 → netcat-openbsd-1.187-1-1.190-1 - set version from 1.187 to 1.190: → https://git.io/vp4nD
<jbo>
Hi there - so I'm trying to install docker, following https://nixos.wiki/wiki/Docker, but following nixos-rebuild switch --upgrade, neither the docker service is running nor is my user in the docker group. What's up?
<kuri0>
Mic92_, for building the vm
pie_ has quit [Ping timeout: 248 seconds]
<Mic92_>
kuri0: it is used.
vcunat has joined #nixos
<jbo>
aha! Looks like I had some dud entries in nixpkgs.config, which led nixos-rebuild to seem to execute, but actually it was silently failing :(
<{^_^}>
[nixpkgs] @Mic92 pushed commit from @ookhoi to release-18.03 « netcat-openbsd: 1.187 -> 1.190: (#39605) »: https://git.io/vp4cJ
<{^_^}>
[nixpkgs] @Mic92 closed pull request #39642 → netcat-openbsd-1.187-1-1.190-1 - set version from 1.187 to 1.190: → https://git.io/vp4nD
qknight has joined #nixos
<unlmtd>
im setting up a nix-channel. I have the `nixexprs.tar.xz` and `binary-cache-url` there, both reachable with wget, but `nix-channel --update` fails with `HTTP error 404 (curl error: No error)`
<unlmtd>
does it expect https?
maingo has quit [Quit: Connection closed for inactivity]
<{^_^}>
→ 1c235abd by @davidak: elementary-xfce-icon-theme: 2017-11-28 -> 0.11
asuryawanshi has joined #nixos
<{^_^}>
→ 6078f019 by @Mic92: Merge pull request #39639 from davidak/elementary-xfce-icon-theme-update
bpa has quit [Remote host closed the connection]
nixname_ has quit [Quit: Page closed]
sigmundv__ has joined #nixos
lhart has joined #nixos
lhart has quit [Client Quit]
<unlmtd>
Hallelujah! It works!
<sphalerite>
what was it?
pcr has joined #nixos
<kuri0>
Mic92_, i built the virtual machine and its stuck at mke2fs 1.44.1 (24-Mar-2018)
coot has joined #nixos
<pcr>
I'd like to make a derivation out of a URL that hosts a zipped binary. After fetchurl, unpackPhase fails because "sourceRoot doesn't contain a directory". So I changed phases to use a single installPhase script that unzips src to $out/bin. Is there a neater way to do this?
<kuri0>
also it didn't seem to use the custom kernel i told it to and i'm doing this on ubuntu with nix installed
<sphalerite>
pcr: if you're not using any other stdenv stuff then there's not really any reason not to take that approach. Although since you're not using the phases mechanism really you could just use buildCommand
jbo has quit [Quit: leaving]
asuryawanshi has quit [Remote host closed the connection]
<LnL>
pcr: add nativeBuildInputs = [ unzip ];
<kuri0>
if I put boot.kernelPackages = linuxPackages_4_14; in my nix vm config why do I get error: undefined variable 'linuxPackages_4_14'
<kuri0>
do i need to include something ?
<sphalerite>
kuri0: pkgs.linuxPackages_4_14
<kuri0>
sphalerite, undefined variable pkgs
<kuri0>
it says
<sphalerite>
kuri0: add it to the arguments at the top of the file
<sphalerite>
like { config, pkgs, ... }:
<kuri0>
ok
oida has quit [Remote host closed the connection]
oida has joined #nixos
<kuri0>
why does my vm hang at mke2fs 1.44.1 when booting ?
<jbetz_>
Which makes sense, but now how do I configure this ESP for dual boot?
<sphalerite>
just mount it on /boot in nixos and it should work
<sphalerite>
with os-prober and grub, or with systemd-boot, AFAIK
<kuri0>
jbetz_, but why do you have two ESPs
<kuri0>
its probably confusing the firmware / os
pxc has joined #nixos
<{^_^}>
[nixpkgs] @Mic92 pushed 6 commits to release-18.03: https://git.io/vp4W1
<{^_^}>
→ d68f32e4 by @etu: php70: 7.0.28 -> 7.0.29
<{^_^}>
→ ca82aa9b by @etu: php71: 7.1.15 -> 7.1.16
<jbetz_>
kuri0: oh, I don't anymore
<{^_^}>
→ 0ec977f5 by @etu: php56: 5.6.34 -> 5.6.35
<kuri0>
jbetz_, can you show the contents of your esp using tree then ?
<[Leary]>
systemd-boot should automatically load the windows efi file in the ESP without configuration or os-prober
<jbetz_>
kuri0: sure, once I get back into nixos. stalled by not having function keys to load the boot menu ATM.
hotfuzz_ has quit [Ping timeout: 240 seconds]
pxc has quit [Ping timeout: 240 seconds]
<rizary>
i'm using this `https://nixos.wiki/wiki/PostgreSQL` initialScript in my configuration.nix, but the role that created still doesn't exist
<kuri0>
I have a folder of kernel patches and i put it into pkgs/os-specific/linux/kernel/mypatches/ and i want to know how to write a default.nix in that folder that uses all the patches in the folder ?
Jetien has quit [Ping timeout: 260 seconds]
hotfuzz_ has joined #nixos
<jbetz_>
Damnit, windows boot loader isn't finding the nixos partition, so I don't know how to get in
<kuri0>
jbetz_, it isn't supposed to
<kuri0>
you need to have a bootloader which can boot nixos
<kuri0>
or maybe efi stub
<jbetz_>
Okay, so I how to I switch to one that can?
pie_ has joined #nixos
Arcaelyx_ has joined #nixos
<qknight>
what was the nox command for reverse depedency builds?
<sphalerite>
qknight: nox-review wip/pr ?
Arcaelyx has quit [Ping timeout: 265 seconds]
<sphalerite>
jbetz_: both grub and systemd-boot support it
<sphalerite>
jbetz_: and you already have one of those enabled, so it should just be a matter of running nixos-rebuild boot
<sphalerite>
oh, you might also need to set boot.loader.efi.canTouchEfiVariables = true;
<qknight>
sphalerite: thanks
<jbetz_>
But how do I do that from inside windows?
<sphalerite>
jbetz_: just mounting the partitions as appropriate and running nixos-install should do it
knupfer has joined #nixos
jacob_ has quit [Remote host closed the connection]
<kuri0>
I have a folder of kernel patches and i put it into pkgs/os-specific/linux/kernel/mypatches/ and i want to know how to write a default.nix in that folder that uses all the patches in the folder ?
<kuri0>
how could i do that
ixxie has joined #nixos
<sphalerite>
kuri0: you don't need to put it in nixpkgs, you basically just need attr sets containing `name` and `file` attributes
<sphalerite>
err, sorry, `patch`, not `file`
<sphalerite>
kuri0: have a look at pkgs/os-specific/linux/kernel/patches.nix for some examples
spear2 has quit [Ping timeout: 248 seconds]
<jbetz_>
sphalerite: getting this error from nixos-install: `File system "/boot" has the wrong type for an EFI System Partition (ESP)."
<sphalerite>
did you mount the EFI system partition to /boot?
<sphalerite>
or /mnt/boot
<jbetz_>
Hmm
<jbetz_>
To /mnt/boot
<sphalerite>
huh ok :/
<jbetz_>
Although, that might be the EFI I deleted "|
<jbetz_>
*:|
<jbetz_>
It didn't error when I true though
<jbetz_>
The label was still valid
<sphalerite>
oh right, since the mounts have changed you may also need to re-run nixos-generate-config --root /mnt
<sphalerite>
it won't touch your configuration.nix, only overwrite hardware-configuration.nix
acarrico has joined #nixos
<jbetz_>
Okay, I tried relabeling the old EFI to boot, then regenerating the hardware config, and now it has two entries for boot
<jbetz_>
Which causes nixos-install to fail
<sphalerite>
did you mount it twice?
<jbetz_>
Why would it think there are two? I only have one ESP
<jbetz_>
only mounted once now, but fails in a different way :/
coot has joined #nixos
coot has quit [Remote host closed the connection]
<jbetz_>
I think boot is symlinked to the wrong partition
<{^_^}>
[nixpkgs] @grahamc opened pull request #39649 → NixOS docs: making it easier to hack on → https://git.io/vp44U
<jbetz_>
`File system "/boot" is not a FAT EFI System Partition (ESP) file system.`
<jbetz_>
does the windows ESP need to be reformatted?
<jbetz_>
I already ran `mkfs.fat -F 32 -n boot /dev/sda1`
<sphalerite>
there shouldn't be a symlink at all
<jbetz_>
where /dev/sda1 is the EFI System
<jbetz_>
sphalerite: the symlink was in /dev/disk/by-label
freeman42x]NixOS has joined #nixos
<jbetz_>
`dev/disk/by-label/boot --> ../../sda1`
bbsl has joined #nixos
<bbsl>
could anyone help me write this override please? I found some snippets before to set up a haskellenv which has been working for me so far but now I want to add rust/rustNightly to it and I cant understand whats not working. Error: undefined variable 'pkgs' at /home/bbsl/.config/nixpkgs/config.nix:20:26 config.nix: https://pastebin.com/cxCRwsdG
<bbsl>
<srhb>
bbsl: That's a little weird. Normally, what you're calling super _is_ packageOverrides. It looks like you're mixing up some patterns from packageOverrides and overlays.
<sphalerite>
jbetz_: you shouldn't have reformatted the ESP, that probably broke windows :|
<bbsl>
it works if I remove the rustPackages = "blerg" assign but Im not sure if it only works because fetchFromGithub is lazy and work do anything if its not needed
<srhb>
Er, super _is_ pkgs
<srhb>
bbsl: But regardless, the error says that you do not have pkgs in scope
<srhb>
bbsl: And if you look closely, that is indeed the case.
<srhb>
bbsl: super.fetchFromGitHub would exist, for instance. As would super.pkgs.fetchFromGitHub
<bbsl>
srhb: ow I mucked about with it too much and didnt read the error msg when I copied ... it used to be: undefined variable 'rustNightlyNixRepo' at /home/bbsl/.config/nixpkgs/config.nix:27:41, thats what I ment to paste.
<srhb>
bbsl: You just need a rec in your attrset for that to work
<srhb>
bbsl: Normally attributes cannot refer to each other, ie. this is an error: { foo = bar; bar = 42; }
<srhb>
bbsl: But this is fine: rec { foo = bar; bar = 42; }
<{^_^}>
→ 192352ff by @bgamari: nixos/cupsd: Introduce services.printing.logLevel option
<{^_^}>
→ f9fe798f by @Mic92: Merge pull request #39612 from bgamari/cupsd-loglevel
<bbsl>
yea that is my problem, why is that a thing?
<srhb>
bbsl: Being able to define nonrecursive attrsets is nice when you want to _not_ accidentally introduce recursion.
<jbetz_>
sphalerite: damnit :|
<srhb>
bbsl: This would look cleaner with overlays instead of packageOverrides.
<jbetz_>
sphalerite: what should I do now?
hakujin has joined #nixos
Arcaelyx_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<bbsl>
srhb: ty it works now
<bbsl>
srhb: maybe I shoudl learn about overlays while Im at it, know of a good resource? can I define them in this config.nix that I refer to from configuration.nix like Im doing with this file?
hakujin has quit [Ping timeout: 256 seconds]
<srhb>
bbsl: It's described quite well in the nixpkgs manual :)
<srhb>
bbsl: And I prefer overlays over packageOverrides for basically everything.
<srhb>
bbsl: What it does is allow you to refer to the modified set itself, like so: `modified: old: { ... }` meaning in this case modified.rustNightlyNixRepo would exist.
<bbsl>
srhb: ok ty for the link to the doc. I dont doubt that is is a usefull feature. Overlays that is, but the docs kindof suck at explaining how to get started.
daveo has joined #nixos
pxc has joined #nixos
<jbetz_>
sphalerite: yup, that fixed it
testuser has joined #nixos
<testuser>
hi, I've build my system with nixos-rebuild build and it has rebuilt the /nix/store/5clbbvqvscmdx6iyd8lkxwbg7zl4vdrl-dhcpcd.exit-hook.drv derivation, but how can I find the actual script that was generated to see what is contained in the file ?
pxc has quit [Ping timeout: 264 seconds]
<jbetz_>
and now I'm pretty much back to where I started, where only nixos shows up on the systemd bootloader
<sphalerite>
jbetz_: well yeah by formatting the ESP you deleted the windows bootloader.
franksinistra has joined #nixos
franksinistra has quit [Remote host closed the connection]
<jbetz_>
sphalerite: yeah :| any way to install it on my linux ESP?
<jbetz_>
e.g., dump it into /boot/efi ?
<sphalerite>
yeah. Not reformatting the ESP.
<sphalerite>
Just reusing the same one without destroying anything already on it.
aarvar has quit [Ping timeout: 248 seconds]
franksinistra has joined #nixos
franksinistra has quit [Client Quit]
romildo has joined #nixos
<kuri0>
sphalerite, yeah that would work but it to automaticly add all the patches in the folder
<kuri0>
so i don't have to manually add each one
<[Leary]>
Hopefully that's what windows will do if you boot into a recovery environment and ask it to 'repair' your windows install. I wouldn't trust it not to clobber anything though; might be worth backing up the current contens jbetz_.
<[Leary]>
Also, it will probably set itself first in the boot order, meaning you wont see the systemd-boot loader.
<bbsl>
testuser: if the script is linked in your path you can use readlink to find it. there are probably other better ways but I tend to go to /nix/store and use silver searcher if the first one is'nt the case. PS: "nix --help" it has a bunch of new stuff and you might be able to use some of it to navigate the store, playwith with it yesterday to find the .lua file that weechat-matrix-"whatever" puts in your store
<kuri0>
jbetz_, if you want to install the windows boot loader back you usually install it by clicking repair when you boot a windows install media
<{^_^}>
[nixpkgs] @obadz merged pull request #38962 → buildFHSUserEnv: change to root directory after chroot → https://git.io/vpebM
<{^_^}>
[nixpkgs] @obadz pushed commit from @lheckemann to master « buildFHSUserEnv: change to root directory after chroot »: https://git.io/vp4EC
<sphalerite>
thanks obadz !
<obadz>
np
Mateon3 is now known as Mateon1
franksinistra has left #nixos [#nixos]
<epta>
I've just switched from ghc822 to ghc842 on darwin, and it seems that all the packages need to be built locally. There is no cache for ghc842, or am I missing something?
<sphalerite>
epta: there's no cache for non-default stuff AFAIK
ivanivan has joined #nixos
<ivanivan>
Is there a way to tell a globally-installed Ruby where to find a gem packaged in the nix store?
<tilpner>
You could globally set GEM_HOME and GEM_PATH. Look at nixpkgs/pkgs/applications/networking/cluster/panamax/ui/default.nix for an example, but note that I have not tried this, nor am I familiar with Nix Ruby packaging
<{^_^}>
[nixpkgs] @ayyjayess opened pull request #39653 → Polybar: make call to i3 pure → https://git.io/vp4uJ
jacob_ has joined #nixos
logzet has quit [Remote host closed the connection]
<Slabity>
Hey guys, I'm getting this error when trying to deploy an EC2 instance from NixOps: `No handler was ready to authenticate. 1 handlers were checked. ['HmacAuthV4Handler'] Check your credentials`
<Slabity>
Any idea what this could mean? I'm pretty certain my access keys are correct.
<ivanivan>
tilpner: thanks, I'll take a look at that
johnw has quit [Ping timeout: 264 seconds]
daveo has joined #nixos
<toppler>
Ut-oh. Just upgrading and rebooted and now my Dell is randomly freezing on me.
iyzsong has joined #nixos
jperras has joined #nixos
<Slabity>
Nevermind with my EC2 issue. I just had a typo I couldn't find.
kreetx has joined #nixos
<sphalerite>
toppler: boot into an older generation and check the logs?
<epta>
Is there a way to pass '--allow-newer' cabal flag in a nix expression?
<toppler>
sphalerite: Will do.
<sphalerite>
epta: I don't know much about the haskell packaging but there's this doJailbreak thing which might be what you want
<jbetz_>
anyone know why opera would crash whenever it tries to open a file dialogue?
<jbetz_>
the application closes then reopens immediately
<jbetz_>
no such issue in firefox
boxscape has joined #nixos
<boxscape>
is there some easy way to see for which packages in a channel pre-built binaries exist?
Notkea has joined #nixos
coot has joined #nixos
<toppler>
sphalerite: One problem is that I expect it might be a kernel issue. I'm right in thinking that, if that's the case, I'll have to specifically pick a generation with a different kernel, since I'll have gone through a few generations without ever rebooting into a new kernel?
<sphalerite>
toppler: that's possible
FPtje has joined #nixos
Guest16581 has joined #nixos
iyzsong has quit [Ping timeout: 240 seconds]
coot has quit [Quit: coot]
Slabity has quit [Remote host closed the connection]
acarrico has quit [Ping timeout: 240 seconds]
dx_ has joined #nixos
<FPtje>
Does anyone have experience with 2-in-1s on NixOS? Something is disabling Wacom and touch screen input when I fold the lid backwards. I already figured that folding the lid backwards is considered the same as closing the lid, so logind.lidSwitch and logind.lidSwitchDocked are set to "ignore"
iqubic has quit [Remote host closed the connection]
jperras has quit [Ping timeout: 268 seconds]
klntsky has quit [Remote host closed the connection]
<FPtje>
Regarding lid fold backwards touch disabled: xinput doesn't seem to know that touchpad and stylus input are disabled as it still lists them as enabled
Ariakenom has quit [Read error: Connection reset by peer]
ThatDocsLady has joined #nixos
tyoma has joined #nixos
jperras has joined #nixos
bionicmac[m] has joined #nixos
jperras has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @AmineChikhaoui opened pull request #39654 → GCE: pull the ssh host keys from the metadata service as expected by NixOps. → https://git.io/vp4aw
LysergicDreams has quit [Ping timeout: 260 seconds]
<bbsl>
could someone please explain to me, if I clone the nixpkgs repo and try to build any of the packages with nix build from the respective folder containing default.nix why am I getting: "cannot auto-call a function that has an argument without a default value ('stdenv')"
<clever>
bbsl: you have to load that default.nix with callPackage
<clever>
bbsl: if your working in a fork of nixpkgs, edit all-packages.nix to add your package, and use the default.nix in the root of the whole nixpkgs tree
FPtje has quit [Remote host closed the connection]
xy2_ has joined #nixos
pie___ has joined #nixos
pie__ has quit [Ping timeout: 248 seconds]
nckx has quit [Quit: Updating my GNU GuixSD server — gnu.org/s/guix]
nckx has joined #nixos
<aristid>
great
<aristid>
my system refuses to evaluate because mono is marked insecure
civodul has joined #nixos
<aristid>
of course it's also the default version that is marked insecure, because the default points to an ancient version
<aristid>
and i have no idea which of my packages uses mono
<vcunat>
I think I had a .net-capable Delphi (object pascal) ages ago.
<aristid>
and even though i knew nothing about c#, i still couldn't help myself refactoring some code to remove ~5000 lines of c# (it was easier to do that than to add the 3 new columns in the 25 near-identical copies of a function :D)
<{^_^}>
[nixpkgs] @aristidb pushed to master « mono: change default version to "latest" (5.8), because 4.0 does not evaluate due to security »: https://git.io/vp4oD
<benley>
(I needed that to use a binary-only .dll that only worked on win32 and required .net stuff)
<aristid>
vcunat: don't hate the delphi, hate the marginal programmers who made "programs" with it </code snobbery>
<bbsl>
im trying to add a new (fairly simple) package to nixpkgs but i am now stuck on step 5 of the guide testing the build with "$ nix-build -A libfoo" the jar I have access to is going rogue and trying to put the output of the build where its not supposed to https://pastebin.com/iKSM9itT
<aristid>
benley: the culture of using binary dlls even as a developer is one of the weirdest microsoftisms to me
<aristid>
vcunat: 4.8 is also ancient isn't it
<aristid>
vcunat: i'd rather make the default be the latest and change things that need an older version to specify that explicitly
<bbsl>
im guessing I have to fix it by overrding this somehow, anyone have any pointers as to where to look to find the solution?
<vcunat>
bbsl: depends on the project
<vcunat>
often you can just makeFlags = [ "PREFIX=$out" ];
<vcunat>
(or something in configureFlags)
<aristid>
vcunat: ah, mono 4.8 is february 2017, so not crazy old
<aristid>
dhess: staging usually is only merged once in a while due to the mass rebuilds
<aristid>
dhess: can't you use staging for aarch64 for the time being?
<dhess>
aristid: this only affects aarch64 which is broken anyway
<dhess>
aristid: not easily, this is part of a NixOps deployment with other archs
<aristid>
dhess: not sure why Mic92_ merged it to staging rather than master if it only affects aarch64
<dhess>
right, which is why I'm asking :)
boomshroom has joined #nixos
<dhess>
Is there a way to get a .patch file from a GitHub PR URL?
<boomshroom>
Hello!\
<dhess>
oh add .patch :)
<boomshroom>
Has anyone tried building NixOS for Nintendo Switch?
<dhess>
awesome
<benley>
boomshroom: lol I have no idea but I love this idea already
<boomshroom>
benley: Thanks to the power of Nix, it should be possible to make a derivation that builds an existing configuration.nix and outputs a bash or python script that loads everything over usb.
hakujin has joined #nixos
<sphalerite>
boomshroom: the switch runs linux?
jperras has joined #nixos
<aristid>
sphalerite: not by default :)
<benley>
I think people have figured out how to run homebrew code on switch; I don't know if that extends as far as a linux kernel, but I wouldn't be surprised if it does.
<boomshroom>
sphalerite: The switch _can_ run linux. It requires exploiting a buffer overflow in the Tegra usb recovery mode.
<sphalerite>
nice
<benley>
oooooh that makes sense
<benley>
I forgot they are basically a Tegra platform
<benley>
I bet you could build a version of Android for switch if you were sufficiently determined
* sphalerite
is suddenly interested in having one
<sphalerite>
not to the point that I'll actually get one, most likely. But still
<boomshroom>
benley: Someone managed to run an Android emulator in a GNU enviornment on the Switch.
<benley>
nice
<benley>
I haven't been following the switch hacking scene very closely... maybe I ought to :-)
<benley>
all I've been doing is hacking up Breath of the Wild mods
<boomshroom>
There's also Windows 95 in DOSBox running.
Guest16581 has joined #nixos
<dhess>
how about MAME?
<boomshroom>
I really hope Valve would release an ARM version of Steam with a few games. It would be incredible to play even a fraction of my existing PC library on the Switch.
<dhess>
that would be worth it
jperras has quit [Ping timeout: 268 seconds]
<boomshroom>
dhess: If it runs on Linux-AArch64, it will probably work. Gamecube and playstation emulators have already been tested.
boxscape has joined #nixos
<benley>
I would love to see Katamari Damacy running on Switch via ps2 emulation, that would be hilarious and wonderful
<boomshroom>
benley: If PCSX2 ran at a reasonable speed, I would want to try Xenosaga just to say I can have 4 of the 5 Xeno games on the same system: Xenogears (PS1), Xenosaga (PS2), Xenoblade (Wii), and Xenoblade 2 (Switch). The only one missing is a Wii U emulator to play Xenoblade X.
<benley>
Wii U and Switch are practically the same platform under the hood, aren't they? With any luck it would be relatively easy to port something like cemu
<kreetx>
a nixops noob question: if I want to add a folder from my local machine to a deployed nixos (on virtualbox), do I use a buildhook for that?
<kreetx>
would also be good if I could somehow persist this
<clever>
kreetx: where would you want the folder to turn up?
<boomshroom>
benley: BS. The Wii U and Switch don't even use the same ISA. Wii U is a PowerPC while the Switch is AArch64.
<clever>
kreetx: and should tbe folder be read-only, writable, and if writable, what happens to the changes at next boot?
<benley>
Oh, *duh* sorry. I should have remembered that wii U is powerpc.
rosa has quit [Quit: rosa]
<kreetx>
I'm making basically a web image gallery and want to copy the images
mmlb has quit [Read error: Connection reset by peer]
<clever>
kreetx: id just manually scp the images up once the system is deployed
dx_ has quit [Quit: leaving]
<kreetx>
clever: alright, will try that
<boomshroom>
benley: Even though GC and Wii were also PowerPC, the Wii U was different enough to warant its own emulator, where as the Wii was close enough to GC that they were able to bolt on the functionality to the existing GC emulator.
johnw has quit [Ping timeout: 248 seconds]
<kreetx>
clever: I guess persistence of files, databases, etc is a matter outside of core nixops/nixos?
mmlb has joined #nixos
<benley>
boomshroom: makes sense. I totally forgot about the architecture shift, even though we were just talking about Tegra like 5 minutes ago.
<kreetx>
i.e it's up to me to back those up, etc
anton___ has quit [Ping timeout: 276 seconds]
johnw has joined #nixos
<simpson>
boomshroom: The Wii's CPU and GPU have (almost) the same ABI as the GC, just better clock, cache, RAM, etc. The GPU's a little nicer, but I'd still lump it with GC's in terms of programming model.
<benley>
Wii can straight-up run GC games, so that makes sense too :-)
<simpson>
Yeah, ABI compatibility was such a clever trick.
<clever>
kreetx: yeah
<clever>
kreetx: for more static things like the public_html of a website, you can build and manage it with nix, but a gallery is a bit too dynamic to bother with in nix
<boomshroom>
benley: The GBA and Wii U are also backwards compatible, because they contain a smaller version of the previous system inside themselves.
<simpson>
Although programming for those GPUs must be a pain. The GPUs don't have shaders, just 'texture combiners', proto-shaders which are super-limited.
<sphalerite>
MichaelRaskin: the wrapper is only for plugins, not for scripts, so far
<sphalerite>
weechat isn't very nice about separating config, state and data
<kreetx>
another noob question: where could I get info on predefined attributes I could set in my nixos machine? Like how to set if a user should have a home folder and such
<sphalerite>
kreetx: man configuration.nix
<kreetx>
this could be in the manual, but there are probably many which aren't
<kreetx>
yet another noob question: I added createHome = true to a user's config, but the home appears not to be created after a redeploy, do I need to reboot the machine to make it happen?
<kreetx>
and perhaps also, does nixops remove things I remove from the machine's configuration?
symphorien has joined #nixos
<zybell>
ofc
<clever>
kreetx: createHome only applies when its first creating a user
<clever>
kreetx: once the user has been created, it stops checking for that flag
<clever>
kreetx: and some things will be removed automatically if removed from the config, and that includes users
<sphalerite>
There's quite a dearth of matrix clients. Riot is the only one that actually supports all the features of matrix :|
<MichaelRaskin>
Or LuaJIT…
mmlb has joined #nixos
<boomshroom>
clever: Does it come with a touchscreen-compatible desktop environment? And from my understanding, the current loader expects the rootfs to be on the system's SD card.
<Mic92_>
dhess: aristid I was expecting staging to be merged into master earlier
<dhess>
Mic92_: can you push it to master?
<clever>
boomshroom: it has zero gui, and doesnt even run a login prompt on the main tty, but the rootfs stuff can be modified easily
<vcunat>
nixos-unstable updated!
<clever>
boomshroom: but its also a lot smaller then full nixos, and could be usefull to get a kernel+rootfs working before you switch to something bigger
<zybell>
boomshroom:think of it as grub in disguise;-)
<MichaelRaskin>
Are there any weechat plugins that don't work with LuaJIT?
matklad has joined #nixos
<zybell>
therefore rootfs=/boot lol
<boomshroom>
zybell: Sadly, we haven't found a way to overright the bootloader and we don't know the keys to sign the firmware so we can't overright that either.
<boomshroom>
*overwrite. What is wrong with me?
<sphalerite>
you could just say you meant "replace it with the right bootloader" :D
<clever>
boomshroom: is it a 32bit or 64bit arm cpu?
<boomshroom>
There are ways of getting a serial port through the Joy-Con rails, but it would be easier just to boot to desktop.
<clever>
boomshroom: #nixos-aarch64 may also be of interest to you then
<aristid>
MichaelRaskin: lua 5.1 is really old oO
<dhess>
Mic92_: whoot! Thanks!
<aristid>
MichaelRaskin: i guess they had luajit in mind
johnw has quit [Ping timeout: 264 seconds]
<boomshroom>
clever: Thank you.
<MichaelRaskin>
Hm, how interesting
Arcaelyx has joined #nixos
<MichaelRaskin>
The nixpkgs package for weechat-matrix-bridge expects lua5.2
<MichaelRaskin>
But, it has some luajit/lua5.1-only code…
johnw has joined #nixos
<MichaelRaskin>
Ah wait, it might be trivial to fix
<boomshroom>
It looks like the Linux build requires a /lib/firmware/brcm/brcmfmac4356-pcie.txt. I wonder how hard it would be to patch it to take an environment variable or store path instead.
<dhess>
Hmm I could really use a monad in Nix right now
<ixxie>
heya folks; so my recent generations show an error preceeding the boot log stating something like "[Firmware Bug]: TSC_DEADLINE disabled due to Errata"
<pstn>
What is the best way to get a moderately complex ip tunnel command like
<pstn>
`ip tunnel add he-ipv6 mode sit remote xx.xx.xx.xx local xx.xx.xx.xx ttl 255` into the declarative configuration?
<{^_^}>
[nixpkgs] @ilya-kolpakov opened pull request #39660 → pythonPackages.pymc3 init at 3.2 → https://git.io/vp4P3
zybell has joined #nixos
<MichaelRaskin>
So, I connected to matrix.org using nixpkgs version of weechat-matrix-bridge, and it seems to somewhat succeed, but doesn't create matrix quasi-server. Sigh
ixxie has quit [Quit: Lost terminal]
mmlb has quit [Ping timeout: 240 seconds]
<silver_hook>
Did anyone get built-in codecs to work in Falkon?
<silver_hook>
i.e. compile qtwebengine with codecs
<silver_hook>
i.e. compile qtwebengine with codecs
toby1851 has quit [Ping timeout: 260 seconds]
spear2 has joined #nixos
ixxie has joined #nixos
<MichaelRaskin>
Anyone uses weechat from nixpkgs with matrix? I get strange problems…
johnw has quit [Ping timeout: 240 seconds]
<Myrl-saki>
clever: I think I got to an even worse problem.
<Myrl-saki>
It's always the simple things that we miss.
<Myrl-saki>
clever: But hey, at least it's fixed. :P
<elvishjerricco>
ghc --make is kinda bad at parallelism though. I've found that ghc -M (which produces a makefile with the correct dependency graph), followed by regular `make` is usually way more efficient.
<clever>
i would have expected ld to say that it was out of disk
<clever>
elvishjerricco: ive been told that the interface file parsing makes that worse
<elvishjerricco>
clever: In my somewhat limited experience, that's outweighed by how bad ghc -j is at utilizing multiple threads.
mmlb has joined #nixos
PolarIntersect has joined #nixos
iwxzr has quit [Remote host closed the connection]
<Myrl-saki>
clever: At least we get -l?
<boomshroom>
I'm always scared of doing a GC due to how often I end up building custom GCCs.
<elvishjerricco>
Especially if you give GHC more threads than it needs. Like if you give -j8, and it only needs like -j2 for the module graph, it'll still waste a ton of cycles trying to use 8 cores.
<elvishjerricco>
ends up going like +50% slower
mmlb has quit [Client Quit]
<Myrl-saki>
elvishjerricco: I passed it -j32...
<elvishjerricco>
Yea... In my experience that's a costly mistake
<elvishjerricco>
On my 32 thread machine, I default to `nix-build -j 16 --cores 4` for Haskell. Technically exceeding my core count in the worst case, but not usually relevant. Limiting --cores helps a ton with Haskell
<das_j>
Hey everyone, I'm currently trying to perform remote builds (build a remote system on itself and switch to the system on the remote machine). However, all I can find on the internet looks like it's made for nix 1 (calls a perl script and afaik all perl scripts were removed in nix 2). So what is the current way of doing that?
<ixxie>
anybody know what to make of it? I run a Dell XPS 9350 and this only appeared recently; I just ran a BIOS update and it didn't help.
<sphalerite>
das_j: nixos-rebuild's --build-host and --target-host options should still work.
<kreetx>
ixxie: this is not ubuntu channel
<sphalerite>
ixxie: did you try acpi_rev_override=5 ? I needed that on my 9650 to avoid a lockup caused by nvidia trying to talk to the (powered-off) graphics card
<sphalerite>
kreetx: ixxie is a nixos user and presumably encountering this issue with nixos ;)
<das_j>
sphalerite: That looks like what I need. Does this parse the expressions locally and evaluates them on the remote machine? So I don't have to keep the .nix files there as well?
<sphalerite>
das_j: I believe so.
<kreetx>
yeah, just read him/her again :) "error similar to.."
<das_j>
sphalerite: Nice, thank you! I'll try it out
<ixxie>
kreetx: I am aware of that but I am experiencing the error in NixOS xD
<ixxie>
sphalerite: where is that set exactly?
<sphalerite>
ixxie: kernel parameter. So boot.kernelParams in the nixos config, or in the bootloader by pressing e if you're using grub or systemd-boot
<Myrl-saki>
elvishjerricco: Thanks for the link. I turned it down to --cores 4, and it seems less dead to me.
<Myrl-saki>
The problem is that LLVM will now build with a single thread.
<elvishjerricco>
Myrl-saki: Yea... I don't have a good solution for that :/
vandenoever has joined #nixos
<elvishjerricco>
Ideally, GHC just wouldn't be bad at this... But maybe we could tweak the default runtime options in nixpkgs as per that blog post?
<elvishjerricco>
Then we'd just be missing some sort of `-l` for GHC
<MichaelRaskin>
ekleog: you pinged weechat-matrix authors to submit it to weechat main script repository, but I think the real problem is that it doesn't work with 2.0
aarvar has joined #nixos
ixxie has joined #nixos
<ixxie>
sphalerite: that didn't work.... the SO post mentioned GRUB_CMDLINE_LINUX_DEFAULT="quiet splash acpi_rev_override=5" - do you think I should try that?
<sphalerite>
ixxie: no, that's just the way you pass that parameter to the kernel on ubuntu
<MichaelRaskin>
WOW. If weechat is launched with https_proxy set, its behaviour w.r.t. https requests is… interesting.
rosa has joined #nixos
fendor has joined #nixos
<sphalerite>
Weechat is generally… interesting
<sphalerite>
:p
ylwghst has quit [Ping timeout: 264 seconds]
astrofog has quit [Quit: Quite]
<MichaelRaskin>
Erm. I am actually considering using a GUI matrix client in xdummy, if there is one with FIFO control and sane logs…
<aristid>
MichaelRaskin: what does it do with https?
<aristid>
also i didn't realize weechat ever makes any https requests
<sphalerite>
MichaelRaskin: spoiler alert, there isn't
<sphalerite>
aristid: matrix consists almost entirely of http
ixxie has joined #nixos
<MichaelRaskin>
ARGH
<aristid>
sphalerite: kids these days always using http for everything!
blankhart has joined #nixos
<MichaelRaskin>
Well, really HTTPS
<aristid>
MichaelRaskin: that's like http+http!
<aristid>
multiple https
<MichaelRaskin>
And that means that my local homeserver's Web-UI can be used as WebUI to _any_ Matrix server. Well, any small matrix server
* aristid
shows himself out for the bad joke
<ixxie>
sphalerite: that removed the error but also stopped the boot from completing ^^
<MichaelRaskin>
Because you know, that Web-UI just assumes that asking for a list of _all_ public chats is a sane idea
<darkmeson[m]1>
https and json is all anyone will ever need
<MichaelRaskin>
If there is a proxy set, weechat feeds the caller not only the response body, but also headers
<MichaelRaskin>
Imagine how happy are JSON parsers to get HTTP/1.0
<MichaelRaskin>
Spoiler: they are not at all happy
<fendor>
hey guys, i wanna report a bug. I dont know though, where I should report it. It is not a nixos bug, but it occurs mainly on nixos, i think it is actually a maven/protobuf bug, but I want to make it easy to find for other nixos users and how to fix the problem etc.
<sphalerite>
ixxie: well damn. There's probably lots of other ACPI fiddling to try but I don't really know anything about it
<fendor>
I guess, I should report it to the maven plugin guys?
<fendor>
guys/girls
<sphalerite>
ixxie: oh yeah what might also be worth mentioning is that the ACPI error could be a red herring. I get an error looking like that but everything works fine
<ixxie>
fendor: protip: the word folks is gender neutral ;)
<fendor>
ixxie, yeah, people, too, in english at least :D
<darkmeson[m]1>
people works too
<ixxie>
fendor: yeah but I find folks more endearing somehow, maybe because of the saying that "persons are awesome but people suck"
<sphalerite>
ooooooh there's a matrix client that ships with a nix expression!
<MichaelRaskin>
Nix users are expected to be more or less OK with console, so there is a hope…
mmlb has joined #nixos
doyougnu has joined #nixos
<sphalerite>
"that feels more like a mainstream chat app (Riot, Telegram etc) and less like an IRC client" I suspect you're not the target audience though :)
<fendor>
however, should i just report it to the maven plugin that is in my opinion responsible, and is there some additional place for nixos users where i should report the bug to?
ixxie has quit [Quit: Lost terminal]
<MichaelRaskin>
Yeah, doesn't seem to have remote control
<darkmeson[m]1>
there's also a matrix script for weechat, although lack of cjson has been a problem on most distros I've tried it on
<MichaelRaskin>
sphalerite: a GUI app _might_ have remote control (see: gajim)
ixxie has joined #nixos
jhb563 has joined #nixos
arianvp2 has joined #nixos
<arianvp2>
heya
<arianvp2>
is it possible to let nixos containers connect to eachother?
<jhb563>
sphalerite: I get a weird permission error when I try to run the install script. On the "copying to /nix/store" step, it tries to move a directory to the same location (/nix/store/./{hash}-nss-cacert-3.35.33) and says Permission denied. Still happens when run with sudo. Is this an issue you've encountered?
<arianvp2>
oh I should make a bridge
<arianvp2>
got it
<sphalerite>
jhb563: I don't actually know anything about using it, I don't use windows. Just know it's been done.
<sphalerite>
zimbatm might know something about it, he wrote that article
maingo has joined #nixos
johnw has joined #nixos
<pstn>
I even share stuff between containers with sockets on bind mounts. For example my mysql doesn't have network at all that way.
<pstn>
I'd advise against that though if you don't have an idea how sockets work and what the implementators might have presumed about them.
<pstn>
socket auth for example is a bad idea then...
<arianvp2>
hmm I created a bridge without any interfaces and used the 'hostBridge' option
<arianvp2>
but if you create a bridge without any interfaces nixos won't create the bridge...
<{^_^}>
→ f5268994 by @7c6f434c: weechat-matrix-bridge: 2018-01-10 -> 2017-03-29
<{^_^}>
→ 7cd7e4ce by @7c6f434c: weechat: 2.0 -> 2.1
<arianvp2>
I do need internet I guess.
fenedor has joined #nixos
mkoenig has quit [Remote host closed the connection]
<clever>
arianvp2: then youll want to put the host eth0 into the bridge, and move any ip config from eth0 to br0, if you have any
<arianvp2>
ah
<clever>
and at that point, each container will be available to the local LAN, and it will act like normal machines on the lan
<arianvp2>
got it
<clever>
other then all the machines sharing a mac, an outside observer has no way to tell if you have 10 physical computers, or 10 containers
Ariakenom has joined #nixos
mkoenig has joined #nixos
fendor has quit [Ping timeout: 255 seconds]
<darkmeson[m]1>
that's not strictly true
MR[m]1 has joined #nixos
<darkmeson[m]1>
packet timestamps and a few other methods can be used to expose the probable number of devices if the OSes or a firewall aren't doing any randomization
<arianvp2>
clever: I have this now, but the containers aren't getting any IPs
<clever>
arianvp2: the docs imply you need to set localAddress
<arianvp2>
ah and I'll also have to give a netmask
<arianvp2>
that's probably what was missing
<arianvp2>
that's probably what was missing
zybell has joined #nixos
<arianvp2>
ah and I'll also have to give a netmask
<arianvp2>
oh sorry I was having cursor in wrong screen in xmonad and trying to back in my bash history but went through irc history instead
spear2 has quit [Ping timeout: 256 seconds]
<arianvp2>
clever: thanks! containers can speak to eachother now. Just not to the outside network yet
<arianvp2>
Also the host can't talk to the containers either. which is weird as they're bridged
<clever>
that is weird
<bbsl>
I need some help: Im trying to make a package that can go into nixpkgs. I finaly got it to install the way I thought would be correct (with a json-config file in the same directory as the binary) because its the way I installed it when I just did the old "make install" but somehow when its built form my nix store it doesnt work the same way. Works(can find config file): https://pastebin.com/80VBNQCP Does'nt work(cant find config file):
<arianvp2>
bbsl: what happens if you do cd result/bin; ./webdis
<arianvp2>
perhaps it's a relative path issue?
fenedor has quit [Quit: Leaving]
isidore has joined #nixos
<ixxie>
hmm so my XPS trackpad's right click doing a left click, but I can get a right click by click the left and right buttons at once :/
rosa has quit [Quit: rosa]
rosa has joined #nixos
<bbsl>
arianvp2: ah you are correct... it works if I do that albeit with this msg: "Could not open webdis.log: Read-only file system" lol guess I have a ways to go before this works but I guess the solution is it set relative paths during post instalation and in the case of the logpath maybe set it somewhere in /etc. got any pointers as to where I can find info about this though?
<kreetx>
I'm getting 'out of memory' when deploying haskell application while copying ghc. Last time I increased the virtualbox's ram from 1 to 2 gigs, but now after destroying and redeploying the machine I get this even with 2 gigs -- not sure what could be wrong?
ixxie has quit [Quit: Lost terminal]
ixxie has joined #nixos
arianvp2 has quit [Ping timeout: 260 seconds]
mmlb has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @ilya-kolpakov opened pull request #39662 → pythonPackages.pymc3: init at 3.2 → https://git.io/vp4DR
toby1851 has quit [Quit: Konversation terminated!]
<clever>
leo60228: so you just set qemu-user.aarch64 = true;, nixos-rebuild switch, and it can magically run aarch64 binaries, and nix can even run an "aarch64 only" compiler
<kreetx>
alright, got it copied with 3 gigs :)
<kreetx>
I wonder if linking the binary statically would make this go away?
<leo60228>
how do i use a nixos module from outside nixpkgs?
<kreetx>
*and* make the entire thing a lot smaller
Neo-- has joined #nixos
<clever>
leo60228: git clone my repo, then add the path to the qemu.nix to your imports list
<leo60228>
cool
<pstn>
Does anybody have an idea why my containers resolv.conf might be resetting to the hosts value despite `networking.useHostResolvConf` being false?
<dhess>
kreetx: yeah I would start with haskell.lib.justStaticExecutables since you don't need to do anything else on your end
<ixxie>
anybody know anything about libinput?
<kreetx>
clever: I think it does a dynamic compile, because there is a library in the same project (for faster development, basically)
jacob_ has quit [Quit: Lost terminal]
ilyaigpetrov has quit [Quit: Connection closed for inactivity]
ixxie has quit [Quit: Lost terminal]
<kreetx>
is it so that when deploying the *entire* closure gets allocated ram? i.e not only ghc, but all of it needs to fit?
<infinisil>
kreetx: I think so..
isidore has quit [Quit: WeeChat 2.1]
<clever>
kreetx: i think each top-level directory in /nix/store is copied as a single unit, and the whole thing has to fit in ram
<infinisil>
I just had trouble pushing a 400MB derivation to one of my machines because it for some reason can't allocate enough RAM, even though it has 16GB
<clever>
infinisil: id expect it to only use 2 or 3 times the storepath size
<infinisil>
I'm just gonna reboot, that fixed it last time
<infinisil>
Just where does all my RAM go all the time
sigmundv__ has quit [Ping timeout: 260 seconds]
arianvp2 has joined #nixos
<pstn>
Can I have an extra virtual interface on my container without an IPv4 address?
<arianvp2>
clever: I got it working. I had to NAT the bridge :)
<arianvp2>
at which point I could have just directly natted probably ...
<arianvp2>
how do I get dns working inthe container?
<arianvp2>
found it
<arianvp2>
networking.useHostResolvConf
<jbetz_>
kuri0: so, I booted windows from a usb and ran startup repair, and it couldn't do anything :|
<ekleog>
MichaelRaskin: if you want to investigate weechat more, symphorien is running matrix+weechat on nixos; otherwise I packaged nheko for testing, but it's only GUI and no remote access (just tl;dr'd the discussion)
<ekleog>
MichaelRaskin: oh, and if the behaviour of weechat+https_proxy is weird, you can try weechat+redsocks+https_proxy, maybe
daveo has joined #nixos
daveo is now known as mconstant
<ekleog>
(this way weechat wouldn't even know it's using a proxy, hopefully)
<MichaelRaskin>
Well, it is a local proxy for reasons that are not too relevant for WeeChat…
<MichaelRaskin>
But just the idea!
<mconstant>
hello
<MichaelRaskin>
ekleog: Thanks, will try to ask symphorien how to use all that, currently I reached the level where I cannot understand how to use /matrix msg
<papika>
hello, I had a question about scripting inside of /etc/configuration.nix - all of the examples I can find don't have any script-like sections to get me started...
<papika>
I'm looking to be able to create and copy files before it hits a part of the pipeline
mkoenig_ has joined #nixos
ertes has joined #nixos
<papika>
basically, every `nixos-rebuild switch` runs into "cp: cannot stat '/run/keys/initrd-ssh-key': No such file or directory" because I haven't put it in /run yet
<papika>
does anyone know how I can tell nixos to run a script before rebuild