<gjabell>
hey all, does anyone know of a good way to port-forward nixops machines with the libvirtd backend? looks like the de-facto way is using iptables but I'd prefer to let nixops / libvirt handle that automatically since it requires knowing the ip addr of the vm
oborot has joined #nixos
oborot_ has joined #nixos
<freeman42y>
how can I install Spacemacs on Nixos?
eraserhd has quit [Ping timeout: 245 seconds]
eraserhd has joined #nixos
AviD has joined #nixos
<lambda-11235>
Anyone know how to make the NFS server use version 3?
<clever>
lambda-11235: vers=3 in the mount options i believe
<clever>
but thats chosen on the client sid
<clever>
e
<clever>
not sure how to force it on the server
<lambda-11235>
Wierd I'm getting required NFS version not supported when connecting from client side
<{^_^}>
nixos-org-configurations#57 (by rbvermaa, 44 weeks ago, closed): Users are reporting some issues with cache.nixos.org - gather information here
<{^_^}>
[nix] @matthewbauer opened pull request #3054 → Allow empty /nix directory in multi-user installer → https://git.io/fjNWV
drakonis1 has quit [Ping timeout: 250 seconds]
<thoughtpolice>
I should probably update that post with a new one including the exact options for users, tbh
<thoughtpolice>
Aprophecy: To be clear, `nixos-rebuild --option substituters 'https://aseipp-nix-cache.global.ssl.fastly.net' switch` should do the trick, if you'd like to test the improvement I've made to the cache. I can't guarantee it will help, but it may, due to various technical improvements. (If it does help, I'd also be interested in if you're using IPv6 as well)
<ryantm>
freeman42y: You can add that to your NixOS configuration.
<freeman42y>
is it correct?
<freeman42y>
add what to my Nix configuration? I do not follow
<ryantm>
freeman42y: You don't typically edit /etc/nix/nix.conf directly on NixOS
<freeman42y>
ryantm, I did not edit it
<freeman42y>
what I changed is: /home/neo/.config/nix/nix.conf
<clever>
only roots nix.conf file is trusted, so your nix.conf cant affect things
sb0 has quit [Quit: Leaving]
<ryantm>
"We do not recommend using ~/.config/nix/nix.conf unless you are a power user. It's simpler to have only one global config file."
domogled has quit [Quit: domogled]
philr has quit [Ping timeout: 246 seconds]
is_null_1 has quit [Quit: leaving]
kleisli has joined #nixos
teto has quit [Quit: WeeChat 2.5]
zfnmxt[m]1 has left #nixos ["User left"]
atlas_ is now known as Guest85161
Guest85161 has quit [Killed (verne.freenode.net (Nickname regained by services))]
atlas_ has joined #nixos
kleisli has quit [Ping timeout: 272 seconds]
acarrico has quit [Ping timeout: 248 seconds]
<jackdk>
I am trying to recover a nixos system using a chroot. nixos-enter appears to work, but nixos-rebuild switch seems to want to fetch something from the cache. I have network access outside the chroot but get "ping: system error" when I try to do networky stuff within the chroot. What might I be doing wrong?
<jackdk>
(I have bind-mounted /dev /proc /sys into their corresponding paths under /mnt)
<clever>
jackdk: switch will always fail, you want `nixos-rebuild boot`
<clever>
jackdk: nixos-enter will bind-mount things for you
<jackdk>
clever: noted
<clever>
jackdk: can you ping an ip?
atlas_ has quit [Killed (barjavel.freenode.net (Nickname regained by services))]
<jackdk>
I can curl neverssl.com from outside the chroot but not within it
<clever>
what about `ping 8.8.8.8` ?
<jackdk>
I can ping 8.8.8.8 within the chroot and neverssl.com outside the chroot but not neverssl.com within the chroot
<jackdk>
I set up the wireless using nmtui
<clever>
then its a dns problem
<clever>
does /etc/resolv.conf exist in the chroot?
atlas_ has joined #nixos
<clever>
how do the contents differ?
<jackdk>
it's missing a nameserver line that's for certain
<clever>
try re-entering nixos-enter, without the extra bind mounts?
<clever>
jackdk: maybe `--option sandbox false` ?
jaeckel has quit [Ping timeout: 245 seconds]
emptyflask has quit [Ping timeout: 245 seconds]
Wharncliffe has quit [Quit: leaving]
wykurz has quit [Quit: Leaving]
<jackdk>
that successfully built the derivations but failed to update the boot menu, something to do with the crypt setup
iqubic has joined #nixos
Rusty1 has quit [Quit: Konversation terminated!]
<clever>
jackdk: is /boot also mounted correctly?
<clever>
jackdk: is boot.loader.grub.device set correctly?
eraserhd has quit [Ping timeout: 272 seconds]
eraserhd has joined #nixos
<jackdk>
clever: d'oh yeah mounting /boot helps
jaeckel has joined #nixos
jaeckel has joined #nixos
jaeckel has quit [Changing host]
acarrico has joined #nixos
endformationage has quit [Ping timeout: 268 seconds]
<jackdk>
okay next problem there's something wrong with login. I can rebuild, boot into the new config, and the new password I definitely carefully set is not recognised.
<clever>
jackdk: you should be able to use passwd under nixos-enter to change it
<jackdk>
also clever++ for your help so far
<{^_^}>
clever's karma got increased to 175
<clever>
jackdk: but if mutableusers is turned off, it will undo that
<jackdk>
but only on the next build?
<clever>
if `users.mutableUsers = false;` then it will take effect at every boot
<clever>
the default is true
<clever>
when true, passwd will stick and persist
<iqubic>
How do I tell XDG-Open to use Firefox to open web links.
<clever>
iqubic: xfce4-mime-settings
<iqubic>
When I open hexchat, and click on a link, nothing happens.
<iqubic>
I'd expect the link to be opened in Firefox.
<iqubic>
clever: I'm using EXWM, a tiling window manager, and not xfce. I don't have the program xfce4-mime-settings command.
<clever>
iqubic: you can still install it with nix-env, and it should still work
<iqubic>
I see.
<jackdk>
I normally use mutableUsers = false. So I rebuilt with it true, then ran passwd, then rebooted and I'm back. thanks clever
<iqubic>
On a completely unrelated note, I run nixos on a laptop. When I press the power button, my computer shuts off immediantly.
<iqubic>
Is there any way I can make that do nothing at all.
<iqubic>
Like just ignore any presses of that power button.
<iqubic>
if I have a string with '' two single quotes '' on each side, and the contents are have actual new lines in the file, what exactly will that do?
<clever>
iqubic: it will keep it as multiple lines
<iqubic>
The type of services.logind.extraConfig is strings concatenated with "\n"
<iqubic>
Will this work to satisfy that constraint?
<clever>
that means that if you set it in multiple places, it will join the strings with \n
<clever>
but you can set it to a single line of text, with "foo"
<iqubic>
Oh. I see.
<iqubic>
So, it's basically the exact same as programs.zsh.interactiveShellInit.
witchof01 has quit [Quit: WeeChat 2.4]
<clever>
yep
<iqubic>
And I can use lib.mkAfter or lib.mkBefore to fiddle around with the order if I need to.
<clever>
yep
<iqubic>
I have no idea what black magic those functions are performing.
<craige>
I'm going around in circles on this one and was wondering if anyone had any examples switching "src = fetchurl" etc for "src = /my/local/src" - my attempts make this work are coming unstuck. Thanks!
<srhb>
craige: That should literally be all you need.
<srhb>
craige: What's the problem?
<craige>
I get errors such as "error: value is a path while a set was expected"
<{^_^}>
[nixpkgs] @mmahut pushed 0 commits to revert-67130-containers/unprivileged: https://git.io/fjN8n
<craige>
ah. man, thanks for rubber ducking this srhb - The error is actually from the other file being called - not the the one I switched to being a path.
<craige>
The other file is expecting hydaSrc to be a set.
<craige>
Must be close to knock off time :-)
<srhb>
Good good. :)
hyper_ch2 has quit [Remote host closed the connection]
<cinimod>
I know cabal builds something which crashes if I don't have it
<clever>
cinimod: what exactly is the error?
<srhb>
craige: I'm not sure what your source looks like, but if I `nix-build '<nixpkgs>' -A hello.src`, untar that and use the resulting directory as the src for hello, it works fine.
<craige>
I think I may be making a bad assumption clever, srhb. does `fetchurl... mirror` pull from upstream source, such as git.savanah.gnu.org or frmo a NixOS mirror?
<craige>
good call srhb
<srhb>
craige: If the hash matches something in our mirror, it's used. Otherwise upstream.
Okinan has quit [Quit: Konversation terminated!]
<cinimod>
clever: you mean when I build outside nix with cabal without quadmath?
<clever>
cinimod: when you build with nix
<craige>
okay, that must explain the difference. thanks srhb - your last suggestion is going to be the best way.
<srhb>
craige: I'm a bit puzzled by the difference though. But yeah, if this works, something must have changed in the source..
<Aleksejs>
hello, I've actually managed to get my scanner to be visible over wifi on nixos. However, it's visible only for root. I tried to add myself to scanner and lp groups but it didn't change a thing
<clever>
Aleksejs: when you run `id`, are you listed under those groups?
<craige>
I'm using this to test image deploys for both work and my efforts to deploy NixOS to the librem5 (as I'm carrying local packages until my PR's get through), srhb, clever - so that's been a massive break through for me.
<craige>
That's one of many (the others aren't fit to be PR'd yet. I'm starting to discover how much a giant thing I've bitten off.
<Aleksejs>
clever: scanner:x:59:aleksejs
<clever>
Aleksejs: you need to logout and login again for the group change to apply
<srhb>
craige: Heh, good job ;-)
<craige>
giant but fun.
<craige>
As the hardware hasn't shipped, I'm using VMs to prepare a build, as Purism are doing.
<srhb>
craige: Very cool. Sounds like Nix is practical in this way. Also this way. :)
dasj19 has joined #nixos
<craige>
*loving* `-A vm`
<srhb>
Yeah :P
* craige
disappears for a while
<Ariakenom>
I've been trying to install skypeforlinux. A skype .deb link was broken so I replaced it and it built. But when running it's missing libuuid.so.1
<clever>
,locate libuuid
<{^_^}>
Found in packages: kythe
<Aleksejs>
clever: thanks, it helped
teto has joined #nixos
<clever>
why is libuuid in lythe??
hyper_ch2 has joined #nixos
<srhb>
,locate libuuid.so
<{^_^}>
Found in packages: libuuid, utillinux, libossp_uuid
<clever>
thats better
<clever>
Aleksejs: is libuuid being added to the rpath?
hyper_ch2 has quit [Remote host closed the connection]
<srhb>
Ariakenom: See also 1c1b809c1e53c131c6395c2ccf6c6c4ec8bcdfe5
<srhb>
(From yesterday)
<cinimod>
Woohoo - so after finally getting the Haskell package to test via cabal new-test, using cabal2nix, adding gfortran.cc.lib, removing quadmath, the package now builds in nix
<cinimod>
Oh and runs
<srhb>
cinimod: :o
<Ariakenom>
srhb: looks promising
hyper_ch2 has joined #nixos
veske has joined #nixos
<Ariakenom>
should I remove the old build? and how
slack1256 has quit [Remote host closed the connection]
<cinimod>
Do I need to make a PR for a special derivation for the Haskell package lapack?
<Ariakenom>
im guessing you meant me for both of those, clever? :)
<clever>
Ariakenom: oops, yeah
<Ariakenom>
"nix-env -f . -iA skypeforlinux" on a nixpkgs clone
<clever>
Ariakenom: nix-env -e skypeforlinux to remove it
<Ariakenom>
I did have local changes to the nixpkgs, does that matter
<clever>
Ariakenom: nope
<clever>
Ariakenom: nix-env -e operates entirely on the names from `nix-env -q`
<Ariakenom>
ok thanks
<teto>
has anyone managed to make coc-python work ?
drakonis has quit [Ping timeout: 248 seconds]
judson__ has joined #nixos
<Ariakenom>
clever, srhb: success! thanks a bunch
<clever>
Ariakenom: i also prefer testing builds with nix-build, not nix-env
dasj19_ has joined #nixos
<Izorkin>
teto: you reproduce the error?
<Ariakenom>
yeah, I was using that to build the newest now. I've learned :p
judson_ has quit [Remote host closed the connection]
dasj19 has quit [Remote host closed the connection]
civodul has joined #nixos
drakonis has joined #nixos
<asymmetric>
i'm writing a nixos module for pihole, and i have a question: is it better to hide/disable the cli flags that change the service's config (e.g. toggle logging), since they should only be changed via the nix module?
zaeph has quit [Ping timeout: 252 seconds]
orivej has quit [Ping timeout: 245 seconds]
zfnmxt has quit [Quit: Bye!]
<tilpner>
asymmetric: It's going to fail anyway if the config is generated into the store
cfricke has joined #nixos
zaeph has joined #nixos
<asymmetric>
tilpner: right now the config is saved in /etc for this reason, but it could also make use of StateDirectory if r/w were needed
<asymmetric>
so that's the question, whether pihole's subcommands should fail or not
<tilpner>
Are there other ways to configure logging without state changes?
<tilpner>
Is the state also used for semi-legitimate usage, like storing keys in that config file?
jb55 has quit [Remote host closed the connection]
jb55 has joined #nixos
<Ariakenom>
... but wait I didn't pull those changes .. eh whatever
<asymmetric>
tilpner: no, actually no secrets are stored there, so it would be a good fit for the store i guess
<teto>
Izorkin: sry haven't looked at your tar yet
<teto>
from nixFlakes > "warning: the syntax 'nixpkgs.<attr>' is deprecated; use 'nixpkgs:<attr>' instead", finally \o/
___laika has joined #nixos
<Izorkin>
teto: ok
<hpfr[m]>
Nixos-rebuild outputs a ton of substitutestream() earnings whenever udev rules are updated, is this a big
<hpfr[m]>
Bug*
<asymmetric>
tilpner: if the config is saved to the store, and the subcommands are not removed, they will be broken. this is not an issue for the module, but is it ok to have a package with broken functionality?
drakonis has quit [Ping timeout: 244 seconds]
jgeerds has joined #nixos
<tilpner>
asymmetric: There are plenty of packages where not all features are fully supported. I'm not telling you to break functionality if it's regularly used, and you said it's still possible to enable logging declaratively
zupo has joined #nixos
<tilpner>
asymmetric: Editing configuration manually for debugging is a valid usecase, but many other services already disallow this
drakonis has joined #nixos
hamishmack has joined #nixos
<asymmetric>
tilpner: yeah, it's possible to do it declaratively, and i think doing it imperatively should not be supported
hexo_ is now known as dolanbatar
<asymmetric>
tilpner++
<{^_^}>
tilpner's karma got increased to 40
<tilpner>
Only if it's very useful
hamishmack has quit [Client Quit]
<tilpner>
Users of NixOS aren't going to be surprised a service sacrificed some niche feature for stateless-ness
nschoe has joined #nixos
<tilpner>
(Unless that feature is more than a tiny convenience)
<nschoe>
Hi everyone, I have been using KiCad with nixOS for quite a long time, and it workde mostly okay. But for a few weeks now, when I try to 3D components/boards, I am greeted with an error message "OpenGL 2.1 is required".I have recently updated my nix-channel and upsated the whole system, but it did not chang eanything.
<Yaniel>
did you reboot
<nschoe>
Yaniel, Yes.
is_null has quit [Quit: leaving]
<tilpner>
nschoe: What GPU are you using?
<nschoe>
In KiCad's default.nix file I see libGLU_combined as a dependecy, so I suppose that's where OpenGL is.
<tilpner>
It's possible an update to kicad changed the requirements
<nschoe>
tilpner, I'm on a thinkpad laptop, with an embedded Intel chip. No dedicated graphcis crd.
ilmu has joined #nixos
<nschoe>
tilpner, also noet that it has worked for a long time (on the same computer): I could see my boards in 3D, etc.
<tilpner>
It's possible an update to kicad changed the requirements
<Yaniel>
nixos did also change something about LD_LIBRARY_PATH
<tilpner>
nschoe: And yes, with nix-store -r /nix/store/d1h0sxk7ngld9jghrg8m6v642rrhxdz9-kicad-5.0.2
<tilpner>
(Hopefully)
<nschoe>
tilpner, pciutils output is: 00:02.0 VGA compatible controller: Intel Corporation HD Graphics 5500 (rev 09)
<tilpner>
That sounds recent enough
<nschoe>
tilpner, I have just launched the command. But how does this work? o_O This path doesn't exist on my machine. So I guess it';s fetching it somewhere?
<tilpner>
Yes, it would be fetched from cache.nixos.org by default
jgeerds has quit [Ping timeout: 258 seconds]
<nschoe>
tilpner, I guess so. I havent had other issues with OpenGL before. It worked with KiCad, it worke with other things too
<tilpner>
Does it still work with e.g. glxspheres64?
<tilpner>
(From virtualgl)
<nschoe>
tilpner, the command finnished, should I manually make a link to /nix/store/d1h0sxk7ngld9jghrg8m6v642rrhxdz9-kicad-5.0.2/bin/kicad in order to be able to launch it? (Just askign if there is a special nix command or if I should do it manaully)
<tilpner>
You can just enter that path into a shell directly
<nschoe>
yes
<nschoe>
tilpner, I launched it: no 3D viewer either :/
<nschoe>
tilpner, I don't have an error message, but the window is blank, empty.
alex`` has quit [Ping timeout: 272 seconds]
<tilpner>
Aww, what about glxspheres64?
<nschoe>
tilpner, how do I launch/test that? nix search does't return anything
<tilpner>
$(nix-build --no-out-link '<nixpkgs>' -A virtualgl)/bin/glxspheres64
<clever>
virtualglLib.out 37,888 x /nix/store/ikijswyn692h7i10j3lwhqzj6r0q5aad-virtualgl-lib-2.5.1/bin/glxspheres64
<clever>
thats a new one
<clever>
virtualgl.out 0 s /nix/store/j7xql3qsrp1gzdd1msz97bvp9ch4j4pc-virtualgl-2.5.1/bin/glxspheres64
alex`` has joined #nixos
<clever>
` nix run nixpkgs.virtualgl --command glxspheres64` works here
<clever>
getting 60 fps and ~67 mpixel/sec, with a 2d game also running
<tilpner>
Sure, but your graphics setup isn't broken
<clever>
exact same performance on the laptop, with integrated graphics
<clever>
i think the vsync is throttling it
<tilpner>
Yes
<clever>
same as glxgears, 300 fps is a sign of a problem
<nschoe>
tilpner 's command works for me: 60fps, 37.3 Mpixels/sec.
<tilpner>
vblank_mode=0 to unlimit on some setup
<nschoe>
(the animation is quite beautiful ^^)
<tilpner>
What did it say in the "OpenGL Renderer" line?
<clever>
tilpner: that does in the cli args? it seems to have no effect here
<nschoe>
Got this:
<nschoe>
Polygons in scene: 62464 (61 spheres * 1024 polys/spheres)
<nschoe>
The line with LIBGL_DEBUG=verbose did not change a thing, still have only *** Caught unhandled unknown exception; terminating
ng0 has joined #nixos
<tilpner>
That makes sense if it never even reaches libgl
<nschoe>
Yeah I guess so.
<tilpner>
So it would be interesting if it runs with LIBGL_ALWAYS_SOFTWARE=1
<tilpner>
It shouldn't, if that explanation is correct
<nschoe>
Also launching KICAD with LIBGL_ALWAYS_SOFTWARE=1 doesn't change anything: blank screen, no 3D view.
<nschoe>
It doesn't :/
<tilpner>
Huh
<tilpner>
Are you on unstable or 19.03?
<nschoe>
DUnnow if it matters, but I run i3wm (the i3-gaps fork) and I'm using Compton
<Miyu-chan>
Can I;2ujust second that kicad was buggy for me?
<tilpner>
nschoe: I'm on i3 (no gaps), no compton
fendor has joined #nixos
<Miyu-chan>
I didn't care enough to get it fixed, but it was buggy for me around a year ago.
<tilpner>
I don't know if compton can interfere with that, but i3 isn't the problem
<nschoe>
Miyu-chan, the 3D view?
<Miyu-chan>
Even the main view, IIRC.
<nschoe>
tilpner, yeah I';ve always run i3, for years. So that's not it. I did switch from i3 to i3-gaps a few weeks ago, though. I doubt this is the issue.
<Miyu-chan>
Yeah. Some lines would stop appearing, etc.
<nschoe>
Ok
<Miyu-chan>
Just putting that down. I'm not interested in it anymore, since I don't need to use CAD software.
<nschoe>
tilpner, dunow if this is relevant, but in your trace, the openat calls are on libGLX_mesa.so.0, mine are on libGLX_indirect.so.0 (and they end up on ENOENT).
<nschoe>
tilpner, I'm reading the post you linked.
<tilpner>
What is libGL?
<tilpner>
OpenGL-based programs must link with the libGL library. libGL implements the GLX interface as well as the main OpenGL API entrypoints. When using indirect rendering, libGL creates GLX protocol messages and sends them to the X server via a socket. When using direct rendering, libGL loads the appropriate 3D DRI driver then dispatches OpenGL library calls directly to that driver.
<tilpner>
Now you can figure why that was necessary
<nschoe>
xD I have _no idea_
<nschoe>
tilpner, in my configuration.nix I have: # Enable 32 bits support for steam
<nschoe>
hardware.opengl.driSupport32Bit = true;
teser has joined #nixos
<tilpner>
What about just driSupport?
<tilpner>
nixos-option hardware.opengl.driSupport
<tilpner>
(run that in a shell)
o1lo01ol1o has quit [Ping timeout: 248 seconds]
dasj19_ has quit [Quit: dasj19_]
<tilpner>
I guess it defaults to true?
orivej has quit [Ping timeout: 268 seconds]
<teser>
hi, I have vim compiled that will load /nix/store/41p29hqidnaq06qab2xixnpvla3rvj7c-vim_configurable-8.1.0578/share/vim/vim81/autoload/ scripts, which I suspect are the culprit of really slow usage of vim when editing rst files; any ideas how can I disable them, especially the RstFold.vim which is auto-loaded when starting vim =
jgt has quit [Read error: Connection reset by peer]
<immae>
clever: yes, I read the nix source to find the curl equivalent, so it "should" have the same effect as my `curl --netrc-file /etc/nix/netrc url `
<clever>
fetchurl.cc is only for builtins.fetchurl
<immae>
ah
<clever>
download.cc, i think is used by binary cache api's
<immae>
you mean nixpkgs.fetchurl != builtins.fetchurl ?
<clever>
correct
<immae>
ok
<clever>
pkgs.fetchurl is a normal derivation, that runs curl, inside a builder
<clever>
what is actually failing?
<immae>
The fetching of the url (it’s a 404 but it’s a wrong server configuration to say that the permission is not granted)
<immae>
clever: do you know if adapting the nix-build -> nix-daemon protocol to make it able to pass environment variable (in fixed-output context only of course) is something possible and just noone took time to do it, or something that would be refused anyway?
<clever>
immae: this is where it will get the env vars and put them into a map, it will need to somehow query the remote process
<clever>
either there, or sooner
<cinimod>
Anyone know why I get `Setup: The pkg-config package 'blas' is required but it could not be found.`?
<immae>
ok, cool thanks for the pointer
<cinimod>
How do I reproduce this locally so I can debug it?
<clever>
cinimod: open a nix-shell on that drv file, and then run `pkg-config --print-errors blas`
<clever>
cinimod: clone the rev of nixpkgs that is failing, and then `nix-build -A haskellPackages.blas-ffi`
cfricke has quit [Quit: WeeChat 2.5]
<cinimod>
clever: error: Package ‘blas-ffi-0.1’ in /Users/dom/nixpkgs/pkgs/development/haskell-modules/hackage-packages.nix:40226 is marked as broken, refusing to evaluate.
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<clever>
[clever@amd-nixos:~/apps/nixpkgs-master]$ nix-build -A blas
<clever>
[clever@amd-nixos:~/apps/nixpkgs-master]$ ls -l result/lib/
teser has quit [Ping timeout: 245 seconds]
cfricke has joined #nixos
<clever>
cinimod: i think you want to fix the pkgs.blas package, to include a blas.pc file
cfricke has quit [Client Quit]
cfricke has joined #nixos
drakonis has quit [Ping timeout: 258 seconds]
drakonis has joined #nixos
orivej has joined #nixos
<Miyu-chan>
Eep. NixOps is complaining about "REMOTE HOST IDENTIFICATION HAS CHANGED" now.
<Miyu-chan>
Any thoughts?
<Miyu-chan>
Race condition, maybe?
<joepie91>
Miyu-chan: have you recently reinstalled (from scratch) a target host?
<Miyu-chan>
I've been destroy+delete and deploying machines recently, if that's what you're asking.
<Miyu-chan>
This started happening after I advanced my nixpkgs commit.
<Miyu-chan>
I checked my known_hosts, and it's of course complaining about the lines with "NixOps auto-generated key"
<joepie91>
Miyu-chan: that warning occurs when the SSH host key of a system has changed from what it was previously; if you didn't expect that to occur, it could indicate that an attacker is in the middle of the connection trying to intercept your session
<joepie91>
Miyu-chan: however, when you've been destroying and re-instantiating systems under the same hostname or IP, it's a normal thing to happen
<clever>
or it could be that the default key type for the server has changed
<joepie91>
as each fresh installation will generate a new host key
<cinimod>
clever: I will try
<Miyu-chan>
Okay. So I at the very least know *what's* happening.
nschoe has quit [Ping timeout: 252 seconds]
<Miyu-chan>
The host key is replaced on the machines, because they should be.
<Miyu-chan>
(that much is obvious lol)
<Miyu-chan>
*But* on the local known_hosts, some of my entries are not being updated.
<joepie91>
what do you mean with "not being updated"?
* cinimod
has never created a .pc file
<Miyu-chan>
Based on how I understand it, `nixops deploy` has 2 SSH stages.
<Miyu-chan>
The first one is provisioned by AWS, which gets populated on known_hosts.
<Miyu-chan>
Then the machines generate their own keypair, and sends it back to you(?).
<Miyu-chan>
What happens is that there's a desynchronization. The machines generate their keypair, but the deployer is not notified about this.
<Miyu-chan>
Based on how I understand.
<clever>
nixops will create a keypair locally, and put the private into the metadata when spawning a machine
<joepie91>
Miyu-chan: the machine generating a keypair is unrelated to nixops, and generally occurs the first time that the sshd starts on that system. it's also not 'sent to you' until you attempt to connect to it, and then it's up to your system to decide what to do with it
<tilpner>
manveru: So you mean eval/instantiation (there's no unambiguous "running")? Sounds like you already have a benchmark then?
<freedan42x>
tilpner, runCommandNoCC: command not found
<tilpner>
perf stat -r 200 nix-instantiate ...
andreas303 has quit [Remote host closed the connection]
<tilpner>
freedan42x: It's in pkgs
<srhb>
freedan42x: You should probably read a basic "how to create a package in nix" article at this point :)
<freedan42x>
>_<
<srhb>
freedan42x: These instructions are for how to use the nix language and nixpkgs to create a few package containing your font :)
kleisli has joined #nixos
<srhb>
s/few/new
abbiya has quit [Quit: abbiya]
jgt has quit [Read error: Connection reset by peer]
andreas303 has joined #nixos
andreas303 has quit [Remote host closed the connection]
<manveru>
tilpner: well, seems like it's the substituters indeed
<manveru>
if i set `allowSubstitutes = false;` it finishes in 3 seconds instead
<manveru>
`preferLocalBuild = true;` does nothing it seems
<tilpner>
Just cache.nixos.org?
<manveru>
no
<manveru>
a bunch of cachix ones too :|
<gchristensen>
that'll do ya
<manveru>
it's sad that adding more caches makes things so slow
<tilpner>
In what order are they registered?
<gchristensen>
each one costs a penalty
<tilpner>
Won't help with misses, but I vaguely remember they're tried in order
<srhb>
Should be simple enough to race them all
<srhb>
Or at least a bunch at a time.
<tilpner>
That leaks every lookup you do
<gchristensen>
sounds like a great PR, srhb :D
<manveru>
they're in good order, but the hashes won't be in any of them anyway for this case
<tilpner>
Please make it optional if you do
<srhb>
gchristensen: I grew fungus last time I touched C++ :P
<srhb>
Granted, that was Hydra...
andreas303 has joined #nixos
<manveru>
a single cache lookup takes about 500ms afaict
chiiba has joined #nixos
<manveru>
i have 6 caches specified
<srhb>
Doesn't sound too unfair.
<chiiba>
Hey, I have multiple multiple services set up to run docker-compositions, but because they both reside in /nix/store, docker-compose thinks they're somehow the same. I'm not fluent in docker internals, but has anyone experienced something like this?
<tilpner>
Project-specific caches might be sensible?
<srhb>
chiiba: That does not sound familiar to me. Can you elaborate?
<tilpner>
gchristensen: But do you have to pass them manually? It'd be nice if direnv/lorri could do that
<gchristensen>
yes, that would be really nice :)
<gchristensen>
that was actually part of the original project description
<tilpner>
And it would only take a NIX_ env var to provide caches to use
ilmu has quit [Ping timeout: 276 seconds]
<manveru>
heh
<tilpner>
So manveru, are you annoyed enough? c.c
drakonis_ has joined #nixos
<chiiba>
I think the problem I'm running into is that both compositions are under the same namespace. So because both compositions have a container "postgres" - only one is being created, because the fully qualified container name becomes store_postgres.
<zfnmxt>
Does the NIXOS_CONFIG env variable override setting nixos-config=... in NIX_PATH?
drakonis1 has joined #nixos
sb0 has joined #nixos
<manveru>
zfnmxt: seems like it
<tilpner>
manveru: You could push on the Rust-in-Nix thing
<tilpner>
(But that makes no sense here)
<zfnmxt>
manveru: Thanks. Guess I I could've just tested it myself tbh =/
zupo has joined #nixos
<gchristensen>
oh I thought you were talking about extending lorri :)
<tilpner>
(Only to avoid C++ in the far future)
<manveru>
zfnmxt: i just checked the source anyway ;)
<zfnmxt>
Well that's even better!
<srhb>
zfnmxt: There's also -I nixos-config=... :)
drakonis has quit [Ping timeout: 258 seconds]
cfricke has quit [Quit: WeeChat 2.5]
<zfnmxt>
srhb: I know, that's what I've been doing. Time to set some variables so I don't always have to do that!
tmaekawa has joined #nixos
<srhb>
Ah, okay :)
<manveru>
tilpner: i'm not annoyed enough i guess... but if things continue like now, we'll have a spearate cachix cache for every haskell/rust project, grinding things to a halt
<tilpner>
manveru: You could use any language you want to provide a proxy cache to Nix
<srhb>
tilpner: That's a really good idea.
<tilpner>
That could do anything, including aggressively caching misses, parallel lookup
tmaekawa has quit [Client Quit]
<srhb>
In fact, I think nginx can do this out of the box...
<chiiba>
srhb: I found the problem. Docker-composition containers are labeled from two things - project name & container name. If the project name isn't declared using an environment variable COMPOSE_PROJECT_NAME, then the docker-compose.yml directory basename is used instead. https://docs.docker.com/compose/reference/envvars/#compose_project_name
drakonis_ has quit [Ping timeout: 276 seconds]
<srhb>
chiiba: Right right.
<srhb>
chiiba: That makes sense :)
<chiiba>
Thanks for the moral support tho' :D :)
<zfnmxt>
Hm, so there's no variable for setting your nixpkgs path?
<zfnmxt>
(Aside from editing NIX_PATH with the path)
<srhb>
zfnmxt: nix.nixPath
Lukas4452 has quit [Ping timeout: 245 seconds]
<zfnmxt>
srhb: Oh, even better. Thanks.
<srhb>
Be wary that it's one of those things that can be a bit mind-bendy because it changes your system's behaviour wrt. Nix _after_ activation.
<manveru>
tilpner: so that'd basically MitM, so it needs a separate cert, right?
zzz has joined #nixos
<zfnmxt>
Right.
hyper_ch2 has quit [Remote host closed the connection]
cinimod has quit [Ping timeout: 272 seconds]
<zfnmxt>
Oh, you can set nixos-config in nix.nixPath too. Cool.
<srhb>
zfnmxt: Yeah, it's the entire variable.
jgt has joined #nixos
<roberth>
manveru, tilpner: we can solve this problem for cachix in the future. How bad is it for you now?
<aanderse>
is the manual broken in master for anyone else?
dansho has quit [Quit: Leaving]
<gchristensen>
which?
<manveru>
roberth: well, right now i have an average of 3s overhead for every derivation that doesn't specify `allowSubstitutes = false` with 6 caches
<manveru>
roberth: i'm not sure what a good solution would be, other than parallel lookups in nix itself, or a cachix proxy that combines multiple caches...
<srhb>
I guess we never test that entrypoint anymore.
<gchristensen>
srhb: I test it often :P
<srhb>
gchristensen: Sorry, by "we" I meant a robot. :D
<gchristensen>
aye
<aanderse>
i have broken the manual in the past, so i'm just glad it wasn't something i have done :P
<manveru>
roberth: cool would be if one could `cachix use` for only the duration of a single `nix-build` without having to mess with the nix.conf afterwards :)
o1lo01ol1o has quit [Remote host closed the connection]
inkbottle has joined #nixos
drakonis has joined #nixos
o1lo01ol1o has joined #nixos
Okinan has quit [Ping timeout: 276 seconds]
<jonge[m]>
i have a minor and non-intrusive PR open for a month now and i feel it is ignored, which leaves me with having it to rebase every now and then, which again feels like unnecessary work. is there anyone i could nudge about nodejs packages? https://github.com/NixOS/nixpkgs/pull/65306
<srhb>
jonge[m]: Sorry about that. I think it's mostly that people who are not familiar with nodePackages tend to not feel very secure in merging these. I would suggest pinging some of the people who have previously touched the tree to review and approve, that should speed things up.
<jonge[m]>
srhb: i understand that. thank you
domogled has joined #nixos
<srhb>
jonge[m]: If you still get no responses after a week or so after pinging me, feel free to ping me on the PR, and I'll see what I can do :)
<srhb>
er, after pinging them*
<jonge[m]>
srhb: i mentioned some people who commited things in the last weeks to the same files. if nothing happens for a week i will mention you, too. thank you. :)
boeg has joined #nixos
<srhb>
Sure! Fingers crossed :)
domogled has quit [Client Quit]
fusion809 has quit [Remote host closed the connection]
vika_nezrimaya has joined #nixos
ottidmes has quit [Ping timeout: 248 seconds]
<freeman[w]_>
does anyone know which is the NixOS GUI archive manager that can use the executable from p7zip?
<vika_nezrimaya>
What's the correct Inputs array for Python setuptools' setup_requires?
<emptyflask>
is there any explanation why nix-prefetch-git (with a private github repo, using https://xxxx:x-oauth-basic) would return a different hash for the same revision than bundix?
<fresheyeball>
is there anything special I need to do to get hie to see packages supplied by the nix-shell?
<exarkun>
gchristensen: I just applied your nixops/nixPath thing to my system. Took me a minute to realize it can't affect my environment until I log in fresh but now it seems to be working well. Thanks again.
veske has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
<joepie91>
Ariakenom: update as in installing updates, or as in upgrading to a newer NixOS release?
noudle has joined #nixos
<Ariakenom>
joepie91: especially release. but first one too
<joepie91>
Ariakenom: so assuming that you've specified your packages in your configuration.nix' systemPackges (and *not* installed them using `nix-env`), installing updates is basically just a matter of running `nixos-rebuild switch` -- it will re-evaluate your configuration with whatever is the NixOS package set at that time, build anything that's missing, and switch to the new system build... which is functionally equivalent to
<joepie91>
"installing updates"
<{^_^}>
[nixpkgs] @peti pushed to master « Add a build for cabal-install-3.0.0.0 beta release. »: https://git.io/fjN2N
<{^_^}>
[nixpkgs] @peti pushed 34 commits to haskell-updates: https://git.io/fjN2A
<joepie91>
Ariakenom: for switching releases, the process is *almost* the same, but you first need to change what package set your Nix is pointed at as nixpkgs (namely, you switch it to a newer release), and you pass a flag to nixos-rebuild that tells it explicitly to grab the current package set and not use the local cache
<joepie91>
conceptually there's no difference between installing updates and upgrading to a new release; in both cases you re-evaluate the system config with a new package set. the only practical difference is that for a release upgrade, you first change what package set to look at :)
<Ariakenom>
ah yes. i missed the upgrade flag
<Ariakenom>
what about things installed with nix-env?
noudle has quit [Ping timeout: 246 seconds]
iyzsong has quit [Ping timeout: 250 seconds]
<joepie91>
Ariakenom: I /think/ there may be an update flag for that, but I'm not sure. I personally avoid nix-env entirely, as using it means I can't reproduce my system from its config :)
freedan42x has quit [Ping timeout: 272 seconds]
hyper_ch2 has quit [Remote host closed the connection]
Ariakenom has quit [Remote host closed the connection]
<attente[m]>
hi, is anyone here having issues with their external monitor not working? i have a thinkpad x1 carbon, 6th gen running nixos with linuxPackages_latest
freedan42x has joined #nixos
<joepie91>
attente[m]: it has worked fine for me, but multi-monitor support is usually heavily dependent on your desktop environment
Ariakenom has joined #nixos
<Ariakenom>
im not getting it to work? nixos-version still says 18.09.2574.a7e559a5504 (Jellyfish)
<attente[m]>
joepie91: the thing is i can't even get my tty mirrored, so this is before any sort of desktop is launched
<Ariakenom>
I was expecting 19.03 instead
<joepie91>
attente[m]: have you ruled out a hardware / BIOS/UEFI config issue?
<attente[m]>
i haven't, but i'm not sure how to do that, joepie91
<joepie91>
I don't know the specifics but AFAIK most systems default to mirroring on external display ports
<joepie91>
attente[m]: well, try booting into a different system, see if it works there
<joepie91>
something not NixOS
<joepie91>
ideally something like Windows, but even eg. an Ubuntu live CD would give you useful data points
<attente[m]>
joepie91: ok, i'll try a live cd, thanks for the suggestion
<joepie91>
you can also have a quick look through your BIOS/UEFI config on bootup, to see if there are any obvious options regarding display ports that look misconfigured
<attente[m]>
that's also a good idea, i'll try that as well
<joepie91>
attente[m]: btw, to be clear, I'm not saying that it *can't* be a NixOS issue - but no TTY mirroring is sufficiently weird that a hardware/config issue seems more plausible to me :)
<srhb>
Ariakenom: What does `sudo nix-channel --list` say?
<joepie91>
as my understanding is that that is done before the OS even boots
<attente[m]>
yeah, that was my fear too
endformationage has joined #nixos
<srhb>
attente[m]: Have you rummaged around firmware settings?
<{^_^}>
[nixpkgs] @worldofpeace pushed 6 commits to master: https://git.io/fjNa2
<alexbakker>
the libvirtd module does the same for the "onShutdown" option, so I wonder why this doesn't work
nschoe has joined #nixos
<alexbakker>
is there a way to inspect the generated environment file?
<joepie91>
attente[m]: the issue described in that thread does look like a plausible cause
<joepie91>
attente[m]: in that case, can you check if an issue already exists on the nixpkgs repo, and if not, create a new one referencing that thread?
<alexbakker>
so the ON_BOOT environment variable does actually get set properly
<alexbakker>
perhaps this is a libvirt bug
<freeman[w]_>
currently smartgit package is broken (I will try to submit a PR fix). I did manage to install it on another PC, but I do not remember how. Anyone can take a guess what I could have possibly done?
o1lo01ol1o has joined #nixos
ilmu has quit [Ping timeout: 250 seconds]
<{^_^}>
[nixpkgs] @adisbladis opened pull request #67355 → nixos/desktop-managers/xterm: Disable by default → https://git.io/fjNa5
Ariakenom has quit [Remote host closed the connection]
chiefgoat has joined #nixos
chiefgoat has quit [Client Quit]
zupo has joined #nixos
chiefgoat has joined #nixos
<alexbakker>
oh I see, libvirt-guests.sh needs to be patched, as was done for ON_SHUTDOWN
<alexbakker>
I'll submit a PR adding an onBoot option
<{^_^}>
[nixpkgs] @lheckemann pushed 117 commits to openssl-1.1: https://git.io/fjNax
chiefgoat has quit [Client Quit]
<attente[m]>
joepie91: sure, will do
<attente[m]>
once i test with another live cd first
<teser>
hi, I'm using: nix-shell -p 'with import <nixpkgs> { }; [ gcc_multi gnumake automake autoconf glibc z3 ]' to get the shell, but then when running make of certain program, I get this issue:
<teser>
> fatal error: gnu/stubs-32.h: No such file or directory
<{^_^}>
error: syntax error, unexpected ':', expecting ')', at (string):262:12
<teser>
Shouldn't that be included in the glibc ?
obadz has joined #nixos
kmein has joined #nixos
phreedom_ has joined #nixos
<infinisil>
teser: Write a shell.nix file instead with `with import <nixpkgs> {}; stdenv_32bit.mkDerivation { name = "shell"; buildInputs = [ automake autoconf z3 ]; }`
<infinisil>
The normal stdenv which nix-shell -p uses doesn't support 32 bit
<pie_>
tazjin: theres a discussion at the nixos tent right now about like improving the maintainership process or something if you want to join
<pie_>
tazjin: though the goal seems to right now be figuring out what the current state is, whatever that means
iqubic has quit [Ping timeout: 276 seconds]
phreedom has quit [Ping timeout: 260 seconds]
<teser>
infinisil, +1 thanks
fendor has joined #nixos
buckley310 has quit [Quit: leaving]
buckley310 has joined #nixos
buckley310 has quit [Client Quit]
kleisli has joined #nixos
erictapen has joined #nixos
liori has quit [Remote host closed the connection]
npmccallum has quit [Remote host closed the connection]
iqubic has joined #nixos
pie_ has quit [Ping timeout: 250 seconds]
___laika has quit [Ping timeout: 272 seconds]
<colemickens>
why is nix-daemon having its way with my CPU when its just downloading from a binary cache? is nix-daemon doing the extraction or is there a cpu-hungry loop in here or what?
<iqubic>
So I have lorri installed. I have a systemd user service running lorri daemon. How do I tell lorri to watch a particular file?
freedan42x has quit [Remote host closed the connection]
Thra11 has joined #nixos
casaca has quit [Ping timeout: 272 seconds]
casaca has joined #nixos
freeman42x has joined #nixos
<samueldr>
colemickens: uncompressing I think
Ariakenom has joined #nixos
rpg has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<{^_^}>
[nixpkgs] @disassembler pushed 2 commits to master: https://git.io/fjNKb
orivej has quit [Ping timeout: 268 seconds]
<ejpcmac>
Hello! Is there a way to get `mkfs.exfat` while still using `exfat-nofuse` when mounting ExFAT volumes?
pie_ has quit [Ping timeout: 272 seconds]
nschoe has joined #nixos
<alexarice[m]>
Not sure if I am explaining this well but is there a way to append something to an option set with `mkDefault` without just copying it out
<infinisil>
alexarice[m]: Not sure what you mean by copying it out
<alexarice[m]>
infinisil: say there is `config.a = mkDefault [ "foo" ]` somewhere and I want to append "bar" to that list without just writing `config.a = [ "foo" "bar" ]`
<hyperfekt>
alexarice[m]: I have no idea if this would actually work but mkDefault is just mkOverride 1000 so I'd try combining mkAfter with mkDefault for your 'appendage'.
___laika has joined #nixos
<iqubic>
So, I'm finding myself unable to run a 'nixos-rebuild switch' due to this error: error: assertion failed at /nix/store/d3yysrqlm4sf2mhlcpj12b78vcqpgxqa-nixos-19.09pre190234.dfc6d580bcc/nixos/pkgs/os-specific/linux/phc-intel/default.nix:4:1
<alexarice[m]>
I guess `config.a = mkDefault [ "bar" ]` would work though it feels a bit wrong
<alexarice[m]>
I almost want to escalate the priority of the default somehow
<alexarice[m]>
I'll have a think
<iqubic>
Anyone know how to fix my issue?
<infinisil>
alexarice[m]: Yeah there's no easy way to do that
<alexarice[m]>
was worth an ask
alex`` has quit [Ping timeout: 248 seconds]
knupfer has joined #nixos
alex`` has joined #nixos
drakonis has quit [Ping timeout: 250 seconds]
<hyperfekt>
iqubic: The source at that line says your kernel version is below 4.10.
<Thra11>
I figure with the curl error problem there's no point repeatedly trying to ask cache.nixos.org for a source archive which is available elsewhere
nschoe has quit [Ping timeout: 245 seconds]
is_null has quit [Ping timeout: 258 seconds]
<Thra11>
For future reference, is there an option which will try to fetch binaries, but not 'mirrored' sources? I'm guessing not, as nix doesn't make such a distinction?
casaca has quit [Ping timeout: 245 seconds]
<infinisil>
Thra11: Possibly with `NIX_HASHED_MIRRORS="" nix-...`
noam_ has quit [Remote host closed the connection]
<inkbottle>
alexarice[m]: I so far am a debian user, and considering the possibility of exploring other distro; I've been using debian for a very very long time; but have nothing of the "system wizard"
<joepie91>
inkbottle: feel free to ask questions - NixOS is quite different from distros like Debian, so there'll probably be a lot of new concepts to wrap your head around :)
<{^_^}>
[nixpkgs] @peti opened pull request #67366 → WIP: update mailman to version 3.2.2 → https://git.io/fjNil
<joepie91>
(it's worth it, imo, but it's still a learning curve)
<inkbottle>
Hedgework: what is written or printed; but I rather like fonts than fountain pen (as a sort of manly expensive object); I do use a fountain pen though, but probably not of the sort expected for a fountain pen geek
<gchristensen>
does NixOS have an service for something I can put in HTTP_PROXY for testing a PR?
<{^_^}>
[nixpkgs] @jtojnar opened pull request #67367 → nextcloud-client: clean up → https://git.io/fjNiR
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/fjNii
<Miyu-chan>
And it also matters how often you build things, especially ones that are not in the hydra, since you'll also have to populate /nix/store with the source code.
<samueldr>
I think for laptops it makes sense not to subdivide, or if you do, with a FS you can easily expand as needed, like I imagine ZFS does
<Miyu-chan>
Yeah, that's what I'd do. I have a 128/128 split on my desktop, with one 128 being used as a ZFS cache. :)
<samueldr>
splitting partitions in a one-drive scenario I would always end up needing more space for either of the partition for my use cases :)
<gchristensen>
(ZFS doesn't expand by default as it doesn't limit by default)
mawis has quit [Quit: WeeChat 2.4]
kleisli_ has joined #nixos
Zer0xp has quit [Read error: Connection reset by peer]
chr15 has quit [Read error: Connection reset by peer]
selfsymmetric-pa has quit [Remote host closed the connection]
<inkbottle>
ZFS proprietary?
<{^_^}>
[nixpkgs] @etu closed pull request #67369 → sequeler: Fix missing icons while not in gnome → https://git.io/fjNiw
<inkbottle>
I've read sth in the news about that
<inkbottle>
cna't remember though
<inkbottle>
can't*
<joepie91>
inkbottle: +1 on just having one large partition
kleisli has quit [Ping timeout: 276 seconds]
<joepie91>
since state is pretty well-contained in NixOS (that's the point of it, after all!) the whole "persist my home directory across reinstallations" argument for a separate home partition kind of goes out the window
kleisli_ has quit [Remote host closed the connection]
<joepie91>
LVM is also always an option if you want more flexibility
polman has quit [Read error: Connection reset by peer]
kleisli_ has joined #nixos
<inkbottle>
OK
<joepie91>
inkbottle: one very important thing to keep in mind is that your boot partition needs to be sufficiently big (and probably cannot be an LVM partition)
Zer0xp has joined #nixos
<joepie91>
if you have one single non-LVM partition this isn't an issue, but if you use LVM or otherwise need a separate boot partition for technical reasons, you'll really want to give it at least 1GB, due to how the generation/rollback system works
<joepie91>
you'll end up with multiple kernel images until you clean up old system generations
<joepie91>
and this is something that is a bit painful to fix afterwards :)
chr15 has joined #nixos
<joepie91>
so the same applies for LUKS, and probably md-raid, and so on
Havvy has joined #nixos
<inkbottle>
sure; first thing I'll go to the easy way.
LouisDK has joined #nixos
marcinkuzminski_ has quit []
marcinkuzminski_ has joined #nixos
<LouisDK>
I'm trying to build a package from a local repo in NixOS, however I get this error: "./configure: .: line 5935: can't open '/etc/os-release': No such file or directory "
<clever>
LouisDK: that wont exist due to the nix sandbox, what does configure want to do with it?
<LouisDK>
The file does exist. What is the best way to make the builder aware of it in a proper way?
<LouisDK>
clever, 2 sec. Will look.
polman has joined #nixos
growpotkin has joined #nixos
waleee-cl has quit [Quit: Connection closed for inactivity]
shabius has quit [Quit: Leaving]
marcinkuzminski_ has quit [Client Quit]
marcinkuzminski has joined #nixos
<LouisDK>
I want to look for either "centos","debian" or "ubuntu" in it and if untrue do something else.
<LouisDK>
@clever ^
<clever>
LouisDK: does it have an argument to override things?
<clever>
what will it do in each situation?
is_null has quit [Ping timeout: 272 seconds]
shabius has joined #nixos
shabius has quit [Max SendQ exceeded]
<LouisDK>
@Clever: I'm trying to build mythtv 30.0 which have support for libblueray introduced. Output: https://pastebin.com/qsuXrmUj - I can't figure out where VERSION_ID is sat in the config file.
Okinan has quit [Quit: Konversation terminated!]
shabius has joined #nixos
<LouisDK>
@Clever I'll try to include libbluray and see if that'll solve it.
<clever>
LouisDK: if you add libbluray to the buildInputs, the if condition on line 2 will find it, and end of story
<clever>
LouisDK: VERSION_ID is a variable inside os-release
kleisli_ has quit [Ping timeout: 246 seconds]
<clever>
$ cat /etc/os-release
<clever>
VERSION_ID="19.09pre179307.bc94dcf5002"
<clever>
but thats not pure, and why are they even doing that??
Ariakenom has joined #nixos
o1lo01ol1o has joined #nixos
<clever>
i think it just assumes an os of a certain version has a certain library
<clever>
but that wont be true for nix
<LouisDK>
@clever OId hacky sw I guess. There's comments in the configure file like "# This is hopefully temporary". Really nice.
<clever>
heh
pie_ has quit [Ping timeout: 268 seconds]
LouisDK has quit [Remote host closed the connection]
klntsky has quit [Remote host closed the connection]
civodul has quit [Quit: ERC (IRC client for Emacs 26.2)]
<LouisDK>
@clever seems like a have to modify the configure file from the source tar gz. As /nix/store is mounted ro I need some way to do this in the nix file. Can you point be in a way on how to due that?
pie_ has joined #nixos
<clever>
LouisDK: the unpackPhase copies the configure script to . before running it
<clever>
so the patchPhase can just edit "configure" directly, its in the working dir
xkapastel has quit [Quit: Connection closed for inactivity]
<EsperLily>
LouisDK: you can define your own patchPhase
pie_ has quit [Remote host closed the connection]
<ashkitten>
hmm, i wonder if i should package glimpse for nix
<clever>
ashkitten: package all the things
<clever>
LouisDK: also, irc doesnt highlight when you @clever, putting the @ there is making it less noticable
<ashkitten>
yes, but glimpse hasn't had a release yet :p
<EsperLily>
well, defining your own patchPhase will break the existing patches attribute in the derivation you linked to. You could just define an actual patch too, or you could use a postPatch phase to do the custom patching
pie_ has joined #nixos
<EsperLily>
is it just me, or is it pretty weird for a Nix derivation to fetch a patch from AUR?
Zer0xp has quit [Ping timeout: 245 seconds]
<LouisDK>
EsperLily: Yearh. I too think it's kinda weird. I'll move it to pkgs/applications/video/mythtv/ instead.
chr15 has quit [Ping timeout: 246 seconds]
ilikeheaps has quit [Quit: Konversation terminated!]
<jw358>
anyone using github actions
<jw358>
i would like to figure out some solution for having nix available
<jw358>
the problem with using the normal install script is that it segfaults often
<EsperLily>
o_O
<EsperLily>
I haven't really looked at GitHub Actions yet, but if they're using docker containers, can you base your action on a Nixery-provided container? https://nixery.dev
<ashkitten>
hmph. i don't know how to build this, gimp builds from a release tarball and if i add autoreconfHook it complains about gtk-doc.make
<ashkitten>
and there aren't any glimpse release tarballs yet
<ashkitten>
:/
<clever>
ashkitten: the release tarballs are usually just the result of running automake and autoconf in the source dir
<ashkitten>
clever: isn't that what autoreconfHook does?
<clever>
yep
freeman42x has quit [Ping timeout: 252 seconds]
<ashkitten>
sooo why does it complain
<ashkitten>
during automake
<clever>
it needs extra deps
<ashkitten>
i added gtk-doc to the build inputs
<clever>
does gtk-doc actually contain gtk-doc.make?
<ashkitten>
good question
<ashkitten>
yes
<clever>
and how does automake search for it?
<ashkitten>
uhh
<ashkitten>
idk
<ashkitten>
specific error is `automake: error: cannot open < gtk-doc.make: No such file or directory`
<clever>
sounds like its looking in the current dir?
<ashkitten>
i don't know
jgt has quit [Ping timeout: 250 seconds]
<ashkitten>
looks like i3ipc-glib.nix runs gtkdocize in preAutoreconf
<ashkitten>
i'll try that
<ashkitten>
yep, sure enough
<ashkitten>
hopefully this works
<ashkitten>
it's building now
<ashkitten>
and failed
judson_ has quit [Ping timeout: 248 seconds]
judson has quit [Ping timeout: 248 seconds]
<ashkitten>
needed libxslt
work_ has joined #nixos
LouisDK has quit [Remote host closed the connection]
polman has quit [Ping timeout: 245 seconds]
<tazjin>
EsperLily: that should work!
<tazjin>
I haven't received an invite to Github Actions yet to try it though
<gchristensen>
could also use the NixOS/nix images?
<tazjin>
the limitation that some people might run into is that Nixery doesn't yet support `withPackages`-style calls, but I'll be looking into that soon (probably after cccamp) - https://github.com/google/nixery/issues/31
<tazjin>
gchristensen: two different use-cases I think (having Nix in the image vs. having an image with some tools, built by Nix - where the latter would otherwise be accomplished by having Nix in the image and installing tools)
polman has joined #nixos
<gchristensen>
the first use case seemed like the one they wanted?
<EsperLily>
tazjin: I was thinking you'd use a Nixery image that contains your prereqs as a base, and then build on top of that with the customizations you need (like withPackages)
magnus1 is now known as skorpy2009
<EsperLily>
heck, could you use something like nixery.dev/shell/nix as just a generic Nix base image? Dunno if that's any better than the NixOS/nix image though
<tazjin>
EsperLily: Nix makes a bunch of environment assumptions that your image needs to fulfill, if you look at Nixery's own default.nix you'll see those
<tazjin>
so if you ran shell/nix you'd need to set those up in the image
<tazjin>
in which case the other Nix image might make more sense
<tazjin>
There's an argument to be made that the public instance should know about stuff like this and to the right thing for packages like Nix itself