<pie_>
manveru, re: ruby stuff, I got super sidetracked so I made no progress, I guess what I want is to not need to figure out how to do things with other language tooling if i can just do it with nix
<Ralith>
samueldr: uaccess tagging is a lot better than mode 0666, I think, but last time I mucked with it it didn't work in extraRules for some stupid ordering reason
<samueldr>
Ralith: most probably! which is why I was trying to point out NOT to use that :)
<Ralith>
^^
<samueldr>
techieAgnostic: sorry, can't really say; games can sometimes do tricky stuff :/
<Ralith>
ideally I think you have a package that introduces the right rules file in the right sequence and stick that in udevPackages
<Ralith>
or an option that boils down to it, like the recently added u2f one
<samueldr>
Ralith: mind that this was added to my nixos config during the time I was learning nix, nixpkgs and nixos :) copied from a THEN cargo-culted file I dragged for a small while around
spear2 has quit [Quit: Leaving]
<samueldr>
so, yeah, about right for enterprise software
<techieAgnostic>
samueldr: no worries, thanks :)
jperras has quit [Ping timeout: 252 seconds]
jperras has joined #nixos
derped has quit [Quit: Leaving.]
rouma has quit [Ping timeout: 272 seconds]
Zajcev has quit [*.net *.split]
fractal has quit [*.net *.split]
shoogz has quit [*.net *.split]
cross_ has quit [*.net *.split]
mutantmell has quit [*.net *.split]
c_wraith has quit [*.net *.split]
_cyril_ has quit [*.net *.split]
hxrts has quit [*.net *.split]
shlevy has quit [*.net *.split]
Aleksejs has quit [*.net *.split]
rgrau has quit [*.net *.split]
szicari has quit [*.net *.split]
adelbertc has quit [*.net *.split]
msgctl has quit [*.net *.split]
dev2 has quit [*.net *.split]
jonafato has quit [*.net *.split]
ericnoan has quit [*.net *.split]
Guest50286 has quit [*.net *.split]
lvmond_ has quit [*.net *.split]
cwraith has joined #nixos
Zajcev_ has joined #nixos
rgrau_ has joined #nixos
szicari has joined #nixos
shlevy has joined #nixos
dbmikus_ has joined #nixos
adelbertc_ has joined #nixos
dev2 has joined #nixos
alpha_sh has joined #nixos
lvmond has joined #nixos
swistak35 has joined #nixos
jonafato has joined #nixos
oldnborg has joined #nixos
mutantmell has joined #nixos
cross has joined #nixos
msgctl has joined #nixos
Aleksejs has joined #nixos
ericnoan has joined #nixos
exarkun_ has quit [Ping timeout: 252 seconds]
ericnoan has quit [Changing host]
ericnoan has joined #nixos
Unode has quit [Ping timeout: 260 seconds]
hxrts_ has joined #nixos
_cyril_ has joined #nixos
shoogz- has joined #nixos
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Unode has joined #nixos
fractal has joined #nixos
exarkun_ has joined #nixos
wpcarro has quit [Ping timeout: 246 seconds]
Supersonic has quit [Ping timeout: 252 seconds]
shoogz- is now known as shoogz
wpcarro has joined #nixos
dbmikus_ has quit [Ping timeout: 240 seconds]
Supersonic has joined #nixos
<{^_^}>
[nixpkgs] @qolii opened pull request #46635 → nbd: add which to buildInputs → https://git.io/fAixA
__lawlesseel__ has joined #nixos
jbaum98 has quit [Quit: Connection closed for inactivity]
lawlesseel has quit [Remote host closed the connection]
hamishmack has joined #nixos
wpcarro has quit [Ping timeout: 252 seconds]
wpcarro has joined #nixos
patrl has quit [Quit: WeeChat 2.1]
patrl has joined #nixos
<patrl>
Ugh. I'm getting "Warning: Nix search path entry '$HOME/.nix-defexpr/channels' does not exist, ignoring" whenever I sudo nixos-rebuild switch on 18.09
<patrl>
no warning if I `nixos-rebuild switch` while logged in as root
<Myrl-saki>
paIt's an open issue.
<Myrl-saki>
patrl:
<patrl>
Myrl-saki: oh cheers. I guess I missed that
jtojnar has quit [Quit: jtojnar]
<rotaerk>
hmm could someone take a look at PR #46555
<{^_^}>
[nixpkgs] @ju1m opened pull request #46637 → Fix rmilter's default bindSocket.path and bindSocket.address. → https://git.io/fAPf4
rauno has quit [Ping timeout: 250 seconds]
counting1ort has joined #nixos
brejoc has joined #nixos
countingsort has quit [Ping timeout: 244 seconds]
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
dbmikus_ has quit [Ping timeout: 252 seconds]
sbdchd has joined #nixos
sbdchd has quit [Remote host closed the connection]
sbdchd has joined #nixos
oscarvarto has quit [Quit: Leaving]
jperras has joined #nixos
jperras has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @ryantm opened pull request #46638 → vimPlugins: warn if alias is used inside overrides → https://git.io/fAPJI
fragamus has joined #nixos
jtojnar has quit [Quit: jtojnar]
Havvy has joined #nixos
jtojnar has joined #nixos
Havvy has quit [Quit: Computer Restarted or Restarting IRC]
Havvy has joined #nixos
endformationage has quit [Quit: WeeChat 1.9.1]
drewr has quit [Ping timeout: 250 seconds]
<ldlework>
,locate lspci
<ldlework>
...
<ldlework>
,find lspci
<jackdk>
pciutils
<adamantium>
ty
<adamantium>
the jackdk bot has some serious lag
<ldlework>
lol
<ldlework>
i think jackdk is a person
<adamantium>
;)
<jackdk>
I have you all fooled.
<ldlework>
jackdk: How do you feel about you have me all fooled?
drewr has joined #nixos
<jackdk>
ldlework: I would appreciate it if you would continue.
<ldlework>
;)
<Havvy>
I have a bunch of nix files in ~/.nixpkgs/shell that make derivations and I want to make sure that if I `nix-collect-garbage`, that it won't eat up the things that those derivations depend on.
<ldlework>
jackdk: i wonder when we'll be too old to tell that joke and have anyone get it
<ldlework>
Havvy: something about creating a gc-root
<Havvy>
I keep seeing `nix-instantiate . --indirect --add-root $PWD/shell.drv` but I have no clue what those do.
<jackdk>
ldlework: Sooner than I'd like but probably longer than I fear. Just need to mentor the right people in the old ways.
<ldlework>
jackdk: i actually have an elisp irc bot somewhere around here somewhere that will play doctor heh
<jackdk>
ldlework: I generated my responses to you through a M-x doctor session heh
<lesh>
hey, nixops doesn't seem to create a "full nixos install" with configuration and nix tools, how can I create those on a provisioned machine? tnx
<Taneb>
lesh: I think the idea of nixops is you manage the configuration with nixops, rather than on the provisioned machine
Dedalo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
sigmundv__ has joined #nixos
bennofs has joined #nixos
<srhb>
That's right. You can always convert it (very carefully...) to a self-managed machine afterwards.
datakurre has joined #nixos
<lesh>
Taneb: yes I understand, but in this case I know what I'm doing :) I used nixops from a random virtual machine to provision a hetzner machine (to avoid installing manually) which will be a hypervisor, further nixops operations I'm planning to do from there.. unless I can somehow manage libvirt guests via config on my laptop.. but from what I've read I need to have nixops running locally on hypervisor if I'd
<lesh>
like to use libvirtd/kvm
<lesh>
so basically after bootstraping this big machine via nixops, I'd like to "move in", and do furtner nixops from there :)
<sphalerite>
lesh: I'd say stick your config on it in /etc/nixos/configuration.nix, then if you have nixos-rebuild just use that. Otherwise nix-build '<nixpkgs/nixos>' -A system && ./result/bin/switch-to-configuration switch
<lesh>
sphalerite: haha tnx, few mins ago I broke the machine by sticking the config on it and doing nixos-install --root / and crossing my fingers ;)
<lesh>
unsuprisingly that didn't work :)
<sphalerite>
how did it break?
fendor has joined #nixos
<lesh>
I couldn't ssh into it anymore.. not sure what had happened. I'm being more careful now so will know if something goes wrong
fendor has quit [Remote host closed the connection]
<lesh>
aha, both nixos-rebuild (which is what I tried first before) and nix-build fail due to 'error: file 'nixos-config' was not found in the Nix search path'
<symphorien>
add -I nixos-config=/etc/nixos/configuration.nix as an argument
<lesh>
symphorien: amazing, nixos-rebuild is doing it's thing now. thank you and sorry for newbie questions, I'm two days into nixos
graphene has quit [Read error: Connection reset by peer]
xenog has joined #nixos
graphene has joined #nixos
abueide has quit [Ping timeout: 244 seconds]
xchg111 has joined #nixos
trcc has quit [Remote host closed the connection]
trcc has joined #nixos
trcc has quit [Ping timeout: 244 seconds]
mayhewluke has quit [Ping timeout: 244 seconds]
<teto>
I have a nixos VM I installed manually. I want to manage it via nixops, is there some feature to send the nixops key to that machine or on the opposite force nixops to use another set of keys ?
trcc has joined #nixos
mayhewluke has joined #nixos
trcc has quit [Remote host closed the connection]
trcc has joined #nixos
Enzime has left #nixos ["WeeChat 1.4"]
Enzime has joined #nixos
<ikwildrpepper>
teto, but if you can connect to root via ssh, you can just use the none backend
<ikwildrpepper>
and it will just work
<ikwildrpepper>
note that you will need to make sure your network will have all config in its network, such as the hardware-configuration.nix that you manually generated
<ikwildrpepper>
the none backend basically works on any ssh reachable host, and does nothing wrt provisioning etc
<{^_^}>
[nixops] @rbvermaa pushed 0 commits to refs/tags/v1.6.1: https://git.io/fAP8h
martingale has joined #nixos
martingale has quit [Ping timeout: 245 seconds]
simukis has joined #nixos
trcc has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @rbvermaa pushed to release-17.09 « Fix Debian Jessie package list hashes. »: https://git.io/fAP4l
<symphorien>
systemd contains udev, this could do ?
<Baughn>
Oh, and also this only happens if the system has been suspended in a user session.
<Baughn>
If it's suspended at the gdm login screen, that's fine, but if it has been suspended in gnome user session AND you use nvidia AND you have systemd 239 AND wayland isn't explicitly disabled, then gdm will repeatedly crash when started, such as after a reboot.
<Baughn>
Resetting the machine doesn't help, but a power cycle does.
gamble has quit [Quit: Page closed]
<Baughn>
Computers. Were. A. Mistake.
<exarkun1>
Only in the vast majority of cases.
reinzelmann has quit [Quit: Leaving]
xenog has joined #nixos
patrl has quit [Quit: WeeChat 2.2]
patrl has joined #nixos
xchg111 has quit [Quit: Leaving]
patrl has quit [Client Quit]
patrl has joined #nixos
patrl has quit [Client Quit]
patrl has joined #nixos
patrl has quit [Changing host]
patrl has joined #nixos
<MichaelRaskin>
Baughn: so far all your complaints are about the graphical mode, right?
patrl has quit [Client Quit]
<Baughn>
MichaelRaskin: I would also like to raise a complaint about the density of random service crashes and errors in my logs.
<MichaelRaskin>
I wonder how many of them are only needed to support gnome session…
<Baughn>
Oh, most of them.
patrl has joined #nixos
patrl has quit [Changing host]
patrl has joined #nixos
<Baughn>
Unfortunately Gnome is the only thing I can run that doesn't lag.
Dedalo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Baughn>
I have one of the fastest desktop systems ever made, and KDE is laggier than Windows 3.1 was.
<MichaelRaskin>
I would be quite surprised if Ratpoison lagged on that system
patrl has quit [Client Quit]
patrl has joined #nixos
<MichaelRaskin>
I would even be slightly surprised if XFCE was slower than Gnome on any given system
<Baughn>
XMonad does, XFCE does, haven't tried Ratpoison but I bet it would.
<Baughn>
It's not an issue with the DE, really. It's an issue with nvidia.
<Baughn>
Unfortunately I have to use nvidia.
<Baughn>
Gnome gets whatever-it-is right, but not even the gnome devs seem to know why.
<MichaelRaskin>
Hm. XMonad + compton maybe?
<exarkun1>
How do I change system configuration (/etc/nixos/configuration.nix) in a nixops node definition?
<Baughn>
Tried it. It's no good.
<exarkun1>
Or, higher level, how do I enable ccache on a node using nixops?
<Baughn>
exarkun1: Nixops doesn't need or use /etc/nixos.
<sphalerite>
exarkun1: just put programs.ccache.enable = true; in the node definition
<MichaelRaskin>
Pity.
patrl has quit [Client Quit]
<{^_^}>
[nixops] @phryneas opened pull request #1006 → support ipv6 for scp with "none" backend → https://git.io/fAPoD
<sphalerite>
exarkun1: the node definition is the equivalent of configuration.nix
woodson has quit [Ping timeout: 244 seconds]
<exarkun1>
sphalerite: ah, rad. Should I have been able to find that in the docs somewhere?
<sphalerite>
and you can put all the same options in it
fragamus has joined #nixos
<sphalerite>
exarkun1: https://nixos.org/nixops/manual/#idm140737318599728 " The value of each logical machine attribute is a NixOS configuration module, which describes the desired configuration of the corresponding machine."
<MichaelRaskin>
Baughn: and let me guess, you need CUDA so compton + nouveau + whatever-is-light is not an option?
<Baughn>
Well, I've occasionally used it for fractal rendering.
patrl has quit [Client Quit]
<exarkun1>
sphalerite: makes sense, thanks. What about the `programs.ccache.enable = true` part? is there a document about all of the supported options like that?
tzemanovic has quit [Remote host closed the connection]
<Baughn>
Doesn't really solve the g-sync issue.
petrkr has joined #nixos
<MichaelRaskin>
A better question is whether suspend-vs-hibernate makes any difference.
tzemanovic has joined #nixos
patrl has quit [Client Quit]
<MichaelRaskin>
For the specific trouble with resuming
<Baughn>
Good question. I have no idea how to hibernate it.
tzemanovic has quit [Remote host closed the connection]
<Baughn>
(There are some issues. F'r instance, I'm on a full-disk ZFS system. Heh, maybe later.)
<exarkun1>
sphalerite: thanks again
tzemanovic has joined #nixos
tzemanovic has quit [Remote host closed the connection]
tzemanovic has joined #nixos
periklis` has quit [Ping timeout: 252 seconds]
<MichaelRaskin>
Hm. Resume-from-hibernation is actually quite late in the boot process, so if you configure swap inside a ZFS pool you might have a chance to bring up the pool, look at the swap before mounting anything and resume
tzemanovic has quit [Remote host closed the connection]
<b1000101>
Hi guys, I'm updating a package in nixpkgs - am I supposed to open pull request to other branches than master? For example 18.03, or should I leave it up to whoever merges the request? Thanks! BTW any comments are more than welcome (it's my first pull reuquest to this repo) https://github.com/NixOS/nixpkgs/pull/46644
<MichaelRaskin>
Not sure if the current mainline NixOS initramfs will be happy about that, though.
tzemanovic has joined #nixos
tzemanovic has quit [Remote host closed the connection]
wpcarro has quit [Ping timeout: 252 seconds]
patrl has joined #nixos
martingale has joined #nixos
patrl has quit [Client Quit]
patrl has joined #nixos
trcc has quit [Remote host closed the connection]
patrl has quit [Client Quit]
patrl has joined #nixos
trcc has joined #nixos
iyzsong has quit [Ping timeout: 250 seconds]
fendor has quit [Ping timeout: 246 seconds]
<exarkun1>
hmm, not having any luck getting ccache to actually cache anything though. enabled it and tried nix-build'ing with stdenv overridden to cacheStdenv but no observable side-effects ... eg /var/cache/ccache remains empty
martingale has quit [Ping timeout: 252 seconds]
wpcarro has joined #nixos
baimafeima2 has quit [Quit: Leaving]
trcc has quit [Ping timeout: 272 seconds]
<nwspk>
quick question: I have to pack a pre-compiled library which contains a string of a path to it's database
<nwspk>
replacing this string with the path in the nix-store should fail, because the path in the nix-store is longer
<nwspk>
is there a way to handle this?
<Baughn>
Well, ideally you recompile it.
<Baughn>
Assuming that isn't possible, and patchelf can't do it... you may need to use a chroot to run the whatever.
<nwspk>
well, it's closed software where I don't have the source code for
<nwspk>
I'd be very happy to have the source..
<exarkun1>
shouldn't nix store be append-only? is it a read-only database?
<nwspk>
Baughn: do chroots work with libraries to be called from gcc?
<sphalerite>
`nix copy --to ssh-ng://root@10.123.0.10` is complaining "cannot add path '...' because it lacks a valid signature". How can I pass require-sigs=false in the remote store URL?
<Baughn>
exarkun1: It is append-only, modulo the GC. Why?
<exarkun1>
Relating to nwspk's question.
<sphalerite>
hm, using ssh-ng://root@10.123.0.10?no-require-sigs doesn't complain about a nonexistent option, but also doesn't work
<Baughn>
What of it?
<Taneb>
exarkun1: nwspk is trying to change something before it goes into the nix store
<exarkun1>
Having a database in nix store seems weird.
<Baughn>
nwspk: I can't answer that. I think you'll have a lot of trouble, though.
<symphorien>
nwspk: there is also libredirect you canuse
<nwspk>
exarkun1: sorry, perhaps I described it kind of bad; it's a read-only database which will be queried from the library, I wanna pack
<Baughn>
nwspk: What *is* the original path? If it's somewhere in /var, you can probably keep it.
adfaure has joined #nixos
<nwspk>
and in the library there is a link to /var/.., as Baughn guessed
<Baughn>
And yes, it's ugly. That's the cost of using proprietary software.
<nwspk>
symphorien: thanks, that looks good!
<nwspk>
Baughn: trust me, I don't wanna use this - I have to.
<nwspk>
That's the cost of working .. for the state :^)
<adfaure>
Hello! I find a package with missing dependencies, so I cloned the repo, and add the dep. But now, I would like to test my changes. The problem is that it is a Rpackage, and, I can't only install the package, I need to wrap it. I have a shell.nix that do that, but I would like to tell nix-shell to use my folder of nixpkgs instead of the remote ? Can you help me please ?
<srhb>
adfaure: I think we have something like rWithPackages... Let me check
<Baughn>
nwspk: There are a number of other, possibly less ugly options. lxd, for instance; run it in the distribution it was compiled against.
<MichaelRaskin>
nwspk: yes, you will need to run whatever uses the library in that way.
martingale has quit [Ping timeout: 252 seconds]
rauno has quit [Ping timeout: 244 seconds]
Kelppo has joined #nixos
noam__ has quit [Read error: Connection reset by peer]
Tobba_ has quit [Read error: Connection reset by peer]
noam__ has joined #nixos
Tobba_ has joined #nixos
andersk has quit [Remote host closed the connection]
xenog_ has joined #nixos
davidcl has joined #nixos
goibhniu has quit [Ping timeout: 272 seconds]
qyliss has quit [Ping timeout: 272 seconds]
dejanr has quit [Ping timeout: 272 seconds]
helpusobi_ has joined #nixos
andersk has joined #nixos
goibhniu has joined #nixos
acarrico has quit [Ping timeout: 272 seconds]
deni_ has quit [Ping timeout: 272 seconds]
thblt has quit [Ping timeout: 272 seconds]
xenog has quit [Ping timeout: 272 seconds]
aminechikhaoui has quit [Ping timeout: 272 seconds]
PolarIntersect has quit [Ping timeout: 272 seconds]
emacsomancer has quit [Ping timeout: 272 seconds]
helpusobi has quit [Ping timeout: 272 seconds]
qyliss has joined #nixos
Maxdamantus has quit [Ping timeout: 252 seconds]
dejanr has joined #nixos
PolarIntersect has joined #nixos
aminechikhaoui has joined #nixos
acarrico has joined #nixos
Maxdamantus has joined #nixos
Kelppo has quit []
thblt has joined #nixos
deni_ has joined #nixos
baimafeima2 has joined #nixos
sbdchd has joined #nixos
patrl has quit [Ping timeout: 240 seconds]
<sphalerite>
nwspk: what about making the fixed path that the library expects a symlink to the store path?
<d1rewolf>
guys, i'm trying to install rocket.chat via an overlay. I've downloaded the overlay file and placed it in ~/.config/nixpkgs/overlays. A "nixos-rebuild switch" didn't seem to pick it up. Is there another step to make an overlay work?
exarkun_ has quit [Ping timeout: 252 seconds]
<sphalerite>
d1rewolf: ~/.config/nixpkgs/overlays is only for per-user overlays
<sphalerite>
d1rewolf: for a system overlay, set nixpkgs.overlays in your configuration
<d1rewolf>
sphalerite: wouldn't a per-user overlay work in this case? All I want is my main user (d1rewolf) to have access to the app
<sphalerite>
d1rewolf: e.g. nixpkgs.overlays = [(import ./rocket-overlay.nix)];
exarkun_ has joined #nixos
<nwspk>
sphalerite: just realized while reverseing the lib that for whatever reasons the lib also INSERTS, even if it should just read… but thanks anyway (:
<sphalerite>
d1rewolf: yes you can do that, in that case just put the overlay in ~/.config/nixpkgs/overlays as you have done and install it using nix-env
<d1rewolf>
sphalerite: ok, so is the real diff between per-user and system overlay that you can install the system overlay declaratively? And I assume you can put the nix file anywhere as long as your import path is correct?
Mateon2 has joined #nixos
Mateon1 has quit [Ping timeout: 252 seconds]
Mateon2 is now known as Mateon1
<sphalerite>
d1rewolf: sort of. A system overlay affects the packages used for building the system, e.g. in systemPackages (but also for services and such)
emacsomancer has joined #nixos
<sphalerite>
d1rewolf: while the user overlay affects what the user sees when using <nixpkgs>
<sphalerite>
d1rewolf: and yes, any location will do
<d1rewolf>
sphalerite: great, thank you. Does the nixpkgs manual cover this? I didn't find that necessarily in the manual?
<{^_^}>
[nixpkgs] @brainrape opened pull request #46649 → Increase Virtualbox disk image size → https://git.io/fAPDJ
<sphalerite>
d1rewolf: would be useful for the nixos manual to mention it too…
<srhb>
b1000101: Thanks for this!
patrl has joined #nixos
phreedom_ has quit [Ping timeout: 256 seconds]
<srhb>
b1000101: In general, you should PR against master first, and request backports -- in this case to 18.03 and 18.09, and perhaps create those PRs as well (otherwise a maintainer will likely do the cherry picks for you) :)
Ariakenom has quit [Quit: Leaving]
<srhb>
b1000101: You will also need to add yourself to maintainers/maintainer-list.nix in order for evaluation to succeed.
Ariakenom has joined #nixos
<d1rewolf>
guys, I submitted this PR to node2nix (https://github.com/jbwiv/node2nix/pull/1). Not getting a response from the author. This but affects me quite often. Is that git repo the appropriate place to submit a PR, or should I submit it to the nixpkgs repo?
<srhb>
b1000101: (It's actually an attrset of maintainer contact information :))
<b1000101>
srhb happy to help, I actually work for SatoshiLabs and more guys are getting interested in NixOS in general so I thought it would be nice to keep the package up-to-date as we release new versions. I've created a PR to 18.03 already, 18.09 on the way :).
<srhb>
b1000101: That's great to hear! :)
<sphalerite>
d1rewolf: you seem to have opened it against your own repo, not the upstream one
<srhb>
b1000101: I don't have time to properly review until late tonight, but it looks great aside from the comments I already gave you :)
* d1rewolf
facepams
<b1000101>
srhb maintainers = with stdenv.lib.maintainers; [ canndrew jb55 1000101]; does this suffice? I've actually included it
<d1rewolf>
sheesh. My bad
<infinisil>
Haha
<srhb>
b1000101: Look at the file in nixpkgs/maintainers/maintainer-list.nix
<b1000101>
srhb don't know the frist 2 guys though :)
* samueldr
will drop the comment he's writing
<b1000101>
srhb thank's, I'm on it
<samueldr>
everything's been said here :)
<b1000101>
srhb *thanks
<srhb>
b1000101: You might have some fun with your name. You probably need to double quote it.
<srhb>
Because...
<srhb>
> { 1000101 = "42"; }
<{^_^}>
error: syntax error, unexpected INT, at (string):193:3
<hyper_ch2_>
am I the only one who wonders what SatoshiLabs is?
<srhb>
hyper_ch2_: I googled it. :P
<b1000101>
hyper_ch2_ company behind Trezor wallet
phreedom has joined #nixos
<hyper_ch2_>
b1000101: well, that would lead to my next question: what is trezor wallet :) but I googled it meanwhile
<b1000101>
srhb it should be 69, 101010 is 42 but that was already taken :P
<srhb>
Having companies taking active interest in keeping their packages up-to-date makes me happy.
<srhb>
b1000101: Right, the problem is that attribute keys can't be all-numerics :P
<b1000101>
srhb Having NixOS keeps me happy
<b1000101>
srhb I get it, I'll put b in front of it - that should sort it out
<sphalerite>
> {"1000101" = {};}.1000101
<srhb>
But I think you'll be OK if you double quote it.
<{^_^}>
attempt to call something which is not a function but a set, at (string):193:1
<srhb>
> Both places.
<{^_^}>
error: syntax error, unexpected ')', expecting ID or OR_KW or DOLLAR_CURLY or '"', at (string):194:1
<srhb>
oops
<sphalerite>
> {"1000101" = {};}."1000101"
<{^_^}>
{ }
<srhb>
I guess we'll see if CI breaks. ;-)
<b1000101>
srhb so, shoud I even try the double qutoes? :)
<srhb>
b1000101: Yeah, give it a go.
xenog_ has quit [Remote host closed the connection]
<b1000101>
srhb thanks a million
<srhb>
b1000101: "Breaks ci" => Your PR will be red. We should be okay across the rest of the board ;)
<sphalerite>
b1000101: you'll also need to put stdenv.lib.maintainers."1000101" in the list rather than just "1000101" since it'll just evaluate to the string otherwise
patrl has quit [Ping timeout: 252 seconds]
<srhb>
Oh right.
Dedalo has joined #nixos
<srhb>
sphalerite: Thanks :)
patrl has joined #nixos
<b1000101>
sphalerite you mean like this? Sorry for asking such questions, I'm only starting with NixOS but it's fantastic so far. maintainers = with stdenv.lib.maintainers; [ canndrew jb55 stdenv.lib.maintainers."1000101"];
<Myrl-saki>
Holy crap.
<Myrl-saki>
Why is GHC 1.3 GB?
<srhb>
b1000101: Yep, that should be fine.
<Myrl-saki>
That's literally 20% of my /run/current-system.
<sphalerite>
b1000101: yes. It's not beautiful but it ought to work. It might be nicer to put `with stdenv.lib;` in front of the whole meta section, then you can replace `with stdenv.lib.maintainters;` with `with maintainers;` and `stdenv.lib.maintainers."1000101"` with `maintainers."1000101"`.
<b1000101>
sphalerite or just change my silly github name and put at least b in front of the numbers :)
<sphalerite>
b1000101: that is also an option :)
<sphalerite>
kind of boring though
<sphalerite>
you should leave it as is and set your bio to "discoverer of edge cases". :D
<b1000101>
sphalerite yeah...let's fight
<Myrl-saki>
Err
<Myrl-saki>
How slow does lzip compress?
<Myrl-saki>
tar -c n5i1zdpmk2b1s3z96649xh8f9kr3g96s-ghc-8.4.3 | nix run nixpkgs.lzip -c lzip | nix run nixpkgs.pv -c pv > /dev/null
adfaure has quit [Quit: Page closed]
<Myrl-saki>
Getting 300 KB/s at most there.
<b1000101>
sphalerite :D my whole life is just that, so I'll leave it, let's have some fun
<sphalerite>
Myrl-saki: you might want to move the pv one step earlier in the pipeline, since it's not usually the compressed data rate you're interested in
<Myrl-saki>
sphalerite: In this case, it is because I want to know the resulting file size if I compress it.
knupfer has quit [Ping timeout: 252 seconds]
<Myrl-saki>
211MiB 0:00:42 [5.00MiB/s] [ <=> ]
<Myrl-saki>
Apparently, GHC compressed is 211 MiB.
<Myrl-saki>
I think dd would have done the job as well?
<sphalerite>
Myrl-saki: tar | pv | lzip | wc -c ?
<Myrl-saki>
sphalerite: Oh true. Idk why pv was the first thing in my mind.
knupfer has joined #nixos
sbdchd has quit [Remote host closed the connection]
humanoyd has quit [Quit: WeeChat 2.2]
<Myrl-saki>
sphalerite: Anyways, yeah, you're right, that was a huge oversight lol. lzip's being fed at 3 MiB/s, much higher than 300 KB/s. :P
<b1000101>
btw won't my name cause problems in maintainer-list.nix too? :D
<Myrl-saki>
Okay, um, I need an explanation for this.
<Myrl-saki>
[myrl@myrl:/nix/store]$ tar -c n5i1zdpmk2b1s3z96649xh8f9kr3g96s-ghc-8.4.3 | nix run nixpkgs.pv -c pv > /dev/null
<Myrl-saki>
1.27GiB 0:00:00 [2.58GiB/s] [ <=> ]
<vdemeester>
o/
<vdemeester>
small question/problem : anybody running into `value is a set while a list was expected, at /nix/store/xjxc1jzl97cpv9z9rc30w9zcz05vdlnb-nixos-18.03.133224.5f59ab7d4e0/nixos/pkgs/applications/video/mpv/default.nix:187:32` very recently ?
<Myrl-saki>
Okay, I think that's just telling me that "I've been tar -cing that so much, it's already inside the RAM."
<nwspk>
What is the most simple way to create some directories once by a NixOS-module?
<sphalerite>
b1000101: I'd suggest using git locally to make your changes so that you can test them rather than using the github UI (and its awful default commit message)
<nwspk>
I thought about a systemd oneshot-module with a simple if-Guard to check, if the files already exists, but thought that this might be kind of an overkill
<sphalerite>
b1000101: additionally, any commits to the release branches should be cherry-picked from the corresponding master commit
sbdchd has joined #nixos
<sphalerite>
nwspk: I believe that's the usual approach. You don't need the guard if you just use mkdir -p though :)
<symphorien>
nwspk: systemd-tmpfiles ?
<nwspk>
sphalerite: some magic files must be copied there, so a guard seems necessary for me - thanks, than it's the Poettering way
<b1000101>
yeap, it was just more convenient at the moment but you're absolutely right, I've tested it locally but was too lazy to hit the command in shell
<b1000101>
sphalerite yeap, it was just more convenient at the moment but you're absolutely right, I've tested it locally but was too lazy to hit the command in shell
patrl has quit [Client Quit]
<nwspk>
symphorien: I need a database in a hardcoded path
<sphalerite>
b1000101: additionally, it would make sense to put the maintainer addition and the trezor update in separate commits but in the same PR
<nwspk>
clever: a proprietary shared library which does magic things there.. it's stuff I have to do for work
<sphalerite>
Taneb: how are you doing this?
patrl has quit [Client Quit]
<nwspk>
hyper_ch2_: that's kind of what I was hoping for, thanks!
<clever>
nwspk: is that library being ran in a service?
<nwspk>
clever: no, but it needs some kind of magic files which may be changed in a hard coded path
<Taneb>
Figured out what I was doing wrong... I had "nixpkgs {overlays = import ./overlays.nix; }", overlays.nix being "[ import overlays/foo.nix ]" when it should have been "[ (import overlays/foo.nix) ]"
<sphalerite>
hyper_ch2_: please don't recommend adding activation scripts. nwspk: modifying activation scripts is a risky business because they're run very early on and anything going wrong can cause system startup to fail
rouma has quit [Read error: Connection reset by peer]
Kelppo has quit []
lawlesseel has quit [Remote host closed the connection]
<nwspk>
sphalerite: well.. then back to systemd service, thanks
lawlesseel has joined #nixos
<clever>
nwspk: id just use a one-shot systemd service then, yeah
<nwspk>
clever: that's what I thought about, thanks
hyper_ch2_ has quit [Quit: Page closed]
Dedalo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<b1000101>
sphalerite thanks, makes sense, should I close those PR and open new ones? Thank you so much for your help.
<sphalerite>
b1000101: I'd suggest going to the first PR, the trezord update one, and 1. Changing its target branch to master; 2. Rebasing it on master, and in the process altering the commits to have one adding yourself to maintainers-list.nix and one to update trezord and tone to update he udev rules, with appropriate commit messages
frank87 has quit [Quit: No Ping reply in 180 seconds.]
<sphalerite>
d1rewolf: yep as Myrl-saki said I don't see any reason why that would happen. How and where are you importing it? That might be relevant
frank87 has joined #nixos
abueide has quit [Remote host closed the connection]
abueide has joined #nixos
baimafeima2 has quit [Quit: Leaving]
<HoloIRCUser3>
How robust is Dynamic Disnix these days? The idea intrigues me, but even with 2 year old src, there are copious warnings of it being experimental
<Myrl-saki>
Wait, what is Disnix and how does it differ from Nixops?
vaibhavsagar_ has quit [Quit: Leaving]
<HoloIRCUser3>
NixOps deals with intra-dependencies (between packages on the same machine)
<HoloIRCUser3>
Disnix is the same, but for service-oriented architectures
<sphalerite>
Yeah disnix sounds really cool but I feel like it's not very well advertised and stuff, and I don't know of anyone using it besides __Sander__
ChengCat has quit [Quit: Leaving]
graphene has quit [Read error: Connection reset by peer]
graphene has joined #nixos
lord| has quit [Ping timeout: 245 seconds]
<Myrl-saki>
If I understand correctly, based on what I've read and what HoloIRCUser3 is saying, is that it's declaring what services to have instead of what machines should have what services?
<Myrl-saki>
So in a way, inversion of what Nixops does?
<{^_^}>
[nixops] @rbvermaa pushed 0 commits to refs/tags/v1.6.1: https://git.io/fAPFH
<Myrl-saki>
"Provisioning" of services, instead of provisioning of machines?
halfbit has joined #nixos
lord| has joined #nixos
<HoloIRCUser3>
Both really
<HoloIRCUser3>
There's a link to disnix on the Nixon site
<HoloIRCUser3>
From there, take a look at the extensions to read about dynamic disnix
<Myrl-saki>
I'm reading the HTML docs, FWIW.
borisbabic has joined #nixos
oscarvarto has joined #nixos
lawlesseel has quit [Remote host closed the connection]
jperras has quit [Ping timeout: 240 seconds]
lawlesseel has joined #nixos
jperras_ has joined #nixos
<Myrl-saki>
This might be a better question, can Disnix fully replace Nixops?
<Myrl-saki>
I'm not saying "currently,"
nuncanada has joined #nixos
<Myrl-saki>
but rather, does its (possible) scope completely overlap with Nixops
<d1rewolf>
Myrl-saki: sphalerite: I'm importing it in configuration.nix as nixpkgs.overlays = [(import ../apps/rocketchat.nix)];
graphene has quit [Remote host closed the connection]
nuncanada has quit [Client Quit]
<sphalerite>
d1rewolf: that won't work since it's not an overlay
<Myrl-saki>
sphalerite: s/overlay/module? :P
<Myrl-saki>
Either way, he posted the wrong file.
<sphalerite>
d1rewolf: instead you'd want to put (pkgs.callPackage ../apps/rocketchat.nix) in your systemPackages
<Myrl-saki>
You gave up rocketchat/default.nix
<Myrl-saki>
s/up/us/
graphene has joined #nixos
<sphalerite>
Myrl-saki: no, overlay. nixpkgs.overlays should contain overlays :)
<Myrl-saki>
sphalerite: Oh crap. I'm uh, blind.
<Myrl-saki>
sphalerite: Shhh
abueide has quit [Ping timeout: 250 seconds]
<borisbabic>
Is there a way to drop into a shell during a nix-build?
<Myrl-saki>
sphalerite: I saw [(import ...)]; and my brain heurestically told me it's imports = ; ~ ;
<sphalerite>
borisbabic: yes, but it's all kinds of hacky-nasty
<sphalerite>
borisbabic: Mic92 is doing some research that makes it practical
<sphalerite>
borisbabic: but why do you want to do this? To debug a build? I'd suggest trying to do the build in a nix-shell instead
<d1rewolf>
sphalerite: hmmm... I had inquired via github about the status of rocket chat, and received the following response: @jbwiv you could use my nix expression as an overlay and start running Rocketchat. I'm getting there's still 4 weeks until I have time to look into this PR, there's to say if someone doesn't beat me to it.
<Myrl-saki>
sphalerite: borisbabic: Possible use-case of this?
<sphalerite>
d1rewolf: I suppose the person meant "… use my nix expression *in* an overlay…"
<borisbabic>
sphalerite: If there are instructions somewhere I'm fine with hacky-nasty
<sphalerite>
d1rewolf: but you don't need an overlay for it really, you can just callPackage it in your systemPackages, much neater :)
<d1rewolf>
sphalerite: cool. thanks very much. Myrl-saki: out of curiosity (i try to collect obscure nix commands in a file for notes), what does nix-instantiate -E '(import <nixpkgs> {}).callPackage ./. {}' do?
<Myrl-saki>
d1rewolf: That provides a .drv.
<borisbabic>
Myrl-saki: debugging. I'm having issues with a certain phase of a build and I'd like to be able drop into a shell at that stage for faster iteration while debugging
jperras_ has quit [Ping timeout: 240 seconds]
<Myrl-saki>
d1rewolf: I'm not sure how to explain it, sorry.
<d1rewolf>
Myrl-saki: it's ok. thanks
kiloreux has quit [Ping timeout: 252 seconds]
knupfer has quit [Ping timeout: 240 seconds]
<d1rewolf>
sphalerite: when you say put that "in" your systemPackages, what exactly do you mean? Put it directly in the list of pkgs? like "environment.systemPackages = with pkgs; [ (callPackage ../apps/rocketchat.nix) ]" ?
<sphalerite>
borisbabic: I don't really have any very specific instructions, but basically find the PID of a process that's in the build, then use nsenter -m -u -i -n -p -t <PID> with an appropriate shell to enter its namespaces. Then you need to grab its environment
<sphalerite>
d1rewolf: yes, except pkgs.callPackage and you need a {} after rocketchat.nix
__Sander__ has quit [Quit: Konversation terminated!]
<sphalerite>
borisbabic: it will be pretty fiddly though. I'd really recommend using nix-shell instead. You can run nix-shell <same args as you'd pass to nix-build>, then cd to a directory of your choice where you want the build to happen, export out=/some/path/to/install/to, and run genericBuild to have it run all the same stuff as it would in the real build
erasmas has joined #nixos
<d1rewolf>
sphalerite: that worked. thanks. what does the {} do in that case?
exarkun_ has quit [Ping timeout: 272 seconds]
<sphalerite>
d1rewolf: it's the overrides you want passed to the derivation — in this case none. But if you wanted to replace, say, the dpkg that gets paassed into the rocketchat expression, you could put { dpkg = something_else;} in there
xok has joined #nixos
<sphalerite>
d1rewolf: as for the command that Myrl-saki wrote, nix-instantiate instantiates derivations — that is, evaluates the nix expressions defining them, and produces a .drv file which represents the final build specification
<borisbabic>
sphaletire: Then i'll try to make it work with nix-shell. Thanks for your help!
<sphalerite>
d1rewolf: the expression passed in — '(import <nixpkgs> {}).callPackage ./. {}' — means import nixpkgs, and "call" the package in the cwd. That is, import default.nix, which should contain a package function (like the rocketchat one you have there), and fill in all the args with values from nixpkgs ("compose" it)
<sphalerite>
d1rewolf: basically the same thing as what you've put into systemPackages, except that nixpkgs has already been imported and you're just using the existing one rather than importing a new one
<sphalerite>
the nix pills are really useful for getting an in-depth understanding of this
<{^_^}>
[nixpkgs] @lopsided98 opened pull request #46653 → linux_rpi: add support for Raspberry Pi 3 → https://git.io/fAPAw
xok has quit [Ping timeout: 252 seconds]
<d1rewolf>
sphalerite: yeah, I've gone through the pills. I probably need to do them again. It's one thing to fix errors in a sandbox and another to know what to use and when
<sphalerite>
:)
<sphalerite>
NickHu_: your website's TLS cert has expired btw
<HoloIRCUser3>
Myrl-saki: it piggybacks on top of NixOps to my understanding
<{^_^}>
[nixpkgs] @Mic92 pushed commit from @brocking to release-18.09 « scaleway-cli: 1.14 -> 1.17 »: https://git.io/fAPx1
<{^_^}>
[nixpkgs] @geistesk opened pull request #46654 → Walabot SDK and module → https://git.io/fAPx7
<HoloIRCUser3>
From the Disnix extensions page: With DisnixOS you can first deploy a network of machines with NixOps and deploy services inside them with Disnix.
borisbabic has quit [Ping timeout: 252 seconds]
IRCsum has joined #nixos
graphene has quit [Remote host closed the connection]
graphene has joined #nixos
winem_ has quit [Ping timeout: 252 seconds]
civodul has quit [Quit: ERC (IRC client for Emacs 26.1)]
graphene has quit [Remote host closed the connection]
mpu has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @WilliButz opened pull request #46655 → init CodiMD at 1.2.0, add module & test → https://git.io/fAXeu
sanscoeur has joined #nixos
graphene has joined #nixos
sigmundv__ has quit [Ping timeout: 240 seconds]
fendor has joined #nixos
humanoyd has joined #nixos
<{^_^}>
[nixpkgs] @WilliButz closed pull request #32275 → dillinger: init at 2017-10-08 with module and minimal test → https://git.io/vbO3z
worldofpeace has joined #nixos
xok1 has quit [Ping timeout: 272 seconds]
graphene has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @1000101 opened pull request #46656 → trezord: 2.0.14 -> 2.0.19 and nixos/trezord: revised and updated udev rules → https://git.io/fAXvQ
<{^_^}>
[nixpkgs] @1000101 opened pull request #46657 → trezord: 2.0.14 -> 2.0.19 and nixos/trezord: revised and updated udev rules → https://git.io/fAXv5
<{^_^}>
[nixpkgs] @1000101 opened pull request #46658 → trezord: 2.0.12 -> 2.0.19 and nixos/trezord: revised and updated udev rules → https://git.io/fAXvF
graphene has joined #nixos
Avaq has quit [Ping timeout: 252 seconds]
<b1000101>
sphalerite ^^^ I hope it's better this time
kenshinC1 has joined #nixos
fresheyeball has quit [Remote host closed the connection]
<sphalerite>
b1000101: yes this looks better, although the release branches don't have the references to the master commits — you should backport them with git cherry-pick -x. You can also just leave that up to whoever merges your PR though
Tucky has quit [Remote host closed the connection]
<b1000101>
sphalerite dang! I think I'll leave it as it is and use git cherry-pick -x next time. Thanks a lot for your help, kudos to you and have a wonderful day!
<sphalerite>
b1000101: thank you :)
<kalbasit[m]>
how to cleanup systemd-boot entries? I have 250 entries right now O.o
<exarkun1>
with nixops, can I create an EBS volume and then attach it as a block device to an EC2 instance? I see that I can create an EBS volume, apparently, with `deployment.ebsVolumes`. And I see that volumes can be attached via `deployment.ec2.blockDeviceMapping`. But I don't see how to hook the two together.
sbdchd has quit [Ping timeout: 246 seconds]
<{^_^}>
[nixpkgs] @xeji merged pull request #46635 → nbd: add which to buildInputs → https://git.io/fAixA
<{^_^}>
[nixpkgs] @xeji pushed commit from @qolii to master « nbd: add which to buildInputs (#46635) »: https://git.io/fAXTC
mayhewluke has quit [Ping timeout: 252 seconds]
HoloIRCUser3 has quit [Quit: HoloIRCUser3]
<clever>
exarkun1: i'm guessing blockDeviceMapping creates an xvdb and such, then you use normal fileSystems."/foo" to mount it
mayhewluke has joined #nixos
<clever>
exarkun1: fileSystems also supports auto-format
rauno has joined #nixos
dbmikus_ has joined #nixos
sbdchd has joined #nixos
sbdchd has quit [Remote host closed the connection]
sbdchd has joined #nixos
IRCsum has quit [Ping timeout: 272 seconds]
<exarkun1>
that would let me go from a block device to a filesystem but I'm still not sure how to do from an EBS volume to a block device, unless I hand-create the EBS volume and hard-code the volume id in the blockDeviceMapping expression.
graphene has quit [Remote host closed the connection]
<exarkun1>
I imagine something like `deployment.ebsVolumes.foo = { ... }` generating a magic identifier somewhere that can be pulled off and given to blockDeviceMapping, such that the whole thing notices it needs to create a new resource, does so, and then uses that for the block device...
<exarkun1>
clever: I guess I'm not explaining my problem very well.
chaker has joined #nixos
trebuh has quit [Remote host closed the connection]
<chaker>
Hey guys, What can be the reason that two derivation differ only by one inputSrc?
<clever>
chaker: how do the values differ? can you paste them?
<chaker>
I'm trying to reproduce the evalution (not the actuall build) locally and the only diff that I can see is one input in inputSrc
trebuh has joined #nixos
<clever>
chaker: when doing src = ./.; there can be differences like .git and other files
<clever>
lib.cleanSource can help with that
Ericson2314 has joined #nixos
<chaker>
clever: I'm actually passing the output of builtins.FetchMercurial to src, if that's make a difference
<clever>
chaker: ah, that should omit .hg then, are you able to download both versions of inputSrc (run nix-store -r on one) and then just `diff -ru` the 2?
<clever>
chaker: it looks like one side has a source and the other doesnt
<chaker>
Yep :/
<clever>
is src= the same on both of them?, can you paste that nix code?
xok has quit [Ping timeout: 245 seconds]
<{^_^}>
[nixpkgs] @mbode opened pull request #46664 → gitAndTools.grv: fix darwin build by using go 1.9 → https://git.io/fAXL1
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
knupfer has joined #nixos
kai_w has joined #nixos
sbdchd has quit [Ping timeout: 244 seconds]
<chaker>
Yeah, I'm sure it's the same in both cases. "src = repo". Where repo is an argument. That I configure for hg checkout in Hydra and builtins.fetchMercurial in local eval.
<chaker>
I'm supposing that hydra is doing the same as builtins.fetchMercuial :/
<{^_^}>
[nix] @ElvishJerricco opened pull request #2423 → nix build: Print result paths to stdout with --no-link → https://git.io/fAXtL
kenshinC1 has quit [Ping timeout: 240 seconds]
<clever>
chaker: run `nix show-derivation foo.drv` on both drvs, and compare what src is set to
<{^_^}>
[nixpkgs] @Assassinkin opened pull request #46665 → pythonPackages.fuse: 0.2.1 -> 0.3.1 | refactor | python3 support → https://git.io/fAXtR
<zimbatm>
LnL: not really but I am about to on the next project
<zimbatm>
isn't the vault UI packaged with vault nowadays?
abueide has joined #nixos
sbdchd has joined #nixos
<LnL>
not by default, you need to run the release build
<zimbatm>
> Vault UI is not available in this binary
<{^_^}>
error: syntax error, unexpected IN, expecting ')', at (string):194:27
<LnL>
yeah
<chaker>
clever: The issue was that I passed the result of builtins.fetchMercurial in the local eval instead of passing the function it self. Using the first now produced identicals derivations!
<chaker>
Thanks for helping
<zimbatm>
okay, I would consider this to be a bug with the packaging
<chaker>
s/first/second
sbdchd has quit [Ping timeout: 246 seconds]
orivej has joined #nixos
_ris has joined #nixos
<LnL>
zimbatm: I think I was pretty close, but `npm rebuild node-sass` doesn't work because it needs network access and node_modules isn't writable
<zimbatm>
LnL: yarn2nix isn't usable in nixpkgs
<zimbatm>
because it's using IFD :/
<LnL>
even with the yarn.nix there?
<zimbatm>
it could be fixed but it also loads the name and version from the package.json
<zimbatm>
is your prototype in a branch somewhere?
<LnL>
nix-build -A vault --option allow-import-from-derivation false works fine
cloudcell28 has quit [Killed (Sigyn (Spam is off topic on freenode.))]
sigmundv has joined #nixos
sbdchd has quit [Remote host closed the connection]
sbdchd_ has joined #nixos
mounty has joined #nixos
jtojnar has quit [Ping timeout: 252 seconds]
sbdchd has joined #nixos
sbdchd__ has joined #nixos
tzemanovic has quit [Ping timeout: 252 seconds]
spear2 has joined #nixos
sbdchd_ has quit [Ping timeout: 272 seconds]
sbdchd_ has joined #nixos
sigmundv__ has joined #nixos
<d1rewolf>
do most of you use automatic updates, or do you update manually?
nicomachus has joined #nixos
<nicomachus>
Alⅼah iѕ ⅾoiᥒg
<nicomachus>
s∪n іs nⲟt dοinɡ Αⅼⅼɑh is dοіng
nicomachus has quit [Killed (Sigyn (Spam is off topic on freenode.))]
<samueldr>
manually
sbdchd has quit [Ping timeout: 240 seconds]
georges-duperon has joined #nixos
<samueldr>
I even "pin" channel versions by pushing the head of the branches to a private repo, so I can rebuild another day on another machine and have the same known working state
sigmundv has quit [Ping timeout: 252 seconds]
sbdchd__ has quit [Ping timeout: 252 seconds]
sbdchd has joined #nixos
sbdchd_ has quit [Ping timeout: 244 seconds]
Fare has joined #nixos
jtojnar has joined #nixos
xok has joined #nixos
<{^_^}>
[docker] @domenkozar pushed to nix-2.1 « Upgrade to Nix 2.1.1 »: https://git.io/fAXcL
<{^_^}>
[docker] @domenkozar opened pull request #2 → Upgrade to Nix 2.1.1 → https://git.io/fAXcY
_0bitcount has joined #nixos
<kreisys>
... so I deleted an S3 bucket that I've been using as a cache
<kreisys>
and now every time I try to do `nix-channel --update` I get: error: AWS error fetching 'bk9zkj3xldm75plyy2ixnjqhz009pqm7.narinfo': The specified bucket does not exist
sigmundv__ has quit [Ping timeout: 252 seconds]
<kreisys>
I removed the s3:// url from /etc/nix/nix.conf. what am I missing? :|
<kai_w>
clever: I think I need to use nixpkgs.crossSystem.system, as I am deploying from x86 - does that sound right? It's currently building a whole bunch of packages.
<clever>
kai_w: yeah
mr_trousers has joined #nixos
martingale has joined #nixos
simukis has quit [Ping timeout: 240 seconds]
b1000101 has quit [Ping timeout: 252 seconds]
<kai_w>
cool. I thought most packages should be on hydra, so I'm supprised its trying to build glibc. but if it runs at least that'll be an improvement :)
realrokka has joined #nixos
patrl has quit [Quit: WeeChat 2.2]
graphene has quit [Remote host closed the connection]
martingale has quit [Ping timeout: 272 seconds]
patrl has joined #nixos
patrl has quit [Client Quit]
<d1rewolf>
all, I have a disk space problem and am trying to remove some older packages. For example, I have 4 versions of android-studio in the store. nix-store -q --roots `which android-studio ` gives me https://pastebin.com/0HJi6Wuy
<d1rewolf>
I can't recall, though, how to view which generations are causing the older versions to stay around, nor can I recall how to then remove those older generations.
graphene has joined #nixos
patrl has joined #nixos
<d1rewolf>
can someone trigger my memory?
<MichaelRaskin>
You can remove the generation symlinks in /nix/var/nix/profiles or run nix-collect-garbage -d to delete old generations
<d1rewolf>
MichaelRaskin: how do I know which generation, say, system-15-link is tied to?
fendor has quit [Ping timeout: 245 seconds]
<d1rewolf>
I tried nix-env --delete-generations 1 2 3 4 5
<d1rewolf>
but it didn't seem to remove them. --list-generations still shows them there
<MichaelRaskin>
You can use readlink to see the corresponding store path; or nix-store -qR system-15-link (in the corresponding directory)
<MichaelRaskin>
System generations and default profile generations are different generations, too
<d1rewolf>
nix-env --delete-generations 30d doesn't seem to work either
<MichaelRaskin>
You need -p to say _which_ profile you want to prune
<d1rewolf>
MichaelRaskin: k, thank you
<d1rewolf>
MichaelRaskin: adding -p /nix/var/nix/profiles/system fixed it
jtojnar has joined #nixos
mr_trousers has quit [Ping timeout: 252 seconds]
<clever>
d1rewolf: you can also use `nix-collect-garbage --delete-older-than`
<clever>
which prunes all profiles
<elvishjerricco>
clever: * All profiles you have permissions over
<d1rewolf>
clever: perfect, thx
<clever>
yeah
xok has quit [Quit: Leaving.]
<{^_^}>
[nixpkgs] @costrouc opened pull request #46671 → pythonPackages.jupyter-repo2docker: init at 0.6.0 → https://git.io/fAXWj
<eisbaer-north>
hello, i'm trying to build a coreboot image and used this (https://nixos.wiki/wiki/Coreboot) from the wiki but i can't build the compiler because a gnat compiler is missing. i searched the packages but i couldn't find one in the packages is there a flag i need to set in gcc ?
<grp>
clever: I've defined a couple of packages that create $out/lib/$language/$libname/ but when ~/.nix-profile/lib is updated, packages seem to conflict at ~/.nix-profile/lib/$language/, they compete with each other and at any one point I only have one sub-item there. Is there any way to tell nix to merge at the next folder level?
Supersonic has quit [Disconnected by services]
Supersonic112 has joined #nixos
Supersonic112 is now known as Supersonic
<clever>
grp: are they all installed with nix-env?
<grp>
yes
<clever>
by default, it wont let you install conflicting files like that
<grp>
by default... so there is a way?
<clever>
and when you force it to ignore the conflict, it does what you dont want
<grp>
but can it be specified in the nix expression?
<grp>
or only for nix-env invocation?
humanoyd has joined #nixos
<clever>
it should merge those subdirs together, and only conflict on the last file
<grp>
just in case I'll re-state the problem: currently the subdirs are not merged. There's no conflict warning either but they seem to be mutually exclusive, so I only get the content of one lib at a time
<grp>
if it were to merge them, there'd be no conflict with files
<{^_^}>
[nixpkgs] @LnL7 pushed commit from @mbode to release-18.09 « gitAndTools.grv: fix darwin build by using go 1.9 »: https://git.io/fAXEu
graphene has quit [Remote host closed the connection]
graphene has joined #nixos
haslersn has joined #nixos
<haslersn>
Hi, what should I do if i think there's a mistake in the wiki? In particular, I don't think that https://nixos.wiki/wiki/Wine still holds since the wine default.nix has: wineBuild ? if stdenv.hostPlatform.system == "x86_64-linux" then "wineWow" else "wine32"
kisik21 has quit [Ping timeout: 240 seconds]
<jasongrossman>
haslersn: It's a wiki, so you should edit it! If you're not positive then still edit it, just say you're not positive.
<jasongrossman>
haslersn: And thank you for doing this. :-)
<haslersn>
Ok
patrl has quit [Quit: WeeChat 2.2]
<catern>
how big in TB is the full Nix package set?
<catern>
Approximately - everything built on Hydra
HauM127 has joined #nixos
<HauM127>
Aⅼⅼаh is ԁoiᥒɡ
HauM127 has quit [Killed (Sigyn (Spam is off topic on freenode.))]
fragamus has joined #nixos
<grp>
clever: found the problem
<grp>
nix does not merge folders called "log"
<clever>
that is very very weird
<grp>
I renamed the package to "syslog" instead of "log" and now it merges fine
<grp>
I was banging my head against the wall for a whole evening... to think it was something like that
<jasongrossman>
The ZFS people have just announced a serious bug in their latest release. Can someone please remind me how to quickly look up which version of ZFS is planned for 18.09?
NightTrain has joined #nixos
jperras has quit [Ping timeout: 244 seconds]
<clever>
jasongrossman: what are the details of the bug? i'm running nixos-unstable with zfs
<colemickens>
wait... I built a bunch of packages on my laptop and then added a couple commits and did --build-host=beefy-server.
oldandwise has joined #nixos
oldandwise has quit [Client Quit]
<colemickens>
IT looks like it sent derivations for ALL packages to the remote build server and is building all of them, instead of the ones that are missing.
<colemickens>
Am I doing something wrong? I even checked my nixpkgs to make sure I was building and considering the right revisions...
oldandwise has joined #nixos
petrkr has quit [Ping timeout: 264 seconds]
kyren has joined #nixos
<catern>
tilpner: no, it was made by a core Nix committer and it's fairly low-level
<rain1>
I think I did something wrong. I want to try making my own nix package definition so I followed this https://nixos.wiki/wiki/Nixpkgs/Create_and_debug_packages I cloned the repo and am switching to it... but now it's building some C++ stuff from source
sir_guy_carleton has joined #nixos
georges-duperon has quit [Ping timeout: 252 seconds]
kiloreux has joined #nixos
<rain1>
nixos-rebuild -I nixpkgs=$NIXPKGS switch this command
erasmas has quit [Quit: leaving]
patrl has joined #nixos
<rain1>
is there a way I can write and test my own package without.. it seems to be compiling libc++ and icu from source
<catern>
tilpner: and it's not on google from my much googling :)
brejoc has quit [Ping timeout: 245 seconds]
haslersn has quit [Ping timeout: 252 seconds]
Growpotkin has joined #nixos
oldandwise has quit [Client Quit]
<Growpotkin>
eyyy
oldandwise has joined #nixos
<Growpotkin>
Has anybody has issues with their laptop lids acting up?
<Growpotkin>
When I close and reopen I get a black screen, and certain services fail to restart. BT and wifi
<colemickens>
@growpotkin yes! on nixos-unstable
<colemickens>
@growpotkin it's probably not going to sleep. I get a message about a busy device. I meant to file an Issue.
<clever>
growpotkin: sounds like its going into standby automatically
oldandwise has quit [Client Quit]
<colemickens>
Check journalctl and look for lid events and then errors around there.
oldandwise has joined #nixos
<Growpotkin>
I tried going down a rabbit hole of acpid scripts that some dudes has posted a while ago, but no luck there.
<Growpotkin>
I will check journalctl
<colemickens>
Sep 07 12:51:05 xeep systemd-sleep[9018]: Failed to write /sys/power/state: Device or resource busy
<colemickens>
^ That's what I've been getting a lot lately, along with lid problems, overheating in my bag, etc.
<catern>
tilpner: no
<tilpner>
well, I tried
<colemickens>
clever, have you ever seen something like that? I couldn't find hardly *anything* when I googled it.
<catern>
it was a kubernetes controller or something
haslersn has joined #nixos
<colemickens>
what does "nix deps alongside pods" even mean?
<Growpotkin>
whats strange is mine is definitely going to sleep. It never feels hot or seems like it's running. It just doesn't want to wake up unless I force it to by holding my power button. And even then the wifi and BT just don't wanna play. I have to restart
<catern>
it was still in development when the person mentioned it here; you attached labels to pods which described a Nix store path (or maybe a package?) and it would deploy those store paths to the hosts when the pod was deployed there
<colemickens>
that makes me want to cry, but interesting.
<catern>
it should make you want to cry from joy
<colemickens>
I am the "pods should never mount host paths" kinda guy :)
<catern>
because this allows you to run without Docker/mount namespacing
<colemickens>
I mean, it's a huge security hole, but I guess it makes ... it easier for random pods to poop into your host
<colemickens>
then why bother with kubernetes at all, just use NixOps?
<catern>
I just want to find this thing, not debate it
<colemickens>
fair enough
oldandwise has quit [Client Quit]
<colemickens>
actually, I think gchristensen might have tweeted about that? I might be making that up though.
oldandwise has joined #nixos
<sphalerite>
colemickens: what's wrong with a read-only binding?
<sphalerite>
surely for nix store paths it'll be read-only
<colemickens>
sphalerite: I'm quoting "deploy those store paths to the hosts".
<colemickens>
1. No, just no. Pods should not have host access, r/w access is even more scarey. 2. What's the point, k8s can't execute outside of docker anyway, so what's the point of writing into the store?
<catern>
ah you misunderstood
<colemickens>
But I also remember reading about this and being intrigued, so I'm probably forgetting pertinent details.
<catern>
you run nix on the host and it deploys the store paths before the pods start
<jasongrossman>
Or is that only 18.09 and not unstable? Anyway.
<sphalerite>
jasongrossman: but that's on master, not unstable
tmaekawa has quit [Quit: tmaekawa]
<jasongrossman>
sphalerite: Oh, I see. Quick work anyway!
<sphalerite>
indeed
<jasongrossman>
sphalerite: And the bug was only in master not unstable so it's all good.
<sphalerite>
convenient x)
<jasongrossman>
sphalerite: :-) ... The new zfs 0.7.11 is almost exactly the same as 0.7.9, with the regression from 0.7.10 fixed and I think just one tiny other change.
Kelppo has joined #nixos
<infinisil>
Currently trying to get module system submodules to work with option declarations in definitions
<infinisil>
If anybody happens to know that this is impossible, please let me know lol
jperras has joined #nixos
pie__ has joined #nixos
pie_ has quit [Remote host closed the connection]
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]