<avn>
adisbladis: you can't set modules dir at runtime as I know. So it need patch to allow full paths in `module-load` command
<adisbladis>
avn: Yeah that's currently not possible. What you could do is letting the module system create a directory with aggregated modules
<adisbladis>
(the module system == nixos module system)
<adisbladis>
Hehe, module is a bit overloaded in this context
<avn>
adisbladis: but no way to pass this directory to pulseaudio daemon (at least w/o rebuild)
<infinisil>
mdash: Nix is pretty reproducible for a specific revision of nixpkgs, so if you're getting the error, everybody else will get the error too. If that happens, try it on master to see if it's still broken, and if it is, you can file an issue
<CMCDragonkai>
What do you guys think if packaging upstream transfer learning weights? Right now common weights are being moved around in google cloud or aws s3, but these weights were trained and distributed along with high impact deep learning papers and everybody uses them?
<adisbladis>
avn: You can set it in the config which is read at startup
<CMCDragonkai>
i know of ones for resnet101, resnet50... all of the keras weights.. etc
<infinisil>
CMCDragonkai: I fail to see what this has to do with Nix
<CMCDragonkai>
infinisil: As in packaging it in nixpkgs.
<avn>
adisbladis: really? I though it was only via ./configure and/or patching
<CMCDragonkai>
I saw there was a package on nixpkgs that distributed data sources for particle physics
<jackdk>
Either it's sent to the wrong channel or CMCDragonkai intends to use nix and references to specific nixpkgs versions to help with reproducability in science
<adisbladis>
avn: Wait a second. I'll try to locate it again.
<CMCDragonkai>
infinisil: There is also the libpostal package which requires a weights file to be packaged along for it work
<adisbladis>
avn: dl-search-path I think it is
<avn>
adisbladis: looks cool, if it would work.
<benley>
anyone know where I can find the BSD "dump" and "restore" utilities in nixpkgs?
<adisbladis>
avn: I have the day off today so I have some time to experiment :)
<avn>
My own work is on rid-off pulseaudio (and other server related) stuff from $PATH, if it live as system-wide
<avn>
adisbladis: and I definelly know, whom to mention for review ;)
<infinisil>
CMCDragonkai: Ah, well it would be okay with me
<adisbladis>
avn: Go ahead =)
<CMCDragonkai>
infinisil: Is it ok with you because of precedents, or is it ok with you morally and philosophically?
<infinisil>
But one needs to make sure to not make hydra download it
<CMCDragonkai>
infinisil: How does one do that?
<adisbladis>
CMCDragonkai: hydraPlatforms = [];
<infinisil>
CMCDragonkai: Just with how nixpkgs is used I guess
<infinisil>
adisbladis: I'd have used `preferLocalBuild = true;` and `allowSubstitutes = false;`
<infinisil>
Or maybe only the former
<adisbladis>
infinisil: Yeah that's better
<adisbladis>
`hydraPlatforms = []` is a bit of a weird construct
cdepillabout has joined #nixos
<cdepillabout>
Is it possible to find the path to the /nix/store/*.drv file for a given derivation in /nix/store?
<{^_^}>
[nixpkgs] @fpletz pushed commit from @clefru to release-18.09 « spl/zfs: 0.7.9 -> 0.7.10 »: https://git.io/fAw0v
<ogle>
I am trying to modularize the build expressions in my companies monorepo. Currently we have a single default.nix that includes the deps for all of our subprojects. Instead I would like to have a default.nix per project, so I can create a top level release.nix that allows for installing a specific project. That said is there a way I can compose all of these in a shell.nix file such that during development,
<ogle>
everyone can run nix-shell and have the dependencies to be able to work across the entire repo?
rauno has joined #nixos
worldofpeace has joined #nixos
<kalbasit[m]>
mpickering: I figured out the bazel stuff. I'm cleaning it up, I'll comment on your discourse post when done
Hotkeys has joined #nixos
yodeler has joined #nixos
ydlr_ has joined #nixos
ydlr has quit [Ping timeout: 240 seconds]
ydlr has joined #nixos
yodeler has quit [Ping timeout: 252 seconds]
<ogle>
anybody have any ideas regarding my monorepo stuff?
graphene has quit [Remote host closed the connection]
graphene has joined #nixos
<{^_^}>
[nixpkgs] @kalbasit opened pull request #46509 → bazel-watcher: init at 4d5928e → https://git.io/fAwEm
<Edes>
Does anyone know how to use a python virtualenv on jupyter? When I run `ipython kernel install --user --name whatever` while in the virtualenv it errors out saying [Errno 13] Permission denied: '/tmp/tmpyv049grd_kernels/python3/kernel.json'
<Edes>
Perhaps you want `sudo` or `--user`?
<Edes>
this is the output of nix-info system: "x86_64-linux", multi-user?: yes, version: nix-env (Nix) 2.0.4, channels(erobl): "", nixpkgs: /nix/var/nix/profiles/per-user/root/channels/nixos/nixpkgs
<adamantium>
clever: got a minute. have a nixos zfs question for you.
Ariakenom has joined #nixos
<LnL>
Edes: if you want to use pip directly you'll need something like virtualenv, take a look at this for the nix approach https://nixos.org/nixpkgs/manual/#python
reinzelmann has joined #nixos
jasongrossman has quit [Ping timeout: 246 seconds]
<Edes>
that's what I did
hamishmack has quit [Ping timeout: 272 seconds]
ogle has left #nixos [#nixos]
mounty1 has quit [Ping timeout: 252 seconds]
georges-duperon has joined #nixos
fyber has quit [Ping timeout: 240 seconds]
fyber has joined #nixos
phreedom_ has quit [Remote host closed the connection]
phreedom has joined #nixos
sbdchd has joined #nixos
jackdk_ has joined #nixos
jackdk has quit [Disconnected by services]
jackdk_ is now known as jackdk
sbdchd has quit [Ping timeout: 240 seconds]
Edes has quit [Ping timeout: 245 seconds]
<LnL>
I don't know what that command does, bu it seems to try to write files it shouldn't
abueide has quit [Remote host closed the connection]
<adisbladis>
ogle: You could just import each project in your top-level shell.nix
abueide has joined #nixos
<adisbladis>
adamantium: IRC protip: Just ask your question, dont ask to ask.
<adisbladis>
adetokunbo: Hey. Not at the moment.
mayhewluke has quit [Ping timeout: 245 seconds]
mayhewluke has joined #nixos
acarrico has quit [Ping timeout: 252 seconds]
<xgirl>
ok, does anyone know how to create jobset for each of PR that is in the spec.json generated using the githubpulls plugin?
acarrico has joined #nixos
<adamantium>
adisbladis: yea
<adamantium>
:) Anyone know if it is possible to specify which zfs root dataset is to be loaded from the grub menu in nixos? E.g. in gentoo with dracut, you can use a kernel parameter like this: root=zfs:zroot/ROOT/my-other-root-dataset
<adamantium>
I do not think it is possible, it seems that nixos is baking the root dataset into the initramfs or something.
jackdk has quit [Ping timeout: 240 seconds]
<adamantium>
(baking //the path// in to the root dataset...)
graphene has quit [Read error: Connection reset by peer]
graphene has joined #nixos
<realrokka>
adamantium: have u looked at zedenv ? i don't use it, but maybe it's kinda what you want ʅ(́◡◝)ʃ
<adamantium>
that won't work for us
<adamantium>
it is just a beadm alternative.
<adamantium>
i mean it will create a boot environment, sure.
<adisbladis>
adamantium: It doesn't take stateful things like databases into account but other than that I can't see why I would need one on nixos
phry has quit [Ping timeout: 272 seconds]
Ariakenom has joined #nixos
jensens has joined #nixos
<adamantium>
using real zfs boot environments would still be quite powerful on nixos
<adamantium>
you could have a boot env for unstable , all kinds of test environments you can just thrash around in without a sweat
jasongrossman has joined #nixos
<adamantium>
But it is true, nixos is miles ahead of other distros in having what it does with rollbacks, it is quite awesome!
jasongrossman has quit [Remote host closed the connection]
<Hotkeys>
tonight I dove right in and installed nixos for the first time on my laptop
<Hotkeys>
it's pretty neat
jasongro` has joined #nixos
georges-duperon has quit [Ping timeout: 252 seconds]
<aminechikhaoui>
Hotkeys: ++
<{^_^}>
Hotkeys:'s karma got increased to 1
jluttine has quit [Ping timeout: 252 seconds]
tzemanovic has quit [Remote host closed the connection]
adetokunbo has quit [Quit: This computer has gone to sleep]
Boomerang has joined #nixos
<adamantium>
cool Hotkeys. made a great choice.
tzemanovic has joined #nixos
palo has quit [Ping timeout: 250 seconds]
nckx has quit [Quit: Updating my GNU GuixSD server — gnu.org/s/guix]
WilliButz has quit [Read error: Connection reset by peer]
WilliButz has joined #nixos
goibhniu has joined #nixos
lawlesseel has quit [Ping timeout: 256 seconds]
nckx has joined #nixos
bbsl has joined #nixos
Boomerang has quit [Quit: WeeChat 1.9.1]
palo has joined #nixos
palo has quit [Changing host]
palo has joined #nixos
NinjaTrappeur has joined #nixos
<bbsl>
does anyone have any experience with php in nix? Im trying to enable the zmq extention in php but I cant figgure out how it is intended to work.
mounty1 has joined #nixos
phry has joined #nixos
WilliButz has quit [Quit: WeeChat 2.0]
WilliButz has joined #nixos
Phillemann has left #nixos ["WeeChat 2.2"]
mounty1 has quit [Ping timeout: 245 seconds]
graphene has quit [Remote host closed the connection]
NinjaTrappeur has quit [Ping timeout: 252 seconds]
hyper_ch2 has joined #nixos
NinjaTrappeur has joined #nixos
jluttine has joined #nixos
<mpu>
how does nix deal with versions? sometimes a software will bump its version number but its "API" will remain the same; does one have to write a derivation for each version?
<mpu>
(Possibly factoring it out as a function and calling it, I guess.)
<emily>
you can -- yeah.
<bbsl>
goibhniu thank you for the link I guess that shows how once can go about doing it. but is this possible to do on in a shell.nix kind of way? I guess I wont have access to 'system' there.
<Myrl-saki>
mpu: Yes, and that's what glibc does. At least in older versions.
<mpu>
I'm more interested in the idiomatic way I guess.
phry has joined #nixos
mkoenig has quit [Remote host closed the connection]
<Myrl-saki>
mpu: You can also use override{Attrs,Derivation}, but that's less flexible, and might be more ugly for some.
<goibhniu>
bbsl: sorry, I don't know how. It must be possible though.
<Myrl-saki>
Seems like NIX_DEBUG no longer works.
<sphalerite>
mpu: for nix itself, different versions are completely distinct derivations, so any dependent built against two different versions of a library is also two different derivations
hexa- has joined #nixos
<sphalerite>
mpu: as for how nixpkgs handles multiple versions, the approach is to have only the latest of everything and keep older versions when necessary for software that depends on it
<sphalerite>
mpu: for older versions of software that aren't in the latest nixpkgs, typically you'd use an older nixpkgs version
<mpu>
sphalerite, ah thanks; that clarifies things.
<Myrl-saki>
OTOH, the dependency checking is per-derivation, so you could make a super-derivation that wraps all versions, but that's ugly.
<sphalerite>
mpu: one way is to use an overlay that imports an older nixpkgs and replaces a library completely, or uses override to plug it into specific dependents
<Myrl-saki>
Seems like nix-build no longer accepts NIX_DEBUG, and neither does any other nix programs.
<sphalerite>
mpu: if you could provide more concrete details (what do you want an older version of and for what purpose?) it might be easier to explain
<Myrl-saki>
And how do you want to call it!
<mpu>
sphalerite, I'm just building my mental image there :)
Izorkin has joined #nixos
<Myrl-saki>
You could use -I flags, to import the whole nixpkgs, which I'm guessing symphorien's link shows.
Guanin_ has joined #nixos
<Myrl-saki>
OTOH, if you want to get an older derivation, through Nix the language, you can use (import (fetchTarball ...))
<mpu>
sphalerite, I do plan to develop something with LLVM and it'd be handy to stick to a given version given how often they break their api.
<sphalerite>
mpu: there are lots of different ways depending on how you want to combine things
<sphalerite>
mpu: ah in the case of llvm we do actually have multiple versions in nixpkgs
<sphalerite>
bbsl: by the looks of it you'd want to just write a php.ini that points to the extensions you want to use and run php with -c your-php.ini
pie___ has quit [Ping timeout: 244 seconds]
copumpkin has quit [Ping timeout: 244 seconds]
adetokunbo has joined #nixos
<{^_^}>
[nixpkgs] @andir opened pull request #46511 → [18.03] WIP: firefox: provide clang configure flags on i686 → https://git.io/fAwo2
johanot has joined #nixos
<mpu>
sphalerite, I'm using nix on debian and the documentation keeps referring to configuration.nix; where can I find that?
norfumpit_ has joined #nixos
<bbsl>
mpu afaik configuration.nix is only used in nixos not if you use nix on another distro?
sbdchd has joined #nixos
<bbsl>
sphalerite: I think you are correct unfortunately part of the problem here is that I am new to php :) but ty for the pointer
<sphalerite>
mpu: yeah what bbsl said. Are you looking at the nixos manual?
<mpu>
Yeah; the one you linked :P
<sphalerite>
I linked something?
copumpkin has joined #nixos
<mpu>
08:27 > symphorien < ,pinning mpu
norfumpit has quit [Ping timeout: 244 seconds]
<mpu>
oops, nick confusion
<sphalerite>
oh right :)
mkoenig has joined #nixos
<sphalerite>
yeah I'd suggest pinning nixpkgs as described in the article symphorien got the bot to link you :)
sbdchd has quit [Ping timeout: 240 seconds]
fendor has joined #nixos
norfumpit_ has quit [Quit: norfumpit]
norfumpit has joined #nixos
tzemanovic has quit [Remote host closed the connection]
tzemanovic has joined #nixos
johanot has quit [Quit: Lost terminal]
norfumpit has quit [Client Quit]
<hyper_ch2>
question, linux kernel 4.19 will introduce new wifi ax support.... do you also need different wifi cards or can this be used with just software upgrade?
johanot has joined #nixos
knupfer has joined #nixos
norfumpit has joined #nixos
<sphalerite>
hyper_ch2: the hardware has to support it
tzemanovic has quit [Remote host closed the connection]
<hyper_ch2>
sphalerite: I thought if the same spectrums are used (not check if they are with ax) then it's just a matter of the software/firmware whether its supported :(
IRC-Source_31426 has joined #nixos
IRC-Source_31426 is now known as xgirl
justanotheruser has quit [Quit: WeeChat 2.0]
<sphalerite>
I don't understand a great deal about how wireless comms work, but AFAIU a lot of the details of how data is modulated and sent across the air are implemented in hardware
<sphalerite>
I'm far from an expert though
<hyper_ch2>
sphalerite: well, I'm no expert either :)
<sphalerite>
hyper_ch2: 802.11n was able to use 5GHz bands and you still need new hardware for 802.11ac so…
<hyper_ch2>
right... oh well :)
<sphalerite>
If you want support for all the things, get yourself an SDR :p
knupfer has quit [Ping timeout: 252 seconds]
<hyper_ch2>
one more thing: in nginx, can I have two locations entries with reverse proxy?
<sphalerite>
yes
<hyper_ch2>
SDR?
<sphalerite>
software-defined radio
<hyper_ch2>
:)
<sphalerite>
hyper_ch2: also even if it were possible from the hardware perspective, manufacturers would probably not update the (proprietary of course) firmware to support it because then you wouldn't buy their new hardware!
<tobiasBora>
I don't understand something: I put in my user configuration " extraGroups = ["wheel"];", to make it sudo, but when I start a new bash with "bash", "groups" only return "users"...
<hyper_ch2>
tobiasBora: can you "sudo" with it?
<sphalerite>
hyper_ch2: does it say anything useful in the logs?
<tobiasBora>
hyper_ch2: no
<sphalerite>
tobiasBora: you have to log out and back in for group changes to take effect
<hyper_ch2>
sphalerite: where are nginx logs?
<clever>
tobiasBora: you have to logout and log back in
<hyper_ch2>
they aren't in /var/log/....
<sphalerite>
tobiasBora: on all linuxes, not just nixos
AshyIsMe has joined #nixos
<tobiasBora>
sphalerite: clever : even if I start a new shell? In debian I'm pretty sure I used to start a new shell and it worked
<sphalerite>
hyper_ch2: if it's failing to start, journalctl -eu nginx
<sphalerite>
hyper_ch2: access and error logs are in /var/spool/nginx iirc
<sphalerite>
tobiasBora: I'm pretty sure you remember wrong ;)
<clever>
tobiasBora: only if the new shell is launched by something root (sshd, login shells)
<hyper_ch2>
sphalerite: will check, thx
<etu>
tobiasBora: you can launch a new shell and do "su yourself" to get your new groups
<etu>
tobiasBora: ie, a login.
<clever>
etu: ive tried that before, but it somehow breaks pulseaudio
<tobiasBora>
etu: good point, I like your method ;)
<etu>
clever: That doesn't surprise me at all
<tobiasBora>
really?
<tobiasBora>
that's kind of fun^^
<tobiasBora>
Any way, it works, thank you, sorry for the noise
<sphalerite>
A while back, I did a little hack to get an already running nix build to switch to using more cores, using gdb to get make to re-exec itself. It was pretty fiddly, involving calls to malloc and building up the new argv manually. Is there an easier way?
<clever>
sphalerite: impure!!!
<sphalerite>
clever: no impurer than NIX_BUILD_CORES is anyway :)
<xok>
you know what?... I start liking NixOS... :-D
<Taneb>
sphalerite: oh god that sounds... an evil way of doing things
<sphalerite>
Taneb: from a practical perspective it's still pure, and it saved me several hours of building, so… :D
<sphalerite>
xok: ONE OF US
<Taneb>
sphalerite: oh, the end result is nice, sure
<sphalerite>
Taneb: any better suggestions? :D
<Taneb>
No, it just feels like saving the world by blowing up orphanages
<clever>
:D
<sphalerite>
o.O that bad!
<Taneb>
sphalerite: scale both sides down a few orders of magnitude
<sphalerite>
hm, can processes be reparented somehow?
<sphalerite>
besides when the parent exits
<sphalerite>
reparented arbitrarily that is
<clever>
only pid 1 can adopt orphans
<clever>
you would need kernel changes to be able to freely change the parent pid
<clever>
but it does seem like a fairly simple change
<sphalerite>
I'd like to be able to drop a child process on some other process's doorstep
<sphalerite>
:p
<clever>
:D
tzemanovic has joined #nixos
<sphalerite>
Mic92_: iirc your research stuff in principle allows getting an interactive shell with extra tools within a nix build?
<clever>
sphalerite: unshare might be able to do that
<clever>
it can accept a pid to enter the namespace of
<{^_^}>
[nixpkgs] @michaelpj opened pull request #46512 → haskell lib: add --enable-benchmarks in doBenchmark → https://git.io/fAwii
<{^_^}>
[nixpkgs] @srhb pushed commit from @kalbasit to release-18.09 « bazel: 0.16.0 -> 0.16.1 »: https://git.io/fAwXf
<hyper_ch2>
sphalerite: that helped.... log is in /var/spool/nginx/logs and proxy_set_header X-Forwarded-For $proxy_add_xforwarded_for; != proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
<sphalerite>
clever: isn't that only for creating a new PID namespace?
<{^_^}>
[nixpkgs] @srhb merged pull request #46503 → aws-vault: 4.1.0 -> 4.3.0, pass linker flag for specifying app version → https://git.io/fAwn1
<sphalerite>
clever: pretty sure it is — unshare with PID means that the next fork will result in PID1 of the new namespace
<{^_^}>
[nixpkgs] @Mic92 pushed commit from @erikarvstedt to release-18.09 « networking.hostId: fix cmd in description »: https://git.io/fAwXy
<{^_^}>
[nixpkgs] @rembo10 opened pull request #46514 → headphones: init at 0.5.19 → https://git.io/fAwX5
sbdchd has quit [Ping timeout: 252 seconds]
matto has joined #nixos
<bbsl>
Im going to try again because I can not figgure it out: does anyone know how to enable the zmq extention for php in nixos? I cant for the life of me get it to work :o
<rawtaz>
bbsl: in what way can ou not get it to work?
<rawtaz>
+y
<rawtaz>
describe symptoms, otherwise it will be hard to help
baimafeima has joined #nixos
<bbsl>
rawtaz sorry, well its just not enabled. no matter what I try "php --ri zmq" yields "Extension 'zmq' not present"
IRCsum has quit [Ping timeout: 272 seconds]
<srhb>
bbsl: How did you try to enable it?
<etu>
bbsl: I don't think it's enabled by default and I also don't think we have a flag for it currently
<goibhniu>
bbsl: did you try creating the php.ini file?
<etu>
<php/default.nix>
<bbsl>
at first I installed phpPackages.zeromq alongside php and expected it to work .. it didnt ... then I tried finding options for it in the php package, there was'nt .. then I tried solving it using this approach: https://nixos.wiki/wiki/Phpfpm , that isnt applicable to my use case as I need the php binary to work
<etu>
I thought that would work
Twey has joined #nixos
<bbsl>
etu I think that is what I want, but I need to set the path it searches for the file in I think
<clever>
bbsl: i think you need to generate a php.ini that tells it to load the zmq plugin from the correct path
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
carlosdagos has quit [Quit: Connection closed for inactivity]
<bbsl>
ok ty, I think I have an idea what to try now
simukis has joined #nixos
<sphalerite>
I have a bridge defined using networking.bridges, and I want to connect an interface to it *only if it exists* — currently the bridge's creation will fail if the interface to connect to it doesn't exist. How can I do this?
hamishmack has joined #nixos
<clever>
sphalerite: i think i saw a recent pr that makes adding a device silently fail
<steveeJ>
has anyone noticed weird console behavior in nix repl? the prompt 'nix-repl>' is flickering
<LnL>
nixUnstable usually doesn't get bumped when there's a release
<LnL>
since it's the same wanyway
<srhb>
Yeah, I forgot. :-P
<steveeJ>
so somehow the release sneaked in behind unstable's back
<LnL>
yep :)
Dedalo has joined #nixos
orivej has joined #nixos
<tobiasBora>
manveru: yes, people seems to recommand to put in the file "/etc/sddm.conf" a line "Numlock=on" in the section General... But I've no idea how I should do that in Nixos, should I manually edit the file (looks dirty as it is likely to be erased on next update), or change something in my configuration.nix?
<rawtaz>
even if you dont remember all of it, you know where to look for what in the future
<tobiasBora>
clever: rawtaz : Hum, this page is way easier to search than the git sources! Thanks!
<clever>
tobiasBora: you can also just `man configuration.nix`
<{^_^}>
[nixpkgs] @srhb opened pull request #46517 → Revert build host target platform deprecation on master → https://git.io/fAwDX
<tobiasBora>
amazging!
<{^_^}>
[nixpkgs] @srhb opened pull request #46518 → Revert build host target platform deprecation on release-18.09 → https://git.io/fAwDD
booglewoogle has left #nixos ["Leaving"]
adetokunbo has joined #nixos
<joko>
Hello, it appears I am unable to update the nixos channel on my PC (custom channel based on nixos-unstable-small), any ideas what could be wrong?
<joko>
nix-channel --update seems not to download the channel at all
<clever>
joko: how long since you last updated the channels?
<clever>
it may be the ttl on the cache
tzemanovic has joined #nixos
<joko>
clever: just did
<clever>
joko: if you nuke ~/.cache/nix/tarballs/ then it will loose that cache
<joko>
I mean 5-10 minutes
<joko>
clever: you were right. I have forgotten to move the new nixexprs to the HTTPS server and fired up nixos-rebuild switch --upgrade, so it was cached indeed
<joehh>
hello, how do people handle HA on nixos? ie having a cluster of machines and ensuring that a service (timer or regular service) runs
<joehh>
on exactly one machine at a time?
nD5Xjz has joined #nixos
<rawtaz>
joehh: is it different from how they do it on any other linux distro?
<clever>
joehh: for some services like dhcpd, the daemon is able to talk to another instance of itself, and co-operate with it to share the network
<andi->
joehh: like on any other distributions I'd say? :) THere is no magic in nixos that handles failover detection. You could probably have a custom nixos option that enables/disables a bunch of timers and then you toggle that flag?
<magnetop`>
I just enabled hydra and now it's building graphviz, multiple times. is that normal?
hrk has joined #nixos
<clever>
magnetop`: hydra does depend on graphviz
<joehh>
I've used things like redhat cluster suite before, but wondering if anything similar or dependencies for RHCS have been packaged? (pacemaker etc)
<rawtaz>
joehh: it's really not what you are asking for, but somewhat related, do you know about NixOps?
<magnetop`>
clever: I'm just surprised that it's not using an upstream binary
<magnetop`>
clever: also. it looks like it's building a couple of times instead of once
<joehh>
I do :) using it right now for a bunch of remote physical machines and now starting to want to use it in AWS (and spread my services across multiple availability zones)
<clever>
magnetop`: is it building it within the hydra jobs, or during nixos-rebuild?
<joehh>
I'm kind of hoping to be able to specify a service in nixops that will run on exactly one of the machines in the nixops deployment (ie if one fails, another machine will start service)
<ikwildrpepper>
joehh: nixops is not really suited for that
selaux has joined #nixos
<rawtaz>
perhaps the stuff from HashiCorp provides this type of thing? unrelated to nixos though
nD5Xjz has quit [Ping timeout: 252 seconds]
<joehh>
ikwildrpepper: :( yet
<joehh>
rawtaz: I'll check it out
<ikwildrpepper>
in the nix world, what you are looking for sounds more like disnix, however, I have never used it
<ikwildrpepper>
disnix is more service deployment, where nixops is more system config deployment
<lesh>
hey, quick question on deploying to hetzner bare metal via nixops, which OS should I purchase in their ui? asking because I'm not sure which mechanism nixops uses to interact with the running server..
<lesh>
looking to switch from ansible managed kvm machines to nixos as a hypervisor..
<bbsl>
does anyone know how to find the location of a non binary package in the nix store withour resorting to grepping the whole catalog structure? (Iv tried nix-locate and the readlink that I use for packages that I have links to does not work in this case)
<joehh>
ikwildrpepper, rawtaz: looks like my nix stack may be getting deeper :)
<joehh>
thanks
<johanot>
lesh: I configured my hetzner server with nixops by giving nixops credentials to the hetzner control-panel. Nixops can boot the server in rescue mode and run nixos-install from there. No need to install any other OS first.
smolboye has joined #nixos
<lesh>
johanot: yeah just noticed by reading hetzner.py that it uses the rescue mode. I don't think I can select "no os", but I expect this doesn't matter since nixops will just overwrite everything. tnx!
<johanot>
lesh: you should be able to select "Rescue system"
<srhb>
lesh: Interested to hear about your experience with NixOS as a hypervisor host. I know a few people who need a good host for that, and I've been wary about recommending NixOS.
<lesh>
johanot: oh, I see, might have missed that, tnx
<joehh>
srhb: I am running a (very) remote nixos machine as a (small scale) hypervisor. Has been working really nicely (once I enabled the various bios extensions)
<lesh>
srhb: yeah, I'm about to refresh parts of my infrastructure so decided to give nixos a try, I've been unhappy with ansible+kvm (slow/wonky/ugly) and hashicorp and other corporate stuff feels unattractive, this is a huge leap, I hope it works well, it's about 15-20 machines
<joehh>
I am a little scared of upgrading it, but that is mainly due to its remoteness and current critica lreliance on the windows vm it is running.
<{^_^}>
[nixpkgs] @domenkozar pushed to master « dep2nix: fix eval (no IFD) »: https://git.io/fAwF8
<joehh>
I should be able to get a backup machine to that location in the next month or so and also have a replacement for the windows software it is running
<joehh>
which will be a much better position
seku has joined #nixos
<srhb>
joehh: Nice, thanks :)
<srhb>
lesh: I'll try to remember to poke you in a few weeks about sharing your insights. ;-)
<lesh>
srhb: sure, I'll stick around here :)
<srhb>
lesh: Great!
norfumpit has quit [Quit: norfumpit]
norfumpit has joined #nixos
orivej has joined #nixos
nD5Xjz has quit [Ping timeout: 245 seconds]
patrl has quit [Ping timeout: 252 seconds]
patrl has joined #nixos
nD5Xjz has joined #nixos
<rawtaz>
joehh: :D
<joehh>
rawtaz: yeah - nixops is awesome :)
<rawtaz>
are there any suepr duper great videos or guides showing off nixops in a very interesting way (although not too big scale, so it's still for us common folks)?
<joehh>
the way we use it is interesting (I think anyway...) - each deployment is kept in a git repo with a corresponding buildkite pipeline
vaibhavsagar has joined #nixos
<joehh>
making a commit and push triggers a build in the pipeline which deploys changes to test machines in AWS
nD5Xjz has quit [Ping timeout: 252 seconds]
<joehh>
once our support/devops team are happy the changes are succesful (sadly mostly semi manual testing at this stage)
xok has joined #nixos
sbdchd has joined #nixos
jasongro` has quit [Remote host closed the connection]
<joehh>
they unblock the pipeline (with a comment as to testing done) which first copies the changes to the remote physical systems
<joehh>
rawtaz: how are you trying to get it/them working?
<joehh>
vmwareGuest has always just worked (tm) for me
<rawtaz>
joehh: i would just try to make the contents of the strings in the file smaller, vmware guest tools probably dont need all of it to determine the os.
<rawtaz>
joehh: yeah, im not sure why it doesnt work here. im on 18.03
<rawtaz>
but this is the error message i get and the code seems straight forward, not sure what else it would be causing it
<rawtaz>
it's IMO at a point where it would be worth a try editing the file just to see if it starts working with that.
<rawtaz>
from a debugging perspective.
<joehh>
I've always just done: services.vmwareGuest.enable = true;
<clever>
joehh: i had made the above autoreboot.nix to handle kexec'ing into a nixos that lacks the proper network config
<rawtaz>
joehh: i have those two, and open-vm-tools in system packages. the service tries to start, but i get those error messages. seems straight forward to me?
<rawtaz>
the error message and the code corresponds perfectly fine.
worldofpeace has quit [Quit: worldofpeace]
<joehh>
I've never had open-vm-tools in system packages, but I doubt that is the issue
<matto>
quit
<rawtaz>
if i remove it it will probably just be added by the service enable directive i guess
logzet has joined #nixos
matto has quit [Quit: WeeChat 1.4]
<joehh>
quite probably
<rawtaz>
anyway ill try remounting that mountpoint where the file resides as rw and see if i can edit it then
silver has joined #nixos
<joehh>
clever: interesting so just reboot if nothing else changed
<rawtaz>
eh, there's no -u flag to `mount`, what the heck..
<clever>
joehh: yeah
<clever>
joehh: so you can kexec that image on ubuntu for ex, and if you fail to get into it within an hour, it reboots, and nothing has been installed, so ubuntu just boots back up
<joehh>
nice
<joehh>
very nice
<clever>
but once you are in, you still need to be very careful
<clever>
a recent user in here had formatted the host hdd, before noticing that he lacked zfs unstable support in kexec
<clever>
and then after finishing the entire install with legacy boot, he mentioned he was on an apple machine, which can only boot uefi
<clever>
but thats still recoverable with the ISO on a USB
<clever>
my original goal for kexec, was to use it on remote baremetal machines, which has no undo :P
<joehh>
at least no flights involved :)
sbdchd has joined #nixos
<joehh>
have you succeeded with that goal?
<tobiasBora>
I'm trying to add a .desktop to the "emacseditor" command that connects directly to the server. I saw that people usually use makeDesktopItem, so I tried to write my own makeDesktopItem... However I don't see how I could "plug" this into my emacs derivation: http://paste.debian.net/1041790
<clever>
joehh: the first machine had the raid array go south for some weird reason, and i had to debug it via ipmi, which required running 2 versions of windows in 2 types of vm's to deal with activex, lol
<tobiasBora>
let me see if i can do something with it ;)
<clever>
joehh: i then used that first machine as an example for the 2nd machine, which happened to have different network config, just because the datacenter likes to make every machine its own special little butterfly :P
<joehh>
:( that sort of thing is why nixos is infectious
sbdchd has quit [Ping timeout: 240 seconds]
<rsa>
running .../bin/switch-to-configuration switch activates a new configuration, but it doesn't symlink the profile
<clever>
joehh: the 2nd machine turned out to be a virtualbox vm inside a windows host, lol
<clever>
joehh: so, i just gave up on doing a remote install, did a local qemu install, and then converted it to a virtualbox disk image, and gave them the URL
Dedalo has joined #nixos
<rsa>
so i thought nix-env -p /nix/var/nix/profiles/system --set ... would fix that, which it does
<clever>
joehh: after booting it up, they asked for the root password, so they could fix the network config, then they noticed it was already working
<rsa>
but it doesn't add the boot entry, hmm...
<rawtaz>
hm, so by making a copy of /etc/os-release to /etc/release and editing that file to be much smaller, i got vmwareGuest to start. but when i enable shared folders in fusion i instead now get this: https://pastebin.com/yjnen1Sd
<rawtaz>
joehh: are you using shared folders in fusion and nixos vm_
<rawtaz>
?
<joehh>
:) I've had similar experiences when giving various it depts my vmware nixos images
<joehh>
rawtaz: no
<joehh>
normally just access network resources as regular network shares or tcp sockets
<rawtaz>
is there anyone in here maintaining the open-vm-tools package thats interested in debugging it?
patrl has joined #nixos
<sphalerite>
rawtaz: I haven't seen the named maintainer, joamaki, on IRC before. Probably easier to get hold of them by opening an issue on github and @ing them
<rawtaz>
ok
<rawtaz>
thanks
<{^_^}>
[nixpkgs] @alyssais opened pull request #46522 → zsh-history-substring-search: init at 1.0.1 → https://git.io/fAwp2
<rawtaz>
is there any point in trying the current configuration i have on a completely new install of nixos, to seee if that makes a difference? the current nixos i have was upgraded from 17.x to 18.03.
<rawtaz>
i would guess its rather pointless
<joehh>
rawtaz: I suspect your guess is right (ie pointless), but if it is easy to try, then it may shed some light
Twey has quit [Ping timeout: 240 seconds]
<rawtaz>
sphalerite: would i open such an issue in the NixOS/nixpkgs repository?
adetokunbo has quit [Quit: This computer has gone to sleep]
spacefrogg has quit [Remote host closed the connection]
aw has quit [Remote host closed the connection]
sbdchd has joined #nixos
jtojnar has quit [Remote host closed the connection]
xok has quit [Read error: Connection reset by peer]
xenog has joined #nixos
sbdchd has quit [Remote host closed the connection]
<rawtaz>
hm. after upgrading to 18.03, should i change system.stateVersion to that from 17.09 ?
<srhb>
rawtaz: Never ever do that unless the release notes tell you to
<rawtaz>
good. i havent seen it in the release notes for 18.03
patrl has quit [Ping timeout: 252 seconds]
<symphorien>
srhb: has it ever happened ?
<srhb>
symphorien: The release notes telling us to? Not while I've used NixOS.
<rawtaz>
if it never happened, how come it's on 17.09 in my install (that started on that version)? one can only presume that installing 18.03 from scratch would set it to 18.03.
<jasongrossman>
Me neither.
<srhb>
rawtaz: It's a migration protection
<rawtaz>
in which case changing this one to 18.03 should be fine, no?
adetokunbo has joined #nixos
<srhb>
Nope
<jasongrossman>
I'm afraid the name system.stateVersion is super misleading. It's completely accurate, but still.
<rawtaz>
hmm ok :o
<srhb>
rawtaz: It will prevent eg. databases from being moved to another location that a newer version of eg. the postgres module uses
<srhb>
rawtaz: So changing it without doing the migration by hand will just cause your databases to disappear, your cats to die and your soul to be void.
<rawtaz>
i see. so it's more of a "stuff was installed and set up according to this version" setting
<srhb>
rawtaz: It's always fine to _start_ a system at some version.
<srhb>
rawtaz: Exactly.
<rawtaz>
cool, thanks for the explanation.
<jasongrossman>
rawtaz: Exactly, and that would be a better name for it.
<rawtaz>
it might not be a FAQ, but perhaps that explanation should be added to the FAQs anyway?
gspia_ has joined #nixos
<srhb>
The name is pretty spot on. The system is at the current state-version. ie. the stateful stuff was setup according to this version.
semilattice has joined #nixos
<srhb>
rawtaz: Newer installers add a big fat note to the configuration.nix about it
<jasongrossman>
rawtaz: It is in the documentation. The problem, I think, is that the name looks self-documenting but isn't.
<rawtaz>
yeah i see a "dont change this or your cat's die" note here..
<jasongrossman>
srhb: The name is spot on, and yet it constantly misleads people. Both true, sadly.
<srhb>
Hm, okay.
<samueldr>
IIRC there was an update (or proposed update) to the stateVersion comment
<srhb>
Experience trumps principle. :-)
<samueldr>
,stateVersion
<{^_^}>
Setting stateVersion to the latest release doesn't update anything and can only break your setup at best. If you want to regardless, Ctrl-F for "stateVersion" in https://nixos.org/nixos/manual/release-notes.html to see things that need to be manually upgraded with the new value
<jasongrossman>
rawtaz: Syntax matters in NixOS. Try searching for "cats die" without the apostrophe.
<samueldr>
hm! I thought it was going to be a wiki FAQ entry :/
<johanot>
srhb: I don't have a cat. will it still die? :P
<samueldr>
life... finds a way
<srhb>
johanot: 'fraid so! It's a pretty dangerous thing.
<rawtaz>
i think a good thing would be if the note in the configuration.nix added something like "see the FAQ for more info", so people are directed to https://nixos.wiki/wiki/FAQ/When_do_I_update_stateVersion to understand this fully
<samueldr>
i should add "4. you have the time to fix unintended consequences of changing stateVersion"
<johanot>
srhb: uff scary stuff...
<samueldr>
though, in that case, anything in the system should be found in the manual, so not in the wiki, but migrated from the wiki to the manual I guess
<symphorien>
the factoid is not completely right, it can update postgresql for example
<samueldr>
right, it'll change the version installed
<srhb>
Hopefully only through versions that don't require pg_upgrade?
<samueldr>
so not "doesn't update anything" but "won't upgrade your system"
<samueldr>
srhb: using different stateful dirs IIRC
<srhb>
That's surely a bug.
<samueldr>
from 17.03 -> 17.09 I had a container without stateVersion and "lost" all data
<srhb>
That's terrible.
<samueldr>
(turns out everything was still in the stateful dir and put the stateVersion and everything was back)
<samueldr>
it's a loaded airsoft-footgun; it may sting, but at least the data isn't touched
<srhb>
I see, the dataDir is in fact dependent on the package version which may change despite stateVersion.
<srhb>
That's not good.
<srhb>
But yes, at least it's recoverable.
<srhb>
I'm fine with breakage on stateVersion changes.
<samueldr>
hmm, may change despite stateVersion? (haven't verified anything)
<srhb>
Huh, no idea really. Clearly it can't touch /etc without root. :P
<srid>
the nix commands work though
<symphorien>
~/.config/nix/nix.conf iirc
patrl has joined #nixos
<srhb>
Speaking of postgres, did the module rewrite not make the cut for 18.09? :(
<srid>
symphorien: I think you are right. I just created that file, added my config, and verified that they come through the output of `nix show-confing`
<symphorien>
:) nix show-config is really helpful when debugging
<joehh>
rawtaz, srhb: system.stateVersion is the sort of thing that makes me reluctant to upgrade remote hypervisors - probably ok, but not willing to risk consequences if it goes bad
<srhb>
joehh: Just don't touch the value unless you have to... It's really no worse than that :P
<joehh>
that is what I've learned tonight :)
<srhb>
And if you have to, there's probably someone in here to hold hands with for comfort and guidance. ^^
Ariakenom has quit [Ping timeout: 250 seconds]
patrl has quit [Ping timeout: 252 seconds]
endformationage has joined #nixos
tzemanovic has quit []
<joehh>
:) here is unbelievably helpful. People at work can't believe that someone called "clever" helped us with a problem instantly, but I'm still going to make sure I have some physical continencies (and someone friendly local) in place before I upgrade my more remote systems in the short term
<srhb>
Just wait till #nixos becomes subscription only, it will be a gold mine.
<joehh>
I think your nick is half of the amazement :) the smartest/most experienced people at $dayjob were stuck. We reached out to IRC and clever solves our problem in 30 seconds
reinzelmann has quit [Quit: Leaving]
<clever>
joehh: i think part of my problem is that i am too eager to help people, even if it means loosing sleep or starving, lol
<{^_^}>
[nixpkgs] @samueldr pushed 2 commits to release-18.09: https://git.io/fArTG
<clever>
i sometimes have to force myself to not answer questions
<joehh>
:) and :(
sbdchd has joined #nixos
<Taneb>
clever: what time zone are you in?
<clever>
Taneb: Tue Sep 11 11:01:13 ADT 2018
<joehh>
I know the feeling - hasn't been an issue here (my knowledge is still too low), but back when I was using salt, it was definitely a risk
<jasongrossman>
joehh: I'm in Australia too.
<joehh>
jasongrossman: qld? seems to be lots of queenslanders due to data61
<joehh>
I'm in Melb
rauno has joined #nixos
<Taneb>
clever: so, not too unreasonable right now
npmccallum has joined #nixos
<clever>
Taneb: i woke up at 5am, and while debugging hydra issues, got an email about declarative jobsets in hydra, and responded within minutes
<Taneb>
Maybe unreasonable *then*, but I'm not going to tell you to go to bed right now
<clever>
lol
sbdchd has quit [Ping timeout: 272 seconds]
patrl has joined #nixos
<rawtaz>
just curious, is there anything like NixOps for GuixSD ?
<jasongrossman>
joehh: NSW, and I don't knwo what data61 is. Well, I'm glad we have the states fairly evenly covered.
<rawtaz>
i can reproduce the vmware guest tools issue with a completely new 18.03 nixos vm :) issue coming up.
lesh has quit [Ping timeout: 240 seconds]
garbas has joined #nixos
bbsl has quit [Remote host closed the connection]
<joehh>
jasongrossman: :) good to hear - data61 is the "disruptive" offshoot from CSIRO
<jasongrossman>
Oh I see!
<joehh>
they do a fair bit of haskell and functional programming which includes nix
<jasongrossman>
Hardware purchase time. My old MacBook has just died, and I'd like something non-Apple but don't really know where to start. Any NixOS hardware issues I need to know about, please, anybody?
Guanin_ has quit [Remote host closed the connection]
<jasongrossman>
joehh: Cool.
mayhewluke has quit [Ping timeout: 245 seconds]
<srhb>
jasongrossman: Pretty nice generation of Thinkpads this year..
<jasongrossman>
I forgot to mention that I'd like to be able to put in 2.5" drives easily.
<jasongrossman>
I mean SATA drives.
<clever>
jasongrossman: the kudu has room for a single sata drive, and a single nvme drive
spear2 has quit [Ping timeout: 244 seconds]
<jasongrossman>
:-)
Twey has joined #nixos
<clever>
i got mine with only nvme on it, so there are no sensitive moving parts to deal with
Guanin has joined #nixos
<jasongrossman>
Right, wise. I just want to put in a 2.5" SSD that I already have.
orivej has quit [Ping timeout: 245 seconds]
<Henson>
hello everyone. Is there a way I can specify in a derivation a runtime dependence on other derivations? I know Nix does some automatic dependency checking, but I'm compiling Python to an executable using Nuitka, and need to specify that the resulting derivation depends on a bunch of Python packages. I've tried adding them to propagatedBuildInputs, but that doesn't help.
jtojnar has joined #nixos
<clever>
,runtimeDeps Henson
<{^_^}>
Henson: In order of preference: Patch source OR ((if it uses PATH -> wrap with new $PATH) AND (if it uses dlopen, (patchelf --set-rpath in postFixup OR wrap with new LD_LIBRARY_PATH)))
xok has joined #nixos
grumble has quit [Ping timeout: 600 seconds]
joehh has quit [Ping timeout: 252 seconds]
* Henson
tries it out
grumble has joined #nixos
Yaniel has quit [Max SendQ exceeded]
wpcarro has quit [Remote host closed the connection]
wpcarro has joined #nixos
<{^_^}>
[nixpkgs] @Chiiruno opened pull request #46525 → nixos/zeronet: Fix TOR permissions, add torAlways option → https://git.io/fArLy
<Henson>
clever: thanks for the tip. I found that runtimeDeps was used in the azure-agent.nix derivation in conjunction with wrapProgram, and it perfectly suited my use case.
<rawtaz>
yeah but this is the size of the original (failing) file: -r--r--r-- 1 root root 318 Jan 1 1970 /nix/store/1ylqnn8c0zx91szrq03i0gq62m6djkk9-etc-os-release
<rawtaz>
just 318 bytes
<samueldr>
yea, I was going to say I must be wrong
<NinjaTrappeur>
Hey I'm trying to create a derivation for a electron application I want to build from the source. I'm looking for an example in the nixpkgs archive without any success so far... Do you have any potential example in mind?
<xok>
I use fetchgit and it asks me for a hash...
<rawtaz>
im pondering if there could be something with the patches nixos presumably does that changes the size of that buffer to some really small value?
<xok>
how do I get that hash, does anyone know?..
<symphorien>
,tofu xok
<{^_^}>
xok: To get a sha256 hash of a new source, you can use the Trust On First Use model: use probably-wrong hash (for example: 0000000000000000000000000000000000000000000000000000) then replace it with the correct hash Nix expected.
<rawtaz>
xok: that can then be used with sha256 = "1z48rvlzwi7bq7j55rnb0gg1a4k486yj910z2cxz1p46lxk332j1"; in your config
<samueldr>
NinjaTrappeur: sorry, just thought I'd mention, haven't yet read the issue
<sheyll>
hi all!
<Myrl-saki>
Okay, I'm stupid, that doesn't work.
<ldlework>
hi sheyll
<rawtaz>
sphalerite: may be! unfortunately i dont know how to modify it to replace the link with a regular file. the mount command doesnt have -u to update the mount to rw
<Myrl-saki>
I'm confused now. How do I use a .drv in Nix?
<NinjaTrappeur>
no problem :)
<Myrl-saki>
Like, what if I don't wanna build the .drv beforehand?
<sheyll>
anyone using haskell-ide-engine (hie-nix)?
<sphalerite>
rawtaz: in the last bit rather than ln, do cp $path /etc/os-release
<Myrl-saki>
error: expression does not evaluate to a derivation, so I can't build it
<Myrl-saki>
symphorien: "Myrl-saki | Like, what if I don't wanna build the .drv beforehand?"
<Myrl-saki>
symphorien: I'm guessing I could technically us nix-store -q --outputs
jensens has quit [Ping timeout: 250 seconds]
<symphorien>
ah I thought "build" as in "nix-build"
<rawtaz>
sphalerite: oh.. so the /etc isnt mounted read only.. lol
<rawtaz>
xok: let me know if that helped
<Myrl-saki>
Yeah, I tried nix-store -q --outputs and that also requires it to be built beforehand.
<Myrl-saki>
Welps, not really a problem.
<Myrl-saki>
You'd often want to refer to the file iteslf anyway.
<Myrl-saki>
Oh wait, now I remember why.
<sphalerite>
rawtaz: nope. /etc is populated statefully by nixos's activation script with symlinks into the (read-only) store
<Myrl-saki>
I want to use traditional Nix stuff, such as `overrideAttrs`
<Myrl-saki>
But I'm guessing that that's already lost when it's saved to a .drv.
xok has quit [Read error: Connection reset by peer]
<rawtaz>
sphalerite: i verified that it doesnt make a difference if the /etc/os-release file is a regular file or a symlink, same symptoms
<sphalerite>
rawtaz: now that it's a regular file, you could try shortening it as well to see if it makes a difference maybe
<rawtaz>
sure, it will probalby work if i remove the last three lines. wil try it.
<rawtaz>
thing is, it looks through both os-release and release
<rawtaz>
yeah, removing the last three lines to make /etc/os-release shorter/smaller works, just like with /etc/release as a substitute
<clever>
Myrl-saki: nix-store --query --binding out /nix/store/foo.drv
<rawtaz>
so yes, it should be considered established that the size of that file is one of the problems
<Myrl-saki>
clever: Right, but it errors since the drv fails the dependency check.
<Myrl-saki>
the drv, being the resulting file.
<sphalerite>
rawtaz: maybe patching it to increase the buffer size will help then?
<clever>
Myrl-saki: ah, yeah, nix is out of the story once you have .drv files
<rawtaz>
sphalerite: maybe. i guess one should first investigate what the current size limit is, and why it's so low. i mean, if as you say it should be 1024, then it's really weird that a file 340 bytes large fails
<Myrl-saki>
clever: Look at the bright side, this means that all .drv files are effectively GPL. :)
<clever>
lol
<sphalerite>
rawtaz: maybe it decodes it to UTF-32 or something, making each character take up 4 bytes rather than one. Just a completely random guess of course, but 340*4 = 1360 > 1024
<Myrl-saki>
I tried compiling my webkit project...
<Myrl-saki>
judson: I mean what's wrong with using nix-build to add a gcroot?
<Myrl-saki>
nix-build then nix-shell, basiclaly.
<ldlework>
Is anyone here using qutebrowser? The hacks to make it find the stuff it usually puts in ~/.local/share/ into the nix store doesn't seem to work properly.
<rawtaz>
sphalerite: sorry but where in my configuration.nix would i place that? i currently have almost a default configuration.nix
sbdchd has quit [Remote host closed the connection]
<Myrl-saki>
"Even if you are using Nix only occasionally, it is easy for /nix/store to go beyond 50G."
<sphalerite>
rawtaz: yep that should work
<Myrl-saki>
This never happened to me until I increased my disk size.
<rawtaz>
thanks then. i will try that when i get some time. i need to continue focusing on getting other things to work :)
<Myrl-saki>
OTOH, maybe it's because it literally couldn't happen to me until I increased my disk size.
rfold has joined #nixos
<judson>
You can see why I might wish that nix-shell had a flag to Just Do This.
<ldlework>
no qutebrowser users?
<judson>
sphalerite, that doesn't look so crazy. I might update my ~/.direnvrc...
sbdchd_ has joined #nixos
<sphalerite>
judson: it doesn't have any facility for cleaning up or updating though. Also it creates a whole bunch of invalid roots which the GC will complain about (but continue working)
<tobiasBora>
I'm trying to fix my ntp issue.
<tobiasBora>
So I installed ntp
<tobiasBora>
But I can't even do "ntpdate 0.europe.pool.ntp.org
<tobiasBora>
(or any server, debian, nixos, google...)
<tobiasBora>
I've the error 11 Sep 20:30:16 ntpdate[2975]: no server suitable for synchronization found
sbdchd__ has joined #nixos
xok has joined #nixos
<tobiasBora>
Do you think it is a nixos issue (port to open?)
<tobiasBora>
or a sys admin error?
<tobiasBora>
(someone blocked the port for the computer)
jedahan has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<cransom>
most likely, the port is blocked.
<sphalerite>
tobiasBora: yeah sounds about right. Are you at work?
<sphalerite>
maybe you have a local ntp server you can use
<cransom>
it became common to block ntp and require use of your own internal ntp servers due to DoS prevalance
sbdchd_ has quit [Ping timeout: 244 seconds]
<sphalerite>
clever: any idea if there's a way around "syscall kexec_file_load not available." which kexec reports whenever I pass a kernel command line to it?
sbdchd__ has quit [Ping timeout: 252 seconds]
<sphalerite>
well, baking my desired command line into the kernel Iguess. But that's not feasible for my use case
<xok>
I also want to create a virtual host and set its root directory to a relative path of the nginx conf files...
<ldlework>
What would be the best way to do a little bash interpolation thing such that you could get a file from a nix package for the current location in the store for that file
<xok>
how do I do something like this in a main file: include ${cfg.package}/conf/fastcgi_params;
<xok>
the nixops alerts cfg variable is not found...
<ldlework>
like if there is a file $out/share/foobar.txt how can I get the current store path for that package
revtintin has quit [Quit: WeeChat 1.9.1]
jperras has quit [Ping timeout: 252 seconds]
xenog has quit [Ping timeout: 246 seconds]
stanibanani has joined #nixos
<symphorien>
xok: cfg is a let binding higher in the file
stanibanani has left #nixos [#nixos]
<ldlework>
"$(nix-build '<nixpkgs>' --no-build-output -A fooPkg)/share/foobar.txt" maybe
<ldlework>
yeah it works nice
<xok>
symphorien: yes, and that works if I use it in a separate file...
<xok>
symphorien: but in the main file I can't to add: { config, pkgs, ... }:
<symphorien>
"main file" ?
<xok>
yes, the main file for the nixops ...
<symphorien>
if it is nixops specific I don't know
<elvishjerricco>
tilpner: I might just try implementing it and seeing if it gets merged :P
<tilpner>
With or without GC safety?
<elvishjerricco>
tilpner: I guess I can make sure it prints before the process releases the lock or something, but ultimately I'm not going to focus on making it safer than `nix-build` currently is
<ldlework>
QuteBrowser has userscripts, and one of them uses the Python package tldextract. Isn't this typically enough to get an application to have access to an additional Python package at runtime?
<teej>
I just wanted to use the Nix package manager on my Raspberry Pi 2 Model B.
johnw_ has joined #nixos
jedahan has joined #nixos
<ldlework>
Is my problem that qutebrowser is being passed a python3 nix-package and the tldextract python package I'm adding to propagatedBuildInputs is not seen by the passed in python interpreter used by qutebrowser?
<elvishjerricco>
teej: You might be able to build it from source and bootstrap the installation yourself.
knupfer has quit [Remote host closed the connection]
<ldlework>
i have no damn idea lol
_ris has joined #nixos
knupfer has joined #nixos
jedahan has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jedahan has joined #nixos
vaibhavsagar has quit [Remote host closed the connection]
jedahan has quit [Client Quit]
vaibhavsagar has joined #nixos
jedahan has joined #nixos
jedahan has quit [Client Quit]
rihards has joined #nixos
<exarkun1>
I am trying to build a new librustzcash but it has a Cargo.lock that references a git+https uri (blake2-rfc) which I think causes the build to explode trying to ... eh, I don't know, do something. I don't know Rust. What do I do about that?
jedahan has joined #nixos
jedahan has quit [Client Quit]
jedahan has joined #nixos
jedahan has quit [Client Quit]
knupfer has quit [Client Quit]
ma27 has quit [Quit: WeeChat 2.2]
<xenog>
I have made sure that some Haskell packages I built work well against Nix. How is the process by which NixOS includes new Haskell packages?
<xenog>
I have seen that there is a huge file with Haskell packages listed that is auto-generated. I don't understand how it is auto-generated, and if there is a process I can follow to make sure that my packages get included.
ma27 has joined #nixos
exarkun_ has joined #nixos
matthewbauer has joined #nixos
knupfer has joined #nixos
logzet has joined #nixos
<{^_^}>
[nixpkgs] @Ericson2314 opened pull request #46534 → misc pkgs: Darwin -> Linux cross compilation → https://git.io/fArrm
hotfuzz has quit [Read error: Connection reset by peer]
sbdchd__ has quit [Remote host closed the connection]
hotfuzz has joined #nixos
sbdchd has joined #nixos
sigmundv has joined #nixos
sbdchd has quit [Ping timeout: 246 seconds]
xenog has quit [Remote host closed the connection]
sigmundv__ has joined #nixos
xenog has joined #nixos
sbdchd has joined #nixos
sbdchd_ has joined #nixos
logzet has quit [Remote host closed the connection]
<ldlework>
I have a system Python 3 interpreter installed but I don't know where it comes from
<ldlework>
lol
sbdchd has quit [Ping timeout: 244 seconds]
rihards has quit [Quit: rihards]
rihards has joined #nixos
sbdchd_ has quit [Ping timeout: 264 seconds]
<exarkun1>
I stuck a patch in cargoPatches and it didn't get applied :/ Anymore. On the first build host it worked fine, at least once.
<exarkun1>
When could one reasonably expect that PR to be accepted? And appear in the ... stable channel?
<jmc>
Hello, a haskell package I'm trying to use was updated two weeks ago on hackage, but nixpkgs still has an old version. Should I just be patient, or stake this up with the maintainer of the package? nixpkgs is usually fairly up-to-date, no?
<jmc>
s/stake/take
<symphorien>
no idea. it will at best appear in 18.09 at the end of the month.
kenshinC1 has joined #nixos
rprije has quit [Read error: Connection reset by peer]
<jmc>
symphorien: alright, patience it is, thanks
<exarkun1>
It probably makes more sense for me to figure out a work-around meanwhile. Setting aside this project for 3 weeks isn't really an option.
<symphorien>
jmc: this was for exarkun1 , sorry
<symphorien>
exarkun1: you can use the head of the PR as nixpkgs
<exarkun1>
Why doesn't `buildRustPackages` apply the patches I give it in `cargoPatches`"?
<jmc>
symphorien: oh oops :^)
rprije has joined #nixos
<cocreature>
jmc: major upgrade?
<{^_^}>
[nixpkgs] @The-Compiler opened pull request #46535 → qutebrowser: Fix patching of standarddir.py → https://git.io/fAr64
<exarkun1>
symphorien: maybe but I don't know how
<jmc>
cocreature: yes
logzet has joined #nixos
<cocreature>
jmc: afaik nixpkgs mostly follows stackage lts for packages that are in stackage so it doesn’t do major upgrades
graphene has quit [Read error: Connection reset by peer]
<symphorien>
ldlework: this happens automatically as part of the fixupPhase on all $out
<symphorien>
on all executable files
<ldlework>
symphorien: ok, how do I add a python package to that interpreter?
* ldlework
prays you know the answer
<symphorien>
replace python in the buildinputs by python.withPackages(ps: [ps.foo])
tmaekawa has joined #nixos
<symphorien>
patchShebang will use what it finds on the PATH
<sphalerite>
teej: I'm home now so I can you get nix binaries working on your pi
<sphalerite>
teej: if you have an existing nix installation on another machine it's pretty simple to cross-compile a nix which you can then copy over to "bootstrap" it
<{^_^}>
[nixpkgs] @timokau merged pull request #46367 → Fixes the lldpd service not starting on NixOS 18.09 → https://git.io/fA2nW
<{^_^}>
[nixpkgs] @rembo10 opened pull request #46538 → sickgear: init at 0.17.5 → https://git.io/fArPs
<symphorien>
well it is, by python3Packages.buildPythonApplication
<symphorien>
in this specific case I don't know how to solve your problem
jmc has quit [Ping timeout: 252 seconds]
<exarkun1>
with `preUnpack = "set -x";` it still doesn't show me anything about patches. with --keep-failed, I see there is no Cargo.lock created anywhere in the source tree.
rprije has quit [Read error: Connection reset by peer]
<ldlework>
lol crap
<sphalerite>
clever: it seems CONFIG_KEXEC_FILE can only be set to y or n, not m :(
rprije has joined #nixos
<clever>
sphalerite: dang
<ldlework>
clever: a package takes python3Packages via callPackage and uses python3Packages.buildPythonApplication to provide the derivation. How do you provide an extra Python library to this package?
waleee has quit [Quit: WeeChat 2.2]
<xok>
I am using services.mysql
<adamantium>
clever: In nixos is there a way to command on-the-fly which pool/ROOT/dataset should be used for root from grub.cfg? e.g. from grub edit menu in gentoo, set kernel parameter root=zfs:zroot/ROOT/somedataset and somedataset would be used for the root filesystem.
<xok>
and especially: initialDatabases
<exarkun1>
nixpkgs is from 18.03 on this host, which of course does not have the `cargoPatches` feature.
<xok>
can anyone explain what path should I provide to it?..
<exarkun1>
(yay silent errors)
<xok>
the database doesn't get populated...
graphene has quit [Read error: Connection reset by peer]
<sphalerite>
xok: initialDatabases is only created on the first start of the mysql server I think
<sphalerite>
xok: maybe you want ensureDatabases
<xok>
sphalerite: well, I am aware of it and I do a complete remove and reinstall...
<xok>
including rm -rvf /var/db/mysql/
<xok>
and "ensureDatabases" just ensures there's a database...
<xok>
I want to import a dump into the newly created database...
<exarkun1>
when `nixops deploy ...` creates a new host, does it download necessary packages from the binary cache to the host where it's running and then upload them to the new host?
<exarkun1>
(signs point to yes :()
<ldlework>
that's my impression exarkun1
<exarkun1>
"It builds the NixOS machine configurations locally. For instance, here Valgrind is built or downloaded because our machine configuration has a dependency on it." crap
graphene has joined #nixos
<aminechikhaoui>
exarkun1: which backend are you using ?
<tobiasBora>
sphalerite: yes I was at work indeed. (now I'm back home) I'll try to ask the sys admin then. Or I heard that DHCP could provide a NTP server, is it possible to use it in nixos? I saw this issue, that is now quite old, and I'd like to know if you have some update: https://github.com/NixOS/nixpkgs/issues/27874
<{^_^}>
#27874 (by copumpkin, 1 year ago, open): Support configuring NTP from DHCP
<exarkun1>
What difference does it make? EC2 at the moment.
ogle has joined #nixos
<tobiasBora>
exarkun1: yes, it download everything on the "running host", and then upload to the new host.
<tobiasBora>
(at least that's what people told me here, and I had this same problem with a raspberry pi host, that didn't even had the same architecture)
<aminechikhaoui>
exarkun1: ec2 has deployment.hasFastConnection = false by default which means it will try to substiture first if things are in the binary cache
<aminechikhaoui>
substitute *
<ogle>
Anyone have experience maintaining a monorepo with nix? I would love to get some advice
<clever>
aminechikhaoui: but it will still want the entire closure on the nixops machine first
<tobiasBora>
exarkun1: in my case, the recommended solution was to ssh to my rasp, and then deploy from it... But I guess that you can't do that on EC2... So you may want to join me in this fight ;-) https://github.com/NixOS/nixops/issues/976
<{^_^}>
nixops#976 (by tobiasBora, 8 weeks ago, open): Add a way not to use the user's computer as a cache
<aminechikhaoui>
clever: right
<aminechikhaoui>
but the copying should be fast as it will only copy what's not in cache.nixos.org
trevthedev has quit [Ping timeout: 250 seconds]
matthewbauer has joined #nixos
<clever>
aminechikhaoui: it has to first download everything to the local machine, then have the remote machine download the same things from cache.nixos.org
<sphalerite>
Wondering about how to set up my backlight correctly — X doesn't seem to autodetect it (xbacklight says "No outputs have backlight property"), but I can control it via /sys/class/backlight/backlight/brightness. Anyone know how to do this?
sbdchd has quit [Remote host closed the connection]
matthewbauer has quit [Ping timeout: 240 seconds]
<infinisil>
ogle: Not sure what you mean
<aminechikhaoui>
clever: yeah, I personally don't see a way to prevent that with the current way nixops is designed though (nix-build then copy closure), not sure if there is a better way though
<exarkun1>
/nix/store/fjxk3k59n102sqs572kkspdjy303kv45-stdenv-linux/setup: line 1292: cargo: command not found :/
jonge has joined #nixos
<{^_^}>
[nixpkgs] @andir pushed commit from @marsam to release-18.03 « nodejs-6_x: 6.14.3 -> 6.14.4 »: https://git.io/fAryz
<jonge>
hey there, when i want to create a zip archive of some files in a nix expression... what would be the best way? the zip functions in hte lib do something else obviously
knupfer has quit [Quit: knupfer]
<vaibhavsagar>
jonge: that sounds like something you want to do in a bash script/body of a runCommand
<jonge>
vaibhavsagar: so that means i would execute the bash script in terms of its own derivation, right?
<ogle>
infinisil: my company has a large repo contaning several subprojects. some of these projects depend on each other for shared functionality. ultimately what I would like is for each project to have its own default.nix and at the top level directory release.nix that allows for building any one of the targets and toplevel shell.nix that provides an environment to hack across project boundaries for atomic
winem_ has quit [Ping timeout: 272 seconds]
<ogle>
refactoring.
matthewbauer has joined #nixos
mounty has joined #nixos
<exarkun1>
What is this "librustzcash-unstable-2017-03-17-vendor" drv/package that keeps showing up? I have no such thing defined in my nix expressions.
<symphorien>
it is an archive of the source code of all the dependencies of the rust package you build
<ogle>
vaibhavsagar: funny seeing you here! I was sad to hear you decided not to join our team awhile back, hope you are well.
<jonge>
vaibhavsagar: sry i did not realize that runCommand is a library function. thank you for pointing me there, this looks helpful!
<vaibhavsagar>
ogle: hi!
<infinisil>
ogle: I think you could maybe use lib.makeScope, you could have your subprojects nix files like `{ dep1, dep2 }: stdenv.mkDerivation ...`
<infinisil>
And callPackage them with the changed scope that includes all subprojects themselves
<infinisil>
So they can freely reference each other, similar to how nixpkgs works
<clever>
infinisil: this is an example of using makeScope + newScope + overrideScope
<vaibhavsagar>
ogle: hope you are well too! it was a really tough decision to make
<clever>
it allows getting a callPackage like you said, and also inserting overlays
<infinisil>
Neat, I never used it myself
<magnetop`>
I have a headless machine that installed libX11, so I ran "nix why-depends /run/current-system /nix/store/hash_libX11" it said it's because of dbus, iiuc: https://pastebin.com/uPD45Xcx
<magnetop`>
what gives?
<clever>
magnetop`: yeah, dbus-launch depends on libX11
<ogle>
infisil: interesting I havent come across makeScope before. going to take a look at the link from clever
<clever>
magnetop`: one of the things it does, is remove x11 support from dbus
cnidario has joined #nixos
humanoyd has quit [Quit: WeeChat 2.2]
gilligan_ has joined #nixos
<ldlework>
Python stuff is soooo complicated
<ogle>
vaibhavsagar: IIRC you ended up taking a haskell role, which definitely sounds up your alley.
<vaibhavsagar>
ogle: yup, I'm here now :)
<ldlework>
I don't suppose there is a way to figure out what part of your config is adding a given package to systemPackages
<clever>
ldlework: grep
<magnetop`>
clever: so am I understanding correctly that libX11 comes standard with any nixos box? and if I enable that option, I'll probably have to compile more pkgs,right?
<clever>
magnetop`: yeah
mounty has quit [Ping timeout: 252 seconds]
<magnetop`>
clever: OK, then I'll leave it be. I thought I got X cause I had installed the wrong pkgs or option. Thanks for the clarification! :)
<{^_^}>
[nixpkgs] @LnL7 merged pull request #46540 → perl-TestRunCmdLine: disable tests on darwin → https://git.io/fArXV
<clever>
ldlework: that looks like its in a nix-shell?
graphene has quit [Read error: Connection reset by peer]
<ldlework>
it isn't
<ldlework>
i think qtile is badly packaged
matthewbauer has quit [Ping timeout: 252 seconds]
<ldlework>
with bad practices
Mateon1 has joined #nixos
<clever>
that could be it
<ogle>
vaibhavsagar: congrats! if you are ever in SF would be happy to catch up
<ldlework>
I'm never going to figure out where this python3 is getting installed
<ldlework>
Or how to add libs to it
graphene has joined #nixos
<gilligan_>
huh.. konsole magically stopped working on 18.03 - also, holy *crap* konsole has an insanely huge closure size.. I get that there is some random kde stuff in there but spidermonkey and postgres.. x__X
<clever>
gilligan_: spidermonkey is part of consolekit
knupfer has joined #nixos
<symphorien>
gilligan_: you can use nix why-depends to figure out why
<gilligan_>
clever, i see.. what else do we have here.. cups, systemd, ..
<ldlework>
clever: when I cat /nix/store/6kz7wk44ad64w6ag15g0l2ansy3rv04y-my-qtile/bin/.qtile-wrapped
<ldlework>
it has a bunch of path stuff
<clever>
what about the non -wrapped one?
<ldlework>
so I guess since qtile is my window manager, all my terminal emulators and shell's inherit this stuff?
<clever>
yeah
matthewbauer has joined #nixos
<clever>
,runtimeDeps
<{^_^}>
In order of preference: Patch source OR ((if it uses PATH -> wrap with new $PATH) AND (if it uses dlopen, (patchelf --set-rpath in postFixup OR wrap with new LD_LIBRARY_PATH)))
<clever>
for a window manager, dont patch PATH, lol
<{^_^}>
[nixpkgs] @Ericson2314 pushed 4 commits to staging-18.09: https://git.io/fAr5c
sheyll has quit [Quit: sheyll]
<{^_^}>
[nixos-homepage] @globin opened pull request #239 → Support Page: Mayflower will also provide support → https://git.io/fAr5D
Edes has joined #nixos
lsyoyom has quit [Ping timeout: 245 seconds]
<gilligan_>
so I'm stracing konsole which doesn't start anymore and it repeatedly tries to read /proc/0/cwd ?! readlink("/proc/0/cwd", 0x7fffc42304c0, 4096) = -1 ENOENT (No such file or directory)
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
hamishmack has joined #nixos
hamishmack has quit [Client Quit]
ThatOtherGuy has quit [Remote host closed the connection]
<sphalerite>
does preferLocalBuild somehow override the effect of --max-jobs 0? some git fetching seems to be happening locally even though I passed --max-jobs 0 :|
cnidario has quit [Remote host closed the connection]
spear2 has quit [Remote host closed the connection]
Edes has quit [Ping timeout: 252 seconds]
<bigvalen>
Does anyone know how to change the resolution of the linux console ? I'm looking to set it at 640x480, but though I can do it in Grub, something resets it late in the boot process to a higher resolution.
matthewbauer has quit [Ping timeout: 240 seconds]
<samueldr>
bigvalen: just making sure, you want it AT 640x480?
<sphalerite>
bigvalen: I think if you pass nomodeset to the kernel it'll leave it as is
<bigvalen>
Yeah. I'm using HP's iLO, which seems to freak out if the console is 1024x786, which is the nixos default.
Edes has joined #nixos
<samueldr>
or an alternative could be video=640x480@60
<{^_^}>
[nixos-homepage] @domenkozar pushed commit from @globin to master « Mayflower will also provide support (#239) »: https://git.io/fArF0
<{^_^}>
[nixos-homepage] @domenkozar merged pull request #239 → Support Page: Mayflower will also provide support → https://git.io/fAr5D
<bigvalen>
I'll give it a go. I was using boot.loader.grub.gfxmodeBios="text" which sets it in the kernel & grub. But later on, during boot, it switches to a highres mode.
lsyoyom has joined #nixos
<samueldr>
yeah, that's probably modesetting there activating
<samueldr>
KMS, kernel modesetting*
<{^_^}>
[nixpkgs] @Ericson2314 opened pull request #46544 → glibc: Use lib.enableFeature and friends → https://git.io/fArF9
<{^_^}>
[nixpkgs] @xeji pushed commit from @r-ryantm to master « debootstrap: 1.0.107 -> 1.0.108 (#46294) »: https://git.io/fArAj
<sphalerite>
clever: I find nix-store -q --graph a lot easier to navigate than --tree
theunknownxy has quit [Quit: theunknownxy]
simukis has quit [Quit: simukis]
rihards has quit [Quit: rihards]
<ldlework>
OK so I have ranger installed, which provides Python 3 since it depends on it, and then I have qutebrowser which also depneds on python and somehow they depend on the same exactly python package? How is that possible if they use different Python libraries? https://gist.github.com/dustinlacewell/2226b9a87c3f35b40ad773832876701f
<{^_^}>
[nixpkgs] @Ericson2314 merged pull request #46544 → glibc: Use lib.enableFeature and friends → https://git.io/fArF9
georges-duperon has quit [Ping timeout: 252 seconds]
<magnetop`>
I have a VM without /proc/sys/kernel/modprobe, and it fails to activate because of that. any ideas?
Ariakenom has quit [Quit: Leaving]
<ldlework>
clever: if there is some package based on buildPythonApplication and it has a script who's shebang gets converted to /nix/store/hy65mn4wjswqih75gfr6g4q3xgqdm325-python3-3.6.6/bin/python3 that isn't nessecarilly enough to have that script be able to import various python libraries right?
<ldlework>
Because this python package is shared by ranger and qutebrowser, but they use different python libraries
<ldlework>
which means python libraries must be associated with a package through other means like wrappers using the python path and stuff
<ldlework>
which means that these plugin scripts _will never work_ because the only thing that gets done to them, is that their shebangs get set to a python interpreter in the store
<ldlework>
but that doesn't help them at all find the python libraries they depend on
<ldlework>
is that all right?
<ldlework>
Like I added tldextract to qutebrowser's propagatedBuiltInputs, and I can see it listed in nix-support/ but that doesn't mean some random python scripts with their shebang set to the parent package's python interpreter is going to mean that those scripts have access to the same libraries as qutebrowser
<ldlework>
I seem to confirm that by noticing that "cat /nix/store/sfbz2xc8sd05wpjk06n601ji7p3lyyyq-myqutebrowser/bin/.qutebrowser-wrapped | grep tld" shows me that tldextract is indeed being added for qutebrowser in this wrapper
<{^_^}>
[nixpkgs] @knedlsepp opened pull request #46551 → scs: Fix darwin build → https://git.io/fAoe8
graphene has joined #nixos
jperras has quit [Ping timeout: 252 seconds]
<hrk>
i am trying to run nixos (with nixops) on a system without a bootloader, but it seems it is not updating the /init symlink
<clever>
kalbasit[m]: i havent seen the content-addressed mode before, so i dont know how to fix it
<kalbasit[m]>
I see
<clever>
hrk: nixos doesnt use a /init symlink, you must obey the kernel cmdline that nixos put into the bootloader config
<hrk>
clever: there is no way to change the kernel cmdline
<hrk>
how does that work for example when running in a container?
<hrk>
(there is a init symlink)
<clever>
hrk: the nixos container runs it with a full storepath to the init script
<hrk>
the nixos install is the container tarrball
<jasongrossman>
ldlework: I use qutebrowser. Sorry I'm late to the party.
abueide has quit [Remote host closed the connection]
<jasongrossman>
ldlework: And I've been wondering the same thing. I don't know much about python environments though.
<ldlework>
lol damn it
<ldlework>
sphalerite: I can't really override the middle of a phase
<ldlework>
i guess I can just copy pasta the whole thing
<sphalerite>
ldlework: just stick it in preFixup if you're doing this only via overrideAttrs
<ldlework>
ok
tzemanovic has joined #nixos
nahamu has quit [Quit: WeeChat 1.9.1]
<ldlework>
sphalerite: hmm /nix/store/gpy3r9ss5ngfkib8ylx7jzgahq7m0x5b-patch-shebangs.sh: line 21: find: command not found
exarkun_ has quit [Read error: Connection reset by peer]
<ldlework>
maybe need to add to path rather than overriding it completely
<sphalerite>
ldlework: whoops looks like it relies on what's already in PATH. So maybe the same but prepend rather than replacing
<ldlework>
hehe ok
erasmas has quit [Quit: leaving]
gilligan_ has quit [Quit: Leaving]
<ldlework>
holy crap it works!
<ldlework>
sphalerite: good job!
<sphalerite>
\o/
rfold has quit [Ping timeout: 252 seconds]
<rawtaz>
hmm. gvim, not existing in Nix ?
<rawtaz>
also, question, does 18.03 use Wayland?
exarkun_ has joined #nixos
justanotheruser has joined #nixos
adetokunbo has joined #nixos
clever_ has joined #nixos
<ldlework>
ugh sphalerite i love you
<jasongrossman>
\o/
clever has quit [Ping timeout: 268 seconds]
clever_ has quit [Changing host]
clever_ has joined #nixos
<sphalerite>
rawtaz: yes it does exist, I think it's in vimHugeX
<rawtaz>
haha, what the.. why's that (the name, why not name it gvim)? :)
xenog has quit [Quit: Leaving]
<rawtaz>
works, thanks, that was it
<ldlework>
jasongrossman: rnhmjoj[m] is going to be submitting fixes to all of this
<ldlework>
hopefully to make it in the next release
<jasongrossman>
ldlework++
<{^_^}>
ldlework's karma got increased to 2
<ldlework>
rnhmjoj[m]++
<{^_^}>
rnhmjoj[m]'s karma got increased to 1
<ldlework>
sphalerite++
<{^_^}>
sphalerite's karma got increased to 16
<jasongrossman>
:-)
<hrk>
the symlink works now, but system activation complains about there not being module support in the kernel, is there some way to tell nixos not to try to load or unload kernel modules?
worldofpeace has joined #nixos
<kalbasit[m]>
clever_: restarting the build fixed it, interesting...
<sphalerite>
Is there a way to see the log of a running build started with `nix build`? the lines are getting truncated, and I don't want to restart the build with nix-build because it's a loooong one
<sphalerite>
the lines are getting truncated and I want to see them to compare progress between two concurrently-running copies of the build*
<kiloreux>
I am trying to install php extensions with php. But it seems like php interepreter just ignores they exist for some reason. Found this already, https://nixos.org/nix-dev/2015-September/018240.html But it didn't help.
<sphalerite>
kiloreux: php interpreter on the command line?
<kiloreux>
I meant php-fpm on my web app
<sphalerite>
kiloreux: could you paste the relevant bits of your config?
<sphalerite>
kiloreux: where you wrote it in your nixos config, instead of extension=/nix/var/nix/profiles/default/..., do extension=${phpPackages.yaml}/lib/php/extensions/yaml.so
<sphalerite>
infinisil: {^_^} misses out PR references if they have an anchrot in the url like what k albasit just wrote
<infinisil>
Yeah that's intensional
<sphalerite>
why is that?
jperras has joined #nixos
<infinisil>
Hmm.. I think I thought it linked a bit too often, people link to comments occasionally while discussing the pr itself