<kuznero>
I just created a simple nix package for serving markdown slides based on Reveal.js. I hope it can be useful for some: https://github.com/kuznero/mdslider I also hope to get some feedback on my package definitions :)
kreisys has joined #nixos
ma27 has quit [(Ping timeout: 255 seconds)]
hexamod has quit [(Ping timeout: 248 seconds)]
<pikajude>
hmmm
<pikajude>
why doesn't lua-mpack build on darwin anymore........hmmmm
jgertm has quit [(Ping timeout: 268 seconds)]
asuryawanshi has joined #nixos
<pie_>
i find it weird that its a modulenotfound error and not an importerror though
dieggsy has joined #nixos
dieggsy has quit [(Remote host closed the connection)]
asuryawanshi has quit [(Ping timeout: 240 seconds)]
<pie_>
ok i dont get how tihs can find python but not the packages
digitus has quit [(Quit: digitus)]
erasmas has quit [(Quit: leaving)]
hariel has quit [(Remote host closed the connection)]
kuznero has quit [(Quit: WeeChat 1.9.1)]
<pie_>
huh apparently it didnt like withPackages...
hariel has joined #nixos
<pie_>
ok i can now import while in nix-shell but it still fails at build time
<joepie91>
where would I find arm-linux-gnueabi-gcc ?
<joepie91>
is it packaged in nixpkgs at all?
hariel has quit [(Max SendQ exceeded)]
noobly has joined #nixos
<noobly>
Can one of you wise users tell me what's wrong with this: environment.systemPackages = with pkgs; [ ..... ];, when running nixos-rebuild switch, it says there an error with the "=", and it was expecting ";"
<noobly>
nvm found the missing ";" up much higher, I am not a programmer :-)
noobly has quit [(Quit: Page closed)]
slack1256 has quit [(Ping timeout: 268 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 1 new commit to master: https://git.io/vF1Mv
<NixOS_GitHub>
[nixpkgs] mekeor opened pull request #31722: mu: correctly install mu4e by moving it to emacs' load-path. (master...master) https://git.io/vF1Mu
NixOS_GitHub has left #nixos []
markus1189 has quit [(Ping timeout: 250 seconds)]
mekeor has quit [(Quit: digital revolution = cryptoanarchism + cybercommunism)]
asuryawanshi has quit [(Ping timeout: 268 seconds)]
ilyaigpetrov has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ericson2314 opened pull request #31723: stdenv setup: Run setup hooks and other processing after accumulating deps (staging...stdenv-accum-them-setup) https://git.io/vF1MM
NixOS_GitHub has left #nixos []
asuryawanshi has joined #nixos
jb55 has quit [(Ping timeout: 248 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 1 new commit to master: https://git.io/vF1D3
<fearlessKim[m]>
it seems like the kernel is compiled from my nix-shell with tons of stuff. Any advice on how to tune that down ? like a setting 'minimal-config' ? or should I disable manually stuff, like override the kernel config ?
Arcaelyx has quit [(Read error: Connection reset by peer)]
Arcaelyx has joined #nixos
lambdamu has joined #nixos
lambdamu_ has quit [(Ping timeout: 255 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 1 new commit to master: https://git.io/vF1QK
<dhess>
I think I'm going to pin the kernel on my zfs fileserver right now, before my next channel update :)
<hyper_ch>
adisbladis: nice
<hyper_ch>
and soon my encrypted dataset receiving issue should be resolved as well :)
<dhess>
hyper_ch: besides the send/receive issues, how is encrypted ZFS working so far for you?
<hyper_ch>
perfectly fine
<adisbladis>
dhess: Working out fine :)
<dhess>
cool. I think maybe I'll encrypt my offsite backup next pass but hold off on encrypting the production FS for a few more months :)
<hyper_ch>
even remote rebooting is working
<hyper_ch>
dhess: why only offsite backup?
<hyper_ch>
don't you encrypt everything?
<dhess>
hyper_ch: I've got terabytes of data and I want to make sure it's really battle-tested before I switch that FS
<adisbladis>
dhess: There is a breaking format change coming up but migrating should be pretty straight forward
asuryawanshi has joined #nixos
<dhess>
hyper_ch: I did before I switched to ZFS. When I switched to ZFS I traded off all of its benefits for lack of encryption at rest.
<hyper_ch>
migrating only takes space :)
<hyper_ch>
dhess: well, we had a breakin at out office 10 days ago... server, computers, external hdds, usb sticks etc.... all gone.... but I enforce encryption on everything
<hyper_ch>
and that would be some really, really sensitive information
<dhess>
hyper_ch: anything with personal data is encrypted already. This is just video.
<dhess>
but I would of course prefer to encrypt it anyway.
<hyper_ch>
most people seem to use luks/dm-crypt -> zfs
<dhess>
yeah I didn't bother with that since native encryption was in the works.
<hyper_ch>
also remote unlocking of encrypted root works fine
<dhess>
and I wanted ZFS to be working on raw sectors
<hyper_ch>
on the new server I did buy a small hdd for /boot
<hyper_ch>
so that I can use full hdds for zfs
<hyper_ch>
3x 8TB mirrored
<dhess>
yeah mine is more like... 2x6x8TB :) So I want to be extra careful.
asuryawanshi has quit [(Ping timeout: 260 seconds)]
<hyper_ch>
(also server got migrted from debian + proxmox + mdadm + luks/dm-crypt + ext4) to root encrypted zfs nixos
<hyper_ch>
I still need to get two things to run
<hyper_ch>
stupid java stuff
<hyper_ch>
just replayed incron with systemd.path now
vcunat has joined #nixos
<dhess>
hyper_ch: anyway, sorry to hear about the break-in. That sucks.
<hyper_ch>
it happens
<dhess>
Do you have any video cameras in the office that captured it?
<hyper_ch>
too bad they didn't also took a bunch of folders that were sent to me just the day before the breakin
<hyper_ch>
that would have been good for my client
<hyper_ch>
nope, no cameras in the office
<dhess>
ouch
MercurialAlchemi has joined #nixos
p4cman has joined #nixos
<hyper_ch>
unfortunately they didn't take that bunch of folders :(
<dhess>
I don't want to know any more :)
<hyper_ch>
it would have been beneficial to my client if those would have been taken as well :(
asuryawanshi has joined #nixos
asuryawanshi has quit [(Ping timeout: 250 seconds)]
matthewbauer has quit [(Ping timeout: 268 seconds)]
adisbladis has quit [(Changing host)]
adisbladis has joined #nixos
reinzelmann has joined #nixos
matthewbauer has joined #nixos
reinzelmann has quit [(Client Quit)]
reinzelmann has joined #nixos
leat has joined #nixos
fusion809 has joined #nixos
<fusion809>
Is there a command like nix-build -A package that after it's installed the package adds it to your user .nix-profile?
rogue_koder has quit [(Quit: Konversation terminated!)]
<fusion809>
Installed it based on a local nix (e.g. I want to build Atom 1.22.1, which is newer than that in the nixpkgs repo, locally and add it to my user nix profile)
<vcunat>
fusion809: you can nix-env -f . -iA package
<vcunat>
or even nix-env -i /nix/store/somePath
<fusion809>
Thanks, that worked perfectly!
Kim has quit [(Quit: WeeChat 1.9.1)]
Kim has joined #nixos
asuryawanshi has joined #nixos
asuryawanshi has quit [(Remote host closed the connection)]
asuryawanshi has joined #nixos
Itkovian has joined #nixos
<vcunat>
:-)
dmj` has quit [(Ping timeout: 252 seconds)]
a6a3uh has quit [(Quit: Bye)]
dmj` has joined #nixos
matthewbauer has quit [(Ping timeout: 248 seconds)]
matthewbauer has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] gnidorah opened pull request #31730: rambox: allow user to use other auth0 credentials (master...rambox) https://git.io/vFMUo
NixOS_GitHub has left #nixos []
a6a3uh has joined #nixos
fusion809 has quit [(Quit: Leaving)]
vcunat has quit [(Quit: Leaving.)]
miko__ has joined #nixos
page_ is now known as page
datakurre has quit [(Remote host closed the connection)]
datakurre has joined #nixos
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
aloiscochard has quit [(Quit: Connection closed for inactivity)]
<srhb>
Is adding nixos/release.nix as a jobset sane when you want to maintain a custom fork of nixpkgs including modules? I'm assuming hydra will use binary substitutes from upstream when available and just overlay whatever is necessary.
asuryawanshi has quit [(Remote host closed the connection)]
<srhb>
Though it will always have to rebuild failing jobs since they can't be cached upstream.
Itkovian has quit [(Ping timeout: 248 seconds)]
davidak has quit [(Quit: Leaving.)]
<joko>
srhb: I'd say I'm using lib/eval-config.nix that release.nix also uses. I don't want to build all the packages, test cases that release.nix does
davidak has joined #nixos
<srhb>
joko: You don't end up with a full binary cache of nixpkgs then though, do you?
<joko>
srhb: no, just the packages I want to. E.g. Firefox 57 still isn't in nixos-unstable (or at least yesterday), so I've pulled it on my custom nixpkgs and Hydra built it
<clever>
dhess: can you gist your current configuration.nix or nixops deployment and i can check it?
<srhb>
Do people have some other approach for locally mirroring nixpkgs (with changes?) If this works and just binary substitutes it seems very low maintenance, aside from having to rerun all the failing tests since they're not cached.
<dhess>
clever: yeah I can gist a facsimile of it
<dhess>
anyway I'm getting the same infinite recursion error that Infinisil did in that thread
cfricke has quit [(Client Quit)]
cfricke has joined #nixos
<clever>
dhess: one thing to note, is that you cant use the pkgs attribute when building the imports list
<clever>
because the value of the overrides in pkgs depends on the value of imports
<clever>
dhess: you may need to use (import <nixpkgs> {}).fetchFromGitHub to get around that problem
<dhess>
yeah I am doing that as you advised in the IRC log
<clever>
dhess: your adding this file to the imports list
<clever>
yeah
<dhess>
how do I get that
<clever>
systemd-digitalocean is a variable pointing to the git clone
<dhess>
right
<clever>
so just put this into imports, "${systemd-digitalocean}/module.nix"
<dhess>
ahh of course
<dhess>
it's just a path
<dhess>
ok cool
<clever>
the map should also be removed
<dhess>
right
<dhess>
that worked great, thanks!
<clever>
yep
<dhess>
now let's hope the service works and doesn't clobber my droplet
<sphalerite>
I can't seem to get back into my tmux session :( tmux a says "no sessions", even though ps reveals that there is a tmux running and it has a bunch of child processes which I started
Tucky has quit [(Remote host closed the connection)]
<clever>
sphalerite: ive sometimes seen screen break badly when it gets upgraded
<clever>
sphalerite: the screen client isnt compatible with the screen server
<clever>
but /proc/PID/exe lets me run the old version
<sphalerite>
clever: tmux complains more specifically about a protocol incompatibility if that's the case
<clever>
that magic symlink even works if the binary was deleted and replaced by an imperative package manager
<dhess>
it works! Whoot!
<clever>
screen can still type, but silently looses all output, leaving you blind and thinking it hung
<sphalerite>
clever: yeah, I know. But that doesn't see m to be what's causing the problem in this instance
<dhess>
now I can remove all this hard-wired IP cruft from the config. Excellent.
<clever>
sphalerite: what about stracing tmux, to see how it connects to the daemon
<clever>
and what its not finding
peacememories has joined #nixos
<sphalerite>
good idea
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
<clever>
oh, and strace on curses apps is messy
<clever>
strace -ff -o logfiles
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
erictapen has joined #nixos
ylwghst_ has joined #nixos
zzamboni has quit [(Quit: Leaving.)]
<sphalerite>
huh, it tries to connect to the socket but gets connection refused
<sphalerite>
then unlinks the socket..?
<clever>
unix socket?
p4cman has quit [(Quit: Connection closed for inactivity)]
<clever>
check /proc/PID/fd/ on the other end, to see what its listening on, and also the lsof -p PID output
peacememories has quit [(Client Quit)]
ylwghst has quit [(Ping timeout: 248 seconds)]
<sphalerite>
yeah it's listening on that path but presumably that's an old version of the socket?
zzamboni has joined #nixos
Tucky has joined #nixos
<clever>
that sounds plausible
<clever>
check the docs to see if a unix signal can make it reopen the socket?
goibhniu1 is now known as goibhniu
<clever>
systemd has a flag like that, to repair things after pivot_root
<sphalerite>
I just tried doing `tmux -S /proc/673/fd/6 attach` but that hasn't helped either
<sphalerite>
worst case I guess I can use gdb to get it to reopen the socket x)
zzamboni has quit [(Client Quit)]
<steveeJ>
can someone point out to me where stdenv.cc is defined? I'm trying to figure out how to change the name of the prefixed ar binary
cfricke has quit [(Quit: WeeChat 1.9.1)]
ylwghst_ has quit [(Ping timeout: 248 seconds)]
a6a3uh has quit [(Quit: Bye)]
zzamboni has joined #nixos
<sphalerite>
yeeees tmux accepts SIGUSR1 for restoring the socket
<sphalerite>
That did the trick! Thanks clever
<clever>
yep
ylwghst has joined #nixos
zzamboni has quit [(Client Quit)]
zzamboni has joined #nixos
zzamboni has quit [(Ping timeout: 246 seconds)]
kuznero has joined #nixos
<kuznero>
Hi All!
<kuznero>
Is there any way to install 32-bit packages on 64-bit NixOS?
<clever>
kuznero: yeah
<kuznero>
clever: do you have an example? I need xml2 and pam to patchelf libdb2.so :) Some mainframe stuff :D
<clever>
kuznero: nix-env -iA nixos.hello --argstr system i686-linux
<clever>
ah, nix-build if you just want the paths
<kuznero>
YES :) Thanks you, clever!
<kuznero>
s/Thanks/Thank/
<clever>
or import the <nixpkgs> { system = "i686-linux"; } variant
<clever>
there is also a pkgsi686 atrribute
<clever>
pkgsi686Linux
<clever>
and callPackage_i686
<kuznero>
excellent
<clever>
only problem with pkgsi686Linux and callPackage_i686, they force 32bit linux, even if used on darwin
<clever>
so installing something old like skype on darwin, gives you the 32bit linux binaries!
<kuznero>
what do you mean? It won'r force the whole system to be 32-bit all of a sudden, right?
<clever>
it wont force the whole system to change
<clever>
but it returns 32bit linux versions of paths for you to use
<clever>
and it always returns 32bit LINUX versions, even if you use pkgsi686Linux on a darwin machine
<clever>
wont be a problem if your only working on linux
a6a3uh has joined #nixos
aw has quit [(Quit: Quitting.)]
spacefrogg has quit [(Quit: Gone.)]
<kuznero>
I see, thanks
aw has joined #nixos
spacefrogg has joined #nixos
cfricke has joined #nixos
erictapen has quit [(Ping timeout: 268 seconds)]
<steveeJ>
bgamari: I can't find why the native toolchain in your cross-fix PR produces x86_64-unknown-gcc-ar instead of x86_64-unknown-ar. do you have any pointers? The rustc build bails out cause it can't find ar
kier has joined #nixos
matthewbauer has joined #nixos
<kuznero>
clever: interesting is that it requires `allosBroken = true` :)
roberth has quit [(Ping timeout: 250 seconds)]
seancjscherer[m] has left #nixos ["User left"]
ylwghst has quit [(Ping timeout: 246 seconds)]
ylwghst has joined #nixos
matthewbauer has quit [(Ping timeout: 268 seconds)]
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
erictapen has joined #nixos
<srhb>
Hrm, I have a lot of jobs from a jobset evaluating nixos/release.nix that remain queued. According to Hydra it's done evaluating. How do I figure out why it thinks it's done?
orivej has joined #nixos
cfricke has quit [(Quit: WeeChat 1.9.1)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dywedir opened pull request #31735: fd: 5.0.0 -> 6.0.0, install shell completions (master...fd) https://git.io/vFMCs
<NixOS_GitHub>
[nixpkgs] domenkozar force-pushed lib+paths from 7897128 to 05f5ab7: https://git.io/vFMBW
<NixOS_GitHub>
nixpkgs/lib+paths 05f5ab7 Domen Kožar: lib: add isInStore and isPath, slighly change isStorePath...
NixOS_GitHub has left #nixos []
dywedir has joined #nixos
<hyper_ch>
hmmm, I need to send emails from the cli (script) and also need to attach files... in the past I used heirloom-mailx.... what's recommended on nixos?
peacememories has joined #nixos
universalpackage has joined #nixos
<srhb>
LnL: Right, I'm just looking for the attribute to evaluate to build nars similar to what cache.nixos.org has
<LnL>
the store_uri setting
<LnL>
so it will put the narinfo/nar files in /mnt/nix-cache and you can serve that as a cache
<srhb>
That can't be right.
<srhb>
I'm looking for an answer that's essentially "evaluate pkgs/top-level/all-packages.nix" -- only correct. :P
<srhb>
Or "create a jobset that evaluates that."
universalpackage has quit [(Remote host closed the connection)]
<LnL>
not following
universalpackage has joined #nixos
<LnL>
are you looking for the jobset configuration?
<srhb>
Yeah.
<srhb>
Ah, I can actually see it on hydra.nixos.org
<srhb>
pkgs/top-level/release.nix
<srhb>
There we go :)
<LnL>
yeah, you can see it in the config tab
peacememories has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
matthewbauer has quit [(Read error: Connection reset by peer)]
peacememories has joined #nixos
a6a3uh has quit [(Quit: Hi)]
matthewbauer has joined #nixos
peacememories has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
universalpackage has quit [(Remote host closed the connection)]
avn has quit [(Remote host closed the connection)]
<clever>
sphalerite: yeah, the dedup is increasing the usage
peacememories has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<sphalerite>
yeah so definitely not worth using
ertes-w is now known as supercynic
hellrazor has quit [(Ping timeout: 268 seconds)]
<sphalerite>
but the question about having nix-store use filesystem-provided hashes still stands
<sphalerite>
nix-store --optimise*
<clever>
sphalerite: the main difference between a content addressed FS, and auto-optimize, is that nix will wait until the build is done and becoming immutable to hash things
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] zimbatm pushed 1 new commit to master: https://git.io/vFMEh
<NixOS_GitHub>
nixpkgs/master 97cc469 Jan Tojnar: sublime3: Use consistent program name (#31720)...
NixOS_GitHub has left #nixos []
<clever>
sphalerite: while a content-addressed FS doesnt know how long to wait
<sphalerite>
clever: yes, I get that. But that's not my question
<sphalerite>
could nix-store --optimise ask the filesystem for the hashes rather than reading all the files in and hashing them itself
<clever>
sphalerite: not all filesystems store the hash of the contents
<sphalerite>
yes, but for those that do
<clever>
sphalerite: and random writes to part of the file would require rehashing the entire thing
<kuznero>
How can I test if executable binary is 64 or 32 bit in NixOS?
<sphalerite>
I mean, even just to find equal files, then just use the regular hash
<sphalerite>
clever: what's that got to do with it?
<sphalerite>
kuznero: file path/to/executable
<clever>
sphalerite: if you wanted to get the hash over the entire file
<kuznero>
sphalerite: it is something I downloaded from internet.
<sphalerite>
kuznero: yes, just run the program `file` on it
<kuznero>
sphalerite: ah, thanks!
peacememories has joined #nixos
peacememories has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<makefu>
is it possible for channels to mark some packages as "has to build", because with the latest nixpkgs-unstable i had a lot of issues with qtwebkit. if i'd knew that the packages do not build i wouldn't have upgraded in first place
peacememories has joined #nixos
Cheng has quit [(Remote host closed the connection)]
matthewbauer has quit [(Remote host closed the connection)]
Cheng has joined #nixos
<hyper_ch>
hmmm, I need to send emails from the cli (script) and also need to attach files... in the past I used heirloom-mailx.... what's recommended on nixos?
matthewbauer has joined #nixos
peacememories has quit [(Client Quit)]
<steveeJ>
hyper_ch: is msmtp something that could help?
Neo-- has quit [(Ping timeout: 240 seconds)]
<hyper_ch>
steveeJ: I looked at ssmtp which works fine but doesn't seem to be able to send emails with attachments
peacememories has joined #nixos
<hyper_ch>
steveeJ: will test msmtp :)
orivej_ has joined #nixos
orivej has quit [(Ping timeout: 258 seconds)]
<hyper_ch>
steveeJ: msmtp doesn't provide mail or sendmail...
<sphalerite>
I've used msmtp to send emails without attachments which worked nicely FWIW
<icetan>
me too ^_^
peacememories has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
tommyangelo_ has quit [(Remote host closed the connection)]
tommyangelo_ has joined #nixos
<Unode>
is there any gvim package in nixos compiled with +python3 support? I'm running into configuration issues... neovim seems to only include +python3 whereas vim/gvim only +python
<icetan>
I'm having problems with nixops when deploying to virtualbox on non nixos machines
<icetan>
anyone else trying this?
universalpackage has quit [(Remote host closed the connection)]
<hyper_ch>
sphalerite: I need to send attachments - depending to which folder I do make a scan
<danbst[m]>
iceTanner: what's the error message?
<icetan>
danbst: right now I just keep getting "error: string index out of range"
<icetan>
but it's been a bit of a dice roll
<pie_>
hi guys, im trying t obuild krita 4.0 which adds python scripting support and im having some trouble in the configure phase, it finds python just fine but it cant find necessary packages: https://pastebin.com/31vBpHLh
<icetan>
I get the error when running `nixops deploy`
* pie_
pokes domenkozar (?)
<danbst[m]>
@ice
<icetan>
The VM gets created but not provisioned
<pie_>
err, the propagatednativebuildinputs was me basically just trying everything, i also tried propagatedbuildinputs
<danbst[m]>
iceTanner: so, it fails on evaluating your machine expression?
<icetan>
danbst: hmmm, maybe. I mean it parses it fine at least
<icetan>
danbst: running the trivial.nix example from the tutorial, same error
FRidh has joined #nixos
<hyper_ch>
mutt & msmtp work :) thx sphalerite steveeJ
<pie_>
domenkozar, nvm sorry for the ping, turns out i did something wrong while saving the fix yesterday, and i was in fact correct; the cmake files mangled pythonpath by replacing instead of appending
<sphalerite>
Anyone know stuff about videos? I've recorded something using gstreamer with vp8enc and webmmux, and the length on the resulting video file seems to be all wrong and kdenlive doesn't like it. Any ideas what to do about this? (sorry OT)
<pie_>
what does vlc say on the codec info window?
ylwghst has quit [(Ping timeout: 240 seconds)]
<pie_>
ffmpeg/avenc or whatever its calle probably also has something for getting info but i have no idea how to use it :P
<adisbladis>
sphalerite: You might try re-muxing the streams with ffmpeg?
BioSpider has quit [(Read error: Connection reset by peer)]
matthewbauer has quit [(Remote host closed the connection)]
FRidh has quit [(Read error: Connection reset by peer)]
<sphalerite>
pie_: that it's VP8
<sphalerite>
adisbladis: just the one stream, no audio
matthewbauer has joined #nixos
BioSpider has joined #nixos
hellrazor has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] periklis opened pull request #31737: WIP docker: enable darwin support (master...init-docker-darwin-support) https://git.io/vFMrt
NixOS_GitHub has left #nixos []
<adisbladis>
sphalerite: Ahh check. Still worth trying to create a new container. Just do 'ffmpeg -i file.whatever -vcodec copy -an out.whatever'
kuznero has joined #nixos
<sphalerite>
yep that's better
<sphalerite>
why didn't gstreamer just get it right? :(
<sphalerite>
thanks!
<danbst[m]>
iceTanner: does nixops deploy --build-only works for you?
<kuznero>
When I `patchelf --set-rpath` a binary with one of the paths that is relative to current binary. It will also be packaged together with that executable binary. During build phase ldd works great showing that it resolved local so file, but when ldd from /nix/store/... it fails. I guess this is expected - how could I patchelf with absolute path for l
<kuznero>
ocal so file assuming it is not yet installed?
<clever>
kuznero: in the fixup phase is a call to patchelf --shrink-rpath
<clever>
kuznero: which runs ldd, and deletes any rpath entires that appear unused
<icetan>
and I had to manually set the network adapter of the VM to vboxnet0
<icetan>
I'm not sure what I'm doing wrong, I've got this running although a bit shakely as well before on virtualbox when messing about with hydra
Cheng has quit [(Quit: Leaving)]
<icetan>
would be soo nice if I could get this working well, setting up a new project and I'm trying to convince my team that nixops is the way to go for deployment
<mbrock>
is there a way to escape so that bash syntax like ${FOO?} works inside a ''string''?
<mbrock>
maybe adding Nix escaping to the already confusing Bash syntax is a true horror
<icetan>
mbrock (IRC): ''some bash scripts ''${BASH_ENV_VAR} more stuff''
<mbrock>
whoa, ok
<adamt>
icetan: Aren't you missing a set of '' after the }?
<adisbladis>
mbrock: Wouldn't it just be ''\${FOO?}''
<icetan>
adamt: nope, it confused me too first
<icetan>
''$ is the escape for $
<adamt>
Ah. That's kinda horrible. :-)
<icetan>
well, makes sense
<mbrock>
yeah it works, hehe thanks
jtojnar has quit [(Remote host closed the connection)]
<icetan>
you can't but a $ aftar the end of your string any way
<icetan>
s/but/put
<adisbladis>
I like the single backslash more. Much less weirdness.
<icetan>
adisbladis: you quickly go down backslash hell when it's the same esacape in nix as in bash
Jackneillll has joined #nixos
<mbrock>
adisbladis: if only it worked!
<icetan>
think it's better this way
<icetan>
had the same reaction as you at first
Jackneilll has quit [(Ping timeout: 240 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 1 new commit to master: https://git.io/vFM1d
patrl has quit [(Read error: Connection reset by peer)]
<pie_>
while running cmake: sip: Unable to find file "QtCore/QtCoremod.sip" results in error: build of ‘/nix/store/45r91p7m2flmvxkn6lgdbxywp3ikln30-krita-4.0.0-pre2c.drv’ failed
<pie_>
that sip file is supposed to be contained in pyqt5, which i have included (and cmake finds) in propagateBuildInputs
<pie_>
do i need to use some other kind of buildinputs or something?
<pie_>
unless im misunderstanding what youre saying
ravloony has joined #nixos
<clever>
yeah, that does sound like a bug
<pie_>
nproc returns 4 when i run it
<clever>
i think the problem, is that the cores option is handled by nix-build/nix-daemon, and nix-shell and nix-instantiate arent dealing with it
lavironx has joined #nixos
<clever>
sounds like an issue that should be opened on nix
cfricke has quit [(Quit: WeeChat 1.9.1)]
<pie_>
ok ill open an issue
<joko>
clever: could you give me some info on cross compiling for arm systems, ideally on a hydra farm? I'm planning to install NixOS on my Wandboard this weekend
tommyangelo_ has quit [(Ping timeout: 268 seconds)]
<clever>
joko: the cross-compile stuff in nixpkgs was still a bit wonky last time i looked at it, and doesnt support enough to build a full NixOS, and has more issues down the road when using nix-env
<clever>
joko: so its simpler to always do a "native" arm->arm build, but that could itself be done under emulation
<mbrock>
I'm playing with multiple output derivations to store some ancillary scripts, but I'm confused about how to programmatically find a non-default output of something I build e.g. via nix-build
<pie_>
clever, setting the variable manually works but now the build is complaining :/
<pie_>
make: the '-j' option requires a positive integer argument
<mbrock>
ah yeah I mean finding the store path
<clever>
mbrock: bin is the default output for "${lzma}" because its first
<pie_>
oh right of course, im supposed to set it to the correct amount, not 0, yes?
<clever>
pie_: setup.sh will turn 0 into a number based on your cores, during the startup
<clever>
but manually setting it happens after that startup
<clever>
you might be able to set it before nix-shell is even ran
<pie_>
ok setting it to 4 worked :)
<pie_>
buildPhase runs now
<clever>
pie_: make also has a "feature" where just -j with no number, runs infinite jobs!
<pie_>
clever, ah yeah going by the source that should maybe work because it checks if it exists
<pie_>
i suppose id have to clear the var for that?
<pie_>
well time to wait for an hour long build to break again...
<pie_>
i should have done this from nix-shell the first time so i dont have to wait for the whole thing
<clever>
pie_: also keep in mind that setting it to an empty string, and unset have different mechanics
<pie_>
yeah
<joko>
clever: is x86 emulating arm faster than my barebone pc?
<clever>
joko: its somewhat slower then a normal arm machine, but you have more cores, and can throw more x86 machines at it, which can balance out
<joko>
clever: should I try with 2 single-core machines or 1 dual-core?
<clever>
joko: more cores and more machines will always help, i believe
<joko>
clever: thanks, I will do some experiments and tell you my results :D
<clever>
joko: to use the stuff in that github, you need to nix-build the qemu and run the register script in result/bin/
<clever>
joko: then the system can just run any arm binary
<joko>
clever: so, I have to use qemu-user-arm somehow or what?
<joko>
ah, ok
<clever>
joko: applying that nix override to the nix used by the host (via configuration.nix nix.package) will add support for a new nix.conf flag
<clever>
joko: and then with the right nix.extraConfig, you can convince nix that the machine supports arm binaries (read the patch for the config name)
<joko>
clever: just saw it, yeah
<joko>
So that I can use it as a buildMachine on my wandboard, right?
<clever>
yeah
Neo-- has joined #nixos
<joko>
cool :D
<clever>
you can turn every spare x86 machine into an arm build slave
ma27 has joined #nixos
<clever>
and then either point the weak arm board at them, or point a hydra at them to pre-build
<joko>
clever: what I don't get is where you define the target platform in qemu
<gchristensen>
not the least of which since rust requires the previous version of rust to build the new version of rust
<LnL>
and then we can also have ff use the non default version on stable if neccecary
<catern>
it is this kind of thing that shows how badly Nix is needed - otherwise people think libraries have to be hard!
<gchristensen>
+1 +1
ssmike has quit [(Remote host closed the connection)]
ssmike has joined #nixos
<pie_>
clever, could the cores issue have anything to do with me using fish instead of bash/
<clever>
pie_: nix-shell always runs bash on the inner shell
<pie_>
ok, i wasnt sure about that but i hoped thats the case
<pie_>
i didnt see a #! line so i figured thats what should be happening
<pie_>
otherwise id expect a lot more to be broken
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jyp closed pull request #29920: Revert "cudatoolkit: Move libcudart to a separate output" (master...fix-tensorflow-runtime) https://git.io/vd3Ar
NixOS_GitHub has left #nixos []
<joko>
Anyone familiar with Hydra? 2 questions: I would like to create a job which runs a simple shell script and then I would like to run this job only if a previous one has succeed. I think for the latter I could use pkgs.releaseTools.aggregate, but I would feel more confident if I saw how others are using it
BioSpider has quit [(Ping timeout: 246 seconds)]
orivej has joined #nixos
<clever>
joko: running a command is easy, just pkgs.runCommand, but it will be in a sandbox and cant do too much "fun" stuff, and everything is pure, so it will never know the previous outcome
<joko>
clever: then perhaps have a shell.nix and run the command inside a nix-shell?>
<clever>
hydra only ever uses nix-build
<gchristensen>
also "only if the previous one succeeded" means it can never start, because there is no previous one ;)
<larsvm>
recursion to the rescue
<joko>
ditto
<joko>
So I guess there are no such patterns in Hydra jobs for the time being?
<gchristensen>
can you be more specific about what you actually want to do?
<larsvm>
joko: offtopic: your nick means in finnish "already?" :D
<joko>
gchristensen: sure, I'm building my own nixos channel ( https://git.joko.gr/joko/etc-nixos/src/master/jobs.nix#L48 ) and I would like to do so only if the other jobs (nixos.config.system.build.toplevel mainly for a couple of systems) succeed
<joko>
larsvm: lol, I didn't know :D
<clever>
joko: nixos handles that with an external script
<clever>
joko: it just always builds the channel tarball, but only if the tests pass will that be copied to the website
<DIzFer[m]>
knowing nothing about how hydra works... isn't what joko wants to have the second derivation depend on the first?
<joko>
clever, gchristensen: and afterwards I want to copy the channel to my site
<gchristensen>
for that you'll need a thing outside of hydra
<clever>
joko: yeah, lines 44-54 control what hydra jobset each channel should wait on
<clever>
joko: and only if that jobset passes, will it update the channel
<clever>
joko: line 26 runs mirror-nixos-branch, and gives it the channel name and a link to that hydra job
<joko>
Got it
Neo-- has quit [(Ping timeout: 268 seconds)]
<pie_>
is it possible to contiue a build instead of restarting?
<pie_>
nvm, it was only redoing the beginning
<joko>
And one more question, regarding binary cache this time... I was checking how to setup a binary cache and after seeing zimbatm's presentation I noticed this: https://github.com/zimbatm/todomvc-nix/blob/master/scripts/nix-cloud-push. In practice one can do a simple rsync as well, right? But how the cache is cleaned from older generations?
jasom has quit [(Ping timeout: 240 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 closed pull request #31705: android-udev-rules: use proper path for udev/rules.d (master...android-udev-rules) https://git.io/vFXr9
NixOS_GitHub has left #nixos []
<clever>
joko: there is a hydra-s3-backup-collect-garbage script in hydra
<joko>
clever: I would like the same functionality for normal hosts, hosts running nginx and ssh
sigmundv__ has joined #nixos
<joko>
as the binary cache is no longer hydra
playX has joined #nixos
<playX>
Hello, i want to install firefox 57, i copied default nix of firefox 56 from github and edited it for use tar ball, but i get problem https://pastebin.com/qThmzkSR It's ask to add gconf in red only directory or add to default.nix, but when i add to default nix it says duplicate formal aplication argument
ssmike has quit [(Remote host closed the connection)]
universalpackage has quit [(Remote host closed the connection)]
universalpackage has joined #nixos
tgunb has quit [(Ping timeout: 240 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej closed pull request #23233: nixos/virtualisation/gcp: Use lib/make-disk-image for building Google… (master...gcp-make-image) https://git.io/vyIub
NixOS_GitHub has left #nixos []
universalpackage has quit [(Remote host closed the connection)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ericson2314 pushed 2 new commits to staging: https://git.io/vFDJf
<NixOS_GitHub>
nixpkgs/staging 6a5cda5 John Ericson: stdenv setup: Run setup hooks and other processing after accumulating deps...
<NixOS_GitHub>
nixpkgs/staging b26038f John Ericson: Merge pull request #31723 from obsidiansystems/stdenv-accum-them-setup...
NixOS_GitHub has left #nixos []
universalpackage has joined #nixos
ssmike has quit [(Ping timeout: 264 seconds)]
adamt has joined #nixos
adamt has quit [(Changing host)]
adamt has joined #nixos
sigmundv has joined #nixos
universalpackage has quit [(Read error: Connection reset by peer)]
sigmundv has quit [(Max SendQ exceeded)]
sigmundv has joined #nixos
universalpackage has joined #nixos
adamt has quit [(Ping timeout: 264 seconds)]
<hyper_ch>
goibhniu: can you create containers also in other locations than /var/lib/containers?
universalpackage has quit [(Remote host closed the connection)]
<goibhniu>
hyper_ch: I don't think so, but you can bind mount them wherever you like
<hyper_ch>
goibhniu: it's ok, I just make a zfs dataset to it
<goibhniu>
hyper_ch: I bind mount them to a path that's the same inside and outside the container, which can be handy
<goibhniu>
(I don't think I mentioned that in the screencast)
<hyper_ch>
goibhniu: I'll still have to see if I can get thet java app to run
cfricke has joined #nixos
hellrazor has joined #nixos
jb55 has quit [(Ping timeout: 240 seconds)]
<sphalerite>
goibhniu: tried using kdenlive today, the results were disastrous 😱
universalpackage has quit [(Remote host closed the connection)]
<goibhniu>
sphalerite: oh?
universalpackage has joined #nixos
ma27 has quit [(Ping timeout: 240 seconds)]
ma27 has joined #nixos
<hyper_ch>
goibhniu: ups... I did run nixos-container create in a vm instead of the host... can can I undo all that nixos-container install stuff?
<goibhniu>
hyper_ch: destroy?
<hyper_ch>
when I run nixos-container it did download and install some nix packages... how to remove those?
<goibhniu>
sphalerite: did it run? It might need some wrapper thing if you're not using plasma/kde.
ylwghst has quit [(Ping timeout: 248 seconds)]
<goibhniu>
hyper_ch: does destroy not do the trick?
iyzsong has quit [(Ping timeout: 264 seconds)]
<hyper_ch>
goibhniu: it destroyed the container
<hyper_ch>
but not sure if everything was removed... don't worry about it
<sphalerite>
goibhniu: it worked fine for a while, then it crashed and now it eats all my memory when I try to recover the project :(
<sphalerite>
it seems to be accessing the same file over and over
<goibhniu>
sphalerite: oh, that's no fun :/
adisblad` has joined #nixos
ylwghst has joined #nixos
<sphalerite>
I appreciate an autosave feature that lets me recover my work if something goes wrong… not so much if it doesn't actually work xD
adisbladis has quit [(Ping timeout: 240 seconds)]
vaibhavsagar has quit [(Ping timeout: 240 seconds)]
hellrazor has quit [(Ping timeout: 255 seconds)]
<hyper_ch>
sphalerite: zfs snapshot? :)
<goibhniu>
sphalerite: I noticed someone updated shotcut recently, that might be a good alternative
adisblad` is now known as adisbladis
adisbladis has quit [(Changing host)]
adisbladis has joined #nixos
<sphalerite>
hyper_ch: I don't have autosnapshots on my laptop, not enough space D:
vaibhavsagar has joined #nixos
hellrazor has joined #nixos
<hyper_ch>
sphalerite: time to ugprade to a Samsung 960 Pro 2TB ;)
universalpackage has quit [(Remote host closed the connection)]
Tucky has quit [(Remote host closed the connection)]
<sphalerite>
goibhniu: started trying with pitivi earlier as well, the UI isn't as good as kdenlive's (typical GNOME oversimplification). I'll give shotcut a shot!
<goibhniu>
:D
<sphalerite>
goibhniu: but this seems to be a common theme across all my attempts at editing video in the past 10 years or so: the software tends to be so unreliable that I can barely get a small video project done because it crashes
<sphalerite>
I didn't use pitivi long enough for it to crash actually, but I suspect it would have done so anyway :p
<sphalerite>
hmm
<sphalerite>
error: Package ‘shotcut-17.02’ in /nix/store/03kpq0vgrv1vj52lx44lbq07zgiypaaw-nixos-17.09.2034.78eed74497/nixos/pkgs/applications/video/shotcut/default.nix:37 is marked as broken, refusing to evaluate.
* sphalerite
tries nixos-unstable
<goibhniu>
That's terrible. I've found kdenlive to be very stable.
<goibhniu>
what video format?
<adisbladis>
Heh. I always end up doing any editing with plain ffmpeg from the shell. My needs are usually very simple.
<goibhniu>
yeah, there's a newer one than that, I'm not sure if it has made it into the channel yet, but you could grab it straight from hydra
<sphalerite>
goibhniu: vp8 in webm, and some static PNG images
<NixOS_GitHub>
[nixpkgs] ttuegel opened pull request #31743: wrapFirefox: make overridable, support browserpass (master...firefox/wrapper) https://git.io/vFDI4
NixOS_GitHub has left #nixos []
<goibhniu>
gah, you'd expect that to be fine :/
<sphalerite>
yeah that's what I got from nixos-unstable
<goibhniu>
nice!
Itkovian has joined #nixos
universalpackage has joined #nixos
rogue_koder has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 5 new commits to master: https://git.io/vFDIP
<NixOS_GitHub>
nixpkgs/master dc8857e Orivej Desh: clickhouse: remove boost version override
<NixOS_GitHub>
nixpkgs/master 0b87f3d Orivej Desh: clementine: remove boost version override
<NixOS_GitHub>
nixpkgs/master ee7c81f Orivej Desh: hhvm: remove boost version override...
NixOS_GitHub has left #nixos []
<edef>
dhess: glad to hear it's working for you [=
Itkovian_ has joined #nixos
Itkovian has quit [(Read error: Connection reset by peer)]
Itkovian_ has quit [(Client Quit)]
universalpackage has quit [(Remote host closed the connection)]
<hyper_ch>
goibhniu: just tried to nix-channel --update in the container but says unabel to check
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 1 new commit to master: https://git.io/vFDLX
<NixOS_GitHub>
nixpkgs/master a5df25c Orivej Desh: scantailor: remove boost version override
NixOS_GitHub has left #nixos []
<goibhniu>
hyper_ch: did you add a channel and all that?
<sphalerite>
goibhniu: shotcut looks neat, UI is a bit confusing but it's working… so far ;)
<goibhniu>
you should write a nasty comment on youtube
sigmundv has quit [(Ping timeout: 240 seconds)]
<hyper_ch>
will that interfere with the bridging?
tommyangelo_ has joined #nixos
universalpackage has quit [(Remote host closed the connection)]
<goibhniu>
sorry, I don't know so much about networking
universalpackage has joined #nixos
ravloony has quit [(Ping timeout: 240 seconds)]
<hyper_ch>
things still seem to run
universalpackage has quit [(Remote host closed the connection)]
universalpackage has joined #nixos
erasmas has joined #nixos
<hyper_ch>
goibhniu: yeah, can't get it to work :(
<pie_>
anyone here know who i should ask about pyqt/sip things?
<pie_>
ping FRidh ?
Turion has joined #nixos
<Turion>
Since I've switched to 17.09, enigmail (thunderbird GPG plugin) throws strange errors and complains about "end of file" when I want to encrypt, decrypt or sign
<mpickering>
If I install a haskell package by "nix-shell -p haskellPackages.pkg" then any executables it builds should be available on the path in the shell?
seppellll has quit [(Ping timeout: 255 seconds)]
<sphalerite>
goibhniu: the project from earlier crashes shortcut just as effectively as kdenlive :D
<goibhniu>
amazing
Itkovian has joined #nixos
<sphalerite>
aaaaaaah
<sphalerite>
renaming the file fixed it
<sphalerite>
apparently it had a reference to itself in it, which renaming it broke
<sphalerite>
I think it was trying to load the file recursively or something
<sphalerite>
Not entirely convinced of MLT's greatness though :p
<goibhniu>
eep
Itkovian has quit [(Read error: Connection reset by peer)]
_ts_ has joined #nixos
<_ts_>
hi, does anyone have a nixos module for kanboard? (@fpletz @shlevy as you've had some commits related to kanboard)
universalpackage has quit [(Remote host closed the connection)]
universalpackage has joined #nixos
a6a3uh has joined #nixos
<mpickering>
I think I'm completely misunderstanding how nix-shell works
<hyper_ch>
Turion: I've had issues with gpg in qtpass in 17.09 but current unstable small all is fine again
<mpickering>
If I have python in my buildDepends then in the shell I should be able to run python right?
universalpackage has quit [(Excess Flood)]
ma27 has quit [(Ping timeout: 255 seconds)]
universalpackage has joined #nixos
ma27 has joined #nixos
<Turion>
hyper_ch, alright, I'll try and update then!
<hyper_ch>
Turion: not sure if that actually will help you
<mpickering>
alright I had buildDepends rather than buildInputs
<hyper_ch>
Turion: the good thing is, if you upgrade, you can revert back ;)
adamt has joined #nixos
adamt has quit [(Changing host)]
adamt has joined #nixos
earldoug1as has joined #nixos
universalpackage has quit [(Remote host closed the connection)]
universalpackage has joined #nixos
<Turion>
hyper_ch: Yes, or I could revert even one step further to when it still worked on 17.03. But I don't know whether my gpg database might be messed up then :/
<hyper_ch>
Turion: for that reason you have backups as well, right?
pie_ has quit [(Remote host closed the connection)]
pie_ has joined #nixos
<Turion>
hyper_ch: Yes and no. I'm not used to gpg actually messing something up. That's a new experience brought to me by nixos... I coincidentally do have my keys on my phone. But gpg doesn't even manage to import them from there. I'll just have to hope for the update.
ertes-w has quit [(Ping timeout: 255 seconds)]
<Turion>
Does anyone else have problems with konsole not accepting UP/DOWN/etc. keys?
universalpackage has quit [(Remote host closed the connection)]
<Turion>
Or the standard X terminal not accepting certain upper case letters (X, W, ...)?
ma27 has quit [(Ping timeout: 252 seconds)]
seppellll has joined #nixos
roberth has quit [(Ping timeout: 248 seconds)]
<sphalerite>
Turion: the standard x terminal? Do you mean xterm?
<adamt>
Where about would I go if I wanted to hack on the installers autodetection features? I want it to add a default boot.kernelParams when running on vmware. :-)
<Turion>
sphalerite, yes
jensens has quit [(Ping timeout: 240 seconds)]
universalpackage has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 1 new commit to master: https://git.io/vFDGI
<Turion>
I have that problem since installing nixos (one or two months back?), but didn't bother doing anything about it since there were too many other issues. I should maybe add that I'm using the NEO keyboard layout. But I can't see how it could be an issue of the keyboard layout, since it worked fine in konsole (until I upgraded to 17.09)
seppellll has quit [(Ping timeout: 240 seconds)]
python47` has joined #nixos
davidak has quit [(Ping timeout: 250 seconds)]
yegortimoshenko has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] elitak opened pull request #31745: fix ACME TOS hash mismatch (master...simple) https://git.io/vFDZe
NixOS_GitHub has left #nixos []
<yegortimoshenko>
is there an easy way to override an attribute in some scope in nixpkgs overlay so that all other packages in this scope also use the overrided package?
<yegortimoshenko>
say i want to patch xfce.garcon
<yegortimoshenko>
it is referred to by a bunch of other packages in xfce attrset (scope)
<clever>
yegortimoshenko: all of those packages use a new instance of callPackage, and will search within the deps // xfce_self set
MP2E has quit [(Read error: Connection reset by peer)]
MP2E has joined #nixos
universalpackage has quit [(Remote host closed the connection)]
<clever>
yegortimoshenko: and yeah, there is no way to override values within xfce_self, so your only option is to .override everything you want to impact
universalpackage has joined #nixos
FRidh has quit [(Quit: Konversation terminated!)]
kalhauge_ has quit [(Ping timeout: 248 seconds)]
universalpackage has quit [(Remote host closed the connection)]
zzamboni has quit [(Remote host closed the connection)]
ylwghst has joined #nixos
srdqty has quit [(Quit: WeeChat 1.9.1)]
<greymalkin>
Nope, that doesn't do it.
universalpackage has quit [(Remote host closed the connection)]
universalpackage has joined #nixos
jgertm has joined #nixos
<orivej>
what is the error that you see?
tmaekawa has joined #nixos
seppellll has joined #nixos
tmaekawa has quit [(Client Quit)]
<greymalkin>
KeyError: 'getpwuid(): uid not found: 1003'
fragamus has joined #nixos
<greymalkin>
It appears to be the correct issue, but the fix of removing 'compat' from nsswitch.conf didn't help.
<ldlework>
greymalkin where did you get your name
<greymalkin>
Ah. it's because my user is in 'compat' -- the OS's `whoami` stops working when I remove 'compat' from nsswitch.conf
<greymalkin>
ldlework: Shakespeare
universalpackage has quit [(Remote host closed the connection)]
universalpackage has joined #nixos
universalpackage has quit [(Remote host closed the connection)]
<pie_>
anyone know why python cant access the packages if i do this? propagatedBuildInputs = with python36.withPackages(ps: [ps.sip ps.pyqt5]); [];
universalpackage has joined #nixos
<vcunat>
pie_: that only specifies an empty list.
<vcunat>
`with` keyword brings stuff into scope but doesn't add it anywhere by itself
<pie_>
ok i kind of assumed the with is some kind of magic thing that adds it but right with is a scoping thing
<pie_>
i dabble in functional programming but my mind just shuts down with nix x)
<pie_>
so can i do python = ...; [python]; ?
<pie_>
* with python = ...
<Turion>
The gpg problem seems to be related to pinentry. When I do "gpg --decrypt --pinentry-mode ask", it complains about "empty file" (referring to the passphrase), but "-pinentry-mode loopback" asks for the password on the console alright
<Turion>
But pinentry is clearly installed, and I've configured the gpg-agent to look for pinentry in the correct place
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to release-17.09: https://git.io/vFDuJ
<NixOS_GitHub>
nixpkgs/release-17.09 17d60d2 Tim Steinbach: linux-copperhead: 4.13.12.a -> 4.13.13.a...
NixOS_GitHub has left #nixos []
tobiasBora has quit [(Quit: WeeChat 1.6)]
<Turion>
Could the problem be that thunderbird is built without a pinentry dependency?
<vcunat>
pie_: you probably mean let python = something; in [ python ];
<vcunat>
Turion: I had a problem with gpg keys that have an empty passphrase
<Turion>
vcunat: Did you solve it? My key doesn't have an empty passphrase though
universalpackage has quit [(Remote host closed the connection)]
<vcunat>
It doesn't work on my 17.09 machine but now I see it does on my 18.03 machine
<vcunat>
but those with passphrase work on both
<Turion>
Are the pinentry or gpg-agent versions different?
universalpackage has joined #nixos
<Turion>
vcunat: What's your gpg.conf and gpg-agent.conf like? Do you specify pinentry's path?
<vcunat>
no, I don't
<vcunat>
I don't even have it on $PATH, apparently
<vcunat>
gpg (GnuPG) 2.2.1
<Turion>
vcunat: Yeah, that's the solution! For some reason, specifying the executable's path wasn't what gpg-agent liked
<Turion>
Thanks a lot
<vcunat>
:-)
rogue_koder has quit [(Quit: Konversation terminated!)]
sigmundv__ has quit [(Ping timeout: 268 seconds)]
sigmundv has joined #nixos
rogue_koder has joined #nixos
ixxie has joined #nixos
Thra11 has quit [(Quit: IRC for Sailfish 0.9)]
darlan has joined #nixos
nixer has joined #nixos
darlan has quit [(Client Quit)]
cfricke has joined #nixos
<nixer>
Has anyone got js/webpack to work within nix? I'm aware of node2nix.
<joepie91>
nixer: can you explain more about what you're trying to accomplish with webpack and how that relates to Nix?
universalpackage has quit [(Remote host closed the connection)]
<nixer>
joepie91: I'm trying to get my custom javascript app bundled with webpack within a nix derivation.
universalpackage has joined #nixos
<nixer>
I'm using node2nix to resolve dependencies, but running into trouble with getting webpack to use those dependencies.
<joepie91>
nixer: my node_modules path for a node2nix-deployed Node.js application is /nix/store/7chf9fjzirx32kshrs2i83mqwwkcpb4w-node-pastebin-stream-1.0.0/lib/node_modules/pastebin-stream/node_modules
<joepie91>
the application code itself exists in the one-level-higher lib/ folder
universalpackage has quit [(Remote host closed the connection)]
<joepie91>
so *in theory* once you have a node2nix-built application, you should be able to point webpack at your entry point as normal, and have it produce a bundle
<joepie91>
since it will search for node_modules relative to the entry point, and that's all still in the same place
universalpackage has joined #nixos
<joepie91>
(the project I linked doesn't build the browser bundle from Nix, I do that locally on my dev system)
<joepie91>
oops
<joepie91>
the application code itself exists in the one-level-higher pastebin-stream/ folder *
universalpackage has quit [(Remote host closed the connection)]
katyucha has joined #nixos
fragamus has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
hellrazor has quit [(Ping timeout: 255 seconds)]
<pie_>
btw whats the status on reproducible builds these days?
universalpackage has joined #nixos
<nixer>
joepie91: Cool, I couldn't find any reference to webpack in there? It looks like the builder is just tar-ing everything up?
endformationage has joined #nixos
<joepie91>
nixer: yeah, like I said, I'm not webpack-building in that expression; I just linked it to demonstrate where stuff ends up after node2nix is done with it
<joepie91>
you should then be able to write an expression that uses webpack pointed at the result
balsoft has quit [(Read error: Connection reset by peer)]
hellrazor has joined #nixos
ylwghst has quit [(Ping timeout: 240 seconds)]
universalpackage has quit [(Remote host closed the connection)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] mbld opened pull request #31753: nixos-container: Make show-ip work together with ipv4 + netmask (master...show-ip) https://git.io/vFDVD
NixOS_GitHub has left #nixos []
katyucha has quit [(Quit: Lost terminal)]
katyucha has joined #nixos
universalpackage has joined #nixos
mahalel_ has quit [(Ping timeout: 240 seconds)]
hiratara has quit [(Read error: Connection reset by peer)]
hiratara_ has joined #nixos
MP2E has quit [(Remote host closed the connection)]
hiratara_ is now known as hiratara
jasom has quit [(Ping timeout: 240 seconds)]
mahalel_ has joined #nixos
tv1 has joined #nixos
jasom has joined #nixos
Turion has quit [(Ping timeout: 248 seconds)]
<pikajude>
i can nixops ssh into my server, but i can't build using that server
<pikajude>
:(
universalpackage has quit [(Remote host closed the connection)]
universalpackage has joined #nixos
ylwghst has joined #nixos
tv has quit [(Ping timeout: 268 seconds)]
<pikajude>
oh, i can't remotely build on it at all
<pikajude>
how do i debug this?
<pikajude>
all it tells me is "permission denied, try again" when trying to SSH
<pikajude>
but i can SSH into that server using the exact same arguments
<Dezgeg>
can you ssh from root?
<pikajude>
yes
<pikajude>
i had to add the host to known_hosts, but doing that didn't make the remote nix build work
<Dezgeg>
does it run non-nixos? does 'ssh thathost nix-env --version' give output?
srdqty has joined #nixos
erictapen has joined #nixos
<pikajude>
let me see
<pikajude>
it runs nixos
<pikajude>
but the disk is full, i didn't realize
<pikajude>
i'll try to address that first
universalpackage has quit [(Remote host closed the connection)]
<steveeJ>
Sonarpulse: I've been wanting to pull in zlib.static for the cross compiler but that has actually never worked so far, in a sense that the compiler won't find it at runtime neither will pkg-config. do you know what's missing to make it work?
<sitwon>
I am using the nix package manager on Ubuntu. I have noticed that graphical applications installed with Nix seem to be using different system fonts than the native applications. Is there a way to influence the system fonts used by packages installed with Nix?
<Sonarpulse>
steveeJ: what's the contents of that?
<Sonarpulse>
maybe it's near empty
tgunb has joined #nixos
<Sonarpulse>
also maybe it doesn't contain headers
<Sonarpulse>
so you still need zlib.dev or something
danharaj has joined #nixos
<steveeJ>
Sonarpulse: I actually use zlib.static and zlib.dev.static for native, which works, but the same doesn't work for cross
<Sonarpulse>
steveeJ: huh
<Sonarpulse>
yeah i don't know off hand
<Sonarpulse>
and don't have builds cached
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 2 new commits to master: https://git.io/vFDKd
<NixOS_GitHub>
nixpkgs/master 22c85de José Romildo Malaquias: enlightenment: fix names of setuid wrapped files
<steveeJ>
Sonarpulse: I'd appreciate some insight how the compiler manages to find the installed libraries in general
universalpackage has quit [(Remote host closed the connection)]
<steveeJ>
It's a blackbox to me still, since in the nix-shell there's no LD_LIBRARY_PATH, and I don't know how the native compiler manages to find libraries that are installed in addition
universalpackage has joined #nixos
<Sonarpulse>
steveeJ: so in the static case
<Sonarpulse>
it's very simply
<Sonarpulse>
-L flags
<Sonarpulse>
in the dynamic case it's more complex
<Sonarpulse>
because the RPATH/RUNPATH/ whatever --rpath-link makes of dynamic libraries
<Sonarpulse>
can effectively introduce new -L flags for transitive deps
<steveeJ>
Sonarpulse: some more insight. pkg-config fails to find static libraries, but gcc/ld is able to find it
<Sonarpulse>
steveeJ: missing .pc file
<Sonarpulse>
for pkg-config
<Sonarpulse>
maybe it didn't make it into .dev.static ?
<Sonarpulse>
(it would normally be some sort of dev output)
<Sonarpulse>
make sure that is also a build-input
<Sonarpulse>
even though you just need headers at build-time
<mekeor>
are there also nixos-options which are not listed on nixos.org/nixos/options.html because they require unfree packages? if so, how can i view them?
<steveeJ>
Sonarpulse: must be the case. how does pkg-config even find these files scattered in the /nix/store ?
<Sonarpulse>
(header outputs are a wart of the current thing)
<Sonarpulse>
steveeJ: see the pkg-config setup hook
<Sonarpulse>
it registers an "environment hook"
ambro718 has quit [(Ping timeout: 248 seconds)]
<Sonarpulse>
which gets applied to each pkg (bash function)
<steveeJ>
Sonarpulse: alright, thanks for patiently answering all my newbie questions :-D
<vcunat>
well, in my experience mixing static and dynamic libs is problematic, due to our not wanting them in the same directory
MP2E has quit [(Remote host closed the connection)]
<Sonarpulse>
steveeJ: np
<Sonarpulse>
your comment is very thorough
<vcunat>
mekeor: I don't think they would be hidden
<vcunat>
(from the manual or man configuration.nix)
<Sonarpulse>
vcunat: maybe the cross jobset could be musl static only or something
<steveeJ>
Sonarpulse: somehow it got partially lost when editing it on github though. I had a link to my edited version of the code
<Sonarpulse>
steveeJ: oh bummer
<Sonarpulse>
that is a dead link or edit text problem?
tv4 has joined #nixos
<steveeJ>
text edit problem. I had multiple occurrences of GitHub telling me that "the text you are editing has changed"
<vcunat>
generally I certainly find it easier to build completely separate derivations for static and dynamic libs
<vcunat>
(I'm not sure about your cross case)
<steveeJ>
vcunat: glibc{,.static} and zlib{,.static} have different output paths AFAIU
<greymalkin>
Ah... now I'm back to "'NoneType' object has no attribute 'get'" for nixops -- using debug it seems to be trying to get 'region' from the list of machines
<vcunat>
steveeJ: they have, but they're built from one derivation with the dynamic ones
<tgunb>
@pstn: can i have your openage nix-expr.?
<vcunat>
steveeJ: then you can't have $libdir that contains both, etc.
ylwghst has quit [(Ping timeout: 255 seconds)]
erictapen has quit [(Ping timeout: 246 seconds)]
Lisanna has quit [(Ping timeout: 260 seconds)]
<steveeJ>
vcunat: why not? the paths aren't identical
<steveeJ>
vcunat: in the case of zlib, I needed to override { static = true } in addition to using '.static'. IMHO the first one should be the correct and suffice
<adamt>
how do i refer to the actual path to an installed package? I'm trying to serve the content of riot-web on a webserver, so i figured i could use a path like nixos.riot-web
<pie_>
has anyone considered that it would be nice to use -A with nix-env -e ?
goibhniu has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej pushed 2 new commits to master: https://git.io/vFDQt
<NixOS_GitHub>
nixpkgs/master 440ce4c Frank Doepper: shotcut: 17.09 -> 17.11
simendsjo has quit [(Quit: ERC (IRC client for Emacs 25.3.1))]
hiratara has quit [(Ping timeout: 240 seconds)]
ssmike has quit [(Ping timeout: 248 seconds)]
ma27 has joined #nixos
<mekeor>
i would like to install only one certain package (foobar) from unstable channel. but all the solutions i found so far, suggest to to set `environment.systemPackages = [ unstable.foobar ];`¹ but i don't want to modify my system-config only because one certain user wants to install that unstable package.
<Ralith>
in a nix expression, is there a clean way to say "foo.bar, or if foo does not have that attribute, baz?"
<tilpner>
Ralith - x.foo or x.bar
<mekeor>
you mean `foo.bar or baz`?
<Ralith>
well that's convenient
<Ralith>
thanks
<LnL>
yeah or works for missing attributes
<LnL>
the alternative would be if foo ? "bar" ...
<mekeor>
any idea why `nix-env -iA nixos-unstable.flashplayer` tries to install version 27.0.0.183 although nixpkgs
<mekeor>
any idea why `nix-env -iA nixos-unstable.flashplayer` tries to install version 27.0.0.183 although nixpkgs' master-branch uses version 27.0.0.183?
<tilpner>
master has 27.0.0.187
kier has quit [(Remote host closed the connection)]
<tilpner>
That commit is 3 days old, nixos-unstable is displayed to be 2 days old, but that's close enough to consider an error
<mekeor>
err, that's what i meant, yeah
<mekeor>
ah, hmm, i see.
<mekeor>
tilpner: how do you know the age of nixos-unstable?
<steveeJ>
clever: Sonarpulse: vcunat: thanks for the lessons on the library/toolchain knowledge earlier!
<Sonarpulse>
steveeJ: glad to help!
<Sonarpulse>
steveeJ: also I'm in the process of finally rebasing some of that stuff
<rihards>
hey, how do i install a nix derivation from a local .nix file?
<rihards>
nix-env -i somefile.nix or anything similar i've tried doesn't work
<tilpner>
Try -if somefile.nix
<rihards>
somefile.nix basically just contains a call to mkDerivation
<rihards>
that didn't work. i got error: cannot auto-call a function that has an argument without a default value (‘mkDerivation’)
<rihards>
which makes sense
<rihards>
because the function specified in the file has a bunch of arguments which i don't know the values for
iMatejC has quit [(Remote host closed the connection)]
<rihards>
ok then, let's take a step back - what's the intended way to install nix packages that come from somewhere else instead of nixpkgs - like haskell packages from hackage for example?
<tilpner>
You don't want to take arguments for your top-level file. People usually put "with import <nixpkgs> {};" or "let pkgs = import <nixpkgs> {}; in ..." at the top of their file in that case
civodul has quit [(Quit: ERC (IRC client for Emacs 25.3.1))]
<tilpner>
If you do want to keep the arguments, perhaps because you want to later put this into nixpkgs, you can use nix-build -E 'with import <nixpkgs> {}; callPackage ./somefile.nix {}'
<rihards>
the thing I'm trying to install is a nix expression i got after running cabal2nix
iMatejC has joined #nixos
<rihards>
hmm, i could try to import nixpkgs and then inherit the args that i need to fill in
dywedir has quit [(Remote host closed the connection)]
iMatejC has quit [(Remote host closed the connection)]
<MP2E>
rihards: you've got the right idea, the piece missing is that you don't directly use the nix expression that is generated from cabal2nix when installing, it needs to be referenced in your ~/.nixpkgs/config.nix file, as well
<tilpner>
I gave an example for (practically) that
<MP2E>
here is my ~/.nixpkgs/config.nix file, for reference
p4cman has quit [(Quit: Connection closed for inactivity)]
<rihards>
oh, it turns out i don't have the ~/.nixpkgs/config.nix file
iMatejC has joined #nixos
ssmike has quit [(Ping timeout: 260 seconds)]
<rihards>
thank's, i'll probably be able to piece together what is needed from it :)
<rihards>
but what's the way I should be interacting with my config.nix? going over the manual I only saw mentions of nix-env -i ... for installing packages
kreisys has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<MP2E>
nix-env will refer to your ~/.nixpkgs/config.nix file automatically, assuming it's in the home folder of the user running nix-env, no need to explicitly include
<MP2E>
so for instance, in my config.nix file there, I am overriding xmonad with a local nix script I have in that same repo, which simply uses the latest git version of xmonad
<MP2E>
if i wanted to install this with nix-env, i would type 'nix-env -iA myHaskellPackages.xmonad'
nixer has quit [(Ping timeout: 260 seconds)]
<MP2E>
in this case, I seperated out my overrides of haskell packages into a seperate namespace, myHaskellPackages. Definition is on line 42. But you can simply override 'in place' as well
<MP2E>
a bit lower, I create a new attribute for a new nix expression of a simple C program, doom64ex. Installing that is as simple as typing 'nix-env -iA doom64ex'
orivej has joined #nixos
<MP2E>
I gotta get back to work but hopefully that helps some!
mekeor has quit [(Quit: digital revolution = cryptoanarchism + cybercommunism)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] LnL7 opened pull request #31760: gstreamer: fix darwin build (master...darwin-gstreamer) https://git.io/vFyTj
NixOS_GitHub has left #nixos []
hke has joined #nixos
mkoenig has joined #nixos
<nhill>
Anyone run an install off of USB, rather than just using it to install?
mkoenig_ has quit [(Ping timeout: 248 seconds)]
<nhill>
I'm wanting to setup a LiveUSB that I actually use as my everyday setup (desktop with no HD)
_ts_ has quit [(Quit: Leaving)]
<danbst[m]>
nhill: I've tried to do some daily work off from live-USB, but it was a bit slow to start
<tilpner>
A persistent live USB is perfectly usable, if all machines you intend to use it with have unlocked boot settings, and a USB3 port
<nhill>
tilpner: yep
<nhill>
But, I'd like to have a normal filesystem setup
<nhill>
with a persisten home dir
<danbst[m]>
how large is your USB drive?
<nhill>
64, but I'll probably buy a 128
<nhill>
My plan is to have IP67 (waterproof/dustproof) USB drive and a large SSD (128GB/2TB).
<danbst[m]>
yeah, 64 is too little for active usage/develop
<nhill>
Hopefully will contain all the software I use, all the configurations and all of my data
<nhill>
danbst[m]: My current laptop is 128, with no problems
tgunb has quit [(Ping timeout: 260 seconds)]
<nhill>
I'm asking, because I know abunch of ditros install images do funky things with their filesystems
<nhill>
Since they don't want any changes to be persistent and they don't rely on other storage
<avn>
nhill: 64 is good for swap, L2ARC cache
<nhill>
I guess I just need to know how to partition and configure the USB to boot properly.
<nhill>
Is that tricky?
<danbst[m]>
not much. Do you plan UEFI?
<nhill>
I'm 10 years behind on booting, haven't paid any attention to UEFI
<nhill>
Whatever is simplest and will allow my USB install to boot on the most hardware
barrucadu has quit [(Read error: Connection reset by peer)]
spietz has quit [(Ping timeout: 248 seconds)]
<nhill>
will following the normal install procedure be fine, or would USB be different?
<danbst[m]>
USB should be no different
<jack[m]>
Hrm. I don't suppose anyone has implemented code to generate Hadamard matrices in nix?
<nhill>
okay, so the bits on the stick can be exactly the same as a normal drive and the bios will be fine booting it up?
<danbst[m]>
jack: Oo
<danbst[m]>
nhill: if you specify proper boot.loader.grub.device, then yes
<nhill>
danbst[m]: Okay, great to know
<jack[m]>
danbst: useful in experiments where you have a lot of inputs and you want to try and figure out which ones affect the outputs.
<nhill>
I'll just do the install with two usb sticks plugged in :)
<danbst[m]>
nhill: are you on NIxOS?
<danbst[m]>
it has nixos-install command built-in
<nhill>
danbst[m]: not yet
<disasm>
trying to get a debian container running in systemd-nspawn to boot on a bridge interface on nixos... any suggestions? So far setup bridge br0 with my real interface in it. Static'd that IP. Networking works great on host, then run systemd-nspawn -b -D /debian --network-veth --network-bridge=br0
<disasm>
the end goal here is to move everything to nixos... but trying to run this server in parallel until I get everything working
MP2E has quit [(Read error: Connection reset by peer)]
amfl has quit [(Ping timeout: 255 seconds)]
<disasm>
i tried ip addr add 192.168.0.254/24 dev host0@if8 but it says cannot find device host0@if8