ryantrinkle has quit [(Ping timeout: 248 seconds)]
erasmas has quit [(Quit: leaving)]
<NixOS_GitHub>
[nixpkgs] abbradar pushed 1 new commit to master: https://git.io/vDIyT
<NixOS_GitHub>
nixpkgs/master 98b0195 Nikolay Amiantov: video services: don't install OpenCL files to /etc...
Shou has quit [(Ping timeout: 255 seconds)]
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
thc202 has quit [(Ping timeout: 258 seconds)]
Itkovian has joined #nixos
<Ankhers>
Would anyone be able to share how they go about using node in NixOS? I specifically need to install brunch, but everything I read suggests installing it globally, which I can't do because that would try to save it into the nix store.
twopoint718 has joined #nixos
tempeh has joined #nixos
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
athan has joined #nixos
Itkovian has joined #nixos
<tempeh>
hi. i just installed nixos today, and i'm having some bumpes. i'm guessing they're pretty common so i hope someone here can help
<gchristensen>
ask away! it is a quieter time in #nixos but you can ask anyway. many people here are in CET time, I my self can't type much more today.
<unlmtd[m]>
MichaelRaskin: Thanks a lot !
<unlmtd[m]>
not-os not so easy to find
<yorick_>
what's the difference between symlinkjoin and buildenv?
<tempeh>
at the moment i'm trying to xmonad to compile my config. 'xmonad --recompile' gives a few compilation errors that make me think it can't see contib packages (can't find module EZConfig, etc.), but I'm using nixos-unstable and xmonad.enableContribAndExtras
<yorick_>
Ankhers: use node2nix
mudri has quit [(Quit: WeeChat 1.6)]
<yorick_>
Ankhers: node2nix is still in development, what you can also do is install it locally, then put ./node_modules/.bin in your PATH
<unlmtd[m]>
wow. I need to load not-os on qemu for sure !
<Ankhers>
tempeh: Meta-q should restart (and recompile) XMonad. If there are issues with the config, you will get a new tile telling you so and what the error is.
<tempeh>
Ankhers: yeah, meta-q runs 'xmonad --recompile', which is where i'm seeing the errors
<Ankhers>
yorick_: And what is the suggested way of adding something to my PATH? .bashrc doesn't seem to be executed.
athan has quit [(Remote host closed the connection)]
<NixOS_GitHub>
[nixpkgs] abbradar pushed 1 new commit to master: https://git.io/vDI7h
<NixOS_GitHub>
nixpkgs/master dbd4a35 Benjamin Saunders: vulkan-loader: 1.0.26.0 -> 1.0.39.1
eacameron has quit [(Remote host closed the connection)]
digitalmentat has quit [(Ping timeout: 255 seconds)]
<tnks>
savanni: I often say I'm going to help you out and never get around to it... hopefully this time is better: https://github.com/shajra/example-nix
<tnks>
also, anyone else doing a Nix/Haskell combo, please feel free to review this repository.
justan0theruser has joined #nixos
justanotheruser has quit [(Ping timeout: 264 seconds)]
athan has joined #nixos
MP2E has joined #nixos
derjohn_mob has quit [(Ping timeout: 240 seconds)]
wak-work has quit [(Ping timeout: 255 seconds)]
Mic92 has quit [(Ping timeout: 248 seconds)]
eacameron has joined #nixos
derjohn_mob has joined #nixos
eacameron has quit [(Ping timeout: 240 seconds)]
<stepcut>
bazel is vexing
eacameron has joined #nixos
wak-work has joined #nixos
eacameron has quit [(Ping timeout: 255 seconds)]
Bluebell_ has joined #nixos
mbrgm has quit [(Ping timeout: 240 seconds)]
sellout- has joined #nixos
systemfault has quit [(Quit: Bye!)]
mbrgm has joined #nixos
sellout- has quit [(Ping timeout: 240 seconds)]
Drakonis has joined #nixos
<NixOS_GitHub>
[nixpkgs] jbaum98 opened pull request #22305: smartmontools: 6.4 -> 6.5 and fix segfault on darwin (master...smartmontools) https://git.io/vDIAw
<Drakonis>
wow 75 commits until 100k nixpkgs
<gchristensen>
:D
<gchristensen>
I've been counting down for a few days now
twopoint718 has joined #nixos
mizu_no_oto has quit [(Quit: ["Textual IRC Client: www.textualapp.com"])]
sellout- has joined #nixos
<alphor>
it's kind of crazy that all derivations for the channel are in one repo. surely there is some way to have derivations be in separate repos.
<gchristensen>
why, though?
<simpson>
alphor: Think of it like a ports tree.
<Drakonis>
when is nix getting new commands
<alphor>
all I know about the ports tree is from that linux -> bsd rant I read a few weeks ago.
<Drakonis>
or just a whole cleanup
<Drakonis>
link your rant, thanks in advance
twopoint718 has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<Drakonis>
or maybe it is only the website's appearance
<Drakonis>
i can't tell
<alphor>
gchristensen: to be honest I don't know. It seems working, just I imagine logs are a huge mess
<NixOS_GitHub>
[nixpkgs] vbgl opened pull request #22306: luaPackages.luazlib: fix on darwin (master...luazlib-darwin) https://git.io/vDIpT
<gchristensen>
commit logs?
<alphor>
yes.
<tilpner>
Centralised logs seem a lot easier than having to chase commit messages over all of GitHub
<gchristensen>
who cares, more importantly: you can `git bisect` *your whole operating system*
Bluebell_ has left #nixos []
<Drakonis>
i wonder why keepassxc isn't packaged yet
<Drakonis>
nix has quite a lot obscure packages already
Wizek has joined #nixos
<Drakonis>
a question though
<Drakonis>
why do we still not have the ability to pull from git
<Drakonis>
compile from git
<gchristensen>
sorry?
<Drakonis>
my bad
<Drakonis>
compile from master
* gchristensen
is failing to understand
<Drakonis>
always compile the latest version
<Drakonis>
the latest commit
<Drakonis>
similar to arch's -git builds
<gchristensen>
because it isn't reproducible
<pikajude>
well, it is, with builtins.fetchgit
Wizek_ has joined #nixos
<pikajude>
it's just not a good idea
<Drakonis>
of course not
<Drakonis>
but its a "why not" situation
<gchristensen>
because it isn't reproducible and nix enforces reproducibility
<Drakonis>
no, yes i get it.
<Drakonis>
but there is a use to it.
<Drakonis>
using hydra for local development
Wizek_ has quit [(Ping timeout: 258 seconds)]
eacameron has joined #nixos
hexagoxel has quit [(Ping timeout: 240 seconds)]
Wizek has quit [(Ping timeout: 252 seconds)]
_rht has quit [(Quit: Connection closed for inactivity)]
mguentner has quit [(Quit: WeeChat 1.6)]
hexagoxel has joined #nixos
Drakonis has quit [(Read error: Connection reset by peer)]
roconnor has joined #nixos
Sonarpulse has quit [(Ping timeout: 255 seconds)]
<thomassgn>
hi, I'm having trouble running context after just upgrading my system. Keep getting something about resolvers scanning /nix/store and then it just works for a long time. No Idea if it actually does anything.
<thomassgn>
anyone running context successfully on an up-to-date system?
mguentner has joined #nixos
<MP2E>
do you mean using the latest nixpkgs master tree to build off of?
<MP2E>
or unstable or stable or something else
<MP2E>
couldn't tell by the context, but I updated my nixpkgs master tree today and had some issues on boot
<MP2E>
had to revert to an older profile
<MP2E>
might be related
<thomassgn>
context as in the texlive/tex/latex thing
<MP2E>
ah okay, not related then. hm
<thomassgn>
hehe, np. just ran nix-channel --update and nixos-rebuild switch for the heck of it, and suddenly my context install won't do shiet...
<thomassgn>
(the rebuild switch was with an --upgrade)
systemfault has joined #nixos
hamishmack has quit [(Quit: hamishmack)]
hamishmack has joined #nixos
hamishmack has quit [(Client Quit)]
Wizek has joined #nixos
mguentner2 has joined #nixos
mguentner has quit [(Ping timeout: 240 seconds)]
georges-duperon has quit [(Ping timeout: 252 seconds)]
georges-duperon has joined #nixos
eacameron has quit [(Remote host closed the connection)]
<srhb>
Is there an issue tracking terminal corruption in nix-build and friends on unstable? I haven't been able to find it.
<srhb>
Actually I suppose it's not terminal corruption, just a bunch of weird newlines...
<srhb>
Maybe it's curl's fault.
<stepcut>
any idea how to use cabal2nix to generate a default.nix for a repo on github, when the cabal package is in a subdirectory?
<NixOS_GitHub>
[nixpkgs] fpletz closed pull request #22305: smartmontools: 6.4 -> 6.5 and fix segfault on darwin (master...smartmontools) https://git.io/vDIAw
AllanEspinosa has quit [(Ping timeout: 248 seconds)]
MP2E has quit [(Quit: brb)]
hamishmack has joined #nixos
eacameron has joined #nixos
georges-duperon has quit [(Quit: Leaving)]
obadz has quit [(Ping timeout: 240 seconds)]
Wizek has quit [(Ping timeout: 240 seconds)]
eacameron has quit [(Ping timeout: 240 seconds)]
obadz has joined #nixos
<NixOS_GitHub>
[nixpkgs] vbgl pushed 1 new commit to master: https://git.io/vDLUZ
<NixOS_GitHub>
nixpkgs/master 61c48d9 Vincent Laporte: luaPackages.luaexpat: fix build on darwin
eacameron has joined #nixos
eacameron has quit [(Ping timeout: 240 seconds)]
ryantrinkle has joined #nixos
endformationage has quit [(Quit: WeeChat 1.6)]
twopoint718 has joined #nixos
athan has quit [(Remote host closed the connection)]
eacameron has joined #nixos
eacameron has quit [(Ping timeout: 240 seconds)]
eacameron has joined #nixos
hamishmack has quit [(Read error: Connection reset by peer)]
hamishmack has joined #nixos
twopoint718 has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
eacameron has quit [(Ping timeout: 255 seconds)]
systemfault has quit [(Quit: Bye!)]
Wizek has joined #nixos
hamishmack has quit [(Ping timeout: 276 seconds)]
hamishmack has joined #nixos
<nhooyr>
I've got a digital ocean VPS on which I have a fileserver/webserver and whatnot all running. I just login and use /etc/nixos/configuration.nix to declare the state of the system and what services should be running, what packages should be installed etc. What can nixops do for me? I'm not quite understanding the descriptions.
ebzzry has quit [(Ping timeout: 245 seconds)]
<roconnor>
nixops is mostly useful for coordinating several instances of cloud servers running nixos. If you just have one, there are fewer benefits of using nixops. Nixops will still let you provision that once instance from your local computer though, if that is beneficial.
<roconnor>
*that one instance
<roconnor>
Basically you would keep the configuration on your local computer instead of on the cloud server.
Wizek has quit [(Ping timeout: 245 seconds)]
<nhooyr>
ah ok thanks
athan has joined #nixos
cstrahan has joined #nixos
moet has joined #nixos
<NixOS_GitHub>
[nixpkgs] orivej opened pull request #22308: pidgin-xmpp-receipts: init at 0.7 (master...pidgin-xmpp-receipts) https://git.io/vDLLi
twopoint718 has joined #nixos
edvorg has quit [(Ping timeout: 240 seconds)]
MercurialAlchemi has joined #nixos
sdothum has quit [(Read error: Connection reset by peer)]
<srhb>
You could probably do something after unpackPhase to get at it manually, but yea
<srhb>
It should be easier to figure out now ;-)
xeviox has joined #nixos
raydeo has quit [(Quit: until next time!)]
athan has quit [(Ping timeout: 255 seconds)]
fritschy has joined #nixos
raydeo has joined #nixos
jgertm has joined #nixos
MichaelRaskin has quit [(Quit: MichaelRaskin)]
<nhooyr>
srhb: when I run genericBuild it all works great. but if I try and run each phase myself, neovim tries to download the dependencies into a .deps folder and then everything fails with many errors. I'm not running unpackPhase because I already have it's git repository cloned but even if I do and cd into it and then run configurePhase (there is none)
<NixOS_GitHub>
[nixpkgs] benley opened pull request #22309: vim: make it possible to build with --enable-gui=gtk3 (master...vim-gtk3) https://git.io/vDLm8
Wizek has joined #nixos
<hyper_ch>
hmmm, I just tried to burn a cd with K3B but its telling me that cdrecord has no permission to open the device
<NixOS_GitHub>
[nixpkgs] lsix force-pushed python-wip from ef5f376 to c8e4049: https://git.io/vzaOS
<NixOS_GitHub>
nixpkgs/python-wip c8e4049 Lancelot SIX: pythonPackages.pytest: use pytest_30
<sphalerite>
Aah, it does want the caps
<sphalerite>
I'm not sure nixpkgs currently has a mechanism for giving executables caps, but it should work if you just make it setuid
<sphalerite>
by putting in the system configuration security.extraSetuidPrograms = [pkgs.cdrecord];
panostimos has joined #nixos
<hyper_ch>
what's the difference to: security.setuidPrograms = [ "mount.cifs" ]; ?
<hyper_ch>
between setuidprograms and extrasetuidprograms?
goibhniu has quit [(Ping timeout: 245 seconds)]
<hyper_ch>
trace: warning: The option `security.extraSetuidPrograms' defined in `/etc/nixos/configuration.nix' has been renamed to `security.setuidPrograms'.
<hyper_ch>
this works: security.setuidPrograms = [ "mount.cifs" "pkgs.cdrecord" ];
<NixOS_GitHub>
[nixpkgs] vcunat pushed 3 new commits to master: https://git.io/vDLGA
<NixOS_GitHub>
nixpkgs/master 3dd2a27 Benjamin Staffin: vim: make it possible to build with --enable-gui=gtk3...
<NixOS_GitHub>
nixpkgs/master e01c15d Benjamin Staffin: nixos: if gnome3 is installed, build gvim for gtk3 too
<NixOS_GitHub>
nixpkgs/master 01751a9 Vladimír Čunát: Merge #22309: vim: allow building with gtk3
<NixOS_GitHub>
[nixpkgs] vcunat closed pull request #22309: vim: make it possible to build with --enable-gui=gtk3 (master...vim-gtk3) https://git.io/vDLm8
nschoe has joined #nixos
<hyper_ch>
sphalerite: ok, the cdrecord in /var/setuid-wrappers appears in K3B's permissions settings screen... but still error
<sphalerite>
:( then I really don't know anymore, sorry
<clever>
hyper_ch: i think if you run k3b under strace, as root, then it will just naturaly gain root without the setuid
<clever>
then you should be able to reproduce the issue
<hyper_ch>
I tried gksu k3b but then I couldn't access any folders to pick files to burn
<clever>
throw a file in /tmp?, or was the entire browse window broken?
<hyper_ch>
no, I have entries Home and Root
<hyper_ch>
but I can't unfold them and no files appear
glines has quit [(Ping timeout: 258 seconds)]
<clever>
ah
<clever>
sounds like the browse ui is broken under gksu
<clever>
you can usualy just type/paste in the full path to the ISO
<edwtjo>
gchristensen: aha, great!
<hyper_ch>
clever: I want to burn files onto a cd
<clever>
ah, thats more complex then
<clever>
hmmm, simplest thing i can think of is to run k3b without root, purely in a "group of files" to "iso file" mode
<clever>
then re-run it as root, and paste in that iso path, and debug the burning issue there
<hyper_ch>
that's easy?
<hyper_ch>
shouldn't a cd-brun program just run?
<clever>
i would expect k3b to have both modes in it
<clever>
hyper_ch: oh, is cdrecord in /run/current-system/sw/bin/ ?
<hyper_ch>
creating iso and then running gksu k3b to burn the iso onto cd works
<hyper_ch>
which k3b
<hyper_ch>
/run/current-system/sw/bin/k3b
<hyper_ch>
sphalerite also said to add some setuid stuff
thaega has joined #nixos
johann__ has joined #nixos
<clever>
next simplest option i can think of is to just read the man pages for mkisofs and cdrecord, then do it purely via CLI
johann__ has quit [(Client Quit)]
goibhniu has joined #nixos
thaega has left #nixos ["ERC Version 5.3 (IRC client for Emacs)"]
<clever>
you can also loopback mount the iso youve created, to confirm it looks right, before doing the final burn
* clever
heads off to bed
<hyper_ch>
phreedom's on the case now... so relax everyone :)
<NixOS_GitHub>
[nixpkgs] vbgl pushed 1 new commit to master: https://git.io/vDLcD
<NixOS_GitHub>
nixpkgs/master 3ccc139 Matthias Beyer: cataract: init at 1.1.0 and HEAD
<NixOS_GitHub>
[nixpkgs] vbgl closed pull request #21307: cataract: init at 1.1.0 and HEAD (master...add-cataract) https://git.io/v1NlM
ThatDocsLady has joined #nixos
jgertm has joined #nixos
thc202 has joined #nixos
ilja_kuklic has joined #nixos
derjohn_mob has quit [(Ping timeout: 240 seconds)]
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
Itkovian has joined #nixos
Wizek has quit [(Ping timeout: 240 seconds)]
BluntObject has quit [(Quit: Lost terminal)]
JagaJaga has joined #nixos
athan has joined #nixos
rly has joined #nixos
twopoint718 has joined #nixos
<NixOS_GitHub>
[nixpkgs] DamienCassou created remove-maintainer-DamienCassou (+1 new commit): https://git.io/vDLlS
<NixOS_GitHub>
nixpkgs/remove-maintainer-DamienCassou 58dc8e3 Damien Cassou: Remove myself from maintainers
derjohn_mob has joined #nixos
<nhooyr>
yay, finally got neovim working
<hyper_ch>
congrats
<hyper_ch>
what's neovim?
roconnor has quit [(Ping timeout: 255 seconds)]
<NixOS_GitHub>
[nixpkgs] DamienCassou opened pull request #22310: Remove myself from maintainers (master...remove-maintainer-DamienCassou) https://git.io/vDLlb
ronny has quit [(Ping timeout: 248 seconds)]
ronny has joined #nixos
twopoint718 has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<avn>
btw, would be nice to have some meta flag/standard comment, if maintainer willing to pass maintainership to someone other...
<ronny>
fml and wtf
<avn>
I have few packages which I do/will do for my job, and I am ok to support them, but also I want to flag them, that more involved maintainer will be welcomed
<ronny>
nixos grew new collisions :/
<avn>
(is abstract question, not about any real stuff I added/plan to add ;)
<avn>
well, tbh about plans to add something ;)
<goibhniu>
avn: hrm, but we can have multiple maintainers for a package, and people can remove themselves ... the way I see it, if you care about a package, and update it, fix it when it's broken etc. you might as well add yourself as a maintainer
<avn>
goibhniu: I just not sure, if PRing new packages with emptry maintainer list is a good idea.
Shou has joined #nixos
<goibhniu>
avn: AFAIK people are encouraged to add themselves as the maintainer when they add a new package
eacameron has joined #nixos
<avn>
goibhniu: I see, but for example... I need to build few rdp solutions, and evaluate them. I'll will have 4-5 good done packages. After few weeks evaluations, I continue with 2 of them. I won't throw away 3 others, but I welcome if someone overtake them.
civodul has joined #nixos
mudri has joined #nixos
<avn>
(is just example, just to let you catch my question)
<goibhniu>
avn: in general, I think the NixOS community tries to avoid too many policies on things like that, so you could either remove yourself as maintainer (with the risk that someone could remove the package if it breaks in the future, without letting you know), or wait until someone else steps up to maintain it, and then remove yourself
nwspk has joined #nixos
<avn>
goibhniu: I use second way for most stuff I add.
<ronny>
a newMaintainerWanted flag may be usefull
<avn>
ronny: yes, it what about I asked ;)
Shou has quit [(Ping timeout: 252 seconds)]
mudri has quit [(Ping timeout: 240 seconds)]
mkoenig has quit [(Read error: Connection reset by peer)]
athan has quit [(Read error: Connection reset by peer)]
<goibhniu>
hrm, I think it's pretty cool to add new packages for things you're evaluating, but if you decide not to use them afterwards yourself I think you could remove yourself as the maintainer and it would mean the same thing as "maintainer wanted/remove at will".
mkoenig has joined #nixos
ilja_kuklic has quit [(Ping timeout: 245 seconds)]
<NixOS_GitHub>
[nixpkgs] peterhoeg opened pull request #22313: rockbox_utility: use qt5 instead of qt4 (master...f/rb) https://git.io/vDLRn
<manveru>
that way adding new plugins just needs two lines, though i should probably also generate the configuration.nix
<nhooyr>
neat
Shou has quit [(Ping timeout: 258 seconds)]
civodul has joined #nixos
<jophish>
manveru: update_plugins.rb looks a lot like update-nix-fetchgit
<jophish>
haskell package to do the same
<manveru>
i guess so, i do the same with bundix
mizu_no_oto has joined #nixos
<jophish>
sphalerite: I can't really remember. I do remember thinking that vam was too hard to use, but looking at manveru's configuration I can't imagine how I got that impression, it looks really easy :)
<sheenobu>
jophish, neat! I've been doing the same neovim nix thing
<sheenobu>
but without the makeRtp stuff which looks really useful
<jophish>
sphalerite: I've not found neovim to be slow starting up at all
<manveru>
i found it mostly depends on the plugins and language you're using...
athan has quit [(Ping timeout: 255 seconds)]
<manveru>
neovim itself is fast as hell :)
mudri has quit [(Ping timeout: 240 seconds)]
<manveru>
i got hooked with neomake... used to have to wait 10 seconds after opening/saving a file for linters and the like
<NixOS_GitHub>
[nixpkgs] peterhoeg opened pull request #22314: krename-qt5: init at 20161228 (master...u/krename) https://git.io/vDLix
mkoenig has quit [(Remote host closed the connection)]
sellout- has quit [(Quit: Leaving.)]
<jophish>
manveru: deoplete is super for the same reason
<jophish>
nothing hangs before completions appear
<manveru>
i shall try that
pi3r has joined #nixos
morsicus1 has quit [(Quit: Leaving.)]
lsix has quit [(Remote host closed the connection)]
ebzzry has joined #nixos
<unlmtd[m]>
im hooked. Cant tolerate the dirty pile of files method
<unlmtd[m]>
Just pulled myself from a week of openwrt , what a nightmare
<gchristensen>
nice
<unlmtd[m]>
push this button, reboot, change this, restart network
<unlmtd[m]>
wow
<unlmtd[m]>
just lik in 1985 man!
<sphalerite>
I only started using nixos a couple of months ago, but now SSHing into debian servers makes me so sad
<unlmtd[m]>
then I looked at the voiid linux installer to do a setup, and it failed, and now Im just burning a nixos image
<sphalerite>
It's what I always wanted
<unlmtd[m]>
nixos is just the tip of the icebeeg
<unlmtd[m]>
im a OS snob now
<sphalerite>
(actually it's always made me sad. But I didn't realise that what I was looking for already existed)
mudri has joined #nixos
<unlmtd[m]>
wow what a difference. I have all my tools on the live image, even my editor plugins!
<unlmtd[m]>
ftw!
<gchristensen>
did you succumb to systemd?
eacameron has joined #nixos
<LnL>
TIL. the AMIs don't have any swap, that explains a few things
<gchristensen>
:) swap on an EBS volume is verrryyy nasty
<unlmtd[m]>
runit is made for nixos
<LnL>
0 free memory isn't great either :)
<unlmtd[m]>
and then you almost magically get nixos-bsd and nixos-tiny
<gchristensen>
LnL: well ... fair, but (for example) I would prefer my nodes crash than use swap
ebzzry has quit [(Ping timeout: 258 seconds)]
<unlmtd[m]>
in a way nixos is one of the things that makes systemd lose a lot of its attraction
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
eacameron has quit [(Ping timeout: 248 seconds)]
<Yaniel>
how ironic
<unlmtd[m]>
nixos has solved the problem of arranging a lot of modular systems
<unlmtd[m]>
and it thrives on them actually
<unlmtd[m]>
so systemd is in the way for nix to expand past linux
<unlmtd[m]>
for now*
panostimos has quit [(Ping timeout: 258 seconds)]
<unlmtd[m]>
Ill play with not-os when I go on my vacation
<gchristensen>
I don't think that is a direction nixos will be moving any time soon
<jophish>
where's clever?
<ixxie>
unlmtd[m]: I have been thinking about this too, mainly because there are possibilities to do really cool things if a nix-native init+dm system is made
<unlmtd[m]>
runit is perfect for that
<ixxie>
unlmtd[m]: one could, for example, select particular modules to be loaded before login
<gchristensen>
you already can
<unlmtd[m]>
ya or swap the kernel for something else
<unlmtd[m]>
nixos is made for distributed OS
<ixxie>
gchristensen: I am aware you can do so via bootloader profiles, but I meant that such a section would be done with a display manager
<ixxie>
unlmtd[m]: I am not familiar with runit (or init systems in general tbh) but I recently read about GNU dmd which is being used in Guix
<ixxie>
unlmtd[m]: the natural thing - something simpson apparently considered doing but had no time for - would be to have the init system configurable in Nix, so you could configure init processes in your configuration
<sphalerite>
One thing I still don't understand is why service managers run as PID 1
<sphalerite>
Rather than having a process that does only PID 1's responsibility (reaping zombies) and fork-and-execs another process that can do everything else
<sphalerite>
(like sinit)
<unlmtd[m]>
suckless is a bit extreme
<sphalerite>
Certainly, but what's wrong with putting service management in another process?
<unlmtd[m]>
nothing
<sphalerite>
Then why wouldn't dmd for instance do so?
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to release-16.09: https://git.io/vDL5X
<NixOS_GitHub>
nixpkgs/release-16.09 dc96f03 Vladimír Čunát: Merge #22310: Remove myself from maintainers...
mguentner2 is now known as mguentner
<manveru>
sphalerite: you like runit? :)
<sphalerite>
I don't know, I haven't looked into it much yet
<sphalerite>
I'm not really happy with systemd's model, but it does work and I have a finite amount of time :p
<manveru>
i used it a lot for work, and ran void linux for a while
<manveru>
but well, my ideal would be to either build void linux with nix, or implement runit as alternative for systemd on nixos...
<manveru>
neither of which i have time for :(
<sphalerite>
Still, the worst that's out there is probably bash-as-init. Thanks, docker.
<manveru>
don't give them any ideas :)
<sphalerite>
give them ideas? It's already happened
<manveru>
you can still use custom entrypoint instead of bash
<gchristensen>
damn ...
<sphalerite>
yeah but huge numbers of images do use bash
<gchristensen>
anyone here use Gnome? We need people who are using and are interested in gnome to step up and help maintain the package set
<LnL>
I've also seen stuff like this RUN ["bash","-c","foo &; bar &; baz"]
<sphalerite>
even if it's just running a script that starts up the service that the container is actually meant to provide
<sphalerite>
LnL: yep
<sphalerite>
the worst about that is probably not that it doesn't reap zombies, but that it doesn't actually allow terminating the container without sigkill
<manveru>
well, you can't fix other people
<sphalerite>
if only...
<gchristensen>
iirc docker was moving towards providing their own init for all containers by default, to handle thaht garbafge
<sphalerite>
sinit except with runtime-configurable fork command!
<manveru>
i'm afraid of anything docker does at this point :P
<gchristensen>
yap
<LnL>
I thought the philosophy was that there should only be a single process, even if it's not feasible
jkemet has quit [(Remote host closed the connection)]
<sphalerite>
nah I'm pretty sure it was just one *service*
<gchristensen>
philosophy doesn't always match up with reality :(
<gchristensen>
nope, The Docker Way was a single process (lol)
<sphalerite>
docker is pretty stupid
<sphalerite>
design-wise
rmeis has joined #nixos
<gchristensen>
hey now
<sphalerite>
but I like being able to docker run -p 30000:30000/udp linuxserver/minetest
Shou has joined #nixos
<sphalerite>
(for example)
<manveru>
hmm, deoplete requires a file called pyuv... how on earth do i add that :|
<unlmtd[m]>
so whats the best way to deploy a big infrastructure? Iv seen some guys took haskell into a dsl for that
<sphalerite>
I also don't see why it has to make volumes so complicated. They could have just been directories on the host in all cases, but no — they had to complicate things by reinventing mount and hiding the actual volume data
<unlmtd[m]>
but its closed source I think
<sphalerite>
nixops! (caveat: have never actually used it)
<manveru>
how big is your big? :)
<NixOS_GitHub>
[nixpkgs] jpierre03 opened pull request #22316: nix-exec: patch to compile with nix-0.12pre (master...pr-nix-exec) https://git.io/vDLFM
<jophish>
What's the process for building profiled haskell packages?
<manveru>
nixops is definitely the way to go though, it's a bit bumpy to get started, and you'll probably need a separate tool to handle mutable data
<jophish>
I've build all the dependencies with 'enableLibraryProfiling = true'
<jophish>
however I'm still getting linker errors when I come to use those libraries
nschoe_ has quit [(Quit: Program. Terminated.)]
nschoe has joined #nixos
twopoint718 has joined #nixos
rmeis_ has joined #nixos
<gchristensen>
you's guys ... `nix-store -l`
ixxie has quit [(Ping timeout: 260 seconds)]
<domenkozar>
jophish: what kind of errors?
andesil has joined #nixos
rmeis has quit [(Ping timeout: 255 seconds)]
andesil is now known as andsild
<LnL>
gchristensen: yeah I discovered that recently
<jophish>
domenkozar: a whole bunch of undefined references to symbols in dependent libraries
<jophish>
everything in my 'haskellPackages' has been build with enableLibraryProfiling, but when I try to use this package set to compile something with 'enableExecutableProfiling' things fall apart
<domenkozar>
strange
twopoint718 has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<gchristensen>
is there a more up to date guide on installing nix-daemon on other linuxes?
<gchristensen>
or is the wiki article best?
<jophish>
gchristensen: I wrote a guide a while ago, there's also the .deb output from nix itself
<LnL>
dunno, I just looked at the nixos module definition
<gchristensen>
it is almost as if we have a fairly transparent and repeatable system configuration tool :|
<NixOS_GitHub>
[nixpkgs] peterhoeg opened pull request #22317: wireshark: use cmake and move to gtk3/qt5 (master...f/ws) https://git.io/vDLA6
<manveru>
i added latest pyuv to the neovim extraPython3Packages, now i get an even more cryptic error :P
<jophish>
I used this on my current machine
<jophish>
and it's been working fine, except that I'm stuck on nixUnstable!
<jophish>
gchristensen: next time I'll use my script again though, as I don't like having nix-build and friends in /usr/bin
<gchristensen>
great :D
<jophish>
just waiting to go out of date
<jophish>
manveru: I don't remember having to add pyuv
<gchristensen>
I need to make this a puppet manifest thing
systemfault has joined #nixos
<manveru>
hmm
Wizek has quit [(Ping timeout: 240 seconds)]
<manveru>
it's actually a dependency of neovim/msgpack_rpc
<manveru>
which is used to talk to deoplete
<manveru>
maybe i can try upgrading that one
<manveru>
nope, that one is up to date
<manveru>
i dunno... python always confuses the shit out of me
<manveru>
i'll just remove the plugin until someone implements it in a sane language :)
rmeis has joined #nixos
ryantrinkle has quit [(Ping timeout: 264 seconds)]
<jophish>
gchristensen: Please let me know if you make that markdown into a script as your bash-foo is clearly much stronger than mine :)
<gchristensen>
hehe that bash is juuuust fine
<jophish>
domenkozar: it was a ghc nondeterminsm error I think
akaWolf has quit [(Ping timeout: 240 seconds)]
<jophish>
gchristensen: I had endless problems with the ssl certificates. What's in the script was working three months ago. I hope it's still working now :)
rmeis_ has quit [(Ping timeout: 240 seconds)]
jonasfj has quit [(Ping timeout: 256 seconds)]
Wizek has joined #nixos
data8 has joined #nixos
jonasfj has joined #nixos
vj4 has joined #nixos
Wizek has quit [(Ping timeout: 248 seconds)]
Rotaerk has quit [(Quit: Leaving)]
twopoint718 has joined #nixos
xadi has quit [(Quit: Leaving.)]
xadi has joined #nixos
reinzelmann has quit [(Ping timeout: 240 seconds)]
simon__ has joined #nixos
eacameron has joined #nixos
simon__ has quit [(Client Quit)]
simon__ has joined #nixos
simon__ is now known as sweber
sweber has quit [(Client Quit)]
sweber has joined #nixos
<gchristensen>
jophish: I didn't closely follow the script and ended up in ssl-hell too
MercurialAlchemi has quit [(Ping timeout: 256 seconds)]
twopoint718 has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<gchristensen>
jophish: ok, so here is a winner: deleting nix is easy ;)
fritschy has quit [(Quit: Leaving.)]
danharaj has joined #nixos
Shou has quit [(Ping timeout: 245 seconds)]
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to master: https://git.io/vDtJw
<NixOS_GitHub>
nixpkgs/master a3517bf Vladimír Čunát: lmdb: get rid of $out/man...
erasmas has joined #nixos
eacameron has quit [(Remote host closed the connection)]
grenade has quit [(Ping timeout: 252 seconds)]
eacameron has joined #nixos
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
khajvah_ has joined #nixos
grenade has joined #nixos
sdothum has quit [(Quit: ZNC - 1.6.0 - http://znc.in)]
ryantrinkle has joined #nixos
Shou has joined #nixos
sdothum has joined #nixos
<k0001>
I'm seeing error messages like the following in Hydra: https://gist.github.com/k0001/b4a401720a9f250719b69a32ff4d8bfb --- `x86_64-linux.docker-images` there is an attribute set, and these do not fail locally, just up on Hydra. Anybody seen this before?
eacameron has quit [(Remote host closed the connection)]
twopoint718 has joined #nixos
twopoint718 has quit [(Read error: Connection reset by peer)]
eacameron has joined #nixos
vandenoever has joined #nixos
vandenoever has quit [(Changing host)]
vandenoever has joined #nixos
twopoint718 has joined #nixos
andsild has quit [(Remote host closed the connection)]
eacameron has quit [(Remote host closed the connection)]
panostimos has joined #nixos
<ronny>
on nixos - how do i get debug builds of stuff so i can gdb hook into it?
<jophish>
gchristensen: oh no!
<jophish>
I hope you've got it working now
eacameron has joined #nixos
eacameron has quit [(Remote host closed the connection)]
eacameron has joined #nixos
<gchristensen>
yeah
<jophish>
just out of interest, what did you do wrong?
<jophish>
might be worth putting a note in the gist
<ronny>
how do i set the priority when using buildenv
<gchristensen>
I accidentally ignored a bunch of steps :D
<jophish>
gchristensen: It doesn't actually say to follow all the steps in order, so you're excused :)
<NixOS_GitHub>
nixpkgs/release-16.09 a1dae25 Franz Pletz: bitlbee: 3.5 -> 3.5.1 for multiple CVEs...
<NixOS_GitHub>
nixpkgs/release-16.09 b68b156 Franz Pletz: libarchive: add patch to fix CVE-2017-5601...
Drakonis has joined #nixos
akaWolf has joined #nixos
aminechikhaoui has joined #nixos
rmeis has quit [(Ping timeout: 260 seconds)]
khajvah__ has joined #nixos
khajvah_ has quit [(Ping timeout: 252 seconds)]
<NixOS_GitHub>
[nixpkgs] fpletz pushed 1 new commit to master: https://git.io/vDtsM
<NixOS_GitHub>
nixpkgs/master 44cbb0f Franz Pletz: tcpdump: 4.8.1 -> 4.9.0 for multiple CVEs...
<gchristensen>
fpletz: oh you found it?
<NixOS_GitHub>
[nixpkgs] fpletz pushed 3 new commits to release-16.09: https://git.io/vDtsA
<NixOS_GitHub>
nixpkgs/release-16.09 71bcd8b Franz Pletz: libpcap: 1.7.4 -> 1.8.1...
<NixOS_GitHub>
nixpkgs/release-16.09 81d57f0 Franz Pletz: tcpdump: 4.7.4 -> 4.8.1...
<NixOS_GitHub>
nixpkgs/release-16.09 1756a5d Franz Pletz: tcpdump: 4.8.1 -> 4.9.0 for multiple CVEs...
<fpletz>
gchristensen: yeah, took it from debian… it's trustworthy enough imho
<gchristensen>
LGTM
<gchristensen>
also that list of CVEs :o
<LnL>
fpletz: O_O
<gchristensen>
thank goodness for debian
<gchristensen>
they do so much mostly-thankless work
<fpletz>
probably went through the distros mailing list and debian just went ahead and packages it
<fpletz>
*packaged
reinzelmann has joined #nixos
JagaJaga has joined #nixos
johnsonav has quit [(Ping timeout: 255 seconds)]
Shou has quit [(Ping timeout: 240 seconds)]
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
eacamero_ has joined #nixos
eacameron has quit [(Read error: Connection reset by peer)]
Itkovian has joined #nixos
<NixOS_GitHub>
[nixpkgs] ttuegel pushed 1 new commit to master: https://git.io/vDtcD
<NixOS_GitHub>
nixpkgs/master fa1e8ff Thomas Tuegel: spotify: 1.0.47 -> 1.0.48
<NixOS_GitHub>
[nixpkgs] ttuegel pushed 1 new commit to release-16.09: https://git.io/vDtch
<NixOS_GitHub>
nixpkgs/release-16.09 7ae9da4 Thomas Tuegel: spotify: 1.0.47 -> 1.0.48...
rsynnest has joined #nixos
binaryphile has joined #nixos
<binaryphile>
i've got a redhat server behind a firewall with no access to the outside...presuming /nix exists (and is empty), if i wanted to install nix on it, is it possible to rsync down an installation from my machine?
<binaryphile>
i'm curious whether a nix install is user- or machine-specific
<simpson>
binaryphile: Yikes, this sounds like a bad situation. Nix on non-NixOS is machine-wide but set to be owned only by the installing (non-root) user.
<binaryphile>
it's ok if it's single-user
<binaryphile>
i only have access to one user there anyway
<binaryphile>
i'm familiar with how to set up the .nix-profile directory manually if necessary too
<binaryphile>
i'm more interested whether /nix/store and /nix/var, or the install database are going to be unhappy about the user or machine being moved out from under them
<binaryphile>
i know it's not ideal. :)
<simpson>
The store itself doesn't care, but the database that keeps the store consistent might care. Not sure.
<binaryphile>
that's what i thought might be the case.
<binaryphile>
i think i can run some experiments, i thought someone here might tell me whether it just can't work
<sphalerite>
FWIW I've been using nix-generated binaries on a non-nix system, running nix only on the host machine, and it's been working
<sphalerite>
but I'm guessing that's not what you want
<binaryphile>
sphalerite, how do you make sure you have all the dependencies in the tree, just copy the whole store?
<sphalerite>
so I just copy over the closure that I need
<simpson>
binaryphile: nix-copy-closure and etc. can help with that.
<binaryphile>
thx
<sphalerite>
simpson: not if the target system doesn't have nix working
<yorick>
I'm trying to set up hydra, running http://sprunge.us/dBWg as config. the idea is that it would build on localhost, but it queues the thing and then doesn't build on localhost
<yorick>
what more should I do?
<binaryphile>
i don't want to go too far off the path, i think i can get nix working on the remote
rly has quit [(Ping timeout: 245 seconds)]
<sphalerite>
alright
<sphalerite>
I'm off now, good luck!
<simpson>
sphalerite: Well, in this case, the target machine is firewalled from networks; I imagine that many hoops must be jumped through.
<binaryphile>
thx guys
<binaryphile>
i'm sure i'll have more questions later
<gchristensen>
yorick: you need to add localhost to the build-machines
<yorick>
gchristensen: it lists localhost as idle on /machines
nschoe has quit [(Quit: Program. Terminated.)]
<gchristensen>
yeah but I think you need to add it again anyway
<contrapumpkin>
niksnut: I can't remember if I've asked this before, but (as someone interested in potentially hosting an S3 binary cache for local use) how do you GC your S3 bucket?
<gchristensen>
contrapumpkin: we don't
<contrapumpkin>
so it's append-only? :)
<gchristensen>
yes
<niksnut>
well
<niksnut>
there is a GC script
<contrapumpkin>
I imagine that's a "now" thing and not necessarily what will be true in 10 years
<niksnut>
but I don't run it regularly
<contrapumpkin>
ah
<contrapumpkin>
where's that live?
<gchristensen>
I've fetched artifacts from S3 as old as 2014 I think?
<niksnut>
and it's not to up to date
<contrapumpkin>
mostly just curious for the general shape of the logic
<niksnut>
gchristensen: yes, it deletes paths not reachable from a release
<contrapumpkin>
i.e., do you still retain a notion of GC roots and ask nix for all the nar IDs and subtract those from the total set?
<gchristensen>
ah
<contrapumpkin>
I guess it can be done as a giant iterator over ListBucket, checking if something is reachable from a precomputed set of releases
<niksnut>
pretty much
<gchristensen>
I'm considering making a lifecycle rule which just deletes things older than ~60 days
<gchristensen>
I think that makes nix as a client grumpy, though
<niksnut>
well, that would be dangerous because it could leave holes in closures
<contrapumpkin>
gchristensen: I actually quite like having all sorts of ancient crap in the binary cache :) but I imagine niksnut's wallet (or logicblox's) doesn't
<contrapumpkin>
niksnut: oh, nix doesn't magically fill in holes in closures?
<gchristensen>
contrapumpkin: well ... yeah, and my wallet (prs.nix.gsc.io)
<niksnut>
contrapumpkin: I don't remember what the current behaviour is wrt to fallback to building from source
<contrapumpkin>
that'd be cool
<contrapumpkin>
would allow the S3 to be a true cache
<niksnut>
at the very least you get a message like "don't know how to build XXX"
<contrapumpkin>
ah yes
<contrapumpkin>
I remember seeing that one before
<contrapumpkin>
but using the built-in S3 lifecycle stuff would be cool
<contrapumpkin>
especially for private binary caches
reinzelmann has quit [(Quit: Leaving)]
<gchristensen>
I've seen a nasty failure, don't remember exactly what it was, where nix expected a file to exist in the cache and it didn't, and wouldn't try anything else
<niksnut>
we have a lifecycle rule now to transit to infrequent access
sellout- has quit [(Read error: Connection reset by peer)]
<NixOS_GitHub>
nixpkgs/master fbd69aa Benjamin Staffin: Merge pull request #22321 from pradeepchhetri/minio...
<Ralith>
letting you speak bgp
<cransom>
i'm not a user, but if they don't filter your advertisements to what you are permitted to advertise, that would be insane.
<gchristensen>
you also have to pony up the cash to pay for it
<gchristensen>
and I'm confident they restrict your access, indeed :)
<cransom>
yeah. $100 up front and then a monthly fee if you are doing the global advertisement deal. thats not bad. it is a unique feature that gives you a good reason to not have to run your own network/links/data center
<Ralith>
okay, so that type of filtering is standard then
<cransom>
i assume the only abuse you could run with is flapping. if thats global, everyone dampens your advertisements. if local, it likely doesn't matter much.
pi3r has joined #nixos
<vandenoever>
what's the best way to run rust nightly atm? package rustNightly.rustc fails in master
twopoint718 has joined #nixos
<pikajude>
do you have the sandbox enabled vandenoever
<vandenoever>
pikajude: what sandbox?
<pikajude>
ok, so no
mudri has joined #nixos
<vandenoever>
pikajude: this is the first time i hear about a sandbox in connection with rust
<vandenoever>
pikajude: i'm running nixos with some packages from nixpkgs master
<pikajude>
vandenoever: i mean the sandbox for building nix packages not anything rust specific
<pikajude>
iirc it disables network access and the rustNightly package requires network access because it's a very irresponsible package
<vandenoever>
pikajude: i'm on nixos and i call nix-shell -p rustNightly.rustc
<pikajude>
have you tried using rustNightlyBin?
<vandenoever>
pikajude: just found it and that's compiling prereqs now
<pikajude>
ok
<vandenoever>
pikajude: the non-bin version had problem accessing github, perhaps that was a sandbox restriction
<pikajude>
that is a sandbox restriction yes
Itkovian has joined #nixos
aminechikhaoui has quit [(Quit: leaving)]
<vandenoever>
pikajude: do you know what the restriction is exactly, normally i think it's ok to download data with known checksum
<pikajude>
oh
<pikajude>
well the rustc nightly builder just runs "git clone"
<pikajude>
and i don't think it has a fixed output
<vandenoever>
ah, without a fixed checksum
<pikajude>
normally the source-getting phase and the package building phase are separate
<vandenoever>
you can run clone for a particular sha1
<pikajude>
yea but unless the derivation has a fixed output hash it can't access the network
<vandenoever>
ah
<vandenoever>
nice
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
hamishmack has joined #nixos
eacameron has quit [(Read error: Connection reset by peer)]
nwspk has quit [(Remote host closed the connection)]
nwspk has joined #nixos
jluttine has joined #nixos
Drakonis has joined #nixos
vandenoever has joined #nixos
vandenoever has quit [(Changing host)]
vandenoever has joined #nixos
xwvvvvwx_ has joined #nixos
xwvvvvwx__ has joined #nixos
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
xwvvvvwx has quit [(Ping timeout: 256 seconds)]
xwvvvvwx_ has quit [(Ping timeout: 260 seconds)]
Itkovian has joined #nixos
<NixOS_GitHub>
[nixpkgs] vbgl pushed 1 new commit to master: https://git.io/vDtAD
<NixOS_GitHub>
nixpkgs/master 551c52f Vincent Laporte: feh: 2.17.1 -> 2.18.1
Itkovian has quit [(Client Quit)]
<maurer>
I have systemd.coredump.enable = true set, but this only appears to collect cores coming from the root user
<maurer>
(or possibly only system users)
<maurer>
What do I need to set to collect cores from my normal user too?
rmeis has joined #nixos
<maurer>
(I know I've SIGSEGV'd programs on my normal user, but there are no cores when the user runs coredumpctl list, and when root runs coredumpctl list, it pretty much exclusively lists cores generated during a nix-build procedure
<clever>
maurer: ive found that you still need to "ulimit -c unlimited" when using /proc/sys/kernel/core_pattern
<clever>
and its probably /proc/sys/kernel/core_pattern that systemd is using to capture the cores
Itkovian has joined #nixos
georges-duperon has joined #nixos
<Dezgeg>
sounds like core_pattern is already correct as it captures root's cores
<clever>
and you can run cat on core_pattern to confirm what its doing
<maurer>
btw, is the reason we're not at stable chromium that we have some kind of delayed thing to help hydra, or is it just that nobody's run the update script in a while?
<gchristensen>
FWIW, Hydra has tons of capacity, so please don't hold back on updates on account of hydra
<yorick>
personally, I'm afraid of the update script
<gchristensen>
so, I hope it is because nobody has updated chromium in a while
<maurer>
OK, well, I just ran the update script
danharaj has joined #nixos
<maurer>
if it builds and works locally I'll send a PR
<gchristensen>
great :)
<yorick>
gchristensen: what does that do?
<gchristensen>
yorick: it uses hydra from unstable, but patches unstable first
<yorick>
I currently just build it against 16.09
<gchristensen>
yes, I think you will have a much better experience if you use unstable's hydra + that patch
<yorick>
I'll try, thanks
Ivanych has joined #nixos
danharaj has quit [(Ping timeout: 256 seconds)]
AllanEspinosa has joined #nixos
<yorick>
gchristensen: so are you having fun with the s3 cost?
<pi3r>
Hi, I am trying this code https://paste.ee/p/6ecvE using 'pkgs.haskellPackages.callHackage' and I have got this error
<pi3r>
sed: can't read /nix/store/aqg3rjr6qbabmb0nh9ph6k9ig6hfivxv-all-cabal-hashes-ee101d34ff8bd59897aa2eb0a124bcd3fb47ceec-src/language-puppet/1.3.5/language-puppet.json: No such file or directory
<pi3r>
Any idea ?
<pi3r>
I have been trying this with two pinned recent version of nixpkgs (one in December and another one in January) with no luck
<yorick>
gchristensen: i686 things, it turns out
<gchristensen>
you can configure your x86_64 nodes to build i686 too
athan has quit [(Remote host closed the connection)]
<yorick>
yeah, it does work now
sheenobu has quit [(Quit: Leaving)]
acarrico has quit [(Ping timeout: 255 seconds)]
<pi3r>
maybe callHackage won't work with recent hackage version ? I need to wait nixpkgs to update from Hackage ?
danharaj has joined #nixos
ryantrinkle has joined #nixos
simendsjo has quit [(Quit: ERC (IRC client for Emacs 24.5.2))]
<NixOS_GitHub>
[nixpkgs] peti pushed 1 new commit to master: https://git.io/vDqmy
<NixOS_GitHub>
nixpkgs/master 10349e7 Peter Simons: nixos: drop unused 'haskellPackages' option from ihaskell service...
<dmj`>
pi3r: that haskell package forgot to include language-puppet.json file in the list of its extra source files. So it was never included in the tarball that got sent to hackage.
<dmj`>
You can either submit a patch to the cabal file, hope the author reuploads the new version to hackage, or try to figure out how to get fetchFromGitHub and cabal2nix working together in a nix expression (hairy and ugly)
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<gchristensen>
I have 1 systemd job and when that restarts, I'd like to be certain that 5 other jobs get restarted. How could I do this? right now I have the other "jobs" as ExecStartPost scripts, but want them to run in parallel (thusly thinking I'll make them jobs)
<Yaniel>
learn to stop worrying and love the night(lies)
<Havvy>
Ralith: So, how do we apply this fix to nixpkgs?
<morsicus_>
tilpner, thanks. :)
<Ralith>
Havvy: dump the patch in the package's dir and add it to the package's patches list?
<contrapumpkin>
noo, fetch it if you can :)
<contrapumpkin>
don't junk up the repo with patches we don't own
<contrapumpkin>
I know people do it all the time but I wish they wouldn't
<contrapumpkin>
then they stop being necessary and people always forget to delete them
<maurer>
contrapumpkin: The opposite problem happens with fetched patches sometimes (it did with the ocaml compiler)
<contrapumpkin>
upstream deletes it?
<maurer>
where people point at the fetched patch, it gets merged into upstream, but not released yet, and then the patch gets deleted
<contrapumpkin>
oh, well I'd fetch it from github programmatically
<contrapumpkin>
statically hosted patches are a bit iffy
<maurer>
Yeah, that's what we did, but when the pull request got rebased and closed, the url changed
<maurer>
and then the new url didn't apply
<contrapumpkin>
hah
<contrapumpkin>
ok
<maurer>
I'm not saying you're always wrong
<Havvy>
Well, this patch is just on a mailing list.
<maurer>
I'm just saying it's not black/white for patch fetching vs direct inclusion
<contrapumpkin>
yeah
<morsicus_>
tilpner, i'm starting with NixOs; how can I add this nix expression in my Nix conf ? any documentation about this ?
<Ralith>
contrapumpkin: just leave a clear comment in the package, imo :p
<contrapumpkin>
pfft
<contrapumpkin>
don't make humans do jobs they're bad at :)
<Yaniel>
morsicus_: the readme in the repo has instructions
sellout- has joined #nixos
<Havvy>
morsicus_: Put a copy of rust-nightly.nix into ~/.nixpkgs and then in ~/.nixpkgs/config.nix put packageOverrides = pkgs: { rustNightly = pkgs.callPackage ./rust-nightly.nix {}; }
<Havvy>
At least, that's what I did.
<morsicus_>
Havvy, but this nix expression manage only nightly rust ? I can't switch easily with stable channel, right ?
<Yaniel>
or if you want it system-wide, in /etc/nixos
<Yaniel>
nope
<Havvy>
I've been prodding scott to add support for every stable release of Rust. Though I haven't done that in a while.
<Havvy>
Although that should have pinged them.
<Havvy>
Ralith: I don't see any other patches. Can you add it per-chance?
<morsicus_>
hum. so we can't use rustup on NixOs ?
<Ralith>
Havvy: I'm not a good person to do that, but it's very simple; just write 'patches = [ ./whatever.patch ];' inside the mkDerivation call's args
<Ralith>
morsicus_: the link you were given subsumes rustup
<Yaniel>
morsicus_: well you can use it like on any other platform
<Yaniel>
bit it's heavily advised against
<Yaniel>
as it completely defeats the point of nix
<morsicus_>
Ok... Need to read some documentation about nix expression … ^ ^
<Ralith>
Havvy: or take copumpkin's advice and do some clever dynamic fetch dance instead
<Yaniel>
if rust-nightly-nix gets support for stable releases and more toolchain options it'll get the same functionality as rustup anyway
<contrapumpkin>
damn, did I suggest that?
<Yaniel>
especially if you combine it with shell.nix
<Yaniel>
that is, nix-shell
<Havvy>
Ralith: Yeah, I think leaving a comment saying to remove the patch when updating for the next minor release of R19 would be easier.
jgertm has quit [(Ping timeout: 264 seconds)]
<Havvy>
Because really, there's no file to dynamically fetch.
<Ralith>
Yaniel: <3 shell.nix for keeping my Cargo.lock clippy versions in sync with my compiler versions
<Havvy>
So we could create one, but at that point, that's the same as putting it into nixpkgs repo itself.
<Ralith>
Havvy: well, presumably it's landed in version control *somewhere* by now, and generally you can get VCSs to spit out patches
<Ralith>
and if it hasn't landed in official version control it's probably premature to ship
danharaj has quit [(Ping timeout: 258 seconds)]
ebzzry has joined #nixos
<morsicus_>
Yaniel, Ok, I understand that. I'll read some infos about how Nix works :D
civodul has quit [(Quit: ERC (IRC client for Emacs 25.1.1))]
Wizek has quit [(Ping timeout: 255 seconds)]
<Havvy>
morsicus_: If it helps, rustup is also anathema to how other distro's package management works.
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<MoreTea>
From time to time, I update the hash in the github URL to the newest commit in github.com/nixos/nixpkgs
<MoreTea>
morsicus_, and if you notice that the compiler is out of date, you can just open an issue on GitHub to ask someone to update it
digitus has joined #nixos
Wizek has joined #nixos
fabian_a has quit [(Ping timeout: 255 seconds)]
xwvvvvwx has quit [(Ping timeout: 264 seconds)]
<Acou_Bass>
eey folks, just a query... when i used to use arch on my laptop, i would have the drive automatically decrypt on boot by having the keyfile on a USB stick, so when i stuck the USB stick in and booted, grub would read the key from the stick and decrypt the drive with it. is there any way to replicate this with nixos? it's a fairly basic LUKS setup with encrypted / and unencrypted /boot
mkoenig has joined #nixos
Shou has joined #nixos
data8 has quit [(Quit: nite nite)]
<Acou_Bass>
ook i think ive found it... i would assume this option: boot.initrd.luks.devices.<name?>.keyFile
<Havvy>
MoreTea: Except that pins you on random versions, and if there's an issue on a nightly that gets fixed in the next nightly the next night, you're out of luck.
<morsicus_>
See you tomorrow. thanks for helping me :) Havvy, MoreTea … :)
wkennington has joined #nixos
eacamero_ has joined #nixos
<MoreTea>
@Havvy, very much true. So most of the time I just depend on the latest rustc that is shipped in `unstable`.
<MoreTea>
I just wanted to experiment a bit with procederal macro's that are going to be release in the next version of rust
eacameron has quit [(Ping timeout: 240 seconds)]
eacamero_ has quit [(Ping timeout: 240 seconds)]
stepcut has quit [(Remote host closed the connection)]
mog has joined #nixos
derjohn_mob has joined #nixos
t7 has quit [(Remote host closed the connection)]
djvdorp has quit [(Ping timeout: 252 seconds)]
Sonarpulse has joined #nixos
sellout- has quit [(Ping timeout: 240 seconds)]
AllanEspinosa has quit [(Ping timeout: 248 seconds)]