<{^_^}>
[nixpkgs] @grahamc pushed 3 commits to release-18.03: https://git.io/vpu3V
<{^_^}>
→ 5eae581d by @grahamc: nixpkgs docs: format =)
<{^_^}>
→ 760be190 by @grahamc: nixos docs: ignore generated files
<{^_^}>
→ d3644fc2 by @grahamc: nixos docs: format =)
<samueldr>
awesome
cristian2 has quit [Ping timeout: 240 seconds]
justanotheruser has joined #nixos
<{^_^}>
[nixpkgs] @grahamc closed pull request #39700 → WIP: automagically format the docs → https://git.io/vpBrd
goibhniu has quit [Ping timeout: 248 seconds]
<hodapp>
err, why does org-mode in 18.03 appear to be the 3-years-old 8.2.10?
cristian2 has joined #nixos
<palo>
I have a problem with the hpodder package (a Haskell package) it depends on HaXml 1.20.x, but in nixos only HaXml 1.25.x exist. I try to give hpodder a specific version (via overridesAttrs) but it does not work, I was not even able to install HaXml 1.20.2 because it ignores the overridesAttrs (how I usually do them). These haskellPackages are build in a different way than usuall nixos packages it seems.
<gchristensen>
samueldr: hopefully that makes it nicer to contribute docs ...
<palo>
ah there is something called overrideCabal
<palo>
Yezzz that works (kinda)
thc202 has quit [Ping timeout: 240 seconds]
cristian2 has quit [Ping timeout: 248 seconds]
drakonis has quit [Remote host closed the connection]
<palo>
ajmccluskey: I see. Hoped there was an easier way to do that :( I mean overrideing already existing packages. But thanks, I will try to solve it this way.
<stumble>
Now that I think about it, that warning is useful, although a bit strangely worded!
<gchristensen>
like maybe it should also say "preferring <mychannel> over <rootchannel>?"
<stumble>
Ah yes! That would be a great wording IMO.
<stumble>
I guess it's a little complicated because there could be any number of channels being skipped.
<gchristensen>
I _think_ here it is guaranteed to top out at 2
hamishmack has joined #nixos
ryantrinkle has quit [Ping timeout: 264 seconds]
<gchristensen>
oh I see, yeah you're right
<stumble>
Hm.
jrolfs has quit [Ping timeout: 264 seconds]
<stumble>
I think the problem is that it's a general error message at a relatively low level, but it's being triggered very frequently (I presume) by what must be a very common way of doing things, which is to have exactly two channels in use, one for root and one for an unprivileged user.
<gchristensen>
it might be
<gchristensen>
you could instead, for example, create a channel as your user named nixpkgs, and avoid the error I think
<stumble>
Ah!
<stumble>
Let's see what the docs recommend ...
<gchristensen>
out of curiosity, which docs are you checking?
<stumble>
I guess "nixpkgs" must be a hard-coded default.
<gchristensen>
some things are harder to improve the docs on than others :P
<stumble>
So in fact the warning was alerting me to an error I'd made, which was to name two channels with the same name for no good reason. I thought "nixos" was the defaut but I was wrong.
<stumble>
Yes!
<samueldr>
> If name is omitted, it defaults to the last component of url, with the suffixes -stable or -unstable removed.
<stumble>
Oh, OK, good.
<samueldr>
(from the manpage)
<samueldr>
I knew I read it somewhere :)
<stumble>
:-)
Arcaelyx has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
lektrik has left #nixos ["Leaving"]
stumble- has joined #nixos
ericsagnes has joined #nixos
stumble- has quit [Client Quit]
stumble has quit [Ping timeout: 240 seconds]
semilattice has joined #nixos
jason has joined #nixos
jason is now known as stumble
<ajmccluskey>
I think I found my issue. It seems that `nix.binaryCaches` used to append entries to the default(s) (i.e. `cache.nixos.org`), but now setting that overrides the whole list.
tyrus82 has quit [Quit: tyrus82]
tyrus82 has joined #nixos
<stumble>
That's an interesting change!
<gchristensen>
a good change
<stumble>
Because it's purer?
nur0n0 has joined #nixos
<gchristensen>
because it is good to be able to say "don't talk to cache.nixos.org"
<gchristensen>
for example, people who don't want to trust it
<gchristensen>
anyway, bed time. night!
<elvishjerricco>
gchristensen: Meh. I think it violates the principle of least astonishment. Might have been better to just have a different option to disable that
<stumble>
Good night.
Sonarpulse_ is now known as Sonarpulse
<hodapp>
something needs libgfortran.so.3, and I see it in (e.g.) /nix/store/3m8wl9f0jq3w6xb94br7dnnhi3jdl2h4-gfortran-5.4.0-lib/lib/libgfortran.so.3 but how do I go about getting at this path from something like shell.nix??
<hodapp>
the package gfortran doesn't seem to be the correct one
<ajmccluskey>
All for providing an easy way to not use the default cache(s), but changing that behaviour hurt.
<ryantm>
hodapp: try gfortran.lib
<ajmccluskey>
Didn't see it mentioned in the release notes, but that might be on me.
<ryantm>
hodapp: I see gfortran.cc.lib a bunch in nixpkgs
<ajmccluskey>
Upgrade of packages was very fast and seemingly successful now. All in all I'm still a very happy NixOS user. Just in case anyone took my comments as being whiney :)
<hodapp>
ryantm: that's the incorrect version, libgfortran.so.4 rather than .3
colemickens has joined #nixos
<ryantm>
all-packages.nix has gfortran48, 49, 5, 6, 7
<hodapp>
ooh...
colemickens has quit [Client Quit]
colemickens has joined #nixos
<hodapp>
didn't realize those were there since none were showing up in searches
spacefrogg has quit [Remote host closed the connection]
aw has quit [Quit: Quitting.]
aw has joined #nixos
spacefrogg has joined #nixos
<{^_^}>
[nixpkgs] @Ericson2314 opened pull request #39844 → cctools, apple libiconv: Apple clean up for iOS cross → https://git.io/vpu82
semilattice has quit [Ping timeout: 276 seconds]
semilattice has joined #nixos
drakonis_ has quit [Read error: Connection reset by peer]
drakonis has quit [Read error: Connection reset by peer]
gmarmstrong has joined #nixos
gmarmstrong has quit [Remote host closed the connection]
schoppenhauer has quit [Read error: Connection reset by peer]
schoppenhauer has joined #nixos
jtojnar has quit [Ping timeout: 260 seconds]
justan0theruser has joined #nixos
semilattice has quit [Ping timeout: 264 seconds]
justanotheruser has quit [Ping timeout: 260 seconds]
<justanotheruser>
struggling finding this in the docs: What's the best way to set environment variables, not for build, but to be used in the environment?
<jackdk>
around a particular program that you're building? you could create a wrapper?
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
[Leary] has joined #nixos
rauno has joined #nixos
jedai|2 has joined #nixos
<teto1>
I want to override preConfigure while having autoreconfHook as a buildInput. It seems like the autoreconf is called before which generates an error because of a missing file ? Is there a workaround ?
xy2_ has joined #nixos
jedai has quit [Ping timeout: 255 seconds]
DavidEGr1yson has quit [Remote host closed the connection]
spear2 has quit [Ping timeout: 240 seconds]
Ariakenom has joined #nixos
spear2 has joined #nixos
hyper_ch2 has joined #nixos
siel has quit [Ping timeout: 245 seconds]
robogoat has quit [Ping timeout: 264 seconds]
siel has joined #nixos
robogoat has joined #nixos
xy2_ has quit [Ping timeout: 256 seconds]
hamishmack has joined #nixos
siel has quit [Ping timeout: 256 seconds]
ericsagnes has joined #nixos
justbeingglad has joined #nixos
semilattice has quit [Ping timeout: 260 seconds]
krey has joined #nixos
stephenjudkins has quit [Remote host closed the connection]
stephenjudkins has joined #nixos
siel has joined #nixos
wchresta has quit [Remote host closed the connection]
<teto1>
I rewrote the postUnpack phase and it works so that will be it I guess :)
<{^_^}>
→ b95a24fd by @peterhoeg: nixos-container: support restarts
<{^_^}>
→ 236a9189 by @peterhoeg: Merge pull request #39717 from peterhoeg/f/container
dvim has joined #nixos
jackdk has quit [Ping timeout: 240 seconds]
hyper_ch2 has quit [Ping timeout: 260 seconds]
thc202 has joined #nixos
goibhniu has joined #nixos
hyper_ch2 has joined #nixos
Shados_ has joined #nixos
Shados has quit [Read error: Connection reset by peer]
jaeckel has joined #nixos
<Myrl-saki>
clever: I think I'll just use qemu-i386 with Nix. Too much hassle at this point.
jacob_ has joined #nixos
civodul has joined #nixos
abbradar has joined #nixos
ericsagnes has quit [Ping timeout: 264 seconds]
hyper_ch2 has quit [Quit: Page closed]
<abbradar>
Hi all! I have a server with 2GB of RAM and 1GB zram on which I want to `nix-build -A ghc` -- but get "out of memory". Memory usage indeed creeps up when `nix-build` is working; it doesn't try to use swap at all though. Any ideas what can I do with this? It's overall strange that Nix seemingly loads full derivation into memory...
siel has quit [Ping timeout: 240 seconds]
<woffs>
I also experienced things like this, I had to close all other applications go get it working. Was no problem in 17.09
<rawtaz>
abbradar: i have *no* idea really, i dont even know what compilation goes on there, but is there any way you can limit it to e.g. one process at a time (in case it currently uses multiple)?
<rawtaz>
i dont suppose there's any garbage collection involved that you can tweak to be more aggressive?
MichaelRaskin has left #nixos [#nixos]
TimePath has quit [Ping timeout: 245 seconds]
slabity[m] has quit [Ping timeout: 245 seconds]
jack[m] has quit [Ping timeout: 245 seconds]
dpc has quit [Ping timeout: 245 seconds]
trikl[m] has quit [Ping timeout: 245 seconds]
telent[m] has quit [Ping timeout: 245 seconds]
spacekitteh[m] has quit [Ping timeout: 245 seconds]
<abbradar>
rawtaz: I don't think it's during eval so no GC should be involved (I'm not sure though, maybe nix-daemon uses libgc too). No compilation runs, it's Nix fetching cached derivation.
Kallegro[m] has quit [Ping timeout: 245 seconds]
wervenyt[m] has quit [Ping timeout: 245 seconds]
sphalerit has quit [Ping timeout: 245 seconds]
bdimcheff has quit [Ping timeout: 245 seconds]
florianjacob has quit [Ping timeout: 245 seconds]
bendlas has quit [Ping timeout: 245 seconds]
<abbradar>
I have several people reporting it to me before and experience it myself but that time it was more difficult to reproduce
Dezgeg[m] has quit [Ping timeout: 245 seconds]
yochai[m] has quit [Ping timeout: 245 seconds]
uvnikita[m] has quit [Ping timeout: 245 seconds]
revoltmedia[m] has quit [Ping timeout: 245 seconds]
<abbradar>
The server has just 60mb RAM used now so I don't have anything to close :D
Khorne[m] has quit [Ping timeout: 245 seconds]
lukeadams[m] has quit [Ping timeout: 245 seconds]
baconicsynergy[m has quit [Ping timeout: 245 seconds]
wzy8L-B2[m] has quit [Ping timeout: 245 seconds]
scott1 has quit [Ping timeout: 245 seconds]
wilornel has quit [Ping timeout: 245 seconds]
nh2[m] has quit [Ping timeout: 245 seconds]
lord| has quit [Ping timeout: 240 seconds]
TimePath has joined #nixos
revoltmedia[m] has joined #nixos
yochai[m] has joined #nixos
Dezgeg[m] has joined #nixos
uvnikita[m] has joined #nixos
lukeadams[m] has joined #nixos
baconicsynergy[m has joined #nixos
wilornel has joined #nixos
Kallegro[m] has joined #nixos
spacekitteh[m] has joined #nixos
wzy8L-B2[m] has joined #nixos
Khorne[m] has joined #nixos
nh2[m] has joined #nixos
scott1 has joined #nixos
trikl[m] has joined #nixos
telent[m] has joined #nixos
slabity[m] has joined #nixos
jack[m] has joined #nixos
dpc has joined #nixos
bdimcheff has joined #nixos
wervenyt[m] has joined #nixos
sphalerit has joined #nixos
bendlas has joined #nixos
florianjacob has joined #nixos
hiberno has quit [Quit: WeeChat 1.6]
hyper_ch2 has joined #nixos
peacememories has joined #nixos
peacememories has quit [Read error: Connection reset by peer]
goibhniu has quit [Ping timeout: 260 seconds]
peacememories has joined #nixos
goibhniu has joined #nixos
Jetien_ has joined #nixos
periklis has quit [Remote host closed the connection]
iyzsong has joined #nixos
siel has joined #nixos
periklis has joined #nixos
<woffs>
abbradar, you need a separate build server :-/
<{^_^}>
→ a0dd9713 by @peterhoeg: vlc: 2.2.8 -> 3.0.1
<{^_^}>
→ 490e6fc9 by @peterhoeg: vlc_npapi: drop it as firefox no longer support NPAPI
<{^_^}>
→ a3f5c424 by @vaibhavsagar: phonon-backend-vlc: 0.9.0 -> 0.10.1
Ariakenom has joined #nixos
<symphorien>
arianvp2: either shellHook for arbitrary shell commands, or just define your env variable as an appribute to the argument of mkDerivation
<symphorien>
supernanovoid: NIX_CFLAGS_COMPILE="-Werror" or makeFlags="CFLAGS=-Werror" or any other variable the makefile/underlying build system will understand
JosW has joined #nixos
<supernanovoid>
Thanks!
niklob has joined #nixos
joehh has joined #nixos
wacke has joined #nixos
<supernanovoid>
But can i disable -Werror=format-security somehow, using cmake?
<wacke>
hi all, i have services.acpid enabled and also defined one handler for xrandr, but the xrandr command is not executed, my log file is empty, someone an idea? relevant config part: https://pastebin.com/8eC8BK4L
<{^_^}>
[nix] @edolstra pushed to master « Fix bzip2 compression of files > 4 GiB »: https://git.io/vpudF
tyrus82 has quit [Quit: tyrus82]
<wacke>
ps: system.stateVersion = "18.03";
<arianvp2>
symphorien: oh attributes. of course. thanks
<supernanovoid>
I have some problem again: fatal error: SDL_mixer.h: No such file or directory #include "SDL_mixer.h". But SDL2_mixer is in buildInputs :(
seanparsons has quit [Read error: Connection reset by peer]
<Akii>
Hi! I would like to get this modeller working https://camunda.com/download/modeler/ -- it says "Linux" but I think they meant "Ubuntu".. if I'm interpreting the output of `ldd` correct I'm missing a whole bunch of stuff
<Akii>
kinda sounds like it would be an awesome fit for Nix to just resolve those dependencies but no idea if that's possible
<Akii>
I also lack the linux-fu for that
<srhb>
nikivi: Haskell is an excellent language for being principled about _other_ languages. And the people involved are fairly big-time Haskellers, so probably also because it's just nicer to work in a language they know, rather than C++/perl
kahiru has quit [Remote host closed the connection]
knupfer has quit [Remote host closed the connection]
knupfer has joined #nixos
kahiru has joined #nixos
Henson has joined #nixos
mmlb has quit [Read error: Connection reset by peer]
mmlb has joined #nixos
slyfox_ has joined #nixos
<__monty__>
I have a myEnvFun for my haskell dev environment, building it failed, this is the tail of the log (history didn't go back further): http://ix.io/19ef
kisik21 has quit [Ping timeout: 264 seconds]
<__monty__>
Does this mean my machine just doesn't have sufficient memory? Any ways around this?
<__monty__>
And should I report this as a ghc bug like the error says?
kisik21 has joined #nixos
slyfox has quit [Ping timeout: 248 seconds]
mmlb has quit [Read error: Connection reset by peer]
<avn>
__monty__: does it git master? If so -- possible you have `stylish` in your env
<__monty__>
Maybe I'll have to stick to nix's haskellPackages.
<avn>
__monty__: then idk. My build breaks on stylish atm.
<__monty__>
avn: Already tried throwing dontCheck or doJailbreak at it?
<sphalerite>
peterhoeg: will you backport the syncthing update to 18.03 as well?
hlolli has joined #nixos
<andrewmiller1>
__monty__: report the error to the developer who committed the error handling or the project with the error handling that gave you the error output
<avn>
__monty__: not yet, but possible dontCheck should be enough.
<{^_^}>
[nixpkgs] @dotlambda pushed commit from @r-ryantm to master « abcmidi: 2018.04.18 -> 2018.04.24 (#39796) »: https://git.io/vpzTE
<andrewmiller1>
__monty__: if i understood your question correctly
phreedom has quit [Ping timeout: 255 seconds]
kisik21 has quit [Ping timeout: 276 seconds]
kisik21 has joined #nixos
<__monty__>
andrewmiller1: That'd be GHC but I'm not sure this is actually a bug in GHC, maybe 4G is just too little memory to have it build all those packages.
<andrewmiller1>
__monty__: the developers for the zig programming language might be experts on memory problems too
blankhart has quit [Ping timeout: 264 seconds]
<andrewmiller1>
__monty__: though i understand zig has nothing to do with it. it might be easy for them to consult you on the problem
<hlolli>
What could cause my xrandr from not detecting my HDMI output? I have intel/nvidia drivers working fine via bumblebee.
<hlolli>
xrandr --listmonitors ~
<hlolli>
Monitors: 1
<hlolli>
0: +*eDP1 1920/340x1080/190+0+0 eDP1
<zybell>
__monty__:Nix does all .o in RAM iirc . When you have linked (and dont need them later again)rm ing a few could help.At least it would nail the error to mem if you got farther because of that step.
<__monty__>
zybell: Where do I find the .o's to remove?
orivej has joined #nixos
<zybell>
the paths are in the log you posted
phreedom has joined #nixos
MercurialAlchemi has quit [Ping timeout: 260 seconds]
<sphalerite>
__monty__: do you have any swap? Adding some would probably help. You could also try reducing the number of simultaneous jobs and build cores, if it's not already 1
simpson has joined #nixos
<zybell>
sphalerite:tmpfs on swap?
<sphalerite>
zybell: what's tmpfs got to do with it?
<zybell>
I think he does a tmpfs build
kisik21 has quit [Ping timeout: 276 seconds]
<wchresta_>
There is no lsb-release on darwin. Does anybody have an example of a derivation where the linux version uses lsb-release but was circumvented for darwin?
freeman42x]NixOS has joined #nixos
<sphalerite>
zybell: that's not default
<sphalerite>
__monty__: but yes, if you have boot.tmpOnTmpfs enabled, that's not a good idea for nix on so little RAM
<sphalerite>
unless you have plenty of swap to back it up I guess
kisik21 has joined #nixos
<__monty__>
I just ran nix-env -iA nixpgks.ghc82Env locally on an arch system (not nixos).
<sphalerite>
ah ok. Do you have any swap?
<andrewmiller1>
__monty__: the idea of where to open the issue isn't to be where the bug is - it's to get information on your next steps. if you know where the bug is, then of course open it there. but if you're still unsure, the one who coded the error should be able to tell you it's not them and probably <insert their advice here>. so take my advice as a safety net if you end up still stuck
<{^_^}>
→ 11a244af by @dotlambda: vulnix: drop obsolete patch
odi has joined #nixos
krey has quit [Ping timeout: 256 seconds]
<__monty__>
sphalerite: No swap no, I'll try adding some but it's spinning metal so I imagine a new build to test will take another half a day.
<sphalerite>
__monty__: if you're lucky it won't be using it most of the time
<__monty__>
Just 1/1 ram/swap or is that an old rule of thumb?
<andrewmiller1>
__monty__: i think it's 8gb if you don't even have that much ram
<mkaito>
is anyone working on improving Emacs' nix-mode indentation?
Ariakenom has quit [Ping timeout: 240 seconds]
<gchristensen>
I am constantly improving its indentation, but at the output level, not at the logic level ;)
<gchristensen>
:(
<mkaito>
I feel bad saying this, but it's just kinda bad. I have to manually indent nearly every line. Which is frustrating when I have to write nix full-time.
<mkaito>
then again, indentation in Emacs is black magic, so at least it does *something* I guess
<andrewmiller1>
oh the question mark at the end didn't break it. i guess that makes sense
jperras has joined #nixos
<mkaito>
huh that's neat
<hlolli>
the nix-mode indentation should be fixed, I opened a ticket about it last week, hope to make a PR in next weeks to fix it. It's not really black magic, one just has to write elisp code that looks backwards and counts the nesting levels.
<andrewmiller1>
mkaito: our age, if you have to do a lot of work manually, you're probably missing out on current solutions
<{^_^}>
→ bb122770 by @veprbl: pythonPackages.pyslurm: 20170302 -> 20180427
<{^_^}>
→ ac75ebf1 by @dotlambda: Merge pull request #39694 from veprbl/pr/pyslurm_fix_for_17.11.5
Ariakenom has joined #nixos
<andrewmiller1>
gchristensen: why not make here docs?
<andrewmiller1>
gchristensen: maybe jq in between
<andrewmiller1>
gchristensen: depending on what output you're dealing with
<mkaito>
I'm just going to use vim until nix-mode behaves properly. Writing any Elisp beyond the necessary for configuration is not something I'm very good at anyway.
<mkaito>
I'm happy to see that someone's working on it :)
<andrewmiller1>
mkaito: you can also check out kakoune or micro
<mkaito>
huh I had never seen kakoune
rauno has joined #nixos
ericsagnes has joined #nixos
<gchristensen>
andrewmiller1: huh?
<andrewmiller1>
mkaito: in micro you can use the command line while in the editor. you open a little line at the bottom on a hotkey
<andrewmiller1>
gchristensen: "I am constantly improving its indentation, but at the output level, not at the logic level"
<mkaito>
I'm actually not sure what he means with output/logic
<gchristensen>
I mean I'm not fixing the mode, I'm just manually reindenting
<andrewmiller1>
gchristensen: ah i see
<mkaito>
well yeah, that's what I do. All the damn time :P hence, complaining, because complaining > fixing :P
<gchristensen>
:d
lord| has quit [Quit: WeeChat 2.1]
colemickens has quit [Quit: Connection closed for inactivity]
<mkaito>
then I looked briefly into "writing indentation for a major mode" and just walked away shaking my head.
<{^_^}>
[nixpkgs] @Ericson2314 merged pull request #39844 → cctools, apple libiconv: Apple clean up for iOS cross → https://git.io/vpu82
<{^_^}>
[nixpkgs] @Ericson2314 pushed 3 commits to staging: https://git.io/vpzt9
<{^_^}>
→ 4b3cc9ed by @Ericson2314: cctools: Cleanup `useOld` conditions slightly
<{^_^}>
→ 62fd669a by @Ericson2314: darwin.libiconv: Start to fix for cross
<mkaito>
I love emacs, except I hate emacs.
<{^_^}>
→ c377b21f by @Ericson2314: Merge pull request #39844 from obsidiansystems/cctools-cleanup
<gchristensen>
mkaito: most of #emacs is the same way
ericsagnes has quit [Ping timeout: 265 seconds]
aarvar has quit [Ping timeout: 268 seconds]
shabius has quit [Quit: Leaving]
<mkaito>
they shake their heads a lot?
Ben_______ has joined #nixos
joehh has quit [Ping timeout: 264 seconds]
shabius has joined #nixos
kisik21 has quit [Ping timeout: 260 seconds]
<__monty__>
andrewmiller1: What's special about this in micro? Vim does the same, no? And I'd be baffled if emacs didn't.
iqubic` has left #nixos ["ERC (IRC client for Emacs 25.3.1)"]
<sphalerite>
__monty__: yes, M-! is like vim's :! and M-& can run stuff in the background as well
<andrewmiller1>
__monty__: not special, just could be an overlooked because it feels more like a mix of nano and notepad++, which neither have such functionalities
Ariakenom_ has joined #nixos
chimay has joined #nixos
<{^_^}>
[nixpkgs] @WilliButz opened pull request #39856 → prometheus-nginx-exporter: fix bool to string coercion → https://git.io/vpzOj
Ariakenom has quit [Ping timeout: 264 seconds]
shabius has quit [Quit: Leaving]
wangoe has joined #nixos
shabius has joined #nixos
wangoe has quit [Client Quit]
raynold has quit [Quit: Connection closed for inactivity]
wangoe has joined #nixos
jethroksy has quit [Remote host closed the connection]
MercurialAlchemi has joined #nixos
shabius has quit [Client Quit]
andrewmiller1 has quit [Ping timeout: 260 seconds]
<{^_^}>
[nixpkgs] @globin pushed commit from @WilliButz to master « prometheus-nginx-exporter: fix bool to string coercion »: https://git.io/vpzsQ
erasmas has joined #nixos
odi_ has quit [Remote host closed the connection]
odi has joined #nixos
MercurialAlchemi has quit [Ping timeout: 248 seconds]
Ariakenom_ has quit [Quit: Leaving]
Ariakenom_ has joined #nixos
<azahi>
is it possible to run xorg with xinit?
<azahi>
without using display manager
<Ankhers>
I'm attempting to use nixops to update a deployment. I am getting the following error: `attribute 'nixosVersion' missing, at /nix/store/b5rnlvijmca1zd8q2a4fmarjb1s00jni-nixops-1.5.2/share/nix/nixops/ec2.nix:153:41`. Is `nixosVersion` something that I should be adding to my derivation?
arianvp2 has quit [Quit: Page closed]
<aminechikhaoui>
Ankhers: which version of nixops ?
<zybell>
azahi yes startx is a script that shows how. (or use startx directly)
fuzzy-id has quit [Ping timeout: 255 seconds]
<aminechikhaoui>
Ankhers: oh 1.5.2, can you check with 1.6, I think it's fixed there
<Ankhers>
aminechikhaoui: Will I need to build it from source?
freeman42x[nix] has joined #nixos
drakonis has joined #nixos
<azahi>
aminechikhaoui: I know how to run that stuff on other distros but on NixOS I'm experiencing that issue in the end of the log file: http://ix.io/19eC
odi has quit [Ping timeout: 265 seconds]
<azahi>
.xinitc is only has "exec xmonad" in it
<azahi>
I can post my configuration.nix if that's necessary
<goibhniu>
I don't think you can run startx directly on NixOS
<aminechikhaoui>
Ankhers: no just use the 18.03 channel I guess
dbmikus__ has joined #nixos
krey has joined #nixos
odi has joined #nixos
<Ankhers>
aminechikhaoui: That fixed it. Thanks.
<aminechikhaoui>
cool
Ariakenom_ has quit [Ping timeout: 248 seconds]
<gchristensen>
disasm: 'round?
selaux has quit [Ping timeout: 256 seconds]
xy2_ has joined #nixos
<Ben_______>
What's the best way to explore current configuration tree interactively, possibly also discovering available options (like available window managers and such)?
xAFFE has left #nixos ["Error from remote client"]
fendor has joined #nixos
<drakonis>
ah hell, i don't have wineWow as my default anymore
<fendor>
hi, I opened an issue about stack, and I received the answer that it seems to be an issue with NixOs, or rather not an issue at all. Can someone look over my issue and tell me, if he is right? https://github.com/commercialhaskell/stack/issues/3961
wchresta_ has quit [Ping timeout: 276 seconds]
logzet has quit [Remote host closed the connection]
<mkaito>
fendor: sounds more like it's an issue with the nix part of stack, rather than nixos itself.
Ariakenom has joined #nixos
* hodapp
stabs CUDA in the face
<shapr>
bara CUDA bites back!
<hodapp>
that pun hurt me
<shapr>
death of a thousand parallel cuts
oida has quit [Ping timeout: 255 seconds]
civodul has quit [Quit: ERC (IRC client for Emacs 25.3.1)]
<hodapp>
owwwww
<{^_^}>
[nixpkgs] @Ma27 opened pull request #39860 → qscintilla: revert bump to fix the build → https://git.io/vpzBb
kreisys has quit [Read error: Connection reset by peer]
ryantrinkle has joined #nixos
griff_ has quit [Quit: griff_]
__monty__ has quit [Read error: Connection reset by peer]
winem_ has quit [Ping timeout: 260 seconds]
periklis has joined #nixos
jrolfs has joined #nixos
<Ben_______>
Hey, how can I activate a user that's define in configuration.nix? I've declared a user "ben" via the extrausers option, but when i run "passwd ben" as root, it says "user does not exist".
Tucky has quit [Quit: WeeChat 2.1]
<Ben_______>
What am I missing?
<tilpner>
Have you run "nixos-rebuild switch"?
<Ben_______>
Yes
<tilpner>
And "cat /etc/passwd" lists ben?
__monty__ has joined #nixos
<__monty__>
I'm getting this warning when running nix-env: http://ix.io/13Vc The locale in question is not missing, what could be the actual problem?
<Ben_______>
Yes, but as "Ben". I had set users.extrausers.ben.name = "Ben".
<iqubic>
I'm just supposed to modify the .XResources file? No settings I can change in configuration.nix?
<mkaito>
srhb: I feel like that should be in the manual
<srhb>
mkaito: Hmm, possibly. There's so many ways to do it though. It's a programming language after all.
<tmplt>
srhb: oh neat. I suppose derivations in the overlay are specified via overlays.dervinationName then?
<srhb>
tmplt: Uh, never heard of that.
<mkaito>
srhb: I'm sure an example section would be fine though
<srhb>
mkaito: Yeah.
<iqubic>
Trying to download anything on school wifi is SLOOOOW. ASpell + 3 dictionaries are going to take a while to download for me.
<iqubic>
Actually, no. They didn't.
<tmplt>
So, a library I need produce some so-files with undefined references to some GLIBC_2.27 functions. This is not the case when building the library with the PKGBUILD in Arch Linux. Could this be caused by something in the library's derivation?
<tmplt>
The PKGBUILD does have some CMake options enabled for packages not yet in nixpkgs, but before writing derivations for those I figured I'd check if it was some known issue first.
<{^_^}>
[nixpkgs] @matthewbauer pushed 4 commits to master: https://git.io/vpzPJ
<{^_^}>
→ caec3b9e by @Twey: PyQt5: 5.10.0 -> 5.10.1
<{^_^}>
→ fc7ed300 by @Twey: plover: 3.1.0 -> 3.1.1
<{^_^}>
→ 51a8848f by @Twey: plover: add 4.0.0.dev6
<__monty__>
sphalerite: Thanks btw, I only had 2G to spare for swap but I guess it was enough, or maybe the updated channel had some fixes.
<{^_^}>
[nixpkgs] @matthewbauer merged pull request #28206 → switch-to-configuration: use Net::DBus to retrieve the list of units → https://git.io/v7Hcv
<{^_^}>
[nixpkgs] @matthewbauer pushed 2 commits to master: https://git.io/vpzPq
<{^_^}>
→ 2fa1165e by @edef1c: switch-to-configuration: use Net::DBus to retrieve the list of units
<{^_^}>
→ e508f0ee by @matthewbauer: Merge pull request #28206 from edef1c/net-dbus-list-units
asuryawanshi has joined #nixos
shabius has quit [Quit: Leaving]
muzzy has joined #nixos
<muzzy>
Hey is nixos.org down for anybody else?
<{^_^}>
[nixpkgs] @joncojonathan opened pull request #39873 → PhpStorm: from 2017.3.6 to 2018.1.2 → https://git.io/vpzXc
UNIcodeX_ has joined #nixos
<makefu>
muzzy: working for me
<UNIcodeX_>
I've heard that NixOS does a mirrored /boot for a mirrored install. Does NixOS use ZFS as FS or btrfs or other?
jrolfs has joined #nixos
<symphorien>
you have the choice
<etu>
UNIcodeX_: NixOS use whatever FS you want it to use
<{^_^}>
[nixpkgs] @xeji merged pull request #39863 → Fix typo in help message → https://git.io/vpzuN
<{^_^}>
[nixpkgs] @xeji pushed commit from @alexfmpe to master « terraform/providers/update-all: Fix typo in help message (#39863) »: https://git.io/vpzXX
d6e has quit [Quit: WeeChat 2.0.1]
<etu>
muzzy: works for me
d6e has joined #nixos
<{^_^}>
[nixpkgs] @xeji pushed commit from @rszibele to master « unixODBCDrivers.mariadb: fix library file reference required for ODBC (#39841) »: https://git.io/vpz1v
<{^_^}>
→ 1eea1d6b by @xeji: Merge pull request #39870 from bachp/gitlab-runner-10.7.1
<clever>
boot nixos form usb, mount the rootfs, run nixos-enter, nixos-rebuild boot
<clever>
after fixing /boot and the config that refered to it
<maurer>
UNIcodeX_: I have a crypto layer for part of my home directory but yeah, no proper crypto
<UNIcodeX_>
My first exposure to ZFS was with testing TrueOS and FreeNAS. I'm hooked on it now, but I like and am more familiar with linux land
<UNIcodeX_>
NixOS has a similar feature to TrueOS that they call "boot environments" where you can boot into a previous state. So that's what piqued my interest to come in here
<UNIcodeX_>
what about BTRFS? I don't find it as intuitive as ZFS
<clever>
i find zfs better then btrfs
<__monty__>
Any reasons that apply to laptops?
ixxie has joined #nixos
<UNIcodeX_>
when I was reading about / testing btrfs snapshot ability, I didn't like how you had to create a different folder / dataset manually for snapshots... It just felt weird. Maybe I misunderstood it. ZFS is much more clean... zfs snapshot dataset@name ... done
<maurer>
The biggest issue for me with btrfs is that if you get anywhere near a full disk
<maurer>
1.) Actually hitting full is unpredictable
<maurer>
2.) If you do hit full, it's hard to save yourself
seanparsons has joined #nixos
<iqubic>
I use zfs, but I don't use any of the special zfs features.
<UNIcodeX_>
I'm worried that due to licensing and other factors, that btrfs is being pushed more than ZoL... a whole betamax (ZFS) vs VHS (BTRFS) thing...
<__monty__>
maurer: I know that used to be true but are you sure it still is?
<avn>
iqubic: even compression?
<__monty__>
I haven't run into problems.
<iqubic>
I'm not sure if I use compression. How do I check that?
<__monty__>
I don't agree ZFS is the betamax in that comparison : )
<goibhniu>
wolftune: and `nix-env -iA nixos.gitAndTools.gitFull` will "install" it for your user
<clever>
yeah, docker requires config changes and a rebuild
<wolftune>
tilpner: I was hoping that would be it, but I do then need to run a `./launcher… rebuild app` command for a docker thing
<wolftune>
so it's not just `nix-shell -p docker` ?
<tilpner>
We can't tell for sure
<clever>
wolftune: what is the contents of launcher?
<tilpner>
It might be installed and running already, or it might need to be installed
<wolftune>
to be specific, it's a container for a Discourse forum
<wolftune>
the forum is running
<wolftune>
I'm trying to update it
__monty__ has quit [Quit: leaving]
<wolftune>
is it safe to just try `nix-shell -p docker` then `./launched rebuild app` ?
<wolftune>
wouldn't that just say not available and fail or just work probably?
<clever>
wolftune: it would likely fail, try just running the launcher script without docker first and see what fails
__monty__ has joined #nixos
<wolftune>
clever: I get a bunch of `which…` lines with one error at the end: "Docker is not installed, you will need to install Docker in order to run Launcher"
<clever>
wolftune: are there any .nix files in that directory?
<wolftune>
oh it looks like I need three things: docker.io docker and git
<wolftune>
clever: I don't see any .nix files in the directory where I'm working
<clever>
it sounds like this machine has never built that container before
<clever>
so i'm wondering if there is some special step the old admin forgot to mention
<wolftune>
it pretty surely has built it…
<wolftune>
but I wasn't the one doing it, so I can't say exactly
<iqubic>
Clever: How's that zfs setup gist coming along?
<__monty__>
wolftune: If it had then everything you need would've been available, no? Sounds like the other admin built the container locally then deployed it to the server.
<{^_^}>
[nixpkgs] @dtzWill merged pull request #39877 → ltrace: fix warnings and bugs using patches from debian → https://git.io/vpzDP
<gchristensen>
iqubic: presumably you want to use ZFS because it is good at keeping your data healthy... yeah?
<iqubic>
Yeah.
<iqubic>
That's right.
<hedning[m]>
jtojnar: there, tried freenode first, and matrix is being slow
<gchristensen>
iqubic: then having it perform its own maintenance to identify and (try to) resolve problems is a good part of that
<clever>
bbl
acarrico has quit [Ping timeout: 268 seconds]
<__monty__>
Hmm, I installed dunst but the dbus service file that should be included is missing. Is it only provided on nixos systems?
<iqubic>
Oh, I see. I'll enable that straight away.
<maurer>
__monty__: Files that'd normally land in /etc/ won't show up on non-nixos
<rawtaz>
gchristensen: really nice article :)
jrolfs_ has quit [Ping timeout: 268 seconds]
<rawtaz>
gchristensen: did you write it?
<__monty__>
maurer: It's /usr/share/... in this case but I guess the same applies?
* rawtaz
&
<gchristensen>
rawtaz: no
<iqubic>
Now, do I really want compression on ALL my data. "/", "/home/iqubic" and "/nix" are all compressed with lz4 compression. Is that fine?
<gchristensen>
iqubic: I turn it on everywhere I'm not storing a ton of poorly-compressable data
<maurer>
__monty__: Basically, building and installing a nix package only writes to the nix store
<__monty__>
Does zfs support zstd compression?
<maurer>
__monty__: Your nix-profile links to a spot in the nix store for per-user stuff
<gchristensen>
* you probably don't want to use dedup fwiw
<maurer>
__monty__: "System" tier stuff only happens on nixos-rebuild
<maurer>
So you may well find that a file showing up in /usr/share normally is at .nix-profile/share or similar
<maurer>
but your system dbus won't know to look there
<iqubic>
gchristensen: Is there a way to check if I'm using dedup?
<gchristensen>
you'd have to have turned it on explicitly
<maurer>
iqubic: You need to turn dedup on, it's not default
<maurer>
iqubic: A lot of people turn it on because they hear it's a cool feature
supernanovoid has joined #nixos
<gchristensen>
and then one day they can't mount their disk anymore
<maurer>
but unless your data is _heavily_ redundant, it actually just eats all your RAM
<steveeJ>
offlinehacker[m]: is there any chance you could have a look at the latest kubernetes udpate PR? https://github.com/NixOS/nixpkgs/pull/38445 the tests are failing and there's a lack of kubernetes skills :-D
Ariakenom has quit [Quit: Leaving]
<iqubic>
I may have done that. I'm not sure. Its been a while since I've looked at my zfs setup.
<{^_^}>
→ 73684872 by @nlewo: dockerTools.pullImage: Skopeo pulls images by digest
<{^_^}>
→ d35dcb12 by @nlewo: dockerTools.pullImage: documentation and release note
<gchristensen>
good :)
knupfer has quit [Ping timeout: 260 seconds]
<gchristensen>
people come in to #zfsonlinux all the time and want to turn it on and it is a mad dash to convince them to not. same with encryption, right now, tbh.
<__monty__>
maurer: .nix-profile/share is in my XDG_DATA_DIRS but I guess dbus doesn't care about that variable. How do I get it to look there?
<supernanovoid>
Hey. Can someone sya what I did wrong? fatal error: SDL_mixer.h: No such file or directory #include "SDL_mixer.h". But SDL2_mixer in my buildInputs.
<dash>
:o
<supernanovoid>
say*
<iqubic>
I don't have encryption.
<maurer>
__monty__: I don't know, sorry. I wouldn't even have dbus installed if nixos wasn't dependent on systemd
knupfer has joined #nixos
<maurer>
__monty__: ~/.local/share/dbus-1/services may work?
<maurer>
XDG_DATA_DIRS may not be being set early enough for dbus to get it
<iqubic>
So I installed ASpell and a few dictionaries. Where did the dictionaries go? Clearly not into /usr/share/dict/words because I don't even /usr/share
<{^_^}>
→ 2cca7e68 by @globin: rocket-errbot: fix hash
__monty__ has quit [Read error: Connection reset by peer]
__monty_1 has joined #nixos
<{^_^}>
[nixpkgs] @knedlsepp opened pull request #39879 → whois: Fix darwin build → https://git.io/vpzN7
__monty_1 is now known as __monty__
<{^_^}>
[nixpkgs] @globin closed pull request #39873 → PhpStorm: from 2017.3.6 to 2018.1.2 → https://git.io/vpzXc
MichaelRaskin has joined #nixos
johnw has joined #nixos
<{^_^}>
[nix] @gleber opened pull request #2132 → docs: mention source of env variables used by `impureEnvVars`. → https://git.io/vpzNp
Ariakenom has quit [Read error: Connection reset by peer]
Ariakenom has joined #nixos
LysergicDreams has quit [Quit: leaving]
coot has quit [Quit: coot]
aziei has joined #nixos
jrolfs has joined #nixos
<gchristensen>
"nix is a good build tool" is a bit confusing to new users I think, because they immediately jump to make
MP2E has joined #nixos
Rovanion has quit [Ping timeout: 276 seconds]
lopsided98 has quit [Ping timeout: 240 seconds]
<MichaelRaskin>
Do we have any examples where Nix is actually used as a build tool?
jrolfs has quit [Ping timeout: 240 seconds]
<gchristensen>
how would you describe Nix?
<samueldr>
bundler, composer, npm, [right python equivalent], but for your system or part of your system; that's how I explained it to my colleagues
aziei has quit [Ping timeout: 264 seconds]
<samueldr>
where system could either mean your OS or your $project
<ryantrinkle>
gchristensen: it's a generic system-level memoizer for computations
<gchristensen>
omg
<ryantrinkle>
i wouldn't describe it to *everyone* that way
<samueldr>
^- that would have scared everyone!
<ryantrinkle>
sorry, just barreled into this conversation without reading any scrollback :P
<gchristensen>
how would you describe it to python devs?
<ryantrinkle>
ah, to python devs, interesting
<simpson>
ryantrinkle: I wouldn't describe Nix that way; from that angle, Nix is *terrible* at what it does, since its granularity is necessarily pretty poor.
<gchristensen>
damn
<gchristensen>
oops wrong channel
<symphorien>
virtualenv but also for clibs ?
<ryantrinkle>
how about this
<ryantrinkle>
"microcontainerization"
<ryantrinkle>
in the sense of microservices
<ryantrinkle>
that sort of leaves out the purity aspect, though
<ryantrinkle>
tbqh, i generally *don't* explain nix to the uninitiated
<samueldr>
micropuritanization?
<ryantrinkle>
i try to use it as a hidden piece of whatever value i'm really trying to deliver to them
<ryantrinkle>
nix is rarely the feature; it's almost always the implementation of the feature
<ryantrinkle>
e.g. i used nix for reflex-platform because i wanted my users to be able to use the whole haskell ecossytem, in both ghc and ghcjs, without needing to build anything
<simpson>
If I'm allowed to use terms from the literature, I would describe Nix as an automatically-taming package builder adhering to the principle of least authority (POLA).
<ryantrinkle>
nix was the only effective way to give them prebuilt binaries that would actually work
<simpson>
POLA and taming are the parts of Nix which trip up newbies IMO.
<simpson>
ryantrinkle: I wouldn't want to memoize per-object compilations; I don't have the disk space for that. I think that it's only practical when your builds have big granularity and relatively compact outputs.
__monty__ has joined #nixos
jrolfs has joined #nixos
<simpson>
I totally see your point though.
<ryantrinkle>
hm, maybe better heuristics for auto-gc could be useful for that
<ryantrinkle>
it's about 6x faster to enter an existing sandbox than it is to create a new one
<ryantrinkle>
4.5ms versus 24ms
coot has joined #nixos
olto has quit [Ping timeout: 240 seconds]
<gchristensen>
seems riskier that way
<Dezgeg>
well, the real problem is that with e.g. C programs you don't know the dependency graph of your program beforehand, which is what nix assumes that you do
<dtz>
I mentioned this on #nixos-chat the other day, but on the subject of memoizing builds more granularly in a nix-like manner: https://github.com/StanfordSNR/gg
<dtz>
they have a preprint online somewhere, FWIW
jrolfs_ has quit [Ping timeout: 256 seconds]
<supernanovoid>
how to give compiler possibility to #include SDL_mixer.h from SDL2, using cflags or pkgconfig or something else?
<clever>
supernanovoid: id try pkgconfig first
jrolfs has quit [Ping timeout: 248 seconds]
<Dezgeg>
shouldn't you do this rather: #include <SDL2/SDL_mixer.h>
<clever>
Dezgeg: the issue is that a lot of packages expect -I/usr/include/SDL2 and try to include their own files without the SDL2 prefix
<clever>
because nix wasnt a thing and they had to work around the conflicts
<clever>
and now they are making nix harder to use, lol
<Dezgeg>
well, arguably those packages are broken
<clever>
they currently rely on either pkgconfig or `-I${SDL}/include/SDL2`
<clever>
i would prefer to just ditch the SDL2 prefix in the postInstall, so it works out of the box
<Dezgeg>
the documented way seems to be `sdl2-config --cflags`
<{^_^}>
[nixpkgs] @globin pushed commit from R. RyanTM to master « imagemagick7: 7.0.7-27 -> 7.0.7-29 »: https://git.io/vpgmK
ericsagnes has joined #nixos
ambro718 has joined #nixos
dbmikus__ has quit [Ping timeout: 256 seconds]
TonyTheLion has joined #nixos
mkoenig__ is now known as mkoenig
<dtz>
whoa xD
<gchristensen>
dtz: are you in acadamia?
<gchristensen>
does Nixpkgs have one of them tools which can upload data super fast, like way faster than scp, etc?
aarvar has joined #nixos
<dtz>
yes, and not sure sorry! :)
<gchristensen>
bummer :)
<simpson>
Replace scp with rsync.
<gchristensen>
rsync is quite slow too
<simpson>
Like, not just verbatim, you need to do the flags too. But it's better in almost every way.
<simpson>
Oh, what's the bottleneck?
<gchristensen>
but there are some magic tools which are very efficient at uploading data, uses udp
<Dezgeg>
isn't there some high-performance patchset on top of ssh?
hamishmack has joined #nixos
<Dezgeg>
hpn-ssh or something, haven't tried though
<simpson>
I mean, I use Dat and magic-wormhole all the time, and they're pretty awesome, although the Dat tooling's a little iffy and the Dat community's unhelpful at best.
<gchristensen>
the root of the problem here is I'm very inconveniently testing PXE images and they suck to upload a lot
<clever>
gchristensen: upload the nix expression and build on the far end?
<{^_^}>
[nixpkgs] @globin pushed commit from R. RyanTM to master « insomnia: 5.16.0 -> 5.16.1 »: https://git.io/vpgY0
<gchristensen>
even worse, the remote is a very slow builder ... _however_ I could provision a fast box, then build on that :)