hellrazo1 has quit [(Read error: Connection reset by peer)]
hellrazo1 has joined #nixos
joehh has joined #nixos
jdz1200 has joined #nixos
joehh has quit [(Ping timeout: 248 seconds)]
vaibhavsagar has quit [(Ping timeout: 248 seconds)]
Supersonic112 has quit [(Disconnected by services)]
Supersonic112_ has joined #nixos
Supersonic112_ is now known as Supersonic112
newhoggy_ has joined #nixos
newhoggy has quit [(Ping timeout: 246 seconds)]
LnL has quit [(Quit: exit 1)]
LnL has joined #nixos
lambdamu_ has joined #nixos
lambdamu has quit [(Ping timeout: 240 seconds)]
<Infinisil>
Do you think it makes sense to put the system config (configuration.nix etc., separate from all user config stuff) in /root?
<Infinisil>
I want to have a git repo for that stuff, but I don't want to put it in /home/infinisil
<Infinisil>
And putting it it /etc feels a bit odd
<ottidmes>
Infinisil: I have that seperation, but I do not want to have to be the root user or sudo to edit the files, so I have a special top level directory with my user permissions, but which contains all the system wide stuff
<ottidmes>
Infinisil: ~/my = user stuff, /my = system stuff
zeus_ has joined #nixos
<Infinisil>
ottidmes: Sounds good, probably gonna use that
<Infinisil>
"my" seems a bit arbitrary though, maybe gonna use "config" or "system" instead
<ottidmes>
Infinisil: Agreed, I am not fully set on that yet, I wanted it to be something short 2-3 chars, and it contains more then just my config files, so it has to be generic enough
<Infinisil>
Hmm
<ottidmes>
Infinisil: Although my sound nice when you type things like: /my/proj/nixcfg
<Infinisil>
heh, I'll probably use cfg instead
<ottidmes>
Infinisil: Sure, if you will only use it for config, cfg is the best in my opinion
<Infinisil>
ottidmes: What are you using this for other than configs?
<ottidmes>
Infinisil: Currently it also contains the global npm and composer folder
zeus_ has quit [(Ping timeout: 240 seconds)]
s33se has joined #nixos
<ottidmes>
Infinisil: And some shared web stuff, and my nixpkgs checkouts (which I do not really count as config, but you could label them as such)
<Infinisil>
Yeah I'm just thinking about where to put nixpkgs
<ottidmes>
Infinisil: Basically anything system-wide that does not already has a defined location, such as /etc
<ottidmes>
Infinisil: It currently resides in /my/dev/proj/nixpkgs, but I am thinking about dropping the dev folder
<Infinisil>
ottidmes: Well actually it wouldn't be too bad to put nixpkgs in my system config as a git submodule
s33se_ has quit [(Ping timeout: 268 seconds)]
<Infinisil>
Because it really should be part of it, my cherry-picks and such
mbrgm has quit [(Ping timeout: 260 seconds)]
mbrgm has joined #nixos
<ottidmes>
Infinisil: You could do that, personally I like to keep them seperate
mbrgm has quit [(Changing host)]
mbrgm has joined #nixos
jdz1200 has quit [(Remote host closed the connection)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 4 new commits to master: https://git.io/vQFxP
<NixOS_GitHub>
nixpkgs/master e4eea75 Franz Pletz: zfs: 0.6.5.10 -> 0.6.5.11
<NixOS_GitHub>
nixpkgs/master 7f0994c Franz Pletz: zfsUnstable: 0.7.0-rc4 -> 0.7.0-rc5
<NixOS_GitHub>
nixpkgs/master 77a0b04 Franz Pletz: darktable: 2.2.4 -> 2.2.5
NixOS_GitHub has left #nixos []
<ottidmes>
nixcfg = systems configurations, nixpkgs I try to keep generic (no machine specific stuff)
<Infinisil>
ottidmes: I'd like to keep a single system config repo for all machines (currently only 1 but will be 2 eventually)
<boomshroom>
I have a bootstrapped nix installation on my university student account. I can run nix-env and nix-channel (albeit with build errors for individual packages), but when I try to debug one package with nix-shell, it complains about nixpkgs not being found.
boj has quit [(Quit: byte byte)]
<Infinisil>
Do you have a ~/.nix-profile directory with stuff in it?
<Infinisil>
boomshroom:
<boomshroom>
Infinisil: on the system in question: no.
<Infinisil>
boomshroom: Well that's where it looks for nixpkgs and other stuff
<Infinisil>
boomshroom: Does nix-env -iA nixos.hello work?
<boomshroom>
Infinisil: I realized I was using a nixos channel for nixpkgs, I'm not downloading the nixpkgs channel.
<ottidmes>
Infinisil: I did have that, but moved to a 3 repo structure, shared, personal, and machine-specfic, because once you start to have to manage a machine that is not personal, you still want to use some shared basis, but do not want all you personal stuff to be included as well, but that is just what I ended up with
boj has joined #nixos
<Infinisil>
ottidmes: That sounds reasonable, how are you including the shared one in the others?
<Infinisil>
Via symlinks or submodules?
<ottidmes>
Infinisil: neither, I just have the nixcfg dir and in it, checkout those that are relevant
<timclassic>
ottidmes: I used to be a Solaris admin and ZFS has never treated me badly.
<Infinisil>
ottidmes: Ahh
<dash>
timclassic: i hear it's not so great for large collections of small files, any truth to that
<ottidmes>
Infinisil: And the nixcfg dir can be found on my NIX_PATH, so I do everything through <nixcfg/shared/...>
newhoggy_ has quit [(Remote host closed the connection)]
<ottidmes>
timclassic: Good to know, seems the safest bet for my backup server then. I really want snapshots, so it basically comes down to either BTRFS or ZFS
<Infinisil>
As in, using compress on /home /root and /nix, using the pool root dataset as the root directory, and the bootfs thing (i have no idea what it does but it seems like a thing i want)
vaibhavsagar has joined #nixos
<Infinisil>
I'm going to be snapshotting both /home and /root regularily, where I want the .zfs directory to be visible
<timclassic>
gchristensen: Infinisil: That worked, thank you
<gchristensen>
timclassic: woohoo!
<Infinisil>
Nice
<gchristensen>
timclassic: I hope you found the process fairly pleasant
<timclassic>
ottidmes: I honestly have zero experience with btrfs. ZFS does certainly feel slower than say, ext4, for some uses. But it's also safer, so I generally just accept it as the cost of safe data. I don't remember many anecdotes but I could probably help with specific problems to a degree.
<ottidmes>
dash: I think that holds for any file system with heavy usage of metadata, the overhead of the metadata is more problematic when you have many small files
<Infinisil>
Ah I just read that in the manpage before, zfs stores metadata redundantly, which i assume just means more than once
<ottidmes>
timclassic: From what I read ZFS focuses on stability and features, in that order, and BTRFS has been experimental for much longer, although it is mostly stable now from what I understood, just not long enough for all features to really be considered stable
<gchristensen>
it is a bit shocking that ZFS is basically the modern rosetta stone of file systems
hyphon81 has quit [(Remote host closed the connection)]
rigelk has quit [(Quit: leaving)]
ottidmes has quit [(Ping timeout: 276 seconds)]
<disasm>
gchristensen: I've been contemplating redoing my laptop with zfs... Still using btrfs currently though.
<Infinisil>
disasm: Are you having problems with btrfs
<Infinisil>
?
Wizek__ has quit [(Quit: Connection closed for inactivity)]
Guest94705 has joined #nixos
ryanartecona has joined #nixos
<Guest94705>
Evening everyone. Currently setting up nixos and have kde wallpapers installed. I can find them via the /nix/store but I thought I remembered that there was supposed to be a linked path in $HOME somewhere to installed wallpapers and such. Cant seem to find it. Is my memory off or can someone point me to the correct path?
<boomshroom>
Infinisil: Sorry I took so long (had dinner and watched some Penn & Teller) from the looks of it, hello will probably fail because it still depends on stdenv.
<boomshroom>
Sidenote: apparently one of my CSC profs is logged in at the same time as me.
nslqqq has quit [(Read error: Connection reset by peer)]
<Infinisil>
Damnit, nixos-install doesn't work: chroot: failed to run command ‘/nix/store/d8lp1q56wp4562bygzglrllifgqxwl63-nix-1.11.11/bin/nix-env’: No such file or directory
<Infinisil>
Well damn, I just picked the installer from master and it works
rigelk has joined #nixos
reinzelmann has quit [(Ping timeout: 240 seconds)]
Ivanych has joined #nixos
<bennofs>
when you open a PR as a tab in the browser for review and it gets merged in the mean time :+1: :)
vincent_vdk has left #nixos []
<srhb>
bennofs: Nice :D
lambdael has quit [(Quit: WeeChat 1.7.1)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] bennofs closed pull request #27000: pulseaudio: Resolve conflicting asound.conf of pulseaudio and alsa (master...fix/pulseaudio-alsa-conf) https://git.io/vQRce
NixOS_GitHub has left #nixos []
slack1256 has quit [(Remote host closed the connection)]
reinzelmann has joined #nixos
leothrix has quit [(Ping timeout: 268 seconds)]
freusque has joined #nixos
leothrix has joined #nixos
<boomshroom>
Infinisil: I finally have the results of trying to build hello: Same test failure on findutils.
<boomshroom>
Yes, it's been going since you first asked. The joys of local compilation.
<boomshroom>
Actually, while it failed on the same package, it gave a different error than before. This time it failed at test-lock.
freusque has quit [(Quit: WeeChat 1.7.1)]
bennofs has quit [(Ping timeout: 255 seconds)]
freusque has joined #nixos
<peti>
niksnut, ikwildrpepper: FYI, hydra.nixos.org has run out of disk space.
<boomshroom>
peti: Sounds like it's time to run a nix-store --optomize; nix-store --gc. :P
<makefu>
apparently it does this sometimes but other times the issue cannot be solved with this. thought i've read something like this on the ml
<makefu>
and nix-store --optimize literally takes forever on a big store
ytn07 has joined #nixos
ytn07 has quit [(Client Quit)]
<Infinisil>
makefu: I don't think you mean literally literally
<Infinisil>
I still don't get why hydra just throws out some old stuff
<makefu>
Infinisil: if your store is infinitely big, then yes ;)
<Infinisil>
But if it were infinitely big it wouldn't run out of disk space :P
<makefu>
but maybe you want to save some!
<makefu>
because infinite disk space is expensive you know!
* peti
has been using "auto-optimise-store = true" on all his machines for ages.
<Infinisil>
peti: Just turned this on yesterday
<makefu>
someone here on the channel wrote about the experience of waiting 24hours or something for optimize to finish
<peti>
It makes garbage-collection slow, but I feel it's worth the tradeoff.
<Infinisil>
I should really look through all nix.conf options and configuration.nix options as well at some time
Wizek has joined #nixos
Wizek_ has joined #nixos
<makefu>
peti: not sure why as it simply removes the hardlinked reference no? once all references to an object are deleted the space will be marked as free
Wizek__ has joined #nixos
justbeingglad has joined #nixos
justbeingglad has left #nixos []
katyucha_ is now known as katyucha
elninja44 has quit [(Remote host closed the connection)]
mudri has joined #nixos
nschoe has joined #nixos
boomshroom has quit [(Ping timeout: 260 seconds)]
Khetzal has quit [(Remote host closed the connection)]
Guest28496 has joined #nixos
ertes has quit [(Ping timeout: 248 seconds)]
Guest28496 has quit [(Remote host closed the connection)]
Khetzal_ has joined #nixos
freusque has quit [(Ping timeout: 246 seconds)]
FRidh has joined #nixos
Itkovian has joined #nixos
jensens has joined #nixos
civodul has joined #nixos
<hyper_ch>
nixos-unstable channel still hasn't updated?
newhoggy has quit [(Ping timeout: 246 seconds)]
ertes-w has joined #nixos
MoreTea has joined #nixos
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
sibi_ has quit [(Quit: Connection closed for inactivity)]
lsix has joined #nixos
Itkovian has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to staging: https://git.io/vQbOM
<NixOS_GitHub>
nixpkgs/staging 8e98811 Frederik Rietdijk: python36: 3.6.1 -> 3.6.2
NixOS_GitHub has left #nixos []
spinus has quit [(Ping timeout: 240 seconds)]
vcunat has joined #nixos
vcunat has quit [(Client Quit)]
takle has joined #nixos
hotfuzz_ is now known as hotfuzz
<Infinisil>
hyper_ch: Nope, 4 days old
<hyper_ch>
Infinisil: ?
<Infinisil>
hyper_ch: unstable is from 4 days ago, right?
<hyper_ch>
shows as 13 days old
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<Infinisil>
hyper_ch: That's just what channels are
<Infinisil>
I think
<Infinisil>
Not?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to master: https://git.io/vQb32
<NixOS_GitHub>
nixpkgs/master c0cf196 Frederik Rietdijk: aspellWithDicts: create derivation with aspell and selected dictionaries...
NixOS_GitHub has left #nixos []
<hyper_ch>
it's not the same
<Infinisil>
hyper_ch: What's the difference?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh closed pull request #26622: aspellWithDicts: create derivation with aspell and selected dictionaries (master...aspell) https://git.io/vHxsW
NixOS_GitHub has left #nixos []
<hyper_ch>
no idea
<Infinisil>
hyper_ch: Heh, well there obviously is one, since one is older than the other
<hyper_ch>
nixos unstable is updated when nixos tests pass (e.g. installer works, kde works...).
<hyper_ch>
nixpkgs unstable is updated when some critical packages build.
<hyper_ch>
They are both nixpkgs checkouts, but at different commits.
<Infinisil>
hyper_ch: Ah interesting
<ertes-w>
i notice that python packages are inconsistently named, usually ridiculously outdated, etc., which basically just hints at the lack of something like cabal2nix/haskellPackages for python… is there any solution in sight, or are python packages just too non-uniform?
jgertm has joined #nixos
johnw has quit [(Ping timeout: 246 seconds)]
freusque has joined #nixos
freusque has quit [(Client Quit)]
<FRidh>
ertes-w: it likely won't change in the coming years. While you can use `pypi2nix` it won't cover all cases. The main issue is that there is no proper index with packages and their dependencies like you have with Haskell and stack.
ebzzry has quit [(Ping timeout: 240 seconds)]
<FRidh>
Furthermore, with Python there is no effort for creating and maintaining a package set of packages that work together.
<ertes-w>
yeah… apparently there are even different and conflicting build/deployment systems for python
<ertes-w>
with haskell all build systems (cabal-install, stack and even nix) use cabal
<FRidh>
There are indeed several, but they're working on providing a common front-end (via pip)
freusque has joined #nixos
<ertes-w>
last week someone from #python told me that even the way you invoke test suites is non-uniform
<FRidh>
that's right
<FRidh>
more packages do seem to use pytest nowadays, but they may still invoke the testrunner differently
jgeerds has joined #nixos
<ij>
Shouldn't the config passed down to dockerTools.buildImage, that ends up in the *docker image json*, be checked for valid keys? Maybe not restrict to, but at least warn about unrecognized keys?
__Sander__ has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to master: https://git.io/vQbZs
<NixOS_GitHub>
nixpkgs/master d2aa638 Frederik Rietdijk: update-python-libraries script: don't forget to break when we actually find an update
NixOS_GitHub has left #nixos []
<ertes-w>
ij: probably
<ij>
I added WorkingDir+Environment, but there's no Environment, just Env. Ended up looking inside of docker .nix'es to find out what should be there and ended up reading docker json image spec.
<ertes-w>
ij: unfortunately it's difficult to even find proper documentation on the config.json format
<ertes-w>
all i found is a half-assed wiki page
tmaekawa has joined #nixos
newhoggy has joined #nixos
<ertes-w>
is there a way to encode NIX_PATH fallbacks? imagine <mynixpkgs|nixpkgs>, such that if mynixpkgs is undefined, use nixpkgs
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 4 new commits to python-wip: https://git.io/vQbnm
<NixOS_GitHub>
nixpkgs/python-wip 9bb85d1 Frederik Rietdijk: python.pkgs.numpy: 1.12.1 -> 1.13.1
<NixOS_GitHub>
nixpkgs/python-wip 9df0766 Frederik Rietdijk: python.pkgs.scipy: 0.19.0 -> 0.19.1
<NixOS_GitHub>
nixpkgs/python-wip 64c2476 Frederik Rietdijk: python.pkgs.pandas: 0.20.2 -> 0.20.3
NixOS_GitHub has left #nixos []
<Guest70706>
what is the default backend for matplotlib installed inside a nix-shell?
<FRidh>
Guest70706: try it. Just be aware that if you use --pure it may not find certain things like fonts.
<Guest70706>
I jus tried it
<Guest70706>
and nothing happened
<Guest70706>
plt.show() nothing happens
<the-kenny>
I usually use plt.savefig("foo.svg")
newhoggy has quit [(Quit: Leaving...)]
<the-kenny>
Iirc the problem is that our plt isn't built with gtk/qt support
<FRidh>
indeed
<FRidh>
in order to keep the closure size small
<Guest70706>
Ok so what is the right way to get an interactive backend working?
<Guest70706>
Does it need an extra buildInput?
<the-kenny>
I think you can override matplotlib with matplotlib.override { enableGtk3 = true; } or enableQt = true
<the-kenny>
Haven't tried it
<FRidh>
Which backend do you want? You can add override the matplotlib expression and enable backends, but that requires rebuilds of dependents. Easier is to just add the dependencies to your shell.
<Guest70706>
lets try gtk
<Guest70706>
so instead of rebuilding, it is possible to add it to the environment of the shell, what needs to be written in the shell.nix for this to work?
<Guest70706>
Right now I only have stdenv.mkDerivation buildInputs
<Guest70706>
which has python and matplotlib
<FRidh>
add your gtk there
<Guest70706>
what's the package called?
Itkovian has joined #nixos
<Guest70706>
gnome3.gtk?
<the-kenny>
for gtk2 it's using pythonPackages.pygtk
<Guest70706>
pygtk right
<the-kenny>
trying it myself now
<Guest70706>
oh i'm using python 3
<Guest70706>
no pygtk3 listed
<Guest70706>
for python3
<Guest70706>
ok I'm going to try pyqt
Infinisil has quit [(Ping timeout: 276 seconds)]
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<Guest70706>
does there need to be an override on matplotlib expression or is there an environment variable that can be set for this?
MoreTea has quit [(Ping timeout: 240 seconds)]
samae has quit [(Quit: #okthxbye)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jfrankenau opened pull request #27441: solvespace: Fix path in desktop entry file (master...fix-solvespace) https://git.io/vQbCy
NixOS_GitHub has left #nixos []
ThatDocsLady has joined #nixos
samae has joined #nixos
samae has quit [(Changing host)]
samae has joined #nixos
Coyin has joined #nixos
samae has quit [(Client Quit)]
samae has joined #nixos
samae has quit [(Changing host)]
samae has joined #nixos
samae has quit [(Client Quit)]
<Guest70706>
How do I find out what kind of dependencies are required for all the various backends for matplotlib?
johann__ has quit [(Remote host closed the connection)]
nslqqq has quit [(Ping timeout: 240 seconds)]
<S0rin>
Hey, I am using two channel (stable and unstable) to be able to get a few package latest version (mostly wine) but as I need both 32 and 64 version I use a package override command wich worked before (in plain unstable) but now only work on the wine64 part
<S0rin>
Here a small part of my configuration.nix
<S0rin>
And the problem I get is 'wine --version wine-2.0.1' despite 'wine64 --version wine-2.10 (Staging)' and it should be staging at the same version in both case
<S0rin>
It seems that the override is not working on the 32 bit part so maybe I need to change something in the configration
<avn>
S0rin: wineStaging mean wine with wine-staging patchset applied
<avn>
although output above show that you have mixed wine from different channels
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] globin pushed 1 new commit to master: https://git.io/vQbzk
<NixOS_GitHub>
nixpkgs/master 8669fb1 Nadrieril: tinc service: BindToAddress and ListenAddress are different options, they should not be mistaken
NixOS_GitHub has left #nixos []
iyzsong has joined #nixos
<S0rin>
avn: Yes but I tryed to install it with unstable only but it seems to take the stable for the i386 part
<avn>
S0rin: I am co-maintainer, but I mainly use i386 part only. Just haven't any 64bit app to test on
<S0rin>
avn: My use case for using both is most application need i386 but some big (wow) need amd64
<S0rin>
It was working fine before in plain unstable, but I would like to try to keep most of the basic system to stable and upgrade only the needed package to unstable (ie wine)
<ottidmes>
S0rin: Should you not be explicit about your wine32 build as well: (unstable.wine.override { wineBuild = "wine32"; wineRelease = "staging"; })
<S0rin>
ottidmes: well maybe this override part has changed in the wiki they ask only to add 'wineBuild = "wineWow"' to get both 32/64 https://github.com/nixos-users/wiki/wiki/Wine
<ottidmes>
S0rin: why then have unstable.wine besides your override?
lsix has quit [(Read error: Connection reset by peer)]
<S0rin>
ottidmes: because I want to install that version
<ottidmes>
S0rin: Maybe you need to do the override on the nixpkgs level, so in you nipxkgs config.nix have in your packageOverrides wine = unstable.wine.override { ... }; so if it is referenced as a dependecy elsewhere, it will refer to your unstable wine, not the stable wine
<S0rin>
here the 'old method' I used to get the both 32/64 version in plain unstable http://dpaste.com/1B190V0
<S0rin>
ottidmes: yes I had that part before so maybe I need to add it back
<ottidmes>
S0rin: Yeah, so that is probably the problem, it likely is referenced elsewhere via pkgs.wine outside your systemPackages, and then you will get both
<S0rin>
ottidmes: yes I will try to add that part back
<ottidmes>
S0rin: you say you have both unstable.wine and unstable.wine.override { ... }; because you want to install that version. I am confused, should you not have just one? Especially you said wine installs both versions with the override one
magnetophon has quit [(Ping timeout: 240 seconds)]
<S0rin>
ottidmes: before I used 'unstable.wine.overrid' I used plain unstable so I did *not* use unstable.wine just wine so other package did not confuse I suppose
<ottidmes>
S0rin: Anyway, so what I should do, add your unstable wine to packageOverrides, similar to what you had before, but obviously now using unstable.wine, and then reference just regular wine in your systemPackages (not unstable.wine, since that will not reference your packageOverrides variant)
lsix has joined #nixos
AlanPearce[m] has joined #nixos
<S0rin>
ottidmes: ah I will try that
lsix has quit [(Ping timeout: 255 seconds)]
ertes-w has quit [(Read error: Connection reset by peer)]
ertesx-w has joined #nixos
ertesx-w is now known as ertes-w
Guest70706 has quit [(Quit: WeeChat 1.7)]
tmaekawa has quit [(Quit: tmaekawa)]
simukis has joined #nixos
<S0rin>
ottidmes: So basically changing the names of package from wine to {pkgs,unstable} broke some sort of alias in the wine overall name right ?
<S0rin>
and the only way to override the alias naming is to let the default name then use unstable.wine.override { ... }; syntax ?
noam_ has joined #nixos
<ottidmes>
S0rin: It did not break any alias. It is just pkgs.wine != your unstable.wine.override { ... }, so if there is any reference to pkgs.wine in the system other than were you have used the unstable.wine.override { ... }, it will mean that both variants will be installed
arjen-jonathan has joined #nixos
ibrahims has quit [(Ping timeout: 248 seconds)]
noam__ has quit [(Ping timeout: 240 seconds)]
<S0rin>
ottidmes: ok
seppellll has joined #nixos
<S0rin>
ottidmes: I thought even if some package rely on pkgs.wine my user will have the overrided version
ibrahims has joined #nixos
<ottidmes>
S0rin: It will via packageOverrides, but not via systemPackages, that only means it will be linked in you /run/current-system/sw/... so e.g. their executables are on the PATH
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] mbrock closed pull request #26456: boost: enable static libraries on Darwin (master...master) https://git.io/vH64g
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to staging: https://git.io/vQbVW
<NixOS_GitHub>
nixpkgs/staging 3eceecb Frederik Rietdijk: Merge remote-tracking branch 'upstream/master' into HEAD
NixOS_GitHub has left #nixos []
<S0rin>
ottidmes: ah ok thank you for the details !
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 4 new commits to staging: https://git.io/vQbVr
<NixOS_GitHub>
nixpkgs/staging ca0f2d8 Frederik Rietdijk: python.pkgs.numpy: 1.12.1 -> 1.13.1
<NixOS_GitHub>
nixpkgs/staging c2df8a2 Frederik Rietdijk: python.pkgs.scipy: 0.19.0 -> 0.19.1
<NixOS_GitHub>
nixpkgs/staging ac3cae1 Frederik Rietdijk: python.pkgs.pandas: 0.20.2 -> 0.20.3
<S0rin>
ottidmes: I had hard time finding the good syntax but it works now 'wine --version wine-2.10 (Staging)' and 'wine64 --version wine-2.10 (Staging)'
<ottidmes>
S0rin: Great to hear!
<S0rin>
ottidmes: but finally isn't it a better way to do it for each package I want from unstable to be sure to not have the same problem ?
<ottidmes>
S0rin: As most questions can answered: It depends. However, yes, generally it is best to define them in the same way. Examples when you would not want to do it are when you truly want to have two versions, so e.g. a unstable version for systemPackages, but you would still want the stable packages that depended on it, to refer to the stable version to prevent breakage
<ottidmes>
S0rin: So in short, end user packages, generally yes, intermediate packages mostly used as dependency for others, generally no
<S0rin>
ottidmes: very clear thank you !
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 6 new commits to staging: https://git.io/vQboy
<NixOS_GitHub>
nixpkgs/staging e0f988f Lancelot SIX: pythonPackages.kombu: fix tests for python3.6
<NixOS_GitHub>
nixpkgs/staging 1e06824 Lancelot SIX: pythonPackages.olefile: init at 0.44
<ottidmes>
S0rin: The latter case (the intermediate packages), would for example result in having to build them locally, since you changed their build inputs
snikkers has joined #nixos
<ottidmes>
S0rin: with build them locally, I mean all the packages that have the overriden intermediate package as their build input
<S0rin>
ottidmes: yes -bindist in gentoo style building locally is not too much a pain, but flexibility is important
<ottidmes>
S0rin: I do not want to build things like chromium locally, do you not have to do that with this -bindist gentoo style of building locally?
<S0rin>
ottidmes: in gentoo with -bindist you build everything locally but as you may/sould add some gcc optimisation and other customisation it take time sure but it is with your own changes (and faster)
<S0rin>
But yes install gentoo with -bindist take a couple of day and nixos less than 1 hour but if you need to locally rebuild thing it may become close
vaibhavsagar_ has joined #nixos
<ottidmes>
S0rin: Ah alright. I like that you can choose whether you want to do so in NixOS, most flexible OS I have yet to see
<S0rin>
ottidmes: I have to aggree a bit hard (to me) to bend to my needs but really promising and powerfull
<gchristensen>
one thing is nixos requires more building than gentoo
<ottidmes>
gchristensen: Due to having multiple versions of stuff?
<gchristensen>
no, because if we change a teeny tiny shell script at the root of the dependency tree we have to rebuild everything :P whereas gentoo would just call it good
<ottidmes>
Ah like that
takle has quit [(Remote host closed the connection)]
<S0rin>
gchristensen: yes sure but it has a price respect names and version to let the dependencie solver work and it is not perfect
<gchristensen>
of course
<gchristensen>
but it is a common error to assume it is the same amount of recompiling as gentoo
<S0rin>
gchristensen: yes comming from gentoo I tend to compare with what I know :)
<gchristensen>
hehe
<ottidmes>
I gone the route of Windows -> Ubuntu -> Arch Linux -> NixOS
<FRidh>
Do we have duplicate expressions for gstreamer plugins?
<gchristensen>
duplicate? I don't think so
<gchristensen>
we have the legacy ones and new ones
<S0rin>
In the end I had really like gentoo to be able to configure everything and control, but in the end you get a lot of repetition of task to check (aka the maintenance) and the build time tend to be a problem to use on small device (rpi) or server but is a very good system to learn
<FRidh>
gchristensen: I see two expressions, both downloading from same repo/url but different minor version.
<NixOS_GitHub>
[nix] grahamc opened pull request #1469: Show users how to source the profile, to get a working Nix. (1.11-maintenance...domen-feedback-nix) https://git.io/vQbXf
NixOS_GitHub has left #nixos []
reinzelmann has quit [(Quit: Leaving)]
<FRidh>
so without quotation marks
<S0rin>
ottidmes: but this notation will override the other channel right ?
<ottidmes>
S0rin: That will set the default nixpkgs location
<S0rin>
if I just want to add a local source and use it with the local. prefix like unstable
<ottidmes>
S0rin: So I use that nix.nixPath setting to define the location to my local stable checkout, and the import trick to define a reference to my unstable checkout
<S0rin>
Like that if I want to test something I can change it in the local copy of nixpkgs and use it directly in the config file
<ottidmes>
S0rin: Ow, then use what you have for unstable, but change it to your local variant
<S0rin>
ottidmes: ah yes so it will overide only the default (stable) one
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] edolstra pushed 2 new commits to 1.11-maintenance: https://git.io/vQb13
<NixOS_GitHub>
nix/1.11-maintenance 4b45f94 Graham Christensen: Show users how to source the profile, to get a working Nix.
pie__ has quit [(Remote host closed the connection)]
<S0rin>
ottidmes: ok will try that
<S0rin>
ottidmes: Thank you :D
<ottidmes>
S0rin: I would not complicate things with having 3 versions, just stick with a local checkout of stable and unstable, and modify the default location of nixpkgs to your local stable checkout, and reference the unstable checkout via the unstable = import ... (which I BTW should put in your packageOverrides as well, then you can access it everywhere)
jgertm has joined #nixos
k0001 has joined #nixos
<S0rin>
ottidmes: yes I aggree I will override the unstable with my local one and override the needed package accordingly
<domenkozar>
niksnut: with that I think we can release Nix 1.11.13
<domenkozar>
another thing that comes to mind is root single user install on linux
<domenkozar>
we currently warn it's not supported and then if fails
<niksnut>
domenkozar: ok
<domenkozar>
so we should decide if it should be supported
<domenkozar>
and fix it
<domenkozar>
or disable it all together
<gchristensen>
multi-user installs for all! :)
<domenkozar>
of course on the long run that's the solution
<NixOS_GitHub>
[nixpkgs] fpletz pushed 2 new commits to master: https://git.io/vQb7O
<NixOS_GitHub>
nixpkgs/master 0e9676f Franz Pletz: imagemagick: 6.9.8-10 -> 6.9.9-0
<NixOS_GitHub>
nixpkgs/master cff27d5 Franz Pletz: imagemagick7: 7.0.6-0 -> 7.0.6-1
NixOS_GitHub has left #nixos []
<domenkozar>
but that's still not /tmp :D
<domenkozar>
niksnut: is the -o really necessary in that case?
<domenkozar>
nothing is referencing it
alphor has quit [(Quit: Bye!)]
alphor has joined #nixos
zraexy has joined #nixos
radvendii has joined #nixos
alphor has quit [(Quit: Bye!)]
alphor has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] edolstra pushed 2 new commits to master: https://git.io/vQbdW
<NixOS_GitHub>
nix/master 49304ba Eelco Dolstra: Make the hashes mirrors used by builtins.fetchurl configurable...
<NixOS_GitHub>
nix/master bf6792c Eelco Dolstra: Always use base-16 for hashed mirror lookups...
NixOS_GitHub has left #nixos []
jsgrant has quit [(Quit: Peace Peeps. o/ If you need me asap, message me at msg[(at)]jsgrant.io & I'll try to get back to you within 24-36 hours.)]
bennofs has quit [(Ping timeout: 255 seconds)]
jsgrant has joined #nixos
jsgrant has quit [(Remote host closed the connection)]
<gchristensen>
{.
<gchristensen>
oops
<hodapp>
hm, just installed zulu and now getting "Exception in thread "main" java.lang.Error: Probable fatal error:No fonts found." when trying to run a Java application
jsgrant has joined #nixos
jgertm has quit [(Ping timeout: 276 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] mbrock opened pull request #27445: dapp: init at v0.5.2 (master...dapp) https://git.io/vQbFj
NixOS_GitHub has left #nixos []
<hodapp>
but I already ran this application once - but since then upgraded to 17.03 and did nix-collect-garbage
<gchristensen>
btw was it copumpkin lookin to register a hydra account but couldn't w/ google? turns out you have to disable ublock origin / privacy badger first :)
<__Sander__>
someone interested in making NixOS available for this machine? :D
<hodapp>
graaah. fc-match finds nothing amiss, fc-cache -rv doesn't help
<copumpkin>
Sander: if you give me one I'll try it out
<gchristensen>
"New instructions in the single instruction, multiple data(SIMD) facility offers a boost for traditional workloads usingdecimal operations (i.e. COBOL 6.2, PL/I 5.2) "
<__Sander__>
copumpkin: where are you going to put it? in your garage? haha :D
<__Sander__>
I think you get a huge power bill
<__Sander__>
and also I have heard that the case is also quite heavy :D
<copumpkin>
I don't have a garage, but I have a spare room
<gchristensen>
"Each z14 can have up to 32 TB memory" a modest amount
<copumpkin>
and yes I'll only turn it on when I'm using it
<__Sander__>
I don't see any prices
<__Sander__>
but
<__Sander__>
I have heard (~10 years ago) that the entry level model costs around 250K $
<hodapp>
one does not simply ask for prices with IBM
<MoreTea>
that's quite a hefty write-off.
<__Sander__>
I have worked on one of these machines during an internship
<__Sander__>
more than years ago
<__Sander__>
backend then it had like 1 TiB of RAM
<dash>
o/` biiiig iron, big iron o/`
<__Sander__>
*back then
<gchristensen>
"Canonical, Red Hat and SUSE with their latest supported releases and versions;for the minimum and recommended levels please see IBM Tested platforms page: ibm.com/systems/z/os/linux/resources/testedplatforms.html" ouch
<gchristensen>
super ouch. can you imagine being RedHat and being showed up by Canonical?
<hodapp>
friend worked around some of these years ago, and reported that with the support contracts they were under, they'd come in on a Monday and see that the machine had something fail at 3 AM on Saturday and someone from IBM already showed up to replace the part while they were out because the machine had just phoned home
<dash>
gchristensen: they just forgot to bribe that one copy editor
<__Sander__>
yeah
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] ttuegel opened pull request #27446: vcsh: Patch for Git 2.13.2 (master...bugfix/vcsh-shared) https://git.io/vQbAs
<clever>
gchristensen: i think each build takes 8 hours
<gchristensen>
yowz!
aurorum has joined #nixos
nslqqq has joined #nixos
<aurorum>
hello. i'm very interested in nixos.. however, dauntingly, i'm confused why so many distros have the zeitgeist package.
erictapen has quit [(Ping timeout: 240 seconds)]
<MoreTea>
aurorum, because of demand?
<aurorum>
what demand?
<MoreTea>
no clue, I don't know the package ;)
<aurorum>
i think it's one of the more important packages out there in regards to the fundamental founding ideas of linux, adopted or intended. perhaps you should read about it.
lsix has joined #nixos
<srhb>
aurorum: What, how do you figure that?
<aurorum>
the file description for one
<srhb>
Logging everything you do as a user is not really a fundamental linux feature. :P
<aurorum>
it's definitely counter intuitive to the fundamentals. i guess i said that backwards.
<srhb>
Ah.
<srhb>
That makes more sense :-)
<aurorum>
so, why is everyone using it?
<srhb>
People like getting handy suggestions from their history of usage, I guess.
<srhb>
Or they don't care enough about the implications to remove it.
<srhb>
It's an extended form of "I like my browser history" :)
<aurorum>
i guess but, programs do their own history logging already. why have a program that centralizes all of that data?
<srhb>
To make it available between programs, I suppose
<srhb>
but really, I (we) are probably the wrong people to ask
<MoreTea>
could be a required dependency of gnome?
<srhb>
Could be
<aurorum>
of lots of things i hear. many distros have problems removing it due to dependancies
hke has quit [(Quit: bye)]
<srhb>
aurorum: Go to nixpkgs/pkgs, grep -R zeitgeist
<srhb>
aurorum: That should give you an idea which packages use it
<srhb>
Or depend (possibly optionally) on it, at least
<srhb>
It's not a lot.
<aurorum>
i haven't decided to install nix yet. i was just wondering if some one infinitely knowledgeable could explain why distros are using it vs letting programs do their own logging.
<srhb>
aurorum: It's just a git repo, you can clone it without installing anything.
<srhb>
aurorum: Nope, no such luck. :)
<srhb>
Ask the creators, they're probably the people who are best able to argue for its existence.
<aurorum>
i figured they'd be here
hke has joined #nixos
<srhb>
I meant the creators of Zeitgeist, not zeitgeist the nixpkgs package :)
<srhb>
Most contributions to nixpkgs are accepted if a person is willing to put in the work to maintain it.
<aurorum>
ah, well, ive already read one devs response to it's validity in a post regarding ubuntu.
<srhb>
(Oftentimes not even that)
<srhb>
So perhaps just one person wanted it.
<MoreTea>
aurorum, nixpkgs will probably accept a patch to disable this by default, if by some chance it is logging stuff automagically.
<aurorum>
that sounds nice
erictapen has joined #nixos
<srhb>
aurorum: And if you clone nixpkgs, you can git log the file and find the people who have created/altered it.
<aurorum>
i'll remember that. thanks for the advice. o/
<srhb>
:)
aurorum has quit [(Quit: Page closed)]
erictapen has quit [(Ping timeout: 240 seconds)]
bennofs has joined #nixos
seppellll has quit [(Remote host closed the connection)]
erictapen has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jwiegley pushed 1 new commit to master: https://git.io/vQNJr
<NixOS_GitHub>
[nixpkgs] luigy opened pull request #27449: Fix "mv: argument list too long" for fetchzip in darwin. (master...ll-fix-fetchzip-darwin) https://git.io/vQNJA
Supersonic112 has quit [(Ping timeout: 246 seconds)]
Coyin has quit [(Quit: Leaving)]
snikkers has quit [(Ping timeout: 255 seconds)]
Supersonic112 has joined #nixos
<etu>
Hello, I'm trying to set up jedi with emacs on nixos. Running latest unstable. But got issues when it came to setting up the jedi-server so emacs can speak to python and get suggestions and stuff. Then I realized that letting emacs install jediepcserver.py with virtualenv isn't really the nix way and then I got lost in how to solve this...
<etu>
And the jediepcserver.py script needs python36Packages.epc and python36Packages.jedi to work, but if I'm in my home-environment run python and try to "import jedi" and "import epc", both will fail. But if I run nix-shell -p python36Packages.epc python36Packages.jedi --run "python" and try to import them, the imports works.
<etu>
And that's where I'm stuck at the moment, I don't know why the imports doesn't work in my regular home environment.
vaibhavsagar_ has quit [(Ping timeout: 240 seconds)]
MoreTea has quit [(Quit: Leaving)]
erictapen has quit [(Ping timeout: 255 seconds)]
<bennofs>
etu: nix-shell will run python's setup hooks, which will set PYTHONPATH
<bennofs>
etu: this doesn't happen when using nix-env
<bennofs>
etu: to install python with nix-env, I use something like: nix-env -iE '_: with (import <nixpkgs> {}); python3.withPackages (ps: with ps; [ requests ipython lxml beautifulsoup4 nclib dns ])'
<bennofs>
(you can also create your pythonEnv inside nixpkgs' config.nix, which will allow you to just install it with nix-env -iA)
<srhb>
TIL you can install an expression. Because of course you can. I just never considered it.
<bennofs>
srhb: well, for some reason you need a lambda, but yes, you can
<bennofs>
srhb: /me lobbies for just always accepting exprs everywhere: nix install <expr>, nix build <expr> ...
<srhb>
Indeed :)
<etu>
bennofs: hmm, I see
<bennofs>
preferably in `with (import <nixpkgs> {})` scope by default
<srhb>
Yeah, that would be very flexible.
<srhb>
I basically abuse my shell history to do the same thing, which is rather crappy.
<bennofs>
perhaps even start a REPL on empty `nix install` ? :)
<etu>
bennofs: But is there a good nixy way to get jediepcserver.py in a virtualenv?
<bennofs>
etu: hmm sorry, i don't know enough about jedi yet
<bennofs>
currently trying to slowly figure out how to best setup emacs for all languages on NixOS though
<bennofs>
so eventually I'll look at python too :)
<etu>
:)
<pietranera>
etu in nix you probably would use nix-shell instead of virtualenv
takle has quit [(Remote host closed the connection)]
takle has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] copumpkin pushed 1 new commit to master: https://git.io/vQNOO
<NixOS_GitHub>
nixpkgs/master b4ee24d Dan Peebles: pants13-pre: remove from pythonPackages...
NixOS_GitHub has left #nixos []
<etu>
pietranera: I see
<etu>
srhb: That could work yes
takle has quit [(Remote host closed the connection)]
takle has joined #nixos
takle has quit [(Remote host closed the connection)]
<etu>
pietranera: How did you configure that? When I try to just use jedi:install-server, pip will fail to do stuff in /tmp/pip-some-hash-build/ because that directory is readonly but owned by me
__Sander__ has quit [(Quit: Konversation terminated!)]
erictapen has joined #nixos
digitus has joined #nixos
erictapen has quit [(Ping timeout: 255 seconds)]
peacememories has joined #nixos
nschoe has quit [(Quit: Program. Terminated.)]
arjen-jonathan has quit [(Ping timeout: 240 seconds)]
Filystyn has joined #nixos
Filystyn has quit [(Changing host)]
Filystyn has joined #nixos
alx741_ has joined #nixos
Supersonic112 has quit [(Ping timeout: 268 seconds)]
radvendii has quit [(Ping timeout: 240 seconds)]
alx741 has quit [(Ping timeout: 268 seconds)]
peacememories has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
Supersonic112 has joined #nixos
sibi_ has quit [(Quit: Connection closed for inactivity)]
<pietranera>
etu I have not encountered that problem
<pietranera>
I use jedi:install-server-block and it has worked OK
<pietranera>
I am on NixOS though.
<etu>
pietranera: Me as well
ibrahims has joined #nixos
<etu>
pietranera: What is jedi:install-server-block?
Biappi has quit [(Ping timeout: 240 seconds)]
<pietranera>
it's just a blocking call instead of an aync one
<dash>
pietranera: i think he is asking what it is, not what it does
blahdodo has quit [(Ping timeout: 240 seconds)]
<pietranera>
it's another function defined in the jedi package
<etu>
dash: I know what the non-block one is :)
<etu>
So it's fine
Biappi has joined #nixos
<dash>
:)
blahdodo has joined #nixos
<pietranera>
etu I am not sure that calling the blocking function would make a difference
<etu>
pietranera: Don't think so... The permission issue would be the same.
ertes-w has quit [(Ping timeout: 255 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] globin pushed 1 new commit to master: https://git.io/vQNZ2
<NixOS_GitHub>
nixpkgs/master b8d92a7 Robin Gloster: programs.gnupg: use extraInit instead of interactiveShellInit...
NixOS_GitHub has left #nixos []
zeus_ has joined #nixos
posco has joined #nixos
pie__ has quit [(Read error: Connection reset by peer)]
<bennofs>
these errors happen quite rarely. generally, I hit one error on the first run of the program. Then I need to wait 5 minutes for the cache TTL to expire and on the second run everything is fine
* dtzWill
just went through adventure of (S3+CloudFront)-ifying his hydra bits
<bennofs>
maybe it is because I have such an unusual usage pattern? (I generate loads of 404's :P)
<dtzWill>
idk if this is great advice but IIRC you can use the S3 url directly to avoid the cloudfront cache... useful (perhaps near-necessary) for hydra builders to use
<dtzWill>
since they always hit it up beforee trying to build; which means they'd all be generating negative cache entries as part of normal operation lol
<dtzWill>
bennofs: what's your usage? :D
<qknight>
how can i evaluate a mkOption of a service using nix-instantiate?
<bennofs>
dtzWill: the problem is, I don't know which hashes are built by hydra upfront. so I just try all hashes that a nixpkgs eval gives
<clever>
qknight: nix-build '<nixpkgs/nixos>' -A config.systemd.services.nginx.runner -I nixos-config=./configuration.nix
<bennofs>
dtzWill: and fetch their narinfo / .ls / .ls.xz
<clever>
qknight: this would generate a perl script that runs nginx with all of the right args, based on the service config in the configuration.nix in the current dir
<clever>
qknight: and if you know what attribute the config is at, you can just build/instantiate that directly
<dtzWill>
bennofs: wait, the nar link you gave should have /nar/ before it
peacememories has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<dtzWill>
bennofs: they're not in top-level, only narinfo's are
tvon has joined #nixos
<dtzWill>
bennofs: it works as a narinfo
<dtzWill>
bennofs: at the IP (using "Host" header) as well as cache.nixos.org
<qknight>
clever: i get a runner where a config is referenced which is already pretty good. could i use that to evaluate a single mkOption also? nix-repl seems to know the value of services.nginx.enable for instance
<bennofs>
dtzWill: oh right copy-paste error
<bennofs>
dtzWill: i'm only interested in narinfos
<dtzWill>
yep yep
* bennofs
starts run 3
<bennofs>
dtzWill: well, the 5 min cache TTL is expired, so it works for me too now
<bennofs>
dtzWill: cloudfront caches 500 responses for 5 min
<qknight>
clever: if you answer, please write my name for the notificaten -> afk
<qknight>
clever: and thanks a lot!
<dtzWill>
bennofs: I'm not sure if it's a nice thing to do re:request pricing (dunno!) but if you hit the S3 resource directly you'll not get the 5 min negative cache
jensbin has joined #nixos
<gchristensen>
it is not a nice thing to do, please don't
<dtzWill>
bennofs: and since you're querying hashes that are likely to-be-built (since they're from a nixpkgs eval) might be doing everyone a favor to avoid causing the 5min
<dtzWill>
gchristensen: apologies!
<dtzWill>
gchristensen: since i'm starting on a similar adventure, is it ... because of the request pricing?
<clever>
qknight: you can also just eval/instantiate -A config.services.nginx.foo to target any option
<gchristensen>
more expensive, yeah
<clever>
qknight: but you need to still pass a nixos-config to it
<disasm>
Infinisil: no problems with btrfs, just seems like zsh is the favorite of most around here.
<dtzWill>
bennofs: nevermind my ideas, let's not run up NixOS infra bills unnecessarily! D:
<bennofs>
i don't know why i'm getting these 500's in the first place
<dtzWill>
and I think it's a bit of a feature, in a way, that you can't ask cache.nixos.org to list its contents
<dtzWill>
both because I imagine that list is just a silly thing to generate/transmit but also for security reasons (although maybe our hash density is so high that hardly leaks anything lol)
<bennofs>
dtzWill: i'm querying for about 30k hashes
<dtzWill>
but that does make indexing things a bit more difficult :)
<dtzWill>
bennofs: sounds like at worst your index will be ~5min behind the builders.. which seems reasonable. Or is there a problem you're encountering?
<bennofs>
dtzWill: well, I don't like to tell my users to "just rerun this if you get 500's". seems ugly
<bennofs>
or I could put a sleep 5min in the retries but this makes the indexing slow :(
cornu[m] has joined #nixos
sirius[m] has joined #nixos
jyp[m] has joined #nixos
baconicsynergy[m has joined #nixos
jesper has joined #nixos
regnat[m] has joined #nixos
indefini has joined #nixos
konfou[m] has joined #nixos
xj9[m] has joined #nixos
primeos[m] has joined #nixos
berot3[m] has joined #nixos
scott2 has joined #nixos
bachp has joined #nixos
NickHu has joined #nixos
aspiwack[m] has joined #nixos
hendrik[m]1 has joined #nixos
mith[m] has joined #nixos
TimePath has joined #nixos
davidar has joined #nixos
M-liberdiko has joined #nixos
timclassic has joined #nixos
sudoreboot[m] has joined #nixos
AlanPearce[m] has joined #nixos
ArdaXi[m] has joined #nixos
wak-work[m] has joined #nixos
Kallegro[m] has joined #nixos
teh[m] has joined #nixos
octalsrc[m] has joined #nixos
hl has joined #nixos
Exee7uvo[m] has joined #nixos
Drakonis[m] has joined #nixos
krav_ has joined #nixos
Elephant454[m] has joined #nixos
copumpkin has joined #nixos
spacekitteh[m] has joined #nixos
herzmeister[m] has joined #nixos
reactormonk[m] has joined #nixos
florianjacob has joined #nixos
sargon[m] has joined #nixos
Magnap[m] has joined #nixos
harlock[m] has joined #nixos
spawnthink[m] has joined #nixos
edef[m] has joined #nixos
Wysteriary[m] has joined #nixos
pstn has joined #nixos
offlinehacker[m] has joined #nixos
dtz has joined #nixos
benkolera[m] has joined #nixos
bennofs[m] has joined #nixos
icetan has joined #nixos
Yaniel has joined #nixos
Sovereign_Bleak has joined #nixos
WinterFox[m] has joined #nixos
BurNiinTRee[m] has joined #nixos
bhipple[m] has joined #nixos
Guest64705 has joined #nixos
zimbatm has joined #nixos
mdash has joined #nixos
mtncoder[m] has joined #nixos
jascot[m] has joined #nixos
musicmatze[m] has joined #nixos
qrilka[m] has joined #nixos
StuK[m] has joined #nixos
DIzFer[m] has joined #nixos
Ralith has joined #nixos
frio[m] has joined #nixos
danielrf has joined #nixos
peterhoeg has joined #nixos
bendlas has joined #nixos
jack[m] has joined #nixos
wmertens[m] has joined #nixos
jlle[m] has joined #nixos
jsv[m] has joined #nixos
chominist[m] has joined #nixos
adisbladis[m] has joined #nixos
rycee[m] has joined #nixos
Oo[m] has joined #nixos
Dezgeg[m] has joined #nixos
tommyangelo[m] has joined #nixos
trikl[m] has joined #nixos
seif[m] has joined #nixos
myklam[m] has joined #nixos
necronian has joined #nixos
a123123123[m] has joined #nixos
Criena[m] has joined #nixos
etcinit[m] has joined #nixos
Khorne[m] has joined #nixos
sziszi[m] has joined #nixos
corngood has joined #nixos
fpletz[m] has joined #nixos
taktoa[m] has joined #nixos
eqyiel[m] has joined #nixos
alain[m] has joined #nixos
sphalerite[m] has joined #nixos
admin[m] has joined #nixos
lfont[m] has joined #nixos
ninegua[m] has joined #nixos
<dtzWill>
bendlas: why 500 not 404?
<bennofs>
dtzWill: I get a 500 where I should get a 404
<bennofs>
dtzWill: cloudfront caches the 500 response code
<bendlas>
dtzWill: wat
<qknight>
clever: does work for nginx but not for my service...
<bennofs>
bendlas: probably wanted to reply to me
<dtzWill>
I had that until I gave cloudfront listbucket permissions.. but that said I also actually got 404's when trying your URL's above when they were ".nar" isntead of ".narinfo"
<dtzWill>
bennofs: you have an example of a 500 req? just curious
<bennofs>
dtzWill: I could treat every 500 as a 404 but that seems risky
<bennofs>
dtzWill: the 500's only happen sporadically, on a second run everything is fine
<bennofs>
dtzWill: and they only stay 500's for 5 mins, the cloudfronts cache expires and they are 404's again as expected
<clever>
qknight: can you link to your service on github?
jensbin has quit [(Quit: Leaving)]
Supersonic112 has quit [(Ping timeout: 258 seconds)]
<bennofs>
dtzWill: i will see if this run of nix-index produces another 500
ryanartecona has joined #nixos
<dtzWill>
bennofs: is there anything more in the response beyond 500? I think if it says "internal server error" or somethign you'er supposed to try again
<bennofs>
dtzWill: yes it says internal server error
<dtzWill>
and that's just how AWS works for whatever reason
<dtzWill>
lol
<bennofs>
dtzWill: but I cannot try again because cloudfront caches the 500
<dtzWill>
nix has code to retry when it sees those, for example
<bennofs>
dtzWill: Well I can try again but I'll receive the cached 500 for 5 mins
Supersonic112 has joined #nixos
<dtzWill>
bennofs: I believe those errors should be retried with exp backoff until you get a different response code or retry count is hit; hopefully this is better documented in AWS docs or something
<bennofs>
dtzWill: yes S3 says I should retry
<bennofs>
dtzWill: the thing is, I *cannot* retry because S3 is behind cloudfront, and cloudfront does not allow me to retry before 5 minutes are up because it caches the 500 response for exactly 5 minutes
<dtzWill>
bennofs: hmm maybe nixos's cloudfront caches 500's when it shouldn't (since for the S3 origin I don't think that's ever a valid response?). Shrug, folks who maintain it probably know more, I'm also very new to clcoudfront/s3 myself so :3
<dtzWill>
clever: oooo I was thinking about something similar myself
<bennofs>
clever: what is that
<dtzWill>
clever: since I have a problem with wanting to use S3/etc but keep a local copy when possible to avoid every build being sent to S3 and then requested back for all derivations using it as an input
<clever>
bennofs: this program will proxy all requests to the binary cache, and cache the replies in ram
<bennofs>
clever: doesn't nix already have a narinfo cache?
<LnL>
cache all the things!
<clever>
bennofs: nix will cache things in a sqlite database, but the ttl is fairly short, and it isnt shared between many machines
<clever>
and cache.nixos.org is often "slow"
siel has quit [(Remote host closed the connection)]
siel has joined #nixos
erictapen has quit [(Ping timeout: 255 seconds)]
<dtzWill>
clever: yeah prop'ing this up in my build cluster sounds great. Just curious: when would this be better than some general-purpose caching proxy?
<dtzWill>
well I do like that I can read its implementation lol
<qknight>
clever: i send you a link to it later have to run...
<qknight>
clever: thanks!
<dtzWill>
yo dawg I heard you likes caches, so I put a cache in your cache so you can cache while you cache
<clever>
dtzWill: the code is designed around a binary cache, it will treat the narinfo url's differently from the nar urls
<dtzWill>
*cache builds while you build caches
<dtzWill>
clever: oh, neat!
<clever>
dtzWill: all narinfo files go into a hashmap in ram, while the nar files are currently uncached (i plan to cache them to disk)
<gchristensen>
anyone have sonarpulse's contact info?
<gchristensen>
thank you bennofs ooh nice call lnl
Supersonic112 has quit [(Ping timeout: 246 seconds)]
<bennofs>
it seems sonarpulse is not in maintainers.nix :O
<gchristensen>
:o
<LnL>
hmm apparently not?
yegortimoshenko has joined #nixos
<bennofs>
gchristensen: but the git commits list the same email
<gchristensen>
`git show 3cb745d5a69018829ac15f7d5a508135f6bda123`
<bennofs>
oh that's another one
<gchristensen>
that is the one I wanted :
<gchristensen>
:)
Supersonic112 has joined #nixos
<yegortimoshenko>
how do i build a standalone nix derivation outside of tree? no matter what i try (`nix-build -a packageAttr`, `nix-build`) it fails with:
<yegortimoshenko>
error: cannot auto-call a function that has an argument without a default value (‘stdenv’)
<yegortimoshenko>
bennofs, clever: thanks! and by the way, about `import <nixpkgs> {}`; i use that idiom a lot, and as i understand it it relies on channel name. please correct me if i'm wrong. is there a more clear way to import this?
<aristid>
yegortimoshenko: what it relies on is the NIX_PATH
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] bcdarwin opened pull request #27451: simpleitk: init at 1.0.0 (note this only builds the C++ libs, not the Python/Lua/... bindings) (master...simpleitk) https://git.io/vQN2q
NixOS_GitHub has left #nixos []
<aristid>
yegortimoshenko: for example, on one of my machines, this is the NIX_PATH:
<aristid>
yegortimoshenko: so then in nix expressions, <nixpkgs> will simply be resolved to /nix/var/nix/profiles/per-user/root/channels/nixos/nixpkgs (which is a channel)
<aristid>
you can add your own directories and files in there, too :)
<clever>
tnks: everything ran after this point up setup.sh will then be visible
<yegortimoshenko>
tilpner: when my nixpkgs commits get merged, i'd like to run `nix-channel --update` and immediately get the version with my patches, and nixos.org/channels/nixpkgs-unstable seems to lag quite a bit
<tnks>
clever: nice, thanks.
<clever>
tnks: for both python and gcc, nix will mutate the search path (the include path or pythonpath), based on the buildInputs
<clever>
tnks: but ghc is very different, you use the ghcWithPackages function to create a special instance of ghc, that has those packages in its default search path
<tnks>
clever: right, this is all in line with what I've either looked up or inferred. Nice to hear it from another person.
<tilpner>
yegortimoshenko - Running directly from master isn't recommended, though it might work for you if you're lucky. You would have less pain if you backported your changes onto a local stable nixpkgs checkout
<yegortimoshenko>
tilpner: i run nix-stable on production machines, but i want bleeding edge on my macOS laptop. i don't use too many packages, so i should be fine. thanks :-)
<clever>
yegortimoshenko: there are no stable channels for darwin, only nixpkgs-unstable is officialy supported
jsgrant has quit [(Quit: Peace Peeps. o/ If you need me asap, message me at msg[(at)]jsgrant.io & I'll try to get back to you within 24-36 hours.)]
<yegortimoshenko>
clever: true, but my production machines are NixOS, not macOS :-)
Ivanych has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] aristidb pushed 1 new commit to wireguard-private-keys: https://git.io/vQN1H
<NixOS_GitHub>
nixpkgs/wireguard-private-keys 7cb69d6 Aristid Breitkreuz: wireguard: do not use '' quotes unnecessarily
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] TomSmeets opened pull request #27455: hid-listen: init at 1.01 (master...hidlisten) https://git.io/vQN1Q
NixOS_GitHub has left #nixos []
erictapen has joined #nixos
<tnks>
clever: I forget my exact motivations, but I merged a whole bunch of Haskell environments into one.
<tnks>
I think it related to getting everything working with cabal new-build.
rigelk has quit [(Quit: Lost terminal)]
rigelk has joined #nixos
oida_ has joined #nixos
oida has quit [(Ping timeout: 260 seconds)]
<srhb>
tnks: Does new-build add anything to an already nixified haskell build?
Nobabs27 has quit [(Quit: Leaving)]
ertes has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] aristidb pushed 1 new commit to wireguard-private-keys: https://git.io/vQNDy
<NixOS_GitHub>
nixpkgs/wireguard-private-keys 580607e Aristid Breitkreuz: wireguard: cannot use bashism in ExecStart
NixOS_GitHub has left #nixos []
[0x4A6F] has joined #nixos
tokudan has quit [(Quit: Leaving)]
Neo--- has joined #nixos
Neo-- has quit [(Ping timeout: 240 seconds)]
MoreTea has quit [(Ping timeout: 255 seconds)]
<tnks>
srhb: a lot of people are used to sandboxes, which I really think need to die.
<tnks>
srhb: so in that regard "new-build" is a step towards fixing them more properly.
<tnks>
clever: if you didn't already look at it, I wouldn't mind your review of that project (particularly the Haskell part).
yegortimoshenko has quit [(Remote host closed the connection)]
<copumpkin>
niksnut: I think mac2 is screwed up
<copumpkin>
ikwildrpepper: I think mac2 is screwed up :)
<copumpkin>
or at least there's a job failing with "cannot connect to ‘root@mac2’: ssh: connect to host 213.125.166.74 port 6002"
<copumpkin>
oh the disk is full
Ivanych has quit [(Ping timeout: 255 seconds)]
<sphalerite[m]>
Is there a way to create a temporary udev rule? I don't want to rebuild my system and then have that rule that I really don't want in effect permanently still sitting around
<clever>
sphalerite[m]: either add it to configuration.nix and "nixos-rebuild test", then undo the config change
<clever>
oh, and rules.d is read-only, no other option
tvon_ has quit [(Remote host closed the connection)]
bennofs has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] aristidb pushed 1 new commit to wireguard-private-keys: https://git.io/vQNSh
himmAllRight has quit [(Quit: No Ping reply in 180 seconds.)]
tvon has joined #nixos
ryanartecona has quit [(Quit: ryanartecona)]
himmAllRight has joined #nixos
ryanartecona has joined #nixos
<sphalerite[m]>
Oh no, I'm XYing
<sphalerite[m]>
I want to pass a USB stick through to a VM
MoreTea has joined #nixos
m0rphism has joined #nixos
seppellll has quit [(Ping timeout: 240 seconds)]
<sphalerite[m]>
I've tried using qemu's `-usbdevice host:1234:abcd` option to pass it through, and using `setfacl` to give myself access to the device, but it gives a bunch of I/O errors in the VM
<clever>
i would just chown the dev node on the host, no need to mess with udev
tvon has quit [(Ping timeout: 246 seconds)]
<sphalerite[m]>
yeah, that's what I did (except using ACLs instead of chown)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peti pushed 4 new commits to master: https://git.io/vQN9h
<NixOS_GitHub>
nixpkgs/master dd78249 Peter Simons: LTS Haskell 8.23
<NixOS_GitHub>
nixpkgs/master 6de60af Peter Simons: hackage-packages.nix: automatic Haskell package set update...
<NixOS_GitHub>
nixpkgs/master 47feb45 Peter Simons: haskell-hsyslog: update version override...
NixOS_GitHub has left #nixos []
<clever>
sphalerite[m]: also, i would just use -drive to create a virtual drive, then dd it to the usb stick after its shutdown
<sphalerite[m]>
is there a way that having r/w access but not ownership could affect it?
bennofs has quit [(Ping timeout: 246 seconds)]
<clever>
qemu does also have an option to emulate a usb drive if it still needs to be over usb
<clever>
dont think so
<sphalerite[m]>
I don't want to write all of 8GB to the slow USB stick every time I make a change
<sphalerite[m]>
hence not using an image
<clever>
ah
<clever>
you can also -drive /dev/sdc i think
<clever>
to just directly expose the block device as a whole
<sphalerite[m]>
"must specify either driver or file"
<sphalerite[m]>
so file=/dev/sda
<clever>
-drive index=3,id=usb,file=/dev/sdc
<clever>
yeah
<clever>
may also need a format=raw
<sphalerite[m]>
yes
<clever>
in theory, if the guest writes the correct headers to the file, qemu can start to interpret it as a copy-on-write image of something else, and read an attacker controlled path
<sphalerite[m]>
oh dear, it's sda on the host and sdb in the VM. That's definitely not likely to result in a disaster :D
<clever>
i try to do everything with uuid or label, so it doesnt matter
<sphalerite[m]>
well it restricts write access on block 0 for that reason doesn't it
<clever>
and given that its usb, its very likely to jump around anyways
<NixOS_GitHub>
nixpkgs/wireguard-private-keys 04d7fc9 Aristid Breitkreuz: wireguard: improve phrasing of warning message as suggested by zx2c4
NixOS_GitHub has left #nixos []
<sphalerite[m]>
Well. I was right about the whole mismatch-resulting-in-disaster thing
<sphalerite[m]>
is there, by any chance, a way to get the partition table currently loaded in the kernel back onto the disk when it's been overwritten? :D
<the-kenny>
sphalerite[m]: you can just recreate the previous one and write it again
takle has joined #nixos
<sphalerite[m]>
what if I have no idea what the previous one was?
<the-kenny>
Then you can still use data recovery stuff to find the preambels of the filesystems on there :) As to your original question I'm not sure if that's possible
<sphalerite[m]>
now I just have to hope the GPT didn't also overwrite the filesystem headers. But I don't think it'll go all the way to block 34?
<sphalerite[m]>
actually, looks like the first block of the first partition is 2048. Maybe to avoid just the kind of destruction I was just worrying about
<the-kenny>
that's the default behavior of fdisk iirc
<clever>
the-kenny: ah nice, logs of juicy details in there
arjen-jonathan has joined #nixos
<sphalerite[m]>
clever: /proc/partitions doesn't seem to include the start, unlike /sys/block
<clever>
yeah
bennofs has joined #nixos
roconnor has joined #nixos
<roconnor>
Anyone know how I can customize my personal prompt without clobbering nix_shell's prompt?
<sphalerite[m]>
roconnor: personally I wouldn't bother avoiding clobbering it, and just use `$IN_NIX_SHELL` and `$name` to customise it
<sphalerite[m]>
(yay multiple negation)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] aristidb pushed 2 new commits to master: https://git.io/vQN5M
<NixOS_GitHub>
nixpkgs/master f74a1e6 Nicolas Pouillard: Keybase updates...
<Infinisil>
clever: But the nix store could be located somewhere else
<globin>
Infinisil: theoretically yes, but that would break 90% of nixpkgs currently
<LnL>
sore paths can't have spaces afaik
<Infinisil>
globin: Well we better don't make it worse then!
<gchristensen>
by all means, code defensively
<gchristensen>
I'm not going to say no :P
simukis has quit [(Ping timeout: 255 seconds)]
<Infinisil>
Better safe than sorry
<globin>
Infinisil: ack, nobody will say no, but I can hardly imagine we'll ever support that :)
<Infinisil>
I actually imagine supporting spaces in store paths wouldn't be too difficult, just replace every $out with "$out" and build everything to find other such scenarios
Coyin has quit [(Ping timeout: 260 seconds)]
<gchristensen>
as long as we keep in it our imagination and not reality I'm okay with that
<gchristensen>
:D
<globin>
and $dev and $lib and $man etc %)
<Infinisil>
++
Coyin has joined #nixos
<LnL>
Infinisil: there are other issues, like the way our cc wrapper works
<Infinisil>
LnL: That I have no idea about
[0x4A6F] has joined #nixos
<LnL>
like how include flags are propagated to gcc
<Infinisil>
LnL: Oh I see, the -lc and -I<path>
<Infinisil>
type of thing
<clever>
what about how the stdenv treats $buildInputs as a space seperated list of inputs?
<clever>
i'm pretty sure nix flattens the array down to a space seperated list, and then your hosed
<Infinisil>
clever: Ouch
<Infinisil>
Damn spaces, why don't we have AST editors yet, these would render these issues irrelevant
<LnL>
and it's very likely a lot of projects would break even if everything within the nix world works
zraexy has quit [(Ping timeout: 240 seconds)]
<Infinisil>
Do you think we'll ever get rid of strings as source code?
<Infinisil>
Would take a long time if ever
<LnL>
I don't think so
<clever>
i have noticed that out of the box, template-haskell can only insert a plain String (a linked list of characters!) into the source, but not any other type
Ivanych has quit [(Ping timeout: 268 seconds)]
alex-v has joined #nixos
<bennofs>
clever: with TH you build an AST
tdotdoe has quit [(Ping timeout: 260 seconds)]
<alex-v>
Where in the code nix-env resets the timestamp of the files it copies to /nix/store to 1969-12-31? Thanks
<clever>
probably need to dive into more complex parts of it, or use the th-lift thing i saw somewhere
<bennofs>
clever: inserting Haskell as plain String isn't even easily possible with TH afaik
<clever>
bennofs: pretty sure only string worked with lift when i was messing with it
<bennofs>
you'd first have to parse the String into some AST and then convert that to the TH AST
<bennofs>
clever: that's just if you want to generate literals of certain types
ThatDocsLady has quit [(Quit: Arma-geddin-outta-here!)]
<clever>
yeah
<alex-v>
clever: Thanks! I am having this weird issue with a brand new RHEL7 server with non-root Nix admin user. The user owns /nix/store but can't install anything involving compilation (permission denied during chmod). If there is no compilation involved the packages install fine...
<clever>
alex-v: does the user own everything in the store?
<alex-v>
clever: yes, that's what is puzzling
<clever>
alex-v: if something was ran as root in the past, the ownership may be mixed
<clever>
alex-v: can you gist the entire error?
<alex-v>
clever: well binary packages install fine
radvendii has quit [(Ping timeout: 240 seconds)]
<alex-v>
clever: no access to gist from my network but I can try pasting...
<Infinisil>
alex-v: Only if it's like 1 or 2 lines
<clever>
alex-v: why can you not access gist?
cfricke has quit [(Quit: WeeChat 1.9)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] gebner pushed 1 new commit to master: https://git.io/vQAfm
<NixOS_GitHub>
nixpkgs/master 7ba1c7a Gabriel Ebner: iprover: 0.8.1 -> 2.5
NixOS_GitHub has left #nixos []
[0x4A6F] has quit [(Ping timeout: 240 seconds)]
gm152 has joined #nixos
<alex-v>
clever: example, I have this for git:
Filystyn has quit [(Quit: Konversation terminated!)]
<babs>
so I don't have to edit the xorn or i3 config essentially
<babs>
*xorg
<alex-v>
clever: which is unrelated
<clever>
alex-v: anything like apparmour/gesecurity/selinux active?
<alex-v>
clever: how do I check?
<clever>
those things are usually off by default
<alex-v>
clever: sestatus disabled
<clever>
i cant think of anything else to check
<alex-v>
clever: any way to put a breakpoint to where it tries to run that chmod?
Sonarpulse has joined #nixos
<clever>
alex-v: id just throw strace at it, but that will generate 1000's of lines of logs, and then you need a pastebin
<tnks>
clever: is there a good way to merge a Python environment (.env derivation) with a Haskell environment (.env derivation)?
sidpatil has joined #nixos
<tnks>
I'm not sure there is, because I remember there not being a great way to merge two Haskell environments, which is why I ended up writing the code I did.
<clever>
tnks: no way to merge derivations, you would need to put the ghcWithPackages into the buildInputs of a new derivation, that also depends on the python things
<avn>
Folks, anyone have ready solution for NetworkManager + NFS mounts in nixos?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] globin pushed 1 new commit to master: https://git.io/vQATq
<NixOS_GitHub>
nixpkgs/master 46ead28 Robin Gloster: gnome3.gexiv2: 0.10.4 -> 0.10.6
NixOS_GitHub has left #nixos []
cpennington has quit [(Remote host closed the connection)]
<tnks>
clever: yeah. I need to look at the code for .env and see what it might do beyond a ghcWithPackages. I think it also puts in stand-alone tools needed for compilation.
phinxy has quit [(Read error: Connection reset by peer)]
<tnks>
the docs say something about build-time versus run-time.
<tnks>
I didn't get that.
<clever>
tnks: so if somebody ever tries to cross-compile your package 2 years down the road, it will try to make a build of cmake for the wrong platform, then run that on a cpu that cant run it
<yegortimoshenko>
clever, jtojnar: got it. i will try to get rid of it along with other changes i'm preparing for fpc (version bump, new platforms). update walker seems to be defunct, maybe there is a way to make it run again but i honestly think that small set of heuristics might be a better way than a per-derivation recipe
digitus has quit [(Quit: digitus)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] luigy opened pull request #27459: Fix #27448 argument list too long in fetchzip for darwin. (release-17.03...ll-fix-fetchzip-darwin_17.03) https://git.io/vQAmp
NixOS_GitHub has left #nixos []
Coyin has quit [(Remote host closed the connection)]
Coyin has joined #nixos
<jtojnar>
clever: The package should be gcr, there is gcr-base-3.pc in /nix/store/*-gcr-3.20.0/lib/pkgconfig, and I have it in buildInputs and it is being built, not sure why it is not propagated to the expression
yegortimoshenko has left #nixos ["ERC (IRC client for Emacs 25.2.1)"]
<clever>
jtojnar: is pkgconfig also in the builtInputs?
phinxy has quit [(Read error: Connection reset by peer)]
phinxy has joined #nixos
phinxy has quit [(Read error: Connection reset by peer)]
phinxy has joined #nixos
phinxy has quit [(Read error: Connection reset by peer)]
<jtojnar>
clever: in pure nix-shell in `configurePhase`, I see "configure: WARNING: pkg-config could not find the module gcr-base-3" but gcr-base-3 is listed in `pkg-config --list-all`
<clever>
jtojnar: what if you try to ask pkg-config for the ldflags of gcr-base-3?