<digitalmentat> asymmetric, you should make it -I nixpkgs=~/code/nixpkgs
<digitalmentat> NIX_PATH is composed of k/v pairs
teh__ has quit [(Ping timeout: 255 seconds)]
Intensity has quit [(Ping timeout: 256 seconds)]
<asymmetric> digitalmentat: true. but i still get the same error
<digitalmentat> :-/
<digitalmentat> hmm
<digitalmentat> oh wait
<digitalmentat> well, that shouldn't matter
<digitalmentat> hmm
<digitalmentat> it might be that trailing slash?
<digitalmentat> I just tried and it works fine for me
<digitalmentat> or it might be your ~ qualification
<asymmetric> nix-env -I nixpkgs=/home/asymmetric/code/nixpkgs -i polybar
<asymmetric> error: selector ‘polybar’ matches no derivations
<asymmetric> :(
<digitalmentat> and it's added to the nixpkgs tree in your fork??
<tilpner> asymmetric - Does it have a name = "polybar; line?
<tilpner> Oops, missing quote"
<gchristensen> trying to grok the remote buuilder perl is insane
<digitalmentat> gchristensen, yeah...
<Dezgeg> isn't there the equivalent c++ version nowadays? ;)
<gchristensen> don't think so
<asymmetric> tilpner: it's name = "polybar-${version}";
<Dezgeg> in nix master
<digitalmentat> asymmetric, have you added the derivation to the packages tree and in the top-level hierarchy too?
<asymmetric> digitalmentat: yes, i'm working off of this guy's PR https://github.com/NixOS/nixpkgs/pull/22222/files
<gchristensen> ah, yes
<joachifm> asymmetric: have you tried nix-env -f /foo/bar -iA polybar
<joachifm> asymmetric: notice the -A to select by attrpath instead of name
<NixOS_GitHub> [nixpkgs] fpletz pushed 7 new commits to staging: https://git.io/vDTti
<NixOS_GitHub> nixpkgs/staging 27d21e9 Franz Pletz: npth: 1.2 -> 1.3
<NixOS_GitHub> nixpkgs/staging 4b2731d Franz Pletz: lmdb: 0.9.18 -> 0.9.19
<NixOS_GitHub> nixpkgs/staging 4236364 Franz Pletz: libgpgerror: 1.24 -> 1.26
<asymmetric> joachifm: that worked
<NixOS_GitHub> [nixpkgs] fpletz pushed 5 new commits to master: https://git.io/vDTtX
<NixOS_GitHub> nixpkgs/master ff7c6fe Franz Pletz: youtubeDL: 2017.01.24 -> 2017.01.28
<NixOS_GitHub> nixpkgs/master eadb9ce Franz Pletz: notmuch: 0.23.4 -> 0.23.5
<NixOS_GitHub> nixpkgs/master 0adf664 Franz Pletz: neomutt: 20161126 -> 20170128
mizu_no_oto has joined #nixos
<asymmetric> joachifm: but still, when i do `nix-env -u`, i get
<asymmetric> error: attribute ‘polybar’ missing, at /home/asymmetric/code/dotfiles/config.nix:45:15
<asymmetric> if i leave the overrides in the main config.nix, which is the main issue i was trying to solve :)
<gchristensen> shlevy: I more deeply appreciate the work you did re perl now
<digitalmentat> asymmetric, `nix-env -I test-nixpkgs=/home/asymmetric/code/nixpkgs -i polybar -f '<test-nixpkgs>'` just worked for me; though I got a undefined variable error
mizu_no_oto has quit [(Ping timeout: 255 seconds)]
<digitalmentat> well, that path to nixpkgs was my own path but that seems to work
<digitalmentat> `error: undefined variable ‘libmsgpack_1_4’ at /home/parnell/Development/oss/nixpkgs-crough/pkgs/top-level/all-packages.nix:15349:18`
<digitalmentat> asymmetric, ^
<Ralith> is there a way to force nix to rebuild a package?
<Ralith> I need to debug something not outputting what it's supposed to
<asymmetric> digitalmentat: nope, still get the warning "warning: Nix search path entry ‘~/code/nixpkgs’ does not exist, ignoring", then error "error: file ‘test-nixpkgs’ was not found in the Nix search path (add it using $NIX_PATH or -I)"
<Tarinaky> Ralith: nix-env -e <foo> && nix-garbage-collect && nix-env -i <foo> ?
<gchristensen> w00t watching TU Delft participate in the hyperloop: http://www.spacex.com/hyperloop
<digitalmentat> asymmetric, can you show me your full invocation plz because it's working for me (I checked out the repo and tracked the branch to test)
<Ralith> Tarinaky: without side-effecting my entire system, preferably :P
<asymmetric> digitalmentat: nix-env -I test-nixpkgs=~/code/nixpkgs -i polybar -f '<test-nixpkgs>'
eacameron has joined #nixos
<asymmetric> commit 458f0d9acea9b22f2dbd745f6ec1aff87f02b68e
<Ralith> Corngood: got a new patch tested, just sorting out some changes in glslang now
<corngood> Ralith: build in nix-shell?
<digitalmentat> try using a path that doesn't require expanding ~ just in case, I'm using `nix-env -I test-nixpkgs=/home/parnell/Development/oss/nixpkgs-crough ...`
p0le has quit [(Quit: WeeChat 1.6)]
<corngood> Ralith: sweet. Can you tag me when you make a PR?
<Ralith> will do
<digitalmentat> asymmetric, nix-env -I test-nixpkgs=/home/parnell/Development/oss/nixpkgs-crough -i polybar -f '<test-nixpkgs>'
<digitalmentat> that is what mine looks like and I'm getting an error from all-packages.nix within that fork
<Ralith> Corngood: how can I test the install process in a shell, though? I don't have write access
<Ralith> I guess I can manually clobber DESTDIR
<corngood> Ralith: that is a good question. I want to say change $out, but that doesn't work, does it?
<Ralith> no idea
<Ralith> I'll try that first
<Ralith> nope
<asymmetric> digitalmentat: yeah that was it, the ~ expansion
eacameron has quit [(Ping timeout: 240 seconds)]
markus1189 has joined #nixos
<taktoa> gchristensen: UIUC also has a team there, I think :)
<Ralith> DESTDIR worked fine
<corngood> Ralith: I'll have to remember that
<Ralith> issue reproduced
* Ralith digs
<Ralith> Corngood: to be clear: make DESTDIR=foo install
<Ralith> I don't think it works as an env var
<gchristensen> taktoa: y'all are everywhere!
<taktoa> I only know that because I used to be an aerospace major
markus1199 has quit [(Ping timeout: 240 seconds)]
digitalmentat has quit [(Quit: Leaving)]
<gchristensen> ah
<gchristensen> :(
Intensity has joined #nixos
<taktoa> funnily enough, back when I was an aerospace engineering major (I'm in CS now), I thought about studying abroad at TU Delft (it would have reduced my tuition significantly)
Sonarpulse has quit [(Ping timeout: 255 seconds)]
<gchristensen> haha
<gchristensen> gosh!
ebzzry has joined #nixos
<gchristensen> what a small world
mizu_no_oto has joined #nixos
<gchristensen> well ... damn ...
<gchristensen> I was hoping to use NIX_BUILD_HOOK and "proxy" to `nix-store --realize` as a way of pretty-printing build output, but no such luck. the store locks stymies me.
Drakonis has joined #nixos
mizu_no_oto has quit [(Ping timeout: 260 seconds)]
<Benrob0329[m]> Browsing samba almost works in Dolphin
<Benrob0329[m]> it can find workgroups, but cant go any further
<hodapp> okay, for some reason I cannot figure out what the binary for HandBrake is called
<hodapp> and HandBrakeCLI won't run (error about libx265.so.79)... yet, XFCE is able to find the binary for HandBrake just fine
<hodapp> but I have no idea how to query XFCE for what it actually runs...
<hodapp> oh, thanks, ghb
mudri has quit [(Quit: WeeChat 1.6)]
MP2E has joined #nixos
<hodapp> okay, trying to figure out if Handbrake is just broken out of the box on NixOS, or if this DVD is just weird
mizu_no_oto has joined #nixos
Benrob0329[m] has left #nixos ["User left"]
Havvy has joined #nixos
<NixOS_GitHub> [nixpkgs] Ralith opened pull request #22284: vulkan-loader: 1.0.26.0 -> 1.0.39.1 (master...vulkan-loader) https://git.io/vDTYI
mizu_no_oto has quit [(Ping timeout: 264 seconds)]
mizu_no_oto has joined #nixos
<hodapp> ugggh nevermind, it's just region coding bullshit on the DVD drive
shanemikel has joined #nixos
<shanemikel> what are the first steps I can take to reconcile a collision? So far the output gives paths to the actual colliding derivations, but It's a dependency, not directly in my environment
<shanemikel> It'd be nice to trace back and list all the explicitly installed derivations that depend on either of the conflicting ones
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
justan0theruser has quit [(Ping timeout: 276 seconds)]
justanotheruser has joined #nixos
mizu_no_oto has joined #nixos
<hexa-> does any bootloader in nix support ia32 efi with x86_64 hostsystem?
<Ralith> also attempting to upstream our glslang patch while I'm at it: https://github.com/KhronosGroup/glslang/pull/706
<corngood> Ralith: looks good. Is that glslang change what broke between 1.0.26 and ~1.0.30?
<Ralith> Corngood: it is a thing that broke
<tippenein> trying to determine what this nixops issue is related to. Imagine it's an EC2 issue but anyone had nixops hang on "waiting for SSH" ? In the EC2 terminal it shows the machine as successfully built
<Ralith> I think the function your patch applied to got substantially refactored in trivial ways so I just rewrote it
<corngood> Ralith: Just looking at your updated PR now
<corngood> *update
shanemikel has quit [(Read error: Connection reset by peer)]
shanemikel has joined #nixos
mbrgm has quit [(Ping timeout: 240 seconds)]
mbrgm has joined #nixos
Drakonis has quit [(Ping timeout: 256 seconds)]
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
rmeis has joined #nixos
derjohn_mobi has joined #nixos
rmeis has quit [(Ping timeout: 245 seconds)]
derjohn_mob has quit [(Ping timeout: 240 seconds)]
<corngood> Ralith: any problems with the new loader? Do the demos and smoke test work?
<Ralith> Corngood: smoke test works, vulkaninfo works, cube demo segvs but so does the old one so I have no idea what's going on there
<corngood> Needs to have cwd set to bin/
avn has quit [(Ping timeout: 240 seconds)]
<corngood> Maybe other problems too
<Ralith> oh right
<Ralith> cube works too
<corngood> Smoketest is the important one
<Ralith> assuming it's supposed to be spinning stupidly fast anyway
<corngood> Which driver to you use?
<Ralith> intel
<corngood> AMD is still a bit of a mess right now, and that's all I can test on.
Wizek has joined #nixos
<Ralith> I shouldn't be touching any particularly fragile code paths, anyway
<Ralith> either it finds the manifests or it doesn't
<corngood> Yeah, I'm sure it's fine. Nothing driver specific.
<Ralith> hm
<Ralith> I actually broke the demo output, sec
shanemikel has quit [(Quit: Textual IRC Client: www.textualapp.com)]
<corngood> The demos can take --validate to test the layers
shanemikel has joined #nixos
<Ralith> Corngood: fixed the demo build
dbmikus has joined #nixos
<Ralith> hmm
<Ralith> still slightly broke
<Ralith> apparently multiple outputs can't contain the same binary; not sure why that is
<Ralith> but easily fixed
<Ralith> hokay
<Ralith> Corngood: now all the demos definitely work and vulkaninfo can be installed and so forth
Fare has joined #nixos
Wizek has quit [(Ping timeout: 260 seconds)]
Wizek has joined #nixos
eacameron has joined #nixos
mbrgm has quit [(Ping timeout: 276 seconds)]
mbrgm has joined #nixos
nhooyr has quit [(Remote host closed the connection)]
nhooyr has joined #nixos
acarrico has quit [(Ping timeout: 264 seconds)]
nhooyr has quit [(Remote host closed the connection)]
nhooyr has joined #nixos
<NixOS_GitHub> [nixpkgs] fpletz pushed 1 new commit to master: https://git.io/vDTGM
<NixOS_GitHub> nixpkgs/master c225750 Franz Pletz: weechat: 1.6 -> 1.7
<nhooyr> anyone know where the source of pkgs.fetchFromGitHub is?
<corngood> Ralith: sounds good
<fpletz> nhooyr: it's in pkgs/top-level/all-packages.nix
<nhooyr> ah, thanks
<fpletz> but it just uses fetchgit or fetchzip
<fpletz> those are in pkgs/build-support
<nhooyr> yup thanks
cstrahan has quit [(Quit: Connection closed for inactivity)]
acarrico has joined #nixos
sellout- has quit [(Quit: Leaving.)]
<ToxicFrog> Welp. Looks like digikam is broken in 16.09
<ToxicFrog> Cannot mix incompatible Qt library (version 0x50601) with this library (version 0x50602)
<ToxicFrog> Looks like it's trying to load both qtbase 5.6.2 and qtbase 5.6.1 for some reason
<ToxicFrog> Is there some way I can query the system dependency graph? I want to see why 5.6.1 is ending up in the current system configuration.
mizu_no_oto has joined #nixos
<ToxicFrog> Hmm. Digikam is linked against 5.6.2, but /run/current-system/sw/lib contains only 5.6.1.
alx741_ has joined #nixos
<clever> it shouldnt care about whats in /run/current-system
<clever> RPATH i what matters
<ToxicFrog> clever: well, strace shows it loading both and the error message is consistent with that
<clever> what does ldd say?
<ToxicFrog> At a guess it's picking up a plugin or something that's linked against 5.6.1
alx741 has quit [(Ping timeout: 240 seconds)]
<ToxicFrog> ldd shows it linked only against 5.6.2, like I said
<clever> can you pastebin the strace log?
shanemikel has quit [()]
shanemikel[m] has joined #nixos
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
<ToxicFrog> clever: http://ix.io/1RsE
<ToxicFrog> That's with -e trace=file | fgrep .so, I can paste the whole sordid thing if you want
<ToxicFrog> But at the end you see it reading both the 5.6.1 and 5.6.2 versions of a bunch of libraries
<Ralith> Corngood: glslang patch got merged already :D
hexagoxel has quit [(Ping timeout: 240 seconds)]
<Ralith> that'll be one less bit of maintennace next time we update the loader
<Ralith> (until it inevitably gets broken by some other committer, anyway...)
<Ralith> (judging by the git logs, the cmake files are not tested very carefully)
<clever> ToxicFrog: ah, only the plugins are being loaded via /run/current-system
Wizek has quit [(Ping timeout: 240 seconds)]
<clever> ToxicFrog: and then the rpath of those plugins comes into play, and i see
<ToxicFrog> This still raises the question of why *anything* in /run/current-system is still depending on qt 5.6.1
<clever> nix-store -qR /run/current-system/sw/lib/qt5/plugins/platforms/libqxcb.so
<ToxicFrog> I thought everything just had pkgs.qtbase or similar as its dependency -- indeed, I didn't think it was possible to depend on a specific version without creating a separate derivation for it
<ToxicFrog> clever: output includes: /nix/store/icqimlklxskxhgg4x5q6zay0m3c3a19d-qtbase-5.6.1-1
mguentner has quit [(Quit: WeeChat 1.6)]
<clever> cant think of what else to check right now
Wizek has joined #nixos
hexagoxel has joined #nixos
<ToxicFrog> It looks like that file is provided by qtbase, so now the question is, why does the system-path derivation depend on qtbase 5.6.1?
mguentner has joined #nixos
<tilpner> How do I quickly find out which package provides e.g. libasound.so.2?
<tilpner> It's easy for executables, but I don't know where to look for libraries
<ToxicFrog> It would be nice if Nix had a "why does X depend on Y" tool.
<simpson> ToxicFrog: $ nix-store -q ...
<simpson> ToxicFrog: Depends on the kind of query you want.
<ToxicFrog> simpson: I want it to show me the chain of dependencies that results in 'qtbase-5.6.1' being dependended on by 'system-path'
<ToxicFrog> Or, equivalently, I want to know what chain of dependencies from systemPackages results in a dependency on that specific version of qtbase.
<ToxicFrog> Because `pkgs.qtbase` is version 5.6.2, and AFAICT that means only 5.6.2 should be part of the system derivation.
<ToxicFrog> Huh. nix-env -qa --description qtbase lists six different versions of qtbase o.O
<simpson> ToxicFrog: Is --tree not enough? I know that it's a little hard to read, but it's easy enough to scroll rapidly through if you're just looking for that one chain.
<ToxicFrog> I didn't know --tree existed! That's why I'm asking :)
<simpson> I didn't see anything else in the man page.
<simpson> I mean, I saw many things in there, but not anything closer to what you want.
<ToxicFrog> --graph maybe, but graphviz isn't easy to grep
<ToxicFrog> If I'm reading this right qtbase-5.6.1 depends on itself, which is kind of odd
<ToxicFrog> But it looks like it's pulled in by a bunch of things -- ksshaskpass, ki18n, kwallet, kcoreaddons, kwindowsystem
mizu_no_oto has joined #nixos
* ToxicFrog digs deeper
<ToxicFrog> ksshaskpass depends on kdelibs
Wizek_ has joined #nixos
<ToxicFrog> kdelibs...depends on qt4? I must be looking at the wrong file
<ToxicFrog> Oh, wait, no, kdelibs is an alias for kde4.kdelibs
<ToxicFrog> But that's QT 4.8.7...this is a red herring
<ToxicFrog> Ok, yes, I was looking at the wrong file
<ToxicFrog> wow github search is awful, I'm going to need to clone this
* ToxicFrog bonks head against desk
<ToxicFrog> So it looks like the root of the problem
<clever> ToxicFrog: you can also nix-store -q --tree /run/current-system
<clever> ToxicFrog: and then earch for the problem version
<ToxicFrog> is that I'm using digikam5, which is in unstable
<ToxicFrog> But the rest of my system is from 16.09
kier has joined #nixos
<clever> ah
<clever> then you may want to --tree both and confirm where the conflicting version is from
<ToxicFrog> This could have been solved if I'd just looked at the right packages.nix
<ToxicFrog> clever: I'm not done tracking everything down, but it looks like the problem is that KDE5 in unstable bumped to QT 5.6.2, and KDE5 in 16.09 is still on 5.6.1
<ToxicFrog> So unstable.digikam5 pulls in one version and pkgs.plasma-desktop (or any other qt app) pulls in another
<clever> ah
<ToxicFrog> And I can't install digikam5 from 16.09 because it's not *in* 16.09
<clever> copy its .nix file to your packageOverrides
<clever> thenyou can build it against the libs in 16.09
mguentner2 has joined #nixos
mguentner has quit [(Ping timeout: 248 seconds)]
* ToxicFrog nods
<ToxicFrog> I decided to see if I could get clever and instead do pkgs.lib.callPackageWith pkgs (import <nixos-unstable>/pkgs/applications/graphics/digikam/5.nix) {}
<ToxicFrog> Sadly, it blows up in makeOverrideable with "value is a path while a set was expected"
<clever> skip the import
<clever> the callPackage function can import things for you
<clever> pkgs.callPackage <nixos-unstable>/pkgs/applications/graphics/digikam/5.nix {}
<clever> this should probably do it
<ToxicFrog> That's what I tried first and it died in exactly the same way
<clever> what does <nixos-unstable> eval to in nix-repk?
<ToxicFrog> /nix/var/nix/profiles/per-user/root/channels/nixos-unstable
<ToxicFrog> huh
<ToxicFrog> If I replace <nixos-unstable> in the call to callPackageWith with that path in full, it works
<ToxicFrog> Or at least, it gets as far as importing the file and then exploding because 16.09 makeQtWrapper and unstable makeQtWrapper have different signatures ;.;
<ToxicFrog> I'm going to have to copy this into overrides and make edits anyways, it looks like
<clever> yeah
<ToxicFrog> Hmm, no, misread the error message -- it's complaining that I'm not passing 5.nix makeQtWrapper in its arguments.
eacameron has quit [(Remote host closed the connection)]
shanemikel_ has joined #nixos
shanemikel_ has quit [(Client Quit)]
shanemikel_ has joined #nixos
<clever> ToxicFrog: ah, you want pkgs.qt5.callPackage
<NixOS_GitHub> [nixpkgs] nand0p opened pull request #22285: ElectricSheep: 2.6.8 -> 2.7b33 (master...electricsheep) https://git.io/vDTCP
eacameron has joined #nixos
sellout- has joined #nixos
eacameron has quit [(Ping timeout: 240 seconds)]
rmeis has joined #nixos
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
georges-duperon has quit [(Ping timeout: 252 seconds)]
rmeis has quit [(Ping timeout: 256 seconds)]
<taktoa> ToxicFrog: you nearly always want <nixos-unstable/foo/bar/baz> rather than <nixos-unstable>/foo/bar/baz
<NixOS_GitHub> [nixpkgs] dochang opened pull request #22286: yank 0.7.1 -> 0.8.0 (master...yank) https://git.io/vDTWn
eacameron has joined #nixos
eacameron has quit [(Remote host closed the connection)]
systemfault has quit [(Quit: Bye!)]
acarrico has quit [(Ping timeout: 260 seconds)]
ebzzry has quit [(Remote host closed the connection)]
eacameron has joined #nixos
Fare has quit [(Ping timeout: 255 seconds)]
acarrico has joined #nixos
<NixOS_GitHub> [nixpkgs] orivej opened pull request #22287: telegram-purple: 2016-03-17 -> 1.3.0 (master...telegram-purple) https://git.io/vDTl0
ebzzry has joined #nixos
<NixOS_GitHub> [nixpkgs] pikajude closed pull request #22247: Enable Vulkan for nvidia driver (master...nvidia-vulkan) https://git.io/vDUnP
ebzzry has quit [(Ping timeout: 276 seconds)]
fritschy has joined #nixos
eacameron has quit [(Remote host closed the connection)]
sdothum has quit [(Quit: ZNC - 1.6.0 - http://znc.in)]
ebzzry has joined #nixos
systemfault has joined #nixos
ebzzry has quit [(Ping timeout: 252 seconds)]
eacameron has joined #nixos
MercurialAlchemi has joined #nixos
reinzelmann has joined #nixos
eacameron has quit [(Ping timeout: 256 seconds)]
<NixOS_GitHub> [nixpkgs] veprbl closed pull request #22278: add applgrid, bump sherpa, add pyhepmc (master...sherpa_fix) https://git.io/vDUjp
<NixOS_GitHub> [nixpkgs] joachifm pushed 2 new commits to master: https://git.io/vDTBu
<NixOS_GitHub> nixpkgs/master 42d3f67 Desmond O. Chang: yank 0.7.1 -> 0.8.0
<NixOS_GitHub> nixpkgs/master b596cd5 Joachim F: Merge pull request #22286 from dochang/yank...
<taktoa> anyone know why https://nixos.org/irc/logs/ only goes up to August 2016?
michas_ has joined #nixos
acarrico has quit [(Ping timeout: 255 seconds)]
<c74d> taktoa: I think that logging service is no longer active; the channel topic includes a URL for the current logs
<taktoa> I know about botbot; it's too bad that nixos.org/irc/logs is no longer running, though, since I was going to run an ircbrowse instance off of them
t7 has quit [(Remote host closed the connection)]
rly has joined #nixos
rly has quit [(Ping timeout: 240 seconds)]
Kingsquee has joined #nixos
rly has joined #nixos
panostimos has joined #nixos
<NixOS_GitHub> [nixpkgs] vcunat deleted nvidia-vulkan at ffc498c: https://git.io/vDTRh
nathan7 has quit [(Ping timeout: 240 seconds)]
rly has quit [(Ping timeout: 255 seconds)]
nathan7 has joined #nixos
MichaelRaskin has quit [(Remote host closed the connection)]
<NixOS_GitHub> [nixpkgs] vcunat closed pull request #22268: ihaskell: remove service configuration. (master...remove-ihaskell) https://git.io/vDU1i
jgertm has joined #nixos
derjohn_mobi has quit [(Remote host closed the connection)]
pi3r has joined #nixos
johann__ has joined #nixos
johann__ has quit [(Read error: Connection reset by peer)]
derjohn_mob has joined #nixos
rly has joined #nixos
proteusguy has quit [(Ping timeout: 255 seconds)]
hotfuzz_ has joined #nixos
hotfuzz has quit [(Ping timeout: 240 seconds)]
joelpet has joined #nixos
civodul has joined #nixos
rly has quit [(Ping timeout: 240 seconds)]
johann__ has joined #nixos
johann__ has quit [(Quit: Leaving.)]
johann__ has joined #nixos
samvher has joined #nixos
teknico has joined #nixos
goibhniu has joined #nixos
MP2E has quit [(Quit: brb)]
eacameron has joined #nixos
samvher has quit [(Quit: leaving)]
derjohn_mob has quit [(Remote host closed the connection)]
eacameron has quit [(Ping timeout: 256 seconds)]
systemfault has quit [(Quit: Bye!)]
jgertm has quit [(Ping timeout: 240 seconds)]
ThatDocsLady has joined #nixos
uralbash has joined #nixos
ebzzry has joined #nixos
rmeis has joined #nixos
thc202 has joined #nixos
ebzzry has quit [(Ping timeout: 258 seconds)]
<mbrock> I don't understand how to use emacsWithPackages to get agda2-mode...
__Sander__ has joined #nixos
<NixOS_GitHub> [nixpkgs] vcunat pushed 1 new commit to master: https://git.io/vDT2E
<NixOS_GitHub> nixpkgs/master fd32b16 Vladimír Čunát: knot-dns: another attempt to fix build on Darwin...
<mbrock> oh, I should use "emacs25PackagesNg" instead of "emacsPackages"
derjohn_mob has joined #nixos
mongodb-native has joined #nixos
<NixOS_GitHub> [nixpkgs] frontsideair opened pull request #22289: flow: 0.37.1 -> 0.37.4 (master...flow-37.4) https://git.io/vDTaO
rly has joined #nixos
Kingsqueee has joined #nixos
Kingsquee has quit [(Read error: Connection reset by peer)]
t7 has joined #nixos
fadenb has quit [(Quit: WeeChat 1.6)]
fadenb has joined #nixos
<mongodb-native> Is there support for the native mongodb driver in nodejs when using NixOS? Does anyone have experience with this?
teh__ has joined #nixos
Tarinaky has quit [(Read error: Connection reset by peer)]
<NixOS_GitHub> [nixpkgs] pSub pushed 2 new commits to master: https://git.io/vDTrB
<NixOS_GitHub> nixpkgs/master e2f7d68 Orivej Desh: telegram-purple: 2016-03-17 -> 1.3.0
<NixOS_GitHub> nixpkgs/master 92a3ecf Pascal Wittmann: Merge pull request #22287 from orivej/telegram-purple...
xeviox has joined #nixos
eacameron has joined #nixos
<xeviox> is it possible to use docker-compose in current stable Nixos?
<tilpner> xeviox - Does it not work for you? I only installed it, didn't test it locally yet
<xeviox> tilpner: e. g. nixos.python35Packages.docker_compose?
<viric> hm I can't find when thin_provisioning_tools were deleted in master. They are there in 16.09
<viric> AHH
<viric> thin-provisioning-tools
<tilpner> xeviox - I installed the python27 version, but yes
<xeviox> I just wondered about the package name... "pythonXXPackages"
<xeviox> will give it a try, thanks
<mongodb-native> How come nodePackages.mongoose does not work even though I see it referenced in top-level/node-packages.json?
eacameron has quit [(Ping timeout: 245 seconds)]
<__Sander__> mongodb-native: library packages are not directly accessible
jgertm has joined #nixos
<__Sander__> the Nix expressions that are being generated basically capture the entire closure of dependencies of a specific tool
<__Sander__> node-packages.json only provides end user software, e.g. cli tools
<__Sander__> if you want to use mongoose in a project, then you should add it to your package.json config file and use node2nix directly with that package.json file
rly has quit [(Quit: leaving)]
<mongodb-native> __Sander__: because the issue I am having is that I am working around Nix for now without any issue, except I think I get quite a performance hit from not using the binary mongodb driver
<__Sander__> mongodb-native: I also use mongoose in a private project
<mongodb-native> __Sander__: but since it requires a correct build environment, it will fail to compile when working around Nix (obviously), I hoped I could install those binary packages through Nix
<__Sander__> but for me it works fine
johann__ has quit [(Quit: Leaving.)]
<__Sander__> which version of mongoose are you using? and which nodejs?
michas_ has quit [(Ping timeout: 256 seconds)]
<mongodb-native> __Sander: I am not even using mongoose, I just want the binary driver of mongodb, I just mentioned mongoose, because it depends on it and I hoped by installing it I would gain access to the binary driver as well
<mongodb-native> __Sander__: node v4.6.0
georges-duperon has joined #nixos
johann__ has joined #nixos
<mongodb-native> __Sander__: This is the warning I want to resolve (and hopefully get the application to be faster): js-bson: Failed to load c++ bson extension, using pure JS version
georges-duperon has quit [(Client Quit)]
<__Sander__> mongodb-native: and which version of the mongodb driver
<__Sander__> btw
<mongodb-native> __Sander__: I have been using npm by installing nodejs and setting globalconfig to something outside the nix store
rly has joined #nixos
<__Sander__> mongodb-native: I once created an example for disnix btw https://github.com/svanderburg/disnix-stafftracker-nodejs-example
<__Sander__> uses the mongodb driver directly
<__Sander__> but are you using node2nix to generate the expressions?
<__Sander__> or are you running npm directly?
<NixOS_GitHub> [nixops] rbvermaa pushed 1 new commit to master: https://git.io/vDTKj
<NixOS_GitHub> nixops/master 4671155 Rob Vermaas: Do not evaluate on 'nixops backup'. Allow backup to run on stopped/paused machines.
<asymmetric> how does one make edits to the nix manual?
<mongodb-native> __Sander__: directly because the project I am using is a mess, had to patch a lot of code manually
<__Sander__> ah I see
<__Sander__> then probably some build-time dependencies are missing
<__Sander__> what do you see in npm's output?
<__Sander__> you need to figure out what it's missing and add these dependencies to your environment
civodul has quit [(Quit: ERC (IRC client for Emacs 25.1.1))]
<musicmatze> Oh man, we have PRs from mid-2015 still open ... most of them are abandoned... maybe someone could mass-close these?
rmeis_ has joined #nixos
<goibhniu> good question asymmetric, I know the nixos manual has instructions, but I see nothing for nix: https://github.com/NixOS/nixpkgs/tree/master/nixos/doc/manual
<asymmetric> goibhniu: i see this https://github.com/NixOS/nix/tree/master/doc/manual, but it's not the same as https://nixos.org/nixpkgs/manual
<NixOS_GitHub> [nixpkgs] edolstra pushed 7 new commits to release-16.09: https://git.io/vDT6F
<NixOS_GitHub> nixpkgs/release-16.09 0d9cfd7 Shea Levy: linux: add 4.9...
<NixOS_GitHub> nixpkgs/release-16.09 8258b5c Tim Steinbach: kernel: 4.9.0 -> 4.9.1...
<NixOS_GitHub> nixpkgs/release-16.09 20999c1 Franz Pletz: linux: 4.9.1 -> 4.9.2...
<musicmatze> -.- we even have 3 PRs from 2014 -.-'
rmeis has quit [(Ping timeout: 258 seconds)]
<goibhniu> asymmetric: I guess this is it: https://github.com/NixOS/nix/blob/master/release.nix#L60
<NixOS_GitHub> [nixpkgs] edolstra closed pull request #22274: Backport Linux 4.9 to release-16.09 (release-16.09...linux-49-in-16.09) https://git.io/vDU7D
<goibhniu> and you can surely use shell.nix these, to get the docbook stuff
<mongodb-native> __Sander__: I will try that, already installed node-gyp, because it is mentioned here https://github.com/mongodb/node-mongodb-native but I am having trouble getting build feedback, since it seems to fail silently
<goibhniu> ^these^there
mudri has joined #nixos
<asymmetric> goibhniu: but it's not the same content as on https://nixos.org/nixpkgs/manual right?
<mongodb-native> __Sander__: Ah right, npm does not install binaries unless you specify it to do so globally, however it then tries to install in the lib folder of nodejs, which ofcourse is readonly: EROFS: read-only file system, mkdir '/nix/store/74pwx8kajpqr43d4dhfqs065c495wsqk-nodejs-4.6.0/lib/node_modules/mongodb'
<NixOS_GitHub> [nixops] rbvermaa pushed 1 new commit to master: https://git.io/vDTPB
<NixOS_GitHub> nixops/master 04ab6c5 Rob Vermaas: If security group is not found on destroy, assume it has been destroyed already.
<goibhniu> asymmetric: that's correct ... it should be the same as https://nixos.org/nix/manual/
mudri has quit [(Ping timeout: 264 seconds)]
ebzzry has joined #nixos
<asymmetric> so how does one edit https://nixos.org/nixpkgs/manual/ ?
<asymmetric> ah ok, it's at the bottom: https://nixos.org/nixpkgs/manual/
<asymmetric> makes sense
athan has quit [(Quit: leaving)]
<asymmetric> i was getting confused between nix manual and nixpkgs
<asymmetric> :)
Seichi has joined #nixos
<goibhniu> cool!
hamzaboulaares has joined #nixos
hamzaboulaares has quit [(Quit: Konversation terminated!)]
<mongodb-native> __Sander__: Ah, my bad, that was because on the server I had not configured globalconfig yet, so it still pointrf to the nix store, on my correctly configured desktop I do not get any errors though, but still no binary driver
hamzaboulaares has joined #nixos
hamzaboulaares has quit [(Client Quit)]
hamzab has joined #nixos
herzmeister has joined #nixos
finnegans_wake has joined #nixos
mudri has joined #nixos
herzmeister has quit [(Client Quit)]
<NixOS_GitHub> [nixpkgs] asymmetric opened pull request #22291: Make more visible which fetch approach is best (master...patch-4) https://git.io/vDT1L
<hamzab> Hello everyone. I am working on setting up a VLAN a nixos. I successfully created the VLAN interface a assigned an IP to it. The issue is that when I reboot the machine, the IP is no longer mapped to the interface. With some digging, I found that there is a service named network-addresses-vlan50.service in a failed state. When I start it manually the IP is fixed. Any idea about this?
<NixOS_GitHub> [nixpkgs] rasendubi pushed 2 new commits to master: https://git.io/vDT19
<NixOS_GitHub> nixpkgs/master 0ecde73 Lorenzo Manacorda: siji: init at 2016-05-14
<NixOS_GitHub> nixpkgs/master 2dc8788 Alexey Shmalko: Merge pull request #22283 from asymmetric/siji...
eacameron has joined #nixos
alx741 has joined #nixos
alx741_ has quit [(Ping timeout: 255 seconds)]
eacameron has quit [(Ping timeout: 264 seconds)]
<NixOS_GitHub> [nixpkgs] edolstra pushed 5 new commits to master: https://git.io/vDTM5
<NixOS_GitHub> nixpkgs/master c20cc6d Eelco Dolstra: Excise use of importJSON...
<NixOS_GitHub> nixpkgs/master 12dd086 Eelco Dolstra: Revert "gup: 0.5.5 -> 0.6.0 and derivation update"...
<NixOS_GitHub> nixpkgs/master 35349b6 Eelco Dolstra: gup: 0.5.5 -> 0.6.0
<NixOS_GitHub> [nixpkgs] edolstra pushed 2 new commits to master: https://git.io/vDTMh
<NixOS_GitHub> nixpkgs/master 4e0bb26 Lorenzo Manacorda: Make more visible which fetch approach is best...
<NixOS_GitHub> nixpkgs/master d478768 Eelco Dolstra: Merge pull request #22291 from asymmetric/patch-4...
lsix has joined #nixos
rmeis_ has quit [(Ping timeout: 264 seconds)]
<NixOS_GitHub> [nixpkgs] peti pushed 2 new commits to master: https://git.io/vDTDE
<NixOS_GitHub> nixpkgs/master 522ab7e Peter Simons: LTS Haskell 7.18
<NixOS_GitHub> nixpkgs/master a2018ce Peter Simons: hackage-packages.nix: automatic Haskell package set update...
rmeis has joined #nixos
Kingsqueee has quit [(Quit: https://i.imgur.com/qicT3GK.gif)]
Itkovian has joined #nixos
<NixOS_GitHub> [nixpkgs] vcunat pushed 1 new commit to master: https://git.io/vDTym
<NixOS_GitHub> nixpkgs/master 196b87f Vladimír Čunát: knot-dns: another attempt to fix build on Darwin
cpennington has joined #nixos
xadi has joined #nixos
xadi has quit [(Client Quit)]
xadi has joined #nixos
finnegans_wake has quit [(Quit: Leaving)]
<mongodb-native> __Sander__: I have managed to get build message after rebuild my node_modules folder, fixed the dependencies so far, but have trouble fixing ../lib/kerberos.h:5:27: fatal error: gssapi/gssapi.h: No such file or directory, I have installed libkrb5, which should fix it according to others, it is probably looking in the wrong lib folder, but how to fix that in this case
<__Sander__> mongodb-native: try adding the path to the C_INCLUDE_PATH environment
<__Sander__> variable
<__Sander__> maybe that helps
xadi has quit [(Quit: Leaving.)]
<hamzab> Any help with this issue? : I am working on setting up a VLAN a nixos. I successfully created the VLAN interface a assigned an IP to it. The issue is that when I reboot the machine, the IP is no longer mapped to the interface. With some digging, I found that there is a service named network-addresses-vlan50.service in a failed state. When I start it manually the IP is fixed. Any idea about this?
<mongodb-native> __Sander__: that does not work unfortunately, but I just tried installed bson again after fixing the deps for kerberos, and it worked! Since kerberos is optional in the latest version and since I do not really need it, it does not really matter.
<goibhniu> hamzab: I have no idea about that, but maybe posting your configuration.nix and the details of the error (from journalctl) will help
<mongodb-native> __Sander__: This has been an annoyance ever since I started this project. Thank you very much for you help!
<hamzab> goibhniu: this is the vlan config part
<hamzab> networking.vlans = {
<hamzab> vlan50 = {
<hamzab> id = 50;
<hamzab> interface = "eth2";
<hamzab> };
<hamzab> };
<hamzab> networking.interfaces.vlan50.ipAddress = "10.0.4.50";
<hamzab> networking.interfaces.vlan50.prefixLength = 24;
<hamzab> networking.defaultGateway = "10.0.2.2";
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<hamzab> this is the journalctl logs
<hamzab> [root@nixos-machine:~]# journalctl -u network-addresses-vlan50.service -f
<hamzab> -- Logs begin at Thu 2016-11-10 10:25:12 UTC. --
<hamzab> Jan 30 11:14:24 nixos-machine systemd[1]: Stopping Address configuration of vlan50...
<hamzab> Jan 30 11:14:24 nixos-machine network-addresses-vlan50-pre-stop[1075]: deleting 10.0.4.50/24...
<hamzab> Jan 30 11:14:24 nixos-machine systemd[1]: Stopped Address configuration of vlan50.
<hamzab> -- Reboot --
<hamzab> Jan 30 11:14:37 nixos-machine systemd[1]: Starting Address configuration of vlan50...
<hamzab> Jan 30 11:14:37 nixos-machine network-addresses-vlan50-start[688]: bringing up interface...
<hamzab> Jan 30 11:14:37 nixos-machine network-addresses-vlan50-start[688]: RTNETLINK answers: Network is down
<hamzab> Jan 30 11:14:37 nixos-machine systemd[1]: network-addresses-vlan50.service: Main process exited, code=exited, status=2/INVALIDARGUMENT
<hamzab> Jan 30 11:14:37 nixos-machine systemd[1]: Failed to start Address configuration of vlan50.
<hamzab> Jan 30 11:14:37 nixos-machine systemd[1]: network-addresses-vlan50.service: Unit entered failed state.
<hamzab> Jan 30 11:14:37 nixos-machine systemd[1]: network-addresses-vlan50.service: Failed with result 'exit-code'.
<goibhniu> hamzab: please use some pastebin service
<goibhniu> (in the future)
rmeis has quit [(Ping timeout: 260 seconds)]
<NixOS_GitHub> [nixpkgs] zimbatm pushed 3 new commits to master: https://git.io/vDT91
<NixOS_GitHub> nixpkgs/master 51dc13b Rahul Gopinath: ruby: cleanup unneeded files
<NixOS_GitHub> nixpkgs/master 1918095 Rahul Gopinath: ruby_1_9: remove package...
<NixOS_GitHub> nixpkgs/master 66b6c2c Rahul Gopinath: sdformat: mark as broken...
<NixOS_GitHub> [nixpkgs] zimbatm closed pull request #18385: ruby: remove 1.9 (WIP) (master...ruby) https://git.io/vinpx
johann__ has quit [(Quit: Leaving.)]
michas_ has joined #nixos
teh[m] has joined #nixos
mkoenig has joined #nixos
<gchristensen> niksnut++ for json excision
<LnL> gchristensen: I git a bit closer with the jobset evals, but it's still not showing up properly
<gchristensen> oh?
<gchristensen> how so?
<LnL> the jobset knows about all the builds, but the eval doesn't show up
<gchristensen> O.o
<LnL> so hydra thinks all the jobs where "removed"
<LnL> and I don't really see anything strange in the database
<gchristensen> that is very weird
<Havvy> Is there a way to make nix-shell check and update the packages it uses?
<LnL> I'm also a bit confused why all the builds and jobs reference belong to a jobset directly
<gchristensen> Havvy: close and reopen it?
<teh[m]> Havvy: you mean from within the shell? if so: no, you'll need to leave the shell and re-enter
Seichi has quit [(Ping timeout: 240 seconds)]
Seichi has joined #nixos
<NixOS_GitHub> [nixpkgs] florianjacob opened pull request #22293: tinc: 1.0.29 -> 1.0.31 (master...master) https://git.io/vDT7B
<Havvy> Okay, did a GC (and I guess I didn't set it as a GC root), and now it's redownloading everying.
ertesx has joined #nixos
<Havvy> But for some reason it's downloading Elixir v1.3.1 instead of Elixir 1.4.0 (which is what is attempted to be downloaded when I do `nix-env -iA nixpkgs.elixir`.
<gchristensen> you don't need to do a gc for it to update itself, just leave and re-enter... of course, it'll use the packages from the current version of channel you're on
<Havvy> Or maybe I did?
<gchristensen> sorry?
<Havvy> I dunno. :/ They're both using `nixpkgs` AFAICT.
<teh[m]> Havvy: do you have a custom shell.nix or are you using `nix-shell -p bla` ?
<gchristensen> 16.09 has 1.3.1 and unstable has 1.4.0
<Havvy> `with import <nixpkgs> {};` is at the top of the .nix file.
<Havvy> teh[m]: Custom.
ertes has quit [(Ping timeout: 240 seconds)]
ertesx is now known as ertes
khajvah_ has joined #nixos
<Havvy> nix-channel --list only shows "nixpkgs https://nixos.org/channels/nixpkgs-unstable"
<gchristensen> what if you run that as root?
Fare has joined #nixos
<teh[m]> Havvy: did you override NIX_PATH by any chance?
avn has joined #nixos
<Havvy> $NIX_PATH says "nixpkgs=/nix/var/nix/profiles/per-user/root/channels/nixos/nixpkgs:nixos-config=/etc/nixos/configuration.nix:/nix/var/nix/profiles/per-user/root/channels"
<Havvy> Which shows NixOS...
Itkovian has joined #nixos
<Havvy> So is nix-shell using $NIX_PATH and not nix-channel ?
<teh[m]> `nixpkgs=/nix/var/nix/profiles/per-user/root/channels/nixos/nixpkgs` should be from your channel, so both shell and nix-env would pick that up
rmeis has joined #nixos
<NixOS_GitHub> [nixpkgs] globin pushed 1 new commit to staging: https://git.io/vDT51
<NixOS_GitHub> nixpkgs/staging eaa9e01 Robin Gloster: Revert "binutils: Add lib output"...
<Havvy> teh[m]: But /per-user/root/ is using root's channel, not my user's channel.
<teh[m]> Havvy: ah right, I thought you were running as root
<Havvy> My user on the system is "havvy"
<NixOS_GitHub> [nixpkgs] layus closed pull request #22137: python2Packages.legitfs: init at 0.3 (master...add-legitfs) https://git.io/vMx7l
<ToxicFrog> Huh. I got all the callPackage stuff sorted out, but now it dies with "cannot convert a set to a string
<ToxicFrog> " on the lines:
<ToxicFrog> name = "digikam-${version}";
<ToxicFrog> version = "5.3.0";
<ToxicFrog> That sure looks like a string to me....
<gchristensen> can you paste your whole thing?
<Havvy> So why is $NIX_PATH using root instead of my own user?
<Havvy> I haven't sudoed or messed with environment variables since restarting the OS.
<Havvy> Nor do I have anything in .bashrc touching it.
<ToxicFrog> gchristensen: the whole thing is https://github.com/NixOS/nixpkgs/blob/master/pkgs/applications/graphics/digikam/5.nix, but I'm calling it in a nonstandard way
<teh[m]> Havvy: I don't use nix-channels so I have to defer to people who do, apologies. (I set my NIX_PATH explicitly everywhere)
<ToxicFrog> With `digikam5 = pkgs.kde5.callPackage /nix/var/nix/profiles/per-user/root/channels/nixos-unstable/pkgs/applications/graphics/digikam/5.nix {};`
<ToxicFrog> Since I'm trying to build it in the context of 16.09, since unstable digikam5 doesn't work on a 16.09 system
<gchristensen> hrmmm you probably need to build with --show-trace
ebzzry has quit [(Ping timeout: 240 seconds)]
<ToxicFrog> I am
JagaJaga has joined #nixos
<ToxicFrog> Oh, hrm
<ToxicFrog> Looking at the trace, it reports this "while evaluating the attribute 'nativeBuildInputs' of the derivation 'digikam-5.3.0'
<ToxicFrog> Which implies that the problem is not actually at line 47
<ToxicFrog> Or anywhere near it
<gchristensen> yeah, that
<ToxicFrog> Found the problem. I was passing marble to the digikam5 package wrong.
* ToxicFrog replaces it with "${pkgs.coreutils}/bin/false" for now and gambles that digikam will never actually want to call it
<gchristensen> lol
<ToxicFrog> welp, that bet failed, apparently it's needed for the build
mguentner2 is now known as mguentner
<ToxicFrog> Hrm. So I can pass it the marble pkg, but it complains that it can't find marble.unwrapped.
<ToxicFrog> I must be doing something wrong here.
<ToxicFrog> Ahaaaa! It's because the signature for kdeWrapper changed between 16.09 and master.
sellout- has quit [(Quit: Leaving.)]
johann__ has joined #nixos
<NixOS_GitHub> [nixpkgs] florianjacob opened pull request #22294: os-prober: 1.65 -> 1.71 (master...update_os-prober) https://git.io/vDTb5
<ToxicFrog> Ok, it looks like marble is only needed for the geolocation features, which I never use anyways
<ToxicFrog> Maybe it will be easier to remove the marble dep from digikam
sellout- has joined #nixos
<NixOS_GitHub> [nixops] rbvermaa pushed 4 new commits to master: https://git.io/vDTNO
<NixOS_GitHub> nixops/master 9569110 Amine Chikhaoui: adding cloudwatch log group/stream resources...
<NixOS_GitHub> nixops/master 5511613 Amine Chikhaoui: fixing indentation
<NixOS_GitHub> nixops/master ac49487 Rob Vermaas: Merge branch 'cloudwatch' of https://github.com/AmineChikhaoui/nixops into AmineChikhaoui-cloudwatch
<Havvy> teh[m]: `The nix-env command uses ~/.nix-defexpr` -- http://lethalman.blogspot.com/2014/09/nix-pill-15-nix-search-paths.html
<Havvy> So nix-env and nix-shell are looking in different places.
<teh__> Havvy: ah, good to know! Terrible user experience though ..
<Havvy> What isn't terrible UX in Nix though?
<gchristensen> :(
<gchristensen> could you open a bug for this?
fritschy has quit [(Ping timeout: 240 seconds)]
fritschy has joined #nixos
mudri has quit [(Ping timeout: 276 seconds)]
georges-duperon has joined #nixos
ebzzry has joined #nixos
<taktoa> ToxicFrog: did you see my message earlier about the difference between `<nixos-unstable/foo/bar/baz>` and `<nixos-unstable>/foo/bar/baz`? I don't think it's related to your issues, but it's a pretty subtle point that's good to know about
<ToxicFrog> taktoa: I did, and thank you! Using the latter worked.
sdothum has joined #nixos
<ToxicFrog> Havvy: I'm seriously considering writing a shell script ("nx"?) that sits on top of nix-env/nixos-rebuild/nix-channel/nix-repl/nix-store and gives you a nicer UI
proteusguy has joined #nixos
<taktoa> the difference is that the former will look in every definition of nixos-unstable=... in your NIX_PATH, checking if /foo/bar/baz is valid in each until it finds a valid one, whereas the latter will just get the first instance of nixos-unstable=... and then append /foo/bar/baz to that string (which may cause it to stop being a valid path)
<taktoa> so <nixos-unstable/foo/bar/baz> will be valid strictly more often than <nixos-unstable>/foo/bar/baz
<ToxicFrog> No-one wants to run `nix-store --query --references $(nix-instantiate '<nixpkgs>' -A foo)` rather than `nix-pkg depends foo`
<ToxicFrog> I like Nix, but UX-wise right now it feels a lot like using git using only plumbing commands
<ToxicFrog> taktoa: aah, that makes sense
<Havvy> So, what actually sets NIX_PATH?
<taktoa> it's an environment variable
<taktoa> set by /etc/profile.d/nix in non-NixOS systems
<Havvy> And on NIXOS?
<Havvy> NixOS **
<taktoa> not sure, but obviously somewhere in the bash-sourcing-tree
<taktoa> strace is your friend there
<avn> ToxicFrog: have nix/nixpkgs/nixos commands, with same dispatching style as git do-- nixos-$subcommand, etc
<avn> ToxicFrog: but better to wait when 1.12 will be released, it have new nix command
* ToxicFrog nods
<taktoa> I guess it could also be present in the environment variables somewhere else (e.g.: systemd gets run with those env vars, so all subprocesses get them too)
Itkovian has quit [(Ping timeout: 260 seconds)]
johann__ has quit [(Quit: Leaving.)]
<ToxicFrog> At some point I also want to add proper progress output to nix-build, but it looks like other people are also working on that and it may be finished before I get a chance to poke at it :)
<taktoa> this is why I don't like environment variables
<taktoa> btw, ToxicFrog, have you looked at https://github.com/NixOS/nix/tree/master/src/nix
<taktoa> that's the next-gen nix interface currently being worked on
<taktoa> the command will just be called "nix" and you can play with it by building pkgs.nixUnstable, I think
Itkovian has joined #nixos
<ToxicFrog> taktoa: I haven't, but I'll check it out when I get a chance
<taktoa> oh lol I just noticed that avn mentioned it too
<NixOS_GitHub> [nixpkgs] vcunat pushed 1 new commit to staging: https://git.io/vDTp2
<NixOS_GitHub> nixpkgs/staging db89615 Vladimír Čunát: binutils: minimize rebuilds after parent revert...
johann__ has joined #nixos
acarrico has joined #nixos
peti has quit [(Quit: WeeChat 1.6)]
peti has joined #nixos
ebzzry has quit [(Ping timeout: 260 seconds)]
ebzzry has joined #nixos
zml has quit [(Ping timeout: 276 seconds)]
civodul has joined #nixos
eacameron has joined #nixos
morsicus1 has joined #nixos
morsicus1 has quit [(Ping timeout: 240 seconds)]
eacameron has quit [(Remote host closed the connection)]
<gchristensen> garbas: so I spent some time and hacked out a custom nix-build for TravisCI: https://asciinema.org/a/9rj8spgomqb1ge3d5jj0fr2g6
<gchristensen> (the `sudo` isn't needed on travis)
pi3r has quit [(Ping timeout: 240 seconds)]
michas_ has quit [(Quit: Verlassend)]
teh__ has left #nixos []
mongodb-native has quit [(Quit: Page closed)]
rmeis has quit [(Ping timeout: 240 seconds)]
rmeis has joined #nixos
michiel_l has joined #nixos
<LnL> gchristensen: ooh, what's that?
<gchristensen> trying to make "normal" build tools be friendlier towards nix
<LnL> what's the difference with nox?
hamzaboulaares has joined #nixos
hamzab has quit [(Ping timeout: 240 seconds)]
Seichi has quit [(Ping timeout: 240 seconds)]
Seichi has joined #nixos
<gchristensen> it doesn't attempt to calculate any differences or guess what changed and isn't suitable for nixpkgs
<LnL> ah right
rmeis has quit [(Ping timeout: 258 seconds)]
<gchristensen> the target is people building 10 or so packages on their company jenkins
<gchristensen> or using nix as a test runner
c0bw3b has joined #nixos
pi3r has joined #nixos
<NixOS_GitHub> [nixpkgs] pSub closed pull request #22294: os-prober: 1.65 -> 1.71 (master...update_os-prober) https://git.io/vDTb5
<NixOS_GitHub> [nixpkgs] pSub pushed 2 new commits to master: https://git.io/vDkU9
<NixOS_GitHub> nixpkgs/master 97c3ee9 Florian Jacob: tinc: 1.0.29 -> 1.0.31
<NixOS_GitHub> nixpkgs/master 4396d43 Pascal Wittmann: Merge pull request #22293 from florianjacob/master...
viric_ has joined #nixos
neeasade has joined #nixos
viric has quit [(Ping timeout: 255 seconds)]
ryantrinkle has quit [(Ping timeout: 252 seconds)]
edvorg has joined #nixos
<jophish> grenade: I'm going to be writing a small haskell package for doing exactly that soon (for use with hercules). Mind if I use dobuilds.sh as a jumping off point?
<jophish> gchristensen: *
<gchristensen> not at all, lol, let me give it a license.
<jophish> oh, it's tiny!
<gchristensen> yeah
<gchristensen> the important thing about it is how it uses the nix build hook
<gchristensen> ok I added a license
xadi has joined #nixos
<jophish> thanks!
<gchristensen> you're welcome! :)
<gchristensen> please, take the idea and make something good :P
georges-duperon has quit [(Ping timeout: 252 seconds)]
<jophish> aye aye
c0bw3b has quit [(Quit: http://www.kiwiirc.com/ - A hand crafted IRC client)]
eacameron has joined #nixos
c0bw3b has joined #nixos
eacameron has quit [(Ping timeout: 255 seconds)]
nschoe has joined #nixos
<NixOS_GitHub> [nixpkgs] vcunat pushed 1 new commit to staging: https://git.io/vDkIy
<NixOS_GitHub> nixpkgs/staging cc07ec8 Vladimír Čunát: protobuf-3.x: fixup build with default gcc version...
viric_ is now known as viric
fritschy has quit [(Quit: Leaving.)]
digitalmentat has joined #nixos
mkoenig has quit [(Ping timeout: 276 seconds)]
mkoenig has joined #nixos
nschoe has quit [(Ping timeout: 256 seconds)]
nschoe has joined #nixos
Nadrieril has joined #nixos
mudri has joined #nixos
nschoe has quit [(Quit: Program. Terminated.)]
mkoenig has quit [(Read error: Connection reset by peer)]
<tnks> holy crap. I just realized that automating cabal2nix in a Nix expression is actually really easy and straight-forward.
ebzzry has quit [(Ping timeout: 240 seconds)]
<tnks> I was doing it as a manual and separate step.
<tnks> and with overriding, you can get as much programmatic manipulation of the generated derivation as you need.
ryantrinkle has joined #nixos
<gchristensen> neat!
<savanni> @tnks Really? What does that end up looking like? Also, hi, y'all!
<gchristensen> hiya savanni
mkoenig has joined #nixos
sellout- has quit [(Quit: Leaving.)]
sellout- has joined #nixos
nschoe has joined #nixos
thc202 has quit [(Ping timeout: 240 seconds)]
sellout- has quit [(Ping timeout: 255 seconds)]
hamzaboulaares has quit [(Remote host closed the connection)]
hamzaboulaares has joined #nixos
Rotaerk has quit [(Quit: Leaving)]
<gchristensen> anyone have a good puppet module thing for deploying nix with a daemon? :D
MercurialAlchemi has quit [(Ping timeout: 240 seconds)]
rmeis has joined #nixos
danharaj has joined #nixos
<tnks> savanni: I'm tinkering around with it still, but it definitely builds.
<tnks> I'm curious how much crap it puts in /nix/store.
<tnks> I'd hate for it to put my .git/ repo metadata in there.
<savanni> Maybe. But it doesn't seem so bad so long as the garbage cleanup deals with it correctly.
<NixOS_GitHub> [nixpkgs] vcunat pushed 1 new commit to master: https://git.io/vDkOc
<NixOS_GitHub> nixpkgs/master 7f7faab Vladimír Čunát: knot-dns: yet another attempt to fix build on Darwin
<tnks> savanni: yeah, I might accept it, but I'd prefer to prevent it.
<digitalmentat> tnks, is this for private Haskell packages or ones hosted on Hackage?
<tnks> digitalmentat: private ones.
<tnks> otherwise, I'd use hackage2nix.
<tnks> or callHackage
<tnks> (which is where I got the idea in the first place)
nschoe has quit [(Quit: Program. Terminated.)]
c0bw3b has quit [(Quit: http://www.kiwiirc.com/ - A hand crafted IRC client)]
thc202 has joined #nixos
eacameron has joined #nixos
mkoenig_ has joined #nixos
tv1 has joined #nixos
Nadrieril has left #nixos ["Goodbye"]
mkoenig has quit [(Ping timeout: 255 seconds)]
c0bw3b has joined #nixos
erasmas has joined #nixos
tv has quit [(Ping timeout: 246 seconds)]
eacameron has quit [(Read error: Connection reset by peer)]
k0001 has joined #nixos
<digitalmentat> tnks, how do you solve the problem of having hydra integrate from the individual project?
<digitalmentat> don't you need to run it anyway in the repo if you're going to build GC roots for your cache with Hydra?
<digitalmentat> I'm very curious to see your solution because if it does work, that would be really nice (we have the same problem, of course)
nschoe has joined #nixos
<k0001> If I run Hydra HEAD on a NixOS host, am I expected to run nixpkgs HEAD on that host as well (seeing as how Hydra depends on aversion of Nix that is not the one in 16.09)?
<k0001> s/nixpkgs HEAD/NixOS HEAD/ (although you get the idea...)
trupanka has joined #nixos
<gchristensen> k0001: no
xeviox has quit [(Ping timeout: 240 seconds)]
<gchristensen> k0001: here is how I run hydra on 16.09: https://github.com/grahamc/hydra-prs/blob/master/modules/hydra-master/default.nix where I do indeed use hydra's package from unstable (see: the top of the file, and line147)
<tnks> digitalmentat: we don't use Hydra, but what I'm doing is really straight-forward.
<gchristensen> k0001: note that my use of the nixpkgs unstable _requires a patch_, and yes, it really does
<NixOS_GitHub> [nixpkgs] rycee pushed 1 new commit to master: https://git.io/vDkGy
<NixOS_GitHub> nixpkgs/master 89150d3 Robert Helgesson: theme-vertex: 20161009 -> 20170128
<k0001> gchristensen: Thanks. I'm doing something similar, i was just checking. But.. isn't there any chance that having two different Nix versions around (the one coming from NixOS-16.09 and the one being used by Hydra) will screw up some state in `/nix`?
<gchristensen> no, because it talks to the daemon, and doesn't touch the store itself directly
<gchristensen> at worst it'll fail to work
mkoenig_ has quit [(Ping timeout: 256 seconds)]
<tnks> does /nix/store ever have garbage that won't clean with `nix-collect-garbage -d`?
<tnks> I can't find any roots pointing something in there.
<k0001> gchristensen: super useful info, thanks!
<NixOS_GitHub> [nixpkgs] rycee pushed 1 new commit to master: https://git.io/vDkZZ
<NixOS_GitHub> nixpkgs/master c603d2c Robert Helgesson: svtplay-dl: 1.9 -> 1.9.1
<gchristensen> it will clean up all garbage, anything it doesn't clean is because it isn't considerde garbage
<gchristensen> k0001: you're welcome! :)
<tnks> gchristensen: yeah, but is there a question I can ask as to what's keeping it around?
mkoenig has joined #nixos
<gchristensen> I think `nix-store -q --references <thepath>`?
<tnks> I figured it out. It was a terminal keeping alive a nix-shell.
<gchristensen> ah! :)
<tnks> that was super confusing.
edvorg has quit [(Ping timeout: 255 seconds)]
Sonarpulse has joined #nixos
estewei has joined #nixos
<trupanka> How to override configureFlags = ... in package with config.nix?
<tnks> okay, my method seems to work, but it copies the source into /nix/store twice.
<tnks> seems like a small price to pay.
<tnks> I think I'm gearing up for a code review after I clean this up.
<tnks> savanni: I'll make sure you see the link.
<NixOS_GitHub> [nixpkgs] rbvermaa pushed 1 new commit to release-16.09: https://git.io/vDkcF
<NixOS_GitHub> nixpkgs/release-16.09 5fff5a9 Rob Vermaas: boto: upgrade 2.41.0 to 2.45.0
<NixOS_GitHub> [nixpkgs] rbvermaa pushed 1 new commit to master: https://git.io/vDkCn
<NixOS_GitHub> nixpkgs/master b8fec23 Rob Vermaas: boto: upgrade 2.42.0 to 2.45.0...
eacameron has joined #nixos
rasen has quit [(Ping timeout: 276 seconds)]
octe has joined #nixos
<taktoa> does anyone have opinions on a good name for `foo = bar;` in an attribute set context (`{ foo = bar; }`)?
<taktoa> I've come up with "attribute pair" (kind of vague) and "attribute binding" (bad b/c nonrec attr sets don't affect scope)
rmeis_ has joined #nixos
<gchristensen> there probably is a real name for it
pi3r has quit [(Ping timeout: 252 seconds)]
<simpson> taktoa: The manual says that the correct name is "attribute".
<simpson> "Sets are just a list of name/value pairs (called attributes) enclosed in curly brackets, where each value is an arbitrary expression terminated by a semicolon."
<taktoa> ah, I was thinking that "names" should be called attributes, but that makes sense too
<simpson> Most languages would call them "keys"; I'm a little irked that Nix calls them "names" but whatever.
<NixOS_GitHub> [nixpkgs] alibabzo opened pull request #22295: neofetch: 3.0 -> 3.0.1 (master...update/neofetch) https://git.io/vDk4k
rmeis has quit [(Ping timeout: 240 seconds)]
<gchristensen> `passAsFile` is a super cool derivation attribute!
sellout- has joined #nixos
<taktoa> I'm putting together a kind of "nixonomicon" of weird shit about the nix expression language (and package manager) that usually takes a while to learn
<taktoa> and I want to have a really clear glossary
<gchristensen> great, taktoa!
rmeis_ has quit [(Ping timeout: 256 seconds)]
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<taktoa> sometimes I've had to make up some terms though. for instance, what do you call a literal like `<nixpkgs>`?
<taktoa> what do you call the "type" of strings that can go into NIX_PATH=
pmade has joined #nixos
<taktoa> (respectively, I'm calling those "environment literals" and "environment definitions"; I don't want to have NIX_PATH in the name because `-I` affects them too)
eacameron has quit [(Remote host closed the connection)]
<simpson> Indeed, '<nixpkgs>' is undocumented.
<simpson> taktoa: Have you considered contributing to the Nix manual instead?
<taktoa> I would be fine with merging this into the Nix manual down the road.
<symphorien> is there an equivalent of python.withPackages with ocaml ?
alx741 has quit [(Quit: alx741)]
alx741 has joined #nixos
alx741 has quit [(Client Quit)]
eacameron has joined #nixos
<taktoa> simpson: I want to proceed in a slightly different manner from the Nix manual for now though; the manual tends towards pedagogical explanations where I want bulletproof definitions with tons of examples and edge cases. The style is kind of like the C++ definition.
alx741 has joined #nixos
<taktoa> *C++ reference
<simpson> taktoa: 'k, best of luck.
<gchristensen> the more the merrier!
reinzelmann has quit [(Quit: Leaving)]
jmeredith has joined #nixos
<symphorien> I am new to ocaml and the compiler tells me "Error: Unbound module Core" but I don't know how to install this dependeny with nix
<taktoa> I've already learned new things about Nix while doing this; for instance, that you can use antiquotation in the name of a let binding (`let ${"foo"} = 5; in foo`)
<simpson> taktoa: Commingled syntax and semantics is not a good start IMO. "…" is confusing given that "..." is valid syntax. Nix doesn't call quasiliterals "quasiliterals"; they're just literals with antiquotations.
<taktoa> simpson: yeah the "syntax" section is really a glossary
<LnL> taktoa: nice job!
<simpson> It's not obvious that "./" must precede relative paths with only one path segment. Nothing in the manual suggests that URL literals are deprecated or discouraged.
<taktoa> the ... issue can be fixed with an easy `sed` command once I come up with a suitable replacement
<LnL> I feel like a whole bunch of people are making pretty nice resources, but we don't have a place to reference them properly
<simpson> And finally you're missing stuff like implication.
<taktoa> simpson: "this is what I have _so far_"
<gchristensen> simpson: perhaps digging on taktoa's efforts aren't a productive way of making it better
<taktoa> implication would go under "Built-in Functions and Operators"
<taktoa> since it's a simple rewrite rule
<simpson> taktoa, gchristensen: I'm not a saying-nice-things person, sorry. Also I *really* would like to see contributions to the manual instead of new manuals.
<LnL> so anything that doesn't end up in the manual doesn't get enough visibility
<gchristensen> please keep doing good work, taktoa, I'm excited to see where it goes
<LnL> I think we should at least add a references section or something
<simpson> taktoa: I'm sorry for being rude. Please continue to contribute to Nix.
<taktoa> URL literals are not discouraged in the manual, true, but there was a github discussion about it and the conclusion was that URL literals should not exist (because they add basically nothing and cause problems complicate parsing)
<LnL> you can't just dump everything in the manual, it's already hard for new people to find stuff in there
<taktoa> I think there's value in having a "manual for arcane knowledge that most users don't need to know initially"
<LnL> I really like the manual as a reference guide for stdenv.mkDerviation or how to use the language integrations
<taktoa> regarding quasiliterals, you are right that quasiquotation and quotation are not distinguished in the AST, but they are useful terms to introduce in case I need to exclude/include strings with quasiquotations in some context
eacameron has quit [(Remote host closed the connection)]
<LnL> but there's a whole group of things that don't work for a manual IMHO
Seichi_ has joined #nixos
Seichi has quit [(Ping timeout: 255 seconds)]
<ToxicFrog> I wonder if I should just switch to unstable for my laptop. Mixing unstable and release-16.09 is bringing me a lot of pain.
<taktoa> simpson: don't get me wrong, I would really like for this to be a declarative parser specification + executable semantics, but that's too time consuming (plus there are serious questions about which tools to use), so I am just trying to hammer down all these weird things in a text document. A lot of this stuff is only documented in source comments or github comments.
schneefux has quit [(Quit: ZNC 1.6.4 - http://znc.in)]
<simpson> taktoa: I grok that. I'm just not sure why holes in the manual aren't being plugged in the manual.
<gchristensen> I'm looking at it as a way to invoke Vcunat's Offe9r
eacameron has joined #nixos
<taktoa> simpson: `pandoc -f markdown_github -t docbook nixonomicon.md >> manual.xml`
<taktoa> that's my intention
<taktoa> anyway I gotta go to class
osxi has joined #nixos
johann__ has quit [(Quit: Leaving.)]
eacameron has quit [(Remote host closed the connection)]
hotfuzz_ is now known as hotfuzz
Itkovian has joined #nixos
rly has quit [(Ping timeout: 240 seconds)]
<avn> Can nix accept "normal" sha256/md5 sums for fixed? not base32 encoded/...?
<manveru> avn: no
<gchristensen> eh? yes
<manveru> then i guess my normal is different :)
<gchristensen> not the raw binary
<simpson> avn: No; Nix doesn't compute the same hash as sha256sum/md5sum.
<simpson> avn: Or do you mean using hex instead of the base36 encoding? That's fine.
<avn> boo ;(
<digitalmentat> simpson, it actually does for flat file hashes
<digitalmentat> but it prints it out with a different encoding
<simpson> digitalmentat: Really? Then we should really figure out how to get .asc signatures verified.
<manveru> is there a way to make nix cache fetchTarball by giving it a sha256?
<gchristensen> simpson: I've prototyped a few solutions, it is difficult
<gchristensen> I have a PR in now to do it for firefox-bin
<gchristensen> (on the updater)
nschoe has quit [(Quit: Program. Terminated.)]
nschoe has joined #nixos
jgertm has quit [(Ping timeout: 240 seconds)]
__Sander__ has quit [(Quit: Konversation terminated!)]
Seichi_ has quit [(Ping timeout: 240 seconds)]
Seichi has joined #nixos
teknico has left #nixos []
simendsjo has joined #nixos
hamzaboulaares has quit [(Remote host closed the connection)]
hamzaboulaares has joined #nixos
<LnL> dunno but 1.11.6 was a security update, I don't think it has new stuff compared to 1.11.5
hamzaboulaares has quit [(Client Quit)]
<manveru> well, that commit is from summer
<manveru> according to that, this should work
<manveru> nix-instantiate --eval -E 'fetchTarball { url = https://github.com/svanderburg/node2nix/archive/b6545937592e7e54a14a2df315598570480fee9f.tar.gz; sha256 = "0ys15ag40jb09il2wymgyb1316w5m8y3fz1nzjvci6w55rn2jx3n"; name = "foo"; }'
<manveru> so i don't have to download the bloody tarball every time i open the project
<gchristensen> master is for nixUnstable, ie: 1.12
<gchristensen> IIRC
<LnL> ^^
<manveru> :(
<k0001> Is there any straightforward way to get the sha256 for a fetchFromGitHub output besides just trying to run fetchFromGitHub and copying the sha256 from the error output?
<LnL> you can use it, hydra also uses nixUnstable
<LnL> but note that there has been a schema change
<gchristensen> that isn't totally correct
<gchristensen> hydra builds against nixUnstable, not sure it actually uses nixUnstable
<manveru> yeah, i can't use that without changing the nix version all over the place...
<gchristensen> as like, the system daemon
<manveru> any other way to achieve this using mkDerivation or the like?
<gchristensen> k0001: -A yourpackage.src
<LnL> gchristensen: I'm pretty sure the daemon is also 0.12
<gchristensen> ah hrm
<k0001> gchristensen: I mean... from the commandline.
<gchristensen> derp
<gchristensen> k0001: nix-prefetch-url -A yourpackage.src
<LnL> gchristensen: otherwise it wouldn't be a problem to downgrade (like a while back)
<NixOS_GitHub> [nixpkgs] FRidh pushed 2 new commits to master: https://git.io/vDkwA
<NixOS_GitHub> nixpkgs/master 43753ea Lancelot SIX: pythonPackages.matplotlib: 1.5.3 -> 2.0.0
<NixOS_GitHub> nixpkgs/master e4f739f Frederik Rietdijk: Merge pull request #22003 from lsix/update_matplotlib...
<k0001> gchristensen: Hmmm ... not exactly what I'm looking for, but it gives me some ideas. Thanks.
<gchristensen> aye
<gchristensen> k0001: how is it not what you're looking for? I don't understand
<NixOS_GitHub> [hydra] domenkozar pushed 1 new commit to master: https://git.io/vDkrU
<NixOS_GitHub> hydra/master 63d8c8d Renzo Carbonara: Revert "Sort inputs table" (#448)...
civodul has quit [(Quit: ERC (IRC client for Emacs 25.1.1))]
<k0001> gchristensen: I want to pass the owner/repo/rev arguments on the commandline, not pick them up from some derivation.
<manveru> yeah, same error on 1.12
<manveru> at least with nixos-16.09 channel
<gchristensen> ahh
<gchristensen> k0001: nix-prefetch-git in `nix-shell -p nix-prefetch-scripts`
<k0001> gchristensen: last time I tried that worked for `fetchgit`, not from `fetchFrom`GitHub`, but maybe I'm just having some bad memory issues. I'll try again, thanks.
<gchristensen> oh you're probably right :(
<manveru> yeah, nix-prefetch-github is still missing
endformationage has joined #nixos
FRidh has joined #nixos
Itkovian has quit [(Ping timeout: 245 seconds)]
<FRidh> gchristensen: how about adding https://prs.nix.gsc.io/project/nixos to the PR template?
schneefux has joined #nixos
<gchristensen> sure
<gchristensen> and when it merges with Hydra Official, then we can update
morsicus1 has joined #nixos
<gchristensen> "99,891 commits" eee
<LnL> FRidh: I'm still working on a change to improve the diff
<gchristensen> I don't think we need to wait, though, do you?
<LnL> no I can create a pr for the changes we are using now
<NixOS_GitHub> [nixpkgs] vcunat pushed 3 new commits to master: https://git.io/vDkKm
<NixOS_GitHub> nixpkgs/master c3f56e9 Marc Weber: Add nixpkgs documentation about how to create a derivation with Vim + plugins
<NixOS_GitHub> nixpkgs/master 69348ba Vladimír Čunát: doc vim: nitpicks
<NixOS_GitHub> nixpkgs/master 49b1363 Vladimír Čunát: Merge #16501: add a guide about vim plugins
<NixOS_GitHub> [nixpkgs] vcunat closed pull request #16501: Add nixpkgs documentation about how to create a derivation with Vim plugins (master...submit/nixpkgs-manual-vim) https://git.io/vo9Kr
<FRidh> a pity you can't put the PR number in the template though.
mudri has quit [(Ping timeout: 255 seconds)]
trupanka has left #nixos []
<FRidh> gchristensen: Wait with that?
<LnL> I'm going to spend an hour to figure out what's going wrong with jobseteval, otherwise I'll push the changes
<gchristensen> FRidh: LnL is working to make it easier to tell what changed (is now broken / working) but I don't think we should wait for it
<LnL> FRidh: currently the pr jobsets start out empty, so you don't have a correct references of the state of master to compare with
<gchristensen> if we can get more people wanting it to be great, maybe we'll get more people making it great :)
derjohn_mob has quit [(Ping timeout: 258 seconds)]
morsicus1 has quit [(Ping timeout: 252 seconds)]
<FRidh> ahh, I see.
nschoe has quit [(Quit: Program. Terminated.)]
<avn> Another silly question, are nixos-container able to spawn more than one instance of same container?
Seichi has quit [(Quit: Konversation terminated!)]
<gchristensen> FRidh: we could have a "bot" that posts a comment with a link
<gchristensen> if someone writes the code, I'll execute the code (and provide github credentials ) when I create the jobset
<FRidh> so you want to check against the most recent common ancestor and compare with that. Maybe, until that is in place, compare to nixpkgs-unstable? I guess that is used now?
<gchristensen> yeah that'd be good, but to be honest getting _something_ out there is better than the almost-nothing we have, so I think we should just plow forward with what we've got :D
<NixOS_GitHub> [nixpkgs] vcunat pushed 1 new commit to staging: https://git.io/vDk6c
<NixOS_GitHub> nixpkgs/staging 9cd2dbc Vladimír Čunát: Merge branch 'master' into staging...
glines has joined #nixos
<FRidh> gchristensen: I saw there is a jobset for master, could you add a jobset for nixpkgs-unstable so we can refer to that as well?
<FRidh> refer -> compare
<gchristensen> would that be the channel?
<gchristensen> what would nixpkgs-unstable be building?
<FRidh> gchristensen: https://github.com/NixOS/nixpkgs-channels so nothing in practice
<LnL> that could be another solution
kampfschlaefer has joined #nixos
<FRidh> I typically rebase on nixpkgs-unstable, unless there are breaking changes
<gchristensen> we could probably do that, yeah
<gchristensen> domenkozar: could we get another hook added?
<LnL> that would make the diff easier since it doesn't update that often
<NixOS_GitHub> [nixpkgs] nand0p opened pull request #22297: [WIP] buildbot: 0.9.0.post1 -> 0.9.3 (master...buildbot-0.9.3) https://git.io/vDkiw
<gchristensen> FRidh: why not just link to it now and we can improve
<gchristensen> "your pr will be assimilated, I am graham of borg" http://imgur.com/P9bY6fa
goibhniu has quit [(Ping timeout: 240 seconds)]
<FRidh> ha
<FRidh> yes, will do
<FRidh> also nice to have, a check status, like we have for Travis, showing the PR is build https://developer.github.com/v3/repos/statuses/
<gchristensen> yeah
<gchristensen> if someone writes the code, I'll happily hook it up to run with credentials :)
<gchristensen> not sure how to report status of "500 things failed but that is sort of normal"
<gchristensen> I guess failed :P
mudri has joined #nixos
<FRidh> well, just report whether it finishes the job or not. I guess that is sufficient for now.
c0bw3b has left #nixos []
bennofs has joined #nixos
morsicus1 has joined #nixos
morsicus1 has quit [(Client Quit)]
<taktoa> what does builtins.match do
<taktoa> ah, regex matching
mkoenig_ has joined #nixos
morsicus1 has joined #nixos
mkoenig has quit [(Ping timeout: 260 seconds)]
mkoenig_ has quit [(Remote host closed the connection)]
tokudan has joined #nixos
tokudan has quit [(Remote host closed the connection)]
<taktoa> I wonder how feasible/useful it would be to add a primop that, given two nix expressions returns a bool that represents whether or not the first nix expression (interpreted as JSON) validates against the second nix expression (also interpreted as JSON)
seagreen has joined #nixos
<taktoa> *interpreted as a JSON Schema
<taktoa> I mean it seems like the nixos module system already has something like that hacked in, so the value proposition is questionable
tokudan has joined #nixos
<gchristensen> hrm. can i log in to kde as root somehow?
<tokudan> i believe there is a setting in kdm in the system settings. been some time since I've looked there, though and I'm not using kde anymore
<taktoa> json schema actually kind of sucks usually because you can't define macros to generate types or schemas, but if you're writing Nix-that-can-be-run-through-toJSON, then that problem goes away
<gchristensen> I got this issue in my email today :D https://github.com/NixOS/nixpkgs/issues/17407
georges-duperon has joined #nixos
FRidh has quit [(Ping timeout: 258 seconds)]
mudri has quit [(Ping timeout: 264 seconds)]
SovereignBleak__ has joined #nixos
SovereignBleak__ has quit [(Client Quit)]
BluntObject has joined #nixos
srhb has quit [(Quit: Quit)]
Guest81798 has joined #nixos
mkoenig has joined #nixos
Guest81798 has left #nixos []
edvorg has joined #nixos
srhb has joined #nixos
mkoenig has quit [(Remote host closed the connection)]
<k0001> Has anyone started seeing 100% memory and CPU usage in Hydra in the last few days? I submitted an issue here but I'm not sure that's the problem: https://github.com/NixOS/nixpkgs/issues/22296 -- bottom line: hydra-evaluator is not working here :(
<LnL> k0001: are you using a custom expression?
<k0001> The issue doesn't seem to be in the Hydra setup itself, but rather in the stuff I'm trying to build with it.
<k0001> LnL: what do you mean by "custom"? I have a release.nix file building some software of my own on top of `release-16.09` from some days ago (fd1dbe551cf6338c5f4e4f80c2f5dde9f9e6a271)
xadi has quit [(Ping timeout: 276 seconds)]
<k0001> Hydra is at the latest HEAD.
simendsjo has quit [(Remote host closed the connection)]
simendsjo has joined #nixos
Ankhers has quit [(Quit: ZNC 1.6.3 - http://znc.in)]
<LnL> are you sure you didn't introduce recursion into the jobs attrset
<k0001> LnL: Ah! Maybe I did. Let me check that.
mkoenig has joined #nixos
Ankhers has joined #nixos
sellout- has quit [(Quit: Leaving.)]
<Ankhers> I'm trying to use `nix-env -i git`, It hangs for a moment and just says "Killed". What can I do to debug this?
<Ankhers> If it makes a difference, this box was setup using nixops.
<NixOS_GitHub> [nixpkgs] vcunat pushed 1 new commit to master: https://git.io/vDkxY
<NixOS_GitHub> nixpkgs/master 612333a Vladimír Čunát: knot-resolver: yet another attempt to fix build on Darwin
mkoenig has quit [(Ping timeout: 255 seconds)]
alx741 has quit [(Ping timeout: 276 seconds)]
mkoenig has joined #nixos
<Ralith> Ankhers: your system is probably running out of memory trying to search the full package set
<Ralith> run nix-env -iA nixos.git instead
<Ankhers> Ralith: Thanks. I'll give that a go.
ThatDocsLady has quit [(Ping timeout: 255 seconds)]
<Ralith> this looks up the package by path instead of name, which means it doesn't have to look at any others
shepheb has joined #nixos
<shepheb> I found the Raspberry Pi builds at https://www.cs.helsinki.fi/u/tmtynkky/nixos-arm/installer/ but they don't seem to have the right DTB for RasPi 3
<shepheb> is there a more up-to-date path, or a guide to building either those full systems or just the kernel locally, to grab those DTBs?
morsicus1 has quit [(Quit: Leaving.)]
vandenoever has joined #nixos
<NixOS_GitHub> [nixpkgs] edwtjo pushed 1 new commit to master: https://git.io/vDkhD
<NixOS_GitHub> nixpkgs/master b08524b Edward Tjörnhammar: nixos: nylon, use named instances
justanotheruser has quit [(Ping timeout: 240 seconds)]
<NixOS_GitHub> [nixpkgs] fadenb opened pull request #22298: emby: 3.0.8500 -> 3.1.5 (master...emby_3.1.5) https://git.io/vDkjC
morsicus1 has joined #nixos
reinzelmann has joined #nixos
morsicus1 has quit [(Client Quit)]
<NixOS_GitHub> [nixpkgs] pradeepchhetri opened pull request #22299: nomad: 0.4.1 -> 0.5.3 (master...nomad) https://git.io/vDkjw
jensbin has joined #nixos
ilja_kuklic has joined #nixos
jensbin has quit [(Client Quit)]
JagaJaga has quit [(Ping timeout: 255 seconds)]
rmeis has joined #nixos
<NixOS_GitHub> [nixpkgs] rasendubi closed pull request #22295: neofetch: 3.0 -> 3.0.1 (master...update/neofetch) https://git.io/vDk4k
alx741 has joined #nixos
_rht has joined #nixos
georges-duperon has quit [(Ping timeout: 252 seconds)]
x7C3 has joined #nixos
<x7C3> how come nixos doesn't use systemd-boot?
<gchristensen> we do
<gchristensen> if you enable it, I think by default we use grub
<ToxicFrog> x7C3: it does -- it uses systemd-boot by default on UEFI systems and grub on BIOS ones,.
shanemikel_ has quit [(Quit: Connection closed for inactivity)]
<x7C3> just asking because a lot of the wiki docs refer to grub
<gchristensen> hrm. the nix install script is broken for me.
<NixOS_GitHub> [nixpkgs] rasendubi closed pull request #22299: nomad: 0.4.1 -> 0.5.3 (master...nomad) https://git.io/vDkjw
<gchristensen> /nix/store/qdkzm17csr24snk247a1s0c47ikq5sl6-nix-1.11.6/bin/nix-channel: unable to check ‘https://nixos.org/channels/nixpkgs-unstable
<cheshircat> Hello, has anyone managed to get stack, nix and yesod all to play nicely with each other? Should I ditch stack and just use cabal2nix?
eacameron has joined #nixos
mudri has joined #nixos
rmeis has quit [(Ping timeout: 256 seconds)]
<k0001> cheshircat yes
Wizek_ has quit [(Ping timeout: 255 seconds)]
Wizek has quit [(Ping timeout: 240 seconds)]
eacameron has quit [(Remote host closed the connection)]
<cheshircat> OK, sounds good
<k0001> cheshircat: there's really no need for `stack`, it mostly becomes an obstacle once you are using `nix`.
eacameron has joined #nixos
<cheshircat> Hmm, OK
<cheshircat> It all works!
<cheshircat> Thank you!
<cheshircat> oh, nvm, spoke too soon...
<cheshircat> is there a way to specify which haskell compiler you want?
khajvah_ has quit [(Ping timeout: 240 seconds)]
thc202 has quit [(Ping timeout: 240 seconds)]
<gchristensen> has anyone installed Nix in an env with an HTTP proxy? nix-channel is saying "unable to check <channel>" but curl does just fine
fresheyeball has quit [(Ping timeout: 255 seconds)]
<edwtjo> gchristensen: no problems here atm, but I would say it is kind of intermittent, did you set http_proxy?
derjohn_mob has joined #nixos
<Ankhers> How can I go about installing a recent version of node? It seems to be a really old version that gets installed.
georges-duperon has joined #nixos
binaryphile has quit [(Quit: Leaving)]
<gchristensen> edwtjo: where would I do that? I have env vars http_proxy, https_proxy, HTTPS_PROXY, and HTTP_PROXY
hotfuzz has quit [(Quit: WeeChat 1.6)]
hotfuzz has joined #nixos
ilja_kuklic has quit [(Ping timeout: 240 seconds)]
Fare has quit [(Quit: Leaving)]
<edwtjo> gchristensen: I have a script which follows https://nixos.org/wiki/Proxies#Shell (omitting NIX_CURL_FLAGS).. which reminds me that I should migrate that page..
<edwtjo> gchristensen: so just setting http_proxy *should* work, I however did https://github.com/NixOS/nix/pull/1204 when I was fiddling with it
<teto> hum nixos noob here, how can I launch an X application with the correct python module (more precisely i3bar + i3pystatus, right now the i3pystatus python module can't be found) ?
<edwtjo> gchristensen: but that might just be pebkac
mudri has quit [(Ping timeout: 240 seconds)]
<gchristensen> edwtjo: hrm. it doesn't, though
<gchristensen> this isn't actually even nix though, it is perl, perl is failing to do the channel update
<NixOS_GitHub> [nixpkgs] lsix force-pushed python-wip from 9116b62 to ef5f376: https://git.io/vzaOS
<NixOS_GitHub> nixpkgs/python-wip c6c2137 Lancelot SIX: pythonPackages.hypothesis: 3.5.2 -> 3.6.0
<NixOS_GitHub> nixpkgs/python-wip ef5f376 Lancelot SIX: pythonPackages.pytest: use pytest_30
<edwtjo> so it is an older version of nix where download wasn't using the c curl lib but the perl curl lib?
<edwtjo> so just use a checked out repo
<NixOS_GitHub> [nixpkgs] goetzst opened pull request #22300: youtube-dl: 2017.01.24 -> 2017.01.29 (master...youtube-dl) https://git.io/vDILu
eacameron has quit [(Remote host closed the connection)]
<taktoa> I just read the recent(ish) fractalide thread on nix-dev and am pretty surprised by the hostility about his email; did no one in that thread realize that fractalide has done some very interesting (and potentially nixpkgs-mergable) work with nix, e.g.: nixcrates which seems to be a hackage2nix-style system for rust crates? from what I've read of nix-dev, I've seen it as similar to haskell-cafe, and that kind of post is pretty much
<taktoa> acceptable there (granted, brazenly asking for maintainers may have been a bit too intense).
<gchristensen> edwtjo: what do you mean? it is nix 1.11.6, just installed moments ago
<gchristensen> edwtjo: `nix-channel (Nix) 1.11.6` , `/home/graham/.nix-profile/bin/nix-channel: Perl script, UTF-8 Unicode text executable, with very long lines`
ntc has quit [(Ping timeout: 240 seconds)]
<Profpatsch> Lots of stuff on nixpkgs going on right now.
<Profpatsch> Is there a holiday or something?
AllanEspinosa has joined #nixos
leothrix has quit [(Quit: ZNC 1.6.4 - http://znc.in)]
x7C3 has left #nixos ["WeeChat 1.7"]
<edwtjo> gchristensen: I wasn't thinking straight.. but I don't use nix-channel
eacameron has joined #nixos
<simpson> taktoa: It appears that nixcrates was commissioned.
<gchristensen> niksnut: any information you might have? can't install nix in an env with an http(s) proxy due to erroring out at nix-channel --update
<simpson> Unfortunately, Fractalide's docs appear to have gotten *worse*; I can't understand this table: https://github.com/fractalide/fractalide#features
leothrix has joined #nixos
<taktoa> simpson: I think the idea is kind of like Nix + Rust + CloudABI-style composability
<simpson> taktoa: Ideas are great.
<taktoa> simpson: firstly, I'm not affiliated with the project. secondly, it has nearly 1000 commits, so I think it's more than just an idea :^)
<simpson> taktoa: I'm not affiliated either. I'm just confused about what it offers. I'm trying to figure out whether it's worth adopting.
<NixOS_GitHub> [nixpkgs] goetzst closed pull request #22300: youtube-dl: 2017.01.24 -> 2017.01.29 (master...youtube-dl) https://git.io/vDILu
<NixOS_GitHub> [nixpkgs] fpletz pushed 2 new commits to master: https://git.io/vDImf
<NixOS_GitHub> nixpkgs/master e07f1f2 Franz Pletz: gitlab: 8.16.1 -> 8.16.3
<NixOS_GitHub> nixpkgs/master 1a45dde Franz Pletz: yotubeDL: 2017.01.28 -> 2017.01.29
<NixOS_GitHub> [nixpkgs] fpletz pushed 1 new commit to release-16.09: https://git.io/vDIm3
<NixOS_GitHub> nixpkgs/release-16.09 140ff04 Franz Pletz: gitlab: 8.16.1 -> 8.16.3...
<taktoa> simpson: I think the idea is that instead of having unix garbage with various command line flags, config options, etc, servers would use file descriptor passing and a standardized config format. i.e.: make all your services polymorphic on a socket interface. any communication goes through capnproto, the services are written in rust, and the services are configured with nix in a composable manner (so you can wire together services
<taktoa> arbitrarily, as long as the capnproto message types fit together). I don't think the value proposition is great in the short term, but it seems like a worthy goal long-term.
<simpson> taktoa: Oh. If that's what they're doing, then I am a hard no; Capsicum is the same thing, but with an understanding of capability theory.
<simpson> Fractalide's authors clearly don't understand cap theory; they don't think that capn is "safe" and don't mention caps.
osxi has quit [(Quit: leaving)]
<simpson> Also, I frankly think that *my* capability-safe environment is better than theirs~
<NixOS_GitHub> [nixpkgs] rasendubi closed pull request #22281: wmutils-opt: init at v1.0 (master...master) https://git.io/vDTIe
<taktoa> maybe the "safe" column is just referring to the fact that capnproto is written in C (IIRC)
<taktoa> oh, C++. either way, could be construed as "not safe"
<simpson> Ah, memory-safety.
<simpson> Well, people in Rusty houses shouldn't throw stones~
<taktoa> :^) agreed
civodul has joined #nixos
<gchristensen> "A list of URL prefixes (such as http://hydra.nixos.org/log) from which nix-store -l will try to fetch build logs if they’re not available locally." whoa
<taktoa> there is some kooky stuff in their README though... comparing their programming model to TensorFlow (!) and those giant quotes from Steve Traugott.
<taktoa> (the comparison to TensorFlow is crazy because TF is zero-copy and uses automatic differentiation on neural graphs; this is like claiming that cabal://async is a good replacement for cabal://accelerate)
tokudan has quit [(Quit: Leaving)]
eacameron has quit [(Remote host closed the connection)]
eacameron has joined #nixos
goibhniu has joined #nixos
reinzelmann has quit [(Quit: Leaving)]
<gchristensen> edwtjo: nix's curl seems to be mucking up SSL: SSL routines:SSL23_GET_SERVER_HELLO:unknown protocol
mindB has left #nixos ["User left"]
vandenoever has quit [(Read error: Connection reset by peer)]
vandenoever has joined #nixos
vandenoever has quit [(Changing host)]
vandenoever has joined #nixos
simendsjo has quit [(Quit: ERC (IRC client for Emacs 24.5.2))]
mkoenig has quit [(Read error: Connection reset by peer)]
mkoenig has joined #nixos
vandenoever has quit [(Ping timeout: 240 seconds)]
acarrico has quit [(Ping timeout: 255 seconds)]
thc202 has joined #nixos
vandenoever has joined #nixos
vandenoever has quit [(Changing host)]
vandenoever has joined #nixos
vandenoever has quit [(Ping timeout: 255 seconds)]
<phorse> what's the recommended alternative to npm install -g?
<phorse> since that doesn't seem to work
hiratara has quit [(Ping timeout: 255 seconds)]
twopoint718 has joined #nixos
hiratara has joined #nixos
<twopoint718> Does anyone have experience creating derivations for Rails projects?
AllanEspinosa has quit [(Ping timeout: 256 seconds)]
AllanEspinosa has joined #nixos
<twopoint718> I'm having trouble debugging what's going wrong when I get a "gem failed to build native extension." I've used `nix-build -K` and then inspected the saved tempdir. The problem is that the only thing in the directory is `env-vars`. I was sort of expecting to see a half-built gem or something.
MichaelRaskin has joined #nixos
ebzzry has joined #nixos
<gchristensen> I need to tell nix to use `--proxy1.0 fooo:1234` / use CONNECT proxying. anyone know how to do that? the current use of setting http_proxy / https_proxy isn't doing it.
hamishmack has quit [(Quit: hamishmack)]
<NixOS_GitHub> [nixpkgs] joachifm pushed 2 new commits to master: https://git.io/vDIBJ
<NixOS_GitHub> nixpkgs/master d70c26d Tristan Helmich: emby: 3.0.8500 -> 3.1.5...
<NixOS_GitHub> nixpkgs/master 04a7c61 Joachim F: Merge pull request #22298 from mayflower/emby_3.1.5...
mkoenig has quit [(Read error: Connection reset by peer)]
<NixOS_GitHub> [nixpkgs] LnL7 opened pull request #22301: khd: 1.1.4 -> 2.0.0 (master...khd) https://git.io/vDIBX
mkoenig has joined #nixos
cpennington has quit [(Remote host closed the connection)]
acarrico has joined #nixos
ambro718 has joined #nixos
eacameron has quit [(Remote host closed the connection)]
AveryC has quit [(Quit: WeeChat 1.5)]
eacameron has joined #nixos
hiratara has quit [(Remote host closed the connection)]
eacameron has quit [(Ping timeout: 255 seconds)]
hiratara has joined #nixos
hamishmack has joined #nixos
kampfschlaefer has quit [(Ping timeout: 256 seconds)]
hamishmack has quit [(Quit: hamishmack)]
hamishmack has joined #nixos
Itkovian has joined #nixos
AllanEspinosa has quit [(Ping timeout: 248 seconds)]
sigmundv has joined #nixos
<unlmtd[m]> do you think I could change another system's configs with custom nixpkps, Id love to combine runit and nix on my system
AllanEspinosa has joined #nixos
goibhniu has quit [(Ping timeout: 245 seconds)]
hamishmack has quit [(Quit: hamishmack)]
goibhniu has joined #nixos
hamishmack has joined #nixos
<unlmtd[m]> kinda like nixos modules to define systemd services, only for runit.
georges-duperon has quit [(Ping timeout: 252 seconds)]
hamishmack has quit [(Ping timeout: 240 seconds)]
<unlmtd[m]> like services.acpid.lidEventCommands
pmade has left #nixos ["Using Circe, the loveliest of all IRC clients"]
<unlmtd[m]> I wanna hack on that but with a different init. Id just run nix on void, and give write permission to the right spot for nix users
twopoint718 has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<unlmtd[m]> or should I install runit on nixos?
<unlmtd[m]> nixos-rebuild switch hangs my system... ah!
<MichaelRaskin> 1. Look at Not-OS. 2. NixOS can convert a NixOS-generated systemd unit to a runner script. 3. Yes you can run void and nix-build generated configs and symlink the results wherever you want
<gchristensen> can? or could?
<MichaelRaskin> NixOS can create a runner script right now and I use it. NixOS just cannot do anything useful with the resulting script afterwards
<gchristensen> how do I do that?
<MichaelRaskin> Look up my config!
<gchristensen> hehe
<MichaelRaskin> Which is not a NixOS config, obviously
<MichaelRaskin> config.systemd.services.service-name.runner
sitedyno has quit [(Ping timeout: 258 seconds)]
goibhniu has quit [(Ping timeout: 245 seconds)]
civodul has quit [(Quit: ERC (IRC client for Emacs 25.1.1))]
goibhniu has joined #nixos
danharaj has quit [(Ping timeout: 240 seconds)]
<yorick_> nixos.org is giving an NXDOMAIN for me
<yorick_> I think my dns is just broken
<cransom> functional for me and also the root servers
<gchristensen> doesn't expire for another month
<yorick_> I stopped nscd, now it works
<yorick_> very peculiar
<gchristensen> edwtjo: figured out the problem :)
<gchristensen> http_proxy="http://foobar:1111" https_proxy="https://foobar:1111" and newer curl added support for https proxies, older curl ignored https and used http anyway.
sitedyno has joined #nixos
tg` has joined #nixos
tg has quit [(Ping timeout: 252 seconds)]
tg` is now known as tg
the-kenny has quit [(Ping timeout: 256 seconds)]
ntc has joined #nixos
georges-duperon has joined #nixos
goibhniu has quit [(Ping timeout: 245 seconds)]
the-kenny has joined #nixos
bennofs has quit [(Quit: WeeChat 1.6)]