<Sonarpulse> manveru: Mic92
<Sonarpulse> I thought PST
<Sonarpulse> maybe its doing pdt for some reason?
<Sonarpulse> "America/Los_Angeles"
mkoenig has quit [(Read error: Connection reset by peer)]
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
mkoenig has joined #nixos
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
seanz has quit [(Quit: Leaving.)]
vanrein has quit [(Ping timeout: 245 seconds)]
michiel_l has quit [(Quit: No Ping reply in 210 seconds.)]
<ToxicFrog> How do I deal with programs that are hard-coded to look in /usr/bin?
<ToxicFrog> ST3 wants to use /usr/bin/pkexec to sudo, but of course it's installed in /run/current-system/sw/bin/pkexec
<ToxicFrog> (is there a particular reason that /usr/bin isn't a symlink to /run/current-system/sw/bin ?)
joepie91 has quit [(Excess Flood)]
joepie91 has joined #nixos
joepie91 has quit [(Changing host)]
joepie91 has joined #nixos
michiel_l has joined #nixos
<Mic92> ToxicFrog: this helps to detect missing dependencies for example.
<Mic92> but I don't know the original intention.
<NixOS_GitHub> [nixpkgs] mdaiter opened pull request #21293: python/pillow: Disable tests on Darwin (master...fixNixOpsDarwin) https://git.io/v1b3I
markus1199 has joined #nixos
markus1189 has quit [(Ping timeout: 260 seconds)]
msd has joined #nixos
<msd> LnL: I'd attempt to find a more elegant solution for the Pillow case, but for now, it just seems logical to disable testing on Mac
<msd> LnL: in the future, we might want to look back into it
<LnL> msd: indeed, it's not worth spending hour on it just to fix the tests :p
<msd> LnL: agreed
<LnL> msd: oh I meant !stdenv.isDarwin
redmq has quit [(Ping timeout: 258 seconds)]
<msd> LnL: my b for not catching that
<msd> LnL: pushed
<ToxicFrog> In this case it's a closed-source program hard-coded to look in that path
<ToxicFrog> If I create a symlink in /usr/bin, will it get blown away next time I nixos-rebuild?
<NixOS_GitHub> [nixpkgs] LnL7 closed pull request #21293: python/pillow: Disable tests on Darwin (master...fixNixOpsDarwin) https://git.io/v1b3I
<NixOS_GitHub> [nixpkgs] LnL7 pushed 2 new commits to master: https://git.io/v1bsG
<NixOS_GitHub> nixpkgs/master 6bce0dc Matthew Daiter: python/pillow: Disable tests on Darwin
<NixOS_GitHub> nixpkgs/master 405507a Daiderd Jordan: Merge pull request #21293 from mdaiter/fixNixOpsDarwin...
erasmas has quit [(Ping timeout: 258 seconds)]
ntc has quit [(Ping timeout: 260 seconds)]
siel has joined #nixos
<MoreTea> ToxicFrog, you should patch the binary, or use runInFHSChroot or something like that.
<MoreTea> 'night night
MoreTea has quit [(Quit: Leaving)]
sigmundv has joined #nixos
siel has quit [(Ping timeout: 258 seconds)]
sigmundv has quit [(Client Quit)]
sigmundv_ has joined #nixos
<msd> LnL: https://github.com/NixOS/nixpkgs/issues/17896 <- seems to break only because a sub-package is listed as x86-64_linux only. Should we change selenium to an only Linux package?
sigmundv_ is now known as sigmundv
<NixOS_GitHub> [nixpkgs] copumpkin pushed 2 new commits to master: https://git.io/v1bGe
<NixOS_GitHub> nixpkgs/master 377cef8 Dan Peebles: apple-sdk: 10.9 -> 10.11
<NixOS_GitHub> nixpkgs/master f3b65f6 Dan Peebles: xhyve: update and fix to use our Hypervisor framework
<msd> LnL: the confusing thing is that https://chromedriver.storage.googleapis.com/index.html?path=2.26/ <- lists that there's a Mac version
jazzencat has joined #nixos
siel has joined #nixos
<msd> LnL: we should be able to fix that package
latk has quit [(Ping timeout: 265 seconds)]
<LnL> msd: I expect so, and that will most likely also fix the other package
<msd> Right
<msd> On it
<jazzencat> Oh, boy, there goes the wifi again. Now it doesn't work after a reboot...
<jazzencat> ip link shows the interface down, but ifup spits an error saying that ifup can't open /etc/network/interfaces: no such file or directory. Naturally, it's NixOS so I'd expect that... so what to I need to tell the system to make it behave itself usefully?
<msd> LnL: is there a case statement in Nix?
<copumpkin> I want to build the smallest nixos image ever
<copumpkin> what do I do?
<LnL> msd: no if else or attrs."${name}"
<msd> kk
<LnL> copumpkin: you'll probably want to mkForce environment.systemPackages since a bunch of stuff gets included there
<LnL> by default
<LnL> unless you don't want it to be *that* tiny
<copumpkin> I do
<Sonarpulse> copumpkin: clever had something
<Sonarpulse> though not normal nixos
<copumpkin> oh yeah, that's not-os
<Sonarpulse> yeah
<copumpkin> I want nixos but with no junk
<Sonarpulse> k
<Sonarpulse> I'd love to see minimal true nixos shrink
<Sonarpulse> good challenge
<copumpkin> is there a simple "configuration.nix -> image file" script/entry point?
<clever> copumpkin: yes
warbo has quit [(Ping timeout: 260 seconds)]
siel has quit [(Ping timeout: 258 seconds)]
<clever> copumpkin: this creates a qemu disk image with the whole nixos installed, then the postVM script converts it to a vbox machine
* ToxicFrog bonks head against desk
<ToxicFrog> It's hardcoded to try to pkexec /bin/cp as well
<ToxicFrog> Yeah, FHS is clearly the way to go
<jazzencat> What's the command to get a basic wpa_supplicant.conf file again? It seems to be missing from /etc
<clever> jazzencat: wpa_passphrase
<jazzencat> Thanks.
siel has joined #nixos
<jazzencat> For some reason the file vanished either because of the reboot or after a garbag-collect
<jazzencat> And /net/technology/wifi: Not supported again.
<copumpkin> clever: how do I reach make-disk-image.nix from a file outside of nixpkgs?
<copumpkin> I guess I can ask for .path on it
<copumpkin> my module takes modulesPath, but that's one level down from where I need to be
<clever> copumpkin: import <nixpkgs/nixos/lib/make-disk-image.nix>
<copumpkin> oh fair enough
<copumpkin> although I dislike angle brackets
<jazzencat> Okay, ip link is showing the interface as down and ifup wlp3s0 is giving me an error about not being able to open /etc/network/interfaces because the directory doesn't exist.
<msd> LnL: aha!
<msd> LnL: also it requires gtk2 -- something that isn't really meant for Mac
<LnL> msd: you don't need the patchelf stuff on darwin
<LnL> msd: and if you need to patch shared libs you can use install_name_tool
siel has quit [(Ping timeout: 258 seconds)]
<Mic92> jazzencat: it is actually ip link set wlp3s0 up
<Mic92> jazzencat: but I think you want wap_supplicant
redmq has joined #nixos
<msd> LnL: is there a glibc equivalent on Mac?
<jazzencat> Ooh, thanks Mic92, that ip link command did the trick.
siel has joined #nixos
<jazzencat> And BusyBox doesn't like the systemctl status connman command. Complains about less: --'R' being an invalid option.
<jazzencat> Guess busybox is getting tossed out on it's ear again.
<LnL> msd: libSystem but it's not pure, so it doesn't matter
Rotaerk has joined #nixos
thc202 has quit [(Quit: thc202)]
jazzencat has quit [(Ping timeout: 264 seconds)]
<msd> LnL: cool. I'll just make the package imports and the patchelf statements conditional if the system's Linux or not
[0x4A6F] has quit [(Ping timeout: 264 seconds)]
[0x4A6F] has joined #nixos
siel has quit [(Ping timeout: 250 seconds)]
<msd> LnL: also wrapProgram is or isn't our program, and should or should not be used on Mac?
siel has joined #nixos
ambro718 has quit [(Ping timeout: 252 seconds)]
MinceR has quit [(Ping timeout: 240 seconds)]
Shou has quit [(Ping timeout: 248 seconds)]
MinceR has joined #nixos
herzmeister has quit [(Quit: Leaving)]
herzmeister has joined #nixos
willeh has quit [(Ping timeout: 258 seconds)]
acertainkind has joined #nixos
siel has quit [(Ping timeout: 246 seconds)]
jazzencat has joined #nixos
edwtjo has quit [(Ping timeout: 244 seconds)]
paroneayea has quit [(Ping timeout: 258 seconds)]
edwtjo has joined #nixos
siel has joined #nixos
<jazzencat> Is there an easy way to install either a deb, rpm or slackware tgz on NixOS?
zraexy has joined #nixos
paroneayea has joined #nixos
AllanEspinosa has quit [(Ping timeout: 256 seconds)]
Fare has joined #nixos
siel has quit [(Ping timeout: 258 seconds)]
jazzencat has quit [(Quit: Leaving)]
pdobrogost has quit [(Quit: Connection closed for inactivity)]
mudri has quit [(Quit: WeeChat 1.6)]
AllanEspinosa has joined #nixos
siel has joined #nixos
mizu_no_oto has joined #nixos
redmq has quit [(Quit: leaving)]
siel has quit [(Read error: Connection reset by peer)]
nil has joined #nixos
<mguentner> is there a way to force a remote build or avoid local builds in general when nix.distributed = true and buildMachines is populated?
jazzencat has joined #nixos
phanimahesh has quit [(Ping timeout: 246 seconds)]
Fare has quit [(Ping timeout: 247 seconds)]
siel has joined #nixos
phanimahesh has joined #nixos
siel has quit [(Ping timeout: 246 seconds)]
Fare has joined #nixos
mbrgm has quit [(Ping timeout: 260 seconds)]
siel has joined #nixos
mbrgm has joined #nixos
siel has quit [(Ping timeout: 260 seconds)]
slack1256 has joined #nixos
systemfault has joined #nixos
modulistic has quit [(Ping timeout: 256 seconds)]
abrar has joined #nixos
aj__ has joined #nixos
siel has joined #nixos
<gchristensen> Mic92: wow no kidding ... not very easy to work with.
derjohn_mobi has quit [(Ping timeout: 252 seconds)]
jazzencat has quit [(Read error: Connection reset by peer)]
zraexy has quit [(Ping timeout: 264 seconds)]
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
zraexy has joined #nixos
<stepcut> I successfully hacked up nixpkgs to build GHC 8.0.2 on armv7 \o/. Next step is to do it right :-/
modulistic has joined #nixos
siel has quit [(Ping timeout: 245 seconds)]
<sheenobu> stepcut, are you using the armv7 cache? been wondering how to disable that...
siel has joined #nixos
<stepcut> sheenobu: I have no idea
<sheenobu> ahh. i think only a few things get grabbed from that when you specify system.whatever = 'armlv7', assuming that is what you did
mkoenig has quit [(Read error: Connection reset by peer)]
siel has quit [(Ping timeout: 258 seconds)]
NodeGuy has joined #nixos
<NodeGuy> I'm up and running on NixOS for the first time and it's a dream! Thank you for your help getting me started!
siel has joined #nixos
mkoenig has joined #nixos
herzmeister has quit [(Quit: Leaving)]
herzmeister has joined #nixos
siel has quit [(Ping timeout: 240 seconds)]
PragCyphr has joined #nixos
siel has joined #nixos
PragCyph1 has quit [(Ping timeout: 245 seconds)]
<stepcut> sheenobu: No.. I started from scratch pretty much. I install nixpkgs on raspbian and built nixos-install and then used it to bootstrap my RPi3
<sheenobu> ahhh
<sheenobu> hrm
msd has quit [()]
eacameron has joined #nixos
ttuegel has quit [(Ping timeout: 258 seconds)]
mizu_no_oto has joined #nixos
siel has quit [(Ping timeout: 250 seconds)]
siel has joined #nixos
oida has quit [(Ping timeout: 246 seconds)]
siel has quit [(Ping timeout: 258 seconds)]
SovereignBleak has quit [(Quit: Connection closed for inactivity)]
hexagoxel has quit [(Ping timeout: 258 seconds)]
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
siel has joined #nixos
jsgrant_ has joined #nixos
hexagoxel has joined #nixos
amarsman has quit [(Quit: amarsman)]
amarsman has joined #nixos
siel has quit [(Ping timeout: 258 seconds)]
mguentner has quit [(Quit: WeeChat 1.6)]
siel has joined #nixos
mguentner has joined #nixos
siel has quit [(Ping timeout: 250 seconds)]
mkoenig_ has joined #nixos
mkoenig has quit [(Ping timeout: 250 seconds)]
<NixOS_GitHub> [nixpkgs] vbgl opened pull request #21295: ocamlPackages.containers: 0.20 -> 0.22 (master...containers-0.22) https://git.io/v1buL
oida has joined #nixos
siel has joined #nixos
mkoenig has joined #nixos
mkoenig_ has quit [(Ping timeout: 268 seconds)]
siel has quit [(Ping timeout: 245 seconds)]
phreedom has joined #nixos
mkoenig has quit [(Ping timeout: 245 seconds)]
<Rotaerk> is there some way to make nix-prefetch-url available without using the nix package
siel has joined #nixos
mkoenig has joined #nixos
<sheenobu> Rotaerk, the file is a binary which links against nix libraries. Probably statically compile it?
sdothum has quit [(Quit: ZNC - 1.6.0 - http://znc.in)]
<simpson> Rotaerk: You've had a bunch of questions along these lines. What are you working on?
<sheenobu> Rotaerk, you could probably fetch a url and pipe it through sha. I'd check if their results are the same, though
<Rotaerk> trying to allow cabal:// style cabal2nix calls to work from within a nix derivation
<Rotaerk> cabal2nix seems to look for nix-prefetch-url
<Rotaerk> but if I add nix to my build depends, it tries to write somewhere it shouldn't, when I run the derivation
siel has quit [(Ping timeout: 240 seconds)]
<Rotaerk> particularly, it gives me an error saying that /nix/var/nix/db is not writable ... not sure why using the nix package would cause that though
<NixOS_GitHub> [nixpkgs] Ralith opened pull request #21296: matrix-synapse: 0.18.4 -> 0.18.5 (master...matrix-synapse) https://git.io/v1bzs
RayNbow`TU has joined #nixos
RayNbow` has quit [(Ping timeout: 246 seconds)]
<Rotaerk> actually, the solution may be to find a way to remove cabal2nix's dependency on the nix-prefetch stuff
<Rotaerk> wonder why it needs them...
<sheenobu> I assume since it generates hashes it needs to fetch URLs and generate hashes out of them.
<sheenobu> elm2nix.rb is a good, simple example in the nixpkg tree
<Rotaerk> yea probably; thinking the directory it fetches those URLs into could be customizable though, so that I can put it into $out instead
<simpson> Rotaerk: BTW I think that you might be hitting some sort of recursive Nix edge case, but I'm not sure; Nix itself is beyond me.
<sheenobu> Rotaerk, can you post a simple reproduction of this issue?
<Rotaerk> hmm, give me some time; will see if I can
siel has joined #nixos
<Rotaerk> btw, I've been testing my derivation by doing this in a nix-repl: c2n = import ./cabal2nixResult.nix {}; import (c2n "cabal://mtl")
<Rotaerk> the second import is to force it to actually run... is there a better way?
<sheenobu> interesting... c2n is a function which takes in a cabal URL
<Rotaerk> yep, and produces a derivation
<Rotaerk> or would if it actually worked
mguentner2 has joined #nixos
siel has quit [(Ping timeout: 240 seconds)]
mguentner has quit [(Ping timeout: 240 seconds)]
<Rotaerk> oh... I'm dumb
<Rotaerk> I was assuming that the dependency on nixpkgs.nix was failing somehow, but it was successful; cabal2nix actually ran, tried to use nix-prefetch-url, and that was causing the error
<Rotaerk> so yea, if I want this to work, I need to find a way to get it not to use that at all
AllanEspinosa has quit [(Ping timeout: 258 seconds)]
<sheenobu> and you can't know the hash in advance?
<Rotaerk> looks like cabal2nix *does* support a --sha256 parameter; lemme test if passing that in causes it to skip the prefetch
<Rotaerk> that would be good
<Rotaerk> \o/ it works
mkoenig has quit [(Read error: Connection reset by peer)]
mkoenig has joined #nixos
mkoenig has quit [(Read error: Connection reset by peer)]
marusich has joined #nixos
Fare has quit [(Ping timeout: 250 seconds)]
mkoenig has joined #nixos
MichaelRaskin has quit [(Quit: MichaelRaskin)]
systemfault has quit [(Quit: Bye!)]
acertainkind has quit [(Ping timeout: 244 seconds)]
<mkoenig> If I have to sets, e.g. haskellPackages1 and haskellPackages2. How can I merge them into one?
Wizek has joined #nixos
Wizek_ has joined #nixos
Wizek__ has joined #nixos
<mkoenig> Do I have to make a new derivation and give them as buildInputs?
Wizek_ has quit [(Ping timeout: 256 seconds)]
Wizek has quit [(Ping timeout: 265 seconds)]
siel has joined #nixos
filterfish__ has quit [(Ping timeout: 258 seconds)]
<Rotaerk> mkoenig, what do haskellPackages1/2 contain?
<mkoenig> ghc.override and ghcjs.override (same overrides)
<mkoenig> i want to use ghc and ghcjs in one shell
<mkoenig> nix-shell -A ghc --run "nix-shell -A ghcjs" works but it is a bit dirty
siel has quit [(Ping timeout: 260 seconds)]
Wizek_ has joined #nixos
andymandias has quit [(Ping timeout: 256 seconds)]
siel has joined #nixos
Wizek has joined #nixos
<Rotaerk> hmm trying something
Wizek__ has quit [(Ping timeout: 265 seconds)]
zagy has joined #nixos
<mkoenig> If I use mkDerivation with buildInputs as shell env, nix builds my project before entering the shell
<Rotaerk> mkoenig, I think you can't have ghcjs without ghc, but vice versa is possible
<Rotaerk> nix-shell -p haskell.compiler.ghcjs
<Rotaerk> in that environment, I have access to ghc 7.10.3 and ghcjs 0.2.0
<Rotaerk> nix-shell -p haskell.compiler.ghc801 gives me ghc 8.0.1 but no ghcjs
filterfish has joined #nixos
Wizek__ has joined #nixos
<Rotaerk> not sure how they would coexist, given that that would result in two ghcs
<mkoenig> I have tried that before. But it doesn't work for me.
<mkoenig> I use reflex-platform
filterfish has quit [(Read error: Connection reset by peer)]
siel has quit [(Ping timeout: 258 seconds)]
<mkoenig> which has a shell with both working ghc AND ghcjs
filterfish has joined #nixos
<Rotaerk> actually this gives me ghc 8.0.1 and ghcjs: nix-shell -p haskell.compiler.ghc801 haskell.compiler.ghcjs
<mkoenig> It is possible but I don't understand how it is done here
<Rotaerk> order there matters in terms of which ghc is available
siel has joined #nixos
<Rotaerk> not sure if ghcjs will like the ghc that's in the path being the wrong version for it though
Wizek has quit [(Ping timeout: 265 seconds)]
slack1256 has quit [(Remote host closed the connection)]
<mkoenig> nix-shell shell.nix -p my-ghcjs my-ghc
<mkoenig> this won't work
<mkoenig> nix-shell shell.nix -A my-ghcjs
<mkoenig> works
<Rotaerk> you could also try nix-shell -E
Wizek_ has quit [(Ping timeout: 258 seconds)]
<Rotaerk> look at the example of -E in the man page
siel has quit [(Ping timeout: 245 seconds)]
Wizek_ has joined #nixos
jazzencat has joined #nixos
Wizek__ has quit [(Ping timeout: 265 seconds)]
amarsman has quit [(Quit: amarsman)]
<jazzencat> When I select logout in my DE I get a message saying "Failed to execute login command" before it drops briefly to console and then back to SLiM. Any way to fix this? It doesn't impact usability, but for aesthetics would be nice to correct.
<mkoenig> Using nix-shell -E not very different to shell.nix. the problem with buildInputs is, that it tries to compile everything before entering the shell. And this often fails
katyucha has joined #nixos
Wizek__ has joined #nixos
<mkoenig> this is my shell.nix
<mkoenig> unfortunately: /run/current-system/sw/bin/nix-shell: a single derivation is required
wkennington has joined #nixos
jazzencat has quit [(Quit: Leaving)]
Wizek_ has quit [(Ping timeout: 265 seconds)]
Wizek has joined #nixos
<NixOS_GitHub> [nixpkgs] arch1tect0r opened pull request #21298: systemd: add slice support (master...add-slice-support) https://git.io/v1bwz
<mkoenig> jazzencat: which DE?
Wizek_ has joined #nixos
edvorg has joined #nixos
Wizek has quit [(Ping timeout: 240 seconds)]
qknight has quit [(Remote host closed the connection)]
Wizek__ has quit [(Ping timeout: 265 seconds)]
MercurialAlchemi has joined #nixos
siel has joined #nixos
Wizek has joined #nixos
amarsman has joined #nixos
siel has quit [(Ping timeout: 250 seconds)]
jasom has quit [(Ping timeout: 245 seconds)]
paperd has joined #nixos
qknight has joined #nixos
Wizek__ has joined #nixos
alx741_ has joined #nixos
jasom has joined #nixos
alx741 has quit [(Ping timeout: 248 seconds)]
Wizek_ has quit [(Ping timeout: 265 seconds)]
ebzzry has quit [(Ping timeout: 245 seconds)]
Wizek has quit [(Ping timeout: 248 seconds)]
<NixOS_GitHub> [nixpkgs] phanimahesh opened pull request #21299: python:html5lib: 0.999 -> 0.999999999 (master...fix-LWN-vuln-709146) https://git.io/v1bKT
Wizek has joined #nixos
Wizek has quit [(Ping timeout: 250 seconds)]
fritschy has joined #nixos
Wizek_ has joined #nixos
NodeGuy has quit [(Quit: Page closed)]
rardiol has quit [(Ping timeout: 240 seconds)]
acertainkind has joined #nixos
Wizek__ has quit [(Ping timeout: 252 seconds)]
spacekitteh has joined #nixos
jacob_ has joined #nixos
jacob_ is now known as Guest441
mjacob has quit [(Ping timeout: 260 seconds)]
deepwater has quit [(Remote host closed the connection)]
mjacob has joined #nixos
Wizek has joined #nixos
siel has joined #nixos
siel has quit [(Ping timeout: 250 seconds)]
reinzelmann has joined #nixos
siel has joined #nixos
Wizek has quit [(Ping timeout: 250 seconds)]
zraexy has quit [(Ping timeout: 252 seconds)]
aj__ has quit [(Ping timeout: 252 seconds)]
paperd has quit [(Quit: WeeChat 1.5)]
siel has quit [(Ping timeout: 245 seconds)]
jgertm_ has joined #nixos
filterfish has quit [(Ping timeout: 265 seconds)]
siel has joined #nixos
Wizek has joined #nixos
filterfish has joined #nixos
filterfish has quit [(Remote host closed the connection)]
filterfish has joined #nixos
oida has quit [(Read error: Connection reset by peer)]
oida has joined #nixos
siel has quit [(Ping timeout: 258 seconds)]
Wizek__ has joined #nixos
ckauhaus has joined #nixos
Wizek has quit [(Ping timeout: 265 seconds)]
Wizek_ has quit [(Ping timeout: 252 seconds)]
rly has joined #nixos
Itkovian has joined #nixos
<NixOS_GitHub> [nixpkgs] FRidh closed pull request #21296: matrix-synapse: 0.18.4 -> 0.18.5 (master...matrix-synapse) https://git.io/v1bzs
danielrf1 has quit [(Read error: Connection reset by peer)]
danielrf1 has joined #nixos
Wizek has joined #nixos
civodul has joined #nixos
nschoe has joined #nixos
Wizek_ has joined #nixos
Wizek has quit [(Ping timeout: 240 seconds)]
johann__ has joined #nixos
justbeingglad has joined #nixos
Wizek__ has quit [(Ping timeout: 252 seconds)]
xwvvvvwx has joined #nixos
eacameron has quit [(Remote host closed the connection)]
jgertm_ has quit [(Ping timeout: 252 seconds)]
ambro718 has joined #nixos
Wizek has joined #nixos
jensens has joined #nixos
aj__ has joined #nixos
justbeingglad has left #nixos []
johann__ has quit [(Quit: Leaving.)]
oleks has quit [(Quit: leaving)]
oleks has joined #nixos
wkennington has quit [(Quit: Leaving)]
hotfuzz_ is now known as hotfuzz
ambro718 has quit [(Ping timeout: 245 seconds)]
ThatDocsLady has quit [(Ping timeout: 245 seconds)]
thc202 has joined #nixos
<NixOS_GitHub> [nixpkgs] aszlig pushed 1 new commit to master: https://git.io/v1bS1
<NixOS_GitHub> nixpkgs/master ccbce6b aszlig: python/hetzner: 0.7.4 -> 0.7.5...
<viric> kernelnewbies.org stopped summarizing linux NEWS?
johann__ has joined #nixos
<viric> (at 4.79
<viric> (at 4.7)
<viric> also https://btrfs.wiki.kernel.org/index.php/Status refers only to 4.7.
<NixOS_GitHub> [nixpkgs] aszlig pushed 1 new commit to release-16.09: https://git.io/v1bSS
<NixOS_GitHub> nixpkgs/release-16.09 2a97e14 aszlig: python/hetzner: 0.7.4 -> 0.7.5...
xwvvvvwx has quit [(Ping timeout: 250 seconds)]
ThatDocsLady has joined #nixos
phorse has quit [(Ping timeout: 258 seconds)]
jgertm_ has joined #nixos
phorse has joined #nixos
akaWolf has quit [(Ping timeout: 245 seconds)]
eacameron has joined #nixos
akaWolf has joined #nixos
jophish has quit [(Remote host closed the connection)]
eacameron has quit [(Ping timeout: 260 seconds)]
jophish has joined #nixos
<NixOS_GitHub> [nixpkgs] Mic92 pushed 1 new commit to master: https://git.io/v1b72
<NixOS_GitHub> nixpkgs/master efd5508 Markov Dmitry: systemd: add slice support
mkoenig has quit [(Ping timeout: 260 seconds)]
<NixOS_GitHub> [nixpkgs] edolstra pushed 4 new commits to master: https://git.io/v1b7j
<NixOS_GitHub> nixpkgs/master 8a0843c Eelco Dolstra: qemu-kvm: Mark the version for tests...
<NixOS_GitHub> nixpkgs/master aad5d1f Eelco Dolstra: virtualisation.qemu.program: Remove...
<NixOS_GitHub> nixpkgs/master f173da3 Eelco Dolstra: Use only one build of qemu in VM tests...
nekroze has joined #nixos
<nekroze> I am having troubles compiling emby. It compiles fine on my ubuntu machine with a nix single user install (even upgraded it to a multi user install and it compiled) but my nixos machine just fails on `xbuild` due to a segmentation fault
Wizek__ has joined #nixos
mkoenig has joined #nixos
<nekroze> has anyone seen anything like this or have any suggestions
berce has joined #nixos
Samvh has joined #nixos
jensens has quit [(Ping timeout: 240 seconds)]
Wizek_ has quit [(Ping timeout: 250 seconds)]
<Samvh> hi
<Samvh> I just started using nixOS, and I like it
<Samvh> but now I think I noticed I have the same version of ghc installed twice (both 1gb), even though my installation is still very minimal
<Samvh> and I'm not sure why
Ralith_ has quit [(Ping timeout: 250 seconds)]
<nekroze> Samvh: Hi, while the version is the same, nix doesn't care too much about versions in this case. I would guess the hashes are different and they are both in your nix store consuming the space, not necisarily installed.
<Samvh> they are in /nix/store, yes
<Samvh> that's not where they are installed?
<Samvh> and yes the hashes are different
<nekroze> For this you can use `nix-store --delete /nix/blah`
<nekroze> that will remove the one you dont want anymore
<Samvh> that won't break dependencies?
<Samvh> because I ran garbage-collection, and I assumed that that would remove everything that's unneeded
<nekroze> that is where the packages are stored but with nix that is different from being in an active environment or configuration
<nekroze> its possible you have for example installed one of the GHC versions via nix-env -i in your user and have different version installed say via system.environmentPackages in /etc/nixos/*.nix
<nekroze> in which case you should nix-env -q the user installed one or rebuild without it mentioned in enviornmentPackages
<nekroze> or something similar
latk has joined #nixos
<Samvh> hm ok, I don't entirely understand the user-installed - environmentPackages distinction
<nekroze> see with nix the entire dependency tree of a package (as I understand it) contributes to its hash. so while both GHC apps are the same version they may be built with a different compiler version or use a different lib for something
<Samvh> I assumed if I nixos-rebuild, it will go to a sort of clean install, where maybe user-installed packages are not in that configuration
latk has quit [(Client Quit)]
<nekroze> Samvh: Each user, and the system itself, gets a nix profile. This is what makes the packages available. A user can access any applications installed in the system via environmentPackages but no other user can use the profile (set of apps) of any given user
<Samvh> at some point I installed stack (haskell package manager) from the root user, later I found out I could install it using environmentPackages, so I did that because it seemed cleaner
<nekroze> so you can have nothing in environmentPackages leaving a very minimal system and the users can install whatever packages they want and never affect any other user
<Samvh> aha ok, so if I install a package as root, but not in environmentPackages, it is not globally available?
<nekroze> so then just go back to root and ` nix-env -e ghc` or whatever the package is called and then a nix-collect-garbage will clean it
Itkovian_ has joined #nixos
<nekroze> its very different compared to other distros (bar guix to some degree) but after you get used to it, the flexibility is increadible.
<Samvh> I ran nix-env -e haskellPackages.stack now (that gives no output), then running nix-collect-garbage -d says 0 MB deleted
<Samvh> well yes I really like it so far
<nekroze> but as you have noticed there is a tendency to consume large amounts of space for less then obvious reasons
<Samvh> but things seem to be happening that I don't understand
<Samvh> and that, yes
Itkovian has quit [(Ping timeout: 248 seconds)]
<Samvh> so is there an easy way to list the user installed packages, and the environment packages, separately?
<nekroze> I am in the process of replacing all my servers with nixos and orchestrating it with nixops. I hate it when it doesn't work but its really cool when it does
<nekroze> nix-env -q
<LnL> Samvh: nix-store --gc --print-roots
<nekroze> shows just user installed apckages
<LnL> what does that show
<Samvh> nekroze: thanks
<LnL> wait that's not going to be usefull
<Samvh> LnL: 5 lines with some things that look like links
<Samvh> ok
<Samvh> :)
goibhniu has joined #nixos
<LnL> that's everything that's alive, but the question here is why a specific path is alive right?
<Samvh> ok, I ran nix-env -e on all the packages listed from nix-env -q, and then garbage collected
<Samvh> that seems to have removed the user installed packages
<Samvh> the tools are still available (They are in my environment) and 2 GB was freed up
<Samvh> so that seems to have worked
<Samvh> LnL: I think so - alive means not removed when I garbage collect?
jensens has joined #nixos
<LnL> yes, you don't want it to remove stuff that you are using :)
<nekroze> also it wouldn't be classified as garbage then
<Samvh> ok it seems that my issue is solved, thanks
<Samvh> by the way can I add any package to environmentPackages? i.e. also packages from a signal (I think that's the term?)
<Samvh> no not signal
phreedom has quit [(Ping timeout: 250 seconds)]
<Samvh> when you have a set of packages that's not the base set
Wizek has quit [(Ping timeout: 256 seconds)]
<Samvh> like nixpkgs packages
<sphalerite> Samvh: I also found that running nix-store --optimise saves a lot of the space that the two ghc installations use
<Samvh> never mind I think I found the answer, I have to specify a path: nixos-rebuild switch -I nixpkgs=/path/to/my/nixpkgs
<Samvh> sphalerite: I'll try that
<nekroze> you can install any packages you can see on the site nixos.org/nixos/packages.html and you can install others but you mast call the package file itself which is what nixpkgs does for you
<nekroze> sphalerite: I'll have to look into that one. Does it do some kind of merging for the profiles that use similar versioned packages?
<sphalerite> nekroze: it finds identical files across all packages and creates hard links so that each file is only actually stored in one place on disk (even though it has multiple locations on the filesystem)
johann__ has quit [(Quit: Leaving.)]
<nekroze> with nixops, is there a way to have different nix-channels for each machine in a deployment?
<sphalerite> It doesn't do anything with profiles
mkoenig has quit [(Read error: Connection reset by peer)]
<LnL> nekroze: -I works for nixops create
<LnL> nekroze: oh within a deployment, no idea
<Samvh> hm now as a test, I installed mysql as root (nix-env -i mysql), but actually it *is* available for other users as well
<nekroze> Seperate things sorry, first one I meant the nix-store --optimise. How does it save space with two similar packages?
<nekroze> Samvh: I am not sure if root is special... I never use it
<Samvh> I see
<sphalerite> nekroze: any identical files in two packages are deduplicated. As it happens, when I had two copies of ghc installed, most of the 1GB worth of library files was the same across the two, so rather than using 2GB for the same stuff twice it ended up using 1GB for the same stuff once
<sphalerite> But the two distinct ghc packages still existed
mkoenig has joined #nixos
Wizek has joined #nixos
<nekroze> sphalerite: Ah cool, didn't know about that one. Love nix
dgn has joined #nixos
Ralith_ has joined #nixos
<chris|> can I specify a different nixos channel for specific declarative nix containers?
Samvh has quit [(Quit: Page closed)]
<LnL> nekroze: chris: you might be able to override the pkgs argument of the nixos modules, not sure if there is another way
Itkovian_ has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
Ralith_ has quit [(Ping timeout: 260 seconds)]
Wizek_ has joined #nixos
Wizek has quit [(Ping timeout: 268 seconds)]
Wizek has joined #nixos
Itkovian has joined #nixos
Wizek__ has quit [(Ping timeout: 250 seconds)]
<rly> How do I properly pass "$@" to a process started via nix-shell --run 'myprocess <???>' ? # fill in <???>
Wizek_ has quit [(Ping timeout: 250 seconds)]
sitedyno has quit [(Ping timeout: 250 seconds)]
WilliButz has quit [(Ping timeout: 244 seconds)]
Wizek_ has joined #nixos
kahiru has quit [(Ping timeout: 268 seconds)]
<rly> Every example of nix-shell I see doesn't use arguments :/
<rly> I could generate a script and then call that, but that seems ridiculously complicated.
Wizek_ has quit [(Ping timeout: 260 seconds)]
mudri has joined #nixos
nh2_ has joined #nixos
<nekroze> rly: I am not sure but doing with a script is pretty easy. Just put two shebangs at the top "#! /usr/bin/env nix-shell" and #! nix-shell -i bash -p nixops
Shou has joined #nixos
jensens has quit [(Ping timeout: 246 seconds)]
sitedyno has joined #nixos
<rly> nekroze: the problem is that running the resulting script with e.g. --help is then interpreted by nix-shell, not by the program I want to run.
iyzsong has joined #nixos
<rly> nekroze: I just want to pass "$@" through.
<nekroze> rly: huh. that is odd. sorry I am not sure
michiel_l has quit [(Ping timeout: 245 seconds)]
<rly> #!/usr/bin/env nix-shell
<rly> #!nix-shell -I nixpkgs=../nixpkgs -p terraform --pure -i bash -p terraform
<rly> terraform "$@"
<rly> nekroze: so, I had expected that to work, but it doesn't.
<nekroze> rly: I would expect it to work as well...
berce has quit [(Quit: Lost terminal)]
<rly> nekroze: hmm, actually it's another issue now.
phreedom has joined #nixos
rardiol has joined #nixos
WilliButz has joined #nixos
johann__ has joined #nixos
<rly> #! /usr/bin/env nix-shell
<rly> #! nix-shell -i bash -I nixpkgs=../nixpkgs -p terraform
<rly> echo hi
<rly> OK, I had expected this to run 'echo hi' in the context of bash.
<rly> However, it says that 'echo' is an undefined variable.
Wizek_ has joined #nixos
<the-kenny> Can I read the public IPv4 address inside a nixops deployed container somehow reliably?
<rly> Never mind, I probably already know what's wrong.
<nekroze> the-kenny: I just had to wrestle with this issue so I can say yes.
olejorgenb has quit [(Ping timeout: 268 seconds)]
<nekroze> add resources to the files parameters ( as in {config, resources, ...}:) and then you can use resources.machines.testmachine.networking.privateIPv4
<nekroze> the-kenny: now there is some caveats, depending on your deployment type
<nekroze> the-kenny: none (ssh/metal) does not fille in that variable so I set it to deployment.targetHost in that circumstance. Also containers get privateIPv4 automatically but not a publicIPv4, I believe aws would have a different IP for each of the variables.
<nekroze> however if a machine has not yet been deployed then nixops does't fill in the variable so you may need to set a default or as I know have to do, just deploy everything at once for the first run.
<NixOS_GitHub> [nixpkgs] joachifm opened pull request #21301: make-initrd: create reproducible initrds (master...deterministic-initramfs-v2) https://git.io/v1bjd
<the-kenny> nekroze: Thanks, that's very helpful :)
<nekroze> the-kenny: I came here to the freenodes asking just this like 4 days ago but I had to figure it out myself. I could't subject you to that pain lol.
Wizek_ has quit [(Ping timeout: 252 seconds)]
Shou has quit [(Ping timeout: 245 seconds)]
<the-kenny> heh
<the-kenny> This should likely be added to the documentation
Wizek_ has joined #nixos
Wizek has quit [(Ping timeout: 252 seconds)]
nekroze has quit [(Quit: Lost terminal)]
olejorgenb has joined #nixos
<jophish> I'm trying to build the nix deb file from release.nix however it seems to require access to the kvm module
johann__1 has joined #nixos
<jophish> I've tried building it as root, but no change
johann__ has quit [(Ping timeout: 260 seconds)]
<jophish> lsmod shows that kvm is loaded
ThatDocsLady has quit [(Quit: Arma-geddin-outta-here!)]
Wizek__ has joined #nixos
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<jophish> ah, adding nixbld* to the kvm group did it
Itkovian has joined #nixos
Wizek_ has quit [(Ping timeout: 252 seconds)]
Wizek__ has quit [(Ping timeout: 265 seconds)]
<NixOS_GitHub> [nixpkgs] LnL7 pushed 1 new commit to master: https://git.io/v1Nv6
<NixOS_GitHub> nixpkgs/master fb2a7cd Daiderd Jordan: Revert "apple-sdk: 10.9 -> 10.11"...
<gchristensen> good morning #nixos, looks like there was some awesome work on the vulnerability roundup last night :)
<NixOS_GitHub> [nixpkgs] joachifm opened pull request #21302: libfaketime: make the build reproducible (master...deterministic-libfaketime) https://git.io/v1Nv7
<LnL> gchristensen: that's great!
<gchristensen> thank you so much, niksnut!
Fare has joined #nixos
Shou has joined #nixos
* gchristensen lets out a huge sigh of relief
gilligan_ has joined #nixos
johann__1 has quit [(Ping timeout: 268 seconds)]
warbo has joined #nixos
fadenb has quit [(Quit: WeeChat 1.6)]
fadenb has joined #nixos
<LnL> oh the hydra stuff is fixed?
mkoenig has quit [(Read error: Connection reset by peer)]
<gchristensen> looks like it, lots of failures on hydra still, but for other reasons: https://hydra.nixos.org/jobset/nixos/release-16.09
mkoenig has joined #nixos
<gchristensen> and if you go back intime there was never an eventbreaking 8,000 packages
<unlmtd[m]> test
<gchristensen> pass
<unlmtd[m]> can anyone read this
eacameron has joined #nixos
<gchristensen> yes
Wizek__ has joined #nixos
siel has joined #nixos
Shou has quit [(Ping timeout: 250 seconds)]
<ToxicFrog> mkDerivation question -- there's `buildInputs` for build-time dependencies, and `patchelf` for runtime library dependencies
eacameron has quit [(Ping timeout: 250 seconds)]
<ToxicFrog> What's the best way to specify runtime *program* dependencies?
<gchristensen> ToxicFrog: 1. if the built program references anything else in the store, it will be discovered when the build is finished and marked as a run-time dependency
<gchristensen> ToxicFrog: 2. you can also do propagatedBuildInputs
<joachifm> ToxicFrog: a common approach is to generate a wrapper that sets an appropriate PATH, if you're packaging an app that shells out or smth
<gchristensen> oh, good point, that might be their question
<ToxicFrog> gchristensen, joachifm: yeah, this is a closed-source program that shells out to python and pkexec
<joachifm> wrappers are to be avoided if possible, I think, but it usually gets the job done.
<ToxicFrog> Hmm
<ToxicFrog> actually I think what I really want here is buildFHSUserEnv with appropriate targetPkgs, isn't it
<gchristensen> now that one is definitely avoided, but you can
<ToxicFrog> since at least for cp I'm pretty sure it doesn't respect $PATH at all
<ToxicFrog> It assumes cp == /bin/cp always
FRidh has joined #nixos
siel has quit [(Ping timeout: 258 seconds)]
jophish has quit [(Remote host closed the connection)]
<ToxicFrog> But, hypothetically
<ToxicFrog> If I had a program that did respect $PATH
phanimah1sh has joined #nixos
<ToxicFrog> My question isn't "how do I get PATH set properly"
<ToxicFrog> It's "how do I ensure that installing this program also causes polkit and python3 to be installed"
<mbrock> ToxicFrog: basically, by ensuring that the derivation for this program depends on polkit and python3
<joachifm> ToxicFrog: setting PATH accomplishes that and also makes sure the app calls the versions you built against.
jensens has joined #nixos
Fare has quit [(Ping timeout: 250 seconds)]
siel has joined #nixos
<joachifm> ToxicFrog: generally, what gchristensen said answers your question, I think
<gchristensen> ToxicFrog: well it is related, because if the user doesn't have python3 in their path, you'll need to be either adding it to the path via a wrapper script, or hard-coding the path to your python3
phanimahesh has quit [(Ping timeout: 250 seconds)]
jsgrant_ has quit [(Read error: Connection reset by peer)]
<NixOS_GitHub> [nixpkgs] bjornfor pushed 1 new commit to master: https://git.io/v1NUl
<NixOS_GitHub> nixpkgs/master 6b4a41a Alexander Kjeldaas: libfaketime: make the build reproducible...
<ToxicFrog> joachifm: I'm not sure what you mean by "built against". They aren't build-time dependencies. The program only needs them at runtime. That's why I don't want to just slap them into the buildInputs.
<gchristensen> I think maybe you need an example
<NixOS_GitHub> [nixpkgs] bjornfor closed pull request #21302: libfaketime: make the build reproducible (master...deterministic-libfaketime) https://git.io/v1Nv7
<NixOS_GitHub> [nixpkgs] LnL7 pushed 1 new commit to master: https://git.io/v1NUR
<NixOS_GitHub> nixpkgs/master 49e3190 Daiderd Jordan: Revert "xhyve: update and fix to use our Hypervisor framework"...
oida has quit [(Ping timeout: 245 seconds)]
JagaJaga has joined #nixos
Wizek__ has quit [(Ping timeout: 250 seconds)]
<joachifm> ToxicFrog: as in the versions you passed to the build ... you do realize that all buildInputs does is add stuff to the build environment? whether they are included in the runtime closure depends on whether the build captures any references to them
JagaJaga has quit [(Client Quit)]
jsgrant_ has joined #nixos
ryanartecona has joined #nixos
<joachifm> ToxicFrog: so writing e.g., exec("${python}/bin/python") into your output ensures that 1) that python is in the closure; 2) and that exec calls *your* specific python and not some random python it happens to find at runtime
siel has quit [(Ping timeout: 250 seconds)]
<joachifm> same principle applies for libraries &c
<mbrock> ToxicFrog: if the derivation involves a wrapper that adds, say, the bin/ directory of python3 to PATH, then python3 will be installed as a dependency. That's kind of how dependencies work in Nix, and it's somewhat different from the typical package manager's lists of dependencies
jsgrant- has joined #nixos
jsgrant_ has quit [(Ping timeout: 248 seconds)]
ryanartecona has quit [(Ping timeout: 250 seconds)]
<ToxicFrog> Aah. Ok. That makes sense.
<mbrock> ToxicFrog: I think the Nixpkgs feature you want is probably `wrapProgram`, which you can call in the `postInstall` script to easily create a PATH-manipulating wrapper
<mbrock> here's an example of a postInstall script in one of my packages, wrapping a program called "lambda" to add the awscli and jshon tools to that program's PATH
<mbrock> postInstall =
<mbrock> let path = pkgs.lib.makeBinPath [
<mbrock> pkgs.awscli pkgs.jshon
<mbrock> ]; in ''
<mbrock> wrapProgram "$out/bin/lambda" --prefix PATH : "${path}"
<mbrock> '';
<ToxicFrog> mbrock: that will be useful for some other things, but not, I think, for this
<ToxicFrog> Because it hard-codes some paths
<ToxicFrog> E.g. it expects cp to always be /bin/cp, ignoring PATHJ
jsgrant- has quit [(Ping timeout: 252 seconds)]
<ToxicFrog> and now, to work! thanks for all the help
<mbrock> ah, I think the most straightforward way to fix that is to simply substitute the string "/bin/cp" in the affected files (hopefully it's not too many)
<mbrock> you can search for "substituteInPlace" in nixpkgs to see examples
Wizek__ has joined #nixos
<mbrock> example from one random package: `substituteInPlace src/config.c --replace /usr/bin/tar ${gnutar}/bin/tar`
jsgrant- has joined #nixos
<mbrock> that might not work with a binary blob program though -- I haven't dealt with that situation myself
siel has joined #nixos
<joachifm> if the replacement is shorter than the original it might work (so /bin/cp\0 -> cp\0000)
<joachifm> as I recall Profpatsch talked about doing something like that a while back
mkoenig has quit [(Read error: Connection reset by peer)]
siel has quit [(Ping timeout: 260 seconds)]
siel has joined #nixos
Wizek__ has quit [(Ping timeout: 258 seconds)]
<gchristensen> domenkozar: did you try those commands?
mkoenig has joined #nixos
* Profpatsch steps out of the shadows
<Profpatsch> joachifm: Somebody wants to patch a path in an .so file?
<Profpatsch> 1) Is there absolutely no way to patch that before the .so is created?
<Profpatsch> 2) Is the new path longer than the old path?
<Profpatsch> That is: with the store hash and everything?
<LnL> lol
<Profpatsch> If yes: You might be lucky and nothing crucial is behind the path (check with hexdump)
<Profpatsch> But in any case that’s very fragile and you should try creating an LD_PRELOAD script.
<gchristensen> Profpatsch: !!!
<gchristensen> amazing
<Profpatsch> aszlig told me that LD_PRELOAD is a saner alternative.
<Profpatsch> Depends on what your .so does of course.
<Profpatsch> gchristensen: It’s probably the most dirty trick I’ve ever pulled.
<Profpatsch> And it might break on the next update.
<LnL> ToxicFrog: since it's a binary it depends on the syscall if you can use PATH
siel has quit [(Ping timeout: 260 seconds)]
<dtzWill> LD_PRELOAD to intercept paths seems quite a bit more difficult than one of those FHSEnv chroot wrappers or something
<dtzWill> hahaha "There’s another string afterwards, but we just hope or the best." lmao
<Profpatsch> dtzWill: I don’t think these are possible for .so files?
<dtzWill> oh not without also putting the app in the chroot env, yes
<dtzWill> well AFAIK anyway lol
<Profpatsch> In that case it’s an ODBC “driver” which gets loaded into various driver managers.
<dtzWill> oh I see, oh god.
<Profpatsch> Correct reaction. :P
* dtzWill wants to unsee
<dtzWill> :P
<Profpatsch> Well, if you have a path /a/b/c\0anotherString\0
<Profpatsch> And your new path is /a/b/c/d/e
<Profpatsch> Then it’s /a/b/c/d/e\0therString\0
<dtzWill> hopefluly no one wants to use "anotherString" and/or doesn't mind getting "e\0" instead
<dtzWill> haha
<Profpatsch> So the application will see e\0 instead, right3
<Profpatsch> In that case it was a user message from what I saw.
<Profpatsch> So that was the trade-off.
siel has joined #nixos
<dtzWill> haha, and assuming that no compile-time length constants/limits are used (surprisingly seems to happen less than I would have thought? haha)
Wizek__ has joined #nixos
<gchristensen> dtzWill: why on earth would a compiled-in string change? no need :P
<dtzWill> the LD_PRELOAD you'd have to intercept enough calls to basically pretend /a/b/c was /a/b/c/d/e right? (from the .so's perspective)
<dtzWill> gchristensen: xD
<NixOS_GitHub> [nixpkgs] jgillich opened pull request #21303: docker: disable socket activation by default (master...patch-6) https://git.io/v1NLf
<Profpatsch> You don’t have to set LD_PRELOAD btw, you can prepend the compiled .c file to the ld path of the .so directly
<Profpatsch> dtzWill: Exactly, yes.
<dtzWill> wonder if there's a toolkit for doing that 'easily'? I guess kinda like what proot and similar do...
<Profpatsch> And hope the .so uses exactly the calls you overwrite.
<dtzWill> haha right
<dtzWill> and doesn't use syscalls directly
<dtzWill> or unset LD_PRELOAD and fork
<dtzWill> etc
<Profpatsch> So you have to become an strace ninja first.
<dtzWill> although tbh unless it's malware or a security product (ssh) it'll play nice lol
<dtzWill> ...hopefully
<Profpatsch> It’s a Microsoft binary blob.
<Profpatsch> I never expected it to play nice, so I didn’t.
<gchristensen> so both?
<dtzWill> :D
* gchristensen kids
<gchristensen> microsoft does good work :)
<Profpatsch> :DDDDDDDDDDDDDDDDDDDDDDDDDDDDDD
Fare has joined #nixos
<Profpatsch> Anyway, back to work.
<Profpatsch> I hope this helps. :)
siel has quit [(Ping timeout: 260 seconds)]
sdothum has joined #nixos
<dtzWill> Profpatsch: do you rem offhand the bit to bake LD_PRELOAD-esque feature into the .so? only other way I rem to do it is via ld.so.conf
<dtzWill> oh, just like add it as a library dependency? and maybe with like BINDNOW or w/e to force it to load immediately? (errrr?)
<dtzWill> just curious O:)
Ralith_ has joined #nixos
Wizek__ has quit [(Ping timeout: 258 seconds)]
FareTower has joined #nixos
FareTower has quit [(Read error: Connection reset by peer)]
abcrawf has quit [(Remote host closed the connection)]
<NixOS_GitHub> [nixpkgs] joachifm closed pull request #21301: make-initrd: create reproducible initrds (master...deterministic-initramfs-v2) https://git.io/v1bjd
FareTower has joined #nixos
fare__ has joined #nixos
jensens has quit [(Ping timeout: 258 seconds)]
Fare has quit [(Ping timeout: 265 seconds)]
mkoenig has quit [(Ping timeout: 268 seconds)]
abcrawf has joined #nixos
mkoenig has joined #nixos
FareTower has quit [(Ping timeout: 245 seconds)]
<NixOS_GitHub> [nixpkgs] aristidb pushed 1 new commit to staging: https://git.io/v1Nqp
<NixOS_GitHub> nixpkgs/staging bb18e10 Aristid Breitkreuz: Merge remote-tracking branch 'origin/master' into staging
<gchristensen> FRidh: gsnedders means well (w.r.t. html5lib) :)
filterfish has quit [(Ping timeout: 240 seconds)]
Wizek__ has joined #nixos
<NixOS_GitHub> [nixpkgs] aristidb closed pull request #21290: dropbox: 15.4.22 -> 16.4.29 (master...u/db) https://git.io/v1bJH
jensens has joined #nixos
<NixOS_GitHub> [nixpkgs] aristidb pushed 1 new commit to staging: https://git.io/v1NmH
<NixOS_GitHub> nixpkgs/staging ec4d735 Aristid Breitkreuz: boost: default version 1.62.0
jsgrant- has quit [(Read error: Connection reset by peer)]
jsgrant- has joined #nixos
filterfish has joined #nixos
filterfish has quit [(Remote host closed the connection)]
willeh has joined #nixos
<NixOS_GitHub> [nixpkgs] aristidb closed pull request #21255: syncthing: 0.14.13 -> 0.14.15 (master...u/syncthing-0.14.15) https://github.com/NixOS/nixpkgs/pull/21255
<gchristensen> anyone want to make an easy contribution? upgrade nagios to 4.2.4 and put "for CVE-2016-9565, CVE-2016-9566" in the commit message :)
<gilligan_> is it possible that the deployment.keys feature of nixos is somehow broken in some way? the files/dirs belong to root:root and I cannot access any of it with a user belonging to the "keys" group
<gilligan_> s/nixos/nixops
Wizek__ has quit [(Ping timeout: 250 seconds)]
alx741_ has quit [(Ping timeout: 256 seconds)]
alx741 has joined #nixos
ThatDocsLady has joined #nixos
<LnL> no idea, I didn't know you could do that with nixops
<NixOS_GitHub> [nixpkgs] ttuegel pushed 2 new commits to release-16.09: https://git.io/v1NO6
<NixOS_GitHub> nixpkgs/release-16.09 99352b0 Peter Hoeg: dropbox: 15.4.22 -> 16.4.29
<NixOS_GitHub> nixpkgs/release-16.09 97b3de9 Thomas Tuegel: Merge pull request #21291 from peterhoeg/u/db_stable...
Wizek has joined #nixos
jophish has joined #nixos
jophish is now known as Guest66912
fare__ has quit [(Ping timeout: 245 seconds)]
filterfish has joined #nixos
<gchristensen> "nix-prefetch-url -A package.src" is such a timesaver.
filterfish has quit [(Read error: Connection reset by peer)]
<LnL> gchristensen: howso?
joshie_ is now known as joshie
<gilligan_> LnL, actually everything is working just fine and this key service thingy is pretty cool ;)
<gchristensen> lnl no need to manually reconstruct the URL and nix-prefetch-url, or nix-build, copy the sha, nix-build again
<LnL> gilligan_: yeah sounds pretty useful
<Profpatsch> dtzWill: Not sure I understand the question.
<Profpatsch> But you can use patchelf to modify the RPATH of ELF binaries.
<LnL> gchristensen: well it doesn't work with other fetch* functions (or does it?)
Wizek__ has joined #nixos
<Profpatsch> So if you overwrite e.g. execve in your file and prepend it to the RPATH your execve symbol should be seen by the linker before the glib one.
<gchristensen> LnL: it does
<gchristensen> LnL: the exception is fetchpatch
<Profpatsch> And as far as I know there is a directive for „call the next symbol called xy“ to jump to the actual execve from your wrapper.
<Guest66912> gchristensen: I didn't know that nix-prefetch-url could take an -A argument!
<NixOS_GitHub> [nixpkgs] ttuegel pushed 1 new commit to master: https://git.io/v1N36
<NixOS_GitHub> nixpkgs/master b9d3898 Thomas Tuegel: qt57: 5.7.0 -> 5.7.1
<gchristensen> Guest66912: same, until eelco showed the chan a few weeks ago :)
Guest66912 is now known as jophish___
<LnL> same, but I still prefer using nix-build to avoid issues
<LnL> I fixed darwin.libunwind last week just by changing the sha
<LnL> it had the same as the linux source so nix just used that
<gchristensen> wait ... what?
<LnL> the source doesn't matter for fetchurl
<gchristensen> there was a collision?
<aristid> it sounds more like somebody changed the source but copied the wrong sha
<LnL> no, somebody just used the sha of the linux source
<gchristensen> ohh I see
eacameron has joined #nixos
<LnL> you don't run into that stuff if you nix-build with sha = "0000000000000000000000000000000000000000000000000000";
reinzelmann has quit [(Quit: Leaving)]
<aristid> LnL: yeah when i change the source i do it that way nowadays, usually
<aristid> nix-prefetch-url is a little more cumbersome
<aristid> if there's multiple sources i use 000000, 111111, 22222 and so on :D
<LnL> if would be nice if we had a tool that could do that for us given an expression
<aristid> :)
<gchristensen> what, blow away shas?
<LnL> basically yes
<aristid> or patch them up
<gchristensen> given a fairly known length and pattern "..." it should be fairly doable
<aristid> however, if you have upstream shas, please do try to use those
<LnL> at the moment we don't really have anything that can correctly verify them
eacameron has quit [(Ping timeout: 248 seconds)]
filterfish has joined #nixos
Shou has joined #nixos
<jophish___> you may be interested in this: https://github.com/expipiplus1/update-nix-fetchgit
<jophish___> I write the nix file and leave rev and hash as "", update-nix-fetchgit sets it up to HEAD
iyzsong has quit [(Quit: night.)]
<LnL> cool, it's a different usecase but it's pretty similar to what I'm talking about
<dtzWill> Profpatsch: well rpath is just /where/ things are searched, how do you get it to load your library at all? and yeah RTLD_NEXT is what you'd want for finding next symbol in search
<dtzWill> anyway just trying to get details on your comment re:replicating LD_PRELOAD by modifying the .so
<dtzWill> anyway
oleks has quit [(Quit: leaving)]
oleks has joined #nixos
jensens has quit [(Ping timeout: 260 seconds)]
slack1256 has joined #nixos
Wizek__ has quit [(Ping timeout: 258 seconds)]
<jophish___> The nix-build behaviour has changed and it no longer shows curl in action (or which cache it's using). Is the old behavior still available?
ryanartecona has joined #nixos
<gchristensen> jophish___: fetchurl hasn't changed in ages? https://github.com/NixOS/nixpkgs/commits/master/pkgs/build-support/fetchurl or are you talking about nix proper, using binary substitutoin?
ryanartecona has quit [(Ping timeout: 245 seconds)]
<jophish___> gchristensen: When I ran nix-build in the past I used to see the curl output when it was fetching inputs, and some string like 'signed by "cache name"'. Not all I seem to get is "fetching path ..."
<niksnut> jophish___: what Nix version?
<niksnut> that sounds like Nix 1.12
xwvvvvwx has joined #nixos
Wizek__ has joined #nixos
<jophish___> This is the difference
<gchristensen> niksnut: thank you so much for working on hydra and nix, I really appreciate it. I'm feeling much less anxious about it all now.
<jophish___> 1.11.2 has the old comforting behavior
<jophish___> 1.12 has new scary output :)
<jophish___> I think what I miss is any kind of progress indication
cpennington has joined #nixos
<aristid> niksnut: on a totally unrelated note, what's the main reason for switching from chroot to user namespacing?
mkoenig has quit [(Ping timeout: 245 seconds)]
mkoenig has joined #nixos
<NixOS_GitHub> [nixpkgs] joachifm opened pull request #21305: make-system-tarball: create reproducible tarballs (master...deterministic-tarball-v2) https://git.io/v1NCd
warbo has quit [(Ping timeout: 264 seconds)]
johann__ has joined #nixos
jsgrant- has quit [(Quit: Konversation terminated!)]
jsgrant- has joined #nixos
slack1256 has quit [(Ping timeout: 246 seconds)]
johann__ has quit [(Ping timeout: 246 seconds)]
johann__ has joined #nixos
<NixOS_GitHub> [nixpkgs] joachifm opened pull request #21306: Remove uses of gzip that might capture build time (master...gzip-9n) https://git.io/v1Nlz
cpennington has quit [(Ping timeout: 258 seconds)]
<NixOS_GitHub> [nixpkgs] matthiasbeyer opened pull request #21307: cataract: init at 1.1.0 and HEAD (master...add-cataract) https://git.io/v1NlM
<manveru> i suspect that they want more isolation than chroot provides
xwvvvvwx has quit [(Ping timeout: 248 seconds)]
Rotaerk has quit [(Quit: Leaving)]
jophish___ has quit [(Quit: leaving)]
ryantrinkle has joined #nixos
cpennington has joined #nixos
<NixOS_GitHub> [nixpkgs] edolstra pushed 4 new commits to master: https://git.io/v1N4R
<NixOS_GitHub> nixpkgs/master 6758d15 Joachim Fasting: multipath-tools: ensure gzip does not capture timestamp...
<NixOS_GitHub> nixpkgs/master 8246a84 Joachim Fasting: zerotierone: gzip -9 -> gzip -9n...
<NixOS_GitHub> nixpkgs/master a34eae0 Joachim Fasting: stun: gzip -9 -> gzip -9n...
<gchristensen> domenkozar: can yourestart https://hydra.nixos.org/build/44839000? I'm thinking it may be resolved at this point
<domenkozar> sure
mkoenig has quit [(Remote host closed the connection)]
<gchristensen> thank you :) *crosses fingers*
Wizek_ has joined #nixos
<gchristensen> joachifm: nice patches!
Shou has quit [(Ping timeout: 250 seconds)]
<LnL> yeah, it would be very nice if more stuff works with --check
Wizek has quit [(Ping timeout: 250 seconds)]
xwvvvvwx has joined #nixos
<gchristensen> domenkozar: hrm ... disappointing.
Wizek__ has quit [(Ping timeout: 240 seconds)]
<gchristensen> niksnut: I thought you reverted the changes with nix, but https://hydra.nixos.org/build/44839000 (go_1_7.x86_64) failed due to expecting an EPERM. did I misunderstand?
eacameron has joined #nixos
eacameron has quit [(Remote host closed the connection)]
Wizek has joined #nixos
eacameron has joined #nixos
Wizek_ has quit [(Ping timeout: 250 seconds)]
edvorg has quit [(Remote host closed the connection)]
Wizek_ has joined #nixos
edvorg has joined #nixos
<joachifm> gchristensen: it's the oldest PR still open, I think :)
Wizek has quit [(Ping timeout: 252 seconds)]
<niksnut> gchristensen: I reverted use of seccomp, but it still uses user namespaces
<gchristensen> oh I see
<niksnut> it just doesn't run as root inside the namespace
<niksnut> don't know why that confuses go though
<gchristensen> niksnut: user namespaces seem to allow this to happen: mkdir foo; chmod 0000 foo; touch foo/bar
<gchristensen> which is what broke pytest as well, for example
Wizek__ has joined #nixos
<niksnut> gchristensen: when running as root I assume
<gchristensen> I think even as a user
oida has joined #nixos
<niksnut> possibly relevant: https://github.com/golang/go/issues/10703
jophish has joined #nixos
<unlmtd[m]> where to configure git credentials
<aristid> that feeling when you added a dependency to the function arguments but forgot to add it to the buildInputs
<niksnut> gchristensen: I get permission denied
<aristid> niksnut: is there a way to let nix warn on unused parameters?
<gchristensen> niksnut: interesting, I see
eacamero_ has joined #nixos
<niksnut> the SCM thing suggests some interaction with sending credentials over unix domain sockets
* aristid proceeds with updating cmake to 3.7.1 for aesthetic reasons
eacamero_ has quit [(Remote host closed the connection)]
<niksnut> aristid: not currently
<aristid> i assume updating cmake in staging is fine?
<niksnut> sure
eacameron has quit [(Read error: Connection reset by peer)]
JagaJaga has joined #nixos
eacameron has joined #nixos
<aristid> niksnut: if it doesn't run as uid 0 anymore, can we re-enable the parted tests?
eacamero_ has joined #nixos
<gchristensen> niksnut: that is definitely relevant looking, though our error message is fairly different... also -- 1.7 should have it, being only a few months old.
xwvvvvwx has quit [(Ping timeout: 250 seconds)]
Wizek__ has quit [(Ping timeout: 256 seconds)]
<aristid> doing too much c++ must have brainwashed me, but i think cmake is great
erasmas has joined #nixos
<Biappi> great... maybe not. the best we have... maybe
<Biappi> =)
eacamero_ has quit [(Remote host closed the connection)]
<aristid> Biappi: that is possible... my main gripes are that it silently ignores typos in target_link_libraries, and that the variable scoping and syntax are weird
eacamero_ has joined #nixos
<Biappi> yeah
eacameron has quit [(Ping timeout: 252 seconds)]
<niksnut> aristid: sure, though arguably builds should work even under uid == 0
<aristid> niksnut: the test skips device-mapper tests if not run as root
<aristid> niksnut: but we have never properly set things up for device-mapper tests, i think, so they fail
<NixOS_GitHub> [nixpkgs] aristidb pushed 1 new commit to staging: https://git.io/v1NuC
<NixOS_GitHub> nixpkgs/staging 76c4b5f Aristid Breitkreuz: cmake: 3.6.2 -> 3.7.1
oida has quit [(Ping timeout: 245 seconds)]
eacameron has joined #nixos
eacameron has quit [(Remote host closed the connection)]
eacamer__ has joined #nixos
eacamero_ has quit [(Ping timeout: 252 seconds)]
zagy has quit [(Read error: Connection reset by peer)]
zagy has joined #nixos
ryanartecona has joined #nixos
<gchristensen> niksnut: I wish there were a way to have release-16.09 be built with a slightly-more-stable version of nix than unstable, but I'm not imagining that would be very easy.
Wizek__ has joined #nixos
eacamer__ has quit [(Remote host closed the connection)]
eacameron has joined #nixos
eacamero_ has joined #nixos
zagy has quit [(Quit: Leaving.)]
eacameron has quit [(Ping timeout: 245 seconds)]
Wizek__ has quit [(Ping timeout: 264 seconds)]
johann__1 has joined #nixos
<pierron> cstrahan: What is your opinion on https://github.com/NixOS/nixpkgs/pull/21243 ; Should I continue it and add documentation to the PR?
johann__ has quit [(Ping timeout: 268 seconds)]
<pierron> domenkozar: Do you need any help for the security-update branch on Hydra?
siel has joined #nixos
jensens has joined #nixos
oida has joined #nixos
Wizek__ has joined #nixos
danharaj has joined #nixos
<NixOS_GitHub> [nixpkgs] aristidb pushed 1 new commit to staging: https://git.io/v1NVY
<NixOS_GitHub> nixpkgs/staging 8a5fb68 Aristid Breitkreuz: mariadb: use latest boost
amarsman has quit [(Ping timeout: 264 seconds)]
amarsman has joined #nixos
jazzencat has joined #nixos
aj__ has quit [(Ping timeout: 260 seconds)]
siel has quit [(Ping timeout: 260 seconds)]
ThatDocsLady has quit [(Quit: Arma-geddin-outta-here!)]
siel has joined #nixos
alx741 has quit [(Ping timeout: 250 seconds)]
Wizek has joined #nixos
alx741 has joined #nixos
Wizek_ has quit [(Ping timeout: 250 seconds)]
<jazzencat> I have a binary of Telegram Messenger (a tar.xz with binary in them) but when I try to run it with right-click it says Enlightenment couldn't run the application.
<jazzencat> Is there something I need to do on NixOS to run a binary?
<copumpkin> Dezgeg: what did you mean by https://github.com/NixOS/nixpkgs/issues/21294?
aj__ has joined #nixos
<aristid> jazzencat: yes, binaries typically need to be patched before they can be run with nixos
<aristid> jazzencat: telegram client is open source, isn't it? i think it'd be better to package the telegram client
<jazzencat> This one is meant so that you can just run it from the home directory.
<jazzencat> All I can find from their site is the tar with the two binaries in it (app and updater)
<jazzencat> I can run it on other distros with no special actions.
<aristid> most distros follow the filesystem hierarchy standard (FHS), at least vaguely
<aristid> nixos does not at all.
<aristid> which is why most external binaries and sources need patching
<aristid> jazzencat: also, it seems like somebody already packaged telegram for nixos. why not use that?
<jazzencat> Oh, I didn't know that.
<jazzencat> Which channel is it in? Unstable?
<jazzencat> Or do I need to add another channel..
siel has quit [(Ping timeout: 258 seconds)]
Wizek__ has quit [(Ping timeout: 258 seconds)]
<jazzencat> Other question: could nixos be made to package it's stuff a more like AppImage? That way programs could be moved around.
<gchristensen> jazzencat: appimage is a poor imitation of what nixos does, nixos programs are already portable
<simpson> jazzencat: If the program is sufficiently terrible, then there's an FHS environment tool which can put the program into an FHS cage.
Wizek_ has joined #nixos
<jophish> quit
jophish has quit [(Quit: leaving)]
<jazzencat> Oh, nice. So I can copy my nixos binary of vlc on a usb drive and use it on a different computer.
jophish has joined #nixos
<aristid> jazzencat: it should be in unstable, i think it's called cutegram or so
siel has joined #nixos
<jazzencat> I'm not finding telegram with nix-env -qa in 16.09 or unstable. Do I need to add another channel?
<jazzencat> Oh, that's why.
<jazzencat> We are talking about the program you get from www.telegram.org, right?
Wizek has quit [(Ping timeout: 252 seconds)]
<gchristensen> jazzencat: cutegram is from telegram.org
<gchristensen> there is also telegram-cli and tdesktop
alx741 has quit [(Quit: alx741)]
<jazzencat> Ooh, thanks
alx741 has joined #nixos
george__ has joined #nixos
<george__> hello
<george__> how to list all installed packages in user environment?
<jazzencat> Downloading all of them and test them out. I have broken my NixOS a few times and manage to fix it by rebooting to an earlier generation if necessary... or else just run a rollback from console.
<jazzencat> Thanks, aristid and gchristensen ^,^
<george__> ah, it's nix-env -e
jensens has quit [(Ping timeout: 256 seconds)]
<aristid> george__: no, nix-env -q
siel has quit [(Ping timeout: 246 seconds)]
<NixOS_GitHub> [nixpkgs] lsix opened pull request #21310: gnupg21: 2.1.16 -> 2.1.17 (master...update_libgcrypt) https://git.io/v1Nim
Wizek__ has joined #nixos
siel has joined #nixos
<jazzencat> Well, telegram-cli is a bit of a pain to figure out.
<jazzencat> Not really sure why I need it, since I have my phone with me if I don't have access to my computer's DE
gilligan_ has quit [(Ping timeout: 250 seconds)]
Wizek__ has quit [(Ping timeout: 256 seconds)]
ttuegel has joined #nixos
oida has quit [(Ping timeout: 240 seconds)]
oida has joined #nixos
siel has quit [(Ping timeout: 258 seconds)]
Wizek__ has joined #nixos
siel has joined #nixos
Wizek_ has quit [(Ping timeout: 250 seconds)]
Wizek has joined #nixos
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
<george__> hm, trying to update nixos and it tells me " undefined variable ‘nodejs-0_10’"
<george__> was nodejs 0.10 removed?
jazzencat has quit [(Quit: WeeChat 1.6)]
jgertm_ has quit [(Ping timeout: 258 seconds)]
siel has quit [(Ping timeout: 258 seconds)]
jsgrant- has quit [(Remote host closed the connection)]
jazzencat has joined #nixos
warbo has joined #nixos
civodul has quit [(Quit: ERC (IRC client for Emacs 25.1.1))]
zagy has joined #nixos
siel has joined #nixos
<LnL> george__: I remember something about rhat
siel has quit [(Changing host)]
siel has joined #nixos
Wizek_ has joined #nixos
ckauhaus has quit [(Quit: Leaving.)]
oida has quit [(Ping timeout: 256 seconds)]
Wizek has quit [(Ping timeout: 250 seconds)]
Wizek__ has quit [(Ping timeout: 250 seconds)]
siel has quit [(Ping timeout: 258 seconds)]
xwvvvvwx has joined #nixos
oida has joined #nixos
<manveru> 0.10 is ancient
johann__1 has quit [(Quit: Leaving.)]
Wizek has joined #nixos
ryanartecona has quit [(Quit: ryanartecona)]
<manveru> well, almost 4 years old anyway, but seems like they still do security fixes on it
goibhniu has quit [(Ping timeout: 252 seconds)]
ebzzry has joined #nixos
Wizek__ has joined #nixos
rly has quit [(Ping timeout: 258 seconds)]
Wizek_ has quit [(Ping timeout: 250 seconds)]
<eacamero_> Is it possible to use the nixops options stuff for my own non-system projects?
eacamero_ is now known as eacameron
ryanartecona has joined #nixos
<jophish> Is it possible to get some sort of progress indication on input fetching. Nix used to do this in 1.11.2 however this has changed in 1.12 see: https://gist.github.com/expipiplus1/211977214458bbb8ff83cd81d9bb8a87
<gchristensen> jophish: probably best to open a ticket on nixos/nix
<jophish> gchristensen: will do, thanks
<jophish> interestingly this was opened earlier https://github.com/NixOS/nix/issues/1159
<jophish> which seems to suggest that the curl output is available somewhere
ebzzry has quit [(Ping timeout: 264 seconds)]
siel has joined #nixos
edvorg has quit [(Ping timeout: 258 seconds)]
kampfschlaefer has joined #nixos
zagy has quit [(Quit: Leaving.)]
<NixOS_GitHub> [nixpkgs] makefu opened pull request #21311: services.logstash: default options, examples and address update (master...services/logstash) https://git.io/v1NQI
pie_ has joined #nixos
siel has quit [(Ping timeout: 246 seconds)]
civodul has joined #nixos
<dtzWill> Ralith_, gchristensen: Dell's new XPS 15: http://www.dell.com/en-us/shop/productdetails/xps-15-9560-laptop
<dtzWill> we'll see how pricing/reviews/availability works out but looks pretty great to me! :)
oida has quit [(Read error: Connection reset by peer)]
<jophish> dtzWill: it runs Linux pretty well
siel has joined #nixos
oida has joined #nixos
xwvvvvwx has quit [(Ping timeout: 252 seconds)]
jgeerds has joined #nixos
<dtzWill> awesome! :D I know it's silly but excited that compared to last year's it supports more RAM (max 8GB, really?) and optional discrete graphics card (I know... but I don't play games all that much and don't want a separate machine for when I do...)
<dtzWill> actually kinda surprised they can fit a 1050 in the same package re:power/heat--I know the 10 series is supposed to be much better in that regard but @.@
Wizek has quit [(Ping timeout: 250 seconds)]
mguentner2 is now known as mguentner
<NixOS_GitHub> [nixpkgs] NeQuissimus opened pull request #21312: minecraft-server: 1.11 -> 1.11.1 (master...mc-server_1_11_1) https://git.io/v1N7D
<gchristensen> dtzWill: holy crap, is this brand new? I had one of these _ordered_ a week ago or so, and cancelled the order.
<NixOS_GitHub> [nixpkgs] NeQuissimus closed pull request #21312: minecraft-server: 1.11 -> 1.11.1 (master...mc-server_1_11_1) https://git.io/v1N7D
<jophish> oh, is that a new one?
<dtzWill> gchristensen: AFAIK absolutely brand new, yes, don't think it's actually sold anywhere yet (??). I got the link from HN O:)
<gchristensen> dtzWill: so glad I got cold feet :P
<dtzWill> and yeah I've been eye'ing all the XPS lines for a while now haha
<dtzWill> gchristensen: me too, work is buying me a laptop and this looks like the ticket ^.^
<jophish> In that case I don't know how well it runs linux :)
<gchristensen> jophish: since XPS is designed to run linux, probably pretty well?
<dtzWill> haha indeed, honestly only real thing I want to /know/ works with linux and don't have patience to muck about with is the wireless card
<jophish> We have the last generation "pretty xps 15", and that has 16GB
<dtzWill> spotty/crash-prone/intermittent wireless ruins everything lol, I can deal with most other things xD
<dtzWill> jophish: oh it supported 16GB? oops I missed that :3 good to know
<jophish> dtzWill: I think it might have even gone to 32gb with the 1tb drive
<jophish> can't remember now though
<dtzWill> np, ty for info though :)
<jophish> There are probably some really good machines for Linux now, why compromise? :)
<gchristensen> best one I've ever had is the macbookpro 11,3 :)
zagy has joined #nixos
siel has quit [(Ping timeout: 260 seconds)]
<jophish> I shudder thinking about how many hours of my life I've poured into fixing stupid things like wireless drivers and graphics cards
<gchristensen> (the reason the macbook pro series is so good is there are so few options, and people like running them, so they get what I perceive to be more hours of attention at the kernel level)
<jophish> gchristensen: really? Last time I used linux on a mac it was terrible (8 years ago perhaps)
<jazzencat> Is the format for defining an sddm theme the same as the theme option for slim?
<gchristensen> jophish: https://lwn.net/Articles/707616/
ambro718 has joined #nixos
<jophish> gchristensen: good to see that it's better supported now :)_
<jophish> as long as OSX can run Nix I'm happy there
<gchristensen> :) it is just very disappointing that they mucked up their new line
<dtzWill> jophish: haha but somehow I can tolerate graphics things more readily--a)closer to things I understand and b)"worst case" if I can just get it to blit text I still have an otherwise solid machine for dev work lol. Broken networking stack is either very limiting or just /constantly/ annoying (constantly losing connection == sadface)
<jazzencat> Yeah, OS X with Nix and Homebrew is really neat. :)
jgertm_ has joined #nixos
<jophish> jazzencat: what do you use homebrew for?
<dtzWill> haha I read something somewhere that pointed out how a /slow/ and/or /unreliable/ internet connection was somehow *so much worse* than simply no internet, which I found amusingly accurate
<dtzWill> no internet? -> decide to do non-internet activities, all is well
<dtzWill> xD
<dtzWill> ooo linux on the mac article is from nov 2016-- neat!
<jazzencat> I had it installed before I discovered Nix, and I have used it to install some applications when there isn't a nix package for it.
eacameron has quit [(Remote host closed the connection)]
siel has joined #nixos
<gchristensen> jophish: also, homebrew does a great job with Applications.app, and nix does not
joepie91 has quit [(Quit: Konversation terminated!)]
<LnL> heh I still can't run linux on mine :p
joepie91 has joined #nixos
joepie91 has quit [(Changing host)]
joepie91 has joined #nixos
eacameron has joined #nixos
<NixOS_GitHub> [nixpkgs] aszlig pushed 1 new commit to master: https://git.io/v1NdO
<NixOS_GitHub> nixpkgs/master 5ed1aee aszlig: python/pypillowfight: Disable tests entirely...
<jazzencat> And like nix puts everything in one location, so aside from the com.whatever in /Users/user/Library and /Library
eacamero_ has joined #nixos
Wizek has joined #nixos
<dtzWill> oh it appears the xp15 /does/ have a line (current gen, 9550) with both (up to) 32GB ram /and/ a dedicated gpu (960M), neat!
eacamero_ has quit [(Remote host closed the connection)]
eacameron has quit [(Read error: Connection reset by peer)]
<gchristensen> dtzWill: do they have one with 32gb of ram and no dedicated GPU?
eacameron has joined #nixos
<dtzWill> gchristensen: doesn't look like it?
<gchristensen> damn :(
<dtzWill> gchristensen: but other than price for something you don't want/need, with switching and such you can probably just not use the dedicated graphics
joepie91 has quit [(Client Quit)]
<jophish> cool, thanks
<dtzWill> they all come with the intel graphics which are actually pretty solid these days
nschoe has quit [(Quit: Program. Terminated.)]
<gchristensen> yeah
eacameron has quit [(Remote host closed the connection)]
<gchristensen> I just don't like dealing with switching
<dtzWill> hell my current laptop mostly fits my gaming needs and it's got intel hd 4000, my beloved lenovo x230 hehe
<dtzWill> yeah, you can probably just force it to one and leave it that way lol
eacameron has joined #nixos
<dtzWill> i've heard switching can be a bit difficult but haven't had the hardware to check it out personally yet
<gchristensen> maybe I'm just jaded after how impossible it is to gpu-switch on the mbp
joepie91___ has joined #nixos
joepie91___ has quit [(Changing host)]
joepie91___ has joined #nixos
<jazzencat> "The option value 'services.xserver.displayMangaer.sddm.theme' in '/etc/nixos/configuration.nix' is not a string." What am I missing?
<jazzencat> services.xserver.displayManager.sddm.theme = pkgs.fetchurl { url =
<jazzencat> sha256 = "017kahj8y1f7nh9jgnb91qg996yhra7nc6369zm92m2bpcwbvgaq"; };
<dtzWill> AFAICT if you're willing to reboot between switches (lol) you can usually make that work, which is good enough for my needs at least as a baseline.
<jazzencat> That's what I have in the configuration.nix file, but I am obviously missing something or the rebuild wouldn't be throwing me an error.
<gchristensen> dtzWill: on the xps?
amarsman has quit [(Quit: amarsman)]
joepie91___ is now known as joepie91
eacameron has quit [(Remote host closed the connection)]
<dtzWill> gchristensen: that comment is regarding an inspiron 7000 my coworker picked up a few months ago, but I'd expect the same of the XPS
* dtzWill reads the "gpu switching" section of the linked LWN article to see what the status of things on MBP's are
<gchristensen> ah
<gchristensen> dtzWill: "disaster"
eacameron has joined #nixos
<dtzWill> gchristensen: phew thanks for the tl;dr :D
<gchristensen> you have to trick the system in to thinking you're booting to OSX to access it
<gchristensen> doing that and loading the driver triggers an interrupt storm
<dtzWill> first sentence there suggests that might be specific to MBP's hardware, since they are apparently "the only remaining vendor to build a multiplexer into hybrid graphics laptops"
<dtzWill> oh MAN that's
<dtzWill> kinda excellent in a very terrible way
<gchristensen> yeah
<dtzWill> I remember some wireless card I had only worked if I booted windows and soft-rebooted into linux lol
Wizek_ has joined #nixos
proteusguy has joined #nixos
<dtzWill> which similarly simultaneously made me giggle but also cry
<jophish> dtzWill: I've had that too!
<gchristensen> classic...!
<jazzencat> Grr...
siel has quit [(Ping timeout: 240 seconds)]
dgn has quit [(Ping timeout: 265 seconds)]
<jazzencat> I am trying to define an sddm theme, but it's not a string, and I can't figure out what I'm missing based on the searchable options list on the NixOS website.
Wizek__ has quit [(Ping timeout: 252 seconds)]
Shou has joined #nixos
<jazzencat> The sddm.theme option doesn't have any example like slim.theme option does.
mudri has quit [(Ping timeout: 265 seconds)]
<jazzencat> Well, guess no theme for sddm for now until I can figure out what the hell the damn thing's problem is.
<dtzWill> jazzencat: that's how slim converts its "theme" argument into a "string" (directory of unpacked theme)
oida has quit [(Ping timeout: 252 seconds)]
<dtzWill> so if you don't mind the kludge you can probably copy that and replace 'cfg.theme' with your theme (fetchurl {...}) and away you go
<dtzWill> but sounds like maybe sddm.nix should have similar functionality or something too, and an example in the manual O:)
<jazzencat> What's the problem with sddm.theme then?
Wizek has quit [(Ping timeout: 258 seconds)]
<jazzencat> The example in slim shows default value as '' while sddm it's ""
<dtzWill> jazzencat: oh, sddm.nix doesn't unpack it first (like slim.nix does ta the linked line)
<dtzWill> *at
siel has joined #nixos
<jazzencat> So do I need to paste in the nix-prefetch-url into sddm then?
eacamero_ has joined #nixos
eacameron has quit [(Read error: Connection reset by peer)]
zraexy has joined #nixos
<dtzWill> jazzencat: err so I've never used SDDM so *lots of salt* but here's basically what I had in mind: http://dtz.so/tofoq
jazzencat has quit [(Quit: WeeChat 1.6)]
<dtzWill> unfortunately looking into things
jazzencat has joined #nixos
<dtzWill> looks like SDDM's theme parameter is meant to only select an existing theme (the value is used to set "Current=${cfg.theme}")
<dtzWill> the snippet I pasted well make "theme" be a directory containing your fetched and extracted theme but I'm not sure that works that way
<dtzWill> sorry O:)
siel has quit [(Ping timeout: 246 seconds)]
<jazzencat> Can I put that in configuration.nix?
<dtzWill> OH 'themes' is a parameter! List of packages containing themes ^.^
<dtzWill> that's how you do it xD
evangeline has quit [(Ping timeout: 256 seconds)]
<dtzWill> and 'theme' is a string of the name of the theme to you
<simpson> niksnut, goodwill: Hi! A friend is putting together a thing at SCALE next year: http://scale.opensourceinfra.org/ He's looking for somebody with Nix infrastructure experience (Nix core, nixpkgs Hydra) to talk about it. Any suggestions? Should this go to the mailing list?
<dtzWill> jazzencat: my snippet was intended for configuration.nix, yes, but it won't work--sddm wants you to add your themes as packages and select one with 'theme'
Wizek has joined #nixos
<jazzencat> I extracted the file to my ~ and added the "path/to/theme" to configuration.nix and it works.
<jazzencat> It's a bit unweildy, but it will work.
<jazzencat> I could move it and alter the path if needed.
<jazzencat> That needs to be fixed as well.
<jazzencat> Busybox is still screwing with nix-env -qa and such.
xwvvvvwx has joined #nixos
oida has joined #nixos
<dtzWill> jazzencat: wait does it actually use the theme? If so than the snippet I posted should work! lol
<NixOS_GitHub> [nixpkgs] abbradar pushed 3 new commits to staging: https://git.io/v1NNh
<NixOS_GitHub> nixpkgs/staging a60500c Nikolay Amiantov: cyrus_sasl: 2.1.26 -> 2.5.10
<NixOS_GitHub> nixpkgs/staging f84344e Nikolay Amiantov: libkrb5: 1.14.3 -> 1.15
<NixOS_GitHub> nixpkgs/staging b0a1028 Nikolay Amiantov: samba4: 4.4.6 -> 4.5.3
<dtzWill> but I'm not in a position to test and wasn't to be conservative
<dtzWill> the snippet I post just unpacks it as a derivation and uses that path
<jazzencat> I haven't checked yet.
<phorse> what causes collisions and what's the best way to deal with them?
<jazzencat> One sec.
jazzencat has quit [(Quit: WeeChat 1.6)]
<dtzWill> jazzencat: np, GL sir
<dtzWill> eep
<phorse> I get them every once in a while, and I usually remove one of the conflicting packages
<phorse> but sometimes I'd like both
<phorse> for instance, I'm experimenting with emacs x window manager
<phorse> and that causes collisions with emacs
<phorse> is a collision even something I should avoid?
jazzencat has joined #nixos
<jazzencat> No, it doesn't work. Rolling back again.
siel has joined #nixos
ryanartecona has quit [(Quit: ryanartecona)]
<jazzencat> Meh, doesn't matter, SLiM works.
siel has quit [(Ping timeout: 240 seconds)]
mudri has joined #nixos
eacamero_ has quit [(Read error: Connection reset by peer)]
eacameron has joined #nixos
Wizek has quit [(Ping timeout: 260 seconds)]
xwvvvvwx has quit [(Ping timeout: 248 seconds)]
eacameron has quit [(Read error: Connection reset by peer)]
eacameron has joined #nixos
eacameron has quit [(Remote host closed the connection)]
eacamero_ has joined #nixos
eacamero_ has quit [(Remote host closed the connection)]
eacameron has joined #nixos
jgertm_ has quit [(Ping timeout: 268 seconds)]
<rardiol> is python-packages.nix supposed to be sorted?
siel has joined #nixos
<gchristensen> rardiol: "yes"
<jazzencat> Okay, when I log out of Enlightenment, SLiM gives me a "failed to execute login command"
Wizek has joined #nixos
amarsman has joined #nixos
eacameron has quit [(Read error: Connection reset by peer)]
siel has quit [(Ping timeout: 240 seconds)]
eacameron has joined #nixos
eacameron has quit [(Read error: Connection reset by peer)]
eacameron has joined #nixos
<rardiol> gchristensen: that didn't sound very confident :)
reinzelmann has joined #nixos
<gchristensen> rardiol: well yes they are, but it is very difficult :)
eacamero_ has joined #nixos
<copumpkin> humans are bad at sorting and it's silly to expect them to
eacameron has quit [(Read error: Connection reset by peer)]
eacamero_ has quit [(Read error: Connection reset by peer)]
eacameron has joined #nixos
eacameron has quit [(Read error: Connection reset by peer)]
eacameron has joined #nixos
<gchristensen> eacameron: yo useem to be having network issues :)
<eacameron> gchristensen: :( tell me about it.
<eacameron> I'm wondering if I can use the same "options" stuff found in nixos packages for my own projects. I have a lot of configuration.
eacameron has quit [(Read error: Connection reset by peer)]
eacameron has joined #nixos
siel has joined #nixos
<gchristensen> domenkozar: can you restart the kde tests on 16.09 and unstable? I think they're failing "for no reason"
grw has quit [(Ping timeout: 256 seconds)]
zagy has quit [(Quit: Leaving.)]
Wizek has quit [(Ping timeout: 258 seconds)]
siel has quit [(Ping timeout: 245 seconds)]
Wizek__ has joined #nixos
servilio has joined #nixos
<servilio> hi all!
<servilio> is there a way to load the content of a file as a string?
<servilio> i.e.: having an external file to pass to writeScript ( https://github.com/NixOS/nixpkgs/blob/master/pkgs/build-support/trivial-builders.nix#L54 )
<gchristensen> servilio: builtins.readFile
<ToxicFrog> :ie: LnL, joachifm: re the syscall using $PATH or not: based on strace, it's calling execve(["/usr/bin/pkexec", "/bin/cp", ...], ...). So no $PATH respect.
<ToxicFrog> Profpatsch: in this case it's not a .so but a closed-source text editor (Sublime Text)
<servilio> gchristensen: thanks!!!
<LnL> ToxicFrog: that's what I was afraid of
<LnL> ToxicFrog: and /nix/store/.../bin/cp is definitely going to be longer then /bin/cp
Wizek_ has quit [(Ping timeout: 252 seconds)]
<ToxicFrog> I *pkexec* respects $PATH, so replacing "/bin/cp" with "cp" would work fine as long as I could get it to the point of invoking pkexec
<ToxicFrog> But the same is true of /nix/store/.../pkexec :P
<Profpatsch> ToxicFrog: Well, you can easily change the path of an executable by using mkWrapper
<LnL> oh it's sublime
<LnL> you probably don't want to run that in a chroot :p
<Profpatsch> sublime isn’t packaged yet?
<Profpatsch> fascinating
<ToxicFrog> Profpatsch: I thought mkWrapper was only useful for things like setting PATH=?
<ToxicFrog> Profpatsch: Sublime 2 is. Sublime 3 isn't; that's what I'm working on.
<ToxicFrog> I have it mostly working.
<dtzWill> can you chroot it + bind-mount over X11 socket and misc workspace dirs? lol O:)
<dtzWill> but of course I'd rather see it working in a more..compatible way lol
<ToxicFrog> Editor loads up, plugins load and run, etc -- not many changes needed from the ST2 package.
<LnL> dtzWill: lol, I guess that would work
<ToxicFrog> The bit I'm working on now is editing root files as not-root
servilio has left #nixos ["ERC (IRC client for Emacs 24.5.2)"]
servilio has joined #nixos
<ToxicFrog> If you attempt that in ST3, it (tries to) use pkexec to do the edits
<ToxicFrog> I.e. it writes the new file to /tmp and then invokes pkexec cp to overwrite the root-owned file
<ToxicFrog> Which then prompts you to sudo.
<ToxicFrog> A nicer editing experience than, say, 'xdg-su -c subl /etc/nixos/'
<ToxicFrog> bwuh
<ToxicFrog> I went looking and couldn't find that
<dtzWill> (does that work/not work? idk I just found it lol)
<ToxicFrog> Not in nix-env and not in the online nix packages search
<ToxicFrog> (but the online nix packages search doesn't seem to list *anything* that requires unfree -- I should fileabug)
siel has joined #nixos
<ToxicFrog> I'll give that a shot, thanks!
<ToxicFrog> (what's libredirect.so?)
hotfuzz has quit [(Read error: Connection reset by peer)]
hotfuzz has joined #nixos
<ToxicFrog> gchristensen: aah.
Wizek has joined #nixos
<gchristensen> ToxicFrog: how long did you work on packaging sublime?
<dtzWill> haha it looks like what I was saying I wished existed--a reusable component targetted at the common task of 'redirecting' through interposition
ixxie has joined #nixos
<dtzWill> ToxicFrog: looks like it uses libredirect to convince sublime '/usr/bin/pkexec' is something else
<dtzWill> transparently/automagically
<dtzWill> (it looks neat, investigating...)
<gchristensen> ToxicFrog: (let me know so I can document it as very painful)
<ToxicFrog> gchristensen: based on my commit logs, a few hours
<gchristensen> ToxicFrog: 3? more?
pie_ has quit [(Ping timeout: 256 seconds)]
<dtzWill> Profpatsch: libredirect looks awfully related to what we were discussing this morning! Just passing along in case it helps with things you're hacking on :)
<ToxicFrog> 4 but probably at around 50-75% load, so figure 2-3 hours of actual work
<ToxicFrog> Much of it spent staring at the output of strace
<dtzWill> haha it also rewrites the "/bin/cp" string to "cp\0\0\0\0..." xD
<ToxicFrog> Interestingly I didn't have the issue with plugin_host crashing without wrapProgram
<ToxicFrog> But possibly that issue would have arisen as I fixed other stuff :)
<dtzWill> ToxicFrog: sorry for the news, but glad I helped you find out sooner rather than even later... O:)
siel has quit [(Read error: Connection reset by peer)]
siel has joined #nixos
Wizek_ has joined #nixos
<ToxicFrog> Well, I learned more things about Nix while working on it, so it's not like the time was completely wasted
<gchristensen> yeah but you could have spent that time on other stuff
<ToxicFrog> Hmm. Looks like the web-based package index is generated here: https://github.com/NixOS/nixos-homepage/blob/4ba15f08d17cc679d99a08231ccc1a84574bb694/Makefile#L198
<ToxicFrog> So yeah, it's just the output of nix-env -qa --json
<gchristensen> good digging
<gchristensen> do you know about search.nix.gsc.io?
<LnL> does nix-env -qa --json --arg config '{ allowUnfree = true; }' work?
<ToxicFrog> That is a good question! let me try it
* dtzWill doesn't, *clicky*
<ToxicFrog> neither do I
<gchristensen> that isn't how it works, that would list stuff which you already have installed
Wizek__ has quit [(Ping timeout: 252 seconds)]
<ToxicFrog> gchristensen: no, it appears to list everything, including things I don't have installed
<dtzWill> gchristensen: do you host this? Looks like plain 'nix.gsc.io' has a cert error (serves cert for search.nix.gsc.io), just FYI. (maybe redirect/update cert?)
<gchristensen> erm, sorry, I'm wrong
<LnL> wait
<ToxicFrog> And with `--arg config '{ allowUnfree = true; }'` it does indeed list unfree software!
<gchristensen> dtzWill: I do host it
<ToxicFrog> So this seems like a *really* simple patch to the nixos website generator
<LnL> ToxicFrog: yup :D
<gchristensen> dtzWill: can't do a redirect since the issue is before SSL, the cert would need an extra name for nix.gsc.io
<LnL> same for allowBroken perhaps?
<ToxicFrog> LnL: do you want to go ahead and submit the PR, or shall I, or...?
<ToxicFrog> What does allowBroken do?
<dtzWill> gchristensen: oh, okay, apologies for bad suggestions anyway NBD just thought I'd mention :)
<gchristensen> lets you try to install known-broken software
<ToxicFrog> (or rather, I can guess what it does, but what makes a package count as 'broken'?)
<gchristensen> dtzWill: it is a good point, I can get an additional name...
<dtzWill> yeah, I always thought broken might include derivations that fail to evaluate
<gchristensen> ToxicFrog: meta.broken = true
<LnL> we probably also want that because of stuff that's darwin only for example
<dtzWill> oh it looks like let's encrypt, so yeah that shouldn't be a /huge/ deal (compared to a purchsed SSL cert you have to call up or something lol), but I say that as someone who carefully setup his LE certs and hope they don't break lol
<gchristensen> no, yeah, it really should be easy lol
<dtzWill> anyway, ooo that search looks nice :D
<dtzWill> gchristensen: is that search service using some public/available thing? (like, if I wanted to host a similar instance for random other code?)
Shou has quit [(Ping timeout: 265 seconds)]
siel has quit [(Ping timeout: 258 seconds)]
<dtzWill> AND IT LINKS TO GITHUB?! EEeeeeeeeee
<dtzWill> lol
<dtzWill> ..*ahem*
<gchristensen> dtzWill: it is github.com/etsy/hound (available in nixpkgs / nixos as a service)
<dtzWill> awww yiss
<dtzWill> ty sir! :)
<gchristensen> :)
<simpson> Fancy.
<gchristensen> it even searches all-packages.nix and python-packages.nix, which github dares not do
<LnL> I just grep through nixpkgs when I'm looking for stuff :p
oida has quit [(Ping timeout: 258 seconds)]
<dtzWill> LnL: haha yep, more recently I'm replacing my kludge shell aliases with 'ripgrep' :D
<ToxicFrog> Anyways, it looks like adding that will fix the package search, and license information is already included, so if people want to avoid using unfree software it'll tell them right there on the website
<dtzWill> shell aliases mostly just doing find/xargs/grep but ignoring version control lo
<gchristensen> all the "Search" links on my roundups (https://github.com/NixOS/nixpkgs/issues/21289 for example) use hound, it is just better
<ToxicFrog> ...I may have a hard time submitting patches to the website because they don't have a LICENSE or COPYING
<dtzWill> ToxicFrog: just curious, what do you mean?
slack1256 has joined #nixos
<LnL> ToxicFrog: I'll do it
<gchristensen> niksnut: can you add a license to https://github.com/NixOS/nixos-homepage?
<ToxicFrog> dtzWill: when sending patches to a new (== not patched by you before), OSPO (the Open Source Projects Office) likes to review the first few
<ToxicFrog> This is an automated rubber stamp *as long as the project has an appropriate OSS license*
<ToxicFrog> If it doesn't an actual human needs to look at it
<dtzWill> oh! interesting, and ty for understanding I was sincere about being curious haha and not suggesting anything else :)
Wizek has quit [(Ping timeout: 246 seconds)]
nixy has quit [(Quit: WeeChat 1.0.1)]
Wizek__ has joined #nixos
<mbrock> dtzWill: I don't know what WiFi card the new XPS 15 has, but on the generation of XPS 13 I have from 2015, they came with either Broadcom or Intel. The Intel chip is just fine with Linux, but the Broadcom is kind of sucky and needs a blob driver. I had the Broadcom, but I recently changed it for an Intel 7265 at a cost of around 20 to 25 USD and ten
<mbrock> minutes of easy work (small Torx T-15 screwdriver required)
<gchristensen> LnL: looks like none of the packages on https://nixos.org/nixos/packages.html actually show a license?
<LnL> oh strange, that's probably a bug
Wizek_ has quit [(Ping timeout: 252 seconds)]
<gchristensen> I was going to say you should note in the PR that the search shows the license of the package, so it should be okay to show the package
<gchristensen> but, in fact, it does not
<mbrock> other than the Broadcom Wi-Fi, I'm very happy with the XPS (developer edition) running NixOS... the keyboard is nice, the carbon fibre is sweet, the 3200x1800 resolution (even higher on the new XPS 15) is beautiful, the touch screen works out of the box, etc
alex`` has joined #nixos
<alex``> hello
<Ralith> mbrock: there's a new XPS 15?
jensens has joined #nixos
<ToxicFrog> gchristensen: that information is definitely in the JSON
<alex``> ha i have xps15 :)
Wizek has joined #nixos
<ToxicFrog> So I guess that's an issue with the JS that renders the site
<Ralith> mbrock: the last gen is using a fancy broadcom thing that is theoretically supported, but I'm not aware of the next gen (w/ kaby lake) being out yet
<gchristensen> well, yes, but this change to the json is almost definitely not going to get merged without the fix to the JS, ToxicFrog
<alex``> i want to try a simple nodejs server on 127.0.0.1:8080 but i have a exception (i can listen port)
<Ralith> hype
<ToxicFrog> gchristensen: I figured
<dtzWill> mbrock: ooo ty for the info, and yeah broadcom has a bad track record.. $20-25 is /np/ glad it's reasonably priced lol
<dtzWill> not like $150 or some silly amount
<Ralith> mbrock: the wifi card is an upgradable M.2 deal then?
<alex``> Why is a lambda user unable to run a daemon on an 8080 port?
Wizek_ has joined #nixos
<Ralith> mbrock: hm, doesn't look like this is actually on sale yet, just announced
Wizek__ has quit [(Ping timeout: 252 seconds)]
<Ralith> or not even announced properly
<LnL> gchristensen: well it's fine anyway, people will notice when they try to install it
<alex``> on my xps15 Broadcom Corporation BCM43602 802.11ac Wireless
<ToxicFrog> gchristensen: found the problem, I think
<dtzWill> lol yeah that's my understanding as well, Ralith
<mbrock> hmm, yeah, the product page for me is pretty much empty right now. a few hours ago it was a nice announcement...
<ToxicFrog> var license = info['meta']['license'];
<ToxicFrog> if (typeof license == 'string')
<ToxicFrog> $('.license', details).empty().text(license);
<ToxicFrog> but License isn't a string; it's a map:
<ToxicFrog> packageData.ant.meta.license
<ToxicFrog> Object {fullName: "Apache License 2.0", shortName: "asl20", spdxId: "Apache-2.0", url: "http://spdx.org/licenses/Apache-2.0"}
<viric> really... why the nixos install disk doesn't have 'man'?
<LnL> ToxicFrog: I don't think there's a fixed type for any of the meta fields, but most should look like that
<alex``> my only problems is Gnome crash with touch-screen. but i'm use xmonad and its work like a charm
<ToxicFrog> LnL: there's a TODO for "handle lists and attrsets of licenses"
<ToxicFrog> Based on spot checking, everything I've looked at has had this form
<gchristensen> LnL: I don't think you understand how serious people are about this stuff, heh
<ToxicFrog> It should probably use as-is if it's a string, and if not try ['fullName', 'name', 'shortName'] in that order, I guess? Aggravating that there isn't a formal schema for this
Wizek has quit [(Ping timeout: 250 seconds)]
<LnL> gchristensen: for serious people it will fail to install :)
Wizek has joined #nixos
<ToxicFrog> But at any rate, this *also* looks like a pretty quick and easy fix, at least for the 90% case
<gchristensen> LnL: for some, it is fairly egregious to even _have_ the packages, much less show them.
siel has joined #nixos
Wizek__ has joined #nixos
<NixOS_GitHub> [nixpkgs] Profpatsch opened pull request #21313: picard: use lower mutagen version (master...picard-fix) https://git.io/v1Amx
bytemonger has joined #nixos
<ToxicFrog> I can submit a patch for that if the website has a license by the time I get home, but for now, I must away from work. Laters!
Wizek_ has quit [(Ping timeout: 252 seconds)]
bytemonger has quit [(Client Quit)]
Wizek_ has joined #nixos
jmiven has quit [(Quit: co'o)]
Wizek__ has quit [(Ping timeout: 252 seconds)]
emmanuel_erc has joined #nixos
<emmanuel_erc> hello there
<gchristensen> hello
Wizek has quit [(Ping timeout: 248 seconds)]
siel has quit [(Ping timeout: 260 seconds)]
<emmanuel_erc> I get this error when I try to install the nix package manager: http://lpaste.net/350238
<emmanuel_erc> has anyone here run into this error?
<LnL> :(
<LnL> never seen that before
oida has joined #nixos
<aristid> emmanuel_erc: do you have full permissions for /nix ?
<gchristensen> emmanuel_erc: "error: setting synchronous mode: disk I/O error" do you have disk errors in dmesg?
oida has quit [(Read error: Connection reset by peer)]
Wizek__ has joined #nixos
<emmanuel_erc> should I have added sudo when I used "curl https://nixos.org/nix/install | sh" to install nix
AllanEspinosa has joined #nixos
Wizek_ has quit [(Ping timeout: 250 seconds)]
siel has joined #nixos
ryanartecona has joined #nixos
<emmanuel_erc> gchristensen: how would I know if there is a disk error by checking dmesg?
<LnL> I don't think so
reinzelmann has quit [(Quit: Leaving)]
pikarudolph has quit [(Quit: Quit)]
<gchristensen> usually there are errors in there about, well, your disk :P
<gchristensen> like sd* or something
<emmanuel_erc> gchristensen: the only error I get is "[60108.663650] Chrome_ChildThr[12091]: segfault at 0 ip 000055b97b2d8129 sp 00007fc55dffe390 error 6 in plugin-container[55b97b2cf000+3e000]"
pikajude has joined #nixos
jmiven has joined #nixos
george__ has quit [(Quit: leaving)]
Wizek has joined #nixos
Wizek_ has joined #nixos
ixxie has quit [(Ping timeout: 246 seconds)]
siel has quit [(Ping timeout: 258 seconds)]
siel has joined #nixos
Wizek has quit [(Ping timeout: 240 seconds)]
Wizek__ has quit [(Ping timeout: 252 seconds)]
<NixOS_GitHub> [nixpkgs] abbradar pushed 1 new commit to release-16.09: https://git.io/v1AGb
<NixOS_GitHub> nixpkgs/release-16.09 b2e80a5 Nikolay Amiantov: samba4: 4.3.11 -> 4.3.13
oida has joined #nixos
siel has quit [(Ping timeout: 246 seconds)]
<eacameron> What's the easiest way to see which haskell packages (and versions) are in a nixpkgs version?
<jazzencat> i3, awesome or xmonad?
siel has joined #nixos
<kmicu> Is standard ‘nix-env -qaP '.*' f '<nixpkgs>' --description -A haskellPackages’ not enough?
Wizek has joined #nixos
nixy has joined #nixos
wkennington has joined #nixos
ixxie has joined #nixos
jgeerds has quit [(Quit: Leaving)]
<kmicu> 🙟 𝕹𝖎𝖝 𝕿𝖎𝖒𝖊𝖘 🙜 “…new feature in cabal called Nix–style local builds. It’s great. You should use it. It saved my PhD. Like I, literally, would not have been able to develop cabal changes without this.” ― https://youtu.be/A3ehG4GQpxU?t=2156
siel has quit [(Ping timeout: 246 seconds)]
<eacameron> kmicu: Wow.
Wizek has quit [(Ping timeout: 258 seconds)]
siel has joined #nixos
ixxie has quit [(Ping timeout: 258 seconds)]
<NixOS_GitHub> [nixpkgs] LnL7 opened pull request #21317: libupnp: 1.6.20 -> 1.6.21 (master...libupnp-1.6.21) https://git.io/v1AWn
<NixOS_GitHub> [nixpkgs] aristidb pushed 1 new commit to master: https://git.io/v1AWw
<NixOS_GitHub> nixpkgs/master 06cfff7 Aristid Breitkreuz: parted: re-enable checks
jensens has quit [(Ping timeout: 250 seconds)]
siel has quit [(Ping timeout: 250 seconds)]
alex`` has quit [(Ping timeout: 268 seconds)]
oida has quit [(Ping timeout: 265 seconds)]
Wizek has joined #nixos
siel has joined #nixos
oida has joined #nixos
nixy has quit [(Quit: Outtie)]
nixy has joined #nixos
<cstrahan> Did I see my name?
cpennington has quit [(Remote host closed the connection)]
<nahamu> I think I saw it go by.
<copumpkin> I saw your name when you said something a minute ago
<nahamu> someone asking you about a PR, IIRC.
<nahamu> found it:
<nahamu> <pierron> cstrahan: What is your opinion on https://github.com/NixOS/nixpkgs/pull/21243 ; Should I continue it and add documentation to the PR?
siel has quit [(Ping timeout: 260 seconds)]
roconnor has joined #nixos
ixxie has joined #nixos
siel has joined #nixos
aneeshusa has joined #nixos
<NixOS_GitHub> [nixpkgs] leenaars opened pull request #21318: ssrc: init -> 1.33 (master...ssrc) https://git.io/v1ABR
Wizek has quit [(Ping timeout: 258 seconds)]
<gchristensen> kmicu: !!
<gchristensen> kmicu: what is our user-record?
siel has quit [(Ping timeout: 258 seconds)]
siel has joined #nixos
slack1256 has quit [(Remote host closed the connection)]
Wizek__ has joined #nixos
Wizek has joined #nixos
siel has quit [(Ping timeout: 250 seconds)]
Wizek_ has quit [(Ping timeout: 250 seconds)]
Itkovian has joined #nixos
filterfish has quit [(Ping timeout: 260 seconds)]
siel has joined #nixos
newdan has joined #nixos
ixxie has quit [(Ping timeout: 250 seconds)]
Wizek_ has joined #nixos
Guest441 has quit [(Remote host closed the connection)]
hiratara has quit [(Ping timeout: 258 seconds)]
sigmundv has quit [(Quit: ZNC 1.6.3 - http://znc.in)]
sigmundv has joined #nixos
Wizek__ has quit [(Ping timeout: 250 seconds)]
hiratara has joined #nixos
filterfish has joined #nixos
mkoenig has joined #nixos
siel has quit [(Ping timeout: 260 seconds)]
Wizek has quit [(Ping timeout: 250 seconds)]
Itkovian has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
aj__ has quit [(Ping timeout: 250 seconds)]
gilligan_ has joined #nixos
c0bw3b has joined #nixos
siel has joined #nixos
<gchristensen> ToxicFrog: ping
aneeshusa has quit [(Ping timeout: 250 seconds)]
<stepcut> any idea how I install "collection-cjk" using the new texlive system? I can only find documentation on the old way
JagaJaga has quit [(Ping timeout: 268 seconds)]
<jack[m]> qq: how big is /nix on the hydra servers?
<gchristensen> jack[m]: not sure we know about all of them
<gchristensen> jack[m]: but the AWS ones have a 100GB /
<jack[m]> Huh. Wow.
<gchristensen> hrm?
<stepcut> oh, it is in the manual, nm.
<jack[m]> (that's small, I've got TiBs of /nix stores)
<LnL> stuff gets garbage collected pretty quickly once it's uploaded to the cache
<gchristensen> jack[m]: they get pruned regularly
<gchristensen> and the AWS servers die frequently
<LnL> the s3 bucket is pretty huge
<jack[m]> Mmm. Pruning the binary cache is another open question of mine...
<gchristensen> it doesn't get pruned, I don't think
jdanek has joined #nixos
<LnL> we probably should, at some point...
xwvvvvwx has joined #nixos
Wizek_ has quit [(Ping timeout: 250 seconds)]
siel has quit [(Ping timeout: 258 seconds)]
<gchristensen> LnL: it isn't a problem that needs solving right now, is all :P
<jdanek> hi, how do I set enableDebugInfo = true in ~/.nixpkgs/config.nix for a couple of chosen packages? I cannot do that systemwide, I do not have enough disk space.
mkoenig has quit [(Read error: Connection reset by peer)]
<jack[m]> Hrm. I should move my binary cache to something cheaper.
<LnL> I have a cache on my local nas :)
<gchristensen> jack[m]: how big is yours?
siel has joined #nixos
<gchristensen> jack[m]: one mostly-valid strategy is to just delete stuff older than N days
sigmundv_ has joined #nixos
sigmundv has quit [(Disconnected by services)]
sigmundv has joined #nixos
sigmundv has quit [(Disconnected by services)]
sigmundv_ is now known as sigmundv
sigmundv_ has joined #nixos
mkoenig has joined #nixos
siel has quit [(Ping timeout: 258 seconds)]
Shou has joined #nixos
<jazzencat> If I have a case where I need to install a system on multiple computers, can I generate hardware-config individually on each and then copy a configured configuration.nix and run the install that and get the same environment set up on each machine?
<jack[m]> gchristensen (IRC): ~500GiB. Which is not big, but at $20/GiB (don't ask, enterprise) I get some flak.
<jack[m]> LRU based on atime or something would probably be good, but I put some precious stuff in there, which is useful in archeology.
<gchristensen> jack[m]: O.o may I know where this is? also: use more space! fix the problems with economy of scale!
<jack[m]> gchristensen (IRC): intel
<gchristensen> jack[m]: were you the one doing chip verification?
<NixOS_GitHub> [nixpkgs] bjornfor pushed 1 new commit to master: https://git.io/v1Ao0
<NixOS_GitHub> nixpkgs/master 3fb785b Bjørn Forsman: nixos: unbreak prometheus test...
<jack[m]> gchristensen (IRC): yes, that is one of the hats I wear. The verif stuff isn't terribly big compared to the netlists, though..
<gchristensen> may I PM?
<jack[m]> gchristensen (IRC): okay, I'm on the matrix bridge, so lesse how this works!
<gchristensen> :)
hiratara has quit [(Quit: ZNC - http://znc.in)]
jdanek has left #nixos ["Leaving"]
hiratara has joined #nixos
siel has joined #nixos
sellout- has joined #nixos
aj__ has joined #nixos
sellout- has quit [(Client Quit)]
<NixOS_GitHub> [nixpkgs] teh opened pull request #21320: Update docs for Nginx + ACME. Closes #20698. (master...master) https://github.com/NixOS/nixpkgs/pull/21320
kantor has joined #nixos
mkoenig has quit [(Read error: Connection reset by peer)]
aneeshusa has joined #nixos
Shou has quit [(Ping timeout: 258 seconds)]
siel has quit [(Ping timeout: 246 seconds)]
<NixOS_GitHub> [nixpkgs] rardiol opened pull request #21321: Worldengine (master...worldengine) https://github.com/NixOS/nixpkgs/pull/21321
kampfschlaefer has quit [(Ping timeout: 258 seconds)]
mkoenig has joined #nixos
aneeshusa has quit [(Ping timeout: 265 seconds)]
siel has joined #nixos
<gilligan_> hrm.. i am having some problem with /run/keys ;/
<gilligan_> seems to have correct user & group but i still get Permission denied and can only read contents with root user ;(
xwvvvvwx has quit [(Ping timeout: 264 seconds)]
Fare has joined #nixos
siel has quit [(Ping timeout: 250 seconds)]
mkoenig has quit [(Read error: Connection reset by peer)]
Shou has joined #nixos
mkoenig has joined #nixos
siel has joined #nixos
<spacekitteh> how cna i define a shell environment in my configuration.nix?
<newdan> How can I use nix-env to install Python 2.6.9? I see the expression at ~/.nix-defexpr/channels/nixpkgs/pkgs/development/interpreters/python/cpython/2.6/default.nix
Shou has quit [(Ping timeout: 248 seconds)]
civodul has quit [(Quit: ERC (IRC client for Emacs 25.1.1))]
<newdan> But `nix-env -i python-2.6.9` doesn't work, "error: selector ‘python-2.6.9’ matches no derivations"
<NixOS_GitHub> [nixpkgs] mimadrid opened pull request #21322: meld: 3.16.2 -> 3.16.4 (master...update/meld-3.16.4) https://git.io/v1A1g
ok2 has joined #nixos
<jazzencat> I love it when I mistype things in configuration.nix.
<AtnNn> newdan: I would try something like `nix-env -qa 'python.*2.*6.*'` or `nox python`
<AtnNn> newdan: as a last resort, I take a look at `pkgs/top-level/all-packages.nix'
<jazzencat> Hm, the package for Lumina Desktop exists, but no services.xserver.desktopManager.lumina.enable. Can I still make use of it if I install it through nix-env?
<NixOS_GitHub> [nixpkgs] FRidh closed pull request #21322: meld: 3.16.2 -> 3.16.4 (master...update/meld-3.16.4) https://git.io/v1A1g
<NixOS_GitHub> [nixpkgs] FRidh pushed 2 new commits to master: https://git.io/v1A19
<NixOS_GitHub> nixpkgs/master 8021c1f mimadrid: meld: 3.16.2 -> 3.16.4
<NixOS_GitHub> nixpkgs/master 48fa032 Frederik Rietdijk: Merge pull request #21322 from mimadrid/update/meld-3.16.4...
siel has quit [(Ping timeout: 258 seconds)]
<AtnNn> jazzencat: perhaps start it in `services.xserver.displayManager.session.script'
<jazzencat> Okay.
siel has joined #nixos
<jazzencat> Though looking at the nixpkg for it, the services.xserver option is listed...
isHavvy has joined #nixos
FareTower has joined #nixos
fare__ has joined #nixos
<jazzencat> Meh, I'll just remove it for now.
Havvy has quit [(Ping timeout: 268 seconds)]
<jazzencat> Not up to messing around with the system too much.
Fare has quit [(Ping timeout: 256 seconds)]
Fare has joined #nixos
sigmundv_ has quit [(Quit: ZNC 1.6.3 - http://znc.in)]
FareTower has quit [(Ping timeout: 246 seconds)]
sigmundv_ has joined #nixos
Shou has joined #nixos
fare__ has quit [(Ping timeout: 252 seconds)]
danharaj has quit [(Quit: Leaving.)]
siel has quit [(Ping timeout: 250 seconds)]
ebzzry has joined #nixos
filterfish has quit [(Ping timeout: 240 seconds)]
<ToxicFrog> gchristensen: I'll be around in about an hour and a half, once the toddlebot is in bed
<ToxicFrog> jazzencat: I assume so
<NixOS_GitHub> [nixpkgs] bjornfor created prometheus (+1 new commit): https://git.io/v1AyL
<NixOS_GitHub> nixpkgs/prometheus 577b61b Bjørn Forsman: nixos/prometheus: add services.prometheus.configText option...
<NixOS_GitHub> [nixpkgs] bjornfor deleted prometheus at 577b61b: https://git.io/v1AyW
<NixOS_GitHub> [nixpkgs] bjornfor opened pull request #21323: nixos/prometheus: add services.prometheus.configText option (master...prometheus) https://git.io/v1AyK
Fare has quit [(Ping timeout: 268 seconds)]
c0bw3b has left #nixos []
Fare has joined #nixos
tmobile has joined #nixos
ryanartecona has quit [(Quit: ryanartecona)]
FRidh has quit [(Remote host closed the connection)]
FareTower has joined #nixos
Fare has quit [(Ping timeout: 250 seconds)]
sellout- has joined #nixos
fare__ has joined #nixos
<tmobile> Could someone point me to an example of a nixpkg derivation that adds a udev rule(if one exists)?
<tmobile> I'm curious about the Nix equivalent of copying a rules file to /etc/udev/rules.d
FareTower has quit [(Ping timeout: 256 seconds)]