<elvishjerricco>
gchristensen: There are a lot more arguments to that `release.nix` than there are inputs for that job set. Are those just using the defaults when hydra builds it?
<elvishjerricco>
But also, it definitely doesn't just accept any expression that `nix-build` would accept. For example, `nix-build` will accept an attrset with functions, and it'll just skip them and build only the derivations. But hydra will complain if it encounters a function.
<gchristensen>
it should do okay with functions, I think, and just not build them?
orivej has quit [(Ping timeout: 248 seconds)]
csingley has quit [(Ping timeout: 240 seconds)]
<elvishjerricco>
gchristensen: Nah. Hydra throws a bunch of "unsupported value" errors if you try to leave in any non-derivation values.
dhess has quit [(Remote host closed the connection)]
<gchristensen>
but does it build other things anyway?
<elvishjerricco>
I don't think so
ryantrinkle1 has quit [(Ping timeout: 252 seconds)]
<NixOS_GitHub>
[nixpkgs] abbradar pushed 5 new commits to staging: https://git.io/vd9bN
<NixOS_GitHub>
nixpkgs/staging 8f43c56 Nikolay Amiantov: libtorrentRasterbar: 1.1.4 -> 1.1.5...
<NixOS_GitHub>
nixpkgs/staging 3e4737d Nikolay Amiantov: bastet: fix build with new boost
<NixOS_GitHub>
nixpkgs/staging 593c6a8 Nikolay Amiantov: povray: 3.7.0.3 -> 3.7.0.4...
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] copumpkin opened pull request #1617: [Nasty sketch] Stop crashing High Sierra (master...shitty-kill-processes) https://git.io/vd9Ne
NixOS_GitHub has left #nixos []
<disasm>
wow... installing unifi service to patch my access point just worked with no tweaking! Impressed :) I thought that was going to be a huge hassle I'd give up and reboot into osx to resolve
thc202 has quit [(Ping timeout: 255 seconds)]
<gchristensen>
disasm: me too :o
Wizek has quit [(Quit: Connection closed for inactivity)]
shanemikel has joined #nixos
glenn is now known as scribbler
shanemikel has quit [(Read error: Connection reset by peer)]
<yurrriq>
Are there any nixpkgs+(Quick)Lisp users in here that can show me a working example? My endgame is to run system tests using prove on Travis with `language: nix`. Any advice, code, or blog post links would be greatly appreciated!
<yurrriq>
Aside: Iām loving the stable Darwin channel. Many thanks to @LnL and whoever else is involved.
<dhess>
yurrriq: I saw that recently. Does it just track 17.09?
mizu_no_oto has joined #nixos
Guest92359 has quit [(Ping timeout: 240 seconds)]
<yurrriq>
dhess: yeah, but (most) everything Just Works TM
<yurrriq>
My lilypond-with-fonts package seems to have broken, but I think that has more to do with gcc and/or guile_1_8
MarcelineVQ has joined #nixos
pie_ has joined #nixos
pxc has quit [(Ping timeout: 260 seconds)]
<ldlework>
Do actual nix package expressions typically utilize a package distribution's distributed build scripts/mechanisms or are they generally completely replaced with descriptions in the Nix expressions?
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
<copumpkin>
we usually reuse build scripts
<copumpkin>
although it varies a bit with package ecosystem
<dhess>
Dezgeg: we discussed this briefly in the past, but I can't find the conversation. So sorry to repeat myself, but is it possible in NixOS to build for armv7l-linux on an aarch64-linux system?
<dhess>
I tried to do that with my Jetson TX1 build host but it complained that it wasn't armv7l
<Dezgeg>
there's still an open issue on how to trick uname() to return armv7l inside 32-bit builds
<dhess>
pretty sure that you can build armv7l binaries on a Jetson TX1 running L4 whatever
<samueldr>
dhess: I have the notes around
<dhess>
ahh ok, seemed like it might be a NixOS issue, yes
<dhess>
samueldr: can you make them availble online somewhere?
<dhess>
hmm that's weird, that 2017-09-11 convo with you and clever indicates that clever is using a multi-arch box (armv6l and armv7l)
<dhess>
I wonder how he go that to work
Arcaelyx_ has joined #nixos
<samueldr>
I believe he has a couple of tricks
jb55 has joined #nixos
<samueldr>
if you look at the nix-misc/default.nix file I linked from clever
<samueldr>
there's qemu-user-arm
<dhess>
When I tried that with armv7l-linux,aarch64-linux I got the "I'm not an armv7l-linux"
<samueldr>
this allows executing arm binaries on the host system through qemu
<dhess>
samueldr: yeah I'd like to avoid qemu
<samueldr>
afaik, it's still not possible without "actual arm"
<samueldr>
be it through emulation or hardware
<dhess>
I have an armv7l build machine that works fine, it's just slow, so I'd like my aarch64-linux to help out with armv7l builds
<samueldr>
that was the state as of september
<dhess>
samueldr: oh so he's actually using qemu-user-arm on an arm?
<samueldr>
I don't believe so
Arcaelyx has quit [(Ping timeout: 246 seconds)]
<dhess>
ok
<samueldr>
gchristensen might know more about that, but pretty sure you'd need an armv7l distro running on your aarch64 linux for this to work
<dhess>
yeah that kind of defeats the point for me.
<samueldr>
though, thinking about it, if your board can run armv7l binaries, maybe using clever's patch that adds "build-extra-platforms" could be enough
<Dezgeg>
no, it should all just be getting uname() to return the right thing
<dhess>
Dezgeg: I assume clever's patch there is roughly the nix patch you were thinking of?
<Dezgeg>
yes
<dhess>
right
<dhess>
ok you two, that's very useful and should give me something to play with. Thanks!
<samueldr>
hope it works!
<dhess>
here I am trying to unwind all the various hacks I've got going in my Nix/nixpkgs environment and now I'm about to add some more :)
<ben-coman>
greetings, I'm 2 days into using NixOS for the first time, and wanted to discuss some lagging of the UI I'm experiencing. I've had a quick search of the bug tracker and not noticed anything related.
<ben-coman>
I installed per https://chris-martin.org/2015/installing-nixos on a Dell Lattitude E6520 (Intel graphics) except I re-did it without the encrypted drive after I experienced the UI lagging.
<samueldr>
I'm assuming you installed 17.09 though, right?
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
<ben-coman>
Usually after booting the system is okay for hours, then when moving between windows (e.g. Chromium to Konsole) there is a 3 to 5 second lag before I can type anything
iqubic has joined #nixos
<samueldr>
using plasma5 then I presume
<ben-coman>
@samueldr, yes. full install "17.09.1716.68ef4b14bc (Hummingbird)"
<ben-coman>
yes
<ben-coman>
yes, plasma5
saf_spa has joined #nixos
<ben-coman>
my next step will be to switch to something else, but first I wanted see if there was any useful details I can report on to you guys.
<saf_spa>
Hi! How do I list installed system packages? Not user/root-profile packages, but those pulled by system profile? Thanks!
<samueldr>
I have a hunch, but I'm not even sure I'm right ben-coman
<samueldr>
give me a minute or two to check something on my laptop
<saf_spa>
"nix-env -q" does not cut it
<ben-coman>
I've experienced the problem with Chromium & Firefox & Konsole.
<ben-coman>
thx samuel
<gchristensen>
saf: what it means to be "installed" gets a bit fuzzy. what're you looking to do?
<ldlework>
Is it possible to just fetchFromGithub and provide the ref as "master" and not provide the sha?
<gchristensen>
no
<ben-coman>
another observation, after last reboot a few minutes ago and started Firefox, typing into the address bar two characters appeared and then a 4 second pause.
<ben-coman>
before the remaining dozen characters I typed appeared.
<saf_spa>
gchristensen: those pulled by environiment.systemPackages essentially, a system profile.
lambdamu has joined #nixos
<samueldr>
ben-coman: I think, one or two years ago when I used kwin, that I saw something similar due to its compositing settings
<ben-coman>
and btw, after I close Chromium and Firefox, and have only Konsole open, the problem remains, but to a lesser degree.
<samueldr>
in kde's settings, there should be desktop effects
<samueldr>
in that panel, there probably will be the advanced panel
<samueldr>
with compositing type
csingley has joined #nixos
<samueldr>
I believe I previously used xrender, or something like that
<samueldr>
it's been a while since I used kwin
lambdamu_ has quit [(Ping timeout: 248 seconds)]
<samueldr>
you can try the different options there first
<dgpratt>
I suppose that if I am installing a utility like stylish-haskell, it would be better installed via e.g. nix-env than stack because nix gives me a bit better control over installing/removing stuff, does that make sense?
<dhess>
dgpratt: That's the way I do it.
<ben-coman>
samueldr: I found the dialog "Desktop Effects - System Settings Module". I looks a bit different to yours. No tabs and more candy-like. When I type "composit" into search, nothing shows.
<samueldr>
hmm, I'm basing this on google search screenshots of kwin settings
<samueldr>
if you didn't know, you can create options in your configuration.nix file, (and imported files) that you can then use
<ldlework>
Yeah I've been doing that
<ldlework>
I'm writing plugins for HM and then importing them with fetchFromGithub and then using them in my configuration
<ldlework>
its freaking awesome
ebzzry has joined #nixos
S0rin has joined #nixos
<ben-coman>
samueldr: This seems related... https://bugs.kde.org/show_bug.cgi?id=371731, particularly the last comment by jeconom, but I can't find that setting "Provide buttons to close the windows"
<ben-coman>
except I hadn't been closing many windows to observe that part of their comment.
<ben-coman>
anyway, I'll continue for now testing with the compositor disabled.
<ben-coman>
what is best practice? should I copy that gist into my configuration.nix, or hardware-configuration.nix, or create new hardware-gpu-intel.nix and import it to one of the others?
<samueldr>
that, I don't know
<jeaye>
I'd create a separate one, ben-coman, to better organize your configs.
<samueldr>
I've removed other stuff from that file for the gist
pxc has joined #nixos
<samueldr>
(it wasn't relevant to our discussion)
<samueldr>
you backlight works properly, right?
oida has quit [(Ping timeout: 248 seconds)]
<ben-coman>
backlight is good
<samueldr>
I don't know if it was needed for real, or if it was only one of the backlight tool that wanted it, but I had some system packages in that file
riclima has joined #nixos
<samueldr>
(if you refresh the gist, I put everything back in)
<samueldr>
and there's glamor, that I don't even really know what it does
<samueldr>
yes, I meant that I don't know when you would really need this
<samueldr>
it's accelerated 2D; what apps benefits from that? dunno
riclima has quit [(Ping timeout: 255 seconds)]
riclima has joined #nixos
<disasm>
any ideas how I can fix this? can't load package: package github.com/go-gitea/gitea: code in directory /tmp/nix-build-gitea-1.2.1.drv-0/go/src/github.com/go-gitea/gitea expects import "code.gitea.io/gitea" The app is using custom import paths for flexibility to change code hosting at any point.
endforma1 has joined #nixos
<disasm>
oh wait, might have found it
<ben-coman>
My reading just now is Glamor is accelerating 2D using 3D GPUs. It probably not useful for my IntelHD3000 chipset
endformationage has quit [(Ping timeout: 240 seconds)]
frankpf has quit [(Quit: Leaving)]
<samueldr>
not sure though, it looks like it might actually be useful
<ben-coman>
for the curious... "GLAMOR is an acceleration library merged into the X.Org Server 1.16 that provides 2D acceleration implemented over OpenGL so X.Org DDX drivers no longer need to implement 2D-specific code-paths. GLAMOR is designed to be universal amongst OpenGL / OpenGL ES supported hardware."
<ben-coman>
somewhere I got the feeling that /nix/var/nix/profiles is related (??) and in there are are several links "system-N-link" for N=1-->19
<ben-coman>
does that sound usual?
shanemikel has quit [(Read error: Connection reset by peer)]
hexamod has joined #nixos
hexamod has quit [(Client Quit)]
<ben-coman>
samueldr: tracking... I didn't "disable compositor on startup", just changed to DRI2, but the problem just reoccurred. Next I'm trying RenderingBackend=OpenGL3.1.
<bencoman>
This is my first build I'm trying on NixOS, but I had heard such changes won't work. Although it does untar, jsut all in the base directory.
<srhb>
I'm a bit puzzled that you even get the same hash
<srhb>
Anyway, I just added the file locally and this works
<srhb>
the mkdir src, cd src, tar xf $src thing
<srhb>
Confirmed by setting configurePhase = "ls";
<bencoman>
I see not nix-prefetch-url gives a different ahs each time. I didn't check that before.
<srhb>
Yeah, that will break everything.
<srhb>
Or rather, it will only work for you, since you prefetched it.
<srhb>
(Until you collect garbage)
<bencoman>
Well its a git repo. The tarball must be generated dynamically. I could just copy the tarball to a http server and it would remain the same.
<srhb>
If you have access to the git repo, perhaps use fetchgit instead of fetchurl
<srhb>
And lock it to some ref.
<bencoman>
How do I "add the file locally" to match what you did? Just sit the tarball next to the default.nix file?
<srhb>
src = ./relative/path/to/file
<bencoman>
ahhh, I cam across some discussion in the bug tracker about how fetching from git wasn't giving identical hashes. Maybe that discussion is outdated. Let me see if I can find it again.
<srhb>
But the fetchgit solution really should be preferable.
<srhb>
bencoman: I have to get to work now. Good luck. I'll check in in about an hour or so :)
<bencoman>
thx
JosW has joined #nixos
<bencoman>
I try local first as a bit of learning, then fetchgit.
bkchr has joined #nixos
bkchr has quit [(Remote host closed the connection)]
<rycee>
ldlework: In your osx.nix you can try replacing `nixpkgs = (import <nixpkgs>{});` by changing the argument from `{ ... }:` to `{ pkgs, ... }:`ā¦
hotfuzz has quit [(Read error: Connection reset by peer)]
hotfuzz has joined #nixos
<hyper_ch>
can you actually pm someone on github?
<srhb>
hyper_ch: Not anymore.
<hyper_ch>
too bad
jtojnar has quit [(Read error: Connection reset by peer)]
Oida has quit [(Ping timeout: 248 seconds)]
Oida has joined #nixos
notdaniel has joined #nixos
thc202 has joined #nixos
bkchr has joined #nixos
<bkchr>
adisbladis: Hi, did you see my comment yesterday in github? Regarding plasma
leat has quit [(Quit: WeeChat 1.9.1)]
goibhniu has joined #nixos
<adisbladis>
bkchr: Yes I did see it :)
<adisbladis>
Building as we speak
<bkchr>
ahh nice :)
ThatDocsLady has joined #nixos
jtojnar has joined #nixos
ma27 has joined #nixos
<srhb>
If I create networking.firewall.extraCommands, it seems the rules there are duplicated on each rebuild that touches the firewall. How do I avoid this?
<srhb>
I suppose I could get around it by creating a new chain, but then I'm not sure how to avoid trying to create _that_ each rebuild
jtojnar has quit [(Read error: Connection reset by peer)]
ssmike has joined #nixos
Piece_Maker has joined #nixos
davidak has quit [(Read error: Connection reset by peer)]
puckipedia has quit [(Ping timeout: 248 seconds)]
hydraz has quit [(Ping timeout: 248 seconds)]
pbogdan has quit [(Ping timeout: 248 seconds)]
himmAllRight has quit [(Quit: No Ping reply in 180 seconds.)]
shymega has quit [(Ping timeout: 248 seconds)]
todder has quit [(Ping timeout: 248 seconds)]
andrewmcveigh has quit [(Ping timeout: 248 seconds)]
kosmikus has quit [(Ping timeout: 248 seconds)]
Acou_Bass has quit [(Ping timeout: 248 seconds)]
Piece_Maker is now known as Acou_Bass
mortum5 has joined #nixos
lsix_ has joined #nixos
lsix has quit [(Ping timeout: 252 seconds)]
nil has quit [(Ping timeout: 260 seconds)]
nil has joined #nixos
HappyEnte has quit [(Ping timeout: 260 seconds)]
<srhb>
rycee: I got it! Thanks a lot.
pbogdan_ has quit [(Ping timeout: 260 seconds)]
Bane^ has quit [(Ping timeout: 260 seconds)]
Vorpal has quit [(Ping timeout: 260 seconds)]
<srhb>
Reversing the order of whatever the extraCommands was seems like a plan of sort.
markus1199 has quit [(Ping timeout: 260 seconds)]
hydraz_ is now known as hydraz
hydraz has quit [(Changing host)]
hydraz has joined #nixos
markus1189 has joined #nixos
MP2E has quit [(Remote host closed the connection)]
Vorpal has joined #nixos
Vorpal has quit [(Changing host)]
Vorpal has joined #nixos
HappyEnte has joined #nixos
Bane^ has joined #nixos
pbogdan has joined #nixos
<rycee>
Cool :-)
nikivi has quit [(Ping timeout: 255 seconds)]
Acou_Bass has quit [(Ping timeout: 246 seconds)]
ssmike has quit [(Ping timeout: 240 seconds)]
nikivi has joined #nixos
ma27 has quit [(Ping timeout: 240 seconds)]
ison111 has quit [(Ping timeout: 240 seconds)]
Acou_Bass has joined #nixos
hamishmack has joined #nixos
<rycee>
srhb: Actually, I noticed that the firewall reload runs the extraCommand script but not the extraStopCommands one so rules will still build up :-(
<srhb>
What? That's not what I'm seeing.
<srhb>
restarting the firewall.service arbitrarily many times with the correct stop commands does keep the count of rules constant for me
<rycee>
Yeah, but if you try `systemctl reload firewall.service` I think you'll see a buildupā¦
<srhb>
But isn't that explicitly telling systemd to skip the stop script?
<rycee>
I haven't noticed that before. I had 7 duplicates of port redirects :-)
qzd has joined #nixos
<rycee>
srhb: Well, I think it still should keep it in a sane state. The module even marks it with `reloadIfChanged = true;` so nixos-rebuild will use reload instead of restart.
<rycee>
ā¦ and the builtin start script will reset rules before setting them again.
proteusguy has joined #nixos
<ixxie>
is there a quick way to check which channels a commit made it to?
<rycee>
I put `${config.networking.firewall.extraStopCommands}` at the top of my `extraCommands` and now it doesn't add rules when I reload :-)
<rycee>
ixxie: Under the commit message it should show the branch the commit is in.
<ixxie>
right
<ixxie>
and there are branches corresponding to each channel?
<rycee>
ixxie: Yeah. I'm not sure if this is a reliable or very convenient way to do it but it's the best idea I have at the momentā¦
lsix_ has quit [(Remote host closed the connection)]
lsix_ has joined #nixos
<ixxie>
thanks Rycee
mkoenig has quit [(Ping timeout: 248 seconds)]
__Sander__ has joined #nixos
<hyper_ch>
hmmm, still can't find "sigc++/adaptors/bind.h" when building....
mkoenig has joined #nixos
<rycee>
srhb: I noticed that the nat module does `networking.firewall.extraCommands = mkBefore flushNat;` so my hacky solution will duplicate the nat flushingā¦
nschoe has joined #nixos
<rycee>
srhb: So my final hacky solution is to simply put `${extraStopCommands}` at the top of `extraCommands` and use the `rec` keyword
dannyg has joined #nixos
<rycee>
Now I'll go grab a bite to eat and try to forget all I learned about iptables and firewalling :-)
hke has quit [(Ping timeout: 240 seconds)]
CodeWarrior has quit [(Ping timeout: 248 seconds)]
<Isorkin>
Hi. New kernel - not work nftables rules - reject. nft add rule inet filter input tcp dport \{ 22 \} ip saddr \@f2b-SSH reject - error. Rules drop - work. diff config kernel - https://pastebin.com/f1FQM0RV kernel 4.13.5 - nft worked, 4.13.7 - not work(
ssmike has joined #nixos
<pi3r>
Short little question: I am pinning all user derivations to a specific nixpkgs commit and wonder if "staging" from hydra might be any "better" than trunk ?
hke has joined #nixos
<tilpner>
What do you mean by pinning all user derivations?
<tilpner>
(Are you running a Hydra?)
csingley has joined #nixos
ssmike has quit [(Ping timeout: 255 seconds)]
<pi3r>
I mean derivation that I install with nix-env -i xxxx -f pin.nix (where pin defined a specific commit of nixpkgs)
<pi3r>
I usually looks on hydra.org to see if my packages build correctly before changinng the commit reference.
Lisanna has joined #nixos
<pi3r>
And on hydra.nixos.org I can pick trunk, staging and stable and wonder if staging is any "better" than trunk for my purpose
CodeWarrior has joined #nixos
hamishmack has quit [(Quit: hamishmack)]
<bkchr>
adisbladis: I disabled gold linker in qtbase and now the compilation works again. Maybe gold linker should be deactivated, until the error is resolved upstream in gold linker
<adisbladis>
bkchr: Cool :) I was gonna do the same but it's taking too long.
<adisbladis>
Let's see if 5.11.1 builds OK before changing Qt.
<bkchr>
Yeah it needs to recompile almost everything... :(
<Lisanna>
Anyone here familiar with the vmTools? Wondering if it's possible to use them to generate image files from the result of running something in e.g. runInLinuxVM
ssmike has joined #nixos
<sphalerite>
Lisanna: yes, it is
<sphalerite>
hang on
aloiscochard has joined #nixos
<sphalerite>
nixos/lib/make-disk-image.nix does exactly that
<bkchr>
adisbladis: ahh okay, with the fix it still works ^^
<bkchr>
adisbladis: so we should disable gold linker and wait for a new release
p4cman has joined #nixos
<adisbladis>
bkchr: I think so.
<samae>
Does anyone use the libreswan service successfuly?
<Lisanna>
sphalerite: cool, thanks. looks like I can hack on this to build what I want
<tilpner>
My config contains services.locate = { enable = true; locate = pkgs.mlocate; }, and I get the warning that "mlocate does not support searching as user other than root"
juriq has quit [(Ping timeout: 248 seconds)]
<tilpner>
The source suggests this should only be printed when I specify localuser? Does this work for anyone else?
<sphalerite>
tilpner: the default value for localuser is "nobody" and not null
<tilpner>
Ugh, yes. Thanks, sphalerite! :)
<sphalerite>
so you'd have to specify it to be null
<adisbladis>
bkchr: Will you do it?
catch22 has quit [(Remote host closed the connection)]
hke has quit [(Ping timeout: 240 seconds)]
<sphalerite>
tilpner: out of curiosity, why mlocate rather than findutils?
<bkchr>
adisbladis: the patch for qt? or plasma 5.11.1?
<Svarog>
i've been trying to find some info on the best way to get buildRustPackage to use a more up to date registry - is there a way to do it without having to write a bunch of .nix files?
<adisbladis>
bkchr: The Qt one
<tilpner>
sphalerite - I was running findutils with root anyway, and the merging aspect makes it faster when forcing a manual update
<sphalerite>
ah ok
<bkchr>
adisbladis: I would integrate the patch to your plasma patches? A single commit, but without the pr would not work :)
<sphalerite>
Svarog: I did that just yesterday! Hang on
<Svarog>
ooh
<tilpner>
The default of nobody is a little weird, no interesting files are visible to nobody, so why would you want to index with it?
<Lisanna>
sphalerite: my ultimate goal is to write a function which permutes some specified partition in a given .img file and produces a new .img file, so not sure if that's relevant, but thanks for the additional reference ^^
<sphalerite>
Svarog: just update rev and sha256 in pkgs/top-level/rust-packages.nix
<adisbladis>
bkchr: You could send me a link to your branch? I'll merge it to my working branch for the PR
<tilpner>
Svarog, sphalerite - You can also pass a rustRegistry to buildRustPackage
<sphalerite>
oooh ok
<Svarog>
yes i've read about that - let the build fail, then copy the sha256 sig from the error message and plop that into depsSha256 if i understood correctly
<Svarog>
tilpner, oh
<Svarog>
how does that work?
bencoman has quit [(Quit: Konversation terminated!)]
MercurialAlchemi has joined #nixos
<Svarog>
do you need to create a new registry expression similar to what's in rust-packages.nix and then pass that along?
<Svarog>
interesting - will have to try but that looks close to what i was hoping to find
<Svarog>
thanks tilpner and sphalerite
<viric>
gaa. nix-store --export wants write-access to the nix db
<sphalerite>
tilpner: ooooh, I like the manual use of rustChannelOf as well! I was wondering how I could get a "stable nightly" that doesn't want to redownload every day
<tilpner>
sphalerite - I should note that this default.nix doesn't actually build. This is only for development :/
<tilpner>
But the rustRegistry argument should work everywhere
<tilpner>
(Otherwise weechat, rustfmt and rustracer wouldn't be nativeBuildInputs)
<sphalerite>
tilpner: and the rustChannelOf bit works too, right?
<tilpner>
Run with: s rkt run --debug --stage1-path=$(nixpkgs rktStage1Kvm) docker://redis --insecure-options=image
<adisbladis>
bkchr: Also rebased on latest master
<bkchr>
adisbladis: hopefully, this time everything works :D
puck1pedia is now known as puckipedia
<Isorkin>
Please help. New kernel - not work nftables rules - reject. nft add rule inet filter input tcp dport \{ 22 \} ip saddr \@f2b-SSH reject - error. Rules drop - work. diff config kernel - https://pastebin.com/f1FQM0RV kernel 4.13.5 - nft worked, 4.13.7 - not work(
simukis has joined #nixos
thc202 has quit [(Ping timeout: 240 seconds)]
<adisbladis>
bkchr: Not the smoothest pr
thc202 has joined #nixos
steshaw has joined #nixos
<adisbladis>
bkchr: I saw that some people had reported similar sddm issues that were supposedly fixed with 5.11.1
<adisbladis>
So I'm hopeful your issue will be resolved too :)
<bkchr>
Ohh that would be nice :)
hke has joined #nixos
ertes has quit [(Ping timeout: 255 seconds)]
TweyII has joined #nixos
Neo-- has joined #nixos
TweyII is now known as Guest37279
csingley has quit [(Quit: WeeChat 1.9.1)]
ylwghst has joined #nixos
hke has quit [(Ping timeout: 248 seconds)]
<ylwghst>
hi, i'm not able to stop x11 gnome. How I should do it?
<srhb>
ylwghst: Are you opening it via a display manager (eg. GDM?) In that case, the unit is display-manager.service, afaik
<mpickering>
Is it normal for contributors to force push to my own fork of nixpkgs to update a pull request?
TheAppleMan has quit [(Quit: No Ping reply in 120 seconds.)]
<ylwghst>
srhb: Yes I do. Im not able to kill it using systemctl stop display-manager
notdaniel has quit [(Quit: Leaving)]
<srhb>
ylwghst: Odd.
<ylwghst>
it is
<ylwghst>
it just doesnt work, nothing happens when u run it
<srhb>
mpickering: No idea, but that would definitely surprise me.
<srhb>
As in "arrrgh!"
<kragniz>
mpickering: it's a bit rude to do it without asking
<Mic92>
mpickering: I only override the commit I added for transparency
<Mic92>
what is still missing there?
<mpickering>
That is true.. but it also broke the derivation
<Mic92>
that was clearly not my intention
<kragniz>
gchristensen: ping about stickers for freenode.live (I only just remembered :( )
sigmundv__ has joined #nixos
<Mic92>
mpickering: why is it still broken btw?
<hyper_ch>
so, just wrote a little nix expression to auto-resilver my external usb harddrives :)
<hyper_ch>
next I'll need some guinea pigs... I mean highly valued testers ;)
ylwghst has quit [(Quit: Lost terminal)]
hke has joined #nixos
<Mic92>
mpickering: ?
hke has quit [(Ping timeout: 255 seconds)]
hke has joined #nixos
ylwghst has joined #nixos
<ylwghst>
how I should upgrade nixos from 17.03 to 17.09?
<ylwghst>
should I nix-channel add; nixos-rebuild switch?
<Unode>
ylwghst: you might also have a reference on your configuration to 17.03. But yes, that would be the drill.
hke has quit [(Ping timeout: 248 seconds)]
leat has quit [(Quit: WeeChat 1.9.1)]
<hyper_ch>
hmmm, whats the best way to run a script before shutdown on nixos? with writscriptbin it's easy to create the script and make it available, but how to run it before shutdown? Do I need to created a systemd unit file or something?
<NixOS_GitHub>
[nixpkgs] grahamc pushed 2 new commits to master: https://git.io/vdHVX
<NixOS_GitHub>
nixpkgs/master edcdb95 Anthony Cowley: isync: 1.2.1 -> 1.3.0
<NixOS_GitHub>
nixpkgs/master df77869 Graham Christensen: Merge pull request #30527 from acowley/isync-1.3.0...
NixOS_GitHub has left #nixos []
<Unode>
noob question: when using nix-repl, do I need to import functions into scope to use them? Was trying to play with: concatStringsSep "\n" { "one" "two" } but get both a syntax error and otherwise concatStringsSep isn't defined.
<NixOS_GitHub>
nixpkgs/master 7d5c5eb Graham Christensen: Merge pull request #30523 from xvapx/playonlinux-update...
NixOS_GitHub has left #nixos []
<Unode>
can nix and nixUnstable coexist in the same environment?
ylwghst has joined #nixos
<Mic92>
mpickering: fixed by using: "SHAREDIR=$(out)/share/" instead
<Unode>
seemslike the answer to my last question is yes. Great!
<Guanin>
hi, I am trying to setup a service (gogs) that uses libpam to authenticate different users. I have noticed that setting security.pam.services."gogs".unixAuth = true will cause it to use the unix_chkpwd helper binary to check the supplied password. But as long as /etc/shadow remains unreadable to the user running gogs, the authentication will not work. However, the problem can be fixed (with a _minor_ loss of
<Guanin>
security that I am well aware of) when setting the binary setgid bit, and owning it by a group that also owns /etc/shadow. Now my question is: how can I make a derivative of pam that does this, and how can I set the group and mode of /etc/shadow subsequently?
<Mic92>
mpickering: I will not apply this change this time
hke has joined #nixos
Guanin has quit [(Remote host closed the connection)]
Guanin has joined #nixos
<Guanin>
sorry, had to move a bit and got disconnected
requisitebits has quit [(Ping timeout: 255 seconds)]
ssmike has quit [(Remote host closed the connection)]
Guanin has quit [(Remote host closed the connection)]
ssmike has joined #nixos
andymandias has quit [(Quit: ZNC 1.6.5 - http://znc.in)]
efx has joined #nixos
andymandias has joined #nixos
FRidh has quit [(Ping timeout: 248 seconds)]
hke has quit [(Ping timeout: 248 seconds)]
hke has joined #nixos
hotfuzz has joined #nixos
ixxie has joined #nixos
erictapen has joined #nixos
jbgi has quit [(Ping timeout: 240 seconds)]
FRidh has joined #nixos
requisitebits has joined #nixos
FRidh has quit [(Read error: Connection reset by peer)]
<infinisil>
gchristensen: I've changed the usbmuxd module (locally) to use the -U option to switch to a different user after start
<infinisil>
But I forgot my iPhone cable at home, so I can't test it :-)
erratum has quit [(Quit: Connection closed for inactivity)]
efx has quit [(Ping timeout: 240 seconds)]
ssmike has quit [(Remote host closed the connection)]
<trikl[m]>
If a build script contains hardcoded dependencies to /bin/bash, /bin/pwd and such, what is the proper way of proceeding if inside a nix-shell?
ssmike has joined #nixos
<infinisil>
patchShebangs
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 2 new commits to master: https://git.io/vdHKj
<NixOS_GitHub>
nixpkgs/master 339d2e8 geistesk: openrct2: init at 0.1.1
<Unode>
it's a start. Now if only that little chunk of documentation could be displayed
Myrl-saki has joined #nixos
ssmike has quit [(Remote host closed the connection)]
ssmike has joined #nixos
<infinisil>
How about nix-instantiate --eval -E 'let lib = import <nixpkgs/lib>; in builtins.attrNames lib' --json | jq '.[]' -r
FRidh has quit [(Quit: Konversation terminated!)]
FRidh has joined #nixos
itorres has joined #nixos
indefini has joined #nixos
WinterFox[m] has joined #nixos
hl has joined #nixos
bendlas has joined #nixos
colmarius has joined #nixos
<hyper_ch>
I'm trying to write a writeScriptBin nix expression. However it keeps complaining about vars - it just doesn't like bash parameter expansion.... I need to loop through an array: for i in ${var[@]}; do ... so it keeps complaining about ${var}
<hyper_ch>
how to fix that?
FRidh has quit [(Ping timeout: 240 seconds)]
efx has joined #nixos
<srhb>
You mean how do you escape ${foo} in strings?
<hyper_ch>
yeah
<hyper_ch>
you've formulated it much better
<srhb>
In a multiline string, precede it with ''
<hyper_ch>
it is
<srhb>
so '' ''${foo} ''
<hyper_ch>
for i in ''${var[@]}; do --- that seems to work... need to wait till I see the finished product :)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] edolstra pushed 1 new commit to master: https://git.io/vdHMM
<hyper_ch>
and there it is: for i in ${devices[@]}; do
<hyper_ch>
nice
frankqux1 has joined #nixos
freeman42x has quit [(Ping timeout: 264 seconds)]
<hyper_ch>
srhb: thx... I'm automating the re-adding of usb drives to zpools for mirroring/resilvering
ylwghst has joined #nixos
<ylwghst>
anyone here uses gdm+gnome3 here?
<ylwghst>
it looks im currently encoutering a new bug
frankqux has quit [(Ping timeout: 260 seconds)]
<ylwghst>
gdm+gnome3 alwayes crashes and gets stuck on blank screen with flashing _ after nixos-rebuild switch run from gnome in terminal
pxc has joined #nixos
Ivanych has quit [(Read error: Connection reset by peer)]
__shymega__ is now known as shymega
shymega has quit [(Changing host)]
shymega has joined #nixos
<jtojnar>
ylwghst: For me it returns to login screen after entering the password
Ivanych has joined #nixos
<ixxie>
ylwghst: this is a known issue
<jtojnar>
though not in vm
<ixxie>
but I see you found the thread on github
<ixxie>
:)
jushur has quit [(Ping timeout: 246 seconds)]
<ylwghst>
yup
<ixxie>
it looks like there is a fix committed
<jtojnar>
oh, I misread, mine is a different issue
<ylwghst>
it worked fine until I upgraded to latest packages
<ixxie>
it will just take some time to get to our channels
<ylwghst>
yup
freeman42x has joined #nixos
<ixxie>
you can do nixos-rebuild boot in the meantime
<ixxie>
annoying but effective
pxc has quit [(Ping timeout: 240 seconds)]
<ylwghst>
thx
euniarte has joined #nixos
<ylwghst>
what is actually the differnce between switch and boot?
<ylwghst>
ah i see
jushur has joined #nixos
jtojnar has quit [(Quit: jtojnar)]
<euniarte>
Hi, I'm using a cloned nixpkgs in a local package (because nixos' pyqt5 is missing the QtBluetooth module, I'm going to submit a PR or an issue if I got it working), but nix-build fails with a collision between two different versions of harfbuzz.
<euniarte>
Is there a possibility that pyqt5 -> qt5-base -> harfbuzz and qt5.qtcore -> harfbuzz resolve to different versions?
<euniarte>
But it also confirmed my assumption that qtconnectivity is broken :(
bkchr has quit [(Ping timeout: 248 seconds)]
cement has joined #nixos
glines has quit [(Quit: leaving)]
glines has joined #nixos
acowley_work has joined #nixos
glines has quit [(Client Quit)]
ylwghst has quit [(Ping timeout: 240 seconds)]
fendor has quit [(Ping timeout: 240 seconds)]
Myrl-saki has quit [(Read error: Connection reset by peer)]
jtojnar_ has quit [(Ping timeout: 240 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] globin opened pull request #30541: python.pkgs.cufflinks: init at 0.12.0 (master...cufflinks) https://git.io/vdHjx
NixOS_GitHub has left #nixos []
<fearlessKim[m]>
does nixos mess with clang to automatically add "-internal-isystem" ? It causes a pb (breaks a test regex) with a pkg I am using. I 've grepped a bit but come empty-handed :'(
zzamboni has joined #nixos
<jluttine>
how do i access stdenv in my nixos configuration?
<jluttine>
ixxie fearlessKim[m]: yep, thanks. that's what i tried but i get "error: attribute āstdenvā missing, at /etc/nixpkgs/lib/modules.nix:163:28". guess i'm doing something wrong elsewhere or something, gotta study
nh2 has joined #nixos
ssmike has quit [(Remote host closed the connection)]
jushur has quit [(Quit: The Borg joined forces with Skynet, Resistance is futile! Uploading has begun!)]
erasmas has joined #nixos
dustmote has quit [(Client Quit)]
Guest37279 is now known as TweyII
TweyII has quit [(Changing host)]
TweyII has joined #nixos
proteusguy has quit [(Ping timeout: 252 seconds)]
<TweyII>
So { "foo bar" = 3; } is a valid set expression, but { "foo bar" }: 3 is not a valid set pattern. How do I match against an attribute whose name isn't a valid identifier?
<jluttine>
hmm.. how can i find the package containing executable "run"? bash doesn't tell me directly when i try to run "run" because i'm using local nixpkgs git repo
erictapen has quit [(Read error: No route to host)]
jensens has joined #nixos
<michaelpj>
jluttine: try installing nix-index, or use "nix search" if you're using nixUnstable
<ldlework>
TweyII great question.
jtojnar_ has joined #nixos
zzamboni has quit [(Quit: Leaving.)]
<keta_suki>
Hi all! In my previous nixos config, I've set boot.initrd.network.ssh.hostRSAKey to a file path that existed when I ran nixos-rebuild. Now I've removed that line from my config, and removed the corresponding file but nixos-rebuild now fails as it tries to cp the file I removed, even though I removed the reference from the config file. Is there some configuration cache I need to clear to resolve this error?
<ldlework>
If I'm writing a NixOS module, what is the conventional way to support an option that represents some kind of file transformation for multiple files?
<keta_suki>
clever: Should I maybe try to rebuild without activating i.e. nixos-rebuilt boot ?
<ldlework>
I have a basic one that takes the path to an infile and spits out the content of the transformation, but what if I want to do that for multiple files?
<clever>
keta_suki: the bootloader generation is what is causing the problem
<clever>
keta_suki: oh, i think i see the problem
<ldlework>
I thought about taking a list of infiles, but how would I get specific files from the result so that different dependent modules can write out their transformed files to different places?
<clever>
keta_suki: for each generation you have built, it will fetch the secrets that generation wants
<clever>
keta_suki: so its trying to make sure your old rollback options still have the secrets they are asking for
<clever>
keta_suki: what do these 2 commands output: nix-store --query --roots /nix/store/sbr38724gcj1iznqj4p6y31wb55k6lb7-append-initrd-secrets
<clever>
keta_suki: ls -l /nix/var/nix/profiles/
<keta_suki>
Aha, makes sense. nix-store verifed without any problems btw
erictapen has quit [(Read error: Connection reset by peer)]
<jluttine>
oh, actually some other issue. not finding yarn, so nvm
<keta_suki>
clever: Thank you so much! So I'll create an empty file, generate the config and let nixos invoke a garbage collection to remove the older generations?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] globin opened pull request #30546: WIP: Update to Mariadb 10.2 (master...mariadb-10.2) https://git.io/vdQtJ
NixOS_GitHub has left #nixos []
<clever>
keta_suki: no need to create an empty file if you manualy delete those generations
<clever>
keta_suki: and if those generations are gone, you dont even need to GC
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus deleted vscode_1_17_2 at b269452: https://git.io/vdQtU
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 1 new commit to release-17.09: https://git.io/vdQt3
<NixOS_GitHub>
nixpkgs/release-17.09 0990021 Tim Steinbach: vscode: 1.17.1 -> 1.17.2...
NixOS_GitHub has left #nixos []
acowley_` has joined #nixos
acowley_` is now known as acowley_work`
katyucha has quit [(Ping timeout: 260 seconds)]
<wmertens[m]>
argh I have a server down due to brtfs issue and I can't get into initrd - hints?
<wmertens[m]>
tried init=/bin/sh but not there, tried init=..../bin/sh instead of .../init and that gives kernel panic
<clever>
wmertens[m]: take the existing init that ends in /nix/store/<hash>/init and change it to end in <hash>/sw/bin/sh
<TweyII>
wmertens[m]: It won't be under the same hash as init
<TweyII>
Panic is probably āinit doesn't existā
acowley_work has quit [(Ping timeout: 252 seconds)]
<TweyII>
And /bin/sh is on your root filesystem, which is presumably the one that btrfs just mangled
<wmertens[m]>
hmm trying
<wmertens[m]>
sw/bin/sh did the trick š
<clever>
wmertens[m]: there is also a rdinit= to change the init within the initrd, but the stage-1 script has other stuff thats better
<clever>
ah, then your rootfs is still good
<TweyII>
If the rootfs is good, why doesn't /bin/sh work?
<clever>
it may be partially corrupt
<TweyII>
Hmm
<wmertens[m]>
yes it's weird, df reports 7GB free but no writes work
<keta_suki>
clever: Manually deleting generations is usually done with nix-env --delete-generations ? --nix-env --list-generations appears to be empty before even removing generations. Should I point it to the system profile?
troglotit has quit [(Ping timeout: 248 seconds)]
<keta_suki>
Whoops I meant nix-env without the -- prefix
<clever>
keta_suki: you also need to give it -p, to tell it which profile to list the generations in
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
erictapen has joined #nixos
<keta_suki>
clever: Thanks! And that should be /nix/var/nix/profiles/system to refer to the most current profile?
<clever>
thats the profile for nixos-rebuild
<keta_suki>
or should I just follow the symlink in /root/.nix-profile ?
<clever>
thats for normal nix-env use, and wont have the nixos generations
acowley_work` has quit [(Ping timeout: 255 seconds)]
<keta_suki>
clever: Alright, I managed to resolve the issue thanks to your help as it rebuilt the generation without any problems. Thanks again :)
<clever>
yep
__Sander__ has quit [(Quit: Konversation terminated!)]
keta_suki has quit [(Remote host closed the connection)]
mortum5 has quit [(Ping timeout: 248 seconds)]
keta_suki has joined #nixos
simukis has quit [(Ping timeout: 264 seconds)]
<wmertens[m]>
clever: TweyII thanks guys, system is back, phew!
<wmertens[m]>
now I want a grub entry for a nice rescue shell...
<clever>
wmertens[m]: i have been thinking of a simple option for that, how big is your /boot?
jtojnar_ has quit [(Remote host closed the connection)]
jtojnar_ has joined #nixos
erictapen has quit [(Ping timeout: 248 seconds)]
<wmertens[m]>
clever: 43MB
Myrl-sak1 has joined #nixos
Myrl-saki has quit [(Killed (barjavel.freenode.net (Nickname regained by services)))]
Myrl-sak1 is now known as Myrl-saki
<clever>
wmertens[m]: eek, my idea for a rescue shell needs ~300mb
<wmertens[m]>
not actually on its own filesystem mind you
<wmertens[m]>
I would be happy with a busybox
<clever>
ah, what is the total size of the /boot partition?
keta_suki has quit [(Remote host closed the connection)]
<wmertens[m]>
i guess on AWS I would have attached the root disk as an extra disk to another VM
<wmertens[m]>
clever: good idea!
<clever>
it would probably be a custom initrd with dropbear that just kexec's the default kernel in the config
<clever>
and presents a gui over ssh for selecting a different kernel
<wmertens[m]>
so actually, having the nixos minimal installer image available as a boot option would work, come to think of it. Disk space is not an issue
<ldlework>
If I have a NixOS module that declares some options which take an infile and write out a transformed version of this file - how could I go about supporting multiple transformations? Basically, the module's Options are currently only able to be defined by one other module. But I'd like to be able to define these options as many times as I'd like...
suvash_away has quit [(Ping timeout: 240 seconds)]
<clever>
wmertens[m]: basicaly, you want to import a 2nd instance of <nixpkgs/nixos> (as i do in the netboot_server.nix above), and then refer to the kernel, initrd, and kernelParams like netboot.nix does
<clever>
wmertens[m]: that 2nd instance of nixos should also get a simple configuration.nix with things like boot.supportedFilesystems = [ "btrfs" ]; and a way to login
<wmertens[m]>
Ok, need to look at that on friday. now deadlines to meet. THanks!
hexamod has joined #nixos
<TweyII>
sphalerite: Well, that's just optional id
Wizek has joined #nixos
<TweyII>
sphalerite: Actually I wanted all of lib.optional anyway :)
suvash_away has joined #nixos
CodeWarrior has quit [(Ping timeout: 248 seconds)]
<sphalerite>
TweyII: so you want something that will take a boolean and return [true] if it's true and [] if it's falseā¦?
<sphalerite>
fair enough
CodeWarrior has joined #nixos
<TweyII>
sphalerite: Basically I have a bunch of boolean arguments like { foo ? true, bar ? false } and I want to produce a list like [ "foo" ], where "x" is in the list if x is true
<TweyII>
I've written that as builtins.concatLists [ (optional foo "foo") (optional bar "bar") ] but I'm open to suggestions
<sphalerite>
right
<ldlework>
š¤ nix-mode for emacs is kinda crap
<TweyII>
(also, those aren't the *only* arguments, so I can't loop through th e argument set)
<TweyII>
ldlework: Yeah :(
<TweyII>
ldlework: I updated it yesterday, and it's at least now better than fundamental-mode
<ldlework>
TweyII it's indentation logic is just broken
<TweyII>
ldlework: The old one used to never dedent ā so if you pasted (or yanked!) in a chunk of Nix, it would just march away across the page
<TweyII>
So I gave up on it and just used fundamental-mode
Neo-- has quit [(Ping timeout: 252 seconds)]
<TweyII>
But the new one mostly works, albeit with some rather odd indentation decisions sometimes
jtojnar has quit [(Remote host closed the connection)]
katyucha has quit [(Ping timeout: 248 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] rycee pushed 1 new commit to master: https://git.io/vdQGJ
<NixOS_GitHub>
nixpkgs/master 4233b83 Robert Helgesson: perl-CryptX: 0.050 -> 0.054
NixOS_GitHub has left #nixos []
sigmundv__ has quit [(Ping timeout: 260 seconds)]
keta_suki has quit [(Remote host closed the connection)]
keta_suki has joined #nixos
notdaniel has quit [(Quit: Leaving)]
Neo-- has joined #nixos
fusion809 has joined #nixos
mma has joined #nixos
ryantrinkle has joined #nixos
proteusguy has quit [(Ping timeout: 260 seconds)]
LysergicDreams has quit [(Quit: leaving)]
glenn has joined #nixos
keta_suki has quit [(Ping timeout: 248 seconds)]
alanz has joined #nixos
johnw has quit [(Ping timeout: 255 seconds)]
glenn is now known as scribbler
fusion809 has quit [(Client Quit)]
<Isorkin>
Please help. New kernel - not work nftables rules - reject. nft add rule inet filter input tcp dport \{ 22 \} ip saddr \@f2b-SSH reject - error. Rules drop - work. diff config kernel - https://pastebin.com/f1FQM0RV kernel 4.13.5 - nft worked, 4.13.7 - not work(
<chakra>
how would i test if the URL itself supports the headers Range (for curl -r) and/or modified since and/or last modified however i do not know if it is possible for that to change depending on the website url in accordance to the base url kinda like directory permissions for sub folders in linux, i want to test for those to reduce bandwidth and download times as much as possible for example: https://www.url1.com, https://www.url1.
<toppler>
I'm new to virtualisation stuff. I've got a little VM hosted by bytemark, but I'm trying to run VMs inside it. Is this likely to fail?
<toppler>
Right now, I'm trying virtualbox and it's stalling on "waiting for IP address."
<disasm>
inception!
<toppler>
disasm: Indeed!
<fusion809>
Hey folks is it relevant that on my Arch Linux system where I've got Nix installed that $HOME/.nix-profile for my user is linked to /nix/var/nix/profiles/per-user/fusion809/profile which does not exist?
<fusion809>
Should that file exist?
aloiscochard has quit [(Quit: Connection closed for inactivity)]
<toppler>
disasm: Maybe it's not the right approach. I was just hoping to keep my various servers sandboxed from one another.
goibhniu has quit [(Ping timeout: 264 seconds)]
pxc has joined #nixos
<ArdaXi[m]>
It's not, no. You probably want to look at containers, instead
<disasm>
toppler: how sandboxed do you you require? nixos-container works well :)
<eacameron>
How do tell nix to build haskell code with -j10
<toppler>
Well, it's for fairly obsessive security concerns.
<gchristensen>
try nix-build --cores 10
<toppler>
Maybe I just need to drop the paranoia.
<eacameron>
gchristensen: Does that change how GHC is invoked?
<fusion809>
ArdaXi[m]: you talking to me? Nix seems to work though, the programs I've installed with it are running A OK. Just seems odd.
<clever>
eacameron: i believe --cores is the one that gets routed to NIX_BUILD_CORES
<clever>
eacameron: also look at line 99-102 in the same file
<eacameron>
clever: You are a wiz at GitHub links.
<clever>
eacameron: ive memorized an unreasonable percentage of the nixpkgs codebase ...
<eacameron>
clever: Where unreasonable percentage is 0.1%
keta_suki has joined #nixos
<eacameron>
And up
jtojnar_ has joined #nixos
jtojnar has quit [(Ping timeout: 240 seconds)]
<jtojnar_>
1 minute seems pretty short. Is it possible that the frequent renewals are causing connection interruptions?
<clever>
jtojnar_: if the dhcp server is crap, and changes the IP you are given, yes
<jtojnar_>
clever: the IP address remains the same
TweyII has quit [(Ping timeout: 248 seconds)]
<eacameron>
What's a good tool to check the state of my harddrive? My computer seems to lock up randomly (but not the mouse or hover effects, just the actual task)
<clever>
jtojnar_: the NAT could also be the problem, is it only happening with idle connections?
<eacameron>
I'm almost certain it's the drive.
<clever>
eacameron: smartctl -a
ShalokShalom_ has quit [(Ping timeout: 248 seconds)]
ShalokShalom has joined #nixos
<eacameron>
clever: How do I install that?
<clever>
eacameron: nix-env -iA nixos.smartctl
<jtojnar_>
clever: no, no new web pages load in the web browser
<NixOS_GitHub>
[nixpkgs] CarlOlson opened pull request #30548: tensorflow: dont include cudatoolkit unless cudaSupport is used (master...tensorflow-without-cudatoolkit) https://git.io/vdQW0
NixOS_GitHub has left #nixos []
<ldlework>
or is it
<eacameron>
disasm: clever: Ah smartmontools is it
<disasm>
ldlework: does hm have --show-trace?
jtojnar has quit [(Remote host closed the connection)]
<jtojnar_>
clever: The interruption seems to coincide with "dhcp4-change' [enp3s0]: start running ordered scripts..." in the journal
jtojnar has quit [(Ping timeout: 240 seconds)]
<argo>
Hi, can someone replicate this error using plasma 5: When I try to open the configuration dialog for the digital clock widget in the panel (right click on clock -> Digital clock settings) the opening window body is transparent and empty, system log throws this error "xsession[1699]: file:///nix/store/hm4czv681dnk9ngfmw5g67b9jnlknyyk-plasma-desktop-5.10.5/share/plasma/shells/org.kde.plasma.desktop/contents/configuration/AppletConfiguration.qml:20:1: plugin
<argo>
cannot be loaded for module "QtQuick.Dialogs": Namespace 'QtQuick.Dialogs' has already been used for type registration"
fusion809 has quit [(Quit: leaving)]
proteusguy has joined #nixos
ma27 has joined #nixos
<argo>
I don't know if something is wrong with my configuration (nixos 17.09), or if it is a packaging problem or upstream error
<jtojnar_>
but the scripts are run after every renewal and the connection is not interrupted every time
ylwghst has joined #nixos
jgertm has joined #nixos
jtojnar_ has quit [(Remote host closed the connection)]
jtojnar_ has joined #nixos
ThatDocsLady has quit [(Remote host closed the connection)]
ThatDocsLady has joined #nixos
MichaelRaskin has joined #nixos
riclima has quit [(Quit: riclima)]
<clever>
wmertens[m]: got the rescue shell working
riclima has joined #nixos
riclima has quit [(Client Quit)]
<disasm>
jtojnar_: maybe wireshark it?
<jtojnar_>
disasm: I cannot even ping google's IP address, not sure what I would be looking for
<taaperotassu>
How do I set neovim as my default editor?
katyucha has joined #nixos
<wmertens[m]>
clever: does that boot locally or does it try a netboot?
<clever>
wmertens[m]: locally
<clever>
wmertens[m]: i just grabbed the netboot kernel&initrd, and jammed them into /boot, with a menuentry
<Unode>
Hi all, I've created a nixos service module in nixos/modules/services/backup/burp.nix . When I try to run nixos-rebuild -I /path/to/my/nixpkgs (folder containing nixos/) switch it tells me services.burp defined in configuration.nix doesn't exist. Do I need another file somewhere?
katyucha has quit [(Client Quit)]
<wmertens[m]>
taaperotassu: `export EDITOR=neovim` in your profile maybe
<ldlework>
Anyone suffering with nix-mode in emacs? I wish I new elisp..
<ldlework>
knew too
ma27 has quit [(Ping timeout: 258 seconds)]
phreedom has joined #nixos
<Unode>
anyone has some hints on the burp question above?
<clever>
Unode: you need -I nixpkgs=/path/to/nixpkgs
<clever>
Unode: and you also need to add it to nixos/modules-list.nix i believe
<Unode>
ok the first one I did the second not
ssmike has quit [(Ping timeout: 255 seconds)]
<Unode>
lemme try that.
ylwghst has quit [(Quit: Lost terminal)]
ylwghst has joined #nixos
<Unode>
clever: that helped, thanks. Now some issues to debug.
<taaperotassu>
Thanks clever
zzamboni has quit [(Quit: Leaving.)]
<ldlework>
In a Nix module is there anyway to extend the Nix path so you can make git fetched expressions importable?
tmaekawa has joined #nixos
<taaperotassu>
Any good ways to make capslock into esc key? Normally I use xmodmap.. but it seems that I didnt manage to get it work. I did install xorg.xmodmap
<clever>
taaperotassu: there is a nixos option for that
<yrashk>
and the only way I can start a VM is `VBoxManage startvm <name> --type sdl`
<clever>
yrashk: what does "type VirtualBox" return?
<yrashk>
VirtualBox is /run/wrappers/bin/VirtualBox
<ldlework>
clever I'm trying to figure out how to easily import pluginsĀ for Home Manager and I want to have a top-level expression import them and make them available to further imported modules
<ldlework>
I don't think patching stuff is a good solution
mortum5 has joined #nixos
<clever>
yrashk: it looks like virtualbox is somehow mixing qt 5.9.1 and 5.9.1, youll need to investigate the output of "nix-store -qR" ran against the virtualbox binary, and some other things
<clever>
ldlework: i havent read the source for home manager yet, so not sure what else you can do to make it work that way
<Unode>
are systemd units not allowed to spawn children?
<eacameron>
clever: How do I get results from that smartctl command?
mma has quit [()]
<clever>
eacameron: smartctl -a will check everything and print it all to stdout
ryantrinkle has joined #nixos
<eacameron>
clever: Hmm, it says passed
<clever>
ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE
<Unode>
I'm starting a process that then launched another command. This second command exits with 0 but output is not visible. It doesn't seem like the script had any effect. If I run the script manually, everything works as expected.
<clever>
eacameron: other numbers like this can also be of use
<cransom>
fwiw, systemd 235 does some fancy stuff with being able to allocate it's own temporary uid/gids per service which would be nice.
<infinisil>
Ohh..
<clever>
cransom: i dont think that plays nicely with having state that must be owned by the service
<clever>
but its great if you just want to delete the user and all traces of the service when it stops
k2s has quit [(Client Quit)]
<infinisil>
It actually doesn't really matter what gid i use
<infinisil>
since no files are preserved
<infinisil>
In my case
<tnks>
I kind of wish metadata (most-specifically licensing) were stored in /nix/store.
<tnks>
might that come with nixUnstable? Does anyone talk about that?
<tnks>
It's kind of like Debian's policy about what goes in /usr/share/doc
<clever>
tnks: packages could still do that in $out/share/doc/
betaboon has joined #nixos
<tnks>
clever: yeah, in that regard, it's just infrastructure code and factoring it into Nixpkgs, yeah?
<gchristensen>
$pkg.license could always be a file in the store of the license's text, with some simple stdenv work
<clever>
tnks: yeah
zzamboni has joined #nixos
<disasm>
infinisil: gid and uid, I've been told not to assign statically in PR's I've submitted
<tnks>
Yeah, it sounds like it could ostensibly be a small tweak if done in the right place.
<betaboon>
hello #nixos, is virtualbox-headless broken in 17.09 ?
<tnks>
but with nixpkgs... I've learned there's devils in details.
<infinisil>
disasm: I'll keep that in mind
Isorkin has quit [(Ping timeout: 240 seconds)]
<Unode>
I'm trying to create a service for burp. https://gist.github.com/567574f6119abab4a6e200c4cc64751e . So far I've managed to do most of the configuration however the burp.service unit fails to execute burp_ca and the required certificates aren't generated. If I run the script as root everything works as expected. I'm assuming the problem is burp_ca calls hostname and openssl and while running from systemd these
<Unode>
are not in the PATH and aren't found. Is there any way to configure the service such that all required binaries are found?
<symphorien>
ad those pacakges to systemd.services.<name>.path
<clever>
infinisil: if its usb, just unplug and replug the device, after the switch has finished
<infinisil>
Yeah, but I'm trying to avoid that :P
<yrashk>
woffs: I am trying to resolve this Qt issue, however, it seems like there is no qtwebkit 5.9.2, only 5.9.1; what should we do?
<infinisil>
Would be neat not having to do that
<yrashk>
woffs: ./maintainers/scripts/fetch-kde-qt.sh simply drops qtwebkit from srcs.nix altogether
<infinisil>
Probably it's possible with udevadm trigger, but I doubt that'll get accepted in a PR
Ivanych has quit [(Ping timeout: 248 seconds)]
<infinisil>
Although..
<ldlework>
Is there really no way to fetch from the web without providing a md5, etc?
<Ralith>
there are way to download dynamic data, they just cannot be used in nixpkgs proper for obvious reasons
<ldlework>
Like what
hexamod has quit [(Quit: My MacBook has gone to sleep. ZZZzzzā¦)]
scribbler has joined #nixos
<LnL>
builtins.fetchurl "..."
<ldlework>
that requires an md5 right?
<ldlework>
oh I guess not
<ldlework>
nice
hexamod has joined #nixos
ma27 has quit [(Ping timeout: 248 seconds)]
<Ralith>
it doesn't, but if it did require a hash it would be a sha256, not md5; md5 is broken
<ldlework>
oh neither does fetchFromGithub
<Ralith>
cstrahan: seeing as your PR does the exact thing I want--makes explicitly supplied options work reliably--I am indeed pleased with it as it stands, yeah
<ldlework>
If a function takes `...` is there a way to pass those args to another function?
ma27 has joined #nixos
<LnL>
f = args@{ foo, ... }: args.bar
scribbler has quit [(Ping timeout: 248 seconds)]
<woffs>
yrashk: you could fetch the old 5.9.1 version
scribbler has joined #nixos
<ldlework>
LnL anyway to pass them all?
<yrashk>
woffs: well, that's what current srcs.nix does
Neo-- has joined #nixos
<infinisil>
So, now the usbmuxd module is truly magical
<ldlework>
LnL alternatively, would "foo { inherit args; }" work?
<ldlework>
To pass all of the outer function's args?
<LnL>
ldlework: args here is an attrset with all of the arguments so you can just use it to call another function
hexamod has quit [(Quit: My MacBook has gone to sleep. ZZZzzzā¦)]
<ldlework>
oh nice
hexamod has joined #nixos
<LnL>
args: args.foo is the same as { foo, ... }: foo
<Unode>
Can I have a file that is generated by nix/nixos using writeFile but is not kept in the store? Seems like burp tries to rewrite the .conf file once it receives some parameters from the server (1st connection)
zzamboni has quit [(Ping timeout: 248 seconds)]
reinzelmann has quit [(Ping timeout: 255 seconds)]
<symphorien>
create it in execStartPre
zzamboni has joined #nixos
<symphorien>
possibly by copying it from a nix-generated template
riclima has joined #nixos
<symphorien>
but if you let burp managing its own conf file, how can you update it when nixos options are changed ?
<Guest19986>
I wanted to check in real quick about the status of ruby packages. It sounds like the new builder got released as part of 17.09? Is that so? There's a PR against bundix lingering, as well.
Guest19986 is now known as Judson
<Unode>
symphorien: I know. But don't know how to solve this. Needs a patch upstream most likely.
<yrashk>
woffs: I am not sure I fully understand. So far, I figured out a way to resolve the issue locally. In my home-manager setup, I also installed dropbox from unstable, which (I think) brought qt5 5.9.2 and (5.9.1 qtwebkit) from unstable. The moment I reverted to stable dropbox, VirtualBox and wpa_gui work just fine. So, now I am not sure what's the best way to proceed with unstable -- should we revert it back to 5.9.1 for
<yrashk>
the time being, remove qtwebkit until matching version is released, do nothing, etc :)
<MichaelRaskin>
shlevy: is it official now, Nix as a general-pure-computational-purpose functional programming language?
<Unode>
yup pretty sure now. It's rewriting what ends up being the same.
<shlevy>
michalrus: Yep!
<woffs>
yrashk: maybe report an issue or wait a week ;)
<shlevy>
MichaelRaskin: Yep!
<shlevy>
No need to write code in anyother language now
<yrashk>
woffs: haha, ok :) I am still a newbie with nix, so I am trying to figure the best course of action
<Unode>
when doing pull requests against nixpkgs, Should unstable be the target?
<woffs>
yrashk: me too
<woffs>
Unode: even master, afaik
<MichaelRaskin>
Yeah, the creator's attempts to pretend that Nix is domain-specific get heavier blows each timeā¦
reinzelmann has joined #nixos
Oida has quit [(Ping timeout: 248 seconds)]
<rnhmjoj[m]>
I have been trying to build firefox 56 with a few patches but I get: output ā/nix/store/[...]-firefox-56.0.1ā is not allowed to refer to path ā/nix/store/[...]-gcc-wrapper-6.4.0ā. If I remove the patches the error goes away. how come?
<shlevy>
The fact that there's not yet any easy way to, say, manipulate derivations in any other language (besides guile) and have them fit into the ecosystem, or even *not* fit into the ecosystem but have it work decently well, means by necessity we're going to squeeze more and more absurdity out of the nix expression language :D
<clever>
rnhmjoj[m]: use grep -r to find the gcc-wrapper path within the firefox output, after it fails
<tnks>
huh... how do I commit fixes for the autogenerated pkgs/development/haskell-modules/hackage-packages.nix file?
<sphalerite>
Unode: neither of those apply to modules
<infinisil>
gchristensen: I'd love it if you could test the new version once more
<noam>
Hi, there are some broken links in the Disnix documentation page: "https://nixos.org/disnix/docs.html". The links "Disnix manual" and "DisnixOS" do not work.
chessai has quit [(Remote host closed the connection)]
<Unode>
sphalerite: so what options do I have left?
<sphalerite>
Unode: either keep a local checkout that you rebase on upstream to update (until your PR is merged, if applicable) or keep the module outside the nixpkgs tree and add it to your imports yourself (if it's a separate module and not a change to an existing one)
<tnks>
clever: I did not figure out your word puzzle... but I'll take the answer.
<clever>
tnks: "found it, one min" with my hand being offset a column
<infinisil>
lol
<Unode>
sphalerite: it's a new module but it touches a few internal things, for instance the uids file mentioned above that handles user/group id assignments.
<infinisil>
We now know that clever uses Qwerty
<gchristensen>
it is a shame more americans / canadians aren't going to be at nixcon, I have at least one references ~0% of the audience will get
<tnks>
clever: cool, well sounds like you're answer conforms with pbogdan's. I'll submit the patch.
<MichaelRaskin>
Are you sure you are not underestimating global reach of memetic hazards?
<MichaelRaskin>
I mean, if it reached Canada, it's in the wild.
<samueldr>
or overestimating?
<gchristensen>
I dunno, what do you know about Wendy's commercials from the 1980's
<infinisil>
gchristensen: You american?
<gchristensen>
infinisil: I am
cement has quit [(Ping timeout: 240 seconds)]
scribbler has joined #nixos
<MichaelRaskin>
I won't be at NixCon anyway, but I would have troubles recognizing _any_ commercials from 1980's
<samueldr>
too young probably for 80s wendy's
<MichaelRaskin>
Yup, a temporal problem more than spacial.
<gchristensen>
at least one of those, certainly
<samueldr>
though, canada is weird for those kind of ads
<samueldr>
could have been too local
<ldlework>
aww, fetchFromGithub DOES require rev
<infinisil>
ldlework: Of course it does, otherwise it wouldn't know what to fetch
<ldlework>
could use master by default
<infinisil>
master isn't fixed
<infinisil>
usually you don't want revisions to change in nix
<rnhmjoj[m]>
clever: the path seems appears in a file called env-vars, the config.log e some more automake related files during the build. I don't know what is left after the build is finished yet. I have restarted the build with --keep-failed but it takes a while
<Unode>
sphalerite: I'll look into keeping a clone of the channel. Likely the easiest option.
<clever>
rnhmjoj[m]: the problem is in the storepath, not the tmp directory
<clever>
rnhmjoj[m]: so you need to grep for /nix/store/[...]-gcc-wrapper-6.4.0 within /nix/store/[...]-firefox-56.0.1
<Judson>
I wanted to check in real quick about the status of ruby packages. It sounds like the new builder got released as part of 17.09? Is that so? There's a PR against bundix lingering, as well. (I'm still using my dev copy of it to build gemset.nix's)
<infinisil>
ldlework: And if you used master, it won't ever update it anyways, the first time you build it, that master is used
<infinisil>
which wouldn't work on another machine, because the other machine would fetch a different master and the hash would mismatch
<infinisil>
(assuming it's built at a different time with different masters)
<infinisil>
master: no good
riclima has quit [(Quit: riclima)]
keta_suki has quit [(Quit: WeeChat 1.9.1)]
erictapen has joined #nixos
Guest45531 has quit [(Ping timeout: 252 seconds)]
riclima has joined #nixos
ubuntu has joined #nixos
ubuntu is now known as Guest30360
MercurialAlchemi has quit [(Ping timeout: 240 seconds)]
scribbler has quit [(Ping timeout: 248 seconds)]
scribbler has joined #nixos
<rnhmjoj[m]>
clever: is that path created when the build fails? I don't because the last attempt was probably garbage-collected
<ldlework>
LjL, it seems like if the function I'm calling requires `args` but the outer function is defined as `args@{ ... }: ` then I cannot simply pass `args` to the inner function
<rnhmjoj[m]>
*don't know
<infinisil>
ldlework: I don't know the context or who LjL is, but args is just a name, you can put anything there
<Profpatsch>
How would you handle secrets given to a service files
<Profpatsch>
s/files/file?/
<ldlework>
yes, but I'm saying that calling a function that takes `pkgs` for example can't be called like `some_func args`
<Profpatsch>
Pass a file with EnvironmentFile, yes, but how to construct that file and where to put it?
<ldlework>
I meant LnL
<ldlework>
lol
<infinisil>
ldlework: You can always just use an attrset: some_func { pkgs = ... }
<ldlework>
Yeah but thats what I was hoping to avoid
<ldlework>
fetching modules from github is already hella verbose
<infinisil>
what are you trying to do that makes you do this?
<infinisil>
the some_func thing
<ldlework>
Just as I said, load a module from github
<ldlework>
My toplevel HM expression loads an expression from github
ma27 has quit [(Ping timeout: 260 seconds)]
<clever>
rnhmjoj[m]: after it fails again, grep its output
<ldlework>
since it is a module, it needs to be called with the pkgs, config, lib, etc
<clever>
rnhmjoj[m]: nix keeps the $out of failed builds, but doesnt flag them as valid, so they are the first target for GC
<infinisil>
ldlework: Umm, no?
<sphalerite>
nix-adt looks really neat
<sphalerite>
!m shlevy
<[0__0]>
You're doing good work, shlevy!
<ldlework>
infinisil yeah, because the modules depend on those names
<infinisil>
Modules can either be just an attrset, or a function with pkgs, config, lib, etc. as arguments
<rnhmjoj[m]>
clever: ok, i will. thank you.
<ldlework>
infinisil yeah an attrset if it only defines options, I know.
<ldlework>
the code of the plugin module depends on the pkgs and config attrs, etc
ertes-w has quit [(Read error: Connection reset by peer)]
<ldlework>
so they are forumlated as functions which take them
<ldlework>
once you fetch from github, you have to call the imported function with those names
ertes-w has joined #nixos
<ldlework>
I am not passing them directly to imports = [] because I actually need some non-module functions from them
pie_ has quit [(Ping timeout: 255 seconds)]
<ldlework>
IE, they also act is library code
kuznero has joined #nixos
<infinisil>
Oh, well then it's not a module, but a function that returns one
<kuznero>
Hi All!
<ldlework>
so the plugin module returns an attr set like { func1 = ..., func2=..., module=...}
<ldlework>
And I pass imported_thing.module to imports
<LnL>
ldlework: hmm, fetching expressions like that at eval time will make it pretty slow and you loose reproducability
<kuznero>
Are there any examples of a derivation with several fetchFrom calls?
<infinisil>
LnL: Even with fetchFromGitHub with rev and hash?
<LnL>
ldlework: do you have an example of what you are fetching
<LnL>
infinisil: no only builtins.fetchurl happens at eval time
<infinisil>
LnL: Ah right, didn't see you say eval time
<ldlework>
When you import this plugin, you can use functions defined in the returned attrset, and/or import the module attribute to load its declared Options
<ldlework>
s/import this plugin/fetch this plugin
ma27 has joined #nixos
<infinisil>
ldlework: Hmm, well with an overlay attribute this would be much nicer
<infinisil>
overlays option*
<infinisil>
But home-manager currently doesnt' have that
<ldlework>
explain
<ldlework>
heh
<infinisil>
then you could make a lib overlay and have all your functions available from the lib in all modules in the argument list
<LnL>
how do you intend to use the build attribute?
Guest30360 has quit [(Ping timeout: 248 seconds)]
scribbler has quit [(Remote host closed the connection)]
cwre has quit [(Read error: Connection reset by peer)]
<kuznero>
LnL: are you asking me? sorry the confusion
<infinisil>
Ohh, also, you could define a readOnly, internal option, which is like having a variable as an option ldlework
<infinisil>
kuznero: (he asked ldlework)
<kuznero>
sorry
<LnL>
also it doesn't have the correct format for a module so you'd have to wrap it in a function if you want to import it
<srhb>
Hrm. My nodes in my new nixops kubernetes deployment ended up with a taint: unschedulable=true:NoSchedule
<srhb>
Any idea how this happens?
<ldlework>
LnL, the module attribute is a valid importable module
<kuznero>
LnL: I don't think that will help. I am packaging something that in turn tries to download intermediate bits itself. What I am trying to do basically now is to merge 2 sources of sources ;) into one and then substituteInPlace --replace to avoid trying to download already downloaded/merged packages.
ma27 has joined #nixos
ubuntu___ has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] danharaj opened pull request #30552: pysc2: init at 1.2 (master...pysc2-init) https://git.io/vdQHd
NixOS_GitHub has left #nixos []
erictapen has quit [(Ping timeout: 240 seconds)]
<kuznero>
Is that an ok thing to do, LnL?
ma27 has quit [(Ping timeout: 258 seconds)]
jensens has joined #nixos
<LnL>
not sure I understand, but if you move stuff to the expected location and patch the downloads during the build it might work
<NixOS_GitHub>
[nixpkgs] unode opened pull request #30553: WIP: Add support for burp client/server services (master...master) https://git.io/vdQ7G
NixOS_GitHub has left #nixos []
zraexy has joined #nixos
simpson has joined #nixos
<simpson>
Hi! I want to use a new LyX, because my old LyX is old and I can't pull 2016 texlive tarballs any longer. So I checked out the master branch of nixpkgs, but LyX won't start, and tells me this instead; any ideas?
<simpson>
`This application failed to start because it could not find or load the Qt platform plugin "xcb"`
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ma27 opened pull request #30554: shmig: init at 2017-07-24 (master...init/shmig-2017-07-24) https://git.io/vdQ7D
NixOS_GitHub has left #nixos []
<maurer>
simpson: I saw "19:31 <woffs> all qt5-webkit things in unstable are broken because of 5.9.2 vs. 5.9.1 version clash, like https://github.com/NixOS/nixpkgs/issues/28214 but n+1" earlier today
<maurer>
I don't know anything about it, but it might point you in the right direction
<simpson>
maurer: Interesting. I wonder if the stable 17.09 branch is any better.
<Unode>
disasm: https://github.com/NixOS/nixpkgs/pull/30553 the client part works for me despite all the gotchas. I haven't tested the server. Mine is running on a non-nixos machine.
efx has joined #nixos
<simpson>
euniarte, maurer: Thanks, I'm trying out a patch now to see whether this fixes it.
euniarte has quit [(Remote host closed the connection)]
<clever>
nix-repl> let foo = 5; in rec { inherit foo; }
<clever>
{ foo = 5; }
cwre has quit [(Client Quit)]
<clever>
inherit doesnt get tripped up, and without the ref, foo=foo; does what you would expect in nix
<clever>
rec*
<LnL>
ldlework: you could fetch the sources like you did before and use "${hm-org-build}" and "${hm-org-build}/release.nix" instead
<CrazedProgrammer>
is there any way to make nixos-install fetch everything into /mnt/nix/store instead od /nix/store and then copying it over? 9gb of applications doesn't fit in my ram
<ldlework>
what is release.nix in this case?
<LnL>
a separate file with the build expression
cwre has joined #nixos
<euniarte>
CrazedProgrammer: Maybe mount /mnt/nix to /nix?
<euniarte>
Everything else would involve rebuilding everything if I have understood nix correctly
<clever>
euniarte, CrazedProgrammer: thats a bug in the current nixos-install
<simpson>
maurer, euniarte: Hm, adding `qtx11extras` to buildInputs didn't fix it. Trying adding `qtbase` to propagatedBuildInputs. Any other ideas?
<clever>
CrazedProgrammer: currently, you need to add some swap, or start with a smaller installation
ariutta has joined #nixos
<CrazedProgrammer>
all right, not that much of a hassle
<euniarte>
simpson: Sounds to me like the issure I linked before. I don't have any more ideas, sry
<clever>
voiceftp: i see why, that .emacs-wrapped is just a symlink
S0rin has joined #nixos
<clever>
voiceftp: and i found a bug
<clever>
voiceftp: "emacs" is a bash script, that runs .emacs-wrapped, which is a symlink to emacs-25.3, which itself is an identical wrapper, over .emacs-25.3-wrapped
<clever>
so its being wrapped with 2 bash scripts
k2s has joined #nixos
<voiceftp>
clever, thanks for explanation
<voiceftp>
that type of bug does not cause any inconvenience, I think
k2s has quit [(Client Quit)]
<ariutta>
I've got a question about NixOS and the www-data user. My OS is currently Ubuntu (only option allowed ATM), and I used nix to install the command line program "jq". As a normal user, I can run jq just fine, but I don't know how to let the "www-data" user call jq (www-data is the user for PHP, and I'd like to call jq from PHP via shell_exec). The home directory for www-data is /var/www, and the permissions aren't the same as for a norma
c0ffee152 has joined #nixos
<clever>
ariutta: run "type jq" to see where jq is coming from $PATH
<clever>
ariutta: you may need to get the path to be set better
<ariutta>
Path is currently "/home/ariutta/.nix-profile/bin/jq"
Neo-- has quit [(Ping timeout: 264 seconds)]
goibhniu has quit [(Ping timeout: 240 seconds)]
<ariutta>
Should I use "sudo visudo" and add the line "www-data ALL=NOPASSWD: /home/ariutta/.nix-profile/bin/jq"
<clever>
ariutta: no
<clever>
ariutta: what does "ls -lh /home/ariutta/.nix-profile" say?
<gchristensen>
(you won't want to paste that output)
<pxc>
looks good
<gchristensen>
:) hopefully that makes it easier to get answers we need during support.
<gchristensen>
"what channel are you on?" "unstable" ".. yeah but which one?"
<pxc>
yeah, it's a great idea
efx[m] has joined #nixos
<pxc>
a simple thing but a big help for new users
<gchristensen>
yeah
<gchristensen>
and reduce the burden on people giving support
<gchristensen>
win-win
<pxc>
haha just noticed that it's even markdown-formatted for pasting into github
<pxc>
nice
<gchristensen>
:)
<pxc>
that's the -m ?
<gchristensen>
markdown
<gchristensen>
--help
<rnhmjoj[m]>
clever: I found the culprit: the gcc full path is in an html page listing the compile flags and stuff, I think it's about:buildconfig page. I still don't understand what this has to do with the patches though
<efx[m]>
I tried the nix-info command, but got the following error: https://dpaste.de/dYzK
<pxc>
efx[m]: update your nixpkgs?
<efx[m]>
roger
<gchristensen>
ah yeah your nixpkgs is too old
<elvishjerricco>
whatever happened to `--sign` being removed from `nix-copy-closure`? It's still there in NixOS 17.09
<efx[m]>
hmm, I should just need to do nix-channel --update?
<gchristensen>
ahh, you'll need to run the channel update as root: sudo su -, then nix-channel --update efx[m] (there are a few issues with multi-user nix and macos right now, so the sudo su - being a separate command is necessary :( )
<gchristensen>
this is because nix-shell looks at <nixpkgs> which is provided by the root's channel
<rnhmjoj[m]>
clever: I just did that. thanks for the help
<efx[m]>
ah, thanks gchristensen . unfortunately I need to ask for admin access so will have to try that later.
jensens has quit [(Ping timeout: 260 seconds)]
catch22_ has joined #nixos
catch22 has quit [(Ping timeout: 240 seconds)]
riclima has joined #nixos
markus1199 has joined #nixos
catch22_ has quit [(Ping timeout: 258 seconds)]
iqubic has joined #nixos
markus1189 has quit [(Ping timeout: 248 seconds)]
<gchristensen>
tbttttthnnthh
<gchristensen>
oops, heh
<gchristensen>
efx[m]: oh then you can delete your nixpkgs-latest channel and make a new one, just name it nixpkgs
thc202 has quit [(Ping timeout: 248 seconds)]
<efx[m]>
gchristensen: huh, I tried that and updated but still got the error: https://dpaste.de/qOiq. I did just do nix-env -i and get the expected output! https://dpaste.de/WWe3
<dhess>
Hi, question about NixOps. I'm having problems deploying to DigitalOcean. On the initial deploy, while waiting for the "infect" stage, I get this:
<dhess>
do.host..........> waiting for SSH.........error: global name 'ConnectionError' is not defined