<romanofskiWork>
kalbasit[m]: does nix-build --arg withExtraPackages "[golint, gotools]" yourelease.nix work? (I'm a n00b myself so take the advice with a grain of salt)
detran has joined #nixos
kai_w has joined #nixos
kai_w_ has quit [Ping timeout: 240 seconds]
<yl[m]>
this worked for me in nix-shell: `nix-shell -I nixpkgs=. -p 'with import <nixpkgs> {}; neovim.override { withExtraPackages = [ gotools ]; }'`
<{^_^}>
[nixpkgs] @kalbasit opened pull request #49498 → neovim: take withExtraPackages and add them to the PATH → https://git.io/fxFn9
<{^_^}>
[nixpkgs] @romildo opened pull request #49499 → deepin: add update script and update some packages → https://git.io/fxFnj
<__red__>
ah, I deleted the ; as it was a syntax error
<__red__>
and I didn't realize the "with" was part of the command ;-)
<__red__>
trying again
<__red__>
take 73
<__red__>
;-)
<__red__>
holy crap - it built!
<__red__>
well - the first one - the second one takes **forever**
Peetz0r has quit [Ping timeout: 252 seconds]
<tnks>
I've avoided buildStackProject... but I'm playing around with it... I'm a bit confused about what I'm seeing using it straight from a nix-build. Are Stack-build dependencies being cached by it in /nix/store?
<__red__>
Wow - this is amazing - thanks !
<elvishjerricco>
tnks: I can't recommend actually building `buildStackProject`. I think it's only useful for defining a `nix-shell` that `stack` can use outside of Nix. But actually trying to use it for builds is really bad because it does *no* caching of the ~/.stack files, so *everything* is rebuilt every time.
lopsided98 has quit [Ping timeout: 252 seconds]
acarrico has quit [Ping timeout: 272 seconds]
<tnks>
elvishjerricco: perfect... that is in-line with what I was thinking.
<tnks>
and also... I think the stuff I'm seeing going into /nix/store is the build of Stack itself!
<elvishjerricco>
tnks: I think it'd be possible to cache the ~/.stack files, but it'd take some doing.
<__red__>
now to work out how to get these cached files published
<tnks>
which was confusing me... I don't think it's even gotten to building my project.
lopsided98 has joined #nixos
<tnks>
elvishjerricco: yeah, perfect. I was just confusing myself.
<elvishjerricco>
tnks: Yea, buildStackProject puts the `stack` from nixpkgs onto the PATH for the build (and therefore the shell)
Peetz0r has joined #nixos
mkoenig has quit [Ping timeout: 244 seconds]
romanofskiWork has joined #nixos
<tnks>
elvishjerricco: I'm going to stop this build then... It's uninteresting to me.
<__red__>
cachix time(!)
mkoenig has joined #nixos
<__red__>
oh funny - looks like I didn't have to use hydra after all
<__red__>
heh
zolk3ri has quit [Remote host closed the connection]
<__red__>
ah well - I'm glad I did!
<tnks>
it seems cabal2nix is broken for ghc802... I tried to repair it... but no luck. Maybe ghc802 is a lost cause in Nixpkgs now?
<elvishjerricco>
tnks: How is it broken?
<tnks>
actually, I have an idea now that I say it out-loud... cabal2nix is just a build-tool... I don't need to use the one built for ghc802.
erictapen has quit [Ping timeout: 268 seconds]
thc202 has quit [Ping timeout: 272 seconds]
<tnks>
elvishjerricco: no type-checking. it refers to self.Cabal_2_0_1_1, which doesn't even exist.
<tnks>
when I change it to Cabal_2_2_0_1, the build fails.
<elvishjerricco>
yuck
romanofskiWork has quit [Ping timeout: 252 seconds]
<tnks>
elvishjerricco: from the perspective of a library author, I'd imagine wanting to support ghc802, though as a library consumer, I personally wouldn't want to be that behind.
<__red__>
oh - apparently adding the release.nix broke nix-env locally
<__red__>
since I now have release.nix and default.nix in my actual overlay directory... any execution of nix-env -iA is failing because it now thinks that my function is a set
drakonis_ has joined #nixos
drakonis has quit [Ping timeout: 264 seconds]
<clever>
__red__: the hydra server is also a binary cache
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/fxFl9
romanofskiWork has joined #nixos
drakonis has joined #nixos
<__red__>
clever: I guess I need to move my release.nix out of the root of the overlays directory or elsewhere?
<__red__>
since the plan was to have my end-users install the overlay in their local .config/nixpkgs/overlays
<__red__>
having both a release.nix and default.nix in there is confusing the local nix-env terribly
<clever>
__red__: you could also clone it elsewhere, and then symlink foo.nix to your default.nix
<__red__>
in the release.nix I just included default.nix :-/
<__red__>
I have created a circle that eats itself
drakonis_ has quit [Ping timeout: 264 seconds]
<detran>
does anybody have any suggestions for how to do an encrypted drive? I'm going to do a reinstall because the 18.03->18.09 upgrade can't read my boot drive, and I want something more stable
<clever>
detran: i use luks on all of my installs now
<__red__>
maybe dropping this in a let block will solve that problem
Lisanna has joined #nixos
<detran>
clever: do you have a writeup anywhere? I followed this last time, and I was disappointed it doesn't handle the upgrade:
<clever>
detran: your fdisk doesnt support GPT, so the output isnt complete
<detran>
oh, do you know a tool that does?
<clever>
gdisk i think, or parted maybe
<clever>
on nixos, fdisk supports both seamlessly
<detran>
that's what I'm on, though
<clever>
ah, you have that messy setup with a cryptkey device
<clever>
thats not needed at all on modern nixos
<clever>
stage-1 will remember the passphrase you enter, and try it again on all other disks
<detran>
oh, that's good, it was a real pain to set up initially
<clever>
so if you can use that old cryptkey to add a password to both devices, you can then change the nixos config and upgrade
<clever>
and just ignore it
<clever>
i also refused to do that, and just put lvm on luks
silver_ has quit [Read error: Connection reset by peer]
<detran>
can I transition to lvm on luks with my current install?
<clever>
nope
<__red__>
reading up
<clever>
but you can easily transition to just a normal passphrase on both devices
<detran>
with cryptsetup, right? I'll have to do some reading :)
<clever>
yeah
<__red__>
oh wait, you're not using initrd
<__red__>
oh hang on, you are
<__red__>
you can add another keyslot for a text password
<__red__>
then just use boot.initrd.luks.devices."cryptroot".device = "/dev/disk/by-uuid/27e72959-dcfe-4f16-8d9e-9581f9e07e94";
<__red__>
(and lose the whole boot.initrd.luks.devices block underneath I think)
<__red__>
then duplicate for your swap
<clever>
yep
<detran>
what do you mean by keyslot?
<clever>
in the output of luksDump
<clever>
your devices have 8 keyslots
<clever>
slot 0 is used on all, and slot 1 on one of them
<clever>
each slot can be used by a passphrase or a keyfile
<clever>
and any slot can be used to unlock the disk
<detran>
okay, so I can use luksAddKey for that
<clever>
yeah
<detran>
do I need to alter the hardware-configureation.nix at all then? Can I just add the keyslots to root and swap, and provide that key when I boot?
<clever>
if a keyFile is specified in the nix, it wont ask for a password, i think
<clever>
so the keyFile=..; has to be removed
<clever>
and then the cryptkey = ...; device becomes useless, so it can be removed
lassulus_ has joined #nixos
<detran>
okay, that makes sense. Alright, that will help clear up the encrypted setup, but I think 18.09 is getting stuck finding my /boot partition, the device = "/dev/disk/by-uuid/5BAD-DF26" part
<Lisanna>
"hydra-queue-runner[70983]: nix-store: error while loading shared libraries: libsqlite3.so.0: cannot open shared object file: No such file or directory"
<Lisanna>
...does hydra use the system nix or the one that was passed in at build time?
ng0 has quit [Read error: Connection reset by peer]
ng0 has joined #nixos
Thra11 has quit [Ping timeout: 250 seconds]
<Lisanna>
oh, nvm, problem is unrelated to hydra
turion has quit [Ping timeout: 252 seconds]
lsyoyom has quit [Read error: Connection reset by peer]
pikajude has quit [Remote host closed the connection]
pikajude has joined #nixos
Tucky has joined #nixos
<jD91mZM2>
,help
<{^_^}>
Use `,` to list all commands, `,foo = Foo!` to define foo as "Foo!", `,foo =` to undefine it, `,foo` to output "Foo!", `,foo somebody` to send "Foo!" to the nick somebody
<jD91mZM2>
,
<{^_^}>
Special commands: find locate tell - Commands sorted by use count, page 0 (use ,<n> to view page <n>): library tofu dnw pr -A imperative pills profiling ping pinning unfree unstable ask declarative error escape'' overlay paste stateVersion upgrade IFD NUR allah callPackage cloudfront dentalplan escape" exec help info loot nixlang++ runtimeDeps tias timer whomademe wololo wrapper arm bootfull channels context escape-special fancy-uninstall github
<gchristensen>
anyone using networkd able to give me a walk-through on NixOS + networkd configs? or am I overcomplicating it and all I do is do networking.useNetworkd = true
<jD91mZM2>
gchristensen: Unless it bricks your system :^)
<gchristensen>
it would have to work pretty hard to brick my system
<gchristensen>
but, yes
<etu>
If you have databases etc it can be hard to rollback if changes have been made to things with a state...
<gchristensen>
that is true, etu.
<{^_^}>
[nixpkgs] @TerrorJack opened pull request #49508 → Update nodejs v10 to 10.13.0 and add nodejs v11 → https://git.io/fxFyj
<gchristensen>
speaking of which. IHave this weird problem where I accidentally marked an ancient boot option as default in systemd-boot, and I can't manage to undo that. I don't really want to be able to set a default, but the best option I have is to re-set the default on every boot after I create a new system version.
erictapen has joined #nixos
hyperfekt has quit [Ping timeout: 256 seconds]
o1lo01ol1o has quit [Ping timeout: 264 seconds]
o1lo01ol1o has joined #nixos
mayhewluke has quit [Ping timeout: 240 seconds]
mayhewluke has joined #nixos
szicari has quit [Ping timeout: 245 seconds]
bachp has joined #nixos
bennofs[m] has joined #nixos
Purple-mx has joined #nixos
Ericson2314 has joined #nixos
fgaz has joined #nixos
clacke[m] has joined #nixos
dtz has joined #nixos
cornu has joined #nixos
Mikaela- has joined #nixos
Ralith has joined #nixos
simbergm has joined #nixos
leons has joined #nixos
schmittlauch[m] has joined #nixos
ejpcmac has joined #nixos
thefloweringash has joined #nixos
tokudan[m] has joined #nixos
sphalerit has joined #nixos
roberth has joined #nixos
timclassic has joined #nixos
yangm97 has joined #nixos
koschei[m] has joined #nixos
yegortimoshenko has joined #nixos
nyanloutre[m] has joined #nixos
stammon has joined #nixos
MarkOtaris has joined #nixos
rycee has joined #nixos
ysndr_ has joined #nixos
<michas>
Hi, my nixos system currently does not boot, because stage1 does not find init. According to https://github.com/NixOS/nixpkgs/blob/master/nixos/modules/system/boot/stage-1-init.sh#L557 it tries to make sure that init either exists or is *not* a symlink. - Does that make any sense? It refers to the nixstore, which of course contains symlinks. - Where is my error in reasoning?
civodul has joined #nixos
<sphalerite>
michas: it might be an absolute path symlink, which isn't valid before the chroot
<sphalerite>
as the comment says
sigmundv has joined #nixos
<sphalerite>
michas: err, actually, it makes sure that either it *does* exist *or* it's a symlink
<sphalerite>
oh right, I see what you mean now.
<michas>
sphalerite: Yes, it *is* an absolute path to the nix store. But shouldn't that be absolutely valid? Mind the exclamation mark in the command.
<locallycompact>
I tried the nixos variant on my nixos virtualbox and it didn't error but it still seems to be building an awful lot, is the point of this cache to skip building lots?
<srhb>
Yes, it is.
<locallycompact>
hmm
<locallycompact>
it's even building bash 4.4
<gchristensen>
please share the output of `nix show-config`
<aanderse>
this verion of simgear is from 2017.3.1
<aanderse>
hmm
<aanderse>
i'm betting a version bump might fix that...
civodul has quit [Quit: ERC (IRC client for Emacs 26.1)]
clefru has joined #nixos
<srhb>
Right. Maybe I'm not being helpful. Just looked like it might help in testing :-)
justanotheruser has quit [Ping timeout: 240 seconds]
<aanderse>
ah simgear stopped received updates in 2017.4
Ridout has quit [Ping timeout: 245 seconds]
vk3wtf has joined #nixos
* Taneb
actually reads the implementation of some primops
<Taneb>
builtins.derivation does a lot more than I thought
<clever>
and its half written in nix!
<Taneb>
I know!
<infinisil>
Taneb: For example?
<clefru>
I can't execute <nixpkgs-unstable>.androidenv.android_8_0/libexec/tools/emulator. This will give "./emulator: line 5: /nix/store/[...]-android-sdk-26.1.1/libexec/tools/.emulator-wrapped: No such file or directory". The file is clearly there however. This smells like a linker error. "ldd -d"-ing the file shows no missing dependencies. "objdump -x" shows no crazy rpath's. Running that wrapped file with LD_DEBUG=all doesn't even produce a single line of
<Taneb>
infinisil: in particular, I didn't realise it wrote the .drv to the store
<clefru>
output. strace-ing shows the execve call fails right there. I am bit puzzled.
xy2_ has joined #nixos
xy2_ has quit [Remote host closed the connection]
<clever>
clefru: what about file on the binary?
<clefru>
file-ing "/nix/store/[...]-android-sdk-26.1.1/libexec/tools/.emulator-wrapped" gives ELF 64-bit LSB executable, x86-64, version 1 (SYSV), dynamically linked, interpreter /lib64/ld-linux-x86-64.so.2, for GNU/Linux 2.6.15, stripped, so I don't even have an arch mismatch
xy2_ has joined #nixos
<clever>
clefru: yep, its /lib64/ld-linux-x86-64.so.2 thats not found
<clefru>
oh indeed!
<clever>
patchelf and --set-interpreter have to be ran
<clefru>
clever: thanks, I missed that!
<Taneb>
I now know more about nix than I did earlier
<clefru>
I would have expected "ldd" to bark at me. Thanks!
<clefru>
clever++
<{^_^}>
clever's karma got increased to 31
<clever>
clefru: internally, ldd is just a bash script, that runs the right ld-linux on the binary, with some env vars set
<gchristensen>
Mic92: first a laptop, later: a server with two NICs and one bond, then:a server with three NICs, no bonds, and many VLANs
<Mic92>
gchristensen: what are using at the moment on your laptop?
<gchristensen>
networkmanager... usually wifi, sometimes ethernet over USB or Thunderbolt
<{^_^}>
[nixpkgs] @clefru opened pull request #49515 → androidsdk: emulator(-check) binaries still require patchElf. → https://git.io/fxbeR
<Synthetica>
clever: So if I understand you correctly, there is no way to do that via the configuration tab?
WilliamHamilton[ has joined #nixos
<Synthetica>
infinisil: It's not a publicly accessible hydra server, so I'm not worried about legal ramifications
<clever>
Synthetica: it has to be done in the nix files
<Mic92>
gchristensen: personally I also use networkmanger on laptop as I found it more convinient especially for wifi. You can actually use both together networkd if you want something special with networkd it was designed to coexist. For example I setup my tinc interface with networkd: https://github.com/Mic92/dotfiles/blob/master/nixos/vms/modules/retiolum.nix#L66
<Mic92>
the server sounds like a more compelling use case for networkd
<gchristensen>
Mic92: right. my goal is to try it in all my scenarios, not just compelling use cases, and see how it goes :)
<Mic92>
gchristensen: I also have a more sophisticated networkd example for dhcp/laptop, I have something in my history
<infinisil>
Mic92: So if I ever set up my own router running NixOS, you'd recommend networkd over the standard stuff?
agander has joined #nixos
<gchristensen>
if you use bonds and vlans, and you like your computer to reliably boot, and networkd works as advertised, I would :)
<Mic92>
gchristensen: I never used that. Fpletz did more in this direction I think.
<gchristensen>
aye
<gchristensen>
vlans and bonds in the `scripted` mechanism frequently results in broken boots for me
<andi->
I have used that with networkd for a while.. It brings up the bond and adds vlans.. Just not with nixos :/
patrl has quit [Ping timeout: 252 seconds]
<Mic92>
infinisil: you might still want to use a different dhcp server if you need static leases. The dhcpserver in networkd is a bit basic.
dramforever has quit [Ping timeout: 256 seconds]
<Mic92>
andi-: what do you mean by that?
patrl has joined #nixos
<andi->
It was back when I still had to deal with puppet
<Mic92>
We still broken udev rules that we ship for predictable network interfaces.
<andi->
I usually only matched via pci path or mac address since that seems to be the only thing you can really depend on.
<Mic92>
upstream systemd uses different one. This broke at least the matching rules for systemd-nspawn that are shipped with systemd.
orivej has joined #nixos
nikivi has quit [Quit: ZNC is awesome]
<Mic92>
gchristensen: I can tell that networkd at least does not suffer from race condition.
<gchristensen>
a great improvement already
<gchristensen>
samueldr: I can confirm your channel going back in time patch has been deployed, too
<samueldr>
gchristensen: more context? or just telling me you also verified?
<gchristensen>
I verified
<samueldr>
ok
<gchristensen>
afair we weren't sure if it had been deployed yet or not
<samueldr>
let's wait and never have concrete proof it worked :)
<samueldr>
in the issue eelco confirmed it wasn't and it now is
<gchristensen>
I don't think we can prove a negative like this
<gchristensen>
beyond looking at the service on the machine, reading the source of the program it calls, and comparing to see: yep, this is the patched version
<sgraf>
Hey, I'm using the stable channel of nixpkgs, but like to upgrade one particular package in my home-manager configuration (think configuration.nix). To make it conrete, it's haskellPackages.stack that we are talking about. I'd like to have version 1.9.1 instead of 1.7.1 installed. How would I do that?
<clever>
emerij: its a bool that gets passed into this derivation
<clever>
but its not clear what it does from there
<clever>
just that its a nixpkgs level option, so it has to go under nixpkgs.config
<clever>
if your using nixos
justanotheruser has joined #nixos
nuncanada has joined #nixos
patrl has quit [Ping timeout: 252 seconds]
<qyliss^work>
Does anybody know why OfBorg would be getting a different hash checksum from GitHub than I am? I've tried on two different machines, on different networks, and they all get a hash that's different to what OfBorg gets... https://github.com/NixOS/nixpkgs/pull/49158#issuecomment-434677633
detran has quit [Remote host closed the connection]
<clever>
qyliss^work: and if you temporarily make the hash invalid (just change 1 digit) what hash do you get?
<qyliss^work>
the ofborg one
<qyliss^work>
oops :D
<clever>
you gave it the hash of something else, with a similar name, and it kept using that file, because you claimed thats what the hash was
<qyliss^work>
I was using nix-prefetch-url
<clever>
nix-prefetch-url doesnt do the cleanup fetchpatch does
<clever>
so thats the hash of the raw patch, not the cleaned up patch
<qyliss^work>
oh
<qyliss^work>
is there a similar command for fetching patches?
<clever>
not really
<clever>
just give it a clearly wrong hash and read the error
<gchristensen>
,tofu
<{^_^}>
To get a sha256 hash of a new source, you can use the Trust On First Use model: use probably-wrong hash (for example: 0000000000000000000000000000000000000000000000000000) then replace it with the correct hash Nix expected.
<arianvp>
ExecStop is changed => Daemon reload => Stop is triggered, but ignogred because unit changed
<arianvp>
I literally have thousands of matches for "Current command vanished from unit file" in my journal on my desktop
Rusty1 has joined #nixos
<arianvp>
happens every time I upgrade.
<nh2>
arianvp: I just commented, there's no ExecStop in my unit file
goibhniu has quit [Ping timeout: 252 seconds]
<arianvp>
then it could be the ExecStart that is ignored, because it's changed? Anyhow, systemd doesn't seem to deal well with Units changing the Exec paths between daemon-reloads
<arianvp>
but that's problematic, as that's the way we do things in NixOS
<arianvp>
:/
nD5Xjz has joined #nixos
<arianvp>
Okay, I think I got it
<arianvp>
i'll write up a bit more in the issue
<symphorien>
arianvp: you override Execstart ?
<symphorien>
you must do: ExecStart = [ "" "new-value" ]
<arianvp>
That error message gets printed when:
patrl has joined #nixos
<symphorien>
otherwise system just appends instead of ignoring the old value
<symphorien>
*systemd
<arianvp>
I can reliably trigger the "Vanished" error if the unit file is deleted prior to daemon-reload
<arianvp>
so in your log file, you do see that consul is being terminated
<arianvp>
consul Caught signal: terminated
mayhewluke has joined #nixos
<arianvp>
nh2: could you dig up the logs of a consul server that did stop properly?
<nh2>
arianvp: no, I posted 2 log files there in the issue description. One where stuff is working correctly and got stopped, and one where it didn't get stopped at all.
<arianvp>
thanks
<arianvp>
First thing I note that we should fix... Systemd sends a SIGTERM but consul expects a SIGINT for graceful shutdown according to the docs
<arianvp>
if you look at the logs, you see consul is actually hard-killed and exits with exit-code 1 instead of 0
<arianvp>
but that has nothing to do with our problem though
Dedalo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
reinzelmann has quit [Quit: Leaving]
<arianvp>
nh2: usually switch-to-configuration logs what services it's gonna stop, start and restart
<arianvp>
does consul show up in there?
<clever>
arianvp: ive also been wanting to know, how to make it restart, not stop+start
<__red__>
clever: I solved my release.nix problem by moving the release.nix file out of the root directory of the repo - so overlays no longer parses it and my hydra configuration just references releases/release.nix instead (!)
<clever>
__red__: thats one way to fix it
<nh2>
arianvp: right, I don't really care for graceful shut down, it's totally fine for me if it exits 1 (also it's not a hard-shutdown, consul still quits by itself vs being killed)
<__red__>
Now I just need to work out how to get my users to configure my hydra server as a binary cache
<__red__>
and I'm done
<arianvp>
clever: only unit files with `X-RestartIfChanged` are restarted
<arianvp>
others are `Stopped` then `Started`
<clever>
arianvp: ahh
<arianvp>
You can also set `X-ReloadIfChanged`
<arianvp>
for for example nginx
<nh2>
arianvp: I think I did write the switch-to-configuration output into the issue description, it's the nixops output at the very bottom
<clever>
arianvp: no reload support in this service
<nh2>
so yes, consul.service shows up there
<arianvp>
clever: perhaps we should document this somewhere. I found out by ... reading source code
goibhniu has joined #nixos
fragamus has joined #nixos
<clever>
arianvp: same, but i forgot that minor detail
iyzsong has quit [Read error: Connection reset by peer]
jperras_ has quit [Quit: WeeChat 2.2]
<arianvp>
woohoo debugging perl code
<arianvp>
fu
<arianvp>
fun*
<arianvp>
we should formally verify switch-to-configuration.pl with TLA+ or something
<gchristensen>
on 18.09 the probably immediate thing to do is revert the backport and we'll release pretty soon. then diagnose what is going on, and see about a proper patch to backport.
<gchristensen>
18.09 release fairly regularly.
<ekleog>
does it? it's still on evaluations ~1 week ago right noww
<schopp0r>
gah, how can I find out the checksum of a git repo again?
<ekleog>
(after a bump from ~2/3 weeks ago in the last few days)
<gchristensen>
ekleog: the latest bump of 18.09 is from yesterday
CcxWrk has quit [Max SendQ exceeded]
<ekleog>
gchristensen: yeah, but with a commit from 1 week ago
<gchristensen>
ah I see. at any rate, we can make sure it gets through
CcxWrk has joined #nixos
szicari has joined #nixos
<arianvp>
clever: do you think there's ever a good reason to use "stop-and-start" instead of "restart" ?
<arianvp>
I can only imagine "stop-and-start" to allow for buggy situations, because the "stop" and the "start" are executed by different unit files
<clever>
arianvp: maybe when updating things in /etc/ can break a running service
<Mic92>
oh man touching our network stuff only causes headaches.
<ekleog>
Mic92: Thanks! :) do you notify the issue so they know it's not fixed in 18.09 or should I?
<schopp0r>
ok, I need some assistance … I am not sure how to do this. so in my configuration.nix, I can use overrideDerivation on android-udev-rules I think. But will this package then also be used? because I did not explicitly install it, it gets automatically installed, so I wonder whether adb.enable=true; won't still trigger the installation of the non-patched version.
CcxWrk has quit [Excess Flood]
<arianvp>
I hate mutable state
<arianvp>
headaches
<clever>
arianvp: also, when you nixos-rebuild under xen, and i tries to restart xen services, they fail in weird ways
<clever>
arianvp: and 5 minutes later, due to a lack of services in dom0, the hypervisor watchdog assumes the world is ending, and forces a hard reset
<gchristensen>
arianvp: a service's STOP action may change from revA to revB, so using revB's stop or restart directive to restart revA's service might be invalid.
<Mic92>
ekleog: I guess we add a test or though in future.
<Mic92>
s/though/so/
<arianvp>
=(
<ekleog>
Mic92: yeah that sounds good, nixos needs more tests
<{^_^}>
[systemd] @fpletz pushed 326 commits to nixos-v239: https://git.io/fxbsw
<{^_^}>
[nixpkgs] @timokau pushed 6 commits to release-18.09: https://git.io/fxbnH
<__red__>
So I'm building a gcc port for a microcontroller and nothing compiles because the .a files seem to have been munged. Is that something that buildPhase would do? maybe with patchelf? Secondary hint is this: http://localhost:3000/build/6/nixlog/1 == look at the very end of that log for a mass of "wrong ELF type" messages
<simpson>
wirew0rm: Very cool. Are you anticipating putting this into the cross-compilation subsystem in nixpkgs?
Dedalo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Dedalo has joined #nixos
<wirew0rm>
simpson: For now the toolchain is built in a very nix-unfriendly way (not modular, using crosstool-NG), but I've allready looked into it.
<wirew0rm>
Not sure if i can pull it off, that's why I first built it this way, to have some early sucess :)
nD5Xjz has joined #nixos
<kreisys>
... so I tried to change /'s fsType to "btrfs" on an ec2 instance to see what would happen... now nixos-rebuild is building llvm? does this make sense? lol
<azazel>
hi guys, what if I want to use all the functions in the derivation's build shell from inside a Makefile recipe? a Makefile run during the build process I mean
xy2_ has quit [Read error: Connection reset by peer]
<boxscapeR>
So, I have {allowUnfree = true;} in my home-manager config file, and I can use unfree programs if they're listed in the same file, but I try nix-shell -p <some unfree program>, nix still tells me that I have to put { allowUnfree = true; } into my config.nix file. And it does work if I do that. Is that a bug or am I misunderstanding how it should wo
<boxscapeR>
rk?
<clever>
boxscapeR: nix-shell only obeys the main config.nix file
<boxscapeR>
hm, okay
<clever>
so you have to set it in the file the error tells you to set it in
<boxscapeR>
So I guess if I want to set the option fully with home-manager I will have to set it twice, once for the config.nix file and once for nixpkgs.config
<boxscapeR>
hm, when I try to install steam, or use it in nix-shell, it tells me that curl gets a 404 for various .deb files. But when I try to manually run curl with those same URLs, I can download the files. Any idea what the reason for that could be?
<Lisanna>
I wish hydra showed *all* the build steps for a job, so I could get an idea of how close the job is to being done
<Lisanna>
instead of apparently generating them on the fly
detran has joined #nixos
jperras has joined #nixos
<Mic92>
gchristensen: I guess you need to hurry up if you want a rust port :)
<yl[m]>
What's the best way to make available binaries to neovim? I opened https://github.com/NixOS/nixpkgs/pull/49498/files yesterday for my attempt but I'm still unsure if this is the best way. cc clever, infinisil, Mic92.
<Mic92>
yl[m]: adding go to the PATH might be not always the best solution, because there might be different versions required depending on the projects.
<gchristensen>
Mic92: of hydra? :)
<Mic92>
gchristensen: for activation scripts.
<gchristensen>
aw
justanotheruser has quit [Read error: Connection reset by peer]
<yl[m]>
Mic92: I'm not trying to add go, but goimports, gotools and such. These are required by vim-go vim plugin. Are you suggesting that every project should provide neovim dependencies?
mayhewluke has quit [Ping timeout: 246 seconds]
<Mic92>
yl[m]: ok. this might make sense. We probably need a way to propagated packages from plugins to vim's PATH.
<Mic92>
like this was done for ruby.
<Mic92>
the current approach was to patch plugins instead.
<Mic92>
yl[m]: it could be also set in the vimrc configuration
<Mic92>
if that is supported
<yl[m]>
it might `let bin_path = go#config#BinPath()`
<gchristensen>
Mic92: just not so convinced on porting from perl -> python, really. porting is dangerous I think, and I'd rather do it as few times as possible.
<yl[m]>
I have to track down where that is in the config
mayhewluke has joined #nixos
<gchristensen>
especially when the dbus bindings are apparently actively discouraged and not great.
<Mic92>
gchristensen: To be honest I don't expect them to break.
<sphalerite>
use gdbus via python's gobject bdingins instead? :D
<symphorien>
gi is not easy to cross compile, apparently
<Mic92>
gchristensen: also if we really want, we can still shell-out busctl instead.
<Mic92>
or systemctl for that matter
<{^_^}>
[nixpkgs] @kalbasit closed pull request #49498 → neovim: take withExtraPackages and add them to the PATH → https://git.io/fxFn9
dramforever has quit [Quit: Page closed]
<gchristensen>
that does address the dbus concern but not the general danger
<gchristensen>
I would at least like @edolstra to assign someone to review it
<Mic92>
gchristensen: this sounds like FUD too me. The responsibility is still contained. The systemd update that just went through changed way more things and had less reviews.
masterdonx has joined #nixos
masterdonx has quit [K-Lined]
alex`` has quit [Ping timeout: 268 seconds]
alex`` has joined #nixos
<gchristensen>
I have been bitten by rewrites. This is the most critical juncture of nixos. I'm not saying no, I'm asking for great care
<gchristensen>
have to go
<Taneb>
I'd like it if there were more bindings to Nix the tool like the perl bindings
<chris|>
is there a way to append to an existing list in a configuration, as in eg. systemd.services.foo.wants += [ "bar.service" ] adds "bar.service" to the existing list of foo.wants ?
<clever>
chris|: all list options are append by default
<{^_^}>
[nixpkgs] @xeji pushed commit from @baracoder to master « smartgithg: 18.1.4 -> 18.1.5 (#49527) »: https://git.io/fxbPk
orivej has quit [Ping timeout: 268 seconds]
goibhniu has joined #nixos
<clever>
chris|: the only time its not, is when the service has a default value, any non-default will overwrite it
emerij has quit [Ping timeout: 268 seconds]
<clever>
chris|: but then multiple non-defaults will concat to eachother
goibhniu has quit [Read error: Connection reset by peer]
emerij has joined #nixos
<chris|>
got it
<chris|>
one more question, can I get the name of an attribute as a string, eg. "${config.systemd.service.postgresql}.service" becomes "postgresql.service" ?
<{^_^}>
[nixpkgs] @xeji merged pull request #49503 → hub: 2.5.1 -> 2.6.0, cleanup a bit → https://git.io/fxFu8
<{^_^}>
[nixpkgs] @xeji pushed commit from @dtzWill to master « hub: 2.5.1 -> 2.6.0, cleanup a bit (#49503) »: https://git.io/fxbDB
<Lisanna>
Wow, nixops can't fall back to my second binary cache if the first one has an error?
goibhniu has joined #nixos
<Lisanna>
Very convenient. Can't fix the binary cache without nixops deploy, can't do nixops deploy without the first binary cache being healthy x_x
<steveeJ>
rycee: I'm having a tricky issue. I want to share importable home-manager snippets between systems. some of the systems use the nixos module and others use the tool. the issue I'm facing is that nixos module expects all the options prefixed with "home-manager.users.<user?>" while the tool doesn't
<steveeJ>
I'd need the imports to be prefixed, can nix do that?
<rycee>
steveeJ (IRC): You can use imports.
<infinisil>
Lisanna: Whoa really..
<rycee>
steveeJ (IRC): I also use `home-manager.users.youruser = import youruser.nix;` where youruser on some other system is linked to `~/.config/nixpkgs/home.nix`
<infinisil>
Either that's a bug and nobody thought of it or it has some deeper reason we don't know yet
arjen-jonathan has joined #nixos
<steveeJ>
rycee: ah, so if I reduce all the imports in one file that would work! awesome, thanks a bunch!
<steveeJ>
rycee: I didn't think of using `import <path>.nix` vs `imports = [ <path1>.nix <path2>.nix ]` :-D
<Lisanna>
infinisil well, removing it from the substituter list makes nixops deploy work
<Lisanna>
but this seems silly
goibhniu has quit [Ping timeout: 252 seconds]
<schopp0r>
Hi. I am (still) trying to compile nix with another store-dir on a non-nixos-computer. now I have a nix-environment in the directory /tmp/blargh3 which seems to work, but now I need to compile it with a store-dir in this environment. I get the error message
<schopp0r>
error: file 'nixpkgs' was not found in the Nix search path (add it using $NIX_PATH or -I), at /home/senjak/nix-nix-2/inner/default.nix:1:13
<schopp0r>
I am not quite sure how to create such a directory. Just cloning nixpkgs from git and pointing to it will probably not suffice, right?
boxscapeR has quit [Remote host closed the connection]
<schopp0r>
infinisil, yes, but how to create the nixpkgs-directory?
<infinisil>
Clone nixpkgs
Vskilet has quit [Ping timeout: 256 seconds]
Rajsun has joined #nixos
aleph- has quit [Quit: WeeChat 2.2]
sigmundv has joined #nixos
o1lo01ol1o has joined #nixos
aswanson has joined #nixos
<yl[m]>
is there a way in nixpkgs to find if nixpkgs have a particular commit hash? In https://github.com/NixOS/nixpkgs/commit/f1987fb58f57828944ca822bbb39b3de87f01863 `colemak/en-latin9` was renamed to `colemak` and I'd like to have my configuration.nix work for both 18.09 and unstable which means I have to figure out which keyboard mapping to use
<{^_^}>
[nixpkgs] @xeji pushed commit from @fadenb to master « saxon-he: 9.8.0.6 -> 9.9.0.1 (#49541) »: https://git.io/fxbSt
<Lisanna>
is there any way on hydra, either on the web view or a JSON API, to query how many steps are remaining in a job?
<schopp0r>
omg who wrote the nixos.org page? it is UNUNSABLE with mobile internet. and above that, for example the page https://nixos.org/nixpkgs/ is USELESS. why is there no link to the github page?
<Lisanna>
schopp0r pretty much all of the interesting information about nixpkgs is in the manual, which is linked at the top
<yl[m]>
schopp0r: what's useless about the nixpkgs page?
<o1lo01ol1o>
I have a couple executables linked in /root/ from the /nix/store/ that I need to run as another user, however, I can't chmod them ("Read-only file system"). Does anyone know why?
<schopp0r>
yl[m], does it say anything really interesting?
<yl[m]>
schopp0r: like Lisanna said, all the info are in the manual (except it's not visible on the top, but in the menu accessible with the button on the top right)
<Lisanna>
but, it would probably be more useful if that page on the site just redirected directly to the github page, since the README is far more up-to-date than that static page will ever be
turion has joined #nixos
<schopp0r>
yl[m], to me it looks like one of those bullshit-buzzword-bingo-pages from oracle about java, or from m$ about whatever.
<schopp0r>
or from SAP
<yl[m]>
schopp0r: thankfully it's open-source. If you think there's more information to put there why not submit a PR?
<steveeJ>
zimbatm: how do you put rust crates (into your shell.nix) for direnv to consume?
<schopp0r>
lolwut, there is a link to github ... but only for the license. ROFL
<schopp0r>
well I'll look into it as soon as I have solid internet (I am currently in a train … in germany … which means that I only occasionally have internet \o/ )
<samueldr>
~/.config/nixpkgs/overlays/ will import according to the same rules than nix does an import AFAIK; this means that ~/.config/nixpkgs/overlays/file.nix will be loaded, and ~/.config/nixpkgs/overlays/any_directory/default.nix would
<samueldr>
https://nixos.org/nixpkgs/manual/#sec-overlays-install -> If the path is a directory, then we take the content of the directory, order it lexicographically, and attempt to interpret each as an overlay by: (1) Importing the file, if it is a .nix file. (2) Importing a top-level default.nix file, if it is a directory.
<__red__>
oh, that's awesome
<__red__>
thank you(!)
<samueldr>
you're welcome, the section is terse, though I think everything is laid out
<{^_^}>
[nixpkgs] @xeji pushed commit from @grische to master « openh264: 1.7.0 -> 1.8.0 (#49544) »: https://git.io/fxbdW
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #49552 → setup.sh: avoid running the same hook twice → https://git.io/fxbbp
drakonis_ has quit [Ping timeout: 264 seconds]
aanderse has joined #nixos
<{^_^}>
[nixpkgs] @dywedir opened pull request #49553 → subversion_1_11: init at 1.11.0 → https://git.io/fxbNa
peacememories has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
peacememories has joined #nixos
erasmas has quit [Quit: leaving]
vk3wtf has quit [Ping timeout: 264 seconds]
jD91mZM2 has quit [Quit: WeeChat 2.2]
drakonis has quit [Read error: Connection reset by peer]
drakonis has joined #nixos
toppler has joined #nixos
<toppler>
Hey. I'm getting "selector 'i686-linux' matches no derivations" trying to do "nixos-install" on a VPS. I don't think I've seen that one before. Any clues?
xy2_ has quit [Ping timeout: 268 seconds]
oida has quit [Remote host closed the connection]
vk3wtf has joined #nixos
oida has joined #nixos
EarlDeLaWarr has joined #nixos
galen has quit [Quit: WeeChat 2.2]
<{^_^}>
[nixpkgs] @manveru opened pull request #49554 → kexpand: init at 2017-05-12 → https://git.io/fxbpK
drakonis_ has joined #nixos
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<{^_^}>
[nixpkgs] @xeji pushed commit from @qolii to master « linux-hardkernel: 4.14.73-149 -> 4.14.78-150 (#49535) »: https://git.io/fxbhV
aswanson has quit [Quit: WeeChat 2.2]
aswanson has joined #nixos
peacememories has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zolk3ri has joined #nixos
judson has joined #nixos
<judson>
I've got a package I'm trying to use for development that I believe is not building because of sandboxing.
<judson>
Is there a way to get an exception for nix-shell?
<toppler>
Ah. Seem that nixos-install doesn't want to be called in a nix-shell.
<judson>
Specifically the Ruby gem tiny_tds fails during miniportile with ruby/2.3.0/net/ftp.rb:243:in `initialize': getaddrinfo: Name or service not known (SocketError)
<symphorien>
Judson: if you are a trusted user, --option sandbox false
toppler has quit [Remote host closed the connection]
o1lo01ol1o has quit [Remote host closed the connection]
<Zajcev_>
bsima: pam_pgsql I've found, it works with mysql as well?
tertl3 has joined #nixos
<bsima>
Zajcev_: no you'll have to modify the nix expression
<Zajcev_>
oh ...
<bsima>
that nix expression will install pam_pgsql, you should change it to install pam_mysql by pointing it to the correct git repo and giving it any other deps it needs
<Zajcev_>
it's out of my borders :-0
<bsima>
would be a great contribution to nixpkgs though :)
<bsima>
the PR review process is very fast and friendly, in my experience