<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/fjELx
<DigitalKiwi>
hey nh2 i've been using the nginx-binary-cache-proxy (well slighly modified) a lot and wanted to thank you
<nh2>
DigitalKiwi: awesome, great it's working for you!
<DigitalKiwi>
i don't use it with nixops or amazon because i haven't started using nixops and i don't use amazon so that's some of the modifications i made
<DigitalKiwi>
but yeah really happy with it
mexisme_ has quit [Ping timeout: 272 seconds]
shibboleth has quit [Remote host closed the connection]
toppler`` has joined #nixos
mexisme_ has joined #nixos
Glider_IRC__ has joined #nixos
Glider_IRC_ has quit [Ping timeout: 244 seconds]
goibhniu1 has quit [Remote host closed the connection]
rprije has joined #nixos
zeta_0 has joined #nixos
zeta_0 has quit [Client Quit]
Matthieu_ has joined #nixos
MmeQuignon has quit [Ping timeout: 272 seconds]
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/fjEtL
<{^_^}>
[nixpkgs] @matthewbauer pushed 2 commits to staging: https://git.io/fjEtG
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
<taktoa[c]>
arahghghghghgh someone deleted a repository where they implemented exactly the thing I'm working on now https://webcache.googleusercontent.com/search?q=cache:_zqQJOpRopAJ:https://github.com/mirrexagon/nixpkgs-esp-dev/blob/master/shell-esp32.nix+&cd=3&hl=en&ct=clnk&gl=us
arguser has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @worldofpeace pushed 3 commits to master: https://git.io/fjEtR
o1lo01ol1o has quit [Remote host closed the connection]
groodt has joined #nixos
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
<groodt>
Hi. Any Darwin experts online? I'm trying to build bazel-0.25.1 on Darwin, and I'm getting an error that I don't understand. ld: file not found: /nix/store/2936mcj9h3g3fipbqf99yj3463sp66lj-clang-7.1.0/lib/arc/libarclite_macosx.a clang-7: error: linker command failed with exit code 1 (use -v to see invocation)
<{^_^}>
[nixpkgs] @matthewbauer pushed 3 commits to master: https://git.io/fjEti
o1lo01ol1o has joined #nixos
shabius has quit [Read error: Connection reset by peer]
shabius has joined #nixos
iyzsong has quit [Ping timeout: 252 seconds]
<matthewbauer>
groodt: yeah unfortunately it looks like it needs clang 6 for some reason
<matthewbauer>
`stdenv = llvmPackages_6`
adamantium has quit [Ping timeout: 245 seconds]
dansho has quit [Quit: Leaving]
o1lo01ol1o has quit [Ping timeout: 272 seconds]
<groodt>
matthewbauer: thanks, I'll give that a try. The error I see on 0.26 is something to do with pthread. /nix/store/mffyf74zc4h2wic5lwx0kyx7gvwif0c5-Libsystem-osx-10.12.6/include/pthread/spawn.h:65:34: error: unknown type name 'qos_class_t'; did you mean 'au_class_t'?
<pie_>
taktoa[c], any idea how to figure out why python ends up eating a bunch of ram
<pie_>
c74d, ah, well, good job finding a solution
InFlames has quit [Ping timeout: 258 seconds]
Gohla has quit [Ping timeout: 246 seconds]
ajp has quit [Ping timeout: 276 seconds]
adamantium has joined #nixos
<Ashy>
so bit of a random question, has anyone here tried to get ms-sqlserver running on nixos?
rtjure has joined #nixos
Gohla has joined #nixos
<Ashy>
there's ms provided ubuntu apt packages
<adamantium>
nh2: i would be interested in your, or anyone's, nix-files regarding using WM + gnome-settings, etc. Or perhaps a way to use gnome flashback?
alphor has joined #nixos
alphor is now known as alphor_
alphor_ is now known as alphor2
<alphor2>
I've set nixpkgs in my nixPath in configuration.nix to a local checkout of nixpkgs. It respects changes, but nix instantiate --eval -E '<nixpkgs>' or nix repl thinks <nixpkgs> is a channel
<alphor2>
By it respects changes I mean that a system rebuild will pick up the local changes. But nix repl (or anything that uses <nixpkgs>, like a shell.nix won't
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
ajp has joined #nixos
alphor2 has quit [Ping timeout: 256 seconds]
liff has joined #nixos
ent has joined #nixos
datakurre has joined #nixos
selfsymmetric-mu has joined #nixos
<selfsymmetric-mu>
Hi! I have `programs.ssh.startAgent = false;` but I still have an `ssh-agent` when I rebuild, switch, and reboot. What gives? Is it possible that some other service is covertly requiring it? How can I try that down?
<selfsymmetric-mu>
My eventual goal is to replace ssh-agent with gpg-agent so that I can use a Yubikey.
<clever>
selfsymmetric-mu: `ps -eH x` and see what the parent process is
hhefesto has joined #nixos
<selfsymmetric-mu>
clever: Thank you! No parent process, looks like. Looks like it's right underneath systemd.
<colemickens>
configure gpg, start it, update the env var, won't matter whether or not ssh-agent is running somewhere
<clever>
selfsymmetric-mu: that means the parent process has since quit
<selfsymmetric-mu>
Whoa!
<selfsymmetric-mu>
I was somehow missing the ssh socket earlier.
<selfsymmetric-mu>
But I have it now.
<selfsymmetric-mu>
I'm happy but I'm also really frustrated that I didn't see what made the difference.
<clever>
selfsymmetric-mu: i'm not really sure how i setup gpg-agent, i cant find it in ~/.bashrc or anywhere else obvious, and i know i didnt turn on programs.gpg-agent, that causes even more trouble
<clever>
selfsymmetric-mu: if you turn on the gpg-agent stuff in nixos, it will auto-restart it, any time you connect over ssh
<clever>
and then the passphrase prompts open in that ssh session, not in xfce!
<selfsymmetric-mu>
gosh, what?
<selfsymmetric-mu>
That's confusing. What are the implications of that?
<selfsymmetric-mu>
What does XFCE have to do with it?
<clever>
selfsymmetric-mu: its more that the nixos options tie it into systemd/logind, and rebind to any login sessions
<clever>
selfsymmetric-mu: so the unix socket maps to wherever you last "logged in"
<clever>
if you turn it off, then it just goes to the first place you logged in, which is typically your X11 session
<clever>
the problem it causes, is that any time it wants a password, the prompt opens in a terminal in the other room
<clever>
and you dont notice it, so ssh just appears to lock up
<selfsymmetric-mu>
Oh yikes.
<selfsymmetric-mu>
Thanks for the heads up!
<selfsymmetric-mu>
I appear to be working okay for the time being. :)
<{^_^}>
clever's karma got increased to 134, colemickens's karma got increased to 8
<clever>
agent key RSA SHA256:Xn2yZ3FAQeRqcF7EDiXsPtiuptNLzEAxtOo1DRDzhDM returned incorrect signature type
<clever>
oh, and gpg-agent has been giving this error for months, any idea what the solution is?
<colemickens>
mine has recently started requiring me to physically remove and replace my yubikey, after reboots, in order for `gpg --card-status` to work, and/or for other gpg operations to work. that's been a real blast.
<selfsymmetric-mu>
No idea on my end.
<clever>
i just have normal keys in my gpg state, no hardware tokens
<{^_^}>
[nixpkgs] @khumba opened pull request #62340 → filezilla: add wrapper to set necessary FZ_DATADIR → https://git.io/fjEqU
<clever>
colemickens: ah, from reading that gpg page, i'm guessing nixos runs `UPDATESTARTUPTTY` any time you login over ssh
<clever>
colemickens: and because i run screen, that just makes things worse, it maps to the tty above screen, not inside screen
<colemickens>
I'm not sure I understand the "runs anytime you login over SSH part", since the PIN UI needs to pop wherever the key is, which remains on the client machine generally speaking. But yeah, somehow it seems that have gnome's pinentry program installed system-wide and runnign updatestartuptty is enough for "everything to just work"
<colemickens>
and then I have a gpgssh alias that forwards my gpg-socket in addition to my ssh-socket, so I can still sign remotely with my local yubikey protected key, but I only use that with trusted hosts temporarily.
<colemickens>
right, yeah, I think my connections are always outbound from my laptop so I've never really hit this. And/or now that you're pointing it out to me, I think I've hit it before and rebooted and chalked it up to a fluke.
<lukego>
Emacs question. I want to run some Elisp code at startup on my emacsWithPackages. I've tried to follow the example in the nixpkgs manual but the code doesn't seem to run. I've read the standard site-init.el and I don't see any code that would run it, either. I'm trying to understand, how is this default.el supposed to be invoked?
<gentauro>
just create a folder and add a `default.nix`. Run `nix-shell` from there and suddently I have access to some tools in that `shell` (which aren't accesible from the global profile).
<gentauro>
this is ideal for customer releated projects :o
<bahamas>
immae: I tried that and I got an error: error: anonymous function at /nix/store/35b5f2dc6s37cyc6lgvvzsaas98djck1-source/pkgs/development/haskell-modules/generic-builder.nix:13:1 called with unexpected argument 'LIBRARY_PATH', at /nix/store/35b5f2dc6s37cyc6lgvvzsaas98djck1-source/lib/customisation.nix:74:12
<immae>
bahamas: Ah so it’s not exactly a mkDerivation here
<bahamas>
immae: well, I'm setting them in `mkDerivation`
<immae>
Hmm can you show your derivation? The error seem to say that it’s not the case
<hyper_ch>
hmmm, suddenly qemu (virt manager) complains about "spiceclientgtk" missing
<immae>
it seems like your mkDerivation is actually the function defined in pkgs/development/haskell-modules/generic-builder.nix instead of the standard one
HangoverGenius has quit [Ping timeout: 246 seconds]
<immae>
maybe it has the same name
<bahamas>
immae: I'm calling it with haskellPackages.callPackage ./sparkle.nix" {}
<immae>
Ok
<immae>
Then that’s it
<bahamas>
yeah, I guess that's it
<immae>
So in your case you cannot do it as simplly as what I said, but there is a solution still
<immae>
you can do `preConfigure = ''\nexport FOO=bar\n'';`
<immae>
this is not as clean as having them in mkDerivation directly, but it will work too
<immae>
actually preSetupCompilerEnvironment seems to be dedicated to that but you have no way to provide that information as far as I see
<immae>
(probably forgotten)
<bahamas>
immae: you mean not even with `preConfigure`?
schjetne has joined #nixos
iyzsong has joined #nixos
<bahamas>
I guess, I'll try
mexisme_ has joined #nixos
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
InFlames has joined #nixos
cinimod` has quit [Ping timeout: 252 seconds]
<immae>
I mean that the derivation tries to run an optional preSetupCompilerEnvironment script, but that you have no way to provide that script as it’s not an accepted argument
<bahamas>
immae: yes, I noticed it's only referenced once in the file
rfold has joined #nixos
<immae>
if you need the environment really early postPatch is probably the earliest you can do
thc202 has joined #nixos
hoijui has quit [Ping timeout: 252 seconds]
<bahamas>
immae: it's not clear to me from the code of generic-builder.nix, if it uses cabal or straight ghc. one tells cabal or ghc where to find the native jvm library. the other tells javac where to find some Java packages
<talqu>
hi, https://pastebin.com/HRihcvy5 this is my purs.nix file. Then I have (callPackage ./purs.nix {}) expression in my environment.systemPackage. When I change the version of tar package in line 19 and the hash on line 20, I do nixos-rebuild switch. However after installing, purs --version still gives me previous version. What am i missing?
<immae>
talqu: I don’t see any errors in your process. Can you check that `which purs` is the one you think it is, and not overriden by some other configuration?
<immae>
(talqu: note that you didn’t mention changing the "version" at line 15, even though it shouldn’t make a difference)
<talqu>
I don't get it then, why it doesn't install the new version
<immae>
talqu: Hmm I don’t expect environment.systemPackage to install anything in /home/talqu/.nix-profile
<immae>
If I remember correctly, you’ll find systemPackage installs in /run/current-system/sw/bin
Jetien_ has quit [Ping timeout: 258 seconds]
<immae>
yep confirmed
<immae>
you’ll find the purs version in /run/current-system/sw/bin/purs talqu
shiver has quit [Remote host closed the connection]
<talqu>
immae: I see, thank you. Then my approach to put derivations in systemPackage in configuration.nix file is not correct? If after nixos-rebuild switch my user doesn't see the new version, it does not do what I want. How should I do this?
<immae>
talqu: no, I think it is correct. /run/current-system/sw/bin/ are systemwide packages, and you (as talqu) overrode that systemwide package to install your own
<immae>
There is nothing wrong in that
<immae>
If you want to use the system purs, then you shouldn’t have installed it as user (i.e. in /home/talqu/.nix-profile) in the first place ;)
<dweller>
hmm, not sure, but i think systemd-tmpfiles shuffleing broke creation of user dir in /run
kvda has joined #nixos
<dweller>
cause after logging in with sddm into kde, im missing /run/user/1000 and user mode systemd session is dead
<talqu>
immae: ok, i just nix-env --uninstall purs-simple (which I obviously had installed with my user) :) Now my bash sees the correct system purs. Thank you for explanation and your help again
<immae>
dweller: systemd-tmpfiles shouldn’t remove anything in /run, merely add things. Do you maybe have an aggressive configuration in that regards?
<immae>
(NB: /tmp in contrary is cleaned up at startup by systemd-tmpfiles, and it’s intended, though)
<dweller>
immae: i know, tbf i encountered that before and i couldnt find cause of such behavior
<dweller>
at least now i know its missing /run/user/<pid> durig session creation
<dweller>
uid*
<immae>
you may have encountered this kind of issue with postgresql (which has his socket files in /tmp), but if you have it with things in /run then there is an issue in your configuration, I see nothing suspicious in the default one
<immae>
I can help to dig if you want
cinimod` has joined #nixos
<immae>
(well, actually I’m looking at 19.03, it’s possible that something broke in unstable )
HangoverGenius has quit [Ping timeout: 268 seconds]
<{^_^}>
[nixpkgs] @abbradar pushed commit from @pierreis to master « folly: 2019.04.22.00 -> 2019.05.27.00 »: https://git.io/fjEOd
<immae>
dweller: you can start with `cat /etc/tmpfiles.d/* | grep /run` and check that the only thing that touches /run/user is a `d /run/user 0755 root root -` (and that nothing touches /run directly)
cinimod` has quit [Ping timeout: 245 seconds]
iyzsong has quit [Read error: Connection reset by peer]
<immae>
Hey there, is there a shorter version of `if builtins.hasAttr "some_key" some_set then some_set.some_key else default_value` (I looked in lib/attrsets.nix but couldn’t find anything related)
<immae>
Right, it was actually a module "options" and I didn’t set a default value, that’s why it wasn’t working
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<andi->
lib.hasAtteByPath sounds like what you want
kvda has joined #nixos
<andi->
*hasAttrByPath
Neo-- has joined #nixos
<immae>
andi-: a, you mean attrByPath maybe
<immae>
didn’t think of it
<andi->
Yeah, both exist. Depends on your exact use case
fendor_ has joined #nixos
<immae>
I wanted a default value. But now that I understood I had the wrong issue, my problem has changed to `if isNull some_var then default_var else some_var` :D
fendor_ is now known as fendor
toppler`` has quit [Ping timeout: 272 seconds]
<andi->
isNull is deprecated as far as I remember, just use != null
alex`` has quit [Ping timeout: 248 seconds]
o1lo01ol1o has joined #nixos
chrisabc has joined #nixos
<immae>
Right, it’s still a complicated way of saying "take the variable if it’s not null otherwise take that other default value"
civodul has joined #nixos
<jtojnar>
immae: if the variable comes from args, you can or on args
<chrisabc>
Hello Nix IRC! How can I 'explore' nixpkgs? I think I've done this before using nix repl, but don't remember the commands used. For example I'd like to determine what attributes(?) a package provides.
Glider_IRC__ has quit [Ping timeout: 248 seconds]
<tilpner>
chrisabc: Run: nix repl '<nixpkgs>'
<tilpner>
chrisabc: That's the same as running just "nix repl" and then ":l <nixpkgs>"
<ghasshee>
chrisabc: I do not know that part well. What compiler do I use when I write "hs = pkgs.haskell.packages.ghc844.override" and "hs.ghcWithPackages" in configuration.nix
iqubic has quit [Ping timeout: 252 seconds]
domogled has joined #nixos
<ghasshee>
If I was using ghc844 compiler with it, I has tried three compilers ghc822, ghc844, and ghc864.
<o1lo01ol1o>
genvulkan needs the above defined overrides, but I'm not using the correct function to add them, I think, as it says it can't find those dependencies
<o1lo01ol1o>
does anyone know what the issue might be?
<nh2>
clever: another update: I have now put (in BIOS mode) `/` (includes /boot) directly on a RAID1ed partition on the SATA disks. Still black screen on boot.
mexisme_ has quit [Ping timeout: 252 seconds]
iqubic` has left #nixos ["ERC (IRC client for Emacs 26.1)"]
<{^_^}>
[nixpkgs] @worldofpeace pushed 4 commits to master: https://git.io/fjEnk
<yorick>
my release.nix calls nixos/eval-config 250 times, which takes 32GB of ram. any way I can get this lower?
<{^_^}>
[nixpkgs] @worldofpeace pushed 3 commits to release-19.03: https://git.io/fjEnL
abathur has quit [Ping timeout: 248 seconds]
<yorick>
gchristensen: this bot is getting a bit spammy
<hyper_ch>
it just appears that way
<simpson>
We have a lot of stuff going on. (Something something *flakes*)
<yorick>
merged version update PRs aren't very interesting I think
<yorick>
simpson: any idea when flakes will be ready for rfc?
orivej has joined #nixos
<simpson>
yorick: No idea.
qqlq has joined #nixos
Lears has joined #nixos
[Leary] has quit [Ping timeout: 248 seconds]
<yorick>
a lot of people are waiting until that stage to give their comments, maybe it should happen before too much of it is implemented
<simpson>
AIUI there's a paying customer for the work, so the work will be done either way, and I have faith in the community to reject flakes if it turns out to not be good.
<fasd>
to keep my packages up to date should I be running nix-channel --update and nix-env -u or should I be using nixos-rebuild switch --upgrade?
<fasd>
and will these update packages installed with home-manager?
<immae>
fasd: nixos-rebuild will not touch your "private" (user) packages, and home-manager will only touch those
emptyflask has joined #nixos
<immae>
(well, nixos-rebuild may touch your private packages if you explicitely told so)
<fasd>
will home-manager auto-update those?
ee1943 has quit [Ping timeout: 272 seconds]
ee1943 has joined #nixos
<immae>
I don’t know home manager, it depends on what it provides. If it is linked to a given channel, then upgrading that channel will probably make home manager upgrade that
<fasd>
should I be updating the channels manually or should I let nixos-rebuild do that for me?
<philipp[m]>
I'm looking for rust crates, openssl-sys for example. I can find them in nixpkgs but not their actual path, how do I address them?
<philipp[m]>
I thought e.g. `cratesIO.crates.openssl_sys` would work, but that just returns nix errors.
<immae>
hodapp: about your question, is https://hydra.nixos.org/search?query=strings something like what you’re looking for? (it’s far from complete compared to ,locate, but who knows)
<hodapp>
that might be it
<immae>
Though given how quick locate is to answer, there is probably an indexed database somewhere, so I would ask the admins of the channel where it is
<__monty__>
Isn't there a nix-locate?
toppler`` has joined #nixos
<__monty__>
Or is nix search supposed to work like this?
<ar>
hodapp: looks surprisingly familiar on photos/videos on the internet, but with a bit more space than we have
<hodapp>
__monty__: don't know, but it would be awesome if things like this ever turned up in an Internet search
<immae>
__monty__: nix-locate would only work locally with locally built package, no?
<immae>
in this channel ,locate seems to take some hydra result somewhere
<hodapp>
ar: spaces in the US vary widely with the amount of space they have. There are spaces in Detroit that had probably 3-4 times as much space, but for a similar price, because of the economic depression in Detroit that left huge industrial space without much demand
o1lo01ol1o has joined #nixos
<__monty__>
immae: I thought it built a database out of evaluating the whole of nixpkgs, not sure how it works tbh.
o1lo01ol1o has quit [Remote host closed the connection]
<immae>
__monty__: evaluating the nixpkgs doesn’t give the files that will appear in bin, as this information is not in derivations. The only way to have a comprehensive list is to build all the packages of nixpkgs
<ar>
hodapp: yeah. being located in warsaw (the capital city) is a bit challenging for us in terms of space. we've already had to relocate twice; once because the building we were located in changed ownership (and they didn't want us there for prolonged periods of time), and once because some real estate developer bought the building and wanted to build some offices in its place
arjen-jonathan has quit [Ping timeout: 252 seconds]
o1lo01ol1o has joined #nixos
o1lo01ol_ has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
<__monty__>
immae: Mmm, yeah, that does make sense considering all the bash "mv ... $out" etc.
<immae>
Well, that’s actually the cases where you "may" guess what the output will look like :D (when it’s explicitely given in the installHook or so)
<immae>
but yes, in general you don’t, so ",locate" has other means :p
<ar>
hodapp: open/closed switch; it's in the timestamp
<hodapp>
ar: oh, youtube loves to mangle the timestamp that you link to so that it can show me a bunch of video ads
<hodapp>
that is indeed still a thing
<ar>
hodapp: we've automated this problem. we have an opt-in system that checks for dhcp leases of registered member devices, and https://hackerspace.pl/ has in the upper-left corner this green "Sorry. We're OPEN" sign which changes to red and "Sorry. We're CLOSED" when noone's here
knupfer has joined #nixos
<hodapp>
our reason for not having it be automatic is that sometimes members wanted to be there relatively undisturbed, and the open/closed switch was more meant to indicate "someone is here who is willing to let you in" than "someone is here"
<ar>
right. which is why ours is opt-in and you can easily toggle between being "visible" or not
<__monty__>
I still use &t=hh:mm:ss
<hodapp>
oh, I thought you meant opt-in per-device or something
linarcx has quit [Read error: Connection reset by peer]
linarcx has joined #nixos
siel has quit [Remote host closed the connection]
siel has joined #nixos
<ar>
hodapp it is per-device, but you can easily toggle between them being visible or not
<zeta_0>
rycee: in home.nix i enabled service.mpd, now when i turn mpc in emacs it works but it does not detect my music directory, which mpd option do i need to change to fix this?
HangoverGenius has quit [Ping timeout: 248 seconds]
HangoverGenius has joined #nixos
<zeta_0>
maybe i need to change: Default: "\$HOME/music" to "\$HOME/Music"?
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/fjEWf
zeta_0 has quit [Quit: fixing mpd]
<jb55>
iqubic: you have to do something like this to generate the nuget-packages.json: https://jb55.com/s/758f1ff055e1256c.txt since nuget can be quirky with its cache. yeah it's a bit of a shitshow at the moment
shibboleth has joined #nixos
<iqubic>
Do I need to use Nuget?
HangoverGenius has quit [Ping timeout: 248 seconds]
<jb55>
iqubic: this is for self-contained builds of dotnetcore projects. I think there is some infrastructure in nixpkgs for doing simpler things
drakonis has joined #nixos
zeta_0 has joined #nixos
<clever>
nh2: it sounds like your grub binary has memory leaks in some of its FS drivers
<iqubic>
my project is fairly simple. I don't think it uses many libraries.
<nh2>
clever: ah yes, I have figured out that error by now, it was that I had accidentally given it the lvm on my LUKS. I should edit the gist. (No idea still why the binary memory leaks in that situation.) I afterwards put `/` just on a partition directlya after the BIOS boot partition, but it still didn't work, just black screen
chrisabc has quit [Quit: Lost terminal]
<nh2>
clever: maybe if grub encounters a some LUKS stuff it does a `goto fail` and doesn't free memory correctly
<Ralith>
dhess: I find the emacs direnv mode seems to do a world of good
<clever>
nh2: grub does actually support /boot being on luks, but ive not tried using it
<zeta_0>
so in home.nix i installed mpd and left the options to the default config, now when i turn on mpc in emacs it works now, but my music is not being detected, what options do i need to change in mpd to git fix this problem?
<clever>
thats saying you currently have 2 chnanels on your user, home-manager and nixpkgs, try running `nix-channel --update` again and see if it changes anything?
o1lo01ol1o has quit [Ping timeout: 252 seconds]
<clever>
i can also see that <nixos> and <nixpkgs> (as root) will map to the channel you want, with the new version
<clever>
but home manager is likely using <nixpkgs> (as nonroot)
arjen-jonathan has joined #nixos
<drakonis>
i'm starting to think that the lustrate process might need some upgrading, because its retaining out of date user channel?
<clever>
drakonis: double-check the contents of ~/.nix-channels
<clever>
i dont think lustrate has much to do with this
<drakonis>
this hash is from when i installed nix
<drakonis>
which was over a month ago
<clever>
oh
<clever>
try deleting /home/drakonis/.nix-defexpr/channels then, just simple rm
rprije has quit [Remote host closed the connection]
rprije has joined #nixos
<drakonis>
i'll see you in twenty minutes, this is going to take some time to run
lunik1 has quit [Read error: Connection reset by peer]
<infinisil>
,ask zeta_0
<{^_^}>
zeta_0: Just ask your question. It's the best way to know if anybody can help. Or look at other ways of getting in touch with the Nix community here: https://nixos.wiki/wiki/Get_In_Touch
<Guest74468>
I've added a custom package I've made to my `environment.systemPackages`. nixos-rebuild appeared to build the package, but the executable is not found, and the rebuild also seems to warn me about this - `/nix/store... is not in the system search path.`. How would I add it to the search path?
o1lo01ol1o has quit [Ping timeout: 248 seconds]
lunik1 has joined #nixos
<etu>
Guest74468: Does your package correctly place binaries in store-path/bin/binary and that it's chmodded to +x?
o1lo01ol1o has joined #nixos
<Guest74468>
etu: Yes. Permissions are executable (555), I can use the absolute path and it'll execute.
<clever>
Guest74468: are the binaries in /run/current-system/sw/bin/ ?
shibboleth has quit [Quit: shibboleth]
lunik1 has quit [Read error: Connection reset by peer]
lunik1 has joined #nixos
<Guest74468>
etu, clever: Ohhh it is actually working. My mistake. Thanks!
<etu>
Well then :)
<clever>
its probably cabal being paranoid, checking $PATH, and complaining that $out/bin isnt in $PATH (yet)
<zeta_0>
infinisil: well, i enabled mpd in home.nix, and left the default options, now when i open mpc in emacs it works but my music is not being detected, my guess is that the capitalization of at least some of the mpd options is wrong, for example services.mpd.musicDirectory is Default: "\$HOME/music" but my music is in "/home/wolf/Music"?
<nh2>
clever: yes, though grub on LUKS isn't my goal, that was just an accident.
<nh2>
So the legacy boot didn't work at all, screen just stays black, no chance. Who knows what the hell the problem is.
<nh2>
The problem seems to be that the UUID it generates for / into hardware-configuration.nix "disappears" at the next reboot. The /dev/md125 is still there, but it now has a different UUID.
<nh2>
I managed to boot into NixOS with some manual fixup but the installer doesn't work.
<nh2>
I'm trying full UEFI boot now.
<nh2>
You'd think UUIDs staying the same is their whole purpose
<nh2>
I'm not sure why it changes
<clever>
nh2: double-check that the uuid's are correct, or use labels instead
<gchristensen>
labels are good
siel has joined #nixos
<nh2>
clever: I checked that in the OVH rescue mode, the generated path for `/` is `/dev/disk/by-uuid/645dad8a-104a-4e94-9497-5068fcb5ead5`, which indeed pointed at my intended device. But rebooting (into either NixOS or the rescue mode again) changes the UUID
<{^_^}>
[nixpkgs] @worldofpeace opened pull request #62427 → Some GNOME3 updates (again) → https://git.io/fjE8t
<clever>
nh2: have you tried addding boot.allow_shell to the kernel args, so you can get a shell in the initrd?
<o1lo01ol1o>
I'm getting a cabal configure error: This package indirectly depends on multiple versions of the same package
<o1lo01ol1o>
and then it lists some conficits. To resolve, it should be possible to update the conflicting packages to all point to the same dep and then override my derivation with the updated versions, correct?
KRMX has joined #nixos
silver has joined #nixos
Matthieu_ has quit [Ping timeout: 268 seconds]
<clever>
o1lo01ol1o: which packages is it getting doubles of?
<o1lo01ol1o>
hfmt, hlint, stylish-haskell and haskell-src-exts
<infinisil>
inquisitiv3: That, and in combination I auto-import many such files that declare options. So I can just create a new file in a special folder, declare the options I want there (like an .enable one), and use that option from anywhere, without having to worry about importing what from where
<infinisil>
inquisitiv3: This also allows you to customize modules much better, because you can declare any options you want. `imports` in comparison is really just like a single `enable` option without any customizability
Synthetica has quit [Quit: Connection closed for inactivity]
<inquisitiv3>
infinisil: Thanks for your insights! That seems sensible, and I'll look into that.
<nh2>
clever: OK some of the confusion slowly clears up. It turns out `mdadm` has a lot of logic that I was not aware of https://bugzilla.redhat.com/show_bug.cgi?id=606481#c14. I'm not sure that logic is "good" because it feels very convoluted, but at least explains the renaming of devices (not yet the UUID change).
<NYXT>
I'm trying to compile with gcc in a buildFHSUserEnv but I've found that I can only get the gcc headers when I include the gcc-unwrapped derivation, but I'd like to be able to use gcc9. Is there a way to have the gcc9 package add the headers to the FHS environment?
<inquisitiv3>
Btw, isn't `with lib;` redudant if you've `lib` in the function signature?
<nh2>
clever: useful are comments 14, 16 and 18. mdadm writes down the creation-time host name into the RAID devices. You can also override that name (which I do). During a boot, if the host name at the time mdadm's assembly command is run doesn't match what's written in there, it'll name devices as `/dev/md/old_hostname:name`.
<nh2>
Also, if a disk is configured directly as an ARRAY line in mdadm.conf, then the above logic does not apply.
<nh2>
The hostname it uses if there's no mdadm.conf is the system host name, but if it exists, it reads it as what's configured in that file.
<nh2>
this hostname approach is called `--homehost` in mdadm
<clever>
nh2: ah, that weirdness reminds me of zfs and hostid's
linarcx has quit [Quit: WeeChat 2.4]
<clever>
nh2: with zfs, when you import(/assemble/mount) a pool, it marks the hostid that is currently "using" it
Anton-Latukha has quit [Quit: Leaving.]
v0|d has joined #nixos
<clever>
nh2: at normal shutdown, that hostid is still using it, so other machines cant import it without using the force flag
<clever>
and that makes it safe to have zfs on a network block device, since you cant accidentally import a pool thats still in-use by another box
<clever>
to allow importing from the wrong hostid, you must `zpool export` to fully close it
hoijui has quit [Ping timeout: 252 seconds]
<nh2>
clever: maybe that is the worst possible answer -- after messing around with partitions+mdadm+LUKS+lvm now for another 4 days (even though I've done that stuf already many times in the last years), I had thought "no wonder" people like ZFS if it makes this a lot simpler. But if it's similar ...
zeta_0 has quit [Quit: taking a break]
<infinisil>
inquisitiv3: `with lib;` brings everything *inside* of lib.* into scope, having lib in the function args just brings lib itself into scope
<nh2>
I understand that there are probably good motivations behind all of this, but the fact that no online tutorial deals with it and tha you need to learn it from bugtracker pages that take hours to read feels very wrong
<clever>
nh2: zfs lets you clear that binding by just exporting the pool when your done, and you can ignore it with a force import
<clever>
nh2: and this just stops you from importing it while in-use, it doesnt mess with the name after the import is done
<clever>
nh2: if you dont want swap, you can also remplace the mdadm+lvm+ext4 layers all at once, with just zfs
o1lo01ol1o has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @matthewbauer pushed 2 commits to master: https://git.io/fjE4L
<nh2>
clever: it's really an unfortunate situation with the license, and the corresponding result that upstream kernels "like" to break zfs. I have the feeling just using computer with disks inside is 50x harder than it should be
<nh2>
clever gchristensen: OK I have now figured it all out. `mkfs` creates new FS UUIDs. Nothing ensures that `/dev/disk/by-uuid` learns about them. You need to manually `udevadm trigger` after your `mkfs`. Otherwise nixos-install will put old now-nonexistent UUIDs into hardware-configuration.nix
<nh2>
I didn't solve why booting via legacy BIOS showed black screen only. But with EFI it works now. Next up is figuring out how the machine should get its IP at OVH.
o1lo01ol1o has quit [Ping timeout: 272 seconds]
<clever>
nh2: that sounds like a bug in nixos-generate-config, where its using the old uuid, rather then getting the real one with blkid
<clever>
nh2: if ovh doesnt have dhcp services, youll need to staticly configure it in nixos as normal
<nh2>
clever: should blkid be able refresh automatically without `udevadm trigger`?
<clever>
blkid will just read the device you point it to, it doesnt really cache things much
<clever>
i think the cache is only to allow non-root users to query things
MinceR has quit [Remote host closed the connection]
<nh2>
clever: how does OVH networking work on our soyoustart?
selfsymmetric-mu has joined #nixos
<clever>
nh2: double-checking...
<nh2>
clever: I haven't figured out yet how the OVH rescue mode gets its IP. It uses systemd-networkd. I can't find any config for it. The logs show only
<m1crOman>
Hey folks. I'm trying to set a bunch of values in services.postfix.extraConfig and I can see making a big string with \n is getting unmaintainable fast. Is there a better way to do this, ideally loading a string value from another file with newlines "automatically" added as per the imported file?
<clever>
nh2: nothing directly setting the ip there, but you will want to read the config in the initrd too, which could be tricky to access
<nh2>
m1crOman: where do the "\n"s have to go, can you show an example of what you mean?
<clever>
m1crOman: just put regular old newlines inside a ''string'';
<clever>
m1crOman: this is a chunk of haskell code, right in a nix string
<m1crOman>
Well I might be configuring 10+ machines from a modular set of nix configs which I'll only be controlling for a year so I want to keep it as maintainable as possible.
jackdk_ has joined #nixos
<nh2>
m1crOman: note the double-single quotes (2x ') in what clever said about ''string''